Sei sulla pagina 1di 9

Bioresource Technology 288 (2019) 121518

Contents lists available at ScienceDirect

Bioresource Technology
journal homepage:

Natural freezing-thawing pretreatment of corn stalk for enhancing anaerobic T

digestion performance

Juan Lia, Akiber Chufo Wachemoa,b, Hairong Yuana, Xiaoyu Zuoa, Xiujin Lia,
Department of Environmental Science and Engineering, Beijing University of Chemical Technology, No. 15 Beisanhuan East Road, Chaoyang District, Beijing 100029, PR
Department of Water Supply and Environmental Engineering, Arba Minch University, P.O. Box 21, Arba, Ethiopia



Keywords: Natural freezing-thawing (NFT) was proposed as a low energy input and alternative pretreatment method for
Natural freezing-thawing high biomethane production from corn stalk (CS) by anaerobic digestion (AD). The CS was pretreated by
Biomethane freezing-thawing in winter season using different pretreatment time periods (7d, 14d, 21d and 28d) and solid-to-
Corn stalk liquid ratios (1:2, 1:4, 1:6, 1:8 and 1:10). The results showed that CS pretreated for 21d coupled with a solid-to-
Anaerobic digestion
liquid ratio of 1:6 achieved the best result among all pretreatment conditions. In this case, the biomethane yield
and VS removal rate of CS reached the highest values of 253 mL·gvs−1 and 58.6%, respectively, which were
40.5% and 27.4% higher than that of the untreated. It was also found that the predominant bacterial and
archaeal at genus level in AD were Clostridium_sensu_stricto_1 (36.1%) and Methanobacterium (54.0%), re-
spectively. This study provided that NFT is a simple pretreatment strategy for efficient AD bioconversion of CS to

1. Introduction and heating, automobile fuel, for power generation and petrochemical
uses (Ma et al., 2017). Anaerobic digestion (AD) has been considered as
The search for sustainable and alternative energy is of critical im- an environmentally friendly method for producing biomethane from
portance with ever-growing energy demands and environmental con- agricultural biomass, which has high content of lignocellulose and is
cerns, together with diminishing fossil fuel reserves (Tadesse and widely distributed in the world (Kumar and Sharma, 2017;
Luque, 2011). One example of these alternatives is the widespread use Sawatdeenarunat et al., 2018).
of biomethane, which can be applied in city gas supplies for cooking However, the recalcitrant structure of agricultural biomass limits its

Corresponding author.
E-mail address: (X. Li).
Received 24 March 2019; Received in revised form 15 May 2019; Accepted 17 May 2019
Available online 19 May 2019
0960-8524/ © 2019 Elsevier Ltd. All rights reserved.
J. Li, et al. Bioresource Technology 288 (2019) 121518

utilization in biomethane production, resulting in a low energy yield technology. This method does not use any chemicals or catalysts, which
and level of digestibility (Chen et al., 2018). Hydrolysis is often be- makes it an environmentally-friendly and economically feasible pre-
lieved to be the rate-limiting step in the AD of lignocellulosic biomasses treatment method. This method can be used in accordance with natural
(Dahadha et al., 2017). Therefore, to increase the efficiency of AD in seasonal changes when the partially moistened biomass is left in the
treating biomass and improve the energy yield, an efficient pretreat- field and the moisture within the biomass freezes and thaws due to
ment process is required to enhance the digestibility of lignocellulosic variations in atmospheric temperature (Liu et al., 2017). In this case,
biomass by microorganisms (Zheng et al., 2014). Over years, several there is no additional energy input into the pretreatment process. The
pretreatment processes have been developed and researched and can be objectives of this study were to: (1) investigate natural temperature
categorized as physical, chemical and biological pretreatment processes changes and its effects on the degradation of lignocellulosic compo-
(Mulat et al., 2014; Yun et al., 2016). Among these methods, physical nents in CS; (2) evaluate the capability of NFT pretreatment of CS for
pretreatments are some of the most basic that can be used to pretreat enhancing AD performance and determine optimal pretreatment con-
lignocellulosic substrates, and include comminution (e.g. milling and ditions; and (3) analyze the influence of NFT pretreatment on microbial
grinding), steam-explosion (autohydrolysis), liquid hot water pretreat- community structure changes in digester.
ment (hydrothermolysis), extrusion, heating, and irradiation (ultra-
sound and microwave) (Baral and Shah, 2017). Liu et al. (2012) pre- 2. Materials and methods
treated corn stalk using microwaves at 35 °C and 680 W intensity for
24 min, which achieved 20% more methane (CH4) production than 2.1. Feedstock and inoculum
untreated samples. Menardo et al. (2012) used conventional heating to
treat rice straw under 120 °C for 30 min, however, the methane pro- The sample biomass material of CS was collected in October 2018
duction showed no change as compared to untreated samples. Applying from a farmland at Beijing suburb of Shunyi District, China. The col-
a steam explosion to treat wheat straw under 180 °C for 15 min and lected CS was ground to the size of 20-mesh using a lab mill (YSW-180,
2 MPa, Bauer et al. (2009) obtained a 19.7% increase of methane pro- Zhengde, China). The role of grinding was a physical size reduction to
duction in relation to untreated samples. Notably, the associated with increase the surface area of the CS for the subsequent pretreatment
physical pretreatment rely on energy inputs, require high capital costs process. The inoculum was collected from a continuously operated
and limited the methane yields. Thus, developing a green, efficient and stable biogas plant under mesophilic condition at Shunyi District,
economically feasible pre-treatment approach is essential and pros- Beijing, China. The main components and characteristics of the CS and
pected to avoid or reduce the shortcomings of physical methods. inoculum are listed in Table 1.
Recently, the method of freezing and thawing pretreatment with
simple operating procedures has been applied using relatively new
2.2. Pretreatment process
approaches at low temperature to produce glucose and ethanol from
biomass materials. For instance, the yields of cellulose rich fractions
The pretreatment experiment took place during the winter season of
dissolved from bamboo at −12 °C ranged from 75.1% to 77.7% in the
2018–2019 in Beijing, China. The CS collected in the 2018 autumn
study conducted by Li et al. (2010). Smichi et al. (2016) found that the
season was naturally frozen outdoors from December 10, 2018 to
maximum concentration of glucose was released from J. maritimus
January 8, 2019. The humid stalk was packed in sealed bags and was
(53.78 ± 3.24 g L−1) by freezing-thawing pretreatment. Zhao et al.
left out in the field. The bags were then collected after every 7d and the
(2008) obtained a 70% glucose yield from spruce at a low pretreatment
treated-CS naturally dried indoors for later AD use. Because this was
temperature (−15 °C). Particularly, in 2014–2016, Rooni et al. (2017)
first time field pretreatment was utilized, the moisture content, ac-
pretreated barley straw in field to produce glucose and ethanol, and the
cording to different solid-to-liquid ratios was considered in the natural
highest hydrolysis efficiency was 10.28% from straw which had been
freezing-thawing process. Therefore, five solid-to-liquid ratios based on
gathered in March and stored in a swathe. Moreover, freezing and
the dry weight of CS were selected to investigate the effect of moisture
thawing pretreatment has also been researched on the pretreatment of
content on the biomethane production of CS.
sludge for enhancing biodegradation or AD efficiency (Cano et al.,
The specific pretreatment parameters are shown in Table 2, and the
2015; Jan et al., 2008; Yang et al., 2015). For example, in the study of
temperature of the natural environment was recorded every after every
Montusiewicz et al. (2010), the sludge was frozen at −25 °C for 24 h in
2 h by an automatic temperature recorder (RC-4, Jiangsu, China) during
a laboratory freezer and then thawed for another 12 h at 20 °C in the
the entire freezing-thawing stage. It automatically recorded the outdoor
indoor air. The biogas yield obtained from frozen-thawed sludge was
temperature, and then the recorded temperature data was imported
1.31 m3·kg−1 of removed VS. It is clear consensus that pretreatment at
into the computer via a Universal Serial Bus (USB) connection.
low temperatures may play an effective role at enhancing the solubility
of lignocellulosic fiber or sludge for bioenergy productions. However,
Table 1
freezing and thawing pretreatment has not yet been applied for pre-
Characteristics of corn stalk and inoculum.*
treating agricultural biomass in natural condition for increasing bio-
methane production. Parameter Value (%)
In addition, microorganisms are critical to the AD process as they
Corn stalk Inoculum
convert the feedstock to biomethane, therefore, the feedstock condition
could directly affect the microbial community structure, diversity and Total solids (%) 93.12 ± 0.23 10.24 ± 0.10
activity of specific populations (Gannoun et al., 2016). It is also known Volatile solids (%)a 86.54 ± 0.17 4.23 ± 0.02
MLSS (g L−1)a ND 105.84 ± 1.28
that pretreatment methods could significantly influence the condition
Total carbon (%)b 41.36 ± 0.35 28.21 ± 0.04
of feedstock. In the study of Zhou et al. (2017), the overall performance Total nitrogen (%)b 0.92 ± 0.01 2.31 ± 0.01
of waste activated sludge digestion was significantly dependent on the Cellulose (%)b 42.33 ± 0.12 ND
initial chemical pretreatments, which in turn influenced and was re- Hemicellulose (%)b 23.57 ± 0.97 ND
lated to the microbial community structures. According to the litera- Lignin (%)b 9.47 ± 0.06 ND
LCH (%)b 74.81 ± 0.21 ND
ture, there are no reported investigations into the correlation between
microbial community structure and conversion of feedstock pretreated * Values are means ± SD (n = 3).
by freezing-thawing method has never been studied in literature. a
Content of fresh matter.
Therefore, in this work, natural freezing-thawing (NFT) was used to b
Content of dry matter; ND: not determined; MLSS: Mixed Liquid Suspended
pretreat corn stalk (CS) for biomethane production through AD Solids; LCH: Total of Lignin, Cellulose and Hemicellulose.

