Enzyme penetration into plant cell wall is widely acknowledged to be a key bar­rier to economical and effective biochemical conversion of lignocellulosic biomass [5, 49]. In fact, the primary function of pretreatment of lignocellulosic biomass is to assist subsequent enzymatic digestibility by making cell walls more accessi­ble to saccharifying enzymes [1, 4, 44]. However, an accurate description of the methods by which enzymes penetrate cell walls and accomplish cellulose degra­dation has been lacking. A recent study by Donohoe and coworkers provided, for the first time, direct visual evidence of loosening of plant cell wall structure due to dilute acid pretreatment and the subsequently improved access by cellulases [49]. Figure 3c-f further demonstrate the penetration of cellulases into pretreated cell walls as detected by nano-gold labeled antibodies to Cel7A and other cellulases. This study shows that penetration of enzymes into mildly pretreated cell walls is minimal and that cells stay largely intact even after prolonged exposure to cellu­lases (Fig. 3a, b). In moderately pretreated cell walls, cellulases are able to partially penetrate and disintegrate the inner secondary layers (S3) only (Fig. 3c, d); whereas the outer layers (S1 and S2) remain impervious to enzymes. In severely pretreated cell walls, enzymes penetrate throughout (Fig. 3e, f). These data suggest that enzy­matic digestibility of biomass is restricted by transport of enzymes into cell walls. While not directly evidenced by this study, these results also suggest that thermal pretreatments (and possibly others) “loosen” cell walls in layers providing enzymes access only to these structurally compromised zones of the cell walls. Kinetic data on thermal pretreatments by several research groups also suggests likely mass trans­fer limited xylan removal that can be modeled as parallel fast and slow reactions [44, 50, 51] and the fundamental observations made by Donohoe and coworkers [49] support this hypothesis.

Technically, a variety of different liquid fuels and chemicals can be made from high quality syngas (Fig. 4). The production of liquid fuel, either a thermochemical — catalyzed conversion or a microbial fermentation process (under development),

Olefins

Gasoline

Aldehydes

Alcohols

Fig. 4 A diversity of chemicals can be produced from syngas (from page 3 of Drs. Spath and Dayton’s 2003 NREL Technical Report, NREL/TP-510-34929, with modification)

Data source [21]. EPI: Energy Products of Idaho. GTI: Gas Technology Institute. SEI: Southern Electric International. BCL/FERCO: Battelle Columbus Laboratory/ Future Energy Resources Corporation. MTCI: Manufacturing and Technology Conversion International.

may be used to convert syngas into liquid fuels (methanol, ethanol, gasoline, and FT diesel). The catalytic conversion of syngas to ethanol can occur under high — temperature and high-pressure conditions (~250°C, 60-100 atm) with a molar ratio of H2 to CO at 2-3:1. However, most syngas (Table 3) does not contain such a high H2/CO ratio. Also, the catalysis reaction is not specific, resulting in a final mixture of methanol, ethanol, some other higher alcohols, and reactant gases. Considerable technical progress is required to generate ethanol from syngas at a viable commercial scale and various projects continue to explore possible options. For example, Range Fuels in Georgia (Table 1), is in the process of building a 20 million gallon pilot plant to evaluate using this approach for lignocellulose to ethanol conversion. Syngas can also be converted into gasoline or diesel through the so called MTG (methanol-to-gasoline) or the more common FT process. While these methods have been utilized for many years in the fossil fuel industry (coal or natural gas feedstocks), the utilization of lignocellulosic biomass is not yet viewed as being commercial [23]. Two DOE-funded companies (Table 1) are in the pro­cess of building demonstration scale plants to further explore the feasibility of the gasification-FT process for biofuel production.

In the microbial fermentation process, anaerobic bacteria such as Clostridium ljungdahlii are used to convert cleaned syngas into ethanol [24]. Reactions involved in the biological conversion process are as the follows:

CO + 3H2O ^ C2H5OH + 4CO2

6H2+2CO2 ^ C2H5OH + 3H2O

In general, conditions for microbial conversion of syngas to ethanol are mild and specific, and the H2:CO ratio is not critical. However, microbial toler­ance to ethanol concentration in the fermentation broth is currently a limitation. Several public and private R&D projects are underway to address the issue (e. g. http://www. coskata. com; http://www. ineosbio. com).

The large-scale production of a renewable and environmentally sustainable alter­native fuel faces several technical challenges that need to be addressed to make biodiesel feasible and economical. The two main concerns with any renewable fuel are raw materials and the technologies used for processing. Advances in genetic modification and other biotechnologies are resulting in new or modified feedstocks that have significantly increased the yields of alternative fuels, such as genetically modified Clostridium to improve alcohol production [16]. Technological advance­ments are also being made to convert the feedstocks into fuels by improving techniques or developing completely new and environmentally friendly approaches to biofuel production.

