Combined coagulation-disk filtration process as a pretreatment of ultrafiltration and reverse osmosis membrane for wastewater reclamation: An autopsy study of a pilot plant

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 8 0 3 e1 8 1 6

Available online at www.sciencedirect.com

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

Combined coagulation-disk filtration process as a pretreatment of ultrafiltration and reverse osmosis membrane for wastewater reclamation: An autopsy study of a pilot plant Kangmin Chon a, Seung Joon Kim b, Jihee Moon c, Jaeweon Cho a,* a

School of Environmental Science and Engineering, Gwangju Institute of Science and Technology (GIST), 261 Cheomdan-gwagiro, Buk-gu, Gwangju 500-712, Republic of Korea b Environmental Research Team, Research & Development Institute, Kolon Engineering & Construction, 199-5 Jeondae-ri, Pogok-eup, Cheoin-gu, Yongin, Gyeonggi 449-815, Republic of Korea c Construction Technology Examination Division, Machinery Metals & Construction Examination Bureau, Government Complex-Daejeon, 139 Seonsa-ro, Seo-gu, Daejeon 302-701, Republic of Korea

article info

abstract

Article history:

The effects of the combined coagulation-disk filtration (CCeDF) process on the fouling

Received 19 September 2011

characteristics and behavior caused by interactions between effluent organic matter

Received in revised form

(EfOM) and the membrane surfaces of the ultrafiltration (UF) and reverse osmosis (RO)

7 December 2011

membranes in a pilot plant for municipal wastewater reclamation (MWR) were investi-

Accepted 29 December 2011

gated. The feed water from secondary effluents was treated by the CCeDF process used as

Available online 15 January 2012

a pretreatment for the UF membrane to mitigate fouling formation and the permeate from the CCeDF process was further filtered by two UF membrane units in parallel arrangement

Keywords:

and fed into four RO modules in a series connection. The CCeDF process was not sufficient

Combined coagulation-disk filtra-

to mitigate biofouling but the UF membrane was effective in mitigating biofouling on the

tion process

RO membrane surfaces. Fouling of the UF and RO membranes was dominated by hydro-

Effluent organic matter

philic fractions of EfOM (e.g., polysaccharide-like and protein-like substances) and inor-

Membrane autopsy

ganic scaling (e.g., aluminum, calcium and silica). The desorbed UF membrane foulants

Ultrafiltration

included more aluminum species and hydrophobic fractions than the desorbed RO

Reverse osmosis membrane

membrane foulants, which was presumably due to the residual coagulants and aluminum-

Municipal wastewater reclamation

humic substance complexes. The significant change in the surface chemistry of the RO membrane (a decrease in surface charge and an increase in contact angle of the fouled RO membranes) induced by the accumulation of hydrophilic EfOM onto the negatively charged RO membrane surface intensified the fouling formation of the fouled RO membrane by hydrophobic interaction between the humic substances of EfOM with relatively high hydrophobicity and the fouled RO membranes with decreased surface charge and increased contract angle. ª 2012 Elsevier Ltd. All rights reserved.

* Corresponding author. Tel.: þ82 62 715 2443; fax: þ82 62 715 2434. E-mail addresses: [email protected] (K. Chon), [email protected] (J. Cho). 0043-1354/$ e see front matter ª 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2011.12.062

1804

1.

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 8 0 3 e1 8 1 6

Introduction

Sustainable water resource development is an emerging issue in water and/or wastewater treatment due to the serious water shortage problems induced by the combination of rapid population growth from urbanization, industrialization and climate changes (Asano et al., 2007). Therefore, various membrane processes, including microfiltration (MF), ultrafiltration (UF), membrane bioreactor (MBR), nanofiltration (NF), and reverse osmosis (RO) membrane, have recently gained recognition as viable technologies for the sustainable water resource development (e.g., wastewater reclamation, seawater and brackish water desalination) due to their reliable water quality and cost effectiveness (Bellona and Drewes, 2007; Kamp et al., 2000). However, membrane fouling, which can cause serious flux decline and affect the produced water quality, is still a major concern in the application of membrane technology to sustainable water resource development. The fouling characteristics and behavior on the membrane surfaces can be determined by feed water characteristics (e.g., pH and ionic strength, levels of key constituents), membrane properties (e.g., contact angle, surface charge, roughness and molecular weight cut off (MWCO)), operating conditions (e.g., permeate flux and recovery rate), and pretreatment procedures (Xu et al., 2010; Kim et al., 2008). Furthermore, the fouling mechanisms of membrane processes can be governed by the concentration of possible membrane foulants in the feed water due to cake layer formation (particulate and colloidal materials), concentration polarization (dissolved inorganic materials), electrostatic repulsion and/or hydrophobic interactions (dissolved organic materials), and biofouling (microbial attachment and growth) (Gabelich et al., 2005; Boerlage et al., 1999; Jarusutthirak et al., 2002; Ivnitskya et al., 2005). Although many studies have applied laboratory-scale membrane fouling testing to increase understanding of fouling development in membrane processes for municipal wastewater reclamation (MWR; Jarusutthirak et al., 2002; Lee et al., 2006), the mechanism of fouling formation has not been clearly verified because laboratory-scale fouling experiments conducted by measuring flux decline as a function of time under controlled operating conditions cannot truly reflect the fouling phenomenon occurring in practical applications of membrane processes (i.e., pilot-scale and fullscale). Therefore, practical membrane applications are required to consider variations in pretreatment efficiency, water characteristics, and operating conditions. In addition, autopsies of fouled membranes need to be conducted to elucidate fouling characteristics and behavior associated with feed water characteristics and membrane properties (Schneider et al., 2005). However, autopsies of fouled membranes from pilot-scale and/or full-scale membrane applications have rarely been performed due to their high cost and lengthy duration. Most of the previous research on membrane autopsies was mainly focused on the seawater or brackish water desalination processes (Tran et al., 2007; AlAmoudi and Faroque, 2005). Hence, knowledge on the fundamentals of membrane fouling remains incomplete, including dominant types of fouling and possible fouling

