Biotechnd. Rog.

1991, 7, 369-378 309

Kinetic Model for Substrate Utilization and Methane Production in the

Anaerobic Digestion of Organic Feeds
Archana Barthakur, Munindra Bora, and H. Devendra Singh'
Biochemistry Division, Regional Research Laboratory, Jorhat 785 006, Assam, India

A kinetic model for anaerobic digestion of organic substrates is presented. Two

complementary kinetic equations, one each for substrate utilization and methane
production, are derived. The model equations take into account extracellular hydrol-
ysis of complex substrates and consider the hydrolyzed products as the limiting substrates
for cell growth and product formation according to Monod kinetics. The equations are
applied to the experimental data on the anaerobic digestion of a variety of feeds reported
in the literature with reasonable agreement of the predicted and observed values. When
substrate hydrolysis is poor and rate limiting, the equations are reduced to a Contois-
type equation. On the other hand, if no hydrolysis of the substrate is involved as in
the case of simple and soluble substrate, the equations take the form of a Monod-type
relationship. Thus, the equations represent a generalized kinetic expression for
fermentation relating the Monod and Contois equations, which hold good in two extreme
cases.

Introduction concentration (X)as

Various kinetic models have been advanced for anaer-
obic digestion of waste biomass and methane production.
Most of the early models were based on a single-culture
system and used the Monod equation or its variations
(1-3). Recently, several dynamic simulation models were
developed on the basis of a continuousmulticulture system
corresponding to major bioconversion steps in anaerobic
digestion with the assumption of culture growth according
to Monod-type kinetics (4-7). Doubt was, however,
expressed by several investigators on the validity of
applying the Monod equation in waste treatment (8-10).
In the Monod equation ( I I ) , specific growth rate is
expressed as a function only of the concentration of the
limiting substrate in the reactor. It does not contain any
term for influent substrate concentration, implying that
effluent substrate concentration is independent of the
influent substrate concentration. Experimental results
of various investigatorson anaerobic and aerobic digestion
of waste, however, were not in accord with this implication
of the Monod equation. Dependence of the effluent
substrate concentration on the influent was observed on
examination of the experimental results of the anaerobic
digestion of dairy manure (2), of beef cattle manure at
mesophilic and thermophilic temperatures (12,13),of rice
straw (14), of poultry litter (15), and of water hyacinth
(16). In aerobic chemostat operations with glucose as the
growth-limiting substrate, Grady et al. (8) showed that
effluent concentration of the growth-limiting substrate
expressed as chemical oxygen demand (COD) was not
independent of the concentration of substrate entering
the reactor when pure or heterogenouscultures were used.
However, when the substrate was measured as glucose,
the Monod equation was found to be true in pure cultures
but not in heterogenous cultures.
In the models proposed by Contois (17) and Fujimoto
(181,specific growth rate ( p ) was considered as a function
of the growth-limiting nutrient in both the influent and
effluent and it was expressed in terms of microbial
Correspondence should be addressed to this author.
875&7938/Q113007-0369$02.50/0 0 19s1 American Chemlcal Soclety and American Institute of Chemical Englnwrs
where S is the growth-limitingsubstrate concentration in
the reactor, p m is the maximum specific growth rate, and
C is an empirical constant. On the basis of the Contois
equation, Chen and Hashimoto (10,19)developed kinetic
models for substrate utilization and methane production
and showed that the models were in accord with the
experimental data of various investigators on aerobic and
anaerobic digestion of various organic feeds. A Contois-
type relationship was also used to describe the kinetics of
cell growth on nonmiscible liquid hydrocarbon (20, 21)
where substrate transfer could be a limiting factor. The
Contois equation has a drawback as it contains an empirical
constant C that is ill-defined and ill-understood.
The basic validity of the Monod equation was demon-
strated in chemostata using pure cultures and simple
substrates (8,22). Deviation from the Monod relationship
in many digestion systems may be due to the complexity
of the systems. In the digestion of organic wastes, growth-
limiting substrate, due to necessity, was generally ex-
pressed in terms of COD or volatile solids (VS), which
may not truly reflect the nature of the growth-limiting
substrate. Digestion of complex organic wastes involve
the hydrolysis of polymeric compounds (cellulose, hemi-
cellulose, protein, etc.) constituting the waste by exoen-
zymes in the extracellular medium or on the surface/
vicinity of the cells. Only the hydrolyzed assimilable
compounds may be considered as the growth-limiting
