Inevitably, an essential facet of the public discussions on costs and subsidies of biodiesel production has been that of its potential amelioration of greenhouse gas emissions. If significant, this would augment the case for production and consump­tion incentives to offset higher production costs than for conventional diesel. At the scientific level, this debate has mirrored that for bioethanol (chapter 1, section 1.6) and has proved equally contentious and acrimonious.

In the early 1990s, net energy balance (NEB) values of up to 3.8:1 were calculated for rapeseed-derived biodiesel, depending on how the coproducts and crop straw were assessed in the calculations (figure 6.6).4 Unpublished reports and communica­tions quoted in that report were from 1.3 to 2.1 without coproduct credits and from 2 to 3 if thermal credits for the meal and glycerol coproducts were included. Radi­cally different conclusions were reached in a 2005 publication: biodiesel production

2.0

from soybean oil required 27% more fossil energy than the biodiesel energy content, whereas sunflower oil was even less viable (requiring 118% more fossil energy than in the product).49

Midway (in time) between these conflicting estimates was a report from the National Renewable Energy Laboratory whose main conclusion was that biodiesel (from soybean oil) yielded 3.2 units of fuel product energy for every unit of fossil energy consumed in its life cycle, whereas conventional diesel yielded only 0.83 unit per unit of fossil fuel consumed, that is, that biodiesel was eminently “renew­able.”55 Direct comparison of these conflicting results shows that the disagreements are major both for the stages of soybean cultivation and biodiesel production (fig­ure 6.6). As so often in biofuel energy calculations, part of the discrepancy resides in how the energy content of coproducts is allocated and handled in the equations (chapter 1, section 1.6.1). As the authors of the 2005 study pointed out, if the energy credit of soybean meal is subtracted, then the excess energy required for biodiesel production falls to 2% of the biodiesel energy content.49 A paper posted on the Uni­versity of Idaho bioenergy site suggests other factors:56 [60]

• Even adding in energy requirements for oil transport and transesterification as well as biodiesel transport produces a favorable energy balance of 2.9:1.

The energy balance, in any of these scenarios, is highly dependent on viewing the process as a biorefinery producing coproducts as well as biodiesel. If the energetic (and economic) value of the soybean meal cannot be realized, then the balance will be negative — even using the soybean meal as a “green manure” spread on the soybean fields would only partially offset the major loss of replaced fossil energy in the total process. If biodiesel production from oil seed crops is viewed as an “opportunity” over their alternative uses as, for example, foodstuffs, then the NEB can be recalculated to be favorable.57 As discussed previously (chapter 1, section 1.6.1), this is a contentious argument, and much media comment in Europe dur­ing 2006-7 has pointed to increased areas of arable land being devoted to oilseed rape as demand for the crop as a source of biodiesel increases, thus fulfilling the prediction of a subsidized cash crop.4 A focus of future attention may be that of realizing an economic return on the greatly increased amounts of seed meal and of finding a viable use for glycerol — refining the glycerol coproduct to a chemically pure form is expensive, and alternative uses of glycerol for small — and medium — scale biodiesel facilities are being explored, for example, its use as an animal feed supplement.58

As the number of industrial units producing biodiesel increases, assessments of energy balances should be possible from collected data rather than from calculations and computer simulations. A report on activities in six Brazilian and Colombian biodiesel facilities using palm oil as the agricultural input attempted precisely this.59 NEBs were in the range of 6.7-10.3, with differences arising because of

• Different rates of fertilizer application

• Different uses of plant residues as fertilizers or as boiler fuel for electric­ity production

• On-site electricity generation at some sites, whereas others were entirely dependent on purchased electricity

• Differing efficiencies in the generation of coproducts and the recovery of unused palm oil

Taken together as a group, these palm oil biodiesel producers were assessed as being more energy efficient than reference manufacturers in Europe or the United States — the most recent (2006) detailed estimate of biodiesel from soybean oil in the United States arrived at a NEB of 1.93:1, but this was critically dependent on full credits being taken for soybean meal and glycerol coproducts (without them, the balance decreases to only 1.14:1).60

The energy balance is an important parameter that defines the extent of the biodiesel’s capacity to reduce greenhouse gas emissions, because, in the extreme case, if biodiesel requires more fossil energy in its production than can be usefully recovered in the product, no savings could possibly accrue.49 With a favorable energy balance for soybean biodiesel, its use could displace 41% of the greenhouse gas emissions relative to conventional diesel.60 As headline statements, the National

Renewable Energy Laboratory study on biodiesel use for public transport concluded the following:55

1. Substituting 100% biodiesel (B100) for petroleum diesel reduced the life cycle consumption of petroleum by 95%, whereas a 20% blend (B20) reduced consumption by 19%.

2. B100 reduced CO2 emissions by 74.5%, B20 by 15.7%.

3. B100 completely eliminated tailpipe emissions of sulfur oxides and reduced life cycle emissions of CO, sulfur oxides, and total particulate matter by 32%, 35%, and 8%, respectively.

4. Life cycle emissions of NO* and hydrocarbons were higher (13.4% and 35%, respectively) with B100, but there were small reductions in methane emissions.

