International Biodeterioration & Biodegradation xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

International Biodeterioration & Biodegradation

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

Biogas desulfurization under anoxic conditions using synthetic wastewater

and biogas slurry
Yong Zenga,b, Lifan Xiaoa, Xueying Zhangb,∗∗, Jun Zhoub, Gaosheng Jia, Sarah Schroederd,
Guoxiao Liue, Zhiying Yana,c,∗
a
Key Laboratory of Environmental and Applied Microbiology, CAS, Environmental Microbiology Key Laboratory of Sichuan Province, Chengdu Institute of Biology, Chinese
Academy of Sciences, Chengdu 610041, China
b
Nanjing University of Technology, Nanjing 211816, China
c
Hunan Co-Innovation Center of Animal Production Safety, CICAPS, Changsha 410128, China
d
Technological University of Berlin, Berlin 10623, Germany
e
Aerospace Kaitian Environmental Technology Co., Lid, Changsha 410129, China

A R T I C LE I N FO A B S T R A C T

Keywords: A feasibility study was conducted to determine whether aerated biogas slurry is a suitable nutrient solution for
Hydrogen sulfide biogas desulfurization systems with a biotrickling filter. At a loading rate of 36.20 g-H2S m−3h−1, the H2S and
Biogas desulfurization NOx−-N removal efficiencies were 84.7% (average elimination capacity of 30.67 g-H2S m−3h−1) and 60.9%,
Wastewater denitrification respectively, when utilizing synthetic wastewater in a simultaneous biogas desulfurization and wastewater de-
Aerated biogas slurry
nitrification system. However, these efficiencies were just 61.9% (average elimination capacity of 22.42 g-H2S
Microbial community
m−3h−1) and 49.2%, respectively, when biogas slurry was used. High-throughput sequencing revealed that the
Thiobacillus and Sulfurimonas genera were the main functional bacteria. Alpha- and beta-diversity analyses
showed that the H2S loading rate significantly affected the microbial community structure, especially in the
system utilizing aerated biogas slurry. Finally, based on the results, we describe a feasible approach to using
biogas slurry for biogas desulfurization.

1. Introduction Currently, physicochemical and biological methods are typically

Biogas, which comprises a mixture of different gases, is produced by
anaerobic digestion (Lastella et al., 2002). Manure from livestock and
poultry is the main raw material for the production of biogas in China
(Li et al., 2009a). Manure contains large quantities of proteins and
other sulfur-containing compounds. As a consequence, the biogas pro-
duced by anaerobic digestion will contain hydrogen sulfide (H2S)
(Pokorna et al., 2015; Potumarthi et al., 2009), which is a toxic gas with
a strong corrosive effect on combustion power equipment and metal
pipes (Vikromvarasiri et al., 2017). Furthermore, during the combus-
tion of biogas, H2S is converted into sulfur oxides, which are released
into the atmosphere, causing air pollution (Watsuntorn et al., 2017;
Zhou et al., 2015). Therefore, the H2S in biogas must be removed before
used.
adopted for the removal of H2S from biogas (Kao et al., 2012). How-
ever, physicochemical methods such as chemical oxidation, physical
adsorption, and cryogenic separation require large quantities of che-
mical agents and adsorbents, and they consume significant amounts of
energy (Zhang et al., 2016). Moreover, these adsorbents are expensive
to dispose of and may cause secondary pollution (Chaiprapat et al.,
2011). Hence, the most popular method for removing H2S from biogas
is biological desulfurization (Pokorna et al., 2015; Rattanapan et al.,
2010), of which there are two main types: aerobic desulfurization and
anoxic desulfurization. Aerobic desulfurization technology has been
studied extensively (Rattanapan et al., 2010). With this method, how-
ever, oxygen is used as the electron acceptor, and mixing oxygen with
biogas can dilute the biogas and lead to an increased risk of explosion.
Anoxic desulfurization has lower limitations on the mass transfer for

Abbreviations: H2S, hydrogen sulfide; BTF, biotrickling filter; WWTP, Wastewater Treatment Plants; SDD, simultaneous biogas desulfurization and wastewater
denitrification system; LR, loading rate; RE, removal efficiency; EC, elimination capability; HRT, hydraulic retention time; EBRT, empty bed residence time
∗
Corresponding author. Key Laboratory of Environmental and Applied Microbiology, CAS, Environmental Microbiology Key Laboratory of Sichuan Province,
Chengdu Institute of Biology, Chinese Academy of Sciences, Chengdu 610041, China.
∗∗
Corresponding author.
E-mail address: yanzy@cib.ac.cn (Z. Yan).

https://doi.org/10.1016/j.ibiod.2018.05.012
Received 10 March 2018; Received in revised form 17 May 2018; Accepted 18 May 2018
0964-8305/ © 2018 Elsevier Ltd. All rights reserved.