mechanisms in the practical applications of membrane processes for MWR. Effluent organic matter (EfOM) containing large fractions of soluble microbial products produced during biological wastewater treatment (e.g., activated sludge and MBR) has been identified as a major foulant of membrane processes for MWR (Chon et al., 2011; Jarusutthirak et al., 2002; Lee et al., 2006). In addition, recent research reported that hydrophilic fractions of EfOM and colloidal materials that are able to pass through pretreatment play critical roles in the severe fouling of high-pressure membranes (e.g., NF and RO membranes) depending on the type of pretreatment (Li et al., 2007; Jarusutthirak and Amy, 2006). Although it is important to understand the interactions between possible membrane foulants and membrane surfaces, variations in the fouling mechanism have not been adequately investigated due to the inherent heterogeneity and complexity of EfOM. Pretreatment is known as a key step to control the fouling of highpressure membranes by EfOM. During the past decade, coagulation and low-pressure membrane processes (e.g., MF and UF membranes) have been widely used as a pretreatment of the high-pressure membrane. Therefore, most previous studies have mainly focused on the fouling characteristics of high-pressure membranes fed with the coagulation of MF and/or UF membrane-treated waters. The granular activated carbon process was effective in reducing the organic fouling of RO membranes (Gur-Reznik et al., 2008) and the flux decline of NF membranes was positively affected by the hydrophilic fractions of feed water treated by an MF membrane (Kim et al., 2007). The major foulants of NF and RO membranes fed with the UF membrane-treated water were calcium bound to inorganic materials and silica bound with organic materials (Gwon et al., 2003). However, no comprehensive study has yet been performed on the fouling characterization of RO and UF membranes receiving secondary effluents treated by the combined coagulation-disk filtration (CCeDF) process. The main objective of this study was to evaluate the viability of the CCeDF process as a pretreatment of UF and RO membranes and the effects of this process on the fouling formation by EfOM in order to provide valuable insight into the fouling characteristics and behavior of UF and RO membranes in an MWR pilot plant. The changes in membrane surface chemistry and morphology associated with varying feed water characteristics (e.g., organic and inorganic constituents) caused by the CCeDF process were rigorously analyzed through membrane autopsies and directly correlated to the observed fouling characteristics to elucidate the fouling behavior and dominant fouling mechanisms of UF and RO membranes in the MWR pilot plant.

2.

Materials and methods

2.1. Description of the pilot plant for municipal wastewater reclamation (MWR) The MWR pilot plant was operated for nearly 1 year under continuous operating conditions at Ansan wastewater

1805

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 8 0 3 e1 8 1 6

treatment plant (WWTP) (Ansan, Korea) with an anaerobicanoxic-oxic (A2O) activated sludge process for biological removal of nitrogen and phosphorus. In this work, the CCeDF process and UF membrane were used as pretreatments to minimize fouling of the RO membrane, as shown in Fig. 1. Prior to passage through the RO membranes, secondary effluents (498.5 m3/day) from Ansan WWTP were coagulated using aluminum sulfate, directly filtered by DF with double layers of polyester open mesh fabrics (PETEX, Sefar, Heiden, Switzerland), and then further treated by two polyvinylidene fluoride (PVDF) hollow fiber UF membrane modules, with a nominal pore size of 0.1 mm (Cleanfil-75R, Kolon Membrane Corporation, Gwacheon, Korea), that were arranged in parallel, operated in dead-end mode (recovery rate: 92%), backwashed using NaOCl (concentration: 13.2 mg/L) with an air stripping every 30 min for 30 s to remove aggregates on the surface of the UF membrane, subsequently drained for 20 s to remove any remaining NaOCl solution and then refilled with fresh water. Permeated water from the UF membrane modules was additionally filtered through a 4-stage RO system (recovery rate: 73%) using spiral wound polyamide type thinfilm composite (TFC) membrane modules with MWCO in the range from 100 to 400 daltons (Da) (FL, Woongjin Chemical Corporation, Seoul, Korea) (Lee et al., 2008).

2.2.

Preparation of samples

2.2.1.

Collection of water samples

Four different water samples, including secondary effluents of Ansan WWTP (Feed water), the permeate from coagulation-DF (DF permeate), the permeate from the UF membrane (UF permeate), and the permeate from the RO membrane (RO permeate), were collected weekly from the tested MWR pilot plant to investigate the removal of particulate and/or dissolved water contaminants through the pilot system and the characteristics of EfOM in the feed and treated waters. Prior to all analyses, the feed and treated water samples were filtered with mixed cellulose ester MF filters (nominal pore size: 0.45 mm) (Advantec, Tokyo, Japan).

2.2.2.

Virgin and fouled membranes

Virgin UF and RO membranes were rinsed using deionized (DI) water and soaked in DI water for 2 days to remove membranecoating materials (i.e., humectants). The fouled UF membranes were collected at the inlet of the UF module because more serious fouling was found in the inlet region than in the outlet region. The fouled RO membranes were collected from both the first (ROi) and fourth (ROf) stages of the RO modules to investigate differences in fouling characteristics and behavior according the stage of the RO system. The soaked virgin and fouled UF and RO membranes were dried in

Feed water Secondary Effluents of Ansan WWTP

a closed desiccator for 3 days for analysis of the membrane characteristics.

2.2.3.

Extraction of membrane fouling

Three different solutions (DI water, 0.1 M NaOH, and 0.1 M HCl) were used to desorbed foulants from the fouled UF and RO membranes. The fouled UF (effective surface area: 320 cm2) and RO (effective surface area: 570 cm2) membranes were consecutively soaked in 500 ml each of DI water, 0.1 M NaOH, and 0.1 M HCl for 6 h each with moderate stirring to desorb foulants from the fouled UF and RO membranes according to their physicochemical properties: UFeDI, UF-base, UF-acid, ROieDI, ROi-base, ROi-acid, ROfeDI, ROf-base, and ROf-acid. The pH of the desorbed UF and RO membranes foulants was adjusted using NaOH and HCl to a range of 5.0e7.0 and then pretreated using glass fiber filters (pore size: 0.7 mm) (GF/F, Whatman, Clifton, NJ, USA).

2.3.

Analytical methodologies

The chemical oxygen demand of the water samples was measured by oxidation of organic materials using trivalent manganese, the concentration of dissolved organic carbon (DOC) and total nitrogen (TN) was determined by a total organic carbon analyzer (TOC-VCPH, Shimadzu, Kyoto, Japan) equipped with a TN analyzer (TNM-1, Shimadzu, Kyoto, Japan) using a combustion catalytic oxidation method, and the concentration of total phosphorus was determined by the acid persulfate digestion method at 890 nm using DR/2010 spectrophotometer (Hach, Loveland, Colorado, USA). The UV absorbance at 254 nm (UVA254) was measured with a 1 cm quartz cell (Hellma, Mu¨llheim, Germany) to evaluate the aromatic contents of EfOM in water and the foulant samples and the specific UV absorbance (SUVA) value (as an indicator of aromaticity) was calculated from the ratio of UVA254 to DOC concentration. The ammonium, calcium, nitrate, nitrite, magnesium and phosphate and were quantified using an ion chromatograph (DX-120, ICS-90, Dionex, Sunnyvale, CA, USA) equipped with a IonPac CS12A and a IonPac AS14 column (Dionex, Sunnyvale, CA, USA) and the amount of organic nitrogen was calculated from the difference between the concentration of TN and the sum of nitrate, nitrite, and ammonium. Inductively coupled plasmamass spectrometry (ICPeMS) (7500ce, Agilent, Santa Clara, CA, USA) was used to identify the concentration of inorganic materials in the water and foulant samples. The EfOM in water and desorbed foulant samples was characterized using highperformance size-exclusion chromatography (HPSEC) equipped with a UV detector at 254 nm (SPD-10AVP, Shimadzu, Kyoto, Japan), a fluorescence detector at an excitation (Ex) wavelength of 278 nm and an emission (Em) wavelength of 353 nm (RF-10AxL, Shimadzu, Kyoto, Japan), a Protein-Pak 125 column (Waters, Milford, MA, USA) (Her et al., 2002) and three-

Combined coaulation and disk filtration process Coagulation Settling Tank (Aluminum sulfate)

Distk Filtration First layer (15 µ m)

Second layer (10 µ m)

Reverse osmosis membrane Ultrafiltration (0.1 µ m) First stgae

Second stgae

Third Fourth stgae stgae

Ultrafiltration (0.1 µ m)

2 2 (Final concentration: 3 mg/L) (Effective surface area: 10.5 m ) (Effective surface area of each module: 73 m )

(Effective surface area of each module: 37.2 m2)

Fig. 1 e Schematic diagram of the MWR pilot plant comprising of the CCeDF process, UF, and RO membrane.

