A biological solution that bypasses the severe practical difficulties posed by growing ethanologens in concentrated solutions of potentially toxic hydrolysates of lignocel — lulosic materials is to replace physicochemical methods of biomass substrate hydro­lysis with enzymic breakdown (cellulase, hemicellulase, etc.) under milder conditions — especially if enzyme-catalyzed hydrolysis can be performed immediately before the uptake and utilization of the released sugars in a combined hydrolysis/fermenta — tion bioprocess. Extrapolating back up the process stream and considering a totally enzyme-based hydrolysis of polysaccharides, an “ideal” ethanol process has been defined to include193

• Lignin removal during pretreatment to minimize unwanted solids in the substrate

• Simultaneous conversion of cellulose and hemicellulose to soluble sugars

• Ethanol recovery during the fermentation to high concentrations

• Immobilized cells with enhanced fermentation productivity

An even closer approach to the ideal would use enzymes to degrade lignin suffi­ciently without resort to extremes of pH to fully expose cellulose and hemicellulose before their degradation to sugars by a battery of cellulases, hemicellulases, and ancillary enzymes (esterases, etc.) in a totally enzymic process with only a minimal biomass pretreatment, that is, size reduction. No such process has been devised, but because pretreatment methods could solubilize much of the hemicellulose, two dif­ferent approaches were suggested in 1978 and 1988 with either cellulolytic microbes (whole cell catalysis) or the addition of fungal cellulase and hemicellulase to the fermentation medium.194 195 These two options have become known as direct micro­bial conversion (DMC) and SSF, respectively.

DMC suffers from the biological problems of low ethanol tolerance by the (usually) clostridial ethanologens and poor ethanol selectivity of the fermentation (see section 3.3.2.5).196 Commercialization has been slow, few studies progressing beyond the laboratory stage. The phytopathogenic[41] fungus Fusarium oxysporium is the sole nonbacterial wild-type microbe actively considered for DMC; the ability of the organism to ferment xylose as well as hexose sugars to ethanol was recognized in the early 1980s, and several strains can secrete cellulose-degrading enzymes.197198 Hemicellulose sugars can also be utilized in acid hydrolysates, although with low conversion efficiencies (0.22 g ethanol per gram of sugar consumed).199 Extensive metabolic engineering of F. oxysporium is, therefore, likely to be required for an efficient ethanologen, and detailed analysis of the intracellular biochemical networks have begun to reveal potential sites for intervention.200-202

Metabolic engineering of S. cerevisiae to degrade macromolecular cellulose has been actively pursued by research groups in South Africa, the United States, Can­ada, Sweden, and Japan; fungal genes encoding various components of the cellulase complex have successfully been expressed in ethanologenic S. cerevisiae, yielding strains capable of utilizing and fermenting either cellobiose or cellulose.203-206 Cal­culations show that, based on the growth kinetics and enzyme secretion by cellulose degraders such as H. jecorina, approximately 1% of the total cell protein of a recom­binant cellulase-secreting S. cerevisiae would be required, perhaps up to 120-fold more than has been achieved to date.207,208

In contrast, SSF technologies were installed in North America in the early 1990s in production plants generating between 10 million and 64 million gallons of ethanol/year from starch feedstocks.122 In addition to starch breakdown and sugar fermentation, the technology can also include the stage of yeast propagation in a cas­caded multifermentor design (figure 4.11). Extensive research worldwide has defined some factors for successful process development:

• If yeasts are to act as the ethanologens, thermotolerant strains would per­form more in harmony with the elevated temperatures at which cellulases work efficiently.209,210

• Bacteria are more readily operated in high-temperature bioprocesses, and recombinant Klebsiella oxytoca produced ethanol more rapidly under SSF conditions than did cellobiose-utilizing yeasts; coculturing K. oxytoca and S. pastorianus, K. marxianus, or Z. mobilis resulted in increased ethanol production in both isothermal and temperature-profiled SSF to increase the cellulase activity.211

• Both K. oxytoca and Erwinia species have the innate abilities to trans­port and metabolize cellobiose, thus reducing the need to add exogenous P-glucosidase to the cellulase complex; moreover, chromosomally integrat­ing the E. chrysanthemi gene for endoglucanase and expressing the gene at a high level results in high enzyme activities sufficient to hydrolyze cellulose and even produce small amounts of ethanol in the absence of added fungal

cellulase.212,213

• Although high ethanol concentrations strongly inhibit fermentations with recombinant E. coli and glucose or xylose as the carbon substrate, SSF with cellulose and added cellulase showed a high ethanol yield, 84% of the theoretical maximum.214

• Simulations of the SSF process to identify the effects of varying the operat­ing conditions, pretreatment, and enzyme activity highlight the importance of achieving an efficient cellulose digestion and the urgent need for contin­ued R&D efforts to develop more active cellulase preparations.215

• Reducing the quantity of cellulase added to ensure efficient cellulose diges­tion would also be beneficial for the economics of the SSF concept; add­ing nonionic surfactants, polyethylene glycol, and a “sacrificial” protein to decrease nonproductive absorption of cellulase to lignin binding sites have also been demonstrated to increase cellulase action so that cellulose diges­tion efficiency can be maintained at lower enzyme:substrate ratios.8216

The importance of the quantity of cellulase added was underlined by a Swedish study that showed that reducing the enzyme loading by 50% actually increased the production cost of ethanol in SSF by 5% because a less efficient cellulose hydro­lysis reduced the ethanol yield.217 At low enzyme loading, there are considerable advantages by growing the yeast inoculum on the pretreated biomass material (bar­ley straw); the conditioned cells can be used at a reduced concentration (2 g/l, down from 5 g/l), and with an increased solids content in the SSF stage.218

