Desalination 295 (2012) 1–10

Contents lists available at SciVerse ScienceDirect

Desalination

j ourna l homepage: www.e lsev ie r .com/ locate /desa l

Impact of feed spacer and membrane modification by hydrophilic, bactericidal andbiocidal coating on biofouling control

P.A. Araújo a, D.J. Miller b, P.B. Correia a, M.C.M. van Loosdrecht c, J.C. Kruithof a, B.D. Freeman b,D.R. Paul b, J.S. Vrouwenvelder a,c,d,⁎a Wetsus, centre of excellence for sustainable water technology, Agora 1, P.O. box 1113, 8900 CC Leeuwarden, The Netherlandsb University of Texas at Austin, Department of Chemical Engineering, Center for Energy and Environmental Resources, 10100 Burnet Road, Building 133, Austin, TX 78758, USAc Department of Biotechnology, Faculty of Applied Sciences, Delft University of Technology, Julianalaan 67, 2628 BC Delft, The Netherlandsd King Abdullah University of Science and Technology, Water Desalination and Reuse Center, Thuwal, Saudi Arabia

⁎ Corresponding author at: Department of BiotechnoloDelft University of Technology, Julianalaan 67, 2628 BC D15 2784169; fax: +31 15 2782355.

E-mail address: j.s.vrouwenvelder@tudelft.nl (J.S. Vr

0011-9164/$ – see front matter © 2012 Elsevier B.V. Alldoi:10.1016/j.desal.2012.02.026

a b s t r a c t

a r t i c l e i n f o

Article history:Received 29 December 2011Received in revised form 22 February 2012Accepted 23 February 2012Available online 24 April 2012

Keywords:Ultrafiltration (UF)Nanofiltration (NF)Reverse osmosis (RO)Membrane fouling simulator (MFS)Adenosine tri phosphateTotal organic carbon

The influence of polydopamine- and polydopamine-graft-poly(ethylene glycol)-coated feed spacers andmembranes, copper-coated feed spacers, and commercially-available biostatic feed spacers on biofoulinghas been studied in membrane fouling simulators. Feed spacers and membranes applied in practical mem-brane filtration systems were used; biofouling development was monitored by feed channel pressure dropincrease and biomass accumulation. Polydopamine and polydopamine-g-PEG are hydrophilic surface modifi-cation agents expected to resist protein and bacterial adhesion, while copper feed spacer coatings andbiocides infused in feed spacers are expected to restrict biological growth. Our studies showed that polydo-pamine and polydopamine-g-PEG coatings on feed spacers and membranes, copper coatings on feed spacers,and a commercial biostatic feed spacer did not have a significant impact on feed channel pressure drop in-crease and biofilm accumulation as measured by ATP and TOC content. The studied spacer and membranemodifications were not effective for biofouling control; it is doubtful that feed spacer and membrane modi-fication, in general, may be effective for biofouling control regardless of the type of applied coating.

© 2012 Elsevier B.V. All rights reserved.

1. Introduction

Membrane filtration processes like ultrafiltration (UF), nanofiltra-tion (NF) and reverse osmosis (RO) produce high-quality drinkingwater from sources such as seawater and sewage. Because the globaldemand for clean freshwater is growing, these membrane technolo-gies are increasing in importance. Biofouling – excessive growth ofbiomass which impacts membrane performance [1] – is one of themost serious problems in membrane applications, influencing (i) theamount and quality of purified water, (ii) reliability of water produc-tion, and (iii) operating costs [2]. Numerous authors have describedbiofouling problems in membrane installations [3–9]. The study ofmembrane biofouling and its control has recently intensified [10].

Presently, several strategies to prevent and control biofouling arepursued: (i) nutrient removal by biological pre-treatment (e.g. sand fil-tration), (ii) metabolic inactivation of bacteria (e.g. biocide dosage orUV radiation), (iii) membrane surface modification, and (iv) chemicalcleaning [11]. In most cases, these strategies realize a short term effect

gy, Faculty of Applied Sciences,elft, The Netherlands. Tel.: +31

ouwenvelder).

rights reserved.

only and are unsuitable for biofouling prevention or complete control[2,10,12]. If biofouling cannot be prevented, biomass must be removedfrom the module upon cleaning; otherwise rapid bacterial regrowthmay occur [9,13–15]. Membranemodification has focused on themem-brane surface itself despite the results of several studies showing thatfeed spacers play an important role in membrane system biofouling.Baker et al. [16] reported that initial fouling deposits were found accu-mulating alongside the membrane feed channel spacer; these depositseventually encroached upon the remaining free membrane area. VanPaassen et al. [17] observed an exponential increase of the feed channelpressure drop caused by biofouling on the membrane module feedspacers. The biofilm accumulation they observed proved to be causedby impure hydrochloric acid dosed to the feed water to prevent scaling.Tran et al. [18] found that fouling occurredmost readily in the vicinity ofthe feed spacer strands. Feed spacer biofouling has been addressed byperiodic air/water cleaning [19] and altering feed spacer geometry[20]; spacermodificationmay provide an alternatemeans of preventingor controlling biofouling.

The objective of this study was to determine the potential of feedspacer and membrane modification to control biofouling. Membranefouling simulators (MFS), which mimic the flow conditions and foul-ing patterns found inside commercial spiral-wound modules [9],were used to carry out studies of (i) polydopamine- and poly(ethylene

Nomenclature

ATP adenosine tri phosphate (measure for active biomass)[pg ATP cm−2]

FCP feed channel pressure drop [mbar]MFS Membrane Fouling Simulator [−]NF nanofiltration [−]PD polydopamine [−]PDPEG polydopamine-graft-poly(ethylene glycol) [−]PEG poly(ethylene glycol) [−]PP poly propylene [−]PSf polysulfone [−]PVC poly(vinyl chloride) [−]RO reverse osmosis [−]SMO254 austenitic stainless steel [−]TOC total organic carbon [mg cm−2]UV ultraviolet radiation [−]UF ultrafiltration [−]

2 P.A. Araújo et al. / Desalination 295 (2012) 1–10

glycol)-coated feed spacers and membranes, (ii) a copper-coated feedspacer and (iii) a biostatic (triclosan containing) feed spacer. Polydo-pamine is a hydrophilic surface modification agent which spontane-ously precipitates from buffered, alkaline dopamine solution underaerobic conditions [21]. Hydrophilic surfaces have previously shownresistance to protein and bacteria adhesion [22]. Poly(ethylene glycol)(PEG) may be grafted to polydopamine-modified surfaces as polymerbrushes [23]. Decreases in microbial adhesion are thought to reducethe proliferation of biofilm and, thereby, prevent or control biofouling[24,25]. Copper ions are known to possess bactericidal effects and,therefore, application of a copper coating is proposed as an antifoulingstrategy [26]. The biostatic feed spacer comprises a biocidal agent tocontrol biofouling [27]. The effect of feed spacer and membrane mod-ifications for biofouling control is discussed based on results withthese three modification techniques.

