Waste Management 34 (2014) 2007–2013

Contents lists available at ScienceDirect

Waste Management
journal homepage: www.elsevier.com/locate/wasman

Modeling organic micro pollutant degradation kinetics during sewage

sludge composting
Yumna Sadef a,b,c,d,⇑, Tjalfe Gorm Poulsen a,d,1, Kai Bester b
a
Department of Chemistry, Biotechnology and Environmental Engineering, Aalborg University, Sohngaardsholmsvej 57, 9000 Aalborg, Denmark
b
Department of Environmental Science, Aarhus University, Frederiksborgvej 399, 4000 Roskilde, Denmark
c
College of Earth & Environmental Sciences, University of the Punjab, 54000 Lahore, Pakistan
d
Department of Civil Engineering, Xi’an Jiaotong-Liverpool University, Suzhou 215123, China

a r t i c l e i n f o a b s t r a c t

Article history: Degradation of 13 different organic micro-pollutants in sewage sludge during aerobic composting at 5
Received 22 January 2014 different temperatures over a 52 day period was investigated. Adequacy of two kinetic models: a single
Accepted 1 July 2014 first order, and a dual first order expression (using an early (first 7 days) and a late-time (last 45 days)
Available online 28 July 2014
degradation coefficient), for describing micro-pollutant degradation, and kinetic constant dependency
on composting temperature were evaluated. The results showed that both models provide relatively good
Keywords: descriptions of the degradation process, with the dual first order model being most accurate. The single
Organic micro-pollutant degradation
first order degradation coefficient was 0.025 d1 on average across all compounds and temperatures. At
Kinetic models
Aerobic composting
early times, degradation was about three times faster than at later times. Average values of the early and
Composting temperature late time degradation coefficients for the dual first order model were 0.066 d1 and 0.022 d1, respec-
tively. On average 30% of the initial micro-pollutant mass present in the compost was degraded rapidly
during the early stages of the composting process. Single first order and late time dual first order kinetic
constants were strongly dependent on composting temperature with maximum values at temperatures
of 35–65 °C. In contrast the early time degradation coefficients were relatively independent of compost-
ing temperature.
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction documented to reduce the concentrations of these and other

Composting is one of the most widely adopted methods for con-
verting sewage sludge into a more stable organic material. Organic
chemicals of toxicological concern are often present in sewage
sludge, for instance, personal care products, detergents, plasticizers
etc. originating both from industrial and domestic uses (Aparicio
et al., 2009; Poulsen and Bester, 2010; Sadef et al., 2014a). Sewage
sludge often contains several unwanted organic compounds in ele-
vated concentrations such as Triclosan (an antimicrobial) and its
metabolite Methyl-triclosan (TCS, Me-TCS), detergents as Nonyl-
phenol (NP), plasticizers ((di (2-ethylhexyl) phthalate (DEHP),
tri-iso-butylphosphate (TiBP), tri-n-butyl phosphate (TnBP)),
fragrances (such as the musk fragrances (AHTN), and galaxolide
(HHCB), the HHCB metabolite HHCB-Lactone and the ‘‘non-musk
fragrance (OTNE)’’. Aerobic composting has been extensively
⇑ Corresponding author. Permanent address: Department of Chemistry, Biotech-
nology and Environmental Engineering, Aalborg University, Sohngaardsholmsvej
57, 9000 Aalborg, Denmark.
1
Permanent address: Department of Civil Engineering, Xi’an Jiaotong-Liverpool
University, Suzhou 215123, China.
organic chemicals via biological degradation (Agnew and
Leonard, 2003; Chen and Bester, 2009; Cheng et al., 2008; Haug,
1993; Jones and Westmoreland, 1998; Michel et al., 2001;
Poulsen and Bester, 2010; Sadef et al., 2014b; Semple et al.,
2001). As compost is often used for land application it is important
that it meets required standards for these compounds. However,
the degradation kinetics for these organic pollutants during com-
posting has received comparatively little attention. This is espe-
cially the case with respect to the relationship between
degradation kinetics and composting process conditions such as
temperature, oxygen concentration, or water content. Microbial
degradation of trace organic compounds in soils have often been
described assuming first order kinetics and several studies have
confirmed the applicability of first order kinetics in many cases.
Some examples are: Yan et al. (1995), Staples et al. (1997),
Gavala et al. (2003). For compost, however, the quantity of litera-
ture focusing on degradation kinetics is rather limited. Poulsen
and Bester (2010) studied degradation of 12 organic micro pollu-
tants in sewage sludge during full-scale sludge composting under
thermophilic conditions and used first order degradation kinetics

