Modem industrial processes are often very complex and it is almost impossible to evaluate every alternative by experiment alone. The process steps nearly always depend on each other, and a change in operating conditions in one step affects the performance of other units. Computer simulations can be used to identify problem areas which require further exploration. However, computer simulations can never replace experiments, but constitute an important tool when planning and evaluating laboratory or pilot-plant experiments.

Computer simulation of industrial processes requires mathematical models that are able to predict, for example, reaction rates, yields, vapour-liquid equilibria and energy requirements. The simulation can be performed either with a tailor-made program for the process of interest, or by using a general flowsheeting program, such as Aspen Plus (Aspen Tech, USA), ChemCad (Chemstations Inc., USA), or Process (Simulation Sciences Inc., USA). Another possibility is to use spreadsheet programs such as Excel or Lotus to create the simulation models. Most process simulators are not capable of handling complex solid and heterogeneous materials such as wood or yeast. These materials cannot be characterised in terms of well-defined physical properties, e. g. molecular weight or enthalpies. However, these problems are encountered when simulating the production of ethanol, and the components must be characterised in terms of average physical properties for the material.

We have chosen to use the commercial process simulator Aspen Plus to perform simulations of different process configurations. Aspen Plus is capable of handling heterogeneous and solid materials, and it is also possible to incorporate user-defined components and unit operation modules into the program. When defining user modules, results from experimental investigations, such as laboratory-scale or bench — scale trials, are fitted to empirical or mechanistic models and included in the process simulation.

Cost estimations are also important to determine the economic feasibility of a process. It is valuable to obtain the distribution of costs within a process to identify high-cost steps so that research and development efforts can be directed towards the most expensive process steps. By means of sensitivity analysis it is possible to predict how changes in a specific parameter influence the yield and the economy of the overall process. A simulation program, BioEconomics, developed by von Sivers and Zacchi (48) is used to estimate the production cost for ethanol from lignocellulosic materials.

In the following, three examples of how techno-economic simulations combined with experimental investigations in the bench-scale unit can be used for process development are presented.

The discovery of photosynthetic H2 production is based on the classic work of Gaffron and Rubin in 1942 (12). However, only since 1973, at the time of the energy crisis, has photosynthetic H2 production been investigated as a potential source of energy (13-41). Sustained photoevolution of H2 and 02 by microalgae was first demonstrated by Greenbaum in 1980 (15). Under anaerobic conditions, sustained photoevolution of H2 and 02 in microalgae can be readily demonstrated using a reactor-flow-detection system (Figure 2). Figure 3 presents a typical measurement of H2 and 02 production in Chlamydomonas in a helium atmosphere using the reactor-flow-detection system. The data clearly demonstrate that photoevolution of H2 and 02 can occur stably with a stoichiometric ratio of H2 to 02 of nearly 2:1 as expected for water splitting. Photoevolution of H2 and 02 can be sustained for weeks.

The advantage of the simultaneous H2 and 02 photoevolution is that it can potentially have a high energy conversion efficiency since electrons energized by the light reactions are used directly in the reduction of protons to produce H2 by the Fd/hydrogenase pathway (Figure 1). However, H2 production by this mechanism requires gas product separation since both H2 and 02 are produced simultaneously in the same volume. Furthermore, until an 02-insensitive hydrogenase is developed (42, 43), the 02 concentration in the algal suspension has to be kept low to maintain H2 production since the hydrogenase is sensitive to 02. Therefore, an efficient and inexpensive technique to separate and remove gas products is needed.

Another important aspect is that the photoevolution of H2 and 02 in microalgae is often saturated at a relatively low actinic intensity. This is probably due to three factors: (1) accumulation of a back-proton gradient because of the limited permeability of the thylakoid membrane to protons and the loss of ATP utilization due to the inactivation of the Calvin cycle, (2) partial inactivation of PSII activity owing to loss of C02 binding at a regulatory site on the PSII reaction center in the absence of C02, and (3) the normal nonlinear response of the light saturation curve of photosynthesis. Therefore, the potential still exists to improve the efficiency of photosynthetic H2 production. Further research is needed to eliminate these limiting factors. The first limitation may be solved on a short-term basis by using an appropriate proton uncoupler that dissipates the proton gradient across the thylakoid membrane (44), whereas the second could be overcome by eliminating the requirement of C02 binding through molecular engineering. The third limitation can, in principle, be overcome by reducing the antenna size of the photosynthetic reaction centers.

