Thin film composite forward-osmosis membranes with enhanced internal osmotic pressure for internal concentration polarization reduction

Chemical Engineering Journal 249 (2014) 236–245

Contents lists available at ScienceDirect

Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej

Thin film composite forward-osmosis membranes with enhanced internal osmotic pressure for internal concentration polarization reduction Zhengzhong Zhou, Jim Yang Lee ⇑, Tai-Shung Chung Department of Chemical & Biomolecular Engineering, National University of Singapore 10 Kent Ridge Crescent, Singapore 119260, Singapore

h i g h l i g h t s

g r a p h i c a l a b s t r a c t

 FO membranes were prepared by

incorporating SPPO in substrate.  IOP was generated in the SPPO/PSf

membranes.  The effects of IOP were demonstrated

via structural parameter analysis.  ICP was mitigated in the SPPO/PSf

membranes.  High water fluxes were demonstrated

in FO seawater desalination.

a r t i c l e

i n f o

Article history: Received 3 January 2014 Received in revised form 6 March 2014 Accepted 13 March 2014 Available online 24 March 2014 Keywords: Forward osmosis Sulfonated polymeric membrane Desalination Water purification Concentration polarization

a b s t r a c t Thin-film composite (TFC) forward-osmosis (FO) membranes with enhanced internal osmotic pressure (IOP) were used to reduce internal concentration polarization in this study. These TFC membranes contained a selective polyamide layer deposited by interfacial polymerization on a support substrate cast from a polymer blend of polysulfone (PSf) and sulfonated poly(phenylene oxide) (SPPO). The immobilized counter ions (Na+) in SPPO gave rise to an IOP which facilitated water transport in the AL–FS operating mode (i.e., the active layer is facing the feed solution, also referred to as the FO mode) but retarded water transport in the AL–DS operating mode (i.e., the active layer is facing the draw solution, also called as the pressure retarded osmosis (PRO) mode). An optimized TFC membrane could draw a water flux of 39 LMH (Lm2 h1) in the AL–FS mode, which is among the highest in the current literature; and 57 LMH in the AL–DS mode, which is comparable to other published works using deionized water as the feed and 2 M NaCl as the draw solution. The optimized SPPO/PSf TFC membrane also outperformed other published FO membranes in simulated seawater desalination. Extremely high water fluxes of 25 and 19 LMH could be obtained in the AL–DS and AL–FS modes respectively. The impressive high water flux in the AL–FS mode makes this membrane particularly suitable for FO operations where internal concentration polarization (ICP) and membrane fouling are major concerns. Ó 2014 Elsevier B.V. All rights reserved.

1. Introduction Forward osmosis (FO), an osmotic driven membrane process which harvests fresh water from seawater, brackish and municipal ⇑ Corresponding author. Fax: +65 67791936. E-mail address: [email protected] (J.Y. Lee). http://dx.doi.org/10.1016/j.cej.2014.03.049 1385-8947/Ó 2014 Elsevier B.V. All rights reserved.

wastewater, is a potential technological solution to global water shortage where more than 1.2 billion people have no access to clean water [1–6]. In FO desalination, seawater and a concentrated draw solution are separated by a semi-permeable membrane (the FO membrane); the osmotic pressure difference between the two solutions drives water flow from the seawater to the draw solution across the membrane without any externally applied pressure;

Z. Zhou et al. / Chemical Engineering Journal 249 (2014) 236–245

237

Nomenclature Abbreviations AL–DS active layer facing the draw solution AL–FS active layer facing the feed solution DI de-ionized ECP external concentration polarization FO forward osmosis ICP internal concentration polarization IEC ion exchange capacity IOP internal osmotic pressure IP interfacial polymerization MPD M-phenylene diamine PSf polysulfone RO reverse osmosis SPPO sulfonated poly(phenylene oxide) TFC thin-film composite TMC trimesoyl chloride Symbols A Am B C D dh DF f

water permeability coefficient membrane area salt permeability coefficient concentration salt diffusivity hydraulic diameter of FO cell channel free energy of mixing number of ions per chain

the draw solute is then regenerated with the concurrent production of pure water. In comparison with other desalination processes such as reverse osmosis [7–10], membrane distillation [11], capacitive deionization [12] and electrodialysis [13] where high hydraulic pressure, high temperature or electrical energy is needed, FO processes have some notable advantages: no energy input is required to transport water across the membrane in the FO unit, and draw solute regeneration may make use of renewable energy or low grade waste heat if thermo-responsive or photon-responsive draw solutes [2–4,14] are used. One scenario involves the precipitation of the draw solute after moderate heating, lowering the osmotic pressure of the draw solution and hence the hydraulic pressure required in a subsequent filtration process [15]. In addition, membrane fouling in FO can also be managed more easily than in RO [16–18] to result in the reduction of the operating cost. All of these considerations could lead an overall smaller environmental footprint. Apart from the design of novel draw solutes which is essential to the economic viability of FO desalination, the FO membrane design should also be improved to enable process intensification based on productivity (a high water flux) and quality (a low salt leakage rate). One critical issue in FO processes is the diminution of water flux by internal concentration polarization (ICP) [19–21]. ICP is caused by the membrane resistance to diffusion, resulting in a serious dilution of the draw solution in the AL–FS mode or a slight increase in the feed concentration in the AL–DS mode. Consequently, the osmotic pressure difference across the membrane active layer is reduced, as shown in Fig. 1 (solid curves). Unlike external concentration polarization (ECP) which occurs outside the membrane, ICP resides inside the porous support and as such cannot be mitigated by increasing the water flow rate or turbulence. FO membranes have to be designed differently from the RO membranes with ICP mitigation as the key consideration. Recent studies on TFC FO membranes have made notable progress in meeting the FO requirements [22–27]. An important feature of the TFC membranes is

