Bioresource Technology 96 (2005) 169–180

A respirometric method for characterising the organic

composition and biodegradation kinetics and the
temperature influence on the biodegradation kinetics, for a mixture
of sludge and bulking agent to be co-composted
a,*
A. Tremier , A. de Guardia b, C. Massiani c, E. Paul d, J.L. Martel a

a
SUEZ Environment, Organic Effluents Division, CIRSEE, 38 avenue du President, Wilson 78230 Le Pecq, France
b
Cemagref, Livestock and Municipal Waste Management Research Unit, 17 av. de Cucille, CS 64427 35044 Rennes Cedex, France
c
Universite de Provence, Laboratory of Chemistry and Environment, Centre St. Charles, pl. Victor Hugo, 13331 Marseille Cedex 3, France
d
INSA Toulouse, Laboratory of Environmental Process Engineering (LIPE, EA833), 135 av. de Rangueil, 31 077 Toulouse Cedex 4, France
Received 28 May 2003; received in revised form 2 December 2003
Available online 28 July 2004

Abstract
A respirometric method was set up to study kinetics of biological reactions involved in the treatment of organic wastes––sludge
mixed with pine barks––by composting. Oxygen consumption rates of this type of mixture were monitored during 10–20 days, using
a 10 l respirometric cell kept at constant temperature and moisture. Oxygen consumption kinetics were modelled and organic matter
composition was characterised as biomass, easily-biodegradable, slowly-biodegradable and non-biodegradable organic matter. The
influence of temperature on kinetics was tested. Results show that this respirometric method is a useful tool for the characterisation
of solid organic matter biodegradability and for the modelling of the biological kinetics of the composting process.

2004 Elsevier Ltd. All rights reserved.

Keywords: Composting; Sludge; Respirometry; Organic matter characterisation; Biodegradation kinetics; Modelling; Biodegradability

1. Introduction metabolism. During the second phase of the composting

Composting, or aerobic biological treatment of or-
ganic wastes, is an ancestral way to reduce wastes and to
reuse organic matter. Among the range of existing or-
ganic wastes, sewage sludge composting with the use of
woody bulking agents, such as pine barks, enables the
production of a quality product that may be used as a
soil conditioner or as an organic fertiliser. However, the
quality of the end-product largely depends on the way to
perform the initial mixture and to manage reactions
occurring during the composting treatment.
The composting process includes two major phases.
The first one, called the ‘‘active phase’’, mainly develops
degrading reactions: dissolved organic matter is used as
carbon and energy source by microorganisms for their
*
Corresponding author.
E-mail address: anne.tremier@cemagref.fr (A. Tremier).
process, called the ‘‘curing phase’’, organic macromole-
cules such as humic substances are synthesised. As we
already mentioned, degrading reactions form the central
phenomena of the ‘‘active phase’’ of the treatment. These
reactions, based on numerous biological, thermal and
physico-chemical phenomena, involve oxygen con-
sumption, as well as heat, water and carbon dioxide
production. As all the above phenomena interfere with
one another, they are quite difficult to characterise on the
sole experimental basis obtained from composting trials.
Thus, describing each phenomenon and all the pheno-
mena interferences, thanks to mathematical equations
and, particularly, modelling the biological reactions of
the active phase of composting, should lead to a simu-
lation of the behaviour of the substrates treated by this
process and then should allow a better understanding
and optimisation of the composting treatments.
Knowing the kinetics of biological reactions and the
influencing environmental parameters is essential to

