Accepted Manuscript

Bio-hydrogen and bio-methane production from food waste in a two-stage anaerobic

digestion process with digestate recirculation

Dalal E. Algapani, Wei Qiao, Marina Ricci, Davide Bianchi, Simon M. Wandera,
Fabrizio Adani, Renjie Dong

PII: S0960-1481(18)31029-2
DOI: 10.1016/j.renene.2018.08.079
Reference: RENE 10506

To appear in: Renewable Energy

Received Date: 23 February 2018

Revised Date: 4 July 2018
Accepted Date: 23 August 2018

Please cite this article as: Algapani DE, Qiao W, Ricci M, Bianchi D, M. Wandera S, Adani F, Dong R,
Bio-hydrogen and bio-methane production from food waste in a two-stage anaerobic digestion process
with digestate recirculation, Renewable Energy (2018), doi: 10.1016/j.renene.2018.08.079.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

1 Bio-hydrogen and bio-methane production from food waste in a

2 two-stage anaerobic digestion process with digestate recirculation

3 Dalal E. Algapani a,b, Wei Qiao a,c,*, Marina Ricci d, Davide Bianchi d, Simon M.

PT
4 Wandera a, Fabrizio Adani d, Renjie Dong a,c

RI
5 a College of Engineering, China Agricultural University, Beijing 100083, China

6 b College of Agricultural Technology and Fish Science, Al-Neelain University, Khartoum, Sudan

SC
7 c State R&D Center for Efficient Production and Comprehensive Utilization of Biobased Gaseous Fuels, Energy

8 Authority, National Development, and Reform Committee (BGFuels), Beijing 100083, China

U
9 d Gruppo Ricicla – DiSAA – University of Milan, Via Celoria 2, 20133 Milano, Italy
AN
10 Abstract
11 Bio-H2 and bio-CH4 production from food waste in a two-stage temperature phased system were
M

12 investigated to determine the effects of digestate recirculation on energy efficiency and process

13 stability. Different recirculation ratios (RR), i.e. 0.3, 0.5, and 1.0, were tested. Maximum H2
D

14 production of 3 L-H2 L-1d-1 and yield of 135 L-H2 kg-1VSin were achieved for an RR of 0.3 at
TE

15 HRT 5 d and OLR of 18 kg-VS m-³d-1. The RR of 0.3 was also the best for producing CH4 and

16 gave results of 2.9 L-CH4 L-1 d-1, i.e. 510 L-CH4 kg-1VSin at HRT 9 d and OLR of 5.7 kg-VS m-³
EP

17 d-1. The energy recovered from the recirculation process increased the H2 production by 8% and
C

18 decreased the CH4 production by 3%; the total energy production did not change. Digestate
AC

19 recirculation in comparison with a no-recirculation system reduced the need for alkali addition to

20 maintain pH in the H2-reactor by 54%.

21 Keywords:

22 Anaerobic digestion; Bio-hydrogen and bio-methane; Digestate recirculation; Food waste; Two

23 stage

1
 ACCEPTED MANUSCRIPT

24 1. Introduction

25 Global energy demand is expecting to increase by about one third for the period 2013-2040 [1].

26 Currently about 88% of this demand is provided by fossil fuels which are expected to become

27 depleted within a century. As consequence of that, the production of renewable energy is being

PT
28 encouraged in order to save fossil fuels [2].

RI
29 Biomass can be used to produce renewable energy via anaerobic digestion, producing H2 and/or

30 CH4. Food waste (FW) is an organic waste collected from numerous sources i.e. cafeterias and

SC
31 restaurants, domestic and commercial kitchens and food processing plants. According to the

32 Food and Agriculture Organization (FAO), approximately 1.3 billion tonnes of food are wasted

33
U
annually [3]. Within Asian countries, the FWs are expected to increase in the next 25 years due
AN
34 to economic and population growth [4], leading to critical problems such as health risks,

environmental pollution and scarcity of dumping land, if they are not correctly managed. On the
M

35

36 other hand, FW can be used as biomass in anaerobic digestion (AD) processes to produce
D

37 bioenergy and biofertilizers [5,6]. Managing the food waste by AD reduces the environmental
TE

38 impact [7], as AD of FW has been considered the most advantageous technology to maximise

39 energy recovery [8].

EP

40 A two-stage AD process is more advantageous than a single stage process because the optimum

41 conditions for bacterial activity during each stage can be maintained. For example, optimal pH
C

42 for H2 production in the first stage is of 5-6, which is lower than that of pH 7-8 required for CH4
AC

43 production in the second stage [9,10]. In a single stage and two stage process methane was

44 obtained at 0.78 and 0.81 L g-1 VSin from FW respectively [11]. Also, it was reported that waste

45 activate sludge (WAS) from single and two-stage biogas yields were of 0.34 and 0.47 L g-1VSin

46 [12]. In particular, combining the advantages of rapid thermophilic (H2 producing) and stable

2
 ACCEPTED MANUSCRIPT

47 mesophilic anaerobic digestion has drawn wide attention in the two-stage system because of the

48 benefits which can be achieved [11].

49 One of the key factors required for H2 production is the maintenance of the pH at an optimal

50 value (pH of 5.5) through the addition of alkaline substances [13]. Commonly the volume of

PT
51 alkalies used to maintain the pH affects operational conditions, i.e. OLR, HRT, the acidity of the

RI
52 substrate and the entire process. As reported in earlier work pH can be maintained by adding

53 alkalies at the average quantities of 18, 28, 26, and 44 mL Lreactor-1 for HRT of 15, 10, 5 and 3

SC
54 days [14]. NaOH was used as it was the best alkali compared with respect [e.g.Ca(OH)2] [15].

55 Although alkali improves anaerobic digestion performance, the use of alkali has a cost and

56
U
moreover, increasing the alkali dose reduces biogas production because of bacterial inhibition
AN
57 due to the high concentrations of Na+ or Ca2+ [16]. However, digestate recirculation can be

proposed to maintain the pH in the H2 producing stage. It was reported that recirculation in
M

58

59 between the two stages was able to control pH and to reduce chemical costs, especially when the
D

60 process was treating large amounts of waste [17].

