Journal of Membrane Science

Transcription

1 Journal of Membrane Science 437 (2013) Contents lists available at SciVerse ScienceDirect Journal of Membrane Science journal homepage: Novel hydrophilic nylon 6,6 microfiltration membrane supported thin film composite membranes for engineered osmosis Liwei Huang a, Nhu-Ngoc Bui a, Mark T. Meyering b, Thomas J. Hamlin b, Jeffrey R. McCutcheon a,n a Department of Chemical and Biomolecular Engineering, Center for Environmental Sciences and Engineering, University of Connecticut, Storrs, CT , USA b 3M Purification Inc., Meriden, CT 06450, USA article info Article history: Received 5 October 2012 Received in revised form 15 December 2012 Accepted 13 January 2013 Available online 4 February 2013 Keywords: Nylon Forward osmosis Pressure-retarded osmosis Microfiltration Internal concentration polarization abstract Previous investigations of engineered osmosis (EO) concluded that hydrophobic support layers of thin film composite membrane causes severe internal concentration polarization due to incomplete wetting. Incomplete wetting reduces the effective porosity of the support, inhibiting mass transport and thus water flux. In this study, novel thin film composite membranes were developed which consist of a poly(piperazinamide) or polyamide selective layer formed by interfacial polymerization on top of a nylon 6,6 microfiltration membrane support. This intrinsically hydrophilic support was used to increase the wetted porosity and to mitigate internal concentration polarization. Reverse osmosis tests showed that the permselectivity of our best poly(piperazinamide) and polyamide thin film composite membranes approached those of a commercial nanofiltration and a commercial reverse osmosis membrane, respectively. The osmotic flux performance of the new polyamide thin film composite membrane showed matched water flux, 10 fold lower salt flux and 8 28 fold lower specific salt flux than the standard commercial cellulose triacetate forward osmosis membrane from Hydration Technology Innovations TM. The relatively good performance in osmotic flux tests of our thin film composite membranes was directly related to the high permselectivity of the selective layers coupled with the hydrophilicity of the nylon 6,6 support. These results suggest that these nylon 6,6 supported thin film composite membranes may enable applications like forward osmosis or pressure retarded osmosis. & 2013 Elsevier B.V. All rights reserved. 1. Introduction Clean water and energy are essential for public health and economic prosperity. Engineered Osmosis (EO) is an emerging platform technology that has the potential to sustainably produce clean drinking water and electric power. It has therefore garnered great interest amongst the membrane science community within the past half-decade [1 3]. Unlike hydraulically driven membrane processes, EO exploits the natural phenomenon of osmosis, which occurs when two solutions of differing concentration are placed on two sides of a semi-permeable membrane. The generated osmotic pressure difference drives the permeation of water across the membrane from the dilute solution to the concentrated solution. EO has applications in direct osmotic concentration (DOC) for concentrating high-value solutes, forward osmosis (FO) for seawater desalination and pressure retarded osmosis (PRO) for electric power generation [4 6]. n Corresponding author. Tel.: þ address: jeff@engr.uconn.edu (J.R. McCutcheon). Despite the potential to address water and energy scarcity, EO processes have not yet become commercialized on a large scale. One major obstacle to commercialization is the lack of an appropriately designed membrane. Early work concluded that thin film composite (TFC) membranes specially designed for nanofiltration (NF) and reverse osmosis (RO) membranes, the best salt rejecting membranes commercially available, were unsuitable for osmotic separations due to their poor flux performance [4 6]. A typical RO membrane is comprised of an aromatic polyamide thin film formed in-situ on top of an asymmetric polysulfone (PSu) mid-layer casted by phase inversion over a polyester (PET) nonwoven backing layer [7 9]. TFC-NF or TFC RO membranes fail to perform well in EO processes because the thick support layers that, while necessary to withstand large hydraulic pressures in NF or RO, create resistance to solute mass transfer in FO or PRO. This mass transfer resistance is known as internal concentration polarization (ICP) and occurs within the thick support layers. ICP adversely affects the performance by reducing the osmotic pressure difference across the TFC membranes [2,4,6,10 13]. ICP may even be enhanced by the hydrophobic nature of typical TFC support layers. The intrinsic hydrophobicity of the PSu and PET layers prevent complete wetting of the pore structure. The unwetted areas of the support layer are not available for solute and /$ - see front matter & 2013 Elsevier B.V. All rights reserved.