0960-8524/$ - see front matter
2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biortech.2004.05.005
170 A. Tremier et al. / Bioresource Technology 96 (2005) 169–180
reach the objective of simulating biological degradation
of organic matter, without consideration of any mass
transfer limitation. In such a case, the kinetics of the
organic matter degradation and the quantity of de-
graded matter are directly linked with biomass growth
kinetics (Bailey and Ollis, 1986):
1 metric method respectful of the real physical structure of
rX ¼ lX and rS ¼ lX
YX =S
with rX : biomass growth kinetics (mol O2 /kg/h), l: bio-
mass specific growth rate (h1 ), X : microorganisms
concentration (mol O2 /kg), rS : substrate degradation
kinetics (mol O2 /kg/h), YX =S : biomass growth yield (–).
There is also a relationship between biomass growth,
substrate degradation and oxygen consumption:
rO2 ¼ ð1  YX =S ÞrS ¼ ½ð1  YX =S Þ=YX =S rX
with rO2 : oxygen consumption rate (mol O2 /kg/h).
As biomass or substrate concentrations are difficult to
measure, biological reactions can be studied through
respirometric methods by monitoring instantaneous
oxygen consumption rate and by knowing the biomass
growth yield. Indeed, respirometry is the measurement
and interpretation of the biological oxygen consumption
under well-defined experimental conditions (Spanjers
et al., 1998). Numerous respirometric methods are used
to characterise organic matter composition and biodeg-
radation in liquid effluents (wastewater, slurries, etc.).
Such methods have also been developed to study
organic solid wastes, in order to provide a maturity or
stability index. Ianotti-Frost et al. (1992), Ianotti et al.
(1994) evaluated compost stability through a gaseous
dissolved oxygen measurement in a closed bottle. The
AT4 (German index for respiration activity in 4 days)
evaluates stability with regard to the cumulative quan-
tity of consumed oxygen measured for 96 h (Binner and
Zach, 1999). The DRI (Dynamic Respiration Index),
proposed by Adani et al. (2001, 2002), considers, as a
stability index, the average instantaneous oxygen con-
sumption kinetics, measured on a sample for 24 h after
the maximum respirometric activity has been reached.
Finally, stability has been studied by Lasaridi and
Stentiford (1998) and Stentiford (2002), by placing
compost in a liquid medium and by measuring the
maximal oxygen uptake rate (SOUR). Some authors,
such as Lasaridi et al. (1996), have also applied a res-
pirometric method to solid wastes so as to study organic
matter biodegradation kinetics as a function of envi-
ronmental parameters such as temperature.
The most important limit concerning the previous
methods was the measurement of oxygen consumption
on small quantities of solid sample (30–100 g). More-
over, these samples were often ground and sometimes
they were suspended in aqueous solution before mea-
surement. As a consequence, the result concerning sta-
bility level or biodegradation kinetics would not take
into account the influence of the solid matrix charac-
teristics of the substrate (moisture, granulometry, etc.).
Thus, the results of stabilisation or biodegradation
kinetics might be quite different when considering the
treatment of a real waste.
The first aim of this study was to develop a respiro-
samples and to apply it to the special case of a solid
mixture of sludge and bulking agent, in which homo-
geneous oxygen supply might be a problem. While using
this method, two objectives were pursued: (1) To char-
acterise the initial organic composition of the mixture as
biomass concentration and as easily-biodegradable or-
ganic matter concentration, as well as slowly-biode-
gradable and non-biodegradable organic matter
concentration, thanks to the validation of a kinetic
model simulating the oxygen uptake rate by considering
biomass growth and organic matter hydrolysis limita-
tion; (2) To study the influence of temperature on the
biodegradation kinetics at constant moisture.
2. Methods
2.1. Respirometric device design
The respirometric device (Fig. 1) was set up on the
basis of Aguilar-Juarez (2000). In order to model
household wastes biodegradation kinetics during the
aerobic phase of a landfill cell exploitation period, Ag-
uilar-Juarez developed a 1.8-l closed respirometric cell.
This cell, filled with a representative sample of house-
hold wastes, was continuously aerated and the oxygen
consumption rate within the substrate was recorded.
Considering the specificity of the studied substrate
(sludge and bulking agent), the respirometric device
developed in this study was bigger than Aguilar-Juarez’s
device so that the reacting medium would be compara-
ble to a composting one. The proposed respirometer
consisted of a 10-l hermetically-closed glass cell, filled
with the substrate mixture placed on a grid. Ambient air
was blown into the mixture through a pipe placed at the
bottom of the cell. The flow rate (Q) was maintained
constant and precisely measured through a volumetric
gas counter. In order to provide homogeneous aeration
conditions, a rapid recirculation of the gas in the cell
was carried out. So, considering the solid organic sub-
strate, this device was a closed vessel reactor. Consid-
ering the gas phase, the respirometric device could be
used as an open reactor (continuous injection of ambi-
ent air and continuous air exit) or as a closed reactor
(only recirculation of the initial gas volume).
Temperature and moisture were kept constant during
each experiment, by placing the respirometric cell in a
thermostatic water bath and by humidifying the
incoming air and condensing the water in the exhaust
 A. Tremier et al. / Bioresource Technology 96 (2005) 169–180
Flowmeter
Pump
Volumic
counter
Ambiant air entry
Recirculation
Fig. 1. Respirometric measurement device.
gas. Temperature was controlled by a Pt 100 probe
placed in the mixture. The oxygen consumption rate was
obtained by measuring oxygen concentration both in the
incoming and exhaust gas, thanks to a paramagnetic
oxygen gas analyser.
2.2. Substrates and respirometric measurements charac-
teristics
Mixtures of wastewater treatment sludge and pine
barks were used as substrates for the respirometric
study.
Sludges were produced during the biological treat-
ment of wastewater mainly from food processing
industries. They were directly sampled at the wastewater
treatment plant after centrifugation. Sludges were then
sub-sampled at the laboratory and stored by freezing
them (1–10 kg) in order to work with constant substrate
quality. Pine barks were used as bulking agent in the
mixture. They were sieved in order to remove large
pieces (longer than 10 cm and larger than 4 cm), which
were too big for the respirometric cell. Mixtures were
always performed on a 1/1 wet mass ratio in order to
simulate a mixture to be composted.
Twelve respirometric measurements were performed
with an average of 2.3 kg of wet mixture each time.
Temperatures between 19.6 and 56.9 C were tested.
Oxygen consumption rate was measured, as described
above, for 10–20 days––until it became low and con-
stant.
2.3. Chemical analyses
At the beginning of the experiments, sludge and
mixture were analysed according to the following
parameters:
171
Gas
analyser
Condenser Ambiant air
Thermostated water bath
Air exit
• Dry matter content (DM): wet samples (three repli-
cates of about 1 kg) dried at 80 C to constant weight.
A higher temperature was not used because of the fire
hazard with this type of sample.
• Organic matter content (OM): dried ground samples
calcinated at 550 C using the standard method NF U
44-160 (AFNOR, 1985).
• Total organic carbon content (TOC): dried ground
samples oxidised to CO2 and measurement of CO2
by infrared spectrometry (SKALAR device), accord-
ing to the standardised method NF-EN-13137 (AF-
NOR, 2001).
• Chemical oxygen demand (COD): COD was deter-
mined by dichromate oxidation on 50 mg of dried
ground sample, according to an adaptation of the
standard method described in the norm NF T 90-
101 (AFNOR, 1971) for powdered samples.
• Kjeldhal nitrogen content (TKN): TKN was quanti-
fied on 50 mg of dried ground sample by mineralisa-
tion within a strong acid medium (sulphuric acid
98%), followed by steam distillation and then titri-
metric determination (adaptation of the standard
method NF ISO 11261 (AFNOR, 1995) for pow-
dered samples.
The mean chemical characteristics of sludges and
tested mixtures, and the coefficient of variation are
summarised in Table 1.
Dry matter content was also measured on the mixture
at the end of the respirometric test.
2.4. Respirometric method validation
In order to study and model the biodegradation
kinetics, the respirometric measurement had to rely only
on the intrinsic parameters of the microbiological
172 A. Tremier et al. / Bioresource Technology 96 (2005) 169–180

