Science of the Total Environment 724 (2020) 138168

Contents lists available at ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

Temperature-phased anaerobic co-digestion of food waste and paper

waste with and without recirculation: Biogas production and
microbial structure
Lu Li a, Zhe Kong a, Yu Qin a, Jing Wu a, Aijun Zhu a, Benyi Xiao b, Jialing Ni a, Kengo Kubota a, Yu-You Li a,⁎
a
Laboratory of Environmental Protection Engineering, Department of Civil and Environmental Engineering, Graduate School of Engineering, Tohoku University, 6-6-06 Aza-Aoba, Aramaki, Aoba
Ward, Sendai, Miyagi 980-8579, Japan
b
Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing 100085, China

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• Recirculation helped co-digestion sys-

tem achieved phase separation.
• Recirculated system obtained higher re-
moval efficiency than non-recirculated
one.
• Recirculated system achieved stable H2
production in the first reactor.
• Thermophilic cellulose-degrading bac-
teria were enriched due to recirculation.
• Archaea activity was inhibited in the
first reactor thus ensured H2
accumulation.

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

Article history: Two temperature-phased anaerobic digestion (TPAD) systems (55 °C in the first reactor and 35 °C in the second
Received 29 December 2019 reactor) with and without recirculation were operated in parallel for the co-digestion of food waste and paper
Received in revised form 22 March 2020 waste. A long-term experiment was carried out for these two systems with the paper waste ratios elevated
Accepted 22 March 2020
from 0 to 50%. The removal efficiencies of COD, TS, VS, carbohydrate and protein in the recirculated TPAD system
Available online 24 March 2020
were higher than those of the non-recirculated system. The successful acclimation of thermophilic cellulose-
Editor: Yifeng Zhang degrading bacteria in the first reactor (RT1), partly due to recirculation, ensured the effective degradation of cel-
lulose when the paper waste ratio was higher than 40%, resulting in the production of large amounts of hydrogen
Keywords: in reactor RT1. In the absence of recirculation, the main substance produced in the first reactor of the non-
Anaerobic co-digestion recirculated system (T1) was lactic acid. This gradually led to over-acidification and a low degradation efficiency
Temperature-phased and no methane or hydrogen was produced in T1. Recirculation helped to establish a stable bacterial community
Recirculation capable of producing bio-hydrogen in reactor RT1. The relatively low pH of 5.5 in the RT1 inhibited the activity of
Microbial structure hydrogenotrophic archaea without consuming hydrogen, facilitating high hydrogen production levels.
Biogas
© 2020 Elsevier B.V. All rights reserved.
Paper waste

1. Introduction

Food waste (FW) is one of the most abundant components of the

⁎ Corresponding author. organic fraction of municipal solid waste (OFMSW) in both devel-
E-mail address: gyokuyu.ri.a5@tohoku.ac.jp (Y.-Y. Li). oped and developing countries, with 1.3 billion tonnes produced

https://doi.org/10.1016/j.scitotenv.2020.138168
0048-9697/© 2020 Elsevier B.V. All rights reserved.
2 L. Li et al. / Science of the Total Environment 724 (2020) 138168

