Biocommodity Engineering
Biocommodity Engineering
Biocommodity Engineering
ARTICLES
Biocommodity Engineering
10.1021/bp990109e CCC: $15.00 © 1999 American Chemical Society and American Institute of Chemical Engineers
Published on Web 09/17/1999
778 Biotechnol. Prog., 1999, Vol. 15, No. 5
Table 1. Comparison of Commodity and High-Value tion volume has been observed many times, with oil and
Bioproductsa utility industries offering prominent examples (9, 19).
high-value commodity Indeed, even production of pharmaceuticals must follow
product(s) (e.g. product(s) (e.g. this evolution at least to some degree as patents on early
feature biopharmaceuticals) ethanol) generation products expire, product development costs
product value are recovered, and competition increases.
$/g g104 10-4
$/L unseparated 10 to 100 10-5 III. Feedstocks
broth
Feedstocks for biocommodity processes have a large
contribution to selling price (%) and often dominant impact on process economics, siting
raw materials 0.1 to 1 up to 75% of commercial facilities, environmental benefits and
cost of production <30b >90
impacts, and process development. A diversity of poten-
potential markets (for individual products, assuming mature tial feedstocks are available in the form of residues from
technology) the forest products, agricultural, and other established
$billion/yearc 1-10 s potentially 10 s
industries as well as from dedicated crops for the purpose
Kg/year usually <1 potentially >1011
of providing feedstocks for biocommodity processes. Feed-
a Values given are approximate, with order of magnitude
stocks with both large-scale availability and amenability
variation possible and even likely in some cases. High-value to biological processing generally fall into the categories
product data are based on personal communication of David
of sugar-rich crops such as sugar cane and sweet sor-
DeLucia (formerly of Verax Inc., Lebanon, NH) except where
otherwise noted and are representative of mammalian cell produc- ghum, oils of plant or animal origin, corn or other starch-
tion systems in the biopharmaceutical industry. Commodity rich grains, and cellulose-rich materials in either woody
product values are the author’s estimates based on mature or herbaceous form. Because of their high degree of
technology such as that envisioned in ref 15. b From Angus reductance compared to carbohydrates, oils offer some
Macdonald (Macdonald & Associates, Providence RI). Includes interesting and unique biological processing possibilities.
costs for capital and capital-related costs (insurance, maintenance, The question of whether oils can be cost-effective feed-
taxes) and operating costs including labor and raw materials.
stocks for biocommodity processes is impacted strongly
Excludes fees associated with licensing, clinical trials, obtaining
FDA approval, and recovery of R&D costs. c The largest markets by the extent to which value can be found for the nonoil
for individual biopharmaceutical products are just over $1 billion portions of oil-producing plants. a situation not unlike
currently (16). Biocommodity market sizes are consistent with the use of transgenic plants (Section IV-C). Materials rich
projected markets for biologically produced transportation fuels in soluble sugars suffer from seasonal availability and
(17) and bulk plastics (18). higher price as compared to other sources of carbohydrate
and thus have limited potential as biocommodity feed-
from that of biocommodity products, with 8 orders of stocks in most locations.
magnitude being a representative ratio. Although com- Because of low cost, plentiful supply, and amenability
modity products are usually present in higher concentra- to biotechnology, carbohydrates appear likely to be the
tions prior to separation as compared to pharmaceuticals, dominant source of feedstocks for biocommodity process-
product value per volume of unseparated broth is still ing. Starch-rich and cellulosic materials each have im-
much greater for biopharmaceuticals, typically by over portant advantages in this context, and we expect that
5 orders of magnitude. Raw materials usually account both have important, although probably distinct, roles
for a very small fraction of the selling price of pharma- to play. Advantages to grains include an established
ceuticals, whereas raw materials are large and are often feedstock production and processing infrastructure and
dominant factors in determining the price of commodity the presence of carbohydrate in a form that is both more
products. The cost of production including capital recov- homogeneous and more reactive than that found in
ery is usually by far the dominant factor determining the cellulosic materials. Because of these advantages, corn
price of commodity products, whereas the cost of produc- is by far the dominant feedstock for biological production
tion is not nearly so important for pharmaceuticals. of commodity products today. Advantages to cellulosic
Measured in dollars, markets for individual biopharma- materials include much larger ultimate supply, lower
ceutical and biocommodity products are of relatively purchase cost and lower anticipated transfer cost (see
similar magnitude and very large. However, tremendous Section IV), less erosivity (20), and lower inputs of
differences exist with respect to market size on a mass chemicals and energy required for production (17, 20).
basis, with the largest commodity markets exceeding These features make cellulosic materials the preferred
pharmaceutical markets by approximately 11 orders of long-term feedstock for large-scale biocommodity prod-
magnitude. The production of high-volume/low-value ucts provided that cost-effective and environmentally
biocommodity products has an absolute requirement for benign technologies for overcoming the recalcitrance of
high-volume/low-value feedstocks and must be responsive cellulosic biomass can be developed (Section V-A). Corn
to the availability and characteristics of feedstocks, has an important transition role to play as an established
whereas no such requirement exists for the production feedstock for biocommodity processes while cellulose-
of pharmaceuticals. based technology matures. Furthermore, the corn plant
It was natural that the nascent biotechnology field is a potentially significant source of cellulosic feedstocks
focused first on high-value products, since such products in the form of crop residues (6). A likely permanent
have less stringent requirements for low-cost processing advantage of corn as compared to cellulosic materials is
technology as compared to commodity products, and since that carbohydrate-rich process streams of sufficient
they have the potential to be produced profitably on a purity to accommodate production of food, food additives,
relatively small scale. We anticipate continued rapid and pharmaceutical products for human consumption are
advances in biotechnology related to health care. In more easily obtained.
addition, we anticipate the emergence of biocommodity Although carbohydrate represents from 2/3 to 3/4 of the
engineering as a significant area of endeavor that will dry weight of most plant materials, substantial diversity
become increasingly distinctive over time. The evolution is exhibited among different types of biomass with respect
of technology toward lower unit value and larger produc- to individual carbohydrate components (Table 2). For
780 Biotechnol. Prog., 1999, Vol. 15, No. 5
hexan and 1.14 for pentan due to the water of hydrolysis. Fiber incluces cellulose and hemicellulose. The diversity of compositional
categories used reflects that in the literature. Corn composition from ref 21. Soybeans from ref 22. Corn stover and wheat straw from ref
23. Hardwood data are from ref 24 except for ash and extractives, which are from ref 23. Switch data are from ref 25 except for protein
which is from Bruce Dale (personal communication).
corn kernels, most of the carbohydrate fraction is made the presence of lignin in many cellulosic materials can
up of D-glucose molecules joined by alpha linkages to form lead to the export of power and the elimination of
starch, whereas the carbohydrate fraction in cellulosic external inputs of processing energy, which in turn has
biomass consists of cellulose and hemicellulose. Cellulose a large beneficial impact on net greenhouse gas emissions
is a homopolymer of β-linked glucose usually present in (2). The coproduction of animal feed protein and feed-
a highly ordered crystalline structure that impedes stocks for bioprocessing potentially offers large benefits
hydrolysis. Hemicellulose, also found in cellulosic materi- in terms of land-use efficiency.
als, is an amorphous polymer and typically contains five U.S. production of primary building blocks for the
different sugars: L-arabinose, D-galactose, D-glucose, synthesis of organic chemicals (ethylene, propylene,
D-mannose, and D-xylose. Hemicellulose often also con- benzene, methanol, toluene, xylene, butadiene) totaled
tains smaller amounts of nonsugar components such as about 64 million tons in 1997 (26). This may be compared
acetyl groups. The fiber fractions of both corn and oil to the current rate of biomass consumption by the corn
crops are composed of cellulose and hemicellulose. refining industry (52 million tons/year, ref 27) and the
Whereas the noncarbohydrate fraction of corn kernels pulp and paper industry (100 million tons/year, ref 28).
contains protein and oils, the noncarbohydrate portion Annual availability of collectable waste cellulosic bio-
of cellulosic materials is comprised of mostly lignin with mass, with allowance for maintaining soil fertility, at a
lesser amounts of ash and soluble substances termed price e$45/ton has been estimated at 140 million tons/
extractives. Lignin is a complex phenyl propene material, year (2); many estimates for the potential availability of
while the ash consists of inorganics such as silica, dedicated cellulosic crops are substantially larger than
potassium, and sodium. Early-cut herbaceous materials this value (see refs 2, 17). These observations support
can contain a significant amount of protein. the conclusion that the magnitude of the sustainably
Careful attention to the utilization of all feedstock harvestable biomass resource is sufficient to meet the
components is important in terms of process technology, demand for all petrochemicals produced in the United
economic, and environmental considerations. In the case States. The possible sufficiency of the biomass resource
of corn, oil can be recovered in the wet milling process to meet much larger fuel needs is an important and
and the presence of protein adds value to byproducts such contentious issue. We believe that there are responsible
as corn gluten meal, and corn gluten feed. In the case of scenarios in which biomass-derived fuels can meet U.S.
cellulosic materials, development of microorganisms mobility demands but defer analysis of this complex
capable of converting hemicellulose-derived pentose sug- question to a future paper.
ars (Section V-A) has resulted in significant cost benefits.