Table 1
Mean characteristics of sewage sludge and sewage sludge and pine barks mixtures
DM (%) OM (% DM) COD (mg O2 /g DM) TOC (mg C/g DM) TKN (mg N/g DM)
Sewage sludge Mean value 17.1 83.9 1347 461 75.9
CVa (%) 1.5 0.1 1.6 4.9 3.6
Sewage sludge and Mean value 47.0 94.1 1439 510 15.4
pine barks mixtures CVa (%) 1.5 1.9 4.1 4.9 8.6
a
Coefficient of variation.

activity such as the biomass concentration, the bio-
degradable organic matter concentration, etc. The
environmental conditions (temperature, moisture, etc.)
and the oxygen supply have to be as constant and as
homogeneous as possible in order not to influence
kinetics during measurement. Hence, the first step of the
following work was to check that the method was con-
trolled as far as the oxygen supply, temperature and
moisture are concerned, and that it could be reproduced.
The homogeneity of the aeration in the reacting cell
and the airtightness of the cell were controlled through a
characterisation of the gas flow patterns in the respiro-
metric device. In this purpose, known amounts of
methane were injected as a gas tracer in the respiro-
metric device closed on the gas recirculation loop. Four
recirculation flow rates (3Q, 4.5Q, 6Q, 7.5Q where Q is
the incoming air flow rate in case of trials with the open
gas system) were tested. The evolution of the methane
concentration was monitored through an infrared gas
analyser.
The temperature and moisture maintenance was
verified by calculating their coefficient of variation
during each of the twelve respirometric measurements
performed.
Finally, measurement repeatability was tested by
comparing three experiments (A–C) carried out on
identical substrate mixtures at comparable temperatures
It is assumed that a solid organic substrate is a three
phase matrix (De Guardia, 2002; Mustin, 1987). The
first phase is a dry solid composed of mineral and or-
ganic matter. The second phase is the aqueous phase
formed by the moisture of the material, and composed
of mineral and organic soluble matter, where biological
reactions take place. The heterotrophic biomass X is
contained in this phase. The third phase is the gaseous
phase due to the material porosity, which allows gaseous
exchanges between the aqueous phase and the environ-
ment. Within this matrix, the organic matter is to be
described in three fractions: the easily-biodegradable
fraction MB which is already soluble and which may be
immediately consumed by microorganisms; the slowly-
biodegradable fraction MH which is composed of solid
or soluble macromolecules that are to be hydrolysed
before biological consumption; the inert (or non-bio-
degradable) fraction MI which is not or very little con-
cerned by biological reactions according to the time
scale of the composting process active phase. Kinetic
model’s equations are presented as follows.
dX MBðtÞ
¼ lm X ðtÞ  bX ðtÞ ð1Þ
dt KB þ MBðtÞ
MHðtÞ
dMH X ðtÞ
¼ Kh MHðtÞ X ðtÞ ð2Þ
dt þ KMH
X ðtÞ