There are many feedstocks for biodiesel production such as virgin oils, biomass, algae, and waste oils, to name a few. Feedstocks also vary with climate and location and what might be a great source in one place may not be a good source in another. A considerable amount of research has been done using edible sources of virgin oils from vegetables, like soybean, rapeseed, sunflower seed, and canola oils, to produce biodiesel. However, oil with water or high free fatty acid content can result in the formation of soap as a by-product. Therefore, additional steps must be taken to prevent soap formation, which requires the utilization of more resources.

The production of biodiesel has increased demand for soybean oil from 1.56 bil­lion pounds in 2005-2006, to 2.8 billion pounds in 2006-2007 [17]. The increasing demand for virgin vegetable oil stocks has lead to an increase in price of these oils. The profitability of biodiesel relies heavily on the cost of its feedstock. The costs of soybean oil can account for up to 75% of the final cost per gallon of biodiesel. This has resulted in crops being sold as fuel crops, reducing the food supply and leading to an increase in food prices around the world.

To help with this issue, many oil-bearing non-edible plants have been investi­gated for the production of biodiesel. These are mainly tree species that can grow in harsh environments, such as Jatropha curcas, Pongamia pinnata, Castor, Mohva, Neem, Sal, etc. Jatropha curcas has the most significant potential due to its char­acteristics and growth requirements [16, 18]. It requires very little fertilizer and water (as little as 25 cm a year), is pest resistant, and can survive in poor soil conditions such as stony, gravelly, sandy or saline soils. Most important, it is fast growing, and can bloom and produce fruit throughout the year with a high seed yield. Optimized production has been found to yield an average of more than 99% of Jatropha biodiesel [19], which has comparable fuel properties to that of diesel from petroleum. It is expected that some varieties of Jatropha can produce as much as 1,600 gal of diesel fuel per acre-year compared to the wild variety that produces about 200 gal/acre-year [20]. Jatropha trees can capture four tons of carbon dioxide per acre and the fuel emits negligible greenhouse gases.

There is a growing interest in using algae as a feedstock for biodiesel production within the United States. Algae have become an appealing feedstock due to their aquatic environment providing them an abundant supply of water, CO2, and other nutrients. This results in a photosynthetic efficiency that is significantly higher than the average land based plants [21]. However, the power required to use artificial lighting to grow an aquatic species, such as microalgae, for the production of a biofuel would greatly reduce the overall efficiency of the process [22]. As the algae convert carbohydrates into triglycerides, the reproduction rate slows down so that the higher oil storing strains of algae reproduce at a much slower rate than lower oil storing strains [23]. This was shown by the Department of Energy’s (DOE) Aquatic Species Program, which found the overall yield to decrease as the algae’s oil storage increased.

Recently, Vasudevan and Briggs [21] summarized research on biodiesel pro­duction in a review article. According to them, a crude analysis of the quantum efficiency of photosynthesis can be done without getting into the details of the Calvin cycle; rather simply by looking at the photon energy required to carry out the overall reaction, and the energy of the products. In general, eight photons must be absorbed to split 1 CO2 and 2 H2O molecules, yielding one base carbohydrate (CH2O), one O2 molecule, and one H2O (which, interestingly, is not made of the same atoms as either of the two input H2O molecules.)

With the average energy of “Photosynthetically Available Radiation” (PAR) pho­tons being roughly 217 kJ, and a single carbohydrate (CH2O) having an energy content taken to be one-sixth that of glucose ((CH2O)6), or 467 kJ/mole, we can cal­culate a rough maximum efficiency of 26.9% for converting captured solar energy into stored chemical energy. With PAR accounting for 43% of incident sunlight on earth’s surface [24], the quantum limit (based on eight photons captured per CH2O produced) on photosynthetic efficiency works out to roughly 11.6%. In real­ity, most plants fall well below this theoretical limit, with global averages estimated typically at between 1 and 2%. The reasons for such a difference generally revolve around rate limitations due to factors other than light (H2O and nutrient availabil­ity, for example), photosaturation (some plants, or portions of plants receive more sunlight than they can process while others receive less than they could process), and photorespiration due to Rubisco (the protein that serves ultimately as a catalyst for photosynthesis) also accepting atmospheric O2 (rather than CO2), resulting in photorespiration.

In the US, the average daily incident solar energy (across the entire spectrum) reaching the earth’s surface ranges from 12,000 to 22,000 kJ/m2 (varying primarily with latitude). If the maximum photosynthetic efficiency is 11.6%, then the max­imum conversion to chemical energy is around 1,400-2,550 kJ/m2/day, or 3.8 x 1012 J/acre-year in the sunniest parts of the country. Assuming the heating value of biodiesel to be 0.137 GJ/gal, the maximum possible biodiesel production in the sun­niest part of the US works out to be approximately 28,000 gal/acre-year, assuming 100% conversion of algae biomass to biodiesel, which is infeasible.

It is important to keep in mind that this is strictly a theoretical “upper limit” based on the quantum limits to photosynthetic efficiency, and does not account for factors that decrease efficiency and conversion. Based on this simple analysis though, it is clear that claims of algal biodiesel production yields in excess of 40,000 gal/acre — year or higher should be viewed with considerable skepticism. While such yields may be possible with artificial lighting, this approach would be very ill-advised, as at best only about 1% of the energy of the energy used to power the lights would ultimately be turned into a liquid fuel (clearly, one needs to look at the overall efficiency).

