Direct Microscopic Observation of Forward Osmosis Membrane

Aug 24, 2010 - Forward osmosis (FO) is an osmotically driven membrane process in ... FO setup coupled with a direct microscopic observation facility u...
3 downloads 0 Views 6MB Size
Download PDF
Environ. Sci. Technol. 2010, 44, 7102–7109

Direct Microscopic Observation of Forward Osmosis Membrane Fouling Y I N I N G W A N G , †,‡ F I L I C I A W I C A K S A N A , ‡ C H U Y A N G Y . T A N G , * ,†,‡ A N D A N T H O N Y G . F A N E †,‡ School of Civil and Environmental Engineering and Singapore Membrane Technology Centre, Nanyang Technological University, Singapore 639798

Received June 12, 2010. Revised manuscript received August 13, 2010. Accepted August 13, 2010.

This study describes the application of a noninvasive direct microscopic observation method for characterizing fouling of a forward osmosis (FO) membrane. The effect of the draw solution concentration, membrane orientation, and feed spacer on FO fouling was systematically investigated in a cross-flow setup using latex particles as model foulant in the feedwater. Higher draw solution (DS) concentrations (and thus increased flux levels) resulted in dramatic increase in the surface coverage by latex particles, suggesting that the critical flux concept might be applicable even for the osmotically driven FO process. Under identical draw solution concentrations, the active-layerfacing-the-feed-solution orientation (AL-FS) experienced significantly less fouling compared to the alternative orientation. This may be explained by the lower water flux in AL-FS, which is consistent with the critical flux concept. The use of a feed spacer not only dramatically enhanced the initial flux of the FO membrane, but also significantly improved the flux stability during FO fouling. Despite such beneficial effects of usingthefeedspacer,asignificantamountofparticleaccumulation was found near the spacer filament, suggesting further opportunities for improved spacer design. To the best of the authors’ knowledge, this is the first direct microscopic observation study on FO fouling.

1. Introduction Forward osmosis (FO) is an osmotically driven membrane process in which a water flux is drawn spontaneously from a low-osmotic-pressure feed solution (FS) to a high-osmoticpressure draw solution (DS) across a dense solute-rejecting membrane (1–3). Compared to traditional pressure-driven membrane processes, FO offers opportunities for significant energy savings because the process does not require the application of hydraulic pressure (2–4). However, the performance of an FO process can be severely limited by membrane fouling (5–8) as well as concentration polarization inside the porous support structure of the FO membrane (i.e., internal concentration polarization or ICP) (7–13). Although the effect of ICP on FO flux behavior has been extensively reported in the literature (7–13), only a handful of studies have focused on FO membrane fouling (5–8, 14). * Corresponding author phone: +65 6790 5267; fax: +65 6791 0676; e-mail: [email protected]; mailing address: 50 Nanyang Avenue #N11b-35, School of Civil and Environmental Engineering, Nanyang Technological University, Singapore 639798. † School of Civil and Environmental Engineering. ‡ Singapore Membrane Technology Centre. 7102

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 18, 2010

In addition, most of the existing FO fouling studies focused primarily on flux decline behavior, with limited attempts to characterize the foulant deposition process on the membrane (7, 14). In this regard, foulant characterization methods, in particular, noninvasive direct observation techniques, are important to further advance our understanding of FO membrane fouling. Such methods can potentially provide deep insights into FO fouling and its dependence on operational conditions. Among the various noninvasive fouling characterization methods, direct microscopic observation has been widely used to study the development of fouling in pressure-driven membrane systems. For example, this method has been successfully applied to study the development of biofouling of nanofiltration membranes (15, 16). Similarly, Fane et al. (17–19) have developed an optical microscopic method (i.e., the “direct observation through the membrane” (DOTM) method) to visualize the deposition of particulate and/or colloidal foulants onto a transparent anodized aluminum microfiltration membrane. DOTM has been proven to be a valuable tool to study particle transport and deposition at the membrane-solution interface. By relating the surface cake layer coverage to the membrane permeate flux level, Fane et al. (17–20) were also able to use this method to study the critical flux phenomenon for porous membranes. Despite the prior success of direct microscopic observation to elucidate fouling in pressure-driven membrane processes, its application to characterize FO fouling has not been reported in the literature. The objective of the current study was to develop a direct microscopic observation method for FO fouling characterization. In addition, the effect of draw solution concentration, membrane orientation, and feed spacer were systematically studied under cross-flow FO condition using monodispersed latex particles as a model foulant. To the best of our knowledge, this is the first direct microscopic study on FO membrane fouling.

