Biomass — as described in the preceding section — has a calorific value that can be at least partly “captured” in other, more immediately useful forms of energy. Two of the examples included in figure 2.2 were fermentations, biological processes involving microbial bioenergetics in the chemical transformations. “Fermentation” has come a long way semantically since its coining by Louis Pasteur as life without air (la vie sans l’air).8 Pasteur was referring to yeast cells and how they altered their respiratory functions when air (or oxygen) became limiting. The modern use of the term is best understood by considering how microbes oxidize (or metabolize or combust) a substrate such as glucose released by the hydrolysis of, for example, starch or cellulose (figures 1.13 and 1.23). The complete biological oxidation of glucose can be written as

+ 6O2 ^ 6CO2 + 6H2O

As written, this chemical transformation is of little biological use to a population of microbial cells (yeast, fungi, or bacteria) because it implies only the generation of heat (metabolic heat) by the cells and that all the carbon is evolved as CO2 and cannot be utilized in cell replication. In a real microbiological scenario, glucose represents a valuable organic carbon supply as well as an energy source; part of that carbon is transformed by biochemical reactions inside the cells into the materials for new cells (before cell division), and much of the free energy that is released by the energetically favorable (exothermic) oxidation of the remainder of the glucose is used to drive the energetically unfavorable (endothermic) reactions of biosynthetic pathways; only a small portion of the total energy available is “wasted” as heat by biochemical reactions having mechanisms with thermodynamic efficiencies less than 100%:

nC6H12O6 + mO2 ^ cells + pCO2 + qH2O + evolved heat (where p, q << 6n).

Such a system is, to the microbial physiologist, simply “growth.” If, however, a metabolic output takes the form of a chemical product — especially where that product is commercially desirable or useful — then the process generates cells and product, whereas the CO2 evolved is (from an industrial or bioprocess standpoint) “waste.” In large industrial fermentors (with volumes in excess of 500,000 L), the generation of “metabolic” heat by dense cultures of microbial cells necessitates the expenditure of large amounts of energy to cool the liquid mass and, in turn, stimulates considerable efforts to recycle the “waste” heat on the industrial plant.9 Microbial processes producing antibiotics, enzymes, amino acids, and other products, recombinant proteins, flavors, and pharmaceutical active ingredients are generally termed “fermentations” although large volumes of oxygen-bearing air are supplied (via heavy-duty compressors) at high rates to the vessels so that microaerobic or anaerobic conditions are deliberately prevented, large quantities of glucose (or other carbon and energy sources) are utilized, and commercial products are accumulated as rapidly as the nutritional and physical conditions allow.

The primary fermentation in ethanol production is, on the other hand, a classic fermentation, that is, one where the ability of the cells to absorb glucose and other sugars from the medium (and to complete the early steps in their metabolic oxidation) exceeds the supply of oxygen required to fully complete the oxidative reactions. Under such conditions, yeasts such as Saccharomyces cerevisiae will accumulate (mostly) ethanol, whereas other yeasts, fungi, and bacteria produce combinations of the following:10

• Alcohols (ethanol, glycerol, я-propanol, я-butanol)

• Acids (formic, acetic, lactic, propionic, butyric)

• Decarboxylated acids (acetoin, acetone, diacetyl, 2,3-butanediol)

All these products (including ethanol) are the products (or intermediates) of a cluster of closely linked metabolic pathways and most of them are formed biosyn­thetically by reduction, thus regenerating the finite pool of redox carriers — in par­ticular, nicotinamide adenine dinucleotide (NAD) — in their oxidized forms inside the cell (figure 2.3). With the exception of ethanol (a much earlier development), all of these products were also industrially produced by microbial processes that were commercialized in the twentieth century; they represent “overflow” products of metabolism, given the supply of a high-value carbon and energy source such as glucose and an imbalanced nutritional environment where only part of the available carbon and energy can be fully used in cell replication and cell division. In particular, any cell population has, as determined by its precise genetic profile, a finite maximum specific oxygen uptake rate (i. e., grams of O2 per grams of cells per hour); feeding glucose as the growth-limiting nutrient to induce specific growth rates requiring the cells to exceed this maximum specific oxygen uptake rate causes the accumulation of “overflow” metabolites such as acetic and formic acids (figure 2.3).11 Ethanol produc­tion by S. cerevisiae, however, exhibits a highly important relationship: even under aerobic conditions (where the capacity to fully metabolize, or combust, glucose is not compromised by the lack of O2), feeding glucose at an increasingly rapid rate eventually swamps the ability of the cells to both grow and respire the glucose, that is, to provide the energy required to support growth and all the required ana­bolic, biosynthetic reactions. Below this crucial rate of glucose supply (most readily demonstrated in continuous cultures where the rate of glucose entry, and therefore, growth, is determined by the experimenter without any limitation of O2 or other nutrient supply), no ethanol is formed;[11] above this threshold, ethanol is accumulated, the “Crabtree effect” (see section 3.1.1), which is so vital to the long historical use of “wine” and other yeasts for ethanol production:10

яC6H12O6 + mO2 ^ cells + pCO2 + qH2O + rC2H5OH + evolved heat

Much has been written quasi-philosophically (and from a distinctly anthropo­centric viewpoint) on how the central metabolic pathways in microbes and plants have evolved and on their various thermodynamic inefficiencies.13 The single most pertinent point is, however, very straightforward: in a natural environment, there is no selective advantage in microbes having the capability of extracting the full thermo­dynamic energy in foodstuffs; the problem is that of kinetically utilizing (or using up) any available food source as rapidly as possible, thereby extracting as much energy

glucose

Enzyme pathways for glucose metabolism in microbial fermentations (+[2H], -[2H] represent reductive and oxidative steps,

On

On

and nutritional benefit in the shortest possible time in periods of alternating “feast and famine.”14 Many microbes are adventitious feeders, being capable of adapting even to extremes of oxygen availability in aerobic and anaerobic lifestyles. It is merely a human description that overspill products such as ethanol, acetic and lactic acids, and solvents such as acetone are “useful”; in a natural ecosystem, each might be the food­stuff for successive members of microbial consortia and lower and higher plants — in that sense of the term, nature is never wasteful. Axenic cultures of microbes in shake flasks and industrial fermentors are (in the same way as fields of monoculture crops such as soybean and sugarcane) “unnatural” because they are the products of modern agricultural practice and industrial microbiology and offer the economic advantages and potential environmental drawbacks of both of those twentieth-century activities.

Once the growth rate in a fermentation of glucose by an ethanol-forming organism becomes very low, and if glucose continues to be supplied (or is present in excess), the biochemistry of the process can be written minimally:

C6H12O6 ^ 2CO2 + 2C2H5OH

This represents a fully fermentative state where no glucose is respired and no fermentation side products are formed: for every molecule of glucose consumed, two molecules of ethanol are formed; for every 180 g of glucose consumed, 92 g of ethanol are formed, or 51.1 g of ethanol are produced per 100 g of glucose utilized. Any measurable growth or side-product formation will reduce the conversion efficiency of glucose to ethanol from this theoretical maximum.