substrate in terms of the Monod relationship. There are
various reports that extracellular hydrolysis is the rate-
limiting step in the anaerobic digestion of organic wastes
(7,23-25). It is necessary to take into account the hy-
drolysis process in model building for organic waste
digestion.
From theoretical consideration, kinetic treatment of
anaerobic digestion on the basis of a multiculture system
may be desirable in view of the heterogenous nature of the
microbial population bringing about the various biocon-
version steps involved. However,the kinetic models based
970
on multiculture systems necessarily involve a good number
of kinetic equations and coefficients, which makes the
models highly complex, as shown by the reported models
(4-7). Further, the incorporation of all the cultures
involved in the anaerobic digestion in a design of the mul-
ticulture system cannot be easily done.
Considering these problems, there is a strong case in
favor of a simpler kinetic treatment of anaerobic digestion
on the basis of a single culture system. Methanogenesis
may be specially suited for such treatment as there is a
strong holistic characteristic in the process. Various
cultures and bioconversion steps in methanogenesis are
interdependent and the whole process has certain self-
regulatory characteristics within ita process limits. The
most important factor in the treatment of the process is
the identification of the rate-limiting step.
In this paper, a kinetic model for substrate utilization
and methane production in the anaerobic digestion of
organic feeds is presented that takes into account the
processes of hydrolysis, substrate uptake, and cell growth.
Derivation of two kinetic equations from the model, one
each for substrate utilization and methane production, is
described. It is also shown that the basic derivation from
the model represents a generalized equation that takes
the form of the Monod and Contois equations in two
extreme conditions.
Development of the Model
The following considerations were taken into account
in the development of the kinetic model.
(1)Complex polymeric compounds cannot be taken up
by the cells without initial hydrolysis into assimilable
compounds. The direct substrate for cell growth and
product formation is the hydrolyzed assimilable com-
pounds.
(2) Transport of hydrolyzed assimilable compounds into
the cells is not rate limiting. In the anaerobic digestion
of cellulose, Pavlostathis et al. (7) did not find any
significant pool of hydrolysis products and concluded that
the conversion of particulate cellulose to soluble products
governed the rate of fermentation. Lee and Donaldson
(5) also observed a low pool of soluble carbon in cellulose
fermentation.
(3)The whole process of anaerobic digestion is assumed
to be carried out by a multiculture complex that utilizes
the hydrolyzed assimilable compounds. The various parts
constituting this multiculture complex interact with each
other, behaving as an organic whole. As a corollary from
this assumption, the kinetic constants in the model are
related to this complex and are global in scope.
The digestion process is assumed to take place in three
stages: (I) extracellular hydrolysis of complex compounds
into soluble assimilable substrates, (11)transport of soluble
assimilable substrates into cells, and (111) utilization of
assimilable substrates for cell growth and product for-
mation.
Stage I. Hydrolysis is assumed to be a first-order
reaction with respect to the concentration of hydrolyzable
substrate S (mass/volume):
where s h is the concentration of hydrolyzed substrate
(mass/volume) and Kh is the hydrolysis rate coefficient
(time-'). The validity of this assumption was experimen-
tally demonstrated by Pavlostathis and Gossett (6).
Stage 11. Internalization/transport of hydrolyzed
substrate into the cell is considered to be directly related
Bbtechnol. h g . , 1991, Vol. 7, No. 4
to the difference in concentrations of the hydrolyzed
substrate outside and inside the cells and to the concen-
tration of the active cell biomass X (mass/volume). It is
assumed that hydrolyzed substrate entering into the cells
is metabolized very fast so that the intracellular concen-
tration S, = 0. When the uptake of hydrolyzed substrate
is not rate limiting with respect to hydrolysis, the following
relationship can be written:
(3)
where k is the hydrolyzed substrate transport rate coef-
ficient (time-').
Equations 2 and 3 give
(4)
Stage 111. Cell growth on hydrolyzed (assimilable)
substrate is assumed to follow Monod kinetics and is
expressed as
where K, is the half-saturation constant with respect to
hydrolyzed substrate (mass/volume).
Upon substitution of the value of s h from eq 4, eq 5
becomes
Under steady-state conditions of continuous digestion in
a completely mixed reactor without recycling, the following
relationship holds:
p = 110 (7)
F = (So- S)/O
(8)
where 0 is the hydraulic retention time, F is the volu-
metric substrate removal rate [mass/ (volume-time)], SO
is the influent hydrolyzable (biodegradable) substrate