The cost of commercially used fungal cellulase has decreased by over an order of magnitude because of the efforts of enzyme manufacturers after 1995.219 Multiple efforts have been made to increase the specific activity (catalytic efficiency) of cel — lulases from established and promising novel microbial sources (see section 2.4.1), and recently, the National Center for Agricultural Utilization Research, Peoria, Illi­nois, has focused on the cellulase and xylanase activities from the anaerobic fungus Orpinomyces, developing a robotic sampling and assay system to improve desirable gene mutations for enzymic activity.220-222 Inserting genes for components of the cellulase complex into efficient recombinant ethanol producers has also continued as part of a strategy to reduce the need to add exogenous enzymes; such cellulases can be secreted at levels that represent significant fractions of the total cell protein and increase ethanol production capabilities.223-227 This is of particular importance for the accumulation of high concentrations of ethanol because ethanol at more than 65 g/l inhibits the fungal (H. jecorina) cellulase commonly used in SSF studies.228

SSF has been shown to be superior to independent stages of enzymic hydrolysis and fermentation with sugarcane bagasse, utilizing more of both the cellulose and hemicelluloses.229 A continued industry-wide commitment to SSF is evident in the numbers of publications on SSF technologies applied to ethanol production with a wide variety of lignocellulosic feedstocks (table 4.3).32, 230-238 Issues of process eco­nomics are discussed in chapter 5. Prominent in the list of lignocellulosic feedstocks in table 4.3 is corn stover, a material that has the unique distinction of having a specific biocatalyst designed for its utilization.239 This fusion of the biochemical abilities of Geotrichum candidium and Phanerochaete chrysosporium points toward a long-term option for both SSF and DMC, that of harnessing the proven hypercapabilities of some known microbes to degrade lignocellulose (see section 2.4.1) and converting them to ethanologens by retroengineering into them the ethanol biochemistry of Z. mobilis (see section 3.3.2). Before then, attempts will without doubt continue to introduce fungal genes for starch degradative enzymes into candidate industrial ethanologens and explore the possible advantages from combining genetic backgrounds from two microbes into a single hybrid designed for high amylase secretion.240-242 On a paral­lel track, commercial use of food wastes such as cheese whey, a lactose-rich effluent stream, has prompted the construction of strains with P-galactosidase to hydrolyze lactose extracellularly and use both the released glucose and galactose simultaneously for ethanol production under anaerobic conditions.243

As a final option — and one that mimics the evolution of natural microbial com­munities in soils, forest leaf litter, water-logged areas, and stagnant pools — coculti­vation of a good ethanologen together with an efficient secretor of enzymes to degrade polymeric carbohydrates and/or lignocelluloses is a route avoiding introducing genetically manipulated (GM) organisms and could be adapted to continuous tech­nologies if a close control of relative growth rates and cell viabilities can be achieved. One or more of the microbial partners can be immobilized; table 4.4 includes two examples of this approach together with the cocultivation of different ethanologens to ferment glucose/xylose mixtures and pretreated lignocellulosics.244-249

Simultaneous Saccharification and Fermentation Applied to Fuel Ethanol Production from Lignocellulosic Feedstocks

TABLE 4.4

Cocultivations of Ethanologenic and Ethanologenic Plus Enzyme-Secreting Microbes for DMC/SSF Processes

The International Energy Agency’s 2006 analysis concludes that biodiesels are not price competitive with conventional diesels if all subsidies to crops and production are excluded; if the biodiesel source is animal fat, however, the derived biodiesel would be competitive at crude oil prices below $50-55/barrel.2 By 2030, assum­ing various process improvements and economies of scale, biodiesel from vegetable oils were also predicted to be competitive at crude oil prices below $50-55/barrel; European biodiesel would continue to be more expensive than U. S. biodiesel, with feedstock costs being the largest contributor (figure 6.4). This continues a tradition of cost assessments that commenced in the 1990s. Rapeseed oil-derived biodiesel was estimated to require a total subsidy of between 10% and 186% of the price of conventional diesel in 1992, the variation reflecting the price of the seeds sown to grow the crop; this was equivalent to the cost of biodiesel being between 11% and 286% of the refinery gate cost of conventional diesel — the contemporary price for seeds would have resulted in biodiesel being 243% of the conventional diesel

^ 0.6 G

и

I 0.5

CD

0.4

їй

-З

0.3

CD

0.2

їй

О

0.1

0

cost.4 These calculations required both coproducts (glycerol and rapeseed meal) to be income generators; in addition, the spring-sown crop (although with lower yields per unit of land) had lower production costs.

An evaluation of costs from U. S. soybean and sunflower in 2005 concluded that soybean biodiesel would have production costs 2.8-fold that of conventional diesel (using 2003 price data), whereas sunflower-derived biodiesel was more than five­fold more expensive to produce than petroleum diesel fuel.49 Sunflower seeds were accepted to have a higher oil content (25.5%) than soybeans (18%) but to also have a lower crop productivity (1500 versus 2700 kg/hectare); in both cases, oil extraction was calculated to be highly energy intensive.

