**Part 2**

**Catalysis and Reaction Engineering** 

182 Advances in Chemical Engineering

Zhao, Y.; Ma, C.-Y.; Yuen, S.-N. & Phillips, D.L. (2004b). Study of acetylated food proteins by

Zhou, P.; Liu, X. & Labuza, T.P. (2008). Effects of moisture-induced whey protein

Zidar, J. & Merzel, F. (2011). Probing amyloid-beta fibril stability by increasing ionic

Zinoviadou, K.G.; Koutsoumanis, K.P. & Biliaderis, C.G. (2009). Physico-chemical properties

aggregation on protein conformation, the state of water molecules, and the microstructure and texture of high-protein-containing matrix. *J. Agric. Food Chem.*

of whey protein isolate films containing oregano oil and their antimicrobial action

raman spectroscopy. *J. Food Sci.* 69: 206-213.

strengths. *J. Phys. Chem. B* 115: 2075-2081.

against spoilage flora of fresh beef. *Meat Sci.* 82: 338-345.

56: 4535-4540.

**6** 

**Rational Asymmetric Catalyst** 

Ramzan Naveed2 and Abdullah Alqahtani1

*Saudi Basic Industries Corporation (SABIC), Riyadh*

*University of Engineering and Technology, Lahore* 

Zeeshan Nawaz1, Faisal Baksh1,

*1Chemical Technology Development,* 

*2Department of Chemical Engineering,* 

*1Kingdom of Saudi Arabia* 

*2Pakistan* 

**Design, Intensification and Modeling** 

The development of new catalysts for a variety of chemical processes, molecular-level fundamental understanding of how it works and knowledge of surface science, chemistry, materials science, process modeling, process systems engineering, etc. is needed. Catalyst design and kinetic modeling have long been based on chemical intuition, i.e., the combination of a large empirical database and qualitative concepts of chemical reaction engineering and surface science. Recently the first principle kinetic modeling has become an important tool to investigate catalytic reactions and catalyst structures for superior operational benefits by integrating first principle guided exploration and experimental data. Nanocatalysis design is the building blocks of the concept of micro-level controlled reaction engineering, and the size and shape dependant material properties are the key enabling factor of the emerging technology. The qualitative difference is infect the material properties those changes in 1-10nm scale. Drastic enhancement in capabilities in nanomaterial synthesis with increasing control in size and shape were observed in last two decades but the rational catalyst design criteria is still not mature. Consideration of reaction brings intrinsic complexity in nanomaterial topology development (for a catalyst design) and

emphasize on systematic multiscale simulation method to design and forecast.

In the field of heterogeneous catalysis, desired topological improvements on the basis of experimental information is well known and tedious [1]. Advances in surface science enable data manipulation of individual atoms on the catalyst surface with experiments provides initial guess towards systematic design. Number of model based catalyst design strategies were reported in open literature like use of qualitative reasoning and knowledge-based systems [2, 3], computational models [4-6] and detailed microkinetic modeling for catalytic systems [7]. In order to get the job done, compositional analysis of physical, chemical, and/or biological properties and validation of appropriate molecular structure (robust forward model linked with material description and knowledge extraction from experiment). Reaction modeling systems, optimization, and artificial intelligence based tools

**1. Introduction** 

### **Rational Asymmetric Catalyst Design, Intensification and Modeling**

Zeeshan Nawaz1, Faisal Baksh1,

Ramzan Naveed2 and Abdullah Alqahtani1 *1Chemical Technology Development, Saudi Basic Industries Corporation (SABIC), Riyadh 2Department of Chemical Engineering, University of Engineering and Technology, Lahore 1Kingdom of Saudi Arabia 2Pakistan* 

#### **1. Introduction**

The development of new catalysts for a variety of chemical processes, molecular-level fundamental understanding of how it works and knowledge of surface science, chemistry, materials science, process modeling, process systems engineering, etc. is needed. Catalyst design and kinetic modeling have long been based on chemical intuition, i.e., the combination of a large empirical database and qualitative concepts of chemical reaction engineering and surface science. Recently the first principle kinetic modeling has become an important tool to investigate catalytic reactions and catalyst structures for superior operational benefits by integrating first principle guided exploration and experimental data. Nanocatalysis design is the building blocks of the concept of micro-level controlled reaction engineering, and the size and shape dependant material properties are the key enabling factor of the emerging technology. The qualitative difference is infect the material properties those changes in 1-10nm scale. Drastic enhancement in capabilities in nanomaterial synthesis with increasing control in size and shape were observed in last two decades but the rational catalyst design criteria is still not mature. Consideration of reaction brings intrinsic complexity in nanomaterial topology development (for a catalyst design) and emphasize on systematic multiscale simulation method to design and forecast.

In the field of heterogeneous catalysis, desired topological improvements on the basis of experimental information is well known and tedious [1]. Advances in surface science enable data manipulation of individual atoms on the catalyst surface with experiments provides initial guess towards systematic design. Number of model based catalyst design strategies were reported in open literature like use of qualitative reasoning and knowledge-based systems [2, 3], computational models [4-6] and detailed microkinetic modeling for catalytic systems [7]. In order to get the job done, compositional analysis of physical, chemical, and/or biological properties and validation of appropriate molecular structure (robust forward model linked with material description and knowledge extraction from experiment). Reaction modeling systems, optimization, and artificial intelligence based tools

Rational Asymmetric Catalyst Design, Intensification and Modeling 187

etc. For catalyst design Katare et al. developed a predictive model for surface reactions studies and claims to be rational and robust [6]. The key steps involved in this model building are generation of the simplest plausible reaction mechanism, translation of the reaction mechanism to a computationally tractable mathematical model, solving the model to estimate the parameters in light of high throughput and/or insufficient experimental data, refining the model to better fit the data by altering the mechanism, suggesting new experiments that could help in discriminate among multiple models. These software tools

Accurate representation of the large number of elementary reactions, overall reactions

Robust compiler that fallows the generic reaction rules to generate network generator.

Reduce the number of thermodynamic and/or kinetic parameters to optimize based on

Robust solvers that can handle large number of differential-algebraic equations and

Evaluation of the multiple solutions for the parameters that explain the experimental

Pruning the reaction mechanism to get it in simpler and sophisticated.

Fig. 1. Inventory of model building process

needed state of the art features for specific purposes [6]:

thermokinetic and experimental data manipulation.

**Reaction mechanism generation and networking** 

validate parameter estimation techniques.

and species.

data also.

**Kinetic parameter estimation** 

were extensively developed in automated environment to avoid reaction engineering expertise in building robust kinetic models. These tools allow rigorous analysis of multiple reaction mechanisms in the light of experimental data. Katare et al. [6] demonstrated kinetic model development for propane aromatization on H-ZSM-5 zeolite catalyst with a proof-ofconcept in his studies.

Similarly, organometalic catalyst has interests in the broad areas of homogeneous catalysis, and catalyst modeling and mechanism elucidation. The main thrust of the field is in catalyst design and investigation of mechanistic aspects includes organic synthetic, organometallic and inorganic chemistry, molecular modeling and reaction monitoring. Fundamental research provide a theoretical basis and looks at the catalyst modeling using both experimental and theoretical techniques, providing information on active species and likely transition states. Ultimately helps in designing real experimental systems to be modeled with reliability and confidence. Mechanistic investigations using spectroscopic and kinetic methods help elucidate catalytic pathways and possible catalyst decomposition routes. Assessment and optimization of catalyst performance factors effecting catalyst behavior are important. The study of heterocyclic carbenes and synthesis of mixed carbene-donor ligands (transition metal complexes) were under extensive consideration. The structure, stoichiometric reaction behaviour and catalytic potential of the new complexes are the key goals to elaborate ways of manipulating the chemistry of the complexes in catalysis. The synthetic work is coupled with theoretical studies on orbital behavior and effects on the energetics of decomposition reactions. Synthesis of new polydentate ligands and their transition metal complexes offer prospects for the formation of new alkene polymerisation, copolymerisation and oligomerisation catalysts. Ligands with phosphorus, nitrogen and/or oxygen donors are the main classes of ligands and their selection depends on the metal being investigated. Therefore, methodology for catalyst discovery targets suitable strategy based on the conceptual tools of surface mechanism, molecular, bio and solid state chemistry.

#### **2. Reaction modeling suite**

The model building procedure needs rapid screening of reaction mechanistic hypothesis that could explain the experimental data. Number of user driven tools were available those facilitates knowledge archiving and retrieval to an automated reaction kinetic modeling. These modeling suites develop reaction mechanism and directly construct kinetic model. The architectural philosophy behind cumbersome mathematical modeling is shown in figure 1. Modelling can provide reaction energies (ΔHr), energy barriers (ΔEa), location of intermediates, etc. and it can explore catalyst composition, poisons, promoter's effect, efficiencies of catalysts using alternative routes, etc. In homogeneous catalyst design control of activity, ligand type, concentration of co-catalyst, molecular weight, tacticity (depends upon equilibrium), etc are also important.

The automated kinetic models development software tools of reaction networking, parameter optimization and overall kinetic modeling etc., are available in bulk like Reactor, Reaction Modeling Suite (RMS), KINAL A, Aspen Custom Modeler, gPROMS, Presto Kinetics, GREGPAK, Matlab, MLAB, ParaMetra, Scientist, Eurokin, polymath, Chemkin, Mitsubishi, MKM, IBM CKS, NetGen, XMG, Dynetica, Forcite, COMPASS, Sorption, CASTEP, DMol3, Gepasi, Athena Visual Studio parameter estimation tool, E-cell, DBsolve,

were extensively developed in automated environment to avoid reaction engineering expertise in building robust kinetic models. These tools allow rigorous analysis of multiple reaction mechanisms in the light of experimental data. Katare et al. [6] demonstrated kinetic model development for propane aromatization on H-ZSM-5 zeolite catalyst with a proof-of-

Similarly, organometalic catalyst has interests in the broad areas of homogeneous catalysis, and catalyst modeling and mechanism elucidation. The main thrust of the field is in catalyst design and investigation of mechanistic aspects includes organic synthetic, organometallic and inorganic chemistry, molecular modeling and reaction monitoring. Fundamental research provide a theoretical basis and looks at the catalyst modeling using both experimental and theoretical techniques, providing information on active species and likely transition states. Ultimately helps in designing real experimental systems to be modeled with reliability and confidence. Mechanistic investigations using spectroscopic and kinetic methods help elucidate catalytic pathways and possible catalyst decomposition routes. Assessment and optimization of catalyst performance factors effecting catalyst behavior are important. The study of heterocyclic carbenes and synthesis of mixed carbene-donor ligands (transition metal complexes) were under extensive consideration. The structure, stoichiometric reaction behaviour and catalytic potential of the new complexes are the key goals to elaborate ways of manipulating the chemistry of the complexes in catalysis. The synthetic work is coupled with theoretical studies on orbital behavior and effects on the energetics of decomposition reactions. Synthesis of new polydentate ligands and their transition metal complexes offer prospects for the formation of new alkene polymerisation, copolymerisation and oligomerisation catalysts. Ligands with phosphorus, nitrogen and/or oxygen donors are the main classes of ligands and their selection depends on the metal being investigated. Therefore, methodology for catalyst discovery targets suitable strategy based on the conceptual tools of surface mechanism, molecular, bio and solid state

The model building procedure needs rapid screening of reaction mechanistic hypothesis that could explain the experimental data. Number of user driven tools were available those facilitates knowledge archiving and retrieval to an automated reaction kinetic modeling. These modeling suites develop reaction mechanism and directly construct kinetic model. The architectural philosophy behind cumbersome mathematical modeling is shown in figure 1. Modelling can provide reaction energies (ΔHr), energy barriers (ΔEa), location of intermediates, etc. and it can explore catalyst composition, poisons, promoter's effect, efficiencies of catalysts using alternative routes, etc. In homogeneous catalyst design control of activity, ligand type, concentration of co-catalyst, molecular weight, tacticity (depends

The automated kinetic models development software tools of reaction networking, parameter optimization and overall kinetic modeling etc., are available in bulk like Reactor, Reaction Modeling Suite (RMS), KINAL A, Aspen Custom Modeler, gPROMS, Presto Kinetics, GREGPAK, Matlab, MLAB, ParaMetra, Scientist, Eurokin, polymath, Chemkin, Mitsubishi, MKM, IBM CKS, NetGen, XMG, Dynetica, Forcite, COMPASS, Sorption, CASTEP, DMol3, Gepasi, Athena Visual Studio parameter estimation tool, E-cell, DBsolve,

concept in his studies.

chemistry.

**2. Reaction modeling suite** 

upon equilibrium), etc are also important.

Fig. 1. Inventory of model building process

etc. For catalyst design Katare et al. developed a predictive model for surface reactions studies and claims to be rational and robust [6]. The key steps involved in this model building are generation of the simplest plausible reaction mechanism, translation of the reaction mechanism to a computationally tractable mathematical model, solving the model to estimate the parameters in light of high throughput and/or insufficient experimental data, refining the model to better fit the data by altering the mechanism, suggesting new experiments that could help in discriminate among multiple models. These software tools needed state of the art features for specific purposes [6]:

#### **Reaction mechanism generation and networking**


#### **Kinetic parameter estimation**


Rational Asymmetric Catalyst Design, Intensification and Modeling 189

∆S is entropy change of the reaction, respectively. For a linearly independent set of reactions network, if L is the size of reaction mechanism (unique in size) having I number of reactions, then L ≤ I. The linear dependent reactions constraints can be expressed as cij, those are the

Computer aided catalyst design and intensification is an integrated approach in catalysis research and development, as chemical engineering is currently re-emerging under the label process intensification. The concept of process intensification is catalysis is new and not well established in the catalyst design industry. It is particularly well suited for the design and development of new catalyst. The incorporation of appropriately designed micro-structured catalytic having good control over the surface chemistry of the reaction is presented. Modelling of reactions at the active sites of catalysts also performed some times using

Reaction Modeling Suite (RMS) was developed by Purdue University, USA [6]. The quantitative kinetic description of a reaction mechanism involves specification of a differential and algebraic equation for each reaction component. RMS can develop reaction mechanism that typically involves more than 10 differential equations for a simple reaction and for complex mechanisms it can be 100 or more differential equations. Each set of differential equations includes a number of kinetic parameters that must be regressed from experimental data. It is a tedious, time-consuming task to develop code and solve the appropriate set of differential equation, where there is a real possibility of mistakes due to the complexity. The key steps involved in this process are reaction mechanism form chemistry rules, conversion of mechanism to a tractable mathematical model in computer language, integration of differential/algebraic set of equations with the suitable initial conditions, optimum model parameters estimation and finally analysis of reliability of the

Each of the above steps requires considerable expertise, time and effort. Moreover, there are typically a number of potential kinetic models that need to be evaluated, and the model building/solving tasks must then be repeated for each new physical model. RMS presents the appropriate tools so that chemical experts can focus on chemistry without be overly burdened with the implementation of the required numerical methods. Purdue University, has successfully developed a set of systems tools, called the Reaction Modeling Suite (RMS) for rapid generation of complex kinetic models and evaluates these models with large, diverse sets of experimental data. These system tools are essential if model building and evaluation are to keep a pace commensurate with the rate of data production from high throughput experimentation [6]. In RMS, Chemistry Compiler takes a set of chemistry rules (i.e. any kinetic model is defined by a set of reaction rules), ensures that the rule set is chemically consistent, and finally generates the appropriate kinetic expressions. Next, an Equation Generator takes the kinetic expressions and develops the appropriate differential and algebraic equations in a form that can be solved numerically. Parameter Optimization is the next step, where the best set of parameters for a given model needed to describe the experimental data was determined i.e. non-trival for the large, nonlinear, coupled sets of

coefficients of i-th linearly decomposed reaction onto the reaction basis.

density functional theory, and we will not discuss this here.

**4. Computer aided catalyst design** 

**5. Reaction modeling suite** 

model parameters.

#### **Results and forecasting**


The most critical and important task in building a robust model is development of kinetic model and its also the most-time-consuming (excepting experimental or pilotplant campaigns). Without a satisfactory reaction mechanism and kinetics, a model may be applicable to narrow ranges of conditions and dangerous to use for predictions. The open literature information may not exactly match your conditions or catalyst formulation, therefore, formulation or selection of reaction-mechanism and rate expressions is best suggested. Establishing the reaction kinetics involves selection of rate expressions, and determination of rate parameters. Rate expressions involves four parameters; preexponential factor, activation energies, reaction order to each component and adsorption constants. The pre-exponential factors must be determined from the actual experimental data for catalyst or reaction conditions. This chapter give pathway to model information for catalyst design and intensification from state of the art reaction kinetics data generation, to the developing of computer-based knowledge organization for catalyst development, with brief review said techniques and models.

#### **3. Thermodynamical consistency**

In mechanism development and rate parameter optimization, thermodynamic consistency at both the enthalpic and entropic levels was often overlooked. An inconsistency in enthalpies gives incorrect solutions to the energy conservation equation, which lead towards ambiguous predictions of heat exchange, conversion, selectivity, etc.; while entropic inconsistency explains fundamental inconsistency (i.e. the pre-exponential factors). As there is no specific criterion for thermodynamic consistency evaluation and it distorts underlying equilibrium constant, only few published mechanisms are thermodynamically consistent [9- 12]. In general, for any i-th reaction in a, the following equations form the basis of the enthalpic and entropic constraints [13]. For thermodynamically consistent mechanism the ratio of equilibrium constants (determined by gas phase thermochemistry) should be closer to 1 and for rigorous account of temperature variations data its necessary.

$$\begin{aligned} \Delta H\_i &= E\_i^f - E\_i^b = \sum\_{j=1}^L \mathbf{c}\_{ij} \left( E\_j^f - E\_j^f \right), \qquad \mathbf{i} = L + \mathbf{1}\_i I \\\\ A\_i^f \ne A\_i^b &= e^{\Delta S\_i / R} \\\\ \frac{\Delta S\_i}{R} &= \ln(A\_i^f \ne A\_i^b) = \ln \prod\_{j=1}^L (A\_j^f \ne A\_j^b)^{c\_{\bar{\eta}}}, \quad \mathbf{i} = L + \mathbf{1}\_i I\_i \end{aligned}$$

where f stands for the forward, b for backward reaction steps, A is the pre-exponential factor, E is the activation energy, R is the universal gas constant, ∆H is enthalpy change and ∆S is entropy change of the reaction, respectively. For a linearly independent set of reactions network, if L is the size of reaction mechanism (unique in size) having I number of reactions, then L ≤ I. The linear dependent reactions constraints can be expressed as cij, those are the coefficients of i-th linearly decomposed reaction onto the reaction basis.

#### **4. Computer aided catalyst design**

188 Advances in Chemical Engineering

 Discrepancies identification and optimization between the key features of model and experimental data that leads to design of experiment and questions on generated

Explain robustness of the model development with boundaries and shows statical

The most critical and important task in building a robust model is development of kinetic model and its also the most-time-consuming (excepting experimental or pilotplant campaigns). Without a satisfactory reaction mechanism and kinetics, a model may be applicable to narrow ranges of conditions and dangerous to use for predictions. The open literature information may not exactly match your conditions or catalyst formulation, therefore, formulation or selection of reaction-mechanism and rate expressions is best suggested. Establishing the reaction kinetics involves selection of rate expressions, and determination of rate parameters. Rate expressions involves four parameters; preexponential factor, activation energies, reaction order to each component and adsorption constants. The pre-exponential factors must be determined from the actual experimental data for catalyst or reaction conditions. This chapter give pathway to model information for catalyst design and intensification from state of the art reaction kinetics data generation, to the developing of computer-based knowledge organization for catalyst development, with

In mechanism development and rate parameter optimization, thermodynamic consistency at both the enthalpic and entropic levels was often overlooked. An inconsistency in enthalpies gives incorrect solutions to the energy conservation equation, which lead towards ambiguous predictions of heat exchange, conversion, selectivity, etc.; while entropic inconsistency explains fundamental inconsistency (i.e. the pre-exponential factors). As there is no specific criterion for thermodynamic consistency evaluation and it distorts underlying equilibrium constant, only few published mechanisms are thermodynamically consistent [9- 12]. In general, for any i-th reaction in a, the following equations form the basis of the enthalpic and entropic constraints [13]. For thermodynamically consistent mechanism the ratio of equilibrium constants (determined by gas phase thermochemistry) should be closer

*H E E cE E iL I*

/ / *<sup>i</sup> <sup>f</sup> <sup>b</sup> S R AAe <sup>i</sup> <sup>i</sup>* 

1 ln( / ) ln ( / ) , 1, , *ij*

where f stands for the forward, b for backward reaction steps, A is the pre-exponential factor, E is the activation energy, R is the universal gas constant, ∆H is enthalpy change and

*L f f b b c i i j i j j <sup>S</sup> A A A A iL I*

, 1,

to 1 and for rigorous account of temperature variations data its necessary.

*R*

1

 

*L f b f f i i i ij j j j*

**Results and forecasting** 

mechanism.

brief review said techniques and models.

**3. Thermodynamical consistency** 

analysis.

Computer aided catalyst design and intensification is an integrated approach in catalysis research and development, as chemical engineering is currently re-emerging under the label process intensification. The concept of process intensification is catalysis is new and not well established in the catalyst design industry. It is particularly well suited for the design and development of new catalyst. The incorporation of appropriately designed micro-structured catalytic having good control over the surface chemistry of the reaction is presented. Modelling of reactions at the active sites of catalysts also performed some times using density functional theory, and we will not discuss this here.

#### **5. Reaction modeling suite**

Reaction Modeling Suite (RMS) was developed by Purdue University, USA [6]. The quantitative kinetic description of a reaction mechanism involves specification of a differential and algebraic equation for each reaction component. RMS can develop reaction mechanism that typically involves more than 10 differential equations for a simple reaction and for complex mechanisms it can be 100 or more differential equations. Each set of differential equations includes a number of kinetic parameters that must be regressed from experimental data. It is a tedious, time-consuming task to develop code and solve the appropriate set of differential equation, where there is a real possibility of mistakes due to the complexity. The key steps involved in this process are reaction mechanism form chemistry rules, conversion of mechanism to a tractable mathematical model in computer language, integration of differential/algebraic set of equations with the suitable initial conditions, optimum model parameters estimation and finally analysis of reliability of the model parameters.

Each of the above steps requires considerable expertise, time and effort. Moreover, there are typically a number of potential kinetic models that need to be evaluated, and the model building/solving tasks must then be repeated for each new physical model. RMS presents the appropriate tools so that chemical experts can focus on chemistry without be overly burdened with the implementation of the required numerical methods. Purdue University, has successfully developed a set of systems tools, called the Reaction Modeling Suite (RMS) for rapid generation of complex kinetic models and evaluates these models with large, diverse sets of experimental data. These system tools are essential if model building and evaluation are to keep a pace commensurate with the rate of data production from high throughput experimentation [6]. In RMS, Chemistry Compiler takes a set of chemistry rules (i.e. any kinetic model is defined by a set of reaction rules), ensures that the rule set is chemically consistent, and finally generates the appropriate kinetic expressions. Next, an Equation Generator takes the kinetic expressions and develops the appropriate differential and algebraic equations in a form that can be solved numerically. Parameter Optimization is the next step, where the best set of parameters for a given model needed to describe the experimental data was determined i.e. non-trival for the large, nonlinear, coupled sets of

Rational Asymmetric Catalyst Design, Intensification and Modeling 191

Catalyst deactivation is defined as the steady degradation of the performance of the catalyst over time. Deactivation can be modeled by describing a deactivation correlation which relates the catalyst performance to its age. The deactivation factor varies from 1.0 (fresh catalyst) to 0.0 (fully depleted catalyst). If catalyst deactivation is significant it is important to consider its effect when optimising the catalyst properties. Properties which are optimal when the catalyst is fresh may become highly non-optimal by the time catalyst has degraded. By considering catalyst deactivation during optimization it is possible to design a catalyst that, while non-optimal initially, becomes optimal when its performance is averaged out over its lifespan. An additional factor which may be considered is the possibility of modifying the reactor temperature over time, to help minimize the impact of the catalyst

Center of Process Integration, University of Manchester, United Kingdom provides a unique software suite have all above features named "REACTOR". Figure 2, 3 and 4 shows its user interface for optimizing catalyst properties. It also develops rate expressions and optimized the reactor at desired operating conditions. Heterogeneous catalyst's fluid-particle interface

deactivation.

is modeled in mass and energy balance below

Fig. 2. Catalyst properties options in REACTOR

DAEs associated with complex kinetic mechanisms. Finally, a Statistical Analyzer has been developed to quantitatively assess the reliability of the model parameters with the ability to use nonlinear statistics. The RMS tools enable the nearly automatic analysis of multiple reaction mechanism and kinetics that assist catalyst design.

### **6. Catalyst intensification and optimization**

The following is an overview of the catalyst modeling and optimization features necessary in developing attractive suite. Catalysts can be modeled in several different ways, with increasing levels of complexity by considering pellet balances, pellet factor effectiveness factor using is basic activity profiles, etc. With out pellet balances optimization is a very basic modelling level, it is not necessary to make any extra calculations due to the presence of a catalyst by expressing reaction kinetics in terms of composition of the catalyst. The catalyst composition can then be made an optimization variable. In the case of plug flow reactors, the composition can be varied along the length of the reactor, and viewed using the catalyst profile. Pellet factor calculation involves balances between component concentrations in the bulk, and within the catalyst pellet. The activity distribution inside the pellet is represented by a "Dirac-delta" function (infinite concentration at a single location). Although this cannot be achieved in practice, the Dirac-delta function can be closely approximated by a narrow step distribution of active material around the optimal location of the delta function without noticeable loss in performance.

Catalyst effectiveness factor calculations may involve Thiele modulus and effectiveness factor for the pellet. The effectiveness factor is a function of temperature, pressure and catalyst properties such as shape, size and diffusivity of components. When using this method the catalyst properties become optimizable parameters, and optimized profiles can be obtained. This method is sub-divided into basic activity profile as represented below:


The activity profiles are represented as infinitely variable profiles and by definition, when using the effectiveness factor method, the intrapellet mass balances are solved to determine the conentration profile throughout the catalyst pellet. There are also three further levels of complexity which can be selected to improve the model at the expense of computational time:


DAEs associated with complex kinetic mechanisms. Finally, a Statistical Analyzer has been developed to quantitatively assess the reliability of the model parameters with the ability to use nonlinear statistics. The RMS tools enable the nearly automatic analysis of multiple

The following is an overview of the catalyst modeling and optimization features necessary in developing attractive suite. Catalysts can be modeled in several different ways, with increasing levels of complexity by considering pellet balances, pellet factor effectiveness factor using is basic activity profiles, etc. With out pellet balances optimization is a very basic modelling level, it is not necessary to make any extra calculations due to the presence of a catalyst by expressing reaction kinetics in terms of composition of the catalyst. The catalyst composition can then be made an optimization variable. In the case of plug flow reactors, the composition can be varied along the length of the reactor, and viewed using the catalyst profile. Pellet factor calculation involves balances between component concentrations in the bulk, and within the catalyst pellet. The activity distribution inside the pellet is represented by a "Dirac-delta" function (infinite concentration at a single location). Although this cannot be achieved in practice, the Dirac-delta function can be closely approximated by a narrow step distribution of active material around the optimal location

Catalyst effectiveness factor calculations may involve Thiele modulus and effectiveness factor for the pellet. The effectiveness factor is a function of temperature, pressure and catalyst properties such as shape, size and diffusivity of components. When using this method the catalyst properties become optimizable parameters, and optimized profiles can be obtained. This method is sub-divided into basic activity profile as represented below:

Egg-shell - The active material is concentrated at the pellet surface, and rarefied

Egg-yolk - The active material is concentrated at the pellet centre, and rarefied towards

Peaked - The active material is concentrated at a location partway between the centre

 Layered - The active material is distributed in a layer of thickness T, and location L from the pellet centre, where T and L are expressed in fractions of the pellet diameter.

The activity profiles are represented as infinitely variable profiles and by definition, when using the effectiveness factor method, the intrapellet mass balances are solved to determine the conentration profile throughout the catalyst pellet. There are also three further levels of complexity which can be selected to improve the model at the expense of computational

Interphase mass balances causes the concentration profile between the pellet surface

Interphase heat balances calculates the temperature profile between the pellet surface

Intrapellet heat balances calculates the temperature profile within the catalyst pellet

Uniform - The active material is uniformly distributed through the pellet.

and surface, and rarefied towards the centre and surface.

and the bulk material to be calculated.

and the bulk material

reaction mechanism and kinetics that assist catalyst design.

of the delta function without noticeable loss in performance.

towards the centre.

the surface.

time:

**6. Catalyst intensification and optimization** 

Catalyst deactivation is defined as the steady degradation of the performance of the catalyst over time. Deactivation can be modeled by describing a deactivation correlation which relates the catalyst performance to its age. The deactivation factor varies from 1.0 (fresh catalyst) to 0.0 (fully depleted catalyst). If catalyst deactivation is significant it is important to consider its effect when optimising the catalyst properties. Properties which are optimal when the catalyst is fresh may become highly non-optimal by the time catalyst has degraded. By considering catalyst deactivation during optimization it is possible to design a catalyst that, while non-optimal initially, becomes optimal when its performance is averaged out over its lifespan. An additional factor which may be considered is the possibility of modifying the reactor temperature over time, to help minimize the impact of the catalyst deactivation.

Center of Process Integration, University of Manchester, United Kingdom provides a unique software suite have all above features named "REACTOR". Figure 2, 3 and 4 shows its user interface for optimizing catalyst properties. It also develops rate expressions and optimized the reactor at desired operating conditions. Heterogeneous catalyst's fluid-particle interface is modeled in mass and energy balance below


Fig. 2. Catalyst properties options in REACTOR

Rational Asymmetric Catalyst Design, Intensification and Modeling 193

Fig. 4. Catalyst deactivation modeling window in REACTOR

Microchannel has distinct identical channels and shape with characteristic dimensions in the order of micrometers and surface area-to-volume ratios ca. 50-100 times higher than those of their conventional counterparts. Due to the presence of high surface areas, submillimeter dimensions and metallic substrates usage, heat can be distributed quickly and uniformly distributed along the channel walls of the catalyst. High heat integration opertunities are available to drive endothermic and exothermic reactions independently and/or in parallel. The representative units can be modeled by incorporating twodimensional continuity, momentum, energy conservation and species mass transport equations for the fluid and catalytic wall phase, and energy conservation equation for the metallic/catalytic wall phase. Simultaneous solution of these equations may carry out using finite elements methods. The model is then used to figure out the effects of various configurationally parameters such as wall thickness, type of wall material and the presence of micro-baffles on temperature distribution. The results forecast the parameters

**7. Microchannel design and modeling** 

Mass balance:

$$k\_{g,i} \left(\mathbf{C}\_{s,i} - \mathbf{C}\_{j,i}\right) = -D\_{e,i} \frac{d\mathbf{C}\_i}{dz} = \eta \sum \left(V\_{i,j} - R\_{s,j}\right) \rho \eta \mathbf{S}^\circ \tag{1}$$

Energy balance:

$$
\Delta h \left( T\_s - T\_g \right) = -\lambda \frac{dT}{dz} = \eta \sum \left( \Delta H\_j - R\_{s,j} \right) \rho \eta S' \tag{2}
$$


Fig. 3. Catalyst physical properties options in REACTOR

 , ,, , , , , *<sup>i</sup> gi si ji ei ij sj dC kC C D V R pS dz* 

 , , *s g* Δ *j sj dT hT T H R pS dz* 

Fig. 3. Catalyst physical properties options in REACTOR

 

 

(1)

(2)

Mass balance:

Energy balance:


Fig. 4. Catalyst deactivation modeling window in REACTOR

#### **7. Microchannel design and modeling**

Microchannel has distinct identical channels and shape with characteristic dimensions in the order of micrometers and surface area-to-volume ratios ca. 50-100 times higher than those of their conventional counterparts. Due to the presence of high surface areas, submillimeter dimensions and metallic substrates usage, heat can be distributed quickly and uniformly distributed along the channel walls of the catalyst. High heat integration opertunities are available to drive endothermic and exothermic reactions independently and/or in parallel. The representative units can be modeled by incorporating twodimensional continuity, momentum, energy conservation and species mass transport equations for the fluid and catalytic wall phase, and energy conservation equation for the metallic/catalytic wall phase. Simultaneous solution of these equations may carry out using finite elements methods. The model is then used to figure out the effects of various configurationally parameters such as wall thickness, type of wall material and the presence of micro-baffles on temperature distribution. The results forecast the parameters

Rational Asymmetric Catalyst Design, Intensification and Modeling 195

The suitable model can account for heterogeneity in terms of chemisorption by introducing a sufficient number of different adsorption states. The readsorption formulated and tested against experimental data by describing the intrinsic dynamics of an adsorption state as a quasiequilibrium adsorption/desorption between the gas phase and the surface. Conceptually it is potentially useful for the elucidation of fundamental reaction mechanistic information for design and intensification of catalyst. The transient response methods place similar requirements on experimentation as conventional methods of kinetic analysis. In slow transient input, all or part of the elementary reactions may proceed in quasiequilibrium and their individual rates then can not properly elucidated. In general this is based on nonlinear regression analysis, and omits

In our extensive studies of catalyst intensification we conduct series of kinetic analysis for TPx experiments in phenomenological way [15-17]. Modelling of TPD data is based on adsorption and desorption theories, while TPR and TPO call for gas-solid surface reaction mechanisms, which may include topochemical characteristics. Further the rate of surface reaction divided into intrinsic rate of reaction per area of reaction interface and change in the reaction interface in reaction [18]. In general kinetic analysis of thermoanalytical data where there is a linear temperature rise (β) represented in the rate equation, where, rate of reaction is proportional to Arrhenius temperature dependence, A is pre-exponential factor,

*d E Aexp f dT RT*

( ) *d E ln ln Af dT RT*

The plot of ln(βdα/dT) versus 1/T results straight line at a selected conversion level, the slope of the line gives -E/R at certain conversion level. This also did not explain reaction mechanism but by varing activation energy E=E(α) results from the Friedman analysis explains that eq. 3 with a normal physico-chemical interpretation will not describes the

**9. Design and intensification of Pt-Sn-based alkane dehydrogenation catalyst**  The control of chemical reaction pathways at molecular level presents undoubtedly the most important scientific challenge on the way to fully sustainable, thermodynamically efficient chemical processes. Integrated molecular reaction control is one of the major breakthroughs in achieving sustainable and efficient chemical processes design. Numerous fundamental works on this topic have been reported in literature and it is now the key challenge to advance them in an interdisciplinary way towards the industrial catalysis. The molecular alignment and geometry of collisions as well as selectivity of desired product was tackled by catalyst topology. In this example we discusses Pt-Sn based catalyst design and intensification for light alkane dehydrogenation to light alkenes. Constraints of this reaction

  (3)

(4)

E is activation energy, and these are function of degree of conversion, f(α).

The transforming eq. 3 by taking natural logarithms on both sides gives

are very complex due to endothermic limitation and endothermicity.

 

> 

assumptions of pseudo-steady state.

reaction kinetics.

affects the degree of heat flow between channels and provide useful insights for the design of microchannels/catalysts.

#### **8. Thermoanalytical technique**

Thermoanalytical techniques were considered as transient response techniques, those relate characteristic properties of solid catalysts to its temperature programmed heating. Exchanges of matter and/or energy between and/or on surfaces of the solid catalysts provide means to follow physical or chemical transformations. The response obtained against certain temperature (thermogram) reflects the nature of the system and suitable experimental conditions. This analysis is used as a tool for quantitative and qualitative analysis for evaluating the influence of different activity factors. The most common thermoanalytical techniques are listed in Table 1. Number of methods exists in the open literature for extraction of kinetics information from thermoanalytical data [14]. Particularly in the heterogeneous catalysis, these techniques are powerful tool for investigating the influence of composition, preparation method and pretreatment on the reactivity of the surface or bulk with gases.


Table 1. Thermal analysis techniques and their uses

affects the degree of heat flow between channels and provide useful insights for the

Thermoanalytical techniques were considered as transient response techniques, those relate characteristic properties of solid catalysts to its temperature programmed heating. Exchanges of matter and/or energy between and/or on surfaces of the solid catalysts provide means to follow physical or chemical transformations. The response obtained against certain temperature (thermogram) reflects the nature of the system and suitable experimental conditions. This analysis is used as a tool for quantitative and qualitative analysis for evaluating the influence of different activity factors. The most common thermoanalytical techniques are listed in Table 1. Number of methods exists in the open literature for extraction of kinetics information from thermoanalytical data [14]. Particularly in the heterogeneous catalysis, these techniques are powerful tool for investigating the influence of composition, preparation method and pretreatment on the reactivity of the surface or bulk with gases.

desorption

Gas composition at the reactor outlet Characterisation of adsorptive properties of materials, surface acidity, total adsorption and

desorption capacity with temperature,

mechanism and kinetics of adsorption and

formation in bimetallic catalysts

Specific volume of solid sample

kinetic mechanism

DMC Differential microcalorimetry Enthalpy difference between sample and reference

temperatures of rate maxima, metal surface area and dispersion, surface energetic heterogeneity, binding states and energies of adsorbed molecules,

Characterisation of redox properties of materials, consumption of reducing agent with temperature range, temperatures of rate maxima, valence states of metal atoms in zeolites and metal oxides, metal oxide and support interaction, indication of alloy

Characterisation of coke species in deactivated catalysts and redox properties of metal oxides, total coke content in deactivated catalysts, mechanism and kinetics of oxidation reactions, deactivation

Thermal analysis techniques Characteristic factor monitored

design of microchannels/catalysts.

**8. Thermoanalytical technique** 

TPx Temperature-programmed

Temperature-programmed desorption

Temperature-programmed reduction

Temperature-programmed oxidation

Thermomechanical analysis (TMA),

Thermogravimetry (TG) Weight of sample Differential thermogravimetry (DTG) Rate of weight changes

Table 1. Thermal analysis techniques and their uses

Thermomagnetic analysis (TMA) Magnetic susceptibility

reaction

(TPD)

(TPR)

(TPO)

Dilatometry

The suitable model can account for heterogeneity in terms of chemisorption by introducing a sufficient number of different adsorption states. The readsorption formulated and tested against experimental data by describing the intrinsic dynamics of an adsorption state as a quasiequilibrium adsorption/desorption between the gas phase and the surface. Conceptually it is potentially useful for the elucidation of fundamental reaction mechanistic information for design and intensification of catalyst. The transient response methods place similar requirements on experimentation as conventional methods of kinetic analysis. In slow transient input, all or part of the elementary reactions may proceed in quasiequilibrium and their individual rates then can not properly elucidated. In general this is based on nonlinear regression analysis, and omits assumptions of pseudo-steady state.

In our extensive studies of catalyst intensification we conduct series of kinetic analysis for TPx experiments in phenomenological way [15-17]. Modelling of TPD data is based on adsorption and desorption theories, while TPR and TPO call for gas-solid surface reaction mechanisms, which may include topochemical characteristics. Further the rate of surface reaction divided into intrinsic rate of reaction per area of reaction interface and change in the reaction interface in reaction [18]. In general kinetic analysis of thermoanalytical data where there is a linear temperature rise (β) represented in the rate equation, where, rate of reaction is proportional to Arrhenius temperature dependence, A is pre-exponential factor, E is activation energy, and these are function of degree of conversion, f(α).

$$
\beta \frac{d\alpha}{dT} = A \exp\left[\frac{-E}{RT}\right] f(\alpha) \tag{3}
$$

The transforming eq. 3 by taking natural logarithms on both sides gives

$$
\ln\left[\beta \frac{d\alpha}{dT}\right] = \ln\left(Af(\alpha)\right) - \frac{E}{RT} \tag{4}
$$

The plot of ln(βdα/dT) versus 1/T results straight line at a selected conversion level, the slope of the line gives -E/R at certain conversion level. This also did not explain reaction mechanism but by varing activation energy E=E(α) results from the Friedman analysis explains that eq. 3 with a normal physico-chemical interpretation will not describes the reaction kinetics.

#### **9. Design and intensification of Pt-Sn-based alkane dehydrogenation catalyst**

The control of chemical reaction pathways at molecular level presents undoubtedly the most important scientific challenge on the way to fully sustainable, thermodynamically efficient chemical processes. Integrated molecular reaction control is one of the major breakthroughs in achieving sustainable and efficient chemical processes design. Numerous fundamental works on this topic have been reported in literature and it is now the key challenge to advance them in an interdisciplinary way towards the industrial catalysis. The molecular alignment and geometry of collisions as well as selectivity of desired product was tackled by catalyst topology. In this example we discusses Pt-Sn based catalyst design and intensification for light alkane dehydrogenation to light alkenes. Constraints of this reaction are very complex due to endothermic limitation and endothermicity.

Rational Asymmetric Catalyst Design, Intensification and Modeling 197

distribution of Si in the framework [39, 40]. Its Si based acid sites make it resistant to hydrothermal treatment and gives superiority over ZSM-5, where dealumination occur frequently [33, 34]. These observations suggest that the weak acidic support SAPO-34, which can sophisticatedly control the stereo-chemistry and gives shape selectivity effect for

The influences of the support on the reduction properties of Pt-Sn/SAPO-34 catalyst were analyzed by H2-TPR and results are shown in Fig. 5. The reduction peaks at higher temperatures (above 450 oC) indicated that the Pt was well interacted with the support in the presence of Sn [5, 6, 14, 34]. The peak height and width increases with the increase in Sn loading, between 400 to 600 oC that may be due to joint reduction and valuable interaction between Pt and Sn on the SAPO-34 [24]. At higher temperatures towards Sno species forms. But the oxidation state of Sn (+2) in a bimetallic Pt–Sn catalyst was important, while the Pt-Sn alloy formation results in permanent deactivation [16, 24, 30]. The Pt attached to SnO is believed to be active for dehydrogenation reaction. The reduction peaks at the higher temperature indicated that the Pt had a strongly interaction with the SAPO-34 incomparative to other supports. The H2 reduction results from Pt-Sn/SAPO-34 displayed sharp reduction peaks that confirmed the joint reduction and valuable interaction between

Fig. 5. H2-TPR profiles of PtSn based different support catalysts of identical Pt and Sn

propylene in addition to being a bi-functional catalyst, Pt-Sn/SAPO-34.

Pt and Sn on the SAPO-34.

composition.

Bringing more molecules at the energy levels exceeding the activation energy threshold by conductive heating but it offers only a macroscopic control over the reaction and is thermodynamically inefficient. There are two broader approaches to control stereochemistry of reaction. First consists of immobilizing the molecules during reaction in confined nanospaces by imposing "hard walls"of structures or of other molecules, or to force reacting species to assume certain position inside the offered structure. The second approach consists in acting upon moving molecules with spatially oriented external fields. Methods for controlling molecular alignment and orientation via nano-structural confinement invite shape-selective and imprinted catalyst. Shape-selective zeolites (having distinct sized nanocage) discriminate between reactants and products and between tran-sition states molecules within the pores [19, 20].

Pt-Sn-based catalysts supported on amorphous supports or zeolites, for propane dehydrogenation have been discussed in many studies [15-29]. The selection of Sn as a promoter has been explained in terms of geometric and/or electronic effects [30]. In the geometric effect, Sn decreases the size of platinum ensembles, which reduces hydrogenolysis and coking reactions. Sn also modifies the electronic density of Pt, either by positive charge transfer from the Snn+ species or the different electronic structures in Pt-Sn alloys [31]. The deactivation of Pt-based catalyst during propane dehydrogenation is fatal, and mainly due to the aggregation of Pt particles and carbon deposition [31, 32]. However, the mechanistic understanding is still controversial and the reaction results were not conclusive, especially on stability and selectivity. In early days much focus has been given to Pt-Sn-based catalysts supported on Al2O3 for propane dehydrogenation. But the lifetime of these catalysts is too short. Therefore, there is a growing interest in developing zeolite supported catalysts, e.g. Pt-Sn/ZSM-5 zeolite for propane dehydrogenation, because of longer activation times and higher conversions [15-17]. Many studies have improved the performance of Pt-Sn/ZSM-5 catalysts by adding a more metallic promoter (particularly from alkali or alkaline earth metals) like Na, Fe, Zn, etc. [33, 34] and by increasing the Si/Al ratio of the ZSM-5 support [16, 34]. However, ZSM-5 supported catalysts were detrimentally affected by frequent regeneration with steam [35] and have lower propylene selectivity [16]. ZSM-5 supported catalysts can further take part in cracking and their large pore size fails to create shape selectivity effect for propylene. Moreover comparing the effects of supports and additional promoters, it was observed that the performance of propane dehydrogenation reaction significantly dependent on support. In particular it was noted that the control of intermediates products to desired product over the bi-functional catalyst was significantly support dependent.

Detailed TPx experimentation and kinetic studies with continuous experimental validity took us towards the innovative design of Pt-Sn-based zeolite supported catalyst. In this case, competitive reactions present a picture of a nano-snooker game. Unfortunately, in practicing industrial reactors has very limited degree of control of molecular-level events, therefore the catalyst should be robust enough. Extensive analysis suggests SAPO-34 as support (almost inert to dehydrogenation), with good shape selective opportunities and hydrothermally stable. The silico-aluminophosphate zeolite SAPO-34 is a microporous sieve with a chabasite-like structure (structural code CHA), has an extremely good shape selective effect for propylene [36-38]. Their Bronsted acid sites are responsible for their activities and based on two distinct hydroxyls (OH-bridges to Al and Si) on the structural framework. The structure, acidity and catalytic properties of SAPO-34 depend on the number and

Bringing more molecules at the energy levels exceeding the activation energy threshold by conductive heating but it offers only a macroscopic control over the reaction and is thermodynamically inefficient. There are two broader approaches to control stereochemistry of reaction. First consists of immobilizing the molecules during reaction in confined nanospaces by imposing "hard walls"of structures or of other molecules, or to force reacting species to assume certain position inside the offered structure. The second approach consists in acting upon moving molecules with spatially oriented external fields. Methods for controlling molecular alignment and orientation via nano-structural confinement invite shape-selective and imprinted catalyst. Shape-selective zeolites (having distinct sized nanocage) discriminate between reactants and products and between tran-sition states molecules

Pt-Sn-based catalysts supported on amorphous supports or zeolites, for propane dehydrogenation have been discussed in many studies [15-29]. The selection of Sn as a promoter has been explained in terms of geometric and/or electronic effects [30]. In the geometric effect, Sn decreases the size of platinum ensembles, which reduces hydrogenolysis and coking reactions. Sn also modifies the electronic density of Pt, either by positive charge transfer from the Snn+ species or the different electronic structures in Pt-Sn alloys [31]. The deactivation of Pt-based catalyst during propane dehydrogenation is fatal, and mainly due to the aggregation of Pt particles and carbon deposition [31, 32]. However, the mechanistic understanding is still controversial and the reaction results were not conclusive, especially on stability and selectivity. In early days much focus has been given to Pt-Sn-based catalysts supported on Al2O3 for propane dehydrogenation. But the lifetime of these catalysts is too short. Therefore, there is a growing interest in developing zeolite supported catalysts, e.g. Pt-Sn/ZSM-5 zeolite for propane dehydrogenation, because of longer activation times and higher conversions [15-17]. Many studies have improved the performance of Pt-Sn/ZSM-5 catalysts by adding a more metallic promoter (particularly from alkali or alkaline earth metals) like Na, Fe, Zn, etc. [33, 34] and by increasing the Si/Al ratio of the ZSM-5 support [16, 34]. However, ZSM-5 supported catalysts were detrimentally affected by frequent regeneration with steam [35] and have lower propylene selectivity [16]. ZSM-5 supported catalysts can further take part in cracking and their large pore size fails to create shape selectivity effect for propylene. Moreover comparing the effects of supports and additional promoters, it was observed that the performance of propane dehydrogenation reaction significantly dependent on support. In particular it was noted that the control of intermediates products to desired product over the bi-functional catalyst

Detailed TPx experimentation and kinetic studies with continuous experimental validity took us towards the innovative design of Pt-Sn-based zeolite supported catalyst. In this case, competitive reactions present a picture of a nano-snooker game. Unfortunately, in practicing industrial reactors has very limited degree of control of molecular-level events, therefore the catalyst should be robust enough. Extensive analysis suggests SAPO-34 as support (almost inert to dehydrogenation), with good shape selective opportunities and hydrothermally stable. The silico-aluminophosphate zeolite SAPO-34 is a microporous sieve with a chabasite-like structure (structural code CHA), has an extremely good shape selective effect for propylene [36-38]. Their Bronsted acid sites are responsible for their activities and based on two distinct hydroxyls (OH-bridges to Al and Si) on the structural framework. The structure, acidity and catalytic properties of SAPO-34 depend on the number and

within the pores [19, 20].

was significantly support dependent.

distribution of Si in the framework [39, 40]. Its Si based acid sites make it resistant to hydrothermal treatment and gives superiority over ZSM-5, where dealumination occur frequently [33, 34]. These observations suggest that the weak acidic support SAPO-34, which can sophisticatedly control the stereo-chemistry and gives shape selectivity effect for propylene in addition to being a bi-functional catalyst, Pt-Sn/SAPO-34.

The influences of the support on the reduction properties of Pt-Sn/SAPO-34 catalyst were analyzed by H2-TPR and results are shown in Fig. 5. The reduction peaks at higher temperatures (above 450 oC) indicated that the Pt was well interacted with the support in the presence of Sn [5, 6, 14, 34]. The peak height and width increases with the increase in Sn loading, between 400 to 600 oC that may be due to joint reduction and valuable interaction between Pt and Sn on the SAPO-34 [24]. At higher temperatures towards Sno species forms. But the oxidation state of Sn (+2) in a bimetallic Pt–Sn catalyst was important, while the Pt-Sn alloy formation results in permanent deactivation [16, 24, 30]. The Pt attached to SnO is believed to be active for dehydrogenation reaction. The reduction peaks at the higher temperature indicated that the Pt had a strongly interaction with the SAPO-34 incomparative to other supports. The H2 reduction results from Pt-Sn/SAPO-34 displayed sharp reduction peaks that confirmed the joint reduction and valuable interaction between Pt and Sn on the SAPO-34.

Fig. 5. H2-TPR profiles of PtSn based different support catalysts of identical Pt and Sn composition.

Rational Asymmetric Catalyst Design, Intensification and Modeling 199

 PtSn/Al2O3 PtSn/ZSM-5 PtSn/SBA-15 PtSn/SAPO-34

02468

Time-on-Stream (hr)

02468

 PtSn/Al2O3 PtSn/ZSM-5 PtSn/SBA-15 PtSn/SAPO-34

Time-on-Stream (hr)

Fig. 8. Catalysts selectivity of propylene over different supported PtSn-based catalysts

0

50

60

70

Propylene Selectivity (wt.%)

80

90

100

Fig. 7. Catalysts activity at 586 oC for propane conversion

15

Conversion (wt.%)

30

45

Coke formation during propane dehydrogenation is the inherent factors that adversely affect catalyst performance. The amount of coke formed over different catalysts (after 8 hr reaction) was analyzed by TPO. The results are shown in Fig. 6. The typical TPO profile shows two successive peaks, indicating coke deposited on the metallic sites and support, while the peak height represents the intensity of deposition. The peaks between 300-400 oC and 550 – 650 oC corresponds to coke deposited on metallic site and support, respectively. It was noted that the strong electronic attachment over SAPO-34 facilitated the transfer of carbon deposits on metal sites to support. It's only possible when the metals are well interacted with the support. Moreover Sn modifies Pt ensembles by decreasing surface area and reduces carbon–metal bonds/interaction, and allows the intermediates to mobile easily to support [41].

Fig. 6. TPO profiles of coke deposits on different supported PtSn-based catalysts after 8 hr propane dehydrogenation reaction under identical operating conditions.

The superiority of novel support SAPO-34 performance was experimentally verified (results are shown in Fig. 7 and 8), and it was observed that Pt-Sn/SAPO-34 catalyst controls the stereo-chemistry of propane dehydrogenation reaction much better than PtSn-based catalysts supported on ZSM-5 and Al2O3.

Coke formation during propane dehydrogenation is the inherent factors that adversely affect catalyst performance. The amount of coke formed over different catalysts (after 8 hr reaction) was analyzed by TPO. The results are shown in Fig. 6. The typical TPO profile shows two successive peaks, indicating coke deposited on the metallic sites and support, while the peak height represents the intensity of deposition. The peaks between 300-400 oC and 550 – 650 oC corresponds to coke deposited on metallic site and support, respectively. It was noted that the strong electronic attachment over SAPO-34 facilitated the transfer of carbon deposits on metal sites to support. It's only possible when the metals are well interacted with the support. Moreover Sn modifies Pt ensembles by decreasing surface area and reduces carbon–metal bonds/interaction, and allows the intermediates to mobile easily

Fig. 6. TPO profiles of coke deposits on different supported PtSn-based catalysts after 8 hr

The superiority of novel support SAPO-34 performance was experimentally verified (results are shown in Fig. 7 and 8), and it was observed that Pt-Sn/SAPO-34 catalyst controls the stereo-chemistry of propane dehydrogenation reaction much better than PtSn-based

propane dehydrogenation reaction under identical operating conditions.

catalysts supported on ZSM-5 and Al2O3.

to support [41].

Fig. 7. Catalysts activity at 586 oC for propane conversion

Fig. 8. Catalysts selectivity of propylene over different supported PtSn-based catalysts

Rational Asymmetric Catalyst Design, Intensification and Modeling 201

[12] Aghalayam, P.; Park, Y. K.; Vlachos, D. G., AIChE J., 2000, 46, 2017-2029, Chem. Eng.

[13] Ashish B. Mhadeshwar, Hai Wang, and Dionisios G. Vlachos, THERMODYNAMIC

[15] Zeeshan Nawaz and Fei Wei, Pt-Sn-based catalyst's intensification using Al2O3-SAPO-

[16] Zeeshan Nawaz, Tang Xiaoping, Qiang Zhang, Wang Dezheng and Fei Wei, A highly

[17] Zeeshan Nawaz, Xiaoping Tang, Yao Wang and Fei Wei, Parametric characterization

[18] Haber, J., Kinetics and mechanism of the reduction of VI-b transition metal oxides and

[19] M.E. Davis, Katz A, W.R. Ahmad, Rational catalyst design via imprinted

[22] G.D. Angel, A. Bonilla, Y. Pena, J. Navarrete, J.L.G. Fierro, D.R. Acosta, J. Catal., 219

[20] M. Stöcker, Gas phase catalysis by zeolites, Microp. Mesop. Mater. 82 (2005) 257–292.

[26] M.S. Kumar, D. Chen, J.C. Walmsley, A. Holmen, Catal. Commun. 9 (2008) 747-750. [27] S.M. Stagg, C.A. Quenrini, W.E. Alvarez, D.E. Resasco, J. Catal. 168 (1997) 75-94. [28] Y. Zhang, Y. Zhou, A. Qiu, Y. Wang, Y. Xu, P. Wu, Catal. Comm. 7 (2006) 860–866. [29] G. Aguilar-Rios, P. Salas, M.A. Valenzuela, H. Armendariz, J.A. Wang, Salmones, Catal.

[30] J. Llorca, N. Homs, J. Leon, J. Sales, J.L.G. Fierro, P. Ramirez, Appl. Catal. A. Gen. 189

[31] C. Kappenstein, M. Guierin, K. Liaziar, K. Matusek, Z. Paial, J. Chem. Faraday Trans. 94

[33] Y. Zhang, Y. Zhou, A. Qiu, Y. Wang, Y. Xu, P. Wu, Acta Phys.-Chim. Sin., 22 (2006)

[34] Y. Zhang, Y. Zhou, H. Liu, Y. Wang, Y. Xu, P. Wu, Appl. Catal. A. Gen. 333 (2007) 202-

[35] Y. Zhang, Y. Zhou, K. Yang, Y. Li, Y. Wang, Y. Xu, P. Wu, Micropor. Mesopor. Mater. 96

[36] M.A. Anderson, B. Sulikowski, B.J. Barrie, J. Klinowski, J. Phys. Chem. 94 (1990) 2730-

[32] M.S. Kumar, D. Chen, A. Holmen, J.C. Walmsley, Catal. Today. 142 (2009) 17-23.

thermal analysis, Thermochim. Acta 7 (1973) 333-556.

and Engineering Chemistry, 17 (2011) 389-393.

Catalysis Communications, 10 (2009) 1925-1930.

their oxysalts, J. Less-Comm. Met. 54 (1977) 243-261.

nanostructured materials, Chem. Mater. 8 (1996) 1820–1839.

[23] T. Waku, J.A. Biscardi, E. Iglesia, Chem. Commun., (2003) 1764-1772. [24] F.B. Passos, M. Schmal, M.A. Vannice, J. Catal. 160 (1996) 106-117.

[25] P.L.D. Cola, R. Gläser, J. Weitkamp, Appl. Catal. A. Gen. 306 (2006) 85-97.

CONSISTENCY IN MICROKINETIC DEVELOPMENT OF SURFACE REACTION MECHANISMS, Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem. 2003, 48(2), 514. [14] Sestak, J., Satava V., Wendlandt, W. W., The study of heterogeneous processes by

34 as a support for propane dehydrogenation to propylene, Journal of Industrial

selective Pt-Sn/SAPO-34 catalyst for propane dehydrogenation to propylene,

and influence of Tin on the performance of Pt-Sn/SAPO-34 catalyst for selective propane dehydrogenation to propylene, Industrial & Engineering Chemistry

Progress, Oct. issue, 2000.

Research, 49 (2010) 1274–1280.

[21] H. Lieske, J. Volter, J. Catal. 90 (1984) 96-105.

(2003) 63-73.

Lett. 60 (1999) 21-25.

(1999) 77-86.

672−678.

210.

2734.

(1998) 2463-2474.

(2006) 245–254.

#### **10. Conclusion**

Tight time and cost constraints force catalysis research to continuously reduce their experimental efforts during catalyst development and to facilitate the application of modelbased strategy. Nevertheless, the effort of setting up a sophisticated model or criteria in coupling with novel TPx analysis for the design and intensification of catalysts is still higher. To overcome bottlenecks in catalyst modelling, considerable effort has to be spent with the systematization of kinetic models, formalization of reaction pathways, and development of knowledge-based software tools. This type of contribution claimed by number of software developers as discussed, while for the catalyst design we must need TPx analysis to achieve a robust design. It will be shown, how informal textual or metal-support information defining requirements on catalyst for a specified use. To find the optimal design under these uncertainties, a stochastic optimization method can be employed and verified by experimentation. In this work, the optimal properties of a catalyst for direct dehydrogenation of propane to propylene is presented, and uncertainties associated with the reactions and their parameters are modelled to design sustainable catalyst.

#### **11. References**


Tight time and cost constraints force catalysis research to continuously reduce their experimental efforts during catalyst development and to facilitate the application of modelbased strategy. Nevertheless, the effort of setting up a sophisticated model or criteria in coupling with novel TPx analysis for the design and intensification of catalysts is still higher. To overcome bottlenecks in catalyst modelling, considerable effort has to be spent with the systematization of kinetic models, formalization of reaction pathways, and development of knowledge-based software tools. This type of contribution claimed by number of software developers as discussed, while for the catalyst design we must need TPx analysis to achieve a robust design. It will be shown, how informal textual or metal-support information defining requirements on catalyst for a specified use. To find the optimal design under these uncertainties, a stochastic optimization method can be employed and verified by experimentation. In this work, the optimal properties of a catalyst for direct dehydrogenation of propane to propylene is presented, and uncertainties associated with

the reactions and their parameters are modelled to design sustainable catalyst.

Selective Catalysts. Nature 1998, 393, 52-54.

Catalysts. J. Am. Chem. Soc. 2001, 123, 8404-8405.

Catalyst Design, Ind. Eng. Chem. Res. 2004, 43, 3484-3512.

Chem. Eng. 1987, 11 (3), 265-277.

Chem. Eng. 1988, 12 (9-10), 923-938.

Catal. Today 1999, 50, 511-515.

Washington, DC, 1993; p 316.

Washington, D.C., 1993. [10] Stoltze, P., Prog. Surf. Sci., 2000, 65, 65-150.

[11] Schmitz, G., J. Chem. Phys., 2000, 112, 10714-10717.

[1] Jones, C. W.; Tsuji, C.; Davis, M. E. Organic-Functionalized Molecular Sieves as Shape-

[2] Banares-Alcantara, R.; Westerberg, A. W.; Ko, E. I.; Rychener, M. D. Decadesa Hybrid

[3] Banares-Alcantara, R.; Ko, E. I.; Westerberg, A. W.; Rychener, M. D., Decadesa Hybrid

[4] Van Santen, R. A. Theory, Spectroscopy and Kinetics of Zeolite Catalyzed Reactions.

[5] Jacobsen, J. H. C.; Dahl, S.; Clausen, B. S.; Bahn, S.; Logadottir, A.; Norskov, J. K. Catalyst

[6] Santhoji Katare, James M. Caruthers, W. Nicholas Delgass, and Venkat

[7] Dumesic, J. A.; Rudd, D. F.; Aparicio, L. M.; Rekoske, J. E.; Trevino, A. A. The

[8] J.M. Caruthers, J.A. Lauterbach, K.T. Thomson, V. Venkatasubramanian, C.M. Snively,

Expert System for Catalyst Selection-I. Expert System Consideration. Comput.

Expert System for Catalyst Selection-Ii. Final Architecture and Results. Comput.

Design by Interpolation in the Periodic Table: Bimetallic Ammonia Synthesis

Venkatasubramanian, An Intelligent System for Reaction Kinetic Modeling and

Microkinetics of Heterogeneous Catalysis; American Chemical Society:

A. Bhan, S. Katare, and G. Oskarsdottir, Catalyst design: knowledge extraction from high-throughput experimentation, Journal of Catalysis 216 (2003) 98–109. [9] Dumesic, J. A.; Rudd, D. F.; Aparicio, L. M.; Rekoske, J. E.; Trevino, A. A., The

Microkinetics of Heterogeneous Catalysis, American Chemical Society:

**10. Conclusion** 

**11. References** 


**7** 

Jairo Cubillos

*Colombia* 

*Department of Chemistry Sciences,* 

**Preparation, Catalytic Properties and** 

*Pedagogical and Technological University of Colombia, Tunja* 

**Recycling Capabilities Jacobsen's Catalyst** 

Metal salen complexes, such as the Jacobsen's catalyst, have attracted much interest in the last few decades because of their unique catalytic activity, especially as olefin epoxidation catalysts, in the presence of terminal oxidants like iodosylbenzene (PhIO), sodium hypochlorite (NaOCl), and *meta*-chloroperoxybenzoic acid (*m*-CPBA) (Gladasi, 2007). Salen ligands of Jacobsen's catalyst bind manganese ions through four atoms, two nitrogen and two oxygen atoms. This tetradentate-binding motif is reminiscent of the porphyrin framework in the heme-based oxidative enzymes (Canali & Sherrington, 1999). Nonetheless, salen derivatives are more easily synthesized than porphyrins and their structures are more easily manipulated to create an asymmetric environment around the active metal site (Canali & Sherrington, 1999). However, this homogeneous catalyst cannot be separated from the reaction media and, subsequently, cannot be recycled. Moreover, it suffers deactivation in homogeneous phase by either formation of dimeric μ-oxo-manganese (IV) species, which are inactive in the olefin epoxidation or oxidative degradation of the salen ligand through the imine group (Figure 1) (Xia et al., 2005). The conventional ways to solve these problems are to immobilize Jacobsen-type catalysts onto solid supports. The last decade has witnessed an intense research effort to heterogenise Jacobsen-type catalysts, and in general chiral manganese(III) salen complexes, using several types of supports in order to make them recyclable as well as economical (Murzin et al., 2005). Reports on the heterogenization of Jacobsen-type catalysts have been centered on their covalent binding to organic polymers (Clapham et al., 2001) and on their encapsulation, entrapment, adsorption and covalent attachment to porous inorganic supports, such as zeolites, Si-MCM-41, Al-MCM-41 and clays (Cubillos, 2010; Dasa, 2006; Kureshy, 2006), and also on activated carbon (Mahata et al., 2007). Unfortunately, the catalytic activity of the recovered catalyst decreases during the catalytic tests of reuse. It is found that isolating Mn(III) salen complexes onto a solid support increases the catalyst stability (Baleizaõ & García, 2006) by suppressing the formation of inactive dimeric μ-oxo manganese(IV) species. However, the deactivation route by ligand oxidation cannot be avoided by anchoring the catalyst to a solid matrix, since it depends on the oxidation conditions. On the other hand, the immobilized catalysts usually lead to partial loss of activity and/or enantioselectivity as compared to their analogous homogeneous catalysts (Baleizaõ & García, 2006). It is well known that in homogeneous phase, the catalyst acquires an appropriate geometric configuration that promotes both, the

**1. Introduction** 


### **Preparation, Catalytic Properties and Recycling Capabilities Jacobsen's Catalyst**

#### Jairo Cubillos

*Department of Chemistry Sciences, Pedagogical and Technological University of Colombia, Tunja Colombia* 

#### **1. Introduction**

202 Advances in Chemical Engineering

[38] J. Chen, P.A.Wright, J.M. Thomas, S. Natarajan, L. Marchese, S.M. Bradley, G. Sankar,

[39] G. Sastre, D.W. Lewis, C. Richard, A. Catlow, J. Phys. Chem. B 101 (1997) 5249-5262.

C.R.A. Catlow, P.L. Gai-Boyes, R.P. Townsend, C.M. Lok, J. Phys. Chem. 98 (1994)

[37] Y. Xu, C.P. Grey, J.M. Thomas, A.K. Cheetham, Catal. Lett. 4 (1990) 251-260.

[40] E. Dumitriu, A. Azzouz, V. Hulea, Micropor. Mesopor. Mater. 10 (1997) 1-8. [41] F.N. Thomas, V. Hannelore, A.L. Johannes, J. Catal., 157 (1995) 388-395.

10216-10224.

Metal salen complexes, such as the Jacobsen's catalyst, have attracted much interest in the last few decades because of their unique catalytic activity, especially as olefin epoxidation catalysts, in the presence of terminal oxidants like iodosylbenzene (PhIO), sodium hypochlorite (NaOCl), and *meta*-chloroperoxybenzoic acid (*m*-CPBA) (Gladasi, 2007). Salen ligands of Jacobsen's catalyst bind manganese ions through four atoms, two nitrogen and two oxygen atoms. This tetradentate-binding motif is reminiscent of the porphyrin framework in the heme-based oxidative enzymes (Canali & Sherrington, 1999). Nonetheless, salen derivatives are more easily synthesized than porphyrins and their structures are more easily manipulated to create an asymmetric environment around the active metal site (Canali & Sherrington, 1999). However, this homogeneous catalyst cannot be separated from the reaction media and, subsequently, cannot be recycled. Moreover, it suffers deactivation in homogeneous phase by either formation of dimeric μ-oxo-manganese (IV) species, which are inactive in the olefin epoxidation or oxidative degradation of the salen ligand through the imine group (Figure 1) (Xia et al., 2005). The conventional ways to solve these problems are to immobilize Jacobsen-type catalysts onto solid supports. The last decade has witnessed an intense research effort to heterogenise Jacobsen-type catalysts, and in general chiral manganese(III) salen complexes, using several types of supports in order to make them recyclable as well as economical (Murzin et al., 2005). Reports on the heterogenization of Jacobsen-type catalysts have been centered on their covalent binding to organic polymers (Clapham et al., 2001) and on their encapsulation, entrapment, adsorption and covalent attachment to porous inorganic supports, such as zeolites, Si-MCM-41, Al-MCM-41 and clays (Cubillos, 2010; Dasa, 2006; Kureshy, 2006), and also on activated carbon (Mahata et al., 2007). Unfortunately, the catalytic activity of the recovered catalyst decreases during the catalytic tests of reuse. It is found that isolating Mn(III) salen complexes onto a solid support increases the catalyst stability (Baleizaõ & García, 2006) by suppressing the formation of inactive dimeric μ-oxo manganese(IV) species. However, the deactivation route by ligand oxidation cannot be avoided by anchoring the catalyst to a solid matrix, since it depends on the oxidation conditions. On the other hand, the immobilized catalysts usually lead to partial loss of activity and/or enantioselectivity as compared to their analogous homogeneous catalysts (Baleizaõ & García, 2006). It is well known that in homogeneous phase, the catalyst acquires an appropriate geometric configuration that promotes both, the

Preparation, Catalytic Properties and Recycling Capabilities Jacobsen's Catalyst 205

other oxidizing agents (NaOCl and *m*-CPBA) most commonly used, were employed for R- (+)-limonene epoxidation with R,R-Jacobsen as catalyst. The influence of catalyst optical

The main reagents used for the synthesis of Jacobsen-type catalysts were 2,4-di-tertbutylphenol, tin(IV) chloride (99%, Aldrich®), paraformaldehyde (99.5%, Aldrich®), L-(+) tartaric acid (99.5%, Aldrich®), D-(–)-tartaric acid (99, Aldrich®), *cis*/*trans* 1,2 diaminocyclohexane (99%, Aldrich®), potassium carbonate (99.995%, Aldrich®), manganese acetate tetrahydrate (99%, Aldrich®) and lithium chloride (99.99%, Aldrich®). All these materials were directly used as received. The solvents were absolute ethanol, chloroform, acetone and dichloromethane. 3,5-di-tert-butyl salicylaldehyde was prepared from the formylation of 2,4-di-tert-butylphenol using paraformaldehyde as reactant and tin(IV) chloride as catalyst (Deng & Jacobsen, 1992). (R,R)-1,2-diammoniumcyclohexane and (S,S)- 1,2-diammoniumcyclohexane mono-(+)-tartrate salt were prepared from the racemic resolution of *cis*/*trans* 1,2-diaminocyclohexane with L-(+)-tartaric acid and D-(–)-tartaric acid, respectively (Deng & Jacobsen, 1992). Other reagents included R-(+)-limonene (97%, Aldrich®), ethyl phenylpropiolate (98%, Aldrich®), Lindlar's catalyst (5 wt.% on calcium carbonate, Aldrich®), Oxone® (2KHSO5•KHSO4, K2SO4, Aldrich®), manganese(II) acetate tetrahydrate (99%, Aldrich®), and sodium bicarbonate (Merck), aqueous sodium hypochlorite (NaOCl, 7% active chlorine basis, Carlo Erba), 4-phenylpyridine N-oxide (98%, Aldrich®), *meta*-chloroperoxybenzoic acid (*m*-CPBA, 77%, Aldrich), 4-methylmorpholine Noxide (97%, Aldrich®). *Cis*-ethyl cinnamate was prepared from the partial hydrogenation of

configuration on catalytic performance for this reaction is also revised.

ethyl phenylpropiolate using the Lindlar's catalyst (Larrow & Jacobsen, 1994).

Different salen ligands R,R-Jacobsen, S,S-Jacobsen and racemic were prepared according to the methods previously described (Deng & Jacobsen, 1992), by condensing the appropriate diamine (3.5 mmol) and 3,5-di-tert-butyl salicylaldehyde (7.0 mmol), for 2 h, using ethanol as a solvent. Upon reaction, completion of a yellow–orange precipitate appeared. These precipitates were filtered and dried under vacuum. The obtained solids were characterized by FT-IR (Bahramian et al., 2006). The Jacobsen's catalysts R,R-Jacobsen, S,S-Jacobsen and Jacobsen racemic (Figure 2), were prepared as previously described (Deng & Jacobsen, 1992). Ethanolic solutions of 3.0 mmol of different ligands, and 3.3 mmol of manganese (II) acetate tetrahydrate were refluxed, for 1 h. During reflux, a change in the color of the solution from yellow–orange to dark-brown was observed. The manganese complexes were recrystallized in heptane and dried under vacuum. The FT-IR and DR UV–Vis spectral bands of all complexes were found to be identical to those reported (Bahramian, 2006;

R,R-Jacobsen, S,S-Jacobsen and Jacobsen racemic catalysts were then used as catalysts for the epoxidation of R-(+)-limonene (1) and *cis*-ethyl cinnamate (2). In a standard procedure, 1.0 mmol of substrate, 1.2 mmol of sodium bicarbonate and 0.05 mmol of catalyst were dissolved in 4 mL of acetone. A buffer solution (aqueous NaHCO3, 5 % wt) was added to bring the pH in the range between 8.0-8.5 (mixture A). In another vessel, 2.0 mmol of KHSO5 (Oxone®) was dissolved in 4 mL of water (mixture B). While mixture A was being stirred, mixture B was slowly added, keeping the pH in the range 8.0 - 8.5 using NaHCO3 solution (5 % w/w aqueous). When mixture B was completely added, stirring was stopped and the

**2. Experimental section** 

Chaube, 2005).

M = Mn, Cr, Co, Cu, V, Ru, Fe, Al, Ti R1, R2, R3, R4 = H, alkyl or phenyl group R'1, R'2, R'3, R'4 = H, nitro, alkyl, aryl or alcoxy group

Fig. 1. Chemical structure of metal salen complexes.

oxygen transfer from oxo-Mn(V) active species to the double bond of olefin and chiral induction (Adam et al., 2000). In contrast, when the catalyst is not free to move due to the influence of the solid support, an inappropriate geometrical configuration can be obtained (Fan et al., 2002). In addition to that, structural modification of either the catalyst or the solid support during the immobilization process, generally lead to diminished catalytic activity (Mastrorilli & Nobile, 2004). The most common immobilization methods by covalent bond modify one or two tert-butyl groups of the salen ligand. It is known that the tert-butyl groups are very important, since they define the optimal trajectory of the incoming olefin towards the oxo-Mn(III) active species (Linde et al., 2005). In summary, compared with the homogeneous counterparts, some of the immobilized complexes often suffer from various disadvantages, such as poor activity, leaching of the active species into the reaction medium, and low substrate accessibility (Baleizaõ & García, 2006). Therefore, for industrial practical merit and academic interest of homogeneously catalyzed reactions, the development of an efficient strategy for catalyst recovery is still challenging.

An alternative convenient and economical strategy is to adjust the solubility of homogeneous catalysts by varying the reaction conditions, resulting in the direct separation of the catalysts during the reaction (Kureshy et al., 2001). Recently, I and my colleagues described that using dimethyldioxirane (DMD) as oxidizing agent for the epoxidation of R-(+)-limonene and *cis*ethyl cinnamate with Jacobsen-type catalysts could facilitate product isolation and catalyst recovery by segregating the catalysts into a solid phase during the reaction (Cubillos et al., 2009). Limonene di-epoxide was the main product from limonene, whereas an epoxide was obtained with 78% e.e in the case of *cis*-ethyl cinnamate. Moreover, it was reported that catalyst stability towards the oxidative degradation could be enhanced in the reaction medium, using DMD as oxidizing agent (Cubillos et al., 2009). On the other hand, it was found that the stereogenic center of the pure enantiomerically Jacobsen's catalyst did not influence the catalytic activity for the epoxidation of R-(+)-limonene (Cubillos et al., 2010).

Here I report the preparation and the catalytic properties as well as recycling capabilities of pure enantiomerically Jacobsen's catalysts (R,R-Jacobsen and S,S-Jacobsen) and its racemic form for epoxidation of R-(+)-limonene and *cis*-ethyl cinnamate with *in situ* generated DMD as oxidant. In order to compare the activity, recovery and reuse of the catalysts with DMD, other oxidizing agents (NaOCl and *m*-CPBA) most commonly used, were employed for R- (+)-limonene epoxidation with R,R-Jacobsen as catalyst. The influence of catalyst optical configuration on catalytic performance for this reaction is also revised.

#### **2. Experimental section**

204 Advances in Chemical Engineering

R4

R'4

C C N N

R2 R3

Imine group Imine group

R1

M = Mn, Cr, Co, Cu, V, Ru, Fe, Al, Ti R1, R2, R3, R4 = H, alkyl or phenyl group

R'1

Fig. 1. Chemical structure of metal salen complexes.

efficient strategy for catalyst recovery is still challenging.

O O

Cl

R'2 R'3

R'1, R'2, R'3, R'4 = H, nitro, alkyl, aryl or alcoxy group

oxygen transfer from oxo-Mn(V) active species to the double bond of olefin and chiral induction (Adam et al., 2000). In contrast, when the catalyst is not free to move due to the influence of the solid support, an inappropriate geometrical configuration can be obtained (Fan et al., 2002). In addition to that, structural modification of either the catalyst or the solid support during the immobilization process, generally lead to diminished catalytic activity (Mastrorilli & Nobile, 2004). The most common immobilization methods by covalent bond modify one or two tert-butyl groups of the salen ligand. It is known that the tert-butyl groups are very important, since they define the optimal trajectory of the incoming olefin towards the oxo-Mn(III) active species (Linde et al., 2005). In summary, compared with the homogeneous counterparts, some of the immobilized complexes often suffer from various disadvantages, such as poor activity, leaching of the active species into the reaction medium, and low substrate accessibility (Baleizaõ & García, 2006). Therefore, for industrial practical merit and academic interest of homogeneously catalyzed reactions, the development of an

An alternative convenient and economical strategy is to adjust the solubility of homogeneous catalysts by varying the reaction conditions, resulting in the direct separation of the catalysts during the reaction (Kureshy et al., 2001). Recently, I and my colleagues described that using dimethyldioxirane (DMD) as oxidizing agent for the epoxidation of R-(+)-limonene and *cis*ethyl cinnamate with Jacobsen-type catalysts could facilitate product isolation and catalyst recovery by segregating the catalysts into a solid phase during the reaction (Cubillos et al., 2009). Limonene di-epoxide was the main product from limonene, whereas an epoxide was obtained with 78% e.e in the case of *cis*-ethyl cinnamate. Moreover, it was reported that catalyst stability towards the oxidative degradation could be enhanced in the reaction medium, using DMD as oxidizing agent (Cubillos et al., 2009). On the other hand, it was found that the stereogenic center of the pure enantiomerically Jacobsen's catalyst did not influence

Here I report the preparation and the catalytic properties as well as recycling capabilities of pure enantiomerically Jacobsen's catalysts (R,R-Jacobsen and S,S-Jacobsen) and its racemic form for epoxidation of R-(+)-limonene and *cis*-ethyl cinnamate with *in situ* generated DMD as oxidant. In order to compare the activity, recovery and reuse of the catalysts with DMD,

the catalytic activity for the epoxidation of R-(+)-limonene (Cubillos et al., 2010).

M

The main reagents used for the synthesis of Jacobsen-type catalysts were 2,4-di-tertbutylphenol, tin(IV) chloride (99%, Aldrich®), paraformaldehyde (99.5%, Aldrich®), L-(+) tartaric acid (99.5%, Aldrich®), D-(–)-tartaric acid (99, Aldrich®), *cis*/*trans* 1,2 diaminocyclohexane (99%, Aldrich®), potassium carbonate (99.995%, Aldrich®), manganese acetate tetrahydrate (99%, Aldrich®) and lithium chloride (99.99%, Aldrich®). All these materials were directly used as received. The solvents were absolute ethanol, chloroform, acetone and dichloromethane. 3,5-di-tert-butyl salicylaldehyde was prepared from the formylation of 2,4-di-tert-butylphenol using paraformaldehyde as reactant and tin(IV) chloride as catalyst (Deng & Jacobsen, 1992). (R,R)-1,2-diammoniumcyclohexane and (S,S)- 1,2-diammoniumcyclohexane mono-(+)-tartrate salt were prepared from the racemic resolution of *cis*/*trans* 1,2-diaminocyclohexane with L-(+)-tartaric acid and D-(–)-tartaric acid, respectively (Deng & Jacobsen, 1992). Other reagents included R-(+)-limonene (97%, Aldrich®), ethyl phenylpropiolate (98%, Aldrich®), Lindlar's catalyst (5 wt.% on calcium carbonate, Aldrich®), Oxone® (2KHSO5•KHSO4, K2SO4, Aldrich®), manganese(II) acetate tetrahydrate (99%, Aldrich®), and sodium bicarbonate (Merck), aqueous sodium hypochlorite (NaOCl, 7% active chlorine basis, Carlo Erba), 4-phenylpyridine N-oxide (98%, Aldrich®), *meta*-chloroperoxybenzoic acid (*m*-CPBA, 77%, Aldrich), 4-methylmorpholine Noxide (97%, Aldrich®). *Cis*-ethyl cinnamate was prepared from the partial hydrogenation of ethyl phenylpropiolate using the Lindlar's catalyst (Larrow & Jacobsen, 1994).

Different salen ligands R,R-Jacobsen, S,S-Jacobsen and racemic were prepared according to the methods previously described (Deng & Jacobsen, 1992), by condensing the appropriate diamine (3.5 mmol) and 3,5-di-tert-butyl salicylaldehyde (7.0 mmol), for 2 h, using ethanol as a solvent. Upon reaction, completion of a yellow–orange precipitate appeared. These precipitates were filtered and dried under vacuum. The obtained solids were characterized by FT-IR (Bahramian et al., 2006). The Jacobsen's catalysts R,R-Jacobsen, S,S-Jacobsen and Jacobsen racemic (Figure 2), were prepared as previously described (Deng & Jacobsen, 1992). Ethanolic solutions of 3.0 mmol of different ligands, and 3.3 mmol of manganese (II) acetate tetrahydrate were refluxed, for 1 h. During reflux, a change in the color of the solution from yellow–orange to dark-brown was observed. The manganese complexes were recrystallized in heptane and dried under vacuum. The FT-IR and DR UV–Vis spectral bands of all complexes were found to be identical to those reported (Bahramian, 2006; Chaube, 2005).

R,R-Jacobsen, S,S-Jacobsen and Jacobsen racemic catalysts were then used as catalysts for the epoxidation of R-(+)-limonene (1) and *cis*-ethyl cinnamate (2). In a standard procedure, 1.0 mmol of substrate, 1.2 mmol of sodium bicarbonate and 0.05 mmol of catalyst were dissolved in 4 mL of acetone. A buffer solution (aqueous NaHCO3, 5 % wt) was added to bring the pH in the range between 8.0-8.5 (mixture A). In another vessel, 2.0 mmol of KHSO5 (Oxone®) was dissolved in 4 mL of water (mixture B). While mixture A was being stirred, mixture B was slowly added, keeping the pH in the range 8.0 - 8.5 using NaHCO3 solution (5 % w/w aqueous). When mixture B was completely added, stirring was stopped and the

Preparation, Catalytic Properties and Recycling Capabilities Jacobsen's Catalyst 207

comparing the chromatogram of our products with those of *cis*-ethyl cinnamate isomers available in literature (Steiner et al., 2002). Also, isolated yields of the major epoxide product for either substrate were calculated. Thus, the major epoxide originating from the *cis*-ethyl cinnamate oxidation was purified by short-path distillation (110 °C and 0.5 mmHg), while in

In order to compare the catalytic activity, recovery and reuse of catalysts with DMD, other oxidizing agents such as NaOCl and *m*-CPBA were explored for R-(+)-limonene epoxidation according to reported conventional methods (Wang et al., 2006), using the same molar ratio of substrate/oxidant (0.5 mmol/mmol), substrate/catalyst (20 mmol/mmol), reaction time

Figure 3 shows the FTIR spectra of the salen ligands and their corresponding catalysts. The salen ligands show a characteristic band around 1620 cm-1, which is associated with the vibrations of the imine group (HC=N) (Bahramian et al., 2006). In the catalyst samples, this band is displaced towards lower wavelengths (1600-1590 cm-1) as the first evidence of the formation of the organometallic complex. Additionally, characteristic bands at 1530 (C-O), 550 (Mn-O) and 480 cm-1 (Mn-N) are also associated with the complexation of manganese by

Fig. 3. FT-IR Spectra of (R,R) Jacobsen's ligand (a), (R,R) Jacobsen's catalyst (b), (S,S) Jacobsen's ligand (c), (S,S) Jacobsen's catalyst (d), Jacobsen's racemic ligand (e), Jacobsen's

Figure 4 shows DR UV-vis spectra of the salen ligands and their corresponding catalysts. The salen ligands exhibit absorption bands at 265 nm and 335 nm. These bands are attributed to π→π\* transitions. The band at 265 nm has been assigned to the benzene ring and the one at 335 nm, to the imino groups (Chaube et al., 2005). The imino π→π\* transitions in the Mn salen complexes is shifted to larger wavelengths due to metal coordination,

the case of R-(+)-limonene, diepoxide was collected at 140 °C and 0.5 mmHg.

(30 min) and reaction temperature (25 °C) as DMD.

**3. Results and discussion** 

racemic catalyst (f).

the salen ligand (Bahramian et al., 2006).

Fig. 2. Chemical structure of Jacobsen type catalysts. A: Racemic Jacobsen, B: R,R-Jacobsen, C: S,S-Jacobsen.

formed solids separated by filtration and/or centrifugation. The liquid phase was extracted with dicloromethane in a separation funnel and analyzed by GC. The solid phase was washed with sufficient water up to reach a constant weight in the obtained residue. This dark brown residue (catalyst) was easily dissolved in acetone and thus ready for recycling. Also, the chemical identity of this residue was analyzed by FTIR.

An Agilent Technologies 7890A gas chromatograph (GC), equipped with a DB-1 capillary column (50 m long, 0.32 mm ID and 1.20 mm film thickness) and a FID detector was used for the analysis of solvent purity, olefin and oxidation products. Ultra high pure helium was used as carrier gas (30 mL/min). The injection port temperature was kept at 300 C. For separation of R-(+)-limonene the column temperature was programmed between 80 and 140 C while for *cis*-ethyl cinnamate it was kept isothermal at 140 C. The area normalization method was used to determine conversion, selectivity and relative yield. The enantiomeric excess (ee) for the single epoxide derived from *cis*-ethyl cinnamate epoxidation was determined by GC using a chiral capillary column, i.e. Betadex GTA (60 m long, 250 mm ID and 0.25 mm film thickness). In this case, a commercially available 3-ethyl-phenylglycidate (*cis*/*trans* = 10/90, Aldrich) racemic mixture was used. In the case of the chiral epoxides derived from R-(+)-limonene, 1,2 limonene oxide (97%, mixture of *cis* and *trans*, Aldrich) was used. Limonene diepoxide was prepared by oxidation of R-(+)-limonene (1.0 mmol) oxide using *m*-CPBA (5.0 mmol) as oxidizing agent and confirmed by GC–MS. The optical configuration was assigned by comparing the chromatogram of our products with those of *cis*-ethyl cinnamate isomers available in literature (Steiner et al., 2002). Also, isolated yields of the major epoxide product for either substrate were calculated. Thus, the major epoxide originating from the *cis*-ethyl cinnamate oxidation was purified by short-path distillation (110 °C and 0.5 mmHg), while in the case of R-(+)-limonene, diepoxide was collected at 140 °C and 0.5 mmHg.

In order to compare the catalytic activity, recovery and reuse of catalysts with DMD, other oxidizing agents such as NaOCl and *m*-CPBA were explored for R-(+)-limonene epoxidation according to reported conventional methods (Wang et al., 2006), using the same molar ratio of substrate/oxidant (0.5 mmol/mmol), substrate/catalyst (20 mmol/mmol), reaction time (30 min) and reaction temperature (25 °C) as DMD.

#### **3. Results and discussion**

206 Advances in Chemical Engineering

C N N C

H H \* \*

O O Mn

Cl

**(A)**

C N C

H H

O O Mn N

Cl

**(B)**

C N N C

H H

O O Mn

Cl

**(C)** Fig. 2. Chemical structure of Jacobsen type catalysts. A: Racemic Jacobsen, B: R,R-Jacobsen,

formed solids separated by filtration and/or centrifugation. The liquid phase was extracted with dicloromethane in a separation funnel and analyzed by GC. The solid phase was washed with sufficient water up to reach a constant weight in the obtained residue. This dark brown residue (catalyst) was easily dissolved in acetone and thus ready for recycling.

An Agilent Technologies 7890A gas chromatograph (GC), equipped with a DB-1 capillary column (50 m long, 0.32 mm ID and 1.20 mm film thickness) and a FID detector was used for the analysis of solvent purity, olefin and oxidation products. Ultra high pure helium was used as carrier gas (30 mL/min). The injection port temperature was kept at 300 C. For separation of R-(+)-limonene the column temperature was programmed between 80 and 140 C while for *cis*-ethyl cinnamate it was kept isothermal at 140 C. The area normalization method was used to determine conversion, selectivity and relative yield. The enantiomeric excess (ee) for the single epoxide derived from *cis*-ethyl cinnamate epoxidation was determined by GC using a chiral capillary column, i.e. Betadex GTA (60 m long, 250 mm ID and 0.25 mm film thickness). In this case, a commercially available 3-ethyl-phenylglycidate (*cis*/*trans* = 10/90, Aldrich) racemic mixture was used. In the case of the chiral epoxides derived from R-(+)-limonene, 1,2 limonene oxide (97%, mixture of *cis* and *trans*, Aldrich) was used. Limonene diepoxide was prepared by oxidation of R-(+)-limonene (1.0 mmol) oxide using *m*-CPBA (5.0 mmol) as oxidizing agent and confirmed by GC–MS. The optical configuration was assigned by

Also, the chemical identity of this residue was analyzed by FTIR.

C: S,S-Jacobsen.

Figure 3 shows the FTIR spectra of the salen ligands and their corresponding catalysts. The salen ligands show a characteristic band around 1620 cm-1, which is associated with the vibrations of the imine group (HC=N) (Bahramian et al., 2006). In the catalyst samples, this band is displaced towards lower wavelengths (1600-1590 cm-1) as the first evidence of the formation of the organometallic complex. Additionally, characteristic bands at 1530 (C-O), 550 (Mn-O) and 480 cm-1 (Mn-N) are also associated with the complexation of manganese by the salen ligand (Bahramian et al., 2006).

Fig. 3. FT-IR Spectra of (R,R) Jacobsen's ligand (a), (R,R) Jacobsen's catalyst (b), (S,S) Jacobsen's ligand (c), (S,S) Jacobsen's catalyst (d), Jacobsen's racemic ligand (e), Jacobsen's racemic catalyst (f).

Figure 4 shows DR UV-vis spectra of the salen ligands and their corresponding catalysts. The salen ligands exhibit absorption bands at 265 nm and 335 nm. These bands are attributed to π→π\* transitions. The band at 265 nm has been assigned to the benzene ring and the one at 335 nm, to the imino groups (Chaube et al., 2005). The imino π→π\* transitions in the Mn salen complexes is shifted to larger wavelengths due to metal coordination,

Preparation, Catalytic Properties and Recycling Capabilities Jacobsen's Catalyst 209

The results of catalytic activity and recyclability are collected in Tables 1 and 2. In the absence of catalyst, reaction was observed for R-(+)-limonene (conversion 53 % and relative rates of reaction 0.145 mol/Lt×h). In this case, 79% selectivity to limonene diepoxide with 46% relative yield and 30% isolated yield was achieved. 1,2-limonene oxide was obtained as secondary product with very low selectivities (Table 1, entry 1,). In contrast, no reaction was observed using *cis*-ethyl cinnamate as substrate (Tables 1 and 2, entry 8). It is known that unfunctionalized olefins are more reactive than its functionalized counterparts when

(%)

1 (**1**)a None 58 18e 79g 46i 2 (**1**)a I 100 5e 93g 93i 3 (**1**)a II 100 3e 93g 93i 4 (**1**)a III 100 4e 95g 95i 5 (**1**)b I 99 7e 90g 89i 6 (**1**)c I 98 5e 87g 85i 7 (**1**)d I 98 8e 90g 88i 8 (**2**)a None 0 - - 0

11 (**2**)a I 15 90f - 14j

13 (**2**)c II 8 83f - 7j 14 (**2**)d II 6 82f - 5j a Reaction conditions: substrate = 1.0 mmol; KHSO5 = 2.0 mmol (from Oxone®), catalyst = 0.05 mmol; acetone = 4.0 mL; water = 4.0 mL; reaction time = 30 min. b First reuse. c Second reuse. d Third reuse. e Selectivity to 1,2-limonene oxide. f Selectivity to ethyl-3- phenylglycidate (sum of *cis* and *trans*). g Sum of the four diepoxides of R-(+)-limonene. h Relative yield = Conversion×Selectivity/10,000. i Relative yield

Relative yield to ethyl-3-phenylglycidate (sum of *cis* and *trans*).

Similar yields (relative yields 93-95% and isolated yields 56-60%) to limonene diepoxide were reached with all catalysts, while in the absence of catalyst 46% and 30% respectively, were obtained. This finding suggests that the chiral center of enantiomerically pure (R,R and S,S-Jacobsen) appear to have little or no influence on the catalytic activity; rather the state of coordination given by the salen ligand to the manganese appears to be crucial. Here, the contribution of the catalyst is proven once again by an increase of both R-(+)-limonene conversion (or relative rates of reaction) and selectivity to limonene diepoxide (sum of the four diepoxides of R-(+)-limonene). None of the four diepoxides of R-(+)-limonene predominated. In contrast, the catalytic oxidation of *cis*-ethyl cinnamate offers the possibility

Epoxide selectivity (%)

Diepoxide selectivity (%)




Relative yieldh (%)

Jacobsen type catalysts are used (Porter & Skidmore, 2000).

9 (**2**)a II 18 88f

10 (**2**)a III 20 84f

12 (**2**)b II 12 82f

Table 1. Results of catalytic activity and reutilization

to the diepoxides. j

Entry Substrate Catalyst Conversion

Fig. 4. DR UV-vis Spectra of (R,R) Jacobsen's ligand (a), (R,R) Jacobsen's catalyst (b), (S,S) Jacobsen's ligand (c), (S,S) Jacobsen's catalyst (d), Jacobsen's racemic ligand (e), Jacobsen's racemic catalyst (f).

confirming the formation of Mn(III) salen complex (Chaube et al., 2005). UV–Vis and FT-IR spectra revealed that the salen ligands are unaffected and are not decomposed upon coordination of the organo-funtional groups with manganese (Chaube et al., 2005).

R,R-Jacobsen, S,S-Jacobsen and Jacobsen racemic catalysts were examined in the liquid phase oxidation reaction of *cis*-ethyl cinnamate and R-(+)-limonene using *in situ* generated DMD as oxidizing agent. Different oxidation products were obtained, depending on the substrate type. Figures 5 and 6 outline the test reactions. Di-epoxides appear as the main products from R-(+)-limonene oxidation (Figure 5), while a single epoxide ((2R,3R)-*cis* ethyl-3-phenylglycidate) was produced from *cis*-ethyl cinnamate oxidation (Figure 6). Similarly to the previously reported experiments of olefins oxidation (Cubillos, 2009, 2010), the Jacobsen type catalyst was easily separated from the obtained solid phase at the end of reaction.

Fig. 5. Selective oxidation of R-(+)-limonene (1).

Fig. 6. Enantioselective oxidation of *cis*-ethyl cinnamate (2).

Fig. 4. DR UV-vis Spectra of (R,R) Jacobsen's ligand (a), (R,R) Jacobsen's catalyst (b), (S,S) Jacobsen's ligand (c), (S,S) Jacobsen's catalyst (d), Jacobsen's racemic ligand (e), Jacobsen's

confirming the formation of Mn(III) salen complex (Chaube et al., 2005). UV–Vis and FT-IR spectra revealed that the salen ligands are unaffected and are not decomposed upon

R,R-Jacobsen, S,S-Jacobsen and Jacobsen racemic catalysts were examined in the liquid phase oxidation reaction of *cis*-ethyl cinnamate and R-(+)-limonene using *in situ* generated DMD as oxidizing agent. Different oxidation products were obtained, depending on the substrate type. Figures 5 and 6 outline the test reactions. Di-epoxides appear as the main products from R-(+)-limonene oxidation (Figure 5), while a single epoxide ((2R,3R)-*cis* ethyl-3-phenylglycidate) was produced from *cis*-ethyl cinnamate oxidation (Figure 6). Similarly to the previously reported experiments of olefins oxidation (Cubillos, 2009, 2010), the Jacobsen type catalyst was easily separated from the obtained solid phase at the end of reaction.

O O

Catalyst NaHCO3/H2O 8.0<pH<8.5

Catalyst NaHCO3/H2O pH = 8.0

O

C C COC2

H5

Limonene diepoxide

Ph O

H O

H H

coordination of the organo-funtional groups with manganese (Chaube et al., 2005).

H3 C C CH3

O

Acetone

Acetone

CH3

*Cis*-ethyl cinnamate (2R,3R)-*cis* ethyl-3-phenylglycidate

KHSO5

+ +

Potassium Monopersulfate

KHSO5

+ +

Potassium Monopersulfate

Fig. 6. Enantioselective oxidation of *cis*-ethyl cinnamate (2).

H3 C C

R-(+)-Limonene

COC2H5

O

C C

H H

Ph

Fig. 5. Selective oxidation of R-(+)-limonene (1).

racemic catalyst (f).

The results of catalytic activity and recyclability are collected in Tables 1 and 2. In the absence of catalyst, reaction was observed for R-(+)-limonene (conversion 53 % and relative rates of reaction 0.145 mol/Lt×h). In this case, 79% selectivity to limonene diepoxide with 46% relative yield and 30% isolated yield was achieved. 1,2-limonene oxide was obtained as secondary product with very low selectivities (Table 1, entry 1,). In contrast, no reaction was observed using *cis*-ethyl cinnamate as substrate (Tables 1 and 2, entry 8). It is known that unfunctionalized olefins are more reactive than its functionalized counterparts when Jacobsen type catalysts are used (Porter & Skidmore, 2000).


a Reaction conditions: substrate = 1.0 mmol; KHSO5 = 2.0 mmol (from Oxone®), catalyst = 0.05 mmol; acetone = 4.0 mL; water = 4.0 mL; reaction time = 30 min. b First reuse. c Second reuse. d Third reuse. e Selectivity to 1,2-limonene oxide. f Selectivity to ethyl-3- phenylglycidate (sum of *cis* and *trans*). g Sum of the four diepoxides of R-(+)-limonene. h Relative yield = Conversion×Selectivity/10,000. i Relative yield to the diepoxides. j Relative yield to ethyl-3-phenylglycidate (sum of *cis* and *trans*).

Table 1. Results of catalytic activity and reutilization

Similar yields (relative yields 93-95% and isolated yields 56-60%) to limonene diepoxide were reached with all catalysts, while in the absence of catalyst 46% and 30% respectively, were obtained. This finding suggests that the chiral center of enantiomerically pure (R,R and S,S-Jacobsen) appear to have little or no influence on the catalytic activity; rather the state of coordination given by the salen ligand to the manganese appears to be crucial. Here, the contribution of the catalyst is proven once again by an increase of both R-(+)-limonene conversion (or relative rates of reaction) and selectivity to limonene diepoxide (sum of the four diepoxides of R-(+)-limonene). None of the four diepoxides of R-(+)-limonene predominated. In contrast, the catalytic oxidation of *cis*-ethyl cinnamate offers the possibility

Preparation, Catalytic Properties and Recycling Capabilities Jacobsen's Catalyst 211

2 NaOCla 65 75d Vacuum distillation 3 *m*-CPBAb 35 95d Vacuum distillation 4 DMDc 100 93e Centrifugation and filtration

Table 3. R-(+)- limonene epoxidation using R,R-Jacobsen as catalyst . Effect of the oxidizing

about 95% with *m*-CPBA. In the case of DMD, R-(+)-limonene diepoxides, were the major products. These differences in selectivities can be associated to the easy segregation of the catalyst. Presumably, this phenomenon creates active sites in the precipitated solid that promotes double epoxidation. It has been reported that the catalyst is not separable by means of physical-mechanical methods, when NaOCl and *m*-CPBA were used as oxidizing agents (Abdi et al., 2004). In these cases, a distillation process under vacuum (160 °C and 0.08 MPa) was required in order to separate the catalyst from reaction products, whereas with DMD the catalyst was isolated by physical-mechanical separation methods. On the other hand, figure 7 reveals that the catalyst exhibited the best stability to the oxidative degradation when *in situ* generated DMD was used as the oxidizing agent, since the main band associated to the Mn(III) salen complexes (1530 cm-1) is still present in the FTIR spectrum, where DMD was used as oxidizing agent. Other important bands associated to the Mn-O (550 cm-1) and Mn-N (480 cm-

**2000 1800 1600 1400 1200 1000 800 600 400**

Fig. 7. FT-IR Spectra of the (R,R) Jacobsen's catalyst used with different oxidizing agents:

**Wavenumbers/cm-1)**

1) stretching vibrations are slightly displaced towards a lower wavelength.

**1530 cm-1**

**1530 cm-1**

fresh catalyst (a), DMD (b), NaOCl (c) and m-CPBA (d).

selectivity (%)

Process of catalyst recovery

**a**

**b**

**550 cm-1 480 cm-1**

**500 cm 460 cm-1 -1**

**c**

**d**

Conversion (%)

1 None 0 - -

a R-(+)-limonene = 1.0 mmol; NaOCl = 2.0 mmol (0.05 M aqueous Na2HPO4 solution) ; 4-phenyl pyridine N-oxide = 0.4 mmol; R,R-Jacobsen = 0.05 mmol; dichloromethane = 4 mL Reaction time = 30 min. b R-(+)-limonene = 1.0 mmol; *m*-CPBA= 2.0 mmol; 4-methylmorpholine N-oxide = 5 mmol; R,R-Jacobsen = 0.05 mmol; dichloromethane = 4 mL; Reaction time = 30 min. c R-(+)-limonene = 1.0 mmol; KHSO5 = 2.0 mmol (from Oxone®); catalyst = 0.05 mmol; acetone = 4.0 mL; water = 4.0 mL; Reaction time = 30 min. d Selectivity to 1,2-limonene oxide; e Selectivity to R-(+)-limonene diepoxides (sum of the

Entry Oxidizing

four diepoxides of R-(+)-limonene).

**T r a n s m i s s i o n / a . u**

agent.

agent


a Reaction conditions: substrate = 1.0 mmol; KHSO5 = 2.0 mmol (from Oxone®);catalyst = 0.05 mmol; acetone = 4.0 mL; water = 4.0 mL. Reaction time = 30 min. b First reuse. c Second reuse. d Third reuse. e Relative rate of reaction = Conversion×initial mol/reaction volume×reaction time. f Enantiomeric excess of (2R,3R)-cis-ethyl-3-phenylglycidate over (2S,3S)-cis-ethyl-3-phenylglycidate. g Enantiomeric excess of (2S,3S)-*cis*-ethyl-3-phenylglycidate over (2R,3R)-*cis*-ethyl-3-phenylglycidate. h Isolated yield to R-(+) limonene diepoxides. i Isolated yield to ethyl-3-phenylglycidate (sum of *cis* and *trans*).

Table 2. Results of catalytic activity and reaction rates

to epoxidize the unique C=C double bond located in its chemical structure. In order to perform enantioselective epoxidation, a pure enantiomerically catalyst is required, which can be reached either using R,R-Jacobsen or S,S-Jacobsen. In general, good selectivities to ethyl-3-phenylglycidate (sum of *cis* and *trans*, 84-90%) and good enantiomeric excesses to (2R,3R)-*cis*-ethyl-3-phenylglycidate (78%) were obtained with R,R-Jacobsen. In contrast, lower yields were reached (relative yields 16 and isolated yields 11). Additionally, it is worth to note that the product stereochemistry is strongly dependent on the absolute configuration of catalyst. Thus, *cis*-ethylcinnmate with R,R-Jacobsen gives(2R,3R)-*cis*-ethyl-3-phenylglycidate (Table 2, entry 9), whereas R-(+)-limonene with S,S-Jacobsen gives (2S,2S)-*cis*-ethyl-3-phenylglycidate (Table 2, entry 10). This shows clearly the specificity of a pure enantiomeric catalyst for inducing the preferential formation of the observed product.

Table 3 shows the results of catalytic activity and catalyst recovery for R-(+)-limonene epoxidation using the R,R-Jacobsen catalyst and three oxidants: DMD, NaOCl and *m*-CPBA. For these experiments, Figure 7 shows the spectra of the fresh and used catalyst after reaction with either oxidant. As listed in Table 3, the largest conversion was obtained with DMD (100%), although different selectivities were obtained. Thus, 1,2-limonene oxide was the major product when NaOCl and *m*-CPBA were used as oxidizing agents, obtaining

1 (**1**)a None 0.145 - 30h 2 (**1**)a I 0.25 - 56h 3 (**1**)a II 0.25 - 57h 4 (**1**)a III 0.25 - 60h 5 (**1**)b I 0.25 - 52h 6 (**1**)c I 0.245 - 53h 7 (**1**)d I 0.245 - 50h 8 (**2**)a None 0 - 0

10 (**2**)a III 0.050 55g 9i 11 (**2**)a I 0.038 0 7i 12 (**2**)b II 0.030 76f 7i 13 (**2**)c II 0.020 74f 5i 14 (**2**)d II 0.015 73f 4i a Reaction conditions: substrate = 1.0 mmol; KHSO5 = 2.0 mmol (from Oxone®);catalyst = 0.05 mmol; acetone = 4.0 mL; water = 4.0 mL. Reaction time = 30 min. b First reuse. c Second reuse. d Third reuse. e

of (2R,3R)-cis-ethyl-3-phenylglycidate over (2S,3S)-cis-ethyl-3-phenylglycidate. g Enantiomeric excess of (2S,3S)-*cis*-ethyl-3-phenylglycidate over (2R,3R)-*cis*-ethyl-3-phenylglycidate. h Isolated yield to R-(+)-

to epoxidize the unique C=C double bond located in its chemical structure. In order to perform enantioselective epoxidation, a pure enantiomerically catalyst is required, which can be reached either using R,R-Jacobsen or S,S-Jacobsen. In general, good selectivities to ethyl-3-phenylglycidate (sum of *cis* and *trans*, 84-90%) and good enantiomeric excesses to (2R,3R)-*cis*-ethyl-3-phenylglycidate (78%) were obtained with R,R-Jacobsen. In contrast, lower yields were reached (relative yields 16 and isolated yields 11). Additionally, it is worth to note that the product stereochemistry is strongly dependent on the absolute configuration of catalyst. Thus, *cis*-ethylcinnmate with R,R-Jacobsen gives(2R,3R)-*cis*-ethyl-3-phenylglycidate (Table 2, entry 9), whereas R-(+)-limonene with S,S-Jacobsen gives (2S,2S)-*cis*-ethyl-3-phenylglycidate (Table 2, entry 10). This shows clearly the specificity of a pure enantiomeric catalyst for inducing the preferential formation of the observed product. Table 3 shows the results of catalytic activity and catalyst recovery for R-(+)-limonene epoxidation using the R,R-Jacobsen catalyst and three oxidants: DMD, NaOCl and *m*-CPBA. For these experiments, Figure 7 shows the spectra of the fresh and used catalyst after reaction with either oxidant. As listed in Table 3, the largest conversion was obtained with DMD (100%), although different selectivities were obtained. Thus, 1,2-limonene oxide was the major product when NaOCl and *m*-CPBA were used as oxidizing agents, obtaining

Isolated yield to ethyl-3-phenylglycidate (sum of *cis* and *trans*).

9 (**2**)a II 0.045 78f

Relative rate of reaction = Conversion×initial mol/reaction volume×reaction time. f

Table 2. Results of catalytic activity and reaction rates

limonene diepoxides. i

reactione ( mol/lt×h) Enantiomeric excess (%)

Isolated yield (%)

11i

Enantiomeric excess

Entry Substrate Catalyst Relative rate of


a R-(+)-limonene = 1.0 mmol; NaOCl = 2.0 mmol (0.05 M aqueous Na2HPO4 solution) ; 4-phenyl pyridine N-oxide = 0.4 mmol; R,R-Jacobsen = 0.05 mmol; dichloromethane = 4 mL Reaction time = 30 min. b R-(+)-limonene = 1.0 mmol; *m*-CPBA= 2.0 mmol; 4-methylmorpholine N-oxide = 5 mmol; R,R-Jacobsen = 0.05 mmol; dichloromethane = 4 mL; Reaction time = 30 min. c R-(+)-limonene = 1.0 mmol; KHSO5 = 2.0 mmol (from Oxone®); catalyst = 0.05 mmol; acetone = 4.0 mL; water = 4.0 mL; Reaction time = 30 min. d Selectivity to 1,2-limonene oxide; e Selectivity to R-(+)-limonene diepoxides (sum of the four diepoxides of R-(+)-limonene).

Table 3. R-(+)- limonene epoxidation using R,R-Jacobsen as catalyst . Effect of the oxidizing agent.

about 95% with *m*-CPBA. In the case of DMD, R-(+)-limonene diepoxides, were the major products. These differences in selectivities can be associated to the easy segregation of the catalyst. Presumably, this phenomenon creates active sites in the precipitated solid that promotes double epoxidation. It has been reported that the catalyst is not separable by means of physical-mechanical methods, when NaOCl and *m*-CPBA were used as oxidizing agents (Abdi et al., 2004). In these cases, a distillation process under vacuum (160 °C and 0.08 MPa) was required in order to separate the catalyst from reaction products, whereas with DMD the catalyst was isolated by physical-mechanical separation methods. On the other hand, figure 7 reveals that the catalyst exhibited the best stability to the oxidative degradation when *in situ* generated DMD was used as the oxidizing agent, since the main band associated to the Mn(III) salen complexes (1530 cm-1) is still present in the FTIR spectrum, where DMD was used as oxidizing agent. Other important bands associated to the Mn-O (550 cm-1) and Mn-N (480 cm-1) stretching vibrations are slightly displaced towards a lower wavelength.

Fig. 7. FT-IR Spectra of the (R,R) Jacobsen's catalyst used with different oxidizing agents: fresh catalyst (a), DMD (b), NaOCl (c) and m-CPBA (d).

Preparation, Catalytic Properties and Recycling Capabilities Jacobsen's Catalyst 213

Adam, W.; Mock-Knoblauch, C.; Saha-Mo1ller, C. R. & Herderich, M (2000). Are MnIV

Bahramian, B.; Mirkhani, V.; Moghadam, M. & Amin, A. H (2006). Water-soluble

Canali, C. & Sherrington, D (1999). Utilisation of homogeneous and supported chiral

Chaube, V. D.; Shylesh, S. & Singh, A. P (2005). Synthesis, characterization and catalytic

Clapham, B.; Reger, T. S. & Janda K. D (2001). Polymer-supported Catalysis in Synthetic

Cubillos, J.; Montilla, I. & Montes de Correa, C (2009). Easy separation and reutilization of

Cubillos, J.; Vargas, M.; Reyes, J.; Villa, A. & Montes de Correa, C (2010). Effect of the

Cubillos, J.; Vasquéz, S. & Montes de Correa, C (2010). Salen manganese (III) complexes as

Dasa, P.; Kuźniarska-Biernacka, I.; Silva, A. R.; Carvalho, A. P.; Pires, J. & Freire, C (2006).

Fan, Q. H.; Li, Y. M. & Chan A. S. C (2002). Recoverable catalysts for asymmetric organic

Gladiali, S. (2007). Guidelines and Methodologies in Asymmetric Synthesis and Catalysis. *C.* 

Kureshy, R. I.; Ahmad, I.; Khan, N. H.; Abdi, S. H. R.; Pathak, K. & Jasra R. V (2006). Chiral

*R. Chimie*, Vol.10, No.3, (March 2007), pp. 220-231, ISSN 1631-0748.

*A: Chem*. Vol.248, No.1-2, (January 2006), pp. 135-143, ISSN 1381-1169. Deng, L. & Jacobsen, E. N (1992). A practical, highly enantioselective synthesis of the Taxol

(September 2006), pp. 3987-4043, ISSN 0009-2665.

(September 2009). pp. 348–352, ISSN 0926-960X.

(January 2010). pp. 57–65, ISSN 0926-960X.

4320–4323, ISSN 0022-3263.

(1999), pp. 85-93, ISSN 0306-0012.

ISSN 0002-7863.

0040-4020.

1520-636X.

2665.

species involved in Mn(salen)-catalyzed Jacobsen−Katsuki epoxidations? a mechanistic elucidation of their formation and reaction modes by EPR spectroscopy, mass-spectral analysis, and product studies: chlorination versus oxygen transfer. *J. Am. Chem. Soc*, Vol.122, No.40, (September 2000), pp. 9685–9691,

manganese(III) salen complex as a mild and selective catalyst for oxidation of alcohols. *Appl. Catal. A: Gen*, Vol.315, (November 2006), pp. 52-57, ISSN 0926-960X. C. Baleizaõ. & H. García (2006). Chiral salen complexes: an overview to recoverable and

reusable homogeneous and heterogeneous catalysts. *Chem. Rev*, Vol.106, No.9,

metal(salen) complexes in asymmetric catalysis. *Chem. Soc. Rev*, Vol.28, No.2,

activity of Mn(III)- and Co(II)-salen complexes immobilized mesoporous alumina. *J. Mol. Catal. A: Chem*, Vol.241, No.1-2, (November 2005). pp. 79-87, ISSN: 1381-1169.

Organic Chemistry. Tetrahedron, Vol.57, No.22, (May 2001), pp. 4637-4662, ISSN

the Jacobsen's catalyst in olefin oxidation. *Appl. Catal. A: Gen*, Vol. 366, No.2,

substrate and catalyst chirality on the diastereoselective epoxidation of limonene using Jacobsen-type catalysts. *Chirality*, Vol.22, No.4, (June 2009), pp. 403-410, ISSN

catalysts for R-(+)-limonene oxidation. *Appl. Catal. A: Gen*, Vol. 373, No. 1-2.

Encapsulation of chiral Mn(III) salen complexes into aluminium pillared clays: Application as heterogeneous catalysts in the epoxidation of styrene. *J. Mol. Catal.* 

side chain via asymmetric catalysis. *J. Org. Chem*, Vol. 57, No.15, (July 1992), pp.

synthesis. *Chem. Rev*., Vol.102, No.10, (September 2002). pp. 3385-3466, ISSN 0009-

Mn(III) salen complexes covalently bonded on modified MCM-41 and SBA-15 as

Finally, catalyst reuse was explored for both reactions. As can be observed in Tables 1 and 2, catalysts experienced a slight decrease in their initial catalytic activity through three consecutive runs in both reactions. The catalyst was segregated into a solid phase, while the reaction products remained in the liquid phase. This allowed the easy separation of catalyst and reaction products. On the other hand, Figure 8 shows the spectra of the catalyst used in three cycles. Clearly, it is observed that the main bands associated to the Mn(III) salen complexes are retained in the used catalysts. It indicates that the catalyst is very stable to the reaction conditions during three consecutive runs, whereas with oxidizing agents like NaOCl and *m*-CPBA the catalyst was deactivated in the first run (Figure 7). Therefore, the slight loss of catalytic activity with DMD is associated to physical loss of the catalyst during the isolation process rather than to oxidative degradation.

Fig. 8. FT-IR Spectra of the (R,R) Jacobsen's catalyst used in various run: fresh catalyst (a), second run (b) third run (c) fourth run.

#### **4. Conclusion**

Diepoxides were the main products from the oxidation of R-(+)-limonene, whereas a monoepoxide is obtained in the case of the catalytic oxidation of *cis*-ethyl cinnamate. In the latter case, a pure enantiomerically catalyst is required, while the Jacobsen racemic catalyst was sufficient in the case of the catalytic oxidation of R-(+)-limonene. Given that the catalytic activity of the three catalysts is very similar for R-(+)-limonene epoxidation and considering that the unique difference among the three catalysts is its stereogenic center located in the bond C1-C2 of the 1,2-diamino cyclohexane component, I conclude that the catalytic activity is not dependent on the stereogenic center. The catalyst could be recycled three times without appreciable loss of its initial activity.

#### **5. References**

Abdi, S. H. R.; Kureshy, R. I.; Khan, N. H. & Jasra, R. V (2004). Asymmetric epoxidation of non-functionalized alkenes using transition metal complexes. *Catalysis Surveys from Asia*, Vol.8, No.3, (September 2004), pp. 187-197, ISSN 1574-9266.

Finally, catalyst reuse was explored for both reactions. As can be observed in Tables 1 and 2, catalysts experienced a slight decrease in their initial catalytic activity through three consecutive runs in both reactions. The catalyst was segregated into a solid phase, while the reaction products remained in the liquid phase. This allowed the easy separation of catalyst and reaction products. On the other hand, Figure 8 shows the spectra of the catalyst used in three cycles. Clearly, it is observed that the main bands associated to the Mn(III) salen complexes are retained in the used catalysts. It indicates that the catalyst is very stable to the reaction conditions during three consecutive runs, whereas with oxidizing agents like NaOCl and *m*-CPBA the catalyst was deactivated in the first run (Figure 7). Therefore, the slight loss of catalytic activity with DMD is associated to physical loss of the catalyst during

**2000 1800 1600 1400 1200 1000 800 600 400**

Diepoxides were the main products from the oxidation of R-(+)-limonene, whereas a monoepoxide is obtained in the case of the catalytic oxidation of *cis*-ethyl cinnamate. In the latter case, a pure enantiomerically catalyst is required, while the Jacobsen racemic catalyst was sufficient in the case of the catalytic oxidation of R-(+)-limonene. Given that the catalytic activity of the three catalysts is very similar for R-(+)-limonene epoxidation and considering that the unique difference among the three catalysts is its stereogenic center located in the bond C1-C2 of the 1,2-diamino cyclohexane component, I conclude that the catalytic activity is not dependent on the stereogenic center. The catalyst could be recycled

Abdi, S. H. R.; Kureshy, R. I.; Khan, N. H. & Jasra, R. V (2004). Asymmetric epoxidation of

*Asia*, Vol.8, No.3, (September 2004), pp. 187-197, ISSN 1574-9266.

non-functionalized alkenes using transition metal complexes. *Catalysis Surveys from* 

Fig. 8. FT-IR Spectra of the (R,R) Jacobsen's catalyst used in various run: fresh catalyst (a),

**Wavenumbers/cm-1**

**a**

**480 cm-1**

**480 cm-1**

**480 cm-1**

**480 cm-1**

**550 cm-1**

**550 cm-1**

**550 cm-1**

**550 cm-1**

**b**

**c**

**d**

the isolation process rather than to oxidative degradation.

**1530 cm-1**

**1530 cm-1**

**1530 cm-1**

**1530 cm-1**

**T r a n s m i s s i o n / a . u**

second run (b) third run (c) fourth run.

three times without appreciable loss of its initial activity.

**4. Conclusion** 

**5. References** 


**8** 

*University of Liege* 

*Belgium* 

**Carbohydrate-Based Surfactants:** 

Hary Razafindralambo, Christophe Blecker and Michel Paquot

Carbohydrate-based surfactants (CBS) are, today, among the most important classes of amphiphilic compounds (Dembintsky, 2004; Queneau et al, 2008; Ruiz, 2009). Their structure is the result of the saccharide and lipid combination, naturally biosynthesized within living cells, or synthetically prepared by sequential reactions using carbohydrate and fatty materials, through one or several bonds. The growing interests of such compounds arise from many reasons of fundamental, practical, economical, and environmental orders (Razafindralambo et al, 2009, 2011a; Hill, K. & LeHen-Ferrenbach, 2009; Kitamoto et al, 2009). First, they can be easily prepared from the most abundant renewable vegetable raw materials (cellulose, pectin, hemicellulose, starch, etc.) in a wide range of structure and geometry by modular synthesis thanks to the presence of numerous reactive hydroxyl groups. Second, such a structural diversity makes them, on the one hand, as excellent models for getting insight into the surfactant mechanisms in modifying interfacial properties, which control the formation and the stability of colloidal systems such as micelles, vesicles, foams, emulsions, and suspensions (Razafindralambo et al, 2011b). One the other hand, numerous properties and functionalities would be expected from such a quasi-unlimited number of various compounds that can find specific applications in different industrial areas. Third, their compatibility to the environment, for instance, a higher biodegradability and lower toxicity, is an excellent criterion for their uses as alternatives to surfactants from petrochemical sources**.** Owing to the two former reasons, a systematic investigation of structure-activity relationships, which has rarely been carried out in the past, appears valuable for increasing knowledge on the impact of each CBS structural entity on their activities-functionalities, and ultimately, for achieving successfully a rational design in selecting and combining suitable compounds for further developments and applications. In the present chapter, we report the results of dynamic and equilibrium surface properties of homologous and analogous series of uronic acid derivatives, and evidence, consequently, the impact of different structural entities on their fundamental

The aim of the present contribution is: (1) to review CBS in terms of structural classification based on their molecular size (mono-, oligo-, polymeric surfactants), geometry (standard,

**1. Introduction** 

properties at the air-water interface.

**2. Scope of the contribution** 

**Structure-Activity Relationships** 

efficient catalysts for enantioselective epoxidation of nonfunctionalized alkenes. *J. Catal*, Vol.238, No.1, (January 2006), pp. 134-141, ISSN 0021-9517.


### **Carbohydrate-Based Surfactants: Structure-Activity Relationships**

Hary Razafindralambo, Christophe Blecker and Michel Paquot *University of Liege Belgium* 

#### **1. Introduction**

214 Advances in Chemical Engineering

Kureshy, R. I.; Khan, N-u. H.; Abdi, S. H. R.; Patel, S. T. & Jasra, R. V (2001). Dimeric chiral

Larrow, J. F. & Jacobsen, E. N (1994). A Practical process for the large-scale preparation of (R,R)-N,N'-bis(3,5-di-tert-butylsalicylidene)-1,2-cyclohexanediaminomanganese (III) chloride, a highly enantioselective epoxidation catalyst. *J. Org. Chem*, Vol.59,

Linde, C.; Anderlund, F. M. & Åkermark, B (2005). The effect of phenolates in the (salen)Mn-

Mahata, N.; Silva, A. R.; Pereira, M. F.; Freire, C.; de Castro, B. & Figueiredo, J. L (2007).

Murzin, D. Y.; Mäki-Arvela, P.; Toukoniitty, E. & T. Salmi (2005). Asymmetric

Porter, M. J. & Skidmore, J (2000). Asymmetric epoxidation of electron-deficient olefins.

Steiner, D.; Ivison, L.; Goralski, C. T.; Appell, R. B.; Gojkovic, J. R. & Singaram, B (2002). A

Wang, D.; Wang, M.; Zhang, R.; Wang, X.; Gao, A.; Ma, J. & Sun, L (2006). Asymmetric

Xia, Q. H.; Ge, H. Q. ; Ye, C. P.; Liu, Z. M. & Su, K. X (2005). Advances in homogeneous and

*Interface Sci*. Vol.311, No.1, (July 2007), pp. 152-158, ISSN 0021-9797. Mastrorilli, P. & Nobile, C. F (2004). Supported catalysts from polymerizable transition

*Catal*, Vol.238, No.1, (January 2006), pp. 134-141, ISSN 0021-9517.

No.7, (April 1994), pp. 1939–1940, ISSN 0022-3263.

(February 2007), pp. 175-256, ISSN 1520-5703.

pp. 2359-2363, ISSN: 0957-4166.

pp. 1603-1662, ISSN 0009-2665.

127, ISSN 0926-960X.

*Chem. Commun*, No.14, pp. 1215-1225, ISSN 1359-7345.

ISSN 0040-4039.

ISSN 0010-8545.

5597-5600, ISSN 0040-4039.

efficient catalysts for enantioselective epoxidation of nonfunctionalized alkenes. *J.* 

Mn(III) Schiff base complex-catalysed enantioselective epoxidation of nonfunctionalised alkenes. *Tetrahedron Lett*, Vol.42, No.15. (April 2001). pp. 2915-2918,

catalyzed epoxidation reactions. *Tetrahedron Lett*, Vol.46, No.33, (August 2005), pp.

Anchoring of a [Mn(salen)Cl] complex onto mesoporous carbon xerogels. *J Colloid* 

metal complexes. *Coord. Chem. Rev*. Vol.248, No.3-4, (February 2004), pp. 377-395,

heterogeneous catalysis: science and engineering. *Catal. Rev., Sci. Eng*, Vol.47, No.2,

facile and efficient method for the kinetic separation of commercially available *cis*and *trans*-limonene epoxide. *Tetrahedron asymmetry*, Vol.13, No.21, (October 2002),

epoxidation of styrene and chromenes catalysed by dimeric chiral (pyrrolidine salen)Mn(III) complexes. *Appl. Catal. A. Gen*, Vol.315, (November 2006), pp. 120-

heterogeneous catalytic asymmetric epoxidation. *Chem. Rev*, Vol.105, No.5, (2005),

Carbohydrate-based surfactants (CBS) are, today, among the most important classes of amphiphilic compounds (Dembintsky, 2004; Queneau et al, 2008; Ruiz, 2009). Their structure is the result of the saccharide and lipid combination, naturally biosynthesized within living cells, or synthetically prepared by sequential reactions using carbohydrate and fatty materials, through one or several bonds. The growing interests of such compounds arise from many reasons of fundamental, practical, economical, and environmental orders (Razafindralambo et al, 2009, 2011a; Hill, K. & LeHen-Ferrenbach, 2009; Kitamoto et al, 2009). First, they can be easily prepared from the most abundant renewable vegetable raw materials (cellulose, pectin, hemicellulose, starch, etc.) in a wide range of structure and geometry by modular synthesis thanks to the presence of numerous reactive hydroxyl groups. Second, such a structural diversity makes them, on the one hand, as excellent models for getting insight into the surfactant mechanisms in modifying interfacial properties, which control the formation and the stability of colloidal systems such as micelles, vesicles, foams, emulsions, and suspensions (Razafindralambo et al, 2011b). One the other hand, numerous properties and functionalities would be expected from such a quasi-unlimited number of various compounds that can find specific applications in different industrial areas. Third, their compatibility to the environment, for instance, a higher biodegradability and lower toxicity, is an excellent criterion for their uses as alternatives to surfactants from petrochemical sources**.** Owing to the two former reasons, a systematic investigation of structure-activity relationships, which has rarely been carried out in the past, appears valuable for increasing knowledge on the impact of each CBS structural entity on their activities-functionalities, and ultimately, for achieving successfully a rational design in selecting and combining suitable compounds for further developments and applications. In the present chapter, we report the results of dynamic and equilibrium surface properties of homologous and analogous series of uronic acid derivatives, and evidence, consequently, the impact of different structural entities on their fundamental properties at the air-water interface.

#### **2. Scope of the contribution**

The aim of the present contribution is: (1) to review CBS in terms of structural classification based on their molecular size (mono-, oligo-, polymeric surfactants), geometry (standard,

Carbohydrate-Based Surfactants: Structure-Activity Relationships 217

Bioglycolipids Structure Specificities

O

OH OH OH

**Low molecular mass** 

<sup>O</sup> <sup>O</sup> COOH O

**High molecular mass** 

(headgroup, tail, linkage)

nonionic, mono/di, ester, (Dembitsky, 2005)

nonionic, mono, ester, (Dembitsky, 2005)

nonionic, di, ester, (Dembitsky, 2005)

nonionic, mono, ester, (Kitamoto et al. 2009)

nonionic, di, ester, (Infante et al. 1997)

nonionic, mono, esteramide, (Dembitsky, 2005)

nonionic, multi, ester/amide, (Desai et Banat, 1997)

**3.1 Natural glycolipids** 

Rhamnolipids

Sophorolipids

Trehalolipids

Mannosylerythriol lipid

Monogalactosyldiacylglycerol

Cyclo-oligosaccharides

(Polymeric biosurfactant)

Emulsan

**3.1.1 Microbial biosurfactants** 

bipolar or bolaform, and gemini surfactants), and the nature of the polar headgroup (charged or not, cyclic or not), the apolar tail (number and length of alkyl chain), and the linker (amide, ester, ...) and/or the spacer; (2) to present systematically results on structureactivity relationships of uronic acid derivatives (UADs), a particular class of carbohydratebased surfactants. These concern the impact of each structural entity including the polar headgroup (stereochemistry), apolar tail (chain length, number, and unsaturation), and linkage/spacer, on the performance of UADs to change surface properties, and possibly, to form and stabilize colloidal systems.

### **3. Classification of carbohydrate-based surfactants**

Carbohydrate-based surfactants, currently known as glycolipids (Chester, 1997; Hato et al., 1999), are constituted by a saccharide unit (mono-, di-, oligo-, or polysaccharide) linked to a hydrophobic part of one, two, or multi-hydrocarbon chains by a single, or several bonds. These may be an ester, thioester, ether, amine, or/and amide group (Stubenrauch, 2001). Generally, glycolipids may be classified according to their amphiphilic structure, which depends on the polar headgroup, the apolar tail, but also the linkage between these main entities. Based on these three structural parameters, glycolipids are grouped in different categories, as presented in Table 1.


Table 1. Main classes of glycolipids, based on the amphiphilic structure and geometry

Besides the nature ionic, nonionic, and amphoteric of the polar headgroup, as for all surfactants, glycolipids can also be classified in small, medium, and large compounds, in relation to their molecular mass and size. According to these criteria, three main classes belonging to glycolipid-based monomeric, oligomeric, and polymeric surfactants may be distinguished. In addition to structural criteria, glycolipids may have two origins, those from microbial fermentation, crops, and animals, which are natural products, and those synthetized by chemical, enzymatic, and chemo-enzymatic routes, globally considered as synthetic compounds. In the following section, we try to illustrate by a few representative examples the most important classes of glycolipids with their chemical structure on the basis of their source, molecular mass, and geometry while considering the other classification aspects.

#### **3.1 Natural glycolipids**

216 Advances in Chemical Engineering

bipolar or bolaform, and gemini surfactants), and the nature of the polar headgroup (charged or not, cyclic or not), the apolar tail (number and length of alkyl chain), and the linker (amide, ester, ...) and/or the spacer; (2) to present systematically results on structureactivity relationships of uronic acid derivatives (UADs), a particular class of carbohydratebased surfactants. These concern the impact of each structural entity including the polar headgroup (stereochemistry), apolar tail (chain length, number, and unsaturation), and linkage/spacer, on the performance of UADs to change surface properties, and possibly, to

Carbohydrate-based surfactants, currently known as glycolipids (Chester, 1997; Hato et al., 1999), are constituted by a saccharide unit (mono-, di-, oligo-, or polysaccharide) linked to a hydrophobic part of one, two, or multi-hydrocarbon chains by a single, or several bonds. These may be an ester, thioester, ether, amine, or/and amide group (Stubenrauch, 2001). Generally, glycolipids may be classified according to their amphiphilic structure, which depends on the polar headgroup, the apolar tail, but also the linkage between these main entities. Based on these three structural parameters,

Glycolipids Hydrophilic part Hydrophobic part Linkers

Table 1. Main classes of glycolipids, based on the amphiphilic structure and geometry

Besides the nature ionic, nonionic, and amphoteric of the polar headgroup, as for all surfactants, glycolipids can also be classified in small, medium, and large compounds, in relation to their molecular mass and size. According to these criteria, three main classes belonging to glycolipid-based monomeric, oligomeric, and polymeric surfactants may be distinguished. In addition to structural criteria, glycolipids may have two origins, those from microbial fermentation, crops, and animals, which are natural products, and those synthetized by chemical, enzymatic, and chemo-enzymatic routes, globally considered as synthetic compounds. In the following section, we try to illustrate by a few representative examples the most important classes of glycolipids with their chemical structure on the basis of their source, molecular mass, and geometry while considering the other classification

One tail Two tails Multi tails

> Ester, thioester, Ether, amine, amide

form and stabilize colloidal systems.

Monocatenary Bicatenary Multicatenary

aspects.

**3. Classification of carbohydrate-based surfactants** 

glycolipids are grouped in different categories, as presented in Table 1.

One headgroup

Geminis One spacer/two tails

Alkylpolyglucosides Multi-headgroups One or multi-tails

Glycoglycerolipids Acylglycerol Glycosphingolipids Sphingoides Bolaforms Two headgroups One spacer

#### **3.1.1 Microbial biosurfactants**


Carbohydrate-Based Surfactants: Structure-Activity Relationships 219

Glu(n)-2-Glu(n) nonionic, bi, amides

Uronic Acid-Derivative Surfactants (UADs) are a particular class of monomeric glycolipids (Fig.1). Their polar headgroup represents a considerable part of carbohydrate components, widely distributed in natural plant polysaccharides (Langguth & Benet, 1992). In addition, uronic acids are the result of the primary alcohol oxydation into a carboxylic group within a monosaccharide like glucose and galactose. Thus, they contain, in their structure, both hydroxyl and carboxylic groups that are highly reactive, explaining their potential as a basic unit for generating closely related surfactant compounds. The polar headgroup configurations according to the stereochemistry of OH groups and geometry (cyclic or not, bipolarity), the hydrophobic tail (number and length of alkyl chain), and the type of linker (ester, ether and amide, etc) are among the main variables in their structural entities. Therefore, they represent a

set of ideal compounds for investigating structure-surface activity relationships.

L: NH or O R1: (CH2)*n*-CH3 R2: (CH2)*n*-CH3 R3: OH or H R4: H or OH

Fig. 1. General structure of uronic acid (left) and uronolactone (right) derivatives

**Bolaform** 

**Gemini** 

 α,-diamino-(oxa) [Spacer: -(CH2)n-]

**Oligomeric/polymeric**  Alkylpolyglucoside

Inulin carbamate

**4. Uronic acid derivatives** 

L

O R1

R3 OH R4

O

OH

O

R2

Glycolipids Structure Specificities (headgroup, tail, linkage)

nonionic, bi, amide (Wagenaar&Engberts, 2007)

(Sakai, 2008)

nonionic, mono, ether (Queneau et al. 2008)

nonionic, mono, ester (Tadros, 2004)

O

OH O RX

O

O

OH

R: (CH2)*<sup>n</sup>* X: OH or CH3


#### **3.1.2 Plant biosurfactants**

#### **3.1.3 Animal biosurfactants**


#### **3.2 Synthetic glycolipids**


Bioglycolipids Structure Specificities (headgroup, tail, linkage)

Bioglycolipids Structure Specificities (headgroup, tail, linkage)

Glycolipids Structure Specificities (headgroup, tail, linkage) **Monomeric** 

nonionic, multi, ester, (Dembitsky, 2005)

nonionic, multi, ester, (Dembitsky, 2005)

nonionic, di, amide, (Dembitsky, 2005)

cationic, di, amine, (Dembitsky, 2005)

nonionic, mono, ester, (Hill &LeHen- Ferrenbach, 2009)

nonionic, mono, ester (Hill &LeHen- Ferrenbach, 2009)

> nonionic, mono, amide (Laurent et al. 2011)

nonionic, di, ester-ether (Richel et al. 2010)

nonionic, mono, ether (Vulfson, 1990)

**3.1.2 Plant biosurfactants** 

 Butanolide glycosides

Sucrose ester

Cerebrosides

Gangliosides

**3.2 Synthetic glycolipids** 

Sorbitan monoester

 Isosorbide derivative

 Uronic acid derivatives: *a. Monocatenary* 

*b. Bicatenary* 

O O O OH OH

O OH

Alkylglucoside

**3.1.3 Animal biosurfactants** 


#### **4. Uronic acid derivatives**

Uronic Acid-Derivative Surfactants (UADs) are a particular class of monomeric glycolipids (Fig.1). Their polar headgroup represents a considerable part of carbohydrate components, widely distributed in natural plant polysaccharides (Langguth & Benet, 1992). In addition, uronic acids are the result of the primary alcohol oxydation into a carboxylic group within a monosaccharide like glucose and galactose. Thus, they contain, in their structure, both hydroxyl and carboxylic groups that are highly reactive, explaining their potential as a basic unit for generating closely related surfactant compounds. The polar headgroup configurations according to the stereochemistry of OH groups and geometry (cyclic or not, bipolarity), the hydrophobic tail (number and length of alkyl chain), and the type of linker (ester, ether and amide, etc) are among the main variables in their structural entities. Therefore, they represent a set of ideal compounds for investigating structure-surface activity relationships.

Fig. 1. General structure of uronic acid (left) and uronolactone (right) derivatives

Carbohydrate-Based Surfactants: Structure-Activity Relationships 221

representative experimental data, using appropriate mathematical models. All physical parameters such as critical micelle concentration (CMC), surface tension at the CMC (cmc), surface excess (Γmax), and minimum area per molecule(Amin), can be calculated using Gibbs

1

max <sup>1</sup> *Acmc <sup>N</sup>*

where *C*o is the bulk solution concentration; R = 8.314 J.K-1.m-2, the gas constant; T, the

By Langmuir approach, a small volume of surfactant organic solution is spread drop wise onto the clean milli-Q water surface by means of Hamilton micro-syringe using a reproducibility adapter. After a given laps time, allowing the solvent evaporation, the surface pressure (π) of the monolayer is determined by measuring the horizontal force per unit of length, for a delimited area, during the compression. With an automatic Langmuir film Waage LFW2 3″5 (Lauda, Königshofen, Germany), a total trough area of 927 cm² is available. This is progressively reduced by moving a single barrier made in Teflon, and the force per length unit detected at a float gives the surface pressure of the monolayer at a constant temperature. For characterizing uronic acid derivatives, mainly bicatenary surfactants, the research of optimized

The effect of the octyl glucuronolactone anomeric form is observed for the equilibrium surface properties (Table 2). The anomer α CMC is about the half of that of the anomer β, all other related parameters being quite similar. cmc and Amin are quite similar, indicating that the anomeric form does not affect the molecular arrangement at the air-water interface, but changes that in the bulk water. In other words, the anomer α self-aggregates readily in water while adopting the same close-packed configuration at the air-water interface than the anomer β. Therefore, the anomer α appears more hydrophobic than the anomer β, which is attributed to the difference in the steric effect of their polar headgroup. This interacts stronger with water molecules for the anomer β, delaying its self-aggregation process. Concerning their dynamic surface properties, the anomer β appears slightly more efficient than the anomer α (data not shown), the difference being less important than that observed

ln *<sup>o</sup> d RT d C* 

(1)

(2)

max

temperature in Kelvin, and *N,* the Avogadro's number.

experimental conditions is required (Razafindralambo et al. 2011).

**6. Structure-activity relationships** 

**6.1 Polar headgroup effect** 

for equilibrium surface properties.

**6.1.1 Stereochemistry** 

**Anomeric forms** 

equations for nonionic surfactants :

**5.2.2 Langmuir approach** 

and

Among several activities that CBS are able to develop (Ruiz, 2009), those related to the modification of surface and interface fundamental properties are undeniably the most important to evaluate and understand. Through several research works during last decades, it is established that fundamental properties such as dynamic and equilibrium surface tensions play a key role in the formation and stability of colloidal systems like foams, emulsions, and suspensions (Razafindralambo et al. 2011b). In this chapter, we are only focused on the relationships between chemical structure of UADs and their surface-activities at the air-water interface. UADs are a set of glucuronic and galacturonic acid-based surfactants varying in the polar headgroup configuration, including a cyclic (lactone bond) or non-cyclic structure and a α- or β- anomeric form, the hydrophobic tail chain length (C8 to C14) and number (mono- and bicatenary), the presence of a double bond, as well as an OH group at the terminal carbon, and in the type of linkage, ester, amide in C6, or ether in C1. These compounds have been synthesized by chemical or enzymatic routes, purified, and their chemical structure has been confirmed by means of various spectroscopic techniques (Nuclear Magnetic Resonance, Infrared, and Mass Spectrometry).

#### **5. Surface-active properties**

Surface-active properties of any surfactants can be measured through their capacity, either to reduce the surface tension of aqueous solution (o) to any value (), or to increase the surface pressure (π = o-) of the air-water interface. Measurements can be performed in both dynamic and static/quasi-static modes. Based on the solubility of the coumpound to characterize, one of the two following approaches must be used. The Gibbs'approach, measuring by adsorption, i.e. by migration of surfactant to the surface, is more convenient for short chain derivative surfactants (high solubility in water) whereas the Langmuir approach, determining π after spreading and compressing surfactant molecules at the surface, is rather appropriate for long chain derivative surfactants (low solubility in water). Surface activities of uronic acid derivatives may be evaluated in both dynamic and static/quasi-satic modes using a series of complementary techniques at room temperature comprised between 20-25°C (Razafindralambo et al., 1995; 2009; 2011).

#### **5.1 Dynamic surface properties**

In dynamic mode, surface tensions (dyn) are measured using a drop volume tensiometer (TVT1, Lauda, Germany) by determining the critical volume (maximum pending size) of a series of drop created at different growing rates from 3 to 120 s (Razafindralambo et al., 2004).

#### **5.2 Equilibrium surface properties**

#### **5.2.1 Gibbs approach**

In static and quasi-static modes, the equilibrium surface tensions (e) are measured by means of Wilhelmy plate (Tensimat N3, Prolabo, France), drop volume (TVT1, Lauda, Germany), or pending drop (Tracker, IT-Concept, France)-based tensiometers as detailed in previous papers (Razafindralamboet al., 1995; 2009). These methods consist of measuring continuously e as a function of time, for surfactant solutions at different concentrations. The equilibrium surface tension (e) can be determined as the fitting value of from representative experimental data, using appropriate mathematical models. All physical parameters such as critical micelle concentration (CMC), surface tension at the CMC (cmc), surface excess (Γmax), and minimum area per molecule(Amin), can be calculated using Gibbs equations for nonionic surfactants :

$$\Gamma\_{\text{max}} = -\frac{1}{RT} \left( \frac{d\mathcal{Y}}{d\ln \mathcal{C}\_o} \right) \tag{1}$$

and

220 Advances in Chemical Engineering

Among several activities that CBS are able to develop (Ruiz, 2009), those related to the modification of surface and interface fundamental properties are undeniably the most important to evaluate and understand. Through several research works during last decades, it is established that fundamental properties such as dynamic and equilibrium surface tensions play a key role in the formation and stability of colloidal systems like foams, emulsions, and suspensions (Razafindralambo et al. 2011b). In this chapter, we are only focused on the relationships between chemical structure of UADs and their surface-activities at the air-water interface. UADs are a set of glucuronic and galacturonic acid-based surfactants varying in the polar headgroup configuration, including a cyclic (lactone bond) or non-cyclic structure and a α- or β- anomeric form, the hydrophobic tail chain length (C8 to C14) and number (mono- and bicatenary), the presence of a double bond, as well as an OH group at the terminal carbon, and in the type of linkage, ester, amide in C6, or ether in C1. These compounds have been synthesized by chemical or enzymatic routes, purified, and their chemical structure has been confirmed by means of various spectroscopic techniques

Surface-active properties of any surfactants can be measured through their capacity, either to reduce the surface tension of aqueous solution (o) to any value (), or to increase the surface pressure (π = o-) of the air-water interface. Measurements can be performed in both dynamic and static/quasi-static modes. Based on the solubility of the coumpound to characterize, one of the two following approaches must be used. The Gibbs'approach, measuring by adsorption, i.e. by migration of surfactant to the surface, is more convenient for short chain derivative surfactants (high solubility in water) whereas the Langmuir approach, determining π after spreading and compressing surfactant molecules at the surface, is rather appropriate for long chain derivative surfactants (low solubility in water). Surface activities of uronic acid derivatives may be evaluated in both dynamic and static/quasi-satic modes using a series of complementary techniques at room temperature

In dynamic mode, surface tensions (dyn) are measured using a drop volume tensiometer (TVT1, Lauda, Germany) by determining the critical volume (maximum pending size) of a series of drop created at different growing rates from 3 to 120 s (Razafindralambo et al.,

In static and quasi-static modes, the equilibrium surface tensions (e) are measured by means of Wilhelmy plate (Tensimat N3, Prolabo, France), drop volume (TVT1, Lauda, Germany), or pending drop (Tracker, IT-Concept, France)-based tensiometers as detailed in previous papers (Razafindralamboet al., 1995; 2009). These methods consist of measuring continuously e as a function of time, for surfactant solutions at different concentrations. The equilibrium surface tension (e) can be determined as the fitting value of from

(Nuclear Magnetic Resonance, Infrared, and Mass Spectrometry).

comprised between 20-25°C (Razafindralambo et al., 1995; 2009; 2011).

**5. Surface-active properties** 

**5.1 Dynamic surface properties** 

**5.2 Equilibrium surface properties** 

**5.2.1 Gibbs approach** 

2004).

$$A\_{cmc} = \left(\frac{1}{\Gamma\_{\text{max}}N}\right) \tag{2}$$

where *C*o is the bulk solution concentration; R = 8.314 J.K-1.m-2, the gas constant; T, the temperature in Kelvin, and *N,* the Avogadro's number.

#### **5.2.2 Langmuir approach**

By Langmuir approach, a small volume of surfactant organic solution is spread drop wise onto the clean milli-Q water surface by means of Hamilton micro-syringe using a reproducibility adapter. After a given laps time, allowing the solvent evaporation, the surface pressure (π) of the monolayer is determined by measuring the horizontal force per unit of length, for a delimited area, during the compression. With an automatic Langmuir film Waage LFW2 3″5 (Lauda, Königshofen, Germany), a total trough area of 927 cm² is available. This is progressively reduced by moving a single barrier made in Teflon, and the force per length unit detected at a float gives the surface pressure of the monolayer at a constant temperature. For characterizing uronic acid derivatives, mainly bicatenary surfactants, the research of optimized experimental conditions is required (Razafindralambo et al. 2011).

#### **6. Structure-activity relationships**

#### **6.1 Polar headgroup effect**

#### **6.1.1 Stereochemistry**

#### **Anomeric forms**

The effect of the octyl glucuronolactone anomeric form is observed for the equilibrium surface properties (Table 2). The anomer α CMC is about the half of that of the anomer β, all other related parameters being quite similar. cmc and Amin are quite similar, indicating that the anomeric form does not affect the molecular arrangement at the air-water interface, but changes that in the bulk water. In other words, the anomer α self-aggregates readily in water while adopting the same close-packed configuration at the air-water interface than the anomer β. Therefore, the anomer α appears more hydrophobic than the anomer β, which is attributed to the difference in the steric effect of their polar headgroup. This interacts stronger with water molecules for the anomer β, delaying its self-aggregation process. Concerning their dynamic surface properties, the anomer β appears slightly more efficient than the anomer α (data not shown), the difference being less important than that observed for equilibrium surface properties.

Carbohydrate-Based Surfactants: Structure-Activity Relationships 223

Fig. 3. Dynamic surface tension vs. time of dodecyl glucuronolactones (0.2 g/L), with and

0 20 40 60 80

Drop age [s]

CMC [mM]

GlcA-lactone-C12-OH GlcA-lactone-C12

> 0.7 0.6

Amin [Ų/mol]

> 65 34

ϒcmc [mN/m]

> 43.9 47.2

equilibrium (long time) surface properties, based on their CMC and cmc values.

Table 3. Physical parameters of dodecyl glucuronolactones ending with, and without

It increases the area occupied per molecule (cross-sectional area), and the performance in the adsorption time by reducing faster the dynamic surface tension. This difference is easy to understand, and may be explained by the fact that the two polar headgroups are directed into the aqueous phase, leaving the hydrophobic spacer in the gaseous phase, and forming, as a consequence, a "convex" configuration. In contrast, it has no significant effect on the

The impact of the alkyl chain length on all of interfacial properties can be evaluated with a homologous series of monocatenary and bicatenary uronic acid derivatives. As for all monomeric surfactants, whatever their nature, the general trend is respected with uronic acid derivatives, that is, the longer the alkyl chain length, the lower the CMC of the uronate and uronamide derivatives (Blecker et al., 2002; Laurent et al., 2011). About dynamic surface properties, an optimum chain length of eleven carbon atom gives the best performance, regarding the adsorption time and the maximum reduction of the surface tension. Concerning

without a hydroxyl group at the end, measured by TVT1 at 20°C

OH-C12-Glucuronolactone C12-Glucuronolactone

58

60

62

64

66

Dynamic Surface Tension [mN/m]

68

70

72

74

**6.2 Hydrophobic effect 6.2.1 Alkyl chain length** 

hydroxyl group


Table 2. Physical parameters of octyl glucuronolactone with α and β anomers

#### **4-hydroxyl group (4-OH) axial or equatorial position effect**

The second parameter related to the polar headgroup stereochemistry is the axial or equatorial position of the hydroxyl group of the fourth carbon within bicatenary derivatives of galacturonic and glucuronic acids. This stereochemistry effect impacts their configuration and behaviour at the air-water interface, as observed with the (-A) isotherms (Fig.2). The glucuronic acid derivative (GlcA-C14/14) is more expanded when it is spread at the air-water around 20°C, and shows a transition phase under compression, which does not occur for the galacturonic acid derivative (GalA-C14/14). This difference is in agreement with film morphologies and thicknesses, and is also supported by molecular models (Razafindralambo et al., 2011). In contrast, the 4-OH axial or equatorial position has no effect when the molecules are vertically oriented, i.e. within a film at a condensed state. Such results have been attributed to the configuration of two alkyl chains, which is in "open" or "close" structure according to the film is in expanded or in condensed state.

Fig. 2. (-A) isotherms of GalA-C14/14 and GlcA-C14/14 spread at the air-water interface 20°C

#### **6.1.2 Impact of the bipolarity: (OH group ending the alkyl chain)**

Adding an OH group at the end of the alkyl chain changes the geometry of uronic acid derivatives, which becomes a bipolar asymmetric. Such a modification impacts some dynamic (Fig.3) and equilibrium surface properties (Table 3) of glucuronic acid derivatives.

The second parameter related to the polar headgroup stereochemistry is the axial or equatorial position of the hydroxyl group of the fourth carbon within bicatenary derivatives of galacturonic and glucuronic acids. This stereochemistry effect impacts their configuration and behaviour at the air-water interface, as observed with the (-A) isotherms (Fig.2). The glucuronic acid derivative (GlcA-C14/14) is more expanded when it is spread at the air-water around 20°C, and shows a transition phase under compression, which does not occur for the galacturonic acid derivative (GalA-C14/14). This difference is in agreement with film morphologies and thicknesses, and is also supported by molecular models (Razafindralambo et al., 2011). In contrast, the 4-OH axial or equatorial position has no effect when the molecules are vertically oriented, i.e. within a film at a condensed state. Such results have been attributed to the configuration of two alkyl chains, which is in "open" or

Fig. 2. (-A) isotherms of GalA-C14/14 and GlcA-C14/14 spread at the air-water interface 20°C

0 25 50 75 100 125

Area (Ų/molecule)

Adding an OH group at the end of the alkyl chain changes the geometry of uronic acid derivatives, which becomes a bipolar asymmetric. Such a modification impacts some dynamic (Fig.3) and equilibrium surface properties (Table 3) of glucuronic acid derivatives.

**6.1.2 Impact of the bipolarity: (OH group ending the alkyl chain)** 

CMC [mM]

> 1.3 2.4

GalA-C14/14 (20.3°C) GlcA-C14/14 (20.6°C)

Amin [Ų/mol]

> 43 40

ϒcmc [mN/m]

> 32.8 35.2

Table 2. Physical parameters of octyl glucuronolactone with α and β anomers

"close" structure according to the film is in expanded or in condensed state.

**4-hydroxyl group (4-OH) axial or equatorial position effect**

α-C8- Glucuronolactone β-C8- Glucuronolactone

0

10

20

30

40

Surface pressure (mN/m)

50

60

70

80

Fig. 3. Dynamic surface tension vs. time of dodecyl glucuronolactones (0.2 g/L), with and without a hydroxyl group at the end, measured by TVT1 at 20°C


Table 3. Physical parameters of dodecyl glucuronolactones ending with, and without hydroxyl group

It increases the area occupied per molecule (cross-sectional area), and the performance in the adsorption time by reducing faster the dynamic surface tension. This difference is easy to understand, and may be explained by the fact that the two polar headgroups are directed into the aqueous phase, leaving the hydrophobic spacer in the gaseous phase, and forming, as a consequence, a "convex" configuration. In contrast, it has no significant effect on the equilibrium (long time) surface properties, based on their CMC and cmc values.

#### **6.2 Hydrophobic effect**

#### **6.2.1 Alkyl chain length**

The impact of the alkyl chain length on all of interfacial properties can be evaluated with a homologous series of monocatenary and bicatenary uronic acid derivatives. As for all monomeric surfactants, whatever their nature, the general trend is respected with uronic acid derivatives, that is, the longer the alkyl chain length, the lower the CMC of the uronate and uronamide derivatives (Blecker et al., 2002; Laurent et al., 2011). About dynamic surface properties, an optimum chain length of eleven carbon atom gives the best performance, regarding the adsorption time and the maximum reduction of the surface tension. Concerning

Carbohydrate-Based Surfactants: Structure-Activity Relationships 225

GlcA-O-C14 (24.6°C) GlcA-C14/14 (26.7°C) GlcA-lactone-C14 (25.2°C)

> GlcA-O-C11:0 GlcA-O-C11:1

Fig. 5. (π-A) isotherms of GlcA-O-C14, GlcA-lactone-C14, and GlcA-C14/14 spread at the air-

0 25 50 75 100 125

Area (Ų/molecule)

Fig. 6. Dynamic surface tension vs. time of undecanyl (C11:0) and undecenyl (C11:1)-

One of the spectacular effects of the linkage between the polar headgroup and the hydrophobic tail is the change of the ester bond orientation within the octyl sugar

0 10 20 30 40 50

Drop age [s]

glucuronolactones ( 0.1 g/L) measured by TVT1 at 20°C.

water interface 25°C.

Dynamic Surface Tension [mN/m]

Surface pressure (mN/m)

**6.3 Linkage effect** 

**6.3.1 Ester bond direction** 

bicatenary derivatives, the comparison of GalA-C10/10 and GalA-C14/14 monolayer properties shows that the shorter the two alkyl chains (C10/10), the more expanded and compressible the spread molecules, and the less stable the molecular film (Fig.4). Such behaviour of the shorter chain derivatives is attributed to stronger interactions with the subphase, which is comparable with the case of the dimyristoyl-phosphatidylcholine (DMPC), an ionic phospholipid having the same alkyl chains than that of GalA-C14/14 (data not shown).

Fig. 4. (π-A) isotherms of GalA-C10/10 and GlcA-C14/14 spread at the air-water interface

#### **6.2.2 Alkyl chain number (Monocatenary vs. bicatenary)**

The alkyl chain number of uronic acid derivatives also appears crucial on their surface properties characterized by the Langmuir approach. When a second alkyl chain is added to the polar headgroup, that is, the compound becomes bicatenary, the mechanical properties of the film at the air-water interface are improved, regarding its collapse pressure (Fig.5). It is to be noted, from this figure, that the inclusion of a cyclic ester bond into the polar headgroup affects slightly the behaviour of the monolayer.

#### **6.2.3 Unsaturation**

By including a double bond at the end of the alkyl chain, no significant effect has been observed on the equilibrium surface properties, the cmc, CMC, and Amin values being similar and equal to 34.2 ± 1.1 mN/m, 0.12 ± 0.01 g/L, and 39.0 ± 2.0 Ų/mol., respectively. In contrast, it impacts extensively the dynamic surface properties of the undecanoyl glucuronate (Fig.6). The derivative with an unsaturated chain migrates faster at the airwater interface, and reduces readily the dynamic surface tension, compared to the saturated one. It is attributed to the "shortening effect" of the double bond inclusion (Milkereit et al. 2005), reducing the hydrophobicity of the surfactant, and therefore, its adsorption time.

bicatenary derivatives, the comparison of GalA-C10/10 and GalA-C14/14 monolayer properties shows that the shorter the two alkyl chains (C10/10), the more expanded and compressible the spread molecules, and the less stable the molecular film (Fig.4). Such behaviour of the shorter chain derivatives is attributed to stronger interactions with the subphase, which is comparable with the case of the dimyristoyl-phosphatidylcholine (DMPC), an ionic phospholipid having

> GalA-C10/10 (22.0°C) GalA-C14/14 (23.2°C)

Fig. 4. (π-A) isotherms of GalA-C10/10 and GlcA-C14/14 spread at the air-water interface

The alkyl chain number of uronic acid derivatives also appears crucial on their surface properties characterized by the Langmuir approach. When a second alkyl chain is added to the polar headgroup, that is, the compound becomes bicatenary, the mechanical properties of the film at the air-water interface are improved, regarding its collapse pressure (Fig.5). It is to be noted, from this figure, that the inclusion of a cyclic ester bond into the polar

0 25 50 75 100 125

Area (Ų/molecule)

By including a double bond at the end of the alkyl chain, no significant effect has been observed on the equilibrium surface properties, the cmc, CMC, and Amin values being similar and equal to 34.2 ± 1.1 mN/m, 0.12 ± 0.01 g/L, and 39.0 ± 2.0 Ų/mol., respectively. In contrast, it impacts extensively the dynamic surface properties of the undecanoyl glucuronate (Fig.6). The derivative with an unsaturated chain migrates faster at the airwater interface, and reduces readily the dynamic surface tension, compared to the saturated one. It is attributed to the "shortening effect" of the double bond inclusion (Milkereit et al. 2005), reducing the hydrophobicity of the surfactant, and therefore, its adsorption time.

**6.2.2 Alkyl chain number (Monocatenary vs. bicatenary)** 

headgroup affects slightly the behaviour of the monolayer.

**6.2.3 Unsaturation** 

0

10

20

30

40

Surface pressure (mN/m)

50

60

70

80

the same alkyl chains than that of GalA-C14/14 (data not shown).

Fig. 5. (π-A) isotherms of GlcA-O-C14, GlcA-lactone-C14, and GlcA-C14/14 spread at the airwater interface 25°C.

Fig. 6. Dynamic surface tension vs. time of undecanyl (C11:0) and undecenyl (C11:1) glucuronolactones ( 0.1 g/L) measured by TVT1 at 20°C.

#### **6.3 Linkage effect**

#### **6.3.1 Ester bond direction**

One of the spectacular effects of the linkage between the polar headgroup and the hydrophobic tail is the change of the ester bond orientation within the octyl sugar

Carbohydrate-Based Surfactants: Structure-Activity Relationships 227

In the present chapter, we contribute to a better understanding of the structure-surface activity relationships of uronic acid derivatives, a promising class of carbohydrate-based surfactants. Each structural element impacts either their dynamic performances, measured over a short period range, or their equilibrium activities, evaluated after a longer period. Besides scientific interests of such fundamental information, the approach also leads to the identification of some suitable structures for practical performances in forming and stabilizing colloidal systems like foams, emulsions, and suspensions, which are encountered

This work carried out in the framework of the "TECHNOSE" excellence Program was supported by Belgian Walloon Region. The authors thank Drs Aurore Richel, Gaëtan Richard, Pascal Laurent for the syntheses and characterization of uronic acid derivatives,

Blecker, C.; Piccicuto, S.; Lognay, G.; Deroanne, C.; Marlier, M. & Paquot, M. (2002).

Boyd, B. J.; Drummond, C.; Krodkiewska, I. & Grieser, F. (2000). How Chain Length,

Dembitsky, V. (2004). Astonishing diversity of natural surfactants: 1. Glycosides of fatty

Dembitsky, V. (2005). Astonishing diversity of natural surfactants: 4. Fatty acid amide

Desai, J.& Banat, I. (1997). Microbial production of surfactants and their commercial

Hato, M.; Minamikawa, H. ; Tamada, K. ; Baba, T. & Tanabe, Y. (1999). Self-assembly of

Hill, K. & LeHen-Ferrenbach, C. (2009). Sugar-Based Surfactants for Consumer Products and

Infante, M. ; Pinazo, A.& Seguer, J. (1997). Non-conventional surfactants from amino acids

Kitamoto, D.; Morita, T. ; Fukuoka, T. ; Konishi, M. & Imura, T. (2009). Self-assembling

synthetic glycolipid/water systems. *Advances in Colloid and Interface Science,* 80, 233-

Technical Applications, In: *Sugar-based surfactants fundamentals and Applications*,

and glycolipids: Structure, preparation and properties. *Colloids and Surfaces A:* 

properties of glycolipid biosurfactants and their potential applications. *Current* 

glycosides, their analogs and derivatives. *Lipids,* 40, 641-660.

potential. *Microbiology and Molecular Biology Reviews,* 61, 47-64.

C.C. Ruiz (Ed.), 1-20, CRC Press, ISBN 978-1-4200-5166-7.

*Physicochemical and Engineering Aspects,* 123-124, 49-70.

*Opinion in Colloid & Interface Science*14, 315-328.

Enzymatically prepared *n*-alkyl esters of glucuronic acid: the effect of hydrophobic chain length on surface properties.*Journal of colloid and interface science*, 247, 424-428.

Headgroup Polymerization, and Anomeric Configuration Govern the Thermotropic and Lyotropic Liquid Crystalline Phase Behavior and the Air-Water Interfacial Adsorption of Glucose-Based Surfactants. *Langmuir, 16*, 7359-7367 Chester, A. (1997). Nomenclature of glycolipides. *Pure and Applied Chemistry,* 69, 2475-2487. Dembitsky, V. (2005). Astonishing diversity of natural surfactants: 3. Carotenoid glycosides

**7. Conclusion** 

virtually in all soft manufactured products.

and Mr. Alexandre Schandeler for technical assistance.

and isoprenoid glycolipids. *Lipids,* 40, 535-557.

acids and alcohols. *Lipids,* 39, 933-953.

**8. Acknowledgments** 

**9. References** 

270.

derivatives. By comparing the dynamic and equilibrium surface properties of glucose octanoate (C8-**CO**-O-Glc) and octyl glucuronate (C8-O-**CO**-GlcA), significant differences are observed (Table 4). When the carbonyl group is closed to the polar headgroup constituted by glucuronic acid (C8-O-**CO**-GlcA), the surfactant is more hydrophobic, and the polar head group is larger. The consequence is the increase of dynamic properties, and the decrease of the CMC (10.7 mM vs. 19.1 mM). When the carbonyl group is close to the alkyl chain (C8- **CO**-O-Glc), there is an "interruption" of the hydrophobic tail, which reduces its hydrophobicity (Razafindralambo et al. 2009). The molecular area becomes smaller and the CMC is higher. On the other hand, the adsorption time rate characterized by t\* and vmax are increased, reducing its performance on dynamic surface properties.


Table 4. Physical parameters of glucose octanoate and octyl glucuronate; (1) Compared at 1.63 mM (25°C).

#### **6.3.2 Ester vs. amide bond**

By comparing the surface properties of three derivatives with an octyl chain, the impact of the type of the linkage bond can be evidenced, despite the difference in the ratio alpha/beta of anomeric forms (Table 5). It appears that the uronamide is the most effective in terms of equilibrium surface tensions. This result may be attributed to the difference in the molecular area occupied by the two octyl derivatives. The amide bond has a smaller bulk size than the ester ones. In comparison with a unique form of octyl glucoside, the glucuronamide is more effective. Even though the linkage is not at the same position, on the C1 for glucoside, and on the C6 for glucuronamide, the impact may be significant, considering the similarity in the molecular mass, and the number of the OH group available.


Table 5. Physical parameters of glucuronic acid and glucose derivatives containing different linkages at the air-water interface. (1) By modeling; (2) by linear fit

#### **7. Conclusion**

226 Advances in Chemical Engineering

derivatives. By comparing the dynamic and equilibrium surface properties of glucose octanoate (C8-**CO**-O-Glc) and octyl glucuronate (C8-O-**CO**-GlcA), significant differences are observed (Table 4). When the carbonyl group is closed to the polar headgroup constituted by glucuronic acid (C8-O-**CO**-GlcA), the surfactant is more hydrophobic, and the polar head group is larger. The consequence is the increase of dynamic properties, and the decrease of the CMC (10.7 mM vs. 19.1 mM). When the carbonyl group is close to the alkyl chain (C8- **CO**-O-Glc), there is an "interruption" of the hydrophobic tail, which reduces its hydrophobicity (Razafindralambo et al. 2009). The molecular area becomes smaller and the CMC is higher. On the other hand, the adsorption time rate characterized by t\* and vmax are

[Glc-O-C8]

36.1 0.14 0.024

25.1 19.1 37

By comparing the surface properties of three derivatives with an octyl chain, the impact of the type of the linkage bond can be evidenced, despite the difference in the ratio alpha/beta of anomeric forms (Table 5). It appears that the uronamide is the most effective in terms of equilibrium surface tensions. This result may be attributed to the difference in the molecular area occupied by the two octyl derivatives. The amide bond has a smaller bulk size than the ester ones. In comparison with a unique form of octyl glucoside, the glucuronamide is more effective. Even though the linkage is not at the same position, on the C1 for glucoside, and on the C6 for glucuronamide, the impact may be significant, considering the similarity in the

Table 5. Physical parameters of glucuronic acid and glucose derivatives containing different

C8-glucuronamide [C6-Amide linkage]

α/ β=0.6

24.0 3.3 25 ; 35(1)

Table 4. Physical parameters of glucose octanoate and octyl glucuronate; (1) Compared at

Octylglucuronate [GlcA-O-C8]

> 3.4 0.56 0.0081

> > 28.0 10.7 44

C8-glucoside (25°C) [C1-Ether linkage] α ; β [Boyd et al, 2000]

> 35.2 ; 31.3 12.0; 18.2 42 ; 42(2)

increased, reducing its performance on dynamic surface properties.

Parameters Glucose octanoate

molecular mass, and the number of the OH group available.

C8-glucuronate [C6-Ester linkage]

α/ β=1.5

28.0 10.7 44 ; 45(1)

linkages at the air-water interface. (1) By modeling; (2) by linear fit

*Dynamics* (1) : t\* (10-3 s) *n* 

vmax(mN/m/s)

1.63 mM (25°C).

**6.3.2 Ester vs. amide bond** 

Physical parameters

ϒcmc (mN/m) CMC (mM) Amin (Ų/mol)

*Equilibrium* : ϒcmc (mN/m) CMC (mM) Amin (Ų/mol) In the present chapter, we contribute to a better understanding of the structure-surface activity relationships of uronic acid derivatives, a promising class of carbohydrate-based surfactants. Each structural element impacts either their dynamic performances, measured over a short period range, or their equilibrium activities, evaluated after a longer period. Besides scientific interests of such fundamental information, the approach also leads to the identification of some suitable structures for practical performances in forming and stabilizing colloidal systems like foams, emulsions, and suspensions, which are encountered virtually in all soft manufactured products.

#### **8. Acknowledgments**

This work carried out in the framework of the "TECHNOSE" excellence Program was supported by Belgian Walloon Region. The authors thank Drs Aurore Richel, Gaëtan Richard, Pascal Laurent for the syntheses and characterization of uronic acid derivatives, and Mr. Alexandre Schandeler for technical assistance.

#### **9. References**


**9** 

*Curtin University* 

*Australia* 

**CO2 Biomitigation and Biofuel Production** 

Increased concerns about global warming and greenhouse gas emissions as well as the exhaustion of easily accessible fossil fuel resources are calling for effective carbon dioxide (CO2) mitigation technologies and clean and renewable energy sources. One of the major gases leading to global warming is carbon dioxide. CO2 makes up 68% of the estimated total

There have been several approaches proposed for managing the levels of CO2 emitted into the atmosphere, including ocean sequestration such as deep ocean injection or increasing the amount of CO2 dissolved in the ocean. Another proposed form of sequestration is to sequester the CO2 into terrestrial ecosystems (Halmann, 1993). The short term options of sequestration by direct injection into geologic or oceanic sinks are recognized as methods to reduce the CO2 levels but do not address issues of sustainability (Stewart & Hessami, 2005). Carbon sequestration can also be accomplished through chemical approaches; some problems with these approaches are that they must be safe for the environment, stable for long- term storage, and cost - competitive to other sequestration options. Other technologies have been considered, such as chemical absorption, membrane separation, cryogenic fractionation and adsorption using molecular sieves, but they are even less energy efficient

One of the most understudied methods for CO2 mitigation is the use of biological processes (via microalgae) in a direct CO2 to biomass conversion from point source emissions of CO2 in engineered systems such as photobioreactors. Microalgal biofixation of carbon dioxide (CO2) in photobioreactors has recently gained renewed interest as a promising strategy for CO2 mitigation. The use of photobioreactors for microalgal CO2 sequestration offers the principal advantages of increased microalgae productivity, owing to controlled environmental conditions, and optimized space/volume utilization and, thus, more efficient use of costly land. In fact, the photosynthetic solution when scaled up would present a far superior and sustainable solution under both environmental and economic considerations. Fig. 1 shows the importance of the microalgae photobioreactor and its general applications, microalgae used to capture waste CO2 utilizing the nutrients in wastewater and natural

greenhouse gas emissions (Harrington & Foster, 1999).

as to be considered economically viable (Stewart & Hessami, 2005).

**1. Introduction** 

**Using Microalgae: Photobioreactors** 

**Developments and Future Directions** 

Hussein Znad, Gita Naderi, H.M. Ang and M.O. Tade


### **CO2 Biomitigation and Biofuel Production Using Microalgae: Photobioreactors Developments and Future Directions**

Hussein Znad, Gita Naderi, H.M. Ang and M.O. Tade *Curtin University Australia* 

#### **1. Introduction**

228 Advances in Chemical Engineering

Laurent, P.; Razafindralambo, H. ; Wathelet, B. ; Blecker, C.; Wathelet, J. & Paquot, M.

Milkereit, G.; Garamus,V.; Veermans, K.; Willumeit, R. &Vill,V. (2005). Structures of micelles

Queneau, Y.; Chambert, S. ; Besset, C. & Cheaib, R. (2008). Recent progress in the synthesis

Razafindralambo, H.; Richel, A.; Wathelet, B.; Blecker, C.; Wathelet, J.; Brasseur, R.; Lins, L.;

Razafindralambo,H.; Blecker, C.& Paquot, M.(2011b). Screening of Basic Properties of

*and Applications*, Nagarajan, R., Ed. ACS: Washignton; Vol. 1070, p53-66. Razafindralambo, H.; Blecker, C.; Mezdour, S.; Deroanne, C.; Crowet, J.; Brasseur, R.; Lins,

Razafindralambo, H.; Thonart, P. & Paquot, M. (2004). Dynamic and equilibrium surface

Razafindralambo, H.; Blecker, C.; Delhaye, S.& Paquot, M. (1995). Application of the Quasi-

Ruiz, C. (2009). Sugar-Based Surfactants Fundamentals and Applications. CRC Press.

Sakai, K.; Tamura, M.; Umezawa, S.; Takamatsu, Y.; Torigoe, K.; Yoshimura, T.; Esumi, K.;

Stubenrauch, C. (2001). Sugar surfactants - Aggregation, interfacial, and adsorption phenomena. *Current Opinion in Colloid and Interface Science,* 6, 160-170. Tadros, T.; Vandamme, A.; Levecke, B.; Booten, K. & Stevens, C. (2004). Stabilization of

Vulfson, E.; Patel, R. & Law, B. (1990). Alkyl-α-glucoside synthesis in a water-organic two-

Wagenaar, A.& Engberts, J. (2007). Synthesis of nonionic reduced-sugar based bola

Surfactant Properties. *Journal of colloid and interface science*, 174, 373-377. Richel, A.; Laurent, P.; Wathelet, B.; Wathelet, J. & Paquot, M. (2010). Microwave-assisted

Evidences. *The journal of physical chemistry. B,* 113, 8872-8877.

*A: Physicochemical and Engineering Aspects,* 328, 100-106.

*Detergents,* 14, 51-63.

*and Interface Science*, 284, 704–713

*chemical physics: PCCP*, 13, 15291–15298.

*Tetrahedron letters,* 51, 1356-1360.

*Interface Science,* 108-109, 207-226.

*Tetrahedron,* 63, 10622-10629.

phase system. *Biotechnology Letters,* 12, 397-402.

ISBN978-1-4200-5166-7.

isomaltulose. *Carbohydrate Research,* 343, 1999-2009.

(2011). Synthesis and Surface-Active Properties of Uronic Amide Derivatives, Surfactants from Renewable Organic Raw Materials.*Journal of Surfactants and* 

formed by synthetic alkyl glycosideswith unsaturated alkyl chains. *Journal of Colloid* 

of carbohydrate-based amphiphilic materials: the examples of sucrose and

Miñones, J. & Paquot, M. *(*2011a). Monolayer properties of uronic acid bicatenary derivatives at the air-water interface: effect of hydroxyl group stereochemistry evidenced by experimental and computational approaches. *Physical chemistry* 

Amphiphilic Molecular Structures for Colloidal System Formation and Stability: The Case of Carbohydrate-Based Surfactants. In *Amphiphiles: Molecular Assembly* 

L.; Paquot, M. (2009). Impacts of the Carbonyl Group Location of Ester Bond on Interfacial Properties of Sugar-Based Surfactants: Experimental and Computational

tensions of surfactin aqueous solutions.*Journal of Surfactants and Detergents, 7* (1), 41-46.

Static Mode of the Drop Volume Technique to the Determination of Fundamental

synthesis of D-glucuronic acid derivatives using cost-effective solid acid catalysts.

Sakai, H. & Abe, M. (2008). Adsorption characteristics of sugar-based monomeric and gemini surfactants at the silica/aqueous solution interface. *Colloids and Surfaces* 

emulsions using polymeric surfactants based on inulin. *Advances in Colloid and* 

amphiphiles and gemini surfactants with an α,-diamino-(oxa) alkyl spacer.

Increased concerns about global warming and greenhouse gas emissions as well as the exhaustion of easily accessible fossil fuel resources are calling for effective carbon dioxide (CO2) mitigation technologies and clean and renewable energy sources. One of the major gases leading to global warming is carbon dioxide. CO2 makes up 68% of the estimated total greenhouse gas emissions (Harrington & Foster, 1999).

There have been several approaches proposed for managing the levels of CO2 emitted into the atmosphere, including ocean sequestration such as deep ocean injection or increasing the amount of CO2 dissolved in the ocean. Another proposed form of sequestration is to sequester the CO2 into terrestrial ecosystems (Halmann, 1993). The short term options of sequestration by direct injection into geologic or oceanic sinks are recognized as methods to reduce the CO2 levels but do not address issues of sustainability (Stewart & Hessami, 2005). Carbon sequestration can also be accomplished through chemical approaches; some problems with these approaches are that they must be safe for the environment, stable for long- term storage, and cost - competitive to other sequestration options. Other technologies have been considered, such as chemical absorption, membrane separation, cryogenic fractionation and adsorption using molecular sieves, but they are even less energy efficient as to be considered economically viable (Stewart & Hessami, 2005).

One of the most understudied methods for CO2 mitigation is the use of biological processes (via microalgae) in a direct CO2 to biomass conversion from point source emissions of CO2 in engineered systems such as photobioreactors. Microalgal biofixation of carbon dioxide (CO2) in photobioreactors has recently gained renewed interest as a promising strategy for CO2 mitigation. The use of photobioreactors for microalgal CO2 sequestration offers the principal advantages of increased microalgae productivity, owing to controlled environmental conditions, and optimized space/volume utilization and, thus, more efficient use of costly land. In fact, the photosynthetic solution when scaled up would present a far superior and sustainable solution under both environmental and economic considerations.

Fig. 1 shows the importance of the microalgae photobioreactor and its general applications, microalgae used to capture waste CO2 utilizing the nutrients in wastewater and natural

CO2 Biomitigation and Biofuel Production Using

**2. Mechanism of the photosynthesis and biophotolysis** 

production.

microbes (Amos, 2004).

Calvin cycle for cell growth.

Microalgae: Photobioreactors Developments and Future Directions 231

chapter will present possibly new insights that could be gained in the future for the potential commercial exploitation of microalgae for CO2 biomitigation and biofuel

Photoautotrophic microorganisms like eukaryotic green microalgae, possess chlorophyll and other pigments to capture sunlight energy and use photosynthetic systems (PSII and PSI) to carry out plant-like oxygenic photosynthesis (Kruse et al. 2005). The pigments in PSII (P680) absorb the photons with a wavelength shorter than 680 nm, generating a strong oxidant capable of splitting water into protons (H+), electrons (e-) and O2 as shown in Fig. 2.

Fig. 2. Schematic mechanisms of photosynthesis and biophotolysis of photoautotrophic

The electrons or reducing equivalents are transferred through a series of electron carriers and cytochrome complex to PSI. The pigments in PSI (P700) absorb the photons with a wavelength under 700 nm, which further raises the energy level of the electrons to reduce the oxidized ferredoxin (Fd) and/or nicotinamide adenine dinucleotide phosphate (NADP+) into their reduced forms. The proton gradient formed across the cellular (or thylakoid) membrane drives adenosine triphosphate (ATP) production via ATP synthase. CO2 is reduced with ATP and NADPH via a reductive pentose phosphate pathway or

The excess reduced carbon is stored inside the cells as carbohydrates (CH2O) and/or lipids. The type of carbohydrate product produced depends on the type of strain being

Fig. 1. Microalgae photobioreactors applications

solar light. The microalgae biomass produced can be used for biofuel production (such as biodiesel and methane) and other by products (such as animal feeds and polymers).

Broadly, photo-bioreactors can be classified as open (pond) systems or closed systems. Considering all the limitations and shortcomings of the pond systems, most researchers, had oriented their research works towards the development of an unconventional way for microalgae culture, which should be fully closed and compact with high surface-to-volume ratio and all the growth factors be optimized.

With these desired characteristics as the main goals, research on tubular and airlift photobioreactors were the right orientation and some forms of designs had in certain aspects succeeded when used in the lab scale. However, few of these forms could be really applicable in the pilot production scale, due to serious obstacles of operational problems and growth limitations. Amongst them, were primarily the oxygen build-up in the growth medium, photoinhibition, light saturation effect and the overheating inside the tube walls by the intensive solar radiation when operating in summer seasons especially in the midday light hours. Besides, the poor circulation of the growth medium causes the algal staining on the inner walls of the tubes, gave eventually an uneconomic results. Over the years, several solutions have been proposed to overcome these fundamental limitations to productivity. However, these systems are complicated to scale up and may be suitable for small-scale cultivation. Moreover, there is little knowledge about the feasibility of photobioreactors scale-up and developments.

The developmental state of the photobioreactor technology for CO2 mitigation and biofuel production will be reviewed in this chapter, focusing on several essential issues, such as effective and efficient provision of light; supply of carbon dioxide while minimizing losses; removal of photo-synthetically generated oxygen that may inhibit metabolism or otherwise damage the culture if allowed to accumulate; sensible scalability of the photobioreactor technology; harvesting the microalgae biomass and biofuel production. The theoretical background of microalgae cultivation will be summarized in this chapter as well. The

solar light. The microalgae biomass produced can be used for biofuel production (such as

Broadly, photo-bioreactors can be classified as open (pond) systems or closed systems. Considering all the limitations and shortcomings of the pond systems, most researchers, had oriented their research works towards the development of an unconventional way for microalgae culture, which should be fully closed and compact with high surface-to-volume

With these desired characteristics as the main goals, research on tubular and airlift photobioreactors were the right orientation and some forms of designs had in certain aspects succeeded when used in the lab scale. However, few of these forms could be really applicable in the pilot production scale, due to serious obstacles of operational problems and growth limitations. Amongst them, were primarily the oxygen build-up in the growth medium, photoinhibition, light saturation effect and the overheating inside the tube walls by the intensive solar radiation when operating in summer seasons especially in the midday light hours. Besides, the poor circulation of the growth medium causes the algal staining on the inner walls of the tubes, gave eventually an uneconomic results. Over the years, several solutions have been proposed to overcome these fundamental limitations to productivity. However, these systems are complicated to scale up and may be suitable for small-scale cultivation. Moreover, there is little knowledge about the feasibility of photobioreactors

The developmental state of the photobioreactor technology for CO2 mitigation and biofuel production will be reviewed in this chapter, focusing on several essential issues, such as effective and efficient provision of light; supply of carbon dioxide while minimizing losses; removal of photo-synthetically generated oxygen that may inhibit metabolism or otherwise damage the culture if allowed to accumulate; sensible scalability of the photobioreactor technology; harvesting the microalgae biomass and biofuel production. The theoretical background of microalgae cultivation will be summarized in this chapter as well. The

biodiesel and methane) and other by products (such as animal feeds and polymers).

Fig. 1. Microalgae photobioreactors applications

ratio and all the growth factors be optimized.

scale-up and developments.

chapter will present possibly new insights that could be gained in the future for the potential commercial exploitation of microalgae for CO2 biomitigation and biofuel production.

#### **2. Mechanism of the photosynthesis and biophotolysis**

Photoautotrophic microorganisms like eukaryotic green microalgae, possess chlorophyll and other pigments to capture sunlight energy and use photosynthetic systems (PSII and PSI) to carry out plant-like oxygenic photosynthesis (Kruse et al. 2005). The pigments in PSII (P680) absorb the photons with a wavelength shorter than 680 nm, generating a strong oxidant capable of splitting water into protons (H+), electrons (e-) and O2 as shown in Fig. 2.

Fig. 2. Schematic mechanisms of photosynthesis and biophotolysis of photoautotrophic microbes (Amos, 2004).

The electrons or reducing equivalents are transferred through a series of electron carriers and cytochrome complex to PSI. The pigments in PSI (P700) absorb the photons with a wavelength under 700 nm, which further raises the energy level of the electrons to reduce the oxidized ferredoxin (Fd) and/or nicotinamide adenine dinucleotide phosphate (NADP+) into their reduced forms. The proton gradient formed across the cellular (or thylakoid) membrane drives adenosine triphosphate (ATP) production via ATP synthase. CO2 is reduced with ATP and NADPH via a reductive pentose phosphate pathway or Calvin cycle for cell growth.

The excess reduced carbon is stored inside the cells as carbohydrates (CH2O) and/or lipids. The type of carbohydrate product produced depends on the type of strain being

CO2 Biomitigation and Biofuel Production Using

a polymer of carbohydrates known as chyrsolaminarin.

and carbohydrates as storage compounds.

could be extracted and converted into biodiesel.

**4.1 Tubular photo-bioreactors** 

**4. Photo-bioreactor for carbon dioxide sequestration** 

widely used photobioreactors for CO2 biofixation and biofuel production:

Microalgae: Photobioreactors Developments and Future Directions 233

1. The diatoms (Bacillariophyceae). These algae dominate the phytoplankton of the oceans, but are also found in fresh and brackish water. Approximately 100,000 species are known to exist. Diatoms contain polymerized silica (Si) in their cell walls. All cells store carbon in a variety of forms. Diatoms store carbon in the form of natural oils or as

2. The green algae (Chlorophyceae). These are also quite abundant, especially in freshwater. They can occur as single cells or as colonies. Green algae are the evolutionary progenitors of modern plants. The main storage compound for green

3. The blue-green algae (Cyanophyceae). Much closer to bacteria in structure and organization, these algae play an important role in fixing nitrogen from the atmosphere.

4. The golden algae (Chrysophyceae). This group of algae is similar to the diatoms in pigmentation and biochemical composition. They have more complex pigment systems, and can appear yellow, brown or orange in color. Approximately 1,000 species are known to exist, primarily in freshwater systems. The golden algae produce natural oils

All algae primary comprise of the following, in varying proportions (Table 1): proteins, carbohydrates, fats and nucleic acids. While the percentages vary with the type of algae, there are algae types that are comprised of up to 40% of their overall mass by fatty acids that

Photobioreactors for microalgae cultivation can be classified as open systems or closed systems. Open systems are ponds, constructed on the large open areas, in rows with growth medium exposed to environment and sunlight. Closed systems are those where growth medium enclosed from the environment. Open systems have many disadvantages over closed system, for instance they are hard to control and, contamination from external environment is high and could cause the microalgae mutate (Camacho Rubio et al., 1999). Closed systems are easy to monitor, less chances of contamination, better mass transfer (varies based on the type of bioreactor), occupy less space for the same algal growth. Closed systems can be classified as tubular photobioreactors; stirred photobioreactors; flat plate photobioreactors; hollow fiber membrane photobioreactors; airlift and sparged bubble column photobioreactors. Unfortunately, none of the these bioreactor configurations is able to control effectively all process parameters that are required for maximum CO2 biofixation, microalgal growth and metabolic rates, particularly at large scale production. Below is a brief description for the most

Tubular photo-bioreactors consist of long thin tubes arranged in different geometrical patterns (helical, straight tubes) to optimize irradiance from a point light source (sun). Generally liquid growth medium is circulated in these tubes by air bubbling and by injection of air into one end of the system and degassed at the other end. Construction, light regime, mass transfer and scale up issues of these photo-bioreactors have been discussed (Molina Grima, 2000). Experiments have showed that a large-scale tubular photo-bioreactor

algae is starch, though oils can be produced under certain conditions.

There are approximately 2,000 known species found in a variety of habitats.

used. The reducing power (Fd) could also be directed to hydrogenase (Hase) for hydrogen evolution.

#### **3. Microalgae and microalgae cultivation**

Algae are a large and diverse group of simple, typically autotrophic organisms, ranging from unicellular to multicultural forms. The largest and most complex marine forms are called seaweeds. They are photosynthetic, like plants, and simple because they lack the many distinct organs found in land plants. Some unicellular species rely entirely on external energy sources and have limited or no photosynthetic apparatus. All algae have photosynthetic machinery ultimately derived from the cyanobacteria, and so produce oxygen as a byproduct of photosynthesis.


Table 1. Chemical composition of algae expressed on a dry matter basis (%) (Becker,1994)

Algae can be classified into two types based on their sizes, microalgae and macroalgae. Microalgae are microscopic photosynthetic organisms (less than 2 mm in diameter). However, macroalgae, these organisms that are found in both marine and freshwater environments. Biologists have categorized microalgae in a variety of classes, mainly distinguished by their pigmentation, life cycle and basic cellular structure (Amos, 2004). The most frequently cited microalgae as carrying one or more of the desirable features for efficient and economical combination of CO2 biofixation, wastewater treatment and lipid synthesis toward biofuel production are:

used. The reducing power (Fd) could also be directed to hydrogenase (Hase) for hydrogen

Algae are a large and diverse group of simple, typically autotrophic organisms, ranging from unicellular to multicultural forms. The largest and most complex marine forms are called seaweeds. They are photosynthetic, like plants, and simple because they lack the many distinct organs found in land plants. Some unicellular species rely entirely on external energy sources and have limited or no photosynthetic apparatus. All algae have photosynthetic machinery ultimately derived from the cyanobacteria, and so produce

*Scenedesmus obliquus* 50-56 10-17 12-14 3-6 *Scenedesmus quadricauda* 47 - 1.9 - *Scenedesmus dimorphus* 8-18 21-52 16-40 - *Chlamydomonas rheinhardii* 48 17 21 - *Chlorella vulgaris* 51-58 12-17 14-22 4-5 *Chlorella pyrenoidosa* 57 26 2 - *Spirogyra sp.* 6-20 33-64 11-21 - *Dunaliella bioculata* 49 4 8 - *Dunaliella salina* 57 32 6 - *Euglena gracilis* 39-61 14-18 14-20 - *Prymnesium parvum* 28-45 25-33 22-38 1-2 *Tetraselmis maculata* 52 15 3 - *Porphyridium cruentum* 28-39 40-57 9-14 - *Spirulina platensis* 46-63 8-14 4--9 2-5 *Spirulina maxima* 60-71 13-16 6-7 3-4.5 *Synechoccus sp.* 63 15 11 5 *Anabaena cylindrica* 43 - 56 25-30 4-7 -

Table 1. Chemical composition of algae expressed on a dry matter basis (%) (Becker,1994)

Algae can be classified into two types based on their sizes, microalgae and macroalgae. Microalgae are microscopic photosynthetic organisms (less than 2 mm in diameter). However, macroalgae, these organisms that are found in both marine and freshwater environments. Biologists have categorized microalgae in a variety of classes, mainly distinguished by their pigmentation, life cycle and basic cellular structure (Amos, 2004). The most frequently cited microalgae as carrying one or more of the desirable features for efficient and economical combination of CO2 biofixation, wastewater treatment and lipid

**Strain Protein Carbohydrates Lipids Nucleic acid** 

evolution.

**3. Microalgae and microalgae cultivation** 

oxygen as a byproduct of photosynthesis.

synthesis toward biofuel production are:


All algae primary comprise of the following, in varying proportions (Table 1): proteins, carbohydrates, fats and nucleic acids. While the percentages vary with the type of algae, there are algae types that are comprised of up to 40% of their overall mass by fatty acids that could be extracted and converted into biodiesel.

#### **4. Photo-bioreactor for carbon dioxide sequestration**

Photobioreactors for microalgae cultivation can be classified as open systems or closed systems. Open systems are ponds, constructed on the large open areas, in rows with growth medium exposed to environment and sunlight. Closed systems are those where growth medium enclosed from the environment. Open systems have many disadvantages over closed system, for instance they are hard to control and, contamination from external environment is high and could cause the microalgae mutate (Camacho Rubio et al., 1999). Closed systems are easy to monitor, less chances of contamination, better mass transfer (varies based on the type of bioreactor), occupy less space for the same algal growth. Closed systems can be classified as tubular photobioreactors; stirred photobioreactors; flat plate photobioreactors; hollow fiber membrane photobioreactors; airlift and sparged bubble column photobioreactors. Unfortunately, none of the these bioreactor configurations is able to control effectively all process parameters that are required for maximum CO2 biofixation, microalgal growth and metabolic rates, particularly at large scale production. Below is a brief description for the most widely used photobioreactors for CO2 biofixation and biofuel production:

#### **4.1 Tubular photo-bioreactors**

Tubular photo-bioreactors consist of long thin tubes arranged in different geometrical patterns (helical, straight tubes) to optimize irradiance from a point light source (sun). Generally liquid growth medium is circulated in these tubes by air bubbling and by injection of air into one end of the system and degassed at the other end. Construction, light regime, mass transfer and scale up issues of these photo-bioreactors have been discussed (Molina Grima, 2000). Experiments have showed that a large-scale tubular photo-bioreactor

CO2 Biomitigation and Biofuel Production Using

temperature, pH and nutrient requirements.

**5.1 Light provision** 

darkness (Pulz, 2001).

polysaccharide production.

Microalgae: Photobioreactors Developments and Future Directions 235

dioxide level, photo-synthetically generated oxygen, Gas transfer, mixing rates,

Light is the basic energy source for phototrophic microorganisms. The intensity and utilization efficiency of the light supplied are thus of crucial importance in microalgal bioreactors. Light intensity decreases deeper within the culture medium, especially in highdensity cultures; hence, the issue of optical depth, which measures the proportion of radiation absorbed or scattered along a path through a partially transparent medium,

Both sunlight and artificial light have been used via outer surface exposure as well as inner volume exposure, through the placement of lighting devices (e.g. LEDs or optical fibers) inside the bioreactor itself (Suh & Lee, 2003). The photosynthetically active radiance is normally assumed to be 43–45% in the wavelength range of 400–700 nm (Laws et al. 1987). The light intensity available to microalgae in high-density cultures is significantly attenuated by mutual shading; to maximize light absorbance and minimize light attenuation, bioreactors should be designed with a high surface area-to-volume ratio,

Good microalgal growth rates have been reported (Hu et al. 1998) under a light intensity of 4000 μmol m-2 s-1; this intensity is twice the solar flux in a medium latitude spot at midday during summer. However, a strong species-dependence exists that should be taken into account. By contrast, light above a saturation point causes light inhibition, which can be counterbalanced by exposing microalgal cells to very short cyclic periods of light and

The ratio of light to dark (or low-intensity light) periods in a cycle is crucial for microalgal productivity (Munoz & Guieysse, 2006). Similar overall numbers of moles of photons do not necessarily produce equal growth rates of (or CO2 assimilation by) microalgae. When the light/dark cycle period approaches the photosynthetic unit turnover time (equal to the dark reaction time, estimated to lie within 1–15 ms), maximum photosynthetic efficiencies can be achieved (Richmond et al. 2003). Moreover, compared with periodic darkness, periods of low light intensity significantly increase growth, CO2 assimilation and lipid productivity in microalgae for a given whole light level (Cuello et al. 2008). This type of lighting design can be achieved via artificial light, such as hybrid lighting systems (Muhs, 2000). Different lamps generate distinct spectra, and different microalgal species possess dissimilar absorption optima; therefore, each individual case should be studied before deciding on the set point of this important operational parameter. Variation of the exponential growth rates of *Phorphyridium cruentum* have been recorded (Suh & Lee, 2003) with variable radiation energies and light spectra, concluding that blue light (400–500 nm) increases cell growth and

In terms of artificially illuminated bioreactors, the need for small reactor diameters to increase the illuminated surface area per unit volume of culture can be circumvented through provision of internal illumination. High biomass yields are more crucial in the case of artificially illuminated reactors, because the light provided adds to the overall operational cost of the underlying process. Such costs can be kept below acceptable

should be considered in microalgal bioreactor design (Kumar et al., 2010).

coupled with a short light path (Richmond et al. 2003).

has failed and the main reason attributed for its failure was the large dissolved oxygen in the system (Molina Grima, 2000). Hence a system should not at any stage be over saturated with oxygen as this would cause algae to shutdown photosynthesis and growth (Camacho Rubio et al., 1999). It was also reported that tubular photo-bioreactors are difficult to build and maintain, and have limited scalability.

#### **4.2 Mechanically stirred photo-bioreactors**

Mechanically stirred photo-bioreactors use baffles to stir the growth medium to attain a mass transfer of air/CO2 into liquid. A drawback of the stirred medium is if stirred vigorously the algae cell wall would be damaged by the high fluid shear forces (Molina Grima et al., 1996). If it is stirred slowly, eddy currents will not be established that move the algae toward the light source thereby decreasing the efficiency of light available for the photosynthetic process and also reducing mass transfer of nutrients from the air/CO2 to the liquid in the systems.

#### **4.3 Airlift photo-bioreactors**

Airlift photo-bioreactors are basically a column divided into two parts, air/CO2 is bubble through only one side of the partition which causes a liquid current pattern to develop with the air bubble side called the riser and other part called the downcomer (Sánchez Mirón et al., 2000). These bioreactors are extensively investigated for fermentation process and wastewater treatment (Znad et al. 2004, Znad et al. 2006) but have not been looked at as a replacement for the popular tubular photo-bioreactors until recent times (Sánchez Mirón et al., 2000). The airlift photo-bioreactor characterized by; high mass transfer, good mixing with low shear stress, low energy consumption, high potentials for scalability, easy to sterilize, readily tempered, good for immobilization of algae, reduced photo-inhibition. However, the main limitations are; the small illumination surface area and decrease of illumination surface area upon scale-up. It has become clear that biological carbon sequestration and hydrogen production technologies have been poorly studied in the airlift photo-bioreactors and are in their infancy of development.

