For practical application of PHAs on commercial scale, consideration has to be given to the origin of carbon source used as well as the reducing the cost of production through the use of economical carbon sources. Carbon source such as 4- hydroxyvaleric acid derived from fossil fuel is often used as raw material for producing PHAs. However, it is known that the consumption of a large amount of fossil fuel results in the increase in CO2 concentration in atmosphere. Waste materials such as lignocellulose is desirable from the viewpoints of utilization of unexplored resources and solution of the green house effect. Xylose is one of main components of hemicellulose contained in wood waste. At present, few types of commercial products have been produced from xylose. Young et. al. reported that Pseudomonas cepacia could produce P(3HB) from xylose and lactose, but the productivity was very low(29). Lactococcus lactis 10-1 isolated in our laboratory is able to produce L-lactic acid and acetic acid at a high production rate from xylose (30,31). L-Lactate also has the potential to be used as raw material in the manufacture of a biodegradable plastic, poly(L-lactate). A. eutrophus cannot utilize xylose but utilize lactate as carbon source. We therefore developed a culture method for the production of PHA from xylose employing these two bacteria. This culture method consisted of an initial fermentative production of L-lactate from xylose employing L. lactis 10-1 and a conversion of L-lactate into PHA by A. eutrophus. Flask culture experiment showed that the growth rate of A. eutrophus decreased according to the increase in L-lactate concentration in the medium and the cells could not grow above 30 g*dnr3 of L-lactate. In pH-controlled batch fermentations, a maximum specific growth rate of 0.6 h_1 was obtained when 5 g»dnr3 of L-lactate was used. The growth of microorganisms is generally inhibited by the presence of lactate, however, the specific growth rate of A. eutrophus when using L-lactate was higher than when other types of carbon source were used. According to our study for A. eutrophus, the maximum specific growth rate with using fructose was about 0.2 h_1 and the maximum specific growth rate in autotrophic condition was 0.42 h_1. Such growth characteristic of A. eutrophus on L-lactate is favorable for production of P(3HB) by the two-stage method. The accumulation of P(3HB) by A. eutrophus was next investigated using the culture supernatant containing L-lactate and acetate converted from xylose by L lactis IO-1. A pH-controlled batch culture of L. lactis 10­1 was first anaerobically carried out using 30 g»dnr3 of xylose as carbon source. When xylose in the culture was completely consumed, the culture broth was aseptically centrifuged and the supernatant was returned to the fermenter. A. eutrophus cells was then inoculated and the second stage cultivation for P(3HB) accumulation was aerobically carried out. The initial L-lactate concentration in the second stage culture was adjusted to 10 g*dnr3. After 24 hours of cultivation, 8.5 g*dm~3 of cells were produced. The final percentage of P(3HB) in the cell reached 55 %(w/w) without nitrogen source on medium being limited (32). The growth of A. eutrophus is inhibited by high lactate concentrations, therefore high cell density cultivation can be achieved by pH-stat batch culture with substrate feeding to control

L-lactate concentration in medium at low level. As the C/N ratio for the consumption of L-lactate and ammonium by the cells was determined to be 10 (mol/mol) by a standard-type batch culture, the feed solution in which the C/N ratio was prepared to 10, was first used in the pH-stat batch culture as feed substrate. However, it was impossible to control L-lactate concentration at a constant level by using this feed solution. It was observed that the microorganism accumulated P(3HB) in the cell even during exponential growth phase and excreted a small amount of an unknown organic acid, then the acid-base equilibrium was not balanced in the culture system. The C/N ratio in the feed solution was, therefore, changed from 10 to 23.3(mol/mol) after 12 h of cultivation and phosphate and other organic nutrients were also supplied. As a result, cell concentration increased to 102 g»dnr3 (Fig. 8). The P(3HB) content in the cells reached about 60 %(w/w) although nitrogen source in culture medium was not limiting. We are now investigating substrate feeding strategy to increase P(3HB) accumulation.

1. Conclusion

The practical cultivation systems for hydrogen-oxidizing bacterium, A. eutrophus to produce a biodegradable plastic, P(3HB) from CO2 and xylose were developed and P(3HB) accumulation was improved by incorporating new strategies. The application of such culture systems should contribute to the solution of the global environmental pollution problems caused by increased CO2 level in atmosphere, disposal of non-degradable plastics and utilization of industrial waste materials. For practical application of biodegradable plastics, obviously considerable technological challenges must be overcome, especially in the reduction in production cost and improvement in extraction and refining process of the product. We are tackling this difficult problem and also investigating the conversion of various types of industrial waste materials to other useful compounds.

The PRODUCTION OF FUELS AND CHEMICALS from biomass faces signi­ficant technical and economic challenges at present. Its success depends largely on the development of environmentally friendly pretreatment pro­cedures, a highly effective multienzyme system for conversion of pre­treated biomass to fermentable sugars, and efficient microorganisms to ferment mixed sugar substrates to fuels and chemicals. It is timely to pro­vide a book that can guide further research and development in this area.

This volume was developed from a symposium presented at the 211th National Meeting of the American Chemical Society, titled “Fuels and Chemicals from Biomass”, sponsored by the ACS Biotechnology Secre­tariat and the Division of Biochemical Technology, in New Orleans, Louisiana, March 24-28, 1996. This book presents a compilation of ten manuscripts from that symposium plus nine solicited manuscripts that represent recent advances in the production of fuels and chemicals from biomass. The chapters in this book are organized in two sections: fuels (ethanol, biodiesel, and hydrogen) and chemicals (lactic acid, succinic acid, 1,3-propanediol, 2,3-butanediol, polyhydroxybutyrate, and xylitol). A chapter on synthesis-gas fermentation is included in the chemicals section.

We hope that this book will serve as a valuable interdisciplinary con­tribution to the continually expanding field of the production of fuels and chemicals from biomass.