2. Materials and Methods 2.1. Chemicals. Unless otherwise specified, all chemicals and reagents used in this study were of analytical grade. Sodium chloride was used as the draw solute. Ultrapure water with a resistivity of 18.2 MΩ.cm (Millipore Integral 10 Water Purification System) was used to prepare all working solutions. Latex microparticles based on polystyrene with a mean diameter of 3 µm (Fluka-Sigma Aldrich) were used as model foulants in the feedwater due to their well-defined size and shape (nearly monodispersed spherical particles). Such particles have been used previously in direct microscopic observation of fouling phenomena in pressure-driven membrane systems (17). 2.2. FO Membrane. The forward osmosis (FO) membrane obtained from Hydration Technology Inc. (Hydrowell Filter, HTI) is made of cellulose triacetate (CTA) supported by an embedded polyester mesh (2, 7). According to our previous study (7), the thickness of the membrane varied from 30 to 50 µm depending on the relative location to the polyester mesh fibers. Thus, the HTI membrane is significantly thinner than conventional thin film composite (TFC) reverse osmosis (RO) (thickness ∼ a few hundred micrometers (21)), which is believed critical to reduce internal concentration polarization (7, 9, 13). The contact angles of the active layer and the support layer were 76° and 87°, respectively (7). Atomic force microscopic (AFM) measurement showed that the mean roughness of the active layer was ∼36 nm (7), which is significantly smoother than typical TFC reverse osmosis 10.1021/es101966m

 2010 American Chemical Society

Published on Web 08/24/2010

membranes (21, 22). The reported water permeability (tested in RO mode) was about 0.80 L/m2 · hr · bar (7, 11). Additional scanning electron microscopic (SEM, Zeiss Evo 50) characterization was performed in the current study. This was to allow a comparison between the optical microscopic observation and the SEM observation. To study the internal membrane pore structure, the polyester mesh was carefully removed from the membrane in some cases. The original membrane (with the polyester mesh intact) as well as the separated layers (CTA layer and polyester mesh) were freezedried (Christ Alphr 1-4 LD, Germany) before being sputtercoated with a uniform layer of gold (Emitech SC7620 sputter coater). The accelerating voltage used to scan all samples was 20 kV.

2.3. Bench-Scale FO Setup and Direct Microscopic Observation System. Figure S1 (Supporting Information) illustrates the bench-scale FO setup coupled with a direct microscopic observation facility used in this study. A crossflow membrane filtration cell was made of perspex which allowed light transmission through the transparent membrane. A membrane coupon with an effective area of 29.2 cm2 was housed in the cell. Cross flow was maintained for both the draw solution and the feedwater using Masterflex peristaltic pumps (Cole Parmer). The FO water flux was determined by measuring the weight changes of the feed solution at regular time intervals with a precision balance (Mettler-Toledo) connected to a computer and a data logging system (LabVIEW).

FIGURE 1. SEM (a-d) and optical (e and f) microscopic characterization of the HTI FO membrane: (a) dense skin layer (active layer) of the FO membrane, (b) reverse side of the FO membrane with the polyester mesh peeled off, (c) isolated polyester mesh, (d) cross-sectional image of the FO membrane, (e) clean membrane with its support layer facing the optical microscope, (f) clean membrane with the active layer facing the optical microscope. VOL. 44, NO. 18, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7103