TE

61 The recirculation of digestate is effective for two reasons: firstly, the renovation of H2 producing

62 bacteria and secondly, the reduction of alkali required to keep the pH optimal [13, 18]. On the
EP

63 other hand, digestate recirculation can affect the whole process as well. As reported in a previous

64 pilot study [19], digestate recirculation did not affect H2 production using wastewater, whereas it
C

65 was also found that recirculation decreased H2 yield from food waste in a two-stage process [20].
AC

66 Considering these different results, the two-stage recirculation approach in the AD process still

67 needed more efforts to clarify the principles and practices of hydrogen and methane production

68 [21].

3
 ACCEPTED MANUSCRIPT

69 The present study aimed to achieve superior bio-hydrogen and bio-methane production by

70 developing a two-stage recirculation system for food waste, taking into consideration process

71 stability, the optimization of the digestate recirculation ratio, and the reduction of the alkaline

72 additions to maintain optimal conditions during the process.

PT
73 2. Material and Method

RI
74 2.1 Characteristics of food waste and inoculum

75 The FW in this study was collected every 2-3 weeks from the lunch residues of a canteen at the

SC
76 China Agricultural University (Beijing, China). Undesirable components such as bones, waste

77 paper and plastics were removed manually. Thereafter, the FW was ground for 5 min by using a

78
U
blender (Joyoung, JYLC012, China). The FW slurry was then sealed in plastic bags and stored in
AN
79 a refrigerator at 4˚C before being used as feed for the anaerobic digestion process. The

inoculums used for the AD process were taken from a full-scale thermophilic (55°C) AD-plant
M

80

81 treating maize straw and from a full-scale mesophilic (35°C) AD-plant treating municipal
D

82 sewage sludge. The inoculums were put without any pre-treatment into the first reactor (H1)
TE

83 producing hydrogen. The characteristics of the FW slurry and inoculums sampled at the two full-

84 scale plants are listed in Table 1.

EP

85 Table 1 Characteristics of food waste and inoculums

Feeding food waste Inoculums 55 °C Inoculums 35 °C

Parameter
C

Average SD ± n Average SD ± n Average SD ± n

TS (%) 10.5 1.5 19 4.5 0.8 4 10.4 0.8 4
AC

VS (%) 9.1 1.3 19 3 0.5 4 6.4 1.2 4

SS (%) 4.6 0.7 19 3.6 0.3 4 3.7 0.6 4
VSS (%) 4.5 0.7 19 2.3 0.2 4 2.4 0.7 4
-1
TCOD (gL ) 152 47.5 19 53.7 8 4 84.2 7 4
SCOD (gL-1) 84 15.6 19 20 2.5 4 14.8 3.2 4
NH4+-N ( mgL-1) 40 8.2 10 165 26.8 4 1267 22.8 4
-1
TVFA (mgL ) 1,882 262 14 / / / / / /
86 Note: ‘/’ means data not available

4
 ACCEPTED MANUSCRIPT

87 2.2 Continuous two-stage anaerobic systems

88 The entire process was run using two laboratory-scale continuously stirred tank reactors (CSTR).

89 In the first CSTR (H1), hydrogen fermentation took place while in the second CSTR reactor (M1)

90 methane fermentation occurred. During the first stage, the H1 reactor was maintained at 55°C

PT
91 while the second stage M1 was maintained at 35°C. The reactor temperatures were maintained

RI
92 using a water bath (HH-60, China) which provides an accuracy of ± 0.1°C. Working volumes for

93 H1 and M1 reactors were of 2 L and 4.5 L, respectively. The slurries inside the reactors were

SC
94 mixed by using timer-controlled stirrers operating automatically (50-90 rpm for 30 min every 2

95 h). The gas volumes were recorded with gas meters (LmL-1). The continuous two-stage

96
U
recirculation system is shown in the Supplementary Materials.
AN
97 When the digestate was not recirculated, the H1 was fed with FW diluted with tap water to

obtain a total solid (TS) content of about 8 to 9% (w/w). When the digestate of M1 was
M

98

99 recirculated into H1, the FW was not diluted by water. The effluent from H1 was manually
D

100 added into the M1. During the whole experiments, the fluctuation temperature in M1 was not
TE

101 observed, and the pH was not controlled. Based on M1 effluent volume, the recirculation rate

102 (RR) (equation 1) was calculated as the ratio of the recycle volume of methanogens effluent to
EP

103 the volume of the hydrogen reactor influent [22,23], was of 0.3, 0.5, and 1.0.
ொ
104 ܴܴ = ொ೐ ---------------------- (1)
C

೑
AC

105 Where the Qe is the recycle volume of methanogenic effluent (L), Qf is the volume of new

106 influent (FW) (L).

107 During the whole process, an HRT of 5 days and an OLR ranging from 15-18 kgVS m-3 d-1 were

108 provided for H1, while the M1 used an HRT of 9 days and an OLR of 5.7 kg-VS m-3 d-1. The pH

109 in H1 was set to a value of around 5.5 after each feeding, by the addition of NaOH 1.25 mol L-1

5
 ACCEPTED MANUSCRIPT

110 concentrations. A pH correction was provided for the methanogenic stage, as the values were

111 always around the optimal (Table 2).

112 According to the previous study [24] optimal H2 and CH4 production by a two-stage approach

113 without a recirculation system was achieved by adopting the following process parameters: HRT

PT
114 of 5 and 15 days and OLR of 18 and 3.2 kgVS m-3 d-1.