2. Materials and methods

2.1. Experimental set-up

An overview of the three biofouling control studies is shown inTable 1. Polydopamine, polydopamine-g-PEG, and copper coatingswere applied in our laboratory; the commercial biostatic feed spacerwas modified by the manufacturer and used as received. Previousstudies have shown significant improvement in organic fouling bypolydopamine and polydopamine-g-PEG modifications on ultrafiltra-tion membranes used for oil water separation. Because the polydopa-mine coating is extraordinarily thin and conformal, it may beapplied to the porous membrane without significantly decreasing

Table 1Scheme for the experimental conditions of studies on modification of feed spacers and memsame for all experiments. The substrate was dosed to a concentration in the monitor feed wcentration of 0.5 mg/L acetate C for the other experiments: polydopamine and polydopami

Study Modified material Membranetype

Spacermaterial

Water te°C

Effect of PD and PD-g-PEG coating#

In situ coating Membrane+spacer UF+NF PP 12In situ coating Membrane+spacer UF+NF PP 20Ex situ coating Membrane+spacer UF+NF PP 20

Effect of copper coating Spacer NF PP 20Effect of commercialbiostatic spacer

Spacer (as delivered) NF PP+0.5 wt.%triclosan

20

* 1 mil=25.4 μm; 31 mil=787 μm; # PD = polydopamine; PDPEG = polydopamine-g-pol

the membrane flux; polydopamine and polydopamine-g-PEG modifi-cations were therefore applied to both the ultrafiltration and nanofil-tration membranes and feed spacers. Copper-coated feed spacers andthe commercial biostatic feed spacer were used in conjunction withunmodified nanofiltration membranes.

2.2. Materials

NF and UF membranes were used in the studies. NF polyamidemembranes and polypropylene (PP) feed spacers were harvestedfrom a spiral-wound TS80 element obtained from Trisep Corp (Goleta,CA). A flat-sheet PS20 polysulfone (PSf) UF membrane was purchasedfrom Sepro Membranes, Inc (Oceanside, CA). The NF membraneswere used in the studies of Sections 3.1, 3.2 and 3.3 (Table 1); UFmembranes were used in the study of Section 3.1 (Table 1). The poly-propylene feed spacer was 31 mil (787 μm) thick with strands inter-secting at 90° and a porosity of ~0.85; these specifications match mostcommonly-used feed spacers in spiral-wound modules for water treat-ment in The Netherlands. This spacer was used for the coating studiesdescribed in Sections 3.1, 3.2 and 3.3 (Table 1). Only new membranesand spacers of all types were used. All membrane and feed spacer cou-pons (40.0 mm×200.0 mm) were cut with a punch and hydraulicpress.

Dopamine, Trizma HCl, isopropanol, ethanol, sodium acetate, sodi-um dihydrogen phosphate, and sodium nitrate were purchased fromSigma-Aldrich (St. Louis, Mo.). Sodium hydroxide was obtained fromFisher Scientific (Pittsburgh, PA). Poly(ethylene glycol) monoamine(5000 Da)was purchased from JenKemTechnology (Allen, TX). All che-micals were used as received. Tris buffer (15 mM), used in the polydo-pamine coating and poly(ethylene glycol) grafting (Sections 2.3.1 and2.3.2), was prepared by dissolving Trizma HCl (2.634 g/L) in deionizedwater and the pH of the resultant solutionwas adjusted to 8.8 using so-dium hydroxide.

2.3. Coated membranes and feed spacers

2.3.1. Polydopamine coatingPSf UF and NF membrane coupons were prepared by soaking in

isopropanol for 10 min followed by deionized water for 30 min. Tocoat membranes with polydopamine, the selective side of membranecoupons was put in contact with dopamine solution (2 g/L in Trisbuffer) for 1 h with gentle agitation to aerate the solution and to en-sure complete surface coverage. PP feed spacer coupons(0.04 m×0.20 m) were coated by immersion in stirred dopamine so-lution (2 g/L in Tris buffer) for 1 h. Modified membranes and feedspacers were soaked in ethanol for 10 min to remove residual,weakly-bound polydopamine and stored in deionized water untiluse. The characteristic brown colour of the polydopamine coatingwas imparted on membranes and feed spacers as shown in Fig. 1.

branes to evaluate effectiveness for biofouling control. The feed water flow rate was theater of 0.1 mg/l acetate C for the experiments with copper coated spacers and to a con-ne-g-PEG and the commercial biostatic spacer.

mperature Linear flowvelocitym s−1

Substrate loadmg c m−2 s−1

Feed spacerthicknessmil*

Section

0.16 Constant 31 3.10.16 Constant 31 Supplementary material0.16 Constant 31 Supplementary material0.16 Constant 31 3.2Varying Varying 31, 34 3.3

y(ethylene glycol); PP = polypropylene; UF = ultrafiltration; NF = nanofiltration.

Fig. 1. Unmodified membrane (A), polydopamine-coated membrane (B), unmodified feed spacer (C) and polydopamine-coated feed spacer (D). The brown colour indicates fullcoverage of membrane and spacer.

3P.A. Araújo et al. / Desalination 295 (2012) 1–10

2.3.2. Polydopamine-g-poly(ethylene glycol) coatingHalf of the polydopamine-modified membrane and feed spacer

coupons were further modified by grafting poly (ethylene glycol) tothe polydopamine coating (see Section 2.3.1). Polyethylene glycolmonoamine (PEG-NH2) (1 g/L) was dissolved in Tris buffer at 60 °C.Polydopamine-modified membrane coupons were put in contactwith the PEG-NH2 solution for 1 h at 60 °C with gentle agitation.Polydopamine-modified feed spacer coupons were immersed in thePEG-NH2 solution at 60 °C for 1 h with gentle mixing. Residual PEG-NH2 solution was rinsed from modified membranes and feed spacerswith deionizedwater. The in-situ coating procedure for polydopamine(Fig. S1) and polydopamine-g-poly(ethylene glycol) (Fig. S2) is shownin the Supplementary material.

2.3.3. Copper coatingPP feed spacer coupons were sputter-coated with copper by AIMT

Nederland/H&ST Heat and Surface Treatment BV, Eindhoven, TheNetherlands. The thickness of the coating was 210±5 nm (reportedby AIMT). The thicknesses of uncoated and coated feed spacers, deter-mined using an accurate thickness gauge (Mitutoyo, accuracy of0.001 mm),were both 780±10 μm, indicating that the copper coatingdid not appreciably change the feed spacer thickness. Copper-coatedfeed spacers were used in the experiments described in Section 3.2.