To avoid gas product separation and to increase efficiency, we have previously proposed a PSI and PSII reactor system that can potentially produce H2 and 02 in two separate compartments (32). This reactor system is based on the structure and function of isolated PSI and PSII reaction centers and on the catalytic activity of metallic platinum and osmium for H2 production (45-49). As illustrated in Figure 4, 02 and

VENT

ICE WATER

Figure 2. Schematic illustration of a dual-reactor-flow system for simulta­neous detection of H2, 02, and C02. The dual-reactor system permits two independent experiments to be performed simultaneously.

Figure 3. Sustained photoassimilation of H2 and 02 in Chlamydomonas 137c under anaerobic conditions and in the absence of C02. (Reproduced with permission from Ref. 35 Copyright 1996 Macmillan Magazines Limited)

Figure 4. A photosynthetic reactor system made of PSI and PSII electro­optical cells for production of H2 and 02 in separate compartments. (Reproduced with permission from Ref. 32 Copyright 1996 Human Press Inc.)

protons are produced by water splitting by an array of PSII reaction centers in the PSII compartment. The electrons acquired from water splitting are wired to the reducing side of a PSI array where the electrons are energized again by PSI photochemistry. The PSI-energized electrons are then used to evolve H2 by platinum — (or osmium-) catalyzed reduction of protons that come from the PSII compartment through a proton-conducting channel. As described previously, this system should be able to operate continuously since the number of protons and electrons generated can be balanced with the number consumed. We believe this is an important direction for future research. In this laboratory, research progress has been made in this direction (50-52). Recently, we have constructed a two-dimensional spatial array of PSI reaction centers on a gold surface at nanometer scale by a platinization anchoring technique (52).

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, TL 61604

Advances in enzymes and lignocellulosic biomass processing are necessary to lower the cost of fuels and chemicals production from biomass. Recent developments in lignocellulosic biomass conversion enzymology and process technology are reviewed. Current problems of these multi-enzymes based complex processes, economic assessment, regulatory issues, strategies for development of improved enzymes and processes, and directions of future research are discussed. Results of our endeavor to develop novel enzymes for biomass conversion are presented.

Currently, more than one billion gallons of ethanol are produced annually in the United States, with approximately 95% derived from fermentation of com starch (7). Enzymes play an important part in the conversion of com starch to glucose that is then fermented to ethanol by yeast. In fact, application of amylases in starch conversion is a great example of the successful use of enzymes in biotechnology. With increased attention to clean air and oxygenates for fuels, opportunities exist for rapid expansion of the fuel ethanol industry. Various lignocellulosic biomass such as agricultural residues, wood, municipal solid wastes and wastes from pulp and paper industry have potential to serve as low cost and abundant feedstocks for production of fuel ethanol or chemicals. Right now, the use of lignocellulosic biomass to produce fuel ethanol represents significant technical and economic challenges, and its success depends largely on the development of highly efficient and cost-effective biocatalysts for conversion of pretreated biomass to fermentable sugars. In this article, we describe briefly current knowledge on the application of enzymes in various lignocellulosic biomass conversion.

A preliminary economic evaluation of the MixAlco process shows that from free biomass, the mixed alcohols could be sold for about $0.19/L ($0.72/gal) with a 15% before-tax return on investment. This is a very attractive price compared to other fuel oxygenates such as methyl tertiary butyl ether ($0.24/L or $0.90/gal) and ethanol ($0.29/L or $1.10/gal).

An obvious question is "where does one obtain free biomass?" The answer is from MSW which currently has a disposal cost ranging from $10/wet tonne (Nevada) to over $110/wet tonne (Northeast) with an average of $50/wet tonne. Processing MSW to separate the refractories (e. g., metals and glass) from the organics (e. g., paper, cardboard, food scraps) costs about $21 to $62/wet tonne (19).

Rather than sending the organics to the MixAlco Process, an alternative use for the separated organics is "refuse derived fuel" (RDF); however, chlorine — containing plastics within RDF are very corrosive requiring expensive metallurgy for the boiler tubes and scrubbers to treat the exhaust gases. As a consequence, the energy value of RDF is negative; it actually costs about $55/wet tonne to combust MSW (20). Therefore, this competing use for the separated organics is not attractive.

To avoid the combustion problems described above, the undigested residue from the MixAlco process would likely be landfilled. The undigested residues are dense (specific gravity = 1.2 at 50% moisture) compared to unprocessed MSW in a landfill (specific gravity = 0.71), so the life of a landfill can be significantly increased by employing the MixAlco Process. Assuming that metals and glass are recovered and that 63% of the remaining material is digested, the life of a landfill increases by about 3.5 times.

