In Section 7.1.2.2., the principle and applications of fed-batch fermentation tech­nologies for ethanol production were discussed. The operation of fermenters under fed-batch conditions is very difficult to model due to the fact that micro­bial cells grow under permanently changing conditions including volume changes explained by feeding of fresh medium into the bioreactor (Cardona and Sanchez, 2007). To tackle this problem, da Silva Henriques et al. (1999) developed a hybrid neural model for alcoholic fermentation by Z. mobilis in a fed-batch regime. The model uses all the information available about the process to deal with the dif­ficulties in its development and could be the basis for formulation of the optimal feed policy of the reactor.

Precisely, the optimization of feeding policy plays a crucial role for increas­ing both productivity and ethanol yield of fed-batch fermentations. Kapadi and Gudi (2004) developed a methodology for determination of optimal feed rates of fresh culture medium during fed-batch fermentation using differential evolu­tion that resulted in a predicted augment of ethanol concentration at the end of each cultivation cycle. Wang and Jing (1998) developed a fuzzy-decision-mak — ing procedure to find the optimal feed policy of a fed-batch alcoholic fermenta­tion using recombinant yeasts able to assimilate glucose and xylose. The kinetic model involved expressions that take into account the loss of plasmids. To solve this problem, a hybrid differential evolution (HDE) method was utilized. The application of HDE has also been carried out in order to simultaneously deter­mine the optimal feed rate, fed glucose concentration, and fermentation time for the case of S. diastaticus during ethanol production. The optimal trade-off solution was found using a fuzzy goal attainment method that allowed obtain­ing a good agreement between experimental and computed results (Chen and Wang, 2003). HDE also has been used for estimation of kinetic parameters dur­ing batch culture of the mentioned yeast (Wang et al., 2001). Other strategies of optimization have demonstrated their usefulness for evaluation of hybrid con­figurations involving reaction-separation integration (Cardona and Sanchez, 2007). For instance, vacuum fermentation technology has been modeled, simu­lated, and optimized by means of factorial design and response surface analysis (da Silva et al., 1999).

One of the approaches highlighted in Chapter 2 for accomplishing process synthe­sis consists of the application of the principles of topologic thermodynamics (see Section 2.2.6). In the work mentioned above (Sanchez et al., 2006), a preliminary short-cut method to analyze the extractive co-fermentation was developed. In par­ticular, the main components (substrate-water-product-solvent) can be represented in a quaternary diagram in order to locate initial conditions. For representation of the reaction trajectory and considering that the overall fermentation process is irreversible, a stoichiometric approach was used. Therefore, the fermentation is described as follows:

C6Hi2O6 г > 2C2H5OH + 2CO2

Glucose Ethanol

ЗС5Н10О5 Z mobi“s > 5C2H5OH + 5CO2

Xylose Ethanol

Having determined the proportion between substrate (expressed as the sum of both sugars) and ethanol, using a maximum stoichiometric yield from substrate of 0.511 g/g, the inlet conditions are located in the ternary diagram water-ethanol — я-dodecanol, where the liquid-liquid equilibrium is represented as well. Finally, the balance lines corresponding to the operating conditions for the steady-states were drawn in order to define the zone where the performance of extractive fer­mentation process is more stable and advantageous. The procedure for locating the steady-states in the concentration simplex using a short-cut approach is illustrated in Figure 9.17. The process is ideally divided into two steps: microbial conversion and liquid-liquid extraction. The initial sugar concentration is represented in the quaternary diagram by the point A (see Figure 9.17a). The transformation of sugars into ethanol is shown by the line AB, B being the state of the system where the total amount of produced ethanol is represented. This point is the starting mixture for