2.4. Biostatic feed spacer

A 34 mil thick, blue coloured, polymeric biostatic feed spacer wasprovided by a membrane module manufacturer. This commercially-available spacer contained 0.5 wt.% triclosan, an anti-biofouling com-pound, as prepared by the manufacturer. The commercial biostaticspacer had strands intersecting at 90°. A colourless feed spacer notcontaining an anti-biofouling compound but with identical geometryas the biostatic feed spacer was also provided as a control. The bio-static spacer and its similar control spacer were used in the experi-ments described in Section 3.3 (Table 1). The reference MFS feedspacer channel height is 31 mil. Plastic MFSs were made with feedspacer channel heights of 31 and 34 mil. The manufacturing accuracyof the flow channel height is ~0.04 mil (1 μm).

2.5. Membrane fouling simulator

2.5.1. DescriptionThe Membrane Fouling Simulator (MFS) has been developed to

enable the study of spiral-wound module biofouling on a smallscale. Once assembled with membrane and feed spacer coupons –

loaded with the same spatial orientation relative to feed flow as in afull-size module – the MFS reproduces the geometry of a spiral-woundmodule feed channel [28]. As a result, the hydrodynamic envi-ronment and biofouling development of a spiral-wound module arereproduced [9]. The external dimensions of all MFSs were 0.07 m×0.30 m×0.04 m; membrane and feed spacer coupon dimensionswere 0.04 m×0.20 m. A steel (SMO254) MFS, designed for 31 mil-

thick feed spacers, was used for the copper-coated feed spacer studiesdescribed in Section 3.2 (Table 1). A polyvinylchloride (PVC) MFS wasdeveloped to enable biofouling studies with feed spacers of varyingthicknesses. The PVC MFS was used for the studies described inSection 3.1 (31 mil feed spacer thickness) and in Section 3.3 (31 and34 mil feed spacer thicknesses) (Table 1). The PVC MFS showed thesame hydrodynamic behaviour and the same fouling development asthe steel MFS regardless of whether a 31 mil-thick or 34 mil-thickfeed spacer was used. Fouling development was monitored by mea-suring the pressure drop increase down the length of the MFS feedspacer channel and by analysis of membrane and spacer couponstaken from the monitor after fouling. During operation, the sight win-dow of the steel MFS was covered to prevent growth of phototrophicorganisms; the PVC MFSs do not have sight windows.

The experimental apparatus, described in detail elsewhere [29],was capable of operating several membrane fouling simulators inparallel. Equipment connected to each MFS consisted of two pressurereducing valves, a manometer, a dosing point (for biodegradablecompounds), and a flow controller. The MFSs were operated at a pres-sure of 120 kPa to avoid gas bubble formation in the MFS that mayoccur when using depressurized water. Unless mentioned otherwise,the feed water flowwas 16 L h−1 (equivalent to a linear flow velocityof 0.16 m s−1), representative of industrial practice [30]. The MFSswere operated single pass (without recirculation). The MFSs were op-erated without permeate production; previous studies demonstratedthat the presence or absence of permeate production did not affectbiofouling development [9].

2.5.2. Tap water and substrate dosageTap water prepared from anaerobic groundwater (subsequently

treated by aeration, rapid sand filtration, deacidification, softeningand rapid sand filtration at the Spannenburg treatment plant in TheNetherlands) and distributed without disinfectant residual was usedas feed water to the simulators. The feed water was passed througha heating coil in a hot water bath to maintain a temperature at 20 °C.The tap water contained indigenous bacteria, 105 bacteria cells/mLwhich is a common cell number for drinking water.

High concentrations of biodegradable organic compounds, used bymicrobes as growth substrates, in feed water resulted in rapid bio-mass accumulation and feed channel pressure drop increase [30].Thus, to ensure that the biofouling rate is suitable for laboratorystudy, the MFS feed water was dosed with organic substrate.

A solution of concentrated substrate containing sodium acetate(4.608 g), sodium dihydrogen phosphate (0.518 g), and sodium ni-trate (1.623 g) was prepared with ultrapure water in a sterilized 5 Lbottle. The C:N:P mass ratio of the resultant solution, 100:20:10, rep-resents an abundance of phosphorous relative to the typical biomassC:N:P ratio of ~100:20:4.3, ensuring that available phosphorous doesnot limit biofouling [31]. To restrict bacterial growth in the substratedosage bottle, the pHwas adjusted to 12with NaOH. The concentratedsubstrate solution was dosed into the feed water prior to the MFS by aperistaltic pump (Masterflex, Cole Parmer, VernonHills, IL) at a flow of

4 P.A. Araújo et al. / Desalination 295 (2012) 1–10

0.03 L h−1. Because the substrate dosage flow rate (0.03 L h−1) waslow compared to the feed water flow rate (16 L h−1), the high pH ofthe concentrated substrate solution did not affect the pH of the feedwater. The dosage of substrate was checked periodically by mass.Stock solution bottles were replaced after 5 days of operation. Thesubstrate was dosed to a concentration in the monitor feed water of0.1 mg/L acetate C for the experiments with copper coated spacers(Section 3.2) and to a concentration of 0.5 mg/L acetate C for the otherexperiments: polydopamine and polydopamine-g-PEG (Section 3.1)and the commercial biostatic spacer (Section 3.3).

2.6. Pressure drop

Pressure drop measurements were performed with a differentialpressure transducer (Deltabar S PMD70, Endress+Hauser, Reinach,Switzerland). The calibrated measuring range was 0–50 kPa [32]. Avalve system, developed to enable measurement and logging thepressure drop over several monitors simultaneously, was used in con-junction with the differential pressure transducer.

2.7. Sampling and study of membrane and spacer coupons

At the conclusion of each study, the MFSs operated in parallelwere simultaneously autopsied. The same sampling methods and an-alyses were performed for all studies. ATP and TOC concentrationmeasurements, which are indicators of living biomass and total accu-mulated biomass, are well-established metrics of biofouling [33–35].Samples (4 cm2) were cut from membrane and feed spacer couponsupon MFS disassembly. Each section (4 cm2) was placed in 20 mL ofautoclaved tap water (for ATP analysis) or ultrapure water (for TOC

A

C

00 5 100

100

200

300

400

500

pres

sure

dro

p(m

bar)

time (days)

: -S, :+S:PD+S, :PDPEG+S

-S +S102

103

104

105

106

biom

ass

conc

entr

atio

n(p

g A

TP

/cm

2 )

PD PDPEG

Fig. 2. Pressure drop in time (A), pressure drop increase after 11 days (B) and biomass concwithout dopamine coating (PD and PDPEG) fed with water with (+S) and without organic scoated in the constructed monitors (in-situ). The feed water temperature was 12 °C.

analysis) in a capped tube. To dislodge the accumulated biomass,the tubes were placed in an ultrasonic cleaning bath (Bransonic,model 5510E-DTH, output 135 W, 42 kHz). A low-energy sonic treat-ment (2 min), followed by mixing on a Vortex (few seconds), was re-peated three times. If the liquid was not visually homogeneous or ifbiomass was not sufficiently removed from the coupons, a higher-energy sonic treatment in an ice bath was carried out. Aliquots col-lected from the tubes were used to determine ATP and TOC concen-trations as described elsewhere [9].

