Bioprocess Biosyst Eng DOI 10.1007/s00449-017-1810-2
RESEARCH PAPER
The performance and microbial communities of biodegradationelectron transfer with sulfur metabolism integrated process for flue gas desulfurization wastewater treatment Chao Wei1 • Wenjie He2 • Li Wei1 • Jun Ma1 • Chunying Li3
Received: 26 February 2017 / Accepted: 29 June 2017 Ó Springer-Verlag GmbH Germany 2017
Abstract The biodegradation-electron transfer with sulfur metabolism integrated (BESIÒ) process was used for the treatment of real flue gas desulfurization wastewater. The BESIÒ process consists of an anaerobic activated sludge reactor, an anoxic activated sludge reactor, and an aerobic bio-film reactor. The performance of the integrated process was evaluated by the removal efficiencies of organics and nitrogen pollutants. The sulfate in the wastewater was used as an abundant sulfur source to drive the integrated process. The removal efficiencies of chemical oxygen demand, total organic carbon, ammonia nitrogen, and total nitrogen of the integrated process were 87.99, 87.04, 30.77, and 45.17%, respectively. High-throughput 454-pyrosequencing was applied for the analysis of microbial communities in the integrated process. From the anaerobic activated sludge (Sample 1), anoxic activated sludge (Sample 2), and aerobic bio-film (Sample 3), totals of 1701, 1181, and 857 operational taxonomic units were obtained, respectively. The sulfur cycle was associated with the removal of organics and nitrogen pollutants. The sulfate-reducing bacteria participated in the organics removal in the anaerobic reactor, and the sulfide oxidation was related with the
Electronic supplementary material The online version of this article (doi:10.1007/s00449-017-1810-2) contains supplementary material, which is available to authorized users. & Li Wei
[email protected] 1
State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology, Harbin, Heilongjiang, China
2
Tianjin Waterworks Group Co., Ltd., Tianjin, China
3
School of Energy and Civil Engineering, Harbin University of Commerce, Harbin, Heilongjiang, China
denitrification in the anoxic reactor. A complete nitrogen degradation chain was built in the integrated process. Through the degradation chain, the nitrogenous organic pollutants, ammonia nitrogen, and nitrate could be removed. The participant functional bacteria were also detected by pyrosequencing. Keywords Flue gas desulfurization (FGD) wastewater BESIÒ process 454-pyrosequncing Sulfate-reducing bacteria (SRB) Denitrification
Introduction The biodegradation of sulfate in wastewater relies on the conversion of S with different valences, where biological sulfate reduction (BSR) and biological reduced sulfur oxidation (BSO) participate in S reduction and S oxidization, respectively [1–3]. The two processes are spatially separated to make use of different features to achieve different functions. In an anaerobic environment, sulfate-reducing bacteria (SRB) obtain biochemical energy by reducing sulfate to sulfide, and the organic compounds as electron donors could be removed simultaneously [4]. The sulfide generated from the sulfate reduction tends to dissolve in water as the pH increases. The minimal chemical oxygen demand (COD) requirement in sulfate reduction is two grams of COD consumed per gram of SO42--S reduced [1, 5]. Compared with methanogens, SRB have a much wider spectrum of substrate. From the thermodynamic point of view, the reduction of sulfate to sulfide by SRB releases more energy than the production of methane by methanogens, thereby enabling SRB to out-compete methanogens [6, 7]. The sulfide generated from the sulfate reduction could be oxidized under anoxic conditions.
123
Bioprocess Biosyst Eng
Under oxygen-limited circumstances, sulfur is the major end-product of the sulfide oxidation, whereas sulfate is formed under sulfide-limited circumstances [8–10]. According to the biological reduction and oxidation of sulfur, the biodegradation-electron transfer with sulfur metabolism integrated (BESIÒ) process was designed for the treatment of petrochemical wastewater [11]. The contained sulfate was used to drive the integrated process. The BSR process was used for organics’ removal at anaerobic condition, and the BSO process was used for denitrification at anoxic condition. In the study, it consists of an anaerobic activated sludge reactor, an anoxic activated sludge reactor, and an aerobic bio-film reactor. In the anaerobic reactor, the sulfate was reduced to sulfide by SRB, and the organics were removed simultaneously. When nitrate exists in wastewater, the sulfide generated in the anaerobic reactor is able to serve as an electron donor for autotrophic denitrification in an anoxic reactor [12, 13]. The ammonia nitrogen (AN) could be nitrified to nitrate in an aerobic reactor, and the nitrate recirculated to an anoxic reactor for denitrification. Thereby, organic compounds, nitrate, and sulfate could be simultaneously removed in the BESIÒ process. Sulfate-abundant wastewater could serve as a low-cost sulfur source to drive the BESIÒ process for wastewater treatment. In this research, the BESIÒ process was applied for the treatment of flue gas desulfurization (FGD) wastewater. FGD systems emerged in the industrial field of the coal-burning power plants and for some industrial processes in the early 1970s in the United States (US) and Japan for the reduction of SO2 emissions from large electric utility boilers [14, 15]. In the FGD process, FGD wastewater was produced as aqueous by-product. Filtered water is the effluent produced as a result of the gypsum slurry filtration. In the FGD systems, the downflow** gypsum slurry contacts with the up-flow flue gas, and absorbs the SO2 into water stream. The absorption solution accumulates at the bottom of the scrubber, and the absorption solution is recirculated. To avoid the saturation of pollutants, part of the concentrate should be removed, and fresh water should be replenished to the recirculated water. The removed concentrate need to be dehydrated [14, 16]. Biological treatment of FGD wastewater through sulfate reduction can remove SO2 as well as heavy metals such as mercury and lead from flue gas [17, 18]. The aqueous phase from the dehydrated concentrate was used in our research. There have already been some studies on SRB in sulfate-rich biological reactors and the impact of sulfate on the performance of the biological reactor [19–21]. In the mixed culture, the interaction between SRB and other anaerobic bacteria result in the reduction of sulfate to sulfide, so it is necessary to analyze the microbial
