This chapter focuses on the biochemical conversion of biomass to liquid transportation fuels. It addresses the questions raised in the statement of task related to the application of biochemical conversion to the production of alternative liquid transportation fuels from biomass by discussing the following:

• The technology alternatives for converting biomass to liquid transportation fuels.

• The status of development of biochemical conversion of lignocellulosic biomass to ethanol.

• The projected costs, performance, environmental impact, and barriers to deployment of biochemical conversion of lignocellulosic biomass to ethanol.

• Challenges and needs in research and development (R&D), including basic-research needs for the long term.

• Other technologies for converting biomass to liquid fuels that are not likely to be ready for commercial deployment before 2020.

Liquid fuels can be derived from biomass through biochemical processing, chemical catalysis, or thermochemical conversion. Biochemical conversion and chemical conversion typically transform the biomass into sugars as intermediates. In contrast, thermochemical conversion uses heat to convert the biomass into building

FIGURE 3.1 Comparison of biochemical and thermochemical routes for converting biomass to fuels.

Source: Dayton, 2007.

blocks, such as carbon monoxide (CO) and hydrogen (H₂), which can be used for the synthesis of fuels (Figure 3.1). Other thermochemical conversion processes include pyrolysis and liquefaction.

Biochemical conversion uses enzymes to break down structural carbohydrates (for example, the cellulose¹ and hemicellulose² found in plant cell walls) into sugars, which are transformed into alcohols, organic acids, or hydrocarbons by microorganisms in fermentation. The conversions typically take place at atmospheric pressure and temperatures ranging from ambient to 70°C.

Early ethanol production technology based on biochemical conversion of sugar and starch has been deployed commercially. In that technology, ethanol is

produced when wild-type yeast ferments six-carbon sugars. Sugar can be obtained directly from sugarcane (Brazil) and sugar beets (Europe) or indirectly from the hydrolysis of starch-based grains, such as corn (United States) and wheat (Canada and Europe). In the latter case, the starch feedstock needs to be ground to a meal that is hydrolyzed to glucose by enzymes. The resulting mash is fermented by natural yeast and bacteria. Finally, the fermented mash is separated into ethanol and residues (for feed production) via distillation and dehydration (Figure 3.2).

Corn grain is the major source of ethanol in the United States, and its potential for growth is defined by production efficiencies, food-versus-fuel debates, and the question of sustainability and carbon footprint. Developments aimed at future processes are targeting cellulose conversions that could address those issues by providing a growth potential, a low carbon footprint, and sustainability. The infrastructure that was established by the corn grain ethanol industry will benefit the future cellulosic-ethanol industry because the use of ethanol as a transportation fuel has been proved to be feasible, a distribution system exists, and automobiles with internal-combustion engines that use ethanol efficiently are on the road.

Recent analyses of the full life cycle of corn grain ethanol have indicated

FIGURE 3.2 Schematic representation of bioprocessing elements.

that it provides society with small net energy gains over the fossil energy needed to produce it (Farrell et al., 2006; Hill et al., 2006) and might lead to only small net greenhouse gas advantages (Farrell et al., 2006; Hill et al., 2006; Fargione et al., 2008; Searchinger et al., 2008) or might release more greenhouse gas than do production and combustion of an energetically equivalent amount of gasoline once direct and indirect land-use changes are taken into account (Fargione et al., 2008; Searchinger et al., 2008). Issues with corn grain ethanol have led to increased interest in second-generation biofuel feedstocks—including switchgrass, Miscanthus, hybrid poplar, and the other lignocellulosic feedstocks—and in conversion methods that potentially can make biofuels that, relative to corn ethanol, offer larger energy gains and greenhouse gas benefits and reduced competition with food crops.

The development of biofuels needs to move toward conversion of lignocellulosic materials (so-called second-generation biofuels) that are unused agricultural or forestry residues, agricultural cover crops, dedicated perennial crops grown on marginal lands that are not suitable for commodity-crop production even with high commodity prices, or municipal solid wastes. The need to move away from corn grain ethanol is highlighted by the renewable fuel standard (RFS) as amended in the 2007 Energy Independence and Security Act. The RFS mandates that production of ethanol from corn grain level off from 2008 to 2015 and that production of cellulosic and other advanced biofuels increase from 2008 to 2020. The key differences in production between grain ethanol and cellulosic ethanol are the pretreatment of the biomass and the use of by-products (Figure 3.2). This chapter focuses on the conceptual design, conversion technologies, and economics of the biochemical conversion of cellulosic biomass to ethanol. It will also discuss other technologies to produce advanced biofuels that use nonfood renewable feedstocks. The other technologies could produce fuels more desirable than ethanol—for example, lipids, higher alcohols, hydrocarbons, and other products that can be separated by low-energy distillation. New routes of biochemical conversion of biomass to liquid fuels will probably encounter complications as they are being developed and scaled up; these issues will have to be addressed in a continuous R&D program.

In contrast with biochemical conversion, chemical conversion uses inorganic catalysts in a series of aqueous-phase reactions to convert sugars to hydrocarbons that

can be used as fuels. It is a developing technology that will not be ready for commercial deployment by 2020, but it is discussed later in this chapter.

In what is currently the most developed thermochemical route, biomass is initially converted into CO and H₂ via gasification. The gas stream can be cleaned of impurities and shifted to the needed H₂:CO ratio, and CO₂ can be removed to produce a gas stream that can be catalytically converted to liquid fuels by several routes, including Fischer-Tropsch (FT) and methanol synthesis followed by methanol-to-gasoline (MTG) conversion. Thermochemical conversion is discussed in Chapter 4. Other thermochemical conversion routes involve production of bio-oil by pyrolysis or liquefaction and refinement of the bio-oil (Huber et al., 2006); this technology is not as well developed as FT or MTG.

This section discusses the biochemical processes for converting cellulosic biomass to ethanol in a biorefinery. The processes discussed here occur at the end of the supply chain, when the biomass has been delivered to the biorefinery (Figure 3.3). The process economies are those within the biorefinery.

The biochemical conversion of cellulosic biomass involves six major steps: feedstock preparation, pretreatment to release cellulose from the lignin shield, saccharification (breaking down of the cellulose and hemicellulose by hydrolysis to sugars, such as glucose and xylose), fermentation of sugar to ethanol, and distillation to separate the ethanol from the dilute aqueous solution (Figure 3.4). In the sixth step, the residues, primarily lignin, can be combusted to provide energy (Figure 3.4). The integration of those steps with each other and with the living microorganisms and enzymes that carry out the catalytic conversions in a biorefinery is essential to the development of cost-effective processes.

FIGURE 3.3 Logistics of bioprocessing to convert cellulosic biomass to ethanol.

FIGURE 3.4 Unit operations of a biorefinery. A biomass-based biorefinery should be energy self-sufficient or could even sell excess power to the grid. CO₂ is recycled into plant matter through biomass production.

Some feedstocks have to be washed to remove inorganic and other undesirable materials before pretreatment. Whether washing is needed depends on the source and the manner of storage before the feedstock is delivered to the conversion facility. The biomass is then chopped or ground to the desirable size range to feed into

the pretreatment stage. The extent of grinding and size reduction will depend on the type of biomass and the pretreatment technology being used. Cellulosic feedstock can be chopped or ground with existing forestry or agricultural techniques.

Producing fuel ethanol from lignocellulosic feedstocks has been challenging because of the recalcitrant nature of the cellulose that is embedded in the plant cell-wall structure. Therefore, pretreatment is a key step in production of cellulosic ethanol. Pretreatment greatly increases the rates and extents of enzyme action in breaking down cellulose to fermentable sugars (Ladisch et al., 1978) by improving the accessibility of the structural carbohydrates in the cell wall (Figure 3.5). Yields of fermentable sugars from untreated native lignocellulosic materials are low because of the highly packed cellulose structure and the presence of hemicellulose and lignin, which shield cellulose from acid or enzymatic hydrolysis.

Maximizing the use of all lignocellulosic material that is capable of yielding simple (six- and five-carbon) sugars is essential for improving ethanol yield and

FIGURE 3.5 Schematic of pretreatment to disrupt the physical structure of biomass. Reprinted from Mosier et al., 2005. Copyright 2005, with permission from Elsevier.

lowering the cost of ethanol production. Hence, pretreatment of lignocellulosic material is required to improve the hydrolytic efficiency of cellulose by removing and hydrolyzing hemicellulose, by separating the cellulose from the lignin, and by loosening the structure of cellulose and thereby increasing its porosity. The pretreatment of lignocellulosics is particularly important for enzymatic hydrolysis to reduce the amount of enzyme and the time required to convert cellulose to glucose.

Among the various pretreatment methods, hydrothermolysis with steam or water has been shown to be effective in removing and solubilizing hemicellulose and thus in improving hydrolytic efficiency (Mosier et al., 2005; Wyman et al., 2005a,b). Hot-water pretreatment of lignocellulosic biomass at a controlled pH effectively dissolves hemicellulose and some of the lignin and minimizes the formation of monosaccharides and other coproducts that could interfere with biological processes downstream (Yang and Wyman, 2008). For example, monosaccharides inhibit cellulase in the hydrolysis of cellulose downstream. The sugars could degrade further to form such toxic substances as furfural during the pretreatment step (Ladisch et al., 1998; Kim and Ladisch, 2008; Hendriks and Zeeman, 2009). Other pretreatments are similarly effective, and they use acid, bases, ammonia, or other materials (Mosier et al., 2005; Jorgensen et al., 2007; Murnen et al., 2007; Sendich et al., 2008; Yang and Wyman, 2008; Hendriks and Zeeman, 2009). Several of the promising pretreatment methods have been demonstrated on a pilot scale, but the lowest-cost approach is yet to be determined.