The production costs for biodiesels also depend on the production route. With waste cooking oil as the feedstock, although an alkali-catalyzed process using virgin vegetable oil had the lowest fixed capital cost, an acid-catalyzed process using the waste oil was more economically feasible overall, providing a lower total manufac­turing cost, a more attractive after-tax rate of return, and a lower biodiesel breakeven price; in addition, plant capacity was found to be a significant factor affecting the economic viability of biodiesel manufacture.50 51 The U. S. Department of Agricul­ture’s Agricultural Research Service has developed a computer model to estimate the capital and operating costs of a moderately sized industrial biodiesel production facility (annual production capacity, 10 million gallons):52

• Facility construction costs were calculated to be $11.3 million.

• The largest contributors to the equipment cost (accounting for nearly one — third of expenditures) were storage tanks to contain a 25-day capacity of feedstock and product.

• At a value of $0.52/kg for feedstock soybean oil, a biodiesel production cost of $0.53/l ($2.00/gallon) was predicted.

• The single greatest contributor to this value was the cost of the oil feed­stock, which accounted for 88% of total estimated production costs. An analysis of the dependence of production costs on the cost of the feedstock indicated a direct linear relationship between the two.

• Process economics included the recovery of coproduct glycerol generated during biodiesel production, and its sale into the commercial glycerol mar­ket, which reduced production costs by approximately 6%.

Waste cooking oils, restaurant grease, and animal fats are inexpensive feedstocks; they represent 30% of total U. S. fats and oil production but are currently devoted mostly to industrial uses and animal feed, and because the free fatty acids may represent more than 40% of the material, the production process may be complex.53 Nevertheless, such unconventional feedstocks may become increasingly important because soybean oil prices reached a peak not seen since 1984: an article posted online in Biodiesel Magazine traced the rapid inflation in soybean oil from early 2006.54 This surge in the price of biodiesel feedstock has occurred despite the stocks of soybean oil being at near-record levels — and in only three individual months between January 2006 and March 2007 did soybean oil use exceed production (figure 6.5). With its main feedstock being increasingly expensive, U. S. biodiesel is

not competitive on price with diesel fuels or heating oil; tax incentives may be neces­sary to overcome these production price issues.54

Confounding these feedstock problems, the great expansion of biodiesel produc­tion both in Europe and the United States has caused such a glut of glycerol-containing waste (or coproduct) that, in the absence of glycerol valorization mechanisms in place and on site, disposing of this glycerol is proving an increasingly expensive disposal cost outlay. Because of the enormous potential of this renewable source of a potentially valuable chemical intermediate, however, biodiesel waste glycerol is best considered an example (rather premature) of “biocommodity engineering” and is discussed at length in chapter 8 (section 8.3.3) when the broader topic of replacing petrochemicals by biobased products is considered.

Although a few pioneering studies attempted cost estimates of wood-derived ethanol from the 1980s onward, they focused on aspects of the technological processes required rather than making firm conclusions about market prices.24,25 Swedish studies appear to have been the first to present detailed cost breakdowns for ethanol production from accessible large-scale woody biomass sources.26,27 The first of the two to be published was highly unusual in that it used recent advances in pentose utilization by recombinant bacteria to model a pentose stream process, that is, using the solubilized sugars from the pretreatment of wood (willow) feedstock — the procedure involved the impregna­tion of the material with SO2 and subsequent steaming — and included a detoxification procedure to reduce the levels of inhibitors from the hydrolysate.26 The fermentation with Escherichia coli KO11 (chapter 3, section 3.3.2.1) was assumed to consume 96%

[49] The authors did not discriminate between xylose and xylooligosaccharides or between xylose and arabinose, and the implicit assumption was that more than 95% of the available pentoses were in a readily metabolizable form, that is, monomeric xylose.

A very similar analysis of three different approaches to utilizing the full carbo­hydrate potential of pine wood, digesting the cellulose component with concentrated acid, dilute acid, and enzymatic methods, calculated a full manufacturing costs for ethanol of between 500 and 530/l ($1.89-2.01/gallon).27 The bulk of the production cost (up to 57.5%) was accounted for by the financial costs of installing the hardware for generating fermentable carbohydrates (hexoses as well as pentoses) in more com­plex total processes with longer cycle times.

The bacterium K. oxytoca was isolated from paper and pulp mills and grows around other sources of wood; in addition to growing on hexoses and pentoses, it can uti­lize cellobiose and cellotriose but does not secrete endoglucanase.162 A University of Florida research group transformed strain M5A1 with the xylose-directing PET operon; unlike experience with E. coli, lower plasmid copy numbers gave higher ethanol productivity than with higher plasmid copy number.218 A PET transformant could produce ethanol at up to 98% of the theoretical yield and was highly suitable for lignocellulose substrates because it utilized xylose twice as fast as glucose — and twice as fast as did E. coli strain KO11.

Stabilizing the PET operon was accomplished by chromosomal integration at the site of the PFL (pfl gene); screening for mutants hyperresistant to the selectable chloramphenicol marker resulted in the P2 strain with improved fermentation kinetics capable of producing 44-45 g/l of ethanol from glucose or cellobiose (100 g/l) within 48 hr.219 Strain P2 has been demonstrated to generate ethanol from the cellulosic and lignocellulosic materials sugarcane bagasse, corn fiber, and sugarbeet pulp.167,220,221

As a candidate industrial strain for bioethanol production, P2 can utilize a wide range of low-molecular-weight substrates, including the disaccharides sucrose, malt­ose, cellobiose, and xylobiose, the trisaccharides raffinose, cellotriose, and xylotriose, and the tetrasaccharide stachyose.172,181,219,222 This relatively nonspecific diet has led to the cloning and expression of a two-gene K. oxytoca operon for xylodextrin utilization in E. coli strain KO11; the gene product of xynB is a xylosidase (which also has weak arabinosidase activity), whereas that of the adjacent gene in the K. oxytoca genome (xynT) is a membrane protein previously found in Na+/melibiose symporters[28] and related proteins functioning in transmembrane export and import.223 The enhanced recombinant E. coli could metabolize xylodextrins containing up to six xylose residues; unexpectedly, xylodextrin utilization was more rapid than by the donor K. oxytoca.

