- Email: [email protected]
Fouling potential evaluation of soluble microbial products (SMP) with different membrane surfaces in a hybrid membrane bioreactor using worm reactor for sludge reduction
Bioresource Technology 140 (2013) 111–119
Contents lists available at SciVerse ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Fouling potential evaluation of soluble microbial products (SMP) with different membrane surfaces in a hybrid membrane bioreactor using worm reactor for sludge reduction Zhipeng Li b, Yu Tian a,b,⇑, Yi Ding b, Lin Chen c, Haoyu Wang b a b c
State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology (SKLUWRE, HIT), Harbin 150090, China School of Municipal and Environmental Engineering, Harbin Institute of Technology, Harbin 150090, China School of Civil and Environmental Engineering and Singapore Membrane Technology Centre, Nanyang Technological University, Singapore 639798, Singapore
h i g h l i g h t s The fouling potential of S-SMP was investigated with different types of membranes. The worm reactor reduced fouling propensity of SMP regardless of membrane properties. The weakened fouling potential of SMP was confirmed by interaction energy analysis. The more second minimums surrounding each asperity enhancing the attachment of SMP. The different membranes fouling were predicated by interaction energy and roughness.
a r t i c l e
i n f o
Article history: Received 16 March 2013 Received in revised form 18 April 2013 Accepted 20 April 2013 Available online 29 April 2013 Keywords: Membrane bioreactor (MBR) Worm reactor Soluble microbial products Membrane surface properties Interaction energy
a b s t r a c t The fouling characteristics of soluble microbial products (SMP) in the membrane bioreactor coupled with Static Sequencing Batch Worm Reactor (SSBWR-MBR) were tested with different types of membranes. It was noted that the flux decrements of S-SMP (SMP in SSBWR-MBR) with cellulose acetate (CA), polyvinylidene fluoride (PVDF) and polyether sulfones (PES) membranes were respectively 6.7%, 8.5% and 9.5% lower compared to those of C-SMP (SMP in Control-MBR) with corresponding membranes. However, for both the filtration of the C-SMP and S-SMP, the CA membrane exhibited the fastest diminishing rate of flux among the three types of membranes. The surface morphology analysis showed that the CA membrane exhibited more but smaller protuberances compared to the PVDF and PES. The second minimums surrounding each protruding asperity on CA membrane were more than those on the PVDF and PES membranes, enhancing the attachment of SMP onto the membrane surface. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction Membrane bioreactors (MBRs) have been increasingly and widely used for wastewater treatments in the last decade (Drews, 2010; Hwang et al., 2010; Meng et al., 2009). However, membrane fouling is a major obstacle to the wide application of MBRs. Additionally, the excess sludge withdrawn from the MBR creates a concurrent sludge management problem (Oh et al., 2007). Recently, a MBR coupled with a Static Sequencing Batch Worm Reactor (SSBWR-MBR) was designed, in which excellent sludge reduction efficiency and membrane permeability were obtained (Tian et al., 2012a,b,c). Moreover, the decreased content and changed charac⇑ Corresponding author at: School of Municipal and Environmental Engineering, Harbin Institute of Technology, Harbin 150090, China. Tel.: +86 451 8628 3077/+86 13804589869; fax: +86 451 8628 3077. E-mail address: [email protected] (Y. Tian). 0960-8524/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.biortech.2013.04.078
teristics of extracellular polymeric substance (EPS) due to the predated sludge recycle was one reason for the improved membrane permeability of the mixed liquor (Tian et al., 2012b,c). In the literature, Geng and Hall (2007) found that the bound EPS in a mixed liquor influences membrane filtration indirectly through the release of EPS components into the liquid phase. Ng et al. (2006) indicated that the release of EPS into the bulk solution that would be further hydrolyzed into soluble microbial products (SMP). With improved understanding of fouling mechanisms, SMP have received more attention because of their important role in sludge filterability and membrane fouling (Evenblij and Van der Graaf, 2004; Geng and Hall, 2007). The research of (Tian et al., 2013) was only to investigate the effect of worm reactor on the fouling potential of SMP in the MBR coupled with Worm Reactor (SSBWR-MBR); it was found that the fouling potential of SMP was weaken with the predated sludge recycle. The membranes employed for filtration were polyvinylidene fluoride (PVDF)
112
Z. Li et al. / Bioresource Technology 140 (2013) 111–119
membrane. However, the effect of predated sludge recycle on fouling potential of the SMP with other material membranes in the SSBWR-MBR has never been investigated. It is widely known that the formation of an organic layer during initial stages of membrane filtration leads to subsequent fouling layer development on the membrane surface (Ng and Ng, 2010). In the literature, Le-Clech et al. (2006) reported that the main parameter affecting the initial fouling would be the irreversible deposition of the SMP of the biomass suspension. Organic foulant deposition is usually interpreted through extended Derjaguin–Landau–Verwey–Overbeek (XDLVO) models (Kang et al., 2004; Subramani and Hoek, 2008; Wang et al., 2005). Thus, we focused on physicochemical interactions governing initial SMP deposition, but important stage of the much more complex overall process of membrane fouling. Recent studies found that the surface topological properties of membranes had serious implications on membrane performance and fouling propensity (Elimelech et al., 1997; Park et al., 2005). Rough surfaces fouled more easily because roughness may have an equivalent short-range effect on colloidal interactions (Bhattacharjee et al., 1998) and affect surface properties (Wong et al., 2009). The motivation of this research was, therefore, to contribute towards a better understanding the fouling behaviors of SMP in the Control-MBR (C-SMP) and SSBWR-MBR (S-SMP) using different types of membranes. Firstly, the changed filtration characteristics of SMP due to worm predation was investigated by means of batch filtration experiments with different types of membranes; secondly, the physicochemical interactions between the different types of membranes and SMP were characterized through extended XDLVO models; finally, the roughness of the fouled different types of membranes were detected using atomic force microscopy (AFM). This study would be useful for further confirming whether the worm predation reduced the fouling potential of SMP regardless of the types of membrane.
2. Methods 2.1. SSBWR-MBR system The schematic of the Control- and SSBWR-MBRs is presented in Fig. 1: a MBR (C-MBR) coupled with a SSBWR without worms and a MBR (S-MBR) combined with a SSBWR with worms. The set-up of the MBR system was similar to that described in previous studies (Tian et al., 2012a,b). The configuration of SSBWR was provided in (Tian et al., 2010), and the SSBWR (working volume: 39 L) without and with worms had the same operational condition. The two MBRs, each with a working volume of 40 L, were fed with synthetic wastewater. Synthetic wastewater was used to avoid big fluctuations in the feed concentration. Many researchers have used synthetic wastewater to investigate the relationship between the characteristics of sludge and the membrane fouling in MBR (Shariati et al., 2011; Sun et al., 2011; Tian et al., 2011; Wu and Huang, 2010; Xuan et al., 2010). In this study, the synthetic wastewater was mainly composed of glucose and starch as a carbon source. Stock phosphate solution was prepared to serve as a phosphate source, which was diluted to achieve the desired total phosphorus (TP) concentration. Ammonium sulfate and carbamide were used as nitrogen source. The synthetic influent was composed of the following components: glucose 200 mg/L; starch 200 mg/L; NaHCO3 300 mg/L; CO(NH2)2 32.1 mg/L; NH4Cl 95.5 mg/L; KH2PO4 47 mg/ L; and other trace metals (MgSO4 40 mg/L and CaCl2 5 mg/L). One membrane module which was made of hollow fibers of PVDF with a surface area of 1 m2 and a mean pore size of 0.22 lm (Motimo, China) was immersed in each MBR. An aeration system was placed at the bottom of each MBR to maintain desired dissolved oxygen
(DO) concentration, and the hydraulic retention times (HRT) and sludge retention time (SRT) were controlled at 5 h and 30 days, respectively. A suction pump was used to collect the effluent from the membrane module, and a manometer was fixed between the membrane module and the suction pump to monitor the TMP. The membrane flux was maintained at 8 L/ (m2 h) with a suction mode of 8 min ‘‘on’’ and 2 min ‘‘off’’. Once a day, the same amount of excess sludge withdrawn from the C-MBR and S-MBR were fed to the SSBWR without worms and SSBWR with worms, respectively. About one third of the excess sludge was consumed in the SSBWR with worms (Tian et al., 2010). Then the non-consumed sludge and possible metabolites in the SSBWR with worms were recycled to the S-MBR. In the SSBWR without worms, one third of excess sludge was discharged to maintain the same SRT in both MBRs and subsequently, the residual sludge was returned to the C-MBR. Prior to the batch experiments, the C-MBR and S-MBR were operated at a steady state for 300 days with the MLSS concentration of 9000 mg/L. 2.2. Extraction of SMP from mixed liquor The activated sludge samples were collected from the two MBRs. The SMP sample was separated from the sludge mixed liquor by centrifugation (4000 rpm for 5 min) and membrane filtration (0.45 lm, cellulose acetate membrane). The obtained SMP was then stored in the refrigerator at 4 °C, and the calculated ionic strength was 0.01 M. 2.3. Filtration apparatus The fouling propensities of the C-SMP and S-SMP were evaluated using a stirred dead-end cell (MSC300, Mosu Corp., China) operated at room temperature (20 ± 1 °C). Three types of commercial membranes with cellulose acetate (CA), PVDF and polyether sulfones (PES) were employed for filtration with nominal pore size of 0.22 lm and an effective membrane area of 19.62 cm2. The main characteristics of SMP and three types of membranes are shown in Supplementary Table S1. Before each experiment, deionized (DI) water was filtered through the membrane prior to the fouling experiment for 1 h to stabilize the filtration system. The filtration pressure was maintained constant at 10 kPa, and the stirring speed in the cell was set at 250 rpm to prevent the formation of a concentration polarization layer and to provide a simulation of cross-flow velocity. Permeate flux data were continuously logged using a toploading electronic balance (BL-1200S, Setra Systems, USA) connected to a personal computer. 2.4. Surface thermodynamics analysis Surface thermodynamic properties, such as the surface free energy, were estimated using a contact angle approach. Surface tension components are determined from the extended Young equation (Van Oss, 2006).
