Fuel Etanol from Cellulosic Biomass

Fuel Etanol from Cellulosic Biomass

(Parte 1 de 3)

Fuel Etanol from Cellulosic Biomass LEE R. LYND, JANET H. CusHmAN, ROBERTA J. NICHOLS, CHARLES E. WYMAN

Ethanol produced from cellulosic biomass is examined as a large-scale transportation fuel. Desirable features include ethanol's fuel properties as well as benefits with respect to urban air quality, global climate change, balance of trade, and energy security. Energy balance, feedstock supply, and environmental impact considerations are not seen as significant barriers to the widespread use of fuel ethanol derived from cellulosic biomass. Conversion economics is the key obstacle to be overcome. In light ofpast progress and future prospects for research-driven improvements, a cost-competitive process appears possible in a decade.

from sugar cane in Brazil and from corn and other starchrich grains in the United States, ethanol also can be made from cellulosic materials such as wood, grass, and wastes. The technology for ethanol production from cellulosic materials is fundamentally different from that for production from food crops. Failure to appreciate this difference has resulted in misconceptions about the potential of ethanol as a large-scale transportation fuel in the United States. This artide reviews the current state and future potential of technology for producing ethanol from cellulosic biomass. The focus is on the use of ethanol as the primary fuel component on a scale exceeding that possible with low-level ethanol-gasoline blends.

Ofthe four major energy sources in the United States, petroleum supplies the largest share of total energy used and has the highest fraction imported, both by significant margins (Table 1). The domestic supply ofconventional petroleum is also the most limited ofour major energy sources. Imported oil accounted for about 4% of the 1989 foreign trade deficit (1), and total petroleum expenditures were equal to about 2% of the gross national product (2, 3). This already prominent role for petroleum in the national economy is expected to increase as domestic oil exploration and production become more expensive and as the cost and volume of imports increase (4). Energy use by the transportation sector totaled 2 quad

(1 quad = 1015 Btu) in 1989 and accounted for more than 60% of total petroleum consumption (2). Furthermore, the transportation sector, with its nearly total dependence on petroleum, has virtually no capacity to switch to other fuels in the event of a supply disruption (5). Air pollution is an important factor motivating interest in alter-

L. R. Lynd is with the Thayer School of Engineering, Dartmouth College, Hanover, NH 03755. J. H. Cushman manages the Biofuels Feedstock Development Program Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN

37831. R. J. Nichols is with the Environmental and Safety Engineering Staff, Ford Motor Company, Dearborn, MI 48121. C. E. Wyman manages the Biotechnology Research Branch, Solar Energy Research Institute, Golden, CO 80401.

native fuels. At the local level, about 100 areas in the United States exceed national ambient air quality standards (NAAQS) for ozone

(6), affecting more than half the population (7). Limits set by the NAAQS for carbon monoxide are exceeded in more than 40 areas

(6). At the global level, carbon dioxide (CO2) is responsible for more than halfthe projected anthropically mediated climate change

(8). Transportation fuels account for 27% ofthe 3.3 billion metric tons ofCO2 released annually in the United States from combustion of fossil fuels (9). Vehicles account for 4.7% of total worldwide anthropic CO2 emissions, with U.S. vehicles being responsible for 2.5% of total emissions (10).

Ethanol as a Fuel

Production and utilization. Fuel ethanol production by fermentation of starch crops is about 0.8 billion gallons (-0.06 quad) (1) in the United States, with ethanol selling for about $1.20 per gallon (12). The effective price to the blender is lowered by more than $0.50 per gallon by federal and state tax incentives (13, 14), without which fuel ethanol would not now be cost competitive.

Low-level ethanol-gasoline blends, consisting predominantly of gasoline, may use ethanol directly or indirectly, the latter in the form ofethyl tert-butyl ether (1, 15). About 7% ofall gasoline sold in the

United States currently contains fermentation-derived ethanol, and

10% blends are covered by the warranty of all U.S. automobile manufacturers. Both direct and indirect blends increase octane and also increase fuel oxygen content, facilitating more complete combustion in older cars.

Ethanol may be used as a primary fuel either in neat (unblended) form or with small amounts ofgasoline. E1oo and E85 refer to neat ethanol and an 85% ethanol-15% gasoline blend, respectively; similar terms are used for methanol. About 3 billion gallons of ethanol are used annually in Brazil, primarily as a neat fuel (14). Ethanol was used sporadically as a primary fuel in the first halfofthe 20th century in both the United States and Europe (16). Fiat, Ford, General Motors, and Volkswagen have marketed automobiles designed for use ofhydrous (water-containing) ethanol in Brazil (17). Alcohols are in many respects superior to gasoline as fuels for spark-ignited engines (18-20). Ethanol has fuel properties similar to those ofmethanol; differences between the alcohols and gasoline are much greater than differences among the alcohols (20-2). Combustion of ethanol in internal combustion engines designed for alcohols will give higher thermal efficiency and power than combustion of gasoline in conventional engines (19, 20, 2). Ford has concluded that cold-starting problems have been solved for E85 and

