Forward Osmosis with a Novel Thin-Film Inorganic Membrane

Jul 5, 2013 - *Phone: +86-451-86282008; fax: +86-451-86282110; e-mail: ... Forward osmosis (FO) represents a new promising membrane technology for ...
1 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/est

Forward Osmosis with a Novel Thin-Film Inorganic Membrane Shijie You,*,† Chuyang Tang,‡ Chen Yu,† Xiuheng Wang,*,† Jinna Zhang,† Jia Han,† Yang Gan,§ and Nanqi Ren† †

State Key Laboratory of Urban Water Resource and Environment (SKLUWRE), Harbin Institute of Technology (HIT), Harbin 150090, P.R. China ‡ School of Civil and Environmental Engineering (CEE), Nanyang Technological University (NTU), Singapore Membrane Technology Centre (SMTC) 639798, Singapore § Department of Catalysis Science and Engineering, School of Chemical Engineering and Technology, Harbin Institute of Technology (HIT), Harbin 150001, P.R. China S Supporting Information *

ABSTRACT: Forward osmosis (FO) represents a new promising membrane technology for liquid separation driven by the osmotic pressure of aqueous solution. Organic polymeric FO membranes are subject to severe internal concentration polarization due to asymmetric membrane structure, and low stability due to inherent chemical composition. To address these limitations, this study focuses on the development of a new kind of thin-film inorganic (TFI) membrane made of microporous silica xerogels immobilized onto a stainless steel mesh (SSM) substrate. The FO performances of the TFI membrane were evaluated upon a lab-scale cell-type FO reactor using deionized water as feed solution and sodium chloride (NaCl) as draw solution. The results demonstrated that the TFI membrane could achieve transmembrane water flux of 60.3 L m−2 h−1 driven by 2.0 mol L−1 NaCl draw solution at ambient temperature. Meanwhile, its specific solute flux, i.e. the solute flux normalized by the water flux (0.19 g L−1), was 58.7% lower than that obained for a commercial cellulose triacetate (CTA) membrane (0.46 g L−1). The quasi-symmetry thin-film microporous structure of the silica membrane is responsible for low-level internal concentration polarization, and thus enhanced water flux during FO process. Moreover, the TFI membrne demonstrated a substantially improved stability in terms of mechanical strength, and resistance to thermal and chemical stimulation. This study not only provides a new method for fabricating quasi-symmetry thin-film inorganic silica membrane, but also suggests an effective strategy using this alternative membrane to achieve improved FO performances for scale-up applications.

■

INTRODUCTION Recently, forward osmosis (FO) has drawn increasing research attention because it shows great promise for its potential in a variety of applications such as seawater or brackish water desalination,1 water or wastewater treatment,2 food processing,3 power production,4 and pharmaceutical concentration.5 Fundamentally, forward osmosis works based on a dense hydrophilic semipermeable membrane that separates two solutions with different osmotic pressures (OPs). The water molecules migrate spontaneously across the membrane from the lowOP side (the feed solution; FS) to the high-OP side (the draw solution; DS) driven by the transmembrane OP difference in the absence of external hydraulic pressure. This results in several unique advantages in terms of low energy consumption, efficient water recovery, less membrane fouling, and easy fouling removal, in comparison with conventional pressuredriven membrane processes such as reverse osmosis (RO), nanofiltration (NF), and ultrafiltration (UF).6 However, there remain several challenges for scale-up applications of FO © 2013 American Chemical Society

technology. These bottlenecks are related to the structure and properties of the FO membrane and the way it interacts with feed solution, draw solution, and operational parameters. Apart from some natural biomembranes such as animal bladders, collodion, rubber, and porcelain, the membranes for lab-scale FO studies are commonly prepared by phase-inversion method,7 thin-film composite preparation,8 and chemical modification.9 Polymeric membranes produced in these manners typically involve a dense selective active layer (AL) attached onto a porous supporting layer (SL), which allows water permeation and salt rejection. Such asymmetric structure, however, can cause severe internal concentration polarization (ICP) within the supporting layer, leading to a substantial loss of effective driving force and water flux.6 ICP is undesirable but Received: Revised: Accepted: Published: 8733

April 10, 2013 July 3, 2013 July 5, 2013 July 5, 2013 dx.doi.org/10.1021/es401555x | Environ. Sci. Technol. 2013, 47, 8733−8742