Jv Js k kB L l m N P R Rs Re S Sc Sh T t x V

e p s q t u

water flux salt flux mass transfer coefficient Boltzman constant length of FO cell channel membrane thickness mass Avogadro’s number hydraulic pressure gas constant salt rejection Reynolds number structural parameter Schmidt number Sherwood number temperature time number of chains per unit volume volume membrane porosity osmotic pressure membrane tortuosity density molar volume of solvent volume fraction of polymer network

independent tailorability of the selective and support layers to meet different application demands. For a prospective TFC-FO membrane, the TFC layer, formed by the interfacial polymerization (IP) of an amine and a carboxylic acid or acid chloride, should be designed for high solute rejection and low water resistance by optimizing the IP conditions [28], the substrate surface chemistry [22,29] and post-synthesis treatments [30]. The substrate, formed via phase inversion, should be thin and porous for ICP reduction, and yet has sufficient mechanical, chemical and thermal stability to withstand industrial operations. Research over the years has shown that hydrophilic substrates such as sulfonated polysulfone [31], sulfonated poly (ether ketone) [30] and sPES-co-sPPSf (sulfonated polyethersulfone and polyphenylsulfone copolymer) [23,32] are more capable of high water flux. Substrate hydrophilicity aside, a factor that has previously been overlooked is the effect of cations (e.g. Na+) associated with the sulfonated material in the membrane. In an early study on polymer gel swelling, Flory equated the swelling pressure to the net osmotic pressure of the polymer gel [33], which Amiya and Tanaka described as IOP [34]. IOP is significantly higher in ionic than non-ionic polymer networks as a result of charge localization [33] and the translational degree of freedom of counterions [34] in a polymer network. Hence IOP is likely to be universal in polymeric membranes containing ionizable groups. In osmotic pressure membrane processes such as FO, however, the effect of IOP has thus far not been included in the analysis of the FO performance. If the substrate of an FO-TFC membrane contains an ion-exchange polymer, IOP can be generated by the counterions immobilized in the substrate, which increases the effective driving force in the AL–FS mode and decreases the effective driving force in the AL–DS mode (the dotted curves in Fig. 1); resulting in a higher water flux for the former and a lower water flux for the latter. It should be mentioned that the AL–FS mode is the preferred operating mode for desalination because feed water foulants are deposited on the TFC layer; and are much easier to remove than

238

Z. Zhou et al. / Chemical Engineering Journal 249 (2014) 236–245

Fig. 1. Concentration profiles of draw solute in a typical TFC membrane with (dotted lines) and without (solid lines) an IOP. Subscripts 1, 2, m and s refer to the concentration on the feed side, the draw solution side, active layer facing the outer surface and active layer facing the substrate, respectively. Superscript ‘O’ refers to the concentrations in the presence of IOP. The shaded areas are the active layers.

foulants deposited in the membrane supporting layer by an AL–DS operation [35]. The IOP effect would be more pronounced if a higher concentration of counterions (e.g. Na+) can be accommodated in the membrane substrate by using, for example, sulfonated polymers with a sufficiently high ion exchange capacity (IEC). Polysulfone (PSf), however, cannot be extensively sulfonated without some loss of its membrane formation properties. Our previous work on fuel cell proton exchange membranes suggests that sulfonated poly(phenylene oxide) (SPPO) may provide a solution. SPPO is a low cost membrane forming material with good thermal and chemical stability [36], high water uptake [37] and water permeability [38–39], and the highest IEC value (meq/g) among polymers with the same degree of sulfonation. The TFC substrate in this study was therefore fabricated by blending PSf with a SPPO with an IEC value of 2.8. A comparison of water flux and salt rejection properties with the literature in both AL–FS and AL–DS operating modes, and the analysis of membrane structural parameters, confirmed not only the presence of IOP, but also the benefits of incorporating a high IEC polymer in the support layer in FO applications.

solution was a 15 wt% solution of PSf-SPPO (with 0%, 25% and 50% SPPO content) in NMP, which had been degased overnight at room temperature. The solution was evenly cast onto a clean glass plate and thinned by a 100 lm casting knife. The glass plate was then immersed in a DI water coagulation bath where the polymers were precipitated by phase inversion into a thin membrane. The membrane substrate was then washed and stored in DI water and ready for modification. 2.4. Ion exchange capacity (IEC) measurements The IEC values of the membranes were measured by the back titration method. Specifically a SPPO/PSf blend membrane was first converted to the H+ form by acidification with dilute HCl. The membrane after thorough washing with DI water was equilibrated in 25 ml 0.01 M NaOH for 24 h. The membrane was then removed and washed. The NaOH and wash solutions were titrated against 0.01 M HCl. The IEC value was calculated by Eq. (1), where CHCl and VHCl are the concentration and volume of HCl; with superscript ‘o’ representing the corresponding quantities for neutralizing 25 ml 0.01 M NaOH. mdry is the mass of the dried membrane:

2. Experimental 2.1. Materials Poly (phenylene oxide) (PPO) from Sigma Aldrich, chlorosulfonic acid from Merck, and UdelÒ polysulfone (PSf) from Solvay were used for the preparation of the membrane substrate. Solvents N-methyl-2-pyrrolidone (NMP), n-hexane, chloroform and methanol; and inorganic chemicals sodium chloride and sodium hydroxide were supplied by Merck. Magnesium chloride was purchased from Sigma Aldrich. M-Phenylenediamine (MPD) and trimesoyl chloride (TMS) monomers used for IP were supplied by Alfa Aesar. All chemicals were used as received. 2.2. Sulfonation of PPO PPO was sulfonated by the procedure of Fu et al. [40]. Specifically 10 g PPO powder was dissolved in chloroform to a 20 wt% solution and cooled in an ice bath. 2.5 ml refrigerated chlorosulfonic acid was added slowly dropwise to the PPO solution under vigorous stirring. The reaction mixture was allowed to stand for 30 min at the end of addition before it was quenched and precipitated with methanol. The recovered solid product after washing with DI water was vacuum dried and stored for further use. 2.3. Preparation of SPPO/PSf membrane substrate The preparation of the membrane substrate was based on the classical Loeb–Sourirajan phase inversion technique. The casting

