- Email: [email protected]
Membrane fouling behavior in anaerobic baffled membrane bioreactor under static operating condition
Bioresource Technology 214 (2016) 582–588
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Membrane fouling behavior in anaerobic baffled membrane bioreactor under static operating condition Jiadong Liu ⇑, Xiaolan Jia, Bo Gao, Longli Bo, Lei Wang School of Environmental and Municipal Engineering, Xi’an University of Architecture and Technology, Yan Ta Road, No. 13, Xi’an 710055, China
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
The membrane filtration was
combined with anaerobic baffled reactor. The ABMBR without turbulence intensifying strategy was developed. Membrane fouling developed slowly under static operating condition. The polysaccharide accounted for 79.12% of total area of filter cake. Submicron particles in supernatant were the source of fouling in ABMBR.
a r t i c l e
i n f o
Article history: Received 22 March 2016 Received in revised form 3 May 2016 Accepted 5 May 2016 Available online 7 May 2016 Keywords: Anaerobic membrane bioreactor Anaerobic baffled reactor Membrane filtration Static operation
a b s t r a c t A novel AnMBR combined with ABR as the anaerobic baffled membrane bioreactor (ABMBR) was developed for membrane fouling mitigation without any turbulence intensifying strategy to reduce the energy consumption further. The filtration time of this system lasted 14–25 days under stable condition only with back-flushing every 48 h. The polysaccharide accounted for 6.85 ± 3.1% amount of total filter cake and the protein accounted for 4.12 ± 2.1%, which took 79.12% and 11.12% of total area in laser scanning confocal microscope (CLSM) image. After filtration, 83.72 ± 10.97% of turbidity, 59.28 ± 16.46% of polysaccharide, 16.51% of tryptophan and 37.61% of humic-like substrates were rejected, respectively. The total membrane resistance at the end of each cycle was (4.47 ± 0.99) 1013 m 1. And the resistance from filter cake was (4.15 ± 1.00) 1013 m 1, which accounted for of 92.6 ± 3.4% of total membrane resistance. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction The research and development of anaerobic membrane bioreactor (AnMBR) have been carried out comprehensively in recent years, because of its advantage in energy conservation, anaerobic microorganism retention and higher quality of anaerobic effluent (Lin et al., 2013; Liao et al., 2006). Although the effluent from AnMBR still contains ammonia nitrogen, phosphate from influent and high concentration of soluble organic matters (Gouveia et al., 2015; Skouteris et al., 2012), its potential application in agricul⇑ Corresponding author. E-mail address: [email protected] (J. Liu). http://dx.doi.org/10.1016/j.biortech.2016.05.016 0960-8524/Ó 2016 Elsevier Ltd. All rights reserved.
tural irrigation and hydroponic system still makes this kind of wastewater treatment system popular in research. The high concentration of foulants and low intensity of turbulence without air aeration in anaerobic system can cause more serious membrane fouling phenomenon during filtration. Similar to membrane bioreactor (MBR), the configurations of AnMBR includes external cross-flow, internal submerged, or external submerged (Smith et al., 2012). For cross-flow system, the membrane fouling can be mitigated by the high speed of sludge flow (Skouteris et al., 2012), and additional turbulence is always applied in submerged AnMBR for fouling elimination (Ozgun et al., 2013). As there is artificially intensified turbulence, additional energy is necessary, which will weaken the advantage of AnMBR in energy
J. Liu et al. / Bioresource Technology 214 (2016) 582–588
consumption. And so far, many of the relative reports mainly focus on the novel membrane fouling mitigation technology in AnMBR for efficiency enhancement and energy consumption reduction during the operation (Skouteris et al., 2012; Ozgun et al., 2013). For cross-flow AnMBR, the acceleration of sludge flow can intensify the fouling mitigation, but there is a balance between the speed and efficiency (Skouteris et al., 2012). For submerged AnMBR, biogas aeration is the most popular approach for fouling elimination (Ozgun et al., 2013), but the special equipments and extra energy consumption for biogas aeration are essential, besides the security must be concerned if it is applied in practice. In AnMBR, filter cake on membrane surface is considered as the main factor for permeate flux decrease (Skouteris et al., 2012), so the turbulence enhancement for filter cake removal is the most direct and effective approach. Some novel ways for membrane fouling mitigation have been proposed, which include ultrasound (Xu et al., 2013), transverse vibration (Kola et al., 2014), rotary disk stirring (Kim et al., 2014) and fluidized media (Aslam et al., 2014) or granule activated carbon (Gao et al., 2014) addition, etc. Those kinds of turbulence enhancement approaches still work based on energy consumption. And the addition of fluidized media or granule activated carbon also needs the assistance of biogas aeration. So, there are few AnMBR without additional energy consumption for fouling mitigation had been reported previously. As the filter cake takes most of the fouling quotient, the turbulence is necessary for cake removal during operation. If no turbulence is applied, the cake formation could be reduced directly by reduction of the circumstance concentration of foulants (Metcalfe et al., 2016), which can also attain the goal of membrane fouling mitigation. In this case, the structure of the reactor should be improved. The anaerobic baffled reactor (ABR) is one of effective anaerobic wastewater treatment system (Hu et al., 2009). The ABR is composed by series of up-flow anaerobic sludge blanket (UASB), and the influent traverses the whole system in the form of reciprocation. The pollutants can be removed by the microorganisms when the wastewater flows through the sedimental sludge blanket (Lay et al., 2016). The footprint of wastewater is lengthened and the hydraulic retention time (HRT) can be separated from sludge retention time (SRT), so that the wastewater treatment efficiency is high and the operation is steerable (Hahn and Figueroa, 2015). The most important is that there is no turbulence during the treatment process, and there is no suspended sludge in supernatant at the top of each cell. If the ABR is combined with AnMBR, the membrane module assembled in the supernatant, which can be more efficient in mitigation of membrane fouling comparing with traditional submerged AnMBR, as the foulants concentration in supernatant is much lower than that in bulk sludge. And the total suspended solids in supernatant is reduced constantly when the treatment cell is added in ABR (Hahn and Figueroa, 2015), which means the concentration of membrane foulants in ABR would be reduced at the later cell. So the membrane fouling should be controlled effectively if the cell number of ABR and operation condition is appropriate. In this study, a novel AnMBR combined with ABR as the anaerobic baffled membrane bioreactor (ABMBR) was proposed. The raw domestic wastewater was used as influent. The membrane module was set in the supernatant of ABR’s later cell. The long-term operation was carried out to identify the membrane fouling behavior. The composition of membrane resistance was analyzed during the physical and chemical cleaning process, the structure of filter cake and fouled membrane were characterized, and the features of supernatant were revealed to identify the membrane fouling behavior in ABMBR. The parameters of water quality of supernatant in each cell were analyzed to identify the relationship between the pollutants transformation and membrane fouling behavior in this novel ABMBR.