Microbial energetics and biochemistry can, therefore, predict maximum yields of ethanol and other fermentation products from the sugars available in cane juices, starch, and cellulose (figure 2.3). How closely this theoretical maximum is reached is a function of fermentation design, control, engineering, and the biochemistry of the microorganism used for ethanol production. If a lignocellulosic substrate is to be used, however, the range of biochemical pathways required increases greatly both in scope and complexity as more sugars (pentoses as well as hexoses, plus sugar acids such as glucuronic acid) are potentially made available; this introduces substantially more variables into the putative industrial process, and preceding even this, there is the major hurdle of physically and chemically/enzymically processing biomass materials to yield — on an acceptable mass and time basis — a mixture of fermentable sugar substrates.

The previous sections have itemized the various technologies that have been developed to process lignocellulosic materials to mixtures of glucose, pentose sugars, and oligosaccharides suitable for fermentation by microbes with the production of ethanol. But what commercial and economic forces are presently acting to determine (or limit) the choice of lignocellulosic materials for the first large-scale bioethanol facilities?

The most detailed description so far available for the demonstration plant con­structed by the Iogen Corporation in Canada candidly lists the possibilities for start­ing materials:146

• Straws (wheat, barley, etc.) and corn stover as the leading candidates

• Cane bagasse as a localized leading candidate for some tropical locations

• Grass “energy crops” as possible second-generation candidates

• Native forest wood but difficult to process

• Tree farms too expensive because of demands of other markets

• Bark tree waste — cellulose and hemicellulose contents too low

• Sawdust and other mill waste — too expensive because of pulp and paper market demands

• Municipal solid waste and waste paper — too expensive because of paper demand

For a start-up lignocellulosic ethanol facility, there are crucial issues of cost and availability. An industrial plant may require, for example, close to a million tons of feedstock a year, that feedstock should be (for operational stability) as uniform as possible and as free from high levels of toxic impurities and contaminations as possible. Some materials (e. g., wood bark) have compositions that are incompatible with the high yields achieved in starch — and sugar-based materials, and some soft­woods demand unattractively high inputs of cellulase.147 Any lignocellulosic mate­rial is subject to some competitive use, and this may dictate cost considerations (table 2.10). Some of these direct competitors are long-established, mature industries, whereas others have unarguably “green” credentials for recycling waste materials or in renewable energy generation in OECD and non-OECD economies.148 Agricultural waste materials have, in addition, great potential as substrates for the “solid-state” fermentative production of a wide spectrum of fine chemicals, including enzymes, biopesticides, bioinsecticides, and plant growth regulators.149150

The enormous size of the potential supply of lignocellulose is frequently asserted; for example, “Lignocellulose is the most abundant renewable natural resource and substrate available for conversion to fuels. On a worldwide basis, terrestrial plants produce 1.3 x 1010 metric tons (dry weight basis) of wood/year, which is equivalent to 7 x 109 metric tons of coal or about two-thirds of the world’s energy requirement.”69

Some of this “wood,” however, represents trees grown (or harvested) as food crops and used as direct domestic or even industrial energy resources, while most is intimately involved in the global carbon cycle and in stabilizing the CO2 balance in the global ecosystem. Much of what is calculable as available biomass may not, therefore, be commercially harvestable on a short-term basis without the large — scale planting of dedicated “energy crops” such as fast-growing willow trees that are presently planted and harvested in Sweden for burning in district, local, and domestic heating systems.

TABLE 2.10

Competing Uses for Lignocellulosic Biomass Materials Considered for Bioethanol Production

Material Source Uses

Indeed, these were important issues addressed in a multiauthor projection of biomass options for 2030 to replace 30% of U. S. fuel demands.151 Ignoring questions of cost and efficiency of bioconversion, this study identified both forestry and agricultural resources for use in biofuels production. Three types of forest resources were quantified:

1. Primary (logging residues, removal of excess biomass in timberland fuel treatments, and fuel wood extracted from forestlands)

2. Secondary (mill residues)

3. Tertiary (urban wood residues from construction/demolition and recycling)

Similarly, lignocellulosic agricultural resources were divided among

1. Primary (crop residues, perennial grasses, and perennial woody crops)

2. Secondary (food/feed processing residues)

3. Tertiary (municipal solid waste recycling)

Approximately 280 million tons (dry weight) of such resources were estimated to be available by the time of the report on an annual basis (figure 2.12).[17] Augmentation

□ Agriculture ■ Forestry

of this supply could arise from programs to thin native forests strategically so as to reduce fire hazards — in California, for example, more than 750,000 dry tonnes/year could be generated by such activities.152 Data from Sweden (with its relatively low population density and high degree of forestation) suggest that 25% of the country’s gasoline requirements could be substituted by lignocellulose-derived ethanol from existing biomass resources.153

Of the two major immediately available sources of lignocellulosic material, however, field crop residues have the distinct advantage of being generated in close proximity to cereal crops intended (partly or entirely) for ethanol produc­tion. In August 2005, Abengoa Bioenergy (www. abengoa. com) began constructing the world’s first industrial-scale cellulosic bioethanol plant (to use wheat straw as the feedstock) immediately adjacent to its existing 195 million liters/year, Cereal Ethanol Plant (Biocarburantes de Castilla y Leon, BcyL), at Babilfuente, Salamanca, Spain, to dovetail supply trains and technologies. The biomass plant will process more than 25,000 tonnes of wheat straw and other materials to produce 5 million liters of ethanol annually in addition to preparing lignin, pentose sugars, and animal feed products as manufacturing outputs.

As an “energy crop” feedstock for bioethanol production in North America, a grass such as switchgrass (Panicum virgatum) has persuasive advantages—economic, social, and agricultural.154 But the technology for harvesting and processing grasses must be scaled up considerably from that used in, for example, silage fermentation. In the absence of such purpose-dedicated crops, intelligent choices will be mandatory to access sufficiently large supplies of suitable cellulosic feedstocks for start-up facilities.

How much biomass can any nation or region abstract without harming the environment? In a European context, this is an urgent question because any expansion of biomass use for industrial purposes brings in the threat of placing additional pressures on soil and water resources, farmland and forestry biodiversity, and may run counter to extant legislation aiming to encourage environmentally sound farming practices. The European Environment Agency naturally became interested in this issue as projections for large-scale biomass harvesting for bioenergy began to be more ambitious.155 Taking a cautious view of “harming the environment,” that is, with protected forest areas being maintained, residue removal excluded, with no grasslands or olive groves transformed into arable land, and with at least 20% of arable land maintained under environmentally friendly cultivation, the EEA estimated that by 2030 large amounts of biomass could be made available for ambitious bioenergy programs, reaching approximately 300 million tonnes of oil equivalent annually, or 15% of the total primary energy requirements. The calculated increase during the 27 years from 2003 were assumed to be the result of improved agricultural and forestry productivity and liberalization steps leading to more cultivable land being used with higher oil prices and imposed carbon taxes encouraging this expansion. The most revealing aspect of the data presented by the EEA was, however, a ranking of different plant species planted as annual crops with bioenergy as a significant end use: only a mixture of species could ensure that no increased environmental risks were likely, and it is interesting that maize (corn) ranks highest (most potentially damaging) as a monoculture (figure 2.13). The implication was that some of the

Maize Sugar Beet Potato Rape Seed Sunflower Grasses Wheat Mustard Seed Hemp Alfalfa Linseed

favored biofuel crops (maize, sugarbeet, and rapeseed) were in the highest risk category, and careful monitoring of the areas planted with those crops would be advised to minimize environmental damage; grasses were middle ranking purely on the grounds of increased fire risk.