concentration (mass/volume) , and S is the hydrolyzable
substrate concentration (mass/volume) in the effluent or
in the digester.
To achieve simplification of the model, maintenance
energy and microbial decay are considered small so that
the biomass yield coefficient Y (cell mass/substrate mass)
is constant. Then
- - - Y(S0- S)
x = FY
c1
By use of eq 9, eq 6 can be expressed as
pm KakY So-S K,
-=- -+ - + 1
P K h S s
This is the basic equation for substrate utilization in the
anaerobic digestion of complex organic substances.
K, for hydrolyzed substrate is expected to have a small
value. K, for sugar in methane fermentation was reported
to be 0.24 g/L (6). When S >> K,, which may be the case
in the practical anaerobic digestion of complex feeds, the
second term of the right-hand side of eq 10 becomes
negligible and the equation degenerates to
This equation has the structure of the Contois equation
Botectmd. Rog., 1991, Vol. 7, No. 4
as described in eq 1. It is similar to the Chen-Hashimoto
equation (10, 19):
where K = YC. From eqs 11and 12, K = K,kY/Kh and
C = K,k/Kh. The dimensionless kinetic parameter K of
the Chen-Hashimoto equation and the ill-defined em-
pirical constant C of the Contois equation are thus
completely expressed in terms of the metabolically iden-
tifiable parameters K,, k , Y , and Kh.
In the case of readily hydrolyzable substrates, Kh may
be very large in relation to other values. In the extreme
case of soluble and assimilable substrates, e.g., glucose,
Kh = m. In such cases, the first term of the right-hand
side of eq 10becomesnegligible and the equation is reduced
to
which is the Monod equation. It shows that eq 10 is a
generalized equation relating the Monod and Contois
equations, which hold good in two extreme cases.
By use of eq 7, eq 10 can be rearranged as
where A = K,kY/Kh, to show that the effluent substrate
concentration is dependent on the influent substrate
concentration.
At washout, So = S, then eq 10 can be written as
It shows that if S >> K,, p = p m at critical retention time.
Equation 10 can be rearranged as
Spm
P=
A(So-S)+K,+S
to show that as S approaches zero, p approaches zero.
Incorporationof RefractoryCoefficient. In the case
of complex organic substrates, which are generally ex-
pressed as COD or VS, a part of the substrate is usually
refractory to biodegradation. The refractory coefficient
R is expressed as S,/STO,where S,and Sm are respectively
the refractory and total COD or VS concentrations in the
influent feed. When STdenotes the effluent total COD
or VS concentration, the following expressions can be
written:
By use of eqs 7,17, and 18, the basic equation 10 can be
expressed as
Equation 19 can be rearranged as
to show the relationship between influent and effluent
concentrations of total COD or VS.
371
Volumetric substrate removal rate F may be expressed
as
F3--So-S - Sm(1-R) pmO-l K0
e e p,e + A + Qpme +A- 1)
(21)
where Sm/8 denotes the loading rate of total COD or VS.
Methane Production Kinetics. In anaerobic waste
stabilization, the ultimate oxygen demand of the waste
being degraded was reported to be equal to the ultimate
oxygen demand of the methane gas produced for a wide
variety of wastes (26). Through stoichiometry, the de-
struction of l g of waste COD was shown to be equal to
the production of 0.35 L of methane at STP (26). From
this relationship McCarty (26) showed that waste COD
destroyed could be calculated from the methane produced.
If it is assumed that the ratio of biodegradable COD
reduction and biodegradable VS reduction remains con-
stant throughout the anaerobic process, the above rela-
tionship can be used for calculation of methane production
in terms of VS destruction after suitable correction of the
COD/VS ratio, which is generally constant for a given
residue.
Following the derivation of Chen and Hashimoto (19),
if B denotes the specific methane yield in liters at STP per
gram of COD or VS added to the digester and Bo denotes
the maximum specific methane yield in liters at STP per
gram of COD or VS at infinite retention time, the
biodegradable COD or VS in the digester will be directly
proportional to Bo - B and Bo will be directly proportional
to the biodegradable COD or VS loading. Then the
following relationship can be derived
-- B
S0-s --
S Bo-B
By use of eq 22, eq 10 may be expressed as
%=A- B + -K+,1
p Bo-B S
The above equation representa the kinetic equation for
methane fermentation. By use of eq 17,the above equation
may be rearranged to give
or
Equations 24 and 25 show that, at constant influent
substrate concentration, as 0 approaches m, BIB0 ap-
proaches unity.
Volumetric methane production rate M,,in volume of
CH4 per digestion volume per unit time, may be expressed
by use of eq 25 as
+ K,/ ( S , - RS,)
M,=-=-
e e
A
] (26)
Estimation of the Kinetic Parameters. For the
estimation of the kinetic parameters pm, A, K,, and Bo and
the refractory coefficient R, the nonlinear least-squares
method as described by Draper and Smith (27) was used
by minimizing the combined sum of squares (5”) for
372 Biotedmol. R ~ J1991,
, Vol. 7, No. 4