http://dx.doi.org/10.1016/j.wasman.2014.07.001
0956-053X/Ó 2014 Elsevier Ltd. All rights reserved.
2008 Y. Sadef et al. / Waste Management 34 (2014) 2007–2013
to describe organic micro-pollutant degradation. The adequacy of
first order kinetics for describing degradation, however, was not
evaluated. Also no assessment of the relationship between degra-
dation kinetics and compost process conditions such as tempera-
ture, moisture content, was carried out.
A limited number of studies on organic micro-pollutant degra-
dation in soils, such as Roslev et al. (1998) and Madsen et al. (1999)
found that simple first order kinetics were not always sufficient
and suggested to use multiple-term models allowing kinetics to
vary over time. This type of model, however, has not been applied
to describe degradation of organic contaminants during compost-
ing and its applicability is, thus, not known.
It is well known that temperature plays a major role in degrada-
tion of organic micro pollutants and organic matter both during
active composting (Cheng et al., 2008) and later when compost is
applied as soil amendment (Gonzalez et al., 2012). Increases in
degradation rate are generally observed at elevated temperatures
(Eiland et al., 2001; Epstein, 1997; Goyal et al., 2005; Sadef et al.,
2014b) although unfavorable conditions with respect to aeration
and moisture content can inhibit degradation (Caceres et al., 2006).
Significant degradation of 4-Nonylphenols at composting tem-
peratures between 35 and 55 °C and optimal degradation of Linear
Alkyl Sulfonate (LAS) in sewage sludge at 55 °C has been reported
by Moeller and Reeh (2003a,b) while an optimal temperature for
biodegradation of LAS at 40 °C was observed by Sanz et al. (2006).
Apart from these studies, very little is known about the kinetics
of organic micro pollutant degradation during aerobic sludge
composting and their dependency on composting temperature.
Furthermore to the knowledge of the authors no existing studies
have linked organic micro pollutant degradation kinetics with
composting process temperature.
The objectives of this study were therefore (1) to investigate the
degradation kinetics for organic micro-pollutants during compost-
ing, (2) to describe the kinetics using commonly applied kinetic
models, and (3) to assess the impact of composting temperature
on the values of the kinetic parameters. Degradation kinetics
were investigated using degradation data for 13 organic micro-
pollutants representing the most common groups of organic chem-
icals present in sewage sludge. The micro-pollutants were selected
such that they also represent compounds commonly used in
households and industry.
2. Theory
Microbial degradation of trace compounds has widely been
described using first order degradation kinetics with respect to
compound concentration:
C pollutants (Table 1) in the compost were measured at 4 occasions
¼ eK 1 t ð1Þ
C0
where C0 and C are initial and actual micro pollutant concentra-
tions, respectively, t is time and K1 is the first order degradation
coefficient. Roslev et al. (1998) observed that degradation of DEHP
in sludge applied to soil only followed first order kinetics initially
but later degradation became independent of DEHP concentration.
These authors therefore proposed the following dual-term model
for DEHP degradation in soil:
C number of measurements necessary as measurement of micro-pol-
¼ eK 1 t ; t < t a ð2aÞ
C0
C measurements were used yielding 260 concentration data points.
¼ K N tN ; t  t a ð2bÞ
Ca
where ta is the time at which degradation kinetics becomes Nth
order with respect to time (and independent of concentration) Ca
is the concentration at that point in time and KN is the Nth order
degradation coefficient.
Microbial activity (and degradation rate) typically increases
with increasing temperature until a certain temperature (Tmax)
where it reaches its maximum. At temperatures above Tmax degra-
dation rate decreases with increasing temperature due to inactiva-
tion of the microorganisms. Sadef et al. (2014b) proposed to model
the relative decrease in micro-pollutant concentration over time as
a function of temperature using a Weibull probability distribution
function as:
 a1

