Performance and microbial communities of a batch anaerobic reactor treating liquid and high-solid sludge at thermophilic conditions

Abstract

To investigate the balance between volumetric methane production and substrate loading on anaerobic digestion (AD), long-term batch experiments were conducted at different total solids (TS) contents of 5% (R1), 10% (R2), 15% (R3) and 20% (R4), respectively. The results showed that maximum cumulative methane production (233 mL g_VS⁻¹) and VS removal efficiency (48.49%) was achieved in R1. Volatile fatty acids (VFAs) accumulation in R4 was 57.2% higher than that in R2 and 5.69 times higher than that in R1. Glucose degradation rate (GDR) for R3 and R4 declined fast indicating lower microbial activities than those in other two reactors. Microbial community analysis revealed that Coprothermobacter for proteins degradation and Methanosarcina for methanogenesis were enriched in R1 with relative abundance of 25.0% and 34.5%, respectively. Hydrogenotrophic methanogens was dominant in treatment of high-solid AD, in accordance with VFA accumulation. The assessment of methane production performance showed similar reactor utilization efficiency between R1 and R4.

---

1. Introduction

Massive increase of sewage sludge from municipal wastewater treatment plants (WWTPs) has caused giant environmental risks as well as enormous costs for disposal.¹ From the perspective of energy recovery, anaerobic digestion (AD) has been considered as one of the most widely-used technologies for efficient stabilization of sludge and production of renewable energy source.² However, some limitations restricted the application of AD, including long sludge retention time (SRT), large-volume reactor and difficulties in professionally operating to keep a stable status.³ Moreover, for most WWTPs in China, sewage sludge has already been dewatered to a moisture content of about 80% for the convenience of transportation and centralized treatment.⁴ Thus, high-solid anaerobic digestion (HS-AD), by which the biological waste were treated with total solids (TS) content of more than 15% (w/w), has provided a practical option for sludge disposal.⁵ Due to its advantages of higher volumetric processing capacity for the equivalent solids loading, smaller reactor volume, fewer residual products and less consumption for diluting, heating and mixing, HS-AD for sludge stabilization has attracted increasing attentions.⁶

The feasibility of HS-AD has already been proved according to previous studies,⁷ however, the process was at the expense of unsatisfactory stability performance as well as decrease in methane yield, mainly on account of excess release and accumulation of intermediate products. Benbelkacem et al. operated the HS-AD of municipal solid waste with TS over 22%, observing the low level of biochemical methane potential (by unit mass), which was ascribed to the inhibition of volatile fatty acids (VFAs) (mostly propionate) during the hydrolysis step.⁸ Study conducted by Duan et al. suggested that HS-AD could support a higher organic loading rate (OLR), however, significant inhibition was also observed with free ammonia-nitrogen (FAN) concentration over than 600 mg L⁻¹, leading to the drop of volatile solids reduction (VS reduction) (below 30% when OLR achieving 4 kg_VS m⁻³ d⁻³).⁴ In addition, it was generally believed HS-AD was more difficult to run and control at start-up stage due to the lower mass transfer efficiency.^9,10 Currently, several strategies were taken to improve the HS-AD performance, including alkaline or/and thermal pretreatment,¹¹ anaerobic co-digestion with other feedstocks,¹² inoculation at ideal ratio¹⁰ and even the addition of trace elements for microbes.¹³ Though the strategies mentioned had been proved positive for HS-AD, most of the studies simply focused on process optimization while the inhibition of methanogenic activity in HS-AD still remains unclear, especially in thermophilic condition with a high hydrolysis rate.¹⁴ The knowledge of response relationship among TS contents, compositions of organic matters, variations of VFAs and FAN in supernatant were very limited when discussing the differences between liquid anaerobic digestion (L-AD) and high-solid anaerobic digestion (HS-AD) for sludge in previous studies. Besides, detailed comparison of microbial communities and diversities (for both bacteria and archaea) between L-AD and HS-AD remain to be clarified. Moreover, most of current studies were mainly focused on the short-term response of HS-AD, and the useful information on long-term impacts was definitely limited.