RI
115 Table 2 pH, NaOH, and total alkalinity during different recirculation period
NaOH addition Total Alkalinity
pH
Recirculation (mL Lreactor-1) (g-CaCO3 L-1)
Reactor

SC
Ratio Min Max Average Min Max Average Min Max Average
0 4.5 5.7 5.2 8 56 26 / / /
0.3 4.9 5.4 5.1 7 17 12 / / /

U
H1
0.5 4.8 5.5 5.4 5 25 16 / / /
AN
1.0 5.1 5.7 5.4 5 30 13 / / /
0 7.5 7.9 7.8 / / / 6.9 12.6 9.5
0.3 7.3 7.9 7.6 / / / 4.7 10.6 8.3
M

M1
0.5 7.5 7.9 7.8 / / / 17.8 19.2 18.4
1.0 7.2 7.5 7.3 / / / 13.1 15.7 14.3
Note: ‘/’ means data not available
D

116

117 2.4 Chemical analyses

TE

118 Total solids (TS), volatile solids (VS), COD, soluble COD, and NH4+-N were measured
EP

119 following the standard methods of the American Public Health Association [25]. Volatile fatty

120 acids (VFAs) were analyzed using the Shimadzu GC-2010 Plus (Japan) equipped with a flame
C

121 ionization detector and capillary column type Rtx-wax (30 m ×0.25 mm × 0.25 µm). The column
AC

122 had an initial temperature of 60˚C (5-min hold time), ramp 10˚C min-1 to 140˚C (2-min hold

123 time), and ramp 20˚C min-1 to 230 C (2-min hold time). Temperatures of the injector and

124 detector were of 230˚C and 250˚C, respectively. Methane and carbon dioxide were measured

125 with GC SP-2100 (Beifen-Ruili Analytical Instrument, China) equipped with a stainless steel

126 column (10 m x 2 mm), working with a hydrogen pressure 0.4 Mpa and a speed of 60 mL min-1.

6
 ACCEPTED MANUSCRIPT

127 The temperatures of the inlet, column oven, and TCD were of 130˚C, 130˚C, and 116˚C,

128 respectively. Hydrogen in biogas was detected with Shimadzu GC-8AIT 220 V (Japan) (Column

129 10 m x 2 mm) and adopting a nitrogen pressure speed of 30 mL min-1. The temperatures of the

130 TCD and column oven were of 120˚C and 60˚C, respectively.

PT
131 The calculation of energy balance based on COD was provided in the Supplementary Material.

RI
132 3. Results and discussion

133 3.1 The hydrogen yield with and without recirculation

SC
134 Hydrogen production was investigated by considering both different recirculation ratios (RR)

135 (Table 3) and the different additions of NaOH, of which the amount depended on the features of

136
U
the effluent from M1 to maintain the pH at around 5.5.
AN
137 In this two-stage recirculation system, the hydrogen reactor was working for almost 120 days at a

constant OLR. At the 40th day, H2 yield was of 135 L kg-1VSin when adopting the RR of 0.3, 113
M

138

139 L kg-1VSin for RR of 0.5 at 80th day and 43.8 L kg-1VSin for RR 1.0 at 120th day. H2 contents in
D

140 the biogas produced were of 60.8, 56.4, and 42.2% (v/v), the CO2 contents were of 32.8, 30.6,
TE

141 and 27.9% (v/v), and the CH4 content were of 1.5, 1.9, and 1.1% (v/v), at the respective

142 recirculation ratios (RR) of 0.3, 0.5, and 1.0. Yeshanew et al. (2016) obtained from a two-stage
EP

143 process with recirculation system treating artificial FW, a volumetric hydrogen production of

144 0.178 L L-1d-1 and hydrogen yield of 89 L kg-1VSin with a hydrogen content in the biogas ranging
C

145 between 32% and 42% (v/v) for RR between 0.24–0.48 [22]. A hydrogen production of 0.8 L L-
AC

1 -1
146 d and an hydrogen yield of 51 L kg-1VSin was reported with a hydrogen content in the biogas

147 of 37% (v/v) by adopting a RR of 1.0, from recirculation of bio-waste [26].

148 The effect of the increase of RR on H2 production is illustrated in Figure 1. Increasing the RR

149 decreased the H2 production and yield, i.e. for RR of 0.3, 0.5, and 1.0 the H2 production was of 3,

7
 ACCEPTED MANUSCRIPT

150 2.5, and 1.5 L L-1d-1, respectively. As previously studied, the methanogenic community in the

151 second stage methane reactor was dominated by hydrogenotrophic methanogens, i.e

152 Methanobacterium and Methanoculleus [24]. In the current study, it was supposed that

153 increasing the recirculation ratio possibly induced the consumption of hydrogen by the

PT
154 hydrogenotrophic methanogens. The hydrogen production in first reactor therefore decreased. As

RI
155 reported, adopting a recirculation system of RR 2.9 and using FW, an H2 production of 117 L kg-
1
156 VSin, with 47.7% of H2 (v/v); when the RR increased to 4.3 obtained, the H2 yield and content

SC
157 in the biogas decrease to 1.0 L kg-1VSin and 0.5% (v/v), respectively [27].

U
Hydrogen Fermentation (55 ˚C)
recycle
ratio Without 0.3 0.5 1.0
12
AN
Gas production (l/l/d)

(a)
9

3
M

0 0 20 40 60 80 100 120 140 160 180

90 (b)
CO
H

Gas Composition (%)

60
TE

30

0 0 20 40 60 80 100 120 140 160 180

65
(c)
EP

NaOH pH
NaOH (ml/Lreactor)

6 45
pH

25
C

4 0 20 40 60 80 100 120 140 160 180

5
AC

(d) Acetate
12000
Butyrate
TVFA
VFA (mg/l)

6000

0
1 21 41 61 81 101 121 141 161 181

Duration (d)
158
159 Figure1 Gas, pH and VFA for hydrogen fermentation (1st stage)

8
 ACCEPTED MANUSCRIPT

160 To provide higher performance in H2 production the RR needs to be evaluated because the
161 recirculation was found to inhibit the hydrogen-producing microbial population [10, 20, 28]. In
162 this work, the H2 production without recirculation was of 2 L L-1d-1 that was 125 L kg-1VSin with
163 H2 content in the gas of 64.3% (v/v) and no detectable CH4. In the current study, the hydrogen
164 reactor was operated for more than 200 days and the pH was maintained around 5.5. The

PT
165 methanogenesis activities in the hydrogen reactor would be almost inactivated. The methane in
166 the first reactor was therefore not observed. This result can be compared with that from literature,

RI
167 i.e. 106 L kg-1VSin without recirculation [29] and it was higher than that of 65 L kg-1VSin
168 obtained by a two-stage co-digestion of FW and waste activated sludge [30].