In the saccharification step, the cellulose polymers (long chains of sugar) are broken down by hydrolysis into five-carbon and six-carbon sugars (xylose and glucose) for fermentation into alcohol (Figure 3.6). The enzymes used for hydrolysis are referred to as cellulolytic enzymes, and they are classified into three main groups: cellobiohydrolases, endoglucanases, and beta-glucosidases. The cellobiohydrolases and endoglucanases are modular proteins with two distinct independent domains; the first domain is responsible for the hydrolysis of the cellulose chain, and the second is a cellulose-binding domain (CBD) that has the dual activity of increasing adsorption of cellulolytic enzymes onto insoluble cellulose and affecting cellulose structure. A schematic of the action is shown in Figure 3.7. By intercalating between fibrils and surface irregularities of the cellulose surface, CBDs help to reduce particle size and increase specific surface area. Microscopy of cellulose

FIGURE 3.6 Schematic diagram of bioprocessing of sugars to ethanol through enzymatic hydrolysis (catalytic step that frees the sugars) and microbial conversion of sugars to ethanol and CO₂, which are formed in approximately equal parts. Lignin remains unconverted.

FIGURE 3.7 Schematic representation of mechanisms of enzyme action.

Source: Reprinted from Mosier et al., 1999. Copyright 1999, with permission from Springer.

treated with isolated CBDs generated from recombinant organisms has shown the release of small particles from insoluble cellulose with no detectable hydrolytic activity and an increase in the roughness of highly crystalline fibers.

The cellobiohydrolases are the most important cellulolytic enzyme group in that cellobiohydrolase I makes up 60 percent of the protein mass of the cellulo-

lytic system of Trichoderma reesei, and its removal by gene deletion reduces overall cellulase system activity on crystalline cellulose by 70 percent. The concerted effects of pretreatment and enzymatic hydrolysis affect the plant at the cellular level as illustrated in Figure 3.8. According to the prevailing understanding, cello-biohydrolases attack the chain ends of cellulose polymers to release cellobiose, the repeat unit of cellulose. Endoglucanases decrease the degree of polymerization of cellulose by attacking amorphous regions of cellulose through random scission of

FIGURE 3.8 Scanning electron microscopic images of enzymatically hydrolyzed 425–710 mm corn stover pretreated with hot water (at 500X magnification). (A) 3-h enzymatic hydrolysis, 43.3 percent glucose conversion. (B) 24-h enzymatic hydrolysis; 56.8 percent glucose conversion. (C) 72-hour enzymatic hydrolysis; 64.2 percent glucose conversion. (D) 168-h enzymatic hydrolysis; 63.1 percent glucose conversion. The images from a laboratory experiment illustrate how enzymatic hydrolysis of corn stover pretreated with hot water is connected to pore formation (during pretreatment) and enlargement (during hydrolysis).

Reprinted from Zeng et al., 2007. Copyright 2007, with permission from Wiley-Blackwell.

the cellulose chains. Beta-glucosidase completes the process by hydrolyzing cellobiose to glucose. Cellulolytic systems, such as those in filamentaous T. reesei, have enzymes in all three groups: two cellobiohydrolases, four endoglucanases, and one beta-glucosidase (Mosier et al., 1999). The mechanism by which cellulolytic enzymes act in hydrolyzing cellulose is complex and requires a system of different enzymes to achieve deploymerization of the oligosaccharides to monosaccharides, such as glucose and xylose. Most studies have been done with cellulases, which are produced industrially from T. reesei.

The development of the cellulases has resulted in effective systems that are capable of hydrolyzing cellulose to glucose almost completely. Similar studies are being done on hemicellulases, enzymes that are responsible for breaking down hemicellulose to xylose. Hemicellulases are not as well developed as cellulases.

Despite the complexities, much progress has occurred in the development of enzymes for the hydrolysis of pretreated cellulose. Costs are being reduced, with the ultimate goal of combining cellulases with glucose- and xylose-fermenting microorganisms in a concept referred to as consolidated bioprocessing (Lynd et al., 2005). Hydrolysis and fermentation (combined bioprocessing) are being demonstrated on a pilot scale with the goal of reducing costs.

Pretreatment and enzymatic hydrolysis of plant matter—such as wood, corn stover, or grasses—result in a mixture of five-carbon and six-carbon sugars. Many microorganisms, particularly yeasts, will ferment glucose to ethanol. Typically, however, 25–30 percent of the sugar derived from likely candidates as cellulosic feedstocks for bioprocessing (for example, hardwoods, agricultural residues, and some types of grasses) are pentoses—sugars that have five carbon atoms rather than six carbon atoms. Other potential sources of biomass, such as softwoods, have a lower proportion of hemicelluloses and hence fewer pentoses. Pentoses are not readily fermented to ethanol, so yeasts or bacteria that have been genetically modified to ferment both hexoses (six-carbon sugars) and pentoses are needed to maximize the yield of ethanol from cellulosic materials. Some researchers have been successful in engineering microorganisms that are able to use pentose efficiently but cannot do so naturally to produce ethanol. An alternative would be to supply ethanol-producing microorganisms with pentose-using pathways (Nevoigt, 2008). Development of such microorganisms presents a number of challenges. They have to be capable of fermenting the sugars to ethanol, and they have to

be sufficiently robust to withstand antimicrobial agents released by the pretreatment and hydrolysis of the lignocellulose and to withstand relatively high alcohol concentrations.

The glucose and xylose that result from the saccharification step are fermented into ethanol by microorganisms. Traditionally, fermentation is a separate step from saccharification. As noted above, the ideal development would be organisms that could do both simultaneously. Although Saccharomyces cerevisiae (wild-type yeast) has been used for fermentation in ethanol production from corn grain, it cannot ferment xylose sugars obtained from lignocellulose unless it is genetically modified (Sonderegger et al., 2004). Organisms that ferment xylose and glucose have been developed through metabolic engineering (Aristidou and Penttila, 2000).

The composition of the biomass feedstock determines the amount of ethanol that can be produced per ton of biomass (Table 3.1). The ultimate yield is determined by the maximum yield of sugars that can be obtained from a given biomass type, and the yield of sugars is determined by the combined starch, cellulose, and hemicellulose content of the biomass. The ethanol yield can range from 105 to 135 gal/ton (on a dry-weight basis) if all bioprocessing steps occur at 100 percent efficiency, that is, all the structural carbohydrates—starch, cellulose, and hemicellulose—are used to produce ethanol (Table 3.1). Because the efficiency is typically less than 100 percent—ranging from about 95 percent overall for a corn-grain-to-ethanol plant to 50 percent with some current cellulose-conversion technologies—the actual yields are substantially lower. Ethanol yields can be improved with a combination of advanced pretreatments and enzymes to improve cellulose-conversion efficiency and with the application of fermenting microorganisms that are able to convert glucose, xylose, and other pentoses to ethanol (Wyman et al., 2005a).

Cellulolytic enzymes are subject to product inhibition in which the rate of an enzymatic reaction is inhibited by the end products of the reaction (Gong et al., 1977, 1979; Ladisch et al., 1981). Some work has been done on process strategies and modified enzymes to reduce that inhibition, but more is needed. Simultaneous saccharification, in which the enzyme and the microorganism are used in the same tank, could enable reduced enzyme use because the reaction is not inhibited by the product of enzymatic action, glucose or cellobiose (Gauss et al., 1976; Takagi et al., 1977; Wright et al., 1977; Saddler et al., 1982a,b; Wyman et al., 1986; Spindler et al., 1987; Lynd et al., 2002). The ethanol product that is formed also inhibits enzymatic activity, but to a smaller extent than glucose and cellobiose do.

The reduction in inhibition allows the reaction to proceed to the final desired product—ethanol in this case—more rapidly.

The ethanol solution from the fermentation step is distilled to produce 95 percent ethanol. Ethanol is further dried with molecular sieves to produce the required purity. The distillation requires large quantities of energy (Katzen et al., 1981; Shapouri et al., 2002). Distillation of ethanol is a well-established technology and is used in commercial corn grain ethanol plants. However, recovery of solids will require improved engineering design.

The bottoms from distillation are centrifuged to concentrate them. Solids from other portions of the process might also be added. The residual solids are rich

in lignin and can be burned in a boiler to generate steam and electricity for the biorefinery. Pathways of using lignin as a fuel with low carbon emission are well defined.

A substantial amount of ethanol has been produced in the United States since the 1970s, and the feedstock of choice has been corn grain. The process of producing ethanol from corn grain has been steadily improved in efficiency, and costs have been greatly reduced. The process can be considered fully commercial, well understood, and optimized. There were more than 130 corn grain ethanol plants in the United States in 2008. In contrast, there are no large-scale commercial cellulosic-ethanol plants, although many of the components are in pilot demonstration. In February 2007, the Department of Energy announced that it would invest up to $385 million for six biorefinery projects over 4 years to bring cellulosic ethanol to market (DOE, 2007b), and a number of private companies are actively pursuing pilot demonstration of integrated cellulosic plants, which could lead to commercial-scale demonstrations and eventual deployment of cellulosic-ethanol plants. If the current effort in development and demonstration of cellulosic ethanol is sustained or accelerated, cellulosic-ethanol plants could be ready for commercial demonstration in the next 3–5 years, and the first plant could be built by 2015.

As discussed above, the key issues associated with cellulosic ethanol are related to the ability to develop enzymes and microorganisms that can break cellulose bundles and depolymerize cellulose and hemicellulose to produce soluble sugars, and also to the ability to produce the enzymes and microorganisms at a reasonable cost. Microorganisms capable of fermenting five-carbon and six-carbon sugars in separate steps are available, and their performance is rapidly being proved in the laboratory. The challenge will be to demonstrate their robustness under industrial conditions. A longer-term challenge involves development of microorganisms and enzymes that can break down cellulose structures, depolymerize them to sugars, and ferment the sugars in the same vessel. If hydrolysis and fermentation are to be carried out separately, the development of microorganisms that can withstand and ferment sugars at higher temperatures would provide opportunities to increase the rate of fermentation and reduce the need for cooling the fermentation itself. The

amount of energy that would be required to heat the broth to distillation temperature would also be reduced.

