**1. Introduction**

384 Mass Transfer - Advanced Aspects

Kolpin, D., Furlong, E., Meyer, M., Thurman, E., Zaugg, S. & Barber, L. (2002).

Lietz, A. & Meyer, M. (2006). Evaluation of emerging contaminants of concern at the South

Ottmar, K., Colosi, L. & Smith, J. (2010a). Sorption of statin pharmaceuticals to wastewater

Richardson, S & Ternes, T. (2005). Water analysis: Emerging contaminants and current issues. *Analytical Chemistry*, Vol. 77, No. 12, (December 2005), pp. 3807-3838, ISSN Vanderford, B. & Snyder, S. (2006). Analysis of pharmaceuticals in water by isotope dilution

Yu, C. & Chu, K. (2009). Occurrence of pharmaceuticals and personal care products along

Yu, J.; Bouwer, E. & Coelhan, M. (2006). Occurrence and biodegradability studies of selected

*Management*, Vol. 86, No. 1-2, (January/February 2006), pp. 72-80, ISSN

*Technology*, Vol. 40, No. 23, (December 2006), pp. 7312-7320, ISSN

Vol. 40, No. 5, (May 2006), pp. 1686-1696, ISSN

2006-5240. Reston, Virginia.

2010), pp. 507-512, ISSN

No. 5, (May 2009), pp. 1281-1286, ISSN

*Technology*, Vol. 36, No. 6, (June 2002), pp. 1202-1211, ISSN

municipal wastewater treatment: Proposing a classification scheme. *Water Research*,

Pharmaceuticals, hormones, and other organic wastewater contaminants in U.S. streams, 1999-2000: A national reconnaissance. *Environmental Science and* 

District Wastewater Treatment Plant based on seasonal sampling events, Miami-Dade County, Florida, 2004. *U.S. Geological Survey, Scientific Investigations Report* 

treatment biosolids, terrestrial soils, and freshwater sediment. *Journal of Environmental Engineering*, Vol. 136, No. 2, (February 2010), pp. 256-264, ISSN Ottmar, K., Colosi, L. & Smith, J. (2010b). Development and application of a model to

estimate wastewater treatment plant prescription pharmaceutical influent loadings and concentrations. Bulletin of Environmental Contamination, Vol. 84, No. 5, (May

liquid chromatography/tandem mass spectrometry. *Environmental Science and* 

the West Prong Little Pigeon River in east Tennessee, USA. *Chemosphere*, Vol. 75,

pharmaceuticals and personal care products in sewage effluent. *Agricultural Water* 

The wine sector has faced increasing pressure in order to fulfill the legal environmental requirements, maintaining a competitive position in a global market. The rising costs associated have stimulated the sector to seek sustainable management's strategies, focussing on controlling the demand for water and improving its supply. These can be accomplished by defining the best practical techniques, using technological means (Best Available Technologies) (Duarte *et al*., 2004). Some EU Directives were implemented concerning water protection and management. These included in particular the Framework Directive in the field of water policy and environmental legislation about specific uses of water and discharges of substances. The disposal of the untreated waste from the wine sector is considered an environmental risk, causing salination and eutrophication of water resources; waterlogging and anaerobiosis and loss of soil structure with increased vulnerability to erosion (Schoor, 2005). The winery wastewater is seasonally produced and is generated mainly as the result of cleaning practices in winery, such as washing operations during crushing and pressing grapes, rinsing of fermentations tanks, barrels washing, bottling and purges from the cooling process. As a consequence of the working period and the winemaking technologies, volumes and pollution loads greatly vary over the year. Each winery is also unique in wastewater generation, highly variable, 0.8 to 14 L per litre of wine (Schoor, 2005; Moletta, 2009). Consequently, the treatment system must be versatile to face the loading regimen and stream fluctuation. During the peak season (vintage), the winery wastewater has a very high loading of solids and soluble organic contaminant, but after this period, contaminant load decreases substantially. The high concentration of ethanol and sugars in winery wastewater justifies often the choice of a biological treatment (Bolzonella & Rosso, 2007). But the different wine processing method of each winery generates wastewater with specific properties, causing the impossibility to meet a general agreement on the most suitable cost-effective alternative for biological treatment of this wastewater.

