L. O. Ingram, X. Lai, M. Moniruzzaman, В. E. Wood, and S. W. York

Department of Microbiology and Cell Science, Institute of Food and
Agricultural Sciences, University of Florida, Gainesville, FL 32611

The technologies to convert carbohydrate components of lignocellulose into fuel ethanol are currently available. Today’s challenge is to assemble these into a cost-effective process for commercial demonstration while mounting a vigorous research and development program to achieve incremental improvements. Bacteria such as Escherichia coli КО 11 have been engineered to produce ethanol from all of the sugars present in lignocellulose by adding the genes (pdc, adhB) encoding the ethanol pathway from Zymomonas mobilis. These Z. mobilis genes have been integrated into the E. coli chromosome to produce a stable organism for industrial applications. The effectiveness of KOll has been demonstrated at 150-L scale with hemicellulose syrups and at 10,000-L scale with laboratory sugars, producing over 40 g ethanol/L within 48 h (greater than 90% of theoretical yield). Additional organisms have been engineered for the fermentation of cellulose using an analogous approach. One of these, Klebsiella oxytoca P2, has been investigated in some detail. Cellulase enzymes represent a major cost associated with ethanol production from lignocellulose. K. oxytoca P2 contains native enzymes for the uptake and metabolism of cellobiose and cellotriose, eliminating the need for one of the major cellulase components, 8-glucosidase. Additional studies have shown that the cellulase requirement for ethanol production can be reduced by enzyme recycling during simultaneous saccharification and fermentation. With mixed waste office paper, projected cellulase costs (production on site) can be as low as $0.085 cents (U. S.)/L with a yield of 539 L/metric ton. Recent studies have identified inexpensive nutrients which can be used for industrial processes and opportunities for synergy between lignocellulose to ethanol plants and grain-based or cane-based ethanol plants.

© 1997 American Chemical Society

The United States is dependent upon a secure supply of inexpensive oil. Today, over half of this oil is imported at a direct cost equivalent to half of the national trade imbalance and a military cost exceeding the annual trade imbalance. Although much of the U. S. money used to purchase foreign oil returns through the sale of arms, this proliferation of arms has created further problems and expenses. The strategic and societal benefits from replacing imported oil in the U. S. with renewable domestic fuels such as ethanol would be tremendous (1-4). Lignocellulosic residues from landfills, agriculture, and wood-processing represent both a disposal problem and a potential resource for bioconversion processes. Many excellent books and reviews have recently been published concerning progress in this area. The reader is referred to these for information concerning progress with other systems (5-13). The current paper will be restricted to our studies with genetically engineered bacteria and comparison to cellulose fermentation by yeasts.

Lignocellulose is a complex material composed of cellulose fibers embedded in a molecular matrix of lignin and hemicellulose. In native lignocellulose, cellulose typically represents 45% and hemicellulose approximately 30% of the dry weight. Both of these carbohydrate polymers must be converted to soluble sugars prior to fermentation. Lignin is a phenolic polymer which can be used as boiler fuel for product recovery. In wood-based ethanol plants dating back to the 1900s, high concentrations of mineral acids were used for depolymerization. However, environmental problems associated with the use and disposal of large amounts of mineral acids prevent a similar process today. Variations of this process could be used, however, provided cost-effective methods of acid recovery are employed. Total enzymatic solubilization of lignocellulose represents the other extreme. Both cellulose and hemicellulose can be depolymerized using microbial enzymes but may include high costs for enzymes. Our focus has been to develop a hybrid approach in which dilute sulfuric acid is used to hydrolyze hemicellulose to monomer sugars and to expose cellulose for enzymatic hydrolysis (Figure 1).

Fermentation of Hemicellulose Sugars after Dilute Acid Pretreatment of Lignocellulose

Effect of Dilute Acid Pretreatment. Dilute acid pretreatment (0.2-2.0% sulfuric acid, 120-220°C) of native lignocellulose serves three important functions in a lignocellulose conversion process: 1) hydrolysis of the hemicellulose components to produce a syrup of monomeric sugars; 2) exposure of cellulose for enzymatic digestion (removal of hemicellulose and part of the lignin); and 3) solubilization of heavy metals which may be contaminating feedstocks. This treatment also causes potential problems such as the production of an acid stream which must either be recovered for reuse or neutralized to produce inert solids such as gypsum (sold or disposed). However, with feedstocks which are contaminated with heavy metals (ie., lignocellulosics from municipal solid waste), these metals are concentrated and recovered in the gypsum fraction providing an environmental advantage. A second problem associated with dilute acid pretreatments is the formation of compounds in the syrups which are toxic to biocatalysts such as bacteria and yeasts. Indeed, considerable art exists in the area of toxin amelioration and optimization of hydrolysis conditions to minimize the generation of these toxins.

The production of a hemicellulose syrup which contains high concentrations of sugars is particularly challenging (Figure 2). Since the sugar yield (hemicellulose) from lignocellulosic feedstocks is typically 0.2-0.3 g sugar/g feedstock (dry weight), hydrolysis must be carried out with 25%-35% solids and little free liquid (Table I). One exception is com hulls and fibers which have an unusually high content of hemicellulose plus starch. Hemicellulose hydrolysis of woody substrates under more dilute conditions is potentially more expensive since it would increase the amount of required acid (neutralization and disposal as well) and necessitate the addition of process steps for the concentration of sugars prior to fermentation.

Table I. Hemicellulose Syrups from Dilute Acid Hydrolysis

Metabolic Engineering of Recombinant Escherichia coli for Pentose Fermentations. The complex mixtures of sugars which comprise hemicellulose hydrolysates pose a special problem for bioconversion into useful products. Typically, these contain predominantly pentose sugars although in some cases equal amounts of hexose and pentose sugars are present. In 1987, we developed E. coli strains which efficiently ferment all of the hexose and pentose sugars present in hemicellulose syrups at near theoretical yields (14,15). This was done by expressing the two Zymomonas mobilis genes encoding the ethanol pathway (pdc, adhB) at high levels in E. coli using plasmids (16-18). In the initial constructs, both the native transcriptional promoter and terminator were removed from the pdc gene. The remaining pdc coding region was ligated upstream from a promoterless adhB gene (including the adhB transcriptional terminator). This artificial operon was completed by its insertion behind the lac promoter in pUC18. The resulting construct contains a portable ethanol production pathway which, with minor modifications, can be engineered for expression in many different host organisms. The rationale behind this approach lies in the relative simplicity in genetically engineering the production of soluble, cytoplasmic enzymes which can effectively redirect central metabolism in microorganisms that contain other useful properties such as the utilization of many different types of sugars, secretion of hydrolases, tolerance to environmental extremes or toxins, etc.

190

A variety of E. coli hosts strains were investigated for ethanol and environmental tolerance prior to selecting ATCC11303 as the host for further investigation (19,20). This host, strain B, is quite safe and has been the subject of research investigations for over 50 years. An improved biocatalyst was developed by integration of the Z. mobilis ethanol genes into the chromosome of E. coli В to produce strain КОЗ (21). These genes were integrated behind the pfl promoter, a strong promoter which is always active in E. coli. Prior to integration, a chloramphenicol gene was added downstream the adhB gene to facilitate selection which proved very fortuitous. The original integrated strain did not express ethanol pathway genes at sufficiently high levels to effectively divert metabolism. Single-step mutants were subsequently selected by plating on solid medium containing 600 mg chloramphenicol/L. Resulting mutants expressed the ethanol production genes at a 10-fold higher level and effectively diverted pyruvate metabolism to ethanol. One of these, strain K04, was further improved by inserting a frd mutation (fumarate reductase) to eliminate the production of succinate, creating strain KOll. Studies were conducted to optimize fermentation conditions and investigate nutrient requirements (22,23). A variety of subsequent mutations and selections have yielded strains incapable of fermenting glucose while retaining the ability to ferment galactose and all pentose sugars, and other mutants which ferment sugar mixtures with minimal diauxie (24). Additional mutants have been selected with improved ethanol tolerance.

Fermentation of hemicellulose syrups with К coli KOll. A pilot plant for hemicellulose fermentation was constructed in Gainesville, Florida by BioEnergy International (University of Florida licensee purchased by B. C. International, Hingham, MA in 1995). Both a rotating batch reactor similar to a small cement mixer (6 kg charge) and a continuous, dual horizontal screw reactor (1 ton per day) were developed and used to define process conditions needed to yield fermentable hemicellulose syrups containing high sugar concentrations (Table I). Many substrates were tested including bagasse, com stover, com cobs, com fiber from wet milling, sawdust, etc. (25-30). Sugar content was maximized and acid utilization minimized by working at very high solids (Figure 2). After proprietary treatments to ameliorate toxins, hemicellulose syrups were fermented to ethanol at a 150-L scale with high yields (Figure 3; Table II) (26). Other studies confirmed effective fermentations with laboratory sugars at a 10,000-L scale. Com steep liquor (26), crude yeast hydrolysates (31) or crude soy hydrolysates which can be produced on site (32), and other industrial nutrients have been investigated and optimized to minimize costs of ethanol production. In many cases, ethanol concentration was limited by sugar content rather than ethanol tolerance. Beer concentrations above 40 g ethanol/L were readily produced when sufficient sugars were present. Materials costs for acid, base, and nutrients are estimated to be from $0.05-0.10/L ethanol and represent opportunities for improvements.

