Methane-rich biogas production from waste-activated sludge with the addition of ferric chloride under a thermophilic anaerobic digestion system

Abstract

Thermophilic anaerobic digestion for methane production could realize simultaneously both energy recovery from waste-activated sludge and pollution control; however, low methane production and an imbalance between the hydrolysis and methanogenesis processes are often encountered. Three ferric salts, Fe(NO₃)₃, Fe₂(SO₄)₃ and FeCl₃, were introduced to test the potential effects on sludge anaerobic digestion performance under a thermophilic system. Enhanced methane production was achieved using FeCl₃ as the additive: this had a cumulative methane production of 117.44 mL CH₄ per g of volatile solid (VS), an increase of 98.9% over that in the control experiment (59.05 mL CH₄ per g VS). Both Fe(NO₃)₃ and Fe₂(SO₄)₃ caused some negative effects, meaning that the type of anion should also be considered. The introduction of FeCl₃ created a favorable environment resembling positive precipitation and biocatalysis. The succession of microbial communities before and after the introduction of ferric salts was investigated and compared through pyrosequencing analysis, and in particular the dominance of Methanosarcina increased from 1.3% to 63.2% in effective reads with the addition of FeCl₃.

---

1. Introduction

Waste-activated sludge (WAS) is the inevitable by-product from waste water treatment, and the amount produced increases greatly with the growing quantity of waste water and evermore stringent sewage discharge standards.¹ Anaerobic digestion is an attractive disposal process for sludge, that not only reduces the sludge mass and volume but also generates a valuable energy source in the form of biogas.² Compared to mesophilic digestion, thermophilic anaerobic digestion has additional benefits including a high degree of waste stabilization, and more thorough destruction of viral and bacterial pathogens.³ However, the methanogenic phase is the rate-limiting step of the process due to the slow growth rate of the methanogenic bacteria, while the acceleration of hydrolysis in thermophilic anaerobic digestion results in a greater accumulation of volatile fatty acids (VFAs), which greatly inhibit methanogenesis.⁴

There are some methanogenesis disinhibition methods applied to prevent VFA accumulation in anaerobic digestion, such as adjusting the C/N ratio,⁵ employing a two-stage digestion system to the separate the methanogenesis stage from the hydrolysis and acidification stages,⁶ and adding trace elements.⁷ Trace metals are necessary for the anaerobic digestion process, and it has been confirmed that the introduction of iron could enhance mesophilic anaerobic digestion, by improving the physiological and biochemical performance through preventing sulfide inhibition and controlling phosphate in waste water and kitchen waste treatment.^8,9 During anaerobic digestion, sulfate and phosphate are produced from the decomposition of proteins, which react with ferrous ion to precipitate as ferrous sulphide and phosphate. The formation of iron precipitates is effective for phosphate collection, which is beneficial for reducing the cost of the subsequent processes for treating the digested effluent.¹⁰ Besides, iron is also an essential constituent of some cofactors and enzymes that stimulate and stabilize biogas process performance.¹¹ The supplementation of ferric chemical compounds is operationally easy in anaerobic digestion, and the choice of metal species and anion can influence the operational conditions in the anaerobic reactor, such as the total metal concentration, pH, oxidation–reduction potential (ORP) and temperature.^11,12 Based on the positive effects of iron, it was assumed that the introduction of ferric salts into a waste-activated sludge thermophilic anaerobic digestion system would have a direct influence on biological oxidation and be expected to create a more favourable environment for methane production.

The application of trace metals in anaerobic digestion of WAS has been reported in some works, while the availability of different ferric salts and their potential routes in sludge treatment are still limited, especially in thermophilic anaerobic digestion. To clarify these issues, Fe(NO₃)₃, Fe₂(SO₄)₃ and FeCl₃ were selected as additives for the improvement of the thermophilic anaerobic process in this study. The objectives of this work were to investigate the effects of these three ferric salts on the biogas production process, anaerobic digestion conditions, and the fate of iron ion in the system. Finally, the microbial community structure and relative abundance were also studied by 454 GS-FLX pyrosequencing technology to gain insight into the thermophilic anaerobic digestion system directly.

2. Experimental

2.1 Sludge sampling and characterization

The WAS used in this study was obtained from the secondary sedimentation tank of a municipal waste water treatment plant (MWWTP) in Shanghai, China, where waste water is treated by an anaerobic–anoxic–aerobic process with a capacity of 50 000 m³ per day. The sludge obtained was filtered with a 1.0 mm mesh to eliminate large particles and hair before thickening to the required solid concentration. The pre-treated samples were then stored at 4 °C for further analysis. The seed sludge (inoculum) was collected directly from a long-term continuous lab-scale anaerobic bioreactor in our lab, fed with activated sludge. Compared to the raw sludge, the seed sludge had a healthy population of methanogens and other microorganisms needed for the efficient start-up of new digesters. The main characteristics of the raw sludge and seed sludge are shown in Table 1.