This upper limit also allows us to assess how truly inefficient many crops are when viewed strictly as biofuel producers. With soybeans yielding on average 60 gal of oil (and hence biodiesel) per acre-year, the actual fuel production is stag­geringly small in comparison to the amount of solar energy available. This should further make it clear that using typical biofuels for the purpose of electricity gener­ation (as opposed to the transportation sector) is an inefficient means of harnessing solar energy. Considering that photovoltaic panels currently on the market achieve net efficiencies (for solar energy to electrical energy) on the order of 15-20%, with multi-layer photovoltaics and solar thermal-electric systems achieving efficiencies of twice that in trial runs, biomass to electricity production falls far behind (con­sidering typical plant photosynthetic efficiencies of 1-2%), with conversion of that biomass energy to electrical energy dropping the net efficiency to well under 1%.

Currently, the research for algae growth for fuel production is being done using photobioreactors. Unfortunately, current designs demand a high capital cost, which makes large-scale production uneconomical until a low cost design or new method of production is discovered. Storing energy as oil rather than as carbohydrates slows the reproduction rate of any algae, so higher oil strains generally grow slower than low oil strains. The result is that an open system (such as open raceway ponds) is readily taken over by lower oil strains, despite efforts to maintain a culture of higher oil algae. Attempts to grow higher oil extremophiles, which can survive in extreme conditions (such as high salinity or alkalinity) that most other strains cannot tolerate, have yielded poor results, in terms of the net productivity of the system. While an extremophile may be able to survive in an extreme condition, that doesn’t mean it can thrive in such conditions.

Many research groups have therefore turned to using enclosed photobioreac­tors of various designs as a means of preventing culture collapse or takeover by low oil strains, as well as decreasing the vulnerability to temperature fluctuations. The significant downside is the much higher capital cost of current photobioreactor designs. While such high costs are not prohibitive when growing algae for pro­ducing high value products (specialty food supplements, colorants, pharmaceutical products, etc.), it is a significant challenge when attempting to produce a low value product such as fuel. Therefore, substantial focus must be placed on designing much lower cost photobioreactors and tying algae oil production to other products (animal feed or fertilizer from the protein) and services (growing the algae on waste stream effluent to remove eutrophying nutrients, or growing nitrogen fixing algae on power plant emissions to remove NOx emissions).

An additional challenge, when trying to maximize oil production with algae, is the unfortunate fact that higher oil concentrations are achieved only when the algae are stressed — in particular due to nutrient restrictions. Those nutrient restrictions also limit growth (thus limiting net photosynthetic efficiency, where maximizing that is a prime reason for using algae as a fuel feedstock). How to balance the desire for high growth and high oil production to the total amount of oil produced is no small task. One of the goals of DOE’s well-known Aquatic Species Program was to maximize oil production through nutrient restriction; however their study showed that while the oil concentration went up, there was a proportionally greater drop in reproduction rate, resulting in a lower overall oil yield.

One approach to balancing these issues has been successfully tested on a small commercial scale (2 ha) by Huntley and Redalje [25], using a combination of photobioreactors and open ponds. The general approach involves using large photo­bioreactors for a “growth stage”, in which an algal strain capable of high oil content (when nutrient restricted) is grown in an environment that promotes cell division (plentiful nutrients, etc.) — but which is enclosed to keep out other strains. After the growth stage, the algae enter an open raceway pond with nutrient limitations and other stressors, aimed at promoting biosynthesis of oil. The nutrient limitations discourage other strains from moving in and taking over (since they also require nutrients for cell division).

Waste oils, such as restaurant grease and spent frialator oil, can also be used in the production of biodiesel. This eliminates the “food or fuel” debate that affects virgin edible oil sources. These waste oils normally cost money for restaurants and other establishments to dispose off. This can have a negative feedstock cost which reduces the overall cost of production. However, like virgin oils, traditional pro­cesses of converting waste oils to biodiesel can result in soap formation due to the presence of water and free fatty acids. The waste oils usually contain particulates that require filtration or separation prior to processing. Demand for waste oil as a biodiesel feedstock has already resulted in companies now paying restaurants for their waste vegetable oil (WVO). Quantities of WVO are limited (it is estimated to be about 1.1 billion gallons per year in the US), but it is certainly a good option for producing biodiesel.

2.3 Comparison of Technologies

A conventional base-catalyzed reaction is used in the majority of transesterifica­tion processes to produce biodiesel. Sodium hydroxide is used as the catalyst when methanol is the acyl acceptor, and potassium hydroxide is used when ethanol is the acyl acceptor, due to solubility considerations [15]. The ethyl esters have a slightly higher energy value than the methyl esters due to the presence of the additional carbon atom, and ethanol can be more easily produced from renewable sources, such as corn. Typical reactions take place with a high molar ratio of alcohol to oil of about 6:1 with methanol, and 12:1 for ethanol [15]. The excess alcohol allows for complete conversion of the triglycerides to the fatty acid esters. An advantage of base-catalyzed transesterification is the relatively short reaction time to achieve conversion levels of 98% or greater, compared to other processes. The reaction is a direct process, needing no intermediate steps, and operates at a relatively low tem­perature and pressure of about 66°C and 1.4 atm, respectively. However, a major disadvantage of the base catalyzed process is the formation of soap when water or free fatty acids are present in the feedstock. Thus the feedstock should be anhy­drous but the process still requires a large amount of base to be added to neutralize the fatty acids [15]. Soap formation results in additional downstream separation problems combined with a reduction in the fatty acid ester yield. The process also requires two steps and uses large amounts of chemicals as catalysts.

