Biorefining of biomass to liquid
fuels and organic chemicals
P. F. LEVY, J. E. SANDERSON, R. G. KISPERT and D. L. WISE
Dynatech R I D company, 99 Erie Street, Cambridge,
Massachusetts 02139, USA
Summary. The production o f liquid and gaseous fuels and
industrial chemicals from selected biomass by a process
known as b~orefining is reviewed. Four broad categories o f
biomass appear to be suitable feedstocks: woody biomass
and forest residues, agricultural residues, directly fermentable
crop-grown biomass, and municipal solid waste and sewage
sludge. Through the development o f suppressed methane
fermentation techniques, it is possible to produce valuable
organic chemicals such as acetic acid and ethyl acetate, and
liquid fuel (rather than fuel gas) by exercising various
processing alternatives. Thus the entire field o f methane
fermentation has been broadened. In the petroleum refining
industry, it is usually desirable to produce from crude oil
an optimal mixture o f industrial organic chemicals and
fuels, a concept known as coproduction. The biorefining
process reviewed appears to be adaptable to this same
concept o f coproduction using biomass as a feedstock.
Introduction
The basic processing concept combines a microbial
suppressed methane fermentation of fermentable organic
matter to produce a variety of organic acids, followed by
electrolysis of those acids to produce liquid fuels or
industrial chemicals. Biomass feedstock is used in a generic
sense to include a wide array of organic material. Feedstock
biomass includes woody biomass and forest residue, agri-
culture residues, directly fermentable crop-grown biomass,
and municipal solid waste and sewage sludge, all of which
either require some pretreatment or are directly ferment-
able. As a baseline process for discussion, corn slops, a
large-volume industrial waste, may be considered. In the
US, this waste is produced at a level of 30 000 ppm BOD
and up to 2 million gallons/day. In the process reviewed
in this paper, anaerobic fermentation is carried out under
conditions which suppress methane production to yield
higher molecular weight organic acids (organic acids up to
hexanoic or caproic). The higher organic acids may then be
extracted directly from the fermenter broth into kerosene,
or another suitable solvent. The lower molecular weight
organic acids are recirculated to the fermenter. The recovered
organic acids are transferred to an aqueous base and then
undergo electrolysis to produce liquid fuels or chemicals;
for example, caproic acid yields decane, a C10 straight-chain
liquid hydrocarbon suitable as a refinery feedstock. This
baseline process, however, has much broader potential.
Products from the fermentation of biomass include not
only liquid fuel for use as a feedstock but also diesel fuel
and industrial organic chemicals. By carrying out the
fermentation to produce even higher molecular weight
0141 --0229/81/030207--09 $02.00
© 1981 IPC BusinessPress
organic acids such as octanoic, it is possible to produce a
C 14 straight-chain hydrocarbon. Since C 10-C2ohydro-
carbons are, in fact, diesel fuels, the process may be used
to produce diesel fuel directly from biomass.
Industrial chemicals also may be produced by the direct
recovery of acetic acid from the fermenter broth, and
subsequent conversion to ethyl acetate. Alcohol fuels may
be produced electrolytically at high concentrations, thereby
reducing much of the cost and energy requirements for
distillation. Further, the fermentation employed uses all
the carbon in the organic feedstock, unlike more traditional
yeast fermentations.
As shown in Figure 1, the basic biorefining process is a
combination of fermentation and electrolysis. It has been
shown, however, that the processing concept is extremely
versatile, having the potential for using a broad range of
general biomass resources and producing a wide variety of
fuels and chemicals. Thus, the overall process of suppressed
methane fermentation and aqueous electrolysis is referred
to as 'biorefining.'
Process overview
The biorefining process for the conversion of biomass to
liquid fuels and industrial chemicals is comprised of three
steps. The biomass feedstock stream (see Figure I for the
array of feedstocks; feedstock pretreatment, if required,
is not a part of this review) containing dissolved and
suspended fermentable substances can be fed directly to
an anaerobic digester. In the initial processing step, the
organic material is converted by a non-sterile, mixed
culture, anaerobic fermentation to acetic and higher
molecular weight acids. The microorganisms are introduced
into the system through sewage sludge inoculum. Methane
formation is suppressed by the use of a specific methane
inhibitor.
The second processing step is the removal of the aliphatic
acids from the fermenter broth by liquid-liquid extraction.
For alkane fuel production, the higher organic acids may
be extracted directly from the fermenter broth into kero-
sene. The lower molecular weight acids are recirculated
back to the fermenter. If acetic or propionic acids are the
desired fermentation product, they can be removed by
extraction into kerosene containing 20% trioctylphosphine
oxide (TOPO). In any case, the acid product is then
reextracted in a second extractor into aqueous alkali. This
second extraction serves to concentrate the acids. If acetic
acid is the desired end-product of the process, it can be
separated from the kerosene-TOPO solvent by distillation
and then purified to produce glacial acetic acid.
Enzyme Microb. Technol., 1981, Vol. 3, July 207
 Reviews