Earlier assessments indicated that only 55% of the CO2 emitted from fossil diesel could be saved if biodiesel were to be used because of the CO2 emissions inher­ent in the production of biodiesel and that, other than a marked reduction in sulfur oxides, effects on CO, hydrocarbons, NO*, and polyaromatic hydrocarbons were inconsistent.4 As the use of biodiesel has widened globally, the number of publi­cations exploring individual pollutants or groups of greenhouse gas emissions has expanded, especially after 2000 (table 6.5).61-68 The report of increased mutagenicity in particulate emissions with a biodiesel is unusual as two earlier reports from the same research group in Germany found reduced mutagenicity with rapeseed oil — and soybean-derived biodiesels.66 6970 A high sulfur content of the fuel and high engine speeds (rated power) and loads were associated with an increase in mutagenicity of diesel exhaust particles. This is in accord with the desirability of biodiesels because of their very low sulfur contents, zero or barely detectable, as compared with up to 0.6% (by weight) in conventional diesels.4 There are suggestions that exhaust emis­sions from biodiesels are less likely to present any risk to human health relative to petroleum diesel emissions, but it has been recommended that the speculative nature of a reduction in health effects based on chemical composition of biodiesel exhaust needs to be followed up with thorough investigations in biological test systems.71

5.2.7.1 Complete Process Cost Models

Acknowledging the great uncertainties in establishing guideline costs for lignocel — lulosic feedstocks, a Swedish review of the bioethanol industry in Scandinavia and elsewhere in 1999 noted that large-scale processes and improved overall ethanol yield would be highly desirable for future economic production of biofuels.28 The first trend probably heralded the demise of the farm-scale ethanol plant (section 5.2.1.4), a production model that probably is only relevant to local and private consumption

[50] Data published by the DOE’s Energy Information Administration (www. eia. doe. gov and www. tonto. eia. doe. gov) point to refinery gate prices of gasoline being 70% of the retail price (excluding taxes), although this increases to 90% or more during crude oil price surges; figure 5.5 was constructed from the U. S. conventional gasoline bulk prices after January 1994 and, between 1983 and 1994, the cor­responding retail prices x 0.70.

[51] Willow already has considerable commercial experience in Sweden, the United Kingdom, and elsewhere and has strong potential in Eastern Europe where growing conditions and economics are favorable.

• Poplar is already grown for pulp production, with typical rotation cycles of eight to ten years.

[52] Fuel ethanol production was initiated from cereal grain feedstocks, but a government-set selling price of $1.65/gallon required large subsidies because the production costs were more than $1.80/gallon.

• Total biomass production could exceed the International Energy Agency’s prediction for transportation fuel needs by 2030 at a low feedstock cost ($22/dry ton).

[53] Crop productivity is projected to increase slightly at mid to high latitudes for local mean temperature increases of up to 1-3°C depending on the crop and then decrease beyond that in some regions.

• At lower latitudes, especially seasonally dry and tropical regions, crop pro­ductivity is projected to decrease for even small local temperature increases (1-2°C), which would increase risk of hunger.

• Globally, the potential for food production is projected to increase with increases in local average temperature over a range of 1-3°C, but above this, it is projected to decrease.

• Adaptations such as altered cultivars and planting times allow low and mid to high cereal yields to be maintained at or above baseline yields for modest warming.

• Increases in the frequency of droughts and floods are proj ected to affect local production negatively, especially in subsistence sectors at low latitudes.

[54] South African businesses are planning to open 18 corn-based ethanol production sites within the national borders by 2012.

a counterbalance to the lure of accelerated growth and development as absolute, uncontested priorities in emerging/developing/transitional economies. The reality, however, is more likely to be close insistence on a reliability of bioethanol supply to Europe, Japan, and the United States, despite fluctuating growing conditions from year to year (as occurred, for example, in the dip in ethanol production in Brazil in 2001-2002): consumers with flexibly fueled vehicles will, under those conditions, turn back to gasoline and diesel, thus inhibiting the progress of biofuels programs.106 Although short-lived oil gasoline price bubbles are tolerated (in the absence of any real choice), shortfalls in fuel ethanol supply are unlikely to be; considerations such as those of encroachment on to virgin and marginal land in the tropics, deforestation, and loss of soil carbon may be much lower down the agendas of publics and policy­makers inside and outside the OECD.