The aim of this study was to investigate the effects of different TS contents on the performance of sludge anaerobic digestion under thermophilic conditions. Comparison between HS-AD and L-AD was assessed with special attention paid to composition of organic matters, VFA accumulation as well as pH variation and free-ammonia inhibition at different stages in the long-term AD process. Furthermore, the acidogenesis activities and microbial communities were respectively analyzed by testing degradation rate of specific substrate as well as high-throughput sequencing technology, which both provided a deep insight into the evolution of microorganisms and further explanation of the effects of different TS contents on HS-AD system.

2. Materials and methods

2.1 Characteristics of substrates and inoculum

The sludge samples that dewatered by centrifugation, were obtained from Minhang municipal WWTP in Shanghai, China. The collected dewatered sludge was stored at 4 °C for analysis and preheated to 55 °C before mixed with inoculum. The inoculum was collected from an anaerobic digester of previous long-term thermophilic lab-scale AD system. The characteristics of dewatered sludge and inoculum are summarized in Table 1.

Table 1 Characteristics of substrates and inoculums^a

Parameters	Dewatered sludge	Seed sludge
a TS, total solid; VS, volatile solid; SCOD, soluble chemical oxygen demand; VFAs, volatile fatty acids; TAN, total ammonia nitrogen; FAN, free ammonia nitrogen.
pH	6.37–6.42	7.19–7.23
TS (%, w/w)	20.53–20.71	12.19–12.58
VS/TS (%)	70.01–70.21	59.91–60.82
SCOD (mg L^−1)	1282–1301	7134–7162
Soluble VFAs (mg L^−1)	66.9–68.0	4440.0–4456.4
Soluble proteins (mg L^−1)	61.3–63.1	271.9–275.8
Soluble carbohydrates (mg L^−1)	81.6–84.2	343.5–352.1
TAN (mg L^−1)	269.4–272.3	1338.6–1341.5
FAN (mg L^−1)	2.5–2.7	78.3–82.7

---

2.2 Batch experiment

Batch experiments were carried out in identical horizontal stainless-steel reactors with working volume of 7.2 L. Aim to evaluate the effects of different TS contents on thermophilic AD, dewatered sludge was adjusted to designed TS level (5%, 10%, 15% and 20% for R1, R2, R3 and R4, respectively) and then mixed with inoculum at a ratio of 5 : 1 (based on dry weight). The reactors were purged with nitrogen gas (99.99%) for 10 min to remove the head-space oxygen. During the digestion, all the reactors were at a thermophilic temperature of 55.0 ± 1 °C and equipped with helix-type stirrers, which could run for 10 min and stand by for 50 min periodically at a rate of 35 rpm. The biogas was measured by a gas counter and then collected into gasbag for further analysis.

2.3 Analytical methods

Substrate samples extracted from the reactors were analyzed for pH, TS and VS, according to the Standard Methods.¹⁵ The supernatant of sludge was obtained by centrifugation at 12 000 rpm for 5 min and then filtered through a microfiber membrane (0.45 μm). The filtrate was collected for analyzing soluble chemical oxidation demand (SCOD), soluble proteins, soluble carbohydrates and total ammonia nitrogen (TAN) in accordance with Standard Methods.¹⁵ FAN concentration was calculated as described by Østergaard.¹⁶ The filtrate was acidified by phosphoric acid before VFAs (including acetic acid, propionic acid, iso-butyric acid, butyric acid, iso-valeric acid and valeric acid) were analyzed by a Shimadzu gas chromatograph (GC-2010), which was equipped with a chromato-graphic column (30 m × 0.25 mm × 0.25 mm) and a flame ionization detector (FID).¹⁷ The methane concentration was detected by GC-2010 with a 1.5 m stainless-steel column (Molecular Sieve, 80/100 mesh) and a thermal conductivity detector (TCD). The results of soluble proteins, soluble carbohydrates and VFAs were calculated in terms of COD by unit dry mass of sludge.

2.4 Glucose degradation rate

Glucose degradation rate (GDR) was applied to determine the activities of the microbial communities.¹⁸ To remove the soluble carbohydrates from fermented sludge, 5.0 mg of the samples from each reactor were washed with 0.9% stroke-physiological saline solution for 3 times and then centrifuged at 5000 rpm for 3 min. Glucose solution (100 mg L⁻¹, 20 mL) was added in the sediment as substrate. All samples were conducted in triplicate and cultivated in a water-bath shaker with temperature of 55 ± 1 °C. GDR was then calculated according to the concentration of the specific substrate in the supernatant.