1.1 ETHANOL FROM NEOLITHIC TIMES

There is nothing new about biotechnology. Stated more rigorously, the practical use — if not the formal or intuitive understanding of microbiology — has a very long history, in particular with regard to the production of ethanol (ethyl alcohol). The development of molecular archaeology, that is, the chemical analysis of residues on pottery shards and other artifacts recovered from archaeological strata, has begun to specify discrete chemical compounds as markers for early agricultural, horticultural, and biotechnological activities.1 Among the remarkable findings of molecular archae­ology, put into strict historical context by radiocarbon dating and dendrochronology techniques, as well as archaeobotanical and archaeological approaches, are that

• In western Asia, wine making can be dated as early as 5400-5000 BC at a site in what is today northern Iran and, further south in Iran, at a site from 3500 to 3000 BC.1

• In Egypt, predynastic wine production began at approximately 3150 BC, and a royal wine-making industry had been established at the beginning of the Old Kingdom (2700 BC).2

• Wild or domesticated grape (Vitis vinifera L. subsp. sylvestris) can be traced back to before 3000 BC at sites across the western Mediterranean, Egypt, Armenia, and along the valleys of the Tigris and Euphrates rivers. This is similar to the modern distribution of the wild grape (used for 99% of today’s wines) from the Adriatic coast, at sites around the Black Sea and southern Caspian Sea, littoral Turkey, the Caucasus and Taurus mountains, Lebanon, and the islands of Cyprus and Crete.3

• Partial DNA sequence data identify a yeast similar to the modern Saccha — romyces cerevisiae as the biological agent used for the production of wine, beer, and bread in Ancient Egypt, ca. 3150 BC.2

The occurrence of V. vinifera in regions in or bordering on the Fertile Crescent that stretched from Egypt though the western Mediterranean and to the lower reaches of the Tigris and Euphrates is crucial to the understanding of Neolithic wine making. When ripe, grapes supply not only abundant sugar but also other nutrients (organic and inorganic) necessary for rapid microbial fermentations as well as the causative yeasts themselves — usually as “passengers” on the skins of the fruit. Simply crushing (“pressing”) grapes initiates the fermentation process, which, in unstirred vessels (i. e., in conditions that soon deplete oxygen levels), produces ethanol at 5-10% by volume (approximately, 50-100 g/l).

In China, molecular archaeological methodologies such as mass spectroscopy and Fourier transform infrared spectrometry have placed “wine” (i. e., a fermented mixture of rice, honey, and grape, as well as, possibly, other fruit) as being produced in an early Neolithic site in Henan Province from 6500 to 7000 BC.4 Geographically, China lies well outside the accepted natural range of the Eurasian V. vinifera grape but is home to many other natural types of grape. Worldwide, the earliest known examples of wine making, separated by more than 2,000 km and occurring between 7000 and 9000 years ago, were probably independent events, perhaps an example on the social scale of the “convergent evolution” well known in biological systems at the genetic level.

The epithet “earliest” is, however, likely to be limited by what physical evidence remains. Before domestication of cereals and the first permanent settlements of Homo sapiens, there was a long but unrecorded (except, perhaps, in folk memory) history of hunter-gatherer societies. Grapes have, in some botanical form or other, probably been present in temperate climates for 50 million if not 500 million years.3 It would seem entirely possible, therefore, that such nomadic “tribes” — which included shamans and/or observant protoscientists — had noted, sampled, and replicated natural fermentations but left nothing for the modern archaeologist to excavate, record, and date. The presently estimated span of wine making during the last 9000 years of human history is probably only a minimum value.

Grape wines, beers from cereals (einkorn wheat, one of the “founder plants” in the Neolithic revolution in agriculture was domesticated in southeastern Turkey, ca. 8000 BC), and alcoholic drinks made from honey, dates, and other fruits grown in the Fertile Crescent are likely to have had ethanol concentrations below 10% by volume. The concentration of the ethanol in such liquids by distillation results in a wide spectrum of potable beverages known collectively as “spirits.” The evolution of this chemical technology follows a surprisingly long timeline:56 [1]

Distillation yields “95% alcohol,” a binary azeotrope (a mixture with a constant composition) with a boiling point of 78.15°C. “Absolute” alcohol, prepared by the physical removal of the residual water, has the empirical formula C2H6O and molecular weight of 46.07; it is a clear and colorless liquid with a boiling point of 78.5°C and a density (at 20°C) of 0.789 g/mL. Absolute alcohol absorbs water vapor rapidly from the air and is entirely miscible with liquid water. As a chemical known to alchemists and medicinal chemists in Europe and Asia, it found many uses as a solvent for materials insoluble or poorly soluble in water, more recently as a topical antiseptic, and (although pharmacologically highly difficult to dose accurately) as a general anesthetic. For the explicit topic of this volume, however, its key property is its inflammability: absolute alcohol has a flash point of 13°C.7

By 1905, ethanol was emerging as the fuel of choice for automobiles among engineers and motorists,* opinion being heavily swayed by fears about oil scarcity, rising gasoline prices, and the monopolistic practices of Standard Oil.8 Henry Ford planned to use ethanol as the primary fuel for his Model T (introduced in 1908) but soon opted for the less expensive alternative of gasoline, price competition between ethanol and gasoline having proved crucial. The removal of excise duty from dena­tured ethanol (effective January 1, 1907) came too late to stimulate investment in large-scale ethanol production and develop a distribution infrastructure in what was to prove a narrow window of opportunity for fuel ethanol.8

Ford was not alone in considering a variety of possible fuels for internal com­bustion engines. Rudolf Diesel (who obtained his patent in 1893) developed the first prototypes of the high-compression, thermally efficient engine that still bears his name, with powdered coal in mind (a commodity that was both cheap and readily available in nineteenth-century Germany). Via kerosene, he later arrived at the use of crude oil fractions, the marked variability of which later caused immense practi­cal difficulties in the initial commercialization of diesel engines.9 The modern oil industry had, in effect, already begun in Titusville, Pennsylvania, in the summer of 1859, with a drilled extraction rate of 30 barrels a day, equivalent to a daily income of $600.10 By 1888, Tsarist Russia had allowed Western European entrepreneurs to open up oil fields in Baku (in modern Azerbaijan) with a productive capacity of 50,000 barrels a day. On January 10, 1901, the Spindletop well in Texas began gushing, reaching a maximum flow of 62,000 barrels a day. Immediately before the outbreak of World War I, the main oil-producing countries could achieve out­puts of more than 51 million tons/year, or 1 million barrels a day. In 1902, 20,000 vehicles drove along American roads, but this number had reached more than a million by 1912. These changes were highly welcome to oil producers, including (at least, until its forced breakup in 1911) the Standard Oil conglomerate: kerosene intended for lighting domestic homes had been a major use of oil but, from the turn of the century, electricity had increasingly become both available and preferable (or fashionable). The rapid growth in demand for gasoline was a vast new market for J. D. Rockefeller’s “lost” oil companies.

Greatly aiding the industry’s change of tack was the dominance of U. S. domestic production of oil: in 1913, the oil produced in the United States amounted to more [2]

than 60% of the worldwide total (figure 1.1). The proximity within national boundar­ies of the world’s largest production line for automobiles (in Detroit) and oil refining capacities firmly cast the die for the remainder of the twentieth century and led to the emergence of oil exploration, extraction, and processing, and the related petrochemi­cal industry as the dominant features of the interlinked global energy and industrial feedstock markets.