Several winery wastewater treatments are available, but the development of alternative technologies is essential to increase their efficiency and to decrease the investment and exploration costs (Coetzee *et al.*, 2004). So criteria should be considered in the selection of the adequate technology, such as maximization of removal efficiency, flexibility in order to deal with variable concentration and loads, moderate capital cost, easy to operate and maintain, small footprint, ability to meet discharge requirements for winery wastewater and also low sludge production. On the other hand, small producer with relatively modest financial

Winery Wastewater Treatment - Evaluation of the Air Micro-Bubble Bioreactor Performance 387

post-treatment, to return the treated water to environment (Moletta 2005) in addition, the

The MBRs are very compact systems and offer an alternative to conventional activated sludge processes. The COD efficiency achieved is above 97%. The electricity consumption and the operating life of the membranes are higher than those associated with traditional activated sludge systems (Artiga et al., 2005), what may constitute a constraint to its

The subsurface-flow constructed wetland is described as suitable for treating these wastewaters, but frequently wineries do not have available area for setting up such plants (Grismer *et al*., 2003). However, phytotoxicity bioassays carried out with *Phragmites, Juncus and Schoenoplectus* at different wastewater dilutions showed that at greater than 25% wastewater concentration all the macrophytes died (Arienzo *et al*., 2009a). Nevertheless, the same authors showed that this system when combined with a previously sedimentation/aerobic process could be used for small wineries located in rural areas,

The wastewater treatment of small wineries (less than 15,000 hL of wine per year) can be also performed using a SBR, fed once a day. The SBR system is a modified design of conventional activated sludge process and it has been widely used in industrial wastewater treatment. The COD removal efficiency is between 86-99% (Torrijos & Moletta, 1997). The on-line monitoring of dissolved oxygen concentration appeared as a good indicator of the progress in the COD biodegradation (Andreottola *et al*., 2002). Some modifications have been done in order to improve the reactor performance. The opportunity of combining the advantages of the SBR with fixed biomass was investigated (Andreottola & *et al*., 2002). This system permits the treatment of high organic loads, 6.3 kg COD m-3 d-1 with high biofilm grown (4-5 kg TSS m-3), allowing the reduction of the required volume for biological treatment and avoiding bulking problems. However, the degradation of organic matter present in a winery wastewater sometimes require the addition of extra nutrients, to balance the C/N/P ratio and some oxygen efficiency transfer problems were detected when higher

The fixed bed biofilm reactor or the air-bubble column bioreactor using self-adapted microbial population either free or immobilized can achieve 90% of COD removal (Petruccioli *et al*., 2000). In order to overcome the energetic costs associated with the aeration systems, a Venturi injector was used in the JLRs. This system achieves COD removal efficiency near 90%. Though, the high shear stress applied on the Venturi influence the composition of the microbial population (Petruccioli *et al*., 2000; Eusébio *et al*., 2005) leading to settling sludge problems. A similar technology that utilizes also a Venturi injector is the AMBB. This technology is very promising because it consists in a vertical reactor with good oxygen transfer and high biological conversion capacity. To optimise the mass transfer, a highly efficient Venturi injector coupled with multiplier nozzles was patented (AirJection®) and was applied in a lagoon system (Meyer *et al*., 2004) and in a vertical reactor (Oliveira *et al.,*

2009), at pilot scale, to treat winery wastewater with a treatment efficiency of 90 %.

The maintenance and enhancement of a biological reactor is highly dependent on the microbial population that changes with time and winery activity (Jourjon *et al*., 2005). A deep understanding on the microbial population involved in the process is crucial to address any strategy for treatment system management (Tandoi *et al*., 2006). Although some researchers have been developed (Eusébio *et al*., 2004; Eusébio *et al*., 2005; Jourjon *et al*.,

process control needs specialized personnel (Malandra *et al*., 2003).

achieving 72% of COD removal rate (Arienzo *et al*., 2009b).

organic loads were applied (Lopez-Palau & Mata-Alvarez, 2009).

application.

capacity are interested in simple treatment systems with low maintenance and manpower requirements (Andreottola *et al*., 2009).

Most treatment systems have been designed with large oxidation tanks and oversizing the aeration system to deal with the peak load with a very high oxygen demand, during the vintage period. As a result, wastewater treatment plants are quite large and difficult to manage. One of the most promising technologies appears to be the vertical reactors characterised by high oxygen mass transfer improving the biological conversion capacity. To optimise the mass transfer, a highly efficient Venturi injector coupled with multiplier nozzles were developed (AirJection®), in order to increase the treatment efficiency.

The main goals of the present paper are the comparison of different biological treatment systems, in particular fixed and suspended biomass, operating under aerobic conditions. Since the accurate design of the bioreactor is dependent on many operational parameters, aspects related to hydraulic retention time; oxygen mass transfer and contact time, energetic costs; sludge settling and production; response time during startup, flexibility and treated wastewater reuse, in crop irrigation, with the aim of closing the water cycle in the wine sector, will be addressed. A new treatment system will be presented as a case study, an air micro-bubble bioreactor (AMBB), that will highlight the advantages and constraints on its performance at bench-scale and full-scale, in order to fulfill the gaps associated with the implemented winery wastewater treatment systems. The data presented was collected during four years monitoring plan and used to develop a tool to support the selection of the best available technology. The present study will also contribute to the implementation of an integrated strategy for sustainable production in the wine sector, based on a modular and flexible technology that will facilitate compliance with environmental regulations and potential reuse for crop irrigation. This approach will contribute to the development of a bio-based economy in the wine sector that should be integrated in a Green Innovation Economy Cycle.