Please cite this article as: Zeng, Y., International Biodeterioration & Biodegradation (2018), https://doi.org/10.1016/j.ibiod.2018.05.012
Y. Zeng et al.
nitrate compared to oxygen absorption in aerobic biotrickling filters
(BTFs). However, the cost and the need for large quantities of nitrate
can limit the application of anoxic desulfurization (Almenglo et al.,
2016b). The anoxic method utilizes NOx− as the electron acceptor,
whereby H2S is oxidized to elemental sulfur or sulfates (Almenglo et al.,
2016a). The reaction conditions are mild and require little energy, and
simultaneous desulfurization of biogas and denitrification of waste-
water can be achieved by exploiting the principle of treating waste with
waste (Mahmood et al., 2007a). The main reactions in the systems are
as follows (Li et al., 2009b):
5S2− + 2NO3− + 12H+ → 5S + N2 + 6H2O ΔGθ = −955 kJ/reaction
(1)
5S + 6NO3− + 2H2O → 5SO42− + 3N2 + 4H+ ΔGθ = −2738 kJ/
reaction (2)
3S2− + 2NO2− + 8H+ → 3S + N2 + 4H2O ΔGθ = −917 kJ/reaction
(3)
3S + 6NO2− → 3SO42− + 3N2 ΔGθ = −2027 kJ/reaction (4)
Unfortunately, biogas plants are typically located in remote loca-
tions, where there are no available wastewater resources containing
NOx− ions, such as a domestic wastewater or swine wastewater (Pirolli
et al., 2016). The use of synthetic wastewater is uneconomical, and it
increases operational and maintenance complexity (Arespacochaga
et al., 2014). As is known, ammonia nitrogen can be converted into
nitrate/nitrite nitrogen by a nitrification reaction, and this reaction has
been widely employed for removing ammonia from Wastewater
Treatment Plants (WWTPs) (Deng et al., 2009). Therefore, if biogas
slurry could be used for biogas desulfurization plant, it would greatly
reduce construction costs and simplify the operational process. How-
ever, because the composition of biogas slurry is complex, there have
been few reports of its use in desulfurization procedures (Almenglo
et al., 2016b; Deng et al., 2009). Moreover, even fewer studies have
compared synthetic wastewater and biogas slurry as a nutrient solution
in a simultaneous biogas desulfurization and wastewater denitrification
system (SDD).
This study utilizes synthetic wastewater and aerated biogas slurry as
a nutrient solution to investigate the influence of the biogas loading
rate (LR) on the H2S removal efficiency (RE), and on the structure of the
key microbial communities during the reaction process. We also com-
pare the effectiveness of two types of nutrient solutions, synthetic
wastewater and aerated biogas slurry. Based on the results, some sug-
gestions for how biogas slurry can be used for biogas desulfurization are
provided.
2. Materials and methods
2.1. Experimental reactors of biogas desulfurization
The experimental reactors, packed with pall rings
(r × H = 12.5 mm × 25 mm), were constructed from poly(methyl me-
thacrylate), with a total volume of 7.0 L and a working volume of 5.0 L.
Clear poly(methyl methacrylate) tubes, located at the top and bottom of
the reactor, were respectively used as the inlet and outlet for the biogas
(Fig. 1 (6)). The nutrient solution was circulated using a peristaltic
pump (LongerPump BT 300-2J), and the precise biogas flow was con-
trolled using a glass rotor flowmeter (LZB-4, China). A centrifuge pump
(MP-10RN, China) was used to mix the nutrient solution in the reactor
before sampling. The operations were carried out at a constant tem-
perature of 30 ± 1 °C.
2.2. Inoculum sources and biofilm culture
The inoculants were a mixture of an aerobic activated sludge from a
WWTP and anaerobic digestion slurry (Wang et al., 2015). In the
2
International Biodeterioration & Biodegradation xxx (xxxx) xxx–xxx
Fig. 1. Schematic diagram of the experimental reactor.
biomass immobilization stage, the mineral medium was used to culti-
vate and enrich functional microorganisms. The mineral medium was
replaced every day. Once the NO3− removal rate was stable, the raw
biogas was added into the reactor through the inlet (biogas flow
rate = 0.5 L/min). When the H2S RE reached stable values above 90%,
this operation was stopped.
The composition of the mineral medium was: Na2S2O3·5H2O (5.0 g/
L), KNO3 (4.0 g/L), KH2PO4 (2.0 g/L), NaHCO3 (1.0 g/L), MgCl2·6H2O
(0.5 g/L), FeSO4·7H2O (0.01), and trace element solution (1 mL) (Xu
et al., 2016). The pH of the medium was adjusted to 7.0 using aqueous
solutions of NaOH (1 mol/L) and HCl (1 mol/L) before the overall vo-
lume was increased to 1 L using tap water (Wang et al., 2015).
The composition of the trace element concentrate was as follows:
EDTA (0.5 g/L), FeSO4·7H2O (0.2 g/L), and 100 ml/L of trace element
solution SL-6. The trace element solution SL-6 was composed of the
following (g/L): ZnSO4·7H2O (0.1 g/L), MnCl2·4H2O (0.03 g/L), H3BO3
(0.3 g/L), CoCl2·6H2O (0.2 g/L), CuCl2·2H2O (0.01 g/L), NiCl2·6H2O
(0.02 g/L), and Na2MoO4·H2O (0.03 g/L) (Almenglo et al., 2016b).
2.3. Experimental setup
In this study, two reactors were set up, A0 (without inoculum) and A
(with inoculum). The inoculum of the BTF A0 was 5 L of tap water, and
the BTF A was inoculated with a mixture of anoxic sludge and tap
water. The study was divided into two stages. In stage I, the influent
was synthetic wastewater, and 1 L synthetic wastewater was exchanged
every day. In stage II, the influent was aerated biogas slurry, and 0.5 L
aerated biogas slurry was exchanged every day. The second stage was
initiated based on the promising results obtained from stage I. The
operating parameters for the two stages are shown in Table 1.
The biogas used in this study was obtained from an anaerobic di-
gester with a volume of 200 m3, which contained 66–69% (v/v) me-
thane, 25–30% carbon dioxide, and 0.2% H2S. The compositions of the
synthetic wastewater, aerated biogas slurry, and raw biogas slurry are
shown in Table 2.
2.4. Influence of the main operational variables
The LR, hydraulic retention time (HRT), elimination capability (EC),
empty bed residence time (EBRT), and RE are respectively derived
using Equations (5)–(9):
Y. Zeng et al. International Biodeterioration & Biodegradation xxx (xxxx) xxx–xxx

Table 1
The operating parameters.
Stage Influent Time Hydraulic retention time Biogas flow rate Empty bed residence time Volume of influent and effluent Loading rate
(d) (d) (L/min) (min) (mL−1) g-H2Sm−3h−1

