**Part 1**

## **Fundamentals**

**1** 

*Iran* 

**Application of Chebyshev** 

**Polynomials to Calculate Density and** 

Seyyed Alireza Tabatabaei-Nejad and Elnaz Khodapanah

**Fugacity Using SAFT Equation of State to** 

 *Chemical Engineering Department, Sahand University of Technology, Tabriz* 

**Predict Asphaltene Precipitation Conditions** 

Equations of state are the essential tools to model physical and chemical processes in which fluids are involved. The majority of PVT calculations carried out for oil and gas mixtures are based on a cubic equation of state (EoS). This type of equations dates back more than a century to the famous Van der Waals equation (Van der Waals, 1873). The cubic equations of state most commonly used in the petroleum industry today are very similar to the Van der Waals equation, but it took almost a century for the petroleum industry to accept this type of equation as a valuable engineering tool. The Redlich and Kwong EoS (Redlich & Kwong, 1949) was modified from the VdW with a different attractive term, the repulsive term being the same. Since 1949 when Redlich and Kwong (RK) formulated their two-parameter cubic EoS, many investigators have introduced various modifications to improve ability of RK-EoS. Two other well-known cubic equations are Soave-Redlich-Kwong (SRK), (Soave, 1972) and Peng-Robinson (PR) (Peng & Robinson, 1976) equations which have different formulation of the attractive term and are popular in the oil industry in the thermodynamic

There are thousands of cubic equations of states, and many noncubic equations. The noncubic equations such as the Benedict-Webb-Rubin equation (Benedict et al., 1942), and its modification by Starling (Starling, 1973) have a large number of constants; they describe accurately the volumetric behavior of pure substances. But for hydrocarbon mixtures and crude oils, because of mixing rule complexities, they may not be suitable (Katz & Firoozabadi, 1978). Cubic equations with more than two constants also may not improve the volumetric behavior prediction of complex reservoir fluids. In fact, most of the cubic equations have the same accuracy for phase behavior prediction of complex hydrocarbon

Hydrocarbons and other non-polar fluid vapor–liquid equilibrium properties can be satisfactorily modeled using a symmetric approach to model both, the vapor and the liquid phase fugacity with the use of a Van der Waals type equation model (Segura et al., 2008), the Soave–Redlich–Kwong or Peng–Robinson equations being the most popular ones. When

**1. Introduction** 

modeling of hydrocarbon fluids.

systems; the simpler often do better (Firoozabadi, 1999).

## **Application of Chebyshev Polynomials to Calculate Density and Fugacity Using SAFT Equation of State to Predict Asphaltene Precipitation Conditions**

Seyyed Alireza Tabatabaei-Nejad and Elnaz Khodapanah  *Chemical Engineering Department, Sahand University of Technology, Tabriz Iran* 

#### **1. Introduction**

Equations of state are the essential tools to model physical and chemical processes in which fluids are involved. The majority of PVT calculations carried out for oil and gas mixtures are based on a cubic equation of state (EoS). This type of equations dates back more than a century to the famous Van der Waals equation (Van der Waals, 1873). The cubic equations of state most commonly used in the petroleum industry today are very similar to the Van der Waals equation, but it took almost a century for the petroleum industry to accept this type of equation as a valuable engineering tool. The Redlich and Kwong EoS (Redlich & Kwong, 1949) was modified from the VdW with a different attractive term, the repulsive term being the same. Since 1949 when Redlich and Kwong (RK) formulated their two-parameter cubic EoS, many investigators have introduced various modifications to improve ability of RK-EoS. Two other well-known cubic equations are Soave-Redlich-Kwong (SRK), (Soave, 1972) and Peng-Robinson (PR) (Peng & Robinson, 1976) equations which have different formulation of the attractive term and are popular in the oil industry in the thermodynamic modeling of hydrocarbon fluids.

There are thousands of cubic equations of states, and many noncubic equations. The noncubic equations such as the Benedict-Webb-Rubin equation (Benedict et al., 1942), and its modification by Starling (Starling, 1973) have a large number of constants; they describe accurately the volumetric behavior of pure substances. But for hydrocarbon mixtures and crude oils, because of mixing rule complexities, they may not be suitable (Katz & Firoozabadi, 1978). Cubic equations with more than two constants also may not improve the volumetric behavior prediction of complex reservoir fluids. In fact, most of the cubic equations have the same accuracy for phase behavior prediction of complex hydrocarbon systems; the simpler often do better (Firoozabadi, 1999).

Hydrocarbons and other non-polar fluid vapor–liquid equilibrium properties can be satisfactorily modeled using a symmetric approach to model both, the vapor and the liquid phase fugacity with the use of a Van der Waals type equation model (Segura et al., 2008), the Soave–Redlich–Kwong or Peng–Robinson equations being the most popular ones. When

Application of Chebyshev Polynomials to Calculate Density and

bonding and hydrogen bonding (Chapman et al., 2004).

unstable (Verdier et al., 2006).

Fugacity Using SAFT Equation of State to Predict Asphaltene Precipitation Conditions 5

experimental system. A third advantage of SAFT-type equations versus other approaches is that, as they are based on statistical mechanics, parameters have a clear physical meaning; when carefully fitted they can be used with predictive power to explore other regions of the phase diagram far from the data and operating conditions used in the parameter regression, performing better than other models for interacting compounds like activity coefficient models (Prausnitz et al., 1999). In SAFT a chain molecule is characterized by the diameter or volume of a segment, the number of segments in the chain, and the segment–segment dispersion energy. For an associating or hydrogen bonding molecule, two more physical parameters are necessary: the association energy related to the change in enthalpy of association and the bond volume related to the change in entropy on association. The SAFT equation has found some impressive engineering applications on those fluids with chain

Asphaltenes are operationally defined as the portion of crude oil insoluble in light normal alkanes (e.g., n-heptane or n-pentane), but soluble in aromatic solvents (e.g., benzene or toluene). This solubility class definition of asphaltenes suggests a broad distribution of asphaltene molecular structures that vary greatly among crude sources. In general, asphaltenes possess fused ring aromaticity, small aliphatic side chains, and polar heteroatom-containing functional groups capable of donating or accepting protons interand intra-molecularly. Although asphaltene fractions can be complex molecular species mixtures, they convey, as a whole, an obvious chemical similarity, irrespective of crude geographic origin (Ting, 2003). Asphaltene stability depends on a number of factors including pressure, temperature, and compositions of the fluid; the latter incorporates the addition of light gases, solvents and other oils commingled operation or charges due to contamination. During pressure depletion at constant temperature, asphaltene aggregate formation is observed within a range above and below the bubble point. As pressure drops during production from the reservoir pressure, asphaltene precipitatin can appear due to changes in the solubility of asphaltene in crude oil. The maximum asphaltene precipitation occurs at or around the bubble point pressure. Below the bubble point light gases come out of the solution increasing the asphaltene solubility again (Ting, 2003). Temperature changes also affect asphaltene precipitation, For hydrocarbons deposited in shallow structure, the wellhead flowing temperatures are typically not excessive, 110-140 °F. However, sea bottom temperature in deep water is cold, often near or below 40 °F, even in equatorial waters. Cooling of flow streams during transportation can lead to asphaltene precipitation (Huang & Radosz, 1991). Increases in temperature at constant pressure normally stabilize the asphaltene in crude oil. Depending on the composition of the oil, it is possible to find cases where precipitation first decreases and then increases with increasing temperature (Verdier et al., 2006). Also, depending on the temperature level, significant temperature effects can be observed (Buenrostro-Gonzales & Lira-Galeana, 2004). Changes in composition occur during gas injection processes employed in Enhanced Oil Recovery (EOR). Gas injection includes processes such as miscible flooding with CO2 , N2 or natural gas or artificial gas lifting. The dissolved gas decreases asphaltene solubility and the asphaltene becomes more

The tendency of petroleum asphaltenes to associate in solution and adsorb at interfaces can cause significant problems during the production, recovery, pipeline transportation, and refining of crude oil. Therefore, it is necessary to predict the conditions where precipitation

polar fluids are involved at moderate pressures, activity coefficient models are more suitable for modeling the liquid phase. When a higher pressure range is also a concern, a symmetric EoS approach with complex mixing rules including an excess Gibbs energy term from an activity coefficient model can provide good results. Unfortunately, even those approaches show limitations for complex fluids and can drastically fail near the critical region, unless a specific treatment is included (Llovell et al., 2004, 2008).

Since the early 1980's, there has been increased interest in developing an EoS for pure fluids and mixtures of large polyatomic molecules that does not rely on a lattice description of molecular configurations. A rigorous statistical-mechanical theory for large polyatomic molecules in continuous space is difficult because of their asymmetric structure, large number of internal degrees of freedom, and strong coupling between intra- and intermolecular interactions. Nevertheless, a relatively simple model represents chain-like as freely joined tangent hard spheres (Chapman et al., 1984; Song et al., 1994; Wertheim, 1984). A hard-sphere-chain (HSC) EoS can be used as the reference system in place of the hardsphere reference used in most existing equations of state for simple fluids. Despite their simplicity, hard-sphere-chain models take into account some significant features of real fluids containing chain-like molecules including excluded volume effects and chain connectivity. To describe the properties of fluids consisting of large polyatomic molecules, it is necessary to introduce attractive forces by adding a perturbation to a HSC EoS. Assuming that the influence of attractive forces on fluid structure is week, a Van der Waals type or other mean-field term (e.g. square-well fluids) is usually used to add attractive forces to the reference hard-sphere-chain EoS (Prausnitz & Tavares, 2004).

Molecular-based equations of state, also routed in statistical mechanics, retain their interest in chemical engineering calculations as they apply to a wide spectrum of thermodynamic conditions and compounds, being computationally much less demanding than molecular simulations. Among them, the Statistical Associating Fluid Theory (SAFT) EoS has become very popular because of its capability of predicting thermodynamics properties of several complex fluids, including chain, aromatic and chlorinated hydrocarbons, esters alkanols, carboxylic acids, etc. (Huang & Radosz, 1990). SAFT was envisioned as an application of Wertheim's theory of association (Wertheim, 1984, 1986) through the use of a first-order thermodynamic perturbation theory (TPT) to formulate a physically based EoS (Chapman et al., 1990; Huang & Radosz, 1991). The ambition of making SAFT an accurate equation for engineering purposes has promoted the development of different versions that tried to overcome the limitations of the original one (Economou, 2002; Muller & Gubbins, 1995).

SAFT has a similar form to group contribution theories in that the fluid of interest is initially considered to be a mixture of unconnected groups or segments. SAFT includes a chain connectivity term to account for the bonding of various groups to form polymers and an explicit intermolecular hydrogen bonding term. A theory based in statistical mechanics offers several advantages. The first advantage is that each of the approximations made in the development of SAFT has been tested versus molecular simulation results. In this way, the range of applicability of each term in the EoS has been determined. The second advantage is that the EoS can be systematically refined. Since any weak approximations in SAFT can be identified, improvement is made upon the EoS by making better approximations or by extending the theory. Like most thermodynamic models, SAFT approaches require the evaluation of several parameters relating the model to the

polar fluids are involved at moderate pressures, activity coefficient models are more suitable for modeling the liquid phase. When a higher pressure range is also a concern, a symmetric EoS approach with complex mixing rules including an excess Gibbs energy term from an activity coefficient model can provide good results. Unfortunately, even those approaches show limitations for complex fluids and can drastically fail near the critical region, unless a

Since the early 1980's, there has been increased interest in developing an EoS for pure fluids and mixtures of large polyatomic molecules that does not rely on a lattice description of molecular configurations. A rigorous statistical-mechanical theory for large polyatomic molecules in continuous space is difficult because of their asymmetric structure, large number of internal degrees of freedom, and strong coupling between intra- and intermolecular interactions. Nevertheless, a relatively simple model represents chain-like as freely joined tangent hard spheres (Chapman et al., 1984; Song et al., 1994; Wertheim, 1984). A hard-sphere-chain (HSC) EoS can be used as the reference system in place of the hardsphere reference used in most existing equations of state for simple fluids. Despite their simplicity, hard-sphere-chain models take into account some significant features of real fluids containing chain-like molecules including excluded volume effects and chain connectivity. To describe the properties of fluids consisting of large polyatomic molecules, it is necessary to introduce attractive forces by adding a perturbation to a HSC EoS. Assuming that the influence of attractive forces on fluid structure is week, a Van der Waals type or other mean-field term (e.g. square-well fluids) is usually used to add attractive forces to the

Molecular-based equations of state, also routed in statistical mechanics, retain their interest in chemical engineering calculations as they apply to a wide spectrum of thermodynamic conditions and compounds, being computationally much less demanding than molecular simulations. Among them, the Statistical Associating Fluid Theory (SAFT) EoS has become very popular because of its capability of predicting thermodynamics properties of several complex fluids, including chain, aromatic and chlorinated hydrocarbons, esters alkanols, carboxylic acids, etc. (Huang & Radosz, 1990). SAFT was envisioned as an application of Wertheim's theory of association (Wertheim, 1984, 1986) through the use of a first-order thermodynamic perturbation theory (TPT) to formulate a physically based EoS (Chapman et al., 1990; Huang & Radosz, 1991). The ambition of making SAFT an accurate equation for engineering purposes has promoted the development of different versions that tried to overcome the limitations of the original one (Economou, 2002; Muller & Gubbins, 1995).

SAFT has a similar form to group contribution theories in that the fluid of interest is initially considered to be a mixture of unconnected groups or segments. SAFT includes a chain connectivity term to account for the bonding of various groups to form polymers and an explicit intermolecular hydrogen bonding term. A theory based in statistical mechanics offers several advantages. The first advantage is that each of the approximations made in the development of SAFT has been tested versus molecular simulation results. In this way, the range of applicability of each term in the EoS has been determined. The second advantage is that the EoS can be systematically refined. Since any weak approximations in SAFT can be identified, improvement is made upon the EoS by making better approximations or by extending the theory. Like most thermodynamic models, SAFT approaches require the evaluation of several parameters relating the model to the

specific treatment is included (Llovell et al., 2004, 2008).

reference hard-sphere-chain EoS (Prausnitz & Tavares, 2004).

experimental system. A third advantage of SAFT-type equations versus other approaches is that, as they are based on statistical mechanics, parameters have a clear physical meaning; when carefully fitted they can be used with predictive power to explore other regions of the phase diagram far from the data and operating conditions used in the parameter regression, performing better than other models for interacting compounds like activity coefficient models (Prausnitz et al., 1999). In SAFT a chain molecule is characterized by the diameter or volume of a segment, the number of segments in the chain, and the segment–segment dispersion energy. For an associating or hydrogen bonding molecule, two more physical parameters are necessary: the association energy related to the change in enthalpy of association and the bond volume related to the change in entropy on association. The SAFT equation has found some impressive engineering applications on those fluids with chain bonding and hydrogen bonding (Chapman et al., 2004).

Asphaltenes are operationally defined as the portion of crude oil insoluble in light normal alkanes (e.g., n-heptane or n-pentane), but soluble in aromatic solvents (e.g., benzene or toluene). This solubility class definition of asphaltenes suggests a broad distribution of asphaltene molecular structures that vary greatly among crude sources. In general, asphaltenes possess fused ring aromaticity, small aliphatic side chains, and polar heteroatom-containing functional groups capable of donating or accepting protons interand intra-molecularly. Although asphaltene fractions can be complex molecular species mixtures, they convey, as a whole, an obvious chemical similarity, irrespective of crude geographic origin (Ting, 2003). Asphaltene stability depends on a number of factors including pressure, temperature, and compositions of the fluid; the latter incorporates the addition of light gases, solvents and other oils commingled operation or charges due to contamination. During pressure depletion at constant temperature, asphaltene aggregate formation is observed within a range above and below the bubble point. As pressure drops during production from the reservoir pressure, asphaltene precipitatin can appear due to changes in the solubility of asphaltene in crude oil. The maximum asphaltene precipitation occurs at or around the bubble point pressure. Below the bubble point light gases come out of the solution increasing the asphaltene solubility again (Ting, 2003). Temperature changes also affect asphaltene precipitation, For hydrocarbons deposited in shallow structure, the wellhead flowing temperatures are typically not excessive, 110-140 °F. However, sea bottom temperature in deep water is cold, often near or below 40 °F, even in equatorial waters. Cooling of flow streams during transportation can lead to asphaltene precipitation (Huang & Radosz, 1991). Increases in temperature at constant pressure normally stabilize the asphaltene in crude oil. Depending on the composition of the oil, it is possible to find cases where precipitation first decreases and then increases with increasing temperature (Verdier et al., 2006). Also, depending on the temperature level, significant temperature effects can be observed (Buenrostro-Gonzales & Lira-Galeana, 2004). Changes in composition occur during gas injection processes employed in Enhanced Oil Recovery (EOR). Gas injection includes processes such as miscible flooding with CO2 , N2 or natural gas or artificial gas lifting. The dissolved gas decreases asphaltene solubility and the asphaltene becomes more unstable (Verdier et al., 2006).

The tendency of petroleum asphaltenes to associate in solution and adsorb at interfaces can cause significant problems during the production, recovery, pipeline transportation, and refining of crude oil. Therefore, it is necessary to predict the conditions where precipitation

$$A^R = A\_{hs} + A\_{ch} + A\_{dlsp} + A\_{assoc} \tag{1}$$

$$Z = \frac{P}{\rho RT} = Z^{ld} + Z\_{hs} + Z\_{ch} + Z\_{slsp} + Z\_{Assoc} \tag{2}$$

$$Z\_{hs} = \frac{6}{\pi N\_A \rho} \left[ \frac{\xi\_0 \underline{\xi}\_3}{1 - \underline{\xi}\_3} + \frac{3 \underline{\xi}\_1 \underline{\xi}\_2}{(1 - \underline{\xi}\_3)^2} + \frac{(3 - \underline{\xi}\_3) \underline{\xi}\_2^3}{(1 - \underline{\xi}\_3)^2} \right] \tag{3}$$

$$\ddot{\varepsilon}\_{\mathbf{k}} = \frac{\pi N\_A \rho}{6} \sum\_{l=1}^{N\_c} z\_l r\_l (d\_l)^k \qquad \qquad \qquad \qquad \qquad \qquad \qquad \qquad \qquad \qquad (4)$$

$$d\_l = \sigma\_l \left[ 1 - 0.12 \exp(-3\varepsilon\_l/kT) \right] \tag{5}$$

$$Z\_{hs} = \frac{h\mathbf{s}\_1\rho + h\mathbf{s}\_2\rho^2 + h\mathbf{s}\_3\rho^3}{1 + h\mathbf{s}\_4\rho + h\mathbf{s}\_5\rho^2 + h\mathbf{s}\_6\rho^3} \tag{6}$$

$$h\mathbf{s}\_1 = (\vec{r}\mathbf{S}\_3 + 3\mathbf{S}\_1\mathbf{S}\_2)P\_{na} \tag{7}$$

$$h\mathbf{S}\_2 = \{-2\vec{r}\mathbf{S}\_3^2 - 3\mathbf{S}\_1\mathbf{S}\_2\mathbf{S}\_3 + 3\mathbf{S}\_2^3\}\mathbf{P}\_{na}^2\tag{8}$$

$$\mathbf{h}\mathbf{s}\_3 = \{\overline{r}\mathbb{S}\_3^3 - \mathbb{S}\_2^3\mathbb{S}\_3\} P\_{na}^3 \tag{9}$$

$$h\mathbf{s}\_4 = -\mathbf{3}\mathbf{S}\_3 P\_{na} \tag{10}$$

$$h\mathbf{s}\_5 = 3\mathbf{S}\_3^2 P\_{na}^2\tag{11}$$

$$h\mathbf{s}\_6 = -\mathbf{S}\_3^3 P\_{na}^3 \tag{12}$$

$$\vec{r} = \sum\_{l=1}^{N\_C} z\_l r\_l \tag{13}$$

$$\mathcal{S}\_1 = \sum\_{l=1}^{N\_c} z\_l r\_l d\_l \tag{14}$$

$$\mathcal{S}\_2 = \sum\_{l=1}^{N\_c} z\_l r\_l d\_l^2 \tag{15}$$

$$\mathcal{S}\_3 = \sum\_{l=1}^{N\_c} z\_l r\_l \, d\_l^3 \tag{16}$$

$$P\_{na} = \frac{\pi N\_A}{6} \tag{17}$$

$$Z\_{ch} = \sum\_{l=1}^{N\_c} z\_l (1 - r\_l) L(d\_l) \tag{18}$$

$$L(d\_l) = \frac{2\xi\_3 + 3d\_l\xi\_2 - 4\xi\_3^2 + 2d\_l^2\xi\_2^2 + 2\xi\_3^3 + d\_l^2\xi\_2^2\xi\_3 - 3d\_l\xi\_2\xi\_3^2}{\left(1 - \xi\_3\right)\left(2 - 4\xi\_3 + 3d\_l\xi\_2 + 2\xi\_3^2 + d\_l^2\xi\_2^2 - 3d\_l\xi\_2\xi\_3\right)}\tag{19}$$

$$Z\_{ch} = \sum\_{l=1}^{N\_c} z\_l (1 - \tau\_l) \frac{c h\_1(d\_l) \rho + c h\_2(d\_l) \rho^2 + c h\_3(d\_l) \rho^3}{2 + c h\_4(d\_l) \rho + c h\_5(d\_l) \rho^2 + c h\_6(d\_l) \rho^3 + c h\_7(d\_l) \rho^4} \tag{20}$$

$$ch\_1(d\_l) = [2S\_3 + 3d\_lS\_2]P\_{na} \tag{21}$$

$$ch\_2(d\_l) = \left[ -4S\_3^2 + 2d\_l^2 S\_2^2 \right] P\_{na}^2 \tag{22}$$

$$ch\_3(d\_l) = \left[d\_l^2 \mathcal{S}\_2^2 \mathcal{S}\_3 + 2\mathcal{S}\_3^3 - 3d\_l \mathcal{S}\_2 \mathcal{S}\_3^2\right] P\_{na}^3 \tag{23}$$

$$ch\_4(d\_l) = [-6S\_3 + 3d\_lS\_2]P\_{na} \tag{24}$$

$$ch\_5(d\_l) = \left[6S\_3^2 + d\_l^2 S\_2^2 - 9d\_l S\_2 S\_3\right] P\_{na}^2\tag{25}$$

$$ch\_6(d\_l) = \left[ 9d\_l \mathcal{S}\_2 \mathcal{S}\_3^2 - 2\mathcal{S}\_3^3 - d\_l^2 \mathcal{S}\_2^2 \mathcal{S}\_3 \right] P\_{na}^3 \tag{26}$$

$$ch\_7(d\_l) = [-3d\_l \mathbb{S}\_2 \mathbb{S}\_3^3] P\_{na}^3 \tag{27}$$

$$Z\_{\rm disp} = -2\pi\rho \frac{\partial \left(\underline{\xi}\_3 I\_1\right)}{\partial \left(\underline{\xi}\_3\right)} \overline{r^2 \varepsilon \sigma^3} - \pi\rho \vec{r} \left[\mathcal{C}\_1 \frac{\partial \left(\underline{\xi}\_3 I\_2\right)}{\partial \left(\underline{\xi}\_3\right)} + \mathcal{C}\_2 \underline{\xi}\_3 I\_2\right] \overline{r^2 \varepsilon^2 \sigma^3} \tag{28}$$

$$\mathcal{C}\_1 = 1 + \vec{r} \frac{8\xi\_3 - 2\xi\_3^2}{\left(1 - \xi\_3\right)^4} + (1 - \vec{r}) \frac{20\xi\_3 - 27\xi\_3^2 + 12\xi\_3^3 - 2\xi\_3^4}{\left[\left(1 - \xi\_3\right)\left(2 - \xi\_3\right)\right]^2} \tag{29}$$

$$\mathcal{C}\_2 = -\mathcal{C}\_1^2 \left[ \vec{r} \frac{-4\xi\_3^2 + 20\xi\_3 + 8}{\left(1 - \xi\_3\right)^5} + (1 - \vec{r}) \frac{2\xi\_3^3 + 12\xi\_3^2 - 48\xi\_3 + 40}{\left[\left(1 - \xi\_3\right)\left(2 - \xi\_3\right)\right]^3} \right] \tag{30}$$

$$\overrightarrow{r^2 \varepsilon \sigma^3} = \sum\_{l=1}^{N\_c} \sum\_{j=1}^{N\_c} z\_l z\_{l'} r\_{l'} \left(\frac{\varepsilon\_{lj}}{kT}\right) \sigma\_{lj}^3 \tag{31}$$

$$\overline{r^2 \varepsilon^2 \sigma^3} = \sum\_{l=1}^{N\_c} \sum\_{j=1}^{N\_c} z\_l z\_{l} r\_{l} r\_{j} \left(\frac{\varepsilon\_{lj}}{kT}\right)^2 \sigma\_{lj}^3 \tag{32}$$

$$I\_1 = \sum\_{j=0}^{6} a\_j(\vec{r}) \,\xi\_3^j \tag{33}$$

$$I\_2 = \sum\_{j=0}^{6} b\_j(\vec{r}) \,\vec{\xi}\_3^j \tag{34}$$

$$
\varepsilon\_{lj} = \sqrt{\varepsilon\_l \varepsilon\_j} (1 - k\_{lj}) \tag{35}
$$

$$
\sigma\_{lj} = \frac{1}{2} \left( \sigma\_l + \sigma\_j \right) \tag{36}
$$

$$a\_j(\vec{r}) = a\_{0\parallel} + \frac{\vec{r} - 1}{\vec{r}} a\_{1\parallel} + \frac{\vec{r} - 1}{\vec{r}} . \frac{\vec{r} - 2}{\vec{r}} a\_{2\parallel} \; , \quad j = 0, 1, \ldots, 6 \tag{37}$$

$$b\_{\vec{l}}(\vec{r}) = b\_{0\vec{l}} + \frac{\vec{r} - 1}{\vec{r}} b\_{1\vec{l}} + \frac{\vec{r} - 1}{\vec{r}} . \frac{\vec{r} - 2}{\vec{r}} b\_{2\vec{l}} \; , \quad \vec{j} = 0, 1, \dots, 6 \tag{38}$$


$$d\text{disp}\_1 = 4(\Im \vec{r} - 4)\mathbb{S}\_3 P\_{na} \tag{40}$$

$$\text{disp}\_2 = \{27\vec{r} - 6\} \mathcal{S}\_3^2 P\_{na}^2 \tag{41}$$

$$\text{disp}\_3 = (\\$2 - 70\vec{r}) S\_3^3 P\_{na}^3 \tag{42}$$

$$d\text{lsp}\_4 = (51\vec{r} - 52)S\_3^4 P\_{na}^4 \tag{43}$$

$$disp\_5 = 16(1 - \vec{r})S\_3^5 P\_{na}^5 \tag{44}$$

$$disp\_6 = \mathcal{Z} \{ \vec{r} - 1 \} \mathcal{S}\_3^b P\_{na}^b \tag{45}$$

$$\text{disp}\_7 = -2\mathbf{0}\_3 \mathbf{S}\_3 P\_{na} \tag{46}$$

$$disp\_{\mathfrak{B}} = 4\,\mathbf{1}\mathbf{S}\_3^2 P\_{na}^2\tag{47}$$

$$disp\_9 = -44S\_3^3 P\_{na}^3\tag{48}$$

$$d\text{disp}\_{10} = 2\,\text{6S}\_3^4 P\_{na}^4\tag{49}$$

$$disp\_{11} = -8S\_3^5 P\_{na}^5\tag{50}$$

$$d\text{disp}\_{12} = \mathbb{S}\_3^6 P\_{na}^6 \tag{51}$$

$$\text{disp}\_{13} = 40 + 24\vec{r} \tag{52}$$

$$\text{disp}\_{14} = \{192\overline{r} - 128\} \mathbb{S}\_3 P\_{na} \tag{53}$$

$$d\text{lsp}\_{15} = (148 - 372\vec{r})S\_3^2 P\_{na}^2 \tag{54}$$

$$d\text{lsp}\_{16} = \{230\vec{r} - 70\} \mathcal{S}\_3^3 P\_{na}^3 \tag{55}$$

$$disp\_{17} = (8 - 52\vec{r})S\_3^4 P\_{na}^4\tag{56}$$

$$dl\_2 = -12P\_{na}.A\text{disp}\_1.dl\,\text{disp}\_1\tag{79}$$

$$dl\_3 = -12P\_{na}.A\,\text{disp}\_1.dl\,\text{disp}\_2\tag{80}$$

$$d\mathcal{I}\_4 = -12P\_{na}.A\text{disp}\_1.d\text{ldisp}\_3\tag{81}$$

$$dl\_5 = -12P\_{na}.ddisp\_1.dldisp\_4\tag{82}$$

$$dl\_6 = -12P\_{na}.ddisp\_1.dIdisp\_5\tag{83}$$

$$dl\_7 = -12P\_{na}.ddisp\_1.dldisp\_6\tag{84}$$

$$dI\_8 = \mathcal{Z}b\_1\mathcal{S}\_3P\_{na} \tag{85}$$

$$dI\_9 = 3b\_2 \mathcal{S}\_3^2 P\_{na}^2 \tag{86}$$

$$dl\_{10} = 4b\_3 \mathcal{S}\_3^3 P\_{na}^3 \tag{87}$$

$$dI\_{11} = \mathbb{S}b\_4 \mathbb{S}\_3^4 P\_{na}^4 \tag{88}$$

$$dl\_{12} = 6b\_5 \mathcal{S}\_3^5 P\_{na}^5 \tag{89}$$

$$dI\_{13} = 7b\_6 \mathbb{S}\_3^6 P\_{na}^6 \tag{90}$$

$$AdSp\_1 = \overline{r^2 \varepsilon \sigma^3} = \sum\_{l=1}^{N\_c} \sum\_{j=1}^{N\_c} \mathbf{z}\_l \mathbf{z}\_j r\_l r\_l \{\varepsilon\_{lj} / kT\} \sigma\_{lj}^3 \tag{91}$$

$$AdSp\_2 = \overline{r^2}\overline{\varepsilon^2}\overline{\sigma^3} = \sum\_{l=1}^{N\_c} \sum\_{j=1}^{N\_c} \mathbf{z}\_l \mathbf{z}\_j r\_l r\_l \left(\varepsilon\_{lj} / kT\right)^2 \sigma\_{lj}^3 \tag{92}$$

$$Z\_{assoc} = \rho \sum\_{l=1}^{N\_c} z\_l \left[ \sum\_{S\_l} \left( \frac{1}{X^{S\_l}} - \frac{1}{2} \right) \frac{\partial X^{S\_l}}{\partial \rho} \right] \tag{93}$$

$$X^{S\_l} = \left(1 + N\_A \sum\_{f=1}^{N\_c} \sum\_{Y\_f} z\_l \rho X^{Y\_f} \,\mathcal{W}\_{lj}\right)^{-1} \tag{94}$$

$$\mathcal{W}\_{lj} = \left| \frac{1}{1 - \underline{\zeta}\_3} + \frac{3d\_l d\_{\bar{l}}}{d\_l + d\_{\bar{l}}} \frac{\underline{\zeta}\_2}{\left(1 - \underline{\zeta}\_3\right)^2} + 2 \left(\frac{d\_l d\_{\bar{l}}}{d\_l + d\_{\bar{l}}}\right)^2 \frac{\underline{\zeta}\_2^2}{\left(1 - \underline{\zeta}\_3\right)^3} \right| \left(\sigma\_{lj} k^{S\bar{l}\bar{l}}\right) \left[\exp\left(\frac{\varepsilon^{S\bar{l}\bar{l}}}{kT}\right) - 1\right] \tag{95}$$

Application of Chebyshev Polynomials to Calculate Density and

where � refers to the total number of moles of the known phase.

fugacity of the components:

sphere contribution of SAFT:

*hs*

*n*

equations:

, ,

*k TVn*

���� = (����̅��

*j k*

*k*

 

Fugacity Using SAFT Equation of State to Predict Asphaltene Precipitation Conditions 15

���� (99)

�� (100)

� � �.

 

> 

�)���

� (106)

� (105)

(103)

�� <sup>=</sup> �

�� <sup>=</sup> ��

To use equation (98), we require a suitable EoS that holds for the entire range of possible mole fractions *z* at the system temperature and for the density range between 0 and �

Application of the SAFT EoS in Eq. (98) yields the following equation for calculating the

*V n V n*

*RT nZ RT nZ RTln Z dV dV V n V n*

 

*V V k k TVn TVn*

*disp assoc*

*RT nZ RT nZ dV dV*

1,2, ,

The following equations are derived for the first term in Eq. (101) accounting for the hard

*hs hs V k k TVn TVn*

*c*

11 22 33

� � ��̅����� � �������� � �������� � �������� � ������

2. 3. 4.

. 2. 3. . . .

 

\_

1,2, ,

1 2 2 3 2 3

 

*kk k*

44 5 5 6 6 1 2 3

The parameters used in Eq. (103) are given by equations (7) – (17) and the following

���� � ���

*hs hs hs hs hs hs hs hs hs*

*nZ hs hs hs hs hs hs*

45 6 45 6

 

� � ��̅��

1. . . 1. . .

*hs hs hs hs hs hs*

 

*kk k*

���� = (��̅��

*k N*

*RT nZ RT nZ IR HS dV d V n n k N*

, , , ,

*j k j k*

, , , ,

*c*

*hs ch*

*V V k k TVn TVn*

*j k j k*

, , 0 , ,

 

*j k j k*

2 3 2 3

*hs r S rS S S S S P* 1 3 3 1 2 12 *kk k k* 3 3 *k na* (104)

������ � ��

 

> 

����)���

���� = �������� (107)

2 3

(102)

(101)

parameter ����� characterize, respectively, the association (���) and solvation (���) energy and volume for the specific interaction between sites � and �. These parameters are adjustable. Equation (93) requires no mixing rules. As it can be seen in Eq. (94), ��� ' s satisfy a non-linear system of equations which can be solved using any iterative technique such as Gauss-Seidel, Successive-Over-Relaxation (SOR) or Jacobi iterative method. The derivative

of the function ��� with respect to � yields the following equation:

$$\begin{aligned} \left(\frac{\partial X^{S\_i}}{\partial \rho}\right) = -\left(X^{S\_i}\right)^2 N\_A \left[ \begin{aligned} &\sum\_{j=1}^{N\_c} \sum\_{Y\_j} x\_j X^{Y\_j} W\_{\vec{\eta}j} + \sum\_{j=1}^{N\_c} \sum\_{Y\_j} z\_j \rho X^{Y\_j} \left(\frac{\partial \mathcal{W}\_{\vec{\eta}j}}{\partial \rho}\right) \\ &+ \sum\_{j=1}^{N\_c} \sum\_{Y\_j} z\_j \rho \mathcal{W}\_{\vec{\eta}j} X^{Y\_j} \left(\frac{\partial X^{Y\_j}}{\partial \rho}\right) \end{aligned} \tag{96}$$

where,

$$
\begin{aligned}
\left(\frac{\partial \mathcal{W}\_{\vec{\eta}}}{\partial \rho}\right) &= \left(\sigma\_{\vec{\eta}} k^{S, \mathcal{Y}\_{\vec{\eta}}}\right) \left[\exp\left(\frac{\mathcal{E}^{S, \mathcal{Y}\_{\vec{\eta}}}}{kT}\right) - 1\right] \\
\times \begin{bmatrix}
\left(1 - \xi\_{3}\right)^{-2} \frac{\partial \xi\_{3}}{\partial \rho} + \frac{3d\_{i}d\_{j}}{d\_{i} + d\_{j}} \left[\left(1 - \xi\_{3}\right)^{-2} \left(\frac{\partial \xi\_{2}}{\partial \rho}\right) + 2\left(1 - \xi\_{3}\right)^{-3} \xi\_{2} \left(\frac{\partial \xi\_{3}}{\partial \rho}\right)\right] \\
+ 2\left(\frac{d\_{i}d\_{j}}{d\_{i} + d\_{j}}\right)^{2} \left[2\xi\_{2} \left(\frac{\partial \xi\_{2}}{\partial \rho}\right) \left(1 - \xi\_{3}\right)^{-3} + 3\left(1 - \xi\_{3}\right)^{-4} \left(\frac{\partial \xi\_{3}}{\partial \rho}\right) \xi\_{2}^{2}\right]
\end{aligned} \tag{97}
$$

As it can be seen from Eq. (96), (����⁄ ) �� *'*s are solutions of a linear system of equations which can be estimated using a known technique such as Gaussian Elimination, Gauss-Jordan or Least Square method (Burden et al., 1981).

#### **2.2 Derivation of fugacity using SAFT EoS**

The fugacity of component *i* in terms of independent variables *V* and *T* is given by the following equation for a given phase (Danesh , 1998; Prausnitz et al., 1999; Tabatabaei-Nejad, & Khodapanah, 2009):

$$RT\ln\rho\_k^{\alpha} = $$

$$= RT\ln\frac{f\_k^{\alpha}}{z\_k^{\alpha}P} = \int\_{V^{\alpha}}^{\alpha} \left[ \left(\frac{\partial P}{\partial n\_k}\right)\_{T,V,n\_{j \neq k}} - \frac{RT}{V} \right] dV - RT\ln Z^{\alpha} \tag{98}$$

$$k = 1,2,...,N\_c \quad \alpha = L\_r V$$

where �� , �� , ��, *V* , *Z* and � are fugacity, fugacity coefficient and the number of moles of component *k* , volume, compressibility factor, and pressure, respectively. The superscript denotes liquid (*L*) and vapor phases (*V*).

The compressibility factor is related to the volume by the following equations:

$$Z^{\alpha} = \frac{P}{\rho^{\alpha}RT} \tag{99}$$

$$
\rho^{\alpha} = \frac{n^{\alpha}}{V^{\alpha}} \tag{100}
$$

where � refers to the total number of moles of the known phase.

14 Advances in Chemical Engineering

parameter ����� characterize, respectively, the association (���) and solvation (���) energy and volume for the specific interaction between sites � and �. These parameters are

a non-linear system of equations which can be solved using any iterative technique such as Gauss-Seidel, Successive-Over-Relaxation (SOR) or Jacobi iterative method. The derivative

1

*j Y*

*j*

2 23 3 3 2 3 3 3 2

*k exp kT*

*<sup>A</sup> <sup>N</sup> <sup>Y</sup>*

*c c*

*N N*

 

*j j*

*Y Y ij*

1

 

(98)

*Y*

*c j j*

*i j*

*S Y*

*<sup>X</sup> z WX*

*j ij*

2 3 4 3 2 2 3 32

*zX W z X*

s satisfy

(96)

(97)

 

*W*

adjustable. Equation (93) requires no mixing rules. As it can be seen in Eq. (94), ��� '

<sup>2</sup> 1 1

ξ ξ <sup>3</sup> <sup>ξ</sup> <sup>1</sup> <sup>ξ</sup> <sup>1</sup> <sup>ξ</sup> 2 1 ξ ξ

*i j*

 

<sup>ξ</sup> <sup>ξ</sup> 2 2<sup>ξ</sup> <sup>1</sup> <sup>ξ</sup> 3 1 ξ ξ

As it can be seen from Eq. (96), (����⁄ ) �� *'*s are solutions of a linear system of equations which can be estimated using a known technique such as Gaussian Elimination, Gauss-

The fugacity of component *i* in terms of independent variables *V* and *T* is given by the following equation for a given phase (Danesh , 1998; Prausnitz et al., 1999; Tabatabaei-

, ,

*<sup>f</sup> P RT RTln dV RTlnZ*

*RTln*

*k V k TVn*

The compressibility factor is related to the volume by the following equations:

*z P n V*

 

1,2, , , *j k*

where �� , �� , ��, *V* , *Z* and � are fugacity, fugacity coefficient and the number of moles of component *k* , volume, compressibility factor, and pressure, respectively. The superscript

 

*k N LV*

*c*

*k*

*i j i j*

*d d d d*

*ij S Y ij*

2

*W*

*i j i j*

*d d d d*

Jordan or Least Square method (Burden et al., 1981).

*k*

**2.2 Derivation of fugacity using SAFT EoS** 

denotes liquid (*L*) and vapor phases (*V*).

Nejad, & Khodapanah, 2009):

*<sup>i</sup> j j <sup>i</sup>*

*j ij <sup>j</sup> <sup>S</sup> j Y j Y <sup>S</sup>*

of the function ��� with respect to � yields the following equation:

*<sup>X</sup> X N*

where,

To use equation (98), we require a suitable EoS that holds for the entire range of possible mole fractions *z* at the system temperature and for the density range between 0 and � � � �. Application of the SAFT EoS in Eq. (98) yields the following equation for calculating the fugacity of the components:

$$RT\text{Tr}\left(Z^{\alpha}\phi\_{k}^{\alpha}\right) = \left[\frac{RT}{V}\frac{RT}{V}\left[\frac{\partial\left(nZ\_{hs}\right)}{\partial n\_{k}}\right]\_{T,V,n\_{j\neq k}}dV + \left[\frac{RT}{V}\frac{T}{V}\left[\frac{\partial\left(nZ\_{ch}\right)}{\partial n\_{k}}\right]\_{T,V,n\_{j\neq k}}dV\right.\tag{101}$$

$$+\frac{\alpha}{V}\frac{RT}{V}\left[\frac{\partial\left(nZ\_{disp}\right)}{\partial n\_{k}}\right]\_{T,V,n\_{j\neq k}}dV + \left[\frac{RT}{V}\left[\frac{\partial\left(nZ\_{asoc}\right)}{\partial n\_{k}}\right]\_{T,V,n\_{j\neq k}}dV\right.\tag{102}$$

$$k = 1,2,...,N\_{c}$$

The following equations are derived for the first term in Eq. (101) accounting for the hard sphere contribution of SAFT:

$$I\Omega\\_HS = \oint\_V \frac{RT}{V} \left[\frac{\partial \left(nZ\_{hs}\right)}{\partial n\_k}\right]\_{T,V,n\_{jrk}}dV = \oint\_0 \frac{RT}{\rho} \left[\frac{\partial \left(nZ\_{hs}\right)}{\partial n\_k}\right]\_{T,V,n\_{jrk}}d\rho\tag{102}$$

$$k = 1,2,...,N\_c$$

$$
\begin{split}
&\left[\frac{\partial\left(nZ\_{\rm hs}\right)}{\partial n\_{k}}\right]\_{T,V,n\_{j\neq k}} = \left[\left(\text{hs}\_{1k}+2\text{hs}\_{1}\right),\rho+\left(\text{hs}\_{2k}+3\text{hs}\_{2}\right),\rho^{2}+\left(\text{hs}\_{3k}+4\text{hs}\_{3}\right),\rho^{3}\right] \\
&\times\left[\text{I}+\text{hs}\_{4},\rho+\text{hs}\_{5},\rho^{2}+\text{hs}\_{6},\rho^{3}\right]^{-1}-\left[\text{I}+\text{hs}\_{4},\rho+\text{hs}\_{5},\rho^{2}+\text{hs}\_{6},\rho^{3}\right]^{-2} \\
&\times\left[\left(\text{hs}\_{4k}+\text{hs}\_{4}\right),\rho+\left(\text{hs}\_{5k}+2\text{hs}\_{5}\right),\rho^{2}+\left(\text{hs}\_{6k}+3\text{hs}\_{6}\right),\rho^{3}\right] \times\left[\text{I}\text{s}\_{1},\rho+\text{hs}\_{2},\rho^{2}+\text{hs}\_{3},\rho^{3}\right]
\end{split}
\tag{103}
$$

The parameters used in Eq. (103) are given by equations (7) – (17) and the following equations:

$$\text{rhs}\_{1k} = \left(\overline{r}\_k \text{S}\_3 + \overline{r} \text{S}\_{3k} + \Im \text{S}\_{1k} \text{S}\_2 + \Im \text{S}\_1 \text{S}\_{2k}\right) P\_{ma} \tag{104}$$

$$\mathbf{h}s\_{2k} = \left(-2\overline{\tau}\_k \mathbf{S}\_3^2 - 4\overline{\tau} \mathbf{S}\_{3k} \mathbf{S}\_3 - 3\mathbf{S}\_{1k} \mathbf{S}\_2 \mathbf{S}\_3 - 3\mathbf{S}\_{2k} \mathbf{S}\_1 \mathbf{S}\_3 - 3\mathbf{S}\_1 \mathbf{S}\_2 \mathbf{S}\_{3k} + 9\mathbf{S}\_{2k} \mathbf{S}\_2^2\right) \mathbf{P}\_{\mathbf{n}\mathbf{n}}^2 \tag{105}$$

$$h\mathbf{s}\_{3k} = \{\vec{r}\_k\mathbf{S}\_3^3 + 3\vec{r}\mathbf{S}\_3^2\mathbf{S}\_{3k} - 3\mathbf{S}\_2^2\mathbf{S}\_{2k}\mathbf{S}\_3 - \mathbf{S}\_2^3\mathbf{S}\_{3k}\}P\_{na}^3\tag{106}$$

$$
\hbar h \mathbf{s}\_{4k} = -3 \mathbf{S}\_{3k} P\_{na} \tag{107}
$$

$$h\mathbf{s}\_{5k} = 6\mathbf{S}\_3\mathbf{S}\_{3k}P\_{na}^{\mathbf{s}}\tag{108}$$

$$h\mathbf{S}\_{6k} = -3\mathbf{S}\_{\overline{3}}^2 \mathbf{S}\_{3k} P\_{na}^3 \tag{109}$$

$$
\vec{r}\_k = (-\vec{r} + r\_k) \tag{110}
$$

$$S\_{1k} = -S\_1 + r\_k d\_k \tag{111}$$

$$\mathcal{S}\_{2k} = -\mathcal{S}\_2 + \eta\_k d\_k^2 \tag{112}$$

$$\begin{aligned} S\_{3k} &= -S\_3 + r\_k d\_k^3\\ k &= \mathbf{1}, \mathbf{2}, \dots, \mathbf{N}\_\circ \end{aligned} \tag{113}$$

$$IR\\_Chani = \int\_V^{\circ} \frac{RT}{V} \left[\frac{\partial \langle nZ\_{ch}\rangle}{\partial n\_k}\right]\_{T,V,n\_{f\bullet k}} dV = \int\_0^{\rho} \frac{RT}{\rho} \left[\frac{\partial \langle nZ\_{ch}\rangle}{\partial n\_k}\right]\_{T,V,n\_{f\bullet k}} d\rho \tag{114}$$

$$k = 1.2 \qquad N$$

$$ch\_{1k}(d\_l) = (2S\_{3k} + 3d\_lS\_{2k})P\_{na} \tag{116}$$

$$ch\_{2k}(d\_l) = \left(-8\mathcal{S}\_3\mathcal{S}\_{3k} + 4d\_l^2\mathcal{S}\_2\mathcal{S}\_{2k}\right)P\_{na}^2\tag{117}$$

$$ch\_{3k}(d\_l) = \left(2d\_l^2 \mathcal{S}\_2 \mathcal{S}\_{2k} \mathcal{S}\_3 + d\_l^2 \mathcal{S}\_2^2 \mathcal{S}\_{3k} + 6\mathcal{S}\_3^2 \mathcal{S}\_{3k} - 3d\_l \mathcal{S}\_{2k} \mathcal{S}\_3^2 - 6d\_l \mathcal{S}\_2 \mathcal{S}\_3 \mathcal{S}\_{3k}\right) P\_{na}^3 \tag{118}$$

$$ch\_{4k}(d\_l) = (-6S\_{3k} + 3d\_lS\_{2k})P\_{na} \tag{119}$$

$$ch\_{5k}(d\_l) = \left(12S\_3S\_{3k} + 2d\_l^2S\_2S\_{2k} - 9d\_lS\_{2k}S\_3 - 9d\_lS\_2S\_{3k}\right)P\_{na}^2\tag{120}$$

$$ch\_{6k}(d\_l) = \left\{ 9d\_l \mathcal{S}\_{2k} \mathcal{S}\_3^2 + 18d\_l \mathcal{S}\_2 \mathcal{S}\_3 \mathcal{S}\_{3k} - 6\mathcal{S}\_3^2 \mathcal{S}\_{3k} - 2d\_l^2 \mathcal{S}\_2 \mathcal{S}\_{2k} \mathcal{S}\_3 - d\_l^2 \mathcal{S}\_2^2 \mathcal{S}\_{3k} \right\} P\_{na}^3 \tag{121}$$

$$\begin{aligned} \text{ch}\_{\mathsf{T}k} \left( d\_i \right) &= \left( -\Im d\_i S\_{2k} S\_3^3 - \Re d\_i S\_2 S\_3^2 S\_{3k} \right) P\_{na}^4\\ k &= \mathbf{1}, \mathbf{2}, \dots, \mathbf{N}\_c \end{aligned} \tag{122}$$

$$IR\\_Disp = \int\_V^\alpha \frac{RT}{V} \left[\frac{\partial \left(nZ\_{d\text{disp}}\right)}{\partial n\_k}\right]\_{T,V,n\_{f\text{ast}}}dV = \int\_0^\rho \frac{RT}{\rho} \left[\frac{\partial \left(nZ\_{d\text{disp}}\right)}{\partial n\_k}\right]\_{T,V,n\_{f\text{ast}}}d\rho\tag{123}$$

$$k = 1,2,...,N\_c$$

$$
\left[\frac{\partial \left(n Z\_{\rm disp}\right)}{\partial n\_k}\right]\_{T, V, n\_{j \neq k}} = Z\_{\rm disp} + \left(Z\_{\rm disp}\right)\_k\tag{124}
$$

$$
\left(Z\_{\rm disp}\right)\_k = A\_{1k} + A\_{2k} A\_3 + A\_2 A\_{3k}\,,\quad k = 1, 2, \dots, N\_c
$$

$$A\_1 = d\mathcal{I}\_1.\rho + d\mathcal{I}\_2.\rho^2 + d\mathcal{I}\_3.\rho^3 + d\mathcal{I}\_4.\rho^4 + d\mathcal{I}\_5.\rho^5 + d\mathcal{I}\_6.\rho^6 + d\mathcal{I}\_7.\rho^7\tag{125}$$

$$\begin{aligned} A\_{1k} &= \left( d\mathbf{I}\_{1k} + d\mathbf{I}\_1 \right) \rho + \left( d\mathbf{I}\_{2k} + 2d\mathbf{I}\_2 \right) \rho^2 + \left( d\mathbf{I}\_{3k} + 3d\mathbf{I}\_3 \right) \rho^3 \\ &+ \left( d\mathbf{I}\_{4k} + 4d\mathbf{I}\_4 \right) \rho^4 + \left( d\mathbf{I}\_{5k} + 5d\mathbf{I}\_5 \right) \rho^5 + \left( d\mathbf{I}\_{6k} + 6d\mathbf{I}\_6 \right) \rho^6 + \left( d\mathbf{I}\_{7k} + 7d\mathbf{I}\_7 \right) \rho^7 \\ &\quad k = \mathbf{1}\_r \mathbf{2}, \dots, \mathbf{N}\_c \end{aligned} \tag{126}$$

$$A\_2 = -\Theta P\_{na} \overline{r} A \text{disp}\_2 \tag{127}$$

$$\begin{aligned} A\_{2k} &= -\mathsf{6}P\_{\mathsf{nat}} \Big[ \overline{r}\_k .A \mathrm{disp}\_2 + \overline{r} .A \mathrm{disp}\_{2k} \Big] \\ k &= 1, 2, \dots, N\_c \end{aligned} \tag{128}$$

$$A\_3 = \left(\frac{B\_1}{B\_2}\right) \cdot B\_3 - \left(\frac{B\_1}{\rho B\_2}\right)^2 \cdot \left(\frac{B\_4}{B\_5}\right) \cdot B\_6 \cdot B\_7 \tag{129}$$

$$B\_1 = 4\,\rho + \operatorname{disp}\_1 \cdot \rho^2 + \operatorname{disp}\_2 \cdot \rho^3 + \operatorname{disp}\_3 \cdot \rho^4 + \operatorname{disp}\_4 \cdot \rho^5 + \operatorname{disp}\_5 \cdot \rho^6 + \operatorname{disp}\_6 \cdot \rho^7 \tag{130}$$

$$\begin{aligned} B\_{1k} &= 4\,\rho + \left(d\_{1k} + 2\,\mathrm{disp}\_{1}\right) \cdot \rho^{\,^2} + \left(d\_{2k} + 3\,\mathrm{disp}\_{2}\right) \cdot \rho^{\,^3} + \left(d\_{3k} + 4\,\mathrm{disp}\_{3}\right) \cdot \rho^{\,^4} \\ &+ \left(d\_{4k} + 5\,\mathrm{disp}\_{4}\right) \cdot \rho^{\,^5} + \left(d\_{5k} + 6\,\mathrm{disp}\_{5}\right) \cdot \rho^{\,^6} + \left(d\_{6k} + 7\,\mathrm{disp}\_{6}\right) \cdot \rho^{\,^7} \\ &\qquad k = 1, 2, \dots, N\_c \end{aligned} \tag{131}$$

$$B\_2 = 4 + \operatorname{disp}\_7 \cdot \rho + \operatorname{disp}\_8 \cdot \rho^2 + \operatorname{disp}\_9 \cdot \rho^3 + \operatorname{disp}\_{10} \cdot \rho^4 + \operatorname{disp}\_{11} \cdot \rho^5 + \operatorname{disp}\_{12} \cdot \rho^6 \tag{132}$$

$$\begin{aligned} B\_{2k} &= \left( d\_{7k} + \text{disp}\_7 \right) \cdot \rho + \left( d\_{8k} + 2 \text{disp}\_8 \right) \cdot \rho^2 + \left( d\_{9k} + 3 \text{disp}\_9 \right) \cdot \rho^3 \\ &+ \left( d\_{10k} + 4 \text{disp}\_{10} \right) \cdot \rho^4 + \left( d\_{11k} + 5 \text{disp}\_{11} \right) \cdot \rho^5 + \left( d\_{12k} + 6 \text{disp}\_{12} \right) \cdot \rho^6 \end{aligned} \tag{133}$$

$$B\_3 = b\_0 + d\mathbf{I}\_8 \cdot \boldsymbol{\rho} + d\mathbf{I}\_9 \cdot \boldsymbol{\rho}^2 + d\mathbf{I}\_{10} \cdot \boldsymbol{\rho}^3 + d\mathbf{I}\_{11} \cdot \boldsymbol{\rho}^4 + d\mathbf{I}\_{12} \cdot \boldsymbol{\rho}^5 + d\mathbf{I}\_{13} \cdot \boldsymbol{\rho}^6 \tag{134}$$

$$\begin{aligned} B\_{3k} &= b\_{0k} + \left( d\mathbf{I}\_{8k} + d\mathbf{I}\_{8} \right) \cdot \boldsymbol{\rho} + \left( d\mathbf{I}\_{9k} + 2d\mathbf{I}\_{9} \right) \cdot \boldsymbol{\rho}^{2} + \left( d\mathbf{I}\_{10k} + 3d\mathbf{I}\_{10} \right) \cdot \boldsymbol{\rho}^{3} \\ &+ \left( d\mathbf{I}\_{11k} + 4d\mathbf{I}\_{11} \right) \cdot \boldsymbol{\rho}^{4} + \left( d\mathbf{I}\_{12k} + 5d\mathbf{I}\_{12} \right) \cdot \boldsymbol{\rho}^{5} + \left( d\mathbf{I}\_{13k} + 6d\mathbf{I}\_{13} \right) \cdot \boldsymbol{\rho}^{6} \end{aligned} \tag{135}$$

$$B\_4 = \text{disp}\_{13} + \text{disp}\_{14} \cdot \rho + \text{disp}\_{15} \cdot \rho^2 + \text{disp}\_{16} \cdot \rho^3 + \text{disp}\_{17} \cdot \rho^4 + \text{disp}\_{18} \cdot \rho^5 \tag{136}$$

$$\begin{aligned} B\_{4k} &= d\_{13k} + \left( d\_{14k} + \operatorname{disp}\_{14} \right) \cdot \rho + \left( d\_{15k} + 2 \operatorname{disp}\_{15} \right) \cdot \rho^2 \\ &+ \left( d\_{16k} + 3 \operatorname{disp}\_{16} \right) \cdot \rho^3 + \left( d\_{17k} + 4 \operatorname{disp}\_{17} \right) \cdot \rho^4 + \left( d\_{18k} + 5 \operatorname{disp}\_{18} \right) \cdot \rho^5 \end{aligned} \tag{137}$$

$$\begin{aligned} B\_5 &= 8 + \operatorname{disp}\_{19} \cdot \rho + \operatorname{disp}\_{20} \cdot \rho^2 + \operatorname{disp}\_{21} \cdot \rho^3 + \operatorname{disp}\_{22} \cdot \rho^4 + \operatorname{disp}\_{23} \cdot \rho^5 \\ &+ \operatorname{disp}\_{24} \cdot \rho^6 + \operatorname{disp}\_{25} \cdot \rho^7 + \operatorname{disp}\_{26} \cdot \rho^8 \end{aligned} \tag{138}$$

$$\begin{aligned} B\_{5k} &= \left( d\_{19k} + \operatorname{disp}\_{19} \right) \cdot \rho + \left( d\_{20k} + 2 \operatorname{disp}\_{20} \right) \cdot \rho^2 \\ &+ \left( d\_{21k} + 3 \operatorname{disp}\_{21} \right) \cdot \rho^3 + \left( d\_{22k} + 4 \operatorname{disp}\_{22} \right) \cdot \rho^4 + \left( d\_{23k} + 5 \operatorname{disp}\_{23} \right) \cdot \rho^5 \\ &+ \left( d\_{24k} + 6 \operatorname{disp}\_{24} \right) \cdot \rho^6 + \left( d\_{25k} + 7 \operatorname{disp}\_{25} \right) \cdot \rho^7 + \left( d\_{26k} + 8 \operatorname{disp}\_{26} \right) \cdot \rho^8 \end{aligned} \tag{139}$$

$$B\_6 = b\_0 \cdot \rho^2 + \text{Idisp}\_1 \cdot \rho^3 + \text{Idisp}\_2 \cdot \rho^4 + \text{Idisp}\_3 \cdot \rho^5 + \text{Idisp}\_4 \cdot \rho^6 + \text{Idisp}\_5 \cdot \rho^7 + \text{Idisp}\_6 \cdot \rho^8 \tag{140}$$

$$\begin{aligned} B\_{6k} &= \left(b\_{0k} + 2b\_0\right) \cdot \rho^2 + \left(I \text{disp}\_{1k} + 3I \text{disp}\_1\right) \cdot \rho^3 \\ &+ \left(I \text{disp}\_{2k} + 4I \text{disp}\_2\right) \cdot \rho^4 + \left(I \text{disp}\_{3k} + 5I \text{disp}\_3\right) \cdot \rho^5 \\ &+ \left(I \text{disp}\_{4k} + 6I \text{disp}\_4\right) \cdot \rho^6 + \left(I \text{disp}\_{5k} + 7I \text{disp}\_5\right) \cdot \rho^7 + \left(I \text{disp}\_{6k} + 8I \text{disp}\_6\right) \cdot \rho^8 \end{aligned} \tag{141}$$

$$B\_7 = P\_{\rm nu} S\_3 \tag{142}$$

$$\begin{aligned} B\_{7k} &= P\_{na} S\_{3k} \\ k &= \mathbf{1}\_r \mathbf{2}\_r \dots \mathbf{N}\_r \end{aligned} \tag{143}$$

$$d\_{1k} = 12[\vec{r}\_k \mathbb{S}\_3 + \vec{r} \mathbb{S}\_{3k}] P\_{na} \tag{144}$$

$$d\_{2k} = [27\overline{r}\_k\mathbb{S}\_3^2 + 2\langle 27\overline{r} - 26\rangle \mathbb{S}\_3 \mathbb{S}\_{3k}] P\_{na}^2 \tag{145}$$

$$d\_{3k} = [70\vec{r}\_k \mathcal{S}\_3^3 + 3\mathcal{S}\_3^2 \mathcal{S}\_{3k} (70\vec{r} + 4\mathcal{Q})] P\_{na}^3 \tag{146}$$

$$d\_{4k} = [\mathbf{51}\vec{r}\_k\mathbf{S}\_3^4 + \mathbf{3S}\_3^3\mathbf{S}\_{3k}(\mathbf{51}\vec{r} - \mathbf{27})]P\_{na}^4\tag{147}$$

$$d\_{5k} = \left[ -16\vec{r}\_k \mathbb{S}\_3^5 + 5 \mathbb{S}\_3^4 \mathbb{S}\_{3k} \left( 8 - 16\vec{r} \right) \right] P\_{na}^5 \tag{148}$$

$$d\_{6k} = \left[2\overline{r}\_k \mathbb{S}\_3^6 + 6\mathbb{S}\_3^5 \mathbb{S}\_{3k} (2\overline{r} - 1)\right] P\_{na}^6 \tag{149}$$

$$d\_{7k} = -20 \mathcal{S}\_{3k} P\_{na} \tag{150}$$

$$d\_{\otimes k} = \mathsf{BZ}\_3 \mathsf{S}\_{3k} P\_{na}^2 \tag{151}$$

$$d\_{9k} = -132S\_3^2 S\_{3k} P\_{na}^3 \tag{152}$$

$$d\_{10k} = 104 \text{S}\_3^3 \text{S}\_{3k} P\_{na}^4 \tag{153}$$

$$d\_{11k} = -40S\_3^4 S\_{3k} P\_{na}^\natural \tag{154}$$

$$d\_{12k} = \text{6S}\_3^5 \text{S}\_{3k} P\_{na}^6 \tag{155}$$

$$d\_{13k} = 24\vec{r}\_k\tag{156}$$

$$d\_{14k} = [192\overline{r}\_k\mathbb{S}\_3 + (192\overline{r} - 12\mathbb{S})\mathbb{S}\_{3k}]P\_{na} \tag{157}$$

$$d\_{15k} = [-372\overline{r}\_k \mathcal{S}\_3^2 + 2\mathcal{S}\_3 \mathcal{S}\_{3k} (148 - 372\overline{r})] P\_{na}^2 \tag{158}$$

$$d\_{16k} = [230\overline{r}\_k \mathbb{S}\_3^3 + 3\mathbb{S}\_3^2 \mathbb{S}\_{3k} (230\overline{r} - 70)] P\_{na}^3 \tag{159}$$

$$d\_{17k} = [-52\vec{r}\_k \mathcal{S}\_3^4 + 4\mathcal{S}\_3^3 \mathcal{S}\_{3k} (8 - 52\vec{r})] P\_{na}^4 \tag{160}$$

$$d\_{18k} = \left[2\vec{r}\_k \mathbb{S}\_3^5 + 10 \mathbb{S}\_3^4 \mathbb{S}\_{3k} (1+\vec{r})\right] P\_{na}^5 \tag{161}$$

$$d\_{19k} = -52S\_{3k}P\_{na} \tag{162}$$

$$d\_{20k} = 292S\_3S\_{3k}P\_{na}^2\tag{163}$$

$$d\_{21k} = -693 S\_3^2 S\_{3k} P\_{na}^3 \tag{164}$$

$$d\_{22k} = 900 \mathcal{S}\_3^3 \mathcal{S}\_{3k} P\_{na}^4 \tag{165}$$

$$d\_{23k} = -690 S\_3^4 S\_{3k} P\_{na}^5 \tag{166}$$

$$d\_{24k} = \mathbf{3} \mathbf{1} \mathbf{2} \mathbf{S}\_3^5 \mathbf{S}\_{3k} P\_{na}^6 \tag{167}$$

$$d\_{25k} = -77 \mathcal{S}\_3^6 \mathcal{S}\_{3k} P\_{na}^7 \tag{168}$$

$$\begin{aligned} \mathcal{U}\_{26k} &= 8 S\_3^7 S\_{3k} P\_{na}^8\\ k &= 1, 2, \dots, N\_c \end{aligned} \tag{169}$$

$$AdSp\_{1k} = 2 \times \left[ -\sum\_{l=1}^{N\_c} \sum\_{l=1}^{N\_c} \mathbf{z}\_l \mathbf{z}\_l r\_l r\_l \{\mathbf{e}\_{lj}/kT\} \sigma\_{lj}^3 + \sum\_{l=1}^{N\_c} \mathbf{z}\_l r\_l r\_k \{\mathbf{e}\_{lk}/kT\} \sigma\_{lk}^3 \right] \tag{170}$$

$$AdS\_{2k} = 2 \times \left[ -\sum\_{l=1}^{N\_c} \sum\_{f=1}^{N\_c} z\_l z\_f r\_l r\_l \{\varepsilon\_{lj}/kT\}^2 \sigma\_{lj}^3 + \sum\_{l=1}^{N\_c} z\_l r\_l r\_k \{\varepsilon\_{lk}/kT\}^2 \sigma\_{lk}^3 \right] \tag{171}$$

$$a\_{jk}(\vec{r}) = \begin{bmatrix} \left[ \frac{\vec{r}\_k}{\vec{r}} - \frac{\vec{r}\_k \cdot (\vec{r} - 1)}{\vec{r}^2} \right] a\_{1j} \\ + \left[ \frac{\vec{r}\_k \cdot (\vec{r} - 2)}{\vec{r}^2} + \frac{\vec{r}\_k \cdot (\vec{r} - 1)}{\vec{r}^2} - 2 \frac{\vec{r}\_k \cdot (\vec{r} - 1) \cdot (\vec{r} - 2)}{\vec{r}^3} \right] a\_{2j} \end{bmatrix} \tag{172}$$

$$b\_{jk}(\vec{r}) = \begin{bmatrix} \frac{\vec{r}\_k}{\vec{r}} - \frac{\vec{r}\_k \cdot (\vec{r} - 1)}{\vec{r}^2} \Big| b\_{1j} \\\\ + \left[ \frac{\vec{r}\_k \cdot (\vec{r} - 2)}{\vec{r}^2} + \frac{\vec{r}\_k \cdot (\vec{r} - 1)}{\vec{r}^2} - 2 \frac{\vec{r}\_k \cdot (\vec{r} - 1) \cdot (\vec{r} - 2)}{\vec{r}^3} \right] b\_{2j} \end{bmatrix} \tag{173}$$

$$j = 0, 1, \dots, 6 \qquad k = 1, 2, \dots, N\_c$$

$$Idisp\_{1k} = [b\_{1k} \mathbb{S}\_3 + b\_1 \mathbb{S}\_{3k}] P\_{na} \tag{174}$$

$$M \text{disp}\_{2k} = [b\_{2k} \mathbb{S}\_3^2 + 2b\_2 \mathbb{S}\_3 \mathbb{S}\_{3k}] P\_{na}^2 \tag{175}$$

$$Idisp\_{3k} = [b\_{3k} \mathcal{S}\_3^3 + 3b\_3 \mathcal{S}\_3^2 \mathcal{S}\_{3k}] P\_{na}^3 \tag{176}$$

$$Idisp\_{4k} = [b\_{4k} \mathbb{S}\_3^4 + 4b\_4 \mathbb{S}\_3^3 \mathbb{S}\_{3k}] P\_{na}^4 \tag{177}$$

$$Idisp\_{5k} = \left[b\_{5k} \mathcal{S}\_3^5 + 5b\_5 \mathcal{S}\_3^4 \mathcal{S}\_{3k}\right] P\_{na}^5 \tag{178}$$

$$Idisp\_{6k} = \left[b\_{6k} \mathcal{S}\_3^6 + 6b\_6 \mathcal{S}\_3^5 \mathcal{S}\_{3k}\right] P\_{na}^6 \tag{179}$$

$$dId \text{is} \! p\_{7k} = 2[a\_{1k}\mathbb{S}\_3 + a\_1\mathbb{S}\_{3k}]P\_{na} \tag{180}$$

$$dIdisp\_{2k} = 3[a\_{2k} \mathcal{S}\_3^2 + 2a\_2 \mathcal{S}\_3 \mathcal{S}\_{3k}]P\_{na}^2 \tag{181}$$

$$dIdisp\_{3k} = 4[a\_{3k} \mathbb{S}\_3^3 + 3a\_3 \mathbb{S}\_3^2 \mathbb{S}\_{3k}]P\_{na}^3 \tag{182}$$

$$dIdisp\_{4k} = 5[a\_{4k} \mathcal{S}\_3^4 + 4a\_4 \mathcal{S}\_3^3 \mathcal{S}\_{3k}]P\_{na}^4 \tag{183}$$

$$Mdisp\_{5k} = 6 \left[ a\_{5k} \mathcal{S}\_3^5 + 5 a\_5 \mathcal{S}\_3^4 \mathcal{S}\_{3k} \right] P\_{na}^5 \tag{184}$$

$$dId \text{is} \! p\_{6k} = 7 \left[ a\_{6k} \text{S}\_3^6 + 6a\_6 \text{S}\_3^5 \text{S}\_{3k} \right] P\_{na}^6 \tag{185}$$

$$dI\_{1k} = -12[AdSp\_{1k} \cdot a\_0 + AdSsp\_1 \cdot a\_{0k}]P\_{na} \tag{186}$$

$$dI\_{2k} = -12[AdSsp\_{1k} \cdot dIdisp\_1 + AdSsp\_1 \cdot dIdisp\_{1k}]P\_{na} \tag{187}$$

$$dI\_{3k} = -12[AdSsp\_{1k} \cdot dIdisp\_2 + AdSsp\_1 \cdot dIdisp\_{2k}]P\_{na} \tag{188}$$

$$dI\_{4k} = -12[AdSsp\_{1k} \cdot dIdisp\_3 + AdSsp\_1 \cdot dIdisp\_{3k}]P\_{na} \tag{189}$$

$$dI\_{5k} = -12[AdSip\_{1k} \cdot dIdisp\_4 + AdSip\_1 \cdot dIdisp\_{4k}]P\_{na} \tag{190}$$

$$dI\_{6k} = -12[AdSsp\_{1k} \cdot dIdisp\_5 + AdSsp\_1 \cdot dIdisp\_{5k}]P\_{na} \tag{191}$$

$$\text{dI}\_{7k} = -12[\text{Addisp}\_{1k} \cdot \text{dI} \text{disp}\_6 + \text{Adisp}\_1 \cdot \text{dI} \text{disp}\_{6k}]P\_{na} \tag{192}$$

$$dI\_{8k} = \mathcal{Z}[b\_{1k}\mathcal{S}\_3 + b\_1\mathcal{S}\_{3k}]P\_{na} \tag{193}$$

$$dI\_{9k} = \Im[b\_{2k}\mathbb{S}\_3^2 + 2b\_2\mathbb{S}\_3\mathbb{S}\_{3k}]P\_{na}^2\tag{194}$$

$$dI\_{10k} = 4[b\_{3k} \mathcal{S}\_3^3 + 3b\_3 \mathcal{S}\_3^2 \mathcal{S}\_{3k}] P\_{na}^3 \tag{195}$$

$$dI\_{11k} = \mathbb{S}[b\_{4k}\mathbb{S}\_3^4 + 4b\_4\mathbb{S}\_3^3\mathbb{S}\_{3k}]P\_{na}^4\tag{196}$$

$$dI\_{12k} = 6 \left[ b\_{5k} \mathcal{S}\_3^5 + 5 b\_5 \mathcal{S}\_3^4 \mathcal{S}\_{3k} \right] P\_{na}^5 \tag{197}$$

$$\mathbb{T}\,dI\_{13k} = \mathbb{T}\Big[b\_{6k}\mathbb{S}\_3^6 + 6b\_6\mathbb{S}\_3^5\mathbb{S}\_{3k}\Big]P\_{na}^6\tag{198}$$

$$IR\\_Assoc\, = \int\_{V}^{v} \frac{RT}{V} \left[\frac{\partial(nZ\_{assoc})}{\partial n\_k}\right]\_{T,V,n\_{j\neq k}} dV = \int\_{0}^{\rho} \frac{RT}{\rho} \left[\frac{\partial(nZ\_{assoc})}{\partial n\_k}\right]\_{T,V,n\_{j\neq k}} d\rho\tag{199}$$

$$k = 1,2,...,N\_c$$

$$
\left[\frac{\partial (nZ\_{assoc})}{\partial n\_k}\right]\_{T,V,n\_{f\neq k}} = Z\_{assoc} + \rho \langle A\_k \rangle + \rho^2 \sum\_{l=1}^{N\_c} z\_l \left(\frac{\partial A\_l}{\partial n\_k}\right)\_{T,V,n\_{f\neq k}} \tag{200}
$$

$$k = 1, 2, \dots, N\_c$$

$$\left(\frac{\partial A\_l}{\partial n\_k}\right)\_{T,V,n\_{j\neq k}} = \sum\_{S\_l} -\frac{1}{\{X^{S\_l}\}^2} \left(\frac{\partial X^{S\_l}}{\partial n\_k}\right)\_{T,V,n\_{j\neq k}} \cdot \left(\frac{\partial X^{S\_l}}{\partial \rho}\right) + \left(\frac{1}{X^{S\_l}} - \frac{1}{2}\right) \left[\frac{\partial}{\partial n\_k} \left(\frac{\partial X^{S\_l}}{\partial \rho}\right)\right]\_{T,V,n\_{j\neq k}} \quad (201)$$
 
$$k = 1, 2, \dots, N\_c$$

$$A\_l = \sum\_{S\_l} \left(\frac{1}{X^{S\_l}} - \frac{1}{2}\right) \left(\frac{\partial X^{S\_l}}{\partial \rho}\right) \tag{202}$$

$$\mathbb{E}\left(\frac{\partial X^{\mathbb{S}\_{l}}}{\partial n\_{k}}\right)\_{\mathbf{T},\mathbf{V},n\_{j}\neq\mathbf{k}} = -(X^{\mathbb{S}\_{l}})^{2}N\_{A}\left[\sum\_{j=1}^{N\_{c}}\sum\_{\mathbf{Y}\_{j}}\frac{1}{N}\mathcal{S}\_{lk}X^{\mathbf{Y}\_{j}}W\_{lj} + \sum\_{j=1}^{N\_{c}}\sum\_{\mathbf{Y}\_{j}}z\_{j}\rho X^{\mathbf{Y}\_{j}}\left(\frac{\partial W\_{lj}}{\partial n\_{k}}\right)\_{\mathbf{T},\mathbf{V},n\_{j}\neq\mathbf{k}} + \right] \tag{203}$$

$$k = 1, 2, \dots, N\_c$$

$$
\begin{split}
\left(\frac{\partial \mathcal{W}\_{\boldsymbol{\hat{y}}}}{\partial \boldsymbol{n}\_{k}}\right)\_{\boldsymbol{T},\boldsymbol{V},\boldsymbol{n}\_{j\neq k}} &= \left(\sigma\_{\boldsymbol{\hat{y}}}k^{S\_{\boldsymbol{Y}}\boldsymbol{Y}\_{j}}\right) \left[\exp\left(\frac{\boldsymbol{\mathcal{E}}\_{\boldsymbol{S}}\boldsymbol{Y}\_{j}}{kT}\right) - 1\right] \\
\times \begin{bmatrix}
\left(\boldsymbol{1}-\boldsymbol{\xi}\_{3}\right)^{-2}\frac{\partial \boldsymbol{\xi}\_{3}}{\partial \boldsymbol{n}\_{k}} + 3\frac{d\_{i}d\_{j}}{d\_{i}+d\_{j}} \left[\left(\boldsymbol{1}-\boldsymbol{\xi}\_{3}\right)^{-2}\left(\frac{\partial \boldsymbol{\xi}\_{2}}{\partial \boldsymbol{n}\_{k}}\right) + 2\left(\boldsymbol{1}-\boldsymbol{\xi}\_{3}\right)^{-3}\boldsymbol{\xi}\_{2}\left(\frac{\partial \boldsymbol{\xi}\_{3}}{\partial \boldsymbol{n}\_{k}}\right)\right] \\
\times \left[\begin{matrix}
\boldsymbol{d}\_{i}\boldsymbol{d}\_{j} \\
\boldsymbol{d}\_{i}+\boldsymbol{d}\_{j}
\end{matrix}\right]^{2} \left[\left(\boldsymbol{2}\left(\boldsymbol{1}-\boldsymbol{\xi}\_{3}\right)^{-3}\boldsymbol{\xi}\_{2}\left(\frac{\partial \boldsymbol{\xi}\_{2}}{\partial \boldsymbol{n}\_{k}}\right) + 3\left(\boldsymbol{1}-\boldsymbol{\xi}\_{3}\right)^{-4}\left(\frac{\partial \boldsymbol{\xi}\_{3}}{\partial \boldsymbol{n}\_{k}}\right)\boldsymbol{\xi}\_{2}^{2}\right] \\
\times \left[\begin{matrix}
\boldsymbol{N}\_{k}\boldsymbol{\xi}\_{1} \\
\boldsymbol{n}\_{k}
\end{matrix}\right]^{2} \left[\left(\boldsymbol{\xi}\_{3}\right)^{-2}\boldsymbol{\xi}\_{$$

$$\mathbf{k} = \mathbf{1}, \mathbf{2}, \dots, N\_c$$

$$\frac{\partial \xi\_2}{\partial \rho} = P\_{na} S\_2 \tag{207}$$

$$\frac{\partial \xi\_3}{\partial \rho} = P\_{na} \xi\_3 \tag{208}$$

$$\left(\frac{\partial \xi\_2}{\partial n\_k}\right)\_{T,V,n\_{f\neq k}} = \frac{1}{V} P\_{na} r\_k d\_k^2 \tag{209}$$

$$
\left(\frac{\partial \xi\_3}{\partial n\_k}\right)\_{T, V, n\_{j \neq k}} = \frac{1}{V} P\_{na} r\_k d\_k^3 \tag{210}
$$

$$\left(\frac{\partial}{\partial n\_k} \left(\frac{\partial \xi\_2}{\partial \rho}\right)\right)\_{T, V, n\_{f \neq k}} = P\_{na} \cdot \sum\_{l=1}^{N\_c} \left(\frac{\delta\_{lk} - z\_l}{n}\right) r\_l d\_l^2 = \frac{1}{V} \cdot P\_{na} \cdot \frac{1}{\rho} [r\_k d\_k^2 - S\_2] \tag{211}$$

Application of Chebyshev Polynomials to Calculate Density and

different temperatures.

Fugacity Using SAFT Equation of State to Predict Asphaltene Precipitation Conditions 25

Figure 2 shows the interpolation error using Chebyshev polynomials of degree 15 for approximating pressure vs. density of a binary mixture of ethanol and toluene containing 37.5 mole% ethanol. Figure 3 shows the error in interpolation for another system (oil

Fig. 2. Interpolation error using Chebyshev polynomials for approximating preesure vs. density of a binary mixture of ethanol and toluene containing 37.5 mole% of ethanol at

Fig. 3. Interpolation error using Chebyshev polynomials for approximating preesure vs. density at different temperatures for an oil sample of the composition given in Table 2.

After approximating the �(�) function using Chebyshev polynomials, it is necessary to find solutions for density values at the given pressure(s) and select those which are physically interpretable. In doing so, the complex and negative solutions and those which make �� �� ⁄

sample) for which the composition is given in Table 2 (Jamaluddin et al., 2000).

$$
\left(\frac{\partial}{\partial n\_k} \left(\frac{\partial \boldsymbol{\xi}\_3}{\partial \rho}\right)\right)\_{\mathcal{T}, \boldsymbol{V}, n\_{j \star k}} = P\_{\boldsymbol{n} \boldsymbol{a}} \cdot \sum\_{l=1}^{N\_c} \left(\frac{\delta\_{lk} - \mathbf{z}\_l}{n}\right) \boldsymbol{r}\_l \boldsymbol{d}\_l^3 = \frac{1}{\boldsymbol{V}} \cdot P\_{\boldsymbol{n} \boldsymbol{a}} \cdot \frac{1}{\rho} \left[\boldsymbol{r}\_k \boldsymbol{d}\_k^3 - \mathbf{S}\_3\right] \tag{212}
$$
 
$$k\boldsymbol{k} = \mathbf{1}, \boldsymbol{2}, \dots, N\_{\boldsymbol{c}\_c}$$

In the above equations ,���� refers to dirac delta function which is defined as following:

$$
\delta\_{jk}\delta\_{lk} = \begin{cases} 0 & j \neq k \\ 1 & j = k \end{cases} \tag{213}
$$

#### **3. Application of Chebyshev polynomials to calculate density**

The integration of the terms used in equations (102), (114), (123) and (119) for calculating the fugacity coefficients are performed numerically using Gaussian quadrature method. We found that five point quadrature method leads to a result with acceptable accuracy. As it can be seen from Eq. (101) the fugacity coefficient is a function of temperature, pressure, composition and the properties of the components. In order to calculate the fugacity coefficient of each component, we should first calculate the density of mixture at a given pressure, temperature and composition using Eq. (2). As it can be seen, from the mentioned equation, the density as function of the pressure is not known explicitly. Therefore, the estimation of the density at a given pressure should be performed using an iterative procedure, starting from initial guesses because of the multiplicity of the solution. A solution which is obtained by an iterative technique depends on the choice of the initial guess. Therefore, iterative procedures can not cover all acceptable roots unless the number of roots and the approximate values of the solutions (i.e. initial guesses) had already been known. Hence, an alternative, robust, fast and accurate technique that can predict all acceptable solutions is proposed. The proposed method is based on a numerical interpolation using Chebyshev polynomials in a finite interval (Burden et al., 1981).

It should be pointed out that Chebyshev series provide high accuracy and can be transformed to power series which are suitable for root finding procedure. More general accounts of root finding through Chebyshev approximations are given in (Boyd, 2006). The aforementioned method enables us to calculate all possible solutions and select among them those which are physically interpretable.

It should be considered that using Chebyshev polynomials to approximate a given function will become more efficient when it has non-zero values at both end points of the interval. It can be shown that the pressure vs. density function in SAFT EoS linearly goes to zero for negligible values of the density. In order to avoid this problem, � �⁄ vs. density using Chebyshev polynomials was interpolated.

Another advantage of using Chebyshev polynomials for approximating a function is that for a specific number of basic functions, it always leads to a well-conditioned matrix during the calculation of the unknown coefficients of the basis functions, which is more accurate than the other interpolation techniques.

� ������

�

= 1 � . ���. 1 � �����

1 � �=� (213)

� � ���

(212)

= ���.����� � ��

�� = 1� �� � � ���

In the above equations ,���� refers to dirac delta function which is defined as following:

������ = �� � ���

The integration of the terms used in equations (102), (114), (123) and (119) for calculating the fugacity coefficients are performed numerically using Gaussian quadrature method. We found that five point quadrature method leads to a result with acceptable accuracy. As it can be seen from Eq. (101) the fugacity coefficient is a function of temperature, pressure, composition and the properties of the components. In order to calculate the fugacity coefficient of each component, we should first calculate the density of mixture at a given pressure, temperature and composition using Eq. (2). As it can be seen, from the mentioned equation, the density as function of the pressure is not known explicitly. Therefore, the estimation of the density at a given pressure should be performed using an iterative procedure, starting from initial guesses because of the multiplicity of the solution. A solution which is obtained by an iterative technique depends on the choice of the initial guess. Therefore, iterative procedures can not cover all acceptable roots unless the number of roots and the approximate values of the solutions (i.e. initial guesses) had already been known. Hence, an alternative, robust, fast and accurate technique that can predict all acceptable solutions is proposed. The proposed method is based on a numerical interpolation using Chebyshev polynomials in a finite interval (Burden et al.,

It should be pointed out that Chebyshev series provide high accuracy and can be transformed to power series which are suitable for root finding procedure. More general accounts of root finding through Chebyshev approximations are given in (Boyd, 2006). The aforementioned method enables us to calculate all possible solutions and select among them

It should be considered that using Chebyshev polynomials to approximate a given function will become more efficient when it has non-zero values at both end points of the interval. It can be shown that the pressure vs. density function in SAFT EoS linearly goes to zero for negligible values of the density. In order to avoid this problem, � �⁄ vs. density using

Another advantage of using Chebyshev polynomials for approximating a function is that for a specific number of basic functions, it always leads to a well-conditioned matrix during the calculation of the unknown coefficients of the basis functions, which is more accurate than

��

���

**3. Application of Chebyshev polynomials to calculate density** 

� � ��� � �ξ� �� ��

1981).

those which are physically interpretable.

Chebyshev polynomials was interpolated.

the other interpolation techniques.

��������

Figure 2 shows the interpolation error using Chebyshev polynomials of degree 15 for approximating pressure vs. density of a binary mixture of ethanol and toluene containing 37.5 mole% ethanol. Figure 3 shows the error in interpolation for another system (oil sample) for which the composition is given in Table 2 (Jamaluddin et al., 2000).

Fig. 2. Interpolation error using Chebyshev polynomials for approximating preesure vs. density of a binary mixture of ethanol and toluene containing 37.5 mole% of ethanol at different temperatures.

Fig. 3. Interpolation error using Chebyshev polynomials for approximating preesure vs. density at different temperatures for an oil sample of the composition given in Table 2.

After approximating the �(�) function using Chebyshev polynomials, it is necessary to find solutions for density values at the given pressure(s) and select those which are physically interpretable. In doing so, the complex and negative solutions and those which make �� �� ⁄

Application of Chebyshev Polynomials to Calculate Density and

Fugacity Using SAFT Equation of State to Predict Asphaltene Precipitation Conditions 27

Fig. 4. A typical plot of pressure versus density of SAFT EoS at different temperatures.

Fig. 5. A typical plot of pressure versus density of SAFT EoS at different temperatures.

density.

corresponds to the vapor phase density.

the calculated root is identified as the vapor phase density.

4. If two physically meaning roots are obtained at the given pressure, the smaller root corresponds to vapor phase density and the larger one corresponds to the liquid phase

5. If the system has only a single root at the given pressure and �� �� ⁄ has two zeros, if the obtained root is larger than the larger root of �� �� ⁄ , it is identified as the liquid phase density, otherwise, if the estimated root is smaller than the smaller root of �� �� ⁄ , it

6. If the system has only a single root at the given pressure and �� �� ⁄ has not any zero,


Table 2. Composition (mole%) and properties of the oil sample used to investigate the effect of temperature and pressure on asphaltene precipitation (Jamaluddin et al., 2000).

negative, are discarded because they have no physical meaning. Figure 4 shows a typical plot of pressure versus density for SAFT EoS in the positive region of density. As it can be seen in Figure 4, the derivative of pressure with respect to density (�� �� ⁄ ) has two zeros in this region for different values of the shown temperatures. For pressures between the maximum and minimum of the �(�) function (e.g. the pressure region between two parallel lines passing through the maximum and minimum of the middle curve), the system has three zeros one of which is not acceptable. The smaller root corresponds to the vapor phase density and the larger root corresponds to the liquid phase density. At pressures below the minimum of �(�), the function has only a single root which is identified as the vapor phase density. At pressures above the maximum of �(�), only a single zero is detected for the function which is identified as the liquid phase density. By increasing the temperature (Figure 5), the roots of �� �� ⁄ approaches to each other. At some temperature they coincide above which �� �� ⁄ has not any zero. At these temperatures the system has only a single root for any value of the pressure which is identified as the vapor phase density. Therefore, the procedure for finding roots of the SAFT EoS at the given pressure can be summarized as the following:


N2 0.49 CO2 11.37 H2S 3.22 C1 27.36 C2 9.41 C3 6.70 iC4 0.81 nC4 3.17 iC5 1.22 nC5 1.98 C6 2.49 C7+ 31.79

Component and Properties Oil

C7+ molecular weight 248.3 C7+ density (g/cm3) 0.877

of temperature and pressure on asphaltene precipitation (Jamaluddin et al., 2000).

the following:

polynomials.