The microscopic observation system was modified from the DOTM facility reported elsewhere (17) to suit the FO fouling study. In the original DOTM technique (17), the microscope objective lens was placed on the permeate side of the transparent anodized aluminum microfilter, and it was adjusted to focus through the membrane and at the membrane-feed solution interface. However, this was not the optimal configuration in the current study due to the potential interference from the FO membrane itself as a result of the embedded polyester fibers as well as the unique membrane pore structure (Section 3.1). Consequently, the objective lens of the microscope (10× magnification, Axiolab, Carl Zeiss) was placed on the feed side, and particle deposition was observed directly on the membrane surface throughout the current study. This significantly improved the contrast between the foulant particles and the FO membrane. The feed solution turbidity did not appear to be a problem due to the relatively low latex concentration (0.01% w/v) and the sufficient frame rate of the camera used in this study (30 frames/second, high-resolution color CCD camera, JVC, model TK-C921BEG). In addition, the effect of gravitational settling of the particles was shown to be negligible for small latex particles (23). 2.4. FO Fouling Tests. For each fouling test, a new FO membrane coupon was used. The membrane was first equilibrated with a 10 mM NaCl feed solution for 30 min under cross-flow. A latex suspension was then added into the feed tank to achieve a particle concentration of 0.01% (w/v) in the feedwater. The ionic strength of the foulantcontaining feedwater was adjusted to 10 mM on the basis of conductivity measurement (Ultrameter II, Myron L Company, Carlsbad, CA). The water flux was recorded during the entire fouling experiment, with the initial flux defined as the flux right before the addition of latex particles. Latex deposition was observed with the direct microscopic observation system. The effect of draw solution concentration (0.5-4 M), membrane orientation, and the presence of a spacer in the feed flow channel were systematically investigated in the current study (Table S1, Supporting Information). Unless otherwise specified, the following reference conditions were applicable for the tests: • diamond-patterned spacer placed in draw solution flow channel;

• membrane dense rejection layer (active layer) facing the draw solution (except where the effect of membrane orientation was evaluated); • crossflow velocity maintained at 9 ( 1.5 cm/s on both sides of the membrane; • feed solution pH ∼ 5.8; • temperature of both feed and draw solutions 23 ( 1 °C. As the FO water flux can be affected by factors other than fouling (such as the volumetric dilution of draw solution (7)), baseline tests were also performed as control experiments. Baseline tests were conducted under testing conditions otherwise identical to the respective fouling tests, except latex particles were not introduced to the feed solution. Thus, the flux difference between a fouling test and the corresponding baseline test can be attributed to FO fouling. 2.5. Optical Microscopic Image Analysis. The optical microscopic images of latex particle deposition on the membrane surface were saved in a Tagged-Image File Format (TIFF). The images were then analyzed using image analysis software (MediaCybernetics, Inc.). Prior to analysis, the images were converted to a greyscale image that contained 8 bits of data per pixel. The greyscale image was then transformed to a binary image (i.e., a black and white image) by adjusting the threshold value. A series of image editing steps were performed to minimize the impact of background noise. The membrane surface coverage was determined as the percentage of membrane area covered by latex particles to the total membrane surface area viewed under the microscope (0.17 mm2).

3. Results and Discussion 3.1. Microscopic Observation of FO Membrane Structure. The morphology and structure of the clean FO membrane were characterized by both SEM and the optical microscope (Figure 1a-f). Figure 1a shows the SEM image of the rejection layer of the FO membrane, while Figure 1b shows the support side of the membrane with the polyester mesh peeled away. The FO active surface (i.e., the rejection layer) appeared to be rough, with peak-to-peak distance slightly over 100 µm (Figure 1a). The surface roughness was likely caused by the embedded polyester mesh, as the peak-to-peak distance matches well with the mesh fiber-to-fiber distance (Figure

FIGURE 2. Images of particle deposition captured at 20, 40, and 70 min after latex addition: (a) 2.0 M DS; (b) 3.0 M DS. Other experimental conditions: AL-DS orientation, cross-flow velocity ∼9 cm/s for both FS and DS, spacer used in the DS flow channel but not in the FS flow channel, and temperature at 23 °C. The feed solution contained 10 mM NaCl at pH 5.8. 7104