#### **4.4. Bubble column photo-bioreactors**

Bubble column bioreactors are vertical columns either cylindrical or rectangular filled with growth medium and air is bubble through a sparged system installed at the bottom. These systems have the highest gas hold ups rates which means they have the best mass transfer compared to other systems (Miron et al. 2000; Kommareddy & Anderson , 2003). A modified version of these bubble column bioreactors is porous membrane reactors, which have efficient aeration, give smaller bubbles, and pressure drop across the membrane is low compared with other rigid sparged bubble column reactors. These characteristics are achievable at high gas flow rates, with low energy costs (Poulsen & Iversen, 1997).

#### **5. Factors affecting the photobioreactor performance**

The key parameters that affect the photobioreactor performance, i.e., the growth of the microalgae in the photo-bioreactor, are the effective and efficient provision of light, carbon dioxide level, photo-synthetically generated oxygen, Gas transfer, mixing rates, temperature, pH and nutrient requirements.

#### **5.1 Light provision**

234 Advances in Chemical Engineering

has failed and the main reason attributed for its failure was the large dissolved oxygen in the system (Molina Grima, 2000). Hence a system should not at any stage be over saturated with oxygen as this would cause algae to shutdown photosynthesis and growth (Camacho Rubio et al., 1999). It was also reported that tubular photo-bioreactors are difficult to build

Mechanically stirred photo-bioreactors use baffles to stir the growth medium to attain a mass transfer of air/CO2 into liquid. A drawback of the stirred medium is if stirred vigorously the algae cell wall would be damaged by the high fluid shear forces (Molina Grima et al., 1996). If it is stirred slowly, eddy currents will not be established that move the algae toward the light source thereby decreasing the efficiency of light available for the photosynthetic process and also reducing mass transfer of nutrients from the air/CO2 to the

Airlift photo-bioreactors are basically a column divided into two parts, air/CO2 is bubble through only one side of the partition which causes a liquid current pattern to develop with the air bubble side called the riser and other part called the downcomer (Sánchez Mirón et al., 2000). These bioreactors are extensively investigated for fermentation process and wastewater treatment (Znad et al. 2004, Znad et al. 2006) but have not been looked at as a replacement for the popular tubular photo-bioreactors until recent times (Sánchez Mirón et al., 2000). The airlift photo-bioreactor characterized by; high mass transfer, good mixing with low shear stress, low energy consumption, high potentials for scalability, easy to sterilize, readily tempered, good for immobilization of algae, reduced photo-inhibition. However, the main limitations are; the small illumination surface area and decrease of illumination surface area upon scale-up. It has become clear that biological carbon sequestration and hydrogen production technologies have been poorly studied in the airlift

Bubble column bioreactors are vertical columns either cylindrical or rectangular filled with growth medium and air is bubble through a sparged system installed at the bottom. These systems have the highest gas hold ups rates which means they have the best mass transfer compared to other systems (Miron et al. 2000; Kommareddy & Anderson , 2003). A modified version of these bubble column bioreactors is porous membrane reactors, which have efficient aeration, give smaller bubbles, and pressure drop across the membrane is low compared with other rigid sparged bubble column reactors. These characteristics are achievable at high gas flow rates, with low energy costs (Poulsen &

The key parameters that affect the photobioreactor performance, i.e., the growth of the microalgae in the photo-bioreactor, are the effective and efficient provision of light, carbon

and maintain, and have limited scalability.

liquid in the systems.

**4.3 Airlift photo-bioreactors** 

**4.2 Mechanically stirred photo-bioreactors** 

photo-bioreactors and are in their infancy of development.

**5. Factors affecting the photobioreactor performance** 

**4.4. Bubble column photo-bioreactors** 

Iversen, 1997).

Light is the basic energy source for phototrophic microorganisms. The intensity and utilization efficiency of the light supplied are thus of crucial importance in microalgal bioreactors. Light intensity decreases deeper within the culture medium, especially in highdensity cultures; hence, the issue of optical depth, which measures the proportion of radiation absorbed or scattered along a path through a partially transparent medium, should be considered in microalgal bioreactor design (Kumar et al., 2010).

Both sunlight and artificial light have been used via outer surface exposure as well as inner volume exposure, through the placement of lighting devices (e.g. LEDs or optical fibers) inside the bioreactor itself (Suh & Lee, 2003). The photosynthetically active radiance is normally assumed to be 43–45% in the wavelength range of 400–700 nm (Laws et al. 1987). The light intensity available to microalgae in high-density cultures is significantly attenuated by mutual shading; to maximize light absorbance and minimize light attenuation, bioreactors should be designed with a high surface area-to-volume ratio, coupled with a short light path (Richmond et al. 2003).

Good microalgal growth rates have been reported (Hu et al. 1998) under a light intensity of 4000 μmol m-2 s-1; this intensity is twice the solar flux in a medium latitude spot at midday during summer. However, a strong species-dependence exists that should be taken into account. By contrast, light above a saturation point causes light inhibition, which can be counterbalanced by exposing microalgal cells to very short cyclic periods of light and darkness (Pulz, 2001).

The ratio of light to dark (or low-intensity light) periods in a cycle is crucial for microalgal productivity (Munoz & Guieysse, 2006). Similar overall numbers of moles of photons do not necessarily produce equal growth rates of (or CO2 assimilation by) microalgae. When the light/dark cycle period approaches the photosynthetic unit turnover time (equal to the dark reaction time, estimated to lie within 1–15 ms), maximum photosynthetic efficiencies can be achieved (Richmond et al. 2003). Moreover, compared with periodic darkness, periods of low light intensity significantly increase growth, CO2 assimilation and lipid productivity in microalgae for a given whole light level (Cuello et al. 2008). This type of lighting design can be achieved via artificial light, such as hybrid lighting systems (Muhs, 2000). Different lamps generate distinct spectra, and different microalgal species possess dissimilar absorption optima; therefore, each individual case should be studied before deciding on the set point of this important operational parameter. Variation of the exponential growth rates of *Phorphyridium cruentum* have been recorded (Suh & Lee, 2003) with variable radiation energies and light spectra, concluding that blue light (400–500 nm) increases cell growth and polysaccharide production.

In terms of artificially illuminated bioreactors, the need for small reactor diameters to increase the illuminated surface area per unit volume of culture can be circumvented through provision of internal illumination. High biomass yields are more crucial in the case of artificially illuminated reactors, because the light provided adds to the overall operational cost of the underlying process. Such costs can be kept below acceptable

CO2 Biomitigation and Biofuel Production Using

produced.

**5.4 Gas transfer** 

**5.5 Mixing rates** 

requiring a large energy input.

Microalgae: Photobioreactors Developments and Future Directions 237

Unfortunately, the efficiencies of most techniques used to date for dissolved oxygen removal from microalgal cultures are still not satisfactory. As a result, the classical bubbling mode of operation has been employed to avoid the costlier need for degassing devices. The use of several small bioreactors instead of one large unit also alleviates this problem. Scale-up is indeed easier for facilities that use many small reactors in parallel, even though investment costs might be higher than with fewer large equipment units. Continued research is needed to accurately match the amount of CO2 supplied to the actual uptake requirement of the metabolizing microalgae, as well as the amount of O2 removed to the actual amount of O2

Gases introduced into bioreactors serve a number of purposes in microalgal cultivation, including: supply of CO2 as sources of carbon for biomass primary and secondary metabolism; provision of internal mixing, which avoids nutrient concentration gradients; promotion of exposure of all cells to light (especially in high density cultures), while minimizing self-shading and phototoxicity; control of pH by assuring dissolution of CO2 and avoiding gradients thereof; and stripping of accumulated dissolved oxygen, hence

Among the various alternatives, bubbling CO2-enriched air into the bottom of the bioreactor with bubble diffusers has been the most frequently used approach. Moderate overall transfer efficiencies (13–20%) can be achieved by this mode of gas delivery (Carvalho et al. 2006); however, associated drawbacks are loss of CO2 to the atmosphere, biofouling of diffusers, and poor mass transfer rates owing to a relatively low interfacial specific surface area. Better overall efficiencies are expected for hollow-fiber membrane bioreactors in which the slightly lower mass transfer coefficients that arise from a less turbulent local hydrodynamic pattern are compensated by the much larger area per unit volume available for mass transfer. In addition, the area of mass transfer is well defined, and the pressure on the gas side can be controlled so as to supply only the required amount of CO2, hence permitting more accurate control of the transfer rate and a dramatic reduction in the amount

Mixing is a key parameter for acceptable performance of microalgal bioreactors. Low mixing rates hamper gaseous mass transfer and might even permit biomass settling. In either case, poor mixing leads to emergence of stagnant zones, where light and nutrients are insufficiently available and anoxic/anaerobic conditions will thus prevail, which results in a decrease of productivity (Kumar et al., 2010). Culture viability might also be compromised by production and accumulation of toxic compounds in stagnant zones (Becker, 1994). Conversely, high mixing rates can cause shear damage to cells (Carlsson et al. (2007), besides

The most common methods of mixing in microalgal bioreactors are pumping, mechanical stirring and gas injection. Pumping offers fair mixing efficiency, but low gas transfer rates; the associated hydrodynamic stress increases with the rotation speed of the pumps, or the number of passes of the microalgal suspension through the pump units (Jaouen et al. 1999).

reducing its toxicity to microalgae ((Kumar et al., 2010, Pulz, 2001).

of CO2 lost to the atmosphere (Carvalho & Malcata, 2001).

thresholds via in situ growth-monitoring and associated online control of the intensity of light supplied.

#### **5.2 Carbon uptake**

Biological CO2 fixation can be carried out by higher plants and microalgae, yet the latter possess a greater ability to fix CO2 (Li et al. 2008; Chisti, 2007; Tredici 2010). Usual sources of CO2 for microalgae include atmospheric CO2; CO2 from industrial exhaust gases (e.g. flue gas and flaring gas); and CO2 chemically fixed in the form of soluble carbonates (e.g. NaHCO3 and Na2CO3) (Kumar et al., 2010). The tolerance of various microalgal species to the concentration of CO2 is variable; however, the CO2 concentration in the gaseous phase does not necessarily reflect the CO2 concentration to which the microalga is exposed during dynamic liquid suspension, which depends on the pH and the CO2 concentration gradient created by the resistance to mass transfer. Under heterotrophic or mixotrophic conditions, some microalgal species can metabolize a variety of organic compounds, including sugars, molasses and acetic acid, as well as compounds present in wastewater and petroleum (Becker, 1994). Atmospheric CO2 levels (0.0387% (v/v)) are not sufficient to support the high microalgal growth rates and productivities needed for full-scale biofuel production.

Waste gases from combustion processes, however, typically contain >15% (v/v) CO2; this percentage indicates, in principle, that combustion processes will provide sufficient amounts of CO2 for large-scale production of microalgae (Doucha et al. 2005). Owing to the cost of upstream separation of CO2 gas, direct utilization of power plant flue gas has been considered in microalgal biofuel production systems (Lackne, 2003). Flue gases that contain CO2 at concentrations ranging from 5 to 15% (v/v) have indeed been introduced directly into ponds and bioreactors of various configurations that contain several microalgal species (Kumar et al., 2010).

#### **5.3 Oxygen generated**

Another specific issue of microalgal bioreactors is the accumulation of photosynthetically generated oxygen that may inhibit metabolism or otherwise damage the culture if allowed to accumulate, especially when the rate of photosynthesis, which often correlates with the rate of CO2 transfer, is high (as typical in horizontal tubular reactors) (Kumar et al., 2010). Most solutions to this problem rely on the use of a degasser (or gas exchange unit), where dissolved oxygen can be released (Morita et al. 2000). However, to attain effective separation between the gas and liquid phases, the path through the degasser should be such that the smallest bubbles have sufficient time to disengage from the liquid.

In tubular bioreactors, connections between tubes can incorporate a tube specifically for oxygen degassing, or a layer of parallel tubes connected by two manifolds: the lower manifold is used to inject air into the culture, and the higher one acts as the degasser (Kumar et al., 2010). Nevertheless, microalgal productivities were lower than expected in these tubular systems, possibly because of build-up of dissolved oxygen during high light intensity periods and along the bioreactor path between manifolds. In systems with exhaust gas recirculation, dissolved oxygen accumulation can be avoided by bubbling exhaust gas through a sodium sulfite solution before its return to the bioreactor (Cien-Fernandez et al. 2005).

Unfortunately, the efficiencies of most techniques used to date for dissolved oxygen removal from microalgal cultures are still not satisfactory. As a result, the classical bubbling mode of operation has been employed to avoid the costlier need for degassing devices. The use of several small bioreactors instead of one large unit also alleviates this problem. Scale-up is indeed easier for facilities that use many small reactors in parallel, even though investment costs might be higher than with fewer large equipment units. Continued research is needed to accurately match the amount of CO2 supplied to the actual uptake requirement of the metabolizing microalgae, as well as the amount of O2 removed to the actual amount of O2 produced.

#### **5.4 Gas transfer**

236 Advances in Chemical Engineering

thresholds via in situ growth-monitoring and associated online control of the intensity of

Biological CO2 fixation can be carried out by higher plants and microalgae, yet the latter possess a greater ability to fix CO2 (Li et al. 2008; Chisti, 2007; Tredici 2010). Usual sources of CO2 for microalgae include atmospheric CO2; CO2 from industrial exhaust gases (e.g. flue gas and flaring gas); and CO2 chemically fixed in the form of soluble carbonates (e.g. NaHCO3 and Na2CO3) (Kumar et al., 2010). The tolerance of various microalgal species to the concentration of CO2 is variable; however, the CO2 concentration in the gaseous phase does not necessarily reflect the CO2 concentration to which the microalga is exposed during dynamic liquid suspension, which depends on the pH and the CO2 concentration gradient created by the resistance to mass transfer. Under heterotrophic or mixotrophic conditions, some microalgal species can metabolize a variety of organic compounds, including sugars, molasses and acetic acid, as well as compounds present in wastewater and petroleum (Becker, 1994). Atmospheric CO2 levels (0.0387% (v/v)) are not sufficient to support the high

microalgal growth rates and productivities needed for full-scale biofuel production.

Waste gases from combustion processes, however, typically contain >15% (v/v) CO2; this percentage indicates, in principle, that combustion processes will provide sufficient amounts of CO2 for large-scale production of microalgae (Doucha et al. 2005). Owing to the cost of upstream separation of CO2 gas, direct utilization of power plant flue gas has been considered in microalgal biofuel production systems (Lackne, 2003). Flue gases that contain CO2 at concentrations ranging from 5 to 15% (v/v) have indeed been introduced directly into ponds and bioreactors of various configurations that contain several microalgal species

Another specific issue of microalgal bioreactors is the accumulation of photosynthetically generated oxygen that may inhibit metabolism or otherwise damage the culture if allowed to accumulate, especially when the rate of photosynthesis, which often correlates with the rate of CO2 transfer, is high (as typical in horizontal tubular reactors) (Kumar et al., 2010). Most solutions to this problem rely on the use of a degasser (or gas exchange unit), where dissolved oxygen can be released (Morita et al. 2000). However, to attain effective separation between the gas and liquid phases, the path through the degasser should be such that the

In tubular bioreactors, connections between tubes can incorporate a tube specifically for oxygen degassing, or a layer of parallel tubes connected by two manifolds: the lower manifold is used to inject air into the culture, and the higher one acts as the degasser (Kumar et al., 2010). Nevertheless, microalgal productivities were lower than expected in these tubular systems, possibly because of build-up of dissolved oxygen during high light intensity periods and along the bioreactor path between manifolds. In systems with exhaust gas recirculation, dissolved oxygen accumulation can be avoided by bubbling exhaust gas through a sodium sulfite solution before its return to the bioreactor (Cien-Fernandez et al.

smallest bubbles have sufficient time to disengage from the liquid.

light supplied.

**5.2 Carbon uptake** 

(Kumar et al., 2010).

2005).

**5.3 Oxygen generated** 

Gases introduced into bioreactors serve a number of purposes in microalgal cultivation, including: supply of CO2 as sources of carbon for biomass primary and secondary metabolism; provision of internal mixing, which avoids nutrient concentration gradients; promotion of exposure of all cells to light (especially in high density cultures), while minimizing self-shading and phototoxicity; control of pH by assuring dissolution of CO2 and avoiding gradients thereof; and stripping of accumulated dissolved oxygen, hence reducing its toxicity to microalgae ((Kumar et al., 2010, Pulz, 2001).

Among the various alternatives, bubbling CO2-enriched air into the bottom of the bioreactor with bubble diffusers has been the most frequently used approach. Moderate overall transfer efficiencies (13–20%) can be achieved by this mode of gas delivery (Carvalho et al. 2006); however, associated drawbacks are loss of CO2 to the atmosphere, biofouling of diffusers, and poor mass transfer rates owing to a relatively low interfacial specific surface area. Better overall efficiencies are expected for hollow-fiber membrane bioreactors in which the slightly lower mass transfer coefficients that arise from a less turbulent local hydrodynamic pattern are compensated by the much larger area per unit volume available for mass transfer. In addition, the area of mass transfer is well defined, and the pressure on the gas side can be controlled so as to supply only the required amount of CO2, hence permitting more accurate control of the transfer rate and a dramatic reduction in the amount of CO2 lost to the atmosphere (Carvalho & Malcata, 2001).

#### **5.5 Mixing rates**

Mixing is a key parameter for acceptable performance of microalgal bioreactors. Low mixing rates hamper gaseous mass transfer and might even permit biomass settling. In either case, poor mixing leads to emergence of stagnant zones, where light and nutrients are insufficiently available and anoxic/anaerobic conditions will thus prevail, which results in a decrease of productivity (Kumar et al., 2010). Culture viability might also be compromised by production and accumulation of toxic compounds in stagnant zones (Becker, 1994). Conversely, high mixing rates can cause shear damage to cells (Carlsson et al. (2007), besides requiring a large energy input.

The most common methods of mixing in microalgal bioreactors are pumping, mechanical stirring and gas injection. Pumping offers fair mixing efficiency, but low gas transfer rates; the associated hydrodynamic stress increases with the rotation speed of the pumps, or the number of passes of the microalgal suspension through the pump units (Jaouen et al. 1999).

CO2 Biomitigation and Biofuel Production Using

productivities (Lardon et al. 2009).

effective cultivation (Becker, 1994).

**6.1 Microalgae biomass harvesting** 

(Brennan & Owende, 2010).

the dewatering process of choice.

overall energy efficiency.

**6. Microalgae harvesting and conversion to fuels** 

techniques (Carlsson et al. 2007; Kumar et al., 2010; Uduman et al. 2010).

operating conditions and design new processes (Mallick, 2002).

Microalgae: Photobioreactors Developments and Future Directions 239

deprivation, microalgae grow at lower rates, but produce significantly more lipids, which are reserve compounds synthesized under stress conditions, even at the expense of lower

Phosphorus is the third most important nutrient for microalgal growth, and should be supplied to significant excess as phosphates because not all phosphorus compounds are bioavailable (e.g. those combined with metal ions) (Kumar et al. 2009). In the case of marine microalgae, seawater supplemented with commercial nitrate and phosphate fertilizers is commonly used for production of microalgae (Green & Durnford, 1996). Nevertheless, trace species, such as metals (Mg, Ca, Mn, Zn, Cu and Mb) and vitamins, are typically added for

Harvesting of the microalgae biomass, i.e., concentrating microscopic algal cells from the dilute solutions of the algal mass culture, is an essential step to secure high-quality effluents and to prevent cell washout (Richmond et al. 2003, Munoz & Guieysse, 2006). The main difficulties encountered in harvesting microalgae arise from the relatively low biomass concentration in conventional bioreactors, coupled with the small size of its constituent microalgal cells. Harvesting typically contributes to 20–30% of the total cost of microalgal biomass production (Carlsson et al. 2007). The major techniques presently applied in the harvesting of microalgae include coagulation, flocculation, sedimentation, centrifugation, foam fractionation, ultrasonic separation, flotation, membrane filtration, and electrophoresis

Selection of the harvesting method mainly depends on the properties of microalgae, such as density, size, the value of the desired products. Microalgae harvesting can generally be divided into a two-step process, bulk harvesting, to separate microalgal biomass from the bulk suspension, in this method, the total solid mater can reach 2–7% using flocculation, flotation, or gravity sedimentation; and the second step is thickening, to concentrate the slurry, using filtration and centrifugation. This step needs more energy than bulk harvesting

Microalgal cell immobilization has been proposed to circumvent the harvesting issue, but large-scale applications are limited. Further investigation is clearly needed to optimize

Following biomass harvest by centrifugation or filtration, microalgal paste traditionally consists of 90% (w/w) water, which meets the requirements for anaerobic digestion. However, it is necessary to reduce this value to a maximum of 50% (w/w) water for efficient oil extraction (Kumar et al., 2010). Despite its energy-intensive nature, drying has often been

Almost 90% of the energy required for biodiesel production is indeed accounted for by harvesting and dewatering of biomass, besides lipid extraction itself (Lardon et al. 2009). In addition to lipid extraction for biodiesel production, a novel process that gasifies biomass to methane and concentrated CO2 has recently been proposed (Stucki et al. 2009) for improved

Mechanical stirring has been reported to provide good mixing efficiency and gas transfer; however, it is likely to produce significant hydrodynamic stress (Tredici, 2003), which can be managed via adequate use of baffles to create a controlled turbulence pattern. Gas injection (bubbling) produces lower hydrodynamic stress, while providing good gas transfer and reasonable mixing efficiency (Richmond & Cheng-Wu, 2001); however, cell damage in sparged cultures increases as the biomass concentration increases, because exponentially higher degrees of stirring are needed to maintain a high-density culture at a predefined level of mixing (Pulz, 2001). One approach to minimize this problem is to maintain a low gas input per nozzle, so as to reduce shear stress and consequent cell damage (Barbosa et al. 2003).

#### **5.6 Temperature effects**

Temperature is one of the major factors that regulate cellular, morphological and physiological responses of microalgae: higher temperatures generally accelerate the metabolic rates of microalgae, whereas low temperatures lead to inhibition of microalgal growth (Munoz & Guieysse, 2006). The optimal temperature varies among microalgal species (Ono & Cuello, 2003); however, optimal temperatures are also influenced by other environmental parameters, such as light intensity. Optimal growth temperatures of 15–26 oC have been reported for some species, with maximum cell densities obtained at 23 oC. Only daytime higher temperatures were observed to have clearly favorable effects on microalgal growth rates due to photosynthesis, except when the night temperature was as low as 7 oC (Tamiya, 1957).

#### **5.7 pH effects**

Most microalgal species are favored by neutral pH, whereas some species are tolerant to higher pH (e.g. Spirulina platensis at pH 9 ) (Hu et al. 1998) or lower pH (e.g. Chlorococcum littorale at pH 4) (Kodama et al. 1993). There is a complex relationship between CO2 concentration and pH in microalgal bioreactor systems, owing to the underlying chemical equilibria among such chemical species as CO2, H2CO3, HCO3 and CO32-. Increasing CO2 concentrations can lead to higher biomass productivity, but will also decrease pH, which can have an adverse effect upon microalgal physiology. By contrast, microalgae have been shown to cause a rise in pH to 10–11 in open ponds because of CO2 uptake (Oswald, 1988). This increase in pH can be beneficial for inactivation of pathogens in microalgal wastewater treatment, but can also inhibit microalgal growth. Similarly, the speciation of NH3 and NH4 + in microalgal bioreactors is strongly dependent on pH – NH3 uncouples electron transport in the microalgal photosystem and competes with water molecules in oxidation reactions, thus leading to release of O2 (Hu et al. 1998).

#### **5.8 Nutrient requirements**

In addition to the carbon, nitrogen is the most important element that is required for microalgal nutrition (Becker, 1994) and, as a constituent of both nucleic acids and proteins, nitrogen is directly associated with the primary metabolism of microalgae. Fast-growing microalgal species prefer ammonium rather than nitrate as a primary nitrogen source (Green & Durnford, 1996); intermittent nitrate feeding, however, will enhance microalgal growth if a medium that lacks nitrate is used (Jin et al. 2006). Under partial nitrogen deprivation, microalgae grow at lower rates, but produce significantly more lipids, which are reserve compounds synthesized under stress conditions, even at the expense of lower productivities (Lardon et al. 2009).

Phosphorus is the third most important nutrient for microalgal growth, and should be supplied to significant excess as phosphates because not all phosphorus compounds are bioavailable (e.g. those combined with metal ions) (Kumar et al. 2009). In the case of marine microalgae, seawater supplemented with commercial nitrate and phosphate fertilizers is commonly used for production of microalgae (Green & Durnford, 1996). Nevertheless, trace species, such as metals (Mg, Ca, Mn, Zn, Cu and Mb) and vitamins, are typically added for effective cultivation (Becker, 1994).

#### **6. Microalgae harvesting and conversion to fuels**

#### **6.1 Microalgae biomass harvesting**

238 Advances in Chemical Engineering

Mechanical stirring has been reported to provide good mixing efficiency and gas transfer; however, it is likely to produce significant hydrodynamic stress (Tredici, 2003), which can be managed via adequate use of baffles to create a controlled turbulence pattern. Gas injection (bubbling) produces lower hydrodynamic stress, while providing good gas transfer and reasonable mixing efficiency (Richmond & Cheng-Wu, 2001); however, cell damage in sparged cultures increases as the biomass concentration increases, because exponentially higher degrees of stirring are needed to maintain a high-density culture at a predefined level of mixing (Pulz, 2001). One approach to minimize this problem is to maintain a low gas input per nozzle, so as to reduce shear stress and consequent cell damage (Barbosa et al.

Temperature is one of the major factors that regulate cellular, morphological and physiological responses of microalgae: higher temperatures generally accelerate the metabolic rates of microalgae, whereas low temperatures lead to inhibition of microalgal growth (Munoz & Guieysse, 2006). The optimal temperature varies among microalgal species (Ono & Cuello, 2003); however, optimal temperatures are also influenced by other environmental parameters, such as light intensity. Optimal growth temperatures of 15–26 oC have been reported for some species, with maximum cell densities obtained at 23 oC. Only daytime higher temperatures were observed to have clearly favorable effects on microalgal growth rates due to photosynthesis, except when the night temperature was as low as 7 oC

Most microalgal species are favored by neutral pH, whereas some species are tolerant to higher pH (e.g. Spirulina platensis at pH 9 ) (Hu et al. 1998) or lower pH (e.g. Chlorococcum littorale at pH 4) (Kodama et al. 1993). There is a complex relationship between CO2 concentration and pH in microalgal bioreactor systems, owing to the underlying chemical

concentrations can lead to higher biomass productivity, but will also decrease pH, which can have an adverse effect upon microalgal physiology. By contrast, microalgae have been shown to cause a rise in pH to 10–11 in open ponds because of CO2 uptake (Oswald, 1988). This increase in pH can be beneficial for inactivation of pathogens in microalgal wastewater treatment, but can also inhibit microalgal growth. Similarly, the speciation of NH3 and NH4

in microalgal bioreactors is strongly dependent on pH – NH3 uncouples electron transport in the microalgal photosystem and competes with water molecules in oxidation reactions,

In addition to the carbon, nitrogen is the most important element that is required for microalgal nutrition (Becker, 1994) and, as a constituent of both nucleic acids and proteins, nitrogen is directly associated with the primary metabolism of microalgae. Fast-growing microalgal species prefer ammonium rather than nitrate as a primary nitrogen source (Green & Durnford, 1996); intermittent nitrate feeding, however, will enhance microalgal growth if a medium that lacks nitrate is used (Jin et al. 2006). Under partial nitrogen


and CO32-. Increasing CO2

+

equilibria among such chemical species as CO2, H2CO3, HCO3

thus leading to release of O2 (Hu et al. 1998).

**5.8 Nutrient requirements** 

2003).

**5.6 Temperature effects** 

(Tamiya, 1957).

**5.7 pH effects** 

Harvesting of the microalgae biomass, i.e., concentrating microscopic algal cells from the dilute solutions of the algal mass culture, is an essential step to secure high-quality effluents and to prevent cell washout (Richmond et al. 2003, Munoz & Guieysse, 2006). The main difficulties encountered in harvesting microalgae arise from the relatively low biomass concentration in conventional bioreactors, coupled with the small size of its constituent microalgal cells. Harvesting typically contributes to 20–30% of the total cost of microalgal biomass production (Carlsson et al. 2007). The major techniques presently applied in the harvesting of microalgae include coagulation, flocculation, sedimentation, centrifugation, foam fractionation, ultrasonic separation, flotation, membrane filtration, and electrophoresis techniques (Carlsson et al. 2007; Kumar et al., 2010; Uduman et al. 2010).

Selection of the harvesting method mainly depends on the properties of microalgae, such as density, size, the value of the desired products. Microalgae harvesting can generally be divided into a two-step process, bulk harvesting, to separate microalgal biomass from the bulk suspension, in this method, the total solid mater can reach 2–7% using flocculation, flotation, or gravity sedimentation; and the second step is thickening, to concentrate the slurry, using filtration and centrifugation. This step needs more energy than bulk harvesting (Brennan & Owende, 2010).

Microalgal cell immobilization has been proposed to circumvent the harvesting issue, but large-scale applications are limited. Further investigation is clearly needed to optimize operating conditions and design new processes (Mallick, 2002).

Following biomass harvest by centrifugation or filtration, microalgal paste traditionally consists of 90% (w/w) water, which meets the requirements for anaerobic digestion. However, it is necessary to reduce this value to a maximum of 50% (w/w) water for efficient oil extraction (Kumar et al., 2010). Despite its energy-intensive nature, drying has often been the dewatering process of choice.

Almost 90% of the energy required for biodiesel production is indeed accounted for by harvesting and dewatering of biomass, besides lipid extraction itself (Lardon et al. 2009). In addition to lipid extraction for biodiesel production, a novel process that gasifies biomass to methane and concentrated CO2 has recently been proposed (Stucki et al. 2009) for improved overall energy efficiency.

CO2 Biomitigation and Biofuel Production Using

improved algal harvesting and dewatering.

microalgae biofuels industry (Kumar et al., 2010).

promising approaches (Kumar et al., 2010).

optimized at large scale.

infrastructure.

systems.

**8. References** 

193–200.

Microalgae: Photobioreactors Developments and Future Directions 241

Several challenges, addressed in this chapter, for microalgal based CO2 sequestration and biofuel production remain. Most studies reported to date have been performed on the bench-scale, and were conducted under strictly controlled conditions. As a result, little is known about the feasibility of the photobioreactor scale-up. Factors, such as supply of adequate amounts of CO2, nutrients and light to microalgal cells, should be investigated and

Co-digestion of microalgae with wastewater sludge for biogas production should also be considered, because this strategy could be integrated into the existing wastewater

Microalga-based CO2 fixation and biofuel production can be more sustainable by coupling microalgal biomass production with existing power generation and wastewater treatment infrastructures. Microalgae can utilize low-quality water, such as agricultural runoff or municipal, industrial or agricultural wastewaters, as a source of water for the growth medium as well as a source of nitrogen, phosphorus and minor nutrients (Becker, 1994). Hence, an additional economic and environmental incentive exists as a result of the decreased cost of water and chemicals required for the formulation of the growth medium, while providing a pathway for wastewater treatment (Kumar et al., 2010; Mallick, 2002; Demirbas, 2004). Such a coupling technology need to be further investigated. A number of crucial research gaps remain that must be overcome to achieve full-scale operation such as improved algal growth and nutrient uptake rates; integration of biosystems with waste gas, wastewater and water reclamation systems; improved gas transfer and mixing; and

Harvesting, dewatering and lipid extraction from microalgal biomass are still challenging issues because they consume large amounts of energy – mainly because of the small cell size and relatively low biomass levels of microalgal cultures. Therefore, more efficient and economic harvesting technology should be developed to enhance the commercial viability of

A key challenge for microalgal biodiesel production is the use of microalgal species that can maintain a high growth rate in addition to a high metabolic rate, thus leading to significant lipid yields. This major challenge can be duly addressed via extensive bio-prospecting or target oriented genetic engineering – currently such approaches starting to appear as

Finally, it seems there is a lack of fundamental information needed to rationally optimize the performance of existing bioreactors. Novel bioreactor configurations and designs are also needed that promote microalgal growth and CO2 biofixation, characterized by volumetric productivities at least one order of magnitude above those of conventional open pond

Acién Fernández, FG.; Fernández Sevilla, JM.; Egorova-Zachernyuk, TA.; Molina Grima, E.

(2005). Cost-effective production of 13C, 15N stable isotope-labelled biomass from phototrophic microalgae for various biotechnological applications. *Biomol. Eng*. 22,

Most microalgae exhibit the phenomenon of bioflocculation, which is the spontaneous aggregation of algal cells into large flocs. These flocs will then settle rather rapidly. The process yet not fully understood and need more investigation. It depends on the elaboration of polymers by the algal cells that makes the cells stick together (Benemann & Pedroni, 2008). Sufficient experience exists to suggest that bioflocculation, possibly in combination with centrifugation, could achieve the cost goals for efficient CO2 biofixation and biofuel production. Further study and development of this process remains a central problem, next to productivity and controlled cultivation of specific algal species in the designed photobioreactor (Benemann & Pedroni, 2008).

#### **6.2 Microalgae biomass conversion to fuels**

For economic and environmental reasons the demand for liquid energy from renewable resources will have an ascending trend in the coming year. The advantages of biomass include that it is biodegradable, sustainable and also causes less pollution when compared with fuels being used. Microalgae with high lipid content produces higher biodiesel than commercially used oilseed crops (rapeseed, soybean oil) utilizing less amount of water (Sheehan et al. 1998).

Converting the harvested biomass to a biofuel considered the least difficulty step. The high water content of the harvested biomass makes drying or any thermochemical conversion process (e.g. combustion, gasification, pyrolysis) impractical, and an even more critical problem is the high nitrogen content of algal biomass. Any thermochemical processing would result in unacceptable NO*x* generation and loss of this valuable nutrient and resource. Thus, microalgae biomass fuel conversion processes are dependent on fermentations to produce methane or ethanol, or the metabolism of the algae themselves, to produce oils and hydrocarbons, useable for conversion to biodiesel, or to evolve hydrogen. Methane production from microalgae biomass is technically and economically feasible, but still requires some research and development to improve yields and overall efficiency.

Compared to anaerobic digestion, very little work has been done on ethanol fermentations of algal biomass. The reason is that ethanol fermentations, typically carried out by yeast, are restricted to sugars, starches and similar easily degraded carbohydrates. Microalgae typically contain only about 20% or less of such carbohydrates, present as starch in green algae and glycogen in cyanobacteria. For practical ethanol production, an algal biomass with very high fermentable carbohydrate content, preferably over 60% on a dry weight basis, is required. Such high starch or glycogen accumulation is only observed under conditions of nitrogen limitation, where cell growth is reduced and much or most of the photosynthetically-fixed CO2 is diverted to storage reserves. Thus the issue is whether it is possible to optimize for both high carbohydrate content and high productivity (e.g. CO2 fixation) using nitrogen limitation (Benemann & Pedroni, 2008).

#### **7. Conclusion**

Microalgae have attracted a great deal of attention for CO2 fixation and biofuel production because they can convert CO2 into biomass via photosynthesis at much higher rates than conventional biofuel crops.

Most microalgae exhibit the phenomenon of bioflocculation, which is the spontaneous aggregation of algal cells into large flocs. These flocs will then settle rather rapidly. The process yet not fully understood and need more investigation. It depends on the elaboration of polymers by the algal cells that makes the cells stick together (Benemann & Pedroni, 2008). Sufficient experience exists to suggest that bioflocculation, possibly in combination with centrifugation, could achieve the cost goals for efficient CO2 biofixation and biofuel production. Further study and development of this process remains a central problem, next to productivity and controlled cultivation of specific algal species in the designed

For economic and environmental reasons the demand for liquid energy from renewable resources will have an ascending trend in the coming year. The advantages of biomass include that it is biodegradable, sustainable and also causes less pollution when compared with fuels being used. Microalgae with high lipid content produces higher biodiesel than commercially used oilseed crops (rapeseed, soybean oil) utilizing less amount of water

Converting the harvested biomass to a biofuel considered the least difficulty step. The high water content of the harvested biomass makes drying or any thermochemical conversion process (e.g. combustion, gasification, pyrolysis) impractical, and an even more critical problem is the high nitrogen content of algal biomass. Any thermochemical processing would result in unacceptable NO*x* generation and loss of this valuable nutrient and resource. Thus, microalgae biomass fuel conversion processes are dependent on fermentations to produce methane or ethanol, or the metabolism of the algae themselves, to produce oils and hydrocarbons, useable for conversion to biodiesel, or to evolve hydrogen. Methane production from microalgae biomass is technically and economically feasible, but still requires some research and development to improve yields and overall

Compared to anaerobic digestion, very little work has been done on ethanol fermentations of algal biomass. The reason is that ethanol fermentations, typically carried out by yeast, are restricted to sugars, starches and similar easily degraded carbohydrates. Microalgae typically contain only about 20% or less of such carbohydrates, present as starch in green algae and glycogen in cyanobacteria. For practical ethanol production, an algal biomass with very high fermentable carbohydrate content, preferably over 60% on a dry weight basis, is required. Such high starch or glycogen accumulation is only observed under conditions of nitrogen limitation, where cell growth is reduced and much or most of the photosynthetically-fixed CO2 is diverted to storage reserves. Thus the issue is whether it is possible to optimize for both high carbohydrate content and high productivity (e.g. CO2

Microalgae have attracted a great deal of attention for CO2 fixation and biofuel production because they can convert CO2 into biomass via photosynthesis at much higher rates than

fixation) using nitrogen limitation (Benemann & Pedroni, 2008).

photobioreactor (Benemann & Pedroni, 2008).

**6.2 Microalgae biomass conversion to fuels** 

(Sheehan et al. 1998).

efficiency.

**7. Conclusion** 

conventional biofuel crops.

Several challenges, addressed in this chapter, for microalgal based CO2 sequestration and biofuel production remain. Most studies reported to date have been performed on the bench-scale, and were conducted under strictly controlled conditions. As a result, little is known about the feasibility of the photobioreactor scale-up. Factors, such as supply of adequate amounts of CO2, nutrients and light to microalgal cells, should be investigated and optimized at large scale.

Co-digestion of microalgae with wastewater sludge for biogas production should also be considered, because this strategy could be integrated into the existing wastewater infrastructure.

Microalga-based CO2 fixation and biofuel production can be more sustainable by coupling microalgal biomass production with existing power generation and wastewater treatment infrastructures. Microalgae can utilize low-quality water, such as agricultural runoff or municipal, industrial or agricultural wastewaters, as a source of water for the growth medium as well as a source of nitrogen, phosphorus and minor nutrients (Becker, 1994). Hence, an additional economic and environmental incentive exists as a result of the decreased cost of water and chemicals required for the formulation of the growth medium, while providing a pathway for wastewater treatment (Kumar et al., 2010; Mallick, 2002; Demirbas, 2004). Such a coupling technology need to be further investigated. A number of crucial research gaps remain that must be overcome to achieve full-scale operation such as improved algal growth and nutrient uptake rates; integration of biosystems with waste gas, wastewater and water reclamation systems; improved gas transfer and mixing; and improved algal harvesting and dewatering.

Harvesting, dewatering and lipid extraction from microalgal biomass are still challenging issues because they consume large amounts of energy – mainly because of the small cell size and relatively low biomass levels of microalgal cultures. Therefore, more efficient and economic harvesting technology should be developed to enhance the commercial viability of microalgae biofuels industry (Kumar et al., 2010).

A key challenge for microalgal biodiesel production is the use of microalgal species that can maintain a high growth rate in addition to a high metabolic rate, thus leading to significant lipid yields. This major challenge can be duly addressed via extensive bio-prospecting or target oriented genetic engineering – currently such approaches starting to appear as promising approaches (Kumar et al., 2010).

Finally, it seems there is a lack of fundamental information needed to rationally optimize the performance of existing bioreactors. Novel bioreactor configurations and designs are also needed that promote microalgal growth and CO2 biofixation, characterized by volumetric productivities at least one order of magnitude above those of conventional open pond systems.

#### **8. References**

Acién Fernández, FG.; Fernández Sevilla, JM.; Egorova-Zachernyuk, TA.; Molina Grima, E. (2005). Cost-effective production of 13C, 15N stable isotope-labelled biomass from phototrophic microalgae for various biotechnological applications. *Biomol. Eng*. 22, 193–200.

CO2 Biomitigation and Biofuel Production Using

1, 21–25.

28, 371–380

69

693–698.

*Engineering*, pp. 229–237.

*Technol. Biotechnol*. 85, 387-394

*Biotech. Prog*. 24, 815–820.

*Engineers Journal.* 46(9): 1872-1887.

Microalgae: Photobioreactors Developments and Future Directions 243

Jin, HF.; Lim, BR.; Lee K. (2006). Influence of nitrate feeding on carbon dioxide fixation by microalgae. *J. Environ. Sci. Health A Tox. Hazard Subst. Environ. Eng.* 41, 2813–2824. Kodama, M.; Ikemoto, H.; Miyachi, S. (1993). A new species of highly CO2-tolerant

Kommareddy, A. & Anderson G. (2003). Mixing and Eddie Currents in a Modified Bubble

Kruse, O.; Rupprecht, J.; Mussgnug, J.R.; Dismukes, G.C. & Hankamer, B. (2005).

Kumar, A.; Ergas, S.; Yuan, X.; Sahu, A.; Zhang, Q.; Dewulf, J.; Xavier Malcata, F. &

Kumar, A.; Yuan, X.; Sahu, A. K.; Dewulf, J.; Ergas, S. J.; Van Langenhove, H. (2010).

Lackne, K.S. (2003), Climate Change: A guide to CO2 sequestration. *Science* 300, 1677–1678. Lardon, L.; Hélias, A.; Sialve, B.; Steyer, JP.; Bernard, O. (2009). Life-cycle assessment of biodiesel production from microalgae. *Environ. Sci. Technol.* 43, 6475–6481. Laws, E.A.; Taguchi, S.; Hirata, J.; Pang, L. (1986). Continued studies of high algal

Li Y.; Horsman M.; Wu N.; Lan CQ.; Dubois-Calero N. (2008). Biofuels from microalgae.

Mallick, K. (2002). Biotechnological potential of immobilized algae for wastewater N, P and

Miron, A. S.; Camacho, F. G.; Gomez, A. C.; Grima, E. M. & Chisti, Y. (2000). Bubble-Column

Molina Grima, E.; Acién Fernández, F.G.; Garca Camacho, F.; Camacho Rubio, F. & Chisti, Y. (2000). Scale-up of tubular photobioreactors, *J Appl Phycol* 12:355–68 Molina Grima, E.; Fernandez Sevilla, J. M.; Sanchez Perez, J. A. & Garca Camacho, F. (1996).

Morita, M.; Watanabe Y.; Saiki H. (2000). Investigation of photobioreactor design for

Muhs, J. (2000). Design and analysis of hybrid solar lighting and fullspectrum solar energy

Munoz, R. & Guieysse, B. (2006). Algal-bacterial processes for the treatment of hazardous

and Airlift Photobioreactors for Algal Culture. *American Institute of Chemical* 

A study on simultaneous photolimitation and photoinhibition in dense microalgal cultures taking into account incident and averaged irradiances. *J. Biotechnol.* 45, 59–

enhancing the photosynthetic productivity of microalgae. *Biotechnol. Bioeng.* 69,

systems. Proceedings of the International Solar Energy Conference, *Solar* 

North Fargo, North Dakota, USA October 3-4, 2003

productivities in a shallow flume. *Biomass* 11, 39–50.

metal removal: a review. *Biometals* 15, 377–390.

contaminants: a review. *Water Res.* 40, 2799– 2815

*Photochemical & Photobiological Sciences*, 4 , 957.

fastgrowing marine microalga suitable for high-density culture. *J. Marine Biotechnol.*

Column Reactor . Paper Number: RRV03-0043 An ASAE Meeting Presentation. Written for presentation at the 2003 CSAE/ASAE Annual Intersectional Meeting Sponsored by the Red River Section of ASAE Quality Inn & Suites 301 3rd Avenue

Herman van Langenhove. (2010). Enhanced CO2 fixation and biofuel production via microalgae: recent developments and future directions. *Trends in Biotechnology*

Hollow fiber membrane photo-bioreactor for CO2 sequestration from combustion gas coupled with wastewater treatment: a process engineering approach. *J. Chem.* 


Amos, W.A. (2004). Updated cost analysis of photobiological hydrogen production from

Barbosa, MJ.; Albrecht, M.; Wijffels, RH. (2003). Hydrodynamic stress and lethal events in

Becker, E.W. (1994). In "Microalgae: biotechnology and microbiology" Ed. Baddiley, J. et al.,

Benemann, J.R. & Pedroni, P.M. (2008). Biofixation of Fossil CO2 by Microalgae for

Brennan, L. & Owende, P. 2010. Biofuels from microalgae – a review of technologies for

Camacho Rubio, F.; Acie´n Ferna´ndez, F.G.; Sa´nchez Pe´rez, J.A.; Garca Camacho, F. &

Carlsson, A. S.; Beilen, J. V.; Möller, R.; Clayton, D. (2007). Micro and Micro-algae: Utility for

Carvalho AP.; Meireles, LA.; Malcata FX. (2006). Microalgal reactors: a review of enclosed

Carvalho, A.P. & Malcata, F.X. (2001). Transfer of carbon dioxide within cultures of

Cuello, J.L., et al. (2008). Hybrid solar and electric lighting (HYSEL) for space light support.

Demirbas, A. (2004). Current technologies for the thermo-conversion of biomass into fuels

Doucha, J.; Straka, F.; Lívanský, K. (2005). Utilization of flue gas for cultivation of

Green, B.R. & Durnford, D.G. (1996). The chlorophyll-carotenoid proteins of oxygenic

Halmann, M. (1993). Chemical fixation of carbon dioxide: methods for recycling CO2 into

Harrington, L. Foster, R. (1999). Australian residential building sector greenhouse gas

Hu, Q., Zarmi, Y. and Richmond, A. (1998). Combined effects of light intensity, light path

Jaouen, P.; Vandanjon, L.; Quéméneur, F. (1999). The shear stress of microalgal cell

photosynthesis. *Ann. Rev. Plant Physiol. Plant Mol. Biol*. 47, 685–714.

microalgae: plain bubbling versus hollow-fiber modules. *Biotechnol. Prog.* 17, 265–

microalgae (Chlorella sp.) in an outdoor open thin-layer photobioreactor*. J. Appl.* 

emissions 1990–2010. Final Report, *Energy Efficient Strategies*. Australian

and culture density on output rate of Spirulina platensis (Cyanobacteria). *Eur. J.* 

suspensions (Tetraselmis suecica) in tangential flow filtration systems: the role of

systems design and performances. *Biotechnol. Prog.* 22, 1490–1506.

Chisti, Y. (2007). Biodiesel from microalgae. *Biotechnol. Adv.* 25, 294– 306.

*Presentation at Carbon Recycling Forum* 2008, Arizona

and chemicals. *Energy Source* 26, 715–730

useful products. *CRC Press*, Ann Arbor, MI.

pumps. *Biores. Technol.* 68, 149–154.

Renewable Energy Laborotary, Golden, Colorado, USA.

sparged microalgae culture. *Biotechnol. Bioeng*. 83, 112–120.

178 (1994) Cambridge Univ. Press, Cambridge, New York


272

*Energ. Rev.* 14, 557–577.

Industrial Applications, University of York.

*Bioeng* 62, 71–86.

*Phycol.* 17, 403–412

Greenhouse Office

*Phycol.* 33, 165–171.

chlamydomonas reinhardtii green algae, NREL/MP-560-35593, National

Greenhouse Gas Abatement. *Treccani Encylopedia of Hydrocarbons*, Volume III, p 837

production, processing, and extractions of biofuels and co-products. *Renew. Sust.* 

Molina Grima, E. (1999). Prediction of dissolved oxygen and carbon dioxide concentration profiles in tubular photobioreactors for microalgal culture. *Biotechnol.* 


**10** 

*Canada* 

**Production of Biodiesel from Microalgae** 

Since the early 70s, several major energy crises have forced the scientific community to find alternative sources of power. In July 2008, the price of crude petroleum reached 145\$US/barrel, the highest price ever achieved in 30 years (BP, 2011; Ervin & Associated, 2011). In July 2011, the price of crude petroleum was still relatively high at 108\$US/barrel (Ervin & Associated, 2011). Moreover, the 2008 world economic crisis encouraged the United States government to develop biofuels in order to blend them with petroleum fuels without engine modifications or distribution process changes (Bindraban *et al.*, 2009; The White House, 2010). Furthermore, in Canada, by 2012, the objective of adding biodiesel into transportation diesel and heating fuels is 2% (v/v) (Natural Ressources Canada, 2011b).

Consequently, in July 2007, a nine-year investment of 1.5 billion CAD\$ (the ecoENERGY for Biofuels Initiative) was announced by the Canadian government to stimulate the production

In recent years, most of industrial biodiesels are made from oil (triglycerides) of raw materials (rapeseed, sunflower, soybean, etc.). With the intent to change their physicochemical properties similar to petroleum-based diesel, triglycerides are transesterifed into fatty acid alkyl esters, which can be used in a conventional engine without modifications (Knothe, 2010). On the ecological side, in addition to the ability of oleaginous plants to reduce pollutant emissions of greenhouse gases (GHG) by their capacity to trap and use the carbon dioxide (CO2), using biodiesel also reduces net emissions of pollutants. Typically, the addition of 20% (v/v) of soybean-based biodiesel in petrodiesel reduces emissions of carbon monoxide (CO), CO2, particulate matter (PM) and hydrocarbons (HC) by 11%, 15.5%, 10% and 21%, respectively (Sheehan *et al.*, 1998a; United

The raw materials are also necessary to feed humans and animals. A large demand for raw materials to produce biodiesel could thus increase their price. Moreover, the culture of conventional vegetable material requires an important amount of water, chemical fertilizers

and pesticides, which have a negative impact on the environment (Smith *et al.*, 2009).

of biofuels in Canada (Natural Ressources Canada, 2011a).

States Environmental Protection Agency, 2002).

**1. Introduction** 

 \*

Corresponding Author

Marc Veillette, Mostafa Chamoumi, Josiane Nikiema,

*Chemical Engineering and Biotechnological Engineering Department,* 

Nathalie Faucheux and Michèle Heitz\*

*Université de Sherbrooke* 


## **Production of Biodiesel from Microalgae**

Marc Veillette, Mostafa Chamoumi, Josiane Nikiema, Nathalie Faucheux and Michèle Heitz\* *Chemical Engineering and Biotechnological Engineering Department, Université de Sherbrooke Canada* 

#### **1. Introduction**

244 Advances in Chemical Engineering

Ono, E. & Cuello, J. L. (2003). Selection of optimal microalgae species for CO2 sequestration.

Oswald, W.J. (1988). Large-scale systems (engineering aspects). In Micro-algal

Poulsen, B. R. & J. J. L. Iversen. (1997). Characterization of Gas Transfer and Mixing in a

Pulz, O. (2001). Photobioreactors: production systems for phototrophic microorganisms.

Richmond, A. & Cheng-Wu, Z. (2001). Optimization of a flat plate glass reactor for mass production of Nannochloropsis sp. outdoors. *J. Biotechnol.* 85, 259–269 Richmond, A. et al. (2003). Efficient use of strong light for high photosynthetic productivity:

Sánchez Mirón, A.; Garc´a Camacho, F.; Contreras Gómez, A.; Molina Grima, E. & Chisti,

Sheehan, J.; Dunahay, T.; Benemann, J.; Roessler, P. (1998). *A look back at the US Department of* 

Suh, I.S. & Lee, S.B. (2003). A light distribution model for an internally radiating

Tredici, M. R. (2003). Closed photobioreactors: basic and applied aspects. In Proceedings of Marine Biotechnology: Basics and Applications, p. 1, Matalascan as, Spain Tredici, M.R. (2010). Photobiology of microalgae mass cultures: understanding the tools for

Uduman, N.; Qi, Y.; Danquah, M.K.; Forde, G.M. & Hoadley, A. 2010. Dewatering of

Znad, H.; Báleš, V. & Kawase, K. (2004). Modeling and scale up of airlift bioreactor.

Znad, H.; Kasahara, N. & Kawase, Y. (2006). Biological decomposition of herbicides (EPTC) by activated sludge in a slurry bioreactor. *Process Biochemistry* 41, 1124–1128.

microalgal cultures: a major bottleneck to algae-based fuels. *J. Renew. Sust. Energ.* 2,

Stewart, C. & Hessami, M. A. (2005). Energy Conversion and Management, 46, 403-420. Stucki, S.; Vogel, F.; Ludwig, C.; Haiduc, A. G.; Brandenberger, M. (2009). Catalytic

Renewable Energy Laboratory, July, NERL/TP 580-24190.

Tamiya, H. (1957). Mass culture of algae. *Ann. Rev. Plant Physiol*. 8, 309–334.

VA

46:1872–87

012701.

Cambridge University Press.

*Bioengineering.* 58(6): 633-641.

*Appl. Microbiol. Biotechnol*. 57, 287–293.

growth inhibition. *Biomol. Eng.* 20, 229–236.

capture. *Energy Environ. Sci.* 2, 535–541.

photobioreactor. *Biotechnol. Bioeng.* 82, 180–189.

the next green revolution. *Biofuels* 1, 143–162.

*Computers & Chemical Engineering*, 28, 2765-2777.

Proceedings of Second Annual Conference on Carbon Sequestration. Alexandria,

Biotechnology (Borowitzka, M.A. and Borowitzka, L.J., eds), pp. 357–394,

Bubble Column Equipped with a Rubber Membrane Diffuser. *Biotechnology and* 

interrelationships between the optical path, the population density and the cell

Y.(2000). Bubble column and airlift photobioreactors for algal culture. *AIChE J*

*Energy's aquatic species program - Biodiesel from algae*, Golden (CO), National

gasification of algae in supercritical water for biofuel production and carbon

Since the early 70s, several major energy crises have forced the scientific community to find alternative sources of power. In July 2008, the price of crude petroleum reached 145\$US/barrel, the highest price ever achieved in 30 years (BP, 2011; Ervin & Associated, 2011). In July 2011, the price of crude petroleum was still relatively high at 108\$US/barrel (Ervin & Associated, 2011). Moreover, the 2008 world economic crisis encouraged the United States government to develop biofuels in order to blend them with petroleum fuels without engine modifications or distribution process changes (Bindraban *et al.*, 2009; The White House, 2010). Furthermore, in Canada, by 2012, the objective of adding biodiesel into transportation diesel and heating fuels is 2% (v/v) (Natural Ressources Canada, 2011b).

Consequently, in July 2007, a nine-year investment of 1.5 billion CAD\$ (the ecoENERGY for Biofuels Initiative) was announced by the Canadian government to stimulate the production of biofuels in Canada (Natural Ressources Canada, 2011a).

In recent years, most of industrial biodiesels are made from oil (triglycerides) of raw materials (rapeseed, sunflower, soybean, etc.). With the intent to change their physicochemical properties similar to petroleum-based diesel, triglycerides are transesterifed into fatty acid alkyl esters, which can be used in a conventional engine without modifications (Knothe, 2010). On the ecological side, in addition to the ability of oleaginous plants to reduce pollutant emissions of greenhouse gases (GHG) by their capacity to trap and use the carbon dioxide (CO2), using biodiesel also reduces net emissions of pollutants. Typically, the addition of 20% (v/v) of soybean-based biodiesel in petrodiesel reduces emissions of carbon monoxide (CO), CO2, particulate matter (PM) and hydrocarbons (HC) by 11%, 15.5%, 10% and 21%, respectively (Sheehan *et al.*, 1998a; United States Environmental Protection Agency, 2002).

The raw materials are also necessary to feed humans and animals. A large demand for raw materials to produce biodiesel could thus increase their price. Moreover, the culture of conventional vegetable material requires an important amount of water, chemical fertilizers and pesticides, which have a negative impact on the environment (Smith *et al.*, 2009).

<sup>\*</sup> Corresponding Author

Production of Biodiesel from Microalgae 247

The researches on non-renewable energy are mainly based on coal and natural gas. In 2009, 89% of the energy produced in Canada (15 exaJoules) was coming from non-renewable sources (Statistics Canada, 2011). For example, coal can be converted into syngas (CO, H2) by gasification (Naveed *et al.*, 2010), and then into fuel by Fischer-Tropsch reaction, methanol or synthetic natural gas by catalytic processes and hydrogen by water/gas shift

Renewable energy sources used in the world in 2006 were mainly biomass and waste (58%), hydropower (31%) and others (12%), including wind, geothermal and solar (International Energy Agency, 2008). Most of renewable energies in Canada was produced by hydroelectricity (89%), while biomass was responsible for 6% of the 81 GWatt of renewable energy produced in 2008 (Nyboer & Grove, 2010). In United States, the energy generated from biomass is expected to increase by 300% between 2009 and 2030 to reach 153 billion of 12 kWatt-h (U.S. Energy Information Administration, 2011). Biomass is a source of energy,

First generation biofuels correspond to those issued from food-based crops (Antizar-Ladislao & Turrion-Gomez, 2008). They mainly correspond to ethanol-based fuels obtained from the fermentation of sugars (corn, beet, sugar cane, etc.), vegetable oil-based fuels (raw oil, biodiesel and renewable diesel produced from catalytic hydrodeoxygenation) (Knothe, 2010; Natural Ressources Canada, 2011b) from oleaginous plants (colza, palm, canola, etc.)

However, on many levels (environmental, societal), the fact that food resources could be used to produce biofuels shows several limits, as this would create land pollution, a lack of agricultural land (world hunger) and deforestation (Goldemberg & Guardabassi, 2009; National Research Council, 2007). For example, in some European countries such as France, the arable lands available for cultivation of oleaginous plants used for 1st generation biofuels production will not be able to support the biofuels demand by 2015, except by saturating the lands in fallow, which would create soil impoverishment

Second generation biofuels are the cellulosic-based biofuels obtained from non-food crops materials (wood, leaves, straw, etc.). These biofuels include bioalcohols, bio-oil, 2,5 dimethylfuran (BioDMF), biohydrogen, Fischer-Tropsch diesel, wood diesel (Fatih

Third generation biofuels are microorganisms (yeast, fungi) biofuels and algae-based fuels like vegetable oils, bio-oil, jet-fuels, biohydrogen, biodiesel, renewable diesel and many

Second and 3rd generation biofuels are better than 1st generation biofuels for sustainable development as they are carbon neutral, or they reduce atmospheric CO2 as they are carbon negative (Naik et al., 2010).. For example, 1st generation biodiesel (like soybean) only induces a net reduction of GHG emissions by 41% (Hill *et al.*, 2006). In comparison, for each

reaction (Dry, 1996; Longwell *et al.*, 1995).

**2.1.1 First generation** 

problems (Bordet *et al.*, 2006).

**2.1.2 Second and third generation** 

Demirbas, 2009; Román-Leshkov *et al.*, 2007).

others (Fatih Demirbas, 2009; Nigam & Singh, 2011).

which can be used to produce 1st, 2nd and 3rd generation biofuels.

and biogas emitted from raw material or landfills (Naik *et al.*, 2010).

To overcome these problems, researchers are currently exploring a new way of producing biodiesel using microalgae. Indeed, during their growth, photoautotrophic microalgae metabolize inorganic carbon (CO2) through photosynthesis (Chen *et al.*, 2011). According to Cadoret & Bernard (2008), microalgae have also the capacity to absorb other pollutants such as phosphates and nitrates. Furthermore, microalgae can accumulate a high amount of fatty acids and have a culture yield per hectare at least 10 times higher than any oily plants (Chisti, 2007).

In this paper, the different sources of energy such as oil, conventional and unconventional biofuels like microalgae based biodiesel will be discussed. Then, the different aspects of the microalgae valorization into biodiesel will be considered. The culture of microalgae, the extraction of lipids, the transesterification process and the biodiesel characteristics will be especially discussed.

#### **1.1 Background**

Finding new energy resources to compensate the decrease of the world petroleum reserves is an important challenge. The estimation of the world crude oil reserves is a difficult task because it is influenced by political, economic and technological factors (Pirog, 2005). The proven oil reserves represent the amount of petroleum that could be obtained from deposits already discovered with actual technological and economic conditions (Institut Français du Pétrole, 2005; Pirog, 2005). They also include the oil discovered, but that cannot be recovered with current technologies. In 2009, these reserves were estimated to be about 1,376 billion barrels (BP, 2011), which corresponds to a reserve that will last for 35 years (i.e. until 2045) (Shafiee & Topal, 2009).

According to a predictive model from Huber (1956), the proven reserves of oil should have reached a roof value by 2000, as the production of petroleum should begin to fall. In contrast, more recent data show that between 1989 and 2008, the proven oil reserves seem to have increased from 1006 to 1333 billion barrels (BP, 2011). On the other hand, while the world oil consumption was about 86 million barrels/day in 2006, it would reach 107 million barrels/day by 2030. The transportation sector would be responsible for 80% of this increase (U.S. Energy Information Administration, 2009) and would consume 76% of the world oil production by 2030 (International Energy Agency, 2008).

Fossil fuel dependency, mainly in the transportation sector, has encouraged research on biofuels. A recent study shows that microalgae biofuels have the potential to replace 17% of oil imports in the United States used as transportation fuel by 2022 (Wigmosta *et al.*, 2011). Moreover, following the BP oil spill in the Atlantic Ocean, the United States administration is considering reducing his oil imports by 1/3 by 2021 using, among others, biofuels (N. Banerjee, 2011).

#### **2. Biofuels as a main alternative**

#### **2.1 Conventional biofuels**

Intending to replace oil-based fuels, many studies have been conducted on non-renewable and renewable energy alternatives.

To overcome these problems, researchers are currently exploring a new way of producing biodiesel using microalgae. Indeed, during their growth, photoautotrophic microalgae metabolize inorganic carbon (CO2) through photosynthesis (Chen *et al.*, 2011). According to Cadoret & Bernard (2008), microalgae have also the capacity to absorb other pollutants such as phosphates and nitrates. Furthermore, microalgae can accumulate a high amount of fatty acids and have a culture yield per hectare at least 10 times higher than any oily plants

In this paper, the different sources of energy such as oil, conventional and unconventional biofuels like microalgae based biodiesel will be discussed. Then, the different aspects of the microalgae valorization into biodiesel will be considered. The culture of microalgae, the extraction of lipids, the transesterification process and the biodiesel characteristics will be

Finding new energy resources to compensate the decrease of the world petroleum reserves is an important challenge. The estimation of the world crude oil reserves is a difficult task because it is influenced by political, economic and technological factors (Pirog, 2005). The proven oil reserves represent the amount of petroleum that could be obtained from deposits already discovered with actual technological and economic conditions (Institut Français du Pétrole, 2005; Pirog, 2005). They also include the oil discovered, but that cannot be recovered with current technologies. In 2009, these reserves were estimated to be about 1,376 billion barrels (BP, 2011), which corresponds to a reserve that will last for 35 years (i.e. until 2045)

According to a predictive model from Huber (1956), the proven reserves of oil should have reached a roof value by 2000, as the production of petroleum should begin to fall. In contrast, more recent data show that between 1989 and 2008, the proven oil reserves seem to have increased from 1006 to 1333 billion barrels (BP, 2011). On the other hand, while the world oil consumption was about 86 million barrels/day in 2006, it would reach 107 million barrels/day by 2030. The transportation sector would be responsible for 80% of this increase (U.S. Energy Information Administration, 2009) and would consume 76% of the world oil

Fossil fuel dependency, mainly in the transportation sector, has encouraged research on biofuels. A recent study shows that microalgae biofuels have the potential to replace 17% of oil imports in the United States used as transportation fuel by 2022 (Wigmosta *et al.*, 2011). Moreover, following the BP oil spill in the Atlantic Ocean, the United States administration is considering reducing his oil imports by 1/3 by 2021 using, among others, biofuels (N.

Intending to replace oil-based fuels, many studies have been conducted on non-renewable

(Chisti, 2007).

especially discussed.

(Shafiee & Topal, 2009).

Banerjee, 2011).

**2. Biofuels as a main alternative** 

and renewable energy alternatives.

**2.1 Conventional biofuels** 

production by 2030 (International Energy Agency, 2008).

**1.1 Background** 

The researches on non-renewable energy are mainly based on coal and natural gas. In 2009, 89% of the energy produced in Canada (15 exaJoules) was coming from non-renewable sources (Statistics Canada, 2011). For example, coal can be converted into syngas (CO, H2) by gasification (Naveed *et al.*, 2010), and then into fuel by Fischer-Tropsch reaction, methanol or synthetic natural gas by catalytic processes and hydrogen by water/gas shift reaction (Dry, 1996; Longwell *et al.*, 1995).

Renewable energy sources used in the world in 2006 were mainly biomass and waste (58%), hydropower (31%) and others (12%), including wind, geothermal and solar (International Energy Agency, 2008). Most of renewable energies in Canada was produced by hydroelectricity (89%), while biomass was responsible for 6% of the 81 GWatt of renewable energy produced in 2008 (Nyboer & Grove, 2010). In United States, the energy generated from biomass is expected to increase by 300% between 2009 and 2030 to reach 153 billion of 12 kWatt-h (U.S. Energy Information Administration, 2011). Biomass is a source of energy, which can be used to produce 1st, 2nd and 3rd generation biofuels.

#### **2.1.1 First generation**

First generation biofuels correspond to those issued from food-based crops (Antizar-Ladislao & Turrion-Gomez, 2008). They mainly correspond to ethanol-based fuels obtained from the fermentation of sugars (corn, beet, sugar cane, etc.), vegetable oil-based fuels (raw oil, biodiesel and renewable diesel produced from catalytic hydrodeoxygenation) (Knothe, 2010; Natural Ressources Canada, 2011b) from oleaginous plants (colza, palm, canola, etc.) and biogas emitted from raw material or landfills (Naik *et al.*, 2010).

However, on many levels (environmental, societal), the fact that food resources could be used to produce biofuels shows several limits, as this would create land pollution, a lack of agricultural land (world hunger) and deforestation (Goldemberg & Guardabassi, 2009; National Research Council, 2007). For example, in some European countries such as France, the arable lands available for cultivation of oleaginous plants used for 1st generation biofuels production will not be able to support the biofuels demand by 2015, except by saturating the lands in fallow, which would create soil impoverishment problems (Bordet *et al.*, 2006).

#### **2.1.2 Second and third generation**

Second generation biofuels are the cellulosic-based biofuels obtained from non-food crops materials (wood, leaves, straw, etc.). These biofuels include bioalcohols, bio-oil, 2,5 dimethylfuran (BioDMF), biohydrogen, Fischer-Tropsch diesel, wood diesel (Fatih Demirbas, 2009; Román-Leshkov *et al.*, 2007).

Third generation biofuels are microorganisms (yeast, fungi) biofuels and algae-based fuels like vegetable oils, bio-oil, jet-fuels, biohydrogen, biodiesel, renewable diesel and many others (Fatih Demirbas, 2009; Nigam & Singh, 2011).

Second and 3rd generation biofuels are better than 1st generation biofuels for sustainable development as they are carbon neutral, or they reduce atmospheric CO2 as they are carbon negative (Naik et al., 2010).. For example, 1st generation biodiesel (like soybean) only induces a net reduction of GHG emissions by 41% (Hill *et al.*, 2006). In comparison, for each

Production of Biodiesel from Microalgae 249

245, respectively (FAO, 2011b). In the long term, producing biofuels on arable land could contribute to increasing world hunger. The increase of cereal prices could have an impact on the cost of 1st generation biodiesel production as the FAO Oils/Fats Price Index from 2000 to 2011 also increased from 78 to 278 (FAO, 2011b). In North America, the best lands could be used for biodiesel production, because the harvested area for soybean, which is the most used oil for biodiesel production, represents 32% of the 99.5 million ha harvested worldwide

As microalgae industrial culture does not directly compete with food and wood production, it can represent a great potential economic development. In fact, the use of microalgae for biofuel production would permit to reduce deforestation and preserving the forest heritage. Moreover, the development of valorization of microalgae could favour the energetic autonomy of all countries, including developing countries. Thus, the industrial production of microalgae could be considered as a sustainable solution to energetic, environmental and

Microalgae can be found in a large range of places where light and water are present including ocean, lake, soils, ice, rivers, etc. (Deng *et al.*, 2009). Microalgae demonstrate a great biodiversity (between 200 000 and several millions of species) (Natrah *et al.*, 2007), which can be divided into categories depending on their pigmentation, biological structure

Microalgae are small organisms, which can be divided into 4 size categories as the microplankton (20 to 1000 μm), the nanoplankton (2 to 100 μm), the ultraplankton (0.5 to 15 μm) and the picoplancton (0.2 to 2 μm) (Callieri & Stockner, 2002; Gopinathan, 2004). Their small size allows them to do an effective photosynthesis, converting light energy with CO2

Microalgae can be classified into 4 main taxonomic groups: diatoms (*Bacillariophyceae*), green algae (*Chlorophyceae*), cyanobacteria or blue green algae (*Cyanophyceaes*) and golden algae (*Chrysophyceae*). However, there are 6 other groups of microalgae composed of yellow-green (*Xanthophyceae*), golden algae (*Chrysophyceae*), red algae (*Rhodophyceae*), brown algae (*Phaeophyceae*), dinoflagellates (*Dinophyceae*), *Prasinophyceae and Eustigmatophyceae* (Williams & Laurens, 2010). However, these species are not equally interesting for biodiesel production. As an example, green algae taxonomic group includes the most promising species of microalgae like *Botryococcus* sp., *Dunaliella* sp. and *Chlorella* sp. (Garofalo, 2010)

Microalgae can be separated into 4 main types of metabolism called photoautotrophic,

In photoautotrophic microalgae metabolisms, light (source of energy) converts inorganic carbon (CO2) and water to biomass by photosynthesis reaction (Cadoret & Bernard, 2008).

dissolved in water to produce lipids, carbon hydrates, proteins, etc..

heterotrophic, mixotrophic and photoheterotrophic (Chen *et al.*, 2011).

(FAO, 2011a).

food problematic.

and metabolism. **Size classification** 

**Taxonomic groups** 

for biodiesel production. **Metabolism classification** 

**2.3.2 Types of microalgae** 

ton of microalgal biomass produced, some authors estimate that 1.8 tons of CO2 would be consumed (180% reduction) (Chisti, 2007).

#### **2.2 Microalgae as a source of biofuels**

Microalgae can generate diverse biofuels, which are mainly: biomethane produced by anaerobic digestion (Sialve *et al.*, 2009; Spolaore *et al.*, 2006), biohydrogen by photobiological process (Fedorov *et al.*, 2005; Kapdan & Kargi, 2006), bioethanol by fermentation (Choi *et al.*, 2010; Dexter & Pengcheng, 2009), liquid oil by thermal liquefaction (A. Banerjee *et al.*, 2002; Miao & Wu, 2004; Miao *et al.* 2004; Sawayama *et al.*, 1995) and biodiesel (M. B. Johnson & Wen, 2009; Koberg *et al.*, 2011; Nagle & Lemke, 1990; H. Xu *et al.*, 2006).

Even if industrial scale biofuels from microalgae remain at an early stage, they remain a sustainable solution as a transportation fuel.

#### **2.3 Microalgae for biodiesel**

Some microalgae species like *Botryococcus braunii* or *Schizochytrium* sp. can contain up to 80% of their dry weight of lipids (Deng *et al.*, 2009). These species can produce a yield of lipids by acre up to 770 times higher than oleaginous plants (colza, sunflower, etc.) and their high scale production allows to consider developing high-yield biodiesel (Chisti, 2007). Another advantage of using microalgae to produce biodiesel is that microalgae can double from 1 to 3 times in 24 hours (Khan *et al.*, 2009). Consequently, the microalgae biomass can be harvested more than once a year. Microalgae can potentially be used also in food, cosmetic, fertilizing and many other industries (Jacob-Lopes & Teixeira Franco, 2010).

#### **2.3.1 Sustainable development issues**

#### Environmental issues

Microalgae growth is made naturally from sunlight photosynthesis in diverse media (Deng *et al.*, 2009). Replacing 1st generation biodiesel with biodiesel from microalgae would reduce the carbon footprint of the process. In fact, producing microalgal biodiesel requires no mechanical seeding of grains, no watering or "harmful" chemical products spreading and no harvesting using heavy engines powered by fossil fuels.

The microalgae production can add value to the gases emitted by coal fired power plants (or other processes) (de Morais & Costa, 2007) as they are able to absorb CO2, nitrogen dioxide (NO2) and sulphur dioxide (SO2), which are important nutrients for their growth (Malińska & Zabochnicka-Świątek, 2010). Moreover, some studies have shown that some algae can be grown on municipal *(Chlorella* sp.*)* and industrial wastewaters (*Chlamydomonas globosa, Chlorella minutissima or Scenedesmus bijuga)* since they also use compounds like phosphorus (P) and metals (Al, Fe, Mg, Mn, Zn, etc.) (Chinnasamy *et al.*, 2010; Wang *et al.*, 2010).

#### Economic and social issues

The production of 1st generation biofuels is causing a substantial rise of the world food prices. In order to evaluate the monthly variations for different international food commodities, The FAO Food Price Index can be used. From 2000 to 2011, The FAO Food Price Index (FFPI) and The FAO Cereal Price Index increased from 88 to 238 and from 87 to

ton of microalgal biomass produced, some authors estimate that 1.8 tons of CO2 would be

Microalgae can generate diverse biofuels, which are mainly: biomethane produced by anaerobic digestion (Sialve *et al.*, 2009; Spolaore *et al.*, 2006), biohydrogen by photobiological process (Fedorov *et al.*, 2005; Kapdan & Kargi, 2006), bioethanol by fermentation (Choi *et al.*, 2010; Dexter & Pengcheng, 2009), liquid oil by thermal liquefaction (A. Banerjee *et al.*, 2002; Miao & Wu, 2004; Miao *et al.* 2004; Sawayama *et al.*, 1995) and biodiesel (M. B. Johnson &

Even if industrial scale biofuels from microalgae remain at an early stage, they remain a

Some microalgae species like *Botryococcus braunii* or *Schizochytrium* sp. can contain up to 80% of their dry weight of lipids (Deng *et al.*, 2009). These species can produce a yield of lipids by acre up to 770 times higher than oleaginous plants (colza, sunflower, etc.) and their high scale production allows to consider developing high-yield biodiesel (Chisti, 2007). Another advantage of using microalgae to produce biodiesel is that microalgae can double from 1 to 3 times in 24 hours (Khan *et al.*, 2009). Consequently, the microalgae biomass can be harvested more than once a year. Microalgae can potentially be used also in food, cosmetic,

Microalgae growth is made naturally from sunlight photosynthesis in diverse media (Deng *et al.*, 2009). Replacing 1st generation biodiesel with biodiesel from microalgae would reduce the carbon footprint of the process. In fact, producing microalgal biodiesel requires no mechanical seeding of grains, no watering or "harmful" chemical products spreading and

The microalgae production can add value to the gases emitted by coal fired power plants (or other processes) (de Morais & Costa, 2007) as they are able to absorb CO2, nitrogen dioxide (NO2) and sulphur dioxide (SO2), which are important nutrients for their growth (Malińska & Zabochnicka-Świątek, 2010). Moreover, some studies have shown that some algae can be grown on municipal *(Chlorella* sp.*)* and industrial wastewaters (*Chlamydomonas globosa, Chlorella minutissima or Scenedesmus bijuga)* since they also use compounds like phosphorus

The production of 1st generation biofuels is causing a substantial rise of the world food prices. In order to evaluate the monthly variations for different international food commodities, The FAO Food Price Index can be used. From 2000 to 2011, The FAO Food Price Index (FFPI) and The FAO Cereal Price Index increased from 88 to 238 and from 87 to

(P) and metals (Al, Fe, Mg, Mn, Zn, etc.) (Chinnasamy *et al.*, 2010; Wang *et al.*, 2010).

Wen, 2009; Koberg *et al.*, 2011; Nagle & Lemke, 1990; H. Xu *et al.*, 2006).

fertilizing and many other industries (Jacob-Lopes & Teixeira Franco, 2010).

consumed (180% reduction) (Chisti, 2007).

**2.2 Microalgae as a source of biofuels** 

sustainable solution as a transportation fuel.

**2.3.1 Sustainable development issues** 

no harvesting using heavy engines powered by fossil fuels.

Environmental issues

Economic and social issues

**2.3 Microalgae for biodiesel** 

245, respectively (FAO, 2011b). In the long term, producing biofuels on arable land could contribute to increasing world hunger. The increase of cereal prices could have an impact on the cost of 1st generation biodiesel production as the FAO Oils/Fats Price Index from 2000 to 2011 also increased from 78 to 278 (FAO, 2011b). In North America, the best lands could be used for biodiesel production, because the harvested area for soybean, which is the most used oil for biodiesel production, represents 32% of the 99.5 million ha harvested worldwide (FAO, 2011a).

As microalgae industrial culture does not directly compete with food and wood production, it can represent a great potential economic development. In fact, the use of microalgae for biofuel production would permit to reduce deforestation and preserving the forest heritage. Moreover, the development of valorization of microalgae could favour the energetic autonomy of all countries, including developing countries. Thus, the industrial production of microalgae could be considered as a sustainable solution to energetic, environmental and food problematic.

#### **2.3.2 Types of microalgae**

Microalgae can be found in a large range of places where light and water are present including ocean, lake, soils, ice, rivers, etc. (Deng *et al.*, 2009). Microalgae demonstrate a great biodiversity (between 200 000 and several millions of species) (Natrah *et al.*, 2007), which can be divided into categories depending on their pigmentation, biological structure and metabolism.

#### **Size classification**

Microalgae are small organisms, which can be divided into 4 size categories as the microplankton (20 to 1000 μm), the nanoplankton (2 to 100 μm), the ultraplankton (0.5 to 15 μm) and the picoplancton (0.2 to 2 μm) (Callieri & Stockner, 2002; Gopinathan, 2004). Their small size allows them to do an effective photosynthesis, converting light energy with CO2 dissolved in water to produce lipids, carbon hydrates, proteins, etc..

#### **Taxonomic groups**

Microalgae can be classified into 4 main taxonomic groups: diatoms (*Bacillariophyceae*), green algae (*Chlorophyceae*), cyanobacteria or blue green algae (*Cyanophyceaes*) and golden algae (*Chrysophyceae*). However, there are 6 other groups of microalgae composed of yellow-green (*Xanthophyceae*), golden algae (*Chrysophyceae*), red algae (*Rhodophyceae*), brown algae (*Phaeophyceae*), dinoflagellates (*Dinophyceae*), *Prasinophyceae and Eustigmatophyceae* (Williams & Laurens, 2010). However, these species are not equally interesting for biodiesel production. As an example, green algae taxonomic group includes the most promising species of microalgae like *Botryococcus* sp., *Dunaliella* sp. and *Chlorella* sp. (Garofalo, 2010) for biodiesel production.

#### **Metabolism classification**

Microalgae can be separated into 4 main types of metabolism called photoautotrophic, heterotrophic, mixotrophic and photoheterotrophic (Chen *et al.*, 2011).

In photoautotrophic microalgae metabolisms, light (source of energy) converts inorganic carbon (CO2) and water to biomass by photosynthesis reaction (Cadoret & Bernard, 2008).

Production of Biodiesel from Microalgae 251

studies found that P deprivation could have a positive effect on lipid content (Khozin-Goldberg & Cohen, 2006; Xin *et al.*, 2010). For example, increasing the P concentration from 0.14 to 0.37 mg/L, Xin et al. (2010) observed a raise of microalgae concentration from 0.14 to 0.37 g/L, while the lipid content decreased from 53 to 23.5% (g lipid/g dry

Finally, an osmotic shock might also stimulate the lipids production. For example, Takagi *et al.* (2006) enhanced the sodium chloride (NaCl) concentration from 3.5 to 7 g/L (0.5 to 1 mol/L) and found an increase in lipid production from 60 to 67% (g lipid/g dry weight). However, these physicochemical treatments could also favour the synthesis of polar lipids like phospholipids or glycolipids associated with cell walls of the microalgae (Cadoret & Bernard,

The large scale production of microalgae is generally performed with solar energy (photoautotrophic metabolism) in open ponds (raceways), closed systems (photobioreactors)

Open ponds are generally circular with nested loops and are 30 cm deep (Chisti, 2007). However, ponds can have several non neglectable disadvantages. Indeed, as they are open, evaporation and contaminants (protozoa, bacteria or other microalgae) could affect the

Photobioreactors are continuous culture systems which can achieve concentration of microalgae up to 6.7 g/L (Bai *et al.*, 2011; Chisti, 2007; Ranjbar *et al.*, 2008) in fresh or sea water. Different models of photobioreactors (indoor or outdoor) have been developed including tubular, flate plate, airlift, bubble column and stirred tank (L. Xu *et al.*, 2009). Even if the closed photobioreactor has a higher harvesting efficiency (more biomass) and a good control on culture parameters (temperature, pH, CO2 concentration etc.) (Suh & Lee, 2003), its capital costs remain higher (around 10 times) than those of open ponds (Carvalho *et al.*, 2006). However, the combination of ponds and photobioreactors can be profitable because microalgae can be grown in open ponds while reducing contamination by undesired species (Huntley & Redalje, 2008). In this culture process, the first step of microalgae production is conducted in a controlled temperature (e.g. by a sea water bath (16-18C)) photobioreactor. Microalgae are then transferred into an open pond for a 5 days second culture step (Huntley

Fermentors are mainly used to produce heterotrophic microalgae using an organic source of carbon such as glucose, fructose, galactose acetate, glycerol and acetic acid (Cantin, 2010). These bioreactors can reach high biomass concentration (150 g/L) without rheological problems (Wu & Shi, 2008). However, heterotrophic production costs of microalgae in

Microalgae biomass could be harvested by centrifugation, flocculation, gravity sedimentation, filtration, screening, flotation or by electrophoresis techniques (Chen *et al.*,

2008); such lipids are less interesting for biodiesel production (Nagle & Lemke, 1990).

weight).

or fermentors.

**3. Production of biodiesel 3.1 Culture of microalgae** 

microalgae productivity (Blanco *et al.*, 2007).

& Redalje, 2007; Huntley & Redalje, 2008).

**3.2 Harvesting of microalgae** 

fermentors remains relatively high (Wei *et al.*, 2009).

Even if photoautotrophic microalgae contain high level of lipids, their biomass productivity in photobioreactors and open ponds is generally lower than heterotrophic microalgae, between 0.117 and 1.54 kg/m3/day (Chisti, 2007).

Heterotrophic microalgae need organic carbon as a source of carbon and energy. Their production is carried out in closed bioreactors (fermentors). They are more promising than the photoautotrophic species for the production of biodiesel (Martek, 2008; Xiong *et al.*, 2008; H. Xu *et al.*, 2006). For example, Xiong *et al.* (2008) showed that the biomass productivity obtained from *Chlorella* sp. grew under heterotrophic conditions could reach 7.4 kg/m3/day with lipid content ranging from 50 to 58% (g lipid/g dry weight). However, they do not achieve the main goal of producing microalgae, i.e. mitigation of the emissions of CO2.

Some microalgae species also have mixotrophic metabolism as they grow in light or dark using both inorganic and organic carbon sources. Growing *Chlorella vulgaris* in the dark, Liang *et al.* (2009) changed CO2 contained in air for glucose (as a source of carbon) and observed an increase of maximal biomass productivity from 10 to 151 mg/L/day. When they used light and glucose as a source of carbon, the biomass productivity increased to 254 mg/L/day. The use of microalgae with mixotrophic metabolism is relatively rare for biodiesel production (Chen *et al.*, 2011).

Photoheterotrophic metabolism means that microalgae need light as a source of energy and a source of organic carbon (Chen *et al.*, 2011). For example, using glycerol as a source of carbon and a light intensity of 35 E/m2/s1, Yang *et al.* (2011) obtained an increase in *Chlorella minutissima* (UTEX2341) biomass concentration from 1.2 to 8.2 g/L after 15 days of culture. As these microalgae need an inexpensive source of organic carbon and are dependant of the sunlight periods, photoheterotrophic metabolism seems less interesting for biodiesel production.

#### **2.3.3 Strain promising**

Even if microalgae are a large group of microorganisms, not all species are suitable to produce biodiesel. The factors to consider were stated by Grobbelaar (2000): "1-Carbon dioxide tolerance and uptake, 2-Temperature tolerance, 3-Stability for cultivation in specific bioreactors," 4-Secondary valuable products, "5-Specific growth requirements and competing algal, 6-Vulnerability to infection and herbivotory potential, 7-Excretion of autoinhibitor, 8-Harvesting and down-stream processing and 9-Manipulation potential of genetic engineering".

#### **2.3.4 Lipids yield of microalgae**

Intending to increase lipid production, Huntley & Redalje (2007) proposed a method of growth for microalgae in a photobioreactor (PBR) which allows reaching high lipids yield followed by 2 days of nitrogen deficiency in a pond. In general, nutrient deprivation can lead to an increase in lipid content, but not for all species of microalgae. For example, the microalgae *Navicula* (NAVIC1) had the highest lipid content, which raised in exponential phase from 22 to 49% (g lipid/g dry weight) in silicon (Si) deficiency and increased to 58% (g lipid/g dry weight) when nitrogen was limited. On the other hand, nutrient limitation (nitrogen or Si) limitation had less or no significant effect on lipid content of microalgae *Amphora* (AMPHO1) and *Cyclotella* (CYCLO2), respectively (Sheehan *et al.*, 1998b). Other

Even if photoautotrophic microalgae contain high level of lipids, their biomass productivity in photobioreactors and open ponds is generally lower than heterotrophic microalgae,

Heterotrophic microalgae need organic carbon as a source of carbon and energy. Their production is carried out in closed bioreactors (fermentors). They are more promising than the photoautotrophic species for the production of biodiesel (Martek, 2008; Xiong *et al.*, 2008; H. Xu *et al.*, 2006). For example, Xiong *et al.* (2008) showed that the biomass productivity obtained from *Chlorella* sp. grew under heterotrophic conditions could reach 7.4 kg/m3/day with lipid content ranging from 50 to 58% (g lipid/g dry weight). However, they do not achieve the main goal of producing microalgae, i.e. mitigation of the emissions of CO2.

Some microalgae species also have mixotrophic metabolism as they grow in light or dark using both inorganic and organic carbon sources. Growing *Chlorella vulgaris* in the dark, Liang *et al.* (2009) changed CO2 contained in air for glucose (as a source of carbon) and observed an increase of maximal biomass productivity from 10 to 151 mg/L/day. When they used light and glucose as a source of carbon, the biomass productivity increased to 254 mg/L/day. The use of microalgae with mixotrophic metabolism is relatively rare for

Photoheterotrophic metabolism means that microalgae need light as a source of energy and a source of organic carbon (Chen *et al.*, 2011). For example, using glycerol as a source of carbon and a light intensity of 35 E/m2/s1, Yang *et al.* (2011) obtained an increase in *Chlorella minutissima* (UTEX2341) biomass concentration from 1.2 to 8.2 g/L after 15 days of culture. As these microalgae need an inexpensive source of organic carbon and are dependant of the sunlight periods, photoheterotrophic metabolism seems less interesting for

Even if microalgae are a large group of microorganisms, not all species are suitable to produce biodiesel. The factors to consider were stated by Grobbelaar (2000): "1-Carbon dioxide tolerance and uptake, 2-Temperature tolerance, 3-Stability for cultivation in specific bioreactors," 4-Secondary valuable products, "5-Specific growth requirements and competing algal, 6-Vulnerability to infection and herbivotory potential, 7-Excretion of autoinhibitor, 8-Harvesting and down-stream processing and 9-Manipulation potential of

Intending to increase lipid production, Huntley & Redalje (2007) proposed a method of growth for microalgae in a photobioreactor (PBR) which allows reaching high lipids yield followed by 2 days of nitrogen deficiency in a pond. In general, nutrient deprivation can lead to an increase in lipid content, but not for all species of microalgae. For example, the microalgae *Navicula* (NAVIC1) had the highest lipid content, which raised in exponential phase from 22 to 49% (g lipid/g dry weight) in silicon (Si) deficiency and increased to 58% (g lipid/g dry weight) when nitrogen was limited. On the other hand, nutrient limitation (nitrogen or Si) limitation had less or no significant effect on lipid content of microalgae *Amphora* (AMPHO1) and *Cyclotella* (CYCLO2), respectively (Sheehan *et al.*, 1998b). Other

between 0.117 and 1.54 kg/m3/day (Chisti, 2007).

biodiesel production (Chen *et al.*, 2011).

biodiesel production.

**2.3.3 Strain promising** 

genetic engineering".

**2.3.4 Lipids yield of microalgae** 

studies found that P deprivation could have a positive effect on lipid content (Khozin-Goldberg & Cohen, 2006; Xin *et al.*, 2010). For example, increasing the P concentration from 0.14 to 0.37 mg/L, Xin et al. (2010) observed a raise of microalgae concentration from 0.14 to 0.37 g/L, while the lipid content decreased from 53 to 23.5% (g lipid/g dry weight).

Finally, an osmotic shock might also stimulate the lipids production. For example, Takagi *et al.* (2006) enhanced the sodium chloride (NaCl) concentration from 3.5 to 7 g/L (0.5 to 1 mol/L) and found an increase in lipid production from 60 to 67% (g lipid/g dry weight). However, these physicochemical treatments could also favour the synthesis of polar lipids like phospholipids or glycolipids associated with cell walls of the microalgae (Cadoret & Bernard, 2008); such lipids are less interesting for biodiesel production (Nagle & Lemke, 1990).

### **3. Production of biodiesel**

#### **3.1 Culture of microalgae**

The large scale production of microalgae is generally performed with solar energy (photoautotrophic metabolism) in open ponds (raceways), closed systems (photobioreactors) or fermentors.

Open ponds are generally circular with nested loops and are 30 cm deep (Chisti, 2007). However, ponds can have several non neglectable disadvantages. Indeed, as they are open, evaporation and contaminants (protozoa, bacteria or other microalgae) could affect the microalgae productivity (Blanco *et al.*, 2007).

Photobioreactors are continuous culture systems which can achieve concentration of microalgae up to 6.7 g/L (Bai *et al.*, 2011; Chisti, 2007; Ranjbar *et al.*, 2008) in fresh or sea water. Different models of photobioreactors (indoor or outdoor) have been developed including tubular, flate plate, airlift, bubble column and stirred tank (L. Xu *et al.*, 2009). Even if the closed photobioreactor has a higher harvesting efficiency (more biomass) and a good control on culture parameters (temperature, pH, CO2 concentration etc.) (Suh & Lee, 2003), its capital costs remain higher (around 10 times) than those of open ponds (Carvalho *et al.*, 2006). However, the combination of ponds and photobioreactors can be profitable because microalgae can be grown in open ponds while reducing contamination by undesired species (Huntley & Redalje, 2008). In this culture process, the first step of microalgae production is conducted in a controlled temperature (e.g. by a sea water bath (16-18C)) photobioreactor. Microalgae are then transferred into an open pond for a 5 days second culture step (Huntley & Redalje, 2007; Huntley & Redalje, 2008).

Fermentors are mainly used to produce heterotrophic microalgae using an organic source of carbon such as glucose, fructose, galactose acetate, glycerol and acetic acid (Cantin, 2010). These bioreactors can reach high biomass concentration (150 g/L) without rheological problems (Wu & Shi, 2008). However, heterotrophic production costs of microalgae in fermentors remains relatively high (Wei *et al.*, 2009).

#### **3.2 Harvesting of microalgae**

Microalgae biomass could be harvested by centrifugation, flocculation, gravity sedimentation, filtration, screening, flotation or by electrophoresis techniques (Chen *et al.*,

Production of Biodiesel from Microalgae 253

low lipid yield compared to other species such as *Botryococcus* sp. (28.6% g lipid/g dry weight) (J. Lee *et al.*, 2010). Consequently, in opposition to chemical solvent extraction, supercritical CO2 lipid extraction can be stimulated by the presence of water in the blend of

Some physicochemical techniques like microwave, autoclaving, osmotic shock, beadbeating, homogenization, freeze-drying, French press, grinding and sonication can be used for microalgae cell disrupting in order to recover lipids (Cooney *et al.*, 2009; J. Lee *et al.*, 2010; S. Lee *et al.*, 1998). Using microwave or bead-beating seems to be the most promising techniques to increase the lipid yield. As an example, J. Lee *et al.* (2010) increased the lipid extraction yield of *Botryococcus* sp. microalgae in water phase from 7.7 to 28.6% (g lipid/g

Few studies have used biochemical extraction to extract lipids from microalgae. Using a 72 h cellulase hydrolysis pretreatment of the *Chlorella* sp. microalgae, Fu *et al.* (2010) have obtained a hydrolysis yield of sugars of 70% (concentration reducing sugar/concentration total sugar), although the lipids yield has increased only from 52 to 54% (g lipid/g dry

Direct transesterification is a process that blends the microalgae with an alcohol and a catalyst without prior extraction. Number of acid catalysts have been investigated for heterotrophic microalgae biomass including hydrochloric (HCl) or sulphuric acid (H2SO4) but acetyl chloride (CH3COCl) remains the catalyst producing the higher FAME yield of 56% (g FAME /g dry weight) (Cooney *et al.*, 2009). A less polar solvent, like hexane or chloroform, can be added to increase the yield of biodiesel production (M. B. Johnson & Wen, 2009). Direct transesterification using a heterogeneous catalyst could be more effective coupled with microwaves heating. As an example, using microwave with direct transesterification of *Nannochloropsis* in presence of a heterogeneous catalyst (SrO), Koberg *et al.* (2011) reported an increase in the FAME yield from 7 to 37% (g FAME /g dry weight). However, direct transesterification requires a dry biomass, increasing the cost of

The direct use of crude vegetable oils in diesel engines is envisageable, but could lead to numerous technical problems. For example, their characteristics (high viscosity, high density, difficulty to vaporize in cold conditions) cause deposits in the combustion chamber, with a risk of fouling and an increase in most emissions (Basha *et al.*, 2009). These drawbacks can be mitigated, but not without some modifications of the diesel engine (Altin *et al.*, 2001). To overcome all these inconveniences, the transformation of microalgae lipids in

microalgae.

weight).

harvesting.

**3.4 Transesterification** 

corresponding esters is essential.

**3.3.3 Physicochemical extraction** 

**3.3.4 Biochemical extraction** 

**3.3.5 Direct (in situ) transesterification** 

dry weight) using a 5 min microwave pretreatment.

2011). As microalgae are floating in pond at a concentration less than 0.5 g/L, the harvesting costs can represent 20 to 30% of the industrial microalgae biomass production cost (Carlsson *et al.*, 2007) of 2.95 and 3.80 \$US/kg biomass (photobioreactor and raceway, respectively) (Chisti, 2007).

#### **3.3 Extraction techniques**

In order to produce biodiesel from microalgae lipids, the later must be priory extracted. The main lipid extraction techniques are the use of chemical solvents, supercritical CO2, physicochemical, biochemical and direct transesterification.

#### **3.3.1 Chemical solvents extraction**

Chemical solvents method is by far the most commonly used, but less effective when microalgae are still wet (Samorì *et al.*, 2010). Consequently, for laboratory scale extraction of lipids, freeze-drying (J. Lee *et al.*, 2010) is a popular method, but spray-drying (Koberg *et al.*, 2011), oven-drying (Cooney *et al.*, 2009) or vacuum-evaporation (Umdu *et al.*, 2009) have also been used to dry microalgae. However, drying microalgae prior to lipid extraction could require 2.5 times more energy than a process without drying, which makes a process using a prior drying unprofitable (negative balance) (Lardon *et al.*, 2009).

In laboratory scale studies, even if chloroform-methanol blends have been extensively used with high extraction yields up to 83% (g lipid/g dry weight) (Yaguchi *et al.*, 1997), less polar solvent like hexane are often preferred because of their lower toxicity and affinity for nonlipid contaminants (less polar) (Halim *et al.*, 2010). As an example, hexane was used to obtained lipids content up to 55% (g lipid/g dry weight) from a heterotrophic microalgae, *Chlorella protothecoides* (Miao & Wu, 2006). For microalgae lipid extraction on an industrial scale, Soxhlet extraction is not recommended due to high energy requirement (Halim *et al.*, 2010).

Other less toxic solvents like alcohols (ethanol, octanol) or 1,8-diazabicyclo-[5.4.0]-undec-7 ene (DBU) have been tested but the yield of fatty acid methyl ester (FAME) obtained was up to 5 times lower than with n-hexane extraction (Samorì *et al.*, 2010) even if the hydrocarbon (lipid) yield was more than twice higher.

#### **3.3.2 Supercritical carbon dioxide extraction**

Supercritical CO2 (Dhepe *et al.*, 2003) has the advantages of being not toxic, easy to recover and usable at low temperatures (less than 40C) (Andrich *et al.*, 2005). However, this technique requires expensive equipments (Perrut, 2000) and a huge amount of energy to reach high pressures (Tan & Lee, 2011). Few studies used supercritical CO2 extraction to recover microalgae lipids and transformed them into biodiesel (Halim *et al.*, 2010) even if some studies obtained lipid content up to 26% (g lipid/g dry weight) from *Nannocloropsis*  sp. (Andrich *et al.*, 2005). Using supercritical CO2 extraction at operating temperature of 60C and pressure of 30 MPa to extract lipids from *Chlorococcum* sp. microalgae, Halim *et al.* (2010) obtained a higher extraction yield of lipids with supercritical CO2 than hexane Soxhlet extraction (5.8 and 3.2% (g lipid/g dry weight), respectively). Moreover, using supercritical CO2 extraction with wet microalgae, Halim *et al.* (2010) obtained a maximum yield of lipids of 7.1% (g lipid/g dry weight) for the same experimental conditions, which was a relatively

2011). As microalgae are floating in pond at a concentration less than 0.5 g/L, the harvesting costs can represent 20 to 30% of the industrial microalgae biomass production cost (Carlsson *et al.*, 2007) of 2.95 and 3.80 \$US/kg biomass (photobioreactor and raceway, respectively)

In order to produce biodiesel from microalgae lipids, the later must be priory extracted. The main lipid extraction techniques are the use of chemical solvents, supercritical CO2,

Chemical solvents method is by far the most commonly used, but less effective when microalgae are still wet (Samorì *et al.*, 2010). Consequently, for laboratory scale extraction of lipids, freeze-drying (J. Lee *et al.*, 2010) is a popular method, but spray-drying (Koberg *et al.*, 2011), oven-drying (Cooney *et al.*, 2009) or vacuum-evaporation (Umdu *et al.*, 2009) have also been used to dry microalgae. However, drying microalgae prior to lipid extraction could require 2.5 times more energy than a process without drying, which makes a process using a

In laboratory scale studies, even if chloroform-methanol blends have been extensively used with high extraction yields up to 83% (g lipid/g dry weight) (Yaguchi *et al.*, 1997), less polar solvent like hexane are often preferred because of their lower toxicity and affinity for nonlipid contaminants (less polar) (Halim *et al.*, 2010). As an example, hexane was used to obtained lipids content up to 55% (g lipid/g dry weight) from a heterotrophic microalgae, *Chlorella protothecoides* (Miao & Wu, 2006). For microalgae lipid extraction on an industrial scale, Soxhlet extraction is not recommended due to high energy requirement (Halim *et al.*,

Other less toxic solvents like alcohols (ethanol, octanol) or 1,8-diazabicyclo-[5.4.0]-undec-7 ene (DBU) have been tested but the yield of fatty acid methyl ester (FAME) obtained was up to 5 times lower than with n-hexane extraction (Samorì *et al.*, 2010) even if the hydrocarbon

Supercritical CO2 (Dhepe *et al.*, 2003) has the advantages of being not toxic, easy to recover and usable at low temperatures (less than 40C) (Andrich *et al.*, 2005). However, this technique requires expensive equipments (Perrut, 2000) and a huge amount of energy to reach high pressures (Tan & Lee, 2011). Few studies used supercritical CO2 extraction to recover microalgae lipids and transformed them into biodiesel (Halim *et al.*, 2010) even if some studies obtained lipid content up to 26% (g lipid/g dry weight) from *Nannocloropsis*  sp. (Andrich *et al.*, 2005). Using supercritical CO2 extraction at operating temperature of 60C and pressure of 30 MPa to extract lipids from *Chlorococcum* sp. microalgae, Halim *et al.* (2010) obtained a higher extraction yield of lipids with supercritical CO2 than hexane Soxhlet extraction (5.8 and 3.2% (g lipid/g dry weight), respectively). Moreover, using supercritical CO2 extraction with wet microalgae, Halim *et al.* (2010) obtained a maximum yield of lipids of 7.1% (g lipid/g dry weight) for the same experimental conditions, which was a relatively

physicochemical, biochemical and direct transesterification.

prior drying unprofitable (negative balance) (Lardon *et al.*, 2009).

(Chisti, 2007).

2010).

**3.3 Extraction techniques** 

**3.3.1 Chemical solvents extraction** 

(lipid) yield was more than twice higher.

**3.3.2 Supercritical carbon dioxide extraction** 

low lipid yield compared to other species such as *Botryococcus* sp. (28.6% g lipid/g dry weight) (J. Lee *et al.*, 2010). Consequently, in opposition to chemical solvent extraction, supercritical CO2 lipid extraction can be stimulated by the presence of water in the blend of microalgae.

#### **3.3.3 Physicochemical extraction**

Some physicochemical techniques like microwave, autoclaving, osmotic shock, beadbeating, homogenization, freeze-drying, French press, grinding and sonication can be used for microalgae cell disrupting in order to recover lipids (Cooney *et al.*, 2009; J. Lee *et al.*, 2010; S. Lee *et al.*, 1998). Using microwave or bead-beating seems to be the most promising techniques to increase the lipid yield. As an example, J. Lee *et al.* (2010) increased the lipid extraction yield of *Botryococcus* sp. microalgae in water phase from 7.7 to 28.6% (g lipid/g dry weight) using a 5 min microwave pretreatment.

#### **3.3.4 Biochemical extraction**

Few studies have used biochemical extraction to extract lipids from microalgae. Using a 72 h cellulase hydrolysis pretreatment of the *Chlorella* sp. microalgae, Fu *et al.* (2010) have obtained a hydrolysis yield of sugars of 70% (concentration reducing sugar/concentration total sugar), although the lipids yield has increased only from 52 to 54% (g lipid/g dry weight).

#### **3.3.5 Direct (in situ) transesterification**

Direct transesterification is a process that blends the microalgae with an alcohol and a catalyst without prior extraction. Number of acid catalysts have been investigated for heterotrophic microalgae biomass including hydrochloric (HCl) or sulphuric acid (H2SO4) but acetyl chloride (CH3COCl) remains the catalyst producing the higher FAME yield of 56% (g FAME /g dry weight) (Cooney *et al.*, 2009). A less polar solvent, like hexane or chloroform, can be added to increase the yield of biodiesel production (M. B. Johnson & Wen, 2009). Direct transesterification using a heterogeneous catalyst could be more effective coupled with microwaves heating. As an example, using microwave with direct transesterification of *Nannochloropsis* in presence of a heterogeneous catalyst (SrO), Koberg *et al.* (2011) reported an increase in the FAME yield from 7 to 37% (g FAME /g dry weight). However, direct transesterification requires a dry biomass, increasing the cost of harvesting.

#### **3.4 Transesterification**

The direct use of crude vegetable oils in diesel engines is envisageable, but could lead to numerous technical problems. For example, their characteristics (high viscosity, high density, difficulty to vaporize in cold conditions) cause deposits in the combustion chamber, with a risk of fouling and an increase in most emissions (Basha *et al.*, 2009). These drawbacks can be mitigated, but not without some modifications of the diesel engine (Altin *et al.*, 2001). To overcome all these inconveniences, the transformation of microalgae lipids in corresponding esters is essential.

Production of Biodiesel from Microalgae 255

higher using HCl than NaOH (68 and 1.3% (g FAME/g lipids), respectively) as a catalyst for the same conditions of transesterification (0.1h, 70C). The most used acid catalysts are CH3COCl (Cooney *et al.*, 2009), HCl (Tran *et al.*, 2009), or H2SO4 (Miao & Wu, 2006). Using 4 concentrations of H2SO4 (0.56, 1.13, 1.35 or 2.25 mol/L) and a temperature of 90C, Miao & Wu (2006) found that adding 2.25 mol/L catalyst gave lowest SG (0.863) but a slightly

Different homogenous acid-alkaline (H2SO4-CH3OK, KOH-HCl) catalysts can also be used to transesterify lipids from microalgae (Halim *et al.*, 2010; J. Lee *et al.*, 2010). Using a 1st transesterification step of 2 hours with H2SO4 at 50C and a 2nd step of 2 hours with CH3OK at 55C, Halim *et al.* (2010) obtained a maximum biodiesel yield of 44% (g FAME/g lipid). The purpose to add an alkaline catalyst is mainly to increase the FAME yield

Heterogeneous catalyst used for transesterification of triglycerides can be acid, alkali or

Transesterification studies testing heterogeneous catalysts were conducted with *Nannochloropsis* microalgae using alkaline earth oxide base catalyst such as strontium oxide (SrO) (Koberg *et al.*, 2011), calcium oxide (CaO) or magnesium oxide (MgO) (Umdu *et al.*, 2009). As example, performing a direct transesterification, Koberg *et al.* (2011) reported biodiesel yields up to 99.9% (g biodiesel/g lipid) for a time of reaction of 2 min at 60C for

Few studies used enzymatic catalysts to transesterify lipids from microalgae. Generally, the most common used enzyme is lipase. A biodiesel yield of 98% (g FAME/g lipid) is reported for heterotrophic microalgae *Chlorella protothecoides* lipids transesterified using lipase (*Candida* sp. *99-125*) immobilized on macroporous resins at a concentration of 30% (g/g lipid, 12000 U/g lipid) at 38C during 12h (Li *et al.*, 2007; Xiong *et al.*, 2008). However, enzymatic biodiesel production is often seen expensive due to the cost of the catalyst from 236 to 836 \$US/ton biodiesel (Sotoft *et al.*, 2010). In comparison, the price of acid catalysts like H2SO4 (98%, w/w) is as low as 100 \$US/ton H2SO4 (Tao & Aden, 2009) for a price of catalyst for biodiesel production from 1.4 to 5.3 \$US/ton biodiesel (1 and 5% (w/w) catalyst,

Increasing the reaction time has a positive effect on the SG of the biodiesel produced. For example, performing a direct transesterification of *Chlorella* microalgae at 30C with H2SO4 as catalyst, Ehimen *et al.* (2010) found that the SG decreased from 0.914 to 0.884 when the

The temperature seems to have less effect on the microalgae biodiesel production than reaction time except for high temperatures. For example, Miao & Wu (2006) used H2SO4 catalyst (2.25 mol/L) and found similar biodiesel yield of 56 and 58% (g biodiesel/g lipid) at temperatures of 30 and 50C, respectively. At 90C, the biodiesel yield dropped of about 38%

Stirring can have a positive effect on the biodiesel quality. For example, Ehimen *et al.* (2010) observed a decrease of the SG from 0.9032 (stirring at 500 rpm) to 0.8831 (no stirring).

lowest yield of biodiesel (38%, g biodiesel/g lipid).

(transmethylate acylglycerols and phospholipids) (Halim *et al.*, 2010).

enzymatic (immobilized enzymes) (Helwani *et al.*, 2009).

ratio SrO:microalgae of 30% (g catalyst/g microalgae).

respectively) (Lotero *et al.*, 2005).

(g biodiesel/g lipid).

**3.4.5 Reaction time, temperature and stirring** 

reaction time was increased from 0.25 to 12h.

#### **3.4.1 Reaction**

In the transesterification process, a catalyst and an alcohol are added to a blend of microalgae lipids. The reaction reduces the molecular weight, the viscosity and increases the volatility of microalgae lipids. Different parameters can influence the yield of transesterification like the ratio of alcohol-oil, catalyst types and concentration, reaction time, temperature and agitation rate.

#### **3.4.2 Microalgae lipid content**

The yield of the transesterification reaction depends on the nature of lipids. For example, monoglycerides like palmitic acid (C16:0) produced a FAME yield of 93% (g FAME /g lipid) while triglycerides like triolein had a yield of 88% (g FAME /g lipid). Furthermore, phospholipids and glycolipids gave a lower yield of 54 to 65% (g FAME /g lipid) and 47 to 56% (g FAME /g lipid), respectively (Nagle & Lemke, 1990). The nature of lipids is an important data for biodiesel production because some microalgae can contain up to 93% (g /g lipid) of phospholipids and glycolipids (Williams & Laurens, 2010). Moreover, some microalgae can also contain lipids such as unsaponifiable lipids carotenoids and other elements (chlorophyll) which are considered as by-products (Bai *et al.*, 2011).

#### **3.4.3 Alcohol**

Methanol is the most commonly used alcohol because of its low price. However, other alcohols such as ethanol or butanol can also be employed (Chisti, 2007). In traditional alkalibased catalyst transesterification of vegetable oil, the most used methanol to oil molar ratio for transesterification is 6:1 (Marchetti *et al.*, 2007) even if the stoichiometric value is 3:1 for triglycerides (Berriosa & Skelton, 2008). For microalgae lipids transesterification, the optimal methanol to oil ratio is higher. For example, performing a direct transesterification during 8h at 25C, Ehimen *et al.* (2010) obtained a decrease in the specific gravity (SG) of the biodiesel from 0.8887 to 0.8849 when the molar ratio methanol to oil was increased from 105:1 to 524:1.

#### **3.4.4 Catalyst**

Catalysts used for transesterification of microalgae lipids are mainly homogenous or heterogeneous. Another method of transesterification using methanol in the supercritical state (without catalyst) has been developed, but the cost of this technology renders its use impossible to date (Tan & Lee, 2011).

Homogenous alkaline catalysts used for transesterification of vegetable oils mainly include sodium or potassium hydroxide (NaOH or KOH) and sodium or potassium methoxide (CH3ONa or CH3OK)) while homogenous acid-catalysts includes H2SO4, HCl and sulphonic acids (R-SO3H) (Helwani *et al.*, 2009). In industrial vegetable oil biodiesel, homogenous alkali-catalysed transesterification is commonly used because homogenous acid-catalyzed transesterification is around 4000 times slower (Chisti, 2007) and these catalysts (NaOH or KOH) are relatively less expensive (Helwani *et al.*, 2009).

However, acidic catalyst is preferable for homogenous transesterification of microalgae lipids because the biodiesel yield obtained by Nagle & Lemke (1990) was more than 50 times

In the transesterification process, a catalyst and an alcohol are added to a blend of microalgae lipids. The reaction reduces the molecular weight, the viscosity and increases the volatility of microalgae lipids. Different parameters can influence the yield of transesterification like the ratio of alcohol-oil, catalyst types and concentration, reaction

The yield of the transesterification reaction depends on the nature of lipids. For example, monoglycerides like palmitic acid (C16:0) produced a FAME yield of 93% (g FAME /g lipid) while triglycerides like triolein had a yield of 88% (g FAME /g lipid). Furthermore, phospholipids and glycolipids gave a lower yield of 54 to 65% (g FAME /g lipid) and 47 to 56% (g FAME /g lipid), respectively (Nagle & Lemke, 1990). The nature of lipids is an important data for biodiesel production because some microalgae can contain up to 93% (g /g lipid) of phospholipids and glycolipids (Williams & Laurens, 2010). Moreover, some microalgae can also contain lipids such as unsaponifiable lipids carotenoids and other

Methanol is the most commonly used alcohol because of its low price. However, other alcohols such as ethanol or butanol can also be employed (Chisti, 2007). In traditional alkalibased catalyst transesterification of vegetable oil, the most used methanol to oil molar ratio for transesterification is 6:1 (Marchetti *et al.*, 2007) even if the stoichiometric value is 3:1 for triglycerides (Berriosa & Skelton, 2008). For microalgae lipids transesterification, the optimal methanol to oil ratio is higher. For example, performing a direct transesterification during 8h at 25C, Ehimen *et al.* (2010) obtained a decrease in the specific gravity (SG) of the biodiesel from 0.8887 to 0.8849 when the molar ratio methanol to oil was increased from

Catalysts used for transesterification of microalgae lipids are mainly homogenous or heterogeneous. Another method of transesterification using methanol in the supercritical state (without catalyst) has been developed, but the cost of this technology renders its use

Homogenous alkaline catalysts used for transesterification of vegetable oils mainly include sodium or potassium hydroxide (NaOH or KOH) and sodium or potassium methoxide (CH3ONa or CH3OK)) while homogenous acid-catalysts includes H2SO4, HCl and sulphonic acids (R-SO3H) (Helwani *et al.*, 2009). In industrial vegetable oil biodiesel, homogenous alkali-catalysed transesterification is commonly used because homogenous acid-catalyzed transesterification is around 4000 times slower (Chisti, 2007) and these catalysts (NaOH or

However, acidic catalyst is preferable for homogenous transesterification of microalgae lipids because the biodiesel yield obtained by Nagle & Lemke (1990) was more than 50 times

elements (chlorophyll) which are considered as by-products (Bai *et al.*, 2011).

**3.4.1 Reaction** 

**3.4.3 Alcohol** 

105:1 to 524:1.

**3.4.4 Catalyst** 

impossible to date (Tan & Lee, 2011).

KOH) are relatively less expensive (Helwani *et al.*, 2009).

time, temperature and agitation rate.

**3.4.2 Microalgae lipid content** 

higher using HCl than NaOH (68 and 1.3% (g FAME/g lipids), respectively) as a catalyst for the same conditions of transesterification (0.1h, 70C). The most used acid catalysts are CH3COCl (Cooney *et al.*, 2009), HCl (Tran *et al.*, 2009), or H2SO4 (Miao & Wu, 2006). Using 4 concentrations of H2SO4 (0.56, 1.13, 1.35 or 2.25 mol/L) and a temperature of 90C, Miao & Wu (2006) found that adding 2.25 mol/L catalyst gave lowest SG (0.863) but a slightly lowest yield of biodiesel (38%, g biodiesel/g lipid).

Different homogenous acid-alkaline (H2SO4-CH3OK, KOH-HCl) catalysts can also be used to transesterify lipids from microalgae (Halim *et al.*, 2010; J. Lee *et al.*, 2010). Using a 1st transesterification step of 2 hours with H2SO4 at 50C and a 2nd step of 2 hours with CH3OK at 55C, Halim *et al.* (2010) obtained a maximum biodiesel yield of 44% (g FAME/g lipid). The purpose to add an alkaline catalyst is mainly to increase the FAME yield (transmethylate acylglycerols and phospholipids) (Halim *et al.*, 2010).

Heterogeneous catalyst used for transesterification of triglycerides can be acid, alkali or enzymatic (immobilized enzymes) (Helwani *et al.*, 2009).

Transesterification studies testing heterogeneous catalysts were conducted with *Nannochloropsis* microalgae using alkaline earth oxide base catalyst such as strontium oxide (SrO) (Koberg *et al.*, 2011), calcium oxide (CaO) or magnesium oxide (MgO) (Umdu *et al.*, 2009). As example, performing a direct transesterification, Koberg *et al.* (2011) reported biodiesel yields up to 99.9% (g biodiesel/g lipid) for a time of reaction of 2 min at 60C for ratio SrO:microalgae of 30% (g catalyst/g microalgae).

Few studies used enzymatic catalysts to transesterify lipids from microalgae. Generally, the most common used enzyme is lipase. A biodiesel yield of 98% (g FAME/g lipid) is reported for heterotrophic microalgae *Chlorella protothecoides* lipids transesterified using lipase (*Candida* sp. *99-125*) immobilized on macroporous resins at a concentration of 30% (g/g lipid, 12000 U/g lipid) at 38C during 12h (Li *et al.*, 2007; Xiong *et al.*, 2008). However, enzymatic biodiesel production is often seen expensive due to the cost of the catalyst from 236 to 836 \$US/ton biodiesel (Sotoft *et al.*, 2010). In comparison, the price of acid catalysts like H2SO4 (98%, w/w) is as low as 100 \$US/ton H2SO4 (Tao & Aden, 2009) for a price of catalyst for biodiesel production from 1.4 to 5.3 \$US/ton biodiesel (1 and 5% (w/w) catalyst, respectively) (Lotero *et al.*, 2005).

#### **3.4.5 Reaction time, temperature and stirring**

Increasing the reaction time has a positive effect on the SG of the biodiesel produced. For example, performing a direct transesterification of *Chlorella* microalgae at 30C with H2SO4 as catalyst, Ehimen *et al.* (2010) found that the SG decreased from 0.914 to 0.884 when the reaction time was increased from 0.25 to 12h.

The temperature seems to have less effect on the microalgae biodiesel production than reaction time except for high temperatures. For example, Miao & Wu (2006) used H2SO4 catalyst (2.25 mol/L) and found similar biodiesel yield of 56 and 58% (g biodiesel/g lipid) at temperatures of 30 and 50C, respectively. At 90C, the biodiesel yield dropped of about 38% (g biodiesel/g lipid).

Stirring can have a positive effect on the biodiesel quality. For example, Ehimen *et al.* (2010) observed a decrease of the SG from 0.9032 (stirring at 500 rpm) to 0.8831 (no stirring).

Production of Biodiesel from Microalgae 257

oxidative stability, cold flow properties and lubricity (Knothe *et al.*, 2005). Table 1 presents the main properties of microalgae biodiesel compared with diesel and 1st generation

1. The cetane number is an indicator of quality of ignition of a fuel which increases with the number of carbon and decreases with the number of unsaturated carbon bounds (Hart Energy Consulting, 2007). Consequently, a higher unsaturated biodiesel like microalgae biodiesel would have a lower cetane number. Based on our present knowledge, no measurement of the cetane number of microalgae biodiesel has been performed. However, some studies approximated the cetane numbers of many species based on their FAME content and found cetane numbers ranging from 39 to 54 (Stansell *et al.*, 2011), while cetane number of petrodiesel fuel are at least between 47 and 51

2. The heat of combustion shows if a biodiesel is suitable to burn in a diesel engine. The heat of combustion increases with the length of the carbon chain (Knothe, 2005a). Using lipids extracted from heterotrophic microalgae in the presence of H2SO4 in methanol, Miao & Wu (2006) obtained a biodiesel with a heat of combustion of 35.4 MJ/L which is

3. As the cetane number, the viscosity increases with the number of carbon and decreases with the degree of unsaturation (Knothe, 2005b). A higher kinematic viscosity would create engine problems like engine deposits (Knothe & Steidley, 2005a). Transesterification favours a decrease in the viscosity of the oil at values usually

5. Cold flow properties are important parameters for biodiesel production for northern countries like Canada and could be measured by cloud and pour points. The decrease of temperature could lead to the formation of visible crystals (d ≥ 0.5 m) in the biodiesel at a limit called cloud point (Knothe, 2005). Cloud point temperature decreases with the mole fraction of unsaturated compounds and slightly increases with the length of the carbon chain (Imahara *et al.*, 2006). Pour point is defined as the temperature at which biodiesel does not flow anymore. Usually, cloud and pour points increase as a function of the molar ratio of biodiesel in diesel fuel from 0 to 100% (National Renewable Energy Laboratory, 2009). A higher level of polyunsaturated compounds in microalgae biodiesel could be a benefit in terms of cold properties (cold and poor points) for a blend microalgae biodiesel/petrodiesel in

6. The definition of lubricity for a fuel is "the ability to reduce friction between solid surfaces in relative motion" (Chevron Corporation, 2007; Shumacher, 2005). The lubricity of diesel fuel is influenced by the viscosity, the acidity, the water content and the sulphur compounds (Seregin *et al.*, 1975). Even with additives, the measured friction (no unit reported) of biodiesel (0.114 and 0.117) is lower than the one of petrodiesel

between 4 to 6 mm2/s (40oC) (National Renewable Energy Laboratory, 2009). 4. Oxidation of biodiesel could happen when the FAME are in contact with oxygen and are transformed into hydrogenoperoxides, aldehydes, acids and other oxygenates, which might form deposits (Knothe, 2005). Oxidation of the biodiesel increases as a function of the degree of unsaturation (Hart Energy Consulting, 2007). Oxidation stability of microalgae lipids is therefore a real problem (Stansell *et al.*, 2011) that can be overcome by adding antioxidants if the biodiesel blend is stored more than a few

(ASTM Standard D6751-10, 2010; Knothe, 2006).

months (National Renewable Energy Laboratory, 2009)

in the range of diesel fuel (36-38 MJ/L).

cold climates.

biodiesel.

#### **3.4.6 Purification of biodiesel and by-products**

Microalgae biodiesel and by-products must be separated for increasing the biodiesel production. The main separation processes used hot water (50C) (Li *et al.*, 2007), organic solvents such as hexane (Halim *et al.*, 2010; Wiltshire *et al.*, 2000) and water-organic solvent for a liquid-liquid separation (Couto *et al.*, 2010; Lewis *et al.*, 2000; Samorì *et al.*, 2010). When using a non-polar co-solvent for transesterification of lipids, only water is added to separate biodiesel from the by-products (M. B. Johnson & Wen, 2009).

To our present knowledge, there is no study on the purification of biodiesel from microalgae. Based on 1st generation biodiesel (Leung *et al.*, 2010), three mains means of purification on biodiesel could be applied to microalgae biodiesel purification: "1-water washing 2-dry washing 3-membrane extraction." Based on vegetable oil biodiesel production (Berriosa & Skelton, 2008) and microalgae composition, the main by-products could be unreacted lipids, water, alcohol, chlorophyll, metals and glycerol.

Among the by-products obtained from the biodiesel production, glycerol is the most interesting. Glycerol worldwide consumption remains relatively constant, in recent years, with a consumption of 600 kton/year. Twenty-six percent of glycerol consumption was associated with pharmaceutical, cosmetic and soap industries (Bondioli, 2003). From 2004 to 2011, massive biofuel production created a problem of overproduced glycerol and the price of crude glycerol (80% pure) decreased from 110 to 7.5 \$US/ton (The Jacobsen, 2011; Yazdani & Gonzalez, 2007). Gained glycerol can be transformed into added-value products using many paths including chemical, thermochemical or biological conversion.

Among the chemical added-value products, glycerol can be oxidized or reduced to many compounds like propylene glycol, propionic acid, acrylic acid, propanol, isopropanol, allyl alcohol and acrolein but only some of these products are interesting in terms of market or profitability (D. T. Johnson & Taconi, 2007).

Glycerol can be also converted into Fischer-Tropsch fuel at low temperature (225-300C) by catalytic processes (Soares *et al.*, 2006) or transformed into hydrogen (H2) by catalytic (generally nickel, platinum or ruthenium) or non-catalytic reforming (Vaidya & Rodrigues, 2009).

Biological conversion of glycerol includes fermentation into alcohols (ethanol, butanol, 1,3 propanediol) and other products like H2, formate, propionic or succinic acid (Yazdani & Gonzalez, 2007).

Anaerobic digestion of by-products is another possible way to make biodiesel from microalgae cost effective if the lipid content of the microalgae does not exceed 40% (g lipid/g dry weight) (Sialve *et al.*, 2009). For example, Ehimen *et al.* (2010) used anaerobic digestion of microalgae residues issued from a direct transesterification with a constant loading rate of 5 kg volatile solids (VS)/m3, temperatures and carbon-to-nitrogen (C/N) mass ratio varying from 25 to 40C ant 5.4 to 24, respectively. For a temperature of 40C and a C/N mass ratio of 8.53, a maximum methane (CH4) concentration of 69% (v/v) with a specific CH4 yield of 0.308 m3 CH4/kg VS was obtained.

#### **4. Characteristics of biodiesel**

The physicochemical properties of biodiesel are nearly similar to diesel fuel. The most important properties for biodiesel are cetane number, heat of combustion, viscosity,

Microalgae biodiesel and by-products must be separated for increasing the biodiesel production. The main separation processes used hot water (50C) (Li *et al.*, 2007), organic solvents such as hexane (Halim *et al.*, 2010; Wiltshire *et al.*, 2000) and water-organic solvent for a liquid-liquid separation (Couto *et al.*, 2010; Lewis *et al.*, 2000; Samorì *et al.*, 2010). When using a non-polar co-solvent for transesterification of lipids, only water is added to separate

To our present knowledge, there is no study on the purification of biodiesel from microalgae. Based on 1st generation biodiesel (Leung *et al.*, 2010), three mains means of purification on biodiesel could be applied to microalgae biodiesel purification: "1-water washing 2-dry washing 3-membrane extraction." Based on vegetable oil biodiesel production (Berriosa & Skelton, 2008) and microalgae composition, the main by-products

Among the by-products obtained from the biodiesel production, glycerol is the most interesting. Glycerol worldwide consumption remains relatively constant, in recent years, with a consumption of 600 kton/year. Twenty-six percent of glycerol consumption was associated with pharmaceutical, cosmetic and soap industries (Bondioli, 2003). From 2004 to 2011, massive biofuel production created a problem of overproduced glycerol and the price of crude glycerol (80% pure) decreased from 110 to 7.5 \$US/ton (The Jacobsen, 2011; Yazdani & Gonzalez, 2007). Gained glycerol can be transformed into added-value products

Among the chemical added-value products, glycerol can be oxidized or reduced to many compounds like propylene glycol, propionic acid, acrylic acid, propanol, isopropanol, allyl alcohol and acrolein but only some of these products are interesting in terms of market or

Glycerol can be also converted into Fischer-Tropsch fuel at low temperature (225-300C) by catalytic processes (Soares *et al.*, 2006) or transformed into hydrogen (H2) by catalytic (generally

Biological conversion of glycerol includes fermentation into alcohols (ethanol, butanol, 1,3 propanediol) and other products like H2, formate, propionic or succinic acid (Yazdani &

Anaerobic digestion of by-products is another possible way to make biodiesel from microalgae cost effective if the lipid content of the microalgae does not exceed 40% (g lipid/g dry weight) (Sialve *et al.*, 2009). For example, Ehimen *et al.* (2010) used anaerobic digestion of microalgae residues issued from a direct transesterification with a constant loading rate of 5 kg volatile solids (VS)/m3, temperatures and carbon-to-nitrogen (C/N) mass ratio varying from 25 to 40C ant 5.4 to 24, respectively. For a temperature of 40C and a C/N mass ratio of 8.53, a maximum methane (CH4) concentration of 69% (v/v) with a

The physicochemical properties of biodiesel are nearly similar to diesel fuel. The most important properties for biodiesel are cetane number, heat of combustion, viscosity,

nickel, platinum or ruthenium) or non-catalytic reforming (Vaidya & Rodrigues, 2009).

**3.4.6 Purification of biodiesel and by-products** 

profitability (D. T. Johnson & Taconi, 2007).

specific CH4 yield of 0.308 m3 CH4/kg VS was obtained.

**4. Characteristics of biodiesel** 

Gonzalez, 2007).

biodiesel from the by-products (M. B. Johnson & Wen, 2009).

could be unreacted lipids, water, alcohol, chlorophyll, metals and glycerol.

using many paths including chemical, thermochemical or biological conversion.

oxidative stability, cold flow properties and lubricity (Knothe *et al.*, 2005). Table 1 presents the main properties of microalgae biodiesel compared with diesel and 1st generation biodiesel.



a Calculated

Table 1. Properties of microalgae biodiesel, diesel and biodiesel

#### 258 Advances in Chemical Engineering

Production of Biodiesel from Microalgae 259

European (EN 14214) and American standard (ASTM D6751-10) differ on some levels for biodiesel. For example, EN 14214 stipulates that, polyunsaturated (≥ 4 bonds) ester content and linoleic ester content must be less than 1% (mol/mol) and 12% (mol/mol), respectively (Knothe, 2010). These restrictions are important for biodiesel production from microalgae as the following ester composition is often rich in polyunsaturated content and over these levels (M. B. Johnson & Wen, 2009; Koberg *et al.*, 2011). Some other parameters of EN 14214 are not present in ASTM D6751-10 like esters content, density, iodine value and fatty acid

EN 14214 is more restrictive for biodiesel from microalgae, as the standard of cetane number is higher (51 vs 47), the maximum viscosity is lower (5 vs 6 mm2/s) and the oxidative stability must be higher (6 vs 3 h). Consequently, the ASTM D6751 standard seems to be

Despite the fact that many studies produced biodiesel from microalgae, few studies ensure that the latter satisfies the official standards (M. B. Johnson & Wen, 2009; Miao & Wu, 2006). Producing biodiesel from *Schizochytrium limacinum*, Johnson & Wen (2009) reported a biodiesel that failed to meet the Grade S15 ASTM standard of water and sediments content of 0.05% (v/v) max (0.1% (v/v)) and sulfur content of 15 ppm max (69 ppm) but satisfied the ASTM acid number of 0.5 mg KOH/g max (0.11 mg KOH/g), free glycerine of 0.020% (w/w) max (0.003% w/w), total glycerine of 0.240% (w/w) max (0.097% w/w), corrosiveness to copper of no. 3 max (1a), kinematic viscosity (40C) between 1.9-6.0 (3.87 mm2/s) and flash point (closed cup) of 93C min (204C). In another study, Miao & Wu (2006) satisfied to meet the Grade S15 ASTM standard of flash point (closed cup) (115C), viscosity (5.2 mm2/s, 40C) and acid number (0.374 mg KOH/g). However, based on our present knowledge, methanol content (EN 14110), sulphated ash (D874), cloud point (D2500), cold soak filterability, P content (D4951), distillation temperature (D1160), sodium and potassium (EN 14538), oxidative stability (EN 15751) have not been measured for

With the increase of the price of crude oil in the late 00s, blending biodiesel with petrodiesel appears a sustainable solution to reduce the dependency on oil producing countries. For now, the 1st generation of biofuels currently used could have economic, environmental and social negative consequences. To overcome these problems, producing biodiesel from microalgae lipids seems to be a sustainable solution as microalgae could be used to reduce the CO2 emissions from coal power plants or wastewater pollution. Researchers are working to engineer super lipids producing microalgae strain in order to increase the yield of biodiesel. Producing biodiesel by transesterification of lipids remains relatively costly compared to 1st generation biodiesel. Moreover, the polyunsaturated content of these

reported from the literature.

more adapted to biodiesel from microalgae.

microalgae biodiesel, as microalgae biodiesel is relatively new.

**5. Standards** 

contamination.

**6. Conclusion** 

(0.238 and 0.210) for 25 and 60C (Knothe & Steidley, 2005b). Consequently, a benefit of adding biodiesel in conventional low sulphur diesel fuel is to improve lubricity (Muñoz *et al.*, 2011). For microalgae biodiesel, no lubricity study, to our knowledge, was (0.238 and 0.210) for 25 and 60C (Knothe & Steidley, 2005b). Consequently, a benefit of adding biodiesel in conventional low sulphur diesel fuel is to improve lubricity (Muñoz *et al.*, 2011). For microalgae biodiesel, no lubricity study, to our knowledge, was reported from the literature.

#### **5. Standards**

258 Advances in Chemical Engineering

a Calculated

Table 1. Properties of microalgae biodiesel, diesel and biodiesel

European (EN 14214) and American standard (ASTM D6751-10) differ on some levels for biodiesel. For example, EN 14214 stipulates that, polyunsaturated (≥ 4 bonds) ester content and linoleic ester content must be less than 1% (mol/mol) and 12% (mol/mol), respectively (Knothe, 2010). These restrictions are important for biodiesel production from microalgae as the following ester composition is often rich in polyunsaturated content and over these levels (M. B. Johnson & Wen, 2009; Koberg *et al.*, 2011). Some other parameters of EN 14214 are not present in ASTM D6751-10 like esters content, density, iodine value and fatty acid contamination.

EN 14214 is more restrictive for biodiesel from microalgae, as the standard of cetane number is higher (51 vs 47), the maximum viscosity is lower (5 vs 6 mm2/s) and the oxidative stability must be higher (6 vs 3 h). Consequently, the ASTM D6751 standard seems to be more adapted to biodiesel from microalgae.

Despite the fact that many studies produced biodiesel from microalgae, few studies ensure that the latter satisfies the official standards (M. B. Johnson & Wen, 2009; Miao & Wu, 2006). Producing biodiesel from *Schizochytrium limacinum*, Johnson & Wen (2009) reported a biodiesel that failed to meet the Grade S15 ASTM standard of water and sediments content of 0.05% (v/v) max (0.1% (v/v)) and sulfur content of 15 ppm max (69 ppm) but satisfied the ASTM acid number of 0.5 mg KOH/g max (0.11 mg KOH/g), free glycerine of 0.020% (w/w) max (0.003% w/w), total glycerine of 0.240% (w/w) max (0.097% w/w), corrosiveness to copper of no. 3 max (1a), kinematic viscosity (40C) between 1.9-6.0 (3.87 mm2/s) and flash point (closed cup) of 93C min (204C). In another study, Miao & Wu (2006) satisfied to meet the Grade S15 ASTM standard of flash point (closed cup) (115C), viscosity (5.2 mm2/s, 40C) and acid number (0.374 mg KOH/g). However, based on our present knowledge, methanol content (EN 14110), sulphated ash (D874), cloud point (D2500), cold soak filterability, P content (D4951), distillation temperature (D1160), sodium and potassium (EN 14538), oxidative stability (EN 15751) have not been measured for microalgae biodiesel, as microalgae biodiesel is relatively new.

### **6. Conclusion**

With the increase of the price of crude oil in the late 00s, blending biodiesel with petrodiesel appears a sustainable solution to reduce the dependency on oil producing countries. For now, the 1st generation of biofuels currently used could have economic, environmental and social negative consequences. To overcome these problems, producing biodiesel from microalgae lipids seems to be a sustainable solution as microalgae could be used to reduce the CO2 emissions from coal power plants or wastewater pollution. Researchers are working to engineer super lipids producing microalgae strain in order to increase the yield of biodiesel. Producing biodiesel by transesterification of lipids remains relatively costly compared to 1st generation biodiesel. Moreover, the polyunsaturated content of these

Production of Biodiesel from Microalgae 261

Bondioli, P. (2003). From oil seeds to industrial products: present and near future of

Bordet, J.; Michez, J. & Gilot, A. (2006). *Mise en oeuvre du plan biocarburant au regard de la* 

Bozbas, K. (2008). Biodiesel as an alternative motor fuel: Production and policies in the

BP (2011). BP Statistical Review of World Energy, 07.132011, Available from

Cadoret, J. & Bernard, O. (2008). La production de biocarburant lipidique avec des

Callieri, C. & Stockner, J.G. (2002). Freshwater autotrophic picoplankton: A review. *Journal of* 

Cantin, I. (2010). *La production de biodiesel à partir des microalgues ayant un métabolisme* 

Carlsson, A.S.; van Beilen, J.B.; Moller, R. & Clayton, D. (2007). *Micro- and macro-algae: utility* 

Carvalho, A.P.; Meireles, L.A. & Malcata, F.X. (2006). Microalgal reactors: A review of

Chen, C.; Yeh, K.; Aisyah, R.; Lee, D. & Chang, J. (2011). Cultivation, photobioreactor design

Chevron Corporation (2007). *Diesel Fuels Technical Review.* Chevron Products Company, San

Chinnasamy, S.; Bhatnagar, A.; Hunt, R.W. & Das, K.C. (2010). Microalgae cultivation in a

Chisti, Y. (2007). Biodiesel from microalgae. *Biotechnology Advances,* Vol.25, No.3, pp. 294-306 Choi, S.P.; Nguyen, M.T. & Sim, S.J. (2010). Enzymatic pretreatment of *Chlamydomonas* 

Cooney, M.; Young, G. & Nagle, N. (2009). Extraction of bio-oils from microalgae. *Separation* 

Couto, R.M.; Simões, P.C.; Reis, A.; Da Silva, T.L.; Martins, V.H. & Sánchez-Vicente, Y.

*Crypthecodinium cohnii*. *Engineering in Life Sciences,* Vol.10, No.2, pp. 158-164

http://agriculture.gouv.fr/IMG/pdf/plan\_biocarburants.pdf

/statistical\_review\_of\_world\_energy\_full\_report\_2011.pdf

*for industrial applications.* CPL Press, Newbury, UK, pp. 1-86

*Technology,* Vol.102, No.1, (1), pp. 71-81, ISSN 0960-8524

prodserv/fuels/documents/Diesel\_Fuel\_Tech\_Review.pdf

552

201-211

1506

5336

*Limnology,* Vol.61, No.1, pp. 1-14

Sherbrooke, Québec, Canada, pp. 1-82

*Technology,* Vol.101, No.9, pp. 3097-3105

*& Purification Reviews,* Vol.38, pp. 291-325

oleochemistry. *Italian Journal of Agronomy,* Vol.7, No.2, pp. 129-135, ISSN 1125-4718

*protection de la ressource en eau.* Ministère de l'agriculture, de l'alimentation, de la pêche, de la ruralité et de l'aménagement du territoire, pp. 1-39, Available from

European Union. *Renewable and Sustainable Energy Reviews,* Vol.12, No.2, pp. 542-

http://www.bp.com/liveassets/bp\_internet/globalbp/globalbp\_uk\_english/repo rts\_and\_publications/statistical\_energy\_review\_2011/STAGING/local\_assets/pdf

microalgues : promesses et défis. *Journal de la Société de Biologie,* Vol.202, No.3, pp.

*hétérotrophe.* Maîtrise en environnement (M.Env), Université de Sherbrooke,

enclosed system designs and performances. *Biotechnology Progress,* Vol.22, pp. 1490-

and harvesting of microalgae for biodiesel production: A critical review. *Bioresource* 

Ramon, CA, USA, pp. 1-116, Available from http://www.chevron.com/products/

wastewater dominated by carpet mill effluents for biofuel applications. *Bioresource* 

*reinhardtii* biomass for ethanol production. *Bioresource Technology,* Vol.101, pp. 5330-

(2010). Supercritical fluid extraction of lipids from the heterotrophic microalga

biodiesels could be profitable for cold properties, but seems to cause problems with the cetane number and the oxidative stability to meet the American standard (ASTM). Thus, more efforts must be made to reduce the process costs and to increase the biodiesel quality.

#### **7. Acknowledgment**

The authors are grateful to le Fond Québécois de la Recherche sur la Nature et les Technologies (FQRNT) for the grant to Michèle Heitz and Nathalie Faucheux for the research program in partnership contributing to the reduction of greenhouse gases.

The author for correpondence is Michèle Heitz, a full professor at the Chemical Engineering and Biotechnological Engineering Department at Université de Sherbrooke, 2500 boulevard de l'Université, Sherbrooke, QC, J1K 2R1, Canada (phone: +1-819-821-8000 ext 62827; fax: +1-819-821-7955; e-mail: Michele.Heitz@USherbrooke.ca).

#### **8. References**


biodiesels could be profitable for cold properties, but seems to cause problems with the cetane number and the oxidative stability to meet the American standard (ASTM). Thus, more efforts must be made to reduce the process costs and to increase the biodiesel quality.

The authors are grateful to le Fond Québécois de la Recherche sur la Nature et les Technologies (FQRNT) for the grant to Michèle Heitz and Nathalie Faucheux for the

The author for correpondence is Michèle Heitz, a full professor at the Chemical Engineering and Biotechnological Engineering Department at Université de Sherbrooke, 2500 boulevard de l'Université, Sherbrooke, QC, J1K 2R1, Canada (phone: +1-819-821-8000 ext 62827; fax:

Alptekin, E. & Canakci, M. (2008). Determination of the density and the viscosities of biodiesel-diesel fuel blends. *Renewable Energy,* Vol.33, No.12, pp. 2623-2630 Altin, R.; Çetinkaya, S. & Yücesu, H.S. (2001). Potential of using vegetable oil fuels as fuel for diesel engines. *Energy Conversion and Management,* Vol.42, No.5, pp. 529-538 Andrich, G.; Nesti, U.; Venturi, F.; Zinnai, A. & Fiorentini, R. (2005). Supercritical fluid

Antizar-Ladislao, B. & Turrion-Gomez, J.L. (2008). Second-generation biofuels and local bioenergy systems. *Biofuels, Bioproducts and Biorefining,* Vol.2, No.5, pp. 455-469 ASTM Standard D6751-10, Standard specification for biodiesel fuel blend stock (B100) for

Bai, M.; Cheng, C.; Wan, H. & Lin, Y. (2011). Microalgal pigments potential as byproducts in

Banerjee, A.; Sharma, R.; Chisti, Y. & Banerjee, U.C. (2002). *Botryococcus braunii*: a renewable

Banerjee, N. (2011). *U.S. to reduce oil imports by a third by 2021, Obama says. Los Angeles Times,* 

Basha, S.A.; Gopal, K.R. & Jebaraj, S. (2009). A review on biodiesel production, combustion,

Berriosa, M. & Skelton, R.L. (2008). Comparison of purification methods for biodiesel.

Bindraban, P.S.; Bulte, E.H. & Conijn, S.G. (2009). Can large-scale biofuels production be

Blanco, A.M.; Moreno, J.; Del Campo, J.A.; Rivas, J. & Guerrero, M.G. (2007). Outdoor

sustainable by 2020? *Agricultural Systems,* Vol.101, pp. 197-199

*Journal of Lipid Science and Technology,* Vol.107, No.6, pp. 381-386

middle distillate fuels, Available from www.astm.org

*Chemical Engineering Journal,* Vol.144, pp. 459-465

*and Biotechnology,* Vol.73, No.6, pp. 1259-1266

extraction of bioactive lipids from the microalga *Nannochloropsis* sp.. *European* 

lipid production. *Journal of the Taiwan Institute of Chemical Engineers,* Vol.42, No.5,

source of hydrocarbons and other chemicals. *Critical Reviews in Biotechnology,* 

Available from http://articles.latimes.com/2011/mar/31/nation/la-na-obama-

emissions and performance. *Renewable and Sustainable Energy Reviews,* Vol.13, No.6-

cultivation of lutein-rich cells of *Muriellopsis* sp. in open ponds. *Applied Microbiology* 

research program in partnership contributing to the reduction of greenhouse gases.

+1-819-821-7955; e-mail: Michele.Heitz@USherbrooke.ca).

**7. Acknowledgment** 

**8. References** 

pp. 783-786

energy-20110331/2

7, pp. 1628-1634

Vol.22, No.3, (2002), pp. 245-279


Production of Biodiesel from Microalgae 263

Halim, R.; Gladman, B.; Danquah, M.K. & Webley, P.A. (2010). Oil extraction from

Hart Energy Consulting (2007). *Establishment of the Guidelines for the Development of Biodiesel Standards in the APPEC Region.* Asia Pacific Economic Cooperation, pp. 1-136 Helwani, Z.; Othman, M.R.; Aziz, N.; Kim, J. & Fernando, W.J.N. (2009). Solid

Hill, J.; Nelson, E.; Tilman, D.; Polasky, S. & Tiffany, D. (2006). Environmental, economic,

Huntley, M.E. & Redalje, D.G. (2007). CO2 mitigation and renewable oil from photosynthetic

Huntley, M.E. & Redalje, D.C. (2008). *Continous-batch hybrid process for production of oil and* 

Imahara, H.; Minami, E. & Saka, S. (2006). Thermodynamic study on cloud point of biodiesel with its fatty acid composition. *Fuel,* Vol.85, No.12-13, pp. 1666-1670 Institut Français du Pétrole (2005). Réserves de pétrole : des données évolutives en fonction

 http://www.ifp.fr/content/download/57554/1264970/version/1/ file/IFP-ConfPresse-31mai05\_Reserves\_6\_DonneesEvolutivesSurLesReserves.pdf International Energy Agency (2008). *World Energy outlook 2008.* OECD/IEA, Paris, pp. 1-569, Available from http://www.iea.org/textbase/nppdf/free/2008/weo2008.pdf Jacob-Lopes, E. & Teixeira Franco, T. (2010). Microalgae-based systems for carbon dioxide

Johnson, M.B. & Wen, Z. (2009). Production of biodiesel fuel from the microalga

Johnson, D.T. & Taconi, K.A. (2007). The glycerin glut: Options for the value-added

Kapdan, I.K. & Kargi, F. (2006). Bio-hydrogen production from waste materials. *Enzyme* 

Khan, S.A.; Hussain, M.Z.; Prasad, S. & Banerjee, U. (2009). Prospects of biodiesel

Khozin-Goldberg, I. & Cohen, Z. (2006). The effect of phosphate starvation on the lipid and

Hubbert, K. (1956). Nuclear energy and the fossil fuels, 01.07.2010, Available from

review. *Applied Catalysis,* Vol.363, pp. 1-10

http://www.hubbertpeak.com/hubbert/1956/1956.pdf

No.30, pp. 11206-11210

Vol.12, pp. 573-608

Available from

Vol.23, pp. 5179-5183

7770322, p. 1-9. Available from

http://www.freepatentsonline.com/7770322.html

135-146 978-953-307-113-8, Sciyo, Croatia

*Microbiological Technology,* Vol.38, pp. 569-582

*Phytochemistry,* Vol.67, No.7, pp. 696-701

*Progress,* Vol.26, No.4, pp. 338-348

Vol.13, No.9, pp. 2361-2372

185

microalgae for biodiesel production. *Bioresource Technology,* Vol.102, No.1, pp. 178-

heterogeneous catalysts for transesterification of triglycerides with methanol: A

and energetic costs and benefits of biodiesel and ethanol biofuels. *PNAS,* Vol.103,

microbes: a new appraisal. *Mitigation et Adaptation Strategies for Global Change,* 

*other useful products from photosynthetic microbes,* volume US 2008/0118964, n°

de la technique et de l'économie, In: Institut français du pétrole, 15.06.2010,

sequestration and industrial biorefineries, In: *Biomass,* Momba, M. & Bux, F., pp.

*Schizochytrium limacinum* by direct transesterification of algal biomass. *Energy Fuels,* 

conversion of crude glycerol resulting from biodiesel production. *Environmental* 

production from microalgae in India. *Renewable & Sustainable Energy Reviews,* 

fatty acid composition of the fresh water eustigmatophyte *Monodus subterraneus*.


http://www.kentmarketingservices.com/dnn/ PetroleumPriceData.aspx


de Morais, M.G. & Costa, J.A.V. (2007). Isolation and selection of microalgae from coal fired

Demirbas, A. (2008). Relationships derived from physical properties of vegetable oil and

Deng, X.; Li, Y. & Fei, X. (2009). Microalgae: A promising feedstock for biodiesel. *African* 

Dexter, J. & Pengcheng, F. (2009). Metabolic engineering of cyanobacteria for ethanol

Dhepe, P.L.; Fukuoka, A. & Ichikawa, M. (2003). Novel fabrication and catalysis of nano-

Dry, M.E. (1996). Practical and theoretical aspects of the catalytic Fischer-Tropsch process.

Ehimen, E.A.; Sun, Z.F. & Carrington, C.G. (2010). Variables affecting the in situ

Ehimen, E.A.; Sun, Z.F.; Carrington, C.G.; Birch, E.J. & Eaton-Rye, J.J. (2010). Anaerobic

Ervin, M.J. & Associated (2011). Petroleum Price Data, In: kentmarketingservices.com,

Fatih Demirbas, M. (2009). Biorefineries for biofuel upgrading: A critical review. *Applied* 

Fedorov, A.S.; Kosourov, S.; Ghirardi, M.L. & Seibert, M. (2005). Continuous hydrogen

Fu, C.; Hung, T.; Chen, J.; Su, C. & Wu, W. (2010). Hydrolysis of microalgae cell walls for

Garofalo, R. (2010). Algae and aquatic biomass for a sustainable production of 2nd

Gopinathan, C.P. (2004). Marine microalgae, Proceedings of Ocean Life Food and Medicine

Grobbelaar, J.U. (2000). Physiological and technological considerations for optimising mass

algalcultures. *Journal of Applied Phycology,* Vol.12, pp. 201-206

photoproduction by *Chlamydomonas reinhardtii*: Using a novel two-stage, sulfatelimited chemostat system. *Applied Biochemistry and Biotechnology - Part A Enzyme* 

production of reducing sugar and lipid extraction. *Bioresource Technology,* Vol.101,

generation biofuels, 07.15.2011, Available from http://www.aquaculture.ugent.be Goldemberg, J. & Guardabassi, P. (2009). Are biofuels a feasible option? *Energy Policy,* 

*Management,* Vol.48, No.7, pp. 2169-2173

Vol.5, No.24, pp. 5565-5573

23.07.2011, Available from

2361

biodiesel fuels. *Fuel,* Vol.87, No.8-9, pp. 1743-1748

*Applied Catalysis A: General,* Vol.138, pp. 319-344

*Energy,* Vol.86, No.SUPPL. 1, pp. S151-S161

No.22, (November 2010), pp. 8750-8754

ExpoChennai, India, February 2004

Vol.37, No.1, pp. 10-14

*Journal of Microbiology Research,* Vol.3, No.13, pp. 1008-1014

production. *Energy & Environmental Science,* Vol.2, pp. 857-864

*Applied Energy,* Vol.88, No.10, pp. 3454-3463, ISSN 0306-2619

*Engineering and Biotechnology,* Vol.121, No.1-3, pp. 403-412

 http://www.kentmarketingservices.com/dnn/ PetroleumPriceData.aspx FAO (2011a). FAOSTAT: Crops, 07.142011, Available from http://faostat.fao.org FAO (2011b). FAOSTAT: Food Price Index, In: FAO, 23.05.2011, Available from http://www.fao.org/worldfoodsituation/wfs-home/foodpricesindex/en/

thermoelectric power plant for biofixation of carbon dioxide. *Energy Conversion and* 

structured Rh and RhPt alloy particles occluded in ordered mesoporous silica templates using supercritical carbon dioxide. *Physical Chemistry Chemical Physics,* 

transesterification of microalgae lipids. *Fuel,* Vol.89, No.3, pp. 677-684, ISSN 0016-

digestion of microalgae residues resulting from the biodiesel production process.


http://www.freepatentsonline.com/7770322.html


 http://www.ifp.fr/content/download/57554/1264970/version/1/ file/IFP-ConfPresse-31mai05\_Reserves\_6\_DonneesEvolutivesSurLesReserves.pdf


Production of Biodiesel from Microalgae 265

Lotero, E.; Liu, Y.; Lopez, D.E.; Suwannakarn, K.; Bruce, D.A. & Goodwin Jr., J.G. (2005).

Malińska, K. & Zabochnicka-Świątek, M. (2010). Biosystems for air protection, In: *Air* 

Marchetti, J.M.; Miguel, V.U. & Errazu, A.F. (2007). Possible methods for biodiesel production. *Renewable and Sustainable Energy Reviews,* Vol.11, No.6, pp. 1300-1311 Martek (2008). Martek Biosciences Corporation, In: Martek, 17.06.2010, Available from

Miao, X. & Wu, Q. (2004). High yield bio-oil production from fast pyrolysis by

Miao, X.; Wu, Q. & Yang, C. (2004). Fast pyrolysis of microalgae to produce renewable fuels.

Miao, X. & Wu, Q. (2006). Biodiesel production from heterotrophic microalgal oil. *Bioresource* 

Muñoz, M.; Moreno, F.; Monné, C.; Morea, J. & Terradillos, J. (2011). Biodiesel improves

Nagle, N. & Lemke, P. (1990). Production of methyl-ester fuel from microalgae. *Applied* 

Naik, S.N.; Goud, V.V.; Rout, P.K. & Dalai, A.K. (2010). Production of first and second

Natural Ressources Canada (2011a). Government of Canada calls on industry to participate

Natural Ressources Canada (2011b). Initiative de démonstration nationale sur le diesel

Naveed, S.; Ramzan, N.; Ashraf, A.; Usman, J. & Shabbir, H.A., S. (2010). Modelling and

Nigam, P.S. & Singh, A. (2011). Production of liquid biofuels from renewable resources.

Nyboer, J. & Grove, S. (2010). *A review of renewable energy in Canada, 2008.* pp. 1-25, Available

National Renewable Energy Laboratory (2009). *Biodiesel handling and use guide.* pp. 1-56 National Research Council (2007). Water Implications of Biofuels Production, In: National Research Council, 01.07.2010, Available from http://nationalacademies.org/wstb Natrah, F.M.I.; Yusoff, F.M.; Shariff, M.; Abas, F. & Mariana, N.S. (2007). Screening of

metaboliccontrolling of *Chlorella protothecoides*. *Journal of Biotechnology,* Vol.110, pp.

lubricity of new low sulphur diesel fuels. *Renewable Energy,* Vol.36, No.11, pp. 2918-

generation biofuels: A comprehensive review. *Renewable and Sustainable Energy* 

Malaysian indigenous microalgae for antioxidant properties and nutritional value.

in new biofuels initiative, In: Natural Ressources Canada, 25.05.2011, Available

renouvelable, In: Natural Ressources Canada, 23.05.2011, Available from http://oee.nrcan.gc.ca/transports/carburants-remplacement/programmes/

simulation of coal gasification plant. *Journal of Pakistan Institute of Chemical* 

*Progress in Energy and Combustion Science,* Vol.37, No.1, (2), pp. 52-68, ISSN 0360-

*Pollution,* Villanyi, V., pp. 177-194, InTech, 978-953-307-143-5,

http://aboutmartek.martek.com/researchanddevelopment/

*Journal of Applied Pyrolysis,* Vol.71, No.2004, pp. 855-863

*Biochemistry and Biotechnology,* Vol.24, No.24-25, pp. 355-361

*Journal of Applied Phycology,* Vol.19, No.6, pp. 711-718

idndr/rapport-2010/chap1.cfm?attr=16

*Engineers,* Vol.38, pp. 1-6

from http://www.cieedac.sfu.ca

http://www.nrcan.gc.ca/media/newcom/2007/2007124-eng.php

*Research,* Vol.44, No.14, pp. 5353-5363

*Technology,* Vol.97, No.6, pp. 841-846

*Reviews,* Vol.14, No.2, pp. 578-597

85-93

2924

from

1285

Synthesis of biodiesel via acid catalysis. *Industrial and Engineering Chemistry* 


Knothe, G. (2005). Oxidative stability of biodiesel, In: *The biodiesel handbook,* Knothe, G., Krahl, J. & Gerpen, J. V., pp. 122-126 AOCS Press, Campaign, Illinois, USA Knothe, G.; Gerpen, J.V. & Krahl, J. (2005). *The biodiesel handbook,* AOCS Press, Campaign,

Knothe, G. (2005a). Cetane numbers–heat of combustion–why vegetable oils and their

Knothe, G. (2005b). Viscosity of Biodiesel, In: *The biodiesel handbook,* Knothe, G., Krahl, J. &

Knothe, G. (2006). Analyzing biodiesel: standards and other methods. *Journal of the American* 

Knothe, G. (2010). Biodiesel and renewable diesel: A comparison. *Progress in Energy and* 

Knothe, G. & Steidley, K.R. (2005a). Kinematic viscosity of biodiesel fuel components and

Knothe, G. & Steidley, K.R. (2005b). Lubricity of components of biodiesel and petrodiesel. The origin of biodiesel lubricity. *Energy and Fuels,* Vol.19, No.3, pp. 1192-1200 Koberg, M.; Cohen, M.; Ben-Amotz, A. & Gedanken, A. (2011). Bio-diesel production

Lapuerta, M.; Armas, O. & Rodríguez-Fernández, J. (2008). Effect of biodiesel fuels on diesel

Lardon, L.; Hélias, A.; Sialve, B.; Steyer, J. & Bernard, O. (2009). Life-cycle assessment of

Lee, J.; Yoo, C.; Jun, S.; Ahn, C. & Oh, H. (2010). Comparison of several methods for effective lipid extraction from microalgae. *Bioresource Technology,* Vol.101, pp. 575-577 Lee, S.; Yoon, B. & Oh, H. (1998). Rapid method for the determination of lipid from the green alga *Botryococcus braunii*. *Biotechnology Techniques,* Vol.12, No.7, pp. 553-556 Leung, D.Y.C.; Wu, X. & Leung, M.K.H. (2010). A review on biodiesel production using

Lewis, T.; Nichols, P.D. & McMeekina, T.A. (2000). Evaluation of extraction methods for

Li, X.; Xu, H. & Wu, Q. (2007). Large-scale biodiesel production from microalga *Chlorella* 

Liang, Y.; Sarkany, N. & Cui, Y. (2009). Biomass and lipid productivities of *Chorella vulgaris*

Longwell, J.P.; Rubint, E.S. & Wilson, J. (1995). Coal: Energy for the future. *Progress in Energy* 

recovery of fatty acids from lipid-producing microheterotrophs. *Journal of* 

*protothecoides* through heterotrophic cultivation in bioreactors. *Biotechnology and* 

under autotrophic, heterotrophic and mixotrophic growth conditions. *Biotechnology* 

catalyzed transesterication. *Applied Energy,* Vol.87, pp. 1083-1095

*Microbiological Methods,* Vol.42, pp. 107-116

*Bioengineering,* Vol.98, No.4, pp. 764-771

*and Combustion Science,* Vol.21, pp. 269-360

*Letters,* Vol.31, pp. 1043-1049

related compounds. Influence of compound structure and comparison to

directly from the microalgae biomass of *Nannochloropsis* by microwave and

engine emissions. *Progress in Energy and Combustion Science,* Vol.34, No.2, pp. 198-

biodiesel production from microalgae. *Environmental Science and Technology,* Vol.43,

J. & Gerpen, J. V., pp. 72-80 AOCS Press, Campaign, Illinois, USA

Gerpen, J. V., pp. 81-82 AOCS Press, Campaign, Illinois, USA

petrodiesel fuel components. *Fuel,* Vol.84, No.9, pp. 1059-1065

ultrasound radiation. *Bioresource Technology,* Vol.102, pp. 4265-4269

*Oil Chemists' Society,* Vol.83, No.10, pp. 823-833

*Combustion Science,* Vol.36, No.3, (2010), pp. 364-373

derivatives are suitable as a diesel fuel, In: *The biodiesel handbook,* Knothe, G., Krahl,

Illinois, USA

223

No.17, pp. 6475-6481


Production of Biodiesel from Microalgae 267

Srivastava, A. & Prasad, R. (2000). Triglycerides-based diesel fuels. *Renewable & Sustainable* 

Stansell, G.R.; Gray, V.M. & Sym, S.D. (2011). Microalgal fatty acid composition: implications for biodiesel quality. *Journal of Applied Phycology,* , ISSN 0921-8971 Statistics Canada (2011). *Energy statistics handbook: First quarter 2011.* pp. 1-170, Available

Suh, I.S. & Lee, C. (2003). Photobioreactor engineering: Design and performance.

Takagi, M.; Karseno & Yoshida, T. (2006). Effect of salt concentration on intracellular

Tan, K.T. & Lee, K.T. (2011). A review on supercritical fluids (SCF) technology in sustainable

Tao, L. & Aden, A. (2009). The economics of current and future biofuels. *In Vitro Cellular and* 

The Jacobsen (2011). Renewable fuels: The biodiesel bulletin, 07.04.2011, Available from

The White House (2010). Obama announces steps to boost biofuels, clean coal, In: Office of

Tran, H.; Hong, S. & Lee, C. (2009). Evaluation of extraction methods for recovery of fatty

Tyson, S.K. (2001). *Biodiesel handling and use guidelines.* pp. 1-22, Available from http://www.angelfire.com/ks3/go\_diesel/files042803/biodiesel\_handling.pdf U.S. Energy Information Administration (2009). World proved reserves of oil and natural

U.S. Energy Information Administration (2011). *Annual energy outlook - 2011.* pp. 1-246, Available from http://www.eia.doe.gov/forecasts/aeo/pdf/0383(2011).pdf Umdu, E.S.; Tuncer, M. & Seker, E. (2009). Transesterification of *Nannochloropsis oculata*

*Bioresource Technology,* Vol.100, No.11, (6), pp. 2828-2831, ISSN 0960-8524 United States Environmental Protection Agency (2002). *A comprehensive analysis of biodiesel* 

Vaidya, P.D. & Rodrigues, A.E. (2009). Glycerol reforming for hydrogen production: A review. *Chemical Engineering and Technology,* Vol.32, No.10, pp. 1463-1469 Wang, L.; Min, M.; Li, Y.; Chen, P.; Chen, Y.; Liu, Y.; Wang, Y. & Ruan, R. (2010). Cultivation

Wei, A.; Zhang, X.; Wei, D.; Chen, G.; Wu, Q. & Yang, S. (2009). Effects of cassava starch

acids from *Botryococcus braunii* LB 572 and *Synechocystis* sp. PCC 6803. *Biotechnology* 

gas, most recent estimates, In: U.S. Energy Information Administration, 19.03.2011,

microalga's lipid to biodiesel on Al2O3 supported CaO and MgO catalysts.

of green algae *Chlorella* sp. in different wastewaters from municipal wastewater treatment plant. *Applied Biochemistry and Biotechnology,* Vol.162, No.4, pp. 1174-1186

hydrolysate on cell growth and lipid accumulation of the heterotrophic microalgae

http://www.whitehouse.gov/the-press-office/obama-announces-steps-boost-

accumulation of lipids and triacylglyceride in marine microalgae *Dunaliella* cell.

biodiesel production: Potential and challenges. *Renewable and Sustainable Energy* 

http://www.statcan.gc.ca/pub/57-601-x/57-601-x2011001-eng.pdf

*Journal of Bioscience and Bioengineering,* Vol.101, pp. 223-226

*Developmental Biology - Plant,* Vol.45, No.3, pp. 199-217

the Press Secretary, 23.05.2011, Available from

*and Bioprocess Engineering,* Vol.14, No.2, pp. 187-192

http://www.eia.doe.gov/international/reserves.html

*impacts on exhaust emissions.* pp. 1-118

*Biotechnology and Bioprocess Engineering,* Vol.8, No.6, pp. 313-321

*Energy Reviews,* Vol.4, No.2, pp. 111-133

*Reviews,* Vol.15, No.5, pp. 2452-2456

http://www.thejacobsen.com

biofuels-clean-coal

Available from

from


Perrut, M. (2000). Supercritical fluid applications: Industrial developments and economic issues. *Industrial and Engineering Chemistry Research,* Vol.39, No.12, pp. 4531-4535 Pirog, R. (2005). *CRD Report for Congress: World oil demand and its effect on oil prices.* 

Ranjbar, R.; Inoue, R.; Shiraishi, H.; Katsuda, T. & Katoh, S. (2008). High efficiency

Román-Leshkov, Y.; Barrett, C.J.; Liu, Z.Y. & Dumesic, J.A. (2007). Production of

Samorì, C.; Torri, C.; Samorì, G.; Fabbri, D.; Galletti, P.; Guerrini, F.; Pistocchi, R. &

Sawayama, S.; Inoue, S.; Dote, Y. & Yokoyama, S. (1995). CO2 fixation and oil production through microalga. *Energy Conversion and Management,* Vol.36, pp. 721-731 Seregin, E.P.; Gureev, A.A.; Bugai, V.T.; Makarov, A.A.; Sarantidi, P.G. & Skovorodin, G.B.

Shafiee, S. & Topal, E. (2009). When will fossil fuel reserves be diminished? *Energy Policy,* 

Sheehan, J.; Combreco, V.; Duffield, J.; Graboski, M. & Shapouri, H. (1998a). *An overview of* 

Sheehan, J.; Dunahay, T.; Benemann, J. & Roessler, P. (1998b). *A look back at the U.S.* 

Shumacher, L. (2005). Biodiesel Lubricity, In: *The biodiesel handbook,* Knothe, G., Krahl, J. &

Sialve, B.; Bernet, N. & Bernard, O. (2009). Anaerobic digestion of microalgae as a necessary

Smith, V.H.; Sturm, B.S.M.; deNoyelles, F.J. & Billings, S.A. (2009). The ecology of algal biodiesel production. *Trends in Ecology and Evolution,* Vol.25, No.6, pp. 301-309 Soares, R.R.; Simonetti, D.A. & Dumesic, J.A. (2006). Glycerol as a source for fuels and

Sotoft, L.F.; Rong, B.; Christensen, K.V. & Norddahl, B. (2010). Process simulation and

Spolaore, P.; Joannis-Cassan, C.; Duran, E. & Isambert, A. (2006). Commercial applications of microalgae. *Journal of Bioscience and Bioengineering,* Vol.101, No.2, pp. 87-96

Gerpen, J. V., pp. 127-133 AOCS Press, Campaign, Illinois, USA

Golden, Colorado, USA, pp. 1-47, Available from http://www.nrel.gov/docs /legosti/fy98/24772.pdf

http://www.nrel.gov/docs/legosti/fy98/24190.pdf

*International Edition,* Vol.45, No.24, pp. 3982-3985

*Technology,* Vol.101, No.14, pp. 5266-5274, ISSN 0960-8524

production of astaxanthin by autotrophic cultivation of *Haematococcus pluvialis* in a bubble column photobioreactor. *Biochemical Engineering Journal,* Vol.39, No.3, pp.

dimethylfuran for liquid fuels from biomass-derived carbohydrates. *Nature,* 

Tagliavini, E. (2010). Extraction of hydrocarbons from microalga *Botryococcus braunii* with switchable solvents. *Bioresource Technology,* Vol.101, No.9, pp. 3274-

(1975). Lubricity of diesel fuels. *Chemistry and Technology of Fuels and Oils,* Vol.11,

*biodiesel and petroleum diesel life cycles.* National Renewable Energy Laboratory,

*Department of Energy's aquatic species program—biodiesel from algae.* U.S. Department of Energy's Office of Fuels Development, pp. 1-325, Available from

step to make microalgal biodiesel sustainable. *Biotechnology Advances,* Vol.27, pp.

chemicals by low-temperature catalytic processing. *Angewandte Chemie -* 

economical evaluation of enzymatic biodiesel production plant. *Bioresource* 

Congressional Research Service - The Library of Congress, pp. 1-23

575-580

3279

409-416

No.5, pp. 360-363

Vol.37, pp. 181-189

Vol.447, No.7147, pp. 982-985


http://www.statcan.gc.ca/pub/57-601-x/57-601-x2011001-eng.pdf


http://www.eia.doe.gov/international/reserves.html


**11** 

**Sulfonation/Sulfation Processing** 

Jesús Alfonso Torres Ortega

*Universidad de La Salle* 

*Colombia* 

**Technology for Anionic Surfactant Manufacture** 

In 2008, global production of surfactants was 13 million metric tons reaching a turnover of US\$24,33 million at 2009, which means an increment of 2% from the previous year. Moreover, it is projected a strong growth ca. 2,8% annually till 2012 and between 3,5 – 4% thereafter (Resnik et al., 2010). Sulfonation plants are scattered around the globe in production units with capacities varying from 3.000 to 50.000 tons/year, mainly of anionic surfactants. At least 800 sulfonation plants are estimated to be currently in operation around the World. However, about 20% of the global production (2.500.000 tons/year of sulfonated anionic surfactants) is concentrated in the United States, Western Europe and Japan (Acmite

Anionic surfactants are the key component in a detergent formulation. A molecule of anionic surfactant is composed of a lipophilic oil soluble "tail" (typically an organic molecule C12-C14) and a hydrophilic water soluble "head" (such as SO3−). Mixtures of organic molecules, either form non-renewable resources, such as crude oil or from renewable sources, such as vegetable oils, are currently used as raw materials for household detergents. The cleaning process performed by anionic surfactants (active detergents) is

Sulfonation is the term that identifies an electrophilic chemical reaction where a sulfonic group SO3H is incorporated into a molecule with the capacity to donate electrons. The product of this chemical reaction is recognized as sulfonic acid if the electron donor molecule is a carbon. Sulfuric anhydride reacts easily with delocalized electronic densities as those present in aromatics groups or alkenes in general. These reactions produce a variety of products, including derivate polysulfones. On the other hand, the sulfating process involves the incorporation of the SO3H molecules to an oxygen atom in an organic molecule to form C−O−S bonds and the sulfate group (Figure 1). Sulfates acids can be easily hydrolyzed, and for this reason an immediate neutralization is required after the sulfate group is formed (Foster, 1997). Although sulfonation and sulfating processes are employed industrially to obtain a wide range of products from hair dyes to pesticides and

**1. Introduction** 

Market Intelligence, 2010).

described in the following way (de Groot, 1991):

iii. Retaining the dirt in a stable solution or suspension.

ii. Remotion of dirt from substrate;

i. Wetting of the substrate and dirt due to reduction of surface tension;

*Chlorella protothecoides*. *Journal of Industrial Microbiology and Biotechnology,* Vol.36, No.11, pp. 1383-1389


### **Sulfonation/Sulfation Processing Technology for Anionic Surfactant Manufacture**

Jesús Alfonso Torres Ortega *Universidad de La Salle Colombia* 

#### **1. Introduction**

268 Advances in Chemical Engineering

Wigmosta, M.S.; Coleman, A.M.; Skaggs, R.J.; Huesemann, M.H. & Lane, L.J. (2011).

Williams, P.J. & Laurens, L.M. (2010). Microalgae as biodiesel and biomass feedstocks:

Wiltshire, K.H.; Boersma, M.; Möller, A. & Buhtz, H. (2000). Extraction of pigments and fatty

Wu, Z. & Shi, X. (2008). Rheological properties of *Chlorella pyrenoidosa* culture grown

Xin, L.; Hong-ying, H.; Ke, G. & Ying-xue, S. (2010). Effects of different nitrogen and

Xiong, W.; Li, X.; Xiang, J. & Wu, Q. (2008). High-density fermentation of microalga *Chlorella* 

Xu, H.; Miao, X. & Wu, Q. (2006). High quality biodiesel production from a microalga

Xu, L.; Weathers, P.J.; Xiong, X. & Liu, C. (2009). Microalgal bioreactors: Challenges and

Yaguchi, T.; Tanaka, S.; Yokochi, T.; Nakahara, T. & Higashihara, T. (1997). Production of

Yazdani, S.S. & Gonzalez, R. (2007). Anaerobic fermentation of glycerol: a path to economic

*Journal of the American Oil Chemists' Society,* Vol.74, No.11, pp. 1431-1434 Yang, J.; Rasa, E.; Tantayotai, P.; Scow, K.M.; Yuan, H. & Hristova, K.R. (2011). Mathematical

opportunities. *Engineering in Life Sciences,* Vol.9, No.3, pp. 178-189

*Ressources Research,* Vol.47, No.W00H04, pp. 1-13

*Environmental Science,* Vol.3, pp. 554-590

No.11, pp. 1383-1389

Vol.34, No.2, pp. 119-126

*Biotechnology,* Vol.78, pp. 29-36

*Biotechnology,* Vol.126, pp. 499-507

282

pp. 5494-5500

pp. 3077-3082

213-219

*Chlorella protothecoides*. *Journal of Industrial Microbiology and Biotechnology,* Vol.36,

National microalgae biofuel production potential and resource demand. *Water* 

Review and analysis of the biochemistry, energetics and economics. *Energy &* 

acids from the green alga *Scenedesmus obliquus* (Chlorophyceae). *Aquatic Ecology,* 

heterotrophically in a fermentor. *Journal of Applied Phycology,* Vol.20, No.3, pp. 279-

phosphorus concentrations on the growth, nutrient uptake, and lipid accumulation of a freshwater microalga *Scenedesmus* sp. *Bioresource Technology,* Vol.101, No.14,

*protothecoides* in bioreactor for microbio-diesel production. *Applied Microbiology and* 

*Chlorella protothecoides* by heterotrophic growth in fermenters. *Journal of* 

high yields of docosahexaenoic acid by *Schizochytrium* sp. strain SR21. *JAOCS,* 

model of *Chlorella minutissima* UTEX2341 growth and lipid production under photoheterotrophic fermentation conditions. *Bioresource Technology,* Vol.102, No.3,

viability for the biofuels industry. *Current Opinion in Biotechnology,* Vol.18, No.3, pp.

In 2008, global production of surfactants was 13 million metric tons reaching a turnover of US\$24,33 million at 2009, which means an increment of 2% from the previous year. Moreover, it is projected a strong growth ca. 2,8% annually till 2012 and between 3,5 – 4% thereafter (Resnik et al., 2010). Sulfonation plants are scattered around the globe in production units with capacities varying from 3.000 to 50.000 tons/year, mainly of anionic surfactants. At least 800 sulfonation plants are estimated to be currently in operation around the World. However, about 20% of the global production (2.500.000 tons/year of sulfonated anionic surfactants) is concentrated in the United States, Western Europe and Japan (Acmite Market Intelligence, 2010).

Anionic surfactants are the key component in a detergent formulation. A molecule of anionic surfactant is composed of a lipophilic oil soluble "tail" (typically an organic molecule C12-C14) and a hydrophilic water soluble "head" (such as SO3−). Mixtures of organic molecules, either form non-renewable resources, such as crude oil or from renewable sources, such as vegetable oils, are currently used as raw materials for household detergents. The cleaning process performed by anionic surfactants (active detergents) is described in the following way (de Groot, 1991):


Sulfonation is the term that identifies an electrophilic chemical reaction where a sulfonic group SO3H is incorporated into a molecule with the capacity to donate electrons. The product of this chemical reaction is recognized as sulfonic acid if the electron donor molecule is a carbon. Sulfuric anhydride reacts easily with delocalized electronic densities as those present in aromatics groups or alkenes in general. These reactions produce a variety of products, including derivate polysulfones. On the other hand, the sulfating process involves the incorporation of the SO3H molecules to an oxygen atom in an organic molecule to form C−O−S bonds and the sulfate group (Figure 1). Sulfates acids can be easily hydrolyzed, and for this reason an immediate neutralization is required after the sulfate group is formed (Foster, 1997). Although sulfonation and sulfating processes are employed industrially to obtain a wide range of products from hair dyes to pesticides and

Sulfonation/Sulfation Processing Technology for Anionic Surfactant Manufacture 271

PAS are categorized in different groups regarding the number carbon that compose them: The so called lauryl alcohol sulfates C12-C14, the "tallow" alcohol sulfates (TAS) C16-C18, and the broader cut C10-C18 alcohol sulfate comprising coconut fatty alcohol sulfates. The broad cut (C10-C18) alcohol sulfates presents cost/performance equilibrium in terms of detergency, solubility and foaming properties. This product can partially or totally substitute other anionic surfactants either in liquid or powder detergent formulations with adequate biodegradability and low "defatting action", which is important for human tissue and delicate natural or synthetic fibers. The narrow cut (C12-C14) alcohol sulfates find their main application in a wide range of personal care products such as shampoos, bubble bath products, tooth pastes, dishwashing liquid, delicate products for laundry wash. The C16-C18 alcohol sulfates ("tallow") are used as sodium salts in the formulation of heavy duty laundry products for hand and machine washing. Their detergency power is up to 10% higher than LABS in a wide range of detergent formulations (de Groot, 1991). Furthermore, TAS shows controlled foam, which is important mainly at high temperatures, still keeping the advantage of softness in the wash of sensitive natural and synthetic fibers (Rosen, 2005). The physical detergency and biodegradability of primary alcohols can be affected by the carbon chain length distribution. Therefore, each new supply may require testing to determine whether the desired properties in the chosen application can be achieved. The mechanism for alcohol sulfation is thought to be similar to that for linear alkylbenzene

**2.2 Primary alcohol sulfates (PAS)** 

sulfonation with ∆H = —150 kJ/mol (Figure 3).

**2.3 Alcohol ether sulfates (AES)** 

molecules of ethylene oxide (2EO or 3EO).

Fig. 3. Mechanism of alcohol sulfation (adapted from Roberts, 1998)

Primary alcohol ethoxylates are made by the addition of ethylene oxide to a primary alcohol in the presence of an alkaline catalyst (Boskamp & Houghton, 1996). The addition of the second ethylene oxide molecule to the alcohol is kinetically favored in comparison with the addition of the first ethylene oxide; hence the product of ethoxylation contains a distribution of ethylene oxide chain lengths attached to the alcohol along with the starting alcohol itself. Consequently the physical, detergency and biodegradation characteristics are affected not only by the carbon chain length distribution as is the case for primary alcohols, but also by the ethylene oxide distribution which in turn can be supplier depend (de Groot, 1991). The most common alcohol ethoxylates found as feedstocks for sulfation have an average of 2 to 3

During the sulfating of alcohol ethoxylates the by-products 1,4-dioxane may be formed (Figure 4). Although the formation of 1,4-dioxane is governed predominantly by sulfation and neutralization conditions and by the chemical composition of the feedstock, other factors such as the quality of the raw material also contribute. These factors must be

considered during the store and handling of the alcohol ethoxylate feedstock.

organic intermediates, their main applications are in the production of anionic surfactants (Foster, 2004).

Fig. 1. Functional groups: (a) sulfonate ─*SO3H* and (b) sulfate ─*OSO3H* 

#### **2. Main anionic surfactants**

#### **2.1 Linear alkyl benzene sulfonate (LABS)**

Linear alkylbenzene is the most common organic feedstock employee in the detergent industry (Figure 2). LABS of low molecular weight (230 – 245) lay in the category of anionic surfactants most used in all ranges of household detergent formulations. Dishwashing liquids are prepared from LABS of low molecular weight in combination with other anionic surfactant as Lauryl Ether Sulfate (LES) promoting high detergency, foam stability, degreasing capacity, and high stability in hard water (Zhu et al., 1998). Common concentrations of active detergents in liquid products are: LABS 10-15% (30%), Primary alcohol sulfate/LES 3-5% (10%), where values in brackets are the maximum for concentrated products (Table 1). LABS of high molecular weight (245-260) are the anionic surfactants more used in all ranges of household detergents formulation, but especially in heavy duty laundry products, sometimes in combination with nonionics alcohol sulfates from tallow and soaps (Mungray & Kumar, 2009).

Fig. 2. Sulfonation of alkylbenzene (adapted from Foster, 1997)


Table 1. Heavy powders detergents used in all ranges of household detergents formulation (de Groot, 1991)

#### **2.2 Primary alcohol sulfates (PAS)**

270 Advances in Chemical Engineering

organic intermediates, their main applications are in the production of anionic surfactants

Linear alkylbenzene is the most common organic feedstock employee in the detergent industry (Figure 2). LABS of low molecular weight (230 – 245) lay in the category of anionic surfactants most used in all ranges of household detergent formulations. Dishwashing liquids are prepared from LABS of low molecular weight in combination with other anionic surfactant as Lauryl Ether Sulfate (LES) promoting high detergency, foam stability, degreasing capacity, and high stability in hard water (Zhu et al., 1998). Common concentrations of active detergents in liquid products are: LABS 10-15% (30%), Primary alcohol sulfate/LES 3-5% (10%), where values in brackets are the maximum for concentrated products (Table 1). LABS of high molecular weight (245-260) are the anionic surfactants more used in all ranges of household detergents formulation, but especially in heavy duty laundry products, sometimes in combination with nonionics alcohol sulfates from tallow

Fig. 1. Functional groups: (a) sulfonate ─*SO3H* and (b) sulfate ─*OSO3H* 

(Foster, 2004).

**2. Main anionic surfactants** 

and soaps (Mungray & Kumar, 2009).

(de Groot, 1991)

Fig. 2. Sulfonation of alkylbenzene (adapted from Foster, 1997)

Heavy powders detergents (no soapy) High foam Low foam LABS, high molecular weight (245-260) 20 – 30% 5 – 10% Tallow Alcohol Sulfate (TAS) 2 – 5 % Nonionics 2 – 5 % "Tallow" soap 2 – 5 %

Table 1. Heavy powders detergents used in all ranges of household detergents formulation

**2.1 Linear alkyl benzene sulfonate (LABS)** 

PAS are categorized in different groups regarding the number carbon that compose them: The so called lauryl alcohol sulfates C12-C14, the "tallow" alcohol sulfates (TAS) C16-C18, and the broader cut C10-C18 alcohol sulfate comprising coconut fatty alcohol sulfates. The broad cut (C10-C18) alcohol sulfates presents cost/performance equilibrium in terms of detergency, solubility and foaming properties. This product can partially or totally substitute other anionic surfactants either in liquid or powder detergent formulations with adequate biodegradability and low "defatting action", which is important for human tissue and delicate natural or synthetic fibers. The narrow cut (C12-C14) alcohol sulfates find their main application in a wide range of personal care products such as shampoos, bubble bath products, tooth pastes, dishwashing liquid, delicate products for laundry wash. The C16-C18 alcohol sulfates ("tallow") are used as sodium salts in the formulation of heavy duty laundry products for hand and machine washing. Their detergency power is up to 10% higher than LABS in a wide range of detergent formulations (de Groot, 1991). Furthermore, TAS shows controlled foam, which is important mainly at high temperatures, still keeping the advantage of softness in the wash of sensitive natural and synthetic fibers (Rosen, 2005). The physical detergency and biodegradability of primary alcohols can be affected by the carbon chain length distribution. Therefore, each new supply may require testing to determine whether the desired properties in the chosen application can be achieved. The mechanism for alcohol sulfation is thought to be similar to that for linear alkylbenzene sulfonation with ∆H = —150 kJ/mol (Figure 3).

Fig. 3. Mechanism of alcohol sulfation (adapted from Roberts, 1998)

#### **2.3 Alcohol ether sulfates (AES)**

Primary alcohol ethoxylates are made by the addition of ethylene oxide to a primary alcohol in the presence of an alkaline catalyst (Boskamp & Houghton, 1996). The addition of the second ethylene oxide molecule to the alcohol is kinetically favored in comparison with the addition of the first ethylene oxide; hence the product of ethoxylation contains a distribution of ethylene oxide chain lengths attached to the alcohol along with the starting alcohol itself. Consequently the physical, detergency and biodegradation characteristics are affected not only by the carbon chain length distribution as is the case for primary alcohols, but also by the ethylene oxide distribution which in turn can be supplier depend (de Groot, 1991). The most common alcohol ethoxylates found as feedstocks for sulfation have an average of 2 to 3 molecules of ethylene oxide (2EO or 3EO).

During the sulfating of alcohol ethoxylates the by-products 1,4-dioxane may be formed (Figure 4). Although the formation of 1,4-dioxane is governed predominantly by sulfation and neutralization conditions and by the chemical composition of the feedstock, other factors such as the quality of the raw material also contribute. These factors must be considered during the store and handling of the alcohol ethoxylate feedstock.

Sulfonation/Sulfation Processing Technology for Anionic Surfactant Manufacture 273

oligomerisation of ethylene. The physical, detergency and biodegradation characteristics of alfa-olefins are affected by the carbon chain length distribution and therefore each new supply may require testing to determine whether the desired properties for the new chosen application can be achieved. The Lion Corporation, Japan, is one of the principal producers and users of alfa-olefin sulfonates. In addition to fabric washing powders, they also market fabric washing liquid, shampoos, tooth paste and foam bath products containing this active. In the USA, Minnetonka has utilized AOS in hand cleaners/liquid soaps. AOS is a potential replacement for alkyl benzene sulfonates in dish wash detergent liquids formulations with

FAMES are called to be the main feedstock for detergent formulating in the future due to their applicability in detergent formulations (Ingegar & Martin, 2001; Johansson & Svensson, 2001; Roberts & Garrett, 2000; Satsuki, 1998). Moreover, when it is derived from palm oil presents special biodegradable properties that place them over the surfactants derived from petrochemicals compounds. To date, the application of FAMES is under development in various detergent products, and their presence on the market is still highly restricted. The typical cut of FAMES (C16-C18) shows interesting surface activity (about 90% compared to LABS), high detergent, dispersing and emulsifying power in hard water, high lime soap dispersion and moderate foam levels. FAMES show high stability to pH and temperature hydrolysis. Therefore, they can be easily spray dyer and/or incorporated in detergent bars. Methyl ester sulfonates have a wide range of application and important biological properties. As aggregated value the FAMES can be used in cosmetics, as auxiliary agents in the production of fibers, plastics, and rubber, and in leather manufacture (Cohen et al., 2008;

**3. Sulfonation process used for the manufacturing of anionic surfactants** 

Sulfonation reactions can be carried out in different configurations, either liquid-liquid contact, or gas-liquid contact reactors, and a diversity of sulfonating reagents can be applied for the sulfonation process, such as: Sulfuric acid, SO3 from stabilized liquid SO3, SO3 from sulfur burning and subsequent conversion of the SO2 formed, SO3 from boiling concentrated oleum and chlorosulfonic acid. However some reasons why SO3/air in gas-liquid contactor (sulfonator) is becoming the predominant process for the manufacture of anionic surfactants

Fig. 5. Reactions of alfa-olefin sulfonation (adapted from de Groot, 1991)

performance peaking at C16 chain length (de Groot., 1991).

de Groot, 1991; Roberts et al., 2008; Stein & Baumann, 1975).

are (Foster, 1997):

**2.5 Fatty acid methyl esters sulfonates (FAMES)** 

$$\begin{array}{ll} \text{R-O-CH}\_{2}\text{CH}\_{2}\text{(OCH}\_{2}\text{CH}\_{2}\text{)}\_{n}\text{OH} + 2\text{SO}\_{3} \xrightarrow[\text{(fast)}]{} \text{RO}^{+}\text{CH}\_{2}\text{CH}\_{2}\text{(OCH}\_{2}\text{CH}\_{2}\text{)}\_{n}\text{OSO}\_{3} \\ \text{RO}^{+}\text{CH}\_{2}\text{CH}\_{2}\text{(OCH}\_{2}\text{CH}\_{2}\text{)}\_{n}\text{OSO}\_{3} + \text{R}(\text{O-CH}\_{2}\text{CH}\_{2})\_{n+1}\text{-OH} \xrightarrow[\text{(less fast)}]{} 2\text{R}(\text{OCH}\_{2}\text{CH}\_{2}\text{)}\_{n+1}\text{OSO}\_{3} \text{H}^{+} \\ \text{SO}\_{5}^{\cdot} \\ \text{R}(\text{OCH}\_{2}\text{CH}\_{2})\_{n}\text{OSO}\_{3}\text{H} \xrightarrow{\text{H}^{+}} 2\text{R}(\text{OCH}\_{2}\text{CH}\_{2})\_{n+2}\text{OSO}\_{3}\text{H} + \begin{array}{c} \text{O} \\ \text{H}\_{2}\text{C} \overset{\text{I}}{\underset{}{\text{C}}} \\ \text{H}\_{2}\text{C} \overset{\text{I}}{\underset{}{\text{C}}} \\ \text{O} \end{array} \end{array} \text{2} \text{R}(\text{OCH}\_{2}\text{CH}\_{2}\text{)}\_{n+1}\text{OSO}\_{3}\text{H}^{+} $$

Fig. 4. Reaction scheme of ethoxylated alcohol sulfation (adapted from de Groot, 1991)

LES (C12-C14/15 2-3 ethylene oxide) can be considered as the most efficient anionic surfactant in terms of: superior detergency power, good tolerance for water hardness, and mildness on hands and fibers. The application therefore is wide: from household to personal care and cosmetic product. Unfortunately, sulfated alcohol ether sulfates show a limited stability to hydrolysis at high temperatures, and this restricts their use in heavy duty laundry powders, where high temperatures occur in the spray drying process of powder manufacture.

The high stability to calcium ions permits formulation of liquid detergents with limited or no addition of water "softeners" even in case of use in hard water (Matthijs et al., 1999). The optimum compromise of ethylene oxide addition to keep adequate foam levels and solubility/mildness ratio vary from 2 to 3 moles per mole of fatty alcohol. The most important worldwide application of AES 2-3 ethylene oxide (EO) are in dish washing liquid detergent, generally combined with LABS and in shampoos/bubble baths (Table 2).


Table 2. Applications worldwide of AES 2-3EO combined with LABS and in shampoos/bubble baths (de Groot, 1991)

#### **2.4 Alfa-olefins sulfonates (AOS)**

Alfa-olefins are a potential replacement for alkylbenzenes in detergent applications. Olefin sulfonation is highly exothermic with ∆H = -210 kJ/mol (Roberts, 2001). The neutralized product of alfa-olefin sulfonation requires hydrolysis to remove the sultones, which are skin sensitizers (Figure 5). Their exploitation, however, is largely limited to the Far East, Centre on Japan, at present. Commercial supplies of alfa-olefins are produced by the

Fig. 4. Reaction scheme of ethoxylated alcohol sulfation (adapted from de Groot, 1991)

where high temperatures occur in the spray drying process of powder manufacture.

detergent, generally combined with LABS and in shampoos/bubble baths (Table 2).

LES (C12-C14/15 2-3 EO) 5 – 10 % 10 – 30 % LABS (low molecular weight) 15- 20% Coconut Ethanol Amides (CEA) 2 – 3 % 2 – 3 %

Other actives (i.e. amphoteric/nonionic) 5 – 10 % H2O, perfume, color, preservatives Balance Balance Table 2. Applications worldwide of AES 2-3EO combined with LABS and in

Alfa-olefins are a potential replacement for alkylbenzenes in detergent applications. Olefin sulfonation is highly exothermic with ∆H = -210 kJ/mol (Roberts, 2001). The neutralized product of alfa-olefin sulfonation requires hydrolysis to remove the sultones, which are skin sensitizers (Figure 5). Their exploitation, however, is largely limited to the Far East, Centre on Japan, at present. Commercial supplies of alfa-olefins are produced by the

Hydrotopes (Sodium Tolunene Sulfonate – Sodium Xylene Sulfonate) alcohol solvent

shampoos/bubble baths (de Groot, 1991)

**2.4 Alfa-olefins sulfonates (AOS)** 

LES (C12-C14/15 2-3 ethylene oxide) can be considered as the most efficient anionic surfactant in terms of: superior detergency power, good tolerance for water hardness, and mildness on hands and fibers. The application therefore is wide: from household to personal care and cosmetic product. Unfortunately, sulfated alcohol ether sulfates show a limited stability to hydrolysis at high temperatures, and this restricts their use in heavy duty laundry powders,

The high stability to calcium ions permits formulation of liquid detergents with limited or no addition of water "softeners" even in case of use in hard water (Matthijs et al., 1999). The optimum compromise of ethylene oxide addition to keep adequate foam levels and solubility/mildness ratio vary from 2 to 3 moles per mole of fatty alcohol. The most important worldwide application of AES 2-3 ethylene oxide (EO) are in dish washing liquid

Liquids dish wash

Shampoos/bubble

baths

1 – 3 %

detergent

Fig. 5. Reactions of alfa-olefin sulfonation (adapted from de Groot, 1991)

oligomerisation of ethylene. The physical, detergency and biodegradation characteristics of alfa-olefins are affected by the carbon chain length distribution and therefore each new supply may require testing to determine whether the desired properties for the new chosen application can be achieved. The Lion Corporation, Japan, is one of the principal producers and users of alfa-olefin sulfonates. In addition to fabric washing powders, they also market fabric washing liquid, shampoos, tooth paste and foam bath products containing this active. In the USA, Minnetonka has utilized AOS in hand cleaners/liquid soaps. AOS is a potential replacement for alkyl benzene sulfonates in dish wash detergent liquids formulations with performance peaking at C16 chain length (de Groot., 1991).

#### **2.5 Fatty acid methyl esters sulfonates (FAMES)**

FAMES are called to be the main feedstock for detergent formulating in the future due to their applicability in detergent formulations (Ingegar & Martin, 2001; Johansson & Svensson, 2001; Roberts & Garrett, 2000; Satsuki, 1998). Moreover, when it is derived from palm oil presents special biodegradable properties that place them over the surfactants derived from petrochemicals compounds. To date, the application of FAMES is under development in various detergent products, and their presence on the market is still highly restricted. The typical cut of FAMES (C16-C18) shows interesting surface activity (about 90% compared to LABS), high detergent, dispersing and emulsifying power in hard water, high lime soap dispersion and moderate foam levels. FAMES show high stability to pH and temperature hydrolysis. Therefore, they can be easily spray dyer and/or incorporated in detergent bars. Methyl ester sulfonates have a wide range of application and important biological properties. As aggregated value the FAMES can be used in cosmetics, as auxiliary agents in the production of fibers, plastics, and rubber, and in leather manufacture (Cohen et al., 2008; de Groot, 1991; Roberts et al., 2008; Stein & Baumann, 1975).

#### **3. Sulfonation process used for the manufacturing of anionic surfactants**

Sulfonation reactions can be carried out in different configurations, either liquid-liquid contact, or gas-liquid contact reactors, and a diversity of sulfonating reagents can be applied for the sulfonation process, such as: Sulfuric acid, SO3 from stabilized liquid SO3, SO3 from sulfur burning and subsequent conversion of the SO2 formed, SO3 from boiling concentrated oleum and chlorosulfonic acid. However some reasons why SO3/air in gas-liquid contactor (sulfonator) is becoming the predominant process for the manufacture of anionic surfactants are (Foster, 1997):

Sulfonation/Sulfation Processing Technology for Anionic Surfactant Manufacture 275

Figure 7 shows a diagram of film SO3-sulfonation along with an additional step (bleaching) than could be required depending of the feedstock and characteristic of the final product. Depending on the type of organic feedstock and consequent organic acid, further reaction steps may be required before neutralization. Sulfonic acids of LABS are one of those materials that no require an aging step to reach full conversion. Moreover, a hydrolysis or stabilization step is required to convert anhydrides form during the sulfonation process. Alcohol and alcohol ethoxylate sulfonic acids, as well as FAMES, must be neutralized immediately after a delayed aging to avoid undesired by-products

After aging and hydrolysis a stable product is obtained, then the neutralization stage can be carried out with many alkaline chemicals like caustic, ammonia and sodium carbonate. Neutralization with diluted caustic is recognized as instantaneous and highly exothermic it may form gel at high temperatures or undesired reactions may occur if micro-dispersion of organic acid in the diluted caustic phase fails. Various loop-type reactors, consisting of a circulation pump, homogenizer (where the acid is introduced in the circulating alkaline paste), and heat exchanger, are used for the complex neutralization step (Foster, 1997).

