Effects of dissolved organic matters (DOMs) on membrane fouling in anaerobic ceramic membrane bioreactors (AnCMBRs) treating domestic wastewater

Water Research xxx (2015) 1e12

Contents lists available at ScienceDirect

Water Research journal homepage: www.elsevier.com/locate/watres

Effects of dissolved organic matters (DOMs) on membrane fouling in anaerobic ceramic membrane bioreactors (AnCMBRs) treating domestic wastewater Xiaodi Yue a, Yoong Keat Kelvin Koh b, How Yong Ng a, * a b

Centre for Water Research, Department of Civil and Environmental Engineering, National University of Singapore, 1 Engineering Drive 2, 117576, Singapore Public Utilities Board, 40 Scotts Road #22-01, Environment Building, 228231, Singapore

a r t i c l e i n f o

a b s t r a c t

Article history: Received 7 January 2015 Received in revised form 6 July 2015 Accepted 21 July 2015 Available online xxx

Anaerobic membrane bioreactors (AnMBRs) have been regarded as a potential solution to achieve energy neutrality in the future wastewater treatment plants. Coupling ceramic membranes into AnMBRs offers great potential as ceramic membranes are resistant to corrosive chemicals such as cleaning reagents and harsh environmental conditions such as high temperature. In this study, ceramic membranes with pore sizes of 80, 200 and 300 nm were individually mounted in three anaerobic ceramic membrane bioreactors (AnCMBRs) treating real domestic wastewater to examine the treatment efficiencies and to elucidate the effects of dissolved organic matters (DOMs) on fouling behaviours. The average overall chemical oxygen demands (COD) removal efficiencies could reach around 86e88%. Although CH4 productions were around 0.3 L/g CODutilised, about 67% of CH4 generated was dissolved in the liquid phase and lost in the permeate. When filtering mixed liquor of similar properties, smaller pore-sized membranes fouled slower in long-term operations due to lower occurrence of pore blockages. However, total organic removal efficiencies could not explain the fouling behaviours. Liquid chromatography-organic carbon detection, fluorescence spectrophotometer and high performance liquid chromatography coupled with fluorescence and ultra-violet detectors were used to analyse the DOMs in detail. The major foulants were identified to be biopolymers that were produced in microbial activities. One of the main components of biopolymers e proteins e led to different fouling behaviours. It is postulated that the proteins could pass through porous cake layers to create pore blockages in membranes. Hence, concentrations of the DOMs in the soluble fraction of mixed liquor (SML) could not predict membrane fouling because different components in the DOMs might have different interactions with membranes. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Anaerobic membrane bioreactor Ceramic membranes Dissolved organic matters Membrane fouling Low-strength wastewater Pore size

1. Introduction Anaerobic processes have been widely regarded as a possible way to achieve economic sustainability and energy neutrality in wastewater treatment plants (Liao et al., 2006). Much less waste biomass is produced compared to aerobic processes due to the slow growth rates of anaerobic microorganisms (Stuckey, 2012). Nonetheless, their slow growth rates could create challenges in treating wastewater, especially in start-up periods, due to washout of these slow growing microorganisms (Van Haandal and Lettinga, 1994). Anaerobic membrane bioreactors (AnMBRs) that couple

* Corresponding author. E-mail address: [email protected] (H.Y. Ng).

membranes into the anaerobic processes not only fully retain the anaerobic microorganisms but also produce an effluent with high quality (Stuckey, 2012). As solids-liquid separation is perfect and sedimentation can be eliminated, footprints of the treatment plants are greatly reduced (Kanai et al., 2010). In spite of being an attractive option for wastewater treatment, AnMBRs have not been studied extensively with real domestic wastewater. The biggest hindrances in adopting AnMBRs in domestic wastewater treatment include (1) membrane fouling in AnMBRs is not well understood; and (2) the energy produced from low organic contents may not be able to heat up the anaerobic reactors, especially in cold climate (McCarty et al., 2011). This problem can be exacerbated by dissolved CH4 in the permeate. Polymeric membranes are the most widely used membranes in membrane bioreactors (MBRs), including AnMBRs, due to their low

http://dx.doi.org/10.1016/j.watres.2015.07.038 0043-1354/© 2015 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Yue, X., et al., Effects of dissolved organic matters (DOMs) on membrane fouling in anaerobic ceramic membrane bioreactors (AnCMBRs) treating domestic wastewater, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.07.038

2

X. Yue et al. / Water Research xxx (2015) 1e12

Nomenclature AeCMBR AnCMBR AeMBR AnMBR BW COD DOM FL HPO HPI HPLC HRT LC-OCD LMWs

aerobic membrane bioreactor anaerobic ceramic membrane bioreactor aerobic membrane bioreactor anaerobic membrane bioreactor backflushing water chemical oxygen demand dissolved organic matter fluorescence response hydrophilic organics hydrophilic organics high performance liquid chromatography hydraulic retention time liquid chromatography-organic carbon detection low molecular weight substances

manufacturing cost (Skouteris et al., 2012) and availability of operational experience, particularly for aerobic MBRs (AeMBRs). However, new research fields have increasing demands for chemically and thermally stable membranes (Caro et al., 2000). Thus, inorganic membranes such as ceramic membranes have attracted increasing attention in recent years despite their high upfront capital costs. Lee et al. (2013) found that ceramic membranes have lower fouling propensity than polymeric membranes due to the weaker bonding between foulants and the membranes. In addition to their low fouling propensity, more aggressive cleaning agents can be used to shorten cleaning time required due to their excellent stability and integrity. As a result, the physical and chemical cleaning efficiencies could be increased by 75% and 25%, respectively (Lee and Kim, 2014). Other than these advantages, laborious maintenances such as repair and replacement of polymeric hollow fibres could be eliminated (Lee et al., 2013). All these abovementioned advantages of ceramic membranes suggest that anaerobic ceramic membrane bioreactors (AnCMBRs) may have great potential for wide-scale application if the cost of ceramic membranes can be reduced and when life-cycle cost is being considered. Most of the ceramic membranes used in AnMBRs are tubularshaped owing to their low fouling propensity that was resulted from shear produced by cross flow (Lin et al., 2013). However, as these systems were operated under pressure-driven mode (Herrera-Robledo et al., 2009), there might be a risk of more severe fouling due to shearing of microorganisms by high pressure pumps (Choo and Lee, 1996). This can create greater issues in AnMBRs than in AeMBRs as the slow-growing anaerobic microorganisms may not be able to recover fast enough to compensate for the loss. Moreover, the energy consumption for pressure-driven mode is higher than that for the submerged mode due to higher pressure required (Liao et al., 2006). In this sense, submerged configuration that relies on vacuum suction may be more favourable for AnMBRs. Despite the great potentials of AnMBRs and ceramic membranes, membrane fouling remains as the biggest challenge as it increases operating and capital costs (Martinez-Sosa et al., 2011). Therefore, understanding membrane fouling in AnMBRs is essential in order to find proper controlling methods. In membrane fouling studies, several parameters including operating conditions, feedwater-biomass characteristics and membrane characteristics are considered to have major impacts on membrane fouling (LeClech et al., 2006). Studies on the feedwater-biomass characteristics usually attributed fouling to mixed liquor suspended solids (MLSS) and soluble fraction of mixed liquor (SML). In AeMBRs, conflicting opinions on the effects of MLSS on membrane fouling