Nevertheless, Henry Ford continued his interest in alternative fuels, sponsoring conferences on the industrial uses of agricultural mass products (grain, soybeans, etc.) in 1935-1937; the Model A was often equipped with an adjustable carburetor designed to allow the use of gasoline, alcohol, or a mixture of the two.11

2.3.1 Lignocellulose as a Chemical Resource

Cellulose, hemicellulose, and lignin are the polymers that provide the structural rigid­ity in higher plants that grow vertically from a few centimeters to tens of meters — the giant redwood (Sequoiadendron giganteum) reaches up to 300 ft (90 m) in height. Although a multitude of microorganisms can elaborate enzymes to degrade cellulose and hemicelluloses, the success of the lignocellulose architecture in the global eco­system is such that it was only with the advent of Homo sapiens with flint and (later) metal axes that the domination of deciduous and coniferous forests (especially in the Northern Hemisphere) was seriously challenged.

At the cellular level, plants derive their remarkable resilience to physical and microbial weathering and attack from having evolved the means to greatly thicken their cell walls, using cellulose in linear polymers of high molecular weight (500,000-1,500,000) that are overlapped and aggregated into macroscopic fibers.15 Linear strands of cellulose have a close molecular arrangement in fibrillar bun­dles that are sufficiently regular to have X-ray diffraction patterns characteristic of “crystals.” This not only augments the structural cohesion but also limits access by water-soluble components and enzymes, and native cellulose is essentially insolu­ble in water.

Hemicelluloses are diverse in both sugar components and structure, with polymeric molecular weights below 50,000.15 The heterogeneity of lignins is even greater; any estimate of molecular weight is highly dependent on the method used for extraction and solubilization, and average molecular weight distributions may be less than 10,000.16 Lignin and hemicelluloses may form chemically linked complexes that bind water-soluble hemicelluloses into a three-dimensional array, cemented together by lignin, that sheaths the cellulose microfibrils and protects them from enzymic and chemical degradation.151718

From the perspective of 2007, an important strategic crossroads in the means of providing bioenergy has been reached. A wide complement of technologies has been developed to process realistically available amounts of lignocellulosic materials on at least a semi-industrial scale preparatory to bioethanol production.156 A crucial component in this is the massively increased production and continued improvement of the “molecular machines” of cellulases (endoglucanases).157 Chem­ical (pyrolytic and thermal “cracking”) methodologies are being critically devel­oped for biomass substrates to generate synthesis/producer gas both as a direct energy source and as an intermediary stage for “green” chemical refineries.158 Despite these advances, simple wood-burning stoves for domestic use and furnaces for district heating projects still represent far larger energy gains than does any form of biofuel production.159

Nascent industrial facilities for lignocellulosic ethanol are focused on exploiting the supply of cereal crop waste materials — in particular, wheat straw — that together comprise only 30% of the presently available lignocellulosic biomass resources.151 Worldwide, the industry may have to adapt to a succession of different seasonally available feedstocks, each of which presents unique challenges to pretreatment processing; softwood trees are, for example, dominant contributors to vegetation in Canada, northern Europe, Russia, and Scandinavia, and have both dedicated and passionate advocates and a history of several decades of scientific research; their technoeconomic features and factors are discussed in chapter 4.

Biotechnology must, therefore, demonstrate that multiple carbon sources ( hexoses, pentoses, sugar acids, oligosaccharides) can be efficiently converted to ethanol, a proposal that flies in the face of the fermentation industry’s tradition of sim­ple (single), highly reproducible carbon inputs. Moreover, biochemical engineering solutions must be found to maximize the value extracted from the processed cel­lulose, hemicelluloses, and lignin. The former area is that of metabolic engineering, the latter that of bioprocess control; both are considered next as integral parts of evolving models and paradigms of bioethanol production.

Many commentators state that the Oil Crisis of 1973, after the Yom Kippur War, cat­alyzed the interest in and then sustained the development of biofuels on the national and international stages. This is an overly simplistic analysis. The following words were spoken by Senator Hubert Humphrey in May 1973, some five months before war in the Middle East broke out:12

I have called these hearings because … we are concerned about what is going on with gasoline; indeed, the entire problem of energy and what is called the fuel crisis. Gas prices are already increasing sharply and, according to what we hear, they may go much higher. … We were saved from a catastrophe in the Midwest — Wisconsin, Iowa and Minnesota — and in other parts of the country, by the forces of nature and divine provi­dence. We had one of the mildest winters in the past 25 years, and had it not been for the unusually warm weather, we would have had to close schools and factories, we would have had to shut down railroads, and we would have had to limit our use of electrical power.

Security of oil supplies and the pressures of price inflation have, since the 1970s, been major issues that continue to the present day.

Even a cursory glance at figure 1.1 will show how disadvantaged were the Ger­man, Austro-Hungarian, and Ottoman empires in comparison with the Allied powers in World War I, especially after the entry of the United States in 1917, with only Polish and some Romanian oil fields beyond the vagaries of naval blockade and interception; the ingenuity of the German chemical industry was severely stretched by the effort to substitute imports (including fuel oils) by innovations with synthetic, ersatz prod­ucts. Since then, and throughout the twentieth and early twenty-first centuries, any state entering into global or regional wars faces the same strategic imperatives: how to ensure continued oil supplies and how (if possible) to control access to them. From the naval blockades of 1914 to the air strikes of the 2006 Hezbollah-Israel conflict, oil refineries and storage tanks are to be targeted, sea-lanes interdicted, and, if possible, foreign oil fields secured by invasion. In those 90 years, wars and economic depressions often demanded attempts to substitute ethanol for gasoline. In the 1920s and 1930s, several countries (Argentina, Australia, Cuba, Japan, New Zealand, the Philippines, South Africa, and Sweden) used ethanol blends in gasoline; alcohol-fueled vehicles became predominant in Germany during World War II and, by 1944, the U. S. Army had developed a nascent biomass-derived alcohol industry.11 Such programs were, however, mostly of a contingency (or emergency) nature, highly subsidized, and, once oil began flowing in increasingly large amounts after 1945, generally abandoned.

In the decade immediately preceding 1973, the United States had lost its domi­nance of world oil production (figure 1.2). Other major players were expanding (e. g., the Middle East reached 30% of world oil production) and new producers were appear­ing: Africa (Libya, Algeria, and Nigeria) already produced 13% of world oil.13 Allow­ing for inflation, world oil prices slowly decreased throughout the 1960s (figure 1.3). At the time, this was perceived as a “natural” response to increasing oil production, especially with relative newcomers such as Libya and Nigeria contributing signifi­cantly; global production after World War II followed an exponential rate of increase (figure 1.2). Political changes (especially those in Libya) and a growing cooperation between oil-producing states in the Organization of Petroleum Exporting Countries (OPEC) and the Organization of Arab Petroleum Exporting Countries (OAPEC) led to new agreements between oil producers and oil companies being negotiated in Tehran (Iran) and Tripoli (Libya) in 1970 and 1971, which reversed the real oil price erosion.

Then, Libya and Kuwait began to significantly reduce oil output in a structured, deliberate manner. In Libya, average production was reduced from a peak of 3.6 million barrels/day before June 1970 to approximately 2.2 million barrels/day in 1972 and early 1973; the Kuwaiti government enforced a ceiling of 3 million barrels/ day in early 1972, shifting down from peak production of 3.8 million barrels/day.10 Structural imbalances in the global supply of oil had by that time become apparent because of short — and medium-term causes: [3]

• Accidental damage resulted in the prolonged closure of the pipeline carry­ing oil from Saudi Arabia to the Mediterranean.

