Performance of mesoporous adsorbent resin and powdered activated carbon in mitigating ultrafiltration membrane fouling caused by algal extracellular organic matter

Performance of mesoporous adsorbent resin and powdered activated carbon in mitigating ultrafiltration membrane fouling caused by algal extracellular organic matter

Desalination 336 (2014) 129–137 Contents lists available at ScienceDirect Desalination journal homepage: www.elsevier.com/locate/desal Performance ...

1MB Sizes 0 Downloads 26 Views

Desalination 336 (2014) 129–137

Contents lists available at ScienceDirect

Desalination journal homepage: www.elsevier.com/locate/desal

Performance of mesoporous adsorbent resin and powdered activated carbon in mitigating ultrafiltration membrane fouling caused by algal extracellular organic matter Kai Li, Fangshu Qu, Heng Liang ⁎, Senlin Shao, Zheng-shuang Han, Haiqing Chang, Xing Du, Guibai Li State Key Laboratory of Urban Water Resource and Environment (SKLUWRE), Harbin Institute of Technology, 73 Huanghe Road, Nangang District, Harbin 150090, PR China

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• Effects of two adsorbents on UF membrane fouling by algal EOM were evaluated. • MAR mainly adsorbed macromolecules in EOM and effectively reduced cake formation. • Both MAR and PAC pretreatments decreased irreversible adhesion of hydrophobic EOM. • MAR was much more efficient in mitigating EOM fouling than PAC.

a r t i c l e

i n f o

Article history: Received 14 September 2013 Received in revised form 5 December 2013 Accepted 1 January 2014 Available online 25 January 2014 Keywords: Ultrafiltration Membrane fouling Extracellular organic matter Adsorption Mesoporous adsorbent resin Powdered activated carbon

a b s t r a c t This paper focused on the control of ultrafiltration (UF) membrane fouling caused by extracellular organic matter (EOM) extracted from Microcystis aeruginosa through pretreatment with mesoporous adsorbent resin (MAR) and powdered activated carbon (PAC). The influence of MAR and PAC pretreatments on characteristics of EOM was investigated using molecular weight (MW) fractionation, DAX-8/XAD-4 resin fractionation and fluorescence excitation-emission matrix (EEM) spectroscopy. The results suggested that MAR mainly removed high-MW (N100 kDa) fraction of EOM, while PAC primarily adsorbed low-MW (b1 kDa) fraction. Both MAR and PAC adsorption removed more hydrophobic fraction than hydrophilic fraction. UF experiments were carried out to evaluate the efficacies of MAR and PAC pretreatments in EOM fouling control. MAR pretreatment significantly reduced the reversible fouling due to the efficient removal of high-MW fraction and the consequent reduction of cake formation; whereas PAC exhibited little ability in alleviating the reversible fouling. Nevertheless, the irreversible fouling, which accounted for a small part of the total fouling, was mitigated by both MAR and PAC pretreatments because they both reduced irreversible adhesion caused by hydrophobic EOM. Overall, with respect to EOM fouling control, MAR was much more efficient than PAC. © 2014 Elsevier B.V. All rights reserved.

Abbreviations: DOC, dissolved organic carbon; EEM, excitation-emission matrix; EOM, extracellular organic matter; MAR, mesoporous adsorbent resin; MW, molecular weight; NOM, natural organic matter; PAC, powdered activated carbon; PES, polyethersulfone; TMP, trans-membrane pressure; UF, ultrafiltration. ⁎ Corresponding author. Tel./fax: +86 451 86283001. E-mail addresses: [email protected] (K. Li), [email protected] (F. Qu), [email protected] (H. Liang), [email protected] (S. Shao), [email protected] (Z. Han), [email protected] (H. Chang), [email protected] (X. Du), [email protected] (G. Li). 0011-9164/$ – see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.desal.2014.01.001

130

K. Li et al. / Desalination 336 (2014) 129–137

1. Introduction The frequent occurrence of algal blooms in water sources has posed a serious challenge to the conventional drinking water treatment process [1–3]. Algal blooms can not only adversely affect the treatment process but also deteriorate the finished water quality [4]. UF is a promising technology for algal-rich water treatment because it is able to remove algal cells absolutely without cell lysis and is less affected by raw water quality fluctuation [5,6]. However, algal-rich water can cause severe membrane fouling during UF, which leads to the increase of operation cost and the decrease of water production [5,7]. Membrane fouling caused by algal-rich water has been broadly investigated and extracellular organic matter (EOM) released from algal cells has been identified as the major membrane foulants [2,8–11]. Therefore, it is crucial to reduce the EOM fouling for the application of UF in algal-rich water treatment. Many efforts have been made to get more insights into EOM characteristics and its fouling behavior. Characteristics of EOM originating from four different algal species were extensively investigated using various techniques [4]. All EOM was found to be dominated by hydrophilic polysaccharides and hydrophobic proteins and exhibited low specific ultraviolet absorbance value, which suggested that EOM differed greatly from allochthonous natural organic matter (NOM). Qu et al. [12] investigated UF membrane fouling caused by EOM extracted from M. aeruginosa and influence of interfacial characteristics of EOM on membrane fouling. It was found that the high molecular weight (highMW, N100 kDa) fraction of EOM contributed a significant portion of the fouling and that hydrophobic components tended to adhere to membrane surface and cause irreversible fouling. Huang et al. [13] found that EOM generated under different nutrient conditions exhibited different characteristics and membrane fouling potentials and that the membrane fouling potential of EOM was mainly associated with the high-MW polysaccharide-like and protein-like substances. Zhang et al. [14] reported that EOM extracted from M. aeruginosa with longer growth time caused more severe membrane fouling and that highMW biopolymers were the major foulants in EOM. These results indicate that the high-MW fraction of EOM is the major membrane foulants and the removal of this fraction may effectively reduce the EOM fouling. Adsorption can reduce the amount of organic matter in feed water, therefore it was supposed to be able to mitigate UF membrane fouling caused by organic matter. Many studies have been conducted to investigate the efficacy of different types of adsorbents (e.g., powdered activated carbon (PAC), metal oxide particles, anion exchange resin and adsorbent resin) in controlling membrane fouling caused by allochthonous NOM [15,16]. It was demonstrated in many studies that adsorbents reduced the amount of organic matter deposited on membrane and mitigated membrane fouling efficiently [6,17–21]. However, in some studies, adsorbents that removed a large amount of organic matter exerted minor effects on membrane fouling because the major foulant was not removed by the adsorbents [22,23]. Furthermore, adsorbents were reported to even adversely affect membrane fouling in integrated adsorbent/UF systems, since the adsorbent particles were bound to each other and to the membrane surface by organic matter [24,25]. These results suggested that reducing the amount of organic matter did not necessarily alleviate UF membrane fouling and the characteristics of adsorbents and foulants were the more decisive factors in fouling control [19,26]. The characteristics of EOM were quite different from those of allochthonous NOM. However, unlike the extensive attention paid to NOM fouling control, few studies have specifically focused on the performance of adsorbents in EOM fouling control. Campinas and Rosa [20] reported that PAC addition affected neither flux decline nor the reversibility of EOM fouling. However, only one PAC dose (10 mg/L) was investigated in that study, and the knowledge about the adsorption characteristics of EOM on PAC as well as the effects of PAC dose and

