- Email: [email protected]
Effect of seed sludge on nitrogen removal in a novel upflow microaerobic sludge reactor for treating piggery wastewater
Bioresource Technology 216 (2016) 19–27
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Effect of seed sludge on nitrogen removal in a novel upflow microaerobic sludge reactor for treating piggery wastewater Jia Meng, Jiuling Li, Jianzheng Li ⇑, Cheng Wang, Kaiwen Deng, Kai Sun State Key Laboratory of Urban Water Resource and Environment, School of Municipal and Environmental Engineering, Harbin Institute of Technology, 73 Huanghe Road, Harbin 150090, PR China
h i g h l i g h t s Two upflow microaerobic sludge reactors were constructed to treat MFPW. Aerobic and anaerobic activated sludge (S) were the inocula, respectively. Aerobic S was easier being accumulated than anaerobic S, only feasible for low NLR. The accumulated anaerobic AS was feasible for treating MFPW with higher NLR. An increased loading rate changed microbial community structure and removal load.
a r t i c l e
i n f o
Article history: Received 7 April 2016 Received in revised form 8 May 2016 Accepted 11 May 2016 Available online 13 May 2016 Keywords: Piggery wastewater Microaerobic process Nitrogen removal Seed sludge Microbial community structure
a b s t r a c t Anaerobic activated sludge (AnaS) and aerobic activated sludge (AerS) were used to start up a novel upflow microaerobic sludge reactor (UMSR), respectively, and the nitrogen removal in the two reactors were evaluated when treating low C/N ratio manure-free piggery wastewater with a COD/TN ration of about 0.85. With the same hydraulic retention time 8 h and TN loading rate (NLR) 0.42 kg/(m3 d), the UMSR (R2) inoculated with AerS could reach its steady state earlier and obtained a better TN removal than that in the UMSR (R1) inoculated with AnaS. However, the accumulated AnaS made R1 show a better capability in bearing shock load and demonstrated an excellent NH+4-N and TN removal with a NLR as high as 1.07 kg/(m3 d). Microbial community structure of the accumulated AerS and AnaS were observable different. The decreased proportion of nitrifiers restricted the ammonium oxidation in R2, and resulting in a decrease in TN removal. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction With the large-scale and intensive development of pig industry, more and more piggery wastewater is discharged from the pig farms (Liwei et al., 2011). The wastewater is rich in nutrients and can cause severe harm to the natural environment and human health (Daverey et al., 2013; Zhao et al., 2014). There are three main approaches to remove manure from piggery in China, namely urine-free manure (UFM), combined manure with urine (CMU) and soaked manure with urine (SMU) (Li et al., 2016). As a traditional mode, UFM collection is still widely used in China (Zhao et al., 2014). The collected pig manure can be recycled as fertilizer (Zoccarato et al., 1995), but resulting in an increased ammonium (NH+4-N) and a decreased chemical oxygen demand (COD) in the manure-free piggery wastewater (MFPW) (Zhao et al., 2014). The ⇑ Corresponding author. E-mail address: [email protected] (J. Li). http://dx.doi.org/10.1016/j.biortech.2016.05.034 0960-8524/Ó 2016 Elsevier Ltd. All rights reserved.
nitrogen removal is a serious challenge in treating MFPW due to the high NH+4-N and the low C/N ratio (Kishida et al., 2003). Although traditional anaerobic–aerobic combined process is an engineered treatment technology with robust adaptability (Zheng et al., 2012), the cost of which is still high when treating low C/N ratio wastewater because some extra physical chemistry pretreatment has to be used for a feasible C/N ratio to the subsequent traditional biological nitrogen removal process (Bernet et al., 2000). Autotrophic denitrification such as anaerobic ammonium oxidation (anammox) is considered a cost-effective alternative to traditional denitrification process in treating low C/N ratio wastewater because of no need of extra carbon source (Windey et al., 2005). In the process, ammonium, S(0) and Fe(0) instead of organic carbon provide electron for reducing nitrite to N2 and/or N2O (Batchelor and Lawrence, 1978; Strous et al., 1999; Till et al., 1998). Moreover, the autotrophic nitrogen removal process has a lower sludge production compared with heterotrophic process (Van Loosdrecht and Jetten, 1998). Therefore, autotrophic
20
J. Meng et al. / Bioresource Technology 216 (2016) 19–27
denitrification is considered an essential process for nitrogen removal in treating MFPW with low C/N ratio (Sliekers et al., 2002). However, the total nitrogen (TN) in the wastewater is dominated by NH+4-N (Bernet et al., 2000). To obtain a better nitrogen removal, partial nitritation and autotrophic denitrification have to be performed in aerobic–anaerobic combined processes or sequencing batch reactors, but normally resulting in a high cost (Meng et al., 2015). Furthermore, autotrophic nitrification and denitrification would be inhibited when organic carbon source existed in wastewater because of their poor competition with heterotrophic bacteria (Ni et al., 2012). And the anaerobic condition required by autotrophic denitrifers would result in a poor COD removal (Chan et al., 2009). Though autotrophic denitrification has been extensively investigated, few reports on treating raw MFPW can be found up to now. Microaerobic condition, identified by a dissolved oxygen (DO) of 0.3–1.0 mg/L, is a transitional state between anaerobic and aerobic conditions, and allows anaerobic and aerobic bacteria to coexist in a single activated sludge process (Hu et al., 2005; Zitomer, 1998). Microaerobic biological treatment technology has been proved feasible in treating domestic wastewater with low operating cost and less excess sludge produced (Chu et al., 2006; Zheng and Cui, 2012; Zitomer, 1998). But few researches on treating livestock and poultry wastewater were reported (Chu et al., 2006; Chuang et al., 2005; Hu et al., 2005). In previous research, an upflow microaerobic sludge reactor (UMSR) has been developed to treat raw MFPW without supply of extra carbon source (Meng et al., 2015). With a C/N ratio of about 0.84 in the wastewater and operated at a TN loading rate of 1.10 kg/(m3 d), the average COD, NH+4-N and TN removal in the UMSR reached 0.72, 0.76 and 0.94 kg/(m3 d), respectively. It was found that all anaerobic fermentation bacteria, ammonia-oxidizing bacteria (AOB), hetero-
Water lock Thermostat sensor Sampling port
DO controller
Effluent Pump Feed Aerang tank
2. Materials and methods 2.1. Microaerobic treatment processes To investigate the effect of AnaS and AerS as inocula on the startup of microaerobic treatment process treating MFPW, two UMSRs were constructed in the same structure. As illustrated in Fig. 1, each of the UMSR was a 0.5-meter-high plexiglass column with a working volume of 4.9 L. A 0.5 L circular cone was attached to the bottom of the column. A 3 L solid–liquid–gas separator was designed on the top of the column and the off-gas was discharged after a water lock. Four sampling ports were at 100 mm away from the bottom of the solid–liquid–gas separator. The piggery wastewater was introduced into the reactor by a peristaltic pump. The effluent was collected by a 10 L tank. Part of the effluent was aerated to a DO of about 3.0 mg/L and then recirculated into the reactor at a reflux ratio of 45:1 to create an internal microaerobic environment with a DO less than 1.0 mg/L. DO in the reactor was detected by an on-line monitoring instrument and used to control the aeration in the aerating tank. The two reactors were operated parallel and the internal temperatures were kept at the same 35 ± 1 °C by a temperature controller. 2.2. Wastewater and inoculums
Computer
DO sensor
trophic denitrifiers and autotrophic denitrifiers could dominate the UMSR (Meng et al., 2016). This result suggested that the proportion and activity of each functional bacteria in the inocula would affect the startup time and performance of the UMSR. Though the significant impact of seed sludge on microbial community structure, startup and pollutant removal in various biological processes have been extensively investigated, few researches on microaerobic processes could be found (Cao et al., 2013). In view of the wide availability, anaerobic activated sludge (AnaS) and aerobic activated sludge (AerS) were collected as inocula to start up two UMSRs, respectively, in the present study. Effect of the inocula on the reactors’ performance within the startup processes was evaluated. To get a comprehensive insight into the biological mechanism for the effect of the seed sludge on the microaerobic treatment processes, the microbial community structure in the two reactors was also investigated using Illumina Miseq platform.
Aerang equipment
Fig. 1. Schematic representation of the lab-scale UMSR.
Wastewater fed to the two UMSRs was the same raw MFPW collected from a local pig farm in Harbin, China. The quality of the wastewater fluctuated following the breeding seasonality. The average concentration of COD, NH+4-N, NO 2 -N, NO3 -N, total nitrogen (TN) and pH was 307, 299.7, 0.1, 1.0, 366.9 mg/L and 8.0, respectively. One of the two UMSRs (marked as R1) was seeded by anaerobic activated sludge collected from an upflow anaerobic sludge bed (UASB) treating raw piggery wastewater in the State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology, China. The other UMSR (marked as R2) was seeded by aerobic activated sludge from a domestic wastewater
Table 1 Stages and operating parameters of the two UMSRs.
a b c
Stage
day
Dilution ratioa (v/v)
pH
NH+4-N (mg/L)
NO 2 -N (mg/L)
NO 3 -N (mg/L)
TN (mg/L)
COD (mg/L)
NLRb (kg/(m3 d))
OLRc (kg/(m3 d))
1 2 3
1–49 50–94 95–161
3 1 0
7.6 ± 0.2 7.9 ± 0.2 8.0 ± 0.1
107.1 ± 11.4 209.7 ± 16.7 299.7 ± 16.4
0.3 ± 0.2 0.2 ± 0.2 0.1 ± 0.1
1.5 ± 0.5 0.7 ± 0.6 1.0 ± 0.3
132.9 ± 14.0 256.8 ± 20.4 366.9 ± 19.9
116 ± 19 213 ± 26 307 ± 35
0.40 ± 0.04 0.77 ± 0.06 1.10 ± 0.06
0.35 ± 0.06 0.64 ± 0.08 0.92 ± 0.11
The ratio of dilution water to raw wastewater in terms of volume. Nitrogen loading rate in terms of TN. Organic loading rate in terms of COD.
21
J. Meng et al. / Bioresource Technology 216 (2016) 19–27
A
Influent
Effluent R1
Effluent R2
Removal efficiency R1
Removal efficiency R2
80 70
600
60 500 50 400 40 300
30
200
20
100
10
0
0 0
10
20
30
40
50
60
70
80
90
100
110
120
130
140
150
160
Time (d)
800
Influent Effluent R2 Removal efficiency R2
B
Effluent R1 Removal efficiency R1
70 500
60 50
400
40
300
30 200 20 10 0
0 0
10
20
30
40
50
60
70
80
90
100
110
120
130
140
150
160
Time (d) 800
Influent Effluent R2 Removal efficiency R2
C
Effluent R1 Removal efficiency R1
NO 3 -N,
COD, MLSS and MLVSS were detected according to the Standard Methods (APHA, 2005). DO and pH were measured with a dissolved oxygen meter (Taiwan Hengxin, AZ 8403) and pH meter (Switzerland Mettler Toledo, DELTA320), respectively. TN was measured by a total nitrogen analyzer (Germany Analytikjena Multi, N/C 2100S).