a T a
e ð b Þ
T
rðTÞ ¼ ð3Þ
b b
where r is the relative concentration decrease, T is temperature, and
a and b are constants that depend on the chemical being degraded
and the conditions under which degradation occur. The Weibull
probability density function was used as it was found to be able
to fit the degradation data well in comparison with other mathe-
matical expressions of similar complexity. As decrease in relative
micro-pollutant concentration and micro-pollutant degradation
kinetics are related, Eq. (3) was selected in this study to model
temperature–kinetics relationships.
3. Data and statistical methods used
3.1. Sample collection and incubation
Data used in this study were measured by Sadef et al. (2014b)
using compost samples collected from a full scale compost pile at
a commercial composting facility (Komtek composting company,
Southern Denmark). The compost material consisted of 4% straw,
20% yard waste, 10% paper, 10% trolltex (an insulation material
made from wood fibers covered with cement), 15% screening resi-
due (from screening of finished compost) and 41% sewage sludge
by weight. Further details about material collection and treatment
can be found in Sadef et al. (2013). For incubation, 40 compost
samples of 100 g each were taken and placed in separate 500 ml
glass bottles. A loose cotton plug was put in the mouth of each bot-
tle to prevent contamination e.g. via dust of the sample while
allowing oxygen supply from the atmosphere. Samples were then
incubated at 5 different temperatures (18, 30, 37, 50 and 70 °C)
for a period of 56 days. The water content of each bottle was
adjusted by adding demineralized water at a constant water con-
tent of 55% by weight (the typical water content maintained in
the compost during full-scale processing at Komtek). The bottles
were shaken gently by hand once every day to allow sufficient oxy-
gen supply. During this period concentrations of 13 selected micro-
(at 4, 11, 26, and 56 days after initiation of the incubation). Dupli-
cate compost samples (two for each combination of temperature
and time) were incubated. Further details of the measurement
methods applied can be found in Sadef et al. (2014a). An overview
of the 13 compounds is given in Table 1.
The data set therefore contains duplicated values of micro pol-
lutant concentration for 13 micro-pollutants for 5 temperatures
and four points in time equal to a total 520 concentration measure-
ments. This number of combinations were selected to limit the
lutant concentrations is rather time consuming and costly. In the
following analyses average values of the duplicated concentration
To allow for equilibration of the samples and completion of any
lag period in the degradation process, no data prior to the fourth
day of the incubation were used. In the subsequent analyses there-
fore the fourth day of the incubation will be considered the start of
 Y. Sadef et al. / Waste Management 34 (2014) 2007–2013 2009

Table 1
Overview of the 13 organic micro-pollutants used in this study.

Analyte Use Name

AHTN Fragrance 1-(5,6,7,8-Tetrahydro-3,5,5,6,8,8-hexamethyl-2-naphthalenyl)-ethanone
DEHP Plasticizer bis-(2-ethyl-hexyl) phthalate
EHMC UV-blocker 2-Ethylhexyl-4-methoxycinnamate
HHCB Fragrance 1,3,4,6,7,8-Hexahydro-4,6,6,7,8,8 hexamethyl-cyclopenta-(g)-2-benzopyran
MK Fragrance Musk ketone: 1-(4-tert-butyl-2,6-dimethyl-3,5-dinitrophenyl)ethanone
NP Detergent metabolite Technical 4-nonylphenol (consisting of a multitude of different isomers
OTNE Fragrance 1-(1,2,3,4,5,6,7,8-octahydro-2,3,8,8-tetramethyl-2-naphthalenyl)ethanone
TBEP Plasticizer tris(2-butoxyethyl) phosphate
TCPP Flame retardant tris-(chloro-isopropyl) phosphate
TCS Bactericide Triclosan: 2,4,4-trichloro, 2-hydroxy-bisphenylether
TDCP Flame retardant tris(1,3-dichloropropyl) phosphate
TiBP Plasticizer tri-iso-butyl phosphate
TnBP Plasticizer tri-n-butyl phosphate

Table 2
Measured initial concentrations, C (ng g inorganic matter1) and fitted first order, dual first order degradation coefficients (K1, K1a, K1b) at five different temperatures using Eqs. (1
and 4), and estimates of optimal degradation temperature (Topt) using Eq. (3) for the 13 organic micro-pollutants used in this study.