1806

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 8 0 3 e1 8 1 6

dimensional fluorescence Ex-Em matrix (3D FEEM) (fluorescence spectrophotometer, F-2500, Hitachi, Tokyo, Japan) at a photomultiplier tube (TMP) voltage of 700 V with Ex and EM slit width of 5 nm. Fractionation using Amberlite XAD-8/4 resins (Supelco, Bellefonte, PA, USA) and Fourier transform infrared (FTIR) spectroscopy (FT/IR-460 plus, Jasco, Tokyo, Japan) with a KBr pellet (Pike, Madison, WI, USA) were used to identify the relative hydrophobicity/hydrophilicity and functional group compositions of the desorbed membrane foulants. The surface morphological features of the virgin and fouled membranes were investigated using atomic force microscopy in non-contact mode (XE-100, Park System, Suwon, Korea) and a field emission scanning electron microscope (FE-SEM) (S-4700, Hitachi, Tokyo, Japan). An energy dispersive X-ray spectrometer (EDS) (7200-H, Horiba, Kyoto, Japan) was also used to identify the characteristics of the inorganic foulants on the membrane surface. Prior to the FESEM/EDS analysis, the virgin and fouled UF membrane fiber and RO membrane specimens were coated with platinum for 65 s using ion sputter (E-1030, Hitachi, Tokyo, Japan) to minimize the electron-charging effects on the membrane surfaces. An electrophoretic light scattering spectrophotometer (ELS8000, Otsuka Electronics, Osaka, Japan) with a 10 mM KCl electrolyte solution (SigmaeAldrich, St. Louis, MO, USA) and polystyrene latex particles (size: 520 nm) (Otsuka Electronics,

Osaka, Japan) was used to measure the zeta potential of the membrane surfaces at pH 7. The hydrophobicity of the membrane surfaces (i.e., contact angle) was measured by a goniometer (Rame-Hart, Mountain Lakes, NJ, USA) using the sessile drop method. Briefly, a 1.0 ml water droplet was dropped onto the dried membrane surface and then the contact angle was quickly measured within approximately 10 s.

3.

Results and discussion

3.1.

Water characteristics

The feed water characteristics, in terms of pH, ionic strength, hydrophobicity and/or hydrophilicity of organic matter, and presence and concentration of metal ions, have a great influence on the fouling characteristics and behavior of membrane processes. Therefore, changes in water quality through the CCeDF process, and UF and RO membranes were characterized using various analytical methods, as summarized in Table 1. Significant changes in the water quality were observed in the aluminum concentration. The aluminum concentration of the feed water was increased from 12.00 mg/L to 28.68 mg/L after the CCeDF process due to the residual

Table 1 e Changes in water characteristics through the MWR pilot plant (n [ 3). Feed water

DF permeate

UF permeate

RO permeate

Bulk parameters

pH Conductivity (mS/cm) CODMn (mgO2/L) DOC (mgC/L) Turbidity (NTU) UVA254 (cm1) SUVA (L mg1 m1)

7.3 333.8 (78.4) 3.69 (0.39) 3.83 (1.02) 0.94 (0.10) 0.096 (0.08) 2.64 (0.76)

7.5 329.0 (69.0) 3.71 (0.46) 4.28 (0.31) 0.68 (0.01) 0.091 (0.010) 2.14 (0.40)

7.5 271.0 (61.9) 3.01 (0.54) 3.87 (0.36) 0.03 (0.00) 0.076 (0.016) 1.96 (0.28)

6.4 23.9 (2.0) 0.17 (0.04) 0.28 (0.8) N.D. 0.006 (0.001) 2.37 (0.51)

Nutrients

TN (mgN/L) Nitrite (mgN/L) Nitrate (mgN/L) Ammonium (mgN/L) ON (mgN/L) org-N/C (M) TP (mgP/L) Phosphate (mgP/L)

7.11 (0.64) N.D. 6.83 (0.64) 0.00 (0.00) 0.28 (0.04) 0.07 (0.02) 1.65 (0.13) 1.13 (0.21)

7.38 (0.77) N.D. 6.98 (0.57) 0.00 (0.00) 0.40 (0.20) 0.08 (0.04) 1.51 (0.13) 0.84 (0.27)

6.86 (0.96) 0.00 (0.01) 6.54 (0.71) 0.00 (0.00) 0.31 (0.25) 0.07 (0.06) 1.24 (0.18) 0.80 (0.32)

0.96 (0.13) N.D. 1.03 (0.11) 0.00 (0.00) N.D. N.A. 0.04 (0.04) 0.02 (0.03)

Metals

Aluminum (mg/L) Calcium (mg/L) Cadmium (mg/L) Chromium (mg/L) Cobalt (mg/L) Copper (mg/L) Iron (mg/L) Magnesium (mg/L) Manganese (mg/L) Nickel (mg/L) Zinc (mg/L)

12.00 (4.46) 29.08 (4.19) 0.37 (0.25) 0.67 (0.37) 1.85 (0.25) 1.44 (0.51) 29.27 (0.87) 7.60 (1.16) 10.42 (2.29) 9.44 (2.74) 20.77(1.20)

28.68 (7.44) 29.19 (4.22) 0.32 (0.11) 0.45 (0.17) 1.72 (0.20) 1.48 (0.52) 6.85 (1.48) 7.64 (1.16) 9.51 (2.12) 9.03 (2.53) 13.30 (1.59)

17.88 (3.05) 26.15 (3.81) 0.41 (0.16) 0.47 (0.21) 1.65 (0.25) 1.20 (0.21) 7.90 (1.23) 6.87 (1.05) 9.48 (2.72) 8.19 (1.97) 13.18 (3.48)

3.46 (0,68) 0.27 (0.04) 0.16 (0.12) 0.14 (0.11) 0.14 (0.09) 0.41 (0.16) 0.92 (1.08) N.D. 0.30 (0.16) 0.14 (0.11) 3.43 (1.30)

Metalloids

Arsenic (mg/L) Boron (mg/L) Silica (mg/L)

1.72 (0.33) 48.94 (2.84) 7.34 (0.92)

1.53 (0.12) 50.65 (2.26) 7.45 (0.87)

1.37 (0.41) 49.92 (2.39) 6.70 (0.69)

0.32 (0.13) 41.60 (1.73) 0.28 (0.02)

N.D., not detected; N.A., Not available.

