(605d) Building Reverse Osmosis Films One Monomer at a Time Via Molecular Layer Deposition

The conventional preparation of thin film composite (TFC) reverse osmosis membranes involves interfacial polymerization (IP) atop a porous, polymeric support. While commercially successful, IP is limited in its ability to control the selective layer composition, thickness, and morphology which hinders transport properties and operational lifetime. To overcome the limitations, we have developed TFC membranes using molecular layer deposition (MLD) rather than IP. This derivative of atomic layer deposition was used to create semipermeable polymer films with monomer-level control of composition, thickness, and chemistry. MLD is accomplished by exposing a substrate to monomer vapors one at a time in vacuum. The sequential method is self-limiting at each exposure, which allows for precise film growth. Key to fabrication was the ability to grow the MLD film across the pore openings of a porous substrate despite the conformal nature of MLD. We developed a sacrificial pore-filling process that enabled such growth atop a polyethersulfone ultrafiltration support layer.

Selective aromatic polyamide MLD films were made from m-phenylenediamine and trimesoyl chloride in vacuum at 120°C at 3 Å per MLD cycle. From ellipsometry and X-ray reflectivity results, we posit that the surface of the films were comprised of polymer tails up to two monomers in length and connected to a fully crosslinked bulk layer. The structure and surface of MLD and IP polyamide films were compared using X-ray photoelectron spectroscopy, infrared spectroscopy and atomic force microscopy. Layer-by-layer MLD synthesis resulted in properties favorable for salt rejection, water flux and fouling resistance: compared to IP, MLD films were dense, highly crosslinked, smooth, and contained reduced void volumes. In crossflow desalination tests, ultrathin MLD TFC membranes had NaCl rejection as high as 98% and permeance as high as 1.2 L/m²∙h∙bar. Tradeoff between permeance and rejection could be tuned by varying the number of MLD cycles and chemistry of the selective layer. This approach provides a competitive and novel route for fabricating desalination membranes. It enables the design of nanolayers for the optimization of permeance as well as exploration of degradation and fouling resistant materials. This work also has broad implications for other TFC applications, such as gas separations and nanofiltration.