Lignin-rich process residues provide a potential source IV. Economic Framework
of aromatic chemicals at low production volumes and an
attractive fuel for power generation at high volume. For To contribute significantly to sustainable resource
appropriately harvested herbaceous cellulosic materials, supply and improved environmental quality, biocommod-
protein can potentially be recovered and sold as animal ity processes must be attractive from economic as well
feed. Economic impacts of coproduct production are as environmental viewpoints. Economic analysis is also
explored in Sections IV-A and VI. important in targeting opportunities for R&D-driven cost
Both feedstock production and process coproducts can reductions and anticipating the direction of future tech-
have profound impacts with respect to resource and nology. Aspects of the economic framework for biocom-
environmental metrics (see Section VI-C). For example, modity engineering are considered in this section with
Biotechnol. Prog., 1999, Vol. 15, No. 5 781
Table 3. Representative Prices for Selected Biomass and Table 4. Alternative Feedstock Price Metrics
Fossil Resourcesa
purchase price (P) ¢/kg feedstock
$/dry metric ton $/GJ
purchase price of carbohydrate, Pc P
fossil Pc )
Fc
oil
@ $17.5/barrel
@ $12.7/barrel
129
94
3.1
2.3
net price of carbohydrate, Nc Nc ) Pc ) ∑V
i
Pi
@ $6/barrel 44 1.2
natural gas (@ $2.50/1000 scf) 122 2.3 transfer price of carbohydrate, Tc N+o+c
coal 33 1.0 Tc )
biomass
corn % product value, F P
kernels (@ $2.5/bushel) 98 5.0 F)
stover
cellulosic
19 1.0
∑Vj
PjYPj
short rotation poplar, switchgrass 44 2.3
where:
a Prices for fossil resources from ref 29. Corn stover price from Pc ) ¢/kg feedstock
ref 6. Cellulosic price from refs 30, 31. Heating values from ref Fc ) kg PSSa/kg feedstock
32. Nc ) net price of carbohydrate (¢/kg PSS)
VPi ) product value (¢/unit product)b
respect to feedstocks, overall processing, and plant-based YPi ) product yield, carbohydrate basis
production systems. (unit product/kg PSS)
i ) coproduct index, excluding the primary product
A. Cost Competitive Raw Materials. It is logical to j ) index of all products, including the primary product
benchmark the cost of plant biomass against the cost of o ) operating costc (¢/kg PSS)
oil, the dominant source of organic fuels and chemicals c ) annualized capital costc (¢/kg PSS)
currently, as well as other prominent fossil resources, ) carbohydrate conversion efficiency
natural gas and coal (Table 3). Relative to oil at $17.5/ (actual soluble sugar/PSS)
barrel, a representative value during the 1990s (29), the a PSS ) potential soluble sugar. b Fermentable carbohydrate
price of corn at $2.50/bushel (also representative) is valued at its transfer cost in computing F. c For production of
competitive on a mass basis but not on an energy basis. carbohydrate and all coproducts.
Cellulosic materials such as short-rotation poplar and
switchgrass are expected to be very widely available at capital costs for all process steps up to and including the
$40/delivered dry ton (30, 31), a price calculated on the production of fermentable carbohydrates and coproducts
basis of fully compensating farmers for producing cel- (including waste treatment) and also adjusted for the
lulosic crops as opposed to more traditional crops. Waste carbohydrate conversion efficiency. In essence, the trans-
cellulosic materials such as corn stover, sugar cane fer price is the price of carbohydrate that a company
bagasse, waste paper sludge, and municipal solid waste would have to charge itself to recover its cost at a given
are available at many locations for substantially lower return on investment. The transfer price is exclusive of
prices. Even at $40/ton, the price of cellulosic biomass is costs associated with conversion of soluble sugars to
substantially less than that of oil at $17.5/barrel on both products of interest. The percent of product value, F,
a mass and energy basis. The break-even oil price relative provides an indication of the impact of feedstock costs
to cellulosic biomass at $40/ton is $12.7/barrel on an on overall process economics.
energy basis and $6/barrel on a mass basis. Relative to Table 5 compares feedstock price metrics and other
natural gas, at $2.50/1000 scf, cellulosic biomass is price- data relevant to refining of corn (via wet milling; Ron
competitive on an energy basis and has a substantial Landucci, ProForma Systems, Inc., personal communica-
price advantage on a mass basis. The purchase price of tion), cellulose processing via current technology, and
dedicated biomass crops approaches being cost competi- cellulose processing via projected mature technology.
tive with coal on a mass basis but not on an energy basis. Costs for cellulose processing via current and mature
Both corn and cellulosic materials compare more favor- technology are based on the base-case and, most likely,
ably on a mass basis than an energy basis because of the advanced case scenarios for ethanol production from
more reduced and oxygen-poor character of fossil re- poplar defined by Lynd et al. (15). We expect that the
sources. It may be noted that the prices of oil, corn, and cost of processing cellulosic biomass would be similar for
pulpwood all have varied by more than 2-fold during the herbaceous cellulosic feedstocks such as switchgrass. The
1990s. data in Table 5 illustrate the potential of cellulosic
Today’s corn and oil refineries convert these chemically feedstocks as low-cost sources of carbohydrate for biologi-
diverse raw materials into an array of saleable products. cal processing, as the purchase and net price is substan-
We believe this will also be true of cellulose refineries of tially lower for both cellulose scenarios as compared to
the future. Beyond the simple purchase price, other corn. The carbohydrate-transfer price is similar for the
feedstock cost metrics can be defined to reflect the fact corn wet mill and current cellulosic biomass scenarios
that coproducts produced from noncarbohydrate portions but is substantially lower for the mature cellulosic
of biomass feedstocks can impact the effective cost of scenario. For the corn wet mill scenario, comparison of
carbohydrate for biological processing. Alternative price the net and transfer prices indicates that, although
metrics relevant in this context are listed in Table 4. The substantial coproduct value is realized, the costs associ-
purchase price of carbohydrate, Pc, is the purchase price ated with this realization are comparably large. For the
on a dry mass basis divided by the potential soluble sugar current cellulosic scenario, the transfer price is more than
content of the feedstock expressed as a mass fraction. The twice the purchase price, reflecting primarily the high
net price of cabohydrate, Nc, represents the purchase cost of overcoming the recalcitrance of cellulosic biomass
price less the value of coproducts. The transfer price of using current technology (see Section V. A.). Realization
carbohydrate, Tc, represents the net price adjusted for of the transfer price for the current cellulosic scenario
coproduct revenues and for operating and annualized will require that risk and the cost of capital be lowered
782 Biotechnol. Prog., 1999, Vol. 15, No. 5
from Ron Landucci (ProForma Systems Inc). Purchase, net, and transfer costs are defined as in Table 4. Cellulosic values are from Lynd
et al. (15). Costs are calculated from the base case and most likely advanced scenarios for ethanol production, with costs for product
recovery subtracted and associated steam used to generate power. The current/base case scenario involves dilute acid hydrolysis and
SSF. Cellulosic sugar yields are for both hexoses and pentoses and reflect losses due to incomplete hydrolysis, pentose degradation during
pretreatment, and cellulase production. The mature scenario involves liquid hot water pretreatment and consolidated bioprocessing. For
both the current and mature cases, power is derived via burning lignin-rich process residues in a Rankine cycle. A consistent costing
framework is used for both corn and cellulosic plants as specified by Landucci. Features of this framework include 20 year plant life, 3
year construction period, 20% discounted cash flow rate of return, 37% combined federal and state taxes, 3% inflation, MACRS depreciation,
10 year equipment life, 31 year building service life, 25% owner equity financing, 10% effective loan rate.
to values characteristic of corn processing, which is not •Because of their significant level of oxygen substitu-
currently the case. In addition, the transfer price for tion, biomass and most biomass-derived products are
cellulosic biomass implicitly assumes ability to process more amenable to subsequent reactive transformations
both pentose and hexose sugars at high yields, which has than largely aliphatic petroleum.
been realized to date for some but not all products (see •Modern biotechnology, arguably the most powerful
Section V. A.). Realization of the transfer price for mature new development in processing technology, is much more
cellulosic technology will require R&D-driven advances readily applicable to processing plant biomass than
associated with overcoming the recalcitrance of cellulosic petroleum. In particular, biotechnology can be applied
biomass (see ref 15 and subsequent discussion herein). to production of commodity products from biomass not
It is likely that a more complete slate of coproducts, such once but twice: first in development of plants amenable
as those produced from an oil refinery, would lower the to subsequent processing and second in developing
transfer price of carbohydrate in the mature cellulosic improved biological catalysts for producing products of
scenario somewhat further. interest.