(a) (b)

the liquid-liquid equilibrium. The line BC represents the addition of я-dodecanol to the aqueous medium containing ethanol (Figure 9.17b). This line lies in the ternary diagram water-ethanol-«-dodecanol, where the zone of heterogeneous mixtures is also drawn. Vertical lines represent the geometric place of points that represent the operating conditions related to the solvent feed stream/aqueous feed stream (R). The intersection of these vertical lines with the line BC (point D) represents the the­oretical conditions corresponding to the mixtures before the separation in phases (the equivalent to the feed mixture in a liquid extractor). Through tie lines, the com­positions of the extract (E) and the raffinate (W) are obtained. These correspond to the compositions of solvent phase effluent and aqueous phase effluent, respectively. For identical inlet concentrations of sugars in the aqueous stream, the position of the starting point B changes when the inlet dilution rate varies. For example, if DAi increases, the new line B’C will lie below the original line BC, which it is explained by a major dilution of ethanol and, therefore, the line approaches the bottom edge of the ternary diagram.

Using this short-cut approach, the zone of feasible operating points can be easily determined. Let us analyze the extreme case when the feed aqueous stream has the maximum allowable concentration of sugars. From the fermentation stoichiometry, this condition corresponds to an inlet concentration in the feed aqueous stream of about 600 g/L of total sugars. Assuming a 95% yield, the total amount of etha­nol that could be produced is 0.486 g/g, which implies a theoretical starting etha­nol concentration of 291.6 g/L (approximately an ethanol mass fraction of 0.42). This value determines the position of the substrate solubility boundary (point H in Figure 9.18). Because the concentration of ethanol in the aqueous phase (raffinate) should not be above the ethanol inhibition boundary (approximately 10% w/w), the operation conditions represented by the line R3 should be such that the ethanol content in the raffinate corresponding to point D" be equal to the ethanol content of the point I to avoid product inhibition. In this way, the area delimited by the points R3D"EK is the zone of feasible steady-states for given conditions of the process

with the maximum concentrations of sugars in the culture broth. For a given inlet dilution rate of 0.1 h-1, an R ratio of 3.6 (line R2), and a working volume of 1 L, the location of the point D’ and the corresponding composition of the extract, can be found. In this case, the ethanol mass content of the extract and raffinate is of 7.5% and 9.0%, respectively, resulting in a total ethanol productivity of 26.4 g/(L x h). The productivity calculated by the rigorous model using ModELL-R is 28.86 g/(L x h). Hence, the developed short-cut method allows one to determine the feasibility of operating parameters and an estimation of the productivities.

The delimited zones can be taken into account for the development of a pre­liminary strategy of optimization. Since the region of feasible steady-states was determined in the ternary diagram (see Figure 9.18), the values range of such manipulating variables as inlet dilution rate, the R ratio, and the concentration of the sugars in the inlet aqueous stream is known and can be bounded for solving an optimization algorithm. The GAMS system was employed to find the optimal value of above-mentioned variables that maximize the total ethanol productivity using the nonlinear programming (NLP) solver CONOPT3. With this aim, liquid — liquid equilibrium relationships were simplified for generating a way to evaluate the ethanol concentration in both phases during extractive fermentation. For this, the distribution coefficient &EtOH was assumed to be linearly dependent of the total concentration of substrates. A good concordance with the data obtained from the ModELL-R was achieved, especially for high values of substrates concentration. The results of this optimization are presented in Table 9.11. From this table, it is evi­dent that calculated optimal variables are indeed in the zone predicted by the short­cut method, and the predicted increase in total productivity is effectively attained, as is shown by the rigorous model.

For a more accurate solution of the optimization problem, the rigorous descrip­tion of the equilibrium model should be coupled or embedded into the GAMS code. Further analysis of this extractive fermentation process could include the formula­tion of an objective function that considers, besides ethanol productivity, other per­formance indexes like the conversion of sugars (better utilization of the feedstock), or the amount of generated wastewater (evaluation of environmental impact). These issues are very significant considering process synthesis procedures. After obtain­ing a global picture of the space of operating conditions and their optimal values for the studied process, experimental runs should be performed in order to confirm the validity of the given theoretical approach. In this manner, the acquired insight of the process will make possible the reduction of expensive experimental work in the search of the optimal operation.