One research gap is an incomplete understanding of the molecular biology of the plants that provide the feedstock and of the microorganisms that provide the biocatalysts and enzymes for transforming the feedstocks into ethanol and other biofuels. Fundamental studies on the structure and function of the enzymes that catalyze the breakdown of cellulosic components to fermentable sugars are key to improving the rates at which the enzymes operate and to reducing product inhibition. The nature and chemical structure of inhibitors and how they interact with the active catalytic sites of the enzymes need to be defined so that strategies to mitigate the inhibition can be developed at the catalytic level or through engineering approaches to remove the inhibitors by bioseparation.

Another research gap is in the fundamental understanding of how the plant cell-wall structure resists enzyme attack. Genomic studies could provide insights into how lignin (the key agent in resistance of the cell wall to enzymatic conversion) might be modified at the molecular level to enable facile transformation but retain the properties that provide resistance to pathogens. The molecular biology and metabolic pathways involved in directing the flow of carbon into ethanol or other advanced biofuels and the manner in which modified bacteria and other microorganisms are able to break down cellulose are also important topics. Research will help to reduce enzyme costs, a major component of cellulose-conversion costs. Research needs to be pursued on pretreatments that exploit knowledge gained and biochemical engineering that will define design principles of low-cost manufacturing through enzyme-catalyzed reactions, thermal processing, fermentation, and advanced bioseparation. Fundamental research would also address how temperature, pressure, and pretreatment media interact at the microscopic and macroscopic levels so that plant cells are readily attacked by biological or chemical catalysts.

The corn grain ethanol plants have created an industry, and they provide a basis for estimating some of the capital and operating costs of cellulosic-ethanol plants. Several key studies also provide information on and analysis of the design and performance of cellulosic-ethanol plants (Aden et al., 2002; Johnson, 2006; Kwiatkowski et al., 2006; Kim et al., 2008).

On the basis of experience gained from corn-ethanol plants, published studies, and its own expertise, the panel used SuperPro Designer^® to estimate the capital and operating costs of ethanol-production plants (Intelligen, 2009). SuperPro Designer is a chemical-process simulation software that contains a set of unit operations that can be customized to the specific modeling needs of the corn grain-to-ethanol and cellulosic biomass-to-ethanol processes. It has a well-developed economic-evaluation package, which includes an extensive database of chemical components and mixtures, extensive equipment and resource databases, equipment sizing and costing information, and economic-evaluation parameters—such as financing, depreciation, running royalty expenses, inflation rate, and taxes—for cost estimates.

A grain-ethanol plant was used as the baseline to validate the model and input parameter values. Once validated, the model was applied to estimate capital and operating costs for cellulosic-ethanol production. Most of the unit operations for cellulosic-ethanol production are similar to those for grain-ethanol production.

The panel decided to develop its own model for cost estimates. The process-flow diagram used for the panel’s modeling is shown in Figure 3.9. The panel conducted three sets of analyses with the cellulosic-ethanol plant model. First, it assessed the sensitivity of the capital and manufacturing cost estimates with respect to such process parameters as enzyme cost, types of feedstock, plant size, solids loading, and yields in the pretreatment, hydrolysis, and fermentation steps. The results are presented in the form of process cost estimates for three representative scenarios: the most pessimistic, with little advancement in technologies or process efficiencies from where they are today (2008) to reduce costs (low); reasonable advancement (medium); and the most optimistic (high) in terms of technology and process improvement. Second, the panel assessed the sensitivity

FIGURE 3.9 Process-flow diagram for ethanol production from lignocellulose.

Source: Schwietzke et al., 2008a. Reprinted with permission from IEA Bioenergy.

of operating costs to different types of feedstock. As mentioned earlier, process efficiency depends on feedstock composition. Assuming reasonable advancement in process technologies (that is, using the medium case), the panel compared the costs of ethanol produced from different biomass alternatives. Third, the panel assessed whether and how the capital and operating costs of a cellulosic-ethanol plant vary with size; in this analysis, the panel used the medium case to estimate costs and assumed that plants of different size used poplar chips as a feedstock.

The panel used a biorefinery with a capacity of 40 million gallons of ethanol per year as the basis of its calculations and assessed the effect of biorefinery size later. The capacity was used because it is comparable with that of grain-ethanol plants, matches well with previous studies and current cellulosic-ethanol projects in the planning stage, and could serve well those geographic areas that have a lower concentration of biomass. The distribution of biomass could pose a limit on the volumes that could be economically supplied to a biorefinery in some areas, and transportation of large volumes of biomass over a long distance could be expensive and consume more fossil fuel (see Chapter 2).

Producing 40 million gallons of ethanol per year with a fermentor effluent concentration of 4–5 percent ethanol by weight requires nine fermentors, each of which has a capacity of 800,000 gal if a fermentation time of 72 h is assumed. The 800,000-gal capacity of the fermentors was selected on the basis of typical

industry practice. The run time was assumed to be 80 h/batch. The current analysis is based on an output of 40 million gallons per year. Further economic benefits as a result of scale will be realized as the technologies develop and larger plants are constructed.

In the model, each fermentor has a train of pretreatment and saccharification tanks that work in batch mode. Each train supplies its own fermentor, thereby avoiding the need for holding tanks for intermediate storage and minimizing the potential for fouling the saccharified mash. The pretreatment unit, saccharification tank, and fermentor, therefore, scale linearly with the size of the plant. Other units—such as distillation, centrifugation, shredding, drum drying, and boiler and steam generator—operate continuously. The continuously operating units were sized so that they can supply and process the contents of a single fermentor over a period equal to t/N, where t is the duration of each batch and N is the number of processing units. Because the fermentations are staggered for efficient loading and emptying, overall batch time is set at 80 h, which is the time required for cycling the full set of 8–10 fermentors. Although fermentations operate in batch mode, using multiple vessels enables a staged operation so that the upstream processing, including pretreatment, can be carried out continuously.

The major process determinants that affect the size and operation of a plant are

• Yield in saccharification,

• Yield in pretreatment,

• Fermentation time and yield, and

• Solids loading.

With the exception of fermentation time, which was set at 72 h for all simulations, the parameters listed above were varied, with the ranges indicated in Table 3.2 representing low-, medium-, and high-performance scenarios. Once those parameters were defined, the design of the various units, the sizing, and the cost calculations were automatically handled by SuperPro Designer. Details of the modeling and analyses are presented in Appendix I.

The following assumptions were made to simplify the modeling:

• Total dry weight is the sum of the weight of soluble and insoluble components.

• Ash, fat, protein, and lignin do not react in the process.

• There is no contamination with lactic acid or acetic acid bacteria.

• Added enzymes add to protein mass balance.

• All batch fermentations take 72 h.

• Distillation and dehydration recover 100 percent of the ethanol.

• The final concentration of ethanol is 99.5 percent by weight.

Some of the assumptions represent ideal operating conditions that might not be achievable all the time. For example, some ash, fat, protein, and lignin could react, or there could be occasional contamination with lactic acid or acetic acid bacteria. However, most of the assumptions do not affect cost substantially.

A solids loading of 30 percent was used in initial modeling. However, for the majority of the modeling, a range of 18–25 percent was used to provide a more realistic assessment. At the time the report was written, 25 percent solids by weight was considered optimistic. The yield range for pretreatment was assumed to be 80–95 percent; the best case assumes that after pretreatment 95 percent of glucans and five-carbon sugar polymers, such as xylans, are accessible to cellulases for saccharification. For saccharification, 90–95 percent of the accessible sugar

polymers³ are assumed to be hydrolyzed into single sugars (monomers⁴) given the concentration of enzymes and the solids loading. For fermentation, in which the final conversion is a multiple of sugar conversion into products and the percentage of maximum theoretical yield of ethanol produced by yeast, the value for sugar conversion into ethanol was assumed to range from 85 to 90 percent for glucose and from 75 to 81 percent for xylose.

Cellulase cost was calculated as cents per gallon of ethanol produced by using the required amount of cellulase per unit biomass dry weight, the ethanol yield per ton of biomass, and the price per kilogram of cellulose. That gave a range of $0.10–0.40/gal of ethanol produced.⁵ That range was validated by discussions with an industry representative (B. Foody, Iogen Corporation, personal communication, March 2008).

Three scenarios were developed in the form of process cost estimates: current technology (low), reasonable evolutionary advances in technology (medium), and the most optimistic technology advances (high) for the biochemical conversion of cellulosic feedstocks (Table 3.3). The panel judges that the reasonable-advances case best represents where the technology would be for 2020 deployment. The optimistic case shows that considerable potential will remain. The results of the calculations were validated via interviews with representatives of three cellulosic-ethanol companies. For proprietary reasons, the level of detail provided in the validation ranged from the expected cost of individual units and materials to overall plant cost. The work was also validated by comparing plant configurations and cost estimates generated herein with those previously reported in the literature.

The initial analysis is a comparison of three assumed scenarios, from most pessimistic to most optimistic as outlined in Table 3.2. Initially, poplar woodchips were used as the feedstock for all cases to illustrate the differences between scenarios in

process economics. Those analyses, however, generated some unexpected results (for example, high electricity generation from the lignin-rich residue of poplar). Poplar woodchips, rich in lignin, can be considered an outlier as a feedstock with respect to the other biomass types. Therefore, cases were run with a high-sugar/glucan biomass (HGBM) that has a composition closer to those of most of the other cellulosic biomass types. The cost estimates are summarized in Table 3.3. The detailed breakdown of costs of major equipment, fixed capital, materials, and other items are detailed in Appendix I.

The total capital cost required to build a 40-million-gallon biorefinery and put it into operation ranges from $174 million to $223 million ($4.34 to $5.65 per annual gallon) if poplar woodchips are used as feedstock and from $128 to $166 ($3.20 to $4.15 per annual gallon) if HGBM is used as feedstock (Table 3.3). The capital-cost requirements for a biorefinery that uses poplar woodchips is about 35

percent higher than the requirements for one that uses HGBM primarily because of the increased cost of the boiler and steam electrical generator associated with the increased lignin content of the feedstock. However, the higher capital cost is offset by the sale of excess electricity generated by burning the lignin residue. (See discussion on energy cost later in this section.)

