Separation performance and interfacial properties of nanocomposite reverse osmosis membranes

Desalination 308 (2013) 180–185

Contents lists available at ScienceDirect

Desalination j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / d e s a l

Separation performance and interfacial properties of nanocomposite reverse osmosis membranes MaryTheresa M. Pendergast, Asim K. Ghosh 1, E.M.V. Hoek ⁎ Department Civil & Environmental Engineering and California NanoSystems Institute, University of California, Los Angeles, California, USA

a r t i c l e

i n f o

Article history: Received 8 February 2011 Received in revised form 3 May 2011 Accepted 3 May 2011 Available online 31 May 2011 Keywords: Nanocomposite Polyamide Polysulfone Reverse osmosis Water treatment Compaction

a b s t r a c t Four different types of nanocomposite reverse osmosis (RO) membranes were formed by interfacial polymerization of either polyamide (PA) or zeolite A-polyamide nanocomposite (ZA-PA) thin films over either pure polysulfone (PSf) or zeolite A-polysulfone nanocomposite (ZA-PSf) support membranes cast by wet phase inversion. All three nanocomposite membranes exhibited superior separation performance and interfacial properties relative to hand-cast TFC analogs including: (1) smoother, more hydrophilic surfaces (2) higher water permeability and salt rejection, and (3) improved resistance to physical compaction. Less compaction occurred for membranes with nanoparticles embedded in interfacially polymerized coating films, which adds further proof that flux decline associated with physical compaction is influenced by coating film properties in addition to support membrane properties. The new classes of nanocomposite membrane materials continue to offer promise of further improved RO membranes for use in desalination and advanced water purification. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Mixed matrix membranes—in which a filler material is embedded within a polymeric matrix—are already used in a variety of membrane processes including fuel cells, pervaporation, and gas separations [1–5]. This concept has added a new degree of freedom in the development of membranes with novel separation performance, i.e., selection of the unique properties of the filler material, which may include enhanced permeability, selectivity, stability, surface area, or catalytic activity. More recently, nanocomposite mixed matrix membranes have been explored to tailor performance and add novel functionality to membranes for water purification applications. A number of researchers have deposited nanoparticles onto the surface of reverse osmosis (RO) membranes or encapsulated them within RO membrane thin films, noting higher permeability (generally for both water and salts) and in some cases advanced functionality such as antimicrobial activity [6–10]. Jeong et al. [11] demonstrated that the incorporation of zeolite molecular-sieve nanoparticles into polyamide thin films (during interfacial polymerization) could nearly double the water flux without reducing observed rejection of sodium chloride, magnesium sulfate, and

⁎ Corresponding author at: University of California Los Angeles; Department of Civil & Environmental Engineering; 5732 Boelter Hall; PO Box 951593; Los Angeles, CA, 90095–1593. Tel.: + 1 310 206 3735; fax: + 1 310 206 2222. E-mail address: [email protected] (E.M.V. Hoek). 1 Current address: Desalination Division, Bhabha Atomic Research Centre, Trombay, Mumbai 400 085, India. 0011-9164/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.desal.2011.05.005

polyethylene glycol. A key feature of these ‘thin film nanocomposite’ (TFN) membranes is that zeolite nanoparticles are roughly the same diameter as polyamide film thickness, thereby creating a “percolation threshold” through the thin film with a single particle. It is thought that zeolite molecular-sieves with 3-D interconnected pores of ~4 Angstroms improved water permeability while maintaining good salt selectivity by offering preferential flow paths for water transport; however, zeolite-polyamide coating films containing impermeable zeolite nanoparticles (internal pores filled with a polymer) at the same zeolite loading produce fluxes intermediate between the relatively low flux pure polyamide thin film composite (TFC) membrane and the relatively high flux TFN membrane. This result offered an early insight that a mechanism besides molecular-sieving could be responsible for the enhanced water flux. Subsequently, Lind et al. [12] reported on the formation and characterization of pure polyamide TFC and zeolite-polyamide TFN membranes with seawater RO separation performance. Both TFC and TFN membranes were more permeable, hydrophilic, and rougher than a commercially available seawater RO membrane. Salt rejection by TFN membranes was consistently lower than that of hand-cast TFC membranes; however, two TFN membrane formulations exhibited better salt rejections than the commercial membrane (N99.4%). According to spectroscopic analysis, polyamide films formed in the presence of zeolite nanoparticles were less crosslinked than similarly cast pure polyamide films. At the low nanoparticle loadings evaluated, differences between pure polyamide and zeolite-polyamide membrane water and salt permeability correlated weakly with extent of crosslinking of the polyamide film, suggesting that a combination of