Environmental Science & Technology

Article

Preparation of Xerogels and STI Membrane. The preparation of the thin-film inorganic (TFI) membrane was based on layer-by-layer (LBL) deposition of silica xerogels onto the stainless steel mech (SSM) (see SI Figure S2). In detail, the pristine SSM was cut into small pieces (3.0 cm × 3.0 cm) and polished with alumina suspension (0.3-μm particles), followed by being washed with DI-water, acetone, ethanol, and DI-water again. The SSM sheets were pretreated using HNO3 (5.0 mol L−1) solution to produce hydroxyl-enriched or hydroxylated SSM surface.16 In parallel, the silica sol was prepared by HNO3catalyzed hydrolysis and condensation of TEOS in ethanol. The mixture of HNO3 and DI-water was added into the mixture of TEOS and ethanol (molar ratio of TEOS:C2H5OH:H2O:HNO3 = 1:3.8:6.4:0.085) under vigorous magnetic stirring, and then heated at 60 °C in water bath for 3.0 h.17 Following the cooling and dilution at room temperature, the reacted mixture was adhered onto the SSM substrate by dip-coating procedure to produce SSM-xerogels membrane. The as-prepared membrane was calcined at 500 °C in nitrogen atmosphere for 4.0 h in order to remove the organic groups and to produce the porous structure, followed by cooling to room temperature (25 °C) at rate of 0.5 °C min−1. The dip-coating and calcining procedures were repeated up to four times for stepwise repairing the defects via LBL deposition, giving rise to so-called one-, two-, three-, and four-layer coating membranes. For the as-prepared TFI membrane, the side from which the xerogel was casted and deposited was denoted active layer (AL) during FO experiment. Characterization. The morphologies of the as-prepared TFI membranes were characterized by using a field-emission scanning electron microscope (SEM, JEOL-6700F) equipped with energy dispersive X-ray spectroscopy (EDX) analysis and atomic force microscopy (AFM, Bioscope, Veeco, U.S.). The determination of the chemical composition of the electrocatalysts was taken by using an X-ray photoelectron spectrometer (XPS; PH1-5700 ESCA system, U.S.) equipped with a hemispherical analyzer and an aluminum anode (monochromatic Al Kα 1486.6 eV) as source. The mechanical strength of the TFI membrane was measured by using tensile testing system (type QJ211S, China) at testing rate of 5.0 mm min−1. The average Young’s modulus and tensile strength were obtained at the break point of the membrane material upon the tests in triplicate. The permporometry of the TFI membranes was performed on a Porosometry System (ASAP 2020, Global Spec. Inc., U.S.) with a turbo molecular pump system and pressure transducer using water vapor as the condensable gas and nitrogen as the permeating gas.18,19 The Kelvin diameter was calculated from the relative actual vapor pressure to saturated vapor pressure by using the Kelvin equation, and the average pore size was defined as the pore size obtained at the 50% permeability of nitrogen gas by condensed water.20 The water permeation coefficient (A, m s−1 Pa−1), NaCl permeation coefficient (B, m s−1), and NaCl rejection rate (R, %) of the TFI membrane were determined using pressure-induced filtration tests.6 All these details can be found in the SI. FO Setup and Operation. The lab-scale crossflow FO system was employed with the configuration as described in our previous study.21 (see SI Figure S3) The FO cell comprised two Plexiglas-made plates, containing an inner cavity that was separated by the membrane into two symmetric halves to form FS and DS compartments. Both FS and DS were continuously recirculated at a constant flow rate of 260 mL min−1 through a reservoir (2.0 L) by using two individual peristaltic pumps

difficult to control via hydrodynamic means. Thus, if it were feasible to develop a membrane with nonasymmetric structure, one would expect increased effective osmotic driving force and therefore enhanced water flux. Moreover, organic polymeric FO membranes may have additional inherent problems in terms of low mechanical strength and stiffness, low thermal stability (e.g., irreversible damage at temperature higher than 50 °C), and poor resistance to chemical erosion (e.g., acidic or alkaline environment).10 To address these issues, we attempt to prepare a microporous inorganic silica membrane with desired permeability and selectivity for FO application. Previous studies showed that the tetraethylorthosilicate (TEOS)-driven silica membrane had amorphous silica networks with micropores of nanometers (0.3−0.8 nm) for gas separation.11 It is known that the water molecules have a kinetic diameter of 0.26 nm, and the diameters of hydrated sodium cations (Na+) and chloride anions (Cl−) are 0.72 and 0.66 nm, respectively.12 Thus it may be possible to produce silica membrane with suitable pore size for water permeation and solute rejection. However, selfsupporting silica membranes do not yet exist, and thus the silica membranes require use with the support of a substrate material. This gives priority to the selection of suitable supporting materials in view of not only the membrane performances (i.e., water permeation and solute rejection) but also the possibility for scale-up applications (i.e., cost, mechanical strength, stiffness, and flexibility). Although there are several studies reporting the synthesis of asymmetric inorganic membrane using thick porous α-Al2O3,13 γ-Al2O3,14 and TiO215 for pressure-driven membrane processes, it appears unlikely these supporting substrates will be used for FO process because their large thickness (1−5 mm) will promote severe ICP. Moreover, the high cost and low flexibility of these materials also limits their scale-up applications. Herein, we report for the first time the synthesis of a novel thin-film inorganic (TFI) FO membrane by layer-by-layer (LBL) deposition of microporous silica xerogels on an inexpensive stainless steel mesh (SSM) support. The chemical composition, surface properties, pore size, and separation properties (water and salt permeability) were investigated. The FO behavior of TFI membrane was evaluated for different membrane orientations over a wide range of draw solution concentrations. In addition, the stability of TFI membrane was studied under different temperatures and pH values.

■

EXPERIMENTAL SECTION Materials and Chemicals. The type 316L stainless steel mesh (SSM; wire diameter of 18 μm and grid size of 1.0 μm) substrates of 40-μm thickness were purchased from Anping Metal Ltd., Hebei Province, China. The chemical composition of the pristine SSM is given in Supporting Information (SI) Table S1. The tetra-ethyl-ortho-silicate (TEOS; 99.9%, v/v) for TFI preparation was purchased from Sigma-Aldrich Co., St. Louis, MO (see SI Figure S1). The feed solution (FS) was deionized water (DI-water) and the draw solution (DS) was NaCl solution (0.0−2.0 mol L−1). In some cases, the commercially viable cellulose triacetate (CTA)-made hydrophilic Hydrowell membrane sheet (50-μm thickness) purchaed from Hydration Technologies Inc. (HTI, Albany, Oregon) was also included for comparison (HTI-CTA membrane), as it has received a high degree of acceptance as benchmark membrane material in FO studies.6 8734

dx.doi.org/10.1021/es401555x | Environ. Sci. Technol. 2013, 47, 8733−8742

Environmental Science & Technology

Article

Figure 1. Optical photographs of (A) pristine SSM and (B, C) TFI membrane. SEM images of (D) the pristine SSM, the TFI membrane with (E) one-, (F) two-, (G) three-, and (H) four-coating layers of xerogels, (I) the cross section of the TFI membrane with four-coating layers, (J) magnified apertures, and (K) membrane surface. The inserted images in (D) and (E) indicate the enlarged morphology of the stainless steel wire before and after xerogel coating.

into the FS and DS to examine the chemical resistance of the TFI membrane. The forward transmembrane water permeation flux (JV, L m−2 h−1) was determined by recording the FS volume reduction using an online data logging system connetced with a personal computer, and the reverse salt permeation flux (JS, g m−2 h−1) was obtained by measuring the increased amount of

