Вы находитесь на странице: 1из 10

Mesophilic methane fermentation of chicken manure at a wide range

of ammonia concentration: Stability, inhibition and recovery


Qigui Niu
a
, Wei Qiao
a,b
, Hong Qiang
c
, Toshimasa Hojo
a
, Yu-You Li
a,d,
a
Graduate School of Environmental Studies, Tohoku University, 6-6-06 Aza-Aoba, Aramaki, Aoba-ku, Sendai, Miyagi 980-8579, Japan
b
College of Chemical Engineering, China University of Petroleum, Beijing 102249, China
c
College of Resources and Environment, Northwest A&F University, Yangling 712100, China
d
Key Lab of Northwest Water Resource, Environment and Ecology, MOE, Xian University of Architecture and Technology, Xian 710055, PR China
h i g h l i g h t s
High solid methane fermentation of
chicken manur e was investigated.
Ammonia inhibition was the main
problem for stable operation.
The ammonia inhibitive sensitivities
for different reactions were
evaluated.
Methanogenesis, acidogenesis and
hydrolysis were simulated by models.
The seriously inhibited reactor was
successfully recovered by dilution.
g r a p h i c a l a b s t r a c t
a r t i c l e i n f o
Article history:
Received 3 February 2013
Received in revised form 9 March 2013
Accepted 11 March 2013
Available online 21 March 2013
Keywords:
Mesophilic CSTR
Process stability
Ammonia inhibition
Reactor recovery
Chicken manure
a b s t r a c t
A 12 L mesophilic CSTR of chicken manure fermentation was operated for 400 days to evaluate process
stability, inhibition occurrence and the recovery behavior suffering TAN concentrations from 2000 mg/
L to 16,000 mg/L. A biogas production of 0.350.4 L/gVS
in
and a COD conversion of 68% were achieved
when TAN concentration was lower than 5000 mg/L. Ammonia inhibition occurred due to the addition
of NH
4
HCO
3
to the substrate. The biogas and COD conversion decreased to 0.3 L/gVS
in
and 20% at TAN
10,000 mg/L and was totally suppressed at TAN 16,000 mg/L. Carbohydrate and protein conversion
decreased by 33% and 77% after inhibition. After extreme inhibition, the reactor was diluted and washed,
reducing TAN and FA to 4000 mg/L and 300 mg/L respectively, and the recovered biogas production was
0.5 L/gVS
in
. The extended Monod model manifested the different sensitivities of hydrolysis, acidogenes is
and methanogene sis to inhibition. VFA accumulation accompanied an increase in ammonia and exerted a
toxic on microorganism.
2013 Elsevier Ltd. All rights reserved.
1. Introduction
With the increase in intensive and mechanized poultry breed-
ing industries, large amounts of waste are being produced. Annu-
ally, about 13 million tons of chicken manure (CM) is generated
in Japan, which corresponds to 0.65 times the total food processing
waste (MAFF, 2008 ). Since the organic matter in CM is highly bio-
degradab le, methane fermentation is considered the best method
to minimize waste and recover bioenergy.
Efforts to produce much more biogas by treating CM with dry
methane fermentati on methods with high solids resulted in failure
(Gallert and Winter, 1997 ). Actually, CMhas a higher nitrogen con-
tent than cow manure, food waste, pig manure and waste active
sludge (Qiao et al., 2011 ). The excessive ammonia produced due
to the hydrolysis of this nitrogen material exerts a toxic and
0960-8524/$ - see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.biortech.2013.03.080

Corresponding author. Tel.: +81 227957464; fax: +81 227957465.


E-mail address: yyli@epli.civil.tohoku.ac.jp (Y.-Y. Li).
Bioresource Technology 137 (2013) 358367
Contents lists available at SciVerse ScienceDi rect
Biore source Tec hnology
j our nal homepage: www. el sevi er . com/ l ocat e/ bi or t ech
inhibitory effect on microbial activity and results in an unstable
process (Angelidaki and Ellegaard, 2003; Chen et al., 2008 ). At-
tempts to use TS concentratio ns as low as 0.53% were made to
alleviate ammonia inhibition. However, the substantial amount
of efuent produced meant the cost of following treatment was
higher (Bujoczek et al., 2000; Le Hyaric et al., 2011 ). While other
efforts, including co-digestion with low nitrogen waste and dilut-
ing CM with other slurry to low total solid (TS), have been made
to alleviate the ammonia inhibition effects, those methods were
not always effective. Previous researches reported that the co-
digestion of CM with cattle slurry and fruit/vegeta ble waste was
not able to counter high ammonia concentratio ns (Callaghan
et al., 2002; Duan et al., 2012; Hidaka et al., 2012 ). As yet, the rea-
sons for the difculties encounter ed with the sole high solid fer-
mentation of CM have not been revealed, let alone overcome,
according to the limited number of reports in the literature.
Hashimoto (1986) reported that the fermentation of swine
manure alone cannot be maintained stably. The concentrations of
ammonia tolerated in livestock digestion at 30004000 mg/L, were
reported (Angelidaki and Ahring, 1993 ). It was shown that between
8090% of methane production was suppressed when the total
ammonia nitrogen (TAN) concentratio n was 8000 mg/L (Krylova
et al., 1997 ). Hashimoto (1986) reported that both thermop hilic
and mesophilic processes are inhibited at a TAN of 2500 mg/L. Zee-
man et al. (1985) even reported an inhibition at TAN concentra-
tions of 1700 mg/L. Free ammonia (FA) was distinguished as the
cause of inhibition. It was shown that FA diffuses into the cell
membranes and sequentially ionizes the NH
3
to NH