Comparison of values for the percent product value •Oil refineries are unlikely to have significant econo-
represented by feedstock, F, suggests that corn wet- mies of scale advantages relative to mature biomass
milling technology is relatively mature. That is, the refineries. This expectation is supported by the fact that
conversion processes are developed to the point that the the largest existing corn wet mills process a material flow
price of feedstock represents the largest share of total (∼11000 tonnes/day) in the range typical of an oil refinery
product value and processing cost margins are relatively (34) and further by the tremendous amount of plant
small. In particular, F values for both corn wet milling material that can be produced within a reasonable (e.g.,
(F ) 0.66, Table 5) and oil refining (F ) 0.67, based on 50 mile) radius of a processing plant (15). These advan-
the average value for U.S. refining industry for the period tages are counterbalanced to some extent by the greater
1990-1997; refs 29, 33, 34) are about 2/3, which we take difficulty of handling solid-phase biomass as opposed to
to be indicative of mature technologies for production of liquid-phase petroleum, but we think this unlikely to be
commodity products. The realism of the mature cellulosic a dominant economic factor in the long run. This expec-
biomass scenario is supported by its similar F value tation is based in part on the availability of different
(F ) 0.66). The far lower value for the current cellulosic processing paradigms suitable for processing solid ma-
scenario (0.52) is a strong indication of the relative terials, considered in Section V.
immaturity of this technology. C. Plant-Based Production Systems. Consistent
B. Cost-Competitive Processes. We believe that the with the perspective expressed in recent reviews (35-
higher current cost of biologically processing biomass as 37), we believe that the prospect of “fermentorless”
compared to nonbiological processing of petroleum is due biosynthesis in transgenic plants is an exciting develop-
to the fact that modern society has invested vastly less ment that is likely to be the preferred production mode
effort in the former as compared to the latter. Further- for some products. We also offer two cautionary com-
more, we anticipate that R&D-driven improvements in ments relevant to plant-based production systems for
technology for biologically based conversion of plant commodity products as compared to fermentor-based
biomass can, over the long term and given sufficient production from whole plants.
effort, result in commodity products priced competitively For commodity products, potential savings due to
with products from petroleum at oil prices in the range elimination of process steps via use of transgenic plants
seen in the 1990s. In addition to the data presented in can be offset by even small fractional increases in
Table 5, this view is further supported by the following feedstock costs. Consider, for example, fermenting the
observations: entire carbohydrate fraction of a plant to lactic acid or
•In light of the dominance of feedstock cost for com- ethanol, for which respective fermentation yields of 65%
modity products (Table 5), selling price tends to be and 35% based on overall plant dry matter are possible.
relatively insensitive to small differences in processing We think it exceedingly unlikely that such yields can be
costs. Said another way, the cost of biomass processing realized in a transgenic plant and that overall plant
would have to be substantially higher than the cost of productivity will be severely compromised if attaining
processing petroleum in order to have much impact on these yields does prove possible. Lower plant productiv-
product prices for mature technology. ity, P (tons dry matter acre-1 year-1), will raise the
Biotechnol. Prog., 1999, Vol. 15, No. 5 783
effective feedstock cost for transgenic plants as compared cellulosic biomass into reactive intermediates); and (2)
to nontransgenic plants that can subsequently be pro- product diversification (that is, converting reactive in-
cessed by fermentation. For the common case of a fixed termediates into useful products). Before considering
opportunity cost for land, O ($ acre-1 year-1, determined these challenges, two general observations are offered.
by the revenue that could be realized growing some First, processing paradigms most advantageous for solid
conventional crop), the feedstock price required to offset biomass feedstocks may well differ from those most
the opportunity cost of land, C ($/ton), is given by advantageous for liquid petroleum. For example, process-
ing of fluid-phase petroleum is based on conveyance of
C ) O/P fluids between time-invariant spacially discreet process-
ing environments. By contrast, a sequencing batch
For a commodity product resulting from mature conver- operating mode featuring dynamic variation of the pro-
sion technology for which processing represents 1/3 the cess environment with little or no conveyence of feedstock
of total product value (Table 5), savings realized as a between processing environments may be preferred for
result of eliminating processing costs through transgenic upstream unit operations associated with refining solid-
plant production will be canceled entirely by increased phase biomass feedstocks (e.g., pretreatment). Ap-
feedstock costs accompanying a plant productivity de- proaches to bioreactor productivity enhancement provide
crease of 33%. Moreover, this calculation is based on an a second example of different processing paradigms for
upper limit of the savings possible by realization of plant- biomass as compared to soluble substrates. Whereas
based production since in reality some processing will still retention/recycle of cells or enzymes is a standard ap-
be necessary. proach to increasing productivity for soluble substrates,
Our second cautionary comment is that satisfying a differential retention of substrates is more easily imple-
significant fraction of the demand for fuels and plastic mented and more likely to be effective for biological
monomers via processes that produce these compounds processing of solid substrates.
at the low fractional mass yields likely to be realized by Our second general observation is that processes and
plant-based production requires careful identification of biocatalysts for production of biocommodity products are
very large-scale coproducts, which will usually be an typically developed and designed on the basis of feedstock
important factor in overall process economics. Consider, characteristics. This may be contrasted to high-value
for example, transgenic production of monomers for bioproducts, for which carbon sources are usually selected
thermoplastic polymer synthesis from biomass at the on the basis of requirements of the process and biocata-
current U.S. production of over 60 billion pounds (26) and lyst. Consistent with these observations, by far the major
a 15% mass yield on plant dry matter. The 400 billion focus of biopharmaceutical manufacturing R&D is how
pounds of plant residues would exceed, for example, the to make and recover a product of interest. For biocom-
total production of wood pulp in the U.S. It is yet more modity products, such product-focused activity is also
difficult to identify suitably high-volume coproducts for important but no more so than R&D directed toward the
transgenic production of fuels at low mass yield. matter of how to utilize a feedstock of interest. Indeed, a
For products well-suited to production in plants at significant fraction of biocommodity research frontiers are
modest yields (e.g., e15% of plant dry matter), economic defined by feedstock characteristics, including overcom-
and environmental imperatives strongly favor producing ing the recalcitrance of cellulosic biomass, utilization of
an additional useful product from the residue remaining all sugars present in the feedstock, and value recovery
after the transgenic product is recovered. Microbial from noncarbohydrate feedstock fractions.
conversion in a fermentor is one of the most attractive A. Overcoming the Recalcitrance of Cellulosic
options for processing these residues. It is quite possible Biomass. The recalcitrance of cellulosic biomass is a
that sequential application of plant-based and fermentor- generic obstacle impeding the cost-effective production
based biotechnology could, in some cases, provide a viable of both fuels and chemicals from cellulose-rich materials,
means by which to get the maximum product value from as well as realizing value from residues associated with
an acre of land. In such a scenario, the relationship production of starch-rich grains. Approaches for overcom-
between these two production routes becomes comple- ing the recalcitrance of cellulosic biomass include gasifi-
mentary. However, we think that the most natural cation, acid hydrolysis, and pretreatment/enzymatic hy-
complementation is for specialty products to be produced drolysis. Process design studies have indicated that, for
in plants at low mass yield with commodity products all of these approaches, steps associated with overcoming
produced from residues. This has the effect of matching the recalcitrance of cellulosic biomass are typically the
relative production to demand, which becomes increas- most costly (15, 38, 39).
ingly necessary as biological production becomes a larger Gasification and enzymatic hydrolysis were found to
fraction of overall commodity product consumption. have roughly equal costs in a comparative study reported
by Wyman et al. in 1993 (38). Enzymatic and acid
V. Process Technology and Related Research hydrolysis have been seen as being competitive in the
Challenges emergent biomass ethanol industry in terms of both
Plant biomass, the dominant foreseeable source of process design studies and commercial activity. Thus
feedstocks for biological processing of any kind, is a cost- gasification, acid hydrolysis, and enzymatic hydrolysis
competitive raw material with low-cost petroleum whether have been thought to be roughly cost-competitive in the
considered on a mass or energy basis (Table 3) or in terms 1990s. The economics of these alternatives differ, how-
of purchase or transfer price (Table 5). This indicates that ever, with respect to their potential for future improve-
the dominant factor impeding biological production of ment. Enzymatic hydrolysis has been considered in the
commodity products is the high cost of current processing context of modern molecular biology for only a decade or
technology rather than the cost of raw materials. so, and order-of magnitude reductions in the cost of
Technological and research challenges associated with biological processing in a pretreatment/enzymatic hy-
converting plant biomass into commodity products can drolysis scenario have been forecast (15). By contrast,
usefully be grouped into two categories: (1) overcoming gasification and acid hydrolysis have been practiced and
the recalcitrance of cellulosic biomass (that is, converting understood to a substantial degree for a half century or
784 Biotechnol. Prog., 1999, Vol. 15, No. 5
more, and analyses of R&D-driven improvements project and of carbon dioxide (66) has also been considered to
more modest cost impacts. We focus on pretreatment/ reduce chemical and materials of construction costs.
enzymatic hydrolysis here because of its potential for Sulfur dioxide has been found to be effective, although
future research-driven improvements. We note that at higher cost than for sulfuric acid (67).
biomass gasification has a potentially important role to A number of attributes are very important for effective
play in energy-efficient production of electricity from pretreatment (2, 15, 42, 43, 62, 63, 68). It is vital that
biomass and that gasification-based power generation high yields of sugars are realized from the hemicellulose
from lignin-rich residues may well provide an attractive fraction and that the cellulose fiber left be very digestible
way to realize value from the residues of processes by enzymes; yields on the order of 90% and preferably
featuring enzymatic hydrolysis. closer to 100% are important from each hydrolysis
Biomass Pretreatment. For most types of biomass, reaction. A low corrosivity environment is advantageous
the enzymatic digestibility of the cellulose is very low to keep materials of construction costs reasonable. Be-
(<20%) without some type of pretreatment to open up cause energy demands for mechanical size reduction of
the structure and make it accessible to attack by enzymes biomass prior to pretreatment can be one of the largest
(40-43). Removal of either hemicellulose or lignin is in the plant, pretreatment technologies that minimize
thought to create pores that allow the enzyme to pen- this requirement are obviously desirable. Chemicals
etrate into the biomass structure (44), but elimination released as direct products from the hemicellulose hy-
of one barrier to hydrolysis can accentuate the impor- drolysis reaction (e.g., acetic acid) and by biomass
tance of another. For example, removal of hemicellulose degradation during pretreatment (e.g., furfural) can be
can increase the yields of cellulose hydrolysis via enzy- toxic to downstream biological steps, and it is advanta-
matic digestion to over 90% (45, 46), but the rate and geous to develop technologies that reduce or eliminate
yield of cellulose hydrolysis will increase further with conditioning steps, avoiding use of costly chemicals and
removal of lignin (47-49). In addition, mitigating one production of problematic residues. It is also important
barrier may actually alter another and mask its original that pretreatment involve minimal water addition to
impact. For instance, when hemicellulose is hydrolyzed reduce energy demands and produce an acceptable sugar
at high temperatures, the nature of lignin is undoubtedly and consequently final product concentration. These
changed even though it remains on the solid substrate, targets for pretreatment technology have major implica-
impacting digestibility (45-49). tions on not only the direct cost of pretreatment itself
A number of biological, chemical, and physical pre- but also that for upstream and downstream operations
treatment techniques have been investigated (42, 43). (15, 62, 63).
