Fuel ethanol production has increased remarkably because many countries look to reduce oil imports, boost rural economies, and improve air quality. The world ethyl alcohol production was about 64 billion liters in 2008 (Renewable Fuels Association, 2009), with the United States being the first producer (see Table 1.3). On average, 86% of produced ethanol worldwide corresponds to fuel ethanol, 17% to beverage ethanol, and 10% to industrial ethanol. Reported data taken from

different sources indicate that Brazil and the United States account for 82% of world production of fuel ethanol, though this percentage is changing constantly due to the dynamics of this biofuel on the global market. Asia and the EU are following the large producers. China has the biggest plant for ethanol production with an annual capacity of 320 million gallons; this plant is located in the Jilin province and currently produces 240 million gallons per year (Murray, 2005). India is seeking to enhance its ethanol production to meet its auto biofuel pro­gram; it has envisioned the use of ethanol not only for gasoline blends, but also for ethanol-diesel blends and for biodiesel production (Subramanian et al., 2005).

Europe’s fuel ethanol sector was a slow starter. It took over 10 years to grow production from 60 million liters in 1993 to 525 million liters in 2004. In the following two years, we saw a true explosion in production. In 2005 and 2006, there were double-digit growth levels of over 70%. However, it was not a sustain­able growth; in 2007, production increased by only 11%. Total EU production in 2008 was 2.8 billion liters, up from 1.8 billion liters the previous year. This represents a significant increase of 56%. This increase was due to the growth in French production, which almost doubled to 1 billion liters in 2008 (539 million liters in 2007). This made France the biggest EU fuel ethanol producer in 2008, followed by Germany that also expanded its production to 568.5 million liters, which represented a 32.5% increase in internal Germany fuel ethanol produc­tion. The third highest producing country was Spain with 317 million liters. In 2008, fuel ethanol production capacity increased in Belgium. Finland resumed its production in 2008 and, in Austria, fuel ethanol has been produced for the first complete year (European Bioethanol Fuel Association, 2009).

Many countries have implemented or are implementing programs for addition of ethanol to gasoline (Table 1.3). Through the ProAlcool program, Brazil has been utilizing hydrous ethanol as a fuel and anhydrous ethanol as an oxygen­ate. This country produces ethanol from sugarcane. Although, at the beginning of this program, concerns about the disjunction between food production versus fuels (sugar versus ethanol) were expressed, it has been demonstrated that when the food shelves remained empty it was caused by the destination of food produc­tion to the export markets. In the same way, when the shelves remained full, it was because large proportions of the population lacked the purchasing power to buy food (Rosillo-Calle and Hall, 1987). However, due to the liberalization of the Brazilian fuel market, the official end of the ProAlcool program, the gradual elimination of subsidies, and the political and economic conjuncture, production and consumption paces of fuel ethanol diminished at the end of twentieth cen­tury, although a reactivation is expected in the coming years (Rosillo-Calle and Cortez, 1998; Wheals et al. 1999). Brilhante (1997) indicates that the pursuit of ethanol fuel in Brazil was not based on long-term plans with deep-set values, but has been a response to such circumstances as a depressed sugar industry, an ambi­tious attempt to reduce oil dependency, and more recently, “green” arguments.

Brazil is the second largest producer of fuel ethanol in the world and the first in world’s exports (Renewable Fuels Association, 2007). Brazil is considered as the best economical model for ethanol production in the world because it does not implement subsidies for fuel ethanol production. However some authors consider that the successful Brazilian ethanol model is sustainable only in Brazil due to its advanced technology and its enormous amount of arable land available (Sperling and Gordon, 2009). Currently, Brazil is actively seeking export markets for its bioethanol (Thomas and Kwong, 2001), especially in Japan, which will become a net importer of ethanol (Orellana and Neto, 2006). Brazil’s 30-year-old fuel ethanol program is based on the most efficient agricultural technology for sugar­cane cultivation in the world, using modern equipment and cheap sugarcane as feedstock. The residual cane waste (bagasse) is used to process heat and power, which results in a very competitive price and also in a high energy balance (output energy/input energy), which varies from 8.3 for average conditions to 10.2 for best practice production (Macedo et al., 2004).

In the case of the United States, ethanol is used either as a fuel substitute or as an oxygenate. At present, both Ford and Chrysler offer standard models designed to run on either 85% ethanol (E85) or gasoline. The addition of gasoline to the E85 or E95 mixtures is done for improving cold start in engines using these kinds of fuel. However, the best perspectives for bioethanol can be found in the oxy­genate market. Fuel ethanol production was boosted by the passage of the Clean Air Act in 1970, and especially by the Clean Air Amendments in 1990. Current regulations aimed at controlling CO emissions have already produced a signifi­cant demand for ethanol as an oxygenate. The tax credit gives oil companies an economic incentive to blend ethanol and gasoline. It is currently $0.51/gal — lon for pure ethanol, scaled to the amount of ethanol in the blend. Additionally, a $0.54/gal tariff on ethanol from Brazil was established to protect the domes­tic ethanol industry. The 2008 farm bill reduced the blenders credit to $0.46/gal and maintained the tariff through 2010 (Keeney, 2009). The establishment of the Renewable Fuel Standard (biofuels) provided additional incentives for ethanol use, mainly in the form of binding mandates. The 2006 act required that 4.0 bil­lon gallons per year (BGPY) be mixed with gasoline, 6.1 BGPY by 2009, and 7.5 BGPY by 2012. The Energy Independence and Security Act (EISA) of 2007 then revised the mandates to require 15 BGPY by 2015. No doubt the industry is highly subsidized; one analysis estimates that the biofuels industry in the United States receives a $60-billion subsidy each year (Keeney, 2009).