It is vital at this point to differentiate commercial realities from strategic (or geopo­litical) and all other considerations. Although (as discussed in chapter 1) historical, environmental, political, and macroeconomic arguments have all been adduced in support of bioenergy programs, fiscal considerations now play an important role in both encouraging (prompted by political agendas) the take-up of novel alternative fuels and in partitioning the market for first and subsequent generations of rival but not equally readily commercialized biofuels. Indeed, taxation issues were quickly recognized and seized on by proponents of bioethanol, particularly because they were useful to counter the gasoline versus gasohol price differential: for example, in the United States, the indirect costs of regulating air pollution and of military protection for oil supplies from the Middle East are calculable and greatly inflate the nominal price of crude oil but are not (explicitly) passed on to the consumer.5 This distortion of the transportation fuel market by “hidden” subsidies has also led to economics models in which other indirect factors are included in the cost-benefit analysis:

• Technological developments that improve the national scientific base for employment, patents and overseas licensing, and engineering advances that “spillover” into related fields

• Reduced foreign currency payments and associated “banking” costs—highly important for a developing economy such as Brazil’s

• Higher income and sales tax returns from greater rural employment

• Reduced longer-term economic impacts of climate change and air pollution

All these arguments are, to varying degrees, contentious, and skeptics can be found from opposite ends of the economic spectrum, from oil industry analysts to academ­ics who foresee only accelerated land degradation from the industrial agronomy of energy crop cultivation.67

Taxation as an instrument of social and economic policy has, moreover, obvi­ous limitations if wasteful subsidies or punitive levels of taxation on standard gasoline and diesel products are to be avoided. Consider the following three scenarios:

• Bioethanol production can generate a commercial fuel with pump prices no greater than those of standard gasoline grades at equivalent tax rates, the comparison being valid when average prices during a period of one to five years are calculated, thus avoiding false comparisons at peaks and troughs caused by fluctuations in both agricultural feedstock prices (as an important cost input to biofuel production) and oil price movements if they continue to move inside the wide limits evident since the early 1980s (figure 5.1).

• Bioethanol can be produced commercially at a total (production, distribu­tion, and resale) cost that averaged, during a 5- to 10-year period, 10-50% higher than that of gasoline.

• Bioethanol production can only generate an unsubsidized product with a total cost more than twice that of the refinery gate price of standard gasoline (a price differential quoted for the United States in the late 1990s8) — or per­haps, even up to 10 times higher than conventional fuels where, for example, local conditions of climate and biomass availability are consistently much less favorable than for sugarcane production in Brazil or corn in the United States or where only refractory lignocellulosic feedstocks can be accessed with poorly developed bioprocessing technology.

The first (optimistic) case approximates that of Brazilian consumers with flexibly fueled cars after 2000.9 The second case is the conclusion most often reached in technoeconomic studies, whereas the third scenario is parallel to the emergency or “wartime” case discussed when the energy yields of conventional and alternative fuels were considered in chapter 1 (section 1.6.1): even if biofuels are prohibitively expensive now, technical developments may erode that differential or be obviated if (or when) fossil fuel shortages become acute in the present century (see later, section 5.6). In all three cases, taxation policy can (and will) influence consumer choice and purchasing patterns, whether for short — (tactical) or longer-term (strategic) reasons and when legislation enforces alternative or reconstituted fuels to achieve environ­mental targets.

A snapshot of data from October 2002 in Brazil, however, reveals the complexity of the interaction between production/distribution costs and imposed taxation on the final at-pump selling price.10 Although gasohol mixtures, hydrous ethanol, and diesel all had very similar production costs, equivalent to approximately 150/l (570/gal — lon) at that time, the final cost to the user was determined by the much higher taxes applied to gasohol (figure 5.2). Brazil exemplifies the extensive use of taxation to determine and direct the perceived prices of gasoline and alternative fuels as a delib­erate instrument of national policy. Such management of the fuel economy is likely to be instigated in societies where not only economics but social and environmental considerations are taken into account, but runs the risk of experiencing budgetary shortfalls if the total tax raised is severely reduced when the policy is too successful in achieving its aims — this becomes even worse if private transportation is perceived as being subsidized by other taxation sources (e. g., sales tax, income tax). For all the various interest groups in biofuels development, therefore, the priority is to establish viable production processes with the minimum requirement for tax incentives.

5-Hydroxymethylfurfural (HMF) was discussed in chapter 3 as a toxic product of acidic pretreatment techniques for biomass. The boiling point of HMF is too high (291°C) to be considered as a liquid fuel, but if HMF is subject to chemical hydroge — nolysis of two of its C-O bonds, a more volatile product, 2,5-dimethylfuran (DMF) is formed (figure 7.3).12 DMF has a boiling point of 93°C, 20°C higher than ethanol, and has a Research Octane Number of 119 — the by-product 2-methylfuran has an even higher RON (131) but is more water-soluble than DMF.

HMF is most readily formed by the dehydration of fructose, a naturally occur­ring sugar and a straightforward isomerization product of glucose; mineral acids such as hydrochloric acid (HCl) can be used to catalyze the reaction, 88% conversion being achieved at 180°C.12 A solvent such as n-butanol (chapter 6, section 6.3.3) can then be employed to extract the HMF before hydrogenolysis over a mixed Cu-Ru catalyst at 220°C.