SC
169 Again, data reported in this work for trial performance without recirculation (Table 3) were in

170 agreement with a previous study [31] which mentioned that NH4+-N was involved in process

171
U
performance and its concentration was inversely proportional to the hydrogen production and
AN
172 yield. In a hydrogen reactor, proteins are hydrolyzed slowly to soluble amino acids [29]. But a

low NH4+-N concentration in a hydrogen reactor in a two-stage process was previously reported
M

173

174 [26]. It was reported that syntrophic amino acid degradation occurs correctly in reactors where
D

175 hydrogenotrophic methanogens and protein-degrading bacteria are at an ideal balance and under
TE

176 the optimal environmental conditions that guarantee their cooperation [32]. Therefore, the

177 absence of the methanogens group in the H2 reactor possibly decreased the protein hydrolysis.
EP

178 As reported in a previous study [33], carbohydrates are firstly converted to VFAs at low pH 5-6,

179 then proteins are converted to VFAs at neutral pH and finally, VFAs produced are converted into
C

180 CH4 and CO2 in a methane reactor.

AC

181 Table 3. Operational conditions of biohydrogen and biomethane reactors with and without recirculation

Bio-hydrogen reactor (55 °C) Bio-methane reactors (35 °C)

Parameter Units without without
with recirculation with recirculation
recirculation recirculation
Recirculation
/ / 0.3 0.5 1 / 0.3 0.5 1
ratio
Duration days 70 38 40 39 70 38 39 40
TSin % 8.5±1.3 10±1.1 8±0.7 9±0.8 6.2±0.4 6.7±0.8 6.8±0.3 6.6±0.5

9
 ACCEPTED MANUSCRIPT

TSout % 6.2±0.4 6.5±0.9 5±0.5 6.8±0.3 2.7±0.4 3.1±0.3 2.8±0.3 3.4±0.3

VSin % 7.7±1.2 9±0.9 7.5±0.7 8±0.7 4.7±0.3 5.2±0.8 5 ±0.3 5.7±0.3
VSout % 4.7±0.3 5±0.8 4.5±0.3 5.7±0.3 1.1±0.3 1.1±0.3 1.7±0.3 2.2±0.2
TCODin gL-1 116±56.4 101±12.7 94.5±41 92.6±27.8 87.7±65.5 71.5±25 64.9±27.4 80.3±29.7
TCODout gL-1 87.7±28.8 71.5±25 64.9±27.4 80.3±29.7 25±21.6 31±12.5 28±8.3 43±30 .7
Gas production L L-1d-1 3.1±0.9 5±0.8 4.5±1.7 3.3±1.8 2.4±0.6 4.1±0.5 3±0.7 2.2±1.1

PT
CH4 % / 1.5±0.4 1.9±0.5 1.1±0.7 70.1±3.3 71.9±2.7 70.8±2.9 66.6±6.5
CH4 Yield L(kgVSin)-1 / 3.5±1.8 3.8±1.5 1.1±0.9 526±137 510±147 478±96 234±103
CO2 % 32±6.4 32.8±2.7 30.6±9.1 27.9±12.4 26.5±3.4 23.7±2.9 22.3±5.6 26.8±7.3
H2 % 64.3±6.9 60.8±1.5 56.4±6.5 42.2±9.8 / / / /

RI
H2 Yield L(kgVSin)-1 125±59 135±54.2 113±37.8 43.8±41.8 / / / /
NH4+-N mgL-1 41±8 585±71 875±148 1423±279 1035±217 1629±645 1988±358 1583±540
Total Alkalinity g-CaCO3 L-1

SC
/ / / / 9.5±0.2 8.3±1.8 18.3±0.5 14.4±0.7
Total VFAs gL-1 9.6±1.5 3.5±0.7 4.9±0.9 1.9±0.4 2.6±0.9 0.7±0.3 3.6±1.2 5.0±1.3
Acetic acid gL-1 4.0±1.7 2.4±0.5 2.5±0.5 1.2±0.3 2±0.7 0.5±0.2 1.4±0.4 3±0.8
Propionic acid gL-1 0.05±0.04 0±0 0±0 0±0 0.2±0.2 0±0 0.02±0.006 0.02±0.004

U
Butyric acid gL-1 5.5±1.9 1.1±0.3 2.3±0.5 0.8±0.2 0.2±0.3 0.1±0.07 2.1±0.8 1.9±0.9
gL-1
AN
Valeric acid 0.05±0.04 0.05±0.06 0.07±0.04 0±0 0.1±0.1 0.05±0.04 0.3±0.2 0.1±0.2
Energy kJ kg-1VSin 1,625 1,755 1,469 569 20,945 20,488* 19,185* 9,362*
182 Note: ‘/’ means data not available; ‘*’ energy from CH4 in the first and second stage
M

183 In this study, low NH4+-N concentration in H1, i.e. 41 mg L-1 was obtained in the system without

184 digestate recirculation. The introduction of the recirculation led to an increase of NH4+-N, i.e.
D

185 585, 875 and 1,430 mg L-1 for RR values of 0.3, 0.5, and 1, respectively. The ammonium
TE

186 increase agreed with data of Kobayashi et al. [20], who observed NH4+-N concentration

187 increasing in a hydrogen reactor according to the flow rate of the circulated sludge.
EP

188 Acetate and butyrate were the most represented VFAs in the H1 reactor (Table 3). The
C

189 recirculation diluted the acidity present in the H1 reactor, i.e. total VFAs (TVFAs) decreased

from 9.6 g L-1 in the process without the recirculation system to 1.9 g L-1 for RR 1. This result
AC

190

191 agreed with Kraemer and Bagley [34], who found in the H2 reactor the concentration of VFAs

192 decreased from 4.3 g L-1 without recirculation to 2.4 g L-1 with recirculation. During the

193 experiment using RR 0.3, and also during the trial period of RR 0.5, TVFAs increased from 3.5

194 to 4.9 g L-1. The reason for this variation is thought to be the methanogenic effluent flow rate and

10
 ACCEPTED MANUSCRIPT

195 different VFA accumulation in the second stage. When the RR increased to 1.0 the operation was

196 inhibited; the VFA and hydrogen production decreased to 1.9 g L-1 and 1.5 LH2 L-1 d-1. This was

197 probably due to the inhibition of H2 producing microorganism activity, as H2-consuming bacteria

198 concentration in H1 rose when the ratio of RR increased.