(type BT100-1Z, China). The FS and DS temperature (25−70 °C) was controlled by using immersed heater and real-time temperature controllers (type XMT9000, China). The membranes were tested under both forwad osmosis (active layer-feed solution, AL-FS) and pressure retarded osmosis (active layer-draw solution, AL-DS) mode. In some experiments, HCl (1.0 mol L−1) or NaOH (1.0 mol L−1) was added 8735

dx.doi.org/10.1021/es401555x | Environ. Sci. Technol. 2013, 47, 8733−8742

Environmental Science & Technology

Article

Figure 2. C1s and Si2p XPS spectra of the pristine SSM, pretreated SSM, and xerogel-deposited SSM.

solute on FS side. The final concentration of NaCl on FS side was measured by using inductively coupled plasma optical emission spectrometer (ICP-OES, Perkin-Elmer Optima 5300DV, U.S.). All details can be found in the SI.

dimensional AFM images (Figure 3) showing considerable increase in the roughness of SSM surface (Ra = 8.58, Rms = 36.88, Rmax = 39.41) after silica film coating compared with uncoated one (Ra = 2.87, Rms = 14.08, Rmax = 15.93) (see SI,Figure S5 and Table S2). The XPS (C1s and Si2p peaks), EDX, and FT-IR analyses confirmed the formation of siloxane (Si−O bonds) and carbon-silica (C−Si) matrix on the SSM surface after xerogel deposition and calcination (Figure 2 and SI Figures S4 and S6). Although the one-layer-coating membrane shows good intergrown morphological characteristics, the defects between the silica film and steel wires appear unlikely to be fully closed, meaning there still existed large pore size and incapability of solute rejection. This is because the thickness of single-coating layer cannot exceed 0.1−0.3 μm, otherwise the horizontal attaching force will be larger than the bonding force between coating layer and the SSM causing the contracting and deattachment of silica matrix. Thus, the xerogels were repeatedly immobilized onto the as-prepared SSM-supported silica-coating surface via layer-by-layer (LBL) deposition, until all the defects were completely repaired. As can be seen from

■

RESULTS Characterization of TFI Membrane. The pristine SSM had thin steel wires (18 μm in diameter) with long-range regularly interweaved network and smooth surface (Figure 1A and D), characterized by the content of typical metallic elements (see SI Figure S4), and negligible oxygen and silica as shown in XPS spectra (Figure 2). The hydroxylated SSM after oxidative pretreatment was immobilized with abundant oxygen elements, producing hydrophilic surface (C−O bonds, Figure 2 and SI Figure S4) for chemisorptions and deposition of silica coatings. As shown in Figure 1B, C, and E, the sol−gel silica film could be formed on the SSM using a simple dip-coating process and post thermal treatment under 500 °C for 4 h. Following the coating of xerogels, the steel wires were immobilized with uniform, continuous, porous film on the smooth SSM surface. This was confirmed by the three8736

dx.doi.org/10.1021/es401555x | Environ. Sci. Technol. 2013, 47, 8733−8742

Environmental Science & Technology

Article

membrane demonstrated excellent mechanical strength characterized by Young’s modulus and tensile strength being 2−3 orders of magnitude larger than that of polymeric membrane reported in the literature (see SI Table S4).22−25 Overall, unlike conventional methods for preparation of the polymeric FO membranes,6,10 this synthetic strategy eliminates the need for a definable supporting layer (SL)/active layer (AL) interface, producing an inorganic membrane with selfsupporting and thin-film macro structures and three-dimensional nanopores and nanonetwork micro structures. This would be expected to give several advantages in FO process as discussed in detail below. Water and Solute Permeation of TFI Membrane. The water permeation and salt rejection of TFI membrane were evaluated based on pressure filtration of NaCl solution. Figure 4 reveals the strong trade-off between water permeation and

Figure 3. AFM images of the stainless steel wire surface (A) before and (B) after xerogel coating (one-layer). The corresponding section analyses are provided in SI Figure S5, which gives the Ra, Rms, and Rmax values. Figure 4. Water and salt permeability coefficient as function of coating-layer-dependent pore diameter of the TFI membrane. The highlighted lower-left region indicates the available pore size for reasonable A and B values for water permeation and NaCl rejection.

Figure 1E−H, with the increased number of coating layers, the membrane surface tended to become more closed, forming dense long-range intertwined network structure. The one-, two-, three-, and four-layer coating membrane corresponded to the average pore size of 8.6, 5.5, 2.4, and 1.1 nm upon the pore size at 50% permeability of nitrogen gas condensed by water measured by permporometry analysis20 (see SI Figures S7 and S8). The evolution of surface morphology and decreased pore diameter indicated the large-size pores were filled by the toplayer siloxane matrix. For convenient description, these membranes were denoted according to the convection of TFI-MN (N, the pore size measured). The membrane with pore size of 1.1 nm (TFI-M1.1) exhibited well-defined FO performance as discussed in detail below. The LBL deposition of xerogels produced the final membrane (TFI-M1.1) with thickness of approximately 45 μm (Figure 1F), which was thinner than that of the commercial HTI-CTA membrane (50 ± 2.0 μm).21 The apertures between the steel wires and siloxane matrix provided microchannels for permeation of water molecules (Figure 1J). The contact angle of the membrane surface increased from 50.54° to 67.71° with the increased number of coating layers (SI Figure S8), possibly due to the increased amount of CH3-terminated domains and shielding of silanol groups (see SI Figure S6 and Table S3) and surface roughness (Figure 1K). Even so, the finally prepared TFI-M1.1 (contact angle of 67.71°) had relatively good hydrophilic property compared with the HTI-CTA membrane. Thanks to the use of steel material as support, the TFI