4
leading to a
pH imbalance between inside and outside the bacterial cell. This
pH change affects both the transportation of the materials and
leads to lower enzyme activity (Kadam and Boone, 1996; Kayha-
nian, 1999 ). An FA concentration of 7001100 mg/L was shown
to be capable of triggering inhibition in many kinds of substrates
(Angelidaki and Ahring, 1993; Hansen et al., 1998; Sprott et al.,
1984). However, not all of the research supports this nding. One
example is that in pure culture in the lower pH range (6.57.0),
NH

4
ion inhibition effects were identied before the FA threshold
was reached (Jarrell et al., 1987; Sprott and Patel, 1986 ).
In very few articles, recovery strategie s have been proposed for
inhibited reactors. Bujoczek et al. (2000) suggested diluting the
substrate to low TS (0.53%) to reduce the nitrogen content. Niel-
sen and Angelidaki (2008) discussed the effects of lowering the
reactor temperat ure, adding chemicals and increasing the C/N ratio
in attempt to recover an inhibited reactor. A practical approach of
diluting the cow manure fermentation reactor with efuent, fresh
water and raw manure was shown to successfully shorten the
recovery period (Nielsen and Angelidaki, 2008 ).
Some research has been focused on the sole fermentation of CM
and a detailed analysis of the process steps has been made. Still,
more informat ion is required about the reactor operation and man-
agement in order to understand how to effectively recover the pro-
cess. In this work, a long-term methane fermentation process
feeding CM containing 10% TS using a continuously stirred tank
reactor (CSTR) was performed to investigate the stability, inhibi-
tion and recovery of the inhibited reactor. The methane conversio n
efciency, VFA accumulation following ammonia variation and
methanogenes is, acidogenesi s and hydrolysi s were investigated
both with and without inhibition.
2. Methods
2.1. CM properties
Original raw CM, with TS of 44.3%, was taken from farmland
was kept in the refrigerator. CM was grinded with tap water to
10 2% TS using a heavy duty laboratory blender. The CM slurry
was used as the continuous stirred tank reactor (CSTR) substrate.
The shredded CM, hereafter referred to as raw CM, was stored in
a 4 C substrate tank equipped with circulation of cooling water.
Ammonia stripped CM, with reduced nitrogen of raw CM from
the same farmland, was taken from HITACHI company. The TN of
striped CM was 3590 mg/L, much lower than raw CM with TN of
6450 mg/L. Ammonia stripped CM and raw CM were similarly in
TCOD but different in ammonia concentration. The total solid
(TS), volatile solid (VS), suspended solid (SS), volatile suspended
solid (VSS), total COD (TCOD), NH

4
N and total Nitration (TN) were
analyzed to provide substrate propertie s. The characteri stics of the
CM are given in Table 1.
2.2. CSTR operation procedure
A lab-scale CSTR with a working volume of 12 L (total 15 L) was
operated under the mesophilic (35 1 C) condition. The reactor
was warmed by water circulation agitated with a motor (200
300 rpm). A wet gas meter was used to measure the amount of dai-
ly biogas. The substrate tank was stirred (200300 rpm) to keep
the CM at a uniform state. The HRT was set at 30 days. An auto-
matic feed system with a peristaltic inuent pump and a timer
were used to feed substrate 12 times per day. Each feed was lower
than 1% of the reactor working volume. The seed sludge was taken
from the mesophilic anaerobic digestion of the municipal sewage
treatment plant.
2.3. Analytica l methods
The pH, alkalinity, COD, NH

4
N, TS, VS, SS, VSS, carbohyd rate
and protein were analyzed according to Japan Standard Methods
(JSWA, 1997 ). Carbohydra te was analyzed by phenolsulfuric acid
method; protein was analyzed by Lowry method. The VFA was
measure d by gas chromatograp hy (Agilent-6890). The record daily
biogas was calibrated to that under standard conditions (0 C;
1.013 bar). The biogas composition was measure d by a gas chro-
matograp h (Shimaszu GC-8A) equipped with a thermal conductiv-
ity detector. An elemental analyzer (Nario EL III CHNS) analyzed
the elemental composition of C, H, O, N, and S. FA was calculated
accordin g to the equilibrium Eq. (1) (Hansen et al., 1998 ):
Table 1
Characteristics of raw CM and ammonia stripped CM.
Constituent Unit Average (n = 6) SD ()
Raw CM TS (%) 11.2 0.53
VS (%) 8.27 0.83
SS (%) 10.1 0.11
VSS (%) 7.55 0.67
T-COD (mg/L) 103000 3200
TN (mg/L) 6450 810
TAN (mg/L) 3850 200
C % (dry) 35.2 0.45
H % (dry) 4.83 0.05
N % (dry) 5.44 0.24
O % (dry) 30.1 0.18
S % (dry) 0.84 0.10
Ammonia stripped CM TS (%) 8.93 0.14
VS (%) 6.13 0.20
SS (%) 0.79 0.05
VSS (%) 5.59 0.05
T-COD (mg/L) 94,400 8230
TN (mg/L) 3590 570
TAN (mg/L) 2500 100
T-COD: total COD.
TN: total nitrogen.
TAN: total ammonia nitrogen.
C: C element in dry CM.
Q. Niu et al. / Bioresource Technology 137 (2013) 358367 359
NH