583
2. Materials and methods A 49 L polymethyl methacrylate ABR equipped with 5 cells was used in this study. The active volume of one cell was 8 L (for first four cells) and it was 11.5 L for the last cell for membrane module installation, respectively. The membrane used in this experiment was the self-produced hollow fiber polyvinylidene fluoride (PVDF) membrane with polyethylene terephthalate (PET) internal support, which with outside diameter of 2.5 mm and 31 pieces of membrane fiber were sealed in polyethylene tube in the form of ‘‘U” for one membrane module. Two membrane modules with total active area of 0.28 m2 were used in parallel for dead-end filtration. In order to identify filtration and wastewater treatment efficiency of our ABMBR under different operational conditions, the membrane modules were installed in the 5th cell for the first two cycles, and then moved into the 3rd cell during the rest tests. The cover plate was sealed on the top of reactor with silicone gasket and bolts for anaerobic condition. The tube sealed on the cover plate was used for effluent and biogas spilling, and water sealing system was used for biogas collection. The whole system was operated for 130 days. The raw wastewater for influent was drawn from the sewer at the campus of Xi’an University of Architecture and Technology after 3 pm each time, the water quality parameters of influent, supernatant and effluent during each operating condition were shown in Figs. S1–S3, which included chemical oxygen demand (COD), suspended solids (SS), ammonia nitrogen (NH3-N), nitrite and nitrate nitrogen (NO2 -N, NO3 -N), phosphate (PO34 ) and were measured every one week according to the standard methods (APHA, 2005). And the wastewater was filtrated via sifter with 0.9 mm of pore size for big size of particles or suspended solids removal before using. The anaerobic sludge was taken from the bottom of UASB system of Hans brewery (Xi’an, China), which was filled in the ABMBR and the acclimatization process lasted for 30 days with continuous influent of domestic wastewater and without membrane filtration. For five cells of operation, a 5 L of anaerobic activated sludge with mixed liquor suspended solids (MLSS) of 48.05 ± 0.49 g/L was injected in the first four cells (sludge thickness of 35–38 cm), respectively. The initial MLSS was 30.03 ± 0.31 g/L (including the volume of supernatant) and with final MLSS of 32.71 ± 6.63 g/L in the 1st cell and 2nd cell together at the end of operation (including the volume of supernatant). For three cells of operation, most of the sludge was discharged from the 3rd cell (residual sludge thickness of 5–8 cm). For sludge containing cell, the level of sediment sludge was lower than the sampling hole, which was 45 cm from the bottom of reactor. For the cell with membrane module, the sludge level kept lower to make sure that the module stayed away from the sediments. And no regular sludge discharging was carried out during the whole operation. The peristaltic pumps (Langer, China) were used for influent and effluent. The speed of influent was set at 1.11 L/h (7 rpm) with constant working. The speed for effluent was set at 4 L/m2/h (8 rpm) with working mode of 5 min ‘‘on” and 1 min ‘‘off”, which was controlled by the time switcher. The HRT for five cells of operation was 39.19 h and it was 21.62 h for three cells of operation. The transmembrane pressure (TMP) was shown via vacuum meter and the membrane resistance was calculated by Darcy’s law (Liu et al., 2012). The whole study included 9 operating cycles. The back-flushing was carried out by reversed influent pump every 48 h with 1 L of tap water from the 1st cycle to the 6th cycle, which lasted about 1 h and then followed the mode of 5 min ‘‘on” and 1 min ‘‘off”. For the 7th, 8th and 9th cycles, the back-flushing was carried out every 24 h with 500 mL tap water for 30 min, because the
J. Liu et al. / Bioresource Technology 214 (2016) 582–588
membrane fouling was more serious during wintertime. From the 1st cycle to the 6th cycle, the filtration was suspended when the TMP was higher than that of 0.03 MPa; for the 7th, 8th and 9th cycles, the filtration was suspended when 50% of the initial flux was lost. The physical cleaning was carried out after the first, third, forth, sixth and eighth cycles, the chemical cleaning was carried out after the second, fifth and seventh cycles. The physical cleaning was carried out in the form of high pressure tap water flush. The chemical cleaning procedure included: physical cleaning for filter cake removal, immersing in HCl solution with pH = 2 for 4 h, tap water flush, immersing in 0.5 g/L NaOH and 5 mL/L NaClO mixed solution for 4 h and tap water flush. The flux and TMP data of pure water after the physical cleaning, acid and alkali cleaning were recorded for membrane resistance ingredient analysis. The pH, turbidity and UV254 were measured by pH meter (PHS-3C, INESA, China), turbidity meter (GDS-3B, KEDA, China) and ultraviolet–visible spectrophotometer (UV 2102 C, UNIC), respectively. All of the water samples were filtrated via qualitative filter paper before testing. The extracellular polymeric substances (EPS, polysaccharide and protein) in supernatant and filter cake of different cells were extracted according to the following steps: 10 mL of supernatant or filter cake solution was treated by ultrasound for 10 min and then heated in boiling water bath for 30 min, then the natural cooling solution was filtrated by qualitative filter paper after shaking adequately and the permeate was collected for tests. The MLSS and mixed liquor volatile suspended solid (MLVSS) were measured according to the standard methods (APHA, 2005). The concentration of polysaccharide and protein were tested according to previous report (FrØlund et al., 1995). The glucose and bovine serum albumin (A1933-100G, Sigma) were used for standard curve definition. All of the reagents used in this study were analytical pure. The particle size and Zeta potential of suspended solid in unfiltered supernatant were measured with Zeta potential analyzer (Zetasizer Nano ZS, Malvern, England) at the 100th day of operation. The protein and humic-like substrates in unfiltered (raw sample) supernatant were detected qualitatively with fluorescence excitation–emission matrix (EEM) spectroscopy (FP-6500, Jasco, Japan). The filter cake on membrane surface was measured with confocal laser scanning microscope (CLSM, Leica SP8, Germany) at the end of operation following the previous reported method (Kim et al., 2013), and the images were analyzed by the software of Photoshop CS2 for the area calculation of different color zones.
3. Results and discussion 3.1. Membrane fouling behavior in ABMBR For 5 cells of operation, two filtration cycles were carried out, which lasted 14 and 16 days (Fig. 1), respectively. One time of physical cleaning was taken between those two cycles. Seven filtration cycles was carried out for 3 cells of operation, which lasted 25, 22, 14, 18, 10, 6 and 3 days, respectively (Fig. 1). Two times of chemical cleaning were conducted between the 5th and 6th cycles, 7th and 8th cycles, respectively. And only physical cleaning was carried out for the rest cycles. Comparing with 5 cells of operation, there was no deterioration of membrane fouling appeared under 3 cells of operational condition, as the filtration time of the 3rd, 4th and 6th cycles were longer than that of the 1st and 2nd cycles. With shorter HRT, the system was still in smooth operation under static operating condition. In terms of the overall situation, the first 6 cycles were stable and the rising trend of membrane resistance was very gentle. But filtration time was reduced and membrane resistance was
15
Mmebrane resistance (1013 m-1)
584
5 cells
# Chemical cleaning
12
9
*
6
3
0
0
#
*
* Physical cleaning
3 cells
* *
#
#
*
10 20 30 40 50 60 70 80 90 100 110 120 130
Time (day) Fig. 1. Evolution of membrane resistance in ABMBR during the entire period of operation.
increased sharply during the last 3 cycles (Fig. 1) although the back-flushing interval was reduced from 48 to 24 h. The composition of membrane resistance was identified during the physical and chemical cleaning processes. Except for the last three cycles, the total membrane resistance at the end of each cycle was 3.04 1013–5.59 1013 m 1 (average value of (4.47 ± 0.99) 1013 m 1, Table 1). And the resistance from filter cake was 2.83 1013–5.34 1013 m 1 (average value of (4.15 ± 1.00) 1013 m 1), which accounted for 86.4–97.1% (average value of 92.6 ± 3.4%) of total membrane resistance. For the 7th and 8th cycles, the resistance of filter cake accounted for 97.4–98.8% of the total membrane resistance (Table 1). The MLSS and MLVSS of filter cake were measured based on membrane area (Table 1). From the 3rd to 6th cycles, the MLSS was 6.76 ± 0.22 11.05 ± 0.72 g/m2 and the MLVSS was 3.83 10.03 ± 0.03 g/m2. The similar MLSS and MLVSS amount was shown on membrane surface for the 7th and 8th cycles, although the filtration only lasted 6 and 10 days (Table 1). The membrane fouling developing rate was 0.0012– 0.0021 MPa/d under stable condition. Comparing with those fouling mitigation methods with energy consumption (Xu et al., 2013; Gao et al., 2014; Kola et al., 2014; Kim et al., 2014), the membrane fouling behavior in this study was sharper as there was no obvious retarded stage (Fig. 1). That might caused by lacking of continuous membrane cleaning strategy, because those constant in suit cleaning ways will make a balance between foulants attachment and detachment for a certain time (Meng et al., 2009).
3.2. Membrane fouling mechanism in ABMBR Previous reports indicated that the filter cake took most of membrane resistance in traditional AnMBR (Skouteris et al., 2012). In this study, the membrane with 0.1 lm pore size was used, and the average particle size of suspended solid in supernatant of cell 3 was 0.35 ± 0.02 lm (Table 2), so that most of those particles would be rejected during filtration to form filter cake. The particle size of suspended solid in supernatant of cell 1 and cell 2 was 0.36 ± 0.02 and 0.37 ± 0.03 lm (Table 2), respectively, which meant that most of the suspended solid in supernatant of this ABMBR were submicron particles. The Zeta potential of 24.37 ± 0.72 25.20 ± 1.83 mV (Table 2) also indicated that those suspended solid particles were similar to the active sludge or microbial flocs (Deng et al., 2015). Actually, the submicron
585
J. Liu et al. / Bioresource Technology 214 (2016) 582–588 Table 1 The compositions of membrane resistance and filter cake.
TR* (1013 m 1) FR* (1013 m 1) FR/TR* IOR* (1013 m 1) IOR/TR* OR* (1013 m 1) OR/TR* MLSS (g/m2) MLVSS (g/m2) Polysaccharide (mg/ m2) Protein (mg/m2) *
Physical cleaning
Chemical cleaning
Physical cleaning
Physical cleaning
Chemical cleaning
Physical cleaning
Chemical cleaning
Physical cleaning
5.59 5.22 93.36% – – – – – – –
4.13 3.57 86.46% 0.12 2.92% 0.17 4.05% – – –
3.04 2.83 93.12% – – – – 7.42 ± 0.98 3.83 618.64
5.50 5.34 97.10% – – – – 10.84 ± 0.25 8.09 ± 0.35 469.14
4.68 4.36 93.05% 0.03 0.64% 0.05 1.15% 6.76 ± 0.22 6.15 ± 0.44 816.98
3.87 3.58 92.55% – – – – 11.05 ± 0.72 10.03 ± 0.03 565.42
6.07 5.92 97.4% 0.03 0.49% – – 7.71 ± 0.28 6.44 ± 0.26 566.00
13.43 13.28 98.84% – – – – 16.26 ± 0.13 13.98 ± 0.14 632.27
–
–
589.82
67.28
515.42
287.84
188.90
561.42
TR: total membrane resistance; FR: resistance from filter cake; IOR: resistance from inorganic matters; OR: resistance from organic matters.
Table 2 The properties of solid particles in influent and supernatant.