Acknowledgments

We are fortunate to have contributions from many leading authorities in their respective disciplines. We would like to take this opportunity to express our gratitude to the contributing authors, the reviewers who pro­vided excellent comments to the editors, the ACS Biotechnology Secre­tariat and the Division of Biochemical Technology, and the ACS Books Department for making possible the symposium and the publication of this book. We also would like to acknowledge the Oak Ridge National Laboratory, which is managed by Lockheed Martin Energy Research Cor-

IX

poration for the U. S. Department of Energy under contract DE — AC05-960R22464.

badal C. Saha

Fermentation Biochemistry Research Unit

National Center for Agricultural Utilization Research

Agricultural Research Service

U. S. Department of Agriculture

1815 North University Street

Peoria, IL 61604

Jonathan Woodward Chemical Technology Division Oak Ridge National Laboratory Oak Ridge, TN 37831

January 24, 1997

x

The combination of simulation and bench-scale experiments may also be used to investigate the benefits of process integration regarding energy consumption. The integration can be performed internally within the ethanol plant or by integrating various parts of the ethanol plant with another type of plant, such as a pulp mill or a heat-generating plant. The latter requires, of course, detailed knowledge of the plant with which the ethanol plant is to be integrated. The internal integration will be exemplified by the incorporation of the distillation unit with a multiple-effect evaporation unit. Examples of integration of an ethanol plant and a pulp mill, as well as a power plant, will also be presented.

Internal Process Integration. The evaporation and distillation steps are large consumers of steam in the ethanol plant. The steam consumption in the evaporation and the distillation steps can be reduced by an increase in the dry matter content in the stillage or by increasing the ethanol concentration in the feed to the distillation step (60). This can be achieved by recirculation of various streams as described above. This might however, as shown, inhibit fermentation and/or hydrolysis depending on the degree of recirculation. Another way of increasing the ethanol content is to incorporate a stripper into the evaporation line, as shown in Figure 9. Instead of first distilling the fermentation broth and then evaporating the distillation stillage, the fermentation broth is fed to the stripper, evaporator effect 4, which is equipped with a reboiler, while the condensate from effect 5 is fed to the distillation unit.

In such case, the ethanol concentration in the feed to the distillation unit is increased from about 2-3% (wt/wt) to 18-20% (wt/wt), thus reducing the refining cost. With this configuration the steam consumption in the evaporation unit is not affected, while it is lowered by 60% in the distillation unit. For separate distillation and evaporation of a 100 tonnes/h liquid stream, the cost is 0.20 and 0.13 US$/kg product for dry matter contents in the feed of 4 and 6% (wt/wt), respectively (60). In both cases the liquid was concentrated to 65% (wt/wt) DM. With the incorporation of

Figure 7. COD and BOD of various evaporated fractions.

Figure 8. Recycling of the distillation feed. 1: Pretreatment; 2: Filtration; 3: Enzyme production; 4: Hydrolysis; 5: Fermentation; 6: Distillation.

a stripper into the evaporation unit, the cost decreases to 0.15 and 0.10 US$/kg product, as shown in Figure 10, illustrating the cost improvements when recycling is employed.

Further improvement in the economy can be achieved by integrating the energy­demanding steps in the entire process. One example is to release the high-pressure steam used in pretreatment to a back pressure of 3 bar and to use the secondary steam formed in other parts of the process, e. g. pre-steaming of the wood chips, in the distillation or in the evaporation unit (54).

External Process Integration. In an investigation by the Swedish engineering company AF-IPK (61), based on 150,000-250,000 tonnes/year raw material, the benefits of integrating an ethanol plant and existing processes handling biomass were evaluated. The plants investigated were a pulp mill, a power plant and a combined saw mill and peat-drying facility. The common infrastructure and the wood intake and storage are advantageous for all cases. The contribution of the common infrastructure is difficult to quantify in economic terms. The largest overall effects were obtained for a combination of an ethanol plant with either a pulp mill or a power plant. The saw mill has fewer possibilities for integration, since it is almost only the material handling that is common for the plants.

There are several opportunities for integrating redundant equipment or capacity in an ethanol plant and a pulp mill. The most interesting process step is that of steam generation as a pulp mill usually has an excess of low-pressure steam at 4 bar, which can be used in different process steps in the ethanol plant. The second most interesting process is effluent treatment. Due to the increased efforts in recent years to close the processes at pulp mills, the treatment plant will have surplus capacity which can be used by the ethanol plant. Furthermore, there is also the possibility of using a bark-fuelled boiler.

Assuming that the power plant is designed to receive large quantities of wood residue, the raw material handling, or part of it, could be shared with the ethanol plant. Furthermore, the boiler in the power plant can be used for steam generation in the ethanol plant. It is also possible to take advantage of the sorting of incoming wood to obtain a better raw material for ethanol production.

The potential for synergetic effects was determined to be in the range of 0.05­

0. 10 US$/kg ethanol, which should be compared with a calculated total production cost of 0.5-0.67 US$/kg ethanol for a plant with no integration. These benefits must be shared between the ethanol plant and the process with which it is integrated.

Rathin Datta and Shih-Perng Tsai

Waste Management and Bioengineering Section, Energy Systems
Division, Argonne National Laboratory, Argonne, EL 60439

Lactic acid has been an intermediate-volume specialty chemical (world production -50,000 tons/yr) used in a wide range of food processing and industrial applications. Lactic acid has the potential of becoming a very large volume, commodity-chemical intermediate produced from renewable carbohydrates for use as feedstocks for biodegradable polymers, oxygenated chemicals, environmentally friendly "green" solvents, plant growth regulators, and specialty chemical intermediates. The recent announcements of plant expansions and building of new development-scale plants for producing lactic acid and/or polymer intermediates by major U. S. companies, such as Cargill, Chronopol, A. E. Staley, and Archer Daniels Midland (ADM), attest to this potential. In the past, efficient and economical technologies for the recovery and purification of lactic acid from crude fermentation broths and the conversion of lactic acid to the chemical or polymer intermediates had been the key technology impediments and main process cost centers. The development and deployment of novel separations technologies, such as electrodialysis (ED) with bipolar membranes, extractive distillations integrated with fermentation, and chemical conversion, can enable low-cost production with continuous processes in large-scale operations. The use of bipolar ED can virtually eliminate the salt or gypsum waste produced in the current lactic acid processes. Thus, the emerging technologies can use environmentally sound processes to produce environmentally useful products from lactic acid. The process economics of some of these processes and products can also be quite attractive. In this paper, potential products and recent technical advances in lactic and polylactic acid processes are discussed. The technical accomplishments at Argonne National Laboratory (ANL) and the future directions of this program at ANL are discussed.