(around 37 C) and moistures (around 53.0%). Trials A dMB 1 MBðtÞ

¼  lm X ðtÞ
and B were performed at the same time. Trial C was dt Y KB þ MBðtÞ
carried out one month later. MHðtÞ
X ðtÞ
þ Kh MHðtÞ X ðtÞ ð3Þ
X ðtÞ
þ KMH
2.5. Conceptual approach of the biological oxygen con-
sumption ð1  Y Þ MBðtÞ
rO2 ðtÞ ¼ lm X ðtÞ
Y KB þ MBðtÞ
As mentioned in the introduction, the oxygen uptake þ bð1  f ÞX ðtÞ ð4Þ
rate measured in biological reacting systems is directly
linked with the biological activity and the organic mat- As described by Eq. (1), microorganisms use the
ter biodegradation. In the wastewater treatment field, easily-biodegradable fraction as a source of carbon and
the development of the activated sludge theory has led energy. Actually, a part of MB is transformed to give
to propose conceptual models (ASM) to simulate oxy- new biomass cells (Eq. (3)). The coefficient of biomass
gen uptake rate and characterise the organic matter growth yield (ratio between biomass formed and avail-
composition as a function of several biological processes able MB) is called Y . The other part (1  Y ) of MB is
(carbon oxidation, nitrification and denitrification) oxidised in presence of an electron acceptor, which is
(Henze et al., 2000). In the context of solid organic oxygen in the case of an aerobic process such as com-
matter biodegradation, only the concepts concerning posting (Eq. (4)). This oxidation supplies energy and
carbon oxidation by heterotrophic biomass were con- produces heat, water and carbon dioxide. The other
sidered as explained in the following. biodegradable organic fraction MH is reduced to simple
 A. Tremier et al. / Bioresource Technology 96 (2005) 169–180
organic molecules, which become part of the fraction
MB, through enzymatic hydrolysis reactions as exposed
in Eq. (2) (Eliosov and Argaman, 1995). The biomass X ,
which develops along with MB consumption is also lost
by decay. The dead biomass is partly transformed in
inert matter MI (fraction f ) and partly oxidised in
presence of oxygen (fraction (1  f )): it is the endoge-
nous respiration concept. As a consequence, oxygen
consumption is proportional to the MB biodegradation
and the biomass decay (Eq. (4)).
To sum up, the conceptual model of solid organic
matter degradation is composed of:
• three substrate active components: X , MH, MB
expressed in mmol O2 /kg of substrate’s DM or mg
O2 /g of substrate’s DM
• five kinetic parameters: lm (h1 ), KB (mmol O2 /kg of
substrate’s DM or mg O2 /g of substrate’s DM), b
(h1 ), Kh (h1 ), KMH (–)
• two dimensionless stoichiometric parameters: Y , f
2.6. Parameters estimation and model numerical solution
In order to estimate the values of the model para-
meters for the twelve experimental curves, the dynamic
biological model was fitted to the experimental data.
For each experiment, a computer program optimised
model parameters by calibrating the modelled oxygen
consumption kinetics result with the experimental oxy-
gen consumption kinetics, using a least square method.
To this purpose, coupled differential equations involved
in the model were solved with a SCILAB ODE’s pack-
age on the basis of an initial set of parameters. An
identification procedure based on a structural identifi-
ability study (Dochain et al., 1995; Sperandio and Paul,
2000), was developed in order to estimate groups of
initial parameters, using linearisation hypothesis on
800
700
600
[CH4] ppm
500
400
300
200
100
0
0 50 100
Fig. 2. Gas flow behaviour inside the respirometric cell as a function of the recirculation flow rate.
173
growth and hydrolysis periods of the experimental
curves. Then, the kinetics (l and Kh ) and the charac-
terisation of initial values (X0 , MB0 and MH0 ) were
obtained after fixing default values, based on literature
references, for Y, f, b, KB , and KMH .
3. Results and discussion
3.1. Respirometric method validation
3.1.1. Homogenous aeration validation
Fig. 2 shows that whatever the flow rate, the evolu-
tion of the methane concentration in the reacting cell
was identical. After a few seconds the gas analyser de-
tected methane. The concentration grew rapidly until it
reached a constant value. This constant value meant
that the closed gas respirometric device was perfectly
airtight: no loss of gas tracer could be established.
Moreover, this evolution was characteristic of the mixed
flow pattern (Levenspiel, 1999) with a little delay due to
the time required by the tracer to enter the cell just after
being injected. That is to say that the gas flow was
perfectly homogeneous in the cell after a few minutes:
the concentration of the gas tracer in the gas phase
was the same everywhere in the respirometric device.
Finally, the constant methane concentration made it
possible to calculate the gaseous volume in the respiro-
metric device. This volume was the same as the one
calculated thanks to the characteristics of the device and
the porosity of the solid substrate. Therefore it demon-
strated that no dead volume was detected in the system.
According to the above results, the respirometric device
allowed to supply homogeneous gas phase without any
misbehaving in the flow (dead volume, short cut, etc.).
Fig. 2 also shows that mixing time decreased along with
increasing recirculation flow rate. It was decided to work
Recirculation flow rate = 3*Q
Recirculation flow rate = 4.5*Q
Recirculation flow rate = 6*Q
Recirculation flow rate = 7.5*Q
150 200 250 300 350 400 450 500
Time (s)
174 A. Tremier et al. / Bioresource Technology 96 (2005) 169–180
at the fastest recirculation rate (7.5 times faster than the
ambient airflow rate injected during trials with the open
gas system), which homogenised the gas phase within 2
min. As a consequence, homogenization time concern-
ing the gas phase of the reacting cell was roughly five
times higher than the injected air residence time in the
respirometric system.
3.1.2. Temperature and moisture maintenance
For ten of the twelve experiments, the coefficient of
variation of the mean temperature value during the
measurement was less than 5%. Only two higher coeffi-
cients of variation (6.7% and 6.3%) were observed dur-
ing experiments at low temperature. Indeed in these last
two cases, temperature maintenance was much more
influenced by ambient air temperature variation. Con-
cerning moisture maintenance during experiments, var-
iation around an average value between the initial and
the final moisture content was always less than 4%.
3.1.3. Measurement repeatability
Fig. 3 displays the respirometric curves for trials A–
C. Trials A–C were carried out on the same substrate
mixture and the same water bath temperature. While
trials A and B were carried out at the same date, trial C
was carried out one month later.
All curves presented the same behaviour. For the
three trials, the values of the maximum oxygen uptake
rate showed a coefficient of variation of 8.2%. The val-
ues of the total consumed oxygen amount, calculated
between the beginning of the experiment and the same
measurement date (defined as the minimal duration
necessary to observe a low and apparently constant
value of respiration rate in each case), presented a
coefficient of variation of 17.5%. In fact, total amounts
45
40
35
mmol O2.h-1.kg DM-1
30
25
20
15
10
0
0 50
Fig. 3. Respirometric measurement repeatability.
of consumed oxygen were exactly equal for trials A and
B, which were done at exactly the same temperature
(36.6 C). However this amount was higher for trial C
carried out one month later at a temperature higher of 3
C (39.6 C), in spite of a comparable temperature im-
posed in the water bath. One must keep in mind that
oxygen consumption is linked with the growth kinetics
and the death kinetics of microorganisms, which both
depend on temperature. Between trials A, B and trial C,
temperatures were slightly different. Then, the same
stage of organic matter biodegradation might have not
been reached exactly at the same date and it could partly
explain the gap in cumulative oxygen consumption be-
tween the three experiments. Thus, considering the
natural physical heterogeneity of the mixture and the
deviation in temperature, the validation of the mea-
surement repeatability was accepted.
3.2. Experimental respirometric curves and respirometric
modelling
Typical examples of respirometric curves are given on
Fig. 3. The proposed model leads to an interpretation of
the respirometric measurement as follows. After a time
lag, biomass consumes oxygen in order to degrade the
most easily-biodegradable substrate and to sustain its
own growth requirements. With the increasing biomass,
the oxygen consumption accelerates until the whole of
the easily-biodegradable substrate is degraded. Conse-
quently, a break in the exponential growth is observed.
The biomass keeps on developing by consuming the
substrate coming from the hydrolysis of the slowly-
biodegradable substrate. As a consequence, the oxygen
consumption rate decreases regularly down to the
complete disappearance of all the biodegradable sub-
Oxygen uptake rate - Trial A
Oxygen uptake rate - Trial B
Oxygen uptake rate - Trial C
100 150 200 250 300 350 400
Time (h)
 A. Tremier et al. / Bioresource Technology 96 (2005) 169–180 175