Zeta potential (mV)* Particle size (lm)* *
Influent
Cell 1
Cell 2
Cell 3
24.57 ± 1.29 0.66 ± 0.06
24.47 ± 1.06 0.36 ± 0.02
24.37 ± 0.72 0.37 ± 0.03
25.20 ± 1.83 0.35 ± 0.02
Average values (n = 3).
particles determined the membrane flux in traditional AnMBR and MBR (Ozgun et al., 2013; Ivanovic et al., 2008), because that pore plugging/clogging could result in formation of compact filter cake. There was no direct contact between membrane module and bulk sludge in this ABMBR, consequently the submicron suspended solid would be the main component of filter cake and mainly contributed to the membrane fouling. In this ABMBR, the inorganic matters just accounted for 0.49–2.92% of membrane resistance and the rest 1.15–4.05% came from the residual organic matters (Table 1). The inorganic matters, such as MgNH4PO46H2O, K2NH4PO4 and/or CaCO3, etc, were easy to be formed on membrane surface (Ozgun et al., 2013), because there was no obvious removal efficiency for those matters in anaerobic treatment system. And the concentrations of Ca2+, Mg2+, NH4+ and PO34 in domestic wastewater would cause serious membrane fouling in AnMBR (Ozgun et al., 2013). But it was different in this ABMBR, the filter cake and residual organic matters took more than 90% of the membrane resistance. So, it is necessary to indentify the composition of filter cake and suspended solid. The EPS or soluble microbial products (SMP) were considered as the main organic foulant, which included proteins, carbohydrates, nucleic acids, lipids (Liu et al., 2015) and humic acids (Meng et al., 2009). A series of researches had focused on the EPS on membrane fouling in AnMBR (Meng et al., 2009; Ding et al., 2015). The EPS from bulk sludge and filter cake showed different fouling properties and the hydrophobic neutral (HPO-N) in EPS worked as the main foulant on membrane surface (Ding et al., 2015). In this ABMBR, the polysaccharide and protein in filter cake was 469.14–816.98 mg/m2 (611.41 ± 115.9 mg/m2) and 67.28–589.82 mg/m2 (368.45 ± 217.84 mg/m2) (Table 1), respectively. The polysaccharide accounted for 6.85 ± 3.1% of total filter cake and the protein accounted for 4.12 ± 2.1%. For observing the distribution of the protein, polysaccharide and bacteria in filter cake, the cake sample was examined microscopically with CLSM after staining. Those images revealed that the distribution of protein was wider and uniform and the polysaccharide was more concentrated (Fig. S4). The polysaccharide, protein and total cell took 79.12%, 11.12% and 0.12% of total area in combined image, respectively (Fig. S4). Those results corresponded
to filter cake composition that the content of polysaccharide was higher than that of protein (Table 1). There was no unanimous verdict about the efficiency of polysaccharide and protein on membrane fouling, some reports indicated that the contribution from protein was more obvious than that from polysaccharide, but some contradictory versions also had been presented (Meng et al., 2009), because the composition of EPS is variable and complicated under different operational conditions. More polysaccharide was found in filter cake of AnMBR for pulping wastewater treatment through CLSM analysis (Lin et al., 2009). But it was conflict with CLSM results from Gao et al. (2011) that the protein in filter accounted for 37.72 ± 5.77 87.33 ± 1.42% of EPS and the polysaccharide only took 1.56 ± 0.25 7.51 ± 3.25% in AnMBR for thermomechanical pulping whitewater treatment. When the domestic wastewater was used as influent of AnMBR, the protein was the dominant component of EPS in the filter cake, which could be adsorbed by granular activated carbon for fouling mitigation (Gao et al., 2014). Those conflicting conclusions revealed the complexity of EPS/SMP in anaerobic reactor and filter cake. In this study, the amount and area distribution of polysaccharide was more than that of protein, which would account for more membrane resistance ratio in filter cake. For simulating the actual condition, no temperature control system was applied to maintain the temperature of the reactor. The air temperature was reduced constantly after summer and a precipitate dropping of temperature was occurred at the 116th day (Fig. S5), which just corresponded to the beginning of sudden deterioration of membrane fouling (Fig. 1). The increase of turbidity, UV254 and polysaccharide in influent and supernatant indicated that the deterioration of influent quality and treatment efficiency was the reason for sudden deterioration of membrane fouling (Figs. 2 and S6). The EEM height of tryptophan-like and humiclike substances in all cells and effluent also were increased during the last 3 cycles (Fig. S7 and Table S1). That deterioration of influent and supernatant caused fast-forming of filter cake directly. The ambient temperature would affect the metabolic activity and microorganism species of anaerobic system obviously (Bialek et al., 2014; McKeown et al., 2009). Comparing with 20–35 °C, the secretion rate of extracellular organic matter (EOM or EPS)
175 93 372 77 155 257 589 210/400 215/462 220/336 220/472 265/292 275/304 275/342 A A A T T 1471 2460 225/338 275/340
A – humic-like substances; B – tyrosine-like substances; T – tryptophan-like substances.
Cell 2
Substances EX/EM (nm)
Height Cell 1
EX/EM (nm)
Substances Height Influent
Table 3 The peak/height of EEM and relative substances at the 100th day of operation.
In the ABMBR, the foulants on membrane surface came from the supernatant directly, so the relative parameters including turbidity, UV254, humic-like substances, protein and polysaccharide were measured to discover the relationship between the properties of supernatant and membrane fouling. In this study, raw wastewater from the sewer at the campus of University was used as influent. The COD in influent was 225–325 mg/L. After treatment, about 53% COD could be removed under different operating condition and the membrane rejection accounted 10–30% of COD removal (Figs. S1–S3). Actually, the contents of polysaccharide and humic-like substrates in influent were high (Figs. 2, S6, Tables 3 and S1), which were not easy to be degraded (Kim et al., 2013; Miura and Okabe, 2008) but part of them could be rejected by the membrane. And as an anaerobic system, there was no obvious efficiency on NH3-N and PO34 removal (Figs. S1–S3). After treatment process of 3 cells of ABR and membrane filtration, the turbidity and UV254 of the supernatant were decreased (Fig. 2). The turbidity represented the content of all suspended solids and UV254 reflected the aggregation of aromatic humic and fulvic structures (Zhang et al., 2007). The reduction of those parameters in supernatant indicated that the static operation of ABMBR could inhibit the foulants generation after the sequential treatment, which was good for anaerobic membrane filtration without turbulence intensifying strategy. After the filtration, the turbidity was reduced 83.72 ± 10.97% because of membrane rejection. And it was only 28.21 ± 8.29% for UV254 removal because some of aromatic humic and fulvic structures were soluble, which could pass through the membrane during filtration. The polysaccharide could be detected in influent and supernatant but without obvious removal efficiency after the anaerobic sequential treatment (Fig. 2). The polysaccharide was secreted by microorganism constantly during the metabolic process (Meng et al., 2009; He et al., 2015) and part of them could be decomposed by respiratory action of some special bacteria (Kim et al.,
Height
3.3. The relationship between the properties of supernatant and membrane fouling behavior
*
Substances
Cell 3
was accelerated under lower temperature condition of 15 °C (Zhao et al., 2015) and caused the higher SMP concentrations and turbidity in supernatant (Ozgun et al., 2015), which was similar to this study in ABMBR. The influent was draw from the sewer that came from the septic tank, which also under anaerobic condition, so the wastewater quality was deteriorated at the same time.
242 115 196 424 727
Height
Fig. 2. The transformation of pH, turbidity, polysaccharide and UV254 in influent, subsequent 3 cells and effluent from the 2nd to the 5th operational cycle.
205/412 210/450 215/566 220/336 280/338
Effluent
A T* T* T*
Cell 3
182 424 861 861
Cell 2
210/388 220/338 275/340 280/338
Cell 1
A A* T* T*
0.0
Influent
Substances
0
210/402 210/578 220/336 275/338
10
185 164 475 987
Substances Effluent
0.2
T T*
20
Height
0.4
30
EX/EM (nm)
40
*
0.6
*
50
EX/EM (nm)
60
*
0.8 pH Turbidity Polysaccharide UV254
EX/EM (nm)
70
A A T A B B T
J. Liu et al. / Bioresource Technology 214 (2016) 582–588
UV254
pH, Turbidity (NTU), Polysaccharide (mg/L)
586
587
J. Liu et al. / Bioresource Technology 214 (2016) 582–588
500
Influent
Cell 1
Cell2
EX (nm)
450 400 350 300 250 200 500
250 300 350 400 450 500 550 600
EM (nm)
Effluent
Cell3
450
EX (nm)
400 350 300 250 200 250 300 350 400 450 500 550 600 250 300 350 400 450 500 550 600
EM (nm)
EM (nm)
Fig. 3. The EEM spectrogram of supernatant in influent, cell 1–cell 3 and effluent at the 100th day of operation.