4
$NH
3
H

NH
3
TAN
1
10
pH
10
0:9018
2729:92
Tk
_ _
_
_
_
_
_
_
1
1
The continuous experimental data of VFA was simulated in this
study using a modication of the Gompert z model, which can be
written as:
C
V
C
Vmax
exp exp
K
max
e C
A;0
C
A
C
Vmax
C
Vmax
_ _ _ _
2
whereC
A
: ammonia concentration (mg/L)C
V
: VFA accumulat ion
concentrat ion (mg/L)K
max
: the maximum VFA accumulat ion rate;
K
max
is dened as the tangent in the inection point C
A;0
: the TAN
concentrat ion of the initiate accumulat ion under inhibit ion; C
0
, is
dened as the x-axis intercept of this tangent (mg/L)C
Vmax
: the max-
imal value of VFA concentrat ion (mg/L)
2.4. Denition of conversion efciency
The conversion efciency of hydrolysi s, acidogenesi s and meth-
anogenesis were calculated based on the COD balance in Eqs. (3)
(5). TCOD
in
was the total inuent COD and SCOD
in
was the inuent
SCOD. SCOD was the centrifuged efuent supernatant COD. COD
CH4
was calculated based on the principle of 0:35m
3
CH
4
=kg COD
under standard conditions. COD
VFA
was the total VFA concentra-
tion calculated by oxygen demand of individual VFA. COD
VFAin
was the inuent of CODVFA. The conversion efciency was calcu-
lated once the steady state was reached.
Hydrolysis%
SCOD SCOD
in
COD
CH
4

TCOD
in
SCOD
in
3
Acidogensis%
COD
VFA
COD
VFAin
COD
CH
4

TCOD
in
COD
VFAin
4
Methanogenes is%
COD
CH
4
TCOD
in
5
3. Results and discussion
3.1. Long term performance under stable and inhibition
The experiments were divided into two stages. The rst stage
involved the stable (lasting 190 days) and inhibition (lasting
124 days) parts of the process. The other was the recovery experi-
ment, which lasted 46 days. All the experime ntal data were used to
establish an inhibition model, to calculate the inhibition threshold
and describe its occurrence.
In stage one, the experime nts were subdivided into 4 phases. In
phase I, from day 28 to day 75, ammonia stripped CM was used as
the feed to investiga te the performanc e of low nitrogen CM fer-
mentation. In phase II, from day 76 to day 195, raw CM was used
as the feed to analyze ammonia accumulation and inhibition ef-
fects. In phase III, (day196 to day 245) and phase IV (day 246 to
day 310), NH
4
HCO
3
was articially added with the substrate to
bring the TAN to an extremely high concentr ation. Fig. 1 shows
the TAN, FA, pH, alkalinity, biogas production and VFA variation
following operation time. The summary of these four phases is
shown in Table.2.
As shown in Fig. 1, the reactor became stable after 28 days, indi-
cating a successful startup. From day 28 to day 75 in phase I, the
reactor achieved a VS conversion of 61.1% and a biogas conversio n
of 0.350.40 L/gVS
in
with a methane content over 61%. The ammo-
nia stripped substrate resulted in low levels of ammonia , around
3000 mg/L, and no inhibition occurred.
In phase II, from day 76 to day 195, raw CM was fed into the
reactor. The TAN gradually climbed to approximately 6000 mg/L
and kept level. The VFA was around 2000 mg/L. The pH remained
stable and the alkalinity increased from 10,000 mg/L to
20,000 mg/L. The biogas production reduced to 0.30 L/g VS, which
was 80% that of phase I with low TN ammonia stripped CM. The
methane content slightly decreased to 5560%, which was consis-
tent with the slight increase in the VFA. The reactor did not suffer
from inhibition in this phase. It should be noted that Vanvelsen
(1979) reported that TAN concentr ations above 3000 mg/L exert
inhibition on reactor.
In phase III, from day 196 to day 245, with TAN concentratio ns
up to 7000 mg/L thanks to the addition of NH
4
HCO
3
in the sub-
strate
,
the biogas production decreased from 0.30 to 0.10 L/gVS,
correspond ing to a reduction of approximat ely 71% and 66% com-
pared to phase 1 and phase II, respectively.
In phase IV, after day 246, TAN were as high as 10,000 mg/L,
with NH
4
HCO
3
continua lly added both to the reactor and the sub-
strate, until they were over 15,000 mg/L at day 311. From day 260
to day 290, even with VFA levels higher than 16,000 mg/L, biogas
production remained stable. This can be attributed to the escaping
CO
2
which was come from HCO