3. Results

The development of feed channel pressure drop and biomass ac-cumulation in membrane fouling simulators was investigated with:polydopamine and polydopamine-g-PEG coatings of feed spacersand membranes (Figs. 1 and 2), copper coatings of feed spacers(Figs. 3 and 4) and a commercial biostatic spacer containing a biocide(Figs. 5–7). Feed channel pressure drop evolution was monitoredduring biofilm development over time, while accumulated biomasswas measured when the MFSs were autopsied.

3.1. Effect of polydopamine- and polydopamine-g-poly(ethylene glycol)-coated membranes and feed spacers

Biofouling in MFSs containing polydopamine- and polydopamine-g-PEG-coated UF membranes and feed spacers was compared to bio-fouling in an MFS with uncoated UF membranes and feed spacers.An MFS with unmodified membranes and feed spacers was also oper-ated with no growth substrate dosage. Experimental details are out-lined in Table 1. The feed channel pressure drop increased similarly

D

B

-S +S PD PDPEG0

100

200

300

400

500

pres

sure

dro

p in

crea

se(m

bar)

-S +S101

102

103

104

biom

ass

conc

entr

atio

n(m

gTO

C/c

m2 )

PD PDPEG

entration measured as ATP (C) and TOC (D) after 11 days monitor operation with andubstrate dosage (−S) at the same cross flow velocity. The spacers and membranes were

Fig. 3. Unmodified PP feed spacer (A) and copper-coated PP feed spacer (B), showingcomplete and even spacer coverage. Reflections from flash photography are apparentin (B).

0 7 140

100

200

300

400

pres

sure

dro

p(m

bar)

time (days)

:blank-S:Cu:blank

A B

blank-S Cu blank0

25

50

75

100

125

biom

ass

conc

entr

atio

n(p

g AT

P/c

m2

* 10

3 )

C D

test condition

Fig. 4. Pressure drop in time (A), pressure drop increase after 15 days of operation (B), andMFSs operated at 0.16 m s−1 crossflow velocity at 20 °C. Blank-S = unmodified PP feed spblank = unmodified PP feed spacer, with substrate dosage.

5P.A. Araújo et al. / Desalination 295 (2012) 1–10

over 11 days in all of the MFSs receiving growth substrate dosing, re-gardless of whether the membranes and feed spacers were coated ornot. The feed channel pressure drop over the MFS not receiving sub-strate dosing did not increase over the same experimental duration,indicating that little or no biomass accumulated in the feed channel(Fig. 2A and B). After 11 days of operation, the MFSs were openedfor analysis of the accumulated material. All MFSs supplied with sub-strate showed two orders of magnitude greater ATP and TOC concen-trations (Fig. 2C and D) than the MFS receiving no substrate dosing.This result corroborates the feed channel pressure drop measure-ments, suggesting that biomass accumulation in all MFSs fed withgrowth substrate was similar.

Two additional experiments were performed at 20 °C with in-situand ex-situ coating of spacers and UF membranes. As in the first ex-periment, both coated and unmodified membranes and spacers withsubstrate dosage led to an increased pressure drop in theMFS and bio-mass accumulated. The pressure drop increased and biomass amountsremained low in the systemswithout substrate dosage (Figs. S3 and S4in the Supplementary material). Increasing the water temperaturefrom 12 °C to 20 °C reduced the running time from 11 to 4 days, likelydue to higher bacterial growth rates at the higher temperatures. At20 °C, in the MFSs with substrate dosage the pressure drop increasedand biomass accumulated as observed during the experiment at12 °C (Fig. 2). The method of coating (in-situ or ex-situ) had no effecton the pressure drop development and biomass accumulation (resultsfor in-situ coating shown in Figs. 2 and S3 and ex-situ coating shownin Fig. S4).

test condition

blank-S Cu blank0

100

200

300

400

pres

sure

dro

p in

crea

se(m

bar)

blank-S Cu blank0

100

200

300

biom

ass

conc

entr

atio

n(m

g TO

C/c

m2 )

test condition

biomass concentration measured as ATP (C) and TOC (D) after 15 days of operation. Allacer, no substrate dosage; Cu = copper-coated PP feed spacer, with substrate dosage;

Fig. 5. Feed spacers used in the study to determine the effect of the biostatic feed spacer on biofouling: Unmodified PP feed spacer, 31 mil thick (A), commercial biostatic feed spacer,34 mil thick (B) and unmodified feed spacer with geometry identical to biostatic feed spacer, 34 mil thick (C). The commercial biostatic spacer has a distinctive blue colour. The31 mil thick PP feed spacer is the standard feed spacer used in our MFS studies.

6 P.A. Araújo et al. / Desalination 295 (2012) 1–10

The studies performedwith feed spacers and UF (Figs. 2 and S3 andS4 in the Supplementary material) and NF membranes (data notshown) all showed a similar development in pressure drop and similarbiomass accumulation for the unmodified and PD and PDPEGmodifiedmaterials for water containing a biodegradable compound. Withoutsubstrate dosage always a stable pressure drop and low biomass accu-mulation were found. Clearly, polydopamine and polydopamine-g-PEG modifications of membranes and feed spacers did not inhibitbiofouling.

31-S 34-S 34 34B 31102

103

104

105

106

biom

ass

conc

entr

atio

n(p

g AT

P/c

m2 )

feed spacer thickness (mil)

nd

0 2 4 6

0

500

1000

pres

sure

dro

p(m

bar)

time (days)

:31-S:34-S:34:34B:31

C

A

Fig. 6. Pressure drop in time (A), pressure drop increase after 6 days of operation (B) and biooperated at 0.16 m s−1 crossflow velocity at 20 °C. 31-S = unmodified PP feed spacer, 31 mibiostatic feed spacer, 34 mil thick, no substrate dosage; 34 = unmodified feed spacer withcommercial biostatic feed spacer, 34 mil thick, with substrate dosage; 31 = unmodified PP ffeed spacer used in our MFS studies. nd = not determined.