MBR MLSS MLVSS MW MWD PSD Rs Rt SEM SML SRT tDOC TN TOC UV220

membrane bioreactor mixed liquor suspended solids mixed liquor volatile suspended solids molecular weight molecular weight distribution particle size distribution COD removal efficiency in mixed liquor supernatant COD removal efficiency in membrane permeate scanning electronic microscopy soluble fraction of mixed liquor solids retention time total dissolved organic carbon total nitrogen total organic carbon UV absorbance at wavelength of 220 nm

have been reported (Brookes et al., 2006; Chang and Kim, 2005; Rosenberger et al., 2006). Likewise, there is no consensus on the impacts of SML on membrane fouling. Several authors reported that SML had negative impacts on the membrane fouling (Bouhabila et al., 2001; Rosenberger et al., 2006), while others could not find this correlation (Drews et al., 2008; Kimura et al., 2009). The disparity might arrive from the fact that some components in SML had larger impacts on membrane fouling than the others (Miyoshi et al., 2012). These two factors were studied to a much lesser extent in AnMBRs. Nonetheless, similar results on the effect of MLSS concentration could be found in available literatures of AnMBRs (Lin et al., 2010; Robles et al., 2013). Most of the studies that investigated the SML in AnMBRs used synthetic feedwater to facilitate the examination of the soluble microbial products (Stuckey, 2012). However, influent organics could not be ignored when real wastewater was used as membrane fouling rates were reported to be closely related to the influent chemical organic demand (COD) levels (Lin et al., 2010). To date, few studies have been reported on the treatment of real domestic wastewater using ceramic membranes in AnMBRs. Species of the dissolved organic matters (DOMs) in the SMLs were not fully understood, while their effects on membrane fouling were seldom reported in AnMBRs. In this study, AnCMBRs were fed with real domestic wastewater to investigate the effect of membrane pore size on DOMs production and its subsequently effects on membrane fouling. While most the studies for fouling mechanisms relied on statistical correlations between fouling rates and concentrations of foulants in SML, foulants were recovered from the cake layers and the pores, and subsequently, characterised in this study. 2. Materials and methods 2.1. Reactors setup Three acrylic AnCMBRs, each with an effective volume of 3.6 L, were set up in parallel (Fig. 1). They were denoted as R80, R200 and R300 for AnCMBRs, in which 80, 200 and 300 nm pore-sized ceramic (Al2O3) membranes were immersed in, respectively. They were operated under ambient temperature of 25e30
C, which was in the mesophilic range. The HRT and SRT were maintained at 7.5 h and 60 d, respectively, in all the three AnCMBRs. The biogas produced was collected by water displacement method and was recirculated to scour the membrane surface through two gas diffusers by a KNF compressor (KNF, N86 KT 18, Germany) in each

Please cite this article in press as: Yue, X., et al., Effects of dissolved organic matters (DOMs) on membrane fouling in anaerobic ceramic membrane bioreactors (AnCMBRs) treating domestic wastewater, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.07.038

X. Yue et al. / Water Research xxx (2015) 1e12

3

Fig. 1. Schematic diagram of an AnCMBR.

AnCMBR at 2 L/min. A peristaltic pump (Masterflex L/S, Cole & Palmer) was used to extract the permeate at a flux of 6 L/m2∙hr (LMH) in each AnCMBR. Trans-membrane pressures (TMPs) were recorded by digital pressure gauges (SMC, ZSE50F) connected between the membranes and the permeate pumps, and were used to assess the membrane fouling rates. The three AnCMBRs were seeded with identical anaerobic sludge (volatile suspended solids was 9600 mg/L) from a digester in a domestic wastewater reclamation plant in Singapore, and fed with primary effluent collected from the same wastewater treatment plant. The primary effluent had a total COD of 330.4 ± 89.8 mg/L, soluble COD of 68.2 ± 47.6 mg/L, suspended solids of 341.1 ± 94.9 mg/L, total nitrogen (TN) of 68.2 ± 10.2 mg/L and NH3eN of 32.7 ± 5.5 mg/L. Steady states were achieved in about 60 days judging from the stabilization of MLSS concentrations, mixed liquor volatile suspended solids (MLVSS) concentrations and COD removal efficiencies (Supplementary Material Figure S-I). All results presented in this paper were based on data collected after this 60-day stabilisation period. 2.2. Sampling methods Samples were taken in the sequence of permeate, mixed liquor in the reactors and influent for analyses on alternate day. The membranes were deemed as completely fouled and taken out from the AnCMBRs when the TMP reached 30 kPa. Cake layers on the fouled membrane surfaces were taken down by physical cleaning e scrubbing with a piece of clean facial sponge and followed by flushing with ultrapure water. The cake layer suspension was kept at 500 mL for each extraction by topping up with ultrapure water. The backflushing water (BW) containing pore-blocking foulants

was obtained by simultaneously sonicating and backflushing the ceramic membranes in 1.5 L of ultrapure water at 8 LMH for 1.5 h. Ninety-seven percent of the membrane permeability was recovered after this process. Hence, the pore-blocking foulants extractions were deemed as successful. Samples were filtered sequentially with 1- and 0.45-mm filter papers (Whatman, USA) to divide the samples into three categories. Those could be retained by 1-mm filter papers were suspended solids while those could pass through 0.45-mm filter papers were soluble products. Those fell between 0.45 and 1 mm were defined as the colloids. 2.3. Analytical methods The MLSS, MLVSS and COD were tested according to the Standard Methods (APHA et al., 2005). TN was measured by a total organic carbon (TOC) analyser (TOC-VCSH Shimadzu, Japan). Particle size distributions (PSDs) of the bioflocs in the mixed liquid were measured by a laser diffraction particle size analyser (Coulter LS230, Beckman Coulter, USA), while PSDs of SML was measured by a Marlvern counter (Zetasizer Nano, UK). Dried specimens of the clean membranes, the membranes after physical cleaning and the membranes with cake layers attachment were sputtered with platinum before being examined under a Scanning Electronic Microscopy (SEM) (JEOL, JSM 5600 LV, Japan). The liquid chromatography-organic carbon detection (LC-OCD) (Dr. Huber, DOC-LAB), which was equipped with a size exclusion column, Toyopearl HW-50S (Tosoh, USA), was used to characterise the DOMs. The proteins and polysaccharides were quantified by the colorimetric methods according to Lowry et al. (1951) and Dubois et al. (1956). The excitation emission matrices (EEM) were obtained by scanning the samples with a fluorescence spectrometer

Please cite this article in press as: Yue, X., et al., Effects of dissolved organic matters (DOMs) on membrane fouling in anaerobic ceramic membrane bioreactors (AnCMBRs) treating domestic wastewater, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.07.038

4

X. Yue et al. / Water Research xxx (2015) 1e12

lower than 0.38 L CH4/g CODutilized e the theoretical CH4 yield calculated from the complete combustion of CH4 at 25
C. It was suspected that a large portion of the CH4 generated was lost in the permeate as dissolved CH4 in the liquid phase, which resulted in supersaturation whereby the dissolved CH4 largely exceeded the theoretical saturating concentration calculated from the Henry's law. In other words, Henry's law could not provide accurate estimation when gas was released from liquid to atmosphere (Pauss et al., 1990). In this study, the dissolved CH4 tests, as discussed in the Materials and Methods section, were performed in order to quantify the following:

