The principle objectives of this work were fourfold: first, to quantify and control the suspension properties of a metal-organic framework (MOF), MIL-101; second, to use this knowledge to understand the formation of thin film nanocomposite (TFN) membranes made via interfacial polymerisation (IP); third, to identify other secondary filler effects; and fourth, to relate IP reaction conditions to filler incorporation.
Preliminary studies revealed that trimesoyl chloride (TMC) adsorption onto MIL-101 in n-hexane significantly affected membrane performance. More importantly, aggregation led to non-selective defects for TFN membranes synthesised with n-hexane. To control aggregation, the Hamaker constant of a MOF was determined for the first time and Derjaguin, Landau, Verwey, Overbeek (DLVO) theory was related to quantifiable dispersion measurements, showing that hydrophobic solvents like those used to make TFN membranes provide poor dispersion. Surface-selective modification of a MOF increased the concentration of well-dispersed MOF by over an order of magnitude, while the use of nitrobenzene increase it by over two orders.
Thus, nitrobenzene was used to fabricate TFN membranes via IP. These membranes demonstrated pure solvent permeance increases relative to the thin film composite membrane control of up to factors of 1.9 for methanol and 2.7 for water, while maintaining a dye rejection of ~90 % (MW ≈ 314 g·mol-1). Further exploration revealed that aggregates introduce non-selective defects in certain circumstances and semi-selective defects in others, the latter of which can enhance solvent permeance without major reductions in rejection. Additionally, MIL-101 adsorbed less than half of the TMC in nitrobenzene compared to that adsorbed in n-hexane, mitigating this secondary effect. Finally, membrane surface roughness was controlled and correlated with whether or not MIL-101 addition to the organic phase impacted membrane performance, with smoother membranes showing no performance differences and rougher membranes showing differences similar to those observed by previous researchers [1].