Effect of pH on the performance of polyamide/polyacrylonitrile based thin film composite membranes

Abstract

In this study the effect of pH on the performance of thin film composite (TFC) nanofiltration (NF) membranes has been investigated at the relevant pH conditions, in the range of pH 1–13. TFC polyamide NF membranes have been fabricated on a polyacrylonitrile support via interfacial polymerization between piperazine in an aqueous phase and trimesoyl chloride in an organic phase. Membrane characterization has revealed that the produced membranes show a NaCl retention similar to NF-270 and Desal-5DK, a permeance in between those of NF-270 and Desal-5DK, and a slightly higher iso-electric point than NF-270 and Desal-5DK. The molecular weight cut-off of the membranes appeared to be practically constant in acidic and neutral conditions. At extremely alkaline conditions (pH > 11) an increase in molecular weight cut-off and a reduction in membrane flux has been observed. According to the Donnan steric partitioning pore model (DSPM) the change in performance in alkaline conditions originates from a larger effective average pore size and a larger effective membrane thickness as compared to the other pH conditions.

Introduction

Thin film composite (TFC) membranes prepared via interfacial polymerization (IP) have often been used for nanofiltration (NF) due to their superior permeation performance [1]. Although most commercial TFC NF membranes are suitable for treating aqueous streams at pH levels between 2 and 10, many potential applications in the chemical industry involve much more aggressive conditions [2], [3]. Advances in development of stable membranes and their performance characterization in these harsh applications will therefore expand the application window of such membranes in commercial processes.

Several recipes to fabricate polyamide (PA) TFC membranes via IP have been published in literature [2], [4], [5], [6], [7], [8]. Many commercial NF membrane manufacturers seem to favor poly(piperazine-amide) based recipes [2], [3], [9], probably due to a superior chlorine resistance of these membranes [4]. Materials like polysulfone [10], [11], polyacrylonitrile [12], [13], polyethersulfone [14], polypropylene [15], poly(tetrafluoroethylene) [16] have been reported as support membranes for the IP process. Little is known about the performance of these membranes at extreme pH conditions.

In this work composite nanofiltration membranes have been fabricated using an aliphatic amine piperazine (PIP) and acyl chloride trimesoyl (TMC) chloride with a controlled fabrication technique. Ultrafiltration polyacrylonitrile (PAN) membranes (type HV3), supplied by GKSS-Forschungszentrum Geesthacht GmbH (Germany), which have been reported to encompass reasonable chemical stability [17], have been used as support membranes for the IP process. The performance of the membranes has been evaluated in a pH range from 1 to 13, using polyethylene glycol (PEG) [18] and glucose molecules, using a batch filtration setup. The performance has been compared to the performance of commercial nanofiltration membranes at neutral pH.

Changes in the performance of the membranes have been related to structural changes using the Donnan steric partitioning pore model (DSPM) developed by Bowen and coworkers [19], [20]. This model is based on the extended Nernst–Plank equation, and has often been used to characterize commercial NF membranes [21], [22], [23], [24], [25], [26]. Other researchers have also successfully characterized PA TFC membranes using the DSPM [27], [28], [29]. Ahmad et al. have used the model to analyze the effect of the fabrication parameters of PA TFC membranes prepared via the IP route [30], [31], [32]. However, the effect of feed pH on the membrane performance is rarely studied. Labbez and coworkers have studied the effect of pH on the volume charge of a ceramic titania membrane, which encompasses a fixed pore radius [33]. Mänttäri and coworkers have reported the effect of pH on the filtration performance of commercial nanofiltration membranes using a mixture of glucose and NaCl [34]. Kaya [35] and Nanda [36] have studied the change in the fouling potential of commercial membranes with a change in the feed pH. The present work analyzes the retention of uncharged molecules for IP membranes on PAN support in the absence of salt in a range of feed pH. The absence of salt limits the disturbance in retention of an uncharged molecule due to the presence of charged species in the feed [37] mixture. DSPM has been used to determine the effective pore radius and the effective thickness/porosity ratio of a PA TFC membrane prepared via the IP route as function of pH.