Organic liquid flow through of the reactor wall in laminar regimen, the high flow of the gas phase by gravitational effects intensifies the formation of random waves all along the gasliquid interface. Depending on the flow rate of organic liquid and gas stream the thickness of the film can increase or decrease up to twice its average value in the zone where the waves are present (Díaz, 2009). This induced turbulence affects the local values of concentration and temperature in the regions where appears, hence altering the mass transfer and temperature profiles in the film. Mathematical models which describe the sulfonation of tridecylbenzene in FFRs have been developed by Akanksha et al. (2007), Davis et al., (1979), and Johnson & Crynes, (1974), while Dabir et al. (1996), Gutiérrez et al. (1988) and Talens (1999) focused on dodecylbenzene sulfonation. Nevertheless, these models have been subject of debate due to the assumption that either the chemical reaction is limited to the gas liquid interface, the mass transfer of the sulfonating reagent in the gas phase is the rate determining step, and/or the flow profiles in the film are neatly laminar,

Recently Torres et al. (2009b) proposed a model for the methyl esters sulfonation that is appropriate for both laminar and turbulent films and it considers effects of wavy film

formed in side reactions.

Fig. 7. Process diagram for film SO3-sulfonation

**4. Phenomenological description of film sulfonation** 

neglecting the effects of the waves formed at the gas-liquid interface.


Several studies have been done about absorption along with exothermic reaction in a Falling Film Reactor - FFR (Mann & Moyes, 1977; Villadsen & Nielsen, 1986), particularly for dodecylbenzene and tridecylbenzene sulfonation. However, due to de complexity of processes taking place inside the FFR has not been completely elucidated, being of special interest today. The SO3-sulfonation is carried out in tubular reactors where the organic matter (liquid) wets the wall of the tubes while a gas stream containing the sulfonating reagent flows in co-current with the organic matter to avoid over-sulfonation (MacArthur et al., 1999). The simplest FFR configuration can be described as a two concentric tubes arranged in a vertical way (Figure 6).

Fig. 6. Sketch of falling film reactor

Organic matter forms a thin film covering the inner wall of the inner tube. The film descends from the top of the reactor in laminar flow forming an annulus for whose interior a gas stream flows in turbulent regimen. In the first reaction section the concentrated sulfonate reagent get in touch with fresh organic matter. The reaction rate is high as well as the amount of heat released (150 – 170 kJ/mole). A coolant stream flows by the external wall of the inner tube in parallel with the reactant streams. As long as the reaction advances the viscosity of the liquid phase increases (ca. 100 times the initial value). The depletion of reactants reduces the reaction rate and the increase of viscosity slow down the mass transfer process in the film. In this point the co-current coolant, this has already removed a huge amount of heat from the first reactor zone, works as a heating current that controls the viscosity of the film.

Figure 7 shows a diagram of film SO3-sulfonation along with an additional step (bleaching) than could be required depending of the feedstock and characteristic of the final product. Depending on the type of organic feedstock and consequent organic acid, further reaction steps may be required before neutralization. Sulfonic acids of LABS are one of those materials that no require an aging step to reach full conversion. Moreover, a hydrolysis or stabilization step is required to convert anhydrides form during the sulfonation process. Alcohol and alcohol ethoxylate sulfonic acids, as well as FAMES, must be neutralized immediately after a delayed aging to avoid undesired by-products formed in side reactions.

Fig. 7. Process diagram for film SO3-sulfonation

274 Advances in Chemical Engineering

i. Adaptability: All types of organic feedstocks, like alkylbenzenes, primary alcohols, alcohol ethers, alfa-olefins and fatty acid methyl esters, can be successfully transformed to high-quality sulfonate/sulfate active detergents using SO3/air as sulfonating reagent. Sulfonating reagents like sulfuric acid and oleum are less desirable because only alkylbenzene feedstocks can be converted to high-quality alkylbenzene sulfonic acids. ii. Security: Concentrated sulfuric acid, liquid SO3, and oleum (20 or 65%) are hazardous to be handled, transported, and storage. Sulfur, either in liquid or solid form, although

less dangerous option as initial material for the manufacture of SO3, is still risky. iii. Price: SO3 obtained directly from the sulfur combustion is the most economical option among all the others options mentioned above regarding transport, handle and storage. iv. Availability: Liquid SO3, 65% and 20% oleum and even sulfuric acid are not produced everywhere. Even close to sulfuric acid plants, it is not guaranteed the availability of all

Several studies have been done about absorption along with exothermic reaction in a Falling Film Reactor - FFR (Mann & Moyes, 1977; Villadsen & Nielsen, 1986), particularly for dodecylbenzene and tridecylbenzene sulfonation. However, due to de complexity of processes taking place inside the FFR has not been completely elucidated, being of special interest today. The SO3-sulfonation is carried out in tubular reactors where the organic matter (liquid) wets the wall of the tubes while a gas stream containing the sulfonating reagent flows in co-current with the organic matter to avoid over-sulfonation (MacArthur et al., 1999). The simplest FFR configuration can be described as a two concentric tubes

Organic matter forms a thin film covering the inner wall of the inner tube. The film descends from the top of the reactor in laminar flow forming an annulus for whose interior a gas stream flows in turbulent regimen. In the first reaction section the concentrated sulfonate reagent get in touch with fresh organic matter. The reaction rate is high as well as the amount of heat released (150 – 170 kJ/mole). A coolant stream flows by the external wall of the inner tube in parallel with the reactant streams. As long as the reaction advances the viscosity of the liquid phase increases (ca. 100 times the initial value). The depletion of reactants reduces the reaction rate and the increase of viscosity slow down the mass transfer process in the film. In this point the co-current coolant, this has already removed a huge amount of heat from the first reactor

zone, works as a heating current that controls the viscosity of the film.

the gamma of oleum concentrations.

arranged in a vertical way (Figure 6).

Fig. 6. Sketch of falling film reactor

After aging and hydrolysis a stable product is obtained, then the neutralization stage can be carried out with many alkaline chemicals like caustic, ammonia and sodium carbonate. Neutralization with diluted caustic is recognized as instantaneous and highly exothermic it may form gel at high temperatures or undesired reactions may occur if micro-dispersion of organic acid in the diluted caustic phase fails. Various loop-type reactors, consisting of a circulation pump, homogenizer (where the acid is introduced in the circulating alkaline paste), and heat exchanger, are used for the complex neutralization step (Foster, 1997).

#### **4. Phenomenological description of film sulfonation**

Organic liquid flow through of the reactor wall in laminar regimen, the high flow of the gas phase by gravitational effects intensifies the formation of random waves all along the gasliquid interface. Depending on the flow rate of organic liquid and gas stream the thickness of the film can increase or decrease up to twice its average value in the zone where the waves are present (Díaz, 2009). This induced turbulence affects the local values of concentration and temperature in the regions where appears, hence altering the mass transfer and temperature profiles in the film. Mathematical models which describe the sulfonation of tridecylbenzene in FFRs have been developed by Akanksha et al. (2007), Davis et al., (1979), and Johnson & Crynes, (1974), while Dabir et al. (1996), Gutiérrez et al. (1988) and Talens (1999) focused on dodecylbenzene sulfonation. Nevertheless, these models have been subject of debate due to the assumption that either the chemical reaction is limited to the gas liquid interface, the mass transfer of the sulfonating reagent in the gas phase is the rate determining step, and/or the flow profiles in the film are neatly laminar, neglecting the effects of the waves formed at the gas-liquid interface.

Recently Torres et al. (2009b) proposed a model for the methyl esters sulfonation that is appropriate for both laminar and turbulent films and it considers effects of wavy film

Sulfonation/Sulfation Processing Technology for Anionic Surfactant Manufacture 277

As discussed by Knaggs, (2004), even if the liquid film is turbulent and does wavy flow then turbulent diffusivity cannot be neglected, this and turbulent viscosity in the liquid phase can

3

1 exp( ( / ) / ) 1 exp( ( / ) / )

 

0,5 1 10

log *<sup>i</sup>*

0,5 0,5

(5)

(6)

*G Gi N k C mC SO G SO SO* (7)

(9)

(McCready & Hanratty, 1984) (8)

 

*δ ≤ y ≤*

 

2

*<sup>d</sup>* (1)

(3)

0 *≤ y ≤ δ* (2)

0,5 <sup>2</sup> 0,5

(4)

3 3 <sup>3</sup> (D ) *SO SO*

<sup>2</sup> (/ ) 0,5 0,5 1 0,64( ) 1 exp *<sup>T</sup> <sup>w</sup>*

1 *<sup>w</sup> <sup>L</sup>*

 

*<sup>T</sup> <sup>w</sup> <sup>T</sup> T w <sup>v</sup> <sup>y</sup> <sup>A</sup> Sc D y B* 

5

1

solubility. The Henry constant *m*, is determined from the SO3 vapor pressure:

*u*

0,704 0,8 *Gk Sc*

*u*

where the turbulent velocity is defined as:

**4.2 Momentum balance** 

*i i B Sc C Sc* 

with *A+* = 25,1; *C1* = 34,96; *C2* = 28,97; *C3 =* 13,95; *C4* = 6,33 and *C5* = –1,186. For non–volatile liquids such as methyl stearate, the vapor pressure is zero at working temperatures. At the interface, it is assumed that Henry and Raoult's laws are applicable to determine the SO3

3 33

0,5 *G G*

Axial liquid velocity *vz*, can be found from the momentum equation after neglecting the pressure gradient and axial terms (Figure 9). The flow profile of the liquid falling is

*<sup>D</sup> <sup>y</sup> y f <sup>A</sup>* 

*G L*

Turbulent Schmidt number is evaluated from the Cebeci's modification of the van Driest

*y*

 

(D ) *A A <sup>z</sup> A T C C v Dr zy y* 

*z SO T C C v Dr zy y* 

be taken of work suggested by Yih & Liu (1983).

model and is further modified as:

*T w*

flow by using eddy diffusivity parameter. The eddy diffusivity models proposed by Lamourelle & Sandall (1972) for the outer region and modified by van Driest (1956) for the region near the wall were used. Effects of interfacial drag at the gas–liquid interface and the gas–phase heat and mass transfer resistance have also been considered in the proposed model. The model takes into account the variations of physical properties with temperature and predicts conversion profiles, gas–liquid interface temperature in the axial direction, and average liquid film thickness along the reactor length. Knowledge of temperature distribution along with the reactor is important for the product quality control, since for highly exothermic reactions under certain conditions can produce degradation of the products. The equations described in the following section account for the mass, momentum and heat transfer. In the development of these equations was considered the turbulent diffusivity for mass transfer coupled with chemical reaction, according to the theory of Yih & Seagrave (1978), and with heat transfer according with Yih & Liu (1983).

Finally some additional assumptions were made for the mathematical model:


According with these assumptions the mathematical model is showed in the following sections.

#### **4.1 Mass balance**

Only three components are considered in the liquid phase: organic liquid, acid product and sulfonating reagent, therefore two microscopic balances are sufficient to determine the concentration profiles (Figure 8), where *y* varies from *y* = 0 (at the wall surface) to *y* = *δ* (at the liquid surface).

Fig. 8. Mass balance on finite volume includes the boundary conditions at the solid wall and liquid/gas interface

It is assumed that the mass balance for SO3(G) absorbing by the liquid (equation 1) can also be applied to reagent in the liquid phase where reaction occurs, then equation 2 is the steady state mass balance on the absorbing species *A* in liquid phase.

flow by using eddy diffusivity parameter. The eddy diffusivity models proposed by Lamourelle & Sandall (1972) for the outer region and modified by van Driest (1956) for the region near the wall were used. Effects of interfacial drag at the gas–liquid interface and the gas–phase heat and mass transfer resistance have also been considered in the proposed model. The model takes into account the variations of physical properties with temperature and predicts conversion profiles, gas–liquid interface temperature in the axial direction, and average liquid film thickness along the reactor length. Knowledge of temperature distribution along with the reactor is important for the product quality control, since for highly exothermic reactions under certain conditions can produce degradation of the products. The equations described in the following section account for the mass, momentum and heat transfer. In the development of these equations was considered the turbulent diffusivity for mass transfer coupled with chemical reaction, according to the theory of Yih & Seagrave (1978), and with heat transfer according with

Finally some additional assumptions were made for the mathematical model:

ii. Fully developed film (entrance and exit effects to reactor are neglected);

iii. The liquid film is symmetric with respect to the reactor axis.

i. No entrainments of liquid droplets into gas or of gas bubbles into the liquid film occur;

According with these assumptions the mathematical model is showed in the following

Only three components are considered in the liquid phase: organic liquid, acid product and sulfonating reagent, therefore two microscopic balances are sufficient to determine the concentration profiles (Figure 8), where *y* varies from *y* = 0 (at the wall surface) to *y* = *δ* (at

Fig. 8. Mass balance on finite volume includes the boundary conditions at the solid wall and

It is assumed that the mass balance for SO3(G) absorbing by the liquid (equation 1) can also be applied to reagent in the liquid phase where reaction occurs, then equation 2 is the

steady state mass balance on the absorbing species *A* in liquid phase.

Yih & Liu (1983).

**4.1 Mass balance** 

the liquid surface).

liquid/gas interface

sections.

$$\left[v\_z \frac{\partial \mathcal{C}\_{SO\_3}}{\partial z} = \frac{\partial}{\partial y}\right] \left(\mathcal{D}\_{SO\_3} + D\_T\right) \frac{\partial \mathcal{C}\_{SO\_3}}{\partial y}\left[-r \quad \delta \le y \le \frac{d}{2}\right] \tag{1}$$

$$
\sigma\_z \frac{\partial \mathbf{C}\_A}{\partial z} = \frac{\partial}{\partial y} \left[ (\mathbf{D}\_A + \mathbf{D}\_T) \frac{\partial \mathbf{C}\_A}{\partial y} \right] - r \quad \mathbf{0} \le y \le \delta \tag{2}
$$

As discussed by Knaggs, (2004), even if the liquid film is turbulent and does wavy flow then turbulent diffusivity cannot be neglected, this and turbulent viscosity in the liquid phase can be taken of work suggested by Yih & Liu (1983).

$$\frac{D\_T}{\nu\_T} = -0, 5 + 0, 5 \left[ 1 + 0, 64 \left( y^{+2} \right) \frac{\tau}{\tau\_w} \times \left[ 1 - \exp \left( \frac{-y^+ \left( \tau \left/ \tau\_w \right)^{0.5}}{A^+} \right) \right)^2 f \right]^{0.5} \tag{3}$$

$$\frac{\tau\_w}{\tau} = 1 - \left(\frac{\tau\_L}{\tau\_G + \tau\_L}\right)^3 \left(\frac{\mathcal{Y}^+}{\mathcal{S}^+}\right) \tag{4}$$

Turbulent Schmidt number is evaluated from the Cebeci's modification of the van Driest model and is further modified as:

$$Sc\_T = \frac{\upsilon\_T}{D\_T} = \frac{1 - \exp(-y^+ \left(\tau \mid \tau\_w\right)^{0.5} \nearrow A^+)}{1 - \exp(-y^+ \left(\tau \mid \tau\_w\right)^{0.5} \nearrow B^+)}\tag{5}$$

$$\mathcal{B}^{+} = \text{Sc}^{-0.5} \sum\_{i=1}^{5} \mathbb{C}\_{i} \left( \log\_{10} \text{Sc} \right)^{i-1} \tag{6}$$

with *A+* = 25,1; *C1* = 34,96; *C2* = 28,97; *C3 =* 13,95; *C4* = 6,33 and *C5* = –1,186. For non–volatile liquids such as methyl stearate, the vapor pressure is zero at working temperatures. At the interface, it is assumed that Henry and Raoult's laws are applicable to determine the SO3 solubility. The Henry constant *m*, is determined from the SO3 vapor pressure:

$$N\_{SO\_3}^G = k\_G \left( \mathbf{C}\_{SO\_3}^G - m \mathbf{C}\_{SO\_3}^i \right) \tag{7}$$

$$\frac{k\_{\odot}}{\mu} = 0,8Sc^{-0.704} \qquad \text{(McCready \& Handrats, 1984)}\tag{8}$$

where the turbulent velocity is defined as:

$$
\mu = \left(\frac{\tau\_G}{\rho\_G}\right)^{0.5} \tag{9}
$$

#### **4.2 Momentum balance**

Axial liquid velocity *vz*, can be found from the momentum equation after neglecting the pressure gradient and axial terms (Figure 9). The flow profile of the liquid falling is

Sulfonation/Sulfation Processing Technology for Anionic Surfactant Manufacture 279

For exothermic reactions such as sulfonation a large amount of heat may be released, the boundary conditions showed at Figure 10 is applying for the energy balance (equation 17).

> *z L cT <sup>T</sup> v kH <sup>r</sup> zyy*

(17)

(18)

(19)

11 1

*<sup>U</sup> d d k h d d*

Heat transport equations follow the Prandtl analogy and are equivalent to those used for

0,704 0,8 *Gh Sc*

An experimental set was developed to study the effect of follows factors: (i) mole ratio between SO3 and organic liquid, (ii) wall temperature and, (iii) volumetric percentage of SO3 in the phase gaseous. The variables representing the quality of the sulfonated product are: active matter, unsulfonated matter, acid value and color (Ahmad et al., 2007; Inagaki, 2001). The experimental matrix is presented in Table 3 and detailed information about the analysis

*u*

Conditions value SO3/N2 inlet (gaseous sulfonate mixture), % vol/vol 3 – 7 SO3/N2 temperature inlet, °C 50 – 60 SO3 /organic liquid, mole ratio 1:1 – 1,2:1

Table 3. Operating conditions used for the methyl ester sulfonation (Torres et al., 2009a)

**5. Main parameters of film SO3-sulfonation** 

is presented below (Torres et al., 2008b).

*lm ex w w ex in*

Fig. 10. Schematic representation of model for a segment heat balance

**4.3 Heat balance** 

mass transfer.

Fig. 9. Schematic representation of velocity profiles in laminar and turbulent regimes for both liquid and gas phase

predominantly laminar, while SO3 flow is clearly turbulent and consequently the SO3 is absorbed at the gas/liquid interface. In equation 10 for co-currents systems *J* is +1 and for counter current systems *J* is -1.

$$w\_z = \frac{\rho\_L g}{\mu\_L} \left[ y \delta - \frac{y^2}{2} \right] + J \frac{\tau\_G y}{\mu\_L} \tag{10}$$

For high gas flow where the shear force predominates over the gravitational force, the linear velocity distribution is:

$$w\_z = \frac{\tau\_G y}{\mu\_L} \tag{11}$$

Calculation of *τG* based on the relations proposed by Riazi & Faghri (1986) shows that when the gas flow is turbulent, the effects of the interfacial drag cannot be neglected; *τG* can also be verified by means of experimental pressure drop data. In this way a set of parameters can be introduced and adjusted to minimize the deviation from a data set for sulfonation:

$$
\Gamma = \frac{\rho\_{\rm L} g}{3\mu\_{\rm L}} \delta^3 - \frac{\tau\_{\rm G}}{2\mu\_{\rm L}} \delta^2 \tag{12}
$$

$$
\pi\_{\mathcal{G}} = \mathcal{C}\_f \rho\_{\mathcal{G}} \mu^2 \tag{13}
$$

$$\frac{1}{\text{C}\_f^2} = -4\text{Log}\left|\frac{\rho\delta}{3,7d} - \frac{5.02}{\text{Re}\_G}\text{Log}\left(\frac{\rho\delta}{3,7d} + \frac{13}{\text{Re}\_G}\right)\right|\quad\text{ (Tales, 1999)}\tag{14}$$

$$\ln(q) = 3,59 - 5,14 \text{ v}\_{\text{iL}} \text{ (if } v\_{\text{iL}} \le 0,175 \text{ ms} \text{-1)} \text{ or } \ln(q) = 20,55 \text{ v}\_{\text{iL}} - 0,93 \text{ (if } v\_{\text{iL}} \ge 0,175 \text{ ms} \text{-1)} \tag{15}$$

The initial value of the film thickness *δ* can be obtained through equation 16, after this value is calculated by iteration:

$$\mathcal{S} = \left(\frac{3\Gamma\mu}{\mathcal{g}\mathcal{P}\_{\mathcal{L}}}\right)^{\frac{1}{3}}\tag{16}$$

#### **4.3 Heat balance**

278 Advances in Chemical Engineering

Fig. 9. Schematic representation of velocity profiles in laminar and turbulent regimes for

predominantly laminar, while SO3 flow is clearly turbulent and consequently the SO3 is absorbed at the gas/liquid interface. In equation 10 for co-currents systems *J* is +1 and for

 

For high gas flow where the shear force predominates over the gravitational force, the linear

*<sup>G</sup> <sup>z</sup> L <sup>y</sup> <sup>v</sup>* 

Calculation of *τG* based on the relations proposed by Riazi & Faghri (1986) shows that when the gas flow is turbulent, the effects of the interfacial drag cannot be neglected; *τG* can also be verified by means of experimental pressure drop data. In this way a set of parameters can be

> 3 2 *L G L L*

 

 

*ln(φ) = 3,59 – 5,14 viL* (if *viL < 0,175* ms–1) or *ln(φ) = 20,55 viL – 0,93* (if *viL ≥ 0,175* ms–1) (15) The initial value of the film thickness *δ* can be obtained through equation 16, after this value

 

1 3 <sup>3</sup> *<sup>L</sup> g* 

introduced and adjusted to minimize the deviation from a data set for sulfonation:

*g*

1 5.02 13 <sup>4</sup> 3,7 Re 3,7 Re *<sup>f</sup> G G Log Log <sup>C</sup> d d*

  3 2

 

2

 

2 2 *L G <sup>z</sup> L L gy y vy J*

(10)

(11)

(12)

*G fG C u* (13)

(Talens, 1999) (14)

(16)

both liquid and gas phase

velocity distribution is:

counter current systems *J* is -1.

2

is calculated by iteration:

For exothermic reactions such as sulfonation a large amount of heat may be released, the boundary conditions showed at Figure 10 is applying for the energy balance (equation 17).

Fig. 10. Schematic representation of model for a segment heat balance

$$
v\_z \frac{\partial \rho cT}{\partial z} = -\frac{\partial}{\partial y} \left[ -k\_L \frac{\partial T}{\partial y} \right] + (\Delta H)r \tag{17}
$$

$$\frac{1}{M} = \frac{1}{k\_w \frac{d\_{lm}}{d\_{cc}}} + \frac{1}{h\_w \frac{d\_{cc}}{d\_{in}}} \tag{18}$$

Heat transport equations follow the Prandtl analogy and are equivalent to those used for mass transfer.

$$\frac{h\_G}{\mu} = 0.8 Sc^{-0.704} \tag{19}$$

#### **5. Main parameters of film SO3-sulfonation**

An experimental set was developed to study the effect of follows factors: (i) mole ratio between SO3 and organic liquid, (ii) wall temperature and, (iii) volumetric percentage of SO3 in the phase gaseous. The variables representing the quality of the sulfonated product are: active matter, unsulfonated matter, acid value and color (Ahmad et al., 2007; Inagaki, 2001). The experimental matrix is presented in Table 3 and detailed information about the analysis is presented below (Torres et al., 2008b).


Table 3. Operating conditions used for the methyl ester sulfonation (Torres et al., 2009a)

Sulfonation/Sulfation Processing Technology for Anionic Surfactant Manufacture 281

Figure 12 proves that the unsulfonated matter percentage decreases the same as the

Figure 13 shows the effect of the conditions process on acid value. Increase in the sulfur trioxide has a positive effect in agreement with expectations; also the increase of temperature in process enlarges the acidity in the product. Change in the slope can be explained by kinetics effects favored by temperature rising. Both the increase in the mole ratio and sulfur trioxide in the sulfonating mixture can be explained by the effectiveness of the reaction because an excess of SO3 promotes the consumption of the same reactants for

Fig. 13. Acid values obtain by changes of mole ratio, temperature and SO3 inlet

Fig. 14. Impact on coloration in the sulfonated product due to variations of process

Figure 14 presents the trends for the coloration in the sulfonated product which intensifies the values of concentration of SO3 some that the mole ratio and SO3 content in the gas stream. All variables show a direct influence to the increase in color which is associated with

sulfonating reagent in the gas mixture increases.

the generation of over-sulfonated matter.

**5.3 Acid value** 

**5.4 Coloration** 

#### **5.1 Active matter**

Increase of active matter in product was proportional to the increase of SO3/organic liquid mole ratio as well as the increase of SO3 percentage in the gas stream. Slope changes observed with respect to the temperature are probably due to side reactions occurring at higher temperatures; the formations of undesired matters decrease active matter (Figure 11).

Fig. 11. Impact of the operation conditions on the degree of sulfonation

#### **5.2 Unsulfonated matter**

The impact of the experimental factors is initially inverse compared with the effect obtained with the active matter; however the SO3/organic liquid mole ratios beyond 1,1 produce an increase in the quantified unsulfonated matter. This change can be explained by over sulfonation of reactant and formation of side products. This assumption is consistent with the effects of temperature on the percentage of non-sulfonated matter (Figure 12).

Fig. 12. Effects of the experimental factors on the unsulfonated matter

Over-sulfonated products do not have the same characteristics of the washing active substance and therefore are not identified as active matter but yet as free oil (unsulfonated). Figure 12 proves that the unsulfonated matter percentage decreases the same as the sulfonating reagent in the gas mixture increases.

#### **5.3 Acid value**

280 Advances in Chemical Engineering

Increase of active matter in product was proportional to the increase of SO3/organic liquid mole ratio as well as the increase of SO3 percentage in the gas stream. Slope changes observed with respect to the temperature are probably due to side reactions occurring at higher temperatures; the formations of undesired matters decrease active matter (Figure 11).

The impact of the experimental factors is initially inverse compared with the effect obtained with the active matter; however the SO3/organic liquid mole ratios beyond 1,1 produce an increase in the quantified unsulfonated matter. This change can be explained by over sulfonation of reactant and formation of side products. This assumption is consistent with

Over-sulfonated products do not have the same characteristics of the washing active substance and therefore are not identified as active matter but yet as free oil (unsulfonated).

the effects of temperature on the percentage of non-sulfonated matter (Figure 12).

Fig. 11. Impact of the operation conditions on the degree of sulfonation

Fig. 12. Effects of the experimental factors on the unsulfonated matter

**5.1 Active matter** 

**5.2 Unsulfonated matter** 

Figure 13 shows the effect of the conditions process on acid value. Increase in the sulfur trioxide has a positive effect in agreement with expectations; also the increase of temperature in process enlarges the acidity in the product. Change in the slope can be explained by kinetics effects favored by temperature rising. Both the increase in the mole ratio and sulfur trioxide in the sulfonating mixture can be explained by the effectiveness of the reaction because an excess of SO3 promotes the consumption of the same reactants for the generation of over-sulfonated matter.

Fig. 13. Acid values obtain by changes of mole ratio, temperature and SO3 inlet

#### **5.4 Coloration**

Figure 14 presents the trends for the coloration in the sulfonated product which intensifies the values of concentration of SO3 some that the mole ratio and SO3 content in the gas stream. All variables show a direct influence to the increase in color which is associated with

Fig. 14. Impact on coloration in the sulfonated product due to variations of process

Sulfonation/Sulfation Processing Technology for Anionic Surfactant Manufacture 283

diffusivities of reagents are estimated from the Wilke-Chang equations and diffusivities in the mixture are estimated through the Vignes equation (Vignes, 1966). Generally, in methyl ester sulfonation the amount of intermediate III in the final product varies from 10-20% (Foster, 2004), but this amount can be reduced by a long and heated digestion (aging stage). An experimental apparatus showed in Figure 16 was utilized by Torres (2009) for researching the parameters of film SO3-sulfonation of methyl ester derived from hydrogenated stearin palm, this apparatus was designed by Chemical Engineering

Fig. 16. Experimental apparatus for methyl ester sulfonation using a falling film reactor and

High degree of sulfonation is obtained in the aging step controlling simultaneously the temperature and residence time. At higher temperatures it is feasible to obtain higher conversion levels, whereas at low temperatures (below 80°C) the time required to reach high conversions is considerably long. These reactions are highly exothermic in order of 150 – 170 kJ/mole (including 25 kJ/mole of the absorption heat of gaseous SO3). A kinetic model has been development by Roberts (2007) based on the proposal that two major intermediates are

SO3 stripped from 65% oleum with dried air process

involved in aging (Figure 17).

Laboratory from Universidad Nacional de Colombia (Bogotá, Colombia).

higher sulfonation degrees. Although the coloration is identified mostly as an esthetic factor for the commercialization of sulfonates, higher colorations can also be a qualitative indicator of over-sulfonation. Anionic surfactants in aqueous solution have colors ranging from yellow to reddish orange (Inagaki, 2001).

#### **6. Film SO3-sulfonation applied**

#### **6.1 Sulfonation of methyl ester with SO3**

Use of methyl esters (ME) in the industry detergent, although under investigation and development since more than 25 years, has not yet expanded to high levels, mainly because the following reason:


Complex chemistry is not yet fully elucidated, but may be summarized as is shown in Figure 15 (Morales & Martínez, 2009).

$$\text{HSO}\_3 + \text{H}\_2\text{SO}\_4 \xrightarrow{\text{H}\_2\text{SO}\_3^+} \text{H}\_2\text{SO}\_4^+ + \text{HSO}\_4^+ \tag{1}$$

$$\underset{\underset{\text{HOS}}{\text{C}}}{\underset{\text{O}}{\text{C}}} \underset{\underset{\text{-}k\_{2}}{\text{C}}}{\underset{\text{-}k\_{2}}{\text{C}}} \underset{\underset{\text{-}k\_{2}}{\text{C}}}{\underset{\text{-}k\_{2}}{\text{C}}} \leq \underset{\text{-}k\_{2}}{\text{C}}$$

$$\text{(iv)}\qquad\text{Q}\qquad\bigotimes\_{\text{SOH}}\text{B}\qquad\underset{\text{\ast}^{\text{R}}}{\text{Q}}\qquad\text{\ast}^{\text{A}}\qquad\text{Q}\qquad\underset{\text{\ast}^{\text{A}}}{\text{Q}}\qquad\text{Q}$$

$$\alpha\_{\rm (y)}^{\rm \rm \rm \&} \sim \sim \sim \sim\_{\rm \rm \&} \sim \sim\_{\rm \rm \&} \sim\_{\rm \rm \&}$$

Fig. 15. Mechanism of methyl ester sulfonation in sulfonator (Torres et al., 2009b)

The methyl ester molecule is initially di-sulfonated in a relatively fast reaction accompanied with a high amount of heat released (Roberts, 2003). There is a third reaction stage considerably slower than the previous ones, where an SO3 group is liberated (on aging). For some researchers, this SO3 group just released would be especially active and therefore capable of directly sulfonating another methyl ester molecule in an alfa position. The

higher sulfonation degrees. Although the coloration is identified mostly as an esthetic factor for the commercialization of sulfonates, higher colorations can also be a qualitative indicator of over-sulfonation. Anionic surfactants in aqueous solution have colors ranging from

Use of methyl esters (ME) in the industry detergent, although under investigation and development since more than 25 years, has not yet expanded to high levels, mainly because

i. Controversial forecasts about availability of petrochemical feedstocks with related cost

iii. The process to produce high quality α-sulfonated methyl ester (SME) is generally more

iv. Application know-how is not yet completely availed and low FAMES solubility involves some restrictions in application, notably concerning the use in liquid detergent

Complex chemistry is not yet fully elucidated, but may be summarized as is shown in

Fig. 15. Mechanism of methyl ester sulfonation in sulfonator (Torres et al., 2009b)

The methyl ester molecule is initially di-sulfonated in a relatively fast reaction accompanied with a high amount of heat released (Roberts, 2003). There is a third reaction stage considerably slower than the previous ones, where an SO3 group is liberated (on aging). For some researchers, this SO3 group just released would be especially active and therefore capable of directly sulfonating another methyl ester molecule in an alfa position. The

yellow to reddish orange (Inagaki, 2001).

**6.1 Sulfonation of methyl ester with SO3**

comparison via vs. natural sources;

Figure 15 (Morales & Martínez, 2009).

ii. Viability of sufficient quality of sulfonation grade methyl esters;

complex than that for alkyl benzene sulfonates;

products and low temperature washing cycles.

**6. Film SO3-sulfonation applied** 

the following reason:

diffusivities of reagents are estimated from the Wilke-Chang equations and diffusivities in the mixture are estimated through the Vignes equation (Vignes, 1966). Generally, in methyl ester sulfonation the amount of intermediate III in the final product varies from 10-20% (Foster, 2004), but this amount can be reduced by a long and heated digestion (aging stage).

An experimental apparatus showed in Figure 16 was utilized by Torres (2009) for researching the parameters of film SO3-sulfonation of methyl ester derived from hydrogenated stearin palm, this apparatus was designed by Chemical Engineering Laboratory from Universidad Nacional de Colombia (Bogotá, Colombia).

Fig. 16. Experimental apparatus for methyl ester sulfonation using a falling film reactor and SO3 stripped from 65% oleum with dried air process

High degree of sulfonation is obtained in the aging step controlling simultaneously the temperature and residence time. At higher temperatures it is feasible to obtain higher conversion levels, whereas at low temperatures (below 80°C) the time required to reach high conversions is considerably long. These reactions are highly exothermic in order of 150 – 170 kJ/mole (including 25 kJ/mole of the absorption heat of gaseous SO3). A kinetic model has been development by Roberts (2007) based on the proposal that two major intermediates are involved in aging (Figure 17).

Sulfonation/Sulfation Processing Technology for Anionic Surfactant Manufacture 285

Generally, feedstock for the manufacture of FAMES containing unsaturated fatty acids and these has been attributed to the formation of polysulfone in the double bonds (Yamada & Matsutani, 1996). Unsaturated in methyl ester make it an olefin with a carboxyl methyl group at the end of the chain. Olefins are more rapidly sulfonated by SO3 also unsaturated bound produces oversulfonation and oxidation of the olefin which competes with the saturated ester obtain product more colored, however the color can be improved by bleaching. Unsaturated make it an olefin with a carboxyl methyl group at the end of the chain. Olefins are more rapidly sulfonated also unsaturated bound produces oversulfonation and oxidation of the olefin which competes with the saturated ester obtain

product more colored, however the color can be improved by bleaching (Figure 19).

Fig. 19. Reaction scheme for the coloration in aging step (adapted from Roberts et al., 2008)

Mechanism proposed by Roberts (2007) suggests a reversible formation of β-dioxide cycle and CH3SO3H this β-anhydride reacts opening its cycle, sintering itself, and losing a carbon monoxide to become an alkene sulfonic acid. This is formed mainly in reactions of sulfonation of alfa-olefins, these are very intensive in color when aged in the acid form

The input variables more important for the conversion are: the length and diameter reactor, flow of liquid reactant, mole ratio between SO3 and organic liquid, in this case methyl stearate derived of hydrogenated stearin from palm oil (Narváez et al., 2005; Torres

(Clippinger, 1964).

**6.2 Validation of model** 

Fig. 17. Reactions in the aging step (adapted from Roberts, 2001)

$$k = Ac^{\left(-\frac{B}{T}\right)}\tag{20}$$


Table 4. Values for equation 20 on aging stage (Roberts, 2008)

The overall conversion as function of time and mole ratio *M* of SO3/ME is given by:

$$\% \text{ conversion} = 100M \left( \frac{1}{M\_{100}} \right) - 0.25e^{-k\_f t} - 0.167e^{-k\_s t} \tag{21}$$

*M100* is mole ratio for a conversion at 100%, after a delayed aging *M100* = 1,2. These equations, for aging in a batch reactor system or in a plug flow systems, can be used as guidelines when setting initial conditions before fine-tuning plant operation to meet a required specification (Roberts, 1998). Methyl esters are less active than aromatic compounds to sulfonating due to the less electronic density of the aliphatic chains. The methyl ester sulfonation include a neutralization step to obtain monosodium salts of α-sulfo methyl esters as desire products (Kapur et al. 1978). If neutralization is immediate disodium salt is formed (see Figure 18a). However, if neutralization of the acid is delayed, the sulfo ester monodisodium salt is obtained as final product (see Figure 18b).

$$\begin{array}{cccc} \text{R-CH-C}^{-} \text{O-CH-}\_{3} & + & \text{NaOH} \xrightarrow{-} & \text{R}^{-} \text{-CH-} \text{C} \text{-} \text{ONa} & + & \text{CH}\_{3}\text{OSSO}\_{3}\text{Na} \\ \text{^{\text{R}}} \text{SO}\_{3}\text{H} & \text{O-SO}\_{3}^{\cdot} & & \text{SO}\_{3}\text{Na} \\ & & & \text{(Bi-salt)} & \\ \end{array}$$

$$\begin{array}{llll} \text{R}-\text{CH-}\text{C}\text{'}-\text{O}-\text{CH}\_{3} \xrightarrow{\text{H}} \text{R}-\text{CH-}\text{C}\text{-}\text{O}-\text{CH}\_{3} & + \text{SO}\_{3} & \overset{\text{+ NaOH}}{\underset{\text{H}}{\longleftarrow}} \text{R}-\text{CH-}\overset{\text{C}}{\text{C}}-\text{O}-\text{CH}\_{3} & + \text{H}\_{2}\text{C} \\ \text{SO}\_{3}\text{H} & \text{O-SO}\_{3} & \text{SO}\_{3}\text{H} & \overset{\text{H}\_{2}\text{O}\_{2}}{\underset{\text{(a-sulfo methyl ester acid)}}{\longrightarrow}} \text{SO}\_{3}\text{Na} \\ & & & \text{(a-sulfo methyl ester acid)} \\ \text{(monosodium salt)} \end{array}$$

Fig. 18. Neutralization chemistry of SME (Torres et al., 2009b)

*B*

(20)

(21)

*kf*, (s-1) *ks*, (s-1)

*<sup>T</sup> k Ae* 

*LogA* 12,10 11,52 *B* 12,060 12,130

The overall conversion as function of time and mole ratio *M* of SO3/ME is given by:

100 <sup>1</sup> % <sup>100</sup> 0.25 0.167 *<sup>f</sup> <sup>s</sup> k t k t conversion M e e M*

*M100* is mole ratio for a conversion at 100%, after a delayed aging *M100* = 1,2. These equations, for aging in a batch reactor system or in a plug flow systems, can be used as guidelines when setting initial conditions before fine-tuning plant operation to meet a required specification (Roberts, 1998). Methyl esters are less active than aromatic compounds to sulfonating due to the less electronic density of the aliphatic chains. The methyl ester sulfonation include a neutralization step to obtain monosodium salts of α-sulfo methyl esters as desire products (Kapur et al. 1978). If neutralization is immediate disodium salt is formed (see Figure 18a). However, if neutralization of the acid is delayed, the sulfo ester

Fig. 17. Reactions in the aging step (adapted from Roberts, 2001)

Table 4. Values for equation 20 on aging stage (Roberts, 2008)

monodisodium salt is obtained as final product (see Figure 18b).

Fig. 18. Neutralization chemistry of SME (Torres et al., 2009b)

Generally, feedstock for the manufacture of FAMES containing unsaturated fatty acids and these has been attributed to the formation of polysulfone in the double bonds (Yamada & Matsutani, 1996). Unsaturated in methyl ester make it an olefin with a carboxyl methyl group at the end of the chain. Olefins are more rapidly sulfonated by SO3 also unsaturated bound produces oversulfonation and oxidation of the olefin which competes with the saturated ester obtain product more colored, however the color can be improved by bleaching. Unsaturated make it an olefin with a carboxyl methyl group at the end of the chain. Olefins are more rapidly sulfonated also unsaturated bound produces oversulfonation and oxidation of the olefin which competes with the saturated ester obtain product more colored, however the color can be improved by bleaching (Figure 19).

Fig. 19. Reaction scheme for the coloration in aging step (adapted from Roberts et al., 2008)

Mechanism proposed by Roberts (2007) suggests a reversible formation of β-dioxide cycle and CH3SO3H this β-anhydride reacts opening its cycle, sintering itself, and losing a carbon monoxide to become an alkene sulfonic acid. This is formed mainly in reactions of sulfonation of alfa-olefins, these are very intensive in color when aged in the acid form (Clippinger, 1964).

#### **6.2 Validation of model**

The input variables more important for the conversion are: the length and diameter reactor, flow of liquid reactant, mole ratio between SO3 and organic liquid, in this case methyl stearate derived of hydrogenated stearin from palm oil (Narváez et al., 2005; Torres

Sulfonation/Sulfation Processing Technology for Anionic Surfactant Manufacture 287

Battaglini et al., 1986; Schambil & Schwuger, 1990). The progress of the reaction is decisive

Parameters Correlation

*SO ME* <sup>3</sup> *r kC C* (22) 14.350

(23)

2/3 3,12 10 *<sup>L</sup>*

2/3 6,288 10 *<sup>L</sup>*

2/3 2,031 10 1000 *L L*

1949,6 <sup>6</sup> <sup>273</sup> 8 10 *TL e* (29)

*e* (30)

a) 0< *x* ≤ 0,25: (31)

<sup>6</sup> 3,86 2,72 10 *T x <sup>L</sup>*

b) 0,25 < *x* ≤ 0,6: (32) 4850 <sup>9</sup> 1,035 3,7 10 0,003 *T x <sup>L</sup>*

c) 0,6 < *x* < 1: (33) 4850 <sup>9</sup> 5,22 3,7 10 0,003 *T x <sup>L</sup>*

<sup>8</sup> 5,88 1,36 10 *TL*

*e*

*e*

*e*

*SME* 

> *L*

> > *L*

*L* 

5700

2980

*L T* 

(24)

*L T* 

*T* 

(26)

(25)

<sup>19</sup> 1,14 10 *<sup>T</sup> k e*

*DSME-ME =* <sup>11</sup>

*DME-SME =* <sup>11</sup>

*DSO3=* <sup>10</sup>

for sulfonation degree expressed as amount active matter.

Kinetic in sulfonator, kmol/m3s (Torres et al.,

Diffusivity, m2/s (Wilke & Chang, 1955)

Gas thermal conductivity, J/msK (Davis et al.,

Heat capacity of gas mix, J/kmolK (Same at

Liquid mixture density, kg/m3 (Talens &

Heat capacity of liquid, J/kmolK (Broström, 1975;

Liquid thermal conductivity, J/msK (Davis et al.,

Viscosity of methyl esters mix, kg/ms (Torres &

Viscosity of methyl esters sulfonic acid, kg/ms

Viscosity of liquid mix, kg/ms (Broström, 1975;

liquid and gas mixture (Torres & Sánchez, 2009)

(Talens & Gutiérrez, 1995)

Talens & Gutiérrez, 1995)

Sánchez, 2008) *μME =* 

2008a)

Physicals and chemicals properties used in model are listed in Table 5.

1979) *kG* = 0,0279

nitrogen) *cG* = 29,82

1979) *kL* = 0,276 Surface tension, N/m (de Groot, 1991) *σ* = 0,046

Viscosity of gas mix, kg/ms (Same at nitrogen) *μ<sup>G</sup>* = 1,910-5

All thermal conductivity, J/msK (Davis et al., 1979) *kw* = 16,3

Table 5. Correlation to estimate heat and mass transfer coefficients and properties of organic

Reid et al., 1987) *cL =* 507,300 + 101,010*x* (27)

Gutiérrez, 1995; Broström, 1975) *<sup>ρ</sup>L =* 980 + 192*x* – 0,66*TL* (28)

et al., 2005), and amount sulfur trioxide in the gas phase (SO3/N2), finally the temperature of the process. This mathematical model permits to calculate the profiles of interfacial liquid temperature, liquid film density, liquid viscosity for any column height and longitudinal profiles of conversion. The proposed model may be suitable for use in design and operation of industrial film reactors. To ensure convergence of the system of equations then transformation of the equations proposed by Agrawal & Peckover (1980) was chosen following the same development by Talens (1999). The set of equations resulting from the mass, momentum and heat transfer is solved numerically. Figure 20 shows schematic view from the top of a reactor: the liquid is evenly distributed around the wall, and the gas mixture is injected through the center of the column. The interfacial temperature is affected by the SO3 amount in the gas mix. It is clear an increase of interface temperature result of the SO3 excess in the gas flow. The temperature of the reagents is a key control variable to avoid undesirable side impact that damage the product mainly by strong coloration.

Fig. 20. An example of interfacial temperature profiles fall in the reactor longitudinal

Other example of the results provided by the model for longitudinal conversion profile from top of the reactor (expressed as percentage of active matter) is shown in Figure 21(a). The input values of the model are: SO3/N2 percentage at 5%, SO3/methyl stearate mole ratio at 1; *TG*, *TL*, and *Tw* at 343 K, 333 K and 313 K, respectively. This figure shows schematically the fast conversion region at the top of the reactor (associated with gas phase control) and slow conversion region at the bottom (linked with liquid phase control). The washing active substance was determinate using a two titration technique with Hyamine 1622 as the titrant reagent and methylene blue as indicator (Tsubochi et al., 1979; Milwdsky & Gabriel, 1982;

et al., 2005), and amount sulfur trioxide in the gas phase (SO3/N2), finally the temperature of the process. This mathematical model permits to calculate the profiles of interfacial liquid temperature, liquid film density, liquid viscosity for any column height and longitudinal profiles of conversion. The proposed model may be suitable for use in design and operation of industrial film reactors. To ensure convergence of the system of equations then transformation of the equations proposed by Agrawal & Peckover (1980) was chosen following the same development by Talens (1999). The set of equations resulting from the mass, momentum and heat transfer is solved numerically. Figure 20 shows schematic view from the top of a reactor: the liquid is evenly distributed around the wall, and the gas mixture is injected through the center of the column. The interfacial temperature is affected by the SO3 amount in the gas mix. It is clear an increase of interface temperature result of the SO3 excess in the gas flow. The temperature of the reagents is a key control variable to avoid undesirable side impact that damage the

Fig. 20. An example of interfacial temperature profiles fall in the reactor longitudinal

Other example of the results provided by the model for longitudinal conversion profile from top of the reactor (expressed as percentage of active matter) is shown in Figure 21(a). The input values of the model are: SO3/N2 percentage at 5%, SO3/methyl stearate mole ratio at 1; *TG*, *TL*, and *Tw* at 343 K, 333 K and 313 K, respectively. This figure shows schematically the fast conversion region at the top of the reactor (associated with gas phase control) and slow conversion region at the bottom (linked with liquid phase control). The washing active substance was determinate using a two titration technique with Hyamine 1622 as the titrant reagent and methylene blue as indicator (Tsubochi et al., 1979; Milwdsky & Gabriel, 1982;

product mainly by strong coloration.

Battaglini et al., 1986; Schambil & Schwuger, 1990). The progress of the reaction is decisive for sulfonation degree expressed as amount active matter.



Table 5. Correlation to estimate heat and mass transfer coefficients and properties of organic liquid and gas mixture (Torres & Sánchez, 2009)

Sulfonation/Sulfation Processing Technology for Anionic Surfactant Manufacture 289

area. Same phenomena occur with the film thickness. The jump in conversion takes place in the top reactor, the temperature rises considerably and reduces the viscosity of the liquid, even canceling the effect of viscosity then in the bottom reactor increases composition and the interfacial velocity. Subsequently, the reduced generation heat and descent of the

Transfer rates in the gas phase are affected by changes in the tubular reactor. Increases uncontrolled in the gas flow could drag some liquid into the gas phase. Therefore the gas velocity has to be set at the point where no liquid drops can be pulled to the gas phase. Temperature is a critical parameter in the quality control of the sulfonated products. Although inlet streams' temperature should be adjusted above room to enhance the reaction and avoid the solidification of the organic matter, an adequate control is required due to the high release of heat attributed to the sulfonation reaction. The SO3/organic liquid mole ratio requires rigorous control. Excess of SO3 enhance side reactions and extended reaction times

The comparison obtained for this same process with petrochemicals compounds indicates that the model could be applied to any film sulfonation but adjusting the parameters and specific conditions, such as the physicochemical properties of the compounds used, since the sulfonation process described in this work is one of the more complicated cases. Although some of the physical and chemical properties of mixture are obtained of a similar form, these should be tested and approach to achieve convergence of the model; these yielded the

The model predicts two distinct transfer areas. The first is characterized by an abrupt increase in conversion and temperature, in which the controlling step depends initially of the gas phase and in accordance with the extent of the sulfonation reaction, the viscosity fluid increases, the film thickness is also higher and the film velocity decreases, then the liquid phase becomes the controlling stage with a mild increase of the temperature and conversion. The mathematical model proposed for a film SO3-sulfonation fits adequately the trend of experimental results, so it is now possible to make a prediction on the conversion in a falling film reactor, because the profiles of temperature, density, viscosity and conversion are consistent with experimental results that satisfy the conditions to minimize the strictest

temperature is increase the viscosity again.

will also enhance side reactions.

best results in the mathematical model of falling film reactor.

mathematical calculations mistakes due to the usage of numerical solutions.

 

**7. Conclusions** 

**8. Notation** 

*A* pre-exponential factor, s-1 *A+* a van Driest constant *B+* a van Driest parameter

*c* heat capacity, J/(kmolK) *Cf* friction factor, dimensionless

*f* damping factor, 1.66 1 / *<sup>w</sup> f e*

*g* acceleration due to gravity, m/s2

*D* diffusivity, m2/s *d* reactor diameter, m

*C* concentration, kmol/m3; van Driest constants

The model was constructed to predict the sharp increase in the conversion that takes place in the first stage inside of the falling film reactor. It confirms that the mass transfer is initially controlled by the resistance in the gas phase. After due to several factors the resistance occurs in the liquid phase. Figure 21(a) presents the conversion profile in the film reactor from the model, experimental results showed conversions lower than those predicted by the reactor model in upper reactor region. This is due that the model assumes a fully developed flow and entrance effects of the streams to the reactor are neglected. However, the model is able to predict adequately conversions downstream for longer lengths. In the bottom of reactor is a small jump in the conversion predicted by the model, perhaps due to kinetic effects that reach importance by the consumption of reactants.

Fig. 21. (a) Longitudinal conversion profile for SME and (b) density and viscosity model estimated by the model

The most important outlet data obtained by solving the mathematical model are conversion, density and viscosity of the product. The density and viscosity of the effluent, downstream in the reactor film, estimated by the mathematical model is similar to that obtained experimentally, as shown in the Figure 21(b). The increase of temperature produces a decrease in viscosity enhancing the solubility of SO3 in the liquid, and causing a decrease in film thickness. These variations are the result of abrupt change in composition and release of heat in the initial part of the reactor, which produce an increasing of the temperature in this area. Same phenomena occur with the film thickness. The jump in conversion takes place in the top reactor, the temperature rises considerably and reduces the viscosity of the liquid, even canceling the effect of viscosity then in the bottom reactor increases composition and the interfacial velocity. Subsequently, the reduced generation heat and descent of the temperature is increase the viscosity again.

### **7. Conclusions**

288 Advances in Chemical Engineering

The model was constructed to predict the sharp increase in the conversion that takes place in the first stage inside of the falling film reactor. It confirms that the mass transfer is initially controlled by the resistance in the gas phase. After due to several factors the resistance occurs in the liquid phase. Figure 21(a) presents the conversion profile in the film reactor from the model, experimental results showed conversions lower than those predicted by the reactor model in upper reactor region. This is due that the model assumes a fully developed flow and entrance effects of the streams to the reactor are neglected. However, the model is able to predict adequately conversions downstream for longer lengths. In the bottom of reactor is a small jump in the conversion predicted by the model,

perhaps due to kinetic effects that reach importance by the consumption of reactants.

Fig. 21. (a) Longitudinal conversion profile for SME and (b) density and viscosity model

The most important outlet data obtained by solving the mathematical model are conversion, density and viscosity of the product. The density and viscosity of the effluent, downstream in the reactor film, estimated by the mathematical model is similar to that obtained experimentally, as shown in the Figure 21(b). The increase of temperature produces a decrease in viscosity enhancing the solubility of SO3 in the liquid, and causing a decrease in film thickness. These variations are the result of abrupt change in composition and release of heat in the initial part of the reactor, which produce an increasing of the temperature in this

estimated by the model

Transfer rates in the gas phase are affected by changes in the tubular reactor. Increases uncontrolled in the gas flow could drag some liquid into the gas phase. Therefore the gas velocity has to be set at the point where no liquid drops can be pulled to the gas phase. Temperature is a critical parameter in the quality control of the sulfonated products. Although inlet streams' temperature should be adjusted above room to enhance the reaction and avoid the solidification of the organic matter, an adequate control is required due to the high release of heat attributed to the sulfonation reaction. The SO3/organic liquid mole ratio requires rigorous control. Excess of SO3 enhance side reactions and extended reaction times will also enhance side reactions.

The comparison obtained for this same process with petrochemicals compounds indicates that the model could be applied to any film sulfonation but adjusting the parameters and specific conditions, such as the physicochemical properties of the compounds used, since the sulfonation process described in this work is one of the more complicated cases. Although some of the physical and chemical properties of mixture are obtained of a similar form, these should be tested and approach to achieve convergence of the model; these yielded the best results in the mathematical model of falling film reactor.

The model predicts two distinct transfer areas. The first is characterized by an abrupt increase in conversion and temperature, in which the controlling step depends initially of the gas phase and in accordance with the extent of the sulfonation reaction, the viscosity fluid increases, the film thickness is also higher and the film velocity decreases, then the liquid phase becomes the controlling stage with a mild increase of the temperature and conversion. The mathematical model proposed for a film SO3-sulfonation fits adequately the trend of experimental results, so it is now possible to make a prediction on the conversion in a falling film reactor, because the profiles of temperature, density, viscosity and conversion are consistent with experimental results that satisfy the conditions to minimize the strictest mathematical calculations mistakes due to the usage of numerical solutions.

#### **8. Notation**


Sulfonation/Sulfation Processing Technology for Anionic Surfactant Manufacture 291

Adami from Ballestra S.p.A. (Italy) by communications received. Same wish to thank COLCIENCIAS (Departamento Administrativo de Ciencia, Tecnología e Innovación) for providing financial support. The experimental work presented here was finished at 2009 in the Chemical Engineering Laboratory from Universidad Nacional de Colombia (Colombia), under the direction from Professor Francisco J. Sánchez C. Dr. Paulo C. Narváez R., Dr.

Acmite Market Intelligence. (2010). *Market report. World Surfactant Market.* (Market report

http://www.acmite.com/brochure/Brochure-Surfactant-MarketReport.pdf Agrawal, A. K., & Peckover, R. S. (1980). Nonuniform grid generation for boundary-layer problems. *Comput. Phys. Comm.,* Vol.19, No. 2, pp.171-178, ISSN 0010-4655 Ahmad, S., Siwayanan P., Murad Z. A., Aziz H. A., & Soi H. S. (2007). Beyond biodiesel,

Akanksha, Pant, K. K., & Srivastava, V. K. (2007). Modeling of sulphonation of

Battaglini, G., Larson-Zobus, J., & Baker, T. G. (1986). Analytical methods for alpha sulfo methyl tallowate. *JAOCS,* Vol.63, No. 8, pp.1073-1077, ISSN 1558-9331 Boskamp, J. V. & Houghton, M. P. (1996) Detergent compositions containing primary

Broström, A. (1975). A mathematical model for simulating the sulphonation of

Clippinger, E. (1964). Reactions of alpha-olefins. *Ind. Eng. Chem. Proc. D. D.,* Vol.3, No. 1,

Cohen, L., Soto, F., Melgarejo, A., & Roberts, D. W. (2008). Performance of Φ-sulfo fatty

Dabir, B., Riazi, M. R., & Davoudirad, H. R. (1996). Modelling of falling film reactors. *Chem.* 

Davis, J. E., Van Ouwerkerk, M., & Venkatesh, S. (1979). An analysis of the falling film gasliquid reactor. *Chem. Eng. Sci.,* Vol.34, No. 4, pp.539-550, ISSN 0009-2509 de Groot, W. H. (1991). *Sulphonation technology in the detergent industry*. Kluwer Academic

Díaz, L. (2009). Hydrodynamics analysis of a falling film reactor for the sulfonation of

Foster, N. C. (2004). Manufacture of methyl ester sulfonates and other derivates. *Soaps,* 

methyl esters derived from palm oil. *M.Sc. Thesis*, Universidad Nacional de

*Detergents, Oleochemicals and Personal Care Products*, In: Spitz L. (Ed.), AOCS

*Eng. Sci.,* Vol.51, No. 11, pp.2553-2558, ISSN 0009-2509

Publishing, ISBN 9781893997769, Seattle, USA

Publisher, ISBN 0-7923-1202-3, Dordrecht, The Netherlands.

Ratingen, Germany: Acmite Market Intelligence, Retrieved from

Methyl esters as the route for the production of surfactants feedstock. *INFORM,* 

tridecylbenzene in a falling film reactor. *Math. Comp. Model.,* Vol.46, No. 9-10,

alcohol sulfate, ethoxylated nonionic surfactant, and zeolite builder. *Zeolites,* Vol.

dodecylbenzene with gaseous sulphur trioxide in an industrial reactor of votator type. *Transactions of the Institution of Chemical Engineers,* Vol.53, pp.29-33, ISSN 0046-

methyl ester sulfonate versus linear alkylbenzene sulfonate, secondary alkane sulfonate and α-sulfo fatty methyl ester sulfonate. *J. Surfactants Deterg.,* Vol.11, No.

Oscar Y. Suárez P. and MSc. Luis A. Díaz A. assisted with the experiments.

Vol.18, pp.216-221, ISSN 0897-8026

17, Issues 5-6, pp.525, ISSN 0144-2449

pp.1332-1344, ISSN 0895-7177

pp.3-7, ISSN 0196-4305

3, pp.181-186, ISSN 1558-9293

Colombia, Bogotá, Colombia

**10. References** 

9858


#### Greek letters



#### Subscripts


#### **9. Acknowledgment**

Gratefully acknowledge at Dr. Federico I. Talens Alesson from University of Nottingham (UK), Dr. David W. Roberts from Liverpool John Moores University (UK) and Dr. Icilio Adami from Ballestra S.p.A. (Italy) by communications received. Same wish to thank COLCIENCIAS (Departamento Administrativo de Ciencia, Tecnología e Innovación) for providing financial support. The experimental work presented here was finished at 2009 in the Chemical Engineering Laboratory from Universidad Nacional de Colombia (Colombia), under the direction from Professor Francisco J. Sánchez C. Dr. Paulo C. Narváez R., Dr. Oscar Y. Suárez P. and MSc. Luis A. Díaz A. assisted with the experiments.

#### **10. References**

290 Advances in Chemical Engineering

*k* reaction rate constant, m3/kmols; thermal conductivity, J/msK; mass transfer

*m* Henry constant, (kmol of SO3/m3 of gas)/(kmol of SO3/m3 of liquid)

*Re* Reynolds number, dimensionless: *ReG = ρu(d – 2δ)/μ; ReL = 4Г/μ*.

*y* transversal coordinate (from wall toward the liquid free surface)

*Γ* volumetric flow rate of the liquid per unit wetted perimeter, m2/s

*x* conversion expressed as molar fraction of the sulfonic acid, dimensionless

Gratefully acknowledge at Dr. Federico I. Talens Alesson from University of Nottingham (UK), Dr. David W. Roberts from Liverpool John Moores University (UK) and Dr. Icilio

*ΔH* reaction enthalpy, J/kmol *h* heat transfer coefficient, J/(m2sK)

*L* reactor length, m *M* mole ratio SO3/ME

*P* pressure, atm

*r* reaction rate

*T* temperature, K

*z* axial coordinate

*δ* film thickness, m

*A* absorbing specie

*μ* liquid viscosity, kg/ms *ν* kinematic viscosity, m2/s *ρ* liquid density, kg/m3 *τ* interfacial shear stress, N/m2 *σ* Surface tension, N/m

Greek letters

Subscripts

*ex* exterior *G* gas phase *i* interface *in* interior *L* liquid phase *lm* logarithmic mean

*T* turbulent *w* wall

**9. Acknowledgment** 

*Q* heat of reaction, J/mol

coefficient, kmol/m2s

*Sc* Schmidt number, dimensionless

*v* axial velocity of liquid film, m/s

*N* mass flux of gaseous reactant, kmol/m2s

*U* global heat transfer coefficient, J/(m2sK) *u* turbulence characteristic velocity of gas, m/s

*δ*<sup>+</sup> dimensionless film thickness, *δ*+ = *δu/ ν*

*φ* roughness enhancement factor, dimensionless

*y+* non-dimensional distance to the wall: *y*(*τwg*/*ρ*)½/*ν*


Sulfonation/Sulfation Processing Technology for Anionic Surfactant Manufacture 293

Reid, R. C., Prausnitz,J. M., & Poling,B. E. (1987). *The properties of gases and liquids* (4th ed.)

Resnik, G. O., Vishwanath, P., Pynn, M. A., Sitnik, J. M., Todd, J. J., Wu, J., . . . Jarrell, K. A.

Riazi, M. R., & Faghri, A. (1986). Effect of the interfacial drag on gas absorption with

Roberts, D. W. (1998). Sulfonation technology for anionic surfactant manufacture. *Org. Proc.* 

Roberts, D. W. (2001). Manufacture of anionic surfactants. *Oleochemical Manufacture and* 

Roberts, D. W. (2003). Optimization of linear alkyl benzene sulfonation process for

Roberts, D. W. (2007). The origin of colour formation in methyl ester sulfonation. *Communication Journal Com. Esp. Det.,* Vol.37, pp.153-159, ISSN 0017-3495 Roberts, D. W., Giusti, L., & Forcella, A. (2008). Chemistry of methyl ester sulfonates.

Roberts, D. W., & Garrett, P. R. (2000). Ester sulphonates derived from non-methyl esters. In:

Rosen, M. R. (2005). *Delivery System Handbook for Personal Care and Cosmetic Products:* 

Satsuki, T. (1998). Methyl ester sulfonates, new products and applications in surfactant

Schambil, F., & Schwuger, M. J. (1990). Physico-chemical properties of alpha-sulpho fatty

Stein,W., & Baumann, H. (1975). a–sulfonated fatty acids and esters: manufacturing process, properties, and applications. *JAOCS,* Vol.52, No. 9, pp.323-329, ISSN 1558-9331 Talens, F. I., & Gutiérrez, J. M. (1995). Estimation of the viscosity and density of lauryl

Talens, F. I. (1999). The modelling of falling film chemical reactor. *Chem. Eng. Sci.,* Vol.54,

Talens, F. I. (2000). The effect of supercritical gas velocity on wavy films and its use in

McGraw-Hill Book Company. ISBN 0-07-051799-1, NY, USA

*Microbiol. Biot.,* Vol.85, No. 5, pp.1387-1397, ISSN 0175-7598

*Res. Dev.*, Vol. *2, No.3*, pp. 194–202, ISSN 1520-586X

*INFORM,* Vol.19, No. Suppl. 5, pp.2-9, ISSN 0897-8026

Academic Press, ISBN 9780849397851, UK

075-8, Kuala Lumpur, Malaysia, 1999

Academic Press, ISBN 978-084-939-738-7, UK

8155-1504-9, Norwich, N.Y., USA

No. 6, pp.380-385, ISSN 0932-3414

No. 12, pp.1871-1881, ISSN 0009-2509

No.7, pp.629-632, ISSN 0930-7516

0120-5609

1541

586X

0263-8762

presence of mono-, di- and triglycerides. *Ing. Investig.*, Vol. 57, No. 1, pp.58-62, ISSN

(2010). Use of sustainable chemistry to produce an acyl amino acid surfactant. *Appl.* 

chemical reaction in a vertical tube. *AIChE J.,* Vol.32, No. 4, pp.696-699, ISSN 0001-

*Applications*. In: Gustone F. D., & R. J. Hamilton (Eds.), pp. 55-73, Sheffield

surfactant manufacture. *Org. Process Res. Dev.,* Vol.7, No. 2, pp.172-184, ISSN 1520-

*Proceedings of the 1999 PORIM International Palm Oil Congress, "Emerging Technologies and Opportunities in the Next Millenium",* pp. 209-215, ISBN 978-967-961-

*Technology, Applications, and Formulations*. William Andrew Publishing, ISBN: 0-

technology. In: *Annual surfactants review*, Karsa D. R. (Ed.), pp. 123-155, Sheffield

acid methyl esters and alpha-sulpho fatty acid di salts. *Tenside Surfact. Det.,* Vol.37,

alcohol/laurylsulphuric acid mixtures as a function of the temperature and laurylsulphuric and molar fraction. *Chem. Eng. Res. Des.,* Vol.73A, pp.206-207, ISSN

enhancing the performance of falling film reactors. *Chem. Eng. Technol.,* Vol.23,


Foster, N. C. (1997). *Sulfonation and sulfation processes*. The Chemithon Corporation,

http://www.chemithon.com/Resources/pdfs/Technical\_papers/Sulfo%20and%2

Gutiérrez, G. J., Mans, T. C., & Costa, L. J. (1988). Improved mathematical model for a falling

Inagaki, T. (2001). Development of α-Sulfo fatty acid esters. *Proceedings World Conference on Oleochemicals,* ISBN 0-935315-34-9, Kuala Lumpur, Malaysia, October 8th -12th, 2000 Ingegard, J. & Martin, S. (2001). Surfactants based on fatty acids and other natural hydrophobes. *J. Colloid Interf. Sci.*, Vol.6, pp. 178-188, ISSN 0021-9797 Johansson, I., & Svensson, M. (2001). Surfactants based on fatty acids and other natural hydrophobes. *Curr. Opin. Colloid In.,* Vol.6, No. 2, pp.178-188, ISSN 0196-4305 Johnson, G. R., & Crynes, B. L. (1974). Modelling of a thin–film sulfur trioxide sulfonation reactor. *Ind. Eng. Chem. Proc. D. D.,* Vol.13, No. 1, pp.6-14, ISSN 0196-4305 Kapur, B. L., Solomon, J. M. & Bluestein, B. R. (1978) Summary of the technology for the

sulfonation reactor. *Ind. Eng. Chem. Res.,* Vol.27, No. 9, pp.1701-1707, ISSN 1520-

manufacture of higher alpha-sulfo fatty acid esters. *JAOCS*, Vol.55, pp. 549-557,

process for surfactant manufacture" and "Sulfonation technology for anionic surfactant manufacture": Falling film SO3 sulfonation – Laminar or turbulent flow

*PORIM International Palm Oil Congress, "Emerging Technologies and Opportunities in the Next Millenium",* pp. 229-250, ISBN 978-967-961-075-8, Kuala Lumpur, Malaysia

alkylbenzene sulfonate, alcohol ethoxylate, alcohol ethoxy sulfate, alcohol sulfate, and soap. *Environ. Toxicol. Chem*. Vol.18, No. 11, pp.2634–2644, ISSN 1552-8618 McCready, M. J., & Hanratty, T. J. (1984). A comparison of turbulent mass transfer at gas-

liquid and solid-liquid interfaces. *Gas transfer at water surfaces*. In: Brutsaert W., & Jirka G. H. (Eds.), pp. 283-292, Reidel Publishing Company, ISBN 9027716978,

the analysis of chemical engineering phenomena. *CT&F*, Vol. 3, No. 5, pp. 205-224,

environment: A review. *Int. Biodet. Biodeg.*, Vol. 63, No. 8, pp. 981–987, ISSN 0964-

chromatography determination of fatty acid alkyl esters (methyl and ethyl) in the

Knaggs, E. (2004). Comment on "Optimization of the linear alkyl benzene sulfonation

controversy. *Org. Proc. Res. Dev.*, Vol.8, No. 6, pp.1079-1080, ISSN 1520-586X Lamourelle, A., & Sandall, O. (1972). Gas absorption into a turbulent liquid. *Chem. Eng. Sci.,* 

MacArthur, B. W., Brooks, B., Sheats W.B., & Foster, N. C. (1999). In: *Proceedings of the 1999* 

Mann, R., & Moyes, H. (1977). Exothermic gas absorption with chemical reaction. *AIChE J.,* 

Matthijs, E., Holt, M.S., Kiewiet, A., Rijs, G.B.J. (1999). Environmental monitoring for linear

Milwdsky, B. M., & Gabriel, D. M. (1982). *Detergents Analysis – A Handbook for cost-effective quality control*, John Wiley & Sons, Inc., ISBN 0-7114-5735-2, London, UK. Morales, G. M., & Martínez, R. R. (2009). Molecular and multiscale modeling: New tools in

Mungray, A. K., & Kumar, P. (2009). Fate of linear alkylbenzene sulfonates in the

Narváez, R. P. C., Torres, J. A., Sánchez, C. F. J., Ponce de León, Q. L. F. (2005). Gas

Retrieved from:

0Sulfa%201.pdf

ISSN 1558-9331

Vol.27, pp.1035-1043, ISSN 0009-2509

Vol.23, No. 1, pp.17-23, ISSN 0001-1541

Dordrecht, Netherlands

ISSN 0122-5383

8305

5045

presence of mono-, di- and triglycerides. *Ing. Investig.*, Vol. 57, No. 1, pp.58-62, ISSN 0120-5609


**12** 

*Poland* 

Grzegorz Wielgosiński *Technical University of Lodz,* 

*Faculty of Process and Environmental Engineering* 

**Pollutant Formation in Combustion Processes** 

Each combustion process is a source of various emissions. During combustion, are formed not only carbon dioxide and water, but still a lot of other products of combustion and incomplete combustion. Knowledge of the mechanisms and the pathways of formation allow the use of so-called primary methods of reducing emissions and thereby reduce

From the chemical point of view combustion is the oxidation reaction of organic compounds. Organic compounds - a variety of hydrocarbons which have in the molecule the atoms of carbon (*C*) and hydrogen (*H*). The oxidation reaction of hydrocarbons is accompanied by the emission of large amount of heat - that is why the reaction is exothermic. Hydrocarbons and their derivatives containing other atoms in the molecule, such as sulfur (*S*), nitrogen (*N*), oxygen (*O*), chlorine (*Cl*), etc. make a flammable substance which can be called fuel only when it meets certain qualitative conditions. Fuels can be divided according to several methods. One of them is the division according to the state of matter - into solid, liquid and gas fuels. Another method is the division according to the origin - natural fuels (e.g. fossil fuels) and synthetic fuels produced by processing natural

Natural solid fuels include mainly coal, lignite, peat, wood, etc. Natural liquid fuels include first of all oil from which numerous synthetic fuels such as petrol, kerosene, diesel, fuel oil, etc. are produced. In a group of natural gas fuels natural gas is crucial, while synthetic gas fuels include blast furnace gas, generator gas, water gas, city gas or extracted gas fractions

Another type of fuel are wastes and produced from them the so-called alternative fuel which is known as refuse derived fuel (RDF) or secondary recovered fuel (SRF). Generally, fuel, in addition to the information about its form (physical state), is first of all characterized by elemental composition i.e. the contents of basic elements such as carbon (*C*), hydrogen (*H*), sulfur (*S*), nitrogen (*N*), oxygen (*O*), chlorine (*Cl*), etc. Apart form the contents of these elements, the content of water (*W*) and non-flammable substance (A), also known as ash, in the fuel are extremely important. Typical elemental composition of selected solid, liquid and

such as methane, propane or butane as well as their mixtures.

**1. Introduction** 

fuels.

emissions to the atmosphere.

gas fuels is shown in Table 1.

**2. Various fuels composition** 


## **Pollutant Formation in Combustion Processes**

#### Grzegorz Wielgosiński

*Technical University of Lodz, Faculty of Process and Environmental Engineering Poland* 

#### **1. Introduction**

294 Advances in Chemical Engineering

Torres, J. A. (2009). Production of alfa-sulfo methylstearate in a falling film reactor. *PhD* 

Torres, J. A., Díaz, L., & Sánchez, F. J. (2008a). Kinetic study of the sulfonation of methyl

Torres, J. A., Díaz, L., & Sánchez, F. J. (2009a). Falling film reactor for methyl ester

Torres, J. A., Morales, G. M., Suárez, P. O. Y., & Sánchez, F. J. (2009b). Mathematical model

Torres, J. A., Narváez, P. C., Suárez, O. Y., & Sánchez, F. J. (2008b). Sulfonation of methyl

Torres, J. A., & Sánchez, F. J. (2008). Physicochemical properties of the liquid mixture of

Torres, J. A., & Sánchez, F. J. (2009). Small scale SO3-sulfonation of methyl esters. *Program* 

Torres, J. A., Sánchez. C. F. J, Narváez, R. P. C., Ponce de León, Q. L. F. (2005). Production of

van Driest, E. R. (1956). On turbulent flow near a Wall. *Journal of the Aeronautical Sciences,* 

Vignes, A. (1966). Diffusion in Binary Solutions. Variation of Diffusion Coefficient with Composition. *Ind. Eng. Chem. Fundam.*, Vol. 5, pp.189-199, ISSN 0196-4313 Villadsen, J., & Nielsen, H. P. (1986). Models for strongly exothermic absorption and reaction in falling films. *Chem. Eng. Sci.,* Vol.41, No. 6, pp.1655-1671, ISSN 0009-2509 Wilke, C. R., & Chang, P. (1955). Correlation of diffusion coefficients in dilute solutions.

Yamada K, & Matsutani S. (1996). Analysis of the dark colored impurities in sulfonates fatty

Yih, S. M., & Seagrave, R. C. (1978). An analytical solution for mass transfer in turbulent

Yih, S. M., & Liu, J. L. (1983). Prediction of heat transfer in turbulent falling liquid films with or without interfacial shear. *AIChE J.,* Vol.29, No. 6, pp.903-909, ISSN 0001-1541 Zhu, Y. P., Rosen, M. J., Morral, S. W. & Tolls, J. (1998) Surface properties of linear alkyl

falling films with or without chemical reaction. *Chem. Eng. Sci.,* Vol.33, No. 12,

benzene sulfonate in hard river water. *J. Surfartants Deterg*. Vol. 1, No. 2, pp.187-

*Investigaciones de la UNAD*, Vol.4, No. 2, pp. 150-162, ISSN 0124-793X. Tsubochi, M., Yamasaki, N., & Matsuoka, K. (1979). Determination of anionic surfactants by

*JAOCS,* Vol.56, No. 11, pp.921-923, ISSN 1558-9331

*AIChE J.,* Vol.1, No. 2, pp.264-270, ISSN 0001-1541

acid methyl ester. *JAOCS*, Vol.73, pp.121-125, ISSN 1558-9331

esters derived of palm stearin. *Corpoica Cienc. Tecnol. Agropecu.,* Vol.9, No. 2, pp.88-

sulphonation with gaseous sulphur trioxide. *Ing. Investig.,* Vol.29, No. 3, pp.48-53,

of a falling film reactor for methyl ester sulfonation. *Chem. Prod. Process Model.* Vol.

esters derived from palm oil. *Corpoica Cienc. Tecnol. Agropecu.*, Vol.9, No. 2, pp.77-

methyl stearate/sulfo acid methyl stearate. *Publ. Invest.,* Vol.2, No. 1, pp.97-118,

*Book 8th World Congress of Chemical Engineering,* ISBN 0-920804-44-6, Montréal,

methyl stearate derived from stearin hydrogenated derived of palm oil. *Revista de* 

two-phase titration with tetrabromophenolphthalein ethyl ester as indicator.

*Thesis*, Universidad Nacional de Colombia, Bogotá, Colombia

95, ISSN 0122-8706

87, ISSN 0122-8706

ISSN 1900-6608

4, Iss. 5, Art. 12, ISSN 1934-2659

Canada, August 23th - 27th, 2009

Vol.23, pp.1007-1011, ISSN 0095-9812

pp.1581-1587, ISSN 0009-2509

193, ISSN 1558-9293

ISSN 0120-5609

Each combustion process is a source of various emissions. During combustion, are formed not only carbon dioxide and water, but still a lot of other products of combustion and incomplete combustion. Knowledge of the mechanisms and the pathways of formation allow the use of so-called primary methods of reducing emissions and thereby reduce emissions to the atmosphere.

#### **2. Various fuels composition**

From the chemical point of view combustion is the oxidation reaction of organic compounds. Organic compounds - a variety of hydrocarbons which have in the molecule the atoms of carbon (*C*) and hydrogen (*H*). The oxidation reaction of hydrocarbons is accompanied by the emission of large amount of heat - that is why the reaction is exothermic. Hydrocarbons and their derivatives containing other atoms in the molecule, such as sulfur (*S*), nitrogen (*N*), oxygen (*O*), chlorine (*Cl*), etc. make a flammable substance which can be called fuel only when it meets certain qualitative conditions. Fuels can be divided according to several methods. One of them is the division according to the state of matter - into solid, liquid and gas fuels. Another method is the division according to the origin - natural fuels (e.g. fossil fuels) and synthetic fuels produced by processing natural fuels.

Natural solid fuels include mainly coal, lignite, peat, wood, etc. Natural liquid fuels include first of all oil from which numerous synthetic fuels such as petrol, kerosene, diesel, fuel oil, etc. are produced. In a group of natural gas fuels natural gas is crucial, while synthetic gas fuels include blast furnace gas, generator gas, water gas, city gas or extracted gas fractions such as methane, propane or butane as well as their mixtures.

Another type of fuel are wastes and produced from them the so-called alternative fuel which is known as refuse derived fuel (RDF) or secondary recovered fuel (SRF). Generally, fuel, in addition to the information about its form (physical state), is first of all characterized by elemental composition i.e. the contents of basic elements such as carbon (*C*), hydrogen (*H*), sulfur (*S*), nitrogen (*N*), oxygen (*O*), chlorine (*Cl*), etc. Apart form the contents of these elements, the content of water (*W*) and non-flammable substance (A), also known as ash, in the fuel are extremely important. Typical elemental composition of selected solid, liquid and gas fuels is shown in Table 1.

Pollutant Formation in Combustion Processes 297

Fuel Heat of combustion (upper heal value) Heat value (lower heat value)

hydrogen 147,8 119,9 acetylene 49,9 48,2 methane 55,5 50,1 propane 50,3 46,4 butane 49,5 45,7 natural gas 41,1 37,5 blast furnace gas 4,1 4,0 coke oven gas 19,7 17,5 petrol 46,1 42,7 diesel 44,7 41,8 heating oil 44,8 42,7 coal 19-31 lignite 6-14 peat 12-16 coke 28-32 wood 10-18

Table 2. Combustion heat and calorific value of fuel (in MJ/kg)

value can be calculated from the formula (Recknagel et al., 1994):

Heat of combustion can be determined experimentally in an adiabatic calorimeter. However, fuel usually contains also water (moisture) whose presence and evaporation significantly reduces the amount of heat generated during the combustion of the fuel unit. In order to determine the amount of heat that can be effectively generated in the combustion process the concept of heat (calorific) value (*Hu*), called the lower heat value,

The calorific (heat) value is the amount of heat emitted during combustion of mass unit or volume unit of fuel at its complete and perfect combustion, assuming that the steam contained in the exhaust gas does not condense, even though the temperature of the exhaust gases reach the initial temperature of the fuel. Knowing the heat of combustion the calorific

*<sup>u</sup> <sup>o</sup> <sup>w</sup> <sup>H</sup> <sup>H</sup>* <sup>9</sup>*<sup>H</sup> <sup>W</sup> r* (2)

where:

was introduced.

*Ho* – combustion heat [MJ/kg]

*C* – carbon content in the fuel [kg/kg], *H* – hydrogen content in the fuel [kg/kg], *S* – sulfur content in the fuel [kg/kg], *N* – nitrogen content in the fuel [kg/kg], *O* – oxygen content in the fuel [kg/kg].


Table 1. Typical elemental composition of selected solid, liquid and gas fuels [in %]

#### **3. Lower and upper heat value**

The elemental composition determines another very important parameter which characterizes fuel i.e. heat of combustion so called upper heat value. Heat of combustion (*Ho*) - is the amount of heat that is generated during complete and perfect burning of unit weight or unit volume of the analyzed substance in constant volume, wherein:


The term complete combustion means the process in which all the fuel (organic matter) will be oxidized (burned) and perfect combustion occurs when all the combustion products are non-flammable. Data on typical values of combustion heat of the selected fuels are shown in Table 2.

Heat of combustion (*Ho*) can be easily estimated on the basis of the elemental composition. The approximate value for the heat of combustion of solid and liquid fuels (upper heat value) can be calculated from the formula (Recknagel et al., 1994):

$$H\_{\rho} = 34,8\ \text{\$\mathcal{C} \gets 9\\$,9\ \$\mathcal{H} + 10,5\ \text{\$\mathcal{S} \gets 6\$},\\$\cdot\text{\$N\$-10},\\$\cdot\text{\$O\$}\tag{1}$$

where:

296 Advances in Chemical Engineering

Fuel *C H O N S Cl W A*  hydrogen - 100 - - - - - acetylene 92,3 7,7 - - - - - methane 75 25 - - - - - propane 81,8 18,2 - - - - - butane 83 17 - - - - - natural gas 69,4 22,9 0,7 7,0 0,05 - - blast furnace gas 41,6 19,4 9,0 2,2 - - - coke oven gas 15,0 50,0 23,0 12,0 - - 8,5 petrol 85 15 - - 0,05 - - diesel 87 13 - - 0,05 - - heating oil 86 13 - - 0,2 - - coal 72-83 3,4-5,3 1,8-12,5 1,0-1,2 0,1-6,5 0,1-1 3-20 3-30 lignite 25-77 2-6,5 1,5-20 0,1-0,5 0,2-6,5 0,1-1 10-60 3-30 peat 40 5 25 2 1 0,1 20 7 coke 80-90 0,3-1,5 1,0-2,0 1,2-2,2 0,5-1,0 0,5 1,5-8,5 2-17 wood 35-50 5-7 35-45 0,3-07 0,01-0,1 0,01 5-65 0,3-7

Table 1. Typical elemental composition of selected solid, liquid and gas fuels [in %]

weight or unit volume of the analyzed substance in constant volume, wherein:



value) can be calculated from the formula (Recknagel et al., 1994):

The elemental composition determines another very important parameter which characterizes fuel i.e. heat of combustion so called upper heat value. Heat of combustion (*Ho*) - is the amount of heat that is generated during complete and perfect burning of unit

The term complete combustion means the process in which all the fuel (organic matter) will be oxidized (burned) and perfect combustion occurs when all the combustion products are non-flammable. Data on typical values of combustion heat of the selected fuels are shown in

Heat of combustion (*Ho*) can be easily estimated on the basis of the elemental composition. The approximate value for the heat of combustion of solid and liquid fuels (upper heat

*<sup>H</sup> <sup>C</sup> <sup>H</sup> <sup>S</sup> <sup>N</sup> <sup>O</sup> <sup>o</sup>* 34,8 93,9 10,5 6,3 10,8 (1)

**3. Lower and upper heat value** 

Table 2.






Table 2. Combustion heat and calorific value of fuel (in MJ/kg)

Heat of combustion can be determined experimentally in an adiabatic calorimeter. However, fuel usually contains also water (moisture) whose presence and evaporation significantly reduces the amount of heat generated during the combustion of the fuel unit. In order to determine the amount of heat that can be effectively generated in the combustion process the concept of heat (calorific) value (*Hu*), called the lower heat value, was introduced.

The calorific (heat) value is the amount of heat emitted during combustion of mass unit or volume unit of fuel at its complete and perfect combustion, assuming that the steam contained in the exhaust gas does not condense, even though the temperature of the exhaust gases reach the initial temperature of the fuel. Knowing the heat of combustion the calorific value can be calculated from the formula (Recknagel et al., 1994):

$$\left\| H \right\|\_{\mu} = H\_{o} - \left( \Phi \cdot H + \mathcal{W} \right) \cdot r\_{w} \tag{2}$$

Pollutant Formation in Combustion Processes 299

conversion of solid fuels into gas fuels under the influence of gasifying medium e.g. air or steam or both. During the flow through the glowing layers of solid fuel intended for the gasification a reaction of oxygen or water with carbon takes place in result of which carbon monoxide (*CO*) and hydrogen (*H2*) are formed as well as small amounts of methane (*CH*4), carbon dioxide (*CO2*), nitrogen (*N2*) and water steam (*H2O*). The process of gasification of solid hydrocarbons can be described by the following simplified chemical equations

*<sup>C</sup> <sup>O</sup> CO <sup>q</sup>* <sup>2</sup> <sup>2</sup>

which are accompanied at the same time by thermal decomposition according to the

Some of the mentioned above chemical reactions are exothermic (+*q* - proceeding with heat

The stream of flue gases from the combustion process consists of a stream resulting from the oxidation reaction of flue gases and a stream of excess air supplied to the combustion process (with excess air coefficient λ greater than 1). Knowing the elemental composition of the fuel or only its calorific value it is possible to estimate both stoichiometric amount of air needed for combustion process and the amount of flue gases generated in the process.

For liquid and solid fuels, the amount of air needed for combustion reaction (stoichiometric)

emission) and some endothermic (-*q* - requiring supplying heat from the outside).

**5. Volume of flue gases from combustion processes** 

can be calculated from the formula (Recknagel et al., 1994):

*<sup>C</sup> <sup>H</sup> <sup>O</sup> CO <sup>H</sup> <sup>O</sup> CO CH <sup>C</sup> <sup>H</sup> CH <sup>O</sup> <sup>C</sup> <sup>H</sup> <sup>C</sup> <sup>q</sup> <sup>m</sup> <sup>n</sup> heat <sup>x</sup> <sup>y</sup> <sup>z</sup>* ... <sup>2</sup> <sup>2</sup> <sup>4</sup> <sup>2</sup> <sup>6</sup> <sup>2</sup> (12)

1 (4)

*<sup>C</sup> <sup>O</sup> CO <sup>q</sup>* <sup>2</sup> <sup>2</sup> (5)

*<sup>C</sup> <sup>H</sup> <sup>O</sup> CO <sup>H</sup> <sup>q</sup>* <sup>2</sup> <sup>2</sup> <sup>2</sup> <sup>2</sup> 2 (6)

*<sup>C</sup> <sup>H</sup> <sup>O</sup> CO <sup>H</sup> <sup>q</sup>* <sup>2</sup> <sup>2</sup> (7)

*<sup>C</sup> <sup>H</sup> CH <sup>q</sup>* <sup>2</sup> <sup>4</sup> 2 (8)

*CO <sup>H</sup> <sup>O</sup> CO <sup>H</sup> <sup>q</sup>* <sup>2</sup> <sup>2</sup> <sup>2</sup> (9)

*CO <sup>H</sup> CH <sup>H</sup> <sup>O</sup> <sup>q</sup>* <sup>2</sup> <sup>4</sup> <sup>2</sup> 3 (10)

*<sup>C</sup> CO* <sup>2</sup>*CO <sup>q</sup>* <sup>2</sup> (11)

(Thome-Kozmiensky, 1994):

simplified reaction equation:

where:

*Hu* – calorific value [MJ/kg] *W* – water content in the fuel [kg/kg], *rw* – water evaporation heat [MJ/kg].

#### **4. Oxygen excess**

Every combustion process is conducted with an excess of oxygen relative to stoichiometric amount resulting from the oxidation reaction of individual components of the fuel. The measure of this excess is so called coefficient of excess air () defined as a ratio of the actual amount of air supplied to the combustion process to the theoretical amount of air required to carry out the complete and perfect combustion – the amount resulting from the stoichiometry of combustion (oxidation) reaction. If in the combustion process coefficient is equal to 1 then exactly the same amount of air that is required by the stoichiometry of the oxidation reaction is supplied to the process. If the coefficient amounts to less than 1 the combustion process is incomplete. As mentioned above, each actual combustion process is conducted with an excess of air, hence the coefficient is usually more than 1. Practically, its value depends on the type of fuel burned (gaseous and liquid fuels require less excess air than solid fuels) and design of the equipment in which combustion is carried out.

If the thermal process takes place without the presence of oxygen it is the process of pyrolysis. Pyrolysis – is the thermal decomposition of flammable substance without oxygen. This is an endothermic process that requires supplying heat from the outside. Pyrolysis process usually takes place at temperatures of 450-900°C in which the organic part of the fuel is converted into: the gas phase (pyrolysis gas), the liquid phase (pyrolysis oils) and solid phase (pyrolysis coke). Pyrolysis is sometimes also called degassing or carbonization. Depending on the temperature we distinguish: smouldering (below 600°C) and coking (above 600°C).

The composition and quantity of pyrolysis products depend on the type of fuel, their physicochemical properties and the temperature of the process. During the process of pyrolysis the mass of fuel is converted into (Thome-Kozmiensky, 1994):


The process of pyrolysis of hydrocarbons can be described by the following simplified chemical equation:

$$\text{C}\_{x}\text{H}\_{y}\text{O}\_{z} \xrightarrow[\text{heat}]{} \text{CO}\_{2} + \text{H}\_{2}\text{O} + \text{CO} + \text{CH}\_{4} + \text{C}\_{2}\text{H}\_{6} + \text{CH}\_{2}\text{O} + ... + \text{C}\_{m}\text{H}\_{n} + \text{C}-\text{q} \tag{3}$$

However, if the thermal process takes place with limited access of air (oxygen) – when the coefficient is less than 1, often fuel gasification process occurs. Gasification is a total

Every combustion process is conducted with an excess of oxygen relative to stoichiometric amount resulting from the oxidation reaction of individual components of the fuel. The

amount of air supplied to the combustion process to the theoretical amount of air required to carry out the complete and perfect combustion – the amount resulting from the stoichiometry of combustion (oxidation) reaction. If in the combustion process coefficient

is equal to 1 then exactly the same amount of air that is required by the stoichiometry of the

combustion process is incomplete. As mentioned above, each actual combustion process is

value depends on the type of fuel burned (gaseous and liquid fuels require less excess air

If the thermal process takes place without the presence of oxygen it is the process of pyrolysis. Pyrolysis – is the thermal decomposition of flammable substance without oxygen. This is an endothermic process that requires supplying heat from the outside. Pyrolysis process usually takes place at temperatures of 450-900°C in which the organic part of the fuel is converted into: the gas phase (pyrolysis gas), the liquid phase (pyrolysis oils) and solid phase (pyrolysis coke). Pyrolysis is sometimes also called degassing or carbonization. Depending on the temperature we distinguish: smouldering (below 600°C) and coking

The composition and quantity of pyrolysis products depend on the type of fuel, their physicochemical properties and the temperature of the process. During the process of




The process of pyrolysis of hydrocarbons can be described by the following simplified

However, if the thermal process takes place with limited access of air (oxygen) – when the coefficient is less than 1, often fuel gasification process occurs. Gasification is a total

*<sup>C</sup> <sup>H</sup> <sup>O</sup> CO <sup>H</sup> <sup>O</sup> CO CH <sup>C</sup> <sup>H</sup> CH <sup>O</sup> <sup>C</sup> <sup>H</sup> <sup>C</sup> <sup>q</sup> <sup>m</sup> <sup>n</sup> heat <sup>x</sup> <sup>y</sup> <sup>z</sup>* ... <sup>2</sup> <sup>2</sup> <sup>4</sup> <sup>2</sup> <sup>6</sup> <sup>2</sup> (3)

pyrolysis the mass of fuel is converted into (Thome-Kozmiensky, 1994):

sulfide, ammonia, hydrogen chloride and hydrogen fluoride

aldehydes, alcohols and organic acids.

than solid fuels) and design of the equipment in which combustion is carried out.

) defined as a ratio of the actual

amounts to less than 1 the

is usually more than 1. Practically, its

where:

*Hu* – calorific value [MJ/kg]

**4. Oxygen excess** 

(above 600°C).

substances

chemical equation:

*W* – water content in the fuel [kg/kg], *rw* – water evaporation heat [MJ/kg].

measure of this excess is so called coefficient of excess air (

oxidation reaction is supplied to the process. If the coefficient

conducted with an excess of air, hence the coefficient

conversion of solid fuels into gas fuels under the influence of gasifying medium e.g. air or steam or both. During the flow through the glowing layers of solid fuel intended for the gasification a reaction of oxygen or water with carbon takes place in result of which carbon monoxide (*CO*) and hydrogen (*H2*) are formed as well as small amounts of methane (*CH*4), carbon dioxide (*CO2*), nitrogen (*N2*) and water steam (*H2O*). The process of gasification of solid hydrocarbons can be described by the following simplified chemical equations (Thome-Kozmiensky, 1994):

$$\text{C} + \frac{1}{2}\text{O}\_2 \xrightarrow{\text{---}} \text{CO} + q \tag{4}$$

$$\text{C} + \text{O}\_2 \xrightarrow{\text{-}} \text{CO}\_2 + q \tag{5}$$

$$\text{C} + 2\text{H}\_2\text{O} \xrightarrow{\text{C}} \text{CO}\_2 + 2\text{H}\_2 - q \tag{6}$$

$$\text{C} + \text{H}\_2\text{O} \xrightarrow{\text{C}} \text{CO} + \text{H}\_2 - q \tag{7}$$

$$\text{C} + 2\text{H}\_2 \xrightarrow{\text{-}} \text{CH}\_4 + q \tag{8}$$

$$\text{CO} + \text{H}\_2\text{O} \xrightarrow{\text{H}\_2\text{O}} \text{CO}\_2 + \text{H}\_2 + q \tag{9}$$

$$\text{CH} + \text{\text{\textdegree}}H\_2 \xrightarrow{\text{\textdegree}} \text{CH}\_4 + H\_2O + q \tag{10}$$

$$\text{C} + \text{CO}\_2 \xrightarrow{\text{-}} \text{2CO} - q \tag{11}$$

which are accompanied at the same time by thermal decomposition according to the simplified reaction equation:

$$\text{C}\_{x}\text{H}\_{y}\text{O}\_{z} \xrightarrow[\text{heat}]{} \text{CO}\_{2} + \text{H}\_{2}\text{O} + \text{CO} + \text{CH}\_{4} + \text{C}\_{2}\text{H}\_{6} + \text{CH}\_{2}\text{O} + ... + \text{C}\_{m}\text{H}\_{n} + \text{C}-\text{q} \tag{12}$$

Some of the mentioned above chemical reactions are exothermic (+*q* - proceeding with heat emission) and some endothermic (-*q* - requiring supplying heat from the outside).

#### **5. Volume of flue gases from combustion processes**

The stream of flue gases from the combustion process consists of a stream resulting from the oxidation reaction of flue gases and a stream of excess air supplied to the combustion process (with excess air coefficient λ greater than 1). Knowing the elemental composition of the fuel or only its calorific value it is possible to estimate both stoichiometric amount of air needed for combustion process and the amount of flue gases generated in the process.

For liquid and solid fuels, the amount of air needed for combustion reaction (stoichiometric) can be calculated from the formula (Recknagel et al., 1994):

Pollutant Formation in Combustion Processes 301

Taking into account the current temperature (*T*) and flue gases pressure (*P*) it is possible to

273

Heat emitted in a chemical reaction of oxidation in the combustion process is used for heating of flue gases from the process (combustion products and excess air supplied to the combustion process), solid secondary products of the combustion process (slag and ashes) and the equipment in which the combustion process is conducted (plus the obvious loss to

It is obvious that, if the generated amount of heat is too small, the combustion process does not run properly, its temperature is too low and the flue gases do not reach the required temperature. If, however, fuel of high calorific value is burned the amount of heat is sufficient to ensure proper combustion temperature and temperature of flue gases. This is particularly important in the case of waste incineration, where the law (Directive 2000/76/EC on the incineration of waste) requires that the exhaust gases have a minimum temperature of 850°C and remain in this temperature for at least 2 seconds when the waste fuel contains less than 1% chlorine. When the chlorine content in waste exceeds 1% it is required to achieve a minimum temperature of 1100°C. If the required temperature of flue gas (and the combustion process) is reached in the device, in which the combustion process is run, it is called autothermal combustion process, but when due to too low calorific value of fuel the required temperatures cannot be reached, combustion is not an autothermal

As mentioned at the beginning from the chemical point of view combustion is the reaction of oxidation of hydrocarbons. Combustion of gaseous fuel is the simplest combustion, in

*<sup>T</sup> VV <sup>N</sup>*

the environment) up to the process temperature. This is illustrated in Figure 1.

*P*

1013

<sup>273</sup> (21)

estimate flue gas stream in the real conditions from the formula:

**6. Combustion as an autothermal process** 

Fig. 1. Simplified heat balance in the combustion chamber

process and requires additional fuel of higher calorific value.

**7. Combustion as a set of chemical reactions** 

where:

*T* – flue gas temperature [oC] *P* – flue gas pressure [hPa]

$$V\_T = 8,88 \cdot C + 26,44 \cdot H + 3,32 \cdot S - 3,33 \cdot O \tag{13}$$

where:

*VT* – minimal (stoichiometric) amount of air for combustion converted to standard conditions (T = 273K, P = 1013hPa) [m3/kg]

The volume of flue gases can be calculated on the basis of known elemental composition of the fuel from the formula (Recknagel et al., 1994):

$$V\_{\rm S} = 1,85 \cdot \text{C} + 11,11 \cdot H + 0,68 \cdot \text{S} + 0,8 \cdot N + 1,24 \cdot W \tag{14}$$

where:

*VS* – volume of flue gas from the combustion process converted to standard conditions [m3/kg]

*N* – nitrogen content in the fuel [kg/kg],

In case the elemental composition of fuel is not known the minimum amount of combustion air and exhaust gas volume can be calculated from approximate formulas taking into account only the calorific value of fuel (Recknagel et al., 1994):


$$V\_T = 0, \Im 41 \cdot H\_u + 0, \Im \tag{15}$$

$$V\_S = 0, \mathcal{Z}1\mathcal{2} \cdot H\_u + 1, 65\tag{16}$$


$$V\_T = 0, \mathcal{D}\mathbf{0} \cdot H\_{\
u} + \mathcal{D}\_r \mathbf{0} \tag{17}$$

$$V\_S = 0.265 \cdot H\_u \tag{18}$$

The total gas flow (after taking into account the excess air coefficient λ) under normal conditions can be calculated from the formula:

$$\boldsymbol{V}\_{\mathcal{C}} = (\mathcal{A} - \mathbf{1}) \cdot \boldsymbol{V}\_{T} + \boldsymbol{V}\_{\mathcal{S}} \tag{19}$$

Multiplying the value of the total flue gas stream *VC* by the amount of fuel burned per unit of time we can calculate the volumetric flow of flue gases from combustion process converted to standard conditions:

$$\boldsymbol{V}\_{N} = \boldsymbol{B} \cdot \boldsymbol{V}\_{\mathcal{C}} \tag{20}$$

where:

*VN* – volumetric flow of flue gases from the combustion process converted to standard conditions [m3/h]

*B* – fuel consumption [kg/h],

Taking into account the current temperature (*T*) and flue gases pressure (*P*) it is possible to estimate flue gas stream in the real conditions from the formula:

$$V = V\_N \cdot \frac{T + 273}{273} \cdot \frac{1013}{P} \tag{21}$$

where:

300 Advances in Chemical Engineering

*VT* – minimal (stoichiometric) amount of air for combustion converted to standard

The volume of flue gases can be calculated on the basis of known elemental composition of

*VS* – volume of flue gas from the combustion process converted to standard conditions

In case the elemental composition of fuel is not known the minimum amount of combustion air and exhaust gas volume can be calculated from approximate formulas taking into

<sup>5</sup> 0,241 0, *<sup>T</sup> <sup>u</sup> <sup>V</sup> <sup>H</sup>* (15)

<sup>65</sup> 0,212 1, *<sup>S</sup> <sup>u</sup> <sup>V</sup> <sup>H</sup>* (16)

The total gas flow (after taking into account the excess air coefficient λ) under normal

 *<sup>C</sup> <sup>T</sup> <sup>S</sup> <sup>V</sup>* 

Multiplying the value of the total flue gas stream *VC* by the amount of fuel burned per unit of time we can calculate the volumetric flow of flue gases from combustion process

*VN* – volumetric flow of flue gases from the combustion process converted to standard

where:

where:

[m3/kg]



conditions (T = 273K, P = 1013hPa) [m3/kg]

*N* – nitrogen content in the fuel [kg/kg],

conditions can be calculated from the formula:

converted to standard conditions:

where:

conditions [m3/h]

*B* – fuel consumption [kg/h],

the fuel from the formula (Recknagel et al., 1994):

account only the calorific value of fuel (Recknagel et al., 1994):

*<sup>V</sup> <sup>C</sup> <sup>H</sup> <sup>S</sup> <sup>O</sup> <sup>T</sup>* 8,88 26,44 3,32 3,33 (13)

*VS* 1,85*<sup>C</sup>* 11,11*<sup>H</sup>* 0,68*<sup>S</sup>* 0,8*<sup>N</sup>* 1,24*W* (14)

0,203 <sup>2</sup> ,0 *<sup>T</sup> <sup>u</sup> <sup>V</sup> <sup>H</sup>* (17)

*<sup>S</sup> <sup>u</sup> <sup>V</sup>* 0 (18) ,265*<sup>H</sup>*

1 *V V* (19)

*<sup>N</sup> <sup>C</sup> <sup>V</sup> <sup>B</sup>V* (20)

*T* – flue gas temperature [oC] *P* – flue gas pressure [hPa]

#### **6. Combustion as an autothermal process**

Heat emitted in a chemical reaction of oxidation in the combustion process is used for heating of flue gases from the process (combustion products and excess air supplied to the combustion process), solid secondary products of the combustion process (slag and ashes) and the equipment in which the combustion process is conducted (plus the obvious loss to the environment) up to the process temperature. This is illustrated in Figure 1.

Fig. 1. Simplified heat balance in the combustion chamber

It is obvious that, if the generated amount of heat is too small, the combustion process does not run properly, its temperature is too low and the flue gases do not reach the required temperature. If, however, fuel of high calorific value is burned the amount of heat is sufficient to ensure proper combustion temperature and temperature of flue gases. This is particularly important in the case of waste incineration, where the law (Directive 2000/76/EC on the incineration of waste) requires that the exhaust gases have a minimum temperature of 850°C and remain in this temperature for at least 2 seconds when the waste fuel contains less than 1% chlorine. When the chlorine content in waste exceeds 1% it is required to achieve a minimum temperature of 1100°C. If the required temperature of flue gas (and the combustion process) is reached in the device, in which the combustion process is run, it is called autothermal combustion process, but when due to too low calorific value of fuel the required temperatures cannot be reached, combustion is not an autothermal process and requires additional fuel of higher calorific value.

#### **7. Combustion as a set of chemical reactions**

As mentioned at the beginning from the chemical point of view combustion is the reaction of oxidation of hydrocarbons. Combustion of gaseous fuel is the simplest combustion, in

Pollutant Formation in Combustion Processes 303

much more complicated. In the initial period, at sufficiently high temperatures degassing processes (pyrolysis) begin and flammable gases are emitted from the solid fuel. They are relatively easily oxidized (burned) in the gas phase generating heat flux heating the solid phase and accelerating the release of flammable gases. These processes run as long as fuel in a given element does not run out of hydrogen. The solid phase in result of degassing is enriched in the process of carbonization in carbon hence degassing process (the release of flammable gases) is accompanied by solid phase carbonization. At this point a change in the mechanism of the process takes place and the process of gasification of solid fuel begins (using air and/or steam), which results in carbonization phase being converted into flammable gases (carbon monoxide and hydrogen), which then in the gas phase are oxidized (burning). In the gas phase the oxidation reaction dominates, however, at the same

Apart from that we should not forget about other reactions running in parallel such as Boudouard reaction in which carbon dioxide, previously formed during the combustion, at the temperature in the range 400-950°C can be reduced with the participation of carbon to

It is assumed that depending on the temperature in the combustion process of solid fuels

250°C – deoxidation reductive processes, reduction, decomposition of sulfuric acid esters, separation of bound moisture and carbon dioxide, depolymerization, the beginning of the release of hydrogen sulfide, 340°C – formation of aliphatic compounds, including unsaturated, the beginning

400°C – beginning of the formation of the compounds of carbon with oxygen and

600°C – cracking of bituminous substances in the direction of thermally persistent substances (gases, hydrocarbons of short chain structure), formation of new compounds (benzene derivatives in result of cyclization of

> formation, dehydration to butadiene, the formation of cyclohexane *C6H12*, thermal aromatization to benzene and other higher aromatic compounds.

of release of methane and other aliphatic compounds,

400–600°C – transformation of bituminuous substances into smouldering oil or tar,

> 600°C – further course of the hypothetical reaction of dimerization of butylene

However, we should be aware that the actual combustion process is far from ideal - that is complete and perfect combustion. The combustion process (thermal decomposition and

380°C – beginning of carbonization in the smouldering process,

unsaturated aliphatic compounds),

100–200°C – thermal drying process, separation of water (physical process),

*CO <sup>C</sup>* <sup>2</sup> *CO* <sup>2</sup> (26)

time a number of other reactions take place - such as synthesis, cyclization, etc.

less oxidized form - a flammable carbon monoxide according to the equation:

which can participate in further reactions e.g. oxidation.

nitrogen,

**8. Complete and uncompleted combustion** 

several important steps can be defined (Thome-Kozmiensky, 1994):

which simple reaction of hydrocarbons oxidation takes place according to the simplified reaction scheme:

$$\rm C\_xH\_y + \left(x + \frac{y}{4}\right)O\_2 \xrightarrow{\longrightarrow} x \rm CO\_2 + \frac{y}{2}H\_2O \tag{22}$$

$$\rm C\_xH\_y + \left(\frac{x}{2} + \frac{y}{4}\right)O\_2 \xrightarrow{\longrightarrow} x\ CO + \frac{y}{2}H\_2O \tag{23}$$

and the basic products of oxidation are carbon dioxide and water. In case of insufficient oxygenation of the combustion zone the additional reaction product is carbon monoxide. Flammable gas usually mixes well with air, causing a kinetic character of the combustion i.e. controlled by the rate of chemical oxidation reaction. In rare cases, when the flammable gas is mixed with air, the process can be controlled by diffusion of oxygen to the reaction zone. With inadequate oxygenation of the reaction zone the product of the combustion process is carbon monoxide, and elemental carbon (soot), because the reaction rate of water formation is higher than the rate of carbon oxidation.

Combustion of liquid fuels is a more complicated process, because to ensure the oxidation (combustion) it is necessary to evaporate the liquid, because its vapor is burned. Evaporation is promoted by the liquid spray, however, it is very difficult to obtain a homogeneous mixture of liquid vapor and air, hence the combustion of liquid fuels has rarely kinetic character, usually this is a diffusion combustion. Organic liquids generally have a more complex chemical structure than gases, and hence more often we are dealing with the formation of products of incomplete combustion, some of them being other organic compounds:

$$\text{C}\_{\text{x}}\text{H}\_{\text{y}} + \frac{3}{2}\text{O}\_{2} \xrightarrow[\text{-}\text{-}1]{} \text{C}\_{\text{x}-1}\text{H}\_{\text{y}-2} + \text{CO}\_{2} + \text{H}\_{2}\text{O}\tag{24}$$

of course simpler than the compound that is subject to burning.

In extreme cases there can be a situation when due to shortage of air in result of the combustion process the decomposition of organic compounds contained in the fuel will be incomplete and not all carbon will be oxidized. Consequently, in addition to gaseous products from the combustion process a solid product will be also formed - elemental carbon (*C*) or soot:

$$\text{C}\_{\text{x}}\text{H}\_{\text{y}} + \left(\text{x} - \frac{1}{2} + \frac{y}{4}\right)\text{O}\_{2} \xrightarrow{\longrightarrow} \text{C} + (\text{x} - 1)\text{CO}\_{2} + \frac{y}{2}\text{H}\_{2}\text{O}\tag{25}$$

The process of solid fuel combustion runs in even more complicated way. It is a multidirectional and multi-stage process, a combination of combustion (oxidation), gasification, thermal decomposition, including pyrolysis (lack of oxygen process). One should be aware that direct heterogeneous reaction between gaseous oxygen and solid hydrocarbon is problematic and because of the hetero phases it must be slow. Thus, direct combustion of solid phase fuel runs in a minimal degree. The real combustion process is

which simple reaction of hydrocarbons oxidation takes place according to the simplified

*<sup>H</sup> <sup>O</sup> <sup>y</sup> <sup>O</sup> <sup>x</sup> CO <sup>y</sup> <sup>C</sup> <sup>H</sup> <sup>x</sup> <sup>x</sup> <sup>y</sup>* <sup>2</sup> <sup>2</sup> <sup>2</sup> <sup>4</sup> <sup>2</sup> 

*<sup>H</sup> <sup>O</sup> <sup>y</sup> <sup>O</sup> <sup>x</sup> CO <sup>x</sup> <sup>y</sup> <sup>C</sup> Hx <sup>y</sup>* <sup>2</sup> <sup>2</sup> <sup>2</sup> <sup>4</sup> <sup>2</sup> 

and the basic products of oxidation are carbon dioxide and water. In case of insufficient oxygenation of the combustion zone the additional reaction product is carbon monoxide. Flammable gas usually mixes well with air, causing a kinetic character of the combustion i.e. controlled by the rate of chemical oxidation reaction. In rare cases, when the flammable gas is mixed with air, the process can be controlled by diffusion of oxygen to the reaction zone. With inadequate oxygenation of the reaction zone the product of the combustion process is carbon monoxide, and elemental carbon (soot), because the reaction rate of water formation

Combustion of liquid fuels is a more complicated process, because to ensure the oxidation (combustion) it is necessary to evaporate the liquid, because its vapor is burned. Evaporation is promoted by the liquid spray, however, it is very difficult to obtain a homogeneous mixture of liquid vapor and air, hence the combustion of liquid fuels has rarely kinetic character, usually this is a diffusion combustion. Organic liquids generally have a more complex chemical structure than gases, and hence more often we are dealing with the formation of products of incomplete combustion, some of them being other organic

*<sup>C</sup> <sup>H</sup> <sup>O</sup> <sup>C</sup> <sup>H</sup> CO <sup>H</sup> <sup>O</sup> <sup>x</sup> <sup>y</sup>* <sup>2</sup> *<sup>x</sup>* <sup>1</sup> *<sup>y</sup>* <sup>2</sup> <sup>2</sup> <sup>2</sup> <sup>2</sup>

In extreme cases there can be a situation when due to shortage of air in result of the combustion process the decomposition of organic compounds contained in the fuel will be incomplete and not all carbon will be oxidized. Consequently, in addition to gaseous products from the combustion process a solid product will be also formed - elemental

*<sup>H</sup> <sup>O</sup> <sup>y</sup> <sup>O</sup> <sup>C</sup> <sup>x</sup> CO <sup>y</sup> <sup>C</sup> <sup>H</sup> <sup>x</sup> <sup>x</sup> <sup>y</sup>* <sup>2</sup> <sup>2</sup> <sup>2</sup> <sup>2</sup>

The process of solid fuel combustion runs in even more complicated way. It is a multidirectional and multi-stage process, a combination of combustion (oxidation), gasification, thermal decomposition, including pyrolysis (lack of oxygen process). One should be aware that direct heterogeneous reaction between gaseous oxygen and solid hydrocarbon is problematic and because of the hetero phases it must be slow. Thus, direct combustion of solid phase fuel runs in a minimal degree. The real combustion process is

<sup>1</sup>

1

(25)

(22)

(23)

(24)

 

> 

 

> 

3

2 4

   

of course simpler than the compound that is subject to burning.

is higher than the rate of carbon oxidation.

reaction scheme:

compounds:

carbon (*C*) or soot:

much more complicated. In the initial period, at sufficiently high temperatures degassing processes (pyrolysis) begin and flammable gases are emitted from the solid fuel. They are relatively easily oxidized (burned) in the gas phase generating heat flux heating the solid phase and accelerating the release of flammable gases. These processes run as long as fuel in a given element does not run out of hydrogen. The solid phase in result of degassing is enriched in the process of carbonization in carbon hence degassing process (the release of flammable gases) is accompanied by solid phase carbonization. At this point a change in the mechanism of the process takes place and the process of gasification of solid fuel begins (using air and/or steam), which results in carbonization phase being converted into flammable gases (carbon monoxide and hydrogen), which then in the gas phase are oxidized (burning). In the gas phase the oxidation reaction dominates, however, at the same time a number of other reactions take place - such as synthesis, cyclization, etc.

Apart from that we should not forget about other reactions running in parallel such as Boudouard reaction in which carbon dioxide, previously formed during the combustion, at the temperature in the range 400-950°C can be reduced with the participation of carbon to less oxidized form - a flammable carbon monoxide according to the equation:

$$\text{CO}\_2 + \text{C} \xrightarrow{+} \text{2CO} \tag{26}$$

which can participate in further reactions e.g. oxidation.

It is assumed that depending on the temperature in the combustion process of solid fuels several important steps can be defined (Thome-Kozmiensky, 1994):

	- 250°C deoxidation reductive processes, reduction, decomposition of sulfuric acid esters, separation of bound moisture and carbon dioxide, depolymerization, the beginning of the release of hydrogen sulfide,
	- 340°C formation of aliphatic compounds, including unsaturated, the beginning of release of methane and other aliphatic compounds,
	- 380°C beginning of carbonization in the smouldering process,
	- 400°C beginning of the formation of the compounds of carbon with oxygen and nitrogen,
	- 600°C cracking of bituminous substances in the direction of thermally persistent substances (gases, hydrocarbons of short chain structure), formation of new compounds (benzene derivatives in result of cyclization of unsaturated aliphatic compounds),
	- > 600°C further course of the hypothetical reaction of dimerization of butylene formation, dehydration to butadiene, the formation of cyclohexane *C6H12*, thermal aromatization to benzene and other higher aromatic compounds.

#### **8. Complete and uncompleted combustion**

However, we should be aware that the actual combustion process is far from ideal - that is complete and perfect combustion. The combustion process (thermal decomposition and

Pollutant Formation in Combustion Processes 305

On the third pathway carbonized solid phase is gasified with water and/or air to carbon monoxide and hydrogen which in the gas phase are easily oxidized to carbon dioxide and water. The last fourth pathway of the course of combustion processes is a secondary synthesis path. In result of decomposition and gasification processes in the gas phase simple hydrocarbons and hydrocarbon radicals are formed. Since virtually all fuels contain trace admixtures of chlorine (carbon even sometimes even about 1%!) in the gas phase there are also simple chlorinated hydrocarbons (including unsaturated - for example, acetylene *C2H6*) and chlorinated hydrocarbon radicals. They are extremely reactive and in the gas phase appear a number of synthesis reactions, including cyclization. In this way chlorinated and non-chlorinated aliphatic hydrocarbons are formed (methane, ethane, chloromethanes, chloroethanes etc., aldehydes: e.g. formaldehyde and acetaldehyde, alcohols: e.g. methanol, simple carboxylic acids: e.g. formic acid and acetic acid) and aromatic hydrocarbons (benzenes, chlorobenzenes, phenols and chlorophenols, naphthalenes and many others). The latter take part in the synthesis of polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), polychlorinated biphenyls (PCBs), polychlorinated naphthalenes (PCNs) and polycyclic aromatic hydrocarbons (PAHs) (Wielgosiński, 2009).

Practically all fuels solid, liquid and gas contain some amounts of sulphur (Table 3).

Gas fuels: LNG, CNG, LPG below 0,01% Diesel oil below 0,10% Petrol below 0,05% Heating oil 0,2 - 0,3% Wood 0,02 - 0,04% Coal (Poland) 0,5 - 2,5% Coal (World) 0,1 - 4,5% Lignite (Poland) 0,4 - 1,2% Lignite (World) 0,4 - 6,5% Coke 0,2 - 1,0% Municipal waste 0,2 - 0,6%

Sulphur in the combustion process is relatively easily oxidized to sulphur dioxide (*SO2*) in

<sup>2</sup> <sup>2</sup> *S O SO combustion*

The rate of this reaction at temperatures, under which the combustion process is run, is very high. On the basis of many research results it can be assumed that approximately 90-95% of the sulfur contained in fuel in the combustion process will be oxidized to *SO2*. The presence

(27)

Kind of fuel Sulphur content

**9. Formation of sulphur oxides** 

Table 3. Sulphur content in different fuels

accordance with the simplified reaction scheme.

oxidation) of many organic compounds (in particular those contained in the waste) is not a perfect process that produces only carbon dioxide, carbon monoxide and water. In this process usually a large number of intermediate products of decomposition and oxidation are formed which then are not further decomposed. It would seem that in the drastic combustion conditions, at about 1000°C any organic material must be burned. Unfortunately this is not true. Many chemical compounds, often flammable, are not subject to complete destruction during process of combustion. Some of the organic compounds are generated in result of many secondary reactions running in the combustion zone and outside it. Simplified diagram of the formation of pollutants in combustion processes is shown in Figure 2.

Fig. 2. Simplified diagram of the formation of pollutants in the combustion process of solid flammable substance (fuel, waste, etc.)

In the process of fuel combustion there are four main pathways. The first pathway - direct oxidation reactions. If we treat fuel as a collection of various chemicals consisting of carbon, hydrogen, oxygen, nitrogen, sulfur, chlorine, etc. so in result of direct oxidation typical flammable gases will appear such as carbon dioxide (*CO2*), water (*H2O*) in accordance with reaction equation (22) and also sulfur dioxide (*SO2*), nitrogen oxides (*NO*, *NO2*, *N2O*) or hydrogen chloride (*HCl*). At the same time the combustion process will be conducted according to other two pathways. In the process of thermal decomposition flammable gases will be emitted that contain for example carbon monoxide (*CO*), methane (*CH4*) and other simple hydrocarbons, which are relatively easily oxidized to carbon dioxide and water in the gas phase. Thermal decomposition process impoverishes the solid phase of fuel in hydrogen and other volatile substances, causing its carbonization.

oxidation) of many organic compounds (in particular those contained in the waste) is not a perfect process that produces only carbon dioxide, carbon monoxide and water. In this process usually a large number of intermediate products of decomposition and oxidation are formed which then are not further decomposed. It would seem that in the drastic combustion conditions, at about 1000°C any organic material must be burned. Unfortunately this is not true. Many chemical compounds, often flammable, are not subject to complete destruction during process of combustion. Some of the organic compounds are generated in result of many secondary reactions running in the combustion zone and outside it. Simplified diagram of the formation of pollutants in combustion processes is

Fig. 2. Simplified diagram of the formation of pollutants in the combustion process of solid

In the process of fuel combustion there are four main pathways. The first pathway - direct oxidation reactions. If we treat fuel as a collection of various chemicals consisting of carbon, hydrogen, oxygen, nitrogen, sulfur, chlorine, etc. so in result of direct oxidation typical flammable gases will appear such as carbon dioxide (*CO2*), water (*H2O*) in accordance with reaction equation (22) and also sulfur dioxide (*SO2*), nitrogen oxides (*NO*, *NO2*, *N2O*) or hydrogen chloride (*HCl*). At the same time the combustion process will be conducted according to other two pathways. In the process of thermal decomposition flammable gases will be emitted that contain for example carbon monoxide (*CO*), methane (*CH4*) and other simple hydrocarbons, which are relatively easily oxidized to carbon dioxide and water in the gas phase. Thermal decomposition process impoverishes the solid phase of fuel in

hydrogen and other volatile substances, causing its carbonization.

shown in Figure 2.

flammable substance (fuel, waste, etc.)

On the third pathway carbonized solid phase is gasified with water and/or air to carbon monoxide and hydrogen which in the gas phase are easily oxidized to carbon dioxide and water. The last fourth pathway of the course of combustion processes is a secondary synthesis path. In result of decomposition and gasification processes in the gas phase simple hydrocarbons and hydrocarbon radicals are formed. Since virtually all fuels contain trace admixtures of chlorine (carbon even sometimes even about 1%!) in the gas phase there are also simple chlorinated hydrocarbons (including unsaturated - for example, acetylene *C2H6*) and chlorinated hydrocarbon radicals. They are extremely reactive and in the gas phase appear a number of synthesis reactions, including cyclization. In this way chlorinated and non-chlorinated aliphatic hydrocarbons are formed (methane, ethane, chloromethanes, chloroethanes etc., aldehydes: e.g. formaldehyde and acetaldehyde, alcohols: e.g. methanol, simple carboxylic acids: e.g. formic acid and acetic acid) and aromatic hydrocarbons (benzenes, chlorobenzenes, phenols and chlorophenols, naphthalenes and many others). The latter take part in the synthesis of polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), polychlorinated biphenyls (PCBs), polychlorinated naphthalenes (PCNs) and polycyclic aromatic hydrocarbons (PAHs) (Wielgosiński, 2009).

#### **9. Formation of sulphur oxides**

Practically all fuels solid, liquid and gas contain some amounts of sulphur (Table 3).


Table 3. Sulphur content in different fuels

Sulphur in the combustion process is relatively easily oxidized to sulphur dioxide (*SO2*) in accordance with the simplified reaction scheme.

$$S + O\_2 \xrightarrow[\text{conclusion}]{} SO\_2 \tag{27}$$

The rate of this reaction at temperatures, under which the combustion process is run, is very high. On the basis of many research results it can be assumed that approximately 90-95% of the sulfur contained in fuel in the combustion process will be oxidized to *SO2*. The presence

Pollutant Formation in Combustion Processes 307

At high temperature molecular oxygen is decomposed into highly active form of atomic oxygen. Atomic oxygen molecules attack the nitrogen molecules and nitric oxide is formed, while formed in this reaction active nitrogen atoms, by attacking oxygen molecules, also form nitric oxide and active atomic oxygen. So it is a classic example of chain reaction, for which the first reaction (30) is the stage of initiation while the next two reactions (31) and

in fuel-rich mixtures where the concentration of hydroxyl radicals is significant, greater than the concentration of hydrogen and oxygen atoms ( *OHOH* ) the following reaction can

An additional source of nitric oxide formation may be the following reactions in accordance

The rates of formation of nitrogen oxides in the thermal mechanism are relatively high but only at high temperatures. This mechanism becomes negligible at temperatures above

**Fuel mechanism** is directly related to nitrogen content in fuels. As shown in Table 1 most solid and liquid fuels contain nitrogen. So the source of nitrogen in this mechanism is fuel while the source of oxygen is air introduced to the combustion process. The formation of nitric oxide in this mechanism is quite long and it goes through a number of succeeding-

parallel reactions (Bowman et al., 1982, Miller & Bowman, 1989) shown in Figure 3.

Fig. 3. Diagram of nitric acid formation according to fuel mechanism

(32) are the propagation stage:

1400°C.

be considered as the last stage of termination:

with the mechanism described by Bozzelli (Bozzelli et at., 1994):

*MOOMO*2 (31)

*NO NNO* <sup>2</sup> (32)

*ON* <sup>2</sup> *ONO* (33)

*OHN HNO* (34)

*HNNH* <sup>2</sup> <sup>2</sup> (35)

*NHNOOHN*2 (36)

in the burned material of the non-combustible fraction containing some metals (e.g. vanadium - *V*) will result in partial catalytic oxidation of the formed sulfur dioxide to sulfur trioxide (SO3) in accordance with reaction equation:

$$SO\_2 + \frac{1}{2}O\_2 \xrightarrow[\text{catalyst-}-e.g.\,\text{nomainum}]{} SO\_3 \tag{28}$$

Coal as a chemical element, like hydrogen, has reductive properties. In high temperature accompanying the combustion process, in the conditions of limited access of oxygen, in the presence of hydrogen in the gasification reaction, sulfur contained in the fuel can be reduced to hydrogen sulphide according to reaction equation:

$$S + H\_2 \xrightarrow[\text{carbon}]{} H\_2S \tag{29}$$

Hydrogen sulphide formed during the reaction (30) is a flammable gas and at a later stage may be oxidized to sulphur dioxide and water according to reaction equation:

$${}^{H}\_{2}\text{S} + \frac{3}{2} {}^{O}\_{2}\text{O} \xrightarrow{\longrightarrow} \text{SO}\_{2} + {}^{H}\_{2}\text{O} \tag{30}$$

It should be clearly stated that in the flue gases from combustion process both *SO3* and *H2S* may be present in trace amounts (well below 1% of the total emissions of sulfur compounds) and sulphur dioxide - *SO2* will be the main pollutant that is emitted. It is the so called raw material pollutant whose quantity in the exhaust almost in 100% corresponds to the amount of sulphur introduced to the combustion process. This allows relatively accurate calculation of sulphur emissions from combustion processes.

#### **10. Formation of nitrogen oxides**

The studies of the combustion processes lead to the conclusion that the formation of nitrogen oxides (*NOx*) is observed for many fuels. This applies primarily to nitric oxide (*NO*), nitrogen dioxide (*NO2*) and nitrous oxide (*N2O*). Of course there are known many other chemical compounds which are a combination of oxygen and nitrogen, however, in the flue gas from combustion they are in fact absent. The basic nitrogen oxide formed in the combustion process is nitrogen monoxide - *NO*. The extensive literature on the subject gives the information about three mechanisms of the formation of this compound in thermal processes. They are:


**Thermal mechanism** was discovered and described for the first time by Zeldovich in the forties of the last century (Zeldovich, 1946). It includes the direct synthesis of nitric oxide from oxygen and nitrogen contained in air at high temperature. In the first stage there is decomposition of molecular oxygen to two molecules of active atomic oxygen in result of contact with high-energy inert molecule *M*, acting as a catalyst (it may be e.g. a molecule of hot metallic wall of the combustion chamber, burner, etc.):

in the burned material of the non-combustible fraction containing some metals (e.g. vanadium - *V*) will result in partial catalytic oxidation of the formed sulfur dioxide to sulfur

> <sup>3</sup> . . <sup>2</sup> <sup>2</sup> <sup>2</sup> <sup>1</sup> *SO <sup>O</sup> SO*

Coal as a chemical element, like hydrogen, has reductive properties. In high temperature accompanying the combustion process, in the conditions of limited access of oxygen, in the presence of hydrogen in the gasification reaction, sulfur contained in the fuel can be reduced

*S H H S*

Hydrogen sulphide formed during the reaction (30) is a flammable gas and at a later stage

*<sup>H</sup> <sup>S</sup> <sup>O</sup> SO <sup>H</sup> <sup>O</sup>* <sup>2</sup> <sup>2</sup> <sup>2</sup> <sup>2</sup> <sup>2</sup>

It should be clearly stated that in the flue gases from combustion process both *SO3* and *H2S* may be present in trace amounts (well below 1% of the total emissions of sulfur compounds) and sulphur dioxide - *SO2* will be the main pollutant that is emitted. It is the so called raw material pollutant whose quantity in the exhaust almost in 100% corresponds to the amount of sulphur introduced to the combustion process. This allows relatively accurate calculation

The studies of the combustion processes lead to the conclusion that the formation of nitrogen oxides (*NOx*) is observed for many fuels. This applies primarily to nitric oxide (*NO*), nitrogen dioxide (*NO2*) and nitrous oxide (*N2O*). Of course there are known many other chemical compounds which are a combination of oxygen and nitrogen, however, in the flue gas from combustion they are in fact absent. The basic nitrogen oxide formed in the combustion process is nitrogen monoxide - *NO*. The extensive literature on the subject gives the information about three mechanisms of the formation of this compound in thermal

**Thermal mechanism** was discovered and described for the first time by Zeldovich in the forties of the last century (Zeldovich, 1946). It includes the direct synthesis of nitric oxide from oxygen and nitrogen contained in air at high temperature. In the first stage there is decomposition of molecular oxygen to two molecules of active atomic oxygen in result of contact with high-energy inert molecule *M*, acting as a catalyst (it may be e.g. a molecule of

may be oxidized to sulphur dioxide and water according to reaction equation:

3

*catalyst e g vanadium*

(28)

*carbon* <sup>2</sup> <sup>2</sup> (29)

(30)

trioxide (SO3) in accordance with reaction equation:

to hydrogen sulphide according to reaction equation:

of sulphur emissions from combustion processes.

**10. Formation of nitrogen oxides** 



hot metallic wall of the combustion chamber, burner, etc.):

processes. They are:


$$O\_2 + M \xrightarrow{\quad} O + O + M \tag{31}$$

At high temperature molecular oxygen is decomposed into highly active form of atomic oxygen. Atomic oxygen molecules attack the nitrogen molecules and nitric oxide is formed, while formed in this reaction active nitrogen atoms, by attacking oxygen molecules, also form nitric oxide and active atomic oxygen. So it is a classic example of chain reaction, for which the first reaction (30) is the stage of initiation while the next two reactions (31) and (32) are the propagation stage:

$$O + N\_2 \xrightarrow{\longrightarrow} NO + N \tag{32}$$

$$N + O\_2 \longrightarrow NO + O\tag{33}$$

in fuel-rich mixtures where the concentration of hydroxyl radicals is significant, greater than the concentration of hydrogen and oxygen atoms ( *OHOH* ) the following reaction can be considered as the last stage of termination:

$$N + OH \xrightarrow{\longrightarrow} NO + H \tag{34}$$

An additional source of nitric oxide formation may be the following reactions in accordance with the mechanism described by Bozzelli (Bozzelli et at., 1994):

$$H + N\_2 \longleftrightarrow N\_2H \tag{35}$$

$$N\_2H + O \longleftrightarrow NO + NH \tag{36}$$

The rates of formation of nitrogen oxides in the thermal mechanism are relatively high but only at high temperatures. This mechanism becomes negligible at temperatures above 1400°C.

**Fuel mechanism** is directly related to nitrogen content in fuels. As shown in Table 1 most solid and liquid fuels contain nitrogen. So the source of nitrogen in this mechanism is fuel while the source of oxygen is air introduced to the combustion process. The formation of nitric oxide in this mechanism is quite long and it goes through a number of succeedingparallel reactions (Bowman et al., 1982, Miller & Bowman, 1989) shown in Figure 3.

Fig. 3. Diagram of nitric acid formation according to fuel mechanism

Pollutant Formation in Combustion Processes 309

Reactions of nitric oxide formation from nitrous oxide proceed relatively quickly at temperatures below 1200°C in the area of good oxygenation of the combustion zone. The nitrous oxide synthesis itself proceeds according to several possible paths, one of them requires, just as in the thermal mechanism, the contact with the high-energy inert molecule *M*:

The reaction (45), (46) and (47) shows that the synthesis of *N2O* can take place outside the combustion zone with the participation of nitric oxide previously formed in the presence of

The studies of combustion products show above all the presence of nitric oxide (*NO*) and small quantities of nitrous oxide (*N2O*). Nitrogen dioxide (*NO2*), from the oxidation of *NO* formed during combustion, is a component of flue gas as well. The formation of *NO2* has been described by Miller and Bowman in 1989 (Miller & Bowman, 1989). As a result of diffusion of hydrogen radicals beyond flame zone to the area of lower temperatures (below 750°C) in the presence of excess oxygen there are formed, involving high-energy inert particles of *M*, *HO2* radicals which also in the same zone can react with nitric oxide

In parallel to the synthesis reaction, nitrogen dioxide decomposition reactions take place:

Moreover, in the further zone of the installation, after heat recovery systems, at temperatures below 200°C occurs direct oxidation of *NO* to *NO2* by oxygen present in the

<sup>2</sup> <sup>2</sup> <sup>2</sup>

Under proper combustion conditions, the participation of *NO2* in the whole stream of

numerous radicals.

according to the reaction:

emitted nitrogen oxides does not exceed 5-10%.

flue gas:

*<sup>N</sup> <sup>O</sup> CO NO NCO* <sup>2</sup> (43)

*<sup>O</sup> <sup>N</sup>*<sup>2</sup> *<sup>M</sup> N*2*<sup>O</sup> <sup>M</sup>* (44)

*NCO NO <sup>N</sup> <sup>O</sup> CO* <sup>2</sup> (45)

*NH NO <sup>N</sup> <sup>O</sup> <sup>H</sup>* <sup>2</sup> (46)

*NO<sup>N</sup> <sup>O</sup> <sup>O</sup>* <sup>2</sup> 2 (47)

*NO HO NO OH* <sup>2</sup> <sup>2</sup> (48)

<sup>2</sup> <sup>2</sup> *NO <sup>O</sup> NO O* (49)

*NO <sup>H</sup> NO OH* <sup>2</sup> (50)

<sup>1</sup> *NO <sup>O</sup> NO* (51)

In the first stage from organic matter, subject to combustion process (hydrocarbons containing nitrogen), hydrogen cyanide is released which by further oxidation and hydrogenation gives first radical *NCO,* next *NH* and then free atomic nitrogen (*N*). At this point there are three possible further courses of reactions - oxidation reaction in which nitric oxide is formed, the reaction with previously formed nitric oxide in result of which neutral molecular nitrogen is formed and the reaction with hydrocarbon radical returning the reaction to the beginning. Thus, only one course leads to the formation of nitric oxide while the other two effectively reduce its quantity - hence you can see that there are technical possibilities of influencing on the amount of produced nitrogen oxides - through the organization of the combustion process.

**Prompt mechanism** was first described by Fenimore in 1971 (Fenimore, 1971). He noted that in the early stage of flame, in the presence of numerous hydrocarbon radicals, occurs a synthesis of nitric oxide from oxygen and nitrogen introduced into the combustion process. Since *NO* formation reactions occur in the early stage of flame this mechanism was called the prompt mechanism. The prompt mechanism is complementary to the fuel mechanism and some chemical reactions follow the same path. Generally it can be assumed that the following reactions proceed in this mechanism:

$$CH + N\text{-}2 \xrightarrow{\quad} H\text{CN} + N\tag{37}$$

$$\text{CH}\_2 + \text{N}\_2 \xrightarrow{\text{---}} \text{HCN} + \text{NH} \tag{38}$$

$$\text{HCN} + \text{O} \xrightarrow{\text{H}} \text{NCO} + \text{H} \tag{39}$$

$$\text{NCO} + \text{O} \xrightarrow{\text{-}} \text{NO} + \text{CO} \tag{40}$$

The rate of formation of nitric oxide in this mechanism is very high, but the amount of formed *NO* according to this mechanism is relatively small.

Generally, during combustion of fuels, especially solid fuels, the greatest amount of nitrogen oxide is produced by the fuel mechanism (sometimes even 80-90%). Thermal mechanism begins to play a role only after exceeding the temperature of 1400°C, so the contribution of this mechanism usually does not exceed 10-20%. The contribution of the prompt mechanism is usually at the level of 1-5%. Of course the contribution of thermal mechanism increases with the increase of temperature.

In recent years, it was noted that during low-temperature combustion of fuel-poor mixtures with low excess air there appears an additional mechanism of nitric oxide formation. It is connected with the described by Malte and co-workers (Steele et al., 1995) synthesis of nitrous oxide in the combustion conditions. Most likely, this reaction occurs with the participation of active atoms of hydrogen, oxygen or carbon monoxide - a product of incomplete combustion according to the following equations:

$$N\_2O + H \xrightarrow{\cdot \cdot \cdot} NO + NH \tag{41}$$

$$N\_2O + O \xrightarrow{} 2 \text{ NO} \tag{42}$$

In the first stage from organic matter, subject to combustion process (hydrocarbons containing nitrogen), hydrogen cyanide is released which by further oxidation and hydrogenation gives first radical *NCO,* next *NH* and then free atomic nitrogen (*N*). At this point there are three possible further courses of reactions - oxidation reaction in which nitric oxide is formed, the reaction with previously formed nitric oxide in result of which neutral molecular nitrogen is formed and the reaction with hydrocarbon radical returning the reaction to the beginning. Thus, only one course leads to the formation of nitric oxide while the other two effectively reduce its quantity - hence you can see that there are technical possibilities of influencing on the amount of produced nitrogen oxides - through the

**Prompt mechanism** was first described by Fenimore in 1971 (Fenimore, 1971). He noted that in the early stage of flame, in the presence of numerous hydrocarbon radicals, occurs a synthesis of nitric oxide from oxygen and nitrogen introduced into the combustion process. Since *NO* formation reactions occur in the early stage of flame this mechanism was called the prompt mechanism. The prompt mechanism is complementary to the fuel mechanism and some chemical reactions follow the same path. Generally it can be assumed that the

The rate of formation of nitric oxide in this mechanism is very high, but the amount of

Generally, during combustion of fuels, especially solid fuels, the greatest amount of nitrogen oxide is produced by the fuel mechanism (sometimes even 80-90%). Thermal mechanism begins to play a role only after exceeding the temperature of 1400°C, so the contribution of this mechanism usually does not exceed 10-20%. The contribution of the prompt mechanism is usually at the level of 1-5%. Of course the contribution of thermal mechanism increases

In recent years, it was noted that during low-temperature combustion of fuel-poor mixtures with low excess air there appears an additional mechanism of nitric oxide formation. It is connected with the described by Malte and co-workers (Steele et al., 1995) synthesis of nitrous oxide in the combustion conditions. Most likely, this reaction occurs with the participation of active atoms of hydrogen, oxygen or carbon monoxide - a product of

*CH N* 2 *HCN N* (37)

*CH <sup>N</sup> HCN NH* <sup>2</sup> <sup>2</sup> (38)

*HCN O NCO H* (39)

*NCO O NO CO* (40)

*<sup>N</sup> <sup>O</sup> <sup>H</sup> NO NH* <sup>2</sup> (41)

*<sup>N</sup> <sup>O</sup> <sup>O</sup>* <sup>2</sup> *NO* <sup>2</sup> (42)

organization of the combustion process.

following reactions proceed in this mechanism:

formed *NO* according to this mechanism is relatively small.

incomplete combustion according to the following equations:

with the increase of temperature.

$$N\_2O + CO \xrightarrow{-} NO + NCO \tag{43}$$

Reactions of nitric oxide formation from nitrous oxide proceed relatively quickly at temperatures below 1200°C in the area of good oxygenation of the combustion zone. The nitrous oxide synthesis itself proceeds according to several possible paths, one of them requires, just as in the thermal mechanism, the contact with the high-energy inert molecule *M*:

$$O + N\_2 + M \xrightarrow{-} \xrightarrow{-} N\_2O + M \tag{44}$$

$$\text{NCO} + \text{NO} \xrightarrow{\text{-}} \text{N}\_2\text{O} + \text{CO} \tag{45}$$

$$NH + NO \xrightarrow{\longrightarrow} N\_2O + H \tag{46}$$

$$2\,\text{NO} \longleftrightarrow \text{N}\_2\text{O} + \text{O} \tag{47}$$

The reaction (45), (46) and (47) shows that the synthesis of *N2O* can take place outside the combustion zone with the participation of nitric oxide previously formed in the presence of numerous radicals.

The studies of combustion products show above all the presence of nitric oxide (*NO*) and small quantities of nitrous oxide (*N2O*). Nitrogen dioxide (*NO2*), from the oxidation of *NO* formed during combustion, is a component of flue gas as well. The formation of *NO2* has been described by Miller and Bowman in 1989 (Miller & Bowman, 1989). As a result of diffusion of hydrogen radicals beyond flame zone to the area of lower temperatures (below 750°C) in the presence of excess oxygen there are formed, involving high-energy inert particles of *M*, *HO2* radicals which also in the same zone can react with nitric oxide according to the reaction:

$$\text{NO} + \text{HO}\_2 \xrightarrow{\text{-}} \text{NO}\_2 + \text{OH} \tag{48}$$

In parallel to the synthesis reaction, nitrogen dioxide decomposition reactions take place:

$$NO\_2 + O \xrightarrow{-} NO + O\_2 \tag{49}$$

$$NO\_2 + H \xrightarrow{-} NO + OH \tag{50}$$

Moreover, in the further zone of the installation, after heat recovery systems, at temperatures below 200°C occurs direct oxidation of *NO* to *NO2* by oxygen present in the flue gas:

$$\stackrel{1}{2}\stackrel{NO}{\underset{2}{\longrightarrow}}\stackrel{1}{O}\_{2} \longrightarrow \stackrel{NO}{\longrightarrow} \stackrel{0}{O}\_{2}\tag{51}$$

Under proper combustion conditions, the participation of *NO2* in the whole stream of emitted nitrogen oxides does not exceed 5-10%.

Pollutant Formation in Combustion Processes 311

Antimony - Sb 25 75 Arsenic - As 68,6 31,4 Chromium - Cr 92 8 Tin - Sn 50,8 49,2 Zinc - Zn 52 48 Aluminium - Al 87,8 12,2 Cadmium - Cd 9,5 90,5 Cobalt - Co 90,1 9,9 Magnesium - Mg 91,9 8,1 Manganese - Mn 94 6 Copper - Cu 97,3 2,7 Molybdenum - Mo 91,7 8,3 Nickel - Ni 98,2 1,8 Lead - Pb 59 41 Mercury - Hg 0,7 99,3 Titanium - Ti 85,3 14,7 Iron - Fe 99,1 0,9

Table 4. Speciation of the selected metals in the combustion process

included in Table 5.

absence and in 10% presence of chlorine.

The situation is substantially changed, if in the burned area appear halides - especially chlorine and bromine. Melting point and boiling point of heavy metal salts (chlorides) is significantly lower than the melting point and boiling point of pure metal or its oxide.

According to Barton (Barton et al., 1991) and Niessen (Niessen, 2002) it is particularly evident in the case of nickel (*Ni*), thallium (*Tl*) and lead (*Pb*). This is illustrated by the data

0% Cl 10% Cl

Metal Volatility temperature [oC]

Antimony - Sb 660 660 Arsenic - As 32 32 Chromium - Cr 1613 1610 Cadmium - Cd 214 214 Nickel - Ni 1210 693 Lead - Pb 627 -15 Mercury - Hg 14 14 Selenium - Se 318 318 Thallium - Tl 721 138 Table 5. Volatility temperature (reaching resilience 0,1 Pa) for the selected metals in the

Metal Solid phase [%] Gas phase [%]

Nitrogen oxides (*NOx*) understood as the sum of *NO, NO2* and *N2O* are pollutants whose rate of emission is minimally dependent on nitrogen content in the fuel. Their emissions is a direct result of plant construction and organization of the combustion process.

#### **11. Particulate matter emission**

The problem of particulate matter emissions practically applies only to solid fuel combustion process. It occurs in a minimal degree during the combustion of liquid fuels and practically does not occur during the combustion of gaseous fuels. The presence in the burned material of inorganic solid substance - non-flammable, most often called ash makes that in the combustion process solid residue in the form of slag and ash is formed. The flow of air through the combustion zone results in entrainment of fine solid particles and thus dust emission from combustion processes is formed. This emission is the higher, the better is the oxygenation of the combustion zone and the higher are air velocities observed in this zone. Solid particles, lifted in the flue gas stream from the combustion zone, have usually very small diameter and their main component is silicon dioxide (*SiO2* - silica). At the same time these particles are carriers of metals, elemental carbon (soot) as well as the adsorbed products of incomplete combustion and products of secondary synthesis outside the flame zone. In modern solid fuel combustion installations the release of solid particles is usually not more than 60% for the pulverized fuel and fluidal installations, and about 25-40% for the grate installations. The emission of solid particles, as dust from the combustion process is proportional to the content of non-flammable substances in the fuel and the factor of proportionality, the so called release factor, which is different for each installation and depends on its structure. So it depends partly on raw materials characteristics.

#### **12. Metals emission**

The emission of metals from the combustion process is very closely linked with the emission of solid particles (particulate matter) and therefore this problem does not exist during the combustion of gaseous fuels; it occurs to a small extent during combustion of liquid fuels and it applies fully to the combustion of solid fuels containing significant admixtures of non-flammable substance. Metals (mainly heavy metals), due to the speciation in the products of combustion, can be divided roughly into three groups:


The first group includes above all mercury (*Hg*) and cadmium (*Cd*). Similar properties have selenium (*Se*). Most metals belong to the second group – they are mainly iron (*Fe*), magnesium (*Mg*), cobalt (*Co*), chromium (*Cr*), copper (*Cu*), manganese (*Mn*), molybdenum (*Mo*) and nickel (*Ni*). Similar properties also have aluminum (*Al*) and titanium (*Ti*). The third intermediate group, includes mainly arsenic (*As*), lead (*Pb*), tin (*Sn*) and zinc (*Zn*) as well as antimony (*Sb*). The detailed data on the speciation of the selected metals in the combustion process, based on the example of waste incineration plants (Belevi & Moench, 2000, Belevi & Langmaister 2000, Sukrut et al., 2002), are given in Table 4.

Nitrogen oxides (*NOx*) understood as the sum of *NO, NO2* and *N2O* are pollutants whose rate of emission is minimally dependent on nitrogen content in the fuel. Their emissions is a

The problem of particulate matter emissions practically applies only to solid fuel combustion process. It occurs in a minimal degree during the combustion of liquid fuels and practically does not occur during the combustion of gaseous fuels. The presence in the burned material of inorganic solid substance - non-flammable, most often called ash makes that in the combustion process solid residue in the form of slag and ash is formed. The flow of air through the combustion zone results in entrainment of fine solid particles and thus dust emission from combustion processes is formed. This emission is the higher, the better is the oxygenation of the combustion zone and the higher are air velocities observed in this zone. Solid particles, lifted in the flue gas stream from the combustion zone, have usually very small diameter and their main component is silicon dioxide (*SiO2* - silica). At the same time these particles are carriers of metals, elemental carbon (soot) as well as the adsorbed products of incomplete combustion and products of secondary synthesis outside the flame zone. In modern solid fuel combustion installations the release of solid particles is usually not more than 60% for the pulverized fuel and fluidal installations, and about 25-40% for the grate installations. The emission of solid particles, as dust from the combustion process is proportional to the content of non-flammable substances in the fuel and the factor of proportionality, the so called release factor, which is different for each installation and

direct result of plant construction and organization of the combustion process.

depends on its structure. So it depends partly on raw materials characteristics.

products of combustion, can be divided roughly into three groups:

Langmaister 2000, Sukrut et al., 2002), are given in Table 4.

The emission of metals from the combustion process is very closely linked with the emission of solid particles (particulate matter) and therefore this problem does not exist during the combustion of gaseous fuels; it occurs to a small extent during combustion of liquid fuels and it applies fully to the combustion of solid fuels containing significant admixtures of non-flammable substance. Metals (mainly heavy metals), due to the speciation in the



The first group includes above all mercury (*Hg*) and cadmium (*Cd*). Similar properties have selenium (*Se*). Most metals belong to the second group – they are mainly iron (*Fe*), magnesium (*Mg*), cobalt (*Co*), chromium (*Cr*), copper (*Cu*), manganese (*Mn*), molybdenum (*Mo*) and nickel (*Ni*). Similar properties also have aluminum (*Al*) and titanium (*Ti*). The third intermediate group, includes mainly arsenic (*As*), lead (*Pb*), tin (*Sn*) and zinc (*Zn*) as well as antimony (*Sb*). The detailed data on the speciation of the selected metals in the combustion process, based on the example of waste incineration plants (Belevi & Moench, 2000, Belevi &

**11. Particulate matter emission** 

**12. Metals emission** 

products (slag and grate ash)

combustion products.


Table 4. Speciation of the selected metals in the combustion process

The situation is substantially changed, if in the burned area appear halides - especially chlorine and bromine. Melting point and boiling point of heavy metal salts (chlorides) is significantly lower than the melting point and boiling point of pure metal or its oxide.

According to Barton (Barton et al., 1991) and Niessen (Niessen, 2002) it is particularly evident in the case of nickel (*Ni*), thallium (*Tl*) and lead (*Pb*). This is illustrated by the data included in Table 5.


Table 5. Volatility temperature (reaching resilience 0,1 Pa) for the selected metals in the absence and in 10% presence of chlorine.

Pollutant Formation in Combustion Processes 313

Thus the concentration of hydrogen chloride and free chlorine in the flue gas depends on many factors, including temperature and that is why their emission rates cannot be

The term products of incomplete combustion (PICs) means organic compounds introduced to combustion and formed during combustion that were not oxidized in the combustion zone as well as products of secondary synthesis outside the combustion zone. As already mentioned when discussing the combustion process paths shown in Figure No. 2, in the combustion zone and directly next to it, in the gas phase there are many hydrocarbon and chlorinated hydrocarbons radicals as well as simple aliphatic hydrocarbons, often unsaturated, which may participate in secondary, high-temperature synthesis reactions. Among these radicals there is also acetylene, which plays a key role in the later stages of

The key here is a reaction first described by Aubrey and van Wazer (Aubrey & van Wazer, 1964), in which in high temperature aromatic compounds are formed from aliphatic

A special role is played by acetylene. It is always present in the flue gas from the combustion process and is a precursor of the formation of many chloro-aromatic compounds (Lenoir et al., 2001). It is subject in the first stage to chlorination to dichloroacetylene in result of ligand exchange reaction and then is subject either to cyclization to hexachlorobenzene or condensation to hexachlorobutadiene (Lenoir et al., 1998). Subsequent studies have shown that, for example from acetylene at high temperatures chlorobenzenes, chlorophenols and chloronaphthalenes can be formed.

In addition, the presence in the waste of organic substances of unsaturated character, that is containing in the molecule double or triple carbon-carbon bonds (*C C* or *C C* ) causes that at a temperature of about 500-800°C, occurs the synthesis reaction of polycyclic aromatic hydrocarbons (Liow et al., 1997). Polycyclic aromatic hydrocarbons (PAHs) are among the most dangerous substances in the environment. Many of them, such as benzo(a)pyrene, benzo(a)anthracene, benzo(k)fluoranthene, dibenzo(a,h)anthracene, indeno(1,2,3-c,d)pyrene are classified by the International Agency for Research on Cancer (IARC) as substances with proven carcinogenic effect. They are emitted to the atmosphere practically from all combustion

The result of these reactions, which proceed outside the combustion zone, is the presence of


processes, not only from the waste incineration plants (Mastral et al., 2000).

pollutants in the flue gas, among others such as (Eduljee, 1994):

<sup>2</sup> <sup>6</sup> <sup>4</sup> <sup>6</sup> <sup>6</sup> <sup>9</sup> *<sup>C</sup> Cl* <sup>12</sup> *CCl <sup>C</sup> Cl* (55)

predicted on the basis of chlorine content in the fuel.

**14. Products of incomplete combustion** 

synthesis.

compounds (acetylene):


pentachlorophenol,


So the emission of metals from combustion processes largely depends on the type of metal, temperature of the combustion process, but also on the presence of halides - chlorine and bromine which significantly increase the presence of some metals in flue gas.

#### **13. Hydrogen chloride emission**

Chlorine (*Cl*) is an element that is widespread in the environment. Its small, sometimes even trace amounts are found in each fuel. Of course, the least amount of chlorine (almost immeasurable amount) is in fuel gas, slightly more in liquid fuels, while in solid fuels its content may be as high as 2%. In the process of combustion, chlorine - like metals is subject to speciation. The largest part of chlorine (50-60%) is bound in the form of chlorides in the fly ash, a part is also bound in a similar way in slag and grate ashes. The remaining part of chlorine is released into the environment as emissions of hydrogen chloride (*HCl*), while only a small amount is released in form of molecular chlorine (*Cl2*). With the increase of combustion temperature, the amount of chlorine released in the form of emissions increases and at the same time the amount of chlorine present in the slag and ashes decreases. Data on the speciation of chlorine in the combustion process at different temperatures according to the work of Liu (Liu et al., 2000) are shown in Table 6


Table 6. Speciation of chlorine in the combustion process depending on the temperature

The presence of chlorine in fly ash is of great importance in the formation of chloroorganic compounds outside the combustion zone. The presence of free chlorine which can participate in the chlorination and oxychlorination reactions outside the combustion zone is of similar importance. In the flue gas, outside the combustion zone, hydrogen chloride may undergo a catalytic decomposition (Deacon reaction) in accordance with the reaction equation (Griffin, 1986):

$$2\text{ }H\text{Cl} + O\_2 \xrightarrow[\text{catalyst}-\text{Cu}, \text{Fe}, \text{Al}]{} 2\text{ }Cl\_2 + 2\text{ }H\_2O \tag{52}$$

In the case of sulfur dioxide (*SO2*) present in flue gas, at the same time proceeds hydrogen chloride reproduction reaction (Lindbauer et al., 1994):

$$\text{SO}\_2 + \text{Cl}\_2 + \text{H}\_2\text{O} \xrightarrow{\text{e}} \text{SO}\_3 + \text{2HCl} \tag{53}$$

The balance in a chlorine - hydrogen chloride system is also affected by hydroxyl radicals and water present in the flue gas:

$$\text{HCl} + \text{OH} \longleftrightarrow \text{Cl} + \text{H}\_2\text{O} \tag{54}$$

So the emission of metals from combustion processes largely depends on the type of metal, temperature of the combustion process, but also on the presence of halides - chlorine and

Chlorine (*Cl*) is an element that is widespread in the environment. Its small, sometimes even trace amounts are found in each fuel. Of course, the least amount of chlorine (almost immeasurable amount) is in fuel gas, slightly more in liquid fuels, while in solid fuels its content may be as high as 2%. In the process of combustion, chlorine - like metals is subject to speciation. The largest part of chlorine (50-60%) is bound in the form of chlorides in the fly ash, a part is also bound in a similar way in slag and grate ashes. The remaining part of chlorine is released into the environment as emissions of hydrogen chloride (*HCl*), while only a small amount is released in form of molecular chlorine (*Cl2*). With the increase of combustion temperature, the amount of chlorine released in the form of emissions increases and at the same time the amount of chlorine present in the slag and ashes decreases. Data on the speciation of chlorine in the combustion process at different temperatures according to

HCl 7,6 10,7 10,1 19,9 27,4 43,7 Cl2 0,5 0,2 0,4 0,6 0,8 1,3 Fly ash 50,9 47,5 56,3 53,2 62,4 51,9 Slag 40,5 41,3 32,8 25,6 8,5 1,2 Table 6. Speciation of chlorine in the combustion process depending on the temperature

The presence of chlorine in fly ash is of great importance in the formation of chloroorganic compounds outside the combustion zone. The presence of free chlorine which can participate in the chlorination and oxychlorination reactions outside the combustion zone is of similar importance. In the flue gas, outside the combustion zone, hydrogen chloride may undergo a catalytic decomposition (Deacon reaction) in accordance with the reaction

*HCl O Cl H O*

In the case of sulfur dioxide (*SO2*) present in flue gas, at the same time proceeds hydrogen

The balance in a chlorine - hydrogen chloride system is also affected by hydroxyl radicals

600 650 700 750 800 900

*catalyst Cu Fe Al* <sup>2</sup> <sup>2</sup> , , <sup>2</sup> <sup>4</sup> <sup>2</sup> <sup>2</sup> (52)

*SO Cl <sup>H</sup> <sup>O</sup> SO* <sup>2</sup> *HCl* <sup>2</sup> <sup>2</sup> <sup>2</sup> <sup>3</sup> (53)

*HCl OH Cl <sup>H</sup> <sup>O</sup>*<sup>2</sup> (54)

bromine which significantly increase the presence of some metals in flue gas.

**13. Hydrogen chloride emission** 

equation (Griffin, 1986):

and water present in the flue gas:

the work of Liu (Liu et al., 2000) are shown in Table 6

chloride reproduction reaction (Lindbauer et al., 1994):

Presence of chlorine Temperature (°C)

Thus the concentration of hydrogen chloride and free chlorine in the flue gas depends on many factors, including temperature and that is why their emission rates cannot be predicted on the basis of chlorine content in the fuel.

#### **14. Products of incomplete combustion**

The term products of incomplete combustion (PICs) means organic compounds introduced to combustion and formed during combustion that were not oxidized in the combustion zone as well as products of secondary synthesis outside the combustion zone. As already mentioned when discussing the combustion process paths shown in Figure No. 2, in the combustion zone and directly next to it, in the gas phase there are many hydrocarbon and chlorinated hydrocarbons radicals as well as simple aliphatic hydrocarbons, often unsaturated, which may participate in secondary, high-temperature synthesis reactions. Among these radicals there is also acetylene, which plays a key role in the later stages of synthesis.

The key here is a reaction first described by Aubrey and van Wazer (Aubrey & van Wazer, 1964), in which in high temperature aromatic compounds are formed from aliphatic compounds (acetylene):

$$\text{9 C}\_2\text{Cl}\_6 \longleftrightarrow \text{12 CCl}\_4 + \text{C}\_6\text{Cl}\_6 \tag{55}$$

A special role is played by acetylene. It is always present in the flue gas from the combustion process and is a precursor of the formation of many chloro-aromatic compounds (Lenoir et al., 2001). It is subject in the first stage to chlorination to dichloroacetylene in result of ligand exchange reaction and then is subject either to cyclization to hexachlorobenzene or condensation to hexachlorobutadiene (Lenoir et al., 1998). Subsequent studies have shown that, for example from acetylene at high temperatures chlorobenzenes, chlorophenols and chloronaphthalenes can be formed.

In addition, the presence in the waste of organic substances of unsaturated character, that is containing in the molecule double or triple carbon-carbon bonds (*C C* or *C C* ) causes that at a temperature of about 500-800°C, occurs the synthesis reaction of polycyclic aromatic hydrocarbons (Liow et al., 1997). Polycyclic aromatic hydrocarbons (PAHs) are among the most dangerous substances in the environment. Many of them, such as benzo(a)pyrene, benzo(a)anthracene, benzo(k)fluoranthene, dibenzo(a,h)anthracene, indeno(1,2,3-c,d)pyrene are classified by the International Agency for Research on Cancer (IARC) as substances with proven carcinogenic effect. They are emitted to the atmosphere practically from all combustion processes, not only from the waste incineration plants (Mastral et al., 2000).

The result of these reactions, which proceed outside the combustion zone, is the presence of pollutants in the flue gas, among others such as (Eduljee, 1994):


Pollutant Formation in Combustion Processes 315




The main difference in these mechanisms is the source of carbon and the temperature range in which the synthesis takes place. In the light of the research results it seems that the most important mechanism owing to which most dioxins are formed is the third mechanism (*de novo*), then the second and least dioxins are produced by the first mechanism (Huang &

The analysis of the above mentioned paths of appearance of dioxins in the flue gas stream from thermal processes, including combustion of solid fuels shows that the first case is relatively unlikely. PCDD/Fs and PCBs are not chemicals of high thermal stability. In the combustion conditions (presence of oxygen, mixing, flow) practically most organic compounds, including dioxins, is decomposed at the temperature about 850°C. The temperature of 99.9% of the destruction of PCDD/Fs is around 700°C while the temperature of the destruction of other compounds that are precursors of the synthesis of dioxins can reach up to 950°C. However, in the case of lack of oxygen the limit of the decomposition of organic compounds may be increased to around 1000°C. It therefore seems unlikely that the dioxins contained in the material subject to combustion at temperatures reaching and exceeding 1000°C could not be subject to destruction, although in the case of poor construction of the combustion installation and the formation of cold zones in the

The second path of the appearance of PCDD/Fs, PCBs, PCNs in flue gas is a high temperature synthesis which is a condensation of precursor molecules - benzene, naphthalene, phenanthrene, acetone, trichloroethanes, benzaldehyde, dibenzofuran, benzofuran, phenol, mono-, di-, trichlorophenoles, chlorobenzenes, tetrachlorethylene or polychlorinated diphenyl ethers. These precursors may react with each other both at 500- 700°C in the gas phase and on the surface of fly ash at a temperature of 200-500°C. The key to the formation of dioxins in result of synthesis in the gas phase are the conditions for conducting the combustion process. Numerous studies have shown a very close relationship between the formation of chlorinated and non-chlorinated aromatic compounds in the combustion and afterburning chamber (at a temperature 650-900°C) and the parameters of combustion. It was also noted that by-products of gas fuel combustion - such as natural gas,

in the gas phase at temperatures of 500-700°C (Ballschmiter et al., 1985),

called *de novo* synthesis (Vogg & Stieglitz 1986).

combustion chamber this is not excluded.

molecule,

Buekens, 1995).


Subsequent studies conducted by other researchers (Wienecke et al., 1995, Jay & Stieglitz, 1995, Mascolo et al., 1997, Trenholm 1998) showed the presence of more than 350 different types of chemical compounds (organic) in the flue gases from the waste incineration plants in concentrations above 5 µg/m3. Similar results were obtained for burning of wood and other biomass. Unfortunately, there are no such test results for the combustion of coal, although it is expected that the situation is similar.

#### **15. Formation of PCDD/Fs, PCBs, PCNs and PAHs**

Polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), polychlorinated biphenyls (PCBs), polychlorinated naphthalenes (PCNS) and polycyclic aromatic hydrocarbons (PAHs) are also products of incomplete combustion formed outside the combustion zone. Their common feature is the negative biological impact - some of them have proven carcinogenic properties (PAHs), others are xantoestrogens (endocrine disrupters) that disrupt the hormonal balance of living organisms. In recent years many myths referring to them have appeared, however, most of them have nothing to do with reality.

According to the diagram in Figure 2 showing the combustion process of solid fuels, the crucial stage for the synthesis of PCDD/Fs, PCBs, PCNs and PAHs is the formation of hydrocarbon radicals, often including halogenated (mainly chlorinated) ones as well simple unsaturated hydrocarbons e.g. ethylene and acetylene. These reactions proceed at high temperatures and in the next stage their products undergo further reactions such as cyclization - the formation of aromatic often chlorinated compounds - including benzene, naphthalene, chlorobenzenes and chlorophenols. All these chemical compounds are formed in virtually any combustion process of solid and liquid fuels and also, although less frequently, gas fuels. If in the zone of respective temperatures appear chlorine, oxygen and organic matter chlorinated compounds are formed - such as PCDD / Fs, PCBs and PCNs. This phenomenon is observed not only for combustion process but also for most thermal processes running at temperatures of 200-700°C. The overall reaction is as follows:

$$R-Cl + O\_2 \xrightarrow[\text{conclusion}]{} \text{PCDD} / \text{Fs} \text{ } \text{PCBs} \text{ } \text{PCNS} \tag{56}$$

On the basis of the critical analysis of numerous literature data, we can assume the following paths of appearance of dioxins (as well as PCBs and PCNs) in the flue gas stream from combustion:



Subsequent studies conducted by other researchers (Wienecke et al., 1995, Jay & Stieglitz, 1995, Mascolo et al., 1997, Trenholm 1998) showed the presence of more than 350 different types of chemical compounds (organic) in the flue gases from the waste incineration plants in concentrations above 5 µg/m3. Similar results were obtained for burning of wood and other biomass. Unfortunately, there are no such test results for the combustion of coal,

Polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), polychlorinated biphenyls (PCBs), polychlorinated naphthalenes (PCNS) and polycyclic aromatic hydrocarbons (PAHs) are also products of incomplete combustion formed outside the combustion zone. Their common feature is the negative biological impact - some of them have proven carcinogenic properties (PAHs), others are xantoestrogens (endocrine disrupters) that disrupt the hormonal balance of living organisms. In recent years many myths referring to them have appeared, however, most of them have nothing to do with

According to the diagram in Figure 2 showing the combustion process of solid fuels, the crucial stage for the synthesis of PCDD/Fs, PCBs, PCNs and PAHs is the formation of hydrocarbon radicals, often including halogenated (mainly chlorinated) ones as well simple unsaturated hydrocarbons e.g. ethylene and acetylene. These reactions proceed at high temperatures and in the next stage their products undergo further reactions such as cyclization - the formation of aromatic often chlorinated compounds - including benzene, naphthalene, chlorobenzenes and chlorophenols. All these chemical compounds are formed in virtually any combustion process of solid and liquid fuels and also, although less frequently, gas fuels. If in the zone of respective temperatures appear chlorine, oxygen and organic matter chlorinated compounds are formed - such as PCDD / Fs, PCBs and PCNs. This phenomenon is observed not only for combustion process but also for most thermal

processes running at temperatures of 200-700°C. The overall reaction is as follows:

*R Cl O PCDD Fs PCBs PCNs*

On the basis of the critical analysis of numerous literature data, we can assume the following paths of appearance of dioxins (as well as PCBs and PCNs) in the flue gas stream


*combustion* / , , <sup>2</sup> (56)



1,1,2,2-tetrachloroethane,


reality.

from combustion:

molecule (Tosine et al., 1985),


although it is expected that the situation is similar.

**15. Formation of PCDD/Fs, PCBs, PCNs and PAHs** 


The main difference in these mechanisms is the source of carbon and the temperature range in which the synthesis takes place. In the light of the research results it seems that the most important mechanism owing to which most dioxins are formed is the third mechanism (*de novo*), then the second and least dioxins are produced by the first mechanism (Huang & Buekens, 1995).

The analysis of the above mentioned paths of appearance of dioxins in the flue gas stream from thermal processes, including combustion of solid fuels shows that the first case is relatively unlikely. PCDD/Fs and PCBs are not chemicals of high thermal stability. In the combustion conditions (presence of oxygen, mixing, flow) practically most organic compounds, including dioxins, is decomposed at the temperature about 850°C. The temperature of 99.9% of the destruction of PCDD/Fs is around 700°C while the temperature of the destruction of other compounds that are precursors of the synthesis of dioxins can reach up to 950°C. However, in the case of lack of oxygen the limit of the decomposition of organic compounds may be increased to around 1000°C. It therefore seems unlikely that the dioxins contained in the material subject to combustion at temperatures reaching and exceeding 1000°C could not be subject to destruction, although in the case of poor construction of the combustion installation and the formation of cold zones in the combustion chamber this is not excluded.

The second path of the appearance of PCDD/Fs, PCBs, PCNs in flue gas is a high temperature synthesis which is a condensation of precursor molecules - benzene, naphthalene, phenanthrene, acetone, trichloroethanes, benzaldehyde, dibenzofuran, benzofuran, phenol, mono-, di-, trichlorophenoles, chlorobenzenes, tetrachlorethylene or polychlorinated diphenyl ethers. These precursors may react with each other both at 500- 700°C in the gas phase and on the surface of fly ash at a temperature of 200-500°C. The key to the formation of dioxins in result of synthesis in the gas phase are the conditions for conducting the combustion process. Numerous studies have shown a very close relationship between the formation of chlorinated and non-chlorinated aromatic compounds in the combustion and afterburning chamber (at a temperature 650-900°C) and the parameters of combustion. It was also noted that by-products of gas fuel combustion - such as natural gas,

Pollutant Formation in Combustion Processes 317

The emission of dioxins depends primarily on the construction of the combustion unit and burning conditions. The concentration of chlorine in the flue gases significantly affects the profile of formed congeners PCDD and PCDF. The higher the concentration the greater the share of congeners with higher number of chlorine atoms. This suggests that *de novo* synthesis is the reaction of subsequent chlorination of individual congeners of dioxins and furans. Analyzing the above mechanism of *de novo* synthesis it can be stated that on the fly ash containing CuCl2 there can also run the follow-chlorination of lower substituted dioxins generated in the gas phase as well as dechlorination of higher chlorinated dioxins. The optimal temperature assumed for *de novo* synthesis is about 350°C. Oxygen concentration also influences the course of the *de novo* synthesis. This is obvious if we consider that oxygen is present in the molecule of both dibenzo-p-dioxin and dibenzofuran. Studies have confirmed that the increase in oxygen concentration outside the combustion zone significantly increases the amount of generated dioxins, the increase in the concentration of

The catalyst in the *de novo* synthesis is primarily copper chloride. Copper has a very strong effect in supporting the formation and closing the rings such as chlorophenols. Subsequent studies have shown that transition metals catalyze both ring closing and the process of chlorination and *de novo* synthesis may be treated as an electrophilic substitution reaction that runs in two stages - the first stage includes the chlorination of carbon surface while the second stage is oxidation and decomposition in which dioxins and furans are formed. Similar to the catalytic properties of copper, in the *de novo* synthesis similar properties also have aluminum, iron, magnesium, manganese, cobalt as well as sodium, potassium and zinc. Other metals - such as nickel does not exhibit catalytic activity or they decrease the amount of generated dioxins e.g. chromium, wolfram or vanadium. But there is no doubt that the best catalyst for *de novo* synthesis is copper. In reality, however, on the surface of fly ash there are many metals, whose influence on the synthesis of dioxins is different - some of them catalyze the *de novo* synthesis or the chlorination process of low chlorinated dioxins formed earlier (in the synthesis in gas phase) - for example, copper, zinc, iron or aluminum, while others catalyze the decomposition of generated dioxins or their dechlorination – e.g.

After analyzing the emissions of dioxins from combustion processes and all the described paths of the presence of dioxins in the flue gas, especially the paths of their synthesis it should be noted that probably most dioxins are formed outside the combustion zone. According to Goldfarb, who analyzed the problem of dioxin emissions from Canadian

Summing up the problem of paths of dioxins formation in combustion processes we should first mention the decisive influence of the combustion process conditions on the quantity of produced dioxins and furans. Combustion process conditions include first of all temperature, residence time of gases from the combustion in the zone of high temperatures, oxygen concentration in the combustion zone and directly connected with it concentrations of carbon monoxide in the flue gas. Poor combustion conditions i.e. too low temperature and high concentration of carbon monoxide contribute to the formation of incomplete combustion products such as chlorobenzenes and chlorophenols which are precursors of dioxins and can undergo condensation reactions in both high and low temperature zone in the presence of metals catalyzing condensation reactions. Poor combustion conditions also

incineration plants, this is most probably the vast majority (Goldfarb, 1986).

hydrogen chloride and gives the same effect.

chromium, vanadium, tungsten or nickel.

acetylene, ethane and ethylene may be a sufficient source of carbon for the synthesis of dioxins.

Outside the combustion zone at temperatures below 500°C dioxins are formed as a result of a series of catalytic reactions occurring on the surface of dust containing metals. Usually it is assumed that dioxins are formed there by two mechanisms - the catalytic synthesis of precursors such as chlorobenzenes, chlorophenols, polychlorinated naphthalenes and polychlorinated biphenyls (similarly as in the gas phase) and catalytic synthesis of elemental carbon (soot) contained in dust particles or polycyclic aromatic hydrocarbons as well as gaseous chlorine and oxygen.

The first of these two mechanisms was first described in 1987 by Dickson (Dickson & Karasek 1987). Numerous studies conducted in subsequent years have shown that this reaction proceeds in principle as a condensation reaction of chlorobenzenes or chlorophenols or as a reaction of selective oxychlorination of polychlorinated naphthalenes and polychlorinated biphenyls. This reaction can be catalyzed by many metals – e.g. copper, titanium, manganese, cobalt, zinc but definitely copper has by far the largest catalytic ability. The characteristic of the condensation mechanism (high temperature in gas phase or on the surface of dust particles) is the fact that in this way are formed low chlorinated dioxins, containing from one to maximum 3-4 atoms of chlorine. So the problem of the origins of high chlorinated dioxins, containing from 4 to 8 atoms of chlorine per molecule, in flue gas remained unresolved. The third mechanism of dioxin formation offers the solution to this problem - the *de novo* synthesis.

The second mechanism of the formation of dioxins in the zone after the combustion chamber, in the area of flue gas cooling, was discovered in the late eighties of the last century by Stieglitz and co-workers, the so called mechanism of *de novo* synthesis (Vogg & Stieglitz 1986). The *de novo* synthesis is a slow heterogenic catalytic reaction in which dioxins are formed without the formation of gaseous intermediate products. It occurs within the temperature range 200-500°C from carbon particles contained in fly ash, through the formation and closure of benzene rings, which then react with oxygen and chlorine on the catalyst surface (Stieglitz & Vogg, 1987). The source of carbon in *de novo* synthesis is usually so called elemental carbon contained in fly ash particles, as a carbonized residue of thermal decomposition and oxidation in the combustion process. In practice it is impossible to achieve complete burnout of carbon contained in the fuel and small amounts of unburned carbon are present in flue gas in form of soot which is made up of elemental carbon and the polycyclic aromatic hydrocarbons - PAHs. The emission of dioxins from the combustion of pure wood suggests that the carbon source can also be large, complex molecules of organic compounds such as lignin and lignite. The chlorinating agent in *de novo* synthesis can be free, molecular chlorine, chlorinated organic compounds as well as volatile inorganic chlorine salts - such as NaCl or FeCl3. However, the fundamental role in the formation of dioxins plays the concentration of chlorine free radicals in the zone of combustion reaction and cooling. Despite the obvious role of chlorine in *de novo* synthesis the data from over 1 900 research works on an industrial scale (Rigo & Handler, 1998) shows that there is no clear relationship between the amount of chlorine introduced in the fuel to the combustion process and the amount of dioxins emitted. There is, however, relationship between the amount of dioxins emitted and the concentration of chlorine or hydrogen chloride at the outlet of the combustion chamber.

acetylene, ethane and ethylene may be a sufficient source of carbon for the synthesis of

Outside the combustion zone at temperatures below 500°C dioxins are formed as a result of a series of catalytic reactions occurring on the surface of dust containing metals. Usually it is assumed that dioxins are formed there by two mechanisms - the catalytic synthesis of precursors such as chlorobenzenes, chlorophenols, polychlorinated naphthalenes and polychlorinated biphenyls (similarly as in the gas phase) and catalytic synthesis of elemental carbon (soot) contained in dust particles or polycyclic aromatic hydrocarbons as well as

The first of these two mechanisms was first described in 1987 by Dickson (Dickson & Karasek 1987). Numerous studies conducted in subsequent years have shown that this reaction proceeds in principle as a condensation reaction of chlorobenzenes or chlorophenols or as a reaction of selective oxychlorination of polychlorinated naphthalenes and polychlorinated biphenyls. This reaction can be catalyzed by many metals – e.g. copper, titanium, manganese, cobalt, zinc but definitely copper has by far the largest catalytic ability. The characteristic of the condensation mechanism (high temperature in gas phase or on the surface of dust particles) is the fact that in this way are formed low chlorinated dioxins, containing from one to maximum 3-4 atoms of chlorine. So the problem of the origins of high chlorinated dioxins, containing from 4 to 8 atoms of chlorine per molecule, in flue gas remained unresolved. The third mechanism of dioxin formation offers the solution

The second mechanism of the formation of dioxins in the zone after the combustion chamber, in the area of flue gas cooling, was discovered in the late eighties of the last century by Stieglitz and co-workers, the so called mechanism of *de novo* synthesis (Vogg & Stieglitz 1986). The *de novo* synthesis is a slow heterogenic catalytic reaction in which dioxins are formed without the formation of gaseous intermediate products. It occurs within the temperature range 200-500°C from carbon particles contained in fly ash, through the formation and closure of benzene rings, which then react with oxygen and chlorine on the catalyst surface (Stieglitz & Vogg, 1987). The source of carbon in *de novo* synthesis is usually so called elemental carbon contained in fly ash particles, as a carbonized residue of thermal decomposition and oxidation in the combustion process. In practice it is impossible to achieve complete burnout of carbon contained in the fuel and small amounts of unburned carbon are present in flue gas in form of soot which is made up of elemental carbon and the polycyclic aromatic hydrocarbons - PAHs. The emission of dioxins from the combustion of pure wood suggests that the carbon source can also be large, complex molecules of organic compounds such as lignin and lignite. The chlorinating agent in *de novo* synthesis can be free, molecular chlorine, chlorinated organic compounds as well as volatile inorganic chlorine salts - such as NaCl or FeCl3. However, the fundamental role in the formation of dioxins plays the concentration of chlorine free radicals in the zone of combustion reaction and cooling. Despite the obvious role of chlorine in *de novo* synthesis the data from over 1 900 research works on an industrial scale (Rigo & Handler, 1998) shows that there is no clear relationship between the amount of chlorine introduced in the fuel to the combustion process and the amount of dioxins emitted. There is, however, relationship between the amount of dioxins emitted and the concentration of chlorine or hydrogen chloride at the

dioxins.

gaseous chlorine and oxygen.

to this problem - the *de novo* synthesis.

outlet of the combustion chamber.

The emission of dioxins depends primarily on the construction of the combustion unit and burning conditions. The concentration of chlorine in the flue gases significantly affects the profile of formed congeners PCDD and PCDF. The higher the concentration the greater the share of congeners with higher number of chlorine atoms. This suggests that *de novo* synthesis is the reaction of subsequent chlorination of individual congeners of dioxins and furans. Analyzing the above mechanism of *de novo* synthesis it can be stated that on the fly ash containing CuCl2 there can also run the follow-chlorination of lower substituted dioxins generated in the gas phase as well as dechlorination of higher chlorinated dioxins. The optimal temperature assumed for *de novo* synthesis is about 350°C. Oxygen concentration also influences the course of the *de novo* synthesis. This is obvious if we consider that oxygen is present in the molecule of both dibenzo-p-dioxin and dibenzofuran. Studies have confirmed that the increase in oxygen concentration outside the combustion zone significantly increases the amount of generated dioxins, the increase in the concentration of hydrogen chloride and gives the same effect.

The catalyst in the *de novo* synthesis is primarily copper chloride. Copper has a very strong effect in supporting the formation and closing the rings such as chlorophenols. Subsequent studies have shown that transition metals catalyze both ring closing and the process of chlorination and *de novo* synthesis may be treated as an electrophilic substitution reaction that runs in two stages - the first stage includes the chlorination of carbon surface while the second stage is oxidation and decomposition in which dioxins and furans are formed. Similar to the catalytic properties of copper, in the *de novo* synthesis similar properties also have aluminum, iron, magnesium, manganese, cobalt as well as sodium, potassium and zinc. Other metals - such as nickel does not exhibit catalytic activity or they decrease the amount of generated dioxins e.g. chromium, wolfram or vanadium. But there is no doubt that the best catalyst for *de novo* synthesis is copper. In reality, however, on the surface of fly ash there are many metals, whose influence on the synthesis of dioxins is different - some of them catalyze the *de novo* synthesis or the chlorination process of low chlorinated dioxins formed earlier (in the synthesis in gas phase) - for example, copper, zinc, iron or aluminum, while others catalyze the decomposition of generated dioxins or their dechlorination – e.g. chromium, vanadium, tungsten or nickel.

After analyzing the emissions of dioxins from combustion processes and all the described paths of the presence of dioxins in the flue gas, especially the paths of their synthesis it should be noted that probably most dioxins are formed outside the combustion zone. According to Goldfarb, who analyzed the problem of dioxin emissions from Canadian incineration plants, this is most probably the vast majority (Goldfarb, 1986).

Summing up the problem of paths of dioxins formation in combustion processes we should first mention the decisive influence of the combustion process conditions on the quantity of produced dioxins and furans. Combustion process conditions include first of all temperature, residence time of gases from the combustion in the zone of high temperatures, oxygen concentration in the combustion zone and directly connected with it concentrations of carbon monoxide in the flue gas. Poor combustion conditions i.e. too low temperature and high concentration of carbon monoxide contribute to the formation of incomplete combustion products such as chlorobenzenes and chlorophenols which are precursors of dioxins and can undergo condensation reactions in both high and low temperature zone in the presence of metals catalyzing condensation reactions. Poor combustion conditions also

Pollutant Formation in Combustion Processes 319

Such efforts to reduce emissions, which are an interference in the combustion process are called primary methods of reducing emissions. They become more and more important in recent years, because from the economic point of view, they are more cost effective (cheaper) than the secondary methods, called "end of pipe technologies". Their effectiveness is generally lower than that of flue gas cleaning methods, however, they significantly help to improve the work of the latter, allowing to obtain the total degree of purification of gases at

In recent years the determination of the effect of parameters of the combustion process on the emission of metals, PAHs, dioxins and other organic substances from the combustion process became an important issue. Numerous literature reports clearly show that keeping good combustion process parameters significantly affects the reduction of emissions of organic substances (including PAHs and dioxins) to the atmosphere. A parameter which characterizes "good combustion conditions" very well is the concentration of carbon monoxide in the flue gas. Other parameters defining the "good combustion conditions" are above all so called "3T"- temperature, turbulence and time of flue gas residence in proper temperature. Detailed conditions are specified among others in the Directive on the incineration of waste (2000/76/EC) - they are as follows: flue gas residence time at 850°C should be not less than 2 seconds at high gas turbulence (for waste containing less than 1% of chlorine compounds, in the case of waste containing more than 1% of chlorine

It is essential, however, to ensure combustion conditions close to complete and perfect combustion, with minimized amount of formed carbon monoxide. Such conditions prevail in properly oxygenated combustion zone at an optimal concentration of oxygen (for coal combustion 6-8%, excess air coefficient about 1.4-1.8). In the case of waste incineration better oxygenation of the combustion zone and providing excess air coefficient of the order of 2,0- 2,4 (oxygen concentration 10-13%) is required. The effect of oxygen concentration and combustion temperature on the concentration of carbon monoxide in the flue gas is

Fig. 5. The effect of oxygen concentration in the combustion zone on the carbon monoxide

According to the research of Seeker (Seeker, 2001) the concentration of carbon monoxide, that is combustion conditions, has a great impact on the amount of dioxins produced in the

combustion process and their emission. This is illustrated in Table 7.

compounds the temperature should be higher than 1100°C).

illustrated in Figure 5 (Seeker, 2001).

emission.

a level above 99%.

favor the formation of polycyclic aromatic hydrocarbons (PAHs) and unburned coal residues in the dust entrained in the air which is an excellent raw material for *de novo*  synthesis and this in turn means that you can prove the thesis that there is a direct impact of the combustion conditions on the amount of dioxins generated in all three known mechanisms described earlier.

When conducting combustion processes we are always dealing with two opposing reactions (reaction systems) - formation of dioxins as a result of various homo- and heterogenic reactions and decomposition of dioxins as a result of high temperature and catalytic reactions. Graphic illustration of this problem is shown in Figure 4. Most of the dioxins generated in waste incineration plants is formed in result of *de novo* synthesis in the temperature range 200-400°C with maximum temperatures around 300°C. This maximum is the result of synthesis and decomposition processes running in parallel.

Fig. 4. Formation and decomposition of dioxins in a catalytic process on dust particles.

The process of formation of PCDD/Fs, PCBs, PCNS and PAHs during combustion depends on many factors and that is why it is very difficult to predict how much of these compounds will be formed in a specific process with fixed parameters. Momentary dioxin emissions results not only from the current combustion conditions. Due to the fact that probably most of the dioxins is produced by *de novo* synthesis, which is a very slow reaction, the formation of dioxins and their release continues long after the optimization of the parameters of the combustion process and minimizing the amount of generated dioxins. This is so called memory effect occurring in waste incineration plants. In the case of unstable operation of incinerators high concentrations of dioxins in the flue gas can be observed over long periods of time (Hunsinger et al., 2007).

#### **16. Primary methods of emission reduction**

Understanding the mechanisms of pollutants formation in the combustion process enables the development of such technologies and methods of combustion which ensure the smallest amount of generated pollutants. There are well known so called low emission combustion processes relating to carbon monoxide and nitrogen oxides - the use of the combustion zone, in stages, flue gas recirculation and the so-called reburning (introducing fuel to the periphery of flame zone) allows to reduce NOx emissions by about 50-60% compared to the original process (Jarosiński, 1996, Hill & Smoot, 2000).

favor the formation of polycyclic aromatic hydrocarbons (PAHs) and unburned coal residues in the dust entrained in the air which is an excellent raw material for *de novo*  synthesis and this in turn means that you can prove the thesis that there is a direct impact of the combustion conditions on the amount of dioxins generated in all three known

When conducting combustion processes we are always dealing with two opposing reactions (reaction systems) - formation of dioxins as a result of various homo- and heterogenic reactions and decomposition of dioxins as a result of high temperature and catalytic reactions. Graphic illustration of this problem is shown in Figure 4. Most of the dioxins generated in waste incineration plants is formed in result of *de novo* synthesis in the temperature range 200-400°C with maximum temperatures around 300°C. This maximum is

Fig. 4. Formation and decomposition of dioxins in a catalytic process on dust particles.

The process of formation of PCDD/Fs, PCBs, PCNS and PAHs during combustion depends on many factors and that is why it is very difficult to predict how much of these compounds will be formed in a specific process with fixed parameters. Momentary dioxin emissions results not only from the current combustion conditions. Due to the fact that probably most of the dioxins is produced by *de novo* synthesis, which is a very slow reaction, the formation of dioxins and their release continues long after the optimization of the parameters of the combustion process and minimizing the amount of generated dioxins. This is so called memory effect occurring in waste incineration plants. In the case of unstable operation of incinerators high concentrations of dioxins in the flue gas can be observed over long periods

Understanding the mechanisms of pollutants formation in the combustion process enables the development of such technologies and methods of combustion which ensure the smallest amount of generated pollutants. There are well known so called low emission combustion processes relating to carbon monoxide and nitrogen oxides - the use of the combustion zone, in stages, flue gas recirculation and the so-called reburning (introducing fuel to the periphery of flame zone) allows to reduce NOx emissions by about 50-60%

the result of synthesis and decomposition processes running in parallel.

mechanisms described earlier.

of time (Hunsinger et al., 2007).

**16. Primary methods of emission reduction** 

compared to the original process (Jarosiński, 1996, Hill & Smoot, 2000).

Such efforts to reduce emissions, which are an interference in the combustion process are called primary methods of reducing emissions. They become more and more important in recent years, because from the economic point of view, they are more cost effective (cheaper) than the secondary methods, called "end of pipe technologies". Their effectiveness is generally lower than that of flue gas cleaning methods, however, they significantly help to improve the work of the latter, allowing to obtain the total degree of purification of gases at a level above 99%.

In recent years the determination of the effect of parameters of the combustion process on the emission of metals, PAHs, dioxins and other organic substances from the combustion process became an important issue. Numerous literature reports clearly show that keeping good combustion process parameters significantly affects the reduction of emissions of organic substances (including PAHs and dioxins) to the atmosphere. A parameter which characterizes "good combustion conditions" very well is the concentration of carbon monoxide in the flue gas. Other parameters defining the "good combustion conditions" are above all so called "3T"- temperature, turbulence and time of flue gas residence in proper temperature. Detailed conditions are specified among others in the Directive on the incineration of waste (2000/76/EC) - they are as follows: flue gas residence time at 850°C should be not less than 2 seconds at high gas turbulence (for waste containing less than 1% of chlorine compounds, in the case of waste containing more than 1% of chlorine compounds the temperature should be higher than 1100°C).

It is essential, however, to ensure combustion conditions close to complete and perfect combustion, with minimized amount of formed carbon monoxide. Such conditions prevail in properly oxygenated combustion zone at an optimal concentration of oxygen (for coal combustion 6-8%, excess air coefficient about 1.4-1.8). In the case of waste incineration better oxygenation of the combustion zone and providing excess air coefficient of the order of 2,0- 2,4 (oxygen concentration 10-13%) is required. The effect of oxygen concentration and combustion temperature on the concentration of carbon monoxide in the flue gas is illustrated in Figure 5 (Seeker, 2001).

Fig. 5. The effect of oxygen concentration in the combustion zone on the carbon monoxide emission.

According to the research of Seeker (Seeker, 2001) the concentration of carbon monoxide, that is combustion conditions, has a great impact on the amount of dioxins produced in the combustion process and their emission. This is illustrated in Table 7.

Pollutant Formation in Combustion Processes 321

Fig. 6. The effect of combustion conditions (concentration of CO in flue gas) on the emission

Aubrey, N. E., van Wazer, J. J. (1964) Equilibrium Rearrangement of Perchlorocarbon Compounds. *Journal of American Chemical Society*, Vol.86, pp.4380-4383 Ballschmiter, K., Zoller, W., Buchert, H., Clas Th. (1985) Correlation between substitution

pattern and reaction pathway in the formation of polychlorodibenzofurans.

of some organic pollutants from combustion process (Schmöckel & Streit, 1994).

*Fresenius Zeitschrift für Analistische Chemie*, Vol. 322, pp.587-594

**17. References** 


Table 7. The effect of combustion conditions on the rate of PCDD/Fs emission to atmosphere and the concentration in secondary waste from the combustion process (per 1 Mg of incinerated waste).

The research conducted by Schmöckel (Schmöckel & Streit, 1994) of wood combustion process in plants with a capacity of 11 kW to 13.9 MW have shown a very close correlation between the concentration of some pollutants in the flue gas and the concentration of carbon monoxide, which clearly shows the impact of combustion conditions on the emission of organic pollutants. The results of these studies are shown in Figure 6.

Combustion parameters have also great impact on emissions of metals. This is due to the different volatility of metals and their speciation depending on the forms of occurrence. The presence of chlorine and hydrogen chloride in the combustion zone causes that different amounts of metals are in slag, fly ash, ash from de-dusting equipment or in gases emitted to the atmosphere (Chiang et al. 1997). There is also a clear effect of combustion temperature and oxygen concentration in different zones of combustion on the balance of metals in individual secondary products of the combustion process (Wunsh et al., 1995, Wey et al., 1998). For example, controlling the amount of primary air in the individual zones of the grate and the temperature can result in almost zero emission of some metals with almost 100% their content in the slag (Modigell & Liebig, 1999). The use of primary methods of reducing emissions makes that secondary purification of flue gas to the level required by law (e.g. Directive 2000/76/EC) is easier. The positive effect of the use of primary methods of reducing emissions on the concentration of pollutants in flue gas is shown in Table 8.


Table 8. The effect of the primary methods for reducing emissions of pollutants in waste incineration plants (Büttenberger & Hansen 1997).

Table 7. The effect of combustion conditions on the rate of PCDD/Fs emission to

organic pollutants. The results of these studies are shown in Figure 6.

Pollution Unit

incineration plants (Büttenberger & Hansen 1997).

atmosphere and the concentration in secondary waste from the combustion process (per 1

The research conducted by Schmöckel (Schmöckel & Streit, 1994) of wood combustion process in plants with a capacity of 11 kW to 13.9 MW have shown a very close correlation between the concentration of some pollutants in the flue gas and the concentration of carbon monoxide, which clearly shows the impact of combustion conditions on the emission of

Combustion parameters have also great impact on emissions of metals. This is due to the different volatility of metals and their speciation depending on the forms of occurrence. The presence of chlorine and hydrogen chloride in the combustion zone causes that different amounts of metals are in slag, fly ash, ash from de-dusting equipment or in gases emitted to the atmosphere (Chiang et al. 1997). There is also a clear effect of combustion temperature and oxygen concentration in different zones of combustion on the balance of metals in individual secondary products of the combustion process (Wunsh et al., 1995, Wey et al., 1998). For example, controlling the amount of primary air in the individual zones of the grate and the temperature can result in almost zero emission of some metals with almost 100% their content in the slag (Modigell & Liebig, 1999). The use of primary methods of reducing emissions makes that secondary purification of flue gas to the level required by law (e.g. Directive 2000/76/EC) is easier. The positive effect of the use of primary methods of reducing emissions on the concentration of pollutants in flue gas is

Fly ash Dust from

0,33 μg/Mg 0,14 μg/Mg 21,2 mg/Mg 11,2 μg/Mg

0,35 μg/Mg 0,22 μg/Mg 21,2 mg/Mg 1,6 μg/Mg

Installation without the use of primary methods

Dust mg/m3 2 000 – 10 000 1 000 – 1 500 10 Carbon monoxide mg/m3 50 – 80 10 – 15 50 Sum of organic compounds mg/m3 10 – 100 0,5 – 1 10 PCDD/Fs ng TEQ/m3 5 – 12 0,6 – 1,2 0,1 Table 8. The effect of the primary methods for reducing emissions of pollutants in waste

Installation using the primary methods

Allowable concentration according to Directive 2000/76/EC

electrostatic precipitator

Emission to atmosphere

Combustion parameters

Mg of incinerated waste).

shown in Table 8.

T = 830 oC CO = 25 ppm

T = 850 oC CO = 1 ppm

Slag, bottom ash

Fig. 6. The effect of combustion conditions (concentration of CO in flue gas) on the emission of some organic pollutants from combustion process (Schmöckel & Streit, 1994).

#### **17. References**


Pollutant Formation in Combustion Processes 323

Lenoir, D., Wahrmeier, A., Sidhu, S. S., Taylor, P. H. (2001) Formation and inhibition of

Lindbauer, R. L., Wurst, F., Prey, Th. (1994) PCDD/F emission control for MSWI by SO3

Liow, M. Ch., Lee, W. J., Chen, S. J., Wang, L. Ch., Chung, Ch. H., Chen, J. H. (1997)

Liu, K., Pan, W-P., Riley, J. T. (2000) A study of chlorine behavior In a simulated fluidized

Mascolo, G., Spinos,a L., Lotito, V., Mininni, G., Bagnuolo, G. (1997) Lab-scale evaluations

Mastral, A. M., Callen, M. S., Garcia, T. (2000) Toxic organic emissions from coal

Miller, J. A., Bowman, C. T. (1989) Mechanism and modeling of nitrogen chemistry in combustion. *Progress in Energy and Combustion Science*, Vol.15, pp.287-338 Modigell, M., Liebig, D. (1999) Modeling and simulating the waste incineration process. -

Nissen, W. R. (2002) Combustion and Incineration Processes. Marcel Dekker, New York, Recknagel, H., Sprenger, E., Hönmann, W., Schramek, E. R. (1994) Taschenbuch für Heizung

Rigo, H. G., Handler, A. J. (1998) Is there a strong dioxin-chlorine link in commercial scale

Schmöckel, G., Streit, A. (1994) Emission organischer Stoffe bei der Holzfeuerung.