The EU has issued different directives about the addition of renewable oxy­genates to fuels. The oxygenation target for fuels estimates the addition of up to 2 wt.% by 2005 and 5.75% by 2010. However, the implementation of these direc­tives varies too much among the different countries. Spain and France are leading in the production of bioethanol in Europe. In contrast, Germany has developed the production of methyl esters of fatty acids obtained from rapeseed as biodiesel. For this country, it is considered that the production of fuel ethanol is not eco­nomically feasible in comparison to gasoline due to the high costs of feedstocks (grains, sugar beet; Henke et al., 2005; Rosenberger et al., 2002). This situation can change dramatically if volatility and high prices in the oil market remain.

Besides Brazil, other Latin American countries are implementing fuel etha­nol programs. Colombia is a country with important oil reserves. However, the country imports gasoline due to its refining limitations. In addition, the pace of new oil discoveries in the country has decreased. Colombia faces the depletion of its reserves and possible oil imports in 2015. For this reason, besides environ­mental considerations, the country has decided on replacing part of the fossil fuels with alternative energy sources. The production and use of liquid biofuels (bioethanol and biodiesel) is the strategy adopted by the Colombian government to diminish the oil dependency and reduce the polluting gases released by the transportation sector in the Colombian cities. Moreover, it is expected that the implementation of the biofuels programs allows the development of the rural sector that, in certain regions, is attacked by the violence associated with the use of land for illegal crops.

Colombia started the addition of 10% (v/v) ethanol to gasoline in cities with more than 500,000 inhabitants in November 2005 with the Act 693 of 2001 issued by the Congress. This disposition will be extended to the entire country in the coming years. The perspectives for the ethanol market in Colombia are bright. The amount of fuel ethanol needed in the zones where the E10 program is being implemented reaches 1,050,000 L/d. The expansion of the program to the rest of the country requires a total ethanol production of about 1,600,000 L/d. In addi­tion, Colombia has all the conditions to become an ethanol-exporting country considering its agro-ecological conditions, its geographical location, as well as the growth of potential markets in North America, Europe, and Asia. To make the most of these conditions, the country should produce an estimated 3,800,000 L/d of fuel ethanol by 2020 to meet the national demand and export the surplus to new growing markets (Proexport Colombia, 2005).

The implementation of the E10 program depends on raw materials and tech­nological limitations. For this reason and driven by the availability of the feed­stock and the capacity of the well-structured sugar industry, the main cities of the West Southern and West Central departments (administrative regions into which Colombia is divided) started this program. The Cauca River valley is the major sug­arcane production region and it is located in these departments. This region concen­trates the best cultivable lands for sugarcane cropping (about 200,000 Ha) because it has the most appropriate conditions: extensive and fertile alluvial valleys located at an average of 1,000 m above sea level in tropical latitude. Currently, five etha­nol production plants co-located in large sugar mills are operating. However, the synergies of the sugar sector, controlled by a few economic groups, have not been allowed to achieve a great impact on the creation of new rural jobs. According to estimations by the Colombian Ministry of Agriculture (Ministerio de Agricultura y Desarrollo Rural, 2006), the surface cropped with sugarcane in 2006 did not increase compared to data obtained in 2005 (by the Columbian sugar industry). Despite the enhancement in the production of fuel ethanol during these two years, Colombia has not projected any increase in cane plantations. This means that the generation of new rural jobs related to ethanol production is practically nil. In fact, part of the cane produced is diverted to ethanol distilleries, decreasing the volumes of sugar for export. Regulations issued by the government have included the sugar value in the international market as a variable considered in calculating the price of fuel ethanol produced in the country. Thus, future increases in the international price of sugar will not discourage the national production of fuel ethanol. This situ­ation allows the development of new commercial projects for ethanol production in order to guarantee permanent use of ethyl alcohol for the E10 program.

The Colombian government expects that the construction of new ethanol-pro­ducing facilities in other regions of the country will lead to real improvement in the economic situation of the rural communities through new demands for agricultural raw materials. Therefore, other feedstocks for ethanol production are being analyzed considering the future growth of the fuel ethanol market. Corn, cassava, and beets have been considered as potential feedstocks. Sugarcane is also considered for ethanol production in zones other than in the Cauca River valley. This encourages the cropping of cane by specific rural communities not related to the big sugar companies. Many rural communities cultivate sugarcane for producing noncentrifugal sugar called panela (solid brown sugar), a sweet­ener and low-cost beverage base widely used by popular segments in Colombia. However, the quality of life of panela producers is traditionally low within the Colombian context. For this reason, the government is actively encouraging the organization of the communities linked to the panela economy for them to supply the feedstock for new projects of fuel ethanol production.

The relatively mature technology for ethanol production from corn is one of the options to be considered under Colombian conditions because corn is an important crop mainly cultivated in northern and eastern regions. The construc­tion of ethanol plants using corn could offer the production of valuable co-prod­ucts (e. g., dried distillers grains with solubles for cattle food). Furthermore, this technology may be the base for the development of ethanol production processes from other starchy materials such as cassava or potatoes, crops with a significant economic importance in Colombia. Additionally, the projected signing of a free trade agreement with the United States, the first world producer of corn, implies the search for new markets for local corn production.