Table 1 Characteristics of the raw sludge and seed sludge

Parameters	Raw sludge	Seed sludge
pH	6.32–6.40	6.87–6.90
TS (g L^−1)	39.5–39.9	68.7–69.2
VS (g L^−1)	29.0–29.3	49.6–52.2
TCOD (mg L^−1)	32 570–37 640	84 576–87 260
SCOD (mg L^−1)	124.0–535.6	16 240–18 440
STN (mg L^−1)	104.18–121.8	762.6–812.4
STP (mg L^−1)	55.13–98.64	416.16–467.38
C (%)	34.68–34.83	—
H (%)	5.39–5.51	—
N (%)	6.76–6.90	—
S (%)	1.35–1.38	—
Fe (%)	1.72–1.91	1.83–1.94

---

2.2 Experimental design

Batch experiments were carried out in double-walled cylindrical vessels with 6 L working volumes. The seed sludge to WAS ratio was 1 : 3 (volume : volume). After loading the sludge, oxygen was removed from the headspace by the injection of nitrogen gas (99.99%) for 5 min to maintain anaerobic conditions. During the anaerobic digestion process, all reactor vessels were maintained at a thermophilic digestion temperature of 55 ± 2 °C by water circulation, and equipped with stainless-steel stirrers for mixing the contents. The biogas produced was measured using a calibrated sampling syringe. All samples from the reactor vessels were analysed in triplicate. With the aim of investigating the effects of different ferric salts on the thermophilic anaerobic digestion of WAS, a fixed dosage of either Fe(NO₃)₃, Fe₂(SO₄)₃ or FeCl₃ was applied (referred to as R2, R3 and R4 experiments respectively) in the form of ferric salt solutions (50 mL) on the 3^rd day after the experimental start-up, based on a ferric ion equivalent of 200 mg L⁻¹. A control experiment (R1) was also carried out under the same operational conditions but without any ferric salt. No alkalinity or buffering agent was added into the system, and the pH value was not adjusted during the process.

2.3 Analytical methods

50 mL sludge was collected from the reactors each time after the contents of each was completely mixed, and the corresponding supernatant was sampled by centrifugation at 12 000 rpm for 5 min with a subsequent filtration through 0.45 μm microfiber filter paper. The ORP and pH were measured by an ORP meter (ORP-502, Ruosull Technology Co., Ltd., Shanghai) and a pH meter (pHs-3C, Leici Co. Ltd., Shanghai), respectively. Total solids (TS), VS, soluble chemical oxygen demand (SCOD), total chemical oxygen demand (TCOD), NH₄⁺–N, total nitrogen (TN), soluble total nitrogen (STN), total phosphorus (TP) and soluble total phosphorus (STP) were measured according to standard methods.¹³ VFAs (including acetic acid, propionic acid, butyric acid, isobutyric acid, valeric acid and isovaleric acid) were analyzed by a gas chromatograph (GC-2010, Shimadzu) with a chromatographic column (DB-FFAP: 30 m × 0.25 mm × 0.25 mm) and a flame ionization detector (FID). The concentrations of CH₄ and CO₂ were quantified by a gas chromatograph (GC-14B, Shimadzu) equipped with a chromatographic column (TDX-02) and a thermal conductivity detector (TCD).¹⁴ The concentration of iron ion in the supernatant was determined via atomic absorption spectrometry (spectrometer contrAA, Analytik Jena). An energy-dispersive spectrometer (EDS) analysis of the sludge samples was performed by a large area silicon drift detector (Large Area SDD, AZtec X-Max 80). The C, H, N and S content in the WAS were measured by an elemental analyzer (Vario Macro Cube, Elementar) with sulfanilamide (C₆H₈O₂N₂S) as a reference material. All the results represent the mean value of experiments done in triplicate with an accuracy of ±5%.

2.4 DNA extraction and high-throughput 16S rDNA gene pyrosequencing

Mixed sludge was collected from the reactors before start-up and after 28 days operation, when daily methane production (DMP) was at a high level. The samples were washed with phosphate-buffered saline, after which the genomic DNA of the samples was extracted using an extraction kit (Felix bio-tech, USA) according to the manufacturer’s instructions. The quality of the extracted DNA was checked by determining its absorbance at 260 and 280 nm, and agarose gel electrophoresis (AGE) was employed to test the DNA integrity.