Taking credit for the increased life of the landfill and using $41/wet tonne as the MSW sorting cost, free biomass can be obtained from communities that currently pay $58/tonne to landfill unprocessed MSW.

A.-P. Zeng, H. Biebl, and W.-D. Deckwer

Biochemical Engineering Division, Gesellschaft fur Biotechnologische
Forschung mbH, Mascheroder Weg 1, D—38124 Braunschweig, Germany

The microbial conversion of glycerol to 1,3-propanediol has recently re­ceived much attention because of the appealing properties of 1,3-propa­nediol and the anticipated surplus of glycerol on the market. Our knowledge of the metabolic pathway, the intrinsic metabolic potential and kinetic limitation of this bioconversion process has been substantially augmented. Progress has also been made in improving the process perfor­mance and strains both on process engineering and molecular biology levels. These recent advances are summarized in this communication. Further research and development needs are also discussed.

H C-O-COR H COH

2 2

The fatty acid is used by the oleo-chemical industry as a feedstock for the production of detergents and other chemical intermediates. Glycerol has been traditionally used in the production of pharmaceuticals, cosmetics, resins, food, beverages, tabacco as well as cellophan and explosives. The oleo-chemical industry has considerably grown in the last decade due to the relatively low price increase of natural fats compared to petro­chemicals. However, no major new application has been found for glycerol during this period, leading to a surplus of glycerol on the market. This surplus of glycerol is expec­ted to expand further with the envisaged application of rape-seed oil as a diesel substitu — © 1997 American Chemical Society

te. The conversion of rape-seed oil to an appropriate diesel fuel yields about 10% glyce­rol by weight (13,15). The use of glycerol as a stock compound for chemical processing is therefore of industrial interest.

One potential use of surplus glycerol is its conversion to 1,3-propanediol (1,3- PD). Recently, this bioconversion has received world-wide attention. It is mainly driven by two factors. First, 1,3-PD is an appealing product and finds applications in the syn­thesis of heterocycles and polyesters. Polyesters based on 1,3-PD have special proper­ties such as biodegradability, improved light stability, anti-sliding and re-stretching qualities when used as material for the manufacture of carpet ware in combination with terephthalic acid. Recently, two large chemical companies Shell and Degussa anounced the commercialization of 1,3-PD production based on petrochemical feedstocks. Se­cond, 1,3-PD from biological route represents a rare case for a primary chemical the biological production of which is competitive or even more economical compared to the chemical route (12). This fermentation process has been successfully scaled up on a pilot plant scale with batch culture of Clostridium butyricum (17). The recovery and purification of 1,3-PD has also been intensively studied. Recently, Deckwer (12) revie­wed the microbial conversion of glycerol to 1,3-PD, covering major work up to about 1993 and including economical aspects. In this communication advances achieved in the last few years are briefly summarized, focusing mainly on pathway, stoichiometric and kinetic analysis, strain improvement and process optimization.

Studies Using a Genetically Engineered Saccharomyces Yeast

M. S. Krishnan1’2, Y. Xia1, N. W. Y. Ho1, and G. T. Tsao1-2

lLaboratory of Renewable Resources Engineering and 2School
of Chemical Engineering, Purdue University, West Lafayette, IN 47907

Fermentation studies of ethanol production from lignocellulosic sugars using the genetically engineered Saccharomyces yeast 1400 (pLNH33) and its parent Saccharomyces yeast strain 1400 are reported. While the parent strain 1400 is unable to ferment xylose, the recombinant yeast 1400 (pLNH33) ferments xylose and mixtures of glucose and xylose. High ethanol yields upto 84% were obtained by fermentation of glucose-xylose mixtures using the recombinant yeast. The kinetics of ethanol inhibition of yeast cell growth on glucose and xylose are presented. Results of ethanol production from com fiber and com cob by the simultaneous saccharification and fermentation (SSF) process are also reported.