Ethanol can be obtained synthetically from coal and natural gas. The major producer of this type of alcohol is South Africa (Thomas and Kwong, 2001). It can be obtained by oxidation of olefins as well. However, 95% of world ethanol is produced by fermentation from carbohydrate-containing feedstock (Berg, 2004). The main part of ethanol is produced in batches.

Substrate concentration at the beginning of fermentation is 15 to 25% (w/v) solids and the pH is adjusted to a value of 4 to 5 with the aim of reducing infection risks. The process is carried out at 30 to 35°C. Generally, yield reaches 90% of theoretical maximum, expressed in g EtOH/g substrate. The rest of the substrate is converted into biomass and other by-products such as glycerol and acetate. Formed biomass can be utilized many times. Ethanol concentration at the end of fermentation is 80 to 100 g/L (Claassen et al., 1999). In most of the distilleries, fermentation time is 24 h, employing 6 h for yeast sedimentation in the vessels (Pandey and Agarwal, 1993).

Cereal straws are lignocellulosic materials with a high content of hemicellulose compared to cellulose (see Table 3.11). The straws comprise the dry stalks of a cereal plant, crushed or not, after the grain or seed has been removed. The wheat straw presents a great availability due to the vast cultivated volume of this grain in the world (about 813 million ton per year of straw). In the United States, about 33 million ton of wheat straw is available, assuming that 40% of total straw is collectable in a sustainable way (Kadam and McMillan, 2003). This amount of straw is equivalent to 9.6 million L of ethanol if considering an ethanol yield from wheat straw of 261.3 L/ton. Rice straw is another source of lignocellulosic biomass with a high availability in the world (779 million ton). In particular, the increase in environmental controls is phasing out the practice of burning rice straw in the open air. These regulations contribute to the search for new applica­tions for these types of residues.

As mentioned in the Chapter 6, S. cerevisiae converts hexoses into pyruvate through glycolysis, which is decarboxylated to obtain acetaldehyde that is finally reduced to ethanol generating two moles of adenosine triphosphate (ATP) for each mol of consumed hexose under anaerobic conditions. In addition, this micro­organism also has the ability to convert hexoses into CO2 by aerobic respiration favoring the production of yeast cells. Therefore, aeration is an important factor for both cell growth and ethanol production. Although these yeasts have the abil­ity to grow under anaerobic conditions, small amounts of oxygen are needed for synthesis of such substances as fatty acids and sterols. In the case of continu­ous cultures, cell concentration, cell yield from glucose, and yeast viability are enhanced by increasing air supply while decreasing ethanol concentration under both microaerobic and aerobic conditions. Inhibition of cell growth by etha­nol decreases at microaerobic conditions related to fully anaerobic cultivation. Specific ethanol productivity is stimulated with the increase of oxygen percentage in the feed (see, for example, the work of Alfenore et al., 2004). In the case of fed-batch cultures, Alfenore et al. (2004) show that higher ethanol concentrations (147 g/L) can be obtained in cultures without oxygen limitation (0.2 vvm) in only 45 h in comparison to microaerobic conditions. In addition, a 23% increase in the viable cell mass was achieved. Similar studies for fed-batch cultures were performed for assessing the synergistic effect of temperature and ethanol content on the behavior of S. cerevisiae cultures (Aldiguier et al., 2004). Best results were found at 30°C and 33°C for around 120 g/L of ethanol produced in 30 h. Slight benefits for growth at 30°C and for ethanol production at 33°C were observed. These data suggest the possibility of designing two-stage, high-cell-density biore­actors. One proposed method to reduce the inhibition by ethanol of yeast cultures is the preadaptation to a medium with an initial content of ethanol. In this way, yeasts could be acclimatized to acquire ethanol tolerance by serially transferring them into a medium with ethanol. Vriesekoop and Pamment (2005) point out that this approach has not provided important success because which preadaptation to ethanol causes a decrease in the rate of yeast death, it does not prevent it. These authors showed that the addition of acetaldehyde to preadapted cultures of S. cer — evisiae eliminates the long lag-phase of yeasts caused by the sudden exposure to initial ethanol concentrations as high as 50 g/L.

To evaluate the contribution of reaction-reaction integration to the improvement of energy efficiency of the corn-to-ethanol process, two technological configurations were simulated and their energy consumption was calculated. The first process cor­responded to a nonintegrated configuration based on SHF of the dextrins produced during the liquefaction of corn starch. In addition, the considered configuration

TABLE 9.1

Integration of Reaction-Reaction Processes by Simultaneous Saccharification and Fermentation (SSF) for Fuel Ethanol Production from Different feedstocks

Semicontinuous SSF S. cerevisiae + commercial cellulase supplemented with p-glucosidase Paper sludge Special design of solids-fed reactor; EtOH conc. 35-50 g/L; 0.466 g/g EtOH yield; 74-92% cellulose conversion; 1^ months of operation Fan et al. (2003) Continuous SSF S. cerevisiae + microbial amylases Grains Yield 2.75 gal/bushel; industrially implemented Madson and Monceaux (1995) Co-immob. Zymomonas mobilis and glucoamylase on K-carrageenan Liquefied corn liquid Fluidized-bed reactor; approx..100 g/L dextrin feed; conversion 53.6-89.3%; EtOH conc. 22.9-36.44 g/L; productivity 9.1-15.1 g/(L. h) Krishnan et al. (1999) S. cerevisiae + commercial cellulase supplemented with p-glucosidase Dilute-acid pretreated hardwood CSTR; residence time 2-3 d; 83% conversion; EtOH conc. 20.6 g/L South et al. (1993) Continuous SSYPF S. cereveisiae + microbial amylases Corn, milo, wheat Yield 2.75-2.8 gal/bushel; industrially implemented Madson and Monceaux (1995) Source: Adapted from Cardona, C. A., and O. J. Sanchez. 2007. Bioresource Technology 98:2415-2457. Elsevier Ltd.