© 1997 American Chemical Society

Lactic acid (2-hydroxypropionic acid), CH3CHOHCOOH, is the most widely occurring hydroxycarboxylic acid. It was first discovered in 1780 by the Swedish chemist Scheele. Lactic acid is a naturally occurring organic acid that can be produced by fermentation or chemical synthesis. It is present in many foods, both naturally or as a product of in-situ microbial fermentation (as in sauerkraut, yogurt, buttermilk, sourdough breads, and many other fermented foods). Lactic acid is also a major metabolic intermediate in most living organisms — from anaerobic prokaryotes to human beings.

Although lactic acid has been ubiquitous in nature and has been produced by fermentation or chemical synthesis for over 50 years, it has not been a large-volume chemical. Its worldwide production volume by 1995 had grown to approximately 50 x 10 tons/yr with only a few major producers — CCA biochem b. v. of the Netherlands and its subsidiaries in Brazil and Spain, ADM in Decatur, Illinois, as the primary manufacturers. Sterling Chemicals, Inc., in Texas City, used to be a major producer but has recently announced the closure of its plant and exit from the business. Musashino in Japan has been a smaller manufacturer. CCA and ADM uses carbohydrate feedstocks and fermentation technology, while Sterling and Musashino use a chemical technology. Thus, lactic acid was considered a relatively mature fine chemical in that only its use in new applications (such as a monomer in plastics or as an intermediate in synthesis of high-volume oxygenated chemicals) would cause a significant increase in its anticipated demand (7).

For lactic acid to enter these applications, economical, efficient, and environmentally sound manufacturing processes are needed for its production. In the past, efficient and economical technologies for the recovery and purification of lactic acid from crude fermentation broths and conversion of lactic acid to the chemical or polymer intermediates had been the key technology impediments and main process cost centers. The development and deployment of novel separations technologies, such as electrodialysis (ED) with bipolar membranes, extractive distillations integrated with fermentation, and chemical conversion, can enable low-cost production with continuous processes in large-scale operations. The use of bipolar ED can virtually eliminate the salt or gypsum waste produced in the current lactic acid processes. Thus, the emerging technologies can use environmentally sound processes to produce environmentally useful products from lactic acid. Recent announcements of new lactic acid production plants by major chemical and agriprocessing companies may usher new technologies for the efficient, low-cost manufacture of lactic acid and its derivatives for new applications (2-5).

Hemicelluloses are heterogeneous polymers of pentoses (xylose and L-arabinose), hexoses (mannose) and sugar acids. Xylans, major hemicelluloses of many plant materials, are heteropolysaccharides with a homopolymeric backbone chain of 1,4-linked P-D-xylopyranose units. Besides xylose, xylans may contain L-arabinose, D-glucuronic acid or its 4-o-methyl ether, and acetic, p-coumaric, and ferulic acids.

The total hydrolysis of xylan requires endo P-l,4-xylanase (EC 3.2.1.8), P — xylosidase (EC 3.2.1.37), and several accessory enzyme activities such as a-L — arabinosidase (EC 3.2.1.55), a-glucuronidase (EC 3.2.1.1), acetyl xylan esterase (EC 3.2.1.6), feruloyl esterase and p-coumaroyl esterase which are necessary for hydrolyzing various substituted xylans. The endo-xylanase randomly attacks the main chains of xylans and P-xylosidase hydrolyzes xylooligosaccharides to xylose. The a-L — arabinosidase and a-glucuronidase remove the arabinose and 4-O-methyl glucuronic acid substituents, respectively from the xylan backbone. The esterases hydrolyze the ester linkages between xylose units of the xylan and acetic acid (acetyl xylan esterase) or between arabinose side chain residues and phenolic acids such as ferulic acid (feruloyl esterase) and p-coumaric acid (p-coumaroyl esterase). Synergistic action of depolymerizing and side-group cleaving enzymes has been proved using acetylated xylan as substrate (37). Bachmann and McCarthy (38) reported significant synergistic interaction between endo-xylanase, P-xylosidase, a-L-arabinofuranosidase, and acetyl xylan esterase enzymes of the thermophilic actinomycete Thermomonospora fusca. Many xylanases do not cleave glycosidic bonds between xylose units which are substituted. The side chains must be cleaved before the xylan backbone can be completely hydrolyzed (39). On the other hand, several accessory enzymes only remove side-chains from xylooligosaccharides. These enzymes require xylanases to hydrolyze hemicellulose partially before side-chains can be cleaved (40). Although the structure of xylan is more complex than cellulose and requires several different enzymes with different specificities for a complete hydrolysis, the polysaccharide does not form tightly packed crystalline structures and is thus more accessible to enzymatic hydrolysis (41). The yeast-like fungus Aureobasidium is a promising source of xylanase (MW 20 kDa) with an exceptionally high specific activity (2100 U/mg protein) (42). Xylanase represented nearly half the total extracellular protein, with a yield of up to 0.3 g of xylanase per liter (43). A few recent reviews have dealt with the multiplicity, structure and function of microbial xylanases, and molecular biology of xylan degradation (5, 44, 45).

The utilization of hemicellulosic sugars is essential for efficient conversion of lignocellulose to ethanol. The commercial exploitation of the pentose fermenting yeasts for ethanol production from xylose is restricted mainly by their low ethanol tolerance, slow rates of fermentation, difficulty to control the rate of oxygen supply at the optimal level plus sensitivity to microbial inhibitors, particularly those liberated during pretreatment and hydrolysis of lignocellulosic substrates (46, 47). Xylose can also be converted to xylulose using the enzyme xylose isomerase and traditional yeasts can ferment xylulose to ethanol (48, 49). Xylose can be easily converted into xylitol, a potentially attractive sweetening agent by a variety of microorganisms (yeasts, fungi and bacteria) (50).