ð1 þ coshÞcTOT ¼2 l
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
LW cLW þ s cl
pffiffiffiffiffiffiffiffiffiffiffi
cþs cl þ
pffiffiffiffiffiffiffiffiffiffiffi
cAB ¼ 2 cþ c
qffiffiffiffiffiffiffiffiffiffiffi
cs cþl
ð1Þ
ð2Þ
cTOT ¼ cLW þ cAB
ð3Þ TOT
Where h is the contact angle, c is the total surface tension, cLW is the Lifshitz–van der Waals component, and c+ and c- are the electron-acceptor and electron-donor components, respectively. The subscripts s and l represent the solid surface and the
Z. Li et al. / Bioresource Technology 140 (2013) 111–119
113
Fig. 1. A schematic representation of the control-MBR and SSBWR-MBR (Tian et al., 2013).
liquid, respectively. The surface tension components were converted into free energies following the model described by (Brant and Childress, 2002). 2.5. Extended DLVO theory The XDLVO theory Eq. (4) describes the total interaction energy per unit area (E) of particle -surface in terms of Lifshitz–van der Waals force (LW), electrostatic force (EL) energy and acid–base (AB) interaction energy. EL AB EXDLVO ¼ ELW 123 123 þ E123 þ E123
ð4Þ
where the subscripts 1–3 represent the foulant, water and foulant, respectively. The interaction energy per unit area for LW, EL and AB is estimated from Eqs. (5–7) (Van Oss, 2006) as the function of separation distance (h):
ELW 123 ðhÞ ¼
EEL 123 ðhÞ
A 12ph
ð5Þ
2
" # n21 þ n23 1 ¼ ee0 kn1 n3 ð1 cothkhÞ þ sinhkh 2n1 n3
AB EAB 123 ðhÞ ¼ DGh0 exp
h0 h k
ð6Þ
ð7Þ
2
where A ¼ 12ph0 DGLW h0 at the right-hand side of Eq. (2) is the qffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffi r LW rLW rLW rLW is Hamaker constant; DGLW h0 ¼ 2ð 2 1 Þð 3 2 Þ the free LW energy per unit area between the surfaces; h0 is the
minimum cut-off distance due to Born repulsion; rLW is the Lifshitz–van der Waals component; e is the dielectric constant of fluid; qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi e0 is the dielectric permittivity of vacuum, 1k ¼ e0 er Rg T=ð2F 2 IS Þ is the Debye length; T is the absolute temperature in Kelvins; e0 is the permittivity of free space; ew is the dielectric constant; Rg is the gas constant; F is the Faraday’s constant; Is is the ionic strength; n1 and n3 represent the surface potentials of membrane and foulant, respectively; k is the decay length of AB interaction;
qffiffiffiffiffi pffiffiffiffiffi pffiffiffiffiffi pffiffiffiffiffi pffiffiffiffiffi qffiffiffiffiffi qffiffiffiffiffi qffiffiffiffiffi DGAB r þ2 ð r 1 þ r3 r 2 Þ þ 2 r2 ð r þ1 þ r þ3 r þ2 Þ h0 ¼ 2 qffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffi 2ð rþ1 r 3 r1 r þ3 Þ is the acid–base free energy per unit area between the surfaces at contact. The calculated surface tension parameters and free energy of cohesion for each clean membrane and SMP are shown in Supplementary Table S2. 2.6. Analytical methods Total organic carbon (TOC) concentrations in the SMP samples were measured by a TOC analyser (TOC-5000A, Shimadzu, Japan). The zeta potentials (f) and size distribution of the SMP were monitored using a zetasizer (Zetasizer 3000 HS type A, Malverin, England). Atomic force microscopy (AFM) (Veeco, Santa Barbara, CA) was used to describe foulant layer morphology in terms of membrane surface topography. The membrane roughness was determined from the AFM height images of 10 10 lm areas. The Nanoscope control software (Version 5.30r3sr3) was used for image acquisition; the mean average roughness (Ra) and root mean
114
Z. Li et al. / Bioresource Technology 140 (2013) 111–119
square (Rms) of surface roughness values were calculated from at least 7 images for each membrane sample. Contact angle measurements were performed using an NRL Contact Angle Goniometer (Rame Hart, Mountain Lakes, NJ), which is a standard goniometer with image analysis attachments (i.e., video camera, computer with monitor, and image analysis software). 3. Results and discussion 3.1. Fouling propensity of C-SMP and S-SMP with different types of membranes The C-SMP and S-SMP samples derived from the two MBRs were tested with CA, PVDF and PES membranes in a stirred cell with similar initial flux. The flux decline trends by different membranes are shown in Fig. 2. For the C-SMP, the flux of CA membrane was found to drop by 85.4% when filtration experiment stopped and about 300 mL of permeate was collected, while the PVDF and PES membranes resulted in only ca. 73.4% and 49.2% decrement, respectively. With respect to the S-SMP, the flux of CA, PVDF and PES membranes resulted in ca. 78.7%, 64.9% and 39.7% decrement, respectively. It was noted that the flux decrement of S-SMP with CA, PVDF and PES were lower compared to that of C-SMP with corresponding membranes. This meant that the predated sludge recycle could reduce the fouling propensity of the SMP regardless of the characteristics of membrane. However, for both the filtration of the C-SMP and S-SMP, the CA membrane exhibited the fastest diminishing rate of flux among the three types of membranes. Since the same SMP was filtered and similar initial flux was applied, the different rate of permeability decline among the CA, PVDF and PES could be much dependent on the characteristics of membrane surface. Regarding to the C-SMP and S-SMP filtration using the same membrane, the hydrophilic/hydrophobic interaction between the SMP and membrane may be of much importance, which was supported by previous research (Tian et al., 2013). The interfacial free energy of cohesion (DGcoh), represents the free energy (per unit area) when two surfaces of the same material are immersed in a solvent (water). The free energy of cohesion gives a qualitative representation of the hydrophobicity–hydrophilicity of organics and membranes (Van Oss et al., 2007). With respect to a hydrophobic material, the free energy of cohesion is negative, and the free energy of cohesion is positive for hydrophilic materials. It was seen from Supplementary Table S2 that the C-SMP and S-SMP both presented hydrophobic nature. The free energy of CA membrane (53.70 mJ/m2) was lower than that of PVDF (75.40 mJ/m2) and PES (62.75 mJ/m2) membranes, suggesting CA has the lowest SMP fouling tendency in terms of adsorption onto the membrane surface. But the experimental fouling data inconsistently showed that the CA fouled more severely than PVDF and PES membranes. It seemed that the characteristics of hydrophobicity may not be sufficient to predict the effect of membrane property on membrane fouling. Hence, the relative fouling behavior of the three commercial membranes should be predicted from the combined knowledge of physical surface morphology and surface chemical properties. 3.2. Influence of predated sludge recycle on the interaction energy between SMP and different clean membrane surfaces SMP adsorption and deposition on membrane surfaces are affected by the interaction energies between the SMP constituents and the membrane surface that can be predicted implementing the XDLVO theory (Ying et al., 2010). The interaction energy profiles of S-SMP and C-SMP approaching the clean CA, PVDF and
Fig. 2. Normalized flux of the three different membranes during filtration of C-SMP, S-SMP and pure water (filtration pressure: 10 kPa). J stands for flux and J0 is the initial flux.
PES membrane were calculated, exhibiting an energy barrier and a secondary energy minimum (Fig. 3). The secondary energy minimum represents the ability of SMP being sucked onto membrane surface (Hoek et al., 2003). An energy barrier means the SMP in suspension should have sufficient kinetic energy to overcome this barrier to approach the membrane (Redman et al., 2004).
Z. Li et al. / Bioresource Technology 140 (2013) 111–119
115
Fig. 3. Interaction energies profiles between the C-SMP (1) and S-SMP (2) and the clean CA (a), PVDF (b) and PES (c) membrane as a function of separation distance.
The interfacial energies profiles between the SMP and the clean CA membrane surface are shown in Fig. 3a. The C-SMP and S-SMP particle experienced attraction from 50 nm and this attractive energy reached a maximum until the particle came to 15 nm separation distance. The maximum attractive energies (secondary energy minimum) for C-SMP and S-SMP particle were significantly different, with the values of 1.78 KT and 0.72 KT, respectively.