FEEDSTOCK

J 'i WOODYBIOMASS(AND FORESTRESIDUE)p

GASIFICATION TO CO/CO2 B H2 /
/

II AGRICULTURAL RESIDUE Diesel I

fuels
A Primary agricultural residues (Ct4alkane)
(Straws, corn stayer) Extraction Electrolysis
into Kerosene
B Secondaryagricultural residues
(i) cellulosic (pulp &
paper + ~ J Suppressed ~ / . / / / /
for liquid
]
fuel production
/~ Refinery
feedstock
methane " / / / / / / / / / / / / , ~////, "////////////, additives
wastes, beet sugar pulp,etc.)
[ fermentationJ Kolbe (octane,
(ii) sugar/starch (corn stillage, reaction decane)
whey,etc.)
(iii) carbohydrates(tomato I Alcohol
processing wastes,etc.) (' gasohol ')
(methyl &
Extraction into ethyl alcohol,
C Lignaceousresidues (residue pentanol
from biomass-to-olcohol solvents for
facilities) organic chemical
production
Hater- Moest
reaction Industrial
chemicals
]
II[ DIRECTLY FERMENTABLE CROP
GROWN BIOMASS
I
(ethyl
acetate,etc)

A Aquaticplants, algae Industrial

B Short term rotational crops chemicals
( acetic acid,
propionic
acid,higher
IV MUNICIPALSOLID WASTE AND
I
Figure
SEWAGE SLUDGE
1 Biorefiningof biomass: processing alternatives
mol wl
fatty acids)

Production of other products is accomplished through
aqueous electrolysis of the organic acids. In this processing
step, the aqueous alkali from the second extractor which
contains the concentrated organic acids, is fed to an
electrolysis cell. Conditions of the electrolysis can be
adjusted to obtain a variety of products including alkanes,
alkenes, esters, and alcohols. The electrolysis product will
be dependent on the organic acid mixture fed to the
electrolysis cell as well as on the operating conditions of
the cell. Separation of the final product from the aqueous
stream may then be necessary. The alkane product can be
recovered simply after phase separation from the aqueous
layer. Other products require distillation for separation
from the aqueous phase.
The suppressed methane fermentation, digester design,
liquid-liquid extraction, and electrolytic oxidation of
organic acids are discussed below.
Suppressed methane fermentation
Anaerobic fermentation refers to the ability of certain
classes of microorganisms to grow on a number of different
organic compounds in the absence of oxygen. In traditional
anaerobic digestion, the organic compounds are ultimately
converted to the gases carbon dioxide and methane.
Historically, digestion has been used for the treatment of
various liquid wastes or suspensions, such as sewage, packing
house wastes, sulphite waste liquor, and various industrial
wastes. Its function has been primarily to accomplish
changes in the character of noxious wastes and a reduction
of their quantity, rendering them largely inoffensive;
generation of fuel gas or other products has been of
secondary importance.
The microbiology and biochemistry of the methane/
carbon dioxide producing digestion process has been an
important subject of research. The numerous findings are
summarized below.
(a) The process of anaerobic digestion of most organic
substrates proceeds biochemically in two discrete stages.
These stages are mediated by two different major classes of
microorganisms. The first step in digestion is hydrolysis
of the substrate, liquid or solid, and formation of organic
acids from the products of the hydrolysis step. This is
known as the acid formation step, and is accomplished by
the class of microorganisms known as acid formers. The
second step is the biological conversion of these acids,
including acetic, propionic, butyric etc., to methane, which
is accomplished by the class of organisms known as
methanogens.
(b) The acid formers grow and convert the substrate to
acid quite rapidly when sufficient substrate is available, and
will do so over a wide range of pH values, from about 4.5
to 7.5. The methane formers grow and convert these acids
to methane and will do so only over a narrower range of
pH value, from 6.0 to 8.0.
(c) Temperature must be maintained at the correct value
for optimum digester functioning.
(d) Mass transfer must be provided by stirring for rapid
functioning of the system. This is partly to provide a means
for transfer of the acid formed by the acid formers to the
methane formers, and also to provide heat transfer.
Microbial nutrients must be added when organic sub-
strates are deficient.
A consequence of these findings is that standard digester
designs and operational procedures have evolved which are

208 Enzyme Microb. Technol., 1981, Vol. 3, July

 Biomass refining to fuels and chemicals: P. F. Levy et aL

considered optimal for systems treating the most common 7.o

soluble or suspended wastes. In general, digesters for this
purpose are stirred, either by impellers or by bubbling gas
through them. Provision is also made for heat transfer, 6.0
for temperature control, and sometimes for pH control
and nutrient addition.
5.0
Inhibition o f methane formation
6.3g
Methanogenesis from acetate has been demonstrated 1'2
However, the production of methane is deleterious to
fermentations in which the desired products are higher 400
acids, for which the primary precursor is acetate. In fermen-
tations using model aquatic biomass substrates such as the u
fresh water angiosperm Hydrilla, methane formation was 300
observed with a concomitant decrease in acid levels in the g - ~
fermenter. 3 This brought about the need for a methanogenic
inhibitor which does not interfere with the normal meta-
E
~ 200
/ ' _8 • •