Acid-catalyzed transesterification is a viable alternative, in which sulfuric acid is typically used. One advantage over the base-catalyzed method [26] is that it is not as susceptible to soap formation. The resulting downstream product is easily separated and produces a relatively high quality glycerol byproduct. The process also requires only one step, compared to two steps in the base-catalyzed process. However, acid-catalysis reactions are slower and result in lower yields than base- catalysis, ranging from 56.8 to 96.4% depending on the feedstock [27]. A major disadvantage to either base or acid transesterification process is the disposal of the glycerol byproduct. Glycerin is already inexpensive, easily available, and is used in a wide array of pharmaceutical formulations. The major issue is with the purity of the glycerin; the byproduct glycerin from the production of biodiesel is 80-88% while industrial grade is 98% or higher [15]. The low market value of glycerin does not make purification economical. Many researchers are investigating innovative chemical and biological processes for the conversion of glycerin into value-added products including antifreeze agents, hydrogen, and ethanol [28].

A relatively new and promising development in the production of biodiesel is via enzymatic transesterification with lipase as the catalyst. Several microbial strains of lipases have been found to have transesterification activity; Pseudomonas cepa­cia [29], Thermomyces lanuginosus [30], and Candida antarctica [31] are a few that have been reported. The products of an enzyme-catalyzed reaction can easily be collected and separated. Unlike alkali-based reactions, enzymes can be recycled since they are not used up and require much less alcohol to perform the reaction. However, enzyme reactions take much longer to complete and can have lower yields due to inhibition of the enzyme caused by glycerol formation. Methanol, the acyl acceptor, can also strip the essential water from the active site of the enzyme, result­ing in deactivation of the enzyme. Enzymes are also expensive and require treatment such as immobilization, purification, pre-treatment, and modification [32].

New technologies are being developed to produce biodiesel that do not form glycerol as a byproduct. The hydrocracking process uses hydrocracking, hydrotreat­ing, and hydrogenation reactions to convert a wide range of feedstocks to biodiesel with yields of 75-80% [15]. This process is currently being utilized in petroleum refineries and uses a conventional commercial refinery hydrotreating catalyst. However, the hydrocracking process requires hydrogen, which is primarily obtained from natural gas. To reduce the costs of hydrogen, the process could be easily integrated with a refinery.

The production of biodiesel has significantly increased over the past few years. The National Biodiesel Board reports an increase in production from 250 million gallons in 2006 to 450 million gallons in 2007, an increase of 55.6%. European countries produced 5.7 million tons of biodiesel in 2007 (~1.5 billion gallons), which is an increase of 16.8% from 2006 according to the European Biodiesel Board. Germany is the World leader in biodiesel production and produced 2.9 million tons (~790 million gallons) in 2007, which is over 50% of the European biodiesel market.

Lignin is a polymeric material composed of phenylpropanoid units derived primar­ily from three cinnamyl alcohols (monolignols): p-coumaryl, coniferyl, and sinapyl alcohols. Polymer formation is thought to occur via oxidative (radical-mediated) coupling between monolignols and the growing oligomer/polymer [52, 53] and is commonly believed to occur in a near-random fashion [54], although some recent studies suggest an ordered and protein-regulated lignin synthesis [55]. In any case, the resulting polymer is complex, heterogeneous, and recalcitrant to biological degradation. Although lignin loss is minimal during thermal-acidic/neutral pre­treatments, it can undergo structural and chemical changes [56] that significantly influence downstream enzymatic conversion.

Although enzymes thoroughly penetrate cell walls after high severity pre­treatments [49], incomplete cellulose conversion by cellulases suggests additional barriers exist at the ultrastructural level. One potential barrier is occlusion of the cel­lulose microfibrils by residual lignin or hemicellulose that would sterically prevent

cellulases from binding to cellulose [42]. Other indirect mechanisms that impede complete cellulose hydrolysis are also possible such as non-productive binding of cellulases to lignin [34-36], however reports that contradict this theory also exist [57].

