Linking the Microstructural and Separation Properties of Electrically Tuneable Polyaniline Pressure Filtration Membranes

Abstract

The overall objective of this research was to determine if and how a conducting polymer membrane from polyaniline (PANI) can be made to change its separation selectivity in a pressure filtration of a neutral solute through applying an external electrical potential. This was done by determining the relationships between formation variables and the membrane properties used in pressure filtrations. Thus, in this research, firstly PANI membranes were prepared by coating composite polymerisation on PVDF support via vapour, solution and diffusion cell polymerisation techniques. To obtain a membrane that was conductive and within the molecular weight cut-off (MWCO) range of nanofiltration (NF) membranes, a continuous PANI coating on PVDF was required, making the amount of PANI deposited a key optimising parameter. The most PANI coated onto the support was in the order: diffusion cell polymerisation > solution polymerisation > vapour polymerisation. The PANI percentage was: 120 mass% > 13 mass% > 5 mass% respectively under the set of conditions studied. For all membranes, conductivity increased with PANI mass%. A modified dynamic contact angle technique showed that the membranes were electrically tuneable: the water permeation rates through the membranes increased when a potential difference was applied. The potential driven permeation was found to be independent of the initial membrane porous structure (SEM imaged), indicating that the increased rates were largely related to applied potential driven surface energy changes. A complete surface coverage of PANI coated membranes (around 30 to 80 mass%) obtained via the diffusion cell technique were taken forward to pressure filtration testing. PANI membranes were also fabricated from in-house PANI synthesis followed by phase inversion via immersion precipitation. The membrane formation variables (e.g. polymer concentration, casting thickness, heat treatment and type of acid dopant) were linked to the micro-structural properties to find the best NF membrane with the largest tuneability. The PANI membranes all had a typical integrally skinned layered asymmetric membrane structure. The addition of different dopants in PANI membranes changed the thickness of these layers, with the HCl and ASA doped membranes having almost the same structure; these were the best membranes for comparison. Initial tuneable assessment again using the dynamic contact angle technique for all of the membranes doped in different dopants showed that the ASA and HCl doped membranes have the greatest tuneability as the rate of effective contact angle and peak water droplet height tested under 7 V applied voltage was more than three times higher than the rates found in other membranes. This all demonstrated that bigger dopant size/molecular weight and also greater membrane conductivity (resulting from the dopant) have the greatest influence on permeability and tuneability. A rapid, reliable and cheap membrane characterisation method for membrane flux and molecular weight cut off (MWCO) was developed and the MWCO further determined by performing a single run of NF and/or low UF with polyethylene glycols (PEGs) mixture in aqueous solution. Analysis and detection via a low temperature evaporative light scanning detection (ELSD) method in HPLC has resolved peaks to a MW difference of just 44 g mol−1(corresponding to the CH2–O–CH2structural unit difference between the PEG oligomers), allowing the most precise ever one-filtration MWCO determination. MWCO testing of commercial membranes (from Koch, FilmtecTMand Hydranautics) confirmed the method gives MWCO in the expected range. Finally, the tuneability of the best diffusion cell and phase inversion membranes was characterised in a custom-made electrically connected cross-flow filtration rig, using the PEG MWCO method developed. The MWCO of PANI membranes doped in ASA and HCl had the lowest MWCO with no potential applied (< 2000 g mol-1). The MWCO values increased under applied voltage, with the ASA doped PANI showing the greatest increase at 7 V applied (400 – 500 ± 50 g mol-1). This is thought to be due to the dopants size, producing a greater free volume change when it rearranges under the applied potential. However very little tuneability was found for the MA and PMVEA doped membranes. This is thought to be due to the smaller free volume available in the membranes and lower membrane conductivity compared to the ASA and HCl doped membranes. The membranes were tested at lower applied potentials (1 V and 3.5 V), but little tuneability was observed, indicating that there was a large resistance across the membranes formed. Other effects such as the thickness of the top layer cross-section, dopant size, shape and MW and also electrical conductivity may also influence the MWCO changes. Overall, it has been demonstrated that PANI membranes either prepared by composite coating polymerisation of PANI onto non-conducting PVDF or by phase inversion in addition of different dopants, can have their permeability for neutral species electrically tuned under cross-flow conditions. Since PANI can be easily coated/cast onto existing type membrane supports and applied in already commercially available modules, this indicates that these membranes could be used to externally control the separation properties in practical and economically feasible pressure driven membrane processes and applications.