2.5 DNA extraction and high throughput pyrosequencing

Anaerobic sludge was sampled when methane production rate came to be stable. The genomic DNA was then extracted from the sludge samples with an extraction kit (Felix bio-tech, USA) according to the instructions of manufacturer. The quality and quantity of the extracted DNA were determined by its decrease in absorption ratios at 260 and 280 nm. DNA integrity was tested by Agarose Gel Electrophoresis (AGE) and then stored at −20 °C for further use. DNA from each samples were amplified by PCR with primers set 8F (5′-AGAGTTTGATCCTGGCTCAG-3′) and 533R (5′-TTACCGCGGCTGCTGGCAC-3′) for V1–V3 regions of bacterial 16S rRNA gene as well as primers 787F (5′-ATTAGATACCCSBGTAGT CC-3′) and 1059R (5′-GCCATGCACCWCCTCT-3′) for archaeal.¹⁷ The PCR process was as follows: denaturation step (94 °C) for 30 s, followed by 27 cycles (94 °C) for 30 s, annealing step (54 °C) for 30 s, extension (72 °C) for 30 s and a final extension (72 °C) for 5 min. The PCR products were used for sequencing on the Roche Illumina Miseq sequencer (Roche Life Sciences, Branford, CT, USA). Afterwards, effective sequences were clustered into operational taxonomic units (OTUs) by the MOTHUR program, at a distance limit of 0.03. The effective sequences from each OTUs were compared with NCBI taxonomy tool using the green genes 16S rRNA database, and the species distribution diagram was adopted. Alpha diversity statistics of Chao1, Simpson and Shannon diversity index were calculated for each sample according to the method.¹⁹

3. Results and discussion

3.1 Effect of TS contents on thermophilic AD performances

The cumulative methane production of the identical reactors at different TS contents during thermophilic AD was shown in Fig. 1a. The maximum cumulative methane production of 233 mL g_VS⁻¹ was obtained in R1 after 25 days, while the cumulative methane production decreased greatly to 159 mL g_VS⁻¹, 146 mL g_VS⁻¹ and 127 mL g_VS⁻¹ when TS content increased to 10%, 15% and 20%, respectively. It was generally believed that two periods, i.e., start-up and steady state, should be distinguished in the AD process.²⁰ Rapid methane production was observed within first 25 days in R1 while the methane production rate was much lower in R2 (Fig. S1†). However, in spite of being inoculated, a similar prolonged lag phase was observed in R3 and R4 in the start-up stage. After 25 days' digestion, the methane production rates at higher TS contents began to increase and achieved peak values on 37^th day (R2) and 51^st day (R3 and R4), respectively. Longer periods of time were taken at start-up stage for assays at higher TS content, probably resulting from the poor flowability of sludge, which caused compromised mass transfer.⁹ Under thermophilic condition, higher loading rate might also cause great release of organic matters from sludge into supernatant, which was more likely leading to the poor adaptation of microbes as well as the instability even the failure of AD process.²¹

Fig. 1 Cumulative methane production (a) and VS removal efficiency (b) of the reactors at different TS contents.

As shown in Fig. 1b, the VS removal in R1 achieved 40.92% after 37 days digestion, while that in R2 and R3 needed more than 60 days to achieve the EPA Class A requirement of 38% for sludge stabilization.²² The final VS removal efficiency for R1, R2 and R3 were 48.49%, 46.74% and 39.37%, respectively while the value was only 36.99% for R4 at the end of digestion. It was basically feasible for HS-AD to achieve stabilization requirement of sludge in this study, though longer digestion time was taken. Interestingly, an obvious increase of VS removal from 38.69% to 44.34% was observed in R2 after the 87^th day as well as its TS content rapidly dropped to below 5.9%, which might provide a better mass transfer and produce less residue.