TABLE 9.2

Energy Comparison of Two Processes for Fuel Ethanol Production from Corn Grains by Dry-Milling

Integrated Process Non-lntegrated Process Item (ssF) (sHF)

Ethanol produced (kg EtOH/h) 17,837 17,838

Energy consumption (MJ/h) 270,487 290,856

Energy consumption (MJ/L EtOH) 11.67 12.55

comprised the dry-milling of corn grains, ethanol dehydration by adsorption with molecular sieves, and effluent treatment step, which allows the recovery of a co­product called dried distiller’s grain with solubles (DDGS) that is commercialized as an animal feed. The flowsheet for this configuration is illustrated in Figure 11.6 of Chapter 11.

The simulation of the above-described process was performed using the simu­lator Aspen Plus™ using a capacity of 17,839 kg/h. The simulation procedure for distillation columns was the same of that described in Case Study 8.1 (see Chapter 8, Section 8.3).

The second studied process corresponded to the integrated configuration based on SSF of the dextrins produced during the liquefaction of corn starch. This process only differs from the SHF process in that the saccharification and fermentation are carried out simultaneously in the same unit at 33°C (see Figure 11.7 of Chapter 11). The remaining steps are similar for both processes. This configuration has already been simulated in previous works (Cardona et al., 2005a, 2005b; Quintero et al., 2008) applying the same simulation features mentioned above.

The obtained results are shown in Table 9.2. Calculated data allow observing that the effect of the reaction-reaction integration studied on the energy performance of the process is favorable. The integration of the enzymatic reaction (saccharifi­cation) and the microbial transformation (fermentation) allowed the reduction of energy costs by 7%, which is a significant energy savings.

The co-generation process has a strong influence on the environmental indicators of the cane-to-ethanol process, as shown in Chapter 10, Case Study 10.3. In order to get a comprehensive procedure to take into account this process, the simulation of the co-generation unit should be appropriately accomplished. In the previous work cited in the last case study (Quintero et al., 2008), this simulation was performed.

The co-generation technology corresponded to a circulating fluidized bed com- bustor/turbogenerator (CFBC/TG) system. This system has been contemplated in the model process designed by the U. S. National Renewable Energy Laboratory (NREL) for co-generation using the lignin released during the conversion of ligno — cellulosic biomass into ethanol (Wooley et al., 1999). The CFBC/TG technology offers an increased efficiency in the generation of steam and power related to con­ventional co-generation units working with low pressure boilers, which usually generate steam of 280° to 300°C and 20 to 21.7 atm (Aguero et al., 2006; Macedo and Nogueira, 2005). Mass balance data of CFBC/TG systems reported in the work of Wooley et al. (1999) were utilized for conceptual design and simulation of the co-generation unit using cane bagasse. This unit was simulated through several process modules of Aspen Plus. The burner was described through a stoichiomet­ric reactor considering the incomplete combustion of bagasse organic components (lignin, cellulose, hemicellulose, etc.), taking into account the formation not only of CO2, but also of CO. In the same way, reactions for NO formation were included. The boiler was studied as a heat exchanger where the feed water enters at 121°C and 97.5 atm and the generated steam exits at 510°C and 84.9 atm. A pump elevat­ing the pressure of the feed water up to 97.5 atm was included in this analysis. The simulation of the co-generation unit also took into account that the combustion gases leaving the boiler can be used not only for preheating the air required to burn the bagasse, but also for drying the wet bagasse generated in the mills. As studied by Maranhao (1982), a previous drying enables the reduction of bagasse moisture down to 40%, improving the combustion and increasing the amount of generated steam. The analysis of the co-generation system also included a cyclone for separat­ing most particulate matter from the flue gases leaving the bagasse dryer. The elec­tricity production using a turbogenerator was simulated through the compressor module of Aspen Plus considering a negative change of pressure and selecting the

Source: Quintero, J. A., M. I. Montoya, O. J. Sanchez, O. H. Giraldo, and C. A. Cardona. 2008. Energy 33 (3):385-399. Elsevier Ltd. With permission.

isentropic type of compressor. Thus, it is possible to simulate the power generation and calculate the properties of the exhausted steam. In this paper, a multistage tur­bine was taken into consideration for producing three types of steam: high pressure steam (used for preheating the water feeding the boiler), low pressure steam (used for the energy supply of most process units like heaters, sterilizers, and column reboilers), and very low pressure steam (employed for stillage evaporation).