At 8 nm separation, the repulsive electrostatic interactions dominated and the total interaction quickly became repulsive. The maximum repulsive energy (energy barrier) of the S-SMP (7.13 KT) was much higher than that of the C-SMP (3.75 KT), indicating that when the SMP was brought to the clean CA membrane surface by the convective permeate flow, the S-SMP needed to overcome the higher repulsive interaction energy than the C-SMP. At 3 nm
116
Z. Li et al. / Bioresource Technology 140 (2013) 111–119
separation, the acid–base attraction dominated and the total interaction quickly became attractive for S-SMP and C-SMP. The interaction energy curves between the SMP and the clean PVDF membrane surface are presented in Fig. 3b. The C-SMP particle encountered a weak attraction at a distance of about 50 nm due to the present of the van der Waals attraction, and then at 15 nm, the attractive energy reached a maximum (1.58 KT). Due to the dominated electrostatic repulsion, the total interaction became repulsion at 8 nm, and at 5 nm, this repulsive energy reached a maximum (4.04 KT). With respect to the S-SMP approaching the clean PVDF membrane surface, the energy barrier was 76.0% stronger than that between the C-SMP and the clean PVDF membrane surface. The second minimum in the interaction energy profile of S-SMP was 58.9% less than that of C-SMP. With respect to the interaction between the SMP and the clean PES membrane surface (Fig. 3c), the second minimum in the interaction energy profile of S-SMP was 33.9% less than that of C-SMP, however, the energy barrier (8.03 KT) was higher than that of CSMP (7.32 KT). For different clean membranes, the interaction energy profiles of S-SMP exhibited a lower second minimum and a higher energy barrier than those of C-SMP. The lower second minimum showed that the energy of S-SMP being sucked onto the clean CA, PVDF and PES membrane surface was weaker than that of C-SMP being sucked onto the corresponding membrane surface; meanwhile, the higher energy barriers indicated that the S-SMP approaching the membrane surface needed to overcome stronger repulsive interaction energy compared to the C-SMP approaching the corresponding membrane surface. This could further confirm that the
Table 4 Summary of fouled membrane roughness parameters. SMP
Fouled membrane
hwater (°)
hformamide (°)
hdiiodomethane (°)
C-SMP
CA PVDF PES
65.9 80.9 78.79
15.99 53.99 54.2
21.31 35.11 32.59
S-SMP
CA PVDF PES
64.33 78.33 74.89
20.04 51.04 50.34
23.89 28.31 34.11
RMS roughness Average (Rms) roughness (Ra)
Maximum roughness (lm)
C-SMP CA PVDF PES
183.32 342.58 263.38
114.44 251.49 196.28
1.362 2.923 1.948
S-SMP CA PVDF PES
130.81 306.26 133.78
102.87 233.79 105.31
0.986 2.546 1.238
predated sludge recycle reduced the fouling potential of the SMP regardless of the characteristics of membrane. 3.3. Influence of predated sludge recycle on the interaction energy between SMP and different fouled membrane surfaces Adsorption of SMP onto CA, PVDF and PES membranes reduced pure water contact angles and changed contact angles of formamide and diiodomethane (Table 1). This suggested that the chemistry of the membrane surfaces has been significantly altered. However, neither membrane exhibited the exact contact angle behavior of the thick SMP layers analyzed in Supplementary Table S1. It is possible that the surfaces were not completely coated with SMP and some of the clean membrane remained exposed after filtration. Calculated interfacial interaction parameters for the organic-fouled membranes are all listed in Table 2. The zeta potentials of the clean membrane were assumed for the organic-fouled membrane surfaces. Although this assumption may not be fundamentally justified, it introduced very little error in our analyses below because the electrostatic double layer interfacial free energies at contact were negligible. The membranes fouled by C-SMP and S-SMP expressed significantly different characters, but they had similar change tendency. The membranes fouled by C-SMP and S-SMP all showed a slight reduction in apolar character. The free energy of cohesion values became significantly lower after organic adsorption for CA, PVDF and PES membrane. Hence, after adsorption of C-SMP and S-SMP, CA, PVDF and PES become relatively hydrophilic compared to clean membrane surfaces. Moreover, the membranes
Table 1 Measured physicochemical properties of fouled membrane. SMP
Fouled membranes
Table 2 Calculated surface energetic parameters for fouled membranes. SMP
Fouled membranes
cLW
c+
c
DGLW sls
DGAB sls
DGEL sls
DGcoh
C-SMP
CA PVDF PES
47.39 41.98 43.12
2.92 0.18 0.06
4.32 3.74 5.31
9.8124 6.5533 7.2013
39.7078 57.6492 52.7646
0.0576 0.0548 0.069
49.4625 64.1477 59.8969
S-SMP
CA PVDF PES
46.54 44.91 42.44
2.5 0.12 0.21
5.96 4.51 6.78
9.2706 8.2617 6.8121
36.1907 55.0494 44.9215
0.0576 0.0548 0.069
45.4037 63.2563 51.6646
Table 3 Summary of DLVO interaction energy calculations of the clean and fouled membrane. SMP
Clean membranes
Energy barrier (KT)
Secondary minimum (KT)
Fouled membranes
Energy barrier (KT)
Secondary minimum (KT)
Flux declined J/J0 (%)
C-SMP
CA PVDF PES
3.75 4.04 7.32
1.78 1.59 1.21
CA PVDF PES
4.16 5.26 6.86
1.72 1.32 1.36
85.4 73.4 49.2
S-SMP
CA PVDF PES
7.13 7.13 8.03
0.72 0.64 0.80
CA PVDF PES
5.90 6.01 8.24
1.08 1.00 0.85
78.7 64.9 39.7
Z. Li et al. / Bioresource Technology 140 (2013) 111–119
fouled by S-SMP showed weaker hydrophobic than those fouled by C-SMP. The interfacial energies profiles between the SMP and the fouled membrane surface (CA, PVDF and PES) were calculated implementing the XDLVO theory, which are presented in Fig. 4. With respect to the C-SMP, the secondary energy minimum value for CA was approximately 1.72 KT, but changed into 1.32 KT for PVDF; finally, the PES had a potential well of 1.36 KT. The corre-
117
sponding secondary energy minimum values for S-SMP were 1.08 KT, 1.00 KT and 0.85 KT, respectively. The energy barriers of S-SMP with CA, PVDF and PES were respectively 39.4%, 14.2%, and 20.6% bigger than those of C-SMP with corresponding membranes. In order to explain the relationship between the interfacial energies profiles and fouling rate, the secondary energy minimums and energy barriers between the SMP and clean and fouled
Fig. 4. Interaction energies profiles between the C-SMP (1) and S-SMP (2) and the fouled CA (a), PVDF (b) and PES (c) membrane as a function of separation distance.
118
Z. Li et al. / Bioresource Technology 140 (2013) 111–119
membrane are summarized in Table 3. For the same membrane materials, the results of XDLVO theory were accordance with the SMP adhesion; the S-SMP with lower secondary energy minimums and higher energy barriers in the interaction energy profiles showed weak membrane fouling. The same membrane materials have similar property, hence, the interfacial energies is governing factor of SMP adhesion onto membrane surfaces. Apparently, worm predation could reduce fouling potential of S-SMP regardless of the membrane materials through changing the interaction energy. But, the membrane fouling rate did not follow XDLVO predictions with different membrane surfaces. The energy barriers and the secondary energy minimums for S-SMP filtration with CA and PVDF were similar. However, significant difference in fouling rate was detected between CA and PVDF membranes. Therefore, the comparability of the different fouling rates between the different membranes was not mainly determined by the interfacial energies of the membrane and SMP. It was quite possible that the presence of morphological heterogeneities can result in surfaces with very different energy distributions than smooth surfaces, and hence, can result in substantially different fouling behavior compared to XDLVO predictions (Brant and Childress, 2002). 3.4. The effect of roughness on fouling propensity of C-SMP and S-SMP with different types of membranes The surface topological properties of membranes had serious implications on membrane performance and fouling propensity (Lee et al., 2005). The surface morphologies of the CA, PVDF and PES membranes fouled by C-SMP and S-SMP, reflected by Ra and Rms were significantly different (Table 4). For example, for the PVDF membrane fouled by C-SMP, the Rms and Ra were larger than the values for virgin membranes (Supplementary Fig. S1). The increase in membrane surface roughness might be attributed to the surface enrichment of large molecules due to fouling caused by pore blocking or surface adsorption. However, the values of Rms and Ra of fouled PVDF membrane induced by S-SMP showed little lower than that of the virgin membrane The decrease of roughness might be attributed to a result of the deposition of small molecular in the ridge-and-valley structure of the membrane and in pore so as to make the membrane surface flatter and the roughness decreased. A mechanistic explanation for the pronounced influence of surface roughness on colloidal fouling emerges by analyzing the nature of membrane morphology (Vrijenhoek et al., 2001). As permeation rate through membranes is proportional to the thickness of the active (skin) layer, the bottom of a valley presents the ‘‘path-of-least resistance’’ to permeating water. Hence, water convection and particle transport are focused towards the valley bottom. Thus, as the SMP are preferentially convected towards the center of a valley, the valleys rapidly become ‘‘clogged,’’ resulting in significant loss of permeate flux. The XDLVO interaction energy between the SMP and a smooth membrane surface was significantly altered by membrane surface topology (Chen et al., 2012). Both the PVDF and PES exhibited relatively fewer but larger protuberances, compared to the CA with abundant but small asperities (Supplementary Fig. S1). It was obvious that the relative fouling behavior of the three commercial membranes could not be predicted from knowledge of surface roughness (Table 4). Large-scale surface roughness significantly increased the rate of SMP attachment through providing a larger surface area and greater contact opportunities for SMP with the membrane surface. However, in this study, the CA with many small asperities showed the fastest flux decline. The CA membrane exhibited relatively more but smaller protuberances compared to the PVDF and PES with few large asperities and peaks. According to the calculation and drawing of the interaction energy map in our previous study (Chen et al., 2012), the attractive regions on