100 90 80
600 70 500
60
400
50 40
300
30 200
TN removal effifiency (%)
NO 2 -N,
NH4+-N removal effifiency (%)
80 600
100
TN concentraon (mg/L)
NH+4-N,
100 90
700
700
2.4. Analytical methods
100 90
700
COD concentraon (mg/L)
Both the UMSRs were started up at the same hydraulic retention time (HRT) of 8 h, reflux ratio of 45:1 and temperature 35 ± 1 °C. To acclimate the inoculated sludge, both the UMSRs were fed with a mixture of MFPW and molasses to adjust the COD/TN ratio in feed was 6.24 within the first 80 d and 4.52 for the next 20 d (the data were not shown). And then, in the present research, the UMSRs were fed with diluted and raw MFPW with a COD/TN ratio of about 0.85. Since the 100-day accumulation of the inoculated sludge, both the UMSRs were continued to be operated for another 161 d which was divided into three stages. The characteristics of feed and operating parameters in each of the three stages were illustrated in Table 1. The TN loading rate (NLR), COD loading rate (OLR) increased stage by stage, resulted in the decreased dilution ratio of the raw MFPW. The UMSRs would continuously be operated within a certain stage until a steady state was achieved and kept 11 days at least. The influent and effluent were sampled daily for water quality analysis. Activated sludge in the two reactors was sampled at the end of each stage for biomass (MLSS and MLVSS) and microbial community structure analysis.
stage 3
stage 2
800
NH4+-N concentraon (mg/L)
2.3. Acclimation of the inoculated sludge and start-up of the UMSRs
stage 1 900
COD removal effifiency (%)
treatment plant in Harbin, China. The initial biomass in R1 and R2 was 5.68 g/L and 5.34 g/L in terms of mixed liquor suspended solids (MLSS), or 1.95 g/L and 1.65 g/L in terms of mixed liquor volatile suspended solids (MLVSS), respectively.
20 100
10
0
0 0
10
20
30
40
50
60
70
80
90
100
110
120
130
140
150
160
Time (d)
2.5. High-throughput pyrosequencing NOx--N concentraon (mg/L)
The seed sludge and the activated sludge in each steady state of the three stages were sampled and respectively named SR1, SR1-1, SR1-2 and SR1-3 in R1, while with SR2, SR2-1, SR2-2 and SR2-3 for the sampled sludge from R2. The total DNA of the sludge samples were separately extracted using the Bacteria DNA Isolation Kit Components (MOBIO) according to the manufacturer’s instruction. Bacterial 16S rRNA genes for the V3–V4 regions were amplified with primer pairs 341f (50 -CCTACGGGAGGCAGCAG-30 ) and 805r (50 -GACTACHVGGGTATCTAATCC-30 ) (Herlemann et al., 2011; Hugerth et al., 2014). Composition of the PCR products of V4 region was determined by pyrosequencing using the Illumina Miseq sequencer platform (Shanghai Sangon, China).
60
D
NO2 in effluent R1
NO3 in effluent R1
NO2 in effluent R2
NO3 in effluent R2
50
40
30
20
10
0 0
10
20
30
40
50
60
70
80
90
100
110
120
130
140
150
160
Time (d) 9
E
Influent
Effluent R1
Effluent R2
8.5
3. Results and discussion pH
8
3.1. Pollutant removal in the two UMSRs
7.5
Fig. 2 showed the performance of the two UMSRs in COD (Fig. 2A), NH+4-N (Fig. 2B), TN (Fig. 2C), NOx-N (Fig. 2D) and pH (Fig. 2E), respectively. The results showed that R2 reached its steady state more quickly (since day 35) than R1 (since day 38) in stage 1. Here, the steady state refers to a certain period of the reactors’ performance, during which the removal of COD, NH+4-N and TN and pH in effluent are all relatively stable. Whenever there was an increase in NLR and OLR, both the UMSRs were observably
7
6.5 0
10
20
30
40
50
60
70
80
90
100
110
120
130
140
150
160
Time (d)
Fig. 2. Performance of the two UMSRs inoculated different seed sludge in COD (A), NH+4-N (B), TN (C), NOx-N (D) and pH (E).
22
J. Meng et al. / Bioresource Technology 216 (2016) 19–27
influenced in performance. But a new steady state would be obtained in the reactors with time. It was interesting to find that R1 reached its steady state earlier than R2 in stage 2, since the 80th day and the 84th day, respectively. Since the 149th day, both the UMSRs synchronously get steady again in stage 3. The average pollutant removal from the low C/N MFPW in R1 and R2 during each of the three steady phases were illustrated in Table 2. Compared with R1 that was inoculated with AneS, R2 inoculated with AerS illustrated a better pollutant removal in the steady phase (the 38–49th days) of stage 1 (Fig. 2). As shown in Table 2, with a lower NLR 0.42 kg/(m3 d) and OLR 0.30 kg/(m3 d), the COD, NH+4-N and TN removal in R2 averaged 75.0%, 95.2% and 83.4%, observably higher than that of 61.3%, 77.4% and 62.7% in R1, respectively. As a result, the effluent TN of 23.4 mg/L in R2 was lower than that of 52.4 mg/L in R1. Both of the two UMSRs synchronously found an observable accumulation of NOx-N within the steady phase of stage 1, but showed a different profile. The NO 2 -N in effluent of R1 averaged 21.4 mg/L with a NO3 -N as low as about 5.3 mg/L, while with a NO 3 -N about 10.0 mg/L and a NO 2 -N about 7.8 mg/L in the effluent of R2. Obviously, the oxidation of NH+4-N in R2 was more complete than that in R1. The oxidation of NH+4-N and the accumulation of NOx-N resulted in a slight decrease of pH in the effluent comparing with the influent.
Along with the dilution ratio decreased from 3 (in stage 1) to 1 in stage 2, the NLR and OLR in the both UMSRs increased to 0.77 and 0.63 kg/(m3 d), respectively (Table 1). And the increased loading rates showed a remarkable influence on the performance of the two reactors in pollutant removal (Fig. 2). As illustrated in Fig. 2, since the 50th day, pollutant removal efficiency in R1 become better than that in R2, which was opposite to that in stage 1. As shown as Table 2, within the steady phase (84–94th day) of stage 2, the average removal of COD, NH+4-N and TN in R1 was 74.6%, 85.5% and 83.7%, respectively, each of which was obviously higher than that in stage 1. On the contrary, the removal of COD, NH+4-N and TN in R2 within the steady phase were lower than that in stage 1, with an average of 67.0%, 75.1% and 76.5%, respectively. Since the 95th day, all through the stage 3, both the UMSRs were fed with the same raw MFPW with a NLR about 1.07 and OLR about 0.92 kg/(m3 d). And the results showed that R1 kept the better pollutant removal efficiency than R2 (Fig. 2). When R1 reached the third steady phase (149th–161st day), the removal of COD, NH+4-N and TN averaged 77.9%, 86.2% and 87.2%, respectively (Table 2). Though the COD removal in R2 reached 72.6%, its NH+4-N and TN removal was only 56.0% and 61.3%, respectively, remarkably lower than that in R1.
Table 2 Average pollutant removal from low C/N organic wastewater in R1 and R2 during steady phase of the three stages.
NH+4-N
TN
COD
NO 2 -N NO 3 -N pH
3
Stage 1 (38–49th day)
Stage 2 (84–94th day)
Stage 3 (149–161st day)
R1
R1
R1
R2
Volume loading (kg/(m d)) Load removal (kg/(m3 d)) Influent (mg/L) Effluent (mg/L) Removal (%)
0.34 ± 0.03 0.26 ± 0.02 113.7 ± 8.9 25.8 ± 4.7 77.4 ± 3.6
Volume loading (kg/(m3 d)) Load removal (kg/(m3 d)) Influent (mg/L) Effluent (mg/L) Removal (%)
0.42 ± 0.03 0.26 ± 0.03 140.7 ± 11.0 52.4 ± 5.8 62.7 ± 3.4
Volume loading ((kg/(m3d)) Load removal (kg/(m3 d)) Influent (mg/L) Effluent (mg/L) Removal (%)
0.30 ± 0.04 0.19 ± 0.03 101 ± 13 39 ± 5 61.3 ± 4.7
Influent (mg/L) Effluent (mg/L)
0.4 ± 0.1 21.4 ± 5.8
Influent (mg/L) Effluent (mg/L) Influent Effluent
1.3 ± 0.5 5.3 ± 2.6 7.7 ± 0.2 7.5 ± 0.2
Influent COD/TN
0.32 ± 0.03 5.5 ± 2.1 95.2 ± 1.9 0.35 ± 0.03 23.4 ± 5.5 83.4 ± 4.1
R2
0.63 ± 0.04 0.54 ± 0.05 211.1 ± 14.2 30.5 ± 7.0 85.5 ± 3.3 0.77 ± 0.05 0.65 ± 0.05 258.2 ± 17.2 42.0 ± 6.6 83.7 ± 2.5
25 ± 5 75.0 ± 6.2
0.63 ± 0.09 0.47 ± 0.08 211 ± 30 53 ± 10 74.6 ± 4.5
7.8 ± 4.0
0.2 ± 0.1 5.4 ± 3.1
0.23 ± 0.04
10.0 ± 2.5 7.6 ± 0.2
0.87
0.5 ± 0.4 6.2 ± 3.8 7.9 ± 0.2 8.0 ± 0.2
0.48 ± 0.04 52.4 ± 8.6 75.1 ± 4.0 0.59 ± 0.05 60.6 ± 6.8 76.5 ± 2.5
0.88 ± 0.04 0.76 ± 0.04 292.1 ± 13.2 40.2 ± 7.0 86.2 ± 2.5
69 ± 8 67.0 ± 4.7 4.9 ± 1.7
0.1 ± 0.1 2.6 ± 0.9
3.3 ± 2.4 8.2 ± 0.2
0.84
0.49 ± 0.04 128.5 ± 11.8 56.0 ± 3.3
1.07 ± 0.05 0.94 ± 0.05 357.7 ± 16.1 45.8 ± 8.4 87.2 ± 2.5 0.92 ± 0.10 0.72 ± 0.10 308 ± 35 68 ± 9 77.9 ± 3.1
0.43 ± 0.08
R2
0.66 ± 0.04 138.4 ± 11.8 61.3 ± 2.7 0.67 ± 0.09 84 ± 11 72.6 ± 3.0 7.3 ± 1.7
1.0 ± 0.3 3.0 ± 1.7 8.0 ± 0.1 8.3 ± 0.1
2.7 ± 1.4 8.5 ± 0.1
0.84
Table 3 Diversity of microbial community in inoculum and activated sludge sampled from the three steady phases of the two UMSRs. Reactor
Stage
Sample
Reads
3% distance
Raw reads
Clean reads
OUTs
Chao1 richness estimation
ACE index
Shannon diversity index
Good’s coverage (%)
R1
Inoculum 1 2 3
SR1 SR1-1 SR1-2 SR1-3
12958 25735 45750 41342
12526 21690 45740 41328
1240 2718 6254 5632
2526.06 5950.15 14231.96 12481.36
3394.42 8309.56 22614.14 18929.92
5.30 5.95 6.81 6.83
94.42 91.80 90.83 90.92
R2
Inoculum 1 2 3
SR2 SR2-1 SR2-2 SR2-3
43331 40824 36558 36735
43324 40810 36553 36616
7292 6472 5769 6390
15529.57 15935.68 13089.64 14267.46
23326.54 25326.13 19819.39 21431.29
7.25 6.86 6.95 7.07
89.60 88.92 89.79 88.79
SR1, SR1-1, SR1-2 and SR1-3 was the sample of seed sludge, and the activated sludge in each steady state of the three stages, respectively, in R1; while with SR2,SR2-1,SR2-2,SR2-3 for the sampled sludge from R2.