In contrast to gasoline which is spark-ignited, DF after injection is ignited by the heat of compression in a diesel engine. The diesel engine is therefore also termed a compression-ignition (Cl) engine. The differences in the ignition processes entail significant differences in chemical composition and physical properties of the fuels.

Conventional DF is, like gasoline, obtained from cracking of petroleum. It is a fraction boiling at an initial distillation temperature of 160° (90% range of 290-360°C) (7), also termed middle distillates because of its boiling range in the mid-range of cracking products.

The ignition quality of DF is commonly measured by ASTM D613 and reported as the cetane number (CN). Ignition quality is defined by the ignition delay time of the fuel in the engine. The shorter the ignition delay time, the higher the CN. To rank different compounds on the cetane scale, hexadecane (СІ6Нз4; also called cetane), which has a very short ignition delay, has been assigned a CN of 100. At the other end of the scale, 2,2,4,4,6,8,8-heptamethylnonane (HMN; also CI6H34), which has poor ignition qualities, has been assigned a CN of 15. It should be noted that the cetane scale is arbitrary and that compounds with CN > 100 (although the cetane scale does not provide for compounds with CN > 100) or CN < 15 have been identified. The ASTM specification for conventional DF (ASTM D975) requires a minimum CN of 40.

The CN scale clarifies an important aspect of the composition of, or, on a more fundamental level, the molecular structure of the compounds comprising DF. Long — chain, unbranched, saturated hydrocarbons (alkanes) have high CNs and good ignition quality while branched hydrocarbons (and other materials such as aromatics) have low CNs and poor ignition quality.

Since both too high and too low CN can cause operational problems (in case of too high CN, combustion can occur before the fuel and air are properly mixed, resulting in incomplete combustion and smoke; in case of too low CN, engine roughness, misfiring, higher air temperatures, slower engine warm-up and also incomplete combustion occur), most engine manufacturers designate a range of required CN for their engines. In most cases, this range is around CN 40-50.

Conventional DF is classified into different grades by ASTM D 975. This classification is the following: No. 1 diesel fuel (DF1) comprises volatile fuels oils from kerosene to intermediate distillates. They are applicable for high-speed engines whose operation involves frequent and relatively wide variations in engine load and speed. Such fuel is required for use at abnormally low temperatures. No. 2 diesel fuel (DF2) includes distillate gas oils of lower volatility. This grade is suitable for use in high­speed engines under relatively high loads and uniform speeds. DF2 can be used in engines not requiring fuels having the greater volatility and other properties specified for No. 1 diesel fuels. DF2 is the transportation diesel fuel to which biodiesel is usually compared. No. 4 diesel fuel (DF4) covers the more viscous distillates and their blends with residual fuel oils. It is usually satisfactory only for low-speed and medium-speed engines operated under sustained load at nearly constant speed.

Besides the just discussed characteristics of conventional DF, other properties such as heat of combustion, pour point, cloud point, and viscosity are of great significance. These properties also play very important roles in the use of biodiesel.

The two general types of diesel engines are the direct injection (DI) engine and the indirect injection (IDI) engine (5). In DI engines, the fuel is directly injected into the combustion chamber in the cylinder. In IDI engines, the fuel is injected into a prechamber which is connected with the cylinder through a narrow passage. Rapid air transfer from the main cylinder into the prechamber promotes a very high degree of air motion in the prechamber which is particularly conducive to rapid fuel air mixing (8). Combustion beginning in the prechamber produces high pressure and the fuels are subjected to high shear forces. The IDI engine is no longer used for heavy bus and truck engines due to somewhat lower efficiency and higher fuel consumption than the DI system (8). However, for special purposes, such as underground work, IDI engines are still made in the heavier class due to low exhaust emissions. For smaller vehicles such as cars and light trucks, the IDI system is used because of its ability to cover a wider speed range. The low exhaust emissions in combination with the wider speed range may lead to a continued use of IDI engines in urban areas, where the demand for low emissions can be more important than a somewhat higher fuel consumption combined with low annual mileage. The IDI engine is also less sensitive to fuel quality (8). Tests of biodiesel as a fuel have been performed on both DI and IDI engines.

Ammonia Steeping. Dry wood chip was mixed with 20% (w/v) aqueous ammonia with a solid to liquid ratio of 1 to 5 in a 250-ml flask and was incubated on a shaker at 24°C for 24 h. The mixture was then filtered to separate the solid from the ammonia solution. The partial delignined wood chip was collected by filtration, washed, and vacuum dried. The same procedures were also used for grounded com cob, except the ammonia concentration was 10% (w/v). The lignin fraction was recovered as a precipitate after the vacuum removal of ammonia and the evaporated ammonia was collected by cooling trap.

Dilute Acid Hydrolysis. The partial delignined materials were subjected to further treatment by 1% hydrochloric acid solution at 100-108°C for 1 h with a solid to liquid ratio of 1 to 5. The solid cellulose fractions were separated from hemicellulose hydrolysates by filtration and washed in deionized water to remove residual acid.

Enzymatic Hydrolysis of Cellulose Fraction. Cellulose fractions of com cob (20 g) were mixed with 0.05 M phosphate buffer (pH 5.0) and 1.0 ml cellulase enzyme (equivalent to 8.5 IFPU/gm com cob) to a total volume of 50 ml in a 250-ml flask. The saccharification was carried out at 50°C for 48 h.

Simultaneous Saccharification And Fermentation Of Cellulose Fraction.

Wood chip and com cob cellulose fractions obtained after pretreatment were autoclaved for 15 min at 121 °С in a 250-ml Erlenmeyer flask. After cooling, sterile nutrient solution (YMP) was added to a final volume of 80 ml. Initial pH was adjusted to pH 5.5 by potassium phosphate buffer. Concentrated bacterial cells (2 g/L dry equivalent) and cellulase (8.5 IFPU/gm original wood chip or com cob) were added, and the flasks were incubated in a Controlled Environment Incubator Shaker (New Brunswick Scientific Inc.) at 200 rpm at a constant temperature of 32°C. Dissolved oxygen and pH were not monitored, nor were they controlled during fermentation. Samples (0.2 ml) were taken at 12 hour intervals for 96 h. Samples were then centrifuged and analyzed.

Analysis. Glucose, xylose, acetoin, 2,3-butanediol, ethanol, and acetic acid were analyzed by HPLC method (column: HPX-87H, 300×7.8 mm, Bio-Rad, Richmond, CA). Cellulose and hemicellulose were determined according to the procedures described (31). Lignin was determined as Klason lignin by weight method (32). The residual ammonium ion in solution was determined by 05800-05 Solution

Analyzer equipped with ammonium electrode (Cole-Parmer Instrument Co., Niles, IL).

The results of the laboratory fermentations confirm that CO2 is inhibitory to the ethanol fermentation. This inhibition manifests itself in two ways: a lower rate of glucose consumption and ethanol accumulation at higher CO2 levels, associated with a lower peak cell count at 2.5 and 3.5 atm and an increased rate of cell death at

3.5 atm. The lower rate would lead to incomplete conversion of available glucose at the higher CO2 levels if the fermentations were terminated at 45 h. Letting the fermentations continue to 65 h or more, there was little or no difference in the ultimate conversion of glucose except possibly at 3.5 atm. Two reservations accompany these conclusions. 1) Each condition was only tested once, so that no statistical test can be applied. The effects described here are greater, usually substantially greater, than the intrinsic uncertainties of the analytical methods. The reproducibility of the laboratory fermentation itself, however, was not tested due to limitations of time and resources. 2) Not all the carbon can be accounted for at 3.5 atm and in the air-supplemented run, though this may reflect a simple setup error. There was no indication of increased evaporation or acid production. The calculation of the percentage completion was based on the final ethanol concentration and was not influenced by any error in carbon recovery or amount of carbohydrate added. Similar slowing of the fermentation and decreases in the peak cell number have been observed at lower temperature by workers in the brewing industry (6-9). Slowing of industrial alcohol fermentation of molasses by CO2 was observed by Ukrainian workers (28). However, if the concentration rather than the partial pressure of carbon dioxide is the determining factor (6), it is somewhat surprising to observe appreciable slowing of the fermentation at pressures as low as

1.5 atmospheres, since the solubility of carbon dioxide is so much decreased at 32 °С. This slowing of the fermentation must arise in part from causes other than decreased cell count, since cell count was not decreased at 1.5 atm.