(Varian Gary Eclipse, Agilent) at 1200 nm/min. The excitation wavelength was increased at 5-nm interval from 230 to 400 nm, while at each of the excitation wavelength, the emission wavelength was increased at steps of 2 nm from 230 to 500 nm. The excitation and emission slits were 10 nm. The results were plotted to contour maps with a resolution of 25 using a software (Origin 9.0). The dissolved CH4 was measured by a modification of the methods described by Souza et al. (2011). The compositions of headspace gas produced in the dissolved CH4 test and the biogas collected from the AnCMBRs were analysed using a gas chromatograph (GC17A, Shimadzu, Japan). The high performance liquid chromatography (HPLC, Shimadzu Class VP series) was used to characterise the molecular weight distribution (MWD). Fluorescence response and UV absorbance at 220 nm were detected with RF-10AxL and SPD-M10A detectors, respectively. Flowrate of the mobile phased (0.002 M NaH2PO4 þ 0.002 M K2HPO4 þ 0.1 M NaCl) was 1 mL/min. The system was calibrated with polyethylene glycol. The resistance-in-series model proposed by Choo and Lee (1996) was applied to analyse the fouling resistance distributions. The critical fluxes for clean membranes were tested by step-wisely increasing the fluxes at 2 LMH increments every 10 min.

 The degrees of supersaturation (ratios of directly measured dissolved CH4 and the theoretical saturating concentration): They were used to assess the extent of supersaturation. Based on the median methane percentages, the degrees of supersaturation were found to be 2.6, 2.7 and 2.7 for R80, R200 and R300, respectively, which fell in the range of 1.9e6.9 based on the COD balance calculations as summarised by Hartley and Lant (2006).  The true methane yields: COD balance calculation was performed to quantify the true methane yields and confirm the dissolved CH4 measurements. It was assumed that influent COD was distributed into five species in anaerobic process e the COD in the permeate, the COD as CH4 collected, the COD used in sulfate reduction, the COD used in biomass growth and the COD as dissolved CH4. These five species were calculated from direct measurements of relevant parameters in the biogas, permeate and mixed liquors of the three AnCMBRs. Their summation was comparable to the influent COD in each of the three AnCMBRs (refer to Supplementary Material Table S-I), validating the assumption of influent COD distribution and the direct measurements of dissolved CH4. Overall, 61.8, 61.5 and 60.1% of the influent COD was converted to CH4, whereas 67.5, 67.5 and 67.1% of the total CH4 generation was lost in the permeate of R80, R200 and R300, respectively. These figures concurred with Hartley and Lant (2006), who summarized that the dissolved CH4 accounted for 38e85% of the total CH4 generated. The true CH4 yields computed from the CH4 collected from the biogas and dissolved CH4 were found to be 0.3 ± 0.03, 0.3 ± 0.01 and 0.3 ± 0.02 L/g CODutilized for R80, R200 and R300, respectively, which were much closer to the theoretical value.

3. Results and discussion 3.1. Performances of AnCMBRs The MLSS concentrations were 12.8 ± 1.2, 12.9 ± 0.6 and12.8 ± 1.1 g/L for the R80, R200 and R300, respectively. The MLVSS/MLSS ratios were around 0.8 for all the three AnCMBRs. PSDs of bioflocs in the mixed liquor of the three AnCMBRs on day 30 were similar with only slight differences in median particle sizes e 43.7 mm for both the R80 and R300, and 39.8 mm for the R200 (Supplementary Material Figure S-II). The treatment performances of the three AnCMBRs are summarised in Table 1. The COD removal efficiencies in mixed liquor (Rs) were 78.6 ± 6.0, 78.7 ± 6.9 and 79.7 ± 5.5%, while the total COD removal efficiencies (Rt) (i.e., COD removal by mixed liquor and membrane filtration) were 88.6 ± 9.0, 87.9 ± 7.4 and 86.3 ± 9.7% for R80, R200 and R300, respectively. The Rs and Rt found in this study were lower than those reported by Huang et al. (2011), who operated an AnMBR at similar SRT, HRT and flux using a synthetic wastewater. This might be attributed to the higher complexity and lower biodegradability of the organics in the real domestic wastewater used in this study as compared to the synthetic wastewater. Similar to other anaerobic processes, there was no ammonia removal in the AnCMRBs. The ammonia concentrations were slightly higher in the mixed liquor than in the influent due to protein destruction (Moen et al., 2003). CH4 contents in the biogas were around 60e65% in the three AnCMBRs. Similar amount of CH4 were collected daily for the R80, R200 and R300 (i.e., 0.3 ± 0.03, 0.3 ± 0.01 and 0.3 ± 0.02 L, respectively), which were translated into CH4 yields of 0.1 ± 0.02 L CH4/g CODutilized in the three AnCMBRs. These values were much

It could be concluded that system performances in the three AnCMBRs were of no significant difference. Hence, the MLSS concentrations and the operating conditions that were commonly reported to affect membrane fouling (Le-Clech et al., 2006) were ruled out in explaining the fouling behaviours in this study. 3.2. Relationship between fouling rates and membrane pore sizes The TMP profile in each AnCMBR went through a plateau stage before a sudden surge as shown in Fig. 2. The two stages were more prominent in the AnCMBRs mounted with larger pore-sized

Table 1 Treatment performances of the three AnCMBRs. R80 Rs (%)a Rt (%)b Methane production (L/d) Specific methane production (mL/g MLVSS$d1) Methane yield (L CH4/g CODutilised) a b

78.6 88.6 0.3 30.4 0.1

R200 ± ± ± ± ±

6.0 9.0 0.03 0.2 0.02

78.7 87.9 0.3 30.4 0.1

± ± ± ± ±

R300 6.9 7.4 0.01 0.3 0.02

79.7 86.3 0.3 30.7 0.1

± ± ± ± ±

5.5 9.7 0.02 0.2 0.02

Rs: COD removal efficiency in mixed liquor supernatant. Rt: COD removal efficiency in membrane permeate.