Two universal primers for Archaea, 787F (5′-ATTAGATACCCSBGTAGTCC-3′) and 1059R (5′-GCCATGCACCWCCTCT-3′) were used to amplify the Archaea 16S rRNA gene.¹⁵ The PCR program consisted of an initial 5 min denaturation step at 94 °C, 27 cycles of repeated denaturation at 94 °C for 30 s, annealing at 54 °C for 30 s, and extension at 72 °C for 30 s, followed by a final extension step of 5 min at 72 °C. After being purified and quantified, the PCR products of the 16S rRNA gene were determined by pyrosequencing using the Roche 454 FLX Titanium sequencer (Roche 454 Life Sciences, Branford, CT, USA) according to the methodology described by Zhang et al.¹⁶ Subsequently, the MOTHUR program was used to cluster effective sequences into operational taxonomic units (OTUs) at a 3% level. Rarefaction curves, the species richness estimator Chao1 and the Shannon diversity index were also generated via the MOTHUR program to identify the species diversity for each sample. The OTUs defined by a 3% distance level were classified using the RDP classifier at a 50% confidence threshold. The effective sequences obtained from pyrosequencing were compared with the Greengenes 16S rRNA gene database using the NCBI’s BLASTN tool, using the species distribution diagram. In order to further distinguish the dominant species that could account for the different digestion performances among the reactors, phylogenetic relationships of sequences were conducted according to the method described by Ye et al.¹⁷ The main software used for metagenome analysis was MEGAN, which is capable of conducting cluster analysis for the extensive sequencing data.

3. Results and discussion

3.1 Biogas generation performance

To investigate the effects of ferric salts on the thermophilic anaerobic digestion of WAS, DMP, total methane production (TMP) and cumulative methane production (CMP) were determined and calculated, based on the initially loaded volatile solid (VS).

Fig. 1a illustrates the time curves of the CMP of the reactors with ferric salt dosing (R2, R3 and R4) and the control experiment (R1). During the initial stage, R1, R3 and R4 had a similar tendency, while R2 – dosed with Fe(NO₃)₃ – showed no distinct increase. Afterwards, CMP increased greatly in R4 – dosed with FeCl₃ – and reached maximum methane production around the 43^rd day; thus a final CMP of 117.44 mL CH₄ per g VS was obtained, which was 98.9% higher than that in R1 (59.05 mL CH₄ per g VS) and similar to the values (102–145 mL CH₄ per g VS) reported by Sheets et al. under a thermophilic anaerobic system.⁴ It should be emphasized that it is still lower than some literature values: the current best reported maximum CMP of 256.45 mL CH₄ per g VS was achieved after thermal pre-treatment (70 °C),¹⁸ and this result might be due to the differences of sludge and inoculum properties, or operating conditions like the reactors, hydraulic retention time (HRT), or pH. For R2, which had Fe(NO₃)₃ added, biogas production was inhibited completely, and a low CMP of 5.96 mL CH₄ per g VS was observed. With Fe₂(SO₄)₃ dosing, a stable increase of CMP was obtained that finally reached a peak of 46.67 mL CH₄ per g VS. The maximum TMP was also found in R4: this was 95.23% higher than that of R1 (5075 mL), whereas the TMP of R2 and R3 were only 695 mL and 4250 mL, respectively. Nevertheless, the addition of FeCl₃ did contribute to enhancement of the methane production from WAS under the test conditions, which confirmed the positive effects of iron on anaerobic digestion.¹⁹

Fig. 1 (a) Cumulative methane production (CMP) and (b) daily methane production (DMP) during the anaerobic digestion process.

With respect to DMP (Fig. 1b), all reactors had a rapid methane production rate in the first three days, implying that methanogenesis was effectively activated in the reactors,²⁰ and that DMP in R1, R3 and R4 then declined around the 9^th day due to the excessive accumulation of VFAs and low pH resulting from rapid hydrolysis acidification in the thermophilic anaerobic digestion;²¹ DMP increased and reached its peak afterwards. The DMP of R2 was low throughout the whole process as a result of the unsuitable conditions for methanogens. As an indicator for process stability, the ratio of CO₂/CH₄ (Fig. A1†) in R2 remained at a high level (>3), indicating the small proportion of CH₄ in the produced biogas and confirming the process failure after dosing with Fe(NO₃)₃.¹⁴

3.2 Variations of anaerobic digestion conditions

It is well known that sludge anaerobic digestion is a delicate balance between the rate of hydrolysis and methanogenesis because of the far greater sensitivity of methanogenic than acidogenic and fermentative bacteria to VFA accumulation and pH drop.²² Hence, the key factors of pH, ORP, and concentrations of VFAs, SCOD, TN and NH₄⁺–N, were measured to assess the process stability of anaerobic digestion.²³