Enzymatic hydrolysis of biomass pretreated under alkaline conditions, which hydrolyzes less xylan than acidic pretreatments, supports the steric hindrance concept. Elevated cellulolytic activity is observed on alkaline pretreated biomass when cellulases are supplemented with xylanases and other hemicellulose degrading enzymes, likely a function of removing additional barriers to cellulose accessibility [58, 59]. A study in pretreatment variability by Selig and co-workers suggested that cellulose digestibility is improved directly by xylan removal, but only indirectly by lignin removal [47]. Removal of lignin by pretreatment appeared to increase enzymatic removal of xylan, which in turn increased cellulose digestibility. Lignin removal alone had little impact on cellulose digestion. Lignin modifying enzymes, however, have been shown to synergistically work with cellulases during digestion of steam-pretreated biomass, improving sugar yields through at least partial removal of the lignin barrier [60]. In spite of a general consensus in the scientific community about the significance of the lignin barrier to cellulose digestibility, only limited attention has been given to the fate of lignin during widely used high tempera­ture dilute acid, hot water, and steam pretreatments which only partially remove lignin [1,8].

A recent study investigated the fate of lignin during high temperature acid and neutral pretreatments using electron microscopy and spectroscopy techniques [40]. This study revealed that lignin could be mobilized within the cell wall matrix at temperatures as low at 120°C during both neutral and low pH pretreatments, and appears to be, at least in part, dependent on pretreatment severity. On a relatively macro scale, part of the mobilized lignin deposits back on to biomass surfaces as spherical bodies, suggesting that lignin undergoes the following sequence of events during these pretreatments — phase-transition or melting, mobilization into bulk solution, coalescence, and deposition onto solid surfaces. Scanning- and transmis­sion electron microscopy (SEM and TEM) of pretreated cell walls shows that the lignin droplets (stained with KMnO4) take a wide range of sizes (<50 nm to 2 ^m) and shapes (Fig. 4a, b and Fig. 5), though the “free” shapes are uniformly spheri­cal. Other shapes observed appear to be dictated by the physical constraints of the structures surrounding them. In addition to redeposition, there also appears to be a reorganization of lignin structure within the cell walls. A fraction of the lignin remains within the walls during pretreatment. This fraction apparently melts, but is unable to escape into the bulk liquid phase before coalescing back into droplets, as evidenced by the KMnO4 stained lignin droplets that appear between layers in the cell wall (Fig. 4b-d).

Aside from the obvious implications of lignin mobility, coalescence, and rede­position observed during high temperature pretreatments, chemical modification of the lignin should also be considered. These may range from covalent bond break­age and formation to changes in inter — and intramolecular interactions. Although FTIR and NMR studies did not distinctly show chemical changes in the mobilized

Fig. 1.4 TEM micrograph of lignin droplets re-deposited on cellulose surfaces after being transported from the cell wall matrix during high temperature pretreatments (a). Electron tomograph images of coalesced lignin within cell walls. The boxed region in b has been segmented to show the 3D volume of coalesced lignin (c). Large lignin globules can form in openings like pits (arrow b, d). Scale bars = 200 nm a; 500 nm b, c; 200 nm d

lignin in this study, it is possible that chemical alteration could be part of the lignin removal and transport process because lignin can partially dissolve and react in acid solutions under appropriate conditions [56]. It is further possible that part of this mobilized lignin could contain lignin-carbohydrate complexes that might sequester cellulases as observed in some studies [34, 36].

Another recent study [42] showed that purified lignin preparations as well as native lignin from corn stover could be redeposited onto clean cellulose surfaces such as filter paper. More severe pretreatments (higher temperature or acid concen­trations) resulted in finer redeposited droplets. Under these conditions, digestibility of filter paper was lower by up to 15% in comparison with treatments that did not contain lignin. Since these digestions were performed at very high enzyme load­ings to circumvent issues related to non-productive binding to lignin, it appears that physical blockage of the cellulose surface by lignin resulted in lower digestibility. Although redeposited lignin inhibited digestion of pure cellulose substrates in the study by Selig and coworkers [42], it is also probable that the mass transport of lignin could enhance enzymatic cellulose degradation in biomass. For example, we could visualize that as a result of lignin mass transport, the lignin sheath coating cellulose surfaces gets concentrated into droplets rendering a greater cellulose sur­face area available for enzymatic attack. Removal of lignin could also improve cell wall porosity allowing enzymes better access for penetration. Much work needs to be done to completely understand the nature and implications of lignin transport.

1.3 Overview

Theoretically, the basic process for biochemical conversion of lignocellulosic biomass into ethanol or other biofuels is relatively straightforward. First, the lig — nocellulosic matrix must be treated to gain access to and/or separate the main components: lignin, cellulose, hemicellulose, and pectin. The polysaccharides (cel­lulose and hemicelluloses) are then hydrolyzed to sugars, which are fermented to ethanol. This hydrolytic conversion process for lignocellulosic biomass contributes to the technical barriers that currently limit commercial operations. The fermen­tation process for ethanol production from lignocellulosic biomass is also more complex than for corn-based ethanol production. Hydrolysates of lignocellulosic biomass typically contain significant amounts of pentoses (e. g. xylose and arabi — nose). These C5 sugars are not readily fermented to ethanol by the commonly-used yeast (Saccharomyces cerevisiae). Efficiently converting both glucose and pentoses (xylose and arabinose) into ethanol or other biofuels and at reasonably high con­centrations (8-12%) is another challenge for the fermentation microorganisms.