globally in 2016 (Kuczman et al., 2018). The amount of FW produced microorganisms and their functions in the degradation of cellulose
increases annually, with contributions from hotels, restaurants, fam- and the production of hydrogen were investigated and analyzed.
ilies, canteens and companies all over the world (Zhang et al., 2014). The objectives of this study were as follows: 1) To illustrate the effect
Paper waste (PW) is defined as waste mostly comprised of toilet of digestate recirculation on the TPAD system by investigating the long-
paper, newspaper, printer paper, wrapping paper and cardboard term performance of organic removal efficiency as well as methane and
(Fonoll et al., 2016). The trend towards using paper towels or tissues hydrogen production with the increasing of PW ratio; and 2) To deter-
instead of reusable towels at mealtimes has been shown to contrib- mine the reasons for the different performances of NR-TPAD and R-
ute significantly to the volume of PW around the world (Argun and TPAD by determining the impact of key hydrolytic, fermentative and
Onaran, 2018). Because garbage classification systems for municipal methanogenic microorganisms on the entire system. Based on the re-
waste recycling have yet to be implemented in most developing sults of this study, suggestions will also be made to improve the perfor-
countries, and are rare even in developed countries, the complete mance of the co-digestion of FW and PW in the treatment of OFMSW.
separation of FW and PW is difficult to achieve (López et al., 2010;
Romero-Güiza et al., 2014). For this reason, PW is typically accompa- 2. Materials and methods
nied with FW as mixed waste in the OFMSW.
Anaerobic digestion (AD), a sustainable and environmental friendly 2.1. Experimental apparatus and procedure
process, is considered a promising technology for the disposal and min-
imization of organic wastes like FW and PW, and has the added value of The main components of the entire TPAD system for the co-digestion
simultaneously generating recycled fuel in the form of either methane of FW and PW were two continuously stirred tank reactors (CSTR) in
or hydrogen (Hassan et al., 2017; Kong et al., 2018). To achieve stable operating in series. As shown in Fig. S1, the NR-TPAD and R-TPAD sys-
operation in the AD system, the optimal carbon/nitrogen ratio (C/N tems were clearly marked and were operated in parallel. The NR-
ratio) needs to be maintained: a high C/N ratio results in insufficient ni- TPAD system contains a thermophilic (55 °C) CSTR with an effective vol-
trogen, and results in the accumulation of volatile fatty acids (VFAs) ume of 3 L (T1) and a mesophilic (35 °C) CSTR with an effective volume
while a low C/N ratio often results in ammonia inhibition (Wang et al., of 12 L (M2). The R-TPAD system used the same volume as the two
2014). By adopting a co-digestion strategy for the AD process, a suitable CSTRs (RT1 and RM2) with an additional recirculation device pumping
C/N ratio can be maintained in the substrate by mixing nitrogen- the digestate from the second reactor, RM2, back to the first reactor,
enriched organic wastes, such as FW, with nitrogen-poor organic RT1, at a recirculation ratio of 0.4 (Qin et al., 2019). No additional alka-
waste, such as PW (Liu et al., 2015; Wang et al., 2012). Because the linity was provided to the reactors. The same substrate tank was used to
co-digestion of different organic wastes enhances degradation perfor- prepare the mixture of FW and PW for both reactors, operated under an
mance, it is attracting more attention than mono-digestion in the re- organic loading rate (OLR) of approximately 4.4–5.3 g COD L−1 d−1. The
search community (Asato et al., 2016; Kong et al., 2019; Li et al., 2018). total hydraulic retention time (HRT) of the entire system was 30 days
It is acknowledged that the two-phase AD system enhances the sta- since a relative long HRT of 30 days is known to be suitable for the AD
bility and improves substrate degradation because it provides the ideal of solid waste. The HRT of the first reactor (T1 and RT1) was 6 days,
environmental conditions for different groups of microorganisms while that of the second reactor (M2 and RM2) was 24 days, and sepa-
(Algapani et al., 2019; Carvalheira et al., 2018; Wu et al., 2018). Since ration was based on a recirculation ratio of 0.4. It should be noted that,
AD is also significantly influenced by temperature, there has been grow- in this study, the HRT is equivalent to the sludge retention time (SRT).
ing interest in the temperature-phased anaerobic digestion (TPAD) sys- The long-term continuous operation of both NR-TPAD and R-TPAD con-
tem (Buffière et al., 2018; Dooms et al., 2018).A typical TPAD system tinued for 365 days. When the removal performance was stable (with
establishes a phase-separation strategy, with hydrolysis, acidogenesis stable removal and stable biogas production), the PW ratio in the sub-
and acetogenesis typically under a thermophilic condition of 55 °C in strate was increased: this occurred on the 127th, 202nd and 279th days.
the first reactor (Liu et al., 2013), and methanogenesis in the second re-
actor under a mesophilic condition of 35 °C (Xiao et al., 2019). This has 2.2. Feedstock and inoculum sludge
been shown facilitate the effective conversion from organic wastes to
hythane (a mixture of methane and hydrogen) (Jiang et al., 2018). It The FW substrate was artificially prepared every 30 days and was
should also be noted that the recirculation of digestate from the second stored at 4 °C by smashing and mixing the following raw materials: a
reactor to the first reactor has been reported to stabilize the pH without total solid (TS) ratio of 20% wasted staple foods (including rice, bread,
the requirement for any extra alkalinity, and this resulted in higher hy- udon noodles and Chinese noodles), 14% animal products (including
drogen production (Cavinato et al., 2011; Kobayashi et al., 2012). There- wasted meat, fish and eggs), 30% fruit waste and 36% vegetable waste.
fore, for the reliable co-digestion of FW and PW, the recirculated The recipe was based on a local survey result of the real FW content
temperature-phased anaerobic digestion (R-TPAD) system is consid- (Li et al., 2003). The PW substrate was prepared by artificially mixing
ered a more effective system with better outcomes than the non- shredded toilet paper, printed office paper and wasted newspaper
recirculated temperature-phased anaerobic digestion (NR-TPAD). How- with a ratio of 1:1:1 (dry weight). The mixture of FW and PW was pre-
ever, few studies have been focused on the application of the TPAD sys- pared once a week and was mixed into a slurry and stored in the sub-
tem to the co-digestion of PW and FW. To the best of our knowledge, strate tank at 4 °C, with a TS content of around 10%. During the
there has been no study in which the improvement of recirculation operation, the PW ratio in the mixture substrate was periodically ele-
has been considered from the microbial angle. The comparative advan- vated from 0 to 20%, 40% and 50% of the TS in the following runs
tages of R-TPAD can be investigated by analyzing the effect of recircula- (Run-1 to Run-4). Trace elements, including Fe, Ni and Co, were added
tion on the shift of prokaryotic community structure and functional to the substrate to ensure sufficient nutrients: sufficiency was defined
microorganisms. The detailed results obtained from a microbial analysis as 10 mg Fe, 1 mg Co and 1 mg Ni per 1 kg OFMSW (wet weight).
during long-term operation are expected to clearly show the superiority The mesophilic reactors (M2 and RM2) were inoculated with the
of the recirculation strategy. Therefore, in this study, two TPAD systems, mesophilic anaerobic digested sludge from a local wastewater treat-
one with and one without recirculation, were operated in parallel for ment plant, while the thermophilic reactors (T1 and RT1) were inocu-
the co-digestion of FW and PW during a long-term operation of lated with the same anaerobic digested sludge acclimated to 55 °C.
365 days with periodical increases in the PW ratio of the substrate Before initiating the long-term continuous operation, both the NR-
from 0 to 50%. Microbial samples of both the NR-TPAD and R-TPAD sys- TPAD and R-TPAD were given two weeks for the start-up of a stable di-
tems were taken during the stable operation for each PW increment. gestion of FW at an HRT of 100 days and another two weeks at an HRT of
The characterization of the prokaryotic community structure and key 50 days.
 L. Li et al. / Science of the Total Environment 724 (2020) 138168 3