Stage I Synthetic wastewater 1–10 5 0.5 14 1000 18.20

11–20 5 1 7 1000 36.20
21–30 5 2 3.5 1000 54.59
31–35 5 3 2.33 1000 72.88
36–40 5 4 1.75 1000 145.68

Stage II Aerated biogas slurry 41–50 6 0.2 35 500 7.30

51–60 6 0.5 14 500 18.20
61–70 6 0.8 8.75 500 29.21
71–80 6 1 7 500 36.20

Table 2 2.6. Statistical analysis

Composition of synthetic wastewater, aerated biogas slurry, and raw biogas
slurry. The sequence data were processed using QIIME Pipeline-Version
Composition Synthetic Aerated biogas Raw biogas slurry 1.7.0 (http://qiime.org/). All sequence reads were trimmed and as-
wastewater slurry signed to each sample based on their barcodes. All sequences were
clustered into operational taxonomic units (OTUs) at a 97% identity
COD (mg/L) 0 2600 ± 470 3460 ± 500
threshold, and the OTUs were generated for each sample and used for
NO3−-N (mg/L) 499 76 ± 5 0
NO2−-N (mg/L) 0 443.8 ± 10 0 statistical analysis. We calculated the alpha-diversity (Chao1 estimator,
NH4+-N (mg/L) 0 300 ± 50 1700 ± 100 observed species, and Shannon index) for which the rarefaction curves
SO42− (mg/L) 0 0 0 were generated from the observed species. Taxonomic analysis was
S0 (mg/L) 0 0 0 carried out using the Ribosomal Database Project (RDP) classifier. All
graphs were drawn using Origin 9.0.0 software.
Cin − Cout
RE (%) = × 100 3. Results and discussion
Cin (5)

Qg 3.1. Influence of loading rate on H2S removal efficiency

LR (g−H2 Sm−3h−1) = Cin ×
VB
EC ( g− H2 Sm−3h−1) = LR × RE
Vb
HRT (d) =
QL
Vb
EBRT (min) =
Qg
where Cin and Cout respectively denote the inlet and outlet concentra-
tions (g m−3) of H2S, Qg is the biogas volumetric flow (m3 h−1), VB is
the packed volume (m3), and QL is the liquid volumetric flow (m3 h−1).
2.5. Analytical methods
Samples of the liquid were collected from each reactor every day,
and nitrite and nitrate were analyzed by a colorimetric method (4500-
NO2− B) and an ultraviolet spectrophotometric screening method
(4500-NO3− B), respectively (Clesceri et al., 2012). Analysis of the
concentrations of S2− and SO42− employed HACH sulfide 1, 2-reagent
method 4 (HACH DR1900, USA), and HACH sulfaVer® 4-reagent
method (HACH DR1900, USA), respectively. The concentration of ele-
mental sulfur in each reactor was determined by high-performance li-
quid chromatography (HPLC) (Tichý et al., 1994). The obtained water
samples were mixed with CCl4 by 1:1 (v/v) and extracted for 10 min.
The bottom layer was filtered over a 0.45 μm organic membrane. Pure
methanol was used as the mobile phase, and the flow rate was
1 ml min−1. An InertSustain C18 column was used for separation and
analysis at 240 nm wavelengths. The concentration of biogas H2S was
measured using a portable biogas analyzer (Biogas Check, Geotech,
UK).
A MO-BIO PowerSoil® DNA isolation kit (American, 12888-50) was
used to extract genomic DNA from the samples. The extracted DNA was
diluted to a concentration of 10 μg μL−1 and stored at −80 °C for fur-
ther analyses.
(6)
In stage I, five different LRs were examined, ranging from 18.20 to
(7) 145.68 g-H2S m−3 h−1. The influence of the LR on the H2S RE is shown
in Fig. 2a. At an LR of 18.20 g-H2S m−3 h−1, the H2S RE in reactor A
was higher than 96% (EC more than 17.47 g-H2S m−3 h−1) during the
(8)
first 10 days. By contrast, the RE in control reactor A0 gradually in-
creased to 85% during the first 6 days, followed by a gradual drop to
60% by day 10. Hence, high RE in the control reactor was maintained
(9)
only for a few days. This can be explained by the fact that reactor A was
cultivated under constant conditions in terms of biomass immobiliza-
tion time, thereby allowing the functional microbes in reactor A to
adapt to the conditions. By contrast, reactor A0 relied exclusively on
physical absorption and/or chemical oxidation in tap water, and could
thus remove only a fraction of the H2S in the biogas. Fortuny et al.
(2008) also reported that the removal of H2S can be ascribed to the
physical absorption by a mineral medium and/or packing materials
before the reactor eliminates it. When the LR was 36.20 g-H2S m−3 h−1,
the EBRT decreased to 7 min, and the H2S RE gradually decreased from
90.0 to 73.5% in reactor A. Overall, reactor A maintained a high H2S RE
up to days 18–20. The reason for the decline in the H2S RE was the
increase in the flow rate, which caused a decrease in the EBRT. How-
ever, the H2S RE in reactor A0 declined rapidly to 10% in later periods.
This suggested that physical absorption and/or chemical oxidization in
tap water reached saturation in terms of their capacity for H2S removal
as the reaction progressed.
From day 21 onwards, the biogas flow rate and LR were increased to
2 L/min (EBRT = 3.5 min) and 54.59 g-H2S m−3 h−1, respectively. As a
consequence, the EBRT in the reactor decreased, causing a decrease in
the H2S RE. In reactor A, the RE of H2S fluctuated around 30% (EC
about 16.38 g-H2S m−3 h−1). As the LR was increased, the RE of H2S
decreased further to 20%, at which point it stabilized. When the LR
reached 145.68 g-H2S m−3 h−1 (the maximum LR examined in this
study), the poorest results in terms of the RE of H2S (16.8%) were ob-
served in reactor A. These results can probably be attributed to (i) the
high H2S LR exceeding the oxidation limit of the functional microbes