bolism of non-methanogenic microorganisms. o

Inhibitors of methane formation can be classified into I00 ~ oo
two groups. Chemicals that can act as electron acceptors,
i.e. sulphates and propanediot, comprise the first
group.4 The second consists of specific inhibitors of
particular methanogenic steps. 4-8 Work at Dynatech I J I J J
has shown 2-bromoethane sulphonic acid (BES) to be a 0 10 20 30 40 50 60
Fermentation time (days)
suitable suppressor of methane formation in Hydrilla
fermentations .3 Figure 2 Effect of addition of 2-bromoethane sulphonic acid
(coenzyme M analogue) on the conversion of Hydrilla (3%, w/v)
BES is a potent inhibitor of growth and uptake of the into acids. Sewage sludge was used as inoculum. •, Bromoethane
coenzyme-M, 2-mercapto-ethane sulphonic acid (HS-CoM). sulphonic acid (5 x 10-4M); o, control
Coenzyme-M is found predominately as C H 3 - S - C o M and
(S-CoM)2 forms in the cells. Presumably, the former is inhibit methanogenesis in fermentation of a wide variety
the final precursor to methane formation in the cell. In the of substrates.
presence of the inhibitor, the final step is blocked, and
methane biosynthesis is suppressed. As is indicated in Table Formation o f higher molecular weight acids
1, 1 ~M of BrCH2CH2SO3H is enough to exert 100%
inhibition of HS-CoM uptake. 8 A significant result from the ongoing work at Dynatech
is that higher aliphatic acids are produced in these fermen-
The effect of 5 x 10-4M BES on the anaerobic fermen-
tations at increased total acid concentrations. To speculate
tation of Hydrilla is shown in Figure 2. Conversion to the
on the reason for this action of the organisms, once
acid product is greater than in the control sample, and
sufficient acetic acid is produced to drop the pH of the
there is no significant methane formation even after
environment to levels which are inhibitory to their
fermenting for 60 days.
metabolism, further production of acetic acid is counter-
BES has also been employed as an effective suppressor
productive to the well-being of the flora. However,
of methane formation in fermentations with corn meal
conversion of acetic acid to higher aliphatic acids not
(see below). It is anticipated that BES can be used to
only provides energy for ATP formation but also increases
the pH. Higher aliphatic acids may be formed from acetic
Table 1 Effect of metabolic inhibitors a and uncoupling agents on
HS-CoM uptake"
acid by the following set of disproportionation reactions:

Inhibition
7CH3COOH ~ 4CH3CH2COOH + 2C02 + 2H 2 0
Inhibitor Final concn (%) 5CH3COOH ~ 2CH3CH2CH2COOH + 2C02 + 2H20
13CH3COOH -+ 4CH3CH2CH2CH2COOH + 6CO 2
Sodium azide 10 mM 95 + 6H20
Potassium cyanide 10 mM 0
4CH3COOH -+ CH3CH2CH2CH2CH2COOH + 2C02
2,4-Dinitrophenol 1 mM 69
Iodoacetate 100 #M 88 + 2H20
BrCH2CH=SO 3- 1 #M 100
It is also possible that glucose is converted to only two
Br(CH 4) 3SO3- 10/.tM 0
DCCDb 1 #M 0 moles of acetic acid and that the hydrogen produced is
CCCPb 10 #M 0 used to form the higher acids, perhaps in the following
Valinomycin 10 #M 20 manner:
a Inhibitor was added to cells 3 min prior to addition of 0.5#M 3CH3COOH + 2H2 ~ 2CH3CH2COOH + 2H20
H3SS-CoM. Uptake rate was determined from 0.5, 1.0 and 1.5 min 2CH3COOH + 2H 2 ~ CH3CH2CH2COOH + 2H20
time points. Rate is expressed relative to a control which contained
no inhibitor
5CH3COOH + 6H 2 ~ 2CH3CH2CH2CH2COOH + 6H20
b DCCD (N,N'-dicyclohexylcarbodiimide) and CCCP (carbonyl 3CH3COOH + 4H 2 -+ CH3CH2CH2CHzCOOH + 4H20
cyanide-m-chlorophenyl hydrazone) dissolved in methanol. A 10 #1
portion was added to 1 ml of cell suspension to give the final concen-
In any case, if anaerobic fermentation as described is
tration indicated. No inhibition was observed by addition of methanol carried out at high acid concentrations, a mixture of acetic
alone acid and higher organic acids is obtained. The distribution