#### **2. Comparison of different biological treatment systems**

#### **2.1 Biological treatments in winery wastewater**

Several treatment systems, both physico-chemical and biological, have been assayed to reduce the organic load of the winery wastewater. Some of these technologies are based on membrane bioreactors (MBRs), sequencing batch reactor (SBR), upflow anaerobic sludge blanket (UASB), anaerobic sequencing batch reactor (ASBR) and jet loop reactors (JLR). However, most of these methods have some characteristics in common: they are relatively expensive, they are not applicable in all situations, and they are not always able to deal with uctuations in the hydraulic and pollution load. In order to overcome some of these problems, research efforts have been made towards the development of novel bioreactors as alternatives or to improve, the above-mentioned conventional methods. Although the high organic load of this wastewater would recommend the application of an anaerobic treatment for removing its polluting content, several problems have been found in the application of anaerobic processes due to its seasonal nature, its variable volumes and compositions and the difficulties in the monitoring and process control by specialized personnel (Malandra *et al*., 2003). The anaerobic treatments such as UASB and ASBR have successfully been used to treat a variety of effluents including those from wineries. The chemical oxygen demand (COD) removal efficiency is greater than 90%, but a specific microbial community is required. However, normally after this reactor there is an aerobic

capacity are interested in simple treatment systems with low maintenance and manpower

Most treatment systems have been designed with large oxidation tanks and oversizing the aeration system to deal with the peak load with a very high oxygen demand, during the vintage period. As a result, wastewater treatment plants are quite large and difficult to manage. One of the most promising technologies appears to be the vertical reactors characterised by high oxygen mass transfer improving the biological conversion capacity. To optimise the mass transfer, a highly efficient Venturi injector coupled with multiplier

The main goals of the present paper are the comparison of different biological treatment systems, in particular fixed and suspended biomass, operating under aerobic conditions. Since the accurate design of the bioreactor is dependent on many operational parameters, aspects related to hydraulic retention time; oxygen mass transfer and contact time, energetic costs; sludge settling and production; response time during startup, flexibility and treated wastewater reuse, in crop irrigation, with the aim of closing the water cycle in the wine sector, will be addressed. A new treatment system will be presented as a case study, an air micro-bubble bioreactor (AMBB), that will highlight the advantages and constraints on its performance at bench-scale and full-scale, in order to fulfill the gaps associated with the implemented winery wastewater treatment systems. The data presented was collected during four years monitoring plan and used to develop a tool to support the selection of the best available technology. The present study will also contribute to the implementation of an integrated strategy for sustainable production in the wine sector, based on a modular and flexible technology that will facilitate compliance with environmental regulations and potential reuse for crop irrigation. This approach will contribute to the development of a bio-based economy in the wine sector that should be integrated in a Green Innovation

Several treatment systems, both physico-chemical and biological, have been assayed to reduce the organic load of the winery wastewater. Some of these technologies are based on membrane bioreactors (MBRs), sequencing batch reactor (SBR), upflow anaerobic sludge blanket (UASB), anaerobic sequencing batch reactor (ASBR) and jet loop reactors (JLR). However, most of these methods have some characteristics in common: they are relatively expensive, they are not applicable in all situations, and they are not always able to deal with uctuations in the hydraulic and pollution load. In order to overcome some of these problems, research efforts have been made towards the development of novel bioreactors as alternatives or to improve, the above-mentioned conventional methods. Although the high organic load of this wastewater would recommend the application of an anaerobic treatment for removing its polluting content, several problems have been found in the application of anaerobic processes due to its seasonal nature, its variable volumes and compositions and the difficulties in the monitoring and process control by specialized personnel (Malandra *et al*., 2003). The anaerobic treatments such as UASB and ASBR have successfully been used to treat a variety of effluents including those from wineries. The chemical oxygen demand (COD) removal efficiency is greater than 90%, but a specific microbial community is required. However, normally after this reactor there is an aerobic

nozzles were developed (AirJection®), in order to increase the treatment efficiency.

**2. Comparison of different biological treatment systems** 

**2.1 Biological treatments in winery wastewater** 

requirements (Andreottola *et al*., 2009).

Economy Cycle.

post-treatment, to return the treated water to environment (Moletta 2005) in addition, the process control needs specialized personnel (Malandra *et al*., 2003).

The MBRs are very compact systems and offer an alternative to conventional activated sludge processes. The COD efficiency achieved is above 97%. The electricity consumption and the operating life of the membranes are higher than those associated with traditional activated sludge systems (Artiga et al., 2005), what may constitute a constraint to its application.