Described types of integration in Chapter 9 are related to the integration of mate­rial flows either for their transformation or for the separation of their components. Similarly, the energy integration of the different steps for ethanol production is possible. Energy integration, particularly heat integration, looks for the best utili­zation of energy flows (heat, mechanical, and electrical) generated or consumed inside the process with the aim of reducing the consumption of external sources of energy such as electricity, fossil fuels (oil, natural gas) mainly used for steam generation, and cooling water. Pinch technology is one of the most widely applied approaches for heat integration in the process industry, especially in the petro­chemical industry. This technology provides the necessary tools for design of the heat exchanger networks including plant utilities. During preliminary design of the heat exchange network, pinch technology allows one to obtain the best values of many process parameters as the type of utilities and their specifications.

Pinch technology has been utilized for the design of heat exchanger networks (HEN) during ethanol production. For the case of ethyl alcohol production from molasses, a process flowsheet was simulated and optimized by heat integration emphasizing the separation step by distillation (Sobocan and Glavic, 2000). The simulation was made by shortcut and rigorous methods in order to perform the heat integration. For ethanol concentrations up to 95.7% by weight, the optimal configuration corresponded to one single column and not two, as had been pro­posed. This work demonstrates the usefulness of heat integration since the optimal design showed a 27% reduction in the total costs (Cardona and Sanchez, 2007).

Methyl tert-butyl ether (MTBE) has been used since 1979 in the United States. This compound is totally miscible with the gasoline, has a similar volatility, and does not absorb water, which confers a low susceptibility to the water-gasoline phase separation. MTBE was the main oxygenating compound and was added to the gasoline up to 15% (by volume), which is equivalent to 2.7% oxygen. Its uti­lization causes the reduction of ozone concentrations at ground level during sea­sons with high smog and CO formation as well as the decrease in benzene usage for gasoline, which led to the reduction of this powerful carcinogenic substance (Braids, 2001).

The production pace of MTBE has increased remarkably since the begin­ning of a reformulated gasoline program in the United States from 12,000 bar — rels/day in 1970 to 250,000 barrels/day in 1998 (Braids, 2001). In fact, it was the chemical with the greatest growth during the 1990s with annual increments in its production of 10 to 20%. The worldwide MTBE production reached 30 million ton/year at the end of the twentieth century (Oudshoorn et al., 1999). MTBE is produced by the reaction of isobutene with methanol. Both isobutene and methanol are from fossil origins, although the latter can be produced from natural gas and even from renewable sources through biomass gasification to methane followed by its oxidation toward methanol. The isobutene is obtained from fractions with four atoms of carbon (C4 fractions) from the catalytic crack­ing process in oil refineries as well as from processing of ethylene (Ancillotti and Fattore, 1998).

The solubility of MTBE in water is in the range of 50,000 ppm or 50 g/L, which makes this compound the most soluble compared to any other gasoline component. When gasoline contacts water, MTBE has a tendency to be dissolved in water but also continues to be dissolved in gasoline. In addition, MTBE is resistant to degradation by chemical or biological means. These two features suggest that MTBE presents a high mobility in natural water streams, especially in groundwater where it can migrate as a result of leakages in the storage and transport systems of gasoline. Furthermore, MTBE at low concentrations can alter the taste and odor of potable water (Nadim et al., 2001). Numerous cases of water sources contaminated with MTBE without the presence of the remaining gasoline hydrocarbons have been reported. One of the most effective methods for MTBE removal is by means of adsorption using granulated activated coal obtained from coconut peels (Braids, 2001). Due to these negative features, the use of MTBE was banned in California in 2004 and its total prohibition is pro­jected by 2010.