Please cite this article in press as: Yue, X., et al., Effects of dissolved organic matters (DOMs) on membrane fouling in anaerobic ceramic membrane bioreactors (AnCMBRs) treating domestic wastewater, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.07.038

X. Yue et al. / Water Research xxx (2015) 1e12

5

Fig. 2. TMP profiles of the three AnCMBRs.

membranes. In other words, the fouling rates for smaller pore-sized membranes were more incremental while that for larger poresized membranes were more abrupt. Hence, the foulants governing the fouling behaviours might vary for membranes with different pore sizes (Ognier et al., 2002). Despite the average TMPs at the start of the operation were 4.2, 1.9 and 0.3 kPa, it took an average of 25.3, 12.5 and 9.5 days for the R80, R200 and R300, respectively, to reach a TMP of 30 kPa. This implied that membranes with smaller pore sizes fouled slower. These findings were in accordance with He et al. (2005) who observed that membranes with the smallest molecular weight cutoff (MWCO) had highest permeability loss within the first 15 min but lower fouling rate in extended operating duration. They were also in line with those obtained in aerobic ceramic membrane bioreactors (AeCMBRs) where the same membranes were used as in this study (Jin et al., 2010). Nonetheless, membranes fouled much faster in AnCMBRs than in AeCMBRs when similar conditions were used. The critical fluxes were found to be 18.1, 17.9 and 13.5 LMH for R80, R200 and R300, respectively. Despite the debates over critical flux being used in predicting long-term fouling behaviours in membrane systems filtering complex suspensions, the critical fluxes obtained in this study corroborated with the long-term fouling rates of the three AnCMBRs, which implied that the parameters determining the critical fluxes might also contribute to different long-term fouling behaviours. Since critical fluxes largely depended on the foulants that could irreversibly foul the membranes, the determining factors of critical fluxes might be pore narrowing and blockage that were more irreversible than cake layer fouling (Le Clech et al., 2003). Although Kang et al. (2002) suggested that inorganic membranes were more prone to

inorganic fouling, primarily struvite (MgNH4PO4$6H2O), inorganic fouling was unnoticeable in this study since the Ca2þ and Mg2þ concentrations did not differ much among the feedwater, the SML and the permeate. The discrepancy between this observation and the literature might be due to the low ammonia levels in domestic wastewater used in this study. Therefore, only organic fouling is discussed in this paper. PSDs of the SML, tested on day 30, revealed that the percentages of DOMs with sizes smaller than the membrane pores were 5.1, 40.1 and 60.2% for R80, R200 and R300, respectively (Supplementary Material Figure S-III). These DOMs could still pass through the cake layers and cause internal fouling (Vyas et al., 2002). Consequently, there were increased risks of pore blockages for larger pore-sized membranes. Prolonged pore blockages in larger poresized membranes significantly reduced permeability of the membranes and led to earlier TMP surges. Meng et al. (2009) also proposed that pore blockage was responsible for membrane fouling in long-term operation. As DOMs could be adsorbed onto cake layers as well as into membrane pores, examination of the membrane surfaces would be useful to verify their interactions with the membranes and the cake layers. SEM images of the virgin membranes revealed that the larger pore-sized membranes were rougher. Notwithstanding the fact that many studies attributed faster fouling rates to the rougher membrane surfaces due to more deposits (Evans et al., 2008), thicker cake layer were present on the smaller pore-sized membranes with smoother surfaces in this study. The dry weights of cake layers on the 80-, 200- and 300-nm membranes were 53.1, 22.1 and 12.8 mg/m2, respectively (Table 2). In addition, denser cake layers were observed on the smoother and smaller pore-sized

Please cite this article in press as: Yue, X., et al., Effects of dissolved organic matters (DOMs) on membrane fouling in anaerobic ceramic membrane bioreactors (AnCMBRs) treating domestic wastewater, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.07.038

6

X. Yue et al. / Water Research xxx (2015) 1e12

Table 2 Compositions of suspended solids, colloids and soluble substances in the total solids of the cake layers.

Total solids (g/m2) Suspended Solids (g/m2) Colloids (g/m2) Soluble Substances (g/m2)

R80 (%)

R200 (%)

R300 (%)

53.1 48.0 2.9 2.2

22.1 20.0 1.2 0.9

12.8 11.7 0.6 0.5

(100) (90.4) (5.4) (4.2)

(100) (90.6) (5.3) (4.1)

(100) (91.4) (4.7) (3.9)

membranes that fouled slower (Fig. 3). The colloids and DOMs accounted for 9.6, 9.4 and 8.6% of the total solids in the cake layers for R80, R200 and R300, respectively (Table 2). Higher proportions of the colloids and DOMs might help in forming denser cake layers by binding the large particles together on smaller pore-sized membrane surfaces. Thicker cake layers might be beneficial for membrane filtration as they were able to retain the solutes and prevent them from blocking the membrane pores. Examining the membrane surfaces after physical cleaning, it was observed that a thin layer of solutes covered the 80-nm membrane while the porous areas of 200- and 300-nm membranes were filled with solutes (Fig. 3). It indicated that pore blockages were prominent in larger pore-sized membranes. As a consequence, the larger poresized membranes fouled faster as pore blockages were more irreversible compared to cake layer fouling. 3.3. Effects of DOMs on membrane fouling 3.3.1. Total dissolved organic carbon (tDOC) In view of the importance of DOMs, the tDOC concentrations in SML were obtained. It was found that 48.9 ± 4.5, 47.8 ± 6.2 and 46.1 ± 7.8% of tDOC was retained by the 80-, 200- and 300-nm membranes, respectively, showing a somewhat decreasing trend with increasing membrane pore size. In order to elucidate the interactions of organics with the cake layers and membranes, cake layer soluble products and BW were also analysed. The tDOC concentrations of the cake layer soluble products based on the 500-mL suspensions were 38.1, 26.5 and 16.3 mg/L, while those in the 1.2 L BW solutions were 4.1, 5.1 and 7.8 mg/L for R80, R200 and R300, respectively. This suggested that the denser cake layers further decreased the chances of internal fouling caused by DOMs with sizes smaller than membranes pores. As the tDOC removal efficiencies across the membranes could not explain the fouling rates, it was speculated that different components of the tDOC might contribute to fouling in different ways. Hence, LC-OCD was deployed to further analyse the tDOC and the results are presented in Table 3. The hydrophilic organics (HPI) could be eluted from the LC-OCD column, while the hydrophobic organics (HPO) were bound with the column. Therefore, concentration of the HPO could be calculated by subtracting those of the HPI from the tDOC. It could be observed from Table 3 that the concentrations of HPO increased in the permeate compared to the SML, which might be resulted from microbial degradation of organics in the cake layers (Yamamura et al., 2014). HPO were considered as the main foulants in many early membrane fouling studies, especially when surface water was used as feedwater (Chang et al., 1999; Madaeni et al., 1999; Tian et al., 2013b; Yuan and Zydney, 1999). Nonetheless, there have been conflicting opinions in recent years. Considerable attention has been drawn to HPI and many researchers concluded that HPI, rather than HPO, contributed to faster fouling rates and more irreversible fouling (Fang and Shi, 2005; Kimura et al., 2014; Tian et al., 2013a). The discrepancy might be attributed to different feedwater used. Ma et al. (2001) demonstrated that the HPO comprised more than 50% of tDOC in surface water, while their