3
with reaction of VFA. Biogas pro-
duction nally ceased when TAN concentratio ns extended beyond
16,000 mg/L, with the CH
4
composition as low as 14%. The ob-
served inhibition threshold of TAN 16,000 mg/L in this study was
higher than that previousl y reported both in batch experiments
and in full-scale experime nts (Garcia and Angenent, 2009; Nielsen
and Angelidaki, 2008 ).
The results of stage one indicate that ammonia initial inhibition
occurred at around 5000 mg/L after 150 days. Stable methane fer-
mentation was maintained in the case of raw CM with TAN levels
below 5000 mg/L. After prolonged exposure to higher ammonia,
the methane yield was seriously inhibited. The VFA reached
15,000 mg/L and ammonia concentr ation reached 10,000 mg/L.
The FA concentration varied from 900 mg/L to 1600 mg/L with
obvious increases in the VFA concentration from 1678 mg/L to
8747 mg/L. The biogas ceased with FA at around 3000 mg/L at a
TAN concentr ation of 15,000 mg/L.
3.2. Effects of ammonia on COD, carbohydrate and protein removal
efciency
Carbohyd rate, protein and lipids are the main organic materials
contributi ng to methane production. TCOD, carbohydrate and pro-
tein removal efciency were evaluated with and without inhibi-
tion. Fig. 2 illustrated the results. The carbohydrate removal
efciency was 60% at a TAN of 8000 mg/L and 25% at a TAN of
14,000 mg/L. The results demonstrated the possibility of obtaining
stable TCOD removal efciency over 60% and carbohyd rate re-
moval efciency of 80% with TAN concentratio ns less than
5000 mg/L.
The conversion efciency of protein was much lower than TCOD
and carbohydrate, as shown in Fig. 2. The average removal ef-
ciency was 50% with TAN levels lower than 2000 mg/L, and this de-
creased to 30% when TAN increased from 3000 mg/L to 5000 mg/L.
A rapid decrease in protein removal efciency occurred when TAN
levels exceeded 5000 mg/L. Gallert et al. (1998) reported a 50%
ammonia inhibition at a TAN of 3000 mg/L, and that protein re-
moval efciency almost ceased at a TAN of 6000 mg/L in peptone
mesophil ic fermentation.
The protein content in CM was 25% of the TS, which is main
source of ammonia. As Mata-Alvare z (2003) reported, protein deg-
radation can be expresse d by their rst-order kinetics (k, 0.25
0.8 d
1
), which are lower than for carbohydrates (k, 0.52.0 d
1
).
360 Q. Niu et al. / Bioresource Technology 137 (2013) 358367
TCOD, carbohydrate and protein removal efciency gradually de-
creased as TAN increased from 5000 mg/L to 15,000 mg/L. Under
inhibition, the protein conversio n was 80% lower than phase I,
while carbohyd rate conversion decrease d by 25% and COD conver-
sion decreased by 22%. These results indicated that the degradation
of protein was more sensitive to ammonia than the degradation of
carbohydrat e.
3.3. Relations hip between ammonia production and the stoichiometry
of CM
CM has a higher proportio n of biogradable organic matter
including lipids, carbohydrat es, protein and uric acid. Ammonia
is produced from the biological degradation of protein and urea.
The raw CM of elemental composition was C (35.16%), H (4.83%),
Fig. 1. The time course of TAN, pH, biogas production, gas composition and VFA accumulation.
Q. Niu et al. / Bioresource Technology 137 (2013) 358367 361
O (30.12%), S (0.84%), and N (5.44%). The stoichiomet ric
biochemical reaction formula, C
n
H
a
O
b
N
c
+ (n 0.25a 0.5b +
1.75c) H
2
O ?(0.5n + 0.125a 0.25b 0.375c) CH
4
+ (0.5n
0.125a + 0.25b 0.625c) CO
2
+ cNH
4
+ cHCO
3
is used to describe
the element transmission balance (Richards et al., 1991 ). The
CM methane fermentation was identied as stoichiomet ry
Eq. (6):
C
7:5
H
12:4
O
4:8
N 3:89H
2
O !3:7CH
4
2:8CO
2
NH

4
HCO

3
6
From the biochemi cal reaction formula, the degradat ion of 1 kg
VS of CM produces 0.74 m
3
biogas, 0.42 m
3
methane (56%) and
70.9 g of ammonia nitrogen. These theoretical results were consis-
tent with the experimental data at a steady stage with low VFA
production.
Fig. 2. Effects of TAN on the removal efciencies of T-COD, carbohydrate and protein.
Table 2
Average performance at different phases (steady stage, inhibition stage and recovery stage).
Parameters Unit Phase1 Phase 2 Phase3 Phase4 Recovery (with 5% fed)
3075 d 76200 d 201259 d 260314 d 360400 d
TS
in
% 8.93 10.5 11.0 11.5 5.7
VS
in
% 6.13 7.70 8.10 8.40 4.0
pH 8.10 8.15 8.25 8.00 ?8.20 7.80
TAN (mg/L) 2300 5000 8000 10,000 ?15,000 4000
VFA (mg/L) 500 1500 5000 15,000 1000
T-COD (mg/L) 80,000 10,000 11,000 11,000 55,000
Gas production rate (L/gVS
in
) 0.41 0.35 0.30 0.30 ?0.00 0.50
CH
4
(%) 62 60 59 50 ?08 60
CO
2
(%) 38 40 41 40 ?00 40
Removal efciency
T-COD (%) 68 60 55 20 ?0 70
Protein (%) 55 30 18 10 ?0 60
Carbohydrate (%) 85 75 60 40 ?00 86
COD conversion
CH
4
-
COD
/T-COD (%) 71 59 49 70
VFA
-COD
/T-COD (%) 1.2 1.5 12 6.9
S

-COD
/T-COD (%) 6.1 6.2 1.0 3.1
P
-COD
/T-COD (%) 21.0 34.1 37.3 20.1
CH
4
-COD: COD as CH
4
.
VFA-COD: COD as VFA.
S

-COD: SCOD- COD as VFA.