3.1 Background

Over the past few years, butanol made from biomass, popularly known as biobu­tanol, has gained a lot of attention as a biofuel. Butanol is an alcohol-based fuel that contains four carbons and has chemical properties similar to that of gasoline, thus making it an attractive substitute or additive. Biobutanol can be produced from the fermentation of sugars from biomass or by the gasification of cellulosic biomass. It can be blended in any ratio with gasoline and be used in existing automobiles without any need for engine or fuel line modifications. It is an attractive substitute to gasoline because its BTU content is 110,000 BTU’s per gallon, which is very close to the 115,000 BTU per gallon of gasoline, resulting in little change to fuel economy. The Reid vapor pressure (RVP) of butanol (0.33 psi) is low compared to ethanol (2 psi) or gasoline (4.5 psi), resulting in lower evaporative emissions. The octane values and energy density of butanol are also closer to gasoline than is ethanol. Ethanol is 100% soluble in water whereas the solubility of butanol is 9.1% at 25°C [10]; this results in less water absorbed and rust dissolved into the fuel from tanks and pipelines. An added benefit to the low solubility is reducing the spread into groundwater in case of a spill.

However, biobutanol is not a perfect fuel and has several disadvantages. Butanol is more toxic to humans and animals than lower carbon alcohols. The LD50 oral consumption for a rat for butanol is 790 mg/kg compared to 7,060 mg/kg for ethanol [13]. However, it is well known that gasoline contains chemicals such as benzene, which is toxic and carcinogenic. There have been no definitive tests as to whether butanol will degrade the materials in an automobile over time, but current evidence suggests that this is unlikely [10]. Environmental Energy, Inc. tested a 1992 Buick Park Avenue by driving it 10,000 miles on 100% butanol [33]. No modifications were done to the car and it passed all emission tests performed in 10 states with an average increase in gas mileage of 9%. Compared to gasoline, combustion of butanol reduces the amount of hydrocarbons, carbon monoxide, and smog-creating compounds that are emitted [33].

Butanol is used as an industrial solvent and the market demand is about 350 million gallons a year worldwide, with the United States accounting for 63%. The production of butanol via fermentation is the second oldest fermentation process, next only to ethanol. Since the 1950s however, production of butanol via fermen­tation has not been an economically viable alternative due to the historic low cost of petroleum. A new push for renewable alternative fuel sources has been fueled by the increasing cost of petroleum combined with the generation of more green­house gases. These two reasons and the development of new technologies form the underpinnings of the reemergence of the butanol fermentation process.

Uniform distribution of heat, chemical catalysts, and enzymes as well as absence of product gradients within conversion reactors are all dependent on the mixing properties of biomass slurries being processed, which in turn are determined by rheological characteristics. Biomass rheology poses several challenges because of the fibrous nature of the particles, their ability to absorb water and become unsatu­rated at relatively low solid concentrations of 25-35% (w/w), and the continually changing particle chemical/physical properties during flow through the process. Free water content appears to be the largest factor contributing to slurry rheology. This is especially true at the high solid concentrations that are desired to make the overall process economical by lowering equipment volume and thereby cost [27]. At solid concentrations beyond the point of unsaturation, the slurries become wet granular material that agglomerate and can compact under their own weight if not adequately mixed. At lower concentrations, adequate mixing is still required to pre­vent settling. To further complicate matters, as biomass gets broken down into its constitutive sugars, changes occur in particle size as well as chemical properties. Water retaining polymers, such as hemicellulose and pectin, are broken down and the previously hygroscopic biomass has lower capacity for water absorption result­ing in an increased amount of free water, and thereby altered slurry rheology. These dynamic changes in solid properties necessitate studies to understand rheological behavior of slurries through various process treatments.

In simplest terms, biomass slurries can be described as non-Newtonian pseudo­plastic (shear-thinning) fluids [27, 61, 62]. Whereas the exact mechanism leading to pseudoplasticity in biomass slurries is unknown, a possible explanation of the behavior can be ascribed to formation of three dimensional network structure of the fibrous particles and subsequent breakdown of this structure under shear [63]. Previous studies show that while free water is present, apparent viscosity values under continuous shear increase with increasing solid concentrations. These mea­sured apparent viscosities can be modeled with simple Casson, Bingham or Power Law models [27, 61, 62]. Thick slurries with little or no free water do not exhibit a further increase in apparent viscosity with increasing solid concentrations under continuous shear [27]. Other viscoelastic properties, such as storage and loss mod — ulii could continue to change; however, these measurements have not yet been reported for biomass slurries.