Ethanol has received attention recently as an octane booster and a transportation fuel. The economics of fuel ethanol production are significantly influenced by the cost of the raw materials used in the production process. Lignocellulosic materials such as agricultural residues and municipal waste paper have been identified as potential feedstocks, in view of their ready availability and low cost (7). These lignocellulosic hydrolyzates that are produced either chemically or enzymatically contain both pentoses and hexoses. The pentoses are comprised of D-xylose and L-arabinose while the major hexose is D-glucose (2). While the glucose is readily fermented by using Saccharomyces yeasts, few microorganisms have the ability to ferment xylose. For the economics of the biomass to ethanol process, it is necessary to convert the xylose to ethanol as well. Pichia stipitis and Candida shehatae are the best wild type xylose fermenting yeasts that have been reported in the literature (3). Recent advances in molecular biology techniques have led to the development of genetically engineered microorganisms for xylose fermentation. These include recombinant bacterial strains of E. coli (4), Klebsiella oxytoca (5) and Zymomonas mobilis (6).

© 1997 American Chemical Society

Although these strains show good xylose fermentation performance, the low ethanol tolerance of these microorganisms is a limiting factor in the process. Saccharomyces yeasts have a relatively higher ethanol tolerance and hence attempts have been made to develop recombinant Saccharomyces yeasts that can ferment xylose (7,8). However, the ethanol yields and productivities are low. This has been attributed to the cofactor imbalance and an insufficient capacity for xylulose conversion through the pentose phosphate pathway.

A recombinant yeast denoted 1400 (pLNH33) has been developed by Nancy Ho and co-workers at the Laboratory of Renewable Resources Engineering, Purdue University (9,10). This strain was developed using the high ethanol tolerance Saccharomyces yeast 1400 (77) as the host and cloning the xylose reductase, xylitol dehydrogenase genes (both from Pichia stipitis) and xylulokinase gene (from S. cerevisiae) into yeast 1400. The recombinant yeast ferments glucose and xylose simultaneously to ethanol in high yields.

In this paper, we report the fermentation studies conducted on glucose, xylose and their mixtures using this recombinant yeast. Ethanol tolerance is a key factor influencing process economics, motivating us to investigate the kinetics of ethanol inhibition on these genetically engineered yeasts. Results of the simultaneous saccharification and fermentation (SSF) process using com fiber and com cob as model feedstocks are also presented.

Besides the properties discussed above and accompanying operational problems, the question of combustion, emissions, and engine deposits of biodiesel fuels is of extreme significance and will be discussed here.

Generally, similar types of compounds are observed in the exhaust emissions of conventional DF and vegetable oil-derived fuels. This is additional proof of the suitability of fatty compounds as DF because there presumably exist similarities in their combustion behavior.

Emissions from any kind of engine are the result of the preceding combustion within in the engine. The combustion process, in relation to the properties of the fuel, and its completeness are responsible for any problems associated with the use of biodiesel, such as formation of deposits, etc. To understand the formation of emissions and deposits, and possibly direct the combustion to suppress undesirable emissions and deposits, it is essential to study the combustion of the fuel.

Ideally, the products of complete combustion of hydrocarbons are carbon dioxide (C02) and water according to the equation (shown for alkanes (saturated hydrocarbons)):

C„H2^2 + (1.5w + 0.5)O2 — nC02 + (л+1)Н20

Combustion in a diesel engine occurs mainly through a diffusion flame and is therefore incomplete {8). This causes the formation of partially oxidized materials such as carbon monoxide (CO), other oxygenated species (aldehydes, etc.), and hydrocarbons.

In the case of biodiesel, liberation of C02 (decarboxylation), as indicated above, from the ester moiety of the triglyceride or methyl ester occurs besides combustion formation of C02 from the hydrocarbon portions of biodiesel. The formation of C02, an incombustible compound despite its high oxygen content (although mistakenly assumed by some that it can serve as a combustion enhancer because of its high oxygen content), shows that one has to be judicious in choosing oxygenated compounds as combustion enhancers because the combustion-enhancing properties will depend on the nature of the oxygen (bonding, etc.) in those compounds. Therefore, the higher oxygen content of biodiesel does not necessarily imply improved combustion compared to conventional DF because of removal of this oxygen from the combustion process by decarboxylation, but C02 may contribute to combustion in other ways.

Exhaust emissions observed in the combustion of conventional DF and biodiesel are smoke, particulates (particulate matter), polyaromatic hydrocarbons (PAHs), hydrocarbons, CO, and oxides of nitrogen (NOx; also referred to as nitrous oxides, or nitrogen oxides). An important difference are sulfur-containing emissions which are not formed from biodiesel due to its lack of sulfur. Note that rapeseed contains low amounts of sulfur but variations such as canola have not only lower erucic acid content but also reduced sulfur (56).

