You are on page 1of 4

ARTICLE IN PRESS

Bioresource Technology xxx (2008) xxxxxx

Contents lists available at ScienceDirect

Bioresource Technology
journal homepage: www.elsevier.com/locate/biortech

Short Communication

Simple and rapid methods to evaluate methane potential and biomass yield for
a range of mixed solid wastes
P. Shanmugam *, N.J. Horan
Public Health and Environmental Engineering, School of Civil Engineering, University of Leeds, Leeds, LS2 9JT, UK

a r t i c l e

i n f o

Article history:
Received 13 December 2007
Received in revised form 30 May 2008
Accepted 5 June 2008
Available online xxxx
Keywords:
Empirical formula
Biochemical/stoichiometric methane
potential
C/N ratio
Adenosine tri-phosphate
Biomass yield

a b s t r a c t
This paper describes rapid techniques to evaluate the methane potential and biomass yield of solid
wastes. A number of solid wastes were mixed to provide a range of C:N ratios. Empirical formulae were
calculated for each waste based on the results of chemical analysis and these formulae were used to estimate the COD equivalent and stoichiometric methane potential (SMP). The actual COD and biochemical
methane potential (BMP) were determined experimentally for each waste and for both parameters there
was a good agreement between the empirical and experimental values. The potential of adenosine triphosphate (ATP) to act as an indicator of biomass yield (mg VSS mg1 COD removed) was determined
during the anaerobic digestion process. The biomass yield determined from ATP analysis was in the range
0.010.25 mg VSS mg1 COD removed which corroborated well with previously reported studies. Empirical formula based SMP together with ATP measurement were shown to provide rapid methods to replace
or augment the traditional BMP and VSS measurements and are useful for evaluating the bioenergy and
biomass potential of solid wastes for anaerobic digestion.
2008 Elsevier Ltd. All rights reserved.

1. Introduction
The techniques used in UK for disposing of solid wastes are
changing rapidly and largely as a result of the Landll Directive,
biodegradable waste is now being diverted away from landll. This
Directive requires that by July 2010 biodegradable municipal
waste (BMW) going to landll shall be reduced to 75% of its
1995 tonnage, rising to 35% of this by 2020 (Defra, 2007). Diversion
is also being aided by increases in land ll tax, currently around
22/T. Other drivers that contribute to this change include climate
change and the potential scal rewards from the Renewable Obligation Certicates (ROCs). As a result a large tonnage of biodegradable waste stream previously sent to landll now requires dealing
with by other means. The opportunity to convert this waste to energy is obviously an attractive one and current technologies for
achieving energy from waste focus on biological and thermal options. The biological route is essentially the application of anaerobic digestion (AD) which is able to reduce the amount of volatile
solids in a waste feed by up to 70%, whilst at the same time generating a source of renewable energy as methane. The Department of
Environment, Food and Rural affairs (Defra, 2007), has estimated
that the potential annual market for anaerobic digestion is 400

* Corresponding author. Address: Environmental Engineering, Central Leather


Research Institute (CSIR), Adyar, Chennai 600 020, India. Tel.: +91 44 24911386.
E-mail addresses: pashanmugam@yahoo.com (P. Shanmugam), n.j.horan@
leeds.ac.uk (N.J. Horan).

million. In addition AD currently attracts 2 ROCs with a value of


30/MWh.
However, AD is a capital intensive project and in order to undertake a basic process economic feasibility study, it is important to
have an indication of the likely methane yield available from the
digester feedstock. A number of techniques are available to provide
this information including the biochemical methane potential
(BMP), dynamic respiration rate (DR4), and the COD test (Owen
et al., 1979; Environment Agency, 2005). In addition, a number of
indicative ratios can be used such as the soluble COD to volatile organic solids ratio. Of these, the BMP test is the most popular but, as
with any simple batch test that is intended to provide information
on the likely performance of a full-scale continuously operated
process, the results of the BMP require interpretation with caution.
In particular, the BMP test is conducted over a period of 28 days,
whereas conventional AD rarely operates at retention times above
15 days. Many researchers have searched for feasible alternatives
to the BMP test, but with limited success and consequently rather
than attempting to search a replacement, it is perhaps more appropriate to develop additional protocols to complement the results
from the BMP and aid in its interpretation (Shelton and Tiedje,
1984; Mendez and Lema, 1993). This paper intends to evaluate
two such complementary test procedures. The rst of these involves determination of the empirical formula of biomass that
can be used for stoichiometric product estimation (Erickson,
1978, 1980). The empirical formula developed for activated sludge
from aerobic biological treatment was C5H7NO2 (Rittmann and