PT
199 The total amount of organic acids present in the hydrogen reactor depended on operating

RI
200 conditions, i.e. HRT, OLR, and substrates, therefore the alkali requirements significantly varied

201 with the variation of pH even at the same recirculation ratio [20]. Hydrogen production processes

SC
202 with recirculation able to maintain the pH at around 5.5 required the addition of 12, 16 and 13

203 mL Lreactor-1 of NaOH 1.25 mol L-1 for RR of 0.3, 0.5 and 1.0, respectively. The average of alkali

204
U
additions depended on the daily pH survey after feeding the methanogenic effluent. Compared
AN
205 with the volumes of NaOH required for an H2 production process without a recirculation system,

on average, the use of alkaline was reduced by 54% (from 26 to 12 mL Lreactor-1) for RR of 0.3 as
M

206

207 shown in Table 2 and Figure 1. It was reported that recirculation systems can significantly
D

208 decrease alkalinity needs for hydrogen production by 40% [34]. This fact indicated that effluent
TE

209 recirculation can effectively reduce the external alkali addition for pH control in a two-stage

210 hydrogen-producing system. In the current work, the highest hydrogen production of 3 L L-1 d-1
EP

211 was obtained when RR was 0.3. At the same time, an RR increase had negative effects on

212 hydrogen process stability. As previously found, the recirculation system significantly influenced
C

213 the overall performance of the two-stage process by increasing solid waste degradability [26,34].
AC

214 When a digestate recirculation system is applied within the AD process, the main results from

215 the different experimental investigation are summarized in Table 4.

216 Table 4 Comparison of different AD studies with controlling pH by recirculation system

Temperature OLR (kgVS.m-3d- Biogas yield (L kg- pH
Substrate Reactor type 1 1 Reference
(˚C) ) VSin) control

11
 ACCEPTED MANUSCRIPT

Stage Stage H2 CH4

Stage 1 Stage 2 Stage 1 Stage 2
1 2 (stage 1) (stage 2)
Yeshanew et al.
FW CSTR AFBR 55 37 3.4 6.1 115 334 R
(2016)
Kobayashi et al.
FW CSTR ABR 55 55 ns ns 68.4 392 R+C
(2012)
Bio- Cavinato et al.
CSTR CSTR 55 55 16 4 51 416 R
waste (2011)

PT
Chinellato et al.
FW CSTR CSTR 52 52 15 ns 1 484 R
(2013)
Chinellato et al.
FW CSTR CSTR 52 52 20 ns 117 311 R

RI
(2013)
OFMSW CSTR ASUBR 55 35 38.4 6.6 205 464 R Chu et al. (2008)
Cavinato et al.

SC
FW CSTR CSTR 55 55 16.3 4.8 66.7 720 R
(2012)
FW CSTR CSTR 55 35 18 5.7 135 510 R+C This study

U
FW CSTR CSTR 55 35 16 AN 4.2 113 478 R+C This study

FW CSTR CSTR 55 35 15 6.6 43 234 R+C This study

217 Note: AFBR: Anaerobic fixed bed reactor, ABR: Anaerobic baffled reactor, OFMSW: Organic fraction of
218 municipal solid wastes, ASUBR: Anaerobic suspended biofilm reactor, ns: not specified, R: recirculation, R+C:
219 recirculation + controlling pH by chemical agent
M

220 The results presented in Table 4 indicated that the introduction of recycling increased both

221 hydrogen production and yield, and minimized the total operational costs of treatment because of
D

222 a considerable reduction of the alkali additions needed.

TE

223 3.2 The methane yield with and without recirculation

EP

224 The daily methane production for trials performed was of 2.9, 2, and 1.5 L L-1d-1, i.e. 510, 478,

225 and 234 L kg-1VSin for RR of 0.3, 0.5, and 1.0, respectively (Figure 2). A CH4 production rate of
C

226 0.035, 0.139, 0.353 and 0.401 LCH4L-1d-1 was obtained in a recirculation approach using an
AC

227 artificial FW [22]. Wu et al. [12] reported that treating waste activated sludge by a temperature

228 phased process with recycling, 0.161 LCH4L-1d-1 with the content of 69.8% of CH4 could be

229 obtained. Methane production rate and yields obtained without recirculation were of 1.7 L L-1d-1,

230 i.e. 526 L kg-1VSin. With the recirculation system CH4 production increased; on the other hand, it

231 decreased without recirculation. In the current study, VFAs were of 3.6 ±1.2 and 5 ±1.3 gL-1 for

12
 ACCEPTED MANUSCRIPT

232 RR of 0.5 and 1.0 and so higher than 0.7 ±0.3 and 2.6 ±0.9 g L-1 detected for RR of 0.3 and

233 without recirculation. During the period from the 81st d to the 120th d and for an RR value of 1.0,

234 the pH of the effluent from M1 decreased from 7.8 to 7.3 (Table 2); this reduction occurred

235 because NaOH addition was reduced to 13 mL Lreactor-1.

PT
Methane Fermentation (35 ˚C)

RI
recycle
ratio 9 Without 0.3 0.5 1.0
(a)
Gas production (l/l/d)

SC
6

U
0 20 40 60 80 100 120 140 160 180

120
(b)
CH

Gas concentration (%)

AN
CO

80

40
M

0
0 30 60 90 120 150 180

9 30000
(c) pH
Total alkalinity Total alkalinity (mg/l)
D

20000
pH

7
10000
TE

5 0 30 60 90 120 150 180

0

8000
(d)
Acetate
EP
VFA (mg/l)

Butyrate
4000

0
C

1 21 41 61 81 101 121 141 161 181

Duration (d)
AC

236
237 Figure 2 Gas production and composition for methane (2nd stage) fermentation

238 In the second stage (methane-production reactor), the NH4+-N concentration increased because

239 of enhanced protein degradation. In particular, the NH4+-N concentration for RR of 0.3, 0.5, and

240 1.0 was of 1,639, 1,987, and 1,583 mg L-1, respectively, whereas in the case of the absence of

241 recirculation it was of 1,035 mg L-1 NH4+-N. The efficiency of the two-stage recirculation system

13
 ACCEPTED MANUSCRIPT

242 in terms of stability and methane production was affected by the hydrogen fermentation process

243 [34]. In the current study, the increasing of the RR to 1.0 inhibited the hydrogen production in

244 the H1, likewise in the M1 the pH and alkalinity decreased to 7.3 and 14 g L-1 and VFA

245 increased to 5g L-1.