Interest in Erwinia bacteria for ethanol production dates back at least to the late 1950s; in 1971, the explanation for the unusually high ethanol production by Erwinia species was identified as a PDC/ADH pathway, decarboxylating pyruvate to acetaldehyde fol­lowed by reduction to ethanol, akin to that in Z. mobilis; ethanol is the major fermenta­tive product, accompanied by smaller amounts of lactic acid.224 Soft-rot bacteria secrete hydrolases and lyases to solubilize lignocellulosic polymers, and the PET operon was used to transform E. carotovora and E. chrysanthemi to produce ethanol from cellobi — ose, glucose, and xylose; both strains fermented cellobiose at twice the rate shown by cellobiose-utilizing yeasts.42 The genetically engineered E. chrysanthemi could ferment sugars present in beet pulp but was inferior to E. coli strain KO11 in ethanol production, generating more acetate and succinate in mixed-acid patterns of metabolism.221

Lactococcus lactis is another GRAS organism; its use in the industrial produc­tion of lactic acid is supplemented by its synthesis of the bacteriocin nisin, the only such product approved for food preservation.225 When a PDC-encoding gene from Zymobacter palmae was inserted into L. lactis via a shuttle vector, the enzyme was functionally expressed, but, although a larger amount of acetaldehyde was detected, a slightly higher conversion of glucose to ethanol was measured (although glucose was used more slowly), and less lactic acid was accumulated, no increased ethanol production could be achieved, presumably because of insufficient endogenous ADH activity.226 The same group at USDA’s National Center for Agricultural Utilization Research examined L. plantarum as an ethanologen for genetic improvement; strain TF103, with two genes for lactate dehydrogenase deleted, was transformed with a PDC gene from the Gram-positive bacterium Sarcinia ventriculi to redirect carbon flow toward ethanol production, but only slightly more ethanol was produced (at up to 6 g/l).227 Other attempts to metabolically engineer lactic acid bacteria have been similarly unsuccessful (although more ethanol is produced than by the parental strains and the conversion of glucose to ethanol is increased by nearly 2.5-fold), the bacteria remaining eponymously and predominantly lactic acid producers; although Z. mobilis pdc and adh genes in PET operons are transcribed, the enzyme activities can be very low when compared with E. coli transformants.228 229

Zb. palmae was isolated on the Japanese island of Okinawa from palm sap by sci­entists from the Kirin Brewery Company, Yokohama, Japan. A facultative anaerobe, the bacterium can ferment glucose, fructose, galactose, mannose, sucrose, maltose, melibiose, raffinose, mannitol, and sorbitol, converting maltose efficiently to ethanol with only a trace of fermentative acids.230 Its metabolic characteristics indicate potential as an ethanologen; broadening its substrate range to include xylose followed previous work with Z. mobilis, expressing E. coli genes for xylose isomerase, xylulokinase, transaldolase, and transketolase.231 The recombinant Zb. palmae completely cofer­mented a mixture of 40 g/l each of glucose and xylose simultaneously within eight hours at 95% of the theoretical yield. Introducing a Ruminococcus albus gene for P — glucosidase transformed Zb. palmae to cellobiose utilization; the heterologous enzyme was more than 50% present on the cell surface or inside the periplasm, and the recom­binant could transform 2% cellobiose to ethanol at 95% of the theoretical yield.232 The PDC enzyme of the organism is, as discussed briefly above, an interesting target for het­erologous expression in ethanologenic bacteria; it has the highest specific activity and lowest affinity for its substrate pyruvate of any bacterial PDC, and it has been expressed in E. coli to approximately 33% of the soluble protein. Codon usage for the gene is quite similar to that for E. coli genes, implying a facile recombinant expression.233

Cyanobacteria (blue-green algae) have generally lost their fermentative capabilities, now colonizing marine, brackish, and freshwater habitats where photosynthetic metab­olism predominates; of 37 strains in a German culture collection, only five accumulated fermentation products in darkness and under anaerobic conditions, and acids (glycolic, lactic, formate, and oxalate) were the major products.234 Nevertheless, expression of Z. mobilis pdc and adh genes under the control of the promoter from the operon for the CO2-fixing ribulose 1,5-bis-phosphate carboxylase in a Synechococcus strain synthesized ethanol phototrophically from CO2 with an ethanol:acetaldehyde molar ratio higher than 75:1.235 Because cyanobacteria have simple growth nutrient require­ments and use light, CO2, and inorganic elements efficiently, they represent a sys­tem for longer-term development for the bioconversion of solar energy (and CO2) by genetic transformation, strain and process evolution, and metabolic modeling. The U. S. Department of Energy (DOE) is funding (since 2006) DNA sequencing studies of six photosynthetic bacteria at Washington University in St. Louis, Missouri, and the DOE’s own sequencing facility at Walnut Creek, California, using a biodiversity of organisms from rice paddies and deep ocean sources to maximize biochemical and metabolic potential.

Economic considerations have featured in both primary analyses and reviews of the biotechnology of fuel ethanol production published in the last 25 years.11 Because a lignocellulosic ethanol industry has yet to fully mature, most of those studies have been derived from laboratory or (at best) small pilot-plant data, and estimates for feedstock and capital investment costs have varied greatly, as have assumptions on

the scale of commercial production required to achieve any intended price/cost target for the product. As with estimates of net energy yield and greenhouse gas reductions (chapter 1, sections 1.6.1 and 1.6.2), the conclusions reached are heavily influenced by the extent to which costs can be offset by coproduct generation (as a source of income) and of the complexity of the total production process, not only as a primary cause for increased setup costs but also a potential source of process efficiencies and additional, saleable coproducts.