Table I. Kinetic Conmtnntr and Refractory Coefficients in the Anaerobic Digestion of Various Feeds
_, -.
dairy manure 32.5 34.9 (VS) 0.751 0.585 0.45 0.28 0.223
52.3 0.890 0.585 0.45 0.28 0.223 2
69.8 1.220 0.585 0.45 0.28 0.223
87.2 1.500 0.585 0.45 0.28 0.223
cattle waste 35 49 (COD) 0.50 0.40 0.25 0.30 0.37
73 0.50 0.40 0.25 0.30 0.37 12
93 0.64 0.40 0.25 0.30 0.37
cattle waste 55 49 (COD) 0.36 0.40 0.50 0.30 0.34
73 0.53 0.40 0.50 0.30 0.34 12
95 0.57 0.40 0.50 0.30 0.34
116 0.70 0.40 0.50 0.30 0.34
acetic acid 35 3.135 (acetic acid) O.OOO18 O.OOO18 0.440 0.30 0.332 1
propionic acid 35 5.62 (COD) 0.0014 0.0014 0.274 0.25 0.370 1

responses of s T / s T o and B through the best-fit criterion
where the subscript u,exp denotes the uth experimental
observation corresponding to the independent variable
e,, qprddenotes the predicted value from the models
through use of the assigned parameter values at e,, and
n is the total number of observations. The parameter
values were varied to minimize T as defined in eq 27 until
the best values were obtained. Equation 20 was used to
obtain the predicted values Of sT/sTo, and eq 25 was used
for calculating the predicted values of B. Such a method
of multiple parameter estimation was also used by Chiu
et al. (28) in the study of kinetic behavior of mixed
populations of activated sludge.
The comparison between the predicted and experimen-
tally observed values O f S T and B was done by calculating
the coefficient of determination (r2). As r2 approaches
unity, increasing correspondence between the predicted
and observed values is indicated.
Computation was done by using a Horizon Unique
System computer of Hindustan Computer Ltd., Calcutta,
India.
Evaluation of the Kinetic Model with
Experimental Results
Literature data on semicontinuous anaerobic digestion
and methane production from various complex materials
and pure substrates were analyzed and kinetic parameters
were determined on the basis of the above kinetic models.
STand B values as predicted by the model equations were
compared with the experimental values. Data for anaer-
obic digestion of a complex substrate, cattle manure,
reported by Morris (2)and Hashimoto (12)were selected
for analysis as the data were extensive with regard to
influent substrate concentration, retention time, and tem-
perature of digestion. The work on the anaerobic digestion
of acetic and propionic acids reported by Lawrence and
McCarty (1) gave data of an extreme condition where ex-
tracellular hydrolysis of the substrate was not involved.
Examination of the validity of the model equations under
such a condition was considered important.
Anaerobic Digestion of Dairy Manure. Morris (2)
carried out anaerobic digestion of dairy manure at 0 values
of 30,20,10,5,and 2.5days and ST^ values of 34.9,52.3,
69.8,and 87.2g of TVS/L in a semicontinuous mode with
daily feeding and wasting at a digester temperature of
32.5 OC. For the calculation of the kinetic parameters,
the data for 6 = 5and 2.5days were omitted as the methane
production was extremely low, possibly due to the accu-
mulation of volatile fatty acids (VFA).A basic requirement
for eq 25 is that, for a given feed, methane production per
unit mass of the feed digested should be constant. If
methane fermentation is associated with the accumulation
of other products such as VFA,the above condition cannot
be fulfilled. Accumulation of VFA is known to be an
indication of impending reactor upset. A few methane
production data were also omitted in the calculation as
the results were far from satisfying the above condition.
The values of A, R, pm, K,,and BOcalculated according
to the model equations from the data of Morris are given
in Table I. From these parameter values, the values of ST
predicted by eq 20were compared with the experimentally
observed values in Figure 1. The values match closely
with the line of perfect fit with r2= 0.999. The comparison
of observed values of B with the predicted values is shown
in Figure 2. Though there is scattering of observed data
points, the r2 value of 0.748 indicates the acceptability of
the fit. The value of B is highly susceptible to experimental
error, as it is obtained by dividing methane production by
feed loading. Any error in the estimation of methane
production and Sn, is, therefore, magnified in the calcu-
lated value of B.
The value of R for dairy manure was found to be 0.585,
which is similar to the value of 0.584 calculated by Chen
and Hashimoto (9) from the same data of Morris. The
value also agrees with the experimentally determined
values ranging from 0.55 to 0.60 obtained by Morris (2).
The value of 0.45 day1 for pm obtained in this study is
slightly higher than the value of 0.411 day1 obtained by
Chen and Hashimoto (10) from the same data by using
the substrate utilization equation developed by them. They
also reported lower values of 0.284.33 day1 for pm
calculated from the same data by using an equation for
methane production (19). The reason for the latter low
values for pm may be due to the inclusion of questionable
data for methane production, as discussed above, in their
calculation. By fitting the data to a Monod-type equation,
Morris (2) obtained straight-line curves in the plots of
1/Svs 0 for each S.~O and observed that pm was unaffected
by variation in Sn, but K,increased with the increase of
SO.The K,value of 26-101 g of TVS/L obtained by him
cannot be considered real. He obtained a high pm value
of 1.0 day1 but its determination in terms of the Monod
equation is questionable because of the variable KB
obtained in the plots. This exercise by Morris clearly
indicates the inadequacy of the Monod equation in
describing the kinetics of anaerobic digestion of complex
feeds. The low K, value of 0.28 g/L for hydrolyzed
substrate obtained in the present analysis is comparable
with the reported KOvalues of 0.24 g/L for sugar (5) and
~ t e C h n dBog.,
. lQQl,
Vol. 7, No. 4 373
I
*Ol /
70 -
1
60-
/ h
A
70
I