Physical approaches such as irradiation and comminution Although a large literature exists addressing pretreat-
tend to be slow, energy-intensive, and too costly. Biologi- ment on a phenomenological basis, few studies have
cal methods based on lignin-solubilizing organisms are examined the chemistry underlying pretreatment pro-
conceptually inviting because of their simplicity and low- cesses from a fundamental perspective. At this point,
energy demands, but they are slow and they consume gaining more insight into pretreatment fundamentals is
cellulose and hemicellulose in addition to lignin. Steam a particularly important frontier in terms of enabling
heating hydrolyzes hemicellulose with natural acids further applied advances. The knowledge gained will
released during the reaction, but reported sugar yields provide vital directions for technology advancement by
from hemicellulose are less than about 65% of theoretical clarifying cause-and-effect relationships and guiding
(50-52). Pretreatment with liquid hot water appears to selection of process configurations and conditions. It also
exhibit different and more promising behavior relative supports scale-up of pretreatment technologies by provid-
to steam pretreatment; however the mechanistic basis ing a rational basis for plant designs that engineers can
of these differences is not understood and it is not clear apply with less need for expensive and time-consuming
to what extent they are limited to low solid concentra- pilot and demonstration plant studies. Establishing solid
tions (49, 53-56). fundamentals and resultant process concepts for im-
Various chemicals have been incorporated into pre- provement and scale-up of pretreatment technologies is
treatment processes to improve the enzymatic digest- an important opportunity to enable production of com-
ibility of cellulosic materials (42, 43). Solvents such as modity products from plant biomass in much the same
ethanol and methanol or bases such as sodium hydroxide way that chemical engineering enabled petroleum refin-
dissolve lignin, but costs are so high that these methods ing to realize the impact we see today.
are not considered competitive for manufacture of high- Biotechnology for Utilization of Cellulosic Mate-
volume, low-value commodity products. Ammonia has rials. Four biologically mediated events typically occur
been employed in combination with explosive decompres- in the course of biological processing of cellulosic biomass
sion to enhance the digestibility of certain types of some using enzymatic hydrolysis: cellulase production, cel-
grasses and agricultural residues with some success (57). lulose hydrolysis, hexose fermentation, and pentose
Ammonia is relatively easy to recycle, few fermentation fermentation. Process configurations proposed for the
inhibitors are formed, the energy use and capital costs biological steps differ in the degree to which these events
are projected to be reasonable, and opportunities have are integrated. As presented in Figure 1, separate
been defined to improve the economics. However, it is hydrolysis and fermentation (SHF) involves four discrete
not clear whether the process is effective for woody process steps. Simultaneous saccharification and fermen-
materials or waste paper. Dilute acid pretreatment, and tation (SSF) consolidate hydrolysis and hexose fermenta-
particularly dilute sulfuric acid pretreatment, has been tion. Simultaneous saccharification and cofermentation
studied extensively and is considered by many to be the (SSCF) combine hydrolysis, hexose fermentation, and
leading pretreatment option at this time (58-63). How- pentose fermentation. Consolidated bioprocessing (CBP)
ever, dilute sulfuric acid pretreatment requires very accomplishes cellulase production, hydrolysis, and fer-
expensive materials of construction, additional costs for mentation simultaneously in a single step. [The term
neutralization chemicals, and handling and disposal of “consolidated bioprocessing” is synonymous with the term
large amounts of gypsum or other salts formed in “direct microbial conversion” (of DMC) used in earlier
neutralization (15, 62, 63). Use of sulfur dioxide (64, 65) literature. See ref 2 for a discussion of nomenclature.]
Biotechnol. Prog., 1999, Vol. 15, No. 5 785
As reviewed elsewhere (2, 68), detailed process analy- sive research, especially involving the aerobic fungus
ses of ethanol production from cellulosic biomass using Trichoderma reesei (75, 76), and is a substantial com-
the extensive process design framework of the National mercial activity serving the textile, food processing, and
Renewable Energy Laboratory (NREL) support the po- pulp and paper industries. Extensive fundamental lit-
tential for substantially lower processing costs for SSF erature also exists on the genetics, structure, and func-
as compared to SHF (69) and for CBP as compared to tion of cellulase enzyme systems, which is summarized
SSF or SSCF (15). We think it very likely that the trend in recent comprehensive reviews (77-80). Notwithstand-
of decreasing potential processing cost with increasing ing, the cost of cellulase is a major impediment to cost-
consolidation is applicable to most products. However, effectively processing cellulosic biomass via enzymatic
the magnitude of R&D advances required to realize this hydrolysis. A recent comprehensive study by NREL (81)
potential is higher for the more highly consolidated estimated cellulase costs at $3/gallon ethanol ($0.80/L)
strategies and for CBP in particular. for commercial enzyme formulations and $0.50/gallon
SHF, SSF, and SSCF all rely on the production of ethanol ($0.13/L) for less highly processed formulations.
cellulase in a dedicated unit operation separate from the The latter value corresponds to about 5.1¢/kg biomass
unit operation(s) used for producing a desired product. or about 8.6¢/kg carbohydrate (depending on the scenario
Key research challenges common to these configurations assumed) and, thus, roughly doubles the feedstock cost
include developing microorganisms that can utilize bio- on a purchase price basis for both the current and mature
mass-derived soluble sugars in addition to glucose under technology scenarios and on a transfer price basis for the
industrially relevant conditions and lowering the cost of mature scenario (Table 5). Although some members of
cellulase production. the Trichoderma genus are prodigious producers of
Engineering of microorganisms that can utilize xylose cellulase, other cellulases are thought to have substan-
and other nonglucose sugars has received substantial tially higher specific activities (82, 83). Thus one strategy
attention over the past decade and represents one of the for reducing the cost of cellulase production involves the
more extensively studied applications of metabolic engi- heterologous production of high-specific-activity cellu-
neering. Pursued to date largely in the context of ethanol lases. This will most likely involve aerobic processes and
production, such organism development has been based host organisms in light of the higher ATP yields and
on one of two strategies (Figure 2). The “native substrate consequently higher potential protein yields of oxidative
utilization” strategy involves beginning with a microor- phosphorylation as opposed to anaerobic fermentation.
ganism that already utilizes the substrate of interest and Improvement of specific activity via protein engineering
improving selectivity and other desired product-related is another, potentially complementary option.
features. The “recombinant substrate utilization” strategy CBP is differentiated from the other processing strate-
involves beginning with an organism that already has gies in Figure 1 in that both cellulase production and
high product selectivity and other product-related fea- production of the desired product are carried out by a
tures and conferring the ability to utilize substrates of single microbial species or microbial community. This
interest. The work of Ingram and co-workers with Es- requires that cellulose hydrolysis be viewed as a micro-
cherichia coli and Klebsiella oxytoca (70, 71) exemplifies bial phenomenon and not only an enzymatic phenom-
the former strategy, whereas the work of both Ho and enon, a perspective that has received relatively little
co-workers with Saccharomyces cerevisiae (72) and Zhang attention in the literature. It will be desirable in many
and co-workers with Zymomonas mobilis (73, 74) exem- instances to implement the CBP strategy via anaerobic
plifies the latter. Current research frontiers address microorganisms producing C2 to C4 alcohols, ketones,
improving the industrial robustness (e.g., decreased and organic acids as catabolic products. While of par-
sensitivity to inhibitors generated during pretreatment) ticular interest, anaerobic CBP also poses a particular
of recombinant strains, increasing the range of sugars challenge: can high rates be supported given the low ATP
utilized, and increasing the range of products produced. gain of anaerobic metabolism in combination with the
Production of cellulase has been the subject of exten- low substrate turnover number of hydrolytic enzymes?
786 Biotechnol. Prog., 1999, Vol. 15, No. 5
Figure 2. Alternative organism development strategies to obtain organisms useful in processing cellulosic feedstocks.
