Method for the preparation of cellulose acetate flat sheet composite membranes for forward osmosis—Desalination using MgSO₄ draw solution

Abstract

A lab scale method for the preparation of defect free flat sheet composite membranes for forward osmosis (FO) has been developed. Membranes containing a thin layer of cellulose acetate (CA) cast on a nylon fabric of 50 μm thick were prepared by phase inversion in water. Cellulose acetate (CA) membranes with an overall thickness of 70–80 μm have been prepared with lactic acid, maleic acid and zinc chloride as pore forming agents, at different annealing temperatures, for forward osmosis. These membranes have been tested in the desalination of saline feeds (35 g·L^− 1 of NaCl) using magnesium sulphate solution (150 g·L^− 1) as the draw solution. The water flux, and rejection of NaCl, were compared with those of commercially available membranes tested under the same FO conditions. The commercially available FO membrane from Hydration Technologies Inc, OR (M1) has a permeability of 0.13 L·h^− 1·m^− 2·bar^− 1 with a NaCl rejection of 97% when tested with 150 g·L^− 1 of MgSO₄ in the draw solution. Another commercially available membrane for FO from Hydration Technologies Inc, OR, M2 has a water permeability of 0.014 L·h^− 1·m^− 2·bar^− 1 with NaCl rejection of 100%. The flux and rejection of the CA membranes prepared in this work are found to be dependent on the nature of the pore forming agent, and annealing temperature. Impregnation of an inorganic filler, sodium montmorrillonite in CA membranes and coating of CA membranes with hydrophilic PVA did not enhance the flux of base CA membranes. Cellulose acetate membranes cast from dope solutions containing acetone/isopropanol and lactic acid, maleic acid and zinc chloride as pore forming agents have water permeabilities of 0.13, 0.09 and 0.68 L·h^− 1·m^− 2·bar^− 1 respectively, with NaCl rejections of 97.7, 99.3 and 88% when annealed at 50 °C. CA membranes prepared with zinc chloride as a pore forming agent have good permeability of 0.27 L·h^− 1·m^− 2·bar^− 1 with a NaCl rejection of 95% when annealed at 70 °C.

Introduction

In desalination processes, Reverse Osmosis (RO) is the dominant technology. RO is a membrane separation process that recovers pure water from an impure or saline water feed by pressurizing it to a level above its osmotic pressure [1]. The membrane rejects the salt ions from the pressurized solution, allowing only the water to pass. Thin film composite (TFC) membranes fabricated via interfacial polymerization (IP) have been proved to be the most successful RO membranes [2], [3]. Recent advances in RO include the incorporation of zeolite nanoparticles into the RO membranes to provide additional flow channels through the membrane [4]. Further speculative experimental and computational work utilizing molecular dynamics has shown that oriented carbon nanotubes can provide excellent RO flux and selectivity [5]. Major concerns expressed in the desalination of water using RO are (1) high applied hydraulic pressures to create the driving force for osmosis results in high energy consumption, (2) environmental problems in the discharge of concentrated brine, (3) operational costs in the replacement of membrane. For any new desalination technology to be commercially viable it must offer significant improvements over RO in at least one of many performance measures. These measures include energy costs, capital costs, water recovery rates, operability and maintenance requirements, water quality, and the product water cost. Less expensive processes are to be designed to make desalination technologies more competitive.

Forward Osmosis (FO) processes which operates at low or no pressure may serve as alternative processes to RO to desalinate water at reduced cost. FO uses a semi-permeable membrane to separate water from dissolved solutes similar to reverse osmosis process [6]. The semi-permeable membrane acts as a barrier that allows small molecules such as water to pass through while blocking larger molecules. In this process, a concentrated solution having higher osmotic pressure (draw solution) than the feed is circulated on the permeate side of the membrane. The osmotic pressure gradient across the membranes acts as the driving force for the transport of pure water through the membrane. FO operates at low pressures, gives high rejection of dissolved solutes and offers low fouling propensity [7]. The main challenges in the FO process are the selection of a suitable osmotic agent to prepare draw solution and preparation of effective membranes which offer high water flux, high salt rejection with no internal concentration polarization. Another area of concern in the FO process is the recovery of pure water from draw solution which depends upon the performance of the membranes and draw solute used.