3.2 Effects of TS contents on pH and FAN

The variations of pH in anaerobic digestion mainly depend on accumulation of VFA as well as the release of ammonia nitrogen.⁴ As shown in Fig. 2a, the pH of all samples increased rapidly to over 7.0 in the first 6 days. Meanwhile, the TAN concentration rose to over 1160 (R1), 2269 (R2), 2932 (R3) and 4450 mg L⁻¹ (R4), respectively due to the decomposition of protein in the extracellular polymeric substance in thermophilic condition.²³ Afterwards, pH of R1 kept fast growth to 7.82 on the 9^th day and reached the maximum of 8.23, while for R3 and R4 it maintained below 7.2 until the 32^nd day. The increase of pH in R3 and R4 was then observed with the dramatic increase of methane yield (Fig. 1a) as well as the decrease of VFA concentrations (further discussed in following part). It could be inferred that the activities of methanogens were evidently inhibited with the accumulation of VFAs in high-solid reactors, leading to the lag phase of methane production. Additionally, pH descend was observed in R2, R3 and R4 after 60 days' digestion and varied between 7.7 to 7.9 till the end of the digestion, which might be affected by both the accumulation and consumption of VFA, in accompany with the excess release of TAN.

Fig. 2 Variations of pH (a) and FAN concentrations (b) at different TS contents.

FAN was also reported as a toxic compound in anaerobic digestion and may cause the collapse of system when arriving at a certain volumetric concentration (over than 600–800 mg L⁻¹) in traditional L-AD.⁹ The variation of FAN depended on both pH and TAN concentration under thermophilic condition.¹⁶ As shown in Fig. 2b, FAN concentration in R1 increased continuously and achieved the maximum of 895.69 mg L⁻¹ on the 57^th day, before which the VFA accumulation had greatly relieved and the methane production was almost stable. In contrast, the FAN value of R3 and R4 ranged from 191 to 303 mg L⁻¹ during the first 27 days digestion, much lower than that in R1 and R2, even though the TAN concentrations in high-solid reactors (4248 and 5456 mg L⁻¹ for R3 and R4, respectively) were times higher than those in L-AD reactors (1608 and 2852 mg L⁻¹ for R1 and R2, respectively). After that, the FAN concentration of high-solid digestions increased significantly in correlation with the pH rise and peaked at 1583.6 mg L⁻¹ for R3 and 2093.8 mg L⁻¹ for R4 respectively, which might eventually lead to the poor yield of methane in HS-AD. The result indicated that pH had an important influence on FAN concentration under thermophilic condition and the appropriate pH range for methanogens in HS-AD was found between 7.2–7.6 in this study. It could be concluded that both VFA and FAN accumulation had negative impacts on activities of microorganisms and system stability for HS-AD, in accordance with the phenomenon that the cumulative methane production (per gram of VS) decreased in high-solid reactors.

3.3 Effects of TS contents on SCOD and VFAs

The variation of SCOD associated with the process that particulate substrate is rapidly converted into soluble organic matters during the hydrolysis and acidification phases under the thermophilic conditions.²⁴ As shown in Fig. 3a, the concentration of SCOD increased sharply during the first 3 days and kept an ascending trend except that in R1, whose SCOD gradually declined to the minimum level of 161.3 mg_COD g_VS⁻¹, lower than three other assays till the end of digestion. R2, R3 and R4 showed a similar tendency, gradually reaching peak values of SCOD on the 37^th day. The result indicated that the methanogenesis process was the limiting step when thermophilic AD was operated at a higher TS content, probably attribute to that low moisture content restrained both the substrate hydrolysis and mass transfer of hydrolysis product.¹⁰

Fig. 3 Effects of different TS contents on SCOD (a) and supernatant components (b). Acetate, propionate, butyrate and valerate were summed as VFAs.