The composition of particulate matter has been studied for conventional diesel fuels (57). Particulates from conventional DF have a high carbon to hydrogen ratio of approximately 10:1 (38). Thus, particulates are mainly carbon in forms of crystallites. As temperatures decrease below 500°С, the particles are coated with adsorbed and condensed species, which include unbumed hydrocarbons, various oxygenated hydrocarbons, PAHs and nitrogen dioxide (in case of conventional DF, also sulfur — containing species). With rapeseed methyl ester as fuel in DI engines, particulate matter showed large amounts of volatile and extractable compounds adsorbed on the soot, which caused the particulate emissions to be higher than with conventional DF (39).

PAHs are compounds composed of fused aromatic rings that may carry alkyl substituents such as a methyl group. They are of concern because many of them are known carcinogens.

Hydrocarbons represent a broad category of compounds including hydrocarbons and oxygenated species such as aldehydes, ketones, ethers, etc.

Nitrogen oxides (NOJ arise by the reaction of nitrogen and oxygen from air at an early stage in the combustion process (40). NOx emissions are difficult to control because such techniques may increase other emissions or fuel consumption (8).

Emissions of Neat Vegetable Oil Fuel. While neat vegetable oils are competitive with conventional DF in some emission categories, problems were identified for other kinds of emissions. For example, it was shown that PAH emissions were lower for neat vegetable oils, especially very little amounts of alkylated PAHs, which are common in the emissions of conventional DF (41). Besides higher NOx levels (42), aldehydes are reported to present problems with neat vegetable oils. Total aldehydes increased dramatically with vegetable oils (42). Formaldehyde formation was also consistently higher than with DF2. It was reported that component TGs in vegetable oils can lead to formation of aromatics via acrolein (CH2=CH-CHO) from the glycerol moiety (16).

Another author observed significantly lower emissions of C3 aldehydes (for example, acrolein) for methyl esters of rapeseed oil than for the oil itself {43). Another study {44) attributes increased emissions of aldehydes and ketones when using vegetable oils as fuels to the formation of acidic water during decomposition of the oils. This acidic water could be an indication for the formation of short-chain oxygenates which likely ignite poorly compared to the long-chain carbon-rich fatty compounds.

Engine Problems with Neat Vegetable Oil Fuel. Most references in this section report that, at least in short-term trials, neat oils gave satisfactory engine performance and power output, often equal to or even slightly better than conventional DF. However, vegetable oils cause engine problems. This was recognized in the early stages of renewed interest in vegetable oil-based alternative DFs. Studies on sunflower oil as fuel noted coking of injector nozzles, sticking piston rings, crankcase oil dilution, lubricating oil contamination, and other problems {29-31). These problems were confirmed and studied by other authors {45-52). A test for external detection of coking tendencies of vegetable oils was reported (53). The causes of these problems were attributed to the polymerization of TGs via their double bonds which leads to formation of engine deposits as well as the low volatility and high viscosity with resulting poor atomization patterns. An oxidative free-radical mechanism was suggested as governing TG polymerization in lubricating oil contamination when using sunflower oil as fuel {54). Fumigation with propane was studied as a means to reduce injector coking (55). The engine problems have caused neat vegetable oils to be largely abandoned as alternative DF and lead to the research on the aforementioned four solutions (32).

Emissions of esters. Generally, most emissions observed for conventional DF are reduced when using esters. NOx emissions are the exception. In an early paper reporting emissions with methyl and ethyl soyate as fuel (20), it was found that CO and hydrocarbons were reduced but NOx were produced consistently at a higher level than with the conventional reference DF. The differences in exhaust gas temperatures corresponded with the differences in NOx levels. Similar results were obtained from a study on the emissions of rapeseed oil methyl ester (43). NOx emissions were slightly increased, while hydrocarbon, CO, particulate and PAH emissions were in ranges similar to the DF reference. As mentioned above, the esters emitted less aldehydes than the corresponding neat rapeseed oil. Unrefined rapeseed methyl ester emitted slightly more aldehydes than the refined ester, while the opposite case held for PAH emissions. A 31 % increase in aldehyde and ketone emissions was reported when using rapeseed methyl ester as fuel, mainly due to increased acrolein and formaldehyde, while hydrocarbons and PAHs were significantly reduced, NOx increased slightly, and CO was nearly unchanged (56). The study on PAH emissions (41), where also the influence of various engine parameters was explored, found that the PAH emissions of sunflower ethyl ester were situated between DF and the corresponding neat vegetable oil. Reduced PAH emissions may correlate with the reduced carcinogenity of particulates when using rapeseed methyl ester as fuel (57). The general trend on reduced emissions except NOx was confirmed by later studies (58), although some studies report little changes in NOx (59-60). In a DI engine, sunflower methyl ester produced equal hydrocarbon emissions but less smoke than a 75:25 blend of sunflower oil with DF (61). Using a diesel oxidation catalyst (DOC) in conjunction with soy methyl ester was reported to be a possible emissions reduction technology for underground mines (62). Soy methyl esters were reported to be more sensitive towards changes in engine parameters than conventional DF (63).