23
J. Meng et al. / Bioresource Technology 216 (2016) 19–27
3.2. Diversity of microbial community To comprehend the effect of the seed sludge on the performance of the UMSR, activated sludge in R1 and R2 were sampled during each of the three steady phases for DNA extraction and PCR amplification, as well as the inoculated aerobic and anaerobic activated sludge as controls respectively. Reads and diversity indices of microbial community in the inocula and the acclimated sludge of the two reactors were illustrated in Table 3. Among the indices, Chao1 and ACE index was used to indicate the species richness, and the higher index value, the higher richness. While, Shannon diversity index was used to estimate the diversity of community, and a higher value means a higher diversity (Zhou et al., 2014). As illustrated in Table 3, the indices Chao1, ACE and Shannon of the seed AerS to inoculate R2 were remarkably higher than that of the seed AnaS to inoculate R1. The more complex microbial community structure of the inoculated AerS suggested a better ecological self-balancing capacity in R2 than that in R1 inoculated with AnaS, resulting in a faster startup process and a higher pollutant removal during stage 1 (Fig. 2 and Table 2). It was found in the steady phase that the species richness in the two reactors have been remarkably enhanced, though an increase and a decrease in Shannon index was found by R1 and R2, respectively. Since the NLR and OLR were significantly improved in stage 2, both the Shannon index in the two reactor had been enhanced. However, index of Chao1 and ACE showed a difference between
the two reactors. The Chao1 and ACE in R1 reached 14231.96 and 22614.14 during the steady phase, observably higher than that of 13089.64 and 19819.39 in R2, respectively. And the higher species richness in R1 laid a foundation for the better pollutant removal than that in R2 (Table 2). When both the reactors were fed with the same raw MFPW and reached the steady phases in stage 3, Shannon diversity index in R1 and R2 was all further enhanced. However, both the Chao1 and ACE in R1 decreased to 12481.36 and 18929.92, observably lower than that of 14267.46 and 21431.29 in R2. As illustrated in Table 2, the pollutant removal in R1 was much better than that in R2. The results suggested that the pollutant removal efficiency not only could be influenced by species richness and diversity, but also may be related to the microbial community structure (Meng et al., 2016). The Good’s coverage of the eight samples indicated that the sequence libraries constructed in the present research could cover the diversity of microbial community in the sludge samples (Zhou et al., 2014).
50 45 40 35
Percentage (%)
Comparing with the steady phase in stage 1, less effluent NOx-N was found by the two reactors in the steady phases of stage 2 and stage 3 (Fig. 2D and E, Table 2). The better removal of NH+4-N and TN, and less accumulation of NOx-N indicated ammonium oxidation and denitrification were more thorough in the reactors and resulted in an increase of effluent pH (Table 2). Because both the NH+4-N and TN removal in R2 were remarkably lower than that in R1, the effluent pH of R2 was a little higher than that in R1. Above all, the UMSR could be started up whether using AnaS or AerS as inoculum. The accumulated AerS was feasible for treating the low C/N ratio MFPW only at a lower NLR of about 0.42 kg/ (m3 d). With the accumulated AnaS, the UMSR showed a better capability of bearing shock load and demonstrated an excellent NH+4-N and TN removal in treating the raw MFPW with a COD/TN ratio of about 0.84 and a NLR as high as 1.07 kg/(m3 d).
Alphaproteobacteria Betaproteobacteria
30
Gammaproteobacteria 25 Deltaproteobacteria 20
Epsilonproteobacteria
15 10 5 0 SR1
SR1-3
SR1-4
SR1-5
SR2
SR2-3
SR2-4
SR2-5
Fig. 4. Bacterial community structure at class in Proteobacteria of eight samples.
SR1
SR1-3
SR1-4
SR1-5
SR2
SR2-3
SR2-4
SR2-5
Fig. 3. Bacterial community structure at phylum of eight samples.
24
J. Meng et al. / Bioresource Technology 216 (2016) 19–27
Fig. 5. Bacterial community structure at genus level of the inoculum (A), activated sludge sampled at the end of the first stage (B), the second stage (C) and the third stage (D) in R1.
J. Meng et al. / Bioresource Technology 216 (2016) 19–27
25
Fig. 6. Bacterial community structure at genus level of the inoculum (A), activated sludge sampled at the end of the first stage (B), the second stage (C) and the third stage (D) in R2.
26
J. Meng et al. / Bioresource Technology 216 (2016) 19–27
3.3. Microbial community structure Phylogenetic classification of the eight samples (Table 3) at phylum, class and genus level was summarized in Figs. 3–6, respectively. The phylum assignment result (Fig. 3) showed that structure of the microbial community in the seed AnaS (SR1) was significantly different from that in the seed AerS (SR2). As the main phyla, the proportion of Proteobacteria, Firmicutes, Bacteroidetes and Chloroflexi was 33.33%, 19.66%, 15.34% and 13.65% in SR1, while with 40.37%, 12.82%, 9.08% and 5.06% in SR2, respectively, (Fig. 3). The microbial community of activated sludge in R1 changed significantly by stages compared with that in R2. It was found that b-Proteobacteria or c-Proteobacteria was the top class of Proteobacteria in each of the eight sludge samples (Fig. 4). Significantly, b-Proteobacteria and c-Proteobacteria have been considered the critical players for biological removal of nitrogen and phosphorus in wastewater treatment processes (Yang et al., 2014). To reveal the biological mechanism of COD and nitrogen removal with the effect of seed sludge in the UMSR, further analysis of the microbial communities at genus level was carried out. Though many sequences in the eight activated sludge samples have not been classified yet at genus level, a significant difference in the changes of microbial community structures with the increased influent loading rate could be observed between R1 (Fig. 5) and R2 (Fig. 6). The proportion of anaerobic fermentation bacteria in SR1 was 50.42% (Fig. 5A), much higher than that of 8.57% in SR2 (Fig. 5A). The anaerobic heterotrophic (10.84%) and autotrophic (0.21%) denitrifiers in SR1 was also higher than that of 7.46% and 0.08%, respectively, in SR2. However, the proportion of aerobic AOB was about 0.69% in SR1 (Fig. 5A), much lower than that of 1.28% in SR2 (Fig. 6A). Among the 1.28% AOB in SR2 were 0.54% Sphingomonas (Zheng et al., 2005), 0.43% unclassified_Nitrosomonadaceae (Prosser et al., 2014), 0.20% Nitrosococcus (Ward and Perry, 1980) and 0.11% Nitrosomonas (Igarashi et al., 1997). Furthermore, the nitrite-oxidizing bacteria (NOB), i.e. Nitrospira (Daims et al., 2000) and Nitrobacter (Boon and Laudelout, 1962) were found to be 0.22% and 0.01%, respectively, in SR2, but could not be detected in SR1. As shown in Fig. 5B –D, chemoheterotrophic bacteria, mainly responsible for COD removal, were the foremost populations in R1 throughout the operation process with a proportion of 31.19%, 17.74% and 9.40% in SR1-1 (Fig. 5B), SR1-2 (Fig. 5C) and in SR1-3 (Fig. 5D), respectively. Among the heterotrophic bacteria, the proportion of anaerobic fermentation bacterial decreased from 23.89% in SR1-1 to 10.91% in SR1-2, and then to 9.38% in SR1-3 stage by stage. Compared with the seed sludge in R1, the proportion of AOB and NOB increased slightly to 0.77% and 0.03% (Fig. 5B), respectively, in stage 1 with a lower influent nitrogen loading rate (NLR) of 0.42 kg/m3 d, resulting in a poor NH+4-N removal (Table 2). When the NLR increased to 0.77 kg/m3 d in stage 2 and then further to 1.07 kg/m3d in stage 3, the proportion of AOB increased to 1.18% in SR1-2 (Fig. 5C) and 1.07% in SR1-3 (Fig. 5D), with a proportion of NOB of 0.27% and 0.16% in SR1-2 and SR1-3, respectively (Fig. 5C and D). Enrichment of these genera resulted in the excellent NH+4-N removal in stage 2 and stage 3 (Table 2), and supplied NOx-N to the consequent denitrification in the microaerobic system. The proportion of heterotrophic and autotrophic denitrifiers in total increased from 9.57% in SR1-1 to 11.05% in SR1-2 and to 11.44% in SR1-3. The results showed that the denitrifiers had also been enriched in stage 2 and stage 3, resulting in a higher TN removal than that in stage 1 (Table 2). Chemoheterotrophic bacteria in R2 were also the foremost populations as that in R1, with a proportion of 17.27% in SR2-1 (Fig. 6B), 17.69% in SR2-2 (Fig. 6C), and 19.36% in SR2-3 (Fig. 6D), resulting in a good COD removal above 67.0% (Table 2). Because the inoculum of R2 was AerS, a nitrifier proportion of 1.63% was found in stage 1,
which was much higher than that of 0.80% in R1. Among the 1.63% nitrifiers were 1.30% AOB and 0.33% NOB. As a result, R2 showed a better NH+4-N removal than R1 in the steady phase of stage 1 (Table 2). However, the proportion of nitrifiers (AOB and NOB in total) decreased to 1.50% in SR2-2 (Fig. 6C) and then to 1.42% in SR2-3 (Fig. 6D), resulting in a significant decrease of NH+4-N removal. Though the proportion of denitrifiers was approximation in the three stages, the TN removal decreased stage by stage due to the decreased NH+4-N oxidation (Table 2). It was found that the nitrifiers in R1 increased with time and the increased NLR, while a decrease of nitrifiers was found in R2. As known, aerobic activated sludge is rich in AOB and NOB, both of which are poor in anaerobic activated sludge (Chu et al., 2006; Windey et al., 2005; Zheng et al., 2012). When the AerS was inoculated into R2, the microaerobic environment could restrict the growth of AOB and NOB due to the limited DO, and resulted in the decrease of nitrifiers in the reactor. On the other hand, AOB and NOB in the AnaS could be enriched by the microaerobic environment in R1 and showed an increase of nitrifiers with time and the increased NLR. Because the DO less than 1.0 mg/L, the nitrifiers in the seeded AerS became less and less in R2 (Fig. 6). Though the nitrifiers in R2 could balance the NLR in stage 1 and showed a better TN removal, the nitrifiers was not enough to balance the increased NLRs in stage 2 and stage 3, and illustrated a decrease in NH+4-N and TN removal stage by stage (Fig. 2B and C). On the contrary, the nitrifiers in the seeded AnaS was enriched in R1 stage by stage (Fig. 5) and could balance the increased NLRs in the last stages, resulting in the better NH+4-N and TN removal (Fig. 2B and C). Above all, the microbial community structure of the activated sludge had determined the pollutant removal efficiency of the UMSRs. The ammonium oxidation was the rate-limiting step for TN removal (Li et al., 2016). And more NH+4-N oxidized, more TN removed in the UMSRs. 4. Conclusion Both AnaS and AerS could be the inocula to start up the UMSR treating MFPW with a C/N ratio about 0.85. The accumulated AerS was feasible for a lower NLR of about 0.42 kg/(m3 d). The accumulated AnaS made the UMSR bear shock load and demonstrate an excellent NH+4-N and TN removal at a NLR as high as 1.07 kg/(m3 d). Microbial community structure of the accumulated activated sludge determined the pollutant removal efficiency of the UMSR. Ammonium oxidation was the limiting step for TN removal in the microaerobic system, and more NH+4-N oxidized, more TN removed in the UMSRs. Acknowledgements The authors gratefully acknowledge the National Natural Science Foundation of China (Grant No. 51478141), the Major Science and Technology Program for Water Pollution Control and Management (Grant No. 2013ZX07201007), the Science and Technology Department of Heilongjiang Province (Grant No. GC13C303), and the State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology (Grant No. 2016DX06) for valuable financial support. References APHA, 2005. Standard Methods for the Examination of Water and Wastewater,, American Public Health Association, Washington DC, USA. Batchelor, B., Lawrence, A.W., 1978. Autotrophic denitrification using elemental sulfur. Journal (Water Pollution Control Federation), 1986–2001. Bernet, N., Delgenes, N., Akunna, J.C., Delgenes, J., Moletta, R., 2000. Combined anaerobic–aerobic SBR for the treatment of piggery wastewater. Water Res. 34 (2), 611–619.