Seeker, R. (2001) Current understanding of dioxin/furan formation and control*.* - private

Steele, R. C., Malte, P. C., Nicol, D. G., Kramlich, J. C. (1995) NOx and N2O in lean-premixed

Stieglitz, L., Vogg, H. (1987) On formation conditions of PCDD/PCDF in fly ash from

Sukrut, S., Thipse, E., Dreizin, L. (2002) Metal partitioning in products of incineration of

Thome-Kozmiensky, K. J. (1994) Thermische Abfallbehandlung. - EF-Verlag für Energie-

Tosine, H. M., Clemen,t E. R., Ozvacic, V., Wong, G. (1985) Levels of PCDD/PCDF and other chlorinated organics in municipal refuse. *Chemosphere*, Vol.14, pp.821-827 Trenholm, A. (1998) Identification of PICs in hazardous waste combustion emissions. *Waste* 

Vogg, H., Stieglitz, L. (1986) Thermal behavior of PCDD/PCDF in fly ash form municipal

Wey, M.Y., Su, J.L., Yan, M.H., Wei, M.C. (1998) The concentration distribution of heavy

metals under different operation conditions. *The Science of the Total Environment*,

addition. *Organohalogen Compounds*, Vol.19, pp.355-359

*Journal of Aerosol Science*, Vol.28, pp.549-550

bed combustion system. *Fuel*, Vol.79, pp.1115-1124

combustion. *Fuel Processing Technology*, Vol.67, pp.1-10

und Klimatechnik. - R. Oldenburg Verlag, München

Bayerisches Landsamt für Umweltschutz, Augsburg

jet-stirred flames. *Combustion and Flame*, Vol.100, pp.440-449

municipal solid waste. *Chemosphere*, Vol.46, pp.837-849

incinerators. *Chemosphere*, Vol.15, pp.1373-1378

municipal waste incinerators. *Chemosphere*, Vol.16, pp.1917-1922

systems? *Chemosphere*, Vol.37, pp.2031-2046

Vol.43, pp.107-114

pp.219-226

private communication

communication

und Umwelttechnik, Berlin

*Management*, Vol.18, pp.485-492

Vol.212, pp.183-193

chloroaromatic micropollutants formed in incineration processes. *Chemosphere*,

Emission polycyclic aromatic hydrocarbons from medical waste incinerators.

on formation odf products of incomplete combustion in hazardous waste incineration: influence of process variables. *Water Science and Technology*, Vol.36,


Barton, R. G., Seeker, W. R., Bostian, H. E. (1991) The behavior of metals in municipal sludge

Belevi, H., Langmeier, M. (2000) Factors determining the element behavior in municipal

Belevi, H., Moench, H. (2000) Factors determining the element behavior in municipal solid waste incinerators. Field studies. *Environ. Sci. Technol*. Vol.34, pp.2501-2506 Bowman, C.T., Hanson, R.K., Louge, M. Garman, A. (1982) Gas Phase Reaction Kinetics of

*1982 Joint Symposium on Stationary Symposium NOx Control*, Dallas, USA Bozzelli, J. W., Chang, A. Y., Dean, A. M. (1994) Analysis of the reactions H+N2O and

Büttenberger, B., Hansen, H. (1997) Das EVT-Müllverbrennungssystem im Hinblick auf die

Chiang, K.Y., Wang, K.S. Lin, F.L., Chu, W.T (1997) Chloride effects on the speciation and

Dickson, L. C., Karasek, F. W. (1997) Model studies of polychlorinated dibenzo-p-dioxin formation during municipal refuse incineration. - *Science*, Vol.237, pp.754-756 Eduljee, G. H. (1994) Organic micropollutant emissions from waste incineration. - in: Hester

Fenimore, C. P. (1971) Formation of nitric oxide in premixed hydrocarbon flames. *– in:* 

Goldfarb, T. D. (1989) Evidence for post-furnace formation of PCDD and PCDF -

Griffin, R. D. (1986) A new theory of dioxin formation in municipal solid waste combustion.

Hill, S.C., Smoot, L.D. (2000) Modeling of nitrogen oxides formation and destruction in combustion systems. *Progr. in Energy and Combustion Sci*. Vol.16, pp.417-458; Huang, H., Buekens, A. (1995) On the mechanisms of dioxin formation in combustion

Hunsinger, H., Seifert, H., Jay, K. (2007) Control of PCDD/F formation under conditions of

Jay, K., Stieglitz, L. (1995) Identification and quantification of volatile organic components in emission of waste incineration plants. *Chemosphere*, Vol.30, pp.1249-1260 Lenoir, D., Wahrmeier, A., Schramm, K. W., Kaune, A., Zimmermann, R., Taylor, P. H.,

fluctuating combustion performance in MSWI. *Organohalogen Compounds*, Vol.69,

Sidhu, S. S. (1998) Thermal formation of polychlorinated dibenzodioxins and furans: investigation of relevant pathways. *Environmental Engineering Science*,

implications form control. *Chemosphere*, Vol.18, pp.1051-1055

Jarosiński, J. (1996) Techniki czystego spalania (in polish). WNT, Warszawa;

*The Science of the Total Environment*, Vol. 203, pp.129-140

solid waste incinerators. Laboratory experiments. *Environ. Sci. Technol*. Vol.34,

NOx Formation from Fuel Nitrogen in Fossil Fuel Combustion. – *in: Proceedings of* 

NH+NO: Pathways and rate constants over a wide range of temperature and pressure. *in: Proceedings of International Symposium on Combustion*, Vol.25, No.1,

neuesten gesetzlichen Anforderungen. *BWK/TÜV-Umwelt-Special*, October 1997,

partitioning of heavy metals during the municipal solid waste incineration process.

R. E., Harrison R. M. (ed.) - "*Waste incineration and the environment*", Royal Society

*Proceedings of XIIIth International Symposium on Combustion,* The Combustion

incinerators. *Trans. Inst. Chem. Eng*. Vol.69, pp.29-36

pp.2507-2512

pp.965-974

pp.26-38

of Chemistry, London

pp.956-961

Vol.15, pp.37-47*8*

Institute, Pittsburg, pp.373-384

*Chemosphere*, Vol.15, pp.1987-1990

processes. *Chemosphere*, Vol.31, pp.4099-4117


**Part 3** 

**Process Engineering** 


**Part 3** 

**Process Engineering** 

324 Advances in Chemical Engineering

Wielgosiński, G. (2009) Dioxin Emission from Thermal Processes and Methods of Emission

Wienecke, J., Kruse, H., Huckfeldt, U., Eickhoff, W., Wassermann, O. (1995) Organic

Wunsch, P., Matuschek, G., Kettrup, A. (1995) The influence of combustion bed temperature during waste incineration on slag quality. *Thermochimica Acta*, Vol.263, pp.95-100 Zeldovich Y. B. (1946) The Oxidation of Nitrogen in Combustion and Explosions. *Acta* 

compounds in the flue gas of a hazardous incinerator. *Chemosphere*, Vol.30, pp.907-

Reduction. (in polish) - Polish Academy of Science, Lodz

*Physcochimica, Academy of Sciences U.S.S.R.*, Vol.21, 577-628

913;

**13** 

*1Pakistan* 

*3Germany* 

*2Kingdom of Saudi Arabia*

**Systematic Framework for Multiobjective** 

Ramzan Naveed1, Zeeshan Nawaz2, Werner Witt3 and Shahid Naveed1 *1Department of Chemical Engineering,* 

*University of Engineering and Technology, Lahore 2Chemical Technology Development, STCR, Saudi Basic Industries Corporation (SABIC) 3Lehrstuhl Anlagen und Sicherheitstechnik,* 

*Brandenburgicshe Technische Universität, Cottbus* 

**Optimization in Chemical Process Plant Design** 

For solving multiobjective decision making problems, a systematic and effective procedure is required. As far as the process or control system has to be modified process simulators like Aspen PlusTM, Aspen Dynamics are widely used. But these simulators are not designed for investigation of other objectives as environment and safety. Due to complex and conflicting nature of multiobjective decision making an integrated optimization tool should be of value. In this chapter a systematic methodology based on independent modules and its

The methodology is built around several standard independent techniques. These techniques have been suitably modified/adopted and woven together in an integrated plate form. The main aim is to standardize the screening and selection of decisions during design/modification of chemical process plant and optimizing the process variables in order to generate a process with improved economics along with satisfaction of environmental and safety constraints. The methodology (see Figure 1) consists of four

Analysis of alternatives i.e. generation of relevant data for comparison of

Design evaluation stage i.e. decision making from the pareto-surface of non-inferior

different stages to deal this problem is presented in detail.

Generation of alternatives and problem definition;

Environmental, economic and safety objectives Multiobjective decision analysis/ optimization

solution or ranking of alternatives

**1. Introduction** 

layers/stages:

**2. Proposed methodology** 

### **Systematic Framework for Multiobjective Optimization in Chemical Process Plant Design**

Ramzan Naveed1, Zeeshan Nawaz2,

Werner Witt3 and Shahid Naveed1 *1Department of Chemical Engineering, University of Engineering and Technology, Lahore 2Chemical Technology Development, STCR, Saudi Basic Industries Corporation (SABIC) 3Lehrstuhl Anlagen und Sicherheitstechnik, Brandenburgicshe Technische Universität, Cottbus 1Pakistan 2Kingdom of Saudi Arabia 3Germany* 

#### **1. Introduction**

For solving multiobjective decision making problems, a systematic and effective procedure is required. As far as the process or control system has to be modified process simulators like Aspen PlusTM, Aspen Dynamics are widely used. But these simulators are not designed for investigation of other objectives as environment and safety. Due to complex and conflicting nature of multiobjective decision making an integrated optimization tool should be of value. In this chapter a systematic methodology based on independent modules and its different stages to deal this problem is presented in detail.

#### **2. Proposed methodology**

The methodology is built around several standard independent techniques. These techniques have been suitably modified/adopted and woven together in an integrated plate form. The main aim is to standardize the screening and selection of decisions during design/modification of chemical process plant and optimizing the process variables in order to generate a process with improved economics along with satisfaction of environmental and safety constraints. The methodology (see Figure 1) consists of four layers/stages:


Systematic Framework for Multiobjective Optimization in Chemical Process Plant Design 329

about the process and chemicals involved such as thermodynamic and kinetic data can be found from journal articles, patents or handbooks. Current chemical prices can be obtained from market reports if not available in main plant documentation or company central data base. In addition to these sources, some data related to quantification of environmental impacts and material safety data sheets of chemical are also collected from commercial data bases so that an impact assessment and safety analysis can be performed in subsequent design steps. Commercial computer aided tool like ComosPT can be used for plant

This stage is composed of independent modules used to generate relevant information for evaluation of economic, safety and environmental performance objectives. These modules

and a data manager for managing the relevant information generated from these

In the process module, an operation model of the process system has to be developed for evaluating alternatives. The configured simulation model has to be able to reproduce the selected results to an accepted degree of accuracy. This simulation model can be used for design and operation, revamping and debottlenecking of the process under study[7]. Three major integrated simulation systems widely used in the firms and companies for this purpose are Aspen technology (Aspen Plus, Aspen dynamics etc), Hyprotech (Hysys process, Hysys plant etc) and Simulation Sciences (Pro/II etc.). Aspen PlusTM 12.1 is used in this work for development of simulation model and linked in a visual basic platform for integration with safety, economic and environment modules. The most important results available from the process simulation model are material and energy balance information for both streams and units, rating performance of units and tables and graphs of physical properties. A brief description of Aspen PlusTM 12.1 and steps involved in development of

Aspen PlusTM supports both sequential modular and equation oriented computation strategy and allows the user to build and run a steady-state simulation model for a chemical process. It provides a flexible and productive engineering environment designed to maximize the results of engineering efforts, such as user interface mode manager, quick property analysis, rigorous and robust flowsheet modelling, interactive architecture, powerful model analysis tools and analysis and communication of results. Therefore, it lets the user to focus his/her energies on solving the engineering problems, not on how to use the software. It is not only good for process simulation but also allows to perform a wide

documentation and to support stage-I of proposed methodology.

the process simulation model is described here below.

**2.2 Stage II: Analysis of alternatives** 

are:

 Process module Safety module [1,3] Economic module Environment module [2]

modules.

**Aspen PlusTM** 

**2.3 Process module** 

Fig. 1. Simplified block diagram of proposed methodology

#### **2.1 Stage I: Generation of alternatives and problem definition**

The first layer composed of following tasks:

Definition of the scope of the study,

Statement of key assumptions and the performance targets such as quality etc.,

Degree of freedom analysis,

Identification of the key design, control, and manipulated variables,

Definition of the system boundary,

Identification of constraints,

Choice of functional unit for all calculations,

Collection of relevant information about process and chemicals to be handled,

Generation of different alternatives either based on suggestion from independent departments or using the individual objective modules from stage II.

Seader et al. (1999)[6] has described rules for selection of process variables in the book "Process design principles-synthesis, analysis and evaluation". The data and information about the process and chemicals involved such as thermodynamic and kinetic data can be found from journal articles, patents or handbooks. Current chemical prices can be obtained from market reports if not available in main plant documentation or company central data base. In addition to these sources, some data related to quantification of environmental impacts and material safety data sheets of chemical are also collected from commercial data bases so that an impact assessment and safety analysis can be performed in subsequent design steps. Commercial computer aided tool like ComosPT can be used for plant documentation and to support stage-I of proposed methodology.

#### **2.2 Stage II: Analysis of alternatives**

This stage is composed of independent modules used to generate relevant information for evaluation of economic, safety and environmental performance objectives. These modules are:

Process module

328 Advances in Chemical Engineering

Comos PT

Aspen Plus

Access-VB-6.0

Environment module

> Objective functions

**Generation of alternatives**

Economic module

Aspen Dynamics VB6 WAR-VB6

Process module

Data manager

**Design Evaluation-------------------- Final Solution**

**Analysis of alternatives**

**Multiobjective decision analysis (MOO/MADA)**

Safety module

Modify alternatives (Expert,Knowledge base)

MATLAB-VB-Excel

Fig. 1. Simplified block diagram of proposed methodology

The first layer composed of following tasks:

Choice of functional unit for all calculations,

Definition of the scope of the study,

Definition of the system boundary, Identification of constraints,

Degree of freedom analysis,

**2.1 Stage I: Generation of alternatives and problem definition** 

Identification of the key design, control, and manipulated variables,

departments or using the individual objective modules from stage II.

Statement of key assumptions and the performance targets such as quality etc.,

Collection of relevant information about process and chemicals to be handled,

Generation of different alternatives either based on suggestion from independent

Seader et al. (1999)[6] has described rules for selection of process variables in the book "Process design principles-synthesis, analysis and evaluation". The data and information


#### **2.3 Process module**

In the process module, an operation model of the process system has to be developed for evaluating alternatives. The configured simulation model has to be able to reproduce the selected results to an accepted degree of accuracy. This simulation model can be used for design and operation, revamping and debottlenecking of the process under study[7]. Three major integrated simulation systems widely used in the firms and companies for this purpose are Aspen technology (Aspen Plus, Aspen dynamics etc), Hyprotech (Hysys process, Hysys plant etc) and Simulation Sciences (Pro/II etc.). Aspen PlusTM 12.1 is used in this work for development of simulation model and linked in a visual basic platform for integration with safety, economic and environment modules. The most important results available from the process simulation model are material and energy balance information for both streams and units, rating performance of units and tables and graphs of physical properties. A brief description of Aspen PlusTM 12.1 and steps involved in development of the process simulation model is described here below.

#### **Aspen PlusTM**

Aspen PlusTM supports both sequential modular and equation oriented computation strategy and allows the user to build and run a steady-state simulation model for a chemical process. It provides a flexible and productive engineering environment designed to maximize the results of engineering efforts, such as user interface mode manager, quick property analysis, rigorous and robust flowsheet modelling, interactive architecture, powerful model analysis tools and analysis and communication of results. Therefore, it lets the user to focus his/her energies on solving the engineering problems, not on how to use the software. It is not only good for process simulation but also allows to perform a wide

Systematic Framework for Multiobjective Optimization in Chemical Process Plant Design 331

**Event tree / Fault tree** Step-1

Step-2

Step-3

Step-4

**System description and objectives of analysis**

**Extended HAZOP**

**Risk [ Ri ] for specific initiating event**

**Risk potential matrix (HAZOP decision matrix)**

Yes

**>Risk limit i Stop**

No

**Safety / risk analysis**

**Safety / risk assessment**

 **Riski**

**Safety / risk system optimization**

§ Simulation of process related malfunctions

Fig. 2. Simplified block diagram of safety module

i = { financial risk , environmental risk , human health risk }

**alternatives Return to Step-1**

Step 4: Safety/risk system optimization

**Definition and analysis of system optimization**

Step 2: Safety/risk analysis (identification of weak points via Extended HAZOP)

Step 1: System description and objectives of analysis (before starting safety and risk analysis)

Step 3: Safety/risk assessment (categorization of risk via risk potential matrix (HAZOP decision matrix))

**§Dynamic simulation**

range of other tasks such as estimating and regressing physical properties, generating custom graphical and tabular output results, sensitivity analysis, data-fitting plant data to simulation models, costing the plant, optimizing the process, and interfacing results to spreadsheets.

The development of a simulation model for a chemical process using Aspen PlusTM 12.1 involves the following steps (see details in table 1):

	- Unit operations
	- Process streams flowing between the units
	- Unit operation models to describe each unit operation


5. Specify the operating conditions for the unit operations,

Table 1. Developmental process for an Aspen PlusTM simulation model

#### **2.4 Safety module**

Safety module is based on combination of conventional standard risk analysis techniques and process disturbance simulation. This module not only generates relevant information related to safety aspects for multiobjective decision analysis but also used for safety/risk analysis and optimization. The purpose of this module is to determine risk from operational disturbances and to develop effective risk reductions. It can be divided into the following steps (Figure 2):

range of other tasks such as estimating and regressing physical properties, generating custom graphical and tabular output results, sensitivity analysis, data-fitting plant data to simulation models, costing the plant, optimizing the process, and interfacing results to

The development of a simulation model for a chemical process using Aspen PlusTM 12.1

3. Choose a thermodynamic model to represent the physical properties of the components

4. Specify the component flow rates and thermodynamic conditions (i.e. temperature,

streams, unit operations, and product streams

Calculate the temperature, pressure, vapor fraction, molecular weight, enthalpy, entropy and density for the

many of the property parameters required by physical

carry material and energy flows from one block to another. For all process feed streams, we must specify flowrate, composition, and thermodynamic condition

when we define our simulation flowsheet

**Step Used to**  Defining the flowsheet Break down the desired process into its parts: feed

simulation streams Entering components From a databank that is full of common components Estimating property parameters Property Constant Estimate System (PCES) can estimate

property models

Specifying streams Streams connect unit operation blocks in a flowsheet and

Unit operation blocks We choose unit operation models for flowsheet blocks

Safety module is based on combination of conventional standard risk analysis techniques and process disturbance simulation. This module not only generates relevant information related to safety aspects for multiobjective decision analysis but also used for safety/risk analysis and optimization. The purpose of this module is to determine risk from operational disturbances and to develop effective risk reductions. It can be divided into the following

Table 1. Developmental process for an Aspen PlusTM simulation model

spreadsheets.

Unit operations

Specifying stream properties

and units

**2.4 Safety module** 

steps (Figure 2):

2. Specify the chemical components,

and mixtures in the process,

involves the following steps (see details in table 1):

1. Define the process flowsheet configuration by specifying

Unit operation models to describe each unit operation

Process streams flowing between the units

pressure, or phase condition) of the feed streams, 5. Specify the operating conditions for the unit operations,

§ Simulation of process related malfunctions

i = { financial risk , environmental risk , human health risk }

Step 1: System description and objectives of analysis (before starting safety and risk analysis)

Step 2: Safety/risk analysis (identification of weak points via Extended HAZOP)

Step 3: Safety/risk assessment (categorization of risk via risk potential matrix (HAZOP decision matrix)) Step 4: Safety/risk system optimization

Fig. 2. Simplified block diagram of safety module

Systematic Framework for Multiobjective Optimization in Chemical Process Plant Design 333

**Effects Class Financial loss (€) Class related consequences: examples** 

**Effects Class Community Class related consequences: examples** 

3 103 – 104

4 104 – 105

5 105 – 106

<sup>2</sup>Minor irritation effect to

Moderate irritation effect to people and non compliance to laws, local news

Moderate irritation effects to people & environment, single injuries and regional news

Significant effects to people and environment, > 1 injuries & regional news

Major effects to people and environement, multiple injuries, fatality likely, regional news

Severe effects to people and environment, fatality, regional news

Multiple fatalities and process shutdown certain, international news

Table 2. Scoring chart for Consequence Financial consequences [3.4]

\* < 10 : Product quality lowering (brief) 1 101 – 102 : Product quality lowering 2 102 – 103 : Product quality lowering (long term)

6 106 – 107 Fireballs due to catastrophic rupture of vapour

7 107 – 108 Vapour cloud explosion (ignoring domino effect) 8 >108 Vapour cloud explosion along with domino effect

people & local news : Product quality lowering (long term)

\* No effect on people : Product quality lowering (brief) 1 Nuisance effect : Product quality lowering

: Production disturbance (brief) Soil contamination Safe dispersion of material release from vent line

> : Production disturbance Material release from the piping Pump damage (pressure impacts)

: Production disturbance (long term) Jet fire as result of release of material from vent line Pool fire (from pump leakage)

product line

: Production disturbance (brief) Soil contamination Safe dispersion of material release from vent line

> : Production disturbance Material release from the piping Pump damage (pressure impacts)

: Production disturbance (long term) Jet fire as result of release of material from vent line Pool fire (from pump leakage)

Fireballs due to catastrophic rupture of pipe or condenser (vapour product line)

Vapour cloud explosion (ignoring domino effect)

Vapour cloud explosion along with domino effect

**Function impairment** 

**Functional Loss** 

**Safety and Environmental pollution** 

> **Function impairment**

**Functional Loss** 

**Safety and Environmental pollution** 

3

4

5

6

7

8

Environment and Health consequences

#### **2.4.1 Step 1: (Before starting safety/risk analysis) - Description of system and objectives of analysis**

For efficient safety/risk studies, the analyst must have an accurate description of the system to be investigated and a clear objective of the analysis study. Therefore, in this step the purpose, objectives, and scope of the study are clearly defined. The necessary information required for the study such as process flow diagrams, piping and instrumentation diagrams, plant layout schematics, material safety data sheets, equipment data sheets, operating instructions, start up and emergency shutdown procedures, and process limits, etc. is gathered from plant documentation. A team under a trained and experience leader with five to seven people including experts of the design and operation of the subject process may be formulated.

#### **2.4.2 Step 2: Safety/risk analysis (Identification of weak points via extended HAZOP) - Extended HAZOP**

Our intention is to identify weak points due to disturbances in operation, which may or may not be hazardous, in order to improve safety, operability, and/or profit at the same time. Extended HAZOP (HAZOP supported by dynamic simulation, event tree and fault tree techniques and HAZOP decision matrix) is used not only for identification of weak points but also for generation and analysis of optimization proposals [8-11]. Extended HAZOP differs from the standard HAZOP approach in following aspects:

i. Use of dynamic simulation:

In Extended HAZOP, the analysis of the influence of disturbances (failures) on the behaviour of the process is based on shortcut or simplified hand calculations or dynamic simulation. Aspen dynamics is used for this purpose.

ii. Classification of risk related consequences:

Each established consequence (hazard) has to be expressed by a consequence class (C). The plant specific scoring (from 0 (lowest) to 8 (highest)) chart is given in Table 2 (a & b) based on principle consequence analysis. For classification of consequences based on principle release estimates, accident consequence analysis techniques (models for calculation of toxic, fire and explosion effects) and plant location data (capital investment, population density etc.) have to be considered.

#### **Illustrative Example 1**

Figure 3 shows the plant lay out considered for developing plant specific consequence scoring chart. The area around the plant is open fields (rural condition). As weather conditions changed around the year, so certain assumptions are made to results in worse case conditions for consequence analysis. These include weather conditions and wind speed that result in smallest value of dispersion coefficients. Therefore, stability "F" and wind speed as low as possible (1.5 m/s) is selected. It is assumed that 10 workers are present (working 24 h each day), which are not distributed uniformly, on the land in area (100 m x 100 m) around the column under study. Acetone is selected as representative fluid for consequence analysis.

Acetone vapours released from the vent line at a rate of 1616 kg/h due to loss of cooling medium. It is assumed that released vapours form a cloud for 30 minutes before being

For efficient safety/risk studies, the analyst must have an accurate description of the system to be investigated and a clear objective of the analysis study. Therefore, in this step the purpose, objectives, and scope of the study are clearly defined. The necessary information required for the study such as process flow diagrams, piping and instrumentation diagrams, plant layout schematics, material safety data sheets, equipment data sheets, operating instructions, start up and emergency shutdown procedures, and process limits, etc. is gathered from plant documentation. A team under a trained and experience leader with five to seven people including experts of the design and operation of the subject process may be formulated.

**2.4.2 Step 2: Safety/risk analysis (Identification of weak points via extended HAZOP) -** 

Our intention is to identify weak points due to disturbances in operation, which may or may not be hazardous, in order to improve safety, operability, and/or profit at the same time. Extended HAZOP (HAZOP supported by dynamic simulation, event tree and fault tree techniques and HAZOP decision matrix) is used not only for identification of weak points but also for generation and analysis of optimization proposals [8-11]. Extended HAZOP

In Extended HAZOP, the analysis of the influence of disturbances (failures) on the behaviour of the process is based on shortcut or simplified hand calculations or dynamic

Each established consequence (hazard) has to be expressed by a consequence class (C). The plant specific scoring (from 0 (lowest) to 8 (highest)) chart is given in Table 2 (a & b) based on principle consequence analysis. For classification of consequences based on principle release estimates, accident consequence analysis techniques (models for calculation of toxic, fire and explosion effects) and plant location data (capital investment, population density

Figure 3 shows the plant lay out considered for developing plant specific consequence scoring chart. The area around the plant is open fields (rural condition). As weather conditions changed around the year, so certain assumptions are made to results in worse case conditions for consequence analysis. These include weather conditions and wind speed that result in smallest value of dispersion coefficients. Therefore, stability "F" and wind speed as low as possible (1.5 m/s) is selected. It is assumed that 10 workers are present (working 24 h each day), which are not distributed uniformly, on the land in area (100 m x 100 m) around the column under study. Acetone is selected as representative fluid for

Acetone vapours released from the vent line at a rate of 1616 kg/h due to loss of cooling medium. It is assumed that released vapours form a cloud for 30 minutes before being

differs from the standard HAZOP approach in following aspects:

simulation. Aspen dynamics is used for this purpose.

ii. Classification of risk related consequences:

**2.4.1 Step 1: (Before starting safety/risk analysis) - Description of system and** 

**objectives of analysis** 

**Extended HAZOP** 

i. Use of dynamic simulation:

etc.) have to be considered.

**Illustrative Example 1** 

consequence analysis.


Environment and Health consequences

Table 2. Scoring chart for Consequence Financial consequences [3.4]

Fig. 3. Plant lay out for establishing consequence score chart

Systematic Framework for Multiobjective Optimization in Chemical Process Plant Design 335

ignited and leads to vapour cloud explosion. The physical effects of this scenario or event is

*<sup>c</sup> TNT*

Then, using relation 1/3 /( ) *ZR M TNT* and figure 4, scaled distance and overpressure is

*<sup>M</sup> <sup>H</sup> <sup>M</sup>*

*TNT*

*H* 

explosion efficiency ~ 0.05 (Cameron 2005) *Hc* heat of combustion of fuel ~

<sup>10</sup> <sup>1</sup> <sup>100</sup>

Fig. 4. Overpressure versus scaled distance for TNT explosions on flat surfaces (Tweeddale

Scaled distance, Z m / kg1/3

Table 3. Results of physical effects of vapour cloud explosion

10 1.64 90 20 3.28 40 50 8.21 20 100 16.40 7

**Scaled distance Ze [m/kg1/3)**

Overpressure, Δp kPa

Then amount of TNT equivalent to the amount of this flammable material is

calculated as:

Where

3.03 x 104 kJ/kg for acetone

**Overpressure (kPa)**

2003, p. 115)

Distance , R M

10

100

1

Weight of fuel in the cloud= M = 1616 / 2 = 808 kg

estimated. Table 3 presents the results obtained.

*HTNT* TNT blast energy ~ 5420 kJ/kg; so 225.25 *MTNT* kg

ignited and leads to vapour cloud explosion. The physical effects of this scenario or event is calculated as:

Weight of fuel in the cloud= M = 1616 / 2 = 808 kg

334 Advances in Chemical Engineering

Open fields

Reactor location

\$ \$

HCR unit

10 - workers

(non-uniform distribution)

Open fields

T1701

100 m

100 m

Acetone recovery unit

Raw Material Storage

Other plant

sites

4 km

1 km

Rail track

Chemical Plant Boundary wall

Product storage

area

Fig. 3. Plant lay out for establishing consequence score chart

Residential Area

5 km

Road

Location of distillation

column under study

DF

Process

Parking Area

Offices

Effluent treatment plant

River

Open fields

Other plant

sites

Then amount of TNT equivalent to the amount of this flammable material is

$$M\_{TNT} = \alpha \cdot \frac{M \cdot H\_c}{H\_{TNT}}$$

Where explosion efficiency ~ 0.05 (Cameron 2005) *Hc* heat of combustion of fuel ~ 3.03 x 104 kJ/kg for acetone

*HTNT* TNT blast energy ~ 5420 kJ/kg; so 225.25 *MTNT* kg

Then, using relation 1/3 /( ) *ZR M TNT* and figure 4, scaled distance and overpressure is estimated. Table 3 presents the results obtained.

Fig. 4. Overpressure versus scaled distance for TNT explosions on flat surfaces (Tweeddale 2003, p. 115)


Table 3. Results of physical effects of vapour cloud explosion

Systematic Framework for Multiobjective Optimization in Chemical Process Plant Design 337

The Extended HAZOP methodology worksheet for documenting the HAZOP team results is shown in Figure 5. Below consequence the physical effects and risk has to be documented first and next risk has to be classified using score charts (Table 3.2) related to financial, environment and health related consequences. The worst score of each risk has to be documented. For each risk related consequence, frequency class has also to be established.

> 103 – 104

Immediate action needed before further operation

Release of material to atmosphere from vent line or vapour line may disperse safely or has toxic effects or can lead to several outcomes such as flash fire, vapour cloud explosion and fire balls. So documenting consequence class in HAZOP work sheet, the score '8' of the most

Action at next occasion after qualification of analysis for improving

104 – 105

\* 1 2 3 4 5 6 7 8

105 – 106

106 – 107

107 –

<sup>108</sup> >108

102 – 103

Then, frequency of vapour cloud explosion = 6 9 10 0.10 0.01 10

Frequency of fatality of a person exposed to VCE = <sup>9</sup> <sup>10</sup> 10 0.20 2.10

<10 101 – 102

system Optional

No further action needed

Fig. 5. Risk potential matrix (Extended HAZOP decision matrix)

So frequency class for this scenario = 9

So frequency class for this scenario = 9

Consequence, €

Frequency C

1/y F

**Illustrative Example 2** 

severe consequence will be documented.

>100 \* 10-1 – 100 1 10-2 - 10-1 2 10-3 - 10-2 3 10-4 - 10-3 4 10-5 - 10-4 5 10-6 - 10-5 6 10-7 - 10-6 7 10-8 - 10-7 8 <10-8 9

iv. Way of documenting the HAZOP results

It is estimated that severe structural damage and 15 % chance of fatality outdoors or 50 % chance indoor will be experienced out to 20 m and almost complete destruction of all ordinary structures and 100 % chance of fatality indoors to 10 m distance.[8] (see Cameron 2005, p. 268).

iii. Classification of frequencies of risk related consequences:

The frequency of occurring for each possible consequence (hazard) has to be expressed by a frequency class, called (F) according to the scoring chart for frequency (Table 3.4). Definition of frequency class may be supported by Event Tree and/or Fault tree analysis techniques or Layer of protection analysis (LOPA) or historical databases.

**For establishing frequency class:** Estimation / calculation of frequency of vapour cloud explosion and fatality of person because of release of material due to catastrophic rupture of distillation column.


Frequency of catastrophic rupture of column = 10-6 (Taken from table 4)

Table 4. Scoring chart for frequency [3.4]

Probability of ignition of released material = 0.10 (CCPs 2000, Borysiewich 2004)

Probability of VCE if released material ignited = 0.01 (CCPs 2000, Borysiewich 2004)

Probability of fatality of a person exposed to overpressure of 40 kPa due to VCE = 0.20 (Tweeddale 2003, p. 117 Figure 5-14)

Then, frequency of vapour cloud explosion = 6 9 10 0.10 0.01 10

So frequency class for this scenario = 9

336 Advances in Chemical Engineering

It is estimated that severe structural damage and 15 % chance of fatality outdoors or 50 % chance indoor will be experienced out to 20 m and almost complete destruction of all ordinary structures and 100 % chance of fatality indoors to 10 m distance.[8] (see Cameron 2005, p. 268).

The frequency of occurring for each possible consequence (hazard) has to be expressed by a frequency class, called (F) according to the scoring chart for frequency (Table 3.4). Definition of frequency class may be supported by Event Tree and/or Fault tree analysis techniques or

**For establishing frequency class:** Estimation / calculation of frequency of vapour cloud explosion and fatality of person because of release of material due to catastrophic rupture of

**Frequency of occurring incident** 

**1/y Comprehension Examples based on general data bases** 

diameter > 150 mm

Double wall tank leakage

(excluding piping), storage tank rupture

Operators failure under high stress

9 <10-8 Very very small Catastrophic rupture or leakage of pipe of

6 10-6 - 10-5 Less small Pipe residual failure, 100 m full breach,

5 10-5 - 10-4 Moderate Process vessel leakage of 1 mm diameter 4 10-4 - 10-3 Less moderate Pump leakage , Heat exchanger leakage

3 10-3 - 10-2 Less high Safety valve open spuriously, Large external fire 2 10-2 - 10-1 High Cooling water failure, BPCS instrument loop failure 1 10-1 – 100 Very high Operator failure, Regulator failure , Solenoid valve failure

\* >100 Very very high Power failure in developing countries,

Probability of fatality of a person exposed to overpressure of 40 kPa due to VCE = 0.20

Probability of ignition of released material = 0.10 (CCPs 2000, Borysiewich 2004) Probability of VCE if released material ignited = 0.01 (CCPs 2000, Borysiewich 2004)

8 10-8 - 10-7 Very small Catastrophic rupture of pipe of diameter <sup>50</sup> mm 7 10-7 - 10-6 Small Catastrophic rupture of fractionating system

iii. Classification of frequencies of risk related consequences:

Layer of protection analysis (LOPA) or historical databases.

Frequency of catastrophic rupture of column = 10-6 (Taken from table 4)

distillation column.

**Frequency** 

Table 4. Scoring chart for frequency [3.4]

(Tweeddale 2003, p. 117 Figure 5-14)

**Class** 

Frequency of fatality of a person exposed to VCE = <sup>9</sup> <sup>10</sup> 10 0.20 2.10

So frequency class for this scenario = 9

iv. Way of documenting the HAZOP results

The Extended HAZOP methodology worksheet for documenting the HAZOP team results is shown in Figure 5. Below consequence the physical effects and risk has to be documented first and next risk has to be classified using score charts (Table 3.2) related to financial, environment and health related consequences. The worst score of each risk has to be documented. For each risk related consequence, frequency class has also to be established.


Fig. 5. Risk potential matrix (Extended HAZOP decision matrix)

#### **Illustrative Example 2**

Release of material to atmosphere from vent line or vapour line may disperse safely or has toxic effects or can lead to several outcomes such as flash fire, vapour cloud explosion and fire balls. So documenting consequence class in HAZOP work sheet, the score '8' of the most severe consequence will be documented.

Systematic Framework for Multiobjective Optimization in Chemical Process Plant Design 339

Fig. 6. Application of HAZOP decision matrix in Extended HAZOP


1- Short cut calculations 2-Dynamic simulation 3-deterministic models 4- Event tree 5- Fault tree 6- Historic data base

#### **2.4.3 Step 3: Safety/risk assessment - Risk potential matrix (HAZOP decision matrix)**

Figure 6 shows the risk potential matrix (HAZOP decision matrix) used for order of magnitude ranking of events. The rows of the matrix consider frequency class, while the columns show the consequence class. Each cell in the matrix represents a risk category. For the decision process, the matrix is divided into four risk category levels.

Risk level I --- red area --- scenario in this level is intolerable and immediate action (pant or process modification) is needed to reduce that risk category or more detailed quantified analysis has to be carried out in order to find arguments for wrong preliminary decisions.

Risk level II --- grey area --- scenario in this level is tolerable but not acceptable for long period of time so action at next schedule maintenance is needed to reduce that risk category.

Risk level III --- yellow area --- scenario in this level is acceptable and any action to reduce that risk category is optional.

Risk level IV--- green area --- scenario in this level needs no action.

Risk potential matrix (HAZOP decision matrix) may be also used for:


The application of risk potential matrix (HAZOP decision matrix) in the Extended HAZOP is shown in figure 5. Arrows show the transformation of entries from the Extended HAZOP worksheet to the HAZOP decision matrix. The identity number (ID) of each scenario of the Extended HAZOP worksheet is placed in HAZOP decision matrix. Recommended actions for this scenario will be placed from Extended HAZOP sheet to the bottom of HAZOP decision matrix. First HAZOP decision matrix will shows the existing status and second HAZOP decision matrix shows the improved plant status after recommended actions.

Fig. 6. Application of HAZOP decision matrix in Extended HAZOP

Consequences

Physical effects:

1- Short cut calculations 2-Dynamic simulation 3-deterministic models 4- Event tree 5- Fault tree

**2.4.3 Step 3: Safety/risk assessment - Risk potential matrix (HAZOP decision matrix)**  Figure 6 shows the risk potential matrix (HAZOP decision matrix) used for order of magnitude ranking of events. The rows of the matrix consider frequency class, while the columns show the consequence class. Each cell in the matrix represents a risk category. For

Risk level I --- red area --- scenario in this level is intolerable and immediate action (pant or process modification) is needed to reduce that risk category or more detailed quantified analysis has to be carried out in order to find arguments for wrong preliminary decisions. Risk level II --- grey area --- scenario in this level is tolerable but not acceptable for long period of time so action at next schedule maintenance is needed to reduce that risk category. Risk level III --- yellow area --- scenario in this level is acceptable and any action to reduce

The application of risk potential matrix (HAZOP decision matrix) in the Extended HAZOP is shown in figure 5. Arrows show the transformation of entries from the Extended HAZOP worksheet to the HAZOP decision matrix. The identity number (ID) of each scenario of the Extended HAZOP worksheet is placed in HAZOP decision matrix. Recommended actions for this scenario will be placed from Extended HAZOP sheet to the bottom of HAZOP decision matrix. First HAZOP decision matrix will shows the existing status and second HAZOP decision matrix shows the improved plant status after recommended actions.

Risk related:

Document: Page : Date :

Actions FC Resp./

Ref.

FC Recommended

Process:

Function:

Possible causes

the decision process, the matrix is divided into four risk category levels.

Risk level IV--- green area --- scenario in this level needs no action. Risk potential matrix (HAZOP decision matrix) may be also used for:

 Documentation of the status of the plant safety Selection and development of optimization proposals

Documentation of improvement achieved

Detection/ Safeguards

Plant/P&ID : Equipment : Volume :

Guide word

6- Historic data base

that risk category is optional.

Importance of improvement

No

/ Process Parameter

Systematic Framework for Multiobjective Optimization in Chemical Process Plant Design 341

 FCI1 = Fixed capital investment using either cost equations that have been derived by Ulrich or correlations developed by Guthrie

 OC1= Operating cost (including both direct (e.g. raw material, utilities etc.) and indirect

normally d is taken 0.15-25 but can also be computed using depreciation calculation

costs (e.g. taxes, overhead cost etc)

depending on users choice

TAC1 = total annualized cost

= d (FCI1) OC1

methods

equipment [13-15].

**Standard cost calculations Economic module used in this work** 

users choice

cost etc)

risk cost

cost

Table 5. Elements of economic module and difference from standard cost calculations

are used for the estimation of installed cost of the process units in the chemical plant.

instrumentation and control systems, buildings and structures, and auxiliary facilities such as utilities, land and civil engineering work. Several capital cost estimate methods ranging from order of magnitude estimate (ratio estimate) to detailed estimate (contractors estimate)

The most commonly used method that provides estimates within 20-30% of actual cost and widely used at design stage involve the usage of cost charts/correlations (Guthrie's article (1969) and book (1974), chapter 5 of Ulrich's 'A guide to chemical engineering process design and economics' (1984), 'Plant design and economics for chemical engineers' by Peters and Timmerhause (1991)) for estimating the purchase cost of major type of process

These cost charts / correlations were assembled in the 1960's or earlier and are projected to the date of installation using cost indices or escalation factors such as the chemical engineering plant cost index (published biweekly by chemical engineering magazine),

 FCI1= Fixed capital investment using either cost equations that have been derived by Ulrich or correlations developed by Guthrie depending on

 FCI2= FCI1 + fixed capital investment related to safety system (FCISS) OC1= Operating cost (including both direct (e.g. raw material, utilities etc.) and indirect costs (e.g. taxes, overhead

normally d is taken 0.15-25 but can also be computed using depreciation

 RC1 = risk cost 1= Asset risk cost + health risk cost + environmental risk cost RC2 = risk cost2 = RC1+ production loss

RC3 = risk cost3 = process interruption

TRC = total risk cost = RC2+RC3

ECC = Extended costs

TAC1 = total annualized cost

TAC2 = total annualized cost

= d (FCI1) OC1

= d (FCI2) OC1

calculation methods Extended cost calculations

Similarly all results from Extended HAZOP worksheets are transferred to the HAZOP decision matrix. Keeping in view the risk target and depending on the scenario or recommended actions during the Extended HAZOP discussion, analysis team may reach a safety related modification proposal. Next, if safety/ risk optimization is in focus then weak points/scenarios with similar risk are clustered after analyzing HAZOP decision matrix and safety related optimization proposals are developed.

#### **2.4.4 Step 4: Safety/risk system optimization (Development and analysis of optimization proposals)**

In this step, safety related optimization proposals are generated and evaluated using dynamic simulation, Event tree analysis and/or Fault tree analysis. The optimization proposals can be developed at two levels:


The relevant information such as frequency and damage data will be transferred to economic module for safety related cost calculations and multiobjective decision making (if more than one alternatives developed).

#### **2.5 Economic module**

In all stages of design process, economic evaluation is crucial for the evaluation of process alternatives. Various objective functions are available in the literature of chemical engineering economics for economic evaluation of chemical processes. Some quite elegant objective functions, which incorporate the concept of the "time value of money", are net present value (NPV) and discounted cash flow. Business managers, accountants and economists prefer these methods because they are more accurate measures of profitability over an extended time period. However, application of these methods needs certain assumptions[12]. Total annualized cost (TAC) can be used as economic indicators/objective function for the evaluation of design alternatives and economic optimization.

Economic module developed in Visual Basic consists of two distinct sections. First section carries out standard cost calculations (i.e. Fixed capital investment (FC1) and operational cost (OC1)) and compute total annualized cost (TAC1) while second section carries out extended cost calculations i.e. process safety/risk related costs and computes the fixed capital investment related to safety system (FCISS), accident and incident damage related risk cost. Table 3.5 illustrates the difference of cost elements considered in standard practice of cost calculations of chemical process design and in this economic module. Figure 7 shows the simplified block diagram of economic module.

#### **2.5.1 Standard cost calculations**

Standard cost calculations involves fixed capital investment (FCI1) and operational cost (OC1). Fixed capital investment (FCI1) includes the cost of design and other engineering and construction supervision, all items of equipment and their installation, all piping,

Similarly all results from Extended HAZOP worksheets are transferred to the HAZOP decision matrix. Keeping in view the risk target and depending on the scenario or recommended actions during the Extended HAZOP discussion, analysis team may reach a safety related modification proposal. Next, if safety/ risk optimization is in focus then weak points/scenarios with similar risk are clustered after analyzing HAZOP decision matrix and

In this step, safety related optimization proposals are generated and evaluated using dynamic simulation, Event tree analysis and/or Fault tree analysis. The optimization

Simple optimization proposals e.g. addition of pressure alarm or change of location of

Optimization proposals related to severe scenarios by evaluating risk potential matrix

The relevant information such as frequency and damage data will be transferred to economic module for safety related cost calculations and multiobjective decision making (if

In all stages of design process, economic evaluation is crucial for the evaluation of process alternatives. Various objective functions are available in the literature of chemical engineering economics for economic evaluation of chemical processes. Some quite elegant objective functions, which incorporate the concept of the "time value of money", are net present value (NPV) and discounted cash flow. Business managers, accountants and economists prefer these methods because they are more accurate measures of profitability over an extended time period. However, application of these methods needs certain assumptions[12]. Total annualized cost (TAC) can be used as economic indicators/objective

Economic module developed in Visual Basic consists of two distinct sections. First section carries out standard cost calculations (i.e. Fixed capital investment (FC1) and operational cost (OC1)) and compute total annualized cost (TAC1) while second section carries out extended cost calculations i.e. process safety/risk related costs and computes the fixed capital investment related to safety system (FCISS), accident and incident damage related risk cost. Table 3.5 illustrates the difference of cost elements considered in standard practice of cost calculations of chemical process design and in this economic module. Figure 7 shows

Standard cost calculations involves fixed capital investment (FCI1) and operational cost (OC1). Fixed capital investment (FCI1) includes the cost of design and other engineering and construction supervision, all items of equipment and their installation, all piping,

function for the evaluation of design alternatives and economic optimization.

**2.4.4 Step 4: Safety/risk system optimization (Development and analysis of** 

safety related optimization proposals are developed.

sensor within the Extended HAZOP discussion

proposals can be developed at two levels:

(HAZOP decision matrix)

**2.5 Economic module** 

more than one alternatives developed).

the simplified block diagram of economic module.

**2.5.1 Standard cost calculations** 

**optimization proposals)** 


Table 5. Elements of economic module and difference from standard cost calculations

instrumentation and control systems, buildings and structures, and auxiliary facilities such as utilities, land and civil engineering work. Several capital cost estimate methods ranging from order of magnitude estimate (ratio estimate) to detailed estimate (contractors estimate) are used for the estimation of installed cost of the process units in the chemical plant.

The most commonly used method that provides estimates within 20-30% of actual cost and widely used at design stage involve the usage of cost charts/correlations (Guthrie's article (1969) and book (1974), chapter 5 of Ulrich's 'A guide to chemical engineering process design and economics' (1984), 'Plant design and economics for chemical engineers' by Peters and Timmerhause (1991)) for estimating the purchase cost of major type of process equipment [13-15].

These cost charts / correlations were assembled in the 1960's or earlier and are projected to the date of installation using cost indices or escalation factors such as the chemical engineering plant cost index (published biweekly by chemical engineering magazine),

Systematic Framework for Multiobjective Optimization in Chemical Process Plant Design 343

Marshall and Swift Index (also provided in chemical engineering magazine) and Nelson-Fabaar Index (from the oil and gas journal). For the comparison of process design alternatives, these study estimates for purchased cost of process units using cost charts or equations based on them are adequate. Given the purchase cost of a process unit, the installed cost is obtained by adding the cost of installation using factored-cost methods. For each piece of equipment Guthrie (1969, 1974) provides factors to estimate the direct cost of labor, as well as, indirect costs involved in the installation procedure. The cost elements that

are included in the estimation of fixed capital investment are shown in figure 8.

f.o.b ---- freight on board cost i.e. equipment purchase cost plus transport cost

The operating cost (OC1) of a chemical plant is divided into two groups:

The elements in fixed operating cost includes maintenance cost, operating labor cost, laboratory cost, supervision cost, plant overheads, capital charges, taxes, insurance, licence fees and royalty payments while the variable operating cost consists of raw material costs, miscellaneous operating material costs, utilities (services) and shipping and packaging. However this division of operating cost is somewhat arbitrary and depends on the accounting practice of a particular organization. The typical cost elements included in

However, from the existing process optimization point of view, energy cost and raw

Economic module developed in this thesis using Visual basic computes fixed capital expenditure using either cost equations that have been derived by Ulrich or correlations

Fig. 8. Typical cost elements for fixed capital investment

operating cost "OC1" are shown in figure 9.

material costs are more important and often considered.

 Fixed operating cost Variable operating cost

#### **Note:**

In Extended cost calculations, the costs such as insurance cost, market loss cost, loss of image and prestige cost should also be considered in addition. But in this module these costs are not included.

Fig. 7. Simplified block diagram of economic module

Marshall and Swift Index (also provided in chemical engineering magazine) and Nelson-Fabaar Index (from the oil and gas journal). For the comparison of process design alternatives, these study estimates for purchased cost of process units using cost charts or equations based on them are adequate. Given the purchase cost of a process unit, the installed cost is obtained by adding the cost of installation using factored-cost methods. For each piece of equipment Guthrie (1969, 1974) provides factors to estimate the direct cost of labor, as well as, indirect costs involved in the installation procedure. The cost elements that are included in the estimation of fixed capital investment are shown in figure 8.

f.o.b ---- freight on board cost i.e. equipment purchase cost plus transport cost

Fig. 8. Typical cost elements for fixed capital investment

The operating cost (OC1) of a chemical plant is divided into two groups:

Fixed operating cost

342 Advances in Chemical Engineering

Fixed capital cost

> Operating cost

Conventional safety cost

damage,frequency

Incident damage,frequency

interruption cost Incident damage risk cost(RC3)

In Extended cost calculations, the costs such as insurance cost, market loss cost, loss of image and prestige cost should also be considered in addition. But in this module these

Ulrich's cost models

Guthrie 's correlations

Hot & Cold utility cost

Raw material cost

Safety classification

Safety design cost Process control measure cost Add on safety cost

Asset loss

Human health

Environmental damage cost Loss of production cost

Process

cost

loss cost Accident

Standard cost calculations

Extended cost calculations

Fixed capital investment (FCI1)

Operating cost (OC1)

Total annualized cost (TAC1)

Fixed safety system cost (FCISS)

Fixed capital investment (FCI2)

Total annualized cost (TAC2)

Accident damage risk cost (RC2)

Extended cost calculations (ECC)

Fig. 7. Simplified block diagram of economic module

**Note:**

costs are not included.

Variable operating cost

The elements in fixed operating cost includes maintenance cost, operating labor cost, laboratory cost, supervision cost, plant overheads, capital charges, taxes, insurance, licence fees and royalty payments while the variable operating cost consists of raw material costs, miscellaneous operating material costs, utilities (services) and shipping and packaging. However this division of operating cost is somewhat arbitrary and depends on the accounting practice of a particular organization. The typical cost elements included in operating cost "OC1" are shown in figure 9.

However, from the existing process optimization point of view, energy cost and raw material costs are more important and often considered.

Economic module developed in this thesis using Visual basic computes fixed capital expenditure using either cost equations that have been derived by Ulrich or correlations

Systematic Framework for Multiobjective Optimization in Chemical Process Plant Design 345

Here, the first term *CSD* is cost for safety design (i.e. cost related to safety classification, safety requirements and design specification, detailed design and engineering, factory acceptance test or pre-start up acceptance test and start up and correction). Table 3.6 gives

sum of the purchase cost of safety equipment. Here CSE,i is the purchase cost of equipment "i" and NSE,i is the number (count) of that equipment. The costs for these devices are based on the recent detailed survey of available costs from various suppliers conducted by Khan and Annyotte (2004), however in the module the user has the possibility to enter the present

Safety classification cost e.g SIL determination cost **C**SIL Safety requirements and design specifications (SRS) cost **C**SRS Detailed design and engineering cost **C**DE !Miscellaneous Cost: **C**ME Initial training cost **C**TC

(PSAT) cost **C**FAT

Startup and correction cost **C**SCC

*CCC CC CC C SD SIL SRS DE ME TC FAT SCC*

21 *FCI FCI FCISS* (3)

So, the extended total annualized cost will be calculated using extended fixed capital

Maintenance and repair cost of safety system should also be included in this calculation. But

Risk cost (RC1), which is the sum of property risk cost due to asset loss (PRC), health risk cost due to human health loss (HRC) and environmental risk cost due to environmental

damage (ERC). The relations for calculation of these costs used in the module are:

TAC2 d (FCI2) OC1 (4)

Factory acceptance test (FAT)/Installation/Pre-startup acceptance test

Table 6. Typical cost elements included in *CSD* of safety system cost calculations

Then, extended fixed capital investment is calculated by adding FCISS to FCI1.

in this economic module these cost elements are not considered.

!power, wiring, junction boxes, operators interface cost

ii. Fixed capital investment (FCI2)

iii. Total annualized cost (TAC2)

investment.

iv. Risk cost 1 (RC1)

1

*i*

*SE i SE i*

is the

*N C*

*n*

the typical cost elements included in *CSD* calculations. The second term , ,

market costs.

Fig. 9. Typical cost elements for operating cost (OC1)

developed by Guthrie depending on users choice, The significant operating cost for process optimization i.e. energy consumption cost (heating and cooling utilities cost) can also be calculated by using this module. Once the FCI1 and OC1 are calculated, then total annualized cost is obtained using the following equation:

$$\text{Total annualized cost (TAC1)} \ = \text{d} \cdot (\text{FCI1}) + \text{OCI} \tag{1}$$

Here d is depreciation or capital recovery factor and normally taken between 0.15-0.25 but can also be computed using depreciation calculation methods e.g. double declining balance method.

#### **2.5.2 Extended cost calculations**

The second section of the economic module (see Figure 7) carries out Extended cost calculations, which considers the fixed capital investment related to safety system, and risk cost due to accident and incident damage.

i. Fixed capital investment related to Safety system:

The fixed investment related to safety system is calculated by the following equation:

$$FCISS = \mathbf{C}\_{SD} + \sum\_{i=1}^{n} \mathbf{N}\_{SE,i} \cdot \mathbf{C}\_{SE,i} \tag{2}$$

Here, the first term *CSD* is cost for safety design (i.e. cost related to safety classification, safety requirements and design specification, detailed design and engineering, factory acceptance test or pre-start up acceptance test and start up and correction). Table 3.6 gives

the typical cost elements included in *CSD* calculations. The second term , , 1 *n SE i SE i i N C* is the

sum of the purchase cost of safety equipment. Here CSE,i is the purchase cost of equipment "i" and NSE,i is the number (count) of that equipment. The costs for these devices are based on the recent detailed survey of available costs from various suppliers conducted by Khan and Annyotte (2004), however in the module the user has the possibility to enter the present market costs.


!power, wiring, junction boxes, operators interface cost

Table 6. Typical cost elements included in *CSD* of safety system cost calculations

ii. Fixed capital investment (FCI2)

344 Advances in Chemical Engineering

developed by Guthrie depending on users choice, The significant operating cost for process optimization i.e. energy consumption cost (heating and cooling utilities cost) can also be calculated by using this module. Once the FCI1 and OC1 are calculated, then total

Here d is depreciation or capital recovery factor and normally taken between 0.15-0.25 but can also be computed using depreciation calculation methods e.g. double declining balance

The second section of the economic module (see Figure 7) carries out Extended cost calculations, which considers the fixed capital investment related to safety system, and risk

1

*n SD SE i SE i i FCISS C N C* 

The fixed investment related to safety system is calculated by the following equation:

Total annualized cost (TAC1) d (FCI1) OC1 (1)

, ,

(2)

Fig. 9. Typical cost elements for operating cost (OC1)

annualized cost is obtained using the following equation:

method.

**2.5.2 Extended cost calculations** 

cost due to accident and incident damage.

i. Fixed capital investment related to Safety system:

Then, extended fixed capital investment is calculated by adding FCISS to FCI1.

$$\text{FCI} \mathbf{2} = \text{FCI} \mathbf{1} + \text{FCI} \mathbf{SS} \tag{3}$$

iii. Total annualized cost (TAC2)

So, the extended total annualized cost will be calculated using extended fixed capital investment.

$$\text{TAC2} = \text{d} \cdot (\text{FCI2}) + \text{OC1} \tag{4}$$

Maintenance and repair cost of safety system should also be included in this calculation. But in this economic module these cost elements are not considered.

$$\text{iv. Risk cost 1 (RC1)}$$

Risk cost (RC1), which is the sum of property risk cost due to asset loss (PRC), health risk cost due to human health loss (HRC) and environmental risk cost due to environmental damage (ERC). The relations for calculation of these costs used in the module are:

Systematic Framework for Multiobjective Optimization in Chemical Process Plant Design 347

*RC RC PLRC* 2 1

Risk cost 3 (RC3), which is sum of process interruption cost due the spurious trip of the safety system and process interruption cost because of safe shut down to avoid from

> 3 ( ) *<sup>n</sup> trip trip*

*RC F t F t C t*

demand arises, ttrip is down time due to spurious trip and tdR is down time to safe shut

23 *TRC RC RC* (11)

*TAC TRC t risk* / *op*

2 *ECC FCI PVC* (12)

1 (1 ) ( 1 cos )

Besides, the cost elements mentioned above in Extended cost calculation section, the other elements such as warranty/insurance cost, lost of image and prestige cost, market lost cost should also be considered but quantification of these elements is still almost impossible.

Environment module consists of four steps and introduced an environmental performance index (EPI1) for evaluation of environmental performance and environmental pollution index (EPI2) as environmental objective to be integrated along with economics. The environmental performance index (EPI1) is calculated by combining total PEI based on

*S trip R dR o p p*

(10)

is safe shut down frequency when trip system

*ly t*

*R*  is the

Where PLRC is the production loss risk cost, td is the time lost due to accident and *Cp*

accident, accounts for incident damage risk cost and calculated as follow:

1

Total risk cost can be annualized by dividing it with total operation time ( *op t* ):

Extended cost calculations (ECC) is Life cycle related cost and calculated as follow :

Here, PVC is present value of the annual costs (OC1, *TACrisk* ) and calculated as follow:

*risk <sup>R</sup> PVC OC TAC Insurance t*

R is the present interest rate and tly is the number of years (predicted life of system).

*i*

production loss value in \$/h. Thus Risk cost 2 (RC2) is

is spurious trip frequency, *trip FR*

Total Risk cost (TRC) is the sum of all risk costs:

down when trip system demand arises.

vii. Total Risk cost (TRC)

viii. Extended Cost (ECC)

**2.6 Environment module** 

vi. Risk cost 3 (RC3)

Here, *trip SF*

Property risk cost due to asset lost (PRC) is the cost incurred due to lost of physical assets such as damage to property, loss of equipment due to accident/scenario and calculated by the equation below:

$$PRC = \sum\_{i=1}^{n} \dot{F}\_{A,i} \cdot A\_{D,i} \cdot C\_{A,i} \cdot t\_{op} + \sum\_{j=1}^{n} \dot{F}\_{I,j} \cdot C\_{D,j} \cdot t\_{op} \tag{5}$$

*FA i*, is frequency of occurring the hazardous accident, *AD i*, is damage area due to that accident, *CA*,*<sup>i</sup>* is the asset cost per unit area, *op <sup>t</sup>* is total operation time, *FI*, *<sup>j</sup>* is incident occurring frequency and *CD*, *<sup>j</sup>* is incident damage cost.

Health risk cost due to human health lost (HRC) is the cost of fatality and/or injury due to the accident scenario under study.

$$HRC = \sum\_{i=1}^{n} \dot{F}\_{A,i} \cdot N\_{\text{Peap},eff} \cdot \mathbb{C}\_{H, \text{life}} \cdot t\_{op} \tag{6}$$

Here, *Npeop*,*eff* is the number of person affected due to accident and is equal to *Npeop eff* , *pop* . Where *POP* is the population around the area of accident and is the population distribution factor (is 1 if population is uniform distributed (maximum value) and is 0.2 if population is localized and away from the area of accident (minimum value)) and *CH li* , *fe* is dollar value of human life or health. Though attempts to put value on human life have caused criticism and it changes from place to place. But a value for this can be obtained by dividing the annual gross national product by the annual number of births or by estimating how much money the person would have earned if not killed by the accident (Tweeddale 2003). A value for cost of loss of lives, marginal cost to avert the fatality, for the highest category of involuntariness risk 14 x 106 \$ is used in this work (Passman, H.J. et al. 2003).

Environmental risk cost due to environmental damage is the cost incurred due to environmental damage.

$$ERC = \sum\_{i=1}^{n} \dot{F}\_{ED} \cdot A\_{ED,i} \cdot \mathbf{C}\_{ED,i} \cdot t\_{op} \tag{7}$$

Where *AED i*, is the environmental damage area due to scenario "i", *FED* is the frequency of release of material to environment and *CED i*, is the environmental damage cost per unit area.

so the sum of these three risk costs gives:

$$R\text{C1} = \text{PRC} + \text{HRC} + \text{ERC} \tag{8}$$

#### v. Risk cost 2 (RC2)

Risk cost 2 (RC2), which is the sum of risk cost1 (RC1) and production loss risk cost (PLRC), accounts for accident damage risk cost. Here, production loss risk cost due to asset damage (PLRC) accounts for the cost due to the production loss because of accident and given by:

$$PLRC = \sum\_{i=1}^{n} \dot{F}\_{A,i} \cdot t\_d \cdot \dot{C}\_p \cdot t\_{op} \tag{9}$$

Where PLRC is the production loss risk cost, td is the time lost due to accident and *Cp* is the production loss value in \$/h. Thus Risk cost 2 (RC2) is

$$RC\,2 = RC\,1 + PLR\,C$$

#### vi. Risk cost 3 (RC3)

346 Advances in Chemical Engineering

Property risk cost due to asset lost (PRC) is the cost incurred due to lost of physical assets such as damage to property, loss of equipment due to accident/scenario and calculated by

1 1

*FA i*, is frequency of occurring the hazardous accident, *AD i*, is damage area due to that accident, *CA*,*<sup>i</sup>* is the asset cost per unit area, *op <sup>t</sup>* is total operation time, *FI*, *<sup>j</sup>* is incident

Health risk cost due to human health lost (HRC) is the cost of fatality and/or injury due to

*HRC F N C t*

Here, *Npeop*,*eff* is the number of person affected due to accident and is equal to *Npeop eff* , *pop* . Where *POP* is the population around the area of accident and is the population distribution factor (is 1 if population is uniform distributed (maximum value) and is 0.2 if population is localized and away from the area of accident (minimum value)) and *CH li* , *fe* is dollar value of human life or health. Though attempts to put value on human life have caused criticism and it changes from place to place. But a value for this can be obtained by dividing the annual gross national product by the annual number of births or by estimating how much money the person would have earned if not killed by the accident (Tweeddale 2003). A value for cost of loss of lives, marginal cost to avert the fatality, for the highest category of involuntariness risk 14 x 106 \$ is used in this work (Passman, H.J. et al. 2003).

Environmental risk cost due to environmental damage is the cost incurred due to

*ERC F A C t*

release of material to environment and *CED i*, is the environmental damage cost per unit area.

1 *RC PRC HRC ERC* (8)

Risk cost 2 (RC2), which is the sum of risk cost1 (RC1) and production loss risk cost (PLRC), accounts for accident damage risk cost. Here, production loss risk cost due to asset damage (PLRC) accounts for the cost due to the production loss because of accident and given by:

> , 1

*Ai d o p p*

*n*

*i PLRC F t C t* 

1

*i*

Where *AED i*, is the environmental damage area due to scenario "i", *FED*

*n*

, ,

*ED ED i ED i op*

, ,,

*A i Peop eff H life op*

*i j PRC F A C t F C t* 

1

*i*

*n*

occurring frequency and *CD*, *<sup>j</sup>* is incident damage cost.

the accident scenario under study.

environmental damage.

v. Risk cost 2 (RC2)

so the sum of these three risk costs gives:

*n n*

, , , ,,

*A i Di Ai op I j D j op*

(5)

(6)

(7)

(9)

is the frequency of

the equation below:

Risk cost 3 (RC3), which is sum of process interruption cost due the spurious trip of the safety system and process interruption cost because of safe shut down to avoid from accident, accounts for incident damage risk cost and calculated as follow:

$$\text{RCCS} = \left(\sum\_{i=1}^{n} \dot{F}\_{S}^{trip} \cdot t\_{trip} + \dot{F}\_{R}^{trip} \cdot t\_{dR}\right) \cdot \dot{C}\_{p} \cdot t\_{op} \tag{10}$$

Here, *trip SF* is spurious trip frequency, *trip FR* is safe shut down frequency when trip system demand arises, ttrip is down time due to spurious trip and tdR is down time to safe shut down when trip system demand arises.

#### vii. Total Risk cost (TRC)

Total Risk cost (TRC) is the sum of all risk costs:

$$\text{TRC} = \text{RC} \text{2} + \text{RC} \text{3} \tag{11}$$

Total risk cost can be annualized by dividing it with total operation time ( *op t* ):

$$TAC\_{risk} = T\mathcal{RC} / \, t\_{op}$$

viii. Extended Cost (ECC)

Extended cost calculations (ECC) is Life cycle related cost and calculated as follow :

$$\text{ECC} = \text{FCI2} + \text{PVC} \tag{12}$$

Here, PVC is present value of the annual costs (OC1, *TACrisk* ) and calculated as follow:

$$PVC = \left(OC1 + TAC\_{risk} - \text{Insumance} \cos t\right) \cdot \frac{1 - \left(1 + R\right)^{-t\_{ly}}}{R}$$

R is the present interest rate and tly is the number of years (predicted life of system).

Besides, the cost elements mentioned above in Extended cost calculation section, the other elements such as warranty/insurance cost, lost of image and prestige cost, market lost cost should also be considered but quantification of these elements is still almost impossible.

#### **2.6 Environment module**

Environment module consists of four steps and introduced an environmental performance index (EPI1) for evaluation of environmental performance and environmental pollution index (EPI2) as environmental objective to be integrated along with economics. The environmental performance index (EPI1) is calculated by combining total PEI based on

Systematic Framework for Multiobjective Optimization in Chemical Process Plant Design 349

degree of freedom analysis etc.

Generation of alternatives

Energy consumption (Ec)

Fugitive emission (Ef

Resource conservation (RC)

relevant information from process module

)

Total potential environmental impact

Environmental performance index (EPI)

EL Total PEI,Ec,Ef ,Rc

Environmental pollution index

Energy consumption factor refers the total amount of energy consumed in the process per

( )/ *E HM <sup>C</sup> <sup>p</sup>* [kJ / kg product] (14)

the enthalpy of steam per kg [KJ/kg], *EE* is electrical energy consumed per unit time [KJ /

The sensitivity analysis of this factor with respect to optimization variables should also be

Here *H* = <sup>ˆ</sup> *Msteam steam E h E* where *Msteam* is the mass flow rate of steam [kg/h], <sup>ˆ</sup>

 Definition of scope of study i.e. evaluation or optimization Process study and analysis i.e identification of additional waste and emission streams, identification of fugitive emission sources,

Collection of additional environmental related data and retrieval of

Individuals potential environmental impacts calculation (PEIL)

Level-3

Fig. 10. Simplified block diagram of environment module

Level-2 Level-1

**I- Problem definition & Data gathering**

**II- Calculations of individual impact**

**III- Determination of weighting factors**

**IV- Environmental performance & pollution index calculation**

**Energy consumption factor (** *EC* **)** 

h] and *Mp* is product rate [kg/h].

performed.

unit of product and is calculated as follow:

**categories**

Environmental performance index (EPI)

<sup>C</sup> Mp E (H)/

sources s Ef ( <sup>M</sup> <sup>s</sup> <sup>x</sup> )/ <sup>M</sup>

<sup>C</sup> <sup>u</sup> RM Mp R (M M )/

Comps k PEIL (M <sup>b</sup> xkb <sup>Q</sup> )/ <sup>M</sup>

v,s p

<sup>p</sup> <sup>E</sup> kL <sup>r</sup> <sup>L</sup>

EnvCat L Total PEI WL PEI

> EPI1 <sup>1</sup> ( <sup>W</sup> <sup>E</sup> ) <sup>L</sup> EnvCat L <sup>L</sup>

Total PEI Ef EPI2

L

*hsteam* is

Ef PEI EC RC

HTPI HTPE TTP ATP GWP ODP PCOP AP

WAR algorithm[16,17], resource depletion, energy conservation and fugitive emission rates while environmental pollution index (EPI2) is calculated by combining total PEI based on WAR algorithm and fugitive emissions because in this case other factors like resource depletion and energy consumption will be integrated in economic module or objective function. The Analytic hierarchy process (AHP) is used as multicriteria decision analysis tool for combining these different impacts and determination of weighting factors of individual impact categories in total PEI and later on in environmental performance index (EPI1) and environmental pollution index (EPI2) calculations. The module is developed using Microsoft Visual Basic 6.0 and WAR GUI (WAR graphical user interface) is integrated in the user plate form. The steps are:

Step I : Problem definition and data gathering

Step II : Individual impact categories calculation

Step III: Determination of weighting factors

Step IV: Environmental performance index calculation

Figure 10 shows the simplified block diagram of environment module and tasks to be performed.

#### **2.6.1 Step I: Problem definition and data gathering**

The primary task in step 1 is problem framing and scope definition. Information such as material and energy balance information, process conditions, process technology and nature of used materials/chemicals should be retrieved from process module. Process flow diagram is to be re-examined for identification of additional waste and emission streams. Collect additional data and information for environment evaluation to fill gaps. As sources of emissions such as fugitive emission sources, venting of equipment, periodic equipment cleaning, incomplete separations etc. are often missing in process so process is analyzed to identify these sources.

#### **2.6.2 Step II: Individual impact categories calculation (Potential environmental impact calculations based on WAR algorithm)**

The software WAR GUI (waste reduction algorithm graphical user interface) from the US Environmental Protection Agency is used to calculate individual potential environmental impacts. The generalized formula based on WAR algorithm for calculating individual PEI is given in equation 13.

$$PEI\_L = \left(\dot{M}\_b \cdot \sum\_k^{\text{Comp}} \mathbf{x}\_{kb} \cdot \boldsymbol{\nu}\_{kL} + \dot{Q}\_r \cdot \boldsymbol{\nu}\_L^E\right) / \dot{M}\_p \qquad \text{[Impact / kg product]}\tag{13}$$

Where *PEIL* is the potential environmental impact of category L, *Mb* is mass flow rate of base (effluent) stream, *kb x* is the mass fraction of component k in the base stream, *kL* is the normalized impact score of chemical k for category L, *Qr* is energy rate supplied for separation and *<sup>E</sup> <sup>L</sup>* is the normalized impact score of category L due to energy. The sensitivity analysis results of individual potential environmental impact with respect to optimization variables should also be performed.

Fig. 10. Simplified block diagram of environment module

#### **Energy consumption factor (** *EC* **)**

348 Advances in Chemical Engineering

WAR algorithm[16,17], resource depletion, energy conservation and fugitive emission rates while environmental pollution index (EPI2) is calculated by combining total PEI based on WAR algorithm and fugitive emissions because in this case other factors like resource depletion and energy consumption will be integrated in economic module or objective function. The Analytic hierarchy process (AHP) is used as multicriteria decision analysis tool for combining these different impacts and determination of weighting factors of individual impact categories in total PEI and later on in environmental performance index (EPI1) and environmental pollution index (EPI2) calculations. The module is developed using Microsoft Visual Basic 6.0 and WAR GUI (WAR graphical user interface) is integrated

Figure 10 shows the simplified block diagram of environment module and tasks to be

The primary task in step 1 is problem framing and scope definition. Information such as material and energy balance information, process conditions, process technology and nature of used materials/chemicals should be retrieved from process module. Process flow diagram is to be re-examined for identification of additional waste and emission streams. Collect additional data and information for environment evaluation to fill gaps. As sources of emissions such as fugitive emission sources, venting of equipment, periodic equipment cleaning, incomplete separations etc. are often missing in process so process is analyzed to

**2.6.2 Step II: Individual impact categories calculation (Potential environmental impact** 

The software WAR GUI (waste reduction algorithm graphical user interface) from the US Environmental Protection Agency is used to calculate individual potential environmental impacts. The generalized formula based on WAR algorithm for calculating individual PEI is

*E*

Where *PEIL* is the potential environmental impact of category L, *Mb* is mass flow rate of

sensitivity analysis results of individual potential environmental impact with respect to

*<sup>L</sup>* is the normalized impact score of category L due to energy. The

*[Impact / kg product]* (13)

*kL* is the

is energy rate supplied for

 

base (effluent) stream, *kb x* is the mass fraction of component k in the base stream,

in the user plate form. The steps are:

performed.

identify these sources.

given in equation 13.

separation and *<sup>E</sup>*

**calculations based on WAR algorithm)** 

 ( ) / *Comps*

optimization variables should also be performed.

*k*

normalized impact score of chemical k for category L, *Qr*

*L b kb kL r L p*

*PEI M x Q M* 

Step I : Problem definition and data gathering Step II : Individual impact categories calculation

Step III: Determination of weighting factors

Step IV: Environmental performance index calculation

**2.6.1 Step I: Problem definition and data gathering** 

Energy consumption factor refers the total amount of energy consumed in the process per unit of product and is calculated as follow:

$$E\_{\mathbb{C}} = (\dot{H}) / \dot{M}\_p \qquad \text{[kJ / kg product]} \tag{14}$$

Here *H* = <sup>ˆ</sup> *Msteam steam E h E* where *Msteam* is the mass flow rate of steam [kg/h], <sup>ˆ</sup> *hsteam* is the enthalpy of steam per kg [KJ/kg], *EE* is electrical energy consumed per unit time [KJ / h] and *Mp* is product rate [kg/h].

The sensitivity analysis of this factor with respect to optimization variables should also be performed.

Systematic Framework for Multiobjective Optimization in Chemical Process Plant Design 351

fugitive sources such as pump seals, valves, flanges and connection is equal to 1, i.e. fluids

**2.6.3 Step III: Determination of weighting factors (Application of multicriteria decision** 

The integration of these individual impact categories into one index is a hierarchical multicriteria decision analysis problem. The analytic hierarchy process (AHP) is used for this purpose[19] and a computer programme for it is developed in VB 6.0. In this stage, first a hierarchical structure of the problem, which is structured hierarchically similar to a flow chart, is constructed. The overall objective is placed at the top while the criteria and subcriteria are placed below. For example, as shown in figure 11, the overall objective Environmental performance index (EPI1) is placed at the top (level 1), then below (level 2) are criterias Total PEI, Ef , Ec and RC and after this (level 3) sub-criterias as HTPI, HTPE, TTP, ATP, GWP, ODP, PCOP and AP. After this using the numerical scale given in table 2.6, two pairwise comparison matrices (see Table 7 and 8) are constructed for determination of weights for aggregation of individual impact categories of WAR to total PEI and for determination of weights of total PEI, Ef, Ec and RC to Environmental performance index

> Environmental performance index (EPI1)

Ef PEI EC RC

HTPI HTPE TTP ATP GWP ODP PCOP AP

Fig. 11. Hierarchical structuring of multicriteria decision analysis problem for integrating

The right hand upper diagonal information in both matrices is to be provided by the decision maker giving the relative importance of the two criteria using the numerical scale of table 2.6 while the left hand lower diagonal is the reciprocal of the right hand upper diagonal. Once these pair wise comparison matrices are constructed, then developed computer programme using the AHP method, determines the weighting of individual impact categories. The level of inconsistency of decision makers input is checked by

are composed entirely of volatile compounds.

**analysis technique)** 

(EPI1).

Level-2

Level-1

Level-3

individual environmental impacts

#### **Resource conservation factor (RC)**

The resource consumption refers all needed raw materials and utilities used and given by:

$$R\_{\mathbb{C}} = (\dot{M}\_u + \dot{M}\_{RM}) / \dot{M}\_p \qquad \text{[kg / kg product]} \tag{15}$$

Where *RC* is the resource conservation factor, *MU* is utilities consumption rate, *MRM* is raw material consumption rate.

#### **Fugitive emission factor (Ef )**

Fugitive emissions are unplanned or unmanaged, continuous or intermittent releases from unsealed sources such as storage tank vents, valves, pump seals, flanges, compressors, sampling connections, open ended lines etc and any other non point air emissions. These sources are large in number and difficult to identify. These emission rates depends on factors such as the age and quality of components, specific inspection and maintenance procedures, equipment design and standards of installation, specific process temperatures and pressures, number and type of sources and operational management commitment[18]. However, four basic approaches for estimating emissions from equipment leaks in a specific processing unit, in order of increasing refinement, in use are:


All these approaches require some data collection, data analysis and/or statistical evaluation. On the other hand, using fundamental design / engineering calculations for accurate fugitive emission estimations for each source present in the process industry are difficult due to:


As focus in this work is to integrate fuggitve emissions into environmental performance evaluation and optimization objectives so average emission factor approach giving a bit over estimates are used. Average emission factors for estimating fugitive emissions from fugitive sources found in synthetic organic chemical manufacturing industries operations (SOCMI) obtained from the US Environmental Protection Agency L & E Databases are used. The relation used in this work for calculation of fugitive emissions is:

$$E\_f = \sum\_{s}^{\text{sources}} \left( \dot{M} \mathbf{s} \cdot \boldsymbol{\xi} \cdot \mathbf{x}\_{v,s} \right) / \dot{M}\_p \text{ [kg/kg product]} \tag{16}$$

Here *Ef* is fugitive emission factor per unit of product, *Ms* mass flow rate through the source 's', is average emission factor and *v s*, *x* is mass fraction of volatile component through source 's' and *Mp* is product rate. It is assumed *v s*, *<sup>x</sup>* for the process fluids through

The resource consumption refers all needed raw materials and utilities used and given by:

( )/ *R MM M C u RM <sup>p</sup>* [kg / kg product] (15)

Where *RC* is the resource conservation factor, *MU* is utilities consumption rate, *MRM* is

Fugitive emissions are unplanned or unmanaged, continuous or intermittent releases from unsealed sources such as storage tank vents, valves, pump seals, flanges, compressors, sampling connections, open ended lines etc and any other non point air emissions. These sources are large in number and difficult to identify. These emission rates depends on factors such as the age and quality of components, specific inspection and maintenance procedures, equipment design and standards of installation, specific process temperatures and pressures, number and type of sources and operational management commitment[18]. However, four basic approaches for estimating emissions from equipment leaks in a specific

All these approaches require some data collection, data analysis and/or statistical evaluation. On the other hand, using fundamental design / engineering calculations for accurate fugitive emission estimations for each source present in the process industry are

 dependence of emission rates on other factors along with design and operating conditions e.g. installation standards, inspection and maintenance procedure etc.

As focus in this work is to integrate fuggitve emissions into environmental performance evaluation and optimization objectives so average emission factor approach giving a bit over estimates are used. Average emission factors for estimating fugitive emissions from fugitive sources found in synthetic organic chemical manufacturing industries operations (SOCMI) obtained from the US Environmental Protection Agency L & E Databases are used.

Here *Ef* is fugitive emission factor per unit of product, *Ms* mass flow rate through the

through source 's' and *Mp* is product rate. It is assumed *v s*, *<sup>x</sup>* for the process fluids through

is average emission factor and *v s*, *x* is mass fraction of volatile component

[kg/ kg product] (16)

processing unit, in order of increasing refinement, in use are:

large number and type of fugitive emission sources

The relation used in this work for calculation of fugitive emissions is:

*sources*

*s*

, ( )/

*f v s p*

*E Ms x M* 

**Resource conservation factor (RC)** 

raw material consumption rate. **Fugitive emission factor (Ef )** 

Average emission factor approach

Unit-specific correlation approach

 Screening ranges approach EPA correlation approach

difficult due to:

source 's',

fugitive sources such as pump seals, valves, flanges and connection is equal to 1, i.e. fluids are composed entirely of volatile compounds.

#### **2.6.3 Step III: Determination of weighting factors (Application of multicriteria decision analysis technique)**

The integration of these individual impact categories into one index is a hierarchical multicriteria decision analysis problem. The analytic hierarchy process (AHP) is used for this purpose[19] and a computer programme for it is developed in VB 6.0. In this stage, first a hierarchical structure of the problem, which is structured hierarchically similar to a flow chart, is constructed. The overall objective is placed at the top while the criteria and subcriteria are placed below. For example, as shown in figure 11, the overall objective Environmental performance index (EPI1) is placed at the top (level 1), then below (level 2) are criterias Total PEI, Ef , Ec and RC and after this (level 3) sub-criterias as HTPI, HTPE, TTP, ATP, GWP, ODP, PCOP and AP. After this using the numerical scale given in table 2.6, two pairwise comparison matrices (see Table 7 and 8) are constructed for determination of weights for aggregation of individual impact categories of WAR to total PEI and for determination of weights of total PEI, Ef, Ec and RC to Environmental performance index (EPI1).

Fig. 11. Hierarchical structuring of multicriteria decision analysis problem for integrating individual environmental impacts

The right hand upper diagonal information in both matrices is to be provided by the decision maker giving the relative importance of the two criteria using the numerical scale of table 2.6 while the left hand lower diagonal is the reciprocal of the right hand upper diagonal. Once these pair wise comparison matrices are constructed, then developed computer programme using the AHP method, determines the weighting of individual impact categories. The level of inconsistency of decision makers input is checked by

Systematic Framework for Multiobjective Optimization in Chemical Process Plant Design 353

1( ) 1 *EnvCat*

*L*

The higher value of environmental performance index (EPI1) shows that the process is environmentally better and vice versa. While the higher value of environmental pollution

The relevant information generated from process module, safety module and environment module for each alternative is transferred to data manager. This information is used to formulate process diagnostic tables and multiobjective decision-making problem formulation. These tables consist of mass input/output table, energy input/output table, capital and utility annual expense summary, environmental impact summary and frequency

The purpose of this layer/stage is to set up multiobjective decision making/optimization among these conflicting objectives. The aim is to find out the trade-off surface for each alternative and /or complete ranking of alternatives. The calculation loop used for it is

In each independent performance module i.e. economic, environment and safety module, relevant information is generated and transferred automatically or manually to data manager for each alternative generated or under study. Before transferring the values of performance objective functions, each objective function is optimized within their

Process/Economic optimization of each alternative is carried out using SQP

 The lower and upper limits for Environmental objective functions are calculated using environmental module from the material and energy balance information from process

 Safety/risk aspects are optimized in the safety module and information such as hazard occurance frequency, safety cost data (fixed safety system cost, accident and/or incident

Depending on the case under study or objectives of the study, graphical tool box of MatLab and/or multiobjective optimization technique (goal programming) or multiattribute decision analysis technique (PROMETHEE and/or AHP) is used for multiobjective decision analysis. The Data Manager is linked with MatLab 7.0 via Excel link Toolbox in the

damage risk cost) for each alternative is transferred to the data manger.

Where *E Total PEI E E R L c* ,, , *<sup>f</sup> <sup>c</sup>*

**2.7 Data manager** 

shown in figure 12.

model.

independent module such as:

and environmental pollution index (EPI2) is calculated as follow

index (EPI2) shows that the environmental performance of process is worse.

of occurance of an event and their consequence categories and safety cost.

**2.8 Stage III: Multiobjective decision analysis/ optimization** 

optimization algorithm build within Aspen PlusTM.

*L L*

*EPI W E* (18)

2 *EPI Total PEI E <sup>f</sup>* (19)


Table 7. Pairwise comparison matrix for individual impact categories at level 3


Table 8. Pairwise comparison matrix for individual impact categories at level 2

consistency ratio before giving the output. Consistency ratio less than 0.1 is good and for ratios greater than 0.1, the input to pair wise matrix should be re-evaluated.

#### **2.6.4 Step IV: Environmental performance & pollution index calculation**

In the final step, first Total PEI is determined by multiplying each impact category values with its relevant weighting factor WL as given below:

$$\text{Total PEI} = \sum\_{L}^{\text{EnvCat}} \mathcal{W}\_{L} \cdot \text{PEI}\_{L} \tag{17}$$

After calculating Total PEI, Environmental performance index (EPI1) is determined for each alternative by multiplying the values of Total PEI, Ef, EC and RC with its relevant weighting factor WL (table 8) as given below:

$$EPI1 = \bigvee \left(\sum\_{L}^{EnvCat} \mathcal{W}\_{L} \cdot E\_{L}\right) \tag{18}$$

Where *E Total PEI E E R L c* ,, , *<sup>f</sup> <sup>c</sup>*

and environmental pollution index (EPI2) is calculated as follow

$$\text{EPI2} = \text{Total PEI} + \text{E}\_f \tag{19}$$

The higher value of environmental performance index (EPI1) shows that the process is environmentally better and vice versa. While the higher value of environmental pollution index (EPI2) shows that the environmental performance of process is worse.

#### **2.7 Data manager**

352 Advances in Chemical Engineering

Pairwise comparison matrix

HTPI 1 A12 A13 A14 A15 A16 A17 A18 **HTPE** 1 A23 A24 A25 A26 A27 A28 **TTP** 1 A34 A35 A36 A37 A38 **ATP** 1 A45 A46 A47 A48 **GWP** 1 A56 A57 A58 **ODP** 1 A67 A68 **PCOP** 1 A78 **AP** 1 **WL** W1 W2 W3 W4 W5 W6 W7 W8

Table 7. Pairwise comparison matrix for individual impact categories at level 3

Table 8. Pairwise comparison matrix for individual impact categories at level 2

ratios greater than 0.1, the input to pair wise matrix should be re-evaluated.

**2.6.4 Step IV: Environmental performance & pollution index calculation** 

with its relevant weighting factor WL as given below:

factor WL (table 8) as given below:

**Pairwise comparison matrix** 

consistency ratio before giving the output. Consistency ratio less than 0.1 is good and for

In the final step, first Total PEI is determined by multiplying each impact category values

*EnvCat*

*L*

After calculating Total PEI, Environmental performance index (EPI1) is determined for each alternative by multiplying the values of Total PEI, Ef, EC and RC with its relevant weighting

*L L*

*Total PEI W PEI* (17)

 **PEI RC EC Ef PEI** 1 A12 A13 A14 **RC** 1 A23 A24 **EC** 1 A34 **Ef** 1 **WL** W1 W2 W3 W4

**HTPI HTPE TTP ATP GWP ODP PCOP AP** 

The relevant information generated from process module, safety module and environment module for each alternative is transferred to data manager. This information is used to formulate process diagnostic tables and multiobjective decision-making problem formulation. These tables consist of mass input/output table, energy input/output table, capital and utility annual expense summary, environmental impact summary and frequency of occurance of an event and their consequence categories and safety cost.

#### **2.8 Stage III: Multiobjective decision analysis/ optimization**

The purpose of this layer/stage is to set up multiobjective decision making/optimization among these conflicting objectives. The aim is to find out the trade-off surface for each alternative and /or complete ranking of alternatives. The calculation loop used for it is shown in figure 12.

In each independent performance module i.e. economic, environment and safety module, relevant information is generated and transferred automatically or manually to data manager for each alternative generated or under study. Before transferring the values of performance objective functions, each objective function is optimized within their independent module such as:


Depending on the case under study or objectives of the study, graphical tool box of MatLab and/or multiobjective optimization technique (goal programming) or multiattribute decision analysis technique (PROMETHEE and/or AHP) is used for multiobjective decision analysis. The Data Manager is linked with MatLab 7.0 via Excel link Toolbox in the

Systematic Framework for Multiobjective Optimization in Chemical Process Plant Design 355

[1] Ramzan, N., Naveed, S., Feroze, N. and Witt, W.; "Multicriteria decision analysis for

[2] Ramzan, N., Degenkolbe, S. and Witt, W.; "Evaluating and improving environmental

[3] Ramzan, N., Compart, F. and Witt, W.; "Methodology for the generation and evaluation

Journal of Americal Institute of Chemical Engineering), 26(1),35-42 (2007). [4] Ramzan, N., Compart, F. and Witt, W.; "Application of extended Hazop and event tree

[5] Ramzan, N., Witt, W.; "Multiobjective optimization in distillation unit: A case study", The Canadian Journal of Chemical Engineering, 84(5), 604-613(2006). [6] Seader et al. (1999), Process design principles-Synthesis, Analysis, and Evaluation, John

[7] Nawaz, Z. Mahmood, Z. (2006) Importance Modeling, Simulation and Optimization in

[8] Cameron, I., Raman, R. (2005), Process Systems Risk Management, Elsevier Academic

[9] Crowl, D.A., Louvar, J.F. (1999), Chemical Process Safety: Fundamentals with

[10] Kletz, T.A. (1997), Hazop-past and future, Reliability engineering and system safety, 55,

[12] Peters, M. S. and K.D. Timmerhaus (1991), Plant Design and Economics for Chemical

[14] Guthrie, K.M. (1969), Data and techniques for preliminary capital cost estimating,

[15] Guthrie, K.M. (1974), Process plant estimating, evaluation and control, Craftsman,

[16] Cabezas, H. Bare, C. & Mallick, K. (1999), Pollution prevention with chemical process

[17] Cabezas, H. Bare, C. & Mallick, K. (1997), Pollution prevention with chemical process

simulators: the generalized waste reduction (WAR) algorithm-full version.

simulators: the generalized waste reduction (WAR) algorithm, Computers and

safety and economic achievement using PROMETHEE: A case study" Process Safety Progress (A Journal of Americal Institute of Chemical Engineering), 28(1),

performance of HC's recovery system: A case study of distillation unit", Chemical Engineering Journal (Journal published by Elsevier), 40(1-3), 201-213(2008). ISSN:

of safety system alternatives based on extended Hazop" Process Safety Progress (A

analysis for investigating operational failures and safety optimization of distillation column unit" Process Safety Progress (A Journal of Americal Institute of Chemical

Chemical process Design, The Pakistan Engineer (Journal of Institute of Engineers

**3. References** 

68-83 (2009).

1385-8947.

Pakistan), 38-39.

Chem. Eng., 114-1421.

Solano Beach, CA.

263-266.

Engineering), 26(3),248-257 (2007).

Wiley & Sons Inc. New York, 338-370.

applications," Prentice Hall, New York.

[11] Lees, F.P. (1996), Loss prevention in CPI, Butterworth's, London, UK.

[13] Douglas, J. (1998), Conceptual design of chemical processes, McGraw Hill Inc.

Engineers, Ed.2nd, McGraw-Hill, New York, 90-145.

Computers and Chemical Engineering, 23,623-634.

Chemical Engineering, 21s, s305-s310.

Press, NY, ISBN 0-12-156932-2

Fig. 12. Calculation loop for multiobjective optimization

integrated interface. Aspen plusTM is linked in the integrated interface via Visual basic 6.0 and Microsoft Excel. AHP technique is also programmed for the cases under study in Visual basic 6.0. The computer realization of these links in the integrated interface is explained in chapter four.

#### **2.9 Stage IV: Design evaluation**

The purpose of this layer/stage is to select the best alternative and/or find the complete ranking of alternatives under study based on the results of third stage/layer of the developed methodology. Pareto approach (non dominated analysis) or PROMETHEE is used for this purpose.

#### **3. References**

354 Advances in Chemical Engineering

Fig. 12. Calculation loop for multiobjective optimization

chapter four.

used for this purpose.

**2.9 Stage IV: Design evaluation** 

integrated interface. Aspen plusTM is linked in the integrated interface via Visual basic 6.0 and Microsoft Excel. AHP technique is also programmed for the cases under study in Visual basic 6.0. The computer realization of these links in the integrated interface is explained in

The purpose of this layer/stage is to select the best alternative and/or find the complete ranking of alternatives under study based on the results of third stage/layer of the developed methodology. Pareto approach (non dominated analysis) or PROMETHEE is


**14** 

*Brazil* 

**CFD Modelling of Fluidized** 

**Bed with Immersed Tubes** 

*Universidade Estadual de Maringa* 

A.M.S. Costa, F.C. Colman, P.R. Paraiso and L.M.M. Jorge

Fluidized beds are widely used in combustion and chemical industries. The immersed tubes are usually used for enhancement of heat transfer or control of temperature in fluidized beds. By his turn, tubes subjected to the solid particle impact may suffer severe erosion wear. Many investigations have been devoted to erosion in tubes immersed in fluidized beds on the various influencing factors (cf. Lyczkowski and Bouillard, 2002). As pointed by Achim et al. (2002), the factors can be classified as particle characteristics, mechanical design

Some previous experimental studies have focused on bubble and particle behaviors (Kobayashi et al., 2000, Ozawa et al., 2002), tube attrition, erosion or wastage (Bouillard and Lyczkowski, 1991; Lee and Wang, 1995; Fan et al., 1998; Wiman, 1994), heat transfer (Wong and Seville, 2006, Wiman and Almstedt, 1997) and gas flow regimes (Wang et. al, 2002).

Previous numerical studies were also performed using different CFD codes. Recently He et al. (2009, 2004), using the K-FIX code adapted to body fitted coordinates investigated the hydrodynamics of bubbling fluidized beds with one to four immersed tubes. The erosion rates predicted using the monolayer kinetic energy dissipation model were compared against the experimental values of Wiman (1994) for the two tube arrangement. The numerical values were three magnitudes lower than the experimental ones. Also employing a eulerian-eulerian model and the GEMINI numerical code, Gustavsson and Almstedt (2000, 1999) performed numerical computations and comparison against experimental results (Enwald et. al., 1999). As reported for those authors, fairly good qualitative agreement between the experimental and numerical erosion results were obtained, and the contributions to the erosion from the

In the present study, we revisit the phenomena of the immersed tubes in a gas fluidized bed with different immersed tube arrangements employing the eulerian-eulerian two fluid model and the MFIX code. The purposes of the numerical simulations are to compare and

The mathematical model is based on the assumption that the phases can be mathematically described as interpenetrating continua; the point variables are averaged over a region that is

explore some effects not previously investigated in the above-mentioned references.

different fluid dynamics phenomena near the tube were identified.

**2. Two fluid and erosion model** 

**1. Introduction** 

and operating conditions.


## **CFD Modelling of Fluidized Bed with Immersed Tubes**

A.M.S. Costa, F.C. Colman, P.R. Paraiso and L.M.M. Jorge *Universidade Estadual de Maringa Brazil* 

#### **1. Introduction**

356 Advances in Chemical Engineering

[18] Dimian, A. C. (2003), Integrated design and simulation of chemical processes, 1st Eds.,

[19] Dev, P.K. (2004), Analytic hierarchy process helps evaluate project in Indian oil

pipelines industry, International journal of operation and production management,

Elsevier Netherlands, 1-30,113-134.

24(6), 588-604

Fluidized beds are widely used in combustion and chemical industries. The immersed tubes are usually used for enhancement of heat transfer or control of temperature in fluidized beds. By his turn, tubes subjected to the solid particle impact may suffer severe erosion wear. Many investigations have been devoted to erosion in tubes immersed in fluidized beds on the various influencing factors (cf. Lyczkowski and Bouillard, 2002). As pointed by Achim et al. (2002), the factors can be classified as particle characteristics, mechanical design and operating conditions.

Some previous experimental studies have focused on bubble and particle behaviors (Kobayashi et al., 2000, Ozawa et al., 2002), tube attrition, erosion or wastage (Bouillard and Lyczkowski, 1991; Lee and Wang, 1995; Fan et al., 1998; Wiman, 1994), heat transfer (Wong and Seville, 2006, Wiman and Almstedt, 1997) and gas flow regimes (Wang et. al, 2002).

Previous numerical studies were also performed using different CFD codes. Recently He et al. (2009, 2004), using the K-FIX code adapted to body fitted coordinates investigated the hydrodynamics of bubbling fluidized beds with one to four immersed tubes. The erosion rates predicted using the monolayer kinetic energy dissipation model were compared against the experimental values of Wiman (1994) for the two tube arrangement. The numerical values were three magnitudes lower than the experimental ones. Also employing a eulerian-eulerian model and the GEMINI numerical code, Gustavsson and Almstedt (2000, 1999) performed numerical computations and comparison against experimental results (Enwald et. al., 1999). As reported for those authors, fairly good qualitative agreement between the experimental and numerical erosion results were obtained, and the contributions to the erosion from the different fluid dynamics phenomena near the tube were identified.

In the present study, we revisit the phenomena of the immersed tubes in a gas fluidized bed with different immersed tube arrangements employing the eulerian-eulerian two fluid model and the MFIX code. The purposes of the numerical simulations are to compare and explore some effects not previously investigated in the above-mentioned references.

#### **2. Two fluid and erosion model**

The mathematical model is based on the assumption that the phases can be mathematically described as interpenetrating continua; the point variables are averaged over a region that is

CFD Modelling of Fluidized Bed with Immersed Tubes 359

The solid stress model outlined by Eqs. (6) and (7) will be quoted here as the standard model. Additionally, a general formulation for the solids phase stress tensor that admits a

<sup>v</sup> \*

S if ε ε

According to Pannala et al.(2009), two diferent formulations for the weighting parameter

 

*s s*

S + 1 S if ε ε

*p*

Tanh

\* f

\* \* ε ε

ε

2

and the shape factor

goes to zero and

is the momentum interaction term between the solid and

2.65

*<sup>s</sup> P vv <sup>f</sup> <sup>s</sup> <sup>f</sup>* (10)

 

1

(9a,b)

are smaller values less

(Eqs. 11 to 16). The first of the

blending function

*f f s p*

*v vd*

(11)

*f*

equals to unity, the

(8)

f ff <sup>s</sup>

ε

"switch" model as proposed by Syamlal et al. (1993) based on the Schaeffer (1987) can be recovered. The models based on eqs. (9a) and (9b) will be referred in the numerical

On the other hand, the Srivastava and Sundaresan (2003), also called "Princeton model", can

 I = fs 

correlations for the drag coefficient is based on Wen and Yu (1966) work. The Gidaspow drag coefficient is a combination between the Wen Yu correlation and the correlation from Ergun (1952). The Gidaspow blended drag correlation allows controlling the transition from

was originally proposed by Lathowers and Bellan (2000) and the value of parameter *C* controls the degree of transition. From Eq. (14), the correlation proposed by Syamlal and O'Brien (1993) carries the advantage of adjustable parameters C1 and d1 for different minimum fluidization conditions. The correlations given in Eq. (15) and Eq. (16) are based on Lattice-Boltzmann simulations. For detailments of these last drag correlations refer to the

0.687 <sup>f</sup>

3 4

*d*

*f fs f s D f p*

*v v*

 

transition between the two regimes is given by :

S

"" can be employed :

*v s*

\*

1 

ε

In the above equation the void fraction range

be placed on the basis of Eq. (9) Also in equations (4) and (5) Ifs

fluid phases, given by

than unity. It must be emphasized that when

simulations as BLEND S and BLEND T, respectivelly.

There is a number of correlations for the drag coefficient

works by Benyahia et al. (2006) and Wang et al. (2010).

*D*

*C*

the Wen and Yu, and Ergun based correlations. In this correlation the

<sup>24</sup> 1 0.15Re Re<1000 Re Re =

*C*

 

0.44 Re 1000

Wen-Yu

1

\* ε ε ε

 

large compared with the particle spacing but much smaller than the flow domain (see Anderson, 1967). A short summary of the equations solved by the numerical code in this study are presented next. Refer to Benyahia et al. (2006) and Syamlal et al. (1993) for more detailment.

The continuity equations for the fluid and solid phase are given by :

$$\frac{\partial}{\partial t}(\varepsilon\_{\text{f}}\boldsymbol{\varrho}\_{\text{f}}) \; + \; \nabla \cdot (\varepsilon\_{\text{f}}\boldsymbol{\varrho}\_{\text{f}} \; \overleftarrow{\mathbf{v}}\_{\text{f}}) = \; \mathbf{0} \tag{1}$$

$$\frac{\partial}{\partial t}(\varepsilon\_s \wp\_s) + \nabla \cdot (\varepsilon\_s \wp\_s \cdot \overline{\wp}\_s) = 0 \tag{2}$$

In the previous equations f, s, f, s, vf and vs are the volumetric fraction, density and velocity field for the fluid and solids phases.

The momentum equations for the fluid and solid phases are given by:

$$\frac{\partial}{\partial t}(\varepsilon\_{\text{f}}\mathbf{p}\_{\text{f}}\mathbf{\overline{v}}\_{\text{f}}) \; + \; \nabla \bullet (\varepsilon\_{\text{f}}\mathbf{p}\_{\text{f}}\mathbf{\overline{v}}\_{\text{f}}\mathbf{\overline{v}}\_{\text{f}}) = \nabla \bullet \mathbf{\overline{S}}\_{\text{f}} \; + \; \varepsilon\_{\text{f}}\mathbf{p}\_{\text{f}}\mathbf{\overline{g}} \; - \; \mathbf{\overline{I}}\_{\text{fs}} \tag{3}$$

$$\frac{\partial}{\partial t} \left( \varepsilon\_{s} \rho\_{s} \overline{\mathbf{v}}\_{s} \right) \; + \; \nabla \bullet \left( \varepsilon\_{s} \rho\_{s} \overline{\mathbf{v}}\_{s} \overline{\mathbf{v}}\_{s} \right) = \nabla \bullet \overline{\mathbf{S}}\_{s} \; + \; \varepsilon\_{s} \rho\_{s} \overline{\mathbf{g}}\_{s} \; + \; \overline{\mathbf{I}}\_{\mathrm{fs}} \tag{4}$$

Sf Ss are the stress tensors for the fluid and solid phase. It is assumed newtonian behavior for the fluid and solid phases, i.e.,

S 2 *P v I S pI ij* 1 1 2 3 *<sup>T</sup> S vv v ij* (5a,b)

In the above equation *P, ,* are the pressure, bulk and dynamic viscosity, respectively.

In addition, the solid phase behavior is divided between a plastic regime (also named as slow shearing frictional regime) and a viscous regime (also named as rapidly shearing regime). The constitutive relations for the plastic regime are related to the soil mechanics theory. Here they are representated as:

$$\mathbf{p}\_s^{\mathbb{P}} = \mathbf{f}\_1 \left( \mathbf{e}^\*, \mathbf{e}\_f \right). \qquad \mu\_s^{\mathbb{P}} = \mathbf{f}\_2 \left( \mathbf{e}^\*, \mathbf{e}\_f, \boldsymbol{\phi} \right) \tag{6}$$

In the above equation \* ε is the packed bed void fraction and is the angle of internal friction.

A detailing of functions f1 to f4 and f9 can be obtained in Benyahia (2008).

On the other hand, the viscous regime behavior for the solid phase is ruled by two gas kinetic theory related parameters (e, ).

$$\mathbf{p\_s^v = f\_3\left(\varepsilon\_{s'}, \mathfrak{p}\_{s'}, d\_{p'}, \Theta, \mathbf{e}\right)} \qquad \mathfrak{m}\_s^v = \mathfrak{f}\_4\left(\varepsilon\_{s'}, \mathfrak{p}\_{s'}, d\_{p'}, \Theta^{1/2}, \mathbf{e}\right) \tag{7}$$

large compared with the particle spacing but much smaller than the flow domain (see Anderson, 1967). A short summary of the equations solved by the numerical code in this study are presented next. Refer to Benyahia et al. (2006) and Syamlal et al. (1993) for more

<sup>f</sup> f f f f ε ρ + ε ρ v = 0

<sup>s</sup> s s s s ε ρ + ε ρ v = 0

 <sup>f</sup> f f <sup>f</sup> fs f f f f f f ε ρ v + ε ρ v v = S + ε ρ I <sup>t</sup> *<sup>g</sup>* 

 <sup>s</sup> s s <sup>s</sup> fs s s s s s s ε ρ v + ε ρ v v = S + ε ρ I <sup>t</sup> *<sup>g</sup>* 

Sf Ss are the stress tensors for the fluid and solid phase. It is assumed newtonian behavior

In addition, the solid phase behavior is divided between a plastic regime (also named as slow shearing frictional regime) and a viscous regime (also named as rapidly shearing regime). The constitutive relations for the plastic regime are related to the soil mechanics

<sup>p</sup> \* p f s 1 <sup>ε</sup> ,<sup>ε</sup> *<sup>f</sup>* . <sup>p</sup> \* <sup>μ</sup>s 2f <sup>ε</sup> ,<sup>ε</sup> , *<sup>f</sup>*

On the other hand, the viscous regime behavior for the solid phase is ruled by two gas

1 1

and vs

(1)

(2)

are the volumetric fraction, density and

(3)

(4)

are the pressure, bulk and dynamic viscosity, respectively.

<sup>v</sup> p = f s 3 ss ε ,, ,Θ,e *<sup>p</sup> d* <sup>v</sup> 1/2 <sup>μ</sup>s 4 ss <sup>f</sup> <sup>ε</sup> ,, ,<sup>Θ</sup> ,e *<sup>p</sup> <sup>d</sup>* (7)

2 3 *<sup>T</sup> S vv v ij* 

(5a,b)

(6)

is the angle of internal

The continuity equations for the fluid and solid phase are given by :

t

t

The momentum equations for the fluid and solid phases are given by:

In the previous equations f, s, f, s, vf

velocity field for the fluid and solids phases.

S 2 *P v I S pI ij* 

In the above equation \* ε is the packed bed void fraction and

A detailing of functions f1 to f4 and f9 can be obtained in Benyahia (2008).

*,* 

for the fluid and solid phases, i.e.,

theory. Here they are representated as:

kinetic theory related parameters (e, ).

In the above equation *P,* 

friction.

detailment.

The solid stress model outlined by Eqs. (6) and (7) will be quoted here as the standard model. Additionally, a general formulation for the solids phase stress tensor that admits a transition between the two regimes is given by :

$$\begin{aligned} \stackrel{\mathsf{m}}{\mathbf{S}\_{8}} = \begin{cases} \phi(\boldsymbol{\varepsilon}\_{\mathrm{f}}) \stackrel{\mathsf{m}^{\mathsf{v}}}{\mathbf{S}\_{s}} + \left[1 - \phi(\boldsymbol{\varepsilon}\_{\mathrm{f}})\right] \stackrel{\mathsf{m}^{\mathsf{p}}}{\mathbf{S}\_{s}} & \text{if } \,\,\boldsymbol{\varepsilon}\_{\mathrm{f}} < \boldsymbol{\varepsilon}^{\*} + \delta \\ \stackrel{\mathsf{m}^{\mathsf{v}}}{\mathbf{S}\_{s}} & \text{if } \,\,\boldsymbol{\varepsilon}\_{\mathrm{f}} \ge \boldsymbol{\varepsilon}^{\*} + \delta \end{cases} \end{aligned} \tag{8}$$

According to Pannala et al.(2009), two diferent formulations for the weighting parameter "" can be employed :

$$\phi(\varepsilon) = \frac{1}{1 + \frac{\varepsilon - \varepsilon^\*}{2\delta \varepsilon^\*}} \qquad \phi(\varepsilon) = \frac{\text{Tanh}\left(\frac{\pi \left(\varepsilon - \varepsilon^\*\right)}{\delta \varepsilon^\*}\right) + 1}{2} \tag{9a,b}$$

In the above equation the void fraction range and the shape factor are smaller values less than unity. It must be emphasized that when goes to zero and equals to unity, the "switch" model as proposed by Syamlal et al. (1993) based on the Schaeffer (1987) can be recovered. The models based on eqs. (9a) and (9b) will be referred in the numerical simulations as BLEND S and BLEND T, respectivelly.

On the other hand, the Srivastava and Sundaresan (2003), also called "Princeton model", can be placed on the basis of Eq. (9)

Also in equations (4) and (5) Ifs is the momentum interaction term between the solid and fluid phases, given by

$$\vec{\mathbf{I}}\_{\rm fs} = -\boldsymbol{\varepsilon}\_{s} \nabla P\_{f} - \mathcal{J}\left(\vec{v}\_{s} - \vec{v}\_{f}\right) \tag{10}$$

There is a number of correlations for the drag coefficient (Eqs. 11 to 16). The first of the correlations for the drag coefficient is based on Wen and Yu (1966) work. The Gidaspow drag coefficient is a combination between the Wen Yu correlation and the correlation from Ergun (1952). The Gidaspow blended drag correlation allows controlling the transition from the Wen and Yu, and Ergun based correlations. In this correlation the blending function was originally proposed by Lathowers and Bellan (2000) and the value of parameter *C* controls the degree of transition. From Eq. (14), the correlation proposed by Syamlal and O'Brien (1993) carries the advantage of adjustable parameters C1 and d1 for different minimum fluidization conditions. The correlations given in Eq. (15) and Eq. (16) are based on Lattice-Boltzmann simulations. For detailments of these last drag correlations refer to the works by Benyahia et al. (2006) and Wang et al. (2010).

$$\mathbb{B}\_{\text{Ween-Yu}} = \frac{3}{4} \mathbb{C}\_{D} \frac{\rho\_{f} \varepsilon\_{f} \mathbb{a}\_{s} \left| \vec{v}\_{f} - \vec{v}\_{s} \right|}{d\_{p}} \varepsilon\_{f}^{-2.65}$$

$$\mathbb{C}\_{D} = \begin{cases} \frac{24}{\text{Re}} \{ 1 + 0.15 \text{Re}^{0.687} \} & \text{Re} \le 1000\\ 0.44 & \text{Re} \ge 1000 \end{cases} \quad \text{Re} = \frac{\rho\_{l} \varepsilon\_{f} \left| \vec{v}\_{f} - \vec{v}\_{s} \right| d\_{p}}{\mu\_{f}}$$

CFD Modelling of Fluidized Bed with Immersed Tubes 361

For erosion calculations in this work we use the monolayer energy dissipation model (Lyczkowski and Bouillard, 2002). In that model the kinetic energy dissipation rate for the solids phase in the vicinity of stationary immersed surfaces is related to erosion rate in m/s by multiplication with an appropriate constant. This constant is function of surface hardness, elasticity of collision and diameter of particles hitting the surface. The kinetic

s

 

 

: v

MFIX (Multiphase Flow with Interphase eXchanges) is an open source CFD code developed at the National Energy Technology Laboratory (NETL) for describing the hydrodynamics, heat transfer and chemical reactions in fluid-solids systems. It has been used for describing bubbling and circulating fluidized beds, spouted beds and gasifiers. MFIX calculations give transient data on the three-dimensional distribution of pressure, velocity, temperature, and

The hydrodynamic model is solved using the finite volume approach with discretization on a staggered grid. A second order accurate discretization scheme was used and superbee scheme was adopted for discretization of the convective fluxes at cell faces for all equations in this work. With the governing equations discretized, a sequential iterative solver is used to calculate the field variables at each time step. The main numerical algorithm is an extension of SIMPLE. Modifications to this algorithm in MFIX include a partial elimination algorithm to reduce the strong coupling between the two phases due to the interphase transfer terms. Also, MFIX makes use of a solids volume fraction correction step instead of a solids pressure correction step which is thought to assist convergence in loosely packed regions. Finally, an adaptive time step is used to minimize computation time. See Syamlal (1998) for more details. The immersed obstacles were implemented using the cut-cell

The numerical runs were based on experiments of Wiman (1994, 1997) in an air pressurized bed with horizontal tubes for four different tube-bank geometries. The T2 and I4 are portrayed in Fig. 1 whereas the S4 and S4D tubes geometry are given in Figure 3. Figure 2 details the domain for the numerical simulations and circumferential angle coordinate for erosion measurement. In Fig. 2a is detailed the mesh for the T2 arrangement This nonuniform stretched grid, follows Cebeci et al. (2005) approach, has fine spacing close to the surface of the tubes and coarse spacing away from the surface. For all the other arrangements the mesh was uniform as depicted in Fig. 2b. The mesh employed for the bed without tube, and the I4 arrangement was 60 260 cells, and for the S4 and S4D arrangement 60 340. The bed was operated at room temperature (24 C) at pressures between 0.1 and 1.6 MPa and at two different excess velocities: Uf1 – Umf = 0.2 m/s and Uf1 – Umf = 0.6 m/s. Here, Uf1 is the superficial fluidization velocity based on the free bed crosssection. The voidage at minimum fluidization was 0.46 and the minimum fluidization velocities was 0.42 , 0.31 and 0.18 m/s, for 0.1, 0.4 and 1.6 MPa pressures, correspondingly. The particle diameter and density were 700 m and 2600 kg/m3. The bubble parameters

*<sup>s</sup> <sup>s</sup> s s*

2

(19)

v

2

energy dissipation rate s in W/m3 for the solids phase is given by :

**3. Numerical method** 

species mass fractions.

technique available in the code (Dietiker, 2009)

$$\beta\_{\text{Gidaspow}} = \begin{vmatrix} \beta\_{\text{Wen-Yu}} & \varepsilon\_{\text{f}} > 0.8\\ \beta\_{\text{Eygun}} = \frac{150 \ \varepsilon\_{s} (1 - \varepsilon\_{s}) \mu\_{\text{f}}}{\varepsilon\_{f} d\_{p}^{2}} + \frac{1.75 \ \eta \varepsilon\_{s} \left| \vec{v}\_{f} - \vec{v}\_{s} \right|}{d\_{p}} & \varepsilon\_{\text{f}} \le 0.8 \end{vmatrix} \tag{12}$$

$$\beta\_{\text{Gidaspro-blended}} = \chi \ \beta\_{\text{Wen-Yu}} + (1 - \chi)\beta\_{\text{Ergun}} \quad \chi = \frac{\tan^{-1}\left(C \ (\varepsilon\_{\text{f}} - 0.8)\right)}{\pi} + 0.5\tag{13}$$

$$\begin{aligned} \text{p}\_{\text{Syanal-ORiter}} &= \frac{3}{4} \frac{\rho\_f \varepsilon\_f \varepsilon\_s}{V\_r^2 d\_p} \left( 0.63 + 4.8 \sqrt{\frac{V\_r}{\text{Re}}} \right)^2 \left| \vec{v}\_f - \vec{v}\_s \right| \\ V\_r &= 0.5A - 0.03 \,\text{Re} + 0.5 \times \sqrt{\left( 0.06 \,\text{Re} \right)^2 + 0.12 \,\text{Re} \left( 2B - A \right) + A^2} \\ A &= \varepsilon\_f^{4.14} \quad B = \begin{cases} \mathbb{C}\_1 \varepsilon\_f^{1.28} & \varepsilon\_f \le 0.85 \\ \varepsilon\_f^{d\_1} & \varepsilon\_t > 0.85 \end{cases} \end{aligned} \tag{14}$$

$$\beta\_{\text{Hill-Kock-Ladd}} = 18 \mu\_f \left( 1 - \varepsilon\_s \right)^2 \varepsilon\_s \frac{F}{d\_p^2} \qquad F = f\_9 \left( F\_{0'} F\_{1'} F\_{2'} F\_3 \right) \tag{15}$$

$$\beta\_{\text{Besselra}} = 180 \frac{\mu\_t \varepsilon\_s^2}{\text{d}\_{\text{p}}^2} + 18 \frac{\mu\_t \varepsilon\_t^3 \varepsilon\_s \left(1 + 1.5 \sqrt{\varepsilon\_s} \right)}{\text{d}\_{\text{p}}^2} + 0.31 \frac{\mu\_t \varepsilon\_s \text{ Re}}{\varepsilon\_t \text{d}\_{\text{p}}^2} \frac{\left[\varepsilon\_t^{-1} + 3 \varepsilon\_t \varepsilon\_s + 8.4 \text{Re}^{-0.343} \right]}{\left[1 + 10^{3 \varepsilon\_s} \text{ Re}^{-0.5 - 2 \text{e}\_s} \right]} \tag{16}$$

For closing the model, a transport equation for the granular energy provides a way of determine the pressure and viscosity for the solid phase during the viscous regime. Equation (17) is a transport equation for the granular energy . Its solution provides a way of determine the pressure and viscosity for the solid phase during the viscous regime. The terms <sup>s</sup> and gs are the granular energy conductivity, dissipation and exchange, respectively.

$$\frac{3}{2} \left[ \frac{\partial}{\partial t} \varepsilon\_s \rho\_s \Theta \; \; \; + \; \nabla \cdot \mathbf{p}\_s \overline{\mathbf{v}}\_s \Theta \right] = \overline{\overline{\mathbf{S}}}\_s : \nabla \overline{\mathbf{v}}\_s - \nabla \cdot \left( \kappa\_s \nabla \Theta \right) \; \; \; - \gamma + \phi\_{\overline{\mathbf{g}}s} \tag{17}$$

$$\begin{aligned} \mathbf{x}\_s &= \mathbf{f}\_5 \left( \boldsymbol{\varepsilon}\_s, \rho\_s, d\_p, \Theta^{1/2}, \mathbf{e}\_r, \mathfrak{P} \right) \\ \boldsymbol{\gamma} &= \mathbf{f}\_6 \left( \boldsymbol{\varepsilon}\_s, \rho\_s, d\_p, \Theta^{3/2}, \mathbf{e} \right) \\ \boldsymbol{\phi}\_{\mathcal{S}^s} &= \boldsymbol{f}\_7 \left( \boldsymbol{\varepsilon}\_s, \rho\_s, d\_p, \Theta\_r \middle| \vec{v}\_f - \vec{v}\_s \middle| \ \boldsymbol{\rho} \right) \end{aligned} \tag{18}$$

In the algebraic approach, instead solving the full equation (17) , the granular energy is obtained by equating the first term on the right hand side with the dissipation term.

The model where Eqs. (5) to (8) and (17) are solved is the kinetic theory model, termed here as KTGF. Conversely, in the constant solids viscosity model (CVM) the solids pressure is defined as in Eq. (6) and the solids viscosity in either plastic and viscous regimes is set equal to a constant.

For erosion calculations in this work we use the monolayer energy dissipation model (Lyczkowski and Bouillard, 2002). In that model the kinetic energy dissipation rate for the solids phase in the vicinity of stationary immersed surfaces is related to erosion rate in m/s by multiplication with an appropriate constant. This constant is function of surface hardness, elasticity of collision and diameter of particles hitting the surface. The kinetic energy dissipation rate s in W/m3 for the solids phase is given by :

$$\Phi\_s = \left[ \stackrel{\bullet}{\varepsilon\_s \tau\_s} : \stackrel{\bullet}{\nabla \mathbf{v}\_s} + \beta \stackrel{\bullet}{\frac{\mathbf{v}\_s^2}{2}} \right] \tag{19}$$

#### **3. Numerical method**

360 Advances in Chemical Engineering

Wen-Yu f

> 0.8

+ 0.8 *sf s s s*


<sup>2</sup>

1 1.5 Re 3 8.4Re

<sup>f</sup> f f (16)

**:** (17)

*<sup>F</sup> F f FFFF <sup>d</sup>* (15)

2

1 = 0.5

*v v*

*v v*

f

*C*

tan 0.8

2 2

(12)

s s

(13)

(14)

(18)

Ergun 2 f

*f fs <sup>r</sup> f s*

*V*

*f p p*

*d d*

150 1 1.75

Gidaspow f f

Gidaspow-blended Wen-Yu Ergun

Syamlal-OBrien 2

*f d*

*r*

respectively.

to a constant.

*A B*

1

180 18 0.31

*f*

1 f 4.14

*C*

1.28

*f*

*r p*

*V d*

0.85 0.85

f

p p f p

s s s s <sup>s</sup> <sup>3</sup> ρ + ρ v = S v

5 ss

*s p*

6ss

7ss

obtained by equating the first term on the right hand side with the dissipation term.

 

2 t *s s gs*

f ,, ,Θ ,e,

*d*

*p*

*f d vv*

In the algebraic approach, instead solving the full equation (17) , the granular energy is

The model where Eqs. (5) to (8) and (17) are solved is the kinetic theory model, termed here as KTGF. Conversely, in the constant solids viscosity model (CVM) the solids pressure is defined as in Eq. (6) and the solids viscosity in either plastic and viscous regimes is set equal

f ,,,Θ ,e

*d*

*gs p fs*

,,,Θ, ,

1/2

3/2

 

 

<sup>3</sup> 0.63 4.8 4 Re

0.5 0.03Re 0.5 0.06Re 0.12Re 2

*V A BA A*

Hill-Koch-Ladd <sup>2</sup> 9 0123 18 1 , , , *f ss p*

<sup>1</sup> 0.343 f s <sup>s</sup> f fs s s Beestra 22 2 <sup>3</sup> 0.5 2

For closing the model, a transport equation for the granular energy provides a way of determine the pressure and viscosity for the solid phase during the viscous regime. Equation (17) is a transport equation for the granular energy . Its solution provides a way of determine the pressure and viscosity for the solid phase during the viscous regime. The terms <sup>s</sup> and gs are the granular energy conductivity, dissipation and exchange,

dd d [1 10 Re

MFIX (Multiphase Flow with Interphase eXchanges) is an open source CFD code developed at the National Energy Technology Laboratory (NETL) for describing the hydrodynamics, heat transfer and chemical reactions in fluid-solids systems. It has been used for describing bubbling and circulating fluidized beds, spouted beds and gasifiers. MFIX calculations give transient data on the three-dimensional distribution of pressure, velocity, temperature, and species mass fractions.

The hydrodynamic model is solved using the finite volume approach with discretization on a staggered grid. A second order accurate discretization scheme was used and superbee scheme was adopted for discretization of the convective fluxes at cell faces for all equations in this work. With the governing equations discretized, a sequential iterative solver is used to calculate the field variables at each time step. The main numerical algorithm is an extension of SIMPLE. Modifications to this algorithm in MFIX include a partial elimination algorithm to reduce the strong coupling between the two phases due to the interphase transfer terms. Also, MFIX makes use of a solids volume fraction correction step instead of a solids pressure correction step which is thought to assist convergence in loosely packed regions. Finally, an adaptive time step is used to minimize computation time. See Syamlal (1998) for more details. The immersed obstacles were implemented using the cut-cell technique available in the code (Dietiker, 2009)

The numerical runs were based on experiments of Wiman (1994, 1997) in an air pressurized bed with horizontal tubes for four different tube-bank geometries. The T2 and I4 are portrayed in Fig. 1 whereas the S4 and S4D tubes geometry are given in Figure 3. Figure 2 details the domain for the numerical simulations and circumferential angle coordinate for erosion measurement. In Fig. 2a is detailed the mesh for the T2 arrangement This nonuniform stretched grid, follows Cebeci et al. (2005) approach, has fine spacing close to the surface of the tubes and coarse spacing away from the surface. For all the other arrangements the mesh was uniform as depicted in Fig. 2b. The mesh employed for the bed without tube, and the I4 arrangement was 60 260 cells, and for the S4 and S4D arrangement 60 340. The bed was operated at room temperature (24 C) at pressures between 0.1 and 1.6 MPa and at two different excess velocities: Uf1 – Umf = 0.2 m/s and Uf1 – Umf = 0.6 m/s. Here, Uf1 is the superficial fluidization velocity based on the free bed crosssection. The voidage at minimum fluidization was 0.46 and the minimum fluidization velocities was 0.42 , 0.31 and 0.18 m/s, for 0.1, 0.4 and 1.6 MPa pressures, correspondingly. The particle diameter and density were 700 m and 2600 kg/m3. The bubble parameters

CFD Modelling of Fluidized Bed with Immersed Tubes 363

(a) (b)

obtained from simulation were based on methodology described in Almstedt (1987), using numerical probes in the domain, centered at (0.15, 0.55) and separated from 15 mm. The target tube for erosion measurements is the one centered at (0.18, 0.55) for all the tube arrangements.

In this work, the parameters for controlling the numerical solution (e.g., under-relaxation, sweep direction, linear equation solvers, number of iterations, residual tolerances) were kept

More information about the experiments can be accessed from Wiman (1994, 1997).

Fig. 3. Geometry and mesh for (a) S4 and (b) S4D tube arrangement

Fig. 1. Geometry and mesh for T2 and I4 tube arrangement.

Fig. 2. Mesh detailment for (a) T2 arrangement (b) all the others arrangement

(a) (b)

(a) (b)

Fig. 2. Mesh detailment for (a) T2 arrangement (b) all the others arrangement

Fig. 1. Geometry and mesh for T2 and I4 tube arrangement.

Fig. 3. Geometry and mesh for (a) S4 and (b) S4D tube arrangement

obtained from simulation were based on methodology described in Almstedt (1987), using numerical probes in the domain, centered at (0.15, 0.55) and separated from 15 mm. The target tube for erosion measurements is the one centered at (0.18, 0.55) for all the tube arrangements. More information about the experiments can be accessed from Wiman (1994, 1997).

In this work, the parameters for controlling the numerical solution (e.g., under-relaxation, sweep direction, linear equation solvers, number of iterations, residual tolerances) were kept

CFD Modelling of Fluidized Bed with Immersed Tubes 365

Fig. 5. Time averaged kinetic energy dissipation predicted for different gas-solid drag

results the most severe dissipation occurs on the lower parts of the tube for an angle corresponding to 90 degrees. This fact is in agreement with experimental measured values of erosion from Wiman (1994). Analysis of results shows that, except for the BVK drag model, all the drag models predict the highest dissipation rate occurring at 90 degrees. The value predicted using the Syamlal-O'Brien drag model is the highest. There is a noticeable difference in the value of dissipation value between the Gidaspow and Gidaspow blend drag models, whereas the results by the Wen-Yu drag locate in the intermediate range. By his turn, the results by HYS and Koch-Hill models locates near the

Also, analysis of transient simulated results shows the highest values of kinetic energy dissipation rate occur when the particle fraction suddenly changes from a low to a high

Figure 6 presents the result of kinetic energy dissipation for different solid stress models. As shown the peak values of the dissipation rate are close to 180 degrees for the model with (baseline) and without the blending function discussed in section 2. On the other hand, the maximum value predicted by the constant solids viscosity model (NO KTGF) locates around

Figure 7 presents the result of kinetic energy dissipation for different tube surface slip conditions. The baseline case considers the free slip condition for the solids and non-slip condition for the gas. The peak values do not change when considering the solids with the non-slip condition. On the other hand, when considering slip conditions for both phases the value decreased. The above results, suggests the free slip condition for the gas phase

180 degrees, and it is superior to the predicted by the solids kinetic energy theory.

value, which corresponds to the tube being hit by the wake of a bubble.

models – T2 arrangement

Wen-Yu values.

as their default code values. Moreover, for setting up the mathematical model, when not otherwise specified the code default values were used. The computer used in the numerical simulations was a PC with OpenSuse linux and Intel Quad Core processor. The simulation time was 20 s.

For generating the numerical results, e employed the parameters listed above, referred here as baseline simulation. In addition, for the baseline simulation we employed the Syamlal-O´Brien drag model, the standard solid stress model, and slip and non-slip condition for solid and gas phase, correspondingly. The previous set of models will be referred in the result's section as baseline simulation models.

#### **4. Results and discussion**

Figure 4 is a sampling plot showing the instantaneous solids velocities and gas volumetric fraction fields following two different bubbles passage around the obstacle. Analysis of Fig. (4) shows the bubble splitting mechanism taking place and the characteristic time scale for the bubble passage. After the bubble passage, the solid wake has higher solid velocity magnitude around the obstacle. The solids upward movement following the bubble wake and wall downward movement is kept unchanged. Generally, for beds without internal obstacles, bubble size increases with bed height, particle size and superficial velocity. Analysis of CFD results shows that the presence of tubes was found to alter such general trends for bubble growth.

Fig. 4. Detail of the instantaneous voidage and solids velocity vector field for the T2 arrangement

Figure 5 presents the time averaged kinetic energy dissipation as a function of circumferential position on the tube surface. As it would be seen from the majority of

as their default code values. Moreover, for setting up the mathematical model, when not otherwise specified the code default values were used. The computer used in the numerical simulations was a PC with OpenSuse linux and Intel Quad Core processor. The simulation

For generating the numerical results, e employed the parameters listed above, referred here as baseline simulation. In addition, for the baseline simulation we employed the Syamlal-O´Brien drag model, the standard solid stress model, and slip and non-slip condition for solid and gas phase, correspondingly. The previous set of models will be referred in the

Figure 4 is a sampling plot showing the instantaneous solids velocities and gas volumetric fraction fields following two different bubbles passage around the obstacle. Analysis of Fig. (4) shows the bubble splitting mechanism taking place and the characteristic time scale for the bubble passage. After the bubble passage, the solid wake has higher solid velocity magnitude around the obstacle. The solids upward movement following the bubble wake and wall downward movement is kept unchanged. Generally, for beds without internal obstacles, bubble size increases with bed height, particle size and superficial velocity. Analysis of CFD results shows that the presence of tubes was found to alter such general

(c) 11 s (d) 12 s

Figure 5 presents the time averaged kinetic energy dissipation as a function of

on the tube surface. As it would be seen from the majority of

Fig. 4. Detail of the instantaneous voidage and solids velocity vector field for the T2

time was 20 s.

result's section as baseline simulation models.

**4. Results and discussion** 

trends for bubble growth.

arrangement

circumferential position

Fig. 5. Time averaged kinetic energy dissipation predicted for different gas-solid drag models – T2 arrangement

results the most severe dissipation occurs on the lower parts of the tube for an angle corresponding to 90 degrees. This fact is in agreement with experimental measured values of erosion from Wiman (1994). Analysis of results shows that, except for the BVK drag model, all the drag models predict the highest dissipation rate occurring at 90 degrees. The value predicted using the Syamlal-O'Brien drag model is the highest. There is a noticeable difference in the value of dissipation value between the Gidaspow and Gidaspow blend drag models, whereas the results by the Wen-Yu drag locate in the intermediate range. By his turn, the results by HYS and Koch-Hill models locates near the Wen-Yu values.

Also, analysis of transient simulated results shows the highest values of kinetic energy dissipation rate occur when the particle fraction suddenly changes from a low to a high value, which corresponds to the tube being hit by the wake of a bubble.

Figure 6 presents the result of kinetic energy dissipation for different solid stress models. As shown the peak values of the dissipation rate are close to 180 degrees for the model with (baseline) and without the blending function discussed in section 2. On the other hand, the maximum value predicted by the constant solids viscosity model (NO KTGF) locates around 180 degrees, and it is superior to the predicted by the solids kinetic energy theory.

Figure 7 presents the result of kinetic energy dissipation for different tube surface slip conditions. The baseline case considers the free slip condition for the solids and non-slip condition for the gas. The peak values do not change when considering the solids with the non-slip condition. On the other hand, when considering slip conditions for both phases the value decreased. The above results, suggests the free slip condition for the gas phase

CFD Modelling of Fluidized Bed with Immersed Tubes 367

on the surface of the tube plays an important role on the value of the kinetic energy

Figures 8 and 9 shows a comparison between the numerically predicted time averaged values of the kinetic energy dissipation rate, using the baseline simulation models discussed in section 2, and those based on the experimental values for the T2 arrangement in the work of Gustavson and Almstedt (2000). A comparison of the numerical and experimental results for the two different operational pressures (c.f. Figs 8 and 9 ) shows some degree of discordance. However, the lack of minutely agreement with experimental values, the results can be compared for recognizing similar drifts. For instance, from the Fig 8, the higher simulated values of the dissipation rate occurring in the bottom position of the tube, i.e. for < 180o, is in agreement with the experimental counterpart. Similarly, the increase of dissipation rate with increased operational pressure for the simulated results is in agreement with the experimental values. Regarding the erosion and baseline models used for the simulation, some remarks towards better agreement with experimental values can also be done. According to the monolayer erosion model and its discussion above Eq. (19) some degree of uncertainty is associated to the multiplying constant, as the exact value of elasticity of collision is not known. It is also expected, that adjustments in the baseline simulation models, such as those outlined in Figs. (5) to (7) would produce better

Figure 10 is a sampling plot showing the instantaneous gas volumetric fraction fields for different tube arrangements. Analysis of Fig. (10) shows the influence of immersed obstacles on the bubble splitting mechanism taking place and the bubble passage pattern. Above the tube bank, the bubble appears to grow to size similar to the without tube geometry. For the geometry with tubes the bubble encompasses the obstacles but not at the full width of the

Figures 11 and 12 shows a comparison between the numerically predicted values of bubble frequency, using the baseline simulation models discussed in section 2, and those based on the experimental measurements from the works of Almstedt (1987) and Wiman (1995). As shown in the Fig. 10, the calculated values of Nb are underestimated at higher pressures, while at low pressures, there is a quite good agreement between calculated and experimental results. This conclusion holds true both for the I4 and for the S4D tube arrangement. As in the experimental results there are no noticeable differences for the two

Figures 13 and 14 shows a comparison for the bubble frequency for the bed without tube and with the S4 arrangement. As shown in Fig. 13 for the bed without tubes the trend points to a frequency agreement between 0.4 and 0.6 MPa. Up this range the values differences increases as pressures increases up to 1.6 MPa. For the S4 arrangement the trend is similar to the I4 and S4D arrangement. The experimental results depicted in Figs. 11 to 14 suggest that the frequency increases with pressure both with and without tubes. For numerical results this holds true only for low pressures, i.e., 0.1 and 0.4 MPa. The numerical results suggest a maxima occurring between 0.4 and 1.6 MPa. On the other hand, the numerical results corroborate the experimental trend that the mean frequency is higher for

bed. The interaction is stronger for the denser tube geometry.

the bed with tubes than for the freely bubbling bed.

dissipation.

agreement.

tubes arrangements.

Fig. 6. Time averaged kinetic energy dissipation predicted for different solid stress models

Fig. 7. Time averaged kinetic energy dissipation predicted for different tube surface slip conditions

Fig. 6. Time averaged kinetic energy dissipation predicted for different solid stress models

Fig. 7. Time averaged kinetic energy dissipation predicted for different tube surface slip

conditions

on the surface of the tube plays an important role on the value of the kinetic energy dissipation.

Figures 8 and 9 shows a comparison between the numerically predicted time averaged values of the kinetic energy dissipation rate, using the baseline simulation models discussed in section 2, and those based on the experimental values for the T2 arrangement in the work of Gustavson and Almstedt (2000). A comparison of the numerical and experimental results for the two different operational pressures (c.f. Figs 8 and 9 ) shows some degree of discordance. However, the lack of minutely agreement with experimental values, the results can be compared for recognizing similar drifts. For instance, from the Fig 8, the higher simulated values of the dissipation rate occurring in the bottom position of the tube, i.e. for < 180o, is in agreement with the experimental counterpart. Similarly, the increase of dissipation rate with increased operational pressure for the simulated results is in agreement with the experimental values. Regarding the erosion and baseline models used for the simulation, some remarks towards better agreement with experimental values can also be done. According to the monolayer erosion model and its discussion above Eq. (19) some degree of uncertainty is associated to the multiplying constant, as the exact value of elasticity of collision is not known. It is also expected, that adjustments in the baseline simulation models, such as those outlined in Figs. (5) to (7) would produce better agreement.

Figure 10 is a sampling plot showing the instantaneous gas volumetric fraction fields for different tube arrangements. Analysis of Fig. (10) shows the influence of immersed obstacles on the bubble splitting mechanism taking place and the bubble passage pattern. Above the tube bank, the bubble appears to grow to size similar to the without tube geometry. For the geometry with tubes the bubble encompasses the obstacles but not at the full width of the bed. The interaction is stronger for the denser tube geometry.

Figures 11 and 12 shows a comparison between the numerically predicted values of bubble frequency, using the baseline simulation models discussed in section 2, and those based on the experimental measurements from the works of Almstedt (1987) and Wiman (1995). As shown in the Fig. 10, the calculated values of Nb are underestimated at higher pressures, while at low pressures, there is a quite good agreement between calculated and experimental results. This conclusion holds true both for the I4 and for the S4D tube arrangement. As in the experimental results there are no noticeable differences for the two tubes arrangements.

Figures 13 and 14 shows a comparison for the bubble frequency for the bed without tube and with the S4 arrangement. As shown in Fig. 13 for the bed without tubes the trend points to a frequency agreement between 0.4 and 0.6 MPa. Up this range the values differences increases as pressures increases up to 1.6 MPa. For the S4 arrangement the trend is similar to the I4 and S4D arrangement. The experimental results depicted in Figs. 11 to 14 suggest that the frequency increases with pressure both with and without tubes. For numerical results this holds true only for low pressures, i.e., 0.1 and 0.4 MPa. The numerical results suggest a maxima occurring between 0.4 and 1.6 MPa. On the other hand, the numerical results corroborate the experimental trend that the mean frequency is higher for the bed with tubes than for the freely bubbling bed.

CFD Modelling of Fluidized Bed with Immersed Tubes 369

(a) No tubes (b) I4 (c) S4 (d) S4D

Fig. 10. Snapshots of voidage field at 4 s for different tube arrangements. P = 0.4 MPa ,

Uf = 0.6 m/s

Fig. 8. Simulated kinetic energy dissipation at various circumferential angular positions in the surface of the tube and values from experimental results by Gustavsson and Almstedt, 2000. Operational pressures : 0.8 MPa ;

Fig. 9. Simulated kinetic energy dissipation at various circumferential angular positions in the surface of the tube and values from experimental results by Gustavsson and Almstedt, 2000. Operational pressures : 1.6 MPa

Fig. 8. Simulated kinetic energy dissipation at various circumferential angular positions in the surface of the tube and values from experimental results by Gustavsson and Almstedt,

Fig. 9. Simulated kinetic energy dissipation at various circumferential angular positions in the surface of the tube and values from experimental results by Gustavsson and Almstedt,

2000. Operational pressures : 0.8 MPa ;

2000. Operational pressures : 1.6 MPa

Fig. 10. Snapshots of voidage field at 4 s for different tube arrangements. P = 0.4 MPa , Uf = 0.6 m/s

CFD Modelling of Fluidized Bed with Immersed Tubes 371

Fig. 13. Nb versus pressure, numerical X experimental: no tubes, Uf = 0.6 m/s

Fig. 14. Nb versus pressure, numerical X experimental: S4 arrangement, Uf = 0.6 m/s

Fig. 11. Nb versus pressure, numerical X experimental: I4 arrangement, Uf = 0.6 m/s

Fig. 12. Nb versus pressure, numerical X experimental: S4D arrangement, Uf = 0.6 m/s

Fig. 11. Nb versus pressure, numerical X experimental: I4 arrangement, Uf = 0.6 m/s

Fig. 12. Nb versus pressure, numerical X experimental: S4D arrangement, Uf = 0.6 m/s

Fig. 14. Nb versus pressure, numerical X experimental: S4 arrangement, Uf = 0.6 m/s

CFD Modelling of Fluidized Bed with Immersed Tubes 373

Fig. 17. Vb versus pressure, numerical X experimental: : no tubes, Uf = 0.6 m/s

Fig. 18. Vb versus pressure, numerical X experimental: : S4 arrangement, Uf = 0.6 m/s

Figures 19 and 20 shows a comparison for the bubble frequency for Uf = 0.2 m/s. Comparison with results in Figs. 12 and 14 shows that the numerical results, although smaller are closer to the experimental. The tendency holds true for both S4 and S4D arrangements. Comparison with the numerical results for the S4D and S4 arrangements

Figure 15 to 18 shows a comparison for the mean bubble velocity. In all cases, the simulated results are underestimated in relation to the experiments. However, the trend observed for the experimental results with a maxima around 0.4 MPa is verified for the numerical results for all the tube arrangements. For the bed without tubes the experimental increase trend of Vb with pressure is valid for pressures higher than 0.4 MPa.

Fig. 15. Vb versus pressure, numerical X experimental: : I4 arrangement, Uf = 0.6 m/s

Fig. 16. Vb versus pressure, numerical X experimental: S4D arrangement, Uf = 0.6 m/s

Figure 15 to 18 shows a comparison for the mean bubble velocity. In all cases, the simulated results are underestimated in relation to the experiments. However, the trend observed for the experimental results with a maxima around 0.4 MPa is verified for the numerical results for all the tube arrangements. For the bed without tubes the experimental increase trend of

Fig. 15. Vb versus pressure, numerical X experimental: : I4 arrangement, Uf = 0.6 m/s

Fig. 16. Vb versus pressure, numerical X experimental: S4D arrangement, Uf = 0.6 m/s

Vb with pressure is valid for pressures higher than 0.4 MPa.

Fig. 17. Vb versus pressure, numerical X experimental: : no tubes, Uf = 0.6 m/s

Fig. 18. Vb versus pressure, numerical X experimental: : S4 arrangement, Uf = 0.6 m/s

Figures 19 and 20 shows a comparison for the bubble frequency for Uf = 0.2 m/s. Comparison with results in Figs. 12 and 14 shows that the numerical results, although smaller are closer to the experimental. The tendency holds true for both S4 and S4D arrangements. Comparison with the numerical results for the S4D and S4 arrangements

CFD Modelling of Fluidized Bed with Immersed Tubes 375

Fig. 21. Time averaged kinetic energy dissipation predicted for different tube arrangements

Fig. 22. Time averaged kinetic energy dissipation predicted for different tube arrangements

at two different operating pressures : : S4 arrangement, P = 1\_6 MPa

at two different operating pressures: : S4 arrangement, P = 0.1 MPa

Fig. 19. Nb versus pressure, numerical X experimental: : S4 arrangement, Uf = 0.2 m/s

Fig. 20. Nb versus pressure, numerical X experimental: S4D arrangement, Uf = 0.2 m/s

given in Fig. 6 and 7, also shows that Nb increases with increasing excess velocity. The last, is the same drift verified for the experimental values.

Figures 21 and 22 present the results for the time averaged kinetic energy dissipation as a function of circumferential position on the tube surface for the S4 arrangement for two distinct pressures. As it would be seen the highest experimental results are between 90 and

Fig. 19. Nb versus pressure, numerical X experimental: : S4 arrangement, Uf = 0.2 m/s

Fig. 20. Nb versus pressure, numerical X experimental: S4D arrangement, Uf = 0.2 m/s

is the same drift verified for the experimental values.

function of circumferential position

given in Fig. 6 and 7, also shows that Nb increases with increasing excess velocity. The last,

Figures 21 and 22 present the results for the time averaged kinetic energy dissipation as a

distinct pressures. As it would be seen the highest experimental results are between 90 and

on the tube surface for the S4 arrangement for two

Fig. 21. Time averaged kinetic energy dissipation predicted for different tube arrangements at two different operating pressures: : S4 arrangement, P = 0.1 MPa

Fig. 22. Time averaged kinetic energy dissipation predicted for different tube arrangements at two different operating pressures : : S4 arrangement, P = 1\_6 MPa

CFD Modelling of Fluidized Bed with Immersed Tubes 377

Benyahia, S., Syamlal, M., O'Brien, T. J., "Summary of MFIX Equations 2005-4", 1 March

Benyahia, S., Syamlal, M., O'Brien, T. J., 2006, "Extension of Hill–Koch–Ladd drag

Bouillard, J. X., Lyczkowski, R. W., 1991, "On the erosion of heat exchanger tube banks in

Cebeci, T., Shao, J. P., Karyeke, F., Laurendeau, E., 2005, Computational Fluid Dynamics for

Dietiker, J., "Cartesian Grid User Guide", 4 September 2009, Available from https://mfix.netl.doe.gov/documentation/Cartesian\_grid\_user\_guide.pdf. Enwald, H., Peirano, E., Almstedt, A. E., Leckner, B., 1999, "Simulation of the fluid

Ergun, S., 1952, "Fluid-flow through packed columns", Chemical Engineering Progress, Vol

Fan, J. R., Sun, P., Chen, L. H., Cen, K. F., 1998, "Numerical investigation of a new protection method of the tube erosion by particle impingement", Wear, Vol 223, pp. 50-57. Gustavsson, M., Almstedt, A. E., 1999, "Numerical simulation of fluid dynamics in fluidized

Gustavsson, M., Almstedt, A. E., 2000, "Two-fluid modelling of cooling-tube erosion in a

He, Y. R., Lu, H. L., Sun, Q. Q., Yang, L. D., Zhao, Y. H., Gidaspow, D., Bouillard, J., 2004,

He, Y. R., Zhan, W., Zhao, Y., Lu, H. Schlaberg, I., 2009, "Prediction on immersed tubes

Kobayashi, N., Yamazaki, R., Mori, S., 2000, "A study on the behaviour of bubbles and solids in bubbling fluidized beds", Powder Technology, Vol 113, pp. 327-344. Lathowers, D., Bellan, J., 2000, "Modeling of dense gas-solid reactive mixtures applied to

Lee, S. W., Wang, B. Q., 1995, "Effect of particle-tube collision frequency on material

Lyczkowski, R. W., Bouillard, J. X., 2002, "State-of-the-art review of erosion modeling in

Ozawa, M., Umekawa, H., Furui, S., Hayashi, K., Takenaka, N., 2002, "Bubble behavior and

bed model", Experimental Thermal and Fluid Science, Vol 26, pp. 643-652.

fluidized bed", Chemical Engineering Science, 55, 867-879.

beds", Powder Technology, 145, pp. 88-105.

Program Review, NREL/CP-570-28890. USA.

Science, 64, pp. 3072-3082.

pp. 223-229.

602.

fluidized-bed combustor", Powder Technology, Vol 68, pp. 37-51.

correlation over all ranges of Reynolds number and solids volume fraction",

dynamics of a bubbling fluidized bed: Experimental validation of the two-fluid model and evaluation of a parallel multiblock solver", Chemical Engineering

beds with horizontal heat exchanger tubes", Chemical Engineering Science, 55, 857-

"Hydrodynamics of gas-solid flow around immersed tubes in bubbling fluidized

erosion using two-fluid model in a bubbling fluidized bed", Chemical Engineering

biomass pyrolysis in a fluidized bed", Proceedings of the 2000 U.S. DOE Hydrogen

wastage on in-bed tubes in the bubbling fluidized bed combustor", Wear, Vol 184,

fluid/solid systems", Progress in Energy and Combustion Science, Vol 28, pp. 543-

void fraction fluctuation in vertical tube banks immersed in a gas-solid fluidized-

http://www.mfix.org/documentation/MfixEquations2005-4-1.pdf.

2006, Available from

Science, 54, 311-328.

48, n. 2, pp. 91-94.

866.

Powder Technology, 162, 166-174.

Engineers, Horizon Publishing Inc, USA, 402p.

240 degrees, while for the numerical are between 30 and 150. The numerical values are over predicted in the range 0 to 100 degrees and above 240 degrees. In the range from 100 to 240 degrees the numerical are below the experimental. Similar trends are verified for the 1.6 MPa pressure, with the experimental curves less sensitive to pressure variation . By his turn the numerical values, show more sensitivity to pressure, although with the same magnitude order.

#### **5. Conclusion**

In this work was investigated numerically the hydrodynamics of two dimensional beds with immersed tubes. The simulations were based on an experimental bed with different tube bank geometries, operating pressures and gas excess velocities. The objective of this study was two fold: explore and investigate some effects not previously explored in the literature, to verify the feasibility of the MFIX code for such a kind of study.

The simulation's results were framed in terms of averaged solids kinetic energy dissipation rate and bubble parameters (frequency and mean velocity). A comparison between the numerical results for frequency and the experiments shows good agreement for low pressures. Also, some similar drifts were identified, e.g., greater frequency for greater excess velocities. By his turn, the bubble velocity numerical results agrees better with experiments for high pressures. For the bed with tubes, a maxima of bubble velocity for intermediate pressure is identifiable, this could not be verified for the numerical results. Our results points to significant influences in the predicted dissipation rates and consequently, in the erosion rate, when employing different drag models. The dissipation rate is also influenced by either the use of blending functions for the transition between the plastic and viscous regime of solids flow or the use of a constant viscosity model. The results are key sensitive to the slip condition for the gas phase in the surface of the tube. In the case of a free slip condition for the gas phase the lowest values of dissipation are obtained By his turn, for the solids kinetic dissipation energy, the range of angles that gives maximum values are shifted in relation to the experiments. Finally, remarks towards better agreement with experimental values can also be done for the energy dissipation model. Specifically, according to the monolayer erosion model and its discussion above Eq. (19) some degree of uncertainty is associated to the multiplying constant, as the exact value of elasticity of collision is not known.

#### **6. References**


240 degrees, while for the numerical are between 30 and 150. The numerical values are over predicted in the range 0 to 100 degrees and above 240 degrees. In the range from 100 to 240 degrees the numerical are below the experimental. Similar trends are verified for the 1.6 MPa pressure, with the experimental curves less sensitive to pressure variation . By his turn the numerical values, show more sensitivity to pressure, although with the same magnitude

In this work was investigated numerically the hydrodynamics of two dimensional beds with immersed tubes. The simulations were based on an experimental bed with different tube bank geometries, operating pressures and gas excess velocities. The objective of this study was two fold: explore and investigate some effects not previously explored in the

The simulation's results were framed in terms of averaged solids kinetic energy dissipation rate and bubble parameters (frequency and mean velocity). A comparison between the numerical results for frequency and the experiments shows good agreement for low pressures. Also, some similar drifts were identified, e.g., greater frequency for greater excess velocities. By his turn, the bubble velocity numerical results agrees better with experiments for high pressures. For the bed with tubes, a maxima of bubble velocity for intermediate pressure is identifiable, this could not be verified for the numerical results. Our results points to significant influences in the predicted dissipation rates and consequently, in the erosion rate, when employing different drag models. The dissipation rate is also influenced by either the use of blending functions for the transition between the plastic and viscous regime of solids flow or the use of a constant viscosity model. The results are key sensitive to the slip condition for the gas phase in the surface of the tube. In the case of a free slip condition for the gas phase the lowest values of dissipation are obtained By his turn, for the solids kinetic dissipation energy, the range of angles that gives maximum values are shifted in relation to the experiments. Finally, remarks towards better agreement with experimental values can also be done for the energy dissipation model. Specifically, according to the monolayer erosion model and its discussion above Eq. (19) some degree of uncertainty is associated to the multiplying constant, as the exact value of elasticity of collision is not

Achim, D., Easton, A. K., Schwarz, M. P., Witt, P.J., Zakhari, A., 2002, "Tube erosion

Almstedt, A. E., 1987, A study of bubble behaviour and gas distribution in pressurized

Anderson, T. B., 1967, "A fluid mechanical description of fluidized beds: Equations of motion", Industrial Engineering Chemical Fundamentals, Vol 6, pp. 527-539. Benyahia, S., 2008, "Validation study of two continuum granular frictional flow theories",

Chalmers University of Technology, Goteborg. Sweden.

Industrial Engineering Chemical Research, 47, 8926-8932.

modelling in a fluidised bed", Applied Mathematical Modelling, Vol 26, pp. 191-

fluidized beds burning coal. Thesis for the degree of Licentiate of Engineering,

literature, to verify the feasibility of the MFIX code for such a kind of study.

order.

known.

**6. References** 

201.

**5. Conclusion** 

Benyahia, S., Syamlal, M., O'Brien, T. J., "Summary of MFIX Equations 2005-4", 1 March 2006, Available from

http://www.mfix.org/documentation/MfixEquations2005-4-1.pdf.


**15** 

**Optimal Synthesis of Multi-Effect** 

**with a High Boiling Point Rise** 

Jaime Alfonzo Irahola *Universidad Nacional de Jujuy* 

*Argentina* 

**Evaporation Systems of Solutions** 

In the past, optimization had been studied only for typical flowpatterns like forward and backward feed. To make the choice between them, simple rules were applied based on the viscosity and the temperature of the initial dilute solution (TF). Thus, forward feed was usually favored for the evaporation of low-viscous hot solutions featuring a temperature TF>Tp, where TP is the desired temperature of the final product. By doing so, the liquid heating load is largely cut down. In turn, backward feed was recommended for heavy-

Moreover, a few contributions to the optimal synthesis of multiple effect evaporator systems (MEES) have so far been published. Most of the previous papers was focused on the analysis rather than the synthesis of evaporation systems. They generally assumed that important structural variables like vapor and liquid flowpatterns and the number of effects are all

Nishitani and Kunugita (1979) first presented a multiobjective problem formulation to determine the optimal flowpattern of a multiple-effect evaporator system. However, they did not consider stream mixing/splitting. In addition, the solution method performed oneby-one the simulation of the MEES for all possible flowpatterns. More recently, Hillebrand and Westerberg (1988) developed a simple model to explicitly compute the utility consumption for multiple-effect evaporator systems exchanging sensible heat with outside streams. In turn, Westerberg and Hillebrand (1988) introduced the concept of "heat shunt" to derive the best liquid flowpattern in a heuristic way. Nonetheless, major assumptions like constant boiling point elevation, no liquid bypassing and negligible heat of mixing

To resolve the problem presented here has been used a mathematical model rigorous previously developed (Irahola & Cerdá, 1994). It considers the possibility of optimizing the variables that you want to. This has allowed that the model is used to solve various types of problems, namely: simulation, optimization, optimal synthesis and optimum partial reengineering restrictive of multi effect evaporation system (MEES) (Irahola, 2008). The mathematical model developed is the MINLP and solved using commercial software. The approach was successfully applied to three industrial problems. Depending on the feed and

known data though they drastically change the performance of a MEES.

somewhat limit the applicability of their findings.

**1. Introduction** 

viscous cold liquors.


## **Optimal Synthesis of Multi-Effect Evaporation Systems of Solutions with a High Boiling Point Rise**

Jaime Alfonzo Irahola *Universidad Nacional de Jujuy Argentina* 

#### **1. Introduction**

378 Advances in Chemical Engineering

Pannala, S., Daw, C. S., Finney, C. E. A., Benyahia, S., Syamlal, M., O´Brien, T. J., "Modelling

Schaeffer, D. G., 1987, "Instability in the evolution equations describing incompressible

Siravastava, A., Sundaresan, S., 2003, "Analysis of a frictional–kinetic model for gas–particle

Syamlal, M., 1998, "MFIX Documentation, Numerical Techniques", Technical Note,

Syamlal, M., Rogers, W. A., O'Brien, T. J., 1993, "MFIX Documentation, Theory Guide",

Wang, J., van der Hoef, M. A., Kuipers, J. A. M., 2010,"CFD study of the minimum bubbling

Wen, C. Y., Yu, Y. H., 1966, "Mechanics of Fluidization", Chemical Engineering Progress

Wiman, J., 1994, An experimental study of hydrodynamics and tube erosion in a pressurized

Wiman, J., Almstedt, A. E., 1997, "Hydrodynamics, erosion and heat transfer in a

granular flow", Journal Differential Equations, v. 66, pp. 19-50.

Micromechanics of Granular Media, pp. 657-660.

flow", Powder Technology, v. 129, pp. 72-85.

Information Service, Springfield, VA, USA.

Symposium Series, Vol 62, n. 62, pp. 100-111.

beds", AIChe Journal, Vol 52, pp. 4099-4109.

Chalmers University of Technology, Goteborg. Sweden.

Springfield, VA, USA.

Science, 65, pp. 3772-3785.

the collisional-plastic stress transition for bin discharge of granular material", 2009, Powders and Grains 2009 – Proceeding of the 6th International Conference on

DOE/MC-31346-5824, NTIS/DE98002029, National Technical Information Service,

Technical Note, DOE/METC-94/1004, NTIS/DE94000087, National Technical

velocity of Geldart A particles in gas-fluidized beds", Chemical Engineering

fluidized with horizontal tubes. Thesis for the degree of Licentiate of Engineering,

pressurized fluidized bed: influence of pressure, fluidization velocity, particle size and tube bank geometry", Chemical Engineering Science, Vol 52, pp. 2677-2695. Wong, Y. S., Seville, J. P. K., 2006, "Single-particle motion and heat transfer in fluidized In the past, optimization had been studied only for typical flowpatterns like forward and backward feed. To make the choice between them, simple rules were applied based on the viscosity and the temperature of the initial dilute solution (TF). Thus, forward feed was usually favored for the evaporation of low-viscous hot solutions featuring a temperature TF>Tp, where TP is the desired temperature of the final product. By doing so, the liquid heating load is largely cut down. In turn, backward feed was recommended for heavyviscous cold liquors.

Moreover, a few contributions to the optimal synthesis of multiple effect evaporator systems (MEES) have so far been published. Most of the previous papers was focused on the analysis rather than the synthesis of evaporation systems. They generally assumed that important structural variables like vapor and liquid flowpatterns and the number of effects are all known data though they drastically change the performance of a MEES.

Nishitani and Kunugita (1979) first presented a multiobjective problem formulation to determine the optimal flowpattern of a multiple-effect evaporator system. However, they did not consider stream mixing/splitting. In addition, the solution method performed oneby-one the simulation of the MEES for all possible flowpatterns. More recently, Hillebrand and Westerberg (1988) developed a simple model to explicitly compute the utility consumption for multiple-effect evaporator systems exchanging sensible heat with outside streams. In turn, Westerberg and Hillebrand (1988) introduced the concept of "heat shunt" to derive the best liquid flowpattern in a heuristic way. Nonetheless, major assumptions like constant boiling point elevation, no liquid bypassing and negligible heat of mixing somewhat limit the applicability of their findings.

To resolve the problem presented here has been used a mathematical model rigorous previously developed (Irahola & Cerdá, 1994). It considers the possibility of optimizing the variables that you want to. This has allowed that the model is used to solve various types of problems, namely: simulation, optimization, optimal synthesis and optimum partial reengineering restrictive of multi effect evaporation system (MEES) (Irahola, 2008). The mathematical model developed is the MINLP and solved using commercial software. The approach was successfully applied to three industrial problems. Depending on the feed and

Optimal Synthesis of Multi-Effect Evaporation

b. A solid phase never arises in any effect.

f. The steam always condenses completely. g. Subcooling of the condensate is very small. h. Flow of noncondensables is negligible.

d. There is no leakage or entrainment.

transfer liquid to any other one.

the preset temperature (Tp).

no boiling.

**3.2 Mathematical model** 

adopted in the formulation:

Systems of Solutions with a High Boiling Point Rise 381

The scope of the rigorous mathematical model is limited by the following assumptions

i. The concentrated final product is withdrawn from a single effect which in turn does not

j. If necessary, you can use a heat exchanger or condenser so that the product go out to

To solve the mathematical model and find the optimal design should be available before, the mathematical expressions for all dependent variables: enthalpy of steam (H), enthalpy of solution of soda caustic (h), latent heat of vaporization of the water (λ), overall heat transfer coefficient (U), temperature of the solution in the evaporator (T) and costs of forced circulation evaporator, barometric condenser multijet, surface condenser and heat exchanger. In general, useful information is available in graphics, which presented various authors cited in the bibliography, but there are no equations of those curves. These situations and other problems are resolved below. According to Standiford (1963), in forced circulation evaporators, film transfer coefficient (h) on the side of the liquid, can be calculated from the conventional Dittus-Boelter equation for forced circulation when there is

> � � ��� �� �� � � ���

If there is this equation for two points and combine both equations, you can find the

As the overall heat transfer coefficient U is practically determined by the film coefficient h fluid side, the above equation can be used to obtain a correlation for U. In Geankoplis for T=105 °F and X = 0.5, data is U = 400 (Btu/h ft2 °F). T and X is obtained from a graph (Horvath, 1985) μ2 = 22.84 centipoise. Then, for five values of concentration and six temperature values are obtained from graphics (Horvath, 1985) the values of Cp1 and μ1. With the data obtained can be calculated according to the above expression, the overall heat transfer coefficient U1 with reference to U2. Thus, U is plotted vs. X (Fig. 1) and U vs. T (Fig. 2). Finally, using the triple X, T, Ui, can be found by regression, the functionality of U in terms

���

�� �� � ��� � ��� ���� (1)

(2)

a. In each effect, the vapor and liquid phases are in equilibrium.

e. Heat losses from any effect need not be considered.

k. Not consider any type of pump between the effects.

ℎ�

functionality of **h1** with respect to another point (2) as:

of concentration and temperature:

� = ������ ���

ℎ� = ℎ� �

c. The impact of the hydrostatic head on the liquid boiling point is neglected.

the product temperature, the optimal configuration uses a distinct liquid flowpattern that often differs from the conventional forward and backward feed and leads to reasonable savings (Irahola & Cerdá, 1996)

Among the results should be noted that: the splitting of the flow of live steam can be a better alternative than the traditional cascade of steam; the best fixed cost curve is not always a monotonous increasing; the correct distribution of the areas of the effects of the MEES, the appropriate operating conditions and the correct choice of liquid and vapor flowpatterns, are the determining factors in the optimal design of the MEES.

Perhaps the greatest disadvantage found using the method proposed here to solve the formulated mathematical model (MINLP type) for optimal synthesis of the MEES is the presence of numerous local optimal what makes it difficult to obtain optimal Global.

#### **2. Evaporation of an aqueous solution of caustic soda**

Among the solutions of industrial interest that present a high increase in boiling point are sodium hydroxide solutions (caustic soda). The concentration of these substances by evaporation, presents significant disadvantages due to the characteristics of the caustic solutions, namely:


Since the transfer of heat (U) of liquor film coefficient, depending on the speed of the caustic solution through tubes (among other variables), usually, seeks a high speed in order to obtain a large coefficient (Kern, 1999).

According to the literature, it has taken as standard for the concentration of caustic soda, a evaporation system of two or three effects operate in backward feeed (Kern, 1999). In this study, found that the structure in counterflow or backward feed, obviously presents a high performance, but is not the best. In order to confirm what was said, is going to solve a problem.

#### **3. Optimal synthesis of a multi-effect evaporation system for the concentration of caustic soda**

#### **3.1 Problem**

Find the optimal MEES to concentrate 30040 lb/h (13626 kg/h) of an aqueous solution of sodium hydroxide from 10 to 50% by weight. The type of used evaporator is long vertical tube with forced circulation. Available in the plant: live steam boiler to 63.69 Psia (4.48 kg/cm2). The allowable minimum absolute pressure in an effect is of 1.942 Psia (0.1365 kg/cm2) (Geankoplis, 1983).

#### **3.2 Mathematical model**

380 Advances in Chemical Engineering

the product temperature, the optimal configuration uses a distinct liquid flowpattern that often differs from the conventional forward and backward feed and leads to reasonable

Among the results should be noted that: the splitting of the flow of live steam can be a better alternative than the traditional cascade of steam; the best fixed cost curve is not always a monotonous increasing; the correct distribution of the areas of the effects of the MEES, the appropriate operating conditions and the correct choice of liquid and vapor flowpatterns,

Perhaps the greatest disadvantage found using the method proposed here to solve the formulated mathematical model (MINLP type) for optimal synthesis of the MEES is the

Among the solutions of industrial interest that present a high increase in boiling point are sodium hydroxide solutions (caustic soda). The concentration of these substances by evaporation, presents significant disadvantages due to the characteristics of the caustic

Have a high boiling point elevation (BPE) which implies a great loss in the temperature

Concentrated solutions are highly viscous, which severely reduces the rate of heat

 They can have detrimental effects on steel, causing what is called caustic fragility. In addition, they may require removal of large amounts of salt when the solution is

Since the transfer of heat (U) of liquor film coefficient, depending on the speed of the caustic solution through tubes (among other variables), usually, seeks a high speed in order to

According to the literature, it has taken as standard for the concentration of caustic soda, a evaporation system of two or three effects operate in backward feeed (Kern, 1999). In this study, found that the structure in counterflow or backward feed, obviously presents a high performance, but is not the best. In order to confirm what was said, is going to solve a

Find the optimal MEES to concentrate 30040 lb/h (13626 kg/h) of an aqueous solution of sodium hydroxide from 10 to 50% by weight. The type of used evaporator is long vertical tube with forced circulation. Available in the plant: live steam boiler to 63.69 Psia (4.48 kg/cm2). The allowable minimum absolute pressure in an effect is of 1.942 Psia (0.1365

**3. Optimal synthesis of a multi-effect evaporation system for the** 

presence of numerous local optimal what makes it difficult to obtain optimal Global.

are the determining factors in the optimal design of the MEES.

**2. Evaporation of an aqueous solution of caustic soda** 

transfer in natural circulation evaporators.

savings (Irahola & Cerdá, 1996)

solutions, namely:

difference available.

obtain a large coefficient (Kern, 1999).

**concentration of caustic soda** 

kg/cm2) (Geankoplis, 1983).

concentrated.

problem.

**3.1 Problem** 

The scope of the rigorous mathematical model is limited by the following assumptions adopted in the formulation:


To solve the mathematical model and find the optimal design should be available before, the mathematical expressions for all dependent variables: enthalpy of steam (H), enthalpy of solution of soda caustic (h), latent heat of vaporization of the water (λ), overall heat transfer coefficient (U), temperature of the solution in the evaporator (T) and costs of forced circulation evaporator, barometric condenser multijet, surface condenser and heat exchanger. In general, useful information is available in graphics, which presented various authors cited in the bibliography, but there are no equations of those curves. These situations and other problems are resolved below. According to Standiford (1963), in forced circulation evaporators, film transfer coefficient (h) on the side of the liquid, can be calculated from the conventional Dittus-Boelter equation for forced circulation when there is no boiling.

$$\frac{hD}{k} = \ 0.0023 \left(\frac{DG}{\mu}\right)^{0.8} \left(\mu \frac{Cp}{k}\right)^{0.4} \tag{1}$$

If there is this equation for two points and combine both equations, you can find the functionality of **h1** with respect to another point (2) as:

$$h\_1 = \; h\_2 \left(\frac{\mu\_2}{\mu\_1}\right)^{0.4} \left(\frac{\mathcal{C}\_{p1}}{\mathcal{C}\_{p2}}\right)^{0.4} \tag{2}$$

As the overall heat transfer coefficient U is practically determined by the film coefficient h fluid side, the above equation can be used to obtain a correlation for U. In Geankoplis for T=105 °F and X = 0.5, data is U = 400 (Btu/h ft2 °F). T and X is obtained from a graph (Horvath, 1985) μ2 = 22.84 centipoise. Then, for five values of concentration and six temperature values are obtained from graphics (Horvath, 1985) the values of Cp1 and μ1. With the data obtained can be calculated according to the above expression, the overall heat transfer coefficient U1 with reference to U2. Thus, U is plotted vs. X (Fig. 1) and U vs. T (Fig. 2). Finally, using the triple X, T, Ui, can be found by regression, the functionality of U in terms of concentration and temperature:

$$\begin{aligned} U &= 1254.780865 - 4954.6700 \, X + 4832.059524 \, X^2 + 6.321549 \, T \\ &- 4.30974 \, XT \left[ \frac{Btu}{hft\_2 \, ^\circ F} \right] \end{aligned} \tag{3}$$

$$\begin{aligned} h &= -10.250000 \ -319.591837 \ X + 939.795918 \ X^2 \\ &+ 0.963929 \ T \ -0.335714 \ XT \ \left[ \frac{Btu}{lb} \right] \end{aligned} \tag{4}$$

$$C\_{\rm CMB} = 4,753104 + 2,480885 \, w + 0,281818 \, w^2 \, \left(10^3 \, \text{LSD}\right) \tag{5}$$

Optimal Synthesis of Multi-Effect Evaporation

equipment are presented.

fixed.

(°C)

Temperature °F

Condensers

Vaporization

Solvent Latent Heat of

costs of utilities and equipment.

Systems of Solutions with a High Boiling Point Rise 385

The information of the data used in the resolution of the problem is presented in table 1. Also, the thermodynamic properties of the remaining functions and costs of services and

In order to consider probable situations that could be presented in the industry, will study the cases in which the temperature of the weak solution (feed) is equal, higher or lower than the temperature of the strong solution (product). In the three cases, only change the values of the inlet of the weak solution temperature. The rest of the parametric conditions remain

Feed and product conditions

Operation conditions

Cool water (min) 89.6 (32.0) (max) 107.6 (42.0) Effect (min) 125.0 (51.7) (max) 294.6 (145.9)

Thermodynamic properties

Operating temperature at effect i Ti= (1 + 0.1419526 Xi )Tvi - 9.419608 Xi + 271.3627 Xi

Costs of utilities and equipment

Table 1. Data for example and, functional expression of thermodynamic properties and the

λi= 1104 – 0.65 Tvi [Btu/lb]

Case II 130 (54.4) 130 (54.4)

Temperature °F (°C)

Minimum allowable temperature difference °F (°C)

2 [°F]

Item Feed Product Flow rate lb/h (Kg/h) 30040 (13626) 60080 (2725.2)

Weight fraction 0.1 0.5

Case I 180 (82.2)

Case III 80 (26.7)

Steam 296.6 (147.0)

18.0 (10) Heat Exchangers

Specific vapor enthalpy Hi= 1075 +0.3466 Tvi [Btu/lb]

Steam 2.922 10�� (USD/Btu) 2.104E-2 Cool water (305 K) 1.952 10�� (USD/Btu) 1.405E-3

Surface condenser 1092.�� ��.��(USD) Heat exchanger 11��.1� ��.��(USD)

Fig. 4. Installed Cost of multijet barometric condenser.

$$\text{low} = \frac{Q}{500 \text{( $T\_s - T\_w - t\_a$ )}} \text{(Gpm)}\tag{6}$$

where:

*Q* = heat load, Btu / h *Ts* = temperature of saturated steam, °-F *Tw* = temperature of cooling water, °-F *ta* = 15 °-F = degree of approximation at *Ts*

The cost of forced circulation evaporator is obtained based on information reported by Maloney, 2008:

$$CE\_{FC} = 2420.5 \, A^{0.7121} \, U \text{\u0} \,\tag{7}$$

Fig. 5. Evolution of total operating costs.

Fig. 4. Installed Cost of multijet barometric condenser.

0

10

20

**Installed Cost [10³ U\$S]** 

30

40

where:

*Q* = heat load, Btu / h

Maloney, 2008:

*Ts* = temperature of saturated steam, °-F *Tw* = temperature of cooling water, °-F *ta* = 15 °-F = degree of approximation at *Ts*

Fig. 5. Evolution of total operating costs.

� �

Data

Cuadratic regression

� 500��� � �� � ���

0 2 4 6 810

**Flow rate Water [k gal/min]** 

The cost of forced circulation evaporator is obtained based on information reported by

����� (6)

1979

���� � ���0�5������������ (7)

The information of the data used in the resolution of the problem is presented in table 1. Also, the thermodynamic properties of the remaining functions and costs of services and equipment are presented.

In order to consider probable situations that could be presented in the industry, will study the cases in which the temperature of the weak solution (feed) is equal, higher or lower than the temperature of the strong solution (product). In the three cases, only change the values of the inlet of the weak solution temperature. The rest of the parametric conditions remain fixed.


Table 1. Data for example and, functional expression of thermodynamic properties and the costs of utilities and equipment.

Optimal Synthesis of Multi-Effect Evaporation

364.126

Fig. 7. Comparison with Typical Flowpatterns.

330

0

50

100

150

**Annual Cost. (103 USD)**

200

250

300

350

340

350

**Total annual cost (103 USD)**

360

370

Fig. 8. Relative incidence of Operating and fixed Costs

Systems of Solutions with a High Boiling Point Rise 387

347.639

1

**3-effect MEES**

Total cost Utilities Effect Barometric C. Exchanger

340.45

337.835

FFA BFA BF FF SA SO

FFA BFA BF FF SA SO

354.417 353.584

#### **4. Analysis and discussion**

#### **4.1 Case I. Feed temperature higher than the temperature of the product (TF>TP )**

#### **4.1.1 Comparative analysis of the optimal solution found**

Adopted TF = 180 °F (82.2 °C). Coinciding with the generally accepted criterion of optimality for the evaporation of caustic soda, has been found that the optimum number of effects of the MEES is equal to three (Fig. 5). However, a mixed structure {2,1,3} has been found in the path of the current liquid instead of backward feed. The feed stream enters to the effect 2 and then continuous countercurrent to the effect 1. Then go out and circulates in forward feed to the effect 3. (mixed liquid flowpattern). In this new structure that is presented (Fig. 6), we see significant increases in boiling point of the solution: 10.5 °C and 40 °C, in effects 1 and 3 respectively.

Fig. 6. Optimal configuration three-effect MEES. (TAC=337835 USD)

The optimal solution will be the one with the lowest total annual cost (TAC). From this point of view, the classical structure proposed as an optimum solution: evaporation system of three effects of equal area arranged backward feed (BFA), is 4.31% more expensive than the optimal solution (SO) (Fig. 7). More, even if it is allowed to in the structure backward feed, the effects have distinct areas (BF), do not get a better result that the optimal solution found. The difference in cost is 4.03%. In the figure 7, we also present results for a forward feed evaporation system. This structure, in its classic form forward feed with effects of equal area (FFA), is 7.91% more expensive. Which could corroborate, in some way, because in the past the BFA structure was preferred. If is allowed that the effects have different areas in the structure forward feed (FF) is very interesting the result obtained. The correct distribution of areas has led to a decrease in the total cost. The decrease is so great that, now, the FF structure is better than any of the above structures in backward feed. Its total annual cost (TAC) is only 2.95% greater than the of the optimum solution found (Fig. 7). In general, it appears that whatever the structure of MEES, the operating cost is significantly greater than the fixed cost (Fig. 8). It is approximately 65% of the total cost.

Adopted TF = 180 °F (82.2 °C). Coinciding with the generally accepted criterion of optimality for the evaporation of caustic soda, has been found that the optimum number of effects of the MEES is equal to three (Fig. 5). However, a mixed structure {2,1,3} has been found in the path of the current liquid instead of backward feed. The feed stream enters to the effect 2 and then continuous countercurrent to the effect 1. Then go out and circulates in forward feed to the effect 3. (mixed liquid flowpattern). In this new structure that is presented (Fig. 6), we see significant increases in boiling point of the solution: 10.5 °C and 40 °C, in effects 1

**4.1 Case I. Feed temperature higher than the temperature of the product (TF>TP )** 

**4.1.1 Comparative analysis of the optimal solution found** 

Fig. 6. Optimal configuration three-effect MEES. (TAC=337835 USD)

the fixed cost (Fig. 8). It is approximately 65% of the total cost.

The optimal solution will be the one with the lowest total annual cost (TAC). From this point of view, the classical structure proposed as an optimum solution: evaporation system of three effects of equal area arranged backward feed (BFA), is 4.31% more expensive than the optimal solution (SO) (Fig. 7). More, even if it is allowed to in the structure backward feed, the effects have distinct areas (BF), do not get a better result that the optimal solution found. The difference in cost is 4.03%. In the figure 7, we also present results for a forward feed evaporation system. This structure, in its classic form forward feed with effects of equal area (FFA), is 7.91% more expensive. Which could corroborate, in some way, because in the past the BFA structure was preferred. If is allowed that the effects have different areas in the structure forward feed (FF) is very interesting the result obtained. The correct distribution of areas has led to a decrease in the total cost. The decrease is so great that, now, the FF structure is better than any of the above structures in backward feed. Its total annual cost (TAC) is only 2.95% greater than the of the optimum solution found (Fig. 7). In general, it appears that whatever the structure of MEES, the operating cost is significantly greater than

**4. Analysis and discussion** 

and 3 respectively.

Fig. 7. Comparison with Typical Flowpatterns.

Fig. 8. Relative incidence of Operating and fixed Costs

Optimal Synthesis of Multi-Effect Evaporation

0

0

2

4

6

**V (Ton/h)** 

8

10

12

5

10

15

**Area (m^2)** 

20

25

30

35

Systems of Solutions with a High Boiling Point Rise 389

3

4

012345

**Effect** 

Fig. 10. Area profiles in the 1 to 4-effect MEES (n: number of effects).

n=1

n=1

2

Fig. 11. Flow rate profiles in the 1 to 4-effect MEES (n: number of effects).

012345

3

4

**Effect** 

2

#### **4.1.2 Impact of the flow pattern**

In this case, the trajectory of the liquid stream is the determining factor in the performance of a given MEES. The benefit achieved is even greater than obtained by allowing the effects having different areas with each other. That said, is based on the result of the structure SA. This has the same flow pattern that the optimal solution, but the effects of evaporation system are of equal area. The increase in cost is only 0.8% compared to MEES optimum (Fig. 7). From the practical point of view, the alternative SA may be the best option.

#### **4.1.3 Profiles of the structural and parametric variables**

As it will be seen later, only in this case it can be seen some regularity in the curves of the structural and parametric variables. Furthermore, after reaching the optimal point generally the next curve is anomalous with respect to the preceding ones. In the last effect occurs the maximum concentration jump (Δ X) of the solution (Fig. 9). At the same time, it has the maximum area as shown in the curves of the 1 to 4-effect optimal MEES (Fig. 10). The flow rate of steam produced in the effects is approximately the same. However, this does not apply to MEES with greater number of effects than the optimal. (Fig. 11). The profile of the temperature to the optimum MEES of a different number of effects, does not have a regular aspect. However, the maximum temperature jump occurs in the last effect (Fig. 12).

Fig. 9. Concentration profiles in the 1 to 4-effect MEES. (n: number of effects)

In this case, the trajectory of the liquid stream is the determining factor in the performance of a given MEES. The benefit achieved is even greater than obtained by allowing the effects having different areas with each other. That said, is based on the result of the structure SA. This has the same flow pattern that the optimal solution, but the effects of evaporation system are of equal area. The increase in cost is only 0.8% compared to MEES optimum (Fig.

As it will be seen later, only in this case it can be seen some regularity in the curves of the structural and parametric variables. Furthermore, after reaching the optimal point generally the next curve is anomalous with respect to the preceding ones. In the last effect occurs the maximum concentration jump (Δ X) of the solution (Fig. 9). At the same time, it has the maximum area as shown in the curves of the 1 to 4-effect optimal MEES (Fig. 10). The flow rate of steam produced in the effects is approximately the same. However, this does not apply to MEES with greater number of effects than the optimal. (Fig. 11). The profile of the temperature to the optimum MEES of a different number of effects, does not have a regular aspect. However, the maximum temperature jump occurs in the last effect

7). From the practical point of view, the alternative SA may be the best option.

Fig. 9. Concentration profiles in the 1 to 4-effect MEES. (n: number of effects)

012345

3

<sup>2</sup><sup>4</sup>

**Effect** 

**4.1.3 Profiles of the structural and parametric variables** 

n=1

**4.1.2 Impact of the flow pattern** 

0

0.1

0.2

0.3

**X (Kg/Kg)** 

0.4

0.5

0.6

(Fig. 12).

Fig. 10. Area profiles in the 1 to 4-effect MEES (n: number of effects).

Fig. 11. Flow rate profiles in the 1 to 4-effect MEES (n: number of effects).

Optimal Synthesis of Multi-Effect Evaporation

Fig. 13. Relative incidence of Operating and fixed Costs.

0

100

200

300

400

**Annual Cost [10³ USD]**

500

600

700

Fig. 14. Optimal configuration four-effect MEES. (TAC=351082 USD)

Systems of Solutions with a High Boiling Point Rise 391

Total Cost

Utility Cost

Capital Cost

0123456

**Number of effects**

Fig. 12. Temperature profiles in the 1 to 4-effect MEES (n: number of effects).

#### **4.2 Case II. Feed temperature equal than the temperature of the product (TF = TP)**

#### **4.2.1 Impact of the splitting of the live steam flow rate on the optimal solution**

The result obtained when TF = TP, is different, not only to the found for the case I, but also with respect to the classical position. Found structure is highly innovative and simple in its conception.

If you look at the evolution of the total annual cost (TAC) curve of the 1, 2, 3, 4 and 5 effect Optimal MEES, it was found that the four-effect MEES is that of lower cost (optimal quasiglobal) (Fig. 13). The flow pattern is backward feed and in this aspect, this result coincides with the classical motion, but not with the number of effects: proposed here a four-effect MEES, instead of three. However, this new proposal, would not be really the best alternative, if it was not associated to the new steam flow pattern proposed (Fig. 14). It emphasizes, splitting in the live steam flow pattern, it enters parallel to the effects 1 and 2; the by-passing effect 2 by the vapor stream from effect 1 and finally, the mixing of vapor streams from effects 1 and 2 for heating effect 3.

Against, this new trajectory of the flow of steam, first doubt that occurs, is the performance of this configuration against the unifilar cascade of high thermodynamic efficiency.

#### **4.2.2 Comparative study of proposed flow pattern with respect to traditional configurations**

For the purpose of explaining the improvement achieved, compares the structure backward feed with effects of different areas (BF) and the optimal solution (SO). Both structures have equal number of effects and same trajectory of the liquid flow and only differ in the

Fig. 12. Temperature profiles in the 1 to 4-effect MEES (n: number of effects).

2

conception.

**configurations** 

streams from effects 1 and 2 for heating effect 3.

60

70

80

90

n=1

100

110

**T (°C)** 

120

130

140

**4.2 Case II. Feed temperature equal than the temperature of the product (TF = TP) 4.2.1 Impact of the splitting of the live steam flow rate on the optimal solution** 

The result obtained when TF = TP, is different, not only to the found for the case I, but also with respect to the classical position. Found structure is highly innovative and simple in its

012345

3

4

**Effect** 

If you look at the evolution of the total annual cost (TAC) curve of the 1, 2, 3, 4 and 5 effect Optimal MEES, it was found that the four-effect MEES is that of lower cost (optimal quasiglobal) (Fig. 13). The flow pattern is backward feed and in this aspect, this result coincides with the classical motion, but not with the number of effects: proposed here a four-effect MEES, instead of three. However, this new proposal, would not be really the best alternative, if it was not associated to the new steam flow pattern proposed (Fig. 14). It emphasizes, splitting in the live steam flow pattern, it enters parallel to the effects 1 and 2; the by-passing effect 2 by the vapor stream from effect 1 and finally, the mixing of vapor

Against, this new trajectory of the flow of steam, first doubt that occurs, is the performance

For the purpose of explaining the improvement achieved, compares the structure backward feed with effects of different areas (BF) and the optimal solution (SO). Both structures have equal number of effects and same trajectory of the liquid flow and only differ in the

of this configuration against the unifilar cascade of high thermodynamic efficiency.

**4.2.2 Comparative study of proposed flow pattern with respect to traditional** 

**Number of effects**

Fig. 13. Relative incidence of Operating and fixed Costs.

Fig. 14. Optimal configuration four-effect MEES. (TAC=351082 USD)

Optimal Synthesis of Multi-Effect Evaporation

385.85

Fig. 15. Comparison with Typical Flowpatterns.

340

**103 USD**

360

**Total Annual Cost [10³ USD]**

380

Fig. 16. Comparison with Typical Flowpatterns (Case II).

is almost doubled with respect to other effects.

the effect 2 of the four effect MEES (Fig. 18), through the use of live steam in the effect. Thermal jumps that are achieved with the optimum structure are higher that in the triple and quintuple-effect MEES. In addition, it should be noted that the thermal jump in effect 2

**Total Cost Utiliies Cost Fixed Cost**

Systems of Solutions with a High Boiling Point Rise 393

**Estructure MEES**

366.03 365.16

Three effect MEES Four effect MEES

351.08

FFA FF BFA=SA BF SO

FFA FF BFA; SA BF SO

380.26

trajectory of the flow of steam. Two comparisons were made: one relating to the cost of auxiliary services and the other with respect to the cost of the effects, which is almost all of the fixed cost.

#### **4.2.3 Energy efficiency and fixed cost of the traditional structure**

The BF structure has a higher efficiency since the cost of auxiliary services is 174963 USD, 21.30% less than the cost for the **SO**. On the other hand, the cost of the effects is 193008 USD, i.e. 167,8% of the respective cost observed in SO. The net result of the comparison of the total costs, indicates that the BF structure is 8.9% more expensive than SO. With these results, following the classical position we can say, that BF 4 effects is not better than the optimum found (SO), because the MEES should be a structure BFA 3 effects, not four, which was used to compare. Therefore, will be then verified the validity of this rule, for the case study.

#### **4.2.4 Difference in the number of effects due to non-traditional flow patterns**

The optimal number of effects found by the mathematical model does not coincide with the optimal number for BFA and FFA traditional structures or even structures BF and FF. This explains why in the absence of a mathematical model to explore the multiple alternatives of design, the best answer to the problem was until now, a countercurrent system.

It was found that the optimum number of effects to structures backward feed and forward feed is 3. However, the developed model proposes the structure SO of 4 effects as the best solution. Therefore, to verify the quality of it, is advisable to compare the best results found for each structure.

#### **4.2.5 The optimal solution compared to traditional structures**

The total cost of the BFA MEES is 4.2% more than the optimal. Therefore, the structure and number of effects, traditionally proposed do not seem to be the most appropriate. Then, one might think that if you remove the restriction of equal area of the effects, could be improved, significantly, the current result. The results show that the BF structure of three effects is 4.1% more expensive than SO (Fig. 15). However, despite the difference in the number of effects, is convenient to analyze in more detail these recent results.

Structurally, BF and SO differ only in the flow pattern of steam. The cost of the auxiliary services of BF is 1.0% lower than the SO. On the other hand, the fixed cost is 13.4% greater determining to SO submit one minor TAC (Fig. 16).

#### **4.2.6 Profiles of the structural and parametric variables**

This case is characterized because the profiles of the process variables, for the various intermediate optimal MEES, they have no similarity among themselves. In particular, notes that the optimal solution presents the most discordant curve with respect to the others.

The temperature profile is irregular with temperature differences between effects nonuniform, being the most important jump located between 2 and 3 effect, following the drop of temperature effects 3 and 4, both heated with secondary steam. (Fig. 17). The greater temperature difference between the heating steam and the solution to evaporate, occurs in

trajectory of the flow of steam. Two comparisons were made: one relating to the cost of auxiliary services and the other with respect to the cost of the effects, which is almost all of

The BF structure has a higher efficiency since the cost of auxiliary services is 174963 USD, 21.30% less than the cost for the **SO**. On the other hand, the cost of the effects is 193008 USD, i.e. 167,8% of the respective cost observed in SO. The net result of the comparison of the total costs, indicates that the BF structure is 8.9% more expensive than SO. With these results, following the classical position we can say, that BF 4 effects is not better than the optimum found (SO), because the MEES should be a structure BFA 3 effects, not four, which was used to compare. Therefore, will be then verified the validity of this rule, for the case study.

The optimal number of effects found by the mathematical model does not coincide with the optimal number for BFA and FFA traditional structures or even structures BF and FF. This explains why in the absence of a mathematical model to explore the multiple alternatives of

It was found that the optimum number of effects to structures backward feed and forward feed is 3. However, the developed model proposes the structure SO of 4 effects as the best solution. Therefore, to verify the quality of it, is advisable to compare the best results found

The total cost of the BFA MEES is 4.2% more than the optimal. Therefore, the structure and number of effects, traditionally proposed do not seem to be the most appropriate. Then, one might think that if you remove the restriction of equal area of the effects, could be improved, significantly, the current result. The results show that the BF structure of three effects is 4.1% more expensive than SO (Fig. 15). However, despite the difference in the number of effects,

Structurally, BF and SO differ only in the flow pattern of steam. The cost of the auxiliary services of BF is 1.0% lower than the SO. On the other hand, the fixed cost is 13.4% greater

This case is characterized because the profiles of the process variables, for the various intermediate optimal MEES, they have no similarity among themselves. In particular, notes that the optimal solution presents the most discordant curve with respect to the others.

The temperature profile is irregular with temperature differences between effects nonuniform, being the most important jump located between 2 and 3 effect, following the drop of temperature effects 3 and 4, both heated with secondary steam. (Fig. 17). The greater temperature difference between the heating steam and the solution to evaporate, occurs in

**4.2.4 Difference in the number of effects due to non-traditional flow patterns** 

design, the best answer to the problem was until now, a countercurrent system.

**4.2.5 The optimal solution compared to traditional structures** 

is convenient to analyze in more detail these recent results.

**4.2.6 Profiles of the structural and parametric variables** 

determining to SO submit one minor TAC (Fig. 16).

**4.2.3 Energy efficiency and fixed cost of the traditional structure** 

the fixed cost.

for each structure.

Fig. 15. Comparison with Typical Flowpatterns.

Fig. 16. Comparison with Typical Flowpatterns (Case II).

the effect 2 of the four effect MEES (Fig. 18), through the use of live steam in the effect. Thermal jumps that are achieved with the optimum structure are higher that in the triple and quintuple-effect MEES. In addition, it should be noted that the thermal jump in effect 2 is almost doubled with respect to other effects.

Optimal Synthesis of Multi-Effect Evaporation

**V [10³ kg/h]** 

**Area [m²]** 

had been resolved for a fixed range area (100, 1000 ft2).

Fig. 19. Area profiles in the 1 to 5 effect MEES (n: number of effects)

n=1

n=1

4 5

**Effect** 

Fig. 20. Flow rate profiles in the 1 to 5-effect MEES (n: number of effects).

**Effect** 

3 4

Systems of Solutions with a High Boiling Point Rise 395

Similarly, the profile of the area for each MEES, is far from being uniform, with the highest values located in the lower thermal effects. (Fig. 19). Should be mentioned, that the problem

The flow rate of solvent evaporated in each effect is approximately the same in the double and triple-effect MEES (Fig. 20), where the chosen structure is backward feed. But in the optimal solution and after this, the values of vapor flow rate are far from each other. Although it can be seen that, in SO the curve is regular and decreasing with temperature

**[10³ lb/h]** 

**[ft²]** 

Fig. 17. Temperature profile optimal MEES for n=1 to 5 effects.

Fig. 18. Temperature difference profiles between the condensation and evaporation chambers in each effect for 2, 3, 4 and 5 effects.

n=2

**Efecto** 

[°F]

**[°F]** 

Fig. 17. Temperature profile optimal MEES for n=1 to 5 effects.

n=1

**T [°C]** 

Fig. 18. Temperature difference profiles between the condensation and evaporation

**Effect** 

chambers in each effect for 2, 3, 4 and 5 effects.

[°C]

Similarly, the profile of the area for each MEES, is far from being uniform, with the highest values located in the lower thermal effects. (Fig. 19). Should be mentioned, that the problem had been resolved for a fixed range area (100, 1000 ft2).

The flow rate of solvent evaporated in each effect is approximately the same in the double and triple-effect MEES (Fig. 20), where the chosen structure is backward feed. But in the optimal solution and after this, the values of vapor flow rate are far from each other. Although it can be seen that, in SO the curve is regular and decreasing with temperature

Fig. 19. Area profiles in the 1 to 5 effect MEES (n: number of effects)

Fig. 20. Flow rate profiles in the 1 to 5-effect MEES (n: number of effects).

Optimal Synthesis of Multi-Effect Evaporation

areas of effects.

Table 2.).

Systems of Solutions with a High Boiling Point Rise 397

Evolution of cost curves show a singularity with respect to the previously analyzed cases. Now, to increase the number of effects to reduce the TAC, the consumption of live steam begins to increase (Fig. 22) rather than continue to decline with the addition of a new effect, and the cost of the effects decreasing, rather than continue to increase. Isn't traditional behavior: a curve monotone decreasing for the cost of the live steam and one monotone increasing for the cost of the effects. Here is not complied with this scheme because of the significant reduction of the U coefficient and the high increase of the boiling point (BPE) in the effects of greater concentration. It is now "pays" with live steam part of the savings in the

Fig. 22. Evolution of the TAC and its component terms with the number of effects (Case III).

**4.3.2 Structure and distribution of temperature, concentration and heat transfer area**  A structural analysis allows us to appreciate that the cuadruple-effect MEES results have added an effect, between the first and the second of the triple-effect MEES (Figs. 23 and 24,

Looking to reduce the total cost, the new effect requires one of the lowest values of area of heat transfer (Fig. 25). For this purpose, it operates with a large temperature difference; almost double that for the remaining effects (Fig. 26). On the other hand, contrary to expectations, their presence causes a slight increase in the consumption of live steam and allows at the same time, an increase in the thermal jump in the last two effects of, approximately, 6 °F (3.3 °C).

effects, i.e. the greater evaporated flow rate occurs in the effect 4. The profile of the global coefficient U is similar: lower in effect 1 and higher in the effect 4. It is important to clarify that, the values of the flow rate of steam produced in each effect are not directly indicative parameters. Yes it is a relative measure, for example, the percentage of the current liquid evaporates. Thus, the percentage amount of solvent evaporated, with respect to the flow rate of solvent that enters each effect, is maximum in the second effect (41%), followed in decreasing order, the third effect (34%), the first (31%) and finally the fourth (25%).

The concentration curve of the solution is monotone increasing, considering the path of the liquid stream (Fig. 21). The biggest jump in percentage concentration occurs in the second effect (69 % ), then in descending order, the third effect (51 % ), the first (45 %) and the fourth (34 % ).

Fig. 21. Concentration profiles in the 1 to 5-effect MEES

The recent analysis of incremental concentration, the carried out for the steam produced in the effects, and the biggest jump thermal observed in effect 2 of the Optimal MEES show the importance of this effect. Its presence is the root cause of the improvement achieved in SO. This is achieved, thanks to the optimal design of the MEES, which allows an adequate relationship of the variables that define the system.

#### **4.3 Case III. Temperature of the weak solution lower than strong solution (TF < TP)**

#### **4.3.1 Uniqueness of the cost curves**

This case presents great similarity with the Case II. The optimal effect is the same (4) and intermediate structures found for n ≤ 4 are almost identical. The best configuration for the liquid stream, given by {4,3,2,1}, as in Case II, use live steam to heat the first two effects.

effects, i.e. the greater evaporated flow rate occurs in the effect 4. The profile of the global coefficient U is similar: lower in effect 1 and higher in the effect 4. It is important to clarify that, the values of the flow rate of steam produced in each effect are not directly indicative parameters. Yes it is a relative measure, for example, the percentage of the current liquid evaporates. Thus, the percentage amount of solvent evaporated, with respect to the flow rate of solvent that enters each effect, is maximum in the second effect (41%), followed in

The concentration curve of the solution is monotone increasing, considering the path of the liquid stream (Fig. 21). The biggest jump in percentage concentration occurs in the second effect (69 % ), then in descending order, the third effect (51 % ), the first (45 %) and the fourth (34 % ).

The recent analysis of incremental concentration, the carried out for the steam produced in the effects, and the biggest jump thermal observed in effect 2 of the Optimal MEES show the importance of this effect. Its presence is the root cause of the improvement achieved in SO. This is achieved, thanks to the optimal design of the MEES, which allows an adequate

012345

3

4

5

**Effect**

**4.3 Case III. Temperature of the weak solution lower than strong solution (TF < TP)** 

This case presents great similarity with the Case II. The optimal effect is the same (4) and intermediate structures found for n ≤ 4 are almost identical. The best configuration for the liquid stream, given by {4,3,2,1}, as in Case II, use live steam to heat the first two effects.

decreasing order, the third effect (34%), the first (31%) and finally the fourth (25%).

Fig. 21. Concentration profiles in the 1 to 5-effect MEES

2

relationship of the variables that define the system.

**4.3.1 Uniqueness of the cost curves** 

0

0.1

0.2

0.3

**X [kg/kg]**

0.4

0.5

n=1

0.6

Evolution of cost curves show a singularity with respect to the previously analyzed cases. Now, to increase the number of effects to reduce the TAC, the consumption of live steam begins to increase (Fig. 22) rather than continue to decline with the addition of a new effect, and the cost of the effects decreasing, rather than continue to increase. Isn't traditional behavior: a curve monotone decreasing for the cost of the live steam and one monotone increasing for the cost of the effects. Here is not complied with this scheme because of the significant reduction of the U coefficient and the high increase of the boiling point (BPE) in the effects of greater concentration. It is now "pays" with live steam part of the savings in the areas of effects.

#### **4.3.2 Structure and distribution of temperature, concentration and heat transfer area**

A structural analysis allows us to appreciate that the cuadruple-effect MEES results have added an effect, between the first and the second of the triple-effect MEES (Figs. 23 and 24, Table 2.).

Looking to reduce the total cost, the new effect requires one of the lowest values of area of heat transfer (Fig. 25). For this purpose, it operates with a large temperature difference; almost double that for the remaining effects (Fig. 26). On the other hand, contrary to expectations, their presence causes a slight increase in the consumption of live steam and allows at the same time, an increase in the thermal jump in the last two effects of, approximately, 6 °F (3.3 °C).