J. Li, et al. Bioresource Technology 288 (2019) 121518

Table 2
The results of fiber analysis of CS after pretreatment and AD.*
Time Condition After pretreatment (%) After AD (%)

Solid-to-liquid ratio Lignin Hemicellulose Cellulose Lignin Cellulose Hemicellulose

7d 1:6 9.29 ± 0.01 23.19 ± 1.04 41.15 ± 0.27 8.04 ± 0.07 13.29 ± 0.08 8.02 ± 0.08
14d 1:6 9.23 ± 0.04 22.86 ± 0.92 40.70 ± 0.22 8.02 ± 0.04 12.04 ± 0.07 7.65 ± 0.06
21d 1:6 9.16 ± 0.03 22.37 ± 0.81 40.38 ± 0.31 7.87 ± 0.05 10.01 ± 0.06 6.96 ± 0.07
28d 1:6 9.19 ± 0.07 22.34 ± 0.95 40.36 ± 0.24 8.19 ± 0.07 13.73 ± 0.08 7.61 ± 0.07
28d 1:2 9.41 ± 0.05 23.34 ± 0.73 41.37 ± 0.30 8.14 ± 0.05 13.27 ± 0.08 8.05 ± 0.08
28d 1:4 9.38 ± 0.06 23.12 ± 0.76 41.60 ± 0.29 8.11 ± 0.04 13.52 ± 0.05 7.75 ± 0.06
28d 1:6 9.29 ± 0.06 22.48 ± 0.76 41.38 ± 0.29 8.02 ± 0.05 12.14 ± 0.04 7.45 ± 0.05
28d 1:8 9.32 ± 0.03 22.91 ± 0.84 41.49 ± 0.36 8.30 ± 0.06 14.60 ± 0.07 8.18 ± 0.08
28d 1:10 9.33 ± 0.01 23.07 ± 0.75 41.93 ± 0.28 8.41 ± 0.03 14.73 ± 0.09 8.32 ± 0.09
Untreated 9.47 ± 0.06 23.57 ± 0.97 42.33 ± 0.35 8.45 ± 0.08 17.23 ± 0.09 10.19 ± 0.09

* Values are means ± SD (n = 3).

2.3. Batch fermentation process 2013):

In this study, the complete experimental set-up consisted of a 1 L P(t ) = Pm exp ⎧−exp ⎡ (λ − t ) + 1⎤ ⎫
anaerobic digester with a working volume of 0.8 L, a 1 L gas collecting ⎨
⎩ ⎢
⎣ Pm ⎥
⎦⎬⎭ (1)
bottle, and a 1 L beaker. AD took place in a water bath at a constant where P(t) is the cumulative CH4 production (mL CH4) at time t; Pm is
incubation temperature (35 ± 1 °C). The organic loading of 50 g (total the maximum CH4 potential (mL CH4) at the end of incubation time; t is
solid, TS) L−1 pretreated CS was applied. The anaerobic inoculum time (d); Rm is the CH4 production rate (mL CH4/d); λ is the lag phase
seeded in each reactor was 105 g (mixed liquid suspended solid, MLSS) (d) and e is the base of natural logarithms, i.e.2.71828. The three
L−1 (Guan et al., 2018). Subsequently, deionized water was added to parameters Pm, Rm, and λ were estimated with the Nonlinear Curve Fit
each digester to maintain the same working volume of 0.8 L. The di- program in OriginPro 9.0.
gesters were immediately sealed using rubber septa with drilled holes to Each analytical result was the mean value of measurements on the
sure that the anaerobic environment was maintained, and then they replicate samples. The standard deviations and statistical differences
were placed in a water bath. The untreated corn stalk was used as the were analyzed by Microsoft Excel 2013. All the figures in this paper
control. The total duration of AD was 55 days. Each test condition run were drawn by OriginPro 9.0.
in triplicate and the results presented here are the mean values of the
three tests.
2.5. Microbial community analysis