As the starch is a polymer exclusively composed of glucose residues, it has become a very important feedstock for ethanol production. For this, the breakdown of starch into glucose with the subsequent conversion of glucose into ethanol using appropriate fermenting microorganisms is required. The technologies for starch hydrolysis will be discussed in Chapter 5. The grains are the feedstock most used for ethanol production from starchy materials, especially corn and wheat, though bioethanol production from rye, barley, triticale (Wang et al., 1997), and sorghum (Zhan et al., 2003) has been reported. Besides the grains, some tubers, such as the cassava and potatoes, offer significant starch contents. In a similar way, etha­nol from sweet potatoes, eddoes, yam (Hosein and Mellowes, 1989; Nigam and Singh, 1995), plantain, and bore (Aya et al., 2005) has been produced.

Some bacteria have the capacity to produce ethanol in significant amounts mak­ing them potential microorganisms for industrial purposes (see Table 6.2). Among bacteria, the most promising microorganism is Zymomonas mobilis. This facul­tative anaerobe presents higher ethanol yields than yeasts, which is related to the metabolic pathways involved. Z. mobilis makes use of the Entner-Doudoroff pathway converting 1 mol of hexose into 2 mol of ethanol, but releasing only 1 mol of ATP (Jeffries, 2005; Figure 6.3), unlike S. cerevisiae that uses the Embden — Meyerhoff-Parnas pathway (i. e., the glucolytic way) but forms two molecules of ATP (see Figure 6.1). This fact implies a lower cell yield due to the lower energy yield of this bacterium, increasing the amount of ethanol that can be obtained from the same amount of substrate compared to yeasts. There have been reported ethanol yields of 97% of the theoretical yield from glucose. Furthermore, this bacterium has a more rapid fermentation due to its higher ethanol production rate (three to five times higher than in S. cerevisiae) and substrate consumption

rate. Additionally, this bacterium has a high ethanol tolerance (100 g/L) and a higher optimal production temperature. Z. mobilis requires no controlled addition of oxygen and is much more susceptible to genetic manipulations to enhance its yield or transfer new traits than is the case of yeasts.

Among the drawbacks of Z. mobilis, the too narrow range of fermentable sub­strates (glucose, fructose, and sucrose) should be noted (Claassen et al., 1999; Hawgood et al., 1985). Other disadvantage of the use of this bacterium for fermen­tation of sugarcane syrup and other sucrose-based media is the formation of the polysaccharide levan (made up of fructose units), which increases the viscosity of fermentation broth, and of sorbitol, a product of fructose reduction that reduces the efficiency of the conversion of sucrose into ethanol (Doelle and Doelle, 1989; Grote and Rogers, 1985; Lee and Huang, 2000). In addition, preculture conditions have a significant influence on bacterium performance, especially on sucrose hydrolysis rate. Hence, the addition of invertase to the culture medium has been proposed (Doelle and Doelle, 1989). Lee and Huang (2000) studied batch ethano — lic fermentation using Z. mobilis through nonstructured models based on meta­bolic analysis. These models allowed the use of ethanol and sorbitol formation during cultivation on a medium containing glucose and fructose or on a sucrose medium supplemented with immobilized invertase.

Other bacteria that have been investigated in order to implement processes of direct conversion of lignocellulosic biomass into ethanol are thermophilic and saccharolytic clostridia. Clostridium thermohydrosulfuricum, C. thermosaccha- rolyticum, and C. thermocellum can synthesize up to 2 mol EtOH/mol hexose. Likewise, these bacteria may transform pentoses and amino acids into ethanol. Having saccharolytic properties, these microorganisms have the ability to grow on a wide variety of nontreated wastes. C. thermocellum can even directly con­vert lignocellulosic materials into ethanol (McMillan, 1997). In this way, ethanol can be directly obtained from pretreated lignocellulosic biomass without the need of adding costly cellulases. Moreover, the cultivation of these microorganisms at high temperatures offers the possibility of an easier ethanol removal by distilla­tion or pervaporation and reduced cooling expenditures (Claassen et al., 1999). Clostridium thermocellum has been the most studied thermophilic clostridium because it has the capacity to produce cellulases, hydrolyzing the cellulose and fermenting the glucose forming ethyl alcohol. On the other hand, the possibility of employing C. thermosaccharolyticum to produce ethanol from pentoses result­ing from hemicellulose degradation during the pretreatment of biomass has been shown (Wyman, 1994).

