ORIGINAL PAPER
C. Gallert á J. Winter
Mesophilic and thermophilic anaerobic digestion
of source-sorted organic wastes: effect of ammonia
on glucose degradation and methane production
Received: 26 March 1997 / Received revision: 13 May 1997 / Accepted: 19 May 1997
Abstract The wet organic fraction of household wastes
was digested anaerobically at 37 °C and 55 °C. At both
temperatures the volatile solids loading was increased
from 1 g l
A1
day
A1
to 9.65 g l
A1
day
A1
, by reducing the
nominal hydraulic retention time from 93 days to
19 days. The volatile solids removal in the reactors at
both temperatures for the same loading rates was in a
similar range and was still 65% at 19 days hydraulic
retention time. Although more biogas was produced in
the thermophilic reactor, the energy conservation in
methane was slightly lower, because of a lower methane
content, compared to the biogas of the mesophilic re-
actor. The slightly lower amount of energy conserved in
the methane of the thermophilic digester was presum-
ably balanced by the hydrogen that escaped into the gas
phase and thus was no longer available for methano-
genesis. In the thermophilic process, 1.4 g/l ammonia

protein undegraded.
Introduction
Anaerobic fermentation of the organic fraction of wastes
is a suitable method to reduce both volume and mass for
deposition in sanitary land®lls. Approximately 33% of
household wastes (annual amount 70±140 kg/in-
habitant; Lu
È
bben 1994) can be separated as bioorganic
wet waste. The residual semi- dry fraction (67%) may be
disposed of either in sanitary land®lls or be incinerated,
leading to an increased caloric energy output per mass
unit, because of its decreased water content. Biological
treatment of the wet organic fraction can be performed
either in the presence of air (composting) or in the ab-
sence of air (biogas fermentation, sometimes in accu-
rately, called anaerobic composting). In 1996,
composting plants with a total capacity of 2.4 ´ 10
6
tonnes bioorganic wastes were in operation in Germany,
whereas the total capacity for anaerobic fermentation
was only 0.19 ´ 10
6
tonnes (Biehler and Nuding 1996).
Anaerobic fermentation signi®cantly reduces the
total mass of wastes, generates solid or liquid fertilizer
and yields energy. It can be maintained at psychrophilic
(12±16 °C, e.g. in land®lls, swamps or sediments),
mesophilic (35±37 °C, e.g. in the rumen and in anaerobic
digester) and thermophilic condition s (55±60 °C; e.g. in

during anaerobic degradation of urea or proteins.
Livestock manure from pork and poultry contains
about 4 g N/l (Angelidaki and Ahring 1991), that from
cattle about 1.5 g N/l (Angelidaki and Ahring 1993). In
the organic fraction of household waste the organic
nitrogen that was released as ammonia during anaero-
bic fermentation amounted to 2.15 g N/l (Jager et al.
1989). Free ammonia may be inhibitory for anaerobic
fermentation and may be toxic for methanogenic
bacteria (Angelidaki and Ahring 1993). Inhibitory
NH
3
concentrations under mesophilic conditions of
80±150 mg N/l at a pH of 7.5 have been reported
(Koster and Lettinga 1984; Braun et al. 1981). Under
thermophilic conditions at a pH of 7.2±7.3, for ace-
ticlastic methanogens, inhibitory concentrations of free
ammonia of 3.5 g/l NH
4
-N/l (250 mg/l NH
3
-N) were
found and, for hydrogenolytic methanogens, 7 g/l
NH
4
-N/l (500 mg/l NH
3
-N) (Angelidaki and Ahring
1993; Borja et al. 1996).
In this study, we compare the fermentation of the wet