• Supply and demand became very much more closely matched, impos­ing acute pressures on shipping and refinery kinetics; the estimated spare capacity in crude oil shrank from 7 million barrels/day in 1965 to less than 0.5 million barrels/day in early 1973.

A rapid response to the outbreak of war in October 1973 continued the politically motivated reduction in crude oil output: OAPEC proposed with immediate effect to cut back output by 5% with a further 5% each month until a settlement in accord with United Nations resolutions was effected. In addition, the Gulf States of OPEC, together with Iran, imposed unilateral price rises of up to 100%. The immediate effect on world oil prices was severe (figure 1.3). More importantly, however, the effect was

not transitory: although prices decreased from the initial peaks in 1973-1974, prices began a second wave of rapid increase in 1979 after the Iranian revolution, to reach a new maximum in 1981. From more than $50 a barrel in 1981, prices then con­founded industry analysts again, despite the subsequent conflict between Iran and Iraq, and crashed down to $20 by the late 1980s, but for over a decade real oil prices had been continuously threefold more (or greater) than those paid in 1970. Although not reaching the real prices recorded in the 1860s during the American Civil War (when industrialization was a new phenomenon for most of the world), the oil price inflation between 1973 and 1981 represented a markedly different scenario from any experienced during the twentieth century — in dollar or real terms — despite world wars and major depressions (figure 1.3).

Across the industrially developed states of the Organisation for Economic Co­operation and Development (OECD) — the United States, Japan, Germany, France, United Kingdom, Italy, and Canada — while the real price of imported crude oil had decreased between 1960 and 1973 by an average of 1%/annum, the inflation-adjusted price increased by 24.5%/annum between 1973 and 1980; the result was that the oil crisis soon developed into a deep economic crisis even in those economically and technically advanced OECD nations.14 Because gasoline prices were “buffered” by the (frequently high) taxes included in the at-pump prices in the OECD countries, gasoline prices to motorists increased by only two — to threefold between 1970 and 1980, whereas crude oil prices rose by more than eightfold; in contrast, industrial and domestic oil prices increased by approximately fivefold.14

Furthermore, viewed from the perspective of 1973, the future for oil supplies to net oil importers was highly problematic. Although known oil reserves amounted to 88 x 109 tons, more than 55% of these lay in the Middle East, and mostly in OAPEC countries (figure 1.4). In the days of the then-Cold War, the Soviet Union (USSR), Eastern Europe, and China accounted for only 16.3% of world oil production but

were net exporters of both crude oil and oil products, whereas the United States had become a net importer of both (figure 1.5). In the United States, oil represented 47% of total primary energy consumption.15 In other OECD countries, the dependence on oil was even more marked: 64% in Western Europe and 80% in Japan. The developed economies of the OECD countries responded to the oil price “shocks” of the 1970s by becoming more oil-efficient: while total OECD gross domestic product (GDP) increased by 19% between 1973 and 1980, total oil imports fell by 14%, and the oil used to produce each unit of GDP fell by 20% — to offset the reduced use of oil, however, coal and (especially) nuclear energy source utilization increased greatly.16 Energy conservation became a priority (“energy-demand management” measures), and technologies for the improved efficiency of energy use were much developed, advertised, and retrofitted to both domestic and industrial premises. “Fuel switch­ing” was much less obvious in the strategies adopted by OECD countries. While the substitution of gasoline for road transport by alcohol, liquefied gas, and so forth was widely advocated, by 1980, Canada was unique in having adopted a comprehensive policy (the “off oil conversion programme”) covering all aspects of oil use and pro­viding oil reduction targets as well as financial incentives.

For an “emerging” economy like Brazil’s, the economic dislocation posed by sustained oil price rises was potentially catastrophic. In November 1973, Brazil relied on imports for more than 80% of the country’s oil consumption; in the course of the following year, the total import bill rose from $6.2 billion to $12.6 billion, and the trade balance collapsed (figure 1.6). For the preceding decade, the Brazil­ian economy had enjoyed high growth rates (figure 1.6). Industrialization had pro­ceeded well, and the inflation rate had reached its lowest level since the 1950s.17 The Brazilian government opted against economic stagnation; rather, it aimed to pay

15

13

11

9

7

5

3

1

-1

-3

-5

for the higher oil bills by achieving continued growth. To meet the challenges of energy costs, the Second National Development Plan (1975-1979) decreed the rapid expansion of indigenous energy infrastructure (hydroelectricity) as well as nuclear power and alcohol production as a major means of import substituting for gasoline.

In the next decades, some of these macroeconomic targets were successfully realized. Growth rates were generally positive after 1973, and historically massive positive trade balances were recorded between 1981 and 1994. The counterindicators were, however, renewed high rates of inflation (reaching >100%/annum by 1980) and a spiral of international debt to fund developmental programs that made Brazil the
third world’s largest debtor nation and resulted in a debt crisis in the early 1980s. Arguments continue concerning the perceived beneficial and detrimental effects of the costs of developmental programs on political, social, and environmental indices in Brazil.17

Cane sugar was the key substrate and input for Brazil’s national fuel alcohol program. Sucrose production from sugarcane (Saccharum sp.) in Brazil has a long history, from its days as a colony of Portugal. Brazil had become the world’s leading sugar supplier by the early seventeenth century, but sugar production was based initially on slave labor and remained (even in the twentieth century) inefficient. This, however, represented a potential for rapid growth after 1975 because large monocul­ture plantations had been long established in the coastal regions of the northeast and southeast of the country. Expansion of cultivated land was greatly encouraged for the “modern” export crops — sugarcane, cotton, rice, corn, soybeans, and wheat — at the expense of the more traditional crops, including manioc, bananas, peanuts, and cof­fee. Sugarcane cultivation increased by 143% between 1970 and 1989 when expressed as land use, but production increased by 229% as Brazil’s historically low use of fertil­izer began to be reversed.17

Brazil is also the southernmost producer of rum as an alcoholic spirit, but cachaga is the oldest and most widely consumed national spirit beverage, with a yearly produc­tion of ca. 1.3 billion liters.18 The primary fermentation for cachaga uses sugarcane juice, and large industrial plants had been established after the end of World War II; a variety of yeasts had been developed, suitable for continuous or discontinuous fer­mentations, the former reusing and recycling the yeast cells.18 Before distillation, the fermentation is (as are all traditional potable alcohol processes) allowed to become quiescent, the yeast cells settling and then being removed (along with other residual solids) by, in technologically more advanced facilities, centrifugation; batch (“pot still”) and continuous distillation are both used, and final alcohol concentrations are in the 38 to 48% range (by volume). Predating the oil crises of the 1970s and 1980s, the first moves toward using cane sugar as a substrate for industrial ethanol produc­tion independent of beverages dated from 1930, when the Sugar and Alcohol Institute (Instituto do Agucar e do Alcool) was set up; in 1931, a decree imposed the compul­sory addition of 5% ethanol to gasoline, and the blending was increased to 10% in 1936. Four decades of experience had, therefore, been garnered in Brazil before fuel substitution became a priority on the political agenda.19

The final element in Brazil’s developing strategy to produce “gasohol” was, iron­ically, petroleum itself. Brazil had produced oil at a low rate from at least 1955, but the offshore deposits discovered by the state-owned company PETROBRAS were so large that by 1998 domestic oil production equaled 69% of domestic consumption.17 Production continued to increase (figure 1.7), and by 2005, Brazil had become a sig­nificant global producer, accounting for 2.2% of world oil production, equivalent to that of the United Kingdom, considerably higher than either Malaysia or India (both 0.9%) and approaching half that of China (4.6%).20 Indigenous refining capacity also increased during the 1970s and again after 1996 (figure 1.7). The ability to produce alcohol as a fuel or (when mixed with gasoline) as a fuel additive became — if need be, at an unquantified ecological cost (chapter 5, section 5.5.3) — an ongoing feature of Brazilian economic life.