Compound C ng g im1 K1 d1 Topt (°C) K1a d-1 Topt,a (°C) K1b d1 Topt,b (°C)
18a 30a 37a 50a 70a 18a 30a 37a 50a 70a 18a 30a 37a 50a 70a
AHTN 243 0.002 0.016 0.008 0.014 0.019 70 0.041 0.026 0.011 0.026 0.034 18 0.001 0.015 0.008 0.012 0.017 70
DEHP 18,678 0.015 0.024 0.045 0.022 0.015 40 0.064 0.032 0.134 0 0.041 40 0.011 0.023 0.034 0.023 0.013 42
EHMC 132 0.034 0.037 0.036 0.054 0.008 39 0.042 0.039 0.017 0.031 0.034 18 0.032 0.035 0.037 0.056 0.005 41
HHCB 2688 0.012 0.03 0.029 0.04 0.023 49 0.063 0.038 0.037 0.03 0.059 18 0.008 0.03 0.028 0.04 0.02 48
MK 5.21 0.021 0.008 0.029 0.046 0.037 57 0.054 0.08 0.111 0.026 0.095 70 0.016 0.005 0.023 0.048 0.035 58
NP 5341 0.002 0.012 0.003 0.018 0.005 53 0.005 0.018 0.028 0.045 0.103 70 0.001 0.011 0.001 0.017 0.003 51
OTNE 1283 0.026 0.042 0.035 0.054 0.043 54 0.08 0.125 0.058 0.118 0.074 40 0.021 0.034 0.032 0.046 0.041 58
TBEP 143 0.035 0.038 0.046 0.012 0.006 29 0.153 0.102 0.133 0.022 0.055 18 0.025 0.031 0.038 0.008 0.002 30
TCPP 1527 0.007 0.012 0.008 0.012 0.004 39 0.034 0.015 0.067 0.054 0.02 45 0.004 0.013 0.003 0.008 0.002 35
TCS 1336 0.014 0.039 0.03 0.039 0.018 43 0.057 0.023 0.018 0.029 0.059 70 0.012 0.042 0.03 0.04 0.014 42
TDCP 22 0.004 0.013 0.001 0.034 0.036 61 0.055 0.001 0.027 0.065 0.075 64 0.001 0.016 0.003 0.03 0.034 63
TiBP 462 0.044 0.045 0.042 0.086 0.023 45 0.205 0.202 0.269 0.079 0.111 25 0.028 0.03 0.02 0.088 0.013 53
TnBP 28 0.014 0.061 0.001 0.022 0.046 70 0.244 0.072 0.14 0.084 0.063 18 0.001 0.06 0.013 0.015 0.042 18
Average 0.016 0.026 0.02 0.033 0.023 49.923 0.087 0.057 0.068 0.04 0.065 39.538 0.009 0.024 0.016 0.032 0.018 46.846
Stdev. 0.015 0.018 0.019 0.021 0.018 12.426 0.066 0.057 0.08 0.035 0.032 22.255 0.014 0.017 0.018 0.023 0.016 14.259
a
Experiment temperature (°C).

the degradation process (time = 0) and the degradation period,
thus, covers 52 days.
3.2. Statistical methods used
Two different statistical methods for comparing different data
sets were applied in this study: Percentile bootstrapping (Efron
and Tibshirani, 1993) which is used to compare the means of
two data sets or to compare the mean of one data set with a spe-
cific constant, and the Mann – Whitney test (Wilcox, 1997) which
is used to compare the distributions of two data sets. Both methods
are non-parametric which means that they can be used without
any assumptions or restrictions with respect to the distributions
of the data being analyzed. Thus, when using these methods, data
need not be normally distributed. Percentile Bootstrapping is the
non-parametric equivalent of the classical t-test. This method is
based on generating a large number of artificial data sets (typically
1000 or more) each of which contains the same number of data and
has the same statistical characteristics as the measured data. This
is essentially done by random sampling in the set (or sets) of mea-
sured data (repeated until the necessary number of artificial data
sets has been generated). Means and standard deviations for the

