Fabrication and characterization of support layer for thin film nanocomposite desalination membranes

The production of potable water has become a global concern; for many communities, conventional available water resources are inadequate in supplying the demand due to rapid population growth and industrialization. Desalination of brackish or seawater has become an important source of both drinking water production and irrigation. The leading technology is reverse osmosis (RO) membrane process and is accounted for producing more than half of the world's desalination capacity. Significant development in membrane materials and technologies; in terms of module and process design, feed pre-treatment and energy recovery in the past few decades have greatly improved the cost effectiveness and performance capability of membrane processes and have made the RO technology dominating in the desalination market. Owing to showing inherently high salt rejection and durability with small footprint and ease of the operation, polymeric membranes have governed the RO membrane industry. On the other hand, there are several severe problems need to be eliminated such as comparatively low permeability due to their characteristic hyrophobicities, low boron removal and fouling issues. In order to resolve aforementioned problems, different approaches were developed. Among them, thin film composite (TFC) membranes have drawn substantial attention since their invention by Cadotte et al. in the 1970s in which a thin selective polyamide (PA) layer is formed on top of a microporous support by in-situ interfacial polymerization (IP) technique and these membranes have high salt rejection values over 99%. Although, commercially available PA-TFC membranes can remove almost all ions in the seawater, their boron removal values are around 95% under normal test conditions. Recently, a new concept of mixed-matrix composite RO membranes was reported by incorporation of nanoparticles such as carbon nanotubes (CNTs) and zeolites into the thin selective PA barrier in the course of IP process. Both experimental and molecular simulation studies have shown that the use of CNTs in PA layer resulted with enhanced water flux and improved salt rejection values compared with conventional counterparts that are mainly because of their preeminent characteristics, and ease of functionalization with various functional groups. This thesis aims to develop a novel fabrication method of thin film nanocomposite (TFN) membrane for RO technology in which single-walled CNTs are embedded in PA selective layer during IP process on top of a microporous polysulfone (PSf) substrate with the purpose of increasing the boron rejection capacity and improving the water flux as well. Experiments were performed at three stages: i) preparation of porous polysulfone (PSf) substrate by phase inversion method and optimization of the process parameters such as addition of pore-former agents, changing the type of non-solvent and modifying the temperature of precipitation medium, ii) the development of TFC membranes on PSf supports having different pore sizes and optimization of the IP process conditions such as altering the monomer concentrations and reaction time, iii) fabrication of TFN membranes by means of filtration of commercially available –COOH functionalized SWCNTs on PSf support prior to IP process for the semi-alignment of the nanotubes. For the characterization of the resultant membranes field emission scanning electron microscopy (FESEM) and contact angle measurements were employed to investigate both the surface and cross-sectional morphology, and surface hydrophilicity, respectively. Performance evaluations of the fabricated membranes were conducted using a laboratory scale cross-flow membrane test unit. According to the results, the ideal PSf substrate membrane for the incorporation of SWCNT into PA layer was obtained from the recipe of 15 wt.% PSf with the addition of 10 wt.% pore-forming agent which was polyvinylpyrrolidone (PVP) and precipitated in cold isopropyl alcohol. On the other hand, regarding to TFC results; SEM micrographs have shown that pore size of the PSf support has great impact on the resultant PA layer. Relatively smaller pores produced thicker PA layer due to the limited diffusion of aqueous m-phenylenediamine (MPD) solution into the small pore channels, whereas larger pores produced thinner PA layer mainly because of ability of MPD solution to penetrate deep into the pore channels and at the same time these big pores act like a MPD reservoir in which subsequent trimesoyl chloride (TMC) introduction will produce PA layer. Other than characteristics of PSf substrate, the PA layer thickness was affected by the interfacial polymerization reaction time in such a way that longer the polymerization time is applied, the thicker the cross-linked PA layer is formed. The 15 s reaction time produced around 100 nm PA layer, while doubling the reaction time from 15 s to 30 s formed approximately 1 µm-thick PA barrier. In addition to these, the effect of MPD concentration was also studied. Increasing the amount of MPD can caused to form thicker layer than its low concentration counterpart produced. All the results were consistent with the literature. In order to test the fabricated membranes in the testing cell having 140 cm2 test area, various masking techniques with Al tape and foil were tried and finally the most proper method has been developed. This newly developed technique enables to test the performance of the membranes having smaller areas than the test cell. Technique was tested with a membrane having a known specification. Results showed that the developed method works successfully without any leaking. In this regard, one of the fabricated membranes TFC-5 was tested in order to determine the performance of synthesized membranes so that to reflect a representative result. Salt rejection was calculated by measuring conductivity of the solutions before and after, which were feed and permeate, respectively. The rejection value was calculated as ~85% at 2000 ppm NaCl, which was smaller than the commercial membranes. These results have shown that the fabricated membranes can be improved in terms of fabrication technique to obtain better results. Nevertheless, this study contains noteworthy clues with respect to membrane masking and is still a guidance for the fabrication of substrate membranes for high boron rejection TFN membranes that can be used for desalination applications.