Given the refractory nature of native lignocelluloses, it is not surprising that chemical processing techniques using acids or alkalis and elevated temperatures have been essential for their use as industrial materials. Conventionally, the starting point has been feedstock material such as wood chips, sawdust, and chopped stalks and stems from herbaceous plants.19 Mechanical size reduction is unavoidable and, therefore, has both economic and energy costs unless fragmented waste or by-products (e. g., sawdust) is the starting material.20

Diverse techniques have been explored and described for the pretreatment of size-reduced biomass materials with the aim of producing substrates that can be more rapidly and efficiently hydrolyzed — by either chemical or biological (enzymic) means — to yield mixtures of fermentable sugars. Physical and thermochemical methods described in the literature are summarized in table 2.1. These approaches have in common the use of conditions and procedures to greatly increase the surface area to which aqueous reactants and/or enzymes have access, in particular, the percentage of the major cellulosic materials that are opened up to attack and thereby reduced to glucose and oligosaccharides on hydrolysis within feasible time limits in batch or continuous processes.

Milling has been little favored because the fibrous nature of lignocellulosic materials requires lengthy processing times and unacceptably high energy inputs; only compression milling has been taken to a testing scale beyond the laboratory. Nevertheless, several studies have concluded that milling can greatly increase the susceptibility to enzymic depolymerization of cellulose.21 Irradiation with gamma rays and electron beams was a research topic from the 1950s to the 1980s; fragmentation of polysaccharides and lignin was demonstrated to increase the rates of hydrolysis of

cellulose when subsequently treated with acids or enzymes, but contradictory results, differential responses when using different wood species, and high investment costs meant that no irradiation technique progressed to pilot-scale evaluation.

Both milling and irradiation give single product streams with only minor degradation of lignocellulosic polymers. Thermochemical methods, in particular those using steam explosion,[12] can result in extensive degradation of hemicelluloses.22 Potentially, therefore, a twin-product stream process can be devised by separating solid and liquid phases, the former containing the bulk of the cellulose and the l atter the pentose and hexose components of hemicelluloses, although these may be predominantly present in oligosaccharides.23 At temperatures close to 200°C, even short (10-minute) pretreatment times have major impacts on surface area and enzyme accessibility (figure 2.4). Lignin-carbohydrate bonds are disrupted, some of the l ignin is depolymerized, and much of the morphological coherence of the lignified plant cell wall is destroyed.22 In addition, aqueous extraction at elevated temperatures removes much of the inorganic salts — this is of particular importance with feedstocks such as wheat straw whose combustion (or combustion with coal, etc.) is impeded by their high salt content and the consequent corrosion problems.2425

Chemical pretreatment methods have usually implied hydrolytic techniques using acids and alkalis, although oxidizing agents have also been considered (table 2.2). The use of such chemical reactants introduces a much higher degree of polysaccharide breakdown and greater opportunities for separately utilizing the various potential substrates in lignocellulosic materials. In fact, chemical fractionation procedures for plant cell walls have often been described and have been of inestimable value in the separation and structural elucidation of the cell wall polymers in plant cells and plant organs.26 With wheat straw, for example, sequential treatments (figure 2.5) with

FIGURE 2.4 Efficacy of steaming pretreatments with birch wood. (Data from Puls et al.23)

an aqueous methanol, sodium chlorite, and alkali yield distinct extractives, lignin, cellulose, and hemicellulose fractions.27 Similarly, sequential treatments with alkali (lignin removal) and dilute acid (hemicellulose hydrolysis) to leave a highly enriched cellulosic residue have been devised for a variety of feedstocks including switchgrass, corn cob, and aspen woodchip.28 Because the usual intention is to utilize the sugars present in the polysaccharides as fermentation substrates, however, the developments for bioindustrial applications have invariably focused on faster, simpler, and more advanced engineering options, including some that have been progressed to the pilot- plant scale. Pretreatments involving acids (including SO2 steam explosion) primarily solubilize the hemicellulose component of the feedstock; the use of organic solvents and alkalis tends, on the other hand, to cosolubilize lignin and hemicelluloses. As with thermochemical methods, the product streams can be separated into liquid and solid (cellulosic) phases; if no separation is included in the process, inevitably, a complex mixture of hexoses and pentoses will be carried forward to the fermentation step.

Combinations of physical, thermochemical, and chemical pretreatments have often been advocated to maximize cellulose digestibility by subsequent chemical or

enzymic treatments; this usually involves higher capital and processing costs, and the potential economic benefits of increased substrate accessibility have seldom been assessed in detail. Table 2.3 presents a historical sequence of pilot plants in North America, Japan, New Zealand, Europe, and Scandinavia developed for the processing of lignocellulosic feedstocks to illustrate different approaches to pretreatment choices and cellulose processing; not all of these initiatives included fermentation steps producing ethanol, but every one of them was the result of intentions to generate sugar solutions suitable for subsequent fermentative treatments. Different biomass feedstocks may require different technologies for optimized upstream processing; for example, ammonia-based pretreatments (ammonia fiber explosion and ammonia — recycled percolation) are more effective with agricultural residues (including corn stover and corn straw) than with woody materials.28 Hardwoods yield higher degrees of saccharification after steam explosion than do softwoods.22 An organic base, я-butylamine, has been recommended for pretreatment of rice straw on the evidence of efficient delignification, highly enhanced cellulose hydrolysis by cellulase, the ease of recovery of the amine, and the almost complete reprecipitation of the solubi­lized lignin when the butylamine is removed by distillation.2930

Many accounts of pretreatment optimizations have been published; over a decade ago, a report commissioned by the Energy Research and Development Corpora­tion of Australia estimated that “several thousand” articles dealt with physical and chemical pretreatment methods for lignocellulosics (including by-products from the paper and pulping industries).31 There are several reasons for this sustained research effort, including the large number of lignocellulosics of potential industrial use, the multiplicity of pretreatment methods (tables 2.1 and 2.2) singly and in combinations, uncoordinated funding from national and international agencies, and the various scales, from the laboratory bench up to demonstration units with the capacity of processing two tonnes/hr of feedstocks (table 2.3). A multiauthor review in 2005 of four thermochemical methods, two pretreatments with acid, two with ammonia, and one with lime (calcium hydroxide) as an alkali — all described as “promising tech­nologies” — concluded that although all nine approaches gave positive outcomes on increasing accessible surface area and solubilizing hemicelluloses and although all but one altered lignin structure, only five could reduce the lignin content, and only two (the ammonia-based methods) decrystallized cellulose.32 Exceptions and caveats were, however, noted; for example, ammonia fiber/freeze explosion worked well with herbaceous plants and agricultural residues and moderately well with hard­woods, but poorly with softwoods.20