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 18, 2010

1c and d). Similar roughness structure can also be observed on the support side of the membrane (Figure 1d). The crosssectional SEM image (Figure 1d) shows that the overall thickness of the membrane was around 50 µm, which is consistent with previous reports (2, 7). The FO membrane has a unique asymmetric structure, with a porous internal layer sandwiched between the active rejection layer and another “skin-like” layer on the support side. The internal pore sizes were on the order of several tens of micrometers, with smaller pores adjacent to the dense active layer and bigger pores toward the “support skin”. Due to the presence of this “support skin”, the micrometer-sized internal pores were not visible when SEM images were taken directly from the support side (refer to ref 7). However, upon peeling away the polyester mesh together with the “support skin”, the internal pore structure of the FO membrane was clearly visible under SEM (Figure 1b). The pore size observed (for pores adjacent to the “support skin” of the membrane) ranged approximately from 20 to 40 µm. The optical microscopic images are presented in Figure 1e (support layer facing the microscope) and Figure 1f (active layer facing the microscope). The polyester mesh was clearly visible under the optical microscope in both orientations. The mesh fiber diameter (∼40 µm) and the fiber-to-fiber distance (∼140 µm) agreed very well with the SEM observations (Figure 1c and d), which validates the optical microscopic observation method. Interestingly, a bubbly void structure was also observed with the “voids” ranging from 20 to 50 µm in diameter (Figure 1e). This structure had a pattern and characteristic size similar to the pore structure observed in Figure 1b, which suggests that the “voids” were likely the macropores near the support side, since the optical microscope was focused near the support surface in this membrane orientation. In contrast, the “voids” observed were much smaller when the microscope was focused near the active layer of the membrane (Figure 1f), which is consistent with the SEM observation that pore size was smaller close to the active layer (Figure 1d). The current study suggests that the optical microscopic characterization can potentially be a powerful method to characterize the internal FO membrane pore structure, although systematic studies are still needed to further optimize the method. 3.2. Microscopic Observation of FO Fouling. 3.2.1. Effect of Draw Solution Concentration and FO Flux Level. As FO is an osmotically driven membrane process, the draw solution concentration plays a critical role in FO performance (6–8). To study the effect of DS concentration on FO membrane fouling, four different DS concentrations (0.5, 2.0, 3.0, and 4.0 M) were investigated for the activelayer-facing-DS orientation (AL-DS). For each DS concentration, the latex particle deposition on the membrane surface was monitored as a function of time. Compared to the clean membrane (Figure 1e), the deposited latex particles were easily visible for the fouled membrane (Figure 2a and b), suggesting that the direct microscopic observation method developed in the study is suitable for FO fouling investigation. The foulant deposition on the FO membrane was nearly negligible for a 0.5 M DS (Figure 3a). Figure 2a shows optical microscopic images of latex deposition during FO fouling (20, 40, and 70 min) for a 2.0 M DS. A small amount of particle deposition was observed, and the amount of deposition increased slowly with time. At 70 min, the surface coverage by latex particles reached approximately 4% of the total membrane surface (Figure 3a). The corresponding flux decline with respect to the baseline was ∼8%. Most of the deposited particles were found in the valley regions adjacent the embedded polyester fibers. Such preferential deposition was probably caused by the local hydrodynamic conditions. Using the direct microscopic observation method, it was observed that the latex particles moved significantly slower

FIGURE 3. Effect of draw solution concentration and FO flux level on FO fouling: (a) surface coverage as a function of time at different DS concentrations; (b) surface coverage as a function of initial flux (FO flux at time 0). Other experimental conditions: AL-DS orientation, cross-flow velocity ∼9 cm/s for both FS and DS, spacer used in the DS flow channel but not in the FS flow channel, and temperature at 23 °C. The feed solution contained 10 mM NaCl at pH 5.8. in the vicinity of the polyester fibers, which promoted their deposition over these regions. Thus, the direct microscopic observation in the current study revealed further opportunities for membrane surface morphology optimization, with a smoother membrane surface likely preferred over rougher ones (24, 25). Significantly higher surface coverage by latex particles was observed for a 3.0 M DS (Figure 2b). Even for a short fouling duration of 20 min, a few patches of deposition were found on the FO surface (∼2% surface coverage). Once again, these early deposits were located mostly next to the edge of the embedded polyester fibers, consistent with the pattern observed for 2 M DS (Figure 2a). The latex coverage was rapidly increased to ∼32% at 40 min and started to become more uniformly distributed on the membrane surface. At 70 min, a nearly uniform coverage (∼42%) was found on the membrane. Figure 3a plots the membrane surface coverage against fouling duration for different DS concentrations. Clearly, the draw solution concentration played a critical role in FO fouling. While particle deposition was negligible VOL. 44, NO. 18, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7105