Optimal Synthesis of Multi-Effect Evaporation

Systems of Solutions with a High Boiling Point Rise 399

**Alternative Optimal Solutions (Case III)**

**Equal area effects MEES: SA**

**(i) [°C] [10³ kg/h] [kg/kg] [m²]**

**2** 128.93 107.28 0 0 2.979 0.340 15.33

**Backward feed MEES. Optimal area effects: BF Effect T Tv Vs L V X F 1** 130.40 87.83 4.000 0 3.232 0.500 29.47 **2** 74.99 67.71 0 0 3.710 0.198 26.22 **3** 54.76 51.67 0 13.626 3.039 0.129 22.57 **Backward feed MEES. Equal area effects: BFA Effect T Tv Vs L V X F**

**Effect T Tv Vs L V X F**

**1** 137.99 94.92 4.105 0 3.232 0.500

**3** 80.62 73.39 0 0 3.665 0.195 **4** 54.72 51.67 0 13.626 2.971 0.128

**1** 128.39 85.95 3.984 0 4.126 0.500

**3** 54.77 51.67 0 13.626 3.055 0.129

**1** 135.12 130.25 4.870 13.626 3.530 0.135

**3** 91.67 51.67 0 0 3.680 0.500

Steam Flow rate)

**2** 121.62 111.99 0 0 3.691 0.213 24.40

Table 2. Results of alternative evaporation systems. (Vs: Live steam flow rate. V: Secondary

**2** 73.12 65.84 0 0 3.719 0.199 26.54

**Forward feed MEES. Optimal area effects: FF Effect T Tv Vs L V X F 1** 132.77 127.95 4.822 13.626 3.517 0.135 20.37 **2** 116.53 107.02 0 0 3.714 0.213 18.96 **3** 91.67 51.67 0 0 3.671 0.500 32.69 **Forward feed MEES. Equal area effects: FFA Effect T Tv Vs L V X F**

Fig. 23. Optimal Structure for three effects. (Local optimun).

Fig. 24. Optimal Structure for four effects. (Optimal Solution).

**0.500 0.198 0.129**

**4.000 4.152 3.710 3.039**

**67.7 51.7**

**26.22 22.56**

**F**

Fig. 23. Optimal Structure for three effects. (Local optimun).

**A(m )=29.46 2**

**P**

Fig. 24. Optimal Structure for four effects. (Optimal Solution).


Table 2. Results of alternative evaporation systems. (Vs: Live steam flow rate. V: Secondary Steam Flow rate)

Optimal Synthesis of Multi-Effect Evaporation

0

50

100

150

200

**Annual Cost [10³ USD ]**

250

300

350

400

450

Systems of Solutions with a High Boiling Point Rise 401

**FFA FF BFA BF SA SO**

0

10

5

20

**[10³ lb/h]** 

30

Fig. 27. Impact of the liquid flow pattern and the distribution of heat transfer area in

**Total Cost Utilities Effects Condenser Exchanger**

Profile of concentration of the solution in the optimal MESS shows a non-uniform increase of concentration along the evaporator train (Fig. 29). Thus, following the path of the current liquid is seen a percentage increase of 28% in the effect 4, 53% in the effect 3, 73% in effect 2 and 46% in the first effect. Similarly as stated in Case II, the largest concentration jump

**Efect**

012345

3 4

different items of the total annual cost of the MEES (Case III).

n=1

2

Fig. 28. Flow rate steam profiles in the 1 to 5 effect MEES.

0

5

**V [10³ kg/h]** 

10

Fig. 25. Area profiles in the 1 to 5-effect MEES.

Fig. 26. Temperature profiles in the 1 to 5-effect MEES.

The net result is a drastic decrease in the area of thermal transfer, as you can see through the cost of the effects (Fig. 27). Moreover, the distribution of the heat transfer area curve changes dramatically, from being monotone decreasing for a triple-effect MEES to have a nonmonotonic behavior in the Optimal MEES. Presents a maximum in the effect 3 and a minimum in the effect 1, where the product is removed and in which the coefficient U takes its smallest value (Fig. 25). This was achieved by lowering the flow rate of the solvent evaporated in the first two effects, especially in the first (Fig. 28).

3

012345

**Effect**

0

**[°F]** 

50

100

150

200

4 5

**[ft²]** 

250

300

350

400

Fig. 25. Area profiles in the 1 to 5-effect MEES.

n=1

n=1

2

0

45

65

85

105

**T [°C]** 

125

145

10

20

**Area [m²]** 

30

Fig. 26. Temperature profiles in the 1 to 5-effect MEES.

evaporated in the first two effects, especially in the first (Fig. 28).

2

The net result is a drastic decrease in the area of thermal transfer, as you can see through the cost of the effects (Fig. 27). Moreover, the distribution of the heat transfer area curve changes dramatically, from being monotone decreasing for a triple-effect MEES to have a nonmonotonic behavior in the Optimal MEES. Presents a maximum in the effect 3 and a minimum in the effect 1, where the product is removed and in which the coefficient U takes its smallest value (Fig. 25). This was achieved by lowering the flow rate of the solvent

**Effect**

3 4 5

012345

Fig. 27. Impact of the liquid flow pattern and the distribution of heat transfer area in different items of the total annual cost of the MEES (Case III).

Fig. 28. Flow rate steam profiles in the 1 to 5 effect MEES.

Profile of concentration of the solution in the optimal MESS shows a non-uniform increase of concentration along the evaporator train (Fig. 29). Thus, following the path of the current liquid is seen a percentage increase of 28% in the effect 4, 53% in the effect 3, 73% in effect 2 and 46% in the first effect. Similarly as stated in Case II, the largest concentration jump

Optimal Synthesis of Multi-Effect Evaporation

feed (BF).

same.

**5. Conclusion** 

using a mathematical optimizer.

in the last effect, where in addition, the product leaves.

**4.3.3 Comparison with traditional configurations** 

of evaporation, including fixed and operational costs.

of the optimal flow pattern, were considered such as:

Systems of Solutions with a High Boiling Point Rise 403

occurs in effect 2, as opposed to what was observed in Case I, where the largest jump occurs

As in the case II, the optimum number of effects of optimal MEES is four and does not match the traditionally proposed by the classic bibliography: triple-effect countercurrent (BFA). If you seek the best solution between traditional structures with effects of equal area (Table 2), again confirms the superiority of the backward feed configuration and that three is the optimal number of effects. Comparatively, FFA is 11.4% more expensive that the structure BFA (Fig. 30). If in both structures allows you to optimize the distribution of the transfer area between the different effects, the improvement is not important. Of them, the best option (BF) is 4.1% more expensive than SO. However, the best configuration for a MEES whose only structural restriction is equality of areas of its effects (SA): is backward

In this paper, we solve the problem of designing a MEES, for the concentration of caustic soda, developing a rigorous mathematical model non-convex MINLP type and solving it

Unlike previous papers, the new formulation proposal incorporates as decision variables: (a) the trajectories of the steam and liquid flows along the evaporator train whose correct choice determines, in the opinion of different authors, as has been demonstrated in the resolution of the example here presented, the level of operational costs and investment of the MEES. It is also considered (b) the number of effects of the MEES, as was proved in the results presented is another critical design decision and (c) heat transfer area in each effect, without resorting to the hypothesis of equal areas on the effects that in many cases substantially increases the total cost of investment. As an evaluation criterion of alternative designs included in the solutions space of the problem, we used the total annual cost of the system

Other important aspects, not usually treated and much less simultaneously with the search

the rise of the boiling point of the solution and its dependency on temperature and concentration of the same, the variation of the overall heat transfer coefficient along the train following changes in the concentration and the temperature of each effect and the functional dependence of the heat of non-ideal solution with the concentration and temperature.

In addition, we studied other variants non-conventional design that arise by allowing: feed in parallel of the weak solution to two or more effects of the evaporator train, the entry of two or more liquid streams (even of different concentration) to a given effect and the derivation of a liquid solution branch around an effect to avoid its treatment in the

A serious drawback found in the resolution of the mathematical model is the presence of many stationary points (optimal local), and also the great influence of the initial point.

Fig. 29. Concentration profiles in the 1 to 5 effect MEES.

Fig. 30. Impact of the liquid and vapor flow pattern and area effects on the total annual cost of the MEES (Case III).

occurs in effect 2, as opposed to what was observed in Case I, where the largest jump occurs in the last effect, where in addition, the product leaves.

#### **4.3.3 Comparison with traditional configurations**

As in the case II, the optimum number of effects of optimal MEES is four and does not match the traditionally proposed by the classic bibliography: triple-effect countercurrent (BFA). If you seek the best solution between traditional structures with effects of equal area (Table 2), again confirms the superiority of the backward feed configuration and that three is the optimal number of effects. Comparatively, FFA is 11.4% more expensive that the structure BFA (Fig. 30). If in both structures allows you to optimize the distribution of the transfer area between the different effects, the improvement is not important. Of them, the best option (BF) is 4.1% more expensive than SO. However, the best configuration for a MEES whose only structural restriction is equality of areas of its effects (SA): is backward feed (BF).

### **5. Conclusion**

402 Advances in Chemical Engineering

012345

 **Effect** 

3

4

5

Fig. 30. Impact of the liquid and vapor flow pattern and area effects on the total annual cost

3-effect MEES

377.65 376.744

371.169

4-effect MEES

362.084

**FFA FF BFA BF SA SO**

Fig. 29. Concentration profiles in the 1 to 5 effect MEES.

417.664

412.723

n=1

2

0

0.1

0.2

0.3

**X [kg/kg]** 

0.4

0.5

0.6

of the MEES (Case III).

350

370

390

**Total Annual Cost [10³ U\$S]**

410

430

In this paper, we solve the problem of designing a MEES, for the concentration of caustic soda, developing a rigorous mathematical model non-convex MINLP type and solving it using a mathematical optimizer.

Unlike previous papers, the new formulation proposal incorporates as decision variables: (a) the trajectories of the steam and liquid flows along the evaporator train whose correct choice determines, in the opinion of different authors, as has been demonstrated in the resolution of the example here presented, the level of operational costs and investment of the MEES. It is also considered (b) the number of effects of the MEES, as was proved in the results presented is another critical design decision and (c) heat transfer area in each effect, without resorting to the hypothesis of equal areas on the effects that in many cases substantially increases the total cost of investment. As an evaluation criterion of alternative designs included in the solutions space of the problem, we used the total annual cost of the system of evaporation, including fixed and operational costs.

Other important aspects, not usually treated and much less simultaneously with the search of the optimal flow pattern, were considered such as:

the rise of the boiling point of the solution and its dependency on temperature and concentration of the same, the variation of the overall heat transfer coefficient along the train following changes in the concentration and the temperature of each effect and the functional dependence of the heat of non-ideal solution with the concentration and temperature.

In addition, we studied other variants non-conventional design that arise by allowing: feed in parallel of the weak solution to two or more effects of the evaporator train, the entry of two or more liquid streams (even of different concentration) to a given effect and the derivation of a liquid solution branch around an effect to avoid its treatment in the same.

A serious drawback found in the resolution of the mathematical model is the presence of many stationary points (optimal local), and also the great influence of the initial point.

**16** 

*Politecnico di Torino* 

*K&E Srl Italy* 

**Optimization of Spouted Bed Scale-Up** 

**by Square-Based Multiple Unit Design** 

Among several configurations typical of gas-solids fluidization, spouted beds have demonstrated to be characterized by a number of advantages, namely a reduced pressure drop, a relatively lower gas flow rate, the possibility of handling particles coarser than the ones treated by bubbling fluidized beds. Additionally, significant segregation is prevented

Spouted beds appear to go through a revival, testified by a very recent and comprehensive book on the topic (Epstein & Grace, 2011). This renewed interest arises by implementing new concepts in scaling-up spouting contactors and devising potential applications to high temperature processes, noticeable examples being given by pyrolysis and gasification of biomass, kinetically controlled drying of moist seeds to guarantee the requested qualities

Fluidization is a hydrodynamical regime in which a bed of solid particles is expanded and suspended by an upward fluid flow. This regime is established when the fluid velocity reaches a value corresponding to the minimum fluidization. The basic design of a fluidized unit is carried out by considering a vessel having a cross section of any shape (circular, squared or rectangular) with a perforated bottom which separates the volume holding the

the solid particle bulk exhibits a liquid-like behaviour: the surface of the solids remains

if two or more vessels operating in a fluidization regime are connected, the solids reach

in the presence of a side opening under the bed surface, particles gush as a liquid flow;

Fluidization, besides being influenced by the solid characteristics, depends on the physical properties of the fluid and its superficial velocity. When this parameter is very low, the fluid

heterogeneous bodies may float or sink, depending on their actual density.

Fluidized beds show a number of features which are summarized below: forces are in balance and there is no net force acting in the system;

**1. Introduction** 

by the peculiar hydraulic structure.

and polymer upgrading processes.

**2. Generalities on fluidization** 

solids from the lower gas plenum.

horizontal by tilting the vessel;

an identical hydrostatic level;

Giorgio Rovero, Massimo Curti and Giuliano Cavaglià

It has been found that in order to obtain the optimal design of lower total cost annual: (i) not always steam flow pattern should be the traditional unifilar cascade, could be useful to feed steam live in more than one effect, (ii) the heat transfer areas do not necessarily have to be equal, (iii) the fresh feed stream should not always come in the last effect evaporation train.

However, the synthesis of the optimal MEES only be achieved if you are optimized simultaneously structural, parametric, and operation variables.

#### **6. Acknowledgment**

I thank to SeCTER for their help in research.

I thank Liliana, my dear sister, for her love and permanent support throughout my life.

#### **7. References**

Foust, Alan S., *Principles of Unit Operations,* Second Ed. C. Wiley, New York, 1980.


## **Optimization of Spouted Bed Scale-Up by Square-Based Multiple Unit Design**

Giorgio Rovero, Massimo Curti and Giuliano Cavaglià *Politecnico di Torino K&E Srl Italy* 

#### **1. Introduction**

404 Advances in Chemical Engineering

It has been found that in order to obtain the optimal design of lower total cost annual: (i) not always steam flow pattern should be the traditional unifilar cascade, could be useful to feed steam live in more than one effect, (ii) the heat transfer areas do not necessarily have to be equal, (iii) the fresh feed stream should not always come in the last effect evaporation train. However, the synthesis of the optimal MEES only be achieved if you are optimized

I thank Liliana, my dear sister, for her love and permanent support throughout my life.

Systems Using Minimum Utility Insights - I. A Cascaded Heat Representation

Evaporator System. *Fifth International Symposium On Process Systems Engineering* 

Multiefecto para la Concentración de Licor de Caña de Azúcar. *IX Simposio Internacional en Aplicaciones de Informática. Infonor'96.* Antofagasta. Chile. 1996. Irahola Ferreira, Jaime A., Aplicación para Formular y Resolver Modelos Matemáticos de Sistemas de Evaporación Multiefecto. *Información Tecnológica*. Chile, 2008

Horvath, A. L., *Handbook of Aqueous Electrolyte Solutions, Physical Properties, Estimation and* 

Irahola Ferreira, Jaime A., & Jaime Cerdá. Optimal Synthesis Of A Multiple-Effect

Irahola Ferreira, Jaime A., & Jaime Cerdá. Síntesis Optima de un Sistema de Evaporación

Maloney. James O. *Perry's Chemical Engineers Handbook*, 8th Edn. McGraw-Hill, New York

Nishitani, H. & Kunugita, E., The Optimal Flow-Pattern of Multiple Effect Evaporator Systems, *Computers & Chemical Engineering*, Vol. 3 , pp. 261-268, 1979. Peters, Max Stone Klaus D. Timmerhaus. *Plantdesign and economics for chemical engineers.* 

Westerberg, A.W. & Hillebrand Jr., J. B. The Synthesis of Multiple-Effect Evaporator Systems

Using Minimum Utility Insights - II. Liquid Flow-Pattern Selection. *Computers &* 

Foust, Alan S., *Principles of Unit Operations,* Second Ed. C. Wiley, New York, 1980. Geankoplis, Ch. J., *Transport Processes and Unit Operations*, 2nd. Edition, 495-501, 1983. Hillebrand Jr., J. B. & Westerberg, A.W., The Synthesis of Multiple-Effect Evaporator

*Computers & Chemical Engineering*, Vol. 12, pp. 611-624, 1988.

Kern, D. Q. *Process Heat Transfer*. McGraw-Hill, New York .pp 375 - 452, 1999.

McGraw-Hill, Inc. Singapore, ISBN 0-07-100871-3, 1991.

*Chemical Engineering*, Vol. 12, pp. 625-636, 1988.

Standiford, Ferris C. Jr., W. L. Badger, Evaporation. *Chem. Engng,* 70, 158—176, 1963.

*Correlation Methods,* Wiley, C. New York, 1985.

simultaneously structural, parametric, and operation variables.

**6. Acknowledgment** 

**7. References** 

I thank to SeCTER for their help in research.

*(PSE)*, Korea, 1994.

2008.

Among several configurations typical of gas-solids fluidization, spouted beds have demonstrated to be characterized by a number of advantages, namely a reduced pressure drop, a relatively lower gas flow rate, the possibility of handling particles coarser than the ones treated by bubbling fluidized beds. Additionally, significant segregation is prevented by the peculiar hydraulic structure.

Spouted beds appear to go through a revival, testified by a very recent and comprehensive book on the topic (Epstein & Grace, 2011). This renewed interest arises by implementing new concepts in scaling-up spouting contactors and devising potential applications to high temperature processes, noticeable examples being given by pyrolysis and gasification of biomass, kinetically controlled drying of moist seeds to guarantee the requested qualities and polymer upgrading processes.

#### **2. Generalities on fluidization**

Fluidization is a hydrodynamical regime in which a bed of solid particles is expanded and suspended by an upward fluid flow. This regime is established when the fluid velocity reaches a value corresponding to the minimum fluidization. The basic design of a fluidized unit is carried out by considering a vessel having a cross section of any shape (circular, squared or rectangular) with a perforated bottom which separates the volume holding the solids from the lower gas plenum.

Fluidized beds show a number of features which are summarized below:


Fluidization, besides being influenced by the solid characteristics, depends on the physical properties of the fluid and its superficial velocity. When this parameter is very low, the fluid

Optimization of Spouted Bed Scale-Up by Square-Based Multiple Unit Design 407

In addition to properties of the fluidization medium, particle features play an essential role. A simple mapping was proposed (Geldart, 1973) to group particulate solid materials in four well-defined classes according to their hydrodynamic behaviour. To follow this categorization, Figure 2 shows the four different regions, proper of air/solids systems at ambient fluidization conditions and particles of the same shape, excluding pneumatic

A type or "aeratable": solids with small diameter and density lower than about 1400 kg/m3. The minimum fluidization velocity can be reached smoothly; then fine bubble

B type or "sandlike": these particles are coarser than the previous ones, ranging from 40 to 1000 m and densities from 1400 to 4000 kg/m3. A vigorous fluidization with large

C type or "cohesive": very fine powders, with a mean diameter generally lower than 50

D type or "spoutable": coarsest particles within a broad density range. These systems are characterized by a high permeability, which generates severe channelling and uneven gas distribution: the standard fluidization geometry should be modified to give

Fluidization is an operation characterized by several interesting peculiarities with desirable associated to non-optimal features. On one hand this technique guarantees a smooth and

transport conditions.

Fig. 2. Particle classification according to Geldart

The particles of the regions can be described as follows:

fluidization occurs at higher gas velocities.

m. Strong interparticle forces render fluidization difficult.

bubbles may be established.

origin to spouted systems.

**3. Fluidization versus spouting** 

merely percolates through the particles and no movement is induced, this condition being defined as "static bed". By rising the flow rate, frictional forces between particles and fluid increases: when the upward component of force counterbalances the particle weight, the minimum condition to expand the bed is reached. When all the particles are suspended by the fluid, the bed can be considered in a state of "incipient fluidization" and the pressure drop through any bed section equalizes the weight of the fluid and solids in that section. By further increasing the velocity, some phenomena of instability such as "bubbling or turbulent fluidization" may occur, depending on the system geometry and particle properties. In a gas-solid system operated at high fluid velocity, gas bubbles tend to coalesce and grow in volume during their upward travel; if the bed is not wide enough, a gas bubble can take all the vessel cross section, then the solid particles are lifted as a piston, giving origin to the so-called "flat slugging". This undesired occurrence easily happens with coarse particles as well with cohesive powders. Finally, when a critical value is reached, the velocity of the gas is high enough to transport individually or in clusters the bed particles in a "pneumatic conveying" fashion. These hydraulic regimes are schematically shown in Figure 1.

Fig. 1. Schematic representation of various fluidization regimes

merely percolates through the particles and no movement is induced, this condition being defined as "static bed". By rising the flow rate, frictional forces between particles and fluid increases: when the upward component of force counterbalances the particle weight, the minimum condition to expand the bed is reached. When all the particles are suspended by the fluid, the bed can be considered in a state of "incipient fluidization" and the pressure drop through any bed section equalizes the weight of the fluid and solids in that section. By further increasing the velocity, some phenomena of instability such as "bubbling or turbulent fluidization" may occur, depending on the system geometry and particle properties. In a gas-solid system operated at high fluid velocity, gas bubbles tend to coalesce and grow in volume during their upward travel; if the bed is not wide enough, a gas bubble can take all the vessel cross section, then the solid particles are lifted as a piston, giving origin to the so-called "flat slugging". This undesired occurrence easily happens with coarse particles as well with cohesive powders. Finally, when a critical value is reached, the velocity of the gas is high enough to transport individually or in clusters the bed particles in a "pneumatic conveying" fashion. These hydraulic regimes are schematically shown in

Fig. 1. Schematic representation of various fluidization regimes

Figure 1.

In addition to properties of the fluidization medium, particle features play an essential role. A simple mapping was proposed (Geldart, 1973) to group particulate solid materials in four well-defined classes according to their hydrodynamic behaviour. To follow this categorization, Figure 2 shows the four different regions, proper of air/solids systems at ambient fluidization conditions and particles of the same shape, excluding pneumatic transport conditions.

Fig. 2. Particle classification according to Geldart

The particles of the regions can be described as follows:

A type or "aeratable": solids with small diameter and density lower than about 1400 kg/m3. The minimum fluidization velocity can be reached smoothly; then fine bubble fluidization occurs at higher gas velocities.

B type or "sandlike": these particles are coarser than the previous ones, ranging from 40 to 1000 m and densities from 1400 to 4000 kg/m3. A vigorous fluidization with large bubbles may be established.

C type or "cohesive": very fine powders, with a mean diameter generally lower than 50 m. Strong interparticle forces render fluidization difficult.

D type or "spoutable": coarsest particles within a broad density range. These systems are characterized by a high permeability, which generates severe channelling and uneven gas distribution: the standard fluidization geometry should be modified to give origin to spouted systems.

#### **3. Fluidization versus spouting**

Fluidization is an operation characterized by several interesting peculiarities with desirable associated to non-optimal features. On one hand this technique guarantees a smooth and

Optimization of Spouted Bed Scale-Up by Square-Based Multiple Unit Design 409

 reduced pressure drop and lower gas flow rate required to attain solids motion with respect to the minimum fluidization velocity, this result being possible as the gas transfers its momentum to a limited portion of solids constituting the whole bed; wide range of operating conditions starting from a value slightly exceeding the

The term "Spouted Bed" was coined with an early work carried out at the National Research Council of Canada by (Gishler & Mathur, 1954). A comprehensive book by (Mathur & Epstein, 1974) provided a systematic summary of the scientific work done in the years. Very recently (Epstein & Grace, 2011) the most advanced knowledge in the field has been

Spouted beds were originally developed as an alternative method of drying moist seeds needing a prompt and effective processing. Sooner the interest in spouted beds grew and their application included coal gasification and combustion, pyrolysis of coal and oil shale, solid blending, nuclear particle coating, cooling and granulation as well as polymer crystallization and solid state polymerization processes. Fundamental studies were carried out to establish design correlations, the performance of spouted beds as chemical reactors, motion patterns

Some additional improvement in the gas-to-solids contacting can be provided by independently aerating the annulus, thus generating the so-called "Spout-fluid Beds" (Chatterjee, 1970). Again, a perforated draft-tube can be placed to surround the spout, thus

Most studies were carried out in plain cylindrical geometries, either full sectional, half sectional or even in a reduced angular section of a cylinder in order to explore scale-up possibilities. In any case adding a flat transparent wall has been demonstrated to interfere to a moderate extent with the solids trajectory vectors within the annulus (Rovero et al., 1985) while does not affect the measurement of the fundamental parameters of spouted beds (*Ums*, *Hm*, *Ds*). A limited number of examples consider multiple spouting, either in parallel or in

A typical spouted bed scheme is given in Figure 3. This representation depicts the gas inlet and outlet, the upward movement of solids in the spout, their trajectories in the fountain and the subsequent descent in the annulus to reach the spout again at depths which depends on the path-lines originated by the landing position on the bed surface. The particle holdup can be loaded batchwise or, alternatively, in a continuous mode. The latter option depends on process requirements, nevertheless a continuous solids renewal in no way alters the above features. A proper solids feeding should minimize bypass towards the discharge port; in this view a direct feeding over the bed surface appears the best option to guarantee at least one circulation loop in the annulus to all the particles. A direct distributed feed over

and segregation of solids, gas distribution within the complex hydrodynamics.

contributing in terms of stability and operational flexibility (Grbavčić et al., 1982).

series, squared and rectangular cross sections (Mathur & Epstein, 1974).

the fountain is required only in case of particles with high tendency to stick.

possibility of handling coarse particles having a wide size range and morphology.

minimum spouting velocity ;

**4. Spouted beds** 

**5. Design bases** 

updated.

liquid-like flow of solid particulate materials that allows continuous and easily controlled operations. The good mixing of solids provides a large thermal flywheel and secures isothermal conditions throughout the reactor. If some minor wall-effects are neglected, the solids-to-fluid relationship is independent of the vessel size, so this operation can be easily scaled up and large size operations are possible.

On the other hand, fluidization may reach conditions of instability, such as bubbling or slugging, which usually represent a situation of inefficient contact between the two phases. Moreover, in high temperature operations or with sticky particles, solids sintering or agglomeration easily occur. Finally, but not less important, fines generation, erosion of vessel, internals and pipes is a serious problem caused by the random and intense movement of particles (i.e. particle or carry over exerted by the fluid.

Mass and heat transfer related to physical and chemical reactions are kinetically limited by surface area of large particle, either when the operation occurs in the fluid phase or on the solids. In these cases fluidized systems must operate with fine particulate materials; examples are given by heterogeneous catalysis or combustion/gasification of coal fines.

For reasons intrinsic to many processes (agricultural products upgrading, agglomeration, pelletization, etc.), large particle handling is required and fluidization does not represent an optimal technology. Material comminution to reach the size required by conventional fluidization is an additional negative aspect which increases the exergetic overall process cost. A very noticeable gas rate is required to reach fluidization of large particles, which often far exceeds the amount required for the physical or chemical operation considered. It should also be noted that fluidized systems are operated at a gas rate double or triple with respect to the minimum fluidization velocity to confer the system adequate mixing and avoid any dead zone. In conclusion, fluidization appears an interesting operation thanks to an easy scale-up, though its extensive feature (large gas flow rate need) as well as its intensive characteristics (random fluid-to-particle hydraulic interaction) counterbalance its favourable aspects to some extent.

A spouted bed can be realized by replacing the perforated plate distributor typical of a standard fluidized bed with a simple orifice, either located in the central position of a flat bottom or at the apex of a bottom cone, whose profile helps the solids circulation and avoids stagnant zones. Examples of non-axial orifices appear in the scientific literature, too. The fluidizing gas enters the system at a high velocity, generates a cavity which protrudes upward through the "spout", which, having an almost cylindrical shape, can be characterized by its spout diameter value *Ds*. When the gas flow rate is large enough, the spout reaches the bed surface and forms a "fountain" of particles in the freeboard. The fountain can be more or less developed depending on the gas rate and the overall system features. After falling on the bed surface, the solids continue their downward travel in the "annulus" surrounding the spout and reach different depths before being recaptured into the spout. The dual hydrodynamics, good mixing in spout and fountain and piston flow in the peripheral annulus with alternated high and slow interphase transfer, in the spout and in the annulus respectively, makes spouted beds unique reactors.

Consequently, spouted beds offer very peculiar features, which can be summarized as:

very regular circulation of particles and absence of dead zones;


### **4. Spouted beds**

408 Advances in Chemical Engineering

liquid-like flow of solid particulate materials that allows continuous and easily controlled operations. The good mixing of solids provides a large thermal flywheel and secures isothermal conditions throughout the reactor. If some minor wall-effects are neglected, the solids-to-fluid relationship is independent of the vessel size, so this operation can be easily

On the other hand, fluidization may reach conditions of instability, such as bubbling or slugging, which usually represent a situation of inefficient contact between the two phases. Moreover, in high temperature operations or with sticky particles, solids sintering or agglomeration easily occur. Finally, but not less important, fines generation, erosion of vessel, internals and pipes is a serious problem caused by the random and intense

Mass and heat transfer related to physical and chemical reactions are kinetically limited by surface area of large particle, either when the operation occurs in the fluid phase or on the solids. In these cases fluidized systems must operate with fine particulate materials; examples are given by heterogeneous catalysis or combustion/gasification of coal fines.

For reasons intrinsic to many processes (agricultural products upgrading, agglomeration, pelletization, etc.), large particle handling is required and fluidization does not represent an optimal technology. Material comminution to reach the size required by conventional fluidization is an additional negative aspect which increases the exergetic overall process cost. A very noticeable gas rate is required to reach fluidization of large particles, which often far exceeds the amount required for the physical or chemical operation considered. It should also be noted that fluidized systems are operated at a gas rate double or triple with respect to the minimum fluidization velocity to confer the system adequate mixing and avoid any dead zone. In conclusion, fluidization appears an interesting operation thanks to an easy scale-up, though its extensive feature (large gas flow rate need) as well as its intensive characteristics (random fluid-to-particle hydraulic interaction) counterbalance its

A spouted bed can be realized by replacing the perforated plate distributor typical of a standard fluidized bed with a simple orifice, either located in the central position of a flat bottom or at the apex of a bottom cone, whose profile helps the solids circulation and avoids stagnant zones. Examples of non-axial orifices appear in the scientific literature, too. The fluidizing gas enters the system at a high velocity, generates a cavity which protrudes upward through the "spout", which, having an almost cylindrical shape, can be characterized by its spout diameter value *Ds*. When the gas flow rate is large enough, the spout reaches the bed surface and forms a "fountain" of particles in the freeboard. The fountain can be more or less developed depending on the gas rate and the overall system features. After falling on the bed surface, the solids continue their downward travel in the "annulus" surrounding the spout and reach different depths before being recaptured into the spout. The dual hydrodynamics, good mixing in spout and fountain and piston flow in the peripheral annulus with alternated high and slow interphase transfer, in the spout and

Consequently, spouted beds offer very peculiar features, which can be summarized as:

scaled up and large size operations are possible.

favourable aspects to some extent.

movement of particles (i.e. particle or carry over exerted by the fluid.

in the annulus respectively, makes spouted beds unique reactors.

very regular circulation of particles and absence of dead zones;

The term "Spouted Bed" was coined with an early work carried out at the National Research Council of Canada by (Gishler & Mathur, 1954). A comprehensive book by (Mathur & Epstein, 1974) provided a systematic summary of the scientific work done in the years. Very recently (Epstein & Grace, 2011) the most advanced knowledge in the field has been updated.

Spouted beds were originally developed as an alternative method of drying moist seeds needing a prompt and effective processing. Sooner the interest in spouted beds grew and their application included coal gasification and combustion, pyrolysis of coal and oil shale, solid blending, nuclear particle coating, cooling and granulation as well as polymer crystallization and solid state polymerization processes. Fundamental studies were carried out to establish design correlations, the performance of spouted beds as chemical reactors, motion patterns and segregation of solids, gas distribution within the complex hydrodynamics.

Some additional improvement in the gas-to-solids contacting can be provided by independently aerating the annulus, thus generating the so-called "Spout-fluid Beds" (Chatterjee, 1970). Again, a perforated draft-tube can be placed to surround the spout, thus contributing in terms of stability and operational flexibility (Grbavčić et al., 1982).

Most studies were carried out in plain cylindrical geometries, either full sectional, half sectional or even in a reduced angular section of a cylinder in order to explore scale-up possibilities. In any case adding a flat transparent wall has been demonstrated to interfere to a moderate extent with the solids trajectory vectors within the annulus (Rovero et al., 1985) while does not affect the measurement of the fundamental parameters of spouted beds (*Ums*, *Hm*, *Ds*). A limited number of examples consider multiple spouting, either in parallel or in series, squared and rectangular cross sections (Mathur & Epstein, 1974).

### **5. Design bases**

A typical spouted bed scheme is given in Figure 3. This representation depicts the gas inlet and outlet, the upward movement of solids in the spout, their trajectories in the fountain and the subsequent descent in the annulus to reach the spout again at depths which depends on the path-lines originated by the landing position on the bed surface. The particle holdup can be loaded batchwise or, alternatively, in a continuous mode. The latter option depends on process requirements, nevertheless a continuous solids renewal in no way alters the above features. A proper solids feeding should minimize bypass towards the discharge port; in this view a direct feeding over the bed surface appears the best option to guarantee at least one circulation loop in the annulus to all the particles. A direct distributed feed over the fountain is required only in case of particles with high tendency to stick.

Optimization of Spouted Bed Scale-Up by Square-Based Multiple Unit Design 411

the vessel ("cylindrical spouted beds"). In both cases the conical included angle is in the range of 60 to 90°; by further diminishing the angle instability in solid circulation might occur, while increasing excessively the angle decreases solids circulation at the base. A gross criterion that distinguishes conical and cylindrical spouted beds can be given considering the type of reaction to carry out: when solid phase undergoes a fast surface transformation, the optimum residence time of the gaseous phase is very short. This condition is satisfied by shallow beds as in the case of catalytic polymerization or coal gasification and pyrolysis. When the reaction is controlled by heat or mass transfer, the gas-solid contact must be adjusted with a deeper bed

Thanks to this flexibility, the mean residence time of the solids in continuous operations can be regulated by optimizing the solids hold-up in a single vessel, which gives origin to a well mixed unit, or otherwise it is possible to conceive a cascade of several units to have a system approaching a plug flow. In the latter case square based units can have a number of advantages over a conventional cylindrical geometry. Specifically, the construction is cheaper, more compact and the heat dissipation toward the outside lower. Some scientific aspects remain open though: the design correlations that should be validated, the stability of multiple units proved both during the start-up and at steady state conditions and the possibility of fully predicting the solids residence time distribution as a function of

Stable spouting can be obtained by satisfying two hydrodynamic requirements: 1) the bed depth must be lower than the *Hm* value and 2) the gas flow rate has to exceed *Ums*. From an initial condition of a static bed with a nil gas flow, by increasing the gas flow a certain pressure drop is built up through the bed of particles. The graph given in Figure 4 describes this hydrodynamic evolution, which implies a pressure drop/flow rate hysteresis between an increasing flow and the reverse situation. The hysteresis is caused by different packing conditions of the bed particles, that expand to attain a loose state once a spouting condition is reached. Starting from the static bed condition denoted by A, the pressure drop increases with

of the gas velocity, the bed displays a moderate progressive expansion and a corresponding decrease of pressure drop to reach C. Finally an abrupt spouting leads to a sudden decrease of pressure drop which stabilizes at an nearly constant value (D), which is maintained in all the operating range of gas rate. This situation represents a stable spouting. In case of fluid velocity decrease, the pressure drop remains constant down to the spout collapse (E), which compacts the system to some extent and the pressure drop increases again to F, giving origin to the afore

The whole system hydrodynamics is given by knowing *ΔPM, Ums, ΔPs* and the *U/Ums* ratio chosen for a stable spouting. A recent paper has compared data obtained in the mentioned 0.35 m side square-base unit to the existing literature correlations (Beltramo et al., 2009). The design data for the blower are *ΔPs* and *U*, while the maximum pressure drop and the relative transitory flow rate can be easily generated by a side capacitive device, to be used at the start-up only. It is important knowing that the hydrodynamic transient can be as short as a few seconds, so that the capacitive device can be designed with a characteristic time shorter than a fraction of a

minute. In this view, the timing of a spouting process onset is quantified in Figure 15.

*PM* at B). With an additional increase

configuration as in drying, coating, solid phase polymerization, etc..

the fluid velocity and reaches a maximum pressure drop (

mentioned hysteresis. The minimum spouting velocity is recorded at E.

geometry and number of stages.

**6. Spouting regime** 

Fig. 3. Schematic of a spouted bed

The solids discharge from continuous operations is generally carried out with an overflow port, unless special process control is required. This case may be given by coating processes, where the total bed surface area should be controlled. In this case a submerged port preferentially discharges coarse material, due to local segregation mechanisms (Piccinini, 1980), while the entire spouted bed retains good mixing capacity, which can be regulated by the fountain action.

The gas flow distribution between spout and annulus is completely independent whether the solids are batch or continuously fed: part of the gas progressively percolates from the spout into the annulus by moving toward higher elevations in the bed of particles. In case of a bed of sufficient height the gas in the annulus may reach a superficial velocity close to the minimum fluidization velocity of the solids. In this event the annulus is prone to collapse into the spout, thus defining the "maximum spoutable bed depth, or *Hm*". This parameter represents one of the fundamental criteria to design a spouted bed unit (Mathur & Epstein, 1974); *Hm* depends on vessel geometry, fluid and particle properties.

The second fundamental parameter is given by the minimum rate of gas required to maintain the system spouting, the so-called "minimum spouting velocity, or *Ums*". This operating factor can be either determined by an experimental procedure (as described in Figure 10) or can be calculated by the existing correlations.

A spouted bed, thanks to its flexibility, can be operated with a wide range of solids load to fill part or all the cone ("conical spouted beds"), or otherwise to engage also the upper portion of the vessel ("cylindrical spouted beds"). In both cases the conical included angle is in the range of 60 to 90°; by further diminishing the angle instability in solid circulation might occur, while increasing excessively the angle decreases solids circulation at the base. A gross criterion that distinguishes conical and cylindrical spouted beds can be given considering the type of reaction to carry out: when solid phase undergoes a fast surface transformation, the optimum residence time of the gaseous phase is very short. This condition is satisfied by shallow beds as in the case of catalytic polymerization or coal gasification and pyrolysis. When the reaction is controlled by heat or mass transfer, the gas-solid contact must be adjusted with a deeper bed configuration as in drying, coating, solid phase polymerization, etc..

Thanks to this flexibility, the mean residence time of the solids in continuous operations can be regulated by optimizing the solids hold-up in a single vessel, which gives origin to a well mixed unit, or otherwise it is possible to conceive a cascade of several units to have a system approaching a plug flow. In the latter case square based units can have a number of advantages over a conventional cylindrical geometry. Specifically, the construction is cheaper, more compact and the heat dissipation toward the outside lower. Some scientific aspects remain open though: the design correlations that should be validated, the stability of multiple units proved both during the start-up and at steady state conditions and the possibility of fully predicting the solids residence time distribution as a function of geometry and number of stages.

#### **6. Spouting regime**

410 Advances in Chemical Engineering

The solids discharge from continuous operations is generally carried out with an overflow port, unless special process control is required. This case may be given by coating processes, where the total bed surface area should be controlled. In this case a submerged port preferentially discharges coarse material, due to local segregation mechanisms (Piccinini, 1980), while the entire spouted bed retains good mixing capacity, which can be regulated by

The gas flow distribution between spout and annulus is completely independent whether the solids are batch or continuously fed: part of the gas progressively percolates from the spout into the annulus by moving toward higher elevations in the bed of particles. In case of a bed of sufficient height the gas in the annulus may reach a superficial velocity close to the minimum fluidization velocity of the solids. In this event the annulus is prone to collapse into the spout, thus defining the "maximum spoutable bed depth, or *Hm*". This parameter represents one of the fundamental criteria to design a spouted bed unit (Mathur & Epstein,

The second fundamental parameter is given by the minimum rate of gas required to maintain the system spouting, the so-called "minimum spouting velocity, or *Ums*". This operating factor can be either determined by an experimental procedure (as described in

A spouted bed, thanks to its flexibility, can be operated with a wide range of solids load to fill part or all the cone ("conical spouted beds"), or otherwise to engage also the upper portion of

1974); *Hm* depends on vessel geometry, fluid and particle properties.

Figure 10) or can be calculated by the existing correlations.

Fig. 3. Schematic of a spouted bed

the fountain action.

Stable spouting can be obtained by satisfying two hydrodynamic requirements: 1) the bed depth must be lower than the *Hm* value and 2) the gas flow rate has to exceed *Ums*. From an initial condition of a static bed with a nil gas flow, by increasing the gas flow a certain pressure drop is built up through the bed of particles. The graph given in Figure 4 describes this hydrodynamic evolution, which implies a pressure drop/flow rate hysteresis between an increasing flow and the reverse situation. The hysteresis is caused by different packing conditions of the bed particles, that expand to attain a loose state once a spouting condition is reached. Starting from the static bed condition denoted by A, the pressure drop increases with the fluid velocity and reaches a maximum pressure drop (*PM* at B). With an additional increase of the gas velocity, the bed displays a moderate progressive expansion and a corresponding decrease of pressure drop to reach C. Finally an abrupt spouting leads to a sudden decrease of pressure drop which stabilizes at an nearly constant value (D), which is maintained in all the operating range of gas rate. This situation represents a stable spouting. In case of fluid velocity decrease, the pressure drop remains constant down to the spout collapse (E), which compacts the system to some extent and the pressure drop increases again to F, giving origin to the afore mentioned hysteresis. The minimum spouting velocity is recorded at E.

The whole system hydrodynamics is given by knowing *ΔPM, Ums, ΔPs* and the *U/Ums* ratio chosen for a stable spouting. A recent paper has compared data obtained in the mentioned 0.35 m side square-base unit to the existing literature correlations (Beltramo et al., 2009). The design data for the blower are *ΔPs* and *U*, while the maximum pressure drop and the relative transitory flow rate can be easily generated by a side capacitive device, to be used at the start-up only. It is important knowing that the hydrodynamic transient can be as short as a few seconds, so that the capacitive device can be designed with a characteristic time shorter than a fraction of a minute. In this view, the timing of a spouting process onset is quantified in Figure 15.

Optimization of Spouted Bed Scale-Up by Square-Based Multiple Unit Design 413

on spouted beds by (Epstein & Grace, 2011) together with general criteria, though the issue

Both routes must be discussed in terms of advantages and drawbacks. The first approach implies a simple geometry and mechanic construction: some doubts arise on the validity of the existing correlations and the overall hydrodynamics in the unit (gas distribution between spout and annulus, solids circulation, etc.). The use of the existing correlations up to a unit diameter (*Dc*) of about 0.6 m is generally thought fully safe. If a continuous operation is considered, this arrangement gives the solid particulate material a well-mixed

Conversely, if a sequence of multiple beds is realized, achieving a fully independence of the units becomes the fundamental goal. In other words a non-interfering system must be designed, so that it is up to the operator decide which unit to start-up or shut-down first according to process needs. Due to the complexity of a multiple system each unit must mandatorily replicate the foreseen behaviour of the basic component. In this case, the residence time distribution of the solids approaches closely a plug flow to meet most process

Design geometry and regulation criteria are thoroughly discussed in the continuation of this chapter. The design and the construction of a multiple unit implies a careful geometrical optimization to minimize heat loss, investment and operating costs, assure a straightforward start-up, guarantee stability and process performance. According to this key requirement, squared-based units could replace a standard cylindrical section geometry, according to an account presented in literature (Beltramo et al., 2009). A correlation between

This chapter describes a systematic experimentation in a 0.13, 0.20 and 0.35 m side units to correlate the hydrodynamics of square-based spouted beds to the one of a corresponding cylindrical units and define the optimal geometrical configuration to assure solids circulation and transfer to downstream modules when multiple spouted beds in series are considered. The square-based experimental modules were made of wood with a frontal Perspex wall or of AISI 316 SS with a tempered glass window, depending on whether the apparatus had to be operated at room or higher temperatures. A 0.15 m ID cylindrical unit was also used for data comparison. Table 1 provides more geometrical details of the spouted beds. All bases, either frustum shaped or conical were characterized by an included angle of 60°. The spout orifice extended up of 1 mm over the base to improve solids circulation at the bed bottom, according to suggestion existing in the literature. The vessels were 1.5 m or 2 m high to allow the measurement of the maximum spoutable bed depth for all the solids

The tests were carried out with several materials to cover a sufficient range of parameters, as

Scaling-up a spouted and spout-fluid beds can be tackled according two approaches:

behaviour with a broad distribution of particle residence time at the exit of the unit.

remains open.

requirements.

tested.

they appear in Table 2.

1. increasing the size of a single unit, or 2. repeating side by side several units.

cylindrical and square-based units is also needed.

**8. Experimentation on single and multiple square-based units** 

Fig. 4. Hydrodynamic diagram for spouting onset

Figure 5 displays a sequence of pictures that show the spout onset from the initial cavity generation (A) to the full spouting. The third picture qualitatively corresponds to the point B in Figure 4, the fourth picture shows the rapid sequence between C and D, while the last picture may describe any point in the interval E-D, or over.

Fig. 5. Photographic sequence of the evolution of a spouting process performed in a squared-based half sectional 0.2 m side unit

#### **7. Scale-up**

Due to the peculiarities of spouted beds, their application to industrial processes requires a sound experience and a clear vision of their hydrodynamics since scale-up from laboratory experience is required. A summary of industrial implementations appears in the recent book on spouted beds by (Epstein & Grace, 2011) together with general criteria, though the issue remains open.

Scaling-up a spouted and spout-fluid beds can be tackled according two approaches:


412 Advances in Chemical Engineering

Figure 5 displays a sequence of pictures that show the spout onset from the initial cavity generation (A) to the full spouting. The third picture qualitatively corresponds to the point B in Figure 4, the fourth picture shows the rapid sequence between C and D, while the last

Fig. 5. Photographic sequence of the evolution of a spouting process performed in a

Due to the peculiarities of spouted beds, their application to industrial processes requires a sound experience and a clear vision of their hydrodynamics since scale-up from laboratory experience is required. A summary of industrial implementations appears in the recent book

Fig. 4. Hydrodynamic diagram for spouting onset

squared-based half sectional 0.2 m side unit

**7. Scale-up** 

picture may describe any point in the interval E-D, or over.

Both routes must be discussed in terms of advantages and drawbacks. The first approach implies a simple geometry and mechanic construction: some doubts arise on the validity of the existing correlations and the overall hydrodynamics in the unit (gas distribution between spout and annulus, solids circulation, etc.). The use of the existing correlations up to a unit diameter (*Dc*) of about 0.6 m is generally thought fully safe. If a continuous operation is considered, this arrangement gives the solid particulate material a well-mixed behaviour with a broad distribution of particle residence time at the exit of the unit.

Conversely, if a sequence of multiple beds is realized, achieving a fully independence of the units becomes the fundamental goal. In other words a non-interfering system must be designed, so that it is up to the operator decide which unit to start-up or shut-down first according to process needs. Due to the complexity of a multiple system each unit must mandatorily replicate the foreseen behaviour of the basic component. In this case, the residence time distribution of the solids approaches closely a plug flow to meet most process requirements.

Design geometry and regulation criteria are thoroughly discussed in the continuation of this chapter. The design and the construction of a multiple unit implies a careful geometrical optimization to minimize heat loss, investment and operating costs, assure a straightforward start-up, guarantee stability and process performance. According to this key requirement, squared-based units could replace a standard cylindrical section geometry, according to an account presented in literature (Beltramo et al., 2009). A correlation between cylindrical and square-based units is also needed.

#### **8. Experimentation on single and multiple square-based units**

This chapter describes a systematic experimentation in a 0.13, 0.20 and 0.35 m side units to correlate the hydrodynamics of square-based spouted beds to the one of a corresponding cylindrical units and define the optimal geometrical configuration to assure solids circulation and transfer to downstream modules when multiple spouted beds in series are considered. The square-based experimental modules were made of wood with a frontal Perspex wall or of AISI 316 SS with a tempered glass window, depending on whether the apparatus had to be operated at room or higher temperatures. A 0.15 m ID cylindrical unit was also used for data comparison. Table 1 provides more geometrical details of the spouted beds. All bases, either frustum shaped or conical were characterized by an included angle of 60°. The spout orifice extended up of 1 mm over the base to improve solids circulation at the bed bottom, according to suggestion existing in the literature. The vessels were 1.5 m or 2 m high to allow the measurement of the maximum spoutable bed depth for all the solids tested.

The tests were carried out with several materials to cover a sufficient range of parameters, as they appear in Table 2.

Optimization of Spouted Bed Scale-Up by Square-Based Multiple Unit Design 415

Fig. 7A. See from left to right: a) 0.15 m ID cylindrical Perspex spouted bed vessel, b) 0.13 m side square-based wooden unit; equivalent to a), c) 0.20 m side square-based wooden unit

Fig. 7B. 0.35 m side square-based AISI 316 SS unit with tempered glass frontal window.

and d) half-sectional 0.20x0.10 m2 square-based wooden unit.


Table 1. Geometrical characteristics of spouted test apparatuses


Table 2. Physical properties of the particulate material used

Fig. 6. Materials used in the tests

The experimental equipment was composed of several units, whose pictures are given in Figure 7A and 7B. Depending on the flow rate required, air as spouting medium was either provided by two volumetric compressors (total flow rate of about 250 Nm3/hr) or from a blower (flow rate of 350 Nm3/hr). The air flow was cooled through a corrugated pipe heat exchanger to guarantee spouting at a constant room temperature. Two rotameters were used to meter the flow together with gauge pressure recording.

The experimental strategy was directed both to run batch experiments in a single vessel to assess the fundamental spouting parameters (*Hm, Ums*, a stable *U/Ums* ratio), as well as

L-13 squared 0.13 - 0.15 1.50 60 D-15 cylindrical - 0.15 0.15 1.50 60 L-20 squared 0.20 - 0.23 2.00 60 L-35 squared 0.35 - 0.40 2.50 60 half L-20 rectangular 0.20 - - 2.00 60

PET chips 3.04 1336 0.87 35 Turnip seeds 1.50 1081 1.00 27 Corn 7.82 1186 0.80 27 Soya beans 7.23 1144 0.99 29 Sunflower seeds 6.16 696 0.87 37

The experimental equipment was composed of several units, whose pictures are given in Figure 7A and 7B. Depending on the flow rate required, air as spouting medium was either provided by two volumetric compressors (total flow rate of about 250 Nm3/hr) or from a blower (flow rate of 350 Nm3/hr). The air flow was cooled through a corrugated pipe heat exchanger to guarantee spouting at a constant room temperature. Two rotameters were used

The experimental strategy was directed both to run batch experiments in a single vessel to assess the fundamental spouting parameters (*Hm, Ums*, a stable *U/Ums* ratio), as well as

Equivalent diameter (m)

Density

Height of the unit (m)

(kg/m3) Sphericity Repose angle

Cone / Pyramid angle (deg)

(deg)

(m)

Table 1. Geometrical characteristics of spouted test apparatuses

Table 2. Physical properties of the particulate material used

to meter the flow together with gauge pressure recording.

Fig. 6. Materials used in the tests

diameter (mm)

Material Equivalent mean

Module

type: Section Side (m) Diameter

Fig. 7A. See from left to right: a) 0.15 m ID cylindrical Perspex spouted bed vessel, b) 0.13 m side square-based wooden unit; equivalent to a), c) 0.20 m side square-based wooden unit and d) half-sectional 0.20x0.10 m2 square-based wooden unit.

Fig. 7B. 0.35 m side square-based AISI 316 SS unit with tempered glass frontal window.

Optimization of Spouted Bed Scale-Up by Square-Based Multiple Unit Design 417

parameters characteristic (*Hm, Ums, Ds* , mean particle velocity in the annulus, volume of spout and fountain). Figure 8B shows the scheme of the three cell equipment used to optimize geometry of internals, overall structure, effect of continuous solids feeding and

The hydrodynamic behaviour of a square-based spouted beds was evaluated by exploring a wide range of conditions, from a bed depth corresponding to the frustum height to the maximum spoutable bed depth by operating with all the material shown in Figure 6. A reasonable ample scale-up factor (in excess of 7) was considered by running four squarebased units of 0.13, 0.20 and 0.35 m side. Moreover the 0.13 m side unit was compared to the equivalent 0.15 m ID cylindrical unit to identify any difference in terms of *Ums* and *Hm* and then validate the applicability of the existing correlations to the non-standard square-based

The maximum spoutable bed depth is measured by progressively adding solid granulated material in the vessel and verifying that a stable spout could be formed. Some difficult transition may be encountered at *H* ≈ *Hm* to make the spouting process neatly evolve from an internal spout to an external well-formed one; in these cases some subjective uncertainty may be left in the measurements. In this case visualization in half-column can be useful (see

Table 3 presents the experimental results obtained in two equivalent units, namely the cylindrical 0.15 m ID column (D-15) and the 0.13 m side square-based vessel (L-13); additionally, the larger L-20 vessel provided data useful for scale-up and validation of existing correlations. The test results show good and consistent agreement with the predictions given by literature equations (Malek & Lu, 1965; McNab & Bridgwater, 1977).

Material *Hm*, m *Hm*, m *Hm*, m PET chips 0.69 0.65 1.26 Turnip seeds 0.84 0.85 1.44 Corn 0.41 0.46 0.98 Soya beans 0.37 0.36 0.65 Sunflower seeds 0.68 0.69 1.48 Table 3. Experimental values of the maximum spoutable bed depth *Hm* in the 0.13 m side square-based unit (L-13), in the 0.20 m side square-based unit (L-20) and in the 0.15 m ID

Another interesting result derived from the comparison of the D-15 and L-13 units; the maximum spoutable bed depth values are very close for the same particles and identical orifice, so it allows us to assume that this fluid dynamic parameter does not depend on the cross section geometry. For this reason several correlations predicting *Hm* can be

D-15 L-13 L-20

residence time distribution of the solid phase.

**8.1.1 Maximum spoutable bed depth** 

the photographs of Fig.14).

cylindrical spouted bed (D-15).

**8.1 Batch unit tests** 

geometry.

optimize the geometrical spouted bed features (details of the base, orifice diameter, particle traps at the gas exit, etc.). Additionally, cylindrical and square-based vessels were comparatively tested.

Continuously operating experiments were directed to design the geometry of internals required to guarantee easy start-up, spouting stability in a multiple stage unit, discharge facilities to minimize "off-spec" products, adequate solids transfer from the feeding port, through the inter-stage weir, to the final overflow discharge.

Figure 8A provides the schematic view of the spouted bed assemblage of D-15, L-13, L-20, half-L-20 and L-35 units for batchwise measurements of the fundamental operating

Fig. 8A. Schematic of the spouted bed assemblage for batch measurements of operating parameters (see legend in the below Fig. 8B)

Fig. 8B. Schematic of the three module spouted bed for continuous hydrodynamic measurements

parameters characteristic (*Hm, Ums, Ds* , mean particle velocity in the annulus, volume of spout and fountain). Figure 8B shows the scheme of the three cell equipment used to optimize geometry of internals, overall structure, effect of continuous solids feeding and residence time distribution of the solid phase.

#### **8.1 Batch unit tests**

416 Advances in Chemical Engineering

optimize the geometrical spouted bed features (details of the base, orifice diameter, particle traps at the gas exit, etc.). Additionally, cylindrical and square-based vessels were

Continuously operating experiments were directed to design the geometry of internals required to guarantee easy start-up, spouting stability in a multiple stage unit, discharge facilities to minimize "off-spec" products, adequate solids transfer from the feeding port,

Figure 8A provides the schematic view of the spouted bed assemblage of D-15, L-13, L-20, half-L-20 and L-35 units for batchwise measurements of the fundamental operating

Fig. 8A. Schematic of the spouted bed assemblage for batch measurements of operating

Fig. 8B. Schematic of the three module spouted bed for continuous hydrodynamic

through the inter-stage weir, to the final overflow discharge.

parameters (see legend in the below Fig. 8B)

measurements

comparatively tested.

The hydrodynamic behaviour of a square-based spouted beds was evaluated by exploring a wide range of conditions, from a bed depth corresponding to the frustum height to the maximum spoutable bed depth by operating with all the material shown in Figure 6. A reasonable ample scale-up factor (in excess of 7) was considered by running four squarebased units of 0.13, 0.20 and 0.35 m side. Moreover the 0.13 m side unit was compared to the equivalent 0.15 m ID cylindrical unit to identify any difference in terms of *Ums* and *Hm* and then validate the applicability of the existing correlations to the non-standard square-based geometry.

#### **8.1.1 Maximum spoutable bed depth**

The maximum spoutable bed depth is measured by progressively adding solid granulated material in the vessel and verifying that a stable spout could be formed. Some difficult transition may be encountered at *H* ≈ *Hm* to make the spouting process neatly evolve from an internal spout to an external well-formed one; in these cases some subjective uncertainty may be left in the measurements. In this case visualization in half-column can be useful (see the photographs of Fig.14).

Table 3 presents the experimental results obtained in two equivalent units, namely the cylindrical 0.15 m ID column (D-15) and the 0.13 m side square-based vessel (L-13); additionally, the larger L-20 vessel provided data useful for scale-up and validation of existing correlations. The test results show good and consistent agreement with the predictions given by literature equations (Malek & Lu, 1965; McNab & Bridgwater, 1977).


Table 3. Experimental values of the maximum spoutable bed depth *Hm* in the 0.13 m side square-based unit (L-13), in the 0.20 m side square-based unit (L-20) and in the 0.15 m ID cylindrical spouted bed (D-15).

Another interesting result derived from the comparison of the D-15 and L-13 units; the maximum spoutable bed depth values are very close for the same particles and identical orifice, so it allows us to assume that this fluid dynamic parameter does not depend on the cross section geometry. For this reason several correlations predicting *Hm* can be

Optimization of Spouted Bed Scale-Up by Square-Based Multiple Unit Design 419

to direct observations, which are ambiguous in some instances due to bed instability as *Hm* is approached. Moreover careful extrapolation of the *Ums* and *Uonset* curves to detect their intersection is a possible method to estimate *Hm* when the experimental conditions are not

A continuously operating unit was tested to devise a guideline to design, start-up, gain in

A picture of the rig is present in Figure 11A, while a 3D scheme of the same unit is shown

stability and proper hydrodynamics in multiple square-based spouted beds.

Fig. 10. Onset and minimum spouting velocities vs. bed depth

suitable for a complete direct measurement.

**8.2 Continuously operating multiple units** 

Fig. 11A. Picture of the three-module experimental rig.

below, see Figure 11B.

indifferently applied to both geometries. The herein below Malek-Lu (derived in SI units) and McNab-Bridgwater equations, among others, were compared with the experimental results, as it appears in Figure 9. The agreement can be defined fully satisfactory.

Malek-Lu equation:

$$\frac{H\_M}{D\_\odot} = 418 \cdot \left(\frac{D\_\odot}{d\_p}\right)^{0.75} \cdot \left(\frac{D\_\odot}{d\_i}\right)^{0.40} \cdot \left(\frac{\mathcal{k}^2}{\rho\_s^{1.2}}\right) \tag{1}$$

McNab-Bridgwater equation:

$$H\_M = \frac{D\_\odot^2}{d\_p} \cdot \left(\frac{D\_\odot}{d\_i}\right)^{2g} \cdot \frac{700}{Ar} \left(\sqrt{1 + 35.9 \cdot 10^{-6} \cdot Ar} - 1\right)^2\tag{2}$$

Fig. 9. Comparison of experimental data obtained in square-based and cylindrical units with literature correlations

#### **8.1.2 Minimum spouting velocity**

As given in the hydrodynamic diagram of Figure 4, the minimum spouting velocity *Ums* represents the superficial velocity in the vessel below which spouting does not occur. Aiming to work in a stable situation, good practice suggests to moderately exceed this value by defining an operating spouting regime given by *U/Ums* > about 1.05, which is the measure of a very modest excess of gas with respect to the minimum spouting condition. It is worthwhile noting that this value can be further reduced if the spouted bed approaches its maximum spoutable bed depth *Hm*, as the below Figure 10 diagram indicates. The onset spouting and minimum spouting velocities approach as *H* is closer to *Hm*. This occurrence also indicates that passing from a submerged to an external spouting is progressively easier as *Hm* is approached, since the bed of particles is already highly expanded by a very high gas rate (pictures on Figure 14 represent this case). Consistently, these operating conditions generate hydrodynamic diagrams with a much less pronounced hysteresis, as given in a recent paper (Beltramo et al., 2009). To predict the minimum spouting velocity in multiple cell systems an empirical correlation was proposed (Murthy and Singh, 1994). By observing carefully the type of plots appearing in Figure 10, it is possible to note that the maximum bed depth can be inferred from the overlapping of the two curves. This remarkable experimental statement (based on hydrodynamic data) may offer an interesting alternative

Fig. 10. Onset and minimum spouting velocities vs. bed depth

to direct observations, which are ambiguous in some instances due to bed instability as *Hm* is approached. Moreover careful extrapolation of the *Ums* and *Uonset* curves to detect their intersection is a possible method to estimate *Hm* when the experimental conditions are not suitable for a complete direct measurement.

#### **8.2 Continuously operating multiple units**

418 Advances in Chemical Engineering

indifferently applied to both geometries. The herein below Malek-Lu (derived in SI units) and McNab-Bridgwater equations, among others, were compared with the experimental

> 1.2 <sup>418</sup> *M CC C pi s*

 

Fig. 9. Comparison of experimental data obtained in square-based and cylindrical units with

As given in the hydrodynamic diagram of Figure 4, the minimum spouting velocity *Ums* represents the superficial velocity in the vessel below which spouting does not occur. Aiming to work in a stable situation, good practice suggests to moderately exceed this value by defining an operating spouting regime given by *U/Ums* > about 1.05, which is the measure of a very modest excess of gas with respect to the minimum spouting condition. It is worthwhile noting that this value can be further reduced if the spouted bed approaches its maximum spoutable bed depth *Hm*, as the below Figure 10 diagram indicates. The onset spouting and minimum spouting velocities approach as *H* is closer to *Hm*. This occurrence also indicates that passing from a submerged to an external spouting is progressively easier as *Hm* is approached, since the bed of particles is already highly expanded by a very high gas rate (pictures on Figure 14 represent this case). Consistently, these operating conditions generate hydrodynamic diagrams with a much less pronounced hysteresis, as given in a recent paper (Beltramo et al., 2009). To predict the minimum spouting velocity in multiple cell systems an empirical correlation was proposed (Murthy and Singh, 1994). By observing carefully the type of plots appearing in Figure 10, it is possible to note that the maximum bed depth can be inferred from the overlapping of the two curves. This remarkable experimental statement (based on hydrodynamic data) may offer an interesting alternative

*D D <sup>H</sup> Ar*

<sup>2</sup> <sup>2</sup> <sup>3</sup> <sup>2</sup> <sup>700</sup> <sup>6</sup> 1 35.9 10 1 *C C*

0.75 0.40 <sup>2</sup>

(1)

(2)

results, as it appears in Figure 9. The agreement can be defined fully satisfactory.

*H DD D dd*

*d d Ar*

*p i*

*M*

Malek-Lu equation:

McNab-Bridgwater equation:

literature correlations

**8.1.2 Minimum spouting velocity** 

A continuously operating unit was tested to devise a guideline to design, start-up, gain in stability and proper hydrodynamics in multiple square-based spouted beds.

A picture of the rig is present in Figure 11A, while a 3D scheme of the same unit is shown below, see Figure 11B.

Fig. 11A. Picture of the three-module experimental rig.

Optimization of Spouted Bed Scale-Up by Square-Based Multiple Unit Design 421

A spouted bed fountain can be defined underdeveloped, developed or overdeveloped, depending on whether its geometrical margins reach the side of the spouting vessel. Since the volumetric solids circulation through the spout/fountain system can be estimated to range to a few percents of the spouting gas flow rate, the fountain alone can pour very noticeable flows of material onto contiguous stages; then, its action should be limited by some mechanical device to restrain its hydrodynamic effect. Some devices, the so-called "Chinese hats" were presented in the sector literature (Mathur & Epstein, 1974) and tested in this unit. The experimental output was entirely disappointing, as these regulators failed in sufficiently defining the fountain shape, acted as a target for the particles propelled by the spout and interfered hydrodynamically with the gas flow in the freeboard, either when they were made of solid steel plates or wire mesh screen. To conclude, these devices are not

Each stage was segregated from its adjacent ones by side vertical baffles protruding down from the vessel top to a very short elevation over the bed free surface. The gap left had to assure solids flow only, depending on the continuous throughput rate; this gap between the submerged and the freeboard baffles has to be regulated to allow solids transfer by overflow, with restricted particle bounces from the fountain. Also bypass was minimized by this precaution. These flat and inexpensive devices were chosen to completely separate the freeboard into as many stage as the spouted bed design required. As a result, the action of each fountain (independently of its shape, thus gaining in spouting regime flexibility) was limited to its own stage. The use of these simple baffle repartition was observed to be fundamental for minimizing any interference between stages and enormously gain in

In principle, according to the fundamentals of fluidization, a multiple orifice spouted bed does not require a repartition between the annuluses. This consideration is also compatible with the particle vertical streamlines and the side-to-side homogeneous percolation of gas from a spout into the corresponding annulus. This assessment can be fully accepted when the system is operated batchwise and no net solids flow from one stage to the downstream one has to be steadily maintained. Practical reasons (easy start-up, spouting stability over time, independent gas flow rate regulation in each spouting module) have demonstrated that submerged baffles greatly help in defining the solids holdup in each stage. The separation of contiguous annular regions contributes in properly distributing the gas rate and giving origin to fully independent spouts. Conversely, if the holdup of solids is out of

In the recent past, in the frame of industrialization of a novel patented process for polyethylene terephthalate solid state polymerization (Cavaglià, 2003), a unit was conceived as a series of n-fluid beds (where n>5), operated either in turbulent fluidization or in

**8.2.2 Fountain height regulators** 

advisable as internals in multistage spouted beds.

**8.2.3 Freeboard baffles between stages** 

**8.2.4 Submerged baffles between stages** 

control, all the system stability may be affected.

**8.2.5 Overall layout of a multiple spouting unit** 

stability.

Fig. 11B. 3D diagram of the same continuously operating unit with a horizontal layout of stages

Designing a multiple spouted bed system does not represent a "black art", though may be more complex than other gas/solids contactors. In particular the following design aspects were focused in this project:


The following paragraphs illustrate the progressive tuning of a multiple stage spouted bed up to achieve safe know-how and run continuous operations.

#### **8.2.1 Solids inlet design**

Solid particles were stored in an elevated drum, from which they could flow by gravity, being metered by a rotary valve, a screw feeder or a simple calibrated orifice. The particles were distributed along one entire side of the spouted bed, addressed down by a vertical baffle protruding to a small distance from the bed free surface. This simple device avoided any solids bypass promoted by the fountain. Residence time distribution measurements quantified a bypass as high as about 20% of the total flow rate, in case of baffle absence. The elevation of the baffle over the bed may require some regulation, depending on solids feed rate.

#### **8.2.2 Fountain height regulators**

420 Advances in Chemical Engineering

Fig. 11B. 3D diagram of the same continuously operating unit with a horizontal layout of

Designing a multiple spouted bed system does not represent a "black art", though may be more complex than other gas/solids contactors. In particular the following design aspects

The following paragraphs illustrate the progressive tuning of a multiple stage spouted bed

Solid particles were stored in an elevated drum, from which they could flow by gravity, being metered by a rotary valve, a screw feeder or a simple calibrated orifice. The particles were distributed along one entire side of the spouted bed, addressed down by a vertical baffle protruding to a small distance from the bed free surface. This simple device avoided any solids bypass promoted by the fountain. Residence time distribution measurements quantified a bypass as high as about 20% of the total flow rate, in case of baffle absence. The elevation of the baffle over the bed may require some regulation, depending on solids feed

stages

rate.

were focused in this project:

**8.2.1 Solids inlet design** 

 diverting baffle at solids inlet, fountain height regulators, freeboard baffles between units, submerged baffles between units,

overall layout of a multiple spouting unit

up to achieve safe know-how and run continuous operations.

A spouted bed fountain can be defined underdeveloped, developed or overdeveloped, depending on whether its geometrical margins reach the side of the spouting vessel. Since the volumetric solids circulation through the spout/fountain system can be estimated to range to a few percents of the spouting gas flow rate, the fountain alone can pour very noticeable flows of material onto contiguous stages; then, its action should be limited by some mechanical device to restrain its hydrodynamic effect. Some devices, the so-called "Chinese hats" were presented in the sector literature (Mathur & Epstein, 1974) and tested in this unit. The experimental output was entirely disappointing, as these regulators failed in sufficiently defining the fountain shape, acted as a target for the particles propelled by the spout and interfered hydrodynamically with the gas flow in the freeboard, either when they were made of solid steel plates or wire mesh screen. To conclude, these devices are not advisable as internals in multistage spouted beds.

#### **8.2.3 Freeboard baffles between stages**

Each stage was segregated from its adjacent ones by side vertical baffles protruding down from the vessel top to a very short elevation over the bed free surface. The gap left had to assure solids flow only, depending on the continuous throughput rate; this gap between the submerged and the freeboard baffles has to be regulated to allow solids transfer by overflow, with restricted particle bounces from the fountain. Also bypass was minimized by this precaution. These flat and inexpensive devices were chosen to completely separate the freeboard into as many stage as the spouted bed design required. As a result, the action of each fountain (independently of its shape, thus gaining in spouting regime flexibility) was limited to its own stage. The use of these simple baffle repartition was observed to be fundamental for minimizing any interference between stages and enormously gain in stability.

#### **8.2.4 Submerged baffles between stages**

In principle, according to the fundamentals of fluidization, a multiple orifice spouted bed does not require a repartition between the annuluses. This consideration is also compatible with the particle vertical streamlines and the side-to-side homogeneous percolation of gas from a spout into the corresponding annulus. This assessment can be fully accepted when the system is operated batchwise and no net solids flow from one stage to the downstream one has to be steadily maintained. Practical reasons (easy start-up, spouting stability over time, independent gas flow rate regulation in each spouting module) have demonstrated that submerged baffles greatly help in defining the solids holdup in each stage. The separation of contiguous annular regions contributes in properly distributing the gas rate and giving origin to fully independent spouts. Conversely, if the holdup of solids is out of control, all the system stability may be affected.

#### **8.2.5 Overall layout of a multiple spouting unit**

In the recent past, in the frame of industrialization of a novel patented process for polyethylene terephthalate solid state polymerization (Cavaglià, 2003), a unit was conceived as a series of n-fluid beds (where n>5), operated either in turbulent fluidization or in

Optimization of Spouted Bed Scale-Up by Square-Based Multiple Unit Design 423

(snapshot B), τ = 4.5 min) solid rates and at a solids throughput far exceeding the nominal system capacity, as required by any foreseen process (snapshot C), τ = 2.5 min). The bed free surface slant increased to a maximum slope (about 15°, as a mean value between inlet and overflow sides) by increasing the flow rate. This angle is in the range of 1/2 to 1/3 of the solids repose angle, which can be measured following literature recommendation (Metcalf, 1965-66). By further enhancing the feed rate, the downcomer flooded unless further raised. The hydrodynamic slope that builds up at the bed surface caused the first stage to work with a solids depth quite higher with respect the last one, this difference increasing with the number of stages. It follows that the fluid dynamic control was much trickier and the overall

Fig. 13. Continuous operation in the three-module 0.20 m side spouted bed: – – – ideal solids free surface; ─── actual solids surface at various solids flow rates: a) 2 kg/min, b) 6 kg/min

This geometrical limitation was overcome by setting each bed at a minimum difference of

*ΔH = Dc* tan

A rule of thumb suggests to determine the progressive vertical distance between adjacent

As a conclusion, stable operations in a multiple square-based spouted beds require three types of flat internal baffles: one for properly addressing the solids feed to the bed surface, intermediate baffles in the freeboard to confine each fountain action, submerged baffles, each of them setting the solids overflow level from the upstream to the downstream stage. A non-interfering condition between stages was provided by generating a sloped cascade of

A rig corresponding to half of the 0.20x0.20 m2 spouted bed was built to represent a vertical section of the full unit. The axial sectioning included the orifice, the pyramid frustum and the constant cross section sector, thus originating a 0.20x0.10 m2 column. A

is the angle formed by the actual solids surface with the horizontal level.

which, compared to the output of Eq. (3), leaves a safe operating margin.

(3)

*ΔH = 0.5 Dc* (4)

spouting stability impaired.

level equal to:

spouted beds at:

independent spouting units.

**8.3 Half-sectional spouted bed tests** 

where 

and c) 10 kg/min, as given in the scheme of Fig. 11 B

spouting regime, where polyester beads with low intrinsic viscosity are heated-up and solid reacted in one equipment. A sextuple spouting demonstration unit was then built and operated as a prototype equipment for PET chips upgrading (Beltramo et al., 2009). The six modules were placed at identical elevation and positioned according to a 2x3 layout; the solids moved following a chicane path without being hindered by any internal repartition. The multiple spouted bed appeared advantageous in term of heat transfer efficiency (higher gas temperature at the inlet, thanks to a very short contact time between gas and heat sensitive solids) and generated good property polymers. However, the overall operation was troublesome because of difficult control on solids holdup and gas flow rate regulation. From that study sound design hypotheses were drawn to construct the experimental rig appearing in Figure 11A, whose main difference with respect to the previous industrial equipment consists in the possibility of positioning each stage at the desired elevation to facilitate the solids overflow to the downstream stage, as represented in the schematic of Figure 12. The final version of this experimental rig had the possibility of testing all the internals described above.

Fig. 12. 3D diagram of a continuously operating unit with a sloped layout of stages

The difference of level between units suitable for continuous and stable operations was evaluated both by running specific tests and by comparing these results against a simple correlation originated by estimating the angle of internal friction of the solids used in the experiments. The tests performed in the triple spouted bed unit aimed at devising the effect of increasing solids feed rates on bulk mass transfer from stage to stage, by measuring the angle of the bed surface with respect to the horizontal level, as well as the effectiveness of the internal baffle positioning. Figure 13 shows three different operating conditions at a mean (snapshot A), corresponding to a mean particle total residence time τ = 13.5 min), high

spouting regime, where polyester beads with low intrinsic viscosity are heated-up and solid reacted in one equipment. A sextuple spouting demonstration unit was then built and operated as a prototype equipment for PET chips upgrading (Beltramo et al., 2009). The six modules were placed at identical elevation and positioned according to a 2x3 layout; the solids moved following a chicane path without being hindered by any internal repartition. The multiple spouted bed appeared advantageous in term of heat transfer efficiency (higher gas temperature at the inlet, thanks to a very short contact time between gas and heat sensitive solids) and generated good property polymers. However, the overall operation was troublesome because of difficult control on solids holdup and gas flow rate regulation. From that study sound design hypotheses were drawn to construct the experimental rig appearing in Figure 11A, whose main difference with respect to the previous industrial equipment consists in the possibility of positioning each stage at the desired elevation to facilitate the solids overflow to the downstream stage, as represented in the schematic of Figure 12. The final version of this experimental rig had the possibility of testing all the

Fig. 12. 3D diagram of a continuously operating unit with a sloped layout of stages

The difference of level between units suitable for continuous and stable operations was evaluated both by running specific tests and by comparing these results against a simple correlation originated by estimating the angle of internal friction of the solids used in the experiments. The tests performed in the triple spouted bed unit aimed at devising the effect of increasing solids feed rates on bulk mass transfer from stage to stage, by measuring the angle of the bed surface with respect to the horizontal level, as well as the effectiveness of the internal baffle positioning. Figure 13 shows three different operating conditions at a mean (snapshot A), corresponding to a mean particle total residence time τ = 13.5 min), high

internals described above.

(snapshot B), τ = 4.5 min) solid rates and at a solids throughput far exceeding the nominal system capacity, as required by any foreseen process (snapshot C), τ = 2.5 min). The bed free surface slant increased to a maximum slope (about 15°, as a mean value between inlet and overflow sides) by increasing the flow rate. This angle is in the range of 1/2 to 1/3 of the solids repose angle, which can be measured following literature recommendation (Metcalf, 1965-66). By further enhancing the feed rate, the downcomer flooded unless further raised. The hydrodynamic slope that builds up at the bed surface caused the first stage to work with a solids depth quite higher with respect the last one, this difference increasing with the number of stages. It follows that the fluid dynamic control was much trickier and the overall spouting stability impaired.

Fig. 13. Continuous operation in the three-module 0.20 m side spouted bed: – – – ideal solids free surface; ─── actual solids surface at various solids flow rates: a) 2 kg/min, b) 6 kg/min and c) 10 kg/min, as given in the scheme of Fig. 11 B

This geometrical limitation was overcome by setting each bed at a minimum difference of level equal to:

$$
\Delta H = D\_c \tan \alpha \tag{3}
$$

where is the angle formed by the actual solids surface with the horizontal level.

A rule of thumb suggests to determine the progressive vertical distance between adjacent spouted beds at:

$$
\Delta H = 0.5 \, D\_c \tag{4}
$$

which, compared to the output of Eq. (3), leaves a safe operating margin.

As a conclusion, stable operations in a multiple square-based spouted beds require three types of flat internal baffles: one for properly addressing the solids feed to the bed surface, intermediate baffles in the freeboard to confine each fountain action, submerged baffles, each of them setting the solids overflow level from the upstream to the downstream stage. A non-interfering condition between stages was provided by generating a sloped cascade of independent spouting units.

#### **8.3 Half-sectional spouted bed tests**

A rig corresponding to half of the 0.20x0.20 m2 spouted bed was built to represent a vertical section of the full unit. The axial sectioning included the orifice, the pyramid frustum and the constant cross section sector, thus originating a 0.20x0.10 m2 column. A

Optimization of Spouted Bed Scale-Up by Square-Based Multiple Unit Design 425

Fig. 15. The snapshot sequence indicates the motion of a tracer layer: A at t = 0 s; B at t = 1 s;

annulus. The fifth image gives evidence of noticeable shear acting on the particles, though no internal mixing occurs as far as particles enter the spout; then, quite an amount of particles is thoroughly mixed in the spout and fountain as they begin recirculating. The last

Spout diameter and spout profile were measured at the flat transparent wall. The experimental mean values were compared against the predictions given in the literature (McNab, 1972), with a difference of about 10%, which indicates that squared-based spouted

One of the main goals of a spouted bed cascade is to control the mixing degree of the overall system and possibly generate a piston flow of the solid phase to guarantee identical residence time to all particles. These evaluations are carried out by means of stimulusresponse techniques after attaining steady state in a continuously operating unit. Since each individual spouted bed appears to have about 90% of its volume in a perfectly mixed state (Epstein & Grace, 2011), the recycle ratio (ratio of internal circulation in the spout referred to feed rate) leaves a small volumetric fraction of the annulus to operate as plug flow. Residence time distribution (RTD) studies in multiple spouted bed were presented in the literature (Saidutta & Murthy, 2000) in small rectangular columns having two or three spout cells. The absence of internals in this system brought to fountain wandering and excessive fountain heights that caused overall mixing higher than the one corresponding to the number of mixed units in series. A detailed RTD study on stable systems and the correlation of the experimental results with respect to the ones predicted by a model can give a relevant contribution in designing these units. Models has gained increasing importance by making

**9. Fluid dynamics of solids in multiple spouted beds and its modelling** 

use of direct measurements in the half-sectional unit, thus becoming fully predictive.

The RTD curves represent an effective way to interpret the fluid dynamics of the solid phase in a multiphase continuously operating reactor. These functions describe the elapse of time spent by individual solids fractions in the system and can be modelled by a relatively simple combination of ideal systems, each of them describing a basic element (mixed or plug flow

C at t = 8 s ; D at t = 16 s; E at t = 22 s and F at t = 28 s

image indicates an overall good mixing condition.

**9.1 Residence time distribution function** 

system, dead zone, bypass, recycling).

bed findings do match the ones obtained in cylindrical vessels.

flat Perspex wall allowed a direct internal vision of spout, annulus and fountain, as already given in Figure 5.

Specific tests were carried out to demonstrate the close correspondence between data (*Hm* and *Ums*) obtained in full and half-sectional columns. Approaching the maximum bed depth an underdeveloped and stable fountain was obtained with *U* = 1.01 *Ums* (Figure 14 a)). Other runs highlighted a relevant bed expansion surmounting a submerged spout when the bed slightly exceeded the maximum bed depth (Figure 14 b)). Identical conditions also revealed occasional instability identified by a submerged wandering spout and some upper slugging, see Figure 14 c).

Fig. 14. Half sectional column tests at H: a) external spout and fountain formation at *H* ≈ 0.95 *Hm*, b) submerged spouting at *H* ≈ 1.05 *Hm* c) spouting instability with spout wandering at *H* ≈ 1.05 *Hm*

Half-sectional units are suitable to measure particle cycle time, define solids streamlines, as well as visualize, at proper frame frequency, zones characterized by a high mixing degree. As far as the downward particle velocities are concerned, the considerations presented in the literature were taken into account, though obtained in semicylindrical vessels (Rovero et al., 1985). Figure 15 shows a sequence of snapshots which make visible the progressive motion of a tracer layer deposited on a fixed bed before starting the spouting process (first image). The second snapshot indicates that the tracer particles have maintained their position ahead of an external spout be formed. The third and fourth images indicate that particles move in the fountain in a piston flow fashion: then, local trajectories, their envelope (i.e. streamlines) and individual particle velocities can be defined. A minor portion of tracer only has been captured by the spout in the travel along the constant cross section of the

flat Perspex wall allowed a direct internal vision of spout, annulus and fountain, as

Specific tests were carried out to demonstrate the close correspondence between data (*Hm* and *Ums*) obtained in full and half-sectional columns. Approaching the maximum bed depth an underdeveloped and stable fountain was obtained with *U* = 1.01 *Ums* (Figure 14 a)). Other runs highlighted a relevant bed expansion surmounting a submerged spout when the bed slightly exceeded the maximum bed depth (Figure 14 b)). Identical conditions also revealed occasional instability identified by a submerged wandering spout and some upper slugging,

Fig. 14. Half sectional column tests at H: a) external spout and fountain formation at *H* ≈ 0.95 *Hm*, b) submerged spouting at *H* ≈ 1.05 *Hm* c) spouting instability with spout

Half-sectional units are suitable to measure particle cycle time, define solids streamlines, as well as visualize, at proper frame frequency, zones characterized by a high mixing degree. As far as the downward particle velocities are concerned, the considerations presented in the literature were taken into account, though obtained in semicylindrical vessels (Rovero et al., 1985). Figure 15 shows a sequence of snapshots which make visible the progressive motion of a tracer layer deposited on a fixed bed before starting the spouting process (first image). The second snapshot indicates that the tracer particles have maintained their position ahead of an external spout be formed. The third and fourth images indicate that particles move in the fountain in a piston flow fashion: then, local trajectories, their envelope (i.e. streamlines) and individual particle velocities can be defined. A minor portion of tracer only has been captured by the spout in the travel along the constant cross section of the

already given in Figure 5.

wandering at *H* ≈ 1.05 *Hm*

see Figure 14 c).

Fig. 15. The snapshot sequence indicates the motion of a tracer layer: A at t = 0 s; B at t = 1 s; C at t = 8 s ; D at t = 16 s; E at t = 22 s and F at t = 28 s

annulus. The fifth image gives evidence of noticeable shear acting on the particles, though no internal mixing occurs as far as particles enter the spout; then, quite an amount of particles is thoroughly mixed in the spout and fountain as they begin recirculating. The last image indicates an overall good mixing condition.

Spout diameter and spout profile were measured at the flat transparent wall. The experimental mean values were compared against the predictions given in the literature (McNab, 1972), with a difference of about 10%, which indicates that squared-based spouted bed findings do match the ones obtained in cylindrical vessels.

#### **9. Fluid dynamics of solids in multiple spouted beds and its modelling**

One of the main goals of a spouted bed cascade is to control the mixing degree of the overall system and possibly generate a piston flow of the solid phase to guarantee identical residence time to all particles. These evaluations are carried out by means of stimulusresponse techniques after attaining steady state in a continuously operating unit. Since each individual spouted bed appears to have about 90% of its volume in a perfectly mixed state (Epstein & Grace, 2011), the recycle ratio (ratio of internal circulation in the spout referred to feed rate) leaves a small volumetric fraction of the annulus to operate as plug flow. Residence time distribution (RTD) studies in multiple spouted bed were presented in the literature (Saidutta & Murthy, 2000) in small rectangular columns having two or three spout cells. The absence of internals in this system brought to fountain wandering and excessive fountain heights that caused overall mixing higher than the one corresponding to the number of mixed units in series. A detailed RTD study on stable systems and the correlation of the experimental results with respect to the ones predicted by a model can give a relevant contribution in designing these units. Models has gained increasing importance by making use of direct measurements in the half-sectional unit, thus becoming fully predictive.

#### **9.1 Residence time distribution function**

The RTD curves represent an effective way to interpret the fluid dynamics of the solid phase in a multiphase continuously operating reactor. These functions describe the elapse of time spent by individual solids fractions in the system and can be modelled by a relatively simple combination of ideal systems, each of them describing a basic element (mixed or plug flow system, dead zone, bypass, recycling).

Optimization of Spouted Bed Scale-Up by Square-Based Multiple Unit Design 427

Fig. 16. Comparison between the RTD curves in the 3-module spouted bed with *H/Dc*=1.72

The theoretical description of continuous units combines basic elemental models, whose combination gives origin to a system capable of generating an overall response to properly match the actual behaviour of the real system studied. A descriptive model can produce an output without having a strict link with the actual hydrodynamic behaviour and then has to make use of fitting parameters. This approach does not allow sound predictions or extension to more complex reactor structures. A much powerful tool is produced by conceiving a phenomenological model based on experimental observations. These models become fully predictive since all parameters are based on actual measurements. Then, a model interacts with scale-up procedures through the validity of the correlations used rather

Dynamic responses were obtained by making use of a Matlab Simulink tool by generating model schemes as given in the below Figures 17 and 18. The fundamental modelling was based on a one-stage spouted bed; a multiple-cell system was then given by a cascade of

The initial modelling started by considering an early dynamic description of spouted beds, where the overall behaviour can be portrayed by a well-mixed system with a minor (8 to 10%) portion of plug flow. The corresponding scheme given in Figure 17 A) includes the feed rate F1, the bypass to fountain F3 (which become negligible when the inlet diverting baffle is considered, as a consequence F2 ≡ F1 and F5 ≡ F4 ), the total circulation from spout F4, the net discharge rate F7 ≡ F1 (for continuity) and F6 = F4 - F1. From the experimental conditions F1 is known and F4 can be estimated by measuring mean particle velocity at the frontal wall of half-column. As far as the other parameters than appear in Figure 17 B) are concerned, *td* is estimated from particle circulation, *τannulus* follows from holdup in the annulus and *τspout+fountain* is calculated by difference from the known bed holdup. At usual *H/Dc* ratios adopted in cylindrical columns, the ratio of these two time constants approaches

with and without submerged baffles

**9.2 Modelling** 

than its structure.

**9.2.1 Descriptive model** 

one magnitude order.

basic units.

Two types of curves can be studied. The *E(t)* function describes what a system releases instant after instant, i.e. the volumetric (or mass) fraction of particles whose residence time is between t and t+dt. The *F(t)* function provides the integral of *E(t)*·dt and represents the fraction of elements whose residence time is lower than t.

The most direct way to trace an *E(t)* curve makes use of a physical tracer, whose characteristics are identical or very close to the ones characterizing the bulk of solids travelling the system. Usually, two types of stimulus are adopted, a pulse (given by a definite amount of tracer introduced into the system in the shortest time) or a step (an abrupt change from the normal feedstock to an identical feed made of tracer only). The first one is generally the prompter to use. Right away after the introduction of the pulse, samples are taken at the system exit with a proper scrutiny degree and the tracer concentration is measured and recorded.

The two *E(t)* and *F(t)* functions are defined below:

$$E(t) = \frac{C(t) \cdot \stackrel{\bullet}{M}}{M\_{tr} \cdot \rho\_b} \tag{5}$$

$$F(t) = \int\_0^t E(t)dt\tag{6}$$

with *C (t)* being the concentration of solid tracer in the discharge, *Mtr* the mass of tracer injected, *ρb* the bulk density of tracer, *M* the mass flow rate of solids travelling the system.

A pulse function (called Dirac function, δ(t)) is given by an instantaneous but finite entity (equal to unity) entering a system at t =0. The Laplace transforms of these functions allow the use of simple algebraic input/output relationships.

In our experiments, the pulse was obtained by quickly introducing in the feed a small amount of tracer (150 to 300 g, equivalent to about 1.5% of the total solids hold-up) made of PET chips doped with some ferromagnetic powder. After sampling the ferromagnetic PET chips were sorted out from the PET bulk by a magnet and their concentration calculated in each sample.

An example of RTD curves is given by Figure 16 which comparatively reports two *E(t)* curves obtained at steady state from tests in the 3-module spouted bed operated with and without submerged baffles, according to the configuration appearing on Figure 11B. The two curves overlap almost perfectly to demonstrate that internal baffles do not alter at all the solids circulation in a multi-unit cascade. The same test also demonstrates that the external solids streamlines do not have any transversal (horizontal) component, as also visible from half column monitoring. The slope of solids at the free surface was modest in these runs, due to a very low throughput (2 kg/min). Thus, it is straightforward assessing that batch operations do not require any repartition between modules. From these considerations, the use of submerged baffles is beneficial to the start-up phase of continuous processes only and becomes fundamental as geometrical boundary to generate the configuration given in Figure 12.

Fig. 16. Comparison between the RTD curves in the 3-module spouted bed with *H/Dc*=1.72 with and without submerged baffles

#### **9.2 Modelling**

426 Advances in Chemical Engineering

Two types of curves can be studied. The *E(t)* function describes what a system releases instant after instant, i.e. the volumetric (or mass) fraction of particles whose residence time is between t and t+dt. The *F(t)* function provides the integral of *E(t)*·dt and represents the

The most direct way to trace an *E(t)* curve makes use of a physical tracer, whose characteristics are identical or very close to the ones characterizing the bulk of solids travelling the system. Usually, two types of stimulus are adopted, a pulse (given by a definite amount of tracer introduced into the system in the shortest time) or a step (an abrupt change from the normal feedstock to an identical feed made of tracer only). The first one is generally the prompter to use. Right away after the introduction of the pulse, samples are taken at the system exit with a proper scrutiny degree and the tracer concentration is

( ) ( )

*Ct M E t M* 

> 0 () () *t*

with *C (t)* being the concentration of solid tracer in the discharge, *Mtr* the mass of tracer injected, *ρb* the bulk density of tracer, *M* the mass flow rate of solids travelling the system. A pulse function (called Dirac function, δ(t)) is given by an instantaneous but finite entity (equal to unity) entering a system at t =0. The Laplace transforms of these functions allow

In our experiments, the pulse was obtained by quickly introducing in the feed a small amount of tracer (150 to 300 g, equivalent to about 1.5% of the total solids hold-up) made of PET chips doped with some ferromagnetic powder. After sampling the ferromagnetic PET chips were sorted out from the PET bulk by a magnet and their concentration calculated in

An example of RTD curves is given by Figure 16 which comparatively reports two *E(t)* curves obtained at steady state from tests in the 3-module spouted bed operated with and without submerged baffles, according to the configuration appearing on Figure 11B. The two curves overlap almost perfectly to demonstrate that internal baffles do not alter at all the solids circulation in a multi-unit cascade. The same test also demonstrates that the external solids streamlines do not have any transversal (horizontal) component, as also visible from half column monitoring. The slope of solids at the free surface was modest in these runs, due to a very low throughput (2 kg/min). Thus, it is straightforward assessing that batch operations do not require any repartition between modules. From these considerations, the use of submerged baffles is beneficial to the start-up phase of continuous processes only and becomes fundamental as geometrical boundary to generate the

*tr b*

(5)

*F t E t dt* (6)

fraction of elements whose residence time is lower than t.

The two *E(t)* and *F(t)* functions are defined below:

the use of simple algebraic input/output relationships.

measured and recorded.

each sample.

configuration given in Figure 12.

The theoretical description of continuous units combines basic elemental models, whose combination gives origin to a system capable of generating an overall response to properly match the actual behaviour of the real system studied. A descriptive model can produce an output without having a strict link with the actual hydrodynamic behaviour and then has to make use of fitting parameters. This approach does not allow sound predictions or extension to more complex reactor structures. A much powerful tool is produced by conceiving a phenomenological model based on experimental observations. These models become fully predictive since all parameters are based on actual measurements. Then, a model interacts with scale-up procedures through the validity of the correlations used rather than its structure.

Dynamic responses were obtained by making use of a Matlab Simulink tool by generating model schemes as given in the below Figures 17 and 18. The fundamental modelling was based on a one-stage spouted bed; a multiple-cell system was then given by a cascade of basic units.

#### **9.2.1 Descriptive model**

The initial modelling started by considering an early dynamic description of spouted beds, where the overall behaviour can be portrayed by a well-mixed system with a minor (8 to 10%) portion of plug flow. The corresponding scheme given in Figure 17 A) includes the feed rate F1, the bypass to fountain F3 (which become negligible when the inlet diverting baffle is considered, as a consequence F2 ≡ F1 and F5 ≡ F4 ), the total circulation from spout F4, the net discharge rate F7 ≡ F1 (for continuity) and F6 = F4 - F1. From the experimental conditions F1 is known and F4 can be estimated by measuring mean particle velocity at the frontal wall of half-column. As far as the other parameters than appear in Figure 17 B) are concerned, *td* is estimated from particle circulation, *τannulus* follows from holdup in the annulus and *τspout+fountain* is calculated by difference from the known bed holdup. At usual *H/Dc* ratios adopted in cylindrical columns, the ratio of these two time constants approaches one magnitude order.

Optimization of Spouted Bed Scale-Up by Square-Based Multiple Unit Design 429

The relevant limitation contained by the above model consists in the fact that the annulus, representing the massive part of a spouted bed, has not been given a proper description. By observing it through the flat transparent wall of semicylindrical columns, particles show well-defined trajectories with scarce intermixing, according to consolidated findings

The phenomenological description adopted in the updated model assumes that the squaredsection of the annulus (with a cylindrical spout *Ds*) is divided into three axisymmetric zones, each of them having the same width according to the scheme given in Figure 19 A). Each of these regions receives from the fountain a solids flow rate proportional to its cross sectional area (F6A, F6B and F6C, respectively moving from outside towards the spout). The flow fashion in each region is a piston with particle residence time *td,A*, *td,B* and *td,C*, from the bed surface down to the cylinder-frustum junction, according to experimental observations. The mixing component acting in the annulus was concentrated in the frustum, which progressively discharges solids into the spout, depending on local streamline length. As a whole, this section was assimilated as far as its dynamics is concerned to a well-mixed volume, accounting to about 20% of the total holdup of the spouted bed. Also in this case, the effect of the F4/F7 ratio overcomes the sensitivity of other variables on the model, so that the ratio between frustum to parallelepiped volumes (i.e. plug to well-mixed volume ratio)

Fig. 19. Phenomenological model of one stage four-zone continuously operating spouted bed: A): schematic of flow circulation between annulus and spout/fountain regions; B) Simulink model including pulse stimulus, three parallel delay times in annulus, followed by

spout/fountain region and recirculation to bed surface. Small well-mixed volume accounts

Figure 19 B) presents the Matlab Simulink scheme, where the RTDF is generated by introducing the pulse into the sector A of the annulus (due to geometrical constrain of the

three perfectly mixed regions in bottom frustum, one perfectly mixed zone in

**9.2.2 Phenomenological model** 

(Mathur & Epstein, 1974).

is not relevant at all.

for sampling end effects.

Fig. 17. Descriptive model of one stage two-zone continuously operating spouted bed: A): schematic of flow circulation between annulus and spout/fountain regions; B) Simulink model including pulse stimulus, delay at annulus entrance, two perfectly mixed regions for annulus and spout/fountain volumes and recirculation to bed surface.

Accepting this description also this model does not contain any fitting parameter, once the constitutive elements are assumed. Since the F4 (internal circulation) to F7 (net flow) ratio is very large, the overall system approaches a well-mixed unit and in this view the modelling is scarcely sensitive to the hydrodynamic description given to the annulus.

Figure 18 presents the comparison between experimental and modelling results for *H/Dc* = 1.72. The fitting is excellent, considering the time delay given by the minimum residence time of particles (*td*) and the well-mixed key dynamic component brought by spout recirculation.

Fig. 18. Comparison between experimental data and the descriptive model

#### **9.2.2 Phenomenological model**

428 Advances in Chemical Engineering

Fig. 17. Descriptive model of one stage two-zone continuously operating spouted bed: A): schematic of flow circulation between annulus and spout/fountain regions; B) Simulink model including pulse stimulus, delay at annulus entrance, two perfectly mixed regions for

Accepting this description also this model does not contain any fitting parameter, once the constitutive elements are assumed. Since the F4 (internal circulation) to F7 (net flow) ratio is very large, the overall system approaches a well-mixed unit and in this view the modelling

Figure 18 presents the comparison between experimental and modelling results for *H/Dc* = 1.72. The fitting is excellent, considering the time delay given by the minimum residence time of particles (*td*) and the well-mixed key dynamic component brought by spout

annulus and spout/fountain volumes and recirculation to bed surface.

is scarcely sensitive to the hydrodynamic description given to the annulus.

Fig. 18. Comparison between experimental data and the descriptive model

recirculation.

The relevant limitation contained by the above model consists in the fact that the annulus, representing the massive part of a spouted bed, has not been given a proper description. By observing it through the flat transparent wall of semicylindrical columns, particles show well-defined trajectories with scarce intermixing, according to consolidated findings (Mathur & Epstein, 1974).

The phenomenological description adopted in the updated model assumes that the squaredsection of the annulus (with a cylindrical spout *Ds*) is divided into three axisymmetric zones, each of them having the same width according to the scheme given in Figure 19 A). Each of these regions receives from the fountain a solids flow rate proportional to its cross sectional area (F6A, F6B and F6C, respectively moving from outside towards the spout). The flow fashion in each region is a piston with particle residence time *td,A*, *td,B* and *td,C*, from the bed surface down to the cylinder-frustum junction, according to experimental observations. The mixing component acting in the annulus was concentrated in the frustum, which progressively discharges solids into the spout, depending on local streamline length. As a whole, this section was assimilated as far as its dynamics is concerned to a well-mixed volume, accounting to about 20% of the total holdup of the spouted bed. Also in this case, the effect of the F4/F7 ratio overcomes the sensitivity of other variables on the model, so that the ratio between frustum to parallelepiped volumes (i.e. plug to well-mixed volume ratio) is not relevant at all.

Fig. 19. Phenomenological model of one stage four-zone continuously operating spouted bed: A): schematic of flow circulation between annulus and spout/fountain regions; B) Simulink model including pulse stimulus, three parallel delay times in annulus, followed by three perfectly mixed regions in bottom frustum, one perfectly mixed zone in spout/fountain region and recirculation to bed surface. Small well-mixed volume accounts for sampling end effects.

Figure 19 B) presents the Matlab Simulink scheme, where the RTDF is generated by introducing the pulse into the sector A of the annulus (due to geometrical constrain of the

Optimization of Spouted Bed Scale-Up by Square-Based Multiple Unit Design 431

spouted bed. The agreement is fully satisfactory, even though it may appear that experimental data anticipate the model output moderately and then a tail slightly higher than expected is displayed. This analysis could require to consider some direct bypass from fountain to downstream stage through the gap between submerged and freeboard vertical baffles and a small partially stagnant backwater, possibly existing along edges of frustum

Both cold tests on the dual and triple module square-based spouted beds and industrial operations on the sextuple module demonstration unit (Beltramo et al., 2009) show that design criteria, operating conditions and step sequence must be defined carefully. In opposite case, start up and transitory from spouting onset in one module to overall stable

 Cross-sectional area of the gap between vertical baffles: it should be designed to allow a maximum solids flow rate equal to the process solids nominal capacity with an excess of about 50% to avoid solids flooding (if too narrow) or bed emptying (if too large,

 Freeboard baffle height: it has to fully cover the fountain height at spouting onset and, if the structure of an industrial unit is considered, should reach the top deflectors of

 Cross-section of particle deflectors on gas phase path in each module (in the upper part of freeboard, above fountain projections): it has to be designed to set a progressive

 Spouting modules have to be assembled at sufficient different elevation to guarantee a steady and even solids holdup. A rule of thumb drawn from experimentation suggests that ΔH ≈ 0.5 *Dc* surely gives origin to a non-interacting system and prevents bed emptying of downstream modules, or vice versa avoid extra hold up in upstream

Given the fact that gas phase pressure drop is maximum at incipient start-up (up to 2 to 3 times of the gas pressure drop at stable spouting for conical or cylindrical shallow beds), to avoid oversize of circuit blower, with consequential dramatic increase of capital and

a. **Start-up with reduced solids holdup**: fill the first module with 1/3 to 1/2 depth with respect to the design bed load; start injecting gas till fountain formation, then begin filling the bed to reach the operating solids bed depth and then continue feeding solids at nominal flowrate to reach overflow discharge to the second module. Continue with

decrease of gas cross-sectional velocity and avoid any upward solids elutriation. The design of pyramid frustum base should consider a small gap between the orifice circumference and the slanted wall to help particle circulation. This gap is related to the average size of particulate material with a factor larger than 2. Several experiments have demonstrated that in case of local stagnation at the frustum bottom, this dead zone may spread up to the pyramid/parallelepiped junction. A careful bottom design

and/or high gas flow rate counter this undesired phenomenon.

Each module must be provided of independent gas flow regulation.

operating costs, one can proceed according to two start up routes:

the same procedure for all modules.

spouting might be a serious issue. The key design parameters affecting start-up are:

**10. Final remarks on multiple bed start-up and shut-down** 

because of direct discharge from fountain).

particles to avoid upper solids bypass.

walls.

ones.

inlet baffle). The overall solids flow rate from fountain travels three parallel annulus pistons, then each portion of solids enters the corresponding well-mixed portion of frustum, respectively characterized by a mean residence time estimated by observations at the flat frontal wall. The spout collects particles from the annulus and mix them in the fountain. A small well-mixed volume characterizes the solids sampling operation to account for end effects.

Figure 20 compares experimental results to the output of the phenomenological model. Any difference can be hardly noted with respect to the previous descriptive model output. Due to plug flow effect, a certain oscillation matching the cycle time frequency is observed. A short sampling time *τend effects* was sufficient to damp the greatest part of oscillation.

Fig. 20. Comparison between experimental data and the phenomenological model

#### **9.2.3 Model validation for multiple units**

The Matlab Simulink description conceived for the phenomenological model of one-stage spouted bed can be replicated a number of times corresponding to the number of cells included in a multistage system. An example is given in Figure 21 for a three-module

Fig. 21. Comparison between experimental and model RTDs in the 0.20 m side squaredbased three-module spouted bed

inlet baffle). The overall solids flow rate from fountain travels three parallel annulus pistons, then each portion of solids enters the corresponding well-mixed portion of frustum, respectively characterized by a mean residence time estimated by observations at the flat frontal wall. The spout collects particles from the annulus and mix them in the fountain. A small well-mixed volume characterizes the solids sampling operation to account for end

Figure 20 compares experimental results to the output of the phenomenological model. Any difference can be hardly noted with respect to the previous descriptive model output. Due to plug flow effect, a certain oscillation matching the cycle time frequency is observed. A

short sampling time *τend effects* was sufficient to damp the greatest part of oscillation.

Fig. 20. Comparison between experimental data and the phenomenological model

The Matlab Simulink description conceived for the phenomenological model of one-stage spouted bed can be replicated a number of times corresponding to the number of cells included in a multistage system. An example is given in Figure 21 for a three-module

Fig. 21. Comparison between experimental and model RTDs in the 0.20 m side squared-

**9.2.3 Model validation for multiple units** 

based three-module spouted bed

effects.

spouted bed. The agreement is fully satisfactory, even though it may appear that experimental data anticipate the model output moderately and then a tail slightly higher than expected is displayed. This analysis could require to consider some direct bypass from fountain to downstream stage through the gap between submerged and freeboard vertical baffles and a small partially stagnant backwater, possibly existing along edges of frustum walls.

#### **10. Final remarks on multiple bed start-up and shut-down**

Both cold tests on the dual and triple module square-based spouted beds and industrial operations on the sextuple module demonstration unit (Beltramo et al., 2009) show that design criteria, operating conditions and step sequence must be defined carefully. In opposite case, start up and transitory from spouting onset in one module to overall stable spouting might be a serious issue. The key design parameters affecting start-up are:


Given the fact that gas phase pressure drop is maximum at incipient start-up (up to 2 to 3 times of the gas pressure drop at stable spouting for conical or cylindrical shallow beds), to avoid oversize of circuit blower, with consequential dramatic increase of capital and operating costs, one can proceed according to two start up routes:

a. **Start-up with reduced solids holdup**: fill the first module with 1/3 to 1/2 depth with respect to the design bed load; start injecting gas till fountain formation, then begin filling the bed to reach the operating solids bed depth and then continue feeding solids at nominal flowrate to reach overflow discharge to the second module. Continue with the same procedure for all modules.

Optimization of Spouted Bed Scale-Up by Square-Based Multiple Unit Design 433

 demonstrating the achievement of fully stable operations by introducing novel concepts (stage segregation with internal baffles, sloped cascade of stages) in designing a multi-

obtaining a plug flow of solids with a sufficient number of stages, which may

modelling single unit and multiple square-based spouted beds to predict solids

The final structure of a cold model apparatus has demonstrated the achievement of all the

The economical support granted by ENGICO Srl (LT– Italy) allowed construction of the equipment and the research fellowship to one of the authors (M. C.). The authors wish to tank Mr. Alfio Traversino for patiently and carefully constructing the three-module

cell equipment,

hydrodynamics.

**12. Acknowledgments** 

listed goals.

experimental rig.

**13. Nomenclature** 

*E(t)* E function *F(t)* F function *H* bed depth, m

Greek letters:

*Ar* Archimedes number

*Dc* column diameter, m *di* inlet diameter, m *dp* particle diameter, m *Ds* mean spout diameter, m

*Mtr* mass of the tracer, kg

*U* fluid velocity, m / s

*Uonset* fluid onset velocity, m / s

free surface slope, deg

*<sup>f</sup>* fluid density, kg / m3

mean residence time, s

*<sup>b</sup>* bulk solids density , kg / m3

*<sup>s</sup>* actual material density, kg / m3

*(t)* Dirac function

shape factor

*Hm* maximum spoutable bed depth, m H gap between adjacent units, m *M* mass flow rate of solids, kg / s

*PM* maximum pressure drop across bed, Pa

*PS* spouting pressure drop across bed, Pa

*Ums* minimum spouting velocity, m / s

implement process scale-up at the same time,

*C* concentration of solid tracer in the discharge, kg / m3

b. **Start-up with full solids holdup:** fill each module with the design solids bed depth, provide the circuit with a capacitive booster section, suitable to inject high pressure gas to bed orifices for about 10 to 15 seconds and reach external spouting onset; then operate spouting process with the master circuit blower.

In a multiple-bed spouting unit, to have the whole series of spouted beds started up, one has to proceed with a start-up sequence, one by one, from the first bed to the last one, following the solids flow direction. Monitoring the gas phase pressure drop vs. time represent the trigger element for the control system to determine when the start-up of one module is completed and the situation is ready to move and start up next module. The b) procedure is safer since enough head is available to re-start the system in case of failure.

The above defined key design parameters are also suited for shutdown phase. As far as this procedure is concerned, one has to stop solids feeding, allowing holdup of each module to be processed at the steady state operating conditions (as to minimize off-spec), while decreasing each bed depth to a level lower than the overflow weir. At that point, side submerged baffles must be risen, while gas flow continues to be injected to each bed, so to have prompt residual solid emptying.

#### **11. Conclusions**

Spouted beds, throughout over half a century studies, have demonstrated to display very interesting features against bubbling conventional fluidization. Thanks their peculiar hydrodynamic structure a relavant gas rate can be saved also operating at the maximum spoutable bed depth. Again, the total frictional pressure drop across a spouting unit can be as lower as one-third of the one in a corresponding particulate material fluidization. The application of spouted bed to relatively coarse solids overcomes undesirable features characteristic of fluidization, namely random gas channelling and solids circulation, slugging and poor contacting between phases.

The scale-up issue of spouting units has remained open since their initial invention and the debate on whether prefer larger or multiple units has struck the opinion of scientists and technologists every time that this problem required a sound solution. This chapter has tackled the scale-up problem by opting for a square-based spouted bed geometry, since constructing a cascade of these vessels is economically advantageous and much more effective for the solids fluid dynamics as well as for insulation problems.

Following the experience gained during an industrial demonstration project, encouraging results came by evaluating the product quality; the potential performance of a multiple spouted bed was thus confirmed. Nevertheless, this unit required an excessive attention to govern its stability over time and the need of several improvements was highlighted.

A new research project required the construction of several apparatuses to pursue a comprehensive strategy which has aimed at:


The final structure of a cold model apparatus has demonstrated the achievement of all the listed goals.

### **12. Acknowledgments**

432 Advances in Chemical Engineering

b. **Start-up with full solids holdup:** fill each module with the design solids bed depth, provide the circuit with a capacitive booster section, suitable to inject high pressure gas to bed orifices for about 10 to 15 seconds and reach external spouting onset; then

In a multiple-bed spouting unit, to have the whole series of spouted beds started up, one has to proceed with a start-up sequence, one by one, from the first bed to the last one, following the solids flow direction. Monitoring the gas phase pressure drop vs. time represent the trigger element for the control system to determine when the start-up of one module is completed and the situation is ready to move and start up next module. The b) procedure is

The above defined key design parameters are also suited for shutdown phase. As far as this procedure is concerned, one has to stop solids feeding, allowing holdup of each module to be processed at the steady state operating conditions (as to minimize off-spec), while decreasing each bed depth to a level lower than the overflow weir. At that point, side submerged baffles must be risen, while gas flow continues to be injected to each bed, so to

Spouted beds, throughout over half a century studies, have demonstrated to display very interesting features against bubbling conventional fluidization. Thanks their peculiar hydrodynamic structure a relavant gas rate can be saved also operating at the maximum spoutable bed depth. Again, the total frictional pressure drop across a spouting unit can be as lower as one-third of the one in a corresponding particulate material fluidization. The application of spouted bed to relatively coarse solids overcomes undesirable features characteristic of fluidization, namely random gas channelling and solids circulation,

The scale-up issue of spouting units has remained open since their initial invention and the debate on whether prefer larger or multiple units has struck the opinion of scientists and technologists every time that this problem required a sound solution. This chapter has tackled the scale-up problem by opting for a square-based spouted bed geometry, since constructing a cascade of these vessels is economically advantageous and much more

Following the experience gained during an industrial demonstration project, encouraging results came by evaluating the product quality; the potential performance of a multiple spouted bed was thus confirmed. Nevertheless, this unit required an excessive attention to

A new research project required the construction of several apparatuses to pursue a

comparing the fundamental operating parameters of square-based spouted beds with

govern its stability over time and the need of several improvements was highlighted.

the corresponding values characteristic of conventional cylindrical columns,

effective for the solids fluid dynamics as well as for insulation problems.

operate spouting process with the master circuit blower.

have prompt residual solid emptying.

slugging and poor contacting between phases.

comprehensive strategy which has aimed at:

carrying out an adequate experimental scale-up,

**11. Conclusions** 

safer since enough head is available to re-start the system in case of failure.

The economical support granted by ENGICO Srl (LT– Italy) allowed construction of the equipment and the research fellowship to one of the authors (M. C.). The authors wish to tank Mr. Alfio Traversino for patiently and carefully constructing the three-module experimental rig.

#### **13. Nomenclature**



**17** 

*Brazil* 

**Techno-Economic Evaluation of Large Scale 2.5-Dimethylfuran Production from Fructose** 

In an era of increasing oil prices and climate concerns, biofuels have gained more and more attention as potential fuel alternative energy sources. Governments have become active in the target of securing a supply of raw materials and limiting climate change, and many innovative proposals have been made, development work has started and potential

A number of factors must be considered when evaluating biofuels: technical factors (raw materials, supply, conversion and engines), economic (engine modification cost, infrastructure) and ecological/political (greenhouse gases, land use efficiency, oil

An end-user survey assessed car customer acceptance and attitude toward biofuels and revealed that their main demands are: price (48%), biofuel price should not exceed fossil fuels prices and there should be no cost in engine modification; environment (24%);

Since customers consider the final cost as a decisive factor, the economic analysis is an important tool in the assessment of the success of biofuel production process and consequent market success. Achieving economic viability used to be the key to success, but

Leshkov et al. (2007) show a catalytic strategy for the production of 2.5 dimethylfuran (DMF) from fructose (a carbohydrate obtained directly from biomass or by the isomerization of glucose) for use as a liquid transportation fuel. Compared to ethanol, 2.5-dimethylfuran has a higher energy density (by 40 percent), a higher boiling point (by 20K), and is not soluble in water. This catalytic strategy creates a route for transforming abundant renewable biomass resources into a liquid fuel suitable or the transportation sector and it is also a CO2

The first step in production is to convert fructose to hydroxymethylfurfural (HMF) using an acid catalyst (HCl) and a solvent with a low boiling point in a biphasic reactor. The reactive aqueous phase in the biphasic reactor contains acid and sugar, and the extractive phase contains a partially miscible organic solvent (eg, 1-butanol) that continuously extracts HMF. The addition of a salt to the aqueous phase improves the partitioning of HMF into the

candidate fuels have been studied in the energy area (Schaub & Vetter, 2008).

**1. Introduction** 

free process.

dependence reduction) (Festel, 2008).

consumption (19%) and performance (9%) (Festel, 2008).

today, other factors are important, such as sustainability.

Fábio de Ávila Rodrigues and Reginaldo Guirardello *State University of Campinas, School of Chemical Engineerging* 

#### **14. References**


Fábio de Ávila Rodrigues and Reginaldo Guirardello *State University of Campinas, School of Chemical Engineerging Brazil* 

#### **1. Introduction**

434 Advances in Chemical Engineering

Beltramo, C.; Rovero, G. & Cavaglià, G. (2009). Hydrodynamics and thermal

Cavaglià, G. (2003), "Reactor and process for solid state continuous polymerisation of poly-

Chatterjee, A. (1970). Spout-fluid bed technique. *Ind. Eng. Chem. Process Des. Develop*, Vol.9,

Epstein, N. & Grace, J. (2011). *Spouted and spout-fluid beds*. Cambridge Univ. Press*,* ISBN 978-

Gishler, P.E. & Mathur, K.B. (1957). Method of contacting solid particles with fluids. U.S.

Grbavčić, Ž.B.; Vuković, D.V.; Hadžismajlovic, D. E.; Garić, R. V. & Littman, H. (1982). Fluid

Mathur, K.B. & Epstein, N. (1974). *Spouted beds*, Academic Press*,* ISBN 0-12-480050-5, New

McNab. G.S. (1972). Prediction of spout diameter. *Brit. Chem. Eng & Proc. Techn.*, Vol.17, 532 McNab, G.S. & Bridgwater, J. (1977). Spouted beds – estimation of spouting pressure drop

Metcalf, J. R. (1965-66). The mechanics of the screw feeder. *Proc. Inst Mech. Eng*., Vol.180,

Murthy, D.V.R. & Singh, P.N. (1994). Minimum spouting velocity in multiple spouted beds.

Piccinini, N. (1980). Particle segregation in continuously operating spouted beds, In:

Rovero, G.; Piccinini, N. & Lupo, A. (1985). Vitesses des particules dans les lits à jet

Saidutta, M.B. & Murthy, D.V.R. (2000). Mixing behavior of solids in multiple spouted beds.

tridimensional et semi-cylindriques. *Entropie*, Vol.124, 43-49

mechanical behaviour of a spouted bed with draft tube and external annular flow. *2nd Int. Symp. on Spouted Beds, 32nd Can. Chem. Eng. Conf.*, Vancouver, Canada

and the particle size for deepest bed. *Proc. of European Council on Particle Technology*,

*Fluidization III*, J.R Grace & J.M. Matsen*,*(Eds), 279-285, Plenum Press, ISBN 0-306-

Patent No. 2,786,280 to National Research Council of Canada

Geldart, D. (1973). Type of gas fluidization. *Powder Tech.*, Vol.7, 285-292 Malek, M.A. & Lu, B.C.Y. (1965). *I&EC Process. Des. Develop.*, Vol.4, 123-127

scale-up. *The Can. J. Chem. Eng.*, Vol.87, 394-402

ethylene terephthalate (PET)" Patent EP 1576028 B1

experimentation on square-based spouted beds for polymer upgrading and unit

**14. References** 

340-341

York

131-146

0-521-51797-3,New York

Nuremberg, Germany

40458-3, New York, USA

*The Can. J. Chem. Eng.*, Vol.72, 235-239

*The Can. J. Chem. Eng.*, Vol.78, 382-385

In an era of increasing oil prices and climate concerns, biofuels have gained more and more attention as potential fuel alternative energy sources. Governments have become active in the target of securing a supply of raw materials and limiting climate change, and many innovative proposals have been made, development work has started and potential candidate fuels have been studied in the energy area (Schaub & Vetter, 2008).

A number of factors must be considered when evaluating biofuels: technical factors (raw materials, supply, conversion and engines), economic (engine modification cost, infrastructure) and ecological/political (greenhouse gases, land use efficiency, oil dependence reduction) (Festel, 2008).

An end-user survey assessed car customer acceptance and attitude toward biofuels and revealed that their main demands are: price (48%), biofuel price should not exceed fossil fuels prices and there should be no cost in engine modification; environment (24%); consumption (19%) and performance (9%) (Festel, 2008).

Since customers consider the final cost as a decisive factor, the economic analysis is an important tool in the assessment of the success of biofuel production process and consequent market success. Achieving economic viability used to be the key to success, but today, other factors are important, such as sustainability.

Leshkov et al. (2007) show a catalytic strategy for the production of 2.5 dimethylfuran (DMF) from fructose (a carbohydrate obtained directly from biomass or by the isomerization of glucose) for use as a liquid transportation fuel. Compared to ethanol, 2.5-dimethylfuran has a higher energy density (by 40 percent), a higher boiling point (by 20K), and is not soluble in water. This catalytic strategy creates a route for transforming abundant renewable biomass resources into a liquid fuel suitable or the transportation sector and it is also a CO2 free process.

The first step in production is to convert fructose to hydroxymethylfurfural (HMF) using an acid catalyst (HCl) and a solvent with a low boiling point in a biphasic reactor. The reactive aqueous phase in the biphasic reactor contains acid and sugar, and the extractive phase contains a partially miscible organic solvent (eg, 1-butanol) that continuously extracts HMF. The addition of a salt to the aqueous phase improves the partitioning of HMF into the

all molecules in solution, solvent or electrolyte (Aznar, 1996). Therefore, no specific model for electrolytes was used in this study. The UNIQUAC model was used to obtain the activity coefficient. According to Mock et al. (1986), although the contribution of long-range interaction of the equation of Pitzer-Debye-Huckel is important to obtain the ionic activity coefficient in the aqueous phase, it has little effect on the behavior of the equilibrium phase of the water-organic solvent system. Thus, the effect of the electrolyte is considered only for

The binary interaction parameters of UNIQUAC model were estimated from experimental data (Santis et al., 1976a, 1976b), from Catté et al. (1994) and for the UNIFAC method. The tables 2, 3 e 4 show the data experimental used to estimate the binary interactions parameters. A Fortran programming language was used to determine the parameters from experimental data. The UNISIMTM software was used to estimate parameters for the UNIFAC method.

Water - 89.22\* 1543' -153.35\*\* 1160' 1361' -455\* 1-butanol 208\* - 383' 223' 1421' 530' 863\* DMF 249' -90' - 73' - -146' -371' Fructose 324\*\* 91' 892' - -197' 1.412' 160' HCl -674' -179' - 399' - 702' -266' HMF -121' -1155' 564' 162' 2,776' - 2479' NaCl -165\* 1251\* 1793' 354' 2943' 1391' - \* Binary interaction parameters of UNIQUAC model estimated from experimental data from Santis et

Aqueous Phase Organic Phase Water 1-butanol NaCl Water 1-butanol NaCl

92.60 7.4 - 20.4 79.6 - 92.04 6.8 1.16 18.78 81.2 0.025 91.64 6.1 2.26 17.45 82.5 0.045 90.85 5.8 3.35 16.64 83.3 0.061 90.60 5.0 4.40 15.43 84.5 0.074 89.96 4.6 5.44 14.6 85.3 0.086 89.16 4.4 6.44 14.1 85.8 0.095 87.84 3.7 8.46 13.29 86.6 0.110 86.30 3.3 10.4 12.48 87.4 0.122 85.10 2.7 12.2 11.37 88.5 0.130 83.50 2.5 14.0 10.66 89.2 0.138 82.20 2.0 15.8 8.75 90.1 0.144 80.90 1.7 17.4 9.05 90.8 0.148 79.40 1.5 19.1 8.55 91.3 0.153 78.00 1.3 20.7 7.94 91.9 0.156 76.70 1.1 22.2 7.34 92.5 0.159 75.20 0.9 23.9 7.04 92.8 0.162 74.00 0.8 25.2 6.54 93.3 0.164 73.30 0.8 25.9 6.43 93.4 0.167 73.30 0.8 25.9 6.23 93.6 0.167 Table 2. Liquid-liquid equilibrium in the system water-1-butanol-NaCl (Santis et al., 1976a)

Water 1-butanol DMF Fructose HCl HMF NaCl

non-ideality, represented by the adjustable model parameters.

' Binary interaction parameter estimated from UNIFAC method. Table 1. Binary interaction parameters of UNIQUAC model

al. (1976a, 1976b). \*\* Catté et al. (1994).

extracting phase, and leads to increased HMF yields without the use of high boiling point solvents. Following, water, HCl and solvent evaporate, leading to precipitation of NaCl. Then, HMF is converted into DMF under a copper-ruthenium based catalyst. The final step involves the separation of DMF from the solvent and the non-reacted intermediates. As described below, the process also involves two separation steps. A schematic diagram of fructose conversion to DMF was reported by Leshkov et al. (2007).

The purpose of this paper is to evaluate economically the process production of DMF from fructose. In the present work the following analysis were carried out: Firstly, thermodynamic process modeling was investigated. Following this, the Process Flow Diagram (PFD) was developed from schematic diagram reported by Leshkov et al. (2007). The simulation stage makes use data from Leshkov et al. (2007). The thermal energy required for each piece of equipment was assessed with material and energy balances for each system using the UNISimTM software. Each piece of equipment is then approximately sized for economic analysis.

#### **2. Thermodynamic modeling**

The thermodynamic equilibrium of a system consisted of a multicomponent mixture takes place when temperature, pressure and chemical potential of components are equated between the phases, for each component. Although there are other basic criteria for system equilibrium, the minimization of Gibbs free energy is the condition which ensures equilibrium. Salt can affect the solubility of the system components with the formation of complex associations. In general it can be inferred that the particles (molecules, ions, or both) of dissolved salt tend to attract molecules from one of the system components more strongly than others.

The work of Debye and Huckel (1923) was the first important academic contribution and established a model for long-range interactions between ions based on the concept of ionic strength. A different family of models was developed using another extension of the Debye-Huckel model to represent the different contributions to Gibbs free energy excess. Implementation of the local composition to electrolytes means it is governed by local interactions such as short-range solvent/solvent, short-range ion/solvent and long-range ion/ion interactions that exist around the immediate neighborhood of a central ionic species (Aznar, 1996). For the contribution of short-range the following models of local composition can be used: Non-Random Two Liquid model (NRTL) (Renon and Prausnitz, 1968), UNIQUAC (Abrams and Prausnitz, (1975)) or UNIversal Functional Activity Coefficient (UNIFAC) (Fredenslund et al., 1977). The Debye-Huckel term or one of its variations, such as Fowler and Guggnheim (1949) or the Pitzer (1973) are used for long-range interactions. A series of different combinations have been proposed with these elements.

The hypothesis in this work was that salt should be treated as simple molecule, nondissociated, rather than as charged ions distributed in the solution. Most works concerning the phase equilibrium in systems containing electrolytes distinguish long-range contributions due to electrostatic forces between ions and between ion and solvent from short-range contributions due to interactions between molecules. Two different models are then used for each contribution type. Considering salt as a simple molecule eliminates both contributions and requires only one appropriate model to describe the interactions between all molecules in solution, solvent or electrolyte (Aznar, 1996). Therefore, no specific model for electrolytes was used in this study. The UNIQUAC model was used to obtain the activity coefficient. According to Mock et al. (1986), although the contribution of long-range interaction of the equation of Pitzer-Debye-Huckel is important to obtain the ionic activity coefficient in the aqueous phase, it has little effect on the behavior of the equilibrium phase of the water-organic solvent system. Thus, the effect of the electrolyte is considered only for non-ideality, represented by the adjustable model parameters.

The binary interaction parameters of UNIQUAC model were estimated from experimental data (Santis et al., 1976a, 1976b), from Catté et al. (1994) and for the UNIFAC method. The tables 2, 3 e 4 show the data experimental used to estimate the binary interactions parameters. A Fortran programming language was used to determine the parameters from experimental data. The UNISIMTM software was used to estimate parameters for the UNIFAC method.


\* Binary interaction parameters of UNIQUAC model estimated from experimental data from Santis et al. (1976a, 1976b).

\*\* Catté et al. (1994).

436 Advances in Chemical Engineering

extracting phase, and leads to increased HMF yields without the use of high boiling point solvents. Following, water, HCl and solvent evaporate, leading to precipitation of NaCl. Then, HMF is converted into DMF under a copper-ruthenium based catalyst. The final step involves the separation of DMF from the solvent and the non-reacted intermediates. As described below, the process also involves two separation steps. A schematic diagram of

The purpose of this paper is to evaluate economically the process production of DMF from fructose. In the present work the following analysis were carried out: Firstly, thermodynamic process modeling was investigated. Following this, the Process Flow Diagram (PFD) was developed from schematic diagram reported by Leshkov et al. (2007). The simulation stage makes use data from Leshkov et al. (2007). The thermal energy required for each piece of equipment was assessed with material and energy balances for each system using the UNISimTM software. Each piece of equipment is then approximately

The thermodynamic equilibrium of a system consisted of a multicomponent mixture takes place when temperature, pressure and chemical potential of components are equated between the phases, for each component. Although there are other basic criteria for system equilibrium, the minimization of Gibbs free energy is the condition which ensures equilibrium. Salt can affect the solubility of the system components with the formation of complex associations. In general it can be inferred that the particles (molecules, ions, or both) of dissolved salt tend to attract molecules from one of the system components more

The work of Debye and Huckel (1923) was the first important academic contribution and established a model for long-range interactions between ions based on the concept of ionic strength. A different family of models was developed using another extension of the Debye-Huckel model to represent the different contributions to Gibbs free energy excess. Implementation of the local composition to electrolytes means it is governed by local interactions such as short-range solvent/solvent, short-range ion/solvent and long-range ion/ion interactions that exist around the immediate neighborhood of a central ionic species (Aznar, 1996). For the contribution of short-range the following models of local composition can be used: Non-Random Two Liquid model (NRTL) (Renon and Prausnitz, 1968), UNIQUAC (Abrams and Prausnitz, (1975)) or UNIversal Functional Activity Coefficient (UNIFAC) (Fredenslund et al., 1977). The Debye-Huckel term or one of its variations, such as Fowler and Guggnheim (1949) or the Pitzer (1973) are used for long-range interactions. A

The hypothesis in this work was that salt should be treated as simple molecule, nondissociated, rather than as charged ions distributed in the solution. Most works concerning the phase equilibrium in systems containing electrolytes distinguish long-range contributions due to electrostatic forces between ions and between ion and solvent from short-range contributions due to interactions between molecules. Two different models are then used for each contribution type. Considering salt as a simple molecule eliminates both contributions and requires only one appropriate model to describe the interactions between

series of different combinations have been proposed with these elements.

fructose conversion to DMF was reported by Leshkov et al. (2007).

sized for economic analysis.

strongly than others.

**2. Thermodynamic modeling** 

' Binary interaction parameter estimated from UNIFAC method.

Table 1. Binary interaction parameters of UNIQUAC model


Table 2. Liquid-liquid equilibrium in the system water-1-butanol-NaCl (Santis et al., 1976a)

Simulation of DMF plant production was based on the standard conditions by Leshkov et al. (2007) from which it was developed a process flow diagram (PFD). The following unit operations compose the production plant: pumps, heat exchangers, one reactor for conversion of fructose into HMF (CRV-102) and one reactor for conversion of HMF into DMF (CRV-101), two stripping columns (T-100 and T-101), one distillation column (T-102). The volume of feed was of 30% fructose and the ratio between the organic phase and the aqueous phase volume was of 3.1 in the biphasic reactor (CRV-102). The conversion of

Unreacted fructose was recycled back into the biphasic reactor. 1-Butanol was then separated from the water in the organic biphasic reactor. Cezário et al. (2009) proposed a separation system for water and 1-butanol composed by two stripping columns, one cooler and one settling tank. The formation of heterogeneous azeotrope turns this separation process more difficult and two liquid phases are formed in the decanter. This system can separate 98% of 1-butanol. Literature provides various processes for separating 1-butanol from water but the most traditional recovery process is distillation. Other techniques are adsorption, liquid-liquid extraction, evaporation and reverse osmosis. The energy required to recover 1-butanol by adsorption is of 1948 kcal/kg while the stripping column method requires 5789 kcal/kg. Other techniques such as perevaporation requires 3295 kcal/kg 1 butanol (Qureshi et al., 2005). The last step was to separate DMF from 1-butanol. The proposed separation system was composed by a distillation column (T-102) which separated 92% of DMF. The T-102 operates with reflux rate of 85 kgmol/h and top component (DMF)

Thus, material and energy balances were then solved using UNISimTM software and is

Temperature (oC) 25 25 25 25 25 Pressure (kPa) 101 101 101 101 1351 Massic flow (kg/h) 896 1000 864 52 896 Enthalpy (kJ/kgmol) -275200 -290600 -290500 -326600 -275200

water 0.9 0.7 0.5372 - 0.9 1-butanol - - - 1 - Fructose - 0.3 - - - Hmf - - - - - NaCl - - 0.4628 - - HCl 0.1 - - - 0.1 NaOH - - - - - DMF - - - - -

Temperature (oC) 25 25 180 180 180

1h 2h 3h 5h 8

9 10 11 12 13

fructose is 75% and the conversion of HMF to DMF is 100%.

fraction of 0.9. The 1-butanol recovered in the T-102 was recycled.

**3. Simulation** 

showed in Table 5.

Composition (massic fraction)


Table 3. Liquid-liquid equilibrium in the system water-1-butanol-NaCl (Santis et al., 1976a)


Table 4. Liquid-liquid equilibrium in the system water-1-butanol-NaCl (Santis et al., 1976b)

#### **3. Simulation**

438 Advances in Chemical Engineering

Aqueous Phase Organic Phase water 1-butanol NaCl water 1-butanol NaCl

92.90 7.10 - 20.60 79.4 - 92.43 6.42 1.15 18.77 81.2 0.026 92.04 5.70 2.26 17.75 82.2 0.045 91.22 5.44 3.34 16.74 83.2 0.061 90.59 5.00 4.41 15.92 84.0 0.075 89.96 4.59 5.45 15.31 84.6 0.086 89.28 4.24 6.48 14.70 85.2 0.096 87.9 3.60 8.50 13.69 86.2 0.111 86.46 3.04 10.5 12.68 87.2 0.123 85.01 2.59 12.4 11.87 88.0 0.132 83.35 2.45 14.2 11.26 88.6 0.140 82.19 1.81 16.0 1.15 89.7 0.146 80.79 1.51 17.7 9.35 90.5 0.150 79.67 1.23 19.1 8,74 91.1 0.155 77.97 1.03 21.0 7.84 92.0 0.158 76.51 0.89 22.6 7.44 92.4 0.161 75.15 0.75 24.1 6.94 92.9 0.164 73.81 0.69 25.5 6.63 93.2 0.166 73.34 0.68 26.0 6.43 93.4 0.169 Table 3. Liquid-liquid equilibrium in the system water-1-butanol-NaCl (Santis et al., 1976a)

Aqueous Phase Organic Phase water 1-butanol NaCl water 1-butanol NaCl

93.40 6.60 - 21.40 78.6 - 92.85 5.96 1.19 19.68 80.3 0.021 92.44 5.25 2.31 18.36 81.6 0.041 91.88 4.70 3.42 17.54 82.4 0.060 90.86 4.61 4.53 16.83 83.1 0.072 90.21 4.21 5.58 16.31 83.6 0.085 89.49 3.86 6.65 15.50 84.4 0.095 88.04 3.27 8.69 14.49 85.4 0.111 86.62 2.78 10.6 13.58 86.3 0.122 85.19 2.31 12.5 12.77 87.1 0.131 83.70 1.90 14.4 11.96 87.9 0.139 82.10 1.70 16.2 11.25 88.6 0.146 80.92 1.28 17.8 10.35 89.5 0.151 79.59 1.01 19.4 9.44 90.4 0.156 78.18 0.82 21.0 8.84 91.0 0.160 76.74 0.66 22.6 8.34 91.5 0.163 75.36 0.54 24.1 7.83 92.0 0.166 73.9 0.50 25.6 7.53 92.3 0.170 73.42 0.48 26.1 7.03 92.8 0.173 73.42 0.48 26.1 7.03 92.8 0.173 Table 4. Liquid-liquid equilibrium in the system water-1-butanol-NaCl (Santis et al., 1976b)

Simulation of DMF plant production was based on the standard conditions by Leshkov et al. (2007) from which it was developed a process flow diagram (PFD). The following unit operations compose the production plant: pumps, heat exchangers, one reactor for conversion of fructose into HMF (CRV-102) and one reactor for conversion of HMF into DMF (CRV-101), two stripping columns (T-100 and T-101), one distillation column (T-102). The volume of feed was of 30% fructose and the ratio between the organic phase and the aqueous phase volume was of 3.1 in the biphasic reactor (CRV-102). The conversion of fructose is 75% and the conversion of HMF to DMF is 100%.

Unreacted fructose was recycled back into the biphasic reactor. 1-Butanol was then separated from the water in the organic biphasic reactor. Cezário et al. (2009) proposed a separation system for water and 1-butanol composed by two stripping columns, one cooler and one settling tank. The formation of heterogeneous azeotrope turns this separation process more difficult and two liquid phases are formed in the decanter. This system can separate 98% of 1-butanol. Literature provides various processes for separating 1-butanol from water but the most traditional recovery process is distillation. Other techniques are adsorption, liquid-liquid extraction, evaporation and reverse osmosis. The energy required to recover 1-butanol by adsorption is of 1948 kcal/kg while the stripping column method requires 5789 kcal/kg. Other techniques such as perevaporation requires 3295 kcal/kg 1 butanol (Qureshi et al., 2005). The last step was to separate DMF from 1-butanol. The proposed separation system was composed by a distillation column (T-102) which separated 92% of DMF. The T-102 operates with reflux rate of 85 kgmol/h and top component (DMF) fraction of 0.9. The 1-butanol recovered in the T-102 was recycled.


Thus, material and energy balances were then solved using UNISimTM software and is showed in Table 5.

Temperature (oC) 74 71 36 36 71 Pressure (kPa) 50 50 50 50 50 Massic flow (kg/h) 78804 55229 59614 19189 55230 Enthalpy (kJ/kgmol) -224500 -197800 -245800 -284100 -197800

water 0.2585 0.0275 0.0255 0.4822 0.02752 1-butanol 0.1716 0.1647 0.2217 0.01592 0.1647 Fructose - - - - - Hmf 0.00341 - 0.004517 - - NaCl - - - - - HCl - - - - - NaOH - - - - - DMF 0.5665 0.8077 0.7483 0.00177 0.8077

33 34 35 38 39

Temperature (oC) 100 101 219 71 99 Pressure (kPa) 50 1650 1650 50 50 Massic flow (kg/h) 4384 4384 4384 290 4094 Enthalpy (kJ/kgmol) -390900 -308800 -279700 -215100 -310900

water - - - - - 1-butanol 0.9388 0.9388 0.9388 0.07891 0.9999 Fructose - - - - - Hmf 0.0611 0.0611 0.0611 - - NaCl - - - - - HCl - - - - - NaOH - - - - - DMF - - - 0.9221 0.000086

Composition (massic fraction)

Composition (massic fraction)

40

Temperature (oC) - Pressure (kPa) 0.9388 Massic flow (kg/h) - Enthalpy (kJ/kgmol) 0.0611

water - 1-butanol - Fructose - Hmf - NaCl - HCl 0.9388 NaOH - DMF 0.0611

Composition (massic fraction)

Table 5.



DMF 0.0611 Table 5.

water - 1-butanol - Fructose - Hmf - NaCl - HCl 0.9388 NaOH -

440 Advances in Chemical Engineering

Pressure (kPa) 1351 101 1351 1351 1351 Massic flow (kg/h) 1000 864 4142 896 1000 Enthalpy (kJ/kgmol) -290600 -290500 -326600 -263000 -278700

Água 0.7 0.5372 - 0.9 0.9589 1-butanol - - 1 - - Frutose 0.3 - - - 0.0411 Hmf - - - - - NaCl - 0.4628 - - - HCl - - - 0.1 - NaOH - - - - - DMF - - - - -

14 16 15 18 19

Temperature (oC) 180 180 180 180 37 Pressure (kPa) 1351 1351 1351 1351 1351 Massic flow (kg/h) 864 7392 489 7392 7392 Enthalpy (kJ/kgmol) -260600 -291400 -274800 -277200 -294200

water 0.7901 0.3250 0.8825 0.3250 0.3250 1-butanol - 0.5603 - 0.5603 0.5603 Fructose - 0.0484 0.1174 0.0121 0.0121 Hmf - - - 0.0363 0.0363 NaCl 0.2098 0.0541 - 0.0541 0.0541 HCl - 0.0122 - 0.0218 0.0218 NaOH - - - 0 0 DMF - 0.00005 - 0 0

22 29 30 25 24

Temperature (oC) 37 37 37 80 81 Pressure (kPa) 1351 1351 1351 50 50 Massic flow (kg/h) 7592 4876 2716 23575 489 Enthalpy (kJ/kgmol) -296500 -311400 -240600 -282400

water 0.3353 0.0887 0.7781 0.7995 0.8822 1-butanol 0.5456 0.8444 0.0091 0.1876 - Fructose 0.0118 0.0118 0.01168 - 0.1178 Hmf 0.0353 0.0549 - 0.01138 - NaCl 0.0719 - 0.2011 - - HCl - - - - - NaOH - - - - - DMF - - - 0.0015 -

26 27 Organic

phase

Aqueous

phase <sup>32</sup>

Composition (massic fraction)

Composition (massic fraction)

Composition (massic fraction)

number equipment Parameter equipment Equipment cost (\$),

Heaty duty: 10000 KW Height: 4 m; diameter: 1m

Fraction of delivered equipment % chosed

Fluid processing plant (C)

Solid fluid processing plant (B)

E' 1.609 Delivery, fraction of E' 0.10 0.10 0.10 0.10 0.1604

instalation 0.45 0.39 0.47 0.39 0.6903

controls (installed) 0.18 0.26 0.36 0.26 0.460 Piping (installed) 0.16 0.31 0.68 0.31 0.548

(installed) 0.10 0.10 0.11 0.10 0.177

T-101 stripping colunn Height: 4 m; diameter: 1m 15003 T-102 destilation colunn Height: 4 m; diameter: 1m 15003 **Total 1609365** 

P-101 pump flow: 4.97 m3/h 3814 P-102 pump flow: 0.92 m3/h 2910 P-103 pump flow 0.864 m3/h 2898 P-104 pump flow: 0.63 m3/h 2845 P-105 pump flow: 0.54 m3/h 2823 P-100 pump Flow: 5.27 m3/h 4670 E-102 heater Heaty duty: 302, 5 KW 64862 E-103 heater Heaty duty: 160 KW 42107 E-104 heater Heaty duty: 134 KW 37333 E-105 heater Heaty duty: 107 KW 320119 E-106 heater Heaty duty: 8790 KW 637790 E-108 heater Heaty duty: 100 KW 30610 E-100 cooler Water flow: 198 m3/h 11421 E-101 heater Heaty duty:467 KW 87085 CRV-102 reactor Heaty duty: 10000 KW 2917722 CRV-100 reactor Heaty duty: 344,75 KW 22354

CEPCI = 499,6

2917722 15003

> Calculated values, million (\$)

(B)

Equipment

CRV-101 T-100

**Cost directs** 

Purchased equipment,

Purchased equipment

Instrumentation &

Eletrical systems

reactor

Table 6. Equipment parameter

stripping colunn

Solid processing plant (A)

Fig. 1. Material and energy balance for each stream in DMF production plant.

#### **4. Economic evaluation**

The economic evaluation was based on the spreadsheets by Peters & Timmethaus (2003). The following steps were used by performed the economic analysis.


Each piece of equipment was roughly sized from material and energy balance and the approximate cost determined. Costs of equipment operating at ambient pressure and using carbon steel, were estimated by Eq. (1) (Turton et al., 2003).

$$\log \text{Cp}^{\text{o}} = \text{K}\_1 + \text{K}\_2 \log \text{(A)} + \text{K}\_3 \text{(log(A))}^2 \tag{1}$$

Where A is the equipmen t capacity or size parameter and K1, K2 and K3 are the parameters (Turton et al., 2003). The effect of time, operating conditions and material construction on


Table 6. Equipment parameter

442 Advances in Chemical Engineering

Fig. 1. Material and energy balance for each stream in DMF production plant.

The following steps were used by performed the economic analysis.

equipament was entered. The results are showed in Table 7.

carbon steel, were estimated by Eq. (1) (Turton et al., 2003).

The economic evaluation was based on the spreadsheets by Peters & Timmethaus (2003).

i. On the sheet '**Capital Inv**.' The estimated current total purchased cost of the

ii. On the sheet '**Materials & Labor**' the product prices and flowrates, the raw materials prices and flow rates, and the labor requirements were entered. The results are showed

iii. On the sheet '**Utilities**' the quantity of each utility needed annually was entered in appropriate units. The total annual utilities cost is transferred to sheet '**Annual TPC**'; iv. The '**Depreciation**' sheet is used only if the user wishes to change the default (5-year **Modified Accelerated Cost Recovery System (**MACRS) depreciation method); v. On the '**Annual TPC**' sheet, all values were calculated from information available on other sheets. The Calculated annual TPC was transferred to '**Evaluation**'. The results

vi. The sheet '**Evaluation**' used values from other sheets to calculate the common profitability measures. All calculations in 'Evaluation' are made in current (i.e. inflated)

Each piece of equipment was roughly sized from material and energy balance and the approximate cost determined. Costs of equipment operating at ambient pressure and using

Where A is the equipmen t capacity or size parameter and K1, K2 and K3 are the parameters (Turton et al., 2003). The effect of time, operating conditions and material construction on

o 2

12 3 logCp K K log(A) K (log(A)) (1)

**4. Economic evaluation** 

in Table 8.

dollars.

are showed in Table 9.


Basis Cost (million) US\$/year (D)

Cost (milion) US\$/year (E)

Item (A) Factor (B) Basis (C)

 1. Raw materials 4.204 2. operating labor (M) 0.885 3. operating supervision 0.15 de (2) 0.885 0.133 4. utilities 0.55 5. maintenance and repairs (MR) 0.06 de FCI 1.407 0.46 6. operating supplies 0.15 de (5) 0.084 0.07 7. laboratory charges 0.15 de (2) 0.885 0.133 8. Royalties 0.01 de co 3.090 0.08 9. catalysts and solvents 0 2.5

**Variable cost** 6.536 10. taxes (property) 0.02 de FCI 1.407 0.156 11. financing (interest) 0 de FCI 1.407 0 12. insurance 0.01 de FCI 1.407 0.014 13. rent 0 1.407 0.078

**C** 0.233 15. plant overhead, general 0.6 1.102 0.891 16. administration 0.661 17. manufacturing cost 7.660 18. administration 0.2 de (2), (5) 1.102 0.297 19. distribution & selling 0.05 de co 3.090 0.437 20. Research & development 0.04 de co 3.090 0.350

General expense 0.469 Total product cost without depreciation = co 8.744

purchased equipment cost was corrected by time factor (I), material factor (FM) and

Cp = Cp0 FM FP I (2)

14. depreciation Calculated separately

Table 9. Annual total product cost at 100 % capacity

conditions factor (FP). Purchased equipment cost is then expressed by:


Table 7. Estimation of capital investment by percentage of delivered equipment method


Table 8. Annual raw material costs and products values


444 Advances in Chemical Engineering

Solid fluid processing plant (B)

services) 0.25 0.29 0.18 0.29 0.513 Yard improvements 0.15 0.12 0.10 0.12 0.212

(installed) 0.40 0.55 0.70 0.55 0.973 Total direct costs 5.345

supervision 0.33 0.32 0.33 0.32 0.566 Construction expenses 0.39 0.34 0.41 0.34 0.602 Legal expenses 0.04 0.04 0.04 0.04 0.071 Contractor's fee 017 0.19 0.22 0.19 0.336 contigency 0.35 0.37 0.44 0.37 0.655 Total indirect costs 2.23

investimento 7.575 Working capital 0.70 0.75 0.89 0.75 1.327 Total capital investment 8.903 Table 7. Estimation of capital investment by percentage of delivered equipment method

Material classification Price (US\$/kg) Annual amount

DMF Product variable 2.217 0.88 Fructose Raw material 0.78 (variable) 2.484 1.94(62%) NaCl Raw material 0.015 0.05 0.0007(0.02%) HCl Raw material 0.295 0.745 0.22(7.7%) 1-butanol Raw material 1.72 0.431 0.7906(23%) water Raw material 0.08 16.312 1.30(2.6%) NaOH Raw material 0.10 25 2.5(2.7%) H2 Raw material 10 0.13834 1.3834(1.4%)

Solid processing plant (A)

Table 8. Annual raw material costs and products values

Buildings (incluing

Service facilities

**Indirect costs**  Engineering and

Fixed capital

Fraction of delivered equipment % chosed

Fluid processing plant (C)

(million kg/year)

(B)

Calculated values, million (\$)

Annual value (million US\$/year)

Table 9. Annual total product cost at 100 % capacity

purchased equipment cost was corrected by time factor (I), material factor (FM) and conditions factor (FP). Purchased equipment cost is then expressed by:

$$\mathbf{Cp} = \mathbf{Cp}^0 \times \mathbf{F}\_\mathbf{M} \times \mathbf{F}\_\mathbf{P} \times \mathbf{I} \tag{2}$$

Thus, the estimated cost of the equipment was U.S. \$ 12 million, fixed capital investment was U.S. \$58 million, direct cost were U.S. \$41 million, indirect costs were U.S. \$17 million, working capital was U.S. \$10 million and total capital investment was U.S. \$68 million. From economic evaluation the value and cost DMF was 2.68 U.S. \$/kg and 1.95 U.S \$/kg, respectively. For this analysis, the plant is economically feasible for a scale factor of thirty

The following conclusions can be drawn from the facts presented in the above review. In the thermodynamic analysis salt is considered a solute, so it´s possible to use the model UNIQUAC (Mock et al., 1986). The estimation of binary interaction parameters for UNIQUAC in the system water-butanol-salt was carried out with Fortran software from liquid-liquid equilibrium data and UNIFAC (UNIQUAC Functional-group Activity Coefficient) method was used to estimate remain parameters. The separation system (composed by two stripping columns, one cooler and one settling tank) used to separate 1 butanol and water recovery 98% of 1-butanol. The separation system (composed by distillation column) used to separate DMF recovery 92 % of DMF. Economic evaluation showed that a suitable operational plant could work with 12.4 tons/year of fructose. It could produce 11.1 tons/year of DMF. The fixed capital investment in plant and equipment is estimated at U.S. \$ 58 million and U.S. \$ 12 million, respectively. The DMF value was 2.69 U.\$. \$/kg. For this analysis, the plant is economically feasible, from comparison with a reference market of 15.0 %/year (return on investment) with a 3.6 year payback period. This analysis suggests that DMF production from fructose deserves serious consideration by

Abrams, D.S. & Prausnitz, J.M. (1975). Statistical Thermodynamics of Liquid Mixtures: A

Aznar, M. (1996). *Equilíbrio Líquido-Vapor de Sistemas com Eletrólitos Via Contribuição de Grupo,*

Catté, M.; Dussap, C.G.; Achard, C. & Gros, J.B. (1994). Excess Properties and Solid-Liquid

Cezário, G. L.; M. Filho, R. & Mariano, A. P. (2010). Projeto e Avaliação Energética do

Debye, P. & Huckel, E. (1923). Zur Theorie der Elektrolyte. *Physics Zeitsch*, Vol.24, pp. 185-

Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil.

New Expression for the Excess Gibbs Energy of Partly or Completeley Miscible Systems. *American Institute of Chemical Engineers Journal*, Vol.21, pp. 116-128. Allen, D.H. (1980). *A Guide to Economic Evaluation of Projects*, The Institutions Chemical

Equilibria for Aqueous Solutions of Sugars Using a UNIQUAC Model. *Fluid Phase* 

Sistema de Destilação de uma Planta de Fermentação Extrativa para a Produção de Biobutanol, *Proceedings of XVII Congresso de Iniciação Científica da Universidade* 

Where N is the scale factor with values of N = 2, 5, 10, etc.

Engineers, Rugby, United Kingston.

*Estadual de Campinas*, Campinas, Brazil.

*Equilibria*, Vol.96, pp. 33-50.

(N= 30).

investors.

**6. References** 

206.

**5. Conclusions** 

In this work, inflation was account ed for by the Chemical Engineering Plant Cost Index (Lozowski, 2010). According the Table 8 the price of raw material and solvent are the more expensive.

With chemical and utility cost were obtained and a discounted cash flow analysis was performed to determine profitability. The quantities of chemical material, utilities and production of DMF were doubled, tripled, etc, from the simulated plant, to achieve sale price and cost DMF similar to gasoline and ethanol. However, the equipment cost increased according to Eq. (3). For all sale price and cost DMF from Tables 5, 6 e 7 the profitability measures were: 15.0 %/year (return on investment) and a 3.6 year payback period.

The conversion and the price fructose were changed too. The tables 10, 11 and 12 show the results.


Table 10. Sale price of DMF/cost (US \$/kg) of DMF to several conversions in reactor CRV-102 and various prices of fructose (standard plan used in the simulation).


Table 11. Sale price of DMF/cost (US \$/kg) of DMF to several conversions in reactor CRV-102 and various prices of fructose (scale factor 12).


Table 12. Sale price of DMF/cost (US \$/kg) of DMF to several conversions in reactor CRV-102 and various prices of fructose (scale factor 30).

In table 12 observes that the sale price of DMF can be compared with the gasoline. The cost of DMF decreases with the increase of the conversion of fructose to HMF and with the price decrease of fructose.

$$\text{Equipament cost} = \mathbf{N}^{0.6} \tag{3}$$

Where N is the scale factor with values of N = 2, 5, 10, etc.

Thus, the estimated cost of the equipment was U.S. \$ 12 million, fixed capital investment was U.S. \$58 million, direct cost were U.S. \$41 million, indirect costs were U.S. \$17 million, working capital was U.S. \$10 million and total capital investment was U.S. \$68 million. From economic evaluation the value and cost DMF was 2.68 U.S. \$/kg and 1.95 U.S \$/kg, respectively. For this analysis, the plant is economically feasible for a scale factor of thirty (N= 30).

#### **5. Conclusions**

446 Advances in Chemical Engineering

In this work, inflation was account ed for by the Chemical Engineering Plant Cost Index (Lozowski, 2010). According the Table 8 the price of raw material and solvent are the more

With chemical and utility cost were obtained and a discounted cash flow analysis was performed to determine profitability. The quantities of chemical material, utilities and production of DMF were doubled, tripled, etc, from the simulated plant, to achieve sale price and cost DMF similar to gasoline and ethanol. However, the equipment cost increased according to Eq. (3). For all sale price and cost DMF from Tables 5, 6 e 7 the profitability

The conversion and the price fructose were changed too. The tables 10, 11 and 12 show the

Conversion (75%) Conversion (80%) Conversion (85%)

Conversion (75%) Conversion (80%) Conversion (85%)

Conversion (75%) Conversion (80%) Conversion (85%)

Equipament cost = N0.6 (3)

measures were: 15.0 %/year (return on investment) and a 3.6 year payback period.

 Fructose price, 0.78 US\$/kg 6.00/3.94 5.90/3.83 5.80/3.75 Fructose price, 0.50 US\$/kg 5.70/3.59 5.60/3.49 5.40/3.42 Fructose price, 0.10 US\$/kg 5.20/3.10 5.00/3.00 4.90/2.94 Table 10. Sale price of DMF/cost (US \$/kg) of DMF to several conversions in reactor CRV-

 Fructose price, 0.78 US\$/kg 3.70/2.68 3.60/2.59 3.52/2.54 Fructose price, 0.50 US\$/kg 3.30/2.33 3.20/2.26 3.13/2.22 Fructose price, 0.10 US\$/kg 2.70/1.83 2.63/1.78 2.58/1.74 Table 11. Sale price of DMF/cost (US \$/kg) of DMF to several conversions in reactor CRV-

 Fructose price, 0.78 US\$/kg 2.68/1.95 2.60/1.89 2.55/1.85 Fructose price, 0.50 US\$/kg 2.27/1.60 2.20/1.56 2.15/1.52 Fructose price, 0.10 US\$/kg 1.68/1.10 1.63/1.07 1.60/1.05 Table 12. Sale price of DMF/cost (US \$/kg) of DMF to several conversions in reactor CRV-

In table 12 observes that the sale price of DMF can be compared with the gasoline. The cost of DMF decreases with the increase of the conversion of fructose to HMF and with the price

102 and various prices of fructose (standard plan used in the simulation).

102 and various prices of fructose (scale factor 12).

102 and various prices of fructose (scale factor 30).

decrease of fructose.

expensive.

results.

The following conclusions can be drawn from the facts presented in the above review. In the thermodynamic analysis salt is considered a solute, so it´s possible to use the model UNIQUAC (Mock et al., 1986). The estimation of binary interaction parameters for UNIQUAC in the system water-butanol-salt was carried out with Fortran software from liquid-liquid equilibrium data and UNIFAC (UNIQUAC Functional-group Activity Coefficient) method was used to estimate remain parameters. The separation system (composed by two stripping columns, one cooler and one settling tank) used to separate 1 butanol and water recovery 98% of 1-butanol. The separation system (composed by distillation column) used to separate DMF recovery 92 % of DMF. Economic evaluation showed that a suitable operational plant could work with 12.4 tons/year of fructose. It could produce 11.1 tons/year of DMF. The fixed capital investment in plant and equipment is estimated at U.S. \$ 58 million and U.S. \$ 12 million, respectively. The DMF value was 2.69 U.\$. \$/kg. For this analysis, the plant is economically feasible, from comparison with a reference market of 15.0 %/year (return on investment) with a 3.6 year payback period. This analysis suggests that DMF production from fructose deserves serious consideration by investors.

#### **6. References**


**18** 

*1Qatar 2Tunisia* 

**Inland Desalination: Potentials and Challenges** 

Groundwater is the main source of drinking water in many countries all over the world. In absence of surface water supply, the use of groundwater as the main water source for drinking, industrial, and agricultural use becomes essential especially in the case of rural communities. Underground reservoirs constitute a major source of fresh water, in terms of storage capacity; underground aquifers worldwide contain over 95% of the total fresh water available for human use. Typical groundwater supplies have low coliform counts and total bacterial counts, low turbidity, clear color, pleasant taste, and low odor. Accordingly, groundwater has higher quality than surface water, and the quality is quite uniform throughout the year that makes it easy to treat. A disadvantage of groundwater supplies is that many groundwater aquifers have moderate to high dissolved solids such as calcium, magnesium, iron, sulfate, sodium, chloride, and silica. The high concentration of dissolved solids particularly, sodium chloride, makes the water brackish and thus requires to be

With the growth of membrane science, reverse osmosis RO overtook multi stage distillation MSF as the leading desalination technology. In the last two decades, RO processes have advanced significantly, allowing new brackish groundwater desalination facilities to use RO technology much more economically than distillation. RO treatment plants use semipermeable membranes and pressure to separate salts from water. These systems typically use less energy than thermal distillation, leading to a reduction in overall

The reverse osmosis process enables now the massive production of water with a moderate cost, providing flexible solutions to different necessities within the fields of population supply, industry and agriculture. The great development of reverse osmosis (RO) technology has been a consequence of several factors such as reduction in energy consumption and membrane cost. Nevertheless, the major problem of RO desalination

**1. Introduction** 

desalination costs.

Corresponding Author

 \*

desalinated before its use for a certain purpose.

Khaled Elsaid1, Nasr Bensalah1,2,\* and Ahmed Abdel-Wahab1

*1Department of Chemical Engineering, Texas A&M University at Qatar,* 

*2Department of Chemistry, Faculty of Sciences of Gabes,* 

*Education City, Doha* 

*University of Gabes, Gabes* 


## **Inland Desalination: Potentials and Challenges**

Khaled Elsaid1, Nasr Bensalah1,2,\* and Ahmed Abdel-Wahab1

*1Department of Chemical Engineering, Texas A&M University at Qatar, Education City, Doha 2Department of Chemistry, Faculty of Sciences of Gabes, University of Gabes, Gabes 1Qatar 2Tunisia* 

#### **1. Introduction**

448 Advances in Chemical Engineering

Festel, G. W. (2008). Biofuels - Economic Aspects. *Chemical Engineering Technology*, Vol.31,

Fowler, R.H. & Guggenheim, E.A. (1949). *Statistical Thermodynamics,* Cambridge University

Fredenslund, A.A.; Gmehling, J. & Rasmussen, P. (1977). *Vapor-Liquid Using UNIFAC*,

Leshkov, Y.R.; Barrett, C. J.; Liu, Z.Y. & Dumesic, J.A. (2007). Production of Dimethylfuran

Mock, B.; Evans, L.B. & Chen, C.C. (1986). Thermodynamic Representation of Phase

Peters, M.S.; Timmerhaus, K.D. & West, R.E.W. (2003), *Plant Design and Economics for* 

Pitzer, K.S. (1973). Thermodynamics of Electrolytes I. Theoretical Basis and General

Qureshi, N.; Hughes, S.; Maddon, I.S. & Cotta, M.A. (2005). Energy-Efficient Recovery of

Renon, H. & Prausnitz, J. M. (1968). Local Compositions in Thermodynamics Excess

Turton, R.; Bailie, R.C.; Whiting, W.B. & Shauwitz, J.A. (2003), *Analysis, Synthesis, and Design* 

Santis, R.; Marrelli, L. & Muscetta, P.N. (1976a). Liquid-Liquid Equilibria in Water-Aliphatic

Santis, R.; Marrelli, L. & Muscetta, P.N. (1976b). Influence of Temperature on the Liquid-

Schaub, G. & Vetter, A. (2003). Biofuels for Automobiles - An Overview. *Chemical* 

Lozowski, D. (2010). Economic Indicators. *Chemical Engineering*, Vol.117, pp. 55.

*Engineers Journal*, Vol.32, pp. 1655-1664.

*Chemical Engineers*, McGraw-Hill, New York.

*and Biosystems Engineering*, Vol.27, pp. 215-222.

*of Chemical Processes,* Prentice Hall, New Jersey.

*of Chemical and Engineering Data*, Vol.21, pp. 324-327.

*Engineering Technology*, Vol.31, pp. 721-729. UNISim Honeywell (2007), http://www.honeywell.com.

Equation. *Journal Physics Chemical*, Vol.77, pp. 268-277.

for Liquid Fuels from Biomass-Derived Carbohydrates. *Nature*, Vol.447, pp. 982-

Equilibria of Mixed-Solvent Electrolyte Systems. *Association International of Chemical* 

Butanol from Model Solutions and Fermentation Broth by Adsorption. *Bioprocess* 

Functions for Liquid Mixtures. *American Institute of Chemical Engineers Journal*,

Alcohol Systems in the Presence of Sodium Chloride. *Chemical Engineering Journal*,

Liquid Equilibrium of the Water-n Butyl Alcohol-Sodium Chloride System. *Journal* 

pp. 715-720.

986.

Press, Cambridge.

Elsevier, Amsterdam.

Vol.14, pp. 135-144.

Vol.11, pp. 207-214.

Groundwater is the main source of drinking water in many countries all over the world. In absence of surface water supply, the use of groundwater as the main water source for drinking, industrial, and agricultural use becomes essential especially in the case of rural communities. Underground reservoirs constitute a major source of fresh water, in terms of storage capacity; underground aquifers worldwide contain over 95% of the total fresh water available for human use. Typical groundwater supplies have low coliform counts and total bacterial counts, low turbidity, clear color, pleasant taste, and low odor. Accordingly, groundwater has higher quality than surface water, and the quality is quite uniform throughout the year that makes it easy to treat. A disadvantage of groundwater supplies is that many groundwater aquifers have moderate to high dissolved solids such as calcium, magnesium, iron, sulfate, sodium, chloride, and silica. The high concentration of dissolved solids particularly, sodium chloride, makes the water brackish and thus requires to be desalinated before its use for a certain purpose.

With the growth of membrane science, reverse osmosis RO overtook multi stage distillation MSF as the leading desalination technology. In the last two decades, RO processes have advanced significantly, allowing new brackish groundwater desalination facilities to use RO technology much more economically than distillation. RO treatment plants use semipermeable membranes and pressure to separate salts from water. These systems typically use less energy than thermal distillation, leading to a reduction in overall desalination costs.

The reverse osmosis process enables now the massive production of water with a moderate cost, providing flexible solutions to different necessities within the fields of population supply, industry and agriculture. The great development of reverse osmosis (RO) technology has been a consequence of several factors such as reduction in energy consumption and membrane cost. Nevertheless, the major problem of RO desalination

<sup>\*</sup> Corresponding Author

Inland Desalination: Potentials and Challenges 451

AQUASTAT program (Food and Agriculture Organization [FAO]) was launched to form global information system on water and agriculture, the main objective of the program is to collect and analyze information on water resources, water uses, and agriculture water management within different countries. Information on the quantity of major water

(1000 km3)

Oceans/seas 1,338,000 96.54 - Saline/brackish groundwater 12,870 0.93 - Saltwater lakes 85 0.006 -

Glaciers and permanent snow covers 24,064 1.74 68.70 Fresh groundwater 10,530 0.76 30.06 Fresh lakes 91 0.007 0.26 Wetlands 11.5 0.001 0.03 Rivers 2.12 0.0001 0.006

Freshwater is the water naturally found on Earth's surface and in underground aquifers such as surface water, fresh groundwater, and glaciers, and mainly characterized by its low content of dissolved solids. These water sources are considered to be renewable resource, by effect of natural water cycle. The quantity of freshwater present on Earth is around 2.5%

Surface water is the water present in rivers, fresh lakes, and wetlands; the main source of surface water is by precipitation in the form of rain, snow….etc. Surface water is characterized by low content of dissolved salts generally below 500 mg/L. Surface water represents only around 0.3 % of the total freshwater present on Earth's surface. Fresh groundwater is the water located under the Earth's surface i.e. subsurface water which is mainly located in pores or spaces of soil and rocks, or in aquifers below the water table. It is mainly characterized by its low suspended solids. In many places groundwater contains high content of dissolved salts compared to that of surface water; with salinity level around 500-2,000 mg/L (Mickley, 2001). Groundwater represents around 0.76 % of the total water

Water or ice present in glaciers, icebergs, and icecaps represents the vast majority freshwater, this huge amount of water is currently unused and locked up in southern and northern poles. Up to date there is no efforts has been made to make use of such water resources due to the high cost associated with its processing as it is mainly present in very

Brackish groundwater is the water located under the Earth's surface and it is characterized by its higher salinity than that of fresh groundwater with values of 2,000-10,000 mg/L

Percent of total water

Percent of total fresh water

Water Resource Volume,

Table 1. Estimates of major water resources on Earth (Gleick, 2001).

present on Earth, and around 30 % of the freshwater available on Earth.

only of the total water present on Earth.

distant areas or at very high altitudes.

resources is present in table 2.1.

Saline water:

Freshwater:

plants is the generation of a concentrate effluent (brine) that must be properly managed. Disposal of such brines presents significant costs and challenges for the desalination industry due to high cost and environmental impact of brine disposal.

The reject brine from desalination plants not only contains various types of salts at higher concentration but also other types of wastes like pretreatment chemicals (antiscalents, antifouling,…etc). Also, if the feed water includes harmful chemicals such as heavy metals or others, these chemicals are concentrated in the reject brine. Improper disposal of reject brine from inland plants results in several environmental problems.

Although sea disposal of reject brine is a common practice for plants located in coastal areas, it would not be available for inland desalination. Deep well injection is prohibitly expensive and has its own problems such as the possibility of corrosion and subsequent leakage in the well casing, seismic activity which could cause damage to the well and subsequently result in ground water contamination, and uncertainty of the well life. Additionally, when a sewerage system is used for disposal of concentrate brine high in total dissolved solids (TDS) the treated municipal sewage effluent becomes unsuitable for reuse.

While operation and maintenance costs for evaporation ponds are minimal, large land areas are required, and pond construction costs are high. Even in arid climates ideally suited for evaporation, a typical design application rate is only 2 gpm per acre. The construction cost for an evaporation pond with a liner and monitoring system typically ranges between \$100,000 and \$200,000 per acre, exclusive of land cost. Thus a concentrate flow as small as 100 gpm would require a pond area of at least 50 acres and cost \$10 to \$20 million to construct. Consequently, evaporation ponds are often cost prohibitive and impractical for handling any significant concentrate flow. Furthermore, water evaporated from a pond is often a lost resource.

Therefore, the need to protect surface and groundwater resources may in many cases preclude concentrate disposal by the earlier three methods. The alternative is zero liquid discharge (ZLD). In ZLD, concentrate is treated to produce desalinated water and essentially dry salts. Consequently there is no discharge of liquid waste from the process. Most ZLD applications in operation today treat industrial wastewater or power plant cooling water using thermal crystallization, evaporation ponds, or a combination of these technologies. Thermal crystallization is energy-intensive with high capital and operating costs.

Given the need for ZLD and the disadvantages of existing ZLD methods, it is imperative to find alternative ZLD treatment technologies that provide more affordable concentrate management. This article reviews current trends and potential advancements of inland desalination and brine management alternatives.

#### **2. Water and water resources**

Water resources present naturally in the environment can be generally divided into freshwater and saline water according to the amount of dissolved solids it contains. Quality and quantity of different water resources are of high importance, many efforts are being made to have good estimates of water resources at both worldwide and country levels. In 1990s "The comprehensive assessment of the freshwater resources of the world" was launched to have estimates on worldwide water resources (United Nations [UN], 1997). The

plants is the generation of a concentrate effluent (brine) that must be properly managed. Disposal of such brines presents significant costs and challenges for the desalination

The reject brine from desalination plants not only contains various types of salts at higher concentration but also other types of wastes like pretreatment chemicals (antiscalents, antifouling,…etc). Also, if the feed water includes harmful chemicals such as heavy metals or others, these chemicals are concentrated in the reject brine. Improper disposal of reject

Although sea disposal of reject brine is a common practice for plants located in coastal areas, it would not be available for inland desalination. Deep well injection is prohibitly expensive and has its own problems such as the possibility of corrosion and subsequent leakage in the well casing, seismic activity which could cause damage to the well and subsequently result in ground water contamination, and uncertainty of the well life. Additionally, when a sewerage system is used for disposal of concentrate brine high in total dissolved solids

While operation and maintenance costs for evaporation ponds are minimal, large land areas are required, and pond construction costs are high. Even in arid climates ideally suited for evaporation, a typical design application rate is only 2 gpm per acre. The construction cost for an evaporation pond with a liner and monitoring system typically ranges between \$100,000 and \$200,000 per acre, exclusive of land cost. Thus a concentrate flow as small as 100 gpm would require a pond area of at least 50 acres and cost \$10 to \$20 million to construct. Consequently, evaporation ponds are often cost prohibitive and impractical for handling any significant concentrate flow. Furthermore, water evaporated from a pond is

Therefore, the need to protect surface and groundwater resources may in many cases preclude concentrate disposal by the earlier three methods. The alternative is zero liquid discharge (ZLD). In ZLD, concentrate is treated to produce desalinated water and essentially dry salts. Consequently there is no discharge of liquid waste from the process. Most ZLD applications in operation today treat industrial wastewater or power plant cooling water using thermal crystallization, evaporation ponds, or a combination of these technologies.

Given the need for ZLD and the disadvantages of existing ZLD methods, it is imperative to find alternative ZLD treatment technologies that provide more affordable concentrate management. This article reviews current trends and potential advancements of inland

Water resources present naturally in the environment can be generally divided into freshwater and saline water according to the amount of dissolved solids it contains. Quality and quantity of different water resources are of high importance, many efforts are being made to have good estimates of water resources at both worldwide and country levels. In 1990s "The comprehensive assessment of the freshwater resources of the world" was launched to have estimates on worldwide water resources (United Nations [UN], 1997). The

Thermal crystallization is energy-intensive with high capital and operating costs.

desalination and brine management alternatives.

**2. Water and water resources** 

industry due to high cost and environmental impact of brine disposal.

brine from inland plants results in several environmental problems.

(TDS) the treated municipal sewage effluent becomes unsuitable for reuse.

often a lost resource.

AQUASTAT program (Food and Agriculture Organization [FAO]) was launched to form global information system on water and agriculture, the main objective of the program is to collect and analyze information on water resources, water uses, and agriculture water management within different countries. Information on the quantity of major water resources is present in table 2.1.


Table 1. Estimates of major water resources on Earth (Gleick, 2001).

Freshwater is the water naturally found on Earth's surface and in underground aquifers such as surface water, fresh groundwater, and glaciers, and mainly characterized by its low content of dissolved solids. These water sources are considered to be renewable resource, by effect of natural water cycle. The quantity of freshwater present on Earth is around 2.5% only of the total water present on Earth.

Surface water is the water present in rivers, fresh lakes, and wetlands; the main source of surface water is by precipitation in the form of rain, snow….etc. Surface water is characterized by low content of dissolved salts generally below 500 mg/L. Surface water represents only around 0.3 % of the total freshwater present on Earth's surface. Fresh groundwater is the water located under the Earth's surface i.e. subsurface water which is mainly located in pores or spaces of soil and rocks, or in aquifers below the water table. It is mainly characterized by its low suspended solids. In many places groundwater contains high content of dissolved salts compared to that of surface water; with salinity level around 500-2,000 mg/L (Mickley, 2001). Groundwater represents around 0.76 % of the total water present on Earth, and around 30 % of the freshwater available on Earth.

Water or ice present in glaciers, icebergs, and icecaps represents the vast majority freshwater, this huge amount of water is currently unused and locked up in southern and northern poles. Up to date there is no efforts has been made to make use of such water resources due to the high cost associated with its processing as it is mainly present in very distant areas or at very high altitudes.

Brackish groundwater is the water located under the Earth's surface and it is characterized by its higher salinity than that of fresh groundwater with values of 2,000-10,000 mg/L

Inland Desalination: Potentials and Challenges 453

Fig. 2. Worldwide groundwater resources mapping (The Federal Institute for Geosciences

Pure water is a colorless, odorless, and tasteless liquid, water is polar and strong solvent capable of dissolving many natural and synthetic substances, many inorganic and organic compounds, in addition it is able to suspend many solids, and hence it is very hard to find

The quality of groundwater is mainly determined by its content of dissolved solids and gases, presence of suspended solids, and bacteria present. Usually the nature and concentration of dissolved solids present in groundwater source will depend on characteristics of the aquifer and on travelling time or velocity of groundwater flow through

The physical properties of water are mainly, the total, suspended, and dissolved solids, in addition to turbidity, temperature, color, odor and taste. Typical groundwaters, as they undergo natural filtration while passing through sand formations have very low suspended solids content, low turbidity, clear color, pleasant taste, and low odor. However, while water is travelling through the soil formations, groundwater may carry dissolved solids if the soil formations are relatively soluble. The type and concentration of dissolved solids released from soil to groundwater may vary based on the soil composition, travelling time and flow

Chemical characteristics are mainly concerned with pH value, cations and anions, alkalinity, acidity, hardness, dissolved gases, and other contaminants such as organic substances and heavy metals that might be present in water. Groundwaters in general have higher hardness

and Natural Resources).

pure water in nature.

velocities.

**3. Groundwater quality** 

the rock formation (Delleur, 2007).

(Mickley, 2001). It is mainly present in aquifers that are much deeper than that of fresh groundwater. Brackish groundwater represents around 0.93 % of the total water present on Earth.

Saline or salty water is the water that contains considerable amount of salts and it is mainly found in oceans, seas, saline or brackish groundwater, and saltwater lakes. Saline water represents the majority of water resources in terms of quantity with around 97.5 % of the total water present on Earth. the salinity of seawater varying from one location to another, from around 21 g/L in the North Sea to 40 - 45 g/L in the Arabian Gulf and Red Sea, and even up to 300 g/L as in the Dead Sea (Gleick, 2006).

The majority of world population use surface water or groundwater as the main source for domestic, agriculture, and industrial water supplies. The most common surface water sources are rivers, and lakes. However, the most common groundwater sources are pumped wells or flowing artesian wells. In absence of surface water supply, it is clear that the use of groundwater becomes essential especially in the case of rural communities. Underground reservoirs constitute a major source of fresh water, in terms of storage capacity; underground aquifers worldwide contain over 95% of the total fresh water available for human use. In addition when looking to the map of worldwide water stress in figure 2.2, we find that areas that face water stresses are increasing with time with Middle East, North Africa, and Central Asia having the highest water stresses, while when looking to the worldwide groundwater resources map as shown in figure 2.3, we find that most of these areas have access to groundwater resources, which means that groundwater will present the relief to the faced water stress problems, and even can support the different developmental planes of these areas.

Fig. 1. Worldwide water stresses map (United nations Environmental Programme [UNEP]).

(Mickley, 2001). It is mainly present in aquifers that are much deeper than that of fresh groundwater. Brackish groundwater represents around 0.93 % of the total water present on

Saline or salty water is the water that contains considerable amount of salts and it is mainly found in oceans, seas, saline or brackish groundwater, and saltwater lakes. Saline water represents the majority of water resources in terms of quantity with around 97.5 % of the total water present on Earth. the salinity of seawater varying from one location to another, from around 21 g/L in the North Sea to 40 - 45 g/L in the Arabian Gulf and Red Sea, and

The majority of world population use surface water or groundwater as the main source for domestic, agriculture, and industrial water supplies. The most common surface water sources are rivers, and lakes. However, the most common groundwater sources are pumped wells or flowing artesian wells. In absence of surface water supply, it is clear that the use of groundwater becomes essential especially in the case of rural communities. Underground reservoirs constitute a major source of fresh water, in terms of storage capacity; underground aquifers worldwide contain over 95% of the total fresh water available for human use. In addition when looking to the map of worldwide water stress in figure 2.2, we find that areas that face water stresses are increasing with time with Middle East, North Africa, and Central Asia having the highest water stresses, while when looking to the worldwide groundwater resources map as shown in figure 2.3, we find that most of these areas have access to groundwater resources, which means that groundwater will present the relief to the faced water stress problems, and even can support the different developmental

Fig. 1. Worldwide water stresses map (United nations Environmental Programme [UNEP]).

even up to 300 g/L as in the Dead Sea (Gleick, 2006).

Earth.

planes of these areas.

Fig. 2. Worldwide groundwater resources mapping (The Federal Institute for Geosciences and Natural Resources).

#### **3. Groundwater quality**

Pure water is a colorless, odorless, and tasteless liquid, water is polar and strong solvent capable of dissolving many natural and synthetic substances, many inorganic and organic compounds, in addition it is able to suspend many solids, and hence it is very hard to find pure water in nature.

The quality of groundwater is mainly determined by its content of dissolved solids and gases, presence of suspended solids, and bacteria present. Usually the nature and concentration of dissolved solids present in groundwater source will depend on characteristics of the aquifer and on travelling time or velocity of groundwater flow through the rock formation (Delleur, 2007).

The physical properties of water are mainly, the total, suspended, and dissolved solids, in addition to turbidity, temperature, color, odor and taste. Typical groundwaters, as they undergo natural filtration while passing through sand formations have very low suspended solids content, low turbidity, clear color, pleasant taste, and low odor. However, while water is travelling through the soil formations, groundwater may carry dissolved solids if the soil formations are relatively soluble. The type and concentration of dissolved solids released from soil to groundwater may vary based on the soil composition, travelling time and flow velocities.

Chemical characteristics are mainly concerned with pH value, cations and anions, alkalinity, acidity, hardness, dissolved gases, and other contaminants such as organic substances and heavy metals that might be present in water. Groundwaters in general have higher hardness

Inland Desalination: Potentials and Challenges 455

of 880,000 m3/d at Shoaiba 3 project in Saudi Arabia (International Desalination Association [IDA], 2009). By 2009 there were 14,451 desalination plants online, with further 244 known to be on their way, being under contract or under construction, with an additional capacity of 9.1 million m3/d in 130 countries around the world (IDA, 2009). Desalination in general can be mainly classified according to the feed water source into two main classes as seawater desalination and brackish groundwater desalination. In 2006 around 56 % of the world desalination capacity was for seawater desalination, and 24 % for brackish water, with Saudi Arabia KSA, Unite states US, United Arab Emirates UAE, Spain, Kuwait, and Japan having desalination capacity over 1 million m3/d, with Middle East countries holding

Desalination processes are many and can be generally classified according to the technology used and it mainly classified to thermal desalination, and membrane desalination, ion exchange, and electrodialysis. In addition, other new technologies are still under development. The common processes for desalination have been changed from expensive techniques with extensive energy requirements to a sustainable method for drinking water

The selection of desalination process for a certain purpose depends on many factors such as energy cost, final water quality, fouling propensity, temperature, and overall process cost. Table 2 shows the worldwide, US, and Gulf Corporation Council GCC desalination capacity

**Process Type Worldwide United States GCC Countries** 

Table 2. Desalination capacity percentage distribution according to process type (Gleick, 2006).

The quality of desalinated water differs depending on the process used, with thermal desalination producing very high quality product water with salinity about 2-10 mg/L; which usually requires remineralization in the post treatment step (Gabbrielli, 1981). However for membrane process the quality of product water depends on many factors such as the quality of feedwater, design recovery, and membrane properties, but in general the product water will have higher salinity than that produced from thermal desalination.

The quality of the final water product mainly depends on the application in which it is to be used for, ranging from very high quality for process water to specified quality as per the regulation for drinking water, or to a certain quality suitable for agriculture. Most of the desalination plants worldwide are designed to be able to produce high quality water. Drinking water standards vary slightly from country to another. The most widely used is the United States Environmental Protection Agency US-EPA drinking water standards with

**RO** 46 69 28 **MSF** 36 1 54 **ED** 5 9 - **VC** 5 3 - **MED** 3 1 9 **NF** - 15 - **Others** 5 2 9

over 50 % of the worldwide desalination capacity (Gleick, 2006).

supply.

by processes.

when compared to surface waters; this is mainly due to the dissolution of limestone and dolomite formations which in turn increase the content of calcium ions and hence increasing the hardness of water.

Biological characteristics are concerned with the living organisms present in water, mainly bacteria, fungi, algae, and viruses. Microbes are generally absent from groundwater due to natural filtration, and hence groundwater has low coliform and total bacterial counts.

In general groundwater has higher quality than surface waters, and the quality is quite uniform throughout the year making it easy to treat. A disadvantage of groundwater is that many have moderate to high dissolved solids. The high concentration of dissolved solids, particularly sodium chloride, makes the water brackish. The removal of such dissolved solids requires use of desalination process to treat the water to a level of dissolved solids that makes it suitable for a certain use.

Technical and economical evaluations of desalinating groundwater has started early in 1960s. One of the first major brackish groundwater desalination plants was built in Florida, USA with capacity of 2.62 mgd, followed by another one using Reverse Osmosis RO in mid-1970s with capacity of 0.5 mgd (Bart Weiss, 2002). A combination of both field investigations and computer simulation modeling were used to assess the economic suitability of using highly brackish groundwater for large scale abstraction for feeding reverse osmosis desalination plant in many countries around the world (Brich et al., 1985; Hadi, 2002; Sherif et al., 2011; Zubari, 2003). The results from such technical and economical evaluations have shown that brackish groundwater can be considered as high quality source of feedwater for desalination plants, even at higher salinities, which still far below the salinity of seawater. However the evaluations also indicated that the groundwater quality is changing and not constant over long years with continuous increase in dissolved solids content, which should be considered during design stages of such desalination plants.

It was concluded that the major misconception of considering that the groundwater quality is relatively stable over time is very critical for design purposes. The groundwater quality tends to show increase in salinity with time, and hence the initial and future groundwater qualities should be considered (Missimer, 1994). The deteriorating quality of groundwater is becoming of high concern globally, mainly due to human activities such as over abstraction and seawater intrusion. The seawater intrusion is a very common problem worldwide especially, near coastal areas, it is a natural phenomenon in which saline water from sea or ocean moves into the fresh groundwater in coastal aquifers. This behavior is mainly attributed to the density difference and to tidal effects. Seawater intrusion was found to be the main source for the increased salinity of near-coast aquifers in many places worldwide (Amer et al., 2008; Zubari et al., 1994).

#### **4. Inland desalination processes**

Desalination is a water treatment process for removing total dissolved solids (TDS) from water. Desalination of seawater and brackish water has become a reliable method for water supply all-over the world and had been practiced successfully for many decades.

The worldwide desalination capacity increased dramatically from around 35 million m3/d in 2005 (Gleick, 2006) to about 60 million m3/d by 2009, with the largest desalination plant

when compared to surface waters; this is mainly due to the dissolution of limestone and dolomite formations which in turn increase the content of calcium ions and hence increasing

Biological characteristics are concerned with the living organisms present in water, mainly bacteria, fungi, algae, and viruses. Microbes are generally absent from groundwater due to natural filtration, and hence groundwater has low coliform and total bacterial counts.

In general groundwater has higher quality than surface waters, and the quality is quite uniform throughout the year making it easy to treat. A disadvantage of groundwater is that many have moderate to high dissolved solids. The high concentration of dissolved solids, particularly sodium chloride, makes the water brackish. The removal of such dissolved solids requires use of desalination process to treat the water to a level of dissolved solids

Technical and economical evaluations of desalinating groundwater has started early in 1960s. One of the first major brackish groundwater desalination plants was built in Florida, USA with capacity of 2.62 mgd, followed by another one using Reverse Osmosis RO in mid-1970s with capacity of 0.5 mgd (Bart Weiss, 2002). A combination of both field investigations and computer simulation modeling were used to assess the economic suitability of using highly brackish groundwater for large scale abstraction for feeding reverse osmosis desalination plant in many countries around the world (Brich et al., 1985; Hadi, 2002; Sherif et al., 2011; Zubari, 2003). The results from such technical and economical evaluations have shown that brackish groundwater can be considered as high quality source of feedwater for desalination plants, even at higher salinities, which still far below the salinity of seawater. However the evaluations also indicated that the groundwater quality is changing and not constant over long years with continuous increase in dissolved solids content, which should

It was concluded that the major misconception of considering that the groundwater quality is relatively stable over time is very critical for design purposes. The groundwater quality tends to show increase in salinity with time, and hence the initial and future groundwater qualities should be considered (Missimer, 1994). The deteriorating quality of groundwater is becoming of high concern globally, mainly due to human activities such as over abstraction and seawater intrusion. The seawater intrusion is a very common problem worldwide especially, near coastal areas, it is a natural phenomenon in which saline water from sea or ocean moves into the fresh groundwater in coastal aquifers. This behavior is mainly attributed to the density difference and to tidal effects. Seawater intrusion was found to be the main source for the increased salinity of near-coast aquifers in many places worldwide

Desalination is a water treatment process for removing total dissolved solids (TDS) from water. Desalination of seawater and brackish water has become a reliable method for water

The worldwide desalination capacity increased dramatically from around 35 million m3/d in 2005 (Gleick, 2006) to about 60 million m3/d by 2009, with the largest desalination plant

supply all-over the world and had been practiced successfully for many decades.

the hardness of water.

that makes it suitable for a certain use.

(Amer et al., 2008; Zubari et al., 1994).

**4. Inland desalination processes** 

be considered during design stages of such desalination plants.

of 880,000 m3/d at Shoaiba 3 project in Saudi Arabia (International Desalination Association [IDA], 2009). By 2009 there were 14,451 desalination plants online, with further 244 known to be on their way, being under contract or under construction, with an additional capacity of 9.1 million m3/d in 130 countries around the world (IDA, 2009). Desalination in general can be mainly classified according to the feed water source into two main classes as seawater desalination and brackish groundwater desalination. In 2006 around 56 % of the world desalination capacity was for seawater desalination, and 24 % for brackish water, with Saudi Arabia KSA, Unite states US, United Arab Emirates UAE, Spain, Kuwait, and Japan having desalination capacity over 1 million m3/d, with Middle East countries holding over 50 % of the worldwide desalination capacity (Gleick, 2006).

Desalination processes are many and can be generally classified according to the technology used and it mainly classified to thermal desalination, and membrane desalination, ion exchange, and electrodialysis. In addition, other new technologies are still under development. The common processes for desalination have been changed from expensive techniques with extensive energy requirements to a sustainable method for drinking water supply.

The selection of desalination process for a certain purpose depends on many factors such as energy cost, final water quality, fouling propensity, temperature, and overall process cost. Table 2 shows the worldwide, US, and Gulf Corporation Council GCC desalination capacity by processes.


Table 2. Desalination capacity percentage distribution according to process type (Gleick, 2006).

The quality of desalinated water differs depending on the process used, with thermal desalination producing very high quality product water with salinity about 2-10 mg/L; which usually requires remineralization in the post treatment step (Gabbrielli, 1981). However for membrane process the quality of product water depends on many factors such as the quality of feedwater, design recovery, and membrane properties, but in general the product water will have higher salinity than that produced from thermal desalination.

The quality of the final water product mainly depends on the application in which it is to be used for, ranging from very high quality for process water to specified quality as per the regulation for drinking water, or to a certain quality suitable for agriculture. Most of the desalination plants worldwide are designed to be able to produce high quality water. Drinking water standards vary slightly from country to another. The most widely used is the United States Environmental Protection Agency US-EPA drinking water standards with

Inland Desalination: Potentials and Challenges 457

Membrane processes are basic in principles, where semi-permeable membrane is used to allow the passage of water but not the salt to under certain driving forces. Different types of membrane processes include reverse osmosis RO, nanofiltration, and forward osmosis FO;

Osmosis is a natural physical process, where the solvent (i.e. water) moves through semipermeable membrane (i.e. permeable to solvent and impermeable to solute) from low solute concentration to higher concentration creating differential pressure called osmotic pressure. The osmotic pressure depends mainly on the concentration difference, temperature, and nature of solute). The process continues till the hydraulic pressure difference due to the liquid column is equal to the osmotic pressure. In reverse osmosis RO, hydraulic pressure in value greater than the osmotic pressure is applied to the concentrated solution, results in reversing the osmosis process and in net solvent or water flow from the concentrate to the

RO membranes do not have a distinct pore that traverse the membrane; it consists of polymeric material forming a layered, web-like structure characterized by high rejection to most of the dissolved solids present in water, with typical salt rejection above 99 % (Lee et al., 2011). The driving force for RO process is the applied hydraulic pressure which varies considerably from

The Nanofiltration NF term was first introduced by FilmTech in 1980s to describe RO membranes that allow selectively and purposely some ionic solutes to path through the membrane, the membrane's selectivity was towards solute of about 1 nm cutoff, and hence the term nano comes from (Wang, 2008). Nanofiltration is an intermediate between RO membranes which has a low Molecular Weight Cut Off MWCO of about 100 and

NF membranes have higher permeability for monovalent ions such as Na, K, and Cl and very low permeability to multivalent ions such as Ca, Mg, SO4 and organics with MWCO of around 300 (Rautenbach & Groschl, 1990), as a result NF membranes were used mainly for removal of hardness and natural organic matter, as pretreatment before RO and MSD for seawater desalination (Al-shammiri et al., 2004; Hassan et al., 1998), and for groundwater quality enhancement (Burggen & Vandercasteele, 2003; Gorenflo et al., 2002; Saitua et al., 2011; Tahaikt et al., 2007). Nanofiltration membranes operating in a similar fashion to that of RO membranes, except of lower driving pressure and hence lower energy requirements,

Electrodialysis/Electrodialysis Reversal membrane desalination (ED/EDR) is one of the oldest tried desalination processes, with ED process since the 1950s. In ED/EDR electrochemical separation is the main phenomena takes place utilizing electrical power as driving force to separate ions through ion-exchange membranes (Gleick, 2001). In typical ED

15-25 bar for brackish water, and 60-80 bar for seawater (Fritzmann et al., 2007).

ultrafiltration membranes which has MWCO of about 1000 (Eriksson, 1988).

higher flowrates, and lower product water quality (Schaep et al., 1998).

**4.2.3 Electrodialysis/ electrodialysis reversal membrane desalination** 

the later is still under development mainly at laboratory and pilot scales.

**4.2.1 Reverse osmosis RO membrane desalination** 

dilute solution (Fritzmann et al., 2007).

**4.2.2 Nanofiltration NF membrane desalination** 

500 mg/L for total dissolved solids, and 250 mg/L for each of sulfate and chloride(US EPA, 2009). However the World Health Organization WHO guidelines for drinking water quality suggest that total dissolved solids value of 600-1,000 mg/L is generally acceptable (WHO, 2011). On the other hand the FAO suggests that a total dissolved solids up to 2,000 mg/L is acceptable for irrigation purposes (Ayers & Westcot, 1994).

#### **4.1 Thermal desalination processes**

In desalting operations, thermal technologies were the only viable option for long time, and Multi-Stage Flash Distillation (MSF) was established as the baseline technology; however Multiple-Effect Distillation (MED) is now the state-of-the-art thermal technology, but has not been widely implemented yet. Thermal desalination plants have provided the major portion of the world's desalination capacity, as the world's requirements for treated water increase. Thermal desalination is usually used for cases where high salinity feed waters are used i.e. seawater, high recoveries are required, high feedwater temperature, and low energy cost, however the main drawback is the extensive energy consumption (Greenlee et al., 2009).

Thermal desalination or some time called phase change desalination; is a very basic process in concept as it copies the natural process of water cycle, where energy in thermal form through the solar radiation evaporates the water into water vapor, which is condensed later and fall in the form of rains or snow (Gleick, 2001), so in thermal desalination thermal energy or heat is applied to the water present in boiler or evaporator to drive water evaporation, this water vapor is condensed later in condenser by exchanging heat, thus sometimes the process is called phase change desalination as water phase changes from liquid into vapor is encountered. Even though the basic concept is the same however there are many processes which utilize that concept in application today, the main thermal desalination technologies are Multi-Stage Flash Distillation (MSF), Multiple-Effect Distillation (MED), and Vapor Compression (VC).

#### **4.2 Membrane desalination processes**

With the growth of membrane science, RO overtook MSF as the leading desalination technology, membrane desalination processes in general and commercial RO processes in particular have been undergoing appreciable development. Important factors in the expansion of commercial RO applications are their favorably low power requirements and the realization of continuous technical improvements in membranes which are used in RO systems, RO was first applied to brackish groundwater with first large scale plants in late 1960s. A decade later RO membrane after further development was suitable for seawater desalination and become a strong competitive to conventional thermal desalination by the 1980s (Vandercaseele & Bruggen, 2002) and hence was able to expand the water sources used for desalination and utilize the brackish groundwater and dominate its market.

In the last two decades, RO processes have undergone major advancements significantly enabling now the massive production of water with a moderate cost, providing flexible solutions to different necessities within the fields of population supply, industry and agriculture. The great development of reverse osmosis (RO) technology has been a consequence of several factors such as energy consumption reduction, improvement of membrane material, and decrease in membrane cost.

500 mg/L for total dissolved solids, and 250 mg/L for each of sulfate and chloride(US EPA, 2009). However the World Health Organization WHO guidelines for drinking water quality suggest that total dissolved solids value of 600-1,000 mg/L is generally acceptable (WHO, 2011). On the other hand the FAO suggests that a total dissolved solids up to 2,000 mg/L is

In desalting operations, thermal technologies were the only viable option for long time, and Multi-Stage Flash Distillation (MSF) was established as the baseline technology; however Multiple-Effect Distillation (MED) is now the state-of-the-art thermal technology, but has not been widely implemented yet. Thermal desalination plants have provided the major portion of the world's desalination capacity, as the world's requirements for treated water increase. Thermal desalination is usually used for cases where high salinity feed waters are used i.e. seawater, high recoveries are required, high feedwater temperature, and low energy cost, however the main drawback is the extensive energy consumption (Greenlee et al., 2009).