2.4. Analytical methods and data analysis The microbial community structures in each anaerobic digester
treating CS without pretreatment and with NFT pretreatment were in-
2.4.1. Biomethane production analysis vestigated. The functional roles of specific microbial populations were
The daily biomethane production from each digester was recorded evaluated in relation with biomethane production in each digester. The
using the water displacement method. The biogas composition (H2, sample for deoxyribonucleic acid (DNA) extraction was prepared and
CH4, CO2, and N2) was determined using a gas chromatograph (GC) analyzed based on the method adapted from the manufacturer protocol
(SP2100, BeiFen-RuiLi, Beijing, China) equipped with a molecular sieve of Fast DNA Spin Kit (MP Biomedicals, U.S.A.). The sample (slurry) for
(TDX-01) packed in a 2 m × 3 mm stainless-steel column and a thermal DNA analysis was collected from the effluent of AD reactors. The V3–V4
conductivity detector (TCD). The temperatures of the oven, injector region of the bacterial 16S ribosomal ribonucleic acid (RNA) gene was
port, and TCD were 140 °C, 150 °C, and 150 °C, respectively. Argon was amplified by Polymerase Chain Reaction (PCR) (95 °C for 3 min, fol-
used as the carrier gas at flow rate of 30 mL·min−1. Daily biomethane lowed by 27 cycles at 95 °C for 30 s, 55 °C for 30 s, and 72 °C for 45 s and
production was determined from methane content and the amount of a final extension at 72 °C for 10 min) using primers 338F (5′-ACTCCT
daily biogas produced after converting to standard temperature and ACGGGAGGCAGCAG-3′) and 806R (5′-GGACTACHVGGGTWTCT-
pressure conditions, and the corresponding cumulative biomethane AAT-3′). In the same way the archaeal 16S ribosomal RNA gene were
production was calculated. amplified by PCR (95 °C for 3 min, followed by 33 cycles at 95 °C for
30 s, 55 °C for 30 s, and 72 °C for 45 s and a final extension at 72 °C for
2.4.2. Chemical compositions and physical structure analyses 10 min) using primers 524F_10_ext (5′-TGYCAGCCGCCGCGGTAA)-3′)
The TS, volatile solids (VS) and MLSS of the CS, inoculum and their and Arch 958R (5′-YCCGGCGTTGAVTCCAATT-3′). The PCR reactions
mixture were determined according to the American Public Health were performed in triplicate using 20 μL mixture which contained 4 μL
Association (APHA) Standard Methods (1998). The total carbon (TC) of 5× FastPfu Buffer, 2 μL of 2.5 mM dNTPs, 0.8 μL of Forward Primer
and total nitrogen (TN) were determined by the Vario EL/micro cube (5 μM), 0.8 μL of Reverse Primer (5 μM), 0.4 μL of FastPfu Polymerase,
elemental analyzer (Elementar, Germany). The composition of cellu- 0.2 μL of BSA and 10 ng of template DNA. Double-distilled water
lose, hemicelluloses and lignin in the CS and digestate were measured (ddH2O) was then added to complete the 20 μL mixture.
using a fiber analyzer (A2000I, ANKOM, USA) (Guan et al., 2018). Amplicons were extracted from 2% agarose gels and purified using
Scanning electron microscope (SEM, S-4700 N, Hitachi, Japan) was the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City,
used to scan the microstructures of treated and untreated CS samples. CA, U) according to the manufacturer’s instructions and quantified
The CS after pretreatment was dried in an oven at 105 °C for 24 h prior using QuantiFluor™-ST (Promega, U.S.). The purified amplicons were
to SEM analysis. pooled in equimolar and paired-end sequenced (2 × 250) on an
Illumina MiSeq platform according to the standard protocols. The raw
2.4.3. Data analyses reads were deposited into the NCBI Sequence Read Archive (SRA) da-
The modified Gompertz equation was used to describe the cumu- tabase. These data were analyzed by Shanghai Major Bio-Pharm
lative CH4 production curves of AD as shown in Eq. (1) (El-Mashad, Technology Co. Ltd. (Shanghai, China).

J. Li, et al. Bioresource Technology 288 (2019) 121518

Fig. 1. Changes of temperature in field during 28 days in 2018–2019 winter.

3. Results and discussion 3.2. The changes of lignocellulosic components after pretreatment

3.1. Changing of natural temperatures The biomethane production potential and suitability of the specific
biomass can be characterized on the basis of its relative lignocellulose
The CS sample was naturally frozen and thawed outdoor for 28 d (LCH) content (Kikas et al., 2016). The untreated CS collected in Oc-
from December 10, 2018 to January 8, 2019. During the entire pre- tober 2018 had the highest cellulose (42.33%), hemicellulose (23.57%)
treatment period, the outdoor temperatures were recorded as shown in and lignin content (9.47%) (Table 1). The compositions of LCH in the
Fig. 1. It was observed that the lowest and highest temperatures were CS after NFT pretreatment are listed in Table 2.
−13 °C and 10 °C, respectively. The freezing/thawing time for CS was Notably, the reduction of cellulose was more than hemicellulose and
basically between 8:00 am and 8:00 pm (Fig. 1a). The average tem- lignin, while the reductions from total LCH was less than 5%. The de-
perature distributed mainly from −5.5 °C to 2.0 °C (Fig. 1b). In addi- crease in cellulose content was due to the fact that freezing and thawing
tion, every day had temperatures above and below 0 °C, which is fa- can strengthen the hydrogen bonding (HB) force between water and
vorable to the freezing and thawing process. cellulose, whilst the HB force inside the cellulose itself is weakened,
These findings demonstrate that the natural winter environment thereby leading to the reduction in the degree of polymerization and
could replace a freezer as a natural cold container without any energy physical degradation of cellulose (Mohsenzadeh et al., 2012). This is
input. In these natural conditions, temperature fluctuations during the advantageous to dissolving the polymer, and as a result, the dense and
frozen stage and/or thawed stage are often long-standing, and they can hard lignocellulose becomes loose (Li et al., 2018). The internal area of
lead to ice recrystallization and cellular damage, which can affect the samples and the volume of holes are conducive to the attachment,
structure of lignocellulosic materials (Zhang et al., 2018). growth and reproduction of microorganisms (Smichi et al., 2016). This
would be significant for the biomass conversion and therefore, the
improvement of biomethane production. Obviously, the lowest content