FIGURE 4. Effect of membrane orientation on FO fouling: (a) optical micrograph for a fouled membrane in AL-DS at 70 min; (b) optical micrograph for a fouled membrane in AL-FS at 70 min; (c) surface coverage vs time for both AL-DS and AL-FS orientations. Other experimental conditions: 4.0 M NaCl as DS, cross-flow velocity ∼9 cm/s for both FS and DS, spacer used in the DS flow channel but not in the FS flow channel, and temperature at 23 °C. The feed solution contained 10 mM NaCl at pH 5.8. at the low DS concentration of 0.5 M, small patches were observable for a 2 M DS. At higher DS concentrations, the surface coverage was extensive (∼40% for 3 M DS and ∼70% for 4 M DS at 70 min). Corresponding to the dramatic particle deposition at higher DS concentrations, the flux decline was also more severe for the 3 and 4 M DS (Figure S2, Supporting Information). Similar effect of draw solution concentration has been previously reported by Tang and co-workers (7) for FO membranes fouled by humic acid, who observed that the flux reduction was much more severe for more concentrated draw solutions. The greater fouling propensity at higher DS concentrations may be explained by the higher flux levels as a result of increased FO driving force. For pressure-driven membrane processes, it is well-known that elevated flux levels inevitably lead to severe permeability loss (26, 27). In contrast, fouling is minimal when flux is below some threshold values a phenomenon known as the critical flux behavior (26, 27). The current study convincingly demonstrated that the critical flux concept is also applicable to the osmotically driven FO process. The surface coverage is plotted against the initial FO flux (i.e., the flux at time 0) in Figure 3b. FO fouling was negligible at a flux of 15 L/m2 · h. Small amount of surface coverage started to appear at 28 L/m2 · h. On the other hand, surface coverage was drastically increased when the flux was greater than 41 L/m2 · h. This suggests that the FO critical flux during latex particle filtration was probably close to 28 L/m2 · h. Like the DOTM technique which has been applied to determine the critical flux for porous membranes (17–20), the direct microscopic observation method in this study is capable of 7106

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 18, 2010

determining the critical flux for FO membranes. The accurate determination of the FO critical flux requires further refining of the draw solution concentration. 3.2.2. Effect of Membrane Orientation. The effect of membrane orientation was evaluated by performing FO fouling experiments in both the active-layer-facing-the-DS orientation (AL-DS) and the active-layer-facing-the-FS orientation (AL-FS). Compared to the AL-DS orientation, it was more difficult to observe the deposited particles in the AL-FS orientation (Figure 4a and b). As also shown in Figure 1f, the micrograph of the clean FO membrane in AL-FS contained more dark colored bubble-like features, probably due to the presence of numerous smaller pores near the active layer (Section 3.1). These fine structures make it more difficult to identify the latex particles in the AL-FS orientation. Nevertheless, it was still possible to see the dramatic contrast between the two membrane orientations. While an extensive coverage (∼70%) was observed for the AL-DS orientation, the surface coverage in the AL-FS orientation at 4 M DS was nearly zero (Figure 4b and c), indicating that the AL-FS orientation was more fouling resistant compared to AL-DS orientation. This is in good agreement with earlier studies on FO fouling (6, 7). Both Mi and Elimelech (6) and Tang et al. (7) reported better flux stability in the AL-FS orientation, which could be attributed to (1) the lower flux levels in this orientation due to the more severe internal concentration polarization (11), (2) the relatively smoother surface of the active layer compared to the “support skin” (7), and (3) the avoidance of foulant entrapment inside the porous support layer by orienting the dense active layer toward the feed