2.3. Chemical analysis were generated based on a similarity of 97%. OTUs were aligned against
the most recent vision of the Greengenes database, 13_8, (http://qiime.
The chemical reagents used in the analysis were all purchased from org/home_static/dataFiles.html). Chimeras and Singleton OTUs were re-
Wako Co. Ltd., Japan. The values of pH were measured by a HM-30R pH moved and the sequence number of each sample was unified to 55,000
meter (DKK-TOA). The total ammonia nitrogen (TAN, N-NH+ 4 ) concen- with random selection in order to remove the bias of sequencing depth
tration was determined with the indophenol method (Bolleter et al., caused by the DNA concentration and PCR. The Chao1 richness index,
1961). The alkalinity was determined by titration with terminal pH the Shannon's diversity index and the Simpson's diversity index were
values of 4.8. calculated based on the clustered sequence data (Magurran, 1988).
The concentrations of COD, TS and volatile solid (VS) were measured The representative sequence of the main OTUs listed in tables and fig-
according to the standard methods: potassium dichromate method was ures were submitted to GenBank under the accession numbers from
used for COD and dry weight method for TS and VS. Carbohydrate was MT013413 to MT013484.
measured with the phenol-sulfate examination method (Qin et al.,
2019). Protein was measured with the traditional Folin-phenol Lowry
method. 3. Results and discussion
The volatile fatty acids (VFAs) and ethanol (EtOH) concentrations
were detected using gas chromatography equipped with a flame ioniza- 3.1. Physical and chemical parameters
tion detector (GC-FID, Agilent 6890) and a 30 m capillary column (J&W,
DB-WAXetr). The temperature of the injector and detector were both As the PW ratio increased from 0 to 50%, the C/N ratio in the sub-
250 °C. For oven was used as follows: at 125 °C for 5 min and then the strate increased from 17.1 to 25.8. The OLRs of four runs were 5.3, 4.7,
temperature was elevated at a rate of 2.5 °C/min, and held at 180 °C 4.6 and 4.4 g COD L−1 d−1, respectively. While the pH of both systems
for 5 min. The VFAs detected in this study include acetic acid (HAc), gradually decreased, the pH of the mesophilic reactors (above 7.2)
propionic acid (HPr), butyric acid (HBu/i-HBu), valeric acid (HVa/i- was higher than that of thermophilic ones (below 5.2). Note also that
HVa) and caproic acid (HCa). Lactic acid (HLa) concentrations were de- the pH value in reactor RT1 (4.59–5.18) was higher than that in T1
termined by capillary zone electrophoresis (CZE, Agilent 7100) with a (3.76–4.80), with recirculation providing alkalinity to the first reactor.
(D/L)-lactic acid standard. The electrophoresis solution consisted of The pH, TAN and alkalinity concentrations are given in Table 1 and the
Tris, quinolinic acid and HDTMA, with a pH at 7.3. concentrations of COD, TS, VS, carbohydrate and protein are shown in
Biogas produced from the CSTR was collected and recorded by a wet Table S1.
tip gas meter (W-NK-0.5B, Shinagawa) and the volume was normalized The overall removal efficiency of both NR-TPAD and R-TPAD during
with the standard temperature and pressure (Kong et al., 2018). The the entire 365-day experiment is given in Table 2. The general perfor-
component ratios of N2, CH4 and CO2 were measured by a gas chromato- mance of the R-TPAD was better than that of the NR-TPAD, and the per-
graph equipped with a thermal conductivity detector (GC-TCD, formance was notably better when the PW ratio reached 50%: The R-
Shimadzu GC-8A) with 2 m stainless steel column packed with Porapak TPAD system maintained COD removal efficiencies from 86.2% to
Q. The H2 content of the biogas was detected by another GC-TCD 76.5% with the increase of PW ratio from 0 to 50% while those of the
(Shimadzu GC-8A) with a column packed with a molecular sieve 5A NR-TPAD system was from 81.3% to 69.7%. The VS removal efficiencies
(60/80, φ3 mm). The temperature of the injector, column and detector in the R-TPAD decreased from 86.0% to 78.4% as the PW ratio increased,
were 100 °C, 70 °C and 100 °C, respectively. The carrier gas was argon while those in the NR-TPAD decreased from 81.6% to 72.5%. Both the
and was controlled at 100 kPa. NR-TPAD and R-TPAD realized a carbohydrate removal efficiency of
over 86.5%.
2.4. Microbial community structure analysis Fig. 1 illustrates the contribution of each reactor to COD, VS and car-
bohydrate removal. With the COD, VS and carbohydrate contribution
Sludge samples were collected from the outlet of each reactor during ratios all lower than 9%, it is reasonable to conclude that reactor T1
the stable phase of each run on Day 97, 181, 275 and 350, then stored in barely contributed to the degradation of the substrate in the NR-TPAD
−80 °C before DNA extraction. The collected sludge was initially centri- system. The substrate was mainly degraded in the second reactor (con-
fuged at 10,000 rpm for 10 min at 4 °C and 0.15 g of the sediment was tributing for 91% ~ 100%), as is indicated in Fig. 1(a), (c) and (e). How-
used for DNA extraction with ISOIL for Beads Beating kit (Nippon ever, a large proportion of COD, VS and carbohydrate degradation
gene, Japan). The concentrations and quality of the extracted DNA occurred in reactor RT1, accounting for 70%, 74% and 82%, respectively
were measured with NanoDrop 2000 (Nanodrop Inc., USA) and diluted when the PW ratio was 0, as can be seen in Fig. 1(b), (d) and (f). Even
to 10 ng μL−1 for PCR. Former primer 341F (5'-CCT AYG GGR BGC ASC when the PW was elevated to 50%, the COD, VS and carbohydrate deg-
AG-3′) and mixed reverse primer 806R & 806R-P (5′-GGA CTA CHV radation in reactor RT1 was 27%, 45% and 40%, respectively. This result
GGG THT CTA AT-3′ & 5′-GGA CTA CCA GGG TAT CTA AG-3′ in ratio of demonstrates the importance of the role of first reactor in maintaining
30:1) were used for the PCR amplification of V3-V4 fragment in 16S the relatively high removal efficiency of organic wastes in the R-TPAD
rRNA gene (Li et al., 2019; Ni et al., 2019). The PCR amplifications system. This means the OLR of the second reactor (RM2) is much
were performed under the following conditions: 30 cycles at 94 °C for lower, which elevates the total removal efficiencies of the whole
5 s, 50 °C for 5 s, 68 °C for 10 s, and a final extension at 68 °C for 7 min system.
with Low DNA Ex Taq® (TaKaRa, Japan). Triplicate PCR products for From the view of biogas production rate, the R-TPAD system was
each sample were mixed to minimize the PCR bias before the sequenc- also superior to the NR-TPAD system. In the NR-TPAD system, almost
ing barcode was added. The purity and concentration of the amplified no hythane was generated by the first reactor, and all the hythane
DNA was verified by a DNA chip with Bioanalyzer 7500 (Agilent Tech- was produced in reactor M2. Despite the high gas production rate ob-
nologies, USA) and then purified with Agencourt® AMPure® XP tained, the CH4 content in reactor M2 was b60%. Hydrogen was gener-
(Beckman Coulter, Inc., USA) according to the manufacturers' instruc- ated in the R-TPAD system, accounting for approximately 50% of the
tions. Purified DNA was measured by Qubit 3.0 ® (Life technologies, produced biogas in the first reactor. Also, a large amount of methane
USA) and diluted to 4 nmol L−1. Illumina Miseq platform were used was produced from RM2 and the CH4 content was higher than that of
for sequencing. M2. The total combustion heat of methane and hydrogen derived from
The raw data were processed by QIIME (version 1.8.0): Sequences the removed COD in R-TPAD was significantly higher than that in the
were assembled and filtered to remove poor-quality sequences (longer NR-TPAD, and even 10 times higher when the PW was 50%, as shown
than 480 bp or shorter than 400 bp). Operation taxonomic units (OTUs) in Table S2. These results are clear evidence that the degradation
4 L. Li et al. / Science of the Total Environment 724 (2020) 138168

Table 1
The pH, TAN and alkalinity of the sludge in each reactor.

System Stage PW ratio (%) pH TAN (g L−1) Alkalinity (g CaCO3 L−1)

Substrate tank 0 3.99 ± 0.06 0.04 ± 0.02 0

20 4.64 ± 0.37 0.04 ± 0.01 0
40 4.92 ± 0.32 0.05 ± 0.00 0
50 5.23 ± 0.24 0.05 ± 0.01 0.45 ± 0.31
NR-TPAD (non-recirculation) First reactor (T1, 55 °C) 0 3.76 ± 0.16 0.03 ± 0.01 0
20 4.44 ± 0.34 0.19 ± 0.20 0
40 4.59 ± 0.16 0.06 ± 0.00 0
50 4.80 ± 0.14 0.05 ± 0.01 0
Second reactor (M2, 35 °C) 0 7.80 ± 0.03 1.76 ± 0.07 8.03 ± 0.10
20 7.64 ± 0.21 1.39 ± 0.08 6.80 ± 0.34
40 7.22 ± 0.05 0.45 ± 0.05 3.42 ± 0.23
50 7.28 ± 0.10 0.79 ± 0.48 4.31 ± 1.52
R-TPAD (recirculation) First reactor (RT1, 55 °C) 0 4.59 ± 0.13 0.85 ± 0.30 0
20 4.74 ± 0.04 0.76 ± 0.11 0
40 4.92 ± 0.02 0.34 ± 0.08 0
50 5.18 ± 0.11 0.27 ± 0.04 1.15 ± 0.56
Second reactor (RM2, 35 °C) 0 7.81 ± 0.04 2.19 ± 0.36 9.72 ± 0.20
20 7.52 ± 0.20 0.91 ± 0.09 7.38 ± 1.26
40 7.33 ± 0.12 0.66 ± 0.05 4.52 ± 0.28
50 7.26 ± 0.19 0.70 ± 0.20 4.03 ± 0.91