Table 2. Composition (mole%) and properties of the oil sample used to investigate the effect

negative, are discarded because they have no physical meaning. Figure 4 shows a typical plot of pressure versus density for SAFT EoS in the positive region of density. As it can be seen in Figure 4, the derivative of pressure with respect to density (�� �� ⁄ ) has two zeros in this region for different values of the shown temperatures. For pressures between the maximum and minimum of the �(�) function (e.g. the pressure region between two parallel lines passing through the maximum and minimum of the middle curve), the system has three zeros one of which is not acceptable. The smaller root corresponds to the vapor phase density and the larger root corresponds to the liquid phase density. At pressures below the minimum of �(�), the function has only a single root which is identified as the vapor phase density. At pressures above the maximum of �(�), only a single zero is detected for the function which is identified as the liquid phase density. By increasing the temperature (Figure 5), the roots of �� �� ⁄ approaches to each other. At some temperature they coincide above which �� �� ⁄ has not any zero. At these temperatures the system has only a single root for any value of the pressure which is identified as the vapor phase density. Therefore, the procedure for finding roots of the SAFT EoS at the given pressure can be summarized as

1. The pressure versus density of SAFT EoS is approximated using Chebyshev

2. The derivative of pressure with respect to density is calculated to find zeros of �� �� ⁄ .

3. The roots of the fitting polynomial are estimated at the given pressure using a proper root finding algorithm for polynomials. The negative and complex roots and those

The complex and negative zeros are eliminated.

which make �� �� ⁄ negative are eliminated.

Fig. 4. A typical plot of pressure versus density of SAFT EoS at different temperatures.

Fig. 5. A typical plot of pressure versus density of SAFT EoS at different temperatures.


Application of Chebyshev Polynomials to Calculate Density and

Fugacity Using SAFT Equation of State to Predict Asphaltene Precipitation Conditions 29

Fig. 7. Experimental and calculated densities versus pressure at different temperatures using SAFT EoS for binary system composed of ethanol and toluene at 25.0 mole% of ethanol.

Fig. 8. Experimental and calculated densities versus pressure at different temperatures using SAFT EoS for binary system composed of ethanol and toluene at 37.5 mole% of ethanol.

#### **4. Results and discussion**

#### **4.1 Density calculation for binary systems of ethanol and toluene**

The SAFT EoS was first applied to calculate densities of the asymmetrical binary systems composed of ethanol and toluene. Experimental liquid densities for ethanol (1) and toluene (2) and seven of their binary mixtures in the temperature range 283.15-353.15 K at each 10 K and for pressures up to 45 MPa in steps of 5 MPa are given in (Zeberg-Mikkelsen et al., 2005). No density measurements were performed at 353.15 K and 0.1 MPa for ethanol as well as for mixtures containing more than 25 mole% ethanol, since ethanol and all mixtures with a composition higher than 25 mole% ethanol is either located in the two phase region or the gaseous phase (Zeberg-Mikkelsen et al., 2005). A comparison of the experimental density values of the aforementioned binary mixtures and pure compounds with the values calculated using SAFT EoS has been performed in this work. Figure 6 shows plots of the compressibility factor (Z-factor) of ethanol for different pressures of 0.1, 25 and 45 MPa using the SAFT EoS. As can be seen in this figure the contribution from the hard chain term (��� � ��� � ���), the dispersion term �������, and the association term (������) are shown at different pressures versus density. Each point on a constant pressure curve corresponds to a certain temperature. Increasing the temperature, the liquid density decreases. A comparison between experimental and calculated densities using SAFT equation are presented in figures 7-10 versus pressure for different temperatures. The average absolute values of the relative deviations (*AAD*) found between experimental and calculated densities for different compositions of the binary mixtures of ethanol and toluene at different pressures and temperatures is 0.143%. Figure 11 represents relative deviations for different mixtures of ethanol and toluene on a 3D diagram.

Fig. 6. Contributions to Z-factor of ethanol at different pressures and temperatures according to SAFT EoS.

The SAFT EoS was first applied to calculate densities of the asymmetrical binary systems composed of ethanol and toluene. Experimental liquid densities for ethanol (1) and toluene (2) and seven of their binary mixtures in the temperature range 283.15-353.15 K at each 10 K and for pressures up to 45 MPa in steps of 5 MPa are given in (Zeberg-Mikkelsen et al., 2005). No density measurements were performed at 353.15 K and 0.1 MPa for ethanol as well as for mixtures containing more than 25 mole% ethanol, since ethanol and all mixtures with a composition higher than 25 mole% ethanol is either located in the two phase region or the gaseous phase (Zeberg-Mikkelsen et al., 2005). A comparison of the experimental density values of the aforementioned binary mixtures and pure compounds with the values calculated using SAFT EoS has been performed in this work. Figure 6 shows plots of the compressibility factor (Z-factor) of ethanol for different pressures of 0.1, 25 and 45 MPa using the SAFT EoS. As can be seen in this figure the contribution from the hard chain term (��� � ��� � ���), the dispersion term �������, and the association term (������) are shown at different pressures versus density. Each point on a constant pressure curve corresponds to a certain temperature. Increasing the temperature, the liquid density decreases. A comparison between experimental and calculated densities using SAFT equation are presented in figures 7-10 versus pressure for different temperatures. The average absolute values of the relative deviations (*AAD*) found between experimental and calculated densities for different compositions of the binary mixtures of ethanol and toluene at different pressures and temperatures is 0.143%. Figure 11 represents relative deviations for different mixtures of

Fig. 6. Contributions to Z-factor of ethanol at different pressures and temperatures

**4.1 Density calculation for binary systems of ethanol and toluene** 

**4. Results and discussion** 

ethanol and toluene on a 3D diagram.

according to SAFT EoS.

Fig. 7. Experimental and calculated densities versus pressure at different temperatures using SAFT EoS for binary system composed of ethanol and toluene at 25.0 mole% of ethanol.

Fig. 8. Experimental and calculated densities versus pressure at different temperatures using SAFT EoS for binary system composed of ethanol and toluene at 37.5 mole% of ethanol.

$$f\_l^V = f\_l^L \qquad \qquad l = 1, \dots, N\_{\mathbb{C}} \tag{214}$$

$$f\_{N\_c}^L = f\_{N\_c}^S \tag{215}$$

$$\sum\_{l=1}^{N\_c} \frac{\{K\_l^{VL} - 1\} \mathbf{z}\_l}{1 + V\{K\_l^{VL} - 1\} + \mathcal{S}\{K\_l^{SL} - 1\}} = \mathbf{0} \tag{216} \tag{216}$$

Application of Chebyshev Polynomials to Calculate Density and

C7+ 66.68

Table 3. Composition (mole%) and properties of oil ad solvent from Burke et al. (1990).

Test Pressure (psia)

0 600 596.14 3014.7 0.14 0.1578 8.97 20 1050 1053.3 3014.7 0.27 0.2667 7.83 50 2310 2142 3014.7 1.46 1.4700 6.96 70 3750 3467 4214.7 1.65 1.6503 4.34 78 4510 4565 5014.7 3.21 3.4025 7.84 85 5000 5395 5014.7 1.29 1.2714 8.02 90 4250 4578 5014.7 1.10 1.0428 7.17

Table 4. Experimental and calculated values of the amount of asphaltene precipitation and saturation pressure for different mixtures of the oil sample and solvent given in Table 3.

Exp. Calc. Exp. Calc.

Precipitates from oil (wt%)

Total Precipitaes (wt%)

C7+ molecular weight 281 C7+ specific gravity 0.9020 Oil molecular weight 202.4 API gravity of stock tank oil 24.0 Reservoir temperature, °F 218 Saturation pressure, psia 600

> Mixture Saturation Pressure (psia)

Solvent (mol%)

Fugacity Using SAFT Equation of State to Predict Asphaltene Precipitation Conditions 33

N2 0.51 3.17 CO2 1.42 17.76 C1 6.04 30.33 C2 7.00 26.92 C3 6.86 13.09 iC4 0.83 1.26 nC4 3.35 4.66 iC5 0.70 0.77 nC5 3.46 1.26 C6 3.16 0.78

**Component and Properties Oil Solvent** 

$$\sum\_{l=1}^{N\_c} \frac{\left\{K\_l^{SL} - 1\right\} \mathbf{z}\_l}{1 + V\left\{K\_l^{VL} - 1\right\} + S\left\{K\_l^{SL} - 1\right\}} = \mathbf{0} \tag{217} \tag{217}$$

where, � and � are the mole fraction of the gas and solid phases, respectively. �� �� and �� �� are the equilibrium ratios of the vapor-liquid and liquid-solid phase equilibrium, respectively. In order to find the values of � and � using the above system of equations, we used the Newton-Raphson method. The complexity of multiphase flash calculations is due to the fact that the number of phases in equilibrium is not known a priori. The stability test for performing multiphase flash calculations has been performed using the stage-wise method developed by Michelsen (Michelsen, 1982a, 1982b). The parameters of the SAFT EoS including �, � and � have been given by Gross and Sadowski (2001) for N2, CO2 and hydrocarbons from C1 to C20. The values of these parameters for heavier and lumped component, also the volume and energy parameters used in association term have been estimated using a tuning approach. The thermodynamic model was applied to predict the precipitation behavior of petroleum fluids. The SAFT model has been used to describe the oil and gas phases. Table 3 presents the composition of an oil sample and a solvent from Burke et al. (1990). The oil sample was mixed with various amounts of solvent. Table 4 shows the precipitation and saturation pressure data for different concentrations of solvent in the oil sample for which the compositions are given in Table 3. The weight percent corresponds to the percentages with respect to the original mass of the oil. The last column in Table 4 shows the total amount of precipitates. To estimate the fugacity of pure asphaltene phase at a reference state, data at 0 mole% of solvent and �<sup>∗</sup> = 3014.7 psia and � = �1� � was used. The amount of precipitate, 0.14 weight%, was removed from the feed. The fugacity of the asphaltene component in the remaining mixture was then calculated using SAFT equation of state and equated to �� ∗ (reference state fugacity of the asphaltene). The amounts of asphaltene precipitation and saturation pressures are also calculated using WinProp (CMG software) in which the fluid phases are described with a cubic equation of state and the fugacities of components in the solid phase are predicted using the solid model desribed in (Nghiem & Heidemann, 1982). In this work the Peng-Robinson equation of state has been used to describe the fluid phases non-ideality. Figures 12-13 represent a comparison between experimental and calculated asphaltene precipitation amounts and saturation pressures at different concentrations of solvent in the oil sample using SAFT EoS and WinProp software. As it can be seen in Figure 12 and Table 4, for solvent concentration above 78 mole%, the measured values show a substantial decrease in the precipitate. For this case, WinProp does not show a drop in the amount of precipitate. Howevere, it shows that the amount of precipitate levels off at high concentration of the solvent. Burke et al. (1990), and Chaback (1991) attributed the decrease in the precipitate at high solvent concentration to the switching of the mixture from bubble point fluid to a dew point fluid. WinProp does not show a decrease in the precipitate, while SAFT EoS based on the developed method for the calculation of densities and fugacities shows a very good agreement with measured data. The saturation pressures calculated using the proposed model at solvent concentrations above 78 mole % correspond to the upper dew point pressures while those calculated using WinProp are the buuble point pressues (Figure 13). The average relative deviations of saturation pressure and asphaltene precipitation amount using SAFT EoS are 4.6% and 3.8%, respectively. The values of AAD obtained using WinProp are higher than 17% for both parameters.

the equilibrium ratios of the vapor-liquid and liquid-solid phase equilibrium, respectively. In order to find the values of � and � using the above system of equations, we used the Newton-Raphson method. The complexity of multiphase flash calculations is due to the fact that the number of phases in equilibrium is not known a priori. The stability test for performing multiphase flash calculations has been performed using the stage-wise method developed by Michelsen (Michelsen, 1982a, 1982b). The parameters of the SAFT EoS including �, � and � have been given by Gross and Sadowski (2001) for N2, CO2 and hydrocarbons from C1 to C20. The values of these parameters for heavier and lumped component, also the volume and energy parameters used in association term have been estimated using a tuning approach. The thermodynamic model was applied to predict the precipitation behavior of petroleum fluids. The SAFT model has been used to describe the oil and gas phases. Table 3 presents the composition of an oil sample and a solvent from Burke et al. (1990). The oil sample was mixed with various amounts of solvent. Table 4 shows the precipitation and saturation pressure data for different concentrations of solvent in the oil sample for which the compositions are given in Table 3. The weight percent corresponds to the percentages with respect to the original mass of the oil. The last column in Table 4 shows the total amount of precipitates. To estimate the fugacity of pure asphaltene phase at a reference state, data at 0 mole% of solvent and �<sup>∗</sup> = 3014.7 psia and � = �1� � was used. The amount of precipitate, 0.14 weight%, was removed from the feed. The fugacity of the asphaltene component in the remaining mixture was then

asphaltene). The amounts of asphaltene precipitation and saturation pressures are also calculated using WinProp (CMG software) in which the fluid phases are described with a cubic equation of state and the fugacities of components in the solid phase are predicted using the solid model desribed in (Nghiem & Heidemann, 1982). In this work the Peng-Robinson equation of state has been used to describe the fluid phases non-ideality. Figures 12-13 represent a comparison between experimental and calculated asphaltene precipitation amounts and saturation pressures at different concentrations of solvent in the oil sample using SAFT EoS and WinProp software. As it can be seen in Figure 12 and Table 4, for solvent concentration above 78 mole%, the measured values show a substantial decrease in the precipitate. For this case, WinProp does not show a drop in the amount of precipitate. Howevere, it shows that the amount of precipitate levels off at high concentration of the solvent. Burke et al. (1990), and Chaback (1991) attributed the decrease in the precipitate at high solvent concentration to the switching of the mixture from bubble point fluid to a dew point fluid. WinProp does not show a decrease in the precipitate, while SAFT EoS based on the developed method for the calculation of densities and fugacities shows a very good agreement with measured data. The saturation pressures calculated using the proposed model at solvent concentrations above 78 mole % correspond to the upper dew point pressures while those calculated using WinProp are the buuble point pressues (Figure 13). The average relative deviations of saturation pressure and asphaltene precipitation amount using SAFT EoS are 4.6% and 3.8%, respectively. The values of AAD obtained using WinProp are higher than 17%

= 0 � = 1� � � �� (217)

�� and ��

∗ (reference state fugacity of the

�� are

�� � 1�

where, � and � are the mole fraction of the gas and solid phases, respectively. ��

� ���

��

���

for both parameters.

1 � ����

�� � 1���

�� � 1� � ����

calculated using SAFT equation of state and equated to ��


Table 3. Composition (mole%) and properties of oil ad solvent from Burke et al. (1990).


Table 4. Experimental and calculated values of the amount of asphaltene precipitation and saturation pressure for different mixtures of the oil sample and solvent given in Table 3.

Application of Chebyshev Polynomials to Calculate Density and

and WinProp.

and 13.19%).

Fugacity Using SAFT Equation of State to Predict Asphaltene Precipitation Conditions 35

Fig. 14. Comparison between experimental and calculated values of the saturation pressure, upper and lower AOP's for the oil sample given in (Jamaluddin et al., 2002) using SAFT EoS

deviation in calculating the upper AOP using SAFT EoS and WinProp are 0.002% and 5.26%, respectively. Also, shown in Figure 14 are the calculated values of the lower AOP at different temperatures. The amounts of asphaltene precipitation vs. pressure at different temperatures is presented in Figure 15 for SAFT EoS. As it shown (Fig. 15) the maximum values of the asphaltene precipitation occures at the bubble point pressure of the mixture above which the amount of precipitaes decreases by increasing pressure up to upper AOP. In the pressure range below the bubble point pressure, decreasing pressure leads to a decrease in the amount of asphaltene precipitation and becomes infinitesimal at lower AOP. The effect of the injection gas on the asphaltene precipitation conditions has also been investigated for an oil sample given in (Rydahl et al., 1997) (Figures 16-17). Figure 17 shows that by increasing the amount of the injection gas added to the initial oil, the pressure interval of the asphaltene precipitation increases and shifts to the right side of the figure. Also, the amount of asphaltene precipitation increases by increasing the amount of the

The amount of asphalltene precipitation at different dilution ratios of normal heptane (nC7) and stock tank conditions, are also calculated using both models for two Iranian oil samples, Sarvak oil A and Fahliyan oil (Bagheri et al., 2009). Figures 18-19 shows a comparison between experimental and calculated amount of asphaltene precipitation vs. dilution ration of nC7 using SAFT EoS and WinProp. Again an excellent agreement has been observed using SAFT EoS with the experimental data. As it can be seen in these figues using WinProp, the amount of precipitaes increases rapidly at low dilution ratios after which the slope of the curve does not change considerably. The results show that SAFT EoS with average relative deviations of 2.32% and 1.73% for Sarvak oil A and Fahliyan oil, agrees well with the experimental data in comparison with the results obtained using WinProp (10.26%

injection gas. A similar scenario happens when using WinProp.

Fig. 12. Comparison between experimental and calculated values of the amount of asphaltene precipitation for different mixtures of the oil sample and solvent given in Table 3 using SAFT EoS and WinProp.

Fig. 13. Comparison between experimental and calculated values of the saturation pressure for different mixtures of the oil sample and solvent given Table 3 using SAFT EoS and WinProp.

The effect of temperature and pressure on the solid model prediction results has also been investigated. To do so, different hydrocarbon mixtures have been used. Figure 14 shows a comparison between the calculated and experimental values of the bubble point pressure and upper asphaltene onset pressure (AOP) for an oil sample (Jamaluddin et al., 2002). As it can be seen in this figure, excellent agreement is observed between experimental and predicted values of the upper AOP using SAFT EoS. The average values of the relative

Fig. 12. Comparison between experimental and calculated values of the amount of

using SAFT EoS and WinProp.

WinProp.

asphaltene precipitation for different mixtures of the oil sample and solvent given in Table 3

Fig. 13. Comparison between experimental and calculated values of the saturation pressure for different mixtures of the oil sample and solvent given Table 3 using SAFT EoS and

The effect of temperature and pressure on the solid model prediction results has also been investigated. To do so, different hydrocarbon mixtures have been used. Figure 14 shows a comparison between the calculated and experimental values of the bubble point pressure and upper asphaltene onset pressure (AOP) for an oil sample (Jamaluddin et al., 2002). As it can be seen in this figure, excellent agreement is observed between experimental and predicted values of the upper AOP using SAFT EoS. The average values of the relative

Fig. 14. Comparison between experimental and calculated values of the saturation pressure, upper and lower AOP's for the oil sample given in (Jamaluddin et al., 2002) using SAFT EoS and WinProp.

deviation in calculating the upper AOP using SAFT EoS and WinProp are 0.002% and 5.26%, respectively. Also, shown in Figure 14 are the calculated values of the lower AOP at different temperatures. The amounts of asphaltene precipitation vs. pressure at different temperatures is presented in Figure 15 for SAFT EoS. As it shown (Fig. 15) the maximum values of the asphaltene precipitation occures at the bubble point pressure of the mixture above which the amount of precipitaes decreases by increasing pressure up to upper AOP. In the pressure range below the bubble point pressure, decreasing pressure leads to a decrease in the amount of asphaltene precipitation and becomes infinitesimal at lower AOP. The effect of the injection gas on the asphaltene precipitation conditions has also been investigated for an oil sample given in (Rydahl et al., 1997) (Figures 16-17). Figure 17 shows that by increasing the amount of the injection gas added to the initial oil, the pressure interval of the asphaltene precipitation increases and shifts to the right side of the figure. Also, the amount of asphaltene precipitation increases by increasing the amount of the injection gas. A similar scenario happens when using WinProp.

The amount of asphalltene precipitation at different dilution ratios of normal heptane (nC7) and stock tank conditions, are also calculated using both models for two Iranian oil samples, Sarvak oil A and Fahliyan oil (Bagheri et al., 2009). Figures 18-19 shows a comparison between experimental and calculated amount of asphaltene precipitation vs. dilution ration of nC7 using SAFT EoS and WinProp. Again an excellent agreement has been observed using SAFT EoS with the experimental data. As it can be seen in these figues using WinProp, the amount of precipitaes increases rapidly at low dilution ratios after which the slope of the curve does not change considerably. The results show that SAFT EoS with average relative deviations of 2.32% and 1.73% for Sarvak oil A and Fahliyan oil, agrees well with the experimental data in comparison with the results obtained using WinProp (10.26% and 13.19%).

Application of Chebyshev Polynomials to Calculate Density and

Fugacity Using SAFT Equation of State to Predict Asphaltene Precipitation Conditions 37

(a)

(b)

Fig. 17. The amount of asphaltene precipitation vs. pressure at different values of the injection gas for the oil and gas samples given in (Rydahl et al., 1997) using (a) SAFT EoS

and (b) WinProp.

Fig. 15. The amount of asphaltene precipitation vs. pressure at different temperatures for the oil sample given in (Jamaluddin et al., 2002) using SAFT EoS.

Fig. 16. Calculated values of the saturation pressure, upper and lower asphaltene AOP's vs. the amount of injection gas for the oil sample given in (Rydahl et al., 1997) using SAFT EoS and WinProp.

Fig. 15. The amount of asphaltene precipitation vs. pressure at different temperatures for the

Fig. 16. Calculated values of the saturation pressure, upper and lower asphaltene AOP's vs. the amount of injection gas for the oil sample given in (Rydahl et al., 1997) using SAFT EoS

oil sample given in (Jamaluddin et al., 2002) using SAFT EoS.

and WinProp.

(a)

(b)

Fig. 17. The amount of asphaltene precipitation vs. pressure at different values of the injection gas for the oil and gas samples given in (Rydahl et al., 1997) using (a) SAFT EoS and (b) WinProp.

Application of Chebyshev Polynomials to Calculate Density and

Fugacity Using SAFT Equation of State to Predict Asphaltene Precipitation Conditions 39

Fig. 19. Comparison between experimental and calculated values of the amount of asphaltene precipitation vs. dilution ration of nC7 for Fahliyan Oil (Bagheri et al., 2009)

using SAFT EoS and WinProp.

Fig. 18. Comparison between experimental and calculated values of the amount of asphaltene precipitation vs. dilution ration of nC7 for Sarvak Oil A (Bagheri et al., 2009) using SAFT EoS and WinProp.

Fig. 18. Comparison between experimental and calculated values of the amount of asphaltene precipitation vs. dilution ration of nC7 for Sarvak Oil A (Bagheri et al., 2009)

using SAFT EoS and WinProp.

Fig. 19. Comparison between experimental and calculated values of the amount of asphaltene precipitation vs. dilution ration of nC7 for Fahliyan Oil (Bagheri et al., 2009) using SAFT EoS and WinProp.

Application of Chebyshev Polynomials to Calculate Density and

*Chem. Res.*, Vol. 30, pp. 1994–2005.

*Conference*, Calgary, Canada, Jane 4-8, 2000.

pp. 953–962.

Vol. 9, pp. 1-19.

*Phase Equilibria*, Vol. 9, pp. 21-40.

Press, Talor & Francis Group LLC.

*Fundam.*, Vol. 15, pp. 59-64.

1260.

Fugacity Using SAFT Equation of State to Predict Asphaltene Precipitation Conditions 41

Economou, I.G. (2002).A Successful Model for the Calculation of Thermodynamics and

Huang, S.H. & Radosz, M. (1990). Equation of State for Small, Large, Polydisperse and Associating molecules, Ind. Eng. Chem. *Res.*, Vol. 29, No. 11, pp. 2284–2294. Huang, S.H. & Radosz, M. (1991). Equation of State for Small, Large, Polydisperse and

Jamaluddin, A.K.M.; Joshi, N.; Joseph, M.T.; D'Cruz, D.; Ross, B.; Greek, J.; Kabir. C.S. &

Jamaluddin, K.M.; Joshi, N.; Iwere, F. & Gurpinar, F. (2002). An Investigation of Asphaltene

Associated Fluid Theory. *J. Chem. Phys.*, Vol. 121, No. 21, pp. 10715–10724. Llovell, F.; Vega, L.F.; Seiltgens, D.; Meja, A. & Segura, H. (2008). An Accurate Direct

Crossover Treatment. *Fluid Phase Equilibria*, Vol. 264, No. 1–2, pp. 201–210. Michelsen, M.L. (1982a). The Isothermal Flash Problem, I. Stability. *Fluid Phase Equilibria*,

Michelsen, M.L. (1982b). The Isothermal Flash Problem, II. Phase Split Calculation. *Fluid* 

Muller, E.A. & Gubbins, K.E. (1995). An Equation of State for Water from a Simplified Intermolecular Potential*. Ind. End. Chem. Res.*, Vol. 34, pp. 3662–3673. Nghiem, L.X., Heidemann, R.A. (1982). General Accelaration Procedure for Multiphase

Peng, D.Y. & Robinson, D.B. (1976). A New Two Constant Equation of State. *Ind. Eng. Chem.* 

Prausnitz, J.M.; Lichtenthaler, R.N. & de Azevedo, E.G. (1999). *Molecular Thermodynamics of* 

Prausnitz, J.M. & Tavares, F.W. (2004). Thermodynamics of Fluid-Phase Equilibria for Standard Chemical Engineering Operations. *AIChE J.*, Vol. 50, No.4, pp. 739–761. Redlich, O. & Kwong, J.N.S. (1949). On the Thermodynamics of Solutions. V. An Equation of Sate. Fugacities of Gaseous Solutions. *Chem. Rev.*, Vol. 44, pp. 233–244.

*Fluid Phase Equilibria* (Third Edition), Prentice Hall PTR, pp. 390.

Firoozabadi. (1999). *Thermodynamics of Hydrocarbon Reservoirs*, McGraw-Hill, NewYork. Gross, J. & Sadowski, G. (2001). Perturbed-Chain SAFT : An Equation of State Based on

Phase Equilibria Propeties of Complex Fluid Mixtures. *Ind. Eng. Chem. Res.*, Vol. 41,

Perturbation Theory for Chain Molecules. *Ind. Eng. Chem. Res.*, Vol. 40, pp. 1244-

Associating molecules, Ind. Eng. Chem : Extension to Fluid Mixtures. *Ind. Eng.* 

Fadden, J.D. (2000). Laboratory Techniques to Defines the Asphaltene Deposition Envelop, *Proceedings of the Petroleum Society's Canadian International Petroleum* 

Instability under Nitrogen Injection, *Proceedings of SPE International Petroleum Conference and Exhibition*, SPE 74393, Villahermosa, Mexico, February 10-12, 2002. Katz, D.L. & Firoozabadi, A. (1978). Predicting Phase Behavior of Condensate/Crude Oil Systems Using Methane Interaction Coefficients. *J. Pet. Tech.*, Vol. 27, pp. 1649. Llovell, F.; Pamies, J.C. & Vega, L.F. (2004). Thermodynamic Properties of Lennard-Jones

Chain Molecules: Renormalization Group Corrections to a Modified Statistical

Technique for Parametrizing Cubic Equations of State: Part III. Application of a

Flash Calculation with Application to Oil-Gas-Water Systems, *Proceedings of the 2nd European Symposium on Enhanced Oil Recovery*, Paris, France, November 8-10, 1982. Pedersen, K.S. & Christensen, P.L. (2007). *Phase Behavior of Petroleum Reservoir Fluids*, CRC

#### **5. Conclusion**

In this study a model based on statistical association fluid theory (SAFT) has been developed to predict phase behavior of hydrocarbon systems containing asphaltene and associating components. A robust, fast and accurate method based on Chebyshev polynomial approximation was proposed to find density using SAFT EoS which plays an important role in the calculation of the fugacity coefficients. The model was first evaluated using binary systems of ethanol and toluene. A good agreement between experimental and calculated liquid densities at different pressures, temperatures and compositions was obtained. The proposed model was then used to investigate the effect of solvent addition on the amount of asphaltene precipitate. The results showed a good agreement between experimental and calculated values of the amount of precipitate for different solvent–oil mixtures. In addition, the effect of temperature on the onset pressure of asphaltene precipitation and bubble point pressure was investigated. An excellent agreement was observed between experimental and predicted values of the asphaltene onset pressure at different temperatures.

#### **6. References**


In this study a model based on statistical association fluid theory (SAFT) has been developed to predict phase behavior of hydrocarbon systems containing asphaltene and associating components. A robust, fast and accurate method based on Chebyshev polynomial approximation was proposed to find density using SAFT EoS which plays an important role in the calculation of the fugacity coefficients. The model was first evaluated using binary systems of ethanol and toluene. A good agreement between experimental and calculated liquid densities at different pressures, temperatures and compositions was obtained. The proposed model was then used to investigate the effect of solvent addition on the amount of asphaltene precipitate. The results showed a good agreement between experimental and calculated values of the amount of precipitate for different solvent–oil mixtures. In addition, the effect of temperature on the onset pressure of asphaltene precipitation and bubble point pressure was investigated. An excellent agreement was observed between experimental and predicted values of the asphaltene onset pressure at

Bagheri, MB; Kharrat, R. & Mirzabozorg, A. (2009). A Novel Method to Develop a New

Benedict, M.; Webb, G.B. & Rubin, L.C. (1942). An Empirical Equation for Thermodynamic

Boyd, J.P. (2006). Computing Real Roots of a Polynomial in Chebyshev Series form Through

Buenrostro-Gonzales, E. & Lira-Galeana, C. (2004). Asphaltene Precipitation in Crude Oils,

Burden, R.L.; Faures, J.D. & Reynolds, A.C. (1981). *Numerical analysis* (Second Edition),

Burke, N.E.; Hobbs, R.E. & Kashou, S.F.R. (1990). Measurement and Modeling of Asphaltene

Chaback JJ. (1991). Discussion on Measurement and Modeling of Asphaltene Precipitation.

Chapman, W.G.; Jackson, G. & Gubbins, K.E. (1988). Phase Equilibria for Associating Fluids Chain Molecules with Multiple Bonding Sites. *Molec. Phys.*, Vol. 65, pp. 1057. Chapman, W.G.; Gubbins, K.E., Jackson, G. & Radosz, M. (1990). New Reference Equation of State for Associating Liquids. *Ind. Eng. Chem. Res.*, Vol. 29, pp. 1709-1721. Chapman, W.G.; Sauer, Sh.G.; Ting, D. & Ghosh, A. (2004). Phase Behavior Applications of

Danesh, A. (1998). *PVT and Phase Behavior of Petroleum Reservoir Fluids*, Elsevier Science,

SAFT Based Equations of State from Associating Fluids to Polydisperse, Polar

Ethane, Propane, and n-Butane. *J. Chem. Physics*, Vol. 10, pp. 747.

Subdivision. *Appl. Numer. Math.*, Vol. 56, No. 8, pp. 1077–1091.

Ttheory and Experiments. *AIChE J.*, Vol. 50, pp. 2552-2570.

Precipitation. *J. Petroleum Tech.*, Vol. 42, pp.1440.

*Journal of Petroleum Technology*. Vol. 43, pp. 1519-1520.

Copolymers. *Fluid Phase Equilibria*, Vol. 217, pp. 137–143.

Prindle, Weber & Schmidt, Boston.

ISBN 0444 821961, Netherlands.

Scaling Equation for Modeling of Asphaltene Precipitation, *Proceedings of the 2009 SPE/EAGE Reservoir Characterization and Simulation*, Abu Dhabi, UAE, October 19-

Properties of Light Hydrocarbons and Their Mixtures II: Mixtures of Methane,

**5. Conclusion**

different temperatures.

21, 2009.

**6. References**


**2** 

Ailing Yang

*China* 

*Ocean University of China* 

**Based on Common Inverted Microscope** 

**Oil-Gas Inclusions and Colour Analysis** 

Hydrocarbon fluid inclusions occur in a variety of geological environments, most commonly within carbonate rocks from petroliferous sedimentary (Stasiuk & Snowdon,1997). Oil-gas inclusions (OGIs) belong to hydrocarbon inclusions. The sizes of OGIs are usually in the range of 5-20m. A micro-mass (ng-fg) palaeo-oils were trapped in a single OGI. The interaction between OGIs and outside is relative weak although a long geological age elapsing. The OGI can be seen as a closed system. In this way, an intact (not cracked) OGI can be seen as a microoil-gas reservoir. The OGIs take rich information about the palaeo-oils. Generally, in the evolution of the oil-gas reservoir, with the deposition depth increasing, the stratum temperature increases. The organic macromolecules were decomposed into micromolecules. The oil maturity evolutes from low to high. Similarly, in the different digenetic stages, the types, colours and composites of the OGIs are also different. With the maturity of organism from low to high, the types of OGIs are mainly from liquid phase, liquid and gas phase to gas phase. The colours change from colourless, yellow, and brown to black (Liu, Y. R. *et al*., 2003, as cited in Burruss R.C.,1991). Micro-fluorescence properties of OGIs, largely controlled by the aromatic characteristics of the hydrocarbons, are a signature of the organic chemical composition. This phenomenon was often used to distinguish OGIs from saltwater inclusions. By different fluorescence colours of OGIs, the oil and gas charge history and oil-gas maturity are qualitatively determined. And by the abundance of grains containing OGIs (GOI)(Liu, K. Y. & Eadington, 2005, as cited in Eadington, 1996), the petroleum characteristics of the corresponding strata can be qualitatively estimated. The micro-spectroscopy is very important

The VIS spectra technique includes fluorescence micro photometry (FMP), fluorescence alteration of multiple macerals (FAMM) and laser scanning confocal microscope (LSCM).

FMP is a technique combined microscope and photometer (see Fig.1). Generally, the exciting source is UV light (365 nm, called internal light source in this chapter) from a mercury arc

**1. Introduction** 

to OGIs and mainly includes VIS and UV-VIS spectra.

**1.1.1 Fluorescence micro photometry (FMP)** 

**1.1 The VIS spectra technique** 

**to Measure UV-VIS Spectra of Single** 


### **Based on Common Inverted Microscope to Measure UV-VIS Spectra of Single Oil-Gas Inclusions and Colour Analysis**

Ailing Yang *Ocean University of China China* 

#### **1. Introduction**

42 Advances in Chemical Engineering

Rydahl, A.K.; Pedersen, K.S. & Hjermstad, H.P. (1997). Modeling of Live Oil Asphaltene, *Precipitation Proceedings of AIChE Spring Meeting*, Houston, March 9-13, 1997. Segura, H.; Seiltgens, D.; Mejia, A.; Llovell, F. & Vega, L.F. (2008). An Accurate Direct

Soave, G. (1972). Equilibrium Constants from a Modified Redlich-Kwong Equation of State.

Song, Y.; Lambert, S.M. & Prausnitz, J.M. (1994). Equation of State for Mixtures of Hard Sphere Chains Including Copolymers. *Macromolecules*, Vol. 27, pp. 441. Starling, K.E. (1973). *Fluid Thermodynamic Properties for Light Petroleum Systems*, Gulf

Tabatabaei-Nejad, S.A.R. & Khodapanah, E. (2009). An Investigation on the Sensitivity

Ting, D.L., (2003). Thermodynamic Stability and Phase Behavior of Asphaltenes in Oil and

Van der Waals, J.D. (1873). Over de Continuiteit van der Gas en Vloeistoftoestand Leiden,

Verdier, S.; Carrier, H.; Andersen, S.I. & Daridon, J.L. (2006). Study of Pressure and

Wertheim, M.S. (1984). Fluids with Highly Directional Attractive Forces. I. Statistical

Wertheim, M.S. (1986). Fluids with Highly Directional Attractive Forces. III. Multiple

Wertheim, M.S. (1987). Thermodynamic Perturbation Theory of Polymerization. *J. Chem.* 

Zeberg-Mikkelsen, C.K.; Lugo, L.; Garcia, J. & Fernandez*,* J. (2005). Volumetric Properties

under Pressure for the Binary System Ethanol + Toluene, *Fluid Phase Equilibria*, Vol.

Analysis of the Parameters of Proposed Wax Precipitation Model. *Journal of* 

of Other Highly Asymmetric Mixtures, PhD Thesis, Rice University, Houston,

Temperature Effects on Asphaltene Stability in Presence of CO2. *Energy and Fuels*,

pp. 155–172.

USA.

*Chem. Eng. Sci.*, Vol. 27, pp. 1197-1203.

Doctoral Dissertation, The Netherlands.

Vol. 20, pp. 1584-1590.

*Phys.*, Vol. 87 , pp. 7323.

235, pp. 139–151.

*Petroleum Science* & *Engineering*, Vol. 68, pp. 89–98.

Thermodynamics. *J. Stat. Phys.*, Vol. 35, No.1, pp. 19–47.

Attraction Sites. *J. Stat. Phys.*, Vol. 42, No.3, pp. 459–476.

Publishing Company, Houston.

Technique for Parameterizing Cubic Equations of State. Part II. Specializing Models for Predicting Vapor Pressures and Phase Densities. *Fluid Phase Equilibria*, Vol. 265,

> Hydrocarbon fluid inclusions occur in a variety of geological environments, most commonly within carbonate rocks from petroliferous sedimentary (Stasiuk & Snowdon,1997). Oil-gas inclusions (OGIs) belong to hydrocarbon inclusions. The sizes of OGIs are usually in the range of 5-20m. A micro-mass (ng-fg) palaeo-oils were trapped in a single OGI. The interaction between OGIs and outside is relative weak although a long geological age elapsing. The OGI can be seen as a closed system. In this way, an intact (not cracked) OGI can be seen as a microoil-gas reservoir. The OGIs take rich information about the palaeo-oils. Generally, in the evolution of the oil-gas reservoir, with the deposition depth increasing, the stratum temperature increases. The organic macromolecules were decomposed into micromolecules. The oil maturity evolutes from low to high. Similarly, in the different digenetic stages, the types, colours and composites of the OGIs are also different. With the maturity of organism from low to high, the types of OGIs are mainly from liquid phase, liquid and gas phase to gas phase. The colours change from colourless, yellow, and brown to black (Liu, Y. R. *et al*., 2003, as cited in Burruss R.C.,1991). Micro-fluorescence properties of OGIs, largely controlled by the aromatic characteristics of the hydrocarbons, are a signature of the organic chemical composition. This phenomenon was often used to distinguish OGIs from saltwater inclusions. By different fluorescence colours of OGIs, the oil and gas charge history and oil-gas maturity are qualitatively determined. And by the abundance of grains containing OGIs (GOI)(Liu, K. Y. & Eadington, 2005, as cited in Eadington, 1996), the petroleum characteristics of the corresponding strata can be qualitatively estimated. The micro-spectroscopy is very important to OGIs and mainly includes VIS and UV-VIS spectra.

#### **1.1 The VIS spectra technique**

The VIS spectra technique includes fluorescence micro photometry (FMP), fluorescence alteration of multiple macerals (FAMM) and laser scanning confocal microscope (LSCM).

#### **1.1.1 Fluorescence micro photometry (FMP)**

FMP is a technique combined microscope and photometer (see Fig.1). Generally, the exciting source is UV light (365 nm, called internal light source in this chapter) from a mercury arc

Based on Common Inverted Microscope to Measure

UV-VIS Spectra of Single Oil-Gas Inclusions and Colour Analysis 45

itself. This will lead to a decreasing reliability. So one should be cautious to judge the oil and

In 1997, Stasiuk *et al*. (Stasiuk & Snowdon,1997) used Zeiss MPV II to study the fluorescence of OGIs. The wavelength of the exciting light is 365 nm. A 2.5-5 m light spot was obtained by a pin hole. The size of the light spot is less than that of the OGIs. The authors measured the micro-fluorescence spectra of artificially synthesized and natural hydrocarbon inclusions. The fluorescence spectra were related with the chemical constitutes, API degree and other geochemistry parameters of crude oils. The migration of oil and gas was revealed. The crude oils used to synthesize the artificial hydrocarbon inclusions composite a wide range of saturated hydrocarbons, aromatic hydrocarbons and NSO compounds. And also span a long geologic time (Ordovician, Devonian, Carboniferous, Jurassic and Cretaceous). For the artificial hydrocarbon inclusions, the spectra are the real spectra for single hydrocarbon inclusions. And for the natural inclusions, there are some organic compounds in the mineral grains difficult to be observed even under the microscope, so strictly speaking, the fluorescence spectra of natural inclusions are the addition of the inclusions and the mineral matrix. The results showed that most artificially synthesized hydrocarbon inclusions have main colours. A small amount have other colours. The authors gave a detailed analysis between the peak position *L*max, *Q* index (Q=Intensity @ 650 nm / Intensity @ 500 nm) of the fluorescence spectra and saturated hydrocarbons, aromatic hydrocarbons, NSO compounds, nC18/Ph、nC18/Pr , API degree of the crude oils. The results showed the correlative factors between the peak position *L*max and the saturated hydrocarbons, *Q* index and saturated hydrocarbons are all up to 0.84. *Q* index can be used to indicate the maturities of the crude oils. Caja *et al*. (Caja, 2007, 2009) measured the fluorescence spectra of hydrocarbon inclusions. The results showed that *Q*580 index (580-700nm vs 400-580nm areas) is very sensitive to the total composites of the petroleum. The higher the *Q*580 index, the heavier the crude oil and the tighter relation with the API degree. For the hydrocarbon inclusions emitting yellow fluorescence, the API degree is in the range of 15-22. And for the hydrocarbon inclusions

gas charge history and oil-gas maturity by the fluorescence colours of the OGIs.

radiating blue fluorescence, the API degree is in the range of 23-45.

In 1990s, CSITRO of Australia established FAMM. This technique combined laser with microscope. For laser has good monochromaticity and high intensity, a micrometer light spot can be easily obtained, which enhances the space resolution of the setup. By means of micro-Raman setup, assuming Ar+ laser as exciting light source (488 nm), a focal spot as less as several micrometer was obtained. The emitting intensities of source rocks at 625 nm were detected with time changing. Wilkins (Wilkins *et al*., 1992, 1995) probed the relation between vitrinite inhibiting and the fluorescence intensity changing with time. The result showed that FAMM can effectively correct inhibiting effect of vitrinite reflectivity before the stage of high maturity. Lo (Lo, *et al*., 1997) and Veld (Veld *et al*.,1997) used this technique to measure the maturity of source rock. With the help of FAMM, Xiao (Xiao *et al.*, 2002) used a laser induced fluorescence microscopy (LIFM) to study the maturities of carbonate rocks with

FAMM and LIFM techniques are not very fitful to measure the fluorescence spectra of single OGIs. After assuming laser, one can measure the fluorescence spectra, but can not

**1.1.2 Fluorescence alteration of multiple macerals (FAMM)** 

higher maturities.

Fig. 1. A schematic diagram of a FMP. The upright microscope includes in a beam splitting diachronic mirror (with higher reflectivity to the incident light and also a higher transmittance to the fluorescence) and a diachronic mirror (absorbing the scattered light from the sample, with a higher transmission to the fluorescence).

lamp in a microscope. The UV light is focused by an objective and incidents onto a sample. The sample absorbs the exciting photons then emits fluorescence. The wavelengths of the fluorescence are longer than that of the incident light.

Some fluorescence backs (upward for the upright microscope and downward for the inverted microscope) and is collected by the same microscope objective. That light enters into the aperture of the photometer and the intensity is measured. For the inverted microscope, except for the downward fluorescence, one can measure the upward (transmitted) fluorescence because there is a big space above the sample stage to connect a photometer with the body of the microscope.

According to different study contents, one can choose different filter, beam splitting diachronic mirror, diachronic mirror and detector to realize fluorescence, transmission, reflective and absorption measurement. For example, in Fig. 1, if the photometer is an energy meter, one can measure the intensity of the fluorescence. And if the diachronic mirror has a higher absorption to the fluorescence and a higher transmission to the scattered light from the sample, the reflectivity of the sample can be measured. MPV3 microphotometer is an example. It was widely used to measure the vitrinite reflectivity of coal and source rock (Xiao *et al*., 1998, 2000).

The superiors of the above FMP are simple and low cost. But there are still some shortcomings. The first is one can only gain visible spectra (400-780 nm) for the exciting light is at 365nm. But some aromatic hydrocarbons emit UV fluorescence which can not be measured by above method. The second is the pin hole (about 0.15 mm) to limit the incident light spot often larger than the under studied OGI. Not only the under studied OGI emits fluorescence, but also other OGIs, the cements around the mineral grain trapping the OGI and even the mineral grain itself. The other OGIs may not be the same generation as the under studied OGI. What' more, the intensities of the cements are usually stronger than that of the under studied OGI. That means the fluorescence of the under studied OGI often mixes with other fluorescence. This makes the OGI looks like involving into a colourful cloud. So one can see if the size of the incident light spot is larger than that of the under studied OGI, the measured fluorescence spectrum may include the fluorescence of other OGIs, the cements around the mineral grain trapping the OGI and even the mineral grain

beam splitting diachronic mirror

diachronic mirror

photometer

Fig. 1. A schematic diagram of a FMP. The upright microscope includes in a beam splitting

lamp in a microscope. The UV light is focused by an objective and incidents onto a sample. The sample absorbs the exciting photons then emits fluorescence. The wavelengths of the

Some fluorescence backs (upward for the upright microscope and downward for the inverted microscope) and is collected by the same microscope objective. That light enters into the aperture of the photometer and the intensity is measured. For the inverted microscope, except for the downward fluorescence, one can measure the upward (transmitted) fluorescence because there is a big space above the sample stage to connect a

According to different study contents, one can choose different filter, beam splitting diachronic mirror, diachronic mirror and detector to realize fluorescence, transmission, reflective and absorption measurement. For example, in Fig. 1, if the photometer is an energy meter, one can measure the intensity of the fluorescence. And if the diachronic mirror has a higher absorption to the fluorescence and a higher transmission to the scattered light from the sample, the reflectivity of the sample can be measured. MPV3 microphotometer is an example. It was widely used to measure the vitrinite reflectivity of

The superiors of the above FMP are simple and low cost. But there are still some shortcomings. The first is one can only gain visible spectra (400-780 nm) for the exciting light is at 365nm. But some aromatic hydrocarbons emit UV fluorescence which can not be measured by above method. The second is the pin hole (about 0.15 mm) to limit the incident light spot often larger than the under studied OGI. Not only the under studied OGI emits fluorescence, but also other OGIs, the cements around the mineral grain trapping the OGI and even the mineral grain itself. The other OGIs may not be the same generation as the under studied OGI. What' more, the intensities of the cements are usually stronger than that of the under studied OGI. That means the fluorescence of the under studied OGI often mixes with other fluorescence. This makes the OGI looks like involving into a colourful cloud. So one can see if the size of the incident light spot is larger than that of the under studied OGI, the measured fluorescence spectrum may include the fluorescence of other OGIs, the cements around the mineral grain trapping the OGI and even the mineral grain

transmittance to the fluorescence) and a diachronic mirror (absorbing the scattered light

diachronic mirror (with higher reflectivity to the incident light and also a higher

objective sample

from the sample, with a higher transmission to the fluorescence).

filter mecury lamp

fluorescence are longer than that of the incident light.

photometer with the body of the microscope.

coal and source rock (Xiao *et al*., 1998, 2000).

itself. This will lead to a decreasing reliability. So one should be cautious to judge the oil and gas charge history and oil-gas maturity by the fluorescence colours of the OGIs.

In 1997, Stasiuk *et al*. (Stasiuk & Snowdon,1997) used Zeiss MPV II to study the fluorescence of OGIs. The wavelength of the exciting light is 365 nm. A 2.5-5 m light spot was obtained by a pin hole. The size of the light spot is less than that of the OGIs. The authors measured the micro-fluorescence spectra of artificially synthesized and natural hydrocarbon inclusions. The fluorescence spectra were related with the chemical constitutes, API degree and other geochemistry parameters of crude oils. The migration of oil and gas was revealed. The crude oils used to synthesize the artificial hydrocarbon inclusions composite a wide range of saturated hydrocarbons, aromatic hydrocarbons and NSO compounds. And also span a long geologic time (Ordovician, Devonian, Carboniferous, Jurassic and Cretaceous). For the artificial hydrocarbon inclusions, the spectra are the real spectra for single hydrocarbon inclusions. And for the natural inclusions, there are some organic compounds in the mineral grains difficult to be observed even under the microscope, so strictly speaking, the fluorescence spectra of natural inclusions are the addition of the inclusions and the mineral matrix. The results showed that most artificially synthesized hydrocarbon inclusions have main colours. A small amount have other colours. The authors gave a detailed analysis between the peak position *L*max, *Q* index (Q=Intensity @ 650 nm / Intensity @ 500 nm) of the fluorescence spectra and saturated hydrocarbons, aromatic hydrocarbons, NSO compounds, nC18/Ph、nC18/Pr , API degree of the crude oils. The results showed the correlative factors between the peak position *L*max and the saturated hydrocarbons, *Q* index and saturated hydrocarbons are all up to 0.84. *Q* index can be used to indicate the maturities of the crude oils.

Caja *et al*. (Caja, 2007, 2009) measured the fluorescence spectra of hydrocarbon inclusions. The results showed that *Q*580 index (580-700nm vs 400-580nm areas) is very sensitive to the total composites of the petroleum. The higher the *Q*580 index, the heavier the crude oil and the tighter relation with the API degree. For the hydrocarbon inclusions emitting yellow fluorescence, the API degree is in the range of 15-22. And for the hydrocarbon inclusions radiating blue fluorescence, the API degree is in the range of 23-45.

#### **1.1.2 Fluorescence alteration of multiple macerals (FAMM)**

In 1990s, CSITRO of Australia established FAMM. This technique combined laser with microscope. For laser has good monochromaticity and high intensity, a micrometer light spot can be easily obtained, which enhances the space resolution of the setup. By means of micro-Raman setup, assuming Ar+ laser as exciting light source (488 nm), a focal spot as less as several micrometer was obtained. The emitting intensities of source rocks at 625 nm were detected with time changing. Wilkins (Wilkins *et al*., 1992, 1995) probed the relation between vitrinite inhibiting and the fluorescence intensity changing with time. The result showed that FAMM can effectively correct inhibiting effect of vitrinite reflectivity before the stage of high maturity. Lo (Lo, *et al*., 1997) and Veld (Veld *et al*.,1997) used this technique to measure the maturity of source rock. With the help of FAMM, Xiao (Xiao *et al.*, 2002) used a laser induced fluorescence microscopy (LIFM) to study the maturities of carbonate rocks with higher maturities.

FAMM and LIFM techniques are not very fitful to measure the fluorescence spectra of single OGIs. After assuming laser, one can measure the fluorescence spectra, but can not

Based on Common Inverted Microscope to Measure

**1.2 The UV-VIS spectra technique for single OGIs** 

is necessary to develop UV-VIS spectra technique.

range of 220-900 nm and record the UV-VIS spectra.

**2. The geology background of the samples** 

oil at suitable depth.

the range of VIS.

section of this chapter.

UV-VIS Spectra of Single Oil-Gas Inclusions and Colour Analysis 47

LSCM can effectively resolve the cellular tissue in the algae. The author observed that algae have stable outermost cell wall predominance. The result assured kerogen has a selected preservation during the formation. LSCM was also used to obtain images about the space distribution of organism in mineral slides. Stasiuk (Stasiuk, 2001) measured fluorescence spectra of lipoids and chlorophyll originated from diatom. The mineral slides in the Saanich Inlet region are in rich spore oil and reproductive spore, which is relative to the periodic spore blooming in the spring. The results under the LSCM may be the preservation of spore

Although LSCM has a higher space resolution and good image quality, till now the shortest laser wavelength of semiconductor laser is 370nm, the fluorescence of single OGI is still in

As above mentioned, some organic composites in OGIs emit fluorescence at UV range. So it

For overcoming above drawbacks, Kihle (Kihle, 1995) established an UV-VIS microspectroscopy for measuring excitation-emission spectra of single OGIs. The setup includes an upright tri-ocular deep ultraviolet (DUV) microscope, a condensed lens, a collimated objective, two fiber adaptors, two fiber cables, and an UV-VIS spectrometer. The DUV microscope is very expensive comparing with the common fluorescence microscope. But for measuring the UV-VIS spectra of single OGIs, it had to be used. The spectrometer was connected with the microscope by the two fiber cables, the condensed lens, the collimated objective and the two fiber adaptors. The spectrometer supplies the exciting light in the

The setup could successfully measure the UV-VIS spectra of single OGIs. But there are four points to be noted. The first is for obtaining a small focal point, one has to replace the original focusing mirror of the microscope itself with a condensed mirror which was settled down on a 3D adjustable adaptor. The second is the C-mount was connected with the fiber cable not a videograph head as origin. So one can only observe the sample by ocular and can't take photos of the samples or get the sizes of the single OGIs and focusing point precisely. The third is the expensive DUV microscope, its price is 9-20 times of a common fluorescence microscope. This will greatly limit it using. The fourth is that for effective collecting fluorescence an oil immersed objective had to be used, which is trouble to clean. The above experimental system is so expensive and one has to rebuild the microscope. Is it possible to replace the DUV microscope with a common microscope to lower cost and still detect UV-VIS fluorescence spectra of single OGIs? Such kind of micro-spectroscopy system to meet this aim was established (Yang, 2009a, 2009b, 2011). This will be given in the third

The sand rocks came from the core drilling samples of five oil wells (Bai 95#, Hua5#, Jian 22#, Nong 29# and Fu 4# ), Jinlin Oil Filed, Songliao Basin, northeast China. Bai 95# oil well localizes in the west slope of the basin, the other are in the south to the basin. The earth

directively observe the sample by eyes because the laser safety threshold for human eyes is less than 5 W (Sliney D. H. ,1995). For most matters, the absorption bands are in the range of UV. But the wavelengths of Ar+ ion laser are 488 nm and 515 nm, they don't mach well with the absorption bands of materials. So it is not very well to use Ar+ ion laser to excite the fluorescence of single OGI. The focused laser has a higher radiation flux density at the sample surface and easily damages the sample. The changing fluorescence intensities at 625 nm with time may be relevant to the Photolysis (Sanches S., 2011) and photo polymerization of vitrinite in the coal and source rock.

#### **1.1.3 Laser scanning confocal microscope (LSCM)**

LSCM was realized in 1980s. Fig. 2 is a schematic diagram of a LSCM. The illuminating pinhole and the detector pinhole are conjugate to the focal plane of the objective. A point on the focal plane of the objective is focused at the illuminating and detecting pinholes at the same time. The other points out of the focal plane of the objective have not images at the detecting pinhole. This means confocal. A point light source illuminating and a point image are realized by an illuminating pinhole before the laser and a detecting pinhole before the detector. The light passes through the detecting pinhole will be received by a PMT or a cCCD and quickly imaged in a computer. The images have higher space resolution and good quality. By controlling the movement of the objective, LSCM can realize continues optical section by tomography similar to CT. After computer 3D imaging, a 3D profile of a micro sample can be recovered.

Fig. 2. The schematic diagram of a LSCM.

LSCM has been widely used in cytobiology, cytophysiology, neurophysiology and other modern medicine and biology relative to cells (Damaskinos *et al.*, 1995; Kevin, 2003). It has been a forceful tool in the field of bio-science. Recently, LSCM was used to obtain precise volume ratios between liquid and gas phases in OGIs. Combining homogenous temperature, GC-MS analysis and PVTsim software, the pressure of palaeo-liquid can be obtained (Aplin *et al*., 1999; Thiéry *et al*. 2002, Liu, D. H. *et al*., 2003). Further more, the depths of the strata existing palaeo-liquids can also be obtained by the pressure. This is meaningful to study the migration of oil, gas and reservoir formation.

LSCM has special superiors as resolving the fluorescent macerals with micrometer sizes (source rock and oil shale). Stasiuk (Stasiuk, 1999) used LSCM to observe algae in oil shale.

directively observe the sample by eyes because the laser safety threshold for human eyes is less than 5 W (Sliney D. H. ,1995). For most matters, the absorption bands are in the range of UV. But the wavelengths of Ar+ ion laser are 488 nm and 515 nm, they don't mach well with the absorption bands of materials. So it is not very well to use Ar+ ion laser to excite the fluorescence of single OGI. The focused laser has a higher radiation flux density at the sample surface and easily damages the sample. The changing fluorescence intensities at 625 nm with time may be relevant to the Photolysis (Sanches S., 2011) and photo polymerization

LSCM was realized in 1980s. Fig. 2 is a schematic diagram of a LSCM. The illuminating pinhole and the detector pinhole are conjugate to the focal plane of the objective. A point on the focal plane of the objective is focused at the illuminating and detecting pinholes at the same time. The other points out of the focal plane of the objective have not images at the detecting pinhole. This means confocal. A point light source illuminating and a point image are realized by an illuminating pinhole before the laser and a detecting pinhole before the detector. The light passes through the detecting pinhole will be received by a PMT or a cCCD and quickly imaged in a computer. The images have higher space resolution and good quality. By controlling the movement of the objective, LSCM can realize continues optical section by tomography similar to CT. After computer 3D imaging, a 3D profile of a

LSCM has been widely used in cytobiology, cytophysiology, neurophysiology and other modern medicine and biology relative to cells (Damaskinos *et al.*, 1995; Kevin, 2003). It has been a forceful tool in the field of bio-science. Recently, LSCM was used to obtain precise volume ratios between liquid and gas phases in OGIs. Combining homogenous temperature, GC-MS analysis and PVTsim software, the pressure of palaeo-liquid can be obtained (Aplin *et al*., 1999; Thiéry *et al*. 2002, Liu, D. H. *et al*., 2003). Further more, the depths of the strata existing palaeo-liquids can also be obtained by the pressure. This is

detecting pinhole

illuminating pinhole

laser

objective

sample

PMT

LSCM has special superiors as resolving the fluorescent macerals with micrometer sizes (source rock and oil shale). Stasiuk (Stasiuk, 1999) used LSCM to observe algae in oil shale.

meaningful to study the migration of oil, gas and reservoir formation.

of vitrinite in the coal and source rock.

micro sample can be recovered.

Fig. 2. The schematic diagram of a LSCM.

**1.1.3 Laser scanning confocal microscope (LSCM)** 

diachronic mirror

LSCM can effectively resolve the cellular tissue in the algae. The author observed that algae have stable outermost cell wall predominance. The result assured kerogen has a selected preservation during the formation. LSCM was also used to obtain images about the space distribution of organism in mineral slides. Stasiuk (Stasiuk, 2001) measured fluorescence spectra of lipoids and chlorophyll originated from diatom. The mineral slides in the Saanich Inlet region are in rich spore oil and reproductive spore, which is relative to the periodic spore blooming in the spring. The results under the LSCM may be the preservation of spore oil at suitable depth.

Although LSCM has a higher space resolution and good image quality, till now the shortest laser wavelength of semiconductor laser is 370nm, the fluorescence of single OGI is still in the range of VIS.

#### **1.2 The UV-VIS spectra technique for single OGIs**

As above mentioned, some organic composites in OGIs emit fluorescence at UV range. So it is necessary to develop UV-VIS spectra technique.

For overcoming above drawbacks, Kihle (Kihle, 1995) established an UV-VIS microspectroscopy for measuring excitation-emission spectra of single OGIs. The setup includes an upright tri-ocular deep ultraviolet (DUV) microscope, a condensed lens, a collimated objective, two fiber adaptors, two fiber cables, and an UV-VIS spectrometer. The DUV microscope is very expensive comparing with the common fluorescence microscope. But for measuring the UV-VIS spectra of single OGIs, it had to be used. The spectrometer was connected with the microscope by the two fiber cables, the condensed lens, the collimated objective and the two fiber adaptors. The spectrometer supplies the exciting light in the range of 220-900 nm and record the UV-VIS spectra.

The setup could successfully measure the UV-VIS spectra of single OGIs. But there are four points to be noted. The first is for obtaining a small focal point, one has to replace the original focusing mirror of the microscope itself with a condensed mirror which was settled down on a 3D adjustable adaptor. The second is the C-mount was connected with the fiber cable not a videograph head as origin. So one can only observe the sample by ocular and can't take photos of the samples or get the sizes of the single OGIs and focusing point precisely. The third is the expensive DUV microscope, its price is 9-20 times of a common fluorescence microscope. This will greatly limit it using. The fourth is that for effective collecting fluorescence an oil immersed objective had to be used, which is trouble to clean.

The above experimental system is so expensive and one has to rebuild the microscope. Is it possible to replace the DUV microscope with a common microscope to lower cost and still detect UV-VIS fluorescence spectra of single OGIs? Such kind of micro-spectroscopy system to meet this aim was established (Yang, 2009a, 2009b, 2011). This will be given in the third section of this chapter.

#### **2. The geology background of the samples**

The sand rocks came from the core drilling samples of five oil wells (Bai 95#, Hua5#, Jian 22#, Nong 29# and Fu 4# ), Jinlin Oil Filed, Songliao Basin, northeast China. Bai 95# oil well localizes in the west slope of the basin, the other are in the south to the basin. The earth

Based on Common Inverted Microscope to Measure

and c 10 m.

1

3 2 5

4

**single OGIs and colour analysis** 

**3.1 The experimental setup** 

spectrometer and (7) a computer.

(internal light source). The OGI emits light yellow fluorescence.

UV-VIS Spectra of Single Oil-Gas Inclusions and Colour Analysis 49

UV, blue or green light, they often emit fluorescence. Fig.3 shows the typical OGIs in the Bai 95# oil well. Fig.3 a is a micro-photo of the first episode oil inclusions, and Fig. 3b is the second episode OGIs. Fig.3 c is the fluorescence image of Fig.3 b as excited at 365 nm

Fig. 3. The two episodes OGIs in Bai 95# oil well, in which, a is the first episode, the depth is 426.8 m; b the second episode, the depth 420.1 m; and c the fluorescence image of b under excited at 365 nm internal light source in the microscope. The scale in a is 20 m, and in b

a b c

**3. Based on common inverted microscope to measure UV-VIS spectra of** 

Fig.4 shows the schematic diagram of the experimental setup (Yang, 2009a, 2009b), which includes seven parts: (1) an inverted fluorescence microscope (IMF); (2) a reflecting microscope objective (RMO); (3) a 3D adaptor; (4) a micro lens; (5) a fiber cable; (6) a

Fig. 4. Schematic diagram of the micro-spectroscopy system (a) based on common inverted microscope and the reflecting microscope objective (b). In a, 1IMF; 2RMO; 33D adaptor; 4micro-lens; 5 fiber cable; 6UV-VIS spectrometer and 7computer.

6

a b

7

UV-enhanced aluminium films

The IFM includes an internal light source (mercury arc lamp), three (violet, blue and green) standard fluorescence "cubes" bandpass filters, a dichronic mirrors and beam splitters in the optical light path. The mercury arc lamp can be used as an internal excitation source. Using bandpass filters and fluorescence "cubes", a beam with tens of nm band width can be obtained. There are five differences between this system and Kihle's setup (Kihle, 1995). (1)

strata belong to middle to shallow, upper Cretaceous Period. The thickness of inclusion thin slide is about 150 nm.

Table 1 shows the information about the strata, depths, GOIs, average homogeneous temperatures ( ) *T sh* and salinities of the sand rocks.

There are two episodes OGIs in these samples. The first episode is in the earlier stage of quartz overgrowth, the second in the later stage of the quartz overgrowth.


Table 1. Information about the strata, depths, the measurement results of GOIs, *T sh* and salinities of the sand rocks.

The GOIs, *Th* s and salinities are different in the two stages. In the first episode and for all strata, except for Jian 22# oil well, the GOIs belong to middle value (Liu, K. Y. & Eadington, P., 2005, as cited in Eadington, P, 1996). In the second episode, the GOIs of Bai 95# K2y1, K2qn2 and Hua 5# K2n1 decrease. But for Bai 95# K2y2+3 stratum, the GOI doubles. The GOIs in the second episode increase from 10 to 23 times of Jian 22#, Nong 29# and Fu 4 # oil wells, which shows there are lots of oils charging these strata in the second episode.

In the first episode, *Th* s are in the range of 72.3-81.0 C. And in the second episode, *T sh* increase. The increasing amplitude is in the range of 23- 60 C

The salinities of all strata decrease in the second episode. The *Th* s and salinities show that the deposition environment has a great change in the second episode. The second episode may be in the water expanding time.

Under the fluorescence microscope, the OGIs of the first episode are liquid phase and very small. They are dense and often distribute along a line or a stripe. Their colors are often deep brown or grey brown. The fluorescence intensities are very weak. It is very difficult to measure the fluorescence of such kind of oil inclusions. The second episode OGIs are liquid or liquid-gas phases. In a mineral grain, they distribute in scattering or in group. The liquids in the OGIs are light yellow, light brown yellow, brown and grey black. Under excited at UV, blue or green light, they often emit fluorescence. Fig.3 shows the typical OGIs in the Bai 95# oil well. Fig.3 a is a micro-photo of the first episode oil inclusions, and Fig. 3b is the second episode OGIs. Fig.3 c is the fluorescence image of Fig.3 b as excited at 365 nm (internal light source). The OGI emits light yellow fluorescence.

Fig. 3. The two episodes OGIs in Bai 95# oil well, in which, a is the first episode, the depth is 426.8 m; b the second episode, the depth 420.1 m; and c the fluorescence image of b under excited at 365 nm internal light source in the microscope. The scale in a is 20 m, and in b and c 10 m.

#### **3. Based on common inverted microscope to measure UV-VIS spectra of single OGIs and colour analysis**

#### **3.1 The experimental setup**

48 Advances in Chemical Engineering

strata belong to middle to shallow, upper Cretaceous Period. The thickness of inclusion thin

Table 1 shows the information about the strata, depths, GOIs, average homogeneous

There are two episodes OGIs in these samples. The first episode is in the earlier stage of

The first episode The second episode

GOI (%) *Th* (℃) Salinity (%) GOI (%) *Th* (℃) Salinity (%)

422.1 2-3 75.3 9.93 5-6 113.3 4.03

436.5 3 74.6 10.31 2 95.4 2.98

1524.0 2 \ \ 0.5 110.3 3.79

420.0 <1 81.0 6.77 20 104.2 4.89

595.9 2 77.0 11.34 20 116.8 7.86

426.6 2-3 75.0 7.17 70 135.2 4.35

Table 1. Information about the strata, depths, the measurement results of GOIs, *T sh* and

wells, which shows there are lots of oils charging these strata in the second episode.

increase. The increasing amplitude is in the range of 23- 60 C

The GOIs, *Th* s and salinities are different in the two stages. In the first episode and for all strata, except for Jian 22# oil well, the GOIs belong to middle value (Liu, K. Y. & Eadington, P., 2005, as cited in Eadington, P, 1996). In the second episode, the GOIs of Bai 95# K2y1, K2qn2 and Hua 5# K2n1 decrease. But for Bai 95# K2y2+3 stratum, the GOI doubles. The GOIs in the second episode increase from 10 to 23 times of Jian 22#, Nong 29# and Fu 4 # oil

In the first episode, *Th* s are in the range of 72.3-81.0 C. And in the second episode, *T sh*

The salinities of all strata decrease in the second episode. The *Th* s and salinities show that the deposition environment has a great change in the second episode. The second episode

Under the fluorescence microscope, the OGIs of the first episode are liquid phase and very small. They are dense and often distribute along a line or a stripe. Their colors are often deep brown or grey brown. The fluorescence intensities are very weak. It is very difficult to measure the fluorescence of such kind of oil inclusions. The second episode OGIs are liquid or liquid-gas phases. In a mineral grain, they distribute in scattering or in group. The liquids in the OGIs are light yellow, light brown yellow, brown and grey black. Under excited at

K2qn2 494.6 4 72.3 10.95 2 98.7 5.09

quartz overgrowth, the second in the later stage of the quartz overgrowth.

slide is about 150 nm.

Oil well Stratum Depth

K2y2+3

K2y1

Bai 95#

Hua 5# K2n1

Jian 22# K2q4

Nong 29# K2q4

Fu 4 # K2q3

salinities of the sand rocks.

may be in the water expanding time.

temperatures ( ) *T sh* and salinities of the sand rocks.

419.3-

426.8-

1449.8

377.8

571.0-

419.1-

Fig.4 shows the schematic diagram of the experimental setup (Yang, 2009a, 2009b), which includes seven parts: (1) an inverted fluorescence microscope (IMF); (2) a reflecting microscope objective (RMO); (3) a 3D adaptor; (4) a micro lens; (5) a fiber cable; (6) a spectrometer and (7) a computer.

Fig. 4. Schematic diagram of the micro-spectroscopy system (a) based on common inverted microscope and the reflecting microscope objective (b). In a, 1IMF; 2RMO; 33D adaptor; 4micro-lens; 5 fiber cable; 6UV-VIS spectrometer and 7computer.

The IFM includes an internal light source (mercury arc lamp), three (violet, blue and green) standard fluorescence "cubes" bandpass filters, a dichronic mirrors and beam splitters in the optical light path. The mercury arc lamp can be used as an internal excitation source. Using bandpass filters and fluorescence "cubes", a beam with tens of nm band width can be obtained. There are five differences between this system and Kihle's setup (Kihle, 1995). (1)

Based on Common Inverted Microscope to Measure

two slits widths are as same as step (5).

**3.3 The lest focal spot of the RMO** 

Fig. 5. The least light spot

determined.

**3.4 The background fluorescence and the subtract factor** 

permits the green light arrive to the videohead).

advantage to determine the sizes of the OGI and the focal spot.

UV-VIS Spectra of Single Oil-Gas Inclusions and Colour Analysis 51

The exciting and emitting slits are 10 nm and 8 nm respectively. A proper filter should be placed before the emitting window of the spectrometer. (6) Measure the micro fluorescence spectra of OGI excited by the internal light source (UV, blue and green). Note that the exciting slit of the spectrometer should be off. The emitting slit is 5 nm. (7) Measue the background fluorescence spectra when excited by external light source. The exciting wavelengths and the

Fig. 5 is the least focal spot of the RMO. The background is a quartz grain in an inclusion thin slide. The size of focal spot is about 12 m. The halo light around the focal spot is the result of the diffraction. It is difficult to be eliminated for a RMO. The focal spot includes most energy.

20m

The focal spot of the RMO is not small enough. When the size of the focal spot is larger than that of the OGI, not only the OGI emits fluorescence, but also the background. The background fluorescence is complex. It may origins as following: (1) The mineral grain may include in a few even lots of micro oil inclusions (Stasiuk & Snowdon, 1997; Meng, 2009), but they can't be resolved under the IMF. (2) The surface of the mineral grain was contaminated by oils existing in the sand rock as preparing the inclusion thin slide. (3) The mineral grain may contain some thulium, which will emit fluorescence when they are excited by light. (4) The inclusion thin slide may be contaminated by organism from air or by people's hands.

Fig.6 shows the transmission micro-images of a single OGI (a, c, and e) and background (quartz grain)(b, d and f). In Fig.6 a and Fig.6 b, the exciting wavelengths are 250 nm and a violet fluorescence filter cube was used. In Fig.6 c and Fig.6 d, the exciting wavelength is 470 nm and a blue fluorescence filter cube was used. And in Fig.6 e and Fig.6 f, the 550 nm light was as exciting beam, the fluorescence filter cube was violet (The green fluorescence filter cube blocks the green light to get to the videohead, but the violet fluorescence filter cube

At the 550 nm light exciting, the borderline of the OGI and the focal spot are clear. This will

From Fig.6, one can see that the fluorescence of single OGI is the addition of the background and the OGI. For decreasing the influence of the background, a subtract factor was

A common and cheap IFM was used but not an expensive upright DUV microscope. The 3D adjustable adaptor can conveniently connected to the main body of the IFM and not need any rebuilding to the microscope. (2) It is easy to adjust the RMO and the IMF coaxial. When the switch of the IMF is on and kept the least intensity, one can adjust the 3D adjustable adaptor while observe the reflected light intensity from the RMO. When the beam intensity is maximal, the two are coaxial. (3) The RMO has many advantages over refracting objective. It is an all-reflecting construction and free from chromatic aberration. The primary spherical aberration, primary coma and primary astigmatism have been corrected. The specific mirror coating is UV-enhanced aluminium film, which has as high as 89% average reflectivity in the range of 190 nm-10 m and is highly recommended for most UV use. Comparison with the refracting objective, the RMO has stronger focusing ability. (4) The RMO has a relative big numerical aperture (0.65), so it can be used as an excellent focusing element and also a good component to effectively collect fluorescence; (5) One can use the videohead to take photos for the samples and the focusing point in time.

In the setup, the spectrometer is FluroMax-4 (Horiba Jobin Yvon) with a 150 W xeon lamp and a single photon PMT. The RMO is made of Ealing (52X). The micro-lens and the fiber cable are from Avertise. The diameter of the fiber is 400 m and has a 90% transmission in the range of 240-800 nm. There are seven fibers in the fiber cable. The fibres are arranged like a club. The middle one is as exciting fiber to guide the exciting light from the xenon lamp in the spectrometer to the micro-lens. The peripheral six are as emitting fibers to guide the fluorescence into the spectrometer.

The system integrates the functions of micro-area location, DUV light excitation, weak fluorescence detection and real-time taking photos together. A computer program to calculate the chromaticity coordinates of the OGIs by the spectra excited at 365 nm was also established.

#### **3.2 The main measuring steps**

For observing, taking photos and measuring the fluorescence spectra of single OGIs, there are several key steps to be noted in the experiment. To decreasing influence from environment, the experiment was done in the dark. The main steps are as following: (1) Find and localize an OGI in the middle of the visual field. Take photos when white light illuminates the OGI. (2)Take photos as exciting by UV, blue and green light of the internal light source. Be careful not make the CCD satiate. (3) Adjust the 3D adaptor and make the RMO and the objective of the IMF coaxial. When the two are coaxial, one can observe a brightest and roundest spot over the RMO. Note the power of the green light from the mercury arc lamp least and ware a goggle glasses to protect eyes. (4) Obtain a least focal spot. Connect the micro-lens with the 3D adaptor, and also the fibre cable with the microlens. Switch on the spectrometer, let the exciting fibre guide the 550 nm green light out of the spectrometer and get to the micro-lens and then to the RMO. The RMO focuses the exciting light into a small spot and incident on the inclusion thin slide. If the adjustment is good enough, one can see a green focal spot on the slide. According to the position of the focal spot, one should repeatedly adjust x, y, and z axes of the adjustable adaptor and decrease the size of the focal spot. For obtaining a least focal spot, one should also adjust the relative distance between the micro-lens and the RMO. Once a least focal spot appearance, this distance should be fixed and not change it in the whole experiment. (5) Take photos of single OGI excited by external light source (Xeon lamp) and record the fluorescence spectra. The exciting and emitting slits are 10 nm and 8 nm respectively. A proper filter should be placed before the emitting window of the spectrometer. (6) Measure the micro fluorescence spectra of OGI excited by the internal light source (UV, blue and green). Note that the exciting slit of the spectrometer should be off. The emitting slit is 5 nm. (7) Measue the background fluorescence spectra when excited by external light source. The exciting wavelengths and the two slits widths are as same as step (5).

#### **3.3 The lest focal spot of the RMO**

50 Advances in Chemical Engineering

A common and cheap IFM was used but not an expensive upright DUV microscope. The 3D adjustable adaptor can conveniently connected to the main body of the IFM and not need any rebuilding to the microscope. (2) It is easy to adjust the RMO and the IMF coaxial. When the switch of the IMF is on and kept the least intensity, one can adjust the 3D adjustable adaptor while observe the reflected light intensity from the RMO. When the beam intensity is maximal, the two are coaxial. (3) The RMO has many advantages over refracting objective. It is an all-reflecting construction and free from chromatic aberration. The primary spherical aberration, primary coma and primary astigmatism have been corrected. The specific mirror coating is UV-enhanced aluminium film, which has as high as 89% average reflectivity in the range of 190 nm-10 m and is highly recommended for most UV use. Comparison with the refracting objective, the RMO has stronger focusing ability. (4) The RMO has a relative big numerical aperture (0.65), so it can be used as an excellent focusing element and also a good component to effectively collect fluorescence; (5) One can use the videohead to take

In the setup, the spectrometer is FluroMax-4 (Horiba Jobin Yvon) with a 150 W xeon lamp and a single photon PMT. The RMO is made of Ealing (52X). The micro-lens and the fiber cable are from Avertise. The diameter of the fiber is 400 m and has a 90% transmission in the range of 240-800 nm. There are seven fibers in the fiber cable. The fibres are arranged like a club. The middle one is as exciting fiber to guide the exciting light from the xenon lamp in the spectrometer to the micro-lens. The peripheral six are as emitting fibers to guide

The system integrates the functions of micro-area location, DUV light excitation, weak fluorescence detection and real-time taking photos together. A computer program to calculate the chromaticity coordinates of the OGIs by the spectra excited at 365 nm was also established.

For observing, taking photos and measuring the fluorescence spectra of single OGIs, there are several key steps to be noted in the experiment. To decreasing influence from environment, the experiment was done in the dark. The main steps are as following: (1) Find and localize an OGI in the middle of the visual field. Take photos when white light illuminates the OGI. (2)Take photos as exciting by UV, blue and green light of the internal light source. Be careful not make the CCD satiate. (3) Adjust the 3D adaptor and make the RMO and the objective of the IMF coaxial. When the two are coaxial, one can observe a brightest and roundest spot over the RMO. Note the power of the green light from the mercury arc lamp least and ware a goggle glasses to protect eyes. (4) Obtain a least focal spot. Connect the micro-lens with the 3D adaptor, and also the fibre cable with the microlens. Switch on the spectrometer, let the exciting fibre guide the 550 nm green light out of the spectrometer and get to the micro-lens and then to the RMO. The RMO focuses the exciting light into a small spot and incident on the inclusion thin slide. If the adjustment is good enough, one can see a green focal spot on the slide. According to the position of the focal spot, one should repeatedly adjust x, y, and z axes of the adjustable adaptor and decrease the size of the focal spot. For obtaining a least focal spot, one should also adjust the relative distance between the micro-lens and the RMO. Once a least focal spot appearance, this distance should be fixed and not change it in the whole experiment. (5) Take photos of single OGI excited by external light source (Xeon lamp) and record the fluorescence spectra.

photos for the samples and the focusing point in time.

the fluorescence into the spectrometer.

**3.2 The main measuring steps** 

Fig. 5 is the least focal spot of the RMO. The background is a quartz grain in an inclusion thin slide. The size of focal spot is about 12 m. The halo light around the focal spot is the result of the diffraction. It is difficult to be eliminated for a RMO. The focal spot includes most energy.

Fig. 5. The least light spot

#### **3.4 The background fluorescence and the subtract factor**

The focal spot of the RMO is not small enough. When the size of the focal spot is larger than that of the OGI, not only the OGI emits fluorescence, but also the background. The background fluorescence is complex. It may origins as following: (1) The mineral grain may include in a few even lots of micro oil inclusions (Stasiuk & Snowdon, 1997; Meng, 2009), but they can't be resolved under the IMF. (2) The surface of the mineral grain was contaminated by oils existing in the sand rock as preparing the inclusion thin slide. (3) The mineral grain may contain some thulium, which will emit fluorescence when they are excited by light. (4) The inclusion thin slide may be contaminated by organism from air or by people's hands.

Fig.6 shows the transmission micro-images of a single OGI (a, c, and e) and background (quartz grain)(b, d and f). In Fig.6 a and Fig.6 b, the exciting wavelengths are 250 nm and a violet fluorescence filter cube was used. In Fig.6 c and Fig.6 d, the exciting wavelength is 470 nm and a blue fluorescence filter cube was used. And in Fig.6 e and Fig.6 f, the 550 nm light was as exciting beam, the fluorescence filter cube was violet (The green fluorescence filter cube blocks the green light to get to the videohead, but the violet fluorescence filter cube permits the green light arrive to the videohead).

At the 550 nm light exciting, the borderline of the OGI and the focal spot are clear. This will advantage to determine the sizes of the OGI and the focal spot.

From Fig.6, one can see that the fluorescence of single OGI is the addition of the background and the OGI. For decreasing the influence of the background, a subtract factor was determined.

Based on Common Inverted Microscope to Measure

K2y2+3

K2y1

Bai 95#

Hua 5# K2n1

Jian 22# K2q4

Nong 29# K2q4

Fu 4 # K2q3

UV-VIS Spectra of Single Oil-Gas Inclusions and Colour Analysis 53

grain was chosen as background. There are 38 typical OGIs to be measured. Table 2 shows

Oil well Stratum Depth (m) No. Size (m) Phase F (%)

419.3 1 7.0, 12.3 L+G 86.0 420.1 2 7.7, 10.1 L+G 79.0

426.8 3 4.9, 10.7 L+G 94.2 431.0 4 8.3, 9.7 L+G 86.0 434.5 5 7.0, 10.7 L+G 88.7

435.0 7 12.5, 37.1 L 86.4

1449.8 10 4.5, 4.7 L+G 90.0 1522.4 11 17.3, 53.7 L+G 42.7 1523.0-1524.0 12\* 11.1, 12.5 L+G 0.0 1523.3 13 4.0, 4.0 L+G 96.0

377.8 14 16.6, 19.6 L+G 59.8 396.6 15 9.42, 9.38 L+G 49.0

401.9 17 9.6, 15.3 L+G 69.9

420.0 21\* 26.4, 27.1 L+G 38.3

571.0-572.0 23 11.0, 17.3 L+G 72.0

573.0-574.0 25 7.4, 10.6 L+G 82.0

419.1 30 9.8, 11.7 L 81.6

424.3 35 6.8, 7.0 L+G 89.2

425.2 37 15.9, 18.3 L+G 65.7

K2qn2 494.6 8 14.1, 14.6 L+G 57.2

6 7.7, 10.8 L+G 35.3

9 6.5, 10.1 L 75.0

16 15.0, 15.4 L+G 61.9

18 5.7, 6.6 L+G 91.5 19 21.6, 22.9 L+G 49.1 20 10.8, 16.7 L+G 64.7

22 17.1, 17.1 L+G 36.6

24 27.2, 30.0 L+G 0.0

26 8.7, 9.7 L+G 86.0 27 12.7, 25.4 L+G 65.6 28\* 37.2, 37.4 L+G 0.0 29 12.2, 25.3 L+G 50.0

31 20.8, 24.2 L+G 39.0

32 51.1, 71.3 L+G 0.0 33 10.3, 15.7 L+G 62.2 34 7.8, 13.6 L 78.0

36\* 22.6, 23.0 L+G 0.0

38 15.5, 25.3 L+G 55.6

the depths, strata, sizes, phases, F factors and corresponding oil wells of 38 OGIs.

404.7

595.9

422.8

oil wells of 38 OGIs. \* means the OGI is in the cement.

Table 2. The depths, strata, sizes, phases (L is liquid, and G gas), F factor and corresponding

Fig. 6. Transmission micro-images of single OGI (a, c and e) and the background (b, d and f). In which, a and b, c and d, and e and f are excited at 250 nm, 470 nm and 550 nm respectively. In a, b, e, and f, the violet fluorescence filter cube was used, which has a higher transmittance to the green light. In c and d, the blue fluorescence filter cube was used.

The focal spot includes most exciting energy. The energy distribution is approximated to be homogeneous on the spot area. The OGI and the background are seen as surface light sources. Assuming the areas of the OGI and the light spot are *S*1 and *S*2. The two areas are easily obtained from Fig.6 e and Fig.6 f. The luminous intensities from unit surfaces of the OGI and the background are *I*i and *I*b. If the total fluorescence intensity of OGI and part background (*S*2-*S*1) is *I*1(), the intensity of the background is *I*2(). Then

$$I\_1 = S\_1 I\_i + (S\_2 - S\_1) I\_b \tag{1}$$

$$I\_2 = \mathcal{S}\_2 I\_b. \tag{2}$$

From Eq. (1) and (2), the intensity of the OGI is

$$I\_i S\_1 = I\_1 - (1 - \frac{S\_1}{S\_2})I\_2 = I\_1 - FI\_2 \tag{3}$$

$$F = 1 - \frac{S\_1}{S\_2}.\tag{4}$$

Here *F* is the subtract factor. It can be obtained from the micro-photos of the OGI and the background. By the experiment, one can obtain *I*1, *I*2, *S*1, and *S*2. According to Eq. (3) and (4), one can further obtain the fluorescence spectra of single OGI. The following results based on above analysis.

#### **3.5 The UV-VIS spectra of single OGIs**

The first episode oil inclusions are very small and dense. The fluorescence is very weak. The background fluorescence greatly interferes the signal. The OGIs in the second episode are relative big and in scattering or in group in the mineral grains. It is possible to obtain the fluorescence spectra of single OGIs of the second episode, so these inclusions were chosen as samples.

In the experiment, we first observed dozens of OGIs in an inclusion thin slide then chose some typical OGIs to take photos and measure fluorescence spectra. The typical OGIs in an inclusion thin slide often have similar fluorescence colors. A relative clear area in the quartz

Fig. 6. Transmission micro-images of single OGI (a, c and e) and the background (b, d and f).

a b c d e f 20<sup>m</sup>

respectively. In a, b, e, and f, the violet fluorescence filter cube was used, which has a higher transmittance to the green light. In c and d, the blue fluorescence filter cube was used.

The focal spot includes most exciting energy. The energy distribution is approximated to be homogeneous on the spot area. The OGI and the background are seen as surface light sources. Assuming the areas of the OGI and the light spot are *S*1 and *S*2. The two areas are easily obtained from Fig.6 e and Fig.6 f. The luminous intensities from unit surfaces of the OGI and the background are *I*i and *I*b. If the total fluorescence intensity of OGI and part

> 1 11 21 2 2

> > 1 2

*<sup>S</sup> I S I I I FI*

1 . *<sup>S</sup> <sup>F</sup>*

Here *F* is the subtract factor. It can be obtained from the micro-photos of the OGI and the background. By the experiment, one can obtain *I*1, *I*2, *S*1, and *S*2. According to Eq. (3) and (4), one can further obtain the fluorescence spectra of single OGI. The following results based on

The first episode oil inclusions are very small and dense. The fluorescence is very weak. The background fluorescence greatly interferes the signal. The OGIs in the second episode are relative big and in scattering or in group in the mineral grains. It is possible to obtain the fluorescence spectra of single OGIs of the second episode, so these inclusions were chosen as

In the experiment, we first observed dozens of OGIs in an inclusion thin slide then chose some typical OGIs to take photos and measure fluorescence spectra. The typical OGIs in an inclusion thin slide often have similar fluorescence colors. A relative clear area in the quartz

(1 ) *<sup>i</sup>*

11 21 ( ) *i b I SI S S I* (1)

*<sup>S</sup>* (3)

2 2 . *<sup>b</sup> I SI* (2)

*<sup>S</sup>* (4)

In which, a and b, c and d, and e and f are excited at 250 nm, 470 nm and 550 nm

background (*S*2-*S*1) is *I*1(), the intensity of the background is *I*2(). Then

From Eq. (1) and (2), the intensity of the OGI is

**3.5 The UV-VIS spectra of single OGIs** 

above analysis.

samples.


grain was chosen as background. There are 38 typical OGIs to be measured. Table 2 shows the depths, strata, sizes, phases, F factors and corresponding oil wells of 38 OGIs.