2013; Miura and Okabe, 2008). But there was 59.28 ± 16.46% of polysaccharide removed after the filtration (Fig. 2), which was the main source of polysaccharide in filter cake. The protein in supernatant could not be detected via chemical analysis, so the EEM was used to identify the change of humic-like and protein substrates qualitatively. There were two fluorescence zones in spectrogram for influent (Fig. 3) with kex = 225, kem = 338 (height of 1471) and kex = 275, kem = 340 (height of 2460) (Table 3), which were tryptophan-like substances (Henderson et al., 2009; Yue et al., 2015). In cell 1, the area of fluorescence zones were expanded and the humic-like substrates were generated, but the height of tryptophan-like substances were reduced by 63% (Table 3). Then the fluorescence zones were shrunk constantly and some new matters were generated after the treatment, and then that small zone of tryptophan-like substances was almost disappeared after filtration (Fig. 3 and Table 3). The height of humiclike substances was removed of 37.61% and the height of tryptophan-like substances was removed of 16.51% by the membrane rejection in cell 3 (Table 3). The change of fluorescence zones, substrates and its height indicated that the protein and humic-like substrates also underwent a process of generation and decomposition, which was similar to the polysaccharide in supernatant. All of those matters were the main components of EPS/SMP (Meng et al., 2009). It was a dynamic process for the EPS change in the supernatant of ABMBR, and part of them was rejected during filtration that developed the membrane fouling finally. But it did not mean that all of the EPS in filter cake just came from the supernatant, because the filter cake also was a kind of biofilm with the ability of secretion and decomposition of EPS (Meng et al., 2009). So, in this ABMBR the EPS in filter cake came from the supernatant rejection and biofilm generation, while the EPS in supernatant came from the influent and bulk sludge generation. The loose bound filter cake on membrane surface could be removed easily by the physical cleaning (tap water wash), most of the membrane resistance could be eliminated and the flux could be recovered. That guaranteed the operation of subsequent cycles. Actually, there was no biogas aeration, stirring and sludge reflux or any other membrane fouling mitigation approach except the
regular back-flushing. So, the strategy of membrane fouling mitigation in this ABMBR is successful and it is fit for the long term operation of anaerobic system with the least of energy consumption. 4. Conclusions In this study, ABMBR coupling with ABR and membrane filtration was proposed and operated under static condition without biogas aeration, stirring or any other turbulence intensifying strategy. Under room temperature, the membrane fouling developing rate was 0.0012–0.0021 MPa/d with regular back-flushing. The submicron particles in supernatant, with 0.35 ± 0.02 lm of particle size and 25.20 ± 1.83 mV of Zeta potential, were considered as the main membrane foulants, which were composed of polysaccharide, protein and humic-like substrates. In the filter cake, the polysaccharide accounted for more content and bigger area distribution than that of protein. Acknowledgments This study was supported by the science and technology foundation for talents and young researcher from Xi’an University of Architecture and Technology (Project No. RC1440, RC1441, QN1515 and QN1516). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.biortech.2016.05. 016. References APHA, AWWA, WEF, 2005. Standard Methods for the Examination of Waters and Wastewaters, 21st ed. American Public Health Association, Washington DC. Aslam, M., McCarty, P.L., Bae, J., Kim, J., 2014. The effect of fluidized media characteristics on membrane fouling and energy consumption in anaerobic fluidized membrane bioreactors. Sep. Purif. Technol. 132, 10–15. Bialek, K., Cysneiros, D., O’Flaherty, V., 2014. Hydrolysis, acidification and methanogenesis during low-temperature anaerobic digestion of dilute dairy
588
J. Liu et al. / Bioresource Technology 214 (2016) 582–588
wastewater in an inverted fluidised bioreactor. Appl. Microbiol. Biotechnol. 98, 8737–8750. Deng, L., Guo, W., Ngo, H.H., Zuthi, M.F.R., Zhang, J., Liang, S., Li, J., Wang, J., Zhang, X., 2015. Membrane fouling reduction and improvement of sludge characteristics by bioflocculant addition in submerged membrane bioreactor. Sep. Purif. Technol. 156, 450–458. Ding, Y., Tian, Y., Li, Z., Zuo, W., Zhang, J., 2015. A comprehensive study into fouling properties of extracellular polymeric substance (EPS) extracted from bulk sludge and cake sludge in a mesophilic anaerobic membrane bioreactor. Bioresour. Technol. 192, 105–114. FrØlund, B., Griebe, T., Nielsen, P.H., 1995. Enzymatic activity in the activatedsludge floc matrix. Appl. Microbiol. Biotechnol. 43, 755–761. Gao, D.W., Hu, Q., Yao, C., Ren, N.Q., Wu, W.M., 2014. Integrated anaerobic fluidizedbed membrane bioreactor for domestic wastewater treatment. Chem. Eng. J. 240, 362–368. Gao, W.J., Lin, H.J., Leung, K.T., Schraft, H., Liao, B.Q., 2011. Structure of cake layer in a submerged anaerobic membrane bioreactor. J. Membr. Sci. 374, 110–120. Gouveia, J., Plaza, F., Garralon, G., Fdz-Polanco, F., Peña, M., 2015. Long-term operation of a pilot scale anaerobic membrane bioreactor (AnMBR) for the treatment of municipal wastewater under psychrophilic conditions. Bioresour. Technol. 185, 225–233. Hahn, M.J., Figueroa, L.A., 2015. Pilot scale application of anaerobic baffled reactor for biologically enhanced primary treatment of raw municipal wastewater. Water Res. 87, 494–502. He, J.Z., Li, C.C., Wang, D.J., Zhou, D.M., 2015. Biofilms and extracellular polymeric substances mediate the transport of graphene oxide nanoparticles in saturated porous media. J. Hazard. Mater. 300, 467–474. Henderson, R.K., Bakera, A., Murphy, K.R., Hambly, A., Stuetza, R.M., Khana, S.J., 2009. Fluorescence as a potential monitoring tool for recycled water systems: a review. Water Res. 43, 863–881. Hu, S., Yang, F., Liu, S., Yu, L., 2009. The development of a novel hybrid aerating membrane-anaerobic baffled reactor for the simultaneous nitrogen and organic carbon removal from wastewater. Water Res. 43, 381–388. Ivanovic, I., Leiknes, T., ØØdegaard, H., 2008. Fouling control by reduction of submicron particles in a BF-MBR with an integrated flocculation zone in the membrane reactor. Sep. Sci. Technol. 43, 1871–1883. Kim, J., Shin, J., Kim, H., Lee, J.Y., Yoon, M., Wone, S., Lee, B.C., Song, K.G., 2014. Membrane fouling control using a rotary disk in a submerged anaerobic membrane sponge bioreactor. Bioresour. Technol. 172, 321–327. Kim, S.R., Oh, H.S., Jo, S.J., Yeon, K.M., Lee, C.H., Lim, D.J., Lee, C.H., Lee, J.K., 2013. Biofouling control with bead-entrapped quorum quenching bacteria in membrane bioreactors: physical and biological effects. Environ. Sci. Technol. 47, 836–842. Kola, A., Ye, Y., Le-Clech, P., Chen, V., 2014. Transverse vibration as novel membrane fouling mitigation strategy in anaerobic membrane bioreactor applications. J. Membr. Sci. 455, 320–329. Lay, C.H., Huang, C.Y., Chen, C.C., Lin, C.Y., 2016. Biohydrogen production in an anaerobic baffled stacking reactor: recirculation strategy and substrate concentration effects. Biochem. Eng. J. 109, 59–64. Liao, B., Kraemer, J.T., Bagley, D.M., 2006. Anaerobic membrane bioreactors: applications and research directions. Crit. Rev. Environ. Sci. Technol. 36, 489– 530.
Lin, H., Peng, W., Zhang, M., Chen, J., Hong, H., Zhang, Y., 2013. A review on anaerobic membrane bioreactors: applications, membrane fouling and future perspectives. Desalination 314, 169–188. Lin, H., Xie, K., Mahendran, B., Bagley, D.M., Leung, K.T., Liss, S.N., Liao, B.Q., 2009. Sludge properties and their effects on membrane fouling in submerged anaerobic membrane bioreactors (SAnMBRs). Water Res. 43, 3827–3837. Liu, Y., Chang, S., Defersha, F.M., 2015. Characterization of the proton binding sites of extracellular polymeric substances in an anaerobic membrane bioreactor. Water Res. 78, 133–143. Liu, L., Liu, J., Gao, B., Yang, F., 2012. Minute electric field reduced membrane fouling and improved performance of membrane bioreactor. Sep. Purif. Technol. 86, 106–112. McKeown, R.M., Scully, C., Mahony, T., Collins, G., O’Flaherty, V., 2009. Long-term (1243 days), low-temperature (4–15°C), anaerobic biotreatment of acidified wastewaters: bioprocess performance and physiological characteristics. Water Res. 43, 1611–1620. Meng, F., Chae, S.R., Drews, A., Kraume, M., Shin, H.S., Yang, F., 2009. Recent advances in membrane bioreactors (MBRs): membrane fouling and membrane material. Water Res. 43, 1489–1512. Metcalfe, D., Jarvis, P., Rockey, C., Judd, S., 2016. Pre-treatment of surface waters for ceramic microfiltration. Sep. Purif. Technol. 163, 173–180. Miura, Y., Okabe, S., 2008. Quantification of cell specific uptake activity of microbial products by uncultured chloroflexi by microautoradiography combined with fluorescence in situ hybridization. Environ. Sci. Technol. 42, 7380–7386. Ozgun, H., Tao, Y., Ersahin, M.E., Zhou, Z., Gimenez, J.B., Spanjers, H., Lier, J.B., 2015. Impact of temperature on feed-flow characteristics and filtration performance of an upflow anaerobic sludge blanket coupled ultrafiltration membrane treating municipal wastewater. Water Res. 83, 71–83. Ozgun, H., Dereli, R.K., Ersahin, M.E., Kinaci, C., Spanjers, H., Lier, J.B., 2013. A review of anaerobic membrane bioreactors for municipal wastewater treatment: integration options, limitations and expectations. Sep. Purif. Technol. 118, 89– 104. Skouteris, G., Hermosilla, D., López, P., Negro, C., Blanco, A., 2012. Anaerobic membrane bioreactors for wastewater treatment: a review. Chem. Eng. J. 198– 199, 138–148. Smith, A.L., Stadler, L.B., Love, N.G., Skerlos, S.J., Raskin, L., 2012. Perspectives on anaerobic membrane bioreactor treatment of domestic wastewater: a critical review. Bioresour. Technol. 122, 149–159. Xu, M., Wen, X., Huang, X., Yu, Z., Zhu, M., 2013. Mechanisms of membrane fouling controlled by online ultrasound in an anaerobic membrane bioreactor for digestion of waste activated sludge. J. Membr. Sci. 445, 119–126. Yue, X., Koh, Y.K.K., Ng, H.Y., 2015. Effects of dissolved organic matters (DOMs) on membrane fouling in anaerobic ceramic membrane bioreactors (AnCMBRs) treating domestic wastewater. Water Res. 86, 96–107. Zhang, Z., Lei, Z., Zhang, Z., Sugiura, N., Xu, X., Yin, D., 2007. Organics removal of combined wastewater through shallow soil infiltration treatment: a field and laboratory study. J. Hazard. Mater. 149, 657–665. Zhao, F., Su, Y., Tan, X., Chu, H., Zhang, Y., Yang, L., Zhou, X., 2015. Effect of temperature on extracellular organic matter (EOM) of Chlorella pyrenoidosa and effect of EOM on irreversible membrane fouling. In: Colloids Surf. B 136, 431– 439.