contact time on membrane fouling are still lacking. Mesoporous adsorbent resin (MAR) which was made of a hydrophobic membrane material was reported to be able to remove NOM with MW of 20–200 kDa in a lake water [27,28]. Based on the understanding of EOM characteristics and its fouling behavior, we inferred that MAR had the potential for removing the high-MW fraction of EOM and thus effectively controlling the membrane fouling caused by EOM. The main objective of this study was to investigate the performance of MAR and PAC in reducing UF membrane fouling caused by EOM extracted from M. aeruginosa. The removal of EOM by these two adsorbents and the variations in characteristics of EOM after adsorption were systematically examined and compared. UF experiments were performed with raw and treated EOM to investigate the efficacy of MAR and PAC pretreatments in mitigating EOM fouling. In addition, the contributions of MAR and PAC particles themselves to EOM fouling were also examined. 2. Materials and methods 2.1. MAR and PAC MAR was made of polyethersulfone (PES) following the method proposed by Clark et al. [27]. The synthesis and rinse of MAR were conducted in four sequential steps: 1) a polymer solution was prepared by dissolving 4 g of PES (Veradel 3000P, Solvay, USA) in 172 mL of N-methylpyrrolidone (Bench Chemicals, China) and 20 mL of propionic acid (Bench Chemicals, China), 2) the solution was injected into a stirred reactor containing Milli-Q water by a peristaltic pump at a rate of 1.5 mL/min and a suspension of MAR particles was formed during the injection, 3) the suspension was transferred into a stirred filtration cell equipped with a 0.45 μm mixed cellulose filter (Taoyuan, China) and MAR particles were retained by the filter while organic solvents could pass through the filter, 4) sufficient Milli-Q water was used to rinse MAR particles until no release of dissolved organic carbon (DOC) was detected and the MAR particles were concentrated to a mass concentration of 10 g/L and stored in refrigerator for use. Woodbased PAC, which was purchased from Bench Chemicals (Tianjin, China), was used without further purification. 2.2. Algal culture and EOM extraction M. aeruginosa was cultured in batch mode under the conditions as described by Qu et al. [12]. EOM was extracted by centrifuging the algal suspension at 10,000 rpm (11,179 g) and 4 °C for 15 min and subsequently filtering the supernatant by 0.45 μm mixed cellulose filters (Taoyuan, China). The DOC content of EOM was diluted to 3.5 ± 0.4 mg/L with Mill-Q water and the pH was adjusted to 7.0 ± 0.1 using1 M HCl and 1 M NaOH before use. 2.3. Adsorption tests Adsorption of EOM with MAR and PAC was conducted as follows: 1) EOM solutions were mixed with predetermined amounts of MAR and PAC in flasks; 2) the flasks were shaken in a rotary shaker at 120 rpm and 25 °C for predetermined time; 3) mixed solutions of adsorbents and EOM were filtered with 0.45 μm mixed cellulose filters (Taoyuan, China) to remove adsorbent particles. The efficacy of MAR and PAC for EOM removal was evaluated by measuring DOC concentrations of EOM before and after adsorption. Effects of adsorbent dose and contact time were investigated by varying the amount of adsorbent (10–200 mg/L) and the contact time (10 min–12 h). 2.4. UF tests The UF membranes used in this study were flat sheet PES membranes (OM100076, Pall, USA) with an effective surface area of 42 cm2

K. Li et al. / Desalination 336 (2014) 129–137

and a nominal molecular weight cutoff of 100 kDa. New membranes were rinsed carefully to remove preservatives before use. These membranes were soaked in Milli-Q water for 24 h and the water was replaced for 3 times. Then they were filtered with Milli-Q water until the filtrate exhibited a comparable DOC concentration with Milli-Q water. Membrane filtration experiments were conducted in a 400 mL stirred cell (Amicon 8400, Millipore, USA) and operated in constant flux mode at room temperature (25 ± 1 °C). The membrane was placed at the bottom of the cell with glossy side towards the feed solution. A peristaltic pump was used as the suction pump and the permeate flux was kept at 150 L/(m2 h). A pressure transducer (PTP708, Tuopo Electric, Foshan, China), which was connected to a computer, was mounted between the filtration cell and the suction pump to monitor the trans-membrane pressure (TMP) and the data was automatically logged every five seconds. The increase of TMP indicated the accumulation of membrane fouling. Each filtration test consisted of 4 steps: 1) filtration of 100 mL Mill-Q water with the average of TMP values recorded as TMP0, 2) filtration of 300 mL feed solution with the TMP at the end named as TMP1, 3) backwashing with 50 mL Mill-Q water by placing the reverse side of the membrane upwards, and 4) filtration of 100 mL Mill-Q water with the average of TMP values recorded as TMP2. Total fouling, reversible fouling and irreversible fouling could be calculated as follows.