2 142 L. Huang et al. / Journal of Membrane Science 437 (2013) water transport [14]. Existing commercial FO membranes from Hydration Technology Innovations TM (HTI) are specifically designed for EO processes with minimized ICP, which is achieved by eliminating the thick backing layer and using hydrophilic cellulose triacetate (CTA) to make the membranes [15]. However, due to its integrated membrane structure and chemistry, these commercial FO membranes have relatively low permeability and salt rejection. Additionally, CTA is also susceptible to hydrolysis under basic conditions and high temperature [8]. A membrane designed for an osmotically driven process should have a combination of characteristics. It must have a permselectivity to reject most solutes and produce high water fluxes. This is common in today s RO membranes which have a highly selective polyamide layer. For applications in EO, however, this selective layer must also be supported by a thin, highly porous, non-tortuous and hydrophilic support layer to minimize effective ICP [16,17]. For EO applications involving hydraulic pressure (like PRO) the membranes must be able to tolerate the mechanical stresses during operation, though all membranes will need to have a minimum strength requirement for fabrication and installation into elements and modules. One option is to design a TFC membrane with an RO-type selective layer supported by a hydrophillic porous support. Such a membrane would retain the high permselectivity of a commercial RO membrane and exhibit low resistance to mass transfer during osmotic flow due to improved hydrophilicity of the support. However, research using hydrophilic polymers in TFC membrane supports is still in its infancy at the time of this writing. Recent efforts on designing TFC flat sheet and hollow fiber membranes for EO [17 19] focus on support structures but not support chemistry. Hydrophobic PSu and polyethersulfone (PES) are used in these studies. Several groups are also considering the fabrication of integrated asymmetric EO membranes using hydrophilic polymers such as polybenzimidazole and cellulose acetate (CA) [20 22]. However, hydrophilic polymers are not conventionally used to support TFC membranes, even in RO applications. This is due to differences in the fabrication process and the likelihood of plasticization in aqueous environments. An issue that needs to be considered in making hydrophilic polymer supported TFC membrane is support swelling [23]. Water molecules might plasticize the hydrophilic support more severely than the rigid polyamide selective layer, causing postfabrication perforation or delamination of the selective layer. Additionally, the mechanical stability of hydrophilic supports may be impacted by a swelling deswelling equilibrium in the presence of high concentration of salt ions [24]. Choosing a suitable hydrophilic support with low swelling propensity is essential but also challenging. In this work, a nylon 6,6 microfiltration membrane from 3M TM was used as the support for TFC membranes for the first time. Nylon 6,6, a common polymer for textile and plastic industry has good mechanical, thermal and chemical properties due to its semi-crystalline structure. Additionally, it is much more hydrophilic than conventional PSu support but has less swelling propensity than other common hydrophilic polymers, such as cellulose acetate [25]. We selected both poly(piperazinamide) and polyamide as the selective layer in order to suit different types of draw solutes. Poly(piperazinamide) is usually used to fabricate NF membranes, which has high rejection to divalent salts, such as MgSO 4,with relatively high permeate flux [26]. It may therefore be suitable for osmotic processes using draw solutions composed of divalent salts. Polyamide, on the other hand, is usually used to fabricate RO membranes, which has high rejection to monovalent salts like NaCl [26] and might be applied in EO with monovalent salt draw solutions. This investigation demonstrates the performance of both classes of TFC membranes supported by a commercial hydrophilic nylon 6,6 microfiltration (MF). Salt rejection and pure water permeability of the newly fabricated TFC poly(piperazinamide) and polyamide membranes are comparable to commercially available nanofiltration (NF) and reverse osmosis (RO) membranes, respectively. The osmotic flux performance of polyamide TFC membrane is also evaluated and is shown to meet or exceed the performance of the standard commercial FO membranes. We hypothesized that hydrophilic nylon 6,6 that supports the membrane selective layer should enhance osmotic flux by minimizing ICP due to improved wetting. 2. Experimental 2.1. Materials A multi-zoned nylon 6,6 MF membrane designated BLA010 was provided by 3M Purification Inc. (Meriden, CT) as the support for TFC EO membranes. This MF membrane has three regions: (1) a large pore region at the upstream side of the membrane that usually faces the feed in MF as a pre-filter by capturing larger particles; (2) a nonwoven scrim used as a mechanical support and to facilitate manufacturing; and (3) a small-pore region on the downstream side of the membrane that faces the permeate in MF to provide the retention of small contaminents. The mean pore sizes of the small-pore region and large-pore region are 0.1 mm and 0.45 mm, repectively, according to the manufacturer. Diamine monomers piperazine (PIP) and m-phenylenediamine (MPD) were purchased from Acros Organic and Sigma-Aldrich, respectively. Acid chloride monomer trimesoyl chloride (TMC) and acid acceptor triethylamine (TEA) were purchased from Sigma-Aldrich. Hexane, the solvent for TMC, was purchased from Fisher Scientific. Deionized water (DI) obtained from a Milli-Q ultrapure water purification system (Millipore, Billerica, MA) was used as the solvent for diamine monomers. Sodium chloride and magnesium sulfate were purchased from Fisher Scientific. Commercial