1807

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 8 0 3 e1 8 1 6

Feed water DF permeate UF permeate RO permeate

700 780

21000 UV response (mV)

b

28000

80000

1,270 1,680

14000

7000

Feed water DF permeate UF permeate RO permeate

840 1,000

Fluorescence response (mV)

a

60000

1,200 1,810 3,660

40000

26,100 490

20000

0

0

100

1000

10000

100

1000

10000

100000

Molecular weight (Da)

Molecular weight (Da)

Fig. 2 e Changes in the MW distribution through the MWR pilot plant: (a) aromatic substances and (b) protein-like substances.

coagulants (aluminum sulfate) but it was decreased to 17.88 mg/L through the UF membrane, indicating that the large amount of aluminum (>38%) from the residual coagulants was in the form of high molecular weight (MW) complexes

a

b

Maximum intensity: 2643

600

which were larger than pore size of the UF membrane (>0.1 mm). Therefore, it is expected that the aluminum complexation by humic substances has a high fouling potential for the UF membrane.

550

Maximum intensity: 2665

600 550

Humic-like fluorescence

Humic-like fluorescence

(Ex= 330 nm and Em= 410 nm)

500 Excitation (nm)

Excitation (nm)

500 450 400 350

450 400 350

300

300

250

250 250

300

350

400

450

500

550

(Ex= 330 nm and Em= 410 nm)

600

250

300

Emission (nm)

c

550

400

450

500

550

600

500

550

600

Emission (nm)

d

Maximum intensity: 2186

600

350

600 550

Humic-like fluorescence (Ex= 330 nm and Em= 410 nm)

500 Excitation (nm)

Excitation (nm)

500 450 400 350

450 400 350

300

300

250

250 250

300

350

400

450

Emission (nm)

500

550

600

250

300

350

400

450

Emission (nm)

Fig. 3 e 3D fluorescence spectra of (a) Feed water, (b) DF permeate, (c) UF permeate, and (d) RO permeate.

1808

Table 2 e Foulant characteristics of the UF and RO membranes from the MWR pilot plant (n [ 3). UFeDI 2 a

DOC (mgC/m ) UVA254 (cm1)a SUVA (L mg1 m1)a DOC fractions (%)a

Nitrogen properties

TN (mgN/m2)a Nitrite (mgN/m2) Nitrate (mgN/m2) Ammonium (mgN/m2) ON (mgN/m2) Org-N/C (M)

Metals

Aluminum (mg/m2) Calcium (mg/m2) Cadmium (mg/m2) Chromium (mg/m2) Cobalt (mg/m2) Copper (mg/m2) Iron (mg/m2) Lead (mg/m2) Magnesium (mg/m2) Manganese (mg/m2) Nickel (mg/m2) Zinc (mg/m2)

5401.3 57.5 0.6 5.4 2.5 207.6 398.2 1.3 14.0 23.1 26.9 1402.5

Metalloids

Arsenic (mg/m2) Boron (mg/m2) Silica (mg/m2)

7.7 (0.9) 1367.9 (9.4) 147.4 (1.2)

N.D., Not detected; N.A., Not available. a n ¼ 4.

134.1 0.065 0.75 52.8

(1.3) (0.000) (0.01) (0.3)

41.0(0.6) 0.8 (0.1) 5.7 (0.1) 34.7 (0.4) 0.1 (0.8) 0.00 (0.00) (48.1) (0.7) (0.1) (0.1) (0.1) (2.2) (4.5) (0.1) (0.5) (0.6) (1.5) (29.5)

106.1 (0.4) 0.078 (0.001) 1.15 (0.01) 41.8 (0.0)

UF-acid 13.6 0.014 1.58 5.4

(0.8) (0.001) (0.14) (0.3)

ROieDI 102.7 (0.5) 0.068 (0.001) 0.58 (0.01) 47.2 (0.2)

20.0 (0.4) N.D. 4.3 (0.1) N.D. 15.6 (0.2) 0.13 (0.00)

4.1 (0.5) N.D. 2.4 (0.1) N.D. 1.5 (0.4) 0.09 (0.02)

16.7 (0.3) N.D. 0.6 (0.0) 9.7 (0.8) 6.3 (1.0) 0.05 (0.01)

841.5 (6.9) 31.5 (0.4) 3.8 (0.2) 16.8 (0.3) 2.9 (0.1) 170.0 (0.9) 114.8 (9.4) 1.7 (0.1) 7.7 (0.3) 56.7 (1.5) 29.3 (0.2) 2016.8 (29.8)

210.2 (7.6) 64.5 (1.6) 0.4 (0.1) 14.4 (0.2) 1.6 (0.1) 41.7 (0.2) 17.6 (0.6) N.D. 6.2 (0.1) 177.3 (1.7) 39.4 (1.1) 718.8 (11.9)

670.1 (7.0) 2.8 (0.1) 0.3 (0.0) 311.8 (1.5) 1.1 (0.0) 24.8 (0.6) 2876.0 (23.6) 3.5 (0.2) 0.6 (0.0) 13.9 (0.6) 11.6 (0.5) 480.0 (6.5)

7.9 (0.9) 1201.4 (6.6) 606.2 (5.2)

5.3 (0.4) 209.9 (2.6) 149.9 (0.6)

5.7 (0.9) 718.8 (14.7) 24.6 (0.2)

ROi-base 108.8 0.104 0.84 50.1

(0.4) (0.001) (0.00) (0.2)

14.5 (0.4) N.D. 0.2 (0.0) N.D. 14.3 (0.5) 0.11 (0.00) 3051.2 7.5 0.2 422.2 1.9 44.0 660.7 1.7 0.5 52.8 21.5 955.4

(53.7) (0.3) (0.1) (0.7) (0.1) (1.0) (2.1) (0.0) (0.0) (0.6) (0.4) (21.5)

15.7 (0.4) 648.2 (0.6) 181.4 (1.6)

ROi-acid 5.8 (0.1) 0.006 (0.000) 0.89 (0.04) 2.7 (0.0) N.D. N.D. 0.2 (0.0) N.D. N.D. N.A. 60.6 (7.5) 8.3 (0.5) 0.6 (0.1) 14.4 (0.1) 2.9 (0.1) 29.5 (0.4) 46.1 (0.3) 2.6 (0.0) 0.5 (0.1) 257.4 (3.0) 43.4 (1.6) 1089.0 (7.2) 2.6 (0.2) 179.9 (4.3) 10.0 (0.1)

ROfeDI 60.3 0.040 0.58 33.2

(0.2) (0.000) (0.00) (0.0)

9.5 (0.2) N.D. 0.9 (0.1) 3.8 (0.5) 4.9 (0.6) 0.07 (0.01) 526.1 6.1 0.1 32.0 0.6 29.0 333.1 0.7 1.2 4.8 10.5 501.9

ROf-base 108.3 (0.5) 0.083 (0.001) 0.67 (0.01) 59.8 (0.1) 14.5 (0.3) N.D. 0.1 (0.1) N.D. 14.2 (0.2) 0.11 (0.00)