J. Li, et al. Bioresource Technology 288 (2019) 121518

Fig. 2. Daily biomethane production of CS pretreated by NFT with different conditions ((a) different days; (b) different solid-to-liquid ratios).

of cellulose was still above 40%, indicating that due to its abundance, 14.9%, 21.0%, and 26.3% higher than that of the untreated group,
cellulose was the major source of methane production. respectively.
In addition, all of the solid-to-liquid ratio groups displayed a rapid
3.3. Anaerobic digestion performance of CS pretreated by NFT rise in CMP after 10–15 days, and the reason may result from the five
ratio groups operated under the 28d pretreatment condition. It is said
3.3.1. The trend of biomethane production that the readily degradable fraction of organic matter may be washed
The pretreatment experiments in field and solid-to-liquid ratios away if the biomass stayed on the hydrolysis stage for a long period of
were conducted in order to assess the effects of NFT on the biomethane time (Nkemka and Murto, 2010). However, the CMP of the 1:2–1:10
production processes from CS and select an optimal pretreatment con- groups were 11.4%–25.3% higher than that of untreated, among which,
dition. The trend in daily biomethane production (DMP) of the NFT the ratio of 1:6 presented better fruition (8268 mL) (Fig. 3b).
pretreated CS is illustrated in Fig. 2, which show similar tendencies The biodegradability and digestion efficiency of the substrate was
with all experimental groups. Three or four peaks of DMP appeared also analyzed by T80 (Fig. 3), where T80 is defined as the time required
during the period of anaerobic digestion, and then the production of for achieving 80% of the total biomethane production (Liu et al., 2019).
daily biomethane slowly approached zero over time. The T80 for 7 d, 14 d, 21 d and 28d pretreated CS were 28 ± 2 d,
Among these peaks of different pretreatment time (Fig. 2a), the 28 ± 1 d, 26 ± 1 d and 29 ± 1 d (Fig. 3a), respectively, which was
DMP at first peak reached 486 mL, 584 mL, 636 mL, and 275 mL for 7d, 12.1%–21.2% shorter than that of the untreated group (33 ± 2 d). The
14d, 21d and 28d pretreated groups, respectively, which were T80 for 1:2, 1:4, 1:6, 1:8 and 1:10 pretreated CS were 3.0%–12.1%
103.2%–371.1% higher than that of the untreated (135 mL). Besides, shorter than that of the untreated group (Fig. 3b). T80 for 21d and 1:6
the time of the second peaks for pretreated groups (7d, 14d, 21 and pretreated CS groups were shortest along with better digestion effi-
28d) were at 14th–18th days, which were 1–4 days shorter than that of ciency. This also suggests that NFT pretreatment under 21 d combined
the untreated (19 days). These implied that the bio-digestibility of CS with 1:6 solid-to-liquid ratio condition will have significant economic
was improved by NFT pretreatment, among which 21d group produced benefits because the shorter time would lower both operation and
the best result. Likely, among different solid-to-liquid ratio groups, the management costs.
first peak of 1:6 achieved 449 mL, which were 16.0%–74.0% higher These results mean that NFT pretreatment was capable of improving
than that of the 1:2, 1:4, 1:8 and 1:10 pretreated groups (258 mL, the biodegradability of CS and enhancing the biomethane production in
428 mL, 387 mL and 377 mL) (Fig. 2b). This suggested that the 21d and AD process. Moreover, the result was also applicable for practical large-
1:6 groups exhibited better biodegradability than other ratios under scale AD plants because there is no energy and chemical regent input
natural condition. into the pretreatment process. In addition, the collected CS can be once
pretreated using freezing-thawing method in winter season, and then
fermented for biomethane production. The AD plants operated in
3.3.2. Cumulative volume of biomethane at technical digestion time (T80)
spring, summer and autumn can also fed with the CS sample pretreated
It was also observed that all the NFT pretreated CS achieved higher
by NFT in winter.
cumulative biomethane production (CMP) than the untreated CS
(Fig. 3). For the 21d and 14d pretreated groups, the CMP rapidly in-
creased from the day 5 to day 26, and then the trend gradually slowed 3.3.3. Kinetic parameters of biomethane production analyses
down when the digestion time exceeded 27 (Fig. 3a). On the other side, The CMP curves of the NFT pretreated and untreated groups were
the groups with 7 d, 28 d pretreatments and the untreated showed a selected and fitted by the Gompertz model (Fig. 3) and the simulation
rapid rise in CMP after 15d, 10d and 20 d, respectively, which were results were listed in Table 3. The maximal daily biomethane produc-
5–15 d later than other pretreatment groups. The highest CMP of pre- tion (Rm) means the efficiency of AD and the highest Rm values of 7 d,
treated CS achieved at 21 d was 9005 mL, which was 35.8% higher than 14 d, 21 d and 28 d pretreatment condition were 252 mLN·d−1,
that of the untreated CS (6633 mL). This result was 15.3% higher than 253 mLN·d−1, 300 mLN·d−1 and 331 mLN·d−1, which was higher than
the result of Shetty et al. that was from rice straw pretreated with 5% that of untreated by 4.1–36.2%. In addition, the lag-phase time (λ)
sodium hydroxide (Shetty et al., 2017). The CMP of CS produced by 21 signifies that the start speed of the experiment is another important
d pretreatment was also higher than the result of Yang et al. (2014) by indicator (Pan et al., 2019). The calculated λ values of all the NFT
43.2%, who pretreated CS using 4% ammonia. This suggests that NFT pretreated groups was shorter by 17.2–97.2% than that of untreated.
pretreatment can achieve a better effect on biomethane production as This illustrated that NFT pretreatment can improve the biodegrad-
compared to chemical pretreatment. Besides, the CMP of 7 d, 14 d and ability of CS and thus accelerate the AD process. R2 represents the
28d pretreated CS were 7623 mL, 8027 mL and 8380 mL, which were correlation degree of fitting (Wei et al., 2015) and the values ranged

J. Li, et al. Bioresource Technology 288 (2019) 121518

Fig. 3. Cumulative biomethane production of CS pretreated by NFT with different conditions and curve fitting of modified Gompertz model to biomethane pro-
duction data ((a) different days; (b) different solid-to-liquid ratios).