The subsurface-flow constructed wetland is described as suitable for treating these wastewaters, but frequently wineries do not have available area for setting up such plants (Grismer *et al*., 2003). However, phytotoxicity bioassays carried out with *Phragmites, Juncus and Schoenoplectus* at different wastewater dilutions showed that at greater than 25% wastewater concentration all the macrophytes died (Arienzo *et al*., 2009a). Nevertheless, the same authors showed that this system when combined with a previously sedimentation/aerobic process could be used for small wineries located in rural areas, achieving 72% of COD removal rate (Arienzo *et al*., 2009b).

The wastewater treatment of small wineries (less than 15,000 hL of wine per year) can be also performed using a SBR, fed once a day. The SBR system is a modified design of conventional activated sludge process and it has been widely used in industrial wastewater treatment. The COD removal efficiency is between 86-99% (Torrijos & Moletta, 1997). The on-line monitoring of dissolved oxygen concentration appeared as a good indicator of the progress in the COD biodegradation (Andreottola *et al*., 2002). Some modifications have been done in order to improve the reactor performance. The opportunity of combining the advantages of the SBR with fixed biomass was investigated (Andreottola & *et al*., 2002). This system permits the treatment of high organic loads, 6.3 kg COD m-3 d-1 with high biofilm grown (4-5 kg TSS m-3), allowing the reduction of the required volume for biological treatment and avoiding bulking problems. However, the degradation of organic matter present in a winery wastewater sometimes require the addition of extra nutrients, to balance the C/N/P ratio and some oxygen efficiency transfer problems were detected when higher organic loads were applied (Lopez-Palau & Mata-Alvarez, 2009).

The fixed bed biofilm reactor or the air-bubble column bioreactor using self-adapted microbial population either free or immobilized can achieve 90% of COD removal (Petruccioli *et al*., 2000). In order to overcome the energetic costs associated with the aeration systems, a Venturi injector was used in the JLRs. This system achieves COD removal efficiency near 90%. Though, the high shear stress applied on the Venturi influence the composition of the microbial population (Petruccioli *et al*., 2000; Eusébio *et al*., 2005) leading to settling sludge problems. A similar technology that utilizes also a Venturi injector is the AMBB. This technology is very promising because it consists in a vertical reactor with good oxygen transfer and high biological conversion capacity. To optimise the mass transfer, a highly efficient Venturi injector coupled with multiplier nozzles was patented (AirJection®) and was applied in a lagoon system (Meyer *et al*., 2004) and in a vertical reactor (Oliveira *et al.,* 2009), at pilot scale, to treat winery wastewater with a treatment efficiency of 90 %.

The maintenance and enhancement of a biological reactor is highly dependent on the microbial population that changes with time and winery activity (Jourjon *et al*., 2005). A deep understanding on the microbial population involved in the process is crucial to address any strategy for treatment system management (Tandoi *et al*., 2006). Although some researchers have been developed (Eusébio *et al*., 2004; Eusébio *et al*., 2005; Jourjon *et al*.,

Winery Wastewater Treatment - Evaluation of the Air Micro-Bubble Bioreactor Performance 389

dDO/dt

During the assays the DO control was the key to the COD removal, because this treatment was carried out with constant aeration (up to 4.5 hours). When the treatment started the COD decreases as the DO concentration is maintained at low levels (Figure 1). Once the microbial activity decreases by diminishing the organic load, the DO concentration begins to rise until reaching a plateau. At this stage, the process is complete and the cycle can be stopped. The end of each cycle can be calculated based on the first derivative function of the DO concentration *vs* time (Figure 2). With this strategy it was possible to reduce the hydraulic retention time in about three times, which has allowed the treatment of a higher

0123456

Another approach based on dissolved oxygen control was carried out to optimize a SBR cycle for total organic carbon and ammonia removal (Puig *et al*., 2006). In this treatment the aerobic phases of the SBR cycle were initially operated using an On/Off dissolved oxygen

The cycle was divided in reaction phase, under aerobic and anoxic conditions, settling and

Aerobic Anoxic Aerobic

0 30 60 90 120 150

0

10

20

30

DO

OUR

40

50

60

Fig. 2. Time derivative of DO concentration

0

5

10

15

20

25

flow with a similar effluent quality.

control strategy.

Adapted from Puig *et al*. (2006).

Fig. 3. DO and OUR evolution during aerobic and anoxic phases

discharge.

2005), the understanding of the microflora dynamics inside the bioreactor will be of utmost importance for the treatment system optimisation. Moreover, in the aerobic bioreactors the microorganisms are dependent on aeration oxygen supply. Knowledge of mass transfer coefficients between the different phases together with reaction dynamics is utmost importance to design gas-liquid-solid reactor and to predict the microbial metabolism pathway.

#### **2.2 Optimization of operational parameters in aerobic reactors**

The optimization of operational parameters in bioprocesses is based essentially on reducing the volume and footprint, oxygen mass transfer and contact time, energetic costs, sludge settling and production and response time during start-up, while maintain a high removal efficiency of organic matter.