CA membrane were obviously larger than those on the PVDF and PES membrane due to the more second minimums surrounding each protruding asperity. Although the large-scale surface roughness were absent on CA membrane, the more second minimum enhanced the attachment of SMP onto the membrane surface. Compared with the CA membrane, the PVDF membrane exhibited the larger peak-to-peak separation distance, less attractive regions around the protrusions, indicating the lower fouling rate of PVDF membrane. In contrast to other two membranes, the PES membrane was much smoother with the least asperities, leading to relatively smaller attractive regions around the positive asperities. As the SMP approaching to the mean-plane, the probability of SMP being sucked onto the membrane significantly decreased. Consequently, the different fouling rates for the different membranes were attributed to the combination of interaction energy and the number of asperities effect. 4. Conclusions The flux declines of SMP were decreased by 6.7%, 8.5% and 9.5% for CA, PVDF and PES membranes, respectively, with the predated sludge recycle. For the same membrane materials, the weakened fouling potential of SMP due to worm predation was confirmed by the bigger energy barriers and lower secondary energy minimum. With respect to different membrane surfaces, the second minimums surrounding each protruding asperity on CA membrane were more than those on the PVDF and PES membrane. The different fouling rates between the different membranes were determined by the combination of interaction energy and the number of asperities. Acknowledgments This study was supported by the Major Science and Technology Program for Water Pollution Control and Management (No. 2013ZX07201007), the State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology (No. 2011DX01) and the Specialized Research Fund for the Doctoral Program of Higher Education (No. 20112302110060). The authors also appreciate the Funds for Creative Research Groups of China (No. 51121062). 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.2013. 04.078. References Bhattacharjee, S., Ko, C.H., Elimelech, M., 1998. DLVO interaction between rough surfaces. Langmuir 14, 3365–3375. Brant, J.A., Childress, A.E., 2002. Assessing short-range membrane–colloid interactions using surface energetics. J. Membr. Sci. 203, 257–273. Chen, L., Tian, Y., Cao, C.-Q., Zhang, J., Li, Z.-N., 2012. Interaction energy evaluation of soluble microbial products (SMP) on different membrane surfaces: role of the reconstructed membrane topology. Water Res. 46, 2693–2704. Drews, A., 2010. Membrane fouling in membrane bioreactors—Characterisation, contradictions, cause and cures. J. Membr. Sci. 363, 1–28. Elimelech, M., Zhu, X., Childress, A.E., Hong, S., 1997. Role of membrane surface morphology in colloidal fouling of cellulose acetate and composite aromatic polyamide reverse osmosis membranes. J. Membr. Sci. 127, 101–109. Evenblij, H., Van der Graaf, J., 2004. Occurrence of EPS in activated sludge from a membrane bioreactor treating municipal wastewater. Water Sci. Technol. 50, 293–300. Geng, Z., Hall, E.R., 2007. A comparative study of fouling-related properties of sludge from conventional and membrane enhanced biological phosphorus removal processes. Water Res. 41, 4329–4338. Hoek, E.M., Bhattacharjee, S., Elimelech, M., 2003. Effect of membrane surface roughness on colloid–membrane DLVO interactions. Langmuir 19, 4836–4847.
Z. Li et al. / Bioresource Technology 140 (2013) 111–119 Hwang, B.-K., Kim, J.-H., Ahn, C.H., Lee, C.-H., Song, J.-Y., Ra, Y.-H., 2010. Effect of disintegrated sludge recycling on membrane permeability in a membrane bioreactor combined with a turbulent jet flow ozone contactor. Water Res. 44, 1833–1840. Kang, S.T., Subramani, A., Hoek, E., Deshusses, M.A., Matsumoto, M.R., 2004. Direct observation of biofouling in cross-flow microfiltration: mechanisms of deposition and release. J. Membr. Sci. 244, 151–165. Le-Clech, P., Chen, V., Fane, T.A.G., 2006. Fouling in membrane bioreactors used in wastewater treatment. J. Membr. Sci. 284, 17–53. Lee, N., Amy, G., Croué, J.-P., Buisson, H., 2005. Morphological analyses of natural organic matter (NOM) fouling of low-pressure membranes (MF/UF). J. Membr. Sci. 261, 7–16. 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. Ng, T.C.A., Ng, H.Y., 2010. Characterisation of initial fouling in aerobic submerged membrane bioreactors in relation to physico-chemical characteristics under different flux conditions. Water Res. 44, 2336–2348. Ng, H.Y., Tan, T.W., Ong, S.L., 2006. Membrane fouling of submerged membrane bioreactors: impact of mean cell residence time and the contributing factors. Environ. Sci. Technol. 40, 2706–2713. Oh, Y.K., Lee, K.R., Ko, K.B., Yeom, I.T., 2007. Effects of chemical sludge disintegration on the performances of wastewater treatment by membrane bioreactor. Water Res. 41, 2665–2671. Park, N., Kwon, B., Kim, I.S., Cho, J., 2005. Biofouling potential of various NF membranes with respect to bacteria and their soluble microbial products (SMP): characterizations, flux decline, and transport parameters. J. Membr. Sci. 258, 43–54. Redman, J.A., Walker, S.L., Elimelech, M., 2004. Bacterial adhesion and transport in porous media: role of the secondary energy minimum. Environ. Sci. Technol. 38, 1777–1785. Shariati, S.R.P., Bonakdarpour, B., Zare, N., Ashtiani, F.Z., 2011. The effect of hydraulic retention time on the performance and fouling characteristics of membrane sequencing batch reactors used for the treatment of synthetic petroleum refinery wastewater. Bioresour. Technol. 102, 7692–7699. Subramani, A., Hoek, E., 2008. Direct observation of initial microbial deposition onto reverse osmosis and nanofiltration membranes. J. Membr. Sci. 319, 111–125. Sun, F.-Y., Wang, X.-M., Li, X.-Y., 2011. Change in the fouling propensity of sludge in membrane bioreactors (MBR) in relation to the accumulation of biopolymer clusters. Bioresour. Technol. 102, 4718–4725.
119
Tian, Y., Lu, Y., Chen, L., Lin, H., 2010. Optimization of process conditions with attention to the sludge reduction and stable immobilization in a novel Tubificidae-reactor. Bioresour. Technol. 101, 6069–6076. Tian, Y., Chen, L., Zhang, S., Cao, C., Zhang, S., 2011. Correlating membrane fouling with sludge characteristics in membrane bioreactors: an especial interest in EPS and sludge morphology analysis. Bioresour. Technol. 102, 8820–8827. Tian, Y., Li, Z., Chen, L., Lu, Y., 2012a. Role of extracellular polymeric substances (EPSs) in membrane fouling of membrane bioreactor coupled with worm reactor. Bioresour. Technol. 123, 566–573. Tian, Y., Li, Z., Lu, Y., 2012b. Changes in characteristics of soluble microbial products and extracellular polymeric substances in membrane bioreactor coupled with worm reactor: relation to membrane fouling. Bioresour. Technol. 122, 62–69. Tian, Y., Lu, Y., Li, Z., 2012c. Performance analysis of a combined system of membrane bioreactor and worm reactor: wastewater treatment, sludge reduction and membrane fouling. Bioresour. Technol. 121, 176–182. Tian, Y., Li, Z., Ding, Y., Lu, Y., 2013. Identification of the change in fouling potential of soluble microbial products (SMP) in membrane bioreactor coupled with worm reactor. Water Res. 47, 2015–2024. Van Oss, C.J., 2006. Interfacial Forces in Aqueous Media. CRC. van Oss, C.J., 2007. Development and applications of the interfacial tension between water and organic or biological surfaces. Colloids Surf., B 54, 2–9. Vrijenhoek, E.M., Hong, S., Elimelech, M., 2001. Influence of membrane surface properties on initial rate of colloidal fouling of reverse osmosis and nanofiltration membranes. J. Membr. Sci. 188, 115–128. Wang, S., Guillen, G., Hoek, E.M.V., 2005. Direct observation of microbial adhesion to membranes. Environ. Sci. Technol. 39, 6461–6469. Wong, P.C.Y., Kwon, Y.-N., Criddle, C.S., 2009. Use of atomic force microscopy and fractal geometry to characterize the roughness of nano-, micro-, and ultrafiltration membranes. J. Membr. Sci. 340, 117–132. Wu, J., Huang, X., 2010. Use of ozonation to mitigate fouling in a long-term membrane bioreactor. Bioresour. Technol. 101, 6019–6027. Xuan, W., Bin, Z., Zhiqiang, S., Zhigang, Q., Zhaoli, C., Min, J., Junwen, L., Jingfeng, W., 2010. The EPS characteristics of sludge in an aerobic granule membrane bioreactor. Bioresour. Technol. 101, 8046–8050. Ying, W., Yang, F., Bick, A., Oron, G., Herzberg, M., 2010. Extracellular polymeric substances (EPS) in a hybrid growth membrane bioreactor (HG-MBR): viscoelastic and adherence characteristics. Environ. Sci. Technol. 44, 8636– 8643.