0960-8524/$ - see front matter 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biortech.2008.06.027

Please cite this article in press as: Shanmugam, P., Horan, N.J., Simple and rapid methods to evaluate methane potential and biomass yield
for ..., Bioresour. Technol. (2008), doi:10.1016/j.biortech.2008.06.027

ARTICLE IN PRESS
2

P. Shanmugam, N.J. Horan / Bioresource Technology xxx (2008) xxxxxx

McCarty, 2001; Tchobanoglous et al., 2005) and routinely used as


basic design rule to calculate the oxygen requirement for BOD removal. In a similar way Rittmann and McCarty (2001) and Hansen
(2005) have proposed that stoichiometric methane potential (SMP)
determined from the assessment of empirical formula based organic solid wastes will provide information on the energy balance
for fermentation processes.
Even with the SMP test, information is not available as to the
behaviour of bacteria during the AD process and those conditions
which provide an optimum environment for their growth. The direct measurement of volatile suspended solids (VSS) cannot be applied in solid wastes digesters as it cannot distinguish between
bacteria and suspended organic solid wastes particles and so there
is a need for an alternative method of VSS measurement for solid
waste digesters to predict biokinetic design constants. The quantication of adenosine triphosphate (ATP) has been used by a number of researchers as an anaerobic cell viability assay method
(Chung and Neetheling, 1988; Yu et al., 2002; Kim et al., 2007).
The ratio of ATP to VSS (Hwang and Hansen, 1998) has also been
used to assess both cell viability and biomass yield, thus providing
important information on the behaviour of the organisms during
the AD process. Hence, this present study investigates the SMP
and ATP analysis as complementary tests to the BMP and VSS for
evaluating the suitability of a range of organic solid wastes as substrates for AD.

2. Methods
2.1. Characterisation of solid wastes
A number of solid waste sources were used in this study. Municipal solids waste (MSW) was collected in bulk and autoclaved at
130 C to permit storage without deterioration. Leather eshing
(LF) and primary chemically treated sludge (LS) were collected
from the tannery efuent treatment plant at Holmes Hall tannery
in Hull. The leather eshing was minced and homogenised with
a commercial blender to 6 mm diameter before feeding to the digester. Wastewater treatment sludge included primary sedimentation tank sludge (PST), mixed liquor from a sequencing batch
reactor (SBR), and surplus activated sludge (ASP) were collected
from the Knostrop wastewater treatment plant at Leeds. The characterisation of total solids (TS), and volatile solids (VS) were carried
out using Standard Methods (APHA, 1998). Elemental analysis of
carbon, hydrogen, nitrogen and sulphur was undertaken using
CHNS analysers Model Thermo Flash EA 1112 series. Samples were
oven dried at 103 C and then combusted at 1800 C in the CHNS
analyser in a steam of helium with measured amount of oxygen.
This produces N2, CO2, H2O, and SO2 which are then separated
and quantied by gas chromatography using a 5 mm diameter
steel (length of 2 m) packed column, helium carrier gas with a ow
rate of 40 mL/min, detected with a Propack model TCD.
2.2. Development of empirical formula for stoichiometric methane
potential (SMP)
From the calculated values of C, H, N, and O an empirical formula
was computed for each solid waste following the procedure developed by Rittmann and McCarty (2001) and Hansen (2005). This formula was used to determine the COD equivalent of which permitted
the calculation of the stoichiometric methane potential (SMP).
2.3. Biochemical methane potential (BMP) assessment
The BMP was determined in anaerobic batch reactors of 500 ml
capacity Duran bottles with hermetically sealed stoppers and con-