J. Meng et al. / Bioresource Technology 216 (2016) 19–27 Boon, B., Laudelout, H., 1962. Kinetics of nitrite oxidation by Nitrobacter winogradskyi. Biochem. J. 85 (3), 440. Cao, S., Wang, S., Peng, Y., Wu, C., Du, R., Gong, L., Ma, B., 2013. Achieving partial denitrification with sludge fermentation liquid as carbon source: the effect of seeding sludge. Bioresour. Technol. 149, 570–574. Chan, Y.J., Chong, M.F., Law, C.L., Hassell, D.G., 2009. A review on anaerobic–aerobic treatment of industrial and municipal wastewater. Chem. Eng. J. 155 (1–2), 1– 18. Chu, L., Zhang, X., Yang, F., Li, X., 2006. Treatment of domestic wastewater by using a microaerobic membrane bioreactor. Desalination 189 (1), 181–192. Chuang, S., Pai, T., Horng, R., 2005. Biotreatment of sulfate-rich wastewater in an anaerobic/micro-aerobic bioreactor system. Environ. Technol. 26 (9), 993–1002. Daims, H., Nielsen, P., Nielsen, J., Juretschko, S., Wagner, M., 2000. Novel Nitrospiralike bacteria as dominant nitrite-oxidizers in biofilms from wastewater treatment plants: diversity and in situ physiology. Water Sci. Technol. 41 (4– 5), 85–90. Daverey, A., Hung, N.-T., Dutta, K., Lin, J.-G., 2013. Ambient temperature SNAD process treating anaerobic digester liquor of swine wastewater. Bioresour. Technol. 141, 191–198. Herlemann, D.P., Labrenz, M., Jürgens, K., Bertilsson, S., Waniek, J.J., Andersson, A.F., 2011. Transitions in bacterial communities along the 2000 km salinity gradient of the Baltic Sea. ISME J. 5 (10), 1571–1579. Hu, L., Wang, J., Wen, X., Qian, Y., 2005. Study on performance characteristics of SBR under limited dissolved oxygen. Process Biochem. 40 (1), 293–296. Hugerth, L.W., Wefer, H.A., Lundin, S., Jakobsson, H.E., Lindberg, M., Rodin, S., Engstrand, L., Andersson, A.F., 2014. DegePrime, a program for degenerate primer design for broad-taxonomic-range pcr in microbial ecology studies. Appl. Environ. Microbiol. 80 (16), 5116–5123. Igarashi, N., Moriyama, H., Fujiwara, T., Fukumori, Y., Tanaka, N., 1997. The 2.8 Å structure of hydroxylamine oxidoreductase from a nitrifying chemoautotrophic bacterium, Nitrosomonas europaea. Nat. Struct. Mol. Biol. 4 (4), 276–284. Kishida, N., Kim, J.-H., Chen, M., Sasaki, H., Sudo, R., 2003. Effectiveness of oxidationreduction potential and pH as monitoring and control parameters for nitrogen removal in swine wastewater treatment by sequencing batch reactors. J. Biosci. Bioeng. 96 (3), 285–290. Li, J., Meng, J., Li, J., Wang, C., Deng, K., Sun, K., Buelna, G., 2016. The effect and biological mechanism of COD/TN ratio on nitrogen removal in a novel upflow microaerobic sludge reactor treating manure-free piggery wastewater. Bioresour. Technol. 209, 360–368. Liwei, Y., Tinglin, H., Yuexi, Z., Haiyan, W., Zhaoyue, Z., Zhaohui, P., 2011. Technology progresses of swine wastewater treatment from large-scale pig farms. In: Water Resource and Environmental Protection (ISWREP), 2011 International Symposium on IEEE, pp. 947–950. Meng, J., Li, J., Li, J., Antwi, P., Deng, K., Wang, C., Buelna, G., 2015. Nitrogen removal from low COD/TN ratio manure-free piggery wastewater within an upflow microaerobic sludge reactor. Bioresour. Technol. 198, 884–890. Meng, J., Li, J., Li, J., Sun, K., Antwi, P., Deng, K., Wang, C., Buelna, G., 2016. Efficiency and bacterial populations related to pollutant removal in an upflow
27
microaerobic sludge reactor treating manure-free piggery wastewater with low COD/TN ratio. Bioresour. Technol. 201, 166–173. Ni, S.Q., Ni, J.Y., Hu, D.L., Sung, S.W., 2012. Effect of organic matter on the performance of granular anammox process. Bioresour. Technol. 110, 701–705. Prosser, J.I., Head, I.M., Stein, L.Y., 2014. The Family Nitrosomonadaceae. In: The Prokaryotes. Springer, pp. 901–918. Sliekers, A.O., Derwort, N., Gomez, J., Strous, M., Kuenen, J., Jetten, M., 2002. Completely autotrophic nitrogen removal over nitrite in one single reactor. Water Res. 36 (10), 2475–2482. Strous, M., Fuerst, J.A., Kramer, E.H., Logemann, S., Muyzer, G., van de Pas-Schoonen, K.T., Webb, R., Kuenen, J.G., Jetten, M.S., 1999. Missing lithotroph identified as new planctomycete. Nature 400 (6743), 446–449. Till, B.A., Weathers, L.J., Alvarez, P.J., 1998. Fe (0)-supported autotrophic denitrification. Environ. Sci. Technol. 32 (5), 634–639. Van Loosdrecht, M., Jetten, M., 1998. Microbiological conversions in nitrogen removal. Water Sci. Technol. 38 (1), 1–7. Ward, B., Perry, M., 1980. Immunofluorescent assay for the marine ammoniumoxidizing bacterium Nitrosococcus oceanus. Appl. Environ. Microbiol. 39 (4), 913–918. Windey, K., De Bo, I., Verstraete, W., 2005. Oxygen-limited autotrophic nitrification– denitrification (OLAND) in a rotating biological contactor treating high-salinity wastewater. Water Res. 39 (18), 4512–4520. Yang, S., Guo, W., Chen, Y., Zhou, X., Zheng, H., Feng, X., Yin, R., Ren, N., 2014. Simultaneous nutrient removal and reduction in sludge from sewage waste using an alternating anaerobic–anoxic–microaerobic–aerobic system combining ozone/ultrasound technology. RSC Adv. 4 (95), 52892–52897. Zhao, B., Li, J., Leu, S.-Y., 2014. An innovative wood–chip-framework soil infiltrator for treating anaerobic digested swine wastewater and analysis of the microbial community. Bioresour. Technol. 173, 384–391. Zheng, D., Deng, L.W., Fan, Z.H., Liu, G.J., Chen, C., Yang, H., Liu, Y., 2012. Influence of sand layer depth on partial nitritation as pretreatment of anaerobically digested swine wastewater prior to anammox. Bioresour. Technol. 104, 274–279. Zheng, S., Cui, C., 2012. Efficient COD removal and nitrification in an upflow microaerobic sludge blanket reactor for domestic wastewater. Biotechnol. Lett. 34 (3), 471–474. Zheng, X.-S., Yang, H., Li, D.-T., 2005. Change of microbial populations in a suspended-sludge reactor performing completely autotrophic N-removal. World J. Microbiol. Biotechnol. 21 (6–7), 843–850. Zhou, Z., Qiao, W., Xing, C., Shen, X., Hu, D., Wang, L., 2014. A micro-aerobic hydrolysis process for sludge in situ reduction: performance and microbial community structure. Bioresour. Technol. 173, 452–456. Zitomer, D.H., 1998. Stoichiometry of combined aerobic and methanogenic COD transformation. Water Res. 32 (3), 669–676. Zoccarato, I., Benatti, G., Calvi, S.L., Bianchini, M.L., 1995. Use of pig manure as fertilizer with and without supplement feed in pond carp production in Northern Italy. Aquaculture 129 (1), 387–390.