The laboratory experiments were run as fed-batch fermentations with pure glucose as the carbon source. This design was chosen to avoid any uncertainties associated with the use of complex substrates and enzymes in situ. The remainder of the medium was based on a well-characterized defined medium with ammonium chloride and yeast extract added as sources of nitrogen. Although we intended the lab runs to simulate the most important features of plant conditions, there were some noticeable differences. We were unable to get good fermentation rates or cell counts until we increased the level of N in the media. We now think it likely that the lower nitrogen media led to N-limitation of the inoculum; the runs employing the lower-N media are not reported here. Even with the higher N level, we could not match the cell count, the retention of cell vitality or the fermentation rate of the industrial runs. This is possibly due to the leaner semi-defined media used in the lab runs in comparison with the rich, complex ground-corn medium employed in the plant. Also, the base medium we used is not a particularly good match for the inorganic components of the com mash. Supplementation with higher levels of N, lipids, or potassium did not improve the performance in shake flask experiments (not shown). The lab runs were a good match for the industrial runs in the critical variables of pH and temperature, particularly during the latter part of the fermentation when the CO2 inhibition manifests itself.

Glycerol production represents an economically significant diversion of carbon, about 4-5%, from the production of ethanol (18 and our results). Glycerol production under plant conditions is largely growth associated. Under lab conditions, glycerol production was largely non-growth-associated; this may be related to the fact that cell growth in the industrial fermentation was nitrogen limited, while growth in the main set of lab runs was not, or it may be related to the higher rate of cell death in the lab runs. The limiting factor in the lab fermentations was not identified. Glycerol formation was least in the 3.5 atmosphere lab run, and greatest at 0.5 atmospheres. This is contrary to Oura’s hypothesis (18) but in accord with the results of Kuriyama et al. and Bur’yan & Volodrez (19,20). These two papers employed semiaerobic continuous fermentations while our experiments were anaerobic batch fermentations. Oura based his suggestion of increased glycerol production at elevated CO2 on the requirement of pyruvate carboxylase for CO2 in

—°— log viable — glycerol -o— ethanol/10 ■— C02, mM count Figure 3b. Product Levels, Air Supplement 0—23 h

time, h

g/l. h —■— g/vc. h

Figure 3c. Glycerol Productivity, Air Supplement 0-23 h

order to produce oxaloacetate, which is necessary for succinate production.

However the Km of yeast pyruvate carboxylase for KHCO3 is in the range 2-8 mM (29), so the enzyme should be easily saturated with CO2. Kuriyama et al. suggest pyruvate dehydrogenase as the site sensitive to CO2 but their results and ours may also be related to inhibition of pyruvate carboxylase by high CO2 since Foster & Davis (30) found that high CO2 inhibited Rhizopus pyruvate carboxylase. Although high levels of CO2 suppressed some of the glycerol production the concomitant decrease in fermentation rate means that increased CO2 levels would not be a good strategy for increasing yield of conventional batch fermentations under industrial conditions. Oura demonstrated that low levels of aeration in a continuous fermentation led to decreased glycerol production (31). We were able to manipulate the rate of glycerol production with air supplementation, both in the lab and in the plant. The air-supplementation in each case was calculated to be just sufficient to permit the cells to retain redox balance without producing glycerol, assuming 50% oxygen uptake. In the lab, oxygen eliminated the initial spike of growth-associated glycerol productivity and decreased the non-growth-associated productivity on a per-cell basis but the accompanying higher cell count resulted no decrease in overall production. The oxygen uptake in the lab fermentation was substantially more than 50%, but decreases in the flow rate compensated for this so that the total oxygen uptake through the run was close to the intended amount. In the plant, oxygen decreased the growth-associated glycerol production on a per-cell basis but did not eliminate it. Additional glycerol was produced later in the run, resulting in no net improvement. In the lab, aeration was continued until 53 hours and then discontinued without resulting in an increase in glycerol production. It is plausible
to think that an aeration schedule could be developed to decrease glycerol production, though more experimentation would be required to establish whether this could be achieved.

We were unable to influence the carbon dioxide level or the fermentation kinetics in the plant by operating the fermentor at reduced pressure. This might have been more effective with a larger pump which could keep up with the peak CO2 production. The light aeration in the air-supplemented fermentation resulted in a small decrease in carbon dioxide level in the industrial fermentor, but no clear-cut improvement in fermentation kinetics. Other experiments including increased CO2 back pressure and nitrogen sparge were considered and rejected due to cost or complexity.

Conclusions

Carbon dioxide levels as low as 1.5 atmospheres, the range prevailing in the shallow fermentors at Morris Ag-Energy, slows ethanol fermentation appreciably. In these experiments this effect became noticeable after the fermentation was about half complete.

Carbon dioxide at 2.5 and 3.5 atmospheres decreased the peak yeast cell count slightly and at 3.5 atmospheres was associated with more rapid cell death.

Higher carbon dioxide was accompanied by decreased glycerol production.

Air supplementation of the fermentation decreased glycerol productivity on a per — cell basis and led to higher cell counts. These effects combined to leave the overall glycerol production essentially unchanged.

Acknowledgments

This research was supported by a grant from the Minnesota Com Research and Promotion Council, with additional support from AURI (Minnesota’s Agricultural Utilization Research Institute), Morris Ag-Energy, and the Minnesota Department of Public Service. We thank Duaine Flanders and Richard W. Fulmer for encouragement and helpful advice, Gerald Bachmeier for arranging plant access, Nancy Goebel for help with assays and lab fermentations, and Sterling Keller and Craig Bremmon for help with instrumentation.

Animal fats. The most prominent animal fat to be studied for potential biodiesel use is tallow. Tallow contains a high amount of saturated fatty acids (Table II), and it has therefore a melting point above ambient temperature. Blends of tallow esters (methyl, ethyl, and butyl) with conventional DF were studied for this reason (168). Smoke emissions were reduced with the esters, particularly the butyl ester. Other features such as torque, power, and thermal efficiency did not deviate from conventional DF by more than 3% in any case. Specific fuel consumption was higher for the neat esters but only 1.8% higher for a 50:50 blend of butyl tallowate with conventional DF. A study on beef tallow and an inedible yellow grease both neat and a 1:1 (weight ratio) blend of tallow with DF in short-term engine tests with DI and IDI engines was carried out (169). The deposits were softer than those formed with reference cottonseed oil but still excessive. In a 200 h EMA test the deposits caused ring sticking and cylinder wear. Thus animal fats, like vegetable oils, were not suitable for long-term use unless modified.

Other researchers blended methyl tallowate with 35 vol-% ethanol to achieve the viscosity of petrodiesel and the fuel properties were closely related to that of No. 2 diesel fuel (170). In an investigation of blends of DF2 with methyl tallowate and ethanol (777), an 80:13:7 blend of DF2:methyl tallowate:ethanol reduced emissions the most without a significant drop in engine power output. The same authors determined numerous physical properties of blends of DF with methyl tallowate, methyl soyate and ethanol and found them to be similar to the pertinent properties of DF2.

Waste vegetable oils. Vegetable oils have many other applications, notably as food ingredients and cooking oils. Especially the latter use produces significant amounts of waste vegetable oils. These vegetable oils contain some degradation products of vegetable oils and foreign material. However, analyses of used vegetable oils claimed (7 72) that the differences between used and unused fats are not very great and in most cases simple heating and removal by filtration of solid particles suffices for subsequent transesterification. The cetane number of a used frying oil methyl ester was given as 49 (7 73), thus comparing well with other materials, but little demand could be covered by this source. Biodiesel in form of esters from waste cooking oils was tested and it was reported that emissions were favorable (774). Used canola oil (only purified by filtration) was blended with DF2 (775). Fuel property tests, engine performance tests and exhaust emission values gave promising results. Filtered frying oil was transesterified under both acidic and basic conditions with different alcohols (methanol, ethanol, 1-propanol, 2-propanol, 1-butanol, and 2-ethoxyethanol) (775). The formation of methyl esters with base catalysis (KOH) gave the best yields. The methyl, ethyl, and 1-butyl esters obtained here performed well in short-term engine tests on a laboratory high-speed diesel engine.

Hemicellulose is one major component of plant cell wall materials, comprising up to 40% of agricultural residues and hardwood. It can be hydrolyzed by using dilute acids under mild hydrolysis conditions to yield a mixture of sugars (glucose, xylose, L-arabinose, mannose) of which xylose is the major component. These xylose containing hemicellulose hydrolyzates can serve as potential substrates for xylitol production. However, during acid hydrolysis, many potentially toxic compounds such as acetic acid, furfural, phenolic compounds, or lignin-degradation products are formed which inhibit growth of yeast.