P-COD: particulate COD.
Gas production: L/g-VS
inuent
.
362 Q. Niu et al. / Bioresource Technology 137 (2013) 358367
3.4. Distinguishing the inhibition effects of ammonia and VFA
accumulation
The relationship between biogas production and VFA accumula-
tion is clear and is based on the metabolism of the microorgan isms
involved. At a steady stage, hydrolysis, acidogenes is and methano-
genesis maintain a metaboli c balance. DVFA (VFA
production
VFA
consumption
), used to express the VFA accumulation, was
analyzed as the indicator of the process deterioration. As has been
reported, when VFA accumulation concentratio n reached 6000 mg/
L at a TAN inhibition concentration of 5000 mg/L, the biogas was
only 25% that of the no inhibition stage (Garcia and Angenent,
2009).
The long-term experimental results were calculated and estab-
lish in an inhibition model. Biogas production was simulated using
the extended Monod Eq. (2) (Goncalves et al., 1991).
R R
0
1
I
I

_ _
n
7
where R: the biogas product ion ratio (L/g VS); R
0
: the maximu m R
(L/g VS); I: the TAN or FA concentrat ion (mg/L); I

: the TAN or FA
inhibition threshold (mg/L); n: constant.
In this model, inhibition effects were described by R
0
, n and I

.
Constant n determined the curve shape reecting acute or chronic
toxically inhibition pattern. In Eq. (7), nonlinear tting solves the
unknown parameters. In this study, Origin software V 8.1 was used
to conrm R
0
, n, and I by the GaussNewton procedure.
Fig. 3 illustrates the effect of TAN concentratio n on the biogas
production rate and VFA accumulation. The VFA concentr ation
was tted with R
2
= 0.95 based on Eq. (2), with an initial inhibition
concentratio n of C
A;0
= 5135 188 with K
max
= 2.85 0.21 and
C
Vmax
= 17,200 670, respectivel y. Biogas production was simu-
lated with R
2
= 0.85 based on Eq. (7), with a maximum
R
0
= 0.34 0.01, I = 14,500 470 and n = 0.12 0.03, respectivel y.
IC
10
and IC
50
of ammonia in the biogas were 8471 mg/L and
14400 mg/L, respectively.
At TAN concentrations of lower than 4000 mg/L, the VFA was
lower than 1000 mg/L when fed with both ammonia stripped CM
and raw CM. The VFA signicantly increased after TAN reached
5100 mg/L. TAN resulted in VFA accumulation when NH
4
HCO
3
was articially added. When TAN varied from 5000 mg/L to
10,000 mg/L, the VFA accumulati on sharply increased from
2000 mg/L to 15,000 mg/L. The VFA concentratio n became stable
at approximat ely 15,000 mg/L, when the TAN concentr ation had
increased from 10,000 mg/L to 16,000 mg/L.
Biogas production slightly decreased once the TAN concentra-
tion was over 4000 mg/L. Biogas production tended to be unstable
when the TAN concentratio n was higher than 6000 mg/L. When
TAN increased to 10,000 mg/L, the biogas production had de-
creased by1320%, with methane as low as 4550%. A sharp drop
in biogas production with low methane content was observed
when TAN reached 14,000 mg/L, indicating a strong inhibition ef-
fect on the reactor. Biogas production ceased at a TAN concentr a-
tion of 16,000 mg/L, while the simulation result indicated it
would stop at a TAN concentration of 14,500 470 mg/L.
However , biogas decreased gradually after TAN concentratio ns
went over 4000 mg/L until 14,000 mg/L and then decreased shar-
ply when TAN ranged from 14,000 mg/L to 16,000 mg/L. VFA accu-
mulation concentratio n was observed to sharply increase in the
range from 5000 mg/L to 10,000 mg/L and keep stable at
16,000 mg/L even when TAN continued to increase from
14,000 mg/L to 16,000 mg/L with the addition of NH
4
HCO
3
. A sim-
ilar result was found in earlier research (Garcia and Angenent,
2009). The interaction between ammonia , VFA and pH creates a
balance, which is especially notable when treating a high ammonia
concentr ation (Angelidaki and Ahring, 1993 ). This ammonia buffer-
ing system increases the pH stability even with signicant VFA
accumulation and forms a steady inhibition stage. In this steady
stage, not only methanogenet ic activity but also hydrolysis and
acidogenetic activity are suppressed (Nielsen and Angelidak i,
2008). This special steady stage increases the difculties of process
stability judgement and of making recovery strategie s. It has been
reported that acclimate d thermop hilic digestion enables methano-
gens to tolerate up to of 800013,000mg NH

4
N/L, depending on
acclimation condition s and system pH (Sung and Liu, 2003 ). Previ-
ous studies suggested that adapted fermentation of manure is
inhabited at a FA of 215 mg/L with 50% inhibition of CH
4
produc-
tion (El Hadj et al., 2009 ). However, a high tolerance of 3000 mg/
L and 4000 mg/L of FA has also been reported in the thermop hilic
process by (Angelidaki and Ahring, 1993 ).
Excess ammonia caused a primary inhibition phase with VFA
accumulation and biogas production gradually decreasing before
the next phase, characterized by signicant interactive inhibition
together with VFA. In this study, an FA concentratio n of
1000 mg/L started to primarily inhibit methanogenesi s. A sharp in-
crease in VFA accumulation occurred when the FA concentr ation
was 2000 mg/L, indicating strong inhibition on the process. Simul-
taneously , a sharp drop in biogas production was observed when
FA reached 2000 mg/L.
3.5. Simulatio n of hydrolysis , acidogenesis and methanogen esis
sensitiviti es on ammonia
Hydrolysi s, acidogenesis and methano genesis are the main
steps in the process. Most of the earlier research on ammonia
Fig. 3. Simulation of TAN (FA) effect on the biogas production and VFA accumulation.
Q. Niu et al. / Bioresource Technology 137 (2013) 358367 363
inhibition were based on batch experiments and the focus was
mainly on the methano genesis stage. Hydrolysis and acidogenesis
has attracted little attention. In this study, long-term reactor
experiment results were calculated for hydrolysi s, acidogenesis,
and methano genesis, which clearly illustrate the three steps of
methane fermentation process by Eq. (7). The tting curves of
the three stages are shown in Fig. 4. The paramete rs of R
0
, I