At the initiation of the experiment the amount of fresh substrate to be delivered per unit weight of alkali fed (a) was set on the computer. In the initial series of experiments, a was varied to determine its effect on the dilution rate and residual glucose concentration at steady state. The software also incorporated a short program for signal conditioning or smoothing, using a moving average calculation method, to filter out noise generated in the electrical circuits contained in the
digital weighting balances. Continuous plot of the weight of alkali and fresh substrate reservoirs were displayed on the computer screen during the course of the experiment.

Analytical Methods: Ethanol concentration was estimated by gas chromatography (GC-8 APE; Shimadzu Corporation, Kyoto Japan) with PEG 80-100 mesh. The temperature of the column oven and the injector were 70°C and 90°C, respectively. Glucose concentration was determined with a glucose analyzer (Model 23 A, Yellow Spring Instrument Co. Ltd., Ohio USA). Biomass was estimated using correlation between optical density measurements at 562 nm using a spectrophotometer (Uvidec 320 AS Co. Ltd., Tokyo Japan) and dry cell weights.

Biebl (4) studied the inhibition potentials of 1,3-PD, acetic acid, butyric acid and glyce­rol on the growth of C. butyricum in a pH-auxostat culture and found that all these sub­stances are toxic to C. butyricum. The critical concentrations of these substances, i. e., concentrations above which cells cease to grow, were found to be: 27 g/1 for acetic acid (0.49 g undissociated acetic acid/1); 19 g/1 for butyric acid (0.39 g undissociated acid/1), 64.0 g/1 for 1,3-PD, and 97.0 g/1 for glycerol at pH 6.5 in this study with externally ad­ded substances. Zeng et al. (34) further examined the inhibition potentials of substrate and products both self-produced and externally added on C. butyricum and K. pneumo­
niae with the help of a mathematical model. Whereas the inhibition potential of exter­nally added and self-produced 1,3-propanediol reveals to be essentially the same, buty­ric acid produced by the culture is found to be more toxic than that externally added. The same seems to apply for acetic acid. Furthermore, the inhibitory effect of butyric acid is shown to be due to the total concentration instead of its undissociated form.

The inhibition effects of products and substrate in the glycerol fermentation are essentially irrespective of the strains. Thus, a common growth model could be proposed for the two strains anaerobically grown on glycerol at different pH value:

where p is the specific growth rate; H+ is the hydrogen ion concentration; p*max, KH, and KqH are constants; K* is the saturation constant; , C*hbu, C*Et0H> C*PD , C*GIy

are the critical concentrations of acetic acid, butryic acid, ethanol, 1,3-PD and glycerol respectively. In the above model the parameters p*max, KH, and KoH are strain-spe­cific and can be estimated from experimental data for each strain (34). The critical con­centrations of acetic acid, 1,3-PD and glycerol are assumed to be identical for both C butyricum and K. pneumoniae. The best estimates of the critical concentrations (C*pi) are as follows: 0.35 g/1 for undissociated acetic acid, 10.1 g/1 for total butyric acid, 16.6 g/1 for ethanol, 71.4 g/1 for 1,3-PD, and 187.6 g/1 for glycerol. Eq. (13) describes the product and substrate inhibition of both C. butyricum and K. pneumoniae in different types of continous cultures over a pH range of 5.3-8.5 satisfactorily.

Zeng and Deckwer (55) and Zeng (36) studied also the kinetic of substrate con­sumption and product formation of K. pneumoniae. They found that in order to achieve a high substrate uptake rate and a high production rate of 1,3-PD a certain level of gly­cerol excess in the culture is necessary. The dependence of increase of substrate uptake rate and product formation rate on the excess concentration of glycerol can be described with a saturation function similar to that of Michaelis-Menten kinetics.

The maximum achievable concentration of 1,3-PD in the continuous fermenta­tion of glycerol can be predicted from considerations of growth and product formation stoichiometry and product inhibition (Eq.13). As discussed above, the maximum 1,3- PD production would be obtained in both strains if only 1,3-PD and acetic acid are for­med as fermentation products and at the same time there is no formation of hydrogen. Under these conditions the mole ratio of acetic acid to 1,3-PD should be 0.31 mol/mol. At pH = 7 and a residual glycerol concentration of 2-100 mmol/1 (C*GIy » CGiy » Ks) eq. (13) reduces to

^maxd-^Xl-S5) (И)

CHAc CPD

Substituting Сндс by 0.31 CPD Eq. (14) can be used to predict the maximum achievalbe

1,3- PD concentration and productivity under different dilution rate. Since all the para­meters of Eq. 14 are the same for C. butyricum and K. pneumoniae at pH 7.0 (pmax =

0. 67 h’1) the predicted theoretical maximum product concentration and productivity are applied to both strains. Fig. 4 shows the theoretical maximum propanediol concentra­tion and experimental maximum values achieved so far at different dilution rate. At high dilution rate (D > ca. 0.35 h1) nearly the same level of 1,3-PD has been obtained for C. butyricum and K. pneumoniae, being about 70 — 80% of the theoretical maxi­mum. Whereas at low dilution rate (< 0.35 h’1) K. pneumoniae reached 80 — 96% of the theoretical maximum C. butyricum reached only about 40 — 60%. At the lowest dilution rate (0.1 h’1) where experimental data are available for both strains C. butyricum rea­ched less than half of the propanediol concentration and productiviy of K. pneumoniae. Obviously, the maximum values measured for C. butyricum do not necessarily repre­sent the real maximum values achievable at the respective dilution rate. As shown for

K. pneumoniae (Menzel, K.; Zeng, A.-P.; Deckwer, W.-D. Enzyme Microbiol Technol. in press) there exists an optimum residual glycerol concentration for propanediol forma­tion at each dilution rate. So far, no systematical work has been carried out to find the optimum propanediol production at each dilution rate for C. butyricum. Another reason for the lower propanediol production by C. butyricum seems to be due to the less favor­able distribution of the by-products. At lower dilution rates the acetic acid production of C butyricum is far below the acetate level for optimal 13-PD formation. As mentioned above, for an optimal production of 1,3-PD the formation of butyric acid in C. butyri­cum and ethanol in K. pneumoniae should be as low as possible. It can be shown that the butyric acid concentration in C. butyricum culture is often higher than the ethanol concentration in the K. pneumoniae culture. However, butyric acid is nearly twice as toxic as ethanol (34). Thus, 1,3-PD concentration is limited in the C butyricum culture.