131

Total fouling ¼ TMP1 ‐TMP0

ð1Þ

Corp., USA) were used. The separations were conducted in the stirred cell with nitrogen gas at a constant pressure of 0.1 MPa as the drive force and with the stirrer running at the rate of 200 rpm. For each fraction, 200 mL sample was fed and 100 mL permeate was produced for analysis with the initial 10 mL discarded. The size distributions were calculated as differences in DOC concentrations between permeates of membranes with different nominal molecular weight cutoffs. Each fractionation was done in duplicate. Nonionic macroporous resins were used to separate raw, MARtreated and PAC-treated EOM into three fractions based on their hydrophobicity [29,30]. The samples were acidified to pH = 2.0 and then passed through the DAX-8 (Supelite, Sigma, USA) and XAD-4 (Amberlite, Rohm and Haas, USA) columns successively. Hydrophobic fraction was adsorbed onto DAX-8 resin and transphilic fraction was adsorbed onto XAD-4 resin, with hydrophilic fraction passing through both DAX-8 and XAD-4 columns. Before fractionation, the resins were washed by methanol, 0.1 N NaOH, 0.1 N HCl and Milli-Q water according to the procedure described by Her et al. [30]. Each fractionation was conducted in duplicate and quantified by DOC measurements. Fluorescence spectroscopy is a widely applied tool for characterizing NOM and it can offer useful information for identification of membrane foulants [31,32]. A fluorescence spectrophotometer (F-7000, HITACHI, Japan) was used to obtain fluorescence EEM spectra of raw, MARtreated and PAC-treated EOM. The emission spectra were scanned from 250 to 550 nm at 1 nm increment and the excitation spectra were scanned from 200 to 450 nm at 5 nm increments [12,32].

Reversible fouling ¼ TMP1 ‐TMP2

ð2Þ

3. Results and discussion

Irreversible fouling ¼ TMP2 ‐TMP0 :

ð3Þ

3.1. Textural and surface properties of MAR and PAC

UF tests were performed with EOM before and after MAR and PAC pretreatments to investigate their effects on EOM fouling. MAR and PAC pretreatments were carried out following the procedure of adsorption tests and adsorbent particles were removed before UF. Moreover, in order to evaluate the contributions of adsorbent particles to membrane fouling, UF tests were also carried out with MAR and PAC particles present in feed water at the dose of 50 mg/L. Each UF test was conducted in triplicate. For convenience, in this paper, MAR-treated EOM and PAC-treated EOM denote the treated EOM with adsorbent particles removed before UF, while MAR + EOM and PAC + EOM denote the treated EOM with adsorbent particles present in UF feed water. 2.5. Analytical methods DOC was measured using a total organic carbon analyzer (multi N/C 2100, Jena, Germany) and each sample was analyzed in triplicate with errors less than 3%. The particle size distributions of MAR and PAC were measured using MasterSizer 2000 (Malvern, UK). Zeta potentials of MAR and PAC were measured using Nano S90 (Malvern, UK) at the pH of 7.0 ± 0.1. The surface areas and pore size distributions of MAR and PAC were measured using a surface area and porosity analyzer (ASAP 2020, Micromeritics, USA) with N2 adsorption method. Before analysis, stock solutions of MAR and PAC were filtered using 0.45 μm filters and then MAR and PAC particles retained by the filters were dried at 65 °C overnight. All the analyses were done in triplicate. Microstructures of MAR and PAC were observed using a scanning electron microscope (Quanta 200FEG, FEI, USA). Samples of MAR and PAC were dried at 65 °C overnight and then coated with gold using a precision etching coating system (Model 682, Gatan, USA) before observation. MW distributions of raw, MAR-treated and PAC-treated EOM were determined using UF separation method in parallel mode with identical initial samples fed for each membrane [12]. Regenerated cellulose membranes with nominal molecular weight cutoffs of 100, 30, 10, 3, and 1 kDa (Amicon YM100,YM30, YM10, YM3 and YM1, Millipore

The textural and surface properties of MAR and PAC are listed in Table 1. It can be observed that the average particle size of MAR was smaller than that of PAC, but the difference (25.2 versus 32.1 μm) was not remarkable. Zeta potentials of MAR and PAC indicated that they were both negatively charged and had similar surface charge. However, the pore structure characteristics of MAR and PAC were apparently different. MAR had a BET surface area of 108 g/m2, which was significantly smaller than that of PAC (1219 g/m2). The average pore size of MAR was 16.4 nm and the large mesopore (10 b w b 50 nm) volume (0.215 cm3/g) accounted for 78% of the total pore volume. The characteristics of MAR in this study were similar to those of MAR prepared by Clark et al. [27]. With respect to PAC, the average pore size was much smaller and the pore structure was characterized by micropores (w b 2 nm), which was consistent with the results in previous literatures [17,21]. Scanning electron microscope images of MAR and PAC are shown in Fig. 1. As shown in Fig. 1(a), MAR particles are aggregates of nanoscale primary particles which are fairly spherical and uniform and the pores are the interstitial area between the primary particles. Moreover, the

Table 1 Textural and surface properties of MAR and PAC. Parameter

MAR

Average particle size (d50, μm) Zeta potential (mV) BET surface area (m2/g) Average pore size (nm) Micropore (w b 2 nm) volume (cm3/g) Small mesopore (2 b w b 10 nm) volume (cm3/g) Large mesopore (10 b w b50 nm) volume (cm3/g) Macropore (w N 50 nm) volume (cm3/g)

25.2 −22.4 108 16.4 0.031 0.021

PAC ± ± ± ± ± ±

0.8 1.0 9 0.5 0.003 0.003

32.1 ± −23.9 ± 1219 ± 2.2 ± 0.244 ± 0.075 ±

0.7 0.9 13 0.1 0.013 0.004

0.215 ± 0.015

0.037 ± 0.003

0.010 ± 0.002

0.016 ± 0.002

Note: values represent average ± standard deviation, n = 3.