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Effect of seed sludge on nitrogen removal in a novel upflow microaerobic sludge reactor for treating piggery wastewater Jia Meng, Jiuling Li, Jianzheng Li ⇑, Cheng Wang, Kaiwen Deng, Kai Sun State Key Laboratory of Urban Water Resource and Environment, School of Municipal and Environmental Engineering, Harbin Institute of Technology, 73 Huanghe Road, Harbin 150090, PR China
h i g h l i g h t s Two upflow microaerobic sludge reactors were constructed to treat MFPW. Aerobic and anaerobic activated sludge (S) were the inocula, respectively. Aerobic S was easier being accumulated than anaerobic S, only feasible for low NLR. The accumulated anaerobic AS was feasible for treating MFPW with higher NLR. An increased loading rate changed microbial community structure and removal load.
a r t i c l e
i n f o
Article history: Received 7 April 2016 Received in revised form 8 May 2016 Accepted 11 May 2016 Available online 13 May 2016 Keywords: Piggery wastewater Microaerobic process Nitrogen removal Seed sludge Microbial community structure
a b s t r a c t Anaerobic activated sludge (AnaS) and aerobic activated sludge (AerS) were used to start up a novel upflow microaerobic sludge reactor (UMSR), respectively, and the nitrogen removal in the two reactors were evaluated when treating low C/N ratio manure-free piggery wastewater with a COD/TN ration of about 0.85. With the same hydraulic retention time 8 h and TN loading rate (NLR) 0.42 kg/(m3 d), the UMSR (R2) inoculated with AerS could reach its steady state earlier and obtained a better TN removal than that in the UMSR (R1) inoculated with AnaS. However, the accumulated AnaS made R1 show a better capability in bearing shock load and demonstrated an excellent NH+4-N and TN removal with a NLR as high as 1.07 kg/(m3 d). Microbial community structure of the accumulated AerS and AnaS were observable different. The decreased proportion of nitrifiers restricted the ammonium oxidation in R2, and resulting in a decrease in TN removal. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction With the large-scale and intensive development of pig industry, more and more piggery wastewater is discharged from the pig farms (Liwei et al., 2011). The wastewater is rich in nutrients and can cause severe harm to the natural environment and human health (Daverey et al., 2013; Zhao et al., 2014). There are three main approaches to remove manure from piggery in China, namely urine-free manure (UFM), combined manure with urine (CMU) and soaked manure with urine (SMU) (Li et al., 2016). As a traditional mode, UFM collection is still widely used in China (Zhao et al., 2014). The collected pig manure can be recycled as fertilizer (Zoccarato et al., 1995), but resulting in an increased ammonium (NH+4-N) and a decreased chemical oxygen demand (COD) in the manure-free piggery wastewater (MFPW) (Zhao et al., 2014). The ⇑ Corresponding author. E-mail address: [email protected] (J. Li). http://dx.doi.org/10.1016/j.biortech.2016.05.034 0960-8524/Ó 2016 Elsevier Ltd. All rights reserved.
nitrogen removal is a serious challenge in treating MFPW due to the high NH+4-N and the low C/N ratio (Kishida et al., 2003). Although traditional anaerobic–aerobic combined process is an engineered treatment technology with robust adaptability (Zheng et al., 2012), the cost of which is still high when treating low C/N ratio wastewater because some extra physical chemistry pretreatment has to be used for a feasible C/N ratio to the subsequent traditional biological nitrogen removal process (Bernet et al., 2000). Autotrophic denitrification such as anaerobic ammonium oxidation (anammox) is considered a cost-effective alternative to traditional denitrification process in treating low C/N ratio wastewater because of no need of extra carbon source (Windey et al., 2005). In the process, ammonium, S(0) and Fe(0) instead of organic carbon provide electron for reducing nitrite to N2 and/or N2O (Batchelor and Lawrence, 1978; Strous et al., 1999; Till et al., 1998). Moreover, the autotrophic nitrogen removal process has a lower sludge production compared with heterotrophic process (Van Loosdrecht and Jetten, 1998). Therefore, autotrophic
20
J. Meng et al. / Bioresource Technology 216 (2016) 19–27
denitrification is considered an essential process for nitrogen removal in treating MFPW with low C/N ratio (Sliekers et al., 2002). However, the total nitrogen (TN) in the wastewater is dominated by NH+4-N (Bernet et al., 2000). To obtain a better nitrogen removal, partial nitritation and autotrophic denitrification have to be performed in aerobic–anaerobic combined processes or sequencing batch reactors, but normally resulting in a high cost (Meng et al., 2015). Furthermore, autotrophic nitrification and denitrification would be inhibited when organic carbon source existed in wastewater because of their poor competition with heterotrophic bacteria (Ni et al., 2012). And the anaerobic condition required by autotrophic denitrifers would result in a poor COD removal (Chan et al., 2009). Though autotrophic denitrification has been extensively investigated, few reports on treating raw MFPW can be found up to now. Microaerobic condition, identified by a dissolved oxygen (DO) of 0.3–1.0 mg/L, is a transitional state between anaerobic and aerobic conditions, and allows anaerobic and aerobic bacteria to coexist in a single activated sludge process (Hu et al., 2005; Zitomer, 1998). Microaerobic biological treatment technology has been proved feasible in treating domestic wastewater with low operating cost and less excess sludge produced (Chu et al., 2006; Zheng and Cui, 2012; Zitomer, 1998). But few researches on treating livestock and poultry wastewater were reported (Chu et al., 2006; Chuang et al., 2005; Hu et al., 2005). In previous research, an upflow microaerobic sludge reactor (UMSR) has been developed to treat raw MFPW without supply of extra carbon source (Meng et al., 2015). With a C/N ratio of about 0.84 in the wastewater and operated at a TN loading rate of 1.10 kg/(m3 d), the average COD, NH+4-N and TN removal in the UMSR reached 0.72, 0.76 and 0.94 kg/(m3 d), respectively. It was found that all anaerobic fermentation bacteria, ammonia-oxidizing bacteria (AOB), hetero-
Water lock Thermostat sensor Sampling port
DO controller
Effluent Pump Feed Aerang tank
2. Materials and methods 2.1. Microaerobic treatment processes To investigate the effect of AnaS and AerS as inocula on the startup of microaerobic treatment process treating MFPW, two UMSRs were constructed in the same structure. As illustrated in Fig. 1, each of the UMSR was a 0.5-meter-high plexiglass column with a working volume of 4.9 L. A 0.5 L circular cone was attached to the bottom of the column. A 3 L solid–liquid–gas separator was designed on the top of the column and the off-gas was discharged after a water lock. Four sampling ports were at 100 mm away from the bottom of the solid–liquid–gas separator. The piggery wastewater was introduced into the reactor by a peristaltic pump. The effluent was collected by a 10 L tank. Part of the effluent was aerated to a DO of about 3.0 mg/L and then recirculated into the reactor at a reflux ratio of 45:1 to create an internal microaerobic environment with a DO less than 1.0 mg/L. DO in the reactor was detected by an on-line monitoring instrument and used to control the aeration in the aerating tank. The two reactors were operated parallel and the internal temperatures were kept at the same 35 ± 1 °C by a temperature controller. 2.2. Wastewater and inoculums
Computer
DO sensor
trophic denitrifiers and autotrophic denitrifiers could dominate the UMSR (Meng et al., 2016). This result suggested that the proportion and activity of each functional bacteria in the inocula would affect the startup time and performance of the UMSR. Though the significant impact of seed sludge on microbial community structure, startup and pollutant removal in various biological processes have been extensively investigated, few researches on microaerobic processes could be found (Cao et al., 2013). In view of the wide availability, anaerobic activated sludge (AnaS) and aerobic activated sludge (AerS) were collected as inocula to start up two UMSRs, respectively, in the present study. Effect of the inocula on the reactors’ performance within the startup processes was evaluated. To get a comprehensive insight into the biological mechanism for the effect of the seed sludge on the microaerobic treatment processes, the microbial community structure in the two reactors was also investigated using Illumina Miseq platform.
Aerang equipment
Fig. 1. Schematic representation of the lab-scale UMSR.
Wastewater fed to the two UMSRs was the same raw MFPW collected from a local pig farm in Harbin, China. The quality of the wastewater fluctuated following the breeding seasonality. The average concentration of COD, NH+4-N, NO 2 -N, NO3 -N, total nitrogen (TN) and pH was 307, 299.7, 0.1, 1.0, 366.9 mg/L and 8.0, respectively. One of the two UMSRs (marked as R1) was seeded by anaerobic activated sludge collected from an upflow anaerobic sludge bed (UASB) treating raw piggery wastewater in the State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology, China. The other UMSR (marked as R2) was seeded by aerobic activated sludge from a domestic wastewater
Table 1 Stages and operating parameters of the two UMSRs.
a b c
Stage
day
Dilution ratioa (v/v)
pH
NH+4-N (mg/L)
NO 2 -N (mg/L)
NO 3 -N (mg/L)
TN (mg/L)
COD (mg/L)
NLRb (kg/(m3 d))
OLRc (kg/(m3 d))
1 2 3
1–49 50–94 95–161
3 1 0
7.6 ± 0.2 7.9 ± 0.2 8.0 ± 0.1
107.1 ± 11.4 209.7 ± 16.7 299.7 ± 16.4
0.3 ± 0.2 0.2 ± 0.2 0.1 ± 0.1
1.5 ± 0.5 0.7 ± 0.6 1.0 ± 0.3
132.9 ± 14.0 256.8 ± 20.4 366.9 ± 19.9
116 ± 19 213 ± 26 307 ± 35
0.40 ± 0.04 0.77 ± 0.06 1.10 ± 0.06
0.35 ± 0.06 0.64 ± 0.08 0.92 ± 0.11
The ratio of dilution water to raw wastewater in terms of volume. Nitrogen loading rate in terms of TN. Organic loading rate in terms of COD.
21
J. Meng et al. / Bioresource Technology 216 (2016) 19–27
A
Influent
Effluent R1
Effluent R2
Removal efficiency R1
Removal efficiency R2
80 70
600
60 500 50 400 40 300
30
200
20
100
10
0
0 0
10
20
30
40
50
60
70
80
90
100
110
120
130
140
150
160
Time (d)
800
Influent Effluent R2 Removal efficiency R2
B
Effluent R1 Removal efficiency R1
70 500
60 50
400
40
300
30 200 20 10 0
0 0
10
20
30
40
50
60
70
80
90
100
110
120
130
140
150
160
Time (d) 800
Influent Effluent R2 Removal efficiency R2
C
Effluent R1 Removal efficiency R1
NO 3 -N,
COD, MLSS and MLVSS were detected according to the Standard Methods (APHA, 2005). DO and pH were measured with a dissolved oxygen meter (Taiwan Hengxin, AZ 8403) and pH meter (Switzerland Mettler Toledo, DELTA320), respectively. TN was measured by a total nitrogen analyzer (Germany Analytikjena Multi, N/C 2100S).