The SBR system has been widely applied to organic carbon removal in municipal and industrial wastewater treatments, as this system presents different advantages such as space reduction and the ability to make operational changes, during the treatment cycle.

In the SBR system the sludge settlement occur in the same tank as oxidation, so in order to optimize the sludge settling time, the formation of granules could be performed based on feast-famine periods (Lopez-Palau *et al*., 2009). The start-up were performed with the increasing of the COD loading (2.7-20 kg COD. m-3 day-1) in order to reach the feast period. After ten days of operation, the first aggregates were observed. But, the use of a high organic load promotes microbial growth and the reactor reached solids concentration of around 6 g VSS L-1. Consequently, some problems of aeration appeared, and the air supply had to be increased from 13.5 L min-1 to 20 L min-1. This study showed that is possible to cultivate aerobic granular sludge in SBR, improving the sludge settleability. Nevertheless, the aeration must be proportional to the COD load.

The combination of the SBR with fixed biomass SBBR (Sequencing Batch Biofilm Reactor) to treat winery wastewater was studied by Andreottola *et al.* (2002) and revealed the possibility of treating higher organic loads without increasing the required treatment volume, as the biomass grown on plastic media. However, this type of reactor needs a separated settler, as the biomass settlement worsens in the presence of the plastic material. In order to optimize the energetic costs and the SBBR performance, a strategy based on dissolved oxygen (DO) monitoring was developed.

Adapted from Andreottola *et al.* (2002)

Fig. 1. DO concentration and COD dynamic during a typical SBBR cycle

Fig. 2. Time derivative of DO concentration

2005), the understanding of the microflora dynamics inside the bioreactor will be of utmost importance for the treatment system optimisation. Moreover, in the aerobic bioreactors the microorganisms are dependent on aeration oxygen supply. Knowledge of mass transfer coefficients between the different phases together with reaction dynamics is utmost importance to design gas-liquid-solid reactor and to predict the microbial metabolism

The optimization of operational parameters in bioprocesses is based essentially on reducing the volume and footprint, oxygen mass transfer and contact time, energetic costs, sludge settling and production and response time during start-up, while maintain a high removal

The SBR system has been widely applied to organic carbon removal in municipal and industrial wastewater treatments, as this system presents different advantages such as space

In the SBR system the sludge settlement occur in the same tank as oxidation, so in order to optimize the sludge settling time, the formation of granules could be performed based on feast-famine periods (Lopez-Palau *et al*., 2009). The start-up were performed with the increasing of the COD loading (2.7-20 kg COD. m-3 day-1) in order to reach the feast period. After ten days of operation, the first aggregates were observed. But, the use of a high organic load promotes microbial growth and the reactor reached solids concentration of around 6 g VSS L-1. Consequently, some problems of aeration appeared, and the air supply had to be increased from 13.5 L min-1 to 20 L min-1. This study showed that is possible to cultivate aerobic granular sludge in SBR, improving the sludge settleability. Nevertheless,

The combination of the SBR with fixed biomass SBBR (Sequencing Batch Biofilm Reactor) to treat winery wastewater was studied by Andreottola *et al.* (2002) and revealed the possibility of treating higher organic loads without increasing the required treatment volume, as the biomass grown on plastic media. However, this type of reactor needs a separated settler, as the biomass settlement worsens in the presence of the plastic material. In order to optimize the energetic costs and the SBBR performance, a strategy based on dissolved oxygen (DO)

> 0123456 DO concentration COD removal

0

20

40

60 80

100

reduction and the ability to make operational changes, during the treatment cycle.

**2.2 Optimization of operational parameters in aerobic reactors** 

the aeration must be proportional to the COD load.

pathway.

efficiency of organic matter.

monitoring was developed.

Adapted from Andreottola *et al.* (2002)

Fig. 1. DO concentration and COD dynamic during a typical SBBR cycle

During the assays the DO control was the key to the COD removal, because this treatment was carried out with constant aeration (up to 4.5 hours). When the treatment started the COD decreases as the DO concentration is maintained at low levels (Figure 1). Once the microbial activity decreases by diminishing the organic load, the DO concentration begins to rise until reaching a plateau. At this stage, the process is complete and the cycle can be stopped. The end of each cycle can be calculated based on the first derivative function of the DO concentration *vs* time (Figure 2). With this strategy it was possible to reduce the hydraulic retention time in about three times, which has allowed the treatment of a higher flow with a similar effluent quality.

Another approach based on dissolved oxygen control was carried out to optimize a SBR cycle for total organic carbon and ammonia removal (Puig *et al*., 2006). In this treatment the aerobic phases of the SBR cycle were initially operated using an On/Off dissolved oxygen control strategy.

The cycle was divided in reaction phase, under aerobic and anoxic conditions, settling and discharge.

Adapted from Puig *et al*. (2006).