K. Li et al. / Desalination 336 (2014) 129–137

mesoporous structure of MAR was consistent with the pore characteristics measured by N2 adsorption method. The pores of PAC were mainly micropores (w b 2 nm) and they could not be observed at the magnification times of 160,000 ×. Therefore, the number of visible pores in Fig. 1(b) is much less than that in Fig. 1(a).

3.2. Removal of EOM by MAR and PAC adsorption Fig. 2(a) presents the removal of EOM by MAR and PAC over the dose range of 10–200 mg/L. For both MAR and PAC, the removal rate of EOM increased with the adsorbents' dose. Besides, PAC removed more EOM than that of MAR at all doses investigated. At the dose of 10 mg/L, MAR removed only 6% of EOM while PAC achieved a slightly higher removal of 9%. 41% of EOM was adsorbed by MAR at the dose of 200 mg/L, whereas PAC had already achieved a higher removal of 46% at a smaller dose of 100 mg/L. The higher adsorption capacity of PAC might be

a 60 MAR

PAC

50

DOC Removal (%)

132

40 30 20 10 0 0

50

100

150

b 40 DOC Removal (%)

MAR

a

200

Adsorbent dose (mg/L) PAC

30

20

10

0 0

2

4

6

8

10

12

Contact time (h) Fig. 2. Removal of EOM by MAR and PAC adsorption: (a) effects of adsorbent dose (contact time: 12 h); (b) effects of contact time (doses of MAR and PAC: 50 mg/L) (error bars represent standard deviations from the means, n = 3).

b

attributed to its larger surface area because a large surface area was usually in favor of the adsorption process [25,33]. However, the difference between the adsorption capacity of MAR and PAC (e.g. 41% versus 57% at the dose of 200 mg/L) was not as significant as the difference between the surface areas (i.e. 108 versus 1219 g/m2). The adsorption capacity of an adsorbent can be influenced by many factors, including the characteristics of the adsorbent (e.g. surface area, pore size distribution,

2.0

Raw EOM MAR-treated EOM PAC-treated EOM

DOC (mg/L)

1.6 1.2 0.8 0.4 0.0 >100

30~100 10~30

3~10

1~3

<1

Apparent Molecular Weight (kDa) Fig. 1. Scanning electron microscope images of MAR (a) and PAC (b) (magnification = 160,000×).

Fig. 3. Effects of MAR and PAC pretreatments on MW distributions of EOM (doses of MAR and PAC: 50 mg/L, contact time: 30 min) (error bars represent standard deviations from the means, n = 2).

K. Li et al. / Desalination 336 (2014) 129–137

2.5

contact time further extended to 12 h, the removals of EOM by MAR and PAC increased to 27% and 35%, respectively, suggesting that further extension of contact time just resulted in slight increase of EOM removals.

Raw EOM MAR-treated EOM PAC-treated EOM

1.5 3.3. Influence of MAR and PAC pretreatments on the characteristics of EOM

1.0 0.5 0.0 Hydrophobic

Transphilic

Hydrophilic

Fig. 4. Effects of MAR and PAC pretreatments on hydrophobicity distributions of EOM (doses of MAR and PAC: 50 mg/L, contact time: 30 min) (error bars represent standard deviations from the means, n = 2).

carbon content), the properties of the adsorbate (such as molecular size, solubility), and the solution chemistry (pH, ionic strength, etc.)[34–37]. In this study, the adsorbate and the solution chemistry were identical for the two adsorbents, therefore it can be inferred that the factors such as pore size distribution and carbon content might play an important role. Fig. 2 (b) shows the effects of contact time on the removal of EOM at the dose of 50 mg/L. It can be observed that the removal increased sharply during the initial 30 min for both MAR and PAC. The removal of EOM by MAR was 9% at the time of 10 min and it increased to 19% at the time of 30 min. For PAC, the removal rates of EOM were 12% and 29% at the time of 10 min and 30 min, respectively. With the

Ex Wavelength (nm)

a

b 450

450 0 100 200 300 400 500 600 700 800 900

400 350

A

300

T1

C

250 200 250

The MW distributions of raw, MAR-treated and PAC-treated EOM are shown in Fig. 3 in terms of DOC concentrations. It can be observed that the MW of raw EOM distributed in bimodal mode with high-MW (N 100 kDa) and low-MW (b1 kDa) fractions accounting for 47.9% and 27.7%, respectively. This result was well consistent with that reported by Henderson et al. [4] and Qu et al. [12]. After MAR pretreatment, the DOC concentration of high-MW fraction decreased by 40.5%, while the DOC concentrations of the rest portions remained almost the same, indicating that MAR mainly adsorbed high-MW fraction of EOM. Concretely, 88.7% of the EOM adsorbed by MAR was larger than 100 kDa and the percentage of high-MW fraction decreased to 35.5% after MAR pretreatment. In contrast, PAC preferentially removed low-MW fraction of EOM and the DOC concentration of low-MW fraction lowered by 82.1% after PAC pretreatment. Specifically, 78.6% of the EOM adsorbed by PAC was smaller than 1 kDa and the ratio of low-MW portion decreased dramatically to 6.9% after PAC treatment. The different effects of MAR and PAC pretreatments on MW distributions of EOM could be explained by the different pore structures of MAR and PAC. The relationship between the pore size of an adsorbent and the size of adsorbate is one of the most important factors determining the adsorption performance [36,38]. On one hand, size exclusion effect limits the accessibility of adsorbate to the internal pores if the adsorbate is larger than that of the pore of the adsorbent; on the other hand, adsorption strength declines with decreasing adsorbate size [35,36,38].

300

350

400

450

500

Ex Wavelength (nm)

DOC (mg/L)

2.0

133

350

A

300

T1

300

Em Wavelength (nm)

400

450

500

550

d 450

450 0 100 200 300 400 500 600 700 800 900

400 350

A

300

T1

C

250 200 250

350

Em Wavelength (nm)

300

350

400

450

500

Em Wavelength (nm)

550

Ex Wavelength (nm)

Ex Wavelength (nm)

c

C

250 200 250

550

0 100 200 300 400 500 600 700 800 900

400

0 100 200 300 400 500 600 700 800 900

400 350 300

T1 250 200 250

300

350

400

450

500

550

Em Wavelength (nm)

Fig. 5. Fluorescence EEM spectra of raw EOM (a), EOM filtered through a 100 kDa membrane (b), MAR-treated EOM (c) and PAC-treated EOM (d) (doses of MAR and PAC: 50 mg/L, contact time: 30 min).