Enzyme Microb. Technol., 1981, Vol. 3, July 209

 Reviews
of the acid products appears to be dependent on the
fermenter conditions. These conditions (and thus the product
spectrum) can be controlled by varying operating parameters
such as pH and temperature and by selective removal of
particular acid products.
F e r m e n t a t i o n rates and conversion efficiencies
Experiments at Dynatech have shown that the rate of
conversion of model biomass is very rapid. Preliminary
experiments have been performed using corn meal,
composed primarily of starch, as a substrate. A first-order
rate constant of 0.257 day -1 and initial rate of 115 meq
1-1 day - l have been calculated. These data are presented
in a semilogarithmic plot in Figure 3. If the fermenter is
modelled as a CSTR, a mean residence time of 15 days will
be required to ferment 75% of the corn meal (starch)
substrate to acids. Residence times of 15 to 30 days are
used in the economic analyses (given later), depending on
the acid product produced.
Evaluation of fermentable substrates involves exami-
nation of rate as well as conversion efficiency (percent of
substrate converted to product). Conversion efficiencies are
based on total available reducing equivalents. The total
available reducing equivalents, R, is obtained from the
equation:
R = A + 1.75P + 2.5B + 3.25 V + 4.0C
where A, P, B, V and C are the measured concentrations
of acetic, propionic, butyric, valeric and caproic acids,
respectively. The basis for this equation is the dispropor-
tionation reactions showing production of higher acids
presented above. The percent conversion, C, is then
6OO
500I
400 - loop, without the necessity for treating the digesting
i
g
300
3 high.
~8
2O0
100 I I I I 1 [ I I
0 20 40 60 80
Time (h)
Figure 3 First-order rate plot of the conversion of cornmeal to
organic acids. 3% cornmeal fermentation run no. 0 6 2 8 7 - B ;
C ~ = 4 5 0 meq/I; K = 0.011 h -1 = 0.257 day - l ; dc/dt = 115 meq/I
day
210 Enzyme Microb. Technol., 1 9 8 1 , V o l . 3, J u l y
calculated by multiplying R (in meq 1-1) by 0.06 g meq -1
and then dividing by the substrate concentration, S, in
g1-1 on a total solids basis:
C = O.06R/S
Fermentations were run in static reactors or in a 300 litre
fixed packed-bed reactor with continuous acid extraction.
The inoculum was 2% sewage sludge or effluent from
another fermentation. Fermentations in static reactors
employed calcium carbonate as a buffer to maintain a pH
at which the microorganisms remained viable.
The volatile solids content represents the maximum
achievable conversion for a given substrate. Typical
volatile solids conversions achieved to date for the various
biomass fermented are: Chondrus crispus, 90%; Gracilaria,
80%, Hydrilla, 50%; corn meal, 90%.
The limiting factors in the microbiological conversion of
volatile solids are the inaccessibility of portions of the
substrate to the microorganisms and the presence of part
of the volatile material as lignin. (Lignin is not a substrate
for the bacteria unless pretreatment of the liquor reduces
it to lower molecular weight fermentable fragments.)
Digester design
Engineering at Dynatech 9 has shown a significant advantage
of the anaerobic digestion of 'solid' biomass (municipal
solid waste, aquatic biomass, straws) in a reactor system
termed the fixed packed-bed fermenter. In the fixed
packed-bed system, solid substrate can be continuously fed
into the reactor and residue removed allowing the unit to
function in a plug flow manner. A moving stream of liquid,
circulating through the substrate, can perform the function
of mixing ordinarily accomplished by stirring in conven-
tional digesters. With proper arrangement of piping inlets
and outlets, this same liquid stream may be used to control
the temperature, add nutrients, control pH, and the like.
It has also been determined that aqueous toxic components
may be removed from the liquid phase by suitable means,
such as filtering through carbon in an external circulation
material directly. Microorganisms in the digester grow on
the solids remaining in the reactor when the liquid stream
is circulated through the external piping. This technique
allows the bacterial population in the digester to remain
Figure 4a shows one possible arrangement for the fixed
packed-bed fermenter. In this case, it is used as a batch
reactor for organic acid production. The fermenter, A, is
filled with a comminuted solid substrate. Aqueous nutrients
and microorganisms are added in a quantity sufficient to
saturate the substrate and to fill the digester head space and
circulation loop as shown. The pump, B, is used to circulate
the liquid through the loop and through the packed mass of
comminuted solid substrate which contains fermentation
microorganisms. As fermentation proceeds, required
buffers and additional nutrients may be added through the
inlet port, C; the acid is removed in subsystem D, which
may be an ion exchange column, liquid-liquid extractor,
or membrane device. Toxic compounds may also be removed
I from the aqueous phase by passing it through a cleaner, E,
IOO
which may contain, for example, activated charcoal. When
fermentation is economically completed, the undigested
material is partially dewatered by opening valve F and
pumping liquid out by means of pump G. The partially
dewatered material is then removed, fresh comminuted