Precombustion of Triglycerides. As discussed, every DF, conventional or vegetable oil-based, experiences an ignition delay, which is the basis of CN measurements. The fuel passes through a temperature and pressure gradient directly after injection but before combustion begins. Chemical reactions already occur in this precombustion phase. In an initial study (64), the unsaturated TGs triolein, trilinolein, and trilinolenin were studied at temperatures up to 400°С in air or N2 in a reactor simulating conditions in a diesel engine. The compounds arising in this phase were fatty acids of different chain lengths (some even longer than those in the parent fatty acids), various aliphatic hydrocarbons, and smaller amounts of other compounds such as aldehydes. The parent acids were the most prominent compounds in the precombustion mixture. Component patterns were largely independent of the starting material and reaction conditions. In a second study (65), tristearin and tripalmitin were studied besides the three unsaturated TGs at temperatures of450°С in air and N2. Presumably due to the higher temperature, different component patterns were observed. Besides mainly unsaturated aliphatic hydrocarbons and unsaturated aldehydes, various aromatics, including benzene, toluene, compounds with unsaturated side chains, and polyaromatic hydrocarbons were detected. The atmosphere (air or N2) had considerable influence on product formation. The number of components was less for samples of tripalmitin, tristearin and triolein for reactions under N2 than under air while this finding was reversed for trilinolein and trilinolenin. No fatty acids, glycerol or acrolein (as decomposition product of glycerol) were detected. Extensive decarboxylation occurred, showing that the oxygen in biodiesel does not necessarily contribute to its combustion as an oxidizer. The compounds identified are also found in the exhaust emissions of engines running on conventional DF. It is therefore necessary to influence not only combustion but also precombustion to improve the combustion properties and emissions of biodiesel.

Cetane Improvers. Various compounds such as alkyl nitrates are used as cetane­enhancing additives in conventional DF (66). Few studies on such compounds in biodiesel exist. One paper reports (67) that in a turbulence combustion chamber and at an intake air temperature of 105 °С, 8% hexyl nitrate in vegetable oils (cottonseed, rape, palm) was necessary to exhibit the same ignition delay as conventional DF. The use of nitrate esters of fatty acids as cetane improvers in DF was reported in a patent (68).

Fructose is an expensive carbon source in the fermentation industry, hence we next investigated the application of other economical carbon sources for the heterotrophic culture in the two stage method. We used acetic acid as the carbon source and developed a pH-stat continuos substrate-feeding method for the culture. Flask culture experiments showed the optimum concentration of acetic acid for cell growth was very low (ca.1.0 g-dnr3) and the growth was seriously inhibited by a slight increase in acetic acid concentration. It was, therefore, necessary to control the acetate concentration around this level in high cell density cultivation. The ratio of consumption of acetic acid to that of ammonium by A. eutrophus cells was determined by a standard-type batch culture experiment to be about 10 (mol-acetic acid/mol — ammonium). It was therefore expected that the acid-base equilibrium in culture system would be balanced by feeding the substrate solution in which the C/N ratio was 10 (mol/mol) so as to maintain the culture pH at a constant level, in order for acetate concentration in the fermentor to be also controlled at low level. However, in batch culture with such a feeding, acetate concentration of the culture liquid increased after cell concentration reached approximately 5 g*dnr3. The increase in acetate concentration was thought to be due to the depletion of mineral nutrients. Hence, the mineral concentrations in the medium was increased 5 times as that of the basal medium (this was referred as 5-fold medium). As a result, acetate concentration was controlled around 1 gednr3 and cell concentration reached about 25 g*dnr3 after 18 h. Acetate concentration increased after that due to the depletion of phosphate. A pH- stat batch culture was hence carried out by feeding a solution in which the C/P ratio was 118.4 (mol-C/mol-P). Acetate concentration was maintained around 1 gednr3, and cell and protein concentrations increased to 48.6 g*dm-3 and 35.0 gedm~3, respectively after 21 h of cultivation (Fig.4). Autotrophic cultivation for P(3HB) accumulation was then performed using the protein-rich cell mass obtained from a pH-stat batch culture into which the modified substrate solution was fed. When the cell concentration reached to about 5 g*dnrr3, feeding of the substrate solution was stopped and autotrophic cultivation was performed by feeding a substrate gas mixture into the fermentor. P(3HB) accumulated in the cells up to about 60 % by dry cell weight. This feeding method can therefore be used in fermentation process where the cell growth or P(3HB) accumulation is inhibited by high concentrations of the substrate such as propionate, formate, or lactate, etc.