For the materials costs, two items have important effects on the cost of production for the plant: the cellulases for the saccharification step and the lignocellulosic feedstock. If the cellulases are produced in-house in an enzyme-propagation unit for the plant, the cost of cellulases would be equivalent to the operating cost of producing the enzymes. That cost includes the cost of sugar water and of cofactor additives, the cost of utilities consumed by the propagation unit, and the cost of cleaning and sterilizing the seed fermentors after every batch. The panel estimated enzyme cost to be about $0.10–0.40/gal and used the midpoint of $0.25/gal as the medium case.

The purpose of this analysis is to illustrate the sensitivity of the process economics of biochemical conversion itself. The assumed feedstock costs listed in Table 3.2 are used here. As discussed in Chapter 2, they represent the low cost of feedstock. However, the largest input cost for ethanol production is the cost of cellulosic feedstock, which represents about 75 percent of material costs. Feedstock cost represents about 30–45 percent of overall plant operating costs. Discussions with industry representatives confirmed that even though in some rare cases cellulosic biomass can be obtained at low cost, the price is likely to increase as demand increases.

The external energy requirements of a cellulosic biorefinery are expected to be zero because the lignin and other unconverted residue from processing can be used to generate energy. For a lignin content of 20–30 percent of the feedstock (on a dry-weight basis), the steam generated from the lignin-residue combustion is enough to supply the energy requirements for the entire biorefinery and to export electricity to the grid. The lignin is assumed to have an energy content of 11,500 Btu/lb (26.7 kJ/g). The conversion of steam to electricity by the steam turbine generator was assumed to be 80 percent efficient, considering the steam turbine and generator only. If the process efficiencies of all steps from lignin to electricity are taken into account, the conversion of lignin to electricity is estimated to be about 25 percent efficient. That resulted in an estimated 43 MW of gross electrical

power that could be sold to the grid, which is higher than the Aden et al. (2002) estimate of 29 MW of gross electrical power.

The higher estimate of energy generation is a result of the use of poplar woodchips, which have a high lignin content. The difference between the panel’s estimate and the Aden et al. (2002) estimate was smaller when HGBM was used as a feedstock. The credit gained from the export of excess electricity generated by the biorefinery could be $8–16 million per year and could reduce the manufacturing cost by $0.20–0.40/gal of ethanol.

The estimated labor cost was $3–4 million per year on the basis of discussions with industry representatives. Within the range of production capacity of 20–100 million gallons of ethanol per year, the labor cost is expected to be essentially independent of the size of the biorefinery. Other material costs or consumables presumably could affect production costs if a different pretreatment process were used instead of hot-water pretreatment (used in this model). However, the choice of pretreatment process has a large effect on the overall cost because of its relationship to reducing the costs of cellulase. Nonetheless, biomass feedstock costs are likely to dominate overall product cost unless the conversion process is expensive. Water use could also affect costs, but plants will probably be designed to economize on water use and be located with that variable in mind.

On the basis of those considerations, the model projected the production cost of a biorefinery that uses poplar woodchips to be $1.50–2.70/gal of ethanol (including capital charge) with a middle value of $2.00/gal (Table 3.3). The cost estimates assume that excess electricity generated from the lignin residue will be sold at $0.05/kWh.

Solids loading emerges from the model study as a key parameter for the following reasons:

• The theoretical ideal solids loading of 30 percent has the physical consistency of a solid paste, and even though it might be achieved with substantial technological breakthroughs, many operational issues would have to be overcome—for example, achieving good mixing between the pretreated biomass and the cellulases and achieving greater than 80 or 85 percent saccharification.

• Solids loading affects the capital cost of the plant because for equal production, the equipment volumes required for the pretreatment, saccharification, and fermentation processes vary approximately inversely with solids loading.

• Higher solids loading translates into increased ethanol concentration in the fermentations, and to accommodate this increase, ethanol-tolerant yeasts will have to be developed; this is one of the R&D objectives for more efficient future ethanol processes.

As illustrated above, the capital and operating costs vary with the type of feedstock used in the biorefinery. Therefore, the costs were estimated for poplar woodchips, wheat straw, switchgrass, corn stover, and Miscanthus (Table 3.4). A constant feedstock cost is used in this analysis to demonstrate the effect of feedstock type on capital costs and operating costs of the conversion process. The total capital costs range from $123 to $194 million for a capacity of 40 million gallons of ethanol per year. Because cellulosic-ethanol plants will receive biomass from specific and relatively limited geographic areas, plant designs will be specific to the expected biomass type. If cellulosic feedstock other than poplar is used, the capital costs could decrease by up to 37 percent.

Some aspects of a biorefinery can benefit from economies of scale. Table 3.5 summarizes how sensitive capital and operating costs are to the size (capacity) of a biorefinery. Poplar is used as a feedstock, and reasonable improvement in technologies and process efficiencies are assumed in this illustration. In this analysis, a facility with a capacity of 20 million gallons per year would consume feedstock at about 700 tons/day, and a facility with a capacity of 100 million gallons per year would consume about 3500 tons/day. The yield of ethanol per dry ton of feedstock and the expected revenue from electricity do not vary with the size of the biorefinery. In contrast, capital cost is sensitive to the size of the biorefinery and benefits from economies of scale because larger equipment and construction of a larger facility cost less per unit size than do the equipment for and construction of a smaller one. The capital cost per annual gallon of ethanol produced was estimated to be $5.84 for a 20-million-gallon biorefinery compared with $3.49 for a 100-million-gallon biofinery. The operating costs also vary with size and range from about $1.52/gal for a 20-million-gallon facility to $1.30/gal for a 100-million-gallon facility. Increasing the size of a 20-million-gallon biorefinery by a factor of 5 decreases the annual capital costs per annual gallon by about 40 percent and the operating costs by about 15 percent.

Estimates of carbon dioxide (CO₂) emission in different scenarios or with different feedstocks show that it varies widely from 8 kg to 16 kg of CO₂ per gallon of ethanol produced (Tables 3.6 and 3.7). Those values represent CO₂ emission during the biochemical conversion process and do not take into account CO₂ uptake by photosynthesis during the growth of the feedstock. Net CO₂ emissions from cellulosic ethanol from field to wheel are presented in Chapter 6. The range of values presented in Tables 3.6 and 3.7 is tied to the burning of lignin and other residual materials for generating steam and power. The most effective use of the residual lignin is to generate steam and electricity to power the plant. Wheat straw as a feedstock has the lowest lignin content (Table 3.7) and thus produces the lowest CO₂ emission, but it does not provide enough energy for the biorefinery.

Biochemical conversion processes, as configured today, produce a stream of pure CO₂ from the fermentor that can be dried, compressed, and stored in geologic formations. The concept of using carbon capture and storage (CCS) in

cellulosic-to-liquid fuels biochemical conversion processes, although not discussed before, follows directly from the modeling work done for CCS technology applied to thermochemical conversion in Chapter 4. The panel did not analyze that possibility from a technology standpoint but believes that the required CCS technologies could be integrated into the biochemical-conversion-plant design. The potential issues are scale, capital-cost efficiency, and logistics of the biochemical conversion plant relative to the storage site.

In addition to greenhouse gases emitted by the biorefinery during processing, exhaust emissions are associated with ethanol. They are less toxic than those associated with gasoline and have lower atmospheric reactivity (Worldwatch Institute, 2007). Brazil presents an informative case study. The use of ethanol in Brazil reduced CO emissions from 50 g/km to less than 5.8 g/km by 1995. However, aldehyde emission was found to increase with the use of hydrous-ethanol engines. Total aldehyde emission was higher from engines using ethanol (both neat and blended) than from engines using gasoline. Nonetheless, present ambient concentrations of aldehyde in São Paulo are below the recommended reference for human health (Goldemberg et al., 2008).

The production of cellulosic ethanol requires process water for mixing with fermentation substrates and for cooling, heat, electricity, and reagents that are associated with hydrolysis and fermentation. In the case of thermal processing of cellulose, process water is required primarily for cooling. The amount of water required for processing biomass into ethanol or other biofuels is estimated to be 2–6 gal per gallon of ethanol produced (Aden et al., 2002; Pate et al., 2007; Cornell, 2008). The lower value would be approached if the plant were designed for recycling process water. The processing of cellulosics to ethanol will result in a residual water stream that will need to undergo treatment. However, by definition an efficient process will ferment most of the sugars to ethanol and leave only small amounts of organic residue.

Air emissions will result from either bioprocessing or thermal processing. Fermentation processes release CO₂ as a consequence of microbial metabolism; 1 mole

of CO₂ will be formed for each mole of ethanol produced. Thermal processes release CO₂ as a result of partial combustion of the biomass but also form other volatiles, CO, and H₂. CO and H₂ are the desired intermediates in some processes because they can be passed over a catalyst or fermented to biofuels. Other emissions include water vapor, particularly if the lignin coproduct is dried before being shipped from the plant for use as boiler fuel at an off-site power-generation facility. Sulfur and nitrogen content of the fermentation residues would be expected to be low unless chemicals are used in the pretreatment of biomass. Chemicals used in pretreatment would need to be recovered or otherwise to be in the wastewater or in the solid residues. Because the residues are expected to be burned and used as boiler fuel, air emissions (sulfur in the case of acid pretreatment, nitrogen if nitrogen-containing reagents are used for pretreatment) would increase if the solid residues contain pretreatment chemicals. Emissions from the cellulosic feedstock itself are low, and if processes that transform it to ethanol add chemicals only minimally, emissions will be minimized. Minerals in the cellulosic biomass will end up in the wastewater.

Ethanol has 66 percent as much energy as gasoline does. Ethanol is hygroscopic and cannot be transported in existing fuel-infrastructure pipelines because of its affinity for water. (See Chapter 5 for details on distribution of ethanol.)