asymmetric cellulose triacetate (HTI-CTA) FO membrane (Hydration Technology Innovations Inc., Albany, OR), TFC NF membrane NF270 and TFC seawater RO membrane SW30- XLE (Dow Water & Process Solutions Company, Midland, MI) were acquired for comparison Interfacial polymerization of TFC membrane The monomers used for in situ interfacial polymerization of poly(piperazinamide) were PIP and TMC. PIP was dissolved in Milli-Q water at varying concentrations ranged from 0.25% to 3% (w/v). An equivalent amount of TEA was added into aqueous solution as an acid acceptor [27]. PIP has been reported to react slowly to form the poly(piperazinamide) layer because of its inefficient partitioning capacity into the organic phase and tendency to react with hydrochloride. Addition of acid acceptor drives the reaction toward the formation of poly(piperazineamide) [28]. A 0.15% (w/v) solution of TMC in hexane was prepared. All solutions were stirred at room temperature for at least 3 h prior to using. A nylon 6,6 microfiltration membrane was first taped onto a glass plate with the smaller pore side facing up. The support was then immersed into an aqueous PIP solution for 120 s. Excess PIP solution was removed from the support membrane surface using a rubber roller. The membrane was then dipped into a TMC/hexane solution for 60 s to form an ultrathin poly(piperazineamide) film. The resulting composite film was air dried for 120 s and subsequently cured in an air-circulation oven at 80 1C for 5 min for

3 L. Huang et al. / Journal of Membrane Science 437 (2013) attaining the desired stability of the formed structure [28]. The TFC poly(piperazinamide) membrane was thoroughly washed and stored in deionized water at 4 1C before carrying out studies. The procedure for interfacial polymerization of polyamide was similar to that of poly(piperazinamide), except the aqueous phase is pure MPD dissolved in Milli-Q DI. The MPD concentrations ranged from 0.25% to 3% (w/v) Membrane characterization Surface morphology of the BLA010 MF support and the TFC polyamide membranes were qualitatively evaluated with scanning electron microscopy (SEM) using a cold cathode field emission scanning electron microscope JSM-6335F (FEI Company, USA). Before imaging, samples were kept overnight in a desiccator and then sputter coated with a thin layer of platinum to obtain better contrast and to avoid charge accumulation. Cross-sectional structure of the BLA010 support and the selective layers of TFC membranes were also imaged with SEM. These samples were prepared for imaging using a freeze fracture technique involving liquid nitrogen. Due to the difficulty to freeze fracture the reinforced nonwoven scrim, a razor blade was also submerged into liquid nitrogen with the sample strip simultaneously and then used to quickly cut the sample into half once removed from the liquid nitrogen. The prepared samples were sputter coated with a thin layer of gold before imaging. The thickness of the support was measured using a digital micrometer at 5 different locations for each membrane sample. A CAM 101 series contact angle goniometer was used to measure the contact angle of the support Determination of pure water permeability, salt rejection, and solute permeability in cross-flow reverse osmosis Pure water permeability (A) for poly(piperazinamide) TFC membranes made at different PIP concentrations were evaluated by reverse osmosis using DI as a feed. Pure water permeability tests were conducted at four pressures ranging from 50 to 250 psi, at a cross-flow velocity of 0.26 m/s at 25 1C. The salt rejection tests for poly(piperazinamide) TFC membranes were conducted in using a 2000 ppm MgSO 4 feed. Observed salt rejection rate (%R) was characterized by measuring the conductivity of the bulk permeate and the feed. The Dow Water & Process Solutions TM NF270 membrane was used as a control. Pure water permeability and salt rejection tests for polyamide TFC membranes were carried out using DI and 2000 ppm NaCl feed at four pressures ranging from 150 to 300 psi, respectively. The Dow SW30-XLE was used as a control. Other testing conditions are kept the same as that for poly(piperazinamide). The pure water permeability and salt rejection rate of our optimum polyamide TFC membranes was also compared with commercial FO membranes at 20 1C, at which the osmotic flux tests were conducted. The solute permeability coefficient, B, was also determined to calculate structural parameter, S. Pure water permeability, A, was determined by dividing the pure water flux (J w ) by the applied pressure (DP), A¼J w /DP. Salt rejection rate, %R, was determined from the difference in bulk feed (c f ) and permeate (c p ) salt concentrations measured using a conductivity meter. The solute permeability coefficient, B, wasdetermined fromeq.(1) [9]: B ¼ J w 1 R R exp J w k where k, the cross-flow cell mass transfer coefficient, is calculated from correlations for this geometry [29]. ð1þ 2.5. Evaluation of osmotic water flux and reverse salt flux Osmotic water flux and reverse salt flux of polyamide TFC membranes were evaluated using a custom lab-scale cross-flow forward osmosis system. The experimental setup was described in details elsewhere [5,13]. A 1.5 M sodium chloride solution was used as the draw solution while DI water was used as the feed solution. Osmotic flux tests were carried out with the membrane oriented in both PRO mode (the membrane active layer faces the draw solution) and FO mode (the membrane active layer faces the feed solution). The hydraulic pressures of the feed and draw solutions were the same (1.5 psi) and the cross-flow velocities were kept at 0.18 m/s for both the feed and draw solutions. The temperatures of the feed and draw solutions were maintained at C using a recirculation water bath and a heat exchanger. Conductivity of the feed was measured to estimate the reverse salt flux through the membrane. The osmotic water flux, J w, was calculated by dividing the volumetric flux by the membrane area. By measuring the conductivity of the feed solutions at certain time points during the tests, the reverse salt flux, J s, was calculated by dividing the NaCl mass flow rate by the membrane area Determination of specific salt flux and structural parameter The specific salt flux [30,31], J s /J w, was determined as the ratio of the reverse salt flux and the water flux. The structural parameter was determined by using Eq. (2) [17] S ¼ D In BþAp D,b ð2þ J w BþJ w þap F,m where D is the diffusion coefficient of the draw solute, J w is the measured water flux, B is the solute permeability, A is the pure water permeability, p D, b is the bulk osmotic pressure of the draw solution, and p F, m is the osmotic pressure at the membrane surface on the feed side (0 atm for DI feed). 