8
c
30-
g
a
20-
2ot /
10
I lot/ 10 20 30 40 50 60 70
L
80
0 10 20 30 40 50 60 70
OBSERVED sT(gcoD/L)
OBSERVED %( g VSIL)
Figure 3. Comparison of observed STwith predicted values for
Figure 1. Comparison of observed and predicted STvalues for anaerobic semicontinuous digestion of cattle waste at 35 and 55
anaerobic semicontinuous fermentation of dairy manure at 32.5 "C. Data from Hashimoto (12). Sm (g of COD/L) at 35 " C (0)
"C. Data from Morris (2). Sm (g of VS/L): (0) 34.9,(0) 52.3, 49, (0)73,and (A)95. Sm (g of COD/L) at 55 "C: ( 0 )49, (W)
(A)69.8,and (X) 87.2. 73, (A)95,and (X) 116.

this was not so in the prediction of B, which is a very

h
oz5 t sensitive value. It was found by these investigators (12,
19,29) that only different values of K varied in relation
to changes in Sn, could make satisfactory predictions of
B.
0
W
0 The Bo value of 0.223 L/g obtained in this analysis is
0
a equivalent to a methane production of 0.335 L a t STP/g
9 020. of COD digested. This value is close to the stoichiometric
a
-I
methane value of 0.35 L at STP/g of COD destroyed as
Y

m reported by McCarty (26).

0 Anaerobic Digestion of Cattle Waste. Hashimoto
F (12) studied semicontinuous methane fermentation of
u0
g 015'
a of 116 and 142 g/L produced high levels of VFA a t all the
010
0'10 015 02 0'25
OBSERVED B ( L / g VS ADDED)
Figure 2. Comparison of observed methane yield ( B ) with
predicted values for anaerobic fermentation of dairy manure at
32.5 "C. Data from Morris (2). Sm (g of VS/L): (0) 34.9,( 0 )
52.3, (A)69.8,and (X) 87.2.
0.143-0.207 g/L for acetate (I)in the anaerobic methane
fermentation.
For anaerobic digestion of dairy manure, eq 10 is almost
equivalent to eq 11 or 12 as the second term of the right-
hand side of eq 10 has a relatively negligible value. Then
A = K. In the present study A was found to be increased
along with the increase in Sn, (Table I). But Chen and
Hashimoto (10) in their analysis of the data according to
eq 12 obtained a constant value of 1.02 for K a t all Sn.
In the present experiment, the average value of A was
found to be 1.09, which is very close to the above value of
K. The apparent independence of K from S n observed
by them may be due to the fact that prediction of STwas
less affected by the variation in K so that an average value
of K was sufficient for satisfactory prediction of ST.But
cattle waste a t 35 OC and at different Sn, values of 49,73,
95,116, and 142 g of COD/L. At 35 OC, high S n values
6 values studied; therefore, these data were not considered
in the analysis. At a low 8 value of 8 days, S n values of
73 and 95 g/L also produced high levels of VFA and the
data were considered unsuitable for analysis. At 55 "C,
digestion and methane production were very much im-
proved. Even a t a low 8 value of 4 days and a high Sn,
value of 116 g/L, VFA production was within the accept-
able limit. Data for 6 = 5 days and S n = 49 and 73 g/L
were omitted in the analysis, as the methane production
was found to be erratic.
Kinetic parameters and refractory coefficientcomputed
by use of eqs 20, 25, and 27 are shown in Table I.
Comparison of predicted ST with the observed values
showed close correspondence with r2 = 0.94 for data a t 35
"C and r2 = 0.99 for data a t 55 OC (Figure 3). The
correspondence of the predicted B with the observed
values, however, was not as good as the above but within
acceptable limits, with r2 values of 0.77 and 0.70 for data
at 35 and 55 "C,respectively (Figure 4). The effect of B
and S n on ST is depicted in Figure 5. At low 8, ST
decreased rapidly with the increase in 8, but as 8 was further
increased, only a slight decline in S was observed. The
figure also shows that with the increase of Sn,, ST also
increased, showing the dependence of ST on Sn,.
The value of wm at 55 "C expectedly was higher than
that at 35 "C. These values, however, were somewhat
lower than the values of 0.326 and 0.586 day1 at 35 and
974 Bbtechnol. Frog., 1991, Vol. 7, No. 4
i 0'3