Detailed comparisons of different pretreatment methods in controlled, side-by­side studies of multiple technologies using single feedstocks are very rare. A collabo­ration between the National Renewable Energy Laboratory and six universities in the United States compared ammonia explosion, aqueous ammonia recycle, controlled pH, dilute acid, flow-through with compressed hot water and lime approaches to prepare corn stover for subsequent biological conversion to sugars; material balances and energy balances were estimated for the processes, and the digestibilities of the solids were assessed by a standardized cellulase procedure.33-39 With this feedstock (already a major “waste” product resulting from the corn ethanol industry), all six pretreatment options resulted in high yields of glucose from cellulose by subsequent treatment with cellulase; in addition, the use of high-pH methods offered potential

Pilot Plants Developed for the Saccharification of Lignocellulosic Biomass

b Direct microbial conversion by cellulase-secreting ethanol producer

for reducing cellulase amounts required in cellulose hydrolysis.40 Differences were, however, observed in the kinetics of sugar release that were sufficient to influence the choice of process, enzymes, and fermentative organisms. This conclusion was foreshadowed by Swedish research on steam pretreatment of fast-growing willow (Salix) with or without SO2 impregnation that showed that, while glucose yields of more than 90% and overall xylose yields of more than 80% could be obtained both with and without SO2, the most favorable pretreatment conditions for the separate yields of glucose and xylose were closest when using SO2-impregnated wood chips.41 To a large extent, therefore, all pretreatment strategies are likely to include a partial compromise because of the very different susceptibilities to hydrolytic breakdown and solubilization of cellulose and hemicelluloses; highly efficient industrial solutions will require biotechnological approaches to provide fermenting organisms capable of using both hexoses and pentoses and both monomeric and oligomeric (and possibly polymeric) carbohydrates (this is discussed in detail in chapter 3). Even for a single choice of pretreatment method, variation in the biological material (the feedstock) will inevitably occur in, for example, the water content that will either necessitate a flexible technology or extra cost outlay to standardize and micromanage the inflow of biomass material.42

The most recent development in pretreatment technologies has been the dem­onstration that microcrystalline cellulose can be readily solubilized and recovered using a class of chemicals called “ionic liquids.” These are salts that are liquids at room temperature and stable up to 300°C; their extreme nonvolatility would also have minimal environmental impact.43 With one such ionic liquid (an я-butyl-methy- limidazolium chloride), cellulose could be solubilized by comparatively short (<3-hr) treatments at 300°C, the cellulose could then be recovered by the addition of “anti­solvents” such as water, methanol, and ethanol, and the resulting cellulose was 50­fold more susceptible to enzyme-catalyzed hydrolysis as compared with untreated cellulose.42 Chinese researchers have shown that ionic liquids can successfully pre­treat materials such as wheat straw; full commercialization still requires economic synthesis routes and toxicological assessments.44,45

3.1 TRADITIONAL ETHANOLOGENIC MICROBES

The fundamental challenge in selecting or tailoring a microorganism to produce eth­anol from the mixture of sugars resulting from the hydrolysis of lignocellulosic feed­stocks is easily articulated: the best ethanol producers are incompetent at utilizing pentose sugars (including those that are major components of hemicelluloses, that is, D-xylose and L-arabinose), whereas species that can efficiently utilize both pentoses and hexoses are less efficient at converting sugars to ethanol, exhibit poor tolerance of high ethanol concentrations, or coproduce high concentrations of metabolites such as acetic, lactic, pyruvic, and succinic acids in amounts to compromise the efficiency of substrate conversion to ethanol.1-4

Because bioprospecting microbial species in many natural habitats around the global ecosphere has failed to uncover an ideal ethanologen for fuel ethanol or other industrial uses, considerable ingenuity has been exhibited by molecular geneticists and fermentation specialists in providing at least partial solutions for the two most popular “combinatorial biology” strategies of

• Endowing traditional yeast ethanologens with novel traits, including the ability to utilize pentoses

• “Reforming” bacterial species and nonconventional yeasts to be more effi­cient at converting both pentoses and hexoses to ethanol

A third option, that is, devising conditions for mixed cultures to function synergisti — cally with mixtures of major carbon substrates, is discussed in chapter 4 (section 4.5).

Adding to the uncertainty is the attitude of the traditionally conservative alcohol fermentation industry toward the introduction of organisms that lack the accepted historic advantages of the yeast Saccharomyces cerevisiae in being generally regarded as safe (GRAS) and, by extrapolation, capable of being sold as an ingredi­ent in animal feed once the fermentation process is completed.5 At various times in the past 40-50 years, thermotolerant yeast strains have been developed to accelerate the fermentation process at elevated temperature.6 Bioprocesses have been advocated and, to varying degrees, developed, in which polymeric carbohydrate inputs are both hydrolyzed with secreted enzymes and the resulting sugars and oligosaccharides are taken up and metabolized to ethanol by the cell population, the so-called simultane­ous saccharification and fermentation (SSF) strategy.[18] [19] [20] Because the potential advan­tages of SSF are best understood in the light of the differing fermentation hardware requirements of multistage and single-stage fermentations, consideration of SSF technologies is postponed until the next chapter (section 4.5).

As a volatile chemical compound viewed as a gasoline substitute, pure ethanol has one major drawback. Internal combustion engines burn fuels; ethanol, in comparison with the typical hydrocarbon components of refined oils, is more oxygenated, and its combustion in oxygen generates less energy compared with either a pure hydrocarbon or a typical gasoline (table 1.1). This is not mitigated by the higher density of ethanol because liquid volumes are dispensed volumetrically and higher weights in fuel tanks represent higher loads in moving vehicles; a gallon of ethanol contains, therefore, only 70% of the energy capacity of a gallon of gasoline.1121 A review of the relative merits of alternative fuels in 1996 pointed out that ethanol not only had a higher octane number (leading to higher engine efficiencies) but also generated an increased volume of combustion products (gases) per energy unit burned; these factors in optimized ethanol engines significantly eroded the differential advantages of gasoline.21 Similar arguments could not be extended to a comparison between ethanol and diesel fuel, and ethanol had only 58 to 59% of the energy (net heat of combustion) of the latter.21

The high miscibility of ethanol and refined oil products allows a more conserva­tive option, that is, the use of low-ethanol additions to standard gasoline (e. g., E10: 90% gasoline, 10% ethanol) and requires no modifications to standard gasoline­burning vehicles. Dedicated ethanol-fueled cars were, however, the initial favorite of the Brazilian Alcohol Program (PROALCOOL); sales of alcohol-powered vehicles reached 96% of total sales in 1980 and more than 4 million such vehicles were esti­mated to be in the alcohol “fleet” by 1989.22 Such high market penetration was not, however, maintained, and sales of alcohol-powered vehicles had almost ceased by 1996 (figure 1.8). The major reason for this reversal of fortune for ethanol-fueled

vehicles was the collapse in oil prices during the late 1980s and 1990s — by 1998, the real price of crude oil was very similar to that before November 1973 (figure 1.3). Ethanol production from sugarcane in Brazil increased from a low and declining production level in early 1972, by nearly 20-fold by 1986, and then continued to increase (although at a greatly reduced rate) until 1998 (figure 1.9). The government responded to the novel “crisis” of the competing ethanol-gasoline market in several ways23: [4]

• Prices of sugarcane and ethanol were deregulated as of January 1, 1997.

• Tariffs on sugar exports were abolished in 1997.

• In January 2006, the tax rate for gasoline was set to be 58% higher than that for hydrated ethanol (93% ethanol, 7% water), and tax rates were made advantageous for any blend of gasoline and anhydrous ethanol with ethanol contents of more than 13%.