NaCl rejection depending on the pore size of the TFI membrane. With the decreased pore size (8.2−1.1 nm) resulting from the increase in number of silica coating layers (1−4), there was an observation in synchronous decline in water and salt permeation, as indicated by decreased A (6.7 × 10−10 to 3.2 × 10−12 m s−1 Pa−1), and B (5.6 × 10−5 to 1.8 × 10−7 m s−1) values, respectively. The increased coating layers led to the decrease in aperture size and thus the mobility of both molecular and ionic species in the micropores (diffusivity as low as 10−18 m2 s−1).26 The free water molecules within the membrane pores are much fewer than in bulk (feed solution) due to interaction of hydrated solute ions and charged double layers particularly for the double layers overlapping in nanoscale space.27 Hence, the migration of free water molecules is weakened by the slow diffusing hydrated solute ions due to inability of the molecules to pass each other easily in a confined space. In addition, the smaller pores make it easy for hydrated ions to enter the pores and channels and bind onto the pore surface, and hence improving the NaCl rejection but hindering the diffusion of water molecules concurrently. This may explain why both the forward water flux and reverse salt flux are decreased synchronously with the decreased pore size of the membrane. The water permeation coefficient of TFI-M1.1 (3.2 × 10−12 m s−1 Pa−1) was on the same magnitude, and 8737

dx.doi.org/10.1021/es401555x | Environ. Sci. Technol. 2013, 47, 8733−8742

Environmental Science & Technology

Article

approximately 90% higher than that (2.2 × 10−12 m s−1) of commercial HTI-CTA membrane.28 The higher A value should account for lower diffusion resistance to water transport because of self-supporting thin-film structure. The decreased B values accompanied the enhancement in selectivity of the membrane, of which the B value (1.8 × 10−7 m s−1) for the TFI membrane was comparable to that (1.7 × 10−7 to 3.7 × 10−7 m s−1) for HTI-CTA membrane. 29 FO Performances of the TFI Membrane. There was no definable FO performance observed, i.e. severe reverse salt diffusion, using TFI-M8.2, TFI-M5.5, and TFI-M2.4 (data not shown). On the contrary, the TFI-M1.1 exhibited defined water flux and salt rejection during FO process being in line with the dependence of A and B values on pore size as addressed above. As shown in Figure 5A, the water flux was increased substantially with the increased DS concentration in the range of 0.5−2.0 mol L−1, reaching the unprecedentedly great values of 60.3 L m−2 h−1 (AL-DS mode) and 59.8 L m−2 h−1 (AL-FS mode) at DS of 2.0 mol L−1. This amounted to approximately 183.1 and 376.7 times of water flux as great as

that of 21.3 L m−2 h−1 (AL-DS mode) and 11.9 L m−2 h−1 (ALFS mode) for HTI-CTA membrane upon the same conditions. Unlike the asymmetry-structured HIT-CTA membrane where the AL-DS water flux was 77.6% higher than the AL-FS water flux,30 the TFI-M1.1 produced very similar water flux at both membrane-orientating modes, indicating the slight impact of membrane orientation to water flux because of the quasisymmetry structure. The JS/JV value (AL-FS mode at DS of 2.0 mol L−1) obtained for the TFI-M1.1 (average 0.19 g L−1) was 58.7% smaller than that of the HTI-CTA membrane (average 0.46 g L−1; Figure 5B) upon different DS concentrations tested, which implied good selectivity of TFI membrane for FO applications. With lower internal concentration polarization (ICP) limitation as a result of self-supporting thin-film structure, the TFI membrane was expected to deliver higher effective driving force, achieving the water flux much higher than majority values (i.e., JV of 5.0−59.0 L m−2 h−1) obtained at ambient temperature in previous FO studies (Table 1). Furthermore, the hydrophilic nature of the TFI may also help to control ICP.53 Effect of Temperature and pH. To further access the stability of TFI-M1.1 during FO operation, we examined the effect of working temperature and pH on transmembrane water flux. The temperature−pH matrix diagram (Figure 6) shows that the TFI-M1.1 membrane was able to tolerate a wide range of pH values (5.0−9.0) to perform FO water flux durably. The water flux was observed to increase greatly at all pHs when elevating temperature from 25 to 70 °C, arriving at the highest value of 83.23 L m−2 h−1 (AL-FS mode at 70 °C). The enhancement of diffusive kinetics within membrane related to the decline in water viscosity as function of temperature should be the most likely reason for the improvement of FO performances, which was consistent with the observation in our previous studies21 and other literature.54−56 The high chemical and thermal resistance of stainless steel substrate makes the TFI membrane far superior to the organic polymeric membrane in terms of the stability (Figure 6).

■

DISCUSSION Improved Water Permeation and Solute Rejection of TFI Membrane. The improved FO performances obtained here should be attributed to the TFI membrane with selfsupporting nonasymmetry thin-film structure and LBLdepositing microporous xerogels. Currently, it is common practice to prepare polymeric FO membrane by coating a dense active layer (AL) onto the porous supporting layer (SL), which yields the typical asymmetric structure characterized by defined AL/SL interface for either membrane sheet or membrane hollow module.10 This synthetic strategy also applies to the preparation of inorganic membranes such as zeolite membrane and SiO2 membrane for gas separation.13−15 However, the commonly used porous template may be problematic for the pressure-free FO system, due to great thickness of 1−5 mm and the resultant severe internal polarization. We herein used the SSM as the supporting substrate to mitigate this problem. The SSM material serves as thin corrosion-resistant metallic sheet (c.a. 40-μm thickness) made of highly ordered cross-linked wire-steel networks, providing long-range intertwined surface for silica xerogel deposition. The microscale grids (1−2 μm) among the interweaved steel wires yield sufficient surface tension to support the xerogels firmly and facilitate water permeation at low resistance. The simple repeated LBL deposition of xerogels onto SSM substrate results

Figure 5. (A) Transmembrane water flux (JV) and (B) specific reverse salt flux (JS/JV) as function of DS concentration for TFI-M1.1 and HTICTA membrane. The dashed lines in blue and green indicate the ideal water flux of both membranes. The data with error bars ± SD indicate the experiments in triplicate. 8738

dx.doi.org/10.1021/es401555x | Environ. Sci. Technol. 2013, 47, 8733−8742

Environmental Science & Technology

Article

Table 1. Comparison of FO Performances for Different Membrane Materials Operated under AL-FS Mode FO performance (AL-FS)a membranes HTI-CTA membrane TFI membrane HTI flat sheet PDA-PSf TFC Zeolite-PAD TFN PAI-PES double-layer hollow fiber SPEK-TFC CA-NF hollow fiber PAD-PES TFC hollow fiber PAD-PSf TFC flat sheet PAD-PSf TFC flat sheet PAD-PSf TFC flat sheet PAD-PEN hollow fiber positively charged PAI flat sheet double dense layer flat sheet nanofiber composite TFC hollow fiber CAP-I TFC double-skinned cross-linked LBL cellulose ester CA PAI NF hollow fiber CA double-skin PAI-RO hollow fiber (single-skinned) dual-layer (PBI-PES) hollow fiber CTA/dioxane/acetone/acetic acid HTI-CTA