3. Results and discussion 3.1. Characterization of nylon 6,6MF support The surface and cross-sectional SEM images of nylon 6,6MF support are shown in Fig. 1. Both the surface of small-pore region and large-pore region show rough and open porous morphology. The surface porosities of two zones are 51.1% and 48.4%, respectively. The average thickness of the support was measured to be mm and the contact angle of the support is approximately 40.51, which is lower than that of conventional TFC PSu support reported in other studies [32,33]. While these membranes are typically oriented with the larger pores facing the feed during MF, the selective layers of the TFC membranes are built on the small pore size. The small pores region allows for a selective layer to form with fewer defects while the large pores zone decreases the resistance to mass transfer. This finding is consistent with previous reports on TFC membranes; membranes with a dense support layer resulted in high salt rejections yet low water fluxes [33,34] Characterization and performance of poly(piperazinamide) TFC membranes Top surface scanning electron micrographs The top surface SEM images for poly(piperazinamide) selective layers made at different PIP concentrations are shown in Fig. 2. Defect free films were obtained for TFC membranes made at all

4 144 L. Huang et al. / Journal of Membrane Science 437 (2013) Fig. 1. SEM images of (a) cross-section, (b) surface of small-pore region on which the selective layers were formed (magnification 5000 ), and (c) surface of large-pore region of BLA010 nylon 6,6 support (magnification 5000 ). Fig. 2. Top surface SEM images (magnification 2200 ) of poly(piperazinamide) of TFC membranes made at PIP concentrations of (a) 0.25% (w/v); (b) 0.5%; (c) 1.0%; (d) 2.0%; and (e) 3.0%; and of (f) commercial NF270. PIP concentrations ranging from 0.25 to 3%. The surface morphology varied with PIP concentration. Circle-like morphologies can be observed on 0.25% and 0.5% PIP based-tfc membranes, while a more uniform peak-and-valley structure appeared when using 1.0% and 2.0% PIP. These peaks and valleys may be caused from the rough support layer as the lower PIP concentrations yield thinner films. For 3.0% PIP based membrane, it seems that the selective layer entirely covered the features of the support and peak-and-valley structure disappeared. Generally, our PIP based TFC membranes gave rougher surfaces than commercial NF Cross-sectional scanning electron micrographs The cross-sectional SEM images for poly(piperazinamide) selective layers made at different PIP concentrations are shown in Fig. 3. The corresponding thickness as a function of PIP concentration is shown in Table 1. Ultra-thin poly(piperazinamide) layers with the thickness less than 1 mm were obtained. The thickness of the selective layers first gradually increased with increasing PIP concentration up to 2%, and then dramatically increased to 0.9mm at 3% PIP. It is also important to note that at 3% PIP, the poly(piperazinamide) selective layer seems to delaminate from the support. On the other hand, the resulting TFC selective layers made at lower PIP concentration (i.e., less than 1%) were better integrated with the support, indicating better adhesion with the support Reverse osmosis tests Fig. 4 illustrates the MgSO 4 rejection rates for PIP based TFC membranes as a function of PIP concentration. The rejection rate increased with increase in PIP concentration until 2% due to increased density of the film, then unexpectedly drop at 3%. This dramatic performance drop for 3% PIP selective layer may be due to the observed delamination of the selective layer in the SEM images (Fig. 3). It is also seen that the rejection for 0.25% and 0.5% PIP based membranes decreased with increasing hydraulic pressure, indicating the resulting films became more flexible and/or weaker under higher pressures. However, the rejection rates of 1% and 2% PIP based TFC membranes maintained above 95% over a range of hydraulic pressures, which approached the rejection performance of NF270. The good pressure tolerance of our TFC membranes also implies their potential application in pressure retarded osmosis. Table 2 shows the pure water permeability of the PIP based TFC membranes. The water permeability decreased with enhancing PIP concentration due to the increased thickness of the selective layer. Comparing with commercial NF270, our 1% PIP based TFC membrane not only showed a similar rejection rate, but also a matched water flux. Overall performance of our handmade TFC membranes matched up that of a industry standard commercial NF membrane Characterization and performance of polyamide TFC membranes Top surface scanning electron micrographs The top surface SEM images for polyamide selective layers made at different MPD concentrations are shown in Fig. 5. Defect free films with ridge-and-valley structure were obtained for TFC membranes made at all MPD concentrations ranging from 0.25 to 3%. All of the selective layers of our TFC membranes appeared to be rougher than commercial SW30-XLE membranes.