123
community structures in the reactors. The analysis of the microbial community would result in understanding the relationship between the performance of the reactor and the microbial community structures. Recently, high-throughput pyrosequencing has shown advantages for the analysis of microbial taxa, and this method can generate enormous amounts of DNA reads through a massively parallel sequencing-by-synthesis approach. This technology has been widely used to analyze the microbial community in various environmental samples, including permafrost soils around oil pipes, anaerobic sludge for saline wastewater treatment, and samples from nitrification reactors. [6, 22–24]. In the study, the BESIÒ process was used for FGD wastewater treatment, which could provide sulfate and organics. The performance of the integrated process was evaluated by the removal efficiencies of COD, total organic carbon (TOC), AN, and total nitrogen (TN). We analyzed the microbial communities in the process using 454-pyrosequencing, and the functional and abundant phylotypes were detected.
Materials and methods Experimental setup and operation The laboratory-scale synthetic glass anaerobic internal circulation reactor, anoxic reactor, and bio-film aerobic reactor were used as an integrated process in our research (Fig. 1). The anaerobic, anoxic, and aerobic reactors had working volumes of 4, 5, and 20 L, respectively. The activated sludge used in our study, came from Chengfengzhuang wastewater treatment plant in Daqing, China. The sludge was diluted with distilled water and FGD wastewater, and the volume ratio of sludge, distilled water, and FGD wastewater was 1:2.5:2.5. We removed the impurities in the activated sludge, and placed the activated sludge at anaerobic and anoxic conditions for acclimation, respectively. The polyurethane fillers were mixed with the activated sludge, and put them in aerated container. The porosity of the used polyurethane fillers was over 99.5%, and the specification of the used polyurethane fillers was 2 cm 9 2 cm 9 2 cm. The polyurethane fillers have good adhesive performance, and the bacteria could attach on them. After 30 days acclimation, we transferred the activated sludge and fillers into corresponding reactors. The acclimated activated sludge was deposited in the anaerobic reactor and anoxic reactor. In the aerobic reactor, the polyurethane fillers formed a filler layer with the thickness of 20 cm, and a continuous air supply system was set beneath the filler layer. Half of the effluent of the aerobic reactor was returned to the anoxic reactor. The reflux
Bioprocess Biosyst Eng Fig. 1 Experimental setup of the BESIÒ process
wastewater contained high concentration of dissolved oxygen, which affects the operating condition of the anoxic reactor. The mixture of the reflux and anaerobic effluent served as the influent of the anoxic reactor to provide anoxic conditions for the microbial community. An outlet was set for the N2 overflow at the top of the anoxic reactor. The anaerobic and anoxic reactors were equipped with oxidation–reduction potential (ORP) probes to monitor the anaerobic and anoxic conditions, respectively. There were settling sections at the tops of all three reactors that realized the separation of water and activated sludge. In the three reactors, only the supernatant flowed into the next reactor, so the suspended solids were very low in the effluent, and there was no need to set up a secondary sedimentation tank. The anaerobic and anoxic reactors were operated at 37 °C using a thermostatic resistance wire, and the aerobic reactor was operated at 30 °C using electric heaters. The FGD wastewater was fed into the integrated process continuously for 20 days as further acclimation; the HRT for anaerobic reactor was 48 h in the period. To determine the effect of hydraulic retention times (HRTs), the integrated process was operated at five different HRTs, and the operational time lasted 10 days for every HRT. We took the total COD removal efficiency as a measure of the performance of the integrated process, and chose the optimum HRT. Analysis methods The water samples were collected at every 2 days’ intervals and analyzed immediately The COD was measured using the potassium dichromate titrimetric method; ammonia nitrogen was detected using Nessler reagent colorimetry [25]. The dissolved sulfide was measured using the iodometric method with a starch indicator [25]. TOC