In addition, the hydrogen production in C. butyricum appeared to be higher than in К pneumoniae under substrate excess conditions. This also impedes the production of 1,3- PD in C. butyricum. Recently, Barbirato et al. (3) found that the accumulation of 3- hydroxypropionalde-hyde strongly inhibits glycerol fermentation of three enterobacteri­al species including К pneumoniae. Cameron et al. (10) showed that glycerol metabo­lism in E. coli is inhibited by the intracellular accumulation of glycerol-3-phosphate. The identification of possible inhibitory effects of intermediate metabolite(s) in C. buty­ricum is desirable.

Fermentation Studies on Glucose and Xylose. The fermentation performance of the parent yeast and the genetically engineered yeast was studied on glucose and xylose separately. In both these experiments the substrate concentrations selected were well above the growth limiting concentrations and not inhibitory to growth of the yeasts.

The results of the glucose fermentation study using yeasts 1400 and 1400 (pLNH33) is shown in Figures la and lb, respectively. These experiments were performed using approximately 100 g/L glucose as the substrate with an initial cell density in the range of 0.05-0.10 g/L. In both cases, a maximum cell density in the range of 11-11.5 g/1 was obtained after 14 hours. The specific growth rates of the yeast was calculated from the slope of a semi-log plot of cell dry weight versus time. For the yeast 1400, the specific growth rate under the experimental conditions was calculated to be 0.4910.02 hr’1 while for the recombinant yeast 1400 (pLNH33), the specific growth rate was calculated to be 0.4810.02 hr’1 . Thus the specific growth rates of both yeasts on glucose are identical. The specific glucose utilization rates and ethanol yields in both cases were 2.3 g/g-hr and 0.49, respectively. Based on a theoretical yield of 0.51 g/g, these yields correspond to 96 % of the theoretical yield.

The results of the xylose fermentation study using the above yeasts are shown in Figures 2a and 2b. These experiments were performed using approximately 50 g/L xylose as the substrate with an initial cell density of 0.05-0.10 g/L. As seen in Figure 2a, the yeast 1400 is unable to grow and ferment xylose to ethanol. In comparison, the recombinant yeast 1400 (pLNH33) grows to a cell density of 9.4 g/L, and produces 20.45 g/L ethanol from 52.06 g/L initial xylose after 36 hours. Glycerol and xylitol are the by-products that are produced in minor amounts, to the extent of 1.73 g/L and 2.85 g/L respectively. The specific growth rate of the recombinant yeast was calculated, as explained earlier, to be 0.19±0.02 hr"1. The specific xylose utilization rate was calculated to be 0.30 g/g-hr. Based on the xylose consumed, the ethanol yield was calculated to be 0.40 g/g. This corresponds to 78 % of the theoretical yield (based on a theoretical yield of 0.51 g/g). These results indicate that the specific xylose utilization rate and ethanol yield from xylose are lower than those from glucose.

Fermentation Study on a Glucose-Xylose Mixture. From the above studies on single substrates, it is clear that the fusion product 1400 lacks the ability to ferment xylose. On the other hand, the recombinant yeast 1400 (pLNH33) shows a good fermentation performance on xylose. Typically, both glucose and xylose are present in lignocellulosic hydrolysates. This requires that the fermentative microorganism be able to ferment xylose in presence of glucose. The recombinant yeast 1400 (pLNH33) has been genetically designed to ferment both glucose and xylose present in the same medium (9). To demonstrate this, fermentation of a 1:1 mixture of glucose and xylose was performed. The composition of the sugar mixture was 52.8 g/L glucose and 56.3 g/L xylose and the initial cell density was 2.3 g/L. As shown in Figure 3, the recombinant yeast ferments glucose and xylose simultaneously to ethanol. However, the glucose utilization rate at 8.11 g/L-hr is relatively higher than the xylose utilization rate at 1.77 g/L-hr. A final cell density of 11.5 g/L and an ethanol concentration of 47.9 g/L were achieved after 48 hours. Based on a theoretical yield of 0.51 g/g, the ethanol yield from this experiment was calculated to be 80%.

й І В

Figure 1. Growth and fermentation performance of a) yeast strain 1400 and b) recombinant yeast strain 1400 (pLNH33) on glucose.

Time (hours)

Figure 2. Growth and fermentation performance of a) yeast strain 1400 and b) recombinant yeast strain 1400 (pLNH33) on xylose.

Figure 3. Fermentation of a glucose-xylose mixture by recombinant yeast strain 1400 (pLNH33).

Fermentation Kinetics. The effects of substrate and product inhibition on glucose and xylose fermentation were studied in order to determine the optimal fermentation conditions for achieving highest ethanol yields and productivities.

Substrate Inhibition. These experiments were conducted by using different initial glucose or xylose concentrations in the fermentation media. For the glucose fermentation, significant lag times are observed in experiments with an initial concentration greater than 200 g/L (72). This can be attributed to the effect of glucose inhibition on the yeast growth, as all other conditions such as temperature and availability of nutrients are favorable. Similar studies have also been conducted for the xylose fermentation using recombinant yeast 1400 (pLNH33). These studies indicate that the effect of xylose inhibition for yeast growth is significant above 70 g/L (data on specific growth rates not included).

A classical process methodology for concentrated sugar fermentations is to feed the substrate into the fermentation medium in steps, instead of feeding all of it initially (77,72). To achieve higher ethanol concentrations from glucose fermentation, this modified feeding scheme was used as shown in Figure 4. The fermentation was begun with 106 g/L glucose initially, followed by four successive additions of glucose with nutrients later during the fermentation. The arrows in the figure indicate the times at which glucose with nutrients was added into the fermentation medium. The addition of nutrients along with glucose helped in maintaining a high cell viability throughout the fermentation. At the end of 67 hours, an ethanol concentration of 133 g/L was obtained at a productivity of 1.99 g/L-hr. This reflects the high ethanol tolerance of the Saccharomyces strain 1400. Since ethanol tolerance significantly influences the economics of ethanol recovery, this high ethanol tolerance yeast has potential applications in commercial ethanol production.