5 L. Huang et al. / Journal of Membrane Science 437 (2013) Fig. 3. Cross-sectional SEM images of poly(piperazinamide) of TFC membranes made s of (a) 0.25% (w/v); (b) 0.5%; (c) 1.0%; (d) 2.0%; and (e) 3.0%. Table 1 Average thickness of poly(piperazinamide) selective layers made at different PIP concentrations. Table 2 Pure water permeability coefficient of the TFC membranes made at different PIP concentrations. Average thickness of the selective layer (mm) Pure water permeability (LMH/bar) 0.25% PIP % % % % % PIP % % % % 3.1 NF Fig. 4. MgSO 4 rejection of TFC membranes fabricated from different PIP concentrations. Tests were conducted over a range of hydraulic pressures in reverse osmosis. Experimental Conditions: 2000 ppm Mgso 4 as the feed; cross-flow velocity of 0.26 m/s; 25 1C Cross-sectional scanning electron micrographs The cross-sectional SEM images for polyamide selective layers made at different MPD concentrations are shown in Fig. 6. The corresponding measured thickness of the polyamide layers as a function of MPD concentration is shown in Table 3. The thickness slightly increased with increasing MPD concentration (approximately 50 nm over the range of MPD concentrations). This is a different result than the poly(piperazinamide) which exhibited a stronger thickness dependence on monomer concentration. It is due to differences in the diffusivities and the reacting kinetics with TMC of the two amines. The interfacial polymerization is described to take place in three steps: incipient film formation, a fast process followed by slowing down in polymerization depending upon the permeability of the initial film formed; and finally shifting to a diffusion controlled process. The initial layer formed during the incipient film formation is the actual barrier layer. Then film growth takes place until diffusion of monomers starts to be limited [35]. The termination of the reaction is explained by slower diffusion of diamines as well as by hydrolysis of the acid chlorides that competes with the polymerization [36]. For MPD based polyamide, it is possible that at the low concentration of MPD (r0.5%), increase in MPD concentration up to 0.5% leads to the formation of barrier layer with maximum thickness corresponding to 0.15% TMC concentration. Further increase in MPD concentration can lead to some accumulation of MPD in the amino end group rich region of the thin film. This could increase the density of the thin film since the film thickness remains almost unchanged at this stage. However, PIP reacts with TMC much slower than MPD, so it would take longer time to form a relatively thick barrier layer under higher PIP concentration. Furthermore, the addition of acid acceptor consumes most of the acid chlorides formed during reaction, which might postpone the completion of polymerization.

6 146 L. Huang et al. / Journal of Membrane Science 437 (2013) Fig. 5. Top surface SEM images (magnification 2200 ) of polyamide of TFC membranes made at MPD concentrations of (a) 0.25% (w/v); (b) 0.5%; (c) 1.0%; (d) 2.0%; and (e) 3.0%; and of (f) commercial SW-30XLE. Fig. 6. Cross-sectional SEM images of polyamide of TFC membranes made at MPD concentrations of (a) 0.25% (w/v); (b) 0.5%; (c) 1.0%; (d) 2.0%; and (e) 3.0%. Table 3 Average thickness of polyamide selective layers made at differnent MPD concentrations. 0.25% MPD % % % % 0.12 Average thickness of the selective layer (mm) Reverse osmosis tests The NaCl rejection rates for MPD based TFC membranes as a function of MPD concentration are shown in Fig. 7. The rejection rate of 0.25% MPD based TFC membrane dramatically dropped at 300 psi due to the low mechanical strength of the thin selective layer. Similar to PIP based TFC membranes, salt rejection of these membranes increased with MPD concentration (up to 2% MPD). It was due to the enhancement of crosslinking density which leaded to the formation of denser polyamide layers. Also, the chain flexibility of the polyamide may decrease with increasing concentration of MPD [28]. The rejection rate at 3% MPD may have dropped due to a reduced water flux which reduces dilution of the salt flux. As dilution effect implies, high permeate flux dilutes the salt crossing the membrane, which in turns increases the salt rejection rate. In comparison with commercial SW30-XLE, the 1% and 2% MPD based TFC membranes showed matched rejection rates, which maintained above 95%. It is important to note that they can also tolerate hydraulic pressure of at least 300 psi without compromising the rejection performance, meaning they might be suitable for PRO. Table 4 also compares the pure water permeability of MPD based TFC membranes with commercial SW30-XLE. Similar to PIP based TFC membranes, increasing MPD concentration caused a water flux decline due to the increased crosslinking density and decreased chain flexibility of the polyamide film. In comparison with commercial SW30-XLE, 1% MPD seemed to be the optimal concentration because the resulting TFC membrane not only showed an excellent rejection rate, but also a relatively high water flux, which approached the performance of a commercial sea water RO membrane.