Chen and Gong (32) studied the fermentation of sugarcane bagasse hemicellulose hydrolyzate to xylitol by a hydrolyzate-acclimatized yeast strain Candida sp. B-22. With this strain, a final xylitol concentration of 94.74 g/L was obtained from 105.35 g/L xylose in hemicellulose hydrolyzate after 96 h of incubation. C. guilliermondii FTI 20037 was able to ferment a sugar cane bagasse hydrolyzate producing 18.4 g/L xylitol from 29.5 g/L of xylose, at a production rate of 0.38 g/L/h (62). This lower value, compared to that (0.66 g/L/h) of the synthetic medium, may be attributed to the various toxic substances that interfere with microbial metabolism (e. g., acetic acid). Dominguez et al. (63) studied different treatments (neutralization, activated charcoal and neutralization, cation-exchange resins and neutralization) of sugar cane bagasse hemicellulose hydrolyzate to overcome the inhibitory effect on xylitol production by Candida sp. 11-2. The highest xylitol productivity (0.205 g/L/h), corresponding to 10.54 g/L, was obtained from hydrolyzates treated with activated charcoal (initial xylose, 42.96 g/L). To obtain higher xylitol productivity, treated hydrolyzates were concentrated by vacuum evaporation in rotavator to provide higher initial xylose concentration. The rate of xylitol production increased with increasing initial xylose concentration from 30 to 50 g/L, reaching a maximum of 28.9 g/L after 48 h fermentation. The decrease in xylitol production was dramatic with further increases in the initial xylose concentration. Parajo et al. (48) later reported a xylitol production of 39-41 g/L from concentrated Eucalyptus globulus wood acid hydrolyzate containing 58-78 g xylose/L by Debaryomyces hansenii NRRL Y-7426 using an initial cell concentration of 50-80 g/L.

Roberto et al. (64, 65) tested hydrolyzed hemicellulosic fractions of sugar cane bagasse and rice straw for xylitol production in batch fermentation by C. guilliermondii under semi-aerobic condition and compared these with synthetic medium containing xylose. For all media tested, simultaneous utilization of hemicellulosic sugars (glucose and xylose) was observed and the highest substrate uptake rate was attained in sugar cane bagasse medium. Increased xylitol concentration (40 g/L) was achieved in synthetic and rice straw media, although the highest xylitol production rate was obtained in sugar cane bagasse hydrolyzate. They concluded that both hydrolyzates can be converted into xylitol with satisfactory yields and productivities. Roberto et al (66, 67) evaluated xylitol production by C. guilliermondii in a rice straw hemicellulose hydrolyzate under different conditions of initial pH, nitrogen sources and inoculum level. The xylitol yields were

0. 68 g/g for the medium containing ammonium sulfate at pH 5.3 and 0.66 g/g with urea at pH 4.5. Under appropriate inoculum conditions rice straw hemicellulose hydrolyzate was converted into xylitol by the yeast with efficiency values as high as 77% of the theoretical maximum. The production of xylitol from various hemicellulosic hydrolyzates is presented in Table Ш.

Gurgel et al. (68) studied xylitol recovery from fermented sugarcane bagasse hydrolyzate. The best clarifying treatment was found by adding 25 g activated carbon to 100 ml fermented broth at 80°C for 1 h at pH 6.0. The clarified medium was treated with ion-exchange resins after which xylitol crystallization was attempted. The ion exchange resins were not efficient but the crystallization technique showed good performance, although the crystals were involved in a viscous, colored solution.

Table EL Fermentative production of xylitol from hemicellulose hydrolyzates

Concluding Remarks

The demand for xylitol in the food and pharmaceutical industries as an alternative sweetener has created a strong market for the development of low cost xylitol production process. Various xylose rich hemicellulosic materials can serve as abundant and cheap feedstocks for production of xylitol by fermentation. The cellulosic fraction can be converted to glucose, which is then fermented to fuel ethanol by S. cerevisiae. Much research needs to be done to select a suitable microorganism that can convert xylose into xylitol efficiently in presence of other hemicellulosic sugars and to understand the regulation and optimization of xylitol production by fermentation.

Virtually all fermentative microorganisms can ferment glucose to a mixture of fermentation products, but only a few can ferment glucose selectively to ethanol and thereby achieve high ethanol production yields. Even fewer microorganisms can ferment the pentose sugars in biomass (D-xylose and L-arabinose) to ethanol at high yields. A key challenge to developing advanced processes for converting biomass sugars to ethanol at high yield is to identify or construct microorganisms that exhibit the ability to convert all hexose and pentose biomass sugars to ethanol at high yield. Another is to develop bioprocesses using such microorganisms that achieve the high performance levels necessary to realize favorable overall process economics.

The objective of strain development and process development research-to maximize conversion of heterogeneous biomass sugars to ethanol — is discussed and recent advances in biomass sugar conversion are briefly reviewed. The major polysaccharides, which comprise hardwood and herbaceous biomass species and agricultural residues, are described to motivate the fact that glucose and xylose, and in some cases arabinose, must be efficiently fermented to achieve a high process yield and favorable process economics. The various processing options available for biomass saccharification and fermentation are then outlined and recent fermentation-related accomplishments achieved at NREL and other laboratories are reviewed. Finally, overviews of fermentation strain and fermentation process development research now being pursued at NREL are provided.

Introduction to Biomass Fermentation. Lignocellulosic materials that are potential feedstocks for bioethanol production include hardwoods, herbaceous crops, agricultural residues, and municipal solid waste (i. e., wastepaper and other fractions). These types of "biomass" are comprised primarily of cellulose, hemicellulose, and lignin (76, 117). Carbohydrates account for roughly 50% to 70% of the dry mass of such materials, with lignin accounting for approximately another 20%; ash and other minor components make up the balance.

Feedstock costs represent more than one-third of all processing costs in technoeconomic cost projections of large-scale bioethanol production using energy crop feedstocks costing $42 per dry ton (43). High feedstock costs mean that an integrated biomass-to-ethanol process must achieve high ethanol yields on the feedstock to be economical. Maximizing ethanol production requires hydrolyzing all cellulosic and hemicellulosic biomass sugars to fermentable forms (e. g., monomeric) and carrying out fermentation under conditions that produce high ethanol yields. Achieving high process is challenging because of the complex heterogeneous structure and relatively refractory nature of lignocellulosic biomass to complete saccharification and high yield fermentation.

Feedstock Composition. Table 2 lists the compositions of a variety of lignocellulosic biomass types. As this table shows, glucan (from cellulose) and xylan (from hemicellulose) are the two carbohydrates at the highest levels in these types of biomass. The contribution of pentosans (xylan and arabinan) to total carbohydrates varies with biomass type, ranging from about 25% in hardwoods to 33% in herbaceous crops and 40% in agricultural residues. The levels of the minor carbohydrates arabinan, galactan, and mannan also vary with biomass type. Hardwoods typically contain more mannan and are more highly acetylated, whereas herbaceous plants and agricultural residues generally contain higher levels of galactan and arabinan. In some herbaceous crops and agricultural residues arabinan levels are high enough that conversion of arabinan-in addition to glucan and xylan — is required to achieve high overall conversion yields (118).

Processing Options. A variety of processing schemes can be used to convert lignocellulose materials to ethanol. Most importantly, the process must achieve a high conversion yield on the hemicellulosic and cellulosic sugars to be economical (43,119). Processing options differ in the method(s) chosen for hydrolyzing cellulose and hemicellulose to their component monomeric sugars (i. e., glucose, xylose, arabinose, galactose, and mannose). In addition, cellulose hydrolysis can either precede or be carried out simultaneously with fermentation. Processing schemes in which cellulose is hydrolyzed to glucose using either cellulase enzymes or concentrated acids at relatively low temperatures (lower than 100°C) typically degrade less carbohydrate than those based on higher temperature acid hydrolysis of cellulose. Although processes based on complete acid hydrolysis have a long history (120-122) and continue to be investigated (123,124), enzyme hydrolysis-based processes are considered by many to offer the best potential to maximize overall process yields (125). Current efforts at NREL are directed at developing and demonstrating enzyme-based bioethanol process technology.

Economic analyses indicate that the most important factors influencing bioethanol production cost are overall process yield and final ethanol concentration; volumetric productivity is a significant but secondary factor (119). Economic analyses also show that substantial capital and operational savings are possible in advanced process designs in

Composition (wt%, dry basis) Total

Hexosan Pentosan Klason Carbo-

which hexose and pentose sugars are fermented simultaneously (e. g., cofermentation) (Putsche, V., personal communication, 1996).