strates. These phenomena are verified when no other varying between 9.4% and 27.1% of the experimental
limitation than the availability of the organic substrate value. At this stage of the work, this model precision
occurs. was considered as sufficient and the simplest model was
The simulations were performed for the twelve preferred.
experimental results. For each simulation, model
parameters were calibrated. Fig. 4 presents an example
3.3. Parameter estimation and biodegradable organic
of simulation result.
matter fractionation
Whatever the tested temperature, the simulation of
the growing phase was always quite representative. The
Stoichiometric parameters (Y ; f ) were initially esti-
maximum oxygen uptake rate obtained with simulation
mated on the literature basis (ASM values). The factor f
was sometimes higher than the experimental one but the
was kept constant (0.2) as its influence on simulation
difference did not have much influence on the total
results was very low. Y value was calibrated on experi-
amount of consumed oxygen. The decreasing phase was
mental results. A mean value of 0.68 was found with a
generally in accordance with the experimental data. The
coefficient of variation of 11.6%. This result is in con-
most difficult phenomenon to represent was the break
formity with the biomass growth yields (0.56–0.80)
phase. Indeed, the theoretical model can represent a
found for easily-biodegradable substrates such as glu-
sharp decrease for a little time after the maximum
cose or volatile fatty acids (Sperandio, 1998).
oxygen uptake rate, which corresponds to the beginning
The constant parameter KB (saturation constant for
of an exponential decreasing phase. Actually, such a
substrate MB) was set to 3 mmol O2 /kg DM (85 mg O2 /l
sharp break was not observed in each experimental case.
aqueous phase) which is equivalent to the recommended
Experimental break slopes were often more gradual and
value of the ASM (50–150 mg O2 /l). This value was not
as a consequence were quite difficult to simulate pre-
optimised because of its low influence on the simulation
cisely.
results. KMH (hydrolysis saturation constant for the ratio
Such a difference between the experimental data and
MH/X ) was initially set to 6.5 (–). After optimization, a
the model result would perhaps be lowered by consid-
mean value of 6.54 ± 11.9% was found. No real reference
ering a model with several slowly-biodegradable frac-
may be found in the literature for this parameter, which
tions. These fractions would then have different
is quite new even in the activated sludge models (Henze
hydrolysis kinetics. Nevertheless, the simple model
et al., 2000).
proposed in this study made it possible to model the
Other kinetic parameters such as growth rate,
experimental curves with a mean error factor
hydrolysis rate and death rate varied with temperature.
0

1


The growth rate (lm ) ranged from lows of 0.13 to highs

rO2 expðiÞ  rO2 modelðiÞ

B C of 0.34 h1 for temperatures varying between 19.6 and
@ A
rO2 expðiÞ 56.9 C. As a comparison, a typical value of 0.25 h1
at 20 C was found in the ASM models, in the case of

45

40

35
-1
mmol O2.h-1.kg DM

30

25 Simulated oxygen uptake rate

Experimental uptake rate
20

15

10

0
0 50 100 150 200 250 300 350 400 450 500
Time (h)

Fig. 4. Example of simulation result (tested temperature: 39.6 C).