The relatively sparse data and lack of fundamental understanding of rheologi­cal properties of biomass slurries makes calculations on mixing requirements for biomass conversion processes uncertain. Also, transport properties within biomass slurries, such as convective/conductive heat transport and convective/diffusive mass transport, and their effects on conversion are hard to discern or estimate. For exam­ple, Fig. 6 shows enzyme digestibility data obtained during digestion of pretreated corn stover at high solids pretreatments (>15% solids). Each data point was gen­erated as a single measurement from triplicate reactors after 5 days of digestion. As can be seen from Fig. 6a, conversion of cellulose to glucose decreases steadily as solids concentrations increase suggesting inhibition of enzymes, possibly due to poor mass transfer resulting in localized accumulation of sugars as suggested by Hodge and coworkers [22]. Clearly, slurry properties will play a major role in determining these transport parameters that are crucial to determine optimal process performance across multiple scales. As another example, Fig. 7 shows experimental data from tests performed to evaluate heating time in a closed reactor containing biomass slurries of varying concentrations. These data show significant retarda­tion of heat transfer, even with the moderate density slurries containing 10% solids (w/w). Simple heat transfer simulation models have been developed for biomass slurries assuming conductive heat transfer and a one-dimensional system; however, their validity has not been verified with experimental data [64, 65]. In unsatu­rated biomass slurries containing discrete aggregates, the accurate determination and prediction of transport properties might be a challenging exercise.

Many pretreatment processes have been tested for the capability to facilitate lig­nocellulosic biomass component separation and to aid in subsequent access for the hydrolytic enzymes [25, 26]. The more extensively studied methods are listed in Table 4, which includes AFEX (ammonia fiber explosion) and ARP (ammonia recycle percolation) [27, 28], lime [29], organosolv [30], liquid hot water, ionic liquid [31], dilute acid and steam explosion [32, 33], and enzyme treatment [34]. Additional information on pretreatments is available from Taherzadeh and Karimi [35] and Jorgensen, Kristensen, and Felby [27].

An effective practical pretreatment process should meet the following standards for use in future commercial facilities: (a) allow excellent cellulose digestibility by commercial cellulases, (b) good recoveries of cellulose and pentoses from hemicel — luloses, (c) minimal or no microbial inhibitory by-products, (d) good separation of lignin, (e) be easily managed at large volumes, (f) be relatively inexpensive (capex and opex), (g) not require large energy inputs, and (h) have environmentally acceptable features.

Published economic analysis has suggested that the MESP (minimal ethanol selling price) for cellulosic ethanol from corn stover, using different pretreatment technologies, ranges from $1.41/gallon for the AFEX process to $1.7/gallon for hot water treated corn stover [36]. More recently, Sendich et al. [37] indicated that the MESP for AFEX treated corn stover could be as low as $0.81/gallon due to reduced ammonia concentration and a simplified ammonia recycle process. However, we believe the assumptions used are perhaps overly-optimistic. For example, a feed­stock cost of $30/ton is very low, especially given the alternative nutrient and soil texture improvement values for corn stover. More recently, the DOE reported a 2007 cellulosic MESP of $2.43/gallon [38]. In any case, and despite many years of R&D, it is difficult to validate the assumptions since none of the conversion processes have been evaluated at practical scale.

The oldest method of butanol production is the acetone-butanol-ethanol (ABE) bacterial fermentation by Clostridium acetobutylicum, which dates back to Louis Pasteur in 1861 [13]. The bacterial microorganism, C. acetobutylicum, was first isolated by Weizmann [13]. In the ABE fermentation process, C. acetobutylicum produces acetic, butyric, and propionic acids from glucose that can be generated from various biomass sources. Potential feedstocks include corn, molasses, whey permeates, or glucose. An enzyme catalyzed reaction of acetoacetyl-CoA transfers

CoA to acetate forming acetyl-CoA. Through a series of metabolic reactions, butyryl-CoA is produced from acetyl-CoA, which is then converted to butanol in the solventogenic pathway [33]. Acetyl-CoA can also produce ethanol and acetone from acetoacetyl-CoA. A typical process produces acetone, butanol and ethanol in the ratio 3: 6:1.

The butanol yield from the ABE fermentation of glucose is relatively low, about 15-25 wt% typically [33]. This is due to the buildup of acetic, butyric, and propionic acids along with the products acetone, butanol, and ethanol, during the fermenta­tion process. The solvents are toxic to C. acetobutylicum. The butanol destabilizes the cell membrane of the microorganisms ultimately resulting in cell death. Higher yields can be achieved by continuously removing the harmful solvents, mainly butanol, and/or by genetically modifying strains of microorganisms that can tolerate higher concentrations of butanol [33].

A butanol-tolerant mutant strain of C. acetobutylicum has been developed and designated as SA-1 [34]. This strain shows a 121% improvement in butanol tol­erance over the typical strain used in ABE fermentation. The enhancement of the strain results in an overall increase in butanol production of 13.2%. Additional advantages of the mutated strain are an increase in growth rate, more pH resistance, more effective utilization of carbohydrates, and reduction in acetone concentration by 12.5-40% [34]. Other studies using genetic and metabolic engineering have modified strains, which have resulted in an increase of about 320% in the final butanol concentration [35]. The antisense RNA process helps down-regulate genes for butyrate formation by acidogenesis and increases the butanol yield through solventogenesis. The process has resulted in strains with butanol yields of 35% [36].