There are three key challenges that need to be overcome before widespread commercialization: (1) improving the effectiveness of pretreatment—removing and hydrolyzing the hemicellulose, separating the cellulose from the lignin, and loosening the cellulose structure or using other pretreatment methods (Murnen et al., 2007; Sendich et al., 2008); (2) reducing the production cost of enzymes for converting cellulose to sugars; and (3) reducing capital costs by developing more efficient microorganisms for converting the sugar products of biomass deconstruction to biofuels. The size of some biorefineries could be limited by the supply of biomass. The limit in size could result in loss of the economies of scale that can be achieved with large plants. The costs of products of the first-generation commercial plants could be higher than estimated. One company that has been operating a fully integrated demonstration reported that the capital cost of the plant was substantially higher than predicted by models because of unanticipated problems,

such as the complexity of handling mineral matter in feedstocks (B. Foody, Iogen Corporation, personal communication, March 17, 2008). Costs are expected to decrease as experience is gained.

In addition to the mandated 16 billion gallons of cellulosic biofuel, the RFS as amended in the 2007 Energy Independence and Security Act states that advanced biofuels must reduce net greenhouse gas emissions by at least 50 percent relative to those from gasoline. To meet the goals established by the 2007 act, economically viable lignocellulosic-ethanol production is essential, and technological progress needs to be made. If broad deployment (implementation) is to occur by 2020, the facilities would have to obtain permits by about 2015. Although a series of evolutionary changes are likely to occur between 2008 and 2015, successful deployment by 2020 would require a large, sustained effort to improve technologies and process efficiencies. Demonstrations on a commercial scale have to occur at an aggressive pace to ensure sufficient learning from the activity. Much engineering, technical, and operational knowledge can be gained only from designing and building integrated facilities and then operating them for a reasonable period.

The growing biofuel industry is based on well-established technology for producing ethanol via fermentation and separating it by distillation. Fuel ethanol cannot be added to gasoline before pipeline transportation. The cost of fuel-ethanol transportation is estimated to range from $0.13 to $0.18 per gallon, which is as much as 6 times the cost of transporting traditional petroleum-based fuels (GAO, 2007). Cellulosic ethanol derived from nonfood renewable feedstocks could be a transition to or one of many contributors to a diverse portfolio of alternative fuels as other biofuels are proved, demonstrated, and commercialized. This section discusses some of the technologies for producing other biofuels. As is the case with ethanol derived from renewable cellulosic feedstocks, the technologies that would use sugars as part of the conversion process would be attractive only if the feedstocks and the sugars obtained from them are inexpensive compared with ethanol production.

Approaches to produce hydrocarbon fuels directly from biomass that are analogous to production of fuels from petroleum are being explored (Huber et al., 2006).

One approach produces straight-chain hydrocarbons, mostly hexane, via aqueous-phase hydrogenation of biomass-derived sugars followed by dehydration. All the hydrogen consumed in the process can be obtained from biomass processing. The process is exothermic as a result of oxidation of a portion of the biomass-derived carbohydrates. Because the reactants are dissolved in water, the hydrocarbons produced form a separate phase, and distillation is not required. This process has the potential of higher energy efficiency and shorter residence times than the fermentation and distillation steps used in ethanol production, but considerable development is required to confirm that the potential can be realized in a commercially viable process (Huber et al., 2005).

The product consisting of linear hydrocarbons can be isomerized in a conventional refining process to form branched hydrocarbons with higher octane more suitable for gasoline blending. Conventional refinery alkylation technology can be used to process the low-boiling-point straight-chain hydrocarbons to increase octane and boiling point to the extent needed for gasoline blending. If this kind of production of hydrocarbons from biomass were widely commercialized, refining capability for isomerization and alkylation would probably need to be increased.

Another approach to biohydrocarbon fuels produces high-cetane diesel (Huber et al., 2005). Sugars are first dehydrated and then hydrogenated to form cyclic oxygenated molecules that can undergo aldol condensation (self-addition) to form larger oxygenated molecules that remain soluble in water. The condensation products are hydrogenated and then dehydrated to form mostly straight-chain hydrocarbons ranging from 7 to 15 carbons. The final hydrogenation and dehydration reactions are carried out in a four-phase reactor. The feed streams to the reactor in the four phases are water with dissolved oxygenated hydrocarbon reactants, gaseous hydrogen, solid catalyst, and hydrocarbon (required to reduce coke for-

mation on the catalyst). The process can be modified to produce oxygenated compounds in the diesel boiling range that are soluble in the fuel.

Although those processes have been shown to be feasible in the laboratory with pure feedstocks, much development beyond what has been reported remains. The concepts need to be tested by using biomass-derived feedstocks with recycling and with reactors that can be scaled for commercial operation. The keys to success in the processes appear to be achieving sufficient yield of the hydrocarbon product, developing high-activity catalysts with long-term stability, and minimizing coking reactions. There remains a large amount of R&D to be done on these concepts before commercial applications can be undertaken.

Butanol is another potential entrant into the light-duty-vehicle biofuel market. Butanol is a four-carbon alcohol (ethanol has two carbons). When butanol is made from biomass, it is referred to as biobutanol. The longer hydrocarbon chain makes it fairly nonpolar and thus more similar to gasoline. Biobutanol has a number of attractive features as a fuel: its energy content is close to that of gasoline, it has a low vapor pressure, it is not sensitive to water, it is less hazardous to handle and less flammable than gasoline, and it has a slightly higher octane content than gasoline. Thus, it can go directly into the existing distribution system. It has been shown to work in gasoline engines without modification (DuPont, 2008).

Several technologies to produce biobutanol are in the R&D phase. The one receiving the most attention is the acetone-butanol-ethanol process, which uses the bacterium Clostridium acetobutylicum. This process was used initially to produce acetone for making cordite in 1916. The process produced about twice as much butanol as acetone and also produced acetic, lactic, and propionic acids and ethanol and isopropanol. As currently being commercialized, the process involves the biochemical conversion of sugars or starches from sugar beets, sugar cane, corn, wheat, or cassava to butanol. When the mutated strain Clostridium beijernickii BA101 is used, greater selectivity for butanol is achieved. There are also efforts to develop microorganisms that have an increased rate and selectivity in the conversion of sugars to butanol. These include microorganisms that can efficiently convert the different sugars that are obtained from cellulose and hemicelluloses. Because butanol is toxic to the producing organism, its concentration is limited

to about 15–18 g/L even in the native organism that produces it (for example, clostridia). Isobutanol is less toxic and is also a good fuel component, so a more promising approach to improve the process is to seek or engineer organisms that produce isobutanol.

The obvious extension of that technology is the conversion of cellulose to biobutanol. It depends on the development of biotechnologies for the effective, efficient depolymerization of cellulose and hemicelluloses into the basic sugars, which can then be converted to butanol. The most important development would be metabolic engineering of microorganisms that could depolymerize the biomass components into sugars and then convert the sugars to butanol in the same reactor to reduce capital cost. The cellulose approach to biobutanol is being studied, but the technology is in the research stage and is far from commercial.

In another variation, fermentation in a fixed-bed bioreactor using Clostridium tyrobutyricum produces primarily butyric acid and hydrogen. When the butyric acid is fed into another bioreactor with C. acetobutylicum, the butyric acid is converted to butanol with high selectivity.

Biobutanol’s main challenge now is cost. To address the cost and initiate market entry, some companies are working on retrofitting an existing sugar-based bioethanol plant to produce biobutanol (Chase, 2006; DuPont, 2008). An improved next-generation bioengineered organism is projected to be available by 2010 (Chase, 2006).

Large-scale production of photosynthetic microorganisms (algae) to be used as biomass feedstock for liquid transportation fuels has been contemplated for many years, but uncertainties surrounding production costs have resulted in smaller investments in R&D compared with that in cellulosic biofuels. (See Appendix J for details of systems, strains, and resource requirements for production of microbial biomass.) A major program in this field was funded and managed by the U.S. Department of Energy National Renewable Energy Laboratory (Sheehan et al., 1998). Research in the development of algae that have high lipid productivity is being conducted (Briggs, 2004; Pacheco, 2006). Advances in the metabolic engineering and genomics of algae are leading to new strategies for increasing the utility of algae for fuel production.

Several types of fuel potentially can be produced by photosynthetic microorganisms. To date, the emphasis has been on producing biodiesel via transesterifica-

tion of algal glycerolipids to produce fatty acid methyl (or ethyl) esters, which can be used in diesel engines. Cellular lipids can also be converted via catalytic hydrocracking to a mixture of alkanes suitable for use as a jet fuel or gasoline ingredient. Some algae, such as Botryococcus, produce long-chain hydrocarbons that are potentially usable as fuel after hydrocracking to reduce the chain length of the molecules. Production of ethanol from recombinant photosynthetic microorganisms is also being considered. It involves introducing foreign genes that encode ethanol biosynthetic enzymes into cyanobacteria or microalgae. Tens of thousands of species of cyanobacteria and microalgae occur naturally in a wide variety of habitats, but only a small fraction of them are available in public culture collections. Strain collection and characterization programs are needed to enable large-scale production of biofuels from photosynthetic microorganisms.

The basic resources required for large-scale cultivation of photosynthetic microorganisms for fuel production are land with suitable topography, climate, and sunlight; water of acceptable quality and abundance; and concentrated sources of CO₂ (Maxwell et al., 1985). The co-location of the three resources is important for minimizing production costs. One study estimated that 18–34 liters of water is needed to produce 1 MJ of energy from algae (Dismuskes et al., 2008). That water requirement is similar to that of corn grain ethanol (33 L/MJ). Large quantities of saline groundwater are present in the southwestern United States, much of which could be used to support large-scale cultivation of photosynthetic microorganisms for biofuel production. It is imperative, however, to gain a better understanding of how much water can be removed from saline aquifers without adversely affecting the flow and quality of contiguous freshwater aquifers (if present) and without creating wastewater-disposal issues. Closed photoreactors for culturing algae would use less water than would open ponds.