The main drawback of these bacteria consists in their very low ethanol toler­ance compared to yeasts. Consequently the maximum reached ethanol concentra­tions are lower than 30 g/L. In addition, they exhibit a reduced ethanol yield due to the formation of fermentation by-products like acetic acid and lactate that make the final ethanol concentrations very low and cultivation times prolonged (3 to 12 days) (Baskaran et al., 1995; McMillan, 1997; Szczodrak and Fiedurek, 1996; Wyman, 1994). Process integration can play a crucial role for evaluating the most optimal configurations in order to implement them at an industrial level.

For process synthesis procedures intended for the design of fuel ethanol pro­duction processes, the selection of the technological scheme for ethanol separa­tion and dehydration has paramount importance. Thus, the utilization of salts as extractive agents (saline extractive distillation) has demonstrated certain energetic advantages compared to other dehydration schemes, according to some reports (Barba et al., 1985; Llano-Restrepo and Aguilar-Arias, 2003). Cited results indi­cate that energy costs of saline distillation were lower than in the case of azeo­tropic distillation (using benzene, pentane, or diethyl ether), extractive distillation (using ethylene glycol or gasoline), or solvent extraction, being almost the same as the costs of pervaporation. Pinto et al. (2000) used Aspen Plus® for simulation and optimization of the saline extractive distillation for several substances (NaCl, KCl, KI, and CaCl2). This configuration was compared to the simulated scheme of conventional extractive distillation with ethylene glycol and to data for azeotro­pic distillation. Obtained results showed considerably lower energy consumption for the process with salts. However, for this latter case, the recovery of salts was not simulated. Thus, if evaporation and recrystallization of salts is contemplated, energy requirements could significantly increase, taking into account the energy expenditures. In this way, the utilization of commercial simulators shows the viability for predicting the behavior of a given process configuration providing the appropriate thermodynamic models of studied systems, as illustrated in the following case study.

In general, it is considered that the characteristics of the stillage produced from lignocellulosic materials are comparable to those of the stillage obtained when conventional feedstocks are employed, thus the treatment and utilization methods of this latter type of stillage can be applied to the lignocellulosic stillage (Wilkie et al., 2000). The flowsheets proposed for the treatment of the stillage from bio­mass contemplate its evaporation and centrifugation in order to obtain the thin stillage and solid materials (Wooley et al., 1999). The solids recovered have high lignin content so they are sent to the burner of the boiler for generation of the steam required by the overall ethanol production process. The thin stillage is par­tially recirculated, along with the bottoms of the rectification column, as water for washing the pretreated lignocellulosic biomass (see Chapter 11). The remaining thin stillage is directed to the evaporation train in order to concentrate it into a syrup containing about 60% water. Due to this high moisture content, this syrup can undergo anaerobic digestion though its incineration.

Besides the stillage, other liquid waste streams are generated in the process for fuel ethanol production from lignocellulosic biomass. For the treatment of this wastewater, different alternatives have been evaluated that include its anaerobic digestion to remove 90% of its organic matter content, as well as the utilization of the biogas produced for steam generation. The effluent of anaerobic digestion can be sent to an aerobic digestion pond where another 90% of the organic matter content is removed. The aerobic sludge formed in this process is withdrawn by clarification and filtration. The filtered sludge can also be sent to the burner for steam generation. The water from the clarifier can be recirculated in the process (Merrick & Company, 1998; Wooley et al., 1999).