The performed simulation of the co-generation system for the cane-based pro­cess showed a good agreement with the reported industrial data. The CO2, CO, and NOx emissions obtained in this work are close to the average emission factors reported by the U. S. Environmental Protection Agency (EPA) and other authors for combustion of bagasse in sugar mills (EPA, 1993, 1996; Gheewala, 2005; Table 11.6). Also, the presence of moisture in the bagasse has an important influ­ence on the amount of thermal energy released during its combustion. This fact is confirmed by simulation. Thus, the use of nonpreheated wet bagasse (about 50% moisture) implies a 7.3% reduction in the amount of produced steam with a pressure of 84.9 atm compared to the case when the bagasse is dried down to 40% moisture. This very high-pressure steam from the boiler is used to power turbines. In con­trast, the reduction of available steam reaches 13.18% when low-pressure boilers (29 atm) are used (Maranhao, 1982). The total amount of exhausted steam from the turbogenerator covers all the needs of thermal energy required by the ethanol production facility. This fact dramatically improves the economic performance of the overall process. A fraction of the energy released in the turbogenerator is also employed to cover the needs of mechanical energy required during cane milling. Moreover, the power produced (33 MW) meets plant requirements (13.90 kWh/ ton cane, according to simulation results), remaining a significant surplus that can be sold to the electric network. Herein, the electricity surplus was not considered during the economic evaluation. Nevertheless, the co-generation system that was examined generates 99.06 kWh/ton cane of electric energy available for sale. The indicated amount of power is within the range of modern co-generation technolo­gies based on extraction-condensation turbogenerators, which can reach 90 to 150

kWh/ton cane of electricity surplus (Macedo and Nogueira, 2005). This type of co-generation unit has been proposed for new Brazilian sugar mills and distilleries, being an excellent option for new ethanol production facilities in Colombia.

The environmental performance of the technological configuration for fuel ethanol production from sugarcane was carried out applying the WAR (WAste Reduction) algorithm by using the software WAR GUI (WAR graphical user inter­face). For this, the data on mass and energy balances obtained from the simulation with Aspen Plus were employed. As discussed in Chapter 10, Case Study 10.3, highest weightings (10.0) were assigned to the four local toxicological categories in comparison to the weightings of the four global environmental physical categories (2.5). These assignments were done to give higher importance to the local condi­tions taking into account that Colombia is an agricultural country with vast hydric resources and rich biodiversity that should be protected. Results using equally weighted categories can be found in a previous work (Montoya et al., 2006).

The total output rate of environmental impact for cane ethanol process is shown in Figure 11.3 and the potential environmental impact generated within the system is shown in Figure 11.4. The improved environmental performance (in the terms of potential environmental impact (PEI) leaving the system per mass of products) of the sugarcane-based process can be explained by the operation of the co-generation unit. In this unit, one of the process by-products, the bagasse, is employed as renew­able fuel in order to generate all thermal and mechanical energy required by the process as well as the power needed. When bagasse is burned, the negative effect of CO2 released into the atmosphere during its combustion is compensated positively by the CO2 fixed from the atmosphere during the sugarcane growth. Therefore, the utilization of bagasse as a solid biofuel does not necessarily imply a net increase of

Impact Categories

FIGURE 11.3 Total output rate of environmental impact for the studied processes. The impact is expressed as the PEI leaving the system per mass of product streams. (From Quintero, J. A., M. I. Montoya, O. J. Sanchez, O. H. Giraldo, and C. A. Cardona. 2008. Energy 33 (3):385-399. Elsevier Ltd. With permission.)

Impact Categories

FIGURE 11.4 Potential environmental impact generated within the studied processes. The impact is expressed as the PEI generated within the process per mass of product streams. (From Quintero, J. A., M. I. Montoya, O. J. Sanchez, O. H. Giraldo, and C. A. Cardona. 2008. Energy 33 (3):385-399. Elsevier Ltd. With permission.)

CO2 in the atmosphere. However, the environmental impact of the involved mass and energy flows in the whole life cycle is not assessed by the WAR GUI software.

It only performs the evaluation of the potential environmental impact of the streams leaving or generated during the conversion process. To overcome the problem of CO2 emissions, the flue gases stream leaving the co-generation unit was analyzed as if it were free of carbon dioxide. Thus, zero CO2 net emissions were considered on a life cycle basis.

The commercial utilization of press mud and concentrated stillage indicates the drastic reduction of two polluting streams generated during the cane-to-ethanol process. Stillage is a highly contaminating liquid stream due to its high biologi­cal oxygen demand (BOD; 30,000 to 60,000 mg/L). Several methods have been proposed for its treatment, but few have been employed. The evaporation of stillage to obtain a co-product used as a fertilizer for cane plantations is one of the most popular applications, along with its use for irrigation. However, the properties of the soil may be affected by the utilization of huge volumes of concentrated stillage. These effects have not been properly studied within the framework of the life cycle assessment methodology and further studies are required.

It is expected that the newer forms of biofuels, including ethanol, could really be cleaner and more efficient than traditional forms of biofuels. Favorable CO2 bal­ance and other emissions reduction when bioethanol is burned in engines could help mitigate global climate change. Even if aldehyde emissions slowly increase, the use of catalysts normally destroys this contaminant easily.

Biofuel production may introduce new environmental risks and new chal­lenges for the producing countries. This will particularly be the case when natural resource constraints cause greater trade-offs between food production and biofuel production. Respect for the rainforests and all protected areas should be the core of any regulations and policies. There are serious concerns in Latin America and Africa for the deforestation activities.

Fuel ethanol production includes the generation of a large quantity of residues. Effluent treatment will always be an important topic of research. Reduction of air pollution must not cause the increase of soil or water contamination. There are cases of noncompatibility of stillages in Colombia with the soil quality in the sugarcane plantations.