Thermal desalination or some time called phase change desalination; is a very basic process in concept as it copies the natural process of water cycle, where energy in thermal form through the solar radiation evaporates the water into water vapor, which is condensed later and fall in the form of rains or snow (Gleick, 2001), so in thermal desalination thermal energy or heat is applied to the water present in boiler or evaporator to drive water evaporation, this water vapor is condensed later in condenser by exchanging heat, thus sometimes the process is called phase change desalination as water phase changes from liquid into vapor is encountered. Even though the basic concept is the same however there are many processes which utilize that concept in application today, the main thermal desalination technologies are Multi-Stage Flash Distillation (MSF), Multiple-Effect

With the growth of membrane science, RO overtook MSF as the leading desalination technology, membrane desalination processes in general and commercial RO processes in particular have been undergoing appreciable development. Important factors in the expansion of commercial RO applications are their favorably low power requirements and the realization of continuous technical improvements in membranes which are used in RO systems, RO was first applied to brackish groundwater with first large scale plants in late 1960s. A decade later RO membrane after further development was suitable for seawater desalination and become a strong competitive to conventional thermal desalination by the 1980s (Vandercaseele & Bruggen, 2002) and hence was able to expand the water sources

used for desalination and utilize the brackish groundwater and dominate its market.

In the last two decades, RO processes have undergone major advancements significantly enabling now the massive production of water with a moderate cost, providing flexible solutions to different necessities within the fields of population supply, industry and agriculture. The great development of reverse osmosis (RO) technology has been a consequence of several factors such as energy consumption reduction, improvement of

acceptable for irrigation purposes (Ayers & Westcot, 1994).

Distillation (MED), and Vapor Compression (VC).

membrane material, and decrease in membrane cost.

**4.2 Membrane desalination processes** 

**4.1 Thermal desalination processes** 

Membrane processes are basic in principles, where semi-permeable membrane is used to allow the passage of water but not the salt to under certain driving forces. Different types of membrane processes include reverse osmosis RO, nanofiltration, and forward osmosis FO; the later is still under development mainly at laboratory and pilot scales.

#### **4.2.1 Reverse osmosis RO membrane desalination**

Osmosis is a natural physical process, where the solvent (i.e. water) moves through semipermeable membrane (i.e. permeable to solvent and impermeable to solute) from low solute concentration to higher concentration creating differential pressure called osmotic pressure. The osmotic pressure depends mainly on the concentration difference, temperature, and nature of solute). The process continues till the hydraulic pressure difference due to the liquid column is equal to the osmotic pressure. In reverse osmosis RO, hydraulic pressure in value greater than the osmotic pressure is applied to the concentrated solution, results in reversing the osmosis process and in net solvent or water flow from the concentrate to the dilute solution (Fritzmann et al., 2007).

RO membranes do not have a distinct pore that traverse the membrane; it consists of polymeric material forming a layered, web-like structure characterized by high rejection to most of the dissolved solids present in water, with typical salt rejection above 99 % (Lee et al., 2011). The driving force for RO process is the applied hydraulic pressure which varies considerably from 15-25 bar for brackish water, and 60-80 bar for seawater (Fritzmann et al., 2007).

#### **4.2.2 Nanofiltration NF membrane desalination**

The Nanofiltration NF term was first introduced by FilmTech in 1980s to describe RO membranes that allow selectively and purposely some ionic solutes to path through the membrane, the membrane's selectivity was towards solute of about 1 nm cutoff, and hence the term nano comes from (Wang, 2008). Nanofiltration is an intermediate between RO membranes which has a low Molecular Weight Cut Off MWCO of about 100 and ultrafiltration membranes which has MWCO of about 1000 (Eriksson, 1988).

NF membranes have higher permeability for monovalent ions such as Na, K, and Cl and very low permeability to multivalent ions such as Ca, Mg, SO4 and organics with MWCO of around 300 (Rautenbach & Groschl, 1990), as a result NF membranes were used mainly for removal of hardness and natural organic matter, as pretreatment before RO and MSD for seawater desalination (Al-shammiri et al., 2004; Hassan et al., 1998), and for groundwater quality enhancement (Burggen & Vandercasteele, 2003; Gorenflo et al., 2002; Saitua et al., 2011; Tahaikt et al., 2007). Nanofiltration membranes operating in a similar fashion to that of RO membranes, except of lower driving pressure and hence lower energy requirements, higher flowrates, and lower product water quality (Schaep et al., 1998).

#### **4.2.3 Electrodialysis/ electrodialysis reversal membrane desalination**

Electrodialysis/Electrodialysis Reversal membrane desalination (ED/EDR) is one of the oldest tried desalination processes, with ED process since the 1950s. In ED/EDR electrochemical separation is the main phenomena takes place utilizing electrical power as driving force to separate ions through ion-exchange membranes (Gleick, 2001). In typical ED

Inland Desalination: Potentials and Challenges 459

There are many other desalination process that are available today, some of them still at research and development stages, however they did not reach the development level to be commercialized on large scale as the previous processes, although many of them have a very promising features over the now widely used desalination process such as less corrosion and scaling problems, less energy consumption, less need for pre and post treatment. These processes include solar, ion-exchange, freezing, and membrane distillation.

Desalination has been extensively used over the past decades; with the great developments in desalination industry that have led to a higher acceptance and growth worldwide, particularly in arid areas. Although different desalination processes are well established today, further development are needed to resolve its various technical and operational issues which represent the essential key for successful desalination, including feed characterization for fouling and scaling propensity, process development, energy requirements, desalination economics, and finally brine disposal which will be given more

Fouling is a phenomenon that plagues the operation of desalination units, the deposition of foulants on to the heat (in case of thermal desalination) and mass (in case of membrane desalination) transfer surface results in reducing water productivity and decrease product quality, therefore as the fouling deposit builds up the energy consumption increases to accommodate for the required product flow till the unit is cleaned (Hamrouni & Dhahbi,

Fouling of membrane surface can be caused by any of the rejected constitutes, and can be generally classified to chemical fouling or scaling caused by sparingly soluble inorganic salts exceeding their saturation level, physical or colloidal fouling caused by particulate matter, biological fouling or biofouling due to the formation of biofilms of microorganisms, and finally organic fouling caused by natural organic matter NOM (Fritzmann et al., 2007). The water recovery is mainly constrained by fouling, and hence it is paramount to mitigate

The general approach to avoid scaling and fouling by sparingly soluble salts, is to estimate the saturation level of these salts according to the feed water quality, design recovery, and operation conditions and try to operate below such saturation levels where the solution is stable (Sheikholesami, 2004). Extensive work has been done to study the fouling and to determine the saturation levels of common scale forming sparingly soluble salts at different conditions with focus on calcium, barium, and silica particularly calcium sulfate, barium

In seawater desalination main types of fouling is scaling by calcium carbonate, calcium sulfate, and magnesium hydroxide in case of thermal desalination. For membrane desalination, the major fouling of concern is biofouling (Al-Ahmad & Abdil Aleem, 1993). Fouling caused by precipitation of sparingly soluble salts is less likely to occur mainly due

**4.3 Other desalination processes** 

**5.1 Scaling and fouling** 

2001; Sheikholesami, 2004).

**5. Developments in desalination processes** 

attention due to its high importance (Sheikholesami, 2009).

fouling of desalination units (Semiat et al., 2004).

sulfate, calcium carbonate, and silica.

cell, a series of anion- and cation-exchange membranes are a arranged in alternating pattern between the two electrical electrodes, anode and cathode, and hence ion concentrations increase in alternating compartments, and decrease simultaneously in the other compartments (Walha et al., 2007), ED process has been applied successfully but on small scale for brackish water (Adhikary et al., 1991; Brown, 1981; Harkare et al., 1982), and seawater (Sadrzadeh & Mohammadi, 2008; Seto et al., 1978).

In the 1970s, electrodialysis reversal EDR has been introduced as an innovative modification to the conventional electrodialysis, EDR operates on the same principle as ED. However the polarities of the electrodes are reversed for short time at specified time intervals, so that the ions are attracted in the opposite direction, and hence the brine and product channels are switched (Katz, 1979). EDR process has several features that promoted its application such as ability to treat feedwater of different qualities i.e. higher content of dissolved and suspended solids, ability to operate with high salts saturation levels and hence higher scale resistance, chlorine tolerance, cleanability, higher recovery, un affected by non-ionic species such as silica, low chemical pretreatment, and durability (Buros, 1999; Fubao, 1985; Katz, 1979; Katz, 1982; Valcour, 2010). In addition ED/EDR have been integrated successfully with other desalination processes such as RO (Oren et al., 2010).

#### **4.2.4 Forward osmosis FO membrane desalination**

Although RO membrane desalination has the major share in membrane desalination plants, the energy and membrane replacement cost are of major concern, and hence there is a search for new low energy and low fouling membrane processes. Forward osmosis (FO) or called direct osmosis (DO), employs the natural physical osmosis process by increasing the osmotic pressure in the permeate side to balance the pressure in the opposite side.

The FO exploits this natural tendency of water to move through the semi-permeable membrane from the saline water to a more concentrated solution called draw solution, the draw solution has significantly higher pressure than the saline water. Draw solutions of different natures have been tested for FO operation. Volatile solutes or gases such as sulfur dioxide which can be stripped out later to have pure water, perceptible salts such as aluminum sulfate which can be treated by lime to precipitate aluminum hydroxide and calcium sulfate, a two-stage system with sulfur dioxide-potassium nitrate used as the draw solutions has been also evaluated (McCutcheon et al., 2005a).

However recently more attention has been given to ammonia-carbon dioxide system due to several advantages such as the high solubility of ammonia and carbon dioxide gasses in water and the solution of formed ammonium bicarbonate have a high osmotic pressure which in turn provide higher water flux and recovery, followed by ease separation of the gases by moderate heating, which will be recycled back to the process (McCutcheon et al., 2005b; McGinnis & Elimelech, 2007).

Unlike other membrane process, FO does not require any hydraulic pressure to be applied and hence less energy requirements, which in turn results in less capital and operating costs. Furthermore the process has much lower fouling propensity (McGinnis & Elimelech, 2007, Phillip et al., 2007). In addition FO has been integrated successfully with other desalination processes such as RO enabling increased recovery, lower energy consumption (Lee et al., 2009; Martinetti et al., 2009; Tang & Ng, 2008; Yangali et al., 2011).

#### **4.3 Other desalination processes**

458 Advances in Chemical Engineering

cell, a series of anion- and cation-exchange membranes are a arranged in alternating pattern between the two electrical electrodes, anode and cathode, and hence ion concentrations increase in alternating compartments, and decrease simultaneously in the other compartments (Walha et al., 2007), ED process has been applied successfully but on small scale for brackish water (Adhikary et al., 1991; Brown, 1981; Harkare et al., 1982), and

In the 1970s, electrodialysis reversal EDR has been introduced as an innovative modification to the conventional electrodialysis, EDR operates on the same principle as ED. However the polarities of the electrodes are reversed for short time at specified time intervals, so that the ions are attracted in the opposite direction, and hence the brine and product channels are switched (Katz, 1979). EDR process has several features that promoted its application such as ability to treat feedwater of different qualities i.e. higher content of dissolved and suspended solids, ability to operate with high salts saturation levels and hence higher scale resistance, chlorine tolerance, cleanability, higher recovery, un affected by non-ionic species such as silica, low chemical pretreatment, and durability (Buros, 1999; Fubao, 1985; Katz, 1979; Katz, 1982; Valcour, 2010). In addition ED/EDR have been integrated successfully

Although RO membrane desalination has the major share in membrane desalination plants, the energy and membrane replacement cost are of major concern, and hence there is a search for new low energy and low fouling membrane processes. Forward osmosis (FO) or called direct osmosis (DO), employs the natural physical osmosis process by increasing the

The FO exploits this natural tendency of water to move through the semi-permeable membrane from the saline water to a more concentrated solution called draw solution, the draw solution has significantly higher pressure than the saline water. Draw solutions of different natures have been tested for FO operation. Volatile solutes or gases such as sulfur dioxide which can be stripped out later to have pure water, perceptible salts such as aluminum sulfate which can be treated by lime to precipitate aluminum hydroxide and calcium sulfate, a two-stage system with sulfur dioxide-potassium nitrate used as the draw

However recently more attention has been given to ammonia-carbon dioxide system due to several advantages such as the high solubility of ammonia and carbon dioxide gasses in water and the solution of formed ammonium bicarbonate have a high osmotic pressure which in turn provide higher water flux and recovery, followed by ease separation of the gases by moderate heating, which will be recycled back to the process (McCutcheon et al.,

Unlike other membrane process, FO does not require any hydraulic pressure to be applied and hence less energy requirements, which in turn results in less capital and operating costs. Furthermore the process has much lower fouling propensity (McGinnis & Elimelech, 2007, Phillip et al., 2007). In addition FO has been integrated successfully with other desalination processes such as RO enabling increased recovery, lower energy consumption (Lee et al.,

osmotic pressure in the permeate side to balance the pressure in the opposite side.

seawater (Sadrzadeh & Mohammadi, 2008; Seto et al., 1978).

with other desalination processes such as RO (Oren et al., 2010).

solutions has been also evaluated (McCutcheon et al., 2005a).

2009; Martinetti et al., 2009; Tang & Ng, 2008; Yangali et al., 2011).

2005b; McGinnis & Elimelech, 2007).

**4.2.4 Forward osmosis FO membrane desalination** 

There are many other desalination process that are available today, some of them still at research and development stages, however they did not reach the development level to be commercialized on large scale as the previous processes, although many of them have a very promising features over the now widely used desalination process such as less corrosion and scaling problems, less energy consumption, less need for pre and post treatment. These processes include solar, ion-exchange, freezing, and membrane distillation.

#### **5. Developments in desalination processes**

Desalination has been extensively used over the past decades; with the great developments in desalination industry that have led to a higher acceptance and growth worldwide, particularly in arid areas. Although different desalination processes are well established today, further development are needed to resolve its various technical and operational issues which represent the essential key for successful desalination, including feed characterization for fouling and scaling propensity, process development, energy requirements, desalination economics, and finally brine disposal which will be given more attention due to its high importance (Sheikholesami, 2009).

#### **5.1 Scaling and fouling**

Fouling is a phenomenon that plagues the operation of desalination units, the deposition of foulants on to the heat (in case of thermal desalination) and mass (in case of membrane desalination) transfer surface results in reducing water productivity and decrease product quality, therefore as the fouling deposit builds up the energy consumption increases to accommodate for the required product flow till the unit is cleaned (Hamrouni & Dhahbi, 2001; Sheikholesami, 2004).

Fouling of membrane surface can be caused by any of the rejected constitutes, and can be generally classified to chemical fouling or scaling caused by sparingly soluble inorganic salts exceeding their saturation level, physical or colloidal fouling caused by particulate matter, biological fouling or biofouling due to the formation of biofilms of microorganisms, and finally organic fouling caused by natural organic matter NOM (Fritzmann et al., 2007). The water recovery is mainly constrained by fouling, and hence it is paramount to mitigate fouling of desalination units (Semiat et al., 2004).

The general approach to avoid scaling and fouling by sparingly soluble salts, is to estimate the saturation level of these salts according to the feed water quality, design recovery, and operation conditions and try to operate below such saturation levels where the solution is stable (Sheikholesami, 2004). Extensive work has been done to study the fouling and to determine the saturation levels of common scale forming sparingly soluble salts at different conditions with focus on calcium, barium, and silica particularly calcium sulfate, barium sulfate, calcium carbonate, and silica.

In seawater desalination main types of fouling is scaling by calcium carbonate, calcium sulfate, and magnesium hydroxide in case of thermal desalination. For membrane desalination, the major fouling of concern is biofouling (Al-Ahmad & Abdil Aleem, 1993). Fouling caused by precipitation of sparingly soluble salts is less likely to occur mainly due

Inland Desalination: Potentials and Challenges 461

desalination processes together, with main objective of maximizing overall recovery, minimizing energy requirements, and cost reduction. NF has been integrated successfully with RO process mainly as pretreatment step, which resulted in improving the RO performance (Al-Shammiri et al., 2004; Hassan et al., 1998), EDR with RO (Oren et al., 2010), FO with RO (Lee et al., 2009; Martinetti et al., 2009; Tang & Ng, 2008, Yangali et al., 2011) in case of membrane desalination processes, integration of VC with MSF and MED (El-Dessouky et al., 2000; Mabrouk et al., 2007), and even combination of thermal and membrane processes by integrating NF/RO/MSF together (Hamed et al., 2009), and FO and

In membrane processes, development of much better membrane material that can work for wider range of pH, chlorine resistant, high mechanical strength to withstand higher hydraulic pressures, better salt rejection, and scale resistant are considered to be the next breakthrough in membrane desalination development, enabling higher recoveries at lower

Desalination processes are known for their intensive energy consumption especially thermal desalination, and hence energy consumption make up the major part of operation cost of any desalination process, the energy consumption differs according to desalination process in use i.e. thermal or membrane, water source and quality i.e. seawater or brackish water, design recovery, system design, plant capacity, and utilization of energy recovery devices. Energy requirement for desalination processes is generally reported as specific energy consumptions in kWh/m3 of product water. There are a wide range of reported values for energy consumption in desalination with the most recent values of about 1.8 kWh/m3 for seawater desalination using MED, 4 kWh/m3 for MSF with heat recovery mechanism incorporated (Khawaji et al., 2008). However for seawater using RO it went down from 20 kWh/m3 in early 1970s to 1.6-2 kWh/m3 recently, and below 1 kWh/m3 for brackish water

In RO operation, the main energy consumption is mainly for the high pressure pump to provide hydraulic pressure in excess to the osmotic pressure, most of this pressure is retained by the concentrate stream flowing out of the RO unit. Energy recovery devices ERD have been developed mainly for RO operation to recover some of the energy retained in the concentrate before disposal. There are two main classes of ERD, class I which transfer hydraulic energy from the concentrate stream to the feed stream in one step with net energy transfer of more than 95%. Class II transfer hydraulic energy of the concentrate to centrifugal mechanical energy and then to hydraulic energy in the feed in two steps process

Integration of desalination plants with power plants or as called cogeneration, which refers to the use of single energy source for multiple needs; mainly encountered with thermal desalination offers better energy utilization. For example, power plants use high pressure steam for power generation by means of turbines, the steam comes out at low pressure which is very suitable for thermal desalination (Gleick, 2001). In addition there are efforts for implementing cogeneration in RO desalination process in order to address the water-

with energy recovery devices (Fritzmann et al., 2007; Khawaji et al., 2008).

MD to RO (Martinetti et al., 2009).

cost (Sheikholesami, 2009).

**5.3 Energy requirements** 

(Greenlee et al., 2009).

electricity demand trade off (Altmann, 1997).

to the relative lower recovery, higher ionic strength, and low bicarbonate and sulfate concentration (Reverter et al., 2001). The contribution of each type of fouling in typical seawater RO desalination are 48% for biofouling, 18% for inorganic colloids, 15% for organic matter, 13% for silicates, and only about 6% for mineral deposits (Shon et al., 2009). Hence one of the most important steps in seawater pretreatment for desalination is disinfection with optimized dose of biocide, usually chlorine, in order to reduce biofouling propensity (Fujiwara & Matsuyama, 2008).

Brackish groundwater however has higher quality as it is mainly characterized by low content of suspended solids, low bacterial count, and low content of organic matter; as a result the most found type of fouling is scaling by sparingly soluble inorganic salts such as calcium and barium salts and silica.

The dissolved solids present in groundwater results mainly from chemical weathering or dissolution of geological formations i.e. minerals which can be attributed to the direct contact of groundwater with the calcium carbonate, and calcium sulfate rocks forming the aquifer. In addition sulfate may result from biological oxidation of reduced sulfur species. As a result the different aspects of scaling by calcium sulfate and calcium carbonate has been intensively studied (Sheikholesami, 2003a; Sheikholesami, 2003b; Rahardianato et al., 2008).

Silica originates from the dissolution or chemical weathering of amorphous or crystalline SiO2 and the major clay minerals (Faust & Aly, 1998). Crystalline silica has a very low solubility in water, however amorphous silica can have solubility up to 120 mg/L at pH 7 and the solubility increases with pH increase reaching around 889 mg/L at pH 10 (Hamrouni & Dhabi, 2001; Sheikholesami & Tan, 1999). Silica in water can be classified into two categories 1) soluble or dissolved silica which contains monomers, dimmers, and polymers of silicic acid, and 2) insoluble or colloidal silica, which results from high polymerization of silicic acid. Due to its severe effect on membrane desalination performance and lifetime of membranes, great attention has been paid to silica fouling (Al-Shammiri et al., 2000; Ning, 2002; Semiat et al., 2003; Sheikholesami et al., 2001,).

In any desalination process, there are three main factors for sustainable operation: 1) proper design, 2) proper pretreatment, and 3) proper operation and maintenance; with the proper pretreatment as the foundation for successful operation (Neofotistou & Demadis, 2004). The primary goal of pretreatment is to lower the fouling propensity during the desalination process, and the required pretreatments depend mainly on the characteristics of the water resource (Greenlee et al., 2009). Scale inhibitors or antiscalents are chemicals that are added to water during pretreatment to prevent scale formation and usually work synergically with dispersant polymers.

However most of traditional antiscalents are successful in scale control for crystalline mineral precipitates but not silica because it is amorphous (Freeman & Majerle, 1995), and hence control of silica scaling requires chemical pretreatment.

#### **5.2 Process development**

Desalination is a multi unit process, starting from water intake, pretreatment, desalination, and post-treatment. Desalination plants in the past used to contain one type of desalination processes in the past. However attention has been recently paid to hybridization of different

to the relative lower recovery, higher ionic strength, and low bicarbonate and sulfate concentration (Reverter et al., 2001). The contribution of each type of fouling in typical seawater RO desalination are 48% for biofouling, 18% for inorganic colloids, 15% for organic matter, 13% for silicates, and only about 6% for mineral deposits (Shon et al., 2009). Hence one of the most important steps in seawater pretreatment for desalination is disinfection with optimized dose of biocide, usually chlorine, in order to reduce biofouling propensity

Brackish groundwater however has higher quality as it is mainly characterized by low content of suspended solids, low bacterial count, and low content of organic matter; as a result the most found type of fouling is scaling by sparingly soluble inorganic salts such as

The dissolved solids present in groundwater results mainly from chemical weathering or dissolution of geological formations i.e. minerals which can be attributed to the direct contact of groundwater with the calcium carbonate, and calcium sulfate rocks forming the aquifer. In addition sulfate may result from biological oxidation of reduced sulfur species. As a result the different aspects of scaling by calcium sulfate and calcium carbonate has been intensively studied (Sheikholesami, 2003a; Sheikholesami, 2003b; Rahardianato et al., 2008). Silica originates from the dissolution or chemical weathering of amorphous or crystalline SiO2 and the major clay minerals (Faust & Aly, 1998). Crystalline silica has a very low solubility in water, however amorphous silica can have solubility up to 120 mg/L at pH 7 and the solubility increases with pH increase reaching around 889 mg/L at pH 10 (Hamrouni & Dhabi, 2001; Sheikholesami & Tan, 1999). Silica in water can be classified into two categories 1) soluble or dissolved silica which contains monomers, dimmers, and polymers of silicic acid, and 2) insoluble or colloidal silica, which results from high polymerization of silicic acid. Due to its severe effect on membrane desalination performance and lifetime of membranes, great attention has been paid to silica fouling (Al-

Shammiri et al., 2000; Ning, 2002; Semiat et al., 2003; Sheikholesami et al., 2001,).

hence control of silica scaling requires chemical pretreatment.

In any desalination process, there are three main factors for sustainable operation: 1) proper design, 2) proper pretreatment, and 3) proper operation and maintenance; with the proper pretreatment as the foundation for successful operation (Neofotistou & Demadis, 2004). The primary goal of pretreatment is to lower the fouling propensity during the desalination process, and the required pretreatments depend mainly on the characteristics of the water resource (Greenlee et al., 2009). Scale inhibitors or antiscalents are chemicals that are added to water during pretreatment to prevent scale formation and usually work synergically with

However most of traditional antiscalents are successful in scale control for crystalline mineral precipitates but not silica because it is amorphous (Freeman & Majerle, 1995), and

Desalination is a multi unit process, starting from water intake, pretreatment, desalination, and post-treatment. Desalination plants in the past used to contain one type of desalination processes in the past. However attention has been recently paid to hybridization of different

(Fujiwara & Matsuyama, 2008).

calcium and barium salts and silica.

dispersant polymers.

**5.2 Process development** 

desalination processes together, with main objective of maximizing overall recovery, minimizing energy requirements, and cost reduction. NF has been integrated successfully with RO process mainly as pretreatment step, which resulted in improving the RO performance (Al-Shammiri et al., 2004; Hassan et al., 1998), EDR with RO (Oren et al., 2010), FO with RO (Lee et al., 2009; Martinetti et al., 2009; Tang & Ng, 2008, Yangali et al., 2011) in case of membrane desalination processes, integration of VC with MSF and MED (El-Dessouky et al., 2000; Mabrouk et al., 2007), and even combination of thermal and membrane processes by integrating NF/RO/MSF together (Hamed et al., 2009), and FO and MD to RO (Martinetti et al., 2009).

In membrane processes, development of much better membrane material that can work for wider range of pH, chlorine resistant, high mechanical strength to withstand higher hydraulic pressures, better salt rejection, and scale resistant are considered to be the next breakthrough in membrane desalination development, enabling higher recoveries at lower cost (Sheikholesami, 2009).

#### **5.3 Energy requirements**

Desalination processes are known for their intensive energy consumption especially thermal desalination, and hence energy consumption make up the major part of operation cost of any desalination process, the energy consumption differs according to desalination process in use i.e. thermal or membrane, water source and quality i.e. seawater or brackish water, design recovery, system design, plant capacity, and utilization of energy recovery devices.

Energy requirement for desalination processes is generally reported as specific energy consumptions in kWh/m3 of product water. There are a wide range of reported values for energy consumption in desalination with the most recent values of about 1.8 kWh/m3 for seawater desalination using MED, 4 kWh/m3 for MSF with heat recovery mechanism incorporated (Khawaji et al., 2008). However for seawater using RO it went down from 20 kWh/m3 in early 1970s to 1.6-2 kWh/m3 recently, and below 1 kWh/m3 for brackish water with energy recovery devices (Fritzmann et al., 2007; Khawaji et al., 2008).

In RO operation, the main energy consumption is mainly for the high pressure pump to provide hydraulic pressure in excess to the osmotic pressure, most of this pressure is retained by the concentrate stream flowing out of the RO unit. Energy recovery devices ERD have been developed mainly for RO operation to recover some of the energy retained in the concentrate before disposal. There are two main classes of ERD, class I which transfer hydraulic energy from the concentrate stream to the feed stream in one step with net energy transfer of more than 95%. Class II transfer hydraulic energy of the concentrate to centrifugal mechanical energy and then to hydraulic energy in the feed in two steps process (Greenlee et al., 2009).

Integration of desalination plants with power plants or as called cogeneration, which refers to the use of single energy source for multiple needs; mainly encountered with thermal desalination offers better energy utilization. For example, power plants use high pressure steam for power generation by means of turbines, the steam comes out at low pressure which is very suitable for thermal desalination (Gleick, 2001). In addition there are efforts for implementing cogeneration in RO desalination process in order to address the waterelectricity demand trade off (Altmann, 1997).

Inland Desalination: Potentials and Challenges 463

Brine, concentrate, or reject are different names for a stream that is commonly produced from any desalination process. In any desalination process, two streams are produced: 1) product water with high quality, 2) brine or concentrate stream that contains all the salts were originally present in the feed water in addition to the chemicals added in the

To reduce energy consumption, cleaning time and expenditure, loss of production during downtime, it is paramount to mitigate fouling. The general approach is to study the feed water characteristics and couple it with the expected recovery and operating conditions; and to operate at conditions where the solution is stable, hence scaling and fouling is minimized. However working at lower recoveries to avoid the fouling of membrane will increase the flow of brine stream generated, which present the main trade-off for desalination operation. Brine stream does not contains only 2 or 4-5 folds the salinity of the feed water as in case of seawater, or brackish water respectively, but it contains all the chemicals that has been added to the desalination process during pretreatment. Moreover in case of thermal desalination it will be at high temperature, and hence more attention should be taken when

The problem of brine discharge is different in sea water and brackish water desalination. In the case of seawater desalination plants the problem is readily solved since these plants are usually placed near the coast, so the discharge method of choose is usually to discharge it back to the sea through brine pipes or submarine emissaries. Encouraging facts to utilize that option are, first the discharged brine is of similar chemistry, even being more concentrated but only by 50-100 %. Second is that the volume of brine stream relative to the water body being discharged to being very small, hence lower drawbacks are expected. However, there are many criteria to be considered such as having the discharge point far enough from intake at good mixing zone so it can be mixed with the

However, the management of brine from brackish desalination plants i.e. inland desalination can be significant problem in case they are placed far from the coast (inland plants). Some of the conventional options for brine disposal from inland desalination plants are: 1) disposal into surface water bodies, 2) disposal to municipal sewers, 3) evaporation ponds, 4) deep well injection, and 5) irrigation of plants tolerant to high salinities (Ahuja & Howe, 2005). The main factors that influence the selection of a disposal method, among others, are: 1) volume or quantity of concentrate, 2) quality and constitutes present in the concentrate, 3) physical and geographical considerations, 4) capital and operational costs, 5) availability of receiving site, 6) permissibility of the option, facility future expansion plan, and 7) public acceptance. All of these factors together will affect the cost of brine disposal

Brine disposal method should be considered after the necessary studies and investigations have been performed in order to minimize the brine stream to be disposed off, and hence reduce the cost of subsequent disposal. This is mainly achieved by employing the proper feed water pretreatment, proper desalination process, maximizing the system recovery. However attention to the increased salinity and quality of brine should be considered.

that can ranges from 5 -33 % of the total desalination cost (Ahmed et al, 2001).

**6. Brine disposal from inland desalination** 

pretreatment and during desalination such as antiscalents.

considering the brine discharge method (Ahmed et al., 2002).

main body of seawater.

Use of renewable energy sources for driving the desalination process provide another development opportunity, specifically for membrane desalination where less energy is required and for rural communities in which desalination systems are generally small in size with usually non-continuous operation, and hence can be integrated to renewable energy sources. Considerable efforts has been made for integrating desalination processes with different renewable energy sources namely solar (photovoltaic and thermal), wind, and geothermal. The renewable energy can be used in one of two forms thermal or electrical, depending on which one that best match the desalination process (Al-karaghouli et al., 2009; Al-karaghouli et al., 2010; Forstmeier et al., 2007; Mathioulakis et al, 2007).

#### **5.4 Desalination economics**

Economics of any process represent the most crucial part for development and application, and hence economical feasibility is a very important factor when considering desalination processes. In desalination processes it is difficult to standardize the economics of process because it is case specific. There are several factors that affect desalination cost such as water source (brackish or seawater), desalination process used (thermal or membrane), energy source (traditional or renewable), and plant size (Dore, 2005; Karagiannis & Soldatos, 2008).

The source, and hence the quality of the feed water plays important rule for determination of both capital and operating cost, and the overall desalination cost. Brackish water has much low salt content and better water quality than seawater and therefore, it incorporates less capital and operating costs. The most recent data for average investment cost for brackish water was around \$200-450/ (m3/d) with product water cost of \$0.25-0.75/m3 for RO process which is the process dominating the brackish water desalination market (Vince et al., 2008; Yun et al, 2006).

The product water desalination cost varies significantly according to the salinity of the water sources. For example the product water cost for brackish water with salinity around 3,000 mg/L was found to be \$0.32/m3. However, for water with salinity around 10,000 mg/L, the desalination cost was \$0.54 /m3 (Karagiannis & Soldatos, 2008). The same trend was observed also for sea water desalination with desalination cost ranging from \$0.54 /m3 for Mediterranean seawater to \$0.87/m3 for Arabic Gulf seawater (Greenlee et al., 2009).

Capacity or size of desalination plant greatly affects the product water cost; table 3 shows the average desalination cost for brackish and seawater desalination plants of different production capacities.


Table 3. Size of desalination plant and water production cost (Karagiannis & Soldatos, 2008).

Use of renewable energy sources for driving the desalination process provide another development opportunity, specifically for membrane desalination where less energy is required and for rural communities in which desalination systems are generally small in size with usually non-continuous operation, and hence can be integrated to renewable energy sources. Considerable efforts has been made for integrating desalination processes with different renewable energy sources namely solar (photovoltaic and thermal), wind, and geothermal. The renewable energy can be used in one of two forms thermal or electrical, depending on which one that best match the desalination process (Al-karaghouli et al., 2009;

Economics of any process represent the most crucial part for development and application, and hence economical feasibility is a very important factor when considering desalination processes. In desalination processes it is difficult to standardize the economics of process because it is case specific. There are several factors that affect desalination cost such as water source (brackish or seawater), desalination process used (thermal or membrane), energy source (traditional or renewable), and plant size (Dore, 2005; Karagiannis & Soldatos, 2008). The source, and hence the quality of the feed water plays important rule for determination of both capital and operating cost, and the overall desalination cost. Brackish water has much low salt content and better water quality than seawater and therefore, it incorporates less capital and operating costs. The most recent data for average investment cost for brackish water was around \$200-450/ (m3/d) with product water cost of \$0.25-0.75/m3 for RO process which is the process dominating the brackish water desalination market (Vince

The product water desalination cost varies significantly according to the salinity of the water sources. For example the product water cost for brackish water with salinity around 3,000 mg/L was found to be \$0.32/m3. However, for water with salinity around 10,000 mg/L, the desalination cost was \$0.54 /m3 (Karagiannis & Soldatos, 2008). The same trend was observed also for sea water desalination with desalination cost ranging from \$0.54 /m3 for Mediterranean seawater to \$0.87/m3 for Arabic Gulf seawater (Greenlee et al., 2009).

Capacity or size of desalination plant greatly affects the product water cost; table 3 shows the average desalination cost for brackish and seawater desalination plants of different

**Brackish** ≤ 1,000 0.78 – 1.33

**Seawater** < 1,000 2.2 – 11.25

Table 3. Size of desalination plant and water production cost (Karagiannis & Soldatos, 2008).

**Feed water Plant size (m3 / d) Cost (\$ / m3)** 

5,000 – 60,000 0.26 – 0.54

1,000 – 5,000 0.7 – 3.9 12,000 – 60,000 0.44 – 1.62 > 60,000 0.50 – 1.0

Al-karaghouli et al., 2010; Forstmeier et al., 2007; Mathioulakis et al, 2007).

**5.4 Desalination economics** 

et al., 2008; Yun et al, 2006).

production capacities.

#### **6. Brine disposal from inland desalination**

Brine, concentrate, or reject are different names for a stream that is commonly produced from any desalination process. In any desalination process, two streams are produced: 1) product water with high quality, 2) brine or concentrate stream that contains all the salts were originally present in the feed water in addition to the chemicals added in the pretreatment and during desalination such as antiscalents.

To reduce energy consumption, cleaning time and expenditure, loss of production during downtime, it is paramount to mitigate fouling. The general approach is to study the feed water characteristics and couple it with the expected recovery and operating conditions; and to operate at conditions where the solution is stable, hence scaling and fouling is minimized. However working at lower recoveries to avoid the fouling of membrane will increase the flow of brine stream generated, which present the main trade-off for desalination operation.

Brine stream does not contains only 2 or 4-5 folds the salinity of the feed water as in case of seawater, or brackish water respectively, but it contains all the chemicals that has been added to the desalination process during pretreatment. Moreover in case of thermal desalination it will be at high temperature, and hence more attention should be taken when considering the brine discharge method (Ahmed et al., 2002).

The problem of brine discharge is different in sea water and brackish water desalination. In the case of seawater desalination plants the problem is readily solved since these plants are usually placed near the coast, so the discharge method of choose is usually to discharge it back to the sea through brine pipes or submarine emissaries. Encouraging facts to utilize that option are, first the discharged brine is of similar chemistry, even being more concentrated but only by 50-100 %. Second is that the volume of brine stream relative to the water body being discharged to being very small, hence lower drawbacks are expected. However, there are many criteria to be considered such as having the discharge point far enough from intake at good mixing zone so it can be mixed with the main body of seawater.

However, the management of brine from brackish desalination plants i.e. inland desalination can be significant problem in case they are placed far from the coast (inland plants). Some of the conventional options for brine disposal from inland desalination plants are: 1) disposal into surface water bodies, 2) disposal to municipal sewers, 3) evaporation ponds, 4) deep well injection, and 5) irrigation of plants tolerant to high salinities (Ahuja & Howe, 2005). The main factors that influence the selection of a disposal method, among others, are: 1) volume or quantity of concentrate, 2) quality and constitutes present in the concentrate, 3) physical and geographical considerations, 4) capital and operational costs, 5) availability of receiving site, 6) permissibility of the option, facility future expansion plan, and 7) public acceptance. All of these factors together will affect the cost of brine disposal that can ranges from 5 -33 % of the total desalination cost (Ahmed et al, 2001).

Brine disposal method should be considered after the necessary studies and investigations have been performed in order to minimize the brine stream to be disposed off, and hence reduce the cost of subsequent disposal. This is mainly achieved by employing the proper feed water pretreatment, proper desalination process, maximizing the system recovery. However attention to the increased salinity and quality of brine should be considered.

Inland Desalination: Potentials and Challenges 465

caused by improper disposal of concentrates on nearby land can be another disadvantage of

In conclusion, while operation and maintenance costs for evaporation ponds are minimal, large land areas are required which increases as the plant capacity increase, and pond construction costs are high due to lining and monitoring requirements. Consequently, proper evaporation ponds are often cost prohibitive and impractical for handling significant concentrate flow. Furthermore, water evaporated from a pond is often a lost resource.

In deep well injection, the brine is injected back underground to depth ranges from few hundreds of meters to thousands of meters, depending on many factors which should be considered while designing, installing, and operating the system. Deep well injection for brine disposal includes permitting considerations, which look for identification of adequate geologic confining unit to prevent upward migration of effluent from the injection area. While design considerations focus generally on the tubing and packing installed inside the final cemented casing of the injection well, compatibility of the concentrate with the tubing material (to avoid corrosion), expected concentrate flow, and leak detection and monitoring

One of the very attractive options with deep well injection is to use depleted oil and gas fields for brine disposal. This encounter many advantages such as making use of the readily available gas and oil wells, long experience encountered with the operation of such wells. However before applying this option the fields should be tested physically and chemically

Generally site selection for installing of such deep well, is the most important step, and hence hydrological and geological conditions should be considered, as example the wells should never be installed in areas vulnerable to earthquakes (Ahmed et al., 2000). Although of availability of such option to many inland desalination plants, however many factors should be considered with deep well injection for brine disposal which can be summarized

1. Site selection, which is performed through many geological and hydrological studies, to

4. Seismic activity which could cause damage to the well and subsequently result in

6. Pollution of groundwater resources, which may result from high salinity and the

Land application such as use in irrigation systems that was originally developed for sewage effluents, can be used for brine disposal, and hence helps conserve natural resources. In areas where water conservation is of great importance, spray irrigation is especially

for accepting the brine stream [Mace et al., 2006; Nicot & Cjowdhury, 2005]

brine disposal to evaporation ponds.

systems (Skehan & Kwiatkowski, 2000).

as follow (Mickley et al, 2006):

5. Uncertainty of the well life,

**6.4 Land applications of brine** 

leakage,

identify the proper area for installing the well,

presence of other harmful chemicals in the brine.

2. High cost, associated with both capital and operational cost, 3. Possibility of corrosion and subsequent leakage in the well casing,

**6.3 Deep well injection** 

#### **6.1 Disposal to surface water bodies and sewers systems**

Disposal of brine to surface water bodies if a available, present the first option to choose as it represent a ready and good solution to the challenge of brine disposal taking into account that the brine stream is diluted by mixing with the water body. However many consideration should be taken into account. The salinity if the receiving body might increase due to the disposal of the high salinity brine, and hence the self-purification capacity of the receiving water should be considered (Ahmed et al., 2000). As a result disposal to surface water should be permitted only if that will avoid dramatic impact on environment.

Another option is to dispose the brine to the local sewage system, which is usually employed by small membrane desalination plants. This option has many advantages such as use of the ready available and installed sewage system, lowering the BOD of the domestic sewage water. However that should be practiced carefully as the salinity of sewage water might increase which might affect the wastewater treatment facility especially biological treatment step. This might also render treated municipal sewage effluent unsuitable for agriculture use when disposing brines with high salinity (Ahmed et al., 2000).

#### **6.2 Disposal to evaporation ponds**

In evaporation bonds, the brine is discharged into a large surface area pond, where the water is naturally evaporated. Use of evaporation pond technology is practiced primarily in the arid and semi-arid areas, particularly in Middle East and Australia. Evaporation pond is probably the most widespread method for brine disposal from inland desalination plants.

Simple evaporation ponds have many advantages such as being easy to construct, low maintenance and operation cost, no equipment are needed specifically mechanical. Making it the most appropriate method with lower cost, especially in arid areas with high evaporation rates, low rainfall, and low land cost (Ahmed et al., 2000). Use of evaporation ponds for cultivation of brine shrimps has been studies as well, giving ideal place for brineshrimp production as it present mono-culture environment under natural conditions with absence of any food competitors or predators (Ahmed et al., 2001).

The basic concern associated with use of evaporation pond for brine disposal is leakage of brine through soil. This may result in subsequent contamination and increasing salinity of the aquifer. Electrical conductivity and concentration of salts in the evaporation ponds can be used as indicators for leakage in the pond, where insignificant increase is a strong indication of brine leakage through the soil (Amed et al., 2001). Deterioration of soil and groundwater quality in areas nearby evaporation ponds used for brine disposal in KSA, UAE, and Oman was investigated and reported as one of the draw backs to use of evaporation bonds (Al-Faifi et al, 2010; Mohamed et al., 2005).

As a result most of the evaporation ponds installed recently are lined with polymeric sheets. Liner installation should be carried out carefully as joints sealing is very important for leakage prevention. Furthermore double lining is strongly recommended with proper monitoring for leakage.

In addition reduction in production from agricultural lands caused by deposition of airborne salts from dried concentrate of evaporation bonds, and formation of eyesores caused by improper disposal of concentrates on nearby land can be another disadvantage of brine disposal to evaporation ponds.

In conclusion, while operation and maintenance costs for evaporation ponds are minimal, large land areas are required which increases as the plant capacity increase, and pond construction costs are high due to lining and monitoring requirements. Consequently, proper evaporation ponds are often cost prohibitive and impractical for handling significant concentrate flow. Furthermore, water evaporated from a pond is often a lost resource.

#### **6.3 Deep well injection**

464 Advances in Chemical Engineering

Disposal of brine to surface water bodies if a available, present the first option to choose as it represent a ready and good solution to the challenge of brine disposal taking into account that the brine stream is diluted by mixing with the water body. However many consideration should be taken into account. The salinity if the receiving body might increase due to the disposal of the high salinity brine, and hence the self-purification capacity of the receiving water should be considered (Ahmed et al., 2000). As a result disposal to surface water should be permitted only if that will avoid dramatic impact on

Another option is to dispose the brine to the local sewage system, which is usually employed by small membrane desalination plants. This option has many advantages such as use of the ready available and installed sewage system, lowering the BOD of the domestic sewage water. However that should be practiced carefully as the salinity of sewage water might increase which might affect the wastewater treatment facility especially biological treatment step. This might also render treated municipal sewage effluent unsuitable for

In evaporation bonds, the brine is discharged into a large surface area pond, where the water is naturally evaporated. Use of evaporation pond technology is practiced primarily in the arid and semi-arid areas, particularly in Middle East and Australia. Evaporation pond is probably the most widespread method for brine disposal from inland desalination plants. Simple evaporation ponds have many advantages such as being easy to construct, low maintenance and operation cost, no equipment are needed specifically mechanical. Making it the most appropriate method with lower cost, especially in arid areas with high evaporation rates, low rainfall, and low land cost (Ahmed et al., 2000). Use of evaporation ponds for cultivation of brine shrimps has been studies as well, giving ideal place for brineshrimp production as it present mono-culture environment under natural conditions with

The basic concern associated with use of evaporation pond for brine disposal is leakage of brine through soil. This may result in subsequent contamination and increasing salinity of the aquifer. Electrical conductivity and concentration of salts in the evaporation ponds can be used as indicators for leakage in the pond, where insignificant increase is a strong indication of brine leakage through the soil (Amed et al., 2001). Deterioration of soil and groundwater quality in areas nearby evaporation ponds used for brine disposal in KSA, UAE, and Oman was investigated and reported as one of the draw backs to use of

As a result most of the evaporation ponds installed recently are lined with polymeric sheets. Liner installation should be carried out carefully as joints sealing is very important for leakage prevention. Furthermore double lining is strongly recommended with proper

In addition reduction in production from agricultural lands caused by deposition of airborne salts from dried concentrate of evaporation bonds, and formation of eyesores

agriculture use when disposing brines with high salinity (Ahmed et al., 2000).

absence of any food competitors or predators (Ahmed et al., 2001).

evaporation bonds (Al-Faifi et al, 2010; Mohamed et al., 2005).

**6.1 Disposal to surface water bodies and sewers systems** 

environment.

**6.2 Disposal to evaporation ponds** 

monitoring for leakage.

In deep well injection, the brine is injected back underground to depth ranges from few hundreds of meters to thousands of meters, depending on many factors which should be considered while designing, installing, and operating the system. Deep well injection for brine disposal includes permitting considerations, which look for identification of adequate geologic confining unit to prevent upward migration of effluent from the injection area. While design considerations focus generally on the tubing and packing installed inside the final cemented casing of the injection well, compatibility of the concentrate with the tubing material (to avoid corrosion), expected concentrate flow, and leak detection and monitoring systems (Skehan & Kwiatkowski, 2000).

One of the very attractive options with deep well injection is to use depleted oil and gas fields for brine disposal. This encounter many advantages such as making use of the readily available gas and oil wells, long experience encountered with the operation of such wells. However before applying this option the fields should be tested physically and chemically for accepting the brine stream [Mace et al., 2006; Nicot & Cjowdhury, 2005]

Generally site selection for installing of such deep well, is the most important step, and hence hydrological and geological conditions should be considered, as example the wells should never be installed in areas vulnerable to earthquakes (Ahmed et al., 2000). Although of availability of such option to many inland desalination plants, however many factors should be considered with deep well injection for brine disposal which can be summarized as follow (Mickley et al, 2006):


#### **6.4 Land applications of brine**

Land application such as use in irrigation systems that was originally developed for sewage effluents, can be used for brine disposal, and hence helps conserve natural resources. In areas where water conservation is of great importance, spray irrigation is especially

Inland Desalination: Potentials and Challenges 467

There have been many attempts to achieve a successful inland desalination with ZLD, however more attention and further research work and process developments are needed in order develop a full economic-technical feasible ZLD desalination. In the following sections the current efforts for providing a ZLD system, as well as further developments and

Little literature work is available on ZLD systems for inland desalination; however three

 Applying thermal processes directly to the brine generated from the primary desalination process, usually RO, followed by thermal processes for brine concentration, then crystallization or drying for final salt production (Mickley et al,

 Applying chemical treatment to the brine stream, followed by further membrane desalination, brine concentration, and finally crystallization or drying for final salt

 Applying ED/EDR process to the brine stream making use of higher recovery encountered with such units, followed by crystallization or drying for final salt

Similarities between schemes are clear, especially for brine concentration and final salt production, with the difference mainly in brine treatment and further desalination. However brine treatment and further desalination results in significant reduction in the volume of brine to undergo the brine concentration and final drying/crystallization. Options for integrating different units in different setups can be investigated with an overall objective of ZLD desalination and production of salts can be generated, and should be

Conventional inland desalination system usually achieve 70-85 % recovery of the feed water, which is the largest recovery increment in single step, the recovery mainly depend on the quality of the feed water. However this recovery is usually limited due to scaling by sparingly soluble salts, typically calcium salts such as calcium sulfate and carbonate, in addition to silica (Freeman & Majerle, 1995; Rhardianato et al., 2008; Sheikholesami, 2003a; Sheikholesami, 2004). With this recovery range, about 15-30 % of the feed stream will be

In the first ZLD scheme this 15-30 % is fed to thermal processes using single or multiple effect evaporators or vapor compression evaporators for brine concentration which to be followed by crystallization or drying to obtain final dry salts. In the second scheme the brine is treated chemically to remove most of scale forming constitutes, achieving high removal of such constitutes rendering the treated brine suitable for further membrane separation to recover more water. With the two membrane desalination process with the intermediate brine treatment step recovery up to 95% can be achieved, moreover making use of membrane desalination reduces the cost, and minimizes the energy requirements. Moreover

production (Bond & Veerapaneni, 2008; Mohammadesmaeili et al., 2010).

integrated at any stage from design to operation stages.

**7.1 Current zero liquid discharge schemes for inland desalination** 

main schemes can be concluded and summarized as follow:

production (Greenlee et al., 2009, Oren et al., 2010).

evaluated for process optimization (Kim, 2011).

rejected as brine which should be disposed off.

research needs will be discussed.

2006).

on treat the brine resulted from the existing desalination plants, and hence it can be

attractive option. Concentrate can be applied to cropland or vegetation by sprinkling or surface techniques for water conservation when lawns, parks, or golf courses are irrigated and for preservation and enlargement of greenbelts and open spaces. Crops such as watertolerant grasses with low potential for economic return but with high salinity tolerance are generally chosen for this type. However soil sanlinization and groundwater contamination should be carefully considered (Mickley et al, 2006).

### **7. Inland desalination with zero liquid discharge**

In many cases of brackish water desalination, brine management is critical and of high concern, and hence the need for affordable inland desalination has become critical in many regions of the world where communities strive to meet rapidly growing water demands with limited freshwater resources.

Where brine disposal and management is a problem, given the disadvantages of existing brine disposal and management methods, it is imperative to find alternative Zero Liquid Discharge ZLD technologies that provide more affordable concentrate management. In ZLD, brine is treated to produce desalinated water and essentially dry salts; therefore there is no discharge of liquid waste from the site. Most ZLD applications in operation today treat industrial wastewater using thermal or membrane separation processes, or a combination of these technologies.

Thermal desalination is a mature technology that has been practiced for long time especially where energy is relatively inexpensive, while it is a proven process that generates high quality product water, thermal desalination is energy-intensive and its capital and operating costs are high. Membrane processes has been proved to provide high quality water, but also has some limitation concerning scaling and maximum hydraulic pressure and cannot alone provide ZLD solution. Advancement of ZLD science and associated reduction of ZLD costs will be of tremendous benefit and will alleviate the water supply challenges faced by many communities worldwide.

ZLD desalination present the perfect solution for the brine disposal and management problem usually encountered with inland desalination plants. In addition applying inland desalination with ZLD provide several advantages, the main advantages can summarized as below:


attractive option. Concentrate can be applied to cropland or vegetation by sprinkling or surface techniques for water conservation when lawns, parks, or golf courses are irrigated and for preservation and enlargement of greenbelts and open spaces. Crops such as watertolerant grasses with low potential for economic return but with high salinity tolerance are generally chosen for this type. However soil sanlinization and groundwater contamination

In many cases of brackish water desalination, brine management is critical and of high concern, and hence the need for affordable inland desalination has become critical in many regions of the world where communities strive to meet rapidly growing water demands

Where brine disposal and management is a problem, given the disadvantages of existing brine disposal and management methods, it is imperative to find alternative Zero Liquid Discharge ZLD technologies that provide more affordable concentrate management. In ZLD, brine is treated to produce desalinated water and essentially dry salts; therefore there is no discharge of liquid waste from the site. Most ZLD applications in operation today treat industrial wastewater using thermal or membrane separation processes, or a combination of

Thermal desalination is a mature technology that has been practiced for long time especially where energy is relatively inexpensive, while it is a proven process that generates high quality product water, thermal desalination is energy-intensive and its capital and operating costs are high. Membrane processes has been proved to provide high quality water, but also has some limitation concerning scaling and maximum hydraulic pressure and cannot alone provide ZLD solution. Advancement of ZLD science and associated reduction of ZLD costs will be of tremendous benefit and will alleviate the water supply challenges faced by many

ZLD desalination present the perfect solution for the brine disposal and management problem usually encountered with inland desalination plants. In addition applying inland desalination with ZLD provide several advantages, the main advantages can summarized as

 Maximize Water Recovery: with ZLD systems approaches 100 % recovery, when compared to the conventional Inland desalination system with regular recovery of about 70-85%. ZLD systems should be able to provide more product water or less plant

 Preserving Natural Resources: for inland desalination with ZLD systems, the natural resources, which are mainly groundwater and land, are preserved both quantitatively and qualitatively, by avoiding the different problems associated with conventional

 Byproduct salts: the ZLD system results into two stream, product water, and dry salts, these salts can be treated as added value product rather than solid waste, finding a lot

 Integerability and applicability: ZLD system can be integrated to any existing inland desalination plant of any size and location. This is mainly because the system operates

should be carefully considered (Mickley et al, 2006).

with limited freshwater resources.

these technologies.

communities worldwide.

size by 15-30 %.

brine disposal methods.

of applications and beneficial uses.

below:

**7. Inland desalination with zero liquid discharge** 

on treat the brine resulted from the existing desalination plants, and hence it can be integrated at any stage from design to operation stages.

There have been many attempts to achieve a successful inland desalination with ZLD, however more attention and further research work and process developments are needed in order develop a full economic-technical feasible ZLD desalination. In the following sections the current efforts for providing a ZLD system, as well as further developments and research needs will be discussed.

#### **7.1 Current zero liquid discharge schemes for inland desalination**

Little literature work is available on ZLD systems for inland desalination; however three main schemes can be concluded and summarized as follow:


Similarities between schemes are clear, especially for brine concentration and final salt production, with the difference mainly in brine treatment and further desalination. However brine treatment and further desalination results in significant reduction in the volume of brine to undergo the brine concentration and final drying/crystallization. Options for integrating different units in different setups can be investigated with an overall objective of ZLD desalination and production of salts can be generated, and should be evaluated for process optimization (Kim, 2011).

Conventional inland desalination system usually achieve 70-85 % recovery of the feed water, which is the largest recovery increment in single step, the recovery mainly depend on the quality of the feed water. However this recovery is usually limited due to scaling by sparingly soluble salts, typically calcium salts such as calcium sulfate and carbonate, in addition to silica (Freeman & Majerle, 1995; Rhardianato et al., 2008; Sheikholesami, 2003a; Sheikholesami, 2004). With this recovery range, about 15-30 % of the feed stream will be rejected as brine which should be disposed off.

In the first ZLD scheme this 15-30 % is fed to thermal processes using single or multiple effect evaporators or vapor compression evaporators for brine concentration which to be followed by crystallization or drying to obtain final dry salts. In the second scheme the brine is treated chemically to remove most of scale forming constitutes, achieving high removal of such constitutes rendering the treated brine suitable for further membrane separation to recover more water. With the two membrane desalination process with the intermediate brine treatment step recovery up to 95% can be achieved, moreover making use of membrane desalination reduces the cost, and minimizes the energy requirements. Moreover

Inland Desalination: Potentials and Challenges 469

Rehaili, 2003; Cheng et al, 2009; Gilron et al., 2005; Kadem & Zalmon, 1997; Masarawa et al., 1997; Oren et al., 2001; Sheikholeslami & Bright, 2002) showing high efficiency in removal of calcium, magnesium, silica, and heavy metals. However such softening processes were found to be very effective and promising when applied for brine treatment where high calcium, magnesium, and silica removals from brine streams has been achived enabling higher recovery in the subsequent membrane desalination, facilitate reaching zero liquid discharge desalination [Comstock et al., 2011; Gabelich et al., 2007; Ning et al., 2006; Ning &

Sulfate usually present in the brine streams in high concentrations, relative to those of calcium, magnesium, carbonate, and silica. However in presence of calcium, saturation and hence scaling due to calcium sulfate is very likely to happen (Rahardianato et al., 2008; Sheikholesami, 2003a; Sheikholesami, 2004). Precipitative softening processes were found to be very effective in removal of calcium, magnesium, carbonate, and silica. However such process had no success for removal of sulfate, even though removal of calcium from brine stream reduces the scaling potential of calcium sulfate. However it will be paramount to remove sulfate completely or partially, converting the brine chemistry typically to monovalent ions i.e. sodium, potassium, and chloride which has no scaling potential at the

Several works has been performed on removal of sulfate from industrial wastewater streams such as paper mills, mining, and fertilizers, and several attempts have been performed to reach zero discharge with such streams using membrane and thermal separation processes (Ericsson & Hallmans, 1996). However many attempts has been worked to employ precipitation and crystallization removal of sulfate as calcium sulfate, gypsum, by addition of calcium mainly as calcium hydroxide, lime (Tait et al., 2009), or as calcium chloride (Benatti et al., 2009), which was found to be very effective in removal of sulfate. Removal of sulfate as calcium sulfate below 1300 mg/L was found to be very hard due to solubility limits. However addition of aluminum as aluminum sulfate or alum, aluminum chloride, aluminum nitrate (Christoe, 1976), and sodium aluminate (Batchelor et al, 1985) was found to enhance the removal of sulfate far below this value by formation of more complex solids (Batchelor et al, 1985; Christoe, 1976) which has much lower solubility compared to that of

The treated brine after being further concentrated in secondary membrane desalination has to be further concentrated reaching zero liquid. Zero liquid and dry salts cannot be produced by membrane desalination such as RO or NF, and hence the concentrated brine has to be subjected to thermal process such as brine concentration followed by

Thermal processes such as single and multi stage evaporators, or vapor compression evaporators are usually employed for further brine concentration. Such units have dual purpose objectives which are further recovery of water with very high quality with salinity about 10 mg/L, and brine concentration up to 250,000 mg/L, with recovery above 90%. Final salt production can be achieved after brine concentration which usually performed in

Tryoer, 2009; Rahardianto et al., 2007].

normal membrane desalination operating conditions.

calcium sulfate precipitated by lime addition only.

crystallization or drying to produce dry salts.

crystallizers or dryers (Mickley, 2006).

**7.3 Secondary brine concentration and final salt production** 

reduces the volume of brine to be handled by final evaporation step which results in lowering the energy requirements and hence the overall process cost. In the last scheme ED/EDR unit are employed, which can operate at high saturation levels of sparingly soluble salts as in the case of brine streams, and where high recovery up to 97% can be achieved.

#### **7.2 Precipitation softening for brine treatment in zero liquid discharge systems**

Intermediate brine treatment step as employed in the second ZLD scheme is the one receiving large attention recently. The main objective of this step to remove most of the scale forming constitutes typically calcium, magnesium, carbonate, sulfate, and silica. It is hard to find a chemical treatment process that is able to efficiently remove all of these constitutes. Furthermore most of the tested chemical treatment processes were not able to completely remove such constitutes. However the achieved removal efficiency was good enough to prevent such constitutes from limiting the recovery in the secondary membrane desalination process.

Precipitation softening is one of the widely used processes for reduction of hardness (calcium and magnesium) and alkalinity (mainly bicarbonate) in water treatment plants. The reduction of hardness is mainly achieved by removal of calcium as calcium carbonate CaCO3 and magnesium as magnesium hydroxide Mg(OH)2. This is usually achieved by addition of alkali usually lime, calcium hydroxide Ca(OH)2 in lime softening or sodium hydroxide, NaOH in caustic softening and sodium carbonate Na2CO3 depending on the quality of water to be treated (Reynolds & Richards, 1996). Removal of calcium and magnesium present one of the major targets for brine chemical treatment, particularly calcium, as magnesium cause scaling problems only at high pH values forming insoluble magnesium hydroxide Mg(OH)2. However at the normal pH values found in brine streams it will be mainly saturated by calcium sulfate and carbonate (Sheikholesami, 2003a; Sheikholesami, 2004; Rharadianato et al., 2008).

Silica removal during the precipitation softening was extensively studied, and it was found that silica is removal could be by co-precipitation with metal hydroxides, specifically iron, manganese, and magnesium hydroxides, or could be by precipitation as magnesium and calcium silicate (Sheikholesami & Bright, 2002). Furthermore caustic softening using only sodium hydroxide was found to be more effective and more viable in removal of silica over lime-soda softening using lime and soda ash (Al-Rehaili, 2003; Sheikholesami & Bright, 2003). Addition of sodium aluminate and aluminum sulfate was found to enhance the removal of silica during the softening process by co-precipitation with aluminum hydroxide (Cheng et al., 2009; Lindsay & Ryznar, 1939). Conventional softening process is slow, requires extensive space, and generates large volume of sludge which will need dewatering and further treatment later on (Kadem & Zalmon, 1997). As a result a more advanced process designated Compact Accelerated Precipitation Softening CAPS was developed to enhance the performance of the precipitation softening process. In CAPS process the saturated solution is passed through cake of calcium carbonate to enhance crystallization and approach equilibrium rapidly, the process has been found to overcome the different disadvantages encountered in conventional softening process (Gilron et al., 2005; Masarawa et al., 1997; Oren et al., 2001).

Although different precipitation softening processes have been applied basically for surface water treatment and as a pretreatment for membrane processes, specifically NF and RO (Al-

reduces the volume of brine to be handled by final evaporation step which results in lowering the energy requirements and hence the overall process cost. In the last scheme ED/EDR unit are employed, which can operate at high saturation levels of sparingly soluble salts as in the case of brine streams, and where high recovery up to 97% can be achieved.

Intermediate brine treatment step as employed in the second ZLD scheme is the one receiving large attention recently. The main objective of this step to remove most of the scale forming constitutes typically calcium, magnesium, carbonate, sulfate, and silica. It is hard to find a chemical treatment process that is able to efficiently remove all of these constitutes. Furthermore most of the tested chemical treatment processes were not able to completely remove such constitutes. However the achieved removal efficiency was good enough to prevent such constitutes from limiting the recovery in the secondary membrane desalination

Precipitation softening is one of the widely used processes for reduction of hardness (calcium and magnesium) and alkalinity (mainly bicarbonate) in water treatment plants. The reduction of hardness is mainly achieved by removal of calcium as calcium carbonate CaCO3 and magnesium as magnesium hydroxide Mg(OH)2. This is usually achieved by addition of alkali usually lime, calcium hydroxide Ca(OH)2 in lime softening or sodium hydroxide, NaOH in caustic softening and sodium carbonate Na2CO3 depending on the quality of water to be treated (Reynolds & Richards, 1996). Removal of calcium and magnesium present one of the major targets for brine chemical treatment, particularly calcium, as magnesium cause scaling problems only at high pH values forming insoluble magnesium hydroxide Mg(OH)2. However at the normal pH values found in brine streams it will be mainly saturated by calcium sulfate and carbonate (Sheikholesami, 2003a;

Silica removal during the precipitation softening was extensively studied, and it was found that silica is removal could be by co-precipitation with metal hydroxides, specifically iron, manganese, and magnesium hydroxides, or could be by precipitation as magnesium and calcium silicate (Sheikholesami & Bright, 2002). Furthermore caustic softening using only sodium hydroxide was found to be more effective and more viable in removal of silica over lime-soda softening using lime and soda ash (Al-Rehaili, 2003; Sheikholesami & Bright, 2003). Addition of sodium aluminate and aluminum sulfate was found to enhance the removal of silica during the softening process by co-precipitation with aluminum hydroxide (Cheng et al., 2009; Lindsay & Ryznar, 1939). Conventional softening process is slow, requires extensive space, and generates large volume of sludge which will need dewatering and further treatment later on (Kadem & Zalmon, 1997). As a result a more advanced process designated Compact Accelerated Precipitation Softening CAPS was developed to enhance the performance of the precipitation softening process. In CAPS process the saturated solution is passed through cake of calcium carbonate to enhance crystallization and approach equilibrium rapidly, the process has been found to overcome the different disadvantages encountered in conventional softening process (Gilron et al., 2005; Masarawa et al., 1997; Oren et al., 2001).

Although different precipitation softening processes have been applied basically for surface water treatment and as a pretreatment for membrane processes, specifically NF and RO (Al-

**7.2 Precipitation softening for brine treatment in zero liquid discharge systems** 

process.

Sheikholesami, 2004; Rharadianato et al., 2008).

Rehaili, 2003; Cheng et al, 2009; Gilron et al., 2005; Kadem & Zalmon, 1997; Masarawa et al., 1997; Oren et al., 2001; Sheikholeslami & Bright, 2002) showing high efficiency in removal of calcium, magnesium, silica, and heavy metals. However such softening processes were found to be very effective and promising when applied for brine treatment where high calcium, magnesium, and silica removals from brine streams has been achived enabling higher recovery in the subsequent membrane desalination, facilitate reaching zero liquid discharge desalination [Comstock et al., 2011; Gabelich et al., 2007; Ning et al., 2006; Ning & Tryoer, 2009; Rahardianto et al., 2007].

Sulfate usually present in the brine streams in high concentrations, relative to those of calcium, magnesium, carbonate, and silica. However in presence of calcium, saturation and hence scaling due to calcium sulfate is very likely to happen (Rahardianato et al., 2008; Sheikholesami, 2003a; Sheikholesami, 2004). Precipitative softening processes were found to be very effective in removal of calcium, magnesium, carbonate, and silica. However such process had no success for removal of sulfate, even though removal of calcium from brine stream reduces the scaling potential of calcium sulfate. However it will be paramount to remove sulfate completely or partially, converting the brine chemistry typically to monovalent ions i.e. sodium, potassium, and chloride which has no scaling potential at the normal membrane desalination operating conditions.

Several works has been performed on removal of sulfate from industrial wastewater streams such as paper mills, mining, and fertilizers, and several attempts have been performed to reach zero discharge with such streams using membrane and thermal separation processes (Ericsson & Hallmans, 1996). However many attempts has been worked to employ precipitation and crystallization removal of sulfate as calcium sulfate, gypsum, by addition of calcium mainly as calcium hydroxide, lime (Tait et al., 2009), or as calcium chloride (Benatti et al., 2009), which was found to be very effective in removal of sulfate. Removal of sulfate as calcium sulfate below 1300 mg/L was found to be very hard due to solubility limits. However addition of aluminum as aluminum sulfate or alum, aluminum chloride, aluminum nitrate (Christoe, 1976), and sodium aluminate (Batchelor et al, 1985) was found to enhance the removal of sulfate far below this value by formation of more complex solids (Batchelor et al, 1985; Christoe, 1976) which has much lower solubility compared to that of calcium sulfate precipitated by lime addition only.

#### **7.3 Secondary brine concentration and final salt production**

The treated brine after being further concentrated in secondary membrane desalination has to be further concentrated reaching zero liquid. Zero liquid and dry salts cannot be produced by membrane desalination such as RO or NF, and hence the concentrated brine has to be subjected to thermal process such as brine concentration followed by crystallization or drying to produce dry salts.

Thermal processes such as single and multi stage evaporators, or vapor compression evaporators are usually employed for further brine concentration. Such units have dual purpose objectives which are further recovery of water with very high quality with salinity about 10 mg/L, and brine concentration up to 250,000 mg/L, with recovery above 90%. Final salt production can be achieved after brine concentration which usually performed in crystallizers or dryers (Mickley, 2006).

Inland Desalination: Potentials and Challenges 471

requirements, and process economics, which should be given more attention and further

Precipitative softening processes have been widely used for treatment of primary brine stream, however softening process improvements through chemical doses optimization, testing different chemical reagents aiming at high efficiency in removal of scale forming constitutes should help improving the overall process performance. Furthermore other chemical treatment processes should be investigated which can result in better

ZLD system usually employs different units with different nature, such as membrane and thermal process, liquid and solids handling. Process development should look at the different viable and optimum units arrangement and operation conditions with the objective

Thermal processes are usually employed in ZLD systems for further brine concentration up to level that can be handled by crystallizer or dryer. Such processes are known to be energy extensive, and hence reduction in energy requirements and utilization of renewable energy

Economics of ZLD process is very important factor in employing the ZLD for inland desalination. Reaching competitive overall cost for inland desalination with ZLD to that of conventional desalination should help in wider application of the process for inland

In conclusion, as groundwater presents the main source of potable water to communities that do not have access to surface water, the deterioration of groundwater quality, specifically salinity is of high concern, which leads to the use of desalination techniques to overcome such problem. The use of membrane desalination systems in general and reverse osmosis in particular is very beneficial due to capacity flexibility, lower energy requirements, and in turn lower cost for brackish groundwater desalination. However the generation of brine stream is the main problem facing such systems, and which should be managed properly, there are different ways for brine disposal. However each one has certain advantages and disadvantages that are a matter of question. Approaching inland desalination with zero liquid discharge presents the solution for having a perfect inland desalination system. Given such need it is imperative to find a zero liquid discharge treatment technologies that provide more affordable concentrate management at reasonable cost, and hence a very active area of research is going on to provide such

Adhikary, S. K.; Narayanan, P. K.; Thampy,S. K.; Dave, N. J.; Chauhan, D. K.; & Indusekhar,

*Desalination*, Vol. 84, No. 1-3, (October 1991), pp. 189-200, ISSN: 0011-9164.

V. K. (1991). Desalination of Brackish Water of Higher salinity by Electrodialysis.

research and development efforts.

reducing energy requirements and cost.

should help in reducing overall energy requirements.

performance.

desalination systems.

**8. Conclusions** 

solution.

**9. References** 

In addition to the salts produced from final crystallizer/dryer that can be assumed as byproducts from ZLD desalination, the precipitated solids from brine chemical treatment can be considered as another byproduct or added value product. This precipitate is rich in calcium as calcium carbonate, magnesium as magnesium hydroxide, and silicate of calcium and magnesium, in addition to gypsum or calcium sulfoaluminate in case of sulfate removal, such mixture can find a wide range of applications such as road pavement, cement industry, and any other applications where there is a need for mixture of similar composition.

#### **7.4 Cost associated with zero liquid discharge systems**

Reaching inland desalination with zero liquid discharge has to be considered on both scales, technically and economically, while technical ZLD system can be successfully achieved through the different ZLD schemes. However the costs associated with each proposed ZLD system should be carefully considered. It easily noticeable that employing a secondary membrane desalination step is of high importance for reduction of both capital and operating costs over the conventional thermal ZLD systems due to the reduced volume of brine stream to be thermally treated.

A cost comparison for standard bench mark brine treatment by brine concentration and evaporation to advanced brine treatment using secondary RO desalination and final brine concentration and drying for brine of different qualities has been performed. The study showed that a cost reduction ranging from 48-67%, with reduction in energy requirement of 58 - 72% using the advanced ZLD system (Bond & Veerapaneni, 2008) can be achieved. However it worth to mention that the comparison was for the brine management only, not the whole inland desalination system, as the primary RO desalination is a kind of standard step employed for all ZLD schemes.

Inland desalination with zero liquid discharge usually has higher product water cost when compared to conventional inland desalination systems with no brine disposal is employed, but becomes very economically attractive when compared to the different brine disposal methods. The high cost mainly due to the fact that several units such as chemical treatment, secondary membrane desalination, brine concentration, and crystallization/drying are employed to recover only 15-30%. Which increase both capital and operating cost increasing the average product cost compared to single step desalination unit recovering 70-85 % (Greenlee et al., 2009). However due to the different strict regulation on brine disposal using the conventional methods, and the efforts for preserving the groundwater resources, more driving force for advancement of ZLD systems are encouraged (Mickley, 2006).

#### **7.5 Developments and research needs for zero liquid discharge desalination**

Desalination with zero liquid discharge is the ultimate achievement for any inland desalination process. This will help to overcome the brine disposal limitations currently faced for applying inland desalination. Although different zero liquid discharge schemes are currently developed or under development, however further development are needed to resolve its various technical, operational, and economical issues. The essential key for successful ZLD inland desalination are brine treatment, process development, energy

In addition to the salts produced from final crystallizer/dryer that can be assumed as byproducts from ZLD desalination, the precipitated solids from brine chemical treatment can be considered as another byproduct or added value product. This precipitate is rich in calcium as calcium carbonate, magnesium as magnesium hydroxide, and silicate of calcium and magnesium, in addition to gypsum or calcium sulfoaluminate in case of sulfate removal, such mixture can find a wide range of applications such as road pavement, cement industry, and any other applications where there is a need for mixture of similar

Reaching inland desalination with zero liquid discharge has to be considered on both scales, technically and economically, while technical ZLD system can be successfully achieved through the different ZLD schemes. However the costs associated with each proposed ZLD system should be carefully considered. It easily noticeable that employing a secondary membrane desalination step is of high importance for reduction of both capital and operating costs over the conventional thermal ZLD systems due to the reduced volume of

A cost comparison for standard bench mark brine treatment by brine concentration and evaporation to advanced brine treatment using secondary RO desalination and final brine concentration and drying for brine of different qualities has been performed. The study showed that a cost reduction ranging from 48-67%, with reduction in energy requirement of 58 - 72% using the advanced ZLD system (Bond & Veerapaneni, 2008) can be achieved. However it worth to mention that the comparison was for the brine management only, not the whole inland desalination system, as the primary RO desalination is a kind of standard

Inland desalination with zero liquid discharge usually has higher product water cost when compared to conventional inland desalination systems with no brine disposal is employed, but becomes very economically attractive when compared to the different brine disposal methods. The high cost mainly due to the fact that several units such as chemical treatment, secondary membrane desalination, brine concentration, and crystallization/drying are employed to recover only 15-30%. Which increase both capital and operating cost increasing the average product cost compared to single step desalination unit recovering 70-85 % (Greenlee et al., 2009). However due to the different strict regulation on brine disposal using the conventional methods, and the efforts for preserving the groundwater resources, more driving force for advancement of ZLD

**7.5 Developments and research needs for zero liquid discharge desalination** 

Desalination with zero liquid discharge is the ultimate achievement for any inland desalination process. This will help to overcome the brine disposal limitations currently faced for applying inland desalination. Although different zero liquid discharge schemes are currently developed or under development, however further development are needed to resolve its various technical, operational, and economical issues. The essential key for successful ZLD inland desalination are brine treatment, process development, energy

**7.4 Cost associated with zero liquid discharge systems** 

brine stream to be thermally treated.

step employed for all ZLD schemes.

systems are encouraged (Mickley, 2006).

composition.

requirements, and process economics, which should be given more attention and further research and development efforts.

Precipitative softening processes have been widely used for treatment of primary brine stream, however softening process improvements through chemical doses optimization, testing different chemical reagents aiming at high efficiency in removal of scale forming constitutes should help improving the overall process performance. Furthermore other chemical treatment processes should be investigated which can result in better performance.

ZLD system usually employs different units with different nature, such as membrane and thermal process, liquid and solids handling. Process development should look at the different viable and optimum units arrangement and operation conditions with the objective reducing energy requirements and cost.

Thermal processes are usually employed in ZLD systems for further brine concentration up to level that can be handled by crystallizer or dryer. Such processes are known to be energy extensive, and hence reduction in energy requirements and utilization of renewable energy should help in reducing overall energy requirements.

Economics of ZLD process is very important factor in employing the ZLD for inland desalination. Reaching competitive overall cost for inland desalination with ZLD to that of conventional desalination should help in wider application of the process for inland desalination systems.

#### **8. Conclusions**

In conclusion, as groundwater presents the main source of potable water to communities that do not have access to surface water, the deterioration of groundwater quality, specifically salinity is of high concern, which leads to the use of desalination techniques to overcome such problem. The use of membrane desalination systems in general and reverse osmosis in particular is very beneficial due to capacity flexibility, lower energy requirements, and in turn lower cost for brackish groundwater desalination. However the generation of brine stream is the main problem facing such systems, and which should be managed properly, there are different ways for brine disposal. However each one has certain advantages and disadvantages that are a matter of question. Approaching inland desalination with zero liquid discharge presents the solution for having a perfect inland desalination system. Given such need it is imperative to find a zero liquid discharge treatment technologies that provide more affordable concentrate management at reasonable cost, and hence a very active area of research is going on to provide such solution.

#### **9. References**

Adhikary, S. K.; Narayanan, P. K.; Thampy,S. K.; Dave, N. J.; Chauhan, D. K.; & Indusekhar, V. K. (1991). Desalination of Brackish Water of Higher salinity by Electrodialysis. *Desalination*, Vol. 84, No. 1-3, (October 1991), pp. 189-200, ISSN: 0011-9164.

Inland Desalination: Potentials and Challenges 473

Batchelor, B.; McDevitt, M.; Chan, D. (1985) Removal of Sulfate from Recycled Cooling

Bond, R.; Veerapaneni, S. (2008) Zeroing in on ZLD Technologies for Inland Desalination,

Brich, R. P.; Al-Arrayedh M.; Hallmans B. (1985) Bahrain's Fresh Groundwater Situation and

Brown, D. R. (1981) Treating Bahrain Zone C Groundwater using the EDR process, *Desalination*, Vol. 38, (November 1981), pp. 537-547, ISSN: 0011-9164. Bruggen; B., Vandecasteele, C. (2003) Removal of Pollutants from Surface Water and

Buros, O.K. (1999) The ABCs of Desalting, International Desalination Association, accessed 15.08.11, available from: http://www.idadesal.org/pdf/ABCs1.pdf Cheng, H.; Chen, S.; Yang, S. (2009) In-Line Coagulation/Ultrafiltration for Silica Removal

Christoe, J. R. (1976) Removal of Sulfate from Industrial Wastewaters, *Journal of Water* 

Comstock, S. E. H.; Boyer, T. H.; Graf, K. C. (2011) Treatment of Nanofiltration and Reverse

Delleur, J. W. (2nd Ed.). (2007). *The handbook of Groundwater Engineering*, CRC Press, ISBN 0-

Dore, M. H. I. (2005) Forecasting the Economic Costs of Desalination Technology, *Desalination*, Vol. 172, No. 3, (February 2005), pp. 207-214, ISSN: 0011-9164. El-Dessouky, H. T.; Ettouny, H. M.; Al-Juwayhel, F. (2000) Multiple Effect Evaporation –

Eriksson, P. (1988) Nanofiltration Extends the Range of Membrane Filtration, *Environmental* 

Ericsson, B.; Hallmans, B. (1996) Treatment of Saline Wastewater for Zero Discharge at the

*Design,* Vol. 78, No. 4, (May 2000), pp. 662-676, ISSN: 0263-8762.

*Progress.* Vol. 7, No. 1, (February 1988), pp. 58-62, ISSN: 0278-4491.

70, No. 13, (November 2009), pp. 112-117, ISSN: 1383-5866.

*AWWA Research Foundation,* pp. 798-812, Denver*,* Colorado, USA, 1985. Benatti, C. T.; Tavares, C. R. G.; Lenzi, E. (2009) Sulfate Removal from Waste Chemicals by

pp. 504-511, ISSN: 0301-4797.

89, ISSN: 1551-8833.

ISSN: 0269-7491.

0043-1303.

ISSN: 0043-1354.

8493-4316-x, Florida, USA.

115-123, ISSN: 0011-9164.

397-427, ISSN: 0011-9164.

Water by the Ultra-High Lime Process, *Proceedings water Reuse Symposium III,* 

Precipitation, *Journal of Environmental Management,* Vol. 90, No. 1, (January 2009),

*American Water Works Association Journal,* Vol. 100, No. 9, (September 2008), pp. 76-

the Investigations into Using Bahrain's Brackish water Resources as a Feedwater for the Reverse Osmosis Desalination Programme, *Desalination*, Vol. 55, No. 1, pp.

Groundwater by Nanofiltration: Overview of Possible Applications in the Drinking water Industry, *Environmental Pollution*, Vol. 122, No. 3, (April 2003), pp. 435-445,

from Brackish Water as RO pretreatment, *Separation and Purification Technology,* Vol.

*Pollution Control Federation,* Vol. 48, No. 12, (December 1976), pp. 2804-2804, ISSN:

Osmosis Concentrates: Comparison of Precipitative Softening, Coagulation, and anion Exchange, *Water Research*, Vol. 45, No. 16, (October 2011), pp. 4855-4865,

Vapor Compression Desalination Processes, *Chemical Engineering Research and* 

Debiensko Coal Mines in Poland, *Desalination,* Vol. 105, No. 1-2, (June 1996), pp.


http://www.fao.org/DOCREP/003/T0234E/T0234E00.htm#TOC

Bart Weiss, P.G. (2002). Brackish Groundwater Desalination Water Supply, *Proceedings of AMTA 2002 Biennial Conference and Exposition "Water Quality Enhancement through Membrane Technology"*, pp. 0-44, Tampa, Florida, August 6-9, 2002.

Ahmed, M.; Shayya, W. H.; Hoey, D.; Mahendran, A.; Morris, R.; Al-Handaly, J. (2000). Use

Ahmed, M.; Shayya, W. H.; Hoey, D.; Al-Handaly, J. (2001) Brine Disposal from Reverse

Ahmed, M.; Shayya, W. H.; Hoey, D.; Al-Handaly, J. (2002) Brine Disposal from Inland Desalination Plants, *Water International,* Vol. 27, No. 2, (June 2002), pp. 194 -201. Ahuja, N.; Howe, K. (2005). Strategies for Concentrate Management from Inland

Altmann and Water Cogeneration, T. (1997) A New Power Concept with the Application of

Al-Ahmad, M.; Abdil Aleem, F. (1993) Scale Formation and Fouling Problems Effect on the

Al-Faifi, H.; et al. (2010) Soil Deterioration as Influenced by Land Disposal of Reject Brine

Al-Karaghouli, A.; Renne, D.; Kazmerski, L. L. (2009) Solar and Wind Opportunities for

Al-Karaghouli, A.; Renne, D.; Kazmerski, L. L. (2010) Technical and Economic Assessment

Al-Rehaili, A. M. (2003) Comparative Chemical Clarification for Silica Removal from RO

Al-Shammiri, M.; Safar, M.; Al-Dawas, M. (2000) Evaluation of Two Different Antiscalents in

Al-Shammiri, M.; Ahmed, M.; Al-Rageeb, M. (2004) Nanofilteration and Calcium Sulfate

Amer, K. M.; Al-Muraikhi, A.; Rashid, N. (2008) Management of Coastal Aquifers- the Case

Ayers, R.S.; Westcot, D.W., (1994), Water Quality for Irrigation 29 Rev.1, accessed 15.08.11,

Bart Weiss, P.G. (2002). Brackish Groundwater Desalination Water Supply, *Proceedings of* 

*Membrane Technology"*, pp. 0-44, Tampa, Florida, August 6-9, 2002.

http://www.fao.org/DOCREP/003/T0234E/T0234E00.htm#TOC

130, No. 2, (November 2000), pp. 155-168, ISSN: 0011-9164.

Vol. 133, No. 2, (March 2001),pp. 135-147, ISSN: 0011-9164.

761-776, Phoenix, Arizona, USA, March 6-9, 2005.

93, No. 1-3, (August 1993), pp. 287-310, ISSN: 0011-9164.

250, No. 2, (January 2010), pp. 479-484, ISSN: 0011-9164.

(February 2010), pp. 323-328, ISSN: 0960-1481.

2000), pp. 1-16, ISSN: 0011-9164.

167, No. 2, pp. 335-346, ISSN: 0011-9164.

Vol. 13, No. 9, (December 2009), pp. 2397-2407, ISSN: 1364-0321.

pp.139-144, ISSN: 0011-9164.

0011-9164.

23-27, 2008.

available from:

of Evaporation Ponds for Brine Disposal in Desalination Plants, *Desalination,* Vol.

Osmosis Desalination Plants in Oman and the United Arab Emirates, *Desalination,* 

Desalination, *Proceeding of Membrane Technology Conference & Exposition 2005"*, pp.

Reverse osmosis Desalination, *Desalination*, Vol. 114, No. 2 , (December 1997),

Performance of MSF and RO Desalination Plants in Saudi Arabia, *Desalination,* Vol.

from Salbukh Water Desalination Plant at Riyadh, Saudi Arabia, *Desalination,* Vol.

Water Desalination in the Arab Regions, *Renewable and Sustainable Energy Reviews*,

of Photovoltaic-Driven Desalination Systems, *Renewable Energy*, Vol. 35, No. 2,

Groundwater Feed, *Desalination,* Vol. 159, No. 1, (September 2003), pp. 21-31, ISSN:

Real Operation at the Doha Research Plant, *Desalination*, Vol. 128, No. 1, (March

limitation for Top Brine Temperature in Gulf Desalination Plants, *Desalination*, Vol.

of a Peninsula- State of Qatar; *20th Salt water Intrusion Meeting,* Florida, USA, June

*AMTA 2002 Biennial Conference and Exposition "Water Quality Enhancement through* 


Inland Desalination: Potentials and Challenges 475

Hadi, Khaled M. B. (2002) Evaluation of the Suitability of Groundwater Quality for Reverse

Hamed, O. A.; Hassan, A. M.; Al-Shail, K.; Farooque, M. A. (2009) Performance Analysis of a

Hamrouni, B.; Dhahbi, M. (2001) Thermodynamics Description of Saline Waters-Prediction

Hamrouni, B.; Dhabi, M. (2001) Analytical Aspects of Silica in Saline Waters- Application to

Hassan A. M., et al. (1998) A New Approach to Membrane and Thermal Seawater

Harkare, W. P., et al. (1982) Desalination of Brackish Water by Electrodialysis, *Desalination*,

International Desalination Association, (2009). Total World Desalination Capacity close to 60

Karagiannis, I. C.; Soldatos, P. G. (2008) Water Desalination Cost Literature: Review and

Katz, W. E. (1979) The Electrodialysis Reversal (EDR) Process, *Desalination*, Vol. 28, No. 1,

Katz, W. E. (1982) Desalination by ED and EDR – State-of-the-Art in 1981, *Desalination*, Vol.

Kedem, O.; Zalmon, G. (1997) Compact Accelerated Precipitation Softening (CAPS) as a

Khawaji, A. D.; Kutubkhanah, I. K.; Wie, J. (2008) Advances in Seawater Desalination

Kim, D. H. (2011) A Review of Desalting Process Techniques and Economic Analysis of the

Lee, K. P.; Arnot, T. C.; Mattia, D. (2011) A Review of Reverse Osmosis Mambrane Material

Lee, S.; et al. (2009) Toward a Combined System of Forward Osmosis and Reverse Osmosis

*Science*, Vol. 370, No. 1-2, (March 2011), pp. 1-22, ISSN: 0376-7388.

No. 1-3, (January 2009), pp. 215-222, ISSN: 1944-3994.

118, No. 1-3, (September 1998), pp. 35-51, ISSN: 0011-9164.

Vol. 42, No. 1, (July 1982), pp. 97-105, ISSN: 0011-9164.

http://www.desalination.biz/news/news\_story.asp?id=5121

42, No. 2, (August 1982), pp. 129-139, ISSN: 0011-9164.

113, No. 1, (November 1997), pp. 65-71, ISSN: 0011-9164.

*Newsletter,* accessed 15.08.11, available from:

(January 1979), pp. 31-40, ISSN: 0011-9164.

2001), pp. 275-284, ISSN: 0011-9164.

225-232, ISSN: 0011-9164.

0011-9164.

9164.

9164.

8, ISSN: 0011-9164.

246, ISSN: 0011-9164.

Osmosis Desalination, *Desalination*, Vol. 142, No. 3, (March 2002), pp. 209-219, ISSN:

Trihybrid NF/RO/MSF Desalination Plant, *Desalination and Water Treatment*, Vol. 1,

of Scaling Limits in Desalination Processes, *Desalination*, Vol. 137, No. 1-3, (May

Desalination of Brackish Waters, *Desalination*, Vol. 136, No. 1-3, (May 2001), pp.

Desalination Processes using Nanofiltration membranes (Part1), *Desalination*, Vol.

million m3/d. In The *international Desalination and water Reuse Quarterly Industry* 

Assessment, *Desalination*, Vol. 223, No. 1-3, (March 2008), pp. 448-456, ISSN: 0011-

Pretreatment for Membrane Desalination I. Softening by NaOH, *Desalination,* Vol.

Technologies, *Desalination*, Vol. 221, No. 1-3, (March 2008), pp. 47-69, ISSN: 0011-

Recovery of Salts from Retenates, *Desalination,* Vol. 270, No. 1-3, (April 2011), pp. 1-

for Desalination – Development to Date and Future Potential, *Journal of Membrane* 

for Seawater Desalination, *Desalination*, Vol. 247, No. 1-3, (October 2009), pp. 239-


http://www.fao.org/nr/water/aquastat/main/index.stm


Faust, S. D.; Aly, O. M. (1998), *Chemistry of Water Treatment* 2nd Ed., Lewis Publishers CRC

Food and Agriculture Organization of the United Nations FAO, AQUASTAT Program,

Forstmeier, M; et al. (2007) Feasibility Study on Wind–Powered Desalination, *Desalination*,

Freeman, S. D.; Majerle, R. J. (1995) Silica Fouling Revisited, *Desalination*, Vol. 103, No. 1-2,

Fritzmann, C.; Lowenberg, J., Wintgens, T; Melin, T. (2007) State-of-the-Art of Reverse

Fubao , Y. (1985) Study on Electrodialysis reversal (EDR) Process, *Desalination*, Vol. 56, pp.

Fujiwara, N.; Matsuyama, H. (2008) Elimination of Biological Fouling in Seawater Reverse

Gabbrielli, E. (1981) A Tailored Process for Remineralization and Potabilization of Salinated Water, *Desalination*, Vol. 39, (December 1981), pp. 503-520, ISSN: 0011-9164. Gabelich, C. J.; Williams, M. D.; Rahardianto, A.; Franklin, J. C.; Cohen, Y. (2007) High

Gabelich, C. J.; Rahardianto, A.; Northrup, C. R.; Yun, T. I.; Cohen, Y. (2011) Process

Gilron, J.; Daltrophe, N.; Waissman, M.; Oren, Y. (2005) Comparison Between Compact

Gleick, P. H. (2001). *The World's water 2000-2001: The Biennial Report on Freshwater Resources*,

Gleick, P. H. (2006). *The World's water 2006-2007: The Biennial Report on Freshwater Resources*,

Gorenflo, A.; Velazques-Padron, D.; Frimmel, F. H. (2002) Nanofiltration of a German

*Desalination*, Vol. 151, No. 3, (January 2003), pp. 253-265, ISSN: 0011-9164. Greenlee, L. F.; Lawler, D. F.; Freeman, B. D.; Marrot, B.; Moulin, P. (2009) Reverse Osmosis

Groundwater of High hardness and NOM content: Performance and Costs,

Desalination: Water sources, Technology, and Today's Challenges, *Water Research*,

Osmosis Desalination, *Desalination*, Vol. 216, No. 1-3, (October 2007), pp. 1-76, ISSN:

Osmosis Desalination Plants, *Desalination,* Vol. 227, No. 1-3, (July 2008), pp. 295-305,

Recovery Reverse Osmosis Desalting Using Intermediate Chemical Demineralization, *Journal of Membrane Science,* Vol. 301, No. 1-2, (September 2007),

Evaluation of Intermediate Chemical Demineralization for Water Recovery Enhancement in Production-Scale Brackish Water Desalting, *Desalination,* Vol. 272,

Accelerated Precipitation Softening (CAPS) and Conventional Pretreatment in Operation of Brackish Water Reverse Osmosis (BWRO), *Industrial Engineering and Chemistry Research,* Vol. 44, No. 15, (June 2005), pp. 5465-5471, ISSN: 0888-

Vol. 203, No. 1-3, (February 2007), pp. 463-470, ISSN: 0011-9164.

Press LLC, ISBN 1-57504-011-5, Florida, USA.

http://www.fao.org/nr/water/aquastat/main/index.stm

(November 1995), pp. 113-115, ISSN: 0011-9164.

No. 1-3, (May 2011), pp. 36-45, ISSN: 0011-9164.

Island Press, ISBN 1-55963-792-7, Washington, DC, USA.

Island Press, ISBN 1-59726-108-8, Washington, DC, USA.

Vol. 43, No. 9, (May 2009), pp. 2317-2348, ISSN: 0043- 1354.

(n.d.). Accessed 13.08.11, available from:

0011-9164.

5885.

ISSN: 0011-9164.

315-324, ISSN: 0011-9164.

pp. 131-141, ISSN: 0376-7388.


http://www.desalination.biz/news/news\_story.asp?id=5121


Inland Desalination: Potentials and Challenges 477

Neofotistou, E.; Demadis, K. D. (2004) Use of antiscalents for Mitigation of Silica (SiO2)

Nicot, J.; Chowdhury, A. H. (2005) Disposal of Brackish water Concentrate into Depleted Oil

Ning, R. Y. (2002) Discussion of Silica Speciation, Fouling, Control and Maximum

Ning, R. Y.; Tarquin, A.; Trzcinski, M. C.; Patwardhan, G. (2006) Recovery Optimization of

Ning, R. Y.; Troyer, T. L. (2009) Tandom Reverse Osmosis Process for Zero-Liquid

Oren, Y.; Katz, V.; Daltrophe, N. (2001) Improved Compact Accelerated Precipitation

Oren, Y; et al, (2010) Pilot Studies on High Recovery BWRO-EDR for Near Zero Liquid

Phillip, W. A.; Yong, J. S.; Elimelech, M. (2010) Reverse Draw Solute Permeation in Forward

Rahardianto, A.; Gao, J.; Gabelich, C. J.; Williams, M. D.; Cohen, Y. (2007) High Recovery

Rahardianato, A; Mccool, B. C.; Cohen, Y. (2008) Reverse Osmosis Desalting of Inland

Rautenbach, R; Groschl, A. (1990) Separation Potential of Nanofiltration membranes,

Reverter, J. A; Talo, S.; Alday, J. (2001) Las Palmas III – the Success Story of Brine Staging, *Desalination,* Vol. 138, No. 1-3, (September 2001), pp. 207-217, ISSN: 0011-9164. Reynolds T. D.; Richards P. A., (1996), *Unit Operations and Processes in Environmental* 

Sadrzadeh, M.; Mohammadi, T. (2008) Sea Water Desalination Using Electrodialysis, *Desalination*, Vol. 221, No. 1-3, (March 2008), pp. 440-447, ISSN: 0011-9164.

*Desalination*, Vol. 77, (March 1990), pp. 73-84, ISSN: 0011-9164.

No. 13, (July 2010), pp. 5170-5176, ISSN: 0013-936X.

No. 1-2, (February 2007), pp. 123-137, ISSN: 0376-7388.

(June 2008), pp. 4292-4297, ISSN: 0013-936X.

*Desalination*, Vol. 167, (August 2004), pp. 257-272, ISSN: 0011-9164.

ISSN: 1061-4303.

61-74, ISSN: 0011-9164.

2006), pp. 315-322, ISSN: 0011-9164.

9164.

9164.

ISSN: 0011-9164.

ISSN: 0011-9164.

Massachusetts, USA.

Treatment, *Water Environment Research,* Vol. 82, No. 4, (April 2010), pp. 342-350,

Fouling and Deposition: Fundamentals and Applications in Desalination Systems,

and Gas Fields: a Texas Study, *Desalination,* Vol. 181, No. 1-3, (September 2005), pp.

Reduction, *Desalination*, Vol. 151, No. 1, (January 2003), pp. 67-73, ISSN: 0011-

RO Concentrate from Desert Wells, *Desalination,* Vol. 201, No. 1-3, (November

Discharge, *Desalination,* Vol. 237, No. 1-3, (February 2009), pp. 238-242, ISSN: 0011-

Softening (CAPS), *Desalination,* Vol. 139, No. 1-3, (September 2001), pp. 155-159,

Discharge Approach, *Desalination*, Vol. 261, No. 3, (October 2010), pp. 321-330,

Osmosis: Modeling and Experiments, *Environmental Science and Technology*, Vol. 44,

Membrane Desalting of Low-Salinity Brackish Water: Integration of Accelerated Precipitation Softening with Membrane RO, *Journal of Membrane Science,* Vol. 289,

Brackish water of High Gypsum Scaling Propensity: Kinetics and Mitigation of Membrane Mineral Scaling, *Environmental Science and Technology*, Vol. 42, No. 12,

*Engineering*, PWS Publishing Company, ISBN 0-534-94884-7, Boston,


Lindsay, F. K.; Ryznar, J. W. (1939) Removal of Silica from Water by Sodium Aluminate,

Mabrouk, A. A.; Nafey, A. S.; Fath, H. E. S. (2007) Analysis of a New Design of a Multi-

Mace, R. E.; Nicot, J.; Chowdhury, A. H.; Dutton, A. R.; Kalaswad, S., (2006), Please Pass the

Masarawa, A.; Meyerstein, D.; Daltrophe, N.; Kedem, O. (1997) Compact Accelerated

*Desalination,* Vol. 113, No. 1, (November 1979), pp. 73-84, ISSN: 0011-9164. Mathioulakis, E.; Belessiotis, V.; Delyannis, E. (2007) Desalination by Using Alternative

Martinetti, C. R.; Childress, A. E.; Cath, T. Y. (2009) High Recovery of Concentrated RO

McCutcheon J. R.; McGinnis, R. L.; Elimelech, M. (2005a) A novel Ammonia-Carbon Dioxide

McCutcheon J. R.; McGinnis, R. L.; Elimelech, M. (2005b) Desalination by Ammonia-Carbon

McGinnis, R. L.; Elimelech, M. (2007) Energy Requirements of Ammonia-Carbon Dioxide

Mickley, M. C. (2001) Membrane Concentrate Disposal: Practice and Regulations. U.S. Department of the Interior, Bureau of Reclamation, Mickley & Associates. Mickley, M. C., (2006) Membrane Concentrate Disposal: Practice and Regulation,

Missimer T.M. (1994) Groundwater as a Feedwater Source for Membrane Treatment Plants:

Mohamed, A. M. O.; Maraqa, M.; Al-Handaly, J (2005) Impact of Land Disposal of Reject

Mohammadesmaeili, F.;Badr, M. K.; Abbaszadegan, M.; Fox, P. (2010) Byproduct Recovery

98, No. 1-3, (September 1994), pp. 451-457, ISSN: 0011-9164.

1-3, (November 2005), pp. 411-433, ISSN: 0011-9164.

*Science*, Vol. 331, No. 1-2, (April 2009), pp. 31-39, ISSN: 0376-7388.

204, No. 1-3, (February 2007), pp. 482-500, ISSN: 0011-9164.

Texas Water Development Board, Report No. 366, April 2006.

2007), pp. 346-365, ISSN: 0011-9164.

2005), pp. 1-11, ISSN: 0011-9164.

114-123, ISSN: 0376-7388.

370-382, ISSN: 0011-9164.

Treatment Engineering and Research Group.

0888-5885.

*Industrial and Engineering Chemistry,* Vol. 31, No. 7, (July 1939), pp. 859-861, ISSN:

Stage-Flash-Mechanical Vapor Compression Desalination Process, *Desalination*, Vol.

Salt: Using Oil Fields for the Disposal of Concentrate from Desalination Plants,

Precipitation Softening (CAPS) as a Pretreatment for Membrane Desalination II. Lime Softening with Contaminant Removal of Silica and Heavy Metals,

Energy: Review and State-of-the-Art, *Desalination*, Vol. 203, No. 1-3, (February

brines using Forward Osmosis and Membrane Distillation, *Journal of Membrane* 

Forward (Direct) Osmosis Desalination Process, *Desalination*, Vol. 174, No. 1, (April

Dioxide Forward Osmosis: Influence of Draw and Feed Solution Concentrations on Process Performance, *Journal of Membrane Science*, Vol. 278, No. 1-2, (July 2006), pp.

Forward Osmosis Desalination, *Desalination*, Vol. 207, No. 1-3, (March 2007), pp.

Desalination and Water Purification Research and Development Program Report No. 123 2nd Ed., April 2006 (Agreement No. 98-FC-81-0054), Prepared for the U.S. Department of the Interior, Bureau of Reclamation, Technical Service Center, Water

Hydrogeologic Controls on Water Quality Variations with Time, *Desalination*, Vol.

Brine from Desalination Plant on Soil and Groundwater, *Desalination,* Vol. 182, No.

from Reclaimed Water Reverse Osmosis Concentrate Using Lime and Soda-Ash

Treatment, *Water Environment Research,* Vol. 82, No. 4, (April 2010), pp. 342-350, ISSN: 1061-4303.


Inland Desalination: Potentials and Challenges 479

Tait, S.; Clarke, W. P.; Keller, J.; Batstone, D. J. (2009) Removal of Sulfate from High-Strength

Tang, W.; Ng, H. Y. (2008) Concentration of Brine by Forward Osmosis: Performance and

The Federal Institute for Geosciences and Natural Resources, (n.d.). Groundwater Resources

http://www.bgr.bund.de/EN/Themen/Wasser/Bilder/Was\_wasser\_startseite\_g

United Nations UN, (1997). UN Assessment of Freshwater Resources 1997, accessed

United States Environmental Protection Agency, US EPA, (2009) National Primary & Secondary Drinking Water Regulations, accessed 15.08.11, available at:

United Nations Environmental Programme UNEP- GRiD, (n.d.). Increased Global Water

 http://www.grida.no/graphic.aspx?f=series/vg-water2/0400-waterstress-EN.jpg Valcour, H. C. (1985) Recent Applications of EDR, *Desalination*, Vol. 54, pp. 163-183, ISSN:

Vandercaseele, C.; Van Der Bruggen, B. (2002) Distillation vs. Membrane Filtration:

Vince, F.; Marechal, F.; Aoustin, E.; Bréant, P. (2008) Multi-Objective Optimization of RO

Walha, K.; Ben Amar, A.; Firdaous, L.; Quemeneur, F.; Jaouen, P. (2007) Brackish

Wang, L. K (2008), *Handbook of Environmental Engineering V.13: Membrane and Desalination* 

WHO (2011) Guidelines for Drinking-Water Quality, 4th Ed. Accessed 15.08.11, available

Yangali-Quintanilla, V.; Li, Z.; Valladares, R.; Li, Q.; Amy G. (2011) Indirect Desalination of

Yun, T. I; Gabelich, C. J.; Cox, M. R.; Mofidi, A. A.; Lesan, R. (2006) Reducing Costs for

*Desalination*, Vol. 189, No. 1-3, (March 2006), pp. 141-154, ISSN: 0011-9164. Zubari, W. K. (2003) Assessing the Sustainability of Non-renewable Brackish Groundwater

*Technologies*, Humana Press, ISBN 1-58829-940-6, New York, USA.

from: http://water.epa.gov/drink/contaminants/index.cfm

water reuse, *Desalination*, doi:10.1016/j.desal.2011.06.066.

(November 2003), pp. 211-244, ISSN: 0011-9164.

Overview of Process Evolutions in Seawater Desalination, *Desalination*, Vol. 143,

Desalination Plants, *Desalination*, Vol. 222, No. 1-3, (March 2008), pp. 96-118, ISSN:

Groundwater Treatment by nanofiltration, Reverse Osmosis, and Electrodialysis in Tunisia: Performance and cost comparison, *Desalination*, Vol. 207, No. 1-3, (March

Red Sea Water with Forward Osmosis and Low Pressure Reverse Osmosis for

Large-Scale Desalting Plants Using Large-Diameter Reverse Osmosis Membranes,

in Feeding an RO Desalination Plant in Bahrain, *Desalination*, Vol. 159, No. 3,

762-772, ISSN: 0043-1354.

143-153, ISSN: 0011-9164.

13.08.11, available from:

0011-9164.

0011-9164.

w\_erde\_g\_en.html?nn=1548136

of the World, accessed 15.08.11, available from:

http://water.epa.gov/drink/contaminants/index.cfm

No. 3, (June 2002), pp. 207-218, ISSN: 0011-9164.

Stress, accessed 15.08.11, available from:

2007), pp. 95-106, ISSN: 0011-9164.

http://www.un.org/ecosocdev/geninfo/sustdev/waterrep.htm

Wastewater by Crystallization, *Water Research,* Vol. 43, No. 3, (February 2009), pp.

influence of membrane structure, *Desalination*, Vol. 224, No. 1-3, (April 2008), pp.


Saitua, H.; Gil, R.; Padilla, A. P. (2011) Experimental Investigations on Arsenic Removal with

Schaep, J; et al. (1998) Removal of Hardness from Groundwater by Nanofiltration, *Desalination*, Vol. 119, No. 1-3, (September 1998), pp. 295-301, ISSN: 0011-9164. Semiat, R.; Sutzkover, I.; Hasson, D. (2003) Scaling of RO Membranes from Silica

Semiat, R.; Hasson, D.; Zelmanov, G.; Hemo, I. (2004) Threshold Scaling Limits of RO

Seto, T., et al. (1978) Seawater Desalination by Electrodialysis, *Desalination*, Vol. 25, No. 1,

Sheikholesami, R.; Tan, S. (1999) Effects of Water Quality on Silica Fouling of Desalination

Sheikholesami, R.; Al-Mutaz, I. S.; Koo, T.; Young, A. (2001) Pretreatment and the Effect of

Sheikholesami, R; Bright, J. (2002) Silica and Metals Removal by Pretreatment to Prevent

Sheikholesami, R. (2003a) Mixed salts – Scaling limits and Propensity, *Desalination*, Vol. 154,

Sheikholesami, R. (2003b) Nucleation and Kinetics of Mixed Salts in Scaling, *AIChE Journal*,

Sheikholesami, R. (2004) Assessment of the Scaling Potential for Sparingly Soluble Salts in

Sheikholesami, R. (2009) Strategies for Future Research and Development in Desalination-

Sherif, M.; Mohamed, M.; Kacimov, A.; Shetty, A. (2011) Assessment of Groundwater

Skehan, S; Kwiatkowski P. J. (2000) Concentrate Disposal via Injection Wells – Permitting

Tahaikt, M., et al. (2007) Fluoride Removal from Groundwater by Nanofiltration, *Desalination*, Vol. 212, No. 1-3, (June 2007), pp. 46-53, ISSN: 0011-9164.

*Desalination*, Vol. 273, No. 2-3, (June 2011), pp. 436-446, ISSN: 0011-9164. Shon, H. K.; et al. (2009) Physicochemical Pretreatment of Seawater: Fouling Reduction and

274, No. 1-3, (July 2011), pp. 1-6, ISSN: 0011-9164.

*Technology*, Vol. 49, No. 2, pp. 211-219, ISSN: 0273-1223.

(March 1978), pp. 1-7, ISSN: 0011-9164.

(September 2001), pp. 83-95, ISSN: 0011-9164.

No. 2, (April 2003), pp. 117-127, ISSN: 0011-9164.

Vol. 49, No. 1, (January 2003), pp. 194-202, ISSN: 001-1541.

pp. 255-267, ISSN: 0011-9164.

191, ISSN: 0011-9164.

9164.

9164.

ISSN: 0011-9164.

10-21, ISSN: 0011-9164.

2000), pp. 19-21, ISSN: 0896-1794.

a Nanofiltration Pilot Plant from Contaminated Groundwater, *Desalination*, Vol.

Supersaturated Solutions, *Desalination*, Vol. 157, No. 1-3, (August 2003), pp. 169-

Concentrates Flowing in a Long Waste Disposal Pipeline, *Water Science and* 

plants, *Desalination*, Vol. 126, No. 1-3, (November 1999), pp. 267-280, ISSN: 0011-

Cations and Anions on Prevention of Silica Fouling, *Desalination*, Vol. 139, No. 1-3,

Fouling of Reverse Osmosis Membranes, *Desalination,* Vol. 143, No. 3, (June 2002),

RO and NF units, *Desalination*, Vol. 167, (August 2004), pp. 247-256, ISSN: 0011-

challenges ahead, *Desalination*, Vol. 248, No. 1-3, (November 2009), pp. 218-224,

Quality in the Northeastern Coastal area of UAE as Precursor for Desalination,

Membrane Characterization, *Desalination,* Vol. 238, No. 1-3, (November 2009), pp.

and Design Considerations, *Florida Water Resources Journal*, Vol. 52, No. 5, (May


http://www.un.org/ecosocdev/geninfo/sustdev/waterrep.htm


http://www.grida.no/graphic.aspx?f=series/vg-water2/0400-waterstress-EN.jpg


**Part 4** 

**Separation Technology** 

Zubari, W. K.; Madany, I. M.; Al-Junaid, S. S.; Al-Manaii, S. (1994) Trends in the quality of Groundwater in Bahrain with Respect to Salinity, 1941-1992., *Environment International*, Vol. 20, No. 6, pp. 739-746, ISSN: 0160-4120.

**Part 4** 

**Separation Technology** 

480 Advances in Chemical Engineering

Zubari, W. K.; Madany, I. M.; Al-Junaid, S. S.; Al-Manaii, S. (1994) Trends in the quality of

*International*, Vol. 20, No. 6, pp. 739-746, ISSN: 0160-4120.

Groundwater in Bahrain with Respect to Salinity, 1941-1992., *Environment* 

**19** 

**Phase Diagrams in Chemical** 

**Distillation and Solvent Extraction** 

Christophe Coquelet1,2 and Deresh Ramjugernath1

*CEP/TEP - Centre Énergétique et Procédés, Fontainebleau* 

By definition, a phase diagram in physical chemistry and chemical engineering is a graphical representation showing distinct phases which are in thermodynamic equilibrium. Since these equilibrium relationships are dependent on the pressure, temperature, and composition of the system, a phase diagram provides a graphical visualization of the effects of these system variables on the equilbrium behavior between the phases. Phase diagrams are essential in the understanding and development of separation processes, especially in the choice and design of separation unit operations, e.g. knowledge about high pressure phase equilibria is essential not just in chemical processes and separation operations, but is also important for the simulation of petroleum reservoirs, the transportation of petroleum fluids, as well as in the refrigeration industry. In order to utilize the knowledge of phase behavior it is important to represent or correlate the phase information via the most accurate thermodynamic models. Thermodynamic models enable a mathematical representation of the phase diagram which ensures comprehensive and reproducible production of phase diagrams. The measurement of phase equilibrium data is necessary to develop and refine thermodynamic models, as well as to adjust them by fitting or correlating their parameters to experimental data. Generally the measurement of phase equilibria is undertaken using two categories of experimental techniques, viz. synthetic and analytic methods. The choice of the technique depends on the type of data to be determined, the range of temperatures and pressures, the precision required,

and also the order of magnitude of the phase concentrations expected.

**2. Definition of phases, phase transitions, and equilibrium** 

A phase is a homogeneous space, which can be composed of one or more components or chemical species, in which the thermodynamic properties, e.g. density or composition, are

**1. Introduction** 

**Engineering: Application to** 

*1Thermodynamic Research Unit, School of Chemical Engineering, University KwaZulu Natal, Howard College Campus, Durban* 

*2MINES ParisTech,* 

*1South Africa 2France* 

### **Phase Diagrams in Chemical Engineering: Application to Distillation and Solvent Extraction**

Christophe Coquelet1,2 and Deresh Ramjugernath1 *1Thermodynamic Research Unit, School of Chemical Engineering, University KwaZulu Natal, Howard College Campus, Durban 2MINES ParisTech, CEP/TEP - Centre Énergétique et Procédés, Fontainebleau 1South Africa 2France* 

#### **1. Introduction**

By definition, a phase diagram in physical chemistry and chemical engineering is a graphical representation showing distinct phases which are in thermodynamic equilibrium. Since these equilibrium relationships are dependent on the pressure, temperature, and composition of the system, a phase diagram provides a graphical visualization of the effects of these system variables on the equilbrium behavior between the phases. Phase diagrams are essential in the understanding and development of separation processes, especially in the choice and design of separation unit operations, e.g. knowledge about high pressure phase equilibria is essential not just in chemical processes and separation operations, but is also important for the simulation of petroleum reservoirs, the transportation of petroleum fluids, as well as in the refrigeration industry. In order to utilize the knowledge of phase behavior it is important to represent or correlate the phase information via the most accurate thermodynamic models. Thermodynamic models enable a mathematical representation of the phase diagram which ensures comprehensive and reproducible production of phase diagrams. The measurement of phase equilibrium data is necessary to develop and refine thermodynamic models, as well as to adjust them by fitting or correlating their parameters to experimental data. Generally the measurement of phase equilibria is undertaken using two categories of experimental techniques, viz. synthetic and analytic methods. The choice of the technique depends on the type of data to be determined, the range of temperatures and pressures, the precision required, and also the order of magnitude of the phase concentrations expected.

#### **2. Definition of phases, phase transitions, and equilibrium**

A phase is a homogeneous space, which can be composed of one or more components or chemical species, in which the thermodynamic properties, e.g. density or composition, are

Phase Diagrams in Chemical Engineering: Application to Distillation and Solvent Extraction 485

*First order:* In this type of transition there is a discontinuity in the first derivative of the Gibbs free energy with regard to thermodynamic variables. For this type of transition there is an absorption or release of a fixed amount of energy with the temperature remaining constant. *Second order:* For this type of transition, there is continuity in the first derivative of the Gibbs free energy with regard to the thermodynamic variable across the transition, but there is a discontinuity in the second derivative of the Gibbs energy with regard to thermodynamic

The chemical potential is the one of the most important thermodynamic properties used in the description of phase equilibrium. If one considers a phase with volume V, containing nc

1 2

Where G=H-TS is the Gibbs free energy of the phase. The expression for the infinitesimal

*dG VdP SdT dn*

*i VdP SdT n d*

Considering a multicomponent system in equilibrium between two phases (

   and 

*dG dG dG dn dn*

 

temperature T and pressure P, the Gibbs free energy of this system is *G=G*

, , , ,..., , , , ,...,

, ,

*C*

*G PTn n n*

*i*

*n*

*i i*

*Nc i i <sup>i</sup> G n*

0 *i i i i*

 

> 

 

(5)

 

*i i*

for each component *i* between phases

 

 

, , , ,..., 1 2 , , , ,..., 1 2 *<sup>C</sup> <sup>C</sup> i N <sup>i</sup> <sup>N</sup> PTn n n PTn n n*

 

  0 *i i*

*i TPn*

*j i*

(2)

(3)

and

 *+G* ), at

. For each

, where *G*

. The Gibbs-Duhem (Eq. 3)

phases respectively. At equilibrium, the

(4)

 and 

*<sup>i</sup>* of component i in

(1)

There are two types of phase transition:

**2.2 Chemical potentials and equilibrium conditions**

components at temperature T and pressure P, the chemical potential

1 2

*i C*

reversible change in the Gibbs free energy is given by

Moreover, from Euler theorem we can write <sup>1</sup>

equation can be obtained from equations 1 and 2:

, are the Gibbs free energy of the

 = - *<sup>i</sup> dn*

chemical species, the following relationship can written:

Gibbs free energy must be at the minimum, therefore (*dG=0*):

 

**2.2.2 Equilibrium conditions**

For a closed system, *<sup>i</sup> dn*

and *G* *PTn n n*

properties.

**2.2.1 Chemical potential**

the phase is defined by

identical. A system can comprise of one or several phases. Depending on whether one or more phases exist, the system can be defined as a monophasic homogenous system in which the composition and thermodynamic properties are identical in the whole space, or as a multiphase heterogenous system for which thermodynamic properties change distinctly at the phase interface. A phase is characterized by its temperature, density, pressure, and other thermodynamics properties, i.e. Gibbs energy, molar enthalpy, entropy, heat capacity, etc*.* 

The concept of a phase can be employed to distinguish between the different states of matter. Matter is generally accepted to exist in three main states, viz. gas, liquid, and solid. Other types of phases do exist, e.g. nematic and smectic phase transitions in liquid crystral, but this will not be relevant to this chapter. Molecular interactions between the components or chemical species which comprise the system are responsible for the different states of matter. Consequently thermodynamics model used to describe phases have to take into account the molecular interactions, or be used at conditions at which they are relevant, e.g. the ideal gas model can only be applied to gas phase at low pressures and sufficiently high temperatures, as at these conditions molecular interactions are negligible.

#### **2.1 Phases transitions**

There a number of definitions that can be used to describe a liquid. One of the definitions, which is easy to visualize but may not be entirely thermodynamically correct, is that a liquid is a fluid which takes the form of its containment without necessarily filling the entire containment volume. This characteristic distinguishes a liquid from a solid, as well as a liquid from a gas. As stated, it is not an accurate definition and it is necessary to carefully check the thermophysical microscopic (structure) and macroscopic properties of the state of matter.

A phase change is the transition between different states of matter. It is generally characterized by a sudden change of the internal microscopic structure and the macroscopic properties of the environment. It is probably better to refer to it as a phase transition instead of phase change because a phase transition doesn't imply the change of state of the matter, e.g. liquid-liquid or solid-solid phase transitions. For a solid-solid phase transition there can be a change of the structure of the crystal.

A phase transition can be effected by a change of the composition, temperatures and/or pressures of the system, or by application of an external force on the system. Consequently, composition, density, molar internal energy, enthalpy, entropy, refractive index, and dielectric constant have different values in each phase. However, temperature and pressure are identical for all phases in multiphase systems in compliance with thermodynamic principles. As a result, when two phases (or more) exist, we refer to it as phase equilibria. Table 1 lists all the phase transitions between solid, liquid and vapor phases.


Table 1. Phases transitions.

There are two types of phase transition:

*First order:* In this type of transition there is a discontinuity in the first derivative of the Gibbs free energy with regard to thermodynamic variables. For this type of transition there is an absorption or release of a fixed amount of energy with the temperature remaining constant.

*Second order:* For this type of transition, there is continuity in the first derivative of the Gibbs free energy with regard to the thermodynamic variable across the transition, but there is a discontinuity in the second derivative of the Gibbs energy with regard to thermodynamic properties.

#### **2.2 Chemical potentials and equilibrium conditions**

#### **2.2.1 Chemical potential**

484 Advances in Chemical Engineering

identical. A system can comprise of one or several phases. Depending on whether one or more phases exist, the system can be defined as a monophasic homogenous system in which the composition and thermodynamic properties are identical in the whole space, or as a multiphase heterogenous system for which thermodynamic properties change distinctly at the phase interface. A phase is characterized by its temperature, density, pressure, and other thermodynamics properties, i.e. Gibbs energy, molar enthalpy, entropy, heat capacity, etc*.*  The concept of a phase can be employed to distinguish between the different states of matter. Matter is generally accepted to exist in three main states, viz. gas, liquid, and solid. Other types of phases do exist, e.g. nematic and smectic phase transitions in liquid crystral, but this will not be relevant to this chapter. Molecular interactions between the components or chemical species which comprise the system are responsible for the different states of matter. Consequently thermodynamics model used to describe phases have to take into account the molecular interactions, or be used at conditions at which they are relevant, e.g. the ideal gas model can only be applied to gas phase at low pressures and sufficiently high

There a number of definitions that can be used to describe a liquid. One of the definitions, which is easy to visualize but may not be entirely thermodynamically correct, is that a liquid is a fluid which takes the form of its containment without necessarily filling the entire containment volume. This characteristic distinguishes a liquid from a solid, as well as a liquid from a gas. As stated, it is not an accurate definition and it is necessary to carefully check the thermophysical microscopic (structure) and macroscopic properties of the state of matter.

A phase change is the transition between different states of matter. It is generally characterized by a sudden change of the internal microscopic structure and the macroscopic properties of the environment. It is probably better to refer to it as a phase transition instead of phase change because a phase transition doesn't imply the change of state of the matter, e.g. liquid-liquid or solid-solid phase transitions. For a solid-solid phase transition there can

A phase transition can be effected by a change of the composition, temperatures and/or pressures of the system, or by application of an external force on the system. Consequently, composition, density, molar internal energy, enthalpy, entropy, refractive index, and dielectric constant have different values in each phase. However, temperature and pressure are identical for all phases in multiphase systems in compliance with thermodynamic principles. As a result, when two phases (or more) exist, we refer to it as phase equilibria.

> **Phase 1 Phase 2 Transition 1-2**  liquid vapor boiling liquid solid solidification vapor liquid liquefaction vapor solid condensation solid vapor sublimation solid liquid melting

Table 1 lists all the phase transitions between solid, liquid and vapor phases.

temperatures, as at these conditions molecular interactions are negligible.

**2.1 Phases transitions** 

be a change of the structure of the crystal.

Table 1. Phases transitions.

The chemical potential is the one of the most important thermodynamic properties used in the description of phase equilibrium. If one considers a phase with volume V, containing nc components at temperature T and pressure P, the chemical potential *<sup>i</sup>* of component i in the phase is defined by

$$\mu\_i \left( P, T, n\_1, n\_2, \dots, n\_C \right) = \left( \frac{\partial \mathcal{G}\left( P, T, n\_1, n\_2, \dots, n\_C \right)}{\partial n\_i} \right)\_{T, P, n\_{j \neq i}} \tag{1}$$

Where G=H-TS is the Gibbs free energy of the phase. The expression for the infinitesimal reversible change in the Gibbs free energy is given by

$$dG = VdP - SdT + \sum\_{i} \mu\_{i}dn\_{i} \tag{2}$$

Moreover, from Euler theorem we can write <sup>1</sup> *Nc i i <sup>i</sup> G n* . The Gibbs-Duhem (Eq. 3) equation can be obtained from equations 1 and 2:

$$VdP - SdT + \sum\_{i} n\_{i}d\mu\_{i} = 0\tag{3}$$

#### **2.2.2 Equilibrium conditions**

Considering a multicomponent system in equilibrium between two phases ( and ), at temperature T and pressure P, the Gibbs free energy of this system is *G=G +G* , where *G* and *G* , are the Gibbs free energy of the and phases respectively. At equilibrium, the Gibbs free energy must be at the minimum, therefore (*dG=0*):

$$\mathrm{d}\mathbf{G} = \mathrm{d}\mathbf{G}^{\alpha} + \mathrm{d}\mathbf{G}^{\beta} = \sum\_{i} \mu\_{i}^{\alpha} \mathrm{d}n\_{i}^{\alpha} + \sum\_{i} \mu\_{i}^{\beta} \mathrm{d}n\_{i}^{\beta} = \mathbf{0} \tag{4}$$

For a closed system, *<sup>i</sup> dn* = - *<sup>i</sup> dn* for each component *i* between phases and . For each chemical species, the following relationship can written:

$$\mu\_i^{\alpha} \left( \mathcal{P}\_{\prime} \mathcal{T}\_{\prime} n\_1^{\alpha}, n\_2^{\alpha}, \dots, n\_{\mathcal{N}\_{\mathcal{C}}}^{\alpha} \right) = \mu\_i^{\beta} \left( \mathcal{P}\_{\prime} \mathcal{T}\_{\prime} n\_1^{\beta}, n\_2^{\beta}, \dots, n\_{\mathcal{N}\_{\mathcal{C}}}^{\beta} \right) \tag{5}$$

Phase Diagrams in Chemical Engineering: Application to Distillation and Solvent Extraction 487

Were, *F*, is the of degrees of freedom, *C*, is the number of components and *Φ*, is the number of phases present. Table 2 illustrates the degrees of freedom for a single component system.

liquid, vapor, or solid 1 2 *P* and *T* Triple point 3 0 Everything is

Critical point 2 1 *TC* or *PC*

phase, the vapor-liquid equilibrium can be represented as shown in Figure 2.

T>TC

The phase behavior of a pure component can be represented in a plot of pressure versus temperature. Assuming that there are only two co-exisiting phases, i.e. a liquid and vapor

Fig. 2. Pure compound Pressure-Temperature-density phase diagram. Red line: isothermal

**Density**

T=TC T<TC

The critical point of a pure compound is the upper limit (temperature and pressure) of the pure component vapor pressure curve. For temperatures and pressures below the critical

curve. Black line saturation curve (bubble and dew points lines)

**3.1.2 The critical point** 

**Pressure**

Vapor

**freedom** 

2 1 *T* or *P*

**Variables to be specified** 

fixed

Liquid

**Region of the phase diagram Number of phases Degrees of** 

Table 2. Degrees of freedom for a pure component (*C=1*)

vapor pressure, or melting or

sublimation curves

Consequently, the equilibrium condition a of multiphase mixture is equality of temperature, pressure, and chemical potential *<sup>i</sup>* of each component *i* in all the phases in equilibrium.

#### **3. Description of phase diagrams**

#### **3.1 Pure compound**

#### **3.1.1 Description**

The phase diagram of a pure compound is characterized by its critical point, triple point, vapor pressure, and melting and sublimation curves (figure 1). At the triple point, the solid, liquid, and vapor phases coexist. The critical point can be defined as the upper limit for the pure component vapor pressure curve. For temperatures and pressures above the critical conditions, there is no possibility to have vapor-liquid equilibrium. The supercritical state can be considered as a "stable state" with no possibility for phase separation.

Fig. 1. Pure compound Pressure-Temperature phase diagram. (▲) Triple point, (●) : critical point. Lines : Coexistence curves

The number of intensive variables which have to be specified in order to characterize the system is determined with the Gibbs phases rule (Eq. 6).

$$F = \mathbb{C} + \mathbb{Z} - \phi \tag{6}$$


Were, *F*, is the of degrees of freedom, *C*, is the number of components and *Φ*, is the number of phases present. Table 2 illustrates the degrees of freedom for a single component system.

Table 2. Degrees of freedom for a pure component (*C=1*)

486 Advances in Chemical Engineering

Consequently, the equilibrium condition a of multiphase mixture is equality of temperature,

The phase diagram of a pure compound is characterized by its critical point, triple point, vapor pressure, and melting and sublimation curves (figure 1). At the triple point, the solid, liquid, and vapor phases coexist. The critical point can be defined as the upper limit for the pure component vapor pressure curve. For temperatures and pressures above the critical conditions, there is no possibility to have vapor-liquid equilibrium. The supercritical state

Fig. 1. Pure compound Pressure-Temperature phase diagram. (▲) Triple point, (●) : critical

Vapor

The number of intensive variables which have to be specified in order to characterize the

*F C* 2

**Temperature**

(6)

can be considered as a "stable state" with no possibility for phase separation.

*<sup>i</sup>* of each component *i* in all the phases in equilibrium.

Super critical

fluid

pressure, and chemical potential

point. Lines : Coexistence curves

**logarithm of Pressure**

Solid

system is determined with the Gibbs phases rule (Eq. 6).

Liquid

**3.1 Pure compound 3.1.1 Description** 

**3. Description of phase diagrams** 

The phase behavior of a pure component can be represented in a plot of pressure versus temperature. Assuming that there are only two co-exisiting phases, i.e. a liquid and vapor phase, the vapor-liquid equilibrium can be represented as shown in Figure 2.

Fig. 2. Pure compound Pressure-Temperature-density phase diagram. Red line: isothermal curve. Black line saturation curve (bubble and dew points lines)

#### **3.1.2 The critical point**

The critical point of a pure compound is the upper limit (temperature and pressure) of the pure component vapor pressure curve. For temperatures and pressures below the critical

Phase Diagrams in Chemical Engineering: Application to Distillation and Solvent Extraction 489

At the triple point, the three states of the matter (solid, liquid, and vapor) are in equilibrium. The temperature and pressure of the triple point is fixed because the degrees of freedom at this point is equal to zero. Triple point values are very useful for the definition of reference points on temperature scales and for the calibration of experimental equipment. The table 4

> **Compound** *Ttr* **/K** *Ptr* **/kPa**  Water 273.16 0.6117 Ethanol 150 4.3 × 10−<sup>7</sup> Oxygen 54.36 0.152 Methane 90.67 11.69

The first equation of state, which could describe the behaviour of both liquid and vapour states of a pure component, was developed by (van der Waals, 1873). Two types of

> <sup>2</sup> *<sup>a</sup> P v b RT*

(8)

 and 0 *v T c* 

 and 0 *v T <sup>c</sup>*

*v*

2 <sup>2</sup> 0, 0 *T T*

2 <sup>2</sup> 0, 0 *T T*

It can be seen that at the critical point, the isochoric heat capacity has infinite value. Phase diagram describing binary mixtures depend of the behaviour of the species. Van Konynenburg & Scott (1980) have classified the phase behaviour of binary mixtures into six types considering van der Waals EoS and quadratic mixing rules. Figure 3 presents the different types of phase diagrams. The transition between each type of phase diagram can be explained by considering the size effects of molecules and the repulsive interactions between them. Figure 4 illustrates the possible transitions between the different types of

The thermodynamic stability of a mixture determines if it would remain as a stable homogenous fluid or split into more stable phases and therefore produce two or more

*P P v v* 

*P P v v* 

At the critical point these two previous conditions are equal to zero.

Table 4. Triple points of few pure compounds (Ref. : NIST [NIST, 2010 #167]).

interactions (repulsive and attractive) were considered in this equation.

The stability criteria are defined by the following equations:

lists triple point conditions for some for some common chemicals.

**3.1.3 Triple point** 

**3.2 Binary systems** 

phase diagrams.

**3.2.1 Mixture critical point** 

point, phase separation can occur leading to two phases in equilibrium (liquid and vapour). The critical point can be considered as a limit of stability of the supercritical phase.

Critical properties (*TC*, *PC*, ...) characterize the critical point and the value of the properties are particularly influenced by molecular interactions, i.e. if there are strong attractive interaction between the molecules, the value of the critical temperatures and pressures are greatly increased. Using the example of water [NIST, 2010 #167] (*TC*= 647.13 K, PC= 22.055 MPa) one can clearly observe that the values of the critical properties are far greater when compared to those of methane [NIST, 2010 #167] (*TC*= 190.564 K, *PC*= 4.599 MPa). Hydrogen bonding between molecules of water is responsible for association effects between water molecules.

At the critical point, due to the fluctuation of the density, there is a disturbance of light waves across the visual spectrum. This phenomenon is called critical opalescence. At the critical point there is a second order type phase transition. In the Pressue-density diagram (figure 2), the isotherm labeled *TC* has an inflection point and is characterized by the following relations:

$$P > 0 \,, \left(\frac{\partial P}{\partial v}\right)\_T = 0 \,, \left(\frac{\partial^2 P}{\partial v^2}\right)\_T = 0 \,, \left(\frac{\partial^3 P}{\partial v^3}\right)\_T < 0 \,\tag{7}$$

Experimentally, it can be observed that along the coexisting curve liquid and vapor densities diverge at the critical point following the relationship *LV C T T* where *ρC* and *TC* are the critical density and critical temperature respectively and *β* is an universal critical exponent (value is around 0.326, whatever the pure compound). This mathematical relationship is one of thermodynamics relationships and referred to as scaling laws. For pressure, one can write the following relationship along the critical isotherm curve: *P P* 

*C C C C P* , where 4.800 . With regard to heat capacity, the following equation

can be written along the critical isochor: *Ck k <sup>V</sup>* 0 1 for *ρ*= *ρC* with *<sup>C</sup> C T T T* .

For isothermal compressibility the relationship along the critical isochors is *k* <sup>1</sup> 

for *ρ*=*ρC* with *<sup>C</sup> C T T T* . The exponents' and are also referred to as critical

exponents. Simple relationships can be written between these exponents, e.g. 2 =2. All of these laws are also observed experimentally and can be derived from renormalisation theory. Table 3 shows the optimum values for these critical exponents.


Table 3. Critical exponents

#### **3.1.3 Triple point**

488 Advances in Chemical Engineering

point, phase separation can occur leading to two phases in equilibrium (liquid and vapour).

Critical properties (*TC*, *PC*, ...) characterize the critical point and the value of the properties are particularly influenced by molecular interactions, i.e. if there are strong attractive interaction between the molecules, the value of the critical temperatures and pressures are greatly increased. Using the example of water [NIST, 2010 #167] (*TC*= 647.13 K, PC= 22.055 MPa) one can clearly observe that the values of the critical properties are far greater when compared to those of methane [NIST, 2010 #167] (*TC*= 190.564 K, *PC*= 4.599 MPa). Hydrogen bonding between molecules of water is responsible for association effects between water

At the critical point, due to the fluctuation of the density, there is a disturbance of light waves across the visual spectrum. This phenomenon is called critical opalescence. At the critical point there is a second order type phase transition. In the Pressue-density diagram (figure 2), the isotherm labeled *TC* has an inflection point and is characterized by the

2

Experimentally, it can be observed that along the coexisting curve liquid and vapor densities

are the critical density and critical temperature respectively and *β* is an universal critical exponent (value is around 0.326, whatever the pure compound). This mathematical relationship is one of thermodynamics relationships and referred to as scaling laws. For pressure, one can write the following relationship along the critical isotherm curve:

and 

All of these laws are also observed experimentally and can be derived from renormalisation

0.110

0.3255

1.239

4.800

**Critical exponent Values** 

For isothermal compressibility the relationship along the critical isochors is *k* <sup>1</sup>

exponents. Simple relationships can be written between these exponents, e.g.

theory. Table 3 shows the optimum values for these critical exponents.

 

*P v* 

<sup>2</sup> 0 *T*

, 3

 

<sup>3</sup> 0 *T*

(7)

*C T T T*

 

.

2=2.

are also referred to as critical

where *ρC* and *TC*

*P v* 

4.800 . With regard to heat capacity, the following equation

for *ρ*= *ρC* with *<sup>C</sup>*

*P* 0 , 0

can be written along the critical isochor: *Ck k <sup>V</sup>* 0 1

. The exponents'

*C T T T*

*P v* ,

*T*

diverge at the critical point following the relationship *LV C T T*

The critical point can be considered as a limit of stability of the supercritical phase.

molecules.

following relations:

*C C C C*

 

for *ρ*=*ρC* with *<sup>C</sup>*

Table 3. Critical exponents

, where

*P P P*

At the triple point, the three states of the matter (solid, liquid, and vapor) are in equilibrium. The temperature and pressure of the triple point is fixed because the degrees of freedom at this point is equal to zero. Triple point values are very useful for the definition of reference points on temperature scales and for the calibration of experimental equipment. The table 4 lists triple point conditions for some for some common chemicals.


Table 4. Triple points of few pure compounds (Ref. : NIST [NIST, 2010 #167]).

#### **3.2 Binary systems**

The first equation of state, which could describe the behaviour of both liquid and vapour states of a pure component, was developed by (van der Waals, 1873). Two types of interactions (repulsive and attractive) were considered in this equation.

$$\left(P + \frac{a}{v^2}\right)(v - b) = RT\tag{8}$$

The stability criteria are defined by the following equations:

$$\left(\frac{\partial P}{\partial v}\right)\_T < 0, \left(\frac{\partial^2 P}{\partial v^2}\right)\_T < 0 \quad \text{and} \quad \frac{T}{c\_v} > 0$$

At the critical point these two previous conditions are equal to zero.

$$\left(\frac{\partial P}{\partial v}\right)\_T = 0\\\left(\frac{\partial^2 P}{\partial v^2}\right)\_T = 0 \quad \text{and} \quad \frac{T}{c\_v} = 0$$

It can be seen that at the critical point, the isochoric heat capacity has infinite value. Phase diagram describing binary mixtures depend of the behaviour of the species. Van Konynenburg & Scott (1980) have classified the phase behaviour of binary mixtures into six types considering van der Waals EoS and quadratic mixing rules. Figure 3 presents the different types of phase diagrams. The transition between each type of phase diagram can be explained by considering the size effects of molecules and the repulsive interactions between them. Figure 4 illustrates the possible transitions between the different types of phase diagrams.

#### **3.2.1 Mixture critical point**

The thermodynamic stability of a mixture determines if it would remain as a stable homogenous fluid or split into more stable phases and therefore produce two or more

Phase Diagrams in Chemical Engineering: Application to Distillation and Solvent Extraction 491

TYPE V

TYPE II

TYPE VI

Molecular interaction

S + MI

TYPE IV TYPE III

effect (MI)

phases in equilibrium. By definition, a mixture is considered as stable when the Gibbs or Helmholtz free energy are at their minimum. Figure 5 gives an indication of the change of the Gibbs free energy with composition for a given temperature and pressure from stable condition (a) (T, P) to an unstable condition (b). In figure 5 b, one can observe that there are

H bonding

Size effect (S)

a b

**GM -->**

Fig. 5. Mixing Gibbs free energy (G) as a function of molar composition (x1) at given T and P.

1 2

*<sup>G</sup> Gn n n n*

*i i i j i j T P*

*ii i i i j*

*ijk*

,

two minima: this corresponds to phase equilibria. More details on thermodynamic phase stability are given by (Michelsen, 1982a). A Taylor series expansion can be obtained for the

5

0.0 0.2 0.4 0.6 0.8 1.0

**Composition**

(9)

,

2

*n n*

*nnn n*

a): stable condition, b): unstable condition with phase equilibria.

00 0

*i jk ijk T P*

3

*nnn*

*G*

 

Gibbs free energy for a given temperature:

0.0 0.2 0.4 0.6 0.8 1.0

**Composition**

1 6

Fig. 4. Evolution of phase diagrams.

effect (MI)

Size effect (S)

TYPE I

Molecular interaction

**GM -->**

Fig. 3. Six types of phase behaviour in binary fluid systems. C: Critical point, L: Liquid, V: Vapor. UCEP: Upper critical end point, LCEP: Lower critical end point. Dashed curve are critical.

Fig. 4. Evolution of phase diagrams.

2

2

VLE

Mixture critical line

C2

Type II

UCEP

VLE

Type IV

Type VI

VLE

2

C2

2

C2

C2

VLLE

VLE

UCEP

C1

LCEP

VLLE

UCEP

C1

VLLE

LCEP

VLLE

1

Temperature

LCEP

LLE

1 2

C1

UCEP

LLE

Fig. 3. Six types of phase behaviour in binary fluid systems. C: Critical point, L: Liquid, V: Vapor. UCEP: Upper critical end point, LCEP: Lower critical end point. Dashed curve are

C2

critical.

Type I

1

Type III

Pressure

Type V

1

VLLE

C1

<sup>1</sup><sup>2</sup>

VLLE

UCEP C1 C2

C1

LCEP

VLE

UCEP

VLE

phases in equilibrium. By definition, a mixture is considered as stable when the Gibbs or Helmholtz free energy are at their minimum. Figure 5 gives an indication of the change of the Gibbs free energy with composition for a given temperature and pressure from stable condition (a) (T, P) to an unstable condition (b). In figure 5 b, one can observe that there are

a): stable condition, b): unstable condition with phase equilibria.

Fig. 5. Mixing Gibbs free energy (G) as a function of molar composition (x1) at given T and P.

two minima: this corresponds to phase equilibria. More details on thermodynamic phase stability are given by (Michelsen, 1982a). A Taylor series expansion can be obtained for the Gibbs free energy for a given temperature:

$$\begin{aligned} G &= \sum\_{i} n\_{i}^{0} \mu\_{i}^{0} + \sum\_{i} \mu\_{i}^{0} \Delta n\_{i} + \frac{1}{2} \sum\_{i} \sum\_{j} \left( \frac{\hat{\sigma}^{2} G}{\partial n\_{i} \partial n\_{j}} \right)\_{T, P} \Delta n\_{i} \Delta n\_{j} + \\\ &\frac{1}{6} \sum\_{i} \sum\_{j} \sum\_{k} \left( \frac{\hat{\sigma}^{3} G}{\partial n\_{i} \partial n\_{j} \partial n\_{k}} \right)\_{T, P} \Delta n\_{i} \Delta n\_{j} \Delta n\_{k} + \Theta \left( \Delta n^{5} \right) \end{aligned} \tag{9}$$

Phase Diagrams in Chemical Engineering: Application to Distillation and Solvent Extraction 493

It is similar to type I but at low temperatures, the two components are not miscible in the liquid mixtures. Consequently a liquid - liquid equilibrium appears. The mixture critical point line for liquid - liquid equilibrium starts from the UCEP (upper critical end point). At the UCEP, the two liquid phases merge into one liquid phase. Examples of systems which

Generally system which have very large immiscibility gaps exhibit this behaviour, e.g. aqueous systems with hydrocarbons. A liquid – liquid - vapour curve appears and a first mixture critical point line starts from the pure component 1 critical point and ends at the UCEP. The second one starts from the infinite pressure (P → ∞) and ends at the pure component 2 critical point, generally the solvent i.e. water. The slope of this second curve can be positive, negative or positive and negative. Concerning the positive curve, we have two phases at temperatures larger than the critical temperature of pure component two.

Positive and negative slope: nitrogen + ammonia or ethane + methanol binary systems.

nitrobenzene binary systems are examples of systems which show this behaviour.

It is similar to type V behaviour. The vapor-liquid critical point line starts at the critical point of component 2 and ends at the LCEP (lower critical end point). Vapour-liquid-liquid equilibrium (VLLE) exists and is present in two parts. Ethane + n-propanol and CO2 +

It is a modification of type III phase diagram. There are two vapor-liquid critical point lines. One goes from the pure component critical point 1 and ends at the UCEP. The other starts at the pure component critical point 2 and ends at the LCEP. Contrary to type IV systems, below the LCEP the liquids are completely miscible. The ethylene + methanol binary system

There are two critical point curves. The first is similar to one presented with the type II diagram: a connection between the two pure component critical points. The second connects the LCEP and the UCEP. Between these two points, there exists VLLE. The system "water + 2-butanol" is a typical example of a type VI system. In fact the main reason is due to the existence of hydrogen bonding for one or both of the pure components (self association) and in the mixture (strong H bonding between the two components). H bonding favours the heat of solution and so miscibility in the liquid state. Above the LCEP hydrogen bonds break and the liquid becomes unstable and a second

exhibit this behaviour are those with hydrocarbons and fluorinated fluorocarbons.

*Type II phase behaviour* 

*Type III phase behaviour* 

Some examples:

*Type IV phase behaviour* 

*Type V phase behaviour* 

*Type VI phase behaviour* 

liquid phase appears.

is an example of a type V system.

Positive slope: helium + water binary system

Negative slope: methane + toluene binary system

With <sup>0</sup> *nn n ii i* and <sup>0</sup> *VV V* .

The stability condition leads to:

$$G - \sum\_{i} n\_i^0 \mu\_i^0 - \sum\_{i} \mu\_i^0 \Delta n\_i \ge 0 \tag{10}$$

Consequently

$$\frac{1}{2} \sum\_{i} \sum\_{j} \left( \frac{\partial^2 G}{\partial n\_i \partial n\_j} \right)\_{T,P} \Delta n\_i \Delta n\_j + \frac{1}{6} \sum\_{i} \sum\_{j} \sum\_{k} \left( \frac{\partial^3 G}{\partial n\_i \partial n\_j \partial n\_k} \right)\_{T,P} \Delta n\_i \Delta n\_j \Delta n\_k + \Theta \left( \Delta n^5 \right) \ge 0 \tag{11}$$

The critical point can be described as the limit of stability (*x* is the molar fraction) and coordinates (P, v) can be determined considering the following relations. More details are given by (Michelsen, 1980, 1982b), (Heidmann & Khalil, 1998) and (Stockfleth & Dohrn,1980).

$$\mathbf{G}\_{2\mathbf{x}} = \left(\frac{\partial^2 G}{\partial \mathbf{n}\_i \partial \mathbf{n}\_j}\right)\_{T,P} = \mathbf{0} \; \text{; } \mathbf{G}\_{3\mathbf{x}} = \left(\frac{\partial^3 G}{\partial \mathbf{n}\_i \partial \mathbf{n}\_j \partial \mathbf{n}\_k}\right)\_{T,P} = \mathbf{0} \; \text{; } \mathbf{G}\_{4\mathbf{x}} = \left(\frac{\partial^4 G}{\partial \mathbf{n}\_i \partial \mathbf{n}\_j \partial \mathbf{n}\_k \partial \mathbf{n}\_l}\right)\_T > 0 \tag{11}$$

In their paper, (Baker et al., 1982) present several examples of Gibbs energy analysis. Figure 6 compares phase diagrams (pressure/composition) and the Mixing Gibbs free energy at a given temperature.

A) Phase diagram at a given temperature. Dashed line: given pressure. B) Gibbs free energy of mixing at the same temperature and given pressure.

Fig. 6. Example of Mixing Gibbs free energy minimum for a binary system.

#### **3.2.2 Phase diagram classification**

#### *Type I phase behaviour*

It is the simplest type of phase diagram. The mixture critical point line starts at the first pure component critical point and finish at the second pure component critical point. Mixtures which exhibit this behaviour are two components which are chemically similar or have comparable critical properties., e.g. systems with CO2 and light hydrocarbons, systems with HFC refrigerants, and benzene + toluene binary systems.

#### *Type II phase behaviour*

492 Advances in Chemical Engineering

00 0 0 *ii i i*

*n n nnn n*

(10)

4 <sup>4</sup> 0 *<sup>x</sup> ijkl <sup>T</sup>*

*nnn n* 

0.0 0.2 0.4 0.6 0.8 1.0

**Composition**

*<sup>G</sup> <sup>G</sup>*

5

(11)

 

*i i Gn n* 

, ,

3

*nnn* 

0 *<sup>x</sup> ijk T P*

In their paper, (Baker et al., 1982) present several examples of Gibbs energy analysis. Figure 6 compares phase diagrams (pressure/composition) and the Mixing Gibbs free energy at a

a b A) Phase diagram at a given temperature. Dashed line: given pressure. B) Gibbs free energy of mixing

**GM -->**

It is the simplest type of phase diagram. The mixture critical point line starts at the first pure component critical point and finish at the second pure component critical point. Mixtures which exhibit this behaviour are two components which are chemically similar or have comparable critical properties., e.g. systems with CO2 and light hydrocarbons, systems with

Fig. 6. Example of Mixing Gibbs free energy minimum for a binary system.

2 6 *i j ijk i j i j i jk ijk T P T P G G*

*<sup>G</sup> <sup>G</sup>*

1 1 <sup>0</sup>

The critical point can be described as the limit of stability (*x* is the molar fraction) and coordinates (P, v) can be determined considering the following relations. More details are given by (Michelsen, 1980, 1982b), (Heidmann & Khalil, 1998) and (Stockfleth & Dohrn,1980).

 (11)

,

;

2 3

*n n nnn*

With <sup>0</sup> *nn n ii i* and <sup>0</sup> *VV V* .

2

*n n* 

*<sup>G</sup> <sup>G</sup>*

0 *<sup>x</sup> i j T P*

at the same temperature and given pressure.

0.0 0.2 0.4 0.6 0.8 1.0

**Composition**

**3.2.2 Phase diagram classification** 

HFC refrigerants, and benzene + toluene binary systems.

*Type I phase behaviour* 

,

;

3

2

given temperature.

**P-->**

The stability condition leads to:

Consequently

It is similar to type I but at low temperatures, the two components are not miscible in the liquid mixtures. Consequently a liquid - liquid equilibrium appears. The mixture critical point line for liquid - liquid equilibrium starts from the UCEP (upper critical end point). At the UCEP, the two liquid phases merge into one liquid phase. Examples of systems which exhibit this behaviour are those with hydrocarbons and fluorinated fluorocarbons.

#### *Type III phase behaviour*

Generally system which have very large immiscibility gaps exhibit this behaviour, e.g. aqueous systems with hydrocarbons. A liquid – liquid - vapour curve appears and a first mixture critical point line starts from the pure component 1 critical point and ends at the UCEP. The second one starts from the infinite pressure (P → ∞) and ends at the pure component 2 critical point, generally the solvent i.e. water. The slope of this second curve can be positive, negative or positive and negative. Concerning the positive curve, we have two phases at temperatures larger than the critical temperature of pure component two.

#### Some examples:

Positive slope: helium + water binary system

Negative slope: methane + toluene binary system

Positive and negative slope: nitrogen + ammonia or ethane + methanol binary systems.

#### *Type IV phase behaviour*

It is similar to type V behaviour. The vapor-liquid critical point line starts at the critical point of component 2 and ends at the LCEP (lower critical end point). Vapour-liquid-liquid equilibrium (VLLE) exists and is present in two parts. Ethane + n-propanol and CO2 + nitrobenzene binary systems are examples of systems which show this behaviour.

#### *Type V phase behaviour*

It is a modification of type III phase diagram. There are two vapor-liquid critical point lines. One goes from the pure component critical point 1 and ends at the UCEP. The other starts at the pure component critical point 2 and ends at the LCEP. Contrary to type IV systems, below the LCEP the liquids are completely miscible. The ethylene + methanol binary system is an example of a type V system.

#### *Type VI phase behaviour*

There are two critical point curves. The first is similar to one presented with the type II diagram: a connection between the two pure component critical points. The second connects the LCEP and the UCEP. Between these two points, there exists VLLE. The system "water + 2-butanol" is a typical example of a type VI system. In fact the main reason is due to the existence of hydrogen bonding for one or both of the pure components (self association) and in the mixture (strong H bonding between the two components). H bonding favours the heat of solution and so miscibility in the liquid state. Above the LCEP hydrogen bonds break and the liquid becomes unstable and a second liquid phase appears.

Phase Diagrams in Chemical Engineering: Application to Distillation and Solvent Extraction 495

0.1 0.9 0.2 0.8 0.3 0.7

R32

0.4 0.6

0.5 0.5 0.6 0.4

**Vapour**

0.7 0.3

**Liquidvapour**

0.8 0.2 0.9 0.1

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

**Liquid**

A

L

B C

F

L-F

Propane R227ea

Fig. 8. Phase diagram of the ternary system R32-R290-R227ea at T=293 K and P=8.5 bar

L2-V

V

L1-V

L1-L2

L

A

L1-L2-V

B C

A B Fig. 9. Ternary diagram: type 2. A: Existence of Vapor Liquid Liquid Equilibrium, B: if we increase pressure or temperature, we have a supercritical fluid in equilibrium with one

(Coquelet et al., 2004).

**Liquidvapour**

liquid phase (FLE).

#### **3.3 Ternary systems**

There exists a number of different classifications for ternary systems, but we propose that of Weinstsock (1952). According to this classification the system can exhibit vapor-liquid, liquid-liquid, and also vapor-liquid-liquid equilibrium (VLE, LLE and VLLE). Considering vapor-liquid-liquid equilibria, the phase diagrams can be classified into 3 categories or types. For ternary diagrams, in general, temperature and pressure are fixed.

*Type 1 :* 

It is the most common diagram. In figure 7, one can observe VLE and the monophasic regions. The shape of the phase diagram changes with the pressure. In figure 8, one can see the phase diagram for the ternary system comprising R32 + R227ea + R290 (Coquelet et al., 2004).

*Type 2 :* 

It is an evolution of a type 1 phase diagram: VLLE and LLE appear but the two liquids are partially miscible (see figure 9).

*Type 3 :* 

According to figure 10, there exists strong immiscibility between two species. The size of the LLE region increases with pressure. Consequently, LLE region disappears to form a Liquid Fluid Equilibrium region (if there is a supercritical fluid).

There is also classification for Liquid – liquid equilibrium phase diagram. There are generally three distinguishable categories (Figure 11) :


A: Vapor-Liquid Equilibria (VLE), B: if T or P is increased a supercrital fluid (F) appears and so a FLE. Fig. 7. Ternary diagram: type 1.

There exists a number of different classifications for ternary systems, but we propose that of Weinstsock (1952). According to this classification the system can exhibit vapor-liquid, liquid-liquid, and also vapor-liquid-liquid equilibrium (VLE, LLE and VLLE). Considering vapor-liquid-liquid equilibria, the phase diagrams can be classified into 3 categories or

It is the most common diagram. In figure 7, one can observe VLE and the monophasic regions. The shape of the phase diagram changes with the pressure. In figure 8, one can see the phase

It is an evolution of a type 1 phase diagram: VLLE and LLE appear but the two liquids are

According to figure 10, there exists strong immiscibility between two species. The size of the LLE region increases with pressure. Consequently, LLE region disappears to form a Liquid

There is also classification for Liquid – liquid equilibrium phase diagram. There are

A

L

B C

F

L-F

A B A: Vapor-Liquid Equilibria (VLE), B: if T or P is increased a supercrital fluid (F) appears and so a FLE.

diagram for the ternary system comprising R32 + R227ea + R290 (Coquelet et al., 2004).

types. For ternary diagrams, in general, temperature and pressure are fixed.

**3.3 Ternary systems** 

partially miscible (see figure 9).

Fig. 7. Ternary diagram: type 1.

L-V

B C

V

A

L

Fluid Equilibrium region (if there is a supercritical fluid).

generally three distinguishable categories (Figure 11) :


*Type 1 :* 

*Type 2 :* 

*Type 3 :* 

Fig. 8. Phase diagram of the ternary system R32-R290-R227ea at T=293 K and P=8.5 bar (Coquelet et al., 2004).

Fig. 9. Ternary diagram: type 2. A: Existence of Vapor Liquid Liquid Equilibrium, B: if we increase pressure or temperature, we have a supercritical fluid in equilibrium with one liquid phase (FLE).

Phase Diagrams in Chemical Engineering: Application to Distillation and Solvent Extraction 497

The Pressure-Temperature (P-T) envelop is a very interesting way to represent phase diagrams. For a mixture (where composition is known), the P-T envelop represents the limits of the phase equilibrium region. An example of a P-T envelop is illustrated in figure 12. The point which corresponds to the maximum of pressure is called the cricondenbar, and with regard to the maximum of temperature, it is called the cricondentherm. Bubble and dew pressures curves are also presented in such a diagram with the critical point. With such

Critical point

Cricondenbar

Dew point curve, Φ=1

Cricondentherm

a diagram, phenomenon such as retrograde condensation can be easily explained.

Fig. 12. Example of P-T envelop of hydrocarbons (C1 to C5).

Bubble point curve, Φ=0

Distillation is the most well known separation unit operation in the world. Utilizing energy, the objective is to create one or more coexisting zones which differ in temperature, pressure, composition and phase state. In order to design a distillation column, the concept of an equilibrium stage is required: at this stage, the vapor and liquid streams which are leaving the stage are in complete equilibrium with each other and thermodynamic relations can be used to determine the temperature and the concentration of the different species for a given pressure. The equilibrium stage can be simulated as a thermodynamic isothermal FLASH (Michelsen, 1982b). Consequently, for a given distillation column, if the number of

**Temperature** 

**4. Application to distillation** 

**Pressure**

**3.4 Pressure-temperature envelops** 

Fig. 10. Ternary diagram: type 3. A: There exists Liquid-Liquid Equilibria (L1L2E) between species A and B: if we increase pressure or temperature, we have a supercritical fluid in equilibrium with one liquid phase (FLE) and L1L2E.

Fig. 11. Liquid – liquid equilibrium : presentation of the 3 configurations of the type 1, 2 and 3 systems.

#### **3.4 Pressure-temperature envelops**

496 Advances in Chemical Engineering

A

L2

L1-L2

B C

F L1F

A B Fig. 10. Ternary diagram: type 3. A: There exists Liquid-Liquid Equilibria (L1L2E) between species A and B: if we increase pressure or temperature, we have a supercritical fluid in

L1

C

Type 1 Type 2 Type 2a

Type 3 Type 3a Type 3b

Fig. 11. Liquid – liquid equilibrium : presentation of the 3 configurations of the type 1, 2 and

equilibrium with one liquid phase (FLE) and L1L2E.

V

L1-V

L1

B

L2-V L1-L2

L2

A

L1-L2-V

3 systems.

The Pressure-Temperature (P-T) envelop is a very interesting way to represent phase diagrams. For a mixture (where composition is known), the P-T envelop represents the limits of the phase equilibrium region. An example of a P-T envelop is illustrated in figure 12. The point which corresponds to the maximum of pressure is called the cricondenbar, and with regard to the maximum of temperature, it is called the cricondentherm. Bubble and dew pressures curves are also presented in such a diagram with the critical point. With such a diagram, phenomenon such as retrograde condensation can be easily explained.

Fig. 12. Example of P-T envelop of hydrocarbons (C1 to C5).

#### **4. Application to distillation**

Distillation is the most well known separation unit operation in the world. Utilizing energy, the objective is to create one or more coexisting zones which differ in temperature, pressure, composition and phase state. In order to design a distillation column, the concept of an equilibrium stage is required: at this stage, the vapor and liquid streams which are leaving the stage are in complete equilibrium with each other and thermodynamic relations can be used to determine the temperature and the concentration of the different species for a given pressure. The equilibrium stage can be simulated as a thermodynamic isothermal FLASH (Michelsen, 1982b). Consequently, for a given distillation column, if the number of

Phase Diagrams in Chemical Engineering: Application to Distillation and Solvent Extraction 499

Fig. 14. Example of McCabe and Thiele construction: system Extract from Perry's handbook

phase separation. For a homogeneous azeotrope, the relative volatility is equal to one. The

*i i y x* for the Nth component. Moreover, considering only VLE, the mathematical criterion can determine whether there existences an azeotrope in a system which presents VLLE behavior: in this case there exists a heterogeneous azeotrope. Table 5 describes the different

a b Fig. 15. Comparison between y-x phase diagram of N2-O2 (a) and Ar-O2 (b) binary systems

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

**y Ar**

*i T P x* 

0 0.2 0.4 0.6 0.8 1

**x Ar**

 or 0 *i P T x* 

or

definition of the existence of an azeotrope is as following: 0

types or categories of azeotropes with example [12].

0 0.2 0.4 0.6 0.8 1

**x N2**

(Perry & Green, 1997).

at 110 K.

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

**y N2**

equilibrium stages is very large it means that the separation is difficult due a relative volatility close to one. The relative volatility of component *i* with respect to component *j* is the ratio between the partition coefficient (or equilibrium constant) (Eq. 12).

$$\alpha\_{ij} = \frac{\mathbf{K}\_i}{\mathbf{K}\_j} = \frac{y\_i}{y\_j} \overbrace{\begin{pmatrix} \mathbf{x}\_i \\ \mathbf{x}\_j \\ \mathbf{x}\_j \end{pmatrix}}^{y\_i} \tag{12}$$

In distillation, three types of binary equilibrium curves are shown in figure 13.

Fig. 13. Typical binary equilibrium curves. Dashed line: system with normal volatility, dotted line: system with homogenous azeotrope and solid line: system with heterogeneous azeotrope

If the relative volatility is equal to one, it is impossible to separate the two components. The phase diagrams are very important for the design of distillation columns: considering the McCabe and Thiele method (Perry & Green, 1997), the knowledge of the y-x curve together with the stream flowrates lead to the determination of the number of theorical equilibrium stage (NTES). The McCabe-Thiele method is based on the assumption of constant molar overflow and the molar heats of vaporization of the feed components being equal; heat effects such as heats of mixing, heat transfer to and from the distillation column are negligible. If the NTES is very large, the separation is difficult. Figure 15 shows phase diagrams for an air separation unit. It can clearly be seen that the separation between O2 and Ar is more difficult in comparison with O2 and N2.

Considering a distillation column with a total reflux, the closer the values of liquid and vapor compositions, the higher is the number of equilibrium stages and more difficult is the

equilibrium stages is very large it means that the separation is difficult due a relative volatility close to one. The relative volatility of component *i* with respect to component *j* is

> *i i i ij <sup>j</sup> <sup>j</sup>*

*y K x K y*

*j*

(12)

*x*

the ratio between the partition coefficient (or equilibrium constant) (Eq. 12).

In distillation, three types of binary equilibrium curves are shown in figure 13.

Fig. 13. Typical binary equilibrium curves. Dashed line: system with normal volatility, dotted line: system with homogenous azeotrope and solid line: system with heterogeneous

Ar is more difficult in comparison with O2 and N2.

If the relative volatility is equal to one, it is impossible to separate the two components. The phase diagrams are very important for the design of distillation columns: considering the McCabe and Thiele method (Perry & Green, 1997), the knowledge of the y-x curve together with the stream flowrates lead to the determination of the number of theorical equilibrium stage (NTES). The McCabe-Thiele method is based on the assumption of constant molar overflow and the molar heats of vaporization of the feed components being equal; heat effects such as heats of mixing, heat transfer to and from the distillation column are negligible. If the NTES is very large, the separation is difficult. Figure 15 shows phase diagrams for an air separation unit. It can clearly be seen that the separation between O2 and

0 0.2 0.4 0.6 0.8 1

**Liquid composition**

Considering a distillation column with a total reflux, the closer the values of liquid and vapor compositions, the higher is the number of equilibrium stages and more difficult is the

azeotrope

**Vapor composition**

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Fig. 14. Example of McCabe and Thiele construction: system Extract from Perry's handbook (Perry & Green, 1997).

phase separation. For a homogeneous azeotrope, the relative volatility is equal to one. The definition of the existence of an azeotrope is as following: 0 *i T P x* or 0 *i P T x* or *i i y x* for the Nth component. Moreover, considering only VLE, the mathematical criterion can determine whether there existences an azeotrope in a system which presents VLLE behavior: in this case there exists a heterogeneous azeotrope. Table 5 describes the different types or categories of azeotropes with example [12].

Fig. 15. Comparison between y-x phase diagram of N2-O2 (a) and Ar-O2 (b) binary systems at 110 K.

Phase Diagrams in Chemical Engineering: Application to Distillation and Solvent Extraction 501

Liquid-liquid extraction, which is commonly referred to as solvent extraction, involves the separation of the components that constitute a liquid stream by contacting it with another liquid stream which may be insoluble or partially soluble. Due to some of the components being preferentially more soluble in one of the liquid streams, separation can be effected. The separation effected in a single contacting stage is usually not large and therefore multiple contacting stages are needed to produce a significant separation. In these extraction processes the feed stream (which contains components that are to be separated) is contacted with a solvent stream. Exiting the contacting unit would be solvent-rich stream which is generally referred to as the extract and a residual liquid which is commonly called the raffinate. In general, this contacting removes a solute from the feed stream and concentrates it in the solvent-rich stream, i.e. decreasing the concentration of that particular solute in the

The typical triangular diagrams which are used to illustrate ternary liquid-liquid equilibria can be converted into more convenient diagrams for visualization and computations in solvent extraction, e.g the distribution diagram, as seen in figure 17. This is undertaken because the phase relationships are generally very difficult to express conveniently algebraically and as a result solvent extraction computations are usually made graphically. The triangular diagram can also be transformed into rectangular coordinates, as seen in

Fig. 17. Triangular (a) and distribution diagram (b) for liquid-liquid equilibria for a system of three components where one pair is partially miscible (extracted from Treybal, 1981).

The principles of separation and utilization of phase diagrams for the sizing of unit operation in solvent extraction is analogous to that which is seen in distillation, e.g. figure 18 illustrates phase diagrams being used to determine the number of theoretical stages for separation. Just as in distillation, the process can be undertaken with reflux. Reflux at the extract end can produce a product which is greater in composition, as is the case in the rectification section of a distillation column. The concept of the operating lines in the diagram, as well as the "stepping-off" in the diagram to determine the number of theoretical

stages is similar to the McCabe-Thiele method for distillation.

**5. Application to liquid-liquid extraction** 

raffinate stream.

Figure 18.



Figure 16 presents an example of a heterogeneous azeotropic system (2 butanol + water) at 320 K.

Fig. 16. 2-butanol (1) + water(2) binary system. Prediction using PSRK EoS (Holderbaum & Gmehling, 1991) at 320 K. Dashed line: Vapor Liquid Liquid equilibria (Heterogenous azeotrope)

#### **5. Application to liquid-liquid extraction**

500 Advances in Chemical Engineering

Figure 16 presents an example of a heterogeneous azeotropic system (2 butanol + water) at

Fig. 16. 2-butanol (1) + water(2) binary system. Prediction using PSRK EoS (Holderbaum & Gmehling, 1991) at 320 K. Dashed line: Vapor Liquid Liquid equilibria (Heterogenous

0.0 0.2 0.4 0.6 0.8 1.0

**x1, y1**

Minimum of pressure Trichloromethane + 2-butanone

1-propanol + water ethanol + benzene

1-butanol + water water + benzene

2-butanone + water 2-butanol + water

benzene + hexafluorobenzene Diethylamine + methanol

Triethylamine + acetic acid

**Category Type of azeotrope Examples** 

Maximum of pressure maximale

Maximum of pressure maximale

Homogeneous Maximum of pressure with immiscibility zone

Local minimum and maximum of pressure

Homogeneous Minimum of pressure with immiscibility zone

<sup>I</sup>Homogeneous

III Homogeneous

<sup>V</sup>Double azeotrope

Table 5. Classification of azeotropes (fixed temperature).

IV

VI

320 K.

**Pressure /MPa**

azeotrope)

II Heterogeneous azeotrope

Liquid-liquid extraction, which is commonly referred to as solvent extraction, involves the separation of the components that constitute a liquid stream by contacting it with another liquid stream which may be insoluble or partially soluble. Due to some of the components being preferentially more soluble in one of the liquid streams, separation can be effected. The separation effected in a single contacting stage is usually not large and therefore multiple contacting stages are needed to produce a significant separation. In these extraction processes the feed stream (which contains components that are to be separated) is contacted with a solvent stream. Exiting the contacting unit would be solvent-rich stream which is generally referred to as the extract and a residual liquid which is commonly called the raffinate. In general, this contacting removes a solute from the feed stream and concentrates it in the solvent-rich stream, i.e. decreasing the concentration of that particular solute in the raffinate stream.

The typical triangular diagrams which are used to illustrate ternary liquid-liquid equilibria can be converted into more convenient diagrams for visualization and computations in solvent extraction, e.g the distribution diagram, as seen in figure 17. This is undertaken because the phase relationships are generally very difficult to express conveniently algebraically and as a result solvent extraction computations are usually made graphically. The triangular diagram can also be transformed into rectangular coordinates, as seen in Figure 18.

Fig. 17. Triangular (a) and distribution diagram (b) for liquid-liquid equilibria for a system of three components where one pair is partially miscible (extracted from Treybal, 1981).

The principles of separation and utilization of phase diagrams for the sizing of unit operation in solvent extraction is analogous to that which is seen in distillation, e.g. figure 18 illustrates phase diagrams being used to determine the number of theoretical stages for separation. Just as in distillation, the process can be undertaken with reflux. Reflux at the extract end can produce a product which is greater in composition, as is the case in the rectification section of a distillation column. The concept of the operating lines in the diagram, as well as the "stepping-off" in the diagram to determine the number of theoretical stages is similar to the McCabe-Thiele method for distillation.

Phase Diagrams in Chemical Engineering: Application to Distillation and Solvent Extraction 503

As the phase behaviour is affected by the choice of the solvent, and it is highly unlikely that any particular liquid will exhibit all of the properties desirable for solvent extraction, the final choice of a solvent is in most cases a compromise between various properties and parameters, viz. selectivity, distribution coefficient, insolunbility of the solvent, recoverability, density, interfacial tension, chemical reactivity, viscosity, vapor pressure,

Generally, the key parameters (selectivity and distribution coefficient) for determining the best solvent for the separation are calculated from liquid-liquid equilibrium measurements

> wt fraction of solute in raffinate wt fraction of solute in extract

Analogous to the relative volatility in distillation, the selectivity must exceed unity (the greater the value away from unity the better) for separation to take place. If the selectivity is unity, no separation is possible. The distribution coefficient (which is effectively the inverse of selectivity) however is not required to be larger than unity, but the larger the distribution coefficient the smaller the amount of solvent which will be required for the extraction.

Excellent summaries of solvent extraction processes are given in Perry and Green (1997), as

In chemical engineering, the knowledge of the phase behaviour is very important, as the design and the optimization of the separation processes needs a good knowledge of the phase diagrams. Practically, the determination of phase diagrams can be obtained through experimental methods and/or modelling. The readers can refer to the books of Raal and Mühlauber (1998) to have a complete description of experimental methods, and Prausnitz et al. (1999) regarding models and principles. The complexity of phase diagrams is increased drastically if a solid or a polymer phase exists, however the purpose of this chapter was to introduce the reader to phase diagrams encountered by chemical engineers in the most

Baker, L.E., Pierce, A.C., Luks, K.D. (1982) Gibbs Energy Analysis of Phase Equilibria, *Society* 

Coquelet, C., Chareton, A., Richon, D. (2004) Vapour–liquid equilibrium measurements and

Gmehling, J., Menke, J., Krafczyk, J., Fischer, K. (1994) *Azeotropic Data Part I*, Wiley-VCH,

correlation of the difluoromethane (R32) + propane (R290) + 1,1,1,2,3,3,3 heptafluoropropane (R227ea) ternary mixture at temperatures from 269.85 to 328.35

*of Petroleum Engineers Journal*, October, pp 731-742, ISSN: 0197-7520

K. *Fluid Phase Equilibria*, Vol. 218, pp 209-214, ISSN 0378-3812

ISBN : 3-527-28671-3, Weinheim, Germany.

commonly used units operations, viz. distillation and solvent extraction.

(13)

freezing point, toxicity, flammability, and cost.

The selectivity is defined as follows:

**6. Conclusion** 

**7. References** 

for the system of components concerned (including the solvent).

well as in Treybal (1981) and McCabe et al. (2005).

Fig. 18. Illustration of phase diagrams generated for a countercurrent extraction process with reflux (extracted from Treybal, 1981)

As the phase behaviour is affected by the choice of the solvent, and it is highly unlikely that any particular liquid will exhibit all of the properties desirable for solvent extraction, the final choice of a solvent is in most cases a compromise between various properties and parameters, viz. selectivity, distribution coefficient, insolunbility of the solvent, recoverability, density, interfacial tension, chemical reactivity, viscosity, vapor pressure, freezing point, toxicity, flammability, and cost.

Generally, the key parameters (selectivity and distribution coefficient) for determining the best solvent for the separation are calculated from liquid-liquid equilibrium measurements for the system of components concerned (including the solvent).

The selectivity is defined as follows:

$$\beta = \frac{\text{wt/fraction of solute in raffinate}}{\text{wt/fraction of solute in exact}} \tag{13}$$

Analogous to the relative volatility in distillation, the selectivity must exceed unity (the greater the value away from unity the better) for separation to take place. If the selectivity is unity, no separation is possible. The distribution coefficient (which is effectively the inverse of selectivity) however is not required to be larger than unity, but the larger the distribution coefficient the smaller the amount of solvent which will be required for the extraction.

Excellent summaries of solvent extraction processes are given in Perry and Green (1997), as well as in Treybal (1981) and McCabe et al. (2005).

#### **6. Conclusion**

502 Advances in Chemical Engineering

Fig. 18. Illustration of phase diagrams generated for a countercurrent extraction process

with reflux (extracted from Treybal, 1981)

In chemical engineering, the knowledge of the phase behaviour is very important, as the design and the optimization of the separation processes needs a good knowledge of the phase diagrams. Practically, the determination of phase diagrams can be obtained through experimental methods and/or modelling. The readers can refer to the books of Raal and Mühlauber (1998) to have a complete description of experimental methods, and Prausnitz et al. (1999) regarding models and principles. The complexity of phase diagrams is increased drastically if a solid or a polymer phase exists, however the purpose of this chapter was to introduce the reader to phase diagrams encountered by chemical engineers in the most commonly used units operations, viz. distillation and solvent extraction.

#### **7. References**


**20** 

Touhami Mokrani *University of South Africa* 

*South Africa* 

**Organic/Inorganic Nanocomposite** 

**Membranes Development for Low** 

**Temperature Fuel Cell Applications** 

The criteria that are going to influence the evolution of the world energy system in the present century are complex. The most important new factor is the need to preserve the environment, both locally and globally, through the use of new technologies and sustainable use of existing resources. The Kyoto protocol, which put a limit on greenhouse gas emissions (mainly CO2) from the industrialized countries, is a turning point in the global energy chain. On the other hand, the fuel specifications to control automotive exhaust gas emission obligate fuel producers to look for different ways of making clean fuel. Automakers are also obligated to look for alternative technology to internal combustion engines. The interest in studies on energy sources alternative to fossil fuels is linked both to the reduction of their availability and the increasing environmental impact caused by their use (Goodstein, 1999). In the energy field, an important cause of pollutant emissions is linked to ground transportation. In the last 40 years, some economic, social and cultural changes have encouraged a wide proliferation of vehicles. For example, in Europe, private cars have increased from 232 to 435 per 1000 inhabitants in the period 1971-1995 (Santarelli *et al*., 2003). Fuel cells are alternative power sources that can meet global emission regulations, and clean production. Although fuel cells have been used since the 1960's for aerospace and military applications, cost was a strong impediment to terrestrial

Five major types of fuel cells are available and are defined by their electrolyte. These include alkaline (AFC), phosphoric acid (PAFC), molten carbonate (MCFC), solid oxide (SOFC) and proton exchange membrane fuel cells (PEMFC). Table 1 summarizes some characteristics of these fuel cells. Proton exchange membrane fuel cells are the most attractive candidate for alternative automotive and stationary power sources due to their smaller size and much lower operating temperature compared to other fuel cell systems. Low temperature fuel cells are fuel cells operating at temperature less than 100°C. They are H2-proton exchange membrane fuel cell (H2-PEMFC), direct methanol fuel cell (DMFC), direct ethanol fuel cell

**1. Introduction** 

applications.

**2. Fuel cell types** 

(DEFC) and direct DME fuel cell (DDMEFC).


### **Organic/Inorganic Nanocomposite Membranes Development for Low Temperature Fuel Cell Applications**

Touhami Mokrani *University of South Africa South Africa* 

#### **1. Introduction**

504 Advances in Chemical Engineering

Heidemann, R. A. & Khalil, A. M. (1980) The Calculation of Critical Points, *American Institute* 

Holderbaum, T. & Gmehling, J. (1991) PSRK : A group contribution equation of state based on UNIFAC, Fluid Phase Equilibria, Vol. 70, pp 251-265, ISSN 0378-3812 McCabe, W., Smith J., Harriot, P. (2005) Unit Operations of Chemical Engineering 7th

Michelsen, M. L. (1980) Calculation of phase envelopes and critical point for multicomponent mixtures, *Fluid Phase Equilibria*, Vol. 4, pp 1-10, ISSN 0378-3812 Michelsen, M. L. (1982a) The isothermal Flash Problem. Part I Stability. *Fluid Phase Equilibria*,

Michelsen, M. L. (1982b) The isothermal Flash Problem. Part II Phase Split Calculation, *Fluid* 

Perry, R.H. & Green, D.W. (1997) Perry's chemical Handbook 7th , edition McGraw Hill

Stockfleth, R. & Dohrn, R. (1998) An algorithm for calculating critical points in

Treybal, R.E. (1981) Mass-Transfer Operations 3rd edition, McGraw-Hill Book Company,

Van der Waals, J.D. (1899) *Over de Continuiteit van den Gas- en Vloestoftoestand. (Über die* 

Van Konynenburg, P. H. & Scott, R.L. (1980) Critical lines and Phase Equilibria in Binary

Weintsock, J. J. (1952) *Phase equilibrium at elevated pressure in ternary systems of ethylene and water and organic liquids*, Phd dissertation, Princeton University, USA

edition, Prenticd Hall International Series, Upper Saddle River, USA. Raal, J. D. & Mühlbauer A. L. (1997) Phase Equilibria, Measurement and Computation,

Companies, ISBN0-07-049841-5, New York, USAPrausnitz, J.M., Lichtenthaler R.N., de Azevedo, E. G. (1999) Molecular thermodynamics of fluid phase equilbria, 3rd

multicomponent mixtures which can be easily implemented in existing programs to calculate phase equilibria. *Fluid Phase Equilibria*, Vol. 145, pp 43-52, ISSN 0378-

*Kontinuittät des Gas- und Flüssigkeitszustands) 1873*, Dissertation, Universität Leiden,

van der Waals mixtures. *Philosophical Transactions of the Royal Society of London*, Vol.

edition, McGraw-Hill Book Company, ISBN 0072848235.

*Phase Equilibria*, Vol. 9, pp 21-40, ISSN 0378-3812

Taylor & Francis, ISBN 1-56032-550-X, London, UK.

Niederlande, deutsche Übersetzung, Leipzig, Germany.

ISSN (electronic): 1547-5905

Vol. 9, pp 1-19, ISSN 0378-3812

ISBN-0-07-065176-0, Singapore.

298, pp 495-539, ISSN 0264-3820

3812

*of Chemical Engineers Journal*, Vol. 26, No.5, pp 769-779, ISSN (printed): 0001-1541.

The criteria that are going to influence the evolution of the world energy system in the present century are complex. The most important new factor is the need to preserve the environment, both locally and globally, through the use of new technologies and sustainable use of existing resources. The Kyoto protocol, which put a limit on greenhouse gas emissions (mainly CO2) from the industrialized countries, is a turning point in the global energy chain. On the other hand, the fuel specifications to control automotive exhaust gas emission obligate fuel producers to look for different ways of making clean fuel. Automakers are also obligated to look for alternative technology to internal combustion engines. The interest in studies on energy sources alternative to fossil fuels is linked both to the reduction of their availability and the increasing environmental impact caused by their use (Goodstein, 1999). In the energy field, an important cause of pollutant emissions is linked to ground transportation. In the last 40 years, some economic, social and cultural changes have encouraged a wide proliferation of vehicles. For example, in Europe, private cars have increased from 232 to 435 per 1000 inhabitants in the period 1971-1995 (Santarelli *et al*., 2003). Fuel cells are alternative power sources that can meet global emission regulations, and clean production. Although fuel cells have been used since the 1960's for aerospace and military applications, cost was a strong impediment to terrestrial applications.

#### **2. Fuel cell types**

Five major types of fuel cells are available and are defined by their electrolyte. These include alkaline (AFC), phosphoric acid (PAFC), molten carbonate (MCFC), solid oxide (SOFC) and proton exchange membrane fuel cells (PEMFC). Table 1 summarizes some characteristics of these fuel cells. Proton exchange membrane fuel cells are the most attractive candidate for alternative automotive and stationary power sources due to their smaller size and much lower operating temperature compared to other fuel cell systems. Low temperature fuel cells are fuel cells operating at temperature less than 100°C. They are H2-proton exchange membrane fuel cell (H2-PEMFC), direct methanol fuel cell (DMFC), direct ethanol fuel cell (DEFC) and direct DME fuel cell (DDMEFC).

Organic / Inorganic Nanocomposite Membranes

**4. Fuels for low temperature fuel cells** 

The two half reactions for the H2-PEMFC are as follows:

Reduction half reaction: cathode O2 + 4H+ + 4e-

PEMFCs (Gloaguen *et al*., 1998; Paulus *et al*., 2001).

**4.1 Pure hydrogen** 

electrons.

Development for Low Temperature Fuel Cell Applications 507

H2-proton exchange membrane fuel cells have existed since the 1960's; in fact they were used in the Gemini aerospace program of the National Aeronautics and Space Administration (NASA) of the United States. The MEA for H2-PEMFCs consists of five components namely: a porous backing layer, an anode catalyst layer, a proton exchange membrane, a cathode catalyst layer, and a porous backing layer. Hydrogen is oxidized at the anode. The proton formed migrates through the membrane while the electrons flow through the external circuit. In the cathode reaction water is formed from oxygen, protons and

2 H2O

Oxidation half reaction: Anode 2 H2 4H+ + 4e-

Cell reaction 2 H2 + O2 2 H2O

H2-PEMFCs have attracted the most attention due to their high electrochemical reactivity (Gottesfeld & Zawodzinski, 1997; Parthasarathy *et al*., 1991; Ralph, 1997) and very low noble catalyst loading since the development of a method at Los Alamos National Laboratory (LANL) to reduce the platinum loading to ca. 0.1 mg/cm2 (Wilson & Gottesfeld, 1992; Wilson, 1993; Wilson *et al*., 1995) compared to 35 mg/cm2 and 4 mg/cm2 used respectively in the Gemini program and at General Electric in the 1970s (Appleby & Yeager, 1986a,1986b). The efficiency achievable is higher than in power plants and internal combustion engines (Dohle *et al*., 2002) and there is practically zero pollution. However, the H2-PEMFC has several disadvantages including hydrogen storage and transportation and the public acceptance of hydrogen as fuel. It is well known that hydrogen and air mixtures are explosive (e.g. the Challenger disaster). Hydrogen safety measures are still one of the major implications when it comes to the commercialization of H2-PEMFCs. Adequate water content of the membranes is essential to maintain the conductivity of the polymeric proton exchange membrane (Anantaraman & Gardner, 1996; Fontanella *et al*., 1995; Gavach *et al*., 1989; Zawodzinski *et al*., 1991, 1993). During fuel cell operation, water molecules migrate through the membrane under electro-osmotic drag, fluid convection, and molecular diffusion, making it difficult to retain a high water content within the membrane. Generally, humidification is applied to the inlets of the anode and/or cathode in order to supply water to the membrane. However, excessive amounts of liquid water could impede mass transport within the electrode structure (Zawodzinski *et al*., 1991). A thinner membrane is preferred in H2-PEMFCs because it can provide an improvement in water management due to the enhanced back-diffusion of production water from the cathode to the anode side (Finsterwalder & Hambitzer, 2001). The oxygen reduction reaction (ORR) is very slow compared to the hydrogen reaction; typically hydrogen electro-oxidation on Pt is shown by an exchange current density of 10-3 A cm-2 Pt at ambient temperature. This is some 107 to 109 times more facile than the oxygen reduction at the cathode (Ralph & Hogarth, 2002). Thus, oxygen reduction is a rate limiting factor in H2-


Table 1. Fuel cells systems (Carrette *et al*., 2001)

#### **3. Low temperature fuel cells**

A fuel cell is an electrochemical system which converts chemical energy to electrical energy. A fuel cell differs from a battery in that fuels are continuously supplied and the products are continuously removed. There are two distinct fuels for low temperature fuel cells: hydrogen as used in a H2-PEMFC, and methanol as used in a DMFC. These fuel cells consist of six major parts: end plates, current collectors, flow channel blocks, gaskets, gas diffusion layers, and a membrane electrode assembly (MEA). The fuel cell principle enables a separation between power and energy. The maximum power required determines the size of the fuel cell; the energy required determines the amount of fuel to be carried. The specific power (W kg-1) of the H2-PEMFC is roughly twice that of the DMFC (Raadschelders & Jansen, 2001). Because no mobile electrolyte is employed, corrosion problems in low temperature fuel cells are reduced and cell construction is simplified with few moving parts (Bernardi & Verbrugge, 1991). Also, fuel cells operate very quietly, therefore, reducing noise pollution (Kordesch & Simader, 1995). Since the proton exchange membrane used for the electrolyte is a solid phase, it does not penetrate deeply into the electrode as does a liquid one; therefore the reaction area is limited to the contact surface between the electrode and membrane (Shin *et al*., 2002). The advantage of using solid electrolyte is that no electrolyte leakage will occur (Uchida *et al*., 1995; Yi & Nguyen, 1999). To meet the requirements of practical application a large number of single cells are assembled together to form a stack. The performance of a stack is different from that of a single cell. The stack has a much higher operating voltage, a greater power and better fuel-energy efficiency (Chu & Jiang, 1999).

#### **4. Fuels for low temperature fuel cells**

#### **4.1 Pure hydrogen**

506 Advances in Chemical Engineering

**Type Electrolyte Charge carrier in Temperature the electrolyte (°C)** 

Alkaline fuel aqueous KOH OH- <100 ,

Proton exchange proton exchange H+ 60-120

Phosphoric acid concentrated H+ 160-220

Molten carbonate mixture of CO3 2- 600-650

Solid oxide ceramic solid O2 - 800-1000

A fuel cell is an electrochemical system which converts chemical energy to electrical energy. A fuel cell differs from a battery in that fuels are continuously supplied and the products are continuously removed. There are two distinct fuels for low temperature fuel cells: hydrogen as used in a H2-PEMFC, and methanol as used in a DMFC. These fuel cells consist of six major parts: end plates, current collectors, flow channel blocks, gaskets, gas diffusion layers, and a membrane electrode assembly (MEA). The fuel cell principle enables a separation between power and energy. The maximum power required determines the size of the fuel cell; the energy required determines the amount of fuel to be carried. The specific power (W kg-1) of the H2-PEMFC is roughly twice that of the DMFC (Raadschelders & Jansen, 2001). Because no mobile electrolyte is employed, corrosion problems in low temperature fuel cells are reduced and cell construction is simplified with few moving parts (Bernardi & Verbrugge, 1991). Also, fuel cells operate very quietly, therefore, reducing noise pollution (Kordesch & Simader, 1995). Since the proton exchange membrane used for the electrolyte is a solid phase, it does not penetrate deeply into the electrode as does a liquid one; therefore the reaction area is limited to the contact surface between the electrode and membrane (Shin *et al*., 2002). The advantage of using solid electrolyte is that no electrolyte leakage will occur (Uchida *et al*., 1995; Yi & Nguyen, 1999). To meet the requirements of practical application a large number of single cells are assembled together to form a stack. The performance of a stack is different from that of a single cell. The stack has a much higher operating voltage, a greater power and

cells (AFC) solution

membrane fuel membrane

fuel cells phosphoric acid

fuel cells molten carbonates (MCFC) (Li2CO3/K2CO3)

fuel cells ZrO2(Y2O3)

Table 1. Fuel cells systems (Carrette *et al*., 2001)

better fuel-energy efficiency (Chu & Jiang, 1999).

**3. Low temperature fuel cells** 

(PAFC) 

(SOFC)

cells (PEMFC)

H2-proton exchange membrane fuel cells have existed since the 1960's; in fact they were used in the Gemini aerospace program of the National Aeronautics and Space Administration (NASA) of the United States. The MEA for H2-PEMFCs consists of five components namely: a porous backing layer, an anode catalyst layer, a proton exchange membrane, a cathode catalyst layer, and a porous backing layer. Hydrogen is oxidized at the anode. The proton formed migrates through the membrane while the electrons flow through the external circuit. In the cathode reaction water is formed from oxygen, protons and electrons.

The two half reactions for the H2-PEMFC are as follows:


H2-PEMFCs have attracted the most attention due to their high electrochemical reactivity (Gottesfeld & Zawodzinski, 1997; Parthasarathy *et al*., 1991; Ralph, 1997) and very low noble catalyst loading since the development of a method at Los Alamos National Laboratory (LANL) to reduce the platinum loading to ca. 0.1 mg/cm2 (Wilson & Gottesfeld, 1992; Wilson, 1993; Wilson *et al*., 1995) compared to 35 mg/cm2 and 4 mg/cm2 used respectively in the Gemini program and at General Electric in the 1970s (Appleby & Yeager, 1986a,1986b). The efficiency achievable is higher than in power plants and internal combustion engines (Dohle *et al*., 2002) and there is practically zero pollution. However, the H2-PEMFC has several disadvantages including hydrogen storage and transportation and the public acceptance of hydrogen as fuel. It is well known that hydrogen and air mixtures are explosive (e.g. the Challenger disaster). Hydrogen safety measures are still one of the major implications when it comes to the commercialization of H2-PEMFCs. Adequate water content of the membranes is essential to maintain the conductivity of the polymeric proton exchange membrane (Anantaraman & Gardner, 1996; Fontanella *et al*., 1995; Gavach *et al*., 1989; Zawodzinski *et al*., 1991, 1993). During fuel cell operation, water molecules migrate through the membrane under electro-osmotic drag, fluid convection, and molecular diffusion, making it difficult to retain a high water content within the membrane. Generally, humidification is applied to the inlets of the anode and/or cathode in order to supply water to the membrane. However, excessive amounts of liquid water could impede mass transport within the electrode structure (Zawodzinski *et al*., 1991). A thinner membrane is preferred in H2-PEMFCs because it can provide an improvement in water management due to the enhanced back-diffusion of production water from the cathode to the anode side (Finsterwalder & Hambitzer, 2001). The oxygen reduction reaction (ORR) is very slow compared to the hydrogen reaction; typically hydrogen electro-oxidation on Pt is shown by an exchange current density of 10-3 A cm-2 Pt at ambient temperature. This is some 107 to 109 times more facile than the oxygen reduction at the cathode (Ralph & Hogarth, 2002). Thus, oxygen reduction is a rate limiting factor in H2- PEMFCs (Gloaguen *et al*., 1998; Paulus *et al*., 2001).

Organic / Inorganic Nanocomposite Membranes

**4.2.2 Ethanol reforming** 

is then fed to the H2-PEMFC.

Pd-based (Nilsson *et al*., 2007, 2009) were investigated.

**4.2.4 Ammonia reforming** 

**4.2.3 DME reforming** 

Development for Low Temperature Fuel Cell Applications 509

Among other candidate liquid fuels, ethanol is a particular case, since it can be easily produced in great quantity by the fermentation of sugar-containing raw materials. In addition, in some countries (e.g. Brazil) ethanol is already distributed in gas stations for use in conventional cars with internal combustion engines. Hydrogen is produced from ethanol in a process unit consisting of either a steam reformer (SR) or a partial oxidation (POX) reactor in series with a water-gas shift (WGS) reactor and a reactor for selective oxidation (PROX) of CO (Ioannides & Neophytides, 2000). Product gas from the reformer or the POX reactor, which operates at an exit temperature higher than 677°C, contains a mixture of H2, CO, CO2, CH4 and H2O. After cooling, this stream enters the WGS reactor, where a large fraction of CO reacts with H2O towards CO2 and H2 at a temperature of 200°C. The product gas of the WGS reactor contains 0.1-1.5% of residual CO and enters the PROX reactor, where CO is totally oxidized - with the addition of a small amount of air - to CO2 with residual CO being less than 10 ppm. The CO free, hydrogen rich stream

DME (dimethylether) has become a promising candidate as a hydrogen source for the reforming process, because it has a high hydrogen-to-carbon ratio and a high energy density. DME can be easily handled, stored and transported. Furthermore, the infrastructure of LPG can readily be adapted for DME due to their similar physical properties. Furthermore, DME is not toxic and less explosive. DME can be catalytically reformed at relatively lower temperatures than ethanol and methane. DME can be reformed through three ways, namely steam reforming (SR), partial oxidation (POX) and authothermal reforming (ATR). DME SR proceeds via two moderately endothermic reactions in sequence; hydrolysis of DME to MeOH and steam reforming of MeOH to hydrogen and carbon dioxide. Hydrolysis of DME takes place over acid catalysis, e.g. zeolite and alumina, while MeOH SR proceeds over Cu-, Pt-, or Pd based catalyst. Therefore, bi-functional catalyst containing both acidic and metallic sites are generally needed for DME SR (Faungnawakij *et al*., 2010; Ledesma & Llorca, 2009; Nishiguchi *et al*., 2006; Takeishi & Suzuki, 2004). DME POX has been investigated over various metal catalysts such as Pt, Ni, Co and Rh supported on different oxide. Supports such as Al2O3, YSZ, LaGaO3-based and MgO were used at a high reaction temperature ranging from 400 to 700°C (S. Wang *et al*., 2002; Q. Zhang *et al*., 2005). ATR also can be used to produce hydrogen from DME. ATR is a combination of SR and POX, and catalysis such as CuFe2O4-Al2O3 (Faungnawakij & Viriya-empikul, 2010) and

Anhydrous ammonia is a widely used commodity and is available worldwide in liquid form in low pressure tanks. Procedures for safe handling have been developed in every country. Facilities for storage and transport by barges, trucks and pipelines from producer to ultimate consumer are available throughout the world. Therefore, liquid anhydrous ammonia is an excellent storage medium for hydrogen (Hacker & Kordesch, 2003). Studies demonstrate that hydrogen derived from anhydrous liquid ammonia, via a dissociation and followed by hydrogen purifier, offers an alternative to conventional methods of obtaining

#### **4.2 Hydrogen reformate**

The question of whether customers will be fuelling their vehicles directly with hydrogen or via the hydrogen-rich carrier (e.g. methanol, ethanol, gasoline, diesel, etc.) still seems to be unanswered. This is a very important issue not just from a refueling infrastructure perspective but also from the public perception and from the gearing up of production, and developing guidelines for dealing with safety issues that will need to put in place for the new fuel (Adamson & Pearson, 2000). In principle, any type of liquid fuel may be employed as a hydrogen source, e.g. gasoline, diesel, methanol, ethanol, etc. Hydrogen is produced by a reforming process. Four distinguish fuels are discussed namely methanol, ethanol, dimethylether (DME) and ammonia.

#### **4.2.1 Methanol reforming**

Methanol is produced from steam reformed natural gas and carbon dioxide using copperbased catalyst, and also from renewable biomass sources. Methanol is a leading candidate to provide the hydrogen necessary to power a fuel cell, especially in vehicular applications (Ledjeff-Hey *et al*., 1998; Mokrani & Scurrell, 2009; Olah *et al*., 2009). Methanol is currently used as a feed stock for a variety of widely used organic chemicals, including formaldehyde, acetic acid, chloromethane, and methyl tert-butyl ether (MTBE). Methanol is the desired fuel to produce hydrogen on-board. Methanol can be reformed to hydrogen by different processes including steam reforming (Amphlett *et al*., 1985; Breen & Ross, 1999; Duesterwald *et al*., 1997; Emonts *et al*., 1998; C.J. Jiang *et al*., 1993; Takahashi *et al*., 1982; Takezawa *et al*., 1982), partial oxidation (Agrell *et al*., 2001; Cubiero & Fierro, 1998; Velu *et al*., 1999) and autothermal reforming (Edwards *et al*., 1998; Höhlein *et al*., 1996; L. Ma *et al*., 1996; Mizsey *et al*., 2001).

Steam reforming of methanol occurs by two different pathways (Emonts *et al*., 1998). The first one involves the decomposition of methanol into CO and H2 through the following reaction:

CH3OH CO + 2H2

followed by a water gas shift reaction:

CO + H2O CO2 + H2

The second mechanism for methanol steam reforming consists of the reaction of water and methanol to CO2 and hydrogen:

$$\text{CH}\_3\text{OH} + \text{H}\_2\text{O} \xrightarrow{\text{H}\_2\text{O}} \qquad \text{CO}\_2 + 3\text{H}\_2\text{O}$$

which can be followed by a reverse shift reaction to establish the thermodynamic equilibrium:

$$\begin{array}{ccc} \text{CO}\_2 + \text{H}\_2 & \begin{array}{c} \implies \\ \text{O}^- \end{array} & \begin{array}{c} \text{CO}^+ \\ \text{H}^+ \end{array} \end{array}$$

Methanol steam reforming is endothermic and therefore requires that external heat, typically 300°C, is supplied. Steam reforming of methanol is usually catalyzed over Cu/ZnO type catalyst and can be performed in fixed-bed reactors (Duesterwald *et al*., 1997).