and TN were detected using a TOC/TN analyzer (Shimadzu TOC-5000A). Sulfate was analyzed using an ion chromatograph (HIC-20A Super) with a conductivity detector and an IC-SA2 analytical column. Statistical analysis The statistical analysis was performed through one-way ANOVA using a Duncan test (P \ 0.05). Tests were implemented using SPSS 20.0. DNA extraction, amplification and sequencing We collected microbial samples from the three reactors on Day 100. The activated sludge was collected from three different layers in the anaerobic and anoxic reactors, and merged them together, respectively. In the aerobic reactor, we selected the polyurethane fillers from five different parts, and collected the attached bio-film together. The microbial samples collected from the anaerobic, anoxic, and reactors were marked as Sample 1, Sample 2, and Sample 3, respectively. The microbial samples were washed, centrifuged (96009g, 3 min) for four times and treated by ultrasonic waves for cell lyses. DNA was extracted using the PowerSoil DNA extraction kit (MO BIO Laboratories, Inc., Carlsbad, CA), and amplified using universal bacterial primer 8F (50 -30 AGAGTTTGATCCTGGCTCAG) and 533R (50 -30 TTACCGCGGCTGCTGGCAC) covering the V1 and V3 regions. To distinguish different microbial samples, different ten-nucleotide barcode sequences and pyrosequencing adapters were added at the 50 end of the universal bacterial primer. PCR products were purified and quantified, and the processed PCR products were used for 454 pyrosequencing using the Roche 454 FLX Titanium platform. Before the clustering of operational taxonomic
123
Bioprocess Biosyst Eng
units (OTUs), the sequences were examined, including the base mismatches, average base quality, ambiguous ‘‘N’’, the sequences length, repeat bases, and chimeras. The processed sequences were assigned to samples according to barcodes. The sequences were aligned using Mothur ver. 1.17.0 and clustered into operational taxonomic units (OTUs) at 97% similarities. Taxonomic classification of the sequences were performed using the RDP Classifier of the Ribosomal Database Project (RDP), the National Centre for Biotechnology Information (NCBI) BLAST, and the Greengenes databases at 70% confidence threshold. The sequence data have been submitted to the NCBI Sequence Read Archive database (Accession Numbers: SRR2375033, SRR2375034, and SRR2375035 for Sample 1, Sample 2, and Sample 3, respectively).
Results Performance of the integrated process In the period of HRT determination, there were no significant difference of the total COD removal efficiencies at 48, 36, and 24 h HRTs (anaerobic reactor). Then we raised the flow velocity, and the total COD removal efficiencies decreased to 78.80 and 66.26% at 18 and 12 h HRTs, respectively (Fig. 2). We selected the 24 h as optimum HRT for anaerobic reactor based on the results presented in Experimental setup and operation, and the integrated system was operated at steady state conditions for 100 days. Similarly, the optimum HRTs for anoxic and aerobic reactors were 15 and 60 h, respectively. In steady operating period, the ORP of the anaerobic and anoxic reactors were -390 and -186 mv, respectively.
Fig. 2 The total COD removal efficiencies at five different HRTs. The a, b, and c presented different significant differences. Statistically significant differences at P B 0.05
123
Table 1 Characteristics of the FGD wastewater Parameters
Influent mean
TOC (mg/L)
103.01 ± 4.79
COD (mg/L)
266.59 ± 7.83
Sulfate (mg S/L)
504.10 ± 12.54
Ammonia nitrogen (mg/L)
2.86 ± 0.34
Nitrite (mg N/L)
4.76 ± 0.57
Nitrate (mg N/L)
20.25 ± 1.70
TN (mg/L)
25.37 ± 1.81
pH
8.22 ± 0.11
The pH values of the anaerobic, anoxic, and aerobic reactors were 8.06, 8.12, and 8.77, respectively (Fig. S1). The characteristics of the FGD wastewater are shown in Table 1. The TOC, COD, AN, TN, and sulfate concentrations of the influent were 103.01 ± 4.79, 266.59 ± 7.83, 2.86 ± 0.34, 25.37 ± 1.81 mg/L, and 504.10 ± 12.54 mg S/L, respectively. The FGD wastewater was a kind of highsulfate petrochemical wastewater that has low organic concentration. The TOC concentration of the effluent was 13.35 ± 1.72 mg/L, and the total TOC removal efficiency was 87.04% (Fig. 3a). The integrated process could reach an 87.99% removal efficiency of COD, and the COD concentration of the effluent was 32.02 ± 8.00 mg/L (Fig. 3b). The real FGD wastewater was used in our research. The components of COD and TOC were complex in the real wastewater, so the fluctuation existed in the water quality of effluent. The AN and TN concentrations of the effluent were 1.98 ± 0.31 and 13.91 ± 1.13 mg/L, and their total removal efficiencies in the integrated process were 30.77 and 45.17%, respectively (Fig. 3c). Role of SRB in the anaerobic reactor As shown in Fig. 3b, the effluent COD concentration was 51.54 ± 7.24 mg/L after the treatment in anaerobic reactor. In the anaerobic reactor, 215.05 mg/L COD was removed, and the removal efficiency was 80.67%. The effluent sulfate concentration of the anaerobic reactor was 431.03 ± 9.16 mg S/L. Hence, in this anaerobic reactor, 73.07 mg S/L of sulfate was reduced (Fig. 3d). In the metabolic process of SRB, the theoretical COD consumption is 2 grams for 1 gram of sulfate-S reduction. The theoretical value of COD removed by SRB, was 146.14 mg/L (73.07 9 2), so approximately 67.96% (146.14/215.05) COD was removed by SRB in the anaerobic reactor. The FGD wastewater contained enough sulfate, and the ratio of COD to sulfate-S was lower than two in the influent, which means that the sulfate was sufficient for SRB. The COD could provide carbon source for microbial growth. The component of the COD was
Bioprocess Biosyst Eng
Fig. 3 The performance of the integrated process. a Concentration of TOC in influent and effluents of different reactors; b concentration of COD in influent and effluents of different reactors; c TN and AN removal efficiencies of the integrated process; d sulfate concentration
of influent, anaerobic effluent, and anoxic effluent (represented by black lines and symbols) and sulfide concentration of anaerobic and anoxic effluents (represented by blue lines and symbols) (color figure online)
complex in the real FGD wastewater. The different organics could serve as substances for the metabolism of different bacteria.