JVb

JSc

JS/JVd

R (%)e

ref

11.9 60.3 36.4f 8.0 22.3 27.5 35.0 5.0 14.0 18.2 12.0 25.0 12.9 19.2 10.0 59.0 14 18.5 42.3 37.9 14.0 27.4 46.3 15.7 22.7 12.5

5.5 11.4

0.46 0.19

91 92

1.5

0.19

5.5 7

0.20 0.20

1.75

0.13

85 90 89 91 90 91

4.9

0.41

93

4.8 9.6 0.8

0.37 0.50 0.08

1.8 1.9 19.8 11.5 4.2 3.8 7.6 0.47 19.6 5.8

0.13 0.10 0.47 0.32 0. 3 0.14 0.16 0.03 0.86 0.46

this work this work 31 32 29 33 34 35 36 37 30 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52

draw solution 2.0 2.0 5.0 2.0 2.0 0.5 2.0 2.0 0.5 1.5 0.5 1.0 1.5 0.5 2.0 2.0 0.5 2.0 0.5 2.0 0.5 5.0 2.0 1.0 2.0 2.0

M M M M M M M M M M M M M M M M M M M M M M M M M M

NaCl NaCl NH4HCO3 NaCl NaCl MgCl2 NaCl MgCl2 NaCl NaCl NaCl NaCl MgCl2 MgCl2 MgCl2 NaCl NaCl NaCl MgCl2 NaCl Na2SO4 MgCl2 NaCl MgCl2 NaCl NaCl

87 99 97g 89 94

79 90 87 89 91

The reported AL-DS water flux data are not taken into account for comparison because of severe membrane fouling in the supporting layer in desalination and wastewater treatment. The maximum water flux was collected from the reported studies using DI-water as feed solution under room temperature (20−25 °C). bJV: water flux (L m−2 h−1). cJS: salt flux (g m−2 h−1). dJS/JV: specific salt flux (g L−1). eR: the salt rejection efficiency (%). f FS of 0.05 mol L−1 NaCl at 50 °C. gDS of 0.5 mol L−1 NaCl. a

SSM allows the formation of self-supporting thin-film structure lacking defined AL/SL interface. This eliminates the requirement for individual thick supporting layer, and hence enabling short-distance water permeation through the nanometer-scale pores with minimal internal polarization (see SI Figure S9). Based on the methods described previously,57 the S value of the TFI membrane was predicted to be 38 μm (see SI Figure S10), which is significantly smaller than that of HTI membrane, and those reported in prior studies,57−59 indicating the quasisymmetry structure is highly effective in minimizing ICP. However, a typical TFC polymeric membrane has three distinct layers: an ultrathin polyamide layer for rejection, a polysulfone or polyethersulfone porous layer for cushion, and a nonwoven fabric support (or woven support in the case of FO applications) for mechanical strength. According to prior publications,23,60 the mesoporous cushion layer (particularly the skin of the cushion layer) can contribute significantly to the overall structural parameter as a result of its lower porosity and high tortuosity. In the current study, the inorganic rejection layer was coated directly onto the woven stainless steel mesh for mechanical support, thus avoiding the need for a mesoporous cushion layer. Since the stainless steel mesh is relatively thin and has straight passage for water and solutes (tortuosity ∼1), the overall structural parameter of the membrane was low (S = 38 μm). The RO experiments performed for cationic ions with different hydrated-ion sizes and charge densities suggested that the salt rejection efficiency (R) for NH4+ (94.3%) and Mg2+

Figure 6. Water flux diagram with respect to pH values and operating temperatures of TFI-M1.1 at AL-FS mode. The geometric size of the discrete circles characterizes the magnitude of water flux (L m−2 h−1). The region (A) and (B) highlighted in green and gray indicates the available operating conditions for HTI-CTA membrane (20−40 °C and pH 6.5−7.5) and TFI-M1.1 (20−70 °C and pH 5.0−9.0), respectively. The values are given in SI Table S5.

in the formation of a macro-scale continuous, uniform, and porous thin sheet with reinforced-concrete-like structure (Figure 1H and I) for salt rejection. The mechanically rigid 8739