#### **4.2.2 Ethanol reforming**

508 Advances in Chemical Engineering

The question of whether customers will be fuelling their vehicles directly with hydrogen or via the hydrogen-rich carrier (e.g. methanol, ethanol, gasoline, diesel, etc.) still seems to be unanswered. This is a very important issue not just from a refueling infrastructure perspective but also from the public perception and from the gearing up of production, and developing guidelines for dealing with safety issues that will need to put in place for the new fuel (Adamson & Pearson, 2000). In principle, any type of liquid fuel may be employed as a hydrogen source, e.g. gasoline, diesel, methanol, ethanol, etc. Hydrogen is produced by a reforming process. Four distinguish fuels are discussed namely methanol, ethanol,

Methanol is produced from steam reformed natural gas and carbon dioxide using copperbased catalyst, and also from renewable biomass sources. Methanol is a leading candidate to provide the hydrogen necessary to power a fuel cell, especially in vehicular applications (Ledjeff-Hey *et al*., 1998; Mokrani & Scurrell, 2009; Olah *et al*., 2009). Methanol is currently used as a feed stock for a variety of widely used organic chemicals, including formaldehyde, acetic acid, chloromethane, and methyl tert-butyl ether (MTBE). Methanol is the desired fuel to produce hydrogen on-board. Methanol can be reformed to hydrogen by different processes including steam reforming (Amphlett *et al*., 1985; Breen & Ross, 1999; Duesterwald *et al*., 1997; Emonts *et al*., 1998; C.J. Jiang *et al*., 1993; Takahashi *et al*., 1982; Takezawa *et al*., 1982), partial oxidation (Agrell *et al*., 2001; Cubiero & Fierro, 1998; Velu *et al*., 1999) and autothermal reforming (Edwards *et al*., 1998; Höhlein *et al*., 1996; L. Ma *et al*., 1996; Mizsey *et* 

Steam reforming of methanol occurs by two different pathways (Emonts *et al*., 1998). The first one involves the decomposition of methanol into CO and H2 through the following

CH3OH CO + 2H2

CO + H2O CO2 + H2 The second mechanism for methanol steam reforming consists of the reaction of water and

CH3OH + H2O CO2 + 3H2 which can be followed by a reverse shift reaction to establish the thermodynamic

CO2 + H2 CO + H2O Methanol steam reforming is endothermic and therefore requires that external heat, typically 300°C, is supplied. Steam reforming of methanol is usually catalyzed over Cu/ZnO

type catalyst and can be performed in fixed-bed reactors (Duesterwald *et al*., 1997).

**4.2 Hydrogen reformate** 

dimethylether (DME) and ammonia.

followed by a water gas shift reaction:

methanol to CO2 and hydrogen:

**4.2.1 Methanol reforming** 

*al*., 2001).

reaction:

equilibrium:

Among other candidate liquid fuels, ethanol is a particular case, since it can be easily produced in great quantity by the fermentation of sugar-containing raw materials. In addition, in some countries (e.g. Brazil) ethanol is already distributed in gas stations for use in conventional cars with internal combustion engines. Hydrogen is produced from ethanol in a process unit consisting of either a steam reformer (SR) or a partial oxidation (POX) reactor in series with a water-gas shift (WGS) reactor and a reactor for selective oxidation (PROX) of CO (Ioannides & Neophytides, 2000). Product gas from the reformer or the POX reactor, which operates at an exit temperature higher than 677°C, contains a mixture of H2, CO, CO2, CH4 and H2O. After cooling, this stream enters the WGS reactor, where a large fraction of CO reacts with H2O towards CO2 and H2 at a temperature of 200°C. The product gas of the WGS reactor contains 0.1-1.5% of residual CO and enters the PROX reactor, where CO is totally oxidized - with the addition of a small amount of air - to CO2 with residual CO being less than 10 ppm. The CO free, hydrogen rich stream is then fed to the H2-PEMFC.

#### **4.2.3 DME reforming**

DME (dimethylether) has become a promising candidate as a hydrogen source for the reforming process, because it has a high hydrogen-to-carbon ratio and a high energy density. DME can be easily handled, stored and transported. Furthermore, the infrastructure of LPG can readily be adapted for DME due to their similar physical properties. Furthermore, DME is not toxic and less explosive. DME can be catalytically reformed at relatively lower temperatures than ethanol and methane. DME can be reformed through three ways, namely steam reforming (SR), partial oxidation (POX) and authothermal reforming (ATR). DME SR proceeds via two moderately endothermic reactions in sequence; hydrolysis of DME to MeOH and steam reforming of MeOH to hydrogen and carbon dioxide. Hydrolysis of DME takes place over acid catalysis, e.g. zeolite and alumina, while MeOH SR proceeds over Cu-, Pt-, or Pd based catalyst. Therefore, bi-functional catalyst containing both acidic and metallic sites are generally needed for DME SR (Faungnawakij *et al*., 2010; Ledesma & Llorca, 2009; Nishiguchi *et al*., 2006; Takeishi & Suzuki, 2004). DME POX has been investigated over various metal catalysts such as Pt, Ni, Co and Rh supported on different oxide. Supports such as Al2O3, YSZ, LaGaO3-based and MgO were used at a high reaction temperature ranging from 400 to 700°C (S. Wang *et al*., 2002; Q. Zhang *et al*., 2005). ATR also can be used to produce hydrogen from DME. ATR is a combination of SR and POX, and catalysis such as CuFe2O4-Al2O3 (Faungnawakij & Viriya-empikul, 2010) and Pd-based (Nilsson *et al*., 2007, 2009) were investigated.

#### **4.2.4 Ammonia reforming**

Anhydrous ammonia is a widely used commodity and is available worldwide in liquid form in low pressure tanks. Procedures for safe handling have been developed in every country. Facilities for storage and transport by barges, trucks and pipelines from producer to ultimate consumer are available throughout the world. Therefore, liquid anhydrous ammonia is an excellent storage medium for hydrogen (Hacker & Kordesch, 2003). Studies demonstrate that hydrogen derived from anhydrous liquid ammonia, via a dissociation and followed by hydrogen purifier, offers an alternative to conventional methods of obtaining

Organic / Inorganic Nanocomposite Membranes

diffusion (M.W. Mench & C.Y. Wang, 2003).

**4.4 Direct ethanol fuel cell** 

polybenzimidazol (PBI) membrane.

**4.5 Direct DME fuel cell** 

Development for Low Temperature Fuel Cell Applications 511

However, as the DMFC is still in its infancy, many problems need to be overcome to reach the commercialization stage. This includes the very sluggish methanol oxidation reaction, methanol crossover through the polymeric proton exchange membrane, CO2 evolvement at the anode (Argyropoulos *et al*., 1999a,1999b; Nordlund *et al*., 2002; Scott *et al*., 1998), and cathode flooding (Amphlett *et al*., 2001; M. Mench *et al*., 2001; X. Ren & Gottesfeld, 2001; Z.H. Wang *et al*., 2001). The methanol crossover through the polymer electrolyte leads to a mixed potential at the cathode, which results from the ORR and the methanol oxidation occurring simultaneously. This effect causes a negative potential shift at the cathode and a significant decrease of performance in the DMFC. Methanol crossover also causes fuel loses; it had been found that over 40% of methanol can be wasted in a DMFC across Nafion® membranes (Narayanan *et al*., 1996). In a DMFC, cathode flooding, which typically occurs unless high cathode stoichiometries are used, can determine to a great extent overall cell performance (Amphlett *et al*., 2001; M. Mench *et al*., 2001; X. Ren & Gottesfeld, 2001; Z.H. Wang *et al*., 2001). Water management in the DMFC is especially critical because anode water activity is near unity due to contact with liquid methanol solution (M.W. Mench & C.Y. Wang, 2003). Thus, unlike a H2-PEMFC, no back-diffusive flux of water from cathode to anode will occur, and as a result, vapourization into dry cathode flow is the only pathway for removal of excess cathode-side water accumulation from electro-osmotic drag, ORR, and

Direct fuel utilization will be of interest. Besides methanol, other alcohols, particularly those coming from biomass resources, are being considered as alternative fuels. Ethanol as an attractive fuel for electrical vehicles was investigated in direct ethanol fuel cells (Fujiwara *et al*., 1999; Gong *et al*., 2001; Lamy *et al*., 2001; W.J. Zhou *et al*., 2004). However, multimetallic catalysts are necessary to orientate the oxidation reaction selectively in the direction of complete combustion to carbon dioxide (Lamy *et al*., 2001). The reaction mechanisms of anodic oxidation of ethanol are more difficult to elucidate than methanol oxidation, since the number of electrons exchanged greatly increases (12 electrons per ethanol molecule versus 6 electrons for methanol), thus many adsorbed intermediates and products are involved (Lamy *et al*., 2001). Direct ethanol fuel cell was the second most studied fuel cell after methanol. The proton conductor membranes used are mainly Nafion® membranes. However some investigators used high temperature membrane such as Nafion®/Silica by Aricò *et al.* (1998), and also by J. Wang *et al.* (1995) using a phosphoric acid doped

Serov and Kwak (2009) have summarized the recent progress in development of direct DME fuel cell (DDMEFC). DDMEFC could be a valuable direct liquid fuel cell candidate for commercialization. However, compared with PEMFC and DMFC, DDMEFC show poor performances under ambient conditions due to the poor electrooxidation reactivity of DME (Colbow *et al.*, 2000; Kerangueven *et al.*, 2006; Mench *et al.*, 2004; Mizutani *et al.*, 2006; Ueda *et al.*, 2006). On the anode side of DDMEFC, the following oxidation reaction takes place:

CH3OCH3 + 3H2O 2CO2 + 12 H+ + 12 e-

pure hydrogen for small scale use (Strickland, 1984). The dissociation rate depends on temperature, pressure and the catalyst being used. An almost complete decomposition of ammonia can take place at approximately 430°C at atmospheric pressure. The influence and kinetic data of materials like porcelain or silica glass, metals like ion, tungsten, molybdenum, nickel, etc. especially noble metals and metal oxides, have been investigated for the dissociation of ammonia. The most used catalysts are nickel oxide and iron oxide (Papapolymerou & Bontozoglou, 1997) and the better cracking efficiencies were obtained with catalysts based on Zr, Mn, Fe and Al/alloys (Boffito, 1999; Rosenblatt & Cohn, 1952; Shikada *et al*., 1991).

#### **4.3 Direct methanol fuel cell**

DMFC technology is relatively new compared to the H2-PEMFC. However, the direct oxidation of methanol in a DMFC has been investigated over many years and some prototypes were built in the 1960's and early 1970's by the Shell Research Center in England (Glazebrook, 1982; Schatter, 1983) and by Hitachi Research Laboratories in Japan (Tamura *et al*., 1984; Williams, 1966). These studies were abandoned in the mid-1980's due to the low performance (25 mW cm-2 at best) resulting from the use of a liquid acid electrolyte (Glazebrook, 1982; Kordesch & Simader, 1996; Lamy *et al*., 2001). An alkaline electrolyte was also used, but evolved CO2 caused carbonation of the electrolyte resulting in decreased efficiency by reducing the electrolyte conductivity and de-polarizing the cathode (Verma, 2000). Currently all the research in DMFCs focuses on using solid proton exchange membranes as electrolyte, largely due to its proliferation in H2-PEMFCs. The structure of the DMFC is similar to the H2-PEMFC. At the anode methanol is directly oxidized to carbon dioxide, and the reaction at the cathode is similar to the H2-PEMFC.

The two main half reactions for the DMFC can be summarized as follows:


The thermodynamic reversible potential for a DMFC is 1.21V at 25°C (Larminie & Dicks, 2000). This value is comparable to that for a H2-PEMFC, which is 1.23V (Chu & Gilman, 1994; Léger, 2001; Qi & Kaufman, 2002; Scott *et al*., 1998). In practice, a DMFC has a much lower open circuit voltage (OCV) (Qi & Kaufman, 2002) and electrochemical losses at both electrodes lead to a significant reduction in overall performance from the theoretical thermodynamic maximum (Argyropoulos *et al*., 1999a). Since methanol is used directly at the anode, and as a consequence, a DMFC requires less auxiliary equipment and is therefore a more simplified system compared to a H2-PEMFC. Methanol is a liquid made from natural gas or renewable biomass sources, which is relatively cheap. Methanol is also easy to store, transport, and distribute, where advantage can be taken of the existing gasoline infrastructure (Mokrani & Scurrell, 2009; Olah *et al*., 2009). The anodic reaction is exothermic for both the H2-PEMFC and the DMFC; heat management is a problem in H2-PEMFC stacks. In contrast, aqueous methanol acts as a coolant in DMFCs (Hogarth *et al*., 1997; Hogarth & Ralph, 2002; Lim & C.Y. Wang, 2003; Surampudi *et al*., 1994).

pure hydrogen for small scale use (Strickland, 1984). The dissociation rate depends on temperature, pressure and the catalyst being used. An almost complete decomposition of ammonia can take place at approximately 430°C at atmospheric pressure. The influence and kinetic data of materials like porcelain or silica glass, metals like ion, tungsten, molybdenum, nickel, etc. especially noble metals and metal oxides, have been investigated for the dissociation of ammonia. The most used catalysts are nickel oxide and iron oxide (Papapolymerou & Bontozoglou, 1997) and the better cracking efficiencies were obtained with catalysts based on Zr, Mn, Fe and Al/alloys (Boffito, 1999; Rosenblatt & Cohn, 1952;

DMFC technology is relatively new compared to the H2-PEMFC. However, the direct oxidation of methanol in a DMFC has been investigated over many years and some prototypes were built in the 1960's and early 1970's by the Shell Research Center in England (Glazebrook, 1982; Schatter, 1983) and by Hitachi Research Laboratories in Japan (Tamura *et al*., 1984; Williams, 1966). These studies were abandoned in the mid-1980's due to the low performance (25 mW cm-2 at best) resulting from the use of a liquid acid electrolyte (Glazebrook, 1982; Kordesch & Simader, 1996; Lamy *et al*., 2001). An alkaline electrolyte was also used, but evolved CO2 caused carbonation of the electrolyte resulting in decreased efficiency by reducing the electrolyte conductivity and de-polarizing the cathode (Verma, 2000). Currently all the research in DMFCs focuses on using solid proton exchange membranes as electrolyte, largely due to its proliferation in H2-PEMFCs. The structure of the DMFC is similar to the H2-PEMFC. At the anode methanol is directly oxidized to carbon

dioxide, and the reaction at the cathode is similar to the H2-PEMFC.

Ralph, 2002; Lim & C.Y. Wang, 2003; Surampudi *et al*., 1994).

The two main half reactions for the DMFC can be summarized as follows:

Reduction half reaction: cathode 3/2 O2 + 6H+ + 6e- 3 H2O

Oxidation half reaction: Anode CH3OH + H2O CO2 + 6H+ + 6e-

Cell reaction CH3OH + 3/2 O2 CO2 + 2 H2O

The thermodynamic reversible potential for a DMFC is 1.21V at 25°C (Larminie & Dicks, 2000). This value is comparable to that for a H2-PEMFC, which is 1.23V (Chu & Gilman, 1994; Léger, 2001; Qi & Kaufman, 2002; Scott *et al*., 1998). In practice, a DMFC has a much lower open circuit voltage (OCV) (Qi & Kaufman, 2002) and electrochemical losses at both electrodes lead to a significant reduction in overall performance from the theoretical thermodynamic maximum (Argyropoulos *et al*., 1999a). Since methanol is used directly at the anode, and as a consequence, a DMFC requires less auxiliary equipment and is therefore a more simplified system compared to a H2-PEMFC. Methanol is a liquid made from natural gas or renewable biomass sources, which is relatively cheap. Methanol is also easy to store, transport, and distribute, where advantage can be taken of the existing gasoline infrastructure (Mokrani & Scurrell, 2009; Olah *et al*., 2009). The anodic reaction is exothermic for both the H2-PEMFC and the DMFC; heat management is a problem in H2-PEMFC stacks. In contrast, aqueous methanol acts as a coolant in DMFCs (Hogarth *et al*., 1997; Hogarth &

Shikada *et al*., 1991).

**4.3 Direct methanol fuel cell** 

However, as the DMFC is still in its infancy, many problems need to be overcome to reach the commercialization stage. This includes the very sluggish methanol oxidation reaction, methanol crossover through the polymeric proton exchange membrane, CO2 evolvement at the anode (Argyropoulos *et al*., 1999a,1999b; Nordlund *et al*., 2002; Scott *et al*., 1998), and cathode flooding (Amphlett *et al*., 2001; M. Mench *et al*., 2001; X. Ren & Gottesfeld, 2001; Z.H. Wang *et al*., 2001). The methanol crossover through the polymer electrolyte leads to a mixed potential at the cathode, which results from the ORR and the methanol oxidation occurring simultaneously. This effect causes a negative potential shift at the cathode and a significant decrease of performance in the DMFC. Methanol crossover also causes fuel loses; it had been found that over 40% of methanol can be wasted in a DMFC across Nafion® membranes (Narayanan *et al*., 1996). In a DMFC, cathode flooding, which typically occurs unless high cathode stoichiometries are used, can determine to a great extent overall cell performance (Amphlett *et al*., 2001; M. Mench *et al*., 2001; X. Ren & Gottesfeld, 2001; Z.H. Wang *et al*., 2001). Water management in the DMFC is especially critical because anode water activity is near unity due to contact with liquid methanol solution (M.W. Mench & C.Y. Wang, 2003). Thus, unlike a H2-PEMFC, no back-diffusive flux of water from cathode to anode will occur, and as a result, vapourization into dry cathode flow is the only pathway for removal of excess cathode-side water accumulation from electro-osmotic drag, ORR, and diffusion (M.W. Mench & C.Y. Wang, 2003).

#### **4.4 Direct ethanol fuel cell**

Direct fuel utilization will be of interest. Besides methanol, other alcohols, particularly those coming from biomass resources, are being considered as alternative fuels. Ethanol as an attractive fuel for electrical vehicles was investigated in direct ethanol fuel cells (Fujiwara *et al*., 1999; Gong *et al*., 2001; Lamy *et al*., 2001; W.J. Zhou *et al*., 2004). However, multimetallic catalysts are necessary to orientate the oxidation reaction selectively in the direction of complete combustion to carbon dioxide (Lamy *et al*., 2001). The reaction mechanisms of anodic oxidation of ethanol are more difficult to elucidate than methanol oxidation, since the number of electrons exchanged greatly increases (12 electrons per ethanol molecule versus 6 electrons for methanol), thus many adsorbed intermediates and products are involved (Lamy *et al*., 2001). Direct ethanol fuel cell was the second most studied fuel cell after methanol. The proton conductor membranes used are mainly Nafion® membranes. However some investigators used high temperature membrane such as Nafion®/Silica by Aricò *et al.* (1998), and also by J. Wang *et al.* (1995) using a phosphoric acid doped polybenzimidazol (PBI) membrane.

#### **4.5 Direct DME fuel cell**

Serov and Kwak (2009) have summarized the recent progress in development of direct DME fuel cell (DDMEFC). DDMEFC could be a valuable direct liquid fuel cell candidate for commercialization. However, compared with PEMFC and DMFC, DDMEFC show poor performances under ambient conditions due to the poor electrooxidation reactivity of DME (Colbow *et al.*, 2000; Kerangueven *et al.*, 2006; Mench *et al.*, 2004; Mizutani *et al.*, 2006; Ueda *et al.*, 2006). On the anode side of DDMEFC, the following oxidation reaction takes place:

CH3OCH3 + 3H2O 2CO2 + 12 H+ + 12 e-

Organic / Inorganic Nanocomposite Membranes

**Nafion® series (DuPont)** 

**Dow Chemicals Co.** 

**Flemion® series (Asahi Glass Co.)** 

Table 2. Perfluorinated membranes

**Aciplex® series (Asahi Chemicals Industry)** 

**5.2 Partially fluorinated ionomer membranes** 

procedure (Kerres, 2001; Wei *et al.*, 1995b).

Development for Low Temperature Fuel Cell Applications 513

**Membrane Thickness (m) Equivalent Weight** 

Nafion® 117 175 1100 Nafion® 115 125 1100 Nafion® 112 50 1100 Nafion® 111 25 1100 Nafion® 1135 87 1100 Nafion® 1035 87 960 Nafion® 105 125 960

Dow® XUS 13204.10 127 800\_850

Flemion® R 50 900 Flemion® S 80 900 Flemion® T 120 900

Aciplex® 1004 100 1000

**5.2.1 Sulfonated copolymer based on the α,β,β-trifluorostyrene monomer membranes**  The Canadian Ballard company developed proton conductor membranes based on trifluorostyrene monomer, under the trade name BAM1G and BAM2G (Ballard Advanced Materials first and second generation, respectively). The longevity of these polymers was limited to approximately 500 hours under practical fuel cell operating conditions (Savadogo, 1998). Based on the above work, Ballard developed third generation membranes under the trade name BAM3G (Steck, 1995; Steck & Stone, 1997; Wei *et al.*, 1995a,1995b). The BAM3G membranes consist of sulfonated copolymers incorporating α,β,β-trifluorostyrene and a series of substituted α,β,β-trifluorostyrene co-monomers. These membranes have an equivalent weight ranging between 375 and 920. The water content of the sulfonated BAM3G is much higher than that of Nafion® and Dow membranes. BAM3G membranes demonstrated a lifetime approaching 15,000 hours when tested in a Ballard MK5 single cell and also exhibited performances superior to Nafion® and Dow® membranes in a H2/O2 fuel cell. Disadvantages of these membranes include the complicated production process for the monomer α,β,β-trifluorostyrene (Livingston *et al.*, 1956) and the difficult sulfonation

The number of electron transferred for complete oxidation is 12, this result in a reduced theoretical fuel requirement of DME, when compared to methanol with 6 electrons transferred, and hydrogen with 2 electrons (Mench *et al.*, 2004; Serov & Kwak, 2009; K. Xu *et al.*, 2010). Furthermore, DME has the advantage over methanol in that crossover is much less pronounced (Mench *et al.*, 2004; Serov & Kwak, 2009). DDMEFC performance increase with increasing temperature, since DME electrooxidation is favored at high temperature. Furthermore, increasing the temperature will enhance also oxygen reduction reaction (ORR) (K. Xu *et al.*, 2010). Compared with hydrogen as the fuel for PEMFC, more water is needed for DME electrooxidation reaction and H+ migration from anode side to cathode side due to the elctro osmotic force of water (Ferrell *et al.*, 2010; K. Xu *et al.*, 2010; Yu *et al.*, 2005).

#### **5. Organic proton conductor membranes**

Proton exchange membranes or proton conductor membranes are the most important component of low temperature fuel cells. Since the development of a solid polymer electrolyte, all the research on fuel cells focuses on the use of these types of electrolyte.

#### **5.1 Perfluorinated membranes**

The first commercially available perfluorinated membrane material from DuPont was Nafion® 120 (1200 equivalent weight (EW), 250 µm thick) followed by Nafion® 117 (1100 EW, 175 µm thick). These high equivalent weight materials were found to have limited use in fuel cells. In 1988, The Dow Chemical Company developed their own perfluorinated polymer membrane with low equivalent weight, typically in the range of 800-850. Nafion® of DuPont and Dow® membranes have identical backbones and are structurally and morphologically similar, but the side chain is shorter in the Dow polymer. Since the success of Dow Chemical, where it was found that the Dow® membrane performed better than the DuPont membrane in H2/O2 fuel cells, DuPont has been active in further developing their membranes with respect to durability and continuous improvement. They increased power densities by further decreasing the equivalent weight from 1100 to 1000 EW and membrane thickness from 175 to 25 µm. Table 2 shows the latest DuPont membranes with some characteristics. Nafion® 117 is the preferred membrane for DMFCs. In the 1990's, Aciplex® perfluorinated ion exchange membranes were introduced by the Asahi Chemical Industry, and the Flemion® series were introduced by Asahi Glass Co. (Yoshida *et al.*, 1998). In general these membranes are in the category of long chain perfluorinated membranes, like Nafion®. Some characteristics of these perfluorinated membranes are summarized in Table 2.

Nafion® membranes are chemically synthesized in four steps according to the DuPont de Nemours process (Grot, 1978): 1) The reaction of tetrafluoroethylene with SO3 to form the sulfone cycle; 2) The condensation of these products with sodium carbonate followed by copolymerization with tetrafluoroethylene to form an insoluble resin; 3) The hydrolysis of this resin to form a perfluorosulfonic polymer and 4) The chemical exchange of the counter ion Na+ with the proton in an appropriate electrolyte. The Dow® membrane is prepared by the co-polymerisation of tetrafluoroethylene with vinylether monomer. The polymer can be described as having a Teflon-like backbone structure with a side chain attached via an ether group. This side chain is characterized by a terminal sulfonate functional group (Savadogo, 1998).

The number of electron transferred for complete oxidation is 12, this result in a reduced theoretical fuel requirement of DME, when compared to methanol with 6 electrons transferred, and hydrogen with 2 electrons (Mench *et al.*, 2004; Serov & Kwak, 2009; K. Xu *et al.*, 2010). Furthermore, DME has the advantage over methanol in that crossover is much less pronounced (Mench *et al.*, 2004; Serov & Kwak, 2009). DDMEFC performance increase with increasing temperature, since DME electrooxidation is favored at high temperature. Furthermore, increasing the temperature will enhance also oxygen reduction reaction (ORR) (K. Xu *et al.*, 2010). Compared with hydrogen as the fuel for PEMFC, more water is needed for DME electrooxidation reaction and H+ migration from anode side to cathode side due to

the elctro osmotic force of water (Ferrell *et al.*, 2010; K. Xu *et al.*, 2010; Yu *et al.*, 2005).

Proton exchange membranes or proton conductor membranes are the most important component of low temperature fuel cells. Since the development of a solid polymer electrolyte, all the research on fuel cells focuses on the use of these types of electrolyte.

The first commercially available perfluorinated membrane material from DuPont was Nafion® 120 (1200 equivalent weight (EW), 250 µm thick) followed by Nafion® 117 (1100 EW, 175 µm thick). These high equivalent weight materials were found to have limited use in fuel cells. In 1988, The Dow Chemical Company developed their own perfluorinated polymer membrane with low equivalent weight, typically in the range of 800-850. Nafion® of DuPont and Dow® membranes have identical backbones and are structurally and morphologically similar, but the side chain is shorter in the Dow polymer. Since the success of Dow Chemical, where it was found that the Dow® membrane performed better than the DuPont membrane in H2/O2 fuel cells, DuPont has been active in further developing their membranes with respect to durability and continuous improvement. They increased power densities by further decreasing the equivalent weight from 1100 to 1000 EW and membrane thickness from 175 to 25 µm. Table 2 shows the latest DuPont membranes with some characteristics. Nafion® 117 is the preferred membrane for DMFCs. In the 1990's, Aciplex® perfluorinated ion exchange membranes were introduced by the Asahi Chemical Industry, and the Flemion® series were introduced by Asahi Glass Co. (Yoshida *et al.*, 1998). In general these membranes are in the category of long chain perfluorinated membranes, like Nafion®. Some

characteristics of these perfluorinated membranes are summarized in Table 2.

Nafion® membranes are chemically synthesized in four steps according to the DuPont de Nemours process (Grot, 1978): 1) The reaction of tetrafluoroethylene with SO3 to form the sulfone cycle; 2) The condensation of these products with sodium carbonate followed by copolymerization with tetrafluoroethylene to form an insoluble resin; 3) The hydrolysis of this resin to form a perfluorosulfonic polymer and 4) The chemical exchange of the counter ion Na+ with the proton in an appropriate electrolyte. The Dow® membrane is prepared by the co-polymerisation of tetrafluoroethylene with vinylether monomer. The polymer can be described as having a Teflon-like backbone structure with a side chain attached via an ether group. This side chain is characterized by a terminal sulfonate functional group (Savadogo,

**5. Organic proton conductor membranes** 

**5.1 Perfluorinated membranes** 

1998).


Table 2. Perfluorinated membranes

#### **5.2 Partially fluorinated ionomer membranes**

#### **5.2.1 Sulfonated copolymer based on the α,β,β-trifluorostyrene monomer membranes**

The Canadian Ballard company developed proton conductor membranes based on trifluorostyrene monomer, under the trade name BAM1G and BAM2G (Ballard Advanced Materials first and second generation, respectively). The longevity of these polymers was limited to approximately 500 hours under practical fuel cell operating conditions (Savadogo, 1998). Based on the above work, Ballard developed third generation membranes under the trade name BAM3G (Steck, 1995; Steck & Stone, 1997; Wei *et al.*, 1995a,1995b). The BAM3G membranes consist of sulfonated copolymers incorporating α,β,β-trifluorostyrene and a series of substituted α,β,β-trifluorostyrene co-monomers. These membranes have an equivalent weight ranging between 375 and 920. The water content of the sulfonated BAM3G is much higher than that of Nafion® and Dow membranes. BAM3G membranes demonstrated a lifetime approaching 15,000 hours when tested in a Ballard MK5 single cell and also exhibited performances superior to Nafion® and Dow® membranes in a H2/O2 fuel cell. Disadvantages of these membranes include the complicated production process for the monomer α,β,β-trifluorostyrene (Livingston *et al.*, 1956) and the difficult sulfonation procedure (Kerres, 2001; Wei *et al.*, 1995b).

Organic / Inorganic Nanocomposite Membranes

**5.3.2 Sulfonated polyimide membranes** 

drops considerably (Kerres, 2001).

Nafion® 117 (Woo *et al.,* 2003).

1999).

**5.3.3 Phosphazene-based cation-exchange membranes** 

Development for Low Temperature Fuel Cell Applications 515

can only be used with a feed of vapourized methanol, because when a liquid contacts the membrane, the phosphoric acid leaches out of the membrane and the proton conductivity

The sulfonated polyimide (SPI) membranes were obtained by casting on a glass plate the polymer solution and evaporating the solvent (Cornet *et al.,* 2000; Fan *et al.,* 2002; Faure *et al.,* 1996, 1997*;* Gebel *et al.,* 1993; Genies *et al.,* 2001; Woo *et al.,* 2003). The polymer solution synthesis was achieved in different ways: The first way was based on the phthalimide-five member imide (4,4'-diamino-biphenyl 2,2' disulfonic acid (BDSA), 4,4' oxy-diphthalic dianhydride (ODPA) and 4,4'–oxydianiline (ODA)) at 200°C. The second way was based on the naphthalimide-six member imide ring (BDSA, 1,4, 5,8-naphthalene tetracarboxylic dianhydride (NTDA) and ODA) at 160°C (Faure *et al.,* 1996, 1997*;* Gebel *et al.,* 1993). The third way was based on the 3,3`,4,4`-benzophenone-tetracarboxylic dianhydride (BTDA), BDSA and ODA (D`Alelio, 1944). The fourth way was based on BDSA/NTDA/mAPI (bis- [3-(Aminophenoxy)-4-phenyl]isopropylidene) (Genies *et al.,* 2001). The water content of membranes at 25°C for the phthalic and naphthalenic sulfonated polyimide membranes is 26% and 30%, respectively. The water content obtained for Nafion® membranes under the same conditions was 20% (Faure *et al.,* 1997*;* Gebel *et al.,* 1993). It was also claimed that the sulfonated polyimide membranes were 3 times less permeable to hydrogen gas than Nafion® membranes. The lifetime measurements were performed on a 175 m phthalic polyimide and a 70 m naphthalenic sulfonated polyimide film at 60°C, 3 bar pressure for H2 and O2 and under a constant current density. It was found that the membrane based on the phthalic structure broke after 70 hours whereas the membrane based on the naphthalic polyimide was stable over 3000 hours (D`Alelio, 1944). The proton conductivity of SPI was found to be half of Nafion® 117, typically 4.1 10-2 S/cm, and methanol permeability was found to be 7.34 10-8 compared to 2.38 10-6 cm2/s for

It was shown that polyphosphazene-based cation-exchange membranes have a low methanol permeability, low water swelling ratios, satisfactory mechanical properties, and a conductivity comparable to that of Nafion® 117 (Allcock *et al.,* 2002a,2002b; Guo *et al.,* 1999; Tang *et al.,* 1999; Wycisk & Pintauro, 1996; X. Zhou *et al.,* 2003). Polyphosphazene-based membranes have been fabricated from poly[bis(3-methylphenoxy)phosphazene] by first sulfonating the base polymer with SO3 and then solution-casting a thin film (Tang *et al.,*  1999; Wycisk & Pintauro, 1996; X. Zhou *et al.,* 2003). Polymer crosslinking was carried out by dissolving benzophenone photoinitiator in the membrane casting solution and then exposing the resulting films after solvent evaporation to UV light (Guo *et al.,* 1999). The conductivity of the polyphosphazene membranes were either similar to or lower than that of Nafion® 117 membranes (Guo *et al.,* 1999; X. Zhou *et al.,* 2003). However, methanol permeability of a sulfonated membrane was about 8 times lower than that of the Nafion® 117 membrane (X. Zhou *et al.,* 2003). Sulfonated/crosslinked polyphosphazene films showed no signs of mechanical failure (softening) up to 173°C and a pressure of 800 kPa (Guo *et al.,* 

#### **5.2.2 Grafted ionomer membranes**

Partially fluorinated membranes can be obtained by using a simultaneous and pre-radiation grafting of monomers onto a base polymer film, and subsequent sulfonation of the grafted component (Brack *et al.*, 2003; Büchi *et al.*, 1995a,1995b; Gode *et al.*, 2003; Gupta *et al.*, 1993; Hatanaka *et al.*, 2002; W. Lee *et al.*, 1996; Lehtinen *et al.*, 1998; Scherer, 1990). These membranes were prepared by pre-irradiation of fluoropolymer films, such as poly(tetrafluoroethylene-*co*-hexafluoropropylene (FEP) or poly(ethylene-*alt*-tetrafluoroethylene) (ETFE), using an electron beam or gamma irradiation source. The pre-irradiated films were grafted by exposing them to solutions of styrene and other radically polymerizable monomers. The grafted films are sulfonated using chlorosulfonic acid. The grafting mixture was crosslinked with divinylbenzene (DVB) and tri-allyl cyamirate (TAC) (Gupta *et al.*, 1994; Gupta & Scherer, 1994) or poly(vinylidene fluoride) (Sundholm, 1998). A disadvantage of membranes using styrene and divinylbenzene monomers is that their oxidation stability is limited, due to the tertiary C-H bonds which are sensitive to O2 and hydrogen peroxide attack (Kerres, 2001).

#### **5.3 Non-perfluorinated membranes**

#### **5.3.1 Polybenzimidazole (PBI)**

PBI is synthesized from aromatic bis-o-diamines and dicarboxylates (acids, esters, amides), either in the molten state or in solution (Jones & Rozière, 2001). PBI is relatively low cost and is a commercially available polymer known to have excellent oxidation and thermal stability. The commercially available polybenzimidazol is poly-[2,2`-(*m*-phenylene)-5,5` bibenzimidazole], which is synthesized from diphenyl-iso-phthalate and tetraaminobiphenyl. Hoel and Grunwald (1977) reported on proton conductivity values of PBI in the range of 2 10-4 – 8 10-4 S/cm at relative humidities (RH) between 0 and 100%. Other authors observed proton conductivity some two to three orders of magnitude lower (Aharoni & Litt, 1974; Glipa *et al.*, 1997; Powers & Serad, 1986). PBI is a suitable basic polymer which can readily be complexed with strong acids (Jones & Rozière, 2001; Glipa *et al.*, 1997; Y.L. Ma *et al.*, 2004; Powers & Serad, 1986; Samms *et al.*, 1996; Savadogo & B. Xing, 2000; Wainright *et al.*, 1995, 1997; J.T. Wang *et al.*, 1996a,1996b; B. Xing & Savadogo, 1999). The immersion of PBI film in aqueous phosphoric acid leads to an increase in both its conductivity and thermal stability (Powers & Serad, 1986). Savinell and co-workers prepared PBI/H3PO4 via two different routes: a) directly casting a film of PBI from a solution containing phosphoric acid; b) preparation by immersion of a preformed PBI membrane in 11M phosphoric acid for several days (Samms *et al.*, 1996; Wainright *et al.*, 1997). The typical thickness for different films was 75 µm. The conductivity depends on the quantity of phosphoric acid in the membrane. Conductivity in the range 5 10-3 to 2 10-2 S/cm at 130°C and 5 10-2 S/cm at 190°C have been reported (Wainright *et al.*, 1995). The conductivity for type "a" membranes is higher than those of type "b" membranes. At a temperature above 150°C, the conductivity of type "a" membranes is similar to that of Nafion® at 80 °C and 100% RH. It was shown that the methanol crossover through doped PBI type "a" membrane, was at least ten times less than that observed with Nafion®. The disadvantage of these membranes is that the H3PO4 molecules can diffuse out of the membrane towards basic polymer sites because they are in excess. PBI/H3PO4 membranes are suitable for direct methanol fuel cell application at a temperature >100°C. However, they can only be used with a feed of vapourized methanol, because when a liquid contacts the membrane, the phosphoric acid leaches out of the membrane and the proton conductivity drops considerably (Kerres, 2001).

#### **5.3.2 Sulfonated polyimide membranes**

514 Advances in Chemical Engineering

Partially fluorinated membranes can be obtained by using a simultaneous and pre-radiation grafting of monomers onto a base polymer film, and subsequent sulfonation of the grafted component (Brack *et al.*, 2003; Büchi *et al.*, 1995a,1995b; Gode *et al.*, 2003; Gupta *et al.*, 1993; Hatanaka *et al.*, 2002; W. Lee *et al.*, 1996; Lehtinen *et al.*, 1998; Scherer, 1990). These membranes were prepared by pre-irradiation of fluoropolymer films, such as poly(tetrafluoroethylene-*co*-hexafluoropropylene (FEP) or poly(ethylene-*alt*-tetrafluoroethylene) (ETFE), using an electron beam or gamma irradiation source. The pre-irradiated films were grafted by exposing them to solutions of styrene and other radically polymerizable monomers. The grafted films are sulfonated using chlorosulfonic acid. The grafting mixture was crosslinked with divinylbenzene (DVB) and tri-allyl cyamirate (TAC) (Gupta *et al.*, 1994; Gupta & Scherer, 1994) or poly(vinylidene fluoride) (Sundholm, 1998). A disadvantage of membranes using styrene and divinylbenzene monomers is that their oxidation stability is limited, due to the tertiary C-H bonds which are sensitive to O2 and

PBI is synthesized from aromatic bis-o-diamines and dicarboxylates (acids, esters, amides), either in the molten state or in solution (Jones & Rozière, 2001). PBI is relatively low cost and is a commercially available polymer known to have excellent oxidation and thermal stability. The commercially available polybenzimidazol is poly-[2,2`-(*m*-phenylene)-5,5` bibenzimidazole], which is synthesized from diphenyl-iso-phthalate and tetraaminobiphenyl. Hoel and Grunwald (1977) reported on proton conductivity values of PBI in the range of 2 10-4 – 8 10-4 S/cm at relative humidities (RH) between 0 and 100%. Other authors observed proton conductivity some two to three orders of magnitude lower (Aharoni & Litt, 1974; Glipa *et al.*, 1997; Powers & Serad, 1986). PBI is a suitable basic polymer which can readily be complexed with strong acids (Jones & Rozière, 2001; Glipa *et al.*, 1997; Y.L. Ma *et al.*, 2004; Powers & Serad, 1986; Samms *et al.*, 1996; Savadogo & B. Xing, 2000; Wainright *et al.*, 1995, 1997; J.T. Wang *et al.*, 1996a,1996b; B. Xing & Savadogo, 1999). The immersion of PBI film in aqueous phosphoric acid leads to an increase in both its conductivity and thermal stability (Powers & Serad, 1986). Savinell and co-workers prepared PBI/H3PO4 via two different routes: a) directly casting a film of PBI from a solution containing phosphoric acid; b) preparation by immersion of a preformed PBI membrane in 11M phosphoric acid for several days (Samms *et al.*, 1996; Wainright *et al.*, 1997). The typical thickness for different films was 75 µm. The conductivity depends on the quantity of phosphoric acid in the membrane. Conductivity in the range 5 10-3 to 2 10-2 S/cm at 130°C and 5 10-2 S/cm at 190°C have been reported (Wainright *et al.*, 1995). The conductivity for type "a" membranes is higher than those of type "b" membranes. At a temperature above 150°C, the conductivity of type "a" membranes is similar to that of Nafion® at 80 °C and 100% RH. It was shown that the methanol crossover through doped PBI type "a" membrane, was at least ten times less than that observed with Nafion®. The disadvantage of these membranes is that the H3PO4 molecules can diffuse out of the membrane towards basic polymer sites because they are in excess. PBI/H3PO4 membranes are suitable for direct methanol fuel cell application at a temperature >100°C. However, they

**5.2.2 Grafted ionomer membranes** 

hydrogen peroxide attack (Kerres, 2001).

**5.3 Non-perfluorinated membranes** 

**5.3.1 Polybenzimidazole (PBI)** 

The sulfonated polyimide (SPI) membranes were obtained by casting on a glass plate the polymer solution and evaporating the solvent (Cornet *et al.,* 2000; Fan *et al.,* 2002; Faure *et al.,* 1996, 1997*;* Gebel *et al.,* 1993; Genies *et al.,* 2001; Woo *et al.,* 2003). The polymer solution synthesis was achieved in different ways: The first way was based on the phthalimide-five member imide (4,4'-diamino-biphenyl 2,2' disulfonic acid (BDSA), 4,4' oxy-diphthalic dianhydride (ODPA) and 4,4'–oxydianiline (ODA)) at 200°C. The second way was based on the naphthalimide-six member imide ring (BDSA, 1,4, 5,8-naphthalene tetracarboxylic dianhydride (NTDA) and ODA) at 160°C (Faure *et al.,* 1996, 1997*;* Gebel *et al.,* 1993). The third way was based on the 3,3`,4,4`-benzophenone-tetracarboxylic dianhydride (BTDA), BDSA and ODA (D`Alelio, 1944). The fourth way was based on BDSA/NTDA/mAPI (bis- [3-(Aminophenoxy)-4-phenyl]isopropylidene) (Genies *et al.,* 2001). The water content of membranes at 25°C for the phthalic and naphthalenic sulfonated polyimide membranes is 26% and 30%, respectively. The water content obtained for Nafion® membranes under the same conditions was 20% (Faure *et al.,* 1997*;* Gebel *et al.,* 1993). It was also claimed that the sulfonated polyimide membranes were 3 times less permeable to hydrogen gas than Nafion® membranes. The lifetime measurements were performed on a 175 m phthalic polyimide and a 70 m naphthalenic sulfonated polyimide film at 60°C, 3 bar pressure for H2 and O2 and under a constant current density. It was found that the membrane based on the phthalic structure broke after 70 hours whereas the membrane based on the naphthalic polyimide was stable over 3000 hours (D`Alelio, 1944). The proton conductivity of SPI was found to be half of Nafion® 117, typically 4.1 10-2 S/cm, and methanol permeability was found to be 7.34 10-8 compared to 2.38 10-6 cm2/s for Nafion® 117 (Woo *et al.,* 2003).

#### **5.3.3 Phosphazene-based cation-exchange membranes**

It was shown that polyphosphazene-based cation-exchange membranes have a low methanol permeability, low water swelling ratios, satisfactory mechanical properties, and a conductivity comparable to that of Nafion® 117 (Allcock *et al.,* 2002a,2002b; Guo *et al.,* 1999; Tang *et al.,* 1999; Wycisk & Pintauro, 1996; X. Zhou *et al.,* 2003). Polyphosphazene-based membranes have been fabricated from poly[bis(3-methylphenoxy)phosphazene] by first sulfonating the base polymer with SO3 and then solution-casting a thin film (Tang *et al.,*  1999; Wycisk & Pintauro, 1996; X. Zhou *et al.,* 2003). Polymer crosslinking was carried out by dissolving benzophenone photoinitiator in the membrane casting solution and then exposing the resulting films after solvent evaporation to UV light (Guo *et al.,* 1999). The conductivity of the polyphosphazene membranes were either similar to or lower than that of Nafion® 117 membranes (Guo *et al.,* 1999; X. Zhou *et al.,* 2003). However, methanol permeability of a sulfonated membrane was about 8 times lower than that of the Nafion® 117 membrane (X. Zhou *et al.,* 2003). Sulfonated/crosslinked polyphosphazene films showed no signs of mechanical failure (softening) up to 173°C and a pressure of 800 kPa (Guo *et al.,*  1999).

Organic / Inorganic Nanocomposite Membranes

**6. Organic/inorganic nanocomposite membranes** 

(<80°C). The advantages of PFSA membranes are:

polytetrafluoroethylene backbone; and

electrolyte fuel cell.

Membrane electrode assembly (MEA) is the basic component of the single cell of a stack. The proton exchange membrane (PEM) is the key element of this component, which separates the electrode structure to prevent the mixing of reactant gases and the formation of an electrical short. This makes its properties, functionality, cost and reliability very important for real cell operations. Up to now perfluorinated sulfonic acid (PFSA) membranes have been the best choice for commercial low temperature polymer products

i. Their strong stability in oxidative and reduction media due to the structure of the

ii. Their proton conductivity, which can be as high as 0.2 S cm-1 in full hydrated polymer

When used at elevated temperatures, however, PEMFC performances decrease. This decrease is related to: dehydration of the membrane; reduction of ionic conductivity;

2002).

Development for Low Temperature Fuel Cell Applications 517

be carried out directly in concentrated sulfonic acid or oleum - the extent of sulfonation being controlled by the reaction time and temperature (Bailly *et al.,* 1987; B. Bauer *et al.,* 1994, 1995). Direct sulfonation of PEEK can give materials with a wide range of equivalent weights to form SPEEK. However, the complete sulfonation of the polymer results in a fully water-soluble product. A sulfonation level of around 60% was found to be a good compromise between the conductivity and mechanical properties of membranes. The backbone of SPEEK is less hydrophobic than the backbone of Nafion®, and the sulfonic acid functional group is less acidic (Kreuer, 2001). Various studies have been made on the conductivity of SPEEK (Alberti *et al.,* 2001; B. Bauer *et al.,* 2000; Kobayashi *et al.,* 1998; Kreuer, 1997, 2001; Linkous *et al.,* 1998; P. Xing *et al.,* 2004; Zaidi *et al.,* 2000). The conductivity increases as a function of the degree of sulfonation, the ambient relative humidity, temperature and thermal history. The conductivity of these materials was found to be high at room temperature (Soczka-Guth *et al.,* 1999). In SPEEK with 65% sites sulfonated, the conductivity was higher than that of Nafion® 117 measured under the same conditions - the conductivity reaching 4 10-2 S/cm at 100°C and 100% RH (Linkous *et al.,*  1998). SPEEK membranes exhibit at 160°C and 75% RH, sufficiently high values of protonic conductivity - typically 5 10-2 – 6 10-2 S/cm - for possible applications in low temperature fuel cells (Alberti *et al.,* 2001). The dependence of the conductivity on RH is more marked for SPEEK than for Nafion® under the same conditions (B. Bauer *et al.,* 2000). Sulfonated polyaryls have been demonstrated to suffer from hydroxyl radical initiated degradation (Hubner & Roduner, 1999). In contrast, SPEEK was found to be durable under fuel cell conditions over several thousand hours by Kreuer (2001). The brittleness of SPEEK makes their handling difficult and may lead to mechanical membrane failure during operation. These types of membranes become very brittle when drying out. SPEEK can also be chemically cross-linked to reduce membrane swelling and increase its mechanical strength. Materials prepared by cross-linking are comparable to commercial Nafion® in terms of their mechanical strength and proton conductivity (Yen *et al.,* 1998). Kerres and co-workers prepared novel acid-base polymer blend membranes composed of SPEEK as the acidic compound, and of P4VP or PBI as the basic compounds (Kerres *et al.,* 1999; Jörissen *et al.,* 

#### **5.3.4 Sulfonated poly(arylethersulfone) membranes**

Polysulfone (PSU) is a low cost, commercially available polymer (e.g. PSU Udel™ from Amoco) which has very good chemical stability. The synthesis and characterization of sulfonated polysulfone (SPSU) has been achieved by Johnson *et al.* (1984) and Nolte *et al.*  (1993). It was found that membranes cast from SPSU (Udel™ P-1700) solutions were completely water soluble (Nolte *et al.,* 1993) and become very brittle when drying out which can happen in the fuel cell application under intermittent conditions (Kerres *et al.,* 1999).

There are two new but different procedures for the sulfonation of polysulfone. In one procedure, the sodium-sulfonated group was introduced in the base polysulfone via the metalation-sulfination-oxidation process (Kerres *et al.,* 1996, 1998a). In the other procedure, trimethylsilyl chlorosulfonate was used as the sulfonating agent (Baradie *et al.,* 1998). Lufrano *et al.* (2000,2001) prepared SPSU via trimethylsilyl chlorosulfonate with different degrees of sulfonation (DS). Different membranes with sulfonation degree from 23% to 53% (Lufrano *et al.,* 2000) on the one hand and 49%, 61% and 77% (Lufrano *et al.,* 2001) on the other hand were prepared. With a 61% sulfonation degree a proton conductivity for SPSU of 2.7 10-2 S/cm at 25°C was reported (Lufrano *et al.,* 2001). This conductivity was 3.5 times lower than Nafion® 117, but was compensated by the lower thickness, 90 µm vs. 210 µm for Nafion® 117. The cell performance obtained by Lufrano *et al.* (2001) was almost the same for SPSU and Nafion® in a H2/O2 fuel cell. This is higher than that reported previously by Kerres *et al.* (1998a) and Baradie *et al.* (1998). Y.S. Kim *et al.* (2003) prepared sulfonated poly(arylether sulfone) membranes.

Promising alternatives suggested by Kerres and co-workers, include composite membranes made from blends of acidic and basic polymers (Cui *et al.,* 1998; Jörissen *et al.,* 2002; Kerres *et al.,* 1999, 2000; Walker *et al.,* 1999) or modified PSU via the metalation-sulfochlorination and the metalation-amination routes (W. Zhang *et al.,* 2001) or crosslinked SPSU (Kerres *et al.,*  1997, 1998b,1998c). These alternatives are made by blending acidic polymers such as SPSU with basic polymers such as poly(4-vinylpyridine) (P4VP), polybenzimidazole (PBI) or a basically substituted polysulfone. Crosslinked SPSU blend membranes have been produced via a new crosslinking process. The blends have been obtained by mixing PSU Udel™ Nasulfonate and PSU Udel™ Li-sulfinate in *N*-methyl pyrrolidone. The membranes have been crosslinked by S-alkylation of PSU sulfinate groups with di-halogenoalkanes. These membranes show very good performance in H2/O2 fuel cells and DMFCs (Kerres *et al.,* 2000; Kerres, 2001). These membranes also show a markedly reduced methanol permeability (Kerres*,* 2001; Walker *et al.,* 1999).

#### **5.3.5 Sulfonated poly(aryletherketone) membranes**

The poly(arylether ketones) are a class of non-fluorinated polymers consisting of sequences of ether and carbonyl linkages between phenyl rings, that can either "ether-rich" like PEEK and PEEKK, or "ketone-rich" like PEK and PEKEKK. The most common material is polyetheretherketone (PEEK) which is commercially available under the name Victrex™ PEEK from ICI Advanced Materials. A number of groups are developing proton conducting polymer materials based on this classification of materials including ICI Victrex, Fuma-Tech and Axiva/Aventis/Hoechst. Sulfonated-PEEK (SPEEK) membranes were prepared as proton conductors in PEMFCs by Schneller *et al.* (1993). Sulfonation of polyetherketones can

Polysulfone (PSU) is a low cost, commercially available polymer (e.g. PSU Udel™ from Amoco) which has very good chemical stability. The synthesis and characterization of sulfonated polysulfone (SPSU) has been achieved by Johnson *et al.* (1984) and Nolte *et al.*  (1993). It was found that membranes cast from SPSU (Udel™ P-1700) solutions were completely water soluble (Nolte *et al.,* 1993) and become very brittle when drying out which can happen in the fuel cell application under intermittent conditions (Kerres *et al.,* 1999).

There are two new but different procedures for the sulfonation of polysulfone. In one procedure, the sodium-sulfonated group was introduced in the base polysulfone via the metalation-sulfination-oxidation process (Kerres *et al.,* 1996, 1998a). In the other procedure, trimethylsilyl chlorosulfonate was used as the sulfonating agent (Baradie *et al.,* 1998). Lufrano *et al.* (2000,2001) prepared SPSU via trimethylsilyl chlorosulfonate with different degrees of sulfonation (DS). Different membranes with sulfonation degree from 23% to 53% (Lufrano *et al.,* 2000) on the one hand and 49%, 61% and 77% (Lufrano *et al.,* 2001) on the other hand were prepared. With a 61% sulfonation degree a proton conductivity for SPSU of 2.7 10-2 S/cm at 25°C was reported (Lufrano *et al.,* 2001). This conductivity was 3.5 times lower than Nafion® 117, but was compensated by the lower thickness, 90 µm vs. 210 µm for Nafion® 117. The cell performance obtained by Lufrano *et al.* (2001) was almost the same for SPSU and Nafion® in a H2/O2 fuel cell. This is higher than that reported previously by Kerres *et al.* (1998a) and Baradie *et al.* (1998). Y.S. Kim *et al.* (2003) prepared sulfonated

Promising alternatives suggested by Kerres and co-workers, include composite membranes made from blends of acidic and basic polymers (Cui *et al.,* 1998; Jörissen *et al.,* 2002; Kerres *et al.,* 1999, 2000; Walker *et al.,* 1999) or modified PSU via the metalation-sulfochlorination and the metalation-amination routes (W. Zhang *et al.,* 2001) or crosslinked SPSU (Kerres *et al.,*  1997, 1998b,1998c). These alternatives are made by blending acidic polymers such as SPSU with basic polymers such as poly(4-vinylpyridine) (P4VP), polybenzimidazole (PBI) or a basically substituted polysulfone. Crosslinked SPSU blend membranes have been produced via a new crosslinking process. The blends have been obtained by mixing PSU Udel™ Nasulfonate and PSU Udel™ Li-sulfinate in *N*-methyl pyrrolidone. The membranes have been crosslinked by S-alkylation of PSU sulfinate groups with di-halogenoalkanes. These membranes show very good performance in H2/O2 fuel cells and DMFCs (Kerres *et al.,* 2000; Kerres, 2001). These membranes also show a markedly reduced methanol permeability

The poly(arylether ketones) are a class of non-fluorinated polymers consisting of sequences of ether and carbonyl linkages between phenyl rings, that can either "ether-rich" like PEEK and PEEKK, or "ketone-rich" like PEK and PEKEKK. The most common material is polyetheretherketone (PEEK) which is commercially available under the name Victrex™ PEEK from ICI Advanced Materials. A number of groups are developing proton conducting polymer materials based on this classification of materials including ICI Victrex, Fuma-Tech and Axiva/Aventis/Hoechst. Sulfonated-PEEK (SPEEK) membranes were prepared as proton conductors in PEMFCs by Schneller *et al.* (1993). Sulfonation of polyetherketones can

**5.3.4 Sulfonated poly(arylethersulfone) membranes** 

poly(arylether sulfone) membranes.

(Kerres*,* 2001; Walker *et al.,* 1999).

**5.3.5 Sulfonated poly(aryletherketone) membranes** 

be carried out directly in concentrated sulfonic acid or oleum - the extent of sulfonation being controlled by the reaction time and temperature (Bailly *et al.,* 1987; B. Bauer *et al.,* 1994, 1995). Direct sulfonation of PEEK can give materials with a wide range of equivalent weights to form SPEEK. However, the complete sulfonation of the polymer results in a fully water-soluble product. A sulfonation level of around 60% was found to be a good compromise between the conductivity and mechanical properties of membranes. The backbone of SPEEK is less hydrophobic than the backbone of Nafion®, and the sulfonic acid functional group is less acidic (Kreuer, 2001). Various studies have been made on the conductivity of SPEEK (Alberti *et al.,* 2001; B. Bauer *et al.,* 2000; Kobayashi *et al.,* 1998; Kreuer, 1997, 2001; Linkous *et al.,* 1998; P. Xing *et al.,* 2004; Zaidi *et al.,* 2000). The conductivity increases as a function of the degree of sulfonation, the ambient relative humidity, temperature and thermal history. The conductivity of these materials was found to be high at room temperature (Soczka-Guth *et al.,* 1999). In SPEEK with 65% sites sulfonated, the conductivity was higher than that of Nafion® 117 measured under the same conditions - the conductivity reaching 4 10-2 S/cm at 100°C and 100% RH (Linkous *et al.,*  1998). SPEEK membranes exhibit at 160°C and 75% RH, sufficiently high values of protonic conductivity - typically 5 10-2 – 6 10-2 S/cm - for possible applications in low temperature fuel cells (Alberti *et al.,* 2001). The dependence of the conductivity on RH is more marked for SPEEK than for Nafion® under the same conditions (B. Bauer *et al.,* 2000). Sulfonated polyaryls have been demonstrated to suffer from hydroxyl radical initiated degradation (Hubner & Roduner, 1999). In contrast, SPEEK was found to be durable under fuel cell conditions over several thousand hours by Kreuer (2001). The brittleness of SPEEK makes their handling difficult and may lead to mechanical membrane failure during operation. These types of membranes become very brittle when drying out. SPEEK can also be chemically cross-linked to reduce membrane swelling and increase its mechanical strength. Materials prepared by cross-linking are comparable to commercial Nafion® in terms of their mechanical strength and proton conductivity (Yen *et al.,* 1998). Kerres and co-workers prepared novel acid-base polymer blend membranes composed of SPEEK as the acidic compound, and of P4VP or PBI as the basic compounds (Kerres *et al.,* 1999; Jörissen *et al.,*  2002).

#### **6. Organic/inorganic nanocomposite membranes**

Membrane electrode assembly (MEA) is the basic component of the single cell of a stack. The proton exchange membrane (PEM) is the key element of this component, which separates the electrode structure to prevent the mixing of reactant gases and the formation of an electrical short. This makes its properties, functionality, cost and reliability very important for real cell operations. Up to now perfluorinated sulfonic acid (PFSA) membranes have been the best choice for commercial low temperature polymer products (<80°C). The advantages of PFSA membranes are:


When used at elevated temperatures, however, PEMFC performances decrease. This decrease is related to: dehydration of the membrane; reduction of ionic conductivity;

Organic / Inorganic Nanocomposite Membranes

size) will increase the conductivity (Jones & Rozière, 2001).

**6.1 Organic/silica nanocomposite membranes** 

Development for Low Temperature Fuel Cell Applications 519

ion-exchange (Bonnet *et al.,* 2000; Jones & Rozière, 2001). Such approach generally avoid any sedimentation of the inorganic component, intimacy of contact between the inorganic and organic components at the molecular level assures the greatest possible interface and, at such small particle size, the mechanical properties can be improved compared with those of a polymer-only membrane (Jones & Rozière, 2001, 2003). In addition, since in many proton conductors of conductivity suitable for electrochemical applications the proton transfer process takes place on the surface of the particles, increase in surface area (small particle

This concept was suggested by Watanabe *et al.* (1995, 1996) and is based on the development of self-humidifying composite membranes. The membranes are fabricated from the dispersion of nano-particles of Pt in a thin Nafion® film (≈ 50 µm). Membranes fabricated based on this concept should not require external humidification and should suppress the crossover of H2 and O2. The dispersed particles should catalyze the oxidation and the reduction of the crossover H2 and O2 respectively. Water from this reaction is directly used to humidify the membrane. This is supposed to result in a more stable operation of the cell

at 80°C without any external humidification of the membrane (Watanabe *et al.*, 1998).

Silica as an additive to Nafion® was widely studies. Both recast Nafion® (Adjemian *et al.*, 2002a,2002b; Antonucci *et al.*, 1999; Arimura *et al.*, 1999; Dimitrova *et al.*, 2002a,2002b) and Nafion® film (e.g. Nafion® 117) (Adjemian *et al.*, 2002b; Baradie *et al.*, 2000; Jung *et al.*, 2002) are used in the fabrication of the composite membrane. Organic–silica composite membranes have been prepared according to several methods by casting mixtures such as: using silicon dioxide particles (e.g. Aerosil A380 from Degussa) (Antonucci *et al.*, 1999; Aricò *et al.*, 1998; Arimura *et al.*, 1999; Dimitrova *et al.*, 2002a,2002b), diphenylsilicate (DPS) (Liang *et al.*, 2006), the other one is to introduce silica oxide incorporated via *in situ* sol-gel reaction of tetraethoxysilane (TEOS) (Adjemian *et al.*, 2002b; Baradie *et al.*, 2000; Deng *et al.*, 1998; R.C. R.C. Jiang *et al.*, 2006b; Jung *et al.*, 2002; Mauritz *et al.*, 1995, 1998). Nafion®–silica membranes shows good performance at T > 100°C due to low levels of dehydration. Nafion®–silica membranes were prepared by mixing Nafion® ionomer (5 wt.%) with 3 wt.% SiO2 followed by a regular membrane casting procedure. In the final stage, the membranes were heat treated at 160°C for 10 min to achieve both a high crystallinity and high mechanical stability (Antonucci *et al.*, 1999). A DMFC utilizing these membranes was tested under galvanostatic conditions at 500 mA cm-2. The voltage initially decreased from 0.42 to 0.36 V but then remained stable for 8 h. The performance decrease is due to adsorption of poisoning species, which appears to be a reversible process at 145°C (removed by short circuit discharging in the presence of water). These membranes demonstrated higher performance with increasing temperature. A nano-particles possessed a core-shell structure consisting of silica core (< 10 nm) and a densely grafted oligometric ionomer layer was incorporated into Nafion® matrix to form a composite membrane. The polyelectrolytegrafted silica particles [P(SPA)-SiO2] was dispersed in Nafion® solution and a composite membrane was formed by recasting process. The proton conductivity of Nafion® membrane containing P(SPA)-SiO2 (4 wt.%) is significantly higher than that of unmodified recast Nafion® in the range 25 to 80°C. The composite membrane offers superior cell performance over unmodified recast Nafion® at both operating temperature of 50 and 80°C. At 50°C, the

decreases in affinity with water; loss of mechanical strength through a softening of the polymer backbone; and parasitic losses (the high level of gas permeation). There are several reasons for the development of high temperature membranes (Savadogo, 2004):


One of the main drawbacks of DMFC (direct methanol fuel cell) is the slow methanol oxidation kinetics. An increase in the operating temperature of the DMFC from 90 to about 140°C is highly desirable. Also operation at high temperature will enhanced CO tolerant when a reformate hydrogen is used in H2-PEMFC. One approach to achieve water retention at high temperature is to fabricate a composite membrane constituted of organic proton conductor and inorganic materials. The organic/inorganic composite proton conductors are developed to overcome the breakdown of the actual state-of-the-art membranes (i.e. PFSA membranes: Nafion® (DuPont), Dow, Flemion® (Asahi Glass Corporation) and Aciplex® (Asahi Chemicals)). Thus, increasing the operating temperature above 100°C, reduced methanol permeability (methanol crossover), increasing the water retention and also increasing the mechanical and thermal stability of the composite membranes.

The method of inclusion of inorganic proton conductor or inorganic particle has involved a bulk powder dispersed in a polymer solution, leading specifically to particles of highly dispersed inorganic fillers of particle size in the sub-micronic range. These methods make use of mild chemistry technique, including intercalation/exfoliation, sol-gel chemistry, and

decreases in affinity with water; loss of mechanical strength through a softening of the polymer backbone; and parasitic losses (the high level of gas permeation). There are several

i. The operation of PEMFC at temperature above 140°C is receiving world-wide attention because fuel selection remains straightforward, and a number of fuels, including reformed hydrogen with high CO content and light hydrocarbons (alcohol, natural gas, propane, etc.) are still being considered for PEMFC application. Accordingly, cell temperature operation at temperatures above 140°C is of great interest because, in this temperature range, anode catalyst poisoning by CO is less important and the kinetics of fuel oxidation will be improved and the efficiency of the cell significantly enhanced. High temperature cell operation will contribute to reducing the complexity of the hydrocarbon fuel cell system. Some other advantages of operating PEMFC at high temperature are: a reduction in the use of expensive catalysts; and minimization of the problems related to electrode flooding. Light hydrocarbons may be potential energy vectors for PEMFC, which may lead to the development of suitable membranes that are

reasons for the development of high temperature membranes (Savadogo, 2004):

stable in high temperature operating conditions and prevent fuel crossover.

of view.

ii. Enhancement of gas transport in the electrode layers is also expected because no liquid water will be present in the cell at these temperatures. Membrane proton conductivity should be dependent on water content at these temperatures; consequently, it is not necessary to humidify the gas before it enters the stack. This may help improve the kinetics of mass transport and simplify the fuel cell system. In particular, the kinetics of oxygen reduction reaction could be improved, by at least three orders of magnitude, if we increase the operating temperature from 25 to 130°C. PFSA membranes cannot be used in PEMFC operating above temperatures around 100°C, because at these temperatures they will lose their mechanical properties and their swelling properties will be lowered. They do not perform well above 90°C in a hydrocarbon PEMFC and above 85°C in hydrogen PEMFC. The boiling point of water can be raised by increasing the operating pressure above 3 bar, which may correspond to a boiling point of water of about 135°C. But raising the pressure of PEMFC is undesirable from an efficiency point

One of the main drawbacks of DMFC (direct methanol fuel cell) is the slow methanol oxidation kinetics. An increase in the operating temperature of the DMFC from 90 to about 140°C is highly desirable. Also operation at high temperature will enhanced CO tolerant when a reformate hydrogen is used in H2-PEMFC. One approach to achieve water retention at high temperature is to fabricate a composite membrane constituted of organic proton conductor and inorganic materials. The organic/inorganic composite proton conductors are developed to overcome the breakdown of the actual state-of-the-art membranes (i.e. PFSA membranes: Nafion® (DuPont), Dow, Flemion® (Asahi Glass Corporation) and Aciplex® (Asahi Chemicals)). Thus, increasing the operating temperature above 100°C, reduced methanol permeability (methanol crossover), increasing the water retention and also

The method of inclusion of inorganic proton conductor or inorganic particle has involved a bulk powder dispersed in a polymer solution, leading specifically to particles of highly dispersed inorganic fillers of particle size in the sub-micronic range. These methods make use of mild chemistry technique, including intercalation/exfoliation, sol-gel chemistry, and

increasing the mechanical and thermal stability of the composite membranes.

ion-exchange (Bonnet *et al.,* 2000; Jones & Rozière, 2001). Such approach generally avoid any sedimentation of the inorganic component, intimacy of contact between the inorganic and organic components at the molecular level assures the greatest possible interface and, at such small particle size, the mechanical properties can be improved compared with those of a polymer-only membrane (Jones & Rozière, 2001, 2003). In addition, since in many proton conductors of conductivity suitable for electrochemical applications the proton transfer process takes place on the surface of the particles, increase in surface area (small particle size) will increase the conductivity (Jones & Rozière, 2001).

This concept was suggested by Watanabe *et al.* (1995, 1996) and is based on the development of self-humidifying composite membranes. The membranes are fabricated from the dispersion of nano-particles of Pt in a thin Nafion® film (≈ 50 µm). Membranes fabricated based on this concept should not require external humidification and should suppress the crossover of H2 and O2. The dispersed particles should catalyze the oxidation and the reduction of the crossover H2 and O2 respectively. Water from this reaction is directly used to humidify the membrane. This is supposed to result in a more stable operation of the cell at 80°C without any external humidification of the membrane (Watanabe *et al.*, 1998).

#### **6.1 Organic/silica nanocomposite membranes**

Silica as an additive to Nafion® was widely studies. Both recast Nafion® (Adjemian *et al.*, 2002a,2002b; Antonucci *et al.*, 1999; Arimura *et al.*, 1999; Dimitrova *et al.*, 2002a,2002b) and Nafion® film (e.g. Nafion® 117) (Adjemian *et al.*, 2002b; Baradie *et al.*, 2000; Jung *et al.*, 2002) are used in the fabrication of the composite membrane. Organic–silica composite membranes have been prepared according to several methods by casting mixtures such as: using silicon dioxide particles (e.g. Aerosil A380 from Degussa) (Antonucci *et al.*, 1999; Aricò *et al.*, 1998; Arimura *et al.*, 1999; Dimitrova *et al.*, 2002a,2002b), diphenylsilicate (DPS) (Liang *et al.*, 2006), the other one is to introduce silica oxide incorporated via *in situ* sol-gel reaction of tetraethoxysilane (TEOS) (Adjemian *et al.*, 2002b; Baradie *et al.*, 2000; Deng *et al.*, 1998; R.C. R.C. Jiang *et al.*, 2006b; Jung *et al.*, 2002; Mauritz *et al.*, 1995, 1998). Nafion®–silica membranes shows good performance at T > 100°C due to low levels of dehydration. Nafion®–silica membranes were prepared by mixing Nafion® ionomer (5 wt.%) with 3 wt.% SiO2 followed by a regular membrane casting procedure. In the final stage, the membranes were heat treated at 160°C for 10 min to achieve both a high crystallinity and high mechanical stability (Antonucci *et al.*, 1999). A DMFC utilizing these membranes was tested under galvanostatic conditions at 500 mA cm-2. The voltage initially decreased from 0.42 to 0.36 V but then remained stable for 8 h. The performance decrease is due to adsorption of poisoning species, which appears to be a reversible process at 145°C (removed by short circuit discharging in the presence of water). These membranes demonstrated higher performance with increasing temperature. A nano-particles possessed a core-shell structure consisting of silica core (< 10 nm) and a densely grafted oligometric ionomer layer was incorporated into Nafion® matrix to form a composite membrane. The polyelectrolytegrafted silica particles [P(SPA)-SiO2] was dispersed in Nafion® solution and a composite membrane was formed by recasting process. The proton conductivity of Nafion® membrane containing P(SPA)-SiO2 (4 wt.%) is significantly higher than that of unmodified recast Nafion® in the range 25 to 80°C. The composite membrane offers superior cell performance over unmodified recast Nafion® at both operating temperature of 50 and 80°C. At 50°C, the

Organic / Inorganic Nanocomposite Membranes

high current density than did unmodified Nafion®.

**6.2 Organic/heteropolyacid (HPA) nanocomposite membranes** 

membrane.

Development for Low Temperature Fuel Cell Applications 521

casting method. The room temperature proton conductivity of full hydration composite membrane was increased from 0.10 to 0.12 S cm-1 as the M-SiO2-SO3H content increased from 0 to 3 wt.%. Methanol permeability decreases with increasing the content of M-SiO2- SO3H, where methanol permeability was 4.5 × 10-6 cm2 s-1, which was 30% lower than unmodified Nafion®. The current densities measured with composite membranes containing 0, 1, 3 and 5 wt.% M-SiO2-SO3H, were 51, 66, 80 and 70 mA cm-2, respectively, at a potential of 0.2 V. Moreover, all composite membranes containing M-SiO2-SO3H performed better at

Nanocomposite proton exchange membranes were prepared from sulfonated poly(phtalazinone ether ketone) (SPPEK) and various amounts of sulfonated silica nanoparticles (silica-SO3H) (Su *et al.*, 2007). The use of silica-SO3H compensates for the decrease in ion exchange capacity of membranes observed when no-sulfonated nano-fillers are utilized. The strong –SO3H/-SO3H interaction between SPPEK chains and silica-SO3H particles leads to ionic cross-linking in the membrane structure, which increases both the thermal stability and methanol resistance of the membranes. The membrane with 7.5 phr of silica-SO3H (phr = g of silica-SO3H / 100 g of SPPEK in membranes) exhibits low methanol crossover, high bound-water content, and a proton conductivity of 3.6 fold increase to that of the unmodified SPPEK membrane. Nafion®/hydrated phosphor-silicate composite membrane was synthesis by Tung and Hwang (2007). The phosphor-silicate glass, with a nominal composition of 30% P2O5 and 70% SiO2 (molar ratio) (called 30P70Si), was prepared by the accelerated sol-gel process, where tetraethylorthosolicate and trimethyl phosphate are used as precursors. It was found that the methanol permeability decreases dramatically with increased SiO2-P2O5 content and the proton conductivity only decreases slightly , as a consequence the selectivity of the hybrid membranes are higher than unmodified Nafion®

Perfluorosulfonic acid based organic/ inorganic composite membranes with different heteropolyacid (HPA) additives have been investigated as alternate materials for low humidity PEMFC operation (Giordano *et al.*, 1996; Ramani *et al.*, 2004, 2005a,2005b; Tazi & Savadogo, 2000, 2001). Two major factors limiting the performance of Nafion®/HPA composite membranes are (Ramani *et al.*, 2005b): (i) the high solubility of the HPA additive and (ii) the large particle size of the inorganic additive within the membrane matrix (Ramani *et al.*, 2004, 2005a). Stabilization technique have been developed (Ramani *et al.*, 2005a) to limit the solubility of the HPA additive. Recast Nafion® with phosphotungstic acid (PTA) as HPA fillers were prepared by Ramani *et al.* (2005b). Three types of fillers were used PTA with 30-50 nm particle size, PTA with 1-2 µm particle size and TiO2 with 1-2 µm particle size. The composite membranes had hydrogen crossover currents on the order of 1-5 mA cm-2, with the crossover flux decreasing and approaching the value for recast Nafion® as the particle size was reduced. A 25 µm thick composite membrane with PTA (1-2 µm particle size) had an area-specific resistance of 0.22 Ω cm-2 at 120°C and 35% RH, while the corresponding value for a 25 µm thick composite membrane with PTA (30-50 nm) was 0.16 Ω cm-2. The latter membrane compared favorably with recast Nafion®, which had an areaspecific resistance of 0.19 Ω cm-2 under the same conditions. Savadogo and co-workers (Savadogo, 2004; Tazi & Savadogo, 2000, 2001; Tian & Savadogo, 2005) prepared composite

maximum power output of the composite membrane is about 1.8 times greater than of the Nafion® membrane and at 80°C, the ratio becomes 1.5 (Tay *et al.*, 2008).

Adjemian *et al.* (2002b) prepared composite membranes by either impregnating an extruded film via sol-gel processing of tetraethoxysilane (TEOS), or by preparing a recast film, using solubilized PFSA and silicon oxide polymer/gel. TEOS when reacted with water in an acidic medium undergoes polymerization to form a mixture of silicas (SiOx) and siloxane polymer with terminal hydroxide and ethoxide groups(SiOx/-OH/-OEt). When PFSAs are used as the acidic medium, the SiOx/siloxane polymer forms within the membrane. Composite membranes were tested in fuel cell operating with pre-humidified reactant gases at temperature of 130°C and a pressure of 3 atm. The PFSA/silicon oxide composite membranes shows resistivities 50% lower than their respective unmodified PFSA under the same operating conditions. The observed resistivity trend from best to worst is as follows: Aciplex® 1004/silicon oxide > Nafion® 112/silicon oxide > Nafion® 105/silicon oxide > Aciplex® recast/silicon oxide > Nafion® recast/silicon oxide > Nafion® 115/silicon oxide.

Recently a new approach to make composite membrane was introduced, where a functionalized silica is used as a filler to make the composite membranes (Li *et al.*, 2006; Y.F. Lin *et al.*, 2007; Sambandam & Ramani, 2007 ; Su *et al.*, 2007; Tung & Hwang, 2007). Sol-gel derived sulfonated diphenyldimethoxysilane (SDDS) with hydrophilic –SO3H functional groups were used as the additive to reduce the methanol permeability of Nafion® (Li *et al.*, 2006). The Nafion®-SDDS nanocomposite membranes were prepared by mixing Nafion® – dimethyl formamide (DMF) solutions with SDDS sol and casting to membranes. The proton conductivity of composite membrane decreased compared with commercial Nafion® membranes. This is partly because (i) the relative low conductivity of organosilica, (ii) the slightly tortuous path through the membrane which is caused by the embedding of the organosilica into the hydrophilic clusters, (iii) and the hydrophobic phenyl groups of the organosilica which change the distribution of the hydrophilic/hydrophobic phases therefore reduce the water content of the membrane. The proton conductivity decreases with the increase of the fillers. On the other hand, the methanol permeability is reduced with the increase of the SDDS content. The methanol permeability drops by a factor of 0.41, 0.61, 0.67 and 0.71 times for nanocomposite with loading of 5, 10, 20 and 25 wt.%, respectively as compared to bare recast Nafion® (Li *et al.*, 2006).

Sulfonic acid functionalized silica was synthesized by condensation of MPTMS (3 mercaptopropyltrimethoxy silane) precursor through a sol-gel approach. Sulfonated poly(ether ether ketone) composite with sulfonic acid functionalized silica were prepared by casting (Sambandam & Ramani, 2007 ). At 80°C and 75% RH (relative humidity) the measured conductivity was 0.05 S cm-1 for SPEEK containing 10% sulfonic acid functionalized silica and 0.02 S cm-1 for the plain SPEEK membrane. At 80°C and 50% RH the measured conductivity was 0.018 S cm-1 for SPEEK containing 10% sulfonic acid functionalized silica and 0.004 S cm-1 for the plain SPEEK membrane.

L64 copolymer-templated mesoporous SiO2, functionalized with perfluoroalkylsulfonic acid was prepared (Y.F. Lin *et al.*, 2007). A condensation reaction between the surface silanol groups of the mesoporous silicas and 1,2,2-trifluoro-2-hydroxy-1-trifluoromethylethane sulfonic acid Beta-sultone was conducted. Nafion®/functionalized mesoporous silica composite membranes were prepared via homogeneous dispersive mixing and the solvent

maximum power output of the composite membrane is about 1.8 times greater than of the

Adjemian *et al.* (2002b) prepared composite membranes by either impregnating an extruded film via sol-gel processing of tetraethoxysilane (TEOS), or by preparing a recast film, using solubilized PFSA and silicon oxide polymer/gel. TEOS when reacted with water in an acidic medium undergoes polymerization to form a mixture of silicas (SiOx) and siloxane polymer with terminal hydroxide and ethoxide groups(SiOx/-OH/-OEt). When PFSAs are used as the acidic medium, the SiOx/siloxane polymer forms within the membrane. Composite membranes were tested in fuel cell operating with pre-humidified reactant gases at temperature of 130°C and a pressure of 3 atm. The PFSA/silicon oxide composite membranes shows resistivities 50% lower than their respective unmodified PFSA under the same operating conditions. The observed resistivity trend from best to worst is as follows: Aciplex® 1004/silicon oxide > Nafion® 112/silicon oxide > Nafion® 105/silicon oxide > Aciplex® recast/silicon oxide > Nafion® recast/silicon oxide > Nafion® 115/silicon oxide.

Recently a new approach to make composite membrane was introduced, where a functionalized silica is used as a filler to make the composite membranes (Li *et al.*, 2006; Y.F. Lin *et al.*, 2007; Sambandam & Ramani, 2007 ; Su *et al.*, 2007; Tung & Hwang, 2007). Sol-gel derived sulfonated diphenyldimethoxysilane (SDDS) with hydrophilic –SO3H functional groups were used as the additive to reduce the methanol permeability of Nafion® (Li *et al.*, 2006). The Nafion®-SDDS nanocomposite membranes were prepared by mixing Nafion® – dimethyl formamide (DMF) solutions with SDDS sol and casting to membranes. The proton conductivity of composite membrane decreased compared with commercial Nafion® membranes. This is partly because (i) the relative low conductivity of organosilica, (ii) the slightly tortuous path through the membrane which is caused by the embedding of the organosilica into the hydrophilic clusters, (iii) and the hydrophobic phenyl groups of the organosilica which change the distribution of the hydrophilic/hydrophobic phases therefore reduce the water content of the membrane. The proton conductivity decreases with the increase of the fillers. On the other hand, the methanol permeability is reduced with the increase of the SDDS content. The methanol permeability drops by a factor of 0.41, 0.61, 0.67 and 0.71 times for nanocomposite with loading of 5, 10, 20 and 25 wt.%, respectively as

Sulfonic acid functionalized silica was synthesized by condensation of MPTMS (3 mercaptopropyltrimethoxy silane) precursor through a sol-gel approach. Sulfonated poly(ether ether ketone) composite with sulfonic acid functionalized silica were prepared by casting (Sambandam & Ramani, 2007 ). At 80°C and 75% RH (relative humidity) the measured conductivity was 0.05 S cm-1 for SPEEK containing 10% sulfonic acid functionalized silica and 0.02 S cm-1 for the plain SPEEK membrane. At 80°C and 50% RH the measured conductivity was 0.018 S cm-1 for SPEEK containing 10% sulfonic acid

L64 copolymer-templated mesoporous SiO2, functionalized with perfluoroalkylsulfonic acid was prepared (Y.F. Lin *et al.*, 2007). A condensation reaction between the surface silanol groups of the mesoporous silicas and 1,2,2-trifluoro-2-hydroxy-1-trifluoromethylethane sulfonic acid Beta-sultone was conducted. Nafion®/functionalized mesoporous silica composite membranes were prepared via homogeneous dispersive mixing and the solvent

Nafion® membrane and at 80°C, the ratio becomes 1.5 (Tay *et al.*, 2008).

compared to bare recast Nafion® (Li *et al.*, 2006).

functionalized silica and 0.004 S cm-1 for the plain SPEEK membrane.

casting method. The room temperature proton conductivity of full hydration composite membrane was increased from 0.10 to 0.12 S cm-1 as the M-SiO2-SO3H content increased from 0 to 3 wt.%. Methanol permeability decreases with increasing the content of M-SiO2- SO3H, where methanol permeability was 4.5 × 10-6 cm2 s-1, which was 30% lower than unmodified Nafion®. The current densities measured with composite membranes containing 0, 1, 3 and 5 wt.% M-SiO2-SO3H, were 51, 66, 80 and 70 mA cm-2, respectively, at a potential of 0.2 V. Moreover, all composite membranes containing M-SiO2-SO3H performed better at high current density than did unmodified Nafion®.

Nanocomposite proton exchange membranes were prepared from sulfonated poly(phtalazinone ether ketone) (SPPEK) and various amounts of sulfonated silica nanoparticles (silica-SO3H) (Su *et al.*, 2007). The use of silica-SO3H compensates for the decrease in ion exchange capacity of membranes observed when no-sulfonated nano-fillers are utilized. The strong –SO3H/-SO3H interaction between SPPEK chains and silica-SO3H particles leads to ionic cross-linking in the membrane structure, which increases both the thermal stability and methanol resistance of the membranes. The membrane with 7.5 phr of silica-SO3H (phr = g of silica-SO3H / 100 g of SPPEK in membranes) exhibits low methanol crossover, high bound-water content, and a proton conductivity of 3.6 fold increase to that of the unmodified SPPEK membrane. Nafion®/hydrated phosphor-silicate composite membrane was synthesis by Tung and Hwang (2007). The phosphor-silicate glass, with a nominal composition of 30% P2O5 and 70% SiO2 (molar ratio) (called 30P70Si), was prepared by the accelerated sol-gel process, where tetraethylorthosolicate and trimethyl phosphate are used as precursors. It was found that the methanol permeability decreases dramatically with increased SiO2-P2O5 content and the proton conductivity only decreases slightly , as a consequence the selectivity of the hybrid membranes are higher than unmodified Nafion® membrane.

#### **6.2 Organic/heteropolyacid (HPA) nanocomposite membranes**

Perfluorosulfonic acid based organic/ inorganic composite membranes with different heteropolyacid (HPA) additives have been investigated as alternate materials for low humidity PEMFC operation (Giordano *et al.*, 1996; Ramani *et al.*, 2004, 2005a,2005b; Tazi & Savadogo, 2000, 2001). Two major factors limiting the performance of Nafion®/HPA composite membranes are (Ramani *et al.*, 2005b): (i) the high solubility of the HPA additive and (ii) the large particle size of the inorganic additive within the membrane matrix (Ramani *et al.*, 2004, 2005a). Stabilization technique have been developed (Ramani *et al.*, 2005a) to limit the solubility of the HPA additive. Recast Nafion® with phosphotungstic acid (PTA) as HPA fillers were prepared by Ramani *et al.* (2005b). Three types of fillers were used PTA with 30-50 nm particle size, PTA with 1-2 µm particle size and TiO2 with 1-2 µm particle size. The composite membranes had hydrogen crossover currents on the order of 1-5 mA cm-2, with the crossover flux decreasing and approaching the value for recast Nafion® as the particle size was reduced. A 25 µm thick composite membrane with PTA (1-2 µm particle size) had an area-specific resistance of 0.22 Ω cm-2 at 120°C and 35% RH, while the corresponding value for a 25 µm thick composite membrane with PTA (30-50 nm) was 0.16 Ω cm-2. The latter membrane compared favorably with recast Nafion®, which had an areaspecific resistance of 0.19 Ω cm-2 under the same conditions. Savadogo and co-workers (Savadogo, 2004; Tazi & Savadogo, 2000, 2001; Tian & Savadogo, 2005) prepared composite

Organic / Inorganic Nanocomposite Membranes

**6.3 Organic/TiO2 nanocomposite membranes** 

2005).

Development for Low Temperature Fuel Cell Applications 523

employed as an electrolyte in H2/O2 PEMFC, the higher current density values (540 and 320 mA cm-2 at 0.4 V, respectively) were obtained than that of the Nafion®115 membrane (95 mA cm-2), under the operating conditions of 110°C and 70% RH. A similar membrane was prepared by Aricò *et al.* (2003a,2003b). Sulfonic-functionalized heteropolyacid-SiO2 nanoparticles were synthesized by grafting and oxidizing of a thiol-silane compound onto the heteropolyacid-SiO2 nanoparticle surface (H.J. Kim *et al.,* 2006). The composite membrane containing the sulfonic-functionalized heteropolyacid-SiO2 nanoparticles was prepared by blending with Nafion® ionomer. TG-DTA analysis showed that the composite membrane was thermally stable up to 290°C. The DMFC performance of the composite membrane increased the operating temperature from 80 to 200°C. The function of the sulfonic-functionalized heteropolyacid-SiO2 nanoparticles was to provide a proton carrier and act as a water reservoir in the composite membrane at elevated temperature. The power density was 33 mW cm-2 at 80°C, 39 mW cm-2 at 160°C, 44 mW cm-2 at 200°C, respectively. SPEEK-silica membranes doped with phosphotungstic acid (PWA) was synthesized by Colicchio *et al.* (2009). The silica is generated *insitu* via the water free sol-gel process of polyethoxysiloxane (PEOS), a liquid hyperbranched inorganic polymer of low viscosity. PEOS was used as silica precursor instead of the corresponding monomeric tetraethoxysilane (TEOS). At 100°C and 90% RH the membrane prepared with PEOS (silica content = 20 wt.%) shows two times higher conductivity than the pure SPEEK. The addition of small amount of PWA (2 wt.% of the total solid content) introduce in the early stage of membrane preparation brings to a further increase in conductivity (more than three times the pure SPEEK). Different classes of composite membranes containing HPA and silica were developed, namely, phosphomolybdic acid (PMA)/phosphotungstic acid (PWA)- P2O5-SiO2 glass electrolyte (Uma & Nogami, 2007), poly(vinyl alcohol) (PVA)/sulfosuccinic acid (SSA)/silica hybrid membrane (D.S. Kim *et al.,* 2004), PVA/SiO2/ SiW (silicotungstic acid) (Shanmugam *et al.,* 2006), PVA/PWA/SiO2 (W. Xu *et al.,* 2004), polyethylene oxide (PEO)/PWA/SiO2 (Honma *et al.,* 2002), polyethylene glycol (PEG)/4-dodecylbenzene sulfonic acid (DBSA)/SiO2 (H.Y. Chang *et al.,* 2003), PWA-doped PEG/SiO2 (C.W. Lin *et al.,* 

Nanosized titanium oxide was synthesized by sol-gel hydrolyzing an alcoholic solution of Ti(OiPr)4 by Baglio *et al.* (2005). Thermal treatments at different temperature, namely 500, 650 and 800°C, were performed to tailor the oxide powder properties. The crystallite size for the three sample was found to be 12, 22 and 39 nm, respectively. A composite membrane Nafion®/ 5 wt.% TiO2 was prepared by using the recast procedure. The composite membrane thickness was about 100 µm. A maximum power density of 350 mW cm-2 was obtained at 145°C with the cell equipped with the composite membrane containing TiO2 calcined at 500°C. Sacca *et al.* (2005) synthesized TiO2 powder by the sol-gel method starting with a Ti(OiPr)4 and calcined a 400°C. This powder was made of spherical particles with a grain size between 5 and 20 nm. A recast Nafion® with 3 wt.% TiO2 composite membrane was prepared, the thickness of the membrane was 100 µm. The proton conductivity of different membranes were measured at two different values of relative humidity (RH), 100 and 85% RH, respectively, and simulating the cell operating conditions in the temperature range from 80 to 130°C. Nafion® recast (70 µm thickness) has the lower conductivity ranging

membranes constituted of recast Nafion®and mixed with appropriate concentration of HPA, namely, silicotungstic acid (STA), phosphotungstic acid (PTA) and phosphomolybdic acid (PMA). It was shown that the water uptake of the various membranes increases in this order: Nafion®117 (27%) < Nafion®/STA (60%) < Nafion®/PTA (70%) < Nafion®/PMA (95%). The ionic conductivity increases in the order Nafion®117 (1.3 × 10-2 S cm-1)) < Nafion®/PMA (1.5 × 10-2 S cm-1) < Nafion®/PTA (2.5 × 10-2 S cm-1) < Nafion®/STA (9.5 × 10- 2 S cm-1). The tensile strength of the membranes decreases in the order: Nafion®117 (15000 Pa) < Nafion®/STA (14000 Pa) < Nafion®/PMA (8000 Pa) < Nafion®/PTA (3000 Pa), while their deformation (εmax) changes in the order : Nafion®/STA (45%) < Nafion®/PMA (70%) < Nafion®/PTA (170%) < Nafion®117 (384%). The current density at 0.600 V of the PEMFCs based on the various membranes varies in the order: Nafion®117 (640 mA cm-2) < Nafion®/STA (695 mA cm-2) < Nafion®/PTA (810 mA cm-2) < Nafion®/PMA (940 mA cm-2).

Tazi and Savadogo (2000) fabricated Nafion® membranes containing silicotungstic acid and thiophene. They reported an increase of up to 60% of water uptake and a considerable improvement in the fuel cell current density, when compared to the plain Nafion® membrane. Dimitrova *et al.* (2002a) prepared a recast Nafion®-based composite membrane containing molybdophosphoric acid. This composite membrane exhibit significantly higher conductivity in comparison to Nafion® 117 and pure recast Nafion®. An enhancement of a factor of 3 in the conductivity at 90°C was observed. Zaidi *et al.* (2000) prepared a series of composite membranes using SPEEK as polymer matrix and tungstophosphoric acid (TPA), its sodium salt (Na-TPA) and molybdophosphoric acid (MoPA) as inorganic fillers. The conductivity of the composite membranes exceeded 10-2 S/cm at room temperature and reached values of about 10-1 S/cm above 100°C. From the DSC (Differential Scanning Calorimeter) studies, it was indicated that the glass transition temperature of the SPEEK/HPA composite membrane increases due to the incorporation of solid HPA into SPEEK membrane. This increase in the glass transition temperature was attributed to an intermolecular interaction between SPEEK and HPA. Staiti *et al.* (2001) prepared Nafion®(recast)-silica composite membranes doped with phosphotungstic (PWA) and silicotungstic (SiWA) acids for application in direct methanol fuel cell at high temperature (145°C). The phosphotungstic acid-based membrane showed better electrochemical characteristics at high current densities with respect to both silicotungstic acid-modified membrane and silica- Nafion® membrane. The best electrochemical performance is obtained with the PWA-based membrane, which gives a maximum power density of 400 mW cm-2 at current density of about 1.4 A cm-2 under oxygen feed operation at 145°C. Maximum power density of 340 mW cm-2 is obtained from the fuel cell which uses the silica-modified membrane, whereas a lower performance was achieved with the SiWA-based membrane. The maximum power density obtained in air with the PWA-based membrane is 250 mW cm-2 at 145°C, and 210 mW cm-2 with the Nafion-SiO2 membrane at the same temperature.

Shao *et al.* (2004) prepared Nafion®/silicon oxide (SiO2)/phosphotungstic acid (PWA) and Nafion®/silicon oxide composite membranes for H2/O2 proton exchange membrane fuel cells operated above 100°C. It was found that the composite membranes showed a higher water uptake compared with the Nafion® recast membrane. The proton conductivity of the composite membranes appeared to be similar to that of the native Nafion® membrane at high temperatures and 100% relative humidity (RH), however, it was much higher at low RH. When the composite membranes viz. Nafion®/SiO2/PWA and Nafion®/SiO2 were

membranes constituted of recast Nafion®and mixed with appropriate concentration of HPA, namely, silicotungstic acid (STA), phosphotungstic acid (PTA) and phosphomolybdic acid (PMA). It was shown that the water uptake of the various membranes increases in this order: Nafion®117 (27%) < Nafion®/STA (60%) < Nafion®/PTA (70%) < Nafion®/PMA (95%). The ionic conductivity increases in the order Nafion®117 (1.3 × 10-2 S cm-1)) < Nafion®/PMA (1.5 × 10-2 S cm-1) < Nafion®/PTA (2.5 × 10-2 S cm-1) < Nafion®/STA (9.5 × 10- 2 S cm-1). The tensile strength of the membranes decreases in the order: Nafion®117 (15000 Pa) < Nafion®/STA (14000 Pa) < Nafion®/PMA (8000 Pa) < Nafion®/PTA (3000 Pa), while their deformation (εmax) changes in the order : Nafion®/STA (45%) < Nafion®/PMA (70%) < Nafion®/PTA (170%) < Nafion®117 (384%). The current density at 0.600 V of the PEMFCs based on the various membranes varies in the order: Nafion®117 (640 mA cm-2) < Nafion®/STA (695 mA cm-2) < Nafion®/PTA (810 mA cm-2) < Nafion®/PMA (940 mA cm-2). Tazi and Savadogo (2000) fabricated Nafion® membranes containing silicotungstic acid and thiophene. They reported an increase of up to 60% of water uptake and a considerable improvement in the fuel cell current density, when compared to the plain Nafion® membrane. Dimitrova *et al.* (2002a) prepared a recast Nafion®-based composite membrane containing molybdophosphoric acid. This composite membrane exhibit significantly higher conductivity in comparison to Nafion® 117 and pure recast Nafion®. An enhancement of a factor of 3 in the conductivity at 90°C was observed. Zaidi *et al.* (2000) prepared a series of composite membranes using SPEEK as polymer matrix and tungstophosphoric acid (TPA), its sodium salt (Na-TPA) and molybdophosphoric acid (MoPA) as inorganic fillers. The conductivity of the composite membranes exceeded 10-2 S/cm at room temperature and reached values of about 10-1 S/cm above 100°C. From the DSC (Differential Scanning Calorimeter) studies, it was indicated that the glass transition temperature of the SPEEK/HPA composite membrane increases due to the incorporation of solid HPA into SPEEK membrane. This increase in the glass transition temperature was attributed to an intermolecular interaction between SPEEK and HPA. Staiti *et al.* (2001) prepared Nafion®(recast)-silica composite membranes doped with phosphotungstic (PWA) and silicotungstic (SiWA) acids for application in direct methanol fuel cell at high temperature (145°C). The phosphotungstic acid-based membrane showed better electrochemical characteristics at high current densities with respect to both silicotungstic acid-modified membrane and silica- Nafion® membrane. The best electrochemical performance is obtained with the PWA-based membrane, which gives a maximum power density of 400 mW cm-2 at current density of about 1.4 A cm-2 under oxygen feed operation at 145°C. Maximum power density of 340 mW cm-2 is obtained from the fuel cell which uses the silica-modified membrane, whereas a lower performance was achieved with the SiWA-based membrane. The maximum power density obtained in air with the PWA-based membrane is 250 mW cm-2 at 145°C, and 210 mW cm-2 with the Nafion-SiO2 membrane at the same temperature. Shao *et al.* (2004) prepared Nafion®/silicon oxide (SiO2)/phosphotungstic acid (PWA) and Nafion®/silicon oxide composite membranes for H2/O2 proton exchange membrane fuel cells operated above 100°C. It was found that the composite membranes showed a higher water uptake compared with the Nafion® recast membrane. The proton conductivity of the composite membranes appeared to be similar to that of the native Nafion® membrane at high temperatures and 100% relative humidity (RH), however, it was much higher at low RH. When the composite membranes viz. Nafion®/SiO2/PWA and Nafion®/SiO2 were employed as an electrolyte in H2/O2 PEMFC, the higher current density values (540 and 320 mA cm-2 at 0.4 V, respectively) were obtained than that of the Nafion®115 membrane (95 mA cm-2), under the operating conditions of 110°C and 70% RH. A similar membrane was prepared by Aricò *et al.* (2003a,2003b). Sulfonic-functionalized heteropolyacid-SiO2 nanoparticles were synthesized by grafting and oxidizing of a thiol-silane compound onto the heteropolyacid-SiO2 nanoparticle surface (H.J. Kim *et al.,* 2006). The composite membrane containing the sulfonic-functionalized heteropolyacid-SiO2 nanoparticles was prepared by blending with Nafion® ionomer. TG-DTA analysis showed that the composite membrane was thermally stable up to 290°C. The DMFC performance of the composite membrane increased the operating temperature from 80 to 200°C. The function of the sulfonic-functionalized heteropolyacid-SiO2 nanoparticles was to provide a proton carrier and act as a water reservoir in the composite membrane at elevated temperature. The power density was 33 mW cm-2 at 80°C, 39 mW cm-2 at 160°C, 44 mW cm-2 at 200°C, respectively.

SPEEK-silica membranes doped with phosphotungstic acid (PWA) was synthesized by Colicchio *et al.* (2009). The silica is generated *insitu* via the water free sol-gel process of polyethoxysiloxane (PEOS), a liquid hyperbranched inorganic polymer of low viscosity. PEOS was used as silica precursor instead of the corresponding monomeric tetraethoxysilane (TEOS). At 100°C and 90% RH the membrane prepared with PEOS (silica content = 20 wt.%) shows two times higher conductivity than the pure SPEEK. The addition of small amount of PWA (2 wt.% of the total solid content) introduce in the early stage of membrane preparation brings to a further increase in conductivity (more than three times the pure SPEEK). Different classes of composite membranes containing HPA and silica were developed, namely, phosphomolybdic acid (PMA)/phosphotungstic acid (PWA)- P2O5-SiO2 glass electrolyte (Uma & Nogami, 2007), poly(vinyl alcohol) (PVA)/sulfosuccinic acid (SSA)/silica hybrid membrane (D.S. Kim *et al.,* 2004), PVA/SiO2/ SiW (silicotungstic acid) (Shanmugam *et al.,* 2006), PVA/PWA/SiO2 (W. Xu *et al.,* 2004), polyethylene oxide (PEO)/PWA/SiO2 (Honma *et al.,* 2002), polyethylene glycol (PEG)/4-dodecylbenzene sulfonic acid (DBSA)/SiO2 (H.Y. Chang *et al.,* 2003), PWA-doped PEG/SiO2 (C.W. Lin *et al.,*  2005).

#### **6.3 Organic/TiO2 nanocomposite membranes**

Nanosized titanium oxide was synthesized by sol-gel hydrolyzing an alcoholic solution of Ti(OiPr)4 by Baglio *et al.* (2005). Thermal treatments at different temperature, namely 500, 650 and 800°C, were performed to tailor the oxide powder properties. The crystallite size for the three sample was found to be 12, 22 and 39 nm, respectively. A composite membrane Nafion®/ 5 wt.% TiO2 was prepared by using the recast procedure. The composite membrane thickness was about 100 µm. A maximum power density of 350 mW cm-2 was obtained at 145°C with the cell equipped with the composite membrane containing TiO2 calcined at 500°C. Sacca *et al.* (2005) synthesized TiO2 powder by the sol-gel method starting with a Ti(OiPr)4 and calcined a 400°C. This powder was made of spherical particles with a grain size between 5 and 20 nm. A recast Nafion® with 3 wt.% TiO2 composite membrane was prepared, the thickness of the membrane was 100 µm. The proton conductivity of different membranes were measured at two different values of relative humidity (RH), 100 and 85% RH, respectively, and simulating the cell operating conditions in the temperature range from 80 to 130°C. Nafion® recast (70 µm thickness) has the lower conductivity ranging

Organic / Inorganic Nanocomposite Membranes

Development for Low Temperature Fuel Cell Applications 525

based on SPEK and SPEEK for application in direct methanol fuel cell were synthesized by Nunes and co-workers (Nunes *et al.,* 2002). The inorganic fillers were introduced via *in situ* generation of SiO2, TiO2 or ZrO2. The modification with ZrO2 led to a 60-fold reduction of

Recast Nafion® composite membranes containing three different percentages (5%, 10% and 20%, w/w) of commercial ZrO2 as an inorganic filler were tested in fuel cell in a temperature range of 80-130°C in humidified H2/air gases at 3.0 bar abs by Sacca *et al.*  (2006). The introduction of 5 wt.% ZrO2 in Nafion® produces no evidence changes in the cell performance, while a better performance with 10 wt.% ZrO2 in Nafion® was obtained with a power density greater than 600 mW cm-2 at 0.6 V both at 80°C and 110°C. The good performance of 10 wt.% ZrO2 in Nafion® was maintained at 130°C with gas humidification of 85% RH, with a maximum power density of about 400 mW cm-2 was obtained in the potential range of 0.5-0.6 V. Silva *et al.* (2005a,2005b) prepared SPEEK/ZrO2 composite membranes using *insitu* formation of zirconia with zirconium tetrapropylate as alkoxide and acetyl acetone as chelating agent. The water/alkoxide ratio was always maintained higher than 1 to ensure the formation of a finely dispersed inorganic phase in the polymer solution. The thickness of the prepared membranes with 0.0, 2.5, 5.0, 7.5, 10.0, 12.5 wt.% of zirconium oxide were 188, 175, 133, 146, 128, 106 µm, respectively. The proton conductivity of the composite membranes was measured at 25°C and it was found that it decreases continuously with the ZrO2 content. Pervaporation experiments at 55°C showed that the membrane permeability towards methanol decreases with the amount of ZrO2. Composite SPEEK with 5.0, 7.5, 10.0 wt.% of ZrO2 were tested in fuel cell. The performance of 12.5 wt.% ZrO2 composite was very low due to the high ohmic resistance of the corresponding MEA. The membrane 5 wt.% ZrO2 presents the best DMFC performance among all the studied MEAs. Three types of superacidic sulfated zirconia (S-ZrO2) were prepared by different methods using hydrated zirconia and sulfuric acid by Hara and Miyayama (2004). Their proton conductivities were evaluated at 20-150°C under saturated water vapor pressure. It was found that the concentration on S-ZrO2 varied largely depending on the method of preparation. The S/Zr atomic ratio changed from 0.046 for the sample prepared through a mixture of hydrated zirconia powder and sulfuric acid to 0.35 for sample prepared through a mixture of hydrated zirconia sol and sulfuric acid. A powder compact of the former S-ZrO2 showed a proton conductivity of 4 × 10-2 S cm-1 at 70°C and 8 × 10-3 S cm-1 at 150°C, whereas that of the latter S-ZrO2

the methanol flux. However, a 13-fold reduction of conductivity was also observed.

exhibited a high conductivity of 5 × 10-2 S cm-1 at 60-150°C.

S. Ren *et al.* (2006) prepared sulfated zirconia/ Nafion® 115 nanocomposite membrane by ion exchange of zirconium ions into the Nafion® followed by precipitation of sulphated ZrO2 by treatment in H2SO4. The incorporation of sulfated zirconia increases water uptake by the Nafion® membrane, and more water is absorbed than an unmodified membrane at high temperatures. The membrane proton conductivity is decreased slightly by ZrO2 impregnation. The proton conductivity of Nafion® 115 membrane was found to be 1.5 × 10-2 S cm-1 at 25°C, while that of S-ZrO2/ Nafion® 115 membrane is decreased to 5.0 × 10-3 S cm-1 at 25°C. At 110°C and above, the proton conductivity of S-ZrO2/ Nafion® 115 membrane is more than one-half that of the Nafion® 115 membrane. Fine particle superacidic sulfated zirconia (S-ZrO2) was synthesized by ameliorated method, and composite membranes with different S-ZrO2 contents were prepared by a recasting procedure from a suspension of S-

from 0.12 to 0.14 S cm-1, while the composite Nafion®/TiO2 showed highest value than Nafion® 115 (125 µm) were value in the range 0.15-0.18 S cm-1 . A power density of 0.514 W cm-2 for Nafion® / 3 wt.% TiO2 composite against 0.354 W cm-2 for Nafion® 115 at 0.56 V and at T = 110°C was recorded. At 120°C, Nafion® 115 was damaged while the composite Nafion®/TiO2 membrane continued to work up to 130°C by reaching a power density of about 0.254 W cm-2 at 0.5 V.

Hybride membranes based on highly sulfonated poly(ether ether ketone) (SPEEK, DS = 0.9) where titania network was dispersed by *insitu* sol-gel reactions were prepared by Di Vona *et al.* (2007). Titania network was introduced following two routes: route 1 using titanium tetrabutoxide (Ti(OBu)4) and pyridine, and route 2 uses Ti(OBu)4 and 2,4-pentandione. Composite membranes prepared by route 2 showed a good conductivity property that can be attributed to the structural characteristics of the inorganic network generated in the presence of a chelating agent. This membrane shows a stable value ( = 5.8 × 10-2 S cm-1) at 120°C in fully hydrated conditions. Jian-hua *et al.* (2008) prepared a composite Nafion®/TiO2 membranes by carrying out *insitu* sol-gel reaction of Ti (OC4H9)4 followed by hydrolyzationcondensation in Nafion® 112, 1135 and 115. TiO2 prepared with this method was found to be 4 nm. TiO2 contents were 1.23, 2.47 and 3.16 wt.% for Nafion® 112/TiO2, Nafion® 1135/TiO2 and Nafion® 115/TiO2, respectively. The polarization characteristics of all three MEAs with the membranes containing TiO2 were improved significantly comparing with those of pure Nafion® film. A mixture of titanium tetraisopropoxide (TTIP) and PEG 1000 were used to prepare the titania sol by Liu *et al.* (2006). The average particle size of 20 nm was reported. The formed sol was deposited on the surface of Nafion® 112 membranes by spin coating. The TiO2 film is dense and well attached to the membrane, but some cracks in the membrane coated with diluted titania sol (e.g. 0.002 mg cm-2), while the membrane coated with thick titania sol (e.g. 0.021 mg cm-2) are very dense and cracks free. The proton conductivity of nano-TiO2-coated Nafion® membranes at 25 and 80°C were recorded with different TiO2 content. It was found that the maximum conductivity was with uncoated Nafion® 112, with values of 0.027 and 0.041 S cm-1 for 25 and 80°C, respectively. The conductivity of coated Nafion® decreases with increasing titania content. On the other hand, methanol permeability of the coated membranes decreases with increasing TiO2 content, namely from 3.2 × 10-6 to 1.7 × 10-6 cm2 s-1 at 25°C, and from 12.5 × 10-6 to 4.6 × 10-6 cm2 s-1 at 85°C. Thus the rise in temperature leads to a strong increase in permeation by a factor of about 3. The methanol permeability of the unmodified Nafion® 112 membrane was found to be 3.6 × 10-6 and 13 × 10-6 cm2 s-1 at 25 and 85°C, respectively. The cell performance with titania coated membrane with a content of 0.009 mg cm-2 exhibits a higher voltage than cells with Nafion® membrane or with the other coated membranes. Nafion® 112 delivered a maximum power density of 37 mV cm-2 at a current density of 200 mA cm-2. A maximum power density of 44 mW cm-2 is obtained from a fuel cell that employs the titania-coated membrane with 0.009 mg cm-2 content.

#### **6.4 Organic/zirconia and sulfated zirconia nanocomposite membranes**

Zirconia as an inorganic filler was added to polymeric proton conductor membranes (Aricò *et al.,* 2003b, 2004; Nunes *et al.,* 2002; Sacca *et al.,* 2006; Silva *et al.,* 2005a, 2005b). The incorporation of zirconia should increase the working temperature, water retention and mechanical stability of the composite membrane. Organic / inorganic composite membranes

from 0.12 to 0.14 S cm-1, while the composite Nafion®/TiO2 showed highest value than Nafion® 115 (125 µm) were value in the range 0.15-0.18 S cm-1 . A power density of 0.514 W cm-2 for Nafion® / 3 wt.% TiO2 composite against 0.354 W cm-2 for Nafion® 115 at 0.56 V and at T = 110°C was recorded. At 120°C, Nafion® 115 was damaged while the composite Nafion®/TiO2 membrane continued to work up to 130°C by reaching a power density of

Hybride membranes based on highly sulfonated poly(ether ether ketone) (SPEEK, DS = 0.9) where titania network was dispersed by *insitu* sol-gel reactions were prepared by Di Vona *et al.* (2007). Titania network was introduced following two routes: route 1 using titanium tetrabutoxide (Ti(OBu)4) and pyridine, and route 2 uses Ti(OBu)4 and 2,4-pentandione. Composite membranes prepared by route 2 showed a good conductivity property that can be attributed to the structural characteristics of the inorganic network generated in the presence of a chelating agent. This membrane shows a stable value ( = 5.8 × 10-2 S cm-1) at 120°C in fully hydrated conditions. Jian-hua *et al.* (2008) prepared a composite Nafion®/TiO2 membranes by carrying out *insitu* sol-gel reaction of Ti (OC4H9)4 followed by hydrolyzationcondensation in Nafion® 112, 1135 and 115. TiO2 prepared with this method was found to be 4 nm. TiO2 contents were 1.23, 2.47 and 3.16 wt.% for Nafion® 112/TiO2, Nafion® 1135/TiO2 and Nafion® 115/TiO2, respectively. The polarization characteristics of all three MEAs with the membranes containing TiO2 were improved significantly comparing with those of pure Nafion® film. A mixture of titanium tetraisopropoxide (TTIP) and PEG 1000 were used to prepare the titania sol by Liu *et al.* (2006). The average particle size of 20 nm was reported. The formed sol was deposited on the surface of Nafion® 112 membranes by spin coating. The TiO2 film is dense and well attached to the membrane, but some cracks in the membrane coated with diluted titania sol (e.g. 0.002 mg cm-2), while the membrane coated with thick titania sol (e.g. 0.021 mg cm-2) are very dense and cracks free. The proton conductivity of nano-TiO2-coated Nafion® membranes at 25 and 80°C were recorded with different TiO2 content. It was found that the maximum conductivity was with uncoated Nafion® 112, with values of 0.027 and 0.041 S cm-1 for 25 and 80°C, respectively. The conductivity of coated Nafion® decreases with increasing titania content. On the other hand, methanol permeability of the coated membranes decreases with increasing TiO2 content, namely from 3.2 × 10-6 to 1.7 × 10-6 cm2 s-1 at 25°C, and from 12.5 × 10-6 to 4.6 × 10-6 cm2 s-1 at 85°C. Thus the rise in temperature leads to a strong increase in permeation by a factor of about 3. The methanol permeability of the unmodified Nafion® 112 membrane was found to be 3.6 × 10-6 and 13 × 10-6 cm2 s-1 at 25 and 85°C, respectively. The cell performance with titania coated membrane with a content of 0.009 mg cm-2 exhibits a higher voltage than cells with Nafion® membrane or with the other coated membranes. Nafion® 112 delivered a maximum power density of 37 mV cm-2 at a current density of 200 mA cm-2. A maximum power density of 44 mW cm-2 is obtained from a fuel cell that employs the titania-coated

about 0.254 W cm-2 at 0.5 V.

membrane with 0.009 mg cm-2 content.

**6.4 Organic/zirconia and sulfated zirconia nanocomposite membranes** 

Zirconia as an inorganic filler was added to polymeric proton conductor membranes (Aricò *et al.,* 2003b, 2004; Nunes *et al.,* 2002; Sacca *et al.,* 2006; Silva *et al.,* 2005a, 2005b). The incorporation of zirconia should increase the working temperature, water retention and mechanical stability of the composite membrane. Organic / inorganic composite membranes based on SPEK and SPEEK for application in direct methanol fuel cell were synthesized by Nunes and co-workers (Nunes *et al.,* 2002). The inorganic fillers were introduced via *in situ* generation of SiO2, TiO2 or ZrO2. The modification with ZrO2 led to a 60-fold reduction of the methanol flux. However, a 13-fold reduction of conductivity was also observed.

Recast Nafion® composite membranes containing three different percentages (5%, 10% and 20%, w/w) of commercial ZrO2 as an inorganic filler were tested in fuel cell in a temperature range of 80-130°C in humidified H2/air gases at 3.0 bar abs by Sacca *et al.*  (2006). The introduction of 5 wt.% ZrO2 in Nafion® produces no evidence changes in the cell performance, while a better performance with 10 wt.% ZrO2 in Nafion® was obtained with a power density greater than 600 mW cm-2 at 0.6 V both at 80°C and 110°C. The good performance of 10 wt.% ZrO2 in Nafion® was maintained at 130°C with gas humidification of 85% RH, with a maximum power density of about 400 mW cm-2 was obtained in the potential range of 0.5-0.6 V. Silva *et al.* (2005a,2005b) prepared SPEEK/ZrO2 composite membranes using *insitu* formation of zirconia with zirconium tetrapropylate as alkoxide and acetyl acetone as chelating agent. The water/alkoxide ratio was always maintained higher than 1 to ensure the formation of a finely dispersed inorganic phase in the polymer solution. The thickness of the prepared membranes with 0.0, 2.5, 5.0, 7.5, 10.0, 12.5 wt.% of zirconium oxide were 188, 175, 133, 146, 128, 106 µm, respectively. The proton conductivity of the composite membranes was measured at 25°C and it was found that it decreases continuously with the ZrO2 content. Pervaporation experiments at 55°C showed that the membrane permeability towards methanol decreases with the amount of ZrO2. Composite SPEEK with 5.0, 7.5, 10.0 wt.% of ZrO2 were tested in fuel cell. The performance of 12.5 wt.% ZrO2 composite was very low due to the high ohmic resistance of the corresponding MEA. The membrane 5 wt.% ZrO2 presents the best DMFC performance among all the studied MEAs. Three types of superacidic sulfated zirconia (S-ZrO2) were prepared by different methods using hydrated zirconia and sulfuric acid by Hara and Miyayama (2004). Their proton conductivities were evaluated at 20-150°C under saturated water vapor pressure. It was found that the concentration on S-ZrO2 varied largely depending on the method of preparation. The S/Zr atomic ratio changed from 0.046 for the sample prepared through a mixture of hydrated zirconia powder and sulfuric acid to 0.35 for sample prepared through a mixture of hydrated zirconia sol and sulfuric acid. A powder compact of the former S-ZrO2 showed a proton conductivity of 4 × 10-2 S cm-1 at 70°C and 8 × 10-3 S cm-1 at 150°C, whereas that of the latter S-ZrO2 exhibited a high conductivity of 5 × 10-2 S cm-1 at 60-150°C.

S. Ren *et al.* (2006) prepared sulfated zirconia/ Nafion® 115 nanocomposite membrane by ion exchange of zirconium ions into the Nafion® followed by precipitation of sulphated ZrO2 by treatment in H2SO4. The incorporation of sulfated zirconia increases water uptake by the Nafion® membrane, and more water is absorbed than an unmodified membrane at high temperatures. The membrane proton conductivity is decreased slightly by ZrO2 impregnation. The proton conductivity of Nafion® 115 membrane was found to be 1.5 × 10-2 S cm-1 at 25°C, while that of S-ZrO2/ Nafion® 115 membrane is decreased to 5.0 × 10-3 S cm-1 at 25°C. At 110°C and above, the proton conductivity of S-ZrO2/ Nafion® 115 membrane is more than one-half that of the Nafion® 115 membrane. Fine particle superacidic sulfated zirconia (S-ZrO2) was synthesized by ameliorated method, and composite membranes with different S-ZrO2 contents were prepared by a recasting procedure from a suspension of S-

Organic / Inorganic Nanocomposite Membranes

Development for Low Temperature Fuel Cell Applications 527

F. Bauer & Willert-Porada (2004,2005,2006a) impregnated Nafion® with different ZrP contents. The proton conductivity of unmodified Nafion® 117 and Nafion® 117/ZrP (21 wt.%) was measured at three different temperatures, 80, 100 and 130°C. It was found that the presence of ZrP decreased the proton conductivity in all cases. At high humidity the conductivity first increased from 80 to 100°C and decreased at 130°. The conductivity decrease is more pronounced in case of the unmodified Nafion®. DMFC performance was conducted with Nafion® 117 and Nafion® 117/ZrP composite membrane at 130°C and 4.6 bar at the anode and cathode. It was found that the power output of Nafion® was higher than that of the composite membranes (Nafion® 117/13 wt.% ZrP and Nafion® 117/26 wt.% ZrP). At 0.2 A cm-2, a values of 420, 370 and 370 mV were measured for Nafion® 117, Nafion® 117/13 wt.% ZrP and Nafion® 117/26 wt.% ZrP, respectively. The crossover current was reduced by a factor of two as compared to the unmodified Nafion®. Also the two composite membranes tested exhibited a higher OCV than unmodified Nafion®, which also indicates lower methanol permeability. A values of 725, 768 and 760 mV were reported for

Nafion® 117, Nafion® 117/13 wt.% ZrP and Nafion® 117/26 wt.% ZrP, respectively.

about 1500 mA cm-2 at 0.45 V at a temperature of 130°C and a pressure of 3 bar.

Yang and coworkers (Yang *et al.,* 2001a,2001b,2004; Costamagna *et al.,* 2002) introduced ZrP into Nafion® 115 through ion exchange of Zr4+ followed by precipitation of ZrP by treatment with phosphoric acid as described by Grot and Rajendran (1999). An MEA employing Nafion® 115/23 wt.% ZrP gave a H2/O2 PEMFC performance of about 1000 mA cm-2 at 0.45 V at a temperature of 130°C and a pressure of 3 bar, while unmodified Nafion® 115 gave 250 mA cm-2 at 0.45 V when operated under the same conditions of temperature and pressure. Similar experiment performed with recast Nafion® and recast Nafion®/36 wt.% ZrP composite confirmed an analogous improvement of performance of the composite membrane over the unmodified ones. The composite recast Nafion®/36 wt.% ZrP gave

Alberti *et al.* (2005b,2007) prepared a recast Nafion® filled with ZrP according to the procedure described in the patent (Alberti *et al.*, 2005b). Zirconyl propionate was used instead of zirconyl oxychloride and the solutions were dissolved in DMF. The IEC (ion exchange capacity) of the prepared composite membrane was found to be higher than those previously reported for Nafion®/ZrP membranes prepared according to the exchange method (F. Bauer & Willert-Porada, 2006b; Yang *et al.,* 2004). The proton conductivity was found to decrease with increasing the filler loading, which is in agreement with the trend found for Nafion®/ZrP prepared by the exchange method (F. Bauer & Willert-Porada, 2005; Casiola *et al.,* 2008; Yang *et al.,* 2004). At constant RH, the logarithm of conductivity shows approximately the same linear dependence on ZrP loading in the RH range 50-90%. However, at 35% RH, the increase in the ZrP loading results in a larger conductivity decrease than that observed in the above RH range. A similar behavior was also reported for Nafion®/ZrP membranes obtained by the exchange method already at 50% RH (Yang *et al.,*  2004), thus confirming that the same type of filler prepared by using different procedures gives rise to different membrane properties. It was concluded that the main difference between pure Nafion® and composite membranes appear at low RH and high filler loading. It was reported that the Nafion® conductivity undergoes an irreversible decay above certain values of temperature and RH, which was attributed to an anisotropic swelling of the membrane, pressed between the electrodes, in the direction parallel to the membrane surface (Alberti *et al.,* 2001; Casiola *et al.,* 2006). It was also found that, at a given RH value,

ZrO2 powder and Nafion® solution (Zhai *et al.,* 2006). The results showed that the IEC (Ion Exchange Capacity) of composite membrane increased with the content of S-ZrO2 and S-ZrO2 was found to be compatible with the Nafion® matrix. The incorporation of the S-ZrO2 increased the crystallinity and also improved the initial degradation temperature of the composite membrane. The performance of single cell was the best when the S-ZrO2 content was 15 wt.% and achieved 1.35 W cm-2 at 80°C and 0.99 W cm-2 at 120°C based on H2/O2 and at a pressure of 2 atm, the performance of the single cell with optimized S-ZrO2 was far more than that of the Nafion® at the same condition (e.g. 1.28 W cm-2 at 80°C and 0.75 W cm-2 at 120°C). A self-humidifying composite membrane based on Nafion® hybrid with SiO2 supported sulfated zirconia particles (SiO2-SZ) was fabricated and investigated for fuel cell application by Bi *et al.* (2008). The bi-functional SiO2-SZ particles, possessing hygroscopic property and high proton conductivity, was incorporated in recast Nafion® membrane. The proton conductivity of Nafion®/SiO2-SZ, Nafion®/SiO2 and recast Nafion® under dry and wet H2/O2 conditions at 60°C were compared. The two composite membranes showed higher proton conductivity in contrast to the recast Nafion® membrane under 0% RH mode with the order Nafion®/SiO2-SZ > Nafion®/SiO2 > recast Nafion®. Under 100% RH mode, the Nafion®/SiO2-SZ composite membrane also exhibited the highest proton conductivity values among the three membranes. The proton conductivity of Nafion®/SiO2-SZ membrane was 6.95 × 10-2 S cm-1 and the value was higher than Nafion®/ SiO2 membrane of 5.54 × 10-2 S cm-1 and recast Nafion® membrane of 6.55 × 10-2 S cm-1. Single cell performance of these composite membranes were tested with wet H2 and O2 at 60°C. Nafion®/ SiO2 composite membrane exhibited the worst output performance (0.864 W cm-2) due to the increased proton conductive resistance caused by incorporated less proton conductivity of SiO2 particles. In contrast, Nafion®/SiO2-SZ composite membrane showed similar cell performance to recast Nafion® (1.045 W cm-2 vs. 1.014 W cm-2). However, the single cell performance of Nafion®/SiO2-SZ and Nafion®/SiO2 membranes with dry H2 and O2 at 60°C were 0.980 and 0.742 W cm-2, respectively. These results shows that the composite membrane perform better than unmodified Nafion® (i.e 0.635 W cm-2) under dry condition and the composite membranes manifested a good water retention.

#### **6.5 Organic/zirconium phosphate nanocomposite membranes**

Layered zirconium phosphate (ZrP) and phosphonates were used as inorganic fillers of proton conducting polymeric membranes because they are proton conductors with good chemical and thermal stability. Under the most favourable conditions, their conductivity is around 10-2 S cm-1 for high surface ZrP (Alberti *et al.,* 1978; F. Bauer & Willert-Porada, 2005) and 10-1 S cm-1 for zirconium phosphate sulfophenylenphosphonates (Alberti *et al.,* 2004,2005a). ZrP can be added to Nafion® (Alberti *et al.,* 2007; F. Bauer & Willert-Porada, 2004,2005,2006a; Casiola *et al.,* 2008; Costamagna *et al.,* 2002; Grot & Rajendran*,* 1999; Helen *et al.,* 2006,2007; Hou *et al.,* 2008; R. Jiang *et al.,* 2006a; Kuan *et al.,* 2006; H.K. Lee *et al.,* 2004; Mitov *et al.,* 2006; Yang *et al.,* 2001a,2001b,2004), SPEEK (Bonnet *et al.,* 2000; Nunes *et al.,* 2002; Silva *et al.,* 2005b; Tchicaya-Bouckary *et al.,* 2002; Triphathi *et al.,* 2007,2009), SPEK (Nunes *et al.,* 2002; Ruffmann *et al.,* 2003) and difulfonated poly(arulene ether sulfone) (Hill *et al.,* 2006). A similar inorganic material derived from ZrP, named zirconium phosphate sulfophenylen-phophonate was also used as a filler with Nafion® (Y.T. Kim *et al.,* 2004), SPEEK (Krishnan *et al.,* 2006) and PVDF (polyvinyl -lidene fluoride) (Casiola *et al.,* 2005).

ZrO2 powder and Nafion® solution (Zhai *et al.,* 2006). The results showed that the IEC (Ion Exchange Capacity) of composite membrane increased with the content of S-ZrO2 and S-ZrO2 was found to be compatible with the Nafion® matrix. The incorporation of the S-ZrO2 increased the crystallinity and also improved the initial degradation temperature of the composite membrane. The performance of single cell was the best when the S-ZrO2 content was 15 wt.% and achieved 1.35 W cm-2 at 80°C and 0.99 W cm-2 at 120°C based on H2/O2 and at a pressure of 2 atm, the performance of the single cell with optimized S-ZrO2 was far more than that of the Nafion® at the same condition (e.g. 1.28 W cm-2 at 80°C and 0.75 W cm-2 at 120°C). A self-humidifying composite membrane based on Nafion® hybrid with SiO2 supported sulfated zirconia particles (SiO2-SZ) was fabricated and investigated for fuel cell application by Bi *et al.* (2008). The bi-functional SiO2-SZ particles, possessing hygroscopic property and high proton conductivity, was incorporated in recast Nafion® membrane. The proton conductivity of Nafion®/SiO2-SZ, Nafion®/SiO2 and recast Nafion® under dry and wet H2/O2 conditions at 60°C were compared. The two composite membranes showed higher proton conductivity in contrast to the recast Nafion® membrane under 0% RH mode with the order Nafion®/SiO2-SZ > Nafion®/SiO2 > recast Nafion®. Under 100% RH mode, the Nafion®/SiO2-SZ composite membrane also exhibited the highest proton conductivity values among the three membranes. The proton conductivity of Nafion®/SiO2-SZ membrane was 6.95 × 10-2 S cm-1 and the value was higher than Nafion®/ SiO2 membrane of 5.54 × 10-2 S cm-1 and recast Nafion® membrane of 6.55 × 10-2 S cm-1. Single cell performance of these composite membranes were tested with wet H2 and O2 at 60°C. Nafion®/ SiO2 composite membrane exhibited the worst output performance (0.864 W cm-2) due to the increased proton conductive resistance caused by incorporated less proton conductivity of SiO2 particles. In contrast, Nafion®/SiO2-SZ composite membrane showed similar cell performance to recast Nafion® (1.045 W cm-2 vs. 1.014 W cm-2). However, the single cell performance of Nafion®/SiO2-SZ and Nafion®/SiO2 membranes with dry H2 and O2 at 60°C were 0.980 and 0.742 W cm-2, respectively. These results shows that the composite membrane perform better than unmodified Nafion® (i.e 0.635 W cm-2) under dry condition

and the composite membranes manifested a good water retention.

**6.5 Organic/zirconium phosphate nanocomposite membranes** 

(Casiola *et al.,* 2005).

Layered zirconium phosphate (ZrP) and phosphonates were used as inorganic fillers of proton conducting polymeric membranes because they are proton conductors with good chemical and thermal stability. Under the most favourable conditions, their conductivity is around 10-2 S cm-1 for high surface ZrP (Alberti *et al.,* 1978; F. Bauer & Willert-Porada, 2005) and 10-1 S cm-1 for zirconium phosphate sulfophenylenphosphonates (Alberti *et al.,* 2004,2005a). ZrP can be added to Nafion® (Alberti *et al.,* 2007; F. Bauer & Willert-Porada, 2004,2005,2006a; Casiola *et al.,* 2008; Costamagna *et al.,* 2002; Grot & Rajendran*,* 1999; Helen *et al.,* 2006,2007; Hou *et al.,* 2008; R. Jiang *et al.,* 2006a; Kuan *et al.,* 2006; H.K. Lee *et al.,* 2004; Mitov *et al.,* 2006; Yang *et al.,* 2001a,2001b,2004), SPEEK (Bonnet *et al.,* 2000; Nunes *et al.,* 2002; Silva *et al.,* 2005b; Tchicaya-Bouckary *et al.,* 2002; Triphathi *et al.,* 2007,2009), SPEK (Nunes *et al.,* 2002; Ruffmann *et al.,* 2003) and difulfonated poly(arulene ether sulfone) (Hill *et al.,* 2006). A similar inorganic material derived from ZrP, named zirconium phosphate sulfophenylen-phophonate was also used as a filler with Nafion® (Y.T. Kim *et al.,* 2004), SPEEK (Krishnan *et al.,* 2006) and PVDF (polyvinyl -lidene fluoride) F. Bauer & Willert-Porada (2004,2005,2006a) impregnated Nafion® with different ZrP contents. The proton conductivity of unmodified Nafion® 117 and Nafion® 117/ZrP (21 wt.%) was measured at three different temperatures, 80, 100 and 130°C. It was found that the presence of ZrP decreased the proton conductivity in all cases. At high humidity the conductivity first increased from 80 to 100°C and decreased at 130°. The conductivity decrease is more pronounced in case of the unmodified Nafion®. DMFC performance was conducted with Nafion® 117 and Nafion® 117/ZrP composite membrane at 130°C and 4.6 bar at the anode and cathode. It was found that the power output of Nafion® was higher than that of the composite membranes (Nafion® 117/13 wt.% ZrP and Nafion® 117/26 wt.% ZrP). At 0.2 A cm-2, a values of 420, 370 and 370 mV were measured for Nafion® 117, Nafion® 117/13 wt.% ZrP and Nafion® 117/26 wt.% ZrP, respectively. The crossover current was reduced by a factor of two as compared to the unmodified Nafion®. Also the two composite membranes tested exhibited a higher OCV than unmodified Nafion®, which also indicates lower methanol permeability. A values of 725, 768 and 760 mV were reported for Nafion® 117, Nafion® 117/13 wt.% ZrP and Nafion® 117/26 wt.% ZrP, respectively.

Yang and coworkers (Yang *et al.,* 2001a,2001b,2004; Costamagna *et al.,* 2002) introduced ZrP into Nafion® 115 through ion exchange of Zr4+ followed by precipitation of ZrP by treatment with phosphoric acid as described by Grot and Rajendran (1999). An MEA employing Nafion® 115/23 wt.% ZrP gave a H2/O2 PEMFC performance of about 1000 mA cm-2 at 0.45 V at a temperature of 130°C and a pressure of 3 bar, while unmodified Nafion® 115 gave 250 mA cm-2 at 0.45 V when operated under the same conditions of temperature and pressure. Similar experiment performed with recast Nafion® and recast Nafion®/36 wt.% ZrP composite confirmed an analogous improvement of performance of the composite membrane over the unmodified ones. The composite recast Nafion®/36 wt.% ZrP gave about 1500 mA cm-2 at 0.45 V at a temperature of 130°C and a pressure of 3 bar.

Alberti *et al.* (2005b,2007) prepared a recast Nafion® filled with ZrP according to the procedure described in the patent (Alberti *et al.*, 2005b). Zirconyl propionate was used instead of zirconyl oxychloride and the solutions were dissolved in DMF. The IEC (ion exchange capacity) of the prepared composite membrane was found to be higher than those previously reported for Nafion®/ZrP membranes prepared according to the exchange method (F. Bauer & Willert-Porada, 2006b; Yang *et al.,* 2004). The proton conductivity was found to decrease with increasing the filler loading, which is in agreement with the trend found for Nafion®/ZrP prepared by the exchange method (F. Bauer & Willert-Porada, 2005; Casiola *et al.,* 2008; Yang *et al.,* 2004). At constant RH, the logarithm of conductivity shows approximately the same linear dependence on ZrP loading in the RH range 50-90%. However, at 35% RH, the increase in the ZrP loading results in a larger conductivity decrease than that observed in the above RH range. A similar behavior was also reported for Nafion®/ZrP membranes obtained by the exchange method already at 50% RH (Yang *et al.,*  2004), thus confirming that the same type of filler prepared by using different procedures gives rise to different membrane properties. It was concluded that the main difference between pure Nafion® and composite membranes appear at low RH and high filler loading. It was reported that the Nafion® conductivity undergoes an irreversible decay above certain values of temperature and RH, which was attributed to an anisotropic swelling of the membrane, pressed between the electrodes, in the direction parallel to the membrane surface (Alberti *et al.,* 2001; Casiola *et al.,* 2006). It was also found that, at a given RH value,

Organic / Inorganic Nanocomposite Membranes

(negative) effect on the proton conductivity.

S cm-1 (Y.T. Kim *et al.,* 2004).

**6.6 Organic/palladium nanocomposite membranes** 

Development for Low Temperature Fuel Cell Applications 529

unmodified SPEEK (SD = 42%) membrane presented the maximum power density output. It achieved an output power density value of 10.4 mW cm-1 for 51.8 mA cm-2. The unmodified membrane with SD = 68% could not be characterized due to its instability (high swelling or even solubility). However, the SPEEK (SD=68%)/20 wt.% ZrP/11.2 wt.% PBI had even higher power density than the membrane with SD = 42% for current density lower than 25 mA cm-2. When the relative humidity at the cathode feed was increased to 138%, the SPEEK (SD=68%)/20 wt.% ZrP/11.2 wt.% PBI membrane had the best performance, with an output power density value of 14.7 mW cm-1 for 58.8 mA cm-2 (Silva *et al.,* 2005c). However, the filler addition to SPEEK (SD = 42%) besides reducing the crossover had an excessive

Zirconium phosphate sulfophenylphosphate, a functionalized ZrP, was incorporated in Nafion® (Casiola *et al.,* 2008; Y.T. Kim *et al.,* 2004), SPEEK (Bonnet *et al.,* 2000; Krishnan *et al.,* 2006) and PVA (Casiola *et al.,* 2005). ZrP sulfophenylphosphates (ZrSPP) are a class of layered materials exhibiting proton conductivity comparable with that of Nafion membranes (i.e. 0.07-0.1 S cm-1 at 100°C and 100% RH) due to the presence of the –SO3H groups in the interlayer region (Alberti *et al.,* 2005a). The functionalization of the ZrP nanoparticles with SPP is therefore expected to increase the conductivity of the Nafion®/ZrP membranes. These phosphonates are ideally obtained by partial replacement of SPP groups for the phosphate groups of ZrP (Casiola *et al.,* 2008). Nafion® 117/ZrSPP composite membranes was found to have a higher conductivity than the parent Nafion® 117/ 20 wt.% ZrP and pure Nafion® 117 membrane at 100°C and RH between 30-90%, with highest value approaching 0.1 S cm-1 at RH = 90% (Casiola *et al.,* 2008), while appreciable dehydration of Nafion® 117 resulted in drastic reduction of proton conductivity above 100°C (Y.T. Kim *et al.,* 2004). However, the proton conductivity of Nafion®/12.5 wt.% ZrSPP composite membrane slightly increased up to 70°C and remained constant until 140°C, with a conductivity of 0.07

This approach is to utilize the unique properties of palladium which is permeable to protons, but very resistant to methanol transport. It was suggested first by Pu *et al.* (1995) where they used a palladium foil of 25 µm thick sandwiched between two Nafion® 115 sheets. They proved that with this approach methanol crossover can be reduced, but the cell

Choi *et al.* (2001) used the same approach by sputtering metallic palladium on the surface of a Nafion® 117 to plug the pores of Nafion®. The palladium film was found to be 20 nm. Methanol permeability was reduced from 2.392 10-6 cm2 s-1 in unmodified Nafion® 117 to 1.7 10-6 cm2 s-1 in Pd-sputtered membrane and the cell performance at 95°C was improved compared to the unmodified Nafion® 117 membrane. The methanol permeability reduction was confirmed by the high OCV obtained with the modified membranes. Similarly, Yoon *et al.* (2002) used sputtering technique to deposit Pd film on the surface of Nafion® 117. It was found that the Pd films thinner than 300 Å were dense and appeared to be well attached to the membrane, but there were many cracks in the 1000 Å films. The 1000 Å films were very unstable and were easily delaminated from the membrane surface. When the composite membrane is immersed in water, the Nafion® membrane swells very much, but the Pd film can not expand as much as the membrane (Yoon *et al.*, 2002). The proton conductivity was

performance will be lower due to the increase of the membrane thickness.

the decay temperature for composite Nafion®/ZrP membranes was higher than for pure recast Nafion® membranes prepared and thermally treated under the same conditions used for the composite sample. The conductivity of the pure Nafion® starts to decay at temperatures higher than 130°C, while the conductivity of the composite membrane is stable up to 140°C. Nafion® 115/23 wt.% ZrP was prepared by ion exchange and tested for DMFC by Hou *et al.* (2008). It was found that the liquid uptakes of Nafion® 115 and Nafion® 115/23 wt.% ZrP membranes increased linearly with increasing methanol concentration. The slope of the plot for Nafion® 115 was larger than for the composite membrane i.e. the liquid uptake of Nafion® 115 increased from 34.3% in 0M methanol solution to 58.6% in 10M methanol solution, while that of the composite membrane increased from 28.3% to 37.5% in the corresponding methanol solution. When 23 wt.% of ZrP was incorporated into Nafion® 115, the IEC of the resulting membrane increased significantly to 1.93 meq/g from a value of 0.909 meq/g for pure Nafion® 115. The proton conductivity at room temperature of Nafion® 115 and Nafion® 115/23 wt.% ZrP was found to be 0.10 and 0.084 S cm-1, respectively. Also it was found that the methanol crossover through the composite membrane was suppressed. The DMFC test at 75°C and 5M methanol solution shows that the composite membrane performed better that the pure Nafion® 115, with a peak power density of 96.3 and 91.6 mW cm-2, respectively. When the methanol concentration was further increase to 10M, the peak power density of DMFC with composite membrane was 76.19 mW cm-2, which is higher than that for Nafion® (42.4 mW cm-2). However, Bonnet *et al.* (2000) investigated the incorporation of ZrP in SPEEK. It was found that the conductivity of the composite membrane exceeded that of the polymer-only membrane, and increases with the amount of the filler (from 0-30 wt.%) up to 0.08 S cm-1 when measured at 100°C and 100% RH. A similar trend was also observed, when the RH varied from 50 to 100%. At all value of RH, the composite membrane SPEEK/20 wt.% ZrP conductivity was higher than that of nonmodified SPEEK. A similar membrane was prepared by Tchicaya-Bouckary *et al.* (2002). The conductivity of SPEEK/25 wt.% ZrP was found to be weakly temperature dependence, the conductivity increases from 2 × 10-2 to 5 × 10-2 S cm-1 between 20 and 100°C at 100% RH. This composite membrane was tested in H2/O2 fuel cell at 100°C at an oxygen pressure of 3.6 bars absolute. A value of 1 A cm-2 at 0.6 V was reported. These results are much better than that reported for Nafion® 115/ ZrP (Costamagna *et al.,* 2002) which provided ca. 0.7 A cm-2 at 0.6 V, 130°C and 3 bars pressure. Nunes and co-workers studied the incorporation of ZrP in SPEEK and SPEK (Nunes *et al.,* 2002; Ruffmann *et al.,* 2003; Silva *et al.,* 2005c). ZrP was prepared according to the procedure described by Belyakov & Linkov (1999). It was reported that the incorporation of ZrP did not lead to a particular reduction of water and methanol permeability, and the proton conductivity at 25°C was decreased to the same extent (44 mS cm-1 for a SPEK/ 20 wt.% ZrP and 50 mS cm-1 for a pure SPEK membranes). A good values of proton conductivities were measured for membranes with 70/20/10 and 69/17/14 wt.% SPEK/ZrP/ZrO2 where a conductivities of 45 and 35 mS cm-1 were measured, respectively (Nunes *et al.,* 2002). ZrP pretreated with *n*-propylamine and PBI was incorporated with SPEEK (Silva *et al.,* 2005c), the proton conductivity of the composite membranes decreases with the amount of inorganic incorporation. On the other hand, methanol and water permeability in the pervaporation experiments at 55°C showed that it decrease with the amount of inorganic incorporation. Similar trend was found for the composite membranes permeability towards nitrogen, oxygen and carbon dioxide. The SPEEK composite membranes were tested in a DMFC at 110°C, it was found that the

the decay temperature for composite Nafion®/ZrP membranes was higher than for pure recast Nafion® membranes prepared and thermally treated under the same conditions used for the composite sample. The conductivity of the pure Nafion® starts to decay at temperatures higher than 130°C, while the conductivity of the composite membrane is stable up to 140°C. Nafion® 115/23 wt.% ZrP was prepared by ion exchange and tested for DMFC by Hou *et al.* (2008). It was found that the liquid uptakes of Nafion® 115 and Nafion® 115/23 wt.% ZrP membranes increased linearly with increasing methanol concentration. The slope of the plot for Nafion® 115 was larger than for the composite membrane i.e. the liquid uptake of Nafion® 115 increased from 34.3% in 0M methanol solution to 58.6% in 10M methanol solution, while that of the composite membrane increased from 28.3% to 37.5% in the corresponding methanol solution. When 23 wt.% of ZrP was incorporated into Nafion® 115, the IEC of the resulting membrane increased significantly to 1.93 meq/g from a value of 0.909 meq/g for pure Nafion® 115. The proton conductivity at room temperature of Nafion® 115 and Nafion® 115/23 wt.% ZrP was found to be 0.10 and 0.084 S cm-1, respectively. Also it was found that the methanol crossover through the composite membrane was suppressed. The DMFC test at 75°C and 5M methanol solution shows that the composite membrane performed better that the pure Nafion® 115, with a peak power density of 96.3 and 91.6 mW cm-2, respectively. When the methanol concentration was further increase to 10M, the peak power density of DMFC with composite membrane was 76.19 mW cm-2, which is higher than that for Nafion® (42.4 mW cm-2). However, Bonnet *et al.* (2000) investigated the incorporation of ZrP in SPEEK. It was found that the conductivity of the composite membrane exceeded that of the polymer-only membrane, and increases with the amount of the filler (from 0-30 wt.%) up to 0.08 S cm-1 when measured at 100°C and 100% RH. A similar trend was also observed, when the RH varied from 50 to 100%. At all value of RH, the composite membrane SPEEK/20 wt.% ZrP conductivity was higher than that of nonmodified SPEEK. A similar membrane was prepared by Tchicaya-Bouckary *et al.* (2002). The conductivity of SPEEK/25 wt.% ZrP was found to be weakly temperature dependence, the conductivity increases from 2 × 10-2 to 5 × 10-2 S cm-1 between 20 and 100°C at 100% RH. This composite membrane was tested in H2/O2 fuel cell at 100°C at an oxygen pressure of 3.6 bars absolute. A value of 1 A cm-2 at 0.6 V was reported. These results are much better than that reported for Nafion® 115/ ZrP (Costamagna *et al.,* 2002) which provided ca. 0.7 A cm-2 at 0.6 V, 130°C and 3 bars pressure. Nunes and co-workers studied the incorporation of ZrP in SPEEK and SPEK (Nunes *et al.,* 2002; Ruffmann *et al.,* 2003; Silva *et al.,* 2005c). ZrP was prepared according to the procedure described by Belyakov & Linkov (1999). It was reported that the incorporation of ZrP did not lead to a particular reduction of water and methanol permeability, and the proton conductivity at 25°C was decreased to the same extent (44 mS cm-1 for a SPEK/ 20 wt.% ZrP and 50 mS cm-1 for a pure SPEK membranes). A good values of proton conductivities were measured for membranes with 70/20/10 and 69/17/14 wt.% SPEK/ZrP/ZrO2 where a conductivities of 45 and 35 mS cm-1 were measured, respectively (Nunes *et al.,* 2002). ZrP pretreated with *n*-propylamine and PBI was incorporated with SPEEK (Silva *et al.,* 2005c), the proton conductivity of the composite membranes decreases with the amount of inorganic incorporation. On the other hand, methanol and water permeability in the pervaporation experiments at 55°C showed that it decrease with the amount of inorganic incorporation. Similar trend was found for the composite membranes permeability towards nitrogen, oxygen and carbon dioxide. The SPEEK composite membranes were tested in a DMFC at 110°C, it was found that the unmodified SPEEK (SD = 42%) membrane presented the maximum power density output. It achieved an output power density value of 10.4 mW cm-1 for 51.8 mA cm-2. The unmodified membrane with SD = 68% could not be characterized due to its instability (high swelling or even solubility). However, the SPEEK (SD=68%)/20 wt.% ZrP/11.2 wt.% PBI had even higher power density than the membrane with SD = 42% for current density lower than 25 mA cm-2. When the relative humidity at the cathode feed was increased to 138%, the SPEEK (SD=68%)/20 wt.% ZrP/11.2 wt.% PBI membrane had the best performance, with an output power density value of 14.7 mW cm-1 for 58.8 mA cm-2 (Silva *et al.,* 2005c). However, the filler addition to SPEEK (SD = 42%) besides reducing the crossover had an excessive (negative) effect on the proton conductivity.

Zirconium phosphate sulfophenylphosphate, a functionalized ZrP, was incorporated in Nafion® (Casiola *et al.,* 2008; Y.T. Kim *et al.,* 2004), SPEEK (Bonnet *et al.,* 2000; Krishnan *et al.,* 2006) and PVA (Casiola *et al.,* 2005). ZrP sulfophenylphosphates (ZrSPP) are a class of layered materials exhibiting proton conductivity comparable with that of Nafion membranes (i.e. 0.07-0.1 S cm-1 at 100°C and 100% RH) due to the presence of the –SO3H groups in the interlayer region (Alberti *et al.,* 2005a). The functionalization of the ZrP nanoparticles with SPP is therefore expected to increase the conductivity of the Nafion®/ZrP membranes. These phosphonates are ideally obtained by partial replacement of SPP groups for the phosphate groups of ZrP (Casiola *et al.,* 2008). Nafion® 117/ZrSPP composite membranes was found to have a higher conductivity than the parent Nafion® 117/ 20 wt.% ZrP and pure Nafion® 117 membrane at 100°C and RH between 30-90%, with highest value approaching 0.1 S cm-1 at RH = 90% (Casiola *et al.,* 2008), while appreciable dehydration of Nafion® 117 resulted in drastic reduction of proton conductivity above 100°C (Y.T. Kim *et al.,* 2004). However, the proton conductivity of Nafion®/12.5 wt.% ZrSPP composite membrane slightly increased up to 70°C and remained constant until 140°C, with a conductivity of 0.07 S cm-1 (Y.T. Kim *et al.,* 2004).

#### **6.6 Organic/palladium nanocomposite membranes**

This approach is to utilize the unique properties of palladium which is permeable to protons, but very resistant to methanol transport. It was suggested first by Pu *et al.* (1995) where they used a palladium foil of 25 µm thick sandwiched between two Nafion® 115 sheets. They proved that with this approach methanol crossover can be reduced, but the cell performance will be lower due to the increase of the membrane thickness.

Choi *et al.* (2001) used the same approach by sputtering metallic palladium on the surface of a Nafion® 117 to plug the pores of Nafion®. The palladium film was found to be 20 nm. Methanol permeability was reduced from 2.392 10-6 cm2 s-1 in unmodified Nafion® 117 to 1.7 10-6 cm2 s-1 in Pd-sputtered membrane and the cell performance at 95°C was improved compared to the unmodified Nafion® 117 membrane. The methanol permeability reduction was confirmed by the high OCV obtained with the modified membranes. Similarly, Yoon *et al.* (2002) used sputtering technique to deposit Pd film on the surface of Nafion® 117. It was found that the Pd films thinner than 300 Å were dense and appeared to be well attached to the membrane, but there were many cracks in the 1000 Å films. The 1000 Å films were very unstable and were easily delaminated from the membrane surface. When the composite membrane is immersed in water, the Nafion® membrane swells very much, but the Pd film can not expand as much as the membrane (Yoon *et al.*, 2002). The proton conductivity was

Organic / Inorganic Nanocomposite Membranes

only 32 mW cm-2.

**6.7 Organic/montmorillonite nanocomposite membranes** 

Development for Low Temperature Fuel Cell Applications 531

ionomer onto Nafion® 112 membranes. The Pd particles, size of about 1.8 nm in average, are charged by PDDA (polydiallyldimethylammonium chloride) ionomers. The Pd loading of the first-layer MLSA Nafion® membranes was 0.63 µg cm-2, and the surface coverage of the Pd nanoparticles on the Nafion® membrane was estimated as 22%. After 5-double-layer Pd particles/Nafion® ionomers assembling, the Pd loading reached to 2.86 µg cm-2. The methanol crossover current of the original Nafion® membranes and 1-double-layer, 2 double-layer, 3-double-layer, 4-double-layer, 5-double-layer MLSA Nafion® membrane were 0.0495, 3.87 10-3, 1.38 10-3, 7.32 10-4, 5.16 10-4 and 4.25 10-3 A cm-2, respectively, corresponding conductivities of 0.112, 0.110, 0.105, 0.094, 0.087 and 0.081 S cm-2. No DMFC data were provided, however, it was suggested that the 3-double-layer self-assembly membrane is the best suited for DMFC application, since it has a methanol crossover decreased to 0.86%, and a conductivity remaining at 83.9% comparing to original Nafion® membrane. Electroless plating was also used to deposit Pd layer on Nafion® membranes (Hejze *et al.*, 2005; Sun *et al.*, 2005). Palladium layer can reduce methanol crossover when coated on the surface of Nafion®, i.e. limiting current from methanol permeation through a membrane electrode assembly was reduce from 64 to 57 mA cm-2 for 1M methanol, and from 267 to 170 mA cm-2 for 5M methanol, for Nafion® 115 and Pd/Nafion® 115 membranes respectively (Sun *et al.*, 2005). Also it was demonstrated that the DMFC performance increase with the incorporation of Pd in Nafion® 115. When 1M methanol was used, the power density increased from 36 to 45 mW cm-2, for Nafion® 115 and Pd/Nafion® 115, respectively. When 5M methanol was used, the maximum power density on Pd/Nafion® 115 was 72 mW cm-2, while the performance of MEA with pure Nafion® 115 membranes was

Montmorillonite (MMT) is a type of layered silicate composed of silica tetrahedral and alumina octahedral sheets (J. Chang *et al.*, 2003) and its intercalation into Nafion® membrane can decrease successfully the methanol permeability and improve mechanical property (Jung *et al.*, 2003; Song *et al.*, 2004). Research by J. Chang *et al*. (2003) has showed that layered silicates incorporated into SPEEK membranes helped to reduce swelling significantly in hot water and decrease the methanol crossover without a serious reduction of the proton conductivity. Gaowen and Zhentao (2005) prepared organically modified MMT (OMMT) through ion exchange reaction between alkylammonium cations and metal cations. The nanocomposite membranes (SPEEK/OMMT) were prepared using the solution intercalation technique. The water uptake of SPEEK membrane increased rapidly above 50°C, while the SPEEK/OMMT composite membranes posses of rather constant water uptake up to 80°C. This indicates that MMT layers incorporated into SPEEK matrix prevent extreme swelling of the composite membranes due to the cohesion of the functional groups between SPEEK matrix and MMT layers. The proton conductivity of the membrane was measured at temperature ranging from 22 to 110°C. It was found that the conductivity of SPEEK/OMMT composite is lower than that of the pristine SPEEK and decreases sequentially as the content of OMMT increases, which is due to prolonging the transfer route of proton. The conductivity of SPEEK/OMMT (5 wt.%) approaches the value of Nafion® 115 at 90°C and reaches 1.2 10-2 S cm-1. The activation energies of SPEEK/OMMT are higher than that of Nafion® 115, where the value of 32.08 kJ/mol and 10.8 kJ/mol, respectively, were found. Methanol permeability was found to be in the

found to decrease with increasing Pd thickness. For the Pd-1000 Å film, 30% reduction in conductivity was observed. Methanol permeability at 25°C decreased with increasing Pd thickness and they varied from 2.90 10-6 to 2.23 10-6 cm2 s-1 by deposition of Pd film of 1000 Å on the Nafion® 117 membrane. For the Pd-1000 Å-Nafion® 115 membrane, permeability decreased as much as 44% from 2.97 10-6 to 1.67 10-6 cm2 s-1. The cell performance at 90°C exhibited a slight decrease with the Pd layered Nafion® 115 membranes. The methanol crossover through an MEA is inversely proportional to current density and thus, its effect on the performance is more prominent at low current densities. It was found that at low current densities cell performance increased in Pd film of 1000 Å on the Nafion® 115, where the Pd film act as a barrier to methanol crossover. However, these results are different of that of Choi *et al.* who observed significant increase in DMFC performance in Pd-20 nm film on Nafion® 117.

Z.Q. Ma *et al.* (2003) followed a different approach, where Nafion® membrane was modified by sputtering a thin layer of Pt/Pd-Ag/Pt on its surface. The methanol crossover can be reduced by sputtering Pd-Ag alloys over the polymer electrolyte, furthermore, when hydrogen is absorbed and dissolved in the membrane, the palladium-silver alloy membrane not only reduces the possibility of embrittling due to α phase transition at low temperatures (< 150°C) but also leads to a higher permeability for hydrogen than pure palladium membrane. The composite membranes were prepared as follows: on one side of the Nafion® 117 membrane, a 2 nm Pt film was first deposited. This was followed by a Pd-Ag film with three different thicknesses (0.1, 0.2 and 1 µm). On the top of the Pd-Ag, it was coated with another 2 nm Pt film. Before a MEA was manufactured, a 4-5 µm layer of Nafion® polymer was recast over the surface of the sputtered Pt/Pd-Ag/Pt layer with Nafion® solution. The final membrane prepared was in the form of Nafion®117/Pt/Pd-Ag/Pt/Nafion®, containing 0.0086 mg cm-2 Pt, 0.90 mg cm-2 Pd, and 0.27 mg cm-2 Ag. The cell performance and the OCV increased with increasing the sputtering alloy layer thickness and the best performance and the highest OCV were found with the 1 µm Pd-Ag film. Also it was showed that the performance with the 1 µm Pd-Ag film is higher than that of cell with a Nafion® membrane having catalyst loading twice as high. Palladinized Nafion® composite membrane was prepared via ion-exchange and chemical reduction method (Y.J. Kim *et al.*, 2004). Palladium(II) acetylacetonate and tetraammine-palladium(II) chloride hydride were used as palladium precursors. Nafion® 117 samples were immersed in palladium precursor solutions followed by chemical reduction of palladium precursors by sodium borohydride. The use of tetraamminepalladium(II) chloride hydride formed 40-50 nm of palladium particles while palladium(II) acetylacetonate formed 5-10 nm of particles. For all palladinized samples, water uptake was higher than for unfilled Nafion® whereas above a certain amount of incorporated palladium, methanol uptake was lower than for unfilled Nafion®. Incorporating Pd nanoparticles decreases the proton conductivity and methanol permeability, compared to bare Nafion®, simultaneously. However, the conductivity increased and methanol permeability decreased as the amount of incorporated Pd increased, and above a certain amount the rapid increase of conductivity and permeability appeared. The DMFC performance at 40°C was improved by incorporating Pd.

Different approach for the incorporation of Pd in Nafion® membranes was suggested by Tang *et al.* (2005). Multi-layer self-assembly Nafion® membranes (MLSA Nafion® membranes) were prepared by alternately assembling charged Pd particles and Nafion®

found to decrease with increasing Pd thickness. For the Pd-1000 Å film, 30% reduction in conductivity was observed. Methanol permeability at 25°C decreased with increasing Pd thickness and they varied from 2.90 10-6 to 2.23 10-6 cm2 s-1 by deposition of Pd film of 1000 Å on the Nafion® 117 membrane. For the Pd-1000 Å-Nafion® 115 membrane, permeability decreased as much as 44% from 2.97 10-6 to 1.67 10-6 cm2 s-1. The cell performance at 90°C exhibited a slight decrease with the Pd layered Nafion® 115 membranes. The methanol crossover through an MEA is inversely proportional to current density and thus, its effect on the performance is more prominent at low current densities. It was found that at low current densities cell performance increased in Pd film of 1000 Å on the Nafion® 115, where the Pd film act as a barrier to methanol crossover. However, these results are different of that of Choi *et al.* who observed significant increase in DMFC

Z.Q. Ma *et al.* (2003) followed a different approach, where Nafion® membrane was modified by sputtering a thin layer of Pt/Pd-Ag/Pt on its surface. The methanol crossover can be reduced by sputtering Pd-Ag alloys over the polymer electrolyte, furthermore, when hydrogen is absorbed and dissolved in the membrane, the palladium-silver alloy membrane not only reduces the possibility of embrittling due to α phase transition at low temperatures (< 150°C) but also leads to a higher permeability for hydrogen than pure palladium membrane. The composite membranes were prepared as follows: on one side of the Nafion® 117 membrane, a 2 nm Pt film was first deposited. This was followed by a Pd-Ag film with three different thicknesses (0.1, 0.2 and 1 µm). On the top of the Pd-Ag, it was coated with another 2 nm Pt film. Before a MEA was manufactured, a 4-5 µm layer of Nafion® polymer was recast over the surface of the sputtered Pt/Pd-Ag/Pt layer with Nafion® solution. The final membrane prepared was in the form of Nafion®117/Pt/Pd-Ag/Pt/Nafion®, containing 0.0086 mg cm-2 Pt, 0.90 mg cm-2 Pd, and 0.27 mg cm-2 Ag. The cell performance and the OCV increased with increasing the sputtering alloy layer thickness and the best performance and the highest OCV were found with the 1 µm Pd-Ag film. Also it was showed that the performance with the 1 µm Pd-Ag film is higher than that of cell with a Nafion® membrane having catalyst loading twice as high. Palladinized Nafion® composite membrane was prepared via ion-exchange and chemical reduction method (Y.J. Kim *et al.*, 2004). Palladium(II) acetylacetonate and tetraammine-palladium(II) chloride hydride were used as palladium precursors. Nafion® 117 samples were immersed in palladium precursor solutions followed by chemical reduction of palladium precursors by sodium borohydride. The use of tetraamminepalladium(II) chloride hydride formed 40-50 nm of palladium particles while palladium(II) acetylacetonate formed 5-10 nm of particles. For all palladinized samples, water uptake was higher than for unfilled Nafion® whereas above a certain amount of incorporated palladium, methanol uptake was lower than for unfilled Nafion®. Incorporating Pd nanoparticles decreases the proton conductivity and methanol permeability, compared to bare Nafion®, simultaneously. However, the conductivity increased and methanol permeability decreased as the amount of incorporated Pd increased, and above a certain amount the rapid increase of conductivity and permeability appeared.

performance in Pd-20 nm film on Nafion® 117.

The DMFC performance at 40°C was improved by incorporating Pd.

Different approach for the incorporation of Pd in Nafion® membranes was suggested by Tang *et al.* (2005). Multi-layer self-assembly Nafion® membranes (MLSA Nafion® membranes) were prepared by alternately assembling charged Pd particles and Nafion®

ionomer onto Nafion® 112 membranes. The Pd particles, size of about 1.8 nm in average, are charged by PDDA (polydiallyldimethylammonium chloride) ionomers. The Pd loading of the first-layer MLSA Nafion® membranes was 0.63 µg cm-2, and the surface coverage of the Pd nanoparticles on the Nafion® membrane was estimated as 22%. After 5-double-layer Pd particles/Nafion® ionomers assembling, the Pd loading reached to 2.86 µg cm-2. The methanol crossover current of the original Nafion® membranes and 1-double-layer, 2 double-layer, 3-double-layer, 4-double-layer, 5-double-layer MLSA Nafion® membrane were 0.0495, 3.87 10-3, 1.38 10-3, 7.32 10-4, 5.16 10-4 and 4.25 10-3 A cm-2, respectively, corresponding conductivities of 0.112, 0.110, 0.105, 0.094, 0.087 and 0.081 S cm-2. No DMFC data were provided, however, it was suggested that the 3-double-layer self-assembly membrane is the best suited for DMFC application, since it has a methanol crossover decreased to 0.86%, and a conductivity remaining at 83.9% comparing to original Nafion® membrane. Electroless plating was also used to deposit Pd layer on Nafion® membranes (Hejze *et al.*, 2005; Sun *et al.*, 2005). Palladium layer can reduce methanol crossover when coated on the surface of Nafion®, i.e. limiting current from methanol permeation through a membrane electrode assembly was reduce from 64 to 57 mA cm-2 for 1M methanol, and from 267 to 170 mA cm-2 for 5M methanol, for Nafion® 115 and Pd/Nafion® 115 membranes respectively (Sun *et al.*, 2005). Also it was demonstrated that the DMFC performance increase with the incorporation of Pd in Nafion® 115. When 1M methanol was used, the power density increased from 36 to 45 mW cm-2, for Nafion® 115 and Pd/Nafion® 115, respectively. When 5M methanol was used, the maximum power density on Pd/Nafion® 115 was 72 mW cm-2, while the performance of MEA with pure Nafion® 115 membranes was only 32 mW cm-2.

#### **6.7 Organic/montmorillonite nanocomposite membranes**

Montmorillonite (MMT) is a type of layered silicate composed of silica tetrahedral and alumina octahedral sheets (J. Chang *et al.*, 2003) and its intercalation into Nafion® membrane can decrease successfully the methanol permeability and improve mechanical property (Jung *et al.*, 2003; Song *et al.*, 2004). Research by J. Chang *et al*. (2003) has showed that layered silicates incorporated into SPEEK membranes helped to reduce swelling significantly in hot water and decrease the methanol crossover without a serious reduction of the proton conductivity. Gaowen and Zhentao (2005) prepared organically modified MMT (OMMT) through ion exchange reaction between alkylammonium cations and metal cations. The nanocomposite membranes (SPEEK/OMMT) were prepared using the solution intercalation technique. The water uptake of SPEEK membrane increased rapidly above 50°C, while the SPEEK/OMMT composite membranes posses of rather constant water uptake up to 80°C. This indicates that MMT layers incorporated into SPEEK matrix prevent extreme swelling of the composite membranes due to the cohesion of the functional groups between SPEEK matrix and MMT layers. The proton conductivity of the membrane was measured at temperature ranging from 22 to 110°C. It was found that the conductivity of SPEEK/OMMT composite is lower than that of the pristine SPEEK and decreases sequentially as the content of OMMT increases, which is due to prolonging the transfer route of proton. The conductivity of SPEEK/OMMT (5 wt.%) approaches the value of Nafion® 115 at 90°C and reaches 1.2 10-2 S cm-1. The activation energies of SPEEK/OMMT are higher than that of Nafion® 115, where the value of 32.08 kJ/mol and 10.8 kJ/mol, respectively, were found. Methanol permeability was found to be in the

Organic / Inorganic Nanocomposite Membranes

incorporating these zeolites.

by incorporating these inorganic nano materials.

**7. Conclusion** 

be done, which include:

be understood;

leaching; and

humidification and no pressure; Long term fuel cell performance;

Development for Low Temperature Fuel Cell Applications 533

lower than that of pristine Nafion®. A study of using ZSM5 as a filler in Nafion® was study by Byun *et al*. (2006). These composite membranes show higher water uptake than Nafion® 115, while methanol permeability has decrease with increasing zeolite contents. The selectivity of nanocomposite membranes was higher than that of Nafion® 115. Zeolite beta was incorporated in Nafion® (Holmberg *et al*., 2008) and Chitosan (Y. Wang *et al*., 2008) membranes. Zeolite beta/ Nafion® nanocomposite membranes with loading of 2.5 and 5 wt.% posses proton conductivity/methanol permeability (selectivity) ratios as much as 93% higher than commercial Nafion® 117 at 21°C, and 63% higher at 80°C. These composite membranes outperform Nafion® 117 in DMFC (Holmberg *et al*., 2008). The incorporation of zeolite beta in Chitosan reduces the methanol permeability. Furthermore, zeolite beta was sulfonated, therefore the methanol permeability was further reduced as a result of the enhanced interfacial interaction between zeolite beta and Chitosan matrix. The cell performance of the composite membrane were comparable to Nafion® 117 at low methanol concentration (2M) and much better at high methanol concentration (12M) (Y. Wang *et al*., 2008). Other zeolite based nanocomposite membranes investigated are MCM-41 (Bello *et al*., 2008; Gomes *et al*., 2008; Karthikeyan *et al*., 2005; Marschall *et al*., 2007), zeolite Y (Ahmad *et al*., 2006), zeolite BEA (Holmberg *et al*., 2005) and Chabazite and clinoptilolite (Tricoli & Nannetti, 2003). These membranes can maintained the proton conductivity at temperature above 100°C, also methanol permeability is reduced by

High temperature operating fuel cell enhanced the performance, especially for methanol, DME and ethanol fuel cells. Incorporating nano inorganic materials in the organic matrix has several advantages, namely increase membrane thermal and mechanical stabilities, increase the working temperature and water retention. Some inorganic fillers are proton conductor which can increase the conductivity of the composite membrane or at least keep the same conductivity of the pristine organic membrane. Fuel crossover also can be reduced

From all the investigated inorganic fillers, the nano size played a major role to enhance the compatibility and the interaction of the inorganic fillers with the polymeric matrix. Also the optimal inorganic loading was found to be around 5 wt.%, with the majority around 3 wt.%. However, for commercialization of these nanocomposite membranes, more R&D needed to

The interaction between the nano-inorganic materials with the organic matrix need to

High temperature fuel cell performance needed to be done with very little or no

Long term membrane stability (thermal and mechanical) and also long term membrane

A comprehensive comparative study between all the investigated inorganic materials

Systematic study of these nanocomposite, especially inorganic loading;

and the interaction with different organic materials.

following order: Nafion® 115> SPEEK > SPEEK/OMMT. Therefore, incorporating nanosized dispersion of OMMT prevents methanol from migrating through the membrane. Song *et al.* (2004) prepared recast Nafion®/MMT nanocomposite membranes at a loading of 3 wt.% clay. It was found that the strength increased more than 35% and the tensile elongation almost doubled. The thermal decomposition behavior of Nafion®/MMT nanocomposite was similar to that of pristine Nafion®, but the major decomposition temperature of polymer main chain shifted to much higher temperature region. The proton conductivity of pristine recast Nafion® with thickness of ca. 100 µm and dry state reached about 0.08 S/cm at room temperature. For Nafion®/MMT nanocomposite membranes, the room temperature conductivity was almost similar to that of neat Nafion® below MMT loading of 2 wt. % and then decreased gradually with the increase in filler content. The methanol permeability of pristine recast Nafion® was 2.3 × 10-6 cm3 cm/ cm2 s, while for Nafion®/MMT composite membranes with a thickness of 50 µm significantly decreases to 1.6 × 10-7 cm3 cm/ cm2 s by only 1 wt.% organo clay loading, which amounted to more than 90% reduction. Jung *et al*. (2003) prepared dodecylamine-exchanged montmorllonite (m-MMT) by a cation exchange reaction. The thermal resistance of Nafion®/MMT composite was found to be lower than that of the pristine Nafion®. Also the thermal resistance of 5 and 7 wt.% MMT was lower than that of 3 wt.% MMT. On the other hand, Nafion®/MMT displayed higher thermal resistance than that of pristine Nafion®. The thermal resistance of the Nafion®/m-MMT nanocomposite was also increased slightly with increasing the contents of m-MMT in the composite membrane. The methanol permeability of pristine Nafion® was found to be 0.13 mol/l at 1 h. By adding MMT and m-MMT, the methanol permeability decreased to 0.045 and 0.042 mol/l at 1 h, respectively. The proton conductivity of Nafion®/MMT was found to be 8.9 × 10-2, 7 × 10-2, 7.2 × 10-2 and 6.7 × 10-2 S/cm at 110°C for a content of MMT of 0, 3, 5 and 7 wt.%, respectively. However, the conductivity of Nafion®/m-MMT was found to be 8.9 × 10-2, 7.72 × 10-2, 7.57 × 10-2 and 7.4 × 10-2 S/cm at 110°C for a content of m-MMT of 0, 3, 5 and 7 wt.%, respectively. In general the proton conductivity of the composite membranes decreased slightly with increasing the contents of MMT and m-MMT and lower than pristine Nafion®. Pristine Nafion® performances were 385, 410 and 138.1 mA/cm2 (at a potential of 0.4 V) at 90, 110 and 125°C, respectively. For Nafion®/3 wt.% MMT performances were 370, 452.6 and 282.86 mA/cm2 (at a potential of 0.4 V) at 90, 110 and 125°C, respectively. Finally, for Nafion®/3 wt.% m-MMT performances were 367.1, 440 and 290 mA/cm2 (at a potential of 0.4 V) at 90, 110 and 125°C, respectively. Gosalawit *et al*. (2008) prepared sulfonated montmorillonite (SMMT) with SPEEK. It was found that the inorganic aggregation in SPEEK increased with SMMT loading. The stability in water and in methanol aqueous solution as well as the mechanical stability were enhanced with SMMT loading. Whereas thermal stability improvement did not exist significantly. The methanol permeability was reduced when the SMMT loading increased. The proton conductivity was improved with the incorporation of SMMT. SMMT/SPEEK nanocomposite membranes showed significant cell performance for DMFC as compared to pristine SPEEK and Nafion® 117 membranes.

#### **6.8 Organic/zeolites nanocomposite membranes**

Mordenite crystals dispersed in poly(acrylic acid) was prepared by Rao *et al*. (1994) and Libby *et al*. (2001), while mordenite dispersed in Nafion® was prepared by Arimura *et al*. (1999). These membranes displayed proton conductivity about two orders of magnitude lower than that of pristine Nafion®. A study of using ZSM5 as a filler in Nafion® was study by Byun *et al*. (2006). These composite membranes show higher water uptake than Nafion® 115, while methanol permeability has decrease with increasing zeolite contents. The selectivity of nanocomposite membranes was higher than that of Nafion® 115. Zeolite beta was incorporated in Nafion® (Holmberg *et al*., 2008) and Chitosan (Y. Wang *et al*., 2008) membranes. Zeolite beta/ Nafion® nanocomposite membranes with loading of 2.5 and 5 wt.% posses proton conductivity/methanol permeability (selectivity) ratios as much as 93% higher than commercial Nafion® 117 at 21°C, and 63% higher at 80°C. These composite membranes outperform Nafion® 117 in DMFC (Holmberg *et al*., 2008). The incorporation of zeolite beta in Chitosan reduces the methanol permeability. Furthermore, zeolite beta was sulfonated, therefore the methanol permeability was further reduced as a result of the enhanced interfacial interaction between zeolite beta and Chitosan matrix. The cell performance of the composite membrane were comparable to Nafion® 117 at low methanol concentration (2M) and much better at high methanol concentration (12M) (Y. Wang *et al*., 2008). Other zeolite based nanocomposite membranes investigated are MCM-41 (Bello *et al*., 2008; Gomes *et al*., 2008; Karthikeyan *et al*., 2005; Marschall *et al*., 2007), zeolite Y (Ahmad *et al*., 2006), zeolite BEA (Holmberg *et al*., 2005) and Chabazite and clinoptilolite (Tricoli & Nannetti, 2003). These membranes can maintained the proton conductivity at temperature above 100°C, also methanol permeability is reduced by incorporating these zeolites.

#### **7. Conclusion**

532 Advances in Chemical Engineering

following order: Nafion® 115> SPEEK > SPEEK/OMMT. Therefore, incorporating nanosized dispersion of OMMT prevents methanol from migrating through the membrane. Song *et al.* (2004) prepared recast Nafion®/MMT nanocomposite membranes at a loading of 3 wt.% clay. It was found that the strength increased more than 35% and the tensile elongation almost doubled. The thermal decomposition behavior of Nafion®/MMT nanocomposite was similar to that of pristine Nafion®, but the major decomposition temperature of polymer main chain shifted to much higher temperature region. The proton conductivity of pristine recast Nafion® with thickness of ca. 100 µm and dry state reached about 0.08 S/cm at room temperature. For Nafion®/MMT nanocomposite membranes, the room temperature conductivity was almost similar to that of neat Nafion® below MMT loading of 2 wt. % and then decreased gradually with the increase in filler content. The methanol permeability of pristine recast Nafion® was 2.3 × 10-6 cm3 cm/ cm2 s, while for Nafion®/MMT composite membranes with a thickness of 50 µm significantly decreases to 1.6 × 10-7 cm3 cm/ cm2 s by only 1 wt.% organo clay loading, which amounted to more than 90% reduction. Jung *et al*. (2003) prepared dodecylamine-exchanged montmorllonite (m-MMT) by a cation exchange reaction. The thermal resistance of Nafion®/MMT composite was found to be lower than that of the pristine Nafion®. Also the thermal resistance of 5 and 7 wt.% MMT was lower than that of 3 wt.% MMT. On the other hand, Nafion®/MMT displayed higher thermal resistance than that of pristine Nafion®. The thermal resistance of the Nafion®/m-MMT nanocomposite was also increased slightly with increasing the contents of m-MMT in the composite membrane. The methanol permeability of pristine Nafion® was found to be 0.13 mol/l at 1 h. By adding MMT and m-MMT, the methanol permeability decreased to 0.045 and 0.042 mol/l at 1 h, respectively. The proton conductivity of Nafion®/MMT was found to be 8.9 × 10-2, 7 × 10-2, 7.2 × 10-2 and 6.7 × 10-2 S/cm at 110°C for a content of MMT of 0, 3, 5 and 7 wt.%, respectively. However, the conductivity of Nafion®/m-MMT was found to be 8.9 × 10-2, 7.72 × 10-2, 7.57 × 10-2 and 7.4 × 10-2 S/cm at 110°C for a content of m-MMT of 0, 3, 5 and 7 wt.%, respectively. In general the proton conductivity of the composite membranes decreased slightly with increasing the contents of MMT and m-MMT and lower than pristine Nafion®. Pristine Nafion® performances were 385, 410 and 138.1 mA/cm2 (at a potential of 0.4 V) at 90, 110 and 125°C, respectively. For Nafion®/3 wt.% MMT performances were 370, 452.6 and 282.86 mA/cm2 (at a potential of 0.4 V) at 90, 110 and 125°C, respectively. Finally, for Nafion®/3 wt.% m-MMT performances were 367.1, 440 and 290 mA/cm2 (at a potential of 0.4 V) at 90, 110 and 125°C, respectively. Gosalawit *et al*. (2008) prepared sulfonated montmorillonite (SMMT) with SPEEK. It was found that the inorganic aggregation in SPEEK increased with SMMT loading. The stability in water and in methanol aqueous solution as well as the mechanical stability were enhanced with SMMT loading. Whereas thermal stability improvement did not exist significantly. The methanol permeability was reduced when the SMMT loading increased. The proton conductivity was improved with the incorporation of SMMT. SMMT/SPEEK nanocomposite membranes showed significant cell performance for

DMFC as compared to pristine SPEEK and Nafion® 117 membranes.

Mordenite crystals dispersed in poly(acrylic acid) was prepared by Rao *et al*. (1994) and Libby *et al*. (2001), while mordenite dispersed in Nafion® was prepared by Arimura *et al*. (1999). These membranes displayed proton conductivity about two orders of magnitude

**6.8 Organic/zeolites nanocomposite membranes** 

High temperature operating fuel cell enhanced the performance, especially for methanol, DME and ethanol fuel cells. Incorporating nano inorganic materials in the organic matrix has several advantages, namely increase membrane thermal and mechanical stabilities, increase the working temperature and water retention. Some inorganic fillers are proton conductor which can increase the conductivity of the composite membrane or at least keep the same conductivity of the pristine organic membrane. Fuel crossover also can be reduced by incorporating these inorganic nano materials.

From all the investigated inorganic fillers, the nano size played a major role to enhance the compatibility and the interaction of the inorganic fillers with the polymeric matrix. Also the optimal inorganic loading was found to be around 5 wt.%, with the majority around 3 wt.%.

However, for commercialization of these nanocomposite membranes, more R&D needed to be done, which include:


Organic / Inorganic Nanocomposite Membranes

Bauer, B.; Menzel, Th. & Kehl, P. (1994). *EP* 0 645 175 A1

Belyakov, V.N. & Linkov, V.M. (1999). *US Patent* 5,932,361

Breen, J.P. & Ross, J.R.H. (1999). *Catalysis Today,* 51, pp. 521-533

Varzi, A. (2008). *Fuel Cells,* 8, pp. 217-224 Chang, H.Y. & Lin, C.W. (2003). *J. Membr. Sci.,* 218, pp. 295-306

Chu, D. & Jiang, R. (1999). *J. Power Sources,* 80, pp. 226-234

*Mater. Electrochem. Systems,* 3, pp. 33-42

Cubiero, M.L. & Fierro, J.L.G. (1998). *J. Catalysis*, 179, pp. 150-162

Cui, W.; Kerres, J. & Eigenberger, G. (1998). *Sep. Purif. Technol.,* 14, pp. 145-154

Deng, Q.; Moore, R.B. & Mauritz, K.A. (1998). *J. Appl. Polym. Sci.,* 68, pp. 747-763

Carrette, L.; Friedrich, K.A. & Stimming, U. (2001). *Fuel Cells,* 1, pp. 5-39

Choi, W.C.; Kim, J.D. & Woo, S.I. (2001). *J. Power Sources,* 96, pp. 411-414 Chu, D. & Gilman, S. (1994). *J. Electrochem. Soc.*, 141, pp. 1770-1773

pp. 197-203

Boffito, G. (1999). *US Patent* 5,976,723

177, pp. 3233-3243

1023-1033

D`Alelio, G. (1944). *US Patent* 2,340,111

*Sci.,* 296, pp. 156-161

*Ionics,* 176, pp. 2985-2989

Bauer, B.; Rafler, G. & Ulrich, H.-H. (1995). *Ger. Patent* DE 195 38 025

Bauer, F. & Willert-Porada, M. (2004). *J. Membr. Sci.,* 233, pp. 141-149 Bauer, F. & Willert-Porada, M. (2005). *J. Power Sources,* 145, pp. 101-107 Bauer, F. & Willert-Porada, M. (2006a). *Fuel Cells,* 6, pp. 261-269

Bauer, F. & Willert-Porada, M. (2006b). *Solid State Ionics,* 177, pp. 2391-2396 Bello, M.; Javaid Zaidi, S.M. & Rahman, S.U. (2008). *J. Membr. Sci.,* 322, pp. 218-224

Bernardi, D.M. & Verbrugge, M.W. (1991). *AIChE Journal,* 37, pp. 1151-1163

Development for Low Temperature Fuel Cell Applications 535

Baradie, B.; Poinsignon, C.; Sanchez, J.Y.; Piffard, Y.; Vitter, G.; Bestaoui, N.; Foscallo, D.; Denoyelle, A.; Delabouglise, D. & Vaujany, M. (1998). *J. Power Sources,* 74, pp. 8-16

Bauer, B.; Jones, D.J.; Rozière, J.; Tchicaya, L.; Alberti, G.; Casciola, M.; Massinelli, L.; Peraio, A.; Besse, S. & Ramunni, E. (2000). *J. New. Mater. Electrochem. Systems,* 3, pp. 93-98

Bi, C.; Zhang, H.; Zhang, Y.; Zhu, X.; Ma, Y.; Dai, H. & Xiao, S. (2008). *J. Power Sources,* 184,

Bonnet, B.; Jones, D.J.; Rozière, J.; Tchicaya, L.; Alberti, G.; Casciola, M.; Massinelli, L.; Bauer, D.; Peraio, A. & Ramunni, E. (2000). *J. New Mater. Electrochem. Systems,* 3, pp. 87-92

Brack, H.P.; Wyler, M.; Peter, G. & Scherer, G.G. (2003). *J. Membr. Sci.,* 214, pp. 1-19

Büchi, F.N.; Gupta, B.; Haas, O. & Scherer, G.G. (1995a). *Electrochim. Acta,* 40, pp. 345-353 Büchi, F.N.; Gupta, B.; Haas, O. & Scherer, G.G. (1995b). *J. Electrochem. Soc*., 142, pp. 3044-3048 Byun, S.C.; Jeung, Y.J.; Park, J.W.; Kim, S.D.; Ha, H.Y. & Kim, W.J. (2006). *Solid State Ionics,*

Casiola, M.; Alberti, G.; Ciarletta, A.; Cruccolini, A.; Piaggio, P. & Pica, M. (2005). *Solid State* 

Casiola, M.; Alberti, G.; Sganappa, M. & Narducci, R. (2006). *J. Power Sources,* 162, pp. 141-145 Casiola, M.; Capitani, D.; Comite, A.; Donnadio, A.; Frittella, V.; Pica, M.; Sganappa, M. &

Chang, J.; Park, J.H.; Park, G.; Kim, C. & Park, O. (2003). *J. Power Sources,* 124, pp. 18-25

Colbow, K.M.; Zhang, J. & Wilkinson, D.P. (2000). *J. Electrochem. Soc*., 147, pp. 4058-4060 Colicchio, I.; Wen, F.; Keul, H.; Simon, U. & Moeller, M. (2009). *J. Membr. Sci.,* 326, pp. 45-57 Cornet, N.; Diat, O.; Gebel, G.; Jousse, F.; Marsaq, D.; Mercier, R. & Peneri, M. (2000). *J. New* 

Costamagna, P.; Yang, C.; Bocarsly, A.B. & Srinivasan, S. (2002). *Electrochim. Acta,* 47, pp.

Di Vona, L.M.; Ahmed, Z.; Bellitto, S.; Lenci, A.; Traversa, E. & Licoccia, S. (2007). *J. Membr.* 

#### **8. References**

Adamson, K.-A. & Pearson, P. (2000). *J. Power Sources,* 86, pp. 548-555


Adjemian, K.T.; Lee, S.J.; Srinivasan, S.; Benziger, J. & Bocarsly, A.B. (2002a). *J. Electrochem.* 

Adjemian, K.T.; Srinivasan, S.; Benziger, J. & Bocarsly, A.B. (2002b). *J. Power Sources,* 109, pp.

Agrell, J.; Hsselbo, K.; Jansson, K.; Jaras, S.G. & Boutonnet, M. (2001). *Appl. Catal. A,* 211, pp.

Alberti, G.; Casiola, M.; Costantino, U.; Levi, G. & Ricciard, G. (1978). *J. Inorg. Nucl. Chem.,*

Alberti, G.; Casiola, M.; Capitani, D.; Donnadio, A.; Narducci, R.; Pica, M. & Sganappa, M.

Allcock, H.R.; Hofmann, M.A.; Ambler, C.M. & Morford, R.V. (2002a). *Macromolecules,* 35,

Allcock, H.R.; Hofmann, M.A.; Ambler, C.M.; Lvov, S.N.; Zhou, X.Y.; Chalkova, E. &

Amphlett, J.C.; Evans, M.J.; Mann, R.F. & Weir, R.D. (1985). *Can. J. Chem. Eng.,* 63, pp. 605-611 Amphlett, J.C.; Peppley, B.A.; Halliop, E. & Sadiq, A. (2001). *J. Power Sources,* 96, pp. 204-213

Antonucci, P.L.; Aricò, A.S.; Cretì, P.; Ramunni, E. & Antonucci, V. (1999). *Solid State Ionics,*

Appleby, A.J. & Yeager, E.B. (1986a). In: *Assessment of Research Needs for Advanced Fuel Cells*,

Aricò, A.S.; Baglio, V.; Di Blasi, A. & Antonucci, V. (2003a). *Electrochem. Communications*, 5,

Aricò, A.S.; Baglio, V.; Di Blasi, A.; Cretì, P.; Antonucci, P.L. & Antonucci, V. (2003b). *Solid* 

Aricò, A.S.; Baglio, V.; Di Blasi, A.; Modica, E.; Antonucci, P.L. & Antonucci, V. (2004). *J.* 

Baglio, V.; Arico, A.S.; Di Blasi, A.; Antonucci, V.; Antonucci, P.L.; Licoccia, S.; Traversa, E.

Bailly, C.; Williams, D.J.; Karasz, F.E. & MacKnight, W.J. (1987). *Polymer,* 28, pp. 1009-1016

Arimura, T.; Ostrovskii, D.; Okada, T. & Xie, G. (1999). *Solid State Ionics,* 118, pp. 1-10

& Serraino Fiory, F. (2005). *Electrochim. Acta,* 50, pp. 1241-1246

Baradie, B.; Dodelet, J.P. & Guay, D. (2000). *J. Electroanal. Chem.,* 489, pp. 101-105

Argyropoulos, A.; Scott, K. & Taama, W.M. (1999a). *J. Appl. Electrochem*., 29, pp. 661-669 Argyropoulos, P.; Scott, K. & Taama, W.M. (1999b). *Electrochim. Acta,* 44, pp. 3575-3584 Aricò, A.S.; Cretì, P.; Antonucci, P.L. & Antonucci, V. (1998). *Electrochem. Solid State Lett.,* 1,

Alberti, G.; Casciola, M.; Massinelli, L. & Bauer, B. (2001). *J. Membr. Sci.,* 185, pp. 73-81 Alberti, G.; Casciola, M.; D`Alessandro, E. & Pica, M. (2004). *J. Mater. Chem.,* 14, pp. 1910-1914 Alberti, G.; Casciola, M.; Donnadio, A.; Piaggio, P.; Pica, M. & Sisani, M. (2005a). *Solid State* 

Aharoni, S.M. & Litt, M.H. (1974). *J. Polym. Sci., Polym. Chem. Ed.,* 12, pp. 639-650 Ahmad, M.I.; Zaidi, S.M.J. & Rahman, S.U. (2006). *Desalination,* 193, pp. 387-397

Alberti, G.; Pica, M. & Tarpanelli, T. (2005b). PCT Patent WO 2005/105667 A1

Anantaraman, A.V. & Gardner, C.L. (1996). *J. Electroanal. Chem.,* 414, pp. 115-120

(2007). *Electrochim. Acta,* 52, pp. 8125-8132

Weston, J. (2002b). *J. Membr. Sci.,* 201, pp. 47-54

Penner, S.S. (Ed.), Chapter 4, Pergamon, New York

Appleby, A.J. & Yeager, E.B. (1986b). *Energy,* 11, pp. 137-151

Adamson, K.-A. & Pearson, P. (2000). *J. Power Sources,* 86, pp. 548-555

**8. References** 

356-364

239-250

40, pp. 533-537

pp. 3484-3489

125, pp. 431-437

pp. 66-68

pp. 882-866

*State Ionics,* 161, pp. 251-265

*Power Sources,* 128, pp. 113-118

*Soc.,* 149, pp. A256-A261

*Ionics,* 176, pp. 2893-2898


Organic / Inorganic Nanocomposite Membranes

Chapter 4, Wiley-VCH, Weinheim

Gupta, B. & Scherer, G.G. (1994). *Chimia,* 48, pp. 127-137

Hara, S. & Miyayama, M. (2004). *Solid State Ionics,* 168, pp. 111-116

Hoel, D. & Grunwald, E. (1977). *J. Phys. Chem.,* 81, pp. 2135-2136

Hubner, G. & Roduner, E. (1999). *J. Mater. Chem.,* 9, pp. 409-418 Ioannides, T. & Neophytides, S. (2000). *J. Power Sources,* 91, pp. 150-156

Hogarth, M.P. & Ralph, T.R. (2002). *Platinum Metals Rev.,* 46, pp. 146-164

Holmberg, B.A.; Wang, X. & Yan, Y. (2008). *J. Membr. Sci.,* 320, pp. 86-92

& Riedel, E. (1996). *J. Power Sources,* 61, pp. 143-147

*Chem. Ing. Tech.,* 44, pp. 167-169 Grot, W.G. & Rajendran, G. (1999). *US Patent* 5,919,583

(2002). *Fuel,* 81, pp. 2173-2176

*Power Sources,* 140, pp. 21-27

*Energy,* 33, pp. 5686-5690

(1984). *J. Polym. Sci.,* 22, pp. 721-737 Jones, D.J. & Rozière, J. (2001). *J. Membr. Sci.*, 185, pp. 41-58

England

52-58

England

Development for Low Temperature Fuel Cell Applications 537

Gottesfeld, S. & Zawodzinski, T.A. (1997). In: *Advances in Electrochemical Science and* 

Grot, W. (1978). *Chem. Ing. Tech.,* 50, pp. 299-301; (1975). *Chem. Ing. Tech.,* 47, pp. 617; (1972).

Hacker, V. & Kordesch, K. (2003). In: *Handbook of Fuel Cells,* Vielstich, W.; Lamm, A. &

Hatanaka, T.; Hasegawa, N.; Kamiya, A.; Kawasumi, M.; Morimoto, Y. & Kawahara, K.

Hejze, T.; Gollas, B.R.; Sauerbrey, R.K.; Schmied, M.; Hofer, F. & Besenhard, J.O. (2005). *J.* 

Helen, M.; Viswanathan, B. & Srinivasa Murthy, S. (2006). *J. Power Sources*, 163, pp. 433-439 Helen, M.; Viswanathan, B. & Srinivasa Murthy, S. (2007). *J. Membr. Sci.,* 292, pp. 98-105 Hill, M.L.; Kim, Y.S.; Einsla, B.R. & McGrath, J.E. (2006). *J. Membr. Sci.,* 283, pp. 102-108

Hogarth, M.; Christensen, P.; Hamnett, A. & Shukla, A. (1997). *J. Power Sources,* 69, pp. 113-124

Höhlein, B.; Boe, M.; Bøgild-Hansen, J.; Bröckerhoff, P.; Colsman, G.; Emonts, B.; Menzer, R.

Holmberg, B.A.; Hwang, S.J.; Davis, M.E. & Yan, Y. (2005). *Microp. Mesop. Mat.,* 80, pp. 347-356

Honma, I.; Nishikawa, O.; Sugimoto, T.; Nomura, S. & Nakajima, H. (2002). *Fuel Cells,* 2, pp.

Hou, H.; Sun, G.; Wu, Z.; Jin, W. & Xin, Q. (2008). *Int. J. Hydrogen Energy,* 33, pp. 3402-3409

Jiang, C.J.; Trimm, D.L.; Wainwright, M.S. & Cant, N.W. (1993). *Appl. Cat. A*, 93, pp. 245-255

Jian-hua, T.; Peng-fei, G.; Zhi-yuan, Z.; Wen-hui, L. & Zhong-qiang, S. (2008). *Int. J. Hydrogen* 

Johnson, B.C.; Yilgor, I.; Tran, C.; Iqbal, M.; Wightman, J.P.; Llyod, D.R. & McGrath, J.E.

Jones, D.J. & Rozière, J. (2001). In: *Handbook of Fuel Cells,* Vielstich, W.; Lamm, A. &

Jörissen, L.; Gogel, V.; Kerres, J. & Garche, J. (2002). *J. Power Sources,* 105, pp. 267-273 Jung, D.H.; Cho, S.Y.; Peck, D.H.; Shin, D.R. & Kim, J.S. (2002). *J. Power Sources,* 106, pp. 173-177 Jung, D.H.; Cho, S.Y.; Peck, D.H.; Shin, D.R. & Kim, J.S. (2003). *J. Power Sources,* 118, pp. 205-211

Gasteiger, H.A. (Eds.), Vol 3, Part 1, Chapter 35, pp. 447-455, Wiley, Chichester,

Jiang, R.; Kunz, H.R. & Fenton, J.M. (2006a). *Electrochim. Acta,* 51, pp. 5596-5605 Jiang, R.C.; Kunz, H.R. & Fenton, J.M. (2006b). *J. Membr. Sci.,* 272, pp. 116-124

Gasteiger, H.A. (Eds.), Vol 3, Part 1, Chapter 10, pp. 121-127, Wiley, Chichester,

Guo, Q.; Pintauro, P.N.; Tang, H. & O'Connor, S. (1999). *J. Membr. Sci.,* 154, pp. 175-181 Gupta, B.; Büchi, F.N.; Scherer, G.G. & Chapiró, A. (1993). *Solid State Ionics,* 61, pp. 213-218 Gupta, B.; Büchi, F.N. & Scherer, G.G. (1994). *J. Polym. Sci. A: Polym. Chem*., 32, pp. 1931-1938

*Enginering*, Alkire, R.C.; Gerischer, H.; Kolb, D.M. & Tobias, C.W. (Eds.), Vol. 5,


Dimitrova, P.; Friedrich, K.A.; Stimming, U. & Vogt, B. (2002a). *Solid State Ionics,* 150, pp.

Dimitrova, P.; Friedrich, K.A.; Vogt, B. & Stimming, U. (2002b). *J. Electroanal. Chem.,* 532, pp.

Dohle, H.; Divisek, J.; Mergel, J.; Oetjen, H.F.; Zingler, C. & Stolten, D. (2002). *J. Power* 

Duesterwald, H.G.; Höhlein, B.; Kraut, H.; Meusinger, J.; Peters, R. & Stimming, U. (1997).

Edwards, N.; Ellis, S.R.; Frost, J.C.; Golunski, S.E.; van Keulen, A.N.J.; Lindewald, N.G. &

Emonts, B.; Hansen, J.B.; Jørgensen, S.L.; Höhlein, B. & Peters, R. (1998). *J. Power Sources,* 71,

Fan, J.; Guo, X.; Harada, S.; Watari, T.; Tanaka, K.; Kita, H. & Okamoto, K. (2002).

Faungnawakij, K.; Shimoda, N.; Viriya-empikul, N.; Kikuchi, R. & Eguchi, K. (2010). *Appl.* 

Faure, S.; Cornet, N.; Gebel, G.; Mercier, R.; Pineri, M. & Sillion, B. (1997). In: *Proceeding of the* 

Fontanella, J.J.; Wintersgill, M.C.; Chen, R.S.; Wu, Y. & Greenbaum, S.G. (1995). *Electrochim.* 

Gavach, C.; Pamboutzoglou, G.; Nedyalkov, M. & Pourcelly, G. (1989). *J. Membr. Sci.,* 45, pp.

Genies, C.; Mercier, R.; Sillion, B.; Cornet, N.; Gebel, G. & Pineri, M. (2001). *Polymer,* 42, pp.

Giordano, N.; Staiti, P.; Hocevar, S. & Arico, A.S. (1996). *Electrochim. Acta,* 41, pp. 397-403

Glipa, X.; El Haddad, M.; Jones, D.J. & Rozière, J. (1997). *Solid State Ionics,* 97, pp. 323-331 Gloaguen, F.; Convert, P.; Gamburzev, S.; Velev, O.A. & Srinivasan, S. (1998). *Electrochim.* 

Gomes, D.; Marschall, R.; Nunes, S.P. & Wark, M. (2008). *J. Membr. Sci.,* 322, pp. 406-415 Gong, X.; Bandis, A.; Tao, A.; Meresi, G.; Wang, Y.; Inglefield, P.T.; Jones, A.A. & Wen, W.-Y.

Goodstein, E.S. (1999). *Economics and Environment*, Upper Saddle River, NJ, Prentice Hall Gosalawit, R.; Chirachanchai, S.; Shishatskiy, S. & Nunes, S.P. (2008). *J. Membr. Sci.,* 323, pp.