Few of the influential studies are full business models for bioethanol, in particular avoiding any computations for profitability, often because the main driver has been to establish and substantiate grounds for initial or continued investment by national and/or international funding agencies — and with the implicit assumption that any production process for fuel ethanol outside Brazil suffers by that very comparison because of the lack of such favorable climatic and economic features (in particular, land use, labor cost, and the dovetailing of ethanol production with a fully mature sugarcane industry). Nevertheless, a historical survey of key points in the develop­ment of the economic case for bioethanol reveals the convergence toward a set of key parameters that will be crucial for any biofuel candidate in the next 10-50 years.

7.3.1 The Hydrogen Economy and Fuel Cell Technologies

The International Energy Agency, in its 2006 review of world energy trends, fore­casts that by 2030, hydrogen-powered vehicles may have begin to “decarbonize” transportation — if, that is, production from low — and zero-carbon sources develops,
if there are breakthroughs in hydrogen storage and if the necessary infrastructure (requiring huge investments) develops.16 The chemistry of hydrogen combustion entirely avoids greenhouse gas emissions:

2H2 + O2 ^ 2H2O

whether this occurs in thermal power generation or in any of the presently developed types of hydrogen fuel cell (table 7.1).17

In principle, generating H2 from the most abundant potential source—water—is eminently straightforward, that is, the electrolysis of water, but the smallest amount of electricity that can produce 1 mol of H2 from 1 mol of water is 237 kJ, whereas the amount of heat generated by the combustion of H2 is 285.6 kJ.18 Although fossil fuels are the main source of electric power generation, switching to the “hydrogen economy” will only be an inefficient means of reducing the emissions of CO2 and other greenhouse gases because the net energy balance of this route to H2 production is no more favorable than that often calculated for corn-derived ethanol (chapter 1, section 1.6.1). Other routes are known, for example the direct thermal decomposition of water, and thermochemical, photochemical, and photoelectrochemical technolo­gies, but how “green” the resulting H2 production is depends critically on the mix of fossil and nonfossil inputs used for power generation, locally or nationally.

The Division of Technology, Industry, and Economics of the United Nations Environment Programme noted in its 2006 review that publicly funded research into hydrogen technologies was intensive in OECD nations (figure 7.4).19 Both OECD countries and a growing number of developing economies have active “hydrogen economy” targets:

• Japan was the first country to undertake an ambitious fuel cell program,

10 years of R&D funded at $165 million, completed in 2002; following on, the New Hydrogen Project focuses on commercialization, funding reaching $320 million in 2005 and with the aims of producing and supporting 50,000 fuel cell-powered vehicles by 2010 and 5 million by 2020 (with 4000 H2 refueling stations by then), 2,200 MW of stationary fuel cell cogeneration systems by 2010, and 10,000 MW by 2020.

300

• The transition to the hydrogen economy envisaged by the U. S. government (the Hydrogen Fuel Initiative) is set to proceed via four phases, of tech­nology development, initial market penetration, infrastructural investment, and full realization to begin by 2025.

• Funding for the hydrogen economy in the European Union was provided by the Renewable Energy Sixth Framework program from 2003 to 2006, and subsequent plans are expected to generate combined public and private funding of approximately $2.8 billion by 2011.

• Canada’s H2 R&D focuses on the Ballard PEM fuel cell and the Hydrogenics alkaline water electrolyser, with public funding of more than $25 million/year.

• Korea has budgeted $586 million for hydrogen-related projects through to 2011, aiming at the introduction of 10,000 fuel cell vehicles, development of H2 production from renewable resources, and development of a 370-MW capacity stationary fuel cell.

• India allocated $58 million from 2004 to 2007 for projects in universities and governmental research laboratories, with car manufacturers expected to contribute $116 million by 2010.

• Russia began to fund ajoint project between the Russian Academy of Science and the Norlisk Nikel Company at $30 million in 2005 on fuel cell development.

• Brazil’s Hydrogen Roadmap focuses on production from water electrolysis, reforming of natural gas, reforming or gasification of ethanol and other bio­fuels, storage technologies (including metal hydrides), and fuel cells.

In 2007, the first liquid H2-dispensing fuel pump was installed in Norway as the first step in providing the “H2 highway,” a 360-mile route from Stavanger to Oslo
that is expected to be complete by 2009, whereas California launched a H2 highway network to include up to 200 fueling stations by 2010.

As with corn stover, bagasse is the residue from sugarcane juice extraction and, as such, is an obligatory waste product. Development programs for bioethanol produc­tion from bagasse started in Brazil in the 1990s, and Dedini S/A Industrias de Base (www. dedini. com. br) now operates a pilot plant facility with the capacity of produc­ing 5,000 l/day of ethanol from bagasse in Sao Paolo state; up to 109 l of hydrated alcohol can be produced per tonne of wet bagasse, and this could be increased to 180 l/tonne with full utilization of hemicellulose sugars.