3.2. Effect of copper-coated feed spacer

Membrane fouling simulators were also used to study the effect ofcopper feed spacer coatings on biofouling. Three MFSs were operatedin parallel (Table 1), all containing unmodified NF membranes and PPfeed spacers, one of which was modified with a ~0.20 μm-thick cop-per coating (Fig. 3). Over the 15-day experiment, the MFS containingthe copper-coated feed spacer showed a slower feed channel pressuredrop evolution than the MFS containing the unmodified spacer

31-S 34-S 34 34B 31101

102

103

104

biom

ass

conc

entr

atio

n(m

g TO

C/c

m2 )

feed spacer thickness (mil)

nd

31-S 34-S 34 34B 31

0

500

1000

pres

sure

dro

p in

crea

se(m

bar)

feed spacer thickness (mil)

B

D

mass concentration measured as ATP (C) and TOC (D) after 6 days of operation. All MFSsl thick, no substrate dosage; 34-S = unmodified feed spacer with geometry identical togeometry identical to biostatic feed spacer, 34 mil thick, with substrate dosage; 34B =eed spacer, 31 mil thick, with substrate dosage. The 31 mil-thick spacer is the standard

Fig. 7. Fouled biostatic 34 mil feed spacer after 6 days of operation in the opened MFS. Slimy yellowish-coloured material was visible with the naked eye. Analysis of accumulatedmaterial revealed high concentrations of ATP and TOC, indicative of biofouling. The arrow indicates the flow direction.

7P.A. Araújo et al. / Desalination 295 (2012) 1–10

(Fig. 4A and B). Lower levels of biomass were also associated with thecopper-coated spacer relative to the unmodified spacer (Fig. 4C andD). However, compared to the MFS with an unmodified feed spacerwhich was not dosed with substrate, the MFSs containing both theunmodified and copper-coated receiving growth substrate showedmuch higher feed channel pressure increase and biomass concentra-tions (Fig. 4A–D). The MFS which did not receive substrate dosageshowed no feed channel pressure drop increase and very low ATPand TOC levels, indicative of little or no biofouling. The copper feedspacer coating, therefore, was able to delay, but not prevent, biofoul-ing development.

3.3. Effect of a commercial biostatic feed spacer

The influence of a commercially-available biostatic feed spacer onbiofouling was studied. Six MFSs were operated in parallel (Table 1),containing standard 31 mil PP feed spacers, the commercial 34 milbiostatic feed spacer (34B), and 34 mil feed spacers not containing abiocide (Fig. 5). To enable comparison, the MFSs were operated withthe same substrate load [30]. The unmodified 34 mil and commercialbiostatic feed spacers had the same spatial geometry, differingvisually in colour only. According to the manufacturer, the blue-coloured biostatic spacer should reduce or prevent biofouling byvirtue of impregnation with biocide. A rapid increase of pressuredrop over all MFSs receiving organic substrate was observed(Fig. 6A and B), including the MFSs containing the biostatic feedspacer (34B). The feed channel pressure drops of the MFSscontaining both the biostatic and unmodified 34 mil spacers rosesimilarly and were somewhat lower than the feed channel pres-sure drop which developed in the MFS with the unmodified31 mil PP spacer. As in the other studies, no pressure drop increasewas observed in the MFSs not receiving substrate dosage, suggest-ing little to no biofouling. After six days, all MFSs were opened forvisual observations and for measurements of ATP and TOC concen-trations. Biomass accumulated in all MFSs supplied with substrateincluding those containing the biostatic feed spacer (Fig. 7). Thisvisual observation was confirmed by the similarly-high concen-trations of ATP and TOC in MFSs receiving growth substrate relativeto those not dosed with substrate (Fig. 6C and D). The feed channelpressure drop development, visual observations, and accumulatedbiomass analysis all unambiguously indicated that fouling accumula-tion occurred when substrate was present, regardless of the use ofa biostatic feed spacer. Replicate experiments showed similar results(data not shown). Use of this biostatic feed spacer, therefore, didnot prevent biofouling.

4. Discussion

The objective of this study was to determine the potential of feedspacer and membrane modification with coatings and/or materialadditions to control biofouling. Polypropylene feed spacers and NFmembranes harvested from a commercially-available spiral woundmembrane module and flat sheet UF membranes were used in ourstudies. A commercial anti-biofouling feed spacer was also used. Bio-fouling, measured by feed channel pressure drop increase and biomassaccumulation measurements, was not prevented by use of polydopa-mine and polydopamine-g-PEG coatings of feed spacer andmembranes(Fig. 2), by copper feed spacer coating (Fig. 4), or by the biostatic feedspacer (Figs. 6 and 7).

4.1. Relevance of reporting ineffective biofouling control approaches

Yu et al. [36] reported the application of anti-fouling bioactive sur-faces based on poly(ethylene glycol) (PEG) and polymer brushes.Many recent studies have addressed anti-biofouling surfaces in mem-brane systems. Positive effects are often shown only in very short-term or batch studies. In general, pure microbial cultures are used tostudy biofouling inhibition effects, whereas in practice fouling commu-nities are highly complex microbial ecosystems [37,38]. Critical evalua-tions hardly exist. For instance, polymer brushes have been shown todecrease microbial adhesion [39–41], therefore they are considered inanti-biofouling applications. Brzozowska [42], however, reported thatthe structure of adsorbed polymer brush layers may easily be disruptedby shear forces, destroying the complex coacervate-brush structure.Many laboratory studies on biofouling control in membrane systemshave been reported under non-representative conditions for practice.A biofouling control approach effective under non-representativelaboratory conditions may not be predictive for industrial practice. Ingeneral, ineffective approaches to control biofouling are not reported.To develop a comprehensive understanding of the state of the art in con-trolling biofouling, it is essential to address these ineffective approaches.

It is strongly recommended that, for proposed biocidal or inhibitivecoatings on spacers or membranes, long-term tests in full scale mem-brane units or representative membrane fouling simulators are usedinstead of the often used short term or batch tests with pure microbialcultures.

4.2. Impact of feed spacer and membrane modification withpolydopamine and polydopamine-g-poly(ethylene glycol)

Hydrophilic surface modifications have been shown to improvefouling performance of reverse osmosis [43,44] and ultrafiltration

8 P.A. Araújo et al. / Desalination 295 (2012) 1–10

[45] membranes. A thin layer application of polydopamine dramati-cally increased the hydrophilicity of many surfaces, includingnatively-hydrophobic membranes [23]. Hydrophilic surfaces bind awater buffer layer through hydrophilic interactions, effectively mini-mizing the hydrophobic interactions responsible for adhesion of hy-drophobic foulants to the membrane. Poly(ethylene glycol), usedthroughout the anti-fouling literature [22], may be grafted topolydopamine-treated surfaces to further increase hydrophilicity,creating a thick water layer on the membrane as water moleculescluster around the grafted PEG chains. Several membranes haveshown improved oily water anti-fouling characteristics upon coatingwith polydopamine or polydopamine-g-PEG [23] and, as a result,these coatings were of interest for potential use as an anti-biofoulant.