Table 3 was the destruction of the crystallization zone, which was limited by
Kinetic parameters of biomethane production using modified Gompertz equa- the dispersed ice crystals and weak shearing force in the biomass.
tion. The rate of VS removal is an important parameter in measuring the
Condition Maximal daily Bacteria Predicted R2 extent of biodegradability in AD, and it can be used to evaluate the
biomethane growth lag biomethane efficiency of the specific pretreatment (Fig. 4c, d). The VS removal rate
production (Rm) time (λ) (d) production potential for all pretreated CS was 53.4%–58.6%, which was 15.2%–27.4%
(mLN·d−1) (Pm) (mLN) higher than that of the untreated (46.3%). The highest VS removal rate
7d 8142.6 4.1 252 0.9942
was achieved with 21d pretreatment when the highest methane yield
14d 8763.1 1.2 253 0.9947 was also produced. This indicates that the higher VS removal rate had
21d 9371.8 1.2 300 0.9959 higher biomethane production and further highlights the determination
28d 8342.6 10.6 331 0.9940 that NFT pretreatment is a promising method for converting raw CS to a
1:2 8181.0 5.3 253 0.9951
substrate susceptible to AD.
1:4 8679.6 6.1 270 0.9956
1:6 8666.2 2.6 256 0.9982
1:8 8628.9 3.5 202 0.9944 3.4. Physical structure changes of the NFT pretreated CS
1:10 8276.8 1.0 186 0.9943
Untreated 7062.4 12.8 243 0.9940
The effects of NFT pretreatment on the cellular integrity of CS were
illustrated by SEM pictures. The SEM result showed that some struc-
from 0.9942 to 0.9959 for all pretreatment conditions, which were turally intact, orderly arranged fiber bundles were extremely smooth
relatively better than untreated sample (R2 was 0.9939). Specifically, for the untreated CS. Besides, slight ruptures are clearly visible when
the 21d pretreatment group achieved the highest fit level among others. the CS was gradually frozen and thawed to the 14 d. After14 d, the
These results were basically in accordance with the experimental value, surface structure of CS was further destroyed during the natural
which illustrates that the Gompertz growth equation fit the experiment freezing and thawing process because of the shearing force of the
very well (Hu et al., 2015). crystal. Additionally, different solid-to-liquid ratios have varying effects
on the surface structure of CS. According to the SEM pictures, when the
solid-liquid ratio was 1:6, the surface structure of CS was damaged most
3.3.4. Unit VS biomethane yield and VS removal rate with many micro pores. In terms of the 1:8 and 1:10 solid-to-liquid
Unit VS biomethane yield (MY-VS) was used to further compare the ratios, the micro pores were not larger than that of the 1:6 solid-to-
energy conversion efficiency and AD performance. The enhancements liquid ratio. This was attributed to the dispersed ice crystals and weak
of MY-VS over time for the different tests are illustrated (Fig. 4). The shearing force in the biomass, which resulted in the incompletely de-
results showed that during a long period of natural freezing-thawing struction of CS (Hassasroudsari and Goff, 2012). Bhattacharya et al.
pretreatment time, the MY-VS of CS increased up to the highest yield of (1988) also reported that the average size of ice crystals influenced by
253 mL·gvs−1 on 21 d (Fig. 4a). The result was 40.5% higher than that water distribution, and a certain size of ice crystal could make lig-
of untreated (180 mL·gvs−1). With the time of freezing-thawing time nocellulosic materials more susceptible to denaturation and fragmen-
prolonged to 28th day, the MY-VS of CS decreased by 11.4% compared tation.
to the result of 21 d. This was attributed to the reduction of degraded The analysis of the surface structure indicated that the NFT pre-
substances during the long-term freezing-thawing process, which re- treated CS samples suffered severe damage with optimal periods (21d)
sulted in loss of dry matter along with a decrease of LCH (Table 2) and solid-to-liquid ratio (1:6). This caused the original cellulose and
(Nishino et al., 2010). The MY-VS of CS in different solid-to-liquid ra- hemicellulose structures to become loose and porous, which enhanced
tios (1:2, 1:4, 1:6, 1:8 and 1:10) were 210, 219, 231, 212 and the AD ability of the CS. This was also why the highest biomethane
206 mL·gvs−1, respectively (Fig. 4b). Obviously, 1:6 pretreatment group production was obtained by the pretreatment conditions in nature.
achieved the highest biomethane yield, which was 28.3% higher than
that of untreated (180 mL·gvs−1). A decrement in MY-VS can be seen in 3.5. Conversion of lignocellulosic components
the CS when the moisture content was too low or high. The reason for
this occurrence was that too low or high moisture in the biomass will Biomethane is generated from the biological conversion of carbon-
result in smaller or fewer crystalline crystals, which make the CS could based substances. LCH is the main carbon source for anaerobic micro-
not be easily degraded (Hassasroudsari and Goff, 2012). Another cause organisms during AD processing of CS. The conversion rates of LCH

J. Li, et al. Bioresource Technology 288 (2019) 121518

Fig. 4. Unit VS biomethane yield (MY-VS) and VS removal rate of untreated and natural freezing and thawing pretreated CS ((a) MP-VS of different days; (b) MP-VS
of different solid-to-liquid ratios; (c) VS removal rate of different days; (d) VS removal rate of solid-to-liquid ratios).