3
Y. Zeng et al.
Fig. 2. Influence of loading rate on H2S removal efficiency and elimination
capacity; (a) instantaneous hydrogen sulfide removal rate; (b) average hy-
drogen sulfide removal efficiency and elimination capacity in stage I; (c)
average hydrogen sulfide removal efficiency and elimination capacity in stage
II.
after the increase in the flow rate, or (ii) loss of microbial function as a
result of the biogas steam flushing down a portion of the microorgan-
isms on the packing material by daily outflow of the sludge out of the
reactor. According to the results collected in stage I and the improved
understanding of the influence of influent wastewater on the H2S RE,
several changes were made to the experimental parameters employed
in stage II (Table 1).
4
International Biodeterioration & Biodegradation xxx (xxxx) xxx–xxx
In stage II, four LR conditions were examined. At a biogas flow rate
of 0.2 L/min, i.e., EBRT of 35 min, LR of 7.30 g-H2S m−3 h−1, the H2S
RE in reactor A increased gradually to 94.0% (the maximum
EC = 6.86 g-H2S m−3 h−1). In reactor A0, the RE also increased but
only to 57.1% (the maximum EC = 4.17 g-H2S m−3 h−1). This likely
occurred because the high shock load (145.68 g-H2S m−3 h−1) led to
the inhibition of the functional microorganisms’ activity around days
36–40 in stage I. These functional microorganisms could have even-
tually recovered their activity when the LR was decreased to 7.30 g-H2S
m−3 h−1. At a biogas flow rate of 0.5 L/min, EBRT of 14 min, the LR
reached 18.20 g-H2S m−3 h−1 and the H2S RE in reactor A remained
stable at 95–97% (EC from 17.29 to 17.65 g-H2S m−3 h−1). Under the
same conditions, the H2S RE in reactor A0 fluctuated around 60% (EC
about 10.90 g-H2S m−3 h−1). Comparison of these results with those
obtained for stage I clearly revealed that the RE and EC of H2S were
always higher in the condition employing aerated biogas slurry (stage
II) than in the conditions utilizing synthetic wastewater (stage I) as the
nutrient source. At the LR of 29.21 g-H2S m−3 h−1, the EBRT decreased
because of the increased biogas flow rate (0.8 L/min). In turn, the H2S
RE decreased from 90.2 to 65.6% in reactor A. An even more marked
and rapid decrease to 20% was observed in reactor A0 under these
conditions. Following this decrease, the H2S RE remained rather low
but stable. At a biogas flow rate of 1 L/min, EBRT of 7 min, the RE of
H2S in reactor A fluctuated around 70%.
Some previous studies suggested that desulfurization performance
can be enhanced by implementing certain changes in filler type or in
the gas-to-liquid ratio of a reactor (Almenglo et al., 2016a; Deng et al.,
2009). For example, Lopez et al. (2016) found that increasing the
trickling liquid velocity from 4.4 to 18.9 m h−1 (gas-to-liquid from 0.15
to 0.28) resulted in a 10% improvement of the elimination capacity
when using a BTF for biogas desulfurization. Soreanu et al. (2008)
found that an increase in the surface-to-volume ratio of a packing
material enhanced the interphase pollutant partition and the effects of
microbial immobilization.
Fig. 2b and c shows the average RE and EC of H2S determined in
reactors A and A0 at different LRs in stages I and II, respectively. The
H2S RE and EC in reactor A were significantly higher than they were in
reactor A0, irrespective of the LR conditions. The H2S REs and ECs
detected for reactor A at the same LR but in different stages were
compared. Although the average EC of H2S at an LR of 36.20 g-H2S
m−3 h−1 was 30.67 g-H2S m−3 h−1 (average RE = 84.7%) in stage I, it
was only 22.42 g-H2S m−3 h−1 (average RE = 61.9%) in stage II. Si-
milarly, at an LR of 18.20 g-H2S m−3 h−1, the average EC was 17.50 g-
H2S m−3h−1 (average RE = 96%) in stage I, compared with 16.90 g-
H2S m−3 h−1 (average RE = 93%) in stage II.
The observed outcomes can most likely be explained by the pre-
sence of (i) some other microorganism genera derived from the aerated
biogas slurry that can inhibit the growth of the functional microbes in
the SDD system (Soreanu et al., 2008); (ii) organic compounds in the
aerated biogas slurry that are more likely to lose electrons than S2− ions
(Pokorna and Zabranska, 2015); or (iii) a large number of microbial
metabolites or other inorganic compounds including NH4+ ions derived
from the aerated biogas slurry that can inhibit the activity of functional
microbes (Mahmood et al., 2007b). The comparison of synthetic was-
tewater and other wastewater for biogas desulfurization in the litera-
ture is shown in Table 3. In summary, synthetic wastewater was more
suitable as a nutrient solution source for SDD than aerated biogas
slurry.
3.2. Comparative analysis of relevant ions in reactor A as a function of
loading rate
Analysis of the concentrations of relevant ions such as SO42−,
NO3−, and NO2− revealed that they underwent virtually no changes in
the control reactor A0 as a result of varying LR. In reactor A0, only S0
increased in concentration as the LR increased (data not shown), since
Y. Zeng et al. International Biodeterioration & Biodegradation xxx (xxxx) xxx–xxx

Table 3
Summary of literature on biogas biological desulfurization couple with denitrification.
Reference Type bed Fillers Electron Nutrient solution Inlet H2S concentration Maximum EC
(Bed volume) (Inoculation) acceptor