K. Li et al. / Desalination 336 (2014) 129–137

TMP buildup and fouling reversibility during UF of raw EOM and EOM treated with MAR and PAC for different contact time (10 min, 30 min and 2 h) are shown in Fig. 6. It can be observed that raw EOM caused severe membrane fouling and the TMP increased to 69.8 kPa at the end of filtration. For EOM treated with MAR for 10 min, the ending TMP declined to 50.0 kPa, suggesting that MAR pretreatment can mitigate EOM fouling. When the contact time extended to 30 min and 2 h, the ending TMPs were further decreased to 36.8 and 33.8 kPa, respectively. The results indicated that TMP buildup rate decreased with the extension of contact time, and its impact became less pronounced when exceeding 30 min. Besides, PAC pretreatment was also found to reduce EOM fouling with the end TMPs of 66.5, 64.1 and 62.8 kPa for the contact time of 10 min, 30 min and 2 h, respectively. As shown in Fig. 6(b), for raw EOM, reversible fouling (50.9 kPa) accounted for almost 90% of the total fouling (57.3 kPa). When EOM was pretreated with MAR for 10 min, the reversible fouling was significantly reduced to 37.4 kPa. A further decrease in TMP was found with the contact time extended to 30 min (21.6 kPa) and 2 h (18.5 kPa). However, PAC pretreatment exhibited little ability in alleviating the reversible fouling. Both MAR and PAC pretreatments reduced more than 50% of the irreversible fouling and the contact time seemed to exert little effect on irreversible fouling control (as shown in Fig. 6(b)). Overall, the results suggested that MAR pretreatment reduced both reversible and irreversible fouling efficiently while PAC pretreatment only alleviated irreversible fouling which accounted for a small part of the total fouling. Meanwhile, the removal of membrane foulants by MAR was mainly accomplished within 30 min of contact, which was well consistent with the kinetic of EOM adsorption (Fig. 2(b)).

a 70

Raw MAR-10 min MAR-30 min MAR-2 h PAC-10 min PAC-30 min PAC-2 h

60

TMP (kPa)

50 40 30 20 10 0

50

b

UF tests were performed with raw, MAR-treated and PAC-treated EOM to investigate the efficacy of MAR and PAC pretreatments in EOM fouling control.

150

200

irreversible fouling

60

250

300

reversible fouling

50 40 30 20 10

h C-

2

in PA

m

C30

10 C-

PA

RPA

m

in

h 2

in A M

30

m

in A M

A

R-

10

m

Ra

w

0

M

3.4. Efficacy of MAR and PAC pretreatments in mitigating EOM fouling

100

Permeate volume (mL)

R-

The hydrodynamic sizes of organic molecules with MW of 100 kDa and 1 kDa are approximately 10 nm and 1 nm, respectively [39,40]. Therefore, the hydrodynamic size of high-MW EOM was larger than 10 nm while that of low-MW EOM was smaller than 1 nm. As shown in Table 1, most of the pore volume of PAC was composed of pores with widths less than 10 nm. High-MW EOM molecules could not access to these pores because of size exclusion effect, while low-MW molecules was suitable to be adsorbed. As a result, PAC preferentially adsorbed low-MW fraction of EOM. For MAR, the pore volume was mainly composed of pores with widths between 10 and 50 nm, which were large enough for adsorption of both high-MW fraction and low-MW fraction. But the adsorption strength between low-MW molecules and MAR was much smaller than that between high-MW molecules and MAR [35,36]. Accordingly, MAR mainly adsorbed high-MW fraction of EOM. In a word, because of the difference in pore structure, MAR and PAC exhibited preferential adsorption for different EOM fractions. Fig. 4 shows DAX-8/XAD-4 resin fractionation results of raw, MARtreated and PAC-treated EOM in terms of DOC concentrations. Hydrophilic and hydrophobic fractions accounted for a large proportion of raw EOM while transphilic fraction took up a very small part, which was consistent with the results reported by Henderson et al. [4] and Qu et al. [12]. MAR adsorption and PAC adsorption reduced the DOC concentration of hydrophobic fraction by 37.3% and 65.1%, respectively. By comparison, the removals of hydrophilic and transphilic fractions were negligible for both MAR and PAC adsorption. These results indicated that both MAR and PAC removed more hydrophobic fraction than hydrophilic fraction and PAC adsorbed hydrophobic fraction more preferentially than MAR did. Specifically, 69.1% of the EOM adsorbed by MAR was hydrophobic fraction while 82.8% of the EOM removed by PAC was hydrophobic fraction. Both MAR and PAC particles were negatively charged with similar zeta potentials (Table 1), and EOM molecules were reported to be neutral or negatively charged [12]. Therefore, the adsorption of EOM onto MAR and PAC was dominated by non-electrostatic interactions such as hydrophobic interaction, van der Waals forces [38]. Hydrophobic interaction was the strong attraction between hydrophobic molecules and surfaces of adsorbents [36]. MAR was made of PES which was a common hydrophobic material and PAC was also a typical hydrophobic adsorbent. Thus, it was reasonable that both MAR and PAC preferentially adsorbed hydrophobic fraction of EOM. Fluorescence EEM spectra of raw EOM, EOM filtered through a 100 kDa membrane, MAR-treated EOM and PAC-treated EOM are present in Fig. 5. It can be observed in Fig. 5(a) that raw EOM exhibits three major peaks: peak T1 which represents tryptophan-like (protein-like) substances and peaks A and C which represent humic-like substances [12,41]. As shown in Fig. 5(b), the intensity of peak T1 decreased significantly and peaks A and C displayed tiny changes when EOM was filtered through a 100 kDa membrane. The results indicated that the MW of most protein-like substances in EOM were larger than 100 kDa, while the humic-like substances were much lower in MW (b100 kDa). EEM spectrum of MAR-treated EOM (Fig. 5(c)) was similar with that of EOM filtered through a 100 kDa membrane except that the decrease of peak T1 was smaller, suggesting that MAR preferentially adsorbed high-MW protein-like substances rather than humic-like substances. As shown in Fig. 5(d), in contrast, PAC removed almost all the humic-like substances in EOM but removed only a little protein-like substances. These results demonstrated that MAR mainly removed EOM with MW larger than 100 kDa while PAC more preferentially adsorbed EOM with MW smaller than 100 kDa, which was consistent with the results of MW fractionation.