) the feed had to be added at 12-h
intervals twice a day. Biogas production was measured continu-
ously with a gas meter (Ritter, Hanau) and the pH was controlled
on-line with a pH electrode (WTW, Weilheim) inserted into the
recirculation line.
Initially the reactors were ®lled with anaerobic sludge from the
municipal sewage treatment plant of Regensburg (Germany). After
an adaptation time of 2 weeks at 37 °C (mesophilic reactor) and
55 °C (thermophilic reactor), 240-ml aliquots were replaced by a
1:1 mixture of fresh sewage sludge and biowaste slurry. Gas pro-
duction, pH and fatty acids were monitored. When the gas pro-
duction ceased and the pH reached 7.4, 240 ml digester content was
again replaced, this time by only biowaste slurry. When the gas
production ceased again, a fed-batch feeding once per day to
maintain an initial t
HR
of 93 days was started. After process sta-
bilization, the t
HR
in both reactors was reduced stepwise and
concomitantly the loading was increased, as indicated later.
Batch digestion experiments in serum bottles
The euent of the mesophilic and the thermophilic biowaste re-
actor was centrifuged in a WKF-50-K centrifuge (Gesellschaft fu
È
r
elektrophysikalischen Apparatebau, Brandau) at 830 g for 10 min
to separate most of the solid waste particles (about 80%) from the
suspended biomass. The supernatant fraction contained the sus-
pended bacteria and the small-particulate sludge ¯ocs, including

Parameter Minimum Maximum Mixture
for expt.
COD
total
(g/l) 118 215 176
COD
diss
(g/l) 46 93 71
COD
diss
/COD
tot
(%) 38 42 40
DOC (g/l) 16 37 25
Total solids (g/l) 139 200 184
Volatile solids (g/l) 121 178 172
TKN (mg/l) 1650 4040 3107
NH

4
(mg/l) 175 349 260
pH 4 5.5 5.3
406
oxidation of the organic material in the homogenized sample with
potassium dichromate/sulphuric acid according to Wolf and
Nordmann (1977). The dissolved proportion of the COD and the
dissolved organic carbon were measured after ®ltration (pore size
0.45 lm) of samples. The dissolved organic carbon was analysed by
infrared spectroscopy with a Tocor 2 carbon analyser (Maihak,
Hamburg).


4
. The concentration of
free ammonia (NH
3
-N) was calculated according to Anthonisen et
al. (1976):
NH
3
-N 
NH

4
-N Â 10
pH
u
b
au
w
 10
pH
u
b
au
w
 e
6344a273 
where N concentrations are in mg/l and T is in °C.
Glucose was measured photometrically according to Lever
(1972) as p-hydroxybenzoic acid hydrazide at 410 nm. The rela-

the city of Regensburg was acclimated to biowaste di-
gestion at 37 °Cand55°C beginning with an apparent
t
HR
of 93-days as mentioned. In the mesophilic reactor
the COD removal eciency during stepwise reduction of
the t
HR
to 80, 60 and 50 days was around 85% (Fig. 1a).
When t
HR
was further reduced to 19 days, equivalent to
a space loading of 9.65 g COD l
A1
day
A1
, the COD-re-
moval eciency decreased to 64%. In the thermophilic
reactor the COD-removal eciency was 95% at a t
HR
of
93 days. During stepwise reduction of the retention time
to 19 days, equivalent to an increase of the space loading
to 9.65 g volatile solids l
A1
day
A1
, the COD removal
decreased steadily from 95% to 67% (Fig. 1b). The
biogas production increased from 0.3 l l

63% or 67% of the COD was degraded
at 37 °Cor55°C and the volatile solids reduction was
64% and 65% respectively. Although the gas production
per litre of reactor volume and per day was apparently
slightly higher in the thermophilic reactor, because of a
reduced solubility of CO
2
at the higher temperature, in
total little more methane was produced in the mesophilic
reactor. This was due to a methane content of the biogas
of 67% in the reactor run at 37 °C but only 59% in the
Fig. 1a,b Eciency of mesophilic (a) and thermophilic (b) biowaste
fermentation at increasing loading rates. A portion of the c ontent of
each fermenter was manually replaced once per day with fresh
biowaste to maintain the respective loading rate/hydraulic retention
time (t
HR
). Biogas production and chemical O
2
demand (COD)
removal were determined. r Loading rate ( VS volatile solids),
n COD removal, Ð hydraulic retention time, m biogas production
407
reactor run at 55 °C (Table 2). The total energy release
by gases in both reactors may have been identical,
however, since the gas of the thermophilic reactor con-
tained some hydrogen (not quanti®ed). The ammonia
content of the euent of the thermophilic reactor was
notably higher, indicating that apparently more protein
was degraded at 55 °C than at 37 °C.