In the early studies on FO process, researchers used RO membranes from Eastman and cellulose acetate hollow fibre membranes from Dow to desalinate seawater using glucose as the draw solution [8]. Goosens and Van-Haute [9] used cellulose acetate membranes reinforced with mineral fillers such as bentonite-38 syloid-161, syloid-244, Aerosil-0, aluminium oxide-L, aluminium to evaluate whether membrane performance under RO conditions can be predicted through FO testing. Mehta and Loeb [10], [11] used DuPont B-9 and B-10 Permasep RO membranes made of an aromatic polyamide. These membranes have a thin separating layer supported by a highly porous sub layers. The presence of highly porous sub layers lead to internal concentration polarization (ICP) which causes a reduction in the effective osmotic driving force across the membrane, resulting in a reduction of fluxes [11], [12]. Kravath et al. [8] observed improved salt rejections in desalination using hollow fibre cellulose acetate membranes rather than flat sheet cellulose acetate membranes designed for RO applications.

Current research on FO is focussed on (i) the preparation of hollow fibre membranes for FO processes [13], [14], [15] and (ii) studies on the effect of various osmotic agents on the water flux and salt rejection [16], [17], [18], [19]. Studies on using commercial RO membranes containing non woven fabrics indicated that non woven fabric materials prevent the exposure of the membrane surface to water, thereby resulting in a marked decrease in water flux [20]. In order to minimise the resistance offered by the backing material, membranes for FO applications are being prepared in hollow fibre configuration. Hollow fibre membranes made of polybenzimidazole and cellulose acetate are reported for FO applications [13], [14], [15].

Hydration Technologies Inc. (HTI) developed flat sheet membranes using cellulose esters exclusively for FO applications using a special casting procedure [21]. A cross-sectional SEM image of the membrane [17] shows that the polyester mesh which offers mechanical support is embedded in the polymer layer. It is also used successfully in commercial applications of water purification for military, emergency relief [22]. The major contributing factors for the success of this membrane are the relative thinness of the membrane and the internalisation of fabric support layer. Various draw solutes and their recovery have been reported in literature, but only few of them could be considered reasonable in terms of recovery and energy requirements [6]. Keeping in view of the high fluxes and retention of solutes in nanofiltration separation process, nanofiltration (NF) might reduce the energy consumption in the recovery of pure water. The recovery of pure water, reuse of draw solution and the energy consumption in the combined FO and NF process will decide if FO can be an alternative to RO or not. However, due to the low permeability of the current FO membranes the FO–NF process requires far greater membrane area than RO, but at the same time the equipment cost will not be as high because FO is carried out at atmospheric pressure. Membranes offering high mechanical strength, fluxes, salt rejections and low ICP are crucial for the commercial success of FO. Research on the preparation of membranes for FO applications is limited to hollow fibre membranes due to constraints in the casting process.

In this manuscript, we describe a lab scale method for the preparation of FO membranes on woven fabric material. For the development of a reliable and reproducible membrane casting method, dope composition reported earlier comprising CA/dioxane, methanol/acetone/lactic acid has been used initially [21]. Subsequently, polymer dope compositions were developed based on CA, acetone, IPA and various pore forming agents. The work reports the effect of polymer concentration, effect of different dope compositions, annealing and pore forming agents on flux and salt rejection in FO process. The effect of inorganic fillers and hydrophilic polymer coatings on flux and salt rejection were also reported. Membranes are tested in FO processes using magnesium sulphate as the draw solute which can be separated by nanofiltration for reuse.