Such production routes, beginning with enzymic conversion of glucose to fruc­tose and proceeding via entirely thermochemical processes, have been described as “hybrid.”13 They have the advantage of avoiding reliance on large fermentation vessels for the production step(s), therefore being potentially much more rapid. Their economics could be similar to, or an improvement on, those for Fischer-Tropsch liquid fuels (chapter 6, section 6.2). The conversion of glucose to fructose, catalyzed by the enzyme glucose isomerase, has been a major industrial application of enzy- mology since the 1960s, the product (high-fructose corn syrup) being introduced as a substitute for Cuban sugar in the U. S. reduced-calorie sweetener market.14 The enzyme technology has been continuously improved, evolving to immobilized forms of the enzyme; the potential of enzymes from hyperthermophilic microbes has now been explored, with a stability at 80oC rivaling that of conventional enzyme pro­cesses operated at 55-650C.15 Rapid and efficient processing of glucose solutions to high concentrations of fructose is feasible if the desirable biocatalytic and thermo­stability properties of suitable enzymes can be realized.

The widely quoted assessment of softwoods is that, as a biomass substrate, the lig- nocellulose is too highly lignified and difficult to process to yield cellulose easily digested by cellulase — in practical terms, excess enzyme is required and imposes unrealistically high costs and protracted digestion times.170 Nevertheless, the mas­sive resources of softwood trees in the Pacific Northwest of the United States, Can­ada, Scandinavia, northern Europe, and large tracts of Russia maintain softwoods as an attractive potential biomass for fuel alcohol production. Sweden has a particular stake in maximizing the efficiency of ethanol production from softwoods on account of the planned diversion of large amounts of woody biomass to direct heat and power facilities with the phasing out of nuclear generating capacity.70 Much of the published work on softwood utilization for bioethanol indeed derives from Canadian and Scan­dinavian universities and research centers.

Although softwoods are low in xylans in comparison with other biomass crops (table 1.5), their content of glucan polymers is high; the requirement for xylose­utilizing ethanologens remains a distinct priority, whereas mannose levels are high and should contribute to the pool of easily utilized hexoses. To make the potential supply of fermentable sugars fully accessible to yeasts and bacteria for fermentation, attention has been focused on steam explosive pretreatments with or without acid catalysts (SO2 or sulfuric acid) since the 1980s; pretreatment yields a mixed pentose and hexose stream with 50-80% of the total hemicellulose sugars and 10-35% of the total glucose, whereas a subsequent cellulase digestion liberates a further 30-60% of the theoretical total glucose.70 Because only a fraction of the total glucose may be recoverable by such technologies, a more elaborate design has been explored in which a first stage is run at lower temperature for hemicellulose hydrolysis, whereas a second stage is operated at a higher temperature (with a shorter or longer heating time and with the same, higher, or lower concentration of acid catalyst) to liberate glucose from cellulose.71 Such a two-stage process results in a sugar stream (before enzyme digestion) higher in glucose and hemicellulose pentoses and hexoses but with much reduced degradation of the hemicellulose sugars and no higher levels of potential growth inhibitors such as acetic acid (figure 4.5).

Two-stage pretreatment suffers from requiring more elaborate hardware and a more complex process management; in addition, attempting to dewater sulfuric acid- impregnated wood chips before steaming decreases the hemicellulose sugar yield from the first step and the glucose yield after the second, higher-temperature stage.72 Such pressing alters the wood structure and porosity, causing uneven heat and mass transfer during steaming. Partial air drying appears to be a more suitable substrate for the dual-stage acid-catalyzed steam pretreatment.

Not only is the sugar monomeric sugar yield higher with the two-stage hydroly­sis process, the cellulosic material remaining requires only 50% of the cellulase for subsequent digestion.73 This is an important consideration because steam-pretreated softwood exhibits no evident saturation with added cellulase even at extremely high enzyme loadings that still cannot ensure quantitative conversion of cellulose to sol­uble sugars: with steam-pretreated softwood material, even lavish amounts of cel- lulase can only liberate 85% of the glucan polymer at nonviably low ratios of solid material to total digestion volume, and although high temperatures (up to 52°C) and high agitation rates are helpful, enzyme inactivation is accelerated by faster mixing speeds.74 The residual lignin left after steam pretreatment probably restricts cellulase attack and degradation by forming a physical barrier restricting access and by bind­ing the enzyme nonproductively; extraction with cold dilute NaOH reduces the lignin content and greatly increases the cellulose to glucose conversion, the alkali possibly removing a fraction of the lignin with a high affinity for the cellulase protein.75

In addition to the engineering issues, the two-stage technology has two serious economic limitations:

1. Hemicellulose sugar recovery is aided after the first step by washing the slurry with water but the amount of water used is a significant cost factor for ethanol production; to balance high sugar recovery and low water usage, a continuous countercurrent screw extractor was developed by the National Renewable Energy Laboratory that could accept low liquid to insoluble sol­ids ratios.76

2. The lignin recovered after steam explosion has a low product value on account of its unsuitable physicochemical properties; organic solvent extraction of lignin produces a higher value coproduct.77