IEC ¼

C oHCl V oHCl  C HCl V HCl mdry

ð1Þ

2.5. Fabrication of TFC membranes A thin selective polyamide layer was formed on the membrane surface by the IP of MPD and TMC monomers (Fig. 2). Specifically a membrane substrate was first equilibrated in a 2 wt% MPD aqueous solution for 1 min. The excess solution was removed by absorption with a few pieces of dry filter papers. The membrane was then affixed to a plastic frame. The top surface was brought into contact with a 0.1 wt% TMC solution in hexane for a predetermined period of time. Thereafter the membrane was dried in air for 5 min, and then washed with and stored in DI water before FO measurements. The following convention was used to name the fabricated membranes: SPPO/PSf (composition)-IP time. The IP time was determined by the TMC exposure time due to the way IP was conducted. Hence a membrane with 25 wt% SPPO and 30 s of IP was named as SPPO/PSf (25:75)-30. 2.6. Forward osmosis performance tests The measurements of water flux and salt reverse flux through the TFC FO membranes made use of the lab-scale filtration unit shown in Fig. 3. The effective membrane area for the test was 2 cm2. Feed and draw solutions flowed co-currently through the rectangular channels partitioned by the FO membrane at the flow

239

Z. Zhou et al. / Chemical Engineering Journal 249 (2014) 236–245

Fig. 2. Interfacial polymerization of m-phenylene diamine (MPD) with trimesoyl chloride (TMC).

filtration cell and a membrane area of 9.62 cm2. Pure water fluxes which permeated through the TFC membranes under 2–5 bars of hydraulic pressure were recorded and the water permeability coefficients (A) of the membranes were calculated from the gradient of the plot of water flux (Jv) against applied pressure (P) (Eq. (4)). A 1000 ppm NaCl aqueous solution was then permeated through the membrane at 5 bars. The salt rejection (Rs) and salt permeability coefficients (B) of the membranes were calculated from Eqs. (5) and (6) [26] respectively. In these equations Cp, Cf, DP and Dp are respectively, salt concentrations in the permeate and feed solution, the applied hydraulic pressure, and the osmotic pressure of the NaCl solution.

A¼ Fig. 3. Schematic of the FO test unit. (1) circulating pump; (2) flow meter; (3) draw solution reservoir; (4) feed solution container; (5) weighing balance; (6) computer; (7) FO membrane and (8) flow chamber of FO test unit.

rate of 0.1 L/min. The FO performance was evaluated in two operating modes, namely the AL–DS mode where the active TFC layer faced the draw solution, and the AL–FS mode where the active layer faced the feed solution. A data-logging digital balance (EK-4100i, A&D company Ltd., Japan) was used to record the weight changes in the feed solution and a conductivity meter (Lab 960, SI analytics GmbH, Germany) measured the salt concentration changes in the feed solution. All measurements were carried out at room temperature (295 K) in at least two replicates and the average values were reported. FO performance was evaluated from data collected over a 15 min period when the rate of decrease of the feed solution weight was constant. The water flux, Jv and the salt reverse flux, Js were calculated from Eqs. (2) and (3) below where V, C, t and Am are the volume and salt concentration of the feed solution, the measurement time, and the effective membrane area respectively.

Jv ¼

DV Dt  A m

DðC t V t Þ Js ¼ Dt  A m

ð2Þ

Rs ¼ 1  B¼

ð4Þ Cp Cf

ð5Þ

Rs Að1  RÞðDP  DpÞ

ð6Þ

The membrane structural parameters (S) were determined from the FO water flux using Eq. (7) (for AL–FS measurements) and Eq. (8) (for AL–DS measurements) [41,42]:

    9 8 < pD exp  JvDS  pf exp Jkv = h    i Jv ¼ A :1 þ B exp Jv  exp  Jv S ; Jv

k

Jv

D

ð7Þ

D

    9 8 < pD exp  Jkv  pf exp JvDS = h    i Jv ¼ A :1 þ B exp Jv S  exp  Jv ;

ð8Þ

k

D, the salt diffusivity and k, the mass transfer coefficient in these equations were calculated in turn from Eqs. (9) and (10) where Sh is the Sherwood number of laminar flow in a rectangular channel, Re, Sc, dh and L are the Reynolds number, the Schmidt number, the hydraulic diameter and the channel length, respectively:

k¼ ð3Þ

DJv DP

D  Sh dh

  dh Sh ¼ 1:85 Re  Sc L

ð9Þ ð10Þ

2.7. Determination of membrane intrinsic properties 2.8. Membrane characterization Following standard practice, the water and salt permeability coefficients of the membranes (A and B in Eqs. (4) and (6) respectively) were measured in the RO mode using a dead-end

Membrane morphology was examined by scanning electron microscopy (SEM) and field emission scanning election microscopy

240

Z. Zhou et al. / Chemical Engineering Journal 249 (2014) 236–245

(FESEM) on a JEOL-JSM-5600LV and a JEOL JSM-6700F microscope respectively. Membrane samples for microscopic examination were freeze-dried in a freeze drier (ModulyoD, Thermo Electron Corporation, USA), fractured in liquid nitrogen and then coated with platinum using a sputter coater (JEOL LFC-1300). Membrane hydrophilicity was evaluated by water contact angle measurements on a contact angle goniometer (Tame Hart, USA). Measurements were carried out over ten random locations of the membrane using DI water as the probe liquid. The average values were reported. Membrane porosity was calculated from the difference between a membrane in the wet and dry states using Eq. (11) where m1, m2, qw and qp are the weights of wet and dry membranes, and the densities of water and the polymer blend, respectively: m1 m2

e ¼ m2 qP

qw

þ m1qm2

ð11Þ

w

Membrane tortuosity was calculated from Eq. (12) where l is the membrane thickness:

s¼

eS l

ð12Þ

mixing, the elastic deformation of the gel; and the mixing of the mobile ions with water, respectively. Since membrane swelling by water is insignificant in a PSf membrane, u/u0 approaches unity. Since the osmotic pressure of the counterions is much higher than the polymer [43], this equation can, as a first order approximation, be reduced to the last term on the right, which is the same as the Van’t Hoff Eq. (14), where i, c and R represent the Van’t Hoff coefficient, solute concentration and universal gas constant, respectively. Hence the IOP may be estimated by the mobile Na+ concentration, cm, in the substrate, which could be calculated from Eq. (15) using the mass (m), the SPPO content (SPPO %) and the volume (Vm) of a dry membrane. For the SPPO/PSf (50:50) membrane where the IEC of SPPO was 2.8, the calculated Na+ concentration was 0.047 M and the corresponding osmotic pressure was 2.3 bars based on the Van’t Hoff equation. Although this value is small compared with the osmotic pressure of the draw solution (99 bars for 2 M NaCl), it was less vulnerable to ICP since they were partially immobilized and as non-diffusive species, were not subjected to diffusion limitations. Based on the assumption of no dilution effect, 2.3 bar of persistent osmotic pressure could generate a water flux of 8 LMH (A  Dp) and account for 10–20% of the water flux in common FO operations.

 1 DF 2 u 2 kB T m "  1=3 #   1 u u u þ xfkB T  þ xkB T u0 u0 2 u0

p¼ 3. Results and discussion 3.1. Investigating the effect of IOP on FO performance 3.1.1. Theoretical basis The FO performance can be influenced by the IOP generated by the counter cations (Na+) immobilized by SPPO in the membrane support layer. It could compensate, to some extent, the decrease in osmotic pressure difference caused by dilutive ICP in the AL–FS operating mode (Fig. 1). This is because an increased Na+ concentration in the membrane substrate increased the osmotic pressure difference to drive water transport across the selective layer. It should be noted that these Na+ ions were partially immobilized and as such were resistant to dilution by the permeated water (the cause of dilutive ICP). The negligible water swelling of the hydrophobic PSf membrane backbone also dismisses a possible cause of dilution by membrane swelling. Hence one could be assured of an increase in water flux through the membrane in the Al–FS mode with higher IOP. The beneficial IOP in the AL–FS operating mode could however retard water permeation in the AL–DS mode because the increase in salt concentration in the support layer compounded the effect of concentrative ICP, as shown in Fig. 1. The osmotic pressure difference across the selective layer was reduced by the IOP in the support layer; which could lead to water flux decrease relative to the case without the IOP if all other conditions (membrane structure, porosity and hydrophilicity etc.) were kept the same. However, since the IOP was created by incorporating SPPO in the membrane substrate, other membrane properties contributing to FO performance such as hydrophilicity would also be changed simultaneously. Thus a simple comparison of FO water fluxes alone might not be sufficient to evaluate the effect of IOP; a more detailed analysis as shown in the following sections is required. Unlike solution osmotic pressure, the direct measurement and quantification of IOP in the membrane substrate are not possible with current techniques. Eq. (13) shows the Flory–Huggins equation for the net osmotic pressure of an ionic gel, where N, kB, T, t, u, DF, x and f are the Avogadro’s number, Boltzman constant, temperature, molar volume of solvent, volume fraction of polymer network, free energy of mixing, number of chains per unit volume and number of ions per chain, respectively. The three terms on the right side of the equation represent the contributions from polymer

NkB T



u þ lnð1  uÞ þ

p ¼ icRT Cm ¼

m  IEC  SPPO% Vm

ð13Þ ð14Þ ð15Þ

The FO performance of membranes with different SPPO contents was measured to determine the effect of IOP. Since the selective layer is critical to FO performance, the IP conditions for fabricating the selective layer were optimized first prior to the investigation. 3.1.2. Optimization of IP conditions Membranes with different SPPO contents were fabricated and converted to TFC membranes by the IP of 2 wt% MPD and 0.1 wt% TMC. The ridge and valley morphology in Fig. 4(a) and (b) was typical of a membrane after IP and indicates the formation of a polyamide layer. Since MPD was used in excess, the extent of IP and consequently the TFC layer thickness were determined by the TMC exposure time (Fig. 4(c) and (f)). Fig. 5 shows the FO performance of SPPO/PSf (50:50) membranes treated with IP for various lengths of time. The highest water flux was registered with an IP time of 45 s. The polyamide layers formed with a very short IP time (e.g. 15 s) contained a significant amount of structural defects which led to a high salt leakage. Water flux decreased accordingly because of the reduction in osmotic pressure difference across the membrane selective layer [23]. Conversely an unduly long IP time (>45 s) produced a very thick selective layer which increased the transport resistance of water as well as salt. Both the water flux and the salt reverse flux decreased as a result. The IP time for the best balance of water permeation and salt rejection properties in this study was determined to be 45 s. 3.1.3. FO performance of SPPO/PSf membranes The FO performance of three membranes with different SPPO contents (and 45 s of IP) is compared in Fig. 6. The water flux in both AL–FS and AL–DS operating modes increased noticeably with the membrane SPPO content. As IOP increased linearly with SPPO content, the water flux enhancing effect of IOP could be demon-

Z. Zhou et al. / Chemical Engineering Journal 249 (2014) 236–245

241

Fig. 4. FESEM images of the top surface of SPPO/PSf (50:50)-45 at (a) low and (b) high magnifications; and cross-sectional views of (c) SPPO/PSf (50:50)-15; (d) SPPO/PSf (50:50)-30; (d) SPPO/PSf (50:50)-60 and (f) SPPO/PSf (50:50)-90; the numbers in the micrographs are the estimated TFC layer thicknesses.