Buffers, nutrients
/ culationpump(B)
Acid removal(D)
a FIu~:dflow ~ ( ~ Effluent pump(G)
(F)
Buffers, nutrients
Spent solids (
removal
Segment of fermenter
showing plug flow
operation
Organic solid
substrate feed
.I l l~ Toxicmaterial
removal
I A number of extraction methods may be used for removing
b Fluid flow
Figure 4 (a) Schematic diagram of fixed packed bed fermenter.
(b) Continuous flow packed bed fermenter
substrate added, and the cycle above repeated. Clearly, this
batch system is the most simple and serves largely to
describe the concept.
Application of the continuous flow system is presented
in Figure 4b. Here, solid substrate is fed into the bottom of
the fermenter and removed from the top of the fermenter.
As with the batch-type packed bed system, fermenter
liquid is circulated through the bed, and organic acids are
recovered in an external subsystem.
For pyrolysis gases, industrial waste and biomass pre-
treated to form a liquor, the packed bed digester is not
suitable. For example, typical industrial waste feedstreams
are relatively dilute (up to 5 - 7 % solids), with most of the
organic material dissolved. This necessitates the design of a
reactor which can maintain high microbial populations in
spite of rapid fluid flow through the fermenter. Digester
designs appropriate for application to these industrial
waste systems have been described by Ashare and Wilson. 1°
Also, gaseous fermentation of CO, CO s, and H 2 are
described elsewhere II as is the suppressed methane fermen-
tation of these gaseous feedstocks. 12
The packed bed fermenter can be modified for use with
feedstreams containing mainly dissolved solids. In such a
system, the digester must be packed with a suitable material
to provide a solid support for attachment of micro-
organisms, while still allowing liquid flow through the bed.
Similar systems have been described by McCarty 13 and
Young and McCarty. 14 Investigations using the 'anaerobic
Biomass refining to fuels and chemicals: P. F. Levy et aL
filter' have been undertaken for treatment of carbohydrates
in food processing wastes is as well as other industrial
waste streams. 16' 17 If accumulation of organic sludge clogs
the packed-bed reactor, a fluidized, or expanded, bed
reactor may be desired. Jewell and coworkers TM reported
the successful use of the fluidized bed design in the treat-
ment of a 2% solids stream of organic residue, reducing the
retention time from that of a CSTR by 80%.
An alternative to the packed or fluidized bed designs is
an anaerobic contact process known as the Bioenergy
Process, recently developed by Biomechanics Ltd, UK,
under the direction of G. M. Rippon, (personal communi-
cation) for the treatment of wastes on a commercial scale.
The process involves a combined heating and mixing
system, a digester, and a 'Bioenergy Separator.' The
separation device is a cooler and gravity separator which
provides solids separation by subjecting the treated liquid
to a thermal shock and allowing solids to settle. The cooled
bacteria tend to flocculate into relatively dense clumps
and settle to the bottom of the gravity separator. Clear
liquid can then be removed from the top and the sedimented
bacteria removed from the bottom and returned to the
fermenter. The process is well suited to the processing of
waste streams containing dissolved organic material.
The choice of a digester design will depend on the type
of biomass feedstock stream and desired end-product of
the treatment process. Ultimately, operational and economic
considerations (determined by experiment and analysis)
will dictate the optimum digester for the process.
Extraction o f organic acids
organic acids from fermentation broths as they are formed.
Of these, ion exchange with conventional elution, solvent
extraction with recovery by reextraction, and diffusion
membrane processes have been investigated. 3
The use of a conventional strong anion exchange resin
with quarternary ammonium sites was considered as an acid
removal technique. The hydroxylated resin may be inserted
in the recirculation loop of the fixed packed-bed reactor as
shown in Figure 4a. The hydroxyl ions are displaced by the
organic acid anions. When the resin is spent (as indicated by
no pH change across the resin), it is removed and replaced
by fresh resin. The spent resin is then eluted with an
equivalent amount of mineral acid to remove the organic
acids collected. The resin is then regenerated by eluting it
with alkali.
One equivalent of mineral acid and more than one
equivalent of alkali are required to generate each equivalent
of organic acid. More than one equivalent of alkali is
required because the equilibrium constant for the reaction:
NR~ X- + O H - ~ NR~ O H - + X -
is < 1. Weak anion exchange resins with equilibrium constants
>1 are available, but the utility of these has not yet been
investigated for the process under review.
Solvent extraction of organic acids from an aqueous
solution is a technique which has been under consideration
for some years. In fact, Goering is credited with the suggestion,
in 1833,19of ethyl acetate as an extracting solvent for
acetic acid. Ethyl acetate, along with diethyl ether, con-
tinued to be the most efficient extraction medium at low
acid concentrations until recently, when novel solvent
extraction media were developed. This new technology is
Enzyme Microb. Technol., 1981, Vol. 3, July 211
Reviews

based on the use of trioctylphosphine oxide in combination
with other solvents, the details of which are proprietary to
Hydrosciences Inc. 2° Higher molecular weight organic acids
(butyric or larger) can be efficiently extracted into a liquid
hydrocarbon solvent such as kerosene.
An alternative approach to conventional solvent extrac-
tion followed by distillation is solvent extraction followed
by reextraction into an aqueous base. The organic acids are
accumulated in the aqueous base as salts, and when their
concentration reaches 25%, they are neutralized with
mineral acid and distilled. This approach requires only one
equivalent each of acid and base to concentrate the organic
acid product from 1 to 3% in the fermenter to 20 or 25%.
Although some difficulty (such as emulsion formation)
has been experienced in the use of this technique in the
laboratory, it appears that this approach is the most suitable
for use on a large scale. The problems are not fundamental
in nature, but rather are those which are generally amenable
to engineering solutions and, in fact, may have already been
solved by Hydrosciences Inc.
When carboxylic acid mixtures are electrolysed, cross
coupling occurs in addition to the usual Kolbe reaction: 21
R1CO2 _ + R 2 C 0 2 _ -2c0~ R1 R 2 + R I _ R 1 +RZ R 2
-2e-
Higher fatty acids give satisfactory yields of the Kolbe
dimer, and alkanes up t o C 3 4 H 7 0 have been prepared by
this method.
In aqueous alkaline solution, in the presence of foreign
anions (such as perchlorates, sulphates and bicarbonates), at
low current densities, or with carboxylates substituted at
the a position with electron-donating substituents, the
major product was alcohol:
2RCO2H + H20 -~ CO2 + ROH + H 2
Similar reaction conditions will promote ester formation:
2RCO2H -+ RCO2R + C02 + Hz
These anodic oxidations to esters or alcohols are termed
Hofer-Moest reactions a3 and occur on carbon anodes.