fractions were less than 30% in wastewater treatment plant effluent. In this study, HPI removals were entirely responsible for tDOC removals across the membranes based on the formula suggested by Rosenberger et al. (2006). In other words, HPO might not contribute to the membrane fouling. More specifically, Yamamura et al. (2014) concluded that the HPO did not contribute to internally fouling. The HPI could be further categorised according to different peaks in LC-OCD chromatographs (an example of the chromatograph is shown in Fig. 4). The first peak that appeared at 29 min was related to biopolymers. Occasionally, this peak was absent in UV chromatograph due to absence of proteins e one of the main components of biopolymers (Rosenberger et al., 2006). The second peak that was eluted at around 45 min represented humic substances (HS), while the third peak right after it at 47 min represented building blocks that were produced in the breakdown of HS. All other peaks thereafter were associated with low molecular weight acids, HS and neutrals (Simon et al., 2012). Biopolymers had molecular weights (MWs) no smaller than 10 kDa, which were much larger than the rest of low molecular weight substances (LMWs) that had MWs smaller than 1 kDa. Hence, LMWs and biopolymers are discussed separately in Sections 3.3.2 and Section 3.3.3, respectively. In these sections, some techniques were used in conjunction with LC-OCD. UV absorbance at wavelength of 220 nm (UV220) could be detected for proteins and organic acids (Gray et al., 2008). EEM was also regarded as a rapid and reliable way to characterise protein- and humic-like substances (Hudson et al., 2007). Using HPLC together with an UV detector at a wavelength of 220 nm and a detector for fluorescence response (FL) could elucidate MWDs for the substances of interest (Her et al., 2003). 3.3.2. Low molecular weight substances Retention efficiencies of LMWs were 40.7, 35.8 and 33.9% across the 80-, 200- and 300-nm membranes, respectively. Nonetheless, their retentions in the cake layers and in the pores were different. It was obvious that the percentages of LMWs formed a trend of R80 < R200 < R300 in cake layer soluble products (Table 3). Considering the morphologies of cake layers presented in Fig. 3, it might be speculated that the more porous cake layers on larger pore-sized membranes had larger surface areas to adsorb more LMWs. In other words, adsorption of LMWs might be related to porosity. On the other hand, the trend of LMWs' accumulations in membrane pores showed opposite trend to that in the cake layers. Since their percentages were higher for smaller pore-sized membranes that fouled slower, LMWs might not contribute to internal fouling. According to classification of the HPI in LC-OCD, LMWs include a large proportion of fulvic and humic acids. Two peaks in EEM contour maps were related to these organics (Fig. 5). Peak B that was associated with fulvic acid-like substances was only found in the permeate of R80, R200 and R300 at Ex/Em of 284e292/ 405e420, 284e286/405e415 and 270e286/360e415 nm, respectively. This peak probably existed in all samples but was covered by a nearby peak that was related to tryptophan- and protein-like substances (discussed in Section 3.3.3). They were only revealed when the concentrations of tryptophan- and protein-like substances were greatly reduced. Due to the coverage, information on fulvic acid-like substances could not be obtained. Another peak e peak C e was observed in the SML of R80, R200 and R300 in the range of (Ex/Em) 325e335/410e425, 318e346/400e438 and 322e342/405e417 nm, respectively. This peak was associated with hydrophobic acid-as well as humic acid-like substances (Chen et al., 2003). Locations and intensities of peak C did not show significant differences between the soluble products of feedwater and the SML (Supplementary Material Table S-II), implying that

Please cite this article in press as: Yue, X., et al., Effects of dissolved organic matters (DOMs) on membrane fouling in anaerobic ceramic membrane bioreactors (AnCMBRs) treating domestic wastewater, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.07.038

7

Fig. 3. SEM scanning images of the virgin membranes, the membrane surfaces after physical cleaning and the cake layers.

X. Yue et al. / Water Research xxx (2015) 1e12

Please cite this article in press as: Yue, X., et al., Effects of dissolved organic matters (DOMs) on membrane fouling in anaerobic ceramic membrane bioreactors (AnCMBRs) treating domestic wastewater, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.07.038

8

X. Yue et al. / Water Research xxx (2015) 1e12

Table 3 Concentrations and percentages of different fractions in the DOMs.

Fig. 4. An example of organic carbon (OC) and ultraviolet (UV) chromatographs in the SML and the permeate (R80).

Please cite this article in press as: Yue, X., et al., Effects of dissolved organic matters (DOMs) on membrane fouling in anaerobic ceramic membrane bioreactors (AnCMBRs) treating domestic wastewater, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.07.038

X. Yue et al. / Water Research xxx (2015) 1e12

9

Fig. 5. EEM contour maps of soluble products in the feedwater and the three AnCMBRs' SML, permeate, cake layer soluble products and BW.

the hydrophobic acid- and humic acid-like substances experienced minimum transformation in the mixed liquor. This was expected as these organics were often regarded as non-biodegradable (Lee, 2010). Peak C was also present in the permeate of R80, R200 and R300 at Ex/Em of 322e342/408e435, 320e342/405e438 and 322e342/398e435 nm, respectively. In most of the cake layer soluble products and BW samples, peak C was missing. The HPLCFL chromatographs showed that majority of DOMs that were associated with peak B and peak C eluted at retention times between 10 and 13 min, corresponding to MW lower than 3 kDa

(data not shown). This indicated that fulvic acid-, hydrophobic acid- and humic acid-like substances had low MWs so that they could pass through the cake layers and membrane pores. This seemed to be contradictory to the LC-OCD results, which showed that LMWs were still present in the cake layers and membrane pores. However, it must be noticed that most of the DOMs associated with peak B and peak C were hydrophobic (Yamamura et al., 2014). Thus, the EEM results could not be correlated with the concentrations of LMWs obtained in the LC-OCD. In conclusion, LMWs were related to morphologies of the surfaces they attached

Please cite this article in press as: Yue, X., et al., Effects of dissolved organic matters (DOMs) on membrane fouling in anaerobic ceramic membrane bioreactors (AnCMBRs) treating domestic wastewater, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.07.038

10

X. Yue et al. / Water Research xxx (2015) 1e12

to and they might not contribute to the internal fouling, possibly due to their low MWs (Drews et al., 2005). 3.3.3. Biopolymers Looking at different HPI components presented in Table 3, biopolymers were removed by 92.6, 91.7 and 90.4% across the 80-, 200- and 300-nm membranes, respectively. Hence, biopolymers might be the main contributor to membrane fouling. As a matter of fact, they were regarded as the main foulants in MBR and UF/NF filtration of secondary effluent in some studies (Zheng et al., 2010). Polysaccharides and proteins were generally considered as two main components of biopolymers (Huber et al., 2011). Hence they were quantified and the results are summarised in Table 4. Retention efficiencies of polysaccharides were 77.3 ± 7.1, 77.8 ± 6.6 and 77.4 ± 5.7% for the 80-, 200- and 300-nm membranes, respectively, showing no obvious trend. On the other hand, the 80-, 200- and 300-nm membranes could retain 51.5 ± 2.1, 45.5 ± 3.2 and 37.7 ± 3.8% of proteins in the SML. Different retention efficiencies of proteins and polysaccharides resulted in higher proteins/polysaccharides ratios in the permeate compared to those in the SML as shown in Table 4. The biopolymers had higher concentrations in the cake layer soluble products for smaller pore-sized membranes and in the BW of larger pore-sized membranes. Proteins/polysaccharides ratios were 1.37, 1.35 and 1.33 in the cake layer soluble products while those were 4.45, 4.91 and 5.27 in the BW for the R80, R200 and R300, respectively. The proteins/polysaccharides ratios in the cake layer soluble products were slightly lower than the average values in the SML, indicating higher retentions of polysaccharides than proteins in the cake layers. Due to retentions in the cake layers, the effects of polysaccharides on pore blockages were reduced. As a result, the proteins/polysaccharides ratios in the BW were much higher than those in the SML, showing accumulations of proteins in the membrane pores. These observations suggested that the proteins might lead to pore blockages and hence, faster long-term fouling rates. This was in accordance to Hernandez Rojas et al. (2005), who observed that specific resistances increased by 10 times when the concentrations of proteins in the SML were increased from 30 to 100 mg/L Yao et al. (2010) also showed that membrane fouling rates accelerated when proteins/polysaccharides ratios were increased. On the contrary, polysaccharides were found to be the main foulants in MBR in some studies (Chu and Li, 2005; Taimur Khan et al., 2013). However, in these studies, there were clear correlations between cake layer fouling and the concentrations of polysaccharides, which confirmed the importance of polysaccharides in cake layer formations. Different distributions of polysaccharides and proteins in the cake layers and BW might be due to the properties of polysaccharides and proteins such as their structures. Polysaccharides, for instance alginate e a common type of polysaccharides present in the water environment