dx.doi.org/10.1021/es401555x | Environ. Sci. Technol. 2013, 47, 8733−8742

Environmental Science & Technology

Article

(95.5%) were higher than that for Na+ ions (92.1%) upon the same solute molar concentration and pressure applied. Although the polyatomic NH4+ ions have the same hydrated number as Na+, the former can form a larger hydrated shell due to rigid tetrahedral four-point charge model.61 This result, together with the increased salt-rejecting coefficient with decrease in pore size (Figure 4), confirmed the pore size dominates the size-exclusive effect that discriminates the hydrated solute ions (Na+ and Cl−) and water molecules, such that the smaller-size pores accounted for higher selectivity. On the other hand, since the bivalent Mg2+ ions possess higher charge density than monovalent Na+ ions, the Mg2+ ions are more capable of polarizing the neighboring water molecules to produce greater and more rigid hydrated complexes than Na+ ions. Additionally, the smaller-sized pores render the microscale pore walls to carry higher-charged density, which enhances the interactions between the charged ions and overlapping charged double layers.62 These results suggest that the charge-repulsion effect also should be of importance in rejecting solute ions through Donnan exclusion (see zeta potential analysis in SI Figure S11). Although the pore size (1.1 nm) of TFI-M1.1 is larger than that of both water molecules and hydrated Na+ ions, it is actually capable of delivering the FO water flux and solute rejection. This should ascribe to much lower diffusivities and mobility of Na+ and Cl− ions than that of water molecules because of strong interactions between the polar siloxane matrix or the overlapped double-layered networks, and charged solute ions in the nanoscale cages.63,64 Besides, the siloxane matrix between the adjacent layers tends likely to cluster and leave percolating spaces, and thus produce cross-linked hydrophilic nanocapillary networks for much faster forward water permeation than reverse salt diffusion. Applications of TFI Membrane for FO process. Since the FO process was first proposed for potential application several decades ago,65 various types of FO membranes have been developed toward FO water production with limited success. In view of membrane materials, there are two main reasons for these unsuccessful outcomes. First, conventional polymeric FO membranes contain a thick supporting layer, and are highly resistant to water permeation. This fact can be seen in Table 1, illustrating that the currently available FO membranes have a low magnitude of water flux less than 60.0 L m−2 h−1. Second, the polymeric FO membranes are highly sensitive to surroundings, i.e. low thermal resistance and chemical stability because of their intrinsic chemical composition.10 Thus, to prevent any adverse impact induced to membrane, the majority of lab-scale FO experiments have been performed under ambient temperature (lower than 50 °C) and pH-neutral conditions. This implies lower-degree adaptability for this technology to be used for more practical situations. We have demonstrated three unique advantages of inorganic TFI membrane for FO application, i.e. low internal polarization, high stability, and ease of scaling-up. Thanks to the selfsupporting thin-film structure, the TFI membrane shows highefficiency water permeation (JV of 60.3 L m−2 h−1) and good selectivity toward NaCl (JS/JV of 0.19 g L−1) at ambient temperature. In particular, by increasing the operational temperature to 70 °C, unprecedented water flux as great as 83.23 L m−2 h−1 was obtained (pH 7.0). Despite the positive correlation between FO performance and temperature,21,54−56 this appears impossible for polymeric membranes due to their sensitivity to temperatures in excess of 50 °C. The temperature

tolerance of the TFI membrane may facilitate realization of thermal separation of low-boiling-point draw solution (e.g., CO2 + NH31 or NH4HCO331) and water at elevated temperature. The ability of TFI membrane to sustain FO process without any curling, wrinkling, cracking, deformation, or dissolution under harshest conditions also suggests a superior mechanical strength, thermal, and chemical stability over polymeric FO membrane. The industrially viable SSM may offer a new substrate material to synthesize inexpensive, flexible, and stable inorganic silica membrane, making it easy and feasible to scale up the FO systems. Notwithstanding this, the membrane preparation and operation need further optimization to enhance the solute rejection without losing high water flux, in order to render the inorganic membrane more effective in potential application of FO or FO−RO loop systems for desalination, power production, and wastewater treatment.

■

ASSOCIATED CONTENT

S Supporting Information *

Additional information as mentioned in the text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Phone: +86-451-86282008; fax: +86-451-86282110; e-mail: [email protected] (S.Y.); [email protected] (X.W.). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Project supported by the National Natural Science Foundation of China (51108121, 51208142), State Key Laboratory of Urban Water Resource and Environment (Grant 2013TS07), and the Project 51121062 (National Creative Research Groups) supported by National Nature Science Foundation of China, and China Postdoctoral Science special Foundation.

■

REFERENCES

(1) McCutcheon, J. R.; McGinnis, R. L. A novel ammonia carbon dioxide forward (direct) osmosis desalination process. Desalination 2005, 174, 1−11. (2) Holloway, R. W.; Childress, A. E.; Dennett, K. E.; Cath, T. Y. Forward osmosis for concentration of anaerobic digester centrate. Water Res. 2007, 41, 4005−4014. (3) Petrotos, K. B.; Lazarides, H. N. Osmotic concentration of liquid foods. J. Food Eng. 2001, 49, 201−206. (4) Pattle, R. E. Production of electric power by mixing fresh and salt water in the hydroelectric pile. Nature 1954, 174, 660. (5) Santus, G.; Baker, R. W. Osmotic drug delivery: A review of the patent literature. J. Controlled Release 1995, 35, 1−21. (6) Cath, T. Y.; Childress, A. E.; Elimelech, M. Forward osmosis: Principles, applications, and recent developments. J. Membr. Sci. 2006, 281, 70−87. (7) Wang, K. Y.; Chung, T. S.; Qin, J. J. Polybenzimidazole (PBI) nanofiltration hollow fiber membranes applied in forward osmosis process. J. Membr. Sci. 2007, 300, 6−12. (8) Wei, J.; Qiu, C.; Tang, C. Y. Y.; Wang, R.; Fane, A. G. Synthesis and characterization of flat-sheet thin film composite forward osmosis membranes. J. Membr. Sci. 2011, 372, 292−302. (9) Setiawan, L.; Wang, R.; Li, K.; Fane, A. G. Fabrication of novel poly(amide−imide) forward osmosis hollow fiber membranes with a positively charged nanofiltration-like selective layer. J. Membr. Sci. 2011, 369, 196−205.