The commercial Brazilian process uses organic-solvent-treated bagasse. Other reports from Cuba, Denmark, Sweden, Japan, Austria, Brazil, and the United States describe alternative processes, all with some merits:

• Steam explosion — impregnation with SO2 before steam explosion gives high yields of pentose sugars with no additional formation of toxic inhibi­tors when compared with the absence of any acid catalyst84-86

• Liquid hot water pretreatment — probably the cheapest method (requiring no catalyst or chemical) and, when operated at less than 230°C, is effec­tive at solubilizing hemicellulose and lignin while leaving cellulose as an insoluble residue for further processing87,88

• Peracetic acid — alkaline pre-pretreatment followed by the use of peracetic acid gives synergistic enhancements of cellulose digestibility89,90

• Ammonia-water mixtures — vacuum-dried material from the alkaline treatment could be used for enzymatic digestion of cellulose without wash­ing or other chemical procedure34

• Dilute acid — this has not yet been fully tested for ethanol production, but as a method for preparing xylose as a substrate for xylitol production, it is capable of yielding hydrolysates with high concentrations of free xylose91,92

• Wet oxidation — alkaline wet oxidation at 195°C for 15 minutes produces solid material that is 70% cellulase; approximately 93% of the hemicellu — loses and 50% of the lignin is solubilized, and the cellulose can be enzymi — cally processed to glucose with 75% efficiency93

Three different species of yeasts have been demonstrated to ferment pentoses and/or hexoses from chemical hydrolysates of sugarcane bagasse: S. cerevisiae, C. shehatae, Pichia stipitis, and Pachysolen tannophilus.94-97 Acid hydrolysates are

best detoxified by ion exchange materials or activated charcoal; laccase and high pH precipitation methods are less effective.95 The most recent of the reports attempted lime treatment to neutralize the acid but found that the novel technique of electrodi­alysis (migration of ions through membranes under a direct electric field) removed the sulfuric acid and also the acetic acid generated during acid hydrolysis of hemicellu — loses so effectively that the reutilization of the sulfuric acid could be contemplated.[35]97 Recombinant xylose-utilizing yeast has been desensitized to hydrolysates containing increasing concentrations of phenolic compounds, furfuraldehydes, and carboxylic acids without loss of the xylose-consuming capacity and while retaining the ability to form ethanol rather than xylitol.96 Acetic acid and furfural at concentrations similar to those measured in sugarcane bagasse hydrolysates adversely affect both “laboratory” and “industrial” strains (see chapter 3, section 3.2.4) of S. cerevisiae.98 One highly practical solution is that the predominantly pentose-containing hydrolysates from bagasse pretreatments can also be used to dilute the sugarcane juice-based medium for sugar ethanol fermentation while maintaining an equivalent sugar concentration and utilizing a pentose-consuming P. stipitis to coferment the sugar mixture.99

Sugarcane is, however, not entirely without its industrial biohazards. Bagassosis, caused by airborne cells (or fragments) of Thermoactinomyces sacchari, was once very prevalent in workforces exposed to bagasse dust. In the United States and also in Japan (where outbreaks occurred in sugar refineries and lacquerware factories), the disease is thought to have been mostly eradicated during the 1970s by improved product handling and safety practices.100,101

China’s demand for oil as a transportation fuel is forecast to increase more than ten­fold between 1990 and 2030 (from 30 to 396 million tonnes), reaching 50% of that of the United States by that date.66 Economic analysis has shown that sweet sorghum and its bagasse as well as rice hulls and corn stover have extensive availability in northern China and could represent attractive feedstocks for bioethanol production.67

Investigations into gasoline supplementation with endogenously produced etha­nol began in 1999, and by 2004, E90 grades were available in eight provinces; a Renewable Energy Law and a National Key R&D Program for cellulosic ethanol were applied to the energy sector during 2005.68 Other primary factors in China’s newly acquired interest in bioethanol include the following: [52]

• Assuming successful implementation of the types of lignocellulosic ethanol technologies on which recent U. S. and European cost models have been based (section 5.2), production cost estimates for Chinese production sites would be in the range of $0.43-0.95/gallon.

Building and operating commercial cellulosic ethanol plants in China thus appears very feasible and would generate exactly the kind of practical experience and knowledge that would induce other nations to invest. The Chinese government has announced the allocation of $5 billion in capital investment in the coming decade for ethanol production capacity with a focus on noncereal feedstocks.68

Ethanol is readily dehydrated by chemical reactions to ethylene (ethene):

C2H5OH ^ C2H4 + H2O

in a reversal of the chemistry used to manufacture “industrial” alcohol from petrochemical sources. Ethylene per se is a key intermediate in organic chemistry for plastics (polyethylenes); in 2005, its worldwide production was estimated to be 113 million tonnes.9 From ethylene (a compound with no direct end uses), the vast “hinterland” of petrochemical production rapidly opens up by way of

• Ethylene oxide (C2H4O), a starting material for the manufacture of acrylo­nitrile, nonionic surfactants, and others, and a ripening agent for fruits

• Ethylene glycol (C2H5O2), a solvent as well as an intermediate in the syn­thesis of synthetic fibers

• Ethylene chlorohydrin (C2H5ClO), another solvent and an intermediate in the production of agrochemicals