(8.3) (0.1) (0.1) (0.3) (0.0) (0.9) (4.9) (0.1) (0.0) (0.1) (0.8) (34.0)

2212.3 (26.5) 12.1 (0.1) 0.4 (0.0) 97.1 (1.2) 1.3 (0.0) 31.5 (0.6) 250.9 (0.5) 1.1 (0.1) 0.6 (0.0) 35.7 (0.9) 42.3 (0.5) 1776.2 (8.3)

2.3 (0.4) 464.9 (14.7) 28.3 (0.6)

5.3 (0.1) 858.0 (17.0) 379.8 (49.3)

ROf-acid 12.6 0.011 0.78 7.0

(0.2) (0.001) (0.05) (0.1)

1.1 (0.5) N.D. 0.2 (0.1) N.D. 1.0 (0.4) 0.07 (0.03) 154.8 29.1 0.4 21.6 2.9 23.2 161.6 0.4 1.6 216.2 58.4 1634.4

(2.9) (0.6) (0.0) (0.0) (0.1) (0.1) (3.1) (0.1) (0.0) (1.8) (0.2) (8.8)

2.1 (0.1) 303.2 (15.1) 86.9 (3.9)

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 8 0 3 e1 8 1 6

Bulk parameters

UF-base

1809

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 8 0 3 e1 8 1 6

Changes in the MW distribution of the aromatic and proteinlike substances through the pilot plant are represented in Fig. 2. The aromatic substances of EfOM were mainly low MW (700, 780, 1270 and 1680 Da), with the highest intensity of UV detection at 700 Da. In the case of protein-like substances, both low MW (490, 840, 1000, 1200, 1810 and 3660 Da) and high MW (26,100 Da) were found to be major fractions for the protein-like substances of EfOM in the pilot plant. The highest intensity of fluorescence detection corresponded to the 840 Da fraction. The aromatic substances of EfOM (700, 780, 1270 and 1680 Da) could be representative of humic substances because the UV detection for humic substances, including Suwannee River humic acids (SRHA) and Suwannee River fulvic acids (SRFA), revealed a low MW range (<4000 Da) while most of the protein-like substances may have been derived from nonhumic substances as humic substances showed very weak fluorescence intensities at an Ex wavelength of 278 nm and an Em wavelength of 353 nm (Chon and Cho, 2008). HPSEC analysis of EfOM for water samples revealed the CCeDF process to be insufficient for EfOM removal. However, the intensity of the UV and fluorescence detection for the DF permeate was slightly reduced by the UF membrane and substantially removed by the RO membrane. Variations in the fluorescence characteristics of EfOM through the pilot plant are shown in Fig. 3. The maximum peak of bovine serum albumin (BSA) (the reference material of protein-like substances) was found at Ex ¼ 280 and Em ¼ 340 nm, and humic-like fluorescence had two pairs of maximum peaks at Ex ¼ 270 nm/Em ¼ 450 nm (SRHA I), Ex ¼ 315 nm/Em ¼ 440 nm (SRHA II) and Ex ¼ 260 nm/ Em ¼ 450 nm (SRFA I), Ex ¼ 320 nm/Em ¼ 440 nm (SRFA II) (Chon et al., 2011). Strong humic-like fluorescence was commonly observed in water samples at the same Ex and Em wavelength (Ex ¼ 330 nm and Em ¼ 410 nm) but protein-like fluorescence was not observed due to the low protein concentration. The fluorescence intensity of the feed water was unaffected by the CCeDF process but that of the DF permeate (intensity: 2665) was slightly reduced by the UF membrane (intensity: 2186) and completely removed through the RO membrane. The polysaccharides could be not well characterized by HPSEC and 3D FEEM analyses due to their physicochemical properties. However, they were considered to have a great potential for the membrane fouling because polysaccharides produced during the biological process for wastewater treatment are major constituents of EfOM (Lee et al., 2006).

3.3. Foulant characteristics of the UF and RO membranes

characteristics. Ammonium was the dominant nitrogen species for UFeDI (>84%) while ROieDI and ROfeDI were mainly comprised of both ammonium (40e58%) and organic nitrogen species (37e52%). In addition, organic nitrogen species were the major fractions in the membrane foulants desorbed using alkaline and acidic solutions in the range of 36%e98%. The membrane fouling could be categorized into two types based on the adhesion strength of the foulants: membrane fouling removable by physical membrane cleaning and membrane fouling removable by chemical membrane cleaning (Chang et al., 2002; Mahendran et al., 2002). Therefore, in the tested pilot plant, the physically removable organic fouling was defined as the DOC of the membrane foulants desorbed using DI water (UF: 53%, ROi: 47% and ROf: 33%) and the sum of the DOC of the membrane foulants desorbed using alkaline solution (UF: 42%, ROi: 50% and ROf: 60%) and acid solution (UF: 5%, ROi: 3% and ROf: 7%) was considered as the chemically removable organic fouling, which meaning that the flux recovery rate of the fouled membranes might be different according to the type of membranes and desorbing agents.

3.3.2.

Inorganic constituents of the membrane foulants

The inorganic constituents of the desorbed membrane foulants are presented in Table 2. Most of the inorganic materials

a

140 F

Virgin UF Fouled UF

C

120 100 Intensity

Characteristics of EfOM

80 60 Al

O

40 20 B

0.0

0.5

1.0

1.5

2.0

2.5

3.0

KeV

b

75 Virgin RO Fouled ROi Fouled ROf

C

50 O

25

3.3.1.

Si

Zn

Ca

0

Intensity

3.2.

Al

Organic constituents of the membrane foulants

Characteristics of the organic foulants desorbed from the UF and RO membranes are summarized in Table 2. The membrane foulants desorbed using DI water had relatively low SUVA and aromaticity compared to the membrane foulants desorbed using alkaline and acid solutions. Another major difference in the organic constituents of the desorbed membrane foulants was related to the nitrogen

B

0 0.0

Fe

Si

Ca

0.5

1.0

1.5

2.0

2.5

3.0

KeV

Fig. 4 e EDS spectra of the virgin and fouled membranes: (a) UF membrane and (b) RO membranes.

1810

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 8 0 3 e1 8 1 6

were more effectively extracted by DI water and alkaline solution than by acid solution but the reverse was true for the inorganic scalants, including calcium and magnesium. In the tested pilot plant, high levels of aluminum, boron, calcium, iron, magnesium, zinc, and silica were included in the desorbed membrane foulants. As shown in Fig. 4, the EDS spectra of the virgin UF membrane had only strong carbon and fluorine peaks which were associated with membrane composite materials, whereas no distinctive peak was found in the EDS spectra of the virgin RO membrane. The strong peaks of carbon for the fouled UF and RO membranes were likely to be a hydrophilic fraction of EfOM (i.e., polysaccharide-like substances) produced during biological wastewater treatment process. The high levels of aluminum, silica and oxygen supports such as aluminum silicates and aluminum oxide

colloids could have been major inorganic constituents in the fouled membranes. These findings were consistent with the results of the ICPeMS analysis of the desorbed membrane foulants. The ICPeMS and EDS analyses revealed that the fouling formation of the UF and RO membranes could be enhanced by inorganic colloids which were bounded by organic fouling layer and/or multivalent metal ions complexed with accumulated organic matter on the membrane surface (Lee and Elemelech, 2006; Zhao et al., 2010).