Sundholm, G. & Walsby, N. (2003). *Fuel Cells,* 3, pp. 21-27

Gode, P.; Ihonen, J.; Strandroth, A.; Ericson, H.; Lindbergh, G.; Paronen, M.; Sundholm, F.;

Fujiwara, N.; Friedrich, K.A. & Stimming, U. (1999). *J. Electroanal. Chem.,* 472, pp. 120-125

*Second International Symposium on New Materials for Fuel Cell and Modern Battery Systems*, Savadogo, O. & Roberge, P.R. (Eds.), pp. 818, Montréal, Canada, July 6-10,

Faure, S.; Mercier, R.; Aldebert, P.; Pineri, M. & Sillion, B. (1996). *Fr. Patent* 96 05707

Ferrell, J.R.; Kuo, M.C. & Herring, A.M. (2010). *J. Power Sources,* 195, pp. 39-45 Finsterwalder, F. & Hambitzer, G. (2001). *J. Membr. Sci.,* 185, pp. 105-124

Gaowen, Z. & Zhentao, Z. (2005). *J. Membr. Sci.,* 261, pp. 107-113

Gebel, G.; Aldebert, P. & Pineri, M. (1993). *Polymer,* 34, pp. 333-339

Glazebrook, R.W. (1982). *J. Power Sources,* 7, pp. 215-256

(2001). *Polymer,* 42, pp. 6485-6492

115-122

pp. 288-293

1997

37-53

359-373

337-346

*Cat. B,* 97, pp. 21-27

*Acta,* 40, pp. 2321-2326

*Acta,* 43, pp. 3767-3772

*Sources,* 105, pp. 274-282

*Chem. Eng. Technol.*, 20, pp. 617-623

*Macromolecules,* 35, pp. 9022-9028

Reinkingh, J.G. (1998). *J. Power Sources,* 71, pp. 123-128

Faungnawakij, K. & Viriya-empikul, N. (2010). *Appl. Cat. A*, 382, pp. 21-27

75-83


Organic / Inorganic Nanocomposite Membranes

*Energy,* 23, pp. 525-529

pp. 207-211

1257

pp. 47-51

pp. 103-111

183-189

Calif., Jan. 9-12, 1996

*Economy*, Wiley-VCH, Weinheim

*A,* 301, pp. 66-74

215-225

134-145

Development for Low Temperature Fuel Cell Applications 539

Linkous, C.A.; Anderson, H.R.; Kopitzke, R.W. & Nelson, G.L. (1998). *Int. J. Hydrogen* 

Liu, Z.; Guo, B.; Huang, J.; Hong, L.; Han, M. & Ming Gan, L. (2006). *J. Power Sources,* 157,

Lufrano, F.; Squadrito, G.; Patti, A. & Passalacqua, E. (2000). *J. Appl. Polym. Sci.,* 77, pp. 1250-

Lufrano, F.; Gatto, I.; Staiti, P.; Antonucci, V. & Passalacqua, E. (2001). *Solid State Ionics,* 145,

Ma, L.; Jiang, C.; Adesina, A.A.; Trimm, D.L. & Wainwright, M.S. (1996). *Chem. Eng. J.,* 62,

Ma, Y.-L.; Wainright, J.S.; Litt, M.H. & Savinell, R.F. (2004). *J. Electrochem. Soc.,* 151, pp. A8-A16

Mench, M.; Boslet, S.; Thynell, S.; Scott, J. & Wang, C.Y. (2001). In: *Direct Methanol Fuel Cells*,

Mitov, S.; Vogel, B.; Roduner, E.; Zhang, H.; Zhu, X.; Gogel, V.; Jorissen, L.; Hein, M.; Xing,

Narayanan, S.R.; Kindler, A.; Jeffries-Nakamura, B.; Chun, W.; Frank, H.; Smart, M.; Valdez,

Nishiguchi, T.; Oka, K.; Matsumoto, T.; Kanai, H.; Utani, K. & Imamura, S. (2006). *Appl. Cat.*

Olah, G.A.; Goeppert, A. & Surya Prakash, G.K. (2009). *Beyond oil and Gas: The Methanol* 

Parthasarathy, A.; Martin, C.R. & Srinivasan, S. (1991). *J. Electrochem. Soc.,* 138, pp. 916-920 Paulus, U.A.; Schmidt, T.J.; Gasteiger, H.A. & Behm, R.J. (2001). *J. Electroanal. Chem.,* 495, pp.

T.I.; Surampudi, S.; Halpert, G.; Kosek, J. & Cropley, C. (1996). In: *Proceedings of 11th Annual Battery Conference on Applications and Advances*, pp. 113-122, Long Beach,

Mizsey, P.; Newson, E.; Truong, T.-B. & Hottinger, P. (2001). *Appl. Cat. A,* 213, pp. 233-237 Mizutani, I.; Liu, Y.; Mitsushima, S.; Ota, K.I. & Kamiya, N. (2006). *J. Power Sources,* 156, pp.

Mench, M.M.; Chang, H.M. & Wang, C.Y. (2004). *J. Electrochem. Soc*., 151, pp. A144-A150

Narayanan, S.; Zawodzinski, T. & Gottesfeld, S. (Eds.), PV 2001-4, pp. 241, *The* 

Marschall, R.; Bannat, I.; Caro, J. & Wark, M. (2007). *Microp. Mesop. Mat.,* 99, pp. 190-196 Mauritz, K.A.; Stefanithis, I.D.; Davis, S.V.; Scheetz, R.W.; Pope, R.K.; Wilkes, G.L. & Huang,

Livingston, D.I.; Kamath, P.M. & Corley, R.S. (1956). *J. Polym. Sci.,* 20, pp. 485-490

Ma, Z.Q.; Cheng, P. & Zhao, T.S. (2003). *J. Membr. Sci.,* 215, pp. 327-336

*Electrochemical Society Proceeding Series*, Pennington, NJ

Mench, M.W. & Wang, C.Y. (2003). *J. Electrochem. Soc*., 150, pp. A79-A85 Mokrani, T. & Scurrell, M. (2009). *Cat. Rev., Sci. Eng.*, 51, pp. 1-145

Nilsson, M.; Jozsa, P. & Pettersson, L.J. (2007). *Appl. Cat.* B, 76, pp. 42-50

Nilsson, M.; Jansson, K.; Jozsa, P. & Pettersson, L.J. (2009). *Appl. Cat. B*, 86, pp. 18-26

Nolte, R.; Ledjeff, K.; Bauer, M. & Mulhaupt, R. (1993). *J. Membr. Sci.,* 83, pp. 211-220 Nordlund, J.; Roessler, A. & Lindbergh, G. (2002). *J. Appl. Electrochem*., 32, pp. 259-265 Nunes, S.P.; Ruffmann, B.; Rikowski, E.; Vetter, S. & Richau, K. (2002). *J. Membr. Sci.,* 203, pp.

Papapolymerou, G. & Bontozoglou, V. (1997). *J. Mol. Catal. A,* 120, pp. 165-171

D.; Schonberger, F. & Kerres, J. (2006). *Fuel Cells,* 6, pp. 413-424

H.-H. (1995). *J. Appl. Polym. Sci.,* 55, pp. 181-190

Mauritz, K.A. (1998). *Mater. Sci. Eng. C*, 6, pp. 121-133


Karthikeyan, C.S.; Nunes, S.P.; Prado, L.A.S.A.; Ponce, M.L.; Silva, H.; Ruffmann, B &

Kerangueven, G.; Coutanceau, C.; Sibert, E.; Hahn, F.; Leger, J.M. & Lamy, C. (2006). *J. Appl.* 

Kerres, J.; Cui, W. & Reichle, S. (1996). *J. Polym. Sci.: Part A: Polym. Chem.*, 34, pp. 2421-2438 Kerres, J.; Cui, W. & Schnurnberger, W. (1997). *D. Patent* 19622337.7; (1997). *Fr. Patent* F.

Kerres, J.; Zhang, W. & Cui, W. (1998c). *J. Polym. Sci.: Part A: Polym. Chem.,* 36, pp. 1441-1448 Kerres, J.; Ullrich, A.; Meier, F. & Häring, T. (1999). *Solid State Ionics,* 125, pp. 243-249

Kerres, J.; Ullrich, A.; Häring, T.; Baldauf, M.; Gebhardt, U. & Preidel, W. (2000). *J. New* 

Kerres, J. ; Cui, W.; Disson, R. & Neubrand, W. (1998a). *J. Membr. Sci.*, 139, pp. 211-225

Kim, D.S.; Park, H.B.; Rhim, J.W. & Lee, Y.M. (2004). *J. Membr. Sci.,* 240, pp. 37-48

Kim, Y.J.; Choi, W.C.; Woo, S.I. & Hong, W.H. (2004). *Electrochim. Acta,* 49, pp. 3227-3234 Kim, Y.S.; Dong, L.; Hickner, M.A.; Pivovar, B.S. & McGrath, J.E. (2003). *Polymer,* 44, pp.

Kim, Y.T.; Song, M.K.; Kim, K.H.; Park, S.B.; Min, S.K. & Rhee, H.W. (2004). *Electrochim.* 

Kobayashi, T.; Rikukawa, M.; Sanui, K. & Ogata, N. (1998). *Solid State Ionics,* 106, pp. 219-225 Kordesch, K. & Simader, G. (1996). In: *Fuel cells and their Applications*, VCH, Weinheim

Krishnan, P.; Park, J.S.; Yang, T.H.; Lee, W.Y. & Kim, C.S. (2006). *J. Power Sources,* 163, pp. 2-8 Kuan, H.C.; Wu, C.S.; Chen, C.Y.; Yu, Z.Z.; Dasari, A. & Mai, Y.W. (2006). *Electrochem. Solid* 

Ledjeff-Hey, K.; Formanski, V.; Kalk, Th. & Roes, J. (1998). *J. Power Sources,* 71, pp. 199-207

Lee, W.; Shibasaki, A.; Saito, K.; Sugita, K.; Okuyama, K. & Sugo, T. (1996). *J. Electrochem.* 

Lehtinen, T.; Sundholm, G.; Holmberg, S.; Sundholm, F.; Björnbom, P. & Bursell, M. (1998).

Li, C.; Sun, G.; Ren, S.; Liu, J.; Wang, Q.; Wu, Z.; Sun, H. & Jin, W. (2006). *J. Membr. Sci.,* 272,

Libby, B.; Smyrl, W.H. & Cussler, E.L. (2001). *Electrochem. Solid State Lett.,* 4, pp. A197-A199

Lin, Y.F.; Yen, C.Y.; Ma, C.C.M.; Liao, S.H.; Lee, C.H.; Hsiao, Y.H. & Lin, H.P. (2007). *J. Power* 

Lamy, C.; Belgsir, E.M. & Léger, J.-M. (2001). *J. Appl. Electrochem*., 31, pp. 799-809 Larminie, J. & Dicks, A. (2000). In: *Fuel Cell Systems Explained,* Wiley, Chichester

Lee, H.K.; Kim, J.I.; Park, J.H. & Lee, T.H. (2004). *Electrochim. Acta,* 50, pp. 761-768

Liang, Z.X.; Zhao, T.S. & Prabhuram, J. (2006). *J. Membr. Sci.,* 283, pp. 219-224

Lin, C.W.; Thangamuthu, R. & Chang, P.H. (2005). *J. Membr. Sci.,* 254, pp. 197-205

Kerres, J.; Cui, W. & Junginger, M. (1998b). *J. Membr. Sci.,* 139, pp. 227-241

Kim, H.J.; Shul, Y.G. & Han, H. (2006). *J. Power Sources,* 158, pp. 137-142

Kordesch, K.V. & Simader, G.R. (1995). *Chem. Rev*., 95, pp. 191-207

Ledesma, C. & Llorca, J. (2009). *Chem. Eng. J.,* 154, pp. 281-286

Lim, C. & Wang, C.Y. (2003). *J. Power Sources,* 113, pp. 145-150

Léger, J.-M. (2001). *J. Appl. Electrochem.,* 31, pp. 767-771

*Electrochim. Acta,* 43, pp. 1881-1890

Schulte, K. (2005). *J. Membr. Sci.,* 254, pp. 139-146

*Electrochem*., 36, pp. 441-448

Kerres, J.A. (2001). *J. Membr. Sci.,* 185, pp. 3-27

5729-5736

*Acta,* 50, pp. 645-648

Kreuer, K.D. (1997). *Solid State Ionics*, 97, pp. 1-15 Kreuer, K.D. (2001). *J. Membr. Sci.,* 185, pp. 29-39

*State Lett.,* 9, pp. A76-A79

*Soc*., 143, pp. 2795-2798

*Sources,* 171, pp. 388-395

pp. 50-57

9706706; (1997). *US Patent* 08/868943

*Mater. Electrochem. Systems,* 3, pp. 229-239


Ma, Z.Q.; Cheng, P. & Zhao, T.S. (2003). *J. Membr. Sci.,* 215, pp. 327-336


Organic / Inorganic Nanocomposite Membranes

*Membr. Sci.,* 296, pp. 21-28

*Membr. Sci.,* 259, pp. 27-33

*Eng. Aspects,* 262, pp. 65-70

Development for Low Temperature Fuel Cell Applications 541

Steck, A.E. (1995). Membrane materials in fuel cells, In: *Proceedings of the First International* 

Su, Y.H.; Liu, Y.L.; Sun, Y.M.; Lai, J.Y.; Wang, D.M.; Gao, Y.; Liu, B. & Guiver, M.D. (2007). *J.* 

Sun, H.; Sun, G.; Wang, S.; Liu, J.; Zhao, X.; Wang, G.; Xu, H.; Hou, S. & Xin, Q. (2005). *J.* 

Sundholm, F. (1998). New Polymer Electrolytes for Low Temperature Fuel Cells. In:

Takezawa, N.; Kobayashi, H.; Kamegai, Y. & Shimokawabe, M. (1982). *Appl. Cat.,* 3, pp. 381-388

Tang, H.; Pintauro, P.N.; Guo, Q. & O'Connor, S. (1999). *J. Appl. Polym. Sci.,* 71, pp. 387-399 Tang, H.; Pan, M.; Jiang, S.; Wan, Z. & Yuan, R. (2005). *Colloids and Surfaces A: Physicochem.* 

Tay, S.W.; Zhang, X.; Liu, Z.; Hong, L. & Chan, S.H. (2008). *J. Membr. Sci.,* 321, pp. 139-145

Veziroglu, T.N. (Eds.), pp.74, Montréal, Quebec, Canada, July 9-13, 1995 Steck, A. & Stone, C. (1997). In: *Proceedings of the Second International Symposium on New* 

(Eds.), pp. 792-807, Montréal, Canada, July 6-10, 1997

Takahashi, K.; Takezawa, N. & Kobayashi, H. (1982). *Appl. Cat.,* 2, pp. 363-366

Tamura, K.; Tsukui, T.; Kamo, T. & Kudo, T. (1984). *Hitachi Hyoron,* 66, pp. 135-138

Tazi, B. & Savadogo, O. (2001). *J. New Mater. Electrochem. Systems,* 4, pp. 187-196 Tchicaya-Bouckary, L.; Jones, D.J. & Roziere, J. (2002). *Fuel Cells,* 2, pp. 40-45

Triphathi, B.P. & Shahi, V.K. (2007). *J. Colloid and Interface Science,* 316, pp. 612-621 Triphathi, B.P.; Kumar, M. & Shahi, V.K. (2009). *J. Membr. Sci.,* 327, pp. 145-154

Uchida, M.; Aoyama, Y.; Eda, N. & Ohta, A. (1995). *J. Electrochem. Soc*., 142, pp. 463-468 Ueda, S.; Eguchi, M.; Uno, K.; Tsutsumi, Y. & Ogawa, N. (2006). *Solid State Ionics,* 177, pp.

Wainright, J.S.; Wang, J.-T.; Weng, D.; Savinell, R.F. & Litt, M. (1995). *J. Electrochem. Soc*., 142,

Wainright, J.S.; Savinell, R.F. & Litt, M.H. (1997). In: *Proceedings of the Second International* 

Wang, J.-T.; Savinell, R.F.; Wainright, J.; Litt, M. & Yu, H. (1996a). *Electrochim. Acta,* 41, pp.

Wang, J.-T.; Wainright, J.S.; Savinell, R.F. & Litt, M. (1996b). *J. Appl. Electrochem*., 26, pp. 751-756

& Roberge, P.R. (Eds.), pp. 808-817, Montréal, Canada, July 6-10, 1997 Walker, M.; Baurngdrtner, K.-M.; Kaiser, M. *et al.,* (1999). *J. Appl. Polym. Sci.,* 74, pp. 67-73 Wang, J.; Wasmus, S. & Savinell, R.F. (1995). *J. Electrochem. Soc.,* 142, pp. 4218-4224

*Symposium on New Materials for Fuel Cell and Modern Battery Systems*, Savadogo, O.

Strickland, G. (1984). *Int. J. Hydrogen Energy,* 9, pp. 759-766

Takeishi, K. & Suzuki, H. (2004). *Appl. Cat. A,* 260, pp. 111-117

Tazi, B. & Savadogo, O. (2000). *Electrochim. Acta,* 45, pp. 4329-4339

Tricoli, V. & Nannetti, F. (2003). *Electrochim. Acta,* 48, pp. 2625-2633

Velu, S.; Suzuki, K. & Osaki, T. (1999). *Catal. Lett.,* 62, pp. 159-167

Tian, H. & Savadogo, O. (2005). *Fuel Cells,* 5, pp. 375-382

Tung, S.P. & Hwang, B.J. (2007). *Fuel Cells,* 7, pp. 32-39

Uma, T. & Nogami, M. (2007). *Fuel Cells,* 7, pp. 279-284

Verma, L.K. (2000). *J. Power Sources,* 86, pp. 464-468

2175-2178

pp. L121-L123

193-197

*Symposium on New Materials for Fuel Cell Systems* 1, Savadogo, O.; Roberge, P.R. &

*Materials for Fuel Cell and Modern Battery Systems*, Savadogo, O. & Roberge, P.R.

*Proceedings of the* 9th *International Conference on Solid State Protonic Conductors SSPC'98*, pp. 155-158, Extended Abstract Book, Bled, Slovenia, 17-21 August, 1998 Surampudi, S.; Narayanan, S.R.; Vamos, E.; Frank, H.; Halpert, G.; LaConti, A.; Kosek, J.; Surya Prakash, G.K. & Olah, G.A. (1994). *J. Power Sources,* 47, pp. 377-385


Powers, E.D. & Serad, G.A. (1986). In: *High Performance Polymers*: *Their Origin and* 

Pu, C.; Huang, W.; Ley, K.L. & Smotkin, E.S. (1995). *J. Electrochem. Soc.,* 142, pp. L119-L120

Ren, S.; Sun, G.; Li, C.; Song, S.; Xin, Q. & Yang, X. (2006). *J. Power Sources,* 157, pp. 724-726

Ruffmann, B.; Silva, H.; Schulte, B. & Nunes, S.P. (2003). *Solid State Ionics,* 162-163, pp. 269 -275 Sacca, A.; Carbone, A.; Passalacqua, E.; D'Epifanio, A.; Licoccia, S.; Traversa, E.; Sala, E.;

Sacca, A.; Gatto, I.; Carbone, A.; Pedicini, R. & Passalacqua, E. (2006). *J. Power Sources,* 163,

Samms, S.R.; Wasmus, S. & Savinell, R.F. (1996). *J. Electrochem. Soc*., 143, pp. 1225-1232 Santarelli, M.G.L.; Calì, M. & Bertonasco, A. (2003). *Energy Conversion and Management,* 44,

Schatter, M.J. (1983). In: *Fuel Cells*, Young, G.J. (Ed.), Vol. 2, pp. 290, Reinhold, New York

Shanmugam, S.; Viswanathan, B. & Varadarajan, T.K. (2006). *J. Membr. Sci.,* 275, pp. 105-109

Shin, S.-J.; Lee, J.-K.; Ha, H.-Y.; Hong, S.-A.; Chun, H.-S. & Oh, I.-H. (2002). *J. Power Sources,* 

Silva, V.S.; Ruffmann, B.; Silva, H.; Gallego, Y.A.; Mendes, A.; Madeira, L.M. & Nunes, S.P.

Silva, V.S.; Schirmer, J.; Reissner, R.; Ruffmann, B.; Silva, H.; Mendes, A.; Madeira, L.M. &

Silva, V.S.; Ruffmann, B.; Vetter, S.; Mendes, A.; Madeira, L.M. & Nunes, S.P. (2005c).

Soczka-Guth, T.; Baurmeister, J.; Frank, G. & Knauf, R. (1999). *International Patent* WO

Staiti, P.; Aricò, A.S.; Baglio, V.; Lufrano, F.; Passalacqua, E. & Antonucci, V. (2001). *Solid* 

Song, M.K.; Park, T.S.B.; Kim, Y.T. *et al*. (2004). *Electrochim. Acta,* 50, pp. 639-643

Schneller, A.; Ritter, H.; Ledjeff, K.; Nolte, R. & Thorwirth, R. (1993). *EP* 0574791 A2 Scott, K.; Taama, W.M. & Argyropoulos, P. (1998). *J. Appl. Electrochem.,* 28, pp. 1389-1397

Qi, Z. & Kaufman, A. (2002). *J. Power Sources,* 110, pp. 177-185

Ralph, T.R. (1997). *Platinum Metals Rev*., 41, pp. 102-113

Rosenblatt, E. & Cohn, J. (1952). *US Patent* 2,601,221

Savadogo, O. (2004). *J. Power Sources,* 127, pp. 135-161

Serov, A. & Kwak, C. (2009). *Appl. Cat.* B, 91, pp. 1-10

(2005a). *J. Power Sources,* 140, pp. 34-40

*Catalysis Today,* 104, pp. 205-212

*State Ionics,* 145, pp. 101-107

Raadschelders, J.W. & Jansen, T. (2001). *J. Power Sources,* 96, pp. 160-166 Ralph, T.R. & Hogarth, M.P. (2002). *Platinum Metals Rev.,* 46, pp. 117-135

Ramani, V.; Kunz, H.R. & Fenton, J.M. (2004). *J. Membr. Sci.,* 232, pp. 31-44

Ren, X. & Gottesfeld, S. (2001). *J. Electrochem. Soc*., 148, pp. A87-A93

Sambandam, S. & Ramani, V. (2007). *J. Power Sources,* 170, pp. 259-267

Savadogo, O. (1998). *J. New Mater. Electrochem. Systems*, 1, pp. 47-66

Scherer, G.G. (1990). *Ber. Bunsenges. Phys. Chem*., 94, pp. 1008-1014

Shao, Z.-G.; Joghee, P. & Hsing, I.-M. (2004). *J. Membr. Sci.,* 229, pp. 43-51 Shikada, T.; Asanuma, M. & Ikariya, T. (1991). *US Patent* 5,055,282

Nunes, S.P. (2005b). *J. Power Sources,* 140, pp. 41-49

Ramani, V.; Kunz, H.R. & Fenton, J.M. (2005a). *Electrochim. Acta,* 50, pp. 1181-1187 Ramani, V.; Kunz, H.R. & Fenton, J.M. (2005b). *J. Membr. Sci.,* 266, pp. 110-114 Rao, N.; Anderson, T.P. & Ge, P. (1994). *Solid State Ionics,* 72, pp. 334-337

Traini, F. & Ornelas, R. (2005). *J. Power Sources,* 152, pp. 16-21

Savadogo, O. & Xing, B. (2000). *J. New Mater. Electrochem. Systems,* 3, pp. 343-347

Amsterdam

pp. 47-51

pp. 2353-2370

106, pp. 146-152

99/29763

*Development,* Seymour, R.B. & Kirschenbaum, G.S. (Eds.), pp. 355, Elsevier,


**21** 

*Italy* 

**Membrane Operations** 

**for Industrial Applications** 

*1Department of Chemical Engineering and Materials and INSTM Consortium, University of Calabria, Rende (CS) 2Department of Ecology, University of Calabria, Rende (CS)* 

Maria Giovanna Buonomenna1,\*, Giovanni Golemme1,\* and Enrico Perrotta2

A resource-intensive industrial development, particularly in some Asian countries, characterized the last century. Its main causes can be ascribed to the significant elongation of life expectation, and to the overall increase in the standards characterizing the quality of life. The drawback of these positive aspects is the emergence of problems related to the industrial development: water stress, the environmental pollution, and the increase of CO2

The need to achieve a knowledge-intensive industrial development is nowadays well recognized. This will permit the transition from an industrial system based on quantity to

Sustainable development is a development that meets the needs of the present without compromising the ability of future generations to meet their own needs (United Nations General Assembly, 1987). The "three pillars" of sustainability are the environmental, social and economic demands – (United Nations General Assembly, 1987), which are not mutually exclusive and can be mutually reinforced (Figure 1): both economy and society are constrained by environmental limits. Sustainability is the path of continuous improvement, wherein the products and services required by society are delivered with progressively less negative impacts upon the Earth. Figure 2 is a cartoon depicting the road to sustainability. To benchmark sustainability, the Sustainability Index (SI) accounts for key factors that are

one based on quality in the framework of a sustainable development.

fundamental to the industrial process (Cobb *et al*, 2007)




Corresponding Authors


 \*



**1. Introduction** 

emissions into the atmosphere.


## **Membrane Operations for Industrial Applications**

#### Maria Giovanna Buonomenna1,\*, Giovanni Golemme1,\* and Enrico Perrotta2 *1Department of Chemical Engineering and Materials and INSTM Consortium, University of Calabria, Rende (CS) 2Department of Ecology, University of Calabria, Rende (CS) Italy*

#### **1. Introduction**

542 Advances in Chemical Engineering

Watanabe, M.; Uchida, H.; Seki, Y.; Emori, M. & Stonchart, P. (1996). *J. Electrochem. Soc.,* 143,

Xing, P.; Robertson, G.P.; Guiver, M.D.; Mikhailenko, S.D.; Wang, K. & Kaliaguine, S. (2004).

Yang, C.; Costamagna, P.; Srinivasan, S.; Benziger, J. & Bocarsly, A.B. (2001a). *J. Power* 

Yang, C.; Srinivasan, S.; Aricò, A.S.; Cretì, P.; Baglio, V. & Antonucci, V. (2001b). *Electrochem.* 

Yang, C.; Srinivasan, S.; Bocarsly, A.B.; Tulyani, S. & Benziger, J.B. (2004). *J. Membr. Sci.,* 237,

Yen, S.-P.S.; Narayanan, S.R.; Halpert, G.; Graham, E. & Yavrouian, A. (1998). *US Patent*

Yoon, S.R.; Hwang, G.H.; Cho, W.I.; Oh, I.H.; Hong, S.A. & Ha, H.Y. (2002). *J. Power Sources,*

Yoshida, N.; Ishisaki, T.; Watakabe, A. & Yoshitake, M. (1998). *Electrochim. Acta,* 43, pp. 3749-

Zaidi, S.M.J.; Mikhailenko, S.D.; Robertson, G.P.; Guiver, M.D. & Kaliaguine, S. (2000). *J.* 

Zawodzinski, T.A. Jr.; Neeman, M.; Sillerud, L.O. & Gottesfeld, S. (1991). *J. Phys. Chem.,* 95,

Zawodzinski, T.A. Jr.; Springer, T.E.; Davey, J.; Jestel, R.; Lopez, C.; Valerio, J. & Gottesfeld,

Zhou, W.J.; Zhou, B.; Li*,* W.Z.; Zhou, Z.H.; Song, S.Q.; Sun, G.Q.; Xin, Q.; Douvartzides, S.;

Zhou, X.; Weston, J.; Chalkova, E.; Hofmann, M.A.; Ambler, C.M.; Allcock, H.R. & Lvov,

Goula, M. & Tsiakaras, P. (2004). *J. Power Sources,* 126, pp. 16-22

Yu, R.H.; Choi, H.G. & Cho, S.M. (2005). *Electrochem. Commun.,* 7, pp. 1385-1388

Xu, K.; Lao, S.J.; Qin, H.Y.; Liu, B.H. & Li, Z.P. (2010). *J. Power Sources,* 195, pp. 5606-5609 Xu, W.; Liu, C.; Xue, X.; Su, Y.; Lv, Y.; Xing, W. & Lu, T. (2004). *Solid State Ionics,* 171, pp.

Wang, Y.; Yang, D.; Zheng, X.; Jiang, Z. & Li, J. (2008). *J. Power Sources,* 183, pp. 454-463

Watanabe, M.; Uchida, H. & Emori, M. (1998). *J. Electrochem. Soc.,* 145, pp. 1137-1141

Wilson, M.S.; Valerio, J.A. & Gottesfeld, S. (1995). *Electrochim. Acta,* 40, pp. 355-363 Woo, Y.; Oh, S.Y.; Kang, Y.S. & Jung, B. (2003). *J. Membr. Sci.,* 220, pp. 31-45

Xing, B. & Savadogo, O. (1999). *J. New Mater. Electrochem. Systems,* 2, pp. 95-101

Wang, S.; Ishihara, T. & Takita, Y. (2002). *Appl. Cat. A,* 228, pp. 167-176

Wei, J.; Stone, C. & Steck, A.E. (1995b). WO 95/08581, 30 March 1995 Williams, K.R. (1966). In: *An Introduction to Fuel Cells*, pp. 152, Elsevier Wilson, M.S. & Gottesfeld, S. (1992). *J. Electrochem. Soc.,* 139, pp. L28-L30

Wycisk, R. & Pintauro, P.N. (1996). *J. Membr. Sci.,* 119, pp. 155-160

Yi, J.S. & Nguyen, T.V. (1999). *J. Electrochem. Soc.,* 146, pp. 38-45

S. (1993). *J. Electrochem. Soc.,* 140, pp. 1981-1985 Zhai, Y.; Zhang, H.; Hu, J. & Yi, B. (2006). *J. Membr. Sci.,* 280, pp. 148-155 Zhang, Q.; Li, X.; Fujimoto, K. & Asami, K. (2005). *Appl. Cat. A,* 288, pp. 169-174 Zhang, W.; Tang, C.-M. & Kerres, J. (2001). *Sep. Purif. Technol.,* 22-23, pp. 209-221

S.N. (2003). *Electrochim. Acta,* 48, pp. 2173-2180

Wei, J.; Stone, C. & Steck, A.E. (1995a). *US Patent* 5,422,411

Watanabe, M. (1995). *US Patent* 5,472,799

Wilson, M.S. (1993). *US Patent* 5,211,984

*J. Membr. Sci.,* 229, pp. 95-106

*Solid State Lett.,* 4, pp. A31-A34

pp. 3847-3852

121-127

pp. 145-161

5,795,496

3754

106, pp. 215-223

pp. 6040-6044

*Membr. Sci.,* 173, pp. 17-34

*Sources,* 103, pp. 1-9

Wang, Z.H.; Wang, C.Y. & Chen, K.S. (2001). *J. Power Sources,* 94, pp. 40-50

A resource-intensive industrial development, particularly in some Asian countries, characterized the last century. Its main causes can be ascribed to the significant elongation of life expectation, and to the overall increase in the standards characterizing the quality of life. The drawback of these positive aspects is the emergence of problems related to the industrial development: water stress, the environmental pollution, and the increase of CO2 emissions into the atmosphere.

The need to achieve a knowledge-intensive industrial development is nowadays well recognized. This will permit the transition from an industrial system based on quantity to one based on quality in the framework of a sustainable development.

Sustainable development is a development that meets the needs of the present without compromising the ability of future generations to meet their own needs (United Nations General Assembly, 1987). The "three pillars" of sustainability are the environmental, social and economic demands – (United Nations General Assembly, 1987), which are not mutually exclusive and can be mutually reinforced (Figure 1): both economy and society are constrained by environmental limits. Sustainability is the path of continuous improvement, wherein the products and services required by society are delivered with progressively less negative impacts upon the Earth. Figure 2 is a cartoon depicting the road to sustainability. To benchmark sustainability, the Sustainability Index (SI) accounts for key factors that are fundamental to the industrial process (Cobb *et al*, 2007)


 \* Corresponding Authors

Membrane Operations for Industrial Applications 545

Their intrinsic characteristics of efficiency and operational simplicity, high selectivity and permeability for the transport of specific components, compatibility between different membrane operations in integrated systems, low energetic requirement, good stability under operating conditions and environmental compatibility, easy control and scale-up, and large operational flexibility, represent an interesting answer for the rationalization of

The traditional membrane separation operations such as reverse osmosis (RO), microfiltration (MF), ultrafiltration (UF), and nanofiltration (NF), electrodialysis,

> Concentration difference (C)

Gas separation

Pervaporation

Dialysis (D)

Reverse osmosis

Conventional membrane separation processes have at least two phase interfaces: feed fluidmembrane interface and product/permeate fluid-membrane interface on the two sides of the membrane. For example, commercialized membrane separation processes, such as RO, NF, UF, MF, GS, and PV, have two such phase interfaces. Over the last couple of decades, new membranes and membrane-separation techniques have appeared wherein the interface between two bulk phases is allowing the development of a new (improved) membrane, the creation of a novel membrane separation technique, the enhancement of the separation in existing membrane-separation processes, or the enhancement of the separation as such. The impact of such new techniques on conventional equilibrium-based separation processes/techniques is striking. The nature of the phase interface in such techniques is

These innovative membrane systems, the separation principle of which is the phase equilibrium and known as membrane contactors, have been studied, realised, and used in integrated membrane processes. In Table 2, a classification of the different types of

(GS)

(PV)

(RO)

Table 1. Classification of membrane processes according to their driving forces

Darcy's law Fick's law Fourier's law Ohm's law

Temperature difference (T)

Membrane distillation (MD)

Electrical potential difference (f)

Electro-dialysis

(ED)

 Electro-osmosis (EO)

pervaporation, etc. (Table 1), are largely used in many different applications.

industrial productions (Drioli & Romano, 2001).

Pressure difference (P)

Microfiltration

Ultrafiltration

Nanofiltration

Reverse osmosis

(MF)

(UF)

(NF)

(RO)

Driving Force

Phenomenological Equation

> Membrane Operations

often crucial.

membrane contactors is given.

Fig. 1. Relationship between the three pillars of sustainability: environmental limits constrain both society and economy.

Fig. 2. Cartoon showing the features and the requirements of the road to sustainability

Chemical Engineering faces today the crucial challenge of sustainable growth to find solutions to the increasing demand for raw materials, energy and tailor-made products.

In this context, the rational integration and implementation of innovative technologies, able to increase process performance, save energy, reduce costs, and minimize the environment impact represent interesting answers.

Recently, the logic of process intensification has been suggested as the best process engineering answer to the situation. It consists of innovative equipment, design, and process development methods that are expected to bring substantial improvements in chemical and any other manufacturing and processing, such as decreasing production costs, equipment size, energy consumption, and waste generation, and improving remote control, information fluxes, and process flexibility (Charpentier, 2007).

Membrane operations are, in principle, the most attractive candidates to satisfy the process intensification concepts and requirements.

Economy

Fig. 1. Relationship between the three pillars of sustainability: environmental limits

Society

Environment

Fig. 2. Cartoon showing the features and the requirements of the road to sustainability

Chemical Engineering faces today the crucial challenge of sustainable growth to find solutions to the increasing demand for raw materials, energy and tailor-made products.

In this context, the rational integration and implementation of innovative technologies, able to increase process performance, save energy, reduce costs, and minimize the environment

Recently, the logic of process intensification has been suggested as the best process engineering answer to the situation. It consists of innovative equipment, design, and process development methods that are expected to bring substantial improvements in chemical and any other manufacturing and processing, such as decreasing production costs, equipment size, energy consumption, and waste generation, and improving remote control,

Membrane operations are, in principle, the most attractive candidates to satisfy the process

constrain both society and economy.

impact represent interesting answers.

intensification concepts and requirements.

information fluxes, and process flexibility (Charpentier, 2007).

Their intrinsic characteristics of efficiency and operational simplicity, high selectivity and permeability for the transport of specific components, compatibility between different membrane operations in integrated systems, low energetic requirement, good stability under operating conditions and environmental compatibility, easy control and scale-up, and large operational flexibility, represent an interesting answer for the rationalization of industrial productions (Drioli & Romano, 2001).

The traditional membrane separation operations such as reverse osmosis (RO), microfiltration (MF), ultrafiltration (UF), and nanofiltration (NF), electrodialysis, pervaporation, etc. (Table 1), are largely used in many different applications.


Table 1. Classification of membrane processes according to their driving forces

Conventional membrane separation processes have at least two phase interfaces: feed fluidmembrane interface and product/permeate fluid-membrane interface on the two sides of the membrane. For example, commercialized membrane separation processes, such as RO, NF, UF, MF, GS, and PV, have two such phase interfaces. Over the last couple of decades, new membranes and membrane-separation techniques have appeared wherein the interface between two bulk phases is allowing the development of a new (improved) membrane, the creation of a novel membrane separation technique, the enhancement of the separation in existing membrane-separation processes, or the enhancement of the separation as such. The impact of such new techniques on conventional equilibrium-based separation processes/techniques is striking. The nature of the phase interface in such techniques is often crucial.

These innovative membrane systems, the separation principle of which is the phase equilibrium and known as membrane contactors, have been studied, realised, and used in integrated membrane processes. In Table 2, a classification of the different types of membrane contactors is given.

Membrane Operations for Industrial Applications 547

Fig. 3. Water desalination plant based on membrane process using solar energy, developed by the Japanese Water Re-use promotion center, in co-operation with Takenaka Corporation

However, high quality feedwater is required to ensure stable, long term performance, and an effective pre-treatment is essential for an efficient plant operation. In the past conventional, (i.e. conventional and physical) pre-treatment without the use of membrane technologies has been applied. Nowadays, membrane-based pre-treatments (such as MF,

MF is a low energy-consuming technique extensively used to remove suspended solids and to lower chemical oxygen demand (COD)/biochemical oxygen demand (BOD) and silt density index (SDI). UF retains suspended solids, bacteria, macromolecules and colloids and despite of the larger pressure gradient with respect to MF, this membrane separation method is competitive against conventional pre-treatments. In an integrated membrane pretreatment (Figure 4), the benefits of lower fouling rates of RO membranes compensates the

Fig. 4. Integrated (UF/RO) seawater desalination plant: UF as a pre-treatment

and Organo Corporation (Drioli *et al*, 2011).

UF, NF) tend to replace conventional pre-treatment systems.

higher cost membrane pre-treatment equipment.


Table 2. Classification of membrane contactors

At present, redesigning important industrial production cycles by combining various membrane operations suitable for separation and conversion units, thus realizing highly integrated membrane processes, is an attractive opportunity because of the synergic effects that can be attained.

In various fields, membrane operations are already the dominant technology. Their utilizations as hybrid systems, in combination with other conventional techniques or integrated with different membrane operations, is considered the way forward rationale applications.

In this context, interesting examples are in seawater desalination ; in wastewater treatment and reuse ; and in gas separation.

#### **2. Seawater desalination**

Sea and brackish water desalination has been at the origin of the interest for membrane operations, and the research efforts on RO membranes have had an impact on all of the progress in membrane science. Reverse osmosis desalination plants are currently leading the the desalination market, with RO installations representing 60% of the total number of worldwide plants, whereas thermal processes represent just 34.8% (Drioli *et al*, 2011). In Fig. 3 an example of the water desalination processes developed by the Japanese Water Re-Use Promotion Center, in co-operation with Takenaka Corporation and Organo Corporation: this process uses solar energy allowing the installation at location with no external electric energy supply (Drioli *et al*, 2011).

In Table 3 a list of traditional membrane technologies for water treatment is given.

The great flexibility, operational simplicity and mutual compatibility for integration of membrane operations offer the possibility of combining different membrane technologies for minimizing the limits of the single membrane units and for increasing the efficiency of the overall system.

Nowadays the most part of conventional seawater desalination plants use either RO or Multi-Stage Flash (MSF) technology. Thermal desalination is the most frequently applied technology in the Middle East, whilst membrane processes have rapidly developed and now surpass thermal processes in new plant installations due to the lesser energy consumptions (2.2.-6.7 kWh/m3 for seawater RO, with respect to 17-18 kWh/m3 for MSF).

**Membrane emulsifiers** 

**PRESSURE GRADIENT** 

**RESISTANCE IN MEMBR. Or LIQUID** 

**Membrane crystallizers (MCr)** 

**PARTIAL PRESSURE GRADIENT** 

**TEMPERAT. CONCENTR. POLARIZ.** 

**Phase transfer catalysis** 

**CONCENTR. GRADIENT** 

**RESISTAN. IN MEMBR. OR LIQUID** 

**Osmotic distillation (OD)** 

**Phase 1 LIQUID GAS/LIQUID LIQUID LIQUID LIQUID LIQUID Phase 2 LIQUID LIQUID LIQUID LIQUID LIQUID LIQUID** 

> **CONCENTR. POLARIZ.**

At present, redesigning important industrial production cycles by combining various membrane operations suitable for separation and conversion units, thus realizing highly integrated membrane processes, is an attractive opportunity because of the synergic effects

In various fields, membrane operations are already the dominant technology. Their utilizations as hybrid systems, in combination with other conventional techniques or integrated with different membrane operations, is considered the way forward rationale

In this context, interesting examples are in seawater desalination ; in wastewater treatment

Sea and brackish water desalination has been at the origin of the interest for membrane operations, and the research efforts on RO membranes have had an impact on all of the progress in membrane science. Reverse osmosis desalination plants are currently leading the the desalination market, with RO installations representing 60% of the total number of worldwide plants, whereas thermal processes represent just 34.8% (Drioli *et al*, 2011). In Fig. 3 an example of the water desalination processes developed by the Japanese Water Re-Use Promotion Center, in co-operation with Takenaka Corporation and Organo Corporation: this process uses solar energy allowing the installation at location with no external electric

In Table 3 a list of traditional membrane technologies for water treatment is given.

(2.2.-6.7 kWh/m3 for seawater RO, with respect to 17-18 kWh/m3 for MSF).

The great flexibility, operational simplicity and mutual compatibility for integration of membrane operations offer the possibility of combining different membrane technologies for minimizing the limits of the single membrane units and for increasing the efficiency of

Nowadays the most part of conventional seawater desalination plants use either RO or Multi-Stage Flash (MSF) technology. Thermal desalination is the most frequently applied technology in the Middle East, whilst membrane processes have rapidly developed and now surpass thermal processes in new plant installations due to the lesser energy consumptions

**PARTIAL PRESSURE GRADIENT** 

**Membrane distillation (MD)** 

**PARTIAL PRESSURE GRADIENT** 

**TEMPERAT. POLARIZ.** 

and reuse ; and in gas separation.

energy supply (Drioli *et al*, 2011).

the overall system.

**2. Seawater desalination** 

that can be attained.

applications.

**Driving force** 

**Limit to transport**  **Membrane strippers/ scrubbers** 

**CONCENTR. GRADIENT** 

**RESISTANCE IN MEMBR. Or LIQUID** 

Table 2. Classification of membrane contactors

Fig. 3. Water desalination plant based on membrane process using solar energy, developed by the Japanese Water Re-use promotion center, in co-operation with Takenaka Corporation and Organo Corporation (Drioli *et al*, 2011).

However, high quality feedwater is required to ensure stable, long term performance, and an effective pre-treatment is essential for an efficient plant operation. In the past conventional, (i.e. conventional and physical) pre-treatment without the use of membrane technologies has been applied. Nowadays, membrane-based pre-treatments (such as MF, UF, NF) tend to replace conventional pre-treatment systems.

MF is a low energy-consuming technique extensively used to remove suspended solids and to lower chemical oxygen demand (COD)/biochemical oxygen demand (BOD) and silt density index (SDI). UF retains suspended solids, bacteria, macromolecules and colloids and despite of the larger pressure gradient with respect to MF, this membrane separation method is competitive against conventional pre-treatments. In an integrated membrane pretreatment (Figure 4), the benefits of lower fouling rates of RO membranes compensates the higher cost membrane pre-treatment equipment.

Fig. 4. Integrated (UF/RO) seawater desalination plant: UF as a pre-treatment

Membrane Operations for Industrial Applications 549

Fig. 5. Flow diagram of integrated desalination membrane-based plant (Macedonio and

advantageous in many industrial processes and in particular in desalination systems.

power ratio and utilized effectively the available intake/outfall facilities.

single pass RO unit is blended with the MSF product (Hamed, 2006)

Hybrid membrane systems, in combination with conventional separation process, are

In recent years, the concept of simple hybrid multistage flash-reverse osmosis (MSF/RO) configuration has been applied to a number of existing or new commercial desalination plants. The SWCC Jeddah, Al-Jubail and Yanbu existing Power/Water cogeneration plants are expanded for more water production by combining with new SWRO desalination plants. The simple hybrid desalination arrangement enabled the increase of the water-to-

The MSF and RO operate in parallel and are entirely independent. The water product of the

The 100 MGD desalination plant in Fujairah (UAE) is one of the largest hybrid MSF (Multistage Flash evaporators) /RO installation in the worlds: it combines a 62.5 MGD MSF

This hybrid desalination system is designed to provide significant operational savings by reducing fuel consumption by up to 25 per cent compared with a similar-sized plant based only on MSF technology. Other key criteria influencing the design of the desalination plant were feed water quality, product water requirements and compatibility with the

Drioli, 2010).

and 37.5 MGD SWRO (Fig.6).

cogeneration of electricity.


Table 3. Membrane operations used for water treatments

Macedonio and Drioli (2010) analysed an integrated membrane-based desalination plant with membrane crystallization (MCr) as post-treatment for the recovery of salts and water contained in the NF/RO retentate streams of a desalination plant (Fig.5).

The feed water enters into MF membrane modules to be cleaned from suspended solids and large bacteria. After MF, process water is pressurized and then sent to NF membrane modules to be cleaned from turbidity, microorganisms, hardness, multivalent ions and 10-50% of monovalent species. After NF step, the RO step requires that the process water be pressurized to overcome the osmotic pressure. In the RO operation the process water is separated into a permeate and a brine. RO brine enters into the precipitator in which is mixed with Na2CO3 for the removal of the Ca2+ ions of the RO brine as CaCO3. The process stream enters the MCr modules where it is separated into a permeate, a purge and a salt containing stream.

Mode of

size exclusion, convection

size exclusion, convection

size exclusion

dialysation, diffusion

solutiondiffusion mechanism

Donnanexclusion

Donnanexclusion

size exclusion, diffusion, Donnanexclusion

and

separation Applications

water purification, sterilization

separation of molecular mixtures

purification of molecular mixtures, artificial kidney

separation of molecular mixtures

desalination

sea & brackish ,water

water desalination

water softening

concentration of solutions

diffusion water desalination,

and ions

Applied driving force

hydrostatic pressure 0.05-0.2 MPa

hydrostatic pressure 0.1-0.5 MPa

hydrostatic pressure 0.1-0.5 MPa

hydrostatic pressure 0.3-3 MPa

hydrostatic pressure 1-10 MPa

electrical potential

Vapor pressure

contained in the NF/RO retentate streams of a desalination plant (Fig.5).

concentration of ions

Macedonio and Drioli (2010) analysed an integrated membrane-based desalination plant with membrane crystallization (MCr) as post-treatment for the recovery of salts and water

The feed water enters into MF membrane modules to be cleaned from suspended solids and large bacteria. After MF, process water is pressurized and then sent to NF membrane modules to be cleaned from turbidity, microorganisms, hardness, multivalent ions and 10-50% of monovalent species. After NF step, the RO step requires that the process water be pressurized to overcome the osmotic pressure. In the RO operation the process water is separated into a permeate and a brine. RO brine enters into the precipitator in which is mixed with Na2CO3 for the removal of the Ca2+ ions of the RO brine as CaCO3. The process stream enters the MCr

modules where it is separated into a permeate, a purge and a salt containing stream.

Separation process

Microfiltration (MF)

Ultrafiltration (UF)

Diafiltration (DF)

Nanofiltration (NF)

Reverse osmosis (RO)

Electrodialysis (ED)

Donnan Dialysis (DD)

> Membrane Distillation (MD)

Membrane type used

symmetric macroporous

Asymmetric macroporous

asymmetric macroporous

asymmetric mesoporous

asymmetric skin-type, dense or microporous

symmetric ionexchange membrane

symmetric ionexchange membrane

Table 3. Membrane operations used for water treatments

symmetric porous hydrophobic membrane

Fig. 5. Flow diagram of integrated desalination membrane-based plant (Macedonio and Drioli, 2010).

Hybrid membrane systems, in combination with conventional separation process, are advantageous in many industrial processes and in particular in desalination systems.

In recent years, the concept of simple hybrid multistage flash-reverse osmosis (MSF/RO) configuration has been applied to a number of existing or new commercial desalination plants. The SWCC Jeddah, Al-Jubail and Yanbu existing Power/Water cogeneration plants are expanded for more water production by combining with new SWRO desalination plants. The simple hybrid desalination arrangement enabled the increase of the water-topower ratio and utilized effectively the available intake/outfall facilities.

The MSF and RO operate in parallel and are entirely independent. The water product of the single pass RO unit is blended with the MSF product (Hamed, 2006)

The 100 MGD desalination plant in Fujairah (UAE) is one of the largest hybrid MSF (Multistage Flash evaporators) /RO installation in the worlds: it combines a 62.5 MGD MSF and 37.5 MGD SWRO (Fig.6).

This hybrid desalination system is designed to provide significant operational savings by reducing fuel consumption by up to 25 per cent compared with a similar-sized plant based only on MSF technology. Other key criteria influencing the design of the desalination plant were feed water quality, product water requirements and compatibility with the cogeneration of electricity.

Membrane Operations for Industrial Applications 551

Fig. 7. Wastewater treatment facility at Sulaibiya near Kuwait City from http://www.water-

landfill, incineration, or by composting. Membrane filtration was selected to provide robust pretreatment of the secondary-treated municipal effluent before being fed to the RO. Membrane filtration was chosen over conventional tertiary clarification and filtration because it reduced the plant chemical consumption and could guarantee that low turbidity water is fed to the RO. It is expected that better quality pre-treatment to the RO will lead to longer membrane life, lower operating pressure, and reduced cleaning frequency for the RO system. The UF plant utilizes Norit's X-Flow membranes, which are capillary hydrophilic hollow fibers. The UF units are operated individually. Each unit is backwashed regularly, whereby all suspended matter that is being retained by the membranes is removed from the plant. The backwash water is pumped back upstream of the WWTP to achieve the highest possible overall water recovery for the plant. The salinity of the municipal effluent has an average monthly value of 1,280 mg/l TDS, with a maximum value of 3,014 mg/l. RO is used to desalinate the water to 100 mg/l TDS, as well as provide a second barrier to bacteria and viruses. RO technology is well proven for desalinating municipal effluent. The system consists of 42 identical skids in a 4:2:1 array. Approximately 21,000 membrane modules, provided by Toray of America, were required for this project. The RO product passes through a stripper to remove carbon dioxide to adjust pH with a minimum amount of caustic before distribution, and the product is then chlorinated before leaving the plant. RO

Membrane bioreactors (MBR) are a combination of activated sludge treatment and membrane filtration for biomass retention. Low-pressure membrane filtration, either UF or MF, is used to separate effluent from activated sludge. The two main MBR configurations involve either submerged membranes (Fig. 8) or external circulation (side-stream

technology.net/projects/sulaibiya/

Pre-treatment

brine is disposed of into the Persian Gulf.

configuration).

Fig. 6. Fujairah desalination plant in United Arab Emirates, from http://fujairahinfocus.blogspot.com/2011/10/fujairah-power-and-desalination-plant.html

The water production system at the Fujairah desalination plant is comprised of five Doosan MSF units producing 57 million l/day (12.5 MGD) each and one RO unit with a design capacity of 171 million l/day (37.5 MGD). The RO unit was supplied by Ondeo Degremont. For drinking water supply, distillate from the MSF units and desalinated water from the RO plant are mixed in a distillate header and treated in a re-mineralization unit before passing into the potable water storage tanks. Prior to export to the water transmission line, potable water is stored in five potable water tanks, each with a capacity of 91 million l (20 million gallons).

A promising approach for pre-treatment of seawater make-up feed to MSF and SWRO desalination processes using NF membranes has been introduced by the R&D Center (RDC) of SWCC.

NF membranes are capable to reduce significantly scale forming ions from seawater, allow high temperature operation of thermal desalination processes, and subsequently increase water productivity.

The developed fully integrated systems NF/MSF and NF/SWRO/MSF result in high water productivity and enhance thermal performance compared to the currently used simple hybrid desalination arrangements (Hamed, 2006).

#### **3. Wastewater treatment and reuse**

Considerable advances in MF, UF and NF technologies to recover municipal wastewater have been also achieved. Also in this case, the implementation of integrated membrane systems is growing rapidly with excellent results.

In fig. 7 the main treatment steps to recover municipal wastewater from Kuwait City and the surrounding area are reported: A conventional biological wastewater treatment plant (WWTP) treats the effluent to better than secondary effluent quality. The secondary effluent then flows to the water reclamation plant, which uses UF and RO to further treat the water for reuse. Sludge from the wastewater treatment plant is treated to allow for disposal by

http://fujairahinfocus.blogspot.com/2011/10/fujairah-power-and-desalination-plant.html

The water production system at the Fujairah desalination plant is comprised of five Doosan MSF units producing 57 million l/day (12.5 MGD) each and one RO unit with a design capacity of 171 million l/day (37.5 MGD). The RO unit was supplied by Ondeo Degremont. For drinking water supply, distillate from the MSF units and desalinated water from the RO plant are mixed in a distillate header and treated in a re-mineralization unit before passing into the potable water storage tanks. Prior to export to the water transmission line, potable water is stored in five potable water tanks, each with a capacity of 91 million l (20 million gallons).

A promising approach for pre-treatment of seawater make-up feed to MSF and SWRO desalination processes using NF membranes has been introduced by the R&D Center (RDC)

NF membranes are capable to reduce significantly scale forming ions from seawater, allow high temperature operation of thermal desalination processes, and subsequently increase

The developed fully integrated systems NF/MSF and NF/SWRO/MSF result in high water productivity and enhance thermal performance compared to the currently used simple

Considerable advances in MF, UF and NF technologies to recover municipal wastewater have been also achieved. Also in this case, the implementation of integrated membrane

In fig. 7 the main treatment steps to recover municipal wastewater from Kuwait City and the surrounding area are reported: A conventional biological wastewater treatment plant (WWTP) treats the effluent to better than secondary effluent quality. The secondary effluent then flows to the water reclamation plant, which uses UF and RO to further treat the water for reuse. Sludge from the wastewater treatment plant is treated to allow for disposal by

Fig. 6. Fujairah desalination plant in United Arab Emirates, from

of SWCC.

water productivity.

hybrid desalination arrangements (Hamed, 2006).

systems is growing rapidly with excellent results.

**3. Wastewater treatment and reuse** 

Fig. 7. Wastewater treatment facility at Sulaibiya near Kuwait City from http://www.watertechnology.net/projects/sulaibiya/

landfill, incineration, or by composting. Membrane filtration was selected to provide robust pretreatment of the secondary-treated municipal effluent before being fed to the RO. Membrane filtration was chosen over conventional tertiary clarification and filtration because it reduced the plant chemical consumption and could guarantee that low turbidity water is fed to the RO. It is expected that better quality pre-treatment to the RO will lead to longer membrane life, lower operating pressure, and reduced cleaning frequency for the RO system. The UF plant utilizes Norit's X-Flow membranes, which are capillary hydrophilic hollow fibers. The UF units are operated individually. Each unit is backwashed regularly, whereby all suspended matter that is being retained by the membranes is removed from the plant. The backwash water is pumped back upstream of the WWTP to achieve the highest possible overall water recovery for the plant. The salinity of the municipal effluent has an average monthly value of 1,280 mg/l TDS, with a maximum value of 3,014 mg/l. RO is used to desalinate the water to 100 mg/l TDS, as well as provide a second barrier to bacteria and viruses. RO technology is well proven for desalinating municipal effluent. The system consists of 42 identical skids in a 4:2:1 array. Approximately 21,000 membrane modules, provided by Toray of America, were required for this project. The RO product passes through a stripper to remove carbon dioxide to adjust pH with a minimum amount of caustic before distribution, and the product is then chlorinated before leaving the plant. RO brine is disposed of into the Persian Gulf.

Membrane bioreactors (MBR) are a combination of activated sludge treatment and membrane filtration for biomass retention. Low-pressure membrane filtration, either UF or MF, is used to separate effluent from activated sludge. The two main MBR configurations involve either submerged membranes (Fig. 8) or external circulation (side-stream configuration).

Membrane Operations for Industrial Applications 553

the permeate quality is superb, even in very high sludge concentrations (up to 2%) and in difficult applications. Besides TSS, the membranes effectively reject also bacteria and even

High flow rate and pressure feed of the reactor content (MLSS) into the Pleiade membrane module, enable high permeate fluxes (60-80 lit/m2\*hr). As a result, the membrane area required for treatment of a given WW flow is much lower than in most other MBR

High circulation speed over the membranes (2 m/sec) reduces fouling accumulation, and

External membrane systems enable full modularity and easy expansion of WW treatment capacity. Gao *et al* (2011) developed a completely green process based on the integration of MBR with UF by treating micro-polluted source water in drinking water treatment. The removal of organic matter is carried by both a biodegradation mechanism in the MBR and by the MF/UF membrane, while the nitrification in MBR removes ammonia (Fig. 10).

Fig. 10. Schematic diagram of the pilot scale experimental set-up MBR/UF (Gao *et al*, 2011)

MEE could also be decreased by 88% correspondingly.

products are the advantages obtained.

The pulp and paper industry is one of the most water-dependent industries (Nurdan and Emre, 2010). The alkaline peroxide mechanical pulping (APMP) process has been widely applied especially in Asia (Liu *et al,* 2011) for the high yield, and the relatively low pollution. To achieve a closed wastewater loop, several APMP plants in the world have attempted to concentrate the total effluent by using a multi-effect evaporation system. Zhang *et al* (2011) studied a hybrid process UF/Multi-effect-evaporation (MEE) to concentrate effluent from APMP plants. With this new membrane concentration process, 88% of the water in the effluent can be removed, 1.4 bilion KWh power could be saved; the capital investment for

The possibility of redesigning overall industrial production by the integration of various already developed membrane operations is of particular interest: low environmental impacts, low energy consumption, higher quality of final products and new available

The leather industry might be an interesting case study because of (i) the large environmental problems related to is operation; (ii) the low technological content of its

viruses.

membranes.

membrane cleaning (CIP) demand.

Fig. 8. Submerged membrane module for wastewater treatment. From ZeeWeed® Submerged Membrane System, from http://www.gewater.com.

Since the early MBR installations in the 1990s, the number of MBR systems has grown considerably. One key trend driving this growth is the use of MBR system for decentralized treatment and water reuse. The successful introduction of MBR systems into small scale and decentralized applications has led to the development of packaged treatment solutions from the main technology suppliers. The company Conder Products, UK, designed the package treatment plant Clereflo MBR; Zenon Environmental Inc., now a part of General Electric, produced ZeeMod®.

The Pleiade® Plate & Frame membranes produced by Orelis©, France, which is one of Europe's leading membrane manufacturer, are installed in skids mounted outside the bioreactor and the sludge is circulated through the module in high speed, by pumps (Fig. 9). This configuration and the Orelis membranes have several advantages over other MBR systems.

Fig. 9. Pleiade® MBR membrane bioreactor for treating 1000 m3/day of effluent from www.vic-ws.it/site/down/PLEENG0411.pdf

The membranes are in the UF range and offer high separation capabilities (0.02 m, MWCO=40 kD), unlike most MBR membranes which fall under the MF range. As a result,

Fig. 8. Submerged membrane module for wastewater treatment. From ZeeWeed®

Since the early MBR installations in the 1990s, the number of MBR systems has grown considerably. One key trend driving this growth is the use of MBR system for decentralized treatment and water reuse. The successful introduction of MBR systems into small scale and decentralized applications has led to the development of packaged treatment solutions from the main technology suppliers. The company Conder Products, UK, designed the package treatment plant Clereflo MBR; Zenon Environmental Inc., now a part of General Electric,

The Pleiade® Plate & Frame membranes produced by Orelis©, France, which is one of Europe's leading membrane manufacturer, are installed in skids mounted outside the bioreactor and the sludge is circulated through the module in high speed, by pumps (Fig. 9). This configuration and the Orelis membranes have several advantages over other MBR

Fig. 9. Pleiade® MBR membrane bioreactor for treating 1000 m3/day of effluent from

The membranes are in the UF range and offer high separation capabilities (0.02 m, MWCO=40 kD), unlike most MBR membranes which fall under the MF range. As a result,

www.vic-ws.it/site/down/PLEENG0411.pdf

Submerged Membrane System, from http://www.gewater.com.

produced ZeeMod®.

systems.

the permeate quality is superb, even in very high sludge concentrations (up to 2%) and in difficult applications. Besides TSS, the membranes effectively reject also bacteria and even viruses.

High flow rate and pressure feed of the reactor content (MLSS) into the Pleiade membrane module, enable high permeate fluxes (60-80 lit/m2\*hr). As a result, the membrane area required for treatment of a given WW flow is much lower than in most other MBR membranes.

High circulation speed over the membranes (2 m/sec) reduces fouling accumulation, and membrane cleaning (CIP) demand.

External membrane systems enable full modularity and easy expansion of WW treatment capacity. Gao *et al* (2011) developed a completely green process based on the integration of MBR with UF by treating micro-polluted source water in drinking water treatment. The removal of organic matter is carried by both a biodegradation mechanism in the MBR and by the MF/UF membrane, while the nitrification in MBR removes ammonia (Fig. 10).

Fig. 10. Schematic diagram of the pilot scale experimental set-up MBR/UF (Gao *et al*, 2011)

The pulp and paper industry is one of the most water-dependent industries (Nurdan and Emre, 2010). The alkaline peroxide mechanical pulping (APMP) process has been widely applied especially in Asia (Liu *et al,* 2011) for the high yield, and the relatively low pollution. To achieve a closed wastewater loop, several APMP plants in the world have attempted to concentrate the total effluent by using a multi-effect evaporation system. Zhang *et al* (2011) studied a hybrid process UF/Multi-effect-evaporation (MEE) to concentrate effluent from APMP plants. With this new membrane concentration process, 88% of the water in the effluent can be removed, 1.4 bilion KWh power could be saved; the capital investment for MEE could also be decreased by 88% correspondingly.

The possibility of redesigning overall industrial production by the integration of various already developed membrane operations is of particular interest: low environmental impacts, low energy consumption, higher quality of final products and new available products are the advantages obtained.

The leather industry might be an interesting case study because of (i) the large environmental problems related to is operation; (ii) the low technological content of its

Membrane Operations for Industrial Applications 555

40 20

60 20 10

Table 4. Sales estimates and sales predicted for the principal gas and vapour separation

Nowadays, obtaining cheap high purity gases or enriched gas mixtures (the air, in particular) is a very important problem in industry and medicine as well as in everyday life

Methods for air separation or oxygen enrichment can be divided into two groups (Freeman *et al*, 2006): cryogenic and non - cryogenic. The gaseous oxygen and nitrogen market is dominated by cryogenic distillation of air, and vacuum swing adsorption (Koros *et al*, 2000). Of the non – cryogenic methods, selective adsorption on zeolites and carbon adsorbents are available (Lin & Guthrie, 2006), and more and more attention is attracted to the membrane separation techniques (Dhingra & Marand, 1998) . The success of polymeric membranes has been largely based on their mechanical and thermal stability, along with good gas separation properties. The process of membrane separation is continuous, has a low capital cost, low power consumption, and the membranes, at least in gas separation, do not require

Nitrogen production today is the largest GS process in use. Nitrogen gas is used in many applications (e.g., to prevent fires and explosions in tanks and piping systems and to prevent equipment degradation, during shutdown periods, in compressors, pipelines and reactors). Single-stage membrane operation is preferred. Air is pressurized and fed into the membrane separators; faster gases (O2, CO2, water vapor) permeate through the polymeric fiber walls, are collected and vented to the atmosphere while the slower, non-permeate N2

Application Annual membrane system sales (106 U.S. \$)

Year 2010 Year 2020

80 70

100 50 25

In Table 4 the membrane sales involved for industrial gas separations are shown.

Nitrogen from air 100 125 Oxygen from air 10 30

Vapour (C2+)/Gas (N2, Ar) 30 90 Vapour/Vapour (including dehydration) 20 100 Air dehydration/other 30 60 TOTAL 340 730 Annual growth, % 8 8

**4. Gas separation** 

H2/gas (CO, N2, C1, C2) H2/gas (C3+, CO2)

NGL removal and recovery N2 removal, dehydration

Hydrogen

Natural gas CO2 removal

applications

(Bodzek, 2000).

regeneration (Vansant & Dewolfs, 1990).

gas is available at the other end of the separator.

traditional operations; (iii) the tendency to concentrate a large number of small-medium industries in specific districts. In Fig. 11 an ideal process based on integrated membrane operations for the tanning process is showed.

Fig. 11. Integrated membrane process proposed for the tanning process in the leather industry (Drioli & Romano, 2001)

The pollution problems of the leather industry have been minimized one by one at the point where they originate, thereby avoiding the need for huge wastewater treatment plants at the end of the overall production line. In addition, the membrane operations act by physical mechanisms without modification of the chemical procedure at the origin of the final high quality of the leather (Cassano *et al*, 2003)

#### **4. Gas separation**

554 Advances in Chemical Engineering

traditional operations; (iii) the tendency to concentrate a large number of small-medium industries in specific districts. In Fig. 11 an ideal process based on integrated membrane

Fig. 11. Integrated membrane process proposed for the tanning process in the leather

The pollution problems of the leather industry have been minimized one by one at the point where they originate, thereby avoiding the need for huge wastewater treatment plants at the end of the overall production line. In addition, the membrane operations act by physical mechanisms without modification of the chemical procedure at the origin of the final high

industry (Drioli & Romano, 2001)

quality of the leather (Cassano *et al*, 2003)

operations for the tanning process is showed.


In Table 4 the membrane sales involved for industrial gas separations are shown.

Table 4. Sales estimates and sales predicted for the principal gas and vapour separation applications

Nowadays, obtaining cheap high purity gases or enriched gas mixtures (the air, in particular) is a very important problem in industry and medicine as well as in everyday life (Bodzek, 2000).

Methods for air separation or oxygen enrichment can be divided into two groups (Freeman *et al*, 2006): cryogenic and non - cryogenic. The gaseous oxygen and nitrogen market is dominated by cryogenic distillation of air, and vacuum swing adsorption (Koros *et al*, 2000). Of the non – cryogenic methods, selective adsorption on zeolites and carbon adsorbents are available (Lin & Guthrie, 2006), and more and more attention is attracted to the membrane separation techniques (Dhingra & Marand, 1998) . The success of polymeric membranes has been largely based on their mechanical and thermal stability, along with good gas separation properties. The process of membrane separation is continuous, has a low capital cost, low power consumption, and the membranes, at least in gas separation, do not require regeneration (Vansant & Dewolfs, 1990).

Nitrogen production today is the largest GS process in use. Nitrogen gas is used in many applications (e.g., to prevent fires and explosions in tanks and piping systems and to prevent equipment degradation, during shutdown periods, in compressors, pipelines and reactors). Single-stage membrane operation is preferred. Air is pressurized and fed into the membrane separators; faster gases (O2, CO2, water vapor) permeate through the polymeric fiber walls, are collected and vented to the atmosphere while the slower, non-permeate N2 gas is available at the other end of the separator.

Membrane Operations for Industrial Applications 557

Fig. 13. Plant for H2 recovery from ammonia synthesis (Drioli & Romano, 2001)

Catalytic reformer off gases Synthesis gas composition adjustment (IGCC)

In a comparison of three separation technologies (membrane, PSA, cryogenic distillation) applied for H2 recovery from refinery off-gas, Spillman (1989) reported that the use of membranes represent the best choice. An evaluation of these processes is provided in terms of sustainability indexes. The *Energy Intensity* of the membrane system is the lowest; the *Mass Intensity* is less than 50% of that for the conventional separations; the membrane system occupies a tenfold lower area required by PSA and cryogenic distillation of

Coke oven gases

Refining Chemical/Petrochemical

HDT/Recycle purge gases Methanol purge gases Refinery fuel gases Ammonia purge gases PSA off gases Polypropylene purge Polyethylene purge Styrene off gases

Electrolysis gases

corresponding capacity (highest *Productivity/Footprint ratio*).

Table 5. H2 membrane applications

Cyclohexane plant gases

Oxygen production by membrane systems is still underdeveloped, since most of the industrial O2 applications require purity higher than 90%, which is easily achieved by adsorption or cryogenic technologies but not by single-stage membranes.

New materials are being developed that could possibly have higher permeabilities than conventional solid electrolytes. Promising oxygen permeation have been obtained in many perovskite systems (Zhu *et al*, 2009; Stiegel, 1999). The dense perovskite type membranes transport oxygen as lattice ions at elevated temperatures with infinite selectivity ratios in O2 separations.

The oxygen-ion conducting membranes must operate above 700°C: an efficient and costeffective way to recover the energy contained in the non- permeate, oxygen-depleted stream is that of integrating the membrane system with a gas turbine (Fig. 12) ( Drioli & Romano, 2001).

Fig. 12. Integrated system for O2 and power productions ( Drioli & Romano, 2001).

Hydrogen recovery was among the first large–scale commercial applications of membrane GS technology. The commercial success in the mid-1970's in Louisiana of the Permea hollow-fiber Prism system for in-process recycling of hydrogen from ammonia purge gases (Fig. 13) was the starting point of the penetration of membrane technology in large-scale manufacturing. A two-step membrane design was chosen for this ideal application for membrane technology: the ammonia reactor operates at high pressures (ca. 130 bar), thus providing the necessary driving force for separation; the H2/N2 membrane selectivity is high and the feed gas is free of contaminants.

This technology has been extended to other situations. In Table 5, some recent H2 membrane applications are listed. H2 recovery from refinery streams is an emerging field for membrane GS in the petrochemical industry; it is a key approach to meet the increased demand of hydrogen (for hydrotreating, hydrocracking or hydrodesulfurization processes) owing to new environmental regulations. An example is the H2 recovery from high pressure purge gas of a hydrotreater.

Oxygen production by membrane systems is still underdeveloped, since most of the industrial O2 applications require purity higher than 90%, which is easily achieved by

New materials are being developed that could possibly have higher permeabilities than conventional solid electrolytes. Promising oxygen permeation have been obtained in many perovskite systems (Zhu *et al*, 2009; Stiegel, 1999). The dense perovskite type membranes transport oxygen as lattice ions at elevated temperatures with infinite selectivity ratios in O2

The oxygen-ion conducting membranes must operate above 700°C: an efficient and costeffective way to recover the energy contained in the non- permeate, oxygen-depleted stream is that of integrating the membrane system with a gas turbine (Fig. 12) ( Drioli & Romano,

Fig. 12. Integrated system for O2 and power productions ( Drioli & Romano, 2001).

high and the feed gas is free of contaminants.

gas of a hydrotreater.

Hydrogen recovery was among the first large–scale commercial applications of membrane GS technology. The commercial success in the mid-1970's in Louisiana of the Permea hollow-fiber Prism system for in-process recycling of hydrogen from ammonia purge gases (Fig. 13) was the starting point of the penetration of membrane technology in large-scale manufacturing. A two-step membrane design was chosen for this ideal application for membrane technology: the ammonia reactor operates at high pressures (ca. 130 bar), thus providing the necessary driving force for separation; the H2/N2 membrane selectivity is

This technology has been extended to other situations. In Table 5, some recent H2 membrane applications are listed. H2 recovery from refinery streams is an emerging field for membrane GS in the petrochemical industry; it is a key approach to meet the increased demand of hydrogen (for hydrotreating, hydrocracking or hydrodesulfurization processes) owing to new environmental regulations. An example is the H2 recovery from high pressure purge

adsorption or cryogenic technologies but not by single-stage membranes.

separations.

2001).

Fig. 13. Plant for H2 recovery from ammonia synthesis (Drioli & Romano, 2001)


Table 5. H2 membrane applications

In a comparison of three separation technologies (membrane, PSA, cryogenic distillation) applied for H2 recovery from refinery off-gas, Spillman (1989) reported that the use of membranes represent the best choice. An evaluation of these processes is provided in terms of sustainability indexes. The *Energy Intensity* of the membrane system is the lowest; the *Mass Intensity* is less than 50% of that for the conventional separations; the membrane system occupies a tenfold lower area required by PSA and cryogenic distillation of corresponding capacity (highest *Productivity/Footprint ratio*).

Membrane Operations for Industrial Applications 559

Fig. 14. Membrane/amine hybrid plant for the treatment of natural gas (Baker &

work on a composite membrane and a laboratory-scale PSA unit.

CO2 flux in the first membrane unit is enhanced.

(http://www.mtrinc.com/news.html#DOE14mil).

units. Instead of constant-composition regular feed, the PSA is fed with a mixture which is progressively enriched in the more adsorbed component during the pressurization and high-pressure adsorption steps of the cycle. This results in sharper concentration fronts. The hybrid has been applied successfully to the bulk separation of an 30:70 mol% CO2/N2 mixture over activated carbon. Process performance is reported in terms of product recovery and purity at cyclic steady state. Numerical simulations were validated by experimental

MTR (Membrane Technology and Research, Inc.) has proposed an innovative retro-fit process for the post-combustion carbon capture and sequestration (CCS) of existing coalfired power plants (Merkel et al., 2010). According to preliminary calculations, a two stage membrane process (Figure 15) should be able to recover 90% of the CO2 released by the power plant at a competitive cost of 23\$/ton CO2 sequestrated. The selected lower cost configuration uses a blower upstream of the first membrane unit and a vacuum pump downstream in order to boost the driving force for the separation. At the same time, in the second membrane module the air to be fed to the boiler strips part of the CO2 in the N2 enriched flue gas from the first membrane module, thereby recycling CO2 and building a higher CO2 concentration for the first membrane module: as a consequence, the sequestered

MTR is now involved in a contract with the US Department of Energy for the construction of a membrane skid containing MTR's Polaris membranes capable of 90% CO2 capture from a 20 tons-of-CO2/day slipstream of coal fired flue gas. The skid will be operated during a 6 month field test at Arizona Public Service's Cholla power plant.

In the process industry the final choice of a separation process is the result of a balance between the economics, the desired purity and recovery of the product and other conditions and restrictions, such as the desired capacity and the composition of the feed and the possibility of integration with other processes. In this logic, the integration of commercial membrane separation units in the recovery of ethylene or propylene monomer from polyolefin resin degassing vent streams has been proposed by Baker *et al* (1998). A simplified flow scheme of the process that generates such streams is shown in

Lokhandwala, 2008).

Fig. 16(a).

Carbon dioxide removal from natural gas (natural gas sweetening) is mandatory to meet pipeline specifications (e.g., down to 2% vol. in U.S.A.), since CO2 reduces the heating value of natural gas, is corrosive and freezes at a relatively high temperature, forming blocks of dry ice that can clog equipment lines and damage pumps. Membrane technology is attractive for CO2 and H2S removal, because many membrane materials are very permeable to these species (enabling a high recovery of the acid gases without significant loss of pressure in the methane pipeline product gases), and because treatment can be accomplished using the high wellhead gas pressure as the driving force for the separation. A high natural gas recovery (>95%) can be achieved in multi-stage systems.

Cynara-NATCO produces hollow fiber modules for CO2 removal and has provided a membrane system (16 in. modules) for the natural gas sweetening in an offshore platform in the Thailand gulf (830,000 Nm3/h), which is the biggest membrane system for CO2 removal. The Natco Group has been awarded in 2008 a \$24.9 million contract to provide membrane separation technology and equipment to capture CO2 for re-injection in Bouri. Eni Oil Limited Libyan Branch operates the Bouri production platforms in the Mediterranean Sea, one of which will house the pre-treatment equipment and membrane systems intended to treat more than 160MM SCFD gas.

The Cynara(R) membrane system has been designed for the selective capture of CO2 from the gas stream for this retrofit project on the existing platform. Natco will design, engineer, and fabricate four membrane skids and related valving, as well as pretreatment equipment. The contract was awarded in May 2008.

The integration of membranes with other well-established separation processes in the chemical and petrochemical industries, , was considered in different works (Doshi, 1987; Choe *et al*, 1987; Dosh & Dolan, 1995; Bhide *et al,* 1998). Usually, the combination membranes / PSA is considered in H2 separation, while hybrid membranes + amine absorption are applied to the CO2 separation.

A comparison of the separation cost for the membrane process with diethanolamine (DEA) absorption showed that the membrane process is more economical for CO2 concentrations in the feed in the range 5–40 mol% (Bhide *et al*, 1998).

If membrane processes are not economically competitive because of the high H2S concentration in the feed, the separation cost could be significantly lowered by using hybrid membrane processes. In the block diagram of Fig.14 the membrane unit removes two-thirds of the CO2 and the amine plant removes the remaining. The combined plant is significantly less expensive than an all-amine or all-membrane plant.

A hybrid system (Cynara membranes + amine absorption) is operating since 1994 in Mallet (Texas, U.S.A.) to perform the bulk removal of CO2 from associated gas (90% CO2 and heavy hydrocarbons), before downstream treating. The membrane system offered a 30% reduction in operating cost when compared with a methyl diethanolamine (MDEA) system and significantly reduced the size of the subsequent operations (Blizzard *et al*, 2005).

An integrated system combining membrane permeation and pressure swing adsorption (PSA) has been developed for CO2/N2 gas separation (Esteves and Mota, 2007). By using the membrane as a pre-bulk separation unit and coupling it to the PSA, the separation performance of the hybrid scheme is enhanced with respect to that of the two stand-alone

Carbon dioxide removal from natural gas (natural gas sweetening) is mandatory to meet pipeline specifications (e.g., down to 2% vol. in U.S.A.), since CO2 reduces the heating value of natural gas, is corrosive and freezes at a relatively high temperature, forming blocks of dry ice that can clog equipment lines and damage pumps. Membrane technology is attractive for CO2 and H2S removal, because many membrane materials are very permeable to these species (enabling a high recovery of the acid gases without significant loss of pressure in the methane pipeline product gases), and because treatment can be accomplished using the high wellhead gas pressure as the driving force for the separation. A

Cynara-NATCO produces hollow fiber modules for CO2 removal and has provided a membrane system (16 in. modules) for the natural gas sweetening in an offshore platform in the Thailand gulf (830,000 Nm3/h), which is the biggest membrane system for CO2 removal. The Natco Group has been awarded in 2008 a \$24.9 million contract to provide membrane separation technology and equipment to capture CO2 for re-injection in Bouri. Eni Oil Limited Libyan Branch operates the Bouri production platforms in the Mediterranean Sea, one of which will house the pre-treatment equipment and membrane systems intended to

The Cynara(R) membrane system has been designed for the selective capture of CO2 from the gas stream for this retrofit project on the existing platform. Natco will design, engineer, and fabricate four membrane skids and related valving, as well as pretreatment equipment.

The integration of membranes with other well-established separation processes in the chemical and petrochemical industries, , was considered in different works (Doshi, 1987; Choe *et al*, 1987; Dosh & Dolan, 1995; Bhide *et al,* 1998). Usually, the combination membranes / PSA is considered in H2 separation, while hybrid membranes + amine

A comparison of the separation cost for the membrane process with diethanolamine (DEA) absorption showed that the membrane process is more economical for CO2 concentrations in

If membrane processes are not economically competitive because of the high H2S concentration in the feed, the separation cost could be significantly lowered by using hybrid membrane processes. In the block diagram of Fig.14 the membrane unit removes two-thirds of the CO2 and the amine plant removes the remaining. The combined plant is significantly

A hybrid system (Cynara membranes + amine absorption) is operating since 1994 in Mallet (Texas, U.S.A.) to perform the bulk removal of CO2 from associated gas (90% CO2 and heavy hydrocarbons), before downstream treating. The membrane system offered a 30% reduction in operating cost when compared with a methyl diethanolamine (MDEA) system and

An integrated system combining membrane permeation and pressure swing adsorption (PSA) has been developed for CO2/N2 gas separation (Esteves and Mota, 2007). By using the membrane as a pre-bulk separation unit and coupling it to the PSA, the separation performance of the hybrid scheme is enhanced with respect to that of the two stand-alone

significantly reduced the size of the subsequent operations (Blizzard *et al*, 2005).

high natural gas recovery (>95%) can be achieved in multi-stage systems.

treat more than 160MM SCFD gas.

The contract was awarded in May 2008.

absorption are applied to the CO2 separation.

the feed in the range 5–40 mol% (Bhide *et al*, 1998).

less expensive than an all-amine or all-membrane plant.

Fig. 14. Membrane/amine hybrid plant for the treatment of natural gas (Baker & Lokhandwala, 2008).

units. Instead of constant-composition regular feed, the PSA is fed with a mixture which is progressively enriched in the more adsorbed component during the pressurization and high-pressure adsorption steps of the cycle. This results in sharper concentration fronts. The hybrid has been applied successfully to the bulk separation of an 30:70 mol% CO2/N2 mixture over activated carbon. Process performance is reported in terms of product recovery and purity at cyclic steady state. Numerical simulations were validated by experimental work on a composite membrane and a laboratory-scale PSA unit.

MTR (Membrane Technology and Research, Inc.) has proposed an innovative retro-fit process for the post-combustion carbon capture and sequestration (CCS) of existing coalfired power plants (Merkel et al., 2010). According to preliminary calculations, a two stage membrane process (Figure 15) should be able to recover 90% of the CO2 released by the power plant at a competitive cost of 23\$/ton CO2 sequestrated. The selected lower cost configuration uses a blower upstream of the first membrane unit and a vacuum pump downstream in order to boost the driving force for the separation. At the same time, in the second membrane module the air to be fed to the boiler strips part of the CO2 in the N2 enriched flue gas from the first membrane module, thereby recycling CO2 and building a higher CO2 concentration for the first membrane module: as a consequence, the sequestered CO2 flux in the first membrane unit is enhanced.

MTR is now involved in a contract with the US Department of Energy for the construction of a membrane skid containing MTR's Polaris membranes capable of 90% CO2 capture from a 20 tons-of-CO2/day slipstream of coal fired flue gas. The skid will be operated during a 6 month field test at Arizona Public Service's Cholla power plant. (http://www.mtrinc.com/news.html#DOE14mil).

In the process industry the final choice of a separation process is the result of a balance between the economics, the desired purity and recovery of the product and other conditions and restrictions, such as the desired capacity and the composition of the feed and the possibility of integration with other processes. In this logic, the integration of commercial membrane separation units in the recovery of ethylene or propylene monomer from polyolefin resin degassing vent streams has been proposed by Baker *et al* (1998). A simplified flow scheme of the process that generates such streams is shown in Fig. 16(a).

Membrane Operations for Industrial Applications 561

recoverable monomers in a typical polymerization plant is very high. The placement of a

Membrane vapor gas separation systems have been installed worldwide. The main membrane applications in the petroleum industry are vapor recovery in tank farms and hydrogen recovery in refineries and chemical plants. Other developed applications of the membrane technology in refineries include solvent recovery in lube oil manufacturing (Max-DeWax developed jointly by Grace Division and Exxon Mobil) and aromatics removal

The development of innovative processes that follow the process intensification strategy for a sustainable industrial growth is critical to the production by non-polluting, defect-free and safe industrial processes. Membrane operations show a higher efficiency than conventional separation and reaction unit operations. They offer new options for the razionalization of innovative production cycles. Membrane engineering plays a crucial role in water desalination, in municipal water reuse (by MBR), in petrochemicals and in the field of gaseous separations. There are also some interesting opportunities to integrate membrane operations into existing industrial processes to achieve the benefits of process

Baker, R.W., Wijmans, J.G. and Kaschemekat, J.H. (1998). The design of membrane Vapor gas separation systems*. J. Membr. Sci.* Vol. 151 pp.55-62 ISSN 0376-7388 Baker, R. & Lokhandwala, K. (2008). Natural Gas Processing with Membranes: An Overview. *Ind. Eng. Chem. Res*. Vol. 47 No 7 pp.2109-2121 ISSN 0888-5885. Bhide, B.D.; Voskericyan, A.; Stem, S.A. (1998) Hybrid processes for the removal of acid gases from natural gas. *J. Membr. Sci.* Vol. *140* No 1 pp.27-49 ISSN 0376-7388 Blizzard, G.; Parro, G.; Hornback, K. (2005). Mallet gas processing facilities uses membranes to efficiently separate CO2. *Oil & Gas Journal* Vol. 103 No 14 pp 48-53. Bodzek, M. (2000) Membrane techniques in air cleaning. *Pol. J. Environm. Stud.* Vol. 9. No.1 ,

Cassano, A; Adzet, J; Molinari, R.; Buonomenna, M.G.; Roig, J; and Drioli, E. (2003)

the leather industry. *Water Research*. Vol. 37, pp. 2426-2434 ISSN 0043-1354. Charpentier, J.C. (2007). In the frame of globalization and sustainability, process

Choe, J.S.; Auvil, S.R.; Agrawal, R. (1987) Process for separating components of a gas

Cobb, C.; Schuster, D.; Beloff, B.; Tanzil, D. (2007). Benchmarking Sustainability.Chemical

Dhingra , S.S. & Marand, E. (1998) Mixed gas transport study through polymeric

Engineering Progress, Vol. 104, No.6, pp. 38-42, ISSN 0360-7275

membranes , *J. Membr.Sci.* Vol. 141 No 1 pp 45-63 ISSN 0376-7388

Membrane treatment by nanofiltration of exhausted vegetable tannin liquors from

intensification, a path to the future of chemical and process engineering (molecules into money). *Chemical Engineering and Processing: Process Intensification,*Vol. 134, No.

membrane to recover and recycle the gases is shown schematically in Fig. 16(b).

from gasoline.

**5. Conclusion** 

intensification.

**6. References** 

pp. 1-12 ISSN 1230-1485.

1-3, pp.84 92, ISSN 0255-2701

stream" US Patent 4,701,187

Fig. 15. Simplified flow diagram of a two-step counter-flow/sweep membrane process to capture and sequester CO2 in flue gas from a coal-fired power plant. The base-case membrane with a CO2 permeance of 1000 gpu and a CO2/N2 selectivity of 50 was used in the calculations (Merkel *et al*, 2010).

Fig. 16. A schematic diagram of monomer recovery and nitrogen recycle from polyolefin plant resin degassing operations: (a) conventional plant (b) membrane vapor gas separation system (Baker *et al*, 1998)

In a typical polyolefin polymerization plant, after polymerization, the polymer is removed to a low-pressure chamber. amounts of sorbed monomers and processing solvents, which must be removed before the polymer can be used. Therefore, the raw polymer is passed to large resin degassing bins through which nitrogen is circulated.

In the past, no economical method of separating these gases was available, so the stream was used as low grade fuel and the monomer content lost. The value of these potentially recoverable monomers in a typical polymerization plant is very high. The placement of a membrane to recover and recycle the gases is shown schematically in Fig. 16(b).

Membrane vapor gas separation systems have been installed worldwide. The main membrane applications in the petroleum industry are vapor recovery in tank farms and hydrogen recovery in refineries and chemical plants. Other developed applications of the membrane technology in refineries include solvent recovery in lube oil manufacturing (Max-DeWax developed jointly by Grace Division and Exxon Mobil) and aromatics removal from gasoline.

#### **5. Conclusion**

560 Advances in Chemical Engineering

Fig. 15. Simplified flow diagram of a two-step counter-flow/sweep membrane process to capture and sequester CO2 in flue gas from a coal-fired power plant. The base-case membrane with a CO2 permeance of 1000 gpu and a CO2/N2 selectivity of 50 was used in

Fig. 16. A schematic diagram of monomer recovery and nitrogen recycle from polyolefin plant resin degassing operations: (a) conventional plant (b) membrane vapor gas separation

In a typical polyolefin polymerization plant, after polymerization, the polymer is removed to a low-pressure chamber. amounts of sorbed monomers and processing solvents, which must be removed before the polymer can be used. Therefore, the raw polymer is passed to

In the past, no economical method of separating these gases was available, so the stream was used as low grade fuel and the monomer content lost. The value of these potentially

large resin degassing bins through which nitrogen is circulated.

the calculations (Merkel *et al*, 2010).

system (Baker *et al*, 1998)

The development of innovative processes that follow the process intensification strategy for a sustainable industrial growth is critical to the production by non-polluting, defect-free and safe industrial processes. Membrane operations show a higher efficiency than conventional separation and reaction unit operations. They offer new options for the razionalization of innovative production cycles. Membrane engineering plays a crucial role in water desalination, in municipal water reuse (by MBR), in petrochemicals and in the field of gaseous separations. There are also some interesting opportunities to integrate membrane operations into existing industrial processes to achieve the benefits of process intensification.

#### **6. References**


**22** 

*México* 

**Thermal Study on Phase Transitions of** 

*Facultad de Química Universidad Nacional Autónoma de México (UNAM)* 

The block copolymers are an exceptional kind of macromolecules constituted by two or more blocks of different homopolymer chains linked by covalent bonds. These polymeric materials have received much attention over past few years due in large part to their ability to selfassemble in the melted state or in a selective solvent inside a variety of ordered phases or welldefined structures of high regularity in size and shape with characteristic dimensions between 100 and 500 nanometres. These ordered phases and their structural modification are the key to many valuable physical properties which make block copolymers of great industrial and technological interest. The molecular self-assemble and formation of periodic phases in the block copolymers depend of the strength of interblock repulsion and composition, for example, mesoscopic studies of the poly(styrene)-poly(isoprene) (PS-PI) diblock copolymer, have demonstrated that this synthetic material, may generate a series of long-range ordered microdomains when exist a weak repulsion between the unlike monomers isoprene and styrene, as result, the PS-PI diblock copolymer chains tend to segregate below some critical temperature, but, as they are linked by covalent bonds, the phase separation on a macroscopic level is prevented, only a local microphase segregation occurs (Soto-Figueroa et al., 2005, Soto-Figueroa et al., 2007). The phase transition from homogeneous state of polymeric chains to an ordered state with periodic phases is called microphase separation transition (MST) or orderdisorder phase transition (ODT) (Leibler, 1980). The phase segregation and generation of ordered structures in the microscopic level of a diblock copolymer via an order-disorder phase

As result of microphase segregation process, the block copolymers can display ordered structures constituted by homopolymer domains that haves only mesoscopic dimensions corresponding to the size of singles blocks. The microphase separation leads to different classes of well-defined periodic structures in dependence on the ratio between the degrees of

**1. Introduction** 

transition is illustrated in Fig. 1.

Corresponding Authors

 \*

César Soto-Figueroa1, Luis Vicente2

*2Departamento de Física y Química Teórica,* 

*1Departamento de Ciencias Químicas, Facultad de Estudios Superiores Cuautitlán,* 

and María del Rosario Rodríguez-Hidalgo1,\*

*Universidad Nacional Autónoma de México (UNAM)* 

**Block Copolymers by Mesoscopic Simulation** 

Doshi, K.J. (1987) Enhanced gas separation process, US Patent 4,690,695, Union Carbide.


### **Thermal Study on Phase Transitions of Block Copolymers by Mesoscopic Simulation**

César Soto-Figueroa1, Luis Vicente2

and María del Rosario Rodríguez-Hidalgo1,\* *1Departamento de Ciencias Químicas, Facultad de Estudios Superiores Cuautitlán, Universidad Nacional Autónoma de México (UNAM) 2Departamento de Física y Química Teórica, Facultad de Química Universidad Nacional Autónoma de México (UNAM) México* 

#### **1. Introduction**

562 Advances in Chemical Engineering

Drioli, E; Romano, M. (2001). Progress and New Perspectives on Integrated Membrane

Drioli, E; Stankiewicz, A; Macedonio, F. (2011). Membrane Engineering in process intensification-An overview. *J. Membr. Sci.* Vol. 380, pp. 1-8, ISSN 0376-7388 Esteves, I.A.A.C., Mota, J.P.B. (2007). Hybrid Membrane/PSA Processes for CO2/N2 Separation. *Adsorption Science & Technology* Vol. 25, pp. 693-715 ISSN 0263-6174. Freeman , B; Yampolskii , Y; Pinnau , I. (2006) Eds., Materials Science of Membranes for Gas

Gao, W; Liang, H; Wang, L.; Chang, H.-q, , Li, G-b. (2011) Pilot Study of Integrated MF based

Hamed, O.A. (2006) Overview Of Hybrid Desalination Systems - Current Status And Future

Koros , W.J.; Mahajan , R.(2000) Pushing the limits on possibilities for large scale gas

Lin, L. & Guthrie, J.T. (2006) Novel oxygen - enhanced membrane assemblies for biosensors.

Macedonio, F.; Drioli, E. (2010). An exergetic analysis of a membrane desalination system.

Merkel, T.C.; Lin, H.; Wei, X.; Baker, R. (2010). Power plant post-combustion carbon dioxide

Nurdan, B.; Emre, K. (2010) Economic evaluation of alternative wastewater treatment plant

Spillman, R. (1989) Economics of gas separation by membranes. *Chem.Eng. Prog*. Vol. 85,

Stiegel, G. J. (1999) Mixed conducting ceramic membranes for gas separation and reaction.

United Nations General Assembly (1987). Report of the World Commission on Environment

Vansant, E.F. & Dewolfs, R. (1990) *Gas Separation Technology* , Elsevier , 1990 ISBN

Zhang, Y; Cao, C-Y; Feng, W-Y; Xue, G-X; Xu, M. (2011) Performance of a pilot scale

Zhu, X; Sun, S; Cong, Y; Yang, W. (2009) Operation of perovskite membrane under vacuum

membrane process for the concentration of effluent from alkaline peroxide mechanical pulping plants. *BioResources*. Vol.6, No.3, pp. 3044-3054, ISSN 19302126.

and elevated pressures for high-purity oxygen production. *J. Membr. Sci.* Vol.345

Water. *IPCBEE* Vol. 14, pp. 11-16, ISSN 2010-4618

Prospects. *Desalination* Vol. 186, pp. 207–214, ISSN 0011-9164

*J. Membr. Sci.* Vol. 278 No 1-2 pp 173-180. ISSN 0376-7388

*Membr. Technol*. Vol. 1999 No 110 pp 5-7 ISSN 0958-2118

and Development: Our Common Future

*Desalination* Vol. 261, pp. 293-299, ISSN 0011-9164

Operations for Sustainable Industrial Growth. *Ind. Eng. Chem. Res*. Vol. 40, pp.

and Vapor Separation , John Wiley and Sons, Ltd. , Chichester ISBN 9780470029039.

MBR and UF for Drinking Water Production by Treating Micropolluted Source

separation: which strategies? *J. Membr. Sci.* Vol. 175 No. 1 pp. 181-196 ISSN 0376-

capture: an opportunity for membranes, *J. Membr. Sci.* Vol. 359 No 1-2 pp 126-139.

options for pulp and paper industry. *Sci. Total. Environ*. Vol. 408, No.24, pp.6070-

Doshi, K.J. (1987) Enhanced gas separation process, US Patent 4,690,695, Union Carbide. Doshi, K.J.; Dolan, W.B. (1995) Process for the rejection of CO2 from natural gas", US Patent

5,411,72, UOP.

7388

ISSN 0376-7388

0444882308

6078, ISSN 0048-9697.

pp.41-62, ISSN 0360-7275.

pp 47-52 ISSN 0376-7388.

1277-1300, ISSN 0888-5885

The block copolymers are an exceptional kind of macromolecules constituted by two or more blocks of different homopolymer chains linked by covalent bonds. These polymeric materials have received much attention over past few years due in large part to their ability to selfassemble in the melted state or in a selective solvent inside a variety of ordered phases or welldefined structures of high regularity in size and shape with characteristic dimensions between 100 and 500 nanometres. These ordered phases and their structural modification are the key to many valuable physical properties which make block copolymers of great industrial and technological interest. The molecular self-assemble and formation of periodic phases in the block copolymers depend of the strength of interblock repulsion and composition, for example, mesoscopic studies of the poly(styrene)-poly(isoprene) (PS-PI) diblock copolymer, have demonstrated that this synthetic material, may generate a series of long-range ordered microdomains when exist a weak repulsion between the unlike monomers isoprene and styrene, as result, the PS-PI diblock copolymer chains tend to segregate below some critical temperature, but, as they are linked by covalent bonds, the phase separation on a macroscopic level is prevented, only a local microphase segregation occurs (Soto-Figueroa et al., 2005, Soto-Figueroa et al., 2007). The phase transition from homogeneous state of polymeric chains to an ordered state with periodic phases is called microphase separation transition (MST) or orderdisorder phase transition (ODT) (Leibler, 1980). The phase segregation and generation of ordered structures in the microscopic level of a diblock copolymer via an order-disorder phase transition is illustrated in Fig. 1.

As result of microphase segregation process, the block copolymers can display ordered structures constituted by homopolymer domains that haves only mesoscopic dimensions corresponding to the size of singles blocks. The microphase separation leads to different classes of well-defined periodic structures in dependence on the ratio between the degrees of

<sup>\*</sup> Corresponding Authors

Thermal Study on Phase Transitions of Block Copolymers by Mesoscopic Simulation 565

<sup>m</sup> AABB <sup>S</sup> n ln n ln

m A B c <sup>V</sup> H RT 

where <sup>A</sup> and B are volume fraction of A and B components of diblock copolymer, N=nA+nB denotes the total number of molecules or degree of polymerization and χ is the Flory-Huggins interaction parameter. The phase behaviour that exhibits the diblock copolymers during an order-disorder phase transition (microphase segregation) is controlled by both entropic and enthalpic interactions. The enthalpic interactions imply the repulsion magnitude between different species via Flory-Huggin's interaction parameter (χ), which represent the chemical incompatibility between different repetitive units and its magnitude is expressed by the type of monomers which integrate the diblock copolymer

1

the phase segregation behaviour is controlled by the value of *χ*, in this way, positives values of interaction parameter lead to incompatibility between different segments and the entropic interactions (ΔSm) appears to be mostly positives, this generates a positive heat of mixing and therefore a ΔGm > 0. Negative values of χ lead to homogeneous state

Whereas the entropic interactions involve the configurational and translation displacement of polymeric chains, and are regulated through the degree of polymerization *N*, architecture constrains and blocks composition (Bates & Fredrickson., 1990, 1999; Hamley, 1998; Balta-Calleja et al., 2000; Thomas et al., 1995). The microphase segregation degree in the diblock copolymers depends of the enthalpic-entropic balance represented by the reduced parameter *χN*, the ODT occurs at a critical value of *χN*, the melt phase behaviour is thus governed by composition and a reduced parameter (Leibler et al., 1980; Bates et al., 1990). Three segregation regimes have been identified by Matsen and Bates and have been defined depending on the extent of microphase segregation: the weak (*χN* ≈10), intermediate (*χN* >12-100) and strong segregation regimes (*χN* >100) (Matsen & Bates, 1996; Bates & Fredrickson, 1990). In the weak segregation regime, the volume fraction of one of the block varies sinusoidally about the average value generating the formation of ordered microphases, this regime is characterized by a diffuse interface between different components and is capable of to be modified by composition effect or for temperature effect (Hamley, 1998; Bates, 1991). In the intermediate segregation regime the composition profile becomes sharper generating ordered microphases with a narrow interface between blocks. The strong segregation regime due to saturation of the blocks composition contains essentially pure components; in this regime, the phase behaviour

(3)

S K n ln n ln m AABB (5)

H kT N <sup>m</sup> A B (6)

<sup>T</sup> (7)

(4)

R

The equations 3 and 4 can be rewritten as:

and has a strong dependence with the temperature:

and therefore a ΔGm< 0.

Fig. 1. The microphase segregation process occurs when the PS-PI diblock copolymers in melted state are transformed to a periodic inhomogeneous phase of ordered structures when the temperature diminishes.

polymerization of the component blocks. Periodic phases with specific morphologies such as: spherical, perforated layers, cylindrical, lamellar and Gyroid can be generated manipulating the composition or length of the component blocks (Strobl, 1997; Bates & Fredrickson., 1990).

The order-disorder and order-order phase transitions play an important role in the design and modification of new supramolecular materials and are the key to manipulate the physical and mechanicals properties in these polymeric materials.

In this chapter, attention has been concentrated on the order-disorder and order-order phase transitions that display the PS-PI diblock copolymers and in the mesoscopic simulations methods employed to explore the kinetics transformation pathway of well-defined ordered phases.

#### **1.1 Order-disorder phase transition (ODT)**

The order-disorder phase transition is a thermodynamic process controlled by enthalpic and entropic interactions, for example, when a diblock copolymer of type A-B are in a melted state by temperature effect, exhibits a homogeneous phase where all different block segments are completely miscible, in the reverse case when the different block segments are immiscible due to decrease of the temperature displays an heterogeneous state of ordered microphases: this conditions are describes by Gibbs free energy equation of mixing when ΔGm < 0 and ΔGm > 0 respectively.

$$
\Delta \mathbf{G\_m} = \mathbf{G\_{AB}} - (\mathbf{G\_A} + \mathbf{G\_B}) \tag{1}
$$

where GA, GB and GAB, denote the Gibbs free energy of A and B segments in separate states and the mixed state, respectively. Equation (1) in accordance with Flory-Huggins theory can be expressed also as a sum of two thermodynamic contributions (Flory, 1953):

$$
\Delta \mathbf{G\_m} = -\mathbf{T} \Delta \mathbf{S\_m} + \Delta \mathbf{H\_m} \tag{2}
$$

In this equation, ΔHm and TΔSm exhibits the enthalpic and entropic interactions of mixing at temperature T. The entropic and enthalpic interactions of mixing of two component segments of diblock copolymer are given by:

$$\frac{\Delta \mathbf{S}\_{\rm m}}{\overline{\mathbf{R}}} = \overline{\mathbf{n}}\_{\rm A} \ln \phi\_{\rm A} + \overline{\mathbf{n}}\_{\rm B} \ln \phi\_{\rm B} \tag{3}$$

$$
\Delta \mathbf{H}\_{\rm m} = \overline{\mathbf{R}} \mathbf{T} \frac{\mathbf{V}}{\overline{\mathbf{V}}\_{\rm c}} \chi \phi\_{\rm A} \phi\_{\rm B} \tag{4}
$$

The equations 3 and 4 can be rewritten as:

564 Advances in Chemical Engineering

Fig. 1. The microphase segregation process occurs when the PS-PI diblock copolymers in melted state are transformed to a periodic inhomogeneous phase of ordered structures

polymerization of the component blocks. Periodic phases with specific morphologies such as: spherical, perforated layers, cylindrical, lamellar and Gyroid can be generated manipulating the composition or length of the component blocks (Strobl, 1997; Bates & Fredrickson., 1990). The order-disorder and order-order phase transitions play an important role in the design and modification of new supramolecular materials and are the key to manipulate the

In this chapter, attention has been concentrated on the order-disorder and order-order phase transitions that display the PS-PI diblock copolymers and in the mesoscopic simulations methods employed to explore the kinetics transformation pathway of well-defined ordered phases.

The order-disorder phase transition is a thermodynamic process controlled by enthalpic and entropic interactions, for example, when a diblock copolymer of type A-B are in a melted state by temperature effect, exhibits a homogeneous phase where all different block segments are completely miscible, in the reverse case when the different block segments are immiscible due to decrease of the temperature displays an heterogeneous state of ordered microphases: this conditions are describes by Gibbs free energy equation of mixing when

where GA, GB and GAB, denote the Gibbs free energy of A and B segments in separate states and the mixed state, respectively. Equation (1) in accordance with Flory-Huggins theory can

In this equation, ΔHm and TΔSm exhibits the enthalpic and entropic interactions of mixing at temperature T. The entropic and enthalpic interactions of mixing of two component

be expressed also as a sum of two thermodynamic contributions (Flory, 1953):

G G (G G ) m AB A B (1)

G TS H m mm (2)

physical and mechanicals properties in these polymeric materials.

when the temperature diminishes.

**1.1 Order-disorder phase transition (ODT)** 

segments of diblock copolymer are given by:

ΔGm < 0 and ΔGm > 0 respectively.

$$\Delta \mathbf{S}\_{\rm m} = -\mathbf{K} \left[ \mathbf{n}\_{\rm A} \ln \phi\_{\rm A} - \mathbf{n}\_{\rm B} \ln \phi\_{\rm B} \right] \tag{5}$$

$$\mathbf{A}\mathbf{H}\_{\rm m} = \mathbf{k}\mathbf{T}\chi\mathbf{N}\phi\_{\rm A}\phi\_{\rm B} \tag{6}$$

where <sup>A</sup> and B are volume fraction of A and B components of diblock copolymer, N=nA+nB denotes the total number of molecules or degree of polymerization and χ is the Flory-Huggins interaction parameter. The phase behaviour that exhibits the diblock copolymers during an order-disorder phase transition (microphase segregation) is controlled by both entropic and enthalpic interactions. The enthalpic interactions imply the repulsion magnitude between different species via Flory-Huggin's interaction parameter (χ), which represent the chemical incompatibility between different repetitive units and its magnitude is expressed by the type of monomers which integrate the diblock copolymer and has a strong dependence with the temperature:

$$\mathbf{x} \approx \frac{1}{\mathbf{T}} \tag{7}$$

the phase segregation behaviour is controlled by the value of *χ*, in this way, positives values of interaction parameter lead to incompatibility between different segments and the entropic interactions (ΔSm) appears to be mostly positives, this generates a positive heat of mixing and therefore a ΔGm > 0. Negative values of χ lead to homogeneous state and therefore a ΔGm< 0.

Whereas the entropic interactions involve the configurational and translation displacement of polymeric chains, and are regulated through the degree of polymerization *N*, architecture constrains and blocks composition (Bates & Fredrickson., 1990, 1999; Hamley, 1998; Balta-Calleja et al., 2000; Thomas et al., 1995). The microphase segregation degree in the diblock copolymers depends of the enthalpic-entropic balance represented by the reduced parameter *χN*, the ODT occurs at a critical value of *χN*, the melt phase behaviour is thus governed by composition and a reduced parameter (Leibler et al., 1980; Bates et al., 1990). Three segregation regimes have been identified by Matsen and Bates and have been defined depending on the extent of microphase segregation: the weak (*χN* ≈10), intermediate (*χN* >12-100) and strong segregation regimes (*χN* >100) (Matsen & Bates, 1996; Bates & Fredrickson, 1990). In the weak segregation regime, the volume fraction of one of the block varies sinusoidally about the average value generating the formation of ordered microphases, this regime is characterized by a diffuse interface between different components and is capable of to be modified by composition effect or for temperature effect (Hamley, 1998; Bates, 1991). In the intermediate segregation regime the composition profile becomes sharper generating ordered microphases with a narrow interface between blocks. The strong segregation regime due to saturation of the blocks composition contains essentially pure components; in this regime, the phase behaviour

Thermal Study on Phase Transitions of Block Copolymers by Mesoscopic Simulation 567

The body centred cubic (BCC) phase or spherical phase is a classic structure of great thermodynamic stability that exhibits the diblock copolymers of type A-B. This ordered structure shows two specific symmetries: a four-fold symmetry ([100] projection) and the

Fig. 3. Snapshots of BCC phase of PS-PI copolymer: a) ordered microdomains of PS and PI chains (red and green regions represent the PS and PI microdomain respectively), b-c) projections [100] and [111] of poly(styrene) microdomains (the poly(isoprene) matrix has

The BCC phase of diblock copolymers have been studied extensively by both theoretical and experimental methods. For example, the theoretical studies of PS-PI diblock copolymer with linear architecture have demonstrated that this periodic phase can be generated in a composition interval of 0.1 to 0.19 (volume fraction of poly(styrene)) with a stable four-fold symmetry (Soto-Figueroa et al., 2005). The spherical phase of this diblock copolymer is known as a stable phase, however, can be transformed on other ordered structure via an

The hexagonally packed cylinders (HPC) phase is also considered as a classic phase of great thermodynamic stability due to its high packing. This ordered structure is characteristic of PS-PI diblock copolymers with linear architecture and can be generated via an orderdisorder phase transition in a composition interval of 0.2 to 0.26 (volume fraction of

Fig. 4. Snapshots of HPC phase of PS-PI diblock copolymer: a) cylindrical microphase constituted by ordered microdomains of poly(styrene) and poly(isoprene), b) representation of density isosurfaces of microdomains of poly(styrene) and c) poly(styrene) cylinders

The HPC phase displays a tetrahedral arrangement of epitaxially cylinders arranged in a hexagonal lattice. In this ordered phase, the cylindrical microdomains of poly(styrene)

hexagonal symmetry ([111] projection) of cubic array respectively, see Fig. 3.

**1.1.1 Body centred cubic phase** 

been removed for a better visualization).

**1.1.2 Hexagonal packed cylinders phase** 

order-order phase transition.

poly(styrene), see Fig. 4.

arranged in a hexagonal lattice.

depends largely on the copolymer composition. Within these segregation regimes it is possible to predict and modify the phase behaviour of block copolymers given *χ, N* and the segment length (block composition).

A general description of phase behaviour that exhibits the well-defined ordered structures of the PS-PI diblock copolymer via order-disorder phase transitions have been explored by mesoscopic simulation methods. When PS-PI diblock copolymer is in melted state, the polymeric chains assume the lowest free energy configuration, if the diblock copolymer is cooled, the repulsion magnitude expressed by reduced parameter χN increases, when the value of this parameter exceeds a certain value specific (χN≈10.5) for the system under consideration, well-defined periodic structures evolves in the disordered state (Soto-Figueroa et al., 2005).

When the PS-PI diblock copolymer has a symmetric composition (volume fraction of both components are the same) display an ordered phase with lamellar morphology (LAM), Fig. 2(a), however, if the volume fraction of a component increases in relative to other component (asymmetric copolymer), the interface tends to become curve. In this case, the conformational entropy loss of the majority component is too high. Therefore, to gain the conformational entropy, the chains of the majority component tend to expand along the direction parallel to interface. As a result, the PS/PI interface becomes convex towards the minority component. This interface curvature effect is more pronounced when the composition of the diblock copolymer is more asymmetric, Fig. 2(b). The PS-PI diblock copolymers with asymmetric compositions can generate a wide range of ordered structures such as the body-centred-cubic (BCC), hexagonal packed cylinders (HPC), ordered bicontinuous double diamond (OBDD or Gyroid) and lamellar (LAM) arrangement via an order-disorder phase transition process, Fig. 2(c-f).

Fig. 2. Schematic representation of composition effect in the phase behaviour of the PS-PI diblock copolymer: a) lamellar phase (symmetric copolymer), b) curvature of PS/PI interface by effect composition (asymmetric copolymer), c) spherical phase, d) cylindrical phase, e) lamellar phase and f) OBDD phase.

The ordered phases with well-defined morphologies that shows the block copolymers are usually describes by the volume fraction of one block (*f*), the overall degree of polymerization (N), and Flory-Huggins interaction parameter (χ) (Soto- Figueroa et al., 2007). The periodic phases of type BCC, HPC, Gyroid and LAM explored by means of mesoscopic simulations are described in the following sections.

#### **1.1.1 Body centred cubic phase**

566 Advances in Chemical Engineering

depends largely on the copolymer composition. Within these segregation regimes it is possible to predict and modify the phase behaviour of block copolymers given *χ, N* and the segment

A general description of phase behaviour that exhibits the well-defined ordered structures of the PS-PI diblock copolymer via order-disorder phase transitions have been explored by mesoscopic simulation methods. When PS-PI diblock copolymer is in melted state, the polymeric chains assume the lowest free energy configuration, if the diblock copolymer is cooled, the repulsion magnitude expressed by reduced parameter χN increases, when the value of this parameter exceeds a certain value specific (χN≈10.5) for the system under consideration, well-defined periodic structures evolves in the disordered state (Soto-

When the PS-PI diblock copolymer has a symmetric composition (volume fraction of both components are the same) display an ordered phase with lamellar morphology (LAM), Fig. 2(a), however, if the volume fraction of a component increases in relative to other component (asymmetric copolymer), the interface tends to become curve. In this case, the conformational entropy loss of the majority component is too high. Therefore, to gain the conformational entropy, the chains of the majority component tend to expand along the direction parallel to interface. As a result, the PS/PI interface becomes convex towards the minority component. This interface curvature effect is more pronounced when the composition of the diblock copolymer is more asymmetric, Fig. 2(b). The PS-PI diblock copolymers with asymmetric compositions can generate a wide range of ordered structures such as the body-centred-cubic (BCC), hexagonal packed cylinders (HPC), ordered bicontinuous double diamond (OBDD or Gyroid) and lamellar (LAM) arrangement via an

Fig. 2. Schematic representation of composition effect in the phase behaviour of the PS-PI diblock copolymer: a) lamellar phase (symmetric copolymer), b) curvature of PS/PI interface by effect composition (asymmetric copolymer), c) spherical phase, d) cylindrical

The ordered phases with well-defined morphologies that shows the block copolymers are usually describes by the volume fraction of one block (*f*), the overall degree of polymerization (N), and Flory-Huggins interaction parameter (χ) (Soto- Figueroa et al., 2007). The periodic phases of type BCC, HPC, Gyroid and LAM explored by means of

length (block composition).

Figueroa et al., 2005).

order-disorder phase transition process, Fig. 2(c-f).

phase, e) lamellar phase and f) OBDD phase.

mesoscopic simulations are described in the following sections.

The body centred cubic (BCC) phase or spherical phase is a classic structure of great thermodynamic stability that exhibits the diblock copolymers of type A-B. This ordered structure shows two specific symmetries: a four-fold symmetry ([100] projection) and the hexagonal symmetry ([111] projection) of cubic array respectively, see Fig. 3.

Fig. 3. Snapshots of BCC phase of PS-PI copolymer: a) ordered microdomains of PS and PI chains (red and green regions represent the PS and PI microdomain respectively), b-c) projections [100] and [111] of poly(styrene) microdomains (the poly(isoprene) matrix has been removed for a better visualization).

The BCC phase of diblock copolymers have been studied extensively by both theoretical and experimental methods. For example, the theoretical studies of PS-PI diblock copolymer with linear architecture have demonstrated that this periodic phase can be generated in a composition interval of 0.1 to 0.19 (volume fraction of poly(styrene)) with a stable four-fold symmetry (Soto-Figueroa et al., 2005). The spherical phase of this diblock copolymer is known as a stable phase, however, can be transformed on other ordered structure via an order-order phase transition.

#### **1.1.2 Hexagonal packed cylinders phase**

The hexagonally packed cylinders (HPC) phase is also considered as a classic phase of great thermodynamic stability due to its high packing. This ordered structure is characteristic of PS-PI diblock copolymers with linear architecture and can be generated via an orderdisorder phase transition in a composition interval of 0.2 to 0.26 (volume fraction of poly(styrene), see Fig. 4.

Fig. 4. Snapshots of HPC phase of PS-PI diblock copolymer: a) cylindrical microphase constituted by ordered microdomains of poly(styrene) and poly(isoprene), b) representation of density isosurfaces of microdomains of poly(styrene) and c) poly(styrene) cylinders arranged in a hexagonal lattice.

The HPC phase displays a tetrahedral arrangement of epitaxially cylinders arranged in a hexagonal lattice. In this ordered phase, the cylindrical microdomains of poly(styrene)

Thermal Study on Phase Transitions of Block Copolymers by Mesoscopic Simulation 569

The lamellar phase can be generated in the three segregation regimes (*χN* ≈10, *χN* >12-100 and *χN* >100) via order-disorder phase transitions and can displays flat and undulated

The ordered structures of PS-PI diblock copolymer represent under specific conditions the states with the lowest Gibbs free energy, these equilibrium phases can be classified through a phase diagram, Fig. 7. The diagram phase depicts the two regions associates constituted by a homogeneous state and an ordered state of five ordered phases different. Each equilibrium phase shows its predominance zone in terms of the volume fraction of poly(styrene) block

Fig. 7. Phase diagram of PS-PI diblock copolymer, the continuous curve describes the points of phase transition between the homogeneous state and the microphase-separated states. The ordered states split into different classes (BCC, HPC, G, HPL and LAM); the dashed lines show the predominance interval between the different types of ordered phases.

The segregation state in the phases diagram is controlled by the product χN, if χN is mayor than a critical value (typically of the order of 10.5), entropic factors dominate and the diblock copolymers exist in an ordered phases state. On the other hand, an order-todisorder transition takes place for large values of χN. The phase behaviour of each ordered structures kind on the phase diagram of poly(styrene)-poly(isoprene) diblock copolymer can be modified through order-order phase transitions, this will be tackled in

The diblock copolymers display a wide variety of classic phases with morphologies or defined structures such as: body centred cubic, hexagonal packed cylinders, hexagonal perforated layers, lamellar and the ordered bicontinuous double diamond phase, the morphology of these periodic structures can be controlled and modified by two different routes: 1) order-disorder phase transition and 2) thermotropic order-order phase transition. In the order-disorder phase transition as was mentioned in the previous section, the phase behaviour in the block copolymers is governed by three experimentally controllable factors during the synthesis process: i) the overall degree of polymerization, ii) architecture of block

**2. Order-order phase transition in block copolymers** 

interfaces.

and reduced parameter χN.

the following section.

**1.1.5 Phase diagram of PS-PI diblock copolymer** 

exhibits an hexagonal arrangement of undulated microdomains immersed in a poly(isoprene) matrix.

#### **1.1.3 Ordered bicontinuous double diamond phase**

The ordered bicontinuous double diamond (OBDD) phase (or Gyroid phase) is the ordered structures more complex that exhibits some diblock copolymers of type A-B. For example, The PS-PI diblock copolymer can generate this ordered structure via an order-disorder phase transition, see Fig. 5.

Fig. 5. Snapshots of OBDD arrangement of PS-PI diblock copolymer obtained by mesoscopic simulations: a) ordered microdomains of poly(styrene) and poly(isoprene) chains, and b) representation of density isosurfaces of microdomains of poly(styrene) (Soto-Figueroa et al., 2008).

The OBDD phase has a tetrahedral arrangement of epitaxially cylinders interconnected by channels of Ia3d symmetry (Hajduk et al., 1995). The OBDD phase of PS-PI copolymer exists only in a narrow interval of values (0.37 to 0.4 volume fraction of poly(styrene)) between the regimes of the perforated layer and lamellar phases. The Gyroid arrangement is known as a stable phase, however, can be transformed on other ordered structure via an order-order phase transition.

#### **1.1.4 Lamellar phase**

Ordered phase with lamellar morphology are characteristics of block copolymers of symmetric and asymmetric composition. This periodic phase is constituted of alternating layers of different homopolymer microdomains separates by flat interfaces, see Fig 6. For the case of PS-PI diblock copolymer with linear architecture, this displays the lamellar arrangement in a specific predominance interval between 0.47 to 0.68 volume fraction of poly(styrene).

Fig. 6. Snapshots of lamellar phase of PS-PI diblock copolymer, the lamellar structure exhibits undulated microdomains of poly(styrene) and poly(isoprene): a) ordered microdomains of poly(styrene) and poly(isoprene) and b) representation of density isosurfaces of LAM phase.

The lamellar phase can be generated in the three segregation regimes (*χN* ≈10, *χN* >12-100 and *χN* >100) via order-disorder phase transitions and can displays flat and undulated interfaces.

#### **1.1.5 Phase diagram of PS-PI diblock copolymer**

568 Advances in Chemical Engineering

exhibits an hexagonal arrangement of undulated microdomains immersed in a

The ordered bicontinuous double diamond (OBDD) phase (or Gyroid phase) is the ordered structures more complex that exhibits some diblock copolymers of type A-B. For example, The PS-PI diblock copolymer can generate this ordered structure via an order-disorder

Fig. 5. Snapshots of OBDD arrangement of PS-PI diblock copolymer obtained by mesoscopic simulations: a) ordered microdomains of poly(styrene) and poly(isoprene) chains, and b) representation of density isosurfaces of microdomains of poly(styrene) (Soto-Figueroa et al.,

The OBDD phase has a tetrahedral arrangement of epitaxially cylinders interconnected by channels of Ia3d symmetry (Hajduk et al., 1995). The OBDD phase of PS-PI copolymer exists only in a narrow interval of values (0.37 to 0.4 volume fraction of poly(styrene)) between the regimes of the perforated layer and lamellar phases. The Gyroid arrangement is known as a stable phase, however, can be transformed on other ordered structure via an order-order

Ordered phase with lamellar morphology are characteristics of block copolymers of symmetric and asymmetric composition. This periodic phase is constituted of alternating layers of different homopolymer microdomains separates by flat interfaces, see Fig 6. For the case of PS-PI diblock copolymer with linear architecture, this displays the lamellar arrangement in a specific predominance interval between 0.47 to 0.68 volume fraction of

Fig. 6. Snapshots of lamellar phase of PS-PI diblock copolymer, the lamellar structure exhibits undulated microdomains of poly(styrene) and poly(isoprene): a) ordered microdomains of poly(styrene) and poly(isoprene) and b) representation of density

poly(isoprene) matrix.

phase transition, see Fig. 5.

2008).

phase transition.

poly(styrene).

**1.1.4 Lamellar phase** 

isosurfaces of LAM phase.

**1.1.3 Ordered bicontinuous double diamond phase** 

The ordered structures of PS-PI diblock copolymer represent under specific conditions the states with the lowest Gibbs free energy, these equilibrium phases can be classified through a phase diagram, Fig. 7. The diagram phase depicts the two regions associates constituted by a homogeneous state and an ordered state of five ordered phases different. Each equilibrium phase shows its predominance zone in terms of the volume fraction of poly(styrene) block and reduced parameter χN.

Fig. 7. Phase diagram of PS-PI diblock copolymer, the continuous curve describes the points of phase transition between the homogeneous state and the microphase-separated states. The ordered states split into different classes (BCC, HPC, G, HPL and LAM); the dashed lines show the predominance interval between the different types of ordered phases.

The segregation state in the phases diagram is controlled by the product χN, if χN is mayor than a critical value (typically of the order of 10.5), entropic factors dominate and the diblock copolymers exist in an ordered phases state. On the other hand, an order-todisorder transition takes place for large values of χN. The phase behaviour of each ordered structures kind on the phase diagram of poly(styrene)-poly(isoprene) diblock copolymer can be modified through order-order phase transitions, this will be tackled in the following section.

#### **2. Order-order phase transition in block copolymers**

The diblock copolymers display a wide variety of classic phases with morphologies or defined structures such as: body centred cubic, hexagonal packed cylinders, hexagonal perforated layers, lamellar and the ordered bicontinuous double diamond phase, the morphology of these periodic structures can be controlled and modified by two different routes: 1) order-disorder phase transition and 2) thermotropic order-order phase transition.

In the order-disorder phase transition as was mentioned in the previous section, the phase behaviour in the block copolymers is governed by three experimentally controllable factors during the synthesis process: i) the overall degree of polymerization, ii) architecture of block

Thermal Study on Phase Transitions of Block Copolymers by Mesoscopic Simulation 571

Experimental and theoretical studies have confirmed that HPC phase develops transient metastable phases during thermal heating process (Kimishima et al., 2000; Krishnamoorti et al., 2000; Rodríguez-Hidalgo et al., 2009). The order-order transition pathway from HPC to

Fig. 8. Snapshots of order-order phase transition process from HPC phase to spherical arrangement monitored by means of mesoscopic simulations: a) cylindrical phase, b) undulation process of poly(styrene) microdomains by temperature effects, c) breakdown of cylindrical microdomains of poly(styrene) and d) formation and stabilization of body-

Three transitory stages are typical of OOT process from HPC to cylindrical phase: (Stage 1) undulation process of cylindrical microdomains of poly(styrene) in the poly(isoprene) matrix. In this transitory stage the thermal heating induce the thermodynamic instability of HPC phase, where the cylindrical microdomains and the PS/PI interface becomes less rigid due to anisotropic composition fluctuations, this generates the undulation of the poly(styrene) microdomain in the poly(isoprene) matrix. (Stage 2) breakdown of cylindrical microdomains of poly(styrene) by temperature effects. When anisotropic composition fluctuations reach a critical point of thermodynamic instability the cylindrical microdomains become unstable and therefore they are broken in ellipsoids, (Stage 3) formation and stabilization of BCC phase. In this thermally induced stage, the ellipsoids microdomains generated by breakdown of cylindrical microdomains of poly(styrene) evolve to an equilibrium state where the uniform spherical microdomains are stabilized into bodycentred-cubic arrangement. The order-order phase transition from HPC to spherical phase is a thermoreversible process. The inverse process from BCC state to HPC arrangement involves deformation and elongation of spheres into ellipsoids and coalescence of ellipsoids into the cylindrical microdomains. This process is driven by thermodynamic instability of

The Gyroid phase as was mentioned in the section 1.1.3 exhibits a tetrahedral arrangement of epitaxially cylinders interconnected by channels of Ia3d symmetry and is capable of suffering order-order phase transitions when it is submitted to thermal heating cycles. The poly(styrene)-poly(isoprene) diblock copolymer displays this ordered phase in a narrow interval of specific composition. The thermally-induced phase transition from Gyroid to lamellar phase in PS-PI diblock copolymer was explored in detail by mesoscopic simulation methods (Soto-Figueroa et al., 2008). The phase transformation during thermal annealed proceed in several stages via the generation two metastable phases (HPL and cylindrical). The order-order phase transition process from Gyroid to lamellar phase is sketched

centred-cubic arrangement (BCC phase) (Soto-Figueroa et al., 2008).

the spherical interface caused by decrease of temperature.

**2.2 Order-order phase transition of Gyroid structure** 

schematically in Fig. 9.

spherical phase is sketched schematically in Fig. 8.

copolymer, and iii) the interaction parameter between components blocks, of this thermodynamic process can emerge periodic structures with well-defined morphologies (Soto-Figueroa et al., 2005, Soto-Figueroa et al. 2007).

Other way to modify the phase behaviour in the block copolymers is through the temperature; thermally induced phase transitions have the potential to promote the kinetic control in these synthetic materials. Leibler was the first in predicting the phase transition between different ordered structures by temperature effect (Leibler et al., 1980). The thermally induced phase transformations are governed for anisotropic composition fluctuation effects (Ryu et al., 1999), when an ordered phase of specific morphology is subject to a thermal heating process, the homopolymer chains into the ordered microdomains exhibit thermodynamic instability by temperature effect and become less rigid, this entropic process promotes the polymeric chains movement (composition fluctuations) into the homopolymer microdomains and modifies with the time the shape of ordered phase.

Three segregation regimes have been defined to explain the extent of microphase segregation and the thermodynamic stability in the classical phases that exhibits the diblock copolymers, these segregation regimes are : weak regime (*χN* ≈10), intermediate regime (*χN* >10-100) and strong segregation regime (*χN* >100) (section 1.1).

Our interest is concentrated in the weak segregation regime, because in this predominance zone, the diblock copolymers are characterized by a widened interface due to enhanced phase mixing. In the vicinity to this regime, thermotropic phase transition between different kinds of ordered phases can be generating (Bates et al., 1990; Matsen et al., 1996).

The transition from one ordered state to another is nowadays denominated as an orderorder phase transition (OOT) (Sakurai et al., 1993; Kim et al., 1998; Almadal et al., 1992; Sakurai et al., 1996; Sakamato et al., 1997; Modi et al., 1999). In this way, the ordered phases that exhibit the diblock copolymers inside weak segregation regime can exhibit order-order phase transitions when are subjects to thermal heating cycles.

In the past decade the order-order phase transitions that display the classical phases of diblock copolymers (BCC, HPC, Gyroid and Lamellar) have been investigated by both experimental and theoretical studies. The order-order phase transition is thermoreversible process that develops transient metastable states during phase transformation. The orderorder phase transition that exhibits the classical phases with specific morphology such as Gyroid, cylindrical, and lamellar are described next.

#### **2.1 Order-order phase transition of HPC structure**

The ordered phase of hexagonal packed cylinders also known as cylindrical structure is characteristic of diblock and triblock copolymers, for example, the poly(styrene) poly(isoprene) diblock copolymer with linear architecture can generate this ordered arrangement in a specific composition of 0.2/0.8 (volume fraction ) of PS/PI via an orderdisorder phase transition. When the HPC phase of this diblock copolymer is subject to thermal heating cycles, exhibits an order-order phase transition to body-centred-cubic (BCC) structure (or spherical phase). The order-order transition pathway that exhibits the HPC to BCC phase was recently reported (Modi et al., 1999; Krishnamoorti et al., 2000).

copolymer, and iii) the interaction parameter between components blocks, of this thermodynamic process can emerge periodic structures with well-defined morphologies

Other way to modify the phase behaviour in the block copolymers is through the temperature; thermally induced phase transitions have the potential to promote the kinetic control in these synthetic materials. Leibler was the first in predicting the phase transition between different ordered structures by temperature effect (Leibler et al., 1980). The thermally induced phase transformations are governed for anisotropic composition fluctuation effects (Ryu et al., 1999), when an ordered phase of specific morphology is subject to a thermal heating process, the homopolymer chains into the ordered microdomains exhibit thermodynamic instability by temperature effect and become less rigid, this entropic process promotes the polymeric chains movement (composition fluctuations) into the homopolymer microdomains and modifies with the time the shape of

Three segregation regimes have been defined to explain the extent of microphase segregation and the thermodynamic stability in the classical phases that exhibits the diblock copolymers, these segregation regimes are : weak regime (*χN* ≈10), intermediate regime (*χN*

Our interest is concentrated in the weak segregation regime, because in this predominance zone, the diblock copolymers are characterized by a widened interface due to enhanced phase mixing. In the vicinity to this regime, thermotropic phase transition between different

The transition from one ordered state to another is nowadays denominated as an orderorder phase transition (OOT) (Sakurai et al., 1993; Kim et al., 1998; Almadal et al., 1992; Sakurai et al., 1996; Sakamato et al., 1997; Modi et al., 1999). In this way, the ordered phases that exhibit the diblock copolymers inside weak segregation regime can exhibit order-order

In the past decade the order-order phase transitions that display the classical phases of diblock copolymers (BCC, HPC, Gyroid and Lamellar) have been investigated by both experimental and theoretical studies. The order-order phase transition is thermoreversible process that develops transient metastable states during phase transformation. The orderorder phase transition that exhibits the classical phases with specific morphology such as

The ordered phase of hexagonal packed cylinders also known as cylindrical structure is characteristic of diblock and triblock copolymers, for example, the poly(styrene) poly(isoprene) diblock copolymer with linear architecture can generate this ordered arrangement in a specific composition of 0.2/0.8 (volume fraction ) of PS/PI via an orderdisorder phase transition. When the HPC phase of this diblock copolymer is subject to thermal heating cycles, exhibits an order-order phase transition to body-centred-cubic (BCC) structure (or spherical phase). The order-order transition pathway that exhibits the HPC to BCC phase was recently reported (Modi et al., 1999; Krishnamoorti et al., 2000).

kinds of ordered phases can be generating (Bates et al., 1990; Matsen et al., 1996).

(Soto-Figueroa et al., 2005, Soto-Figueroa et al. 2007).

>10-100) and strong segregation regime (*χN* >100) (section 1.1).

phase transitions when are subjects to thermal heating cycles.

Gyroid, cylindrical, and lamellar are described next.

**2.1 Order-order phase transition of HPC structure** 

ordered phase.

Experimental and theoretical studies have confirmed that HPC phase develops transient metastable phases during thermal heating process (Kimishima et al., 2000; Krishnamoorti et al., 2000; Rodríguez-Hidalgo et al., 2009). The order-order transition pathway from HPC to spherical phase is sketched schematically in Fig. 8.

Fig. 8. Snapshots of order-order phase transition process from HPC phase to spherical arrangement monitored by means of mesoscopic simulations: a) cylindrical phase, b) undulation process of poly(styrene) microdomains by temperature effects, c) breakdown of cylindrical microdomains of poly(styrene) and d) formation and stabilization of bodycentred-cubic arrangement (BCC phase) (Soto-Figueroa et al., 2008).

Three transitory stages are typical of OOT process from HPC to cylindrical phase: (Stage 1) undulation process of cylindrical microdomains of poly(styrene) in the poly(isoprene) matrix. In this transitory stage the thermal heating induce the thermodynamic instability of HPC phase, where the cylindrical microdomains and the PS/PI interface becomes less rigid due to anisotropic composition fluctuations, this generates the undulation of the poly(styrene) microdomain in the poly(isoprene) matrix. (Stage 2) breakdown of cylindrical microdomains of poly(styrene) by temperature effects. When anisotropic composition fluctuations reach a critical point of thermodynamic instability the cylindrical microdomains become unstable and therefore they are broken in ellipsoids, (Stage 3) formation and stabilization of BCC phase. In this thermally induced stage, the ellipsoids microdomains generated by breakdown of cylindrical microdomains of poly(styrene) evolve to an equilibrium state where the uniform spherical microdomains are stabilized into bodycentred-cubic arrangement. The order-order phase transition from HPC to spherical phase is a thermoreversible process. The inverse process from BCC state to HPC arrangement involves deformation and elongation of spheres into ellipsoids and coalescence of ellipsoids into the cylindrical microdomains. This process is driven by thermodynamic instability of the spherical interface caused by decrease of temperature.

#### **2.2 Order-order phase transition of Gyroid structure**

The Gyroid phase as was mentioned in the section 1.1.3 exhibits a tetrahedral arrangement of epitaxially cylinders interconnected by channels of Ia3d symmetry and is capable of suffering order-order phase transitions when it is submitted to thermal heating cycles. The poly(styrene)-poly(isoprene) diblock copolymer displays this ordered phase in a narrow interval of specific composition. The thermally-induced phase transition from Gyroid to lamellar phase in PS-PI diblock copolymer was explored in detail by mesoscopic simulation methods (Soto-Figueroa et al., 2008). The phase transformation during thermal annealed proceed in several stages via the generation two metastable phases (HPL and cylindrical). The order-order phase transition process from Gyroid to lamellar phase is sketched schematically in Fig. 9.

Thermal Study on Phase Transitions of Block Copolymers by Mesoscopic Simulation 573

Nevertheless, the lamellar phase generated of block copolymers of asymmetric composition, for example of 0.45/0.55 of PS/PI (volume fraction), it can exhibit a phase transformation to HPL arrangement when is subject to thermal heating via an order-order phase transition process (Mani et al., 2000). The order-order phase transition from lamellar to HPL phase is

Fig. 11. Order-order phase transition stage from lamellar phase to HPL arrangement: a) undulation process of lamellar microdomains, b) interconnection of parallel microdomains

The order-order phase transition process of this ordered arrangement is governed by the thermodynamic instability between the components microdomains, generated by thermal heating, where the anisotropic composition fluctuations play an important role in the phase transformation. In the initial stage the lamellar microdomains exhibit an undulation process due to thermal heating, Fig. 11(a), with the temperature increase, the lamellar phase becomes thermodynamically unstable, in order to reach a thermodynamic stability of minimal energy, the lamellar alternate microdomains of poly(styrene) are interconnected by means of narrow microdomains (parallels to PS/PI interface), finally the HPL phase evolve

**3. Thermal study of double directionality of order-order phase transitions of hexagonally perforated layers (HPL) phases by mesoscopic simulation** 

During the past two decades have been reported theoretical and experimental studies of diblock copolymers, where the ODT and OOT transitions play a main role in the structure control and consequently in the physical properties control of these polymeric materials (Kimishima et al., 2000; Kim et al., 2006; Bodycomb et al., 2000; Krishanamoorti et al., 2000; Court et al., 2006; Soto-Figueroa et al., 2008). It is well-known that the poly(styrene) poly(isoprene) diblock copolymer exhibits a wide variety of classical phases with specific morphologies such as: BCC, HPC, OBDD, LAM and HPL. The order-order phase transitions that exhibit this ordered structures with definite morphologies have been explored by Soto-Figueroa and Rodríguez-Hidalgo, they have confirmed the OOT´s between HPC to BCC, Gyroid to LAM microphase through mesoscopic simulations (Soto-Figueroa et al., 2007, 2008; Rodríguez-Hidalgo et al., 2009), although the OOT´s have been well investigated for the majority of classical phases, the dynamic transformation of HPL phase has not been investigated yet in detail. The HPL phase exhibits a double directionality of order-order phase transition when is subject to thermal process. Experimental evidences suggest that HPL phase of the PS-PI copolymer can evolve to a cylindrical structure and to a Gyroid

The double directionality of order-order phase transitions that exhibits the classical phase of PS-PI copolymer is a topic interesting to control the morphology and physical

of poly(styrene) and c) formation and stabilization of HPL arrangement.

in an energetic equilibrium state (Soto-Figueroa et al., 2007), see Fig. 11(b-c).

structure by temperature effect (Park et al., 2005; You et al., 2007).

showed in Fig. 11.

Fig. 9. Transitory stages and metastable phases generated during the order-order phase transition from Gyroid phase to lamellar arrangement: a) undulation of interconnected microdomains of poly(styrene) into the poly(isoprene) matrix, due thermal heating, b) breakdown of interconnected microdomains of Gyroid phase, c) formation of HPL metastable phase, d) formation of cylindrical metastable phase and e) lamellar phase.

When the Gyroid phase of PS-PI copolymer is subject to thermal heating cycles, the interconnected microdomains of poly(styrene) display an undulation process due to anisotropic composition fluctuations, Fig 9(a), with the increase of temperature the anisotropic composition fluctuations induce the breakdown of side interconnections in the Gyroid arrangement, Fig. 9(b), generating the first metastable phase with HPL arrangement, Fig. 9(c). The HPL metastable phase evolves later to cylindrical arrangement by temperature effect, in this transient stage the interconnections of perforated layer microdomains diminish their volume up to the breakdown of interconnections in the HPL arrangement and the formation of cylindrical phase (second metastable phase), Fig 9(d). Finally, the cylindrical metastable phase also changes over time due to anisotropic composition fluctuations. The cylindrical microdomains in this metastable phase are thermodynamically unstable, in order to reach a thermodynamic stability of minimal energy, the cylinders microdomains are joined together to evolve into undulating lamellar phase, Fig. 9(e), the lamellar phase consisting of alternating layers of PS and PI microdomains.

#### **2.3 Order-order phase transition of lamellar structure**

The lamellar phase is the ordered structure more simple that exhibit the multiblock copolymers is considered a classical phase of great thermodynamic stability. The lamellar phase consisting of alternating layers of different homopolymer microdomains. The PS-PS diblock copolymer exhibits this ordered arrangement in a specific composition interval (see section 1.1.4). When the lamellar phase of this diblock copolymer with an equivalent composition between their constituent blocks (0.5 volume fraction of poly(styrene) and poly(isoprene)) is subject to thermal heating cycles, it does not generate an order-order phase transition, the lamellar arrangement in this case evolve to a homogeneous state (melted), see Fig. 10.

Fig. 10. Snapshots of thermal heating of lamellar phase (diblock copolymer of symmetric composition): a-b) equilibrium phase of lamellar structure (representation of diblock chains and surface isodensities of PS and PI microdomains) and c-d) homogeneous state after of thermal heating (melted phase).

Fig. 9. Transitory stages and metastable phases generated during the order-order phase transition from Gyroid phase to lamellar arrangement: a) undulation of interconnected microdomains of poly(styrene) into the poly(isoprene) matrix, due thermal heating, b) breakdown of interconnected microdomains of Gyroid phase, c) formation of HPL metastable phase, d) formation of cylindrical metastable phase and e) lamellar phase.

When the Gyroid phase of PS-PI copolymer is subject to thermal heating cycles, the interconnected microdomains of poly(styrene) display an undulation process due to anisotropic composition fluctuations, Fig 9(a), with the increase of temperature the anisotropic composition fluctuations induce the breakdown of side interconnections in the Gyroid arrangement, Fig. 9(b), generating the first metastable phase with HPL arrangement, Fig. 9(c). The HPL metastable phase evolves later to cylindrical arrangement by temperature effect, in this transient stage the interconnections of perforated layer microdomains diminish their volume up to the breakdown of interconnections in the HPL arrangement and the formation of cylindrical phase (second metastable phase), Fig 9(d). Finally, the cylindrical metastable phase also changes over time due to anisotropic composition fluctuations. The cylindrical microdomains in this metastable phase are thermodynamically unstable, in order to reach a thermodynamic stability of minimal energy, the cylinders microdomains are joined together to evolve into undulating lamellar phase, Fig. 9(e), the lamellar phase consisting of alternating

The lamellar phase is the ordered structure more simple that exhibit the multiblock copolymers is considered a classical phase of great thermodynamic stability. The lamellar phase consisting of alternating layers of different homopolymer microdomains. The PS-PS diblock copolymer exhibits this ordered arrangement in a specific composition interval (see section 1.1.4). When the lamellar phase of this diblock copolymer with an equivalent composition between their constituent blocks (0.5 volume fraction of poly(styrene) and poly(isoprene)) is subject to thermal heating cycles, it does not generate an order-order phase transition, the lamellar arrangement in this case evolve to a homogeneous state

Fig. 10. Snapshots of thermal heating of lamellar phase (diblock copolymer of symmetric composition): a-b) equilibrium phase of lamellar structure (representation of diblock chains and surface isodensities of PS and PI microdomains) and c-d) homogeneous state after of

layers of PS and PI microdomains.

(melted), see Fig. 10.

thermal heating (melted phase).

**2.3 Order-order phase transition of lamellar structure** 

Nevertheless, the lamellar phase generated of block copolymers of asymmetric composition, for example of 0.45/0.55 of PS/PI (volume fraction), it can exhibit a phase transformation to HPL arrangement when is subject to thermal heating via an order-order phase transition process (Mani et al., 2000). The order-order phase transition from lamellar to HPL phase is showed in Fig. 11.

Fig. 11. Order-order phase transition stage from lamellar phase to HPL arrangement: a) undulation process of lamellar microdomains, b) interconnection of parallel microdomains of poly(styrene) and c) formation and stabilization of HPL arrangement.

The order-order phase transition process of this ordered arrangement is governed by the thermodynamic instability between the components microdomains, generated by thermal heating, where the anisotropic composition fluctuations play an important role in the phase transformation. In the initial stage the lamellar microdomains exhibit an undulation process due to thermal heating, Fig. 11(a), with the temperature increase, the lamellar phase becomes thermodynamically unstable, in order to reach a thermodynamic stability of minimal energy, the lamellar alternate microdomains of poly(styrene) are interconnected by means of narrow microdomains (parallels to PS/PI interface), finally the HPL phase evolve in an energetic equilibrium state (Soto-Figueroa et al., 2007), see Fig. 11(b-c).

#### **3. Thermal study of double directionality of order-order phase transitions of hexagonally perforated layers (HPL) phases by mesoscopic simulation**

During the past two decades have been reported theoretical and experimental studies of diblock copolymers, where the ODT and OOT transitions play a main role in the structure control and consequently in the physical properties control of these polymeric materials (Kimishima et al., 2000; Kim et al., 2006; Bodycomb et al., 2000; Krishanamoorti et al., 2000; Court et al., 2006; Soto-Figueroa et al., 2008). It is well-known that the poly(styrene) poly(isoprene) diblock copolymer exhibits a wide variety of classical phases with specific morphologies such as: BCC, HPC, OBDD, LAM and HPL. The order-order phase transitions that exhibit this ordered structures with definite morphologies have been explored by Soto-Figueroa and Rodríguez-Hidalgo, they have confirmed the OOT´s between HPC to BCC, Gyroid to LAM microphase through mesoscopic simulations (Soto-Figueroa et al., 2007, 2008; Rodríguez-Hidalgo et al., 2009), although the OOT´s have been well investigated for the majority of classical phases, the dynamic transformation of HPL phase has not been investigated yet in detail. The HPL phase exhibits a double directionality of order-order phase transition when is subject to thermal process. Experimental evidences suggest that HPL phase of the PS-PI copolymer can evolve to a cylindrical structure and to a Gyroid structure by temperature effect (Park et al., 2005; You et al., 2007).

The double directionality of order-order phase transitions that exhibits the classical phase of PS-PI copolymer is a topic interesting to control the morphology and physical

Thermal Study on Phase Transitions of Block Copolymers by Mesoscopic Simulation 575

ij

 

B ij

In this way, a connection exists between the molecular character of the coarse-grained model and the DPD parameter. The parameter aij is given in terms of kBT (DPD reduced units). Equation (6) implies that, if the species are compatible, ij 0 and therefore, aij = 25. Established procedures for mapping between the DPD and physical scales and for choosing the system temperature are not yet available. We therefore use Flory–Huggins theory (through the χ(T) dependence) to introduce the temperature into the DPD simulations. To

> <sup>m</sup> <sup>2</sup> 1 2

where Vm and δ are the mean molar volume and solubility parameter, respectively. In the DPD simulation, the dynamic behaviour of order-order phase transition of the HPL structure is followed by integration of the equations of motion of each species using a modified version of the Verlet algorithm (Verlet, 1967). The integration of the equations of motion for each particle generates a trajectory through the system's phase, from which thermodynamic observables may be constructed by suitable averaging. Based on this information, the orderorder phase transition can be observed. In this algorithm, the forces are still updated once per

The molecular structure of PS-PI copolymer was built by means of the polymer builder module of Accelrys (Accelrys, 2006). The polymeric molecule with linear architecture

k Tχ (T)

0.306

ij ij ij ij ij <sup>F</sup> γω (r )( . ) ˆ ˆ **vrr** (11)

<sup>ω</sup> (r ) <sup>ˆ</sup> **<sup>r</sup>** (12)

<sup>R</sup> F forces, respectively, and a ij ij is the maximum

(13)

<sup>V</sup> χ δδ RT (14)

(10)

ij (r ) and

<sup>C</sup> ij ij ij ij ij

D D

R R ij ij ij ij F 

where ij <sup>i</sup> <sup>j</sup> r rr , ij ij r r /r <sup>ˆ</sup> , is the dissipation strength, is the noise strength, <sup>D</sup>

ij ii

a a

calculate the aij values of Equation (6), we use the Hildebrand relation:

integration, thus there is virtually no increase in computational cost.

**3.2 Coarse-grained models and parameterization** 

repulsive force between particle *i* and *j.* For the DPD system to have a well-defined equilibrium state that obeys Boltzmann statistics, the equilibrium temperature is defined as <sup>2</sup> kT 2 <sup>B</sup> . This condition fixes the temperature of the system and relates with the two DPD parameters and (kBT is usually chosen as the reduced unity of energy). The parameter aij (henceforth referred to as the bead–bead repulsion parameter or simply as the DPD interaction parameter) depends on the underlying atomistic interactions and is related

R

ij (r ) are weight functions of <sup>D</sup> F and ij

to the parameter χ through:

a (1 r )r (r 1) <sup>ˆ</sup> <sup>F</sup> 0 (r 1) 

properties of polymeric materials. The order-order phase transitions can be investigated in more detail through mesoscopic simulations than through experimentation. Mesoscopic simulations are efficient methods to investigate the physical processes of soft matter and their interactions with chemical environments (Rodríguez-Hidalgo et al., 2011; Ramos-Rodriguez et al., 2010). Offer a particularly useful way of exploring the matter transfer process and to make predictions that may be of interest for understanding and elucidating complex process such as the double directionality of order-order phase transitions. In the mesoscopic simulations the atoms of each molecule are not directly represented, but they are grouped together into beads (coarse-grained models), where a springs force reproduce the typical nature of them, and therefore can exhibit a real physical behaviour in multicomponent systems. In this section, we explored the double directionality of phase transition of HPL structure of the PS-PI copolymer from a mesoscopic point of view by mesoscopic simulations, where the phase evolution stages and transient ordered states are analysed.

#### **3.1 Model and simulation method**

To explore the order-order phase transition of the HPL phase, we employed Dissipative Particle Dynamics (DPD) simulations. The original DPD method was introduced by Hoogerbrugge and Koelman and was later modified by Groot, R.D. (Hoogerbrugge & Koelman, 1992; Koelman & Hoogerbrugge, 1993; Groot & Warren, 1997; Groot & Madden, 1998, 1999). The DPD method allows the study of high-molecular-weight systems as the polymeric materials. The coarse-graine

$$\mathbf{f}\_{i} = \mathbf{m}\_{i} \frac{d\mathbf{v}\_{i}}{dt} \tag{8}$$

where **r***i*, **v***i*, mi and **f***i* are the position, velocity, mass, and force, respectively, of bead *i*. Dimensionless units are used in DPD simulations; usually, the mass of each bead is set to 1 DPD mass units, which results in an equation between the force acting on a bead and its acceleration. Each particle is subject to soft interactions with its neighbours via three forces: conservative ( <sup>C</sup> Fij ), dissipative ( DFij ) and random ( RFij ). The total force acting on particle *i* is:

$$\begin{aligned} \mathbf{f}\_i &= \sum\_{\mathbf{j}\neq \mathbf{i}} \left( \mathbf{F}\_{\mathbf{ij}}^{\mathbf{C}} + \mathbf{F}\_{\mathbf{ij}}^{\mathbf{D}} + \mathbf{F}\_{\mathbf{ij}}^{\mathbf{R}} \right) \\ &\quad \mathbf{j} \neq \mathbf{i} \end{aligned} \tag{9}$$

The conservative force CFij is a soft repulsive force that acts between particles *<sup>i</sup>* and *j*. The dissipative force <sup>D</sup> Fij corresponds to a frictional force that depends on both the positions and relative velocities of the particles. The random f <sup>i</sup> i dt **r v** orce <sup>R</sup> F is a random interaction ij between a bead *i* and its neighbour bead *j*. All forces vanish beyond a cut-off radius rc, which is usually chosen as the reduced unit of length, rc 1. The <sup>D</sup> F and ij <sup>R</sup> F forces act as a ij thermostat that conserves momentum and gives the correct hydrodynamics at sufficiently large time and length scales. These forces are given by:

properties of polymeric materials. The order-order phase transitions can be investigated in more detail through mesoscopic simulations than through experimentation. Mesoscopic simulations are efficient methods to investigate the physical processes of soft matter and their interactions with chemical environments (Rodríguez-Hidalgo et al., 2011; Ramos-Rodriguez et al., 2010). Offer a particularly useful way of exploring the matter transfer process and to make predictions that may be of interest for understanding and elucidating complex process such as the double directionality of order-order phase transitions. In the mesoscopic simulations the atoms of each molecule are not directly represented, but they are grouped together into beads (coarse-grained models), where a springs force reproduce the typical nature of them, and therefore can exhibit a real physical behaviour in multicomponent systems. In this section, we explored the double directionality of phase transition of HPL structure of the PS-PI copolymer from a mesoscopic point of view by mesoscopic simulations, where the phase evolution stages

To explore the order-order phase transition of the HPL phase, we employed Dissipative Particle Dynamics (DPD) simulations. The original DPD method was introduced by Hoogerbrugge and Koelman and was later modified by Groot, R.D. (Hoogerbrugge & Koelman, 1992; Koelman & Hoogerbrugge, 1993; Groot & Warren, 1997; Groot & Madden, 1998, 1999). The DPD method allows the study of high-molecular-weight systems as the

> i i d m

j i *<sup>i</sup>* 

which is usually chosen as the reduced unit of length, rc 1. The <sup>D</sup> F and ij

and relative velocities of the particles. The random f <sup>i</sup>

large time and length scales. These forces are given by:

where **r***i*, **v***i*, mi and **f***i* are the position, velocity, mass, and force, respectively, of bead *i*. Dimensionless units are used in DPD simulations; usually, the mass of each bead is set to 1 DPD mass units, which results in an equation between the force acting on a bead and its acceleration. Each particle is subject to soft interactions with its neighbours via three forces: conservative ( <sup>C</sup> Fij ), dissipative ( DFij ) and random ( RFij ). The total force acting on particle *i* is:

i

 CDR ij ij ij f FFF

i dt **r**

The conservative force CFij is a soft repulsive force that acts between particles *<sup>i</sup>* and *j*. The

dissipative force <sup>D</sup> Fij corresponds to a frictional force that depends on both the positions

between a bead *i* and its neighbour bead *j*. All forces vanish beyond a cut-off radius rc,

thermostat that conserves momentum and gives the correct hydrodynamics at sufficiently

dt **<sup>v</sup> <sup>f</sup>** (8)

(9)

**v** orce <sup>R</sup> F is a random interaction ij

<sup>R</sup> F forces act as a ij

and transient ordered states are analysed.

**3.1 Model and simulation method** 

polymeric materials. The coarse-graine

$$\mathbf{F}\_{\hat{\mathbf{i}}\hat{\mathbf{j}}}^{\mathbb{C}} = \begin{vmatrix} \mathbf{a}\_{\hat{\mathbf{i}}\dagger}(\mathbf{1} - \mathbf{r}\_{\hat{\mathbf{i}}\dagger})\hat{\mathbf{r}}\_{\hat{\mathbf{i}}\dagger} & \quad \left(\mathbf{r}\_{\hat{\mathbf{i}}} \le \mathbf{1}\right) \\ \mathbf{0} & \left(\mathbf{r}\_{\hat{\mathbf{i}}\dagger} > \mathbf{1}\right) \end{vmatrix} \tag{10}$$

$$\mathbf{F}\_{\mathrm{ij}}^{\mathrm{D}} = \left[ -\chi \mathrm{co}^{\mathrm{D}}(\mathbf{r}\_{\mathrm{ij}}) (\mathbf{v}\_{\mathrm{ij}}, \hat{\mathbf{r}}\_{\mathrm{ij}}) \hat{\mathbf{r}}\_{\mathrm{ij}} \right] \tag{11}$$

$$\mathbf{F}\_{\rm ij}^{\rm R} = \left[ \sigma \boldsymbol{\sigma} \boldsymbol{\alpha}^{\rm R} \left( \mathbf{r}\_{\rm ij} \right) \boldsymbol{\xi}\_{\rm ij}^{\rm r} \hat{\mathbf{r}}\_{\rm ij} \right] \tag{12}$$

where ij <sup>i</sup> <sup>j</sup> r rr , ij ij r r /r <sup>ˆ</sup> , is the dissipation strength, is the noise strength, <sup>D</sup> ij (r ) and R ij (r ) are weight functions of <sup>D</sup> F and ij <sup>R</sup> F forces, respectively, and a ij ij is the maximum repulsive force between particle *i* and *j.* For the DPD system to have a well-defined equilibrium state that obeys Boltzmann statistics, the equilibrium temperature is defined as <sup>2</sup> kT 2 <sup>B</sup> . This condition fixes the temperature of the system and relates with the two DPD parameters and (kBT is usually chosen as the reduced unity of energy). The parameter aij (henceforth referred to as the bead–bead repulsion parameter or simply as the DPD interaction parameter) depends on the underlying atomistic interactions and is related to the parameter χ through:

$$\mathbf{a}\_{\text{ij}} = \mathbf{a}\_{\text{ii}} + \frac{\mathbf{k}\_{\text{B}} \mathbf{T} \chi\_{\text{ij}}(\mathbf{T})}{0.306} \tag{13}$$

In this way, a connection exists between the molecular character of the coarse-grained model and the DPD parameter. The parameter aij is given in terms of kBT (DPD reduced units). Equation (6) implies that, if the species are compatible, ij 0 and therefore, aij = 25. Established procedures for mapping between the DPD and physical scales and for choosing the system temperature are not yet available. We therefore use Flory–Huggins theory (through the χ(T) dependence) to introduce the temperature into the DPD simulations. To calculate the aij values of Equation (6), we use the Hildebrand relation:

$$\chi = \frac{V\_{\rm m}}{RT} (\delta\_1 - \delta\_2)^2 \tag{14}$$

where Vm and δ are the mean molar volume and solubility parameter, respectively. In the DPD simulation, the dynamic behaviour of order-order phase transition of the HPL structure is followed by integration of the equations of motion of each species using a modified version of the Verlet algorithm (Verlet, 1967). The integration of the equations of motion for each particle generates a trajectory through the system's phase, from which thermodynamic observables may be constructed by suitable averaging. Based on this information, the orderorder phase transition can be observed. In this algorithm, the forces are still updated once per integration, thus there is virtually no increase in computational cost.

#### **3.2 Coarse-grained models and parameterization**

The molecular structure of PS-PI copolymer was built by means of the polymer builder module of Accelrys (Accelrys, 2006). The polymeric molecule with linear architecture

Thermal Study on Phase Transitions of Block Copolymers by Mesoscopic Simulation 577

Styrene-isoprene interaction

300 350 400 450 500

Temperature (K)

Fig. 13. Interaction parameters between styrene and isoprene molecules at temperatures

order-disorder transition) to room temperature and the double directionality of the HPL phase (via order-order transition) employing the coarse-grained model with predefined

All DPD simulations were performed in a cubic box that measured 202020 in size, which contained a total of 2.4x104 representatives particles, a spring constant of C=4, and a density of =3. The interaction parameters between identical species were then chosen as **a**ST-ST = **a** PI-PI= 25. Each bead was assigned a radius of 1. The coarse-grained number for each chemical

The equilibrium phases formation of PS-PI copolymer is governed by the composition, temperature and immiscibility between their components blocks, these factors were considered in the coarse-grained model via the architecture and interaction parameters.

All simulations start from a disordered state, where the PS-PI chains are in a homogeneous melted phase. First, we set the interactions parameters at T = 298 K. We then let the simulation proceed for 5x105 steps and during the temperature relaxation; we observed the microphase segregation process and the generation of equilibrium phases via ODT. Several transient stages were detected in the ODT process; a) melt phase, where the copolymer chains move freely; b) microphase segregation by temperature effect (temperature decrease), c) generation of pure microdomains of poly(styrene) and poly(isoprene), in this stage the

To identify the composition region where the HPL structures are formed, we scanned the composition interval from 0.1–0.5 (volume fraction of poly(styrene) with increments of 0.03) at constant temperature of 298 K (i.e. constant N). All DPD simulations generate a coarsegrained system sufficiently large to observe the classical phases formation in the analysed composition interval. The mesoscopic simulation of PS-PI copolymer exhibits a wide variety of structures or equilibrium phases such as LAM, HPC, BCC, Gyroid and HPL. The

ordered phases system attains an equilibrium temperature to room temperature.

different obtained by Monte Carlo molecular simulation (Vicente et al. 2006).

0.10

architecture proposed in this mesoscopic study.

**3.3 Simulation results and discussion** 

species was held constant during the DPD simulations.

**3.3.1 Hexagonally perforated layers structures** 

0.15

0.20

0.25

Interaction parameter (

)

0.30

0.35

0.40

contains a total of 300 repetitive units in the main chain. The molecular weight of the PS-PI copolymer presented an interval of 20798–25842 g/mol. The diblock copolymer molecule was replaced by a coarse-grained model constituted by 30 beads, where each bead represents a statistical segment (characteristic ratio, (Cn≈ 10) (Soto-Figueroa, et al 2005)), see Figure 12. The Equation (15) was used to map the real structure of diblock copolymer to statistical model (mesoscopic model).

$$\mathbf{C\_{SGD}} = \frac{\mathbf{M\_P}}{\mathbf{M\_m}(\text{SSL})} \tag{15}$$

where CSGD, MP, Mm and SSL are bead number with Gaussian distribution, the molar mass of the block copolymer, molar mass of a repeat unit and means statistical segment level (characteristic ratio (Cn), or persistence length (Lp) or statistical Kuhn segment length (ak)) respectively (Soto-Figueroa et al., 2007).

Fig. 12. Schematic representation of PS-PI copolymer: a) chemical structure of a PS-PI chain, n and m represent the polymeric unit number of each block, b) coarse-grained model of PS-PI system with linear architecture.

The chemical and physical nature of the coarse-grained model in mesoscopic simulation is described by interaction parameters (χij). In order to study the order-order transition at temperatures different with DPD simulations we take the temperature dependence χij(T) = χ(T), in this way the real temperature is introduced into the DPD simulation (Rodríguez-Hidalgo et al., 2009). The interaction parameters for Equation (13) were evaluated from bulk atomistic simulations using the Fan, F.C. model (Fan et al., 1992). In this way the parameter interaction was expressed as:

$$\chi(\mathbf{T}) = \frac{\Delta \mathbf{G}(\mathbf{T})}{\mathbf{RT}} = \frac{\mathbf{Z}\_{12}\mathbf{E}\_{12}(\mathbf{T}) + \mathbf{Z}\_{21}\mathbf{E}\_{21}(\mathbf{T}) - \mathbf{Z}\_{11}\mathbf{E}\_{11}(\mathbf{T}) - \mathbf{Z}\_{22}\mathbf{E}\_{22}(\mathbf{T})}{2\mathbf{RT}} \tag{16}$$

where ΔG denotes the Gibbs free energy, χ is interaction parameter, Z and ΔE12 are coordination number and differential energy of interaction of an unlike pair respectively. The Figure 13 show the interaction parameters χ(T), the temperature interval analysed is from 298 K to 500 K.

The tendencies that exhibit interaction parameters in temperature function are in agreement with the Hildebrand relation and are adequate to explore the formation of HPL phases (via

Fig. 13. Interaction parameters between styrene and isoprene molecules at temperatures different obtained by Monte Carlo molecular simulation (Vicente et al. 2006).

order-disorder transition) to room temperature and the double directionality of the HPL phase (via order-order transition) employing the coarse-grained model with predefined architecture proposed in this mesoscopic study.

All DPD simulations were performed in a cubic box that measured 202020 in size, which contained a total of 2.4x104 representatives particles, a spring constant of C=4, and a density of =3. The interaction parameters between identical species were then chosen as **a**ST-ST = **a** PI-PI= 25. Each bead was assigned a radius of 1. The coarse-grained number for each chemical species was held constant during the DPD simulations.

#### **3.3 Simulation results and discussion**

576 Advances in Chemical Engineering

contains a total of 300 repetitive units in the main chain. The molecular weight of the PS-PI copolymer presented an interval of 20798–25842 g/mol. The diblock copolymer molecule was replaced by a coarse-grained model constituted by 30 beads, where each bead represents a statistical segment (characteristic ratio, (Cn≈ 10) (Soto-Figueroa, et al 2005)), see Figure 12. The Equation (15) was used to map the real structure of diblock copolymer to

SGD

<sup>M</sup> <sup>C</sup>

 P

M SSL (15)

m

where CSGD, MP, Mm and SSL are bead number with Gaussian distribution, the molar mass of the block copolymer, molar mass of a repeat unit and means statistical segment level (characteristic ratio (Cn), or persistence length (Lp) or statistical Kuhn segment length (ak))

Fig. 12. Schematic representation of PS-PI copolymer: a) chemical structure of a PS-PI chain, n and m represent the polymeric unit number of each block, b) coarse-grained model of PS-

The chemical and physical nature of the coarse-grained model in mesoscopic simulation is described by interaction parameters (χij). In order to study the order-order transition at temperatures different with DPD simulations we take the temperature dependence χij(T) = χ(T), in this way the real temperature is introduced into the DPD simulation (Rodríguez-Hidalgo et al., 2009). The interaction parameters for Equation (13) were evaluated from bulk atomistic simulations using the Fan, F.C. model (Fan et al., 1992). In this way the parameter

where ΔG denotes the Gibbs free energy, χ is interaction parameter, Z and ΔE12 are coordination number and differential energy of interaction of an unlike pair respectively. The Figure 13 show the interaction parameters χ(T), the temperature interval analysed is

The tendencies that exhibit interaction parameters in temperature function are in agreement with the Hildebrand relation and are adequate to explore the formation of HPL phases (via

Z E Z E Z E Z E

<sup>Δ</sup>G(T) <sup>χ</sup> <sup>T</sup> <sup>12</sup> <sup>12</sup> <sup>21</sup> <sup>21</sup> <sup>11</sup> <sup>11</sup> <sup>22</sup> <sup>22</sup> (T) (T) (T) (T) (16)

2RT

statistical model (mesoscopic model).

respectively (Soto-Figueroa et al., 2007).

PI system with linear architecture.

interaction was expressed as:

from 298 K to 500 K.

RT

#### **3.3.1 Hexagonally perforated layers structures**

The equilibrium phases formation of PS-PI copolymer is governed by the composition, temperature and immiscibility between their components blocks, these factors were considered in the coarse-grained model via the architecture and interaction parameters.

All simulations start from a disordered state, where the PS-PI chains are in a homogeneous melted phase. First, we set the interactions parameters at T = 298 K. We then let the simulation proceed for 5x105 steps and during the temperature relaxation; we observed the microphase segregation process and the generation of equilibrium phases via ODT. Several transient stages were detected in the ODT process; a) melt phase, where the copolymer chains move freely; b) microphase segregation by temperature effect (temperature decrease), c) generation of pure microdomains of poly(styrene) and poly(isoprene), in this stage the ordered phases system attains an equilibrium temperature to room temperature.

To identify the composition region where the HPL structures are formed, we scanned the composition interval from 0.1–0.5 (volume fraction of poly(styrene) with increments of 0.03) at constant temperature of 298 K (i.e. constant N). All DPD simulations generate a coarsegrained system sufficiently large to observe the classical phases formation in the analysed composition interval. The mesoscopic simulation of PS-PI copolymer exhibits a wide variety of structures or equilibrium phases such as LAM, HPC, BCC, Gyroid and HPL. The

Thermal Study on Phase Transitions of Block Copolymers by Mesoscopic Simulation 579

The OOT process was visualized during mesoscopic simulation; the results show that the transformation from HPL to cylindrical phase by temperature effect is generated in several stages. When the HPL phase is annealed below the OOT temperature, we observed a slow dynamic motion of the PS and PI microdomains, the thermodynamic stability between microdomains different they are in an energetic barrier that maintain the HPL phase stable. When the HPL phase is annealed to a higher temperature, T≥ 432 K, the energetic barrier that maintain the HPL phase stable is exceeded, the thermodynamic interactions (enthalpic and entropic) play an important role in the OOT process. The enthalpic interaction is proportional to the Flory-Huggins interaction parameter (Fig. 13), which is found to be

During the temperature increase, the interaction parameter between the PS and PI microdomains diminishes, generating the OOT from HPL arrangement to cylindrical phase, the enthalpic interaction in the OOT is accompanied by an increase in entropy. At higher temperatures the entropic interactions dominates and is cause of anisotropic composition

The HPL phase develops short-lived transient states during the OOT process, because of interface instability (PS/PI), combined with fast molecular motion of PS and PI microdomains. In Figure 15(a-d) are shown snapshots of the thermally induced phase transition from HPL to cylindrical structure during the annealed process at temperature,

Fig. 15. Snapshots of the OOT pathway from HPL to cylindrical phase: a) undulation process of HPL phase, b-c) instability of HPL phase and d) formation of cylindrical phase.

this thermally induced stage the uniform cylindrical microdomains are stabilized.

**3.3.3 Order-order phase transition of HPL to Gyroid phase** 

The snapshot of Figure 15(a) corresponds to the initial equilibrium state of HPL phase, initially, the PS perforated microdomains display an undulation process induced by thermodynamic instability, in this point the PS perforated microdomains and the PS/PI interface become less rigid by temperature effect. The PS microdomains maintain their shape an short-period of time, as time goes, the anisotropic composition fluctuations (thermally-induced) increase quickly, the PI interconnections into PS perforated microdomains enlarge their area, in this stage the PS microdomains are unstable see Figure 15(b-c). When the anisotropic composition fluctuations reach a critical point, the PS unstable microdomains change their structure to a new cylindrical arrangement, see Figure 15(d), in

The HPL phase with a volume fraction of PS, fPS=0.36, (obtained within the first 5x105 time steps) was now subject to a thermal heating process for another 5x105 time steps. The

inversely proportional to temperature.

T= 432 K.

fluctuations into polymeric microdomains (Ryu, et al., 1999).

equilibrium phases of PS-PI system depend primarily on three factors: (i) the volume fraction of PS and PI blocks (*f)*, (ii) the degree of polymerization (*N*), and (iii) the interaction parameter (*χ*). The results obtained by DPD simulations are in accordance with the meanfield theory (Leibler, 1980).

The HPL phase was obtained into the composition intervals of 0.3 to 0.36 (volume fraction of poly(styrene)), see Figure 14. The tendency of PS-PI chains to self-assemble into HPL structures depends of the previous factors and is governed by thermodynamics interactions (enthalpic and entropic) during the microphase segregation process.

Fig. 14. Specific phases of PS-PI copolymer obtained as the blocks composition (PS/PI) is modified: a) cylindrical, b) HPL and c) Gyroid.

The thermodynamic stability of HPL phase in the predominance region vary with the relative chain length (poly(styrene) composition) of the component blocks. The predominance region where the HPL phase is formed has two composition limits that are contiguous with other equilibrium phases of different morphology. In the low composition limit where the volume fraction of poly(styrene) is close to 0.3, the HPL structure has as neighbour the cylindrical phase, whereas in the height composition limit (close to 0.36), the HPL structure has as neighbour the Gyroid structure. These behaviours between different phases are in accordance with the phase diagram reported by Khandpur (Khandpur et al., 1995).

The order-order phase transitions that exhibits the HPL structures with specific compositions of 0.3 and 0.36 (volume fraction of poly(styrene) are investigated. The HPL structures are modified through temperature increase, keeping constant the PS-PI block composition. The OOT of HPL structure for each specific composition shows a selective directionality (from HPL to cylinders and from HPL to Gyroid) during thermal study.

#### **3.3.2 Order-order phase transition of HPL to cylindrical phase**

The HPL phase with a specific composition of 0.3 (volume fraction of poly(styrene) was put at continuous cycles of thermal heating in the temperature interval from 298 to 500 K. A total of 2.0 x 105 time steps with step size of t = 0.03 were allowed in the mesoscopic simulation to reach the thermodynamic balance of PS-PI system in each temperature increment. The thermally induced phase transition from HPL to cylindrical phase was observed at the temperature of 432 K, this corresponds to the OOT, and this fact is in accordance with theoretical and experimental evidences (You et al., 2007).

equilibrium phases of PS-PI system depend primarily on three factors: (i) the volume fraction of PS and PI blocks (*f)*, (ii) the degree of polymerization (*N*), and (iii) the interaction parameter (*χ*). The results obtained by DPD simulations are in accordance with the mean-

The HPL phase was obtained into the composition intervals of 0.3 to 0.36 (volume fraction of poly(styrene)), see Figure 14. The tendency of PS-PI chains to self-assemble into HPL structures depends of the previous factors and is governed by thermodynamics interactions

Fig. 14. Specific phases of PS-PI copolymer obtained as the blocks composition (PS/PI) is

The thermodynamic stability of HPL phase in the predominance region vary with the relative chain length (poly(styrene) composition) of the component blocks. The predominance region where the HPL phase is formed has two composition limits that are contiguous with other equilibrium phases of different morphology. In the low composition limit where the volume fraction of poly(styrene) is close to 0.3, the HPL structure has as neighbour the cylindrical phase, whereas in the height composition limit (close to 0.36), the HPL structure has as neighbour the Gyroid structure. These behaviours between different phases are in accordance with the phase diagram reported by Khandpur (Khandpur et al.,

The order-order phase transitions that exhibits the HPL structures with specific compositions of 0.3 and 0.36 (volume fraction of poly(styrene) are investigated. The HPL structures are modified through temperature increase, keeping constant the PS-PI block composition. The OOT of HPL structure for each specific composition shows a selective directionality (from HPL to cylinders and from HPL to Gyroid) during thermal study.

The HPL phase with a specific composition of 0.3 (volume fraction of poly(styrene) was put at continuous cycles of thermal heating in the temperature interval from 298 to 500 K. A total of 2.0 x 105 time steps with step size of t = 0.03 were allowed in the mesoscopic simulation to reach the thermodynamic balance of PS-PI system in each temperature increment. The thermally induced phase transition from HPL to cylindrical phase was observed at the temperature of 432 K, this corresponds to the OOT, and this fact is in

**3.3.2 Order-order phase transition of HPL to cylindrical phase** 

accordance with theoretical and experimental evidences (You et al., 2007).

(enthalpic and entropic) during the microphase segregation process.

modified: a) cylindrical, b) HPL and c) Gyroid.

1995).

field theory (Leibler, 1980).

The OOT process was visualized during mesoscopic simulation; the results show that the transformation from HPL to cylindrical phase by temperature effect is generated in several stages. When the HPL phase is annealed below the OOT temperature, we observed a slow dynamic motion of the PS and PI microdomains, the thermodynamic stability between microdomains different they are in an energetic barrier that maintain the HPL phase stable.

When the HPL phase is annealed to a higher temperature, T≥ 432 K, the energetic barrier that maintain the HPL phase stable is exceeded, the thermodynamic interactions (enthalpic and entropic) play an important role in the OOT process. The enthalpic interaction is proportional to the Flory-Huggins interaction parameter (Fig. 13), which is found to be inversely proportional to temperature.

During the temperature increase, the interaction parameter between the PS and PI microdomains diminishes, generating the OOT from HPL arrangement to cylindrical phase, the enthalpic interaction in the OOT is accompanied by an increase in entropy. At higher temperatures the entropic interactions dominates and is cause of anisotropic composition fluctuations into polymeric microdomains (Ryu, et al., 1999).

The HPL phase develops short-lived transient states during the OOT process, because of interface instability (PS/PI), combined with fast molecular motion of PS and PI microdomains. In Figure 15(a-d) are shown snapshots of the thermally induced phase transition from HPL to cylindrical structure during the annealed process at temperature, T= 432 K.

Fig. 15. Snapshots of the OOT pathway from HPL to cylindrical phase: a) undulation process of HPL phase, b-c) instability of HPL phase and d) formation of cylindrical phase.

The snapshot of Figure 15(a) corresponds to the initial equilibrium state of HPL phase, initially, the PS perforated microdomains display an undulation process induced by thermodynamic instability, in this point the PS perforated microdomains and the PS/PI interface become less rigid by temperature effect. The PS microdomains maintain their shape an short-period of time, as time goes, the anisotropic composition fluctuations (thermally-induced) increase quickly, the PI interconnections into PS perforated microdomains enlarge their area, in this stage the PS microdomains are unstable see Figure 15(b-c). When the anisotropic composition fluctuations reach a critical point, the PS unstable microdomains change their structure to a new cylindrical arrangement, see Figure 15(d), in this thermally induced stage the uniform cylindrical microdomains are stabilized.

#### **3.3.3 Order-order phase transition of HPL to Gyroid phase**

The HPL phase with a volume fraction of PS, fPS=0.36, (obtained within the first 5x105 time steps) was now subject to a thermal heating process for another 5x105 time steps. The

Thermal Study on Phase Transitions of Block Copolymers by Mesoscopic Simulation 581

different processes: order-disorder phase transition and thermotropic order-order phase transition. In the order-disorder phase transition, the phase behaviour is governed by entropic and enthalpic interactions, whereas the order-order phase transition are controlled by anisotropic composition fluctuations of theses thermodynamic processes can emerge

The ODT and OOT that exhibits of block copolymer can be explored in more detail through mesoscopic simulations than through experimentation. The mesoscopic simulation methods, offer a particularly useful way of exploring the phase behaviour pathway and to make predictions that may be of interest for understanding and elucidating complex process such

The DPD approach has been successfully applied to the investigation of phase transition processes (ODT and OOT) of HPL structure. The mesoscopic simulation outcomes show that the HPL phase of PS-PI diblock copolymer exhibits a double directionality of orderorder phase transition from HPL to cylindrical phase and from HPL to Gyroid phase. This double directionality of OOT is controlled by small variations of poly(styrene) concentration in the poly(styrene) microdomains of HPL phase. Finally, all the simulation outcomes are qualitatively consistent with the experimental results, demonstrating that the DPD method may provide a powerful tool for the investigation and analysis of soft matter transformation

This work was supported by the Universidad Nacional Autónoma de México (UNAM), PAPIIT project No. IN104410-2 and IN109712. We also acknowledges the financial support

Almdal, K.; Koppi, K.A.; Bates, F.S.; Mortensen, K. (1992). Multiple ordered phases in a

Accelrys. (2006), Material studio release, notes, release 5.0; Accelrys Software, Inc.: San

Bates, F.S. & Fredrickson, G.H. (1990). Block copolymer thermodynamics: Theory and

Bates, F.S. & Fredrickson, G.H. (1999), Block copolymers-Designer soft materials. *AIP Phys.* 

Bodycomb, J.; Yamaguchi, D.; Hashimoto, T. (2000). A Small-Angle X-ray Scattering Study

Balta-Calleja, F.J. & Roslaniec, Z. (2000), Block Copolymers; Marcel-Dekker Publishers,

Court, F.; Yamaguchi, D.; Hashimoto, T. (2006). Morphological Studies of Binary Mixtures of

of the Phase Behavior of Diblock Copolymer/Homopolymer Blends.

Block Copolymers: Temperature Dependence of Cosurfactant Effects.

of the Consejo Nacional de Ciencia y Tecnología (CONACYT) Project: 2012.

experiment. *Annu. Rev. Phys. Chem*., Vol. 41, pp. 525-557.

*Macromolecules*, Vol. 33(14), pp. 5187-5197.

*Macromolecules*, Vol. 39(7), pp. 2596-2605.

Bates, F. S. (1991). Polymer – Polymer phase behavior. *Science*, Vol. 251, pp. 898.

block copolymer melt. *Macromolecules*, Vol. 25(6), pp. 1743-1751.

supramolecular structures with well-defined morphologies and specific properties.

as order-disorder of order-order phase transitions.

process by thermal heating effects.

**5. Acknowledgment** 

Diego, CA.

*Today*, 2, 32-38.

(Eds.): New York.

**6. References** 

simulation outcome shows an OOT from HPL to Gyroid phase at the temperature of 438 K. The composition increase of poly(styrene) in HPL microdomains modify the phase transition directionality. The OOT from HPL to Gyroid phase obtained by DPD simulation are consistent with experimental results reported by Insun Park. They have investigated the phase transition behaviour from the hexagonally perforated layer (HPL) to the Gyroid phase in supported thin film of a poly(styrene)-*b*-poly(isoprene) (PS-*b*-PI) diblock copolymer (OOT occur at temperature of 443 K) (Park et al., 2005).

During the thermal heating process, the HPL phase undergoes the microdomains modification of PS and PI, generating transient intermediate stages as are sketched schematically in Figure 16(a-c). Three transient intermediate stages were detected during the OOT: (i) undulation process of pure microdomains and PS/PI interface due to the thermodynamic instability by temperature effect, see Figure 16(a), in this stage, the entropic and enthalpic interactions govern the microphase stability and induce anisotropic composition fluctuations into pure microdomains, (ii) increase of volume of poly(isoprene) interconnections into PS perforates microdomain, Figure 16(b), in this stage the PS perforated microdomains are instable, (iii) formation of parallel interconnections between PS perforated layers, see Figure 16(c), in this stage, the Gyroid phase is formed.

Fig. 16. Snapshots of evolution process of the OOT from HPL to Gyroid phase: a) undulation process of PS microdomains in the PI matrix by anisotropic composition fluctuations effect, b) volume increase of PI interconnections into PS perforates microdomains, c) formation of Gyroid phase by generation of parallel interconnections between PS perforated microdomains.

The HPL phase of PS-PI diblock copolymer exhibits a double directionality of order-order transition from HPL to cylindrical phase and from HPL to Gyroid phase. The anisotropic composition fluctuations of PS and PI microdomains of HPL phase by thermal heating process induce the order-order phase transitions, however the phase transition directionality towards a specific phase (cylindrical or Gyroid) in the thermal process is governed by small variations of poly(styrene) concentration in the poly(styrene) microdomains of HPL phase. Variations of small composition into structured microdomains according to Leibler´s mean-field theory greatly modify the phase thermodynamic behaviour (separation, segregation and phase transformation).

#### **4. Conclusion**

The block copolymers are "smart" materials that have the ability to self-assemble inside a variety of periodic phases of high regularity in size and shape. The phase behaviour (type of structure and morphology) that exhibits these polymeric materials can be controlled by two different processes: order-disorder phase transition and thermotropic order-order phase transition. In the order-disorder phase transition, the phase behaviour is governed by entropic and enthalpic interactions, whereas the order-order phase transition are controlled by anisotropic composition fluctuations of theses thermodynamic processes can emerge supramolecular structures with well-defined morphologies and specific properties.

The ODT and OOT that exhibits of block copolymer can be explored in more detail through mesoscopic simulations than through experimentation. The mesoscopic simulation methods, offer a particularly useful way of exploring the phase behaviour pathway and to make predictions that may be of interest for understanding and elucidating complex process such as order-disorder of order-order phase transitions.

The DPD approach has been successfully applied to the investigation of phase transition processes (ODT and OOT) of HPL structure. The mesoscopic simulation outcomes show that the HPL phase of PS-PI diblock copolymer exhibits a double directionality of orderorder phase transition from HPL to cylindrical phase and from HPL to Gyroid phase. This double directionality of OOT is controlled by small variations of poly(styrene) concentration in the poly(styrene) microdomains of HPL phase. Finally, all the simulation outcomes are qualitatively consistent with the experimental results, demonstrating that the DPD method may provide a powerful tool for the investigation and analysis of soft matter transformation process by thermal heating effects.

#### **5. Acknowledgment**

This work was supported by the Universidad Nacional Autónoma de México (UNAM), PAPIIT project No. IN104410-2 and IN109712. We also acknowledges the financial support of the Consejo Nacional de Ciencia y Tecnología (CONACYT) Project: 2012.

#### **6. References**

580 Advances in Chemical Engineering

simulation outcome shows an OOT from HPL to Gyroid phase at the temperature of 438 K. The composition increase of poly(styrene) in HPL microdomains modify the phase transition directionality. The OOT from HPL to Gyroid phase obtained by DPD simulation are consistent with experimental results reported by Insun Park. They have investigated the phase transition behaviour from the hexagonally perforated layer (HPL) to the Gyroid phase in supported thin film of a poly(styrene)-*b*-poly(isoprene) (PS-*b*-PI) diblock copolymer (OOT

During the thermal heating process, the HPL phase undergoes the microdomains modification of PS and PI, generating transient intermediate stages as are sketched schematically in Figure 16(a-c). Three transient intermediate stages were detected during the OOT: (i) undulation process of pure microdomains and PS/PI interface due to the thermodynamic instability by temperature effect, see Figure 16(a), in this stage, the entropic and enthalpic interactions govern the microphase stability and induce anisotropic composition fluctuations into pure microdomains, (ii) increase of volume of poly(isoprene) interconnections into PS perforates microdomain, Figure 16(b), in this stage the PS perforated microdomains are instable, (iii) formation of parallel interconnections between

Fig. 16. Snapshots of evolution process of the OOT from HPL to Gyroid phase: a) undulation process of PS microdomains in the PI matrix by anisotropic composition fluctuations effect, b) volume increase of PI interconnections into PS perforates microdomains, c) formation of

The HPL phase of PS-PI diblock copolymer exhibits a double directionality of order-order transition from HPL to cylindrical phase and from HPL to Gyroid phase. The anisotropic composition fluctuations of PS and PI microdomains of HPL phase by thermal heating process induce the order-order phase transitions, however the phase transition directionality towards a specific phase (cylindrical or Gyroid) in the thermal process is governed by small variations of poly(styrene) concentration in the poly(styrene) microdomains of HPL phase. Variations of small composition into structured microdomains according to Leibler´s mean-field theory greatly modify the phase thermodynamic

The block copolymers are "smart" materials that have the ability to self-assemble inside a variety of periodic phases of high regularity in size and shape. The phase behaviour (type of structure and morphology) that exhibits these polymeric materials can be controlled by two

PS perforated layers, see Figure 16(c), in this stage, the Gyroid phase is formed.

Gyroid phase by generation of parallel interconnections between PS perforated

behaviour (separation, segregation and phase transformation).

occur at temperature of 443 K) (Park et al., 2005).

microdomains.

**4. Conclusion** 


Thermal Study on Phase Transitions of Block Copolymers by Mesoscopic Simulation 583

Mani, S.; Weiss, R. A.; Cantino, M. E.; Khairallah, L. H.; Hans, S. F.; Williams, C. E. (2000).

Park, I.; Lee, B.; Ryu, J.; Im, K.; Yoon, J.; Ree, M.; Chang, T. (2005). Epitaxial Phase Transition

Ramos-Rodriguez, D. A.; Rodriguez-Hidalgo, M. R.; Soto-Figueroa, C.; Vicente, L. (2010).

Rodríguez-Hidalgo, M.R.; Soto-Figueroa, C.; Vicente, L. (2009). Mesoscopic study of

Ryu, C.Y. & Lodge, T.P. (1999). Thermodynamic Stability and Anisotropic Fluctuations in

Sakurai, S.; Kawada, H.; Hashimoto, T.; Fetters, L.J. (1993).Thermoreversible Morphology

Sakurai, S.; Hashimoto, T.; Fetters, L.J. (1996). Thermoreversible Cylinder−Sphere Transition

Sakamato, N.; Hashimoto, T.; Han, C.D.; Kim, D.; Vaidya, N.Y. (1997). Order−Order and

Soto-Figueroa, C.; Rodríguez-Hidalgo, M.R.; Martínez-Magadán, J.M. (2005). Molecular

Soto-Figueroa, C.; Luis-Vicente.; Martínez-Magadán, J.M.; Rodríguez-Hidalgo, M.R. (2007).

Soto-Figueroa, C.; Luis-Vicente.; Martínez-Magadán, J.M.; Rodríguez-Hidalgo, M.R. (2007).

Soto-Figueroa, C.; Rodríguez-Hidalgo, M.R.; Martínez-Magadán, J.M.; Luis-Vicente. (2008).

Soto-Figueroa, C.; Rodríguez-Hidalgo, M.R.; Martínez-Magadán, J.M.; Luis-Vicente. (2008).

Polymeric Systems. *J. Phys. Chem. B*, Vol. 111, pp. 11756-11764.

transitions by temperature control. *Polymer*, Vol. 50, pp 4596-4601.

in Thin Film. *Macromolecules*, Vol. 38, pp.10532-10536.

environment. *Soft Matter*, Vol. 7, pp 8224-8230.

*Macromolecules*, Vol. 26(21), pp. 5796-5802.

Solutions. *Macromolecules*, Vol. 29(2), pp. 740-747.

Copolymer. *Macromolecules*, Vol. 30(6), pp. 1621-1632.

Copolymer. *Macromolecules*, Vol. 41, pp. 3297-3304.

pp. 215–219.

pp. 7190-7201.

Vol. 46, pp. 7485-7493.

*Polymer*, Vol. 48, pp. 3902-3911.

Evidence for a Thermally Reversible Order-Order Transition between Lamellar and Perforated Lamellar Microphases in a Triblock Copolymer. *Eur. Polym. J*., Vol. 36,

of Polystyrene-b-Polyisoprene from Hexagonally Perforated Layer to Gyroid Phase

Molecular and mesoscopic study of ionic liquids and their use as solvents of active agents released by polymeric vehicles. *Molecular Physics*, Vol. 108(5), pp. 657–665. Rodríguez-Hidalgo, M.R.; Soto-Figueroa, C.; Vicente, L. (2011). Mesoscopic simulation of the

drug release mechanism on the polymeric vehicle P(ST-DVB) in an acid

cylindrical phases of poly(styrene)-poly(isoprene) copolymer: Order-order phase

the Cylinder-to-Sphere Transition of a Block Copolymer. *Macromolecules*, Vol. 32,

Transition between Spherical and Cylindrical microdomains of Block Copolymers.

of Polystyrene-block-polyisoprene Diblock Copolymers in Dioctyl Phthalate

Order−Disorder Transitions in a Polystyrene-block-Polyisoprene-block-Polystyrene

simulation of diblock copolymers; morphology and mechanical properties. *Polymer*,

Self-Organization Process of Ordered Structures in Linear and Star Poly(styrene)- Poly(isoprene) Block Copolymers: Gaussian Models and Mesoscopic Parameters of

Mesoscopic simulation of asymmetric-copolymer/homopolymer blends: Microphase morphological modification by homopolymer chains solubilization.

Dissipative Particle Dynamics Study of Order-Order Phase Transition of BCC, HPC, OBDD, and LAM Structures of the Poly(styrene)-Poly(isoprene) Diblock

Mesoscopic simulation of metastable microphases in the order–order transition


Fan, F.C.; Olafson, B.D.; Blanco, M. (1992). Application of molecular simulation to derive phase diagrams of binary mixtures. *Macromolecules*, Vol. 25, pp. 3667-3676. Flory, P. J. (1953). Principles of Polymer Chemistry, Cornell University Press, Ithaca, New

Groot, R. D. & Warren, P. B. (1997). Dissipative particle dynamics: bridging the gap between atomistic and mesoscopic simulation. *J. Chem. Phys*., Vol. 107, pp. 4423-4435. Groot, R. D. & Madden, T. J. (1998). Dynamic simulation of diblock copolymer microphase

Groot, R. D. & Madden, T. J., Tildesley, D. J. (1999). On the role of hydrodynamic

Hoogerbrugge, P. J. & Koelman, J. M. V. A. (1992). Simulating microscopic hydrodynamic phenomena with Dissipative particle dinamics. *Europhys. Lett*.,Vol. 19, pp.155-160. Hajduk, D. A.; Harper, P.E.; Gruner, S. M.; Honeker, C. C.; Thomas, E. L.; Fetters, L. J. (1995).

Hamley, I.W. (1998). The Physics of Block Copolymers. Oxford University Press, ISBN

Koelman, J. M. V. A. & Hoogerbrugge, P. J.(1993). Dynamic simulation of hard sphere

Kimishima, K.; Koga, T.; Hashimoto, T. (2000). Order−Order Phase Transition between

Kim, J.K.; Lee, H.H.; Gu, Q.J.; Chang, T.; Jeong, Y.H. (1998). Determination of Order−Order

Krishanamoorti, R.; Modi M.A.; Tse, M.F.; Wang H.-C. (2000). Pathway and Kinetics of

Krishanamoorti, R.; Silva A. S.; Modi M.A. (2000). Small-Angle Neutron Scattering Study of

Khandpur, A.K.; Forster, S. ; Bates, F.S.; Hamley, I.W.; Ryan, A.J.; Bras, W.; Almadal, K.;

Matsen, M. W. & Bates, F. S. (1996). Unifying Weak- and Strong-Segregation Block

Modi, M.A.; Krishanamoorti, R.; Tse, M.F.; Wang, H.C. (1999). Viscoelastic Characterization

Copolymer Theories. *Macromolecules*, Vol. 29, pp. 1091-1098.

*Macromolecules*, Vol. 32(12), pp. 4088-4097.

Mechanism of the Transition. *Macromolecules*, Vol. 33(3), pp. 968-977. Kim, E.Y.; Lee, D.J.; Kim, J.K. (2006). Phase Behavior of a Binary Mixture of a Block

sispension under steady shear. *Europhys. Lett*., Vol. 21, 363-368.

interactions in block copolymer microphase separation. *J. Chem. Phys*., Vol. 110,

A Reevaluation of Bicontinuous Cubic Phases in Starblock Copolymers.

Spherical and Cylindrical Microdomain Structures of Block Copolymer. I.

Copolymer with Lower Disorder-to-Order Transition and a Homopolymer.

and Order−Disorder Transition Temperatures of SIS Block Copolymers by Differential Scanning Calorimetry and Rheology. *Macromolecules*, Vol. 31(12), pp.

Cylinder-to-Sphere Order−Order Transition in Block Copolymers. *Macromolecules*,

a Cylinder to Sphere order-order Transition in Block Copolymers. *Macromolecules*,

Mortensen, K. (1995). Polyisoprene-Polystyrene Diblock Copolymer Phase Diagram near the Order-Disorder Transition. *Macromolecules*, Vol 28(26), pp. 8796-8806. Leibler, L. (1980). Theory of Microphase Separation in Block Copolymers. *Macromolecules*,

of an Order−Order Transition in a Mixture of Di- and Triblock Copolymers.

separation. *J. Chem. Phys*., Vol. 108, pp. 8713-8724.

*Macromolecules*, Vol. 28, pp. 2570-2573.

*Macromolecules*, Vol. 39(25), pp. 8747-8757.

York,

pp.9739-9749.

019850218 4.

4045-4048.

Vol. 33, pp. 3810-3817.

Vol. 33, pp. 3803-3809.

Vol. 13(6), pp. 1602-1607.


from gyroid-to-lamellar states of PS–PI diblock copolymer. *Chemical Physics Letters*, Vol. 460, pp. 507-511.


Thomas, E.L. & Lescanec, R.L. (1995). Phase morphology in block copolymer systems, In A.

Verlet, L. (1967). Computer "Experiments" on Classical Fluids. I. Thermodynamical Properties of Lennard- Jones Molecules. *Phys. Rev*., Vol.159, pp. 98-103. Vicente, L.; Soto-Figueroa, C.; Pacheco-Sanchez, H.; Hernadez-Trujillo, J.; Martínez Magadán, J. M. (2006). Fluid Phase Equilibria. Vol. 239, pp. 100-106. You, L. Y.; Chen, L.J.; Qian, H.J.; Lu, Z.Y. Microphase Transitions of Perforated Lamellae of

Strobl, G. (1997). The Physic of Polymers. Springer, ISBN 3-540-63203-4.

Materials, p.147-164. Chapman & Hall, London.

Vol. 460, pp. 507-511.

pp. 5222-5227.

from gyroid-to-lamellar states of PS–PI diblock copolymer. *Chemical Physics Letters*,

Keller, M. Warner and A.H. Windle (Eds.), Self-order and Form in Polymeric

Cyclic Diblock Copolymers under Steady Shear. *Macromolecules*, 2007, Vol. 40(14),

### *Edited by Zeeshan Nawaz and Shahid Naveed*

Chemical engineering applications have been a source of challenging optimization problems in terms of economics and technology. The goal of this book is to enable the reader to get instant information on fundamentals and advancements in chemical engineering. This book addresses ongoing evolutions of chemical engineering and provides overview to the sate of the art advancements. Molecular perspective is increasingly important in the refinement of kinetic and thermodynamic molding. As a result, much of the material was revised on industrial problems and their sophisticated solutions from known scientists around the world. These issues were divided in to two sections, fundamental advances and catalysis and reaction engineering. A distinct feature of this text continues to be the emphasis on molecular chemistry, reaction engineering and modeling to achieve rational and robust industrial design. Our perspective is that this background must be made available to undergraduate, graduate and professionals in an integrated manner.

Advances in

Chemical Engineering

*Edited by Zeeshan Nawaz and Shahid Naveed*

ISBN 978-953-51-0392-9

ISBN 978-953-51-6174-5

Advances in Chemical Engineering

Photo by Grafner / iStock