Contents lists available at SciVerse ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Fouling potential evaluation of soluble microbial products (SMP) with different membrane surfaces in a hybrid membrane bioreactor using worm reactor for sludge reduction Zhipeng Li b, Yu Tian a,b,⇑, Yi Ding b, Lin Chen c, Haoyu Wang b a b c
State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology (SKLUWRE, HIT), Harbin 150090, China School of Municipal and Environmental Engineering, Harbin Institute of Technology, Harbin 150090, China School of Civil and Environmental Engineering and Singapore Membrane Technology Centre, Nanyang Technological University, Singapore 639798, Singapore
h i g h l i g h t s The fouling potential of S-SMP was investigated with different types of membranes. The worm reactor reduced fouling propensity of SMP regardless of membrane properties. The weakened fouling potential of SMP was confirmed by interaction energy analysis. The more second minimums surrounding each asperity enhancing the attachment of SMP. The different membranes fouling were predicated by interaction energy and roughness.
a r t i c l e
i n f o
Article history: Received 16 March 2013 Received in revised form 18 April 2013 Accepted 20 April 2013 Available online 29 April 2013 Keywords: Membrane bioreactor (MBR) Worm reactor Soluble microbial products Membrane surface properties Interaction energy
a b s t r a c t The fouling characteristics of soluble microbial products (SMP) in the membrane bioreactor coupled with Static Sequencing Batch Worm Reactor (SSBWR-MBR) were tested with different types of membranes. It was noted that the flux decrements of S-SMP (SMP in SSBWR-MBR) with cellulose acetate (CA), polyvinylidene fluoride (PVDF) and polyether sulfones (PES) membranes were respectively 6.7%, 8.5% and 9.5% lower compared to those of C-SMP (SMP in Control-MBR) with corresponding membranes. However, for both the filtration of the C-SMP and S-SMP, the CA membrane exhibited the fastest diminishing rate of flux among the three types of membranes. The surface morphology analysis showed that the CA membrane exhibited more but smaller protuberances compared to the PVDF and PES. The second minimums surrounding each protruding asperity on CA membrane were more than those on the PVDF and PES membranes, enhancing the attachment of SMP onto the membrane surface. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction Membrane bioreactors (MBRs) have been increasingly and widely used for wastewater treatments in the last decade (Drews, 2010; Hwang et al., 2010; Meng et al., 2009). However, membrane fouling is a major obstacle to the wide application of MBRs. Additionally, the excess sludge withdrawn from the MBR creates a concurrent sludge management problem (Oh et al., 2007). Recently, a MBR coupled with a Static Sequencing Batch Worm Reactor (SSBWR-MBR) was designed, in which excellent sludge reduction efficiency and membrane permeability were obtained (Tian et al., 2012a,b,c). Moreover, the decreased content and changed charac⇑ Corresponding author at: School of Municipal and Environmental Engineering, Harbin Institute of Technology, Harbin 150090, China. Tel.: +86 451 8628 3077/+86 13804589869; fax: +86 451 8628 3077. E-mail address: [email protected] (Y. Tian). 0960-8524/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.biortech.2013.04.078
teristics of extracellular polymeric substance (EPS) due to the predated sludge recycle was one reason for the improved membrane permeability of the mixed liquor (Tian et al., 2012b,c). In the literature, Geng and Hall (2007) found that the bound EPS in a mixed liquor influences membrane filtration indirectly through the release of EPS components into the liquid phase. Ng et al. (2006) indicated that the release of EPS into the bulk solution that would be further hydrolyzed into soluble microbial products (SMP). With improved understanding of fouling mechanisms, SMP have received more attention because of their important role in sludge filterability and membrane fouling (Evenblij and Van der Graaf, 2004; Geng and Hall, 2007). The research of (Tian et al., 2013) was only to investigate the effect of worm reactor on the fouling potential of SMP in the MBR coupled with Worm Reactor (SSBWR-MBR); it was found that the fouling potential of SMP was weaken with the predated sludge recycle. The membranes employed for filtration were polyvinylidene fluoride (PVDF)
112
Z. Li et al. / Bioresource Technology 140 (2013) 111–119
membrane. However, the effect of predated sludge recycle on fouling potential of the SMP with other material membranes in the SSBWR-MBR has never been investigated. It is widely known that the formation of an organic layer during initial stages of membrane filtration leads to subsequent fouling layer development on the membrane surface (Ng and Ng, 2010). In the literature, Le-Clech et al. (2006) reported that the main parameter affecting the initial fouling would be the irreversible deposition of the SMP of the biomass suspension. Organic foulant deposition is usually interpreted through extended Derjaguin–Landau–Verwey–Overbeek (XDLVO) models (Kang et al., 2004; Subramani and Hoek, 2008; Wang et al., 2005). Thus, we focused on physicochemical interactions governing initial SMP deposition, but important stage of the much more complex overall process of membrane fouling. Recent studies found that the surface topological properties of membranes had serious implications on membrane performance and fouling propensity (Elimelech et al., 1997; Park et al., 2005). Rough surfaces fouled more easily because roughness may have an equivalent short-range effect on colloidal interactions (Bhattacharjee et al., 1998) and affect surface properties (Wong et al., 2009). The motivation of this research was, therefore, to contribute towards a better understanding the fouling behaviors of SMP in the Control-MBR (C-SMP) and SSBWR-MBR (S-SMP) using different types of membranes. Firstly, the changed filtration characteristics of SMP due to worm predation was investigated by means of batch filtration experiments with different types of membranes; secondly, the physicochemical interactions between the different types of membranes and SMP were characterized through extended XDLVO models; finally, the roughness of the fouled different types of membranes were detected using atomic force microscopy (AFM). This study would be useful for further confirming whether the worm predation reduced the fouling potential of SMP regardless of the types of membrane.
2. Methods 2.1. SSBWR-MBR system The schematic of the Control- and SSBWR-MBRs is presented in Fig. 1: a MBR (C-MBR) coupled with a SSBWR without worms and a MBR (S-MBR) combined with a SSBWR with worms. The set-up of the MBR system was similar to that described in previous studies (Tian et al., 2012a,b). The configuration of SSBWR was provided in (Tian et al., 2010), and the SSBWR (working volume: 39 L) without and with worms had the same operational condition. The two MBRs, each with a working volume of 40 L, were fed with synthetic wastewater. Synthetic wastewater was used to avoid big fluctuations in the feed concentration. Many researchers have used synthetic wastewater to investigate the relationship between the characteristics of sludge and the membrane fouling in MBR (Shariati et al., 2011; Sun et al., 2011; Tian et al., 2011; Wu and Huang, 2010; Xuan et al., 2010). In this study, the synthetic wastewater was mainly composed of glucose and starch as a carbon source. Stock phosphate solution was prepared to serve as a phosphate source, which was diluted to achieve the desired total phosphorus (TP) concentration. Ammonium sulfate and carbamide were used as nitrogen source. The synthetic influent was composed of the following components: glucose 200 mg/L; starch 200 mg/L; NaHCO3 300 mg/L; CO(NH2)2 32.1 mg/L; NH4Cl 95.5 mg/L; KH2PO4 47 mg/ L; and other trace metals (MgSO4 40 mg/L and CaCl2 5 mg/L). One membrane module which was made of hollow fibers of PVDF with a surface area of 1 m2 and a mean pore size of 0.22 lm (Motimo, China) was immersed in each MBR. An aeration system was placed at the bottom of each MBR to maintain desired dissolved oxygen
(DO) concentration, and the hydraulic retention times (HRT) and sludge retention time (SRT) were controlled at 5 h and 30 days, respectively. A suction pump was used to collect the effluent from the membrane module, and a manometer was fixed between the membrane module and the suction pump to monitor the TMP. The membrane flux was maintained at 8 L/ (m2 h) with a suction mode of 8 min ‘‘on’’ and 2 min ‘‘off’’. Once a day, the same amount of excess sludge withdrawn from the C-MBR and S-MBR were fed to the SSBWR without worms and SSBWR with worms, respectively. About one third of the excess sludge was consumed in the SSBWR with worms (Tian et al., 2010). Then the non-consumed sludge and possible metabolites in the SSBWR with worms were recycled to the S-MBR. In the SSBWR without worms, one third of excess sludge was discharged to maintain the same SRT in both MBRs and subsequently, the residual sludge was returned to the C-MBR. Prior to the batch experiments, the C-MBR and S-MBR were operated at a steady state for 300 days with the MLSS concentration of 9000 mg/L. 2.2. Extraction of SMP from mixed liquor The activated sludge samples were collected from the two MBRs. The SMP sample was separated from the sludge mixed liquor by centrifugation (4000 rpm for 5 min) and membrane filtration (0.45 lm, cellulose acetate membrane). The obtained SMP was then stored in the refrigerator at 4 °C, and the calculated ionic strength was 0.01 M. 2.3. Filtration apparatus The fouling propensities of the C-SMP and S-SMP were evaluated using a stirred dead-end cell (MSC300, Mosu Corp., China) operated at room temperature (20 ± 1 °C). Three types of commercial membranes with cellulose acetate (CA), PVDF and polyether sulfones (PES) were employed for filtration with nominal pore size of 0.22 lm and an effective membrane area of 19.62 cm2. The main characteristics of SMP and three types of membranes are shown in Supplementary Table S1. Before each experiment, deionized (DI) water was filtered through the membrane prior to the fouling experiment for 1 h to stabilize the filtration system. The filtration pressure was maintained constant at 10 kPa, and the stirring speed in the cell was set at 250 rpm to prevent the formation of a concentration polarization layer and to provide a simulation of cross-flow velocity. Permeate flux data were continuously logged using a toploading electronic balance (BL-1200S, Setra Systems, USA) connected to a personal computer. 2.4. Surface thermodynamics analysis Surface thermodynamic properties, such as the surface free energy, were estimated using a contact angle approach. Surface tension components are determined from the extended Young equation (Van Oss, 2006).