As a majority of SCOD, abundant VFAs, proteins and polysaccharide were released to the supernatant and then mineralized to CH₄ and CO₂.²⁵ The composition of SCOD at different stage was shown in Fig. 3b and the concentrations of organic matters were listed in Table S1.† At the beginning stage (6^th day), VFAs were the principal organic compounds, accounting for about 83.3%, 74.6%, 74.3% and 57.4%, respectively. With the increase of TS content, the unknown organic matters (including long chain fatty acid, nucleic acid and amino acid, etc.) increased from 29.7 to 230.3 mg_COD g_VS⁻¹ on the 6^th day, indicating that complicated hydrolysis product was more likely to form in HS-AD. Similar with the changes of SCOD, the concentrations of both total VFAs and unknown organic matters increasingly moved up to the highest levels on the 37^th day in R2, R3 and R4 while the total VFAs value for R1 declined sharply to 28.4 mg_COD g_VS⁻¹, contributing to about only 11% of SCOD. The results above indicated that VFAs were more likely to be consumed at liquid digestion while the increase of TS content accelerated the process that unknown organic matters were converted into SCOD in hydrolysis–acidification stage. At the end, the concentrations of unknown organic matters on the 122^nd day increased with TS content, ranging from 195.3 (R1) to 298.0 (R4) mg_COD g_VS⁻¹, which might be difficult to be utilized by microbes. Decline of VFAs was observed in reactors of higher TS, though the value in R4 was still 57.2% higher than that in R2 and even 5.69 times higher than R1. During the whole digestion, the most effective utilization of excessive VFAs was obtained in R1, which was in accordance with the positive methane production, probably due to the higher activity and the enrichment of specific microbes at the higher moisture content.

For better understanding the accumulation and consumption of VFAs, time evolution of individual VFA concentration was shown in Fig. 4, including three of the most prevalent products: acetic, propionic and butyric acid. Compared with propionic acid, acetic and butyric acid had greater contributions to VFAs, as important indicators for hydrolysis and fermentation process.⁸ Content of acetic acid in R1 rapidly increased and then declined during the first 20 days, while obvious accumulation was observed in R3 and R4. The result coincided with the variations of pH and SCOD as well as the low proportion of soluble protein and polysaccharide. For butyric acid, similar tendency was observed as acetate, though the decline of concentration occurred about 15 days later. With respect to propionic acid, the values kept increasing till the end of digestion in reactors with higher TS, achieving over 60 mg_COD g_VS⁻¹ (2717, 4332 and 4658 mg L⁻¹ for R2, R3 and R4 respectively when calculated in volumetric concentration, greatly higher than the appropriate value reported in traditional L-AD in previous studies), which might lead to suppression on microorganisms. The phenomenon demonstrated that the excessive propionic acid was difficult to convert into acetic acid in HS-AD probably due to the Gibbs free energy fluctuation (ΔG = +76.1 kJ mol⁻¹).²⁶ Conversely, propionic acid concentration in R1 peaked on the 27^th day (58.2 mg_COD g_VS⁻¹) and then sharply declined to a low level, probably on account of the high moisture content that promoted the conversion from propionate to acetic acid. From the results above, it could be concluded that the degradation rate of individual VFA might depends on moisture content of sludge as well as the length of carbon chain in HS-AD system.

Fig. 4 Variations of acetic acid (a), propionic acid (b) and butyric acid (c) concentrations at different TS contents.

3.4 Acidogenesis activities at different TS contents

Activities of the acidogenic bacteria can be quantified by specific substrate degradation rate.¹⁸ The glucose degradation rate (GDR) was detected at different periods of the anaerobic digestion. As seen in Table 2, the glucose degradation rate of R1 at the beginning stage was 0.635 ± 0.008 mg h⁻¹ mg_DS⁻¹, much higher than those of other assays, demonstrating that sludge of lower TS had higher activities of microbe which promoted the hydrolysis of sludge. During the whole digestion, the GDR of R3 and R4 decreased fast with time, suggesting that the inhibition of microbial activities persisted in HS-AD system, in accordance with the enhancement of carbohydrate and unknown organic matters in SCOD. At the end of AD process, the GDR of each batch assays were 0.327 ± 0.005, 0.179 ± 0.012, 0.065 ± 0.006 and 0.067 ± 0.007 mg h⁻¹ mg_DS⁻¹, respectively, much lower than those at the initial stage.

Table 2 Glucose degradation rate (GDR) after the digestion for 6 d, 47 d, 87 d and 122 d. Mean values were in triplicate

Reactor	Glucose degradation rate (mg h^−1 mgDS^−1)
	6 d	47 d	87 d	122 d
R1	0.635 ± 0.008	0.608 ± 0.007	0.615 ± 0.004	0.327 ± 0.005
R2	0.327 ± 0.003	0.348 ± 0.013	0.324 ± 0.006	0.179 ± 0.012
R3	0.245 ± 0.007	0.194 ± 0.012	0.155 ± 0.011	0.065 ± 0.006
R4	0.197 ± 0.015	0.116 ± 0.004	0.061 ± 0.006	0.067 ± 0.007