• Ethyl bromide (C2H5Br), an ethylating agent in organic syntheses

Approximately 0.9 million tonnes of glycerol is synthesized via chemical routes annu­ally; two-thirds of this could be isolated from bioethanol fermentation broths as glyc­erol-rich stillage.3 Biodiesel wastes and direct fermentations could, however, supply far more glycerol (chapter 6, section 6.3.2) — in fact, so much glycerol is presently being produced as a coproduct of biodiesel production that, although not a biological process, biodiesel manufacture is an imminent development of the biorefinery concept (see section 8.3.3). Important derivatives of glycerol prepared chemically include10

• Glycerol trinitrate, that is, nitroglycerin for explosives

• Epichlorohydrin, the most important material for the production of epoxy resins (and historically the immediate precursor of glycerol in chemical manufacture)11

• Oxidation products as intermediates for pharmaceutical synthesis, including glyceric, tartronic, hydroxypyruvic, and mesoxalic acids and dihydroxyacetone

• Tertbutyl ethers formed by the reaction between glycerol and alkenes such as 2-butene

Approximately 36% of global glycerol production is directed toward such chemi­cal intermediates and products, most of the remainder finding uses in the food and cosmetic industries.

Butanol (chapter 6, section 6.3.3) is a major solvent and finds numerous applica­tions during the manufacture of plastics and textiles.11 Butanol could be catalytically dehydrated to yield l-butene, one of four butene isomers (1, cis-2, trans-2, and iso forms); none of these highly volatile hydrocarbons exist naturally but are produced from petroleum refineries; the complexity of the composition of the butene-contain­ing C4 streams from crude oil “cracking” has limited the chemical exploitation of butanes as intermediates and feedstocks.12 The availability of pure 1-butene in feed­stock quantities may herald new horizons for industrial synthetic chemistry.

These examples demonstrate one fundamental feature of biorefineries, that is, the ability to replace petroleum refineries as sources of both liquid fuels (ethanol, biobutanol, etc.) and (by catalytic transformation of these compounds) feedstock chemicals. This model implies a short linear chain of sequential biochemical and chemical processes:

biomass substrate ^ ethanol, methanol, glycerol, butanol ^ feedstock chemicals

This could operate on an either/or basis, either producing biofuels for transportation uses or proceeding straight to the production of bulk chemicals. The development of a full infrastructure for biofuels distribution would bring in parallel a means of transporting the liquid biofuels to existing chemical industrial sites, thus reducing the total investment cost associated with the transition from oil dependency to a biobased commodity economy, and, although the economic analysis of biorefiner­ies is poorly developed, the price tag will be high: one estimate puts the cost of a biorefinery capable of processing 2000 tonnes/day as $500 million.13 Approximately 500 such facilities will be required to process the “billion tons” annually required for the mass production of lignocellulosic ethanol in the United States (chapter 2, section 2.7), thus representing a $200 billion total investment.

A more elaborate model for biorefineries entails the dual bioproduction of biofuels and other fermentation products.14 The logistic basis for this design is the multiple nature of lignocellulosic and whole-plant carbohydrate streams (chapter 2, section 2.3), that is, hemicellulose-derived pentoses, cellulose-derived glucose and oligoglucans, and starch-derived glucose if grains are processed. Given the wide spectrum of naturally occurring and genetically engineered ethanologens (chapter 3), ethanol could be produced entirely from one of these carbohydrate streams, leaving the others as substrates for different types of fermentations — potentially as wide a choice as that of fermentations known or already used for industrial production. To substantially narrow the field, twelve building block chemicals that can be produced from sugars via biological or chemical conversions and that can be subsequently converted to a number of high-value bio-based chemicals or materials have been identified in a report pre­pared for the U. S. DOE.15 These building block chemicals are molecules with multiple functional groups that can transformed into new families of useful molecules:

• 1,4-Diacids (succinic, fumaric, and malic), all intermediates of the tricar­boxylic acid cycle and easily bioproduced by microbes

• 2,5-Furan dicarboxylic acid, chemically produced by the oxidative dehy­dration of C6 sugars

• 3-Hydroxypropionic acid, a microbial product

• Aspartic acid, an amino acid biosynthesized by all living organisms

• Glutamic acid, another amino acid and one of the major products of indus­trial fermentations for fine chemicals (as monosodium glutamate, MSG)

• Glucaric acid, chemically produced by the nitric acid oxidation of starch

• Itaconic acid, a tricarboxylic acid manufactured on an industrial scale by the fungus Aspergillus oryzae

• Levulinic acid, chemically produced by the acid-catalyzed dehydration of sugars

• 3-Hydroxybutyrolactone, chemically produced by the oxidative degrada­tion of starch by hydrogen peroxide

• Glycerol (chapter 6, section 6.3.2)

• Sorbitol, a sugar alcohol derived from glucose (chemically by hydrogenation) but also known as an enzyme-catalyzed product of glucose metabolism

• Xylitol/arabinitol (chapter 3, section 3.2)

These compounds were chosen by consideration of the potential markets for “build­ing blocks” and their derivatives and the technical complexity of their (bio)synthetic pathways. A second-tier group of building blocks was also identified as viable candi­dates, most of which are microbial products: gluconic acid, lactic acid, malonic acid, propionic acid, the two triacids, citric and aconitic, xylonic acid, acetoin, furfural, levoglucosan, and the three amino acids, lysine, serine, and threonine (table 8.1).