8740

dx.doi.org/10.1021/es401555x | Environ. Sci. Technol. 2013, 47, 8733−8742

Environmental Science & Technology

Article

(10) Zhao, S. F.; Zou, L. D.; Tang, C. Y. Y.; Mulcahy, D. Recent developments in forward osmosis: Opportunities and challenges. J. Membr. Sci. 2012, 396, 1−21. (11) De Vos, R. M.; Verweij, H. High-selectivity, high-flux silica membranes for gas separation. Science 1998, 279, 1710−1711. (12) Lin, J.; Murad, S. A computer simulation study of the separation of aqueous solutions using thin zeolite membranes. Mol. Phys. 2001, 99, 1175−1181. (13) Lee, H. R.; Shibata, T.; Kanezashi, M.; Mizumo, T.; Tsuru, T. Pore-size-controlled silica membranes with disiloxane alkoxides for gas separation. J. Membr. Sci. 2011, 383, 152−158. (14) Campaniello, J.; Engelen, C. W. R.; Haije, W. G.; Vente, J. F. Long-term pervaporation performance of microporous methylated silica membranes. Chem. Commun. 2004, 834−835. (15) De Lange, R. S. A.; Hekkink, J. H. A.; Keizer, K.; Burggraaf, A. J. Gas transport in microporous sol−gel derived ceramic membranes. J. Porous Mater. 1995, 2, 141−149. (16) Bozec, N. L.; Compere, C.; L’Her, M.; Laouenan, A.; Costa, D.; Marcus, P. Influence of stainless steel surface treatment on the oxygen reduction reaction in seawater. Corros. Sci. 2001, 43, 765−786. (17) De Vos, R. M.; Verweij, H. Improved performance of silica membranes for gas separation. J. Membr. Sci. 1998, 143, 37−51. (18) Cuperus, F. P.; Smolders, C. A. Characterization of anisotropic UF-membranes: Top layer thickness and pore structure. J. Membr. Sci. 1991, 71, 73−83. (19) Cuperus, F. P.; Bargeman, D.; Smolders, C. A. Permporometry: The determination of the size distribution of active pores in UF membranes. J. Membr. Sci. 1992, 71, 57−67. (20) Tsuru, T.; Hino, T.; Yoshioka, T.; Asaeda, M. Permporometry characterization of microporous ceramic membranes. J. Membr. Sci. 2001, 186, 257−265. (21) You, S. J.; Wang, X. H.; Zhong, M.; Zhong, Y. J.; Yu, C.; Ren, N. Q. Temperature as a factor affecting transmembrane water flux in forward osmosis: Steady-state modeling and experimental validation. Chem. Eng. J. 2012, 198, 52−60. (22) Han, G.; Chung, T. S.; Toriida, M.; Tamai, S. Thin-film composite forward osmosis membranes with novel hydrophilic supports for desalination. J. Membr. Sci. 2012, 423, 543−555. (23) Wang, R.; Shi, L.; Tang, C. Y. Y.; Chou, S.; Qiu, C. Q.; Fane, A. G. Characterization of novel forward osmosis hollow fiber membranes. J. Membr. Sci. 2010, 355, 158−167. (24) Qin, J. J.; Chung, T. S. Effect of dope flow rate on the morphology, separation performance, thermal and mechanical properties of ultrafiltration hollow fibre membranes. J. Membr. Sci. 1999, 157, 35−51. (25) Chung, T. S.; Qin, J. J.; Gu, J. Effect of shear rate within the spinneret on morphology, separation performance and mechanical properties of ultrafiltration polyethersulfone hollow fiber membranes. Chem. Eng. Sci. 1999, 55, 1077. (26) Ruthven, D. M. Principles of Adsorption and Adsorption Processes; John Wiley & Sons: New York, 1984. (27) Palmeri, J.; Blanc, P.; Larbot, A.; David, P. Theory of pressuredriven transport of neutral solutes and ions in porous ceramic nanofiltration membranes. J. Membr. Sci. 1999, 160, 141−170. (28) Tang, C. Y. Y.; She, Q. H.; Lay, W. C. L.; Wang, R.; Field, R.; Fane, A. G. Modeling double-skinned FO membranes. Desalination 2011, 283, 178−186. (29) Ma, N.; Wei, J.; Liao, R. H.; Tang, C. Y. Y. Zeolite-polyamide thin film nanocomposite membranes: Towards enhanced performance for forward osmosis. J. Membr. Sci. 2012, 405, 149−157. (30) Wei, J.; Qiu, C. Q.; Tang, C. Y. Y.; Wang, R.; Fane, A. G. Synthesis and characterization of flat-sheet thin film composite forward osmosis membranes. J. Membr. Sci. 2011, 372, 292−302. (31) McCutcheon, J. R.; McGinnis, R. L.; Elimelech, M. Desalination by ammonia-carbon dioxide forward osmosis: Influence of draw and feed solution concentrations on process performance. J. Membr. Sci. 2006, 278, 114−123. (32) Han, G.; Zhang, S.; Li, X.; Widjojo, N.; Chung, T. S. Thin film composite forward osmosis membranes based on polydopamine