Table 2. The depths, strata, sizes, phases (L is liquid, and G gas), F factor and corresponding oil wells of 38 OGIs. \* means the OGI is in the cement.

Based on Common Inverted Microscope to Measure

UV-VIS Spectra of Single Oil-Gas Inclusions and Colour Analysis 55

Fig. 7. The transmission (a, e, f, g, h, i and j) and reflective (b, c and d) micro-photos of No.6 OGI (Bai 95#, 434.5 m) when excited by white light (a), internal light source (c:violet; d:blue and e:green) and external light source (e:250 nm; f:365 nm; g:440 nm; h:470 nm; and i:546 nm), j is a transmission image when the 550 nm light exciting and violet fluorescence filter cube was used.

> D C B

a b c d e

f g h i j

Fig. 8. The fluorescence spectra of No.6 OGI excited by internal light source (in a, solid:

**500 550 600 650 700 750 800 850**

f Ex:546nm

**600 700 800 900 1000**

**Wavelength (nm)**

**Wavelength (nm)**

d Ex:440nm

**<sup>300</sup> <sup>400</sup> <sup>500</sup> <sup>600</sup> <sup>700</sup> <sup>800</sup> <sup>900</sup> <sup>1000</sup> <sup>0</sup>**

b Ex:250nm

**Wavelength** 

**0.0 2.0x104 4.0x104 6.0x104 8.0x104 1.0x105**

**5.0x10<sup>3</sup>**

**1.0x10<sup>4</sup>**

**1.5x10<sup>4</sup>**

**0**

**1x104**

**2x104**

**Intensity (CPS)**

**3x104**

**Intensity (CPS)**

violet, short dot: blue, and dash dot: green) and external light source (b-f).

**400 450 500 550 600 650 700**

**500 600 700 800 900**

**Wavelength (nm)**

c Ex:365nm

**400 500 600 700 800 900**

**Wavelength (nm)**

a

**Wavelength (nm)**

e Ex:470nm

**4.0x104 6.0x104 8.0x104 1.0x105 1.2x105**

**0.0 5.0x106 1.0x107 1.5x107 2.0x107**

20m

**2.0x104**

**4.0x104**

**Intensity (CPS)**

**6.0x104**

**8.0x104**

**Intensity (CPS)**

**Intensity (CPS)**

Fig.7-Fig.22 are the transmission, reflective micro photos and fluorescence spectra of OGIs of No.6, No.7, No.12, No.15, No.20, No.25, No.26 and No.34. No.6 and No.7 (Fig.7-Fig.10) are in Bai 95# oil well, the depths are 434.5 m and 435.0 m. No.12 (Fig.11-Fig.12) is in Hua 5# oil well, the depth is 1523-1524 m. No.15 (Fig. 13-Fig.14) and No.20(Fig.15-Fig.16)are in Jian 22# oil well, the depths are 396.6 m and 404.7 m. No. 25 (Fig.17-Fig.18) and No. 26 (Fig.19- Fig.20) are in Nong 29# oil well, the depths are 573-574 m and 595.6 m. No.34 (Fig.21-Fig.22) is in Fu 4# oil well, the depth is 422.8 m.

No.6, No.15, No.20, No.25 and No.26 are two phases (liquid and gas) inclusions. No. 7 and No.34 are single phase (liquid) oil OGIs. No.12, No.21 and No.36 are in cements. The others are all in the quartz grains.

From the micro photos, one can see when the internal light source excites the aimed OGI, not only the OGI emits fluorescence, but also other OGIs and the cements around the mineral grains. The intensities of the cements are often stronger than that of the OGIs. Even the mineral grains also emit weak fluorescence. These are the reasons why the spectra look like each other even the OGIs with different phases and colours. For example, No. 6 and No.7 OGIs are in the same oil well, the former is liquid-gas inclusion, the later is liquid phase. The spectra of the two excited by internal light source (see Fig.8 a and Fi9.10 a) are similar. Obviously, the spectra were influenced by the background. According to the spectra excited by violet light of the microscope (see Fig. 8 a and Fig.10 a, solid line), the peak positions are at 400 nm. The main aromatic hydrocarbons in the OGIs may be three or four cyclic hydrocarbons. But the other aromatic hydrocarbons in the OGIs are difficult to determine because except for the main peak, the other parts of the spectra are almost flat. The peak positions excited by the internal blue and green light are at 500-510 nm and 575 nm-600 nm respectively.

When the external light source (250 nm, 365 nm, 440 nm, 470 nm and 546 nm) excited the aimed OGI, the focal spot size is small. The cements were not illuminated, so the fluorescence spectra of the single OGIs weren't interfered by the spectra of the cements. If the OGIs are not dense in the grain, only one OGI is excited, one can obtain the spectra of single OGI after subtracting the background.

From Fig. 8 b, Fig.10 b, Fig.12 b, Fig.14 b, Fig.16 b, Fig.18 b, Fig.20 b and Fig.22 b, one can see that the spectra are in the range of UV-VIS at 250nm exciting. These results show that the experimental setup can measure the UV-VIS spectra of single OGIs. Combined the spectra excited by different external lights with different wavelengths, the characteristics of the OGIs are as following:

1. For most OGIs, when excited at 250 nm, there are three main peaks in the range of 400-470 nm, which are corresponding to three and four condensed aromatic rings, the palaeo oils are the medium Oils (Liu, K. Y., & Eadington P., 2005; Abbas *et al*., 2006). There are also two secondary peaks near to 535 nm and 610 nm, the former is the fluorescence of four or more condensed aromatic rings, the later comes from resin and asphaltene, the palaeo oils are medium and heavy Oils (Liu, K. Y., & Eadington P., 2005; Abbas *et al*., 2006). The spectra in the range of 280-400 nm belong to two or three condensed aromatic rings, the palaeo oils are light Oils (Liu, K. Y., & Eadington P., 2005; Abbas *et al*., 2006). So the OGIs of No.6, No.7, No.12, No.15, No.20, No.25 and No.34 are all filled with light, medium and heavy oils. And medium oils are the major part. The aromatic hydrocarbons in these OGIs are mainly three, four and five cyclic hydrocarbons. There

Fig.7-Fig.22 are the transmission, reflective micro photos and fluorescence spectra of OGIs of No.6, No.7, No.12, No.15, No.20, No.25, No.26 and No.34. No.6 and No.7 (Fig.7-Fig.10) are in Bai 95# oil well, the depths are 434.5 m and 435.0 m. No.12 (Fig.11-Fig.12) is in Hua 5# oil well, the depth is 1523-1524 m. No.15 (Fig. 13-Fig.14) and No.20(Fig.15-Fig.16)are in Jian 22# oil well, the depths are 396.6 m and 404.7 m. No. 25 (Fig.17-Fig.18) and No. 26 (Fig.19- Fig.20) are in Nong 29# oil well, the depths are 573-574 m and 595.6 m. No.34 (Fig.21-Fig.22)

No.6, No.15, No.20, No.25 and No.26 are two phases (liquid and gas) inclusions. No. 7 and No.34 are single phase (liquid) oil OGIs. No.12, No.21 and No.36 are in cements. The others

From the micro photos, one can see when the internal light source excites the aimed OGI, not only the OGI emits fluorescence, but also other OGIs and the cements around the mineral grains. The intensities of the cements are often stronger than that of the OGIs. Even the mineral grains also emit weak fluorescence. These are the reasons why the spectra look like each other even the OGIs with different phases and colours. For example, No. 6 and No.7 OGIs are in the same oil well, the former is liquid-gas inclusion, the later is liquid phase. The spectra of the two excited by internal light source (see Fig.8 a and Fi9.10 a) are similar. Obviously, the spectra were influenced by the background. According to the spectra excited by violet light of the microscope (see Fig. 8 a and Fig.10 a, solid line), the peak positions are at 400 nm. The main aromatic hydrocarbons in the OGIs may be three or four cyclic hydrocarbons. But the other aromatic hydrocarbons in the OGIs are difficult to determine because except for the main peak, the other parts of the spectra are almost flat. The peak positions excited by the internal

When the external light source (250 nm, 365 nm, 440 nm, 470 nm and 546 nm) excited the aimed OGI, the focal spot size is small. The cements were not illuminated, so the fluorescence spectra of the single OGIs weren't interfered by the spectra of the cements. If the OGIs are not dense in the grain, only one OGI is excited, one can obtain the spectra of

From Fig. 8 b, Fig.10 b, Fig.12 b, Fig.14 b, Fig.16 b, Fig.18 b, Fig.20 b and Fig.22 b, one can see that the spectra are in the range of UV-VIS at 250nm exciting. These results show that the experimental setup can measure the UV-VIS spectra of single OGIs. Combined the spectra excited by different external lights with different wavelengths, the characteristics of the

1. For most OGIs, when excited at 250 nm, there are three main peaks in the range of 400-470 nm, which are corresponding to three and four condensed aromatic rings, the palaeo oils are the medium Oils (Liu, K. Y., & Eadington P., 2005; Abbas *et al*., 2006). There are also two secondary peaks near to 535 nm and 610 nm, the former is the fluorescence of four or more condensed aromatic rings, the later comes from resin and asphaltene, the palaeo oils are medium and heavy Oils (Liu, K. Y., & Eadington P., 2005; Abbas *et al*., 2006). The spectra in the range of 280-400 nm belong to two or three condensed aromatic rings, the palaeo oils are light Oils (Liu, K. Y., & Eadington P., 2005; Abbas *et al*., 2006). So the OGIs of No.6, No.7, No.12, No.15, No.20, No.25 and No.34 are all filled with light, medium and heavy oils. And medium oils are the major part. The aromatic hydrocarbons in these OGIs are mainly three, four and five cyclic hydrocarbons. There

blue and green light are at 500-510 nm and 575 nm-600 nm respectively.

is in Fu 4# oil well, the depth is 422.8 m.

single OGI after subtracting the background.

OGIs are as following:

are all in the quartz grains.

Fig. 7. The transmission (a, e, f, g, h, i and j) and reflective (b, c and d) micro-photos of No.6 OGI (Bai 95#, 434.5 m) when excited by white light (a), internal light source (c:violet; d:blue and e:green) and external light source (e:250 nm; f:365 nm; g:440 nm; h:470 nm; and i:546 nm), j is a transmission image when the 550 nm light exciting and violet fluorescence filter cube was used.

Fig. 8. The fluorescence spectra of No.6 OGI excited by internal light source (in a, solid: violet, short dot: blue, and dash dot: green) and external light source (b-f).

Based on Common Inverted Microscope to Measure

violet light illuminating at the same time.

**0.0 5.0x106 1.0x107 1.5x107 2.0x107**

**8.0x104**

**1.2x105**

**Intensity (CPS)**

**1.6x105**

**2.0x105**

**Intensity (CPS)**

20m

UV-VIS Spectra of Single Oil-Gas Inclusions and Colour Analysis 57

b c d

a e

f g h i j

Fig. 11. The transmission (a, b, f, g, h, i and j) and reflective (b, c, d and e) micro-photos of No.12 OGI (Hua 5#, 1523.4 m) in cement when excited by white light (a and b), internal light source (b and c:violet; d:blue and e:green) and external light source (f:250 nm; g:365 nm; h:440 nm; i:470 nm and j:546 nm), b is a transmission and reflection image when white and

Fig. 12. The fluorescence spectra of No.12 OGI excited by internal light source (in a, solid:

**0.0**

**2.0x104**

**4.0x104**

**Intensity (CPS)**

**6.0x104**

**4.0x104 8.0x104 1.2x105 1.6x105 2.0x105**

**0.0 6.0x103 1.2x104 1.8x104 2.4x104**

**Intensity (CPS)**

**Intensity (CPS)**

**600 700 800 900 1000**

**Wavelength (nm)**

f Ex:546nm

**Wavelength (nm)**

**500 550 600 650 700 750 800**

**300 400 500 600 700 800 900 1000**

d Ex:440nm

b Ex:250nm

**Wavelength (nm)**

violet, long dot: blue, and short dot: green) and external light source (b-f).

**500 600 700 800 900**

e Ex:470nm

**400 450 500 550 600 650 700**

**Wavelength (nm)**

**400 500 600 700 800 900**

c Ex:365nm

**Wavelength (nm)**

a

**Wavelength (nm)**

**0.0 4.0x104 8.0x104 1.2x105 1.6x105 2.0x105**

**Intensity (CPS)**

Fig. 9. The transmission (a, e, f, g, h, i and j) and reflective (b, c and d) micro-photos of No.7 OGI (Bai 95#, 435.0 m) when excited by white light (a and b), internal light source (b:violet; c:blue and d:green) and external light source (e:250 nm; f:365 nm; g:440 nm; h:470 nm and i:546 nm), j is a transmission image when the 550 nm light exciting and violet fluorescence filter cube was used.

Fig. 10. The fluorescence spectra of No.7 OGI excited by internal light source (in a, solid: violet, long dot: blue, and short dot: green) and external light source (b-f).

Fig. 9. The transmission (a, e, f, g, h, i and j) and reflective (b, c and d) micro-photos of No.7 OGI (Bai 95#, 435.0 m) when excited by white light (a and b), internal light source (b:violet; c:blue and d:green) and external light source (e:250 nm; f:365 nm; g:440 nm; h:470 nm and i:546 nm), j is a transmission image when the 550 nm light exciting and violet fluorescence filter cube was used.

f g h i j

a b c d e

**300 400 500 600 700 800 900 1000**

**600 700 800 900 1000**

**Wavelength (nm)**

e Ex:546nm

**500 550 600 650 700 750 800 850**

**Wavelength (nm)**

b Ex:250nm

**Wavelength (nm)**

d Ex:440nm

**0.00 1.50x104 3.00x104 4.50x104 6.00x104**

> **1x105 2x105 3x105 4x105 5x105**

**Intensity (CPS)**

**0.0 3.0x104 6.0x104 9.0x104 1.2x105**

**Intensity (CPS)**

**Intensity (CPS)**

Fig. 10. The fluorescence spectra of No.7 OGI excited by internal light source (in a, solid:

violet, long dot: blue, and short dot: green) and external light source (b-f).

**400 500 600 700 800 900**

c Ex:365nm

**Wavelength (nm)**

**400 450 500 550 600 650 700**

**500 600 700 800 900**

**Wavelength (nm)**

**Wavelength (nm)**

e Ex:470nm

a

**0.0 5.0x106 1.0x107 1.5x107 2.0x107**

20m

**2x105 3x105 4x105 5x105**

**1x10<sup>5</sup> 2x10<sup>5</sup> 3x10<sup>5</sup> 4x10<sup>5</sup>**

**Intensity (CPS)**

**Intensity (CPS)**

Fig. 11. The transmission (a, b, f, g, h, i and j) and reflective (b, c, d and e) micro-photos of No.12 OGI (Hua 5#, 1523.4 m) in cement when excited by white light (a and b), internal light source (b and c:violet; d:blue and e:green) and external light source (f:250 nm; g:365 nm; h:440 nm; i:470 nm and j:546 nm), b is a transmission and reflection image when white and violet light illuminating at the same time.

Fig. 12. The fluorescence spectra of No.12 OGI excited by internal light source (in a, solid: violet, long dot: blue, and short dot: green) and external light source (b-f).

Based on Common Inverted Microscope to Measure

20m

UV-VIS Spectra of Single Oil-Gas Inclusions and Colour Analysis 59

a b c d e

f g h i j

Fig. 15. The transmission (a, b, f, g, h, i and j) and reflective (b, c, d and e) micro-photos of No.20 OGI (Jian 22#, 404.7 m) when excited by white light (a and b), internal light source (b and c:violet; d:blue and e:green) and external light source (f:250 nm; g:365 nm; h:440 nm; i:470 nm and j:546 nm), b is a transmission and reflection image when white and violet light illuminating.

Fig. 16. The fluorescence spectra of No.20 OGI excited by internal light source (in a, solid:

**600 700 800 900 1000**

**Wavelength (nm)**

f Ex:546nm

**500 550 600 650 700 750 800**

**Wavelength (nm)**

**300 400 500 600 700 800 900 1000**

b Ex:250nm

**Wavelength (nm)**

d Ex:440nm

**0 1x104 2x104 3x104 4x104 5x104**

**0**

**4.0x104 8.0x104 1.2x105 1.6x105**

**1x104**

**2x104**

**Intensity (CPS)**

**Intensity (CPS)**

**3x104**

**Intensity (CPS)**

violet, long dot: blue, and short dot: green) and external light source (b-f).

Ex:470nm

**500 600 700 800 900**

**400 450 500 550 600 650**

c Ex:365nm

**400 500 600 700 800 900 1000**

**Wavelength (nm)**

a

**Wavelength (nm)**

**Wavelength (nm)**

**4.0x104**

**8.0x104**

**Intensity (CPS)**

**Intensity (CPS)**

**Intensity (CPS)**

**1.2x105**

e

**1.6x105**

**8.0x104 1.2x105 1.6x105 2.0x105 2.4x105**

**0.0 4.0x106 8.0x106 1.2x107 1.6x107 2.0x107**

Fig. 13. The transmission (a, b, f, g, h, i and j) and reflective (b, c and d) micro-photos of No.15 OGI (Jian 22#, 396.6 m) when excited by white light (a and b), internal light source (b and c:violet, d:blue and e:green) and external light source (e:250 nm; f:365 nm; g:440 nm; h:470 nm and i:546 nm), b is a transmission and reflection image when white and violet light illuminating.

Fig. 14. The fluorescence spectra of No.15 OGI excited by internal light source (in a, solid: violet, long dot: blue, and short dot: green) and external light source (b-f).

Fig. 13. The transmission (a, b, f, g, h, i and j) and reflective (b, c and d) micro-photos of No.15 OGI (Jian 22#, 396.6 m) when excited by white light (a and b), internal light source (b and c:violet, d:blue and e:green) and external light source (e:250 nm; f:365 nm; g:440 nm; h:470 nm and i:546 nm), b is a transmission and reflection image when white and violet light illuminating.

a b c d

f g h j i

Fig. 14. The fluorescence spectra of No.15 OGI excited by internal light source (in a, solid:

**600 700 800 900 1000**

**Wavelength (nm)**

f Ex:546nm

**500 550 600 650 700 750 800**

d Ex:440nm

**300 400 500 600 700 800 900 1000**

b Ex:250nm

e

**Wavelength (nm)**

**Wavelength (nm)**

**0.0 3.0x104 6.0x104 9.0x104 1.2x105**

**1x105 2x105 3x105 4x105**

**0 1x104 2x104 3x104 4x104 5x104**

**Intensity (CPS)**

**Intensity (CPS)**

**Intensity (CPS)**

violet, long dot: blue, and short dot: green) and external light source (b-f).

**500 600 700 800 900**

**Wavelength (nm)**

e Ex:470nm

**400 450 500 550 600 650**

**400 500 600 700 800 900**

a

**Wavelength (nm)**

**Wavelength (nm)**

c Ex:365nm

**0 1x105 2x105 3x105 4x105**

**2x105**

**3x105**

**Intensity (CPS)**

**Intensity (CPS)**

**4x105**

**0.0 4.0x106 8.0x106 1.2x107 1.6x107 2.0x107**

20m

**Intensity (CPS)**

Fig. 15. The transmission (a, b, f, g, h, i and j) and reflective (b, c, d and e) micro-photos of No.20 OGI (Jian 22#, 404.7 m) when excited by white light (a and b), internal light source (b and c:violet; d:blue and e:green) and external light source (f:250 nm; g:365 nm; h:440 nm; i:470 nm and j:546 nm), b is a transmission and reflection image when white and violet light illuminating.

Fig. 16. The fluorescence spectra of No.20 OGI excited by internal light source (in a, solid: violet, long dot: blue, and short dot: green) and external light source (b-f).

Based on Common Inverted Microscope to Measure

UV-VIS Spectra of Single Oil-Gas Inclusions and Colour Analysis 61

a b c d e

f g h i j

Fig. 19. The transmission (a, b, f, g, h, i and j) and reflective (b, c, d and e) micro photos of No.26 OGI (Nong 29#, 595.9 m) when excited by white light (a and b), internal light source(b and c: violet; d: blue and e: green) and external light source (f:250 nm; g:365 nm; h:440 nm; i:470 nm and j:546 nm), b is a transmission and reflection image when white and violet light illuminating.

> **400 0**

**6.0x104 1.2x105 1.8x105 2.4x105 3.0x105**

**0.00 1.50x104 3.00x104 4.50x104 6.00x104 7.50x104**

**Intensity (CPS)**

**Intensity (CPS)**

**Intensity (CPS)**

**800 0**

**1200 0**

**<sup>300</sup> <sup>400</sup> <sup>500</sup> <sup>600</sup> <sup>700</sup> <sup>800</sup> <sup>900</sup> <sup>1000</sup> <sup>0</sup>**

b Ex:250nm

**Wavelength (nm)**

d Ex:440nm

**500 550 600 650 700 750 800**

**Wavelength (nm)**

f Ex:546nm

**600 700 800 900 1000**

**Wavelength (nm)**

Fig. 20. The fluorescence spectra of No.25 OGI excited by internal light source (in a, solid:

violet, long dot: blue, and short dot: green) and external light source (b-f).

e

**400 450 500 550 600 650 700**

**500 550 600 650 700 750 800**

**Wavelength (nm)**

**Wavelength (nm)**

c Ex:365nm

**400 500 600 700 800 900 1000**

a

**Wavelength (nm)**

**1.2x105 1.8x105 2.4x105 3.0x105**

**4.0x104 8.0x104 1.2x105 1.6x105 2.0x105**

**0.0 5.0x106 1.0x107 1.5x107 2.0x107**

20m

20m

**Intensity (CPS)**

**Intensity (CPS)**

**Intensity (CPS)**

Ex:470nm

Fig. 17. The transmission (a, b, f, g, h, i and j) and reflective (b, c, d and e) micro-photos of No.25 OGI (Nong 29#, 573-574 m) when excited by white light (a and b), internal light source (b and c:violet; d:blue and e:green) and external light source (f:250 nm; g:365 nm; h:440 nm; i:470 nm and j:546 nm). b is a transmission and reflection image when white and violet light illuminating.

Fig. 18. The fluorescence spectra of No.25 OGI excited by internal light source (in a, solid: violet, long dot: blue, and short dot: green) and external light source (b-f).

(a)

Fig. 17. The transmission (a, b, f, g, h, i and j) and reflective (b, c, d and e) micro-photos of No.25 OGI (Nong 29#, 573-574 m) when excited by white light (a and b), internal light source (b and c:violet; d:blue and e:green) and external light source (f:250 nm; g:365 nm; h:440 nm; i:470 nm and j:546 nm). b is a transmission and reflection image when white and

f g h i j

a b c d e

b Ex:250nm

d Ex:440nm

**300 400 500 600 700 800 900 1000**

**Wavelength (nm)**

**600 700 800 900 1000**

**500 600 700 800**

**Wavelength (nm)**

**Wavelength (nm)**

Ex:546nm

Fig. 18. The fluorescence spectra of No.25 OGI excited by internal light source (in a, solid:

**5.0x103 1.0x104 1.5x104 2.0x104**

**0.00 1.50x104 3.00x104 4.50x104 6.00x104**

**3 8.0x10<sup>3</sup> 1.2x10 <sup>4</sup> 1.6x10 <sup>4</sup>**

**0.0 4.0x10**

**Intensity (CPS)**

**Intensity (CPS)**

f

**Intensity (CPS)**

violet, long dot: blue, and short dot: green) and external light source (b-f).

**500 600 700 800 900**

**e** Ex:470n

**400 450 500 550 600 650 700**

**Wavelength (nm)**

**400 500 600 700 800 900**

**Wavelength (nm)**

c Ex:365nm

a

**Wavelength (nm)**

**0.0 2.0x104 4.0x104 6.0x104 8.0x104**

**0.0 3.0x104 6.0x104 9.0x104 1.2x105**

**Intensity (CPS)**

**Intensity (CPS)**

violet light illuminating.

20m

**Intensity (CPS)**

**0.0 5.0x106 1.0x107 1.5x107 2.0x107**

Fig. 19. The transmission (a, b, f, g, h, i and j) and reflective (b, c, d and e) micro photos of No.26 OGI (Nong 29#, 595.9 m) when excited by white light (a and b), internal light source(b and c: violet; d: blue and e: green) and external light source (f:250 nm; g:365 nm; h:440 nm; i:470 nm and j:546 nm), b is a transmission and reflection image when white and violet light illuminating.

Fig. 20. The fluorescence spectra of No.25 OGI excited by internal light source (in a, solid: violet, long dot: blue, and short dot: green) and external light source (b-f).

Based on Common Inverted Microscope to Measure

UV-VIS Spectra of Single Oil-Gas Inclusions and Colour Analysis 63

2. In the same oil well, the FWHMs of liquid oil inclusions are wider than that of the liquid-gas OGIs at exciting 250 nm. For example, the FWHMs of No. 6 (liquid and gas) and No.7 (liquid) inclusions are 175 nm and 185 nm respectively. But in different oil wells, this result is not always right. For example, the FWHMs of No.12 and No. 15 (liquids and gas) are 189 nm and 190 nm, which are wider than that of No.7. So one should be cautious to say that the palaeo-oils in the liquid oil inclusions are more

3. For all the 38 testing OGIs, when exciting at 365 nm (external light source), the spectra can be classified into two types. The first type has an obvious "bump" following the main peak. The wavelength range is in 425-650 nm. This range is relative to the medium and heavy hydrocarbons, which indicate that these inclusions including in more medium and heavy hydrocarbons (see Fig.10 c, Fig.12 c, Fig.14 c). The second type has not an obvious "bump", which shows that these OGIs (see Fig. 8 c, Fig. 16 c, Fig.18 c, Fig. 20 c and Fig.22 c) don't include in enough medium and heavy hydrocarbons as in the first type. The palaeo-oil maturities of the second type

4. When exciting at 440 nm, except for the main peak near to 478 nm, there is another peak

5. When exciting at 470 nm, except for the main peak near to 502 nm, there is another peak

6. The results of (4) and (5) show that a kind of characteristic matter exist in the OGIs. 440 nm and 470 nm are the effective exciting wavelengths for this kind of matter. This needs

7. The characteristics of the spectra excited at 546 nm are not obvious. The reason is the absorption bands for the most hydrocarbons are in the range of UV and violet.

For simplicity and intuition, the colours of the OGIs under excited at 365 nm are often to be used to judge the maturities of palaeo-oils in the inclusions. But the judgment by eyes are subjective. The results will be influenced by psychology, physiology or environment. So it is

) is the power spectrum

at 657 nm for all 38 OGIs . No.25 OGI has another peak at 672 nm.

may be two oil sources charging these strata at the same time.

heavier than that of in the liquid and gas OGIs.

to make further analysis to determine what it is.

need to calculate the chromaticity coordinates of the OGIs.

The calculation theory (Xu & Su,2004 ) is as following. Assume P(

distribution of an OGI, the tristimulus values are given by equation (5),

are higher than the first one.

at 706 nm for all 38 OGIs.

**3.6 The colours of single OGIs** 

are still some one and two cyclic aromatic hydrocarbons and heavy hydrocarbons. The spectrum (Fig.20 b) of No. 26 OGI is different from others. There is one main peak at 400 nm. The full width of half maximum (FWHM) is obviously less than the others, which shows the main hydrocarbons in this OGI are light-medium oils. The secondary peaks in 575-605 nm show there are also some heavy oils in this OGI. The aromatic hydrocarbons in No. 26 OGI are mainly three and four cyclic hydrocarbons. There are still some two cyclic aromatic hydrocarbons and heavy hydrocarbons. For all 38 testing OGIs, the spectra of 90% are similar to No.6 when exciting at 250 nm.This result shows that the palaeo-oils in the OGIs may come from two maternal sources with different maturities, one is high, another is low. So in the later time of quartz overgrowth, there

Fig. 21. The transmission (a, e, f, g, h, i and j) and reflective (b, c and d) micro-photos of No.34 OGI (Fu4# 422.8 m) when excited by white light (a), internal light source (b: iolet; c:blue and d:green) and external light source (e:250 nm; f:365 nm; g:440 nm; h:470 nm and i:546nm), j is a transmission image when the 550 nm light exciting and violet fluorescence filter cube was used.

Fig. 22. The fluorescence spectra of No.34 OGI excited by internal light source (in a, solid: violet, long dot: blue, and short dot: green) and external light source (b-f).

a b c d e

Fig. 21. The transmission (a, e, f, g, h, i and j) and reflective (b, c and d) micro-photos of No.34 OGI (Fu4# 422.8 m) when excited by white light (a), internal light source (b: iolet; c:blue and d:green) and external light source (e:250 nm; f:365 nm; g:440 nm; h:470 nm and i:546nm), j is a transmission image when the 550 nm light exciting and violet fluorescence filter cube was used.

f g h i j

**300 400 500 600 700 800 900 1000**

b Ex:250nm

**500 550 600 650 700 750 800**

**600 700 800 900 1000**

**Wavelength (nm)**

**Wavelength (nm)**

f Ex:546nm

**Wavelength (nm)**

Ex:440nm

**0.0 5.0x103 1.0x104 1.5x104 2.0x104**

**0.0 2.0x104 4.0x104 6.0x104 8.0x104**

**0.00 7.50x103 1.50x104 2.25x104 3.00x104** d

**Intensity (CPS)**

**Intensity (CPS)**

**Intensity (CPS)**

Fig. 22. The fluorescence spectra of No.34 OGI excited by internal light source (in a, solid:

violet, long dot: blue, and short dot: green) and external light source (b-f).

c Ex:365nm

a

**400 450 500 550 600 650**

**Wavelength (nm)**

e

Ex:470nm

**500 600 700 800 900**

**Wavelength (nm)**

**400 500 600 700 800 900 1000**

**Wavelength (nm)**

**0.0 5.0x106 1.0x107 1.5x107 2.0x107**

**4.0x10<sup>4</sup> 6.0x10<sup>4</sup> 8.0x10<sup>4</sup> 1.0x10<sup>5</sup> 1.2x10<sup>5</sup>**

**0.00 2.50x104 5.00x104 7.50x104 1.00x105**

**Intensity (CPS)**

20m

**Intensity (CPS)**

**Intensity (CPS)**

are still some one and two cyclic aromatic hydrocarbons and heavy hydrocarbons. The spectrum (Fig.20 b) of No. 26 OGI is different from others. There is one main peak at 400 nm. The full width of half maximum (FWHM) is obviously less than the others, which shows the main hydrocarbons in this OGI are light-medium oils. The secondary peaks in 575-605 nm show there are also some heavy oils in this OGI. The aromatic hydrocarbons in No. 26 OGI are mainly three and four cyclic hydrocarbons. There are still some two cyclic aromatic hydrocarbons and heavy hydrocarbons. For all 38 testing OGIs, the spectra of 90% are similar to No.6 when exciting at 250 nm.This result shows that the palaeo-oils in the OGIs may come from two maternal sources with different maturities, one is high, another is low. So in the later time of quartz overgrowth, there may be two oil sources charging these strata at the same time.


#### **3.6 The colours of single OGIs**

For simplicity and intuition, the colours of the OGIs under excited at 365 nm are often to be used to judge the maturities of palaeo-oils in the inclusions. But the judgment by eyes are subjective. The results will be influenced by psychology, physiology or environment. So it is need to calculate the chromaticity coordinates of the OGIs.

The calculation theory (Xu & Su,2004 ) is as following. Assume P() is the power spectrum distribution of an OGI, the tristimulus values are given by equation (5),

Based on Common Inverted Microscope to Measure

c Jian 22#

a Bai 95#

**0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8**

**X (Red)**

**0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8**

**X (red)**

**0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8**

**X (Red)**

e Fu 4#

68

 B G H

 B C D E

D(C) is in the cement.

49

49

**0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8**

**0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9**

49

470 48

**Y (Green)**

**Y (Green)**

470 48

520 530

540 55 56 57 58 59 60 61 68

**0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9**

**Y (Green)**

47 48

I

<sup>52</sup> 530 540

52 53

54 550 56 57 580 59 60

II

55 560 57 580 590 600 61 68

**0.9**<sup>A</sup>

decreased greatly.

**4. Conclusion** 

UV-VIS Spectra of Single Oil-Gas Inclusions and Colour Analysis 65

49

49

**0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9**

**Y (Green)**

47 48

**0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9**

**Y (Green)**

470 48

520 53

<sup>530</sup> <sup>540</sup> 55 560 57 58 590 60 61 68

I

54 55 560 57 580 59 60 61 68

II

b Hua 5#

d Nong 29#

**0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8**

**X (Red)**

**0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8**

**X (Red)**

Fig. 23. The chromaticity coordinates of OGIs in the five oil wells excited by external and internal light sources, in which, a and b are Bai 95# , Hua 5# respectively, external (I) and internal (II); c―Jian 22#, external (A, B) and internal (G, H), B(G) is in the cement; d―Nong 29#, external (triangle) and internal (circle); e―Fu 4#, external (D, E) and internal (B, C),

The inclusion samples came from the core drilling samples of five oil wells, Jinlin Oil Filed, Songliao Basin, northeast China(upper Cretaceous Period). The microscope testing shows: 1. The experimental setup based on common inverted microscope, no need any rebuilding to the microscope and can measure the UV-VIS spectra of single OGIs. The cost was

$$\begin{aligned} X &= \int\_{380}^{780} P(\mathcal{A}) \overline{x}(\mathcal{A}) d\mathcal{A} \\ Y &= \int\_{380}^{780} P(\mathcal{A}) \overline{y}(\mathcal{A}) d\mathcal{A} \\ X &= \int\_{380}^{780} P(\mathcal{A}) \overline{z}(\mathcal{A}) d\mathcal{A} \end{aligned} \tag{5}$$

In which, *x*( ) , *y*( ) and *z*( ) are the tristimulus values of standard light source. One can get them by looking up the CIE 1931 system. The chromaticity coordinates of an OGI are

$$\begin{aligned} x &= \frac{X}{X+Y+Z} \\ y &= \frac{Y}{X+Y+Z} \\ z &= \frac{Z}{X+Y+Z} \end{aligned} \tag{6}$$

In which, *xyz* 1 (*x*:red; *y*:green and *z*:blue). According to above theory, a computing program based on Matlab was established to calculate the chromaticity coordinates of OGIs.

The chromaticity coordinates show that for all OGIs, the chromaticity coordinates excited by internal light source (365 nm) are larger than that of external light source (365 nm). The main reason is that not only the OGI emitting fluorescence, but also the cements and the mineral grain itself when excited by internal light source, which influenced the spectra of the under studied OGI. The chromaticity coordinates deflect to light blue white or yellow white, and relative centralize for different OGIs. When the external light source excites the OGIs, the focal spot is small. The cements don't emit fluorescence and not influence the spectra of single OGIs. The fluorescence of the mineral grain around the OGI has been subtracted as the background. Such spectra are near to the real spectra of single OGIs.

For the OGIs in the cements in Jian 22# and Fu 4# oil wells, the chromaticity coordinates of these OGIs are less than the OGIs in the mineral grains whatever external or internal light source exciting (Fig.23 c and Fig.23 e), which indicates the hydrocarbons in the cements are lighter than those in mineral grains. The formation time of the OGIs in the cements may be later than that in the mineral grains. But for OGI in cements of Nong 29# oil well, the chromaticity coordinates basic overlap with the OGIs in the mineral grains. The OGIs in the cements and in the mineral grains may be formed at the same geology age.

The chromaticity coordinates are dispersive for single OGIs under external light source excited. This result shows that it is easier to distinguish OGIs with external light source exciting. For OGIs in Bai 95# oil well, the chromaticity coordinates are in the range of light blue white and light yellow white. For OGIs in Hua 5# oil well, the chromaticity coordinates are in the range of blue green and light yellow white. For OGIs in Jian 22# oil well, the chromaticity coordinates are in the range of green blue and light yellow white. For OGIs in Nong 29# oil well, the chromaticity coordinates are in the range of light blue white and light green blue. For OGIs in Fu 4# oil well, the chromaticity coordinates are in the range of light green blue, light blue white and white. Above results indicate that in the later time of quartz overgrowth, there may be two maternal source to charge these strata, one with high maturity, another with low maturity.

()()

 

  (5)

(6)

*X*

*XYZ Y <sup>y</sup> XYZ Z*

*XYZ*

In which, *xyz* 1 (*x*:red; *y*:green and *z*:blue). According to above theory, a computing program based on Matlab was established to calculate the chromaticity coordinates of OGIs. The chromaticity coordinates show that for all OGIs, the chromaticity coordinates excited by internal light source (365 nm) are larger than that of external light source (365 nm). The main reason is that not only the OGI emitting fluorescence, but also the cements and the mineral grain itself when excited by internal light source, which influenced the spectra of the under studied OGI. The chromaticity coordinates deflect to light blue white or yellow white, and relative centralize for different OGIs. When the external light source excites the OGIs, the focal spot is small. The cements don't emit fluorescence and not influence the spectra of single OGIs. The fluorescence of the mineral grain around the OGI has been subtracted as

For the OGIs in the cements in Jian 22# and Fu 4# oil wells, the chromaticity coordinates of these OGIs are less than the OGIs in the mineral grains whatever external or internal light source exciting (Fig.23 c and Fig.23 e), which indicates the hydrocarbons in the cements are lighter than those in mineral grains. The formation time of the OGIs in the cements may be later than that in the mineral grains. But for OGI in cements of Nong 29# oil well, the chromaticity coordinates basic overlap with the OGIs in the mineral grains. The OGIs in the

The chromaticity coordinates are dispersive for single OGIs under external light source excited. This result shows that it is easier to distinguish OGIs with external light source exciting. For OGIs in Bai 95# oil well, the chromaticity coordinates are in the range of light blue white and light yellow white. For OGIs in Hua 5# oil well, the chromaticity coordinates are in the range of blue green and light yellow white. For OGIs in Jian 22# oil well, the chromaticity coordinates are in the range of green blue and light yellow white. For OGIs in Nong 29# oil well, the chromaticity coordinates are in the range of light blue white and light green blue. For OGIs in Fu 4# oil well, the chromaticity coordinates are in the range of light green blue, light blue white and white. Above results indicate that in the later time of quartz overgrowth, there may be two maternal source to charge these strata, one with high

()()

( )( ) .

 

.

are the tristimulus values of standard light source. One can

*x*

*z*

the background. Such spectra are near to the real spectra of single OGIs.

cements and in the mineral grains may be formed at the same geology age.

maturity, another with low maturity.

In which, *x*( )

 , *y*( ) 

 and *z*( ) 

*X Pxd*

*Y P y d*

*X Pzd*

get them by looking up the CIE 1931 system. The chromaticity coordinates of an OGI are

Fig. 23. The chromaticity coordinates of OGIs in the five oil wells excited by external and internal light sources, in which, a and b are Bai 95# , Hua 5# respectively, external (I) and internal (II); c―Jian 22#, external (A, B) and internal (G, H), B(G) is in the cement; d―Nong 29#, external (triangle) and internal (circle); e―Fu 4#, external (D, E) and internal (B, C), D(C) is in the cement.

#### **4. Conclusion**

The inclusion samples came from the core drilling samples of five oil wells, Jinlin Oil Filed, Songliao Basin, northeast China(upper Cretaceous Period). The microscope testing shows:

1. The experimental setup based on common inverted microscope, no need any rebuilding to the microscope and can measure the UV-VIS spectra of single OGIs. The cost was decreased greatly.

Based on Common Inverted Microscope to Measure

2003), pp. 29-43, ISSN 9153-7061

No. 7, (July, 2005), pp. 1023-1036, ISSN 0146-6380

(April, 2009), pp. 517-526, ISSN 0883-2927

5, (n. d. 1995), pp. 279-284, ISSN 0030-3092

1458-1465, ISSN 0304-3894

2009), pp. 12-16, ISSN 1729-8806

Vol.22, No.3, (July, 2003), pp. 245-250, ISSN 1007-2802

*Geology*, Vol.33, No.1, (Jan. , 1997), pp. 61-71, ISSN 0166-5162

0213683X

3495

UV-VIS Spectra of Single Oil-Gas Inclusions and Colour Analysis 67

Caja, M. A., Permanyer, A., Kihle, J., Munz, I. A. & Johansen, H. (2009). Fluorescence

Damaskinos, S., Dixon, A. E., Ellis, K. A., & Diehl-Jones, W. L. (1995). Imaging biological

Kevin, B., Liesbeth, P., Niek, N. S., Stefaan, C. D. S. & Joseph, D. (2003). Three-Dimensional

Kihle, J. (1995). Adaptation of fluorescence excitation-emission micro-spectroscopy for

Liu, D. H., Xiao, X. M., Mi, J. K., Li, X. Q., Shen, J. K., Song, Z. G. & Peng, P. A. (2003).

Liu, K. Y. & Eadington, P. (2005). Quantitative fluorescence techniques for detecting residual

Liu, Y. R., Lu, X. B. & He, M. C. (2003). Progress on Application of Fluid Inclusion in

Lo, H. B., Wilkins, R. W. T., Ellacott, M. V. & Buckingham, C. P. (1997). Assessing the

Meng, D. W., Wu, X.L., Fan, X.Y., Meng, X. & Zheng, J.P. & Mason, R. (2009). Submicron-

Sanches S. , Leitão C., Penetra A., Cardoso V.V. , Ferreira E., Benoliel M.J., Barreto Crespo

Siegwart, R. (2001). Indirect Manipulation of a Sphere on a Flat Disk Using Force

Sliney D. H. (1995). Risk assessment and laser safety. *Optics & Laser Technology*, Vol. 27, No.

Vol.101, No.1, (April, 2009), pp.16, ISSN 0375-6742

Vol.26, No.6, (n.d.,1995), pp. 493-502, ISSN 0968-4328

23, No.11-12, (Nov.-Dec. 1995), pp. 1029-1042,ISSN 0146-6380

*Petroleum Geology*, Vol.16, No.2, (Mar., 1999), pp. 97-110, ISSN 9153-7061 Caja, M. A., Permanyer, A., Munz, I. A. & Johansen, H. (2007). Preliminary data on oil and

composition and physical properties of petroleum in fluid inclusions. *Marine and* 

aqueous fluid inclusions of the fracture-fill in the Corones and Armancies Fms, Eocene, SEPyrenees. *GEOGACETA*, Vol.101, (n. d. 2007), pp. 127-130, ISSN

quantification of oil fluid inclusions and oil shows:Implications for oil migration (Armancies Fm, South-eastern Pyrenees, Spain). *Journal of Geochemica Exploration*,

specimens with the confocal scanning laser microscope/macroscope. *Micron*,

Fluorescence Recovery after Photobleaching with the Confocal Scanning Laser Microscope. *Biophysical Journal*, Vol.85, No.4, (Oct., 2003), pp. 2240-2252, ISSN 0006-

characterization of single hydrocarbon fluid inclusions. *Organic Geochemistry*, Vol.

Determination of trapping pressure and temperature of petroleum inclusions using PVT simulation software—a case study of Lower Ordovician carbonates from the Lunnan Low Uplift, Tarim Basin. *Marine and Petroleum Geology*, Vol.20, No.1, (Jan.,

oils and reconstructing hydrocarbon charge history. *Organic Geochemistry*, Vol. 36,

Petroleum Exploration. *Geology Bullet in of Mineralogy, Petrology and Geoch emistry*,

maturity of coals and other rocks from North America using the fluorescence alteration of multiple macerals (FAMM) technique. *International Journal of Coal* 

sized fluid inclusions and distribution of hydrous components in jadeite, quartz and symplectite-forming minerals from UHP jadeite–quartzite in the Dabie Mountains, China: TEM and FTIR investigation. *Applied Geochemistry*, Vol.24, No.4,

M.T., Pereira V.J. (2011). Direct photolysis of polycyclic aromatic hydrocarbons in drinking water sources. *Journal of Hazardous Materials*, Vol. 192, (Jun. 2011), pp.

Information. *International Journal of Advanced Robotic Systems,* Vol.6, No.4, (Dec.


This optical system can measure UV-VIS spectra of single OGIs. For further decreasing the focal spot, one can try a 74x RMO or decrease the fiber diameter and increase the coupling efficient between the fibre and the spectrometer.

The UV-VIS spectra should be combined with GC-MS analysis to find the characteristic matter in these OGIs.

This setup is promising in measuring the fluorescence spectra of micro areas, such as bioinclusions, micro fractures in mineral slides, special mineral grains and so on.

#### **5. Acknowledgment**

This work was supported by Chinese National Program for High Technology Research and Development (national 863 plan, granted No.2006AA09Z336) and National Natural Science Foundation of China (granted No. 41172110). The author is appreciated for Jilin Oil Field to supply sand rocks. The author greatly thanks professor Zhang, J. L., Dr. Tang., M.M., Ren, W. W. and Yang, Y. for help in the experiment.

#### **6. References**

Abbas, O., Rébufa, C., Dupuy, N., Permanyer, A., Kister, J. & Azevedo, D.A. (2006). Application of chemometric methods to synchronous UV fluorescence spectra of petroleum oils. *Fuel*, Vol. 85, No.17-18, (Dec. 2006), pp. 2653-2661, ISSN 0016-2361

Aplin, A.c., Macleod, G., Larter, S.R., Pedersen, K.S., Sorensen H. & Booth, T. (1999). Combined use of Confocal Laser Scanning Microscopyand PVT simulation for estimating the

2. By the spectra excited at 250 nm, the main aromatic hydrocarbons in an OGI can be qualitative determined. Almost all OGIs contain light, medium and heavy petroleum. The main aromatic hydrocarbons are three, four and five cyclic hydrocarbons. There are also some heavy hydrocarbons and non-hydrocarbons in these OGIs. In the same oil well, the FWHMs of liquid oil inclusions are wider than that of the liquid-gas OGIs at

3. The peak positions of the spectra excited at 365 nm are in the range of 395-400 nm, the main aromatic hydrocarbons are three and four cyclic aromatics. The "bump" (425- 650nm) following the main peak shows there are more medium and heavy hydrocarbons in the OGIs. The OGIs without the "bump" don't have many medium and heavy hydrocarbons. The palaeo-oil maturities of the second type are higher than

4. The special peaks near to 657 nm (440 nm exciting) and 706 nm (470 nm exciting) for all OGIs indicate there may be same special matters in these OGIs. This need further

5. The chromaticity coordinates are more dispersive for single OGIs under external light source excited than internal light source. The calculating colours excited at external light source are more objective than judging by eyes under exciting at internal light source. But one should note the different spectra may have same colour. So the spectra

6. The colours, the spectra excited at 250 nm and chromaticity coordinates show that there may be two maternal sources to charge these strata in the later time of quartz

This optical system can measure UV-VIS spectra of single OGIs. For further decreasing the focal spot, one can try a 74x RMO or decrease the fiber diameter and increase the coupling

The UV-VIS spectra should be combined with GC-MS analysis to find the characteristic

This setup is promising in measuring the fluorescence spectra of micro areas, such as bio-

This work was supported by Chinese National Program for High Technology Research and Development (national 863 plan, granted No.2006AA09Z336) and National Natural Science Foundation of China (granted No. 41172110). The author is appreciated for Jilin Oil Field to supply sand rocks. The author greatly thanks professor Zhang, J. L., Dr. Tang., M.M., Ren,

Abbas, O., Rébufa, C., Dupuy, N., Permanyer, A., Kister, J. & Azevedo, D.A. (2006).

Application of chemometric methods to synchronous UV fluorescence spectra of petroleum oils. *Fuel*, Vol. 85, No.17-18, (Dec. 2006), pp. 2653-2661, ISSN 0016-2361 Aplin, A.c., Macleod, G., Larter, S.R., Pedersen, K.S., Sorensen H. & Booth, T. (1999). Combined

use of Confocal Laser Scanning Microscopyand PVT simulation for estimating the

inclusions, micro fractures in mineral slides, special mineral grains and so on.

exciting 250 nm. But in different oil wells, this result is not always right.

analysis combined with GC-MS and palaeo-biology.

overgrowth. One is high maturity, another is low maturity.

efficient between the fibre and the spectrometer.

W. W. and Yang, Y. for help in the experiment.

the first one.

are important.

matter in these OGIs.

**5. Acknowledgment** 

**6. References** 

composition and physical properties of petroleum in fluid inclusions. *Marine and Petroleum Geology*, Vol.16, No.2, (Mar., 1999), pp. 97-110, ISSN 9153-7061


**3** 

*Japan* 

**Challenging Evaluation of the** 

*2Advanced Technology Institute of Northern Resources,* 

and Masayoshi Kobayashi2

*1Hokkaido Industrial Technology Center, 379 Kikyo-cho, Hakodate, Hokkaido* 

*8-6, Tonden 2-Jo, 2-Chome, Sapporo, Hokkaido* 

**Hybrid Technique of Chemical Engineering –** 

**Proton NMR Technique for Food Engineering** 

Yasuyuki Konishi1,

In the field of food engineering, chemical engineering and bioengineering have commonly used the data obtained from physicochemical techniques. In particular, water species are strongly related to food quality, as has been demonstrated in the International Symposium on Properties of Water (Eds. Rockland and Stewart, 1981). The water species retained in foods, as is well known, expose their multifunctional nature by dynamically responding to environmental conditions such as temperature, pressure, dehydration processes, water content, etc. For scientific analysis of the multifunctional water species retained in foods, of particular difficulty is the influence of nonlinear dynamic change on food quality. To quantitatively evaluate the nature of the water species, a large number of researchers have demonstrated the diversity of their biological and physicochemical nature using various parameters. The water activity (aw) has, for example, commonly been used as a parameter to evaluate the deterioration of foods (Fennema, 1976). The effective water diffusivity parameter (*De*) in foods has also frequently been employed to evaluate the dehydration rate (Jason, 1958). From the chemical engineering point of view, this is a typical procedure in evaluating water mobility in the food tissue matrix. Conversely, for direct identification of the water species at the molecular level, magnetic resonance techniques have been applied to evaluate food quality, as demonstrated in the International Conferences on Applications

In our previous papers (Konishi *et al.*, 2001, 2003), using separately chemical engineering and proton NMR techniques, water species retained in fish paste sausage and squid were roughly distinguished into two different species as a function of water content. The two species were water species A1, to be recognized as the higher water content at *W*0 >120%-d.b. accompanied by higher water diffusivity, *De,* and water species A2, *W*0 <120%-d.b., with lower *De*. This delineation, however, was unclear because of a vagueness of the boundary

of Magnetic Resonance in Food Science (Eds. Belton *et al.*, 2003).

**1. Introduction** 

 

Corresponding Author


### **Challenging Evaluation of the Hybrid Technique of Chemical Engineering – Proton NMR Technique for Food Engineering**

Yasuyuki Konishi1, and Masayoshi Kobayashi2

*1Hokkaido Industrial Technology Center, 379 Kikyo-cho, Hakodate, Hokkaido 2Advanced Technology Institute of Northern Resources, 8-6, Tonden 2-Jo, 2-Chome, Sapporo, Hokkaido Japan* 

#### **1. Introduction**

68 Advances in Chemical Engineering

Stasiuk, L. D. & Snowdon, L. R. (1997). Fluorescence micro-spectrometry of synthetic and

Stasiuk, L.D. (1999). Confocal laser scanning fluorescence microscopy of Botryococcus

Stasiuk, L.D. & Sanei, H. (2001). Characterization of diatom-derived lipids and chlorophyll

Thiéry, R., Pironon, J., Walgenwitz, F. & Montel F. (2002). Individual characterization of

Veld, H., Wilkins, R. W. T., Xiao, X. M. & Buckingham, C. P. (1997). A fluorescence alteration

Wilkins, R. W. T., Wilmshurst, J. R., Russell, N. J., Hladky, G., Ellacott, M.V. & Buckingham

Xiao, X.M., Wilkins R. W. T., Liu Z.F. & Fu, J.M. (1998). A preliminary investigation of the

Xiao, X. M., Wilkins R. W. T., Liu, D. H. & Shen, J.Q. (2002). Laser-induced fluorescence

Yang, A.l., Zhang, J.L., Ren. W.W. & Tang. M.M. (2009a). Micro-spectroscopy system based

Yang A.L.,Tang, M.M.,Ren. W.W.,Yang, Y. & Zhang, J.L. (2011). Investigation of the

*Journal of Coal Geology*, Vol.5, No.2, (July, 2002), pp. 129-141, ISSN 0166-5162 Xu, X.R. & Su, Z. M. (Oct., 2004). Luminescence Theory & Luminescence Materials,

*Proceedings of SPIE 7283*, ISBN 0277-786X, Chengdu, China, Nov., 2008 Yang, A.l., Ren. W.W., Zhang, J.L., & Tang. M.M. (2009b). A micro-spectroscopy system to

Chemical Industry, ISBN: 7502554106, 9787502554101, Beijing

0883-2927

1026, ISSN 0146-6380

2001), pp. 1417-1428, ISSN 0146-6380

2002), pp. 847-859, ISSN 9153-7061

No.1, (Jan. , 1995), pp. 191-209, ISSN 0146-6380

No.10, (Oct., 2000), pp.1041-1052, ISSN 0146-6380

ISBN 0277-786X, Beijing, China, April, 2009

pp. 247-255, ISSN 0146-6380

natural hydrocarbon fluid inclusions:crude oil chemistry, density and application to petroleum. *Applied Geochemistry*, Vol.12, No.3, (May, 1997), pp. 229-241, ISSN

alginite from boghead oil shale, Boltysk, Ukraine: selective preservation of various micro-algal components. *Organic Geochemistry*, Vol.30, No.8, (Aug.,1999), pp. 1021-

within Holocene laminites, Saanich Inlet, British Columbia, using conventional and laser scanning fluorescence microscopy. *Organic Geochemistry*, Vol.32, No. 12, (Dec.,

petroleum fluid inclusions (composition and P–T trapping conditions) by microthermometry and confocal laser scanning microscopy: inferences from applied thermodynamics of oils. *Marine and Petroleum Geology*, Vol.19, No.7, (Aug.,

of multiple macerals (FAMM) study of Netherlands coals with "normal" and "deviating" vitrinite reflectance. *Organic Geochemistry*, Vol.26, No.3-4, (Feb., 1997),

C. (1992). Fluorescence alteration and the suppression of vitrinite reflectance. *Organic Geochemistry,* Vol.18, No.5, (Sept., 1992), pp. 629-640, ISSN 0146-6380 Wilkins, R. W. T., Wilmshurst J.R., Hladky, G., Ellacott, M. V. & Buckingham, C. P. (1995).

Should fluorescence alteration replace vitrinite reflectance as a major tool for thermal maturity determination in oil exploration? *Organic Geochemistry*, Vol.22,

optical properties of asphaltene and their application to source rock evaluation. *Organic Geochemistry*, Vol.28, No. 11, (July,1998), pp. 669-676, ISSN 0146-6380 Xiao, X.M., Wilkins, R. W. T., Liu D.H. Liu, Z. F. & Fu, J. M. (2000). Investigation of thermal

maturity of lower Palaeozoic hydrocarbon source rocks by means of vitrinite-like maceral reflectance — a Tarim Basin case study. *Organic Geochemistry*, Vol.31,

microscopy—application to possible high rank and carbonate source rocks. *International* 

on common inverted microscope to measure UV-VIS spectra of a micro-area.

measure UV-VIS spectra of single hydrocarbon inclusions. *Proceedings of SPIE 7384*,

ultraviolet-visible micro-fluorescence-spectra and chromaticity of single oil inclusion. *ACTA OPTICA SINICA*, Vol.31, No.3, (Mar., 2011), pp. 0318002-1-6, ISSN 0253-2239

In the field of food engineering, chemical engineering and bioengineering have commonly used the data obtained from physicochemical techniques. In particular, water species are strongly related to food quality, as has been demonstrated in the International Symposium on Properties of Water (Eds. Rockland and Stewart, 1981). The water species retained in foods, as is well known, expose their multifunctional nature by dynamically responding to environmental conditions such as temperature, pressure, dehydration processes, water content, etc. For scientific analysis of the multifunctional water species retained in foods, of particular difficulty is the influence of nonlinear dynamic change on food quality. To quantitatively evaluate the nature of the water species, a large number of researchers have demonstrated the diversity of their biological and physicochemical nature using various parameters. The water activity (aw) has, for example, commonly been used as a parameter to evaluate the deterioration of foods (Fennema, 1976). The effective water diffusivity parameter (*De*) in foods has also frequently been employed to evaluate the dehydration rate (Jason, 1958). From the chemical engineering point of view, this is a typical procedure in evaluating water mobility in the food tissue matrix. Conversely, for direct identification of the water species at the molecular level, magnetic resonance techniques have been applied to evaluate food quality, as demonstrated in the International Conferences on Applications of Magnetic Resonance in Food Science (Eds. Belton *et al.*, 2003).

In our previous papers (Konishi *et al.*, 2001, 2003), using separately chemical engineering and proton NMR techniques, water species retained in fish paste sausage and squid were roughly distinguished into two different species as a function of water content. The two species were water species A1, to be recognized as the higher water content at *W*0 >120%-d.b. accompanied by higher water diffusivity, *De,* and water species A2, *W*0 <120%-d.b., with lower *De*. This delineation, however, was unclear because of a vagueness of the boundary

Corresponding Author

Challenging Evaluation of the Hybrid Technique of

**2.2 Evaluation of the parameters** 

time (

Chemical Engineering – Proton NMR Technique for Food Engineering 71

of 30~80%-d.b. (dry base, *W*D), the pork and beef meats had the initial water contents of 230320%-d.b. and 230280%-d.b., respectively, and the initial water content of the salmon and the squid commonly ranged from 300360%-d.b. To evaluate the effective diffusivity (*De*) of water species, each of the samples was placed in a stainless steel net tray (4 meshes) that was mechanically hung from a strain gage transducer in the dryer. The sample weight was, to evaluate the dehydration rate, continuously recorded by the output of strain-gage transducer using a data-logger. With molecular-level analysis, to evaluate the correlation

C) of the water species in the salmon, squid, 12 jerkies, BH, BA, and PH, a nuclear magnetic resonance (NMR) technique was applied to measure the 1H-NMR spectra and the spin-spin relaxation time (*T*2) of water protons. All samples, cut into 2×2×10 mm pieces, were inserted into an NMR sample tube (4mm in inner diameter, 180mm in length). 1H-NMR spectra were obtained using a JEOL A-500 FT-NMR spectrometer operating at 500MHz for protons. The observed frequency width was 20 kHz. The 90º pulse width was 12.5 *μs*, and the number of pulse repetitions was 8. The proton chemical shifts were measured by using a slight amount of water containing deuterium oxide as an external reference. All the NMR measurements were performed at 23.50.5ºC. The spin-spin relaxation times, *T*2, were obtained by the spin locking method. The hardness (N/m2) of the samples was measured using a creep tester equipped with a V-shaped plunger (30 mm

As has been reported by a large number of researchers (Andreu and Stamatopoulos, 1986; Waananen and Okos, 1996; Litchfield and Okos, 1992; Kannan and Bandyopadhyay, 1995), diverse mathematical equations for evaluating the effective diffusion coefficient (*De*) of water in foods have been demonstrated in various forms. In the present study, *De*'s of all the samples were evaluated by applying Equation (1) to the dehydration response curves obtained. In the present experimental drying conditions, it was reconfirmed that the drying operations were within a falling-rate period. Through all the evaluation of *De*, the effect of the shrinkage of the samples (derived from the dehydration) on the values obtained was

 D

273

D -

222

(1)

(2)

wide, 1 mm thick) to press a 60% of the sample size of 2~8×10×50 mm.

**2.2.1 Effective diffusion coefficient (***De***) of water species** 

previously evaluated, and the sizes of the samples were corrected.

*D*

*W We*

food and the activation energy, as shown in Equation (2).

3 2

π 4 *<sup>a</sup>*

8 π

*W We De t*

are the half distances of the sample width, and *t* is dehydration time (h).

2 b c

Where *W* is the water content (%-d.b.) of the sample at *t* = *t* (h), *We* is the equilibrium water content (%-d.b.), *W*D is the water content (%-d.b.) at the given initial condition, *De* is the effective diffusion coefficient (m2/h) of water species retained in the foods, *L*a*, L*b*,* and *L*c (m)

*De* is represented as a function of the structure parameters (porosity and labyrinth factor) of

0

*<sup>E</sup> De D D exp R T*

*exp LLL*

between the two water species. To clearly discriminate between the two water species, a new technique is needed for the design of food products. In the present study, responding to this need, a hybrid technique derived from joint research using chemical engineering techniques and proton NMR techniques was proposed. The aims of this study are: (1) to demonstrate the dynamism of water species influenced by the change of environmental conditions and the dynamic transformation between the two water species, A1 and A2; (2) to quantitatively visualize the limitations of both the chemical engineering technique and the proton NMR technique; (3) to apply a hybrid technique in choosing specified design parameters in order to clearly identify differences between the two water species and among a variety of food products; and (4) to classify twelve commercial jerky products, as a typical example, into a few groups based on the characterization of design parameters using the hybrid technique.

#### **2. Experimental methods of the hybrid technique**

#### **2.1 Methods**

Six commercially distributed beef and pork jerkies, (BJ-1BJ-6) and (PJ-1PJ-6), respectively, a pork meat (designated PH, used as a reference meat) produced in Hokkaido, two beef meats (BA: produced in Australia, and BH: produced in Hokkaido), salmon (SM), and squid (SQ) were chosen as the food samples. All samples used in this study were tabulated in Table 1. The commercial jerkies used were characterized as having a water content range


Table 1. Features of the pork meat (PH), beef meat (BA and BH), squid (SQ), salmon (SM), and the commercial jerkies used for this study. PJ-1~-6: pork jerkies; BJ-1~-6: beef jerkies.

of 30~80%-d.b. (dry base, *W*D), the pork and beef meats had the initial water contents of 230320%-d.b. and 230280%-d.b., respectively, and the initial water content of the salmon and the squid commonly ranged from 300360%-d.b. To evaluate the effective diffusivity (*De*) of water species, each of the samples was placed in a stainless steel net tray (4 meshes) that was mechanically hung from a strain gage transducer in the dryer. The sample weight was, to evaluate the dehydration rate, continuously recorded by the output of strain-gage transducer using a data-logger. With molecular-level analysis, to evaluate the correlation time (C) of the water species in the salmon, squid, 12 jerkies, BH, BA, and PH, a nuclear magnetic resonance (NMR) technique was applied to measure the 1H-NMR spectra and the spin-spin relaxation time (*T*2) of water protons. All samples, cut into 2×2×10 mm pieces, were inserted into an NMR sample tube (4mm in inner diameter, 180mm in length). 1H-NMR spectra were obtained using a JEOL A-500 FT-NMR spectrometer operating at 500MHz for protons. The observed frequency width was 20 kHz. The 90º pulse width was 12.5 *μs*, and the number of pulse repetitions was 8. The proton chemical shifts were measured by using a slight amount of water containing deuterium oxide as an external reference. All the NMR measurements were performed at 23.50.5ºC. The spin-spin relaxation times, *T*2, were obtained by the spin locking method. The hardness (N/m2) of the samples was measured using a creep tester equipped with a V-shaped plunger (30 mm wide, 1 mm thick) to press a 60% of the sample size of 2~8×10×50 mm.

#### **2.2 Evaluation of the parameters**

70 Advances in Chemical Engineering

between the two water species. To clearly discriminate between the two water species, a new technique is needed for the design of food products. In the present study, responding to this need, a hybrid technique derived from joint research using chemical engineering techniques and proton NMR techniques was proposed. The aims of this study are: (1) to demonstrate the dynamism of water species influenced by the change of environmental conditions and the dynamic transformation between the two water species, A1 and A2; (2) to quantitatively visualize the limitations of both the chemical engineering technique and the proton NMR technique; (3) to apply a hybrid technique in choosing specified design parameters in order to clearly identify differences between the two water species and among a variety of food products; and (4) to classify twelve commercial jerky products, as a typical example, into a few groups based on the characterization of design parameters using the

Six commercially distributed beef and pork jerkies, (BJ-1BJ-6) and (PJ-1PJ-6), respectively, a pork meat (designated PH, used as a reference meat) produced in Hokkaido, two beef meats (BA: produced in Australia, and BH: produced in Hokkaido), salmon (SM), and squid (SQ) were chosen as the food samples. All samples used in this study were tabulated in Table 1. The commercial jerkies used were characterized as having a water content range

PJ-1 soy-sauce base taste, about 1.0 mm thick 2 PJ-2 soy-sauce base taste, 1.1~1.4 mm thick 1 PJ-3 light taste, oily, smoked flavor, 0.8~5.9 mm thick 2 PJ-4 salt base taste, smoked flavor, strong taste, hard, 2.2~4.8 mm thick 2 PJ-5 salt base taste, smoked flavor, strong taste, soft, 1.3~2.7 mm thick 3 PJ-6 soy-sauce base taste, light taste, hard, 1.9~3.5 mm thick 3 BJ-1 salt base taste, light taste, hard, 1.4~2.2 mm thick 2 BJ-2 salt base taste, strong taste, soft, about 2.2 mm thick 3 BJ-3 miso-sauce base taste, strong taste, soft, 4.2~5.9 mm thick 3 BJ-4 salt base taste, light taste, soft, 1.6~3.1 mm thick 2 BJ-5 soy-sauce base taste, strong taste, hard, 3.1 mm thick 1 BJ-6 salt base taste, light taste, hard, 0.9~2.8 mm thick 2

pork meat produced in Hokkaido, soy-sauce base taste, hard, about

Table 1. Features of the pork meat (PH), beef meat (BA and BH), squid (SQ), salmon (SM), and

the commercial jerkies used for this study. PJ-1~-6: pork jerkies; BJ-1~-6: beef jerkies.

BH beef meat produced in Hokkaido, about 8±1.0 mm thick BA beef meat produced in Australia, about 8±1.0 mm thick

Features Group No.

hybrid technique.

**2.1 Methods** 

PH

6±1.0 mm thick

SQ squid mantle, 50×50×7±1.5 mm SM salmon, 50×15×12~26 mm

**2. Experimental methods of the hybrid technique** 

#### **2.2.1 Effective diffusion coefficient (***De***) of water species**

As has been reported by a large number of researchers (Andreu and Stamatopoulos, 1986; Waananen and Okos, 1996; Litchfield and Okos, 1992; Kannan and Bandyopadhyay, 1995), diverse mathematical equations for evaluating the effective diffusion coefficient (*De*) of water in foods have been demonstrated in various forms. In the present study, *De*'s of all the samples were evaluated by applying Equation (1) to the dehydration response curves obtained. In the present experimental drying conditions, it was reconfirmed that the drying operations were within a falling-rate period. Through all the evaluation of *De*, the effect of the shrinkage of the samples (derived from the dehydration) on the values obtained was previously evaluated, and the sizes of the samples were corrected.

$$\frac{\mathcal{W} - \mathcal{W}e}{\mathcal{W}\_D - \mathcal{W}e} = \left(\frac{8}{\text{n}^2}\right)^3 \exp\left(\frac{-\text{n}^2 \cdot De \cdot t}{\text{4}} \cdot \left(L\_a^{-2} + L\_b^{-2} + L\_c^{-2}\right)\right) \tag{1}$$

Where *W* is the water content (%-d.b.) of the sample at *t* = *t* (h), *We* is the equilibrium water content (%-d.b.), *W*D is the water content (%-d.b.) at the given initial condition, *De* is the effective diffusion coefficient (m2/h) of water species retained in the foods, *L*a*, L*b*,* and *L*c (m) are the half distances of the sample width, and *t* is dehydration time (h).

*De* is represented as a function of the structure parameters (porosity and labyrinth factor) of food and the activation energy, as shown in Equation (2).

$$D\varepsilon = \left(\frac{\varepsilon}{\mathcal{X}}\right) \cdot D = \delta \cdot D\_0 \cdot \exp\left[\frac{\cdot E\_{\rm D}}{R \cdot \left(T\_{\rm D} + 273\right)}\right] \tag{2}$$

Challenging Evaluation of the Hybrid Technique of

**Protein(Food matrix)**

*Adsorbed moisture amount*

(*Smaller*)

(*Larger*)

*Monolayer adsorption*

than the two water species in regions I and II.

Chemical Engineering – Proton NMR Technique for Food Engineering 73

*Normal isotropic bulk water Weakly restricted water Capillary condensation water*

*Region II*

*Region I*

**Restriction degree**

**Weaker**

**Stronger**

*Free water*

**<sup>C</sup>= 10-12s**

*Region III*

**<sup>C</sup>= 10-6 -11s**

**<sup>C</sup>= 10-6 -7s**

*NMR-Correlation time* **(**

(*Lower* 

 C)

*Curve 1*

 (*Higher* 

 C)

 C **)**

*Free water*

*Free water Freezing water Bonded water Glass-state water Multi-adsorbed water Strongly restricted water*

*Non-freezing water Immobilized water Bond water*

Fig. 1. Schematic explanation of the water species adsorbed on macromolecules in food.

*Water content* **(***W***)** *Region I Region II Region III* ( *Lower-W*) (*Higher-W*)

*Stronger* **Restriction degree** *Weaker*

*Water activity* **(aw)**

Fig. 2. Classification of the adsorbed water species in the food tissue matrix.

**3.1.2 Visualization from the chemical engineering technique** 

*Multilayer adsorption*

*Curve 2*

( aW = 0) (aW = 1.0)

conjunction with the progress of dehydration. In atmospheric conditions, according to the progress of dehydration, the free water species in region III is preferentially removed rather

Taking into account the water activity map presented by Labuza *et al.* (1970) and Schmidt (2004), one may visualize three regions on a water activity space. Fig. 2 shows a schematic explanation of the specified two curves: curve 1 for the water adsorption isotherm of food systems as a function of water activity (the isotherm was divided again into three regions, I, II, and III) and curve 2 for the NMR correlation time (*τ*C) of water species retained in food as

Where *ε* is porosity, *χ* is labyrinth factor, *D* is diffusivity (m2/h), *D*0 is the frequency factor of *D*, is diffusibility (= / ), *E*D is activation energy of *De*, *R* is gas constant, and *T*D is drying temperature.

The pre-exponential factor, *D*0, can be a useful tool to discriminate between the water species retained in different kinds of foods, as is described in upcoming sections.

#### **2.2.2 Correlation time (<sup>C</sup>) of water protons**

The spin-locking pulse technique used was effective in detecting a very fast relaxation signal at low water contents. For evaluation of the relaxation time, *T*2 , the equation *M*t = *M*<sup>0</sup> *exp*( *ts*/*T*2) was used, where *M*t is the magnitude of the magnetization vector after the spin locking pulse, *M*0 is magnitude of the macroscopic equilibrium magnetization vector, and *ts* is the spin locking pulse length. In the present study, the plot of *ln* [*M*t / *M*0] vs. *ts* indicated good linearity (which demonstrates a correlation coefficient higher than 0.99) through all water content of the foods, suggesting that the evaluated *T*2 value was reasonable. From *T*2 , the correlation time of a water proton, *τ*C , was evaluated using Equation (3) as described by Abragam (1963):

$$\frac{1}{T\_2} = \frac{\chi^4 \cdot \hbar^2 \cdot I(l+1)}{5r^6} \left( 3\mathbf{r}\_c + \frac{5\mathbf{r}\_c}{1 + a\_0^2 \cdot \mathbf{r}\_c^2} + \frac{2\mathbf{r}\_c}{1 + 4a\_0^2 \cdot \mathbf{r}\_c^2} \right) \tag{3}$$

where is the gyromagnetic ratio of a proton (= 2.675108 rad·T-1·s-1), *ħ* is the modified Plank's constant (6.6310-34 J· s), *I* is the nuclear-spin quantum number of a water proton (= 0.5), *r* is the proton-proton distance of a water molecule (0.16 nm), 0 is the resonance frequency of NMR (= 3.14109 s-1), and C is the correlation time of a water proton (s).

#### **3. Dynamic transformation of water species retained in foods**

#### **3.1 Variety of water species in foods**

#### **3.1.1 Visualization from the proton NMR technique**

The water species retained in foods are dynamically changed into multifunctional states according to environmental conditions such as temperature, atmospheric pressure, atmospheric relative humidity, seasoning components such as salt and sugar, concentration of seasoning components, etc. Figure 1 is a schematic explanation of the water species adsorbed on macromolecules such as proteins in foods. The water species are classified into three groups as: (1) monolayer adsorbed water localized in region-I, which is referred to as bond water, immobilized water, or non-freezing water; (2) multi-adsorbed water localized in region-II, which is referred to as strongly restricted water, glass-state water, etc.; and (3) free water localized in region-III, which is referred to as normal isotropic bulk water, capillary condensation water, etc. The proton NMR technique produced correlation times of the three regions, I, II, and III, as C= 10-6~-7s, 10-6~-11s, and 10-12s, respectively. This classification is interestingly correlated to the three regions of the adsorption isotherm evaluated by the chemical engineering technique, as shown in Figure 2 by curve 1. As has been reported in previous papers (Konishi *et al.*, 2003, 2010), water species in the three regions are dynamically transformed by each other depending on environmental conditions in

Where *ε* is porosity, *χ* is labyrinth factor, *D* is diffusivity (m2/h), *D*0 is the frequency factor of

The spin-locking pulse technique used was effective in detecting a very fast relaxation signal at low water contents. For evaluation of the relaxation time, *T*2 , the equation *M*t = *M*<sup>0</sup> *exp*( *ts*/*T*2) was used, where *M*t is the magnitude of the magnetization vector after the spin locking pulse, *M*0 is magnitude of the macroscopic equilibrium magnetization vector, and *ts* is the spin locking pulse length. In the present study, the plot of *ln* [*M*t / *M*0] vs. *ts* indicated good linearity (which demonstrates a correlation coefficient higher than 0.99) through all water content of the foods, suggesting that the evaluated *T*2 value was reasonable. From *T*2 , the correlation time of a water proton, *τ*C , was evaluated using Equation (3) as described by

6 22 22

is the gyromagnetic ratio of a proton (= 2.675108 rad·T-1·s-1), *ħ* is the modified

*c c <sup>c</sup>*

*c c*

C is the correlation time of a water proton (s).

0 is the resonance

, (3)

2 0 0 <sup>1</sup> <sup>1</sup> 5 2 <sup>3</sup>

0.5), *r* is the proton-proton distance of a water molecule (0.16 nm),

**3. Dynamic transformation of water species retained in foods** 

5 1 14

Plank's constant (6.6310-34 J· s), *I* is the nuclear-spin quantum number of a water proton (=

The water species retained in foods are dynamically changed into multifunctional states according to environmental conditions such as temperature, atmospheric pressure, atmospheric relative humidity, seasoning components such as salt and sugar, concentration of seasoning components, etc. Figure 1 is a schematic explanation of the water species adsorbed on macromolecules such as proteins in foods. The water species are classified into three groups as: (1) monolayer adsorbed water localized in region-I, which is referred to as bond water, immobilized water, or non-freezing water; (2) multi-adsorbed water localized in region-II, which is referred to as strongly restricted water, glass-state water, etc.; and (3) free water localized in region-III, which is referred to as normal isotropic bulk water, capillary condensation water, etc. The proton NMR technique produced correlation times of the three regions, I, II, and III, as C= 10-6~-7s, 10-6~-11s, and 10-12s, respectively. This classification is interestingly correlated to the three regions of the adsorption isotherm evaluated by the chemical engineering technique, as shown in Figure 2 by curve 1. As has been reported in previous papers (Konishi *et al.*, 2003, 2010), water species in the three regions are dynamically transformed by each other depending on environmental conditions in

*γ I(I ) τ τ <sup>τ</sup> T <sup>r</sup> ωτ ωτ* 

species retained in different kinds of foods, as is described in upcoming sections.

**<sup>C</sup>) of water protons** 

), *E*D is activation energy of *De*, *R* is gas constant, and *T*D is drying

*D*0, can be a useful tool to discriminate between the water

*D*, 

temperature.

is diffusibility (=

The pre-exponential factor,

**2.2.2 Correlation time (**

Abragam (1963):

where  /

4 2

**3.1.1 Visualization from the proton NMR technique** 

frequency of NMR (= 3.14109 s-1), and

**3.1 Variety of water species in foods** 

Fig. 1. Schematic explanation of the water species adsorbed on macromolecules in food.

Fig. 2. Classification of the adsorbed water species in the food tissue matrix.

conjunction with the progress of dehydration. In atmospheric conditions, according to the progress of dehydration, the free water species in region III is preferentially removed rather than the two water species in regions I and II.

#### **3.1.2 Visualization from the chemical engineering technique**

Taking into account the water activity map presented by Labuza *et al.* (1970) and Schmidt (2004), one may visualize three regions on a water activity space. Fig. 2 shows a schematic explanation of the specified two curves: curve 1 for the water adsorption isotherm of food systems as a function of water activity (the isotherm was divided again into three regions, I, II, and III) and curve 2 for the NMR correlation time (*τ*C) of water species retained in food as

Challenging Evaluation of the Hybrid Technique of

the course of a drying operation.

species are not influenced by dehydration.

0

2 4

6 8

10

Intensity(a.u.)

squid muscle.

12 14

(3): the sample of

(1)the sample of

 from W

from W W

W0=18%-d.b.

given by the dehydration

0=360%-d.b.

given by the dehydration

0=45%-d.b.

16

Chemical Engineering – Proton NMR Technique for Food Engineering 75

Fig. 4. Dynamic behavior of 1H-NMR spectra for species A, B, and C in the squid muscle in

restricted water, and lipids, respectively. One may recognize the reduction of only peak-A due to dehydration, whereas peaks-B and -C gave an identical value indicating that the two

Focusing on the reversibility of peak-A, one can easily recognize the dynamism of proton NMR spectra of water species in the squid, as shown in Figure 5. The strength of peak-A was steeply reduced by dehydration (see curve (1)), and the peak height was restored by re-hydration (see curve (2)), whereas the peaks-B and -C kept the same peak height regardless of dehydration and re-hydration. This reproducibility of peak-A strongly demonstrates the ability

0=360%-d.b. (2): the sample of

W0=45%-d.b.

W0=18%-d.b.

given by the re-hydration of the

sample (1) from

012345 -2-1 -3 -4 -5

Chemical shift(ppm)

Fig. 5. Reproduced dynamic behavior of 1H-NMR spectra for species-A, -B and -C in the

a function of water content (*W*0). As can be seen from the comparison of the two curves, the water species adsorbed in region I with lower water activity contributes to the higher correlation time (*τ*C). The higher *τ*C means higher restriction strength derived from the food tissues. The water species adsorbed in the region III with higher water activity, contributing to lower *τ*C, which means lower restriction strength. These results bring about an inverse relation between the two parameters, *τ*C (curve 2), and the adsorbed water amount (curve 1) as shown in Fig. 2. This relation can be usefully employed in the hybrid technique as is described in upcoming sections.

The three water species distributed in regions I, II, and III presented in Figs. 1 and 2 can be further classified as species A1 and species A2. The water species in regions I and III belong to species A2 and species A1, respectively. The water species in region II belongs to an intermediate species between species A1 and species A2. Figure 3 schematically illustrates the water species A1 and A2 retained in foods as a model description. From the schematic visualization, one may recognize water species A1 to be weakly restricted by macromolecules in foods, such as proteins, at greater water content and water species A2 to be strongly restricted due to the progress of dehydration. The space formed due to the dehydration in region III in Fig. 1 would create micropores resulting in higher porosity (ε) and higher labyrinth factor (*χ*) in the food tissue matrix. These micropores would be filled again with water species when re-hydration occurred, indicating a reproducibility of the adsorbed water species as is discussed in upcoming sections (see Fig. 5).

Fig. 3. Schematic explanations of the adsorbed states for water species in foods.

#### **3.2 Dynamism and reversibility of the water species in foods identified by the proton NMR technique**

The proton NMR spectrum clearly demonstrates the dynamism of water species derived from the dehydration of the sample (Konishi *et al.,* 2010). Figure 4 demonstrates the dynamism of the reduced amount of water species (the dehydration of squid) as a steep reduction of the proton NMR spectra of peak-A. The three peeks, A, B, and C, were identified in our previous work (Konishi *et al*., 2010) as weakly restricted water, strongly

a function of water content (*W*0). As can be seen from the comparison of the two curves, the water species adsorbed in region I with lower water activity contributes to the higher correlation time (*τ*C). The higher *τ*C means higher restriction strength derived from the food tissues. The water species adsorbed in the region III with higher water activity, contributing to lower *τ*C, which means lower restriction strength. These results bring about an inverse relation between the two parameters, *τ*C (curve 2), and the adsorbed water amount (curve 1) as shown in Fig. 2. This relation can be usefully employed in the hybrid technique as is

The three water species distributed in regions I, II, and III presented in Figs. 1 and 2 can be further classified as species A1 and species A2. The water species in regions I and III belong to species A2 and species A1, respectively. The water species in region II belongs to an intermediate species between species A1 and species A2. Figure 3 schematically illustrates the water species A1 and A2 retained in foods as a model description. From the schematic visualization, one may recognize water species A1 to be weakly restricted by macromolecules in foods, such as proteins, at greater water content and water species A2 to be strongly restricted due to the progress of dehydration. The space formed due to the dehydration in region III in Fig. 1 would create micropores resulting in higher porosity (ε) and higher labyrinth factor (*χ*) in the food tissue matrix. These micropores would be filled again with water species when re-hydration occurred, indicating a reproducibility of the

*Lower water content Higher water content*

**3.2 Dynamism and reversibility of the water species in foods identified by the proton** 

The proton NMR spectrum clearly demonstrates the dynamism of water species derived from the dehydration of the sample (Konishi *et al.,* 2010). Figure 4 demonstrates the dynamism of the reduced amount of water species (the dehydration of squid) as a steep reduction of the proton NMR spectra of peak-A. The three peeks, A, B, and C, were identified in our previous work (Konishi *et al*., 2010) as weakly restricted water, strongly

Fig. 3. Schematic explanations of the adsorbed states for water species in foods.

**Water species A**<sup>1</sup> (**Weakly restricted water**)

*Protein*

*Protein*

adsorbed water species as is discussed in upcoming sections (see Fig. 5).

**Water species A2** (**Strongly restricted water**)

*Protein*

*Protein*

described in upcoming sections.

**Bonded water**

**NMR technique** 

Fig. 4. Dynamic behavior of 1H-NMR spectra for species A, B, and C in the squid muscle in the course of a drying operation.

restricted water, and lipids, respectively. One may recognize the reduction of only peak-A due to dehydration, whereas peaks-B and -C gave an identical value indicating that the two species are not influenced by dehydration.

Focusing on the reversibility of peak-A, one can easily recognize the dynamism of proton NMR spectra of water species in the squid, as shown in Figure 5. The strength of peak-A was steeply reduced by dehydration (see curve (1)), and the peak height was restored by re-hydration (see curve (2)), whereas the peaks-B and -C kept the same peak height regardless of dehydration and re-hydration. This reproducibility of peak-A strongly demonstrates the ability

Fig. 5. Reproduced dynamic behavior of 1H-NMR spectra for species-A, -B and -C in the squid muscle.

Challenging Evaluation of the Hybrid Technique of

0.E+00

Fig. 7. τC as a function of *W*0 for BA, BH, and PH.

Cτ<sup>C</sup>

indicating again a limitation of the proton NMR technique.

PJ-1

PJ-4

0.0E+00

2.0E-08

4.0E-08

6.0E-08

τC (s)

and the pork (solid line).

8.0E-08

1.0E-07

1.2E-07

PJ-2

BJ-5

BJ-6

5.E-08

1.E-07

τC (s)

2.E-07

Chemical Engineering – Proton NMR Technique for Food Engineering 77

*Species A2 Species A1*

○ *BA* △ *B <sup>H</sup>* □ *P <sup>H</sup>*

0 50 100 150 200 250 *W* 0(%-d.b.)

BJ-3

*Species A2 Species A1*

Reference-PH

0 20 40 60 80 100 120 140 160 *W* <sup>0</sup> (%-d.b.)

PJ-3

Fig. 8. τc as a function of *W*0 for the pork and beef jerkies (open circle and closed triangle)

BJ-4

PJ-5 PJ-6 BJ-1

BJ-2

Figure 8 illustrates *τ*C as a function of water content (*W*0) of the 12 jerkies and the PH. The solid line for PH indicated a steep increase with decreasing *W*0, between the values of *τ*C = 1.0×10-8s and *τ*C = 1.1×10-7s. Since *τ*C indicates rotation time of the water species, the steep increase of *τ*C demonstrates that the restriction strength of the water species in this region increases markedly upon dehydration from *W*0 = 120 to 20%-d.b. Similar results were also obtained for the beef (Konishi and Kobayashi, 2009). The *τ*C for the jerkies, on the other hand, fell around the solid line of PH. From these results, one can recognize that it is difficult to distinguish among the three *τ*C –*W*0 curves for the pork jerkies, beef jerkies, and PH,

of water species A1 and A2 to be reformed in the food tissue matrix through re- hydration. This reversibility is a very important factor for dried food products because, generally speaking, dried foods are frequently re-hydrated before being consumed (dried kelp, dried green onions, and dried noodles, etc.). Exploitation of reversible water species in dried foods could produce the same taste as when the food is freshly prepared. To evaluate the reproducibility of dried food taste, one should focus on C because the C value is closely related to the concentration of seasoning and the kind of food. The reproducibility of C is, therefore, an important factor for taste reproducibility. Figure 6 demonstrates the reproducibility of C as a function of *W*0. The results show positive reproducibility of C, indicating favorable taste reproducibility in the case of dried squid.

Fig. 6. Reproducibility of τC(species A) observed between the dehydration and the waterreadsorption operations for the squid muscle: (a) under the continuous drying operation from *W*0 = 120%-d.b. to 10%-d.b., (b) after the gradual readsorption of water from *W*0 = 18% d.b. to *W*0 = 45%-d.b., (c) from *W*0 = 20%-d.b. to *W*0 = 44%-d.b., (d) from *W*0 = 50%-d.b. to *W*0 = 60%-d.b., (e) from *W*0 = 68%-d.b. to *W*0 = 89%-d.b., and (f) from *W*0 = 94%-d.b. to *W*0 = 111%-d.b.