M.M. Pendergast et al. / Desalination 308 (2013) 180–185

molecular-sieving, crosslinking, and thin film defects likely governs transport through zeolite-polyamide TFN membranes. More recently, composite RO membranes were formed by interfacial polymerization of polyamide thin films over pure polysulfone and nanocomposite-polysulfone porous support membranes [13]. Nanocomposite supports were formed from four different sized amorphous non-porous silica nanoparticles, a zeolite A nanoparticle, and a surface-modified (hydrophobic) zeolite A nanoparticle. Nanocomposite-supported RO membranes generally had higher initial permeability and experienced less initial flux decline than pure polysulfone supported membranes. Cross-sectional SEM images verified that this decline was due to significant reduction in thickness of pure polysulfone supports, whereas minimal compaction was seen in nanocomposites due to enhanced mechanical stability imparted by the nanoparticles. A conceptual model was proposed to explain the mechanistic relationship between support membrane compaction and observed changes in water flux and salt rejection. As the support membrane compacts, surface (i.e., skin layer) pore constriction increases the effective path length of diffusion through polyamide thin films, thereby reducing both water and salt permeability. Nanocomposites that resist compaction also resist irreversible flux decline. This study uses nanocomposite membrane materials to elucidate and decouple the potential roles of coating film and support membrane in RO membrane compaction, and offers a potential route to improved RO membrane performance by eliminating flux decline due to physical compaction. Four different types of nanocomposite RO membranes were formed by interfacial polymerization of either polyamide (PA) or zeolite A-polyamide nanocomposite (ZA-PA) thin films over either pure polysulfone (PSf) or zeolite A-polysulfone nanocomposite (ZA-PSf) support membranes. The hand-cast membranes are described as (a) thin film composite (TFC = PA/PSf), (b) nanocomposite supported thin film composite (nTFC = PA/ZA-PSf), (c) thin film nanocomposite (TFN = ZA-PA/PSf), and (d) nanocomposite supported thin film nanocomposite (nTFN = ZA-PA/ZA-PSf). Membrane interfacial properties, separation performance, and compaction resistance are presented and discussed. 2. Materials and methods

181

MPD solution contained MPD:TEA:CSA:SLS:IPA in ratios of 2.0:2.0:4.0:0.02:10 w/v%. Excess MPD solution was removed from the support membrane surfaces using laboratory gas forced through a custom fabricated air knife. Aqueous MPD saturated support membranes were then immersed into 0.1 wt% trimesoyl chloride (1,3,5-tricarbonyl chloride, Sigma-Aldrich, Milwaukee, Wisconsin, USA) in a proprietary isoparrafin (ExxonMobil Isopar G, Gallade Chemicals, Inc., Santa Ana, California, USA) at 30 °C for 15 seconds to initiate polymerization. The resulting composite membranes were heat cured at 82 °C for 10 minutes, washed thoroughly with deionized water, and stored at 5 °C in opaque containers of deionized water. Both TFN and nTFN membranes were formed by dispersing 0.2 w/w% zeolite-A nanoparticles in the TMC solution as previously described [11,12].