trolled gas opening valves. The effective volume maintained was


400 ml and the gas phase was 100 mL. Nutrient medium was prepared using a modication to the method described by Owen et al.
(1979) and added at 1 mL per litre of reactor volume. An internal
temperature of 35 C was maintained by incubating the reactors
in a temperature controlled mechanical shaker. Samples were
mixed at 140 rpm for a 15 minutes period followed by 15 minutes
with no shaking. The quantity of biogas produced was measured by
connecting the gas opening valve on the reactor to the inlet tube of
a hermetically sealed, water displacement aspirator bottles lled
with 5% NaOH to scrub CO2. A measuring cylinder at outlet of
the aspirator bottles collected displaced water which measures
CH4 at atmospheric temperatures and pressure (Nm L CH4. day1).
The initial and nal characterisation of 500 mL bottles was taken
for mass balance analysis and the serum bottle samples were used
for anaerobic process evaluation analysis. The BMP of seed sludge
was simultaneously carried out in control reactors, and subtracted
from the VS and gas yield with solid waste and the BMP yield of
solid waste was calculated as Nm L gm1 VS removed. The contents of the serum bottle were withdrawn periodically and centrifuged at 6000 rpm for 1 h and the supernatant was collected for
product analysis of total alkalinity (T.Alk), volatile fatty acids
(VFA), and ammoniacal nitrogen (NH3N).
2.4. ATPcell viability and biomass yield
Analysis of ATP was based on quantication of luminescence released from the reaction of luciferase with ATP (Chung and Neethling, 1989). Samples were diluted with 20 mM Tris-EDTA at pH
7.75, boiled for 30 min and equilibrated to room temperature.
Supernatant ATP was recovered by centrifugation at 6000 rpm
for 5 minutes. The luminescence reaction of sample ATP and luciferase was measured as a Relative Luminescence Unit (RLU) in a
Promega GloMaxTM 20/20 model luminometer, following addition
of back titre luciferase supplied by Promega. A calibration graph
was prepared from the RLU at varying ATP concentrations with
luciferase to quantify the sample ATP. The COD and VSS content
over time was calculated using 1.4 mg COD mg1 VS and 250 mg
VSS mg1 ATP (Hwang and Hansen, 1998). This was then used to
calculate biomass yield as mg VSS mg1 COD removed.

3. Results and discussion


3.1. Characterisation of a range of raw solid wastes
The important parameters for characterising the suitability of a
waste for AD are carbon to nitrogen ratio (C/N) and VS content, and
the microbiological performance parameters such as BMP and biomass yield are necessary for reactor design estimations. A range of
solid waste types with differing characteristic were selected for
this study (1) with the volatile organic solids ranging from 60 to
75% and the C:N 3.2 to 21.6. The LS had the highest organic content,
whereas the LF had the highest nitrogen content and thus a lower
C/N ratio. Primary sludge had the highest ash content and this is
likely to be a result of the presence of silt, clay and sand particles.
3.2. Empirical formulae, stoichiometric COD and methane potential
(SMP)
Based on the chemical analysis provided in Table 1, the empirical formula of each waste was calculated (Table 2). The calculated
empirical formula was then used to determine the theoretical COD
(CODe) of the waste. When expressed in terms of the measured volatile solids this gave an average value of 1.4 g CODe g1 VS, a value
that is typical of such wastes, for instance Han et al. (2005) re-

Please cite this article in press as: Shanmugam, P., Horan, N.J., Simple and rapid methods to evaluate methane potential and biomass yield
for ..., Bioresour. Technol. (2008), doi:10.1016/j.biortech.2008.06.027

ARTICLE IN PRESS
3

P. Shanmugam, N.J. Horan / Bioresource Technology xxx (2008) xxxxxx


Table 1
Characterisation of range of solid wastes
Parameters

Leather eshings

ASP sludge

SBR sludge

Leather sludge

PST sludge

MSW

1.
2.
3.
4.
5.
6.
7.
8.

21.9
81.3
35.5
9.4
11.2
25.3
0.6
18.7

0.9
70.3
37.6
5.7
7.0
20.1
0.5
29.7

0.4
62.0
28.3
5.1
4.2
24.5
0.8
38.0

21.5
81.5
39.9
6.4
3.5
31.6
1.3
18.5

3.8
60.3
28.7
4.6
2.4
24.7
0.5
39.7

41.7
63.0
27.5
3.8
1.3
30.5
0.4
37.0

Total dry solids (%)


Volatile solids (% of TS)
Carbon (% of TS)
Hydrogen (% of TS)
Nitrogen (% of TS)
Oxygen (% of TS)
Sulphur (% of TS)
Ash content (% of TS)