As with other biomass substrates, heat pretreatment at extremes of pH gener­ates inhibitors of microbial ethanol production and, before this, cellulase enzyme action.7879 A number of detoxification methods have been proposed, but simply adjusting the pH to 10 to precipitate low-molecular-weight sugar and lignin deg­radation products is effective.80 Six species of yeasts — including S. cerevisiae, Candida shehatae, and species found in forest underbrush in the western United States — were tested for adaptation to softwood (Douglas fir) pretreated with dilute acid, and isolates were selected and gradually “hardened” to hydrolysate toxicity for improved ethanol production.81

Aqueous ethanol pretreatment of softwoods (the lignol process) has been strongly advocated on account of its ability to yield highly digestible cellulose as well as lignin, hemicellulose, and furfural product streams — the extraction is operated at acid pH at 185-198°C, and some sugar degradation does inevitably occur.82 After process optimization, a set of conditions (180°C, 60 minutes of treatment time, 1.25% v/v sulfuric acid, and 60% v/v ethanol) yielded83

• A solids fraction containing 88% of the cellulose present in the untreated wood chips

• Glucose and oligosaccharides equivalent to 85% of the cellulosic glucose was released by cellulase treatment (48 hours) of the treated lignocellulose

• Approximately 50% of the total xylose recovered from the solubilized fraction

• More than 70% of the lignin solubilized in a form potentially suitable for industrial use in the manufacture of adhesives and biodegradable polymers

5.3.1 Near-Future Projections for Bioethanol Production Costs

The persistently high production costs of lignocellulosic ethanol (particularly in Europe) have catalyzed several attempts to predict trends in 5-year, 10-year, and lon­ger scenarios, with the implicit or explicit rationale that only lignocellulosic biomass is sufficiently abundant to offer a means of substituting a sizeable proportion of the gasoline presently used for transportation.

In 1999, the National Renewable Energy Laboratory published a projection of the economic production costs for lignocellulosic ethanol that, starting from a baseline of $1.44/gallon (in 1997 dollars), computed a decrease to $1.16/gallon with a 12% increase in yield (to 76 gallons/ton of feedstock), with a 12% increase in plant produc­tion capacity together with a 12% reduction in new capital costs.31 A price trajectory envisaged this price deceasing to below $0.80/gallon by 2015 based on developments in cellulase catalytic efficiency and production and in ethanologenic production organisms:

• Improved cellulose-binding domain, active site, and reduced nonspe­cific binding

• Improved cellulase producers genetically engineered for higher enzyme production

• Genetically engineered crops as feedstocks with high levels of cellulases

• Ethanologens capable of producing ethanol at temperatures higher than 50°C

• Ethanologens capable of direct microbial conversion of cellulose to ethanol

These topics were covered in chapters 2 to 4; to a large extent, they remain research topics, although cellulase production and costs have certainly been improved greatly at the industrial scale of production.

A European study suggested that by 2010 lignocellulosic ethanol production costs could decrease to $0.53/l ($2.00/gallon), $0.31/l ($1.18/gallon) by 2020, and $0.21/l ($0.79/gallon) after 2025.32 These improvements in process economics were considered to result from the combined effects of higher hydrolysis and fermenta­tion efficiencies, lower specific capital investment, increases of scale, and cheaper biomass feedstock costs, but the prospect of ethanol ever becoming cost-competitive to gasoline was nevertheless considered to be “unlikely” — much of the analysis was, however, undertaken during the early stages of the great surge in oil prices and gasoline production costs after 2002 (figure 5.5). A more recent publication by the same research group (Utrecht, the Netherlands) computed a 2006 production cost for lignocellulosic ethanol of €22/GJ ($2.00/gallon, assuming exchange parity between the currencies) that was anticipated to fall to €11/GJ ($1.00/gallon) by 2030 — the single largest contributor to the production cost in 2006 was capital-related (46%) but this was expected to decrease in both absolute and relative terms so that biomass costs predominated by 2030.39

The International Energy Agency’s most recent prognosis is for lignocellulosic ethanol (from willow, poplar, and Miscanthus biomass sources) to reach production costs below €0.05/l (€0.18/gallon) by 2030 with achieved biomass yield increases per unit land area of between 44% and 100% (table 5.16); biomass-derived etha­nol would, on this basis, enjoy lower production costs than other sources in Europe (cereal grain, sugarbeet, etc.).34 These three biomass options each target different areas: [51]

decrease in the starch content (from 59.5% to 55%, w/w) can similarly negatively impact the productivity of an ethanol production unit.40

Projecting forward from a 1995 baseline, feedstock costs for a mature biomass ethanol technology were anticipated as being within the price range $34-38.6/dry ton.22 Switchgrass farming has been estimated to cost $30-36/dry tonne; in compari­son with straw or corn stover, the collection of switchgrass is probably less expensive because of the high yield of a denser biomass — nevertheless, delivered costs of the switchgrass to a production plant could be as low as $37/dry tonne of compacted material to $47/tonne of bales.41 Similarly, corn stover is difficult to handle on account of its low bulk density; chopped corn stover can be compacted into briquettes that can reach a density of 950 g/l, and these more easily transported briquettes are more durable if produced from stover with low water content (5-10%).42 A more radical option is the pipeline transport of corn stover (e. g., at 20% wet solids concentration); this is cheaper than trucking at more than 1.4 million dry tonnes/year and allows the possibility of conducting a partial saccharification during transport if enzymes are added, thus reducing the need for investment in the fermentation plant and lowering production costs by 7-80/gallon.43

In general, as bioethanol plant capacity is increased to cut unit production costs, the land area required for collection of sufficient biomass feedstock increases; as biomass supplies are sought from larger and larger distances, the costs of moving the raw material increases, introducing possibly diseconomies into production mod­els. One solution is to introduce more flexibility into the feedstock “diet,” taking advantage of whatever surpluses of other biomass material may seasonally occur; for example, a Californian study investigated what biomass supplies could be considered for a 40-million-gallon facility in the San Joaquin Valley: locally grown corn was significantly more expensive than midwestern corn ($1.21/gallon of ethanol versus $0.92/gallon), but surplus raisins and tree fruit and (although much more expensive) grapes and citrus fruit might all be included in a biomass harvest.44 The proximate example of feedstock diversity is, however, that of cane sugar bagasse: the Dedini Hidrolise Rapida (Rapid Hydrolysis) process uses organic solvent extraction of sug­arcane bagasse as a pretreatment method and aims to double the alcohol production per hectare of sugarcane harvested (www. dedini. com. br).