effectiveness and biogas production of the R-TPAD were better than these, phylum Firmicutes and Proteobacteria accounted for over 99.5%
those of the NR-TPAD during the entire operation period. of the entire community, as shown in Fig. S2(a).
The OTUs with a relative abundance of N0.5% and their main func-
tions are listed in Table S4. Among them, the functional lactic acid bac-
3.2. Bacterial community structure in reactor T1 teria (LAB) group predominates, with 13 species, all belonging to order
Lactobacillales. The total number of LAB increased from 74.5% in Run-1
Reactor T1 had an average Shannon diversity of 2.19 (Table S3), to 95.2% in Run-4. However, the species belonging to order Bacillales,
which is much lower than that in a typical thermophilic AD system which has protein-degrading ability, only survived in Run-1. Species be-
(Gagliano et al., 2015). A total of 21 bacteria phyla were identified. Of longing to order Rhodospirillales, which were aligned as acetogenic

Table 2
Removal efficiency, biogas production rate and fermentable product concentrations of the long-term performance of NR-TPAD and R-TPAD.

System Stage PW Removal efficiency (%)a Biogas production rate Fermentable products (mg L−1)
ratio
COD TS VS Carbohydrate Protein GPR CH4 H2 HAc HPr i-HBu HBu HLa EtOH
(%)
(L L−1 d−1) content content
(%) (%)

NR-TPAD First reactor (T1, 0 4.34 4.98 5.15 8.85 2.55 – – – 1358 67 86 26 14,930 6729
(non-recirculation) 55 °C) 20 2.49 5.99 6.19 8.00 15.31 0.14 – – 1904 123 106 675 13,230 3422
± 0.58
40 6.90 6.55 6.84 4.47 14.51 0.06 – – 2036 255 295 388 16,560 2601
± 0.07
50 0.00 0.00 0.21 1.66 19.25 0.06 – – 1373 70 110 438 14,270 1427
± 0.08
Second reactor 0 81.30 77.01 81.58 94.28 49.76 3.22 59.23 – 52 34 8 0 0 0
(M2, 35 °C) ± 0.12 ± 3.77
20 78.96 73.76 79.49 94.32 35.72 2.69 57.95 – 47 20 7 24 0 0
± 0.21 ± 2.06
40 75.53 74.09 78.81 91.08 20.47 2.58 56.79 – 172 0 0 0 0 0
± 0.42 ± 4.04
50 69.73 67.57 72.49 86.45 0.00 2.44 53.99 – 33 15 0 11 0 0
± 0.23 ± 3.43
R-TPAD First reactor 0 60.61 59.29 63.29 78.86 40.80 1.58 0.42 47.75 3101 300 117 1419 9842 3798
(recirculation) (RT1, 55 °C) ± 0.45 ± 0.23 ± 4.76
20 38.30 40.21 44.19 60.87 21.44 2.59 – 49.49 4644 167 96 1535 7493 1813
± 0.49 ± 4.79
40 35.55 38.32 41.74 49.12 24.47 1.74 – 50.81 3415 280 216 2916 7185 1489
± 0.94 ± 4.46
50 20.65 32.48 35.08 35.59 14.47 2.37 – 52.42 2244 97 138 4137 4242 1200
± 0.38 ± 1.75
Second reactor 0 86.22 80.96 85.97 96.08 63.82 2.86 62.25 – 62 7 35 17 0 6
(RM2, 35 °C) ± 0.11 ± 0.40
20 82.58 77.03 83.05 95.26 49.96 2.61 60.25 – 174 49 33 118 0 0
± 0.21 ± 3.51
40 77.60 73.38 79.69 90.96 32.75 2.24 60.28 – 181 64 43 46 0 58
± 0.37 ± 3.37
50 76.55 72.27 78.40 89.11 23.42 2.31 56.14 – 55 14 6 21 0 0
± 0.19 ± 4.71
a
The removal efficiencies of the Second reactor (M2 and RM2) represent the total removal efficiency of the entire system.
 L. Li et al. / Science of the Total Environment 724 (2020) 138168 5

Fig. 1. Variation of contribution ratios first reactors (T1 and RT1) and the second reactors (M2 and RM2) for degrading COD (a), VS (b) and carbohydrate (c) in the NR-TPAD system and the
R-TPAD system, respectively.

bacteria (ACB) (Pitiwittayakul et al., 2016), also decreased from 3.7% in Lactobacillus helveticus, Leuconostoc inhae, Acetobacter fabarum and Lac-
Run-1 and to 0.5% in Run-2 and were not detected in either Run-3 or in tobacillus siliginis) were also predominate in RT1 and showed a similar
Run-4. It has been reported that the high concentration of lactic acid variation trend with the increase of PW ratio, as shown in Fig. 2(a).
produced by LAB significantly inhibit the growth of other bacteria The occupation of IB in reactor RT1 was maintained at around 40%,
(Handous et al., 2019; Lyhs et al., 2015). This finding is consistent with and LAB were the predominant functional group in IB. Besides,
the variation of community structures in our system. Acetobacter fabarum, an acetogenic bacterium was present in Run-1,
As can be seen in Table 1 and Table 2, the pH of the reactor T1 was but then disappeared in the other runs.
very low, in the 3.76–4.80 range. This low pH facilitated the lactic acid
and alcohol fermentation, resulting in higher concentrations of lactic
3.3.2. The recirculated bacteria
acid and ethanol (Fu and Mathews, 1999; Luedeking and Piret, 2000).
Recirculation brought the entire microbial community from reactor
It should also be noted that the acetate concentration of T1 was lower
RM2 to RT1 together with the digestate. In reactor RT1, those bacteria
than that of RT1. This result suggests that the conversion from ethanol
which were sourced from the recirculation were defined as the
to acetate was also inhibited: such an inhibition has been reported to
recirculated bacteria (RB). As shown in Fig. 2(a), RB have different bio-
weaken biogas production (Hwang et al., 2004). This is consistent
logical classifications, which greatly improves the α-diversity of RT1.
with the lower biogas production rate found in T1.
The relative abundance of RB decreased from 26.5% in Run-1 and
Therefore, it is likely that the predomination of these LAB made the
47.2% in Run-2 to 9.7% in Run-3 and 5.3% in Run-4, as can be seen in
system over-acidized and eventually led to a deteriorated and simpli-
Fig. 3. They were mainly hydrolytic bacteria, acidogenic bacteria and
fied microbial community (Banks et al., 2012; Khan et al., 2016). The
syntrophic bacteria. The presence of RB suggests that reactor RT1 was
performance of T1 went into a vicious circle, and only produced lactic
able to effectively degrade carbohydrate (including cellulose and cello-
acid, which greatly reduced the organic removal efficiency of the NR-
biose) and protein to produce hydrogen (detailed information is shown
TPAD system.
in Table S5).
Large quantities of RB were transferred to reactor RT1, establishing a
3.3. Bacterial community structure in reactor RT1
stable microbial consortium in RT1 when PW ratio was lower than 20%.
However, as the cellulose ratio in the substrate continued to rise, the RB
Compared with T1, reactor RT1 obtained a relatively high Shannon
apparently did not provide effective degradation ability, especially
diversity with average value of 3.73. A total of 32 bacterial phyla were
under the inappropriate temperature from the mesophilic to thermo-
detected in the RT1 samples. As mentioned above, the R-TPAD had bet-
philic condition.
ter performance both in removal efficiencies and hydrogen production
than that of the NR-TPAD due to the efficient operation of RT1. There-
fore, by comparing the microbial community structure of RT1 to the 3.3.3. Thermophilic-enriched bacteria
other 3 reactors, the predominant bacteria in RT1 were classified into Apart from the indigenous bacteria and the recirculated bacteria,
the 3 groups to explain the effects of recirculation on removal efficiency some strains of thermophilic-enriched bacteria (TB) were detected in
and hydrogen production. reactor RT1 but not in the other three reactors. The pH of RT1 was higher
than that of T1 because of the alkalinity provided from the recirculated
3.3.1. The indigenous bacteria digestate. Moreover, the elevated microbial diversity of RT1 due to recir-
Since RT1 differed from T1 in terms of recirculation, T1 can be culation served to enrich thermophilic bacteria. Even though only 4
regarded as the blank control of RT1. Those bacteria which existed in OTUs were classified as TB, the total relative abundance among them
both first reactors (T1 and RT1) but absence in the second ones (M2 was high. It is highly likely that the presence of these TB also helped
and RM2) were defined as the indigenous bacteria (IB). The five- to achieve cellulose degradation and hydrogen production in reactor
major species of IB predominating in T1 (Lactobacillus sakei, RT1.
6 L. Li et al. / Science of the Total Environment 724 (2020) 138168