(50 mmol/l, pH 7.8, containing 4.5 mmol/l glucose) and
supplemented with 0±35 g/l NH
4
Cl or NaCl. Glucose
degradation rates were determined for the time span
necessary for 80%±90% degradation of the initial
amount and are shown in Fig. 3a. With increasing
concentration of ammonia, glucose degradation was
slightly more inhibited at 55 °C than at 37 °C. Th e K
i50
for ammonia-N was 3.7 g/l at 37 °C (=0.28 g NH
3
-N)
and 3.4 g/l at 55 °C (=0.68 g NH
3
-N). The K
i50
was
similarly determined for methane production (Fig. 3b).
A 50% inhibition was seen at 3.0 g/l ammonia-N
(=0.22 g/l NH
3
-N) in the mesophilic reactor and at
3.5 g ammonia-N (=0.69 g/l NH
3
-N) in the thermo-
philic reactor (Table 3). In the presence of 5 g/l NaCl,
glucose degradation and methane production were not
signi®cantly in¯uenced, whereas in the presence of 15 g/l
NaCl or more, methane was no longer produced, pre-

day
A1
) 9.65 9.65
(g VS l
A1
day
A1
) 9.4 9.4
(g TKN-N l
A1
day
A1
) 0.17 0.17
Hydraulic retention time (days) 19 19
COD
diss
in euent
a
(%) 13.2 30.7
COD reduction (%) 63 67
VS reduction (%) 64 65
Space productivity (l biogas l
A1
day
A1
) 5.3 5.6
Methane content (%) 67 59
Methane (l l reactor volume
A1
day

;
Gessler and Keller 1995). The volatile solids removal
(64%±65%) was also higher in the laboratory reactors
under mesophilic and thermophilic fermentation condi-
tions, as compared to many full-scale plants, where it
was around 55% (Ku
È
bler 1994; Gessler and Keller
1995). Rivard et al. (1995) reported a very high COD
removal eciency of 77% for municipal solid wastes and
tuna processing wastes at mesophilic fermentation tem-
peratures for loading rates up to 14 g volatile solids
l
A1
day
A1
. Kayhanian (1995 ) observed a degradation of
83% of the volatile solids fraction of wet waste in a
fermentation with a high solids content at 30 days t
HR
.
A highly ecient volatile solids removal may lead to
a reduced viscosity and a better separation of solids. The
supernatant of the euent of our mesophilic and ther-
mophilic biowaste digester contained 13.2% (37 °C) or
30.7% (55 °C) of the non-degraded COD, respectively,
after sedimentation of large solids. Whereas the super-
natant of the mesophilic reactor euent contained
mainly soluble organic components, the supernatant of
the thermophilic reactor euent in addition contained a

day
A1
(thermophilic fermentation)
were reported by Mackie and Bryant (1995).
Although a little less biogas was produced in our
mesophilic biowaste reactor, becau se of the higher
methane content in the biogas the total amount of
methane per litre of reactor volume and per day was
higher than that in the gas of the thermophilic reactor
(Table 2). A similar observation was reported by Mackie
and Bryant (1995) for the digestion of cattle manure.
The advantage of thermophilic digestion was mainly
that the euent was rendered hygienic by inactivation of
bacteria (Ku
È
bler 1994) or viruses (Lund et al. 1996).
The speci®c gas production per gram of COD or per
gram of volatile solids degraded was high in both of our
reactors. This was the result of a high lipid and fat
content of the biowaste input material, caused by the
addition of a batch of spoiled butter. Consequently the
speci®c methane productivity was also high: 0.59 l/g at
37 °C and 0.54 l/g volatile solids degraded at 55 °C
(calculated from data of Table 2). For carbohydrates a
theoretical methane yield of 0.35 l/g would be expected.
High speci®c methane yields of 0.4±0.59 l/g were also
reported for thermophilic household solid waste diges-
tion, by Rintala and Ahring (1994).
One obvious result of thermophilic biowaste diges-
tion was the higher yield of ammonia: 1.4 g/l at 55 °C