Fig. 1. Measured values of Ln(C/C0) as a function of time for six selected combinations of organic micro-pollutant and incubation temperature: (a) micro-pollutants with
degradation following first order kinetics and (b) micro-pollutants with degradation following dual-term first order kinetics.
2010 Y. Sadef et al. / Waste Management 34 (2014) 2007–2013
artificial data sets are then calculated and used to construct a ‘t-
distribution’ that of course is only valid for the measured data. This
‘t-distribution’ is then used to test the mean(s) of the measured
data in a way equivalent to that of the classical t-test. The
Mann–Whitney test is based on merging the two data sets to be
compared and ranking the data values across both data sets. The
sum of the ranks for either data set is then compared to so-called
critical values (that depend on the size of the two data sets) to
assess if the distributions of the two data sets are significantly dif-
ferent. In this study, non-parametric statistical methods were
selected as the measured degradation data could not be assumed
normally distributed, thus eliminating the possibility for using of
traditional parametric statistics.
4. Results and discussion
Values of initial micro-pollutant concentration, C0 (at time
t = 0), are given in Table 2. Concentrations (C) for all 13 micro-pol-
lutants generally decreased with time across all temperatures with
a few exceptions. Values of Ln(C/C0) as a function of time for
selected combinations of micro-pollutant and composting temper-
ature are shown in Fig. 1. It is seen that while degradation in some
cases follow first order kinetics (linear Ln(C/C0) – t relationships,
Fig. 1a), this is not so for others (Fig. 1b). For the compounds in
Fig. 1b the data seem to indicate that degradation approaches first
order kinetics later in the process, as curves tend to become more
linear. The compounds undergo rapid degradation early in the pro-
cess while later degradation becomes slower. A possible reason for
this behavior is that easily accessible micro pollutant molecules
which are dissolved in the water phase or adsorbed lightly to the
outer particle surfaces are degraded initially. Later when these
are exhausted, degradation of the more difficult to access mole-
cules strongly adsorbed to particles or present within the micro-
pores of the particles are degraded.
The data in Fig. 1b suggest that for some micro-pollutants deg-
radation may be modeled using Eq. (2), however, modified such
that Eq. (2b) is replaced by a first order term. As only two data
points are available for modeling the initial, rapid degradation
kinetics it is difficult to judge which kinetic model will be appro-
priate based on the data. Based on the findings of Roslev et al.
(1998) and the fact that microbial degradation most often follows
first order kinetics it was chosen here to model the initial degrada-
tion using first order kinetics. The resulting model is, thus:
C found values of 0–0.033 d1 with an average of 0.018 d1 for 12
¼ eK 1a t ; t < t a
C0
C well. The likely reason that the degradation constants observed in
¼ eK 1b t ; t  t a
Ca
Fig. 2. (a) Fitted (Eq. (1)) and (b) fitted (Eq. (4)) versus measured values of relative micro-pollutant concentrations (C/C0) for all combinations of micro-pollutants (13
compounds), temperature (5 values) and time (4 values).
where C is micro pollutants concentration, t is time and K1a is the
early time first order degradation coefficient, K1b is the late time
degradation coefficient, ta is the point in time when degradation
kinetics changes, and Ca is the micro-pollutant concentration at that
point in time. Based on the measured data, including those in
Fig. 1b, the value of ta was selected as 7 days (the second measure-
ment point). For each combination of micro-pollutant and temper-
ature (65 in total), measured concentration data were fitted to the
first-order model (Eq. (1)) and dual first-order model (Eq. (4)) using
K1, K1a and K1b as fitting parameters. Resulting values of K1, K1a and
K1b as functions of temperature and micro-pollutant are shown in
Table 2.
Fitted versus measured values of C/C0 using Eqs. (1 and 4) are
plotted in Fig. 2. Despite some scatter, measured and fitted values
match relatively well with an average relative deviation of 23% for
the first order (Eq. (1)) and 12% for the dual first-order (Eq. (2))
models. The reason for the increased scatter at C/C0 = 1 in Fig. 2a
is that, Eq. (1) tends to underestimate the measured concentrations
at t = 0. Values of C/C0 > 1 indicate measurements where C was
higher than the initial value (likely due to measurement uncer-
tainty as this tendency was not found significant).
Cumulative probability distributions for the fitted K1, K1a and
K1b – values across all combinations of micro-pollutant and tem-
perature are given in Fig. 3a and b. It is clearly seen that the distri-
butions of the three parameters are not normal but exhibit
extensive tailing for higher K (where K indicates K1, K1a or K1b)
requiring use of non-parametric statistical methods. Average K1
across all temperatures and micro-pollutants was 0.025 d1 with
a standard deviation of (0.017 d1) while average K1a and K1b was
0.066 d1 and 0.022 d1, with standard deviations of 0.058 d1
and 0.017 d1, respectively. Average Ca/C0 was 0.7 with a standard
deviation of 0.21. This means that approximately 30% of the