and are used as surrogates for other polysaccharides in membrane studies, have a chain structure (Mirshafiey et al., 2005) that might make them prone to be entangled in cake layers. While we could not assume that all proteins had similar structures, elsewhere it was demonstrated that proteins, such as bovine serum albumin had spherical shape (Jachimska et al., 2008) that enabled them to reach membrane surfaces after passing through the porous cake layers. Hence, they could contribute more to pore blockages. In EEM contour maps, the peak that was associated with tryptophan- and protein-like substances was denoted as peak A. Peak A appeared at Ex/Em of 275e280/365e375, 278e282/358e380 and 275e282/358e375 nm for the SML of the R80, R200 and R300, respectively. The tryptophan- and protein-like substances were produced in biological degradation of organics as peak A was not detected in the soluble products of the feedwater (Fig. 5). Despite peak A was absent in the permeate of AnCMBRs, a shoulder originated from peak B reached out to usual location of peak A, indicating the existence of low concentrations of tryptophan- and protein-like substances (Liu et al., 2011). In addition, the intrusion of this shoulder was more noticeable in the permeate of the AnCMBRs mounted with larger pore-sized membranes, which implied the existence of higher concentrations of tryptophan- and protein-like substances. Peak A in the cake layer soluble products was located at 280e286/328e342, 274e278/358e378, 272e274/ 355e370 nm for the R80, R200 and R300, respectively. Intensities of this peak formed a trend of R80 > R200 > R300 (Supplementary Material Table S-II), implying higher retentions of tryptophan- and protein-like substances in the cake layers on smaller pore-sized membranes. In the BW, peak A was located at 272e276/365e385, 268e271/360e390 and 269e275/365e385 nm for the R80, R200 and R300, respectively. The intensities of peak A were higher for larger pore-sized membranes (Supporting Material Table S-II), indicating more proteins were trapped in the pores in these membranes. In order to confirm the MWs of the proteins associated with peak A, HPLC-UV220 and HPLC-FL were deployed. The proteins were eluted at around 7 min in HPLC-UV220 chromatographs (Fig. 6) and corresponded to MW of 117 kDa. There were significant reductions for this peak in the permeate compared to the SML for the three AnCMBRs. Similar peaks could also be observed in the cake layer soluble products and the BW. Intensities of the peaks at 7 min displayed a trend of R80 > R200 > R300 in the cake layer soluble products. The proteins that passed through cake layers led to a trend of R80 < R200 < R300 in the BW for the peak at 7 min. These results further illustrated the pore blockages initiated by proteins. Similar to the HPLC-UV220 results, the most significant difference between the permeate and SML was still observed at the peak of 7 min. The intensities of these peaks in HPLC-FL results followed the same trends as those in the HPLC-UV220 results for the cake layer soluble products and BW, which indicated that the HPLC-FL

Table 4 Concentrations of proteins and polysaccharides, and the proteins/polysaccharides ratios in the SML, permeate, BW and cake layer soluble products.

SML

Permeate

BW

Cake layer soluble products

Proteins (mg/L) Polysaccharides (mg/L) Proteins/polysaccharides Proteins (mg/L) Polysaccharides (mg/L) Proteins/polysaccharides Proteins (mg/m2) Polysaccharides (mg/m2) Proteins/polysaccharides Proteins (mg/m2) Polysaccharides (mg/m2) Proteins/polysaccharides

R80

R200

R300

19.8 ± 4.9 13.0 ± 4.1 1.6 ± 0.5 9.6 ± 4.3 3.0 ± 1.9 4.2 ± 2.5 85.4 19.2 4.45 205.4 149.7 1.37

19.1 ± 6.1 13.2 ± 4.0 1.5 ± 0.5 10.4 ± 6.6 2.9 ± 1.1 3.6 ± 1.6 102.3 20.9 4.91 54.2 40.2 1.35

17.1 ± 4.7 11.1 ± 3.6 1.6 ± 0.4 10.7 ± 6.4 2.3 ± 0.8 4.8 ± 2.2 119.4 22.7 5.27 42.3 31.9 1.33

Please cite this article in press as: Yue, X., et al., Effects of dissolved organic matters (DOMs) on membrane fouling in anaerobic ceramic membrane bioreactors (AnCMBRs) treating domestic wastewater, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.07.038

X. Yue et al. / Water Research xxx (2015) 1e12

11

Fig. 6. Results of the HPLC-UV220 for the SML, permeate, BW and cake layer soluble products, and results of the HPLC-FL for the BW and cake layer soluble products at peak A locations identified from EEM.

and the HPLC-UV220 were useful in identifying the proteins in the samples. Both results also confirmed that the denser cake layers on smaller pore-sized membranes could reject more proteins and reduce the risks of pore blockages. In addition, they evidenced that the main pore-blocking foulants were proteins with MW about 117 kDa. In contrast to the results for LMWs presented in Section 3.3.2, the HPLC results were able to correlate with the LC-OCD results, implying the main foulants were HPI.

4. Conclusions In this study, 80-, 200- and 300-nm ceramic membranes were used in AnCMBRs for the treatment of real domestic wastewater. The following conclusions could be drawn:  AnCMBRs were successfully applied in the treatment of real domestic wastewater with high COD removal efficiencies. However, high dissolved CH4 in the liquid phase was observed. In order to minimise energy wastage, dissolved CH4 has to be recovered.  In long-term operation, higher fouling rates were observed in larger pore-sized membranes, which was attributed to higher occurrences of pore blockages. The cake layers did not lead to higher fouling propensity. In contrast, they could control fouling by stopping more DOMs from blocking the pores.  Pore blockages by hydrophilic DOMs were responsible for longterm membrane fouling. With the help of the LC-OCD, EEM and HPLC, biopolymers that had similar sizes with membrane pores were found to be major pore-blocking foulants. Two major components of biopolymers e proteins and polysaccharides e contributed to membrane fouling in different ways. Polysaccharides were more prone to be trapped in the cake layers while proteins were more likely to be responsible for pore blockage that resulted in the final TMP jump. Hence, concentrations of DOMs alone could not explain the fouling behaviours