7 L. Huang et al. / Journal of Membrane Science 437 (2013) Fig. 7. NaCl rejection of TFC membranes fabricated from different MPD concentrations. Tests were conducted over a range of hydraulic pressures in reverse osmosis. Experimental conditions: 2000 ppm NaCl as the feed; cross-flow velocity of 0.26 m/s; 25 1C. Table 4 Pure water permeability coefficient of TFC membranes made at different MPD concentrations. 0.25% MPD % % % % 0.64 SW30-XLE 1.08 Pure water permeability (LMH/bar) 3.4. Osmotic flux performance of polyamide TFC membranes. The cross-flow reverse osmosis tests revealed that 1% PIP and 1% MPD based TFC membranes approached the performance of commercial NF270 and SW30-XLE, respectively and hence were considered as optimum samples. Osmotic flux tests using 1.5 M NaCl as the draw solution were conducted on 1% MPD based TFC membranes (referred to TFC EO) and commercial HTI-CTA membrane (referred to HTI) was used as the control. Fig. 8a presents the osmotic water flux performance of both our TFC EO and commercial FO membranes. It can be seen that our TFC EO membrane yielded higher water flux than HTI in the PRO mode. This is primarily due to the higher permselectivity of the TFC membranes. As can be seen from Table 5, the TFC EO membrane exhibits a higher water permeability coefficient and lower solute permeability than the HTI FO membrane. This observation is expected because TFC membranes generally have higher permselectivity than asymmetric membranes [37] due to their ultra-thin, dense, and crosslinked structure of polyamide. The lower solute permeability reduces salt crossover from the draw solution that induces ICP. In PRO mode with DI as the feed, only solute crossing over the selective layer can induce ICP and hence the flux performance is largely dependent on the rejecting ability of the membrane. In FO mode, both TFC EO and commercial HTI membranes showed lower water flux than that in PRO mode due to the more severe ICP that occurred when the support Fig. 8. Comparison of (a) osmotic water flux and (b) reverse salt flux between TFC EO and HTI membranes in both PRO and FO modes. Experimental conditions: 1.5 M NaCl as the draw solution; DI water as the feed; cross-flow velocity of 0.18 m/s; 20 1C. Table 5 Summary of salt rejection (%R), pure water permeability coefficient (A), solute permeability (B), specific salt flux (J s /J w ) and structural parameter (S) of both TFC EO and HTI membranes. %R a A a (LMH/bar) B a (LMH) J s /J w (g/l) S (mm) PRO mode FO mode TFC EO HTI a These parameters are measured in cross-flow reverse osmosis. Experimental conditions: 2000 ppm NaCl feed solution, 150 psi applied pressure, cross-flow velocity of 0.26 m/s, and temperature of 20 1C. was facing the draw solution. What s more, in FO mode, our TFC EO membrane showed slightly lower water flux than the HTI membrane primarily due to the thickness of the 3M membranes used as supports (180 mm compared to approximately 50 mm of HTI membrane). Thicker supports enhance ICP and are generally not desired in FO membranes. Interestingly, the nylon 6,6 supports exceed the HTI membrane support thickness by a factor of three yet water flux performance is nearly equal. The reverse salt flux performance for both TFC EO and HTI membranes are presented in Fig. 8b. In both PRO and FO mode, our TFC membranes showed approximately 10 times lower reverse salt flux than HTI which is primarily due to the higher selectivity of TFC membranes than integrated asymmetric membranes. The reverse salt passage for both membranes in FO mode was slightly lower than those in PRO mode. This is primarily due to the higher salt concentration difference, and hence effective

8 148 L. Huang et al. / Journal of Membrane Science 437 (2013) driving force, at the selective layer interface in PRO mode than that in FO mode. Overall, in all tests the membrane exhibited consistent performance and showed no performance change due to the presence of high concentrations of salt. No delamination of the selective layer was observed during the tests. It is worth mentioning that this new TFC EO membrane yielded approximately 20 fold higher water flux with the same order of magnitude of salt flux compared to commercial TFC RO membranes (SW30-XLE) reported in other studies under similar experimental conditions in PRO mode [32]. Since our TFC EO membrane showed similar rejection rate and pure water permeability to SW30-XLE, the dramatic improvement in water flux must be attributed to the reduced ICP within the support. We feel the reduced ICP for this nylon 6,6 support is more due to its hydrohphilicity than its structure. The