Today, there are no commercial membranes that can separate small molecules of similar size in the liquid phase based on their chemical properties. Such membranes would transform chemical manufacturing and significantly reduce energy costs associated with chemical separations. In this study, we aim to prepare synthetic polymer membranes that mimic two key features of biological pores such as ion channels and porins: Constricted pores only 1-5 nm in diameter that confine permeation, lined with functional groups lining the pore that interact with the target during passage. As an initial system, we focused on membranes capable of charge-based separation through electrostatic interactions. To build these nanostructured layers using scalable techniques, we deposited packed arrays of polymer micelles whose coronas exhibit carboxylic acid groups onto a porous support membrane. Random copolymers that combine highly hydrophobic fluorinated repeat units of trifluoroethyl methacrylate (TFEMA) with ionizable repeat units of methacrylic acid (MAA) form micelles and vesicles in methanol. When these micelles are coated onto the surface of a porous support membrane whose pores are smaller than the micelles and then immersed into water, a selective layer of micelles packed together is formed. The gaps between the micelles act as nanochannels functionalized with carboxylic acid groups. The membrane showed charge-based selectivity between organic molecules, effectively rejecting negatively charged solutes while allowing the passage of neutral solutes in two-solute separation tests. Interestingly, selectivity was enhanced when a mixture of solutes was used, likely due to competition between solutes to enter the very narrow pores.
To expand the types of separations these membranes can address, the carboxyl groups can be post-functionalized. This can enable the customization of these membranes membrane for desired separations. For instance, pi-pi interactions between aromatic groups can enable us to create membranes that preferentially interact with aromatic solutes. Such functionalization has enabled us to impart significant diffusion selectivity of ~10 between two hormones of identical size and charge, differentiated by the presence of an aromatic ring. This demonstrates the potential of using polymer self-assembly and functionality to design membranes that mimic biological pores while maintaining scalable manufacturing methods. We believe these approaches will eventually lead to novel membranes that can separate molecules of similar size but different chemical structure.