Fig. 3. DO and OUR evolution during aerobic and anoxic phases

Winery Wastewater Treatment - Evaluation of the Air Micro-Bubble Bioreactor Performance 391

empirical equations and also theoretical prediction, most of them developed for bubble

The bioprocesses involves simultaneous transport and biochemical reactions, the oxygen is transferred from a rising gas bubble to the liquid phase and then to the place of oxidative phosphorylation within the cell, considered as a solid particle. The steps related to this mass transfer processes can be represented according to the film theory model for mass transfer,

Liquid Phase

bubble Bulk liquid Liquid film

which describes the flux through the film based on a driving force (Figure 5).

 Liquid film surrounding

Fig. 5. Steps and resistances for oxygen transfer from gas bubble to cell, in three phases

Lenght of transport (Z)

1/k 1/k 1/k 1/k

G L B S

The oxygen mass transfer rate per unit of reactor volume is obtained by a solute mass

As kL and *a* are difficult to measure separately, usually the kLa is evaluated together and this parameter is identified as the volumetric mass transfer coefficient that characterizes the gasliquid mass transfer. The driving force is the gradient between the oxygen concentration at the interface and in the bulk liquid. This gradient varies with the solubility and microbial activity. Also, the gas solubility depends on temperature, pressure, concentration and type

In bioreactors it is essential to determine the experimental kLa to set the aeration efficiency and to quantify the effects of the operating variables on the dissolved oxygen supply. To select the appropriated method, some factors should be taken into account, such as aeration

ܱܴܶ ൌ ݇ܽ ൈ ሺכܥ െ ܥሻ (1)

 L-S Interface

citoplasm

1/k

C

Solid particle

 Site O reaction 2

surrounding Cells

columns and airlifts (Garcia-Ochoa & Gomez, 2009).

 G-L Interface

Gas Phase

Bubbles Gas film

Adapted from Garcia-Ochoa & Gomez (2009)

of salts present in the system.

balance for the liquid phase (Fakeeha *et al.,* 1999):

reactors

Oxygen concentration

Adapted from Puig *et al*. (2006).

Fig. 4. Detection of the ammonia valley in the pH evolution during aerobic phase

During the aerobic phase a fixed DO set-point of 2.0 mg DO L-1 was applied, as a simple On/Off control. The system optimization was based on pH, DO and OUR evolution. This strategy allows the detection of the ammonia valley in the pH profile and also the end of nitrification, through the OUR outline (Figure 3 and Figure 4). The analysis of the OUR profile shows a plateau in the OUR value, in the end of the aerobic phase, which may indicate that the microbial populations are under endogenous conditions and that organic matter and ammonia has been completed degraded.

In fact, one of the most important aspects in many biological systems is the aeration supply. The wastewater treatment is one of these processes that require proper aeration to maintain the growth of the microorganisms responsible for the biodegradation of the organic matter. Most wastewater treatments are aerobic and are carried out in aqueous medium containing inorganic salts and organic substances which can give viscosity to the broth, showing a non-Newtonian behavior. In bioprocessing it is very important to ensure an adequate oxygen distribution to the gas stream and to the fermentation broth. Some of the systems used to supply the oxygen are sparging, free-jet flow and bubbling column, among others. Also the different nozzle geometry, the liquid phase properties, the jet length and diameter influences the oxygen distribution to the system that in many cases is a limiting factor to the success of the treatment process. In this sense, it is important to estimate the mass transfer characteristics in order to predict the kinetic growth reaction constant, and control and optimize the aerobic fermentation processes (Choi *et al*., 1996; Fakeeha *et al*., 1999; Tojabas & Garcia-Calvo, 2000; Garcia-Ochoa & Gomez, 2009). The volumetric mass transfer coefficient, kLa, is the parameter that characterizes the gas-liquid mass transfer in bioreactors. However, this value can vary substantially from those obtained for the oxygen absorption in water or in simple aqueous solutions, and in static systems with invariable composition of the liquid media along time. The transfer rate is very influenced by the nature of pollutants present in the wastewater, for example glucose increases the medium viscosity causing a decrease in the kLa value while the low foam surfactants enhances this value (Fakeeha *et al*., 1999; Tojabas & Garcia-Calvo, 2000). Thus, it is necessary to know the composition of the fermentative broth, at least some of the major compounds, to understand the effect of combination of different pollutants for proper design and operation of aerobic process. Many strategies have been proposed to determine the volumetric mass transfer coefficient, empirical equations and also theoretical prediction, most of them developed for bubble columns and airlifts (Garcia-Ochoa & Gomez, 2009).

The bioprocesses involves simultaneous transport and biochemical reactions, the oxygen is transferred from a rising gas bubble to the liquid phase and then to the place of oxidative phosphorylation within the cell, considered as a solid particle. The steps related to this mass transfer processes can be represented according to the film theory model for mass transfer, which describes the flux through the film based on a driving force (Figure 5).