Fig. 2. A heat map of TOP 15 bacterial OTUs (a) and TOP 5 archaea OTUs (b) in reactor RT1. The listed bacterial OTUs accounted for N68.9% of the total bacterial sequences in RT1 and archaea
OTUs accounted for N90.0% of the total archaea sequences in RT1. The IB are marked in green, the RB are marked in blue and the TB are marked in red. The relative abundance of these OTUs
in reactor T1, M2 and RM2 are also shown for comparison purposes.

The denovo10312 was the most predominant OTU detected in RT1, The other OTU worthy of mention is denovo9795, which had a
with a relative abundance ranging from 19.4% to 34.3%, which increased relative abundance of almost 0 in the first two runs but as much as
as the PW ratio was increased. This OTU obtained a 95% similarity with 7.2% in Run-3 and 12.9% in Run-4. This OTU showed a 100% similarity
Caproiciproducens galactitolivorans, a bacteria reportedly capable of with Thermoanaerobacterium thermosaccharolyticum, a hydrogen-
degrading glucose and lactose to H2, CO2, VFAs and ethanol (Kim et al., producing bacterium capable of degrading cellulose (Zhang et al.,
2015). The presence of this strain is thought to have contributed signif- 2019). It is reported that an enzyme named β-glucosidase (BGL) pro-
icantly to the hydrogen production of reactor RT1. duced by T. thermosaccharolyticum has cellobiose-hydrolyzing ability
and this enzyme exhibits high enzymatic activity over a broad tem-
perature range, from 45 to 70 °C. The BGL also has shown a high tol-
erance to glucose and cellobiose due to its high specific activity (Pei
et al., 2012). The decomposition of cellobiose is a rate-limiting step
for cellulose degradation since cellobiose inhibits the activities of
cellobiohydrolases and endoglucanases (George et al., 2001). Due
to the reduction in cellobiose inhibition after cellobiose is hydro-
lyzed to glucose by BGL, the cellulolytic enzymes function more effi-
ciently. Therefore, it is reasonable to assume that the presence and
enrichment of T. thermosaccharolyticum in this system enhanced
the ability of hydrolyzing cellulose and cellobiose, and the by-
products of hydrolysis were further utilized by C. galactitolivorans
and the RB for hydrogen production.
In addition to denovo9795, there were significant increases in the
abundance of two other OTUs at the high PW ratio, and both of them
are known to be capable of degrading cellulose. The total abundance
of cellulose-degrading bacteria which belong to RB and TB, respectively,
were calculated. As indicated in Fig. 4, the total abundance of the
cellulose-degrading bacteria in RT1 basically remained stable at around
10% during the entire operation. Although the degradation of cellulose
in the system remained stable, the type of cellulose-degrading bacteria
gradually shifted from RB to TB as the PW ratio increased, and the de-
composition pathway of cellulose changed from a cellose and glucose
pathway to a VFA pathway.
The successful enrichment of these TB in reactor RT1 is evidence of a
specific microbial evolution associated with the elevated pH and ther-
mophilic condition in the R-TPAD system. The resultant microbial com-
Fig. 3. The variation of total relative abundance of indigenous bacteria (IB), recirculated munity helps to maintain the cellulose-degradation performance in the
bacteria (RB) and thermophilic-enriched bacteria (TB). R-TPAD system under high PW ratios.
 L. Li et al. / Science of the Total Environment 724 (2020) 138168 7