micro-pollutant mass present in the compost is removed during
the initial rapid stage of degradation.
Values of K > 0 indicate degradation while values less than 0
indicate increasing concentrations with time. Apparent increasing
concentrations are likely a result of measurement uncertainty as
negative K values are few and very minor.
The average values of K1, K1a and K1b across all combinations of
organic compound and temperature were found significantly lar-
ger than zero (p < 0.001). Thus on average degradation takes place
at all temperatures and for all compounds. At elevated tempera-
tures (50–70 °C) K1 range between 0.004 and 0.086 d1 with an
average of 0.026 d1. In comparison Poulsen and Bester (2010)
ð4aÞ
organic chemicals during full-scale composting in the same tem-
perature regime. Thus on average the two data sets seem to agree
ð4bÞ
this study are somewhat higher is that crushed compost samples
 Y. Sadef et al. / Waste Management 34 (2014) 2007–2013
Fig. 3. Cumulative probability distribution function for (a) K1 (F(K1)) as a function of first-order (K1), (b) K1a and K1b, (F(K1a) and F(K1b)) as a function of K1a and K1b,
respectively, (c) F(K1a – K1b) as a function of K1a – K1b, across all temperatures and organic micro-pollutant combinations (65 values).
were used, enabling easier access to the organic compounds, com-
pared to the uncrushed compost used by Poulsen and Bester
(2010). Also constant temperatures were used in this study,
enabling better acclimatization of microorganisms, compared to
using variable temperatures (under full-scale conditions) as done
by Poulsen and Bester (2010).
The cumulative probability distribution for the difference K1a –
K1b calculated for each individual combination of compound and
temperature is shown in Fig. 3c. For 83% of the data, differences
are larger than zero meaning that K1a > K1b. As the data in Fig. 3
are based on relatively few measurements (only two data points
for estimating K1a) the apparent tendency K1a > K1b could be argued
a random result caused by measurement uncertainty. Statistical
analyses were therefore carried out to verify if the result is signif-
icant. As K1a and K1b for any given compound and temperature can-
not be assumed independent, this was evaluated by testing if the
average difference K1a – K1b (Fig. 3c) was significantly larger than
zero using bootstrapping. The test confirmed that K1a was signifi-
cantly larger than K1b (p < 0.001) which means that the tendency
K1a > K1b is real and not due to measurement uncertainty.
Overall the results in Figs. 1–3 suggest that improved descrip-
tion of organic micro-pollutant degradation in compost may be
achieved using a dual first order model equivalent to Eq. (4) rather
than a single first order model (Eq. (1)). As suggested above this
could be because part of the initial organic micro-pollutant mass
(about 30%) may be readily available for degradation while the
remaining 70% of the mass is more difficult to degrade. Readily
degradable micro-pollutants are for instance dissolved in the aque-
ous phase or easily desorbed from the organic matter (adsorbed to
the outer particle surfaces). In this case the microbial degradation
rate is limiting removal. More difficult accessible micro-pollutants
are more strongly adsorbed to the organic matter and adsorbed to
internal particle surfaces. In this case desorption and diffusion out
of the organic particles is limiting removal.
All three kinetic parameters K1, K1a and K1b were dependent on
temperature (T). Fig. 4 shows K1–T relationships for three selected
compounds (TDCP, HHCB and TBEP). In all three cases K1 exhibit a
strong dependency on temperature. This is to be expected as
microbial activity normally is temperature dependent, however,
the K1–T relationships for the three compounds do not follow the
same pattern. It is seen that for each compound there is an opti-
mum temperature at which K1 is at a maximum. This behavior
was generally observed for K1, K1a and K1b across all compounds
(data not shown).
The K1–T, K1a–T, and K1b–T data for each individual compound
were fitted to Eq. (3) substituting K for r and using a and b as fitting
parameters. The fitting was done assuming zero degradation at
T = 0 °C. Best fits using Eq. (3) are shown as curves in Fig. 4 for three
selected chemicals. Optimal values of temperature (Topt) were
identified for each fitted curve as the value of T where the fitted
curve reaches the maximum value of K. Following this approach,
2011
values of Topt were determined as 29, 61, and 49 °C for TBEP, TDCP
and HHCB, respectively. Values of Topt for all 39 combinations of
compound and K are given in Table 2.
The cumulative probability distribution for Topt(K1) across all 13
compounds is shown in Fig. 5a. For comparison corresponding
optimal temperatures determined based on the relative reduction
in micro-pollutant concentrations over the experimental period
(Sadef et al., 2014b) are also shown in Fig. 5a. Both data sets thus
originate from the same measurements. A comparison between
the two datasets in Fig. 5a based on the Mann–Whitney test
showed that they were not significantly different at the 95% confi-
dence level. This supports that first order kinetics are adequate for
describing degradation of the organic compounds in compost
(although improved results may be achieved using a dual first