100 90 80
600 70 500
60
400
50 40
300
30 200
TN removal effifiency (%)
NO 2 -N,
NH4+-N removal effifiency (%)
80 600
100
TN concentraon (mg/L)
NH+4-N,
100 90
700
700
2.4. Analytical methods
100 90
700
COD concentraon (mg/L)
Both the UMSRs were started up at the same hydraulic retention time (HRT) of 8 h, reflux ratio of 45:1 and temperature 35 ± 1 °C. To acclimate the inoculated sludge, both the UMSRs were fed with a mixture of MFPW and molasses to adjust the COD/TN ratio in feed was 6.24 within the first 80 d and 4.52 for the next 20 d (the data were not shown). And then, in the present research, the UMSRs were fed with diluted and raw MFPW with a COD/TN ratio of about 0.85. Since the 100-day accumulation of the inoculated sludge, both the UMSRs were continued to be operated for another 161 d which was divided into three stages. The characteristics of feed and operating parameters in each of the three stages were illustrated in Table 1. The TN loading rate (NLR), COD loading rate (OLR) increased stage by stage, resulted in the decreased dilution ratio of the raw MFPW. The UMSRs would continuously be operated within a certain stage until a steady state was achieved and kept 11 days at least. The influent and effluent were sampled daily for water quality analysis. Activated sludge in the two reactors was sampled at the end of each stage for biomass (MLSS and MLVSS) and microbial community structure analysis.
stage 3
stage 2
800
NH4+-N concentraon (mg/L)
2.3. Acclimation of the inoculated sludge and start-up of the UMSRs
stage 1 900
COD removal effifiency (%)
treatment plant in Harbin, China. The initial biomass in R1 and R2 was 5.68 g/L and 5.34 g/L in terms of mixed liquor suspended solids (MLSS), or 1.95 g/L and 1.65 g/L in terms of mixed liquor volatile suspended solids (MLVSS), respectively.
20 100
10
0
0 0
10
20
30
40
50
60
70
80
90
100
110
120
130
140
150
160
Time (d)
2.5. High-throughput pyrosequencing NOx--N concentraon (mg/L)
The seed sludge and the activated sludge in each steady state of the three stages were sampled and respectively named SR1, SR1-1, SR1-2 and SR1-3 in R1, while with SR2, SR2-1, SR2-2 and SR2-3 for the sampled sludge from R2. The total DNA of the sludge samples were separately extracted using the Bacteria DNA Isolation Kit Components (MOBIO) according to the manufacturer’s instruction. Bacterial 16S rRNA genes for the V3–V4 regions were amplified with primer pairs 341f (50 -CCTACGGGAGGCAGCAG-30 ) and 805r (50 -GACTACHVGGGTATCTAATCC-30 ) (Herlemann et al., 2011; Hugerth et al., 2014). Composition of the PCR products of V4 region was determined by pyrosequencing using the Illumina Miseq sequencer platform (Shanghai Sangon, China).
60
D
NO2 in effluent R1
NO3 in effluent R1
NO2 in effluent R2
NO3 in effluent R2
50
40
30
20
10
0 0
10
20
30
40
50
60
70
80
90
100
110
120
130
140
150
160
Time (d) 9
E
Influent
Effluent R1
Effluent R2
8.5
3. Results and discussion pH
8
3.1. Pollutant removal in the two UMSRs
7.5
Fig. 2 showed the performance of the two UMSRs in COD (Fig. 2A), NH+4-N (Fig. 2B), TN (Fig. 2C), NOx-N (Fig. 2D) and pH (Fig. 2E), respectively. The results showed that R2 reached its steady state more quickly (since day 35) than R1 (since day 38) in stage 1. Here, the steady state refers to a certain period of the reactors’ performance, during which the removal of COD, NH+4-N and TN and pH in effluent are all relatively stable. Whenever there was an increase in NLR and OLR, both the UMSRs were observably
7
6.5 0
10
20
30
40
50
60
70
80
90
100
110
120
130
140
150
160
Time (d)
Fig. 2. Performance of the two UMSRs inoculated different seed sludge in COD (A), NH+4-N (B), TN (C), NOx-N (D) and pH (E).
22
J. Meng et al. / Bioresource Technology 216 (2016) 19–27
influenced in performance. But a new steady state would be obtained in the reactors with time. It was interesting to find that R1 reached its steady state earlier than R2 in stage 2, since the 80th day and the 84th day, respectively. Since the 149th day, both the UMSRs synchronously get steady again in stage 3. The average pollutant removal from the low C/N MFPW in R1 and R2 during each of the three steady phases were illustrated in Table 2. Compared with R1 that was inoculated with AneS, R2 inoculated with AerS illustrated a better pollutant removal in the steady phase (the 38–49th days) of stage 1 (Fig. 2). As shown in Table 2, with a lower NLR 0.42 kg/(m3 d) and OLR 0.30 kg/(m3 d), the COD, NH+4-N and TN removal in R2 averaged 75.0%, 95.2% and 83.4%, observably higher than that of 61.3%, 77.4% and 62.7% in R1, respectively. As a result, the effluent TN of 23.4 mg/L in R2 was lower than that of 52.4 mg/L in R1. Both of the two UMSRs synchronously found an observable accumulation of NOx-N within the steady phase of stage 1, but showed a different profile. The NO 2 -N in effluent of R1 averaged 21.4 mg/L with a NO3 -N as low as about 5.3 mg/L, while with a NO 3 -N about 10.0 mg/L and a NO 2 -N about 7.8 mg/L in the effluent of R2. Obviously, the oxidation of NH+4-N in R2 was more complete than that in R1. The oxidation of NH+4-N and the accumulation of NOx-N resulted in a slight decrease of pH in the effluent comparing with the influent.
Along with the dilution ratio decreased from 3 (in stage 1) to 1 in stage 2, the NLR and OLR in the both UMSRs increased to 0.77 and 0.63 kg/(m3 d), respectively (Table 1). And the increased loading rates showed a remarkable influence on the performance of the two reactors in pollutant removal (Fig. 2). As illustrated in Fig. 2, since the 50th day, pollutant removal efficiency in R1 become better than that in R2, which was opposite to that in stage 1. As shown as Table 2, within the steady phase (84–94th day) of stage 2, the average removal of COD, NH+4-N and TN in R1 was 74.6%, 85.5% and 83.7%, respectively, each of which was obviously higher than that in stage 1. On the contrary, the removal of COD, NH+4-N and TN in R2 within the steady phase were lower than that in stage 1, with an average of 67.0%, 75.1% and 76.5%, respectively. Since the 95th day, all through the stage 3, both the UMSRs were fed with the same raw MFPW with a NLR about 1.07 and OLR about 0.92 kg/(m3 d). And the results showed that R1 kept the better pollutant removal efficiency than R2 (Fig. 2). When R1 reached the third steady phase (149th–161st day), the removal of COD, NH+4-N and TN averaged 77.9%, 86.2% and 87.2%, respectively (Table 2). Though the COD removal in R2 reached 72.6%, its NH+4-N and TN removal was only 56.0% and 61.3%, respectively, remarkably lower than that in R1.
Table 2 Average pollutant removal from low C/N organic wastewater in R1 and R2 during steady phase of the three stages.
NH+4-N
TN
COD
NO 2 -N NO 3 -N pH
3
Stage 1 (38–49th day)
Stage 2 (84–94th day)
Stage 3 (149–161st day)
R1
R1
R1
R2
Volume loading (kg/(m d)) Load removal (kg/(m3 d)) Influent (mg/L) Effluent (mg/L) Removal (%)
0.34 ± 0.03 0.26 ± 0.02 113.7 ± 8.9 25.8 ± 4.7 77.4 ± 3.6
Volume loading (kg/(m3 d)) Load removal (kg/(m3 d)) Influent (mg/L) Effluent (mg/L) Removal (%)
0.42 ± 0.03 0.26 ± 0.03 140.7 ± 11.0 52.4 ± 5.8 62.7 ± 3.4
Volume loading ((kg/(m3d)) Load removal (kg/(m3 d)) Influent (mg/L) Effluent (mg/L) Removal (%)
0.30 ± 0.04 0.19 ± 0.03 101 ± 13 39 ± 5 61.3 ± 4.7
Influent (mg/L) Effluent (mg/L)
0.4 ± 0.1 21.4 ± 5.8
Influent (mg/L) Effluent (mg/L) Influent Effluent
1.3 ± 0.5 5.3 ± 2.6 7.7 ± 0.2 7.5 ± 0.2
Influent COD/TN
0.32 ± 0.03 5.5 ± 2.1 95.2 ± 1.9 0.35 ± 0.03 23.4 ± 5.5 83.4 ± 4.1
R2
0.63 ± 0.04 0.54 ± 0.05 211.1 ± 14.2 30.5 ± 7.0 85.5 ± 3.3 0.77 ± 0.05 0.65 ± 0.05 258.2 ± 17.2 42.0 ± 6.6 83.7 ± 2.5
25 ± 5 75.0 ± 6.2
0.63 ± 0.09 0.47 ± 0.08 211 ± 30 53 ± 10 74.6 ± 4.5
7.8 ± 4.0
0.2 ± 0.1 5.4 ± 3.1
0.23 ± 0.04
10.0 ± 2.5 7.6 ± 0.2
0.87
0.5 ± 0.4 6.2 ± 3.8 7.9 ± 0.2 8.0 ± 0.2
0.48 ± 0.04 52.4 ± 8.6 75.1 ± 4.0 0.59 ± 0.05 60.6 ± 6.8 76.5 ± 2.5
0.88 ± 0.04 0.76 ± 0.04 292.1 ± 13.2 40.2 ± 7.0 86.2 ± 2.5
69 ± 8 67.0 ± 4.7 4.9 ± 1.7
0.1 ± 0.1 2.6 ± 0.9
3.3 ± 2.4 8.2 ± 0.2
0.84
0.49 ± 0.04 128.5 ± 11.8 56.0 ± 3.3
1.07 ± 0.05 0.94 ± 0.05 357.7 ± 16.1 45.8 ± 8.4 87.2 ± 2.5 0.92 ± 0.10 0.72 ± 0.10 308 ± 35 68 ± 9 77.9 ± 3.1
0.43 ± 0.08
R2
0.66 ± 0.04 138.4 ± 11.8 61.3 ± 2.7 0.67 ± 0.09 84 ± 11 72.6 ± 3.0 7.3 ± 1.7
1.0 ± 0.3 3.0 ± 1.7 8.0 ± 0.1 8.3 ± 0.1
2.7 ± 1.4 8.5 ± 0.1
0.84
Table 3 Diversity of microbial community in inoculum and activated sludge sampled from the three steady phases of the two UMSRs. Reactor
Stage
Sample
Reads
3% distance
Raw reads
Clean reads
OUTs
Chao1 richness estimation
ACE index
Shannon diversity index
Good’s coverage (%)
R1
Inoculum 1 2 3
SR1 SR1-1 SR1-2 SR1-3
12958 25735 45750 41342
12526 21690 45740 41328
1240 2718 6254 5632
2526.06 5950.15 14231.96 12481.36
3394.42 8309.56 22614.14 18929.92
5.30 5.95 6.81 6.83
94.42 91.80 90.83 90.92
R2
Inoculum 1 2 3
SR2 SR2-1 SR2-2 SR2-3
43331 40824 36558 36735
43324 40810 36553 36616
7292 6472 5769 6390
15529.57 15935.68 13089.64 14267.46
23326.54 25326.13 19819.39 21431.29
7.25 6.86 6.95 7.07
89.60 88.92 89.79 88.79
SR1, SR1-1, SR1-2 and SR1-3 was the sample of seed sludge, and the activated sludge in each steady state of the three stages, respectively, in R1; while with SR2,SR2-1,SR2-2,SR2-3 for the sampled sludge from R2.