ΔTMP (kPa)

134

Fig. 6. Effects of contact time on efficacy of MAR and PAC pretreatments in EOM fouling control (doses of MAR and PAC: 50 mg/L): (a) TMP buildup (only mean values were reported), (b) Reversibility (error bars represent standard deviations from the means) (n = 3).

K. Li et al. / Desalination 336 (2014) 129–137

Raw 25 mg/L MAR 50 mg/L MAR 100 mg/L MAR 25 mg/L PAC 50 mg/L PAC 100 mg/L PAC

a 70

TMP (kPa)

60 50 40 30 20 10 0

50

100

150

200

250

300

Permeate volume (mL)

b

irreversible fouling

60

reversible fouling

ΔTMP (kPa)

50

135

high-MW fraction of EOM decreased significantly after MAR adsorption while PAC adsorption mainly removed low-MW fraction. Several studies have demonstrated that cake layer formation caused by high-MW fraction dominated UF membrane fouling caused by EOM and lowMW fraction only played an insignificant role [12–14]. Therefore, as illustrated in Fig. 8, MAR pretreatment which removed a large proportion of high-MW EOM obviously reduced the cake layer formation on membrane surface while PAC pretreatment that preferentially removed lowMW EOM had little effect on cake layer formation. As a result, MAR was much more efficient in reducing EOM fouling than PAC. With respect to fouling reversibility, reversible EOM fouling was considered to be due to cake layer formation while irreversible EOM fouling was mainly attributed to hydrophobic adhesion and pore plugging [12]. Therefore, it was easy to understand that MAR pretreatment exhibited much stronger ability in controlling reversible EOM fouling than PAC pretreatment. As shown in Fig. 4, both MAR and PAC adsorbed a proportion of the hydrophobic fraction and consequently reduced the irreversible fouling resulted form hydrophobic adhesion. Besides, MAR removed high-MW fraction of EOM and therefore decreased pore plugging. Thus, both MAR and PAC pretreatments alleviated irreversible EOM fouling to some extent.

40 3.5. Contributions of MAR and PAC particles to EOM fouling

30 20 10

C

C

PA L

L

g/ m 0 10

m 50

g/ m 25

g/

L

M L g/ m 0

10

PA

C

A

PA

R

R A M L g/

m 50

25

m

g/

L

M

Ra

A

w

R

0

Adsorbent particles themselves may also play an important role in membrane fouling when they directly contact with membranes [24,25]. To investigate the contributions of MAR and PAC particles to UF membrane fouling, membrane fouling caused by adsorbent particles in Mill-Q water (MAR and PAC) and treated EOM with adsorbent particles present in UF feed water (MAR + EOM and PAC + EOM) was evaluated and compared with that caused by MAR-treated and PAC-treated

Fig. 7. Effects of adsorbents' dose on efficacy of MAR and PAC pretreatments in EOM fouling control (contact time: 30 min): (a) TMP buildup (only mean values were reported), (b) Reversibility (error bars represent standard deviations from the means) (n = 3).

Fig. 7 presents the membrane fouling caused by EOM which was treated with MAR and PAC under the dose of 25, 50 and 100 mg/L. It can be observed in Fig. 7(a) that MAR and PAC pretreatments alleviated EOM fouling to different degrees. For MAR-treated EOM, the rate of TMP buildup declined significantly with the increase of MAR dose. When MAR dose increased to 100 mg/L, the ending TMP decreased by 61.3% in comparison with that of raw EOM. However, for PAC-treated EOM, with the PAC dose increased from 25 mg/L to 100 mg/L, the TMP at the end of filtration just declined from 64.0 kPa to 58.8 kPa. The results suggested that the ability of PAC in controlling EOM fouling was quite weak, even at the very high dose. As shown in Fig. 7(b), the reversible fouling caused by MAR-treated EOM decreased remarkably as the dose of MAR increased while PAC pretreatment only slightly reduced the reversible fouling at all doses investigated. For instance, at the same dose of 100 mg/L, MAR pretreatment reduced the reversible fouling by 74%, while it just decreased by 8% after PAC pretreatment. Nevertheless, for both MAR and PAC pretreatments, the irreversible fouling decreased steadily as the adsorbents' dose increased. These results suggested MAR can remove the major foulants in EOM while PAC can only remove the foulants responsible for the irreversible fouling which accounted for a small part of the total fouling. MAR pretreatment significantly reduced EOM fouling and its efficacy improved with the increase of MAR dose and the extension of contact time; whereas PAC pretreatment exhibited limited competence in alleviating EOM fouling. These results can be associated with the preferential adsorption of these two adsorbents. As shown in Figs. 3 and 5,

Fig. 8. Schematic diagram of the different effects of MAR and PAC pretreatments on UF membrane fouling caused by EOM.

136

K. Li et al. / Desalination 336 (2014) 129–137

a

4. Conclusions

MAR PAC MAR-treated EOM MAR+EOM PAC-treated EOM PAC+EOM

70

TMP (kPa)

60 50 40

The influence of pretreatment with MAR and PAC on characteristics of EOM and subsequent UF membrane fouling was systematically investigated in this study. The following conclusions can be drawn.