Time course of autotrophic culture of A. entrophtts in bench-plant scale culture system.

Table I. List of autotrophic cultivations of hydrogen-oxidizing bacteria and fermentaive production of P(3HB) using various microorgansims and substrates

Strains Substrate Culture Cultivation method time (h) Cell concentration (g-dm3) Cell productivity (g-dm J-h‘) P(3HB) P(3HB) concentration productivity (g-dm3) (g-dm3-h‘) Ref. Alcaligenes H2/02/C02 Batch 25 25.0 1.00 — — 9 eulrophus Alcaligenes H2/02/C02 Batch 40 91.3 2.28 61.9 1.55 5 Pseudomonas H2/02/CO2 Batch 48 24.0 0.50 — — 10 hydrogenovora Alcaligenes H2/02/CO2 Continuous 0.40 — — 11 eutrophus Alcaligenes H2/02/C02 Batch 70 18.0 0.26 16.0 0.23 12 eulrophus H16 Alcaligenes H2/02/C02 Continuous 0.33 — — 13 hydrogenophilus Pseudomonas H2/02/C02 Continuous 3.00 — — 14 hydrogenothermophila Alcaligenes H2/O2/CO2 Batch 60 60.0 1.00 36.0 0.60 3 CUltUfjrtUA Recombinant Glucose Fed-batch 42 117.0 2.79 89.0 2.11 15 E. coli Protomonas Methanol Fed-batch 121 223.0 1.84 136.0 1.12 16 елігициепь Alcaligenes Glucose Fed-batch 50 164.0 3.28 121.0 2.42 17 eutrophus Alcaligenes Sucrose Fed-batch1) 18 142.0 7.89 68.4 4.0 18 latus Methylo- Methanol Fed-batch 70 250.0 5.57 130 1.85 19 bacterium organophilum Alcaligenes Glucose* Fed-batch 50 — — 90.42) 1.80 20 eutrophus valerate 1) The inoculum cell size was 13.7 g*dm 2)The product was poly(hydroxybutyrate-co-hydroxyvalerate).

Economic calculations for the enzymatic process performed by von Sivers (48) show that the capital cost is the most dominating, amounting to 47% of the totabcost, followed by the raw material cost (30 %). The cost estimate is based on experimental data from the literature using pine as raw material and a plant capacity of 100,000 ton dry matter/year. The calculations are based on separate hydrolysis and fermentation (SHF) and anaerobic fermentation of the waste water. Figure 3 shows the production cost for each process step. The most expensive step, constituting 15% of the total cost, is the steam production, including the lignin dryer, the steam boiler and the anaerobic treatment of the liquid waste, followed by pretreatment (13%) and enzymatic hydrolysis (12%). In the following, four examples of major contributors to the overall process economy are discussed.

Influence of the Raw Material. Since the raw material contributes a large part to the total ethanol production cost, the production cost is greatly affected by the cost of the wood and by the chemical and physical characteristics of the wood. These characteristics determine the difficulty in converting the cellulose and hemicellulose fractions in the wood to fermentable sugars at high yields, which in turn influences capital and operating costs. Softwood has proven to be more difficult to utilise than hardwood (30, 49). The chemical composition of the wood determines how much ethanol can theoretically be produced per tonne of raw material. The difference in composition between hardwood and softwood is that hardwood contains high levels of xylan, low levels of mannan and less lignin than softwood. The high level of xylan in hardwood makes it necessary to include a pentose fermentation step or the production of some co-product such as furfural or methane from the xylose, to make ethanol production from hardwood economically feasible. The pentose fermentation step requires further development before it can be used in a full-scale process, although much progress has been made in recent years. Zymonas mobilis (39), S. cerevisiae (37, 50), and Escherichia coli (38) have successfully been genetically transformed to ferment xylose to ethanol, but the organisms must also tolerate the inhibiting components in the hydrolysates. Otherwise, a detoxification step must be included, which has been shown to be quite expensive (51).