Table 2
Empirical formula, and anaerobic process evaluation for a range of solid wastes
Solid waste

Empirical formula

C:N

VFA (mg L1)

T.Alk. (mg L1)

NH3N (mg L1)

VSSe (g g1 COD eliminated)

Leather eshing
ASP sludge
SBR sludge
Chemical sludge
PST sludge
MSW

C4H11NO2
C6H11NO2
C8H17NO5
C13H26NO8
C14H26NO9
C25H41NO21

3.2
5.4
6.7
11.4
14.0
21.6

3540
4147
4266
2722
3490
4371

9163
6039
7997
7477
5703
6821

1296
241
466
999
305
414

0.09
0.14
0.14
0.01
0.12
0.25

ported values of 1.1 and Rittmann and McCarty (2001) observed


1.4. Good correlation was also observed between the CODe determined from the empirical equations and that calculated experimentally. Direct measurement of the COD of a solid waste is
often thought to produce erroneous results and thus the volatile
solids are generally used to quantify the organic strength of solid
wastes (Han et al. 2005; Lin et al., 1999). However, COD data is required for reactor design to estimate the biokinetic yield. Consequently COD estimation based on empirical formula provides an
attractive alternative to the experimental COD for design purposes.
3.3. Comparison of SMP and BMP test
The BMP was measured using the same range of solid wastes
under controlled conditions of nutrient addition, heating and stirring. The measured BMP ranged from 0.36 Nm L g1 VS for the
chemical sludge to 0.52 Nm L CH4 g1 VS for the ASP (Table. 2).
These values are typical of those observed by others, for instance
Lin et al. (1999) reported a BMP of 0.35 Nm L CH4 g1 of CODe
(equivalent to 0.318 Nm L g1 VS) for chemically pretreated ASP
sludge. A comparison of the SMP and BMP for these solid wastes
gave a good linear relationship (Table 2). The SMP of LF, PST,
SBR, and LS were 0.58, 0.53, 0.51, and 0.56 Nm L g1 VS whereas;
the BMP was recorded as 0.49, 0.38, 0.38 and 0.36 Nm L g1 VS,
respectively. The observed SMP and BMP of ASP sludge was 0.61
and 0.52 Nm L g1 VS, in line with the ndings of Gosset and Belser
(1982) and Wook and Hwang (2000) who reported the BMP of ASP
sludge as 0.360.54. In a similar way, the SMP and BMP of the
MSW was 0.40 and 0.36 Nm L CH4 g1 VS and a number of authors
report similar ndings, for example Davidsson et al. (2007) who reported that MSW has a BMP of 0.30.4 Nm L g1 VS. Thus the SMP,
which can be undertaken within 24 h, provides a more rapid alternative to the 28 days BMP test. It has the additional advantage that
it provides information as to the likely COD of the waste under test
and such information can be used for further biokinetic evaluation.
3.4. Evaluation of factors affecting full-scale digester performance
The SMP and CODe offer rapid procedures to estimate the potential methane yield from a waste. However, they give no indication as to how the waste might degrade in a full-scale digester.
Many other factors inuence actual digester performance. For in-