Bayrakdar et al. UFBR Granular activated carbon and NO3− Synthetic solution 7000 ± 1000 ppm 541 mg-H2S L−1 day−1
(2016) (0.25 L) activated sludge
Li et al. (2016) BTF Polypropylene NO3− Synthetic solution 1559 ± 41 ppm 54.5 g-H2S m−3h−1
(0.65 L)
Deng et al. (2009) Bubble column Rubber o-rings NO2−, NO3− Swine wastewater 1000–1500 ppm 23.27 mg-H2S L−1h−1
(3.94 L)
Almenglo et al. BTF Open-pore polyurethane foam NO3− Charge water in WWTP, 4100–7900 ppm 127.30 g-H2S m−3h−1
(2016b) (90 L) synthetic wastewater
Fernández et al. BTF Open-pore polyurethane foam NO3− Synthetic solution 1400–14,600 ppm 118.80 g-H2S m−3h−1
(2013) (2.4 L)
Pokorna et al. Scrubber Plastic carriers NO3− Synthetic solution 1976–8100 ppm 160 g-H2S m−3 day−1
(2015) (28.3 L)
Baspinar et al. Bioscrubbing bubble column NO2−, NO3− Charge water in WWTP, 13,000–37,000 ppm 78.85 g-H2S m−3h−1
(2011) (2.4 m3) synthetic wastewater
This study BTF Open-pore polyurethane foam NO2−, NO3− Biogas slurry 2000 ppm Average 25.37 g-H2S
(5.0 L) m−3h−1

physicochemical oxidation transforms H2S into S0 but not further to
SO42− (Duan et al., 2007).
The concentrations of SO42−, S0, NO3−-N, and NO2−-N detected in
reactor A utilizing synthetic wastewater and aerated biogas slurry as a
function of the influent are shown in Fig. 3. As is apparent from Fig. 3a,
the concentration of S0 in reactor A observed before day 20 was initially
Fig. 3. Analysis of relevant ions in reactor A with different influent; (a) changes
of sulfate and sulfur concentration; (b) changes of nitrate-N and nitrite-N
concentration.
low (∼2 mg/L). As the LR increased, the EBRT decreased. These
changes caused a rapid increase in S0 concentration, because H2S could
not be oxidized completely into SO42− by microbial oxidation (as the
threshold of biological oxidation was exceeded). Therefore, the in-
complete oxidation resulted in the accumulation of the intermediate
product S0. Simultaneously, the concentration of SO42− decreased and
then remained fairly stable, because the rates at which SO42− was
produced by biological oxidation and removed through daily drainage
were equal.
The changes in NO3−-N and NO2−-N concentrations displayed
stepwise changes with increasing LR, as can be seen in Fig. 3b. In stage
I, which employed synthetic wastewater as the influent, the con-
centration of NO3−-N ions increased gradually while that of NO2− ions
decreased. These changes can be explained by (i) the daily replacement
of the nutrient solution and (ii) the decrease in the contact time be-
tween H2S and functional microbes as a result of increased LR. There-
fore, the efficiency of denitrification also decreased as the LR increased.
According to a previous report (Mora et al., 2014), the complete de-
nitrification process can be described as NO3− → NO2− → N2, where
NO2− → N2 is the rate-limiting step (Baspinar et al., 2011). In this
study, the decrease in concentration of NO2− was consistent with the
decrease in the RE of H2S observed in stage I. These results suggest that
there is a certain connection between NO2− ions and the RE of H2S.
In stage II, which used aerated biogas slurry as the influent, the
concentration of NO3−-N ions underwent almost no changes at dif-
ferent LR conditions. By contrast, the concentration of NO2−-N ions
displayed clear and considerable changes. These patterns illustrate that
NO2−-N is utilized more easily as the electron acceptor than NO3−-N in
the SDD system. These findings were consistent with previous research
(Moraes et al., 2012). Additionally, the increasing LR and decreasing
EBRT caused a drop in the denitrification efficiency, despite the fact
that the aerated biogas slurry was replaced daily. Therefore, the NO2−-
N concentration increased in a stepwise manner until it stabilized.
There were virtually no NO2−-N ions in the synthetic wastewater in
stage I, and the NO3−-N concentration displayed no changes in stage II.
Therefore, the connection between the average RE of NOx−-N and the
LR was studied (Table 4). A decrease in the average RE of NOx−-N as a
result of increasing the LR was observed. This was consistent with the
decrease in the average RE of H2S observed in response to increasing
the LR. In order to ensure good desulfurization efficiency, however,
adequate NOx−-N must be provided. Zieminski and Kopycki (2016)
reported that the minimum concentration of NO3−-N in a nutrient so-
lution should be 50 mg/L in order to obtain the optimal desulfurization
efficiency.