FIGURE 5. Effect of feed channel spacer on FO flux performance: (a) effect of feed spacer on the initial FO water flux; (b) normalized flux vs time for a 2.0 M DS without spacer in the feed channel; (c) normalized flux vs time for a 2.0 M DS with spacer in the feed channel. Other experimental conditions: cross-flow velocity ∼9 cm/s for both FS and DS, and temperature at 23 °C. The feed solution contained 10 mM NaCl at pH 5.8. solution. These earlier studies used macromolecules (such as protein and humic acid) as model foulants so that foulants could penetrate into the porous support layer in the AL-DS orientation due to their small sizes (, 1 µm). In the current study, the particle entrapment inside the porous support layer in AL-DS was unlikely due to the much bigger particle size used (3 µm). Thus, the remarkable antifouling performance of AL-FS was likely a direct result of its lower water flux (15 L/m2 · h in AL-FS vs 45 L/m2 · h in AL-DS) in addition to the effect of membrane surface morphology. As discussed earlier (Figure 3b), the critical flux of the FO membrane in AL-DS was close to 28 L/m2 · h. The water flux of 15 L/m2 · h in AL-FS was significantly lower than this critical value, which explains the low fouling propensity in this orientation in the current study assuming the two orientations had similar critical flux values. 3.2.3. Effect of Spacer. To investigate the effect of spacer on FO fouling behavior, cross-flow FO filtration was performed with and without a spacer in the feed solution channel for the AL-DS orientation. In both cases, a spacer was used on the draw solution side. Figure 5a shows the effect of feed spacer on the initial FO water flux for draw solution concentrations of 0.5 and 2.0 M. Dramatic flux improvements (∼50% at 0.5 M and nearly 100% at 2.0 M) were achieved by

simply including a spacer on the feed solution side. The beneficial effect of feed spacer is well-known for cross-flow pressure-driven membranes thanks to the improved mass transfer over the membrane surface (28). The current study demonstrates even more significant improvements for the osmotically driven FO process. The FO flux is affected by the compound effect of both internal concentration polarization and external concentration polarization (ECP) (2), and the overall concentration polarization modulus in FO can be determined by the product of the ICP modulus and the ECP modulus (8). Where the feed spacer was not used, the FO membrane had probably experienced a severe ECP, which can promote ICP indirectly (by providing the FO membrane with a higher effective concentration from the feed) and thus resulted in a dramatic flux reduction. The FO flux behavior upon the addition of latex particles is presented in Figure 5b (without spacer) and Figure 5c (with spacer). The baseline test results are also shown for comparison. The difference between the fouling and baseline results can be attributed to the addition of foulants. Without the use of feed spacer, ∼10% flux decline was observed at the end of 2-h fouling run. In contrast, no significant flux loss with respect to the baseline was observed when the feed spacer was inserted in the FO test cell, despite its much higher VOL. 44, NO. 18, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7107

Supporting Information Available Figure S1, Cross flow forward osmosis test setup equipped with a direct microscopic observation facility; Figure S2, Effect of draw solution concentration on FO flux behavior; Table S1, Summary of experimental conditions. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited FIGURE 6. Comparison of particle deposition on FO membrane: (a) without the spacer and (b) with the spacer (spacer thickness of ∼0.7 mm). Images captured at 40 min after latex addition. Other experimental conditions: 2.0 M NaCl as DS, cross-flow velocity ∼9 cm/s for both FS and DS, and temperature at 23 °C. The feed solution contained 10 mM NaCl at pH 5.8. In (b), the spacer was located at the upper right corner. The particles and membrane area under the spacer are also observable, although that portion of the image is slightly distorted due to the deflection of light by the spacer filament. initial flux (∼52 L/m2 · h). It seems that the critical flux was enhanced (> 52 L/m2 · h) upon use of the feed spacer, compared to a critical flux of only ∼28 L/m2 · h where no spacer was used (Section 3.2.1 and Figure 3). Similar effects of spacers have been reported for particulates in microfiltration (29) and colloids in RO (30). The current study revealed a strong positive effect of using feed spacer on the FO flux. The optical micrographs on latex particle deposition are shown in Figure 6. As discussed in Section 3.2.1, a few small patches of deposition were observed when feed spacer was not used (Figure 6a). In the presence of a feed spacer, however, significant particle deposition was found near and under the spacer filament (Figure 6b). Such particle accumulation was likely due to the local hydrodynamic condition (low shear region) as well as physical blockage due to the relatively large particle size used in the current study. Particle deposition was negligible away from the spacer. Interestingly, the particle accumulation near the spacer filament did not seem to cause any flux decline (Figure 5c). It may be hypothesized that the membrane area underneath the spacer had relatively insignificant contribution to the overall permeate water such that local deposition of particles had little effect on the flux stability. Further research is needed to test such hypothesis. Meanwhile, particle deposition near the spacer could potentially lead to an increased pressure drop between the module inlet and outlet in a large membrane module. Thus, the FO spacer needs to be optimized to further enhance the mass transfer in FO processes and to reduce the risk of particle accumulation near the spacer. The real-time noninvasive direct microscopic observation method developed in the current study was proven to be a useful method to study both the FO membrane structure and FO fouling behavior. A critical flux behavior was observed for the osmotically driven FO membrane, with more severe particle deposition found at higher flux levels (greater draw solution concentrations). FO membrane was more resistant to fouling in the AL-FS orientation as a result of the associated lower flux. Feed spacer was able to drastically improve both the initial flux and the critical flux in FO processes, although further systematic optimization of spacer design is still needed.