Our studies clearly showed that the polydopamine andpolydopamine-g-PEG coatings of UF and NF membranes and PP feedspacers failed to inhibit biofouling (Figs. 2, S3 and S4 in the Supplemen-tary material). MFSs containing polydopamine and polydopamine-g-PEG surface-modified membranes and feed spacers showed behavioursimilar to those containing unmodified spacer and membrane samples.The hydrophilic coatings were not able to suppress biofouling andthe MFSs with modified membranes and feed spacers fouled in amanner similar to MFSs containing unmodified membranes and spacerreceiving substrate dosing. Although the modified surface may resisthydrophobic interactions with biomacromolecules or microbial cells,especially after short times, diminished adhesion has been shown tobe a poor indicator of biofilm formation [46]. Once some biopolymersand/or microbial cells are able to adsorb to the modified surface, evenin reduced numbers, further adhesion atop this conditioning layer is fa-cilitated and biomass accumulates [11].

4.3. Impact of feed spacer modification with copper

Copper ions are known biocides. The effect of copper-charged feedspacers on biofouling control has been reported by Hausman et al.[26,47]. In batch experiments, fewer bacteria were observed growingnear a copper-charged feed spacer relative to an unmodified spacer[47]. In a 48 h cross flow filtration monitor experiment, a 25% flux re-duction was observed with a copper-coated spacer, lower than the70% flux reduction with an unmodified spacer [26]. In batch experi-ments, a significant leaching of copper was observed [26]. Patil et al.[48] reported that biofilms were not affected by the presence of anti-microbial compounds such as copper and silver ions, showing thatmicrobial biofilms have a high degree of resistance to antimicrobialagents. In our parallel studies, a delay in the increase of the feed chan-nel pressure drop was observed for the copper-coated spacer relativeto the uncoated spacer receiving substrate dosage (Fig. 4). Biofoulingwas not, however, prevented. In general, short delays in fouling accu-mulation will not satisfy long-term operational requirements of full-scale membrane installations. However, slowing down the fouling ac-cumulation rate may give plant operators more time to identify theproblem and to take measures at an earlier stage of fouling. The extra-cellular matrix polymers secreted by the biofilm may mask the toxic-ity of the coating, permitting accumulation of microbes. Somemicrobial strains, resistant to the metal coating, may first populatethe surface, paving the way for other bacteria to accumulate on thecoated material.

4.4. Impact of commercial biostatic feed spacer

A patented, commercially-available biostatic feed spacer [27] is in-fusedwith triclosan, a biocidal agent. In the patent, an experimentwasdescribed where coupons of the biocidal feed spacer containing 0.5wt.% triclosan and a virgin control spacer were placed on an agardish inoculated with bacteria. Less bacterial growth was observed onthe agar surrounding the biocidal feed spacer than on the agar sur-rounding the control spacer [27], suggesting that the biocidal agent

leached out of the feed spacer. The cross flow in membrane moduleswould likely result in more rapid leaching of the active compoundthan observed in laboratory batch and agar plate tests. Our MFS stud-ies with the biocidal feed spacer clearly showed that, within a 6-dayresearch period, biofouling was not prevented (Figs. 6 and 7). Thefeed channel pressure drop evolution of the MFSs containing the bio-static and 34 mil unmodified feed spacers was somewhat slower thanthat of theMFSwith the 31 mil feed spacer receiving substrate dosage,but ATP and TOC measurements showed comparable biomass accu-mulation for all MFSs. As noted in the discussion of copper-coatedfeed spacers, a relatively short delay in biofouling developmentwould provide little benefit to long-term industrial installations. Addi-tionally, leaching of active compounds from modified spacers inpractice means that the (i) effectiveness in preventing biofilm accu-mulation will deteriorate in time, (ii) leached compounds will be pre-sent in the concentrate and, depending upon the membrane used, thepermeate, potentially in conflict with discharge, quality and healthregulations, and (iii) reapplying the active compound in membranemodules is periodically required to maintain biofilm growth control.

Evidently, modification of feed spacers using hydrophilic coatings,such as polydopamine and polydopamine-g-PEG, and biocidal agents,such as copper and triclosan, was not effective in controlling biofouling(Figs. 2, 4 and 6).We postulate that, in general, attempts to prevent mi-crobial proliferation in membrane modules with hydrophilic orbactericidal/biocidal coatings will be not effective. Adaptation of spacergeometry [20] and/or membrane operation may be more effective andsustainable at a lower cost [10].

4.5. Perspective of feed spacer and membrane modification for biofoulingcontrol

Evaluation of the literature and the experiments reported here re-veals that membrane and spacer modifications to reduce biofoulingmust be studied in long-term flow-through tests with a natural micro-bial population, not in short-term, batch tests with puremicrobial cul-tures. Microbial adhesion and biofilm formation are related butinfluenced by very different factors [46]. Applying a surface modifica-tion to reduce adhesion is, therefore, not adequate for biofouling pre-vention or limitation. Moreover, the presence of various biopolymersin pre-treated water (partly due to microbial growth in storage reser-voirs [37,38]), even if microorganisms adhere in reduced numbers rel-ative to an unmodified surface, may result in the development of aconditioning film, making the surface more susceptible to further ad-hesion. Toxic coatings are also not a good solution. If the compound re-mains on the surface (copper, silver etc.) the first layer of microbialcells might be killed, but on top a biofilm can still be formed. Foulingoccurs on surfaces even when exposed to lysed microorganisms.Therefore, in order to be effective, the compounds must leach awayfrom the membrane or feed spacer. As the active compound leachesout of the membrane or feed spacer, its biocidal effect will diminishand, eventually, disappear completely. In addition, the presence ofthe toxic compound in the concentrate or permeate may result in un-acceptable discharge and/or productwater quality. The efficacy of feedspacer andmembranemodificationwith toxic coatings and/ormateri-al additions for biofouling control is, at best, questionable.

5. Conclusions

Feed spacer modification studies to reduce biofouling should beexecuted in longer-term tests mimicking the practical conditions ofmembrane filtration in order to be predictive.

Our researchwas focused on the application of membranes and feedspacers currently applied in spiral wound UF, NF and RO membranemodules for the treatment of pre-treated ground water. Biofouling ischaracterized by feed channel pressure drop and biofilm characteristicsmeasured by ATP and TOC. Under these conditions evaluation of feed

9P.A. Araújo et al. / Desalination 295 (2012) 1–10

spacer and membrane modification with hydrophilic, bactericidal and/or biocidal coatings as an anti-biofouling strategy led to the conclusionthat biofouling is not prevented by using:

(i) polydopamine-coated feed spacers and membranes,(ii) polydopamine-g-poly(ethylene glycol)-coated feed spacers

and membranes,(iii) a copper-coated feed spacer,(iv) a commercially-available, triclosan-based, anti-biofouling feed

spacer.

Supplementary materials related to this article can be found onlineat doi:10.1016/j.desal.2012.02.026.