NH

4
aNH
3
-N
(g/l)
K
i50
NH
3
-N
(g/l)
Methane production, 37 °C 3.0 0.22
Methane production, 55 °C 3.5 0.69
Glucose degradation, 37 °C 3.7 0.28
Glucose degradation, 55 °C 3.4 0.68
409
methanogens. Adaptation from 4 g to 6 g N/l required a
time span of 6 months. We observed a 50% inhibition at
37 °C and 55 °C by 3±3.7 g ammonia/l for glucose fer-
mentation and methanogenesis as well. The inhibitory
eect was presumably due to the free permeability of
NH
3
through the cell membrane. At the actual pH of 7.6
in the reactors, 0.22±0.28 g free NH
3
/l caused a 50%
inhibition of mesophilic glucose fermentation and

Angelidaki I, Ahring BK (1993) Thermophilic anaerobic digestion
of livestock waste: the eect of ammonia. Appl Microbiol
Biotechnol 38: 560±564
Angelidaki I, Ahring BK (1994) Anaerobic thermophilic digestion
of manure at dierent ammonia loads: eect of temperature.
Water Res 28: 727±731
Anthonisen AC, Loehr RC, Prakasam TBS, Srinath EG (1976)
Inhibition of nitri®cation by ammonia and nitrous acid. J
Water Pollut Control Fed 48: 835±849
Barlaz MA, Schaefer DM, Ham RK (1989) Bacterial population
development and chemical characteristics of refuse decompo-
sition in a simulated sanitary land®ll. Appl Environ Microbiol
55: 55±65
Biehler MJ, Nuding R (1996) Vergleich verschiedener Verfahren
zur Verwertung von Bio- und Gru
È
nabfa
È
llen auf kommunaler
Ebene. Abfallwirtschaftsjournal 9: 28±38
Borja R, Sanchez E, Weiland P (1996) In¯uence of ammonia
concentration on thermophilic anaerobic digestion of cattle
manure in up¯ow anaerobic sludge blanket (UASB) reactors.
Process Biochem 31: 477±484
Braun R, Huber P, Meyrath J (1981) Ammonia toxicity in liquid
piggery manure digestion. Biotechnol Lett 3: 159±164
Buswell AM, Mueller HF (1952) Mechanics of methane fermen-
tation. J Ind Eng Chem 44: 550
Deutsche Einheitsverfahren zur Wasser-, Abwasser- und
Schlammuntersuchung (1979±1983), VCH Verlagsgesellschaft

Lu
È
bben S (1994) Organische Abfa
È
lle in einer Groûstadt-Mengen
und Verwertungsmo
È
glichkeiten. In: H Ma
È
rkl, R Stegmann (eds)
Anaerobe Behandlung von festen und ¯u
È
ssigen Ru
È
cksta
È
nden.
Dechema Monogr 130: 207±230
Lund B, Jensen VF, Have P, Ahring BK (1996) Inactivation of
virus during anaerobic digestion of manure in laboratory scale
biogas reactors. Antonie van Leeuwenhoek 69: 25±31
Mackie RI, Bryant MP (1995) Anaerobic digestion of cattle waste
at mesophilic and thermophilic temperatures. Appl Microbiol
Biotechnol 43: 346±350
Merck E (1974) Die chemische Untersuchung von Wassser. 13.
Au¯age, Darmstadt
Richards BK, Cummings RJ, White TE, Jewell WJ (1991) Methods
for kinetic analysis of methane fermentation in high solids
biomass digestors. Biomass Bioenergy 1: 65±73
Rintala JA, Ahring BK (1994) Thermophilic anaerobic digestion of