5
Y. Zeng et al. International Biodeterioration & Biodegradation xxx (xxxx) xxx–xxx

Table 4 Thiobacillus genus (Zeng et al., 2018). Sulfurimonas was the second-
Average removal efficiency of NOX−-N in different influent. largest genus detected in stage I experiments. However, its content
LR Average RE of nitrate in Average RE of nitrite in stage decreased gradually with increasing LR; ultimately, this genus was the
(g-H2S m−3 h−1) stage I II third largest in the SDD system in stage II. These changes over time
(%) (%) could be explained by the fact that high concentrations of sulfide can
inhibit the denitrification and desulfurization processes (Cardoso et al.,
7.30 / 79.65
18.20 74.86 74.36
2006). Nevertheless, despite the variation in its contents, the Sulfur-
29.21 / 63.36 imonas genus played a significant part in the overall sulfur cycle (Han
36.20 60.93 49.17 and Perner, 2015). In general, the Thiobacillus and Sulfurimonas genera,
54.59 39.57 / which accounted for > 60% of the microbial communities in the SDD
72.88 29.12 /
system in stage I, were directly associated with the desulfur-
145.68 20.64 /
ization–denitrification performance (Prakash et al., 2012).
Comparison of the genera detected in stage II with those in stage I
3.3. Comparative analysis of microbial community in reactor A revealed marked changes in the dominant species. The Thiobacillus and
Sulfurimonas genera accounted only for a small part of the total. Instead,
In order to analyze the microbial community structures in each three new genera, namely Aequorivita, B-42, and SHD-231, were ob-
reactor, samples were periodically taken from the reactors exposed to served. Currently, their roles in the SDD system remain unknown. The
different LR conditions for sequencing of 16S rRNA genes. The results Methanosaeta genus, which was also detected, has been reported to be
confirmed that there was almost no growth of microorganisms in re- directly related to biogas production (Rotaru et al., 2014; Smith and
actor A0. Ingramsmith, 2007). Therefore, this genus may be derived from the
The microbial community analysis of the genus levels in reactor A is aerated biogas slurry. The Aequorivita, B-42, SHD-231, and Methano-
shown in Fig. 4. In stage I, Thiobacillus was the most dominant genus at saeta genera detected in stage II may inhibit the activity of the func-
all times (i.e., all LR conditions), accounting for 45–57% of the com- tional microbes as a result of space and nutrient limitations, thus af-
munities present overall. Zhao et al. (2016) reported that Thiobacillus fecting the observed H2S RE. Nevertheless, the relatively high H2S RE in
genus can oxidize sulfide to elemental sulfur, sulfate, or thiosulfate in stage II could be explained by (i) the LR in stage II being lower than that
the presence of nitrite, nitrate, or oxygen as the electron acceptor in stage I and (ii) the fact that physical absorption and chemical oxi-
source. Previous work by our team has demonstrated that the dominant dation had less-significant roles in the system employing synthetic
sulfur-oxidizing bacteria in SDD systems are always from the wastewater (stage I) than in that utilizing aerated biogas slurry (stage
II).
Fig. 5 shows the richness of the microbial communities observed in
the two stages, which reflects the diversity of the species at the alpha
level. In this figure, ‘Observed’ refers to the number of operational
taxonomic units (OTUs) detected; ‘Chao1’ provides an estimate of the
total number of species (in ecology, this is based on an estimate of rare
species); ‘Shannon’ is a measurement index based on information
theory, which provides information about the evenness of species
(higher numbers indicate higher community diversities); ‘Simpson’
provides an estimate of one of the microbial diversity indices of a
sample by taking into account both abundance and evenness of species;
and ‘PD’ represents a measure of phylogenetic diversity on the alpha
level.
The observed OTU decreased with increasing LR in both stages. This
observation could be explained by the gradual decrease in the con-
centration of microorganisms (i.e., death) that could not use the H2S
present in biogas and nitrate/nitrite ions to generate energy. In stage I,
the changes determined in OTU values (Observed) were in good
agreement with the richness estimated by Chao1. By contrast, very
different values were observed in stage II using these two analyses. The
observed differences could stem from the fact that Chao1 is often used
to estimate the total number of species based on an estimate of rare
species, whereas the microbial species present in the aerated biogas
slurry comprise a more complex mixture. The values obtained from
Shannon and Simpson analyses for the two stages were fairly consistent.
Nevertheless, the alpha-diversity values (Shannon and Simpson) were
higher in stage I than in stage II, most likely owing to the methanogens
and nitrifying bacteria present in the aerated biogas slurry. The changes
in PD were similar to those determined for Observed and Chao1 in stage
I. But at the LR of 18.20 g-H2S m−3 h−1 in stage II, it is possible that
some measurement errors at 18.20 g-H2S m−3 h−1.
This analysis revealed that the Thiobacillus and Sulfurimonas genera
were the main functional bacteria in the SDD system, and that the LR
had a significant influence on the structure of the microbial community.
The microbial community structure was more complex in a system
employing aerated biogas slurry as the nutrient solution for biogas
Fig. 4. Analysis of microbial community in reactor; A (a) analysis of genus desulfurization, compared with a system utilizing synthetic wastewater.
levels structure in stage I; (b) analysis of genus levels structure in stage II. Overall, these results suggest that the use of biogas slurry as the

6
Y. Zeng et al. International Biodeterioration & Biodegradation xxx (xxxx) xxx–xxx

Fig. 5. The microbial community analysis of alpha-diversity levels; (a) analysis of alpha-diversity levels in stage I; (b) analysis of alpha-diversity levels in stage II.