modified polysulfone substrates with enhancements in both water flux and salt rejection. Chem. Eng. Sci. 2012, 80, 219−231. (33) Setiawan, L.; Wang, R.; Shi, L.; Li, K.; Fane, A. G. Novel duallayer hollow fiber membranes applied for forward osmosis process. J. Membr. Sci. 2012, 421, 238−246. (34) Su, J. C.; Chung, T. S.; Helmer, B. J.; De Wit, J. S. Enhanced double-skinned FO membranes with inner dense layer for wastewater treatment and macromolecule recycle using Sucrose as draw solute. J. Membr. Sci. 2012, 396, 92−100. (35) Su, J. C.; Yang, Q.; Teo, J. F.; Chung, T. S. Cellulose acetate nanofiltration hollow fiber membranes for forward osmosis processes. J. Membr. Sci. 2010, 355, 36−44. (36) Wang, R.; Shi, L.; Tang, C. Y. Y.; Chou, S.; Qiu, C. Q.; Fane, A. G. Characterization of novel forward osmosis hollow fiber membranes. J. Membr. Sci. 2010, 355, 158−167. (37) Yip, N. Y.; Tiraferri, A.; Phillip, W. A.; Schiffman, J. D.; Elimelech, M. High performance thin-film composite forward osmosis membrane. Environ. Sci. Technol. 2010, 44, 3812−3818. (38) Tiraferri, A.; Yip, N. Y.; Phillip, W. A.; Schiffman, J. D.; Elimelech, M. Relating performance of thin-film composite forward osmosis membranes to support layer formation and structure. J. Membr. Sci. 2011, 367, 340−352. (39) Qiu, C.; Setiawan, L.; Wang, R.; Tang, C. Y. Y.; Fane, A. G. High performance flat sheet forward osmosis membrane with an NFlike selective layer on a woven fabric embedded substrate. Desalination 2012, 287, 266−270. (40) Setiawan, L.; Wang, R.; Li, K.; Fane, A. G. Fabrication of novel poly (amide−imide) forward osmosis hollow fiber membranes with a positively charged nanofiltration-like selective layer. J. Membr. Sci. 2011, 369, 196−205. (41) Zhang, S.; Wang, K. Y.; Chung, T. S.; Chen, H.; Jean, Y. C.; Amy, G. Well-constructed cellulose acetate membranes for forward osmosis: Minimized internal concentration polarization with an ultrathin selective layer. J. Membr. Sci. 2010, 360, 522−535. (42) Song, X. X.; Liu, Z. Y.; Sun, D. R. D. Nano gives the Answer: Breaking the bottleneck of internal concentration polarization with a nanofiber composite forward osmosis membrane for a high water production rate. Adv. Mater. 2011, 23, 3256−3260. (43) Chou, S. R.; Shi, L.; Wang, R.; Tang, C. Y. Y.; Qiu, C. Q.; Fane, A. G. Characteristics and potential applications of a novel forward osmosis hollow fiber membrane. Desalination 2010, 261, 365−372. (44) Li, X.; Wang, K. Y.; Helmer, B.; Chung, T. S. Thin-film composite membranes and formation mechanism of thin-film layers on hydrophilic cellulose acetate propionate substrates for forward osmosis processes. Ind. Eng. Chem. Res. 2012, 51, 10039−10050. (45) Qi, S. R.; Qiu, C. Q.; Zhao, Y.; Tang, C. Y. Y. Double-skinned forward osmosis membranes based on layer-by-layer assembly-FO performance and fouling behavior. J. Membr. Sci. 2012, 405, 20−29. (46) Zhang, S.; Zhang, R. W.; Jean, Y. C.; Paul, D. R.; Chung, T. S. Cellulose esters for forward osmosis: Characterization of water and salt transport properties and free volume. Polymer 2012, 53, 2664−2672. (47) Setiawan, L.; Wang, R.; Li, K.; Fane, A. G. Fabrication and characterization of forward osmosis hollow fiber membranes with antifouling NF-like selective layer. J. Membr. Sci. 2012, 394, 80−88. (48) Wang, K. Y.; Ong, R. C.; Chung, T. S. Double-skinned forward osmosis membranes for reducing internal concentration polarization within the porous sublayer. Ind. Eng. Chem. Res. 2010, 49, 4824−4831. (49) Fang, W. X.; Wang, R.; Chou, S. R.; Setiawan, L.; Fane, A. G. Composite forward osmosis hollow fiber membranes: Integration of RO- and NF-like selective layers to enhance membrane properties of anti-scaling and anti-internal concentration polarization. J. Membr. Sci. 2012, 394, 140−150. (50) Yang, Q.; Wang, K. Y.; Chung, T. S. Dual-layer hollow fibers with enhanced flux as novel forward osmosis membranes for water production. Environ. Sci. Technol. 2009, 43, 2800−2805. (51) Ong, R. C.; Chung, T. S. Fabrication and positron annihilation spectroscopy (PAS) characterization of cellulose triacetate membranes for forward osmosis. J. Membr. Sci. 2012, 394, 230−240. 8741

dx.doi.org/10.1021/es401555x | Environ. Sci. Technol. 2013, 47, 8733−8742

Environmental Science & Technology

Article

(52) Yong, J. S.; Phillip, W. A.; Elimelech, M. Coupled reverse draw solute permeation and water flux in forward osmosis with neutral draw solutes. J. Membr. Sci. 2012, 392, 9−17. (53) McCutcheon, J. R.; Elimelech, M. Influence of membrane support layer hydrophobicity on water flux in osmotically driven membrane processes. J. Membr. Sci. 2008, 318, 458−466. (54) Phuntsho, S.; Vigneswaran, S.; Kandasamy, J.; Hong, S.; Lee, S.; Shon, H. K. Influence of temperature and temperature difference in the performance of forward osmosis desalination process. J. Membr. Sci. 2012, 415, 734−744. (55) Zhao, S.; Zou, L. Effects of working temperature on separation performance, membrane scaling and cleaning in forward osmosis desalination. Desalination 2011, 278, 157−164. (56) McCutcheon, J. R.; Elimelech, M. Influence of concentrative and dilutive internal concentration polarization on flux behavior in forward osmosis. J. Membr. Sci. 2006, 284, 237−247. (57) Phillip, W. A.; Yong, J. S.; Elimelech, M. Reverse draw solute permeation in forward osmosis: Modeling and experiments. Environ. Sci. Technol. 2010, 44, 5170−5176. (58) Tang, C. Y.; She, Q. H.; Lay, W. C. L.; Wang, R.; Fane, A. G. Coupled effects of internal concentration polarization and fouling on fluxbehavior of forward osmosis membranes during humic acid filtration. J. Membr. Sci. 2010, 354, 123−133. (59) Cath, T. Y.; Elimelech, M.; McCutcheon, J. R.; McGinnis, R. L.; Achilli, A.; Anastasio, D.; Brady, A. R.; Childress, A. E.; Farr, I. V.; Hancock, N. T.; Lampi, J.; Nghiem, L. D.; Xie, M.; Yip, N. Y. Standard methodology for evaluating membrane performance in osmotically driven membrane processes. Desalination 2013, 312, 31−38. (60) Ma, N.; Wei, J.; Qi, S.; Zhao, Y.; Gao, Y.; Tang, C. Y. Nanocomposite substrates for controlling internal concentration polarization in forward osmosis membranes. J. Membr. Sci. 2013, 441, 54−62. (61) Palinkas, G.; Radnai, T.; Szasz, G. I.; Heinzinger, K. The structure of an aqueous ammonium chloride solution. J. Chem. Phys. 1981, 74, 3522. (62) Malusis, M. A.; Shackelford, C. D. Chemico-osmotic efficiency of a geosynthetic clay liner. J. Geotechnol. Geoenviron. Eng. 2002, 128, 629−629. (63) Lin, C. X.; Ding, L. P.; Smart, S.; Da Costa, J. C. Metamorphosing the SASW method by 2D wavefield transformation. J. Colloid Interface Sci. 2012, 368, 70−76. (64) Li, L. X.; Dong, J. H.; Nenoff, T. M.; Lee, R. Reverse osmosis of ionic aqueous solutions on a MFI zeolite membrane. Desalination 2004, 170, 309−316. (65) Kravath, R. E.; Davis, J. A. Desalination of sea water by direct osmosis. Desalination 1975, 16, 151−155.

8742

dx.doi.org/10.1021/es401555x | Environ. Sci. Technol. 2013, 47, 8733−8742