Brazilian automobile producers introduced truly flexible-fuel vehicles (FFVs) in 2003, with engines capable of being powered by gasoline, 93% aqueous ethanol, or by a blend of gasoline and anhydrous ethanol.24 In 2004, “flex-fuel” cars sold in Brazil were 16% of the total market, but during 2005, sales of FFVs overtook those of conventional gasoline vehicles (figure 1.10). This was a very “prescient” develop­ment as crude oil prices, which had been only slowly increasing during 2003 and early 2004, surged to new dollar highs in 2005 (figure 1.3). Domestic demand for ethanol-containing fuels became so great that the ethanol percentage was reduced from 25% to 20% in March 2006; this occurred despite the increased production of anhydrous ethanol for blending.25 Brazil had evolved a competitive, consumer — led dual-fuel economy where motorists made rational choices based on the relative prices of gasoline, ethanol, and blends; astute consumers have been observed to buy ethanol only when the pump price is 30% below gasoline blends — equal volumes of ethanol and gasoline are still, as noted above, divergent on their total energy (and, therefore, mileage) equivalents.

Other pertinent statistics collected for Brazil for 2004-2006 are the following:23 [5]

• Real prices for ethanol in Brazil decreased by two-thirds between 1973 and 2006.

• Sao Paolo state became the dominant contributor to national ethanol production and PETROBRAS began the construction of a 1,000-mile pipe­line from the rural interior of the state to the coast for export purposes.

A significant contraindicator is that ethanol-compatible vehicles still remain a minority of the total on Brazilian roads: in 1997, before FFVs became available, ethanol — compatible vehicles were only 21% of the total of ca. 15 million.22 The introduction of FFVs in 2005 is expected to gradually improve this ratio (figure 1.10).

Another predictable but little emphasized problem is that improvements to sugarcane harvesting methods have lead to the unemployment of 8% of seasonal sugarcane workers.23 Since 1998, Brazil has restricted the traditional practice of burning sugarcane crops (to eliminate the leaves) before manual harvesting in favor of the mechanical harvesting of green canes.19 Although far from straight­forward (because the lack of burning requires changes in pest management), this change in agricultural practice has contributed to a growing surplus of energy from sugar/alcohol plants as electricity generated on-site and offered to the distribution grids.19,22

Any overall cost-benefit of Brazil’s 30-year experience of ethanol as a biofuel is inevitably colored by the exact time point at which such an assessment is made. In April 2006, crude oil prices exceeded $70, and this price was exceeded during the summer of 2006, with crude trading briefly at $78/barrel (figure 1.11). Although the emphasis on oil prices may be perceived as one-dimensional,26 it undeniably focuses attention on real historic events, especially those on a short time scale that may, if not counterbalanced by government action and/or fiscal policies, determine the success of embryonic attempts at oil/gasoline substitution — as evidenced (negatively) by the 1990s in Brazil (figure 1.8). A survey published in 2005 by

Brazilian authors summarized many official statistics and Portuguese-language publications; the major impact factors claimed for fuel ethanol production in Brazil were the following:27

• After 1975, fuel ethanol substituted for 240 billion liters of gasoline, equivalent to $56 billion in direct importation costs and $94 billion if costs of international debt servicing are included — after 2004, the severe increases in oil prices clearly acted to augment the benefits of oil substitution (figure 1.11).

• The sugar/ethanol sector presented 3.5% of the gross national product and had a gross turnover of $12 billion, employed (directly and indirectly)

3.6 million people, and contributed $1.5 billion in taxation revenues; approximately half of the total sugarcane grown in Brazil in 2003 was dedicated to ethanol production.

• In 2004, sugarcane production required 5.6 million hectares and represented only 8.6% of the total harvested land, but more than 120 million of low — productivity pasture, natural pastures, and low-density savannas could be dedicated to sugarcane production for ethanol, with a potential ethanol yield of more than 300 billion liters/year.

Ethanol became a major exported commodity from Brazil between 1998 and 2005; exports of ethanol increased by more than 17-fold, whereas sugar exports increased by less than twofold, although price volatility has been evident with both commodities (figure 1.12).25 As a report for the International Bank for Reconstruction and Development and World Bank (first published in October 2005) noted, average

wages in the sugar-ethanol sector are higher than the mean for all sectors in Brazil.28 As a source of employment, sugarcane ethanol production directly employs more than 1 million people and is far more labor-intensive than the petrochemical indus­try: 152 times more jobs are estimated to have been created than would have been the case from an equivalent amount of petroleum products.29

Despite the apparent vibrancy of ethanol production in Brazil, ethanol use amounts to only 20 to 30% of all liquid fuels sold in Brazil.27 True levels of sub­sidies remain difficult to accurately assess; for example, public loans and state — guaranteed private bank loans were estimated to have generated unpaid debts of $2.5 billion to the Banco do Brasil alone by 1997.28 The ban on diesel-powered cars has also artificially increased fuel prices because diesel prices have been generally lower than gasohol blend prices.2728 PROALCOOL had invested $11 billion before 2005 but, by that time, could claim to have saved $11 billion in oil imports.29

Viewed from the perspectives of fermentation technology and biochemical engineering, ethanol production in Brazil improved after 1975; fermentation productivity (cubic meters of ethanol per cubic meter of fermentation tank capacity volume per day) increased by 130% between 1975 and 2000.27 This was because of continuous incremental developments and innovations; no reports of radically new fermentor designs in Brazil have been published (although very large fermentors, up to 2 million liters in capacity, are used), and ethanol concentrations in batch fermentations are in the 6 to 12% (v/v) range; the control of bacterial infection of fermentations has been of paramount importance, and selection of robust wild strains of the yeast S. cerevisiae has systematized the traditional experience that wild strains frequently overgrow “laboratory” starter cultures.30 The use of flocculent yeast strains and the adoption of continuous cultivation (chapter 4, section 4.4.1) have also been technologies adopted in Brazil in response to the increased production of sugarcane ethanol.31 Technical development of downstream technologies have been made in the largest Brazilian provider of distillation plants (Dedini S/A Industrias de Base, www. dedini. com. br): conventional (bubble cap trays), sieve tray, and azeotropic distillation methods/dehydration (cyclohexane, monoethylene glycol, and molecular sieving) processes operate at more than 800 sites — up from 327 sites before 2000.31

On a longer-term basis, genomic analysis of sugarcane promises to identify plant genes for programs to improve sugar plant growth and productivity by genetic engineering.32 33

In contrast to (thermo)chemical pretreatments, the use of microbial degradation of lignin to increase feedstock digestibility has several advantages:19

• Energy inputs are low.

• Hardware demands are modest.

• No environmentally damaging waste products are generated.

• Hazardous chemicals and conditions are avoided.

All of these features have associated economic cost savings. Against this, the need for lengthy pretreatment times and the degradation of polysaccharides (thus reduc­ing the total fermentable substrate) have acted to keep interest in biological pre­processing of lignocellulosic materials firmly in the laboratory. A careful choice of organism (usually a wood-rotting fungus) or a mixture of suitable organisms can, however, ensure a high degree of specificity of lignin removal.4748 Extrapolations of this approach could involve either the preprocessing of in situ agricultural areas for local production facilities or the sequential use of the biomass feedstock first as a substrate for edible mushroom production before further use of the partially depleted material by enzymic hydrolysis to liberate sugars from polysaccharides.