Many of the challenges related to algal biofuel production are engineering matters associated with how and where to grow the algae to achieve needed productivity. In most production schemes, the algal oil is extracted from harvested algae. Because of the metabolic burden associated with the biosynthesis of high-energy lipids, production strains that accumulate large amounts of oil tend to grow and reproduce more slowly than strains that do not accumulate oil. Open cultures are therefore prone to contamination with undesirable species unless the production strain is able to grow in specialized conditions that restrict the growth of other species (for example, high pH). Alternatively, production strains selected for high growth rates and high biomass yields without regard for oil content can often compete satisfactorily with contaminating strains, but the chemical composi-

tion of the algae would be better suited for anaerobic digestion than to liquid-fuel production. The use of closed photobioreactors can lower the risk of culture contamination substantially, but capital costs of such systems are high. Maxwell et al. (1985) made a rough estimate of the sites in the southwestern United States that have suitable terrain, climate, and water availability. About 1 percent of the area considered (that is, 2 million of the 200 million hectares considered) could be suitable sites of algal-production facilities.

Large quantities of biofuels potentially could be produced by photosynthetic microorganisms. With an average productivity of 0.046 lb/yd² per day (25 g/m² per day) and lipid content of 20 percent (both values have been demonstrated by numerous groups), 44 lb (20 kg) of lipid could be produced per acre per day. If the growth facility were in operation 300 days/year, total biomass productivity would be 1,730 gal/acre per year. Thus, to produce enough fuel to replace 10 percent of the current U.S. gasoline consumption, about 8.1 million acres (12,700 square miles) of algal-growth facilities would be required. That is slightly less than 5 percent of the area of Texas. (In comparison, about 300 million acres of soybean or 32 million acres of corn would be required to produce the equivalent volume of vegetable oil or ethanol, respectively.) Coal-fired and gas-fired power plants theoretically could supply all the CO₂ necessary to produce the microorganisms. According to Feinberg and Karpuk (1990), 80 million tonnes of CO₂ are required per quad (10¹⁵ Btu) of algal lipid produced, so the production of 14 billion gallons of lipid (equivalent to 1.8 quads) would require 144 million tonnes of CO₂. That represents about half the CO₂ emitted by electricity-generating power plants in Texas (see http://carma.org). Technologies for developing algal strains with desirable traits for biofuel production that encompass classical strain improvement, metabolic engineering, and synthetic biology will probably enhance biofuel productivity in the future.

Few detailed economic analyses of costs of producing biofuels from algae have been completed. Results of the few analyses varied widely because of differences in such input variables as production-system design, cell- and product-recovery procedures, fuel type, and site characteristics. In one case, the reported cost of algal biofuel was well over $4.00/gal, indicating that much progress in R&D is needed to reduce production cost if this technology is to have utility in the foreseeable future (Pacheco, 2006). In another case, however, fuel-production costs with existing technologies were estimated to be $2.00/gal (Huntley and Redalje, 2006). Large-scale testing will be necessary to validate the assumptions used in those and similar economic analyses.

With the rapid growth of synthetic biology and the enhanced ability to engineer metabolic pathways into organisms to produce specific chemical or fuel products, synthetic biology and metabolic engineering for renewable fuel production have great potential and are receiving renewed interest (Savage, 2007). The approaches being taken include using well-established recombinant-DNA techniques to insert genes into microorganisms to make specific fuel precursors or even direct synthesis of hydrocarbon fuel components. Another approach involves redesigning genes with computer assistance to perform specific reactions and then synthesizing the desired genes for insertion into microorganisms. Yeasts can also be engineered to produce larger amounts of lipids, which with additional metabolic engineering can be converted to useful products, potentially fuels. That work has not progressed as far as the work on bacteria.

Those techniques might make it possible to modify bacteria to produce and excrete specific hydrocarbon molecules that have desired fuel or other chemical properties. Microorganisms that produce and excrete specific hydrocarbons minimize the costs of energy-consuming separation, although developing organisms that excrete the fuel products is a major challenge in that most synthesis products, including hydrocarbons, accumulate in the cell. No specific processes can be considered to be approaching commercial production at this point, but the magnitude of activity and the current rate of progress could change that in the not-too-distant future. Several companies are using synthetic biology to produce bacteria that make increased amounts of fatty acids or other lipids that are then converted to hydrocarbons and excreted. The bacteria make and excrete hydrocarbons of any desired length and structure. The phase-separation of hydrocarbons from the growth medium markedly reduces separation costs. The feedstock for the bacteria is renewable sugars, which can be obtained from sugar cane or grain or from cellulosic biomass (LS9, 2008). It is difficult to project future developments. Some companies are producing fuels, but projected costs of fuels have not been reported (Service, 2008).

Important advances are being made in genomics, molecular breeding, synthetic biology, and metabolic and bioprocess engineering that will probably enable discontinuous innovation and advancement in alternative transportation fuels. Those advances and related technologies have the potential to accelerate the creation of

dedicated or dual-purpose energy crops and microorganisms that can be used for both biofuel production and feedstock conversion.

The sequencing of full genomes continues to become faster and less expensive, and this is enabling the sequencing of energy crops, such as trees, perennial grasses, and such nonedible oilseeds as castor and jatropha. Their sequence data are extremely important for improving overall yields, for enabling improved nutrient and water use, and for understanding and manipulating biochemical pathways to enhance the production of desired materials. Sequence data can also be used to target specific genes for downregulation by classical methods, such as antisense and RNA interference, and via complete inactivation with new and evolving procedures for homologous recombination-based gene disruption. Rapid sequencing of breeding populations of energy crops will enable marker-assisted selection to accelerate breeding programs in ways previously not possible. Furthermore, rapid and inexpensive sequencing of fermentative and photosynthetic microorganisms is redefining and shortening the timelines associated with strain-development programs for converting sugars, lignocellulosic materials, and CO₂ to alternative liquid fuels. Strains generated through classical mutagenesis that have improved biocatalytic properties can now be analyzed at the molecular level to determine the specific genetic changes that result in the improved phenotype, and this allows the changes to be implemented in additional strains. In addition, “metagenome” sequence data obtained by randomly sequencing DNA isolated from environmental samples is providing huge numbers of new gene sequences that can be used in genetic engineering to improve crops and microorganisms.

Improved technologies for synthesizing megabase DNA molecules are being developed to allow the introduction of entire biochemical pathways into energy crops and biofuel-producing microorganisms. The technologies could have a great effect on scientists’ ability to generate plants and microorganisms with specific desirable traits. For example, it is becoming conceivable to replace large portions of, or even complete, chromosomes of microorganisms (including photosynthetic microorganisms) in ways that will focus the vast majority of their cells’ biochemical machinery toward production of next-generation biofuel molecules and thus provide cost and product advantages. Maintaining the purity of such cultures, and

finding ways to put at a disadvantage mutants that gain competitive ability by producing less of the desired secondary chemicals, could be serious hurdles.

In addition to genetic manipulation, new bioengineering technologies that will lower the cost of biofuel formation and recovery are coming on line. Synthetic biology can now provide synthetic DNA for transferring heterologous genes into suitable host cells, but metabolic engineering is the enabling technology for constructing functional and optimal pathways for microbial fuel synthesis. This field has matured in only a few years and has an impressive record of accomplishments, many already being applied in industry (for example, in the production of biopolymers, alcohols, 1,3-propanediol, oils, and hydrocarbons). Microbial strains that secrete hydrophobic fuels that are similar to constituents of diesel fuel and gasoline into the culture medium have been developed. The fuels can be separated from the aqueous phase in a manner that simplifies distillation and thereby reduces energy inputs and facilitates continuous production. By taking a systems view of metabolism, metabolic engineering developed tools for overall biosystems optimization that are now facilitating the optimal construction of biosynthetic pathways and elicitation of novel multigenic cellular properties of critical importance for biofuels production, such as tolerance of fuel toxicity. In bioprocessing, the successful development of membrane-based alcohol separation would greatly reduce energy costs relative to the typically used distillation process (Vane, 2008). Gas-stripping, liquid-liquid extractions of secreted fuel molecules or new adsorbent materials that will allow continuous production modes for fermentation-based products are also being developed (Vane, 2008). For photosynthetic production of biofuels, the development of low-cost photobioreactors and associated recovery systems for algal biofuel production is of great interest and could have substantial beneficial effects on overall process economics.

Grain-based ethanol is a bridge to advanced biofuels that has important potential for greenhouse gas displacement. Advanced biofuels do not directly compete with food and feed supply, and they minimize indirect land-use change if appropriate feedstocks are selected and sustainable practices are used in their production.

Grain ethanol has initiated public awareness of the use of ethanol in the current and future transportation fleet and of the pitfalls of feedstock supply for a new industry. Grain ethanol has helped to establish an industrial infrastructure for advanced biofuels and for distribution and use of fuel ethanol.

Lignocellulosic feedstocks for production of advanced biofuels could be agricultural or forestry residues, agricultural cover crops, dedicated perennial crops grown on marginal lands that are not suitable for commodity-crop production even with high commodity prices, or municipal solid wastes. Biochemical conversion of cellulose to liquid fuels emulates commercial corn grain-to-ethanol technology but might require additional processing steps and could result in other types of alcohols and hydrocarbon-rich fuels.

The technologies for biochemical conversion of cellulosic biomass to ethanol are in the early stages of demonstration and commercial development. Several demonstration plants are expected to be operational by 2012. The panel judges that cellulosic bioethanol will be commercially deployable before 2020, and other advanced biofuels are likely to emerge after 2020.

Finding 3.1

Engineering and operational knowledge can be gained only from designing and building commercial-scale, integrated cellulosic-ethanol facilities and then operating them for a reasonable period. The first few commercial plants will be more expensive than commercial facilities that follow because of the learning that occurs with a first-of-its-kind facility. The initial learning that occurs with first-of-a-kind plants will lead to further cost-reducing improvements in commercial facilities deployed thereafter. The pace of learning is expected to be similar to that in the chemical industry, in which costs have historically decreased by 30–40 percent over several cycles of deployment and concurrent process improvement.