---

3.5 Microbial community analysis

Illumina-based 16S sequencing was capable for investigating the complex microbial communities structures in AD system, providing a more detailed raw data than the 454 high-throughput pyrosequencing.¹⁹ The microbial diversity indices of Shannon, observed OTUs numbers, estimators of Chao1 at cut off level of 3% were calculated for both bacteria and archaea (Table 3). The higher values of observed OTUs and Chao1 for bacteria were shown in R1. The results integrated with the higher GDR value jointly suggested that more species richness and microbial activities together drove the anaerobic digestion at a higher moisture. For archaea, contrary to bacteria, apparent increase was observed in OTUs numbers and Chao1 richness when process was operated at a higher TS content. The result was also corresponding to Shannon index ranging from 1.384 to 4.596 for R1–R4, indicating a selective enrichment of specific archaea in R1. Obviously higher diverse microbial population of archaea in R4 might imply that HS-AD system was not conducive to the selection of predominate methanogens, probably due to the accumulation of hydrolysis product and the compromised mass transfer.

Table 3 Alpha diversity indices of bacteria and archaea at a distance limit of 3%

Sample		Observed OTU	Chao1	Shannon	Coverage
Bacteria	R1	2141	4310 ± 28	6.24 ± 0.05	0.99
	R2	1758	3186 ± 19	6.68 ± 0.03	0.98
	R3	1749	2964 ± 24	5.94 ± 0.07	0.98
	R4	1753	3098 ± 23	6.10 ± 0.01	0.98
Archaea	R1	308	392 ± 15	1.38 ± 0.04	0.99
	R2	369	463 ± 15	1.61 ± 0.09	0.99
	R3	385	483 ± 22	1.72 ± 0.02	0.98
	R4	540	593 ± 14	4.59 ± 0.06	0.99

---

Major bacterial genera with relative abundance of greater than 0.5% in at least one sample were shown in Fig. 5a (original data listed in Table S2†) and other genus were assigned to the unclassified group. Genus Coprothermobacter was found in dominant status and maintained a relative abundance of 25% and 20.8% of total sequences in R1 and R2, respectively. As a proteolytic microbe living in strict syntrophic associations, Corprothermobacter was found to ferment abundant extracellular proteinaceous material for methanogens and promote the methane production.²⁷ However, the absence of Coprothermobacter in HS-AD for R3 (5.5%) and R4 (2.1%) might result in the poor degradation of protein and the higher proportion of accumulated unknown organic matters in SCOD, which accounted for adverse biogas production.

Fig. 5 Major bacterial communities (a) and archaea communities (b) at genus level at different TS contents. Each represented by more than 0.5% total sequences in at least one sample and the rest was defined as “Unclassified”.

The population of archaeal (Fig. 5b, original data listed in Table S3†) was identified at genus level to illuminate the effects of TS content on anaerobic methanogens. Genus Methanothermobacter and Methanoculleus, as hydrogenotrophic methanogens, were known to produce methane only by utilizing H₂ and CO₂ as substrates.²⁷ The highest relative abundance of Methanothermobacter was found in R2 representing 42.0%, while it accounted for a small proportion in R3 (21.8%) and R4 (2.1%). The pH recovery at the early stage probably played an important role in the enrichment of Methanothermobacter due to its capacity of living at a higher pH than other genera.²⁸ With the increase of TS contents, the most dominant OTUs were assigned to genus Methanoculleus, accounting for 39.9% and 43.2% in R3 and R4, respectively. The succession of these two hydrogenotrophic methanogens might be due to the difference in physical state (solid and liquid) that ascribed to the mass transfer, change of pH and/or concentrations of VFAs and TAN. With respect to acetoclastic methanogens, there was non-significant gap found in members of genus Methanosaeta in each reactor, ranging from 19.0% to 22.9%. However, Methanosarcina was found in abundance in R1 (34.5%) while the value greatly decreased to 15.5%, 6.3% and 3.6%, respectively in R2, R3 and R4. It is well known that both hydrogenotrophic and acetoclastic methanogenesis could be available for methane production by Methanosarcina, which could use acetate, methanol, methylamine, H₂ and CO₂ as its substrate.²⁹ The enrichment of Methanosarcina in R1 was consistent with the rapid consumption of acetates with a low concentration. As previous literature reported, hydrogenotrophic methanogens were more observably detected in thermophilic HS-AD while acetoclastic methanogens prevailed at L-AD.³⁰ Though the process of syntrophic acetate oxidation and hydrogenotrophic methanogenesis by specific microbes was considered as an important pathway for methane production,³¹ proving the feasibility of HS-AD, the results in present study suggested that the inhibition of methane production with the increase of TS content was reasonable for the lower relative abundance of acetoclastic methanogens at such a high concentration of VFAs.