during the process of AD were used to evaluate the digestion perfor- diversity of bacteria and archaeal communities was supported by an
mance (Li et al., 2017). The content of LCH after AD were analyzed in operational taxonomic unit (OUT) number. For the three samples ob-
Table 2. In the NFT experiment, the proportions of lignin, cellulose and tained from untreated, 21 d and 1:6 groups, the total 16S rRNA gene
hemicellulose decreased to 7.87%, 10.01% and 6.96%, respectively, sequencing gave the range from 423 to 449 and 28 to 30 for bacterial
which were 6.8%, 41.9% and 31.7% lower than that of untreated. and archaeal OTUs, respectively. Based on the Ace and Chao 1 values
Concurrently, the respective conversion rates of lignin, cellulose, and (Table 4), bacterial and archaeal richness of NFT pretreated samples
hemicellulose were 18.5%, 76.4% and 69.8% for CS pretreated by 21d. (21 d and 1:6 groups) were higher than untreated sample. The alpha
This demonstrates that the highest value obtained at 21d pretreatment diversity indexes such as the Shannon and Simpson indexes which fo-
group, and therefore, it had the highest methane yield of all results. In cuses on the evenness of microbial diversity were calculated for each
addition, the high conversion rate of cellulose (65%–75%) and hemi- sample. Based on the results, bacterial Shannon index of 21d pretreated
cellulose (64%–69%) in all pretreated CS after AD indicates that natural sample (4.40) and 1:6 ratio pretreated sample (4.52) were lower than
freezing and thawing could effectively improve the conversion rate of the untreated group (4.59). The bacterial Simpson index of 21 d pre-
the main components in CS. In general, the natural freezing-thawing treated and at 1:6 ratio pretreated samples was also lower than the
pretreatment is a green and very promising pretreatment method for untreated one. On the other hand, archaea Shannon index of each
enhancing the digestibility of CS during anaerobic digestion, which sample showed similar trends. This indicates that the digester of 21 d
could also be easily applied at larger scales. pretreated CS had the higher richness, and higher diversity of bacterial
and archaeal groups. Various studies also reported that microbial
3.6. Changes in the microbial community structure community descriptors such as diversity and evenness indices are di-
rectly related to anaerobic digester functions (Carballa et al., 2011; Liu
3.6.1. Alpha diversity indices of bacterial and archaeal 16S rRNA gene et al., 2018).
The correlations between the function of each microbial group and 3.6.2. The distributions of bacterial sequences
their performances in different digesters were investigated. The en- In the above mentioned three digesters feed with untreated, 21d-
richment of certain active microorganisms in a digester show that the treated and 1:6 ratio treated CS, the dynamic profiles of bacterial phyla
sensitivity of the anaerobic process for different pretreated CS directly varied significantly during the feeding phase (Fig. 5a). The dominant
affects biomethane production process (Lin et al., 2016). Therefore, a phyla in three digesters were Bacteroidetes and Firmicutes, which are
comparison of the microbial richness and diversity in the different di- important acid hydrolytic microorganisms (Zhi et al., 2019). Although
gesters and statistical analysis of bacterial and archaeal communities in the major phyla coexisted in the two communities (Bacteroidetes and
each AD effluent are shown in Table 4. Firmicutes), their respective abundances were very different. The
As can be seen from Table 4, the coverage index was more than abundances of Bacteroidetes in three digesters were in the range from
99.8% for all samples, thereby confirming representativeness. The 8.4% to 18.1%. Firmicutes were the most predominant phylum in three

J. Li, et al. Bioresource Technology 288 (2019) 121518

Table 4
Alpha diversity indices of the bacterial and archaeal 16S rRNA gene sequences.
Groups Samples Coverages (%) OTU Ace Chao 1 Shannon Simpson

Bacterial Untreated 0.99935 423 441 441 4.59 0.0421

21d 0.99953 449 463 465 4.40 0.0247
1:6 0.99941 436 457 460 4.52 0.0346
Archaea Untreated 1.00000 28 29 29 1.65 0.3796
21d 1.00000 30 31 31 1.52 0.2532
1:6 0.99998 28 30 30 1.58 0.2863

digesters while 21d treated-CS group showed maximum proportion of 2018).

76.3%. Meanwhile, a number of other phyla presented in each reactor The dynamic profiles of the archaeal community at the genus level
may have played important roles, although they had low percentage were significantly different in all reactors (Fig. 5d). In the AD system,
(> 1%), such as Proteobacteria, Fibrobacteres and Spirochaetae. more than 70% of methane in biogas comes from the cleavage of
The genus level in three samples was observed; the result showed acetate. Two common genera are known to use acetate as a substrate for
that the dynamic profiles of bacterial genus differed significantly in methanogenesis, namely, Methanosarcina and Methanosaeta (Zhang
three digesters after AD (Fig. 5b). In all digesters, the total bacterial et al., 2019). The Methanobacterium species known to date are hy-
reads were classified into 20 genera, out of these 17 bacterial genera drogenotrophic methanogens, which are able to utilize carbon dioxide
displayed a relative abundance > 1.0% in all samples which were (CO2) and hydrogen (H2) as substrates for methanogenesis (Tejerizo
considered as high-rank groups. Among the dominant genera, Clos- et al., 2017). As it can be seen in Fig. 5d, the three archaeal commu-
tridium_sensu_stricto_1 (36.1 ± 5.9%), Terrisporobater (9.8 ± 2.3%), nities in digester for three groups demonstrated a clear dominance of
Moblitalea (3.7 ± 5.3%) and Ruminofilibacter (3.7 ± 1.2%) were the Methanobacterium (32.9%–54.0%), Methanosaeta (30.2%–32.3%),
more prevalent than others. Therefore, the genus Clostridium_sensu_s- and Methanosarcina (5.9–14.6%). The number of Methanobacterium
tricto_1 was the most abundant bacterial group in the digester for un- species in three digesters was higher than Methanosarcina and Metha-
treated, 21d and 1:6 ratio groups, where its specific abundance was nosaeta species. This is possibly because the acetate was almost con-
43.4%, 33.6% and 36.8%, respectively. In addition, the bacterial di- sumed at the final stage of AD and resulted the thriving of genus Me-
versity in 21d pretreated CS digester was higher than the other two thanobacterium. It was observed that the proportion of
groups, which was attributed to the high conversion rate of lig- Methanobacterium species in the NFT pretreated (21d) groups was
nocellulosic components that left more degradable substance in the higher than that of untreated samples and 1:6 ratio groups. The data
digester. As a result, the diversity of bacteria increased in the final stage presented in this study showed that NFT pretreatment had great effects
of AD. on microbial community structures in AD digester. It behaved like a
similar microbial community and dominant microorganisms, but the
population levels changed significantly.
3.6.3. The distributions of archaeal sequences
The distributions of archaeal sequences at the phylum level from
each sample are shown in Fig. 5c. Bathyarchaeota and Euryarchaeota 4. Conclusions
were the most dominant groups among the archaeal phyla with an
average relative abundance of 23.7 ± 2.2% and 75.1 ± 2.7%, re- This study proved that natural freezing-thawing was a practical
spectively across all digesters. Archaea are believed to be responsible method to pretreat corn stalk for efficient bioconversion. The corn
for many biochemical reactions in the methanogenic phase (Ko et al., stalks pretreated by natural freezing-thawing achieved 16.7%–40.5%

Fig. 5. Bacterial and archaeal sequence distributions at phylum (a, c) and genus (b, d) level after AD.