PT
246 The overall hydrogen and methane yields decreased with the increase of the RR from 0.3 to 0.5

RI
247 and then to 1.0. The hydrogen yield decreased from 113 L kg-1VSin at a RR of 0.5 to 43.8 L kg-
1
248 VSin at a RR of 1.0. The methane yield also show a decline from 478 L kg-1VSin to 234 L kg-

SC
1
249 VSin similar with hydrogen production. When the effluent from methane reactor was recycled to

U
250 the hydrogen reactor the Methanobacterium or Methanoculleus was supposed to convert
AN
251 hydrogen to methane [24]. The RR effect on recirculation systems, hydrogen and methane

252 productions and yields can be clearly seen in Figure 3.

M

600
Methane (1st+2nd stage)
D
Hydrogen and methane yield

Hydrogen in first stage

TE

400
(L/Kg VSin)

200
EP

0
C

0 0.3 0.5 1

Recirculation ratio
AC

253
254 Figure 3 Effect of recirculation ratio on hydrogen and methane yield
255 Note: “0” means without recirculation

256 When the RR was 0.3, the highest gas yield, i.e. 135 L kg-1VSin for H2 and 510 L kg-1VSin for

257 CH4 was obtained. Therefore the results of this work indicated that there was no significant

14
 ACCEPTED MANUSCRIPT

258 difference between a two-stage anaerobic digestion with and without recirculation system on

259 total methane yield, i.e. from 526 to 510 L kg-1VSin.

260 3.3 Organics removal

261 The total TS removal efficiencies in the two-stage system were of 67%, 64%, and 60% for the

PT
262 processes with RR of 0.3, 0.5, and 1.0, respectively (Figure 4). Ventura et al. [35] obtained from

RI
263 a two-stage process producing H2 and CH4 under thermophilic and mesophilic condition

264 respectively, a TS and VS removal of 62% TS and 79.2% VS. According to the different RRs

SC
265 (RR of 0.3, 0.5 and 1.0) when in the FW-AD process substrate recirculation system was applied,

266 TS removals in the H1 reactor were 35%, 37%, and 24%. Compared to H1, TS removal

267
U
percentages in M1 were higher and of 54%, 51%, and 45%. A better TS conversion was obtained
AN
268 at a RR ratio of 0.3 than at other RRs. The total TS removal in the two-stage AD process without
M

269 recirculation was of 69%.

270 In H1, VS removal in the AD process without recirculation was 39%. A slight higher VS
D

271 removal efficiency in H1 at RR of 0.3 was found. As shown in Fig 4b, the overall VS removal
TE

272 efficiency tended to decrease when the digestate was recirculated. These results showed that both

273 TS and VS removal in H2 reactor was higher at a RR of 0.3 than at other RRs.
C EP
AC

15
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
274
M

275 Figure 4 TS and VS removal efficiencies in the two stages process

D

276 3.4 Mass and energy balance

TE

277 The COD mass balance was conducted to determine the conversion efficiency of the

278 recirculation system (Figure 5). Mass flows show the COD concentrations existed in H2 and CH4
EP

279 in the two reactors, the un-degraded materials and those lost. The percentage of the input COD

280 recovered as H2 and CH4 were as in the following: i. for RR of 0.3 CH4 in the first stage of 1 %
C

281 (v/v), H2 in the first stage of 8% (v/v), and CH4 in the second stage of 55% (v/v), ii. for RR of
AC

282 0.5, CH4 in the first stage of 0.8 % (v/v), H2 in the first stage of 6% (v/v), and CH4 in the second

283 stage of 44% (v/v), iii. for RR of 1.0, CH4 in the first stage of 0.3% (v/v), H2 in the first stage of

284 3% (v/v), and CH4 in the second stage of 40% (v/v). The mass balance without recirculation and

285 with reciculation ratio of 0.3 indicated a low COD loss. While at recirculation ratios of 0.5 and

286 1.0, the COD loss was slightly increased. The COD loss may be due to the inaccuracy of the

16
 ACCEPTED MANUSCRIPT

287 TCOD measurement. Some comparable results about the COD loss were also reported by

288 previously studies [36, 37].

120
TCOD H
CH
(stage2) CH
(stage1) lost
100

PT
80
TCOD (%)

60

RI
40

SC
20

0
FW (Influent) without 0.3 0.5 1.0

U
Two stage CSTR (55-35)°C
289
AN
290 Figure 5 Mass balance of two-stage system

291 Material flowing through two stages with recirculation is shown in Figure 6. For RR of 0.3, the
M

292 VS were converted at 9% into H2 and at 1% into CH4 in the first stage and for 60% into CH4 in

293 the second stage. When RR was increased to 0.5 and 1.0, VS conversion decreased to 1.0 % of
D

294 CH4 in the first stage for the two ratios, to 5.0% and 3.5% of H2 in the first stage and to 52% and
TE

295 36% of CH4 in the second stage, respectively.

296 Since the gross calorific value of hydrogen is 13,000 kJ m-3 and methane 39,820 kJ m-3, the
EP

297 energy recovered from hydrogen produced was of 1,755 kJ kg-1VSin, 1,469 kJ kg-1VSin and 569

kJ kg-1VSin and from methane produced was of 20,488 kJ kg-1VSin, 19,185 kJ kg-1VSin, and 9362
C

298

kJ kg-1VSin for RRs of 0.3, 0.5, and 1.0, respectively. Without recirculation, the energy from
AC

299

300 hydrogen and methane production was of 1,625 kJ kg-1VSin and 20,945 kJ kg-1VSin. At a RR of

301 0.3 condition, the energy recovered from H2 produced in the first reactor increased by 8%

302 compared to with the trial without recirculation. While as shown in Figure 5, the methane energy

303 produced decreased when recirculation was used. As a result, the total energy recovered from the

17
 ACCEPTED MANUSCRIPT

304 two reactors slightly decreased when recirculation RR of 0.3 was used. A total energy production