Using ethanol as a vehicle fuel has the potential to reduce nonrenewable energy consumption and greenhouse gas (GHG) emissions. However, acidification, eutrophication, and photochemical smog could increase when compared to using gasoline as liquid fuel. This information should be analyzed for every feedstock, location, and blending to avoid confusion about the environmental performance of fuel ethanol projects.

Life cycle analysis (LCA) enables one to investigate environmental perfor­mance of fuel ethanol used in different concentrations in gasoline. This analysis can include a serious waste reduction (WAR) algorithm for numerical calculation of potential environmental impacts. This type of strategy is an important compo­nent for identifying practices that will help to ensure that a renewable fuel, such as ethanol, may be produced in a sustainable manner.

The dominant factor determining most environmental impacts, such as green­house gas emissions, acidification, eutrophication, and photochemical smog for­mation, is soil-related nitrogen losses. Usually the source of soil nitrogen can include fertilizers.

However, the most important discussion about environmental advantages of fuel ethanol use is the fact that fluxes of avoided GHG emissions from this biofuel production system are found to be much less than from afforestation or reforesta­tion. In countries like Brazil, Colombia, and Peru, where deforestation is advanc­ing significantly, this issue is a real concern and should be watched closely by the entire world (Table 13.4).

[1] Filter paper units (FPU) allow the indirect quantification of enzymatic activity of cellulases. One FPU is equivalent to the cellulase amount needed to form 1 pmol glucose in one minute measured as reducing sugars (or to form 0.10 mg glucose in one minute) during the cellulose hydrolysis reac­tion employing as the substrate Whatman No. 1 paper filter under determined conditions (pH = 4.8, 0.05 M sodium citrate buffer, 50°C). See Ghose (1987).

Hybrid methods for process synthesis make use of mathematical programming techniques based on optimization in combination with the knowledge-based approach. For instance, the synthesis of heat exchange networks has been improved by optimization employing a heuristic approach based on physical phenomena (Gundersen and Grossmann, 1990). A method for process synthesis exploiting the advantages of mathematical programming that uses MINLP techniques and hier­archical decomposition maintaining the consistency of its fundamental principles has been developed. The method consists of solving the whole technological scheme in each decomposition level. Aggregated models (“black box” models) are utilized for the subsystems that are in the lower hierarchical levels, evaluat­ing in this way their interaction with the detailed MINLP problem of the current subsystem that is being analyzed. The subsystems are: reaction, separation, and heat integration (Daichendt and Grossmann, 1997). In this way, the solution of a large-scale MINLP problem is avoided.

The physical methods employed to remove toxic substances for the subsequent fermentation are based on the separation of these inhibiting compounds by tech­niques such as evaporation, extraction, or adsorption (Table 4.6). During the detoxification by evaporation, the hemicellulose hydrolyzate is evaporated in order to separate the volatile fractions from nonvolatile ones. Acetic acid and phenolic compounds are among the toxic volatile compounds that can be par­tially removed by this technique. The limited increase in the fermentability of the hydrolyzates using the evaporation can be explained by the lower inhibition effect of the removed volatile compounds compared to the phenolic compounds that are not removed (Oliva, 2003). The application of this procedure is mostly car­ried out at laboratory scale using rotoevaporators (Palmqvist and Hahn-Hagerdal, 2000a). The detoxification by solvents employs extractive agents like different organic solvents. The inhibitory substances exhibit greater affinity for these sol­vents than for the aqueous medium of the hydrolyzate allowing their migration to the organic phase. Detoxification with supercritical fluids has been proposed for extracting the inhibitors as well (Persson et al., 2002b). In the case of the detoxi­fication by adsorption, the inhibitors are retained on the surface of a solid mate­rial (adsorbent) decreasing their concentration in the hemicellulose hydrolyzate. However, it is possible that useful substances like the glucose also are adsorbed in the adsorbent bed. The usage of molecular sieves as an adsorbing material for partial removal of acetic acid, furfural, and soluble lignin has been proposed (Olsson and Hahn-Hagerdal, 1996).

During the process synthesis procedures aimed at identifying those technological configurations for fuel ethanol production with better technoeconomic and environ­mental performance, the authors of this book carried out a comprehensive review of the available models for describing ethanolic fermentation kinetics (Sanchez, 2008). Considering that the three main types of feedstocks employed for fuel ethanol pro­duction can be broken down into glucose, the first stage of the analysis included the selection of the most appropriate kinetic models that take into account the growth of S. cerevisiae or Z. mobilis on glucose-rich media. As a kind of “sample” that evidences the great diversity of kinetic models studied, some kinetic relationships for cell growth rate are shown in Table 7.4. In this table, substrate consumption and product formation rates are not shown by space limitations, but these expres­sions depend, in turn, on either cell growth rate or cell concentration. All the pre­sented models consider the substrate limitation, which is mostly expressed through Monod-type equations (Monod, 1949). Similarly, all the models include expressions describing the growth rate inhibition by the ethanol formed. One of the most useful models corresponds to number 4 in Table 7.4, which describes the alcoholic fermen­tation using yeasts from glucose (Lee et al., 1983). In particular, this model has been employed for continuous processes, as reported by Tsuji et al. (1986). It was one of the models employed for the subsequent stages of process synthesis procedures to be discussed in this book. On the other hand, number 5 proposed by Garro et al. (1995) was selected for fermentation using Z. mobilis.