#### **4. Limitations of the proton NMR technique**

To design the food products requested by the markets, the design parameters chosen should be able to clearly discriminate between various kinds of foods depending on requirements. In particular, the water species retained in the foods should be characterized by the kinds of foods and environmental influences such as temperature, atmospheric pressure, water content, etc. Although the proton NMR water species evaluation technique is a useful tool, one should recognize that the method has some difficulty discriminating among water species retained in foods. Figure 7 illustrates the C values as a function of *W*0 for BA, BH, and PH. The C~*W*0 curves obtained could not discriminate among the three meats, BA, BH, and PH (Konishi and Kobayashi, 2009), even though the water species A1 and A2 are roughly distinguished at the critical point of C = CC = 10-8 s. This unclear discrimination between the three curves can be understood as a limitation of the proton NMR technique.

of water species A1 and A2 to be reformed in the food tissue matrix through re- hydration. This reversibility is a very important factor for dried food products because, generally speaking, dried foods are frequently re-hydrated before being consumed (dried kelp, dried green onions, and dried noodles, etc.). Exploitation of reversible water species in dried foods could produce the same taste as when the food is freshly prepared. To evaluate the reproducibility of dried food taste, one should focus on C because the C value is closely related to the concentration of seasoning and the kind of food. The reproducibility of C is, therefore, an important factor for taste reproducibility. Figure 6 demonstrates the reproducibility of C as a function of *W*0. The results show positive reproducibility of C, indicating favorable taste

> 0 20 40 60 80 100 120 *W* 0 (%-d.b.)

Fig. 6. Reproducibility of τC(species A) observed between the dehydration and the waterreadsorption operations for the squid muscle: (a) under the continuous drying operation from *W*0 = 120%-d.b. to 10%-d.b., (b) after the gradual readsorption of water from *W*0 = 18% d.b. to *W*0 = 45%-d.b., (c) from *W*0 = 20%-d.b. to *W*0 = 44%-d.b., (d) from *W*0 = 50%-d.b. to *W*0 = 60%-d.b., (e) from *W*0 = 68%-d.b. to *W*0 = 89%-d.b., and (f) from *W*0 = 94%-d.b. to

To design the food products requested by the markets, the design parameters chosen should be able to clearly discriminate between various kinds of foods depending on requirements. In particular, the water species retained in the foods should be characterized by the kinds of foods and environmental influences such as temperature, atmospheric pressure, water content, etc. Although the proton NMR water species evaluation technique is a useful tool, one should recognize that the method has some difficulty discriminating among water species retained in foods. Figure 7 illustrates the C values as a function of *W*0 for BA, BH, and PH. The C~*W*0 curves obtained could not discriminate among the three meats, BA, BH, and PH (Konishi and Kobayashi, 2009), even though the water species A1 and A2 are roughly distinguished at the critical point of C = CC = 10-8 s. This unclear discrimination between

the three curves can be understood as a limitation of the proton NMR technique.

○ □ △ ● ■

(a) (b) (c) (d) (e) (f)

reproducibility in the case of dried squid.

1.5E-07

0.0E+00

**4. Limitations of the proton NMR technique** 

5.0E-08

*τ*

*W*0 = 111%-d.b.

C (s) - Species A

1.0E-07

Fig. 7. τC as a function of *W*0 for BA, BH, and PH.

Figure 8 illustrates *τ*C as a function of water content (*W*0) of the 12 jerkies and the PH. The solid line for PH indicated a steep increase with decreasing *W*0, between the values of *τ*C = 1.0×10-8s and *τ*C = 1.1×10-7s. Since *τ*C indicates rotation time of the water species, the steep increase of *τ*C demonstrates that the restriction strength of the water species in this region increases markedly upon dehydration from *W*0 = 120 to 20%-d.b. Similar results were also obtained for the beef (Konishi and Kobayashi, 2009). The *τ*C for the jerkies, on the other hand, fell around the solid line of PH. From these results, one can recognize that it is difficult to distinguish among the three *τ*C –*W*0 curves for the pork jerkies, beef jerkies, and PH, indicating again a limitation of the proton NMR technique.

Fig. 8. τc as a function of *W*0 for the pork and beef jerkies (open circle and closed triangle) and the pork (solid line).

Challenging Evaluation of the Hybrid Technique of

0

*a*

0.0

0.2

0.4

0.6

Dehydration rate (g/(g・h))

salmon (*T*D = 45C).

0.8

1.0

*a'*

*b'*

evaluated as a function of *W*0 by applying Equation (1).

*c'*

*d'*

0

100

200

Moisture content(%-d.b.)

300

Fig. 10. Dehydration response curves of the salmon (*T*D = 45°C).

*b c*

50

100

150

*W*

(%-d.b.)

200

250

300

350

Chemical Engineering – Proton NMR Technique for Food Engineering 79

*Exp.* ○ △ □ ● ▲ ■

*Cal.*

0 5 10 15 20 25 *Dring time* (h)

> *b~f* :

*PUP-operation*

**(A)**

**(B)**

*<sup>d</sup> <sup>e</sup> <sup>f</sup>*

0 5 10 15 20 Drying time(h)

*f'*

*a, a' b, b' c, c' d, d' e, e' f, f'*

○ △ □ ◇ ● ▲

*e'*

Fig. 11. Acceleration behavior of dehydration rate derived from the PUP operation for the

curve reaching finally to the curve of *a*' with an increase in the elapsed time. Using the dehydration rates obtained, the effective diffusivity of the water species, *De*, can be

#### **5. Limitation of the chemical engineering technique**

#### **5.1 Unclear discrimination of water activity in foods in the chemical engineering technique**

Water activity (aw) has conveniently been used to design food products because of the ease of evaluating the variable using the chemical engineering technique (Fennema, 1976). Figure 9 illustrates aw as a function of *W*0 for the squid and the salmon. As can be seen from the comparison between the two curves, an identical curve was obtained without detecting any difference between the squid and the salmon. This result strongly indicates a limitation of the chemical engineering technique. To develop a new food product, one should choose another parameter for discriminating among products. To address this issue, a computer simulation method for developing new food design parameters is proposed in the next section.

Fig. 9. aw as a function of *W*0 for the salmon and squid.

#### **5.2 Ambiguous physical meaning of the chemical engineering parameters derived from a computer simulation for the dehydration dynamism of salmon**

#### **5.2.1 Dehydration dynamism of the water species in the salmon**

To characterize the water species retained in the salmon, dehydration response curves could effectively be employed. Figure 10 illustrates the dehydration response curves as a function of the water content of the salmon. The five samples of *W*0 = 54, 99, 144, 200, and 252%-d.b. were prepared from a sample of *W*0 = 342%-d.b., and each of the samples was treated by a PUP-operation (poultice up: the samples were stored in a dark room at *T* = 2ºC for 24 h before use; this operation was conducted to provide homogeneity of water distribution in the sample). The PUP operation contributes to the stepwise increase of the dehydration rate, as shown in Figure 11, by the curves of *b*' ~ *f* '. Focusing on the five points of *b* ~ *f* on the dehydration curve of *W*0 = 342%-d.b. in Fig. 11(A), the dehydration rates of the five samples clearly demonstrate a stepwise increase, as shown in Fig. 11(B), by the curves of *b' ~ f '* due to PUP operation. One can recognize that curves *b*' ~ *f* ' demonstrate a monotonous decay

Water activity (aw) has conveniently been used to design food products because of the ease of evaluating the variable using the chemical engineering technique (Fennema, 1976). Figure 9 illustrates aw as a function of *W*0 for the squid and the salmon. As can be seen from the comparison between the two curves, an identical curve was obtained without detecting any difference between the squid and the salmon. This result strongly indicates a limitation of the chemical engineering technique. To develop a new food product, one should choose another parameter for discriminating among products. To address this issue, a computer simulation method for developing new food design parameters is proposed in the next

> 0 50 100 150 200 250 *W* 0 (%-d.b.)

**5.2 Ambiguous physical meaning of the chemical engineering parameters derived** 

To characterize the water species retained in the salmon, dehydration response curves could effectively be employed. Figure 10 illustrates the dehydration response curves as a function of the water content of the salmon. The five samples of *W*0 = 54, 99, 144, 200, and 252%-d.b. were prepared from a sample of *W*0 = 342%-d.b., and each of the samples was treated by a PUP-operation (poultice up: the samples were stored in a dark room at *T* = 2ºC for 24 h before use; this operation was conducted to provide homogeneity of water distribution in the sample). The PUP operation contributes to the stepwise increase of the dehydration rate, as shown in Figure 11, by the curves of *b*' ~ *f* '. Focusing on the five points of *b* ~ *f* on the dehydration curve of *W*0 = 342%-d.b. in Fig. 11(A), the dehydration rates of the five samples clearly demonstrate a stepwise increase, as shown in Fig. 11(B), by the curves of *b' ~ f '* due to PUP operation. One can recognize that curves *b*' ~ *f* ' demonstrate a monotonous decay

**from a computer simulation for the dehydration dynamism of salmon** 

**5.2.1 Dehydration dynamism of the water species in the salmon** 

○ □

*Dried salmon Dried squid*

**5.1 Unclear discrimination of water activity in foods in the chemical engineering** 

**5. Limitation of the chemical engineering technique** 

0.5

Fig. 9. aw as a function of *W*0 for the salmon and squid.

0.6

0.7

0.8

aw

0.9

1.0

**technique** 

section.

Fig. 10. Dehydration response curves of the salmon (*T*D = 45°C).

Fig. 11. Acceleration behavior of dehydration rate derived from the PUP operation for the salmon (*T*D = 45C).

curve reaching finally to the curve of *a*' with an increase in the elapsed time. Using the dehydration rates obtained, the effective diffusivity of the water species, *De*, can be evaluated as a function of *W*0 by applying Equation (1).

Challenging Evaluation of the Hybrid Technique of

decrease of water species mobility (*De*).

15

restriction strength, and the pre-exponential factor (PF =

20

25

*E*

Fig. 14. *E*D as a function of *W*0 for the salmon.

**curves** 

coefficient, *k*W (h-1) and *k*S (h-1).

D(kJ/mol)

30

35

40

45

Chemical Engineering – Proton NMR Technique for Food Engineering 81

Figure 14 demonstrates *E*D as a function of *W*0. The activation energy obtained, the *E*<sup>D</sup> values, demonstrates again two water species, A1 and A2, divided at *W*0 = 120%-d.b. *E*D in the water species A1 region gave an identical value of 22 (4) kJ/mol and a gradual increase from 20 to 38 (8)kJ/mol with decreasing *W*0. Taking into account the restriction of water species growth that accompanies increased dehydration, one can observe evidence of the activation energy growth (indicating a difficulty of the molecular diffusion of water species in the food tissue matrix due to the growth of the restriction strength) as a result of the

*Species A2 Species A1*

0 50 100 150 200 250 300 350 *W* 0(%-d.b.)

*D*0) of *De* showed steep growth

As for a reason that the activation energy was drastically changed at *W*0 = 120 %-d.b., it can be attributed to a change in the physical structure of the food tissue matrix due to progressive dehydration. The value of *W*0 = 120%-d.b. corresponds exactly to the critical value of C (CC = 10-8s), at which water species A1 and A2 were divided according to their

with decreasing *W*0 (as will be discussed in the upcoming section, see Fig. 16). The steep growth of PF resulted from the drastic change in the porosity and labyrinth factor of the salmon at *W*0 = 120 %-d.b. accompanied by the gradual reduction of *De* decreasing *W*0.

The transient response curves of dehydration in the drying operation of foods have been simulated using various mathematical models (Chhinnan, 1984; Madamba *et al.*, 1996; Thompson *et al.*, 1968; Noomhorm and Verma, 1986). In the present study, the dehydration response curves of the salmon presented in Fig. 10 are simulated using a water tank model as a first approximation. The water species retained in the sample are portioned to two different water tanks and the two tanks are respectively filled by the two different water species, A1 (weakly restricted water species) and A2 (strongly restricted water species), as shown in Figure 15 (Konishi *et al.*, 2001). During the drying operation, the two water species in the tanks are simultaneously drained by regulating the two valves as mass transfer

**5.2.3 Water tank model for the computer simulation of the dehydration response** 

#### **5.2.2 Physicochemical mobility (***De***) of the water species**

One can recognize the *De* value as a physicochemical mobility of the water species retained in the foods. Figure 12 illustrates *De* for the salmon as a function of *W*0 and temperature. Although the data are widely scattered, one can roughly recognize two different water species (species A1 and A2) formed in the salmon, depending on the values of *W*0. The two water species were divided at about *W*0 = 120%-d.b., indicating an identical *De* for the water species A1 region between *W*0 = 120 and 360%-d.b. and a gradual decrease of *De* for the water species A2 region with decreasing *W*0 at 35~55ºC. Using a rough evaluation of *De* obtained at the three temperatures, one would roughly construct the Arrhenius plots as shown in Figure 13. The slopes of the straight lines seem to give different activation energies depending on the water content of the samples.

Fig. 12. *De* as a function of *W*0 in the salmon.

Fig. 13. Arrhenius plots of *De* for the salmon.

One can recognize the *De* value as a physicochemical mobility of the water species retained in the foods. Figure 12 illustrates *De* for the salmon as a function of *W*0 and temperature. Although the data are widely scattered, one can roughly recognize two different water species (species A1 and A2) formed in the salmon, depending on the values of *W*0. The two water species were divided at about *W*0 = 120%-d.b., indicating an identical *De* for the water species A1 region between *W*0 = 120 and 360%-d.b. and a gradual decrease of *De* for the water species A2 region with decreasing *W*0 at 35~55ºC. Using a rough evaluation of *De* obtained at the three temperatures, one would roughly construct the Arrhenius plots as shown in Figure 13. The slopes of the straight lines seem to give different activation energies

*Species A2 Species A1*

0 50 100 150 200 250 300 350 *W* 0(%-d.b.)

0.0030 0.0031 0.0032 0.0033 1/*T* D(K-1)

○ △ □ ● ▲ ■

**5.2.2 Physicochemical mobility (***De***) of the water species** 

depending on the water content of the samples.

0.E+00

Fig. 12. *De* as a function of *W*0 in the salmon.


Fig. 13. Arrhenius plots of *De* for the salmon.




ln(*De* ) -13.5



1.E-06

2.E-06

*De*

(m2/h)

3.E-06

4.E-06

5.E-06

△ ○ □ *T* D(°C) 35 45 55

Figure 14 demonstrates *E*D as a function of *W*0. The activation energy obtained, the *E*<sup>D</sup> values, demonstrates again two water species, A1 and A2, divided at *W*0 = 120%-d.b. *E*D in the water species A1 region gave an identical value of 22 (4) kJ/mol and a gradual increase from 20 to 38 (8)kJ/mol with decreasing *W*0. Taking into account the restriction of water species growth that accompanies increased dehydration, one can observe evidence of the activation energy growth (indicating a difficulty of the molecular diffusion of water species in the food tissue matrix due to the growth of the restriction strength) as a result of the decrease of water species mobility (*De*).

Fig. 14. *E*D as a function of *W*0 for the salmon.

As for a reason that the activation energy was drastically changed at *W*0 = 120 %-d.b., it can be attributed to a change in the physical structure of the food tissue matrix due to progressive dehydration. The value of *W*0 = 120%-d.b. corresponds exactly to the critical value of C (CC = 10-8s), at which water species A1 and A2 were divided according to their restriction strength, and the pre-exponential factor (PF = *D*0) of *De* showed steep growth with decreasing *W*0 (as will be discussed in the upcoming section, see Fig. 16). The steep growth of PF resulted from the drastic change in the porosity and labyrinth factor of the salmon at *W*0 = 120 %-d.b. accompanied by the gradual reduction of *De* decreasing *W*0.

#### **5.2.3 Water tank model for the computer simulation of the dehydration response curves**

The transient response curves of dehydration in the drying operation of foods have been simulated using various mathematical models (Chhinnan, 1984; Madamba *et al.*, 1996; Thompson *et al.*, 1968; Noomhorm and Verma, 1986). In the present study, the dehydration response curves of the salmon presented in Fig. 10 are simulated using a water tank model as a first approximation. The water species retained in the sample are portioned to two different water tanks and the two tanks are respectively filled by the two different water species, A1 (weakly restricted water species) and A2 (strongly restricted water species), as shown in Figure 15 (Konishi *et al.*, 2001). During the drying operation, the two water species in the tanks are simultaneously drained by regulating the two valves as mass transfer coefficient, *k*W (h-1) and *k*S (h-1).

Challenging Evaluation of the Hybrid Technique of

respectively. *De*0 (=

**of** *De*

decreasing *W0*.

Chemical Engineering – Proton NMR Technique for Food Engineering 83

evaluated by a curve fitting simulation between the experimental and the calculated curves as shown in Fig. 10. The curve fitting obtained showed favorable agreement with the

**5.2.4 Discrimination of water species A1 and A2 due to the pre-exponential factor (***De***<sup>0</sup>**

The pre-exponential factor, *De*0, can easily be evaluated by the extrapolation of the Arrhenius plot lines against the perpendicular axes in Fig. 13. Figure 16(A) illustrates *De*0 as a function of *W*. The *De*0 values obtained were of identical value regardless of *W* in the water species A1 region, whereas in the water species A2 region, there was a steep increase with

*Species A2 Species A1*

0 100 200 300

Fig. 16. Pre-exponential factor (*De*0) and the dehydration parameters for the computer

Figure 16(B) illustrates *k*S and *k*W as a function of *W0*. The two values show tendencies similar to *De*0 even though they can be evaluated by the computer simulation. Comparing the two values between *k*S (or *k*W) and *De*0, one can generate a linear relation between the two values as shown in Figure 17. The parameters of *k*S and *k*W roughly demonstrate the equations of *k*S = 0.025(0.008) *De*0 and *k*W = 0.007(0.004) *De*0. Using Equations (8) and (9), one may evaluate β1 = 0.025(0.008) and β2 = 0.007(0.004). Concerning the physical meaning of β1 and β2, since the two parameters were evaluated by the computer fitting simulation, they seem to be meaningless. This evidence indicates a disadvantage of the chemical engineering technique, even though the linear relation between *k*S (or *k*W) and *De*0 would

*W* 0 (%-d.b.)

1 and 

*De* **<sup>0</sup>**

**(A)**

**(B)**

k *S* k*W* 2 are the constants

**)** 

*D*0) is the pre-exponential factor of *De*, and

experimental dehydration curves as shown in Fig. 10 by solid curves.

100

simulation (*k*S, *k*W) as a function of *W*0 (*T*D = 45°C) for the salmon.

cause the presumption of a relationship between the two parameters.

1000

*ks or kw* (h-1)

*De0, Apparent pre-exponential*

*factor*

10000

1.E+05

1.E+06

1.E+07

1.E+04

Fig. 15. Schematic explanation of the water tank model proposed as a dehydration of food.

On the material balance equation in the course of the dehydration operation presented in Fig. 10, the water content (*W*), presented as a perpendicular axis of Fig. 10, can be replaced by the water ratio, *W*R = (*W-We)*/(*W*D-*We*), where *We* is the equilibrium water content (% d.b.) and *W*D is the initial water content of the sample. The dynamism of *W*R for the two tanks is expressed by Equation (4).

$$\left(-\frac{d\mathcal{W}\_{RI}}{dt}\right) = \left(-\frac{d\mathcal{W}\_{Rw}}{dt}\right) + \left(-\frac{d\mathcal{W}\_{Rs}}{dt}\right) \tag{4}$$

$$-\frac{d\mathcal{W}\_{Rw}}{dt} = k\_w \cdot \mathcal{W}\_{Rw} \tag{5}$$

$$-\frac{d\mathcal{W}\_{\rm Rs}}{dt} = k\_s \cdot \mathcal{W}\_{\rm Rs} \quad \text{\(\(\prime\)}\tag{6}$$

where *W*RW is the ratio of the weakly restricted water species (=(*W*W-*We*) /(*W*- *W*D)), *W*RS is the ratio of the strongly restricted water species (=(*W*S-*We*) /(*W*- *W*D)), *W*W is the amount of weakly restricted water species (%-d.b.), and *W*S is the amount of strongly restricted water species (%-d.b.).

The differential equation (4) is easily solved and expressed by Equation (7).

$$\begin{split} \mathcal{W}\_{R} &= \frac{\mathcal{W} - \mathcal{W}e}{\mathcal{W}\_{\text{D}} - \mathcal{W}e} = \mathcal{W}\_{\text{R1}} \exp\left[k\_{s} \exp\left(\frac{-E\_{\text{D}}}{R \cdot \left(T\_{\text{D}} + 2\mathcal{T}\right)}\right) \cdot t\right] \\ &+ \mathcal{W}\_{\text{R2}} \exp\left[k\_{w} \exp\left(\frac{-E\_{\text{D}}}{R \cdot \left(T\_{\text{D}} + 2\mathcal{T}\right)}\right) \cdot t\right] \end{split} \tag{7}$$

$$k\_s = \beta\_1 \cdot De^0 \tag{8}$$

$$k\_w = \beta\_2 \cdot De^0 \,, \tag{9}$$

where *W*R1 and *W*R2 are a proportion of the strongly restricted water amount and the weakly restricted water amount, respectively, and *k*S (h-1) and *k*W (h-1) are the mass transfer coefficient of the strongly restricted water species and of the weakly restricted water species,

Fig. 15. Schematic explanation of the water tank model proposed as a dehydration of food.

tanks is expressed by Equation (4).

species (%-d.b.).

On the material balance equation in the course of the dehydration operation presented in Fig. 10, the water content (*W*), presented as a perpendicular axis of Fig. 10, can be replaced by the water ratio, *W*R = (*W-We)*/(*W*D-*We*), where *We* is the equilibrium water content (% d.b.) and *W*D is the initial water content of the sample. The dynamism of *W*R for the two

> *RI Rw Rs dW dW dW dt dt dt*

> > *w Rw*

*s Rs*

*Rw*

*Rs*

The differential equation (4) is easily solved and expressed by Equation (7).

*R R s*

*R w*

2

*D*

1

*dW k W*

*dW k W*

where *W*RW is the ratio of the weakly restricted water species (=(*W*W-*We*) /(*W*- *W*D)), *W*RS is the ratio of the strongly restricted water species (=(*W*S-*We*) /(*W*- *W*D)), *W*W is the amount of weakly restricted water species (%-d.b.), and *W*S is the amount of strongly restricted water

exp exp <sup>273</sup>

where *W*R1 and *W*R2 are a proportion of the strongly restricted water amount and the weakly restricted water amount, respectively, and *k*S (h-1) and *k*W (h-1) are the mass transfer coefficient of the strongly restricted water species and of the weakly restricted water species,

*<sup>E</sup> W k <sup>t</sup> R T*

*W We <sup>E</sup> W Wk <sup>t</sup> W We R T*

(4)

*dt* (5)

*dt* , (6)

*<sup>s</sup>* <sup>1</sup> *k β De* (8)

*<sup>w</sup>* <sup>2</sup> *k β De* , (9)

(7)

D

D

D

0

0

D

exp exp <sup>273</sup>

respectively. *De*0 (= *D*0) is the pre-exponential factor of *De*, and 1 and 2 are the constants evaluated by a curve fitting simulation between the experimental and the calculated curves as shown in Fig. 10. The curve fitting obtained showed favorable agreement with the experimental dehydration curves as shown in Fig. 10 by solid curves.

#### **5.2.4 Discrimination of water species A1 and A2 due to the pre-exponential factor (***De***<sup>0</sup> ) of** *De*

The pre-exponential factor, *De*0, can easily be evaluated by the extrapolation of the Arrhenius plot lines against the perpendicular axes in Fig. 13. Figure 16(A) illustrates *De*0 as a function of *W*. The *De*0 values obtained were of identical value regardless of *W* in the water species A1 region, whereas in the water species A2 region, there was a steep increase with decreasing *W0*.

Fig. 16. Pre-exponential factor (*De*0) and the dehydration parameters for the computer simulation (*k*S, *k*W) as a function of *W*0 (*T*D = 45°C) for the salmon.

Figure 16(B) illustrates *k*S and *k*W as a function of *W0*. The two values show tendencies similar to *De*0 even though they can be evaluated by the computer simulation. Comparing the two values between *k*S (or *k*W) and *De*0, one can generate a linear relation between the two values as shown in Figure 17. The parameters of *k*S and *k*W roughly demonstrate the equations of *k*S = 0.025(0.008) *De*0 and *k*W = 0.007(0.004) *De*0. Using Equations (8) and (9), one may evaluate β1 = 0.025(0.008) and β2 = 0.007(0.004). Concerning the physical meaning of β1 and β2, since the two parameters were evaluated by the computer fitting simulation, they seem to be meaningless. This evidence indicates a disadvantage of the chemical engineering technique, even though the linear relation between *k*S (or *k*W) and *De*0 would cause the presumption of a relationship between the two parameters.

Challenging Evaluation of the Hybrid Technique of

upcoming section.

next section.

Chemical Engineering – Proton NMR Technique for Food Engineering 85

indicating an identical *De* = 2.2(±0.2)×10-6 m2/h in the species A1 region and *De* = 3.5(±1.0) ×10-7 m2/h in the water species A2 region. All the *De* values for the commercial jerkies are distributed in the water species A2 region, indicating an identical *De* = 3.0(±1.8)×10-7 m2/h, which, even though the data are widely scattered, is the same as PH. Regarding the jerky data, although all the samples distributed in the market commonly use the water species A2 not A1 and the effective diffusivity, *De*, falling in the range of 1.0~5.0×10-7 m2/h, C is widely distributed in the range of 10-8 ~ 1.1×10-7 s. Since each of the jerkies used in this study has own taste different from other, the wide distribution of C has an important meaning in developing different tastes in jerky. This idea is also supported in an

As has been demonstrated in Fig. 12, the *De*~*W*0 plots could not clearly discriminate the water species A1 and A2. The *De*~*τC* plots derived from the hybrid technique, on the other hand, are conveniently able to visualize the discrimination between beef and pork and the two water species, A1 and A2, as shown in Figure 19. In particular, the boundary of the two water species is clearly recognized as C*τC* =10-8 s, indicating an advantage of the hybrid technique. Focusing on the behavior of BA, BH, and PH, the *De* values gave a constant value in the two water species regions: in the A1 region, as 3.1(±0.5) ×10-6, 4.7(±0.1) ×10-6, and 2.2(±0.3) ×10-6 m2/h, respectively; in the A2 region, as 1.8(±0.5) ×10-6, 2.0(±0.4) ×10-6, and 7.0(±1.0) ×10-7 m2/h, respectively, even though *τC* is significantly changed from 10-8 to 4.5×10-8 s. These results suggest the water species A2 has the interesting characteristic of being able to make a wide shift of the molecular mobility, *τC*, without changing in the *De* values. This result indicates that, by changing the value of *τC* as a design parameter for food products, the character of various meats can be significantly altered. The hardness of various meats and salmon, for example, can also be controlled by a change in *τC*, as shown in the

**6.2 Clear discrimination of the two water species due to diffusivity (***De***)** 

1.0E-08

0.0E+00 2.0E-08 4.0E-08

Fig. 19. Comparing behavior of *De* among BA, BH, and PH as a function of τC at 70C.

*Species A1 Species A2*

*τ* C(s)

○

△

□

B A

B H

P H

0.0E+00

1.0E-06

2.0E-06

*De*

(m2/h)

3.0E-06

4.0E-06

5.0E-06

Fig. 17. Dehydration parameters (*k*s and *k*w) as a function of pre-exponential factor (*De*0)(*T*D=45°C) for the salmon.

#### **6. Advantages derived from the hybrid technique**

#### **6.1 Visualized correlation between molecular mobility (C) and physicochemical mobility (***De***)**

Physicochemical mobility, *De*, which is evaluated by the chemical engineering technique (Fig. 12), and molecular mobility, C, which is evaluated by the proton NMR technique (Fig. 8), are important parameters for characterizing the water species retained in foods as the diffusion rate in the three dimensional space of the food tissue matrix and as the rotation time of water molecules, respectively. The two parameters (*De* and C) for the pork meat (PH) and 12 jerkies are interestingly related, as shown in Figure 18. This clearly shows an advantage of the hybrid technique. *De* for PH demonstrates a drastic change at C = 10-8 s,

Fig. 18. *De* as a function of τC for jerkies and PH.

1.E+04 1.E+05 1.E+06 1.E+07 *De* <sup>0</sup> '

Fig. 17. Dehydration parameters (*k*s and *k*w) as a function of pre-exponential factor

**6.1 Visualized correlation between molecular mobility (C) and physicochemical** 

Physicochemical mobility, *De*, which is evaluated by the chemical engineering technique (Fig. 12), and molecular mobility, C, which is evaluated by the proton NMR technique (Fig. 8), are important parameters for characterizing the water species retained in foods as the diffusion rate in the three dimensional space of the food tissue matrix and as the rotation time of water molecules, respectively. The two parameters (*De* and C) for the pork meat (PH) and 12 jerkies are interestingly related, as shown in Figure 18. This clearly shows an advantage of the hybrid technique. *De* for PH demonstrates a drastic change at C = 10-8 s,

PJ-4

BJ-4 BJ-6

0.0E+00 5.0E-08 1.0E-07

*τ* C (s)

PJ-3

0.0E+00

Fig. 18. *De* as a function of τC for jerkies and PH.

2.0E-07

4.0E-07

6.0E-07

*De* (m2/h) 8.0E-07

1.8E-06

1.0E-06

2.0E-06

1.2E-06

2.2E-06

PJ-5 PJ-6

BJ-2 BJ-3

*Species A1 Species A2*

BJ-1 BJ-5

PJ-1 PJ-2

Reference-PH

k *S* k*W*

1.E+01

**6. Advantages derived from the hybrid technique** 

1.E+02

1.E+03

*ks* or *kw*(h-1)

(*De*0)(*T*D=45°C) for the salmon.

**mobility (***De***)** 

1.E+04

1.E+05

indicating an identical *De* = 2.2(±0.2)×10-6 m2/h in the species A1 region and *De* = 3.5(±1.0) ×10-7 m2/h in the water species A2 region. All the *De* values for the commercial jerkies are distributed in the water species A2 region, indicating an identical *De* = 3.0(±1.8)×10-7 m2/h, which, even though the data are widely scattered, is the same as PH. Regarding the jerky data, although all the samples distributed in the market commonly use the water species A2 not A1 and the effective diffusivity, *De*, falling in the range of 1.0~5.0×10-7 m2/h, C is widely distributed in the range of 10-8 ~ 1.1×10-7 s. Since each of the jerkies used in this study has own taste different from other, the wide distribution of C has an important meaning in developing different tastes in jerky. This idea is also supported in an upcoming section.

#### **6.2 Clear discrimination of the two water species due to diffusivity (***De***)**

As has been demonstrated in Fig. 12, the *De*~*W*0 plots could not clearly discriminate the water species A1 and A2. The *De*~*τC* plots derived from the hybrid technique, on the other hand, are conveniently able to visualize the discrimination between beef and pork and the two water species, A1 and A2, as shown in Figure 19. In particular, the boundary of the two water species is clearly recognized as C*τC* =10-8 s, indicating an advantage of the hybrid technique. Focusing on the behavior of BA, BH, and PH, the *De* values gave a constant value in the two water species regions: in the A1 region, as 3.1(±0.5) ×10-6, 4.7(±0.1) ×10-6, and 2.2(±0.3) ×10-6 m2/h, respectively; in the A2 region, as 1.8(±0.5) ×10-6, 2.0(±0.4) ×10-6, and 7.0(±1.0) ×10-7 m2/h, respectively, even though *τC* is significantly changed from 10-8 to 4.5×10-8 s. These results suggest the water species A2 has the interesting characteristic of being able to make a wide shift of the molecular mobility, *τC*, without changing in the *De* values. This result indicates that, by changing the value of *τC* as a design parameter for food products, the character of various meats can be significantly altered. The hardness of various meats and salmon, for example, can also be controlled by a change in *τC*, as shown in the next section.

Fig. 19. Comparing behavior of *De* among BA, BH, and PH as a function of τC at 70C.

Challenging Evaluation of the Hybrid Technique of

labyrinth factor, and ΔS is activation entropy.

0.0E+00

5.0E-04

*δD*

0(m2 h) /

1.0E-03

1.5E-03

2.0E-03

0

a function of C.

The *δD*0 ~

demonstrated a drastic reduction at C

Chemical Engineering – Proton NMR Technique for Food Engineering 87

χ, and ∆S. Since the *δD*0 value can be evaluated from the extrapolation of the Arrhenius plots against the perpendicular axes in Fig. 13, those evaluated for BA, BH, and PH can be plotted as a function of *τC*, as shown in Figure 21. Three *δD*0 curves obtained demonstrate a steep decay at the C*τC* = 10-8 s. Based on the rough evaluation of the steep decay for the three curves to be 1/30 ~1/150, the reduction of the *δD*0 values should be attributed to a simultaneous change in the ε value, which becomes smaller while the χ value becomes larger and the ∆S value becomes smaller because of the steep change in the physical structure of the BA, BH, and PH's tissue matrix at the C*τC* = 10-8 s derived from dehydration.

where *k* is Boltzmann constant, *h* is Planck constant, *δ*(=*ε*/*χ*) is diffusibility, *ε* is porosity, *χ* is

*Species A1 Species A2*

*<sup>E</sup> k T S E De D exp exp exp R T <sup>h</sup> R RT* 

D D 273

D D D

0.0E+00 1.0E-08 2.0E-08 3.0E-08 4.0E-08 τC (s)

*<sup>C</sup>* plots again demonstrate the hybrid technique to be a useful tool for

*<sup>C</sup>* in the water species A2 region. This reduction of *E*<sup>D</sup>

Fig. 21. Comparing behavior of the pre-exponential factor between BA, BH, and PH at 50C as

Figure 22 illustrates the optical microscope photographs of the cross-section for BH. Comparing Figs. 22(A) and (B), one can clearly recognize that a large number of pores (complicated structures) appeared in the sample of *W*0 = 54%-d.b. rather than that of *W*0 = 127%-d.b., indicating a growth of the labyrinth factor (*χ*). This growth of *χ* contributes to the reduction of *δ* and is supported by the experimental evidences of the steep reductions of

discriminating between the two water species and among the kind of foods. The drastic change of *δD*0 should derive the change of water species from A1 to A2, suggesting the diffusion mechanism change between the two regions. This idea is strongly supported by experimental evidence as shown in Figure 23. All *E*D's of water species A2 for BA, BH, and PH

strongly suggests a change in the diffusion mechanism from the water species A1 region.

both *δD*0 in Fig. 21 and *De* in Fig. 19 in the water species A2 region.

273 273

○ *BA* △ *B <sup>H</sup>* □ *P <sup>H</sup>*

(10)

#### **6.3 Clear discrimination of the two water species by hardness**

Figure 20 demonstrates the hardness of BA, BH, PH, and salmon as a function of *τC*. The hardness of all the samples shows a drastic increase at C*τC* = 10-8 s with increasing *τC* in the water species A2 region. This clear discrimination between the water species A1 and A2 is an advantage of the hybrid technique. One may recognize that the *N*P value in the water species A1 region commonly showed 1.0(±0.5) ×106 N/m2 regardless of *τC* in all the samples, whereas in the water species A2 region and at the range of *τC* = 3×10-8 ~ 6×10-8 s, it varied widely depending on the samples, as with *N*P = 1.5×107 for BA, 1.3×107 for BH, and 1×107 N/m2 for PH at *τC* = 5.0×10-8 s. From these results, one can recognize that it is possible to design food products with different *N*P values without changing the *τC* values, and it's possible to design products with different *τC* values without changing the *N*P values. Keeping *N*P = 1.0×107 N/m2, for example, one can have different *τC* values such as *τC* = 2.9×10-8 s for BA, *τC* = 3.5×10-8 s for BH, and *τC* = 4.8×10-8 s for PH. Although the physical meaning of this *τC* value shift is still unclear, it can be attributed to the taste difference between BA, BH, and PH because concrete evidence that the *τC* value changes depending on the kind of food, the concentration of seasonings, and the kind of seasonings has been obtained. Details of these results will be reported in an upcoming work (Konishi and Kobayashi, 2011).

The visualization of *N*P ~*τC* exhibited in Fig. 20 clearly demonstrates the hybrid technique as a useful tool for discriminating between the two water species and among the kinds of food.

Fig. 20. Comparing behavior of *N*P as a function of C between BA, BH, and PH at 50C.

#### **6.4 Physical meaning of CC = 10-8 s derived from the pre-exponential factor (***De***<sup>0</sup> ) of** *De*

Focusing on the boundary (C*τC* = 10-8 s) between water species A1 and A2, one may be interested in what is happening in the food tissue matrix at the C*τC* = 10-8 s. The information for the food tissue matrix can be obtained from Equation (2), and the equation is rewritten by Equation (10). The pre-exponential factor (*δD*0) of Equation (10) consists of a function of ε,

Figure 20 demonstrates the hardness of BA, BH, PH, and salmon as a function of *τC*. The hardness of all the samples shows a drastic increase at C*τC* = 10-8 s with increasing *τC* in the water species A2 region. This clear discrimination between the water species A1 and A2 is an advantage of the hybrid technique. One may recognize that the *N*P value in the water species A1 region commonly showed 1.0(±0.5) ×106 N/m2 regardless of *τC* in all the samples, whereas in the water species A2 region and at the range of *τC* = 3×10-8 ~ 6×10-8 s, it varied widely depending on the samples, as with *N*P = 1.5×107 for BA, 1.3×107 for BH, and 1×107 N/m2 for PH at *τC* = 5.0×10-8 s. From these results, one can recognize that it is possible to design food products with different *N*P values without changing the *τC* values, and it's possible to design products with different *τC* values without changing the *N*P values. Keeping *N*P = 1.0×107 N/m2, for example, one can have different *τC* values such as *τC* = 2.9×10-8 s for BA, *τC* = 3.5×10-8 s for BH, and *τC* = 4.8×10-8 s for PH. Although the physical meaning of this *τC* value shift is still unclear, it can be attributed to the taste difference between BA, BH, and PH because concrete evidence that the *τC* value changes depending on the kind of food, the concentration of seasonings, and the kind of seasonings has been obtained. Details of these results will be reported in an upcoming work (Konishi and

The visualization of *N*P ~*τC* exhibited in Fig. 20 clearly demonstrates the hybrid technique as a useful tool for discriminating between the two water species and among the kinds of food.

*Species A1 Species A2*

0.0E+00 2.0E-08 4.0E-08 6.0E-08 *τ* C (s)

Fig. 20. Comparing behavior of *N*P as a function of C between BA, BH, and PH at 50C.

**6.4 Physical meaning of CC = 10-8 s derived from the pre-exponential factor (***De***<sup>0</sup>**

Focusing on the boundary (C*τC* = 10-8 s) between water species A1 and A2, one may be interested in what is happening in the food tissue matrix at the C*τC* = 10-8 s. The information for the food tissue matrix can be obtained from Equation (2), and the equation is rewritten by Equation (10). The pre-exponential factor (*δD*0) of Equation (10) consists of a function of ε,

○ *BA* △ *B <sup>H</sup>* □ *P <sup>H</sup> Salmon*

**) of** *De*

**6.3 Clear discrimination of the two water species by hardness** 

Kobayashi, 2011).

0.0E+00

5.0E+06

1.0E+07

*N*P(N/m2

)

1.5E+07

χ, and ∆S. Since the *δD*0 value can be evaluated from the extrapolation of the Arrhenius plots against the perpendicular axes in Fig. 13, those evaluated for BA, BH, and PH can be plotted as a function of *τC*, as shown in Figure 21. Three *δD*0 curves obtained demonstrate a steep decay at the C*τC* = 10-8 s. Based on the rough evaluation of the steep decay for the three curves to be 1/30 ~1/150, the reduction of the *δD*0 values should be attributed to a simultaneous change in the ε value, which becomes smaller while the χ value becomes larger and the ∆S value becomes smaller because of the steep change in the physical structure of the BA, BH, and PH's tissue matrix at the C*τC* = 10-8 s derived from dehydration.

$$De = \delta D\_0 \cdot \exp\left(\frac{-E\_\mathrm{D}}{R \cdot \left(T\_\mathrm{D} + 273\right)}\right) = \left(\frac{\varepsilon}{\chi}\right) \cdot \left(\frac{k \cdot \left(T\_\mathrm{D} + 273\right)}{h}\right) \left[\exp\left(\frac{\Delta S}{R}\right)\right] \cdot \exp\left(\frac{-E\_\mathrm{D}}{R \cdot \left(T\_\mathrm{D} + 273\right)}\right) \tag{10}$$

where *k* is Boltzmann constant, *h* is Planck constant, *δ*(=*ε*/*χ*) is diffusibility, *ε* is porosity, *χ* is labyrinth factor, and ΔS is activation entropy.

Fig. 21. Comparing behavior of the pre-exponential factor between BA, BH, and PH at 50C as a function of C.

Figure 22 illustrates the optical microscope photographs of the cross-section for BH. Comparing Figs. 22(A) and (B), one can clearly recognize that a large number of pores (complicated structures) appeared in the sample of *W*0 = 54%-d.b. rather than that of *W*0 = 127%-d.b., indicating a growth of the labyrinth factor (*χ*). This growth of *χ* contributes to the reduction of *δ* and is supported by the experimental evidences of the steep reductions of both *δD*0 in Fig. 21 and *De* in Fig. 19 in the water species A2 region.

The *δD*0 ~ *<sup>C</sup>* plots again demonstrate the hybrid technique to be a useful tool for discriminating between the two water species and among the kind of foods. The drastic change of *δD*0 should derive the change of water species from A1 to A2, suggesting the diffusion mechanism change between the two regions. This idea is strongly supported by experimental evidence as shown in Figure 23. All *E*D's of water species A2 for BA, BH, and PH demonstrated a drastic reduction at C*<sup>C</sup>* in the water species A2 region. This reduction of *E*<sup>D</sup> strongly suggests a change in the diffusion mechanism from the water species A1 region.

Challenging Evaluation of the Hybrid Technique of

values accompanied by drastic increases or decreases.

food products requested by the commercial fields.

BJ-3

0.4

0.5

0.6

0.7

aw (-)

**8. Conclusions** 

reproducible.

*N*P, and *W*0 .

0.8

0.9

1.0

PJ-5

Fig. 24. aw as a function of τC for the beef and pork jerkies.

BJ-2 BJ-1

*Group-3*

PJ-6

BJ-4

PJ-4

BJ-6

BJ-5

PJ-2

*Group-2*

*Group-1*

PJ-1

0.0E+00 5.0E-08 1.0E-07 1.5E-07 C (s)

The hybrid of the proton NMR technique and the chemical engineering technique was applied to discriminate the 12 commercially distributed jerkies, salmon, BA, BH, and PH and to distinguish between the two water species, A1 and A2, by using four parameters: *De*, *τ*C,

1. The dehydration and adsorption of the water species, A1 and A2, retained in the foods were reversibly repeated and the transformation between the two water species was

2. The proton NMR technique and the chemical engineering technique could respectively not discriminate among the *τ*C ~ *W*0 curves and among the aw ~ *W*0 curves obtained for the foods. Although the computer simulation model proposed by the chemical engineering technique generated a good fitting against the dehydration response

Chemical Engineering – Proton NMR Technique for Food Engineering 89

A1 to A2 and the pore structure of the food tissue matrix provides regulation of ε and χ

Focusing on the three straight lines in Fig. 24, line-1 generates Group-1 (BJ-5 and PJ-2), line-2, Group-2 (BJ-4, BJ-1, PJ-4, BJ-6, and PJ-1), and line-3, Group-3 (PJ-5, BJ-3, BJ-2, and PJ-6). Based on the experimental data presented here, although one cannot demonstrate an exact physical meaning of the three groups' difference, it might be presumed that each group should have the specified seasonings different from others. Concerning the seasonings in the three groups, Group-1 is characterized by soy sauce, Group-2, by a mixed seasoning such as soy sauce, smoke flavoring, and salt, and Group-3, by a mixed seasoning such as salt, soysauce, and miso sauce. Although each of the three groups could not be distinguished by the chemical engineering (the aw ~ *W*0 curves and the *De* ~ *W*0 curves) and proton NMR techniques (the C ~ *W*0 curves), the hybrid technique (the aw ~ *τ*C linear relation) clearly distinguished the characteristics of the seasonings for the 12 different food products. This evidence strongly demonstrates the hybrid technique to be a useful tool to design various

Fig. 22. Optical microscope photographs (× 1000) of the cross-section for PH, (A): 127%-d.b., (b): 54%-d.b.

Fig. 23. Comparing behavior of *E*D among BA, BH, and PH as a function of C at 50~70C.

#### **7. A new method for jerky product design using the parameters induced by the hybrid technique**

Our interest is focused on a classification of the 12 jerky products influenced by multifunctional water species. Hills *et al.* (1999) empirically demonstrated a linear relation between the NMR relaxation rates (1/*T*2) and aw. In addition, Hills (1999) derived a theoretical equation to explain the linear relation. Since the relaxation time, *T*2, can be replaced by C using Equation (3), it is proposed that aw should be related to C. Figure 24 demonstrates aw as a function of C for the 12 jerky samples. All the samples fell roughly around three straight lines, indicating the existing of three groups. The three lines started from aw = 0.92, which is an identical value independent of the samples and empirically evaluated, meaning a specified value for all kinds of foods (Konishi and Kobayashi, 2010). The aw = 0.92 value always gave C = 10-8 s independent of the kinds of foods and seasoning species. As has been discussed in the previous section, one can recognize both aw = 0.92 and C = 10-8 s to be critical values for foods in which the water species drastically changes from A1 to A2 and the pore structure of the food tissue matrix provides regulation of ε and χ values accompanied by drastic increases or decreases.

Focusing on the three straight lines in Fig. 24, line-1 generates Group-1 (BJ-5 and PJ-2), line-2, Group-2 (BJ-4, BJ-1, PJ-4, BJ-6, and PJ-1), and line-3, Group-3 (PJ-5, BJ-3, BJ-2, and PJ-6). Based on the experimental data presented here, although one cannot demonstrate an exact physical meaning of the three groups' difference, it might be presumed that each group should have the specified seasonings different from others. Concerning the seasonings in the three groups, Group-1 is characterized by soy sauce, Group-2, by a mixed seasoning such as soy sauce, smoke flavoring, and salt, and Group-3, by a mixed seasoning such as salt, soysauce, and miso sauce. Although each of the three groups could not be distinguished by the chemical engineering (the aw ~ *W*0 curves and the *De* ~ *W*0 curves) and proton NMR techniques (the C ~ *W*0 curves), the hybrid technique (the aw ~ *τ*C linear relation) clearly distinguished the characteristics of the seasonings for the 12 different food products. This evidence strongly demonstrates the hybrid technique to be a useful tool to design various food products requested by the commercial fields.

Fig. 24. aw as a function of τC for the beef and pork jerkies.

#### **8. Conclusions**

88 Advances in Chemical Engineering

(A) (B)

Fig. 22. Optical microscope photographs (× 1000) of the cross-section for PH, (A): 127%-d.b.,

○ *BA* △ *B <sup>H</sup>* □ *P <sup>H</sup>*

*Species A1 Species A2*

0.0E+00 2.0E-08 4.0E-08 6.0E-08 τC (s)

Fig. 23. Comparing behavior of *E*D among BA, BH, and PH as a function of C at 50~70C.

**7. A new method for jerky product design using the parameters induced by** 

Our interest is focused on a classification of the 12 jerky products influenced by multifunctional water species. Hills *et al.* (1999) empirically demonstrated a linear relation between the NMR relaxation rates (1/*T*2) and aw. In addition, Hills (1999) derived a theoretical equation to explain the linear relation. Since the relaxation time, *T*2, can be replaced by C using Equation (3), it is proposed that aw should be related to C. Figure 24 demonstrates aw as a function of C for the 12 jerky samples. All the samples fell roughly around three straight lines, indicating the existing of three groups. The three lines started from aw = 0.92, which is an identical value independent of the samples and empirically evaluated, meaning a specified value for all kinds of foods (Konishi and Kobayashi, 2010). The aw = 0.92 value always gave C = 10-8 s independent of the kinds of foods and seasoning species. As has been discussed in the previous section, one can recognize both aw = 0.92 and C = 10-8 s to be critical values for foods in which the water species drastically changes from

(b): 54%-d.b.

0

5

10

*E*

**the hybrid technique** 

D(kJ/mol)

15

20

The hybrid of the proton NMR technique and the chemical engineering technique was applied to discriminate the 12 commercially distributed jerkies, salmon, BA, BH, and PH and to distinguish between the two water species, A1 and A2, by using four parameters: *De*, *τ*C, *N*P, and *W*0 .


Challenging Evaluation of the Hybrid Technique of

*ε porosity of the food tissue* (-)

**Greek letters** 

*C*

analysis.

**11. References** 

349.

Chemistry.

*WR the water ratio* (= (W-We)/(WD-We))

*WR1 the proportion of the strongly restricted water amount* (-) *WR2 the proportion of the weakly restricted water amount* (-) *WRS the ratio of the strongly restricted water species* (-) *WRW the ratio of the weakly restricted water species* (-) *WS the amount of strongly restricted water species* (%-d.b.) *WW the amount of weakly restricted water species* (%-d.b.)

*β1 constant evaluated by the curve fitting simulation* (m-2) *β2 constant evaluated by the curve fitting simulation* (m-2)

 *gyromagnetic ratio of proton* (=2.675108 radT-1s-1)

*ħ modified Plank's constant* (=6.6310-34 Js)

*C critical correlation time of water proton* (s)

*0 resonance frequency* (=3.14109 s-1)

*C correlation time of water proton* (s)

*χ labyrinth factor of the meat tissue* (-)

*δ diffusibility* (=ε /χ)(-)

*u.-Technol.*, 19 , 448-456.

**10. Acknowledgment** 

*π the ratio of the circumference of a circle to its diameter* (=3.14)

Chemical Engineering – Proton NMR Technique for Food Engineering 91

This work was financially supported by the Cooperation of Innovation Technology and Advanced Research in Evolution Area (City Area) from the Japanese Ministry of Education, Culture, Sports, Science, and Technology. The authors wish to thank Associate Professor Koichi Miura, Kitami Institute of Technology, for his assistance with the proton NMR

Abragam, A., (1963). The Principles of Nuclear Magnetism. *Oxford at the Clarend Press*, p347-

Andreu, J. & Stamatopoulos, A.A. (1986). Durum wheat pasta drying kinetics. *Lebensm. Wiss.* 

Belton, P.S.; Gil, A.M.; Webb, G.A., & Rutledge, D., Eds.(2003). Magnetic resonance in food science-latest developments, *The Royal Society of Chemistry, Cambridge*. Chhinnan, M.S. (1984). Evaluation of selected mathematical models for describing thin-layer

Hills, B.P. (1999). NMR studies of water mobility in foods. In"Water Management in the

Hills, B. P., Manning, C. E. & Godward, J. (1999). A multistate theory of water relations in

Design and Distribution of Quality Foods" ISOPOW 7(Y.H. Roos, R.B. Leslie, and

biopolymer systems. In: *Advances in Magnetic Resonance in Food Science* (edited by P. S. Belton, B. P. Hill and G. A. Webb). Pp.45-62. Cambridge, UK: Royal Society of

drying of in-shell pecans. *Trans. ASAE*, 27(2), 610-615.

P.J.Lillford, eds) Technomic Publishing, Lancaster, PA.

Fennema, O.(1976). In principles of food science, Part 1, Marcel Dekker, New York.

curves, the unknown parameters evaluated by the curve fitting, 1 and 2, were recognized as meaningless.


#### **9. Nomenclature**



#### **Greek letters**

90 Advances in Chemical Engineering

3. The hybrid technique proposed demonstrates a clear discrimination between water species A1 and A2, divided at the critical value of C*τ*C = 10-8 s, where the drastic change in the values of *De*, *E*D, and *N*P appeared. The physical meaning of C*τ*C was understood as a drastic change in the pre-exponential factor of *De* derived from the change in the

4. Using the hybrid technique, a variety of food products could easily be designed by changing the value of *τ*C even though the values of aw are identical. The meat jerkies commercially distributed in Japan were reasonably classified into three groups characterized by the three aw ~ *τ*C straight lines. The advantage of the hybrid technique brought a conclusion that the aw ~ *τ*C straight lines give characteristic slopes depending on the seasonings of the 12 jerkies. This could be useful in designing a variety of food

1 and 2, were

curves, the unknown parameters evaluated by the curve fitting,

recognized as meaningless.

products.

**9. Nomenclature** 

food tissue matrix due to dehydration.

*BA beef meat produced in Australia* (-) *BH beef meat produced in Hokkaido, Japan* (-) *D moisture diffusion coefficient* (m2/h) *D0 frequency factor of D* (m2/h)

*H Planck constant* (erg. s)

*K Boltzmann constant* (erg/deg)

*Mt magnitude of magnetization vector* (-)

*R gas constant* (=8.314J/K·mol)

*ΔS activation entropy* (kJ/mol)

*ts spin locking pulse length* (s)

*TD drying temperature* (C)

*t drying time* (h)

*NP hardness of meat products* (Newton/m2) *PH pork meat produced in Hokkaido, Japan* (-)

*T2 spin-spin relaxation time of water proton* (s)

*W water content at the drying time t* (%-d.b.)

*We equilibrium water content* (%-d.b.)

*De effective water diffusion coefficient* (m2/h) *De0 pre-exponential factor of De* (PF= D0 , m2/h) *ED activation energy of water diffusivity* (kJ/mol)

*I nuclear spin quantum number of water proton* (= 0.5) (--)

*kS mass transfer coefficient of strongly restricted water* (h-1) *kw mass transfer coefficient of weakly restricted water* (h-1) *La half distance of a-axis of the rectangular sample* (m) *Lb half distance of b-axis of the rectangular sample* (m) *Lc half distance of c-axis of the rectangular sample* (m)

*M0 magnitude of macroscopic equilibrium magnetisation vector* (-)

*r proton-proton distance of water molecule* (= 0.16 nm)

*W0 initial water content at the time of PUP operated* (%-d.b.) *WD initial water content of drying flesh sample* (%-d.b.)


#### **10. Acknowledgment**

This work was financially supported by the Cooperation of Innovation Technology and Advanced Research in Evolution Area (City Area) from the Japanese Ministry of Education, Culture, Sports, Science, and Technology. The authors wish to thank Associate Professor Koichi Miura, Kitami Institute of Technology, for his assistance with the proton NMR analysis.

#### **11. References**


**4** 

*1Mexico 2UK* 

**Modelling Approach for** 

*2School of Electrical and Electronic Engineering* 

**Redesign of Technical Processes** 

*1Information Technology Laboratory, Cinvestav-Tamaulipas* 

*University of Nottingham, University Park, Nottingham* 

Ivan Lopez-Arevalo1, Victor Sosa-Sosa1 and Saul Lopez-Arevalo2

Nowadays the design and development of new products or modification of existent ones (redesign) is a key and fundamental element to enhance innovation and competitiveness of industrial companies. Design has an increasing importance to differentiate one product from

In general, design is the process of specifying a description of a product that satisfies a set of requirements (Umeda 90). Redesign is the process of changing the description of an existent product to satisfy a new set of requirements (Brown 98). Design engineering includes both design and redesign. In the literature we can find diverse terms to narrow design and redesign, such as preliminary, conceptual, functional, creative, routinary, non-routinary, personified, parametric, innovative, etc., but the characteristic activities of the global design

 Conceptual (re)design, the phase where the global goals, requirements and operation of the product are established based on abstract concepts. The research presented in this

engineering can be divided as follows [Subba-Rao 99], see Figure 1:

**1. Introduction** 

Fig. 1. Product design path.

chapter deals with this aspect.

another.


## **Modelling Approach for Redesign of Technical Processes**

Ivan Lopez-Arevalo1, Victor Sosa-Sosa1 and Saul Lopez-Arevalo2 *1Information Technology Laboratory, Cinvestav-Tamaulipas 2School of Electrical and Electronic Engineering University of Nottingham, University Park, Nottingham 1Mexico* 

*2UK* 

#### **1. Introduction**

92 Advances in Chemical Engineering

Jason, A.C.(1958). A Study of evaporation and diffusion processes in the drying of fish

Kannan, D. and Bandyopadhyay, S. (1995). Drying characteristics of a tropical marine fish

Konishi, Y.; Horiuchi, J., & Kobayashi, M.(2001). Dynamic evaluation of the dehydration

Konishi, Y.; Horiuchi, J., & Kobayashi, M.(2001). Dynamic evaluation of the dehydration

Konishi, Y.;Miura, K. & Kobayashi, M.(2003). Drying efficiency design using multifunctional

Konishi, Y. & Kobayashi, M.(2009). Quantitative evaluation of the design parameters requested in beef and pork drying operation, *AIDIC Conference Series,* 9, 177-186. Konishi, Y., & Kobayashi, M., (2010). Food product design using the water species as a probe molecule. Annual meeting of Japanese Society of Food Engineering. Konishi, Y., Kobayashi, M., & Miura, K. (2010). Characeterization of water species revealed

analysis, *International Journal of Food Science and Technology.*, 45.1889-1894. Konishi, Y. & Kobayashi, M.(2011). Dynamism of the water species as a probe molecule in food, *Chemical Engineering Transactions*, Ed. Sauro Pierucci, 24, 475-480. Labuza, T.P., Tannenbaum, S.R., & Karel,M. (1970), Water Content and Stability of Low moisture and intermediate moisture foods. Food Technol.24, 543-550. Litchfield, J.B. & Okos, M.R. ,1992. Moisture diffusivity in pasta during drying. *J. Food* 

Madamba, P. S., Driscoll, R. H. & Buckle, K. A.(1996). The thin-layer drying characteristics of

Noomhorm, A. & Verma, L.R.(1986). Generalized single layer rice drying models. *Trans.* 

Rockland, L.B. and Stewart, G.F., Eds.(1981). International symposium on properties of water, water activity: Influences on food quality, *Academic Press Inc., London.*  Schmidt, S.J., Water and Solids Mobility in Foods, (2004). Advances in Food and Nutrition

Thompson, T.L., Peart, R.M. & Foster, G.H. (1968). Mathematical simulation of corn drying

Waananen, K.M. & Okos, M.R. (1996). Effect of porosity on moisture diffusion during

research, volume 48, 2004, Edited by S.L. Taylor, Elsevier

Chemical Industry. *McMillan, London*, pp.103-134.

squid, *AIDIC Conference series,* vol.6, 183-190.

slab. *J.Food Sci. Technol.*, 32(1), 13-16.

*Technology*, 19(7), 1253-1269

1271-1285.

*Engng.*, 17, 117-142.

*ASAE.*, 29(2), 587-591.

garlic slices. *J. Food Engng.*, 29, 75-97.

(A new model). *Trans.ASAE.*,24(3),582.

drying of pasta. *J. Food Engng.*, 28, 121-137.

muscle in fundamental aspects of dehydration of food stuffs, ed. Society of

response curves of foods characterized by a poultice-up process using a fish-paste sausage-I. Determination of the mechanisms for moisture transfer, *Drying* 

response curves of foods characterized by a poultice-up process using a fish-paste sausage-II, A new tank model for a computer simulation, *Drying Technology*, 19(7),

dynamics of water molecules in foods-H-NMR analysis of a fish paste sausage and

in the drying operation of Todarodes pacificus Steenstrup using water proton NMR

Nowadays the design and development of new products or modification of existent ones (redesign) is a key and fundamental element to enhance innovation and competitiveness of industrial companies. Design has an increasing importance to differentiate one product from another.

In general, design is the process of specifying a description of a product that satisfies a set of requirements (Umeda 90). Redesign is the process of changing the description of an existent product to satisfy a new set of requirements (Brown 98). Design engineering includes both design and redesign. In the literature we can find diverse terms to narrow design and redesign, such as preliminary, conceptual, functional, creative, routinary, non-routinary, personified, parametric, innovative, etc., but the characteristic activities of the global design engineering can be divided as follows [Subba-Rao 99], see Figure 1:

Fig. 1. Product design path.

 Conceptual (re)design, the phase where the global goals, requirements and operation of the product are established based on abstract concepts. The research presented in this chapter deals with this aspect.

Modelling Approach for Redesign of Technical Processes 95

4. Human designers can understand the process intuitively identifying its functional

The chapter is organised as follows. In section 2 the contextualisation background about redesign and modelling is given. Section 3 describes the proposed modelling approach for redesign. Section 4 shows the experimentation carried out and the obtained results of a first prototype for the proposed approach. Finally, the section 5 presents some conclusions and

The research work related to the process of redesign is huge. Following the different research approaches are presented from a general point of view to more specific one. Some of the most interesting definitions of design that we have found in the literature are

Design is formally a search problem in a large space of objects that satisfy multiple

Design is the task of devising courses of action to change or create better ones (Simon

 Design starts with an intended activity or use (Maher 97b) and uses available knowledge to arrive at a description of an artefact which will produce those results

Defining design is difficult because the term refers both to a product (the object to be designed) and a process (the process of design). The reasoning process involved in design allows moving from a functional concept as a starting point to a product solution. Therefore, the design activity can be seen as an activity of synthesis, which is strongly influenced by

Regarding the process of redesign we have identified the following definitions in the literature: Redesign is considered as design in which there is a priori knowledge on the general and specialised functions to be performed and on the working principles to be selected

 Redesign is an inherent part of most design processes; in which new requirements or new domain knowledge influence the original design process (Brazier 96); but can also

Redesign is part of design, which proposes suitable modifications free for

Independently of the point of view of redesign, three types of redesign can be identified (Dixon 89): parametric redesign, component redesign, and structural redesign. In order to

2. Complex numerical simulators can be used to model the behaviour of the process. 3. The process is already implanted, which means there is a design solution that satisfies

1. The complexity of the process must be high.

the original requirements of such process.

5. The process can be represented by functional abstract concepts.

sections.

**2. Background** 

summarised as:

96).

(Gero 90b).

[Salomons 95].

**2.1 The process of redesign** 

constraints (Chandrasekaran 90).

the skills and mental models of the designer.

be seen as a family of design methods in itself (Pos 97).

inconvenience of existent artefacts (Kitamura 99).

remarks.

 Detailed (re)design, the phase where the results of the conceptual design are used to physically implement a product.

Design engineering involves a wide range of activities. Design engineering can appear in a broad variety of domains, from the assembly of brakes to complex industrial plants and from simple chips to the most advanced super computers. Both design and redesign consist of two main components: the (re)design process and the (re)design object. The (re)design process involves all the (re)design activities performed over the (re)design object, which is the subject entity to be (re)designed. In engineering domains is common to refer to the (re)design object as the artefact. An artefact is a type of product to denote physical and technical devices. The (re)design process over an artefact is characterised by the map between functional requirements to structural requirements. Design and redesign, both can be considered as a dialectic process between goals (what it is desired) and possibilities (real constraints), directed to the satisfaction of functional specifications and performance (Stephanopoulos 90b).

The industry deals with complex technical processes where their behaviour is only predicted by means of complex numerical simulators. The redesign of a process is sometimes necessary when certain time has passed from its implantation or when they must adapt to economical, technological, or environmental requirements. The redesign is not part of the maintenance stage but must be considered into the process life cycle.

From a general point of view, the redesign is done typically in three steps: design-description acquisition (modelling), problem analysis (diagnosis) and proposal of modifications (generation of alternatives). In real redesign situations, human designers intuitively create mental abstract models by removing superfluous information about the process. Such models are based on functions of the components of the process and its context.

From the early 60's, Artificial Intelligence techniques have been used for design, such as constraint-based systems, case-based reasoning, model-based reasoning, planning, neural networks, and genetic algorithms. Although in these approaches the modelling and simulation of the processes has been solved in acceptable way, another problem has been generated, the used knowledge representations require so detailed information that sometimes it is difficult to understand.

The main objective of this work is to obtain a support framework to assist the human designer in the redesign of complex technical processes. The structure of this framework must be based on the common redesign activities performed by human designers on real redesign situations. Therefore, the framework must able to reduce the complexity of the processes to be redesigned, and therefore facilitate the redesign activities.

The framework was obtained integrating model-based reasoning and case-based reasoning techniques. Using model-based reasoning the original process can be modelled hierarchically. Using case-based reasoning alternative process parts can be obtained from other processes, which have to be adapted into the original process. The framework was implemented in the Chemical Engineering domain due to the complexity of the processes involved and the interaction with experts in the area.

The proposed redesign framework has to be able to deal with complex technical processes. In this sense, the type of processes we are referring to follow some assumptions:


The chapter is organised as follows. In section 2 the contextualisation background about redesign and modelling is given. Section 3 describes the proposed modelling approach for redesign. Section 4 shows the experimentation carried out and the obtained results of a first prototype for the proposed approach. Finally, the section 5 presents some conclusions and remarks.

### **2. Background**

94 Advances in Chemical Engineering

Detailed (re)design, the phase where the results of the conceptual design are used to

Design engineering involves a wide range of activities. Design engineering can appear in a broad variety of domains, from the assembly of brakes to complex industrial plants and from simple chips to the most advanced super computers. Both design and redesign consist of two main components: the (re)design process and the (re)design object. The (re)design process involves all the (re)design activities performed over the (re)design object, which is the subject entity to be (re)designed. In engineering domains is common to refer to the (re)design object as the artefact. An artefact is a type of product to denote physical and technical devices. The (re)design process over an artefact is characterised by the map between functional requirements to structural requirements. Design and redesign, both can be considered as a dialectic process between goals (what it is desired) and possibilities (real constraints), directed to the satisfaction of functional specifications and performance

The industry deals with complex technical processes where their behaviour is only predicted by means of complex numerical simulators. The redesign of a process is sometimes necessary when certain time has passed from its implantation or when they must adapt to economical, technological, or environmental requirements. The redesign is not part

From a general point of view, the redesign is done typically in three steps: design-description acquisition (modelling), problem analysis (diagnosis) and proposal of modifications (generation of alternatives). In real redesign situations, human designers intuitively create mental abstract models by removing superfluous information about the process. Such models are based on functions of the components of the process and its context. From the early 60's, Artificial Intelligence techniques have been used for design, such as constraint-based systems, case-based reasoning, model-based reasoning, planning, neural networks, and genetic algorithms. Although in these approaches the modelling and simulation of the processes has been solved in acceptable way, another problem has been generated, the used knowledge representations require so detailed information that

The main objective of this work is to obtain a support framework to assist the human designer in the redesign of complex technical processes. The structure of this framework must be based on the common redesign activities performed by human designers on real redesign situations. Therefore, the framework must able to reduce the complexity of the

The framework was obtained integrating model-based reasoning and case-based reasoning techniques. Using model-based reasoning the original process can be modelled hierarchically. Using case-based reasoning alternative process parts can be obtained from other processes, which have to be adapted into the original process. The framework was implemented in the Chemical Engineering domain due to the complexity of the processes

The proposed redesign framework has to be able to deal with complex technical processes.

In this sense, the type of processes we are referring to follow some assumptions:

of the maintenance stage but must be considered into the process life cycle.

processes to be redesigned, and therefore facilitate the redesign activities.

involved and the interaction with experts in the area.

physically implement a product.

(Stephanopoulos 90b).

sometimes it is difficult to understand.

#### **2.1 The process of redesign**

The research work related to the process of redesign is huge. Following the different research approaches are presented from a general point of view to more specific one. Some of the most interesting definitions of design that we have found in the literature are summarised as:


Defining design is difficult because the term refers both to a product (the object to be designed) and a process (the process of design). The reasoning process involved in design allows moving from a functional concept as a starting point to a product solution. Therefore, the design activity can be seen as an activity of synthesis, which is strongly influenced by the skills and mental models of the designer.

Regarding the process of redesign we have identified the following definitions in the literature:


Independently of the point of view of redesign, three types of redesign can be identified (Dixon 89): parametric redesign, component redesign, and structural redesign. In order to

Modelling Approach for Redesign of Technical Processes 97

(Bridge 97). Akin (Akin 82) outlines that the representational aspects to determine the utility

 The represented information must be in a level of abstraction suitable for its intention. The contents must be in such a way that they are compatible with the expected results

A substantial amount of research has focused on defining models of design (French 85, Tomiyama 87, Treur 89, Brown 89, Chandrasekaran 90, Gero 90a, Takeda 90a, Alberts 92, Vescovi 93, Ohsuga 97, Brown 97). Most of this research highlights that the modelling of the functionality (or properties) of the design object description is an important aspect of the

It is possible to represent explicit knowledge in (re)design by means of modelling functions of artefacts. This facilitates the systematisation of the reasoning and some tasks of (re)design. The reasoning based on functions allows abstracting information of the design on the same way as it is made in the reasoning of the initial stages of the design. The process of design of an artefact starts with the conceptual or functional design followed by the basic design and the detailed design (Stephanopoulos 90a). Within these, the functional design plays the central role since it guarantees the quality of the design and the innovation of the product (Umeda 97, Culley 99). The idea of function is fundamental in design since the work of the designers is to design artefacts that must achieve explicit functions (Chandrasekaran 00). Functional modelling is useful to model the object of (re)design, this modelling of objects enhances the formulation of (re)design strategies and the overall (re)design process. Functional modelling "hides" sections of the artefact structure at a lower abstraction level

Most of the research work on (re)design considers redesign as a knowledge-intensive field; wherein the processes (e.g., tasks) performed, descriptions of sequencing of processes, descriptions of the information within the system, and knowledge employed to perform a task are explicitly modelled most of the times by means of knowledge-based systems. These modelling frameworks try to model the (re)design so the (re)design object as well as the (re)design process are understandable by humans. To do this, human designers use the object specifications to propose a reasonable (re)design approach need to be understood (Leveson 00). Reasoning strategies employed in (re)design are derived or extensions of the commonly named *problem-solving strategies*. Examples of strategies are hypothesis and test (Hempel 66, White 05), pattern recognition (Doyle 62, Kirsch 64, Mitchell 97), skeletal plan refinement (Friedland 85, Tu 89), heuristic classification (Clancey 85), propose and revise (Goel 89), propose critique modify (Chandrasekaran 90), decision tree search (Raiffa 68, Qi 92), means-ends analysis (Newell 63, Rasmussen 86), and reasoning by analogy (Gick 80,

In the above strategies, the human designer needs to formulate an explicit model of expertise as an integration of two types of models: a domain model and problem solving method. The domain model corresponds to the (re)design object and the problem solving method model corresponds to the (re)design process. Work on domain modelling has only recently attracted the attention of knowledge based system researchers (Stephanopoulos 90a, Schoen 91, Gruber 93, Skuce 93, Sowa 95, Kitamura 98, Fensel 01b, Gomez-Perez 04).

according to the mental representations of the designer. The model must be consistent with the reality that tries to reflect.

facilitating the manipulation of the artefact description.

of a model in design are:

overall design process.

Gentner 83).

perform any of these redesign types, it is essential that some form of knowledge is available that allows the adaptation of existing designs. According to Pos (Pos 97) and based on the previous mentioned definitions of design/redesign, is possible distinguish two general points of view about the relationship between design and redesign, these are:


Adopting any of the above points of view, basically minimal differences can be distinguished. In both cases the important issue is to bridge the gap between a set of requirements and an existing design description. We can see that design starts from scratch, however, redesign starts with an existing design description, which is modified until it fulfills the current requirements. Both points of view can be captured by a single spectrum of problem-solving methods for redesign.

Several authors (Akin 82, Chandrasekaran 93, Eldonk 96, Brazier 96, Bridge 97, Pos 97) state that the required knowledge for redesign is based on the following two principles:


Many systems that solve redesign problems have been described in literature (Mitchell 83, Howe 86, Fischer 87, Daube 89, Goel 91, Smyth 96, Eldonk 96, Kitamura 99). However when one takes a closer look at the different variants of the redesign task, subtle differences exist that have an impact on how the task can be performed and what kinds of knowledge are involved.

There are a variety of research works referring to design or redesign; from (re)design of abstract (for example, components in software engineering) to physical entities (for example, a reactor in chemical engineering), for a general review see (Brown 97), for some details see (Akin 82, Mitchell 83, Howe 86, Fischer 87, Mostow 89, Goel 91, Bras 92, Stroulia 92a, Chandrasekaran 93, French 93, Brazier 96, Eldonk 96, Pos 97, Price 97, Umeda 97, Gero 98, Culley 99, Culley 99, Kitamura 99, Kraslawski 00, Grossmann 00, Arana 01, Maher 01).

Models in design and redesign are particularly important to guarantee they represent the intentions by which they were created. In general, the models are abstractions of the reality that guarantees communication of ideas by joining concepts, aggregations and relations

perform any of these redesign types, it is essential that some form of knowledge is available that allows the adaptation of existing designs. According to Pos (Pos 97) and based on the previous mentioned definitions of design/redesign, is possible distinguish two general

1. Viewing the design as total set which contains redesign as a subset. In order to satisfy this relationship, all the components of the design reasoning process should be satisfied for redesign. However redesign as a specialised subset would not be applicable in the same contexts as the more general notion of design. Here, design is viewed as an iterative process that uses intermediate results to get a final design description that fulfills the requirements. The task of redesign on the basis of a design created earlier produces a new temporary design description that is closer to the specification than the

2. Viewing both design and redesign as independent sets joined by a small common subset. For this relationship to be satisfied there is an expectation that some crossover or overlap will occur, thus only some of the components of design reasoning will be applicable in the redesign context and viceversa. Here, redesign starts with a previously constructed design description and a new set of requirements. The previously constructed design description must now be modified to fulfill the new set of

Adopting any of the above points of view, basically minimal differences can be distinguished. In both cases the important issue is to bridge the gap between a set of requirements and an existing design description. We can see that design starts from scratch, however, redesign starts with an existing design description, which is modified until it fulfills the current requirements. Both points of view can be captured by a single spectrum

Several authors (Akin 82, Chandrasekaran 93, Eldonk 96, Brazier 96, Bridge 97, Pos 97) state

Many systems that solve redesign problems have been described in literature (Mitchell 83, Howe 86, Fischer 87, Daube 89, Goel 91, Smyth 96, Eldonk 96, Kitamura 99). However when one takes a closer look at the different variants of the redesign task, subtle differences exist that have an impact on how the task can be performed and what kinds of knowledge are

There are a variety of research works referring to design or redesign; from (re)design of abstract (for example, components in software engineering) to physical entities (for example, a reactor in chemical engineering), for a general review see (Brown 97), for some details see (Akin 82, Mitchell 83, Howe 86, Fischer 87, Mostow 89, Goel 91, Bras 92, Stroulia 92a, Chandrasekaran 93, French 93, Brazier 96, Eldonk 96, Pos 97, Price 97, Umeda 97, Gero 98, Culley 99, Culley 99, Kitamura 99, Kraslawski 00, Grossmann 00, Arana 01, Maher 01).

Models in design and redesign are particularly important to guarantee they represent the intentions by which they were created. In general, the models are abstractions of the reality that guarantees communication of ideas by joining concepts, aggregations and relations

that the required knowledge for redesign is based on the following two principles:

Maximise existing properties and benefits of the current design.

points of view about the relationship between design and redesign, these are:

former design description.

of problem-solving methods for redesign.

Minimise changes in the current design, and

requirements.

involved.

(Bridge 97). Akin (Akin 82) outlines that the representational aspects to determine the utility of a model in design are:


A substantial amount of research has focused on defining models of design (French 85, Tomiyama 87, Treur 89, Brown 89, Chandrasekaran 90, Gero 90a, Takeda 90a, Alberts 92, Vescovi 93, Ohsuga 97, Brown 97). Most of this research highlights that the modelling of the functionality (or properties) of the design object description is an important aspect of the overall design process.

It is possible to represent explicit knowledge in (re)design by means of modelling functions of artefacts. This facilitates the systematisation of the reasoning and some tasks of (re)design. The reasoning based on functions allows abstracting information of the design on the same way as it is made in the reasoning of the initial stages of the design. The process of design of an artefact starts with the conceptual or functional design followed by the basic design and the detailed design (Stephanopoulos 90a). Within these, the functional design plays the central role since it guarantees the quality of the design and the innovation of the product (Umeda 97, Culley 99). The idea of function is fundamental in design since the work of the designers is to design artefacts that must achieve explicit functions (Chandrasekaran 00). Functional modelling is useful to model the object of (re)design, this modelling of objects enhances the formulation of (re)design strategies and the overall (re)design process. Functional modelling "hides" sections of the artefact structure at a lower abstraction level facilitating the manipulation of the artefact description.

Most of the research work on (re)design considers redesign as a knowledge-intensive field; wherein the processes (e.g., tasks) performed, descriptions of sequencing of processes, descriptions of the information within the system, and knowledge employed to perform a task are explicitly modelled most of the times by means of knowledge-based systems. These modelling frameworks try to model the (re)design so the (re)design object as well as the (re)design process are understandable by humans. To do this, human designers use the object specifications to propose a reasonable (re)design approach need to be understood (Leveson 00). Reasoning strategies employed in (re)design are derived or extensions of the commonly named *problem-solving strategies*. Examples of strategies are hypothesis and test (Hempel 66, White 05), pattern recognition (Doyle 62, Kirsch 64, Mitchell 97), skeletal plan refinement (Friedland 85, Tu 89), heuristic classification (Clancey 85), propose and revise (Goel 89), propose critique modify (Chandrasekaran 90), decision tree search (Raiffa 68, Qi 92), means-ends analysis (Newell 63, Rasmussen 86), and reasoning by analogy (Gick 80, Gentner 83).

In the above strategies, the human designer needs to formulate an explicit model of expertise as an integration of two types of models: a domain model and problem solving method. The domain model corresponds to the (re)design object and the problem solving method model corresponds to the (re)design process. Work on domain modelling has only recently attracted the attention of knowledge based system researchers (Stephanopoulos 90a, Schoen 91, Gruber 93, Skuce 93, Sowa 95, Kitamura 98, Fensel 01b, Gomez-Perez 04).

Modelling Approach for Redesign of Technical Processes 99

The design object is the central "actor" that receives the attention during the overall design process. This can be a model of a component, artefact, process or system. Traditionally, the design object was created by technical drafts; but with the advent of computers, the design object has become a computer model that can be shown, modified and deleted easily. Thus,

Some authors (Rasmussen 86, Douglas 88, Hoover 91, Lind 94, Turton 98, Leveson 00) have observed that abstractions of the design object are important during the design process to

abstractions and refinements are selected opportunistically and are characterised by the

Several research works have been developed about (re)design object manipulation. The most relevant approaches in this issue are model-based design and case-based design.

One of the most used approaches in the manipulation of the (re)design object is model-based design which really is a branch of model-based reasoning (MBR) applied to (re)design. Model-based reasoning constitutes a set of techniques applied in several domains and it is used to create models and reasoning about the domain. Mainly the most used technique from MBR has been *compositional modelling* (Falkenhainer 91, Falkenhainer 92, Nayak and Joskowicz 96), which is an approach to construct a model of an artefact (components, devices, processes, systems, etc.) on the basis of a description of the artefact and a query about the composition of the artefact. The modelling of functions (*functional representation*) (Sembugamoorthy 86, Chittaro 98, Chandrasekaran 00) is crucial in compositional modelling. Functional Representation is a top-down approach to describe functions on devices (function), its structure (structure) and its causal processes (behaviour) of the device that culminate with the achievement of the function. Functional modelling reduces drastically the amount of information if simulation is required (Price 98). The

conceptual, layout, and detailed stages are not distinct steps in the design process.

manipulate design objects. In this sense, Hoover (Hoover 91) has observed that:

the design object evolves through abstractions and refinements.

refinements are made within the framework of abstractions.

approaches of functional modelling can be classified in two groups:

designer focusing on a few aspects of the design object at a time.

agent architectures (Dunskus 95, Berker 96, Lander 97).

Prescriptive models (Salomons 95)

Opportunistic design (French 93)

Human learning process (Gero 04)

**2.4 The design object** 

**2.4.1 Model-based design** 

The following models of the design process can be distinguished:

Descriptive models (Stephanopoulos 90b, Ohsuga 97, Sumi 97)

Decision support problem (Bras 92, Thornton 93, Ullman 91)

Theorem problem solving process (Takeda 94a)

several models of artefacts have been used in design.

Multiagent design (Wood 01, DeLoach 04).

The problem solving method determines how those entities in the model will be used in the actual problem solving process. Domain specific concepts, relationships, and knowledge pertaining to them are captured in the domain model through ontologies (Chittaro 93, Kitamura 99, Fensel 01b, Kuraoka 03).

Independently of models and strategies employed in the (re)design, it is important that such data and knowledge can be recorded in a consistent manner for the future understanding of the (re)design; this constitutes what is called *(re)design rationale*.

#### **2.2 The role of function in the design process**

Functions in design play the central role since it guarantees the quality of the design and the innovation of the product (Umeda 97, Culley 99). Function is regarded as what a design object is supposed to do; it is a manageable representation of the overall behaviour of the object (Price 98). Some authors define function as an abstraction of its intended behaviour strongly related to its context (Gero 90a, Goel 92, Stroulia 92a, Chittaro 93, Brown 97, Chandrasekaran 00). Initially the human designers think in functions before they are concerned with specific properties. Functions can exist at different levels of abstraction, depending on the design phase that one is in and the current focus of the design interest. In preliminary design phases, functions usually are independent of working principle, whereas in later design phases, when the functions have been detailed, they become more and more dependent on the working principle that has been selected. In the following, a distinction between three levels or categories of functions is made:


#### **2.3 The design process**

The design process is a complex and not yet well understood cognitive process conducted by humans (Salomons 95). The design process is related to the process of actions and decisions that are taken during design in order to arrive at completed product design. Models of design processes provide a structured description of a process of design. The models differ in their underlying formalisations and have been represented in structures such as:


agent architectures (Dunskus 95, Berker 96, Lander 97).

The following models of the design process can be distinguished:


#### **2.4 The design object**

98 Advances in Chemical Engineering

The problem solving method determines how those entities in the model will be used in the actual problem solving process. Domain specific concepts, relationships, and knowledge pertaining to them are captured in the domain model through ontologies (Chittaro 93,

Independently of models and strategies employed in the (re)design, it is important that such data and knowledge can be recorded in a consistent manner for the future understanding of

Functions in design play the central role since it guarantees the quality of the design and the innovation of the product (Umeda 97, Culley 99). Function is regarded as what a design object is supposed to do; it is a manageable representation of the overall behaviour of the object (Price 98). Some authors define function as an abstraction of its intended behaviour strongly related to its context (Gero 90a, Goel 92, Stroulia 92a, Chittaro 93, Brown 97, Chandrasekaran 00). Initially the human designers think in functions before they are concerned with specific properties. Functions can exist at different levels of abstraction, depending on the design phase that one is in and the current focus of the design interest. In preliminary design phases, functions usually are independent of working principle, whereas in later design phases, when the functions have been detailed, they become more and more dependent on the working principle that has been selected. In the following, a distinction

 *General functions*. (Keuneke 91, Lind 94, Kitamura 98, Bo 99) proposed a restricted list of general functions dealing with the transformation of matter, energy and/or

*Specialised functions or subfunctions*. Act on flows, forces, moments etc., independent of

 *Working principle dependent function*. Salomons (Salomons 95) defines it as the realisation of a specialised function (by means of physical phenomena). Several alternative solutions for fulfilling working principle dependent functions can exist without

The design process is a complex and not yet well understood cognitive process conducted by humans (Salomons 95). The design process is related to the process of actions and decisions that are taken during design in order to arrive at completed product design. Models of design processes provide a structured description of a process of design. The models differ in their underlying formalisations and have been represented in structures

SOAR (cognitive architecture for developing systems with intelligent behaviour) (Steier

task models or problem solving methods (Brown 89, Brazier 94, Wielinga 97), or

Kitamura 99, Fensel 01b, Kuraoka 03).

the (re)design; this constitutes what is called *(re)design rationale*.

**2.2 The role of function in the design process** 

between three levels or categories of functions is made:

changing the working principle itself.

blackboard architectures (Ball 92),

algorithms (Alberts 93b),

the working principle.

**2.3 The design process** 

such as:

91),

information, which are independent of the working principle.

The design object is the central "actor" that receives the attention during the overall design process. This can be a model of a component, artefact, process or system. Traditionally, the design object was created by technical drafts; but with the advent of computers, the design object has become a computer model that can be shown, modified and deleted easily. Thus, several models of artefacts have been used in design.

Some authors (Rasmussen 86, Douglas 88, Hoover 91, Lind 94, Turton 98, Leveson 00) have observed that abstractions of the design object are important during the design process to manipulate design objects. In this sense, Hoover (Hoover 91) has observed that:


Several research works have been developed about (re)design object manipulation. The most relevant approaches in this issue are model-based design and case-based design.

#### **2.4.1 Model-based design**

One of the most used approaches in the manipulation of the (re)design object is model-based design which really is a branch of model-based reasoning (MBR) applied to (re)design. Model-based reasoning constitutes a set of techniques applied in several domains and it is used to create models and reasoning about the domain. Mainly the most used technique from MBR has been *compositional modelling* (Falkenhainer 91, Falkenhainer 92, Nayak and Joskowicz 96), which is an approach to construct a model of an artefact (components, devices, processes, systems, etc.) on the basis of a description of the artefact and a query about the composition of the artefact. The modelling of functions (*functional representation*) (Sembugamoorthy 86, Chittaro 98, Chandrasekaran 00) is crucial in compositional modelling. Functional Representation is a top-down approach to describe functions on devices (function), its structure (structure) and its causal processes (behaviour) of the device that culminate with the achievement of the function. Functional modelling reduces drastically the amount of information if simulation is required (Price 98). The approaches of functional modelling can be classified in two groups:

Modelling Approach for Redesign of Technical Processes 101

objective of this approach is to find the part of the system, which causes the discrepancy between a formal specification of the system to be redesigned and the description of the existing technical system. Kitamura and Mizoguchi [Sasajima 95, Kitamura 99] proposed a redesign approach based on ontologies of functional concepts. They focus on capturing the rationales of design of an artefact and in organising general strategies of redesign. For the first point, they use an ontology of functional concepts that allows to identify functional structures and to represent automatically part of the design rationale. For the second point, they use an ontology of redesign strategies. This approach consists of the following stages: functional understanding, analysis of requirements, proposal of alternative and evaluation.

Aranna et. al. [Fothergill 95, Forster 96, Forster 97b, Arana 00] proposed a redesign environment called DEKLARE, which supports acquisition, representation and reuse of redesign knowledge. It allows the designer to use design techniques to suggest alternative designs that fulfill specific requirements. Gupta et. al. [Das 94] proposed a methodology that automatically provides suggestions of redesign for reducing setup costs for mechanical parts. This approach is based on the interpretation of the design as a collection of mechanical features. The objective is to generate alternative mechanical features by means of geometric changes of the original parts and adding them to the feature set of the original part. Kim [Kim 93] proposed an approach for redesign of assemblies by means of planning techniques. Kim deals with the absence of required design information using the replay and modify principle. He employs a reverse engineering model to infer information about the process executed when creating a given design, and using the inferred information for design recreation or redesign. The propose model consists of the three stages: knowledge

acquisition, construction of the default design plan, and redesign based on cases.