and
n are shown in Table 3.
Hydrolysis inhibition occurred at TAN concentratio ns of
5500 mg/L (IC
10
), 14,000 mg/L (IC
50
) and 15300 mg/L (IC
90
),
respectivly. Acidogen esis inhibition occurred at TAN concentra-
tions of 6500 mg/L (IC
10
), 14,000 mg/L (IC
50
) and 15,000 mg/L
(IC
90
), respectively . The TAN concentratio ns for IC
10,
IC
50
and IC
90
for methanogenes is were 4800 mg/L, 10,300 mg/L and
13,000 mg/L. The threshold of TAN on hydrolysis, acidogenesis,
and methanogenes is was over 14,000 mg/L based on the simula-
tion results. While methanogenes is was the most sensitive,
hydrolysi s was the most tolerant step and also the rst inhibited
step when exposed to high ammonia. The results also illustrated
that after increases of TAN, acidogenesis and hydrolysi s were
inhibited in sequence, nally resulting in the failure of the process.
Under the mesophil ic condition, the TAN threshold was 6000 mg/L
at a controlled initial pH of 7.5 and was 6500 mg/L under a pH of
8.5 (Lay et al., 1998 ). In this study, a higher IC
50
(TAN 10300 mg/
L) was obtained and a higher maximum threshold (excessing
13,000 mg/L). This can be attributed to the effects of the long-term
accumulati on of microorganism s.
It is also important to mention that even increases in FA at sta-
ble TAN concentr ations also resulted in VFA accumulation, indicat-
ing a synergistic relationship between FA and TAN. Many studies
have demonst rated the inhibitory effect of FA on methoge ns,
which was distinguished as the cause of inhibition, leading to a
supression in methane formation (Duan et al., 2012; Hawkins
et al., 2010 ). Duan et al. (2012) reported 4 degress of FA inhibition
Fig. 4. Simulation of TAN (FA) effect on COD conversion efciencies.
Table 3
Results of model tting and calculation of IC
10
, IC
50
and IC
90
.
Parameter n R
0
I

IC
10
(mg/L) IC
50
(mg/L) IC
90
(mg/L)
TAN Hydrolysis 0.29 0.05 82.55 2.18 15,620 350 5500 14,300 15,300
Acidogenesis 0.22 0.05 75.26 3.00 15,381 366 6500 14,600 15,000
Methanogenesis 0.46 0.13 71.76 2.26 13,359 994 4800 10,300 13,000
FA Hydrolysis 0.67 0.35 86.05 4.32 3530 1005 700 2200 3200
Acidogenesis 0.70 0.55 75.96 4.64 3637 1570 800 2300 3500
Methanogenesis 0.21 0.06 70.33 2.60 1897 31.70 650 1730 1800
364 Q. Niu et al. / Bioresource Technology 137 (2013) 358367
in mesophilic digestion of high solid sewage sludge: a slight inhi-
bition at 400 mg/L, moderate inhibition at 400600 mg/L, and sig-
nicant inhibition at 600800 mg/L, lastly causing fragility in the
process. In the mesophilic range, from temperat ure 25 C to
35 C, FA increased from100 mg/L to 250 mg/L elevating toxicity
effects on methanogen (Garcia and Angenent, 2009 ). On the other
hand, no inhibition occurred at a high FA concentratio n of
1100 mg/L when treating swine manure at 55 C (Hansen et al.,
1998). In the present study, obvious inhibition was noted at an
FA of around 1000 mg/L in all three steps. The FA thresholds on
hydrolysis and acidogenesi s were over 3000 mg/L, however , it
was below 2000 mg/L for methanogenes is. The IC
10
(650 mg/L),
IC
50
(1730 mg/L) and IC
90
(1800 mg/L) of FA for methanogenes is
were lower than those of acidogensis (800 mg/L, 2200 mg/L and
3200 mg/L) and hydrolysis (700 mg/L, 2300 mg/L and 3500 mg/L).
The thresholds of both TAN and FA were higher than those re-
ported in mesophilic condition. The total biogas process threshold
extended over TAN concentratio ns of 15,000 mg/L and FA concen-
trations of 3500 mg/L, respectively.
3.6. Performance of the recovery experiment
The substrate was stopped from day 314 to day 362 and the
recovery experiment was started to investiga te feasible recovery
methods . At day 314, half of the inhibited biomass was dis-
charged and injected with tap water to reduce the amount of
ammonia in the reactor. After dilution, the TAN and FA were
7000 mg/L and 1000 mg/L, respectively . The VFA decreased to
8000 mg/L. By diluting the reactor, the FA consternation was ex-
pected to fall below the initiated inhibition of 1000 mg/L while
the TAN concentratio n was still over the initial inhibition thresh-
old. However, within this period, no obvious biogas was pro-
duced, as shown in Fig. 5. Acetic acid was below 5000 mg/L
and then upgraded to 7000 mg/L at day 315 and varied from
50006000 mg/L. This indicated the ammonia still in reactor
was still toxic to the microorgan isms.
At day 326, tap water was used to wash the biomass in two
times of volume. The CSTR was stirred for 30 min, and then left
for one hour to allow a sediment to form before the top-level water
was siphoned off to retain the microorgan isms. The sludge had a
VFA below 4000 mg/L and a TAN below 4000 mg/L. The FA concen-
tration had dropped to 500 mg/L, which was lower than the initial
threshold. From day 345, biogas production recovered up to 0.4 L/
Ld
1
with a sharp decrease in VFA. The CH
4
composition reached
67% and then stabilized at 60% with a pH around 7.8. In this period
acetic acid decrease d, and then signicantly increased from
2000 mg/L to 5000 mg/L when biogas signicantly increased to
1.0 L/Ld
1
. Neverthel ess, recovery after the washing method was
Fig. 5. The time course of biogas production, VFA, and TAN content in the recovery experiment.
Q. Niu et al. / Bioresource Technology 137 (2013) 358367 365
more efcient than by dilution and controlling the FA concentra-
tion. While the increase in the propionate concentration was de-
layed, it slowly returned to a level of 2000 mg/L, indicating the
syntrophic bacteria degrading propionate were the slowest grow-
ing and the last microorgan isms to recover. This is consistent with
earlier ndings that suggested that propionate gave the best indi-
cation of when the entire process had stabilized (Nielsen and Ahr-
ing, 2006 ).
Many studies have demonst rated the inhibitory effects of free
ammonia (FA) on methanogens . It has been determined to be the
cause of inhibition since it leads to a supression of methane forma-
tion during the process. An FA concentration of 7001100 mg/L has
been shown to trigger inhibition under a wide of pH range for
many kinds of substrate (Angelidaki and Ahring, 1993; Hansen
et al., 1998 ). However, in several of the earlier studies based on
pure culture with lower pH (6.57.0) indicated possibility that
the total ammonia nitrogen (TAN) or NH