2.2. Membrane characterization The surface (zeta) potentials of hand-cast UF and RO membranes were determined using the Helmholtz-Smoluchowski equation and measured streaming potentials in 10 mM NaCl solution at an unadjusted pH of 5.8 ± 0.2 (Zeta PALS, Brookhaven Instrument Corp., New York, New York, USA). Sessile drop contact angles of deionized water were measured on membrane samples in an environmental chamber mounted to a contact angle goniometer (DSA10, KRÜSS GmbH, Hamburg, Germany). The equilibrium values were taken as steady-state averages of left and right angles. Surface roughness of synthesized membranes was measured via AFM (Nanoscope IIIa, Digital Instruments, Santa Barbara, California, USA). Mechanical strength of polysulfone and nanocomposite membranes was characterized by the ultimate tensile strength (σ) measured with a mechanical testing instrument (Instron® 5540 Series Single Column Testing Systems, Instron, Norwood, Massachussetts, USA). In this test, a membrane specimen (4 cm× 1.5 cm) was stretched at a constant rate (0.5 mm/min) until breakage. The ultimate tensile strength is calculated from the maximum load applied at breakage divided by the original cross-sectional area of the test piece. The cross-sectional area was determined by micro-caliper measurements of film thickness and the sample width. The non-woven polyester fabric contributed an ultimate strength of 16.2 MPa.

2.1. Membrane preparation Support membranes were prepared by dissolving 18 g of polysulfone (PSf) beads (Mn-26,000 from Aldrich, St. Louis, Missouri, USA) into 72 mL of N-methyl pyrrolidone (NMP) (Acros Organics, Morris Plains, New Jersey, USA) in airtight glass bottles. For nanocomposite membranes, 3.6 g of 250 nm Linde Type-A zeolite nanoparticles (NanoH2O Inc., Los Angeles, California, USA) were dispersed in NMP prior to PSf addition. The suspensions were then agitated for several hours until complete polymer dissolution was achieved to form the casting solutions. Casting solutions were spread via knife-edge over a polyester non-woven fabric (NanoH2O Inc., Los Angeles, California, USA), previously adhered to a glass plate. The glass plate was immediately immersed into a coagulation bath of 18 MΩ laboratory deionized water maintained at room temperature to induce polymer precipitation. After 30 minutes the non-woven fabric supported polysulfone and nanocomposite membranes were removed from the water bath, separated from the glass plate, washed thoroughly with deionized water, and stored in deionized water in a laboratory refrigerator at 5 °C to inhibit biogrowth. Composite RO membranes were then formed via interfacial polymerization atop the phase inverted structures. Pure polysulfone support membranes were used for TFC and TFN membranes and nanocomposite support membranes were used for nTFC and nTFN membranes. Support membranes were immersed in an aqueous solution of m-phenylenediamine (MPD) (1,3-diaminobenzene, Sigma-Aldrich, Milwaukee, Wisconsin, USA), triethyl amine (TEA), (+)-10-champhor sulfonic acid (CSA), sodium lauryl sulfate (SLS), and isopropanol (IPA) for 15 seconds. The

2.3. Membrane performance experiments Polysulfone and nanocomposite support membrane performances were evaluated using a high-pressure chemical resistant dead-end filtration cell (HP4750 Stirred Cell, Sterlitech Corp., Kent, Washington, USA) at room temperature (~20 °C) under applied pressure of 140 kPa (20 psi). Pure water permeability was determinedby filtering deionized water through the support membranes. The molecular weight cutoff (MWCO) of these membranes was determined from the separation of a series of dextran solutions. Dextran concentrations in the feed and permeate were analyzed with a total organic carbon analyzer (Apollo 9000, Tekmar Dohrmann, Cincinnati, Ohio, USA). MWCO is defined as the molecular weight dextran molecule that was rejected 90% or more. The separation performances of synthesized TFC and TFN membranes were evaluated in terms of water flux and salt rejection. Membranes were washed thoroughly for 45 min under applied pressure of 1550 kPa (225 psi). The permeability was monitored during initial stages of compaction to gain an understanding of the mechanical stability of each. Pure water flux was then measured at room temperature by normalizing the volume of pure water collectedin 30 min increments by the active membrane area (13.85 cm2). For all thin film membranes formed, rejection tests were applied by challenging the membranes with a 10 mM NaCl solution and repeating the permeation test.