of the sulfide was recovered as sulfate, so no elemental sulfur accumulated in the anoxic reactor. Research shows that in the autotrophic denitrification process, sulfide serves as electron donor, and the S/N ratio affects the final product of sulfide oxidation [26]. So, theoretically, the final product of sulfide oxidation was sulfate, which can be described by Eqs. (1) and (2) [13] as follows,
Role of sulfide in the anoxic reactor The sulfide could be used as an electron donor in autotrophic denitrification. In the effluent of the anaerobic reactor, the sulfide concentration was 43.62 ± 4.18 mg/L. The sulfide was transferred into the anoxic reactor, and the sulfide concentration was 25.96 ± 3.31 mg/L in the effluent (Fig. 3d). Thus, 17.66 mg/L sulfide was removed in the anoxic reactor, and the sulfide removal efficiency was 40.49%. The effluent of the anoxic reactor contained 449.73 ± 8.61 mg S/L sulfate, and the influent contained 431.03 ± 9.16 mg S/L. Thus, 18.70 mg S/L sulfate was generated in the anoxic reactor. The results showed that all
þ 2 S2 þ 1:6NO 3 þ 1:6H ! SO4 þ 0:8N2 þ 0:8H2 O DG ¼ 743:9 ðKJ mol Þ;
ð1Þ
þ S2 þ 0:4NO 3 þ 2:4H ! S þ 0:2N2 þ 1:2H2 O DG ¼ 191:0 ðKJ mol Þ:
ð2Þ
According to Eq. (1), 12.36 mg N/L of nitrate was removed, which was related to sulfide oxidation in the anoxic reactor. The TN concentration was not high in the FGD wastewater, and the total removal efficiency
123
Bioprocess Biosyst Eng
of TN was 11.46 mg N/L. Thus, all the TN removal was associated with the oxidation of sulfide in the anoxic reactor. Diversity of microbial communities The rarefaction analysis of the bacterial communities derived from the anaerobic (Sample 1), anoxic (Sample 2), and aerobic (Sample 3) reactors is depicted as having a 97% similarity. The three approached saturation curves indicated that the sequencing covered nearly all of the OTUs in the three samples (Fig. 4a). From the anaerobic activated sludge (Sample 1), anoxic activated sludge (Sample 2), and aerobic bio-film (Sample 3), totals of 54528, 45762, and 53916 valid sequence reads were obtained, respectively. The coverage indices of the three samples were all over 99%, which indicates that the recovered sequences well represent the microbial diversity of the three samples (Table S1).
Fig. 4 Analysis of microbial diversities. a Rarefaction analysis of the different samples. Rarefaction curves are depicted at the 3% dissimilarity level; b rank abundances show pyrosequencing
123
The well-distributed rank-abundance curves show the relative abundance of microbial communities in the three samples, and the width of the x-axis represents the abundance (Fig. 4b). The values of the ACE, Chao, Shannon, and Simpson indices further support this result. The values of the ACE and Chao indices reflect the abundance of the microbial communities, and the values of the Shannon and Simpson indices reflect their diversity (Table S1). The abundances of microbial communities in descending order were Sample 1, Sample 2, and Sample 3, and the diversities were in the same order. The unique and shared OTUs were represented by a Venn diagram, which could reflect the similarity of microbial composition between different samples (Fig. 4c). There were 261 OTUs common for all three samples. Besides the 261 OTUs, there were 530 OTUs shared by Sample 1 and Sample 2, and the quantity was over than the unique OTUs of Sample 2. There were 61 OTUs common for Sample 2 and Sample 3, and 165 OTUs for Sample 1 and Sample 3.