305 of 22.3 MJ kg-1VSin in the current study was one of the biggest values obtained from real FW-

306 AD processes found in the literature, i.e. 17.9 MJ kg-1VSin [38], 12.3 MJ kg-1VSin [22] and 9.58

307 MJ kg-1VSin [39].

PT
RI
U SC
AN
308
M

309 Figure 6: Material flow through two-stage recirculation

310 4. Conclusions
D

311 An integrated two-stage digestate recirculation system showed promising results in producing
TE

312 biohydrogen and biomethane from FW. Stable hydrogen and methane production were achieved

313 by the recirculation rate at 0.3. The maximum H2 and CH4 production rates were of 3 LL-1d-1 and
EP

314 2.9 LL-1d-1 i.e., around 70% of VS was degraded. The energy recovered from hydrogen and

315 methane was 1,755 kJkg-1VSin and 20,488 kJkg-1VSin. The process with digestate recirculation
C

316 showed similar energy efficiency to the AD process without recirculation, but reduced the
AC

317 amount of alkali needed to maintain the process stable by 54% when the ratio of 0.3 was adopted.

318 Acknowledgements
319 This work was supported by Beijing Municipal Science and Technology Plan Project

320 (D161100006016003 and D161100006016001), the Ministry of Science and Technology of

321 China within the China–Italy Cooperation on Food Waste Energy Utilization

18
 ACCEPTED MANUSCRIPT

322 (SQ2013ZOA000017). The stay of David Bianchi and Marina Ricci at China Agricultural

323 University was supported by the Project (SQ2013ZOA000017).

324 Reference
1. Paudel S., Kang Y., Yoo Y.S., Seo G.T., Effect of volumetric organic loading rate (OLR) on

PT
325
326 H2 and CH4 production by two-stage anaerobic co-digestion of food waste and brown water.
327 Waste Manage. 2017,61: 484-493.

RI
328 2. Pan J., Chen X., Sheng K., Yu Y., Zhang C., Ying Y., Effect of ammonia on biohydrogen
329 production from food waste via anaerobic fermentation. Int. J. Hydrogen Energy. 2013,38:
330 12747-12754.

SC
331 3. Petracchini F., Liotta F., Paolini V., Perilli M., Cerioni D., Gallucci F., Carnevale M.,
332 Bencini A., A novel pilot scale multistage semidry anaerobic digestion reactor to treat food
333 waste and cow manure. Int. J. Environ. Sci. Technol. 2017.
334 4.
U
Chen H., Jiang W., Yang Y., Yang Y., Man X., State of the art on food waste research: a
AN
335 bibliometrics study from 1997 to 2014. J. Cleaner Prod. 2017,.140: 840-846.
336 5. Paritosh K., Kushwaha S.K., Yadav M., Pareek N., Chawade A., Vivekanand V., Food
337 waste to energy: An overview of sustainable approaches for food waste management and
M

338 nutrient recycling. BioMed Res. 2017, Int. doi.org/10.1155/2017/2370927.

339 6. Jiang J., Zhang Y., Li K., Wang Q., Gong C., Li M., Volatile fatty acid production from
D

340 food waste: Effects of pH, temperature, and organic loading rate. Bioresour. Technol.
341 2013,143: 525-530.
TE

342 7. Peter L., Peter W., Oakdene H., Waste Arising in the Supply of Food and Drink to
343 Households in the UK. The United Kingdom. 2010.
344 8. Fisgativa H., Tremier A., Dabert P., Characterizing the variability of food waste quality: A
EP

345 need for efficient valorisation through anaerobic digestion. Waste Manage. 2016,50:264-
346 274.
C

347 9. Kondusamy D., Kalamdhad A.S., Pre-treatment and anaerobic digestion of food waste for
348 high rate methane production – A review. J. Environ. Chem. Eng. 2014, 2:1821-1830.
AC

349 10. Gottardo M., Micolucci F., Bolzonella D., Uellendahl H., Pavan P., Pilot-scale fermentation
350 coupled with anaerobic digestion of food waste - Effect of dynamic digestate recirculation.
351 Renew. Energy. 2017,114: 455-463.
352 11. Xiao B., Qin Y., Wu J., Chen H., Yu P., Liu J., LiY.Y., Comparison of single stage and two-
353 stage thermophilic anaerobic digestion of food waste: Performance, energy balance and
354 reaction process. Energy Convers. Manage. 2018,156: 215-223.

19
 ACCEPTED MANUSCRIPT

355 12. Wu L J., Qin Y., Hojo T., Li Y-Y., Upgrading of anaerobic digestion of waste activated
356 sludge by a temperature-phased process with recycling. Energy. 2015,87:381-389.
357 13. Chu C F., Li Y-Y., Xu K Q., Ebie Y., Inamori Y., Kong H N., A pH- and temperature-
358 phased two-stage process for hydrogen and methane production from food waste. Int. J.
359 Hydrogen Energy. 2008,33:4739-4746.

PT
360 14. Algapani D.E., Qiao W., Su M., Pumpo F.D., Wandera S.M., Adani F., Dong R., Bio-
361 hydrolysis and bio-hydrogen production from food waste by thermophilic and
362 hyperthermophilic anaerobic process. Bioresour. Technol. 2016,216:768-777.

RI
363 15. Torres M.L., Llorѐns M.C.E., Effect of alkaline pretreatment on anaerobic digestion of solid
364 wastes. Waste Manage. 2008,28:2229-2234.