Recommendation 3.1

The federal government and industry should aggressively pursue technology demonstration or small-scale commercial plants, which will lead to full-scale commercial production of cellulosic ethanol to define its potential and to provide data on engineering and cost performance to help in preparation for full commercial deployment.

In the immediate term, pretreatment and enzymatic hydrolysis, fermentation, or combined enzymatic hydrolysis and fermentation need to be substantially improved to allow efficient deconstruction of carbohydrate polymers to simple sugars and fermentation of the sugars to ethanol. Research in and improvement of pretreatment, with engineering of appropriate microorganisms for optimal use of the resulting simple sugars in an adverse fermentation environment, will have a direct impact on reducing the cost of transforming cellulosic feedstocks to ethanol. The cost of producing sugars directly affects the cost of ethanol. In addition, the sugars have to be converted to ethanol efficiently to minimize feedstock and operational costs.

Feedstock, pretreatment, and enzymes are key components of a cellulose-to-ethanol process, and they are all related to the goal of preparing lignocellulosic feedstocks (through agronomics, plant molecular genetics, and pretreatment) so that they are readily transformed to sugars and ethanol at low cost. Other targets for improvement include increasing solids loading and developing engineered microorganisms and enzymes that have increased tolerance of toxic compounds in biomass hydrolysates and of the biofuel products themselves. Incremental improvements in biochemical conversion technologies and the learning and experience gained from R&D and demonstration can be expected to reduce nonfeedstock processing costs by 25 percent by 2020 and 40 percent by 2035 (see Table 3.3).

Finding 3.2

Process improvements in cellulosic-ethanol technology are expected to be able to reduce the plant-related costs associated with ethanol production by up to 40 percent over the next 25 years. Over the next decade, process improvements and cost reductions are expected to come from evolutionary developments in technology, from learning gained through commercial experience and increases in scale of operation, and from research and engineering in advanced chemical and biochemical catalysts that will enable their deployment on a large scale.

Recommendation 3.2

The federal government should continue to support research and development to advance cellulosic-ethanol technologies. R&D programs should be pursued to resolve the major technical challenges facing ethanol production from cellulosic

biomass: pretreatment, enzymes, tolerance to toxic compounds and products, solids loading, engineering microorganisms, and novel separations for ethanol and other biofuels. A long-term perspective on the design of the programs and allocation of limited resources is needed; high priority should be placed on programs that address current problems at a fundamental level but with visible industrial goals.

Recommendation 3.3

The pilot and commercial-scale demonstrations of cellulosic-ethanol plants should be complemented by a closely coupled research and development program. R&D is necessary to resolve issues that are identified during demonstration and to reduce costs of sustainable feedstock acquisition. Industrial experience shows that such reductions typically occur as processes go through multiple phases of implementation and expansion.

Finding 3.3

Future improvements in cellulosic technology that entail invention of biocatalysts and biological processes could produce fuels that supplement ethanol production in the next 15 years. In addition to ethanol, advanced biofuels (such as lipids, higher alcohols, hydrocarbons, and other products that are easier to separate than ethanol) should be investigated because they could have higher energy content and would be less hygroscopic than ethanol and therefore could fit more smoothly into the current petroleum infrastructure than ethanol could.

Recommendation 3.4

The federal government should ensure that there is adequate research support to focus advances in bioengineering and the expanding biotechnologies on developing advanced biofuels. The research should focus on advanced biosciences—genomics, molecular biology, and genetics—and biotechnologies that could convert biomass directly to produce lipids, higher alcohols, and hydrocarbons fuels that can be directly integrated into the existing transportation infrastructure. The translation of those technologies into large-scale commercial practice poses many challenges that need to be resolved by R&D and demonstration if major effects on production of alternative liquid fuels from renewable resources are to be realized.

Finding 3.4

Biochemical conversion processes, as configured in cellulosic-ethanol plants, produce a stream of relatively pure CO₂ from the fermentor that can be dried, compressed, and made ready for geologic storage or used in enhanced oil recovery with little additional cost. Geologic storage of the CO₂ from biochemical conversion of plant matter (such as cellulosic biomass) further reduces greenhouse gas life-cycle emissions from advanced biofuels, so their greenhouse gas life-cycle emissions would become highly negative.

Recommendations 3.5

Because geologic storage of CO₂ from biochemical conversion of biomass to fuels could be important in reducing greenhouse gas emissions in the transportation sector, it should be evaluated and demonstrated in parallel with the program of geologic storage of CO₂ from coal-based fuels.

Aden, A., M. Ruth, K. Ibsen, J. Jechura, K. Neeves, J. Sheehan, B. Wallace, L. Montague, A. Slayton, and J. Lukas. 2002. Lignocellulosic Biomass to Ethanol Process. Design and Economics Utilizing Co-Current Dilute Acid Prehydrolysis and Enzymatic Hydrolysis for Corn Stover. Golden, Colo.: National Renewable Energy Laboratory.

Aristidou, A., and M. Penttila. 2000. Metabolic engineering applications to renewable resource utilization. Current Opinion in Biotechnology 11:187-198.

Briggs, Michael. 2004. Widescale biodiesel production from algae. Available at www.unh.edu/p2/biodiesel/article_alagae.html. Accessed October 1, 2008.

Chase, R. 2006. DuPont, BP join to make butanol: They say it outperforms ethanol as a fuel additive. USA Today, June 26. Available at http://www.usatoday.com/money/industries/energy/2006-06-20-butanol_x.htm. Accessed October 11, 2008.

Cornell, C.B. 2008. GM announces new cellulosic ethanol partnership with Mascoma Corp. Available at http://gas2.org/2008/05/01/gm-announcesnew-cellulosic-ethanol-partnership-with-mascoma-corp/. Accessed March 10, 2009.

Dayton, D. 2007. R&D needs for integrated biorefineries: The 30 × 30 Vision. In 4th Annual California Biomass Collaborative Forum, Sacramento, Calif.

Dismuskes, G.C., D. Carrieri, N. Bennette, G.M. Ananyev, and M.C. Posewitz. 2008. Aquatic phototrophs: Efficient alternatives to land-based crops for biofuels. Current Opinion in Biotechnology 19:235-240.

DOE (U.S. Department of Energy). 2007a. Biomass Feedstock Composition and Property Database. Available at http://www1.eere.energy.gov/biomass/feedstock_databases.html. Accessed December 13, 2007.

DOE. 2007b. DOE selects six cellulosic ethanol plants for up to $385 million in federal funding. Available at http://www.energy.gov/print/4827.htm. Accessed October 16, 2008.

DOE/EERE (U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy). 2007. Biomass Program 2007 Accomplishments Report. Available at http://www1.eere.energy.gov/biomass/pdfs/program_accomplishments_report.pdf. Accessed February 10, 2009.

DuPont. 2008. DuPont Fact Sheet on Biobutanol. Available at http://www.dupont.com/ag/news/releases/BP_DuPont_Fact_Sheet_Biobutanol.pdf. Accessed October 16, 2008.

Fargione, J., J. Hill, D. Tilman, S. Polasky, and P. Hawthorne. 2008. Land clearing and the biofuel carbon debt. Science 319:1235-1238.

Farrell, A., R. Plevin, B. Turner, A. Jones, M. O’Hare, and D. Kammen. 2006. Ethanol can contribute to energy and environmental goals. Science 311:506-509.

Feinberg, D., and M. Karpuk. 1990. CO₂ sources for microalgae-based liquid fuel production. Golden, Colo.: Solar Energy Research Institute and National Renewable Energy Laboratory.

GAO (U.S. Government Accountability Office). 2007. Biofuels: DOE Lacks a Strategic Approach to Coordinate Increasing Production with Infrastructure Development and Vehicle Needs. Washington, D.C.: GAO.

Gauss, W.F., S. Suzuki, and M. Takagi, inventors. 1976. Manufacture of Alcohol from Cellulosic Materials Using Plural Ferments. U.S. Patent 3990944.

Goldemberg J., S.T. Coelho, and P.M. Guardabassi. 2008. The sustainability of ethanol production from sugarcane. Energy Policy 36:2086-2097.

Gong, C.S., M.R. Ladisch, and G.T. Tsao. 1977. Cellobiase from Trichoderma viride: Purification properties, kinetics and mechanism. Journal of Biotechnology and Bioengineering 19:959-981.

Gong, C.S., M.R. Ladisch, and G.T. Tsao. 1979. Hydrolysis of cellulose: Mechanisms of enzymatic and chemical catalysis. In Advances in Chemistry Series 181, R. Brown and L. Jurasek, eds. Washington, D.C.: American Chemical Society.

Gulati, M., K. Kohlmann, M.R. Ladisch, R. Hespell, and R.J. Bothast. 1996. Assessment of ethanol production options for corn products. Bioresource Technology 58:253-264.

Hendriks, A.T.W.M., and G. Zeeman. 2009. Pretreatments to enhance the digestibility of lignocellulosic biomass. Bioresource Technology 100:10-18.

Hill, J., E. Nelson, D. Tilman, S. Polasky, and D. Tiffany. 2006. Environmental, economic, and energetic costs and benefits of biodiesel and ethanol biofuels. Proceedings of the National Academy of Sciences USA 103:11206-11210.

Huber, G.W., J.N. Chheda, C.J. Barrett, and J.A. Dumesic. 2005. Production of liquid alkanes by aqueous-phase processing of biomass-derived carbohydrates. Science 308:1446-1450.

Huber, G.W., S. Ibora, and A. Corma. 2006. Synthesis of transportation fuels from biomass: Chemistry, catalysts, and engineering. Chemical Review 106:4044-4098.

Huntley, M.E., and D.G. Redalje. 2006. CO₂ Mitigation and Renewable Oil from Photosynthetic Microbes: A New Appraisal. Mitigation and Adaptation Strategies for Global Change 12:573-608.

Intelligen, Inc. 2009. SuperPro Designer: Cost analysis and project economic evaluation. Available at http://www.intelligen.com/costanalysis.shtml. Accessed January 28, 2009.