M85 for some applications, but not for Eioo or M1oo. A significant development for the use of alcohol fuels is the flexible fuel vehicle, which has the potential to operate on any mix ofethanol, methanol, and gasoline (5, 20). Ford's experience, as well as estimates and data from others (23,

SCIENCE, VOL. 251 on May 13, 2009 w.sciencemag.org Downloaded from

Table 1. Selected data for U.S. energy utilization. Consumption, depen- dence, and import data are from (2) for 1989 (1 quad = 10' Btu = 1.06 x 1015 kI). Oil and gas reserves are from (5) and are for conventional reserves only. Total recoverable reserves are the sum of measured, indicated and inferred, and undiscovered reserves; economically recoverable reserves are a smaller quantity. Coal reserves are from (56).

Ratio of estimatedAnnual Sector with Amount total

Energy con: greatest imported recoverable source sumption dependence (%) reserves to

(quad) utilization rate (years)

Natural gas 19.5 Residential-com- 6.7 35 mercial (3%)

*U.S. uranium reserves are the largest in the world (56).

24), indicates that approximately 1.25 gallons ofethanol are needed to travel the same distance as that obtained from 1 gallon ofgasoline in optimized engines. At the 1989 average wholesale gasoline price of $0.655 per gallon (2), the selling price required for neat ethanol to compete with gasoline on an unsubsidized basis is $0.52 per gallon. In the year 2000, with crude oil at the $28 (1989) per barrel midrange price predicted by the Department of Energy (DOE) (4), gasoline can be expected to have a wholesale price of about $0.8 per gallon (25), and a price of $0.70 per gallon would be required for ethanol to be competitive as a neat fuel.

Air-quality impact. The Environmental Protection Agency (EPA)

(2) has stated that significant long-term environmental benefits are available from the use of ethanol, methanol, or compressed natural gas as pure fuels in engines designed to take full advantage of their combustion properties. The prospect of emission reductions has motivated California to consider widespread substitution of meth- anol for gasoline and diesel fuel (26) and is also the driving force behind amendments to the Clean Air Act. Most air-quality calcula- tions, including Ford's (27), have shown some improvement in urban ozone levels and a decrease in air toxics accompanying methanol use. Similar improvements are expected for ethanol because the differences between ethanol and methanol with respect to air pollution impact are likely to be small relative to the differences between either alcohol and gasoline (24, 28). Although the magni- tude of anticipated improvements is small (probably 0 to 15%, depending on meterologic conditions, the source of pollutants, and the model used), they are still significant because ozone reduction is so difficult to achieve.

Biomass Feedstocks

Feedstock options. Representative feedstocks for ethanol production include hardwood, a cellulosic raw material that can be grown

as an energy crop; municipal solid waste (MSW), a prominent waste material; and corn, the primary raw material for the current U.S. fuel ethanol industry. Table 2 presents ethanol yields and the cost and energy inputs associated with production of these feedstocks. The cost of wood without coproduct credits ($0.29 to $0.40 per gallon) does not preclude selling ethanol at prices expected to be competitive with gasoline in the year 2000. The cost of separated MSW is potentially negative and often small relative to the required price for ethanol.

The cost of feedstock less coproduct credits can be quite low for corn (29), but only at low production levels. At levels higher than about 0.3 quad, the prices of both corn and grain would experience strong upward pressure (30). At the ethanol production potential of the current U.S. corn crop (-1.5 quad), by-product markets are expected to be saturated (31). At higher production levels, and thus without coproduct credits, corn is unlikely to be a feasible fuel ethanol feedstock in the absence of subsidies. Thus, although coproduction ofethanol and animal feed from corn may be desirable at low production levels and paves the way for cellulose-based technologies, economic considerations indicate that ethanol produced from corn cannot displace current transportation fuels to any significant extent. The energy required for production of wood (~15% of the potential ethanol combustion energy) is acceptably small for a process devoted to production of a useful form of energy and is at least two times smaller than that required for corn production

(Table 2). Source-separated MSW has no energy requirements related to its use as a feedstock for ethanol production. Potential ethanol yields per unit mass are nearly equal for corn and hardwood and are somewhat less for MSW.