3.2.1 Variations of pH and ORP. The variations of pH and ORP in the reactors are shown in Fig. 2. After commencement of the experiment, the pH fell immediately and reached a minimum value of 5.8 within the first 3 days due to hydrolysis acidification.²⁰ When the different ferric salts were dosed into R2, R3 and R4, the pH sharply dropped to 5.3–5.4, due to the acidification properties of the three ferric compounds and the bonding of hydroxyl ions during the settlement process. After rapid decline, the pH values in R1 and R4 increased and remained in the range of 6.5–7.2, almost in the desired pH range of 6.8–7.2 for anaerobic digestion.²⁴ The pH value of R2 did not noticeably fluctuate and remained below 6.2, corresponding to the poor methane production. Similarly, the biggest change of ORP was observed in R2: it increased sharply from −491 mV to −59 mV after dosing with Fe(NO₃)₃ due to the strong oxidizing properties of ferric and nitrate ions. The failure of methanogenesis in R2 might be attributable to the huge shock and the sensitivity of the methanogens to the environmental conditions. The ORP levels in R1 and R4 continued to decline and fluctuated within a small range, whereas that in R4 was lower (from −505 mV to −527 mV) and therefore more favorable for the thermophilic methanogenic bacteria.²⁵ In particular, ORP in R3 increased rapidly from −403 mV to −295 mV (from day 24 to 36) and then slowly dropped to −322 mV on day 48, corresponding well with gradual methanogenesis inhibition. As cells ruptured and proteolysis occurred, SO₄²⁻ was constantly released into the supernatant. With this introduction of SO₄²⁻ into R3, iron ion was no longer sufficient for sulfide control and this resulted in an enrichment of sulfate-reducing bacteria (SRB). SRB compete with methanogens for the hydrogen, formate and acetate, and in general they have better kinetic growth properties than methanogens in the presence of sulfate.²⁶ The oxidation of organics by SRB might have accounted for the higher ORP in R3, which was adverse to the growth of methanogens and accounted for the subsequent worse biogas production.^8,27

Fig. 2 (a) Variation of pH during the anaerobic digestion process; (b) variation of ORP during the anaerobic digestion process.

3.2.2 Effect of different ferric salts on VFAs and SCOD. Variations of total VFA (TVFA) in the supernatant (Fig. 3a) were in accordance with the pH, and the hydrolysis of sludge particles and the rapid growth of acidogenic and fermentative bacteria seemed to occur rapidly, which led to immediate VFA accumulation. On the whole, the concentrations of TVFA in all the reactors first experienced a sharp increase during the initial digestion period, and experienced then a moderate variation from day 9 to 23, followed by steady decline, except for that of R4 where there was a high level fluctuation from 6260 to 9396 mg L⁻¹. The maximum TVFA concentrations, 9183, 9396, 8582 and 8370 mg L⁻¹, respectively, were obtained on the 18^th, 28^th, 15^th and 23^rd days in R1, R2, R3 and R4. Meanwhile, the TVFA concentration in R4 underwent a relatively stable reduction, meaning that a balance was established between the acidification process and methanogenic activity. For the constituents of TVFA (Fig. A2†), the dominant constituent of acetic acid, which can be considered as the indicator for the TVFA concentration, decreased to zero in R4 at the end of the experiment.⁷ The effective conversion of VFAs in R4 resulted in a favorable pH to enhance biogas production.

Fig. 3 (a) Variation of VFA concentrations in the supernatant; (b) variation of SCOD in the supernatant.

SCOD of R1, R3 and R4 (Fig. 3b) increased sharply at first then only underwent small fluctuations, followed by a steady decline, the same as the trend of the TVFA curve;²⁸ whereas that in R2 remained at a high level, ranging from 20 550 to 26 374 mg L⁻¹ due to the poor methanogenesis. After the rapid growth of SCOD, peak values were obtained on 18^th day in each reactor, and were 22 483, 25 522, 25 391 and 25 912 mg L⁻¹, respectively. The SCOD growth rate of R4 was the fastest of all the reactors, and a rapid decline was obtained with the enhanced methane production rate, indicating that the addition of FeCl₃ accelerated the conversion of soluble organic matter to biogas. On the other hand, the control experiment (R1) maintained a lower SCOD than the other three reactors in the later digestion period, and the value ranged from 5137 to 5665 mg L⁻¹.

3.2.3 Effect of different ferric salts on TN and NH₄⁺–N in the supernatant. In the thermophilic anaerobic digestion system, the decomposition of protein in heat-labile microbes and extracellular polymeric substances gave rise to a concentration of TN in the initial stage as shown in Fig. 4a. The concentration of TN in R3 and R4 had a rapid increase after dosing with Fe(NO₃)₃ and FeCl₃, and a relatively high level was achieved in R4 throughout the whole process, in accordance with the variations of SCOD and VFAs and further confirming the acceleration effect of ferric salts on the hydrolysis and acidification processes. A rapid increase of TN was observed in R2 due to the introduction of nitrate (∼664 mg L⁻¹) with a highest value of 2265 mg L⁻¹, followed by a declining trend. As shown in Fig. 4b, the similar trend of NH₄⁺–N variation in R3 and R4 was observed, while that in R2 dropped rapidly to 735 mg L⁻¹, and then increased steadily. This could be explained by the existence of anaerobic ammonium oxidation (ANAMMOX), as given in reaction (1):

NH₄⁺ + 1.32NO₂⁻ + 0.066HCO₃⁻ + 0.13H⁺ → 1.02N₂ + 0.26NO₃⁻ + 0.066CH₂O_0.5N_0.15 + 2.03H₂ (1)

Fig. 4 (a) Variations of TN concentrations in the supernatant; (b) variations of NH₄⁺–N concentrations in the supernatant.