order model). Fig. 5a also shows that values of Topt(K1) has slightly
less variance (spread) than Topt values based on the relative con-
centration decrease (Sadef et al., 2014b). The reason for this is
likely that K1 is estimated based on four measurements while the
relative decrease in concentration is based only on 2 measure-
ments, thus being somewhat more uncertain.
Cumulative probability distributions for Topt across all 13 com-
pounds for both Topt(K1a) and Topt(K1b) are shown in Fig. 5b. Fig. 5b
suggests that there is a tendency for Topt,a to be smaller than Topt,b,
however, a test of (Topt,b – Topt,a) > 0 using bootstrapping revealed
that the tendency was not statistically significant at the 95% confi-
dence level. It indicates, however, that at early times lower tem-
peratures may be optimal for micro pollutant degradation while
later in the process higher temperatures are optimal. More data,
however, are needed to verify if this is the case. It is also evident
from Fig. 5b that values of Topt,a has more variance (spread) than
Topt,b. An explanation is that K1a is estimated based on fewer points
than K1b. Fig. 5a further indicates that a composting temperature
Fig. 4. First order degradation constant (K1), for three selected compounds, as a
function of temperature. Symbols indicate measured data while curves are best fit
Weibull probability density functions to the measurements.
2012 Y. Sadef et al. / Waste Management 34 (2014) 2007–2013
Fig. 5. Cumulative probability distribution functions for the optimal degradation temperature (Topt) for 13 organic micro-pollutants based on estimates of: (a) first order
degradation coefficient (K1) and relative concentration reduction (Sadef et al. 2014b), (b) dual term first order degradation coefficients (K1a and K1b), for 13 organic micro-
pollutants.
Table 3
p-Values for comparison of K1), K1a and K1b at different temperatures based on
bootstrapping (N = 13).
Hypothesis
K1 (50 °C) > K1 (18 °C)
K1 (50 °C) > K1 (30 °C)
K1 (50 °C) > K1 (37 °C)
K1 (50 °C) > K1 (70 °C)
K1a (18 °C) > K1a (30 °C)
K1a (18 °C) > K1a (37 °C)
K1a (18 °C) > K1a (50 °C)
K1a (18 °C) > K1a (70 °C)
K1b (50 °C) > K1b (18 °C)
K1b (50 °C) > K1b (30 °C)
K1b (50 °C) > K1b (37 °C)
K1b (50 °C) > K1b (70 °C)
(ns: N not significant).
range of 35–65 °C is adequate for micro-pollutant degradation for
almost all of the compounds investigated here. This temperature
range further corresponds to the thermophilic and mesophilic
phases of the composting process and therefore well matches nor-
mal composting conditions.
Average values of K1 and K1b across all chemicals at each tem-
perature are highest at 50 °C while average K1a is highest at 18 °C
(Table 2). It was therefore, tested (using bootstrapping) whether
the average values of K1 and K1a at 50 °C and K1a at 18 °C were in
fact significantly higher than at the other temperatures considered.
The results of the statistical test showed that K1 and K1b were are
significantly higher (at the 95% confidence level) at 50 °C than at
18 and 70 °C with p values of 0.005 and 0.008, respectively, while
there is generally no statistically significant differences in degrada-
tion constants between 30, 37 and 50 °C. For K1a the results were
mixed and no clear trend could be observed. An overview of the
results of the statistical tests is given in Table 3. Overall the data
indicate that compost process temperatures of 35–60 °C are opti-
mal for the degradation of organic micro pollutants.
5. Conclusions
Concentration measurements for 13 organic micro-pollutants
during composting at 5 different temperatures as a function of
time over a 52 day period (a total of 520 concentration measure-
ments) were used to assess micro-pollutant degradation kinetics
and their dependency on compost process temperature. In general
degradation was found to take place at all temperatures and for all
the compounds (except in a few cases). Two different kinetic mod-
els, a first order (using a single degradation coefficient) and a dual-
term first order (using two separate degradation coefficients)
model were used to describe micro-pollutant degradation. First
order kinetics provided a reasonable description of the degradation
process (average deviation 23%), however, the dual-term first order
expression was more accurate (average deviation 12%). The reason
P
was that degradation for several compounds proceeded very rap-
idly during the first 7 days and relatively slowly during the last
0.005
0.203ns
45 days of the degradation process. The results further showed that
0.08ns about 30% of the initial micro-pollutant mass present in the com-
0.008 post is degraded rapidly (during the first week) while the remain-
0.022 ing 70% is degraded at the slower rate. A reason for this behavior
0.264ns
could be that easily accessible micro-pollutant mass (less strongly
0.015
0.12ns adsorbed) is degraded initially while the more difficult to access
0.001 mass (for which desorption may be limiting) is primarily degraded
0.189ns at later times.
0.014 Degradation coefficients were strongly dependent on tempera-
0.007
ture and optimal temperatures (where degradation was at a max-
imum) were further strongly compound dependent. On average
across all 13 compounds, degradation was most rapid at about
50 °C and significantly higher than at 18 and 70 °C. In general the
results suggested that a temperature range between 35 and 60 °C