3.4. Bacterial community structure in reactors M2 and RM2 a proportion of HAc exist in the system in its undissociated form. It has
been reported that un-HAc is the uncoupler of the plasma membrane,
As can be seen in Table S3, the second reactors M2 and RM2 had a exhibiting biotoxicity and inhibiting the synthesis of ATP (Mawson
higher microbial diversity than the first reactors T1 and RT1. This can et al., 1991). When the concentration of un-HAc exceeds 310 mg L−1,
be concluded from the significantly higher Shannon diversity of all sam- 50% of the methane production becomes inhibited and when the con-
ples from both M2 (4.64–5.67) and RM2 (4.83–5.48) than those of T1 centration exceeds 2360 mg L−1, 90% of the methane production be-
(1.70–2.57) and RT1 (2.64–4.89). comes inhibited. Note also that when the concentration exceeds
In the bacterial community, as shown in Fig. S2(b) and (d), 810 mg L−1, the activity of hydrogenotrophic methanogens is not recov-
Firmicutes and Bacteroidetes were predominant in M2 and RM2. The erable (Zhang et al., 2018).
third predominant phylum, Chloroflexi, obtained a high abundance in With the concentrations of HAc given in Table 2, the concentrations
M2 but not in RM2: this was especially notable in Run-2 and Run-3. Al- of un-HAc in RT1 were calculated as follows:
though the identified predominant OTUs in M2 and RM2 were shown to
have the similar ability of degrading cellulose, protein or saccharide, the C
½HAc ¼ ð1Þ
microbial community of these two reactors differed significantly at the 1 þ 10pH−pKa
genus level, as can be seen in Table S6 and Table S7.
where [HAc] is the undissociated acetic acid concentration (mg L−1), C
3.5. Effect of recirculation on archaeal community is the total acetic acid concentration (mg L−1), Ka is the dissociation con-
stant with the value of 4.76 for HAc. As a result, the concentration of un-
As shown in Fig. 2(b), archaea were barely detected in reactor T1. It HAc in RT1 were 1850.1, 2375.5, 1396.5 and 618.1 mg L−1 as the PW
has been reported that the methanogenesis is significantly inhibited ratio increased, obviously exceeding the inhibition concentration for
when the pH is below 5.5 (Liu et al., 2008). The results indicate that methanogenesis.
the pH of reactor T1 at all PW ratio was below 5.0. As mentioned above, the activity of archaea in reactor RT1 was
However, a certain quantity of archaea was detected in RT1. The pre- inhibited by low pH and the toxicity of undissociated acid. Therefore, al-
dominant archaeal species in RT1 in each run was highly consistent with though a certain amount of hydrogenotrophic archaea was transferred
that in RM2. It mainly consisted of Methanosaeta concilii, from RM2 to RT1 by recirculation, the produced hydrogen in RT1 was
Methanothermobacter wolfei, Methanoculleus receptaculi, not utilized or consumed by these archaea. As a result, the invalidation
Methanobacterium petrolearium and Methanosarcina thermophile. of hydrogenotrophic methanogenesis in RT1 ensured the obvious hy-
Among which, Methanothermobacter wolfei, Methanoculleus receptaculi, drogen accumulation of R-TPAD system.
and Methanobacterium petrolearium were hydrogenotrophic
methanogens. 3.6. Environmental parameters and sample clustering
Hydrogenotrophic methanogen is known to utilize hydrogen and
CO2 to produce methane, but methanogenesis did not occur in reactor The canonical correlation analysis (CCA) diagram illustrated in Fig. 5
RT1, despite there being sufficient available substrate. It is assumed shows the clustering trend of all microbial samples. Recirculation signif-
that the toxicity of undissociated acetic acid (un-HAc) inhibited biolog- icantly influenced the composition and variations in the microbial com-
ical activity in this research. As the pH of RT1 floated between 4.59–5.18, munity: The distance between RT1 and RM2 in the R-TPAD was much

Fig. 4. The comparison of the cellulose degrading bacteria in reactor RT1 between low PW ratio (≤20%) and high PW ratio (≥40%).
8 L. Li et al. / Science of the Total Environment 724 (2020) 138168

closer than that of T1 and M2 in the NR-TPAD. The microbial community the first reactor, and was shown to significantly enhance the co-
structure in the first reactor with short HRT was more likely to be influ- digestion of food waste and paper waste. This was especially the case
enced by recirculation, as can be seen by the clustering of the samples for high-strength paper waste.
from M2 and RM2 but a dispersing of the samples from T1 and RT1.
Among all the environmental factors, temperature is the most impor- CRediT authorship contribution statement
tant one that affects the community structure. It should be noted that al-
though pH was given as a categorical factor in CCA, pH was actually Lu Li: Conceptualization, Software, Formal analysis, Data curation,
influenced by temperature and recirculation. In the case of the thermo- Writing - original draft. Zhe Kong: Writing - original draft, Writing - re-
philic reactors, the decrease in the pH was due to the production of view & editing. Yu Qin: Methodology. Jing Wu: Investigation. Aijun
acetic acid by the ACB but was elevated due to the recirculation of Zhu: Investigation. Benyi Xiao: Methodology. Jialing Ni: Formal analy-
digestate. It can therefore be concluded that the pH did not directly in- sis. Kengo Kubota: Methodology. Yu-You Li: Supervision.
fluence the entire system, but it was just one factor that influenced the
distribution of microorganisms. The variation in the PW ratio exhibited Declaration of competing interest
significant influence on RT1. The clustering of RT1 samples greatly