As the cost of the raw material is high, its maximum utilisation is important to lower the final cost of the ethanol. The overall ethanol yield has proven to be the most important factor for the ethanol production cost and it is necessary to develop the various steps in the process, i. e. pretreatment, hydrolysis and fermentation, to achieve as high a yield as possible. As a consequence of this, we have at Lund University a large R&D program focused on the development of these different process steps. However, it is also very important to examine how the various process steps perform in an integrated process, for example, how the accumulation of inhibitors due to recirculation will affect the ethanol yield. This has been investigated in the bench — scale process development unit where the whole process can be experimentally simulated. This is discussed in more detail in the section "Recycling of process streams".

Capital Cost. The enzyme production and the hydrolysis steps are the major contributors to the capital cost. These steps are rate limiting in the process and the high costs are due to the long residence time which requires numerous and large reactors. Decreased residence times in these two steps must be weighed against reduced enzyme and sugar yields. One way of reducing the capital cost for the enzymatic hydrolysis and fermentation steps is to use the SSF concept. There are several advantages with SSF. Only one reactor is needed for both hydrolysis and fermentation, no product inhibition of the enzymes in the hydrolysis arises when the glucose is converted directly to ethanol, and the risk of contamination decreases for the same reason. According to a study performed by Wright et al. (52) the SSF process leads to a reduction in the total production cost of about 30% compared with SHF. One drawback of SSF which remains to be overcome is recycling of the yeast. If the pretreated material is not delignified, it will be very difficult to separate the yeast from the lignin residue after SSF. The capital and productivity advantages attributed to the SSF configuration may decrease drastically due to the cost of producing new yeast in every fermentation batch or the cost of an additional delignification step. An alternative is to run the SSF continuously at low dilution rates (long residence times) so that the yeast has time to grow; but this will require extra-large fermenters. As the hydrolysis is rate limiting, SSF will, in any case, require large fermenters compared with separate hydrolysis and fermentation, where only the hydrolysis tanks are large. Fermenters are more complex and expensive than hydrolysis tanks and the gain in using one reactor instead of two will decrease. The bench-scale unit will, in the near future, be used to compare the SHF and the SSF methods in an integrated process.

Energy. Some operations, such as pretreatment of the raw material, distillation of the product, drying of the lignin, and, if included, evaporation of the stillage for recirculation of process water, are extremely energy demanding. In the steam pretreatment stage, high-quality steam of 20-30 bar is used, while in the drying section and in distillation steam of 3-6 bar is used. The steam consumption in the distillation step is very dependent on the ethanol concentration in the distillation unit. However, in a carefully designed and energy-integrated plant it is possible to reduce the energy costs by a considerable degree. This is described in more detail in the section "Process Integration".

By-products. The primary by-product from the large-scale production of ethanol will be lignin. Lignin can be utilised for many chemical applications, but due to the large amount of lignin which will be produced in a future transition from fossil fuels to fuel ethanol, the most realistic use of lignin is as a solid fuel. An alternative is to use the lignin for electricity production in a back-pressure power plant. The price obtained for the lignin will affect the cost of the ethanol (53, 54) but not as much as the cost of the raw material. Lignin and other by-products are produced in the bench-scale unit for further characterisation as the quality is of great importance for the price.

An advantageous feature of fermentative H2 production is its temporal separation from photosynthesis (Figure 1). That is, 02 evolution and C02 photoassimilation by photosynthesis occur during the day, whereas fermentative H2 production by degradation of photosynthetic product (starch) can occur during the night. By taking advantage of this temporal separation between 02 evolution (day) and H2 production (night), one can potentially develop an algal H2 production technology that avoids the problem of gas product separation. This approach is a major project by scientists in Japan (22, 29, 33, 34), and important progress has been made. Miura et al. (34) have demonstrated algal fermentative H2 production that is temporally separated from photosynthesis (C02 fixation and 02 evolution), using a combination of green algae and photosynthetic bacteria comprising over 100 L of green algae.

The challenge, as always, in this fermentative H2 production approach is the efficiency. Figure 5 presents a typical measurement of photosynthesis and fermentative H2 production in wild-type Chlamydomonas in a helium atmosphere in the presence of C02 under cycles of 12 h of moderate actinic illumination (PAR, 200 pE — m’2* s1) and 12 h of darkness, using the flow-detection system (Figure 2). From the data, it can be clearly seen that the rate of fermentative H2 production is very slow—less that 5% of the rate of oxygen evolution during the day. Therefore, enhancing the rate of dark fermentative H2 production is the key challenge.