Fig. 5. FO performance of TFC SPPO/PSf (50:50) membranes formed with different IP times. Draw solution: 2 M NaCl solution; feed solution: pure water.

strated in the AL-–FS mode, however, the opposing IOP effect in the AL–DS mode was eclipsed by the favorable effects of other factors (e.g. greater porosity and increased membrane hydrophilicity), which could contribute more to the increase in water flux with membrane SPPO content in both operating modes. The increase in membrane hydrophilicity after compounding PSf with highly hydrophilic SPPO was clearly one of the factors. This was confirmed by contact angle measurements which showed a steady decrease in the water contact angle with SPPO content (Table 1). The increased hydrophilicity and high ion-exchange capability (imparted by SPPO) of the membranes also facilitated salt absorption and transport in the membrane support layer. Consequently the dilutive ICP in the AL–FS mode and the concentrative ICP in the AL–DS mode could be counteracted by the faster diffusion of salt into and away from the substrate-TFC interface [44]. Fig. 7 shows the increase in the number of larger pores in the membrane bottom layer with SPPO content while the membrane top

(A)

(B)

(C)

(D)

Fig. 6. FO performance of membranes with varying SPPO/PSf composition using DI water as feed and various concentrations of NaCl as draw solutions. Figs. A and B are water and salt fluxes in the Al–DS mode; Fig. C and D are water and salt fluxes in the AL–FS mode.

242

Z. Zhou et al. / Chemical Engineering Journal 249 (2014) 236–245

Table 1 Contact angle and porosity of membrane substrates.

Table 2 Intrinsic properties of the SPPO/PSf membranes.

Membranes substrate

Contact angle

e

Membrane

Mode

A

B

S

T

SPPO/PSf (50:50) SPPO/PSf (25:75) SPPO/PSf (0:100)

60 ± 2 63 ± 2 73 ± 1

0.86 0.84 0.80

SPPO/PSf (50:50) – 45 s

AL–DS AL–FS AL–DS AL–FS AL–DS AL–FS

3.55 3.55 3.22 3.22 3.29 3.29

0.74 0.74 0.95 0.95 0.285 0.285

381 ± 98 293 ± 22 768 ± 202 562 ± 149 3804 ± 660 3680 ± 431

5.9 4.5 10 7.6 61 59

layer was less affected. A more open bottom layer not only accelerated water transport across the membrane/solution interface, but also (and perhaps more importantly) the transport of salt from the draw solution to the membrane substrate (in the AL–FS mode) or from the membrane substrate to the feed solution (in the AL–DS mode). Membrane porosity also increased noticeably with the SPPO content (Table 1). The resistance to water diffusion in the support layer was clearly lower in high porosity membranes. It should be noted that the transport of solute across the entire membrane is not strongly affected by the support porosity but by the free volume of the active layer. However, a more porous support does allow a faster diffusion of salt in the membrane support layer to lessen the ICP effect. A higher water flux is therefore possible. Hence the existence and effects of IOP could not be ascertained from water flux measurements alone. We show that the analysis of membrane structural parameters (S) in AL–FS and AL–DS modes (Table 2) could establish the existence of IOP. Structural parameter is a measure of the membrane ICP (low S values for low ICP) and is an intrinsic property of the membrane independent of the mode of operation. Eq. (7) (for the Al–FS mode) and Eq. (8) (for the AL–DS mode) have been derived without the supposition of IOP; and as such should yield the same S value only in the absence of the IOP. This was indeed the case for the SPPO-free membrane – the two calculated S values were about equal and large; indicating a large ICP. For the SPPO-containing membranes, the S values calculated from the AL–FS data were systematically smaller than the S values calculated from the AL–DS data. The different S values and their trend in the two operating modes could however be justified by the existence of an IOP which assisted the AL–FS operation (smaller S) but hindered the AL–DS operation (larger S). Moreover, if the estimated IOP (2.3 bar) was taken to be similar to the hydraulic pressure in pressure retarded osmosis (PRO);

SPPO/PSf (25:75) – 45 s SPPO/PSf (0:100) – 45 s

Table 3 Forward osmosis data of SPPO/PSf membranes using a DI water feed and a 1 M MgCl2 draw solution. Membrane

Mode

Water flux

Reverse salt flux

SPPO/PSf (50:50) – 45 s

AL–DS AL–FS AL–DS AL–FS AL–DS AL–FS

36 ± 4 29 ± 3 23 ± 5 16 ± 2 7±1 5±1

2.8 ± 0.4 1.4 ± 0.2 2.1 ± 0.3 1.1 ± 0.1 1.1 ± 0.1 1.0 ± 0.2

SPPO/PSf (25:75) – 45 s SPPO/PSf (0:100) – 45 s

and used to offset the p values in Eqs. (7) and (8); similar S values were calculated from the two operating modes (336 ± 42 and 337 ± 90 for the AL–FS and AL–DS modes, respectively). This may be taken as another demonstration of the existence of IOP and its impacts on FO processes. FO tests were also conducted using a 1 M MgCl2 draw solution. The results were similar to the use of a NaCl draw solution. Table 3 shows an increasing trend of water flux with increased SPPO content in the membrane substrate. The salt reverse flux was lower because of the larger ionic size and lower diffusivity of MgCl2. The S parameters were also calculated and tabulated in Table 4, from which two conclusions could be made. First, the decrease in the S parameter with the increase in SPPO content indicates the effectiveness of SPPO/PSf membranes in lowering the ICP. Second, the lower S values in the AL–FS mode again demonstrates the positive impact of IOP in the AL–FS mode and the negative impact of IOP in the AL–DS mode in a FO operation.

Fig. 7. The SEM observation of A) SPPO/PSf (0:100); B) SPPO/PSf (25:75) and C) SPPO/PSf (50:50). T, X and B represent top, cross-section and bottom layer, respectively.