182

M.M. Pendergast et al. / Desalination 308 (2013) 180–185

(a)

(b)

30

35

25

Roughness, SAD (%)

Roughness, RMS (nm)

30 25 20 15 10

15 10

5

5

0

0

(d)

(c)

Ultimate Tensile Strength (MPa)

80

Water Contact Angle (°°)

20

75

70

65

60

PSf

ZA-PSf

50 40 30 20 10 0

PSf

ZA-PSf

Fig. 1. Properties of UF membranes with zeolite additions to polysulfone support structure (i.e., ZA-PSf) compared to pure polysulfone membrane. Comparison of (a) root mean squared (RMS) roughness, (b) surface area difference (SAD) roughness, (c) deionized water contact angle, and (d) ultimate tensile strength.

3. Results and discussion

(2) zeolites contribute to the permeability of support membranes via their molecular sieving capabilities.

3.1. Polymer and nanocomposite support membrane properties 3.2. Polymer and nanocomposite RO membrane properties All membranes containing zeolite nanoparticles in the support membrane or the coating film (e.g., nTFC, TFN, and nTFN) were more negatively charged, hydrophilic, smooth, permeable, and selective than the baseline TFC membrane (Fig. 3). Surface characteristics were

A

MWCO

200

1000

160

800 120 600 80 400 40

200 0

MWCO (kDa)

1200

A (µ µm.s-1.MPa-1)

Support membranes were characterized prior to interfacial polymerization of coating films to better understand their physical-chemical properties independent of composite membranes. The ZA-PSf nanocomposite support structure was rougher, more hydrophilic, and more mechanically robust (Fig. 1). The roughness as indicated by both root mean squared (RMS) and surface area difference (SAD) measurements doubled in the nanocomposite membranes, likely due to the presence of zeolite particles at or on the membrane surface. More importantly, the ultimate tensile strength of the nanocomposite membranes was almost double that of the pure polysulfone membranes, implying more stable structure for high-pressure separations. Pure water permeability and MWCO were extracted from water flux and dextran rejection experiments. The ZA-PSf nanocomposite membrane was more permeable and more selective than the PSf counterpart (Fig. 2). In fact, the nanocomposite membrane had a water permeability coefficient 45% higher and had a MWCO 40 kDa lower than that of the pure PSf membrane. Enhanced permeability of the nanocomposites is likely due to a combination of enhanced porosity and defect formation in the polysulfone matrix; however, because the zeolite A nanoparticles act as molecular sieves, allowing the passage of water and excluding large solute particles, the selectivity of these membranes is actually enhanced. This is a unique feature of zeolite nanocomposites since most polymer-nanocomposite phase inversion membranes exhibit either (1) higher flux with lower rejection or (2) lower flux with lower rejection [14]. The higher surface roughness and hydrophilicity combined with the higher flux, but lower MWCO suggest that (1) zeolite nanoparticles are present at the interface of ZA-PSf membranes and

PSf

ZA-PSf

0

Fig. 2. Performance characteristics of UF membranes in terms of water permeability coefficient, A, and molecular weight cutoff, MWCO, of UF membrane with zeolite-A additions (i.e., ZA-PSf) compared to the pure polysulfone membrane.

M.M. Pendergast et al. / Desalination 308 (2013) 180–185

(a)

(b) Water Contact Angle (°°)

Roughness, RMS (nm)

100

80 60

40 20

183

75

70

65

60

55

0

50

(c)

(d)

120

0

100

Zeta Potential (mV)

Roughness, SAD (%)

-2

80 60 40

-4 -6 -8 -10 -12 -14

20 -16 0

TFC

nTFC

TFN

nTFN

-18

TFC

nTFC

TFN

nTFN

Fig. 3. Interfacial properties of RO membranes with zeolite additions to polysulfone support structure (i.e., nTFC and nTFN) and/or polyamide thin film (i.e., TFN and nTFN) compared to pure polyamide coated polysulfone TFC membrane. Comparison of (a) root mean squared (RMS) roughness, (b) surface area difference (SAD) roughness, (c) deionized water contact angle, and (d) surface zeta potential.

salt permeability was comparable to the baseline TFC membrane. The nTFN membrane had the high water permeability of the nTFC membrane plus a good salt permeability—close to the TFC and TFN value.