23
J. Meng et al. / Bioresource Technology 216 (2016) 19–27
3.2. Diversity of microbial community To comprehend the effect of the seed sludge on the performance of the UMSR, activated sludge in R1 and R2 were sampled during each of the three steady phases for DNA extraction and PCR amplification, as well as the inoculated aerobic and anaerobic activated sludge as controls respectively. Reads and diversity indices of microbial community in the inocula and the acclimated sludge of the two reactors were illustrated in Table 3. Among the indices, Chao1 and ACE index was used to indicate the species richness, and the higher index value, the higher richness. While, Shannon diversity index was used to estimate the diversity of community, and a higher value means a higher diversity (Zhou et al., 2014). As illustrated in Table 3, the indices Chao1, ACE and Shannon of the seed AerS to inoculate R2 were remarkably higher than that of the seed AnaS to inoculate R1. The more complex microbial community structure of the inoculated AerS suggested a better ecological self-balancing capacity in R2 than that in R1 inoculated with AnaS, resulting in a faster startup process and a higher pollutant removal during stage 1 (Fig. 2 and Table 2). It was found in the steady phase that the species richness in the two reactors have been remarkably enhanced, though an increase and a decrease in Shannon index was found by R1 and R2, respectively. Since the NLR and OLR were significantly improved in stage 2, both the Shannon index in the two reactor had been enhanced. However, index of Chao1 and ACE showed a difference between
the two reactors. The Chao1 and ACE in R1 reached 14231.96 and 22614.14 during the steady phase, observably higher than that of 13089.64 and 19819.39 in R2, respectively. And the higher species richness in R1 laid a foundation for the better pollutant removal than that in R2 (Table 2). When both the reactors were fed with the same raw MFPW and reached the steady phases in stage 3, Shannon diversity index in R1 and R2 was all further enhanced. However, both the Chao1 and ACE in R1 decreased to 12481.36 and 18929.92, observably lower than that of 14267.46 and 21431.29 in R2. As illustrated in Table 2, the pollutant removal in R1 was much better than that in R2. The results suggested that the pollutant removal efficiency not only could be influenced by species richness and diversity, but also may be related to the microbial community structure (Meng et al., 2016). The Good’s coverage of the eight samples indicated that the sequence libraries constructed in the present research could cover the diversity of microbial community in the sludge samples (Zhou et al., 2014).
50 45 40 35
Percentage (%)
Comparing with the steady phase in stage 1, less effluent NOx-N was found by the two reactors in the steady phases of stage 2 and stage 3 (Fig. 2D and E, Table 2). The better removal of NH+4-N and TN, and less accumulation of NOx-N indicated ammonium oxidation and denitrification were more thorough in the reactors and resulted in an increase of effluent pH (Table 2). Because both the NH+4-N and TN removal in R2 were remarkably lower than that in R1, the effluent pH of R2 was a little higher than that in R1. Above all, the UMSR could be started up whether using AnaS or AerS as inoculum. The accumulated AerS was feasible for treating the low C/N ratio MFPW only at a lower NLR of about 0.42 kg/ (m3 d). With the accumulated AnaS, the UMSR showed a better capability of bearing shock load and demonstrated an excellent NH+4-N and TN removal in treating the raw MFPW with a COD/TN ratio of about 0.84 and a NLR as high as 1.07 kg/(m3 d).
Alphaproteobacteria Betaproteobacteria
30
Gammaproteobacteria 25 Deltaproteobacteria 20
Epsilonproteobacteria
15 10 5 0 SR1
SR1-3
SR1-4
SR1-5
SR2
SR2-3
SR2-4
SR2-5
Fig. 4. Bacterial community structure at class in Proteobacteria of eight samples.
SR1
SR1-3
SR1-4
SR1-5
SR2
SR2-3
SR2-4
SR2-5
Fig. 3. Bacterial community structure at phylum of eight samples.
24
J. Meng et al. / Bioresource Technology 216 (2016) 19–27
Fig. 5. Bacterial community structure at genus level of the inoculum (A), activated sludge sampled at the end of the first stage (B), the second stage (C) and the third stage (D) in R1.
J. Meng et al. / Bioresource Technology 216 (2016) 19–27
25
Fig. 6. Bacterial community structure at genus level of the inoculum (A), activated sludge sampled at the end of the first stage (B), the second stage (C) and the third stage (D) in R2.
26
J. Meng et al. / Bioresource Technology 216 (2016) 19–27
3.3. Microbial community structure Phylogenetic classification of the eight samples (Table 3) at phylum, class and genus level was summarized in Figs. 3–6, respectively. The phylum assignment result (Fig. 3) showed that structure of the microbial community in the seed AnaS (SR1) was significantly different from that in the seed AerS (SR2). As the main phyla, the proportion of Proteobacteria, Firmicutes, Bacteroidetes and Chloroflexi was 33.33%, 19.66%, 15.34% and 13.65% in SR1, while with 40.37%, 12.82%, 9.08% and 5.06% in SR2, respectively, (Fig. 3). The microbial community of activated sludge in R1 changed significantly by stages compared with that in R2. It was found that b-Proteobacteria or c-Proteobacteria was the top class of Proteobacteria in each of the eight sludge samples (Fig. 4). Significantly, b-Proteobacteria and c-Proteobacteria have been considered the critical players for biological removal of nitrogen and phosphorus in wastewater treatment processes (Yang et al., 2014). To reveal the biological mechanism of COD and nitrogen removal with the effect of seed sludge in the UMSR, further analysis of the microbial communities at genus level was carried out. Though many sequences in the eight activated sludge samples have not been classified yet at genus level, a significant difference in the changes of microbial community structures with the increased influent loading rate could be observed between R1 (Fig. 5) and R2 (Fig. 6). The proportion of anaerobic fermentation bacteria in SR1 was 50.42% (Fig. 5A), much higher than that of 8.57% in SR2 (Fig. 5A). The anaerobic heterotrophic (10.84%) and autotrophic (0.21%) denitrifiers in SR1 was also higher than that of 7.46% and 0.08%, respectively, in SR2. However, the proportion of aerobic AOB was about 0.69% in SR1 (Fig. 5A), much lower than that of 1.28% in SR2 (Fig. 6A). Among the 1.28% AOB in SR2 were 0.54% Sphingomonas (Zheng et al., 2005), 0.43% unclassified_Nitrosomonadaceae (Prosser et al., 2014), 0.20% Nitrosococcus (Ward and Perry, 1980) and 0.11% Nitrosomonas (Igarashi et al., 1997). Furthermore, the nitrite-oxidizing bacteria (NOB), i.e. Nitrospira (Daims et al., 2000) and Nitrobacter (Boon and Laudelout, 1962) were found to be 0.22% and 0.01%, respectively, in SR2, but could not be detected in SR1. As shown in Fig. 5B –D, chemoheterotrophic bacteria, mainly responsible for COD removal, were the foremost populations in R1 throughout the operation process with a proportion of 31.19%, 17.74% and 9.40% in SR1-1 (Fig. 5B), SR1-2 (Fig. 5C) and in SR1-3 (Fig. 5D), respectively. Among the heterotrophic bacteria, the proportion of anaerobic fermentation bacterial decreased from 23.89% in SR1-1 to 10.91% in SR1-2, and then to 9.38% in SR1-3 stage by stage. Compared with the seed sludge in R1, the proportion of AOB and NOB increased slightly to 0.77% and 0.03% (Fig. 5B), respectively, in stage 1 with a lower influent nitrogen loading rate (NLR) of 0.42 kg/m3 d, resulting in a poor NH+4-N removal (Table 2). When the NLR increased to 0.77 kg/m3 d in stage 2 and then further to 1.07 kg/m3d in stage 3, the proportion of AOB increased to 1.18% in SR1-2 (Fig. 5C) and 1.07% in SR1-3 (Fig. 5D), with a proportion of NOB of 0.27% and 0.16% in SR1-2 and SR1-3, respectively (Fig. 5C and D). Enrichment of these genera resulted in the excellent NH+4-N removal in stage 2 and stage 3 (Table 2), and supplied NOx-N to the consequent denitrification in the microaerobic system. The proportion of heterotrophic and autotrophic denitrifiers in total increased from 9.57% in SR1-1 to 11.05% in SR1-2 and to 11.44% in SR1-3. The results showed that the denitrifiers had also been enriched in stage 2 and stage 3, resulting in a higher TN removal than that in stage 1 (Table 2). Chemoheterotrophic bacteria in R2 were also the foremost populations as that in R1, with a proportion of 17.27% in SR2-1 (Fig. 6B), 17.69% in SR2-2 (Fig. 6C), and 19.36% in SR2-3 (Fig. 6D), resulting in a good COD removal above 67.0% (Table 2). Because the inoculum of R2 was AerS, a nitrifier proportion of 1.63% was found in stage 1,
which was much higher than that of 0.80% in R1. Among the 1.63% nitrifiers were 1.30% AOB and 0.33% NOB. As a result, R2 showed a better NH+4-N removal than R1 in the steady phase of stage 1 (Table 2). However, the proportion of nitrifiers (AOB and NOB in total) decreased to 1.50% in SR2-2 (Fig. 6C) and then to 1.42% in SR2-3 (Fig. 6D), resulting in a significant decrease of NH+4-N removal. Though the proportion of denitrifiers was approximation in the three stages, the TN removal decreased stage by stage due to the decreased NH+4-N oxidation (Table 2). It was found that the nitrifiers in R1 increased with time and the increased NLR, while a decrease of nitrifiers was found in R2. As known, aerobic activated sludge is rich in AOB and NOB, both of which are poor in anaerobic activated sludge (Chu et al., 2006; Windey et al., 2005; Zheng et al., 2012). When the AerS was inoculated into R2, the microaerobic environment could restrict the growth of AOB and NOB due to the limited DO, and resulted in the decrease of nitrifiers in the reactor. On the other hand, AOB and NOB in the AnaS could be enriched by the microaerobic environment in R1 and showed an increase of nitrifiers with time and the increased NLR. Because the DO less than 1.0 mg/L, the nitrifiers in the seeded AerS became less and less in R2 (Fig. 6). Though the nitrifiers in R2 could balance the NLR in stage 1 and showed a better TN removal, the nitrifiers was not enough to balance the increased NLRs in stage 2 and stage 3, and illustrated a decrease in NH+4-N and TN removal stage by stage (Fig. 2B and C). On the contrary, the nitrifiers in the seeded AnaS was enriched in R1 stage by stage (Fig. 5) and could balance the increased NLRs in the last stages, resulting in the better NH+4-N and TN removal (Fig. 2B and C). Above all, the microbial community structure of the activated sludge had determined the pollutant removal efficiency of the UMSRs. The ammonium oxidation was the rate-limiting step for TN removal (Li et al., 2016). And more NH+4-N oxidized, more TN removed in the UMSRs. 4. Conclusion Both AnaS and AerS could be the inocula to start up the UMSR treating MFPW with a C/N ratio about 0.85. The accumulated AerS was feasible for a lower NLR of about 0.42 kg/(m3 d). The accumulated AnaS made the UMSR bear shock load and demonstrate an excellent NH+4-N and TN removal at a NLR as high as 1.07 kg/(m3 d). Microbial community structure of the accumulated activated sludge determined the pollutant removal efficiency of the UMSR. Ammonium oxidation was the limiting step for TN removal in the microaerobic system, and more NH+4-N oxidized, more TN removed in the UMSRs. Acknowledgements The authors gratefully acknowledge the National Natural Science Foundation of China (Grant No. 51478141), the Major Science and Technology Program for Water Pollution Control and Management (Grant No. 2013ZX07201007), the Science and Technology Department of Heilongjiang Province (Grant No. GC13C303), and the State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology (Grant No. 2016DX06) for valuable financial support. References APHA, 2005. Standard Methods for the Examination of Water and Wastewater,, American Public Health Association, Washington DC, USA. Batchelor, B., Lawrence, A.W., 1978. Autotrophic denitrification using elemental sulfur. Journal (Water Pollution Control Federation), 1986–2001. Bernet, N., Delgenes, N., Akunna, J.C., Delgenes, J., Moletta, R., 2000. Combined anaerobic–aerobic SBR for the treatment of piggery wastewater. Water Res. 34 (2), 611–619.