Z. Zhou et al. / Chemical Engineering Journal 249 (2014) 236–245

243

Table 4 Intrinsic properties of the SPPO/PSf membranes measured from the MgCl2 draw solutions. Membrane

Mode

A

B

S

SPPO/PSf (50:50) – 45 s

AL–DS AL–FS AL–DS AL–FS AL–DS AL–FS

3.55 3.55 3.22 3.22 3.29 3.29

0.33 0.33 0.51 0.51 0.25 0.25

342 ± 71 218 ± 32 622 ± 160 477 ± 80 2277 ± 347 2151 ± 538

SPPO/PSf (25:75) – 45 s SPPO/PSf (0:100) – 45 s

In addition to demonstrating the IOP effect, the SPPO-containing membranes also showed great potential in FO applications. Fig. 6 showed the linearity in the increase of water flux with draw solute concentration, demonstrating the capability of SPPO in ICP mitigation. Without the SPPO in the membrane support layer (e.g. SPPO/PSf (0:100)-45) the increase in water flux would level off after sufficiently high draw solute concentrations due to the domination of ICP in the substrate. The higher water flux and sharper increase in water flux with draw solute concentration in the AL–DS mode are common to FO measurements using DI water as the feed. This is because ICP in AL–DS (if any) is caused by salt leakage. The resulting concentrative ICP is much milder than the dilutive ICP in AL–FS if the salt reverse flux is low. 3.2. Benchmarking the FO performance of SPPO/PSf TFC membranes 3.2.1. FO test with pure water feed The measurements in Table 5 were obtained with a DI water feed and a 2 M NaCl draw solution. Literature data corresponding to the same operating conditions was also used for comparison. The SPPO/PSf (50:50)-45 TFC membrane in this study performed well when benchmarked against other TFC members in the literature – it surpassed the performance of other membranes in the AL–FS operating mode, and was on par with the other membranes in the AL–DS mode. From earlier discussion the higher water flux in the AL–FS mode could be attributed to ICP reduction by the combined effect of increased membrane hydrophilicity and IOP. A higher driving force was therefore available to compensate for some of the losses due to dilutive ICP. In the AL–DS operating mode, however, the performance of the SPPO/PSf TFC membrane was edged by PES-TFC hollow fibers. The insignificant concentrative ICP in AL–DS using a fresh water feed (assuming no salt leakage); and the low ECP in hollow fiber operations, could be the reasons. The PES-TFC hollow fibers were however vulnerable to dilutive ICP resulting in low water fluxes in the AL–FS mode and in simulated seawater desalination (Fig. 8). This shows that SPPO/PSf TFC membranes are more effective than PES-TFC hollow fibers in ICP mitigation. It should be emphasized that AL–FS is

Table 5 Comparison of the FO performance of SPPO/PSf (50:50) membranes with other membranes in the literature. The feed was DI water and the draw solution was 2 M NaCl for all membrane tests. Membrane

AL–FS Water/ LMH

SPPO/PSf (50:50)-45 PES–TFC hollow fiber [6,24] HTI flat sheet membranes [45,46] Cellulose acetate [5] TFC-PESU-co-sPPSU [23] SPEK-PSf-TFC [30] TFN 0.5 [47]

AL–DS Salt/ gMH

Water/ LMH

Salt/gMH

39 6.1 57 29.5/34.5 2.6/9.87 68/65.1 13.0 10.5 11.2

6.6 5.8/12.34 11

6.5 21 35 29.37

7 2.8 9 14.1

5.6 2.2 7 7.3

9 33 50 56.3

Fig. 8. Performance comparison of different membrane systems (cellulose acetate [5], hollow fiber TFC [6], SPEK TFC [24], PES TFC [30], PES/SPSf TFC [31], SPSf TFC [32], commercial HTI [45] and CAP TFC [48]) for seawater desalination. Operating conditions for the tests: 0.6 M NaCl feed solution (simulated seawater) and 2 M NaCl draw solution. Measurements from the AL–DS mode are shown by the bars on the left and measurements from the AL–FS mode are on the bars on the right (some AL–FS data is missing from the literature).

the preferred operating mode for seawater desalination from the membrane maintenance perspective. In the AL–FS mode, the seawater foulants deposited on the membrane TFC layer could be easily removed by backwashing. In the AL–DS mode, the foulants could penetrate deep into the membrane substrate, necessitating more intense backwashing for their removal. Hence, SPPO/PSf TFC membranes with a high water flux in the AL–FS mode are more practical for seawater desalination. 3.2.2. Simulated seawater desalination The potential application of SPPO/PSf TFC membranes for seawater desalination was evaluated using 3.5 wt.% NaCl solution (0.6 M) as simulated seawater and 2 M NaCl as the draw solution. The measured water flux was 19 ± 1 LMH in the AL–FS mode and 25 ± 1 LMH in the AL–DS mode for the SPPO/PSf (50:50)-45 membrane. These values were considerably lower than the values obtained from a DI water feed even though the overall osmotic pressure difference was kept the same (i.e. DI water against 1.5 M NaCl draw solution). This is because a feed solution with osmotic pressure induced concentrative ICP in the AL–DS mode and concentrative ECP (on the feed side) in the AL–FS mode; thereby reducing the osmotic gradient across the selective layer. Nevertheless Fig. 8 shows that the SPPO/PSf (50:50)-45 membrane still surpassed the performance of other membranes in the literature under the same conditions in both AL–FS and AL–DS modes (some AL–FS data was missing in the literature). The loss of osmotic gradient when DI water was substituted by simulated seawater apparently affected the SPPO-containing TFC membranes less than the other membranes and this could be attributed to the effectiveness of the former in ICP reduction through increased hydrophilicity and IOP. It is worthwhile to note that the water flux of the SPPO/PSf (50:50)-45 membrane in the AL–FS mode was higher than most of the other membranes operating in the AL–DS mode. Hence the membranes in this study could be operated in the AL–FS mode with all the benefits of low membrane fouling [35]; and yet provide a comparably high water flux for desalination currently only possible for the AL–DS operations. FO measurements were also conducted with a real seawater feed. The seawater was sampled from the west coast of Singapore and pre-filtered with a PYREX filter funnel. The salt concentration was estimated by conductivity measurements to be 0.56 M. The water flux in the AL–DS and AL–FS modes was 24 ± 1 LMH and