The kinetic and bioenergetic feasibility of anaerobic CBP typically targets oxygen-poor compounds for which theo-
is supported by the high rates of cellulose utilization retical yields from more highly oxygenated biomass
exhibited by naturally hydrolytic anaerobes (84, 85) and feedstocks are low. By contrast, most of the attractive
by quantitative models incorporating fermentative ATP candidates for the substitution strategy have a degree
generation, the ATP requirement for cellulase synthesis, of reduction comparable to that of their feedstocks, which
and cellulase kinetics (86). As with xylose utilization, makes high conversion yields possible.
organism development for CBP can proceed via a native These points are illustrated in Table 6 for production
substrate utilization strategy or a recombinant substrate of polyethylene, the largest volume nonfuel organic
utilization strategy (Figure 2). The native substrate material produced from petroleum. The market price of
utilization strategy involves metabolic engineering of the ethylene is substantially higher than any of the carbo-
end-product metabolism of cellulolytic microorganisms hydrate-transfer prices calculated in Table 5. For the
(e.g., Clostridium thermocellum). The recombinant sub- direct replacement strategy, the stoichiometric yield of
strate utilization strategy involves heterologous expres- polyethylene from carbohydrate with an allowance for cell
sion of cellulases in an organism whose product yield and synthesis is 0.28, reflecting the mass losses in producing
tolerance credentials are well-established (e.g., yeast, an oxygen-poor product from a more oxygen-rich reactant
Lactobacillus). Each strategy has its own advantages and in both fermentation of carbohydrate to ethanol and
challenges, and different strategies may well prove most dehydrogenation of ethanol to ethylene. At such low
advantageous for different products. A key research need yields, the feedstock price advantage of biomass materials
for the native substrate utilization strategy is the devel- is lost for all but the mature cellulosic scenario. Polymers
opment of gene-transfer systems, which are not available of lactic acid, polylactides, have relatively similar physical
for most organisms of interest in this context. The properties (e.g., strength, elongation to break) to poly-
recombinant hydrolytic strategy has been pursued most ethylene (97), and a wide range of properties can be
extensively in Saccharomyces cerevisiea, with simulta- obtained via copolymerization with relatively small
neous expression and secretion of several cellulase com- amounts of other functional momomers (18). Thus poly-
ponents and utilization of soluble cellulose recently lactides represent a potential substitute for polyethylene.
reported (87, 88). Utilization of insoluble cellulose is an Because lactic acid is at the same oxidation state as
important objective for the recombinant hydrolytic strat- carbohydrate and polymerization entails only a 20% yield
egy. loss due to dehydration, the stoichiometric yield from
B. Product Diversification. Most schemes envision- carbohydrate with an allowance for cell synthesis, 0.72,
ing a diversified biomass-based chemicals industry in- is much higher than that for the replacement strategy.
volve a small number of versatile, generally low molec- At this high yield, the feedstock cost contribution (cents
ular weight intermediates, most of which are derived via per kilogram of polymer) for all of the biological produc-
biological conversion (4, 89-95 ), a scheme which closely tion scenarios is less than half that for polyethylene
resembles the structure of the existing petroleum-based synthesis from petroleum, with the mature cellulosic
petrochemicals industry (96). Each of these “platform” scenario nearly 5-fold less than that of ethylene. The
intermediates gives rise to a “family” of derivative analysis presented in Table 6 does not consider the cost
chemicals, with synthesis of these derivatives involving of processing downstream of production of fermentable
either biological or, quite often, nonbiological conversion. carbohydrate; in particular, the cost of separation is not
In some cases, these chemicals directly replace a petro- included. Nor does this analysis consider processing
leum-based intermediate with identical composition. In energy inputs, which are important in the context of
other cases, envisioned biological platform chemicals have lifecycle analysis. This example does, however, illustrate
a distinct composition but substitute for their petrochemi- the potential advantages of a substitution-based strategy
cal counterparts by providing equivalent functional char- for biological production of organic chemicals. It also
acteristics. Although the replacement route has the underscores the point made earlier (Section V) that the
advantage of easy integration into the existing infra- cost of processing, rather than feedstocks, is the primary
structure, we think the substitution strategy will have impediment to cost-effective production of biocommodity
more impact in the long run. Indeed, a reasonable products.
argument could be made that it would be quite surprising In general, there is not yet widespread consensus as
if a given compositionally identical intermediate were the to which biologically derived platform chemicals comprise
most advantageous for chemical synthesis involving both viable substitutes for which current petroleum-derived
nonbiological processing of petroleum and biological chemicals. Developing such consensus is an important
processing of plant biomass. The replacement route challenge related to the substitution strategy. Wide-
Biotechnol. Prog., 1999, Vol. 15, No. 5 787
Table 6. Illustration of Replacement and Substitution Strategies for Synthesis of Commodity Plastics
price of yield feedstock cost
processing scenario intermediate (kg of polymer/ contribution
(feedstock/intermediate(s)/product)a (¢/kg)b kg of intermediate)c (¢/kg polymer)d
petroleum processing
oil/ethylene/pe 44 1 44
biological processing
replacement strategy
corn/fc/ethanol/ethylene/pe 14.1 0.28 50.4
cb/fc/ethanol/ethylene/pe (current) 13.5 0.28 48.2
cb/fc/ethanol/ethylene/pe (mature) 7.6 0.28 27.1
substitution strategy
corn/fc/lactic acid/pl 14.1 0.72 19.6
cb/fc/lactic acid/pl (current) 13.5 0.72 18.8
cb/fc/lactic acid/pl (mature) 7.6 0.72 10.6
a FC ) fermentable carbohydrate, PE ) polyethylene, CB ) cellulosic biomass, PL ) polylactic acid. b Ethylene based on 1998 bulk
pricing data (Chemical Marketing Reporter). For biological processing, based on transfer prices for fermentable carbohydrate (Table 5).
c For biological processing, yields reflect stoichiometric fermentation yields times a 10% allowance for cell synthesis times the stoichiometric
yield of secondary intermediate synthesis (where appropriate) times the stoichiometric polymerization yield. For FC f ethanol f ethylene-
PE: (0.51)(0.9)(0.61)(1) ) 0.28. For FC f lactic acid f PL: (1)(0.9)(0.8) ) 0.72. d Equals (price of intermediate/polymer yield).
spread implementation of the substitution strategy will in E. coli of 1,3-propanediol (102), production of novel
require R&D-driven advances associated with product polyhydroxyalkanoates in Alcaligenes eutrophus (103),
diversification. Cost-effective production of substitute and production of cis,cis-muconic acid, which can be
chemicals will also benefit from, and for some products catalytically converted into adipic acid, the precursor of
require, advances associated with overcoming the recal- nylon (94, 104). Fewer examples exist of applying meta-
citrance of cellulosic biomass. bolic engineering to the modification of cell properties
As pointed out by several substantial studies over the important in the context of biological processing, in part
years (4, 89-95), biomass can in principle provide because the genetic basis of such properties is often
substitutes for the vast majority of fuels, chemicals, and difficult to determine (see below). Similarly, little effort
materials presently derived from petroleum. Develop- has been devoted to detoxification associated with bio-
ment of cost-effective technology necessary to realizing commodity applications, although this could be useful,
this vision represents a technological challenge equally for example, in the case of inhibitors generated during
applicable to both starch-rich and cellulosic feedstocks. pretreatment.
Pursuant advances will be required in the areas of Synthesis of proteins and secondary metabolites of
metabolic engineering, chemical catalysis, and separa- interest for health care products is usually based on ATP-
tions. Our consideration of metabolic engineering focuses requiring anabolic metabolism. For most such products,
on features distinctive to biocommodity processing, and aerobic production systems are likely to be preferred due
only summary perspectives are offered on chemical to the much greater availability of ATP. By contrast, a
catalysis and separations. substantial quantity of the small molecules of interest
Metabolic Engineering. Metabolic engineering in- in the context of biocommodity engineering are available
cludes alteration of metabolic flux via cellular manipula- at potentially high yields from anaerobic catabolism. The
tion using recombinant DNA technology, development of absence of aeration and significantly lower-energy re-
quantitative tools and models to understand flux modi- quirements for cooling characteristic of anaerobic pro-
fication in complex biological systems, and development cesses can have substantial beneficial impacts in the
of laboratory techniques that allow fluxes and/or me- context of lifecycle analysis (Section VI) and process
tabolite concentrations to be determined (98-100). As economics. Important host organism characteristics for
cloning and characterization of genes has become routine production of biocommodity products are listed in Table
and sequence information for entire genomes is becoming 7.
increasingly available, modern science has access to an Bacteria and lower (generally unicellular) eukaryotes
unprecedented numbers of enzymes capable of catalyzing are typically the preferred host organisms for metabolic
a tremendous range of reactions. It is likely that meta- engineering directed toward biocommodity products.
bolic engineers have only started to glimpse, much less Such microorganisms are generally robust and fast-
realize, the potential to combine this array of catalysts growing relative to higher eukaryotes (e.g., animal, plant,
into functional pathways. or insect cells). The chief incentive to use higher eukary-
Cameron and Tong (99) have catalogued over 100 otes, post-translational modification of proteins to render
examples of metabolic engineering applied to extending them functional in humans, is unimportant for biocom-
the range of substrates utilized, improved production of modity products. At the same time, there is incentive to
chemicals already made by the host organism, production consider a broader range of microbial hosts for metabolic
of chemicals new to the host organism, modification of engineering of commodity products as compared to
cell properties, and detoxification of toxic chemicals. All pharmaceuticals. This incentive arises because the prop-
of these objectives are potentially relevant for biocom- erties in Table 6 are often determined by multiple genes
modity engineering. Examples of extended substrate and are thus difficult to access via genetic engineering.
range include developing S. cerevisiae and Z. mobilis In many cases it is likely to be more successful to
strains that utilize nonglucose sugars (Section V-A). engineer a desired pathway into an organism having
Examples among many of improved chemical production useful industrial properties, for example, competitive
include high ethanol yields in Escherichia coli and fitness, robustness in industrial environments, and per-
Klebsiella oxytoca (Section V-A) and enhanced solvent haps the ability to utilize cellulose, rather than trying
production in Clostridium acetobutylicum (101). Ex- to engineer such multigene properties into hosts organ-
amples of new chemical production include production isms that do not already have them. Figure 3 illustrates
788 Biotechnol. Prog., 1999, Vol. 15, No. 5
profitable to manufacture than bulk materials such as higher profit margin of chemical coproducts allows the
gasoline. However, fuel production provides economies bulk products to be more competitively priced.