3.3.3.

HPSEC and 3D FEEM analyses

The MW distribution of the desorbed UF and RO membrane foulants, including aromatic and protein-like substances, is illustrated in Fig. 5. The highest intensity of the UV detection with low MW (880e2260 Da) corresponded to a peak at

Fig. 5 e MW distribution of the desorbed membrane foulants: (a)e(c) aromatic substances of the desorbed UF, ROi and ROf membrane foulants and (d)e(f) protein-like substances of the desorbed UF, ROi and ROf membrane foulants.

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 8 0 3 e1 8 1 6

1230 Da, which is associated with humic substances and organic acids. High MW fractions of aromatic substances were found in the MW range of 24,570e35,590 Da, which are related to non-humic substances (i.e., polysaccharides, proteins and amino sugar) from the secondary effluents (Jarusutthirak et al., 2002). Both low and high MW fractions of aromatic substances were similar for all desorbed membrane foulants. The protein-like substances were mainly composed of both low (410e1800 Da) and high MW (23,620e27,080 Da). Furthermore, relatively weak intensities of the fluorescence detection were also observed in the medium MW range (5120e8010 Da). The membrane foulants desorbed using DI

1811

water and alkaline solution had similar DOC fractions but the intensities of the UV detection for the membrane foulants desorbed using DI water were considerably lower than those of the desorbed membrane foulants using alkaline solutions. Furthermore, in contrast to the UV detection, the membrane foulants desorbed using DI water had relatively high fluorescence intensities in the high MW range (23,840e27,080 Da) compared to the membrane foulants desorbed using alkaline solution. These results implied that the membrane foulants desorbed using DI water could include more non-humic substances compared to the membrane foulants desorbed using alkaline solution.

Fig. 6 e Contour plots of the desorbed UF and RO membrane foulants: (a) UFeDI, (b) UF-base, (c) UF-acid, (d) ROieDI, (e) ROibase, (f) ROi-acid, (g) ROfeDI, (h) ROf-base, and (i) ROf-acid.

1812

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 8 0 3 e1 8 1 6

Fig. 7 e FEEM of the reference materials and the desorbed UF and RO membrane foulants.

Fig. 6 exhibits the fluorescence contour plots of the desorbed membrane foulants. Two different types of fluorescence chromophore were commonly found for the UFeDI, ROieDI and ROfeDI at Ex ¼ 280 nm/Em ¼ 330 nm and Ex ¼ 350 nm/ Em ¼ 430 nm, and the strong fluorescence of UF-base, ROi-base and ROf-base was in the Ex range of 330e340 nm and Em range of 410e420 nm. However, weak fluorescence peaks were detected for UF-acid, ROi-acid and ROf-acid in the Ex range of 300e330 nm and Em range of 410e420 nm. The location of the peaks in Group A (UFeDI I, ROieDI I and ROfeDI I) was similar to BSA (proteinlike fluorescence) while the peaks in Group B (UFeDI II, ROieDI II, ROfeDI II, UF-base, ROi-base, ROf-base, UF-acid, ROi-acid, and ROf-acid) corresponded to humic-like fluorescence at higher Ex and Em wavelengths (i.e., SRHA II and SRFA II), as shown in Fig. 7. The observations from the MW and fluorescence characteristics of the desorbed membrane foulants revealed that hydrophilic fractions, including polysaccharide-like and protein-like substances, were major constituents of the UF and RO membrane foulants in the tested pilot plant.

3.3.4. Fractionation using Amberlite XAD-8/4 resins and FTIR analysis As represented in Fig. 8, the desorbed membrane foulants were fractionated using Amberlite XAD-8/4 resins into three fractions: hydrophobic (HPO) (XAD-8 resin adsorbable), hydrophilic (HPI) (not adsorbable in either XAD-8 resin or XAD-4 resin) and transphilic (TPI) (XAD-4 resin adsorbable) fractions. The relative polarities of the desorbed membrane

foulants might differ according to the type of membrane and desorbing agent solution. The desorbed membrane foulants from the ROi (ROieDI: 15% of HPO, 82% of HPI and 2% of TPI; ROi-base: 13% of HPO, 77% of HPI and 10% of TPI; ROi-acid: 10% of HPO, 89% of HPI and 1% of TPI) and ROf (ROfeDI: 7% of HPO, 87% of HPI and 6% of TPI; ROf-base: 15% of HPO, 79% of HPI and 6% of TPI; ROf-acid: 3% of HPO, 93% of HPI and 4% of TPI) membranes contained slightly more hydrophilic fractions than the desorbed foulants from the UF membrane (UFeDI: 18% of HPO, 74% of HPI and 8% of TPI; UF-base: 33% of HPO, 49% of HPI and 18% of TPI; UF-acid: 8% of HPO, 86% of HPI and 6% of TPI). In the case of ROi-base and ROf-base, their hydrophilic fractions were approximate twice as high as those of UF-base, which was in good agreement with the results of HPSEC and 3D FEEM analyses. The infrared (IR) spectra of the desorbed membrane foulants are shown in Fig. 9. The specific constituents of organic matter (e.g., EfOM and membrane foulants) could not be confirmed by FTIR analysis because of their complexity but it is a very useful method to identify the functional group compositions. The peaks in the range of 3500e3300 cm1, 1490e1440 cm1 and 840e750 cm1 were attributed to the NeH stretching of amides, the peak in the range of 1680e1630 cm1 was indicative of the carbonyl C]O band of primary amides, and the CNH band of secondary amides appeared in the range of 1570e1515 cm1. The amide groups were associated with polypeptides released by cell lysis of microorganisms. Carboxylic acids of the humic substances were detected in the range of 3100e2900 cm1 (OeH stretching) and 1740e1720 cm1 (carbonyl C]O band), and the CH band of aldehydes appeared in the range of 2900e2800 cm1. The CeO stretching and the OH band derived from the polysaccharide-like substances occurred in the range of 1410e1310 cm1, 1350e1260 cm1 and 1125e1090 cm1, 1075e1000 cm1. The functional group compositions were also quite different depending on the membrane type and desorbing agent solution. A distinctive IR peak of carboxylic acids was observed for the UF-base that was most hydrophobic in the range of 1740e1720 cm1. Relatively strong IR intensities of polysaccharide-like and protein-like functional groups were found in the membrane foulants desorbed using DI water compared to the membrane foulants desorbed using alkaline solution, indicating that the functional group compositions of the desorbed membrane foulants were closely interrelated to their hydrophobicity.

Fig. 8 e Fractionation of the desorbed membrane foulants using Amberlite XAD-8/4 resins: (a) UF, (b) ROi and (c) ROf membrane.