Adapted from Garcia-Ochoa & Gomez (2009)

390 Mass Transfer - Advanced Aspects

Fig. 4. Detection of the ammonia valley in the pH evolution during aerobic phase

During the aerobic phase a fixed DO set-point of 2.0 mg DO L-1 was applied, as a simple On/Off control. The system optimization was based on pH, DO and OUR evolution. This strategy allows the detection of the ammonia valley in the pH profile and also the end of nitrification, through the OUR outline (Figure 3 and Figure 4). The analysis of the OUR profile shows a plateau in the OUR value, in the end of the aerobic phase, which may indicate that the microbial populations are under endogenous conditions and that organic

0 30 60 90 120 150

pH

In fact, one of the most important aspects in many biological systems is the aeration supply. The wastewater treatment is one of these processes that require proper aeration to maintain the growth of the microorganisms responsible for the biodegradation of the organic matter. Most wastewater treatments are aerobic and are carried out in aqueous medium containing inorganic salts and organic substances which can give viscosity to the broth, showing a non-Newtonian behavior. In bioprocessing it is very important to ensure an adequate oxygen distribution to the gas stream and to the fermentation broth. Some of the systems used to supply the oxygen are sparging, free-jet flow and bubbling column, among others. Also the different nozzle geometry, the liquid phase properties, the jet length and diameter influences the oxygen distribution to the system that in many cases is a limiting factor to the success of the treatment process. In this sense, it is important to estimate the mass transfer characteristics in order to predict the kinetic growth reaction constant, and control and optimize the aerobic fermentation processes (Choi *et al*., 1996; Fakeeha *et al*., 1999; Tojabas & Garcia-Calvo, 2000; Garcia-Ochoa & Gomez, 2009). The volumetric mass transfer coefficient, kLa, is the parameter that characterizes the gas-liquid mass transfer in bioreactors. However, this value can vary substantially from those obtained for the oxygen absorption in water or in simple aqueous solutions, and in static systems with invariable composition of the liquid media along time. The transfer rate is very influenced by the nature of pollutants present in the wastewater, for example glucose increases the medium viscosity causing a decrease in the kLa value while the low foam surfactants enhances this value (Fakeeha *et al*., 1999; Tojabas & Garcia-Calvo, 2000). Thus, it is necessary to know the composition of the fermentative broth, at least some of the major compounds, to understand the effect of combination of different pollutants for proper design and operation of aerobic process. Many strategies have been proposed to determine the volumetric mass transfer coefficient,

Adapted from Puig *et al*. (2006).

6.65

6.70

6.75

6.80

matter and ammonia has been completed degraded.

Fig. 5. Steps and resistances for oxygen transfer from gas bubble to cell, in three phases reactors

The oxygen mass transfer rate per unit of reactor volume is obtained by a solute mass balance for the liquid phase (Fakeeha *et al.,* 1999):

$$
\hat{\mathbf{L}}\hat{\mathbf{C}}\mathbf{R} = k\_L \mathbf{a} \times \left(\mathbf{C}^\* - \mathbf{C}\_L\right) \tag{1}
$$

As kL and *a* are difficult to measure separately, usually the kLa is evaluated together and this parameter is identified as the volumetric mass transfer coefficient that characterizes the gasliquid mass transfer. The driving force is the gradient between the oxygen concentration at the interface and in the bulk liquid. This gradient varies with the solubility and microbial activity. Also, the gas solubility depends on temperature, pressure, concentration and type of salts present in the system.

In bioreactors it is essential to determine the experimental kLa to set the aeration efficiency and to quantify the effects of the operating variables on the dissolved oxygen supply. To select the appropriated method, some factors should be taken into account, such as aeration

Winery Wastewater Treatment - Evaluation of the Air Micro-Bubble Bioreactor Performance 393

can be minimized using different strategies (Pérez-Elvira *et al.* 2006), such as endogenous metabolism and maintenance metabolism. In this last approach part of energy source is used for maintaining living functions, in this phase the substrate consumption is not used for cellular synthesis. In the endogenous metabolism part of cellular components is oxidized to produce the required energy for maintenance functions, which leads to a decrease in the biomass production. The objective is to reach a natural balance between biomass growth and decay rates. The oxic-settling-anoxic activated sludge process, considered as a sludge feast/famine treatment, is based on alternating exposure of sludge to oxic and anoxic environments. This working principle stimulate catabolic activity and make catabolism dissociate from anabolism. The sludge famine is related to an exposure of the settled sludge to anoxic conditions where the substrate concentration is low. Under these stressful conditions microorganisms are starving which may lead to a depletion of cell energy or nutrients storage. The sludge feasting means that fasted microorganisms return to an oxic environment with enough nutrients. As a consequence, the microorganisms growth may be