Steinberg and Mitchell developed a system to redesign VLSI circuits [Steinberg 85]. This redesign approach is based on planning techniques and causal and teleological reasoning [de Kleer 79]. The subtasks of this approach are: a) focus on an appropriate section of the circuit, b) generate redesign options to the level of proposed specifications for individual modules, c) rank the generated redesign options, d) implement the selected redesign option, and e) detect and repair of side effects resulting from the redesign. Maulik et. al. [Maulik 92] proposed the use of optimisation techniques to redesign CMOS analog circuits. The optimisation approach is guided by three principles. First, equations that describe device characteristics are encapsulated and separated from equations that describe the performance of the circuit topologies. Secondly, constrained optimisation techniques are employed to synthesise the redesigned-scaled CMOS circuit. Finally, constrained optimisation allows the solution of some final constraints over specific variables. Based on the approach of Umeda et. al. [Umeda 92], Tomiyama et. al. [Umeda 94] describe an extension of their approach taking into account the potential functions of the components of an artefact to redesign it. The architecture consists of sensors, which monitor the machine, and a model-based reasoner diagnoses faults and plans repairs. The system generates a FBS model (Function-Behaviour-State) based on the design object, and then searches the model for candidate redundant function. The FBS model consists of a function hierarchy that represents the

**2.5.2 Mechanical engineering** 

**2.5.3 Electrical and electronic engineering** 


#### **2.4.2 Case-based design**

Case-based design is a branch of case-based reasoning (CBR). CBR is a general paradigm to solve problems based on the recovery, reuse, revision and retention of specific experiences (cases) (Aamodt 94). CBR has been applied to component-based systems (Maher 97a) which is, however, mostly concerned with the manipulation of design object descriptions. This paradigm is particularly attractive in domains where explicit models do not exist or its understanding is difficult (Kolodner 93). In CBR, similarities between formal methods implemented in computer programs and informal observations from designers are taken into account (Maher 97a). The applications of CBR can be for classification or synthesis tasks. (Re)design problems are within the synthesis tasks.

The direct or analogical use of previous designs or plans of design can reduce and improve the quality of design because take advantages of previous experiences (Maher 95). CBR is viewed as a redesign process for the adaptation of a case where a new artefact (named goal) is designed to achieve certain function, its physical structure can be inferred in analogical way from some physical, chemical or biological object (named source) whose function is similar to the required function.

CBR has been applied to solve problems of real world; there are several works about casebased design for example (Goel 89, Qian 92, Sycara 92, Bhatta 94, Borner 96, de Silva Garza 96, Maher 01, Price 97).

#### **2.5 Redesign approaches**

Here the division between research on design and redesign is remarked, only research from the redesign perspective is presented.

#### **2.5.1 Generic approaches in engineering**

Goel et. al. [Stroulia 92a, Stroulia 92b, Goel 94b, Goel 97a] presented a control architecture for model-based redesign in the context of case-based redesign. They state that the redesign task is characterised by small differences in the functions desired of and delivered by an existent known design. The redesign is divided in three subtasks: a) generation of modifications to the structure of the old design, b) realisation of the modifications on the structure, and c) evaluation of the new design. Eldonk et. al. [Alberts 93a, Bakker 94, Eldonk 96] presented a redesign approach based on techniques developed in model-based diagnosis. Eldonk et. al. state that redesign activities are diagnosis and respecification. The objective of this approach is to find the part of the system, which causes the discrepancy between a formal specification of the system to be redesigned and the description of the existing technical system. Kitamura and Mizoguchi [Sasajima 95, Kitamura 99] proposed a redesign approach based on ontologies of functional concepts. They focus on capturing the rationales of design of an artefact and in organising general strategies of redesign. For the first point, they use an ontology of functional concepts that allows to identify functional structures and to represent automatically part of the design rationale. For the second point, they use an ontology of redesign strategies. This approach consists of the following stages: functional understanding, analysis of requirements, proposal of alternative and evaluation.

#### **2.5.2 Mechanical engineering**

100 Advances in Chemical Engineering

 *State-based representations*. It uses units of function representation, which are abstractions of behaviour states. Behaviour states and hence functions may be

 *Flow-based representations*. Flow-based representations are based on the concepts of flow and effort. In this approach exists a predefined set of functions, and functions of all existent components are expressed in terms of these primitive functions. This approach is based on the System Theory (Bertalanffy 50) and its derivatives (Abstraction Hierarchy (Rasmussen 86), Qualitative Process Theory (Forbus 84), and Multilevel Flow

Case-based design is a branch of case-based reasoning (CBR). CBR is a general paradigm to solve problems based on the recovery, reuse, revision and retention of specific experiences (cases) (Aamodt 94). CBR has been applied to component-based systems (Maher 97a) which is, however, mostly concerned with the manipulation of design object descriptions. This paradigm is particularly attractive in domains where explicit models do not exist or its understanding is difficult (Kolodner 93). In CBR, similarities between formal methods implemented in computer programs and informal observations from designers are taken into account (Maher 97a). The applications of CBR can be for classification or synthesis

The direct or analogical use of previous designs or plans of design can reduce and improve the quality of design because take advantages of previous experiences (Maher 95). CBR is viewed as a redesign process for the adaptation of a case where a new artefact (named goal) is designed to achieve certain function, its physical structure can be inferred in analogical way from some physical, chemical or biological object (named source) whose function is

CBR has been applied to solve problems of real world; there are several works about casebased design for example (Goel 89, Qian 92, Sycara 92, Bhatta 94, Borner 96, de Silva Garza

Here the division between research on design and redesign is remarked, only research from

Goel et. al. [Stroulia 92a, Stroulia 92b, Goel 94b, Goel 97a] presented a control architecture for model-based redesign in the context of case-based redesign. They state that the redesign task is characterised by small differences in the functions desired of and delivered by an existent known design. The redesign is divided in three subtasks: a) generation of modifications to the structure of the old design, b) realisation of the modifications on the structure, and c) evaluation of the new design. Eldonk et. al. [Alberts 93a, Bakker 94, Eldonk 96] presented a redesign approach based on techniques developed in model-based diagnosis. Eldonk et. al. state that redesign activities are diagnosis and respecification. The

associated even with static objects which do not cause any state change.

Modelling (Lind 90, Lind 94)).

tasks. (Re)design problems are within the synthesis tasks.

**2.4.2 Case-based design** 

similar to the required function.

96, Maher 01, Price 97).

**2.5 Redesign approaches** 

the redesign perspective is presented.

**2.5.1 Generic approaches in engineering** 

Aranna et. al. [Fothergill 95, Forster 96, Forster 97b, Arana 00] proposed a redesign environment called DEKLARE, which supports acquisition, representation and reuse of redesign knowledge. It allows the designer to use design techniques to suggest alternative designs that fulfill specific requirements. Gupta et. al. [Das 94] proposed a methodology that automatically provides suggestions of redesign for reducing setup costs for mechanical parts. This approach is based on the interpretation of the design as a collection of mechanical features. The objective is to generate alternative mechanical features by means of geometric changes of the original parts and adding them to the feature set of the original part. Kim [Kim 93] proposed an approach for redesign of assemblies by means of planning techniques. Kim deals with the absence of required design information using the replay and modify principle. He employs a reverse engineering model to infer information about the process executed when creating a given design, and using the inferred information for design recreation or redesign. The propose model consists of the three stages: knowledge acquisition, construction of the default design plan, and redesign based on cases.

#### **2.5.3 Electrical and electronic engineering**

Steinberg and Mitchell developed a system to redesign VLSI circuits [Steinberg 85]. This redesign approach is based on planning techniques and causal and teleological reasoning [de Kleer 79]. The subtasks of this approach are: a) focus on an appropriate section of the circuit, b) generate redesign options to the level of proposed specifications for individual modules, c) rank the generated redesign options, d) implement the selected redesign option, and e) detect and repair of side effects resulting from the redesign. Maulik et. al. [Maulik 92] proposed the use of optimisation techniques to redesign CMOS analog circuits. The optimisation approach is guided by three principles. First, equations that describe device characteristics are encapsulated and separated from equations that describe the performance of the circuit topologies. Secondly, constrained optimisation techniques are employed to synthesise the redesigned-scaled CMOS circuit. Finally, constrained optimisation allows the solution of some final constraints over specific variables. Based on the approach of Umeda et. al. [Umeda 92], Tomiyama et. al. [Umeda 94] describe an extension of their approach taking into account the potential functions of the components of an artefact to redesign it. The architecture consists of sensors, which monitor the machine, and a model-based reasoner diagnoses faults and plans repairs. The system generates a FBS model (Function-Behaviour-State) based on the design object, and then searches the model for candidate redundant function. The FBS model consists of a function hierarchy that represents the

Modelling Approach for Redesign of Technical Processes 103

The overall redesign process depends on the problem-solving strategy used. In order to start the redesign process, the problem must be specified in terms of objectives that the original artefact must satisfy and the criteria that can be used to rank the alternative designs. Then a synthesis process takes place and the results are a set of alternative designs. Each of these alternatives is analysed and evaluated in terms of the predefined objectives and design criteria. Finally, one alternative is selected to be implemented. The process is highly iterative; the results from later stages are fedback to early stages to modify objectives,

Design alternatives are generated through a process of analysis of system composition. The designer breaks down the system (artefact) into a set of subsystems (components), together with the functions and constraints imposed upon the individual subsystem designs. These aspects are analysed with respect to desired system performance features and constraints. The process is iterative until an acceptable design alternative is achieved. At the end of this process all components must be described in such detail that an implementation of the

The understanding of the redesign object depends strongly on the mental models of the human designer. Usually designers communicate their ideas more easily in terms of abstract, high-level descriptions to describe complex concepts (Price 03). The description of the redesign object can be done in many different ways, depending on the context and purpose for which the description is to be used. In the early phases of redesign, highly abstract descriptions (e.g. qualitative or causal) might be helpful, whereas in later phases,

Considering the notion of function, some researchers (Sembugamoorthy 86, Goel 89, Franke 92, Keuneke 91, Chittaro 93, Iwasaki 93) organise the knowledge in a domain by means of functional concepts. The main claim of these approaches is that functions and intentions can provide important additional information for understanding and reasoning about the structure and behaviour of physical systems. In addition, other researchers have directed their extentions to hierarchical modelling by means of different aggregation levels (Liu 91, Rajamoney 91) or different approximations (Weld 86, Kuipers 87, Struss 91, Falkenhainer 91) to organise the knowledge. Independently of the tools and representations employed, several authors (Fischo. 78, Checkland 81, Jaffe 91, Vicente 92) suggest that two important aspects must be addressed if computer tools are used to tackle activities of complex systems: *Content*, the semantic information that should be contained in the representation given the goals and tasks of the users. The content gives the basic issues to understand the information about the redesign object. Independently of the amount and complexity of the information, the designer can conceive, in general terms, the objectives of the

 *Structure*, how to design the representation to facilitate that the user can extract the required information. The structure concerns to the organisation of the process

more detailed and quantitative descriptions provide more suitable information.

**2.6 Modelling** 

**2.6.1 Modelling the redesign process** 

criteria, design alternatives, and so on.

whole object can be performed.

redesign object.

components.

**2.6.2 Modelling the redesign object** 

designer's intentions, and a behaviour network that describes how the function hierarchy is carried out. The system first tries a control type strategy that adjusts various machine parameters. If the strategy fails the system applies a strategy based on functional redundancy, it uses the potential functions of existing parts in a slightly different way from the original design. Heo et. al. [Heo 98] presented a redesign approach of digital electronic systems by means of evolutive programming. They use directed acyclic graphs known as task flow graphs (TFG) to represent the redesign object. Each node of the graph represents computational tasks; an edge represents a transfer of data. The design process consists of five tiers: a) system-level design, b) architectural design, c) logic design, d) circuit design, and e) physical design.

#### **2.5.4 Chemical engineering**

The (re)design of chemical processes is made with the purpose of adapting existing processes to changes in economic, technological or environmental requirements. In the eighties, there were mainly significant advances on saving energy by means of two constraint-based approaches: a) pinch methodology (Tjoe 86, Smith 87, Linnho. 88) and b) mathematical programming on synthesis and design of processes [Papoulias 83, Pistikopoulos 87, Vaselenak 87]. In the nineties, Gundersen [Gundersen 90] made a revision of systematic methods of redesign of processes, which were broadly tackled. In such revision, he emphasised two important observations:


Doherty et. al. [Fischer 87] developed a systematic procedure of redesign by means of opportunistic searches; the procedure considers modifications in the structure of the flowsheet and in the dimension of equipments. Kirkwood et. al. [Kirkwood 88] implemented a methodology of redesign by means of an expert system by using heuristic rules to construct hierarchical structures. Nelson and Douglas [Nelson 90] developed a systematic procedure considering alternative reaction routes; the procedure is hierarchical and provides guides to identify viable processes. Rapoport et. al. [Rapoport 94] proposed an algorithm to design units of process by means of the redesign of already existing ones. The algorithm consists on hierarchical levels and heuristic rules; this approach is similar to synthesis of processes. Stephanopoulos et. al. [Han 95] developed an approach based on agents to synthesis of processes; they modeled the process of design like a set of tasks that can be executed by agents. Systems have also been developed to satisfy economic, environmental and safety constraints. Kraslawski et. al. [Kraslawski 00] developed a methodology centred on the identification and elimination of bottlenecks in reaction and separation sections. Sylvester et. al. [Sylvester 00] optimised processes within the concept of Greener Process. Hertwig et. al. [Hertwig 01] applied techniques of MINLP (Mixed-Integer Non-Linear Programming) to optimise configuration of processes. Pasanen [Pasanen 01] developed a tool for conceptual design of processes, this is called Phenomenon Driven Process Design (PDPD). This methodology focuses on the systematisation of design conceptual of chemical processes. Uerdingen et. al. [Uerdingen 01] presented a screening method based on an analysis of the flow path pattern. They use performance indicators to rate the economic impact of each component of the flowsheet in the flow path.

#### **2.6 Modelling**

102 Advances in Chemical Engineering

designer's intentions, and a behaviour network that describes how the function hierarchy is carried out. The system first tries a control type strategy that adjusts various machine parameters. If the strategy fails the system applies a strategy based on functional redundancy, it uses the potential functions of existing parts in a slightly different way from the original design. Heo et. al. [Heo 98] presented a redesign approach of digital electronic systems by means of evolutive programming. They use directed acyclic graphs known as task flow graphs (TFG) to represent the redesign object. Each node of the graph represents computational tasks; an edge represents a transfer of data. The design process consists of five tiers: a) system-level design, b) architectural design, c) logic design, d) circuit design,

The (re)design of chemical processes is made with the purpose of adapting existing processes to changes in economic, technological or environmental requirements. In the eighties, there were mainly significant advances on saving energy by means of two constraint-based approaches: a) pinch methodology (Tjoe 86, Smith 87, Linnho. 88) and b) mathematical programming on synthesis and design of processes [Papoulias 83, Pistikopoulos 87, Vaselenak 87]. In the nineties, Gundersen [Gundersen 90] made a revision of systematic methods of redesign of processes, which were broadly tackled. In such

The systematic methods of redesign of processes are based on methods of design of

Doherty et. al. [Fischer 87] developed a systematic procedure of redesign by means of opportunistic searches; the procedure considers modifications in the structure of the flowsheet and in the dimension of equipments. Kirkwood et. al. [Kirkwood 88] implemented a methodology of redesign by means of an expert system by using heuristic rules to construct hierarchical structures. Nelson and Douglas [Nelson 90] developed a systematic procedure considering alternative reaction routes; the procedure is hierarchical and provides guides to identify viable processes. Rapoport et. al. [Rapoport 94] proposed an algorithm to design units of process by means of the redesign of already existing ones. The algorithm consists on hierarchical levels and heuristic rules; this approach is similar to synthesis of processes. Stephanopoulos et. al. [Han 95] developed an approach based on agents to synthesis of processes; they modeled the process of design like a set of tasks that can be executed by agents. Systems have also been developed to satisfy economic, environmental and safety constraints. Kraslawski et. al. [Kraslawski 00] developed a methodology centred on the identification and elimination of bottlenecks in reaction and separation sections. Sylvester et. al. [Sylvester 00] optimised processes within the concept of Greener Process. Hertwig et. al. [Hertwig 01] applied techniques of MINLP (Mixed-Integer Non-Linear Programming) to optimise configuration of processes. Pasanen [Pasanen 01] developed a tool for conceptual design of processes, this is called Phenomenon Driven Process Design (PDPD). This methodology focuses on the systematisation of design conceptual of chemical processes. Uerdingen et. al. [Uerdingen 01] presented a screening method based on an analysis of the flow path pattern. They use performance indicators to

and e) physical design.

processes.

**2.5.4 Chemical engineering** 

revision, he emphasised two important observations:

Most of the projects in the industry of processes were redesign projects.

rate the economic impact of each component of the flowsheet in the flow path.

#### **2.6.1 Modelling the redesign process**

The overall redesign process depends on the problem-solving strategy used. In order to start the redesign process, the problem must be specified in terms of objectives that the original artefact must satisfy and the criteria that can be used to rank the alternative designs. Then a synthesis process takes place and the results are a set of alternative designs. Each of these alternatives is analysed and evaluated in terms of the predefined objectives and design criteria. Finally, one alternative is selected to be implemented. The process is highly iterative; the results from later stages are fedback to early stages to modify objectives, criteria, design alternatives, and so on.

Design alternatives are generated through a process of analysis of system composition. The designer breaks down the system (artefact) into a set of subsystems (components), together with the functions and constraints imposed upon the individual subsystem designs. These aspects are analysed with respect to desired system performance features and constraints. The process is iterative until an acceptable design alternative is achieved. At the end of this process all components must be described in such detail that an implementation of the whole object can be performed.

#### **2.6.2 Modelling the redesign object**

The understanding of the redesign object depends strongly on the mental models of the human designer. Usually designers communicate their ideas more easily in terms of abstract, high-level descriptions to describe complex concepts (Price 03). The description of the redesign object can be done in many different ways, depending on the context and purpose for which the description is to be used. In the early phases of redesign, highly abstract descriptions (e.g. qualitative or causal) might be helpful, whereas in later phases, more detailed and quantitative descriptions provide more suitable information.

Considering the notion of function, some researchers (Sembugamoorthy 86, Goel 89, Franke 92, Keuneke 91, Chittaro 93, Iwasaki 93) organise the knowledge in a domain by means of functional concepts. The main claim of these approaches is that functions and intentions can provide important additional information for understanding and reasoning about the structure and behaviour of physical systems. In addition, other researchers have directed their extentions to hierarchical modelling by means of different aggregation levels (Liu 91, Rajamoney 91) or different approximations (Weld 86, Kuipers 87, Struss 91, Falkenhainer 91) to organise the knowledge. Independently of the tools and representations employed, several authors (Fischo. 78, Checkland 81, Jaffe 91, Vicente 92) suggest that two important aspects must be addressed if computer tools are used to tackle activities of complex systems:


Modelling Approach for Redesign of Technical Processes 105

*goals*, which are the objectives or purposes of the system, i.e., the ends that the designers

*functions*, which are the means by which the goals are obtained, i.e., the powers or

*physical components*, which are the different elements of the system, the equipment of

The concept of goal is central to MFM, as it is the "descriptor object" for teleological information. It is important to be able to recognise and describe goals, as they play an important role in every activity using means-end information. Without knowing the goals, it is virtually impossible to know the available functions. Three different types of goals can be

*production goals*, which are used to express what enables production. For example, a

 *safety goals*, *which are used to express reasons of safe operation. For instance, a particular process variable should be kept above or below some value, or inside or outside an interval. economy goals, which are used to express considerations of overall process optimisation.* 

*The function is the second* important concept on MFM. A function is always associated with a goal, and correspondingly, goals are always associated with functions. MFM describes the functional structure of a system as a set of interrelated flow structures on different abstraction levels. The levels are connected via achievement and condition relations; the flow structures consist of connected flow functions. Thus, the following types of flow

These flows are of completely different types, but they have many properties in common. Most flow functions can appear in each type of flow structure, thus, there are three flow types of flow functions. In MFM plant functions are represented by a set of mass, energy, activity and information flow structures on several levels of abstraction. The levels are interdependent and form means-end structures. Mass and energy flow structures are used to model the functions of the plant and activity and information flow structures are used to model the functions of the operator and the control systems. The mass and energy flow

*source*, the capability of a physical system to act as an infinite reservoir of mass, energy,

*transport*, the capability of a system to transfer mass, energy, or information from one

*barrier*, the capability of a system to prevent the transfer of mass, energy or information

*balance*, the capability of a system to provide a balance between the total rates of

part of the system to another (from one medium to another).

from one part of the system to another (from one medium to another). *storage*, the capability of a system to accumulate mass, energy, or information.

specific process variable should be kept within a given interval.

and operators want that the system reaches.

capabilities of the system.

which it consists.

recognised:

structures can be: mass flows energy flows information flows

functions are:

or information.

incoming and outgoing flows.

Rasmussen (Rasmussen 85) observed that the complexity of a system depends on the level of resolution in which the system is considered. The complexity can only be measured comparing with other systems observed at the same level of abstraction. The complexity can be manageable with more or less detail in the representations: then, hierarchical modelling can be seen as a way to handle complex systems. Models of complex artefacts can be expressed in terms of a hierarchy of levels of organisation, each one more complex than the previous. This modelling approach is named *Hierarchy Theory* (Rasmussen 86). Rasmussen studied the protocols developed by people working on complex systems and found that human users structure the system along two dimensions: a part-whole abstraction and a means-ends abstraction.

Some authors (Umeda 90, Franke 92, Lind 94) propose representation approaches for physical systems which maintain a clear separation between knowledge of structure and behaviour on one side and knowledge of function or purposes on the other side. This feature makes them useful in redesign of technical complex systems. The hierarchical functional modelling approaches employed in the present work are Multilevel Flow Modelling (MFM) (Lind 90, Lind 94, Lind 96, Lind 99) and Multimodelling (Ziegler 79, Praehofer 91, Chittaro 93), which are following described. Both approaches provide a more intuitive vision of reasoning on each task to be performed, and thus the redesign activities are enhanced. These approaches have been applied successfully in diagnosis and control domains.

#### **2.6.2.1 Multilevel flow modelling**

MFM provides a graphical and systematic basis for using means-end and whole-part hierarchical decompositions in the modelling of complex systems such as industrial plants. By the distinction between means and ends, a system is described in terms of goals, functions and the physical components that involves. At the same time, each of these descriptions can be given on different levels of whole-part decompositions. The main types of decomposition are illustrated in Figure 2. These are functional models with a very high level of abstraction, combined with a teleological representation of goals, or purposes, of the modelled system. Lind has suggested a syntax for a formal language and given the general ideas on how to use the MFM representation.

Fig. 2. Means-ends and part-whole dimensions in MFM.

An MFM model is a prescriptive description of a system, a representation of what it has been designed to do, how it should do it, and with which information it should do it. Thus, the three basic concept types of MFM are:


The concept of goal is central to MFM, as it is the "descriptor object" for teleological information. It is important to be able to recognise and describe goals, as they play an important role in every activity using means-end information. Without knowing the goals, it is virtually impossible to know the available functions. Three different types of goals can be recognised:


*The function is the second* important concept on MFM. A function is always associated with a goal, and correspondingly, goals are always associated with functions. MFM describes the functional structure of a system as a set of interrelated flow structures on different abstraction levels. The levels are connected via achievement and condition relations; the flow structures consist of connected flow functions. Thus, the following types of flow structures can be:

mass flows

104 Advances in Chemical Engineering

Rasmussen (Rasmussen 85) observed that the complexity of a system depends on the level of resolution in which the system is considered. The complexity can only be measured comparing with other systems observed at the same level of abstraction. The complexity can be manageable with more or less detail in the representations: then, hierarchical modelling can be seen as a way to handle complex systems. Models of complex artefacts can be expressed in terms of a hierarchy of levels of organisation, each one more complex than the previous. This modelling approach is named *Hierarchy Theory* (Rasmussen 86). Rasmussen studied the protocols developed by people working on complex systems and found that human users structure the system along two dimensions: a part-whole abstraction and a

Some authors (Umeda 90, Franke 92, Lind 94) propose representation approaches for physical systems which maintain a clear separation between knowledge of structure and behaviour on one side and knowledge of function or purposes on the other side. This feature makes them useful in redesign of technical complex systems. The hierarchical functional modelling approaches employed in the present work are Multilevel Flow Modelling (MFM) (Lind 90, Lind 94, Lind 96, Lind 99) and Multimodelling (Ziegler 79, Praehofer 91, Chittaro 93), which are following described. Both approaches provide a more intuitive vision of reasoning on each task to be performed, and thus the redesign activities are enhanced. These approaches have been applied successfully in diagnosis and control

MFM provides a graphical and systematic basis for using means-end and whole-part hierarchical decompositions in the modelling of complex systems such as industrial plants. By the distinction between means and ends, a system is described in terms of goals, functions and the physical components that involves. At the same time, each of these descriptions can be given on different levels of whole-part decompositions. The main types of decomposition are illustrated in Figure 2. These are functional models with a very high level of abstraction, combined with a teleological representation of goals, or purposes, of the modelled system. Lind has suggested a syntax for a formal language and given the general

An MFM model is a prescriptive description of a system, a representation of what it has been designed to do, how it should do it, and with which information it should do it. Thus,

means-ends abstraction.

**2.6.2.1 Multilevel flow modelling** 

ideas on how to use the MFM representation.

Fig. 2. Means-ends and part-whole dimensions in MFM.

the three basic concept types of MFM are:

domains.


These flows are of completely different types, but they have many properties in common. Most flow functions can appear in each type of flow structure, thus, there are three flow types of flow functions. In MFM plant functions are represented by a set of mass, energy, activity and information flow structures on several levels of abstraction. The levels are interdependent and form means-end structures. Mass and energy flow structures are used to model the functions of the plant and activity and information flow structures are used to model the functions of the operator and the control systems. The mass and energy flow functions are:


Modelling Approach for Redesign of Technical Processes 107

The Multilevel Flow Modelling (MFM) and the Multimodelling approaches are able to represent how the human designers behave during the redesign process. MFM is used for high-levels of abstractions and Multimodelling is more suitable for the intermediate and lower levels. Thus, the structure and behaviour of the components (equipments) are abstracted using the Multimodelling approach and then this abstraction is mapped to the MFM approach. The bridge between both approaches is the functions for the equipments in the domain. This functional modelling is the basis to manipulate the process during all the redesign process. The main idea is to model hierarchically the process and reason by using functional abstract concepts. In this way the designer can "navigate" in top-down and bottom-up directions in the representation in similar way as when the designer creates its mental models about the process. From an abstract point of view, there are three actors that

1. *The simulator*. The commercial software used to obtain the design description of the process and to implement and evaluate the generated alternative process designs. 2. *The reasoner*. The software modules required to model the process, identify the suitable equipment/section to be modified and obtain similar equipments/sections based on the

1. Identification of objectives and criteria. This stage covers the design-description

2. Generation of design alternatives. This stage is similar to the obtaining alternatives and

This redesign framework can deal with complex technical processes (the redesign object). The modelling approach was chosen to mimic the behaviour of human designers in real redesign situation of such processes. The final intention is to support human designers, not

The first stage of the framework is to obtain the design description of the process to be redesigned. This description is enough to carry out the redesign activities, and just few adaptations are necessaries to fulfill the redesign objective. This stage is carried out in two

3. Evaluation of alternatives. This stage is carried out manually by the human designer. 4. Implementation of alternatives. This stage is also carried out manually by the human

acquisition and the identification of candidates stages of the framework.

human experts acquire through direct interaction with the system. 4. Aggregation levels. The degree of granularity of the represented knowledge.

**3. The redesign framework** 

play an independent role in this framework:

adaptation stages in the framework.

**3.1 Design-description acquisition** 

designer.

3. *The human designer*. The human user interpreting the results. The stages of the proposed redesign framework (Figure 3) are:

to carry out the redesign automatically without human intervention.

substages: data acquisition and functional identification.

selected equipment/section.

 *Empirical*. The knowledge concerning the explicit representation of the system properties through empirical associations (such as observation, experimentation, and experience). This knowledge may include subjective competence that usually

 *sink*, represents the capability of a system to act as an infinite drain of mass, energy or information.

These functions can be used to describe information flows. There are also some specific information flow functions:


In addition to the flow functions, some organisational functions are used. They are concerned with expressing support and control:


#### **2.6.2.2 Multimodelling**

The Multimodelling approach represents many diverse and explicit models of a system, which are used in a cooperative way in specific problem solving tasks. The fundamental assumptions about knowledge modelling and reasoning mechanisms do not identify a unique way of representing a physical system and reasoning about it. On the contrary, the Multimodelling approach is an abstract and general framework that allows for a variety of specific implementations. The fundamental concepts in Multimodelling are:

	- Object-centred ontology. The real world is made up of individual objects whose properties can be stated in an objective, context independent and general way.
	- System-centred ontology. The real world is made up of systems, intended as organised units, whose elements cannot be defined in isolation.
	- The scope of the model, i.e., the aspects of the real system which are considered relevant to the purpose of the model.
	- The precision of the model, i.e., the degree of accuracy of the representation
	- *Structural*. The knowledge about system topology, i.e., the components that constitute the system and how they are linked.
	- *Behavioural*. The knowledge that describes how components work and interact in terms of the physical quantities (variables and parameters).
	- *Functional*. The knowledge about the role components play in the physical processes in which they take part. This knowledge relates the behaviour of the system to its goals, and deals with functional roles, processes, and phenomena.
	- *Teleological*. The knowledge about the goals assigned to the system by its designer and about the operational conditions that allow their achievement through correct operation.

#### **3. The redesign framework**

106 Advances in Chemical Engineering

*sink*, represents the capability of a system to act as an infinite drain of mass, energy or

These functions can be used to describe information flows. There are also some specific

*actor*, represents the capability of a system to turn information into physical

In addition to the flow functions, some organisational functions are used. They are

*manager*, which describes control and supervisory systems, including human operators.

The Multimodelling approach represents many diverse and explicit models of a system, which are used in a cooperative way in specific problem solving tasks. The fundamental assumptions about knowledge modelling and reasoning mechanisms do not identify a unique way of representing a physical system and reasoning about it. On the contrary, the Multimodelling approach is an abstract and general framework that allows for a variety of

1. Ontologies. An ontology contains the descriptions of entities in the real system. Two

2. Representational assumptions. This issue concerns about what to represent of the real

 The precision of the model, i.e., the degree of accuracy of the representation 3. Epistemological types. The type of knowledge represented in the model. These types

The scope of the model, i.e., the aspects of the real system which are considered

*Structural*. The knowledge about system topology, i.e., the components that

*Behavioural*. The knowledge that describes how components work and interact in

 *Functional*. The knowledge about the role components play in the physical processes in which they take part. This knowledge relates the behaviour of the system to its

 *Teleological*. The knowledge about the goals assigned to the system by its designer and about the operational conditions that allow their achievement through correct

 Object-centred ontology. The real world is made up of individual objects whose properties can be stated in an objective, context independent and general way. System-centred ontology. The real world is made up of systems, intended as

*observer*, the capability of a system to translate physical observations to information.

*decision maker*, represents the decision-making capabilities of a system.

*network*, which is used to group a flow structure and connect it to a goal.

specific implementations. The fundamental concepts in Multimodelling are:

organised units, whose elements cannot be defined in isolation.

information.

information flow functions:

consequences.

**2.6.2.2 Multimodelling** 

can be:

operation.

concerned with expressing support and control:

types of ontologies can be distinguished:

system in the model. This involves two basic aspects:

constitute the system and how they are linked.

terms of the physical quantities (variables and parameters).

goals, and deals with functional roles, processes, and phenomena.

relevant to the purpose of the model.

The Multilevel Flow Modelling (MFM) and the Multimodelling approaches are able to represent how the human designers behave during the redesign process. MFM is used for high-levels of abstractions and Multimodelling is more suitable for the intermediate and lower levels. Thus, the structure and behaviour of the components (equipments) are abstracted using the Multimodelling approach and then this abstraction is mapped to the MFM approach. The bridge between both approaches is the functions for the equipments in the domain. This functional modelling is the basis to manipulate the process during all the redesign process. The main idea is to model hierarchically the process and reason by using functional abstract concepts. In this way the designer can "navigate" in top-down and bottom-up directions in the representation in similar way as when the designer creates its mental models about the process. From an abstract point of view, there are three actors that play an independent role in this framework:


The stages of the proposed redesign framework (Figure 3) are:


This redesign framework can deal with complex technical processes (the redesign object). The modelling approach was chosen to mimic the behaviour of human designers in real redesign situation of such processes. The final intention is to support human designers, not to carry out the redesign automatically without human intervention.

#### **3.1 Design-description acquisition**

The first stage of the framework is to obtain the design description of the process to be redesigned. This description is enough to carry out the redesign activities, and just few adaptations are necessaries to fulfill the redesign objective. This stage is carried out in two substages: data acquisition and functional identification.

Modelling Approach for Redesign of Technical Processes 109

The function of each unit is inferred by analysing their inputs (preconditions) and their outputs (postconditions), the variables involved, and the neighbour units. This process involves the analysis of the behaviour of the unit and its consequences in the surrounding units (the units connected to it). The next classification of functions was obtained based on

*Broad function*. Denotes a process-independent function that can be achieved

 *General function*. Denotes a function that can be achieved by several equipments in a domain. These functions deal with the transformation of mass and energy and are

 *Specific function*. Denotes the abstract function as it is known in the domain of the process. These functions relate flow variables with a specific physical process. According to the domain, these are functions to denote functional sections into the

 *Working function*. Denotes a function that can be achieved by a specific single unit. These functions relate specific flow variables with a specific physical phenomenon.

A unit can have several functions but only one goal (the objective, intention or purpose of the artefact). Several units can have a common goal. Therefore, based on the data extracted

*Functional*. Knowledge about the roles of each unit. The functional knowledge connects

 *Teleological.* Knowledge related to the goals of each unit by considering the required input operational conditions and the output operational conditions that were meant to

The functional knowledge is independent of the process (the same functional knowledge can be found in others processes), while the teleological knowledge are the goals assigned to

As result of this stage, each unit is represented by structural, behavioural, functional, and teleological models. The aim of this stage is to model the process in a higher level of

Based on the functions inferred in the functional unit identification stage, it is possible to identify the functional sections of the process (named *meta-units*). The incremental identification of these functional sections denotes the most important sections of the process. This incremental identification is carried out by generating different representations of the process at different levels of abstraction. The function of a unit is a working function because the unit (representing real equipment) was designed to perform only such function. The functional sections of the process denote specific and general functions by means of meta-units. Meta-units representing general functions are composed by meta-units with specific functions, not necessarily of the same type. A meta-unit represents a functional section at an abstract level. Thus a meta-unit at a higher abstract level can contain several

the behaviour (physical phenomena and processes) of the unit to its goal.

considering only flows of mass, energy, or information.

from the simulator it is possible to infer the following knowledge:

independent of the physical phenomena.

**Functional unit identification** 

MFM and Multimodelling:

process.

be produced.

the units by the designer).

abstraction (respect to the simulator). **Functional meta-unit identification** 

Fig. 3. The proposed redesign framework.

#### **3.1.1 Data acquisition**

It deals with data extracted from the specialised simulator used to implement the process to be redesigned. It was conceived as an appropriate step to reduce human intervention on the introduction of data to the reasoner module. The aim of the data acquisition is to obtain only the most useful data to generate the appropriate knowledge useful for the redesign of the process; thus irrelevant or superfluous data is ignored. Based on this data, the following types of knowledge are generated:


#### **3.1.2 Functional identification**

The data obtained from the simulator is used to model hierarchically the process. To do this, the functions of each equipment in the process must be identified. Based on the identified functions, the functional sections of the process can incrementally be identified. In the rest of the chapter any equipment will be named *unit* and a functional section will be named *metaunit*. This stage is divided into functional unit identification and functional meta-units identification. Here it is necessary to specify an ontology about the functional issues of the existing equipments in the process. By using this ontology, it is also required to specify a priority order of functions and the process variables related to each one. The grouping of functions depends strongly on such priority order.

#### **Functional unit identification**

108 Advances in Chemical Engineering

It deals with data extracted from the specialised simulator used to implement the process to be redesigned. It was conceived as an appropriate step to reduce human intervention on the introduction of data to the reasoner module. The aim of the data acquisition is to obtain only the most useful data to generate the appropriate knowledge useful for the redesign of the process; thus irrelevant or superfluous data is ignored. Based on this data, the following

*Structural*. Knowledge related to the topology of the process, i.e., the equipments

*Behavioural*. Knowledge related to the values of variables and parameters that

The data obtained from the simulator is used to model hierarchically the process. To do this, the functions of each equipment in the process must be identified. Based on the identified functions, the functional sections of the process can incrementally be identified. In the rest of the chapter any equipment will be named *unit* and a functional section will be named *metaunit*. This stage is divided into functional unit identification and functional meta-units identification. Here it is necessary to specify an ontology about the functional issues of the existing equipments in the process. By using this ontology, it is also required to specify a priority order of functions and the process variables related to each one. The grouping of

conforming the process and the connection between them.

characterise the behaviour of each equipment.

functions depends strongly on such priority order.

Fig. 3. The proposed redesign framework.

types of knowledge are generated:

**3.1.2 Functional identification** 

**3.1.1 Data acquisition** 

The function of each unit is inferred by analysing their inputs (preconditions) and their outputs (postconditions), the variables involved, and the neighbour units. This process involves the analysis of the behaviour of the unit and its consequences in the surrounding units (the units connected to it). The next classification of functions was obtained based on MFM and Multimodelling:


A unit can have several functions but only one goal (the objective, intention or purpose of the artefact). Several units can have a common goal. Therefore, based on the data extracted from the simulator it is possible to infer the following knowledge:


The functional knowledge is independent of the process (the same functional knowledge can be found in others processes), while the teleological knowledge are the goals assigned to the units by the designer).

As result of this stage, each unit is represented by structural, behavioural, functional, and teleological models. The aim of this stage is to model the process in a higher level of abstraction (respect to the simulator).

#### **Functional meta-unit identification**

Based on the functions inferred in the functional unit identification stage, it is possible to identify the functional sections of the process (named *meta-units*). The incremental identification of these functional sections denotes the most important sections of the process. This incremental identification is carried out by generating different representations of the process at different levels of abstraction. The function of a unit is a working function because the unit (representing real equipment) was designed to perform only such function. The functional sections of the process denote specific and general functions by means of meta-units. Meta-units representing general functions are composed by meta-units with specific functions, not necessarily of the same type. A meta-unit represents a functional section at an abstract level. Thus a meta-unit at a higher abstract level can contain several

Modelling Approach for Redesign of Technical Processes 111

Then, the overall process is represented by several functional sections denoting flow structures. Incrementally the identification of such functional sections denotes the most important sections of the process. This process finishes until the most important functional sections are identified. This corresponds to the "blackbox" from which the original design

The aim of this stage is to get the suitable unit or meta-unit to be modified to fulfill the redesign objectives. In a first instance, the redesign must be focused on a process variable. Once the variable is identified, a diagnostic algorithm is used to identify the units/metaunits affecting such process variable. This reasoning process is based on the functions identified at the functional analysis stage. This stage is composed of two substages: specification of redesign requirements and identification of the suitable unit/meta-unit for

Fig. 6. Interlevel meta-models.

Fig. 7. Abstraction of a process.

**3.2 Candidate identification** 

modification or substitution.

could begin (Figure 7).

units and other meta-units. Two or more meta-units can generate a more abstract meta-unit. Units/meta-units with lower priority functions are "absorbed" by units/meta-units with higher priority functions, as shows Figure 4.

Fig. 4. Grouping of units/meta-units.

Every functional section forms a hierarchy of meta-units and units where meta-models are connected in a same level (intralevel), as shows Figure 5, and at different levels (interlevel), as shown in Figure 6.

Fig. 5. Intralevel meta-models.

Fig. 6. Interlevel meta-models.

110 Advances in Chemical Engineering

units and other meta-units. Two or more meta-units can generate a more abstract meta-unit. Units/meta-units with lower priority functions are "absorbed" by units/meta-units with

Every functional section forms a hierarchy of meta-units and units where meta-models are connected in a same level (intralevel), as shows Figure 5, and at different levels (interlevel),

higher priority functions, as shows Figure 4.

Fig. 4. Grouping of units/meta-units.

as shown in Figure 6.

Fig. 5. Intralevel meta-models.

Then, the overall process is represented by several functional sections denoting flow structures. Incrementally the identification of such functional sections denotes the most important sections of the process. This process finishes until the most important functional sections are identified. This corresponds to the "blackbox" from which the original design could begin (Figure 7).

Fig. 7. Abstraction of a process.

#### **3.2 Candidate identification**

The aim of this stage is to get the suitable unit or meta-unit to be modified to fulfill the redesign objectives. In a first instance, the redesign must be focused on a process variable. Once the variable is identified, a diagnostic algorithm is used to identify the units/metaunits affecting such process variable. This reasoning process is based on the functions identified at the functional analysis stage. This stage is composed of two substages: specification of redesign requirements and identification of the suitable unit/meta-unit for modification or substitution.

Modelling Approach for Redesign of Technical Processes 113

Starting with the selected unit/meta-unit at the candidate identification stage, similar units/meta-units can be retrieved from other processes. The retrieved unit/meta-unit, which functional and teleological models are the most approximate to the functional and teleological models of the unit/meta-unit of interest, is adapted. This process requires that the performance and operational conditions of the cause and consequence units/meta-units

According to the later stage, the overall target process is modelled as a graph denoting a hierarchy of functions. Therefore hierarchical case-based reasoning (Smyth 01) is required.

The organisation of cases into the library is performed according to the type of functions of the unit/meta-unit. In this way, several groups can be distinguished according to the general function type: *source, transport, barrier, storage, balance, and sink*. Within each functional group, units and meta-units are grouped based on their specific functions. Again, within specific functions groups, the units/meta-units are grouped based on the working function achieved. There are not distinctions between units and meta-units with the same specific function. The case library is organised by an abstract hierarchy based on function groups. This structure denotes the organisation of the functional ontology used in the

To retrieve cases from the case library, a similarity engine is used. Only units/meta-units of the same specific functional group are considered. The similarity engine uses functional and teleological targets to search into the library of cases. Functional and teleological models denote strongly the relationship between the units/meta-units and its neighbours. Two

 *Local similarity*. Similarity between two cases is based on the local similarity between each feature of such cases. The computation depends on the type of the feature and the

 *Global similarity*. Once a set of local similarities has been computed for each known feature-value pair, the CBR system computes the global similarity of the candidate cases

As final result, a set of cases is obtained which contains meta-units (with its corresponding units/meta-units) or units. The set is ranked according to the global similarity between the target case (the unit/meta-unit of interest) and the source cases (the retrieved units/meta-

For this stage the human designer intervention is required since the use of the specialised simulator employed in the data acquisition is mandatory to test the proposed candidates. The adaptation is highly domain-dependent and requires online simulation of the process to verify its correct performance. Since information of abstract cases can not be used directly,

types of similarity are computed, local and global, which are defined as follows:

 *ground cases*. Cases located at the lowest level of abstraction, units (real equipments) *abstract cases*. Cases represented at higher levels of abstraction, meta-units (non-existent

associated with the retrieved unit must be similar to the original case.

"meta-equipments")

value that it could take.

**3.4 Adaptation and evaluation** 

based on such set.

framework.

units).

Thus, based on the levels of abstraction, two kinds of cases are distinguished:

#### **3.2.1 Specification of redesign requirements**

Here the human designer must specify the new requirements that the process must satisfy. Two categories of redesign requirements can be identified: functional requirements and physical requirements. A design specification always contains a single functional requirement; it may also contain a set of physical requirements. A functional requirement represents an abstraction of the intended behaviour of the artefact. It can be a general, specific, or working function. There is no direct association between the function that has to be provided and the physical mechanism that provides it. A physical requirement represents an abstraction of the physical process variables, which satisfy the functional requirement specified in the design specification. It denotes preferences about the designer intentions regarding some aspect of the process. The redesign specification can be represented by the functional and physical requirements or only by physical requirements. A redesign specification is a mean (goal) which is defined in terms of functions that must be embodied in a process in order to provide some higher level functionality.

#### **3.2.2 Identification of candidates**

Here In this stage, a diagnostic algorithm is used, that works on the functional concepts identified in the functional analysis. Diagnosis helps to detect "faulty" components (those that do not satisfy the global performance of the process). The design description and the new specifications are used to identify the possible candidates for modification or substitution. The diagnostic algorithm returns an ordered list of units or meta-units. Because the diagnostic algorithm operates over abstract functional concepts, no simulation is required. The diagnostic algorithm does not return the exact unit or meta-unit responsible for the "faulty" behaviour, it returns a list of units/meta-units that do not fulfill the global performance of the process represented by the redesign specification. The human designer is responsible to choose, from the resulting list, the appropriate unit/meta-unit that has to be modified or substituted into the process. Since this unit/meta-unit is connected to others by a flow path, the "cause" and "consequence" units/meta-units also must be identified. A cause unit/meta-unit is(are) the unit(s)/meta-unit(s) situated before the current unit/metaunit in the flow path. They are responsible to provide the appropriate operational conditions to the involved process variables in the function of the unit/meta-unit of interest. A consequence unit/meta-unit is(are) the unit(s)/meta-unit(s) situated after of the current unit/meta-unit in the flow path. They are the unit(s)/meta-unit(s) affected by the operational conditions given by the unit/meta-unit of interest. Both, the cause and the consequence units/meta-units, are not necessarily the closer neighbours.

#### **3.3 Generation of alternatives**

The aim of this stage is to obtain similar units (equipments) or meta-units (sections) to adapt them into the current process based on the suitable unit/meta-unit identified by the human designer at the last stage. The best way to obtain similar units/meta-units is from similar processes. With the adaptation of any retrieved unit/meta-unit into the process of interest, then the alternative process design is obtained, which is the final goal of the redesign framework. An appropriate approach to perform this stage is case-based reasoning (CBR) for reusing past experiences on new situations.

Here the human designer must specify the new requirements that the process must satisfy. Two categories of redesign requirements can be identified: functional requirements and physical requirements. A design specification always contains a single functional requirement; it may also contain a set of physical requirements. A functional requirement represents an abstraction of the intended behaviour of the artefact. It can be a general, specific, or working function. There is no direct association between the function that has to be provided and the physical mechanism that provides it. A physical requirement represents an abstraction of the physical process variables, which satisfy the functional requirement specified in the design specification. It denotes preferences about the designer intentions regarding some aspect of the process. The redesign specification can be represented by the functional and physical requirements or only by physical requirements. A redesign specification is a mean (goal) which is defined in terms of functions that must be

Here In this stage, a diagnostic algorithm is used, that works on the functional concepts identified in the functional analysis. Diagnosis helps to detect "faulty" components (those that do not satisfy the global performance of the process). The design description and the new specifications are used to identify the possible candidates for modification or substitution. The diagnostic algorithm returns an ordered list of units or meta-units. Because the diagnostic algorithm operates over abstract functional concepts, no simulation is required. The diagnostic algorithm does not return the exact unit or meta-unit responsible for the "faulty" behaviour, it returns a list of units/meta-units that do not fulfill the global performance of the process represented by the redesign specification. The human designer is responsible to choose, from the resulting list, the appropriate unit/meta-unit that has to be modified or substituted into the process. Since this unit/meta-unit is connected to others by a flow path, the "cause" and "consequence" units/meta-units also must be identified. A cause unit/meta-unit is(are) the unit(s)/meta-unit(s) situated before the current unit/metaunit in the flow path. They are responsible to provide the appropriate operational conditions to the involved process variables in the function of the unit/meta-unit of interest. A consequence unit/meta-unit is(are) the unit(s)/meta-unit(s) situated after of the current unit/meta-unit in the flow path. They are the unit(s)/meta-unit(s) affected by the operational conditions given by the unit/meta-unit of interest. Both, the cause and the

embodied in a process in order to provide some higher level functionality.

consequence units/meta-units, are not necessarily the closer neighbours.

The aim of this stage is to obtain similar units (equipments) or meta-units (sections) to adapt them into the current process based on the suitable unit/meta-unit identified by the human designer at the last stage. The best way to obtain similar units/meta-units is from similar processes. With the adaptation of any retrieved unit/meta-unit into the process of interest, then the alternative process design is obtained, which is the final goal of the redesign framework. An appropriate approach to perform this stage is case-based reasoning (CBR)

**3.2.1 Specification of redesign requirements** 

**3.2.2 Identification of candidates** 

**3.3 Generation of alternatives** 

for reusing past experiences on new situations.

Starting with the selected unit/meta-unit at the candidate identification stage, similar units/meta-units can be retrieved from other processes. The retrieved unit/meta-unit, which functional and teleological models are the most approximate to the functional and teleological models of the unit/meta-unit of interest, is adapted. This process requires that the performance and operational conditions of the cause and consequence units/meta-units associated with the retrieved unit must be similar to the original case.

According to the later stage, the overall target process is modelled as a graph denoting a hierarchy of functions. Therefore hierarchical case-based reasoning (Smyth 01) is required. Thus, based on the levels of abstraction, two kinds of cases are distinguished:


The organisation of cases into the library is performed according to the type of functions of the unit/meta-unit. In this way, several groups can be distinguished according to the general function type: *source, transport, barrier, storage, balance, and sink*. Within each functional group, units and meta-units are grouped based on their specific functions. Again, within specific functions groups, the units/meta-units are grouped based on the working function achieved. There are not distinctions between units and meta-units with the same specific function. The case library is organised by an abstract hierarchy based on function groups. This structure denotes the organisation of the functional ontology used in the framework.

To retrieve cases from the case library, a similarity engine is used. Only units/meta-units of the same specific functional group are considered. The similarity engine uses functional and teleological targets to search into the library of cases. Functional and teleological models denote strongly the relationship between the units/meta-units and its neighbours. Two types of similarity are computed, local and global, which are defined as follows:


As final result, a set of cases is obtained which contains meta-units (with its corresponding units/meta-units) or units. The set is ranked according to the global similarity between the target case (the unit/meta-unit of interest) and the source cases (the retrieved units/metaunits).

#### **3.4 Adaptation and evaluation**

For this stage the human designer intervention is required since the use of the specialised simulator employed in the data acquisition is mandatory to test the proposed candidates. The adaptation is highly domain-dependent and requires online simulation of the process to verify its correct performance. Since information of abstract cases can not be used directly,

Modelling Approach for Redesign of Technical Processes 115

chemical compounds), functional roles, devices (equipments and connections), measure

Most of the concepts in the ontology correspond to physical entities. The high-level concepts denote very abstract concepts, which can be found in several domains. The middle-level includes the functional concepts proposed in the Multilevel Flow Modelling and Multimodelling approaches, which are: *source, transport, barrier, storage, balance, and sink*. The low-level functional concepts come from the well-known chemical process design methodologies developed by Douglas (Douglas 88) and Turton (Turton 98). The low-level functional concepts can be grouped as: *reaction, separation, temperature change, pressure change, and flow change*. These concepts are called *general functions*. Each specific function is divided into more specific ones named *specific functions*, which denote the function of the equipment into the process. Also each specific function is divided in more specific ones, called *working functions*. A working function can be associated with one or more units and a unit can be related to more than one function. But from the several working functions, only one is the

Over the identified functions an importance functional order and the variables involved in such functions (carried out by the equipments of the process) have been defined. This order was defined with the aim of forming groups of functions where more important functions

units, tasks, operations, and relations.

main function in the process.

1. Acetaldehide from ethanol

3. Ethyl acetate 4. Vinyl acetate 5. Acetone 6. Acetic acid 7. Acrylic acid 8. Cyanhydric acid 9. Nitric acid 10. Acrolein

2. Acetaldehide from ethylene and oxygen

11. Ammonia from natural gas and pure N2

19. Separation of Chlorine-Bencene and Bencene

12. Ammonia from pure N2 and H2 13. Phthalic anhydride from naphtalene 14. Phthalic anhydride from o-Xylene

18. Bencene, Toluene and Styrene

15. Maleic anhydride 16. Bencene and methane 17. Bencene and o-Xylene

20. Ethyl-Bencene 21. Cumene 22. 1,3-Butadiene 23. Cyclohexane 24. Allyl Chloride

"absorb" functions with minor importance, see Figure 8.

The framework was tested over 50 chemical processes (Lopez-Arevalo 05):

the adaptation and revision of equipment must use information of ground cases (real equipments on the simulator). The human designer must fit the ground cases with the process variables involved.

To facilitate the adaptation, an adaptation cost is computed to suggest the human designer the adaptability of the chosen unit/meta-unit. The adaptation cost is based on the differences of the selected unit (source case) and the cause and consequence units/metaunits identified with the diagnostic algorithm. Thus, the cost is a normalised numerical value denoting the difference on the values of the process variables involved in the performance of the unit and the values of the process variables involved in the performance of the neighbour units/meta-units.

The adaptation cost has a value between 0 and 1. Values close to zero mean the adaptation is difficult. The designer experience is determinant because modifications on equipments may affect the overall performance of the process. The modification of the original process, based on the adaptation of a retrieved case, generates an alternative of the process for every unit/meta-unit adapted.

#### **4. Evaluation and results**

The framework has been applied to the Chemical Engineering because chemical processes are suitable for the proposed approach. A chemical plant can be constituted of one or more chemical processes.

The software modules of the redesign framework have been implemented in Java (Sun 05). Additional libraries have been used such as JESS (JESS 08), Ozone (Ozone 08), and The Selection Engine (Wetzel 00). The interaction with the user is done through a graphical interface.

The implementation of the concepts of the proposed redesign framework is based on the following ontological commitments.


Since the framework requires functional concepts, a crucial point is to define the type of functions by means of an ontology. The functional ontology obtained is formed by highlevel and low-level concepts in a similar way to the SUMO (Suggested Upper Merged Ontology) ontology structure (Niles 01). SUMO structures the concepts using metaconcepts, where terminology of general purpose is situated at higher levels, while terminology to specific domains is situated at lower levels. The ontology developed has extended generic concepts of SUMO such as process, objects and mereological and topological concepts. Additional specific concepts have been defined: physico-chemical processes, thermodynamic processes, substances (mass and energy), substance roles (of chemical compounds), functional roles, devices (equipments and connections), measure units, tasks, operations, and relations.

Most of the concepts in the ontology correspond to physical entities. The high-level concepts denote very abstract concepts, which can be found in several domains. The middle-level includes the functional concepts proposed in the Multilevel Flow Modelling and Multimodelling approaches, which are: *source, transport, barrier, storage, balance, and sink*. The low-level functional concepts come from the well-known chemical process design methodologies developed by Douglas (Douglas 88) and Turton (Turton 98). The low-level functional concepts can be grouped as: *reaction, separation, temperature change, pressure change, and flow change*. These concepts are called *general functions*. Each specific function is divided into more specific ones named *specific functions*, which denote the function of the equipment into the process. Also each specific function is divided in more specific ones, called *working functions*. A working function can be associated with one or more units and a unit can be related to more than one function. But from the several working functions, only one is the main function in the process.

Over the identified functions an importance functional order and the variables involved in such functions (carried out by the equipments of the process) have been defined. This order was defined with the aim of forming groups of functions where more important functions "absorb" functions with minor importance, see Figure 8.

The framework was tested over 50 chemical processes (Lopez-Arevalo 05):


114 Advances in Chemical Engineering

the adaptation and revision of equipment must use information of ground cases (real equipments on the simulator). The human designer must fit the ground cases with the

To facilitate the adaptation, an adaptation cost is computed to suggest the human designer the adaptability of the chosen unit/meta-unit. The adaptation cost is based on the differences of the selected unit (source case) and the cause and consequence units/metaunits identified with the diagnostic algorithm. Thus, the cost is a normalised numerical value denoting the difference on the values of the process variables involved in the performance of the unit and the values of the process variables involved in the performance

The adaptation cost has a value between 0 and 1. Values close to zero mean the adaptation is difficult. The designer experience is determinant because modifications on equipments may affect the overall performance of the process. The modification of the original process, based on the adaptation of a retrieved case, generates an alternative of the process for every

The framework has been applied to the Chemical Engineering because chemical processes are suitable for the proposed approach. A chemical plant can be constituted of one or more

The software modules of the redesign framework have been implemented in Java (Sun 05). Additional libraries have been used such as JESS (JESS 08), Ozone (Ozone 08), and The Selection Engine (Wetzel 00). The interaction with the user is done through a graphical

The implementation of the concepts of the proposed redesign framework is based on the

The chemical processes typically operate at steady-state. That means that values of

 A chemical process is constituted of real and abstract units. The abstract units are the sections of the process that appear as atomic elements in conceptual models. All real

 A generic real equipment can be modelled as an object having four attributes: structure, behaviour, function, and teleology to describe all the properties of any real equipment.

Since the framework requires functional concepts, a crucial point is to define the type of functions by means of an ontology. The functional ontology obtained is formed by highlevel and low-level concepts in a similar way to the SUMO (Suggested Upper Merged Ontology) ontology structure (Niles 01). SUMO structures the concepts using metaconcepts, where terminology of general purpose is situated at higher levels, while terminology to specific domains is situated at lower levels. The ontology developed has extended generic concepts of SUMO such as process, objects and mereological and topological concepts. Additional specific concepts have been defined: physico-chemical processes, thermodynamic processes, substances (mass and energy), substance roles (of

equipments can be viewed as descendants of the generic real equipment.

process variables involved.

of the neighbour units/meta-units.

unit/meta-unit adapted.

chemical processes.

interface.

**4. Evaluation and results** 

following ontological commitments.

variables do not change with respect to time.


Modelling Approach for Redesign of Technical Processes 117

The ammonia production process (Figure 9) has been selected as case study because it is one of the most relevant chemical processes in the industry. Ammonia is one of the most important chemicals commodities because of its role in the production of fertiliser and hence of food. It is produced in over 80 countries worldwide with a volume of 130 million tonnes

Technical changes on process equipments were taken into account and some other issues such as economical costs, changes in pipes, environment impact, etc. were not considered. To illustrate the performance only the ammonia production process is used as case study.

First, the data is extracted from the Hysys simulator (Hysys 04), and then the "roles" of chemical substances are asked to the human designer. The first level of the process (abstraction level 0) is shown in Figure 10. This GUI allows the user to interact in two ways,

42. Hydrogen

43. Separation of metane

47. Oxygen and nitrogen 48. Purification of parafins

50. Vinyl chloride

annually (GIA 04).

44. Separation of metane and ethane 45. Methanol from natural gas 46. Methanol from carbon monoxide

49. Propyleneglycol and dipropylene glycol

**4.1 Modelling the ammonia process** 

Fig. 9. The ammonia production process

Fig. 8. The hierarchy of functions


42. Hydrogen

116 Advances in Chemical Engineering

Fig. 8. The hierarchy of functions

30. Ethyl tert-butylic ether (ETBE) 31. Methyl tert-butylic ether (MTBE) 32. Tert-amyl Methyl ether (TAME)

34. Separation of ethane, n-heptane y n-octane

25. Separation of Ciclohexane

28. Purification of Ethanol

26. Chloroform 27. Ethanol

29. Dimethyl ether

33. Styrene

35. Ethylene 36. Ethylene oxide 37. Formaldehyde 38. Formaline 39. Methyl formate

40. HP gas 41. Heptane


#### **4.1 Modelling the ammonia process**

The ammonia production process (Figure 9) has been selected as case study because it is one of the most relevant chemical processes in the industry. Ammonia is one of the most important chemicals commodities because of its role in the production of fertiliser and hence of food. It is produced in over 80 countries worldwide with a volume of 130 million tonnes annually (GIA 04).

Fig. 9. The ammonia production process

Technical changes on process equipments were taken into account and some other issues such as economical costs, changes in pipes, environment impact, etc. were not considered. To illustrate the performance only the ammonia production process is used as case study.

First, the data is extracted from the Hysys simulator (Hysys 04), and then the "roles" of chemical substances are asked to the human designer. The first level of the process (abstraction level 0) is shown in Figure 10. This GUI allows the user to interact in two ways,

Modelling Approach for Redesign of Technical Processes 119

Flow decrement Splitter TEE-100

Pressure increment Compressor K-100

Temperature increment Cooler E-101

Temperature exchange Heat Exchanger E-102

MIX-100 MIX-101 MIX-102

TEE-101

VLV-100 VLV-101 VLV-10

E-103

E-104

V-101

PFR-100 PFR-101 PFR-102

**General Function Specific Function Working Function Label** 

Flow increment Mixer

Pressure decrement Valve

Separation Distillation Flash Separator V-100

Reaction Reaction Tubular Reactor

Table 1. Equipment and functions in the ammonia production process.

Flow change

Pressure change

Temperature change

Fig. 11. Grouping of flow change units.

Fig. 10. First representation of the ammonia production process.

through menus and panels. The diagram panel (upper-right panel) allows the user to manipulate the process layout; the user can organise the components of the process according to its needs. The navigation panel (upper-left panel) is used to navigate into the levels of the process; abstraction levels and its corresponding components. The information panel (bottom panel) displays information about operations carried out in the prototypes. Table 1 summarises the type of equipments and its corresponding functions.

For the generation of meta-units, the flow change equipments (mixers and splitters) are grouped to other ones in the abstraction level 1, as it is shown in Figure 11.

The different abstract levels were automatically generated until the final abstract models were created at level 6 (see Figure 12). The abstraction process continues until the whole process is represented by just one meta-unit. Only new units and meta-units are represented to illustrate how the functional groups are created. In this sense, connections between units and meta-units in the same level have not been represented. The scheme represents the process by means of groups of the general class of type of equipment; its general function can be deduced from them. Figure 12 represents all the unit and meta-units in each level, in similar way that they are presented to the designer.

Fig. 10. First representation of the ammonia production process.

Table 1 summarises the type of equipments and its corresponding functions.

grouped to other ones in the abstraction level 1, as it is shown in Figure 11.

similar way that they are presented to the designer.

through menus and panels. The diagram panel (upper-right panel) allows the user to manipulate the process layout; the user can organise the components of the process according to its needs. The navigation panel (upper-left panel) is used to navigate into the levels of the process; abstraction levels and its corresponding components. The information panel (bottom panel) displays information about operations carried out in the prototypes.

For the generation of meta-units, the flow change equipments (mixers and splitters) are

The different abstract levels were automatically generated until the final abstract models were created at level 6 (see Figure 12). The abstraction process continues until the whole process is represented by just one meta-unit. Only new units and meta-units are represented to illustrate how the functional groups are created. In this sense, connections between units and meta-units in the same level have not been represented. The scheme represents the process by means of groups of the general class of type of equipment; its general function can be deduced from them. Figure 12 represents all the unit and meta-units in each level, in


Table 1. Equipment and functions in the ammonia production process.

Fig. 11. Grouping of flow change units.

Modelling Approach for Redesign of Technical Processes 121

Therefore, the focus will be on reactors, where the ammonia is originated. Since all the roles of the chemical substances are known, the diagnostic module focuses on the concentration of ammonia (which has the main product role). Thus all values related to this substance are

We are interested in finding where the main product is produced. The search starts at the highest level in the hierarchy -abstraction level 6- following the flow direction, from left to right. As result of the search, all the units and meta-units affecting the concentration variable

analysed.

have been identified, as shows Table 2.

**Abstraction level Identified components**  6 meta-reactor-8 5 meta-reactor-7

1 reactor-2

Fig. 13. Units composing the meta-reactor-3.

Table 2. Identified candidates

4 meta-reactor-6, meta-reactor-4

2 reactor-1, meta-reactor-1, reactor-3

3 meta-reactor-5, meta-reactor-3, meta-reactor-2

Since modifications to the process can be performed only at ground level, the cause and consequence units are searched in this level. To illustrate the cause and consequence identification, assume that we focus on the meta-reactor-3, which at ground level includes the units PFR-101, MIX-100, and VLV-100 (Figure 13). In the state analysis, the state

Fig. 12. Hierarchical representation of the ammonia production process in bottom-up direction.

#### **4.2 Identification of candidates**

According to the new design objective(s) the process must fulfill, modifications on reactors are used to illustrate the results of the framework:

#### Requirement

*The redesign problem is the increase of the production of ammonia by 15% in the plant represented by the scheme of Figure 10.* 

The human designer first needs to identify the variables that may affect directly the production of ammonia. Then the designer identifies that the increase of production can be achieved by modifying any of the next conditions:


This gives an idea on the types of equipment the diagnosis must focus on. Assuming that the concentration variable is selected, this is affected by reactors and separators. Initially, reactors affect the concentration of product because they produce the main product, and separators affect it in secondary manner by incrementing the purity of the product.

Fig. 12. Hierarchical representation of the ammonia production process in bottom-up

According to the new design objective(s) the process must fulfill, modifications on reactors

*The redesign problem is the increase of the production of ammonia by 15% in the plant* 

The human designer first needs to identify the variables that may affect directly the production of ammonia. Then the designer identifies that the increase of production can be

This gives an idea on the types of equipment the diagnosis must focus on. Assuming that the concentration variable is selected, this is affected by reactors and separators. Initially, reactors affect the concentration of product because they produce the main product, and separators affect it in secondary manner by incrementing the purity of the product.

direction.

Requirement

 pressure temperature concentration

**4.2 Identification of candidates** 

are used to illustrate the results of the framework:

achieved by modifying any of the next conditions:

*represented by the scheme of Figure 10.* 

Therefore, the focus will be on reactors, where the ammonia is originated. Since all the roles of the chemical substances are known, the diagnostic module focuses on the concentration of ammonia (which has the main product role). Thus all values related to this substance are analysed.

We are interested in finding where the main product is produced. The search starts at the highest level in the hierarchy -abstraction level 6- following the flow direction, from left to right. As result of the search, all the units and meta-units affecting the concentration variable have been identified, as shows Table 2.


Table 2. Identified candidates

Since modifications to the process can be performed only at ground level, the cause and consequence units are searched in this level. To illustrate the cause and consequence identification, assume that we focus on the meta-reactor-3, which at ground level includes the units PFR-101, MIX-100, and VLV-100 (Figure 13). In the state analysis, the state

Fig. 13. Units composing the meta-reactor-3.

Modelling Approach for Redesign of Technical Processes 123

A CBR module is used to obtain alternatives units/meta-units that may be adapted into the ammonia process. Again, the process representation showed in Figure 12 is used to denote the composition of meta-units; and the process representation showed in Figure 13 to denote

Assuming a threshold of 14 items, the most similar source cases from computing the global similarities are summarised in Table 4. It shows the percentage of similarity, the specific

**Rank Similarity Function Inlet Function Outlet Function** 

1 56% meta-reactor reaction separation 2 43% meta-reactor reaction flow change 3 37% tubular reactor tmp change separation 4 31% plug flow reactor pres change tmp change 5 30% meta-reactor separation pres change 6 29% tubular reactor reaction separation 7 29% meta-reactor tmp change tmp change 8 27% meta-reactor pres change separation 9 25% tubular reactor tmp change tmp change 10 23% tubular reactor tmp change separation 11 20% plug flow reactor flow change pres change 12 20% meta-reactor pres change tmp change 13 16% meta-reactor reaction tmp change 14 15% meta-reactor flow change tmp change

Table 4. Result of the global similarity computation for meta-reactor-3.

comprehensive and clear representations of equipment/sections

The aspects considered in the evaluation of the framework were:

suitable grouping of equipment/sections

easy and intuitive graphical interface

1. Modelling of the process

use of simplified models

intuitive goal-driven approach

The modelling of the 50 processes generated 1590 cases in the case library. Therefore, the software prototypes were continually enhanced according to the needs of these processes.

**4.3 Generation of alternatives** 

the presence of units/meta-units in each abstraction level.

function, the inlet and outlet functions of the source case.

conditions are propagated to the units connected to the meta-reactor-3 in the flow path. The analysis in backward/forward stream directions finishes when a closer primary function is reached.

The state of the meta-reactor-3 is set to low capacity because the production of the main product is not enough. Since this meta-unit is not an initial unit in the path, its state may be originated by the effect of the performance of other units. Then back units are analysed. Considering the stream-4, the function to analyse is a source (PFR-100), which directly affects the concentration variable. Perhaps other back units in the same direction may affect the variable, but this is the closer primary function affecting the concentration variable. It may have low volume state, which originates the low capacity of meta-reactor-3. Therefore, the unit associated to this function is identified as cause unit and this branch of backward analysis in this direction finishes. Considering the stream-7, the function to analyse is a balance (TEE-100), which does not affect the variable. Then, the next function is analysed, which is a balance of temperature (E-102) that again does not affect directly the variable. The next function is storage (V-101), which affects the variable. This is another primary function affecting the concentration variable; it may have low volume state. Then its associated unit is a cause unit and this branch of backward analysis finishes.

Now, forward analysis is carried out following the output stream of the functional group. The low capacity state of meta-reactor-3 originates a low flow state and a low volume state, and consequently affects the following source functions producing a low capacity state in such function. Thus, considering the stream-2, the function to analyse is a balance (MIX-101), which does not affect the concentration variable. The next function is source (PFR-102), which affects the variable; it may have low capacity state originated for the low capacity state of the meta-reactor-3. Then, the unit associated with this function is a consequence unit. Since is the closer primary function affected in the forward stream direction, the forward analysis finishes.


Therefore, the identified cause and consequence units for meta-reactor-3 are shown in Table 3, which also represents the cause and consequence units for all the meta-units.

Table 3. Cause and consequence units.

#### **4.3 Generation of alternatives**

122 Advances in Chemical Engineering

conditions are propagated to the units connected to the meta-reactor-3 in the flow path. The analysis in backward/forward stream directions finishes when a closer primary function is

The state of the meta-reactor-3 is set to low capacity because the production of the main product is not enough. Since this meta-unit is not an initial unit in the path, its state may be originated by the effect of the performance of other units. Then back units are analysed. Considering the stream-4, the function to analyse is a source (PFR-100), which directly affects the concentration variable. Perhaps other back units in the same direction may affect the variable, but this is the closer primary function affecting the concentration variable. It may have low volume state, which originates the low capacity of meta-reactor-3. Therefore, the unit associated to this function is identified as cause unit and this branch of backward analysis in this direction finishes. Considering the stream-7, the function to analyse is a balance (TEE-100), which does not affect the variable. Then, the next function is analysed, which is a balance of temperature (E-102) that again does not affect directly the variable. The next function is storage (V-101), which affects the variable. This is another primary function affecting the concentration variable; it may have low volume state. Then its associated unit

Now, forward analysis is carried out following the output stream of the functional group. The low capacity state of meta-reactor-3 originates a low flow state and a low volume state, and consequently affects the following source functions producing a low capacity state in such function. Thus, considering the stream-2, the function to analyse is a balance (MIX-101), which does not affect the concentration variable. The next function is source (PFR-102), which affects the variable; it may have low capacity state originated for the low capacity state of the meta-reactor-3. Then, the unit associated with this function is a consequence unit. Since is the closer primary function affected in the forward stream direction, the

Therefore, the identified cause and consequence units for meta-reactor-3 are shown in Table

separator-2 (V-101) reactor-3 (PFR-102)

separator-2 (V-101) reactor-3 (PFR-102)

3, which also represents the cause and consequence units for all the meta-units.

**Candidate Cause units Consequence units**  reactor-1 (PFR-100) separator-2 (V-101) reactor-2 (PFR-101) reactor-2 (PFR-101) reactor-1 (PFR-100) reactor-3 (PFR-102) reactor-3 (PFR-102) reactor-2 (PFR-101) separator-1 (V-101)

meta-reactor-2 reactor-2 (PFR-101) separator-1 (V-101)

meta-reactor-4 reactor-2 (PFR-101) separator-1 (V-101) meta-reactor-5 separator-2 (V-101) reactor-2 (PFR-101) meta-reactor-6 separator-2 (V-101) reactor-3 (PFR-102) meta-reactor-7 separator-2 (V-101) separator-1 (V-101)

is a cause unit and this branch of backward analysis finishes.

reached.

forward analysis finishes.

meta-reactor-1 reactor-1 (PFR-100)

meta-reactor-3 reactor-1 (PFR-100)

Table 3. Cause and consequence units.

A CBR module is used to obtain alternatives units/meta-units that may be adapted into the ammonia process. Again, the process representation showed in Figure 12 is used to denote the composition of meta-units; and the process representation showed in Figure 13 to denote the presence of units/meta-units in each abstraction level.

Assuming a threshold of 14 items, the most similar source cases from computing the global similarities are summarised in Table 4. It shows the percentage of similarity, the specific function, the inlet and outlet functions of the source case.


Table 4. Result of the global similarity computation for meta-reactor-3.

The modelling of the 50 processes generated 1590 cases in the case library. Therefore, the software prototypes were continually enhanced according to the needs of these processes.

The aspects considered in the evaluation of the framework were:

	- use of simplified models
	- suitable grouping of equipment/sections
	- intuitive goal-driven approach
	- comprehensive and clear representations of equipment/sections
	- easy and intuitive graphical interface

Modelling Approach for Redesign of Technical Processes 125

This research was partially funded by project number 175537 from "Fondo Mixto Conacyt-

(Aamodt 94) A. Aamodt & E. Plaza. Case-based reasoning: Foundational issues,

(Arana 00) I. Arana H. Ahriz & P. Fothergill. Improving re-design support. In Proceedings of

(Arana 01) I. Arana H. Ahriz & P. Fothergill. Redesign Knowledge Analysis, Representation

(Akin 82) O. Akin. Representation and architecture. Silver Spring, O. Akin and E.F. Weinel

(Alberts 92) L.K. Alberts P.M. Wognum & N.J. Mars. Structuring design knowledge on the

(Alberts 93a) L.K. Alberts. YMIR: An ontology for engineering design. PhD thesis,

(Alberts 93b) L.K. Alberts R.R. Bakker D. Beekman & P.M. Wognum. Model-based redesign

(Bakker 94) R.R. Bakker S.J. van Eldonk P.M. Wognum & N.I. Mars. The use of model-based

(Berker 96) I. Berker & D.C. Brown. Conflicts and Negotiation in Single Function Agent

(Bertalanffy 50) L.von Bertalan.y. An Outline of General Systems Theory. British Journal for

(Bhatta 94) S. Bhatta A. Goel & S. Prabhakar. Innovation in Analogical Design: A Model-

(Bo 99) Y. Bo & F. Salustri. Function Modeling Based on Interactions of Mass, Energy and

(Borner 96) K. Borner E. Pipping E. Tammer & C. Coulon. Structural Similarity and

Kluwer Academic Publishers, Dordrecht, The Netherlands.

Information. Florida Artificial Intelligence Research Society, 1999.

Conference on Artificial Intelligence, pages 647–651, A. Cohn Ed., 1994. (Ball 92) N.R. Ball & F. Bauert. The Integrated Design Framework: supporting the design

Gero, J.S. (Ed.), pages 327–348, 1992. Kluwer academic publishers.

eds., Maryland, Information Dynamics Inc., 1982.

University of Twente, 1993. Netherlands.

the Philosophy of Science, vol. 1, no. 2, 1950.

Reasoning, pages 58–75, 1996. Springer-Verlag.

Gero, J.S. (Ed.), pages 639–656, 1992. Kluwer, Dordrecht.

methodological variations and system approaches. AI Communications, vol. 7,

the Fifth World Conference on Integrated Design and Process Technology

and Reuse. Industrial Knowledge Management A Micro-level Approach, 2001.

basis of generic components. Artificial Intelligence in Design (AID'92), vol. In:

of technical systems. Proceedings of the 4th international workshop on principles of

diagnosis in redesign. In Proc. Reasoning about physical systems. 11th European

process using a blackboard system. Artificial Intelligence in Design (AID'92), vol.

Based Design Systems. Concurrent Engineering: Research and Applications, Journal, Special Issue: Multi Agent Systems in Concurrent Engineering, vol. 4, no. 1, pages 17–33, 1996. Brown, D.C., Landes, S.E. and Petrie, C.J. (Eds.), Technomic

Based Approach. Proceedings of Artificial Intelligence in Design, pages 57–74, 1994.

Adaptation. Proceedings of the Third European Workshop on Case-Based

**6. Acknowledgment** 

**7. References** 

Gobierno del Estado de Tamaulipas".

pages 39–59, 1994.

(IDPT2000), 2000.

Springer-Verlag.

diagnosis, 1993.

Publishing Inc.

	- clear and easy search over simple but consistent concepts
	- module easy to use
	- intuitive interpretation of results
	- suggestions according to purpose-driven strategy
	- appropriate guidelines for modification/substitution
	- reuse of past design solutions
	- easy access to abstract and detailed data of proposed solutions
	- rapid response making agile the creation of alternative prototypes

#### **5. Conclusions**

This chapter describes a redesign support framework for technical processes based on hierarchical modelling. This hierarchical modelling is based on means-end and whole-parts aspects. The hierarchical representation enhances the reasoning mechanism to identify the elements to be modified and the possible alternatives. The framework focused on conceptual redesign issues where abstract models are employed. The processes are modelled hierarchically based on their functions and goals. The framework consists of four stages: design-description acquisition, identification of candidates to redesigned, generation of alternatives, and adaptation and evaluation of alternatives. The implementation of the framework was on the Chemical Engineering domain.

The redesign framework combines model-based reasoning and case-base reasoning techniques. This framework enables the designer to work directly with the conceptual design of an existing process (i.e. a process already in operation) to automatically generate abstract multiple-models, which can be modified to develop alternative process designs. The procedure can be seen as the reverse engineering approach of "replay and modify". This model-based approach provides an appropriate way of combining hierarchical and functional modelling to represent and reason about complex technical processes. The hierarchical case-based approach provides a systematic way of reusing the sections of previous processes.

The framework extends the use of Multimodelling and Multilevel Flow Modelling approaches to integrate mental abstract models about the behaviour of processes in the redesign activities. These models provide a more intuitive vision of reasoning on each task to be performed, and thus the redesign activities are enhanced.

The research has some limitations, some of the major ones are:


#### **6. Acknowledgment**

This research was partially funded by project number 175537 from "Fondo Mixto Conacyt-Gobierno del Estado de Tamaulipas".

#### **7. References**

124 Advances in Chemical Engineering

This chapter describes a redesign support framework for technical processes based on hierarchical modelling. This hierarchical modelling is based on means-end and whole-parts aspects. The hierarchical representation enhances the reasoning mechanism to identify the elements to be modified and the possible alternatives. The framework focused on conceptual redesign issues where abstract models are employed. The processes are modelled hierarchically based on their functions and goals. The framework consists of four stages: design-description acquisition, identification of candidates to redesigned, generation of alternatives, and adaptation and evaluation of alternatives. The implementation of the

The redesign framework combines model-based reasoning and case-base reasoning techniques. This framework enables the designer to work directly with the conceptual design of an existing process (i.e. a process already in operation) to automatically generate abstract multiple-models, which can be modified to develop alternative process designs. The procedure can be seen as the reverse engineering approach of "replay and modify". This model-based approach provides an appropriate way of combining hierarchical and functional modelling to represent and reason about complex technical processes. The hierarchical case-based approach provides a systematic way of reusing the sections of

The framework extends the use of Multimodelling and Multilevel Flow Modelling approaches to integrate mental abstract models about the behaviour of processes in the redesign activities. These models provide a more intuitive vision of reasoning on each task

The framework was implemented only in one domain. The ideas can be applied to

 The framework was tested with simulated plants. We did not have access to real plant information, but the results obtained were validated by a team of chemical engineers

The implementation of the framework is not manageable by novice users because

transparent integration with the numerical simulator

 suggestions according to purpose-driven strategy appropriate guidelines for modification/substitution

framework was on the Chemical Engineering domain.

to be performed, and thus the redesign activities are enhanced. The research has some limitations, some of the major ones are:

important human designer decisions must be taken.

specialised in design of processes.

another domain, but a new implementation will be necessary.

clear and easy search over simple but consistent concepts

 easy access to abstract and detailed data of proposed solutions rapid response making agile the creation of alternative prototypes

2. Identification of candidates

module easy to use

**5. Conclusions** 

previous processes.

 intuitive interpretation of results 3. Suggestion of equipment/sections

reuse of past design solutions


Modelling Approach for Redesign of Technical Processes 127

(Daube 89) F. Daube & B. Hayes-Roth. A Case-Based Mechanical Redesign System. In

(de Silva Garza 96) A. Gomez de Silva Garza & M.L. Maher. Design by interactive

(DeLoach 04) S.A. DeLoach. The MaSE Methodology. Methodologies and Software

(Dixon 89) J.R. Dixon M.J. Guenette R.K. Irani E.H. Nielsen M.F. Orelup & R.V. Welch.

(Douglas 88) J.M. Douglas. Conceptual design of chemical processes. Mc Graw Hill, New

(Dove 2001) Dove R. Response ability: the language, structure, and culture of the agile

(Dunskus 95) B.V. Dunskus D.L. Grecu D.C. Brown & I. Berker. Using Single Function

(Eldonk 96) S.J. van Eldonk L.K. Alberts R.R. Bakker F. Diker & P.M. Wognum. Redesign of

(Falkenhainer 91) B. Falkenhainer & K.D. Forbus. Compositional Modeling: Finding the Right Model for the Job. Artificial Intelligence, vol. 51, pages 95–143, 1991. (Falkenhainer 92) B. Falkenhainer & K.D. Forbus. Composing task-specific models.

(Fensel 01b) D. Fensel. Ontologies: A silver bullet for knowledge management and electronic

(Fischer 87) G. Fischer A.C. Lemke & C. Rathke. From design to redesign. In Proceedings of

(Fischo. 78) B. Fischo. P. Slovic & S. Lichtenstein. Fault trees: Sensitivity of estimated failure

(Forbus 84) K.D. Forbus. Qualitative Process Theory. Artificial Intelligence, vol. 24, pages

(Forster 96) J. Forster I. Arana & P. Fothergill. Redesign knowledge representation with

the 9th International Conference on Software Engineering, pages 369–376, IEEE

probabilities to problem representation. Journal of Experimental Psychology:

DEKLARE. In Proceedings of KEML'96: 6th Workshop on Knowledge Engineering:

Techniques for Reuse of Designs, vol. 9, pages 93–104, 1996.

Automated Modelling, ASME, vol. 41, pages 1–9, 1992.

Human Perception and Performance, vol. 4, 1978.

Methods and Languages, Paris, France, 1996.

commerce. Heidelberg, Germany, 2001.

Computer Society Press, 1987.

85–168, 1984.

(IJCAI-89) Vol. 2, pages 1402–1407, Morgan Kaufmann Publishers, 1989. (de Kleer 79) J. de Kleer. Causal and Teleological Reasoning in Circuit Recongnition. PhD

thesis, Massachusetts Institute of Technology, 1979.

1, pages 151–161, 1996.

Massachusetts, USA.

York, 1988.

Kluwer Academic Publishing.

enterprise. Wiley, New York.

no. 4, pages 299–312, 1995.

Proceedings of the 11th International Joint Conference on Artificial Intelligence

exploration using memory based techniques. Knowledge based systems, vol. 9, no.

Engineering for Agent Systems. The Agent-Oriented Software Engineering Handbook Series : Multiagent Systems, Artificial Societies and Simulated Organizations, vol. 11, 2004. Bergenti, F., Gleizes, M.P., Zambonelli, F. (Eds.)

Computer-based models of design processes: the evaluation of design for redesign. NSF Engineering Design Research Conference, pages 491–506, 1989. University of

Agents to Investigate Con.ict. Artificial Intelligence in Engineering Design and Manufacturing (AIEDAM), Special Issue: Conflict Management in Design, vol. 9,

technical systems. Knowledge-based Systems, Special Issue on Models and


(Bras 92) B.A. Bras. Foundation for designing decision-based design processes. PhD thesis,

(Brazier 94) F.M. Brazier P.H. Van Langen Zs. Ruttkay & J. Treur. On formal speci.-cation of

F. (Eds.), pages 535–552, 1994. Kluwer Academic Publishers, Dordrecht. (Brazier 96) F.M. Brazier P.H. Van Langen J. Treur & N.J. Wijngaards. Redesign and reuse in

Architectural and Design Science, University of Sydney, Australia.

structures and control strategies. Morgan Kaufmann, 1989.

Expert, vol. 12, no. 2, pages 14–16, 1997.

Univer-sity Press. Cambridge, 1991.

form. Design Studies 24(2):173-180.

pages 48–56, 1993.

159– 170, 1994.

1981.

Magazine, vol. 11, no. 4, pages 59–71, 1990.

Engineering, vol. 16, pages 162–177, 2000.

vol. 23, no. 6, pages 1718–1751, 1993.

(Brown 89) D.C. Brown & B. Chandrasekaran. Design problem solving: knowledge

(Brown 97) D.C. Brown & W.P. Birmingham. Understanding the Nature of Design. IEEE

(Brown 98) D.C. Brown. Intelligent Computer-Aided Design. AI in Design Group, Com-puter Science Department, WPI, September 1998. Worcester, MA, USA. (Buckingham 91) S. Buckingham Shum. Cognitive Dimensions of Design Rationale. In

(Chang 2003) Chang WC, Van YT. Researching design trends for the redesign of product

(Chandrasekaran 90) B. Chandrasekaran. Design problem solving: a task analysis. AI

(Chandrasekaran 93) B. Chandrasekaran A. Goel & Y. Iwasaki. Functional Representation as

(Chandrasekaran 00) B. Chandrasekaran & J.R. Josephson. Function in Device

(Checkland 81) P. Checkland. Systems thinking, systems practice. John Wiley and Sons,

(Chen 2009) Li, S. and Chen, L., 2009, Pattern-based Reasoning for Rapid Redesign: A Proactive Approach. Research in Engineering Design, Vol. 21, pp. 25-42. (Chittaro 93) L. Chittaro G. Guida C. Tasso & E. Toppano. Functional and Teleological

(Chittaro 98) L. Chittaro & A.N. Kumar. Reasoning about function and its applications to engineering. Artificial Intelligence in Engineering, vol. 12, pages 331–336, 1998. (Culley 99) S.J. Culley. Final Report -Future Issues For Design Research Workshop. Technical Report, Faculty of Engineering and Design, University of Bath, 1999. (Das 94) D. Das S. Gupta & D. Nau. Reducing setup cost by automated generation of

design tasks. Arti.cial Intelligence in Design (AID'94), vol. Gero, J.S. and Sudweeks,

compositional knowledge-based systems. Knowledge Based Systems, Special Issue on Models and Techniques for Reuse of Designs, vol. 9, no. 2, pages 105–119, 1996. (Bridge 97) C. Bridge. From design to redesign: Revisiting design models to highlight

similarities and differences. Seminar Presentation, 1997. Department of

People and Computers VI: Proceedings of HCI'91, pages 331–344, Cambridge

Design Rationale. IEEE Computer, no. Special Issue on Concurrent Engineering,

Representation. Engineering with Computers, Special Issue on Computer Aided

Knowledge in the Multimodeling Approach for Reasoning about Physical Systems: A case study in diagnosis. IEEE Transactions on Systems Man. and Cybernetics.,

redesign suggestions. Proc. ASME Computers in Engineering Conference, pages

University of Houston, 1992.


Modelling Approach for Redesign of Technical Processes 129

(Grossmann 00) I.E. Grossmann & A.W. Westerberg. Research Challenges in Process Systems Engineering. AICHE Journal, vol. 46, no. 9, pages 1700–1703, 2000. (Gruber 93) T.R. Gruber. A translation approach to portable ontology specifications.