1.0 0.9

Normalized Water Flux, J/Jo

most notably impacted in thin film nanocomposite membranes (i.e., TFN and nTFN) due to the presence of zeolites in the coating film. The RMS roughness—an indication of the mean roughness height— was significantly lower in the nanocomposite membranes; however, the SAD roughness—an indication of the frequency of rough features— for TFN and nTFN membranes increased due to the presence of nanoparticles in the coating films. The propensity of a RO membrane to lose permeability due to physical compaction was evaluated by examining the decline of water flux during pressurization with 18 MΩ deionized water (presumed foulant free). Salt rejections reported for each membrane correspond to the post-compaction plateau region on the normalized flux curves in Fig. 4. The TFC membrane shows the most dramatic reduction (50%) in water permeability during the initial stages of pressurization. All three nanocomposite membranes (i.e., nTFC, TFN, and nTFN) maintained more stable fluxes, implying that these membranes were more mechanically robust and experienced less physical compaction. TFN membranes lost only a very small amount of water flux, whether supported by polysulfone or nanocomposite-polysulfone supports. These results imply that coating film mechanical properties were just as important, if not more important than support membrane mechanical properties for producing RO membranes that resist irreversible flux decline due to physical compaction. Apparent water and salt permeability coefficients were extracted from the post-compaction data for the four composite RO membranes (Fig. 5). The water permeability coefficient increased 6-fold for the nanocomposite supported (i.e., nTFC and nTFN) membranes and increased 3-fold for the TFN membrane. While the nTFC membrane showed a large increase in water permeability, it also had the highest increase in salt permeability. The TFN membrane water permeability increased, while the

0.8 0.7 0.6 0.5 TFC, R=85% nTFC, R=93% TFN, R=94% nTFN, R=92%

0.4 0.3 0

50

100

150

200

250

300

Time (min) Fig. 4. Transient flux normalized by initial flux for each membrane type, when challenged with 10 mM NaCl feed solution at applied pressure of 1550 kPa, indicating performance losses for mechanically weaker materials. Rejections listed are the steady values reached at the end of each curve.

M.M. Pendergast et al. / Desalination 308 (2013) 180–185

A

A (µm.MPa-1.s-1)

Percent Change in Salt Permeability

1.2

B

12

-100

1.0

10

0.8

8 0.6 6 0.4

4 2

0.2

0

0.0

-80

-60

-40

-20

B (µ µm.s-1)

14

0

0

-20

-40 TFC

nTFC

TFN

nTFN

Fig. 5. Performance characteristics of RO membranes indicated by post-compaction water permeability, A, and salt permeability, B, coefficients of TFC versus nTFC, TFN, and nTFN membranes containing zeolite-A nanoparticles.