30 20 10 0

50

100

150

200

250

300

Permeate volume (mL)

b

irreversible fouling

60

reversible fouling

ΔTMP (kPa)

50 40 30

(1) The surface area of MAR was much smaller than PAC and it exhibited lower efficacy in EOM removal than PAC. (2) MAR was abundant in pores with widths between 10 and 50 nm and mainly adsorbed high-MW fraction of EOM, while PAC preferentially removed low-MW fraction. With respect to hydrophobicity, both MAR and PAC adsorbed more hydrophobic fraction than hydrophilic fraction. (3) MAR pretreatment effectively controlled EOM fouling and its efficacy improved with the increase of MAR dose and the extension of contact time, whereas PAC pretreatment exhibited limited ability in mitigating EOM fouling. (4) MAR pretreatment reduced both reversible and irreversible EOM fouling efficiently while PAC pretreatment only reduced irreversible fouling which accounted for a small part of the total fouling. (5) Both MAR and PAC particles, no matter “clean” or adhered by EOM, barely contributed to membrane fouling. Acknowledgment

20

This research was jointly supported by National Natural Science Foundation of China (Grants 51138008), Fundamental Research Funds for the Central University (Grants NSRIF.2014096), State Key Laboratory of Urban Water Resource and Environment (Grants. 2012DX11) and the Funds for Creative Research Groups of China (No. 51121062).

10

M C+

C-

PA

EO

EO

M

M ed at tre

A M PA

M

A

R-

tre

at

ed

R+

EO

EO

M

C PA

M

A

R

0

Fig. 9. Contributions of MAR and PAC particles to UF membrane fouling (doses of MAR and PAC: 50 mg/L, MAR-treated EOM and PAC-treated EOM represent treated EOM with adsorbent particles removed before UF, while MAR + EOM and PAC + EOM represent treated EOM with adsorbent particles present in UF feed water): (a) TMP buildup (only mean values were reported), (b) Reversibility (error bars represent standard deviations from the means) (n = 3).

EOM. As shown in Fig. 9(a), both MAR and PAC particles in Mill-Q water caused negligible TMP buildup; MAR + EOM caused slightly higher TMP buildup than MAR-treated EOM. It can be observed in Fig. 9(b) that both MAR and PAC particles in Mill-Q water led to mainly the reversible fouling. Compared with MAR-treated EOM, MAR + EOM caused slightly increase of reversible fouling and almost no increase of irreversible fouling was observed. PAC + EOM and PAC-treated EOM exhibited the similar trend in both TMP buildup and fouling reversibility. These results demonstrated that MAR and PAC particles, no matter “clean” or adhered by EOM, did not cause significant fouling of the PES membrane used in this experiment. Similar results obtained with MAR and allochthonous NOM have been reported by Koh et al. [28]. The result about PAC was consistent with that reported by Campinas and Rosa [20] but was contrary to the phenomenon found by Lin et al. [24] and Zhang et al. [25]. The contradictory results about PAC might be due to the difference of the characteristics of organic matter in feed water. Lin et al. (2000) and Zhang et al. (2003) used commercial humic acid solution and lake water as feed water, respectively. Both humic acid and NOM in lake water were allochthonous NOM, of which the characteristics were different from those of EOM used by Campinas and Rosa (2010) and this study. The differences of organic matter in some aspects, such as hydrophobicity and functional groups, might contribute to the different fouling behavior of PAC particles [26].

References [1] N. Her, G. Amy, H.-R. Park, M. Song, Characterizing algogenic organic matter (AOM) and evaluating associated NF membrane fouling, Water Res. 38 (2004) 1427–1438. [2] S. Babel, S. Takizawa, Microfiltration membrane fouling and cake behavior during algal filtration, Desalination 261 (2010) 46–51. [3] J. Fang, X. Yang, J. Ma, C. Shang, Q. Zhao, Characterization of algal organic matter and formation of DBPs from chlor(am)ination, Water Res. 44 (2010) 5897–5906. [4] R.K. Henderson, A. Baker, S.A. Parsons, B. Jefferson, Characterisation of algogenic organic matter extracted from cyanobacteria, green algae and diatoms, Water Res. 42 (2008) 3435–3445. [5] H. Liang, W. Gong, G. Li, Performance evaluation of water treatment ultrafiltration pilot plants treating algae-rich reservoir water, Desalination 221 (2008) 345–350. [6] Y. Zhang, J. Tian, J. Nan, S. Gao, H. Liang, M. Wang, G. Li, Effect of PAC addition on immersed ultrafiltration for the treatment of algal-rich water, J. Hazard. Mater. 186 (2011) 1415–1424. [7] B. Kwon, N. Park, J. Cho, Effect of algae on fouling and efficiency of UF membranes, Desalination 179 (2005) 203–214. [8] M.T. Hung, J.C. Liu, Microfiltration for separation of green algae from water, Colloids Surf. B: Biointerfaces 51 (2006) 157–164. [9] Y.-T. Chiou, M.-L. Hsieh, H.-H. Yeh, Effect of algal extracellular polymer substances on UF membrane fouling, Desalination 250 (2010) 648–652. [10] D.A. Ladner, D.R. Vardon, M.M. Clark, Effects of shear on microfiltration and ultrafiltration fouling by marine bloom-forming algae, J. Membr. Sci. 356 (2010) 33–43. [11] F. Qu, H. Liang, J. Tian, H. Yu, Z. Chen, G. Li, Ultrafiltration (UF) membrane fouling caused by cyanobateria: fouling effects of cells and extracellular organics matter (EOM), Desalination 293 (2012) 30–37. [12] F. Qu, H. Liang, Z. Wang, H. Wang, H. Yu, G. Li, Ultrafiltration membrane fouling by extracellular organic matters (EOM) of Microcystis aeruginosa in stationary phase: influences of interfacial characteristics of foulants and fouling mechanisms, Water Res. 46 (2012) 1490–1500. [13] W. Huang, H. Chu, B. Dong, Characteristics of algogenic organic matter generated under different nutrient conditions and subsequent impact on microfiltration membrane fouling, Desalination 293 (2012) 104–111. [14] X. Zhang, L. Fan, F.A. Roddick, Influence of the characteristics of soluble algal organic matter released from Microcystis aeruginosa on the fouling of a ceramic microfiltration membrane, J. Membr. Sci. 425–426 (2013) 23–29. [15] H. Huang, K. Schwab, J.G. Jacangelo, Pretreatment for low pressure membranes in water treatment: a review, Environ. Sci. Technol. 43 (2009) 3011–3019. [16] W. Gao, H. Liang, J. Ma, M. Han, Z.-l. Chen, J.-y. Tian, G.-b. Li, Membrane fouling control in ultrafiltration technology for drinking water production: a review, Desalination 272 (2011) 1–8. [17] M. Tomaszewska, S. Mozia, Removal of organic matter from water by PAC/UF system, Water Res. 36 (2002) 4137–4143. [18] Z. Cai, J. Kim, M.M. Benjamin, NOM removal by adsorption and membrane filtration using heated aluminum oxide particles, Environ. Sci. Technol. 42 (2008) 619–623.