Using paper sludge as a feedstock for ethanol production has been claimed to be profitable as it provides a near ideal substrate for cellulase digestion after a com­paratively easy and low-cost pretreatment; even without xylose conversion to ethanol, such a technology may be financially viable at small scales, perhaps as low as 15 tonnes of feedstock processed/day.45 This is a genuinely low-cost feedstock because, in the absence of any productive use, paper sludge goes into landfills — at a cost to its producer. This has led to several attempts to find viable means of converting such waste into fuel, and in late 2006, New York state contributed $14.8 million to fund the development of a demonstration facility in Rochester (New York) to use paper sludge as well as wood chips, switchgrass, and corn stover as feedstocks, the facility being operated by Mascoma (www. mascoma. com). The pulp and paper industry in Canada is estimated to produce at least 1.3 million tonnes of sludge every year, and at up to 70% cellulose, this raw material is economically competitive with cereal grain as a substrate for ethanol production.46 Hungarian researchers have also identified paper sludge and other industrial cellulosic wastes as being cost-effective routes to bioethanol.47

This illustrates the proposition that biomass ethanol facilities might be designed and constructed to adventitiously utilize a range of biomass substrates as and when they become available. As in the Californian example noted above, investigation of wastes from fresh and processed vegetables defined a sizeable resources of plant material (450,000 tonnes/year) in Spain; easily pretreated with dilute acid, such inputs could be merged with those for starch or lignocellulosic production lines at minimal (or even negative) cost.48

Substantial cost savings in cereal-based ethanol production can be achieved by a more integrated agronomic approach: although fermentor stillage could be used as a substitute for mineral fertilizer, total ethanol production was 45% lower if cereals were grown after a previous nitrogen-fixing legume crop; intensifying cereal yields certainly increased crops per unit land area, but ethanol production costs per liter dropped as the ethanol yield per unit land area outweighed the other costs; field trials also suggest that barley may (under German conditions) be economically favorable in comparison with wheat and rye).49 All these conclusions may be directly applicable to cereal straw as well as cereal grain. Moreover, since one of the strongest candidates for lignocellulosic ethanol production (wheat straw) also has one of the lowest ethanol yields per unit dry mass, the ability to flexibly mix substrates has capacity advantages if feedstock handling and processing regimes can be harmonized (figure 5.6).50

Once in the ethanol facility, biomass pretreatment and hydrolysis costs are important contributors to the total production cost burden; for example, with hard­woods and softwoods, enzymes represent 18-23% of the total ethanol production costs; combining enzyme recycling and doubling the enzyme treatment time might

improve the economic cost by 11%.51 Bench-scale experiments strongly indicated that production costs could be reduced if advanced engineering designs could be adopted, in particular improving the efficiency of biomass pretreatment with dilute acid in sequential cocurrent and countercurrent stages.52 Different pretreatment techniques (dilute acid, hot water, ammonia fiber explosion, ammonia recycle per­colation, and lime) are all capital-intensive, low-cost reactors being counterbalanced by higher costs associated with catalyst recovery or ethanol recovery; as a result, the five rival pretreatment options exhibited very similar production cost factors.53 Microbial pretreatments could greatly rescue the costs and energy inputs required by biomass hydrolysis techniques because enzymic digestibility is increased and hydrodynamic properties improved in stirred bioreactors; a full economic analysis remains to be undertaken in industrial ethanol facilities.54 Continued optimization of chemical pretreatments has resulted in a combined phosphoric acid and organic solvent option that has the great advantage of requiring a relatively low temperature (50°C) and only atmospheric pressure.55

Being able to generate higher ethanol concentrations in the fermentation step would improve overall economic performance by reducing the costs of ethanol recov­ery; one solution is to run fermentations at higher biomass substrate loadings, that is, accomplish biomass liquefaction and saccharification at high solids concentrations; wheat straw could be processed to a paste/liquid in a reactor system designed for high solids content, the material then being successfully fermented by Saccharomy- ces cerevisiae at up to 40% (w/v) dry matter in the biological step.56

What is the optimum processing of the various chemical streams (soluble sugars and oligosaccharides, pentoses and hexoses, cellulose, and lignin) resulting from the pretreatment and hydrolysis of lignocellulosic biomass? The utilization of pentose sugars for ethanol production is certainly beneficial for process economics, a conclu­sion reached as early as 1989 in a joint U. S.-New Zealand study of pine as a source of woody biomass, where ethanol production costs of $0.75/l ($2.83/gallon) were cal­culated, decreasing by 5% if the pentose stream was used for fermentable sugars.57 The stillage after distillation is also a source of carbohydrates as well as nutrients for yeast growth; replacement of up to 60% of the fresh water in the fermentation medium was found to be possible in a softwood process, with consequential reduc­tions in production costs of as much as 17%.58