as different components in the DOMs contributed to membrane fouling in different ways. Acknowledgement The author is supported by the National Research Foundation Singapore under its National Research Foundation (NRF) Environmental and Water Technologies (EWT) Ph.D. Scholarship Programme and administered by the Environment and Water Industry Programme Office (EWI). Appendix A. Supplementary material Supplementary material related to this article can be found at http://dx.doi.org/10.1016/j.watres.2015.07.038. References APHA, AWWA, WEF, 2005. Standard Methods for the Examination for Water and Wastewater, 21st ed. Washington DC, USA. Bouhabila, E.H., Ben Aïm, R., Buisson, H., 2001. Fouling characterisation in membrane bioreactors. Sep. Purif. Technol. 22e23, 123e132. Brookes, A., Jefferson, B., Guglielmi, G., Judd, S.J., 2006. Sustainable flux fouling in a membrane bioreactor: impact of flux and MLSS. Sep. Sci. Technol. 41, 219e228. €lsch, P., Scha €fer, R., 2000. Zeolite membranes e state of their Caro, J., Noack, M., Ko development and perspective. Microporous Mesoporous Mater. 38 (1), 3e22. Chang, I.S., Kim, S.N., 2005. Wastewater treatment using membrane filtration e effect of biosolids concentration on cake resistance. Process Chem. 40 (3e4), 1307e1314. Chang, I.S., Lee, C.H., Ahn, K.H., 1999. Membrane filtration characteristics in membrane-coupled activated sludge system: the effect of floc structure on membrane fouling. Sep. Sci. Technol. 34 (9), 1743e1758. Chen, W., Westerhoff, P., Leenheer, J.A., Booksh, K., 2003. Fluorescence excitation e emission matrix regional integration to quantify spectra for dissolved organic matters. Environ. Sci. Technol. 37, 1570e1579. Choo, K.-H., Lee, C.-H., 1996. Membrane fouling mechanisms in the membranecoupled anaerobic bioreactor. Water Res. 30 (8), 1771e1780. Chu, H.P., Li, X.-Y., 2005. Membrane fouling in a membrane bioreactor (MBR): sludge cake formation and fouling characteristics. Biotechnol. Bioeng. 90 (3), 323e321. Drews, A., Vocks, M., Bracklow, U., Iversen, V., Kraume, M., 2008. Does fouling in MBRs depend on SMP? Desalination 231 (1e3), 141e149.

Please cite this article in press as: Yue, X., et al., Effects of dissolved organic matters (DOMs) on membrane fouling in anaerobic ceramic membrane bioreactors (AnCMBRs) treating domestic wastewater, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.07.038

12

X. Yue et al. / Water Research xxx (2015) 1e12

Drews, A., Vocks, M., Iverson, V., Lesjean, B., Kraume, K., 2005. Influence of Unsteady Membrane Bioreactor Operation on EPS Formation and Filtrtaion Resistance. International Congress on Mmebranes and Membrane Processes (ICOM), Seoul, South Korea. Dubois, M., Gilles, K.A., Hamiltin, J.K., Rebers, P.A., Smith, F., 1956. Colorimetric method for determination of sugars and related substances. Anal. Chem. 28 (3), 350e356. €ki, A., Nystro € m, M., 2008. The influence of hydroEvans, P.J., Bird, M.R., Pihlajama phobicity, roughness and charge upon ultrafiltration membranes for black tea liquor clarification. J. Membr. Sci. 313 (1e2), 250e262. Fang, H.H.P., Shi, X., 2005. Pore fouling of microfiltration membranes by activated sludge. J. Membr. Sci. 264 (1e2), 161e166. Gray, S.R., Ritchie, C.B., Tran, T., Bolto, B.A., Greenwood, P., Busetti, F., Allpike, B., 2008. Effect of membrane character and solution chemistry on microfiltration performance. Water Res. 42 (3), 743e753. Hartley, K., Lant, P., 2006. Eliminating non-renewable CO2 emissions from sewage treatment: an anaerobic migrating bed reactor pilot plant study. Biotechnol. Bioeng. 95 (3), 384e398. He, Y., Xu, P., Li, C., Zhang, B., 2005. High-concentration food wastewater treatment by an anaerobic membrane bioreactor. Water Res. 39 (17), 4110e4118. Her, N., Amy, G., McKnight, D., Sohn, J., Yoon, Y., 2003. Characterization of DOM as a function of MW by fluorescence EEM and HPLC-SEC using UVA, DOC, and fluorescence detection. Water Res. 37 (17), 4295e4303. Hernandez Rojas, M.E., van Kaam, R., Schetrite, S., Albsi, C., 2005. Role and variations of supernatant compounds in submerged membrane bioreactor fouling. Desalination 179, 95e107. Herrera-Robledo, M., Morgan-Sagastume, J.M., Noyola, A., 2009. Biofouling and pollutant removal during long-term operation of an anaerobic membrane bioreactor treating municipal wastewater. Biofouling 26 (1), 23e30. Huang, Z., Ong, S.L., Ng, H.Y., 2011. Submerged anaerobic membrane bioreactor for low-strength wastewater treatment: effect of HRT and SRT on treatment performance and membrane fouling. Water Res. 45 (2), 705e703. Huber, S.A., Balz, A., Abert, M., Pronk, W., 2011. Characterisation of aquatic humic and non-humic matter with size-exclusion chromatography-organic carbon detection-organic nitrogen detection (LC-OCD-OND). Water Res. 45 (2), 879e885. Hudson, N., Baker, A., Reynolds, D., 2007. Fluorescence analysis of dissolved organic matter in natural, waste and polluted waters - a review. River Res. Appl. 23 (6), 631e649. Jachimska, B., Wasilewska, M., Adamczyk, Z., 2008. Characterization of globular protein solutions by dynamic light scattering, electrophoretic mobility, and viscosity measurements. Langmuir 24 (13), 6866e6872. Jin, L., Ong, S.L., Ng, H.Y., 2010. Comparison of fouling characteristics in different pore-sized submerged ceramic membrane bioreactors. Water Res. 44 (20), 5907e5918. Kanai, M., Ferre, V., Wakahara, S., Yamamoto, T., Moro, M., 2010. A novel combination of methane fermentation and MBR d Kubota Submerged Anaerobic Membrane Bioreactor process. Desalination 250 (3), 964e967. Kang, I.-J., Yoon, S.-H., Lee, C.-H., 2002. Comparison of the filtration characteristics of organic and inorganic membranes in a membrane-coupled anaerobic bioreactor. Water Res. 36 (7), 1803e1813. Kimura, K., Naruse, T., Watanabe, Y., 2009. Changes in characteristics of soluble microbial products in membrane bioreactors associated with different solid retention times: Relation to membrane fouling. Water Res. 43 (4), 1033e1039. Kimura, K., Tanaka, K., Watanabe, Y., 2014. Microfiltration of different surface waters with/without coagulation: Clear correlations between membrane fouling and hydrophilic biopolymers. Water Res. 49, 434e443. Le-Clech, P., Chen, V., Fane, T.A.G., 2006. Fouling in membrane bioreactors used in wastewater treatment. J. Membr. Sci. 284 (1e2), 17e53. Le Clech, P., Jefferson, B., Chang, I.S., Judd, S.J., 2003. Critical flux determination by the flux-step method in a submerged membrane bioreactor. J. Membr. Sci. 227 (1e2), 81e89. Lee, S.-J., Dilaver, M., Park, P.-K., Kim, J.-H., 2013. Comparative analysis of fouling characteristics of ceramic and polymeric microfiltration membranes using filtration models. J. Membr. Sci. 432 (0), 97e105. Lee, S.J., Kim, J.H., 2014. Differential natural organic matter fouling of ceramic versus polymeric ultrafiltration membranes. Water Res. 48, 43e51. Lee, Y.H., 2010. Archaeal diversity during composting of pig manure and mushroom cultural waste based on 16S rRNA sequence. J. Korean Soc. Appl. Biol. Chem. 53 (2), 230e236. Liao, B.-Q., Kraemer, J.T., Bagley, D.M., 2006. Anaerobic membrane bioreactors: applications and research directions. Crit. Rev. Environ. Sci. Technol. 36 (6), 489e530. 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, 169e188. Lin, H.J., Xie, K., Mahendran, B., Bagley, D.M., Leung, K.T., Liss, S.N., Liao, B.Q., 2010. Factors affecting sludge cake formation in a submerged anaerobic membrane