(Gundersen 90) T. Gundersen. Retrofit Process Design -Research and Applications of

(Heo 98) D.H. Heo A.C. Parker & C.P. Ravikumar. An Evolutionary Approach to System

(Hertwig 01) T.A. Hertwig A. Xu A.B. Nagy R.W. Pike J.R. Hopper & C.L. Yaws. A

(Howe 86) A. Howe P. Cohen J. Dixon & M.D. Simmons. A domain-independent program

(Hysys 04) Hysys. Introduction to hysys.plant ver. 3.1. AEA Technology Engineering

(JESS 08) JESS. Jess 7.1 manual. http://herzberg.ca.sandia.gov/jess/docs/index.shtml, 2008. (Keuneke 91) A.M. Keuneke. Device Representation: The Significance of Functional

(Kim 93) G.Y. Kim & G. Bekey. Construting Design Plans for DFA Redesign. Proc. IEEE International Conference on Robotics and Automation, pages 312–318, 1993. (Kirkwood 88) R.L. Kirkwood M.H. Locke & J.M. Douglas. A prototype expert system for

(Kitamura 98) Y. Kitamura & R. Mizoguchi. Functional Ontology for Functional

(Kitamura 99) Y. Kitamura & R. Mizoguchi. Towards Redesign based on Ontologies of

(Kolodner 93) J. Kolodner. Case-based reasoning. Morgan Kaufman Publishers, Inc., 1993. (Kraslawski 00) R. Ben-Guang H. Fan-Yu A. Kraslawski & L. Nystrom. Review: study on the

synthesizing chemical process flowsheets. Comp. Chem. Eng., vol. 12, no. 4, pages

Understanding. Proceedings of The Twelfth International Workshop on Qualitative

Functional Concepts and Redesign Strategies. In Second International Workshop On Strategic Knowledge And Concept Formation, page JSPS8, Iwate, Japan, 1999.

methodology for retrofitting chemical processes. Chemical Engineering

240, 1990. J.J. Siirola, I. E. Grossmann and G. Stephanopoulos (Eds.). (Han 95) Ch. Han J.M. Douglas G. Stephanopoulos. Agent-Based Approach to a Design

Chemical Engineering, vol. 19 (Supplement), pages S63–S69, 1995.

Systematic Methods. Foundations of Computer-Aided Process Design, pages 213–

Support System For the Synthesis of Continuous Chemical Processes. Computers in

Redesign. Proc. Eleventh International Conference on VLSI Design, pages 359–362,

prototype system for economic, environmental and sustainable optimization of a chemical complex. Proc. 11th European Symposium on Computer Aided Process Engineering, pages 1017–1022, 2001. R. Gani and S.B. Jorgensen eds, Elsevier. (Hoover 91) S.P. Hoover J.R. Rinderle S. Finger. Models and abstractions in design. In

Proceedings of the International Conf. on Engineering Design, ICED'91, pages 46–

for mechanical engineering design. In Proceedings of the 1st International Conference on Applications of AI in Engineering Problems, pages 289–299,

Knowledge Acquisition, vol. 5, no. 2, pages 199–220, 1993.

1998. Chennai, India.

57, Zurich, 1991.

329–343, 1988.

Southampton University, U.K., 1986.

Software, Hyprotech, Calgary, Canada, 2004.

Knowledge. IEEE Expert, vol. 6, no. 2, pages 22–25, 1991.

Reasoning, AAAI Press, pages 77–87, 1998. Cape Cod, USA.

Technology, vol. 23, no. 6, pages 479–484, 2000.


 (Goel 91) A. Goel. A model-based approach to case adaptation. In In Proceedings of the 13th Annual Conference of the Cognitive Science Society (CogSci'91), pages 143–148, Chicago, Illinois, 1991.


(Forster 97b) J. Forster P. Fothergill & I. Arana. Enabling intelligent variant design using

(Fothergill 95) P. Fothergill J. Forster J. A. Lacunza F. Plaza & I. Arana. DEKLARE: A

(Franke 92) D.W. Franke. A theory of teleology. PhD thesis, University of Texas at Austin,

(French 85) R. French & J. Mostow. Toward Better Models of the Design Process. AI

(French 93) M.J. French R.V. Chaplin & P.M. Langdon. A creativity aid for designers.

(Fricke 2005) Fricke E, Schulz A P. Design for changeability (DfC): principles to enable

(Gero 90a) J.S. Gero. Design prototypes: a knowledge representation schema for design. AI

(Gero 90b) J.S. Gero & A. Rosenman. A conceptual framework for knowledge-based de-sign

(Gero 04) J.S. Gero & U. Kannengiesser. The situated Function-Behaviour-Structure

(Goel 89) A. Goel & B. Chandrasekaran. Functional representation of designs and re-design

(Goel 92) A. Goel & B. Chandrasekaran. Case-Based Design: A Task Analysis. Artificial

(Goel 94b) A. Goel & S. Prabhakar. A control architecture for redesign and design

(Gomez-Perez 04) A. Gomez-Perez R. Gonzalez-Cabero & M. Lama. Development of

pages 165–184, 1992. Academic Press, Tong y D. Sriram (editors).

framework. Design Studies, vol. 25, no. 4, pages 373–391, 2004.

1997. London.

Hague, Netherlands.

359. Wiley Periodicals.

Publishers.

Reasoning, 1997.

1992.

Tierney (Eds.), Vienna. IOS Press.

Magazine, vol. 6, no. 1, pages 44–57, 1985.

Magazine, vol. 11, no. 4, pages 26–36, 1990.

Engineering, vol. 5, no. 2, pages 65–77, 1990.

(GIA 04) GIA. Greener Industry Association: Ammonia Process. http://www.uyseg.org/greener industry/index.htm, 2004.

Web Service (ECOWS'04), pages 72–86, 2004.

143–148, Chicago, Illinois, 1991.

constraints. IEE Intelligent Design Systems Colloquium, vol. Digest No: 97/016,

methodological approach to re-design. Proceedings of Conference on Integration in Manufacturing, pages 109–122, 1995. K. R. von Barisani, P. A. MacConaill and K.

International Conference on Engineering Design, ICED'93, pages 53–59, 1993. The

changes in systems throughout their entire lifecycle. Systems Engineering 8(4):342-

research at Sydney University's design computing unit. Artificial Intelligence in

problem solving. Proc. Eleventh International Joint Conference on Arti.cial Intelligence, pages 1388–1394, 1989. Los Altos, California: Morgan Kaufmann

 (Goel 91) A. Goel. A model-based approach to case adaptation. In In Proceedings of the 13th Annual Conference of the Cognitive Science Society (CogSci'91), pages

Intelligence Approaches to Engineering Design, Innovative Design, vol. II, no. 6,

veri.-cation. Proc. Second Australian and New Zealand Conference on Intelligent Information Systems Conference, pages 377–381, 1994. Brisbane, Qld., Australia. (Goel 97a) A. Goel A. Gomez de Silva Garza N. Grue J.W. Murdock & M.M. Recker.

Functional Explanations in Design. IJCAI-97 Workshop on Modeling and

Semantic Web Services at the Knowledge Level. In The European Conference on


Modelling Approach for Redesign of Technical Processes 131

(Papoulias 83) S.A. Papoulias & I.E. Grossman. A Structural Optimization Approach in

(Pasanen 01) A. Pasanen. Phenomenon-Driven Process Design Methodology. PhD the-sis,

(Pistikopoulos 87) E.N. Pistikopoulos & I.E. Grossman. Optimal Retrofit Design for

(Price 97) C.J. Price I.S. Pegler M.B. Ratcli.e & A. McManus. From troubleshooting to process

(Price 98) C.J. Price. Function-directed electrical design analysis. Artificial Intelligence in

(Price 03) C.J. Price N.A. Snooke & S.D. Lewis. Adaptable Modeling of Electrical Systems. In

(Qian 92) L. Qian & J.S. Gero. A design support system using analogy. Proceedings of the

(Rapoport 94) H. Rapoport R. Lavie & E. Kehat. Retrofit design of new units into an existing

(Rasmussen 85) J. Rasmussen. The Role of hierarchical knowledge representation in decision

(Rasmussen 86) J. Rasmussen. Information processing and human-machine interaction: An

(Salomons 95) O.W. Salomons. Computer Support in the Design of Mechanical Products.

(Sasajima 95) M. Sasajima Y. Kitamura M. Ikeda & R. Mizoguchi. FBRL: A Function and Behavior Representation Language. Proceedings of IJCAI, pages 1830–1836, 1995. (Schoen 91) E. Schoen. Intelligent Assistance for the Design of Knowledge based Systems.

(Sembugamoorthy 86) V. Sembugamoorthy & B. Chandrasekaran. Functional

(Sylvester 00) R.W. Sylvester W.D. Smith & J. Carberry. Information and modelling for greener process design. AIChE Symp. Series, vol. 96, no. 323, pages 26–30, 2000.

Represen-tation of Devices and Compilation of Diagnostic Problem-Solving Systems. Experience, Memory and Reasoning, J.L. Kolodner and C.K. Riesbeck,

Engineering Design Research Center. Carnegie Mellon University, 1987. (Pos 97) A. Pos. Problem Solving for Redesign. 10th European Knowledge Workshop on

Process Synthesis II: Heat Recovery Networks. Computers and Chemical

Department of Process and Environmental Engineering of the University of Oulu,

Improving Process Flexibility in Linear Systems. Technical Report EDRC-06-24-87,

Knowledge Acquisition. Modeling and Management, pages 205–220, 1997. Sant

design: closing the manufacturing loop. In 2nd International Conference on Casebased Reasoning Research and Development, pages 114–121, London, UK, 1997.

17th International Workshop on Qualitative Reasoning, pages 147–153, 2003.

Second International Conference on AI in Design, pages 795–813, 1992. Kluwer

plant: Case study: Adding new units to an aromatics plant. Computers and

making and system management. IEEE Transactions on Systems, Man and

http://www.ozone-db.org/frames/documentation/overview.html, 2008.

(Ozone 08) Ozone. Ozone Developers' Guide.

Feliu de Guíxols, Catalonia.

LNCS Vol. 1266, Springer-Verlag.

Cybernetics, vol. SMC-15, no. 2, 1985.

PhD thesis, University of Twente, 1995.

PhD thesis, Stanford University, 1991.

Engineering, vol. 12, pages 445–456, 1998. Elsevier.

Chemical Engineering, vol. 18, no. 8, pages 743–753, 1994.

approach to cognitive engineering. North Holland, 1986.

eds, Lawrence Erlbaum, Hillsdale, N.J., pages 47–73, 1986.

Finland, 2001.

Brasilia, Brazil.

Academic Publishers.

Engineering, vol. 7, pages 707–721, 1983.


(Ozone 08) Ozone. Ozone Developers' Guide.

130 Advances in Chemical Engineering

(Kuraoka 03) K. Kuraoka & R. Batres. An Ontological Approach to Represent HAZOP

(Lander 97) S.E. Lander. Issues in Multi-agent Design Systems. IEEE Expert, vol. 12, no. 2,

(Leveson 00) N.G. Leveson. Intent specifications: an approach to building human-centered

(Lind 90) M. Lind. Representing Goals and Functions of Complex Systems. Technical Report

(Lind 94) M. Lind. Modelling goals and functions of complex industrial plants. Applied

(Lind 96) M. Lind. Status and challenges of intelligent plant control. Annual Reviews in

(Lind 99) M. Lind. Plant Modeling for Human Supervisory Control. Transactions of the Institute of Measurement and Control, vol. 21, no. 4/5, pages 171–180, 1999. (Linnho. 88) B. Linnho. G.T. Polley & V. Sahdev. General Process Improvements Through

(Maher 95) M.L. Maher B. Balachandran & D.M. Zhang. Case-based reasoning in design.

(Maher 97a) M.L. Maher & A. Gomez de Silva Garza. Case-Based Reasoning in Design. IEEE

(Maher 97b) M.L. Maher S. Simno. & J. Mitchell. Formalising building requirements using an activity/space model. Automation in construction, vol. 6, pages 77–95, 1997. (Maher 01) M.L. Maher & A. Gomez de Silva Garza. GENCAD: A Hybrid

(Maulik 92) P.C. Maulik M.J. Flynn D.J. Allstot & L.R. Carley. Rapid Redesign of Analog

(Mitchell 83) T.M. Mitchell L.I. Steinberg S. Kedar-Cabelli V.E. Kelly J. Shul-Man & T.

Conference on Artificial Intelligence (AAAI-83), Washington, USA, 1983. (Mostow 89) J. Mostow. Design by derivational analogy: Issues in the automated replay of design plans. Artificial Intelligence, vol. 40, no. 1-3, pages 119–184, 1989. (Nelson 90) D.A. Nelson & J.M. Douglas. A systematic procedure for retro.tting chemical

(Niles 01) I. Niles & A. Pease. Towards a Standard Upper Ontology. In 2nd International

(Ohsuga 97) S. Ohsuga. Strategic Knowledge Makes Knowledge Based Systems Truly

Analogi-cal/Evolutionary Model of Creative Design. Proc. of the 4th International Conference on Computational Models of Creative Design, 2001. J.S. Gero and M.L.

Standard Cells Using Constrained Optimization Techniques. Proc. IEEE Custom

Weinrich. An intelligent aid for circuit redesign. In Proceedings of the 3th National

plants to operate utilising different reaction paths. Ind. Eng. Chem. Res., vol. 29,

Conference on Formal Ontology in Information Systems (FOIS-2001), pages 2–9,

In-telligent. In Proceedings of the First International Workshop on Strategic Knowledge and Concept Formation, pages 1–24, Lutchi Research Centre, 1997.

Pinch Technology. Chem. Eng. Progr., vol. 84, pages 51–58, 1988.

Laboratory, Tokyo Institute of Technology, 2003.

Artificial Intelligence, vol. 8, pages 259–284, 1994.

Lawrence Erlbaum Associates, Sydney, 1995.

Integrated Circuits Conference, pages 8.1.1–81.3, 1992.

Expert, vol. 12, no. 2, pages 34–41, 1997.

Control, vol. 20, pages 23–41, 1996.

pages 18–26, 1997.

35, 2000.

1990.

Maher (eds).

pages 819–829, 1990.

2001.

Information. Technical Report TR-2003-01, Process Systems En-gineering

specifications. IEEE Transactions on Software Engineering, vol. 26, no. 1, pages 15–

90-D-38, Institute of Automatic Control Systems, Technical University of Denmark,

http://www.ozone-db.org/frames/documentation/overview.html, 2008.


Modelling Approach for Redesign of Technical Processes 133

(Sycara 92) K. Sycara. CADET: A case-based synthesis tool for engineering design. In-ternational Journal for Expert Systems, no. 2, pages 157–188, 1992. (Takeda 90a) H. Takeda P. Verkaamp T. Tomiyama & H. Yoshikawa. Modelling Design

(Takeda 94a) H. Takeda. Abduction for design. In Proceddings of Formal Design Methods

(Tay 2003) Tay FEH, Gu J. A methodology for evolutionary product design. Engineering

(Thornton 93) A.C. Thornton & A. Johnson. Constraint specification and satisfaction in

(Tjoe 86) T.N. Tjoe & B. Linnho. Using pinch technology for processes retrofit. Chem. Eng.,

(Tomiyama 87) T. Tomiyama & H. Yoshikawa. Extended General Design Theory. In

(Treur 89) J. Treur. A logical analysis of design tasks for expert systems. International

(Turton 98) R. Turton R.C. Bailie W.B. Whiting & J.A. Shaeiwitz. Analysis, synthesis and

(Uerdingen 01) E. Uerdingen U. Fischer & K. Hunderbuhler. A screening method for

(Ullman 91) D.G. Ullman. Design histories: archiving the evolution of products. In

(Umeda 90) Y. Umeda T. Tomiyama & H. Yoshikawa. Function, Behaviour and Structure.

(Umeda 92) Y. Umeda T. Tomiyama & H. Yoshikawa. A design methodology for a

(Umeda 94) Y. Umeda T. Tomiyama H. Yoshikawa & Y. Shomimura. Using Functional

(Umeda 97) Y. Umeda & T. Tomiyama. Functional Reasoning in Design. IEEE Expert, vol.

(Vaselenak 87) J.A. Vaselenak I.E. Grossman & A.W. Westerberg. Optimal Retro.t De-sign of

(Vescovi 93) M. Vescovi & Y. Iwasaki. Device design as functional and structural re.ne-ment. In Working Notes of the IJCAI'93 Workshop on AI in Design, pages 55–60, 1993. (Vicente 92) K.J. Vicente & J. Rasmussen. Ecological interface design: Theoretical foundations. IEEE Trans. on Systems, Man and Cybernetics, vol. 22, no. 4, 1992.

(Wetzel 00) B. Wetzel. Selection Engine. http://selectionengine.sourceforge.net/, 2000.

embodiment design. In Procedings of the International Conference on Engineering

Proceedings of the IFIP WG 5.2 Working Conference on Design Theory for CAD,

identifying economic improvement potentials in retrofit design. In ESCAPE-11 (European Symposium on Computer Aided Process Engineering), pages 573–578,

Proceedings of the DARPA Workshop on Manufacturing, Salt Lake City, USA,

Applications of Artificial Intelligence in Engineering V, pages 177–193, 1990. J.S.

self-maintenance machine based on funtional redundancy. Design Theory and

maintenance to improve fault tolereance. IEEE Expert, vol. 9, no. 3, pages 25–31,

Multiproduct Batch Plant. Industrial and Engineering Chemistry Research, vol. 26,

Processes. AI Magazine, vol. 11, no. 4, pages 37–48, 1990.

for CAD, pages 221–243, Tallinn, Estonia, 1994.

Design, ICED 93, pages 1319–1326, The Hague, 1993.

Journal of Expert Systems, vol. 2, pages 233–253, 1989.

Methodology -DTM'92, pages 317–324, 1992.

12, no. 2, pages 42–48, 1997.

no. 4, pages 718–726, 1987.

design of chemical processes. Prentice-Hall, New Jersey, 1998.

with Computers 19:160-173.

vol. 93, pages 47–60, 1986.

pages 95–125, 1987.

2001.

1991.

1994.

Gero (Ed).


(Simon 96) H.A. Simon. The sciences of the artificial. MIT Press, Cambridge, Massachusetts,

(Skuce 93) D. Skuce. A multi functional knowledge management system. Knowledge

(Smith 87) R. Smith & B. Linnho. Process integration using pinch technology. Proceedings of

(Smyth 96) B. Smyth & M.T. Keane. Using adaptation knowledge to retrieve and adapt

(Smyth 01) B. Smyth M.T. Keane & P. Cunningham. Hierarchical Case-Based Reasoning

(Sowa 95) J. Sowa. Distinctions, Combinations and Constraints. In Proceedings of IJCAI 95

(Steier 91) D. Steier. Automating Algorithm Design within a General Architecture for

(Steinberg 85) L. Steinberg & T. Mitchell. The Redesign System: A Knowledge-base

(Stephanopoulos 90a) G. Henning G. Stephanopoulos & H. Leone. MODEL.LA A modelling

(Stephanopoulos 90b) G. Stephanopoulos G. Henning & H. Leone. MODEL.LA A

Computers and Chemical Engineering, vol. 14, no. 8, pages 813–846, 1990. (Stroulia 92a) E. Stroulia M. Shankar A. Goel & L. Penberthy. A Model-Based Approach to

(Stroulia 92b) E. Stroulia & A. Goel. Generic Teleological Mechanisms and their use in Case

(Struss 91) P. Struss. A theory of model simplification and abstraction for diagnosis. In

(Subba-Rao 99) S. Subba-Rao A. Nahm Z. Shi X. Deng & A. Syamil. Artificial intelligence

(Sumi 97) Y. Sumi. Supporting the Acquisition and Modelling of Requirements in Software

and Concept Formation, pages 205–216, Lutchi Research Centre, 1997.

on AI in Design, pages 519–537, 1992. J.S. Gero (editor).

Science Society, pages 319–324, 1992. Indiana, USA.

(Sun 08) Sun Microsystems. Java Language Specification. http://java.sun.com/docs/index.html, 2008.

design cases. Knowledge-based Systems, Special Issue on Models and Techniques

Integrating Case-Based and Decompositional Problem-Solving Techniques for Plant-Control Software Design. IEEE Trans. on Knowledge and data engineering,

Workshop on Basic Ontological Issues in Knowledge Sharing, Montreal, Canada,

Intelligence. Automating Software Design, pages 577–602, 1991. Lowry, M.R. and

approach to VLSI CAD. IEEE Design and Test of Computers, vol. 2, no. 45-54, 1985.

framework language for process engineering II. Multifaced modelling of processing systems. Computers and Chemical Engineering, vol. 14, no. 8, pages 847–869, 1990.

mod-elling framework language for process engineering I. The formal framework.

Blame Assignment in Design. Proceedings of the Second International Conference

Adaptation. Proceedings of the Fourteenth Anual Conference of the Cognitive

Proceedings of the Fifth International Workshop on Qualitative Reasoning, pages

and expert systems applications in new product development-a survey. Journal of Intelligent Manufacturing, vol. 10, no. 3-4, pages 231–244, 199. Kluwer Academic

Design. In Proceedings of the First International Workshop on Strategic Knowledge

ATEE Symp. Energy Management in Industry, 1987. Paris.

for Reuse of Designs, vol. 9, no. 2, pages 127–136, 1996.

Acquisition, vol. 5, pages 305–346, 1993.

vol. 13, no. 5, pages 793–812, 2001.

Mccartney, R.D. (Eds.).

25– 57, 1991.

Publishers.

1996.

1995.


**5** 

**Application Potential of** 

*1TI Food and Nutrition, AN Wageningen* 

*University of Twente, Enschede* 

*The Netherlands* 

**Food Protein Modification** 

Harmen H.J. de Jongh1 and Kerensa Broersen2,\*

*MIRA Institute for Biomedical Technology and Technical Medicine,* 

Proteins are essential in foods, not only for their nutritional value, but also as modulator of structure and perception of a food product. The functional behavior of a protein is inherently susceptible to physico-chemical conditions as pH, ionic strength, temperature, or pressure, making them also an unpredictable, and at the same time, opportune component in food production. Proteins are generally also industrially costly, and with increasing world population and welfare the pressure on protein-availability for food purposes gives rise to some concerns. In view of a more sustainable use of protein-sources a number of routes have been followed in the past decades that provided big steps forward in protein availability: (i) more efficient production or protein refinery methods, (ii) use of alternative protein sources, and (iii) optimized usage of protein functionality. Especially in wheat production correlations between genetic expression and functional product behavior allowed breeders to optimize cultivars for geographic location (e.g. Payne et al., 1984). Alternatively, one has the ability to express specific proteins in non-original sources, for example human milk proteins in plants, such as rice (e.g. Lönnerdal, 2002). Directed alterations in the genome of food-producing organisms can lead to changes in the primary sequences of relevant proteins and thereby introduce potentially new functionality. If sufficient quantities of the novel protein are synthesized and become admixed with the basal levels of protein in the food, the functional properties of the food system (textureformation) may become improved. Alternatively, the modified protein can be isolated for use as food ingredient. More recently, a number of proteins from less-conventional origin have been identified as human food ingredients that one has started to exploit, e.g. algae, leafs, insects, and various seeds. Successful utilization of these new proteinaceous materials has thus far been rather limited, requiring breakthroughs in extractability, their digestibility, nutritive value, and overall functional and organoleptic properties. More downstream in the process is the modulation of protein functional behavior at an ingredient level. This can be

physical-chemically, enzymatically, or via chemical engineering.

**1. Introduction** 

 \*

Corresponding Author

*2Faculty of Science and Technology, Nanobiophysics,* 


### **Application Potential of Food Protein Modification**

Harmen H.J. de Jongh1 and Kerensa Broersen2,\*

*1TI Food and Nutrition, AN Wageningen 2Faculty of Science and Technology, Nanobiophysics, MIRA Institute for Biomedical Technology and Technical Medicine, University of Twente, Enschede The Netherlands* 

#### **1. Introduction**

134 Advances in Chemical Engineering

(Wielinga 97) B.J. Wielinga & A.Th. Schreiber. Con.guration-design problem solving. IEEE

(Wood 01) M. Wood & S.A. DeLoach. An Overview of the Multiagent Systems Engi-neering

(Xin 01) W. Xin & X. Guangleng. Supporting design reuse based on integrated design

(Yoshikawa 91) H. Yoshikawa F. Arbab & T. Tomiyama. Intelligent CAD III: Selected and

Methodology. Agent-Oriented Software Engineering, LNAI, vol. 1957, 2001. P.

rationale. IEEE International Conference on Systems, Man and Cybernetics, vol. 3,

Reviewed Papers and Reports. In Third International Workshop on Computer

Expert, vol. 12, no. 2, pages 49–56, 1997.

Ciancarini, M. Wooldridge, (Eds.).

pages 1909–1912, 2001.

Aided Design, 1991.

Proteins are essential in foods, not only for their nutritional value, but also as modulator of structure and perception of a food product. The functional behavior of a protein is inherently susceptible to physico-chemical conditions as pH, ionic strength, temperature, or pressure, making them also an unpredictable, and at the same time, opportune component in food production. Proteins are generally also industrially costly, and with increasing world population and welfare the pressure on protein-availability for food purposes gives rise to some concerns. In view of a more sustainable use of protein-sources a number of routes have been followed in the past decades that provided big steps forward in protein availability: (i) more efficient production or protein refinery methods, (ii) use of alternative protein sources, and (iii) optimized usage of protein functionality. Especially in wheat production correlations between genetic expression and functional product behavior allowed breeders to optimize cultivars for geographic location (e.g. Payne et al., 1984). Alternatively, one has the ability to express specific proteins in non-original sources, for example human milk proteins in plants, such as rice (e.g. Lönnerdal, 2002). Directed alterations in the genome of food-producing organisms can lead to changes in the primary sequences of relevant proteins and thereby introduce potentially new functionality. If sufficient quantities of the novel protein are synthesized and become admixed with the basal levels of protein in the food, the functional properties of the food system (textureformation) may become improved. Alternatively, the modified protein can be isolated for use as food ingredient. More recently, a number of proteins from less-conventional origin have been identified as human food ingredients that one has started to exploit, e.g. algae, leafs, insects, and various seeds. Successful utilization of these new proteinaceous materials has thus far been rather limited, requiring breakthroughs in extractability, their digestibility, nutritive value, and overall functional and organoleptic properties. More downstream in the process is the modulation of protein functional behavior at an ingredient level. This can be physical-chemically, enzymatically, or via chemical engineering.

<sup>\*</sup> Corresponding Author

Application Potential of Food Protein Modification 137

Jongh & Wierenga, 2006; Graham & Phillips, 1979a, 1979b). Some publications report that the rearrangement process coincides with structural extension of the protein molecules (reviewed in MacRitchie, 1978) while other results suggest that secondary structure does not undergo variation subsequent to adsorption onto the interface (de Jongh & Wierenga, 2006; Graham & Phillips, 1979c). Despite the finding that local concentrations of 150-300 mg/ml can be reached at the interface (Meinders et al., 2001), proteins are still able to laterally diffuse as was shown by thiolated ovalbumin (de Jongh & Wierenga, 2006), illustrating that chemical modification has been used to underpin the molecular mechanisms of the surface

Emulsions consist of two immiscible liquids, oil and water, in which the droplets are termed dispersed phase and the liquid surrounding the droplets is called continuous phase. Depending on the concentrations of each liquid and the environmental conditions, oil-inwater emulsions or water-in-oil emulsions can be formed. These consist of oil droplets in a continuous water phase and water droplets in a continuous oil phase, respectively. Examples of food related emulsions are milk, vinaigrette, and mayonnaise. Emulsions are often unstable by nature and will phase separate or coalesce over extended time, or storage. To stabilize emulsions, so-called emulsifiers can be used which form a small layer on the surface of the dispersed phase, thereby physically separating the continuous phase from the dispersed phase. Such treatment will allow the incorporation of emulsified solutions in food products which can be stored over an extended time frame. Amphiphilic proteins, containing both hydrophilic and hydrophobic regions, are effective emulsifiers. These proteins adsorb onto the interface between the oil and water phase, and stabilize the oil and water phase by selective interaction with both surfaces, thereby preventing the individual droplets from coalescing (reviewed in Rodríguez Patino et al., 2008). Other types of emulsifiers used in food products include lipids, phospholipids, surfactants or polysaccharides (Dickinson, 1992; McClements, 2005). Apart from a texture perspective, the absorption of proteins at the oil-water interface is of interest to the delivery of nutrients (reviewed in Malaki Nik et al., 2010). Upon digestion, changes occur at the oil-water interface of emulsified food products as a function of emulsifier stability, affecting the digestibility and subsequently the availability of encapsulated nutrients. The amphiphilic character of proteins can be modulated for example by the covalent attachment of lipid chains (see paragraph 3.2). Attachment of lipid chains of various lengths to a protein renders it more hydrophobic resulting in an improved ability to stabilize emulsions and/or foams (Aewsiri et al., 2011a, 2011b). Apart from exposed hydrophobicity, other factors have also been identified to determine the affinity of a protein for an oil-water interface. Aggregation or molecular weight of proteins for example adversely affects the emulsifying activity of proteins (Baldursdottir et al., 2010; Corzo-Martínez, 2011). That the aggregation effect is more complex than originally postulated was shown by single molecule total internal reflectance fluorescence microscopy of bovine serum albumin showing that the orientation at which the aggregates absorb to the interface largely determine the rate of absorbance to the interface (Walder & Schwartz, 2010). The identification of factors contributing to the surface activity of proteins allows the improvement of emulsifying activity by means of

activity of proteins.

**2.1.1 Emulsions** 

targeted chemical modification.

This review will focus on the use of chemical engineering to study or better exploit protein functionality in food products. Reasons to employ chemical modification of proteins will be discussed in the context of its relevance in understanding the fundamental principles of proteins as structuring agents in food. These include improving shelf life and sensory properties as well as the development of new functionalities of food proteins, such as the application of plant proteins as meat-texturizers. Further we discuss how these insights have contributed thus far to a more sustainable utilization of protein, including aspects as consumer acceptance and existing/changing legislations for the use as ingredients.

#### **2. Functional role of proteins as food ingredient**

This paragraph will briefly summarize the functional properties of proteins as food ingredients. The molecular mechanisms of their roles in food products will be highlighted. In the next paragraph these molecular mechanisms will be discussed as target for chemical engineering. On each of these topics extensive reviews have been published which will be pointed out.

#### **2.1 Surface properties: Emulsions and foams**

Surfaces or interfaces in food products are abundantly present in terms of emulsions (oilwater interface) or foams (air-water interface). Examples of foams and emulsions in food products include ice cream, dressings or margarine. In foams and emulsions respectively air cells or oil droplets are dispersed or in an aqueous phase; the latter case also water in oil emulsions exist. As a result of their amphiphilic character, proteins, being composed of polar and non-polar amino acids, can contribute to the formation and stability of such dispersions by adsorbing to the interface and developing stabilizing films by coating the oil or air droplets and hence act as emulsifiers. The ability of proteins to induce film formation depends on a number of parameters of molecular nature which have been studied into detail in the past and is collectively governed by a net energy gain from absorbing at an interface. Milk proteins have been identified as good foaming agents as a result of their aggregation state, molecular stability and flexibility, electrostatics, and (surface) exposed hydrophobicity (Hunter et al., 1991; Luey et al., 1991; Shirahama et al., 1990; Suttiprasit et al., 1992; Waniska & Kinsella, 1985). Chemical modifications have been employed in the past to improve surface activity of less performing proteins. These studies and consequences of the used modifications will be discussed in more detail in the following paragraph. A wide range of methods has been employed in literature to study the chemical and molecular properties of proteins adsorbed at interfaces which lead to a detailed understanding of the principle forces of importance to surface activity. These methods include ellipsometry and infrared reflection absorption spectroscopy (IRRAS) which provide information on denaturation and concentration of adsorbed proteins (de Jongh & Wierenga, 2006; Grigoriev et al., 2007; Martin et al., 2003; McClellan et al., 2003). Stabilization of an air-water or oilwater interface is governed by a multiple step process. First, absorption at the interface requires proteins to diffuse to the interface and their retention at the interface is governed by the kinetic barrier of absorbance which, in turn, is influenced by factors such as exposed hydrophobicity (Wierenga et al., 2003), and net charge (Wierenga et al., 2005). Effective absorption onto the interface is followed by rearrangement of the protein molecules to form a thermodynamically stable but dynamic monolayer of molecules coating the droplets (de Jongh & Wierenga, 2006; Graham & Phillips, 1979a, 1979b). Some publications report that the rearrangement process coincides with structural extension of the protein molecules (reviewed in MacRitchie, 1978) while other results suggest that secondary structure does not undergo variation subsequent to adsorption onto the interface (de Jongh & Wierenga, 2006; Graham & Phillips, 1979c). Despite the finding that local concentrations of 150-300 mg/ml can be reached at the interface (Meinders et al., 2001), proteins are still able to laterally diffuse as was shown by thiolated ovalbumin (de Jongh & Wierenga, 2006), illustrating that chemical modification has been used to underpin the molecular mechanisms of the surface activity of proteins.

#### **2.1.1 Emulsions**

136 Advances in Chemical Engineering

This review will focus on the use of chemical engineering to study or better exploit protein functionality in food products. Reasons to employ chemical modification of proteins will be discussed in the context of its relevance in understanding the fundamental principles of proteins as structuring agents in food. These include improving shelf life and sensory properties as well as the development of new functionalities of food proteins, such as the application of plant proteins as meat-texturizers. Further we discuss how these insights have contributed thus far to a more sustainable utilization of protein, including aspects as

This paragraph will briefly summarize the functional properties of proteins as food ingredients. The molecular mechanisms of their roles in food products will be highlighted. In the next paragraph these molecular mechanisms will be discussed as target for chemical engineering. On each of these topics extensive reviews have been published which will be

Surfaces or interfaces in food products are abundantly present in terms of emulsions (oilwater interface) or foams (air-water interface). Examples of foams and emulsions in food products include ice cream, dressings or margarine. In foams and emulsions respectively air cells or oil droplets are dispersed or in an aqueous phase; the latter case also water in oil emulsions exist. As a result of their amphiphilic character, proteins, being composed of polar and non-polar amino acids, can contribute to the formation and stability of such dispersions by adsorbing to the interface and developing stabilizing films by coating the oil or air droplets and hence act as emulsifiers. The ability of proteins to induce film formation depends on a number of parameters of molecular nature which have been studied into detail in the past and is collectively governed by a net energy gain from absorbing at an interface. Milk proteins have been identified as good foaming agents as a result of their aggregation state, molecular stability and flexibility, electrostatics, and (surface) exposed hydrophobicity (Hunter et al., 1991; Luey et al., 1991; Shirahama et al., 1990; Suttiprasit et al., 1992; Waniska & Kinsella, 1985). Chemical modifications have been employed in the past to improve surface activity of less performing proteins. These studies and consequences of the used modifications will be discussed in more detail in the following paragraph. A wide range of methods has been employed in literature to study the chemical and molecular properties of proteins adsorbed at interfaces which lead to a detailed understanding of the principle forces of importance to surface activity. These methods include ellipsometry and infrared reflection absorption spectroscopy (IRRAS) which provide information on denaturation and concentration of adsorbed proteins (de Jongh & Wierenga, 2006; Grigoriev et al., 2007; Martin et al., 2003; McClellan et al., 2003). Stabilization of an air-water or oilwater interface is governed by a multiple step process. First, absorption at the interface requires proteins to diffuse to the interface and their retention at the interface is governed by the kinetic barrier of absorbance which, in turn, is influenced by factors such as exposed hydrophobicity (Wierenga et al., 2003), and net charge (Wierenga et al., 2005). Effective absorption onto the interface is followed by rearrangement of the protein molecules to form a thermodynamically stable but dynamic monolayer of molecules coating the droplets (de

consumer acceptance and existing/changing legislations for the use as ingredients.

**2. Functional role of proteins as food ingredient** 

**2.1 Surface properties: Emulsions and foams** 

pointed out.

Emulsions consist of two immiscible liquids, oil and water, in which the droplets are termed dispersed phase and the liquid surrounding the droplets is called continuous phase. Depending on the concentrations of each liquid and the environmental conditions, oil-inwater emulsions or water-in-oil emulsions can be formed. These consist of oil droplets in a continuous water phase and water droplets in a continuous oil phase, respectively. Examples of food related emulsions are milk, vinaigrette, and mayonnaise. Emulsions are often unstable by nature and will phase separate or coalesce over extended time, or storage. To stabilize emulsions, so-called emulsifiers can be used which form a small layer on the surface of the dispersed phase, thereby physically separating the continuous phase from the dispersed phase. Such treatment will allow the incorporation of emulsified solutions in food products which can be stored over an extended time frame. Amphiphilic proteins, containing both hydrophilic and hydrophobic regions, are effective emulsifiers. These proteins adsorb onto the interface between the oil and water phase, and stabilize the oil and water phase by selective interaction with both surfaces, thereby preventing the individual droplets from coalescing (reviewed in Rodríguez Patino et al., 2008). Other types of emulsifiers used in food products include lipids, phospholipids, surfactants or polysaccharides (Dickinson, 1992; McClements, 2005). Apart from a texture perspective, the absorption of proteins at the oil-water interface is of interest to the delivery of nutrients (reviewed in Malaki Nik et al., 2010). Upon digestion, changes occur at the oil-water interface of emulsified food products as a function of emulsifier stability, affecting the digestibility and subsequently the availability of encapsulated nutrients. The amphiphilic character of proteins can be modulated for example by the covalent attachment of lipid chains (see paragraph 3.2). Attachment of lipid chains of various lengths to a protein renders it more hydrophobic resulting in an improved ability to stabilize emulsions and/or foams (Aewsiri et al., 2011a, 2011b). Apart from exposed hydrophobicity, other factors have also been identified to determine the affinity of a protein for an oil-water interface. Aggregation or molecular weight of proteins for example adversely affects the emulsifying activity of proteins (Baldursdottir et al., 2010; Corzo-Martínez, 2011). That the aggregation effect is more complex than originally postulated was shown by single molecule total internal reflectance fluorescence microscopy of bovine serum albumin showing that the orientation at which the aggregates absorb to the interface largely determine the rate of absorbance to the interface (Walder & Schwartz, 2010). The identification of factors contributing to the surface activity of proteins allows the improvement of emulsifying activity by means of targeted chemical modification.

Application Potential of Food Protein Modification 139

These observations lead to the understanding that net charge plays a major role in determining aggregate and gel morphology (Krebs et al., 2009; Langton & Hermansson, 1992). Cryo-EM investigation of aggregates formed from ovalbumin which had been succinylated to various degrees provided further evidence that net charge dominantly determines aggregate morphology (Weijers et al., 2008). The propensity of proteins to aggregate, or rate of aggregation, has been shown to vary as a function of protein conformational stability (Chiti et al., 2000; Hurle et al., 1994; Kelly, 1998; Quintas et al., 1997; Ramirez-Alvarado et al., 2000; Siepen & Westhead, 2002), rate of unfolding (Broersen et al., 2007a), net charge (Calamai et al., 2003; DuBay et al., 2004), and secondary structure propensity (Fernandez-Escamilla et al., 2004). Exposed hydrophobicity (Calamai et al., 2003) and the possibility to form disulfide bonds naturally affect the aggregation and gelation propensity of proteins as these two forces are primarily driving the assembly process. This has been exemplified in a study which showed that -lactoglobulin A modified with *N*ethylmaleimide largely resisted aggregation induced by heating (Kitabatake et al., 2001). Extensive knowledge of the molecular factors driving the aggregation process of proteins has lead to the development of a number of algorithms able to predict protein aggregation with high fidelity (Chiti et al., 2003; DuBay et al., 2004; Fernandez-Escamilla et al., 2004;

The energetics and kinetics of protein aggregation have been subject of many publications to date and has been reviewed on numerous occasions (e.g. Luheshi & Dobson, 2009; Straub & Thirumalai, 2011). It has been recognized that, regardless of primary sequence or physicochemical properties, all proteins have an inherent tendency to form aggregates *in vitro* under certain conditions (reviewed in Chiti & Dobson, 2006; Dobson, 1999). Protein aggregation is a multiple step complex process which can be viewed as a cascade of steps of assembly which may vary in molecular detail as a function of the protein studied or the environmental conditions. Nevertheless, the aggregate growth mechanisms of many different proteins share essential characteristics which have been elucidated both by experimental and computational methods (Teplow et al., 2006). The onset of aggregation often requires the (partial) unfolding or conformational rearrangement of proteins (Calamai et al., 2003; Dobson, 1999; Kelly, 1996; Rochet & Lansbury, 2000). Using human lysozyme (Canet et al., 1999) and HypF-N (Marcon et al., 2005) it has been shown that a population of less than 1% of partially folded protein can be sufficient to trigger the onset of the aggregation process. The resulting exposure of hydrophobic regions which normally reside in the core of a folded protein drives the self-assembly process to form small oligomers. These oligomers, or nuclei, are metastable and their transient and short-lived nature dictates that they can dissociate into monomeric protein, which has been shown for various proteins. Many efforts in the field of protein aggregation suggest that the nucleus has to reach a critical size which then allows further assembly by monomer addition to ultimately form mature fibrils or aggregated networks (Jarrett & Lansbury, 1993; Lomakin et al., 1996, 1997; Sorci et al., 2011). An alternative scenario is the formation of intermediate protofibrils along the pathway which subsequently assemble into fibrils (Harper et al., 1997; Walsh et al., 1997). Mature fibrils have classically been viewed as the stable end-stage of the aggregation process which are not susceptible for dissociating conditions. Recently, using calorimetric methods (Morel et al., 2010), molecular dynamics simulations (Zidar & Merzel, 2011), and mechanical deformation studies (Paparcone & Buehler, 2011; Paparcone et al., 2010; Xu et al., 2010) it has been observed that fibrils can be dissociated albeit at high temperature.

Maurer-Stroh et al., 2010).

#### **2.1.2 Foams**

Foams consist of gas bubbles dispersed into a liquid. The stability of the air bubbles in a foam is determined by the foaming agent which forms a layer of adsorbed molecules separating the air bubbles from the continuous liquid phase, similar to the emulsifying activity described in paragraph 2.1.1 (reviewed in Halling, 1981; Wilde, 2000). Adsorption of a protein to the air-water interface induces partial dehydration of the molecule promoting protein-protein interactions. This effect is further amplified by the finding that local protein concentrations at the air-water interface can reach up to 150 to 300 mg/ml (Meinders et al., 2001). The rate of absorption to the air-water interface has been reported to depend largely on the hydrophobic nature of the protein under investigation (Kudryashova et al., 2003; Wierenga et al., 2003). Increasing the exposed hydrophobicity of proteins by means of conjugation with lipid chains was shown to increase the adsorption rate to the air-water interface (Wierenga et al., 2003). Net charge is a second parameter of interest determining adsorption kinetics, in which higher net charge slows down the adsorption process due to the electric repulsive forces involved (de Jongh et al., 2004; Kudryashova et al., 2005; Le Floch-Fouéré et al., 2011). Highly aggregated heat-treated ovalbumin was further shown to induce a ten-fold decrease the diffusion rate of proteins to the interface compared to the native protein (Kudryashova et al., 2005). However, the tendency of aggregated ovalbumin to remain adhered to the interface upon first interaction is significantly larger than for nonaggregated protein, which was found to rapidly desorb from the surface after absorption (Kudryashova et al., 2005). Collectively, it has been shown that surface activity of proteins is not determined by a single molecular characteristic but rather depends on a combination of factors. Hence, several types of chemical modification can be employed to improve the airwater interface activity of proteins.

#### **2.2 Aggregation and gelation**

Protein aggregation is a major topic in the field of food science, the regulation of which is believed to markedly affect the texture of food products (Zhou et al., 2008; reviewed in Doi, 1993). Aggregated protein can act as a nucleation prerequisite to induce gelation, albeit at high protein concentration (Alting et al., 2003; Barbut & Foegeding, 1993; Ju & Kilara, 1998). Processing conditions and storage can induce protein aggregation, even at ambient temperatures (Promeyrat et al., 2010; Santé-Lhoutellier et al., 2008). The resulting protein aggregates can vary widely in size and morphology as a result of the environmental conditions under which they were formed, among other factors. For example, upon inspection using electron microscopy, amylin (Patil et al., 2011), hen egg white lysozyme (Arnaudov & de Vries, 2005), and -lactoglobulin (Arnaudov et al., 2003; Veerman et al. 2002) were found to form negatively stained and long fibrillar aggregates at pH values far away from the isoelectric point of the respective proteins. Near the isoelectric pH (Arnaudov & de Vries, 2005), or at high salt concentration (Arnaudov & de Vries, 2006; Veerman et al. 2002), spherical or amorphous aggregates are formed. That aggregate morphologies cannot always be categorized as either fibrillar or amorphous in a clear-cut manner was shown by various groups observing substantial heterogeneity in aggregate morphology, also called polymorphism, within the same preparation (Bauer et al., 1995; Jiménez et al., 2002, 1999). As a consequence, gels formed under these conditions of -lactoglobulin are particulate and the particle size depends on heating temperature and heating rate (Bromley et al., 2006).

Foams consist of gas bubbles dispersed into a liquid. The stability of the air bubbles in a foam is determined by the foaming agent which forms a layer of adsorbed molecules separating the air bubbles from the continuous liquid phase, similar to the emulsifying activity described in paragraph 2.1.1 (reviewed in Halling, 1981; Wilde, 2000). Adsorption of a protein to the air-water interface induces partial dehydration of the molecule promoting protein-protein interactions. This effect is further amplified by the finding that local protein concentrations at the air-water interface can reach up to 150 to 300 mg/ml (Meinders et al., 2001). The rate of absorption to the air-water interface has been reported to depend largely on the hydrophobic nature of the protein under investigation (Kudryashova et al., 2003; Wierenga et al., 2003). Increasing the exposed hydrophobicity of proteins by means of conjugation with lipid chains was shown to increase the adsorption rate to the air-water interface (Wierenga et al., 2003). Net charge is a second parameter of interest determining adsorption kinetics, in which higher net charge slows down the adsorption process due to the electric repulsive forces involved (de Jongh et al., 2004; Kudryashova et al., 2005; Le Floch-Fouéré et al., 2011). Highly aggregated heat-treated ovalbumin was further shown to induce a ten-fold decrease the diffusion rate of proteins to the interface compared to the native protein (Kudryashova et al., 2005). However, the tendency of aggregated ovalbumin to remain adhered to the interface upon first interaction is significantly larger than for nonaggregated protein, which was found to rapidly desorb from the surface after absorption (Kudryashova et al., 2005). Collectively, it has been shown that surface activity of proteins is not determined by a single molecular characteristic but rather depends on a combination of factors. Hence, several types of chemical modification can be employed to improve the air-

Protein aggregation is a major topic in the field of food science, the regulation of which is believed to markedly affect the texture of food products (Zhou et al., 2008; reviewed in Doi, 1993). Aggregated protein can act as a nucleation prerequisite to induce gelation, albeit at high protein concentration (Alting et al., 2003; Barbut & Foegeding, 1993; Ju & Kilara, 1998). Processing conditions and storage can induce protein aggregation, even at ambient temperatures (Promeyrat et al., 2010; Santé-Lhoutellier et al., 2008). The resulting protein aggregates can vary widely in size and morphology as a result of the environmental conditions under which they were formed, among other factors. For example, upon inspection using electron microscopy, amylin (Patil et al., 2011), hen egg white lysozyme (Arnaudov & de Vries, 2005), and -lactoglobulin (Arnaudov et al., 2003; Veerman et al. 2002) were found to form negatively stained and long fibrillar aggregates at pH values far away from the isoelectric point of the respective proteins. Near the isoelectric pH (Arnaudov & de Vries, 2005), or at high salt concentration (Arnaudov & de Vries, 2006; Veerman et al. 2002), spherical or amorphous aggregates are formed. That aggregate morphologies cannot always be categorized as either fibrillar or amorphous in a clear-cut manner was shown by various groups observing substantial heterogeneity in aggregate morphology, also called polymorphism, within the same preparation (Bauer et al., 1995; Jiménez et al., 2002, 1999). As a consequence, gels formed under these conditions of -lactoglobulin are particulate and the particle size depends on heating temperature and heating rate (Bromley et al., 2006).

**2.1.2 Foams** 

water interface activity of proteins.

**2.2 Aggregation and gelation** 

These observations lead to the understanding that net charge plays a major role in determining aggregate and gel morphology (Krebs et al., 2009; Langton & Hermansson, 1992). Cryo-EM investigation of aggregates formed from ovalbumin which had been succinylated to various degrees provided further evidence that net charge dominantly determines aggregate morphology (Weijers et al., 2008). The propensity of proteins to aggregate, or rate of aggregation, has been shown to vary as a function of protein conformational stability (Chiti et al., 2000; Hurle et al., 1994; Kelly, 1998; Quintas et al., 1997; Ramirez-Alvarado et al., 2000; Siepen & Westhead, 2002), rate of unfolding (Broersen et al., 2007a), net charge (Calamai et al., 2003; DuBay et al., 2004), and secondary structure propensity (Fernandez-Escamilla et al., 2004). Exposed hydrophobicity (Calamai et al., 2003) and the possibility to form disulfide bonds naturally affect the aggregation and gelation propensity of proteins as these two forces are primarily driving the assembly process. This has been exemplified in a study which showed that -lactoglobulin A modified with *N*ethylmaleimide largely resisted aggregation induced by heating (Kitabatake et al., 2001). Extensive knowledge of the molecular factors driving the aggregation process of proteins has lead to the development of a number of algorithms able to predict protein aggregation with high fidelity (Chiti et al., 2003; DuBay et al., 2004; Fernandez-Escamilla et al., 2004; Maurer-Stroh et al., 2010).

The energetics and kinetics of protein aggregation have been subject of many publications to date and has been reviewed on numerous occasions (e.g. Luheshi & Dobson, 2009; Straub & Thirumalai, 2011). It has been recognized that, regardless of primary sequence or physicochemical properties, all proteins have an inherent tendency to form aggregates *in vitro* under certain conditions (reviewed in Chiti & Dobson, 2006; Dobson, 1999). Protein aggregation is a multiple step complex process which can be viewed as a cascade of steps of assembly which may vary in molecular detail as a function of the protein studied or the environmental conditions. Nevertheless, the aggregate growth mechanisms of many different proteins share essential characteristics which have been elucidated both by experimental and computational methods (Teplow et al., 2006). The onset of aggregation often requires the (partial) unfolding or conformational rearrangement of proteins (Calamai et al., 2003; Dobson, 1999; Kelly, 1996; Rochet & Lansbury, 2000). Using human lysozyme (Canet et al., 1999) and HypF-N (Marcon et al., 2005) it has been shown that a population of less than 1% of partially folded protein can be sufficient to trigger the onset of the aggregation process. The resulting exposure of hydrophobic regions which normally reside in the core of a folded protein drives the self-assembly process to form small oligomers. These oligomers, or nuclei, are metastable and their transient and short-lived nature dictates that they can dissociate into monomeric protein, which has been shown for various proteins. Many efforts in the field of protein aggregation suggest that the nucleus has to reach a critical size which then allows further assembly by monomer addition to ultimately form mature fibrils or aggregated networks (Jarrett & Lansbury, 1993; Lomakin et al., 1996, 1997; Sorci et al., 2011). An alternative scenario is the formation of intermediate protofibrils along the pathway which subsequently assemble into fibrils (Harper et al., 1997; Walsh et al., 1997). Mature fibrils have classically been viewed as the stable end-stage of the aggregation process which are not susceptible for dissociating conditions. Recently, using calorimetric methods (Morel et al., 2010), molecular dynamics simulations (Zidar & Merzel, 2011), and mechanical deformation studies (Paparcone & Buehler, 2011; Paparcone et al., 2010; Xu et al., 2010) it has been observed that fibrils can be dissociated albeit at high temperature.

Application Potential of Food Protein Modification 141

and unfolded protein molecules at a certain time. The rate at which the unfolded protein collapses to a folded state is reflected by the folding rate and represents the kinetic stability of a protein. Energy landscape theories and the folding funnel hypothesis have both been used as models to understand the energetic barriers of a protein separating the folded from the unfolded state (reviewed in Onuchic et al., 1997; Plotkin & Onuchic, 2002; Wolynes, 2005). Both models start from a similar principle in which there is an energy difference between the folded state of a protein and its unfolded state and that the folded state is defined as the favored entropic state of a protein. Local energy minima in the folding process can result in the accumulation of transient intermediate structures, which are neither folded, nor unfolded, to a larger or lesser extent, depending on the environmental conditions of folding or the primary sequence of the protein (reviewed in Baldwin, 2008; Englander et al., 2007). These intermediate structures, which often lack biological activity, are sometimes sufficiently stable to allow substantial accumulation (reviewed in Englander et al., 2007). Stable folding intermediates are related to a high propensity of aggregation as the hydrophobic core is not sufficiently shielded while opposing assembly forces are absent (reviewed in Ferreira et al., 2006). Kinetic protein stability is defined by a variety of molecular parameters including the proximity of native contacts in the primary sequence of a protein (Cieplak et al., 2004; Plaxco et al., 1998, 2000), internal friction or the energetic of intrachain interactions, energy barriers to backbone rotations and long-range residue interactions (Pabit et al., 2004; Qiu & Hagen, 2004), rate of diffusional motion of an unfolded peptide chain through the solvent (Pabit et al., 2004), and the presence or absence of intermediate state(s) (Baumketner, 2003; Onuchic et al., 1997). Forces that contribute to thermodynamic stability are the strength of intramolecular hydrogen bonds and solvent-protein interactions, both enthalpic in nature. The entropic contribution is mainly defined by the hydrophobic effect of folding through an increase in disorder of water molecules upon folding. Many of the forces retaining a protein structure intact can be disrupted, removed or introduced by chemical modification. For example, reaction of proteins with sugars by means of the Maillard reaction can lead to distinct changes in protein stability. The stability of proteins in food products can be affected by using proteins which have been modified by means of glycosylation or charge modification. Succinylation with the aim to increase net charge of soy protein hydrolysate lead to improved digestibility of the protein as investigated by a multienzyme method involving trypsin, chymotrypsin, and peptidase (Achouri & Zhang, 2001). This finding suggests that the protein had undergone structural rearrangement as a result of the succinylation process. Succinylation also lead to destabilization of Faba bean legumin (Schwenke et al., 1998). This group used differential scanning calorimetry (DSC) to study protein stability of legumin and found a decreased specific enthalpy for unfolding upon succinylation. Succinylation also resulted in an increased surface hydrophobicity of the protein suggesting at least partial unfolding of the molecule. Kosters and colleagues (2003) compared the effects of many different types of chemical modification on protein stability, including lipophilization using capric acid, glycosylation and succinylation of ovalbumin. Ovalbumin stability was probed by DSC, tryptophan fluorescence and circular dichroism (CD). Lipophilization resulted in a decreased denaturation temperature of ovalbumin reflected in an enthalpy decrease and a lower stability upon guanidine titration. Glycosylation was found to stabilize structural integrity of ovalbumin. A similar finding using CD and DSC to study temperature-induced unfolding has been reported upon glucosylation (Broersen et al., 2004; van Teeffelen et al., 2005) and fructosylation (Broersen et al., 2004) of lactoglobulin or glucosylation of codfish parvalbumin (de Jongh et al., 2011). Interestingly, assays performed at ambient temperature but involving denaturant-induced unfolding

Detailed knowledge of the molecular parameters determining aggregation propensity, rate and morphology resulted in the ability to tune protein aggregation through chemical engineering. The attachment of sugar chains to proteins has been shown to inhibit selfassociation (Marquardt & Helenius, 1992; reviewed in Helenius et al., 1997; Land & Braakman, 2001; Song et al., 2001). This effect was largely attributed to covalently linked sugar moieties affecting kinetic partitioning between folding and aggregation from an (partially) unfolded state. For example, glycosylation was found to increase the folding rate of the protein rapidly shielding exposed hydrophobic regions which could potentially act as a driving force for aggregation at ambient temperature (Broersen et al., 2007b; Shental-Bechor & Levy, 2008; Wang et al., 1996). Interestingly, high temperatures induced more rapid aggregation of glycosylated proteins compared to their non-glycosylated counterparts (Broersen et al., 2007b; Chobert et al., 2006). However, another study involving glycated bovine serum albumin concluded that glycation of the protein inhibited its aggregation upon incubation at moderate temperatures (Rondeau et al., 2010, 2007). Glycosylation of proteins has also been shown to affect gel properties: attachment of a ketohexose to ovalbumin by means of the Maillard reaction resulted in the formation of gels with enhanced breaking strength (Sun et al., 2004). The effects of various types of chemical protein engineering on the physico-chemical functionality of proteins will be discussed in more detail in paragraph 3 of this review.

#### **2.3 Protein structural integrity**

Proteins in food products can lose their native structure as a result of processing conditions including storage, heat treatment, acidification, dehydration, mechanical processing or shear, and microbial hydrolysis. For example, long-term storage of milk powder has been found to induce lactosylation of the proteins present in the preparation which, in turn, results in affected powder solubility and emulsifying and foaming properties (reviewed in Thomas et al., 2004). Unfolded or hydrolyzed protein molecules can exert very different functionality to food products compared to folded proteins, a classical example being the boiling of an egg which converts the liquid-like transparent egg white into an opaque semisolid structure with very different textural properties. This paragraph will shortly discuss the principles of protein folding and structure and the forces that are implied. It was first recognized by Anfinsen (1973) that the primary sequence of a protein dictates the specific folded, or native, conformation a protein will assume to allow functional activity. Following urea-induced denaturation of ribonuclease A, the protein was allowed to refold by removal of urea. The protein was found to regain its native structure and functionality after this treatment suggesting that proteins can adopt their native conformation spontaneously (Anfinsen, 1973). This finding was awarded with the Nobel Prize in Chemistry in 1972 and opened up an avenue of experimental and theoretical work in the field of protein folding and unfolding. The structural insights into the folding and unfolding processes of many proteins have since then been explored using a vast range of biophysical instrumentation, both at the ensemble (reviewed in Buchner et al., 2011; Sanchez-Ruiz, 2011) as well as at single molecule (reviewed in Borgia et al., 2008; Ferreon & Deniz, 2011) level. Protein conformational stability can be defined as the ability of the natively folded structure of a protein to resist unfolding. Two types of stability can be distinguished: the difference in energy content between the folded state and unfolded state of a protein is termed thermodynamic stability. Boltzmann's distribution law defines the distribution of folded

Detailed knowledge of the molecular parameters determining aggregation propensity, rate and morphology resulted in the ability to tune protein aggregation through chemical engineering. The attachment of sugar chains to proteins has been shown to inhibit selfassociation (Marquardt & Helenius, 1992; reviewed in Helenius et al., 1997; Land & Braakman, 2001; Song et al., 2001). This effect was largely attributed to covalently linked sugar moieties affecting kinetic partitioning between folding and aggregation from an (partially) unfolded state. For example, glycosylation was found to increase the folding rate of the protein rapidly shielding exposed hydrophobic regions which could potentially act as a driving force for aggregation at ambient temperature (Broersen et al., 2007b; Shental-Bechor & Levy, 2008; Wang et al., 1996). Interestingly, high temperatures induced more rapid aggregation of glycosylated proteins compared to their non-glycosylated counterparts (Broersen et al., 2007b; Chobert et al., 2006). However, another study involving glycated bovine serum albumin concluded that glycation of the protein inhibited its aggregation upon incubation at moderate temperatures (Rondeau et al., 2010, 2007). Glycosylation of proteins has also been shown to affect gel properties: attachment of a ketohexose to ovalbumin by means of the Maillard reaction resulted in the formation of gels with enhanced breaking strength (Sun et al., 2004). The effects of various types of chemical protein engineering on the physico-chemical functionality of proteins will be discussed in

Proteins in food products can lose their native structure as a result of processing conditions including storage, heat treatment, acidification, dehydration, mechanical processing or shear, and microbial hydrolysis. For example, long-term storage of milk powder has been found to induce lactosylation of the proteins present in the preparation which, in turn, results in affected powder solubility and emulsifying and foaming properties (reviewed in Thomas et al., 2004). Unfolded or hydrolyzed protein molecules can exert very different functionality to food products compared to folded proteins, a classical example being the boiling of an egg which converts the liquid-like transparent egg white into an opaque semisolid structure with very different textural properties. This paragraph will shortly discuss the principles of protein folding and structure and the forces that are implied. It was first recognized by Anfinsen (1973) that the primary sequence of a protein dictates the specific folded, or native, conformation a protein will assume to allow functional activity. Following urea-induced denaturation of ribonuclease A, the protein was allowed to refold by removal of urea. The protein was found to regain its native structure and functionality after this treatment suggesting that proteins can adopt their native conformation spontaneously (Anfinsen, 1973). This finding was awarded with the Nobel Prize in Chemistry in 1972 and opened up an avenue of experimental and theoretical work in the field of protein folding and unfolding. The structural insights into the folding and unfolding processes of many proteins have since then been explored using a vast range of biophysical instrumentation, both at the ensemble (reviewed in Buchner et al., 2011; Sanchez-Ruiz, 2011) as well as at single molecule (reviewed in Borgia et al., 2008; Ferreon & Deniz, 2011) level. Protein conformational stability can be defined as the ability of the natively folded structure of a protein to resist unfolding. Two types of stability can be distinguished: the difference in energy content between the folded state and unfolded state of a protein is termed thermodynamic stability. Boltzmann's distribution law defines the distribution of folded

more detail in paragraph 3 of this review.

**2.3 Protein structural integrity**

and unfolded protein molecules at a certain time. The rate at which the unfolded protein collapses to a folded state is reflected by the folding rate and represents the kinetic stability of a protein. Energy landscape theories and the folding funnel hypothesis have both been used as models to understand the energetic barriers of a protein separating the folded from the unfolded state (reviewed in Onuchic et al., 1997; Plotkin & Onuchic, 2002; Wolynes, 2005). Both models start from a similar principle in which there is an energy difference between the folded state of a protein and its unfolded state and that the folded state is defined as the favored entropic state of a protein. Local energy minima in the folding process can result in the accumulation of transient intermediate structures, which are neither folded, nor unfolded, to a larger or lesser extent, depending on the environmental conditions of folding or the primary sequence of the protein (reviewed in Baldwin, 2008; Englander et al., 2007). These intermediate structures, which often lack biological activity, are sometimes sufficiently stable to allow substantial accumulation (reviewed in Englander et al., 2007). Stable folding intermediates are related to a high propensity of aggregation as the hydrophobic core is not sufficiently shielded while opposing assembly forces are absent (reviewed in Ferreira et al., 2006). Kinetic protein stability is defined by a variety of molecular parameters including the proximity of native contacts in the primary sequence of a protein (Cieplak et al., 2004; Plaxco et al., 1998, 2000), internal friction or the energetic of intrachain interactions, energy barriers to backbone rotations and long-range residue interactions (Pabit et al., 2004; Qiu & Hagen, 2004), rate of diffusional motion of an unfolded peptide chain through the solvent (Pabit et al., 2004), and the presence or absence of intermediate state(s) (Baumketner, 2003; Onuchic et al., 1997). Forces that contribute to thermodynamic stability are the strength of intramolecular hydrogen bonds and solvent-protein interactions, both enthalpic in nature. The entropic contribution is mainly defined by the hydrophobic effect of folding through an increase in disorder of water molecules upon folding. Many of the forces retaining a protein structure intact can be disrupted, removed or introduced by chemical modification. For example, reaction of proteins with sugars by means of the Maillard reaction can lead to distinct changes in protein stability. The stability of proteins in food products can be affected by using proteins which have been modified by means of glycosylation or charge modification. Succinylation with the aim to increase net charge of soy protein hydrolysate lead to improved digestibility of the protein as investigated by a multienzyme method involving trypsin, chymotrypsin, and peptidase (Achouri & Zhang, 2001). This finding suggests that the protein had undergone structural rearrangement as a result of the succinylation process. Succinylation also lead to destabilization of Faba bean legumin (Schwenke et al., 1998). This group used differential scanning calorimetry (DSC) to study protein stability of legumin and found a decreased specific enthalpy for unfolding upon succinylation. Succinylation also resulted in an increased surface hydrophobicity of the protein suggesting at least partial unfolding of the molecule. Kosters and colleagues (2003) compared the effects of many different types of chemical modification on protein stability, including lipophilization using capric acid, glycosylation and succinylation of ovalbumin. Ovalbumin stability was probed by DSC, tryptophan fluorescence and circular dichroism (CD). Lipophilization resulted in a decreased denaturation temperature of ovalbumin reflected in an enthalpy decrease and a lower stability upon guanidine titration. Glycosylation was found to stabilize structural integrity of ovalbumin. A similar finding using CD and DSC to study temperature-induced unfolding has been reported upon glucosylation (Broersen et al., 2004; van Teeffelen et al., 2005) and fructosylation (Broersen et al., 2004) of lactoglobulin or glucosylation of codfish parvalbumin (de Jongh et al., 2011). Interestingly, assays performed at ambient temperature but involving denaturant-induced unfolding

Application Potential of Food Protein Modification 143

improved bactericidal properties even more (del Rosario Moreira et al., 2011) as a result of the ionic interaction between the two biopolymers (Pereda et al., 2008, 2009). Cao-Hoang and colleagues (2010) produced a nisin-containing sodium caseinate film to investigate the antimicrobial activity of both surface- and in-depth *Listeria innocua* inoculated soft cheese. The presence of the film reduced surface contamination with *L. Innocua* significantly, while antimicrobial activity within the cheese matrix depended on the distance from the film-coated surface. Antimicrobial films prepared from a mixture of oregano oil and whey protein isolate showed inhibition of growth of lactic acid bacteria, reduction of pseudomonads, total flora and growth rates when applied to fresh beef (Zinoviadou et al., 2009). Even though in many cases complex formulae have been employed, containing both a protein component as well as a carbohydrate or oil component, no publications are known that show the effects of covalently

Sensory aspects of food products include sensation of flavor, odor, color, and texture. These factors play a large role in consumer acceptation of food products and the effects of various types of protein chemical engineering and their applications will be discussed in this paragraph. Many types of modification target the amino groups of lysine residues, including succinylation, lipidation and glycosylation through the Maillard reaction. Textural properties, including emulsifying, foam and gelling capacities, have been discussed in detail

Even though most proteins are tasteless, ingestion of a small number of proteins is perceived as sweet. These include thaumatin (Ohta et al., 2008; van der Wel & Loeve, 1972), monellin (Morris & Cagan, 1972), brazein (Ming & Hellekant, 1994), and lysozyme (Masuda et al., 2001). The sweetness of lysozyme results from the abundance of lysine residues which was shown by alanine substitution in lysozyme (Masuda et al., 2005a). It is therefore perceivable that modification of lysine residues by conjugation of a chemical group has consequences for the sweetness of the protein. Extensive acetylation and phosphopyridoxylation of lysine residues of lysozyme decreased the perceived sweetness of the protein further demonstrating that lysine residues play a major role in sensory aspects of this protein (Masuda et al., 2005, Kaneko & Kitabatake, 2001). No other reports on flavor

modulating aspects as a result of chemical protein engineering have been reported.

One publication which studied the effect of acylation by acetic and succinic anhydride of flaxseed protein isolates reported no off-odors upon modification, although no results were

Three types of modifications have been reported to affect the color of the protein preparation. First, succinylation was reported to convert the color of soy isolate from tan to chalk-white upon visual inspection (Franzen & Kinsella, 1976a). Upon measurement of

linked lipid or sugar to protein films as a potential edible film.

**2.5 Sensory: Color, flavor, odor, texture** 

presented (Wanasundara & Shahidi, 1997).

in paragraphs 2.1 and 2.2.

**2.5.1 Flavor** 

**2.5.2 Odor** 

**2.5.3 Color** 

reported a decrease in protein stability upon conjugation of glucose to -lactoglobulin. This phenomenon, which appeared unique for a glycosylation reaction, has been studied in further detail by van Teeffelen and colleagues (2005). The observations could be explained in terms of a decreased change in heat capacity upon unfolding as a result of glucosylation indicating that the hydration pattern of proteins upon glycosylation is significantly affected.

#### **2.4 Shelf life**

Proteins can affect shelf life and stability of food products by enhancing antioxidant activity, affecting gas exchange, antimicrobial activity or by stabilization of emulsion or foam-based food products (del Rosario Moreira et al., 2011; Emmambux et al., 2004; Mendis et al., 2005). Protein films can be used as as packaging biomaterials as a result of their ability to form networks with rheologically advantageous characteristics (Arvanitoyannis, 1999; Audic & Chaufer, 2005, Longares et al., 2005). However, mixtures of for example proteins and polysaccharides have been found to exert superior functional properties compared to proteins or polysaccharides in isolation (reviewed by Pogaku et al., 2007). For example, the application of an edible coating of storage proteins obtained from sorghum, called kafirins, has been shown to extend the shelf life of freshly harvested pears (Buchner et al., 2011). The shelf life of meat was shown to be extended upon application of a collagen and gelatin coating and led to reduced decoloration, antioxidant activity, and reduction of microbial spoilage (Havard & Harmony, 1869). Such coatings extend quality and shelf-life by acting as a slow-release gas barrier (Baldwin, 1994; Buchner et al., 2011; Park, 1999). Nanobiocomposites of maize prolamin protein zein have also been employed as a gas barrier by coating tomatoes (Park et al., 1994), and apples (Bai et al., 2003). Even though collagen and gelatin coatings were reported to both effectively retain water in meat products (Antoniewski et al., 2007; Farouk et al., 1990), extensive moisture loss of kafirin coated pears compared to the uncoated product left them unacceptable toward consumers (Buchner et al., 2011). The authors (Buchner et al., 2011) therefore suggested to prepare wax or triglycerides/kafirin mixtures instead of pure kafirin coatings to prevent moisture loss as kafirin films themselves do not function effectively as water barriers (Emmambux et al., 2004; Gillgren & Stading, 2008). Because lipids form a very suitable moisture barrier as a result of their hydrophobic character, lipophilized proteins possibly form more effective coatings for fruit. However, to date, no work has been published to demonstrate the effect of lipid-incorporation into proteins to prepare stable coatings for fruit.

Some proteins and peptides are known to have antimicrobial activity (Nizet, 2006; reviewed by Wimley, 2010). Some of these are also applied as food preservatives such as nisin, which is a potent antibacterial 34 amino acid peptide containing a number of uncommon amino acids. Nisin has been employed as an approved food preservative in cheese (Martins et al., 2010), fish, meat, and beverages (reviewed in Lubelski et al., 2008). Another known antimicrobial peptide is -poly-L-lysine which exhibits antimicrobial activity against bacteria and fungi and is used as a food preservative (reviewed in Hamano, 2011). This asset has also been explored within a food based environment, for example by applying mixed formulations of chitosan, a linear polysaccharide, and casein polymers to a number of food products including carrot, cheese, and salami (del Rosario Moreira et al., 2011). Pure caseinate films applied to squash slices showed limited antimicrobial activity (Ponce et al., 2008). While chitosan alone exerts significant anti-microbial activity, the inclusion of casein polymers into the formulation improved bactericidal properties even more (del Rosario Moreira et al., 2011) as a result of the ionic interaction between the two biopolymers (Pereda et al., 2008, 2009). Cao-Hoang and colleagues (2010) produced a nisin-containing sodium caseinate film to investigate the antimicrobial activity of both surface- and in-depth *Listeria innocua* inoculated soft cheese. The presence of the film reduced surface contamination with *L. Innocua* significantly, while antimicrobial activity within the cheese matrix depended on the distance from the film-coated surface. Antimicrobial films prepared from a mixture of oregano oil and whey protein isolate showed inhibition of growth of lactic acid bacteria, reduction of pseudomonads, total flora and growth rates when applied to fresh beef (Zinoviadou et al., 2009). Even though in many cases complex formulae have been employed, containing both a protein component as well as a carbohydrate or oil component, no publications are known that show the effects of covalently linked lipid or sugar to protein films as a potential edible film.

#### **2.5 Sensory: Color, flavor, odor, texture**

Sensory aspects of food products include sensation of flavor, odor, color, and texture. These factors play a large role in consumer acceptation of food products and the effects of various types of protein chemical engineering and their applications will be discussed in this paragraph. Many types of modification target the amino groups of lysine residues, including succinylation, lipidation and glycosylation through the Maillard reaction. Textural properties, including emulsifying, foam and gelling capacities, have been discussed in detail in paragraphs 2.1 and 2.2.

#### **2.5.1 Flavor**

142 Advances in Chemical Engineering

reported a decrease in protein stability upon conjugation of glucose to -lactoglobulin. This phenomenon, which appeared unique for a glycosylation reaction, has been studied in further detail by van Teeffelen and colleagues (2005). The observations could be explained in terms of a decreased change in heat capacity upon unfolding as a result of glucosylation indicating that

Proteins can affect shelf life and stability of food products by enhancing antioxidant activity, affecting gas exchange, antimicrobial activity or by stabilization of emulsion or foam-based food products (del Rosario Moreira et al., 2011; Emmambux et al., 2004; Mendis et al., 2005). Protein films can be used as as packaging biomaterials as a result of their ability to form networks with rheologically advantageous characteristics (Arvanitoyannis, 1999; Audic & Chaufer, 2005, Longares et al., 2005). However, mixtures of for example proteins and polysaccharides have been found to exert superior functional properties compared to proteins or polysaccharides in isolation (reviewed by Pogaku et al., 2007). For example, the application of an edible coating of storage proteins obtained from sorghum, called kafirins, has been shown to extend the shelf life of freshly harvested pears (Buchner et al., 2011). The shelf life of meat was shown to be extended upon application of a collagen and gelatin coating and led to reduced decoloration, antioxidant activity, and reduction of microbial spoilage (Havard & Harmony, 1869). Such coatings extend quality and shelf-life by acting as a slow-release gas barrier (Baldwin, 1994; Buchner et al., 2011; Park, 1999). Nanobiocomposites of maize prolamin protein zein have also been employed as a gas barrier by coating tomatoes (Park et al., 1994), and apples (Bai et al., 2003). Even though collagen and gelatin coatings were reported to both effectively retain water in meat products (Antoniewski et al., 2007; Farouk et al., 1990), extensive moisture loss of kafirin coated pears compared to the uncoated product left them unacceptable toward consumers (Buchner et al., 2011). The authors (Buchner et al., 2011) therefore suggested to prepare wax or triglycerides/kafirin mixtures instead of pure kafirin coatings to prevent moisture loss as kafirin films themselves do not function effectively as water barriers (Emmambux et al., 2004; Gillgren & Stading, 2008). Because lipids form a very suitable moisture barrier as a result of their hydrophobic character, lipophilized proteins possibly form more effective coatings for fruit. However, to date, no work has been published to demonstrate the effect of

the hydration pattern of proteins upon glycosylation is significantly affected.

lipid-incorporation into proteins to prepare stable coatings for fruit.

Some proteins and peptides are known to have antimicrobial activity (Nizet, 2006; reviewed by Wimley, 2010). Some of these are also applied as food preservatives such as nisin, which is a potent antibacterial 34 amino acid peptide containing a number of uncommon amino acids. Nisin has been employed as an approved food preservative in cheese (Martins et al., 2010), fish, meat, and beverages (reviewed in Lubelski et al., 2008). Another known antimicrobial peptide is -poly-L-lysine which exhibits antimicrobial activity against bacteria and fungi and is used as a food preservative (reviewed in Hamano, 2011). This asset has also been explored within a food based environment, for example by applying mixed formulations of chitosan, a linear polysaccharide, and casein polymers to a number of food products including carrot, cheese, and salami (del Rosario Moreira et al., 2011). Pure caseinate films applied to squash slices showed limited antimicrobial activity (Ponce et al., 2008). While chitosan alone exerts significant anti-microbial activity, the inclusion of casein polymers into the formulation

**2.4 Shelf life** 

Even though most proteins are tasteless, ingestion of a small number of proteins is perceived as sweet. These include thaumatin (Ohta et al., 2008; van der Wel & Loeve, 1972), monellin (Morris & Cagan, 1972), brazein (Ming & Hellekant, 1994), and lysozyme (Masuda et al., 2001). The sweetness of lysozyme results from the abundance of lysine residues which was shown by alanine substitution in lysozyme (Masuda et al., 2005a). It is therefore perceivable that modification of lysine residues by conjugation of a chemical group has consequences for the sweetness of the protein. Extensive acetylation and phosphopyridoxylation of lysine residues of lysozyme decreased the perceived sweetness of the protein further demonstrating that lysine residues play a major role in sensory aspects of this protein (Masuda et al., 2005, Kaneko & Kitabatake, 2001). No other reports on flavor modulating aspects as a result of chemical protein engineering have been reported.

#### **2.5.2 Odor**

One publication which studied the effect of acylation by acetic and succinic anhydride of flaxseed protein isolates reported no off-odors upon modification, although no results were presented (Wanasundara & Shahidi, 1997).

#### **2.5.3 Color**

Three types of modifications have been reported to affect the color of the protein preparation. First, succinylation was reported to convert the color of soy isolate from tan to chalk-white upon visual inspection (Franzen & Kinsella, 1976a). Upon measurement of

Application Potential of Food Protein Modification 145

1975). However, this paragraph will entirely focus on the kind of chemical engineering intentionally brought about to link specific molecules to proteins which act as functional ingredients in food. These molecules change the behavior of the protein and are largely hypothesized to infer characteristics to the protein which are little present in the unmodified protein, such as improved foaming properties, inhibition of aggregation or enhanced surface activity. Rationale for chemical modification of proteins is multiple but can be categorized

i. Waste control: For example the re-use of fish gelatin from waste requires less natural resources for their production. Other examples include the production of a peptide with anti-oxidant activity from algae (Sheih et al., 2009). Protein rich by-products are also recovered upon electrocoagulation of wastewater resulting from egg processing (Xu et al., 2002). Chemical modification can be used to increase the functional properties of

ii. Health considerations: An example of this is the replacement of meat or soy proteins by (other) vegetable proteins (O'Kane et al., 2004; Pedroche et al., 2004; Vioque et al., 1999; reviewed in Moure et al., 2006). Similar to proteins obtained as by-product from waste material, unmodified plant proteins often have limited functional applicability. Hence, chemical modification of plant proteins can be used to improve molecular functionality. iii. Cost effectivity by extending molecular functionality of a protein. Chemical modification can be employed to enhance the functional properties of a protein, such that less material is required to obtain a product with similar structural characteristics. iv. Structure-function relationships: Chemical modification is often used to investigate the contribution of specific molecular parameters, such as surface hydrophobicity, to functionality of the protein at ingredient level. Ample examples of protein functional

The types of modification described in literature are extensive and include phosphorylation (attachment of a phosphate to serine, tyrosine or threonine), methylation (attachment attachment of a methyl group to arginine or the N-terminus of the protein), glycosylation (attachment of carbohydrates to lysine or the N-terminus), acetylation (attachment of acetyl to an amino group such as lysine or the N-terminus), and many more. The various types of modification and their impact on molecular behavior of proteins have been reviewed in a vast number of publications and book chapters (Feeney et al., 1982; Means & Feeney, 1971; Tawfik, 2002). This paragraph aims to shortly discuss the current state-of-the-art of the various reactive groups in proteins which can be targeted by chemical engineering. The chemical basis of these modifications will be discussed and applications from literature will be reviewed merely to illustrate the wide variety of applications of the chemical

Food storage and preparation processes such as heating by pasteurization or sterilization often provide for conditions which induce spontaneous and uncontrolled reaction of a reducing carbohydrate present in the food matrix with proteins. This reaction, termed the Maillard reaction, which is actually a complex cascade of reactions, is responsible for the formation of browning products and can have substantial impact on the flavor and color of food products. The Maillard reaction is initiated by a condensation reaction between the -

extension are discussed in the rest of this paragraph.

into four main reasons.

these proteins.

modification procedure.

**3.1 Glycosylation and deglycosylation** 

surface reflectance using a colorimeter, increasing degrees of succinylation of flax seed protein lead to brighter protein preparations (Wanasundara & Shahidi, 1997). A similar observation has been reported for succinylation of fish muscle (Groninger, 1973), alfa-alfa leaf protein (Franzen & Kinsella, 1976b), and soy bean protein preparations in a U.S. patent for coffee whitener (Melnychym & Stapley, 1973). Acylation by treatment of protein with acetic anhydride also lead to *brighter* flaxseed protein isolate, although the effect of succinylation of the same protein was stronger (Wanasundara & Shahidi, 1997). Franzen and Kinsella (1976a) showed no effect on color upon acetylation of soy isolate. A third type of modification with a strong effect on color is glycosylation through the Maillard reaction. This reaction is a complex cascade of reactions initiated by the interaction of a reducing sugar with an amino group. Colored products are formed only at later stages of the reaction pathway and include aldols and melanoidins which are high molecular weight compounds (reviewed in Zamora & Hidalgo, 2006). Paragraph 3.1 further extends on the formation of browning products and Amadori compounds related to the Maillard reaction.

#### **2.5.4 Texture**

Texture is a complex consumer perception of mouthfeel, tastants and afterfeel. A number of most relevant texture attributes, like 'spreadability' (essential for attributes like 'creamy'), 'crumbliness' or 'separating/wateriness' have been shown to be directly related to the energy household in protein-based products when energy is exerted onto the system. This applied energy may arise from oral processing, exposure to heat, gravity or applied pressure during for example industrial processing. The energy balance in protein-based food structures has been proposed by van Vliet and Walstra (1995): Wapplied = Wstored+ Wdissipated + Wfracture. This model implies that all energy applied to the gel can be used either for fracture, can dissipate or can be stored in the network (and regained after release of exerted forces). At a microstructural level this translates into fracture nucleation points and fracture propagation modes, whereas energy dissipation is often assumed to be controlled by serum flow properties. This latter factor is directly determined by the porosity of the gel as set-up by the microstructure and the pore deformation propensity when the system is put under strain. Van den Berg and colleagues (2008) showed that an attribute like 'spreadability' is directly related to directing as much energy as possible to fracture. When energy can be efficiently stored in the protein matrix, this directs the attribute 'crumbliness'. The effective interaction between protein-building blocks that make up the product matrix can be engineered. Strengthening this interaction, using for example transglutaminase (e.g. Dondero et al., 2006), will direct the energy flow from fracture to storage and gels will become less spreadable and more crumbly. Sala and co-workers (2008) showed that by modulating the interaction between a filler (like a fat-particle) and the protein matrix the texture of these protein gels could be strongly affected. This balance between active-inactive property of the filler could be delicately controlled by means of chemical modification.

#### **3. Types of chemical protein engineering – Exploring potential functionality**

Proteins can be chemically modified *in vitro* by covalently cross-linking the protein with a molecule of interest. Many of these reactions will also take place post-translationally in the strictly regulated environment of the cell, such as phosphorylation of cAMP-dependent protein kinases which plays a role in the enhancement of glycogen degradation (Soderling,

surface reflectance using a colorimeter, increasing degrees of succinylation of flax seed protein lead to brighter protein preparations (Wanasundara & Shahidi, 1997). A similar observation has been reported for succinylation of fish muscle (Groninger, 1973), alfa-alfa leaf protein (Franzen & Kinsella, 1976b), and soy bean protein preparations in a U.S. patent for coffee whitener (Melnychym & Stapley, 1973). Acylation by treatment of protein with acetic anhydride also lead to *brighter* flaxseed protein isolate, although the effect of succinylation of the same protein was stronger (Wanasundara & Shahidi, 1997). Franzen and Kinsella (1976a) showed no effect on color upon acetylation of soy isolate. A third type of modification with a strong effect on color is glycosylation through the Maillard reaction. This reaction is a complex cascade of reactions initiated by the interaction of a reducing sugar with an amino group. Colored products are formed only at later stages of the reaction pathway and include aldols and melanoidins which are high molecular weight compounds (reviewed in Zamora & Hidalgo, 2006). Paragraph 3.1 further extends on the formation of

browning products and Amadori compounds related to the Maillard reaction.

Texture is a complex consumer perception of mouthfeel, tastants and afterfeel. A number of most relevant texture attributes, like 'spreadability' (essential for attributes like 'creamy'), 'crumbliness' or 'separating/wateriness' have been shown to be directly related to the energy household in protein-based products when energy is exerted onto the system. This applied energy may arise from oral processing, exposure to heat, gravity or applied pressure during for example industrial processing. The energy balance in protein-based food structures has been proposed by van Vliet and Walstra (1995): Wapplied = Wstored+ Wdissipated + Wfracture. This model implies that all energy applied to the gel can be used either for fracture, can dissipate or can be stored in the network (and regained after release of exerted forces). At a microstructural level this translates into fracture nucleation points and fracture propagation modes, whereas energy dissipation is often assumed to be controlled by serum flow properties. This latter factor is directly determined by the porosity of the gel as set-up by the microstructure and the pore deformation propensity when the system is put under strain. Van den Berg and colleagues (2008) showed that an attribute like 'spreadability' is directly related to directing as much energy as possible to fracture. When energy can be efficiently stored in the protein matrix, this directs the attribute 'crumbliness'. The effective interaction between protein-building blocks that make up the product matrix can be engineered. Strengthening this interaction, using for example transglutaminase (e.g. Dondero et al., 2006), will direct the energy flow from fracture to storage and gels will become less spreadable and more crumbly. Sala and co-workers (2008) showed that by modulating the interaction between a filler (like a fat-particle) and the protein matrix the texture of these protein gels could be strongly affected. This balance between active-inactive property of the filler could be delicately controlled by means of chemical modification.

**3. Types of chemical protein engineering – Exploring potential functionality**  Proteins can be chemically modified *in vitro* by covalently cross-linking the protein with a molecule of interest. Many of these reactions will also take place post-translationally in the strictly regulated environment of the cell, such as phosphorylation of cAMP-dependent protein kinases which plays a role in the enhancement of glycogen degradation (Soderling,

**2.5.4 Texture** 

1975). However, this paragraph will entirely focus on the kind of chemical engineering intentionally brought about to link specific molecules to proteins which act as functional ingredients in food. These molecules change the behavior of the protein and are largely hypothesized to infer characteristics to the protein which are little present in the unmodified protein, such as improved foaming properties, inhibition of aggregation or enhanced surface activity. Rationale for chemical modification of proteins is multiple but can be categorized into four main reasons.


The types of modification described in literature are extensive and include phosphorylation (attachment of a phosphate to serine, tyrosine or threonine), methylation (attachment attachment of a methyl group to arginine or the N-terminus of the protein), glycosylation (attachment of carbohydrates to lysine or the N-terminus), acetylation (attachment of acetyl to an amino group such as lysine or the N-terminus), and many more. The various types of modification and their impact on molecular behavior of proteins have been reviewed in a vast number of publications and book chapters (Feeney et al., 1982; Means & Feeney, 1971; Tawfik, 2002). This paragraph aims to shortly discuss the current state-of-the-art of the various reactive groups in proteins which can be targeted by chemical engineering. The chemical basis of these modifications will be discussed and applications from literature will be reviewed merely to illustrate the wide variety of applications of the chemical modification procedure.