J. Li, et al. Bioresource Technology 288 (2019) 121518

more biomethane production, 65%–75% more cellulose, and 64%–69% of the surface acid sites of tungsten trioxide for highly selective hydrogenation of
more hemicellulose conversions, as compared to the untreated one. The cellulose to ethylene glycol. Bioresour. Technol. 264, 58–65.
Liu, T., Xu, X., Jie, Y., 2017. Experimental study on the effect of freezing–thawing cycles
highest biomethane production of 253 mL·gvs−1 was obtained when CS on wind erosion of black soil in Northeast China. Cold Reg. Sci. Technol. 36, 1–8.
was frozen and thawed for 21 d along with a 1:6 solid-to-liquid ratio. Liu, T., Zhou, X., Li, Z., Wang, X., Sun, J., 2019. Effects of liquid digestate pretreatment on
biogas production for anaerobic digestion of wheat straw. Bioresour. Technol. 280,
The predominant bacterial and archaeal communities at genus level in 345–351.
AD were Clostridium_sensu_stricto_1 (36.1%) and Methanobacterium Liu, W., Ma, H., Cao, C., Yang, Z., Zhao, M., Kong, X., Hu, X., 2012. Effects of pretreat-
(54.0%), respectively. NFT could provide one of alternative methods for ments with steam-explosion using solar energy and microwave irradiation on biogas
production of corn stalk. Trans. Chinese Soc. Agric. Eng. 28, 227–234.
high-efficient biomethane production from CS. Li, Y., Liu, C., Wachemo, A.C., Yuan, H., Zou, D., Liu, Y., Li, X., 2017. Serial completely
stirred tank reactors for improving biogas production and substance degradation
during anaerobic digestion of corn stover. Bioresour. Technol. 235, 380–388.
Declaration of Competing Interest
Ma, C., Zhang, S., Dong, R., Wang, M., Jia, W., Lu, Z., 2017. Corn stalk fiber-based bio-
mass brick reinforced by compact organic/inorganic calcification composites. ACS
The authors declare no competing financial interest. Sustain. Chem. Eng. 6, 2086–2093.
Menardo, S., Airoldi, G., Balsari, P., 2012. The effect of particle size and thermal pre-
treatment on the methane yield of four agricultural by-products. Bioresour. Technol.
Acknowledgements 104, 708–714.
Mohsenzadeh, A., Jeihanipour, A., Karimi, K., Taherzadeh, M.J., 2012. Alkali pretreat-
ment of softwood spruce and hardwood birch by NaOH/thiourea, NaOH/urea,
The authors are grateful to the fund supports from National Key NaOH/urea/thiourea, and NaOH/PEG to improve ethanol and biogas production. J.
Technologies R & D Program of China (No. 2018YFC1900903). Chem. Technol. Biotechol. 87, 1209–1214.
Montusiewicz, A., Lebiocka, M., Rożej, A., Zacharska, E., Pawłowski, L., 2010. Freezing/
thawing effects on anaerobic digestion of mixed sewage sludge. Bioresour. Technol.
Appendix A. Supplementary data 101, 3466–3473.
Mulat, D.G., Ward, A.J., Adamsen, A.P.S.N., Voigt, V., Nielsen, J.L., Feilberg, A., 2014.
Quantifying contribution of synthrophic acetate oxidation to methane production in
Supplementary data to this article can be found online at https:// thermophilic anaerobic reactors by membrane inlet mass spectrometry. Environ. Sci. Technol. 48, 2505–2511.
Nishino, N., Harada, H., Sakaguchi, E., 2010. Evaluation of fermentation and aerobic
stability of wet brewers’ grains ensiled alone or in combination with various feeds as
References a total mixed ration. J. Sci. Food Agric. 3, 557–563.
Nkemka, V.N., Murto, M., 2010. Evaluation of biogas production from seaweed in batch
tests and in UASB reactors combined with the removal of heavy metals. J. Environ.
Apha, A.W., 1998. Standard Methods for the Examination of Water and Wastewater.
Manage. 91, 1573–1579.
American Water Works Association, pp. 130.
Pan, J., Ma, J., Liu, X., Zhai, L., Ouyang, X., Liu, H., 2019. Effects of different types of
Baral, N.R., Shah, A., 2017. Comparative techno-economic analysis of steam explosion,
biochar on the anaerobic digestion of chicken manure. Bioresour. Technol. 275,
dilute sulfuric acid, ammonia fiber explosion and biological pretreatments of corn
stover. Bioresour. Technol. 232, 331–343.
Rooni, V., Raud, M., Kikas, T., 2017. The freezing pre-treatment of lignocellulosic ma-
Bauer, A., Bösch, P., Friedl, A., Amon, T., 2009. Analysis of methane potentials of steam-
terial: a cheap alternative for Nordic countries. Energy 139, 1–7.
exploded wheat straw and estimation of energy yields of combined ethanol and
Sawatdeenarunat, C., Nam, H., Adhikari, S., Sung, S., Khanal, S.K., 2018. Decentralized
methane production. J. Biotechnol. 142, 50–55.
biorefinery for lignocellulosic biomass: Integrating anaerobic digestion with ther-
Bhattacharya, M., Hanna, M.A., Mandigo, R.W., 1988. Lipid oxidation in ground beef
mochemical conversion. Bioresour. Technol. 250, 140–147.
patties as affected by time-temperature and product packaging parameters. J. Food
Shetty, D., Kshirsagar, J.P., Tapadia-Maheshwari, S., Lanjekar, V., Singh, S.K.,
Sci. 53, 714–717.
Dhakephalkar, P.K., 2017. Alkali pretreatment at ambient temperature: a promising
Carballa, M., Smits, M., Etchebehere, C., Boon, N., Verstraete, W., 2011. Correlations
method to enhance biomethanation of rice straw. Bioresour. Technol. 226, 80–88.
between molecular and operational parameters in continuous lab-scale anaerobic
Smichi, N., Messaoudi, Y., Moujahed, N., Gargouri, M., 2016. Ethanol production from
reactors. Appl. Microbiol. Biotechnol. 89, 303–314.
halophyte Juncus maritimus using freezing and thawing biomass pretreatment.
Cano, R., Pérez-Elvira, S.I., Fdz-Polanco, F., 2015. Energy feasibility study of sludge
Renewable Energy 8, 1357–1361.
pretreatments: a review. Appl. Energy 149, 176–185.
Tadesse, H., Luque, R., 2011. Advances on biomass pretreatment using ionic liquids: an