Electrolytic oxidation of organic acids Reaction mechanism

The potential for the use of electricity as an oxidizing or Electrolysis of monobasic aliphatic organic acids generates
reducing agent was realized early in the nineteenth century paraffins, olefins, esters, alcohols and carbon dioxide at
by Rheinold and Derman, who electrolysed dilute aqueous the anode and hydrogen gas at the cathode. The mechanism
solutions of alcohol. Ludersdorff, in 1830, was the first to of the anode reaction is not well understood, but a likely
undertake a detailed study of the products obtained in the scheme is presented in Figure 5. Carboxylate is adsorbed to
oxidation of alcohol using different electrodes. Faraday the anode surface and oxidized to form an acid radical. The
first realized the potential for using electric current to acid radical decomposes to form CO2 and an alkyl radical,
synthesize or decompose a compound. As early as 1834, he which may then react with another radical to form the
observed that electrolysis of an acetate solution yielded Kolbe dimer, or disproportionate to form alkene and alkane
some hydrocarbon. Kolbe's important discoveries in the with half the carbon atoms of the Kolbe dimer. The alkyl
mid-nineteenth century led to the establishment of the radical also can be oxidized further to a carbonium ion,
principles of electrolytic oxidation and reduction. 2~ which subsequently forms alcohol, alkene or ester. As the
Kolbe completely replaced with hydrogen the chlorine anode reaction occurs, hydrogen ions are reduced at the
in chloromethylsulphonic acid by electrolytic means in cathode, forming hydrogen gas.
1845. He observed that trichloromethylsulphonic acid was 2H + + 2e- ~ H 2
completely decomposed in an aqueous solution at a
platinum electrode, stating: 'This acid, which is stable to
Alkene (R)
the strongest oxidizing media, breaks down with little ,&
difficulty at a platinum anode.' Kolbe then studied the
anodic oxidation of organic compounds. Electrolysis of an
alkali-metal acetate was shown to evolve ethane and two Alkene ~
R (-H~)- \q
volumes of carbon dioxide: Dimer
(R-R)
2CH3COO- -~ 2CO2 + C2H 6 + 2e- ,x
Further investigation of this reaction by Kolbe and
others demonstrated the feasibility of hydrocarbon synthesis
by electrolysis of monobasic aliphatic acid salts. The Kolbe
reaction proceeds by the oxidative removal of CO2 from e-
COz [ !
.14 _ R~R/odsl I
the acid anion, with subsequent dimerization of the remain- RC02 ~ RCO2(ods)- _/4= RCOz(°ds)
ing alkyl radical. For example, valeric acid (pentanoic acid)
forms octane according to the following equation. 22
2CH3(CH2)aCOOH ~ C8H18 + 2C0 2 + H 2
-I
)o
In an aqueous medium, hydrocarbon formation occurs cohO,Ro.
A --+°'- I1R a+ °ds l
to an appreciable extent only at a smooth platinum or
iridium anode. Using a platinum electrode, the Kolbe
dimerization is favoured by high concentrations of +RCO~
carboxylate, current densities >0.25 amps/cm 2, low
temperature (below 50°C), low pH (~<6), and the absence Alkene
of foreign anions. 21 The most important competing reaction R (-H 2) Ester
is the formation of olefin possessing half the carbon atoms RCO2R
of the Kolbe electrolysis product. Figure 5 Reaction mechanism for anodic oxidation of carboxylates

212 Enzyme Microb. Technol., 1981, Vol. 3, July

 Biomass refining to fuels and chemicals: P. F. Levy et al.

Oxygen is not produced when aqueous carboxylic acid II Feed

biomass
solutions are electrolysed. This is noteworthy because the I.o - ~*FI Ash Fermentetion Electrolysis
potential at which water electrolysis occurs (1.5 V) is less 0.9
than that at which Kolbe electrolysis normally occurs. An o8 ~,~,~Lignin
layer of adsorbed carboxylate may cover the electrode
surface 2°-22 inhibiting water electrolysis by preventing ~
0.7
0.6
-
~ d-ce
I IIIu ose

the adsorption of water, g 05 - / / Residue

"" Organic
The H2/CO 2 mixture produced in the electrolysis can be o. 0.4 / / Ash acids
used to produce electricity in a molten carbonate fuel cell 0.3 _ // // • ~-Lignin COz [~
Hydrocarbon
operating at 60% efficiency and a temperature of 650°C. A o.2 - //"~ I'--'-~Cellulose CO 2 p~uct
-- I I Liquid I I I
H2/CO 2 mixture is ideal for this application because the o.I - "" I I HzO waste I I I H2
'' I I ~ ~ I I
fuel cell will require carbon oxides as well as hydrogen, o
44°1., I°lo 5°1o 21°Io I% 12% 17%
The electricity produced will meet about one-third of the
requirements of the electrolysis step. The fuel cell will also Figure 7 Mass balance of liquid fuels f r o m biomass

produce high quality heat. The heat can be used to make

sufficient quantities of high pressure steam to meet all of Feed
Fermentation Electrolysis
0.4 biomass I
the pumping requirements of the plant. Some heat will also
be available for distillation of electrolysis products.
Electrolytic oxidation of organic acids is the topic of
several current projects at Dynatech. Present efforts are g 03 ~///~,f Cellulose
.//.
being devoted to reducing the applied potential of the F/l.
///.
electrolytic reaction. Experimental parameters which are @ Organic