Theoretically, 1.32 moles NO₂⁻ consumes 1 mole NH₄⁺, with accompanying production of 1.02 moles N₂ and 0.26 moles NO₃⁻.²⁹ Due to the introduction of NO₃⁻, its conversion from NO₂⁻ was inhibited during the nitrification process, which resulted in the accumulation of NO₂⁻ and the decrease of NH₄⁺ by the ANAMMOX reaction.

3.3 Pyrosequencing analysis of microbial community

3.3.1 Biodiversity of the microbial community phylotypes. A mixed sludge sample taken before start-up (referred to as R0) and sludge samples taken from R1, R2, R3 and R4 on the 28^th day were collected and used for investigating the succession of microbial communities before and after the introduction of ferric salts by pyrosequencing analysis. As shown in Table 2, a total of 12 340 (R0), 17 985 (R1), 14 992 (R2), 18 372 (R3) and 18 998 (R4) effective sequence tags were obtained through primer and barcode matching with raw reads and a series of filtering process. The observed number of operational taxonomic units (OTUs) at a 3% distance were 412 (R0), 323 (R1), 367 (R2) 305 (R3) and 230 (R4). Chao1 values were investigated as a metric for species richness, together with the corresponding Shannon index for presenting the species diversity; these analyses jointly implied that some microbial communities were enriched selectively.³⁰

Table 2 Biodiversity estimation, using the 16S rRNA gene libraries, from the pyrosequencing analysis

Sample	Effective reads	Observed OTUs	Estimated OTUs by Chao1	Shannon
R0	12 340	412	886	3.375
R1	17 985	323	689	2.531
R2	14 992	367	1076	3.134
R3	18 372	305	536	2.756
R4	18 998	230	570	1.767

---

3.3.2 Bacteria and Archaea taxonomy analyses. To clarify the differences in species diversity between the reactors, the relative microbial community abundances were identified at the phylum level (Fig. 5). Euryarchaeota, Proteobacteria, Actinobacteria, Acidobacteria, Firmicutes, Crenarchaeota, Bacteroidetes, BRC1 and unclassified bacteria dominated in the five samples, and the main differences in diversity were attributed to the distribution of Euryarchaeota, Acidobacteria, Actinobacteria, Proteobacteria and unclassified bacteria. As digestion continued, the microbial community structure significantly shifted in R1, R2, R3 and R4. The relative abundance of unclassified bacteria in R0 was 52.8%, which was significantly higher than that in the other samples (0.70%, 3.80%, 1.90% and 0.60%, respectively). Actinobacteria, as the main fermentative bacteria, had a significant increase in R4 (16.2%) compared to R0 (4.9%); whereas that in R1, R2, and R3 was 5.9%, 6.4% and 4.4%, respectively, contributing to the rapid increase of SCOD and VFAs in R4.³¹ Additionally, the relative abundance of Planctomycetes in R2 (4.8%) was the highest, while that in other samples was very low (near to zero). As previously reported, Planctomycetes, the main type of chemoautotrophic bacteria causing the ANAMMOX process, oxidized ammonium with nitrite as the electron acceptor and CO₂ as the main carbon source. This result identified the existence of ANAMMOX after Fe(NO₃)₃ dosing in the thermophilic anaerobic digestion system, and explained the variation of NH₄⁺–N in the early stage.³² As for Archaea, Methanogens was dominant, and the highest relative abundance was obtained in R4 (93.3%), followed by R1 (91.3%), R2 (74.2%) and R3 (88.4%).

Fig. 5 Taxonomic classification of the dominant phylogenetic groups at the phylum level (the relative abundance of phyla comprising less than 1% of the total composition in the five libraries was defined as “other”).

The sequences from Archaea were also analysed at the genus level as shown in Table 3, and a total of 10 genera were identified as dominating the composition in the five samples. Methanosarcina had its highest relative abundance in R4 (63.2%) as compared with that in R0 (1.3%), R1 (35.7%), R2 (2.7%) and R3 (31.4%), corresponding well with the methane production rate. The relative abundance of the genera Methanosphaerula and Methanomethylovorans had no apparent distinction among R1, R2, R3 and R4, and compared with R4 a higher relative abundance of the genera Methanobacterium, Methanobrevibacter, Methanolinea, Methanospirillum and Methanosaeta in R1, R2 and R3 was identified. At the same time, the genus Methanoculleus was enriched in R4 with its highest relative abundance (5%) as compared with 0.2%, 1%, 4.3% and 1.3% in R0, R1, R2 and R3, respectively.