Case Study 10.1 Preliminary Comparison of Effluent Treatment Alternatives

Process synthesis procedures should provide insight on the more suitable tech­nologies for effluent treatment during fuel ethanol production. In a previous work (Cardona et al., 2006), three effluent treatment configurations were analyzed. The effluent studied was the thin stillage generated during the production of fuel etha­nol from lignocellulosic biomass as well as other liquid effluents (wastewater from

pretreatment and detoxification). In addition, the solid residues generated during the centrifugation of the whole stillage were also studied. These residues are mainly represented by the lignin nontransformed during the process, residual cellulose and hemicellulose, and the cell biomass (in this case Zymomonas mobilis fragments). The configurations were simulated using Aspen Plus® v11.1 applying the general guideline for simulations described in Chapter 8, Case Study 8.1. To analyze the biological transformations taking place during the degradation of the organic mat­ter contained in the stillage, the stoichiometric approach was employed in the framework of the process simulator. The percentage of organic matter removal (degradation) was evaluated based on the mass balance results calculated. In addi­tion, preliminary data on generation of electricity was also taken into account. The aim of this case study was to investigate different flowsheet configurations and their combinations for the treatment of the liquid and solid effluents obtained during fuel ethanol production from lignocellulosic biomass.

The treatment procedures analyzed consider the anaerobic digestion of the thin stillage, the aerobic digestion of this same stream, the combination of these two treatments, and the incineration of solid residues in order to obtain electricity. These options are depicted in Figure 10.1. The anaerobic digestion was analyzed through the following chemical reactions mediated by a consortium (mixed culture) of anaerobic bacteria in the corresponding reactor:

+ h2o ^ СбН12Об

Cellulose Water Glucose

(C5H8O4)„ + H2O ^ C5H„O5 Hemicellulose (xylan) Water Xylose

CHL8Oa5Na2 + 0.5 H2O ^ 0.167 C6H12O6 + 0.2 NH3 + 0.1 H2 Z. mobilis Water Glucose Ammonia Hydrogen

C6H12O6 ^ 2 C2H5OH + 2 CO2 Glucose Ethanol Carbon dioxide

3 C5H10O5 ^ 5 C2H5OH + 5 CO2 Xylose Ethanol Carbon dioxide

2 C2H5OH + CO2 ^ CH4 + 2 CH3COOH Ethanol Carbon dioxide Methane Acetic acid

CH3COOH ^ CH4 + CO2

Acetic acid Methane Carbon dioxide

4H2 + CO2 ^ CH4 + 2 H2O

Hydrogen Carbon dioxide Methane Water

These reactions represent some of the main stages of anaerobic digestion of organic matter. The first three reactions correspond to the hydrolytic processes, the following three reactions are carried out by fermentative bacteria, the sixth reaction represents the acetogenesis stage, and the last two reactions globally describe the action of methanogenic bacteria. The formation of anaerobic bacteria is not shown in this reaction scheme although it has been taken into account.

The aerobic digestion was simulated considering the oxidation reaction for each one of the organic compounds present in the liquid effluent. As an example, the complete oxidation of glucose is presented as follows:

C6H12O6 + O2 ^ 6 CO2 + 6 H2O Glucose Oxygen Carbon dioxide Water

Under aerobic conditions, a large portion of the organic matter contained in wastewater may be oxidized biologically by microorganisms to carbon dioxide and water, thus the formation of activated sludge (aerobic bacteria) was also consid­ered. Approximately 50% reduction in solids content can be achieved through this treatment.

Finally, the incineration of the solid residues was simulated through a stoichio­metric approach where all the organic compounds are completely oxidized into carbon dioxide and water without formation of any cell biomass. The products of this process are the combustion gases and the remaining ash. The main combustion reaction is the burning of lignin whose energy content achieves 25.4 MJ/kg.