176 A. Tremier et al. / Bioresource Technology 96 (2005) 169–180
liquid substrates. Moreover, in its composting model
applied to vegetable substrates, Kaiser (1996) used a lm
value of 0.2 h1 at 40 C, corresponding to the bacteria
growth rate. In the same way, the hydrolysis rate (Kh )
varied between values ranging from 0.10 to 0.20 h1 ,
which were in conformity with the typical ASM cited
value of 0.125 h1 at 20 C. Death rates (b) were cal-
culated between 0.05 and 0.13 h1 . These results may be
compared to the ones reported by Henze et al. (2000),
which widely varied with substrate, and ranged from
0.002 h1 for domestics sewage in the USA to 0.07 h1
for some food-processing wastes with no information
about the considered temperature. It is very interesting
to notice that kinetics concerning solid substrates
treatment and liquid substrates treatment are quite
comparable. Then, the longer process durations ob-
served during biological solid wastes treatments seem to
depend more on the substrates intrinsic characteristics
differences (organic fractions quantities, initial biomass
content) than on the enzymatic hydrolysis rate and on
the biomass growth and death kinetics.
The oxygen uptake rate modelling enabled us to
determine values of organic matter fractions: the initial
biomass content (X0 ), the initial easily-biodegradable
organic matter content (MB0 ), the initial slowly-bio-
degradable organic matter content (MH0 ). These values
did not vary along with the temperature, as they repre-
sent intrinsic characteristics of the organic substrates.
The mean values and their coefficient of variation
calculated on the basis of the 12 trials are summarised in
Table 2. It can be immediately noticed that the biomass
initial content could not be determined precisely (co-
efficient of variation of 65.1%). Several causes for this
large variation must be stressed. In fact, X0 initial con-
tent partly depends on defrosting conditions of the
sludge sample. For example, the weight of the frozen
sample (1 or 10 kg) might lead to a different behaviour
of the biomass during defrosting. Moreover, X0 was so
low in comparison with the total organic content (ex-
pressed as COD) that any error on the oxygen con-
sumption amount had a great influence on the X0 value
estimation. Thus, X0 , as determined by this method, has
to be considered as a qualitative value with a large
standard deviation. MB0 and MH0 were much better
estimated. The coefficients of variation on these values
(respectively 19.7% and 14.5%) were in conformity with
the one accepted for measurement repeatability. It must
be pointed out that the MB0 /X0 ratio is very high (170)
Table 2
Organic matter fractionation for the studied substrate mixture
X0 (mg O2 /g DM)
Mean value 0.07
Coefficient of variation (%) 65.1
X0 : initial biomass content; MB0 : initial easily-biodegradable organic matter content; MH0 : initial slowly-biodegradable organic matter content.
in comparison with the typical one (<1), generally ap-
plied in liquid organic substrate treatment with activated
sludge processes. It explains the delay at the beginning
of the biodegradation process in the case of solid wastes:
X has to grow before significant oxygen consumption
kinetics might be measured.
The sum of MB0 and MH0 equals the total amount of
biodegradable organic matter in the given granulometric
structure of the substrate and the time scale of the
process. In this study, it only represented 6.5% of the
total amount of organic matter (expressed as COD) of
this substrate. Such a result might be considered as
being very low in comparison with any other organic
substrates such as pig slurry. In this latter type of sub-
strate, the only easily-biodegradable organic fraction is
about 29% of the total COD (Andreottola et al., 1997).
Two main facts must be considered in order to explain
this difference. First, the studied substrate is a mixture of
sludge and bulking agent. This bulking agent represents
a large part of the total COD but is much less bio-
degradable than the sludge according to the considered
process time scale. Then, the biodegradable fraction
represents a low fraction of the mixture total COD but
would account much more if only the sludge total COD
was considered. Second, the mixture physical structure
plays an important role here. This was previously
pointed out by Stentiford (2002), who observed that the
maximum oxygen uptake rate measured on ground
sample suspended in aqueous solution was higher than
the one measured on the same substrate in its solid form.
The originality of this respirometric modelling was
to allow the organic matter fractionation from a bio-
degradability point of view, taking into account the
physical specificity of the substrate in the composting
treatment. As fractions are defined according to the
biological reactions theory, and estimated through a
measurement of the biological activity (oxygen con-
sumption rate), the comparison of biodegradability be-
tween several substrates will be more reliable than with
other physico-chemical fractionation methods (for
example the comparison of COD content in water ex-
tract and extracted solid part of a substrate). Then, the
future uses of this method might be as follow: the
comparison of the biodegradability of substrates treated
by composting in the physical way they are usually
processed, or the study of the influence of the physical
matrix characteristic (for example granulometry) for a
particular type of substrate.
MB0 (mg O2 /g DM) MH0 (mg O2 /g DM)
11.9 79.9
19.7 14.5
 A. Tremier et al. / Bioresource Technology 96 (2005) 169–180 177

3.4. Temperature influence on biodegradation kinetics
As presented above, simulation results showed that
the growth kinetic factor (lm ), the hydrolysis kinetic
factor (Kh ) and the death kinetic factor (b) varied sig-
nificantly with the temperature. Kinetics first increased
with increasing temperatures. But beyond approxi-
mately 40 C, kinetics started decreasing (Figs. 5–7).
These variations were well modelled with the CTMI
model (Cardinal temperature model with inflection)
(Rosso et al., 1993).
2 limits, the optimum temperature to mesophilic micro-
lm ðT Þ ¼ ðlopt ðT  Tmax ÞðT  Tmin Þ Þ=ððTopt  Tmin Þ
 ½ðTopt  Tmin ÞðT  Topt Þ  ðTopt  Tmax Þ
 ðTopt þ Tmin  2T ÞÞ
This model makes it possible to determine an opti-
mum kinetic factor and the influence of the temperature
in function of three cardinal points: minimum, optimum
and maximum temperatures. Figs. 5–7 present CTMI
modelling results concerning the evolution of growth
(lm ), hydrolysis (Kh ) and death (b) kinetics as a function
of temperature. CTMI model’s parameters are displayed
in Table 3.
Cardinal temperatures describe all types of groups of
typical microorganisms. Thus, the minimum tempera-
ture corresponds to psychrophilic microorganisms low
organisms optimal range and the maximum temperature
to thermophilic microorganisms highest limits (Bailey
and Ollis, 1986). This result confirms the biological

0.45

0.40

0.35
Experimental growth kinetic (µm)
Simulated µm
0.30
h-1

0.25

0.20

0.15

0.10
0.0 10.0 20.0 30.0 40.0 50.0 60.0
Temperature (°C)

Fig. 5. CTMI model for lm ¼ f ðT Þ.

0.30

0.25

Experimental hydrolysis rate (Kh)

Simulated Kh
0.20
-1
h

0.15

0.10

0.05
0.0 10.0 20.0 30.0 40.0 50.0 60.0

Temperature (°C)

Fig. 6. CTMI model for Kh ¼ f ðT Þ.