The building block compounds were clearly differentiated from two other potential biorefinery products: direct product replacements and novel products. The former might include acrylic acid manufactured from lactic acid rather than from fossil-derived propylene; markets already existed for the compound and the cost structures and growth potential of these markets were well understood, thus reduc­ing the risks involved in devising novel production routes. On the other hand, novel products such as polylactic acid (section 8.1) had no competing routes from fossil reserves, had unique properties (thus rendering cost issues less crucial), and were intended to meet new markets.

Microbial Routes Known for Future Biobased Chemical Building Blocks

In contrast, the building block compounds were envisaged as being the starting points for diverse portfolios of products, both replacing existing fossil-based compounds and offering novel intermediates for chemical syntheses. This combina­tion has three advantages:

• The market potential is expanded.

• Multiple possible markets can reduce risks.

• Capital investment can be spread across different industrial sectors.

To illustrate these points, the 12 “finalists” could give rise to many derivatives that would find immediate or short-term uses in fields as diverse as transportation (polymers for automobile components and fittings, anticorrosion agents, and oxygenates), recreation (footgear, golf equipment, and boats), and health and hygiene (plastic eyeglasses, suntan lotions, and disinfectants).15 Three of them will now be considered in greater detail to explore possible key features and likely variables in the development of biorefineries.

Partly as another explored route for process cost reduction but also as a means to avoid the accumulation of ethanol concentrations inhibitory to cell growth or toxic to cellular biochemistry, technologies to remove ethanol in situ, that is, during the course of the fermentation, have proved intermittently popular.254 Seven different modes of separation have been demonstrated in small-scale fermentors:

• A volatile product such as ethanol can be separated from a fermentation broth under vacuum even at a normal operating temperature; a system with partial medium removal and cell recycling was devised to minimize the accumula­tion of nonvolatile products inhibitory to yeast growth and productivity.152

• If the fermentor is operated normally but the culture liquid is circulated through a vacuum chamber, the ethanol formed can be removed on a con­tinuous basis; this arrangement avoids the need to supply O2 to vessels maintained under vacuum.255

• Solvent extraction with a long chain alcohol (я-decanol) with immobilized cells of S. cerevisiae; up to 409 g/l of glucose (from glucose syrup) could be metabolized at 35°C.256

• As with water removal from concentrated ethanol, ethanol can be selectiv­ity adsorbed by different types of resins with hydrophobic surfaces, includ­ing cross-linked divinylbenzene polystyrene resins widely used in modern chromatographic separations of alcohols, sugars, and carboxylic acids; such resins work efficiently with ethanol at low ethanol concentrations, and the ethanol can be desorbed with warm dry N2 gas at 60-80°C.257258

• Hollow-fiber microfiltration is effective for ethanol and other small-mol­ecule products (such as lactic acid) but is slow and difficult to sterilize.259

• In membrane pervaporation, the cells are retained by a semipermeable membrane while a partial vacuum is applied to the permeate side; etha­nol concentrations could be maintained below 25 g/l for five days while a concentrated ethanol efflux stream of 17% w/v was achieved.260 Polyvinyl alcohol membranes operate better at elevated temperature, and this sug­gests that thermophilic ethanologens would be very suitable in a membrane pervaporative process.261

• Gas stripping of ethanol can be effected in an air-lift fermentor, a type of vessel originally developed for viscous microbial fermentation broths but also used for some of the more fragile and shear-sensitive mammalian cells in culture; this is another example of a technology that would inevitably work better with a thermophilic ethanologen and an elevated fermentation temperature.262 Alternatively (and more economically, with reduced power consumption for gas volume flow), the fermentation broth is circulated through an inert packed column and continuously sparged with a strip­ping gas (see figure 4.10) — such arrangements can result in highly stable continuous fermentations (for >100 days), with near-theoretical yields of ethanol from concentrated glucose solutions (560 g/l) in corn steep water to provide nutrients.263 264

How many (if any) of these advanced downstream technologies become adopted for industrial use will depend heavily on their economics — ethanol stripping is, for example, assessed at providing a significant cost savings for fuel ethanol production from cornstarch.265 With lignocellulosic substrates being used more widely, espe­cially in developing economies, a much simplified technology can provide surpris­ingly elegant solutions. Solid-state fermentations[42] have long been used for fermented foods and sake but can easily be adapted to manufacture (under more stringent con­ditions and with a reduced labor intensity) many fine chemicals and enzymes.266 A continuous process has been engineered to process and ferment feedstocks such as fodder beet and sweet sorghum in a horizontal tubular bioreactor, the ferment­ing material (with a low moisture content) moved along with the aid of a spiral screw.267,268 Some ethanol volatilization will occur at any temperature above ambient (caused by the fermentation process), but the bulk of the product could be recovered by a gas or air flowing through the container before the ethanol is condensed and transferred to a final dehydration step (as in the gas stripping technology). Although originally devised for farm-scale facilities (by the Alcohol Fuel Research Labora­tory, South Dakota State University), this solid-phase bioprocess yielded 87 l of ethanol/tonne of feedstock and was sufficiently productive to allow distillation from 8% v/v outputs. Echoing some of the discussion in chapter 1, the net energy balance (see section 1.6.1) was calculated to be unambiguously positive for fodder beet (2.11 for pasteurized pulped beet fodder, 3.0 for unpasteurized substrate), although much less persuasive for sweet sorghum (1.04 and 1.30, respectively) but was — even in 1984, when world oil prices were unpredictable and high after the price inflation of the 1970s — uncompetitive with then current gasoline prices (figure 1.3).