244

Z. Zhou et al. / Chemical Engineering Journal 249 (2014) 236–245

20 ± 1 LMH, respectively, which are very close to the values obtained from the simulated seawater feed. Hence the results confirmed that the optimized SPPO/PSf membrane is applicable for real seawater FO desalination. 4. Conclusion A forward osmosis membrane with charged polymers in the substrate could generate an IOP that compensates for some osmotic gradient losses due to dilutive ICP to increase the water flux in AL–FS operations. In AL–DS operations, however, the IOP could intensify concentrative ICP but the negative effect is moderated by other ICP reduction mechanisms, e.g. increased membrane hydrophilicity imparted by the charged polymers. The optimized SPPO/PSf TFC membrane in this study with a polyamide selective layer could provide a water flux of 39 LMH in the AL–FS mode (among the highest in the literature) and 57 LMH in the AL–DS mode (on par with literature values) in forward osmosis using DI water feed. The results were even more impressive in seawater desalination, where water flux as high as 25 LMH in the AL–DS and 19 LMH in the AL–FS mode (both representing the best in the literature at this point of writing) could be drawn from simulated seawater (3.5 wt% NaCl solution). Hence, IOP is a factor to consider in the mitigation of ICP in FO membranes. Membranes with high IOP could be promising candidates for desalination in the AL–FS operation mode. Acknowledgements The authors would like to thank the Singapore National Research Foundation for funding this study through the project ‘‘Advanced FO membranes and membrane systems for wastewater treatment, water reuse and seawater desalination’’ (R-279-000336-281). Thanks are also extended to Prof. Donald Paul, Dr. Natalia Widjojo and Dr. Xue Li for their comments and suggestions. References [1] M.A. Montgomery, M. Elimelech, Water and sanitation in developing countries: including health in the equation, Environ. Sci. Technol. 41 (2007) 17–24. [2] T.S. Chung, S. Zhang, K.Y. Wang, J.C. Su, M.M. Ling, Forward osmosis processes: yesterday, today and tomorrow, Desalination 287 (2012) 78–81. [3] S. Zhao, L. Zou, C.Y. Tang, D. Mulcahy, Recent developments in forward osmosis: opportunities and challenges, J. Membr. Sci. 396 (2012) 1–21. [4] J.J. Qin, W.C.L. Lay, K.A. Kekre, Recent developments and future challenges of forward osmosis desalination: a review, Desalination Water Treat. 39 (2012) 123–136. [5] S. Zhang, K.Y. Wang, T.S. Chung, H.M. Chen, Y.C. Jean, G. Amy, Wellconstructed cellulose acetate membranes for forward osmosis: minimized internal concentration polarization with an ultra-thin selective layer, J. Membr. Sci. 360 (2010) 522–535. [6] S. Chou, L. Shi, R. Wang, C.Y. Tang, C.Q. Qiu, A.G. Fane, Characteristics, potential applications of a novel forward osmosis hollow fiber membrane, Desalination 261 (2010) 365–372. [7] R.F. Service, Desalination freshens up, Science 313 (2006) 1088–1090. [8] M.A. Shannon, P.W. Bohn, M. Elimelech, J.G. Georgiadis, B.J. Marinas, A.M. Mayes, Science and technology for water purification in the coming decades, Nature 452 (2008) 301–310. [9] D. Li, H. Wang, Recent developments in reverse osmosis desalination membranes, J. Mater. Chem. 20 (2010) 4551–4566. [10] T. Kaghazchi, M. Mehri, M.T. Ravanchi, A. Kargari, A mathematical modeling of two industrial seawater desalination plants in the Persian Gulf region, Desalination 252 (2010) 135–242. [11] M.M.A. Shirazi, A. Kargari, M. Tabatabaei, Evaluation of commercial PTFE membranes in desalination by direct contact membrane distillation, Chem. Eng. Process 76 (2014) 16–25. [12] J.-B. Lee, K.-K. Park, H.-M. Eum, C.-W. Lee, Desalination of a thermal power plant wastewater by membrane capacitive deionization, Desalination 196 (2006) 125–134. [13] H.M.N. AlMadani, Water desalination by solar powered electrodialysis process, Renew. Energy 28 (2003) 1915–1924.

