The significant role of support layer solvent annealing in interfacial polymerization: The case of epoxide-based membranes

The applicability of state-of-the-art water purification membranes in harsh feed streams is limited due to their insufficient chemical robustness. Epoxide chemistry has been recently introduced to achieve pH- and chlorine-stable nanofiltration (NF). This study further investigated the influence of interfacial polymerization (IP) synthesis parameters on the resulting epoxide-based thin-film composite (TFC) membrane structure and performance. Epoxide polymerization could be initiated by N,N,N',N'-tetramethyl-1,6-hexanediamine and NaOH. Surprisingly, neither the type of initiator, nor the initiator concentration used during IP, influenced membrane rejection of rose bengal (RB) dye (1017 g mol(-1)), which was constant at similar to 90%. This consistent RB rejection was primarily determined by annealing of the cross-linked polyimide support by toluene, which is the solvent used during IP. In contrast, the poly(epoxyether) top-layer determined membrane selectivity for methyl orange, a smaller dye of 327 g mol(-1). The effect of solvent annealing of the membrane support by diethyl carbonate, dimethyl sulfoxide and m-xylene was also investigated and revealed that the changes induced by solvent contact are physical rather than chemical in nature. This study shows, for the first time, the substantial direct impact of the solvents used during IP to influence support properties and the resulting membrane performance. Solvent annealing can therefore be considered as a tool in membrane fabrication-during IP or as a post-treatment step-to further tune the separation performance for specific applications.