Table 3 Taxonomic classification of Archaea based on the effective reads at 3% distance at the genus level

Genus	R0	R1	R2	R3	R4
Methanobacterium	1.60%	4.30%	3.20%	3.30%	1.20%
Methanobrevibacter	6.00%	2.60%	1.60%	1.90%	0.50%
Methanothermobacter	2.60%	12.60%	1.80%	6.20%	3.00%
Methanoculleus	0.20%	1.00%	4.30%	1.30%	5.00%
Methanosphaerula	1.90%	0.70%	0.70%	0.60%	0.60%
Methanolinea	0.10%	0.80%	1.40%	1.00%	0.20%
Methanospirillum	1.40%	6.50%	10.40%	7.70%	2.10%
Methanosaeta	4.10%	17.40%	16.20%	16.70%	5.00%
Methanomethylovorans	0.90%	0.30%	0.20%	0.20%	0.20%
Methanosarcina	1.30%	35.30%	2.70%	31.40%	63.20%
Others	5.30%	9.80%	31.70%	18.10%	12.30%
Archaea	25.40%	91.30%	74.20%	88.40%	93.30%

---

Phylogenetic relationships of sequences from the dominant microbial community (bacteria and Archaea) of each sample are shown in Fig. 6. These sequences were assigned into NCBI taxonomies with BLAST and MEGAN.¹⁶ Pie charts indicate the relative abundance of each phylum, class, family and genus. The relative abundances of the corresponding phylum, class, family and genus in the five samples were defined as the ratio of each coloured area to pie area. Although most of the genera were found simultaneously in each sample, there were still differences in terms of the relative abundances. The main bacteria in R0 was the genus Proteobacteria, while that in R1, R2, R3 and R4 was genus Firmicutes, indicating that the genus Firmicutes was enriched and that the dominant status of genus Proteobacteria was weakened during the digestion process. Some specific species of Archaea were selectively enriched,³³ and the genera Methanobacterium, Methanothermobacter and Methanosaeta were the dominant communities in R0 and R1, while those in R2 were genera Methanoculleus, Methanospirillum and Methanosaeta. It should be emphasized that genera Methanosarcina and Methanoculleus were the dominant communities in R4 (colored yellow), which accounts for the higher methane production.

Fig. 6 Phylogenetic relationships of sequences from the dominant bacterial communities in the five samples.

3.4 Possible mechanisms analysis

Ferric salts have been widely used for their outstanding performance in waste water treatment for sulfide and phosphorus control, COD and color removal as shown in Table 4. Similarly, we have found that the addition of ferric salts affects the thermophilic anaerobic digestion of WAS significantly, and that FeCl₃ was the best choice for the enhancement of methane production, with an increase in methane production rate of 98.9% under the test conditions. The introduction of FeCl₃ in R4 might have helped establish a balance between the hydrolysis and methanogenic phases. As shown in Fig. 3b, rapid hydrolysis and acidification took place at an early stage, and SCOD was converted to VFAs; most of the VFA was acetic acid (Fig. A2†), a suitable substrate for methanogenesis,⁷ such that the VFAs generated were immediately converted to methane. In general, Fe(III) is reduced to Fe(II) chemically or biologically by iron-reducing bacteria (IRB) during the anaerobic digestion process according to the Pourbaix diagram for iron,³⁸ and hence the iron in the anaerobic system was mainly in the form of ferrous ion.³⁹ The concentration of iron ion in the supernatant was analyzed and there was no apparent distinction among the reactors (data not shown) to indicate the formation of iron precipitates (e.g. ferrous sulfide and monobasic ferrous phosphate) or adsorption of iron to granules, as reported by Ge et al. and Firer et al.^9,40 The sludge matrix absorption of metals at specific sites was significant due to the high porosity and large internal surface area of the granules,⁴¹ and it was confirmed by the EDS profile shown in Fig. A3.† The formation of FeS led to a reduction in the release of H₂S, which is toxic to methanogens: this might be one reason for the methane production rate increase.⁴² The content (%) of sulfur in dry sludge solid before and after the anaerobic digestion was investigated, and a higher sulfur content in R2 (1.187%), R3 (1.399%) and R4 (1.277%) was obtained, compared with the control experiment (1.128%), confirming the effect of iron ion on sulfur control.