K. Li et al. / Desalination 336 (2014) 129–137 [19] J. Kim, Z. Cai, M.M. Benjamin, Effects of adsorbents on membrane fouling by natural organic matter, J. Membr. Sci. 310 (2008) 356–364. [20] M. Campinas, M.J. Rosa, Assessing PAC contribution to the NOM fouling control in PAC/UF systems, Water Res. 44 (2010) 1636–1644. [21] S.K. Kang, K.H. Choo, Why does a mineral oxide adsorbent control fouling better than powdered activated carbon in hybrid ultrafiltration water treatment? J. Membr. Sci. 355 (2010) 69–77. [22] H. Humbert, H. Gallard, V. Jacquemet, J.-P. Croué, Combination of coagulation and ion exchange for the reduction of UF fouling properties of a high DOC content surface water, Water Res. 41 (2007) 3803–3811. [23] H. Huang, H.H. Cho, J.G. Jacangelo, K.J. Schwab, Mechanisms of membrane fouling control by integrated magnetic ion exchange and coagulation, Environ. Sci. Technol. 46 (2012) 10711–10717. [24] C.-F. Lin, T.-Y. Lin, O.J. Hao, Effects of humic substance characteristics on UF performance, Water Res. 34 (2000) 1097–1106. [25] M. Zhang, C. Li, M.M. Benjamin, Y. Chang, Fouling and natural organic matter removal in adsorbent/membrane systems for drinking water treatment, Environ. Sci. Technol. 37 (2003) 1663–1669. [26] C. Stoquart, P. Servais, P.R. Bérubé, B. Barbeau, Hybrid membrane processes using activated carbon treatment for drinking water: a review, J. Membr. Sci. 411–412 (2012) 1–12. [27] M.M. Clark, W.-Y. Ahn, X. Li, N. Sternisha, R.L. Riley, Formation of polysulfone colloids for adsorption of natural organic foulants, Langmuir 21 (2005) 7207–7213. [28] L.C. Koh, W.-Y. Ahn, M.M. Clark, Selective adsorption of natural organic foulants by polysulfone colloids: effect on ultrafiltration fouling, J. Membr. Sci. 281 (2006) 472–479. [29] Y. Chen, B.Z. Dong, N.Y. Gao, J.C. Fan, Effect of coagulation pretreatment on fouling of an ultrafiltration membrane, Desalination 204 (2007) 181–188. [30] N. Her, G. Amy, J. Chung, J. Yoon, Y. Yoon, Characterizing dissolved organic matter and evaluating associated nanofiltration membrane fouling, Chemosphere 70 (2008) 495–502.

137

[31] N. Lee, G. Amy, J.-P. Croué, Low-pressure membrane (MF/UF) fouling associated with allochthonous versus autochthonous natural organic matter, Water Res. 40 (2006) 2357–2368. [32] R.H. Peiris, C. Hallé, H. Budman, C. Moresoli, S. Peldszus, P.M. Huck, R.L. Legge, Identifying fouling events in a membrane-based drinking water treatment process using principal component analysis of fluorescence excitation-emission matrices, Water Res. 44 (2010) 185–194. [33] M. Bjelopavlic, G. Newcombe, R. Hayes, Adsorption of NOM onto activated carbon: effect of surface charge, ionic strength, and pore volume distribution, J. Colloid Interface Sci. 210 (1999) 271–280. [34] S.H. Wu, P. Pendleton, Adsorption of anionic surfactant by activated carbon: effect of surface chemistry, ionic strength, and hydrophobicity, J. Colloid Interface Sci. 243 (2001) 306–315. [35] L. Li, P.A. Quinlivan, D.R.U. Knappe, Effects of activated carbon surface chemistry and pore structure on the adsorption of organic contaminants from aqueous solution, Carbon 40 (2002) 2085–2100. [36] C. Moreno-Castilla, Adsorption of organic molecules from aqueous solutions on carbon materials, Carbon 42 (2004) 83–94. [37] I. Bautista-Toledo, M.A. Ferro-García, J. Rivera-Utrilla, C. Moreno-Castilla, F.J. Vegas Fernández, Bisphenol a removal from water by activated carbon. Effects of carbon characteristics and solution chemistry, Environ. Sci. Technol. 39 (2005) 6246–6250. [38] G. Newcombe, M. Drikas, Adsorption of NOM onto activated carbon: electrostatic and non-electrostatic effects, Carbon 35 (1997) 1239–1250. [39] K.J. Howe, M.M. Clark, Fouling of microfiltration and ultrafiltration membranes by natural waters, Environ. Sci. Technol. 36 (2002) 3571–3576. [40] H. Huang, N. Lee, T. Young, A. Gary, J.C. Lozier, J.G. Jacangelo, Natural organic matter fouling of low-pressure, hollow-fiber membranes: effects of NOM source and hydrodynamic conditions, Water Res. 41 (2007) 3823–3832. [41] P.G. Coble, Characterization of marine and terrestrial DOM in seawater using excitation-emission matrix spectroscopy, Mar. Chem. 51 (1996) 325–346.