Cell biomass recirculation is a practice employed to increase cell concentration inside the fermenter in order to diminish the cultivation time in batch fermentation or the residence time in continuous fermentation. Considering that this type of configuration will be analyzed during process synthesis, it is necessary to include

TABLE 7.4

Some Kinetic Models Describing Specific Cell Growth Rate during Ethanologenic Fermentations employing Glucose-Based Media

mathematical expressions taking into account such recirculation in the models to avoid a sham growth of cell biomass concentration within the fermenter. In addi­tion, microorganisms undergo stress when cell concentration increases above 100 g/L in the case of yeasts. Another issue to be considered is the substrate inhibition that limits the utilization of very concentrated media due to factors like catabolic repression and osmotic pressure exerting a negative influence on cell growth rate. Therefore, one of the expressions chosen for yeast growth rate (rX) is as follows:

where S, X, and P are the concentrations (in g/L) of substrate (glucose), cell bio­mass (S. cerevisiae), and product (ethanol), respectively, ,nmax is the maximum specific growth rate (in h-1), KS and Ki are the semisaturation and substrate inhibi­tion constants, Pm is the maximum product concentration, and Xm is the maximum

cell biomass concentration in the medium. The models describing the substrate consumption rate (rS) and product formation rate (rP) are represented by the fol­lowing equations:

*=4 £ jr a2)

= (£)* ^

where Yx/S is the cell biomass yield from substrate (in g cell/g substrate) and YP = YX/S/YP/S (in g cell/g EtOH); YP/S is the product yield from substrate (in g EtOH/g substrate).

Substrate consumption rate and product formation rate for the bacterium Z. mobilis were selected according to the model 5 of Table 7.4, which was slightly modified to consider cell biomass recirculation as follows:

*=^^(1-|:П1-0 (7.4)

rp — a*x + pX

where a and P are kinetic constants.

For sucrose-based media, the Equations (7.1) through (7.6) were also applied using not only glucose as a substrate, but also fructose. In this case, it was con­sidered that the cell growth rate is equal to the sum of the growth rates from both substrates. For starch-based media, the same kinetic expressions for glucose con­tent were used considering the content of this sugar in the saccharified starch.

The stoichiometric approach can be useful as an approximation method to describe the ethanologenic fermentation from media containing glucose, fruc­tose, or sucrose. In this case, the overall biological transformations are considered through specific key chemical reactions including the global reaction for biomass cell biosynthesis. Thus, the fermentation employing sucrose-based media may be described by the following reactions:

C12H22On + H2O ^ C6H12O6 + C6H12O6 Sucrose Water Glucose Fructose

C6H12O6 + 1.14 NH3 ^ 5.71 CH18Oo.5No.2 + 2.57 H2O + 0.29 CO2 Glucose Ammonium Yeast Water Carbon dioxide

C6H12O6 + 1.14 NH3 ^ 5.71 CH18Oo 5No.2 + 2.57 H2O + 0.29 CO2 Fructose Ammonium Ethanol Water Carbon dioxide

C6H!2O6 ^ 2 C2H5OH + 2 CO2 Glucose Ethanol Carbon dioxide

C6H!2O6 ^ 2 C2H5OH + 2 CO2 Fructose Ethanol Carbon dioxide

The yeast cells are assumed to have a “molecular formula” that is derived from the elemental analysis of cell biomass. Despite its simplicity, this approach has the advantage of providing suitable relationships to perform mass and energy balances required to simulate the alcoholic fermentation process. Moreover, other transformations can be considered by adding more stoichiometric reactions describing, for example, the formation of fermentation by-products like acetalde­hyde and glycerol:

C6HUO6 ^ C3H5(OH)3 + C2H4O + CO2 Glucose Glycerol Acetaldehyde Carbon dioxide

The biological transformations can be directed by adjusting the correspond­ing conversions of substrate into products. This approach is particularly useful when modular-sequential process simulators are used, such as Aspen Plus (Aspen Technology, Inc., USA) or SuperPro Designer (Intelligen, Inc., USA). These com­mercial simulators have modules for calculating chemical and biochemical trans­formations according to this approach, though a kinetic approximation can be applied through other different modules.

A reasonable approach for increasing the productivity of alcoholic fermentation is the removal of the product that causes the inhibition through an extractive bio­compatible agent (solvent) that favors the migration of ethanol to solvent phase, a process known as extractive fermentation. Some fatty alcohols have the abil­ity of being used as extractive agents thanks to their ethanol selectivity and low toxicity for cells of ethanol-producing microorganisms. In an early work, Minier and Goma (1982) showed that primary aliphatic alcohols with a chain length hav­ing fewer than 12 carbons inhibit the growth of yeast cells. They chose the fatty alcohol n-dodecanol as a solvent for in situ extraction of ethanol in a special continuous pulse-packed column with immobilized cells of S. cerevisiae. This configuration allowed the utilization of very concentrated glucose feed due to the reduction of ethanol in the culture broth. In addition, immobilization seems to protect the cells against solvent toxicity (Aires Barros et al., 1987; Cardona and Sanchez, 2007; Tanaka et al., 1987).

The simultaneous scheme of the integrated extractive fermentation process is shown in Figure 9.14. In this configuration, the solvent is directly added to the cul­ture broth where it comes in contact with ethanol. Ethanol migrates to the organic (solvent) phase, whereas the other components of the culture broth remain in the aqueous phase (cells, substrates). Liquid medium from this bioreactor is continuously removed in order to carry out the decantation of both phases. An aqueous phase with a reduced content of ethanol is sent through the distillation step and the ethanol-rich solvent phase is flashed for the regeneration of the solvent, which is recycled back to the bioreactor, and the production of a concentrated solution of ethanol.