abundances of different samples; c Venn diagram showing unique and shared OTUs between different samples
Bioprocess Biosyst Eng
Fig. 5 Microbial composition at the phylum and class levels. The microbial phylotypes and the relative abundances were represented by bars of different colors and sizes across the three samples at the phylum (a) and class (b) levels (color figure online)
Microbial community structure The microbial compositions of the three samples at the phylum level are shown in Fig. 5a and Table S2. Proteobacteria was the most dominant (average abundance [10%) phylum in the three samples, accounting for 32.73, 44.25, and 24.71% in Sample 1, Sample 2, and Sample 3, respectively. In Sample 1, phyla Proteobacteria and Firmicutes (accounting for 24.60%) were dominant, followed by a few other abundant phyla, including Actinobacteria (9.51%), Chloroflexi (8.87%), Planctomycetes (6.87%), Synergistetes (5.27%), Bacteroidetes (2.11%), Thermotogae (1.92%), and Cyanobacteria (1.28%). In Sample 2, the dominant phyla Proteobacteria and Firmicutes (accounting for 12.90%) accounted for a total of 57.16%, and the other abundant phyla were Chloroflexi (7.94%), Planctomycetes (7.81%), Actinobacteria (6.72%), Bacteroidetes (5.10%),
and Synergistetes (4.55%). In Sample 3, the abundance of phylum Acidobacteria (22.47%) was similar to that of phylum Proteobacteria, and these two phyla accounted for a total of 47.17%. The phyla Planctomycetes (15.68%) and Actinobacteria (15.42%) were also dominant, and phyla Armatimonadetes (7.27%), Nitrospirae (5.69%), Firmicutes (4.01%), Gemmatimonadetes (3.27%), and Chloroflexi (2.70%) were additionally abundant in Sample 3. At the class level, Clostridia (23.16%) was the most dominant class in Sample 1. The class Clostridia is one group of phylum Firmicutes and accounted for 24.60% of the specimens in Sample 1. Most of the phylum Firmicutes was composed of class Clostridia in Sample 1, and the class accounted for 11.49% and 3.09% in Sample 2 and Sample 3, respectively. Within phylum Proteobacteria, class Alphaproteobacteria (14.43%) was dominant in Sample 1, followed by the Gamma- (10.35%), Beta-
123
Bioprocess Biosyst Eng
(4.26%), and Delta- (3.59%) subdivisions. In Sample 2, Alphaproteobacteria (30.80%) was also the most dominant class within the phylum Proteobacteria, followed by the Beta- (4.88%), Delta- (4.39%), and Gamma- (4.17%) subdivisions. The Alpha-, Beta-, Delta-, and Gamma-subdivisions accounted for 13.57, 8.38, 1.79, and 0.91% of the specimens within the phylum Proteobacteria in Sample 3, respectively. In addition to these classes, the three samples shared other abundant classes, including Planctomycetacia, Actinobacteria Caldilineae, and Phycisphaerae. In Sample 1 and Sample 2, Anaerolineae, Synergistia, and Thermoleophilia were abundant, but the abundances of these three classes were low in Sample 3. In Sample 1, the other abundant classes were Coriobacteriia (3.40%), Thermotogae (1.92%), OM190 (1.61%), SM1D11 (1.08%), and Erysipelotrichia (1.03%). At the class level, the microbial communities of Sample 3 were quite different from those in Sample 1 and Sample 2. In Sample 3, Acidobacteria (22.47%) was the most dominant class, but the abundances of Acidobacteria and phyla Armatimonadetes_norank (7.27%), Nitrospira (5.69%), Gemmatimonadetes (3.27%), and Acidimicrobiia (1.99%) were low in Sample 1 and Sample 2 (Fig. 5b; Table S3). The hierarchical heatmap is based on the abundant genera in each sample (Fig. 6). In Sample 1, Clostridium (16.45%) was the most dominant genus, and the other abundant genera were KCM-B-112_norank (5.92%), Rhodopseudomonas (3.39%), Atopobium (2.64%), Peptostreptococcaceae Incertae Sedis (2.55%), Thermovirga (2.41%), Mesotoga (1.89%), OM190_norank (1.61%), Acinetobacter (1.57%), Desulfobulbus (1.43%), Defluviicoccus (1.32%), and SM1D11_norank (1.08%). The SRB were detected as abundant functional phylotypes in Sample 1, including Desulfobulbus, Desulfonatronum, Desulforhabdus, Desulfovibrio, Desulfuromonas, Desulfotomaculum, Desulfurispora, Desulfofustis, and Desulfococcu. In Sample 2, the abundant genera in order were Azoarcus (3.86%), vadinHA17_norank (3.66%), Pirellula (3.61%), Anoxynatronum (3.24%), Run-SP154_norank (2.96%), Thermovirga (1.91%), Methylocystis (1.46%), SRB2_norank (1.27%), PeM15_norank (1.22%), Peptostreptococcaceae Incertae Sedis (1.17%), Hyphomicrobium (1.15%), and MNG7_norank (1.12%). In Sample 3, the microbial compositions were quite different from those in Sample 1 and Sample 2. The most dominant genus was Blastocatella (20.82%), and the other abundant genera were Armatimonadetes_norank (7.27%), SM1A02 (6.62%), Hyphomicrobium (6.33%), Nitrospira (5.69%), Gordonia (5.04%), KCM-B-112_norank (4.72%), Planctomyces (4.60%), Mycobacterium (2.05%), Clostridium (1.80%), Legionella (1.34%), Sh765B-TzT-29_norank (1.33%), OCS155_marine_group_norank (1.12%), and MSB-1E8_norank (1.09%) (Table S4).