The support membrane structure appears to have played an important role in maintaining high water permeability, while the coating film appeared critical for maintaining high salt rejection. This supports our previous model that states that a higher density of surface pores in the support structure—whether due to defect formation or molecular sieving—will lead to higher passage of both water and salt in the overall membrane. On the other hand, nanocomposite thin films maintained high flux and salt rejection via molecular sieving, and therefore, helped maintain salt rejection even as water flux is enhanced. These results point towards a wide range of opportunities for tailoring RO membrane performance by nanostructuring both the support membrane and coating film materials. Based on these results nanocomposite supports appeared optimal for enhancing intrinsic RO membrane permeability, while nanocomposite coating films were most effective at resisting physical compaction under pressure. 4. Discussion and conclusions Nanostructured composite membranes containing zeolite nanoparticles (i.e., nTFC, TFN, and nTFN) exhibited superior interfacial properties and separation performance to the TFC analog, including (1) smoother, more hydrophilic surfaces, (2) higher water permeability and observed salt rejection, and (3) improved resistance to irreversible flux decline due to physical compaction. Nanocomposite supports, as in nTFC and nTFN membranes, led to the best overall performance with a six-fold increase in pure water permeability over the baseline TFC membrane. Nanocomposite thin films, as in TFN and nTFN membranes, contributed improvements to surface properties and compaction resistance. Our previously proposed compaction model attributes changes in membrane transport properties to an increased path length for diffusion by collapse of support membrane skin layer pores [13]. Implicit in this model is that water and salt permeability must change to the same degree by compaction. However, when the relative changes in water and salt permeability coefficients (calculated as the change in A or B from start to finish in a compaction experiment normalized by the starting value of A or B) are plotted against each other for the hand-cast membranes in this study and our previous compaction study [13] it becomes clear that this is not the case (Fig. 6). Almost all data points lie above the unity line; hence, compaction reduces salt permeability to a much larger extent than water permeability. The one outlier is the LTA-filled nanocomposite supported TFC membrane compacted at low pressure (1,724 kPa) in our previous study [13]. This point falls out of the trend because the salt permeability did not change at all after low pressure compaction. It is not known if this data is anomalous due to experimental error or an unknown mechanism, but—almost universally—salt permeability increases more than water permeability. Therefore, collapse of support membrane skin layer pores must not be the only mechanism that manifests via physical compaction of RO membranes.

-60

TFC nTFC TFN nTFN

-80

Percent Change in Water Permeability

184

-100 Fig. 6. Percent changes in pure water permeability coefficient plotted against percent changes in salt permeability coefficient for hand-cast RO membranes. Dotted line of slope one represents an equivalent degree of change in both quantities. Graph shows that the two quantities do not change at an equal rate, but rather, salt permeability coefficient is impacted to a much greater extent by compaction. Data included from the current study and our previous compaction study [13].

The importance of coating film structure may be partially explained by free volume theory, which relates the flux, Ji, of a species i through a dense polymer film to the fractional free volume, FFV, in the film according to   C2;i ; Ji = C1;i ⋅exp − FFV

ð1Þ

where the pre-exponential constant, C1,i, is related to the size and structure of the species i and the exponential constant, C2,i, is related to the amount of open or free volume that must be locally present for the species to diffuse. It is conceivable that the constant value of C2, salt will be larger than that of C2,water since salt molecules are much larger. In fact, a recent study in which the coating film free volume was systematically controlled during film formation confirms that salt permeability is more strongly governed polymer free volume than is water permeability [15]. As composite RO membranes are compacted, both the support structure and the coating film may be physically compressed. In the case of the coating film, the decrease in free volume appears to contribute most towards changes in composite membrane water and salt permeability. Conceptual illustrations were created to depict the proposed physical changes that occur in each type of RO membrane during physical compaction (Fig. 7). Nanocomposite support membranes undergo far less physical compaction than pure polysulfone supports. All coating films experience some densification. However, coating films containing zeolites maintain relatively uncompacted and selective permeation channels via the molecular sieving action of nanoparticles as well as by minimizing the compaction of the surrounding polyamide film. Both nanocomposite support membranes and coating films appear to combat physical compaction and improve composite RO membrane properties, but nanocomposite coating films are particularly effective because they reduce coating film densification. For this reason, it seems that controlling the thin film structure may be the foremost priority for designing better and better performing RO membranes. While the mechanisms responsible for the enhanced flux of nanocomposite membranes are only beginning to be elucidated, nanocomposite materials continue to promise

M.M. Pendergast et al. / Desalination 308 (2013) 180–185

185

Fig. 7. Schematic drawings showing the proposed physical changes to support and thin film structure during compaction in (a) TFC, (b) nTFC, (c) TFN, and (d) nTFN membranes.