ð1 þ coshÞcTOT ¼2 l
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
LW cLW þ s cl
pffiffiffiffiffiffiffiffiffiffiffi
cþs cl þ
pffiffiffiffiffiffiffiffiffiffiffi
cAB ¼ 2 cþ c
qffiffiffiffiffiffiffiffiffiffiffi
cs cþl
ð1Þ
ð2Þ
cTOT ¼ cLW þ cAB
ð3Þ TOT
Where h is the contact angle, c is the total surface tension, cLW is the Lifshitz–van der Waals component, and c+ and c- are the electron-acceptor and electron-donor components, respectively. The subscripts s and l represent the solid surface and the
Z. Li et al. / Bioresource Technology 140 (2013) 111–119
113
Fig. 1. A schematic representation of the control-MBR and SSBWR-MBR (Tian et al., 2013).
liquid, respectively. The surface tension components were converted into free energies following the model described by (Brant and Childress, 2002). 2.5. Extended DLVO theory The XDLVO theory Eq. (4) describes the total interaction energy per unit area (E) of particle -surface in terms of Lifshitz–van der Waals force (LW), electrostatic force (EL) energy and acid–base (AB) interaction energy. EL AB EXDLVO ¼ ELW 123 123 þ E123 þ E123
ð4Þ
where the subscripts 1–3 represent the foulant, water and foulant, respectively. The interaction energy per unit area for LW, EL and AB is estimated from Eqs. (5–7) (Van Oss, 2006) as the function of separation distance (h):
ELW 123 ðhÞ ¼
EEL 123 ðhÞ
A 12ph
ð5Þ
2
" # n21 þ n23 1 ¼ ee0 kn1 n3 ð1 cothkhÞ þ sinhkh 2n1 n3
AB EAB 123 ðhÞ ¼ DGh0 exp
h0 h k
ð6Þ
ð7Þ
2
where A ¼ 12ph0 DGLW h0 at the right-hand side of Eq. (2) is the qffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffi r LW rLW rLW rLW is Hamaker constant; DGLW h0 ¼ 2ð 2 1 Þð 3 2 Þ the free LW energy per unit area between the surfaces; h0 is the
minimum cut-off distance due to Born repulsion; rLW is the Lifshitz–van der Waals component; e is the dielectric constant of fluid; qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi e0 is the dielectric permittivity of vacuum, 1k ¼ e0 er Rg T=ð2F 2 IS Þ is the Debye length; T is the absolute temperature in Kelvins; e0 is the permittivity of free space; ew is the dielectric constant; Rg is the gas constant; F is the Faraday’s constant; Is is the ionic strength; n1 and n3 represent the surface potentials of membrane and foulant, respectively; k is the decay length of AB interaction;
qffiffiffiffiffi pffiffiffiffiffi pffiffiffiffiffi pffiffiffiffiffi pffiffiffiffiffi qffiffiffiffiffi qffiffiffiffiffi qffiffiffiffiffi DGAB r þ2 ð r 1 þ r3 r 2 Þ þ 2 r2 ð r þ1 þ r þ3 r þ2 Þ h0 ¼ 2 qffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffi 2ð rþ1 r 3 r1 r þ3 Þ is the acid–base free energy per unit area between the surfaces at contact. The calculated surface tension parameters and free energy of cohesion for each clean membrane and SMP are shown in Supplementary Table S2. 2.6. Analytical methods Total organic carbon (TOC) concentrations in the SMP samples were measured by a TOC analyser (TOC-5000A, Shimadzu, Japan). The zeta potentials (f) and size distribution of the SMP were monitored using a zetasizer (Zetasizer 3000 HS type A, Malverin, England). Atomic force microscopy (AFM) (Veeco, Santa Barbara, CA) was used to describe foulant layer morphology in terms of membrane surface topography. The membrane roughness was determined from the AFM height images of 10 10 lm areas. The Nanoscope control software (Version 5.30r3sr3) was used for image acquisition; the mean average roughness (Ra) and root mean
114
Z. Li et al. / Bioresource Technology 140 (2013) 111–119
square (Rms) of surface roughness values were calculated from at least 7 images for each membrane sample. Contact angle measurements were performed using an NRL Contact Angle Goniometer (Rame Hart, Mountain Lakes, NJ), which is a standard goniometer with image analysis attachments (i.e., video camera, computer with monitor, and image analysis software). 3. Results and discussion 3.1. Fouling propensity of C-SMP and S-SMP with different types of membranes The C-SMP and S-SMP samples derived from the two MBRs were tested with CA, PVDF and PES membranes in a stirred cell with similar initial flux. The flux decline trends by different membranes are shown in Fig. 2. For the C-SMP, the flux of CA membrane was found to drop by 85.4% when filtration experiment stopped and about 300 mL of permeate was collected, while the PVDF and PES membranes resulted in only ca. 73.4% and 49.2% decrement, respectively. With respect to the S-SMP, the flux of CA, PVDF and PES membranes resulted in ca. 78.7%, 64.9% and 39.7% decrement, respectively. It was noted that the flux decrement of S-SMP with CA, PVDF and PES were lower compared to that of C-SMP with corresponding membranes. This meant that the predated sludge recycle could reduce the fouling propensity of the SMP regardless of the characteristics of membrane. However, for both the filtration of the C-SMP and S-SMP, the CA membrane exhibited the fastest diminishing rate of flux among the three types of membranes. Since the same SMP was filtered and similar initial flux was applied, the different rate of permeability decline among the CA, PVDF and PES could be much dependent on the characteristics of membrane surface. Regarding to the C-SMP and S-SMP filtration using the same membrane, the hydrophilic/hydrophobic interaction between the SMP and membrane may be of much importance, which was supported by previous research (Tian et al., 2013). The interfacial free energy of cohesion (DGcoh), represents the free energy (per unit area) when two surfaces of the same material are immersed in a solvent (water). The free energy of cohesion gives a qualitative representation of the hydrophobicity–hydrophilicity of organics and membranes (Van Oss et al., 2007). With respect to a hydrophobic material, the free energy of cohesion is negative, and the free energy of cohesion is positive for hydrophilic materials. It was seen from Supplementary Table S2 that the C-SMP and S-SMP both presented hydrophobic nature. The free energy of CA membrane (53.70 mJ/m2) was lower than that of PVDF (75.40 mJ/m2) and PES (62.75 mJ/m2) membranes, suggesting CA has the lowest SMP fouling tendency in terms of adsorption onto the membrane surface. But the experimental fouling data inconsistently showed that the CA fouled more severely than PVDF and PES membranes. It seemed that the characteristics of hydrophobicity may not be sufficient to predict the effect of membrane property on membrane fouling. Hence, the relative fouling behavior of the three commercial membranes should be predicted from the combined knowledge of physical surface morphology and surface chemical properties. 3.2. Influence of predated sludge recycle on the interaction energy between SMP and different clean membrane surfaces SMP adsorption and deposition on membrane surfaces are affected by the interaction energies between the SMP constituents and the membrane surface that can be predicted implementing the XDLVO theory (Ying et al., 2010). The interaction energy profiles of S-SMP and C-SMP approaching the clean CA, PVDF and
Fig. 2. Normalized flux of the three different membranes during filtration of C-SMP, S-SMP and pure water (filtration pressure: 10 kPa). J stands for flux and J0 is the initial flux.
PES membrane were calculated, exhibiting an energy barrier and a secondary energy minimum (Fig. 3). The secondary energy minimum represents the ability of SMP being sucked onto membrane surface (Hoek et al., 2003). An energy barrier means the SMP in suspension should have sufficient kinetic energy to overcome this barrier to approach the membrane (Redman et al., 2004).
Z. Li et al. / Bioresource Technology 140 (2013) 111–119
115
Fig. 3. Interaction energies profiles between the C-SMP (1) and S-SMP (2) and the clean CA (a), PVDF (b) and PES (c) membrane as a function of separation distance.
The interfacial energies profiles between the SMP and the clean CA membrane surface are shown in Fig. 3a. The C-SMP and S-SMP particle experienced attraction from 50 nm and this attractive energy reached a maximum until the particle came to 15 nm separation distance. The maximum attractive energies (secondary energy minimum) for C-SMP and S-SMP particle were significantly different, with the values of 1.78 KT and 0.72 KT, respectively.