PSu-PET nonwoven support for a Dow RO membrane generally has a comparable thickness ( 150 mm) and porosity (approximately 50% [38]) to nylon 6,6 support (180 mm and approximately 5060% porosity). Improved wetting of this layer will dramatically reduce the effective structural parameter and improve osmotic flux [14]. Furthermore, we hypothesize that the higher surface porosity of the 3M support at the interface with the selective layer improves the water flux because the selective layer is not shadowed by the support. The more of the selective layer not blocked by the support facilitates the transport of both water and salts away from the interface. The surface porosity of this support was measured to be approximately 50%, while that of commercial RO membrane is reported to be less than 20% [33]. Table 5 also summarizes the comparison of the specific salt flux and between the TFC membranes and HTI membrane. Specific salt flux represents the amount of draw solute loss per liter of water produced [30,31]. The specific salt flux should be as low as possible in order to reduce the loss of draw solute. It can be seen from the table that our TFC EO membrane exhibited 28 times lower specific salt flux in PRO mode and 8 times lower in FO mode when compared to the HTI membrane. These results indicate our TFC EO membrane would save draw solution in FO and exhibit reduced ICP in PRO. As is common with new FO membrane characterization, the structural parameter, S, can be calculated from osmotic flux tests. The structural parameter implies how severe the ICP effect is and it should also be as low as possible to maximize the water flux. Table 5 shows that our TFC EO membrane has two times higher structural parameter of the commercial HTI membrane. This is attributed largely to the thicker support in our TFC membranes. This result was interesting given that the water flux through our membranes matched or exceeded the HTI membrane. This is further evidence that both the support layer properties and the selectivity substantially impact osmotic water flux performance. 4. Conclusions For the first time, hydrophilic nylon 6,6 supported TFC membranes were successfully fabricated via in-situ interfacial polymerization. 1% PIP based TFC membranes showed matched performance with NF270 and 1% MPD based TFC membranes showed matched performance with SW30-XLE in RO. The osmotic flux tests demonstrated that our TFC EO membranes had higher water flux than commercial HTI in PRO mode and matched water flux in FO mode. Furthermore, our TFC EO membrane also showed 10X lower reverse salt flux and 8 28 specific salt flux in FO or PRO mode. This excellent performance was found even though our membrane has a two-fold larger structural parameter than the HTI membrane. Acknowledgments The authors gratefully acknowledge funding from the National Science Foundation (CBET # ), the Department of Energy, USEPA (Project #R834872), and 3M TM. We thank 3M for providing microfiltration membranes, Dow Water & Process Solutions TM for providing reverse osmosis and nanofiltration membranes, and Hydration Technology Innovations for providing their forward osmosis membrane. References [1] R.L. McGinnis, M. Elimelech, Global challenges in energy and water supply: The promise of engineered osmosis, Environ. Sci. Technol. 42 (2008) [2] T.Y. Cath, A.E. Childress, M. Elimelech, Forward osmosis: Principles, applications, and recent developments, J. Membr. Sci. 281 (2006) [3] K.L. Lee, R.W. Baker, H.K. Lonsdale, for power-generation by pressure-retarded osmosis, J. Membr. Sci. 8 (1981) [4] J.R. McCutcheon, M. Elimelech, Influence of concentrative and dilutive internal concentration polarization on flux behavior in forward osmosis, J. Membr. Sci. 284 (2006) [5] J.R. McCutcheon, R.L. McGinnis, M. Elimelech, A novel ammonia carbon dioxide forward (direct) osmosis desalination process, Desalination 174 (2005) [6] J.R. McCutcheon, R.L. McGinnis, M. Elimelech, Desalination by ammonia carbon dioxide forward osmosis: Influence of draw and feed solution concentrations on process performance, J. Membr. Sci. 278 (2006) [7] I.M. Hutten, Handbook of nonwoven filter media, 1st ed., Butterworth- Heinemann, Oxford; Burlington, MA, [8] R.W. Baker, Membrane Technology and Applications, Wiley, West Sussex, [9] M. Mulder, Basic principles of membrane technology, 2nd ed., Kluwer Academic, Dordrecht; Boston, [10] G.D. Mehta, Further results on the performance of present-day osmotic membranes in various osmotic regions, J. Membr. Sci. 10 (1982) [11] G.D. Mehta, S. Loeb, Internal polarization in the porous substructure of a semipermeable membrane