3.6 Reactor utilization efficiency at different TS contents

Since HS-AD had the advantage of smaller reactor volume and less energy consumption while longer reaction time was required, it was essential to estimate the efficiency of reactors when the experiment operated at different TS content. As shown in Fig. 6, with respect to T80, which represents the digestion time to produce 80% of the maximum methane produced,³² no significant difference was found in R2, R3 and R4 of about 70 days, 72 days and 65 days, respectively, much longer than that in R1 (20 days). The result above might attribute to the extended start-up period and low level of VS removal rate in HS-AD (TS over 10% in present work). Though the total methane yield in each assays increased due to the augment of organic loading, AD performances in terms of utilization efficiency exhibited a distinct result. Volumetric methane productivity was a parameter normalized by reactor volume and digestion time, measuring the amount of methane produced per volume unit per day. The value was applied to express the reactor utilization efficiency in some of the previous studies on HS-AD.^8,10 Volumetric methane productivity in R1 reached 0.376 L L⁻¹ d⁻¹, about twice more than that in R2 (0.157 L L⁻¹ d⁻¹) and slightly higher than that in R4 (0.368 L L⁻¹ d⁻¹). The limiting capacity of methanogenesis performance of HS-AD might account for the mass transfer which was more likely to cause adverse effect on microbial activities. However, higher volumetric methane productivity was observed with the increase of substrate loading from R2 to R4 and similar utilization efficiency was achieved at TS of 20% as that in L-AD, probably due to that TS of sludge decreased rapidly during thermophilic AD and the flowability would be better when system was in stable-state, which might raise the microbial adaption to environment.⁹ As a result, though slightly longer digestion time was needed for stabilizing the extensive residue in HS-AD, sludge could be digested with smaller reactor volume (and also the saving of energy and dilution water) in comparison with L-AD at a similar reactor efficiency. Digestion performance for TS higher than 20% as well as the balance between substrate loading and utilization efficiency remained to be studied, especially the negative influence of rheology on HS-AD process at extremely low moisture content.

Fig. 6 Volumetric methane productivity and T80 of sludge at different TS contents.

4. Conclusion

Among four tests operated at different TS, L-AD (5%) achieved the maximum cumulative methane production and the feasibility of HS-AD process was confirmed despite longer start-up period. High relative abundance of Coprothermobacter and Methanosarcina was found in L-AD (R1) for bacteria of protein degradation and archaea of methane production, accounting for 25.0% and 34.5%, respectively, while hydrogenotrophic methanogens Methanoculleus were dominant in HS-AD. HS-AD (at TS 20%) could support similar reactor utilization efficiency as L-AD, thus providing a viable choice for sludge digestion with smaller reactor volume and less consumption of energy and dilution water.

Acknowledgements

This study was supported by National Natural Science Foundation of China (No. 51178261) and the Key project of Science and Technology Commission of Shanghai Municipality (No. 12231202101).