Local ecological impacts from biodiesel production units are most likely to be severe in areas of the world with lax environmental policies or enforcement. This is a par­ticularly acute problem for biodiesel production because of the potential to generate large volumes of aqueous wastes with high biological demands. Glycerol is a major waste product unless it is exploited as an income-generating stream, but this is not always economically feasible.58 Biodiesel wastes containing glycerol can be utilized by a Klebsiella pneumoniae strain to produce hydrogen by fermentation as a source of locally generated combustible gas for local heating or on-site use for biomass drying.72 A second biohydrogen system was based on an Enterobacter isolated from methanogenic sludge; glycerol-containing biodiesel wastes were diluted with a syn­thetic medium to increase the rate of glycerol consumption and the addition of nitro­gen sources (yeast extract and tryptone) enhanced the rates of both hydrogen and ethanol formation.73 In general, however, wastewater streams from biodiesel plants are not considered suitable for microbial remediation because of the high-pH, hex­ane-extractable oil, low nitrogen concentrations, and the presence of growth inhibi­tors; an oil-degrading Rhodotorula mucilaginosa yeast was found to degrade oil in wastewaters diluted with water to reduce growth inhibition and the content of solid materials, and this was developed into a small-scale treatment method.74

Especially in Europe, however, possible environmental damage is more often viewed as an international issue. This can be traced back to the initial period of biodiesel production in the 1980s, when it became evident that land resources inside the European Union were highly likely to limit biodiesel manufacturing capacity with European feedstocks: by 1992, oilseed rape cultivation in the United King­dom (barely known in the 1970s) was estimated to cover 400,000 hectares of land, enough crop to satisfy no more than a 5% substitution of conventional diesel sales.4 Twenty-five years later, the United Kingdom was devoting 570,000 hectares to grow­ing oilseed rape but 40% of this was for food use (cooking oil, margarine, etc.); even if all this land and whatever land is “set aside” under Common Agricultural Policy, policies could still only support a 5% (or less) substitution of fossil diesel.75 Even if all the U. S. soybean crop were to be devoted to biodiesel production, only 6% of U. S. diesel demand could be met.60 To meet U. S. and EU targets, therefore, importing biodiesel feedstocks is probably unavoidable, and this is (equally probably) depen­dent on supplies from Africa and Asia, with energy crops grown on recently cleared, deforested land; net importers can limit their imports from countries operating under plans such as the Round Table on Sustainable Palm Oil, an initiative to legitimize the trade in sustainably produced feedstock, but have often proved reluctant to ban imports of unsustainable biofuel sources for fear of breaching World Trade rules.76

To increase the pressure on available arable land further, biodiesel from mono­culture crops such as soybean and canola support much poorer energy production rates than does corn-based ethanol.77 The International Energy Agency has predicted that the land requirements for biofuels production will increase from a global figure of 1% in 2004 to 2.5% by 2030 — or, under alterative scenarios, to 3.8% or even 4.2% (58.5 million hectares).2 The search for high-yielding energy crops suitable

225

175

* Brazilian sugarcane-derived ethanol may, depending on the sugarcane price, be cheaper to produce than conventional gasoline and represent a negative abatement cost (figure 6.7).

rural employment, and forms part of the government’s policies to eliminate poverty; targets under “Probiodiesel” include 2% of all transport diesel to be biodiesel by 2008 (when all fuel distributors will be required to market biodiesel) and 5% by 2013.80 As a chemical technology, biodiesel production is easily integrated with the existing infrastructure of heavy chemical industry, sharing sites and power costs: a BASF maleic anhydride plant at Feluy (Belgium) shares waste heat with a biodiesel production unit operated by Neochim.81 Industrial plants that used to produce glyc­erol are now closing down to be replaced by others that use glycerol as a raw material, owing to the large surplus of glycerol formed as a coproduct during the production of biodiesel; in parallel, research efforts to find new applications of glycerol as a low — cost feedstock for functional derivatives have led to the introduction of a number of selective processes for converting glycerol into commercially valued products.82

Although the public face of soybean oil-derived biodiesel in Brazil and elsewhere remains that of sustainable symbiotic nitrogen fixation with the bacterium Bradyrhi- zobium japonicum and soybean cultivars selected to grow in the arid savannah in the hinterland state of Mato Grosso, environmentalist concerns about “deforestation diesel” remain acute: Brazil already exports 20 million tonnes of soybean annually and plans to increase this to 32 million tonnes by 2015.83 To what extent soybean — plantation cultivation encroaches from savannah to adjacent rain forest will largely determine how “sustainable” this major source of biodiesel feedstock will prove.