The first configuration contemplates the anaerobic digestion of the liquid efflu­ent (mostly thin stillage) in an anaerobic reactor producing biogas and a suspension of anaerobic sludge that is settled in a decanter. Then, the sludge is mixed with the predried solids residue in order to be burnt in the co-generation system. This

system consists of a burner where the combustion reaction is carried out, a cyclone for removing the ash, a heat exchanger representing the boiler, and a turbogenera­tor where the electricity is produced employing the high-pressure (HP) steam from the boiler. The low-pressure (LP) steam can be used to cover the thermal energy required in the overall ethanol production process. This configuration is schemati­cally depicted in Figure 10.2.

The second configuration is based on the aerobic digestion of the wastewater.

As in the previous case, the aerobic sludge generated and the solids residue are sent to the cogeneration system (Figure 10.3). Finally, the third configuration comprises the combination of anaerobic and aerobic digestion as well as the delivery of the generated sludge and solids residue to the cogeneration unit, as can be observed in Figure 10.4.

The preliminary results obtained are presented in Table 10.1. The scheme involv­ing the anaerobic digestion of the liquid effluents followed by the aerobic digestion of the liquid stream exiting the anaerobic digester represents higher removal of the organic matter contained in these effluents, as suggested in the work of Merrick & Company (1998). If only one type of digestion is considered, the anaerobic diges­tion gives better results. In the Table 10.1, the sole incineration of the two types of effluent streams (solid and liquid) is also considered as a fourth configuration.

In this case, the power regenerated was considered as the comparison standard

(100%). The configurations involving the separate treatment of the wastewater show a reduced generation of electricity due to the lower availability of organic compounds entering the cogeneration unit. However, the incineration implies that the liquid streams are evaporated before their mixing with the solid residues. For this process, energy is required and the need for thermal energy should be supplied by the low-pressure steam extracted from the turbogenerator. This means that the net energy (thermal and electric) produced can be lower than the net energy of the schemes involving the separate treatment of wastewater. Undoubtedly, process simulation tools are invaluable to assess all these configurations and the energy surplus that may be obtained from these treatment schemes.

Even as world hunger increases and government policy responses have limited effect on high food prices, there is an opportunity for agricultural (including small holders) in developing countries. In this context, the Food and Agriculture Organization (FAO) defines food security based on four dimensions:

1. Availability (production)

2. Access (income and prices)

3. Utilization (nutrition)

4. Stability (price volatility)

But in the context of the bioenergy and food security program (BEFS), the focus consists on availability and access (FAO, 2008). The BEFS project is designed in the framework of FAO and the German Gesellschaft fur Technische Zusammenarbeit (GTZ) collaboration to mainstream food security concerns into national and subnational assessments of bioenergy potential. As reported by the FAO, the key questions the project will address are:

• What are the best types of bioenergy systems to help diversify agricul­tural output (energy feedstock), contribute to rural development, and increase rural employment and incomes?

• How could bioenergy production benefit the environment and increase energy and food security for producers farming biomass as a source of energy for themselves, on-farm use, local communities, or commercial markets?

• How could diversification of domestic energy supply provide increased energy access to rural enterprises and reduce the household energy bur­dens of rural women?

• Is there anything different about bioenergy that could mitigate or over­come factors of exclusion that contribute to food insecurity and rural poverty?

• How can low-income, food deficit countries ensure that food security concerns are addressed, given the complex linkages between agricul­ture, energy, environment, and trade?

• What are the implications on available food supplies and food prices in terms of competition for natural or human resources?

• How will bioenergy affect agricultural systems, particularly for poorer households dependent on their own food production?

• Who (public, private, or civil society) is best placed to deal with potential conflicts arising from competition for food, feed, or fuel use of biomass?

FAO considers that increased competition for land and water resources may result, and higher and less stable food prices may be one of many possible con­sequences. Bioenergy may also provide ways to support rural development and raise farm incomes.

Resuming, the BEFS project will analyze the risks and opportunities of bioen­ergy and how it could affect food security in partner countries.