of scale that reduce the cost of making smaller-volume Potential coproduction benefits extend beyond the
coproducts below levels that they could achieve on their synthesis of multiple organic products from fermentable
own. In essence, coproduction allows fuels and smaller- carbohydrate. In the case of woody biomass, for example,
volume products to be sold for lower prices than would about 25% of the dry weight and 40% of the energy in
be possible if either class of products was produced the feedstock is present as lignin, which is in general not
separately. Coproduction benefits are equally apparent viewed as a promising substrate for biological processing.
and essential in the case of corn wet-milling, where the In addition to relatively low-volume, but potentially high-
slate of products includes ethanol, high fructose corn value, markets for lignin-derived chemicals, power pro-
syrup, corn oil, corn gluten meal, corn gluten feed, an duction represents a very high demand use for lignin-
increasing array of chemicals, and food-related products rich residues remaining after biological processing. Process
such as vitamins and amino acids. designs anticipating advanced conversion technology for
We expect that similar synergies will support integra- woody feedstocks have found that coproduction of power
tion of processing technologies to produce a full slate of with biologically produced fuels or chemicals has the
products from cellulosic feedstocks. The optimally sized potential to be highly significant in terms of process
plant for processing cellulosic biomass corresponds to the economics, environmental benefits (in particular green-
point at which the incremental savings in processing house gas related), and energy supply. Herbaceous cel-
costs due to economies of scale are offset by the incre- lulosic feedstocks (e.g., switchgrass, alfalfa) offer the
mental increases in feedstock costs. For all but very low potential for production of protein-rich animal feeds if
fractions of land harvested for feedstock production, the harvested at a suitable point in their growth cycle. This
optimal plant size is quite large, comparable to the possibility extends the coproduction concept yet further
largest current corn wet mills. Only fuels and a small from the processing facility to the agricultural field.
number of organic chemicals have markets sufficiently Growing a crop expressly for coproduction of animal feed
large to consume the dedicated output of one such plant; and fermentable carbohydrate offers advantages not only
thus most chemicals will of necessity be coproducts of a in terms of maximizing product value but also in terms
mature biocommodity industry. The equipment required of maximizing land-use efficiency.
to produce fermentable carbohydrate from biomass is It is somewhat ironic that the relationship between
essentially identical whether such carbohydrate is con- production of fuels, chemicals, power, and feed (which
verted to fuels, chemicals, or materials. Thus, a chemical commands the largest share of U.S. agricultural land) is
plant that uses a side stream of fermentable carbohydrate often portrayed as competitive. In fact coproduction of
from a larger process facility producing fuel or some other most of these product classes in a single facility is
bulk product will be competitive at a lower price than a expected to be beneficial and, in a majority of cases, is
standalone plant, making the same volume of that likely to be truly necessary for cost-competitiveness in
chemical. The economics of producing bulk products (e.g., the long run. The fundamental reasons for this are the
fuels) benefit equally from such coproduction, since the economies of scale resulting from producing high-demand
790 Biotechnol. Prog., 1999, Vol. 15, No. 5
products such as fuels and power, the higher product/ lifecycle analysis and related tools within a mature
feedstock price margins available from producing chemi- understanding or resource and environmental systems.
cals, the value maximization resulting from using all As health care-motivated biotechnology has matured,
parts of the plant, and integration benefits (e.g., utiliza- increasing attention of educators, researchers, and prac-
tion of waste heat from power generation for process titioners is being afforded to the domain of health care-
energy requirements). related products: the body. Similarly, as biocommodity
Partially mature biocommodity processes with less engineering matures, we think it highly desirable that
than fully diversified product slates and less than fully increasing attention be payed to its domain: the environ-
developed infrastructures face significant challenges ment. Whereas efficacy validation for health care-related
penetrating markets occupied by established petroleum- products is determined by clinical trials, efficacy evalu-
based processes. One useful strategy to address this ation for biocommodity products is determined by a
dilemma is so-called “niche opportunities” consisting of lifecycle analysis.
low-cost feedstocks or established infrastructure elements E. Education. Academia has provided the tools of
(e.g., at a power plant, paper mill, or corn wet mill). A biotechnology, has played a major role in the development
second useful strategy is to emphasize applications for of the health care-motivated biotechnology industry, and
which biological products have performance advantages is well-suited to play an equally important role in relation
and thus command higher prices. The cost benefits of to biocommodity engineering. At the undergraduate level,
niche opportunities and applications with performance we see biocommodity engineering as an appropriate topic
advantages must often be balanced against the cost for an elective course or a thesis but in general not as a
disadvantages of small-scale production. Over time, it is major area of study in lieu of more traditional disciplines
reasonable to expect larger-scale production, increasing such as engineering, biology, and chemistry. Biocommod-
use of substitution as compared to replacement, increas- ity engineering is, in our view, a legitimate focus for
ing competition based on price rather than performance, graduate study at both individual and programmatic
and an increasingly diverse range of coproducts such that levels.
value is extracted from all feedstock fractions and, to the As elaborated in Table 8, we see three major compo-
extent possible, process effluents. nents of graduate study in biocommodity engineering:
C. Process Design and Economic Analysis. Design biotechnology, process engineering, and resource and
of processes incorporating multiple integrated unit op- environmental systems. In most cases, an individual
erations producing multiple products is an essential student would develop primary competence in one of
integrative activity associated with biocommodity engi- these areas, with secondary competence in the other two.
neering. Process design enables evaluation of the price- This mixture of disciplines and perspectives is much
competitiveness of plant-derived products as compared more likely to arise by design than coincidence and is
to established products, identification of beneficial com- unusual to find within a graduate degree program today.
binations of products and processes, and prioritization In particular, most existing biotechnology-focused pro-
of opportunities for R&D-driven cost reductions. The grams emphasize product-focused genetic engineering for
gradual augmentation and eventual replacement of production of medicinal products and single-product
products of integrated petroleum refineries byproducts relatively small-scale processes in which cost of produc-
of integrated biomass refineries is a highly complex tion is not a driving force (Section I) without considering
process that will benefit from economic analysis beyond resource or environmental aspects.
the individual processing facility. For the near term at least, we see biocommodity
D. Resource and Environmental Analysis. Para- engineering as a focus area spanning several depart-
phrasing John Prausnitz: “If engineering is the applica- ments, a model used increasingly by today’s health care-
tion of science for human benefit, then the engineer must oriented biotechnology programs and in other areas (e.g.,
be a student both of the application of science and of microelectrical mechanical systems) as well. Our cor-
human benefit as well” (105). For biocommodity engi- respondence with representatives of industry suggests
neering, the major dimension of human benefit evalua- that significant hiring demands exist in biocommodity-
tion beyond economic factors involves resources and the related industries, that this demand is likely to grow over
environment. For the reasons outlined in Section I, we time, and that it is currently somewhat difficult to find
believe that biocommodity engineering has legitimate degree recipients with the combination of skills most
potential to yield substantial resource and environmental useful in a biocommodity context.
benefits, a view which others have also supported (94,
106). At the same time, we think it important to Acknowledgment
acknowledge that biocommodity engineering is a poten- The authors thank Bruce Dale for stimulating discus-
tial double-edged sword in this context as a result of sions and Sean Casten and Ron Landucci for their
competing uses for land and biomass feedstocks and the contributions to Table 5.
possibility of poor land use practices, the large volume
of biocommodity products, and the energy and material References and Notes
inputs required for both feedstock production and pro- (1) DeLucchi, M. A. Emissions of greenhouse gases from the use
cessing. Guidance to realizing the positive potential of of transportation fuels and electricity; Argonne National
biocommodity engineering is available via application of Laboratory: Argonne, IL, 1991, Vol. I.
Biotechnol. Prog., 1999, Vol. 15, No. 5 791
(2) Lynd, L. R. Overview and evaluation of fuel ethanol produc- (26) Facts and figures for the chemical industry. C&E News,
tion from cellulosic biomass: technology, economics, the 1998, June 29.
environment, and policy. Annu. Rev. Energy Environ. 1996, (27) www.corn.org/web/stats/html.
21, 403-465. (28) North America Pulp and Paper Fact Book; Miller Free-
(3) Tyson, K. S.; Riley, C. J.; Humphreys, K. K. Fuel cycle man: San Francisco, 1993.
evaluations of biomass-ethanol and reformulated gasoline; (29) Annual Energy Outlook with Projections to 2015; Energy
NREL/TP-463-4950. DOE Office of Transportation Technolo- Information Administration, U.S. Department of Energy.
gies, Washington, D.C., 1993, Vol. I. Washington, D.C., 1998.
(4) Morris, D.; Ahmed, I. The Carbohydrate Economy; Institute (30) Perlack, R. D.; Wright, L. L. Technical and economic status
for Local Self-Reliance: Washington, D.C., 1992. of wood energy feedstock production. Energy 20 (4), 1995,
(5) Biotechnology for the 21st Century; Biotechnology Research 279-284.
Subcommittee, Committee on Fundamental Science, National (31) Turhollow, A. The economics of energy crop production.
Science and Technology Council, Office of Science and Tech- Biomass Bioenergy 1994, 6 (3), 229-241.
nology Policy, Executive Office of the President, Washington, (32) Domalski, E. S., Jobe, T. L., Milne, T. A., Eds. Thermody-
D.C., 1995. namic data for biomass materials and waste components;
(6) Dale, B. E., Artzen, C. E., Eds. Biobased Industrial Prod- ASME, 1987.
ucts: Research and Commercialization Priorities; National (33) Basic petroleum data book - petroleum industry statistics;
Research Council National Research Council, Washington, American Petroleum Institute, Washington, 1997; Vol. XVII
In press. (1).