178 A. Tremier et al. / Bioresource Technology 96 (2005) 169–180

0.18

0.16

0.14
Experimental death kinetics (b)
0.12 Simulated b

0.1
h-1

0.08

0.06

0.04

0.02

0
0.0 10.0 20.0 30.0 40.0 50.0 60.0
Temperature (°C)

Fig. 7. CTMI model for b ¼ f ðT Þ.

Table 3 ways optimal between 35 and 50 C. In the same way,

Parameters of the CTMI model for lm ¼ f ðT Þ, Kh ¼ f ðT Þ and
b ¼ f ðT Þ
lm or Kh or b Tmin (C) Topt (C)
optimum (h1 )
lm ¼ f ðT Þ 0.2836 0 38.5
Kh ¼ f ðT Þ 0.1798 0 38.2
b ¼ f ðT Þ 0.0999 0 38.5
model hypothesis of a global biomass regardless of the
behaviour of particular microorganisms. One might be
surprised by the rather low value of the optimum tem-
perature. Indeed, the highest temperatures usually de-
scribed during aerobic treatment of solid organic matter
through composting processes reach 60–70 C. Never-
theless, these temperatures do not correspond to the
optimum biological activity. Indeed, the reaction rates
decrease at high temperature, but the phenomenon of
heat accumulation in the solid substrate produces a
further increase in temperature although biological
activity is already decreasing (Haug, 1993). Thus, the
determination of the optimum temperature has already
been a much debated issue. Optimum temperatures from
35 to 70 C were described, according to the solid sub-
strate (Jeris and Regan, 1973; Nielsen and Berthelsen,
2002) and the moisture level of this substrate (Liang
et al., 2003). Among studies describing the influence of
temperature on kinetics, several authors have already
mentioned optimum temperatures around 40 C. For
example, McKinley and Vestal (1984) studied the
interactions between the temperature and the activity of
microbial communities in the composting of municipal
sewage sludge. They showed that, whatever the com-
posting time, microbial activities, measured as the rate
of acetate incorporation into microbial lipids, were al-
Davis et al. (1992) demonstrated that, during pine barks
composting, mesophilic microorganisms outnumbered
Tmax (C)
thermophilic ones whatever the stage of the process.
These two results confirm the positive trend of the
67.8
optimum temperature (around 38.5 C) found in the
63.0
67.6 present study with sewage sludge and pine barks mix-
tures.
Thus, it should be concluded that our respirometric
method makes it possible to give a good picture of the
behaviour of microorganisms during the aerobic treat-
ment of sewage sludge mixed with pine barks.
4. Conclusion
The respirometric system presented here is a validated
and reproducible method to measure oxygen consump-
tion kinetics, while respecting the solid matrix charac-
teristics of the substrate, along with controlled
conditions of aeration and maintenance of temperature
and moisture.
Thanks to this method, a kinetic model comparable
to the one proposed for wastewater treatment with
activated sludge was calibrated. Such model enables us
to fractionate the organic content of the substrate and to
estimate the kinetic parameters of the aerobic biode-
gradation reactions. In the case of the studied mixture
(sludge coming mainly from food-processing industry
wastewater treatment and pine barks), calculated
kinetics are very closed to the ones generally observed
during liquid effluents activated sludge treatment. The
main difference between the liquid and the solid treat-
ment, in term of duration, seems to be due to the organic
matter composition. The biodegradable fraction of the
 A. Tremier et al. / Bioresource Technology 96 (2005) 169–180
studied solid mixture was less than 10% of the total
COD. Moreover, this organic biodegradable fraction
was mainly composed of slowly-biodegradable organic
substrate. This composition is probably a characteristic
of the substrate’s physical structure and of the nutrients
accessibility to microorganisms. Indeed, without any
consideration of oxygenation limit, the same material
would probably have given different respirometric re-
sults depending on the tested granulometry: a ground
material will probably release more easily and more
slowly-biodegradable organic fraction according to the
considered time scale. Moreover, the biomass content of
the waste mixture is extremely low. The concentration of
microorganisms represents less than 1/1000 of the bio-
degradable organic fraction. It explains the long time
required to degrade the biodegradable organic matter.
The biodegradation kinetics is also largely influenced by
environmental parameters such as temperature. It con-
firms that the kinetic modelling has to account for all
these influences in order to be representative of complex
processes such as composting.
As a conclusion we may assert that our respirometric
method is a simple reliable tool, which helps to estimate
solid substrate biodegradability and to supply a good
simulation of biological reactions of more complex
processing treatment. Any future development of the
method will give the opportunity to study other types of
substrates and the influence of their intrinsic character-
istics and of their physical matrix characteristics, but
also other environmental parameters such as the med-
ium moisture.
Acknowledgements
This study was led within the frame of a Ph.D. work
funded by SUEZ Environment (CIRSEE), the Cemagref
laboratories (Agricultural and Environmental Engi-
neering Research Institute) and ADEME (French
Agency for Environment and Energy Management).
References