4
ion can reach toxic levels
before the FA concentration become toxic (Jarrell et al., 1987;
Sprott and Patel, 1986 ). Only a few of the previous studies have
indicated that ammonia inhibition involves the diffusion of free
ammonia (FA) through cell membranes and sequentially intra-cell
accumulation with parts of NH
3
converte d into NH

4
, causing a pH
difference between the intra-cell and extra-cell which change of
the intracellular pH, increase of maintenance energy requiremen t
and inhibiting species enzyme reactions (Kadam and Boone,
1996; Kayhanian, 1999 ). The ammonia buffering system increased
the pH stability even after signicant VFA accumulation and
formed a steady inhibition stage.
After successfu l recovery, the methane production was 0.5 m
3
/
kg VS, which appeased a much higher inhibition stage. The recov-
ery process illustrated that even after suffering serious ammonia
inhibition, the process still can be recovered with methanogens
gaining a high tolerance of ammonia . This study illustrated that
the initial inhibition of the FA concentratio n was more sensitive
for methanogen recovery than the TAN initiated inhibition. When
TAN was diluted to the initial inhibition value of lower than
4000 mg/L, the FA concentratio n was around 800 mg/L, and further
FA decrease d to 500 mg/L after 15 days as the biogas production
rate increased sharply. From day 363, a reactor feed with 5% TS
CM was used since the recovery conditions was good. And to ob-
tain a high biogas production of 0.7 L/gVS
in
. The VFA concentratio n
was maintained below 4000 mg/L. Following the recovery time,
VFA was consumed keeping a lower level than that in both the
steady stage and inhibition stage.
This recovery study illustrated that ammonia inhibition was a
synthetic action between TAN and FA, not only does FA affect
methane fermentation, but TAN concentratio ns do as well. Dilution
and the washing strategy were shown to be feasiblee methods for
the recovery of seriously inhibited processes, even when TAN is
over 16,000 mg/L.
4. Conclusion
(1) The feasibility of mesophilic fermentation with high solid of
CM was established at TAN lower than 5000 mg/L.
(2) VFA accumulation respondi ng to ammonia with TAN varies
from 5000 mg/L to 10,000 mg/L. The interactiv e inhibition
of VFA and TAN caused the process failure.
(3) Model simulations determined the different tolerance of
hydrolysi s, acidogenesi s and methanogenesi s. The ammonia
inhibition threshold was extended to 15,000 mg/L.
(4) Successful ly recovery was proved from a seriously inhibited
system with a TAN of 16,000 mg/L by dilution and then a
washing strategie s.
Acknowled gements
This work was supported by funding from the Japan Science and
Technolo gy Agency (JST). The authors would like to acknowledge
the Program of Scientic Cooperation with Hitachi Company to de-
velop this work.
References
Angelidaki, I., Ahring, B.K., 1993. Thermophilic anaerobic-digestion of livestock
waste the effect of ammonia. Applied Microbiology and Biotechnology 38 (4),
560564.
Angelidaki, I., Ellegaard, L., 2003. Codigestion of manure and organic wastes in
centralized biogas plants status and future trends. Applied Biochemistry and
Biotechnology 109 (13), 95105.
Bujoczek, G., Oleszkiewicz, J., Sparling, R., Cenkowski, S., 2000. High solid anaerobic
digestion of chicken manure. Journal of Agricultural Engineering Research 76
(1), 5160.
Callaghan, F.J., Wase, D.A.J., Thayanithy, K., Forster, C.F., 2002. Continuous co-
digestion of cattle slurry with fruit and vegetable wastes and chicken manure.
Biomass & Bioenergy 22 (1), 7177.
Chen, Y., Cheng, J.J., Creamer, K.S., 2008. Inhibition of anaerobic digestion process: a
review. Bioresource Technology 99 (10), 40444064.
Duan, N.N., Dong, B., Wu, B., Dai, X.H., 2012. High-solid anaerobic digestion of
sewage sludge under mesophilic conditions: feasibility study. Bioresource
Technology 104, 150156.
El Hadj, T.B., Astals, S., Gali, A., Mace, S., Mata-Alvarez, J., 2009. Ammonia inuence
in anaerobic digestion of OFMSW. Water Science and Technology 59 (6), 1153
1158.
Gallert, C., Winter, J., 1997. Mesophilic and thermophilic anaerobic digestion of
source-sorted organic wastes: effect of ammonia on glucose degradation and
methane production. Applied Microbiology and Biotechnology 48 (3), 405410.