Chen, F., He, H., Zhao, S., Yao, J., Sun, Q., Huang, G., Xiao, D., Tang, L., Leng, Y., Tao, X.,
overview. Energy Environ. Sci. 4, 3913–3929.
2018. Analysis of microbial community succession during methane production from
Tejerizo, G.T., Kim, Y.S., Maus, I., Wibberg, D., Winkler, A., Off, S., Pühler, A., Scherer, P.,
baiyinhua lignite. Energy Fuel 32, 10311–10320.
Schlüter, A., 2017. Genome sequence of Methanobacterium congolense strain
Dahadha, S., Amin, Z., Bazyar, Lakeh, A.A., Elbeshbishy, E., 2017. Evaluation of different
Buetzberg, a hydrogenotrophic, methanogenic archaeon, isolated from a mesophilic
pretreatment processes of lignocellulosic biomass for enhanced biomethane produc-
industrial-scale biogas plant utilizing bio-waste. J. Biotechnol. 247, 1–5.
tion. Energy Fuel 31, 10335–10347.
Wei, Y., Li, X., Yu, L., Zou, D., Yuan, H., 2015. Mesophilic anaerobic co-digestion of cattle
El-Mashad, H.M., 2013. Kinetics of methane production from the codigestion of switch-
manure and corn stover with biological and chemical pretreatment. Bioresour.
grass and Spirulina platensis algae. Bioresour. Technol. 132, 305–312.
Technol. 198, 431–436.
Gannoun, H., Omri, I., Chouari, R., Khelifi, E., Keskes, S., Godon, J.J., Hamdi, M., Sghir,
Yang, C., Liu, W., He, Z., Thangavel, S., Wang, L., Zhou, A., Wang, A., 2015. Freezing/
A., Bouallagui, H., 2016. Microbial community structure associated with the high
thawing pretreatment coupled with biological process of thermophilic Geobacillus sp.
loading anaerobic codigestion of olive mill and abattoir wastewaters. Bioresour.
G1: acceleration on waste activated sludge hydrolysis and acidification. Bioresour.
Technol. 201, 337.
Technol. 175, 509–516.
Guan, R., Li, X., Wachemo, A.C., Yuan, H., Liu, Y., Zou, D., Zuo, X., Gu, J., 2018.
Yang, D., Pang, Y., Yuan, H., Chen, S., Ma, J., Yu, L., Li, X., 2014. Enhancing biogas
Enhancing anaerobic digestion performance and degradation of lignocellulosic
production from anaerobically digested wheat straw through ammonia pretreatment.
components of rice straw by combined biological and chemical pretreatment. Sci.
Chin. J. Chem. Eng. 22, 576–582.
Total Environ. 637, 9–17.
Yun, H., Liang, B., Qiu, J., Zhang, L., Zhao, Y., Jiang, J., Wang, A., 2016. Functional
Hassasroudsari, M.H., Goff, D., 2012. Ice structuring proteins from plants, mechanism of
characterization of a novel amidase involved in biotransformation of triclocarban and
action and food application. Food Res. Int. 46, 425–436.
its dehalogenated congeners in ochrobactrum sp. TCC-2. Environ. Sci. Technol. 51,
Hu, Y., Pang, Y., Yuan, H., Zou, D., Liu, Y., Zhu, B., Chufo, W.A., Jaffar, M., Li, X., 2015.
Promoting anaerobic biogasification of corn stover through biological pretreatment
Zhang, B., Zhang, X.L., Shen, C.L., Deng, S.G., 2018. Understanding the influence of
by liquid fraction of digestate (LFD). Bioresour. Technol. 175, 167–173.
carrageenan oligosaccharides and xylooligosaccharides on ice-crystal growth in
Jan, T.W., Adav, S.S., Lee, D.J., Wu, R.M., Su, A., Tay, J.H., 2008. Hydrogen fermentation
peeled shrimp (Litopenaeus vannamei) during frozen storage. Food Funct. 9,
and methane production from sludge with pretreatments. Energy Fuel 22, 98–102.
Kikas, T., Tutt, M., Raud, M., Alaru, M., Lauk, R., Olt, J., 2016. Basis of energy crop
Zhang, W., Chen, B., Li, A., Zhang, L., Li, R., Yang, T., Xing, W., 2019. Mechanism of
selection for biofuel production: cellulose vs. lignin. Int. J. Green Energy 13, 49–54.
process imbalance of long-term anaerobic digestion of food waste and role of trace
Ko, J.H., Wang, N., Yuan, T., Lü, F., He, P., Xu, Q., 2018. Effect of nickel-containing
elements in maintaining anaerobic process stability. Bioresour. Technol. 275,
activated carbon on food waste anaerobic digestion. Bioresour. Technol. 266,
Zhao, Y.L., Wang, Y., Deng, Y.L., 2008. Enhanced enzymatic hydrolysis of spruce by al-
Kumar, A.K., Sharma, S., 2017. Recent updates on different methods of pretreatment of
kaline pretreatment at low temperature. Biotechnol. Bioeng. 99, 1320–1328.
lignocellulosic feedstocks: a review. Bioresour. Bioprocess. 4, 1–19.
Zheng, Y., Zhao, J., Xu, F., Li, Y., 2014. Pretreatment of lignocellulosic biomass for en-
Li, M., Fan, Y., Xu, F., Sun, R., Zhang, X., 2010. Cold sodium hydroxide/urea based
hanced biogas production. Prog. Energy Combust. Sci. 42, 35–53.
pretreatment of bamboo for bioethanol production: characterization of the cellulose
Zhi, S., Li, Q., Yang, F., Yang, Z., Zhang, K., 2019. How methane yield, crucial parameters
rich fraction. Ind. Crop. Prod. 32, 551–559.
and microbial communities respond to the stimulating effect of antibiotics during
Lin, Q., Vrieze, J.D., He, G., Li, X., Li, J., 2016. Temperature regulates methane pro-
high solid anaerobic digestion. Bioresour. Technol. 283, 286–296.
duction through the function centralization of microbial community in anaerobic
Zhou, A., Zhang, J., Varrone, C., Wen, K., Wang, G., Liu, W., Wang, A., Yue, X., 2017.
digestion. Bioresour. Technol. 216, 150–158.
Process assessment associated to microbial community response provides insight on
Liu, C., Wachemo, A.C., Tong, H., Shi, S., Zhang, L., Yuan, H., Li, X., 2018. Biogas pro-
possible mechanism of waste activated sludge digestion under typical chemical pre-
duction and microbial community properties during anaerobic digestion of corn
treatments. Energy 137, 457–467.
stover at different temperatures. Bioresour. Technol. 261, 93–103.
Li, N., Ji, Z., Wei, L., Zheng, Y., Shen, Q., Ma, Q., Tan, M., Zhan, M., Zhou, J., 2018. Effect