Acknowledgments We acknowledge the Environment and Water Industry Development Council of Singapore (EWI) for the support under project MEWR C651/06/173 (EWI0901-02-01). The Singapore Membrane Technology Centre is supported by both EWI and Nanyang Technological University. Feedback from the anonymous reviewers has greatly improved the clarity of the manuscript. 7108

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 18, 2010

(1) Loeb, S. Osmotic power plants. Science 1975, 189 (4203), 654– 655. (2) Cath, T. Y.; Childress, A. E.; Elimelech, M. Forward osmosis: Principles, applications, and recent developments. J. Membr. Sci. 2006, 281 (1-2), 70–87. (3) Xu, Y.; Peng, X.; Tang, C. Y.; Fu, Q. S.; Nie, S. Effect of draw solution concentration and operating conditions on forward osmosis and pressure retarded osmosis performance in a spiral wound module. J. Membr. Sci. 2010, 348 (1-2), 298–309. (4) Achilli, A.; Cath, T. Y.; Marchand, E. A.; Childress, A. E. The forward osmosis membrane bioreactor: A low fouling alternative to MBR processes. Desalination 2009, 238 (1-3), 10–21. (5) Cornelissen, E. R.; Harmsen, D.; de Korte, K. F.; Ruiken, C. J.; Qin, J. J.; Oo, H.; Wessels, L. P. Membrane fouling and process performance of forward osmosis membranes on activated sludge. J. Membr. Sci. 2008, 319 (1-2), 158–168. (6) Mi, B.; Elimelech, M. Chemical and physical aspects of organic fouling of forward osmosis membranes. J. Membr. Sci. 2008, 320 (1-2), 292–302. (7) Tang, C. Y.; She, Q.; Lay, W. C. L.; Wang, R.; Fane, A. G. Coupled Effects of Internal Concentration Polarization and Fouling on Flux Behavior of Forward Osmosis Membranes during Humic Acid Filtration. J. Membr. Sci. 2010, 354, 123–133. (8) Lay, W. C. L.; Chong, T. H.; Tang, C. Y.; Fane, A. G.; Zhang, J.; Liu, Y. Fouling propensity of Forward Osmosis: investigation of the slower flux decline phenomenon. Water Sci. Technol. 2010, 61 (4), 927–936. (9) Loeb, S.; Titelman, L.; Korngold, E.; Freiman, J. Effect of porous support fabric on osmosis through a Loeb-Sourirajan type asymmetric membrane. J. Membr. Sci. 1997, 129 (2), 243–249. (10) Lee, K. L.; Baker, R. W.; Lonsdale, H. K. Membranes for power generation by pressure-retarded osmosis. J. Membr. Sci. 1981, 8 (2), 141–171. (11) Gray, G. T.; McCutcheon, J. R.; Elimelech, M. Internal concentration polarization in forward osmosis: role of membrane orientation. Desalination 2006, 197 (1-3), 1–8. (12) McCutcheon, J. R.; Elimelech, M. Influence of concentrative and dilutive internal concentration polarization on flux behavior in forward osmosis. J. Membr. Sci. 2006, 284 (1-2), 237–247. (13) Wang, R.; Shi, L.; Tang, C. Y.; Chou, S.; Qiu, C.; Fane, A. G. Characterization of In-House Made Composite Forward Osmosis Hollow Fiber Membranes. J. Membr. Sci. 2010, 355, 158– 167. (14) Mi, B.; Elimelech, M. Organic fouling of forward osmosis membranes: Fouling reversibility and cleaning without chemical reagents. J. Membr. Sci. 2010, 348 (1-2), 337–345. (15) Kang, S. T.; Subramani, A.; Hoek, E. M. V.; Deshusses, M. A.; Matsumoto, M. R. Direct observation of biofouling in crossflow microfiltration: Mechanisms of deposition and release. J. Membr. Sci. 2004, 244 (1-2), 151. (16) Wang, S.; Guillen, G.; Hoek, E. M. V. Direct observation of microbial adhesion to membranes. Environ. Sci. Technol. 2005, 39 (17), 6461. (17) Li, H.; Fane, A. G.; Coster, H. G. L.; Vigneswaran, S. Direct observation of particle deposition on the membrane surface during crossflow microfiltration. J. Membr. Sci. 1998, 149 (1), 83–97. (18) Zhang, Y. P.; Fane, A. G.; Law, A. W. K. Critical flux and particle deposition of bidisperse suspensions during crossflow microfiltration. J. Membr. Sci. 2006, 282 (1-2), 189–197. (19) Zhang, Y. P.; Fane, A. G.; Law, A. W. K. Critical flux and particle deposition of fractal flocs during crossflow microfiltration. J. Membr. Sci. 2010, 353 (1-2), 28–35. (20) Li, H.; Fane, A. G.; Coster, H. G. L.; Vigneswaran, S. Observation of deposition and removal behaviour of submicron bacteria on the membrane surface during crossflow microfiltration. J. Membr. Sci. 2003, 217 (1-2), 29–41. (21) Tang, C. Y.; Kwon, Y.-N.; Leckie, J. O. Probing the nano- and micro-scales of reverse osmosis membranessA comprehensive characterization of physiochemical properties of uncoated and coated membranes by XPS, TEM, ATR-FTIR, and streaming potential measurements. J. Membr. Sci. 2007, 287 (1), 146–156.