Table 4 The main functions of ferric salts used in the waste water treatment

Ferric salt species	Treatment objective	Dosage (mg L^−1)	Waste water categories	Function(s)	Removal efficiency	Reference
FeCl3, Fe2(SO4)3	Biomass removal	150–250	Piggery waste water	Coagulation, flocculation	66–98%	34
FeCl3	Sulfide control	0.4–5.4	Domestic waste water	Oxidation, precipitation	—	35
FeCl3	Phosphorus control	0.44	Domestic sewage	Precipitation	100%	10
FeCl3	COD removal	100–200	Slaughterhouse waste water	Precipitation	45–75%	36
FeCl3	Color removal	3500	Molasses waste water	Co-precipitation	96%	37

---

In addition, iron is an indispensable component of pyruvate-ferredoxin oxidoreductase (POR), which catalyzes pyruvate catabolism with the production of CO₂ and acetyl-CoA; consequently, iron has a direct influence on biological oxidation, and thus methane production was improved in the presence of iron. In this study, it was found that the addition of FeCl₃ into a thermophilic anaerobic digestion system optimized the microbial community structure; the possible mechanisms involving iron through which this is achieved are shown in Fig. 7. After FeCl₃ dosing of R4, certain bacteria and Archaea were selectively enriched; namely, the dominant communities shifted from Bacteroidetes and Proteobacteria to Firmicutes, and the main Archaea identified in R4 was Methanosarcina. According to the literature, Methanosarcina is the only one that utilizes all the three methane production pathways (i.e. the acetoclastic methanogenesis, hydrogenotrophic methanogenesis, and the methyl compound utilizing pathway) and tolerates different adverse factors, such as temperature variation or a high concentration of acetic acid.³⁰ When methanogens became dominant, as the main products of hydrolysis acidification VFAs were degraded to methane in a timely manner and the pH increased, creating a favorable environment for stable methanation.

Fig. 7 Possible mechanisms of different ferric salts in the thermophilic anaerobic digestion of WAS (only the main methanogenesis process and key enzyme assayed are labelled).

It should be emphasized that both the anions and the cations in ferric salts should be considered: it has been observed in this work that not all the ferric salts contribute to methane production because the addition of ferric salts not only introduces micronutrients for microorganisms, but also creates a special environment for anaerobic digestion. Specially, the introduction of NO₃⁻ resulted in a low pH and high ORP, accompanied by an AMMNOX process, creating adverse conditions for methanogens. In terms of SO₄²⁻, the ferric salt introduction was beneficial to the oxidation of organics by SRB, but the simultaneous competition for common organic and inorganic substrates also resulted in a gradual methanogenesis inhibition. Nevertheless, the addition of FeCl₃ enhanced the methane production in the thermophilic anaerobic digestion system, attributed to the positive effects of iron and the favorable anaerobic conditions created.

4. Conclusions

Among the three test ferric salts, FeCl₃ was the best at enhancing methane production from WAS, with a 98.9% higher methane production rate than the control experiment. The reason for the enhanced methane production was the favorable environment created, which we assessed using the pH, ORP and variation of VFAs, SCOD, TN and NH₄⁺–N. Interestingly, the introduction of NO₃⁻ resulted in a low pH and high ORP, accompanied by AMMNOX, and a gradual methanogenesis inhibition occurred with the addition of Fe₂(SO₄)₃. In the thermophilic anaerobic system, apart from the positive effects of precipitation of sulfide and biocatalysis for the methanogenesis process, optimization of the microbial community structure was also identified. Pyrosequencing analysis of microbial communities before and after the introduction of ferric salts showed that some specific species of bacteria and Archaea were selectively enriched. The enhanced methane production in the reactor with FeCl₃ added should be attributed to the enrichment of Actinobacteria for hydrolysis acidification and Methanosarcina for methanogenesis, with the values of 16.2% and 63.2%, respectively.

Acknowledgements

This study was financially supported by the National Hi-Tech Research and Development Program of China (863) (no. 2011AA060906), the National Natural Science Foundation of China (no. 51178261), and the Key Project of Science and Technology Commission of Shanghai Municipality (no. 12231202101, 14DZ1207306).