Figure 3 depicts a generic enzyme hydrolysis-based biomass-to-ethanol processing scheme. The dashed lines represent advanced processing options in which separate biologically mediated steps are combined. Processing begins with a "biomass handling" step wherein the biomass is milled to an appropriate size before being subjected to some type of thermochemical treatment, e. g., with steam and dilute sulfuric acid, to open up the lignocellulose pore structure and increase its susceptibility to enzymatic attack (126). This pretreatment step effectively hydrolyzes and solubilizes the relatively more labile hemicellulosic sugars, thereby producing a hydrolyzate liquor typically rich in pentose sugars and a cellulose-rich solid with greater porosity and dramatically increased enzymatic digestibility.

In the base-case process configuration shown in Figure 3, the insoluble pretreated lignocellulosic solids are converted to ethanol in an SSF bioreactor. Hydrolytic cellulase enzyme(s) and fermentative microorganism(s) are present in the SSF bioreactor. The cellulase enzymes hydrolyze the cellulose to glucose, which is then fermented to ethanol by a suitable fermentative microorganism (e. g., brewers’ yeast). The SSF process is efficient because it reduces end-product inhibition of the cellulases by glucose through the continuous fermentative conversion of glucose to ethanol (127). It also reduces capital costs by reducing the number of vessels relative to those required to separately carry out enzymatic hydrolysis and fermentation. In the base-case scheme, the fermentative microorganism used in SSF does not ferment hemicellulosic sugars (e. g., xylose), so a separate pentose fermentation step is required. An enzyme production bioreactor is also required, because the fermentative microorganism used in SSF does not produce cellulolytic enzymes. Distillation is used to recover the ethanol produced in the separate SSF and pentose fermentation unit operations. The lignin fraction and other residual solids carried through the process are sent to a cogeneration plant to supply steam and electricity for the process (not shown).

The advanced processing option indicated in Figure 3 by the inner dashed rectangle (short dashes) incorporates a (novel) microorganism that can ferment hemicellulosic and cellulosic sugars to ethanol at high yield in an SSCF bioreactor. In this case, only a single step is required to enzymatically saccharify and ferment the process liquor and solid fractions of pretreated biomass. As in the base-case scheme, a separate enzyme production unit operation is required.

The scheme represented in Figure 3 by the outer dashed rectangle (long dashes) shows the most advanced processing option, one in which all biologically mediated steps-enzyme production, enzymatic cellulose hydrolysis, and biomass sugar fermentation-occur in a single bioreactor. This process, also known as direct microbial conversion (DMC), can be carried out to various extents by a number of microorganisms, including fungi such as Fusarium oxysporum (128) and bacteria such as thermophilic Clostridia (129, 130). However, DMC strains exhibit relatively low ethanol selectivity; current strains produce at least one other fermentation product, typically acetic acid, at nearly equimolar levels with ethanol.

The designs shown in Figure 3 are considerably more complicated than those used for traditional starch-to-ethanol (or sugar-to-ethanol) processes already in commercial operation. This is because starch is much easier to hydrolyze to glucose than cellulose. And starch contains only glucose, whereas biomass substrates contain significant levels of

other sugars, most notably xylose. Starch-based processes require neither cellulolytic enzyme production nor pentose fermentation to achieve high conversion yields. Another difference is that solids handling requirements are typically higher for biomass substrates, especially if cellulose hydrolysis is carried out in an SSF-mode rather than before fermentation {131). All these characteristics impose significant challenges for bioethanol process development beyond those encountered in starch-to-ethanol processing.

Trends in Fermentation Strain Development. Whereas a number of microorganisms can efficiently ferment glucose, the product of cellulose hydrolysis, to ethanol, only recently has conversion of the pentose sugars in the hemicellulose fraction become feasible. Until a few years ago, the few known organisms that could utilize either D-xylose or L-arabinose did not perform well, but typically grew slowly on pentoses and achieved relatively low ethanol yields and productivities. Therefore, identifying and developing microorganisms that can selectively convert D-glucose, D-xylose and L-arabinose to ethanol at high yield has been the focus of extensive research during the past 20 years. During the past decade, the sophistication of molecular biology has grown tremendously, and numerous attempts have been made to use recombinant DNA technologies to engineer superior microorganisms for bioethanol production. Only a few of these efforts have been highly successful; many are still in progress {132).

Metabolic pathway engineering is increasingly being recognized as a powerful approach for developing microorganisms that can efficiently convert biomass sugars to ethanol In broad terms, superior ethanol-producing microorganisms can be developed by several metabolic engineering approaches.

• Increasing ethanol product selectivity in strains exhibiting broad substrate range but poor product selectivity

• Broadening the substrate range to include important biomass sugars besides glucose (e. g., arabinose, galactose, mannose or xylose) in strains that exhibit good product selectivity but that cannot ferment sugars other than glucose to ethanol

Of course, beyond these two basic approaches, significant metabolic engineering may also be required to stabilize “improved” strains or to enable such strains to achieve high ethanol yields and fermentation productivities.

Following the first approach, E. coli and Klebsiella oxytoca have been engineered to be highly effective ethanol producers by introducing the genes for ethanol production from Z. mobilis {133-135). Extensive evaluations of these “ethanologenic” strains have been carried out, both in media that contain pure sugars and in pretreatment hydrolysates derived from a variety of feedstocks {136-141).

The second approach-broadening the substrate utilization range of strains that are highly efficient ethanol producers-has been demonstrated at NREL by introducing the xylose assimilation and pentose phosphate pathway genes from E. coli into Z. mobilis to confer the ability to ferment xylose to ethanol at high yield {142). More recently, an arabinose-fermenting Z. mobilis strain has been similarly developed by introducing the arabinose assimilation and pentose phosphate pathway genes from E. coli into Z. mobilis {143). Both of these accomplishments are described in more detail below.

More recently, long term efforts to develop xylose-fermenting Saccharomyces sp. have also been successful, with xylose fermentation reported for strains transformed with the xylose reductase and xylitol dehydrogenase genes from P. stipitis and overexpressing xylulokinase {144-146). Other recent achievements in the metabolic engineering of

superior ethanol producers include initial success at inproving the performance of xylose — fermenting yeasts by optimizing the expression of genes that encode xylose assimilation and ethanol production pathways (147-148). Successful transformation of pentose- fermenting Clostridium thermosaccharolyticum has also been reported (149), which provides a key tool for further developing this microorganism by altering product selectivity to favor ethanol production.

The recombinant E. coli and K. oxytoca and one of NREL’s recombinant Z. mobilis can ferment arabinose to ethanol, unlike recombinant xylose-fermenting Saccharomyces sp. and wild-type xylose-fermenting yeasts such as P. stipitis, which can aerobically grow on arabinose but can not anaerobically ferment arabinose.

Current Thrust of Fermentation Strain Development at NREL. Our near-term goal is to develop microbial catalysts that effectively convert sugar streams from hardwood sawdust in which glucose and xylose are the major sugar substrates. In the intermediate term, our goal is to develop strains for converting herbaceous energy crops such as switchgrass, which contain arabinose in addition to glucose and xylose. Our previous in­house research efforts have focused primarily on increasing the substrate range of Z. mobilis. This bacterium’s high conversion yield, fermentation product selectivity, and ethanol tolerance levels are key attributes for a commercially viable ethanol-producing strain. To develop this organism for conversion of mixed hexose and pentose sugar streams derived from lignocellulosic feedstocks, we successfully engineered this bacterium to utilize both xylose and arabinose (742,143).

Xylose-fermenting Z. mobilis was developed through the coordinated expression of the E. coli xylose isomerase, xylulokinase, transketolase, and transaldolase genes (142). The recombinant Z. mobilis could grow on xylose as the sole carbon source and produced ethanol at 0.44 g/g of xylose consumed, corresponding to 86% of theoretical yield. The xylose assimilation and pentose phosphate metabolism genes were introduced on two operons under the control of strong constitutive promoters, which allowed the strain to rapidly ferment a mixture of glucose and xylose to ethanol at 95% of theoretical yield. In a medium that contained glucose and xylose, the xylose-fermenting strain utilized glucose at a faster rate than xylose, but both sugars were utilized simultaneously (142). This strain also performs well at converting cellulose and xylose to ethanol under SSCF conditions in which cellulase enzymes are present (150).

Arabinose-fermenting Z. mobilis was developed by introducing the L-arabinose isomerase, L-ribulokinase, L-ribulose-5-phosphate-4-epimerase, transaldolase, and transketolase genes from E. coli (143). The recombinant bacterium ferments both arabinose and a combination of glucose and arabinose to ethanol at more than 95% of the maximum theoretical ethanol yield on a consumed sugar basis. Arabinose-fermenting Z. mobilis does not yet perform as well on sugar mixtures as xylose-fermenting Z. mobilis. In media that contain glucose and arabinose, the arabinose-fermenting strain utilizes arabinose at a considerably slower rate than glucose, and only after glucose is nearly depleted. The arabinose-fermenting strain may be used in mixed culture with the xylose — fermenting Z. mobilis strain to efficiently ferment glucose, xylose, and arabinose, the predominant biomass sugars in agricultural residues and herbaceous energy crops (see Table 2).