At 8 nm separation, the repulsive electrostatic interactions dominated and the total interaction quickly became repulsive. The maximum repulsive energy (energy barrier) of the S-SMP (7.13 KT) was much higher than that of the C-SMP (3.75 KT), indicating that when the SMP was brought to the clean CA membrane surface by the convective permeate flow, the S-SMP needed to overcome the higher repulsive interaction energy than the C-SMP. At 3 nm
116
Z. Li et al. / Bioresource Technology 140 (2013) 111–119
separation, the acid–base attraction dominated and the total interaction quickly became attractive for S-SMP and C-SMP. The interaction energy curves between the SMP and the clean PVDF membrane surface are presented in Fig. 3b. The C-SMP particle encountered a weak attraction at a distance of about 50 nm due to the present of the van der Waals attraction, and then at 15 nm, the attractive energy reached a maximum (1.58 KT). Due to the dominated electrostatic repulsion, the total interaction became repulsion at 8 nm, and at 5 nm, this repulsive energy reached a maximum (4.04 KT). With respect to the S-SMP approaching the clean PVDF membrane surface, the energy barrier was 76.0% stronger than that between the C-SMP and the clean PVDF membrane surface. The second minimum in the interaction energy profile of S-SMP was 58.9% less than that of C-SMP. With respect to the interaction between the SMP and the clean PES membrane surface (Fig. 3c), the second minimum in the interaction energy profile of S-SMP was 33.9% less than that of C-SMP, however, the energy barrier (8.03 KT) was higher than that of CSMP (7.32 KT). For different clean membranes, the interaction energy profiles of S-SMP exhibited a lower second minimum and a higher energy barrier than those of C-SMP. The lower second minimum showed that the energy of S-SMP being sucked onto the clean CA, PVDF and PES membrane surface was weaker than that of C-SMP being sucked onto the corresponding membrane surface; meanwhile, the higher energy barriers indicated that the S-SMP approaching the membrane surface needed to overcome stronger repulsive interaction energy compared to the C-SMP approaching the corresponding membrane surface. This could further confirm that the
Table 4 Summary of fouled membrane roughness parameters. SMP
Fouled membrane
hwater (°)
hformamide (°)
hdiiodomethane (°)
C-SMP
CA PVDF PES
65.9 80.9 78.79
15.99 53.99 54.2
21.31 35.11 32.59
S-SMP
CA PVDF PES
64.33 78.33 74.89
20.04 51.04 50.34
23.89 28.31 34.11
RMS roughness Average (Rms) roughness (Ra)
Maximum roughness (lm)
C-SMP CA PVDF PES
183.32 342.58 263.38
114.44 251.49 196.28
1.362 2.923 1.948
S-SMP CA PVDF PES
130.81 306.26 133.78
102.87 233.79 105.31
0.986 2.546 1.238
predated sludge recycle reduced the fouling potential of the SMP regardless of the characteristics of membrane. 3.3. Influence of predated sludge recycle on the interaction energy between SMP and different fouled membrane surfaces Adsorption of SMP onto CA, PVDF and PES membranes reduced pure water contact angles and changed contact angles of formamide and diiodomethane (Table 1). This suggested that the chemistry of the membrane surfaces has been significantly altered. However, neither membrane exhibited the exact contact angle behavior of the thick SMP layers analyzed in Supplementary Table S1. It is possible that the surfaces were not completely coated with SMP and some of the clean membrane remained exposed after filtration. Calculated interfacial interaction parameters for the organic-fouled membranes are all listed in Table 2. The zeta potentials of the clean membrane were assumed for the organic-fouled membrane surfaces. Although this assumption may not be fundamentally justified, it introduced very little error in our analyses below because the electrostatic double layer interfacial free energies at contact were negligible. The membranes fouled by C-SMP and S-SMP expressed significantly different characters, but they had similar change tendency. The membranes fouled by C-SMP and S-SMP all showed a slight reduction in apolar character. The free energy of cohesion values became significantly lower after organic adsorption for CA, PVDF and PES membrane. Hence, after adsorption of C-SMP and S-SMP, CA, PVDF and PES become relatively hydrophilic compared to clean membrane surfaces. Moreover, the membranes
Table 1 Measured physicochemical properties of fouled membrane. SMP
Fouled membranes
Table 2 Calculated surface energetic parameters for fouled membranes. SMP
Fouled membranes
cLW
c+
c
DGLW sls
DGAB sls
DGEL sls
DGcoh
C-SMP
CA PVDF PES
47.39 41.98 43.12
2.92 0.18 0.06
4.32 3.74 5.31
9.8124 6.5533 7.2013
39.7078 57.6492 52.7646
0.0576 0.0548 0.069
49.4625 64.1477 59.8969
S-SMP
CA PVDF PES
46.54 44.91 42.44
2.5 0.12 0.21
5.96 4.51 6.78
9.2706 8.2617 6.8121
36.1907 55.0494 44.9215
0.0576 0.0548 0.069
45.4037 63.2563 51.6646
Table 3 Summary of DLVO interaction energy calculations of the clean and fouled membrane. SMP
Clean membranes
Energy barrier (KT)
Secondary minimum (KT)
Fouled membranes
Energy barrier (KT)
Secondary minimum (KT)
Flux declined J/J0 (%)
C-SMP
CA PVDF PES
3.75 4.04 7.32
1.78 1.59 1.21
CA PVDF PES
4.16 5.26 6.86
1.72 1.32 1.36
85.4 73.4 49.2
S-SMP
CA PVDF PES
7.13 7.13 8.03
0.72 0.64 0.80
CA PVDF PES
5.90 6.01 8.24
1.08 1.00 0.85
78.7 64.9 39.7
Z. Li et al. / Bioresource Technology 140 (2013) 111–119
fouled by S-SMP showed weaker hydrophobic than those fouled by C-SMP. The interfacial energies profiles between the SMP and the fouled membrane surface (CA, PVDF and PES) were calculated implementing the XDLVO theory, which are presented in Fig. 4. With respect to the C-SMP, the secondary energy minimum value for CA was approximately 1.72 KT, but changed into 1.32 KT for PVDF; finally, the PES had a potential well of 1.36 KT. The corre-
117
sponding secondary energy minimum values for S-SMP were 1.08 KT, 1.00 KT and 0.85 KT, respectively. The energy barriers of S-SMP with CA, PVDF and PES were respectively 39.4%, 14.2%, and 20.6% bigger than those of C-SMP with corresponding membranes. In order to explain the relationship between the interfacial energies profiles and fouling rate, the secondary energy minimums and energy barriers between the SMP and clean and fouled
Fig. 4. Interaction energies profiles between the C-SMP (1) and S-SMP (2) and the fouled CA (a), PVDF (b) and PES (c) membrane as a function of separation distance.
118
Z. Li et al. / Bioresource Technology 140 (2013) 111–119
membrane are summarized in Table 3. For the same membrane materials, the results of XDLVO theory were accordance with the SMP adhesion; the S-SMP with lower secondary energy minimums and higher energy barriers in the interaction energy profiles showed weak membrane fouling. The same membrane materials have similar property, hence, the interfacial energies is governing factor of SMP adhesion onto membrane surfaces. Apparently, worm predation could reduce fouling potential of S-SMP regardless of the membrane materials through changing the interaction energy. But, the membrane fouling rate did not follow XDLVO predictions with different membrane surfaces. The energy barriers and the secondary energy minimums for S-SMP filtration with CA and PVDF were similar. However, significant difference in fouling rate was detected between CA and PVDF membranes. Therefore, the comparability of the different fouling rates between the different membranes was not mainly determined by the interfacial energies of the membrane and SMP. It was quite possible that the presence of morphological heterogeneities can result in surfaces with very different energy distributions than smooth surfaces, and hence, can result in substantially different fouling behavior compared to XDLVO predictions (Brant and Childress, 2002). 3.4. The effect of roughness on fouling propensity of C-SMP and S-SMP with different types of membranes The surface topological properties of membranes had serious implications on membrane performance and fouling propensity (Lee et al., 2005). The surface morphologies of the CA, PVDF and PES membranes fouled by C-SMP and S-SMP, reflected by Ra and Rms were significantly different (Table 4). For example, for the PVDF membrane fouled by C-SMP, the Rms and Ra were larger than the values for virgin membranes (Supplementary Fig. S1). The increase in membrane surface roughness might be attributed to the surface enrichment of large molecules due to fouling caused by pore blocking or surface adsorption. However, the values of Rms and Ra of fouled PVDF membrane induced by S-SMP showed little lower than that of the virgin membrane The decrease of roughness might be attributed to a result of the deposition of small molecular in the ridge-and-valley structure of the membrane and in pore so as to make the membrane surface flatter and the roughness decreased. A mechanistic explanation for the pronounced influence of surface roughness on colloidal fouling emerges by analyzing the nature of membrane morphology (Vrijenhoek et al., 2001). As permeation rate through membranes is proportional to the thickness of the active (skin) layer, the bottom of a valley presents the ‘‘path-of-least resistance’’ to permeating water. Hence, water convection and particle transport are focused towards the valley bottom. Thus, as the SMP are preferentially convected towards the center of a valley, the valleys rapidly become ‘‘clogged,’’ resulting in significant loss of permeate flux. The XDLVO interaction energy between the SMP and a smooth membrane surface was significantly altered by membrane surface topology (Chen et al., 2012). Both the PVDF and PES exhibited relatively fewer but larger protuberances, compared to the CA with abundant but small asperities (Supplementary Fig. S1). It was obvious that the relative fouling behavior of the three commercial membranes could not be predicted from knowledge of surface roughness (Table 4). Large-scale surface roughness significantly increased the rate of SMP attachment through providing a larger surface area and greater contact opportunities for SMP with the membrane surface. However, in this study, the CA with many small asperities showed the fastest flux decline. The CA membrane exhibited relatively more but smaller protuberances compared to the PVDF and PES with few large asperities and peaks. According to the calculation and drawing of the interaction energy map in our previous study (Chen et al., 2012), the attractive regions on