This fed-batch experimental methodology was also used for the xylose fermentation using the recombinant yeast 1400 (pLNH33), as shown in Figures 5a and 5b. An ethanol concentration of 50.34 g/L was obtained after 193 hours of fermentation, giving a productivity of 0.26 g/L-hr. Thus in comparison to the glucose fed-batch fermentation, relatively lower ethanol concentrations and productivities were achieved. Ethanol inhibition of the xylose fermentation may be one of the several reasons possible for this observation.

Product Inhibition. In order to determine the effect of ethanol on the cell growth rates and the fermentation rate, experiments with a range of initial ethanol concentrations in the fermentation media were performed. This study was carried using both glucose and xylose as substrates. The sugar concentrations in both experiments were initially 50 g/L, which is well in excess of the saturation constant. This sugar concentration was selected in order to separate ethanol inhibition effects from those of substrate or nutrient limitation. The effects of ethanol concentration as a single independent variable can be clearly discerned using this method as compared to batch studies with produced ethanol.

Figure 6 shows the experimental data of variation in specific growth rate of yeast 1400 on glucose as a function of the initial ethanol concentration. When there is no initial ethanol in the medium, the highest specific growth rate of 0.6 hr’1 is

Time (hours)

Figure 5. a) Xylose and b) ethanol concentration profiles during a xylose fed — batch fermentation using the recombinant yeast strain 1400 (pLNH33).

Figure 7. Effect of ethanol on specific growth rate of recombinant yeast strain 1400 (pLNH33) on xylose.

obtained. There is a decline in the specific growth rate as the initial ethanol increases. The effect of ethanol inhibition becomes significant beyond a concentration of -100 g/L. The cell growth stops completely above an ethanol concentration of 136 g/L.

The above experiment was also performed with the recombinant yeast using xylose as the fermentation substrate. The slower growth rate on xylose is evident from the highest specific growth rate of 0.19 hr’1, which is about 3 times lower than the highest specific growth rate on an equal concentration of glucose. The inhibitory effect of ethanol on the cell growth on xylose is also stronger, in comparison to cell growth on glucose. The cell growth on xylose is inhibited strongly beyond an ethanol concentration of 60 g/L, as seen in Figure 7. However, the genetically engineered yeast has been designed to overcome this ethanol inhibition by being able to use glucose for growth, and then cofermenting the rest of the glucose and xylose to ethanol.

Fermentation Performance of the Recombinant Saccharomyces 1400 (pLNH33) on Lignocellulosic Hydrolyzates. The results which have been discussed above were obtained from studies conducted on pure sugars. These studies are useful, as they provide insight into the fermentation kinetics. In order to study the fermentation performance of this recombinant yeast on “real” substrates, some experiments conducted using com fiber and com cob as model feedstocks are presented.

The pretreatment of the lignocellulosic material is an important process step to achieve higher ethanol yields. Various physical and chemical pretreatment methods have been reported in the literature. These include physical treatments, chemical treatment using strong acids and bases, steam explosion and the low temperature ammonia fiber explosion (AFEX) process (13). Pretreatment studies in our laboratory have resulted in the development of two pretreatment techniques: dilute acid hydrolysis and ammonia steeping followed by dilute acid hydrolysis. Pretreatment of the com fiber was accomplished by acid hydrolysis using 0.5% HC1. The com cob which contains a higher fraction of lignin was pretreated by the ammonia steeping process followed by the dilute acid hydrolysis.

Corn Fiber Studies. Com based starch that is mainly obtained by com wet milling is a predominant feedstock for fuel ethanol production. A low value side stream called “com fiber” is generated in the com wet milling process. This stream contains the hulls, fine fibers and residual starch granules from washing the starch. About 9 to 10% (w/v) of the original dry weight of the com is recovered in the com fiber stream. Thus com fiber is an attractive starting material for ethanol production, in view of its availability, limited use and low cost.

The typical composition of com fiber used in these studies is as follows: 25% starch, 25% hemicellulose and 15% cellulose. Pretreatment of com fiber was accomplished by hydrolysis using 0.5% (w/v) HC1 at 120°C for 45 minutes. This pretreatment by dilute acid hydrolysis readily hydrolyses the starch to glucose, and hemicellulose to primarily xylose. The cellulosic residue is then treated with the cellulase enzyme to hydrolyze it to glucose.

In the simultaneous saccharification and fermentation (SSF) studies, the recombinant yeast 1400 (pLNH33) and the Iogen cellulase are added together to the medium containing the pretreated com fiber. Figure 8 shows the result of a batch SSF process using 10% w/v dry pretreated com fiber with 0.46 FPU/ml (about 6 FPU per gram of cellulose from com fiber) Iogen cellulase enzyme in 50 ml medium. As seen from the substrate profiles, there is a simultaneous utilization of glucose and xylose by the recombinant yeast. An ethanol concentration of 25.1 g/L was obtained at the end of 72 hours of SSF. Based on 15 g/L of glucose released from cellulose, 24.6 g/L glucose and 16 g/L xylose from the pretreated com fiber, the total fermentable sugar is 55.6 g/L. The ethanol yield obtained in this experiment was 0.45 g/g, that corresponds to 88 % of the theoretical yield at 0.51 g/g.

The fed-batch SSF process is an effective process methodology for achieving high ethanol productivities and reducing enzyme costs by lower cellulase loadings. Figure 9a shows a fed-batch SSF process starting with 10% (w/v) pretreated com fiber, followed by an identical feeding to give a total of 20% (w/v) solids. Ethanol concentrations of 38.9 g/L and 41 g/L were obtained after 72 and 96 hours, respectively. Instead of a single feeding, multiple feeds can also be used in the fed — batch SSF process. Figure 9b shows the result of such an experiment, in which the SSF is started with 10% (w/v) pretreated com fiber as before, followed by two identical feedings at 12 hour intervals to give a total of 30% (w/v) solids. Ethanol concentrations of 44.4 g/L and 48.7 g/L were obtained after 72 and 96 hours, respectively. Based on these results, the ethanol productivities during SSF of pretreated com fiber lie in the range of 0.44-0.62 g/L-hr.