Adani, F., Lozzi, P., Genevini, P., 2001. Determination of biological
stability by oxygen uptake on municipal solid waste and derived
products. Compost Sci. Util. 9 (2), 163–178.
Adani, F., Ubbiali, C., Tambone, F., Centemero, M., Genevini, P.L.,
2002. Static and dynamic respirometric indexes––Italian research
and studies. In: Workshop Conference: The Biological Treatment
of Biodegradable Waste––Technical Aspects. The Environment
DG and the Joint Research Centre of the European Commission,
Brussels.
AFNOR, 1971. NF T 90-101––Determination de la demande chimique
en oxygene (DCO)––Methode par le dichromate de potassium.
AFNOR, 1985. NF U 44-160––Amendements organiques et supports
de culture––Determination de la matiere organique totale––
Methode par calcination.
179
AFNOR, 1995. NF ISO 11261––Qualite du sol––Dosage de l’azote
total––Methode de Kjeldahl modifiee.
AFNOR, 2001. NF EN 13137––Caracterisation des dechets––Dosage
du carbone organique total (COT) dans les dechets, boues et
sediments.
Aguilar-Juarez, O., 2000. Analyse et modelisation des reactions
biologiques aerobies au cours de la phase d’exploitation d’un
casier d’un centre d’enfouissement technique. Ph.D. Thesis, Genie
des procedes, Institut National des Sciences Appliquees de
Toulouse.
Andreottola, G., Bortone, G., Tilche, A., 1997. Experimental valida-
tion of a simulation design model for nitrogen removal in
sequencing batch reactors. Water Sci. Technol. 35 (1), 113–120.
Bailey, J., Ollis, D., 1986. Biochemical Engineering Fundamentals.
McGraw-Hill Book Co, Singapore.
Binner, E., Zach, A., 1999. Laboratory tests describing the biological
reactivity of pretreated residual wastes. In: Bidlingmaier, W.,
deBertoldi, M., Diaz, L., Papadimitriou, E. (Eds.), Organic
Recovery and Biological Treatment––ORBIT 99. Rhombos-Ver-
lag, Weimar, pp. 255–261.
Davis, C.L., Hinch, S.A., Donkin, C.J., Germishuizen, P.J., 1992.
Changes in microbial population numbers during the composting
of pine bark. Bioresour. Technol. 39, 85–92.
De Guardia, A., 2002. Traitement biologique aerobie des dechets
solides: le procede de compostage. In: Moletta, R. (Eds.), Gestion
des problemes environnementaux dans les industries agroalimen-
taires. Technique et Documentation, Paris (Chapter 11).
Dochain, D., Vanrolleghem, P.A., Van Daele, M.U., 1995. Structural
identifiability of biokinetic models of activated sludge respiration.
Water Res. 29 (11), 2571–2578.
Eliosov, B., Argaman, Y., 1995. Hydrolysis of particulate organics in
activated-sludge systems. Water Res. 29 (1), 155–163.
Haug, R., 1993. The Practical Handbook of Compost Engineering.
Lewis Publishers, Boca Raton, FL.
Henze, M., Gujer, W., Mino, T., VanLoosdrecht, M., 2000. Activated
Sludge Models ASM1, ASM2, ASM2D and ASM3. IWA Publish-
ing, London.
Ianotti, D.A., Grebus, M.E., Toth, B.L., Madden, L.V., Hoitink,
H.A.J., 1994. Oxygen respirometry to asses stability and maturity
of composted municipal solid waste. J. Environ. Qual. 23, 1177–
1183.
Ianotti-Frost, D., Toth, B.L., Hoitink, H.A.J., 1992. Compost
stability. Biocycle (November), 62–66.
Jeris, J.S., Regan, R.W., 1973. Controlling environmental parameters
for optimum composting––I: Experimental procedures and tem-
perature. Compost Sci. 14, 10–15.
Kaiser, J., 1996. Modelling composting as a microbial ecosystem: a
simulation approach. Ecol. Model. 91, 25–37.
Lasaridi, K.E., Papadimitriou, E.K., Balis, C., 1996. Development and
demonstration of a thermogradient respirometer. Compost Sci.
Util. 4 (3), 53–61.
Lasaridi, K.E., Stentiford, E.I., 1998. A simple respirometric technique
for assessing compost stability. Water Res. 32 (12), 3717–3723.
Levenspiel, O., 1999. Chemical Reaction Engineering. John Wiley and
Sons, New York.
Liang, C., Das, K.C., McClendon, R.W., 2003. The influence of
temperature and moisture contents regimes on the aerobic micro-
bial activity of a biosolids composting blend. Bioresour. Technol.
86 (2), 131–137.
McKinley, V.L., Vestal, J.R., 1984. Biokinetic analyses of adaptation
and succession: microbial activity in composting municipal sewage
sludge. Appl. Environ. Microbiol. 47 (5), 933–941.
Mustin, M., 1987. Le compost, gestion de la matiere organique. Ed.
Francßois Dubusq.
Nielsen, H., Berthelsen, L., 2002. A model for temperature dependency
of thermophilic composting process rate. Compost Sci. Util. 10 (3),
249–257.
180 A. Tremier et al. / Bioresource Technology 96 (2005) 169–180
Rosso, L., Lobry, J.R., Flandrois, J.P., 1993. An unexpected corre-
lation between cardinal temperatures of microbial growth high-
lighted by a new model. J. Theor. Biol. 162 (4), 447–463.
Spanjers, H., Vanrolleghem, P., Olsson, G., Dold, P., 1998. Respi-
rometry in Control of the Activated Sludge Process: Principles.
International Association on Water Quality, London.
Sperandio, M., 1998. Developpement d’une procedure de comparti-
mentation d’une eau residuaire urbaine et application a la
modelisation dynamique de procedes a boues activees. Ph.D.
Thesis, Genie des procedes, Institut National des Sciences App-
liquees de Toulouse.
Sperandio, M., Paul, E., 2000. Estimation of wastewater biodegrad-
able COD fractions by combining respirometric experiments in
various S0 =X0 ratios. Water Res. 34 (4), 1233–1246.
Stentiford, E., 2002. The specific oxygen uptake rate (SOUR). In:
Workshop Conference: The Biological Treatment of Biodegradable
Waste––Technical Aspects. The Environment DG and the Joint
Research Centre of the European Commission, Brussels.