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Membrane fouling behavior in anaerobic baffled membrane bioreactor under static operating condition Jiadong Liu ⇑, Xiaolan Jia, Bo Gao, Longli Bo, Lei Wang School of Environmental and Municipal Engineering, Xi’an University of Architecture and Technology, Yan Ta Road, No. 13, Xi’an 710055, China
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
The membrane filtration was
combined with anaerobic baffled reactor. The ABMBR without turbulence intensifying strategy was developed. Membrane fouling developed slowly under static operating condition. The polysaccharide accounted for 79.12% of total area of filter cake. Submicron particles in supernatant were the source of fouling in ABMBR.
a r t i c l e
i n f o
Article history: Received 22 March 2016 Received in revised form 3 May 2016 Accepted 5 May 2016 Available online 7 May 2016 Keywords: Anaerobic membrane bioreactor Anaerobic baffled reactor Membrane filtration Static operation
a b s t r a c t A novel AnMBR combined with ABR as the anaerobic baffled membrane bioreactor (ABMBR) was developed for membrane fouling mitigation without any turbulence intensifying strategy to reduce the energy consumption further. The filtration time of this system lasted 14–25 days under stable condition only with back-flushing every 48 h. The polysaccharide accounted for 6.85 ± 3.1% amount of total filter cake and the protein accounted for 4.12 ± 2.1%, which took 79.12% and 11.12% of total area in laser scanning confocal microscope (CLSM) image. After filtration, 83.72 ± 10.97% of turbidity, 59.28 ± 16.46% of polysaccharide, 16.51% of tryptophan and 37.61% of humic-like substrates were rejected, respectively. The total membrane resistance at the end of each cycle was (4.47 ± 0.99) 1013 m 1. And the resistance from filter cake was (4.15 ± 1.00) 1013 m 1, which accounted for of 92.6 ± 3.4% of total membrane resistance. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction The research and development of anaerobic membrane bioreactor (AnMBR) have been carried out comprehensively in recent years, because of its advantage in energy conservation, anaerobic microorganism retention and higher quality of anaerobic effluent (Lin et al., 2013; Liao et al., 2006). Although the effluent from AnMBR still contains ammonia nitrogen, phosphate from influent and high concentration of soluble organic matters (Gouveia et al., 2015; Skouteris et al., 2012), its potential application in agricul⇑ Corresponding author. E-mail address: [email protected] (J. Liu). http://dx.doi.org/10.1016/j.biortech.2016.05.016 0960-8524/Ó 2016 Elsevier Ltd. All rights reserved.
tural irrigation and hydroponic system still makes this kind of wastewater treatment system popular in research. The high concentration of foulants and low intensity of turbulence without air aeration in anaerobic system can cause more serious membrane fouling phenomenon during filtration. Similar to membrane bioreactor (MBR), the configurations of AnMBR includes external cross-flow, internal submerged, or external submerged (Smith et al., 2012). For cross-flow system, the membrane fouling can be mitigated by the high speed of sludge flow (Skouteris et al., 2012), and additional turbulence is always applied in submerged AnMBR for fouling elimination (Ozgun et al., 2013). As there is artificially intensified turbulence, additional energy is necessary, which will weaken the advantage of AnMBR in energy
J. Liu et al. / Bioresource Technology 214 (2016) 582–588
consumption. And so far, many of the relative reports mainly focus on the novel membrane fouling mitigation technology in AnMBR for efficiency enhancement and energy consumption reduction during the operation (Skouteris et al., 2012; Ozgun et al., 2013). For cross-flow AnMBR, the acceleration of sludge flow can intensify the fouling mitigation, but there is a balance between the speed and efficiency (Skouteris et al., 2012). For submerged AnMBR, biogas aeration is the most popular approach for fouling elimination (Ozgun et al., 2013), but the special equipments and extra energy consumption for biogas aeration are essential, besides the security must be concerned if it is applied in practice. In AnMBR, filter cake on membrane surface is considered as the main factor for permeate flux decrease (Skouteris et al., 2012), so the turbulence enhancement for filter cake removal is the most direct and effective approach. Some novel ways for membrane fouling mitigation have been proposed, which include ultrasound (Xu et al., 2013), transverse vibration (Kola et al., 2014), rotary disk stirring (Kim et al., 2014) and fluidized media (Aslam et al., 2014) or granule activated carbon (Gao et al., 2014) addition, etc. Those kinds of turbulence enhancement approaches still work based on energy consumption. And the addition of fluidized media or granule activated carbon also needs the assistance of biogas aeration. So, there are few AnMBR without additional energy consumption for fouling mitigation had been reported previously. As the filter cake takes most of the fouling quotient, the turbulence is necessary for cake removal during operation. If no turbulence is applied, the cake formation could be reduced directly by reduction of the circumstance concentration of foulants (Metcalfe et al., 2016), which can also attain the goal of membrane fouling mitigation. In this case, the structure of the reactor should be improved. The anaerobic baffled reactor (ABR) is one of effective anaerobic wastewater treatment system (Hu et al., 2009). The ABR is composed by series of up-flow anaerobic sludge blanket (UASB), and the influent traverses the whole system in the form of reciprocation. The pollutants can be removed by the microorganisms when the wastewater flows through the sedimental sludge blanket (Lay et al., 2016). The footprint of wastewater is lengthened and the hydraulic retention time (HRT) can be separated from sludge retention time (SRT), so that the wastewater treatment efficiency is high and the operation is steerable (Hahn and Figueroa, 2015). The most important is that there is no turbulence during the treatment process, and there is no suspended sludge in supernatant at the top of each cell. If the ABR is combined with AnMBR, the membrane module assembled in the supernatant, which can be more efficient in mitigation of membrane fouling comparing with traditional submerged AnMBR, as the foulants concentration in supernatant is much lower than that in bulk sludge. And the total suspended solids in supernatant is reduced constantly when the treatment cell is added in ABR (Hahn and Figueroa, 2015), which means the concentration of membrane foulants in ABR would be reduced at the later cell. So the membrane fouling should be controlled effectively if the cell number of ABR and operation condition is appropriate. In this study, a novel AnMBR combined with ABR as the anaerobic baffled membrane bioreactor (ABMBR) was proposed. The raw domestic wastewater was used as influent. The membrane module was set in the supernatant of ABR’s later cell. The long-term operation was carried out to identify the membrane fouling behavior. The composition of membrane resistance was analyzed during the physical and chemical cleaning process, the structure of filter cake and fouled membrane were characterized, and the features of supernatant were revealed to identify the membrane fouling behavior in ABMBR. The parameters of water quality of supernatant in each cell were analyzed to identify the relationship between the pollutants transformation and membrane fouling behavior in this novel ABMBR.
583
2. Materials and methods A 49 L polymethyl methacrylate ABR equipped with 5 cells was used in this study. The active volume of one cell was 8 L (for first four cells) and it was 11.5 L for the last cell for membrane module installation, respectively. The membrane used in this experiment was the self-produced hollow fiber polyvinylidene fluoride (PVDF) membrane with polyethylene terephthalate (PET) internal support, which with outside diameter of 2.5 mm and 31 pieces of membrane fiber were sealed in polyethylene tube in the form of ‘‘U” for one membrane module. Two membrane modules with total active area of 0.28 m2 were used in parallel for dead-end filtration. In order to identify filtration and wastewater treatment efficiency of our ABMBR under different operational conditions, the membrane modules were installed in the 5th cell for the first two cycles, and then moved into the 3rd cell during the rest tests. The cover plate was sealed on the top of reactor with silicone gasket and bolts for anaerobic condition. The tube sealed on the cover plate was used for effluent and biogas spilling, and water sealing system was used for biogas collection. The whole system was operated for 130 days. The raw wastewater for influent was drawn from the sewer at the campus of Xi’an University of Architecture and Technology after 3 pm each time, the water quality parameters of influent, supernatant and effluent during each operating condition were shown in Figs. S1–S3, which included chemical oxygen demand (COD), suspended solids (SS), ammonia nitrogen (NH3-N), nitrite and nitrate nitrogen (NO2 -N, NO3 -N), phosphate (PO34 ) and were measured every one week according to the standard methods (APHA, 2005). And the wastewater was filtrated via sifter with 0.9 mm of pore size for big size of particles or suspended solids removal before using. The anaerobic sludge was taken from the bottom of UASB system of Hans brewery (Xi’an, China), which was filled in the