Future directions for NREL’s strain development research include continuing efforts to develop superior Zymomonas strains for bioethanol production. One focus is to use recombinant DNA technologies to combine xylose and arabinose fermentation capabilities into a single Zymomonas strain. Beyond this, we are pursuing development of metabolically engineered cofermenting Zymomonas strains that exhibit the high level of stability and robustness required for a commercial fermentation biocatalyst. We are currently developing methodologies for integrating genes into the Zymomonas chromosome. Ultimately, we plan to introduce the pentose-fermenting genes (without the current antibiotic resistance marker gene) directly into the Zymomonas chromosome to ensure that the strain will be stable and compatible with good large-scale manufacturing practices. We also plan to evaluate the potential to use Zymomonas strains that reportedly exhibit particularly high thermo — and ethanol tolerance as improved hosts for metabolic engineering of pentose fermentation capability. Some of these host strains will likely enable recombinant strains that are more tolerant of inhibitory pretreatment hydrolysates to be developed. The objective is to integrate genes that confer xylose and arabinose fermentation capabilities into the chromosomes of such strains. In addition to these activities, we plan to investigate other innovative molecular approaches to inproving the performance of pentose-fermenting Zymomonas strains, including:

• Identifying and alleviating metabolic bottlenecks that limit fermentation performance

• Introducing improved pentose transport systems to increase sugar utilization rates

• Developing novel strategies to minimize by-product formation.

By implementing these efforts, we hope to eventually develop a “super” Zymomonas strain capable of robust high yield cofermentation of glucose, xylose, and arabinose mixtures to ethanol.

We are also attempting to develop other ethanologens. In a previous survey of prospective host microorganisms for developing superior fermentation strains, we identified Lactobacillus sp. as attractive candidates for metabolic engineering of pentose utilization and ethanol production. (757). Experiments carried out at NREL demonstrated that many Lactobacillus sp. exhibit higher tolerance to hardwood hemicellulose hydrolysates than Zymomonas sp. Although most of our work to date has focused on developing and inproving pentose-fermenting Zymomonas strains, we have initiated research to develop Lactobacillus sp. that can ferment the major biomass sugars to ethanol. We have successfully introduced a xylose assimilation pathway into a homofermentative Lactobacillus strain that previously could not utilize xylose (Franden,

M. A., personal communication, 1996). Our future efforts in this area will be to redirect carbon flow from pyruvate to ethanol.

Trends in Fermentation Process Development. Around the world, many efforts are under way to develop, scale up, and commercialize biomass conversion technologies. Active programs to develop cost-effective biomass production systems and biomass conversion technologies exist in many countries of the Americas, Europe, and Asia (752­756). In the United States, DOE/NREL has a particularly active ethanol project that has made significant accomplishments during the past several years. Several bioethanol development programs, including NREL’s, are at the stage where integrated processes are beginning to be demonstrated at pilot scales (10-1000 kg/day dry biomass) {124, 757). Most processes being developed are based on enzymatic hydrolysis of cellulose. Economic sensitivity analyses of enzyme hydrolysis-based processes show that improvements in biomass carbohydrate hydrolysis (Le., saccharification by thermochemical pretreatment in combination with enzyme hydrolysis) offer the greatest potential to reduce bioethanol production costs (158).

The most notable breakthrough during the past several years is the development of improved fermentative microorganisms that can ferment pentose and hexose sugars to ethanol at high yield (759, 760,142,144-146 ). Several of these novel microorganisms, including transformed strains of E. coli (140), K. oxytoca (161), and Z. mobilis (150) perform well under SSF-type conditions. Current efforts at NREL focus on developing processes in which pentose and hexose sugars are fermented together by xylose-fermenting Z. mobilis using an SSCF process configuration, as described above.

The enzyme hydrolysis-based processing option that arguably remains the least developed-but potentially the most promising-is the DMC process in which enzyme production, cellulose hydrolysis, and fermentation are carried out in a single vessel. NREL continues to fund subcontracts in the area of DMC processing. Although significant progress has been made during the past several years in demonstrating DMC-based conversion of lignocellulosic feedstocks (762), considerable progress in strain development remains to make DMC an economical processing option for bioethanol production.

Current Thrust of Fermentation Process Development at NREL. Current efforts at NREL focus on developing and demonstrating an integrated bioethanol process. Rigorous bench-scale experiments are being conducted to identify and characterize the performance of a hardwood sawdust-to-ethanol process based on an SSCF process configuration. In addition to SSCF, distinct processing elements under active development at NREL include pretreatment, conditioning of pretreated materials (detoxification), production of cellulase enzyme(s), and development of improved fermentation strains, including long-term adaptation of strains to develop strains that perform better under inhibitory processing conditions.

Previous process development efforts to optimize processing conditions for converting biomass to ethanol at NREL and other laboratories have generally focused on specific unit operations rather than on the whole process. As a result, previous studies do not provide sufficient information to predict how an integrated bioconversion process will perform. We believe, as do Rehn and co-workers (163), that integrated process testing is the most effective way to identify and optimize operating conditions for an overall bioethanol production process. Moreover, we believe that integrated process studies are most expedient for identifying important interactions that may exist between linked process components.

The major components of the integrated bioethanol process currently under development at NREL are summarized in Figure 4. These include: (1) dilute acid pretreatment of hardwood sawdust feedstock using a Sunds hydrolyzer pilot-scale pretreatment reactor (124,157); (2) slurry conditioning (detoxification) to improve the fermentability of pretreatment liquors and solids and thereby allow the integration of pretreatment with cellulase production and with enzymatic hydrolysis and fermentation;

(3) cellulase production based on T. reesei utilizing pretreated (conditioned) hydrolysates and pretreated solids; and (4) cellulose saccharification and glucose and xylose cofermentation using cellulase enzyme(s) and NREL’s xylose-fermenting Z. mobilis in an SSCF process configuration.

The SSCF process is at the heart of the overall process and achieving high SSCF

conversion yields is the primary objective of fermentation process development efforts. Improved bioethanol process economics are achieved by maximizing SSCF performance and minimizing SSCF processing costs. Process economics benefit not only from reductions in direct SSCF operating costs, but also (and perhaps more dramatically) from improvements made in other processing elements that enable better SSCF performance to be achieved (e. g., improvements in pretreatment, enzyme production, detoxification).

We have made significant progress at NREL toward demonstrating cost-effective bioethanol process technology. Particular accomplishments related to fermentation process development include successfully combining separate xylose fermentation and cellulose conversion unit operations into a single SSCF process. During the past year, we characterized Zymomonas-based SSCF performance as a function of carbohydrate loading and pH, demonstrating that NREL’s xylose-fermenting Zymomonas strain performs well in an SSCF process configuration (750). Our current focus is to integrate all major processing steps-pretreatment, hydrolyzate conditioning, cellulase enzyme production, and SSCF-at the bench scale. Most recently, we demonstrated a minimum level of performance for SSCF processing of conditioned (detoxified) pretreatment whole slurry. These experiments were carried using sterile-filtered cellulase enzyme broth produced using washed and pretreated hardwood sawdust solids and lactose rather than enzyme produced using only pretreated materials. Nonetheless, successfully integrating the main process technology elements is significant for two reasons: (1) it provides an excellent foundation for examining additional integrated process technology improvement; and (2) it greatly reduces uncertainty in technoeconomic projections of bioethanol manufacturing cost.

We plan to continue integrating the major processing steps at the bench scale. Integration at the mini-pilot scale will be pursued once the integrated process meets established performance criteria. Our immediate goal is to demonstrate an integrated process that achieves NREL’s near-term objectives for bioethanol production cost. Integrated process testing is planned to examine the benefits of advanced processing concepts such as incorporating component recycle streams (e. g., recycle of process water, nutrients, substrates [solids], cells, or enzymes). Integrated process testing will also be carried out to evaluate and compare alternatives to SSCF processing such as sequential (enzymatic) hydrolysis and cofermentation (SHCF) or hybrid enzymatic saccharification/fermentation processes that fall between SSCF and SHCF. Such hybrid processes would only partially separate enzymatic saccharification and cofermentation processing steps. Maximizing performance in such hybrid processes will probably require that the sequentially linked saccharification and cofermentation unit operations be carried out at various temperatures and pHs to maximize cellulose hydrolysis in the first step and ethanol fermentation in the second step. Again, integrated process studies will be used to efficiently characterize interactions that are expected between the linked enzymatic cellulose hydrolysis and microbial ethanol production processing steps.