[14] L. Chekli, S. Phuntsho, H.K. Shon, S. Vigneswaran, J. Kandasamy, A. Chanan, A review of draw solutes in forward osmosis process and their use in modern applications, Desalin. Water Treat. 43 (2012) 167–184. [15] M.M. Ling, T.S. Chung, X. Lu, Facile synthesis of thermosensitive magnetic nanoparticles as smart draw solutes in forward osmosis, Chem. Commun. 47 (2011) 10788–10790. [16] B.X. Mi, M. Elimelech, Organic fouling of forward osmosis membranes: fouling reversibility and cleaning without chemical reagents, J. Membr. Sci. 348 (2010) 337–345. [17] B.X. Mi, M. Elimelech, Silica scaling and scaling reversibility in forward osmosis, Desalination 312 (2013) 75–81. [18] Z.Y. Li, V.Y. Quintanilla, R.V. Linares, Q.Y. Li, T. Zhan, G. Amy, Flux patterns and membrane fouling propensity during desalination of seawater by forward osmosis, Water Res. 46 (2012) 195–204. [19] R.E. Kravath, J.A. Davis, Desalination of seawater by direct osmosis, Desalination 16 (1975) 151–155. [20] J.R. McCutcheon, M. Elimelech, Influence of concentrative and dilutive internal concentration polarization on flux behavior in forward osmosis, J. Membr. Sci. 284 (2006) 237–247. [21] S. Phuntsho, S. Sahebi, T. Majeed, F. Lotfi, J.E. Kim, H.K. Shon, Assessing the major factors affecting the perfromances of forward osmosis and its applications on the desalination process, Che. Eng. J. 231 (2013) 484–496. [22] G. Han, S. Zhang, X. Li, N. Widjojo, T.S. Chung, Thin film composite forward osmosis membranes based on polydopamine modified polysulfone substrates with enhancements in both water flux and salt rejection, Chem. Eng. Sci. 80 (2012) 219–231. [23] N. Widjojo, T.S. Chung, M. Weber, C. Maletzko, V. Warzelhan, The role of sulphonated polymer and macrovoid-free structure in the support layer for thin-film composite (TFC) forward osmosis (FO) membranes, J. Membr. Sci. 383 (2011) 214–223. [24] P. Sukitpaneenit, T.S. Chung, High performance thin-film composite forward osmosis hollow fiber membranes with macrovoid-free and highly porous structure for sustainable water production, Environ. Sci. Technol. 46 (2012) 7358–7365. [25] 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) 3812–3818. [26] 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. 385 (2011) 10–19. [27] R. Wang, L. Shi, C.Y. Tang, S.R. Chou, C.Q. Qiu, A.G. Fane, Characterization of novel forward osmosis hollow fiber membranes, J. Membr. Sci. 355 (2010) 158–167. [28] P.W. Morgan, S.L. Kwolek, Interfacial polycondensation 2. Fundamentals of polymer formation at liquid interfaces, J. Polym. Sci. A1 (34) (1996) 531–559. [29] H.I. Kim, S.S. Kim, Plasma treatment of polypropylene and polysulfone supports for thin film composite reverse osmosis membrane, J. Membr. Sci. 286 (2006) 193–201. [30] G. Han, T.S. Chung, M. Toriida, S. Tamai, Thin-film composite forward osmosis membranes with novel hydrophilic supports for desalination, J. Membr. Sci. 423 (2012) 543–555. [31] K.Y. Wang, T.S. Chung, G. Amy, Developing thin-film-composite forward osmosis membranes based on the PES/SPSf substrate through interfacial polymerization, AIChE J. 58 (2012) 770–781. [32] N. Widjojo, T.S. Chung, M. Weber, C. Maletzko, V. Warzelhan, A sulfonated polyphenylenesulfone (sPPSU) as the supporting substrate in thin-film composite (TFC) membranes with enhanced performance for forward osmosis (FO), Che. Eng. J. 220 (2013) 15–23. [33] P.J. Flory, Principles of Polymer Chemistry, Cornell University Press, 1953. [34] T. Amiya, T. Tanaka, Phase transition in cross-linked gels of natural polymers, Macromolecules 20 (1987) 1162–1164. [35] S. Zhao, L. Zou, D. Mulcahy, Effects of membrane orientation on process performance in forward osmosis applications, J. Membr. Sci. 382 (2011) 308– 315. [36] G. Chowdhury, B. Kruczek, T. Matsuura, Polyphenylene oxide and modified polyphenylene oxide membranes: Gas, Vapor and Liquid Separation, Springer, Boston, 2001. [37] R.Q. Fu, D. Julius, L. Hong, J.Y. Lee, PPO-based acid-base polymer blend membranes for direct methanol fuel cells, J. Membr. Sci 322 (2008) 331–338. [38] T.M. Geeta, Chowdhury, S. Sourirajan, A study of reverse osmosis separation and permeation rate for sulfonated poly(2,6-dimethyl-1,4-phenylene oxide) membranes in different cationic forms, J. Appl. Polym. Sci. 51 (1994) 1071. [39] H. Fu, L. Jia, J. Xu, Studies on the sulfonation of poly (phenylene oxide) (PPO) and permeation behavior of gases and water vapor through sulfonated PPO membranes. II. Permeation behavior of gases and water vapor through sulfonated PPO membranes, J. Appl. Polym. Sci. 51 (1994) 1405–1409. [40] H. Fu, L. Jia, J. Xu, Studies on the sulfonation of poly (phenylene oxide) (PPO) and permeation behavior of gases and water vapor through sulfonated PPO membranes. I. Sulfonation of PPO and characterization of the products, J. Appl. Polym. Sci. 51 (1994) 1399–1404. [41] A. Tiraferri, N.Y. Yip, A.P. Straub, S.R.-V. Castrillon, M. Elimelech, A method for the simultaneous determination of transport and structural parameters of forward osmosis membranes, J. Membr. Sci. 444 (2013) 523–538. [42] N.Y. Yip, A. Tiraferri, W.A. Phillip, J.D. Schiffman, L.A. Hoover, Y.C. Kim, M. Elimelech, Thin-film composite pressure retarded osmosis membranes for

Z. Zhou et al. / Chemical Engineering Journal 249 (2014) 236–245 sustainable power generation from salinity gradients, Environ. Sci. Technol. 45 (2011) 4360–4369. [43] F. Horkay, I. Tasaki, P.J. Basser, Osmotic swelling of polyacrylate hydrogel in physiological salt solutions, Biomacromolecules 1 (2000) 84–90. [44] S. Zhao, L. Zou, Relating solution physiochemical properties to internal concentration polarization in forward osmosis, J. Membr. Sci. 379 (2011) 459–467. [45] 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) 114–123.

245

[46] W.A. Phillip, J.S. Yong, M. Elimelech, Reverse draw solute permeation in forward osmosis: modeling and experiments, Environ. Sci. Technol. 44 (2010) 5170–5176. [47] D. Emadzadeh, W.J. Lau, T. Matsuura, M. Rahbari-Sisakht, A.F. Ismail, A novel thin film composite forward osmosis membrane prepared from PSf-TiO2 nanocomposite substrate for water desalination, Chem. Eng. J. 237 (2014) 70– 80. [48] X. Li, K.Y. Wang, B. Helmer, T.S. Chung, Thin-film composite membranes and formation mechanism of thin-film layers on hydrophilic cellulose acetate propionate substrates for forward osmosis process, Ind. Eng. Chem. Res. 51 (2012) 10039–10050.