ABMBR and the acclimatization process lasted for 30 days with continuous influent of domestic wastewater and without membrane filtration. For five cells of operation, a 5 L of anaerobic activated sludge with mixed liquor suspended solids (MLSS) of 48.05 ± 0.49 g/L was injected in the first four cells (sludge thickness of 35–38 cm), respectively. The initial MLSS was 30.03 ± 0.31 g/L (including the volume of supernatant) and with final MLSS of 32.71 ± 6.63 g/L in the 1st cell and 2nd cell together at the end of operation (including the volume of supernatant). For three cells of operation, most of the sludge was discharged from the 3rd cell (residual sludge thickness of 5–8 cm). For sludge containing cell, the level of sediment sludge was lower than the sampling hole, which was 45 cm from the bottom of reactor. For the cell with membrane module, the sludge level kept lower to make sure that the module stayed away from the sediments. And no regular sludge discharging was carried out during the whole operation. The peristaltic pumps (Langer, China) were used for influent and effluent. The speed of influent was set at 1.11 L/h (7 rpm) with constant working. The speed for effluent was set at 4 L/m2/h (8 rpm) with working mode of 5 min ‘‘on” and 1 min ‘‘off”, which was controlled by the time switcher. The HRT for five cells of operation was 39.19 h and it was 21.62 h for three cells of operation. The transmembrane pressure (TMP) was shown via vacuum meter and the membrane resistance was calculated by Darcy’s law (Liu et al., 2012). The whole study included 9 operating cycles. The back-flushing was carried out by reversed influent pump every 48 h with 1 L of tap water from the 1st cycle to the 6th cycle, which lasted about 1 h and then followed the mode of 5 min ‘‘on” and 1 min ‘‘off”. For the 7th, 8th and 9th cycles, the back-flushing was carried out every 24 h with 500 mL tap water for 30 min, because the
J. Liu et al. / Bioresource Technology 214 (2016) 582–588
membrane fouling was more serious during wintertime. From the 1st cycle to the 6th cycle, the filtration was suspended when the TMP was higher than that of 0.03 MPa; for the 7th, 8th and 9th cycles, the filtration was suspended when 50% of the initial flux was lost. The physical cleaning was carried out after the first, third, forth, sixth and eighth cycles, the chemical cleaning was carried out after the second, fifth and seventh cycles. The physical cleaning was carried out in the form of high pressure tap water flush. The chemical cleaning procedure included: physical cleaning for filter cake removal, immersing in HCl solution with pH = 2 for 4 h, tap water flush, immersing in 0.5 g/L NaOH and 5 mL/L NaClO mixed solution for 4 h and tap water flush. The flux and TMP data of pure water after the physical cleaning, acid and alkali cleaning were recorded for membrane resistance ingredient analysis. The pH, turbidity and UV254 were measured by pH meter (PHS-3C, INESA, China), turbidity meter (GDS-3B, KEDA, China) and ultraviolet–visible spectrophotometer (UV 2102 C, UNIC), respectively. All of the water samples were filtrated via qualitative filter paper before testing. The extracellular polymeric substances (EPS, polysaccharide and protein) in supernatant and filter cake of different cells were extracted according to the following steps: 10 mL of supernatant or filter cake solution was treated by ultrasound for 10 min and then heated in boiling water bath for 30 min, then the natural cooling solution was filtrated by qualitative filter paper after shaking adequately and the permeate was collected for tests. The MLSS and mixed liquor volatile suspended solid (MLVSS) were measured according to the standard methods (APHA, 2005). The concentration of polysaccharide and protein were tested according to previous report (FrØlund et al., 1995). The glucose and bovine serum albumin (A1933-100G, Sigma) were used for standard curve definition. All of the reagents used in this study were analytical pure. The particle size and Zeta potential of suspended solid in unfiltered supernatant were measured with Zeta potential analyzer (Zetasizer Nano ZS, Malvern, England) at the 100th day of operation. The protein and humic-like substrates in unfiltered (raw sample) supernatant were detected qualitatively with fluorescence excitation–emission matrix (EEM) spectroscopy (FP-6500, Jasco, Japan). The filter cake on membrane surface was measured with confocal laser scanning microscope (CLSM, Leica SP8, Germany) at the end of operation following the previous reported method (Kim et al., 2013), and the images were analyzed by the software of Photoshop CS2 for the area calculation of different color zones.
3. Results and discussion 3.1. Membrane fouling behavior in ABMBR For 5 cells of operation, two filtration cycles were carried out, which lasted 14 and 16 days (Fig. 1), respectively. One time of physical cleaning was taken between those two cycles. Seven filtration cycles was carried out for 3 cells of operation, which lasted 25, 22, 14, 18, 10, 6 and 3 days, respectively (Fig. 1). Two times of chemical cleaning were conducted between the 5th and 6th cycles, 7th and 8th cycles, respectively. And only physical cleaning was carried out for the rest cycles. Comparing with 5 cells of operation, there was no deterioration of membrane fouling appeared under 3 cells of operational condition, as the filtration time of the 3rd, 4th and 6th cycles were longer than that of the 1st and 2nd cycles. With shorter HRT, the system was still in smooth operation under static operating condition. In terms of the overall situation, the first 6 cycles were stable and the rising trend of membrane resistance was very gentle. But filtration time was reduced and membrane resistance was
15
Mmebrane resistance (1013 m-1)
584
5 cells
# Chemical cleaning
12
9
*
6
3
0
0
#
*
* Physical cleaning
3 cells
* *
#
#
*
10 20 30 40 50 60 70 80 90 100 110 120 130
Time (day) Fig. 1. Evolution of membrane resistance in ABMBR during the entire period of operation.
increased sharply during the last 3 cycles (Fig. 1) although the back-flushing interval was reduced from 48 to 24 h. The composition of membrane resistance was identified during the physical and chemical cleaning processes. Except for the last three cycles, the total membrane resistance at the end of each cycle was 3.04 1013–5.59 1013 m 1 (average value of (4.47 ± 0.99) 1013 m 1, Table 1). And the resistance from filter cake was 2.83 1013–5.34 1013 m 1 (average value of (4.15 ± 1.00) 1013 m 1), which accounted for 86.4–97.1% (average value of 92.6 ± 3.4%) of total membrane resistance. For the 7th and 8th cycles, the resistance of filter cake accounted for 97.4–98.8% of the total membrane resistance (Table 1). The MLSS and MLVSS of filter cake were measured based on membrane area (Table 1). From the 3rd to 6th cycles, the MLSS was 6.76 ± 0.22 11.05 ± 0.72 g/m2 and the MLVSS was 3.83 10.03 ± 0.03 g/m2. The similar MLSS and MLVSS amount was shown on membrane surface for the 7th and 8th cycles, although the filtration only lasted 6 and 10 days (Table 1). The membrane fouling developing rate was 0.0012– 0.0021 MPa/d under stable condition. Comparing with those fouling mitigation methods with energy consumption (Xu et al., 2013; Gao et al., 2014; Kola et al., 2014; Kim et al., 2014), the membrane fouling behavior in this study was sharper as there was no obvious retarded stage (Fig. 1). That might caused by lacking of continuous membrane cleaning strategy, because those constant in suit cleaning ways will make a balance between foulants attachment and detachment for a certain time (Meng et al., 2009).
3.2. Membrane fouling mechanism in ABMBR Previous reports indicated that the filter cake took most of membrane resistance in traditional AnMBR (Skouteris et al., 2012). In this study, the membrane with 0.1 lm pore size was used, and the average particle size of suspended solid in supernatant of cell 3 was 0.35 ± 0.02 lm (Table 2), so that most of those particles would be rejected during filtration to form filter cake. The particle size of suspended solid in supernatant of cell 1 and cell 2 was 0.36 ± 0.02 and 0.37 ± 0.03 lm (Table 2), respectively, which meant that most of the suspended solid in supernatant of this ABMBR were submicron particles. The Zeta potential of 24.37 ± 0.72 25.20 ± 1.83 mV (Table 2) also indicated that those suspended solid particles were similar to the active sludge or microbial flocs (Deng et al., 2015). Actually, the submicron
585
J. Liu et al. / Bioresource Technology 214 (2016) 582–588 Table 1 The compositions of membrane resistance and filter cake.
TR* (1013 m 1) FR* (1013 m 1) FR/TR* IOR* (1013 m 1) IOR/TR* OR* (1013 m 1) OR/TR* MLSS (g/m2) MLVSS (g/m2) Polysaccharide (mg/ m2) Protein (mg/m2) *
Physical cleaning
Chemical cleaning
Physical cleaning
Physical cleaning
Chemical cleaning
Physical cleaning
Chemical cleaning
Physical cleaning
5.59 5.22 93.36% – – – – – – –
4.13 3.57 86.46% 0.12 2.92% 0.17 4.05% – – –
3.04 2.83 93.12% – – – – 7.42 ± 0.98 3.83 618.64
5.50 5.34 97.10% – – – – 10.84 ± 0.25 8.09 ± 0.35 469.14
4.68 4.36 93.05% 0.03 0.64% 0.05 1.15% 6.76 ± 0.22 6.15 ± 0.44 816.98
3.87 3.58 92.55% – – – – 11.05 ± 0.72 10.03 ± 0.03 565.42
6.07 5.92 97.4% 0.03 0.49% – – 7.71 ± 0.28 6.44 ± 0.26 566.00
13.43 13.28 98.84% – – – – 16.26 ± 0.13 13.98 ± 0.14 632.27
–
–
589.82
67.28
515.42
287.84
188.90
561.42
TR: total membrane resistance; FR: resistance from filter cake; IOR: resistance from inorganic matters; OR: resistance from organic matters.
Table 2 The properties of solid particles in influent and supernatant.