changed when the PW ratio was higher than 40%. This was consistent The authors declare that they have no known competing financial
with the previous description as shown in Fig. 3 and Fig. 4: there were interests or personal relationships that could have appeared to influ-
some thermophilic bacteria enriched in RT1, which substituted for RB ence the work reported in this paper.
and played a role in the degradation of cellulose.
Acknowledgements
4. Conclusions
This work was supported by Japan Society for the Promotion of Sci-
The stable co-digestion of food waste and paper waste was realized ence (JSPS) KAKENHI Grant Number JP 18J11397 (the first author) and
in the R-TPAD system at a PW ratio of 50%, achieving a VS removal effi- Number JP 17J02720 (the second author).
ciency higher than 78.4% and carbohydrate removal efficiency higher
than 89.1%. However, the NR-TPAD system realized the same removal Appendix A. Supplementary data
performance only at a maximum PW ratio of 40%. The R-TPAD also suc-
cessfully produced bio-hydrogen. The variations in the microbial com- Supplementary data to this article can be found online at https://doi.
munity variations were as follows: org/10.1016/j.scitotenv.2020.138168.
1. With the recirculation of digestate, microorganisms were trans-
ferred to the first reactor, helping to establish a functional microbial References
community capable of hydrolyzing organics and producing hydrogen.
This effect was more obvious when PW ratio ≤20%. Algapani, D.E., Qiao, W., Ricci, M., Bianchi, D., Wandera, S.M., Adani, F., Dong, R., 2019. Bio-
hydrogen and bio-methane production from food waste in a two-stage anaerobic di-
2. Because the recirculated digestate provided alkalinity to the first gestion process with digestate recirculation. Renew. Energy 130, 1108–1115. https://
reactor, the pH was maintained at higher than 4.5. At 55 °C, thermo- doi.org/10.1016/j.renene.2018.08.079.
philic carbohydrate-degrading bacteria massively bred in the first reac- Argun, H., Onaran, G., 2018. Dark fermentative hydrogen gas production from lime
treated waste paper towel hydrolysate. Waste Biomass Valorization 9, 801–810.
tor. When PW ≥40%, thermophilic cellulose-degrading bacteria replaced https://doi.org/10.1007/s12649-017-9957-2.
the role of recirculated mesophilic bacteria, becoming the main func- Asato, C.M., Gonzalez-Estrella, J., Jerke, A.C., Bang, S.S., Stone, J.J., Gilcrease, P.C., 2016.
tional group of cellulose degradation. Batch anaerobic digestion of synthetic military base food waste and cardboard mix-
tures. Bioresour. Technol. 216, 894–903. https://doi.org/10.1016/j.
3. Since the utilization of hydrogen by archaea was inhibited by low biortech.2016.06.033.
pH and undissociated acids, the accumulation of hydrogen was realized Banks, C.J., Zhang, Y., Jiang, Y., Heaven, S., 2012. Trace element requirements for stable
in RT1. food waste digestion at elevated ammonia concentrations. Bioresour. Technol. 104,
127–135. https://doi.org/10.1016/j.biortech.2011.10.068.
Recirculation helps to realize the phase separation of R-TPAD sys- Bolleter, W.T., Bushman, C.J., Tidwell, P.W., 1961. Spectrophotometric determination of
tem, converting the hydrolysis product from lactic acid to hydrogen in ammonia as indophenol. Anal. Chem. 33, 592–594. https://doi.org/10.1021/
ac60172a034.
Buffière, P., Dooms, M., Hattou, S., Benbelkacem, H., 2018. The hydrolytic stage in high
solids temperature phased anaerobic digestion improves the downstream methane
production rate. Bioresour. Technol. 259, 111–118. https://doi.org/10.1016/j.
biortech.2018.03.037.
Carvalheira, M., Cassidy, J., Ribeiro, J.M., Oliveira, B.A., Freitas, E.B., Roca, C., Carvalho, G.,
Oehmen, A., Reis, M.A.M., 2018. Performance of a two-stage anaerobic digestion sys-
tem treating fruit pulp waste: the impact of substrate shift and operational condi-
tions. Waste Manag. 78, 434–445. https://doi.org/10.1016/j.wasman.2018.06.013.
Cavinato, C., Bolzonella, D., Fatone, F., Cecchi, F., Pavan, P., 2011. Optimization of two-
phase thermophilic anaerobic digestion of biowaste for hydrogen and methane pro-
duction through reject water recirculation. Bioresour. Technol., Special Issue:
Biofuels-III: Biohydrogen 102, 8605–8611. https://doi.org/10.1016/j.
biortech.2011.03.084.
Dooms, M., Benbelkacem, H., Buffière, P., 2018. High solid temperature phased anaerobic
digestion from agricultural wastes: putting several reactors in sequence. Biochem.
Eng. J. 130, 21–28. https://doi.org/10.1016/j.bej.2017.11.011.
Fonoll, X., Astals, S., Dosta, J., Mata-Alvarez, J., 2016. Impact of paper and cardboard sup-
pression on OFMSW anaerobic digestion. Waste Manag. 56, 100–105. https://doi.
org/10.1016/j.wasman.2016.05.023.
Fu, W., Mathews, A.P., 1999. Lactic acid production from lactose by Lactobacillus
plantarum: kinetic model and effects of pH, substrate, and oxygen. Biochem. Eng. J.
3, 163–170. https://doi.org/10.1016/S1369-703X(99)00014-5.
Gagliano, M.C., Braguglia, C.M., Gianico, A., Mininni, G., Nakamura, K., Rossetti, S., 2015.
Thermophilic anaerobic digestion of thermal pretreated sludge: role of microbial
community structure and correlation with process performances. Water Res. 68,
498–509. https://doi.org/10.1016/j.watres.2014.10.031.
George, S.P., Ahmad, A., Rao, M.B., 2001. Studies on carboxymethyl cellulase produced by
Fig. 5. CCA for all the samples with the environmental factors of recirculation, an alkalothermophilic actinomycete. Bioresour. Technol. 77, 171–175. https://doi.
temperature, pH and PW ratio. org/10.1016/S0960-8524(00)00150-4.
 L. Li et al. / Science of the Total Environment 724 (2020) 138168 9