J
0.4-
ol
z -02
-I
v

-
h
m 0
0 I
J
0.3- 2. A -1
2 v

0
w
f
- 01
LT
a
0.2- /I2= r2=0.77( 35<$
0.70( 5c))

0.1
0 .l 0.2 0.3 0.4
OBSERVED B(L1gCOD)
Figure 4. Comparison of observed and predicted values of B for Md
Figure 6. Comparison of predicted (continuous broken lines)
with observed values of (0) ST, (A)B, and (0) in the anaer-
obic fermentation of acetic acid at 35 O C . Data from Lawrence
anaerobicsemicontinuousfermentation of cattle waste at 35 and and McCarty (1) (digester 2).
55 O C . Data from Hashimoto (12). Sm (g of COD L) at 35 "C:
(0) 49, ( 0 )73, (A)95. sm(g of COD/L) at 55
73, (A) 95, and (X) 116.
4:
( 0 )49, (H)

'is
6104 r2.0939
I
4 03 3

1
0 10 20
e (DAYS)
Figure 5. Effect of Sm and 0 on STin anaerobic digestion of
cattle waste at 35 "C (continuouslines) and 55 O C (brokenlines).
Data from Hashimoto (12). A comparison between predicted
(lines) and observed values (symbols) is shown.
55 "C, respectively, obtained by Hashimoto (12) through
the use of an empirical equation. The R value of 0.4for
cattle waste was much lower than the value of 0.585
obtained for dairy waste as discussed above. This may be
due to the difference in the quality of these two cattle
manures. The waste used by Hashimoto (12) had 1.14 g
of COD/g of VS, whereas the manure used by Morris (2)
had 1.26g of COD/g of VS. Further, the former had high
VFA content (7% of TS), lower pH (4.9-5.01, and lower
alkalinity (32mg/g of TS) in comparison to the latter,
which had pH 7.3 and higher alkalinity (110mg/g of TS).
The waste used by Hashimoto appears to have undergone
considerable predigestion as indicated by the low pH and
high VFA. This might be the reason for overall very good
performance in waste utilization and methane production
as reported by Hashimoto (12). A low R value of 0.43 for
dairy manure was also reported by other investigators (30).
I I
0 2 4 6 8 1 0 1 2
@(DAYS)
Figure 7. Comparison of predicted (continuous broken lines)
with observed values of (0) M/
ST,(A)E , and (0) in the anaer-
obic fermentationof propionic acid at 35 "C. Data from Lawrence
and McCarty (1) (digester 9).
30
The K,value was low as expected for small molecules of
hydrolyzed products. There was a tendency for a rise in
A along with the rise in Sn, though the rise was not as
steep as that observed in the case of the Morris data (Table
I). The values were also much lower than those obtained
from the Morris data as presented above. This may be
due to the character of the feed, which showed a state of
considerable degradation, probably leading to a higher
K h . Bo values for digestion a t 35 and 55 O C were nearly
the same. Hashimoto (12)also obtained a Bovalue of 0.42
L/g of VS (0.37L/gof COD) at both temperatures, which
is close to the values obtained in this study.
Anaerobic Digestion of Acetic and Propionic Acids.
Acetic and propionic acids are soluble and assimilable
substrates for anaerobic digestion and methane produc-
tion. Hydrolysis step prior to substrate transport into the
cells will not be involved in these cases and, therefore, the
value of A will be insignificant. R values also are expected
to be insignificant as these substrates are considered to be
completely digestible. With these substrates, eq 10 is
expected to be reduced to eq 13, which is the Monod
equation. Work of Lawrence and McCarty (1) on the
anaerobic digestion of acetic acid (digester 2) covered
Wtechnol. Rog., 1991, Vol. 7, No. 4 375

known (31). KBvalues are comparable with the reported

values in the literature, and Bo values are close to the
stoichiometric values. For the digestion of acetic acid and
propionic acids, Figures 6 and 7 respectively show the
comparison of the predicted ST,B, and M,values with the
observed values by plotting the values against 6. In the
acetic acid digestion, predicted values matched reasonably
well with the observed values with r2 = 0.873 for STand
0.684 for B. The matching was much better in the case
of propionic acid digestion with r2 = 0.936 for ST and 0.939
for B. With the increase in 6, STfollowed a typical curve
dictated by the Monod relationship.
The above analysis shows that the kinetics of digestion
of these substrates are more akin to a Monod-type
L
O- 5 l0 15 20 2 5 0 2 L 6 8 10 relationship.
e (DAW e (DAYS) Volumetric Substrate Removal Rate (F)and Vol-
Figure 8. Effect of 0 and Sm on Fin the anaerobicdigestion of umetric Methane Production Rate (M,). Volumetric
cattle waste at (A) 35 O C and (B) 55 O C . Data from Hashimoto substrate removal and methane production rates are
(12). Comparisons between predicted F (continuous lines) and important factors in the design and operation of anaer-
observed values (symbols)are shown. obic treatment and methane production systems. Figure
8 shows that in the anaerobic digestion of cattle waste
X
(12),with Sm remaining constant ( A becomes constant),
,
.-- *
. F increases with a decrease in 6 until F reaches a peak
followed by rapid decline near the critical 8 (washout).
/
The figure also shows that, at a given 8, as Sm increases,
,:',*-/
--. ,
\
,\