Corn Cob Studies. Com cob is a low value agricultural residue having limited use. A typical composition of com cob used in these studies is as follows: 41.1% cellulose, 36% xylan, 6.8% lignin and 3.2% acetate. Compared to com fiber, the percentage of xylose in sugars obtained from com cob will be typically higher. This requires the use of a xylose fermenting microorganism for effectively fermenting xylose to ethanol.

Recently, an effective pretreatment process for lignocellulosic biomass has been developed (Cao, N. J., personal communication). This process involves steeping the raw material in 10% ammonium hydroxide solution at ambient temperatures for 24 hours. It has been determined that this ammonia steeping process efficiently removes lignin, acetate and extractives. Following this treatment, a dilute acid hydrolysis is used to hydrolyze the hemicellulose fraction. The residual fraction is primarily comprised of cellulose, that can be subjected to enzymatic hydrolysis. This process methodology systematically separates lignin, hemicellulose, cellulose and enables separate processing of each fraction. Since the lignin is removed in the initial stages of the process, the adsorption of cellulase on the lignin is minimized. Thus this pretreatment method also allows lower cellulase loadings in the SSF process. By coupling this pretreatment method with the use of the recombinant Saccharomyces 1400 (pLNH33) in the fermentation, promising results have been obtained (Cao, N. J., et al. Biotechnol. Lett., in press).

The ammonia steeping pretreatment removes almost all the acetate from the raw material. This is a key step in the process, as acetic acid has been determined to be inhibitory for xylose fermentation (14). As discussed earlier, a dilute HC1 pretreatment hydrolyses the hemicellulose and a xylose rich hydrolyzate is obtained,

Figure 8. Simultaneous saccharification and fermentation (SSF) of 10% (w/v) pretreated com fiber using the recombinant yeast strain 1400 (pLNH33) and Iogen cellulase.

Glucose Xylose Ethanol —-e—

Figure 9a. Single stage fed-batch SSF process of pretreated corn fiber using the recombinant yeast strain 1400 (pLNH33) and Iogen cellulase.

Figure 9b. Two stage fed-batch SSF process of pretreated com fiber using the recombinant yeast strain 1400 (pLNH33) and Iogen cellulase.

that is treated with a weak base anion exchange resin. Iogen cellulase was used to hydrolyze the cellulosic residue, which results in a glucose rich hydrolyzate. Both the hemicellulose and cellulose hydrolyzates were mixed and fermented after adjusting the pH to 5. Figure 10 shows the result of a batch fermentation of this mixed hydrolyzate containing 52.8 g/L glucose and 55.7 g/L xylose using the recombinant yeast. An ethanol concentration of 46.9 g/L was obtained within 36 hours, giving a high yield of 84% (based on a theoretical yield of 0.51 g/g).

Results of the SSF process of cellulosic residue (pretreated by different methods) obtained from 20 g com cob using the Iogen cellulase and the recombinant

Figure 11. Effectiveness of different pretreatment methods for the SSF of cellulosic residue from com cob using the recombinant yeast strain 1400 (pLNH33) and Iogen cellulase.

yeast 1400 (pLNH33) at 35°C are shown in Figure 11. The best results were obtained with the pretreatment involving ammonia steeping followed by dilute acid hydrolysis. An ethanol concentration of 40.7 g/L was obtained after 48 hours with a yield of 86% based on dry cellulose from com cob. Application of the pretreatment method allowed a low cellulase loading (8.5 IFPU/g com cob) to be used in the SSF process.

Conclusions

Fermentation studies using the recombinant yeast 1400 (pLNH33) show good results with high ethanol yields achieved on glucose, xylose and their mixtures. The simultaneous sugar utilization pattern exhibited by this yeast is beneficial for the SSF process, as the xylose fermentation time is reduced significantly. Moreover, this recombinant yeast retains the high ethanol tolerance of its parent (the fusion yeast strain 1400). This improves the economics of ethanol recovery greatly, as high final ethanol concentrations can be achieved in the fermentation broth.

Promising results with the yeast 1400 (pLNH33) have also been obtained on lignocellulosic feedstocks like com fiber and com cob. Significant process improvements have also been made in the biomass pretreatment technology. The use of this recombinant yeast coupled with improved pretreatment techniques provides a firm base for developing a mature biomass to ethanol technology.

Acknowledgments

This work was funded by Amoco Corporation and the Consortium for Plant Biotechnology Research. The com cob studies were conducted by M. S. Krishnan with N. J. Cao and C. S. Gong. We acknowledge the assistance of Jean Lu and Xuezhi Yu for the HPLC analysis.

Early 100 hr tests on transesterified sunflower oil initially showed the improved properties of esters for use in a diesel engine by reducing the viscosity of vegetable oils and solving engine problems (29-31).

Table IV compares the essential fuel properties of some esters. In all cases the viscosity decreases dramatically and is only about twice that of DF2. The CNs are also improved, now being higher than that of DF2.

The methyl and ethyl esters of soybean oil generally compared well with DF2 with the exception of gum formation which leads to problems with fuel filter plugging (20). Another study reports that methyl esters of rapeseed and high-linoleic safflower oils formed equal and lesser amounts of deposits than a DF standard while the methyl ester of high-oleic safflower oil formed more deposits (55). Methyl and ethyl esters of soybean oil were evaluated by 200 hr EMA (Engine Manufacturers Association) engine tests and compared to DF2. Engine performance with soybean esters differed little from that with DF. In that work, also a slight power loss was observed, together with an increase in fuel consumption due to the lower heating values of the esters. The emissions for the two fuels were similar, with the exception of NOx which are higher for the esters as discussed above. Engine wear and fuel-injection system tests showed no abnormal characteristics for any of the fuels. Deposit amounts in the engine were comparable, however, the methyl ester showed greater varnish and carbon deposits on the pistons. Operating DI engines with neat soybean oil esters under certain conditions produced lubricating oil dilution which was not observed with an IDI engine (109). Lubricating oil dilution was estimated by Fourier-Transform infrared spectroscopy combined with a fiber optic probe when using rapeseed methyl ester as a fuel (110). The carbonyl absorption was used for quantitation.