Acknowledgements

This work was performed by Wetsus, Centre of Excellence forSustainable Water Technology and Delft University of Technology.Wetsus is funded by the Dutch Ministry of Economic Affairs. The au-thors like to thank the participants of the Wetsus theme ‘Biofouling’and Evides waterbedrijf for the fruitful discussions and their financialsupport. The input of Marcel Boorsma, Paul Buijs, Florian Beyer, MaudVillain, Harold Hollander, Harm van der Kooi, Wim Borgonje and ArieZwijnenburg is fully acknowledged. The authors gratefully acknowl-edge the National Science Foundation Graduate Research FellowshipProgram (0648993) and the National Science Foundation Scienceand Technology Center for Layered Polymeric Systems (DMR-0423914) for support.

References

[1] H.-C. Flemming, Biofouling in water systems—cases, causes and countermeasures,Appl. Microbiol. Biotechnol. 59 (2002) 629–640.

[2] M.A. Shannon, P.W. Bohn, M. Elimelech, J.G. Georgiadis, B.J. Marinãs, A.M. Mayes,Science and technology for water purification in the coming decades, Nature 452(2008) 301–310.

[3] H.F. Ridgway, A. Kelly, C. Justice, B.H. Olson, Microbial fouling of reverse-osmosismembranes used in advanced wastewater treatment technology: chemical,bacteriological, and ultrastructural analyses, Appl. Environ. Microbiol. 45 (1983)1066–1084.

[4] K. Tasaka, T. Katsura, H. Iwahori, Y. Kamiyama, Analysis of RO elements operatedat more than 80 plants in Japan, Desalination 96 (1994) 259–272.

[5] H.F. Ridgway, H.C. Flemming, Membrane biofouling, in: J. Mallevialle, P.E.Odendaal, M.R. Wiesner (Eds.), Water Treatment Membrane Processes,McGraw-Hill, New York, 1996, pp. 6.1–6.62.

[6] J.S. Baker, L.Y. Dudley, Biofouling in membrane systems—a review, Desalination118 (1998) 81–90.

[7] R.P. Schneider, L.M. Ferreira, P. Binder, E.M. Bejarano, K.P. Góes, E. Slongo, C.R.Machado, G.M.Z. Rosa, Dynamics of organic carbon and of bacterial populationsin a conventional pre-treatment train of a reverse osmosis unit experiencing se-vere biofouling, J. Membr. Sci. 266 (2005) 18–29.

[8] J.S. Vrouwenvelder, S.A. Manolarakis, J.P. Van der Hoek, J.A.M. Van Paassen, W.G.J.Van der Meer, J.M.C. Van Agtmaal, H.D.M. Prummel, J.C. Kruithof, M.C.M. VanLoosdrecht, Quantitative biofouling diagnosis in full scale nanofiltration and re-verse osmosis installations, Water Res. 42 (2008) 4856–4868.

[9] J.S. Vrouwenvelder, D.A. Graf von der Schulenburg, J.C. Kruithof, M.L. Johns,M.C.M. van Loosdrecht, Biofouling of spiral wound nanofiltration and reverse os-mosis membranes: a feed spacer problem, Water Res. 43 (2009) 583–594.

[10] J.S. Vrouwenvelder, J.C. Kruithof, M.C.M. van Loosdrecht, Integrated approach forbiofouling control, Water Sci. Technol. 62 (2010) 2477–2490.

[11] J. Mansouri, S. Harrisson, V. Chen, Strategies for controlling biofouling in mem-brane filtration systems: challenges and opportunities, J. Mater. Chem. 20(2010) 4567–4586.

[12] A. Matin, A. Khan, S.M.J. Zaidi, M.C. Boyce, Biofouling in reverse osmosis mem-branes for seawater desalination: phenomena and prevention, Desalination 281(2011) 1–16.

[13] J.W.N.M. Kappelhof, H.S. Vrouwenvelder, M. Schaap, J.C. Kruithof, D. van derKooij, J.C. Schippers, An in situ biofouling monitor for membrane systems,Water Sci. Technol. Water Supply 3 (2003) 205–210.

[14] L.A. Bereschenko, H. Prummel, G.J.W. Euverink, A.J.M. Stams,M.C.M. van Loosdrecht,Effect of conventional chemical treatment on the microbial population in a biofoul-ing layer of reverse osmosis systems, Water Res. 45 (2011) 405–416.

[15] S.A. Creber, J.S. Vrouwenvelder, M.C.M. van Loosdrecht, M.L. Johns, Chemicalcleaning of biofouling in reverse osmosis membranes evaluated using magneticresonance imaging, J. Membr. Sci. 362 (2010) 202–210.

[16] J. Baker, T. Stephenson, S. Dard, P. Cote, Characterisation of fouling of nanofiltra-tion membranes used to treat surface waters, Environ. Technol. 16 (10) (1995)977–985.

[17] J.A.M. Van Paassen, J.C. Kruithof, S.M. Bakker, F.S. Kegel, Integrated multi-objective membrane systems for surface water treatment: pre-treatment ofnanofiltration by riverbank filtration and conventional ground water treatment,Desalination 118 (1998) 239–248.

[18] T. Tran, B. Bolto, S. Gray, M. Hoang, E. Ostarcevic, An autopsy study of a fouled re-verse osmosis membrane element used in a brackish water treatment plant,Water Res. 41 (2007) 3915–3923.

[19] E.R. Cornelissen, J.S. Vrouwenvelder, S.G.J. Heijman, X.D. Viallefont, D. Van DerKooij, L.P. Wessels, Periodic air/water cleaning for control of biofouling in spiralwound membrane elements, J. Membr. Sci. 287 (2007) 94–101.

[20] K. Majamaa, P.E.M. Aerts, C. Groot, L.L.M.J. Paping,W. van den Broek, S. van Agtmaal,Industrial water reuse with integrated membrane system increases the sustainabil-ity of the chemical manufacturing, Desalin. Water Treat. 18 (2010) 17–23.

[21] H. Lee, S.M. Dellatore, W.M. Miller, P.B. Messersmith, Mussel-inspired surfacechemistry for multifunctional coatings, Science 318 (2007) 426–430.

[22] R.G. Chapman, E. Ostuni, M.N. Liang, G. Meluleni, E. Kim, L. Yan, G. Pier, H.S.Warren, G.M. Whitesides, Polymeric thin films that resist the adsorption of pro-teins and the adhesion of bacteria, Langmuir 17 (2001) 1225–1233.

[23] B.D. McCloskey, H.B. Park, H. Ju, B.W. Rowe, D.J. Miller, B.J. Chun, K. Kin, B.D.Freeman, Influence of polydopamine deposition conditions on pure water fluxand foulant adhesion resistance of reverse osmosis, ultrafiltration, and microfil-tration membranes, Polymer 51 (2010) 3472–3485.