(7) Chem industry begins restructuring in wake of biotech (34) 1992 Census of Manufacturers: Industry Series; Petroleum
revolution. Chem. Mark. Rep. 1998, 254 (20), 22-23. and Coal Products Industries, U.S. Department of Commerce,
(8) Thayer, A. Living and loving life sciences. C&E News 1998, U.S. Government Printing Office, Washington, 1992.
76 (47), 17-24. (35) Moffat, A. S. Plants as chemical factories. Science 1995,
(9) The evolution of the world’s energy systems; Shell Interna- 268, 659.
tional LTD, 1996. (36) Brown, K. S. Life on the molecular farm - transgenic
(10) Report to the President of the Interagency Steering Com- plants are extending the range of chemical production pos-
mittee on the Outcome of the Deliberations of the Policy sibilities in agriculture. Bioscience 1996, 46 (2), 80-83.
Dialogue Advisory Committee to Assist in the Development (37) Timberlake, W. E. Agricultural genomics comes of age. Nat.
of Measures to Significantly Reduce Greenhouse Gas Emis- Biotechnol. 1998, 16, 116-117.
sions from Personal Vehicles. The White House, February, (38) Wyman, C. E.; Bain, R. L.; Hinman, N. D.; Stevens, D. J.
1996. Ethanol and methanol from cellulosic biomass. In Renewable
(11) McCoy, M. Biomass ethanol inches forward. C&E News Energy; Johansson, T. B., Kelly, H., Reddy, A. K. N., Williams,
1998, December 7, 29-31. R. H., Eds.; Island Press: Washington, 1993, pp 865-924.
(12) Lugar, R.; Woolsey, R. J. The new petroleum. Foreign (39) Economic feasibility study of an acid hydrolysis based
Affairs 1999, 78 (1), 88-102. ethanol plant; Badger Engineers Inc., Solar Energy Research
(13) Wang, D. I. C. Role of Biochemical Engineering in the New Institute Report ZX-3-030-96-2. Golden, 1984.
Biotechnology; Biochemical Engineering X, Engineering Foun- (40) Knappert, H.: Grethlein, H.; Converse, A. Pretreatment
dation Meeting, Kananaskis, May 1997. of wood for enzymatic hydrolysis. Biotechnol. Bioeng. Symp.
1980, 11, 67-77.
(14) Basu, P. K. Pharmaceutical process development is differ-
(41) Grohmann, K.; Himmel, M.; Rivard, C.; Tucker, M.; Baker,
ent. Chem. Eng. Prog. 1998, September, 75-82.
J. Chemical-mechanical methods for the enhanced utilization
(15) Lynd, L. R.; Elander, R.; Wyman, C. E. Likely features of
of straw. Biotechnol. Bioeng. Symp. 1984, 14, 137-157.
cost of mature biomass ethanol technology. Appl. Biochem.
(42) McMillan, J. D. Pretreatment of lignocellulosic biomass.
Biotechnol. 1996, 57/58, 741-761.
In Enzymatic conversion of biomass for fuels production;
(16) Thayer, A. Great expectations - Biopharmaceutical in- Himmel, M. E., Baker, J. O., Overend, R. P., Eds.; ACS
dustry expects success from new drug therapies and a full Symposium Series 566, American Chemical Society, Wash-
product development pipeline. C&E News 1998, August 10, ington D.C., 1994, pp 292-324.
19-31.
(43) Hsu, T.-A. Pretreatment of biomass. In Handbook on
(17) Lynd, L. R.; Cushman, J. H.; Nichols, R. J.; Wyman, C. E. Bioethanol: Production and Utilization; Wyman, C. E., Ed.;
Fuel ethanol from cellulosic biomass. Science 1991, 251, Taylor & Francis: Washington, 1996, pp 179-195.
1318-1323. (44) Converse, A. O.: Kwarteng, I., K.; Grethlein, H. E.;
(18) Datta, R.; Tsia, S.-P.; Bonsignore, P.; Moon, S.-H.; Frank, Ooshima, H. Kinetics of thermochemical pretreatment of
J. R. Technological and economic potential of poly(lactic acid) lignocelllosic materials. Appl. Biochem. Biotechnol. 1989, 20/
and lactic acid derivatives. FEMS Microb. Rev. 1995, 16, 21, 63-78.
221-231. (45) Grohmann, K.; Torget, R.; Himmel, M. Optimization of
(19) Yergin, D. The Prize; Simon & Schuster, New York. 1991. dilute acid pretreatment of biomass. Biotechnol. Bioeng.
(20) Ranney, J. W.; Mann, L. K. Environmental considerations Symp. 1985, 15, 59-80.
in energy crop production. Biomass Bioenergy 1994, 6 (3), (46) Torget, R.; Walter, P.; Himmel., M.; Grohmann, K. Dilute
211-228. sulfuric acid pretreatment of corn residues and short rotation
(21) Watson, S. A.; Ramstad, P. Corn: Chemistry and Technol- woody crops. Appl. Biochem. Biotechnol. 1991, 28/29, 75-
ogy, American Association of Cereal Chemists, St Paul, MN, 86.
1987. (47) Torget, R.; Hatzis, C.; Hayward, T. K.; Hsu, T.-A.; Philip-
(22) Wiselogel A.; Tyson S.; Johnson, D. Biomass feedstock pidis G. P. Optimization of reverse-flow, two-temperature,
resources and composition. In Handbook on Bioethanol: dilute-acid pretreatment to enhance biomass conversion to
Production and Utilization; Wyman, C. E., Ed.; Taylor and ethanol. Appl. Biochem. Biotechnol. 1996, 57/58, 85-101.
Francis, Washington, 1996, pp 105-118. (48) Torget, R. W.; Kidam, K. L.; Hsu, T.-A.; Philippidis, G. P.;
(23) Salunkhe, D.; Chavan, J. K.; Adsule, R. N.; Kadam, S. S. Wyman, C. E. Prehydrolysis of lignocellulose; US Patent No.
World Oilseeds: Chemistry, Technology, and Utilization; Van 5705369, 1998.
Nostrand Reinhold: New York, 1992. (49) van Walsum, G. P.; Allen, S. G.; Spencer, M. J.; Laser, M.
(24) Jeffries, T. W. Utilization of xylose be bacteria, yeasts, and S.; Antal, M. J.; Lynd, L. R. Conversion of lignocellulosics
fungi. In Adv. Biochem. Eng./Biotechnol.; Fiechter, A., Ed.; pretreated with liquid hot water to ethanol. Appl. Biochem.
1983; Vol. 27, pp 1-32. Biotechnol. 1996, 57/58, 157-70.
(25) Dale, B. E.; Pham, T. K.; Esquivel, V. M.; Rios, I.; Latimer, (50) Heitz, M.; Capek-Menard, E.; Koeberle, P. G.; Gagne, J.;
V. M. Hydrolysis of lignocellulosics at low enzyme levels: Chornet, E.; Overend, R. P.; Taylor, J. D.; Yu, E. Fractionation
application of the AFEX process. Bioresour. Technol. 1996, of Populus tremuloides at the pilot plant scale: optimization
56, 111-116. of steam explosion pretreatment conditions using the STAKE
792 Biotechnol. Prog., 1999, Vol. 15, No. 5
II technology. Bioresour. Technol. 1991, 35, 23-32. (71) Ingram, L. O.; Gomez, P. F.; Lai, X.; Moniruzzaman, M.,
(51) Abatzoglou, N.; Chornet, E.; Belkacemi, K.; Overend, R. Wood, B. E.; Yomano, L. P.; York, S. W. Metabolic engineering
Phenomenological kinetics of complex systems: the develop- of bacteria for ethanol production. Biotechnol. Bioeng. 1998,
ment of a generalized severity parameter and its application 58 (2 & 3), 204-214.
to lignocellulosics fractionation. Chem. Eng. Sci. 1992, 47 (5), (72) Ho, N. W. Y.; Chen, Z.; Brainard, A. P. Genetically
1109-1122. engineered Saccharomyces yeast capable of effectively cofer-
(52) Ramos, L. P.; Breuil, C.; Saddler, J. N. Comparison of menting glucose and xylose. Appl. Environ. Microbiol. 1998,
steam pretreatment of eucalyptus, aspen, and spruce wood 64 (5), 1852-1859.
chips and their enzymatic hydrolysis. Appl. Biochem. Bio- (73) Zhang, M. C.; Eddy, C.; Deand, K.; Finkelstein, M.;
technol. 1992, 34/35, 37-48. Picatagio, S. Metabolic engineering of a pentose metabolism
(53) Bobleter, O.; Niesner, R.; Rohr, M. The hydrothermal pathway in ethanologenic Zymomonas mobilis. Science 1995,
degradation of cellulosic matter to sugars and their fermenta- 267, 240-243.
tive conversion to protein. J. Appl. Polymer Sci. 1976, 20, (74) Deanda, K.; Zhang, M.; Eddy, C.; Picataggio, S. Develop-
2083-2093. ment of an arabinose-fermenting Zymomonas mobilis strain
(54) Hormeyer, H. F.; Schwald, W.; Bonn, G.; Bobleter, O. by metabolic pathway engineering. Appl. Environ. Microbiol.
Hydrothermolysis of birch wood as pretreatment for enzy- 1996, 62 (12), 4465-4470.
matic saccharification. Holzforschung 1988, 42 (2), 95-98. (75) Esterbauer, H.; Steiner, W.; Labudova, I.; Herman, A.;
(55) Mok, W. S.-L.; Antal, M. J., Jr. Uncatalyzed solvolysis of Hayn, M. Production of Trichoderma cellulases in laboratory
whole biomass hemicellulose by hot compressed liquid water. and pilot scale. Bioresour. Technol. 1991, 36, 51-65.