Zeta potential (mV)* Particle size (lm)* *
Influent
Cell 1
Cell 2
Cell 3
24.57 ± 1.29 0.66 ± 0.06
24.47 ± 1.06 0.36 ± 0.02
24.37 ± 0.72 0.37 ± 0.03
25.20 ± 1.83 0.35 ± 0.02
Average values (n = 3).
particles determined the membrane flux in traditional AnMBR and MBR (Ozgun et al., 2013; Ivanovic et al., 2008), because that pore plugging/clogging could result in formation of compact filter cake. There was no direct contact between membrane module and bulk sludge in this ABMBR, consequently the submicron suspended solid would be the main component of filter cake and mainly contributed to the membrane fouling. In this ABMBR, the inorganic matters just accounted for 0.49–2.92% of membrane resistance and the rest 1.15–4.05% came from the residual organic matters (Table 1). The inorganic matters, such as MgNH4PO46H2O, K2NH4PO4 and/or CaCO3, etc, were easy to be formed on membrane surface (Ozgun et al., 2013), because there was no obvious removal efficiency for those matters in anaerobic treatment system. And the concentrations of Ca2+, Mg2+, NH4+ and PO34 in domestic wastewater would cause serious membrane fouling in AnMBR (Ozgun et al., 2013). But it was different in this ABMBR, the filter cake and residual organic matters took more than 90% of the membrane resistance. So, it is necessary to indentify the composition of filter cake and suspended solid. The EPS or soluble microbial products (SMP) were considered as the main organic foulant, which included proteins, carbohydrates, nucleic acids, lipids (Liu et al., 2015) and humic acids (Meng et al., 2009). A series of researches had focused on the EPS on membrane fouling in AnMBR (Meng et al., 2009; Ding et al., 2015). The EPS from bulk sludge and filter cake showed different fouling properties and the hydrophobic neutral (HPO-N) in EPS worked as the main foulant on membrane surface (Ding et al., 2015). In this ABMBR, the polysaccharide and protein in filter cake was 469.14–816.98 mg/m2 (611.41 ± 115.9 mg/m2) and 67.28–589.82 mg/m2 (368.45 ± 217.84 mg/m2) (Table 1), respectively. The polysaccharide accounted for 6.85 ± 3.1% of total filter cake and the protein accounted for 4.12 ± 2.1%. For observing the distribution of the protein, polysaccharide and bacteria in filter cake, the cake sample was examined microscopically with CLSM after staining. Those images revealed that the distribution of protein was wider and uniform and the polysaccharide was more concentrated (Fig. S4). The polysaccharide, protein and total cell took 79.12%, 11.12% and 0.12% of total area in combined image, respectively (Fig. S4). Those results corresponded
to filter cake composition that the content of polysaccharide was higher than that of protein (Table 1). There was no unanimous verdict about the efficiency of polysaccharide and protein on membrane fouling, some reports indicated that the contribution from protein was more obvious than that from polysaccharide, but some contradictory versions also had been presented (Meng et al., 2009), because the composition of EPS is variable and complicated under different operational conditions. More polysaccharide was found in filter cake of AnMBR for pulping wastewater treatment through CLSM analysis (Lin et al., 2009). But it was conflict with CLSM results from Gao et al. (2011) that the protein in filter accounted for 37.72 ± 5.77 87.33 ± 1.42% of EPS and the polysaccharide only took 1.56 ± 0.25 7.51 ± 3.25% in AnMBR for thermomechanical pulping whitewater treatment. When the domestic wastewater was used as influent of AnMBR, the protein was the dominant component of EPS in the filter cake, which could be adsorbed by granular activated carbon for fouling mitigation (Gao et al., 2014). Those conflicting conclusions revealed the complexity of EPS/SMP in anaerobic reactor and filter cake. In this study, the amount and area distribution of polysaccharide was more than that of protein, which would account for more membrane resistance ratio in filter cake. For simulating the actual condition, no temperature control system was applied to maintain the temperature of the reactor. The air temperature was reduced constantly after summer and a precipitate dropping of temperature was occurred at the 116th day (Fig. S5), which just corresponded to the beginning of sudden deterioration of membrane fouling (Fig. 1). The increase of turbidity, UV254 and polysaccharide in influent and supernatant indicated that the deterioration of influent quality and treatment efficiency was the reason for sudden deterioration of membrane fouling (Figs. 2 and S6). The EEM height of tryptophan-like and humiclike substances in all cells and effluent also were increased during the last 3 cycles (Fig. S7 and Table S1). That deterioration of influent and supernatant caused fast-forming of filter cake directly. The ambient temperature would affect the metabolic activity and microorganism species of anaerobic system obviously (Bialek et al., 2014; McKeown et al., 2009). Comparing with 20–35 °C, the secretion rate of extracellular organic matter (EOM or EPS)
175 93 372 77 155 257 589 210/400 215/462 220/336 220/472 265/292 275/304 275/342 A A A T T 1471 2460 225/338 275/340
A – humic-like substances; B – tyrosine-like substances; T – tryptophan-like substances.
Cell 2
Substances EX/EM (nm)
Height Cell 1
EX/EM (nm)
Substances Height Influent
Table 3 The peak/height of EEM and relative substances at the 100th day of operation.
In the ABMBR, the foulants on membrane surface came from the supernatant directly, so the relative parameters including turbidity, UV254, humic-like substances, protein and polysaccharide were measured to discover the relationship between the properties of supernatant and membrane fouling. In this study, raw wastewater from the sewer at the campus of University was used as influent. The COD in influent was 225–325 mg/L. After treatment, about 53% COD could be removed under different operating condition and the membrane rejection accounted 10–30% of COD removal (Figs. S1–S3). Actually, the contents of polysaccharide and humic-like substrates in influent were high (Figs. 2, S6, Tables 3 and S1), which were not easy to be degraded (Kim et al., 2013; Miura and Okabe, 2008) but part of them could be rejected by the membrane. And as an anaerobic system, there was no obvious efficiency on NH3-N and PO34 removal (Figs. S1–S3). After treatment process of 3 cells of ABR and membrane filtration, the turbidity and UV254 of the supernatant were decreased (Fig. 2). The turbidity represented the content of all suspended solids and UV254 reflected the aggregation of aromatic humic and fulvic structures (Zhang et al., 2007). The reduction of those parameters in supernatant indicated that the static operation of ABMBR could inhibit the foulants generation after the sequential treatment, which was good for anaerobic membrane filtration without turbulence intensifying strategy. After the filtration, the turbidity was reduced 83.72 ± 10.97% because of membrane rejection. And it was only 28.21 ± 8.29% for UV254 removal because some of aromatic humic and fulvic structures were soluble, which could pass through the membrane during filtration. The polysaccharide could be detected in influent and supernatant but without obvious removal efficiency after the anaerobic sequential treatment (Fig. 2). The polysaccharide was secreted by microorganism constantly during the metabolic process (Meng et al., 2009; He et al., 2015) and part of them could be decomposed by respiratory action of some special bacteria (Kim et al.,
Height
3.3. The relationship between the properties of supernatant and membrane fouling behavior
*
Substances
Cell 3
was accelerated under lower temperature condition of 15 °C (Zhao et al., 2015) and caused the higher SMP concentrations and turbidity in supernatant (Ozgun et al., 2015), which was similar to this study in ABMBR. The influent was draw from the sewer that came from the septic tank, which also under anaerobic condition, so the wastewater quality was deteriorated at the same time.
242 115 196 424 727
Height
Fig. 2. The transformation of pH, turbidity, polysaccharide and UV254 in influent, subsequent 3 cells and effluent from the 2nd to the 5th operational cycle.
205/412 210/450 215/566 220/336 280/338
Effluent
A T* T* T*
Cell 3
182 424 861 861
Cell 2
210/388 220/338 275/340 280/338
Cell 1
A A* T* T*
0.0
Influent
Substances
0
210/402 210/578 220/336 275/338
10
185 164 475 987
Substances Effluent
0.2
T T*
20
Height
0.4
30
EX/EM (nm)
40
*
0.6
*
50
EX/EM (nm)
60
*
0.8 pH Turbidity Polysaccharide UV254
EX/EM (nm)
70
A A T A B B T
J. Liu et al. / Bioresource Technology 214 (2016) 582–588
UV254
pH, Turbidity (NTU), Polysaccharide (mg/L)
586
587
J. Liu et al. / Bioresource Technology 214 (2016) 582–588
500
Influent
Cell 1
Cell2
EX (nm)
450 400 350 300 250 200 500
250 300 350 400 450 500 550 600
EM (nm)
Effluent
Cell3
450
EX (nm)
400 350 300 250 200 250 300 350 400 450 500 550 600 250 300 350 400 450 500 550 600
EM (nm)
EM (nm)
Fig. 3. The EEM spectrogram of supernatant in influent, cell 1–cell 3 and effluent at the 100th day of operation.