Jensen, T.H. 2008. Race for key biofuel breakthrough intensifies. Reuters, June 2. Available at http://www.reuters.com/article/reutersEdge/idUSL0248503420080602. Accessed February 4, 2009.

Johnson, J.C. 2006. Technology Assessment of Biomass Ethanol: A Multi-objective, Life Cycle Approach Under Uncertainty. Ph.D. Thesis. Department of Chemical Engineering. Cambridge: Massachusetts Institute of Technology.

Jorgensen, H., J.B. Kristensen, and C. Felby. 2007. Enzymatic conversion of lignocellulose into fermentable sugars: Challenges and opportunities. Biofuels, Bioproducts and Biorefining 1:119-134.

Katzen, R., W.R. Ackley, Jr., G.D. Moon, J.R. Messick, B.F. Brush, and K.F. Kaupisch. 1981. Low-energy distillation systems. In Fuels from Biomass and Wastes, D.L. Klass and G.H. Emert, eds. Ann Arbor, Mich.: Ann Arbor Science Publishers.

Kim, Y., and M.R. Ladisch. 2008. Liquid hot water pretreatment of cellulosic biomass. In Methods in Molecular Biology, J. Meilanz, ed. Totowa, N.J.: Humana Press.

Kim, Y., N. Mosier, and M.R. Ladisch. 2008. Process simulation of modified dry grind ethanol plant with recycle of pretreated and enzymatically hydrolyzed distillers’ grains. Bioresource Technology 99:5177-5192.

Kwiatkowski, J., A.J. McAllon, F. Taylor, and D.B. Johnston. 2006. Modeling the process and costs of fuel ethanol production by the corn dry-grind process. Industrial Crops and Products 23:288-296.

Ladisch, M.R., J. Hong, M. Voloch, and G. Tsao. 1981. Cellulase kinetics: Trends in the biology of fermentations for fuel and chemicals. Basic Life Sciences 18:55-83.

Ladisch, M.R., K. Kohlmann, P. Westgate, J. Weil, and Y. Yang, inventors. 1998. Purdue Research Foundation Office of Technology Transfer, assignee. Processes for Treating Cellulosic Material. U.S. Patent 5,846,787.

Ladisch, M.R., M.C. Flickinger, and G.T. Tsao. 1978. Fuels and chemicals from biomass. Energy 4:263-275.

LS9. 2008. Renewable Petroleum Technology. Available at http://www.ls9.com/technology/. Accessed May 1, 2008.

Lynd, L.R., P.J. Weimer, W.H. van Zyl, and I.S. Pretorius. 2002. Microbial cellulose utilization: Fundamentals and biotechnology. Microbiology and Molecular Biology Review 66:506-577.

Lynd, L.R., W.H. van Zyl, J.E. McBride, and M. Laser. 2005. Consolidated bioprocessing of cellulosic biomass: An update. Current Opinion in Biotechnology 16:577-583.

Maxwell, E.L., A.G. Folger, and S.E. Hogg. 1985. Resource evaluation and site selection for microalgae production systems. Golden, Colo.: Solar Energy Research Institute.

Mosier, N., C. Wyman, B. Dale, R. Elander, Y.Y. Lee, M. Holtzapple, and M.R. Ladisch. 2005. Features of promising technologies for pretreatment of lignocellulosic biomass. Bioresource Technology 96:673-686.

Mosier, N.S., P. Hall, C.M. Ladisch, and M.R. Ladisch. 1999. Reaction kinetics, molecular action, and mechanisms of cellulolytic proteins. Advances in Biochemical Engineering and Biotechnology 65:24-40.

Murnen, H.K., V. Balan, S.P.S. Chundawat, B. Bals, L.D. Sousa, and B.E. Dale. 2007. Optimization of ammonia fiber expansion (AFEX) pretreatment and enzymatic hydrolysis of Miscanthus x Giganteus to fermentable sugars. Biotechnology Progress 23:846-850.

Nevoigt, E. 2008. Progress in metabolic engineering of Saccharomyces cerevisiae. Microbiology and Molecular Biology Reviews 72:379-412.

Pacheco, Michael. 2006. Potential of Biofuels to Meet Commercial and Military Needs. Presentation to the National Research Council Committee on Assessment of Resource Needs for Fuel Cell and Hydrogen Technologies.

Pate, R., M. Hightower, C. Cameron, and W. Einfield. 2007. Overview of Energy-Water Interdependencies and the Emerging Energy Demands on Water Resources. Los Alamos, N.Mex.: Sandia National Laboratories.

Pordesimo, L.O., B.R. Hames, S. Sokhansanj, and W.C. Edens. 2005. Variation in corn stover composition and energy content with crop maturity. Biomass and Bioenergy 28:366-374.

Saddler, J.N., H.H. Brownell, L.P. Clermont, and N. Levitin. 1982a. Enzymatic hydrolysis of cellulose and various pretreated wood fractions. Biotechnology and Bioengineering 24:1389-1402.

Saddler, J.N., C. Hogan, M.K.H. Chan, and G. Louisseize. 1982b. Ethanol fermentation of enzymatically hydrolyzed pretreated wood fractions using Trichoderma Cellulases, Zymomonas-Mobilis, and Saccharomyces-Cerevisiae. Canadian Journal of Microbiology 28:1311-1319.

Savage, N. 2007. Better biofuels. Technology Review 110:1.

Schwietzke, S., M.R. Ladisch, L. Russo, K. Kwant, T. Makinen, B. Kavalov, K. Maniatis, R. Zwart, G. Shahanan, K. Sipila, P. Grabowski, B. Telenius, M. White, and A. Brown. 2008a. Gaps in the research of 2nd generation transportation biofuels. IEA Bioenergy T41:2008:2001.

Schwietzke, S., Y. Kim, E. Ximenez, N. Mosier, and M. Ladisch. 2008b. Ethanol production from maize. In Molecular, Genetic Approaches to Maize Improvement, Biotechnology in Agriculture and Forestry, A.L. Kriz and B.A. Larkins, eds. Heidelberg, Germany: Springer-Verlag Berlin.

Searchinger, T., R. Heimlich, R.A. Houghton, F. Dong, A. Elobeid, J. Fabiosa, S. Tokgoz, D. Hayes, and T-H. Yu. 2008. Use of U.S. croplands for biofuels increases greenhouse gases through emissions from land use change. Science 319:1238-1240.

Sendich, E., M. Laser, S. Kim, H. Alizadeh, L. Laureano-Perez, B. Dale, and L. Lynd. 2008. Recent process improvements for the ammonia fiber expansion (AFEX) process and resulting reductions in minimum ethanol selling price. Bioresource Technology 99:8429-8435.

Service, R.F. 2008. Eyeing oil, synthetic biologists mine microbes for black gold. Science 322:522-523.

Shapouri, H., J.A. Duffield, and M. Wang. 2002. The Energy Balance of Corn Ethanol: An Update. Washington, D.C.: U.S. Department of Agriculture.

Sheehan, J., T. Dunahay, J. Benemann, and P. Roessler. 1998. A Look Back at the US Department of Energy’s Aquatic Species Program: Biodiesel from Algae. Golden, Colo.: U.S. Department of Energy.

Sonderegger, M., M. Jeppsson, C. Larsson, M.F. Gorwa-Grauslund, E. Boles, L. Olsson, I. Spencer-Martins, B. Hahn-Hagerdal, and U. Sauer. 2004. Fermentation performance of engineered and evolved xylose-fermenting Saccharomyces cerevisiae strains. Biotechnology and Bioengineering 87:90-98.

Spindler, D.D., C.E. Wyman, K. Grohmann, and A. Mohagheghi. 1987. Thermotolerant yeast for simultaneous saccharification and fermentation of cellulose to ethanol. Applied Biochemistry Symposium 17:279-293.

Takagi, M., S. Abe, S. Suzuki, G.H. Emert, and N. Yata. 1977. A method for production of alcohol directly from cellulose using cellulase and yeast. In Proceedings of the Bioconversion Symposium. New Delhi: Indian Institute of Technology.

USDA-NASS (U.S. Department of Agriculture, National Agricultural Statistics Service). 2005. Quick Stats, Iowa County Data—Crops. Available at http://www.nass.usda.gov. Accessed December 18, 2007.

Vane, L.M. 2008. Separation technologies for the recovery and dehydration of alcohols from fermentation broths. Biofuels, Bioproducts and Biorefining 2:553-588.

Worldwatch Institute. 2007. Biofuels for Transport: Global Potential and Implications for Sustainable Energy and Agriculture. London: Earthscan.

Wright, J.D., C.E. Wyman, K. Grohmann, and N. Yata. 1977. Simultaneous saccharification and fermentation of lignocellulose. Applied Biochemistry and Biotechnology 18:75-90.

Wyman, C.E., B.E. Dale, R.T. Elander, M. Holtzapple, M.R. Ladisch, and Y.Y. Lee. 2005a. Comparative sugar recovery data from laboratory scale application of lead pretreatment technologies to corn stover. Bioresource Technology 96:2026-2032.

Wyman, C.E., B.E. Dale, R.T. Elander, M. Holtzapple, M.R. Ladisch, and Y.Y. Lee. 2005b. Coordinated development of leading biomass pretreatment technologies. Bioresource Technology 96:1959-1966.

Wyman, C.E., D.D. Spindler, K. Grohmann, and S.M. Lastick. 1986. Simultaneous saccharification and fermentation with the yeast Brettanomyces clausenii. Biotechnology Bioengineering Symposium 17:221-238.

Yang, B., and C.E. Wyman. 2008. Pretreatment: The key to unlocking low-cost cellulosic ethanol. Biofuels, Bioproducts and Biorefining 2:26-40.

Zeng, M., N.S. Mosier, C-P. Huang, D.M. Sherman, and M.R. Ladisch. 2007. Microscopic examination of changes of plant cell structure in corn stover due to cellulase activity and hot water pretreatment. Biotechnology and Bioengineering 97:265-278.