#### **3.1 Glycosylation and deglycosylation**

Food storage and preparation processes such as heating by pasteurization or sterilization often provide for conditions which induce spontaneous and uncontrolled reaction of a reducing carbohydrate present in the food matrix with proteins. This reaction, termed the Maillard reaction, which is actually a complex cascade of reactions, is responsible for the formation of browning products and can have substantial impact on the flavor and color of food products. The Maillard reaction is initiated by a condensation reaction between the -

Application Potential of Food Protein Modification 147

and stabilize emulsions and foams which is the result of improved potential to interact with hydrophobic surfaces, both the air-water and oil-water interface, and including (model)membranes (Nakai, 1983; Wierenga et al., 2003; reviewed in Wilde, 2000; Wilde et al., 2004). Various saturated and unsaturated fatty acids have been employed to induce lipophilization of proteins including caproic acid (Liu et al., 2000), capric acid (Aewsiri et al., 2010; Kosters et al., 2003; Liu et al., 2000), lauric acid (Aewsiri et al., 2010), myristic acid (Aewsiri et al., 2010; Ibrahim et al., 1993; Liu et al., 2000), palmitic acid (Haque et al., 1982; Haque & Kito, 1983a, 1983b; Ibrahim et al., 1991), stearic acid (Djagny et al., 2001; Ibrahim et al., 1993), and oxidized forms of linoleic acid (Aewsiri et al., 2011a, 2011b), and the efficiency of the lipophilization reaction was found to be inversely proportional to the length of the lipid chains used (Liu et al., 2000). Reaction of 28% of the available free amino groups of ovalbumin with activated capric acid was shown to result in retained secondary structure while inducing oligomerization and destabilization of the protein structure as a result of lowering the enthalpy for unfolding (Kosters et al., 2003). The presence of acyl chains was thought to cause significant dehydration of the protein. In another study, hen egg white lysozyme was lipophilized with short and middle chain saturated fatty acids including caproic (C6:0), capric (C10:0), and myristic (C14:0) acid (Liu et al., 2000). Lipophilization of lysozyme was reported to decrease the thermal stability of lysozyme as a result of partial loss of -helical content of the protein, and this molecular destabilization appeared to be proportionally related to the chain length and the number of bound fatty acids. The lysine residues involved in the modification were thought to be located in the helical region and to subsequently induce partial unfolding of the -helical region surrounding these residues (Liu et al., 2000). Lysozyme has also been chemically modified using palmitic acid (Ibrahim et al., 1991, 1993) with the primary aim to study the effect of lipophilization on the antimicrobial effect of the protein. Even though increasing extents of covalent linkage with palmitoyl residues lead to insoluble protein, as spectrophotometrically determined by solution turbidity at 500 nm, foaming stability and emulsifying activity were progressively improved by linkage of palmitic acid to the protein molecule. More groups showed that the foaming or emulsifying activities of a wide range of proteins could be improved upon lipophilization, including soybean glycinin (Haque et al., 1982), s1-casein (Haque & Kito, 1983b), and cuttlefish skin gelatin (Awesiri et al., 2011a, 2011b). A further effect resulting from the incorporation of myristic and stearic acids into lysozyme was related to antimicrobial activity and stearic and palmitic acid conjugation resulted in more effective antimicrobial agents against *E. coli*, than the attachment of myristic acid or the unmodified protein (Ibrahim, 1993). Myristoylation was found to induce lysozyme aggregation resulting in concurrent loss of antimicrobial function. The effects of palmitoylation on the structural and functional properties of s1-casein have also been explored (Haque & Kito, 1983a, 1983b). The conjugation of the -amino groups of s1-casein with palmitic acid lead to micelle formation as a result of increased hydrophobicity while negative net charge was increased (Haque & Kito 1983a). Further work by this group showed that palmitic acid linkage to s1-casein did not lead to large scale structural rearrangement of the molecule, both at a secondary and a tertiary structure level using circular dichroism. Interestingly, Aewsiri and colleagues have also investigated the antioxidative activity of cuttlefish skin gelatin modified with a combination of oxidized linoleic acid and oxidized tannic acid, a potent antioxidant (Aewsiri et al., 2011a). Oxidation of lipids and proteins in foams primarily takes place at the air-water interface and the addition of a hydrophilic antioxidant alone reduces surface activity (Aewsiri et al., 2011a). Co-conjugation of tannic acid

amino group of lysine and the reducing group of a sugar to form Amadori or Heyn's rearrangement products via *N*-substituted glycosylamine. During the advanced stages of this reaction, the Amadori and Heyn's rearrangement products are degraded via a number of pathways (Mossine et al., 1994; Röper et al., 1983). The last stages of the Maillard reaction involve extensive protein cross-linking reactions and the formation of so-called melanoidins (Pellegrino et al., 1999). As a result of the wide range of intermediate chemical structures formed, several of these intermediates can be employed as indicators of the Maillard reaction to monitor the extent of the reaction in food products as a measure of quality control. For example, -*N*-(furoylmethyl)-L-lysine (furosine) formation, an intermediate in the Maillard reaction, was shown to be the result of lactosylation upon storage of milk powder (Le et al., 2011). Also prolonged storage of high-protein nutrition bars showed nonenzymatic Maillard browning as a result of interaction between whey protein isolate and high-fructose corn syrup or sorbitol syrup (McMahon et al., 2009). The baking process of bread was found to affect color formation determined by furosine and hydroxymethylfurfural concentrations (Ramírez-Jiménez et al., 2000). Variation of baking temperature and dough composition determines the extent of furosine formation and loss, acid-released lysine, and carboxymethyllysine formation of cookies (Charissou et al., 2007). Glycosylation via the Maillard reaction has also been brought about intentionally to study the effects of covalent sugar linkage to proteins in terms of e.g. protein stability and aggregation (Feeney et al., 1975). Glycosylation of proteins by means of the Maillard reaction has been observed to both induce and protect against aggregation. Incubation of hazelnut proteins with glucose resulted in the formation of high molecular weight protein aggregates detected by SDS-PAGE (Cucu et al., 2011). At the same time, others have shown that Maillardation with glucose inhibited the aggregation of cod fish parvalbumin (de Jongh et al., 2011). As the Maillard reaction is a reaction involving many steps, one possibility which can be raised to explain the discrepancy between these observations is that the various intermediates may display differences in resistance against aggregation, some of which may be protective, others which may be inducing aggregation. Increased thermal stability upon glycosylation was found for many proteins including the apple allergen Mal d3 upon reaction with glucose (Sancho et al., 2005), and -lactoglobulin reaction with glucose and fructose (Broersen et al., 2005). Alternatively, deglycosylation was shown to induce denaturation and aggregation of ovalbumin (de Groot et al., 2007). Apart from the Maillard reaction, protein glycosylation can be achieved by several other routes. *N*glycosylation takes place by modification of the side chains of asparagine or arginine (Kornfeld & Kornfeld, 1985). *O*-glycosylation is brought about by modification of serine, threonine, or tyrosine (Hart, 1992). Many different glycan structures have been identified to be involved in these two types of modification and they are often necessarily involved in biological function of the protein (Rudd et al., 2001). These last two types of glycosylation take place mainly as a result of post-translational processing of proteins *in vivo* and are not used as means to induce glycosylation of proteins applied in the food industry. Hence, *N*and *O*-glycosylation will not be discussed in this paragraph.

#### **3.2 Lipophilization**

Covalent linkage of lipids to proteins results in increased hydrophobic exposure of a protein with interesting applications related to altered surface properties. Increased exposed hydrophobicity of proteins has for example been related to an improved capacity to form

amino group of lysine and the reducing group of a sugar to form Amadori or Heyn's rearrangement products via *N*-substituted glycosylamine. During the advanced stages of this reaction, the Amadori and Heyn's rearrangement products are degraded via a number of pathways (Mossine et al., 1994; Röper et al., 1983). The last stages of the Maillard reaction involve extensive protein cross-linking reactions and the formation of so-called melanoidins (Pellegrino et al., 1999). As a result of the wide range of intermediate chemical structures formed, several of these intermediates can be employed as indicators of the Maillard reaction to monitor the extent of the reaction in food products as a measure of quality control. For example, -*N*-(furoylmethyl)-L-lysine (furosine) formation, an intermediate in the Maillard reaction, was shown to be the result of lactosylation upon storage of milk powder (Le et al., 2011). Also prolonged storage of high-protein nutrition bars showed nonenzymatic Maillard browning as a result of interaction between whey protein isolate and high-fructose corn syrup or sorbitol syrup (McMahon et al., 2009). The baking process of bread was found to affect color formation determined by furosine and hydroxymethylfurfural concentrations (Ramírez-Jiménez et al., 2000). Variation of baking temperature and dough composition determines the extent of furosine formation and loss, acid-released lysine, and carboxymethyllysine formation of cookies (Charissou et al., 2007). Glycosylation via the Maillard reaction has also been brought about intentionally to study the effects of covalent sugar linkage to proteins in terms of e.g. protein stability and aggregation (Feeney et al., 1975). Glycosylation of proteins by means of the Maillard reaction has been observed to both induce and protect against aggregation. Incubation of hazelnut proteins with glucose resulted in the formation of high molecular weight protein aggregates detected by SDS-PAGE (Cucu et al., 2011). At the same time, others have shown that Maillardation with glucose inhibited the aggregation of cod fish parvalbumin (de Jongh et al., 2011). As the Maillard reaction is a reaction involving many steps, one possibility which can be raised to explain the discrepancy between these observations is that the various intermediates may display differences in resistance against aggregation, some of which may be protective, others which may be inducing aggregation. Increased thermal stability upon glycosylation was found for many proteins including the apple allergen Mal d3 upon reaction with glucose (Sancho et al., 2005), and -lactoglobulin reaction with glucose and fructose (Broersen et al., 2005). Alternatively, deglycosylation was shown to induce denaturation and aggregation of ovalbumin (de Groot et al., 2007). Apart from the Maillard reaction, protein glycosylation can be achieved by several other routes. *N*glycosylation takes place by modification of the side chains of asparagine or arginine (Kornfeld & Kornfeld, 1985). *O*-glycosylation is brought about by modification of serine, threonine, or tyrosine (Hart, 1992). Many different glycan structures have been identified to be involved in these two types of modification and they are often necessarily involved in biological function of the protein (Rudd et al., 2001). These last two types of glycosylation take place mainly as a result of post-translational processing of proteins *in vivo* and are not used as means to induce glycosylation of proteins applied in the food industry. Hence, *N*-

and *O*-glycosylation will not be discussed in this paragraph.

Covalent linkage of lipids to proteins results in increased hydrophobic exposure of a protein with interesting applications related to altered surface properties. Increased exposed hydrophobicity of proteins has for example been related to an improved capacity to form

**3.2 Lipophilization** 

and stabilize emulsions and foams which is the result of improved potential to interact with hydrophobic surfaces, both the air-water and oil-water interface, and including (model)membranes (Nakai, 1983; Wierenga et al., 2003; reviewed in Wilde, 2000; Wilde et al., 2004). Various saturated and unsaturated fatty acids have been employed to induce lipophilization of proteins including caproic acid (Liu et al., 2000), capric acid (Aewsiri et al., 2010; Kosters et al., 2003; Liu et al., 2000), lauric acid (Aewsiri et al., 2010), myristic acid (Aewsiri et al., 2010; Ibrahim et al., 1993; Liu et al., 2000), palmitic acid (Haque et al., 1982; Haque & Kito, 1983a, 1983b; Ibrahim et al., 1991), stearic acid (Djagny et al., 2001; Ibrahim et al., 1993), and oxidized forms of linoleic acid (Aewsiri et al., 2011a, 2011b), and the efficiency of the lipophilization reaction was found to be inversely proportional to the length of the lipid chains used (Liu et al., 2000). Reaction of 28% of the available free amino groups of ovalbumin with activated capric acid was shown to result in retained secondary structure while inducing oligomerization and destabilization of the protein structure as a result of lowering the enthalpy for unfolding (Kosters et al., 2003). The presence of acyl chains was thought to cause significant dehydration of the protein. In another study, hen egg white lysozyme was lipophilized with short and middle chain saturated fatty acids including caproic (C6:0), capric (C10:0), and myristic (C14:0) acid (Liu et al., 2000). Lipophilization of lysozyme was reported to decrease the thermal stability of lysozyme as a result of partial loss of -helical content of the protein, and this molecular destabilization appeared to be proportionally related to the chain length and the number of bound fatty acids. The lysine residues involved in the modification were thought to be located in the helical region and to subsequently induce partial unfolding of the -helical region surrounding these residues (Liu et al., 2000). Lysozyme has also been chemically modified using palmitic acid (Ibrahim et al., 1991, 1993) with the primary aim to study the effect of lipophilization on the antimicrobial effect of the protein. Even though increasing extents of covalent linkage with palmitoyl residues lead to insoluble protein, as spectrophotometrically determined by solution turbidity at 500 nm, foaming stability and emulsifying activity were progressively improved by linkage of palmitic acid to the protein molecule. More groups showed that the foaming or emulsifying activities of a wide range of proteins could be improved upon lipophilization, including soybean glycinin (Haque et al., 1982), s1-casein (Haque & Kito, 1983b), and cuttlefish skin gelatin (Awesiri et al., 2011a, 2011b). A further effect resulting from the incorporation of myristic and stearic acids into lysozyme was related to antimicrobial activity and stearic and palmitic acid conjugation resulted in more effective antimicrobial agents against *E. coli*, than the attachment of myristic acid or the unmodified protein (Ibrahim, 1993). Myristoylation was found to induce lysozyme aggregation resulting in concurrent loss of antimicrobial function. The effects of palmitoylation on the structural and functional properties of s1-casein have also been explored (Haque & Kito, 1983a, 1983b). The conjugation of the -amino groups of s1-casein with palmitic acid lead to micelle formation as a result of increased hydrophobicity while negative net charge was increased (Haque & Kito 1983a). Further work by this group showed that palmitic acid linkage to s1-casein did not lead to large scale structural rearrangement of the molecule, both at a secondary and a tertiary structure level using circular dichroism. Interestingly, Aewsiri and colleagues have also investigated the antioxidative activity of cuttlefish skin gelatin modified with a combination of oxidized linoleic acid and oxidized tannic acid, a potent antioxidant (Aewsiri et al., 2011a). Oxidation of lipids and proteins in foams primarily takes place at the air-water interface and the addition of a hydrophilic antioxidant alone reduces surface activity (Aewsiri et al., 2011a). Co-conjugation of tannic acid

Application Potential of Food Protein Modification 149

modification, no disulfide bond aggregation was observed. Final aggregate morphology, gel formation and stability are affected as a result of rapid covalent network formation which does not allow rearrangement into more stable networks, as illustrated by lower gel Young's

Net charge and local charge densities of proteins have been implicated in the regulation of protein stability, aggregation, and aggregate morphology affecting the visual appearance of food products. These hypotheses have been substantiated by a range of observations which involved charge introduction, removal or reversal through succinylation and methylation reactions (Broersen et al., 2007a; Weijers et al., 2008). The reactions of succinylation and acetylation both lead to blockage of the reactive amino groups of proteins with an acyl residue and are hence collectively termed acylation reactions. The rate of acylation reaction depends on the rate of nucleophilic attack. Succinylation leads to increased net negative charge by the covalent linkage of succinate anions to the cationic amino groups of a protein thereby converting a cationic group into an anionic residue having implications for the distribution of net charge of a protein. Upon acetylation, ammonium cations are replaced by neutral acetyl groups resulting in electrostatically neutral groups. Large extents of succinylation have been reported to affect the integrity of secondary and tertiary structure of soy protein hydrolysate as shown by intrinsic tryptophan fluorescence and circular dichroism (Achouri & Zhang, 2001). Similar conformational rearrangements have been reported upon succinylation of whey protein isolate (Gruener & Ismond, 1997), bovine serum albumin (Jonas & Weber, 1970), canola protein (Lakkis & Villota, 1992), Faba bean legumin (Schwenke et al., 1998), rapeseed 12S globulin (Gueguen et al., 1990), and winged bean protein (Narayana & Rao, 1991). As a result of co-incubation of soy protein hydrolysate with succinic anhydride, which is a common compound used to succinylate proteins, heterogeneous reaction mixtures were obtained. Next to the aimed amine groups, this method of succinylation also commonly results in *O*-succinylation, involving threonine or serine hydroxyl groups or tyrosine succinylation (Achouri & Zhang, 2001; Chang & Sun, 1978; Schwenke et al., 1998). This latter reaction was found to be reversible upon treatment with hydroxylamine (Habeeb & Atassi, 1969), but, when present, to induce substantial expansion of Faba bean legumin as observed by viscometric studies (Schwenke et al., 1998). It was postulated that the high accumulation of negative charge upon extensive succinylation leads to dissociation and expansion of the individual subunits legumin is composed of (Schwenke et al., 1998). Other functional properties are equally affected by succinylation. For example, protein solubility has been reported to increase upon succinylation as has been demonstrated for rapeseed preparations (Dua et al., 1996), flax protein isolate (Wanasundara & Shahidi, 1997), oat protein isolate (Mirmoghatadaie et al., 2009), and soy protein isolate (Franzen & Kinsella, 1976a). Improved solubility has been related to the ability of proteins to perform more efficiently as stabilizers in emulsions and foams (Nakai & Li-Chan, 1988; Waniska & Kinsella, 1979), which, in turn, is greatly affected by their ability to absorb at the air-water interface (Wierenga et al., 2005). It has indeed been shown that treatment of a variety of proteins with succinic anhydride leads to increased foam capacity (Dua et al., 1996; Franzen & Kinsella, 1976a; Mirmoghatadaie et al., 2009), although others suggest that succinylation leads to decreased foam expansion capacity (Wanasundara & Shahidi, 1997). These seemingly contradictive findings may be explained

moduli obtained upon thiolation (Broersen et al., 2006).

**3.4 Charge modification by methylation and succinylation** 

and linoleic acid to gelatin both improves migration of the protein to the air-water or oilwater interface improving foaming and emulsifying activity, respectively, while retaining anti-oxidant activity (Aewsiri et al., 2011a).

#### **3.3 Chemical-reactive groups**

Sulfhydryl groups play an important role in regulating the self-assembly of proteins as well as their stability driven by disulfide interchange reactions (Sawyer, 1968). Hence, the presence of these groups has substantial impact on the aggregation and gelation behavior of a wide variety of proteins which has been confirmed by many researchers (Arntfield et al., 1991; Broersen et al., 2006; Graña-Montes et al., 2011; Hayakawa & Nakai, 1985; Hoffmann & van Mil, 1997; Margoshes, 1990; Mine, 1992; Sawyer, 1968; Shimada & Cheftel, 1989). A variety of modifications can be performed targeting sulfhydryl groups, which are part of the cysteine residues. Sulfhydryl groups are highly reactive against various reactants and are thus suitable targets for modification. Sulfhydryls can be blocked to prevent cross-linking by S-methyl methanethiosulfonate (MMTS), *N*-ethylmaleimide (NEM) (Kitabatake et al., 2001), or iodoacetamide (Anson, 1940; Huggins & Jensen, 1949; Smythe, 1936), or additional sulfhydryl-groups can be attached to primary amines (SATA) of proteins. Further, *N*hydroxysuccinimide esters can react irreversibly with primary amines releasing *N*hydroxysuccinimide.

Disulfide bonds are thought to play a crucial role determining the stability of proteins (Betz, 1993; Zavodszky et al., 2001; reviewed by Creighton, 1988) as well as to impact on the aggregation process and gel formation of various proteins including ovalbumin (Broersen et al., 2006; Kato et al., 1983), vicilin (Arntfield et al., 1991), and -lactoglobulin (Sawyer, 1968). Aggregates and gel networks are often the result of combined action of hydrophobic and electrostatic interactions and covalent interactions, in the form of disulfide bonds, are sometimes present (Kato et al., 1983; Koseki et al., 1989; Sun & Hayakawa, 2002). Thiolation of ovalbumin mediated through the reaction of S-acetylmercaptosuccinic anhydride (S-AMSA) with primary amines results in the formation of acetylthio groups and the acetyl group can be cleaved off to yield reactive sulfhydryl groups by the addition of hydroxylamine (Klotz & Heiney, 1962). A range of modification degrees can be obtained by varying the S-AMSA:lysine ratio (Broersen et al., 2006). Next to the linkage of a sulfhydryl group, additional carboxyl groups are conjugated through this reaction introducing additional charge variation which can lead to an additional parameter which can induce variation in aggregation, gelation or stability of a protein. To circumvent this variation, proteins with activated sulfhydryl groups are best compared with similarly modified proteins with blocked (i.e. not reacted with hydroxylamine) acetylthio groups, rather than directly with the unmodified protein (Broersen et al., 2006). Thiolation of ovalbumin in this way lead to limited changes at a secondary and tertiary structure level at high degrees of modification suggesting that the original molecular fold was largely retained upon modification. High degrees of thiolation resulted in a decrease of thermal stability of ovalbumin while fibril morphology was affected. Interestingly, the rate of aggregate formation was not modified by the presence of additional sulfhydryl groups available for disulfide formation. It was concluded that disulfide formation does not represent the prime driving force for aggregation of ovalbumin which was further illustrated by the finding that at room temperature, where significant sulfhydryl groups are already exposed upon

and linoleic acid to gelatin both improves migration of the protein to the air-water or oilwater interface improving foaming and emulsifying activity, respectively, while retaining

Sulfhydryl groups play an important role in regulating the self-assembly of proteins as well as their stability driven by disulfide interchange reactions (Sawyer, 1968). Hence, the presence of these groups has substantial impact on the aggregation and gelation behavior of a wide variety of proteins which has been confirmed by many researchers (Arntfield et al., 1991; Broersen et al., 2006; Graña-Montes et al., 2011; Hayakawa & Nakai, 1985; Hoffmann & van Mil, 1997; Margoshes, 1990; Mine, 1992; Sawyer, 1968; Shimada & Cheftel, 1989). A variety of modifications can be performed targeting sulfhydryl groups, which are part of the cysteine residues. Sulfhydryl groups are highly reactive against various reactants and are thus suitable targets for modification. Sulfhydryls can be blocked to prevent cross-linking by S-methyl methanethiosulfonate (MMTS), *N*-ethylmaleimide (NEM) (Kitabatake et al., 2001), or iodoacetamide (Anson, 1940; Huggins & Jensen, 1949; Smythe, 1936), or additional sulfhydryl-groups can be attached to primary amines (SATA) of proteins. Further, *N*hydroxysuccinimide esters can react irreversibly with primary amines releasing *N*-

Disulfide bonds are thought to play a crucial role determining the stability of proteins (Betz, 1993; Zavodszky et al., 2001; reviewed by Creighton, 1988) as well as to impact on the aggregation process and gel formation of various proteins including ovalbumin (Broersen et al., 2006; Kato et al., 1983), vicilin (Arntfield et al., 1991), and -lactoglobulin (Sawyer, 1968). Aggregates and gel networks are often the result of combined action of hydrophobic and electrostatic interactions and covalent interactions, in the form of disulfide bonds, are sometimes present (Kato et al., 1983; Koseki et al., 1989; Sun & Hayakawa, 2002). Thiolation of ovalbumin mediated through the reaction of S-acetylmercaptosuccinic anhydride (S-AMSA) with primary amines results in the formation of acetylthio groups and the acetyl group can be cleaved off to yield reactive sulfhydryl groups by the addition of hydroxylamine (Klotz & Heiney, 1962). A range of modification degrees can be obtained by varying the S-AMSA:lysine ratio (Broersen et al., 2006). Next to the linkage of a sulfhydryl group, additional carboxyl groups are conjugated through this reaction introducing additional charge variation which can lead to an additional parameter which can induce variation in aggregation, gelation or stability of a protein. To circumvent this variation, proteins with activated sulfhydryl groups are best compared with similarly modified proteins with blocked (i.e. not reacted with hydroxylamine) acetylthio groups, rather than directly with the unmodified protein (Broersen et al., 2006). Thiolation of ovalbumin in this way lead to limited changes at a secondary and tertiary structure level at high degrees of modification suggesting that the original molecular fold was largely retained upon modification. High degrees of thiolation resulted in a decrease of thermal stability of ovalbumin while fibril morphology was affected. Interestingly, the rate of aggregate formation was not modified by the presence of additional sulfhydryl groups available for disulfide formation. It was concluded that disulfide formation does not represent the prime driving force for aggregation of ovalbumin which was further illustrated by the finding that at room temperature, where significant sulfhydryl groups are already exposed upon

anti-oxidant activity (Aewsiri et al., 2011a).

**3.3 Chemical-reactive groups**

hydroxysuccinimide.

modification, no disulfide bond aggregation was observed. Final aggregate morphology, gel formation and stability are affected as a result of rapid covalent network formation which does not allow rearrangement into more stable networks, as illustrated by lower gel Young's moduli obtained upon thiolation (Broersen et al., 2006).

#### **3.4 Charge modification by methylation and succinylation**

Net charge and local charge densities of proteins have been implicated in the regulation of protein stability, aggregation, and aggregate morphology affecting the visual appearance of food products. These hypotheses have been substantiated by a range of observations which involved charge introduction, removal or reversal through succinylation and methylation reactions (Broersen et al., 2007a; Weijers et al., 2008). The reactions of succinylation and acetylation both lead to blockage of the reactive amino groups of proteins with an acyl residue and are hence collectively termed acylation reactions. The rate of acylation reaction depends on the rate of nucleophilic attack. Succinylation leads to increased net negative charge by the covalent linkage of succinate anions to the cationic amino groups of a protein thereby converting a cationic group into an anionic residue having implications for the distribution of net charge of a protein. Upon acetylation, ammonium cations are replaced by neutral acetyl groups resulting in electrostatically neutral groups. Large extents of succinylation have been reported to affect the integrity of secondary and tertiary structure of soy protein hydrolysate as shown by intrinsic tryptophan fluorescence and circular dichroism (Achouri & Zhang, 2001). Similar conformational rearrangements have been reported upon succinylation of whey protein isolate (Gruener & Ismond, 1997), bovine serum albumin (Jonas & Weber, 1970), canola protein (Lakkis & Villota, 1992), Faba bean legumin (Schwenke et al., 1998), rapeseed 12S globulin (Gueguen et al., 1990), and winged bean protein (Narayana & Rao, 1991). As a result of co-incubation of soy protein hydrolysate with succinic anhydride, which is a common compound used to succinylate proteins, heterogeneous reaction mixtures were obtained. Next to the aimed amine groups, this method of succinylation also commonly results in *O*-succinylation, involving threonine or serine hydroxyl groups or tyrosine succinylation (Achouri & Zhang, 2001; Chang & Sun, 1978; Schwenke et al., 1998). This latter reaction was found to be reversible upon treatment with hydroxylamine (Habeeb & Atassi, 1969), but, when present, to induce substantial expansion of Faba bean legumin as observed by viscometric studies (Schwenke et al., 1998). It was postulated that the high accumulation of negative charge upon extensive succinylation leads to dissociation and expansion of the individual subunits legumin is composed of (Schwenke et al., 1998). Other functional properties are equally affected by succinylation. For example, protein solubility has been reported to increase upon succinylation as has been demonstrated for rapeseed preparations (Dua et al., 1996), flax protein isolate (Wanasundara & Shahidi, 1997), oat protein isolate (Mirmoghatadaie et al., 2009), and soy protein isolate (Franzen & Kinsella, 1976a). Improved solubility has been related to the ability of proteins to perform more efficiently as stabilizers in emulsions and foams (Nakai & Li-Chan, 1988; Waniska & Kinsella, 1979), which, in turn, is greatly affected by their ability to absorb at the air-water interface (Wierenga et al., 2005). It has indeed been shown that treatment of a variety of proteins with succinic anhydride leads to increased foam capacity (Dua et al., 1996; Franzen & Kinsella, 1976a; Mirmoghatadaie et al., 2009), although others suggest that succinylation leads to decreased foam expansion capacity (Wanasundara & Shahidi, 1997). These seemingly contradictive findings may be explained

Application Potential of Food Protein Modification 151

PEGylation however lead to more stable emulsions and improved emulsion activity index as a result of better stabilization of individual droplets against coalescence by the absorption of PEG onto the surface of the sulfitolyzed -lactoglobulin (Losso & Nakai, 2002). From the limited number of studies available it is not possible to derive direct conclusions on the applicability of PEGylation on the advancement of functionality of proteins in food

Deamidation involves the hydrolysis of the amino acids glutamine and asparagine into glutamic and aspartic acid and is achieved by acid, alkaline, or enzymatic treatment (Liao et al., 2009; Shih, 1990; reviewed in Wright & Urry, 1991). Consequently, deamidated protein is often obtained as a by-product of food processing. For example, the extrusion of wheat flour induces deamidation of wheat proteins (Izzo et al., 1993). Deamidation has been shown to affect protein functionality. Functional properties, such as solubility, emulsifying and foaming properties, of gluten have been reported to improve upon low levels of deamidation brought about by mild acid hydrolysis (Hamada & Marshall, 1989; Matsudomi et al., 1982, 1985). Deamidation also was reported to increase exposed hydrophobicity of gluten induced by a conformational change and, subsequently, to increase surface activity (Matsudomi et al., 1982). A number of studies investigated the molecular mechanism for protein structural destabilization upon deamidation. For example, the deamidation treatment has direct implications for charge density and, in turn, affects electrostatic interactions the protein may undergo by interacting with water or upon self-assembly (Finley, 1975; reviewed in Riha et al., 1996). This role of electrostatics to deamidationinduced disruption of protein structure was supported by further observations on wheat gluten: both acetic acid and HCl induced deamidation had substantial consequences for the secondary structure of wheat gluten. It was thus postulated that strong deamidation induced protein unfolding as a result of electrostatic repulsion (Liao et al., 2010). Acetic acid induced deamidation of wheat gluten was further found to inhibit SDS-stable aggregate formation whilst largely retaining its ability to form disulfide bonds (Liao et al., 2010). The rate of the deamidation reaction has been found to depend on primary sequence and pH under which the reaction takes place, but was independent of ionic strength for model peptides (Patel & Borchardt, 1990a, 1990b; Robinson & Rudd, 1974; Tyler-Cross & Schirch,

A variety of aldehydes, including gluteraldehyde, formaldehyde, and hydroxyadipaldehyde, have been used to induce chemical cross-linking of proteins. Between these reagents, gluteraldehyde was found to cross-link bovine serum albumin most efficiently forming large insoluble networks (Hopwood, 1969). Also for other proteins gluteraldehyde has been reported as the most efficient cross-linking agent yielding thermally and chemically stable cross-links (Bowes & Cater, 1968; Nimni et al., 1987). Primary amino groups have been reported to act as prime target to initiate the aldehydeinduced cross-linking reaction (Quiocho & Richards, 1966), next to aromatic amino acids (Hopwood et al., 1970). Depending on environmental conditions, gluteraldehyde can bring about cross-linking through a wide variety of reaction mechanisms (reviewed in Migneault

1991), soy protein and egg white lysozyme (Zhang et al., 1993).

products.

**3.6 Deamidation** 

**3.7 Cross-linking** 

by the demonstration by Wierenga and colleagues (2005) that the likelihood of a protein molecule to adsorb at an interface is the result of a balance between hydrophobic and steric effects: highly charged molecules may be adsorbing to the interface as a result of hydrophobic interaction, but the density at which protein molecules continue adsorbing to the interface is mainly determined by the repulsive nature of the charged proteins. Other effects observed upon charge modification of proteins are related to emulsification properties: methylation increased while succinylation decreased the emulsifying activity of rapeseed preparations (Dua et al., 1996). Other studies show an increased emulsifying activity and stability upon succinylation of soy protein (Franzen & Kinsella, 1976a), and oat protein isolate (Mirmoghatadaie et al., 2009). In terms of gelation, an increase in net charge lead to more transparent gels upon gelation of ovalbumin which is related to the morphology of the aggregated network making up the gel structure (Weijers et al., 2008). Overall, many and detailed efforts have been made employing net charge modification of proteins in the field of food science. These studies have lead to in-depth knowledge of the role of electrostatics to common protein functionalities such as emulsification, foaming, and aggregation propensity.

#### **3.5 PEGylation**

The covalent attachment of a polyethylene glycol (PEG) polymer chain to a protein, also termed 'PEGylation' is mostly applied in the field of pharmaceutics as the conjugation of non-toxic PEG imparts substantial advantages to support drug delivery (reviewed in Damodaran & Fee, 2010; DeSantis & Jones, 1999; Francis et al., 1998). The protein targets for PEG modification are regarded as non-specific and include the -amino groups of lysine and other nucleophilic groups such as glutamic acid, aspartic acid, threonine, serine or tyrosine on the surface of the protein resulting in highly heterogeneous protein-PEG conjugates upon modification (Losso & Nakai, 2002). Commercially available PEG is available as mixtures of different oligomer sizes in various molecular weight ranges enabling the variation of exposed hydrophobicity of proteins. Conjugation of hydrophilic PEG to a hydrophobic protein generally results in an increase in hydrodynamic size and water solubility (Damodaran & Fee, 2010). From a pharmaceutical viewpoint, PEGylation has been reported to enhance circulation life of bovine liver catalase in the blood of mice while the presence of PEG does not induce an immune response upon injection (Abuchowski et al., 1977). Modification of peroxidase from turnip was shown to enhance catalytic activity of the enzyme with increased stability in organic solvents as well as increased temperature resistance (Quintanilla-Guerrero et al., 2008). Similar results were found upon PEGylation of trichosanthin, which showed prolonged plasma half-life and reduced immunogenicity (He et al., 1999). PEGylation of lysozyme similarly lead to stabilization of the protein against pH and temperature variation as well as resistance against proteolysis (Silva Freitas & Abrahão-Neto, 2010). Most of these studies have been carried out in the context of pharmaceutical application. The study of PEGylation as a potential route to bring about modification of physicochemical parameters of proteins applied in food products has been less well explored. The only known effort in the field of food science combined oxidative sulfitolysis with conjugation of 5000 dalton activated PEG to investigate the impact on the emulsifying properties of -lactoglobulin A (Losso & Nakai, 2002). PEG molecules were found to cover the entire surface of unmodified and sulfitolyzed -lactoglobulin and PEGylation alone did not improve emulsifying activity or emulsion stability. The combination of sulfitolysis and PEGylation however lead to more stable emulsions and improved emulsion activity index as a result of better stabilization of individual droplets against coalescence by the absorption of PEG onto the surface of the sulfitolyzed -lactoglobulin (Losso & Nakai, 2002). From the limited number of studies available it is not possible to derive direct conclusions on the applicability of PEGylation on the advancement of functionality of proteins in food products.

#### **3.6 Deamidation**

150 Advances in Chemical Engineering

by the demonstration by Wierenga and colleagues (2005) that the likelihood of a protein molecule to adsorb at an interface is the result of a balance between hydrophobic and steric effects: highly charged molecules may be adsorbing to the interface as a result of hydrophobic interaction, but the density at which protein molecules continue adsorbing to the interface is mainly determined by the repulsive nature of the charged proteins. Other effects observed upon charge modification of proteins are related to emulsification properties: methylation increased while succinylation decreased the emulsifying activity of rapeseed preparations (Dua et al., 1996). Other studies show an increased emulsifying activity and stability upon succinylation of soy protein (Franzen & Kinsella, 1976a), and oat protein isolate (Mirmoghatadaie et al., 2009). In terms of gelation, an increase in net charge lead to more transparent gels upon gelation of ovalbumin which is related to the morphology of the aggregated network making up the gel structure (Weijers et al., 2008). Overall, many and detailed efforts have been made employing net charge modification of proteins in the field of food science. These studies have lead to in-depth knowledge of the role of electrostatics to common protein functionalities such as emulsification, foaming, and

The covalent attachment of a polyethylene glycol (PEG) polymer chain to a protein, also termed 'PEGylation' is mostly applied in the field of pharmaceutics as the conjugation of non-toxic PEG imparts substantial advantages to support drug delivery (reviewed in Damodaran & Fee, 2010; DeSantis & Jones, 1999; Francis et al., 1998). The protein targets for PEG modification are regarded as non-specific and include the -amino groups of lysine and other nucleophilic groups such as glutamic acid, aspartic acid, threonine, serine or tyrosine on the surface of the protein resulting in highly heterogeneous protein-PEG conjugates upon modification (Losso & Nakai, 2002). Commercially available PEG is available as mixtures of different oligomer sizes in various molecular weight ranges enabling the variation of exposed hydrophobicity of proteins. Conjugation of hydrophilic PEG to a hydrophobic protein generally results in an increase in hydrodynamic size and water solubility (Damodaran & Fee, 2010). From a pharmaceutical viewpoint, PEGylation has been reported to enhance circulation life of bovine liver catalase in the blood of mice while the presence of PEG does not induce an immune response upon injection (Abuchowski et al., 1977). Modification of peroxidase from turnip was shown to enhance catalytic activity of the enzyme with increased stability in organic solvents as well as increased temperature resistance (Quintanilla-Guerrero et al., 2008). Similar results were found upon PEGylation of trichosanthin, which showed prolonged plasma half-life and reduced immunogenicity (He et al., 1999). PEGylation of lysozyme similarly lead to stabilization of the protein against pH and temperature variation as well as resistance against proteolysis (Silva Freitas & Abrahão-Neto, 2010). Most of these studies have been carried out in the context of pharmaceutical application. The study of PEGylation as a potential route to bring about modification of physicochemical parameters of proteins applied in food products has been less well explored. The only known effort in the field of food science combined oxidative sulfitolysis with conjugation of 5000 dalton activated PEG to investigate the impact on the emulsifying properties of -lactoglobulin A (Losso & Nakai, 2002). PEG molecules were found to cover the entire surface of unmodified and sulfitolyzed -lactoglobulin and PEGylation alone did not improve emulsifying activity or emulsion stability. The combination of sulfitolysis and

aggregation propensity.

**3.5 PEGylation**

Deamidation involves the hydrolysis of the amino acids glutamine and asparagine into glutamic and aspartic acid and is achieved by acid, alkaline, or enzymatic treatment (Liao et al., 2009; Shih, 1990; reviewed in Wright & Urry, 1991). Consequently, deamidated protein is often obtained as a by-product of food processing. For example, the extrusion of wheat flour induces deamidation of wheat proteins (Izzo et al., 1993). Deamidation has been shown to affect protein functionality. Functional properties, such as solubility, emulsifying and foaming properties, of gluten have been reported to improve upon low levels of deamidation brought about by mild acid hydrolysis (Hamada & Marshall, 1989; Matsudomi et al., 1982, 1985). Deamidation also was reported to increase exposed hydrophobicity of gluten induced by a conformational change and, subsequently, to increase surface activity (Matsudomi et al., 1982). A number of studies investigated the molecular mechanism for protein structural destabilization upon deamidation. For example, the deamidation treatment has direct implications for charge density and, in turn, affects electrostatic interactions the protein may undergo by interacting with water or upon self-assembly (Finley, 1975; reviewed in Riha et al., 1996). This role of electrostatics to deamidationinduced disruption of protein structure was supported by further observations on wheat gluten: both acetic acid and HCl induced deamidation had substantial consequences for the secondary structure of wheat gluten. It was thus postulated that strong deamidation induced protein unfolding as a result of electrostatic repulsion (Liao et al., 2010). Acetic acid induced deamidation of wheat gluten was further found to inhibit SDS-stable aggregate formation whilst largely retaining its ability to form disulfide bonds (Liao et al., 2010). The rate of the deamidation reaction has been found to depend on primary sequence and pH under which the reaction takes place, but was independent of ionic strength for model peptides (Patel & Borchardt, 1990a, 1990b; Robinson & Rudd, 1974; Tyler-Cross & Schirch, 1991), soy protein and egg white lysozyme (Zhang et al., 1993).

#### **3.7 Cross-linking**

A variety of aldehydes, including gluteraldehyde, formaldehyde, and hydroxyadipaldehyde, have been used to induce chemical cross-linking of proteins. Between these reagents, gluteraldehyde was found to cross-link bovine serum albumin most efficiently forming large insoluble networks (Hopwood, 1969). Also for other proteins gluteraldehyde has been reported as the most efficient cross-linking agent yielding thermally and chemically stable cross-links (Bowes & Cater, 1968; Nimni et al., 1987). Primary amino groups have been reported to act as prime target to initiate the aldehydeinduced cross-linking reaction (Quiocho & Richards, 1966), next to aromatic amino acids (Hopwood et al., 1970). Depending on environmental conditions, gluteraldehyde can bring about cross-linking through a wide variety of reaction mechanisms (reviewed in Migneault

Application Potential of Food Protein Modification 153

hence, includes lipidation of ovalbumin (Kosters et al., 2003; Wierenga et al., 2003), lysozyme (Liu et al., 2000), and gelatin (Aewsiri et al., 2010, 2011a, 2011b). Other modifications which can be quantified using the OPA assay are succinylation (Kosters et al., 2003; Wierenga et al., 2005), methylation (Kosters et al., 2003), glycosylation (Broersen et al.,

The second frequently used method to quantify available amino groups upon modification involving the chemical TNBS was developed by Okuyama and Satake (1960) and Satake and colleagues (1960). This method was first employed to study free amino acid groups in trypsin and chymotrypsin inhibitors (Haynes et al., 1967) and for routine screening of protein concentrates for animal feeds (Hall et al., 1973). The chemical TNBS reacts with high preference to free amino groups resulting in the formation of trinitrophenyl derivatives. The reaction product can be quantified spectrophotometrically at 335 nm. A disadvantage of this method is that TNBS also reacts with free sulfhydryl groups, albeit at a slower rate than with amino groups and to form a labile product (Kotaki et al., 1964). The TNBS assay has been used for various types of protein modification including fatty acid incorporation (Andersson et al., 1971; Ibrahim et al., 1991, 1993), glycosylation of proteins as a result of the Maillard reaction (Sun et al., 2004), succinylation (El-Adawy, 2000; Schwenke et al., 1998; Zhao et al., 2004a, 2004b), and

The compound 2,2-dihydroxyindane-1,3-dione (ninhydrin) reacts with -amino groups and ammonium ions into a blue-purple Schiff base product called Ruhemann's purple, that can be colorimetrically detected at a wavelength of 440 nm (Yemm & Cocking, 1955; Schilling et al., 1963; Samejima et al., 1971). The ninhydrin assay has been used to quantify degrees of succinylation of soy protein (Franzen & Kinsella, 1976a), and acylation of flax protein

Comparison of TNBS, OPA or ninhydrin to determine -amino groups in pea protein isolates and hydrolysates thereof lead to the conclusion that TNBS and OPA produced comparable results while ninhydrin detected only half of the amino groups that were

Two commonly used assays are available to evaluate the successful conjugation or blockage of sulfhydryl groups in proteins. The assay which was developed first by Ellman (1959) involves the reaction of 5,5'-dithiobis(2-nitrobenzoic acid) (DTNB) or Ellman's reagent with free sulfhydryl groups yielding colored 3-carboxylato-4-nitrothiophenolate (CNT). Spectrophotometric absorbance intensity at 412 nm provides for a direct measure of the concentration of CNT in solution and cysteine is commonly used as calibration standard. Ellman's assay has been used before to determine the extent of *N*-ethylmaleimide modification of -lactoglobulin A (Kitabatake et al., 2001; Wada & Kitabatake, 2001), thiolation of ovalbumin (Broersen et al., 2006), or acylation of soy protein sulfhydryl groups (Franzen & Kinsella, 1976a). However, the sizeable DTNB at 400 dalton may not be able to detect sulfhydryl groups which are buried inside the folded structure of intact proteins or aggregated proteins (reviewed by Visschers & de Jongh, 2005). A useful alternative for the Ellman's assay has been developed by Owusu Apenten and colleagues (2003) for -

2004; Kosters et al., 2003), and thiolation of proteins (Broersen et al., 2006).

acetylation of faba bean legumin (Krause et al., 1996).

isolates (Wanasundara & Shahidi, 1997).

**3.8.2 Sulfhydryl groups** 

detected by TNBS and OPA (Panasiuk et al., 1998).

et al., 2004). This is caused by the large number of different molecular structures gluteraldehyde can assume in solution (Hardy et al., 1969; Korn et al., 1972; Richard & Knowles, 1968; reviewed in Migneault et al., 2004) although the mechanistic details for this is unknown. As a result of its reported toxicity, gluteraldehyde-induced cross-linking has not been employed in the field of food technology, other than as a tool to enable the investigation of intermediates in the aggregation pathway or to immobilize proteins onto a surface to allow further investigation.

#### **3.8 Measuring the degree of modification**

To evaluate the effect of a specific type of modification on the functional aspects under study, the success of the chemical engineering process on proteins is evaluated by most researchers. To this end, targeted chemical and biophysical assays have been developed which are now widely used. These quantitative assays are mostly based on the formation of a chromogenic or fluorogenic product upon specific interaction with reactive groups of a protein. Some assays will provide information on the average degree of modification in the entire ensemble of protein molecules in a solution. Examples of such assays are Ellman's reagent (Ellman, 1959) or the sulfhydryl-disulfide exchange (SEI) index (Owusu Apenten et al., 2003). Others are also useful to obtain information on the distribution of the degree of modification obtained, such as mass spectrometry. The type of assays developed can be categorized by the type of aimed conjugated chemical group of the protein they probe, such as amine groups, thiol groups or carboxyl groups. Some other researchers use methods which rather probe for the attached molecule, such as the use of gas liquid chromatography (GLC) to determine the degree of lipid incorporation (Haque et al., 1982).

#### **3.8.1 Amine groups**

Amine groups in proteins originate either from free amino groups of proteins or from Nterminal residues of proteins (Skraup & Kaas, 1906; Chibnall, 1942). From 1906 it was recognized that lysines in proteins were largely responsible for the free amino groups present in proteins (Skraup & Kaas, 1906). This finding triggered the development of various assays to study accessible -amino groups of lysines in proteins (Gurin & Clarke, 1934; Sanger, 1945). Three of the most commonly used assays in the field of protein engineering to determine the number of available amino groups of a chemically modified protein involve the use of chemicals 2,4,6-trinitrobenzenesulphonic acid (TNBS) (Fields, 1971), ninhydrin (Yemm & Cocking, 1955; Schilling et al.; 1963; Samejima et al., 1971), and ortho-phtaldialdehyde (OPA) (Roth, 1971). The latter assay is based on the reaction of the OPA compound with free amino groups in proteins in the presence of -mercaptoethanol under alkaline conditions. This reaction results in the formation of highly fluorescent alkyliso-indole derivatives which emit at a wavelength of 455 nm upon excitation at 340 nm. High concentrations (*i.e.* 10%) of sodium dodecyl sulfate (SDS) are often added to the protein solution to aid the exposure of all amino groups which are sometimes buried in the folded protein. The extinction coefficient for the formed adducts of both - and -amino groups are similar with an absorptivity of 6000 M-1 cm-1. The OPA assay was further developed to evaluate the degree of proteolysis of dairy proteins by determining the number of -amino groups released upon hydrolysis (Church et al., 1983). The assay can also be used to quantify all types of reactions involving the modification of lysine and,

et al., 2004). This is caused by the large number of different molecular structures gluteraldehyde can assume in solution (Hardy et al., 1969; Korn et al., 1972; Richard & Knowles, 1968; reviewed in Migneault et al., 2004) although the mechanistic details for this is unknown. As a result of its reported toxicity, gluteraldehyde-induced cross-linking has not been employed in the field of food technology, other than as a tool to enable the investigation of intermediates in the aggregation pathway or to immobilize proteins onto a

To evaluate the effect of a specific type of modification on the functional aspects under study, the success of the chemical engineering process on proteins is evaluated by most researchers. To this end, targeted chemical and biophysical assays have been developed which are now widely used. These quantitative assays are mostly based on the formation of a chromogenic or fluorogenic product upon specific interaction with reactive groups of a protein. Some assays will provide information on the average degree of modification in the entire ensemble of protein molecules in a solution. Examples of such assays are Ellman's reagent (Ellman, 1959) or the sulfhydryl-disulfide exchange (SEI) index (Owusu Apenten et al., 2003). Others are also useful to obtain information on the distribution of the degree of modification obtained, such as mass spectrometry. The type of assays developed can be categorized by the type of aimed conjugated chemical group of the protein they probe, such as amine groups, thiol groups or carboxyl groups. Some other researchers use methods which rather probe for the attached molecule, such as the use of gas liquid chromatography

Amine groups in proteins originate either from free amino groups of proteins or from Nterminal residues of proteins (Skraup & Kaas, 1906; Chibnall, 1942). From 1906 it was recognized that lysines in proteins were largely responsible for the free amino groups present in proteins (Skraup & Kaas, 1906). This finding triggered the development of various assays to study accessible -amino groups of lysines in proteins (Gurin & Clarke, 1934; Sanger, 1945). Three of the most commonly used assays in the field of protein engineering to determine the number of available amino groups of a chemically modified protein involve the use of chemicals 2,4,6-trinitrobenzenesulphonic acid (TNBS) (Fields, 1971), ninhydrin (Yemm & Cocking, 1955; Schilling et al.; 1963; Samejima et al., 1971), and ortho-phtaldialdehyde (OPA) (Roth, 1971). The latter assay is based on the reaction of the OPA compound with free amino groups in proteins in the presence of -mercaptoethanol under alkaline conditions. This reaction results in the formation of highly fluorescent alkyliso-indole derivatives which emit at a wavelength of 455 nm upon excitation at 340 nm. High concentrations (*i.e.* 10%) of sodium dodecyl sulfate (SDS) are often added to the protein solution to aid the exposure of all amino groups which are sometimes buried in the folded protein. The extinction coefficient for the formed adducts of both - and -amino groups are similar with an absorptivity of 6000 M-1 cm-1. The OPA assay was further developed to evaluate the degree of proteolysis of dairy proteins by determining the number of -amino groups released upon hydrolysis (Church et al., 1983). The assay can also be used to quantify all types of reactions involving the modification of lysine and,

(GLC) to determine the degree of lipid incorporation (Haque et al., 1982).

surface to allow further investigation.

**3.8.1 Amine groups** 

**3.8 Measuring the degree of modification** 

hence, includes lipidation of ovalbumin (Kosters et al., 2003; Wierenga et al., 2003), lysozyme (Liu et al., 2000), and gelatin (Aewsiri et al., 2010, 2011a, 2011b). Other modifications which can be quantified using the OPA assay are succinylation (Kosters et al., 2003; Wierenga et al., 2005), methylation (Kosters et al., 2003), glycosylation (Broersen et al., 2004; Kosters et al., 2003), and thiolation of proteins (Broersen et al., 2006).

The second frequently used method to quantify available amino groups upon modification involving the chemical TNBS was developed by Okuyama and Satake (1960) and Satake and colleagues (1960). This method was first employed to study free amino acid groups in trypsin and chymotrypsin inhibitors (Haynes et al., 1967) and for routine screening of protein concentrates for animal feeds (Hall et al., 1973). The chemical TNBS reacts with high preference to free amino groups resulting in the formation of trinitrophenyl derivatives. The reaction product can be quantified spectrophotometrically at 335 nm. A disadvantage of this method is that TNBS also reacts with free sulfhydryl groups, albeit at a slower rate than with amino groups and to form a labile product (Kotaki et al., 1964). The TNBS assay has been used for various types of protein modification including fatty acid incorporation (Andersson et al., 1971; Ibrahim et al., 1991, 1993), glycosylation of proteins as a result of the Maillard reaction (Sun et al., 2004), succinylation (El-Adawy, 2000; Schwenke et al., 1998; Zhao et al., 2004a, 2004b), and acetylation of faba bean legumin (Krause et al., 1996).

The compound 2,2-dihydroxyindane-1,3-dione (ninhydrin) reacts with -amino groups and ammonium ions into a blue-purple Schiff base product called Ruhemann's purple, that can be colorimetrically detected at a wavelength of 440 nm (Yemm & Cocking, 1955; Schilling et al., 1963; Samejima et al., 1971). The ninhydrin assay has been used to quantify degrees of succinylation of soy protein (Franzen & Kinsella, 1976a), and acylation of flax protein isolates (Wanasundara & Shahidi, 1997).

Comparison of TNBS, OPA or ninhydrin to determine -amino groups in pea protein isolates and hydrolysates thereof lead to the conclusion that TNBS and OPA produced comparable results while ninhydrin detected only half of the amino groups that were detected by TNBS and OPA (Panasiuk et al., 1998).

#### **3.8.2 Sulfhydryl groups**

Two commonly used assays are available to evaluate the successful conjugation or blockage of sulfhydryl groups in proteins. The assay which was developed first by Ellman (1959) involves the reaction of 5,5'-dithiobis(2-nitrobenzoic acid) (DTNB) or Ellman's reagent with free sulfhydryl groups yielding colored 3-carboxylato-4-nitrothiophenolate (CNT). Spectrophotometric absorbance intensity at 412 nm provides for a direct measure of the concentration of CNT in solution and cysteine is commonly used as calibration standard. Ellman's assay has been used before to determine the extent of *N*-ethylmaleimide modification of -lactoglobulin A (Kitabatake et al., 2001; Wada & Kitabatake, 2001), thiolation of ovalbumin (Broersen et al., 2006), or acylation of soy protein sulfhydryl groups (Franzen & Kinsella, 1976a). However, the sizeable DTNB at 400 dalton may not be able to detect sulfhydryl groups which are buried inside the folded structure of intact proteins or aggregated proteins (reviewed by Visschers & de Jongh, 2005). A useful alternative for the Ellman's assay has been developed by Owusu Apenten and colleagues (2003) for -

Application Potential of Food Protein Modification 155

Chromatography has been further explored in the shape of cation exchange chromatography to validate the degree of methylation of ovalbumin which was found to provide comparable read-outs as the revised version of Woodward's reagent K method to determine carboxylic acid groups (Kosters & de Jongh, 2003). All these techniques provide insight in the ensemble average degree of modification. Chemical engineering inherently implies the rise of heterogeneous species of proteins. Mass spectrometry, often employed as Matrix-assisted laser desorption/ionization-time of flight (MALDI-tof) mass spectrometry, has proven a powerful method to specifically obtain insight into the distribution of the modification reactions. This method has been employed to derive information on modification distributions of *N*-ethylmaleimide modified -lactoglobulin A (Wada & Kitabatake, 2001), and glycosylation of -lactoglobulin (Broersen et al., 2004; van Teeffelen et al., 2005). These studies demonstrated that degrees of modification obtained upon chemical engineering of proteins are rather broad and show a Gauss distribution profile rather than a single well-defined modification degree (Broersen et al., 2004; van Teeffelen et al., 2005). Some attempts have been made to isolate modified protein fractions with more defined degrees of modification, for example by using ion exchange chromatography of succinylated

From the above it is clear that in the past decades a lot of effort has been spend on better understanding and controlling protein behavior and protein-based microstructure formation by making use of chemical engineering approaches. But how much impact have these insights had on the development of new food applications? There are a number of well-known product categories where engineered protein functionality has led to improved product properties. In the early seventies Unilever produced new lines of margarines that showed better performance in aspects like spreadability, prolonged storage stability and during baking, caused by acetylation of milk proteins leading to better fat emulsification (Evans & Irons, 1970). Also for mayonnaises and salad dressings modification of egg yolk proteins (via *N*-succinylation) provided improved product quality (Evans & Irons 1970). Another example of an application of improved protein functionality is that of the use of succinylation to improve the solubility/dispersability properties of soy proteins in the extraction and refinery process (Melnychyn & Stapley, 1973). It is interesting to evaluate via a patent-literature screening how frequent the wide variety of technological possibilities as described in section 3 to better control protein behavior have led to unique market-

**4.1 Current application of protein modifications in food/feed-related products** 

Figure 1 illustrates a landscape representation of patents (worldwide), filed in the last decade, in the area of food and feed where protein modification has played a crucial role in deriving a new type of product functionality. The height of the contour indicates the activity in that particular area. The distance between patents reflects their commonality. In total only 445 relevant patents (grouped in 157 families) can be found. In comparison, a search on any biobased-product (including products with protein-based technical polymers as in coatings, paints, paper, etc.) revealed more than 8600 patents. Clearly the role of protein engineering to derive new food product specifications are very limited, especially in view of the

ovalbumin (Wierenga et al., 2005).

propositions.

**4. Application potential of food protein modification** 

lactoglobulin and bovine serum albumin. Chemical reactivity of thiol groups is an absolute requirement to enable covalent cross-linking through disulfide bond formation (Hillier et al., 1980). This assay, called the sulfhydryl-disulfide exchange (SEI) index, therefore provides a direct measure of the chemical reactivity of thiol groups as it determines the conversion of substrate in time and kinetically relates the conversion to that of fully exposed thiol groups (Owusu Apenten et al., 2003). Chemical activity of introduced sulfhydryl groups by modification of ovalbumin using the SEI index has been verified for example by Wierenga and colleagues (2006).

#### **3.8.3 Carboxylic acid groups**

The carboxylic acid content of proteins is primarily investigated using the compound 2 ethyl-5-phenylisoxazolium-3'-sulfonate or Woodward's reagent K (Woodward & Olofson, 1961; Woodward et al., 1961; Sinha & Brewer, 1985). The activity of Woodward's reagent K is the result of a multistep process. First, Woodward's reagent K is converted into ketoketenimine at neutral pH. The intermediate compound is then either further disintegrated to form ketoamide or interacts with carboxylic acid groups of a protein. The latter interaction results in the formation of an enol ester (Pétra, 1971) which absorbs at 340 nm with a molar extinction coefficient of 7000 M-1 cm-1 (Sinha & Brewer, 1985). At a later stage, Kosters and de Jongh (2003) revised the extinction coefficient of the product to 3150 M-1 cm-1 at 269 nm to improve specificity of the reaction and eliminate the substantial contribution of side reactions with other nucleophiles in proteins (Llamas et al., 1986), and histidine and cysteine (Bustos et al., 1996; Johnson & Dekker, 1996) to the absorbance at 269 nm. This revised version of the assay employing Woodward's reagent K to estimate the number of carboxylic acid groups of chemically modified proteins has been used by a number of researchers. Wierenga and colleagues (2005) used Woodward's reagent K to estimate the degree of modification of succinylated ovalbumin to study the relation between protein net charge and adsorption to air-water interfaces. Similar net charge variation induced by succinylation was used to investigate colloidal versus conformational stability of ovalbumin to aggregation (Broersen et al., 2007a). To investigate the stability of ovalbumin, the protein was modified by succinylation, methylation, glycosylation, and lipophilization and the degrees of modification were validated using Woodward's reagent K and the OPA assay (paragraph 3.8.1).

#### **3.8.4 Conjugated groups**

An alternative route to obtain information on the degree of protein modification is to selectively probe the conjugated group. This can be achieved for example by incorporating an isotopically labeled reagent or inclusion of a chromophore or fluorophore which can then be quantified by read-out of fluorescence intensity using a fluorimeter or simple absorbance measurements using a standard spectrophotometer. Raman spectroscopy was shown to provide direct insight into degrees of succinylation and acetylation of a range of proteins originating from soy, egg white or whey by distinct contributions of the conjugated groups at 1737 cm-1 and 1420 cm-1 (Zhao et al., 2004a, 2004b). The peak intensities at these wavelengths could be directly converted to obtain information on the degree of modification. Degrees of palmitoylation of soybean glycinin (Haque et al., 1982) and s1 casein (Haque & Kito, 1983a) have been determined using gas liquid chromatography.

lactoglobulin and bovine serum albumin. Chemical reactivity of thiol groups is an absolute requirement to enable covalent cross-linking through disulfide bond formation (Hillier et al., 1980). This assay, called the sulfhydryl-disulfide exchange (SEI) index, therefore provides a direct measure of the chemical reactivity of thiol groups as it determines the conversion of substrate in time and kinetically relates the conversion to that of fully exposed thiol groups (Owusu Apenten et al., 2003). Chemical activity of introduced sulfhydryl groups by modification of ovalbumin using the SEI index has been verified for example by Wierenga

The carboxylic acid content of proteins is primarily investigated using the compound 2 ethyl-5-phenylisoxazolium-3'-sulfonate or Woodward's reagent K (Woodward & Olofson, 1961; Woodward et al., 1961; Sinha & Brewer, 1985). The activity of Woodward's reagent K is the result of a multistep process. First, Woodward's reagent K is converted into ketoketenimine at neutral pH. The intermediate compound is then either further disintegrated to form ketoamide or interacts with carboxylic acid groups of a protein. The latter interaction results in the formation of an enol ester (Pétra, 1971) which absorbs at 340 nm with a molar extinction coefficient of 7000 M-1 cm-1 (Sinha & Brewer, 1985). At a later stage, Kosters and de Jongh (2003) revised the extinction coefficient of the product to 3150 M-1 cm-1 at 269 nm to improve specificity of the reaction and eliminate the substantial contribution of side reactions with other nucleophiles in proteins (Llamas et al., 1986), and histidine and cysteine (Bustos et al., 1996; Johnson & Dekker, 1996) to the absorbance at 269 nm. This revised version of the assay employing Woodward's reagent K to estimate the number of carboxylic acid groups of chemically modified proteins has been used by a number of researchers. Wierenga and colleagues (2005) used Woodward's reagent K to estimate the degree of modification of succinylated ovalbumin to study the relation between protein net charge and adsorption to air-water interfaces. Similar net charge variation induced by succinylation was used to investigate colloidal versus conformational stability of ovalbumin to aggregation (Broersen et al., 2007a). To investigate the stability of ovalbumin, the protein was modified by succinylation, methylation, glycosylation, and lipophilization and the degrees of modification were validated using Woodward's reagent K and the OPA

An alternative route to obtain information on the degree of protein modification is to selectively probe the conjugated group. This can be achieved for example by incorporating an isotopically labeled reagent or inclusion of a chromophore or fluorophore which can then be quantified by read-out of fluorescence intensity using a fluorimeter or simple absorbance measurements using a standard spectrophotometer. Raman spectroscopy was shown to provide direct insight into degrees of succinylation and acetylation of a range of proteins originating from soy, egg white or whey by distinct contributions of the conjugated groups at 1737 cm-1 and 1420 cm-1 (Zhao et al., 2004a, 2004b). The peak intensities at these wavelengths could be directly converted to obtain information on the degree of modification. Degrees of palmitoylation of soybean glycinin (Haque et al., 1982) and s1 casein (Haque & Kito, 1983a) have been determined using gas liquid chromatography.

and colleagues (2006).

assay (paragraph 3.8.1).

**3.8.4 Conjugated groups** 

**3.8.3 Carboxylic acid groups** 

Chromatography has been further explored in the shape of cation exchange chromatography to validate the degree of methylation of ovalbumin which was found to provide comparable read-outs as the revised version of Woodward's reagent K method to determine carboxylic acid groups (Kosters & de Jongh, 2003). All these techniques provide insight in the ensemble average degree of modification. Chemical engineering inherently implies the rise of heterogeneous species of proteins. Mass spectrometry, often employed as Matrix-assisted laser desorption/ionization-time of flight (MALDI-tof) mass spectrometry, has proven a powerful method to specifically obtain insight into the distribution of the modification reactions. This method has been employed to derive information on modification distributions of *N*-ethylmaleimide modified -lactoglobulin A (Wada & Kitabatake, 2001), and glycosylation of -lactoglobulin (Broersen et al., 2004; van Teeffelen et al., 2005). These studies demonstrated that degrees of modification obtained upon chemical engineering of proteins are rather broad and show a Gauss distribution profile rather than a single well-defined modification degree (Broersen et al., 2004; van Teeffelen et al., 2005). Some attempts have been made to isolate modified protein fractions with more defined degrees of modification, for example by using ion exchange chromatography of succinylated ovalbumin (Wierenga et al., 2005).

#### **4. Application potential of food protein modification**

From the above it is clear that in the past decades a lot of effort has been spend on better understanding and controlling protein behavior and protein-based microstructure formation by making use of chemical engineering approaches. But how much impact have these insights had on the development of new food applications? There are a number of well-known product categories where engineered protein functionality has led to improved product properties. In the early seventies Unilever produced new lines of margarines that showed better performance in aspects like spreadability, prolonged storage stability and during baking, caused by acetylation of milk proteins leading to better fat emulsification (Evans & Irons, 1970). Also for mayonnaises and salad dressings modification of egg yolk proteins (via *N*-succinylation) provided improved product quality (Evans & Irons 1970). Another example of an application of improved protein functionality is that of the use of succinylation to improve the solubility/dispersability properties of soy proteins in the extraction and refinery process (Melnychyn & Stapley, 1973). It is interesting to evaluate via a patent-literature screening how frequent the wide variety of technological possibilities as described in section 3 to better control protein behavior have led to unique marketpropositions.

#### **4.1 Current application of protein modifications in food/feed-related products**

Figure 1 illustrates a landscape representation of patents (worldwide), filed in the last decade, in the area of food and feed where protein modification has played a crucial role in deriving a new type of product functionality. The height of the contour indicates the activity in that particular area. The distance between patents reflects their commonality. In total only 445 relevant patents (grouped in 157 families) can be found. In comparison, a search on any biobased-product (including products with protein-based technical polymers as in coatings, paints, paper, etc.) revealed more than 8600 patents. Clearly the role of protein engineering to derive new food product specifications are very limited, especially in view of the

Application Potential of Food Protein Modification 157

**2**

**1**

**5**

Fig. 1. Landscape representation of patent (families) filed between 2000 and 2011 where

Whereas in paragraph 3 it was demonstrated how active and progressing the understanding of protein functionality in complex systems has been in the past decades, it is striking to see how minor the contributions to new applications these insights apparently have been. To understand this better one needs to take into account that the technological developments coincided with an increasing level of legislation. In the United States this is embodied in the Food and Drug Administration (FDA). The European Food Safety Authority (EFSA) is its European counterpart. The FDA is responsible for protecting and promoting public health through the regulation and supervision of food safety. It does so by formulating acts that set the boundaries for implementation of new ingredients, processes or compositions related to foods. There are a few acts that have had a strong impact on the food sector. In 1990 the Nutrition Labeling and Education Act was launched. This required food products to be labeled in terms of composition, allowing traceability of its ingredients to their source. This act also amended that all nutrient content claims (e.g. 'high fiber', 'low fat', etc.) would meet the standards set by the FDA. In the end it meant that every engineered protein would require a new label and would need to be recognized and approved first by the FDA.

engineering/modification of protein functionality has been used to derive new material/product characteristics. The small black dots indicate the position of a patentfamily. The numbered circles represent the areas with the highest activity in patents. The

**4 3**

numbers are explained in the body text.

**4.2 Food legislation** 

potential commercial impact. The observation that the patent-families found are spread rather constant over the plot illustrates that these patents are not directly linked to each other in terms of engineering approach or application area. When evaluating the patent filings of the last ten years five areas can be distinguished with a relatively high patenting activity and these will be discussed below in more detail. These are numbered 1 to 5 in the figure.


potential commercial impact. The observation that the patent-families found are spread rather constant over the plot illustrates that these patents are not directly linked to each other in terms of engineering approach or application area. When evaluating the patent filings of the last ten years five areas can be distinguished with a relatively high patenting activity and these will be discussed below in more detail. These are numbered 1 to 5 in the

1. *Refinery of seed storage proteins (12 patents).* There is an increasing interest to use readily abundant and relatively cheap seed storage proteins as nutritional component in food and feed. The major difficulty of this protein-source is to obtain functional proteins after the refinery steps. Loss of functionality occurs especially when the protein is used as powdered ingredient, because of difficulties in resolubilization and unpredictable caking of the powder during storage/transport. A number of patents have been filed that use mild engineering tools, like Maillardation, or enzymetreatments to preserve functional proteins during refinery steps (e.g. patent EP1370157B1: "Highly soluble, high molecular weight soy protein"). Interestingly, all patents filed in this area pay attention to the in principle reversibility of the

2. *Nano-particles (23 patents).* To better direct properties of protein-based nano-particles enzymatic introduction of lipidic groups like small fatty acids or PEGylation has been employed to encapsulate bioactives, typically as microemulsions (e.g. patent US20070154907A1: "Microemulsions as precursors to solid nanoparticles"). These applications are considered food-grade or have passed medical-ethical approval in their testing/and or usage. There are no indications that these application have found their

3. *Nutritional availability in feeds (13 patents).* A considerable number of patents can be found in the application area to increase the nutritional availability of amino acids in feed (14 patents). Most patent-positions are dealing with destabilizing proteinstructures to promote their digestibility and their (proposed) nutritional uptake. Typical patented approaches are de-amidation and Maillardation (e.g. patent WO2004020977A3: "De-amidation to promote nutritional value of rumen in feed"). No patents can be found that link nutritional value and protein modification in foods. 4. *Protein-based emulsion-stabilizers (two times 7 patents).* A few patents are found where specific protein modification is used to improve emulsification of (oil in water) products. Especially enzymatic glycosylation and lipidation approaches have been used (e.g. patent US7126042B1: "Recombinant oleosins from cacao and their use as flavoring or emulsifying agents"). In view of the small scale examples provided in these patents, it is not likely that these inventions have been implemented in a

5. *Edible coatings (12 patents).* A whole family of patents is present on enzymatic protein cross-linking (mainly by transglutaminase) in relation to the production of edible coatings. Modifications typically act on a microstructural level, and not so much on the protein molecular level, to strengthen the spatial network formed (e.g. patent EP963704B1: "Food containing proteinaceous material treated with a transglutaminase and an oxidoreductase"). As these interventions in a product do not necessarily need to be labeled on the product, it is difficult to evaluate whether they have resulted in

figure.

modification applied.

commercial product.

product development.

way into the food/feed product market.

Fig. 1. Landscape representation of patent (families) filed between 2000 and 2011 where engineering/modification of protein functionality has been used to derive new material/product characteristics. The small black dots indicate the position of a patentfamily. The numbered circles represent the areas with the highest activity in patents. The numbers are explained in the body text.

#### **4.2 Food legislation**

Whereas in paragraph 3 it was demonstrated how active and progressing the understanding of protein functionality in complex systems has been in the past decades, it is striking to see how minor the contributions to new applications these insights apparently have been. To understand this better one needs to take into account that the technological developments coincided with an increasing level of legislation. In the United States this is embodied in the Food and Drug Administration (FDA). The European Food Safety Authority (EFSA) is its European counterpart. The FDA is responsible for protecting and promoting public health through the regulation and supervision of food safety. It does so by formulating acts that set the boundaries for implementation of new ingredients, processes or compositions related to foods. There are a few acts that have had a strong impact on the food sector. In 1990 the Nutrition Labeling and Education Act was launched. This required food products to be labeled in terms of composition, allowing traceability of its ingredients to their source. This act also amended that all nutrient content claims (e.g. 'high fiber', 'low fat', etc.) would meet the standards set by the FDA. In the end it meant that every engineered protein would require a new label and would need to be recognized and approved first by the FDA.

Application Potential of Food Protein Modification 159

the EU there is an increasing attention to social inequality issues. From a governmental regulatory point of view there is increasing support for open innovations and shared responsibilities to produce products of good quality that are *affordable* to all social classes. It is unclear what the role of optimized protein behavior by engineering approaches can be. Consumers are susceptible to additional *health*-aspects when safe and acceptable good quality products are available for reasonable prices. In this information technology era consumers are capable to evaluate the added value of products far better, setting higher demands for health marks of food products. The use of added value to products, like bioactives, has not led to major winners in the food sector yet. The desire (and need) to innovate in food product developments could be fed by the wealth of information on how product functionality relates to microstructural morphologies and how these in turn are dictated by molecular properties of specific proteins. On the other hand there are robust approaches in the field of protein engineering that allows us to direct protein behaviour and their propensities to (self)assemble into spatial networks. Still, one has to recognize that these two aspects have not come together yet to contribute to a more sustainable production

**Safe/ Traceable** 

**Acceptable**

**Affordable**

**Nutritional/ Health** 

While genetic modification of proteins generally results in a homogeneous product, chemical engineering approaches to modify proteins are well-known to result in heterogeneity. A good example is the application of the Maillard reaction to -lactoglobulin (Broersen et al., 2004) or fish parvalbumin (de Jongh et al., 2011). On average one may find that for example 6 of the 12 available lysine residues has become glycosylated, where mass analysis will demonstrate that a significant population of protein molecules is present with 5 or 7 sugar groups, some molecules may contain 4 or 8 groups and even traces of molecules with 3 and 9 groups are

and use of proteins in foods. This will be further discussed in section 5.

**4.4 Considerations of chemical modification of proteins** 

Fig. 2. The Pyramid of Food Innovation

Exceptions were those modified ingredients that could be considered as 'occurring from a natural process' or that were 'reversibly modified' (so temporary). The FDA Modernization Act of 1997 was designed to reduce the time for the approval of new pharmaceutical drugs, but also had an impact on food technology by the acknowledgment of the advancement of technological, trade, and public health complexities. Basically, a 'new' food ingredient needed to be seen and evaluated in the complex role it had considering its production up to its digestion in the food. This act was further refined by the Amendments Act of 2007, leading to much sharper defined criteria in what was considered as safe in food products. Recently, in 2010, The Food Safety Modernization Act was signed. Sections of this act require food producers to enable tracking and tracing of all ingredients used. The use of engineered proteins (either genetically or chemically) requires a separate approval for market-clearance. Summarizing, the FDA (and EFSA) have acknowledged that (future) food production requires innovations at the ingredient level, but also via processing routes, and they are in principle open to protein engineering routes. At the same time it enforces that functionally improved ingredients are checked along the full chain from refinery to nutritional value and human health within the complexity of the product.

#### **4.3 Potentials in the area of food product sustainability**

With increasing world-population and welfare the demand for protein as food-nutritional component is rising sharply. Also the identification of proteins as building blocks in nonfood applications in view of a more sustainable economy, has led to increasing pressure on innovations in production, refinery, and application of proteins from wider sources than in the current economy is provided. For the development of a vision on food quality and especially the role of nutritional impact needs to be seen in the context of the basic requirements set not only by consumer demands, but also by participating industries. This can be presented by a so-called Pyramid of Food Innovation, as shown in figure 2. On top one finds the foods that need to be developed in a most sustainable way; to achieve this prospect one needs to comply with lower levels of restrictions, limitations and concomitant scientific challenges.

Food *safety* forms the most fundamental aspect here. Terminology like for example 'nanotechnology' cannot count on consumer acceptance and also the inclusion of genetically modified ingredients gives rise within Europe to hesitance in applicability by foodproducing companies. Especially for traceability and chemical characterization the demands become exceedingly higher. This is the level though where new protein engineering routes could be contributing most. A major bottleneck today in reformulation-strategies is the occurrence of (sensory) differentiations relative to the original product. Exceptions are reduced sugar or salt products, but from a marketing-technical perspective moderations of structuring components are preferably performed within the frame as novel food. A continuously on-going drive to elucidate structure-texture-product acceptance relations is essential, just like innovations in the area of optimized processing tools to deliver products that are *acceptable* with retained food safety. Typically the efforts on applying protein modifications are focusing in deriving food structures more efficiently or to provide new functional building blocks to create new food structures. In an economically global society there is a strong pressure on both ingredient-prices and commercial acceptable processing. Industrial entrepreneurship is essential and the availability of second-line ingredients for products hampers the implementation of new food production strategies. Especially within

Exceptions were those modified ingredients that could be considered as 'occurring from a natural process' or that were 'reversibly modified' (so temporary). The FDA Modernization Act of 1997 was designed to reduce the time for the approval of new pharmaceutical drugs, but also had an impact on food technology by the acknowledgment of the advancement of technological, trade, and public health complexities. Basically, a 'new' food ingredient needed to be seen and evaluated in the complex role it had considering its production up to its digestion in the food. This act was further refined by the Amendments Act of 2007, leading to much sharper defined criteria in what was considered as safe in food products. Recently, in 2010, The Food Safety Modernization Act was signed. Sections of this act require food producers to enable tracking and tracing of all ingredients used. The use of engineered proteins (either genetically or chemically) requires a separate approval for market-clearance. Summarizing, the FDA (and EFSA) have acknowledged that (future) food production requires innovations at the ingredient level, but also via processing routes, and they are in principle open to protein engineering routes. At the same time it enforces that functionally improved ingredients are checked along the full chain from refinery to

With increasing world-population and welfare the demand for protein as food-nutritional component is rising sharply. Also the identification of proteins as building blocks in nonfood applications in view of a more sustainable economy, has led to increasing pressure on innovations in production, refinery, and application of proteins from wider sources than in the current economy is provided. For the development of a vision on food quality and especially the role of nutritional impact needs to be seen in the context of the basic requirements set not only by consumer demands, but also by participating industries. This can be presented by a so-called Pyramid of Food Innovation, as shown in figure 2. On top one finds the foods that need to be developed in a most sustainable way; to achieve this prospect one needs to comply with lower levels of restrictions, limitations and concomitant

Food *safety* forms the most fundamental aspect here. Terminology like for example 'nanotechnology' cannot count on consumer acceptance and also the inclusion of genetically modified ingredients gives rise within Europe to hesitance in applicability by foodproducing companies. Especially for traceability and chemical characterization the demands become exceedingly higher. This is the level though where new protein engineering routes could be contributing most. A major bottleneck today in reformulation-strategies is the occurrence of (sensory) differentiations relative to the original product. Exceptions are reduced sugar or salt products, but from a marketing-technical perspective moderations of structuring components are preferably performed within the frame as novel food. A continuously on-going drive to elucidate structure-texture-product acceptance relations is essential, just like innovations in the area of optimized processing tools to deliver products that are *acceptable* with retained food safety. Typically the efforts on applying protein modifications are focusing in deriving food structures more efficiently or to provide new functional building blocks to create new food structures. In an economically global society there is a strong pressure on both ingredient-prices and commercial acceptable processing. Industrial entrepreneurship is essential and the availability of second-line ingredients for products hampers the implementation of new food production strategies. Especially within

nutritional value and human health within the complexity of the product.

**4.3 Potentials in the area of food product sustainability** 

scientific challenges.

Fig. 2. The Pyramid of Food Innovation

the EU there is an increasing attention to social inequality issues. From a governmental regulatory point of view there is increasing support for open innovations and shared responsibilities to produce products of good quality that are *affordable* to all social classes. It is unclear what the role of optimized protein behavior by engineering approaches can be. Consumers are susceptible to additional *health*-aspects when safe and acceptable good quality products are available for reasonable prices. In this information technology era consumers are capable to evaluate the added value of products far better, setting higher demands for health marks of food products. The use of added value to products, like bioactives, has not led to major winners in the food sector yet. The desire (and need) to innovate in food product developments could be fed by the wealth of information on how product functionality relates to microstructural morphologies and how these in turn are dictated by molecular properties of specific proteins. On the other hand there are robust approaches in the field of protein engineering that allows us to direct protein behaviour and their propensities to (self)assemble into spatial networks. Still, one has to recognize that these two aspects have not come together yet to contribute to a more sustainable production and use of proteins in foods. This will be further discussed in section 5.

#### **4.4 Considerations of chemical modification of proteins**

While genetic modification of proteins generally results in a homogeneous product, chemical engineering approaches to modify proteins are well-known to result in heterogeneity. A good example is the application of the Maillard reaction to -lactoglobulin (Broersen et al., 2004) or fish parvalbumin (de Jongh et al., 2011). On average one may find that for example 6 of the 12 available lysine residues has become glycosylated, where mass analysis will demonstrate that a significant population of protein molecules is present with 5 or 7 sugar groups, some molecules may contain 4 or 8 groups and even traces of molecules with 3 and 9 groups are

Application Potential of Food Protein Modification 161

employed to modify the nutritional quality, for example to increase the nutritional value of plant proteins (Liao et al., 2010). However, the main objective of most studies employing protein engineering is to investigate the consequences of the modification procedure for a range of functional properties. Only a small number of studies investigates nutritional aspects of chemical protein engineering which is mostly regarded as a convenient or inconvenient side effect of the modification procedure. Nutritional aspects covered in literature are exclusively based on *in vitro* studies and cover protein digestibility, availability

Anti-nutritional factors are related to reduced protein digestibility and amino acid availability (reviewed by Gilani et al., 2005; Salunkhe et al., 1982) and are commonly present in large concentrations in plant products (Kay, 1979; Liener, 1980)*.* One of the studies in this field reports on the nutritional quality of mung bean isolate following the exposure to varying concentrations of acetic or succinic anhydride to induce acylation (El-Adawy, 2000). The concentrations of anti-nutritional factors tannin, phytic acid, and trypsin inhibitor showed a significant loss with increasing degrees of modification suggesting that this type of modification can positively impact the effect of anti-nutritional factors. The concentration of trypsin inhibitor even decreased with 70% of the original level of trypsin inhibitor in unmodified protein. The introduction of bulky and/or negatively charged side groups was postulated to affect the extent of protein-tannin (El-Adawy, 2000), protein-mineral-phytic acid (Dua et al., 1996; El-Adawy, 2000), or protein-phenol (Loomis (1974) interactions. Loomis (1974) further showed that the flour and protein production processes provide for optimal conditions for the conversion of polyphenols into quinone oxidation products which, in turn, may bind covalently with sulfhydryl groups of cysteine and -amino groups of lysine and N-terminal amino groups. Further support was provided by Dua and colleagues (1996) who showed that acylation and methylation of rapeseed meal and its water-soluble fraction resulted in loss of anti-nutritional factors polyphenol, glucosinolates and phytic acid. However, the methylation procedure employed by Dua and colleagues (1996) resulted in very limited degrees of modification compared to the succinylation and acetylation process suggesting that other factors than the chemical conjugation itself may play a role in the loss of anti-nutritional factors upon chemical modification. As El-Adawy (2000) comments, the extensive dialysis of the protein following the acylation procedure may well be primarily responsible for the loss of water-soluble anti-nutritional factors upon modification. No studies are known to date that report on the loss of anti-nutritional factors

*In vitro* digestibility of modified proteins is often assayed through exposure of the proteins to a single or a mixture of enzymes including trypsin and pancreatin (Salgó et al., 1984), a combination of trypsin, chymotrypsin and peptidase (Hsu et al., 1977), or pepsin-pancreatin mixtures (Haque et al., 1982) to simulate (post)gastrointestinal digestion of food proteins. The small increase in digestibility of acylated mung bean protein isolate reported by El-Adawy (2000) was primarily correlated to the concurrent loss of tannin; tannins have been shown to play an important role in the reduction of protein digestibility (Barroga et al., 1985). Alternative factors proposed to induce increased digestibility of modified proteins

of essential amino acids or the presence of anti-nutritional factors.

**4.5.2.1 Anti-nutritional factors** 

upon dialysis.

**4.5.2.2** *In vitro* **digestibility** 

present. This is commonly the result of variation in reactivity between the different amino groups present in a specific protein structure: i.e. some amino acids are more exposed and hence are more prone to rapid modification. Buried amino acids do not present a straight forward target for modification as most chemicals are unable to access the folded or aggregated protein structure. Another effect which likely contributes to the extent of heterogeneity is the location of target amino acids in the primary sequence. For example, modification of an amino acid located next to a negatively charged amino acid by succinic anhydride will be hampered as a result of steric hindrance. Some attempts have been made to fractionate protein molecules with different degrees of modification. Five ovalbumin preparations with different degrees of modification have been purified using ion exchange chromatography (Wierenga et al., 2005). The heterogeneous nature of chemically modified proteins has marked implications on the acceptance of these novel ingredients under current food law. Chemical heterogeneity would imply that each obtained modified species would have to be tested separately on toxicity in order to become regarded as safe to use as novel food ingredient.

#### **4.5 Health risks**

In this paragraph we will discuss potential health implications induced by chemically modified proteins used in food products including allergenic response, the presence of antinutritional factors linked to proteins and the development of potential toxic compounds as a side-effect of the modification reaction. Concerns can be raised regarding the impact of protein modification on the potential health risks from products containing modified protein ingredients. Bernstein and colleagues (2003) identified three possible modes for novel food ingredients to result in adverse health effects. These include toxicity, impaired nutrition and food allergy.

#### **4.5.1 Toxicity**

One of the known toxic compounds to be formed upon chemical engineering of proteins is acrylamide, which results from the Maillard reaction between reducing sugars and asparagine or methionine (Mottram et al., 2002; Stadler et al., 2002). The high solubility of this compound induces rapid absorption and metabolism in the body (reviewed in Dearfield et al., 1988). Following absorption, acrylamide can bind to DNA aiding a genotoxic and carcinogenic response which has been demonstrated in animals (Rudén, 2004). The occurrence of carcinogenic acrylamide in foods has been related to frying and cooking of food products (Rosen & Hellenäs, 2002; Tareke et al., 2002). Two communications published in Nature in 2002 demonstrated the requirement of the reaction between asparagine or methionine and a reducing sugar to intermediate formation of the dicarbonyl reactant followed by Strecker degradation (Mottram et al., 2002; Stadler et al., 2002). Particularly plant proteins are rich in asparagine suggesting that glycosylation of proteins from plant-origin using the Maillard reaction should be carefully considered in terms of the known toxic effects of acrylamide. Other toxic side effects of chemical modification of proteins have not been reported.

#### **4.5.2 Impaired nutrition**

Chemical engineering can theoretically have far-reaching consequences for the nutritional value of proteins. In a limited number of cases protein engineering has been intentionally employed to modify the nutritional quality, for example to increase the nutritional value of plant proteins (Liao et al., 2010). However, the main objective of most studies employing protein engineering is to investigate the consequences of the modification procedure for a range of functional properties. Only a small number of studies investigates nutritional aspects of chemical protein engineering which is mostly regarded as a convenient or inconvenient side effect of the modification procedure. Nutritional aspects covered in literature are exclusively based on *in vitro* studies and cover protein digestibility, availability of essential amino acids or the presence of anti-nutritional factors.