Because larger production fermentors are part of the drive toward economies of scale savings in production costs, reformulating media with cheaper ingredients becomes more important. In the fermentation industry at large, devising media to minimize this operating cost parameter has had a long history; as recombinant eth — anologens are increasingly engineered (chapter 3), suitable media for large scales of production are mandatory. With E. coli KO11, for example, laboratory studies showed that expensive media could be substituted by a soya hydrolysate-containing medium, although the fermentation would proceed at a slightly slower rate; both this and a corn steep liquor-based medium could contribute as little as 60/gallon of ethanol produced from biomass hydrolysates.5960

Operating a membrane bioreactor (in a demonstration pilot plant of 7,000-l capacity) showed that the yearly capital costs could be reduced to $0.18-0.13/gallon, with total operating costs for the unit of $0.017-0.034/gallon.61 Another advanced engineering design included continuous removal of the ethanol product in a gas stream; compared with a conventional batch process, ethanol stripping gave a cost saving of $0.03/gallon with a more concentrated substrate being used, thus resulting in less water to remove downstream.62 Simply concentrating the fermented broth if the ethanol concentration is low in a conventional process is feasible if reverse osmo­sis is employed but not if the water is removed by evaporation with its high energy requirement.63

Incremental savings in bioethanol production costs are, therefore, entirely pos­sible as processes are evolved; many of the steps involved obviously require higher initial investment when compared with basic batch fermentation hardware, and it is unlikely that radical innovations will be introduced until a firm set of benchmark costings are achieved in semiindustrial and fully production-scale units. Efficient utilization and realization of the sales potential of coproducts remains, on the other hand, an immediate possibility. Coproduct credits have long been an essential fea­ture of estimates of ethanol production (section 5.2); among these, electricity genera­tion has been frequently regarded as readily engineered into both existing and new ethanol production facilities, especially with sugarcane as the feedstock for ethanol production. For example, in Brazil steam turbines powered by combustion of sugar­cane bagasse can generate 1 MWh/m3 (1,000 l) of ethanol, and economic analysis shows that this is viable if the selling price for electricity is more than $30/MWh, the sales price in Brazil in 2005.64 When electricity credits cannot be realized or where coproducts can be used as substrates for other chemical or biological processes, dif­ferent criteria come into play — this is discussed in chapter 8 (section 8.2). Pyrolysis of sugarcane wastes to produce “bio-oil” could yield 1.5 tons of saleable products per ton of raw sugar used, but the selling cost of the product will be crucial for establish­ing a viable coproduction process.65

The potential of biorefineries to generate a wider collection of chemical feedstocks presently derived from petroleum is best visualized with biomethanol as the starting point (figure 8.3). Methanol obtained from the gasification of biomass (chapter 6, section 6.3.1) can be transformed by well-known purely chemical reactions to form, among many other chemicals:8

• Formaldehyde (CH2O) is used in the production of resins, textiles, cosmet­ics, fungicides, and others; formaldehyde accounted for 35% of the total worldwide production of methanol in the mid-1990s and is prepared by the oxidation of methanol with atmospheric O2 using a variety of catalysts.

• Acetic acid (CH3COOH) is a major acid in the food industry and as feed­stock for manufacturing syntheses, in the production of some plastics, fibers, and others; acetic acid is manufactured by the carbonylation of methanol with CO.

• Formic acid (HCOOH) is a preservative.

• Methyl esters of organic and inorganic acids are used as solvents and meth — ylation reagents and in the production of explosives and insecticides.

• Methylamines are precursors for pharmaceuticals.

• Trimethylphosphine is used in the preparation of pharmaceuticals, vita­mins, fine chemicals, and fragrances.

• Sodium methoxide is an organic intermediate and catalyst.

• Methyl halides are solvents, organic intermediates, and propellants.

• Ethylene is used for plastics and as an organic intermediate.

3.5.1 Ethanol Recovery from Fermented Broths

The distillation of ethanol from fermented broths remains the dominant practice in ethanol recovery in large and small ethanol production facilities.250 Other physical techniques have been designated as having lower energy requirements than simple distillation, and some (vacuum dehydration [distillation], liquid extraction, super­critical fluid extraction) can yield anhydrous ethanol for fuel purposes from a dilute aqueous alcohol feed (figure 4.12).251 Only water removal by molecular sieving* has, however, been successful on an industrial scale, and all new ethanol plants are built with molecular sieve dehydrators in place.252

Nevertheless, the economic costs of dehydration are high, especially when anhydrous ethanol is to be the commercial product (figure 4.12). In the early 1980s, the energy requirements were so high that the practical basis for fuel ethanol pro­duction was questioned because the energy required for distillation approximated the total combustion energy of the alcohol product.253,254 The investment costs of rivals to distillation were, however, so high (up to 8.5 times that of conventional distillation) that little headway was made and attention was focused on improving

Synthetic zeolite resins are crystalline lattices with pore sizes of 0.3 nm, sufficiently small to allow the penetration of water molecules (0.28 nm in diameter) but exclude ethanol (molecular diameter 0.44 nm).

process efficiencies and energy cycling with the development of low-energy hydrous ethanol distillation plants with 50% lower steam-generating requirements.252

The economics of downstream processing are markedly affected by the concen­tration of ethanol in the fermented broth; for example, the steam required to produce an ethanol from a 10% v/v solution of ethanol is only 58% of that required for a more dilute (5% v/v) starting point, and pushing the ethanol concentration in the fermenta­tion to 15% v/v reduces the required steam to approximately half that required for low conversion broth feeds.253