Ind. Eng. Chem. Res. 1992, 31, 1157-1161. (76) Kadam, K. L. Cellulase production. In Handbook on
(56) Allen, S. G.; Kam, L. C.; Zemann, A. J.; Antal, M. J., Jr. Bioethanol: Production and Utilization; Wyman, C. E., Ed.;
Fractionation of sugar cane with hot, compressed, liquid Taylor & Francis: Washington, 1996, pp 213-252.
water. Ind. Eng. Chem. Res. 1996, 35, 2709-2715.
(77) Bayer, E. A.; Morag, E.; Lamed, R. The cellulosome - a
(57) Wang, L.; Dale, B. E.; Yurttas, L.; Goldwasser, I. Cost
treasure-trove for biotechnology. Trends Biotechnol. 1994, 12,
estimates and sensitivity analysis for the ammonia fiber
379-386.
explosion process. Appl. Biochem. Biotechnol. 1998, 70/72,
51-66. (78) Béguin, P.; Aubert, J.-P. The biological degradation of
(58) Thompson, D. R.; Grethlein, H. E. Design and evaluation cellulose. Microbiol. Rev. 1994, 13, 25-58.
of a plug flow reactor for acid hydrolysis of cellulose. Ind. Eng. (79) Tomme, P.; Warren, R. A. J.; Gilkes, N. R. Cellulose
Chem. Prod. Res. Devel. 1979, 18, 166-169. hydrolysis by bacteria and fungi. Adv. Microb. Physiol. 1995,
(59) Economic feasibility study of an enzyme-based Ethanol 37, 1-81.
Plant; Solar Energy Research Institute Report ZX-3-03098- (80) Warren, R. A. J. Microbial hydrolysis of polysaccharides.
1, Chem Systems Inc. , Golden, 1985. Annu. Rev. Microbiol. 1996, 50, 183-212.
(60) Nystrom, J. M.; Greenwald, C. G.; Hagler, R. W.; Stahr, J. (81) Hettenhaus, J. D. Glassner: Cellulase Assessment for
J. Technical and economic feasibility of enzyme hydrolysis for Biomass Hydrolysis; National Renewable Energy Laboratory,
ethanol production from wood; NYSERDA Report Number 85- 1997.
9, New York State Research and Development Authority: (82) Johnson, E. A.; Sakajoh, M.; Halliwell, G.; Madia, A.;
Albany, NY, 1985. Demain, A. L. Saccharification of complex cellulosic sub-
(61) Stone and Webster Engineering Corporation. Economic strates by the cellulase system of Clostridium thermocellum.
feasibility study of an enzyme-based ethanol plant; Report ZX- Appl. Environ. Microbiol. 1982, 43, 1125-1132.
3-03097-1, Solar Energy Research Institute, Golden, 1985. (83) Wood, T. M.; Wilson, C. A.; McRae, S. I.; Jovlin, K. N. A
(62) Hinman, N. D.; Schell, D. J.; Riley, C. J.; Bergeron, P. W.; highly active extracellular cellulase from the rumen anaerobic
Walter, P. J. Preliminary estimate of the cost of ethanol fungus Neocalimastix frontalis. FEMS Microbiol. Lett.1986,
production for SSF technology. Appl. Biochem. Biotechnol. 43, 37-40.
1992, 34/35, 639-649. (84) Lynd L. R.; Grethlein, H. G.; Wolkin, R. H. Fermentation
(63) Evaluation of a potential wood-to-ethanol process. In of cellulosic substrates in batch and continuous culture by
Assessment of Costs and Benefits of Flexible and Alternative Clostridium thermocellum. Appl. Environ. Microbiol. 1989,
Fuel Use in the U. S. Transportation Sector; Technical Report 55, 3131-3139.
Eleven, DOE/EP-0004, U.S. Department of Energy, Wash- (85) Weimer, P. J. Why don’t ruminal bacteria digest cellulose
ington, 1993. faster? J. Dairy Sci. 1996, 79, 1496-1502.
(64) Mackie, K. L.; Brownell, H. H.; West, K. L.; Saddler, J. N. (86) van Walsum, G. P.; Lynd, L. R. Allocation of ATP to
Effect of sulfur dioxide and sulfuric acid on steam explosion synthesis of cells and hydrolytic enzymes in cellulolytic
of aspen wood. J. Wood Chem. Technol. 1985, 5 (3), 405- fermentative microorganisms: bioenergetics, kinetics, and
425. bioprocessing. Biotechnol. Bioeng. 1998, 58, 316-320.
(65) Schwald, W.; Smaridge, T.; Chan, M.; Breuil, C.; Saddler, (87) Van Rensburg, P.; Van Zyl, W. H.; Pretorius, I. S..
J. N. The influence of sulfur dioxide impregnation and Engineering yeast for efficient cellulose degradation. Yeast
fractionation on product recovery and enzymatic hydrolysis 1998, 14, 67-76.
of steam-treated sprucewood. pP 231-242 In Enzyme Systems
for Lignocellulosic Degradation; Coughlan, M. P., Ed.; Else- (88) Peterson, S. H.; van Zyl, W. H.; Pretorius, I. S. Develop-
vier: London, 1989. ment of a polysaccharide-degrading strain of Saccharomyces
cerevisiae. Biotechnol. Lett. 1998, 12 (8), 615-619.
(66) Zheng, Y.; Lin, H.-M.; Tsao, G. T. Pretreatment for cellulose
hydrolysis by carbon dioxide explosion. Biotechnol. Prog. (89) Palsson, B. O.; Fathi-Afshar, S.; Rudd, D. F.; Lightfoot, E.
1998, 14, 890-896. N. Biomass as a source of chemical feedstocks: An economic
(67) Schell, D.; Torget, R.; Power, A.; Walter, P. J.; Grohmann, evaluation. Science 1981, 213, 513-517.
K., Hinman, N. D. A technical and economic analysis of acid- (90) Ng, T. K.; Busche, R. M.; McDonald, C. C.; Hardy, R. W.
catalyzed steam explosion and dilute acid pretreatment using F. Prod uction of chemical feedstocks. Science 1983, 219, 733-
wheat straw or aspen wood chips. Appl. Biochem. Biotechnol. 740.
1991, 28/29, 87-97. (91) Clements, L. D.; Beck, S. R.; Heintz, C. Chemicals from
(68) Wyman, C. E. Biomass ethanol: Technological progress, biomass feedstocks. Chem. Eng. Prog. 1983, November, 59-
opportunities, and commercial challenges; To appear in 62.
Annual Reviews of Energy and the Environment. (92) Bungay, H. Product opportunities for biomass refining.
(69) Wright, J. D. Ethanol from biomass by enzymatic hydroly- Enzymol. Microb. Technol. 1992, 14, 501-507.
sis. Chem. Eng. Prog. 1988, 84 (8), 62-74. (93) Alternative Feedstocks Program Technical and Economic
(70) Alterthum, F.; Ingram, L. O. Efficient ethanol production Assessment: Thermal/Chemical and Bioprocessing Compo-
from glucose, lactose, and xylose by recombinant Escherichia nents. Bozell, J. J., Landucci, R., Eds.; USDOE Office of
coli. Appl. Environ. Microbiol. 1989, 57, 2810-2815. Industrial Technologies. Washington, D.C., 1993.
Biotechnol. Prog., 1999, Vol. 15, No. 5 793
(94) Frost, J. Renewable feedstocks. In The chemical engineer; (102) Tong, I.-T.; Liao, J. J.; Cameron, D. C. 1,3-Propane diol
May 16, 1996, pp 32-35. production by Escherichia coli expressing genes from the
(95) Biomass refining: a major industrial opportunity; Iogen Klebsiella pneumoniae dha regulon. Appl. Environ. Microbiol.
Corp., Ottawa, 1996. 1991, 57 (12), 3541-3546.
(103) Fukui, T.; Shiomi, N.; Doi, Y. Expression and character-
(96) Szmant, H. H. Organic building blocks of the chemical
ization of (R)-specific enoyl coenzyme A hydratase involved
industry; John Wiley & Sons: New York, 1989.
in polyhydroxyalkanoate biosynthesis by Aeromonas caviae.
(97) Mobley, D. P., Ed. Microbial synthesis of polymers and J. Bacteriol. 1998, 180 (3), 667-673.
polymer precursors; Hanser/Gardner, Cincinnati, OH. (104) Draths, K. M.; Frost, J. W. Environmentally compatible
(98) Bailey, J. E. Toward a science of metabolic engineering. synthesis of adipic acid from D-glucose. J. Am. Chem. Soc.
Science 1991, 252, 1668-1674. 1994, 116, 399-400.
(99) Cameron, D. C.; Tong, I.-T. Cellular and metabolic engi- (105) Prausnitz, J. M. From Appolo to Prometheus and Her-
neering. Appl. Biochem. Biotechnol. 1993, 38, 105-140. cules: Goals and methods of chemical engineering. Chem. Ing.
(100) Stephanopolous, G. N.; Aristidou, A. A.; Nielsen, J. Technol. 1991, 63, 447-457.
Metabolic engineering: principles and methodologies; Aca- (106) Anastas, P.; Warner, J. C. Green chemistry: theory and
demic Press: New York, 1998. practice; Oxford University Press: Oxford, U.K., 1998.
(101) Mitchell, W. J. Physiology of carbohydrate to solvent Accepted August 11, 1999.
conversion by clostridia. Adv. Microb. Physiol. 1998, 39, 31-
130. BP990109E