under pressure-retarded osmosis, J. Membr. Sci. 4 (1978) [12] G.D. Mehta, S. Loeb, Performance of permasep B-9 and B-10 membranes in various osmotic regions and at high osmotic pressures, J. Membr. Sci. 4 (1979) [13] G.T. Gray, J.R. McCutcheon, M. Elimelech, Internal concentration polarization in forward osmosis: role of membrane orientation, Desalination 197 (2006) 1 8. [14] J.R. McCutcheon, M. Elimelech, Influence of membrane support layer hydrophobicity on water flux in osmotically driven membrane processes, J. Membr. Sci. 318 (2008) [15] J. Herron, Asymmetric Forward Osmosis, US B2, [16] S. Loeb, L. Titelman, E. Korngold, J. Freiman, Effect of porous support fabric on osmosis through a Loeb-Sourirajan type asymmetric membrane, J. Membr. Sci. 129 (1997) [17] N.Y. Yip, A. Tiraferri, W.A. Phillip, J.D. Schiffman, M. Elimelech, High performance thin-film composite forward osmosis membrane, Environ. Sci. Technol. 44 (2010) [18] N.-N. Bui, M.L. Lind, E.M.V. Hoek, J.R. McCutcheon, Electrospun nanofiber supported thin film composite membranes for engineered osmosis, J. Membr. Sci (2011) [19] R. Wang, L. Shi, C.Y.Y. Tang, S.R. Chou, C. Qiu, A.G. Fane, Characterization of novel forward osmosis hollow fiber membranes, J. Membr. Sci. 355 (2010) [20] S. Zhang, K.Y. Wang, T.-S. Chung, H. Chen, Y.C. Jean, G. Amy, Well-constructed cellulose acetate membranes for forward osmosis: Minimized internal concentration polarization with an ultra-thin selective layer, J. Membr. Sci. 360 (2010) [21] K.Y. Wang, T.-S. Chung, J.-J. Qin, Polybenzimidazole (PBI) nanofiltration hollow fiber membranes applied in forward osmosis process, J. Membr. Sci. 300 (2007) [22] K.Y. Wang, Q. Yang, T.-S. Chung, R. Rajagopalan, Enhanced forward osmosis from chemically modified polybenzimidazole (PBI) nanofiltration hollow fiber membranes with a thin wall, Chem. Eng. Sci. 64 (2009) [23] H. Yasuda, C.E. Lamaze, Permselectivity of solutes in homogeneous waterswollen polymer membranes, J. Macromol. Sci., Part B 5 (1971) [24] H. Muta, M. Miwa, M. Satoh, Ion-specific swelling of hydrophilic polymer gels, Polymer 42 (2001) [25] G.E. Zaikov, Y.A. Mikheevch, Kinetics And Mechanisms Of Chemical Reactions, Nova Publishers, [26] C.Y. Tang, Y.-N. Kwon, J.O. Leckie, Effect of membrane chemistry and coating layer on physiochemical properties of thin film composite polyamide RO and NF membranes: I. FTIR and XPS characterization of polyamide and coating layer chemistry, Desalination 242 (2009)

9 L. Huang et al. / Journal of Membrane Science 437 (2013) [27] L. Yung, H. Ma, X. Wang, K. Yoon, R. Wang, B.S. Hsiao, B. Chu, Fabrication of thin-film nanofibrous composite membranes by interfacial polymerization using ionic liquids as additives, J. Membr. Sci. 365 (2010) [28] N.K. Saha, S.V. Joshi, Performance evaluation of thin film composite polyamide nanofiltration membrane with variation in monomer type, J. Membr. Sci. 342 (2009) [29] E.M.V. Hoek, A.S. Kim, M. Elimelech, Influence of crossflow membrane filter geometry and shear rate on colloidal fouling in reverse osmosis and nanofiltration separations, Environ. Eng. Sci. 19 (2002) [30] W.A. Phillip, J.S. Yong, M. Elimelech, Reverse Draw Solute Permeation in Forward Osmosis: Modeling and Experiments, Environ. Sci. Technol. 44 (2010) [31] N.T. Hancock, T.Y. Cath, Solute coupled diffusion in osmotically driven membrane processes, Environ. Sci. Technol. 43 (2009) [32] J.T. Arena, B. McCloskey, B.D. Freeman, J.R. McCutcheon, Surface modification of thin film composite membrane support layers with polydopamine: Enabling use of reverse osmosis membranes in pressure retarded osmosis, J. Membr. Sci. 375 (2011) [33] A.K. Ghosh, E.M.V. Hoek, Impacts of support membrane structure and chemistry on polyamide polysulfone interfacial composite membranes, J. Membr. Sci. 336 (2009) [34] P.S. Singh, S.V. Joshi, J.J. Trivedi, C.V. Devmurari, A.P. Rao, P.K. Ghosh, Probing the structural variations of thin film composite RO membranes obtained by coating polyamide over polysulfone membranes of different pore dimensions, J. Membr. Sci. 278 (2006) [35] R. Nadler, S. Srebnik, Molecular simulation of polyamide synthesis by interfacial polymerization, J. Membr. Sci. 315 (2008) [36] V. Enkelmann, G. Wegner, Mechanism of interfacial polycondensation and the direct synthesis of stable polyamide membranes, Die Makromolekulare Chemie 177 (1976) [37] R.J. Petersen, Composite reverse-osmosis and nanofiltration membranes, J. Membr. Sci. 83 (1993) [38] J. Wei, C. Qiu, C.Y. Tang, R. Wang, A.G. Fane, Synthesis and characterization of flat-sheet thin film composite forward osmosis membranes, J. Membr. Sci. 372 (2011)