Gallert, C., Bauer, S., Winter, J., 1998. Effect of ammonia on the anaerobic
degradation of protein by a mesophilic and thermophilic biowaste
population. Applied Microbiology and Biotechnology 50 (4), 495501.
Garcia, M.L., Angenent, L.T., 2009. Interaction between temperature and ammonia
in mesophilic digesters for animal waste treatment. Water Research 43 (9),
23732382.
Hansen, K.H., Angelidaki, I., Ahring, B.K., 1998. Anaerobic digestion of swine
manure: inhibition by ammonia. Water Research 32 (1), 512.
Hashimoto, A.G., 1986. Ammonia inhibition of methanogenesis from cattle wastes.
Agricultural Wastes 17 (4), 241261.
Hawkins, S., Robinson, K., Layton, A., Sayler, G., 2010. Limited impact of free
ammonia on Nitrobacter spp. Inhibition assessed by chemical and molecular
techniques. Bioresource Technology 101 (12), 45134519.
Hidaka, T., Arai, S., Okamoto, S., Uchida, T., 2012. Anaerobic co-digestion of sewage
sludge with shredded grass from public green spaces. Bioresource Technology.
Jarrell, K.F., Saulnier, M., Ley, A., 1987. Inhibition of methanogenesis in pure cultures
by ammonia, fatty-acids, and heavy-metals, and protection against heavy-metal
toxicity by sewage-sludge. Canadian Journal of Microbiology 33 (6), 551554.
JSWA, 1997. Japanese Standard Methods of the Examination of Wastewater. Japan
Sewage Works Association, Tokyo, Japan.
Kadam, P.C., Boone, D.R., 1996. Inuence of pH ammonia accumulation and toxicity
in halophilic, methylotrophic methanogens. Applied and Environmental
Microbiology 62 (12), 44864492.
Kayhanian, M., 1999. Ammonia inhibition in high-solids biogasication: an
overview and practical solutions. Environmental Technology 20 (4), 355365.
Krylova, N.I., Khabiboulline, R.E., Naumova, R.P., Nagel, M.A., 1997. The inuence of
ammonium and methods for removal during the anaerobic treatment of poultry
manure. Journal of Chemical Technology and Biotechnology 70 (1), 99105.
Lay, J.J., Li, Y.Y., Noike, T., 1998. The inuence of pH and ammonia concentration on
the methane production in high-solids digestion processes. Water Environment
Research 70 (5), 10751082.
Le Hyaric, R., Chardin, C., Benbelkacem, H., Bollon, J., Bayard, R., Escudie, R., Bufere,
P., 2011. Inuence of substrate concentration and moisture content on the
specic methanogenic activity of dry mesophilic municipal solid waste
digestate spiked with propionate. Bioresource Technology 102 (2), 822827.
Ministry of Agriculture, Forestry and Fisheries. Development and management of
livestock excreta situation http://www.maff.go.jp/j/chikusan/kankyo/taisaku/
t_mondai/02_kanri/index.html.
Mata-Alvarez, J., 2003. Biomethanization of the Organic Fraction of Municipal Solid
Wastes. IWA Publishing.
Nielsen, H.B., Ahring, B.K., 2006. Responses of the biogas process to pulses of oleate
in reactors treating mixtures of cattle and pig manure. Biotechnology and
Bioengineering 95 (1), 96105.
Nielsen, H.B., Angelidaki, I., 2008. Strategies for optimizing recovery of the biogas
process following ammonia inhibition. Bioresource Technology 99 (17), 7995
8001.
Qiao, W., Yan, X.Y., Ye, J.H., Sun, Y.F., Wang, W., Zhang, Z.Z., 2011. Evaluation of
biogas production from different biomass wastes with/without hydrothermal
pretreatment. Renewable Energy 36 (12), 33133318.
366 Q. Niu et al. / Bioresource Technology 137 (2013) 358367
Richards, B.K., Cummings, R.J., White, T.E., Jewell, W.J., 1991. Methods for kinetic
analysis of methane fermentation in high solids biomass digesters. Biomass &
Bioenergy 1 (2), 6573.
Sprott, G.D., Patel, G.B., 1986. Ammonia toxicity in pure cultures of methanogenic
bacteria. Systematic and Applied Microbiology 7 (23), 358363.
Sprott, G.D., Shaw, K.M., Jarrell, K.F., 1984. Ammoniapotassium exchange in
methanogenic bacteria. Journal of Biological Chemistry 259 (20), 26022608.
Sung, S.W., Liu, T., 2003. Ammonia inhibition on thermophilic anaerobic digestion.
Chemosphere 53 (1), 4352.
Vanvelsen, A.F.M., 1979. Adaptation of methanogenic sludge to high ammonia
nitrogen concentrations. Water Research 13 (10), 995999.
Zeeman, G., Wiegant, W.M., Kostertreffers, M.E., Lettinga, G., 1985. The inuence of
the total ammonia concentration on the thermophilic digestion of cow manure.
Agricultural Wastes 14 (1), 1935.
Q. Niu et al. / Bioresource Technology 137 (2013) 358367 367