[24] X. Fan, L. Lin, P.B. Messersmith, Cell fouling resistance of polymer brushes graftedfrom Ti substrates by surface-initiated polymerization: effect of ethylene glycolside chain length, Biomacromolecules 7 (2006) 2443–2448.

[25] B. Fang, S. Gon, M. Park, K.-N. Kumar, V.M. Rotello, K. Nusslein, M.M. Santore, Bac-terial adhesion on hybrid cationic nanoparticle-polymer brush surfaces: ionicstrength tunes capture from monovalent to multivalent binding, Colloids Surf. BBiointerfaces 87 (2011) 109–115.

[26] R. Hausman, T. Gullinkala, I.C. Escobar, Development of copper-charged polypro-pylene feed spacers for biofouling control, J. Membr. Sci. 358 (2010) 114–121.

[27] I. Shelby, C.R. Bartels, Cross flow apparatus with biocidal feed spacer. UnitedStates Patent Application, Publication 2009/0314713. Dec. 24, 2009.

[28] J.S. Vrouwenvelder, J.A.M. Van Paassen, L.P. Wessels, A.F. Van Dam, S.M. Bakker,The membrane fouling simulator: a practical tool for fouling prediction and con-trol, J. Membr. Sci. 281 (2006) 316–324.

[29] J.S. Vrouwenvelder, S.M. Bakker, M. Cauchard, R. Le Grand, M. Apacandie, M.Idrissi, S. Lagrave, L.P. Wessels, J.A.M. van Paassen, J.C. Kruithof, M.C.M. vanLoosdrecht, The membrane fouling simulator: a suitable tool for prediction andcharacterisation of membrane fouling, Water Sci. Technol. 55 (2007) 197–205.

[30] J.S. Vrouwenvelder, C. Hinrichs, W.G.J. van der Meer, M.C.M. van Loosdrecht, J.C.Kruithof, Pressure drop increase by biofilm accumulation in spiral wound ROand NF membrane systems: role of substrate concentration, flow velocity, sub-strate load and flow direction, Biofouling 25 (2009) 543–555.

[31] J.S. Vrouwenvelder, F. Beyer, K. Dahmani, N. Hasan, G. Galjaard, J.C. Kruithof, M.C.M.van Loosdrecht, Phosphate limitation to control biofouling, Water Res. 44 (2010)3454–3466.

[32] J.S. Vrouwenvelder, J.A.M. van Paassen, J.C. Kruithof, M.C.M. van Loosdrecht, Sen-sitive pressure drop measurements of individual lead membrane elements for ac-curate early biofouling detection, J. Membr. Sci. 338 (2009) 92–99.

[33] O. Holm-Hansen, C.R. Booth, The measurement of adenosine triphosphate in theocean and its ecological significance, Limnol. Oceanogr. 11 (1966) 510–519.

[34] S. Kim, S. Lee, S. Hong, Y. Oh, M. Seoul, J. Kweon, et al., Biofouling of reverse osmo-sis membranes: microbial quorum sensing and fouling propensity, Desalination247 (2009) 303–315.

[35] C. Coufort, N. Derlon, J. Ochoa-Chaves, A. Liné, E. Paul, Cohesion and detachmentin biofilm systems for different electron acceptor and donors, Water Sci. Technol.55 (2007) 421.

[36] Q. Yu, Y. Zhang, H. Wang, J. Brash, H. Chen, Anti-fouling bioactive surfaces, ActaBiomater. 7 (2011) 1550–1557.

[37] C.-L. Chen, W.-T. Liu, M.L. Chong, M.-T. Wong, S.L. Ong, H. Seah, W.J. Ng, Commu-nity structure of microbial biofilms associated with membrane-based water puri-fication processes as revealed using a polyphasic approach, Appl. Microbiol.Biotechnol. 63 (2004) 466–473.

[38] L.A. Bereschenko, G.H.J. Heilig, M.M. Nederlof, M.C.M. van Loosdrecht, A.J.M.Stams, G.J.W. Euverink, Molecular characterization of the bacterial communitiesin the different compartments of a full-scale reverse-osmosis water purificationplant, Appl. Environ. Microbiol. 74 (2008) 5297–5304.

[39] A. Roosjen, H.J. Kaper, H.C. van der Mei, W. Norde, H.J. Busscher, Inhibition ofadhesion of yeasts and bacteria by poly(ethylene oxide)-brushes on glass in aparallel plate flow chamber, Microbiology 149 (2003) 3239–3246.

[40] I. Cringus-Fundeanu, J. Luijten, H.C. Van Der Mei, H.J. Busscher, A.J. Schouten, Syn-thesis and characterization of surface-grafted polyacrylamide brushes and theirinhibition of microbial adhesion, Langmuir 23 (2007) 5120–5126.

[41] J.E. Raynor, T.A. Petrie, A.J. García, D.M. Collard, Controlling cell adhesion to titani-um: functionalization of poly[oligo(ethylene glycol)methacrylate] brushes withcell-adhesive peptides, Adv. Mater. 19 (2007) 1724–1728.

[42] A.M. Brzozowska, E. Spruijt, A. de Keizer, M.A. Cohen Stuart, W. Norde, On the sta-bility of the polymer brushes formed by adsorption of ionomer complexes onhydrophilic and hydrophobic surfaces, J. Colloid Interface Sci. 353 (2011)380–391.

[43] A.C. Sagle, H. Ju, B.D. Freeman, M.M. Sharma, PEG-based hydrogel membranecoatings, J. Membr. Sci. 50 (2009) 756–766.

10 P.A. Araújo et al. / Desalination 295 (2012) 1–10

[44] E.M. Van Wagner, A.C. Sagle, M.M. Sharma, Y.-H. La, B.D. Freeman, Surface modi-fication of commercial polyamide desalination membranes using poly(ethyleneglycol) diglycidyl ether to enhance membrane fouling resistance, J. Membr. Sci.367 (2011) 273–287.

[45] H. Ju, B.D. McCloskey, A.C. Sagle, Y.-H. Wu, V.A. Kusuma, B.D. Freeman, Cross-linked poly(ethylene oxide) fouling resistant coating materials for oil/water sep-aration, J. Membr. Sci. 307 (2008) 260–267.

[46] A. Gjaltema, N. van der Marel, M.C.M. van Loosdrecht, J.J. Heijnen, Adhesion andbiofilm development on suspended carriers in airlift reactors: hydrodynamic con-ditions versus surface characteristics, Biotechnol. Bioeng. 55 (1997) 880–889.

[47] R. Hausman, T. Gullinkala, I.C. Escobar, Development of low-biofoulingpolypropylene feedspacers for reverse osmosis, J. Appl. Polym. Sci. 114 (2009)3068–3073.

[48] S. Patil, F. Harnisch, U. Schröder, Toxicity response of electroactive microbial bio-films—a decisive feature for potential biosensor and power source applications,ChemPhysChem 11 (2010) 2834–2837.