stance poor VS removal efciency and low biogas yield are often
associated with low C/N, high NH3N, high VFA and low digester
buffering capacity (Haandel and Lettinga, 1994; Anderson et al.,
2003). In addition, the concentration of ATP is known to indicate
both the cell viability and the metabolic status of a microorganism
and it has been used to predict biomass levels in anaerobic digestion (Chu et al., 2001; Yu et al., 2002; Kim et al., 2007). In order to
establish the potential of ATP as a simple performance monitor,
ATP concentration was measured over the duration of the BMP
test, together with a number of physical parameters that indicate
reduced performance. By monitoring a range of solid waste types
with differing C/N ratios it was hoped that a clear picture would
emerge as to the status of ATP, as an indicator of reactor stability
and performance. C/N ratio was selected as the variable parameter
because it determines both the NH3N and VFA concentration observed in the digester. Increasing NH3N helps to raise the pH,
whereas by contrast an acid pH is governed by VFAs which neu2
tralise HCO
and acetate ions. However ammonia can be
3 , CO3
toxic: Callaghan et al. (2000) and Salminen and Rintala (2002) have
reported the tolerance level of NH3 in anaerobic digesters acclimatised to treat high protein wastes as 11.6 g L1 and 6.0 g L1,
respectively.
Of the six solid wastes evaluated, the cumulative biogas yield
after 28 days was highest for MSW and the lowest for leather eshing, in line with the C:N ratios for these wastes (Fig. 1 and Table 2).
The maximum NH3N concentration of 1296 mg L1 was also recorded for the leather eshing but this is well below the concentration reported as inhibitory. Alkalinity was in the range of 3800
9700 mg L1 and the maximum alkalinity was coincident with high
NH3N levels (Table 2). The VFA was observed to peak at around 8
days for SBR, ASP and MSW wastes, whereas for LF it was as late as
20 days suggesting that a much longer digester retention time
would be required for this waste (Fig. 2). The late decrease in
VFA was also coincided with high alkalinity. ATP was measured
at a concentration in the range of 2.018.43 mg L1 and similar values have been reported by others, for instance Hwang and Hansen
(1998), Yu et al. (2002), and Chen (2004). The ATP level for LF, ASP,
SBR, LS, PST, and MSW were 2.01, 4.21, 3.22, 2.12, 3.22, and 8.43,
respectively, and observed as linear with respect to cumulative
methane yield measured during the BMP test (Table 2), thus demonstrating the potential value of this parameter. Standard BMP test
can be used to estimate biokinetics design constants. Future work

Please cite this article in press as: Shanmugam, P., Horan, N.J., Simple and rapid methods to evaluate methane potential and biomass yield
for ..., Bioresour. Technol. (2008), doi:10.1016/j.biortech.2008.06.027

ARTICLE IN PRESS
4

P. Shanmugam, N.J. Horan / Bioresource Technology xxx (2008) xxxxxx

References

3000

Biogas (N mL)

2500
SBR
ASP
PST
Chemical Slud
Leather Fl
MSW

2000
1500
1000
500
0
0

10

20
30
Days Elapsed

40

Fig. 1. Cumulative biogas evolution with elapsed time.

9000

SBR Sludge
ASP sludge
PST Sludge
Chemical Sludge
MSW
LF

8000

VFA (mg/L)

7000
6000
5000
4000
3000
2000
1000
0
0

10
20
30
Elapsed time (Days)

40

Fig. 2. The volatile fatty acids (VFA) levels at elapsed time during the BMP test.

will therefore focus on correlating the level of ATP attained in the


BMP test with the biokinetic constants of a continuously fed system treating the same waste type.
4. Conclusion
For all the solid waste examined, there was a good corroboration observed between SMP and BMP, CODe and experimental
COD. The wastes with higher C:N ratio produced methane more
rapidly and also provides a guide to the likely operating retention
time of a continuously fed digester. The ATP measured was indicative of both volatile solids removal, amount of CH4 and VSS (bacteria) produced from the wastes. Thus, the combined application of
empirical formulae, SMP and ATP offers the potential to augment
BMP and VSS tests to estimate both methane yields and biokinetic
design constants for solid wastes anaerobic digesters.
Acknowledgements
The authors would like to acknowledge the help of Mr. Nick
Waudby, Holme Hall Tannery, Hull (UK) for providing samples of
leather eshing and chemically treated sludge.