5'
/ sTo=;9 g CoO/L(O)
35 c
10 20 30
LOADING RATE ( 9 CODIL-D)
Figure 9. Effect of substrate loading rate and Sm on M,in the
anaerobic digestion of cattle waste at 35 O C (continuous lines)
and 55 O C (broken lines). Data from Hashimoto (12). A
comparison between predicted M, (lines) and observed values
(symbols) is shown.
experiments at 6 values of 11.8, 7.9, 5.9, 3.9, 2.8, 2.2, and
1.4 days with Sm = 3.135 g/L measured as acetic acid.
The data for 6 = 3.9 days were omitted in the analysis as
material balance and methane production were poor. Data
for 6 = 2.2 and 1.4 days were also omitted as the
experiments were conducted below the critical 8 (2.3 days).
Propionic acid digestion was conducted at 6 values of 11.5,
8.5,5.6,3.9,2.4, and 1.6 days with Sm = 5.62 g of COD/L
(digester 9). In the analysis, data for 8 = 2.4 and 1.6 days
were omitted as the experiments were conducted below
the critical 8 (3.8 days).
Table I shows the computed kinetic parameters and
refractory coefficients for the above digestions. As ex-
pected, negligible values for A and R were obtained for
both digestions. pm for acetic acid digestion was found to
be higher than that for propionic acid digestion. Relatively
poor growth of propionic acid assimilating organisms is
\
F also increases and F is much higher a t 55 OC than a t 35
"C. Figure 9 shows the effect of substrate loading rate on
\
M, in the methane fermentation of cattle waste (12). It
shows that at constant Sm, M,predicted by eq 26 increases
with the increase in substrate loading rate until it reaches
a maximum followed by decline as washout starts. It also
shows that, at a given retention time, M, increases with
the increase in Sm. In practice, theoretical maximum M v
is generally unattainable due to accumulation of VFA. In
the digestion of acetic and propionic acids, the effect of
6 on M, is shown in Figure 6 and 7 depicting a pattern
similar to that above.
Conclusions
Predictions by the proposed kinetic model agree rea-
sonably well with the experimental results obtained by
various investigators on the anaerobic digestion of a variety
of feeds. This supports the basic validity of the model
equations (eqs 19 and 23). Application of the equations
40 50 to digestion of poorly hydrolyzable substrates has shown
that the equations take the form of a Contois-type
relationship, whereas with soluble substrates the equations
are reduced to a Monod-type expression. Fermentations
that fall between these two extreme cases will also be fully
described by the equations, showing that these equations
represent a generalized expression for fermentations.
Equation 19 will be applicable to aerobic fermentations
as well, indicating the universal character of the kinetic
model.
Notation
A ( = K J z Y / K h )kinetic parameter
B specific methane yield, L of CHd/g of substrate
added
BO maximum specific methane yield, L of CHd/g of
substrate added at infinite 0
C empirical constant in Contois equation
F volumetric substrate removal rate, g/(L-day)
k hydrolyzed substrate transport rate coefficient,
time-'
K (= Y C ) kinetic parameter in Chen-Hashimoto
equation
Kh substrate hydrolysis rate coefficient, time-'