SC
365 16. Li H., Li C., Liu W., Zou S., Optimized alkaline pretreatment of sludge before anaerobic
366 digestion. Bioresour. Technol. 2012,123:189-194.
367 17. Cesaro A., Belgiorno V., Pretreatment methods to improve anaerobic biodegradability of

U
368 organic municipal solid waste fractions: Review. Chem. Eng. J. 2014,240: 24-37.
AN
369 18. Zamanzadeh M., Hagen L.H., Svensson K., Linjordet R., Horn S.J., Anaerobic digestion of
370 food waste e Effect of recirculation and temperature on performance and microbiology.
371 Water Res. 2016,96:246-254.
M

372 19. Cheng S.S., Chao Y.C., Yang K.H., Bai M.D., Process recovery of bio- hydrogenation in a
373 pilot plant from methanogens invasion. Int. J. Hydrogen Energy. 2011,36:8779-8784.
D

374 20. Kobayashi T., Xu K.Q., Li Y-Y., Inamori Y., Effect of sludge recirculation on
375 characteristics of hydrogen production in a two-stage hydrogen-methane fermentation
TE

376 process treating food wastes. Int. J. Hydrogen Energy. 2012,37: 5602-5611.
377 21. Thong S.O., Suksong W., Promnuan K., Thipmunee M., Mamimin C., Prasertsan P., Two-
378 stage thermophilic fermentation and mesophilic methanogenic process for biohythane
EP

379 production from palm oil mill effluent with methanogenic effluent recirculation for pH
380 control. Int. J. Hydrogen Energy. 2016,41:21702-21712.
381 22. Yeshanew M.M., Frunzo L., Pirozzi F., Lens P.N.L., Esposito G., Production of biohythane
C

382 from food waste via an integrated system of the continuously stirred tank and anaerobic
383 fixed bed reactors. Bioresour. Technol. 2016,220:312-322.
AC

384 23. Zuo Z., Wu S., Zhang W., Dong R., Effects of organic loading rate and effluent
385 recirculation on the performance of two-stage anaerobic digestion of vegetable waste.
386 Bioresour. Technol. 2013,146:556-561.
387 24. Algapani D.E., Qiao W., Su M., Pumpo F.D., Wandera S.M., Adani F., Dong R., Long-term
388 bio-H2 and bio-CH4 production from food waste in a continuous two-stage system: Energy
389 efficiency and conversion pathways. Bioresour. Technol. 2018,248:204-213.

20
 ACCEPTED MANUSCRIPT

390 25. APHA, Standard Methods for the Examination of Water and Wastewater. American Public
391 Health Association Inc, Washington, DC.1995.
392 26. Cavinato C., Bolzonella D., Fatone F., Cecchi F., Pavan P., Optimization of two-phase
393 thermophilic anaerobic digestion of bio-waste for hydrogen and methane production
394 through reject water recirculation. Bioresour. Technol. 2011,102:8605-8611.

PT
395 27. Chinellato G., Cavinato C., Bolzonella D., Heaven S., Banks C.J., Biohydrogen production
396 from food waste in batch and semi-continuous conditions: evaluation of a two-phase
397 approach with digestate recirculation for pH control. Int. J. Hydrogen Energy.

RI
398 2013,38:4351-4360.
399 28. Khan M.A., Ngo H.H., Guo W., Liu Y., Zhang X., . Guo J., Chang S.W., Nguyen D.D.,
400 Wang J., Biohydrogen production from anaerobic digestion and its potential as renewable

SC
401 energy. Renew. Energy.2017, http://dx.doi.org/10.1016/j.renene.2017.04.029
402 29. Liu X., Li R., Ji M., Han L., Hydrogen and methane production by co-digestion of waste
activated sludge and food waste in the two-stage fermentation process: substrate conversion

U
403
404 and energy yield. Bioresour. Technol.2013,146:317-323.
AN
405 30. Wang X., Zhao Y.C., A bench-scale study of fermentative hydrogen and methane
406 production from food wastes in the integrated two-stage process. Int. J. Hydrogen Energy.
407 2009,34:245-254.
M

408 31. Salerno M.B., Park W., Zuo Y., Logan B.E., Inhibition of biohydrogen production by
409 ammonia. Water Res. 2006,40:1167-1172.
D

410 32. Tang Y., Shigematsu T., Morimura S., Kida K., Microbial community analysis of
411 mesophilic anaerobic protein degradation process using bovine serum albumin (BSA)-fed
TE

412 continuous cultivation. J. Biosci. Bioeng. 2005,99:150-164.

413 33. Breure A.M., Mooijman K.A., Andel J.G., Protein degradation in anaerobic digestion:
414 influence of volatile fatty acids and carbohydrates on hydrolysis and acidogenic
EP

415 fermentation of gelatin. Appl. Microbiol. Biotechnol. 1986,24:426-431.

416 34. Kraemer J.T., Bagley D.M., Continuous fermentative hydrogen production using a two-
417 phase reactor system with recycles. Environ. Sci. Technol. 2005,39:3819-3825.
C

418 35. Ventura J.S., Lee J., Jahng D., A comparative study on the alternating mesophilic and
AC

419 thermophilic two-stage anaerobic digestion of food waste. J. Environ. Sci. 2014,26:1274-
420 1283.
421 36. Mthethwa N.P., Nasr M., Bux F., Kumari S., Utilization of Pistia stratiotes (aquatic weed)
422 for fermentative biohydrogen: Electron-equivalent balance, stoichiometry, and cost
423 estimation. Int. J. Hydrogen Energy. 2018, 43: 8243-8255.
424 37. Reddy K., Nasr M., Kumari S., Kumar S., Gupta S.K., Enitan A.M., Bux F., Biohydrogen
425 production from sugarcane bagasse hydrolysate: effects of pH, S/X, Fe2+, and magnetite
426 nanoparticles. Environ Sci Pollut Res. 2017, 24: 8790-8804.

21
 ACCEPTED MANUSCRIPT

427 38. Chu C.F., Xu K.Q., Li Y-Y., Inamori Y., Hydrogen and methane potential based on the
428 nature of food waste materials in a two-stage thermophilic fermentation process. Int. J.
429 Hydrogen Energy. 2012,37:10611-10618.
430 39. Lee D.Y., Ebie Y., Xu K.Q., Li Y-Y., Inamori Y., Continuous H2 and CH4 production from
431 high-solid food waste in the two-stage thermophilic fermentation process with the
432 recirculation of digester sludge. Bioresour. Technol. 2010,101- S42-S47.

PT
RI
U SC
AN
M
D
TE
C EP
AC

22
 ACCEPTED MANUSCRIPT

Highlights

• Two-stage AD process was positive for food waste digestion

• Digestate recirculation in the first stage reduced alkaline addition by 54%

• Optimal recirculation ratio was 0.3

PT
• The hydrogen energy obtained from FW was improved by 8% with

RI
recirculation

• Digestate recycling did not increase the overall energy yield

U SC
AN
M
D
TE
C EP
AC