2013; Miura and Okabe, 2008). But there was 59.28 ± 16.46% of polysaccharide removed after the filtration (Fig. 2), which was the main source of polysaccharide in filter cake. The protein in supernatant could not be detected via chemical analysis, so the EEM was used to identify the change of humic-like and protein substrates qualitatively. There were two fluorescence zones in spectrogram for influent (Fig. 3) with kex = 225, kem = 338 (height of 1471) and kex = 275, kem = 340 (height of 2460) (Table 3), which were tryptophan-like substances (Henderson et al., 2009; Yue et al., 2015). In cell 1, the area of fluorescence zones were expanded and the humic-like substrates were generated, but the height of tryptophan-like substances were reduced by 63% (Table 3). Then the fluorescence zones were shrunk constantly and some new matters were generated after the treatment, and then that small zone of tryptophan-like substances was almost disappeared after filtration (Fig. 3 and Table 3). The height of humiclike substances was removed of 37.61% and the height of tryptophan-like substances was removed of 16.51% by the membrane rejection in cell 3 (Table 3). The change of fluorescence zones, substrates and its height indicated that the protein and humic-like substrates also underwent a process of generation and decomposition, which was similar to the polysaccharide in supernatant. All of those matters were the main components of EPS/SMP (Meng et al., 2009). It was a dynamic process for the EPS change in the supernatant of ABMBR, and part of them was rejected during filtration that developed the membrane fouling finally. But it did not mean that all of the EPS in filter cake just came from the supernatant, because the filter cake also was a kind of biofilm with the ability of secretion and decomposition of EPS (Meng et al., 2009). So, in this ABMBR the EPS in filter cake came from the supernatant rejection and biofilm generation, while the EPS in supernatant came from the influent and bulk sludge generation. The loose bound filter cake on membrane surface could be removed easily by the physical cleaning (tap water wash), most of the membrane resistance could be eliminated and the flux could be recovered. That guaranteed the operation of subsequent cycles. Actually, there was no biogas aeration, stirring and sludge reflux or any other membrane fouling mitigation approach except the
regular back-flushing. So, the strategy of membrane fouling mitigation in this ABMBR is successful and it is fit for the long term operation of anaerobic system with the least of energy consumption. 4. Conclusions In this study, ABMBR coupling with ABR and membrane filtration was proposed and operated under static condition without biogas aeration, stirring or any other turbulence intensifying strategy. Under room temperature, the membrane fouling developing rate was 0.0012–0.0021 MPa/d with regular back-flushing. The submicron particles in supernatant, with 0.35 ± 0.02 lm of particle size and 25.20 ± 1.83 mV of Zeta potential, were considered as the main membrane foulants, which were composed of polysaccharide, protein and humic-like substrates. In the filter cake, the polysaccharide accounted for more content and bigger area distribution than that of protein. Acknowledgments This study was supported by the science and technology foundation for talents and young researcher from Xi’an University of Architecture and Technology (Project No. RC1440, RC1441, QN1515 and QN1516). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.biortech.2016.05. 016. References APHA, AWWA, WEF, 2005. Standard Methods for the Examination of Waters and Wastewaters, 21st ed. American Public Health Association, Washington DC. Aslam, M., McCarty, P.L., Bae, J., Kim, J., 2014. The effect of fluidized media characteristics on membrane fouling and energy consumption in anaerobic fluidized membrane bioreactors. Sep. Purif. Technol. 132, 10–15. Bialek, K., Cysneiros, D., O’Flaherty, V., 2014. Hydrolysis, acidification and methanogenesis during low-temperature anaerobic digestion of dilute dairy
588
J. Liu et al. / Bioresource Technology 214 (2016) 582–588
wastewater in an inverted fluidised bioreactor. Appl. Microbiol. Biotechnol. 98, 8737–8750. Deng, L., Guo, W., Ngo, H.H., Zuthi, M.F.R., Zhang, J., Liang, S., Li, J., Wang, J., Zhang, X., 2015. Membrane fouling reduction and improvement of sludge characteristics by bioflocculant addition in submerged membrane bioreactor. Sep. Purif. Technol. 156, 450–458. Ding, Y., Tian, Y., Li, Z., Zuo, W., Zhang, J., 2015. A comprehensive study into fouling properties of extracellular polymeric substance (EPS) extracted from bulk sludge and cake sludge in a mesophilic anaerobic membrane bioreactor. Bioresour. Technol. 192, 105–114. FrØlund, B., Griebe, T., Nielsen, P.H., 1995. Enzymatic activity in the activatedsludge floc matrix. Appl. Microbiol. Biotechnol. 43, 755–761. Gao, D.W., Hu, Q., Yao, C., Ren, N.Q., Wu, W.M., 2014. Integrated anaerobic fluidizedbed membrane bioreactor for domestic wastewater treatment. Chem. Eng. J. 240, 362–368. Gao, W.J., Lin, H.J., Leung, K.T., Schraft, H., Liao, B.Q., 2011. Structure of cake layer in a submerged anaerobic membrane bioreactor. J. Membr. Sci. 374, 110–120. Gouveia, J., Plaza, F., Garralon, G., Fdz-Polanco, F., Peña, M., 2015. Long-term operation of a pilot scale anaerobic membrane bioreactor (AnMBR) for the treatment of municipal wastewater under psychrophilic conditions. Bioresour. Technol. 185, 225–233. Hahn, M.J., Figueroa, L.A., 2015. Pilot scale application of anaerobic baffled reactor for biologically enhanced primary treatment of raw municipal wastewater. Water Res. 87, 494–502. He, J.Z., Li, C.C., Wang, D.J., Zhou, D.M., 2015. Biofilms and extracellular polymeric substances mediate the transport of graphene oxide nanoparticles in saturated porous media. J. Hazard. Mater. 300, 467–474. Henderson, R.K., Bakera, A., Murphy, K.R., Hambly, A., Stuetza, R.M., Khana, S.J., 2009. Fluorescence as a potential monitoring tool for recycled water systems: a review. Water Res. 43, 863–881. Hu, S., Yang, F., Liu, S., Yu, L., 2009. The development of a novel hybrid aerating membrane-anaerobic baffled reactor for the simultaneous nitrogen and organic carbon removal from wastewater. Water Res. 43, 381–388. Ivanovic, I., Leiknes, T., ØØdegaard, H., 2008. Fouling control by reduction of submicron particles in a BF-MBR with an integrated flocculation zone in the membrane reactor. Sep. Sci. Technol. 43, 1871–1883. Kim, J., Shin, J., Kim, H., Lee, J.Y., Yoon, M., Wone, S., Lee, B.C., Song, K.G., 2014. Membrane fouling control using a rotary disk in a submerged anaerobic membrane sponge bioreactor. Bioresour. Technol. 172, 321–327. Kim, S.R., Oh, H.S., Jo, S.J., Yeon, K.M., Lee, C.H., Lim, D.J., Lee, C.H., Lee, J.K., 2013. Biofouling control with bead-entrapped quorum quenching bacteria in membrane bioreactors: physical and biological effects. Environ. Sci. Technol. 47, 836–842. Kola, A., Ye, Y., Le-Clech, P., Chen, V., 2014. Transverse vibration as novel membrane fouling mitigation strategy in anaerobic membrane bioreactor applications. J. Membr. Sci. 455, 320–329. Lay, C.H., Huang, C.Y., Chen, C.C., Lin, C.Y., 2016. Biohydrogen production in an anaerobic baffled stacking reactor: recirculation strategy and substrate concentration effects. Biochem. Eng. J. 109, 59–64. Liao, B., Kraemer, J.T., Bagley, D.M., 2006. Anaerobic membrane bioreactors: applications and research directions. Crit. Rev. Environ. Sci. Technol. 36, 489– 530.
Lin, H., Peng, W., Zhang, M., Chen, J., Hong, H., Zhang, Y., 2013. A review on anaerobic membrane bioreactors: applications, membrane fouling and future perspectives. Desalination 314, 169–188. Lin, H., Xie, K., Mahendran, B., Bagley, D.M., Leung, K.T., Liss, S.N., Liao, B.Q., 2009. Sludge properties and their effects on membrane fouling in submerged anaerobic membrane bioreactors (SAnMBRs). Water Res. 43, 3827–3837. Liu, Y., Chang, S., Defersha, F.M., 2015. Characterization of the proton binding sites of extracellular polymeric substances in an anaerobic membrane bioreactor. Water Res. 78, 133–143. Liu, L., Liu, J., Gao, B., Yang, F., 2012. Minute electric field reduced membrane fouling and improved performance of membrane bioreactor. Sep. Purif. Technol. 86, 106–112. McKeown, R.M., Scully, C., Mahony, T., Collins, G., O’Flaherty, V., 2009. Long-term (1243 days), low-temperature (4–15°C), anaerobic biotreatment of acidified wastewaters: bioprocess performance and physiological characteristics. Water Res. 43, 1611–1620. Meng, F., Chae, S.R., Drews, A., Kraume, M., Shin, H.S., Yang, F., 2009. Recent advances in membrane bioreactors (MBRs): membrane fouling and membrane material. Water Res. 43, 1489–1512. Metcalfe, D., Jarvis, P., Rockey, C., Judd, S., 2016. Pre-treatment of surface waters for ceramic microfiltration. Sep. Purif. Technol. 163, 173–180. Miura, Y., Okabe, S., 2008. Quantification of cell specific uptake activity of microbial products by uncultured chloroflexi by microautoradiography combined with fluorescence in situ hybridization. Environ. Sci. Technol. 42, 7380–7386. Ozgun, H., Tao, Y., Ersahin, M.E., Zhou, Z., Gimenez, J.B., Spanjers, H., Lier, J.B., 2015. Impact of temperature on feed-flow characteristics and filtration performance of an upflow anaerobic sludge blanket coupled ultrafiltration membrane treating municipal wastewater. Water Res. 83, 71–83. Ozgun, H., Dereli, R.K., Ersahin, M.E., Kinaci, C., Spanjers, H., Lier, J.B., 2013. A review of anaerobic membrane bioreactors for municipal wastewater treatment: integration options, limitations and expectations. Sep. Purif. Technol. 118, 89– 104. Skouteris, G., Hermosilla, D., López, P., Negro, C., Blanco, A., 2012. Anaerobic membrane bioreactors for wastewater treatment: a review. Chem. Eng. J. 198– 199, 138–148. Smith, A.L., Stadler, L.B., Love, N.G., Skerlos, S.J., Raskin, L., 2012. Perspectives on anaerobic membrane bioreactor treatment of domestic wastewater: a critical review. Bioresour. Technol. 122, 149–159. Xu, M., Wen, X., Huang, X., Yu, Z., Zhu, M., 2013. Mechanisms of membrane fouling controlled by online ultrasound in an anaerobic membrane bioreactor for digestion of waste activated sludge. J. Membr. Sci. 445, 119–126. Yue, X., Koh, Y.K.K., Ng, H.Y., 2015. Effects of dissolved organic matters (DOMs) on membrane fouling in anaerobic ceramic membrane bioreactors (AnCMBRs) treating domestic wastewater. Water Res. 86, 96–107. Zhang, Z., Lei, Z., Zhang, Z., Sugiura, N., Xu, X., Yin, D., 2007. Organics removal of combined wastewater through shallow soil infiltration treatment: a field and laboratory study. J. Hazard. Mater. 149, 657–665. Zhao, F., Su, Y., Tan, X., Chu, H., Zhang, Y., Yang, L., Zhou, X., 2015. Effect of temperature on extracellular organic matter (EOM) of Chlorella pyrenoidosa and effect of EOM on irreversible membrane fouling. In: Colloids Surf. B 136, 431– 439.