Low-temperature Properties. As discussed above, one of the major problems associated with the use of biodiesel, including methyl esters, is its poor low-temperature properties, documented by relatively high cloud point (CP) and pour point (PP) (Tables III and IV). CPs and PPs of vegetable oils and their esters are a result of these materials being mixtures of various compounds. For example, as seen in Table I, saturated fatty compounds have significantly higher melting points than unsaturated fatty compounds and in a mixture they therefore crystallize at higher temperature than the unsaturates.

The CP, which occurs at a higher temperature than the PP, is the temperature at which a fatty material becomes cloudy due to formation of crystals and saturates solidifying. These solids can clog fuel lines. With decreasing temperature, more material solidifies and the compound approaches the pour point, at which it will no longer flow.

Besides CP (ASTM D2500) and PP (ASTM D97), two test methods exist for examining the low-temperature properties of diesel fuel (as discussed briefly in the section on “Biodiesel Standards”), the Low-Temperature Flow Test (LTFT; used in North America; ASTM D4539) and the Cold-Filter Plugging Point (CFPP; used in Europe). These methods have also been used to evaluate biodiesel. Low-temperature filterability tests are necessary because they correlate better with operability tests than CP or PP (111). Recent results showed that for fuel formulations containing at least 10% methyl esters, both LTFT and CFPP are linear functions of the CP (112). Additional statistical analysis showed a strong 1:1 correlation between LTFT and CP (112).

Five possible solutions to the low-temperature problems of esters have been investigated: blending with conventional DF, additives, branched-chain esters, bulky substituents in the chain, and winterization. Blending of esters is currently the preferred method for improving low-temperature properties and is discussed in the next section.

Numerous additives have been synthesized and reported mainly in the patent literature, which allegedly have the effect of lowering CP and PP. These additives are usually a variety of viscosity-modifying polymers such as carboxy-containing interpolymers (113), styrene-maleic anhydride copolymer (114), polymethacrylates (114-115), polyacrylates (113-115), nitrogen-containing polyacrylates (113), poly[alkyl (meth)acrylates] (116), ethylene-vinyl ester (acetate) copolymers (117-120), fumarate or itaconate polymers and copolymers (comb polymers) (117-118), polyoxyalkylene compounds (113). Polar nitrogen compounds (117) have also been reported as additives. Similar additives have also been tested for conventional diesel fuel (7). The beneficial effect of some additives appears to be limited, however, because they more strongly affect the PP than the CP, and the CP is more important than the PP for improving low-temperature flow properties (121).

Another route is the synthesis of fatty compound-derived materials with bulky substituents in the chain (122). The idea associated with these materials is that the bulky substituents would destroy the harmony of the solids which are usually oriented in one direction. However, these materials had only slight influence on the CP and PP.

The use of secondary alcohols in the transesterification reaction provides branched — chain esters such as isopropyl and 2-butyl instead of the methyl esters (95, 123) . These esters showed a lower crystallization onset temperature (Tco) as determined by differential scanning calorimetry (DSC) for the isopropyl esters of SBO by 7-11 °С and for the 2-butyl esters of SBO by 12-14°C (123). The CPs and PPs were also lowered by the branched-chain esters. However, economics would only permit the isopropyl soyate appear attractive as branched-chain ester, raising the price of a biodiesel blend containing 30% isopropyl soyate by $0.02/L while lowering the Tco by 15 °С.

In the winterization method (121, 124), solids formed during cooling of the vegetable oil are removed by filtration, thus leaving a mixture of more unsaturated fatty compounds with lower CP and PP. This procedure can be repeated to achieve the desired CPs and PPs. Saturated fatty compounds, which have higher CNs, are among the major compounds removed by winterization. Thus the CN of the biodiesel decreases. The Tco of typical methyl soyate was lowered from 3.7°C to -7.1 °С by winterization (124), but the yield was low (26%). Winterization of low-palmitate methyl soyate, however, gave a Tco of -6.5 °С with a yield of 86%. Winterization of typical methyl soyate diluted in hexane gave a Tco of -5.8°С with 77% yield. In the latter method, crystal formation was greatly affected by the nature of the solvent, with acetone and chloroform being unsuitable for winterization.

In a paper on fatty acid derivatives for improving ignition and low-temperature properties (125), it was reported that tertiary fatty amines and amides were effective in enhancing the ignition quality of biodiesel without negatively affecting the low — temperature properties. In that paper, saturated fatty alcohols of chain lengths C,2 and greater increased the PP substantially. Ethyl laurate was weakly decreased the PP.

Badal C. Saha and Rodney J. Bothast

Fermentation Biochemistry Research Unit, National Center
for Agricultural Utilization Research, Agricultural Research Service,
U. S. Department of Agriculture, 1815 North University Street,
Peoria, IL 61604

Xylitol, a five-carbon polyalcohol, has attracted much attention because of its potential use as a natural food sweetener, as a dental caries reducer and as a sugar substitute in diets for diabetics. Currently, it is produced chemically by catalytic reduction of xylose. Various microorganisms can convert xylose to xylitol. The present review describes microbial production of xylitol from xylose and xylose rich hemicellulose fractions present in various lignocellulosic biomass.

Xylitol, a pentitol of xylose, has attracted much attention because of its potential use as a natural food sweetener, as a dental caries reducer and as a sugar substitute for treatment of diabetics (7). It is a normal intermediary product of carbohydrate metabolism in humans and animals. The human body produces 5-15 g of xylitol a day during a normal metabolism (2). Xylitol is widely distributed in the plant kingdom, especially, in certain fruits and vegetables (7, 5, 4). However, extracting it from these sources is impractical because it is generally present in small quantities. Xylitol is currently produced chemically by catalytic reduction of xylose present in hemicellulose (xylan) hydrolyzate in alkaline conditions (5, 6). The recovery of xylitol from the xylan fraction reaches about 50-60% (4). Drawbacks of the chemical process are the requirements of high pressure and temperature, use of an expensive catalyst and use of extensive separation and purification steps to remove the by-products mainly derived from hemicellulose hydrolyzate (7). The bulk of xylitol produced is consumed in various food products such as chewing gum, candy, soft drinks and ice cream (2).