#### **4.5.2.1 Anti-nutritional factors**

160 Advances in Chemical Engineering

present. This is commonly the result of variation in reactivity between the different amino groups present in a specific protein structure: i.e. some amino acids are more exposed and hence are more prone to rapid modification. Buried amino acids do not present a straight forward target for modification as most chemicals are unable to access the folded or aggregated protein structure. Another effect which likely contributes to the extent of heterogeneity is the location of target amino acids in the primary sequence. For example, modification of an amino acid located next to a negatively charged amino acid by succinic anhydride will be hampered as a result of steric hindrance. Some attempts have been made to fractionate protein molecules with different degrees of modification. Five ovalbumin preparations with different degrees of modification have been purified using ion exchange chromatography (Wierenga et al., 2005). The heterogeneous nature of chemically modified proteins has marked implications on the acceptance of these novel ingredients under current food law. Chemical heterogeneity would imply that each obtained modified species would have to be tested separately on toxicity in order to become regarded as safe to use as novel

In this paragraph we will discuss potential health implications induced by chemically modified proteins used in food products including allergenic response, the presence of antinutritional factors linked to proteins and the development of potential toxic compounds as a side-effect of the modification reaction. Concerns can be raised regarding the impact of protein modification on the potential health risks from products containing modified protein ingredients. Bernstein and colleagues (2003) identified three possible modes for novel food ingredients to result in adverse health effects. These include toxicity, impaired

One of the known toxic compounds to be formed upon chemical engineering of proteins is acrylamide, which results from the Maillard reaction between reducing sugars and asparagine or methionine (Mottram et al., 2002; Stadler et al., 2002). The high solubility of this compound induces rapid absorption and metabolism in the body (reviewed in Dearfield et al., 1988). Following absorption, acrylamide can bind to DNA aiding a genotoxic and carcinogenic response which has been demonstrated in animals (Rudén, 2004). The occurrence of carcinogenic acrylamide in foods has been related to frying and cooking of food products (Rosen & Hellenäs, 2002; Tareke et al., 2002). Two communications published in Nature in 2002 demonstrated the requirement of the reaction between asparagine or methionine and a reducing sugar to intermediate formation of the dicarbonyl reactant followed by Strecker degradation (Mottram et al., 2002; Stadler et al., 2002). Particularly plant proteins are rich in asparagine suggesting that glycosylation of proteins from plant-origin using the Maillard reaction should be carefully considered in terms of the known toxic effects of acrylamide.

Other toxic side effects of chemical modification of proteins have not been reported.

Chemical engineering can theoretically have far-reaching consequences for the nutritional value of proteins. In a limited number of cases protein engineering has been intentionally

food ingredient.

**4.5 Health risks** 

**4.5.1 Toxicity** 

nutrition and food allergy.

**4.5.2 Impaired nutrition** 

Anti-nutritional factors are related to reduced protein digestibility and amino acid availability (reviewed by Gilani et al., 2005; Salunkhe et al., 1982) and are commonly present in large concentrations in plant products (Kay, 1979; Liener, 1980)*.* One of the studies in this field reports on the nutritional quality of mung bean isolate following the exposure to varying concentrations of acetic or succinic anhydride to induce acylation (El-Adawy, 2000). The concentrations of anti-nutritional factors tannin, phytic acid, and trypsin inhibitor showed a significant loss with increasing degrees of modification suggesting that this type of modification can positively impact the effect of anti-nutritional factors. The concentration of trypsin inhibitor even decreased with 70% of the original level of trypsin inhibitor in unmodified protein. The introduction of bulky and/or negatively charged side groups was postulated to affect the extent of protein-tannin (El-Adawy, 2000), protein-mineral-phytic acid (Dua et al., 1996; El-Adawy, 2000), or protein-phenol (Loomis (1974) interactions. Loomis (1974) further showed that the flour and protein production processes provide for optimal conditions for the conversion of polyphenols into quinone oxidation products which, in turn, may bind covalently with sulfhydryl groups of cysteine and -amino groups of lysine and N-terminal amino groups. Further support was provided by Dua and colleagues (1996) who showed that acylation and methylation of rapeseed meal and its water-soluble fraction resulted in loss of anti-nutritional factors polyphenol, glucosinolates and phytic acid. However, the methylation procedure employed by Dua and colleagues (1996) resulted in very limited degrees of modification compared to the succinylation and acetylation process suggesting that other factors than the chemical conjugation itself may play a role in the loss of anti-nutritional factors upon chemical modification. As El-Adawy (2000) comments, the extensive dialysis of the protein following the acylation procedure may well be primarily responsible for the loss of water-soluble anti-nutritional factors upon modification. No studies are known to date that report on the loss of anti-nutritional factors upon dialysis.

#### **4.5.2.2** *In vitro* **digestibility**

*In vitro* digestibility of modified proteins is often assayed through exposure of the proteins to a single or a mixture of enzymes including trypsin and pancreatin (Salgó et al., 1984), a combination of trypsin, chymotrypsin and peptidase (Hsu et al., 1977), or pepsin-pancreatin mixtures (Haque et al., 1982) to simulate (post)gastrointestinal digestion of food proteins. The small increase in digestibility of acylated mung bean protein isolate reported by El-Adawy (2000) was primarily correlated to the concurrent loss of tannin; tannins have been shown to play an important role in the reduction of protein digestibility (Barroga et al., 1985). Alternative factors proposed to induce increased digestibility of modified proteins

Application Potential of Food Protein Modification 163

allergenic potential of genetically engineered food ingredients based on sequence homology followed by serum screens (FAO/WHO 2001, reviewed by Bernstein et al., 2003). Based on these findings and the fact that to date there is no indication for a common sequence motif of linear IgE epitopes, it can be concluded that the underlying mechanisms of immunological sensitization to food proteins remain elusive. To be able to assess the potential of chemically modified food proteins to induce allergenic reactions implies extensive knowledge of the underlying mechanical aspects of the allergic response. No publications are known which report on the allergenic effect in humans of modified ingested food proteins. In the pharmaceutical field PEGylation is used to modulate drug delivery of proteins or peptides. An early paper in the late 1970s reported that, upon covalent attachment of methoxypolyethylene glycols of 1900 daltons or 5000 daltons to bovine liver catalase, no evidence was observed of a modulated immune response following repeated injection of the

There is a clear ambition to implement fundamental insights obtained on protein behavior in complex systems more effectively in sustainable food production in the future. There is also the need, considering that one should become more flexible in protein sources and more effective in utilizing the proteins structuring and nutritional potential. The technology to enable this is largely present, as demonstrated by the diversity of literature presented in paragraph 3. The mechanistic insight that has been derived in underlying principles on the relation between (engineered) protein functionality is impressive. And still, the development of new applications using chemical protein engineering to make a better usage of protein functionality or to promote protein source exchangeability is poor. There are a number of reasons to identify. The use of food protein modification has been limited to the domain of food chemistry or food physics. There is no significant literature available where the fine-tuning of microstructure formation as part of product development has been studied at a sensory or (human) physiological level. Compared to genetic modification, chemically engineered proteins pose an inherent heterogeneity. Instead of a single ingredient with an altered functionality, the FDA needs to consider all levels of heterogeneity in their approvals. This is an enormous laborious task as it requires fractionation of derived materials and sufficient stability of the formed products. Moreover, many attempts to reduce heterogeneity in the protein material will lead to non-food grade materials that cannot be studied in relation to e.g. sensory or human physiological aspects. The food safety authorities have amended (see paragraph 4.2) that the protein needs to be seen in the complex context of the food product. This is not just chemical analysis, but also includes product stability, consumer acceptance (sensory) and human physiological data. Especially in these latter two domains there is no public literature readily available to help these authorities to establish objective views on tolerance. This is where the scientific community has a highly needed role in bringing different disciplines together to produce

Aalberse, R.C. & Stapel, S.O. (2001). Structure of food allergens in relation to allergenicity.

modified enzymes into mice (Abuchowski et al., 1977).

literature relevant for authorities to base their legislations on.

*Pediatr. Allergy Immunol.* 12: 10-14.

**5. Future prospects** 

**6. References** 

include the improved access of sites susceptible to enzymatic cleavage as a result of the dissociation of quaternary complexes of proteins or partial unfolding induced by the modification procedure (Achouri & Zhang, 2001). Protein unfolding as a result of modification has been shown for soy protein hydrolysate using techniques to study the secondary and tertiary structure content of the protein upon succinylation (Achouri & Zhang, 2001). However, the acylation of cotton seed flour did not improve *in vitro* protein digestibility (Rahma & Narasinga Rao, 1983) suggesting that results in this area are somewhat controversial and perhaps other factors play a role (see paragraph 4.5.2.1). *In vitro* protein digestibility using a multienzyme system containing trypsin, chymotrypsin and peptidase was not impaired for pea proteins upon acetylation (Johnson & Brekke, 1983) while Ma (1984) reported increased digestibility for acylated pea protein, similar to low degrees of succinylation of soy protein hydrolysate (Achouri & Zhang, 2001). Loss of *in vitro* digestibility has also been reported upon succinylation of a variety of proteins, particularly affecting the release of lysine (Matoba & Doi, 1979; Siu & Thompson, 1982; Wanasundara & Shahidi, 1997) or lipophilization of soy bean glycinin with palmitic acid (Haque et al., 1982). Data on *in vivo* digestibility of ingested proteins upon modification have not been reported.

#### **4.5.2.3 Availability of essential amino acids**

A commonly used target for protein modification are the lysine or cysteine residues, lysine being classified as an essential amino acid, i.e. this amino acid cannot be synthesized *de novo* by humans and should therefore be ingested. Extensive modification of these amino acids can therefore result in a lower availability. Few studies report on the impact of lysine or cysteine modification on the availability of these amino acids, usually assayed through amino acid analysis. Overall only small decreases in lysine were reported upon succinylation of soy protein hydrolysate (Achouri & Zhang, 2001). Similar findings were reported for acylated soy proteins (Franzen & Kinsella, 1976a), acylated sunflower proteins (Kabirullah & Wills, 1982), and succinylation of oat proteins (Ma & Wood, 1987). It is not clear whether these reported effects of chemical engineering inducing the loss of lysine availability ultimately result in a noticeable and substantial loss of nutritional quality for the human population.

#### **4.5.3 Food allergy and intolerance**

An estimated 3-4% of the children and 1-2% of adults in the industrialized world exert an allergenic response to one or more ingested food proteins (Baral & Hourihane, 2005; Jansen et al., 1994). Not all ingested proteins behave as allergens but proteins implicated in allergenic response often share features such as unusual resilience against heat, acid or protease digestion, propensity to bind to lipids and are glycosylated to some degree (Lehrer et al., 2002; Metcalfe et al., 1996). Also posttranslational modifications including *N*glycosylation, and hydroxylation of proline residues have been postulated to affect IgE reactivity to Phl p 1 present in timothy grass pollen (Petersen et al., 1998). The precise mechanism of the effects of glycosylation on the allergenic response are not clear. Attempts have been made to develop and evaluate algorithms which predict protein allergenic response based on sequence homology (Aalberse & Stapel, 2001; Jameson & Wolf, 1988), structural identity, and evolutionary relationship (Jenkins et al., 2007), albeit with limited predicting power. The Food and Agriculture Organization (FAO) of the United Nations together with the World Health Organization (WHO) developed a decision tree to assess the allergenic potential of genetically engineered food ingredients based on sequence homology followed by serum screens (FAO/WHO 2001, reviewed by Bernstein et al., 2003). Based on these findings and the fact that to date there is no indication for a common sequence motif of linear IgE epitopes, it can be concluded that the underlying mechanisms of immunological sensitization to food proteins remain elusive. To be able to assess the potential of chemically modified food proteins to induce allergenic reactions implies extensive knowledge of the underlying mechanical aspects of the allergic response. No publications are known which report on the allergenic effect in humans of modified ingested food proteins. In the pharmaceutical field PEGylation is used to modulate drug delivery of proteins or peptides. An early paper in the late 1970s reported that, upon covalent attachment of methoxypolyethylene glycols of 1900 daltons or 5000 daltons to bovine liver catalase, no evidence was observed of a modulated immune response following repeated injection of the modified enzymes into mice (Abuchowski et al., 1977).

#### **5. Future prospects**

162 Advances in Chemical Engineering

include the improved access of sites susceptible to enzymatic cleavage as a result of the dissociation of quaternary complexes of proteins or partial unfolding induced by the modification procedure (Achouri & Zhang, 2001). Protein unfolding as a result of modification has been shown for soy protein hydrolysate using techniques to study the secondary and tertiary structure content of the protein upon succinylation (Achouri & Zhang, 2001). However, the acylation of cotton seed flour did not improve *in vitro* protein digestibility (Rahma & Narasinga Rao, 1983) suggesting that results in this area are somewhat controversial and perhaps other factors play a role (see paragraph 4.5.2.1). *In vitro* protein digestibility using a multienzyme system containing trypsin, chymotrypsin and peptidase was not impaired for pea proteins upon acetylation (Johnson & Brekke, 1983) while Ma (1984) reported increased digestibility for acylated pea protein, similar to low degrees of succinylation of soy protein hydrolysate (Achouri & Zhang, 2001). Loss of *in vitro* digestibility has also been reported upon succinylation of a variety of proteins, particularly affecting the release of lysine (Matoba & Doi, 1979; Siu & Thompson, 1982; Wanasundara & Shahidi, 1997) or lipophilization of soy bean glycinin with palmitic acid (Haque et al., 1982). Data on *in vivo* digestibility of ingested proteins upon modification have not been reported.

A commonly used target for protein modification are the lysine or cysteine residues, lysine being classified as an essential amino acid, i.e. this amino acid cannot be synthesized *de novo* by humans and should therefore be ingested. Extensive modification of these amino acids can therefore result in a lower availability. Few studies report on the impact of lysine or cysteine modification on the availability of these amino acids, usually assayed through amino acid analysis. Overall only small decreases in lysine were reported upon succinylation of soy protein hydrolysate (Achouri & Zhang, 2001). Similar findings were reported for acylated soy proteins (Franzen & Kinsella, 1976a), acylated sunflower proteins (Kabirullah & Wills, 1982), and succinylation of oat proteins (Ma & Wood, 1987). It is not clear whether these reported effects of chemical engineering inducing the loss of lysine availability ultimately result in a noticeable and substantial loss of nutritional quality for the

An estimated 3-4% of the children and 1-2% of adults in the industrialized world exert an allergenic response to one or more ingested food proteins (Baral & Hourihane, 2005; Jansen et al., 1994). Not all ingested proteins behave as allergens but proteins implicated in allergenic response often share features such as unusual resilience against heat, acid or protease digestion, propensity to bind to lipids and are glycosylated to some degree (Lehrer et al., 2002; Metcalfe et al., 1996). Also posttranslational modifications including *N*glycosylation, and hydroxylation of proline residues have been postulated to affect IgE reactivity to Phl p 1 present in timothy grass pollen (Petersen et al., 1998). The precise mechanism of the effects of glycosylation on the allergenic response are not clear. Attempts have been made to develop and evaluate algorithms which predict protein allergenic response based on sequence homology (Aalberse & Stapel, 2001; Jameson & Wolf, 1988), structural identity, and evolutionary relationship (Jenkins et al., 2007), albeit with limited predicting power. The Food and Agriculture Organization (FAO) of the United Nations together with the World Health Organization (WHO) developed a decision tree to assess the

**4.5.2.3 Availability of essential amino acids**

human population.

**4.5.3 Food allergy and intolerance** 

There is a clear ambition to implement fundamental insights obtained on protein behavior in complex systems more effectively in sustainable food production in the future. There is also the need, considering that one should become more flexible in protein sources and more effective in utilizing the proteins structuring and nutritional potential. The technology to enable this is largely present, as demonstrated by the diversity of literature presented in paragraph 3. The mechanistic insight that has been derived in underlying principles on the relation between (engineered) protein functionality is impressive. And still, the development of new applications using chemical protein engineering to make a better usage of protein functionality or to promote protein source exchangeability is poor. There are a number of reasons to identify. The use of food protein modification has been limited to the domain of food chemistry or food physics. There is no significant literature available where the fine-tuning of microstructure formation as part of product development has been studied at a sensory or (human) physiological level. Compared to genetic modification, chemically engineered proteins pose an inherent heterogeneity. Instead of a single ingredient with an altered functionality, the FDA needs to consider all levels of heterogeneity in their approvals. This is an enormous laborious task as it requires fractionation of derived materials and sufficient stability of the formed products. Moreover, many attempts to reduce heterogeneity in the protein material will lead to non-food grade materials that cannot be studied in relation to e.g. sensory or human physiological aspects. The food safety authorities have amended (see paragraph 4.2) that the protein needs to be seen in the complex context of the food product. This is not just chemical analysis, but also includes product stability, consumer acceptance (sensory) and human physiological data. Especially in these latter two domains there is no public literature readily available to help these authorities to establish objective views on tolerance. This is where the scientific community has a highly needed role in bringing different disciplines together to produce literature relevant for authorities to base their legislations on.

#### **6. References**

Aalberse, R.C. & Stapel, S.O. (2001). Structure of food allergens in relation to allergenicity. *Pediatr. Allergy Immunol.* 12: 10-14.

Application Potential of Food Protein Modification 165

Baldwin, R.L. (2008). The search for folding intermediates and the mechanism of protein

Baldwin, E.A. (1994). Edible coatings for fresh fruits and vegetables: past, present, and

Baral, V.R. & Hourihane, J.O. (2005). Food allergy in children. *Postgrad. Med. J.* 81: 693-701. Barbut, S. & Foegeding, E.A. (1993). Ca2+-induced gelation of preheated whey protein

Barroga, C.F.; Laurena, A.C. & Mendoza, E.M.T. (1985). Polyphenol in mung bean (*Vigna radiata* L. Wilczek) determination and removal. *J. Agric. Food Chem.* 33: 1006-1009. Bauer, H.H.; Aebi, U.; Häner, M.; Hermann, R.; Müller, M.; Arvinte, T. & Merkle, H.P.

Bernstein, J.A.; Bernstein, I.L.; Bucchini, L.; Goldman, L.R.; Hamilton, R.G.; Lehrer, S.; Rubin,

Betz, S.F. (1993). Disulfide bonds and the stability of globular proteins. *Protein Sci.* 2: 1551-

Borgia, A.; Williams, P.M. & Clarke, J. (2008). Single-molecule studies of protein folding.

Bowes, J.H. & Cater, C.W. (1968). The interaction of aldehydes with collagen. *Biochim.* 

Broersen, K.; Weijers, M.; de Groot, J.; Hamer, R.J. & de Jongh, H.H. (2007a). Effect of protein

Broersen, K.; Elshof, M.; de Groot, J.; Voragen, A.G.J.; Hamer, R.J. & de Jongh, H.H.J.

Broersen, K.; Voragen, A.G.; Hamer, R.J. & de Jongh, H.H. (2004). Glycoforms of beta-

Broersen, K.; van Teeffelen, A.M.; Vries, A.; Voragen, A.G.; Hamer, R.J. & de Jongh, H.H.

Bromley, E.H.C.; Krebs, M.R.H. & Donald, A.M. (2006). Mechanisms of structure formation

Buchner, S.; Kinnear, M.; Crouch, I.J.; Taylor, J. & Minnaar, A. (2011). Effect of kafirin

charge on the generation of aggregation-prone conformers. *Biomacromolecules* 8:

(2007b). Aggregation of -lactoglobulin regulated by glucosylation. *J. Agric. Food* 

lactoglobulin with improved thermostability and preserved structural packing.

(2006). Do sulfhydryl groups affect aggregation and gelation properties of

in particulate gels of -lactoglobulin formed near the isoelectric point. *Eur. Phys. J.* 

protein coating on sensory quality and shelf-life of 'Packham's Triumph' pears

genetically modified foods. *Environ. Health Persp.* 111: 1114-1121.

future. In: Edible Coatings and Films to Improve Food Quality. Ed. By Krochta, J.M., Bandwin, E.A., Nisperos-Carriedo, M.O., Technomic Publishing, Lancaster

(1995). Architecture and polymorphism of fibrillar supramolecular assemblies produced by *in vitro* aggregation of human calcitonin. *J. Struct. Biol.* 115: 1-15. Baumketner, A.; Jewett, A. & Shea, J.E. (2003). Effects of confinement in chaperonin assisted

protein folding: rate enhancement by decreasing the roughness of the folding

C. & Sampson, H.A. (2003). Clinical and laboratory investigation of allergy to

folding. *Annu. Rev. Biophys.* 37: 1-21.

solutions. *J. Food Sci.* 58: 867-871.

*Annu. Rev. Biochem.* 77: 101-125.

*Biophys. Acta* 168: 341-352.

*Chem.* 55: 2431-2437.

*E.* 21: 145-152.

*Biotechnol. Bioeng.* 86: 78-87.

ovalbumin? *J. Agric. Food Chem.* 54: 5166-5174.

during ripening. *J. Sci. Food Agric.* In press.

energy landscape. *J. Mol. Biol.* 332: 701-713.

PA, pp. 25-64.

1558.

1648-1656.


Abuchowski, A.; McCoy, J.R.; Palczuk, N.C.; van Es, T. & Davis, F.F. (1977). Effect of

Achouri, A. & Zhang, W. (2001). Effect of succinylation on the physicochemical properties of

Aewsiri, T.; Benjakul, S.; Visessanguan W.; Wierenga, P.A. & Gruppen, H. (2010).

Aewsiri, T.; Benjakul, S.; Visessanguan, W.; Wierenga, P.A. & Gruppen, H. (2011a).

Aewsiri, T.; Benjakul, S.; Visessanguan, W.; Wierenga, P.A. & Gruppen, H. (2011b). Surface

Alting, A.C.; Hamer, R.J.; de Kruif, C.G. & Visschers, R.W. (2003). Cold-set globular protein

Andersson, L.-O.; Brandt, J. & Johansson, S. (1971). The use of trinitrobenezenesulfonic acid

Anfinsen, C.B. (1973). Principles that govern the folding of protein chains. *Science* 181: 223-

Anson, M.L. (1940). The reactions of iodine and iodoacetamide with native egg albumin. *J.* 

Antoniewski, M.N.; Barringer, S.A.; Knipe, L. & Zerby, H. (2007). The effect of a gelatin

Arnaudov, L.N.; de Vries, R.; Ippel, H. & van Mierlo, C.P. (2003). Multiple steps during the

Arnaudov, L.N. & de Vries, R. (2005). Thermally induced fibrillar aggregation of hen egg

Arnaudov, L.N. & de Vries, R. (2006). Strong impact of ionic strength on the kinetics of fibrilar aggregation of bovine -lactoglobulin. *Biomacromolecules* 7: 3490-3498. Arntfield, S.D.; Murray, E.D. & Ismond, M.A.H. (1991). Role of disulfide bonds in

networks from ovalbumin and vicilin. *J. Agric. Food Chem.* 39: 1378-1385. Arvanitoyannis, I.S. (1999). Totally and partially biodegradable polymer blends based on

Audic, J.L. & Chaufer, B. (2005). Influence of plasticizers and crosslinking on the properties of biodegradable films made from sodium caseinate. *Eur. Polym. J.* 41: 1934-1942. Bai, J.; Alleyne, V.; Hagenmaier, R.D.; Mattheis, J.P. & Baldwin, E.A. (2003). Formulation of

Baldursdottir, S.G.; Fullerton, M.S.; Nielsen, S.H. & Jorgensen, L. (2010). Adsorption of

potential as food packaging materials. *Polym. Rev.* 39: 205-271.

shear stress measurements. *Colloids Surf. B Biointerfaces* 79: 41-46.

determining the rheological and microstructural properties of heat-induced protein

natural and synthetic macromolecules: preparation, physical properties, and

zein coatings for apples (*Malus domestica* Borkh). *Postharv. Biol. Technol.* 28: 259-268.

proteins at the oil/water interface – observation of protein adsorption by interfacial

coating on the shelf-life of fresh meat. *J. Food Sci.* 72: E382-E387.

formation of -lactoglobulin fibrils. *Biomacromolecules* 4: 1614-1622.

*N*-hydroxysuccinimide esters of fatty acid. *Food Hydrocoll.* In press.

of bovine liver catalase. *J. Biol. Chem.* 252: 3582-3586.

soy protein hydrolysate. *Food Res. Int.* 34: 507-514.

oxidized linoleic acid. *Int. J. Biol. Macromol.* In press.

*Agric. Food Chem.* 51: 3150-3156.

*Biochem. Biophys.* 146: 428-440.

white lysozyme. *Biophys. J.* 88: 515-526.

*Gen. Physiol.* 23: 321-331.

press.

230.

covalent attachment of polyethylene glycol on immunogenicity and circulating life

Improvement of foaming properties of cuttlefish skin gelatin by modification with

Emulsifying property and antioxidative activity of cuttlefish skin gelatin modified with oxidized linoleic acid and oxidized tannic acid. *Food Bioprocess. Technol.* In

activity and molecular characteristics of cuttlefish skin gelatin modified by

gels: interactions, structure and rheology as a function of protein concentration. *J.* 

in studies on the binding of fatty acid anions to bovine serum albumin. *Arch.* 


Application Potential of Food Protein Modification 167

Dearfield, K.L.; Abernathy, C.O.; Ottley, M.S.; Brantner, J.H. & Hayes, P.F. (1988).

De Groot, J.; Kosters, H.A. & de Jongh, H.H. (2007). Deglycosylation of ovalbumin prohibits formation of a heat-stable conformer. *Biotechnol. Bioeng.* 97: 735-741.

De Jongh, H.H.J.; Kosters, H.A.; Kudryashova, E.; Meinders, M.B.J.; Trofimova, D. &

De Jongh, H.H.J. & Wierenga, P.A. (2006). Assessing the extent of protein intermolecular

De Jongh, H.H.; Taylor, S.L. & Koppelman, S.J. (2011). Controlling the aggregation

Del Rosario Moreira, M.; Pereda, M.; Marcovich, N.E. & Roura, S.I. (2011). Antimicrobial

Dickinson, E. (1992). An introduction to food colloids. Oxford, UK: Oxford University Press. Djagny, K.B.; Wang, Z. & Xu, S. (2001). Conformational changes and some functional

Dobson, C.M. (1999). Protein misfolding, evolution and disease. *Trends Biochem. Sci.* 24: 329-

Dondero, M.; Figueroa, V.; Morales, X. & Curotto, E. (2006). Transglutaminase effects on

Dua, S.; Mahajan, A. & Mahajan, A. (1996). Improvement of functional properties of

DuBay, K.F.; Pawar, A.P.; Chiti, F.; Zurdo, J.; Dobson, C.M. & Vendrusculo, M. (2004).

El-Adawy, T.A. (2000). Functional properties and nutritional quality of acetylated and

Emmambux, M.N.; Stading, M. & Taylor, J.R.N. (2004). Sorghum kafirin film property modification with hydrolysable and condensed tannins. *J. Cereal Sci.* 40: 127-135. Englander, S.W.; Mayne, L. & Krishna, M.M. (2007). Protein folding and misfolding:

Evans, M.T.A. & Irons, L. (1970). N-succinylated egg yolk proteins. German Patent 1 951 247. FAO/WHO (2001). Evaluation of allergenicity of genetically modified foods. Report of a

Joint FAO/WHO Expert Consultation of Allergenicity of Foods Derived from Biotechnology. Available: ftp://ftp.fao.org/es/esn/food/allergygm.pdf [accessed

succinylated mung bean protein isolate. *Food Chem.* 70: 83-91. Ellman, G.L. (1959). Tissue sulfhydryl groups. *Arch. Biochem. Biophys.* 82: 70-77.

mechanisms and principles. *Q. Rev. Biophys.* 40: 287-326.

Doi, E. (1993). Gels and gelling of globular proteins. *Trends Food Sci. Technol.* 4: 1-5.

and carcinogenicity. *Mutat. Res.* 195: 45-77.

cod fish parvalbumin. *J. Biosci. Bioeng.* 111: 204-211.

functionality. *Curr. Opin. Biotechnol.* 10: 324-330.

details. *Biopolymers* 74: 131-135.

*Agric. Food Chem.* 44: 706-710.

*Mol. Biol.* 341: 1317-1326.

08 August 2011].

384-389.

332.

Acrylamide: its metabolism, developmental and reproductive effects, genotoxicity,

Wierenga, P.A. (2004). Protein adsorption at air-water interfaces; a combination of

interactions at air-water interfaces using spectroscopic techniques. *Biopolymers* 82:

propensity and thereby digestibility of allergens by Maillardation as illustrated for

effectiveness of bioactive packaging materials from edible chitosan and casein polymers: assessment on carrot, cheese, and salami. *J. Food Sci.* 76: M54-M63. DeSantis, G. & Jones, J.B. (1999). Chemical modification of enzymes for enhanced

characteristics of gelatin esterified with fatty acid. *J. Agric. Food Chem.* 49: 2987-2991.

gelation capacity of thermally induced beef protein gels. *Food Chemistry* 99, 546-554.

rapeseed (*Brassica campestris* Var. Toria) preparations by chemical modification. *J.* 

Prediction of the absolute aggregation rates of amyloidogenic polypeptide chains. *J.* 


Buchner, G.S.; Murphy, R.D.; Buchete, N.V. & Kubelka, J. (2011). Dynamics of protein

Bustos, P.; Gajardo, M.I.; Gómez, C.; Hughes, G.; Cardemil, E. & Jabalquinto, A.M. (1996).

Calamai, M.; Taddei, N.; Stefani, M., Ramponi, G. & Chiti, F. (2003). Relative influence of

Canet, D.; Sunde, M.; Last, A.M.; Miranker, A.; Spencer, A.; Robinson, C.V. & Dobson, C.M.

Chang, T.-S. & Sun, S.F. (1978). Structural studies on the succinylated bovine serum

Charissou, A.; Ait-Ameur, L. & Birlouez-Aragon, I. (2007). Kinetics of formation of three

Chibnall, A.C. (1942). Bakerian lecture: amino-acid analysis and the structure of proteins.

Chiti, F. & Dobson, C.M. (2006). Protein misfolding, functional amyloid, and human disease.

Chiti, F.; Stefani, M.; Taddei, N.; Ramponi, G. & Dobson, C.M. (2003). Rationalization of the effects of mutations on peptide and protein aggregation rates. *Nature* 424: 805-808. Chiti, F.; Taddei, N.; Bucciantini, M.; White, P.; Ramponi, G. & Dobson, C.M. (2000).

Church, F.C.; Swaisgood, H.E.; Porter, D.H. & Catignani, G.L. (1983). Spectrophotometric

Cieplak, M. (2004). Cooperativity and contact order in protein folding. *Phys. Rev. E Stat.* 

Corzo-Martínez, M.; Soria, A.C.; Villamiel, M.; Olano, A.; Harte, F.M. & Moreno, F.J. (2011).

Cucu, T.; Platteau, C.; Taverniers, I.; Devreese, B.; de Loose, M. & de Meulenaer, B. (2011).

Damodaran, V.B. & Fee, C. (2010). Protein PEGylation: an overview of chemistry and

Creighton, T.E. (1988). Disulphide bonds and protein stability. *Bioessays* 8: 57-63.

temperature and type of sugar. *J. Agric. Food Chem.* 55: 4532-4539.

experiment and theory. *Biochim. Biophys. Acta* 1814: 1001-1020.

*Chem.* 15: 467-472.

*Biochemistry* 42: 15078-15083.

*Proc. Roy. Soc. B* 131: 136-160.

*EMBO J.* 19: 1441-1449.

*Assess.* 28: 1-10.

*Annu. Rev. Biochem.* 75: 333-366.

*Nonlin. Soft Matter Phys.* 69: 031907.

ultrasound. *J. Dairy Sci.* 94: 51-58.

albumin. *Int. J. Peptide Protein Res.* 11: 65-72.

isolated milk proteins. J. Dairy Sci. 66: 1219-1227.

process considerations. *Eur. Pharm. Rev* 15: 18-26.

*Microbiol.* 27: 940-944.

folding: probing the kinetic network of folding-unfolding transitions with

Woodward's reagent K reacts with histidine and Cysteine residues in *Escherichia coli* and *Saccharomyces cerevisiae* phosphoenolpyruvate carboxykinases. *J. Protein* 

hydrophobicity and net charge in the aggregation of two homologous proteins.

(1999). Mechanistic studies of the folding of human lysozyme and the origin of amyloidogenic behavior of its disease-related variants. *Biochemistry* 38: 6419-6427. Cao-Hoang, L.; Chaine, A.; Grégoire, L. & Waché, Y. (2010). Potential of nisin-incorporated

sodium caseinate films to control *Listeria* in artificially contaminated cheese. *Food* 

indicators of the Maillard reaction in model cookies: influence of baking

Mutational analysis of the propensity for amyloid formation by a globular protein.

assay using *o*-phthaldialdehyde for determination of proteolysis in milk and

Effect of glycation on sodium caseinate-stabilized emulsions obtained by

ELISA detection of hazelnut proteins: effect of protein glycation in the presence or absence of wheat proteins. *Food Addit. Contam. Part A Chem. Anal. Control Expo. Risk* 


Application Potential of Food Protein Modification 169

Graham, D.E. & Phillips, M.C. (1979c). Proteins at liquid interfaces. III. Molecular structures

Graña-Montes, R.; de Groot, N.S.; Castillo, V.; Sancho, J.; Velazquez-Campoy, A. & Ventura,

Groninger, H.S. Jr. (1973). Preparation and properties of succinylated fish myofibrillar

Gruener, L. & Ismond, M.A.H. (1997). Effects of acetylation and succinylation on the physicochemical properties of the canola 12S globulin. *Food Chem.* 60: 357-363. Gueguen, J.; Bollecker, S.; Schwenke, K.D. & Raab, B. (1990). Effect of succinylation on some

Gurin, S. & Clarke, H.T. (1934). Allocation of the free amino groups in proteins and

Habeeb, A.F.S.A. & Atassi, M.Z. (1969). Enzymic and immunochemical properties of

Hall, R.J.; Trinder, N. & Givens, D.I. (1973). Observations on the use of 2,4,6-

Halling, P.J. (1981). Protein-stabilized foams and emulsions. *Crit. Rev. Food Sci. Nutr.* 15: 155-

Hamada, J.S. & Marshall, W.E. (1989). Preparation and functional properties of

Hamano, Y. (2011). Occurrence, biosynthesis, biodegradation, and industrial and medical

Haque, Z.; Matoba, T. & Kito, M. (1982). Incorporation of fatty acid into food protein:

Haque, Z. & Kito, M. (1983a). Lipophilization of s1-casein. 1. Covalent attachment of

Haque, Z. & Kito, M. (1983b). Lipophilization of s1-casein. 2. Conformational and

Hardy, P.M.; Nicholls, A.C. & Rydon, H.N. (1969). The nature of gluteraldehyde in aqueous

Harper, J.D.; Wong, S.S.; Lieber, C.M. & Lansbury, P.T. Jr. (1997). Atomic force microscopy

Havard, C. & Harmony, M.X. (1869). Improved process of preserving meat, fowls, fish. June

imaging of seeded fibril formation and fibril branching by the Alzheimer's disease

enzymatically deamidated soy proteins. *J. Food Sci.* 54: 598-601.

palmitoyl soybean glycinin. *J. Agric. Food Chem.* 30: 481-486.

palmitoyl residue. *J. Agric. Food Chem.* 31: 1225-1230.

functional effects. *J. Agric. Food Chem.* 31: 1231-1237.

solution. *J. Chem. Soc. Chem. Commun.* 10: 565-566.

Hart, G.W. (1992). Glycosylation. *Curr. Opin. Cell Biol.* 4: 1017-1723.

amyloid- protein. *Chem. Biol.* 4: 951-959.

8, U.S. Patent 90: 944.

formation: the PI3-SH3 domain case. *Antioxid. Redox Signal.* In press. Grigoriev, D.O.; Derkatch, S.; Krägel, J. & Miller, R. (2007). Relationship between structure

S. (2011). Contribution of disulfide bonds to stability, folding and amyloid fibril

and rheological properties of mixed BSA/Tween 80 adsorption layers at the

physicochemical and functional properties of the 12S storage protein from rapeseed

lysozyme II. Conformation, immunochemistry and enzymic activity of a derivative

trinitrobenzenesulphonic acid for the determination of available lysine in animal

applications of a naturally occurring -poly-L-lysine. *Biosci. Biotechnol. Biochem.* 75:

at adsorbed films. *J. Colloid Interface Sci.* 70: 427-439.

air/water interface. *Food Hydrocoll.* 21: 823-830.

(*Brassica napus* L.). *J. Agric. Food Chem.* 38: 61-69.

modified at tryptophan. *Immunohistochemistry* 6: 555-566.

protein. *J. Agric. Food Chem.* 21: 978-981.

peptides. *J. Biol. Chem.* 107: 395-419.

protein concentrates. *Analyst* 98: 673-686.

203.

1226-1233.


Farouk, M.M.; Price, J.F. & Salih, A.M. (1990). Effect of an edible film overwrap on exudation and lipid oxidation in beef round steak. *J. Food Sci.* 55: 1510-1512, 1563. Feeney, R.E.; Blankenhorn, G. & Dixon, H.B. (1975). Carbonyl-amine reactions in protein

Feeney, R.E.; Yamasaki, R.B. & Geoghegan, K.F. (1982). Chapter 1: Chemical modification of

Fernandez-Escamilla, A.M.; Rousseau, F.; Schymkowitz, J. & Serrano, L. (2004). Prediction of

Ferreira, S.T.; De Felice, F.G. & Chapeaurouge, A. (2006). Metastable, partially folded states

Ferreon, A.C. & Deniz, A.A. (2011). Protein folding at single-molecule resolution. *Biochim.* 

Fields, R. (1971). The measurement of amino groups in proteins and peptides. *Biochem. J.* 

Fraenkel-Conrat, H. (1957). Methods for investigating the essential groups for enzyme

Finley, J.W. (1975). Deamidated gluten: a potential fortifier for fruit juices. *J. Food Sci.* 40:

Francis, G.E.; Fisher, D.; Delgado, C.; Malik, F.; Gardiner, A. & Neale, D. (1998). PEGylation

Franzen, K.L. & Kinsella, J.E. (1976b). Functional properties of succinylated and acetylated

Frister, H.; Meisel, H. & Schlimme, E. (1988). OPA method modified by use of *N*,*N*-

Gerbanowski, A.; Rabiller, C.; Larré, C. & Guéguen, J. (1999a). Grafting of aliphatic and

Gerbanowski, A.; Malabat, C.; Rabiller, C. & Guéguen, J. (1999b). Grafting of aliphatic and

Gilani, G.S.; Cockell, K.A. & Sepehr, E. (2005). Effects of antinutritional factors on protein digestibility and amino acid availability in foods. *J. AOAC Int* 88: 967-987. Gillgren, T. & Stading, M. (2008). Mechanical and barrier properties of avenin, kafirin and

Graham, D.E. & Phillips, M.C. (1979a). Proteins at liquid interfaces. I. Kinetics of adsorption

Graham, D.E. & Phillips, M.C. (1979b). Proteins at liquid interfaces. II. Adsorption

physicochemical characteristics. *J. Protein Chem.* 18: 325-335.

physicochemical properties. *J. Agric. Food Chem.* 47: 5218-5226.

and surface denaturation. *J. Colloid Interface Sci.* 70: 403-414.

biological optimisation of coupling techniques. *Int. J. Hematol.* 68: 1-18. Franzen, K.L. & Kinsella, J.E. (1976a). Functional properties of succinylated and acetylated

of cytokines and other therapeutic proteins and peptides: the importance of

dimethyl-2-mercaptoethylammonium chloride as thiol component. *Fresenius'Z.* 

aromatic probes to bovine serum albumin: influence on its structural and

aromatic probes on rapeseed 2S and 12S proteins: influence on their structural and

proteins: an overview. In 'Modification of Proteins', Advances in Chemistry;

sequence-dependent and mutational effects on the aggregation of peptides and

in the productive folding and in the misfolding and amyloid aggregation of

chemistry. *Adv. Protein Chem.* 29: 135-203.

proteins. *Nat. Biotechnol.* 22: 1302-1306.

*Biophys. Acta* 1814: 1021-1029.

*Anal. Chem.* 330: 631-633.

zein films. *Food Biophys.* 3: 287-294.

isotherms. *J. Colloid Interface Sci.* 70: 415-426.

activity. *Methods Enzymol.* 4: 247-269.

soy protein. *J. Agric. Food Chem.* 24: 788-795.

leaf proteins. *J. Agric. Food Chem.* 24: 914-919.

124: 581-590.

1283-1285.

proteins. *Cell Biochem. Biophys.* 44: 539-548.

American Chemical Society, Washington DC. pp 3-55.


Application Potential of Food Protein Modification 171

Jarrett, J.T. & Lansbury, P.T. Jr. (1993). Seeding "one-dimensional crystallization" of

Jenkins, J.A.; Breiteneder, H. & Mills, E.N.C. (2007). Evolutionary distance from human

Jiménez, J.; Nettleton, E.J.; Bouchard, M.; Robinson, C.V.; Dobson, C.M. & Saibil, H. (2002).

Jiménez, J.; Guijarro, J.I.; Orlova, E.; Zurdo, J.; Dobson, C.M.; Sunde, M. & Saibil, H.R. (1999).

Johnson, E.A. & Brekke, J. (1983). Functional properties of acylated pea protein isolates. *J.* 

Johnson, A.R. & Dekker, E.E. (1996). Woodward's reagent K inactivation of *Escherichia coli* L-

Jonas, A. & Weber, G. (1970). Partial modification of bovine serum albumin with

Ju, Z.Y. & Kilara, A. (1998). Aggregation induced by calcium chloride and subsequent

Kabirullah, M. & Wills, R.B.H. (1982). Functional properties of acetylated and succinylated

Kaneko, R. & Kitabatake, N. (2001). Structure-sweetness relationship in thaumatin:

Kato, A.; Ibrahim, H.; Takagi, T. & Kobayashi, K. (1990). Excellent gelation of egg white

Kato, A.; Nagase, Y.; Matsudomi, N. & Kobayashi, K. (1983). Determination of molecular-

Kay, D.E. (1979). Crop and product digest. No. 3. Food legumes. London, Tropical Products

Kelly, J.W. (1998). The alternative conformations of amyloidogenic proteins and their multi-

Kitabatake, N.; Wada, R. & Fujita, Y. (2001). Reversible conformational change in -

Klotz, I.M. & Heiney, R.E. (1962). Introduction of sulfhydryl groups into proteins using acetylmercaptosuccinic anhydride. *Arch. Biochem. Biophys.* 96: 605-612. Korn, A.H.; Feairheller, S.H. & Filachione, E.M. (1972). Gluteraldehyde: nature of the

Kornfeld, R. & Kornfeld, S. (1985). Assembly of asparagines-linked oligosaccharides. *Annu.* 

preheated in the dry state is due to the decreasing degree of aggregation. *J. Agric.* 

weight of soluble ovalbumin aggregates during heat denaturation using low-angle

lactoglobulin A modified with *N*-ethylmaleimide and resistance to molecular

thermal gelation of whey protein isolate. *J. Dairy Sci.* 81: 925-931.

laser-light scattering technique. *Agric. Biol. Chem.* 47: 1829-1834.

step assembly pathways. *Curr. Opin. Struct. Biol.* 8: 101-106.

aggregation on heating. *J. Agric. Food Chem.* 49: 4011-4018.

sunflower protein isolate. *J. Food Technol.* 17: 235-249.

importance of lysine residues. *Chem. Senses* 26: 167-177.

1055-1058.

120: 1399-1405.

99: 9196-9201.

4729-4735.

*Food Sci.* 48: 722-725.

groups. *Protein Sci.* 5: 382-390.

*Food Chem.* 38: 1868-1872.

reagent. *J. Mol. Biol.* 65: 525-529.

*Rev. Biochem.* 54: 631-664.

Institute, 435 p.

molecular packing. *EMBO J.* 18: 815-821.

amyloid: a pathogenic mechanism in Alzheimer's disease and scrapie? *Cell* 73:

homologs reflects allergenicity of animal food proteins. *J. Allergy Clin. Immunol.*

The protofilament structure of insulin amyloid fibrils. *Proc. Natl. Acad. Sci. U.S.A.* 

Cryo-electron microscopy structure of an SH3 amyloid fibril and model of the

threonine dehydrogenase: increased absorbance at 340-350 nm is due to modification of cysteine and histidine residues, not aspartate or glutamate carboxyl

dicarboxylic anhydrides. Physical properties of the modified species. *Biochemistry* 9:


Hayakawa, S. & Nakai, S. (1985). Contribution of the hydrophobicity, net charge and

Haynes, R.; Osuga, D.T. & Feeney, R.E. (1967). Modification of amino groups in inhibitors of

He, X.-H.; Shaw, P.-C. & Tam, S.-C. (1999). Reducing the immunogenicity and improving

Helenius, A.; Trombetta, E.S.; Hebert, D.N. & Simons, J.F. (1997) Calnexin, calreticulin and

Hillier, R.M.; Lyster, R.L.J. & Cheeseman, G.C. (1980). Gelation of reconstituted whey

Hoagland, P.D. (1968). Acylated -caseins. Effect of alkyl group size on calcium ion

Hoffmann, M.A.M. & van Mil, P.J.J.M. (1997). Heat-induced aggregation of -lactoglobulin: role of the free thiol group and disulfide bonds. *J. Agric. Food Chem.* 45: 2942-2948. Hopwood, D. (1969). A comparison of the crosslinking abilities of gluteraldehyde,

Hopwood, D., Callen, C.R. & McCabe, M. (1970). The reactions between gluteraldehyde and various proteins. An investigation of their kinetics. *Histochem. J.* 2: 137-150. Hsu, H.W.; Vavak, D.L.; Satterlee, L.S. & Miller, G.A. (1977). A multienzyme technique for

Huggins, C. & Jensen, E.V. (1949). Thermal coagulation of serum proteins; the effects of

Hunter, J.R.; Carbonell, R.G. & Kilpatrick, P.K. (1991). Coadsorption and exchange of

Hurle, M.R.; Helms, L.R.; Li, L.; Chan, W. & Wetzel, R. (1994). A role for destabilizing amino

Ibrahim, H.R.; Kato, A. & Kobayashi, K. (1991). Antimicrobial effects of lysozyme against

Ibrahim, H.R.; Kobayashi, K. & Kato, A. (1993). Length of hydrocarbon chain and

Izzo, H.V.; Lincoln, M.D. & Ho, C.-T. (1993). The effect of temperature, feed moisture, and

Jameson, B.A. & Wolf, H. (1988). The antigenic index: a novel algorithm for predicting

Jansen, J.J.N.; Kardinaal, A.F.M.; Huijbers, G.; Vlieg-Boerstra, B.J.; Martens, B.P.M. &

antigenic determinants. *Comp. Appl. Biosci.* 4: 181-186.

population. *J. Allergy Clin. Immun.* 93: 446-456.

formaldehyde and -hydroxyadipaldehyde with bovine serum albumin and casein.

iodoacetate, iodoacetamide, and thiol compounds on coagulation. *J. Biol. Chem.* 179:

lysozyme/-casein mixtures at the air/water interface. *J. Colloid Interface Sci.* 143:

acid replacements in light-chain amyloidosis. *Proc. Natl. Acad. Sci. U.S.A.* 91: 5446-

gram-negative bacteria due to covalent binding of palmitic acid. *J. Agric. Food Chem.*

antimicrobial action to gram-negative bacteria of fatty acylated lysozyme. *J. Agric.* 

pH on protein deamidation in an extruded wheat flour. *J. Agric. Food Chem.* 41: 199-

Ockhuizen, T. (1994). Prevalence of food allergy and intolerance in the adult Dutch

the folding of glycoproteins. *Trends Cell Biology* 7: 193-200.

sensitivity and on aggregation. *Biochemistry* 7: 2542-2546.

estimating protein digestibility. *J. Food Sci.* 42: 1269-1273.

powders by heat. *J. Sci. Food Agric.* 31: 1152-1157.

proteolytic enzymes. *Biochemistry* 6: 541-547.

18: 290-295.

*Histochemie* 17: 151-161.

645-654.

37-53.

5450.

202.

39: 2077-2082.

*Food Chem.* 41: 1164-1168.

368.

sulfhydryl groups to thermal properties of ovalbumin. *Can. Inst. Food Sci. Technol. J.* 

the inv vivo activity of trichosanthin by site-directed PEGylation. *Life Sci.* 65: 355-


Application Potential of Food Protein Modification 173

Lomakin, A.; Chung, D.S.; Benedek, G.B.; Kirscher, D.A. & Teplow, D.B. (1996). On the

quantitation of rate constants. *Proc. Natl. Acad. Sci. U.S.A.* 93: 1125-1129. Lomakin, A.; Teplow, D.B.; Kirschner, D.A. & Benedek, G.B. (1997). Kinetic theory of fibrillogenesis of amyloid -protein. *Proc. Natl. Acad. Sci. U.S.A.* 94: 7942-7947. Longares, A.; Monahan, F.J.; O'Riordan, E.D. & O'Sullivan, M. (2005). Physical properties of

Lönnerdal, B. (2002). Expression of human milk proteins in plants. *J. Am. Coll. Nutr.* 21: 218–

Loomis, W.D. (1974). Overcoming problems of phenolics and quinones in the isolation of

Losso, J.N. & Nakai, S. (2002). Stabilization of oil-in-water emulsions by -lactoglobulin-

Lubelski, J.; Rink, R.; Khusainov, R.; Moll, G.N. & Kuipers, O.P. (2008). Biosynthesis,

Luey, J.-K.; McGuire, J. & Sproull, R.D. (1991). The effect of pH and NaCl concentration on

Luheshi, L.M. & Dobson, C.M. (2009). Bridging the gap: from protein misfolding to protein

Malaki Nik, A.; Wright, A.J. & Corredig, M. (2010). Interfacial design of protein-stabilized

Marcon, G.; Plakoutsi, G.; Canale, C.; Relini, A.; Taddei, N.; Dobson, C.M.; Ramponi, G. &

Margoshes, B.A. (1990). Correlation of protein sulfhydryls with the strength of heat-formed

Martin, A.H.; Meinders, M.B.J.; Bos, M.A.; Cohen Stuart, M.A. & van Vliet, T. (2003).

Martins, J.T.; Cerqueira, M.A.; Souza, B.W.; Carmo Avides, M. & Vicente, A.A. (2010). Shelf

Masuda, T.; Ueno, Y. & Kitabatake, N. (2001). Sweetness and enzymatic activity of

Masuda, T.; Ide, N. & Kitabatake, N. (2005a). Structure-sweetness relationship in egg white

Chiti, F. (2005). Amyloid formation from HypF-N under conditions in which the

Conformational aspects of proteins at the air/water interface studied by infrared

life extension of ricotta cheese using coatings of galactomannans from nonconventional sources incorporating nisin against *Listeria monocytogenes*. *J. Agric.* 

lysozyme: role of lysine and arginine on the elicitation of lysozyme sweetness.

Ma, C.-Y. (1984). Functional properties of acylated oat protein. *J. Food Sci.* 49: 1128-1131. Ma, C.-Y. & Wood, D.F. (1987). Functional properties of oat proteins modified by acylation, trypsin hydrolysis or linoleate treatment. *J. Am. Oil Chem. Soc.* 64: 1726-1731.

emulsions for optimal delivery of nutrients. *Food Funct.* 1: 141-148.

protein is initially in its native state. *J. Mol. Biol.* 347: 323-335.

reflection-absorption spectroscopy. *Langmuir* 19: 2922-2928.

polyethylene glycol conjugates. *J. Agric. Food Chem.* 50: 1207-1212.

1255-1260.

Academic Press: New York, p. 528.

nisin. *Cell. Mol. Life Sci.* 65: 455-476.

*Colloid Interface Sci.* 143: 489-500.

egg white gels. *J. Food Sci.* 55: 1753.

lysozyme. *J. Agric. Food Chem.* 49: 4937-4941.

*Food Chem.* 58: 1884-1891.

*Chem. Senses* 30: 667-681.

misfolding diseases. *FEBS Lett.* 583: 2581-2586.

MacRitchie, F. (1978). Proteins at interfaces. *Adv. Protein Chem.* 32: 283-326.

221

nucleation and growth of amyloid -protein fibrils: detection of nuclei and

edible films made from mixtures of sodium caseinate and WPI. *Int. Dairy J.* 15:

plant enzymes organelles. In: *Methods in Enzymology*; Fleischer, S., Packer, L., Eds.:

immunity, regulation, mode of action and engineering of the model lantibiotic

adsorption of -lactoglobulin at hydrophilic and hydrophobic silicon surfaces. *J.* 


Koseki, T.; Kitabatake, N. & Doi, E. (1989). Irreversible thermal denaturation and formation

Kosters, H.A. & de Jongh, H.H. (2003). Spectrophotometric tool for the determination of the

Kosters, H.A.; Broersen, K.; de Groot, J.; Simons, J.W.; Wierenga, P. & de Jongh, H.H. (2003).

Kotaki, A.; Harada, M. & Yagi, K. (1964). Reaction between sulfhydryl compounds and 2,4,6-

Krause, J.-P.; Mothes, R. & Schwenke, K.D. (1996). Some physicochemical and interfacial

Krebs, M.R.H.; Devlin, G.L. & Donald, A.M. (2009). Amyloid fibril-like structure underlies

Kudryashova, E.V.; Meinders, M.B.J.; Visser, A.J.W.G.; van Hoek, A. & de Jongh, H.H.J

Kudryashova, E.V.; Visser, A.J.W.G., & de Jongh, H.H.J. (2005). Reversible self-association of

Lakkis, J. & Villota, R. (1992). Effect of acylation on substructural properties of proteins: a study using fluorescence and circular dichroism. *J. Agric. Food Chem.* 40: 553-560. Land, A. & Braakman, I. (2001). Folding of the human immunodeficiency virus type 1 envelope glycoprotein in the endoplasmatic reticulum. *Biochimie* 83: 783-790. Langton, M. & Hermansson, A.-M. (1992). Fine-stranded and particulate gels of lactoglobulin and whey protein at varying pH. Food Hydrocoll. 5: 523-539. Le, T.T.; Bhandari, B. & Deeth, H.C. (2011). Chemical and physical changes in milk protein concentrate (MPC80) powder during storage. *J. Agric. Food Chem.* 59: 5465-5473. Le Floch-Fouéré, C.; Pezennec,S.; Pézolet, M.; Rioux-Dubé, J.F.; Renault, A. & Beaufils, S.

adsorbed at the air/water interaface. *Eur. J. Biophys.* 32: 553-562.

interface at pH 6.5 and 8.0. *J. Colloid Interface Sci.* 356: 614-623.

conformation of wheat gluten. *J. Sci. Food Agric.* 90: 409-417.

and middle chain fatty acids. *J. Agric. Food Chem.* 48: 265-269.

proteins. *J. Am. Chem. Soc.* 108: 5543-5548.

total carboxylate content in proteins; molar extinction coefficient of the enol ester from Woodward's reagent K reacted with protein carboxylates. *Anal. Chem.* 75:

Chemical processing as a tool to generate ovalbumin variants with changed

properties of native and acetylated legumin from faba beans (*Vicia faba* L.). *J. Agric.* 

the aggregate structure across the pH range for -lactoglobulin. *Biophys. J.* 96: 5013-

(2003). Structural properties and rotational dynamics of egg white ovalbumin

ovalbumin at air-water interfaces and the consequences for the exerted surface

(2011). Unexpected differences in the behavior of ovotransferrin at the air-water

deamidation-induced modification on functional and nutritional properties and

sulfonate (Woodward's reagent K) as a reagent for nucleophilic side chains of

Lehrer, S.B.; Ayuso, R. & Reese, G. (2002). Current understanding of food allergens. *Ann. N.* 

Liao, L.; Zhao, M.; Ren, J.; Zhao, H.; Cui, C. & Hu, X. (2010). Effect of acetic acid

Liener, I.E. (1980). Toxic constituents of plant foodstuffs. New York and London, Academic

Liu, S.-T.; Sugimoto, T.; Azakami, H. & Kato, A. (2000). Lipophilization of lysozyme by short

Llamas, K.; Ownes, M.; Blakeley, L. & Zerner, B. (1986). *N*-ethyl-5-phenylisoxazolium-3'-

of linear aggregates of ovalbumin. *Food Hydrocoll.* 3: 123-134.

trinitrobenzene-1-sulfonic acid. *J. Biochem.* 55: 553-561.

2512-2516.

5019.

stability. *Biotechnol. Bioeng.* 84: 61-70.

pressure. Protein Science 14: 483-493.

*Food Chem.* 44: 429-437.

*Y. Acad. Sci.* 964: 69-85.

Press, 502 p.


Application Potential of Food Protein Modification 175

Morris, J.A. & Cagan, R.H. (1972). Purification of monellin, the sweet principle of

Mossine, V.V.; Glinsky, G.V. & Feather, M.S. (1994). The preparation and characterization of

Moure, A.; Sineiro, J.; Domínguez, H. & Parajó, J.C. (2006). Functionality of oilseed protein

Nakai, S. (1983). Structure-function relationships of food proteins with an emphasis on the importance of protein hydrophobicity. *J. Agric. Food Chem.* 31: 676-683. Nakai, S. & Li-Chan, E. (1988). Hydrophobic interaction in food systems. CRC Press: Boca

Narayana, K. & Rao, N. (1991). Effect of acetylation and succinylation on the

Niestijl Jansen J.J.; Kardinaal, A.F.M.; Huijbers, G.; Vlieg-Boerstra, B.J.; Martens, B.P.M. &

Nimni, M.E., Cheung, D.; Strates, B.; Kodama, M. & Sheikh, K. (1987). Chemically modified

Nizet, V. (2006). Antimicrobial peptide resistance mechanisms of human bacterial

Ohta, K.; Masuda, T.; Ide, N. & Kitabatake, N. (2008). Critical molecular regions for

O'Kane, F.E.; Happe, R.P.; Vereijken, J.M.; Gruppen, H. & van Boekel, M.A.J.S. (2004). Heat-

Okuyama, T. & Satake, K. (1960). On the preparation and properties of 2,4,6-trinitrophenyl-

Onuchic, J.N.; Luthey-Schulten, Z. & Wolynes, P.G. (1997). Theory of protein folding: the

Owusu Apenten, R.K.; Chee, C. & Hwee, O.P. (2003). Evaluation of a sulphydryl-disulphide

Pabit, S.A.; Roder, H. & Hagen, S.J. (2004). Internal friction controls the speed of protein

Paik, W. & Kim, S. (1972). Effect of methylation on susceptibility of proteins to proteolytic

Panasiuk, R.; Amarowicz, R.; Kostyra, H. & Sijtsma, L. (1998). Determination of -amino

Paparcone, R. & Buehler, M.J. (2011). Failure of A(1-40) amyloid fibrils under tensile

folding from a compact configuration. *Biochemistry* 43: 12532-12538.

energy landscape perspective. *Ann. Rev. Phys. Chem.* 48: 545-600.

physicochemical properties of winged bean (Psophocarpus tetragonolobus)

Ockhuizen, T. (1994). Prevalence of food allergy and intolerance in the adult Dutch

collagen: a natural biomaterial for tissue replacement. *J. Biomed. Mater. Res.* 21: 741-

elicitation of the sweetness of sweet-tasting protein, thaumatin I. *FEBS J.* 275: 3644-

induced gelation of pea legumin: comparison with soy glycinin. *J. Agric. Food Chem.*

exchange index (SEI) for whey proteins – beta-lactoglobulin and bovine serum

nitrogen in pea protein hydrolysates: a comparison of three analytical methods.

some Amadori compounds (1-amino-1-deoxy-D-fructose derivatives) derived from

*Dioscoreophyllum cumminsii*. *Biochem. Biophys. Acta* 261: 114-122.

reaction. *Nature* 419: 448.

Raton, FL.

771.

3652.

52: 5071-5078.

products: a review. *Food Res. Int.* 39: 945-963.

proteins. *J. Agric. Food Chem.* 39: 259-261.

pathogens. *Curr. Issues Mol. Biol.* 8: 11-26.

population. *J. Allergy Clin. Immunol.* 93: 446-456.

amino acids and –peptides. *J. Biochem.* 47: 454-466.

albumin. *Food Chem.* 83: 541-545.

enzymes. *Biochemistry* 11: 2589-2594.

loading. *Biomaterials* 32: 3367-3374.

*Food Chem.* 62: 363-367.

a series of aliphatic omega-amino acids. *Carbohydr. Res.* 262: 257-270. Mottram, D.S.; Wedzicha, B.L. & Dodson, A.T. (2002). Acrylamide is formed in the Maillard


Masuda, T.; Ide, N. & Kitabatake, N. (2005b). Effects of chemical modification of lysine

Matoba, T. & Doi, E. (1979). In vitro digestibility of succinylated protein by pepsin and

Matsudomi, N.; Kato, A. & Kobayashi, K. (1982). Conformation and surface properties of

Matsudomi, N.; Sasaki, T.; Kato, A. & Kobayashi, K. (1985). Conformational changes and

Maurer-Stroh, S.; Debulpaep, M.; Kuemmerer, N.; Lopez de la Paz, M.; Martins, I.C.;

structure using position-specific scoring matrices. *Nature Methods* 7: 237-242. Marquardt, T. & Helenius, A. (1992). Misfolding and aggregation of newly synthesized

McClellan, S.J. & Franses, E.I. (2003). Effect of concentration and denaturation on adsorption

McClements, J. (2005). Food emulsions: principles, practice, and techniques. 2nd Ed. Boca

McMahon, D.J.; Adams, S.L. & McManus, W.R. (2009). Hardening of high-protein nutrition bars and sugar/polyol-protein phase separation. *J. Food Sci.* 74: E312-E321. Means, G.E. & Feeney, R.E. (1971). Chemical modifications of proteins. Holden-Day, San

Meinders, M.B.J.; van den Bosch, G.G.M. & de Jongh, H.H.J. (2001). Adsorption properties of

Melnychyn, P. & Stapley, R. (1973). Acylated protein from coffee whitener formulations.

Mendis, E.; Rajapakse, N. & Kim, S. (2005). Antioxidant properties of a radical-scavenging

Metcalfe, D.D.; Astwood, J.D.; Townsend, R.; Sampson, H.A.; Taylor, S.L. & Fuchs, R.L.

Migneault, I.; Dartiguenave, C.; Bertrand, M.J. & Waldron, K.C. (2004). Gluteraldehyde:

Mine, Y. (1992). Sulfhydryl groups changes in heat-induced soluble egg white aggregates in

Ming, D. & Hellekant, G. (1994). Brazzein, a new high-potency thermostable sweet protein

Mirmoghatadaie, L.; Kadivar, M. & Shahedi, M. (2009). Effects of succinylation and

Moore, S. & Stein, W.H. (1948). Photometric ninhydrin method for use in the

Morel, B.; Varela, L. & Conejero-Lara, F. (2010). The thermodynamic stability of amyloid fibrils studied by differential scanning calorimetry. *J. Phys. Chem. B* 114: 4010-4019.

engineered crop plants. *Crit. Rev. Fd. Sci. Nutr.* 36: S165-S186.

proteins in the endoplasmatic reticulum. *J. Cell Biol.* 117: 505-513.

functional properties of acid-modified soy protein. *Agric. Biol. Chem.* 49: 1251-1256.

Reumers, J.; Morris, K.L.; Copland, A.; Serpell, L.C.; Serrano, L.; Schymkowitz, J.W.H. & Rousseau, F. (2010). Exploring the sequence determinants of amyloid

and surface tension of bovine serum albumin. *Colloids Surf, B Biointerfaces* 28: 63-75.

proteins at and near the air/water interface from IRRAS spectra of protein

peptide purified from enzymatically prepared fish skin gelatin hydrolysate. *J.* 

(1996). Assessment of the allergenic potential of foods derived from genetically

behavior in aqueous solution, reaction with proteins, and application to enzyme

deamidation on functional properties of oat protein isolate. *Food Chem.* 114: 127-131.

residues on the sweetness of lysozyme. *Chem. Senses* 30: 253-264.

pancreatic proteases. *J. Food Sci.* 44: 537-539.

Raton, FL: CRC Press.

U.S. Patent 3,764,711.

*Agric. Food Chem.* 53: 581-587.

solutions. *Eur. Biophys. J.* 30: 256-267.

crosslinking. *BioTechniques* 37: 790-802.

relation to molecular size. *J. Food Sci.* 57: 254-255.

from *Pentadiplandra brazzeana* B. *FEBS Lett.* 355: 106-108.

chromatography of amino acids. *J. Biol. Chem.* 176: 367-388.

Francisco.

deamidated gluten. *Agric. Biol. Chem.* 46: 1583-1586.


Application Potential of Food Protein Modification 177

Ponce, A.G.; Roura, S.I.; Del Valle, C.E. & Moreira, M. (2008). Antimicrobial and antioxidant

Promeyrat, A.; Bax, M.L.; Traoré, S.; Abry, L.; Santé-Lhoutellier, V. & Gatellier, Ph. (2010).

Qiu, L. & Hagen, S.J. (2004). A limiting speed for protein folding at low solvent viscosity. *J.* 

Quintanilla-Guerrero, F.; Duarte-Vázquez, M.A.; Tinoco, R.; Gómez-Suárez, M.; García-

Quintas, A.; Saraiva M.J. & Brito, R.M. (1997). The amyloidogenic potential of transthyretin

Quiocho, F.A. & Richards, F.M. (1966). The enzymic behaviour of carboxypeptidase-A in the

Rahma, E.H. & Narasinga Rao (1983). Effect of acetylation and succinylation of cottonseed

Ramírez-Alvarado, M.; Merkel, J.S. & Regan, L. (2000). A systematic exploration of the

Ramírez-Jiménez, A.; Guerra-Hernández, E. & García-Villanova, B. (2000). Browning

Richard, F.M. & Knowles, J.R. (1968). Gluteraldehyde as a protein cross-linkage reagent. *J.* 

Riha, W.E.; Izzo, H.V.; Zhang, J. & Ho, C.T. (1996). Nonenzymatic deamidation of food

Robinson, A.B. & Rudd, C. (1974). Deamidation of glutaminyl and asparaginyl residues in

Rodríguez Patino, J.M.; Carrera Sánchez, C. & Rodríguez Niño, M.R. (2008). Implications of

Rondeau, P.; Navarra, G.; Cacciabaudo, F.; Leone, M.; Bourdon, E. & Militello, V. (2010).

Rondeau, P.; Armenta, S.; Caillens, H.; Chesne, S. & Bourdon, E. (2007). Assessment of

Röper, H.; Röper, S.; Heyns, K. & Meyer, B. (1983). N.M.R. spectroscopy of N-(1-deoxy-Dfructos-1-yl)-L amino acids ("fructose-amino acids"). *Carbohydr. Res.* 116: 183-195. Rosén, J. & Hellenäs, K.E. (2002). Analysis of acrylamide in cooked foods by liquid

chromatography tandem mass spectrometry. *Analyst* 127: 880-882. Roth, M. (1971). Fluorescence reaction for amino acids. *Anal. Chem.* 43: 880-882.

interfacial characteristics of food foaming agents in foam formulations. *Adv. Colloid* 

Thermal aggregation of glycated bovine serum albumin. *Biochim. Biophys. Acta*

temperature effects on -aggregation of native and glycated albumin by FTIR spectroscopy and PAGE: relations between structural changes and antioxidant

flour on its functional properties. *J. Agric. Food Chem.* 31: 351-355.

indicators in bread. *J. Agric. Food Chem.* 48: 4176-4181.

peptides and proteins. *Curr. Top. Cell. Regul.* 8: 247-295.

proteins. *Crit. Rev. Food Sci. Nutr.* 6: 225-255.

properties. *Arch. Biochem. Biophys.* 460: 141-150.

vivo studies. *Postharvest. Biol. Technol.* 49: 294-300.

time and temperature. *Meat Sci.* 85: 625-631.

*Am. Chem. Soc.* 126: 3398-3399.

*Food Chem.* 56: 8058-8065.

*Sci. U.S.A.* 97: 8979-8984.

*Mol. Biol.* 37: 231-233.

*Interf. Sci.* 140: 95-113.

1804: 789-798.

solid state. *Biochemistry* 5: 4062-4076.

300.

activities of edible coatings enriched with natural plant extracts: in vitro and in

Changed dynamics in myofibrillar protein aggregation as a consequence of heating

Almendárez, B.E.; Vazquez-Duhalt, R. & Regalado, C. (2008). Chemical modification of turnip peroxidase with methoxypolyethylene glycol enhances activity and stability for phenol removal using the immobilized enzyme. *J. Agric.* 

variants correlates with their tendency to aggregate in solution. *FEBS Lett.* 418: 297-

influence of protein stability on amyloid fibril formation in vitro. *Proc. Natl. Acad.* 


Paparcone, R.; Keten, S. & Buehler, M.J. (2010). Atomistic simulation of nanomechanical

Park, H.J.; Chinnan, M.S. & Shewfelt, R.L. (1994). Edible corn-zein film coatings to extend

Park, H.J. (1999). Development of advanced edible coatings for fruits. *Trends Food Sci.* 

Patel, K. & Borchardt, R.T. (1990a). Chemical pathways of peptide degradation. II. Kinetics

Patel, K. & Borchardt, R.T. (1990b). Chemical pathways of peptide degradation. III. Effect of

Patil, S.M.; Mehta, A.; Jha, S. & Alexandrescu, A.T. (2011). Heterogeneous amylin fibril

Payne, P.I.; Holt, L.M.; Jackson, E.A. & Law, C.N. (1984). Wheat storage proteins: their

Pedroche, J.; Yust, M.M.; Lqari, H.; Giron-Calle, J.; Alaiz, M.; Vioque, J. & Millan, F. (2004).

Pellegrino, L.; van Boekel, M.A.J.S.; Gruppen, H.; Resmini, P. & Pagani, M.A. (1999). Heat-

Pereda, M.; Aranguren, M.I. & Marcovich, N.E. (2008). Characterization of

Pereda, M.; Aranguren, M.I. & Marcovich, N.E. (2009). Water vapor absorption and

Petersen, A.; Schramm, G.; Schlaak, M. & Becker, W.M. (1998). Post-translational

Pétra, P.H. (1971). Modification of carboxyl groups in bovine carboxypeptidase A. I.

Plaxco, K.W.; Simons, K.T.; Ruczinski, I. & Baker, D. (2000). Topology, stability, sequence,

Plaxco, K.W.; Simons, K.T. & Baker, D. (1998). Contact order, transition state placement and the refolding rates of single protein domains. *J. Mol. Biol.* 277: 985-994. Plotkin, S.S. & Onuchic, J.N. (2002). Understanding protein folding with energy landscape

Pogaku, R.; Eng Seng, C.; Boonbeng, L. & Kallu, U. (2007). Whey protein isolate-starch

chitosan/caseinate films. *J. Appl. Polymer Sci.* 107: 1080-1090.

(Woodward's reagent K). *Biochemistry* 10: 3163-3170.

theory. Part I: Basic concepts. *Quart. Rev. Biophys.* 35: 111-167.

system – a critical review. *Intl. J. Food Eng.* 3: 104-119.

storage life of tomatoes. *J. Food Process Preserv.* 18: 317-331.

loading. *J. Biomechanics* 43: 1196-1201.

hexapeptides. *Pharm. Res.* 7: 787-793.

pollen. *Clin. Exp. Allergy* 28: 315-321.

*Biochemistry* 39: 11177-11183.

*Biochemistry* 50: 2808-2819.

*Lond. B* 304: 359-371.

*Technol.* 10: 254-260.

703-711.

337-346.

9: 255-260.

111: 2777-2784.

properties of Alzheimer's A(1-40) amyloid fibrils under compressive and tensile

of deamidation of an asparaginyl residue in a model hexapeptide. *Pharm. Res.* 7:

primary sequence on the pathways of deamidation of asparaginyl residues in

growth mechanisms imaged by total internal reflection fluorescence microscopy.

genetics and their potential for manipulation by plant breeding. *Phil. Trans. R. Soc.* 

*Brassica carinata* protein isolates: chemical composition, protein characterization and improvement of functional properties by protein hydrolysis. *Food Chem.* 88:

induced aggregation and covalent linkages in -casein model systems. *Int. Dairy J.*

permeability of films based on chitosan and sodium caseinate. *J. Appl. Polym. Sci.*

modifications influence IgE reactivity to the major allergen Phl p 1 of timothy grass

Inactivation of the enzyme by *N*-ethyl-5-phenylisoxazolium-3'-sulfonate

and length: defining the determinants of two-state protein folding kinetics.


Application Potential of Food Protein Modification 179

Shimada, K. & Cheftel, J.C. (1989). Sulfhydryl group/disulfide bond interchange reactions

Shirahama, H.; Lyklema, J. & Norde, W. (1990). Comparative protein adsorption in model

Siepen, J.A. & Westhead, D.R. (2002). The fibril\_one on-line database: mutations,

Silva Freitas, D. & Abrahão-Neto, J. (2010). Batch purification of high-purity lysozyme from

Sinha, U. & Brewer, J.M. (1985). A spectrophotometric method for quantitation of carboxyl

Siu, M. & Thompson, L.U. (1982). In vitro and in vivo digestibilities of succinylated cheese

Skraup, Z.H. & Kaas, K. (1906). Über die Einwirkung von salpetriger Säure auf Ovalbumin.

Smythe, C.V. (1936). The reaction of iodoacetate and of iodoacetamide with various

Soderling, T.R. (1975). Regulation of glycogen synthetase. Specificity and stoichiometry of

Song, Y.; Azakami, H.; Hamasu, M. & Kato, A. (2001). In vivo glycosylation suppresses the

Sorci, M.; Silkworth, W.; Gehan, T. & Belfort, G. (2011). Evaluating nuclei concentration in

Stadler, R.H.; Blank, I.; Varga, N.; Robert, F.; Hau, J.; Guy, P.A.; Robert, M.-C. & Riediker, S. (2002). Acrylamide from Maillard reaction products. *Nature* 419: 449. Straub, J.E. & Thirumalai, D. (2011). Toward a molecular theory of early and late events in monomer to amyloid fibril formation. *Annu. Rev. Phys. Chem.* 62: 437-463. Sun, Y.; Hayakawa, S. & Izumori, K. (2004). Modification of ovalbumin with a rare

Sun, Y. & Hayakawa, S. (2002). Heat-induced gels of egg white/ovalbumins from five avian

Suttiprasit, P.; Krisdhasima, V. & McGuire, J. (1992). The surface activity of -lactalbumin, -

single-component and mixed solutions. *J. Colloid Interface Sci.* 154: 316-326. Tareke, E.; Rydberg, P.; Karlsson, P.; Eriksson, S. & Törnqvist, M. (2002). Analysis of

protein concentrates. *J. Agric. Food Chem.* 30: 743-747.

*Justus Liebigs. Ann. Chemie* 351: 379-389.

protein kinase. *J. Biol. Chem.* 250: 5407-5412.

properties. *J. Agric. Food Chem.* 52: 1293-1299.

properties. *J. Agric. Food Chem.* 50: 1636-1642.

systems. *J. Colloid Interf. Sci.* 139: 177-187.

*Protein Sci.* 11: 1862-1866.

168.

48: 554-562.

327-333.

612.

5006.

*Lett.* 491: 63-66.

during heat-induced gelation of whey protein isolate. *J. Agric. Food Chem.* 3: 161-

experimental conditions, and trends associated with amyloid fibril formation.

egg white and characterization of the enzyme modified by PEGylation. *Pharm. Biol.*

group modification of proteins using Woodward's reagent K. *Anal. Biochem.* 151:

sulfhydryl groups, with urease, and with yeast preparations. *J. Biol. Chem.* 114: 601-

phosphorylation of the skeletal muscle enzyme by cyclic 3':5'-AMP-dependent

aggregation of amyloidogenic hen egg white lysozymes expressed in yeast. *FEBS* 

amyloid fibrillation reactions using back-calculation approach. *PLoS One* 6: e20072.

ketohexose through the Maillard reaction: effect on protein structure and gel

species: thermal aggregation, molecular forces involved, and rheological

lactoglobulin and bovine serum albumin. I. Surface tension measurements with

acrylamide, a carcinogen formed in heated foodstuffs. *J. Agric. Food Chem.* 50: 4998-


Rudd, P.M.; Colominas, C.; Royle, L.; Murphy, N.; Hart, E.; Merry, A.H.; Hebestreit, H.F. &

Rudén, C. (2004). Acrylamide and cancer risk – expert risk assessments and the public

Sala, G.; de Wijk, R.A.; van de Velde, F. & van Aken G.A. (2008). Matrix properties affect the sensory perception of emulsion-filled gels. *Food Hydrocolloids* 22: 353-363. Salgó, A.; Granzler, K. & Jecsai, J. (1984). Simple enzymatic methods for prediction of plant

Salunkhe, D.K.; Jadhav, S.J.; Kadam, S.S. & Chavan, J.K. (1982). Chemical, biochemical, and

Samejima, K.; Dairman, W. & Underfriend, S. (1971). Condensation of ninhydrin with

Sanchez-Ruiz, J.M. (2011). Probing free-energy surfaces with differential scanning

Sancho, A.I.;. Rigby, N.M.; Zuidmeer, L.; Asero, R.; Mistrello, G.; Amato, S.; González-

Santé-Lhoutellier, V.; Astruc, T.; Marinova, P.; Greve, E. & Gatellier, P. (2008). Effect of meat

Satake, K.; Okuyama, T.; Ohashi, M. & Shinoda, T. (1960). The spectrophotometric

Sawyer, W.H. (1968). Heat denaturation of bovine -lactoglobulins and relevance of

Schilling, E.D.; Burchill, P.I. & Clayton, R.A. (1963). Anomalous reactions of ninhydrin. *Anal.* 

Schwenke, K.D.; Knopfe, C.; Mikheeva, L.M. & Grinberg, V.Y. (1998). Structural changes of

Sheih, I.C.; Wu, T.K. & Fang, T.J. (2009). Antioxidant properties of a new antioxidative

Shental-Bechor, D. & Levy, Y. (2008). Effect of glycosylation on protein folding: a close look at thermodynamic stabilization. *Proc. Natl. Acad. Sci. U.S.A.* 105: 8256-8261. Shih, F.F. (1990). Deamidation during treatment of soy protein with protease. *J. Food Sci.* 55:

285-294.

*Nutr.* 17: 277-305.

*Biochem.* 42: 222-247.

debate. *Food Chem. Tox.* 42: 335-349.

Symp. (pp. 311-321). Budapest: Akadémiai Kiadó.

calorimetry. *Annu. Rev. Phys. Chem.* 62: 231-255.

Sanger, F. (1945). The free amino groups of insulin. *Biochem. J.* 39: 507-515.

from apple, Mal d 3. *Allergy* 60: 1262-1268.

disulfide aggregation. *J. Dairy Sci.* 51: 323-329.

*J. Agric. Food Chem.* 56: 1488-1494.

sulfonic acid. *J. Biochem.* 47: 654.

*Bioresour. Technol.* 100: 3419-3425.

*Biochem.* 5: 1-6.

2080-2086.

127-129.

Dwek, R.A. (2001). A high-performance liquid chromatography based strategy for rapid, sensitive sequencing of N-linked oligosaccharide modifications to proteins in sodium dodecyl sulphate polyacrylamide electrophoresis gel bands. *Proteomics* 1:

protein digestibility. In: R. Lásztity, M. Hidvé. Proc. Int. Assoc. Cereal Chem.

biological significance of polyphenols in cereals and legumes. *Crit. Rev. Food Sci.* 

aldehydes and primary amines to yield highly fluorescent ternary products. *Anal.* 

Mancebo, E.; Fernández-Rivas, M.; van Ree, R. & Mills, E.N.C. (2005). The effect of thermal processing on the IgE reactivity of the non-specific lipid transfer protein

cooking on physicochemical state and in vitro digestibility of myofibrillar proteins.

determination of amine, amino acid and peptide with 2,4,6-trinitrobenzene-1-

legumin from Faba beans (*Vicia faba L.*) by succinylation. *J. Agric. Food Chem.* 46:

peptide from algae protein waste hydrolysate in different oxidation systems.


Application Potential of Food Protein Modification 181

Waniska, R.D. & Kinsella, J.E. (1985). Surface properties of -lactoglobulin: adsorption and rearrangement during film formation. *J. Agric. Food Chem.* 33: 1143-1148. Weijers, M.; Broersen, K.; Barneveld, P.A.; Cohen Stuart, M.A.; Hamer, R.J.; de Jongh, H.H.

Wierenga, P.A.; Kosters, H.; Egmond, M.R.; Voragen, A.G.J. & de Jongh, H.H.J. (2006).

Wierenga, P.A.; Meinders, M.B.J.; Egmond, M.R.; Voragen, F.A.G.J. & de Jongh, H.H.J.

Wierenga, P.A.; Meinders, M.B.; Egmond, M.R.; Voragen, A.G. & de Jongh, H.H. (2005).

Wilde, P.J. (2000). Interfaces: their role in foam and emulsion behaviour. *Curr. Opin. Colloid* 

Wilde, P.J.; Mackie, A.; Husband, F.; Gunning, P. & Morris, V. (2004). Proteins and emulsifiers at liquid interfaces. *Adv. Colloid Interface Sci.* 108-109: 63-71. Wimley, W.C. (2010). Describing the mechanism of antimicrobial peptide action with the

Wolynes, P.G. (2005). Recent successes of the energy landscape theory of protein folding and

Woodward, R.B. & Olofson, R.A. (1961). The reaction of isoxazolium salts with bases. *J. Am.* 

Woodward, R.B.; Olofson, R.A. & Mayer, H. (1961). A new synthesis of peptides. *J. Am.* 

Wright, H.T. & Urry, D.W. (1991). Nonenzymatic deamidation of asparaginyl and glutaminyl residues in proteins. *Crit. Rev. Biochem. Mol. Biol.* 26: 41-52. Xu, L.J.; Sheldon, B.W.; Larick, D.K. & Carawan, R.E. (2002). Recovery and utilization of

Xu, Z.; Paparcone, R. & Buehler, M.J. (2010). Alzheimer's A(1-40) amyloid fibrils feature

Yemm, E.W. & Cocking, E.C. (1955). The determination of amino acids with ninhydrin.

Zamora, R. & Hildalgo, F.J. (2006). Coordinate contribution of lipid oxidation and Maillard reaction to the nonenzymatic food browning. *Crit. Rev. Food Sci. Nutr.* 45: 49-59. Zavodsky, M.; Chen, C.W.; Huang, J.K.; Zolkiewski, M.; Wen, L. & Krishnamoorthi, R.

Zhang, J.; Lee, T.C. & Ho, C.-T. (1993). Comparative study on kinetics of nonenzymatic

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

size-dependent mechanical properties. *Biophys. J.* 96: 2053-2062.

*maxima* trypsin inhibitor-V. *Protein Sci.* 10: 149-160.

by raman spectroscopy. *J. Agric. Food Chem.* 52: 1815-1823.

useful by-products from egg processing wastewater by electrocoagulation. *Poult.* 

(2001). Disulfide bond effects on protein stability: designed variants of *Cucurbita* 

deamidation of soy protein and egg white lysozyme. *J. Agric. Food Chem.* 41: 2286-

ovalbumin aggregates. *Biomacromolecules* 9: 3165-3172.

ovalbumin to the air-water interface. *Langmuir* 19: 8964-8970.

interfacial activity model. *ACS Chem. Biol.* 5: 905-917.

function. *Quart. Rev. Biophys.* 38: 405-410.

adsorption to air-water interface. *J. Phys. Chem. B* 109: 16946-16952.

*Colloid Interface Sci.* 119: 131-139.

*Interface Sci.* 5: 176-181.

*Chem. Soc.* 83: 1007-1009.

*Chem. Soc.* 83: 1010-1012.

*Sci.* 81: 785-792.

*Analyst* 80: 209-213.

2290.

& Visschers, R.W. (2008). Net charge affects morphology and visual properties of

Importance of physical vs. chemical interactions in surface shear rheology. *Adv.* 

(2003). Protein exposed hydrophobicity reduces the kinetic barrier for adsorption of

Quantitative description of the relation between protein net charge and protein


Tawfik, D.S. (2002). Chapter 63: Side chain selective chemical modifications of proteins. In:

Teplow, D.B.; Lazo, N.D.; Bitan, G.; Bernstein, S.; Wyttenbach, T.; Bowers, M.T.;

Thomas, M.E.C.; Scher, J.; Desobry-Banon, S. & Desobry, S. (2004). Milk powders ageing: effect on physical and functional properties. *Crit. Rev. Food Sci. Nutr.* 44: 297-322. Tyler-Cross, R. & Schirch, V. (1991). Effects of amino acid sequence, buffers, and ionic

Van den Berg, L.; Carolas, A.L.; van Vliet, T.; van der Linden, E.; van Boekel, M.A.J.S. & van

Van Teeffelen, A.M.M.; Broersen, K. & de Jongh, H.H.J. (2005). Glucosylation of -

van Vliet, T. & Walstra, P. (1995). Large deformation and fracture behaviour of gels. *Faraday* 

Veerman, C.; Ruis, H.; Sagis, L.M.C.; van der Linden, E. (2002). Effect of electrostatic

Visschers, R.W. & de Jongh, H.H.J. (2005). Disulphide bond formation in food protein

Vioque, J.; Sánchez-Vioque, R.; Clemente, A.; Pedroche, J.; Bautista, J. & Millán, F.J. (1999).

Wada, R. & Kitabatake, N. (2001). -Lactoglobulin A with *N*-ethylmaleimide-modified

heating in the presence of dithiothreitol. *J. Agric. Food Chem.* 49: 4971-4976. Walder, R. & Schwartz, D.K. (2010). Single molecule observations of multiple protein

Walsh, D.M.; Lomakin, A.; Benedek, G.B.; Condron, M.M. & Teplow, D.B. (1997). Amyloid

Wanasundara, P.K.J.P.D. & Shahidi, F. (1997). Functional properties of acylated flax protein

Wang, C.; Eufemi, M.; Turano, C. & Giartosio, A. (1996). Influence of the carbohydrate

Waniska, R.D. & Kinsella, J.E. (1979). Foaming properties of proteins: evaluation of a column

moiety on the stability of glycoproteins. *Biochemistry* 35: 7299-7307.

aeration apparatus using ovalbumin. *J. Food Sci.* 44: 1398-1411.

populations at the oil-water interface. *Langmuir* 26: 13364-13367.

proteins/polysaccharide mixed gels. *Food Hydrocolloids* 22: 1404-1417. Van der Wel, H. & Loeve, K. (1972). Isolation and characterization of thaumatin I and II, the

Press Inc. Totowa, NJ, pp 465-467.

approach. *Acc. Chem. Res.* 39: 635-645.

225.

*Discussions* 101: 359-370.

*Biomacromolecules* 3: 869-873.

*Am. Oil Chem. Soc.* 76: 819-823.

isolates. *J. Agric. Food Chem.* 45: 2431-2441.

272: 22364-22372.

small peptides. *J. Biol. Chem.* 266: 22549-22556.

protein thermodynamics. *Protein Sci.* 14: 2187-2194.

aggregation and gelation. *Biotechnol. Adv.* 23: 75-80.

The Protein Protocols Handbook, 2nd Edition, Edited by J.M. Walker, Humana

Baumketner, A.; Shea, J.E.; Urbanc, B.; Cruz, L.; Borreguero, J. & Stanley, H.E. (2006). Elucidating amyloid- protein folding and assembly: a multidisciplinary

strength on the rate and mechanism of deamidation of asparagines residues in

de Velde, F. (2008). Energy storage controls crumbly perception in whey

sweet-tasting proteins from *Thaumatococcus danielli* Benth. *Eur. J. Biochem.* 31: 221-

lactoglobulin lowers the heat capacity change of unfolding; a unique way to affect

interactions on the percolation concentration of fibrillar -lactoglobulin gels.

Production and characterization of an extensive rapeseed protein hydrolysate. *J.* 

sulfhydryl residue, polymerized through intermolecular disulfide bridge on



**Part 2** 

**Catalysis and Reaction Engineering** 