7
Y. Zeng et al.
Fig. 6. Integration of biogas desulfurization and biogas slurry denitrification.
nutrient source in biogas-desulfurization engineering should be ac-
companied by (i) an appropriate sterilization treatment (pasteurization)
or (ii) regular additions of functional bacteria.
3.4. Integrated simultaneous desulfurization and denitrification technology
and its potential applications
Many small- and medium-sized biogas plants have contributed to
the structure of energy in developing countries, such as Thailand,
Vietnam, etc. However, these plants often involve the use of chemical
desulfurization technology. The utilization of chemical desulfurizers
not only causes economic problems in the local area, but also pollutes
the environment. As an alternative, our research suggests utilizing the
SDD process, by reconstructing the chemical desulfurization tower into
a BTF and using part of the biogas slurry as a nutrient solution. With
this way, biogas can be effectively desulfurized, and nitrogen pollution
can be reduced. This technical route is illustrated in Fig. 6, and it can be
used as a guide for technicians.
However, this technical route still lacks a pilot plant for suitable
automation control technology. For example, it remains unclear when
the nutrient should be replaced and how the LR should be adjusted.
Research on the aspect will be helpful for improving the practical ap-
plicability of the SDD process. Especially, it will supply a feasible
method on biogas desulfurization for developing countries by small-
scale reconstruction.
4. Conclusions
The results of our feasibility study showed that the LR significantly
influences the H2S RE and microbial community structure when syn-
thetic wastewater and biogas slurry are utilized as electron acceptors.
For a given LR, using synthetic wastewater achieves better desulfur-
ization–denitrification efficiency than using biogas slurry as the elec-
tron acceptor. The results of this study indicated that appropriate
sterilization treatment or regular addition of functional bacteria may
improve the availability of biogas slurry. Furthermore, this study pro-
vided a feasible route to biogas desulfurization when biogas slurry is
used as a nutrient solution.
8
International Biodeterioration & Biodegradation xxx (xxxx) xxx–xxx
Acknowledgements
This work was supported by National Key Research and
Development Program of China (2017YFD0800803-02) and National
Natural Science Foundation of China (Grant No. 51408579).
Appendix A. Supplementary data
Supplementary data related to this article can be found at https://
doi.org/10.1016/j.ibiod.2018.05.012.
References
Almenglo, F., Bezerra, T., Lafuente, J., Gabriel, D., Ramirez, M., Cantero, D., 2016a. Effect
of gas-liquid flow pattern and microbial diversity analysis of a pilot-scale biotrickling
filter for anoxic biogas desulfurization. Chemosphere 157, 215–223.
Almenglo, F., Ramirez, M., Gomez, J.M., Cantero, D., 2016b. Operational conditions for
start-up and nitrate-feeding in an anoxic biotrickling filtration process at pilot scale.
Chem. Eng. J. 285, 83–91.
Arespacochaga, N.D., Valderrama, C., Mesa, C., Bouchy, L., Cortina, J.L., 2014. Biogas
deep clean-up based on adsorption technologies for Solid Oxide Fuel Cell applica-
tions. Chem. Eng. J. 255, 593–603.
Baspinar, A.B., Turker, M., Hocalar, A., Ozturk, I., 2011. Biogas desulphurization at
technical scale by lithotrophic denitrification: integration of sulphide and nitrogen
removal. Process Biochem. 46, 916–922.
Bayrakdar, A., Tilahun, E., Calli, B., 2016. Biogas desulfurization using autotrophic de-
nitrification process. Appl. Microbiol. Biotechnol. 100, 939–948.
Cardoso, R.B., Sierra-Alvarez, R., Rowlette, P., Flores, E.R., Gómez, J., Field, J.A., 2006.
Sulfide oxidation under chemolithoautotrophic denitrifying conditions. Biotechnol.
Bioeng. 95, 1148.
Chaiprapat, S., Mardthing, R., Kantachote, D., Karnchanawong, S., 2011. Removal of
hydrogen sulfide by complete aerobic oxidation in acidic biofiltration. Process
Biochem. 46, 344–352.
Clesceri, L., Greenberg, A.E., Eaton, A.D., 2012. Standard methods for examination of
water and wastewater. Health Lab. Sci. 4, 137.
Deng, L., Chen, H., Chen, Z., Liu, Y., Pu, X., Song, L., 2009. Process of simultaneous
hydrogen sulfide removal from biogas and nitrogen removal from swine wastewater.
Bioresour. Technol. 100, 5600–5608.
Duan, H., Yan, R., Koe, L.C., Wang, X., 2007. Combined effect of adsorption and biode-
gradation of biological activated carbon on Hydrogen sulfide biotrickling filtration.
Chemosphere 66, 1684–1691.
Fernández, M., Ramírez, M., Pérez, R.M., Gómez, J.M., Cantero, D., 2013. Hydrogen
sulphide removal from biogas by an anoxic biotrickling filter packed with Pall rings.
Chem. Eng. J. 225, 456–463.
Fortuny, M., Baeza, J.A., Gamisans, X., Casas, C., Lafuente, J., Deshusses, M.A., Gabriel,
D., 2008. Biological sweetening of energy gases mimics in biotrickling filters.
Chemosphere 71, 10–17.
Han, Y., Perner, M., 2015. The globally widespread genus Sulfurimonas: versatile energy
Y. Zeng et al.
metabolisms and adaptations to redox clines. Front. Microbiol. 6, 989.
Kao, C.-Y., Chiu, S.-Y., Huang, T.-T., Dai, L., Hsu, L.-K., Lin, C.-S., 2012. Ability of a
mutant strain of the microalga Chlorella sp. to capture carbon dioxide for biogas
upgrading. Appl. Energy 93, 176–183.
Lastella, G., Testa, C., Cornacchia, G., Notornicola, M., Voltasio, F., Sharma, V.K., 2002.
Anaerobic digestion of semi-solid organic waste: biogas production and its purifica-
tion. Energy Convers. Manag. 43, 63–75.
Li, D., Yuan, Z.H., Sun, Y.M., Ma, L.L., 2009a. China biogas resources: current status quo
of biogas resources and their perspective future of utilization in China. Mod. Chem.
Ind. 29, 1–5 7.