being optimized to reduce the applied potential are" (1) 0.2 //z
i/I acids
Hydrocarbon
producl
temperature, (2) pH, (3) concentration of electroactive o_ /11,
/ / / Residue
species, (4) use of supporting electrolyte, (5) current
density, (6) cell geometry, (7) anode material and (8) 0.1 "//. ~
p-Cellulose
Lignin
C02
cathode material. The effect of these parameters on applied
I Liquid
potential and product distribution is being investigated. /Xd
Figure 6 reproduces a gas chromatograph of electrolysis 0
34°1o :5% 16°/o I1% 39%
products from a Kolbe cell-fed fermenter broth containing
Figure 8 Carbon balance o f l i q u i d fuels f r o m biomass
butyric, valeric and caproic acids. Work on alcohol produc-
tion from organic acids is also in progress. Electrolytic
oxidation of acetate and propionate to methane and ethanol Feed
biomass(in) Fermentation Electrolysis
is being investigated. The operating conditions of this 6500
-_ 77;// / / ~ Lignin
6000
reaction are being optimized to produce maximum alcohol ///
///
concentration at a minimum applied potential. -- // // // ICellulose
5000
///
/// Organic
/// acids
4000 -- /// /// Hyrocorbon
C o m p a r i s o n o f liquid fuels process to /// ///
/// /// product
traditional a l c o h o l fuels /// ///
3000 ///
The biorefining process offers a variety of advantages over /// Residue // / // /
/// -~Lignin
traditional fermentations. In this section the biorefining 2000 -- /// // / // /
/// ///
/// ---~Cellulose
/// // // //
///
1000 - -

/// [ [ Liquid ///

[ [ waste Heat ///

1
///
/.4< I I[---I, , Y~
362% 4.6°/° 23% Z7% 50.8`,,/0

Figure 9 Energy balance (1 Ib biomass basis) o f l i q u i d fuels f r o m

biomass

process for the production of liquid alkane fuels (hexane

through decane) is examined and compared to ethanol
fermentation.
It is first of interest to consider the yield of mass,
energy, and carbon from the proposed process. Figures
7-9 present histograms indicating the fate of one pound of
typical biomass containing 75% cellulose and equivalent
fermentables, 10% lignin and other organic non-ferment-

\ ables, and 15% ash. Percentages of the final components

Figure 6 Gas c h r o m a t o g r a p h i c analysis o f K o l b e electrolysis
products f r o m f e r m e n t e d C h o n d r u s crispus. A, pentane; B, heptane;
C, o c t a n e ; D, nonane; E, decane
are indicated along the abscissa. Such a hypothetical
biomass would contain 6500 BTU (Figure 9). Of this
amount, 2850 BTU would form a solid residue and
150 BTU would be released as heat. The remainder of the
energy would be contained in the mixed organic acid
fermentation products. These acids are extracted into a
non-aqueous phase and then reextracted into an aqueous

Enzyme Microb. Technol., 1981, Vol. 3, July 213

Rev~vvs

base for use in the Kolbe electrolysis step, The fermenter
effluent is recycled, except for a small amount which is
discarded to allow for the water added to the fermenter
with the biomass. This water would contain 0.03 pound
of organic acid with 300 BTU of energy.
In the electrolysis the organic acids are converted to
alkanes (0.17 pound) along with carbon dioxide (0.12
pound) and hydrogen (0.01 pound). The alkane product
contains 50% of the energy of the biomass. Energy not
incorporated into the hydrocarbon product may be further
used. Hydrogen produced can be used to run a molten fuel
cell producing one-third of the electrical requirement of
the Kolbe electrolysis. Low-grade heat produced in the
fermentation to acids can be used to maintain the fermenter
at the desired temperature. A portion of the energy con-
tained in the residue may also be recoverable.
Overall energy balance
There has been a continuing debate concerning the
energy balance for the production of gasohol. A compre-
hensive analysis of the energy requirement and by-product
credits for gasohol production from corn has been
presented by Chambers and coworkers.24 This analysis is
summarized in Figure 10 for the production of fuel-grade
ethanol from corn by traditional fermentation, distillation
and purification processes. Values used are based on
production and processing of one bushel (56 pounds) of
corn. The corn grain is assumed to be 70% converted to
products. Half the energy remaining in the residue is taken
as a feed by-product credit. The final product (14.5 pounds)
represents 2.5 gallons of ethanol. The energy requirement
exceeds the output, resulting in an unfavourable energy
balance (output/input) of 0.72.

184 000 BTU 370 000 BTU

Agricultural input Process energy ~IF

I
ETHANOL
Fermentation
[145 Ib)
distillation and 185 600 BTU
I Bushel corn
purification
to
100% ethanol

Output I IFeed by-product

160 0 0 0 BTU 55 0 0 0 BTU

(Stalks and cobs)

Figure 10 Estimates of energy requirements and credits are from Chambers 24 and include all processes from planting corn to production of
anhydrous ethanol. Net energy: --153 400 BTU/bushel; energy balance: out/in = 0.72

184000 BTU

1
Agricultural input
55 000 BTU
Feed by-product

FUEL
~ a 4 Ib)
t95 500 BTU

i
Bio-refiner t Electricalinput
Output 14 500 BTU

160 OO0 BTU

( Stalks and cobs 501bs)
48 3OO BTU
I 0 3 Conversion efficiency

(Electrical power plant)

Figure 11 Energy requirements and credits are estimated from information provided by Chambers 24 and Sanderson? All processing steps from
crop seeding to liquid fuel production are included in this analysis. Net energy: 178200 BTU/bushel; energy balance: out/in = 1.77

214 Enzyme Microb. Technol., 1981, Vol. 3, July

 Biomass refining to fuels and chemicals: P. F. Levy et aL

100 2 Mah, R. A., Smith, M. R. and Baresi, L. Appl. Environ.