Handous, N., Gannoun, H., Hamdi, M., Bouallagui, H., 2019. Two-stage anaerobic digestion sorting. Resour. Conserv. Recycl. 54, 222–228. https://doi.org/10.1016/j.
of meat processing solid wastes: methane potential improvement with wastewater resconrec.2009.08.003.
addition and solid substrate fermentation. Waste Biomass Valorization 10, Luedeking, R., Piret, E.L., 2000. A kinetic study of the lactic acid fermentation. Batch pro-
131–142. https://doi.org/10.1007/s12649-017-0055-2. cess at controlled pH. Biotechnol. Bioeng. 67, 636–644. https://doi.org/10.1002/
Hassan, M., Ding, W., Umar, M., Rasool, G., 2017. Batch and semi-continuous anaerobic co- (SICI)1097-0290(20000320)67:6b636::AID-BIT3N3.0.CO;2-U.
digestion of goose manure with alkali solubilized wheat straw: a case of carbon to ni- Lyhs, U., Snauwaert, I., Pihlajaviita, S., Vuyst, L.D., Vandamme, P., 2015. Leuconostoc rapi
trogen ratio and organic loading rate regression optimization. Bioresour. Technol. sp. nov., isolated from sous-vide-cooked rutabaga. Int. J. Syst. Evol. Microbiol. 65,
230, 24–32. https://doi.org/10.1016/j.biortech.2017.01.025. 2586–2590. https://doi.org/10.1099/ijs.0.000305.
Hwang, M.H., Jang, N.J., Hyun, S.H., Kim, I.S., 2004. Anaerobic bio-hydrogen production Magurran, A.E., 1988. Ecological Diversity and Its Measurement. Princeton University
from ethanol fermentation: the role of pH. J. Biotechnol. 111, 297–309. https://doi. Press.
org/10.1016/j.jbiotec.2004.04.024. Mawson, A.J., Earle, R.L., Larsen, V.F., 1991. Degradation of acetic and propionic acids in
Jiang, H., Qin, Y., Gadow, S.I., Ohnishi, A., Fujimoto, N., Li, Y.-Y., 2018. Bio-hythane produc- the methane fermentation. Water Res. 25, 1549–1554. https://doi.org/10.1016/
tion from cassava residue by two-stage fermentative process with recirculation. 0043-1354(91)90187-U.
Bioresour. Technol. 247, 769–775. https://doi.org/10.1016/j.biortech.2017.09.102. Ni, J., Hatori, S., Wang, Y., Li, Y.-Y., Kubota, K., 2019. Uncovering viable microbiome in an-
Khan, M.A., Ngo, H.H., Guo, W.S., Liu, Y., Nghiem, L.D., Hai, F.I., Deng, L.J., Wang, J., Wu, Y., aerobic sludge digesters by propidium monoazide (PMA)-PCR. Microb. Ecol. https://
2016. Optimization of process parameters for production of volatile fatty acid, doi.org/10.1007/s00248-019-01449-w.
biohydrogen and methane from anaerobic digestion. Bioresour. Technol. 219, Pei, J., Pang, Q., Zhao, L., Fan, S., Shi, H., 2012. Thermoanaerobacterium
738–748. https://doi.org/10.1016/j.biortech.2016.08.073. thermosaccharolyticum β-glucosidase: a glucose-tolerant enzyme with high specific
Kim, B.-C., Seung Jeon, B., Kim, S., Kim, H., Um, Y., Sang, B.-I., 2015. Caproiciproducens activity for cellobiose. Biotechnol. Biofuels 5, 31. https://doi.org/10.1186/1754-6834-
galactitolivorans gen. nov., sp. nov., a bacterium capable of producing caproic acid 5-31.
from galactitol, isolated from a wastewater treatment plant. Int. J. Syst. Evol. Pitiwittayakul, N., Theeragool, G., Yukphan, P., Chaipitakchonlatarn, W., Malimas, T.,
Microbiol. 65, 4902–4908. https://doi.org/10.1099/ijsem.0.000665. Muramatsu, Y., Tanasupawat, S., Nakagawa, Y., Yamada, Y., 2016. Acetobacter
Kobayashi, T., Xu, K.-Q., Li, Y.-Y., Inamori, Y., 2012. Effect of sludge recirculation on char- suratthanensis sp. nov., an acetic acid bacterium isolated in Thailand. Ann. Microbiol.
acteristics of hydrogen production in a two-stage hydrogen–methane fermentation 66, 1157–1166. https://doi.org/10.1007/s13213-016-1200-z.
process treating food wastes. Int. J. Hydrog. Energy, XII International Symposium on Qin, Y., Wu, J., Xiao, B., Cong, M., Hojo, T., Cheng, J., Li, Y.-Y., 2019. Strategy of adjusting re-
Polymer Electrolytes: New Materials for Application in Proton Exchange Membrane circulation ratio for biohythane production via recirculated temperature-phased an-
Fuel Cells 37, 5602–5611. https://doi.org/10.1016/j.ijhydene.2011.12.123. aerobic digestion of food waste. Energy 179, 1235–1245. https://doi.org/10.1016/j.
Kong, Z., Li, L., Kurihara, R., Kubota, K., Li, Y.-Y., 2018. Anaerobic treatment of N, N- energy.2019.04.182.
dimethylformamide-containing wastewater by co-culturing two sources of inocu- Romero-Güiza, M.S., Peces, M., Astals, S., Benavent, J., Valls, J., Mata-Alvarez, J., 2014. Im-
lum. Water Res. 139, 228–239. https://doi.org/10.1016/j.watres.2018.03.078. plementation of a prototypal optical sorter as core of the new pre-treatment config-
Kong, Z., Li, L., Xue, Y., Yang, M., Li, Y.-Y., 2019. Challenges and prospects for the anaerobic uration of a mechanical–biological treatment plant treating OFMSW through
treatment of chemical-industrial organic wastewater: a review. J. Clean. Prod. 231, anaerobic digestion. Appl. Energy 135, 63–70. https://doi.org/10.1016/j.
913–927. https://doi.org/10.1016/j.jclepro.2019.05.233. apenergy.2014.08.077.
Kuczman, O., Gueri, M.V.D., De Souza, S.N.M., Schirmer, W.N., Alves, H.J., Secco, D., Buratto, Wang, X., Yang, G., Feng, Y., Ren, G., Han, X., 2012. Optimizing feeding composition and
W.G., Ribeiro, C.B., Hernandes, F.B., 2018. Food waste anaerobic digestion of a popular carbon–nitrogen ratios for improved methane yield during anaerobic co-digestion
restaurant in Southern Brazil. J. Clean. Prod. 196, 382–389. https://doi.org/10.1016/j. of dairy, chicken manure and wheat straw. Bioresour. Technol. 120, 78–83. https://
jclepro.2018.05.282. doi.org/10.1016/j.biortech.2012.06.058.
Li, Y.-Y., Mizuno, O., Funaishi, K., Yamashita, K., 2003. High-rate methanation of the food Wang, X., Lu, X., Li, F., Yang, G., 2014. Effects of temperature and carbon-nitrogen (C/N)
wastes and garbage by a two-phase process with circulation of digested sludge. Envi- ratio on the performance of anaerobic co-digestion of dairy manure, chicken manure
ron. Eng. Res. 40, 321–332. https://doi.org/10.11532/proes1992.40.321. and rice straw: focusing on ammonia inhibition. PLoS One 9, e97265. https://doi.org/
Li, W., Siddhu, M.A.H., Amin, F.R., He, Y., Zhang, R., Liu, G., Chen, C., 2018. Methane produc- 10.1371/journal.pone.0097265.
tion through anaerobic co-digestion of sheep dung and waste paper. Energy Convers. Wu, C., Huang, Q., Yu, M., Ren, Y., Wang, Q., Sakai, K., 2018. Effects of digestate recircula-
Manag. 156, 279–287. https://doi.org/10.1016/j.enconman.2017.08.002. tion on a two-stage anaerobic digestion system, particularly focusing on metabolite
Li, L., Qin, Y., Kong, Z., Wu, J., Kubota, K., Li, Y.-Y., 2019. Characterization of microbial com- correlation analysis. Bioresour. Technol. 251, 40–48. https://doi.org/10.1016/j.
munity and main functional groups of prokaryotes in thermophilic anaerobic co- biortech.2017.12.020.
digestion of food waste and paper waste. Sci. Total Environ. 652, 709–717. https:// Xiao, B., Zhang, W., Yi, H., Qin, Y., Wu, J., Liu, J., Li, Y.-Y., 2019. Biogas production by two-
doi.org/10.1016/j.scitotenv.2018.10.292. stage thermophilic anaerobic co-digestion of food waste and paper waste: effect of
Liu, D., Zeng, R.J., Angelidaki, I., 2008. Effects of pH and hydraulic retention time on hydro- paper waste ratio. Renew. Energy 132, 1301–1309. https://doi.org/10.1016/j.
gen production versus methanogenesis during anaerobic fermentation of organic renene.2018.09.030.
household solid waste under extreme-thermophilic temperature (70°C). Biotechnol. Zhang, C., Su, H., Baeyens, J., Tan, T., 2014. Reviewing the anaerobic digestion of food
Bioeng. 100, 1108–1114. https://doi.org/10.1002/bit.21834. waste for biogas production. Renew. Sust. Energ. Rev. 38, 383–392. https://doi.org/
Liu, Z., Zhang, C., Lu, Y., Wu, X., Wang, Lang, Wang, Linjun, Han, B., Xing, X.-H., 2013. States 10.1016/j.rser.2014.05.038.
and challenges for high-value biohythane production from waste biomass by dark Zhang, W., Dai, K., Xia, X.-Y., Wang, H.-J., Chen, Y., Lu, Y.-Z., Zhang, F., Zeng, R.J., 2018. Free
fermentation technology. Bioresour. Technol., Biorefineries 135, 292–303. https:// acetic acid as the key factor for the inhibition of hydrogenotrophic methanogenesis in
doi.org/10.1016/j.biortech.2012.10.027. mesophilic mixed culture fermentation. Bioresour. Technol. 264, 17–23. https://doi.
Liu, L., Zhang, T., Wan, H., Chen, Y., Wang, X., Yang, G., Ren, G., 2015. Anaerobic co- org/10.1016/j.biortech.2018.05.049.
digestion of animal manure and wheat straw for optimized biogas production by Zhang, K., Cao, G.-L., Ren, N.-Q., 2019. Bioaugmentation with Thermoanaerobacterium
the addition of magnetite and zeolite. Energy Convers. Manag. 97, 132–139. https:// thermosaccharolyticum W16 to enhance thermophilic hydrogen production using
doi.org/10.1016/j.enconman.2015.03.049. corn stover hydrolysate. Int. J. Hydrog. Energy 44, 5821–5829. https://doi.org/
López, M., Soliva, M., Martínez-Farré, F.X., Fernández, M., Huerta-Pujol, O., 2010. Evalua- 10.1016/j.ijhydene.2019.01.045.
tion of MSW organic fraction for composting: separate collection or mechanical