The VFA salt concentration exiting the fermentor is approximately 25 to 40 g/L, or approximately 25 to 40 parts of water per part of VFA salt. The pKa of VFA’s is 4.8 so at the fermentation pH (~5.8), only about 10% of the VFA is present as free unionized acid; the rest is ionic salt. Both the salt and free unionized acid are less volatile than water, so distillation is not a viable separation technique.

Playne (5) discusses many techniques for recovering VFA’s from dilute aqueous solutions. Some proposed methods employ immiscible solvents (e. g., tributyl phosphate, trioctyl phosphine oxide, high-molecular-weight amines) that react with the free unionized acid and extract it from the broth. For solvent extraction to be effective, the fermentor pH must be acidic (4.8 to 5.2) which severely inhibits the microorganisms. Alternatively, if the fermentation is operated near neutrality, the fermentation broth can be acidified with mineral acids (which generates wastes) or carbon dioxide (which requires high pressures).

bCaldwell & Bryant (77) media modified by removing acetic, propionic, and butyric acids, but keeping the branched fatty acids. Methanogen inhibitor =1 mM 2-bromoethanesulfonic acid.

°MSW composition = 21.0% cardboard, 13.9% newspaper, 10.7% packaging, 10.1% printer paper, 7.8% leaves, 6.3% wood, 6.3% miscellaneous non-packaging, 5.0% books and magazines, 4.4% grass, 4.0% food waste, 3.1% tissue paper, 2.7% brush, 2.7% fats and oils, 1.9% greens. Pretreatment conditions for MSW: lime loading = 0.1 g Ca(OHyg dry biomass, water loading = 10 g HjO/g dry biomass, time = 1 h, temperature = 12ГС.

dMSW gross composition (72,73) = 14.1% lignin, 4.2% ash, 81.7% carbohydrate + fat + protein.

eSS gross composition (72,73) = 39.5% ash, 60.5% carbohydrate + fat + protein. fError band represents ± two standard deviations.

®Calcium carbonate neturalization.

hResidue washed with acetic acid to remove CaC03. Cells remain in residue.

tertra nutrients added. 0.073 g urea/g biomass. Four-fold increase in Caldwell & Bryant (77) nutrients (vitamins, heavy metals, phosphate) but not buffer, oxygen scavengers, or branched fatty acids.

Multi-effect evaporation can concentrate the salts, but it is too energy intensive. Membrane techniques (e. g. reverse osmosis, electrodialysis, water-splitting electrodialysis, carrier-mediated transport membranes) may also be considered (5), but the cost of membranes makes this prohibitive.

The MixAlco Process uses a proprietary separation procedure that is both capital and energy efficient; it overcomes a major obstacle to the economic use of mixed-acid fermentation.

Anaerobiospirillum succiniproducens. Succinate is also produced by a variety of non-ruminal anaerobic bacteria. Anaerobiospirillum succiniciproducens was isolated from beagles (dogs), and the bacteria are gram negative, anaerobic, motile, spiral­shaped of 0.6 to 0.8 pm by 3 to 8 pm size (80). These organisms grow in a temperature range of 25 to 40°C, although growth is slow at 25°C. Most A. succiniciproducens strains utilize glucose, lactose, or sucrose, while some hydrolyze starch and utilize fructose, dextrose or raffinose. The major fermentation products are succinate and acetate with lactate and formate as minor products (80). A. succiniciproducens has an absolute requirement for carbon dioxide and grows well under 0.1 atmosphere partial pressure of carbon dioxide (81,82). Tryptophan is believed to be necessary for utilization of dextrose and com steep liquor (82).

A. succiniciproducens metabolizes glucose using the pathway shown in Figure 2 (81). The pathways for succinate and acetate production from glucose are identical to those for several rumen anaerobes previously described (see Figure 1). In this case, pyruvate is also the precursor for lactate and ethanol. The formation of lactate from pyruvate is catalyzed by a NADH-dependent lactate dehydrogenase enzyme (81). Pyruvate converted to acetyl-CoA may also be reduced to acetaldehyde then to ethanol. Environmental conditions such as pH and C02-HC03“ concentration affect the catabolic end products of glucose fermentation. At high pH (7.2) and low C02- HC03‘ concentration, lactate production is favored, lactate dehydrogenase and alcohol dehydrogenase activities are detected and low phosphoenolpyruvate-carboxykinase activity is observed in the cell extract. At low pH (6.2) and high C02-HC03′ concentration, succinate is the major product, lactate dehydrogenase and alcohol dehydrogenase activities are absent while high phosphoenolpyruvate-carboxykinase activity is observed (81).

glucose

У

Phosphoenolpyruvate-carboxykinase, a key enzyme in succinate pathway, has been isolated and purified from A. succiniciproducens (83). The enzyme has a pH range for optimal activity of 6.5-7.1 with an isoelectric point of 4.9. This enzyme differs from that isolated from F. succinogenes, using ADP as a cosubstrate instead of GDP. The enzyme has an absolute requirement for a divalent metal ion. In presence of Mn2+ or Co2+ alone the enzyme activity is 7 to 8 times greater than in the presence of Mg2+ alone, while its highest activity is in the presence of either Mn2+ or Co2+ and Mg2+ (83).

Several patents (82,84-86) have been issued for the production of succinate using A. succiniciproducens with dextrose and com steep liquor as substrates. Optimal conditions for this process are 39°C, a pH range of 5.8-6.6 with a carbon dioxide partial pressure of greater than 0.1 atmosphere. Using this process 20 g/L to 40 g/L of succinate production has been demonstrated.

Clostridia species. Clostridia are the most widely dispersed of the anaerobic bacteria that produce succinate. Woods (87) noted that the importance of the Clostridia for biotechnology was recognized at the beginning of the century when Clostridium acetobutylicum was used industrially to produce acetone, butanol, and ethanol. Clostridia comprise approximately hundred anaerobic species which can utilize polysaccharides and proteins to produce industrially important products through fermentation, and some Clostridia can produce succinate from different substrates.

Clostridium thermo succinogenes, a species of thermophilic anaerobic bacteria, has been isolated from several sources, including beet pulp at a sugar refinery, soil
around a Jerusalem artichoke, fresh cow manure, and mud at a tropical pond in a botanical garden (88). C. thermosuccinogenes ferments fructose, glucose or inulin to produce succinate, formate, acetate, ethanol, lactate and hydrogen. The formation of succinate as a major fermentation product distinguishes C. thermosuccinogenes from other thermophilic Clostridia (88), although C. thermosaccharolyticum and C. thermocellum produce trace amounts of succinate (89,90). C. aminophilum can act on amino acid carbon sources to produce traces of succinate and lactate, although the major fermentation products are ammonia, acetate and butyrate (91). C. hobsonii comb. nov. isolated from an anaerobic cattle waste digester ferments glucose to produce ethanol, acetate, formate, lactate and succinate (92). C. aldrichii sp. nov. isolated from a wood fermenting anaerobic digester acts on cellobiose to produce succinate, acetate, propionate, isobutyrate, butyrate, isovalerate, lactate, hydrogen and carbon dioxide (93). Acetate, formate, butyrate and lactate are the major products of

C. innocuum strain Co 15-23 with succinate being a minor fermentation product (94). C. kluyveri ferments pyruvate to succinate (95,96), and the proposed pathway is shown in Figure 3. The amount of succinate produced by the fermentation has been directly correlated to the quantity of carbon dioxide used in the fixation reactions (95).

For C. coccoides, a species which was first isolated from the feces of mice, succinate is the major fermentation product from PYFG broth with acetate produced in moderate amounts. Like the other Clostridia, this species is obligately anaerobic, non-motile, sporeforming, gram-positive, with coccobacilli to rod-shaped morphology (97). C. coccoides has also been isolated from the fecal microflora of elderly persons in rural and urban areas in Japan (98), and from the fecal microflora of laboratory mice, rats, hamsters and rabbits (99).

Propionibacteria. Several Propionibacteria produce succinate as one of their fermentation products. Succinate is produced in the range of 7.9-26.1 mmol/100 mmol glucose with several Propionibacteria species (100). Carbon dioxide is utilized in the formation of succinate during the fermentation of glucose and glycerol by Propionibacterium species (101,102). P. freudenreichii grows anaerobically on lactate (103) to produce succinate in trace amounts (104). A study of the metabolism of aspartate using P. freudenreichii showed that 41.5 mM succinate, 40.3 mM acetate,

40.6 mM propionate and 39.8 mM carbon dioxide were produced from 82 mM lactate and 41 mM aspartate. The metabolism of aspartate to succinate and ammonium ion by P. freudenreichii is influenced by the pathway by which lactate is fermented to propionate, acetate and carbon dioxide (Figure 4) (105).