Notes and references

1. R. R. Wang, M. J. Eckelman and J. B. Zimmerman, Environ. Sci. Technol., 2013, 47, 11189–11198.
2. O. Thygesen, S. G. Sommer, S. G. Shin and J. M. Triolo, Fuel, 2014, 132, 44–46.
3. K. V. Lo, P. H. Liao and A. C. March, Biomass, 1985, 6, 301–315.
4. J. P. Sheets, X. M. Ge and Y. B. Li, Bioresour. Technol., 2015, 180, 296–303.
5. F. X. Ye, Y. Y. Ye and Y. Li, J. Hazard. Mater., 2011, 188, 37–43.
6. Y. Xia, H. H. Fang and T. Zhang, RSC Adv., 2013, 3, 15528–15542.
7. G. Y. Zhen, X. Q. Lu, J. Niu, L. H. Su, X. L. Chai, Y. C. Zhao, Y. Y. Lin, Y. Song and D. J. Niu, Chem. Eng. J., 2013, 233, 274–281.
8. L. Zhang and D. Jahng, Waste Manage., 2012, 32, 1509–1515.
9. H. Q. Ge, L. S. Zhang, D. J. Batstone, J. Keller and Z. G. Yuan, J. Environ. Eng., 2013, 139, 594–601.
10. O. Gutierrez, D. Park, K. R. Sharma and Z. G. Yuan, Water Res., 2010, 44, 3467–3475.
11. H. P. V. Leeuwen, Environ. Sci. Technol., 1999, 33, 3743–3748.
12. J. Gustavsson, S. S. Yekta, C. Sundberg, A. Karlsson, J. Ejlertsson, U. Skyllberg and B. H. Svensson, Appl. Energy, 2013, 112, 473–477.
13. APHA, Standard methods for the examination of water and wastewater, 20st edn, 1998.
14. B. Yu, J. B. Xu, H. P. Yuan, Z. Y. Lou, J. X. Lin and N. W. Zhu, Fuel, 2014, 130, 279–285.
15. Y. Yu, C. Lee, J. Kim and S. Hwang, Biotechnol. Bioeng., 2005, 89, 670–679.
16. J. X. Zhang, Y. B. Zhang, X. Quan and S. Chen, Water Res., 2013, 47, 5719–5728.
17. L. Ye, M. F. Shao, T. Zhang, A. H. Y. Tong and S. Lok, Water Res., 2011, 45, 4390–4398.
18. H. N. Gavala, U. Yenal, I. V. Skiadas, P. Westermann and B. K. Ahring, Water Res., 2003, 37, 4561–4572.
19. D. Mamais, P. A. Pitt, Y. W. Cheng, J. Loiacono and D. Jenkins, Water Environ. Res., 1994, 66, 912–918.
20. H. Bouallagui, Y. Touhami, R. B. Cheikh and M. Hamdi, Process Biochem., 2005, 40, 989–995.
21. X. G. Liu, B. Dong and X. H. Dai, Bioresour. Technol., 2013, 148, 461–466.
22. A. Veeken and B. Hamelers, Bioresour. Technol., 1999, 69, 249–254.
23. S. Xie, P. G. Lawlor, J. P. Frost, Z. Hu and X. Zhan, Bioresour. Technol., 2011, 102, 5728–5733.
24. Y. Santosh, T. R. Sreekrishnan, S. Kohli and V. Rana, Bioresour. Technol., 2004, 95, 1–10.
25. T. Jash and D. N. Ghosh, Energy, 1996, 21, 725–730.
26. J. Weijma, A. J. M. Stams, L. W. H. Pol and G. Lettinga, Biotechnol. Bioeng., 1994, 28, 355–367.
27. H. Harada, S. Uemura and K. Monomoi, Water Res., 1994, 28, 355–367.
28. Y. S. Wong, T. T. Teng, S. A. Ong, N. Morad and M. Rafatullah, RSC Adv., 2014, 4, 64659–64667.
29. N. Chamchoi, S. Nitisoravut and J. E. Schmidt, Bioresour. Technol., 2008, 99, 3331–3336.
30. T. M. Gihring, S. J. Green and C. W. Schadt, Environ. Microbiol., 2011, 14, 285–290.
31. J. D. Vrieze, T. Hennebel, N. Boon and W. Verstraete, Bioresour. Technol., 2012, 112, 1–9.
32. Y. Kimura, K. Isaka and F. Kazama, Bioresour. Technol., 2011, 102, 4390–4394.
33. D. Y. Cheong and C. L. Hansen, Appl. Microbiol. Biotechnol., 2006, 72, 635–643.
34. I. D. Godos, H. O. Guzman, R. Soto, P. A. García-Encina, E. Becares, R. Muñoz and V. A. Vargas, Bioresour. Technol., 2011, 102, 923–927.
35. A. H. Nielsen, P. Lens, J. Vollertsen and T. H. Jacobsen, Water Res., 2005, 39, 2747–2755.
36. L. A. Núñez, E. Fuente, B. Martínez and P. A. García, J. Environ. Sci. Health, Part A: Toxic/Hazard. Subst. Environ. Eng., 1999, 34, 721–736.
37. Z. Liang, Y. X. Wang, Y. Zhou and H. Liu, Chem. Eng. J., 2009, 152, 88–94.
38. H. A. M. Casillas, D. Cocke, J. A. G. Gomes, P. Morkovsky, J. R. Parga and E. Pterson, Sep. Purif. Technol., 2007, 56, 204–211.
39. S. C. Peng, J. Xue, C. B. Shi, J. Wang, T. H. Chen and Z. B. Yue, Fuel, 2014, 117, 1–4.
40. D. Firer, E. Friedler and O. Lahav, Sci. Total Environ., 2008, 392, 145–156.
41. Y. Liu, H. L. Xu, S. F. Yang and J. H. Tay, Water Res., 2003, 37, 661–673.
42. Y. Chen, J. J. Cheng and K. S. Creamer, Bioresour. Technol., 2008, 99, 4044–4064.

---

Footnote

† Electronic supplementary information (ESI) available: The ratio of CO₂ to CH₄, the variation of VFAs (including acetic acid, propionic acid, butyric acid, isobutyric acid, valeric acid and isovaleric acid), and EDS analysis of the sludge samples. See DOI: 10.1039/c5ra02362a

---