J. Meng et al. / Bioresource Technology 216 (2016) 19–27 Boon, B., Laudelout, H., 1962. Kinetics of nitrite oxidation by Nitrobacter winogradskyi. Biochem. J. 85 (3), 440. Cao, S., Wang, S., Peng, Y., Wu, C., Du, R., Gong, L., Ma, B., 2013. Achieving partial denitrification with sludge fermentation liquid as carbon source: the effect of seeding sludge. Bioresour. Technol. 149, 570–574. Chan, Y.J., Chong, M.F., Law, C.L., Hassell, D.G., 2009. A review on anaerobic–aerobic treatment of industrial and municipal wastewater. Chem. Eng. J. 155 (1–2), 1– 18. Chu, L., Zhang, X., Yang, F., Li, X., 2006. Treatment of domestic wastewater by using a microaerobic membrane bioreactor. Desalination 189 (1), 181–192. Chuang, S., Pai, T., Horng, R., 2005. Biotreatment of sulfate-rich wastewater in an anaerobic/micro-aerobic bioreactor system. Environ. Technol. 26 (9), 993–1002. Daims, H., Nielsen, P., Nielsen, J., Juretschko, S., Wagner, M., 2000. Novel Nitrospiralike bacteria as dominant nitrite-oxidizers in biofilms from wastewater treatment plants: diversity and in situ physiology. Water Sci. Technol. 41 (4– 5), 85–90. Daverey, A., Hung, N.-T., Dutta, K., Lin, J.-G., 2013. Ambient temperature SNAD process treating anaerobic digester liquor of swine wastewater. Bioresour. Technol. 141, 191–198. Herlemann, D.P., Labrenz, M., Jürgens, K., Bertilsson, S., Waniek, J.J., Andersson, A.F., 2011. Transitions in bacterial communities along the 2000 km salinity gradient of the Baltic Sea. ISME J. 5 (10), 1571–1579. Hu, L., Wang, J., Wen, X., Qian, Y., 2005. Study on performance characteristics of SBR under limited dissolved oxygen. Process Biochem. 40 (1), 293–296. Hugerth, L.W., Wefer, H.A., Lundin, S., Jakobsson, H.E., Lindberg, M., Rodin, S., Engstrand, L., Andersson, A.F., 2014. DegePrime, a program for degenerate primer design for broad-taxonomic-range pcr in microbial ecology studies. Appl. Environ. Microbiol. 80 (16), 5116–5123. Igarashi, N., Moriyama, H., Fujiwara, T., Fukumori, Y., Tanaka, N., 1997. The 2.8 Å structure of hydroxylamine oxidoreductase from a nitrifying chemoautotrophic bacterium, Nitrosomonas europaea. Nat. Struct. Mol. Biol. 4 (4), 276–284. Kishida, N., Kim, J.-H., Chen, M., Sasaki, H., Sudo, R., 2003. Effectiveness of oxidationreduction potential and pH as monitoring and control parameters for nitrogen removal in swine wastewater treatment by sequencing batch reactors. J. Biosci. Bioeng. 96 (3), 285–290. Li, J., Meng, J., Li, J., Wang, C., Deng, K., Sun, K., Buelna, G., 2016. The effect and biological mechanism of COD/TN ratio on nitrogen removal in a novel upflow microaerobic sludge reactor treating manure-free piggery wastewater. Bioresour. Technol. 209, 360–368. Liwei, Y., Tinglin, H., Yuexi, Z., Haiyan, W., Zhaoyue, Z., Zhaohui, P., 2011. Technology progresses of swine wastewater treatment from large-scale pig farms. In: Water Resource and Environmental Protection (ISWREP), 2011 International Symposium on IEEE, pp. 947–950. Meng, J., Li, J., Li, J., Antwi, P., Deng, K., Wang, C., Buelna, G., 2015. Nitrogen removal from low COD/TN ratio manure-free piggery wastewater within an upflow microaerobic sludge reactor. Bioresour. Technol. 198, 884–890. Meng, J., Li, J., Li, J., Sun, K., Antwi, P., Deng, K., Wang, C., Buelna, G., 2016. Efficiency and bacterial populations related to pollutant removal in an upflow
27
microaerobic sludge reactor treating manure-free piggery wastewater with low COD/TN ratio. Bioresour. Technol. 201, 166–173. Ni, S.Q., Ni, J.Y., Hu, D.L., Sung, S.W., 2012. Effect of organic matter on the performance of granular anammox process. Bioresour. Technol. 110, 701–705. Prosser, J.I., Head, I.M., Stein, L.Y., 2014. The Family Nitrosomonadaceae. In: The Prokaryotes. Springer, pp. 901–918. Sliekers, A.O., Derwort, N., Gomez, J., Strous, M., Kuenen, J., Jetten, M., 2002. Completely autotrophic nitrogen removal over nitrite in one single reactor. Water Res. 36 (10), 2475–2482. Strous, M., Fuerst, J.A., Kramer, E.H., Logemann, S., Muyzer, G., van de Pas-Schoonen, K.T., Webb, R., Kuenen, J.G., Jetten, M.S., 1999. Missing lithotroph identified as new planctomycete. Nature 400 (6743), 446–449. Till, B.A., Weathers, L.J., Alvarez, P.J., 1998. Fe (0)-supported autotrophic denitrification. Environ. Sci. Technol. 32 (5), 634–639. Van Loosdrecht, M., Jetten, M., 1998. Microbiological conversions in nitrogen removal. Water Sci. Technol. 38 (1), 1–7. Ward, B., Perry, M., 1980. Immunofluorescent assay for the marine ammoniumoxidizing bacterium Nitrosococcus oceanus. Appl. Environ. Microbiol. 39 (4), 913–918. Windey, K., De Bo, I., Verstraete, W., 2005. Oxygen-limited autotrophic nitrification– denitrification (OLAND) in a rotating biological contactor treating high-salinity wastewater. Water Res. 39 (18), 4512–4520. Yang, S., Guo, W., Chen, Y., Zhou, X., Zheng, H., Feng, X., Yin, R., Ren, N., 2014. Simultaneous nutrient removal and reduction in sludge from sewage waste using an alternating anaerobic–anoxic–microaerobic–aerobic system combining ozone/ultrasound technology. RSC Adv. 4 (95), 52892–52897. Zhao, B., Li, J., Leu, S.-Y., 2014. An innovative wood–chip-framework soil infiltrator for treating anaerobic digested swine wastewater and analysis of the microbial community. Bioresour. Technol. 173, 384–391. Zheng, D., Deng, L.W., Fan, Z.H., Liu, G.J., Chen, C., Yang, H., Liu, Y., 2012. Influence of sand layer depth on partial nitritation as pretreatment of anaerobically digested swine wastewater prior to anammox. Bioresour. Technol. 104, 274–279. Zheng, S., Cui, C., 2012. Efficient COD removal and nitrification in an upflow microaerobic sludge blanket reactor for domestic wastewater. Biotechnol. Lett. 34 (3), 471–474. Zheng, X.-S., Yang, H., Li, D.-T., 2005. Change of microbial populations in a suspended-sludge reactor performing completely autotrophic N-removal. World J. Microbiol. Biotechnol. 21 (6–7), 843–850. Zhou, Z., Qiao, W., Xing, C., Shen, X., Hu, D., Wang, L., 2014. A micro-aerobic hydrolysis process for sludge in situ reduction: performance and microbial community structure. Bioresour. Technol. 173, 452–456. Zitomer, D.H., 1998. Stoichiometry of combined aerobic and methanogenic COD transformation. Water Res. 32 (3), 669–676. Zoccarato, I., Benatti, G., Calvi, S.L., Bianchini, M.L., 1995. Use of pig manure as fertilizer with and without supplement feed in pond carp production in Northern Italy. Aquaculture 129 (1), 387–390.