(22) Tang, C. Y.; Kwon, Y.-N.; Leckie, J. O. Effect of membrane chemistry and coating layer on physiochemical properties of thin film composite polyamide RO and NF membranes. II. Membrane physiochemical properties and their dependence on polyamide and coating layers. Desalination 2009, 242 (1-3), 168–182. (23) Li, H.; Fane, A. G.; Coster, H. G. L.; Vigneswaran, S. An assessment of depolarisation models of crossflow microfiltration by direct observation through the membrane. J. Membr. Sci. 2000, 172 (1-2), 135–147. (24) Elimelech, M.; Zhu, X. H.; Childress, A. E.; Hong, S. K. Role of membrane surface morphology in colloidal fouling of cellulose acetate and composite aromatic polyamide reverse osmosis membranes. J. Membr. Sci. 1997, 127 (1), 101. (25) Tang, C. Y.; Fu, Q. S.; Criddle, C. S.; Leckie, J. O. Effect of flux (transmembrane pressure) and membrane properties on fouling and rejection of reverse osmosis and nanofiltration membranes treating perfluorooctane sulfonate containing wastewater. Environ. Sci. Technol. 2007, 41 (6), 2008–2014.

(26) Bacchin, P.; Aimar, P.; Field, R. W. Critical and sustainable fluxes: Theory, experiments and applications. J. Membr. Sci. 2006, 281 (1-2), 42–69. (27) Tang, C. Y.; Kwon, Y.-N.; Leckie, J. O. Fouling of reverse osmosis and nanofiltration membranes by humic acid -- Effects of solution composition and hydrodynamic conditions. J. Membr. Sci. 2007, 290 (1-2), 86–94. (28) Schwinge, J.; Neal, P. R.; Wiley, D. E.; Fletcher, D. F.; Fane, A. G. Spiral wound modules and spacers review and analysis. J. Membr. Sci. 2004, 242 (1/2), 129. (29) Neal, P. R.; Li, H.; Fane, A. G.; Wiley, D. E. The effect of filament orientation on critical flux and particle deposition in spacerfilled channels. J. Membr. Sci. 2003, 214 (2), 165–178. (30) Chong, T. H.; Wong, F. S.; Fane, A. G. Implications of critical flux and cake enhanced osmotic pressure (CEOP) on colloidal fouling in reverse osmosis: Experimental observations. J. Membr. Sci. 2008, 314 (1-2), 101–111.

ES101966M

VOL. 44, NO. 18, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7109