CA membrane were obviously larger than those on the PVDF and PES membrane due to the more second minimums surrounding each protruding asperity. Although the large-scale surface roughness were absent on CA membrane, the more second minimum enhanced the attachment of SMP onto the membrane surface. Compared with the CA membrane, the PVDF membrane exhibited the larger peak-to-peak separation distance, less attractive regions around the protrusions, indicating the lower fouling rate of PVDF membrane. In contrast to other two membranes, the PES membrane was much smoother with the least asperities, leading to relatively smaller attractive regions around the positive asperities. As the SMP approaching to the mean-plane, the probability of SMP being sucked onto the membrane significantly decreased. Consequently, the different fouling rates for the different membranes were attributed to the combination of interaction energy and the number of asperities effect. 4. Conclusions The flux declines of SMP were decreased by 6.7%, 8.5% and 9.5% for CA, PVDF and PES membranes, respectively, with the predated sludge recycle. For the same membrane materials, the weakened fouling potential of SMP due to worm predation was confirmed by the bigger energy barriers and lower secondary energy minimum. With respect to different membrane surfaces, the second minimums surrounding each protruding asperity on CA membrane were more than those on the PVDF and PES membrane. The different fouling rates between the different membranes were determined by the combination of interaction energy and the number of asperities. Acknowledgments This study was supported by the Major Science and Technology Program for Water Pollution Control and Management (No. 2013ZX07201007), the State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology (No. 2011DX01) and the Specialized Research Fund for the Doctoral Program of Higher Education (No. 20112302110060). The authors also appreciate the Funds for Creative Research Groups of China (No. 51121062). 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.2013. 04.078. References Bhattacharjee, S., Ko, C.H., Elimelech, M., 1998. DLVO interaction between rough surfaces. Langmuir 14, 3365–3375. Brant, J.A., Childress, A.E., 2002. Assessing short-range membrane–colloid interactions using surface energetics. J. Membr. Sci. 203, 257–273. Chen, L., Tian, Y., Cao, C.-Q., Zhang, J., Li, Z.-N., 2012. Interaction energy evaluation of soluble microbial products (SMP) on different membrane surfaces: role of the reconstructed membrane topology. Water Res. 46, 2693–2704. Drews, A., 2010. Membrane fouling in membrane bioreactors—Characterisation, contradictions, cause and cures. J. Membr. Sci. 363, 1–28. Elimelech, M., Zhu, X., Childress, A.E., Hong, S., 1997. Role of membrane surface morphology in colloidal fouling of cellulose acetate and composite aromatic polyamide reverse osmosis membranes. J. Membr. Sci. 127, 101–109. Evenblij, H., Van der Graaf, J., 2004. Occurrence of EPS in activated sludge from a membrane bioreactor treating municipal wastewater. Water Sci. Technol. 50, 293–300. Geng, Z., Hall, E.R., 2007. A comparative study of fouling-related properties of sludge from conventional and membrane enhanced biological phosphorus removal processes. Water Res. 41, 4329–4338. Hoek, E.M., Bhattacharjee, S., Elimelech, M., 2003. Effect of membrane surface roughness on colloid–membrane DLVO interactions. Langmuir 19, 4836–4847.
Z. Li et al. / Bioresource Technology 140 (2013) 111–119 Hwang, B.-K., Kim, J.-H., Ahn, C.H., Lee, C.-H., Song, J.-Y., Ra, Y.-H., 2010. Effect of disintegrated sludge recycling on membrane permeability in a membrane bioreactor combined with a turbulent jet flow ozone contactor. Water Res. 44, 1833–1840. Kang, S.T., Subramani, A., Hoek, E., Deshusses, M.A., Matsumoto, M.R., 2004. Direct observation of biofouling in cross-flow microfiltration: mechanisms of deposition and release. J. Membr. Sci. 244, 151–165. Le-Clech, P., Chen, V., Fane, T.A.G., 2006. Fouling in membrane bioreactors used in wastewater treatment. J. Membr. Sci. 284, 17–53. Lee, N., Amy, G., Croué, J.-P., Buisson, H., 2005. Morphological analyses of natural organic matter (NOM) fouling of low-pressure membranes (MF/UF). J. Membr. Sci. 261, 7–16. 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. Ng, T.C.A., Ng, H.Y., 2010. Characterisation of initial fouling in aerobic submerged membrane bioreactors in relation to physico-chemical characteristics under different flux conditions. Water Res. 44, 2336–2348. Ng, H.Y., Tan, T.W., Ong, S.L., 2006. Membrane fouling of submerged membrane bioreactors: impact of mean cell residence time and the contributing factors. Environ. Sci. Technol. 40, 2706–2713. Oh, Y.K., Lee, K.R., Ko, K.B., Yeom, I.T., 2007. Effects of chemical sludge disintegration on the performances of wastewater treatment by membrane bioreactor. Water Res. 41, 2665–2671. Park, N., Kwon, B., Kim, I.S., Cho, J., 2005. Biofouling potential of various NF membranes with respect to bacteria and their soluble microbial products (SMP): characterizations, flux decline, and transport parameters. J. Membr. Sci. 258, 43–54. Redman, J.A., Walker, S.L., Elimelech, M., 2004. Bacterial adhesion and transport in porous media: role of the secondary energy minimum. Environ. Sci. Technol. 38, 1777–1785. Shariati, S.R.P., Bonakdarpour, B., Zare, N., Ashtiani, F.Z., 2011. The effect of hydraulic retention time on the performance and fouling characteristics of membrane sequencing batch reactors used for the treatment of synthetic petroleum refinery wastewater. Bioresour. Technol. 102, 7692–7699. Subramani, A., Hoek, E., 2008. Direct observation of initial microbial deposition onto reverse osmosis and nanofiltration membranes. J. Membr. Sci. 319, 111–125. Sun, F.-Y., Wang, X.-M., Li, X.-Y., 2011. Change in the fouling propensity of sludge in membrane bioreactors (MBR) in relation to the accumulation of biopolymer clusters. Bioresour. Technol. 102, 4718–4725.
119
Tian, Y., Lu, Y., Chen, L., Lin, H., 2010. Optimization of process conditions with attention to the sludge reduction and stable immobilization in a novel Tubificidae-reactor. Bioresour. Technol. 101, 6069–6076. Tian, Y., Chen, L., Zhang, S., Cao, C., Zhang, S., 2011. Correlating membrane fouling with sludge characteristics in membrane bioreactors: an especial interest in EPS and sludge morphology analysis. Bioresour. Technol. 102, 8820–8827. Tian, Y., Li, Z., Chen, L., Lu, Y., 2012a. Role of extracellular polymeric substances (EPSs) in membrane fouling of membrane bioreactor coupled with worm reactor. Bioresour. Technol. 123, 566–573. Tian, Y., Li, Z., Lu, Y., 2012b. Changes in characteristics of soluble microbial products and extracellular polymeric substances in membrane bioreactor coupled with worm reactor: relation to membrane fouling. Bioresour. Technol. 122, 62–69. Tian, Y., Lu, Y., Li, Z., 2012c. Performance analysis of a combined system of membrane bioreactor and worm reactor: wastewater treatment, sludge reduction and membrane fouling. Bioresour. Technol. 121, 176–182. Tian, Y., Li, Z., Ding, Y., Lu, Y., 2013. Identification of the change in fouling potential of soluble microbial products (SMP) in membrane bioreactor coupled with worm reactor. Water Res. 47, 2015–2024. Van Oss, C.J., 2006. Interfacial Forces in Aqueous Media. CRC. van Oss, C.J., 2007. Development and applications of the interfacial tension between water and organic or biological surfaces. Colloids Surf., B 54, 2–9. Vrijenhoek, E.M., Hong, S., Elimelech, M., 2001. Influence of membrane surface properties on initial rate of colloidal fouling of reverse osmosis and nanofiltration membranes. J. Membr. Sci. 188, 115–128. Wang, S., Guillen, G., Hoek, E.M.V., 2005. Direct observation of microbial adhesion to membranes. Environ. Sci. Technol. 39, 6461–6469. Wong, P.C.Y., Kwon, Y.-N., Criddle, C.S., 2009. Use of atomic force microscopy and fractal geometry to characterize the roughness of nano-, micro-, and ultrafiltration membranes. J. Membr. Sci. 340, 117–132. Wu, J., Huang, X., 2010. Use of ozonation to mitigate fouling in a long-term membrane bioreactor. Bioresour. Technol. 101, 6019–6027. Xuan, W., Bin, Z., Zhiqiang, S., Zhigang, Q., Zhaoli, C., Min, J., Junwen, L., Jingfeng, W., 2010. The EPS characteristics of sludge in an aerobic granule membrane bioreactor. Bioresour. Technol. 101, 8046–8050. Ying, W., Yang, F., Bick, A., Oron, G., Herzberg, M., 2010. Extracellular polymeric substances (EPS) in a hybrid growth membrane bioreactor (HG-MBR): viscoelastic and adherence characteristics. Environ. Sci. Technol. 44, 8636– 8643.