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 8 0 3 e1 8 1 6

1813

Fig. 9 e IR spectra of the desorbed membrane foulants: (a) UF, (b) ROi and (c) ROf membrane.

Fig. 10 e FE-SEM images of the (a)e(b) virgin UF membrane, (c)e(i) inorganic precipitates, algae and bacteria found on the fouled UF membrane, (j)e(k) virgin RO membrane and (l)e(o) inorganic precipitates found on the fouled RO membranes.

1814

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 8 0 3 e1 8 1 6

Table 3 e Physical and chemical properties of the virgin and fouled RO membranes. Fouled ROi membrane

Fouled ROf membrane

0.3 30.4 35.5

2.0 67.7 17.2

1.5 91.3 6.6

Surface topology (mm) Contact angle ( ) Zeta potential at pH 7 (mV) Roughness (nm)

63.5a

0.6a

0.4a

a Roughness presents the root mean-square-average (Rq) of the roughness profile ordinates.

Changes in the surface morphology of the membranes

The FE-SEM images of the virgin and fouled UF and RO membranes are illustrated in Fig. 10. Even though the feed water was pretreated with the CCeDF process, the DF permeate tended to be free from various algae, bacteria and inorganic precipitates associated with calcium carbonate due to the large mesh size of DF (15 mm and 10 mm). The accumulation of bacteria and algae onto the UF membrane surface (i.e., biofouling) was enhanced in the presence of extracellular polymeric substances (EPS) produced via biological processes because EPS form a gel-like matrix that can provide mechanical stability to microbial biofilm through hydrogen bonds, electrostatic interaction and van der Waals interactions (Herzberg et al., 2010). In contrast to the UF membrane, EfOM and inorganic precipitates mostly accumulated in the fouled RO membranes. A substantial decrease in the pore size caused by the deposition of organic and inorganic materials and inorganic colloids stabilized by humic substances was observed in the fouled UF and RO membranes. In addition, the surface topology of the fouled ROi membrane was considerably raised from 300 nm to 2 mm and that of the fouled ROf membrane was also raised by 1.5 mm, as listed in Table 3, which was presumably due to the deposition of hydrophilic EfOM (i.e., non-humic substances), and organic and inorganic colloids.

3.4.

Fouling behavior of the UF and RO membranes

As illustrated in Fig. 11, a serious flux decline was found for both the UF and RO membranes, which was likely due to either polysaccharide-like or protein-like substances of the

a

2.0

b

2.2 Flux TMP

2.0

14

0.58 0.55

1.8

Flux TMP

1.6

1.6

1.4 1.4

1.2

12 TMP (kgf/cm2)

Flux (m3/m2 day)

1.8 Flux (m3/m2 day)

3.3.5.

0.52 0.49

10

0.46

1.0 1.2 0.8 1.0

0.6 0

10

20

30

Time (day)

40

50

60

8 0.43 0.40

6 0

10

20

30

Time (day)

Fig. 11 e Changes in TMP and flux of the UF and RO membranes.

40

50

60

TMP (kgf/cm2)

Virgin RO membrane

EfOM from secondary effluents. However, significant differences in fouling behaviors of the UF and RO membranes were caused by the use of coagulants (i.e., aluminum sulfate). The water and foulant characteristics clearly revealed the aluminum sulfates in the DF permeate to be influential in the fouling formation of the UF and RO membranes. In addition, the inorganic constituents of the desorbed UF membrane foulants and the EDS spectra of the fouled UF membrane fiber demonstrated that the severe fouling of the UF membrane was caused by aluminum species, including aluminum sulfate, aluminum silicates, aluminum oxides, and aluminum ions-humic substances complexes. The hydrophobic fractions of EfOM (i.e., humic substances) were treated with coagulation, after which the aggregates of humic substances that had been induced by aluminum sulfate were found to be easily adsorbed to the UF membrane surface. Nevertheless, higher levels of hydrophilic fractions and aluminum were observed for UFeDI than for UF-base. The differences in the fouling behavior between UFeDI and UF-base may have been because most of the aluminum species for UF-base was in the form of aluminum ions-humic substances complexes. Although the desorbed ROi and ROf membrane foulants were mainly comprised of hydrophilic fractions, including polysaccharide-like and/or protein-like functional groups, the contact angle of the ROi membrane was significantly increased from 30.4 to 67.7 and the contact angle of the ROf membrane was also increased by 91.3 due to adsorption of organic foulants on the ROi and ROf membrane surfaces. In opposition to the contact angle, the surface charge at pH 7 and surface roughness of the RO membranes were continuously reduced through the serially connected 4 RO membrane modules from 35.5 mV to 6.6 mV and from 63.5 nm to 0.4 nm, respectively, as shown in Table 3. The observed results from changes in the contact angle, surface charge and roughness are evidence that the formation of an initial fouling layer of the RO membranes used for wastewater reclamation could be attributed to adsorption of the uncharged or oppositely charged polysaccharides-like and protein-like substances, which had a slightly negative and a slightly positive charge at neutral pH, respectively, onto the RO membrane surfaces which might have reduced the electrostatic interactions between ion species (i.e., charged organic and/or inorganic materials) and membrane surfaces (Her et al., 2007). As depicted in Fig. 12, after the formation of an

1815

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 8 0 3 e1 8 1 6

: Non charged hydrophilic fractions

: Negatively charged hydrophobic fractions

(–) (–) (–) (–) (–) (–) (–) (–)

RO membranes

(–)

(–)

(–)

(–)

(–)

RO membranes

(–)

(–)

(–)

RO membranes

Fig. 12 e The fouling layer formation of the RO membrane surfaces used for wastewater reclamation.

initial fouling layer by polysaccharides-like and protein-like substances, the RO membrane surfaces might become more hydrophobic since hydrophobic fractions of EfOM were more readily adsorbed on the RO membrane surfaces due to the decreasing the surface charge, which could intensify the hydrophobic interaction between the RO membrane surfaces and hydrophobic fractions of EfOM.

4.

Conclusions

Comprehensive investigation on membrane autopsies of the fouled UF and RO membranes provided valuable insights into the MWR pilot plant. The characteristics of the feed water, EfOM and membranes were investigated in relation to the removal of water contaminants and the effects of water characteristics on foulant characteristics, changes in membrane surface chemistry and fouling behavior. In the tested pilot plant, various bacteria and algae may have prevailed on the UF membrane surfaces in spite of the pretreatment by the CCeDF process but they were not found on the surfaces of the RO membrane, indicating that the CCeDF process was ineffective in mitigating biofouling for the UF membrane due to its larger mesh size while the UF membrane was effective in mitigating biofouling for the RO membranes. The major foulants of the UF and RO membranes were hydrophilic fractions of EfOM comprised of polysaccharidelike and protein-like functional groups produced during biological processes for wastewater treatment and inorganic scaling caused by aluminum, calcium, and silica. The use of aluminum sulfate as a pretreatment caused great differences in fouling behavior between the UF and RO membranes. The higher hydrophobic fractions of the desorbed UF membrane foulants (mainly UF-base) compared to the desorbed ROi and ROf membrane foulants were probably due to the adsorption of aluminum ions-humic substances complexes on the UF membrane surfaces. The rejection rate of hydrophilic fractions (i.e., non-humic substances) by a negatively charged membrane surface was substantially lower than that of hydrophobic fractions because of their charge characteristics. Hence, non-humic substances were easily deposited on the negatively charged RO membrane surface, which significantly decreased the surface charges of the RO membrane. The fouling induced by the hydrophobic fractions of EfOM was not significant in the tested UF and RO membranes due to the relatively high charge repulsion between the humic substances and the RO membrane surfaces, but the hydrophobic fractions of EfOM might have covered the fouling layers with polysaccharide-like and protein-like substances as

the contact angle was increased. This last effect was considered the most important factor in enhancing the subsequent fouling formation of the RO membranes.