Li, W., Zhao, Q.L., Liu, H., 2009b. Sulfide removal by simultaneous autotrophic and
heterotrophic desulfurization-denitrification process. J. Hazard Mater. 162, 848–853.
Li, X., Jiang, X., Zhou, Q.Y., Jiang, W.J., 2016. Effect of S/N ratio on the removal of
hydrogen sulfide from biogas in anoxic bioreactors. Appl. Biochem. Biotechnol. 180,
930–944.
Lopez, L.R., Bezerra, T., Mora, M., Lafuente, J., Gabriel, D., 2016. Influence of trickling
liquid velocity and flow pattern in the improvement of oxygen transport in aerobic
biotrickling filters for biogas desulfurization. J. Chem. Technol. Biotechnol. 91,
1031–1039.
Mahmood, Q., Zheng, P., Cai, J., Wu, D., Hu, B., Islam, E., Rashid Azim, M., 2007a.
Comparison of anoxic sulfide biooxidation using nitrate/nitrite as electron acceptor.
Environ. Prog. Sustain. Energy 26, 169–177.
Mahmood, Q., Zheng, P., Cai, J., Wu, D., Hu, B., Li, J., 2007b. Anoxic sulfide biooxidation
using nitrite as electron acceptor. J. Hazard Mater. 147, 249–256.
Mora, M., Guisasola, A., Gamisans, X., Gabriel, D., 2014. Examining thiosulfate-driven
autotrophic denitrification through respirometry. Chemosphere 113, 1–8.
Moraes, B.S., Souza, T.S.O., Foresti, E., 2012. Effect of sulfide concentration on auto-
trophic denitrification from nitrate and nitrite in vertical fixed-bed reactors. Process
Biochem. 47, 1395–1401.
Pirolli, M., da Silva, M.L., Mezzari, M.P., Michelon, W., Prandini, J.M., Moreira Soares,
H., 2016. Methane production from a field-scale biofilter designed for desulfurization
of biogas stream. J. Environ. Manag. 177, 161–168.
Pokorna, D., Carceller, J.M., Paclik, L., Zabranska, J., 2015. Biogas cleaning by hydrogen
sulfide scrubbing and bio-oxidation of captured sulfides. Energy Fuel. 29, 4058–4065.
Pokorna, D., Zabranska, J., 2015. Sulfur-oxidizing bacteria in environmental technology.
Biotechnol. Adv. 33, 1246–1259.
Potumarthi, R., Anupoju, G.R., Mugeraya, G., Jetty, A., 2009. Hydrogen sulfide removal
in biofilter: evaluation of a new filter material by immobilization of thiobacillus sp.
Int. J. Chem. React. Eng. 7.
Prakash, O., Green, S.J., Jasrotia, P., Overholt, W.A., Canion, A., Watson, D.B., Brooks,
S.C., Kostka, J.E., 2012. Rhodanobacter denitrificans sp. nov., isolated from nitrate-
rich zones of a contaminated aquifer. Int. J. Syst. Evol. Microbiol. 62, 2457–2462.
Rattanapan, C., Kantachote, D., Yan, R., Boonsawang, P., 2010. Hydrogen sulfide removal
using granular activated carbon biofiltration inoculated with Alcaligenes faecalis
International Biodeterioration & Biodegradation xxx (xxxx) xxx–xxx
T307 isolated from concentrated latex wastewater. Int. Biodeterior. Biodegrad. 64,
383–387.
Rotaru, A.E., Shrestha, P.M., Liu, F., Shrestha, M., Shrestha, D., Embree, M., Zengler, K.,
Wardman, C., Nevin, K.P., Lovley, D.R., 2014. A new model for electron flow during
anaerobic digestion: direct interspecies electron transfer to Methanosaeta for the
reduction of carbon dioxide to methane. Energy Environ. Sci. 7, 408–415.
Smith, K.S., Ingramsmith, C., 2007. Methanosaeta, the forgotten methanogen? Trends
Microbiol. 15, 150–155.
Soreanu, G., Béland, M., Falletta, P., Edmonson, K., Seto, P., 2008. Investigation on the
use of nitrified wastewater for the steady-state operation of a biotrickling filter for the
removal of hydrogen sulphide in biogas. J. Environ. Eng. Sci. 7, 543–552.
Tichý, R., Janssen, A., Grotenhuis, J.T.C., Lettinga, G., Rulkens, W.H., 1994. Possibilities
for using biologically-produced sulphur for cultivation of Thiobacilli with respect to
bioleaching processes. Bioresour. Technol. 48, 221–227.
Vikromvarasiri, N., Champreda, V., Boonyawanich, S., Pisutpaisal, N., 2017. Hydrogen
sulfide removal from biogas by biotrickling filter inoculated with Halothiobacillus
neapolitanus. Int. J. Hydrogen Energy 42, 18425–18433.
Wang, L., Wei, B., Chen, Z., Deng, L., Song, L., Wang, S., Zheng, D., Liu, Y., Pu, X., Zhang,
Y., 2015. Effect of inoculum and sulfide type on simultaneous hydrogen sulfide re-
moval from biogas and nitrogen removal from swine slurry and microbial me-
chanism. Appl. Microbiol. Biotechnol. 99, 10793–10803.
Watsuntorn, W., Ruangchainikom, C., Rene, E.R., Lens, P.N.L., Chulalaksananukul, W.,
2017. Hydrogen sulfide oxidation under anoxic conditions by a nitrate-reducing,
sulfide-oxidizing bacterium isolated from the Mae Urn Long Luang hot spring,
Thailand. Int. Biodeterior. Biodegrad. 124, 196–205.
Xu, J.L., Fan, Y.M., Li, Z.X., 2016. Effect of pH on elemental sulfur conversion and mi-
crobial communities by autotrophic simultaneous desulfurization and denitrification.
Environ. Technol. 37, 3014–3023.
Zeng, Y., Zhou, J., Yan, Z., Zhang, X., Tu, W., Li, N., Tang, M., Yuan, Y., Li, X., Cao, Q.,
Huang, Y., 2018. The study of simultaneous desulfurization and denitrification pro-
cess based on the key parameters. Process Biochem. https://doi.org/10.1016/j.
procbio.2018.04.007.
Zhang, J., Li, L., Liu, J., Han, Y., 2016. Temporal variation of microbial population in
acclimation and start-up period of a thermophilic desulfurization biofilter. Int.
Biodeterior. Biodegrad. 109, 157–164.
Zhao, Y.G., Zheng, Y., Tian, W., Bai, J., Feng, G., Guo, L., Gao, M., 2016. Enrichment and
immobilization of sulfide removal microbiota applied for environmental biological
remediation of aquaculture area. Environ. Pollut. 214, 307–313.
Zhou, Q., Liang, H., Yang, S., Jiang, X., 2015. The removal of hydrogen sulfide from
biogas in a microaerobic biotrickling filter using polypropylene Carrier as packing
material. Appl. Biochem. Biotechnol. 175, 3763–3777.
Zieminski, K., Kopycki, W.J., 2016. Impact of different packing materials on hydrogen
sulfide biooxidation in biofilters installed in the industrial environment. Energy Fuel.
30, 9386–9395.