123
The microbial compositions of the three samples were different, even in the shared genera, which had abundances that differed between the three samples. The microbial compositions at the genus level could be divided into different situations. Some genera were detected in all three samples, and their relative abundances were similar, including Peptostreptococcaceae_Incertae_Sedis, Dietzia, Microbacterium, and Leucobacter. Some genera were detected in only one sample or had an abundance in one sample that was much higher than that in the other two samples, including Clostridium (16.45%), Rhodopseudomonas (3.39%), Atopobium (2.64%), Mesotoga (1.89%), Acinetobacter (1.57%), Defluviicoccus (1.32%), Turicibacter (0.94%), Aquabacterium (0.78%), Hydrogenophaga (0.77%), and Aquamicrobium (0.67%) in Sample 1; Azoarcus (3.88%), Pirellula (3.61%), Anoxynatronum (3.24%), Methylocystis (1.46%), Truepera (0.83%), SHA-109_norank (0.54%), and Afipia (0.52%) in Sample 2; and Blastocatella (20.82%), SM1A02 (6.62%), Hyphomicrobium (6.33%), Nitrospira (5.69%), Planctomyces (4.60%), Bauldia (0.81%), Bacillus (0.51%), and Nitriliruptor (0.41%) in Sample 3. The abundances of genera Thermovirga, Desulfobulbus, Reyranella, Fastidiosipila, and Desulfonatronum were similar in Sample 1 and Sample 2 but lower in Sample 3. Meanwhile, the uncultured, unclassified, and uncultured_norank genera accounted for a large proportion, and they may play some yet unknown or less understood role.
Discussion In the research, the BESIÒ process was used to treat petrochemical FGD wastewater. The removal of organics and nitrogen were based on sulfur bioconversion by BSR and BSO. The 454-pyrosequencing revealed the microbial community structures, and it is necessary to understand the relationship between the performance of the integrated process and the microbial community structures. The BESIÒ process consists of activated sludge and biofilm reactors. The initial loading of activated sludge accounted for half of the volume in the anaerobic and anoxic reactors. The continuous flow could lift the level of sludge, and there were sludge and bio-film washout in the reactors. The activated sludge presented good sedimentation ability, and the ascending velocity was lower than the sedimentation velocity of the activated sludge. So, there were settling sections in the three reactors, and the suspend solid of effluent was low at steady state. The sampling time of sequencing was important. There should be no washout of activated sludge and bio-film in the reactors, and performance of the integrated process should be steady. To increase the coverage of the microbial samples, we
Bioprocess Biosyst Eng Fig. 6 Relative abundances of genera in the three samples. In the heatmap, the color intensity depicts the relative abundance of each sample, from low (blue) to high (red), with the legend provided at the bottom (color figure online)
collected the activated sludge from three layers in the anaerobic and anoxic reactors, and bio-film from five different parts in the aerobic reactor, respectively.
At sulfate-rich condition, the sulfate could participate in the SRB metabolism, and the organics were oxidized that the SRB could achieve energy for metabolism. The
123
Bioprocess Biosyst Eng
metabolites of SRB are similar with the fermentation end-products, for instance, acetic acid, metacetonic acid, and butanoic acid. The SRB realized the biodegradation of organic pollutants in the anaerobic reactor with the non-sulfate-reducing chemoorganotrophic bacteria. In the anaerobic reactor, the SRB played an important role in COD removal, but there were many types of organic pollutants in the FGD wastewater that could not be degraded by the SRB. Thus, the cooperation between the SRB and non-SRB groups is important for the integrated process. The SRB realized more COD removal, while the non-sulfate-reducing groups were more diverse in the anaerobic reactor. The phyla Firmicutes, Actinobacteria, Synergistetes, and Bacteroidetes were abundant in the anaerobic reactor. These phyla have wide ecological niches in natural and industrial environments, and most of the close groups are chemoorganotrophic, so that these phyla could participate in the biodegradation of organic pollutants in the process [6]. At the genus level, Clostridium was the most dominant group in the anaerobic reactor, and this genus is a subdivision of class Clostridia, which could function in the biodegradation of the organic pollutants [6]. The genus Clostridium was also detected in the anoxic and aerobic reactors, accounting for 0.86 and 1.80%, respectively. In the aerobic reactor, Blastocatella (accounting for 20.82%) was the dominant genus. This genus is an aerobic, chemoorganotrophic bacterium that contributed to the organic removal in the integrated process [27]. The presence of these microbial phylotypes may be helpful for the degradation of organic pollutants, although the specific rate of COD removal by each of these bacteria could not be measured. The