Supply ofcellulosiwfeedstocks. Sources of cellulosic materials can be divided into wastes from processes undertaken for purposes other than fuel production and crops grown specifically for fuel production. The primary waste categories are agricultural residues, forestry residues, and MSW. Table 3 presents ethanol production potentials for these wastes, which total about 4 quad. Nonwaste cellulosic feedstocks may be woody or herbaceous high-productivity energy crops (HPECs) or may be trees produced by conventional forestry. Categories of land that might supply feedstocks include forest land that is not potential cropland and cannot support HPECs, existing cropland (cropland potentially available for energy crop production as a result ofexcess agricultural capacity), and potential cropland (land now in noncrop uses that could grow crops). For land categories capable of supporting HPECs, a range of ethanol production potentials is presented in

Table 3, with the low value being based on the average productivity believed to be achievable with today's technology and the high value being based on productivities projected for future technology. The considerable ethanol production potential of cropland idled in 1988 (3.0 to 5.9 quad) is likely to be a conservative estimate of future production potentials from existing cropland. A recent report to the Secretary of Agriculture (32) recommended that the devel- opment of new, nonfood products use the productive capacity of at

Table 2. Properties of potential ethanol feedstocks. All values are for potential ethanol calculated as reported in (57), with the fraction of total sugars fermented being 0.95 for corn and 0.9 for wood and MSW and a fermentation yield (mass ethanol per mass carbohydrate fermented) of 0.46.

Cost ($/gallon of ethanol) Energy for feedstock Feed- productiont stock Feedstock* Feedstock less (fraction of ethanol YieldS coproductst combustion energy)

*Wood: HPEC produced for $30 to $45 (1990) per dry ton (58). MSW: a representative price paid for separated recyclable fiber, $15 per ton, seen as a reasonable upper limit for the feedstock cost. Corn: at $1.41 to $3.16 per bushel, the range for the period 1981 to 1989 (59). tWood: reflects value as a boiler fuel; calculated with a lignin content of 21% by weight valued at $0.02 per pound (60). Corn: reflects value of animal feed coproducts (59). *Wood: HPEC (61) calculated from the methodology presented m (57); lower productivity methods have about half the indicated energy requirement. MSW: see text. Corn: see (46, 62). SYield in terms ofmass of potential ethanol per mass of dry feedstock.

ARTICLES 1319 on May 13, 2009 w.sciencemag.org Downloaded from

Table 3. Land availability and production potential for cellulose ethanol.

Land available Production Cellulose source (million acres) potential* (quad)


Agricultural 1.7 Forestry 1.4 MSW 0.7

Croplandt Idled (1988)/excess (2012) 78/150 (3-5.9)/(5.7-1.4)

*Yield assumptions are consistent with those described in Table 2. tData are averages from sources compiled in (57); most sources consider the need to maintain soil fertility in estimating collectible agricultural and forest wastes. tExisting idled cropland for 1988 is from (63); anticipated excess cropland for 2012 is from (32), potential cropland is from (64). Ethanol production levels given correspond to the DOE Biofuels Feedstock Development Program's best estimates for current and anticipated HPEC productivity. Values (dry tons per acre per year, current/anticipated) are 5/10 for idled and excess cropland and 3/8 for potential cropland. §From (31), with calculation as described (51). 1 Low value is for current HPEC productivity and 1988 existing cropland; high value is for anticipated HPEC productivity and projected available cropland.

least 150 million acres in the next 25 years. This large quantity of cropland is an indication that the historic problem of excess agricultural capacity in the United States is expected to continue and worsen. At projected land availability and energy crop productivities, excess cropland has an ethanol production potential of 1.4 quad. A large-scale fuel ethanol industry might be further supported by land in the potential cropland category. Because of the relatively low productivity of forest land and the consequent large land areas and loss ofwildlife habitat accompanying the use of forest land for significant fuel production, this category may be less desirable for production of feedstocks.

Given the gap between current production of cellulosic materials for fuel and the production necessary to support a large-scale fuel ethanol industry, estimates for total ethanol production capacity are uncertain. The data presented above, however, suggest that cellulosic materials potentially available from energy crops, wastes, and conventional forestry could provide an amount ofethanol commensurate with current consumption ofliquid transportation fuels in the United States. Previous lower estimates of the cellulose resource base (3) differ from the estimate presented herein in that they primarily considered wastes. Because HPECs have a time to harvest of less than 1 year to 10 years, depending on the crop selected, production of cellulosic feedstocks could be accelerated rapidly.

Environmental impacts. Perennial cellulosic energy crops can be grown on marginal cropland with much less erosion risk than annual row crops, such as corn. Potential erosion risk should be limited to 1 to 2 years during stand establishment. Stand life for HPECs and perennial grasses is uncertain but is thought to be in the range of 10 to 25 years (34). Longer lived production systems could be used on erosive sites. Annual cellulosic energy crops could be grown on higher quality, less erosive cropland, perhaps in a crop rotation with conventional food crops. On the basis of the field research of the DOE Biofuels Feedstock Development Program, perennial HPECs such as short-rotation hardwoods and grasses require substantially less fertilizer and pesticides than corn (35, 36). Perennial species can translocate and reuse nutrients, and herbicide use is limited to 1 or

2 years at stand establishment.

Available information suggests that perennial cellulosic energy crops are more environmentally benign than conventional annual row crops. More experience with large-scale production is needed to confirm the expectation of investigators in the field that environmental problems accompanying well-managed production of cellulosic energy crops will be relatively minor for most sites.

(Parte 1 de 3)