O
90 Mierobiol. 1978, 35, 1174-1184
3 Sanderson, J. E., Garcia-Martinez, D. V., George, G. S., Dillon,
80 o o J. J. and Wise, D. L. Liquid Fuels Production from Biomass
70 Contract no. EG-77-C-02-4388-7, Dynatech Report no. 1889,
Cambridge, Ma., 1979
o 60
4 Prins, R. A., Van Nevel, C. J. and Demeyer, D. I. Antonie van
2 50 Leeuwenhoek, J. MierobioL Serol. 1972, 38,281-287
g
40
5 Czerkawski, J. W. and Breckergidge, G. Br. J. Nutr. 1975, 34,
429-446
=>
g
50 6 Trei, J. E., Parish, R. C., Singh, Y. K. and Scott, G. C. J.
cJ 20 Dairy Sci. 1974, 54,536-540
7 McAskill,J. in Proc. 27th Annu. Livestock Growth Feed
10 Manuf. Dealers Salesmen Conf. Bozeman, Montana, 1976,
p. 83
0 2 4 6 8 10 12 14 16 8 Balch, W. E. and Wolfe, R. S. J. BacterioL 1979, 137,264-273
Time (doys) 9 Augenstein, D. C., Wise, D. L. and Cooney, C. L. Resource
Figure 12 Conversion of cornmeal to organic acids by mixed culture
Recov. Conserv. 1976, 2,257
anaerobes in a batch fermentation is shown. Product acid concen-
10 Ashare, E. and Wilson, E. H. Analysis o f Digester Design
Concepts Dynatech Report no. EY-76-C-02-2991, Cambridge,
trations were measured by gas chromatography. Sewage sludge was
used as inoculum, o, Acid production in experimental flask Ma., 1979
11 Wise,D. L., Cooney, C. L. and Augenstein, D. C. Biotechnol.
Bioeng. 1978, 20, 1153-1172
Production of liquid fuels from corn can be subjected to 12 Levy, P. F., Barnard, G. W., Garcia-Martinez, D. V., Sanderson,
an analogous analysis. This scenario is presented in Figure J. E. and Wise, D. L. 'Organic Acid Production from CO2/H2
11. Values for agricultural requirements and side-product and CO/H 2 by Mixed Culture Anaerobics', presented at 73rd
credits are the same as those suggested by Chambers in his Annual AIChE Meeting, Chicago 16 22 November 1980
13 McCarty, P. L. in Special Lecture Series: Advances in Water
ethanol analysis. Conversion of 70% of the corn grain to Quality Improvement University of Texas, Austin, 1966
liquid fuel is assumed. (Higher conversions have already 14 Young, J.C. andMcCarty, P.L.J. WaterPollut. Control. Fed.
been achieved at Dynatech in static fermentations with 1969, 41 (5), 160-173
corn meal see Figure 12.) The major process energy 15 Plummer, A. H., Malina, J. F. and Eckenfelder, W. W.
requirement is for electricity in the electrolysis step. A presented at 23rd Annual Purdue Industrial Waste Conference
Layfayette, Indiana, 1968
conversion efficiency of 30% is assumed for the production 16 Clark, R. H. and Speece, R. E. in Advances in Water Pollution
of electricity from other fuels. The energy balance (output/ Research Pergamon Press, San Francisco, 1971, vol. 1, p. 27
input) for the production of liquid hydrocarbon fuels from 17 Pailthrop, R. E., Richter, G. A. and Filbert, J. W. in Proe. 44th
corn is 1.77. Water Pollut. Contro. Fed. Convention San Francisco,
California, 1971
18 Jewell, W. J., Guest, R. W., Loehr, R. C., Price, D. R., Gunkel,
Acknowledgements W. W. and Van Soest, P. J. Anaerobic Fermentation o f Agri-
cultural Residues - Potential for Improvement and Implemen-
Much of the work reviewed here has been carried out at tation US Dept of Energy, Fuels from Biomass Contract no.
Dynatech R / D Company under funding from the US EY-76-S-02-2981-7, Cornell University, Ithaca, New York,
1978
Department of Energy, the Solar Energy Research Institute, 19 Jones, E. L. Chem. Ind. 1967; no. 38, 1590-1592
the Bioenergy Council, the National Science Foundation 20 Helsel,R. W. Chem. Eng. Progr. 1977, 55-59
and Gulf Oil Corporation. 21 Utley, J. H. P. in Techniques o] Chemistry Wiley, New York,
1974, vol. 5, part 1
22 Allen, M. J. Organic Electrode Processes Reinhold, New York.,
References 1958
23 Hofer, H. and Moest, M. Ann. Chem. 1902,323,284
1 Weimer,P. J. and Zeikus, J. G. Arch. Microbiol. 1978, 119, 24 Chambers, R. S., Herendeen, R. A., Joyce, J. J. and Penner,
175-182 P. S. Science 1979, 206,789-793

E n z y m e M i c r o b . T e c h n o l . , 1981, V o l . 3, J u l y 215