FGD wastewater contained many types of nitrogen pollutants, such as nitrogenous organic pollutants, ammonia nitrogen, and nitrate. In the BESIÒ process, the anoxic reactor was used for denitrification, the aerobic reactor was used for ammonia nitrogen oxidation, and the generated nitrate could be refluxed back to the anoxic reactor to serve as a substrate for denitrification. Through the 454-pyrosequencing, the nitrogen degradation-associated bacteria were detected. The Hyphomicrobium genus might play an important role in the biodegradation of nitrogenous organic compounds [28], and this genus accounted for 6.33% in the aerobic reactor. The genus Hyphomicrobium was also detected in both the anaerobic and anoxic reactors, accounting for 0.90 and 1.15%, respectively. The nitrogenous organic pollutants could be degraded by genus Hyphomicrobium in all the three reactors of the integrated process, and the nitrogen released as ammonia nitrogen. The ammonia nitrogen was transferred into the aerobic
123
reactor, and could be oxidized to nitrate nitrogen eventually. In the aerobic reactor, Nitrospira is the most important nitrite-oxidizing bacteria (NOB), and the genus was abundant in the aerobic reactor. Nitrospira is a kind of autotrophic microbe, and was reported to adapt to live under substrate limitation [29, 30]. Nitrospira with nitrite bacteria could convert the ammonia nitrogen to nitrate nitrogen, and the nitrate nitrogen was refluxed to the anoxic reactor. In the anoxic reactor, the denitrification-related genera included Azoarcus and Paracoccus. Azoarcus was the most dominant genus in the anoxic reactor, and this genus has been proven to have the sulfide-oxidizing ability under denitrifying conditions. Paracoccus was confirmed to have high efficiency in denitrification [28, 31]. In the study, all the TN removal was associated with the oxidation of sulfide, so Azoarcus was speculated to play important role on denitrification in the anoxic reactor. The optimum product of denitrification is nitrogen gas. A complete nitrogen degradation chain was built in the integrated process. Through the degradation chain, the nitrogenous organic pollutants, ammonia nitrogen, and nitrate could all be removed. The participant functional bacteria were also detected. The study provides a method for the treatment of petrochemical wastewater with low COD concentration. The sulfur metabolism could drive the integrated process, and the pollutants’ removal was coupled with the sulfur metabolism. The combination of different biological reactors realized the removal of organics and nitrogen.
Conclusion The performance of the integrated process, and the microbial communities were both analyzed. The laboratory-scale BESIÒ process performed stable in the 100-day operational period. The BESIÒ process was effective for the treatment of real sulfate-rich wastewater. Highthroughput 454-pyrosequencing provides sufficient sequencing for the analysis of the microbial community. The SRB were detected as functional abundant phylotypes, the nitrogen biodegradation-related bacteria were detected, including Hyphomicrobium, Azoarcus, Paracoccus, Nitrospira, etc. Acknowledgements This work was supported by the Funds for Creative Research Groups of China (No. 51121062) and State Key Laboratory of Urban Water Resource and Environment (Harbin Institute of Technology) (Nos. 2015TS07 and HC201621-02). The authors acknowledge the support from the Guangdong Provincial Science and Technology Planning Project (2016A050503041) and the Department of Education of Guangdong Province.
Bioprocess Biosyst Eng Compliance with ethical standards Conflict of interest The authors declare that they have no conflict of interest.
References 1. Lens P, Visser A, Janssen A, Pol LH, Lettinga G (1998) Biotechnological treatment of sulfate-rich wastewaters. Crit Rev Environ Sci Technol 28(1):41–88 2. Cardoso RB, Sierra-Alvarez R, Rowlette P, Flores ER, Go´mez J, Field JA (2006) Sulfide oxidation under chemolithoautotrophic denitrifying conditions. Biotechnol Bioeng 95(6):1148–1157 3. Qian J, Lu H, Cui Y, Wei L, Liu R, Chen GH (2015) Investigation on thiosulfate-involved organics and nitrogen removal by a sulfur cycle-based biological wastewater treatment process. Water Res 69:295–306 4. Venzlaff H, Enning D, Srinivasan J, Mayrhofer KJJ, Hassel AW, Widdel F, Stratmann M (2013) Accelerated cathodic reaction in microbial corrosion of iron due to direct electron uptake by sulfate-reducing bacteria. Corros Sci 66:88–96 5. Kleerebezem R, Mendez R (2002) Autotrophic denitrification for combined hydrogen sulfide removal from biogas and post-denitrification. Water Sci Technol 45(10):349–356 6. Wang J, Shi M, Lu H, Wu D, Shao MF, Zhang T, Ekama GA, van Loosdrecht MC, Chen GH (2011) Microbial community of sulfate-reducing up-flow sludge bed in the SANIÒ process for saline sewage treatment. Appl Microbiol Biotechnol 90(6):2015–2025 7. van Lier JB, van der Zee FP, Tan NC, Rebac S, Kleerebezem R (2001) Advances in high-rate