References

1. G. Y. Zhen, X. Q. Lu, J. Niu, L. H. Su, X. L. Chai, Y. C. Zhao, Y. Y. Li, Y. Song and D. J. Niu, Chem. Eng. J., 2013, 233, 274–281.
2. T. Forster-Carneiro, M. Pérez and L. I. Romero, Bioresour. Technol., 2008, 99, 6994–7002.
3. B. K. Ahring, Biomethanation I, 81, Springer, New York, USA, 2003.
4. N. N. Duan, B. Dong, B. Wu and X. H. Dai, Bioresour. Technol., 2012, 104, 150–156.
5. L. Luning, E. H. M. V. Zundert and A. J. F. Brinkmann, Water Sci. Technol., 2003, 48, 15–20.
6. D. Brown, J. Shi and Y. Li, Bioresour. Technol., 2012, 124, 379–386.
7. S. Fujishima, T. Miyahara and T. Noike, Water Sci. Technol., 2000, 41, 119–127.
8. H. Benbelkacem, J. Bollon, R. Bayard, R. Escudié and P. Buffière, Chem. Eng. J., 2015, 273, 261–267.
9. X. H. Dai, X. Gai and B. Dong, Bioresour. Technol., 2014, 174, 6–10.
10. F. Xu, F. Wang, L. Lin and Y. Li, Bioresour. Technol., 2016, 200, 753–760.
11. Y. Xue, H. Liu, S. Chen, N. Dichtl, X. Dai and N. Li, Chem. Eng. J., 2015, 264, 174–180.
12. E. Aymerich, M. Esteban-Gutiérrez and L. Sancho, Bioresour. Technol., 2013, 144, 107–114.
13. Y. Zhang, Y. Feng, Q. Yu, Z. Xu and X. Quan, Bioresour. Technol., 2014, 159, 297–304.
14. Q. Li, W. Qiao, X. C. Wang, K. Takayanagi, M. Shofie and Y. Y. Li, Waste Manage., 2015, 36, 77–85.
15. APHA, AWWA, WEF, Standard Methods for the Examination of Water and Wastewater, American Public Health Association/American Water Works Association/Water Environment Federation, Washington, DC, USA, 21st edn, 2005.
16. N. Østergaard, Biogasproduktion i det thermofile temperaturinterval, STUB rapport nr. 21, Kemiteknik Dansk Teknologisk Institut, Taastrup, 1985.
17. B. Yu, Z. Y. Lou, D. L. Zhang, A. D. Shan, H. P. Yuan, N. W. Zhu and K. H. Zhang, Bioresour. Technol., 2015, 179, 291–298.
18. K. Wang, J. Yin, D. Shen and N. Li, Bioresour. Technol., 2014, 161, 395–401.
19. M. C. Nelson, H. G. Benjamino, S. L. Grim and J. Graf, PLoS One, 2014, 9(4), e94249.
20. M. Kim, Y. H. Ahn and R. E. Speece, Water Res., 2002, 36, 4369–4385.
21. T. Hidaka, F. Wang, T. Togari, T. Uchida and Y. Suzuki, Bioresour. Technol., 2013, 149, 177–183.
22. USEPA, 40CFR PART 503, Standards for the Use or Disposal of Sewage Sludge, United States Environmental Protection Agency, Washington, DC, 1993.
23. Y. T. Ahn, S. T. Kang, S. R. Chae, C. Y. Lee, B. U. Bae and H. S. Shin, Desalination, 2007, 202, 114–121.
24. L. Guo, X. M. Li, B. Xie, Q. Yang, G. M. Zeng, D. X. Liao and J. J. Liu, Bioresour. Technol., 2008, 99, 3651–3658.
25. J. Kim, Y. Yu and C. Lee, Bioresour. Technol., 2013, 144, 194–201.
26. V. A. Vavilin, L. Y. Lokshina, S. V. Rytov, O. R. Kotsyurbenko and A. N. Nozhevnikova, Bioresour. Technol., 1998, 63, 159–171.
27. K. Sasaki, M. Morita, D. Sasaki, J. Nagaoka, N. Matsumoto, N. Ohmura and H. Shinozaki, J. Biosci. Bioeng., 2011, 112, 469–472.
28. A. Wasserfallen, J. Nölling, P. Pfister, J. Reeve and E. C. De Macario, Int. J. Syst. Evol. Microbiol., 2000, 50, 43–53.
29. J. L. Garcia, B. K. C. Patel and B. Ollivier, Anaerobe, 2000, 6, 205–226.
30. T. Hori, S. Haruta, Y. Ueno, M. Ishii and Y. Igarashi, Appl. Environ. Microbiol., 2006, 72, 1623–1630.
31. Y. F. Li, M. C. Nelson, P. H. Chen, J. Graf, Y. Li and Z. Yu, Appl. Microbiol. Biotechnol., 2014, 99, 969–980.
32. H. Tian, N. Duan, C. Lin, X. Li and M. Zhong, J. Biosci. Bioeng., 2015, 120, 51–57.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra21111a

---