further improvements to RO membrane productivity, selectivity, stability, and fouling resistance. Disclosure Authors EMVH and AKG have a financial interest in one of the project co-sponsors, NanoH2O Inc., through stock ownership and consulting agreements. Acknowledgements This publication is based on work supported in part by Award No. KUS-C1-018-02, made by King Abdullah University of Science and Technology (KAUST), in addition to the UCLA California NanoSystems Institute (CNSI) and NanoH2O Inc. Additional financial support for MTMP was provided by the National Science Foundation Graduate Research Fellowship through Grant No. DGE-0707424. The authors wish to express their appreciation to Prof. Ajit Mal and Shri Harsh K. Vaid in the Department of Mechanical & Aerospace Engineering at UCLA for providing access to the Instron® mechanical testing instrument. References [1] P.S. Tin, T.S. Chung, L.Y. Jiang, S. Kulprathipanja, Carbon-zeolite composite membranes for gas separation, Carbon 43 (2005) 2025–2027. [2] C.M. Zimmerman, A. Singh, W.J. Koros, Tailoring mixed matrix composite membranes for gas separations, J. Membr. Sci. 137 (1997) 145–154.

[3] E. Okumus, T. Gurkan, L. Yilmaz, Development of a mixed-matrix membrane for pervaporation, Sep. Sci. Technol. 29 (1994) 2451–2473. [4] D.Q. Vu, W.J. Koros, S.J. Miller, Mixed matrix membranes using carbon molecular sieve. I. Preparation and experimental results, J. Membr. Sci. 211 (2003) 311–334. [5] H.Y. Chang, C.W. Lin, Proton conducting membranes based on PEG/SiO2 nanocomposites for direct methanol fuel cells, J. Membr. Sci. 218 (2003) 295–306. [6] H.S. Lee, S.J. Im, J.H. Kim, H.J. Kim, J.P. Kim, B.R. Min, Polyamide thin-film nanofiltration membranes containing TiO2 nanoparticles, Desalination 219 (2008) 48–56. [7] S.Y. Kwak, S.H. Kim, S.S. Kim, Hybrid organic/inorganic reverse osmosis (RO) membrane for bactericidal anti-fouling. 1. Preparation and characterization of TiO2 nanoparticle self-assembled aromatic polyamide thin-film-composite (TFC) membrane, Environ. Sci. Technol. 35 (2001) 2388–2394. [8] S. Lee, H. Kim, R. Patel, S. Im, J. Kim, B. Min, Silver nanoparticles immobilized on thin film composite polyamide membrane: Characterization, nanofiltration, antifouling properties, Polym. Adv. Technol. 18 (2007) 562–568. [9] P. Singh, V. Aswal, Characterization of physical structure of silica nanoparticles encapsulated in polymeric structure of polyamide films, J. Colloid Interface Sci. 326 (2008) 176–185. [10] G. Jadav, P. Singh, Synthesis of novel silica-polyamide nanocomposite membrane with enhanced properties, J. Membr. Sci. 328 (2009) 257–267. [11] B.H. Jeong, E.M.V. Hoek, Y.S. Yan, A. Subramani, X.F. Huang, G. Hurwitz, A.K. Ghosh, A. Jawor, Interfacial polymerization of thin film nanocomposites: a new concept for reverse osmosis membranes, J. Membr. Sci. 294 (2007) 1–7. [12] M.L. Lind, D. Eumine Suk, T.-V. Nguyen, E.M.V. Hoek, Tailoring the Structure of Thin Film Nanocomposite Membranes to Achieve Seawater RO Membrane Performance, Environ. Sci. Technol. 44 (2010) (2010) 8230–8235. [13] J.M. Nygaard, M.T.M. Pendergast, A.K. Ghosh, E.M.V. Hoek, Using nanocomposite membrane materials to understand and control reverse osmosis membrane compaction, Desalination 261 (2010) 255–263. [14] E.M.V. Hoek, A.K. Ghosh, X. Huang, M. Liong, J.I. Zink, Physical-chemical properties and separation performance of inorganic-organic mixed-matrix ultrafiltration membranes, Desalination (in press), doi:10.1016/j.desal.2011.04.008. [15] F. Peng, Z. Jiang, E.M.V. Hoek, Tuning the molecular structure, separation performance and interfacial properties of poly(vinyl alcohol)-polysulfone interfacial composite membranes, J. Membr. Sci. 368 (2011) 26–33.