is appropriate for micro pollutant degradation. It is therefore
advantageous to maintain compost temperatures above 40 °C for
as long as possible to maximize micro-pollutant removal.
Acknowledgements
The authors acknowledge Komtek A/S for supplying the com-
post used in study. The lead author acknowledges the University
of the Punjab for awarding the scholarship for her PhD studies of
which this study is part. The authors also wish to thank the lab
technicians Helle Blendstrup (Aalborg University), Charlotte Dahl
Schiødt and Birgit Groth (Aarhus University) for their assistance
during the research.
References
Agnew, J.M., Leonard, J.J., 2003. Literature review – the physical properties of
compost. Compost Sci. Utilization 11 (3), 238–264.
Aparicio, I., Santos, J.L., Alonso, E., 2009. Limitation of the concentration of organic
pollutants in sewage sludge for agricultural purposes: a case study of in south
Spain. Waste Manage. 29, 1747–1753.
Caceres, R., Flotats, X., Marfa, O., 2006. Changes in the chemical and
physicochemical properties of the solidification of cattle slurry during
composting using different aeration strategies. Waste Manage. 26, 1081–1091.
Cheng, H.F., Kumar, M., Lin, J.G., 2008. Degradation kinetics of DEHP and organic
matter of sewage sludge during composting. Hazard. Mater. 154, 55–62.
Chen, X., Bester, K., 2009. Determination of organic micro pollutants such as
personal care products, plasticizers and flame retardants in sludge. Anal.
Bioanal. Chem. 395, 1877–1884.
Efron, B., Tibshirani, R., 1993. An Introduction to the Bootstrap. Chapman & Hall/
CRC, Boca Raton, FL, ISBN 0-412-04231-2.
Eiland, F., Leth, M., Klamer, M., Lind, M., Jensen, H.E.K., Iversen, J.J.L., 2001. C and N
turnover and lignocellulose degradation during composting of Miscanthus
straw and liquid pig manure. Compost Sci. Utilization 9, 186–196.
Epstein, E., 1997. The science of composting. Technomic Publishing Company,
Lancaster, pp. 171–194.
 Y. Sadef et al. / Waste Management 34 (2014) 2007–2013
Gavala, H.N., Alatriste-Mondragon, F., Iranpour, R., Ahring, B.K., 2003.
Biodegradation of phthalate esters during the mesophilic anaerobic digestion
of sludge. Chemosphere 52, 673–682.
Gonzalez, M.M., Martin, J., Camacho-Munoz, D., Santos, J.L., Aparicio, I., Alonso, E.,
2012. Degradation and environmental risk of surfactants after the application of
compost sludge to soil. Waste Manage. 32, 1324–1331.
Goyal, S., Dhull, S.K., Kapoor, K.K., 2005. Chemical and biological changes during
composting of different organic wastes and assessment of compost maturity.
Bioresour. Technol. 96, 1584–1591.
Haug, R.T., 1993. The Practical Handbook of Composting Engineering. Lewis
Publishers, USA.
Jones, F.W., Westmoreland, D.J., 1998. Degradation of nonylphenol ethoxylates
during the composting of sludges from wool scour effluents. Environ. Sci.
Technol. 32 (17), 2623–2627.
Madsen, P.L., Thyme, J.B., Henriksen, K., Moldrup, P., Roslev, P., 1999. Kinetics of di-
2-ethylhexyl phthalate mineralization in sludge-amended soil. Environ. Sci.
Technol. 33, 2601–2606.
Michel, F.C., Quensen, J., Reddy, C.A., 2001. Bioremediation of a PCB contaminated
soil via composting. Compost Sci. Utilization 9, 274–284.
Moeller, J., Reeh, U., 2003a. Degradation of nonylphenolethoxylates (NPE) in sewage
sludge and source separated municipal solid waste under bench-scale
composting conditions. Environ. Contam. Toxicol. 70, 248–254.
Moeller, J., Reeh, U., 2003b. Degradation of DEHP, PAHs and LAS in source separated
MSW and sewage sludge during composting. Compost Sci. Utilization 11, 370–
378.
Poulsen, T.G., Bester, K., 2010. Organic micro pollutant degradation in sewage
sludge during composting under thermophilic conditions. Environ. Sci. Technol.
44 (13), 5086–5091.
2013
Roslev, P., Madsen, P.L., Thyme, J.B., Henriksen, K., 1998. Degradation of phthalate
and di-2-ethylhexyl phthalate by indigenous and inoculated microorganisms in
sludge amended soil. Appl. Environ. Microbiol. 64, 4711–4719.
Sadef, Y., Poulsen, T.G., Bester, K., 2013. Optimized sampling strategy for
measurement of biomass properties during full scale composting. Waste
Manage. Res. 31 (8), 775–782.
Sadef, Y., Poulsen, T.G., Bester, K., 2014a. Quantifying measurement uncertainty in
full-scale compost piles using organic micro-pollutant concentrations. Waste
Manage. Res. 32 (5), 371–378.
Sadef, Y., Poulsen, T.G., Bester, K., 2014b. Impact of temperature on degradation of
organic micro-pollutants in compost. Submitted in Science of the Total
Environment (accepted for publication).
Sanz, E., Prats, D., Rodriguez, M., Camacho, A., 2006. Effect of temperature and
organic nutrients on the biodegradation of linear alkylbenzene sulfonate (LAS)
during the composting of anaerobically digested sludge from a wastewater
treatment plant. Waste Manage. 26, 1237–1245.
Semple, K.T., Reid, B.J., Fermor, T.R., 2001. Impact of composting strategies on the
treatment of soils contaminated with organic pollutants. Environ. Pollut. 112,
269–283.
Staples, C.A., Peterson, D.R., Parkerton, T.F., Adams, W.J., 1997. The environmental
fate of phthalate esters: a literature review. Chemosphere 35, 667–749.
Wilcox, R.R., 1997. Introduction to Robust Estimation and Hypothesis Testing.
Academic Press, San Diego, CA.
Yan, H., Ye, C.M., Yin, C.Q., 1995. Kinetics of phthalate ester biodegradation by
chlorella pyrenoidosa. Environ. Toxicol. Chem. 14, 931–938.