Acknowledgments This research was supported by a grant from the National Research Laboratory Program by the Korea Science and Engineering Foundation (NOM laboratory: R0A-2007-000-20055-0), and also supported in part by Korea Ministry of Environment as “The Eco-Innovation Project (Global Top Project)”.

references

Al-Amoudi, A.S., Faroque, A.M., 2005. Performance restoration and autopsy of NF membrane used in seawater pretreatment. Desalination 178, 261e271. Asano, T., Burton, F.L., Leverenz, R., Tchobanoglous, G., 2007. Water Reuse: Issues, Technologies, and Applications. McGraw-Hill, New York, USA. Bellona, C., Drewes, J.E., 2007. Viability of a low-pressure nanofilter in treating recycled water for water reuse application: a pilot-scale study. Water Research 41, 3948e3958. Boerlage, S.F.E., Kennedy, M.D., Witkamp, G.J., van der Hoek, J.P., Schippers, J.C., 1999. BaSO4 solubility prediction in reverse osmosis systems. Journal of Membrane Science 159, 47e59. Chang, I.S., Le-Clech, P., Jefferson, B., Judd, S., 2002. Membrane fouling in membrane bioreactors for wastewater treatment. Journal of Environmental Engineering 128, 1018e1029. Chon, K., Cho, J., 2008. Characterizing natural organic matters from various aquatic ecosystem, using informative parameters. In: Proceedings of International Conference on Natural Organic Matter: From Source to Tap Conference. International Water Association, London, UK. Chon, K., Sarp, S., Lee, S., Lee, J.H., Lopez-Ramirez, J.A., Cho, J., 2011. Evaluation of a membrane bioreactor and nanofiltration for municipal wastewater reclamation: trace contaminants control and fouling mitigation. Desalination 272, 128e134. Gabelich, C.J., Chen, W.R., Yun, T.I., Coffey, B.M., Suffet, I.H., 2005. The role of dissolved aluminum in silica chemistry for membrane processes. Desalination 180, 307e319. Gur-Reznik, S., Katz, I., Dosoretz, C.G., 2008. Removal of dissolved organic matter by granular-activated carbon adsorption as a pretreatment to reverse osmosis of membrane bioreactor effluents. Water Research 42, 1595e1605. Gwon, E.M., Yu, M.J., Oh, H.K., Lee, Y.H., 2003. Fouling characteristics of NF and RO operated for removal of dissolved matter from ground water. Water Research 37, 2989e2997. Her, N., Amy, G., Foss, D., Cho, J., Yoon, Y., Kosenka, P., 2002. Optimization of method for detecting and characterizing NOM

1816

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 8 0 3 e1 8 1 6

by HPLC-size exclusion chromatography with UV and on-line DOC detection. Environmental Science and Technology 36, 1069e1076. Her, N., Amy, G., Plottu-Pecheux, A., Yoon, Y., 2007. Identification of nanofiltration membrane foulants. Water Research 41, 3936e3947. Herzberg, M., Berry, D., Raskin, L., 2010. Impact of microfiltration treatment of secondary wastewater effluent on biofouling of reverse osmosis membranes. Water Research 44, 167e176. Ivnitskya, H., Katza, I., Minzc, D., Shimonid, E., Chene, Y., Tarchitzkye, J., Semiatb, R., Dosoretza, C.G., 2005. Characterization of membrane biofouling in nanofiltration processes of wastewater treatment. Desalination 185, 255e268. Jarusutthirak, C., Amy, G., 2006. Role of soluble microbial products (SMP) in membrane fouling and flux decline. Environmental Science and Technology 40, 960e974. Jarusutthirak, C., Amy, G., Croue´, J.P., 2002. Fouling characteristics of wastewater effluent organic matter (EfOM) isolates on NF and UF membranes. Desalination 145, 247e255. Kim, H.A., Choi, J.H., Takizawa, S., 2007. Comparison of initial filtration resistance by pretreatment processes in the nanofiltration for drinking water treatment. Separation and Purification Technology 56, 354e362. Kamp, P.C., Kruithof, J.C., Folmer, H.C., 2000. UF/RO treatment plant Heemskerk: from challenge to full scale application. Desalination 131, 27e35. Kim, J., DiGiano, F.A., Reardon, R.D., 2008. Autopsy of highpressure membranes to compare effectiveness of MF and UF pretreatment in water reclamation. Water Research 42, 697e706.

Lee, S., Ang, W.S., Elimelech, M., 2006. Fouling of reverse osmosis membranes by hydrophilic organic matter: implications for water reuse. Desalination 187, 313e321. Lee, S., Elemelech, M., 2006. Relating organic fouling of Reverse osmosis membranes to intermolecular adhesion forces. Environmental Science and Technology 40, 980e987. Lee, S., Lee, E., Ra, J., Lee, B., Kim, S., Choi, S.H., Kim, S.D., Cho, J., 2008. Characterization of marine organic matters and heavy metals with respect to desalination with RO and NF membranes. Desalination 221, 244e252. Li, G., Xu, Z., Pinnau, I., 2007. Fouling of reverse osmosis membranes by biopolymers in wastewater secondary effluent: role of membrane surface properties and initial permeate flux. Journal of Membrane Science 290, 173e181. Mahendran, B., Lin, H., Liao, B., Liss, S.N., 2002. Surface properties of biofouled membranes from a submerged anaerobic membrane bioreactor after cleaning. Journal of Environmental Engineering 137, 504e513. Schneider, R.P., Ferreira, L.M., Binder, P., Ramos, J.R., 2005. Analysis of foulant layer in all elements of an RO train. Journal of Membrane Science 261, 152e162. Tran, T., Bolto, B., Gray, S., Hoang, M., Ostarcevic, E., 2007. An autopsy study of a fouled reverse osmosis membrane element in a brackish water treatment plant. Water Research 41, 3915e3923. Xu, P., Bellona, C., Drewes, J.E., 2010. Fouling of nanofiltration and reverse osmosis membranes during municipal wastewater reclamation: membrane autopsy results from pilot-scale investigations. Journal of Membrane Science 353, 111e121. Zhao, Y., Song, L., Ong, S.L., 2010. Fouling behavior and foulants characteristics of reverse osmosis membranes for treated secondary effluent reclamation. Journal of Membrane Science 349, 65e74.