Some configurations for ethanol production by extractive fermentation have been proposed, as presented in Table 9.10. Gyamerah and Glover (1996) imple­mented a process where the fermentation stage was coupled with an apparatus for liquid-liquid extraction in a continuous regime at pilot-scale level. They chose я-dodecanol for its very low toxicity for ethanol-producing microorganisms. However, this solvent has some drawbacks: it tends to form a stable emulsion with the culture broth, its melting point is relatively high (26°C) considering fermenta­tion conditions, and its distribution coefficient related to water is not very high (Kollerup and Daugulis, 1985). In addition, Kirba^lar et al. (2001) experimentally showed that small amounts of water migrate to я-dodecanol in water-ethanol — я-dodecanol ternary systems. Weilnhammer and Blass (1994) proposed a simple model based on the mass balance of different components for the description of extractive fermentation with T. thermohydrosulfuricum using oleyl alcohol as a solvent. This model allowed the evaluation of the economy of the process with and without solvent based on production costs. Kollerup and Daugulis (1985) pro­posed a mathematical model for describing an extractive fermentation process for continuous production of ethanol from a glucose-containing medium. In this model, a simple relationship between ethanol concentration in aqueous phase and ethanol content in solvent phase was considered. Additionally, kinetic description of microbial growth did not take into account the inhibition effect due to high concentrations of substrate. Fournier (1986) developed a rigorous description con­sidering the use of UNIversal Functional Activity Coefficient (UNIFAC) equa­tions for continuous extractive fermentation (Cardona and Sanchez, 2007).

Further integration could include a process where simultaneous saccharifica­tion and extractive fermentation may be carried out, as reported by Moritz and Duff (1996). These authors used oleyl alcohol and demonstrated that this solvent produced no effect either on cellulases or cell biomass, i. e., it was biocompatible for both biological agents involved in the process. Oliveira et al. (1998, 2001) proposed an extractive biocatalytic process in which ethanol produced by yeasts is extracted by oleic acid and used as substrate for lipase-catalyzed esterifica­tion reaction with this same acid. In this way, the combination of the enzymatic reaction and the extractive fermentation in a single vessel improves the product extraction. Acceptable results for broths with a high concentration of glucose (300 g/L) were obtained, although experiments without lipase (only extractive fermen­tation) indicated that oleic acid is not a good extracting agent for ethanol because its concentration in the aqueous phase was higher than in the solvent phase at the end of fermentation. Kang et al. (1990) employed a hollow-fiber membrane reactor for carrying out extractive fermentation with yeast using oleyl alcohol and dibutyl phthalate, obtaining productivities of 31.6 g/(L x h). Fournier (1988) proposed a mathematical model for this type of hollow-fiber membrane extractive ferment­ers that predicts significant improvements in productivity related to conventional CSTR with solvent extraction. The utilization of a liquid-lift external loop biore­actor where the solvent is sparged into the base of the column containing a second liquid (the broth) of higher density has been proposed for the extraction of ethanol as well (Modaressi et al., 1997; Stang et al., 2001).

Reaction-Separation Integration for Alcoholic Fermentation Processes through Ethanol Removal by

TABLE 9.10 (Continued)

Reaction-Separation Integration for Alcoholic Fermentation Processes through Ethanol Removal by Liquid-Liquid extraction

Source: Modified from Cardona, C. A., and O. J. Sanchez. 2007. Bioresource Technology 98:2415-2457. Elsevier Ltd.

Note: HFMEF = hollow-fiber membrane extractive fermenter, SSEF = simultaneous saccharification and extractive fermentation.

The selection of the solvent is a crucial factor for extractive fermentation tech­nology. Bruce and Daugulis (1991) developed a solvent screening program for evaluating possible extracting agents to be used in extractive fermentation con­figurations. This program considered the biocompatibility of the solvent and uti­lized the UNIFAC activity model for predicting liquid-liquid equilibrium data. Using this program, these authors (Bruce and Daugulis, 1992) identified the sol­vent mixture of oleyl alcohol with 5% (v/v) 4-heptanone as a promising extract­ant due to its reduced inhibitory effect and increased distribution coefficient. A large amount of alcohols have been tested in order to examine their ethanol extracting properties in water-ethanol-solvent systems (Offeman et al., 2005a, 2005b). The collected data have allowed insight into the relationship between the structure of the solvent and its extracting characteristics. For this type of work, molecular simulation can provide a deeper understanding on solvent con­formation and associations among water, ethanol, and solvent. Although these works are intended to the selection of proper solvent for ethanol dehydration, the obtained results are of great value for extractive fermentation studies, provided the needed biocompatibility tests are done (Cardona and Sanchez, 2007).

Another approach for extractive fermentation that has been used is the appli­cation of aqueous two-phase fermentation where two phases are formed in the bioreactor as a result of adding two or more incompatible polymers (Banik et al., 2003). In this way, ethanol can be partitioned between both phases accumulating in the upper layer, whereas cell biomass is accumulated in the lower phase. This allows the separation of the ethanol-rich phase and the dis­tillation of this alcohol reducing its inhibition effect. Nevertheless, the com­plexity of the process in the case of continuous regime and the high cost of polymers have delayed further development in this ethanol production technol­ogy (Cardona and Sanchez 2007).