(694d) Nanoscale Shape-Morphing in Polyamide Membranes Enabled By 3D Nanoscale Imaging-Analysis Platform

State-of-the-art reverse osmosis, nanofiltration, and desalination membranes are constructed as a thin film composite: they consist of crosslinked polyamide selective layer with an extremely thin thickness (i.e., layer thickness of about 10 nm and apparent thickness of >100 nm). These thin polyamide membranes are generally prepared via interfacial polymerization between aromatic or aliphatic diamine and aromatic acyl chloride. In general, their performance of polyamide membranes is governed by their morphology. However, the interfacial polymerization process leads to highly irregular and complex crumpled nanostructures that make the quantitative measurement of their physiochemical properties extremely difficult. Here, we demonstrate quantitative nanoscale shape-morphing in polyamide membranes and its relation to the membrane’s mechanical heterogeneity by integrating 3D electron tomography, machine learning-based shape classification, and liquid-phase AFM. From diverse nanoscale structure prepared by varying synthesis parameters, quantitative morphometry extracts large datasets of 3D geometry descriptors including interconnectivity, domain, void architecture, layer thickness, surface curvature, surface area, assembly structure (more than 50 descriptors). A data mining algorithm is employed to evaluate their categorical and numerical attributes and to rank descriptors from the most to least informative to composite functionality. This newly created knowledge on morphological properties at the nanometer resolution is related back to bridge synthesis and functionality. The elucidation of the molecular underpinning of the synthesis–morphology–property relationship would enable a new prediction-based design of polymeric materials, advancing beyond previous “trial-and-error” approaches. We anticipate that this imaging−morphometry platform can be applicable to other nanoscale soft materials and provides engineering strategies based directly on synthesis−morphology−function relationships.