P. freudenreichii grown on the cheese originally made with Lactobacillus bulgaricus produces between 18-39 mmol succinate/kg cheese (106). Succinate synthesized by Propionibacteria is formed by one of two pathways, one that generates carbon dioxide or one that consumes carbon dioxide. The first pathway involves the enzyme carboxytransphosphorylase catalyzing a carbon dioxide fixation step, which results in succinate formation from propionate (107). The second pathway involves the metabolism of aspartate to succinate during the fermentation of lactate by P. freudenreichii subsp. shermanii (105,108). In the presence of added propionate, P. freudenreichii consumes aspartate according to the following equation (109):

3 aspartate + 1 propionate —> 3 succinate + 1 acetate + 1 C02 + 3NH3

Figure 3 Biochemical pathway for succinate production by C. kluyveri.

Growth of P. acidipropionici on lactose or glucose results in production of larger quantities of succinate than when the organism is grown on lactate. At a pH of 6.6, the yields of succinate production on lactose, glucose and lactate are 0.114, 0.073 and 0.009 g/g, respectively, with the production of succinate reduced at lower pH (110). Hsu and Yang also reported that the yields of succinate from growth of P. acidipropionici on lactose are dependent on pH (111).

Lactobacilli. Lactobacillus pentosus grown on limited glucose in the presence of citrate results in the production of acetate, formate, lactate and succinate as the major fermentation products (112). The proposed pathway involves the splitting of citrate into oxaloacetate and acetate by citrate lyase. Oxaloacetate and lactate are metabolized, respectively, to generate succinate and acetate by the pathway shown in Figure 5. The overall reaction is represented by:

2 lactate + 1 citrate —> 3 acetate + 2 formate + 1 succinate

Figure 4 Biochemical pathway for succinate production by P. freudenreichii.

In one study, 9.5 mmol of citrate and 8 mmol of glucose were transformed into 27 mmol of acetate and 9 mmol of succinate (112). The presence of nitrate and nitrite prevented the formation of formate, acetate and succinate.

A number of Lactobacilli isolated from fermented cane molasses in Thailand produce succinate in de Man-Rogosa-Sharpe broth (113). Among the strains that produced succinate were 23 of 39 Lactobacillus reuteri strains, 6 of 18 L. cellobiosus strains and 1 of 6 unidentified strains. Diammonium citrate was found to be a precursor of the succinate. Other Lactobacilli which produce succinate include L. brevis (114) which converts tartarate to succinate through oxaloacetate, malate and fumarate, and L. crispatus (115) which produces succinate from glucose. A few Lactobacilli strains isolated from ciders and perries have been shown to produce succinate from malate (116).

L. plantarum degrades L-lactate in the presence of citrate to form formate, acetate and succinate along with carbon dioxide (117). In the absence of citrate, L. plantarum metabolizes lactate to acetate and formate. Evidence suggests that oxaloacetate formed from citrate acts as an electron acceptor for the production of succinate, a result consistent with those obtained for the mannitol fermentation by L. plantarum to produce succinate (118).

Bacteriodes. Species of Bacteriodes are strict anaerobes isolated from gastrointestinal tracts, including some rumen (119). A study of the role of carbon dioxide in the metabolism of glucose of Bacteriodes fragilis (120) showed that carbon dioxide has no effect on the growth rate or cell yield at concentrations above 30% (equivalent to an available C02-HC03′ of 25.5 mM). A slight decrease in the growth rate and cell yield occurs at 20% and 10% carbon dioxide. When C02-HC03* concentrations are

Figure 5 Biochemical pathway for succinate production by L. pentosus.

below 10 mM, the lag phase lengthens and a decrease in maximal growth rate and cell yield is observed. B. fragilis has been extensively studied in continuous culture to determine the effect of carbon dioxide on the growth rate at a constant physiological state (120). At 100% carbon dioxide or 100% nitrogen, decreasing the dilution rate favors propionate and acetate production over succinate, D-lactate, L-malate and formate. When grown in 100% nitrogen, propionate is formed at a greater concentration than succinate. Except at lowest dilution rates, the reverse is true when grown in 100% carbon dioxide. In 100% nitrogen, the maximum dry cell yield is 67.9 g cells/mol glucose while in 100% carbon dioxide this yield is 59.4 g/mol (120).

In one study, Bacteriodes ovatus grown on starch (2 g/L) or arabinogalactan (5 g/L) was used, respectively, to compare carbon-limited with nitrogen-limited conditions. Succinate produced under nitrogen limitation ranged from 9-14 mM compared with 1-7.1 mM under carbon limitation. The production of succinate and propionate is dependent on dilution rate, with higher dilution rates favoring propionate (121). A similar relationship between propionate/succinate and growth rate has been demonstrated for B. thetaiotaomicron (122) and B. fragilis (120).

Масу and Probst (119) reviewed the biology of gastrointestinal Bacteriodes. B. ureolyticus produces succinate (123,124), B. melaninogenicus grows on aspartate to produce succinate (125), and B. oralis forms succinate, acetate and formate (126,127). As noted earlier, organisms originally named B. succinogenes, В. amylophilus and B. ruminicola have been respectively renamed Fibrobacter succinogenes, Ruminobacter amylophilus and Prevotella ruminicola.

Other bacteria. Desulfobacterium cetonicum can degrade acetone to form succinate from isocitrate by isocitrate lyase (128). D. dehalogenans isolated from pond water can grow on pyruvate in the presence of 3-chloro-4-hydroxyphenylacetate or fumarate as an electron acceptor produce 4-hydroxyphenylacetate and succinate in equimolar amounts (129).

Acetonema longum is a hydrogen-oxidizing and carbon dioxide-reducing acetogenic bacterial species isolated from the gut of the termite Pterotermes occidentis. The organism consumes rhamnose to produce acetate and butyrate as the major fermentation products and propionate, succinate and 1,2-propanediol as minor products (130).

Anaerobic methanogens isolated from the feces of rats fed with high-fiber or fiber-free diets generate acetate, propionate and butyrate as major products and lactate, succinate and formate as minor products. Succinate produced by the fermentations is 1.3 mmol/g dry matter for the high-fiber diet and 5.5 mmol/g dry matter for the fiber-free diet (131).

Different strains of Klebsiella oxytoca grown on 10% xylose (666 mM) for 96 hours produce 41-59 mM succinate (132).

Peptostreptococcus productus strain Co8-4 produces acetate and succinate as major fermentation products (94). Peptostreptococcus micros strain СоЗЗ-6 produces acetate, lactate and succinate as its major fermentation products (94).

Fusobacterium russii strain Co21-3 produces acetate and butyrate as major products and succinate as a minor fermentation product (94).

Eubacterium sp В86, Peptostreptococcus sp 610, Enterococcus faecalis sp 84 and Veillonella ratti sp 36 isolated from conventional rat microflora ferment starch to produce succinate, acetate and propionate (133).

Anaerovibrio burkinabensis sp. nov., a strictly anaerobic bacterium isolated from rice fields soil by using lactate as the sole carbon and energy source, ferments fumarate, malate and aspartate to produce succinate (134). A. lipolytica isolated from both ovine and bovine rumen and grown on a linseed oil/rumen fluid agar media breaks down glycerol into propionate and succinate (135).

The myxobacteria Cytophaga succinicans have been shown to degrade glucose in the presence of carbon dioxide to produce succinate, acetate and formate in a 3:2:1 ratio (136).

The cost of fungal cellulase (purchase or production on site) represents a major barrier for the commercial production of ethanol from cellulose. Due to the crystalline

* Abbreviations: Difco, 5 g Difco Yeast Extract and 10 g Difco Tryptone/L; CSL, corn steep liquor; YA, crude yeast autolysate; Max. EtOH, highest concentration of ethanol achieved during fermentation; nm, not measured; Qp, average volumetric productivity calculated by dividing total ethanol by fermentation time; gTE/gS, g total ethanol divided by the grams of sugar.

b 2N KOH consumed to maintain pH 6.0-6.5 during fermentation.

c Corrected for dilution by base.

d Ethanol Yield (g/g) divided by 0.51 X 100.

nature of the substrate, extensive hydrogen bonding of chains, etc., large amounts of cellulase protein are required (6). Enzymatic hydrolysis is further complicated by the accumulation of soluble products (glucose, cellobiose, cellotriose) which act as competitive inhibitors of hydrolysis. The problem of glucose inhibition was substantially reduced by development of the Gulf Simultaneous Saccharification and Fermentation (SSF) process in 1976 (33,34). In this process, fermentation with yeast and saccharification occur together within the same vessel. Cellobiose and cellotriose accumulation were prevented by supplying high levels of supplemental B-glucosidase, but at an additional cost. With minor modification such as the development of cellobiose-fermenting biocatalysts which do not require supplemental B-glucosidase (35,36), process optimization (37), and enzyme recycle (J8), a Modified Gulf SSF process remains the best available technology for cellulose conversion to ethanol.