half-saturation constant for hydrolyzed substrate,
g/L
volumetric methane production rate, L of CHI/
(Laday)
coefficient of determination
(=S,/Sn) refractory coefficient
concentration of hydrolyzed substrate, g / L
intracellular concentration of hydrolyzed substrate,
EI/L
influent biodegradable substrate concentration,
g/L
biodegradable substrate concentration in the ef-
fluent or in the digester, g/L
refractory substrate concentration in the influent,
BIL
(=SO+ S,) total substrate concentration in the
influent, g / L
(=S + S,) total substrate concentration in the
effluent, g/L
concentration of active cell biomass, g/L
biomass yield coefficient, cell mass/substrate mass
hydraulic retention time, days
specific growth rate of organism, days-'
maximum specific growth rate of organism, days-'
Acknowledgment
Financial support for this work from the Department
of Nonconventional Energy Sources, Government of India,
New Delhi, and from the North Eastern Council, Shillong,
India, is acknowledged.
Literature Cited
(1) Lawrence, A. W.; McCarthy, P. L. Kinetics of methane
fermentation inanaerobic waste treatment. TechnicalReport
No 75,Department of Civil Engineering, Stanford University,
Stanford, CA, 1967.
(2) Morris, G. R. Anaerobic fermentation of animal wastes: A
kinetic and empirical design eualuation. M.S. Thesis, Cor-
ne11 University, Ithaca, NY, 1976.
(3) Ghosh, S.;Klass, D. L. Process Biochem. 1978,13, 15.
(4) Kleinstreuer, C.; Poweigha, T. Biotechnol. Bioeng. 1982,24,
1941.
(5) Lee, D. D.; Donaldson, T. L. Biotechnol. Bioeng. Symp. 1984,
14,503.
(6) Pavlostathis, S.G.; Gossett, J. M. Biotechnol. Bioeng. 1986,
28, 1519.
(7) Pavlostathis, S.G.; Miller, T. L.; Wolin, M. J. Appl. Enoiron.
Microbiol. 1988,54, 2660.
Bbtechnol. Rug., 1991, Vol. 7, No. 4
(8) Grady, C. P. L., Jr.; Harlow, L. J.; Riesing, R. R. Biotechnol.
Bioeng. 1972,24,391.
(9) Pfeffer, J. T.Biotechnol. Bioeng. 1974,16,771.
(10) Chen, Y. R.; Hashimoto, A. G. Biotechnol. Bioeng. 1980,22,
2081.
(11) Monod, J. Ann. Znst. Pasteur 1950, 79, 390.
(12) Hashimoto, A. G. Biotechnol. Bioeng. 1982,24,2039.
(13) Varel, V. H.; Isaacson, H. R.; Bryant, M. P. Appl. Enuiron.
Microbiol. 1977,33, 298.
(14) Lequerica, J. L.; Valles, S.; Flors, A. Appl. Microbiol. Bio-
technol. 1984,19,70.
(15) Webb, A. R.; Hawkes, F. R. Agric. Wastes 1986,13,31.
(16) Pillai,K.R.;Unni,B.G.;Barthakur,A.;Singh,H.D.;Baruah,
J. N. Production of Biogas from Water hyacinth. In Water
hyacinth; Thyagarajan, G. Ed.; UNEP Proceedings Series 7;
United Nations Environmental Programmes: Nairobi, Kenya,
1984;p 715.
(17) Contois, D.E. J. Gen. Microbiol. 1959,21,40.
(18) Fujimoto, Y. J. Theor. Biol. 1963,5,171.
(19) Chen, Y.R.; Hashimoto, A. G. Biotechnol. Bioeng. Symp.
1978,8, 269.
(20) Miura, Y.; Okazaki, M.; Murakawa, S. I.; Hamada, S. I.;
Ohno, K. Biotechnol. Bioeng. 1977,19,715.
(21) Coma, G.; Ribot, D. Biotechnol. Bioeng. 1978,20,1723.
(22) Herbert, D. Continuous culture of microorganisms: Some
theoretical aspects. In Recent Progress in Microbiology;
Symposia held at the Seventh International Congress for Mi-
crobiology; Tunevell, G., Ed., Almquist and Wiksell: Stock-
holm, Sweden, 1958;p 381.
(23) Pfeffer, J. T. J. Water Pollut. Control Fed. 1968,40,1920.
(24) Chan, D. B.; Pearson, E. A. Comprehensiue studies of solid
waste management-hydrolysis rate of cellulose in anaer-
obic fermentation. SERL Report No. 103; University of
California: Berkeley, CA, 1970.
(25) Van Velsen,A. F. M.; Lettinga, G. Effect of feed composition
on digester performance. In Anaerobic Digestion; Stafford,
D. A,, Weatley, B. I., Hughes, D. E. Eds.; Applied Science
Publishers Ltd.: London, England, 1980;p 113.
(26) McCarty, P. L. Public Works 1964,(Sept.), 107.
(27) Draper, N.;Smith, H.Applied Regression Analysis; John
Wiley & Sons, Inc.: New York, 1981.
(28) Chiu, Y.; Fan, L. T.; Kao, I. C.; Erickson, L. E. Biotechnol.
Bioeng. 1972,14, 179.
(29) Hashimoto, A. G.Biotechnol. Bioeng. 1983,25, 185.
(30) Dalrymple, W.; Proctor, D. E. Water Sewage Works 1967,
114,361.
(31) Bryant, M. P. J. Animal Sei. 1979,48,193.
Accepted May 20,1991.
Registry No. Methane, 74-82-8;acetic acid, 64-19-7;propi-
onic acid, 79-09-4.