The selection of the most appropriate technology for the winery wastewater treatment is a difficult step that should be done after a diagnosis process. A proper diagnosis should conduct a survey report that includes all the information required for decision-makers. Regarding the production process, it should address all activities associated with it: vintage, racking and bottling (Figure 6). The knowledge of materials and supplies, as well as byproducts generated during the process is essential in diagnosis. The water uses and water consumption are critical, both in terms of quantity or quality. The survey of sewers in the farm unit, particularly if the drainage system is separated or combined, and the points of wastewater discharge should also be covered. The wastewater flows should be evaluated through the installation of flow meters. The different streams of wastewater generated must

The water consumption in two Portuguese wineries, one small and one medium size are quite different, with regard to quantity. However, the distribution of water consumption has a similar behavior throughout the year (Figure 7 and Figure 8). The data presented show that most water (60%-80%) is consumed in the vintage period that last about a month. So, the collection of water consumption associated with the physicochemical characterization of the wastewater is essential for the proper sizing of any treatment system. In addition, it is

be quantified in order to make an assessment, as rigorous as possible.

limited by energy uncoupling (Chen *et al*., 2001).

**2.3 Diagnosis process** 

Fig. 6. Winery activities

system; bioreactor type and its mechanical design; the composition of the fermentation broth and the possible effect of the microorganisms (Xu *et al*., 2010).

The mass balance for the dissolved oxygen in the well-mixed liquid phase can be established as (Garcia-Ochoa & Gomez, 2009; Irizar *et al*., 2009):

$$\frac{d\mathcal{C}}{dt} = OTR - OUR\tag{2}$$

Where dC/dt is the accumulation oxygen rate in the liquid phase, OTR is the oxygen transfer rate from the gas to the liquid phase, described by equation (1) and OUR is the oxygen uptake rate by microorganisms. The methods that can be applied for the oxygen transfer rate measures can be classified depending on whether the measurement is done in the absence of microorganisms or with dead cells or in the presence of biomass that consumes oxygen at the time of measurement. When biochemical reactions do not take place, OUR=0, then the equation (2) can be simplified to:

$$\frac{d\mathcal{C}}{dt} = k\_L a \times (\mathcal{C}^\* - \mathcal{C}\_L) \tag{3}$$

The dynamic method used to measure the kLa value is based on the dissolved oxygen consumption and supply. In this method the change in the dissolved oxygen concentration is analyzed supplying air until the oxygen saturation concentration in the liquid phase is reached. The oxygen decreasing is then recorded as a function of time. Under these conditions the equation (2) can be expressed as equation (4), but, after the decreasing phase, the oxygen is again supplied and the equation (2) can be written as equation (5). In these cases the kLa values can be determined from the slope of the *ln f(CL) vs* time.

$$\ln\left(\frac{\mathcal{C}\_{L0}}{\mathcal{C}\_{L}}\right) = -k\_L a \times t \tag{4}$$

$$\ln\left(1-\frac{C\_L}{C^\*}\right) = -k\_L a \times t \tag{5}$$

Furthermore kLa is usually expressed at standard conditions of temperature and pressure, 20ºC, 1atm (equation 6).

$$k\_{L}a\_{20} = k\_{L}a\_{T} \times 1.024^{20-T} \tag{6}$$

The determination of the oxygen uptake rate OUR can also be carried out using a dynamic method which measures the respiratory activity of microorganisms that grow in the bioreactor. When the air supply is switching off, the dissolved oxygen concentration will decrease at a rate equal to oxygen consumption due to the microorganisms respiration rate. In this situation the OUR is determined from the slope of the plot of dissolved oxygen concentration *vs* time. The biomass concentration should be known in order to determined the specific oxygen uptake rate (SOUR).

Another important parameter in the aerobic reactors optimization is the sludge settling and production. The large amount of excess sludge generated during activated sludge process is estimated to cost about 40-60 % of the operating cost (Chen *et al*., 2001). This sludge contains volatile solids and retains about 95% of water resulting in a large volume of residual solids produced. The biological sludge production in conventional wastewater treatment plants can be minimized using different strategies (Pérez-Elvira *et al.* 2006), such as endogenous metabolism and maintenance metabolism. In this last approach part of energy source is used for maintaining living functions, in this phase the substrate consumption is not used for cellular synthesis. In the endogenous metabolism part of cellular components is oxidized to produce the required energy for maintenance functions, which leads to a decrease in the biomass production. The objective is to reach a natural balance between biomass growth and decay rates. The oxic-settling-anoxic activated sludge process, considered as a sludge feast/famine treatment, is based on alternating exposure of sludge to oxic and anoxic environments. This working principle stimulate catabolic activity and make catabolism dissociate from anabolism. The sludge famine is related to an exposure of the settled sludge to anoxic conditions where the substrate concentration is low. Under these stressful conditions microorganisms are starving which may lead to a depletion of cell energy or nutrients storage. The sludge feasting means that fasted microorganisms return to an oxic environment with enough nutrients. As a consequence, the microorganisms growth may be limited by energy uncoupling (Chen *et al*., 2001).