bioreactor. J. Membr. Sci. 361 (1e2), 126e134. Liu, T., Chen, Z.-l., Yu, W.-z., You, S.-j, 2011. Characterization of organic membrane foulants in a submerged membrane bioreactor with pre-ozonation using threedimensional excitationeemission matrix fluorescence spectroscopy. Water Res. 45 (5), 2111e2122. Lowry, O., Roseburge, N., Farr, A., Randall, R., 1951. Protein measurements with the folin phenol reagent. J. Biol. Chem. 193 (1), 265e276. Ma, H., Allen, H.E., Yin, Y., 2001. Characterization of isolated fractions of dissolved organic matter from natural waters and a wastewater effluent. Water Res. 35 (4), 985e994. Madaeni, S.S., Fane, A.G., Wiley, D.E., 1999. Factors influencing critical flux in membrane filtration of activated sludge. J. Chem. Technol. Biotechnol. 74 (6), 539e543. Martinez-Sosa, D., Helmreich, B., Netter, T., Paris, S., Bischof, F., Horn, H., 2011. Pilotscale anaerobic submerged membrane bioreactor (AnSMBR) treating municipal wastewater: the fouling phenomenon and long-term operation. Water Sci. Technol. 64 (9), 1804e1811. McCarty, P.L., Bae, J., Kim, J., 2011. Domestic wastewater treatment as a net energy producerecan this be achieved? Environ. Sci. Technol. 45 (17), 7100e7106. 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 (6), 1489e1512. Mirshafiey, A., Khodadadi, A., Rehm, B.H., Khorramizadeh, M.R., Eslami, M.B., Razavi, A., Saadat, F., 2005. Sodium Alginate as a novel therapeutic option in experimental colitis. Scand. J. Immunol. 61 (4), 316e321. Miyoshi, T., Aizawa, T., Kimura, K., Watanabe, Y., 2012. Identification of proteins involved in membrane fouling in membrane bioreactors (MBRs) treating municipal wastewater. Int. Biodeterior. Biodegrad. 75, 15e22. €, R., Ferguson, J.F., 2003. Effect of solids retention Moen, G., Stensel, H.D., Lepisto time on the performance of Thermophilic and mesophilic digestion of combined municipal wastewater sludges. Water Environ. Res. 75 (6), 539e548. Ognier, S., Wisniewski, C., Grasmick, A., 2002. Influence of macromolecule adsorption during filtration of a membrane bioreactor mixed liquor suspension. J. Membr. Sci. 209 (1), 27e37. Pauss, A., Andre, G., Perrier, M., Guiot, S.R., 1990. Liquid-to-gas mass transfer in anaerobic processes: inevitable transfer limitations of methane and hydrogen in the biomethanation process. Appl. Environ. Microbiol. 56 (6), 1636e1644. Robles, A., Ruano, M.V., Ribes, J., Ferrer, J., 2013. Factors that affect the permeability of commercial hollow-fibre membranes in a submerged anaerobic MBR (HFSAnMBR) system. Water Res. 47 (3), 1277e1288. Rosenberger, S., Laabs, C., Lesjean, B., Gnirss, R., Amy, G., Jekel, M., Schrotter, J.C., 2006. Impact of colloidal and soluble organic material on membrane performance in membrane bioreactors for municipal wastewater treatment. Water Res. 40 (4), 710e720. Simon, F.X., Penru, Y., Guastalli, A.R., Esplugas, S., Llorens, J., Baig, S., 2012. NOM characterization by LC-OCD in a SWRO desalination line. Desalination Water Treat. 51 (7e9), 1776e1780.  2012. Anaerobic  pez, P., Negro, C., Blanco, A., Skouteris, G., Hermosilla, D., Lo membrane bioreactors for wastewater treatment: a review. Chem. Eng. J. 198e199, 138e148. Souza, C.L., Chernicharo, C.A., Aquino, S.F., 2011. Quantification of dissolved methane in UASB reactors treating domestic wastewater under different operating conditions. Water Sci. Technol. 64 (11), 2259e2264. Stuckey, D.C., 2012. Recent developments in anaerobic membrane reactors. Bioresour. Technol. 122, 137e148. Taimur Khan, M.M., Takizawa, S., Lewandowski, Z., Habibur Rahman, M., Komatsu, K., Nelson, S.E., Kurisu, F., Camper, A.K., Katayama, H., Ohgaki, S., 2013. Combined effects of EPS and HRT enhanced biofouling on a submerged and hybrid PAC-MF membrane bioreactor. Water Res. 47 (2), 747e757. Tian, J.Y., Ernst, M., Cui, F., Jekel, M., 2013a. Correlations of relevant membrane foulants with UF membrane fouling in different waters. Water Res. 47 (3), 1218e1228. Tian, J.Y., Ernst, M., Cui, F., Jekel, M., 2013b. Effect of different cations on UF membrane fouling by NOM fractions. Chem. Eng. J. 223, 547e555. Van Haandal, A.C., Lettinga, G., 1994. Anaerobic Sewage Treatment: a Practical Guide for Regions with a Hot Climate. Wiley, New York. Vyas, H.K., Bennett, R.J., Marshall, A.D., 2002. Performance of crossflow microfiltration during constant transmembrane pressure and constant flux operation. Int. Dairy J. 12 (7473), 473e479. Yamamura, H., Okimoto, K., Kimura, K., Watanabe, Y., 2014. Hydrophilic fraction of natural organic matter causing irreversible fouling of microfiltration and ultrafiltration membranes. Water Res. 54, 123e125. Yao, M., Zhang, K., Cui, L., 2010. Characterization of proteinepolysaccharide ratios on membrane fouling. Desalination 259 (1e3), 11e17. Yuan, W., Zydney, A.L., 1999. Humic acid fouling during microfiltration. J. Membr. Sci. 157 (1), 1e12. Zheng, X., Ernst, M., Huck, P.M., Jekel, M., 2010. Biopolymer fouling in dead-end ultrafiltration of treated domestic wastewater. Water Res. 44 (18), 5212e5211.

Please cite this article in press as: Yue, X., et al., Effects of dissolved organic matters (DOMs) on membrane fouling in anaerobic ceramic membrane bioreactors (AnCMBRs) treating domestic wastewater, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.07.038