Anderson, G.K., Sallis, P., Uyanik, S., 2003. Anaerobic treatment processes. In: Mara,
D.D., Horan, N.J. (Eds.), Handbook of Water and Wastewater Microbiology.
Academic Press, New York, pp. 391423.
APHA, 1998. Standard Methods for the Examination of Water and Waste Water,
20th ed. American Public Health association, American Water Works
Association Water Pollution control Federation, Washington, DC.
Chen, H., 2004. ATP content and biomass activity in sequential anaerobic/aerobic
reactors. J. Zhejiang Univ. Sci. 5 (6), 727732.
Chung, Y.C., Neetheling, J.B., 1988. ATP as a measure of anaerobic sludge digester
activity. J. Water Pollut. Control Fed. 60 (1), 107112.
Chu, C.P., Lee, D.J., Chang, B.V., Liao, C.S., 2001. Using ATP bioluminescence
technique for monitoring microbial activity in sludge. Biotechnol. Bioeng. 75
(4), 469474.
Callaghan, F.J., Wase, D.A.J., Thayanithy, K., Forster, C.F., 2000. Continuous codigestion of cattle slurry with fruit and vegetable wastes and chicken manure.
Biomass Bioenerg. 27, 7177.
Chung, Y.C., Neethling, J.B., 1989. Microbial activity measurement for anaerobic
sludge digestion. J. Water Pollut. Control Fed. 61 (3), 343349.
Defra, 2007. Department of Environment Food Rural affair and Agriculture. UK
Waste Strategy.
Davidsson, A., Christopher Gruvberger, B., Thomas, H., Christensen, C.T., Lund
Hansen, C., Jes, L., Cour, J.C., 2007. Methane yield in sourcesorted organic
fraction of municipal solid waste. Waste Manage. 27, 406414.
Environment Agency, 2005. Guidance on Monitoring MBT and Other Pretreatment
Processes for the Landll Allowances Schemes. UK (England and Wales)
Erickson, L.E., 1978. Application of Mass and energy balance regularities to product
formation. Biotechnol. Bioeng. 20, 15951621.
Erickson, L.E., 1980. Biomass elemental composition and energy content.
Biotechnol. Bioeng. 21, 451456.
Gosset, J., Belser, R., 1982. Anaerobic digestion of waste activated sludge. J. Env. Eng.
108, 11011120.
Haandel, A.C., Lettinga, G., 1994. Anaerobic Sewage Treatment A Practical Guide
for Regions with Hot Climate. John Wiley and Sons.
Han, S.K., Kim, S.H., Kim, H.W., Shin, H.S., 2005. Pilot-scale two stage process: a
combination of acidogenesis and methanogenesis. Water Sci. Technol. 52 (12),
131138.
Hansen, T.L., 2005. Quantication of environmental effects from anaerobic
treatment of sourcesorted organic house hold wastes, Ph.D. Thesis
September 2005. Institute of Environment & Resources, Technical University
of Denmark, Lyngby.
Hwang, S., Hansen, C.L., 1998. Evaluating a correlation between volatile suspended
solid and adenosine 50 tri phosphate levels in anaerobic treatment of high
organic suspended solid wastes water. Bioresour. Technol. 63, 243250.
Kim, K., Lee, C., Shin, S.G., Hwang, S., 2007. Correlation of microbial mass with ATP
and DNA concentrations in acidogenesis of whey permeates. Biodegradation.
Online.
Lin, J.G., Ma, Y.S., Chao, A.C., Huang, C.L., 1999. BMP tests on chemically pretreated
sludge. Bioresour. Technol. 68, 187192.
Mendez, M.S., Lema, J.M., 1993. Methanogenic and non-methanogenic activity tests.
Theoretical basis and experimental set up. Water Res. 27 (8), 13611376.
Owen, W.F., Stuckey, D.C., Healy, J.B., Young, L.Y., McCarty, P.L., 1979. Bioassay for
monitoring biochemical methane potential and anaerobic toxicity. Water Res.
13, 485492.
Rittmann, B.E., McCarty, P.L., 2001. Environmental Biotechnology Principles and
Applications, second ed. McGraw Hill Publications.
Salminen, E.A., Rintala, J.A., 2002. Semi-continuous anaerobic digestion of solid
poultry slaughterhouse waste: effects of hydraulic retention time and loading.
Water Res. 36, 31753182.
Shelton, D.R., Tiedje, J., 1984. General method for determining anaerobic
biodegradation potential. Appl. Environ. Microb. 47 (4), 850857.
Tchobanoglous, G., Burton, F.L., David Stensel, H.D., 2005. Waste water engineering
treatment and reuse. In: Metcalf, Eddy (Eds.), Civil Engineering Series, fourth ed.
McGraw Hill. International editions..
Wook, K.W., Hwang, S., 2000. Mechanical pre-treatment of waste activated sludge
for anaerobic digestion process. Water Res. 34 (8), 23622368.
Yu, Y., Hansen, C.L., Hwang, S., 2002. Biokinetics in acidogenesis of highly suspended
organic wastewater by adenosine 50 triphosphate analysis. Biotechnol. Bioeng.
78 (2), 147156.

Please cite this article in press as: Shanmugam, P., Horan, N.J., Simple and rapid methods to evaluate methane potential and biomass yield
for ..., Bioresour. Technol. (2008), doi:10.1016/j.biortech.2008.06.027