anaerobic treatment: staging of reactor systems. Water Sci Technol 44(8):15–25 8. Jensen AB, Webb (1995) Treatment of H2S-containing gases: a review of microbiological alternatives. Enzyme Microb Technol 17(1):2–10 9. Kuenen JG (1975) Colorless sulfur bacteria and their role in the sulfur cycle. Plant Soil 43(1):49–76 10. Janssen AJ, Ma SC, Lens P, Lettinga G (1997) Performance of a sulfide-oxidizing expanded-bed reactor supplied with dissolved oxygen. Biotechnol Bioeng 53(1):32–40 11. Wei C, He W, Wei L, Li C, Wei D, Ma J (2016) The analysis of microbial community in the biodegradation, electron transfer based on sulfur metabolism integrated (BESIÒ) process for reverse osmosis concentrate (ROC) treatment by 454 pyrosequencing. Desalin Water Treat 57(60):29303–29315 12. van den Bosch PL, van Beusekom OC, Buisman CJ, Janssen AJ (2007) Sulfide oxidation at halo-alkaline conditions in a fed-batch bioreactor. Biotechnol Bioeng 97(5):1053–1063 13. Fajardo C, Mosquera-Corral A, Campos JL, Me´ndez R (2012) Autotrophic denitrification with sulphide in a sequencing batch reactor. J Environ Manag 113:552–556 14. Co´rdoba P (2015) Status of flue gas desulphurisation (FGD) systems from coal-fired power plants: overview of the physicchemical control processes of wet limestone FGDs. Fuel 144:274–286 15. Kaminski J (2003) Technologies and costs of SO2-emissions reduction for the energy sector. Appl Energy 75(3–4):165–172 16. Co´rdoba P, Font O, Izquierdo M, Querol X, Tobı´as A, Lo´pezAnto´n MA, Ochoa-Gonzalez R, Dı´az-Somoano M, Martı´nezTarazona MR, Ayora C, Leiva C, Ferna´ndez C, Gime´nez A (2011) Enrichment of inorganic trace pollutants in re-circulated water streams from a wet limestone flue gas desulphurisation
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
system in two coal power plants. Fuel Process Technol 92(9):1764–1775 Qian J, Jiang F, Chui HK, van Loosdrecht MC, Chen GH (2013) Industrial flue gas desulphurization waste may offer an opportunity to facilitate SANI application for significant sludge minimization in freshwater wastewater treatment. Water Sci Technol 67(12):2822–2826 Zhang L, Lin XJ, Wang JT, Jiang F, Wei L, Chen GH, Hao XD (2016) Effects of lead and mercury on sulfate-reducing bacterial activity in a biological process for flue gas desulfurization wastewater treatment. Sci Rep 6:30455 Jiang F, Zhang L, Peng GL, Liang SY, Qian J, Wei L, Chen GH (2013) A novel approach to realize SANI process in freshwater sewage treatment—use of wet flue gas desulfurization waste streams as sulfur source. Water Res 47(15):5773–5782 Omil F, Lens P, Visser A, Hulshoff Pol LW, Lettinga G (1998) Long-term competition between sulfate reducing and methanogenic bacteria in UASB reactors treating volatile fatty acids. Biotechnol Bioeng 57(6):676–685 Yamaguchi T, Bungo Y, Takahashi M, Sumino H, Nagano A, Araki N, Imai T, Yamazaki S, Harada H (2006) Low strength wastewater treatment under low temperature conditions by a novel sulfur redox action process. Water Sci Technol 53(6):99–105 Ye L, Shao MF, Zhang T, Tong AH, Lok S (2011) Analysis of the bacterial community in a laboratory-scale nitrification reactor and a wastewater treatment plant by 454-pyrosequencing. Water Res 45(15):4390–4398 Li X, Dai X, Yuan S, Li N, Liu Z, Jin J (2014) Thermal analysis and 454 pyrosequencing to evaluate the performance and mechanisms for deep stabilization and reduction of high-solid anaerobically digested sludge using biodrying process. Bioresour Technol 175C:245–253 Yang S, Wen X, Jin H, Wu Q (2012) Pyrosequencing investigation into the bacterial community in permafrost soils along the China–Russia Crude Oil Pipeline (CRCOP). PLoS One 7(12):e52730 Federation WE, Association American Public Health (2005) Standard methods for the examination of water and wastewater. American Public Health Association (APHA), Washington, DC Wong BT, Lee DJ (2011) Denitrifying sulfide removal and carbon methanogenesis in a mesophilic, methanogenic culture. Bioresour Technol 102(12):6673–6679 Yang Q, Xiong P, Ding P, Chu L, Wang J (2015) Treatment of petrochemical wastewater by microaerobic hydrolysis and anoxic/oxic processes and analysis of bacterial diversity. Bioresour Technol 196:169–175 Liao X, Chen C, Zhang J, Dai Y, Zhang X, Xie S (2015) Operational performance, biomass and microbial community structure: impacts of backwashing on drinking water biofilter. Environ Sci Pollut Res Int 22(1):546–554 Ning X, Qiao W, Zhang L, Gao X (2014) Microbial community in anoxic-oxic-settling-anaerobic sludge reduction process revealed by 454 pyrosequencing analysis. Can J Microbiol 60(12):799–809 Nowka B, Daims H, Spieck E (2015) Comparison of oxidation kinetics of nitrite-oxidizing bacteria: nitrite availability as a key factor in niche differentiation. Appl Environ Microbiol 81(2):745–753 Wang X, Zhang Y, Zhou J, Zhang T, Chen M (2015) Regeneration of elemental sulfur in a simultaneous sulfide and nitrate removal reactor under different dissolved oxygen conditions. Bioresour Technol 182:75–81
123