Tetravitae Bioscience has combined a patented mutant strain of C. beijerinckii and a continuous, integrated fermentation process that utilizes gas stripping. C. bei­jerinckii is a species of rod-shaped anaerobic bacteria that is known for the synthesis of organic solvents, and uses a broader substrate range and better pH range than C. acetobutylicum. The solvent genes of C. beijerinckii are located on the chromo­some, which is more genetically stable than on the plasmid for C. acetobutylicum. The gas stripping process prevents the butanol concentrations from reaching toxic levels by sparging oxygen-free nitrogen or fermentation gases through the fermen­tation solution and the ABE captured in the gas are condensed [13]. The exhaust gas is then recycled back to the reactor to collect more ABE for removal. Advantages of this method are the low energy requirements, the fact that it does not remove important acid intermediates, and that it allows for efficient recovery of butanol [37].

Environmental Energy Inc. (EEI) and Ohio State University (OSU) have devel­oped a two-step anaerobic fermentation process in a joint project to produce butanol from biomass. The first process converts the feedstock carbohydrates into butyric acid through acidogenesis using C. tyrobutyricum. The second step converts the butyric acid, using C. acetobutylicum, into butanol, which results in a significant improvement from conventional processes. The butanol solution requires purifi­cation from a recovery unit after the second step reactor. EEI’s process uses a purification process that takes advantage of the azeotrope formed by butanol (55%) and water (45%), which is used to minimize the energy required for dis­tillation. These processes utilize OSU’s proprietary fibrous-bed bioreactor (FBB) that has demonstrated improvements in long-term production with a scalable pack­ing design. The packing consisted of a spiral-wound, fibrous matrix that allows for a high surface area with large enough voids to allow for a high cell density. Immobilizing the cells in the FBB minimizes the energy consumption required by the cells [33].

British Petroleum (BP) has partnered with DuPont to commercialize biobu­tanol using advanced metabolic pathways for 1-butanol. They have announced plans to produce 30,000 tons per year of biobutanol at the British Sugar facility in Wissington, UK. This will help meet the United Kingdom’s Renewable Fuels Obligation set for 2010. Along with 1-butanol, they plan on developing biocatalysts to produce higher octane isomers such as 2-butanol and iso-butanol, and to increase the interest and utility as a fuels additive or substitute [38]. BP and Dupont plan on initially marketing biobutanol to the current market as an industrial solvent and then implement a larger commercialization into fuel blending by 2010 [38].

A different approach to producing butanol utilizes a thermochemical route for the gasification of biomass by a syngas catalyst. W2 Energy Inc. is working to produce biobutanol from a Gliding Arc Tornado plasma reactor (GAT) for biomass gasifica­tion. The GAT is a non-thermal plasma system, which utilizes reverse vortex flow that allows for a larger gas residence time and ensures a more uniform gas treat­ment. An advantage to the GAT system is that because of the thermal insulation, it does not require high-temperature material, thus reducing costs [39]. The gasifica­tion of biomass is accomplished by the solid biomass undergoing a thermochemical reaction under sub-stoichiometric conditions with an oxidizing fuel. The biomass’s energy is released in the form of CO, CH4, H2, and other combustible gases (syn­gas) [40]. The syngas consists of basic elementary components, which can be made into butanol using various petrochemical techniques. Other advances in gasifica­tion technology have been made by the National Renewable Energy Laboratory’s (NREL) Battelle Labs.

3.2 Summary

Biobutanol is a renewable, biodegradable, alternative fuel, which can be used neat or blended with gasoline. Properties such as energy density, octane value, and Reid vapor pressure (RVP) are similar to gasoline; hence current vehicles can use biobu­tanol without any engine modifications. Biobutanol can be produced from biomass by the fermentation of sugars and starches or by thermochemical routes using gasi­fication. The emergence of butanol as a fuel is growing with companies such as BP, DuPont, EEI, Tetravitae Bioscience, and W2 Energy Inc. investing in new technol­ogy as well as in manufacturing. Worldwide commercialization of biobutanol can replace or enhance blends of gasoline to reduce the dependence on petroleum as well as reduce greenhouse gas emissions.

Significantly greater research and development effort in the conversion of ligno — cellulosic biomass, spurred by economic, national security and climate change concerns over the past few years have led to significant strides in development of a fundamental understanding of transport processes that could appreciably

improve overall performance and make renewable liquid transportation fuels sus­tainable and affordable. A thorough understanding of fundamental issues related to transport processes and the development of predictive models that integrate heat, mass and momentum transport are essential to the design, development and imple­mentation of scale-independent processes. Continued synergism between science and engineering disciplines along with participation by industry is crucial to the development of cost-effective alternative motor fuels by 2012 and the significant displacement of fossil-derived fuels specified by the DOE (Energy Independence and Security Act of 2007) EISA for 2022. Improvements in process equipment,

enzymes and microbial systems, as well as improved understanding of the basis for biomass recalcitrance are critical determinants of the successful implementation of biorefineries.

Acknowledgements This work was funded by the US DOE Office of the Biomass Program. The authors also acknowledge the valuable intellectual insights provided by Dr. James McMillan, National Bioenergy Center, National Renewable Energy Laboratory, on issues related to transport processes in biochemical conversion of lignocellulosic biomass.