The growing interest in industrial separation technologies, recycling and the need of waste water treatment in water stressed areas makes membrane separation a competitive and energy economic alternative to thermally driven separation processes. The development of asymmetric reverse osmosis membranes in the early 1960s and their commercial implementation in the 1970s paved the way to the first polymer gas separation membranes which were commercialized in the 1980s. Despite their successful industrial application, existing membranes suffer from a trade-off between permeability and selectivity as indicated by the Robeson plot. Inorganic membranes show the advantage of good permselectivity but suffer from challenging large-scale defect free fabrication and possible crack formation during operation.
In an attempt to bundle the benefits of polymer based and inorganic membranes, mixed matrix membranes (MMM) in which inorganic fillers are incorporated in a polymer matrix were developed. The prospect of good processibility, durability and improved selective separation performance fuelled research on MMMs from the mid 1980s . Importantly, to ensure enhanced permselectivity, gas-transport through the filler particles rather than through the polymer membrane must be facilitated. This sets restrictions on the design of the particle-matrix interface of these membranes, in which the filler particles must be dispersed homogeneously. Finding inorganic filler particles which are compatible with the organic polymer matrix can therefore be a challenging task.
With the development of porous organic cages (POC) in 2009  a purely organic filler with structural flexibility and solubility became available to replace the inorganic filler in MMMs. First attempts to fabricate POC loaded MMMs resulted in a significant increase in permeability but without noteworthy enhancement in selectivity. Aggregation and defect voids at the particle-polymer interface, where gas molecules can penetrate through the membrane, were identified as the reason for poor selectivity.
In their recent communication published in Angewandte Chemie, Ryan Lively and co-workers bypass the interfacial void formation of POC MMMs by using amorphous scrambled porous organic cages (ASPOC). As shown previously, POCs can undergo covalent scrambling reactions, which desymmetrize the cage structure and prevents them from crystallizing. The increased solubility of these cages facilitates solution-based processing and allows truly molecular mixing of the filler with Matrimid as the polymer matrix.
Lively and co-workers first demonstrate that symmetric cages formed by four triformylbenzene units connected by six cyclohexane linkers (CC3-R) form crystals in the polymer matrix. Exchange of four cyclohexane linkers with ethylene linkers (CC1432) allows homogeneous dispersion of the cages up to a filler loading of 20 %, as evidenced by EDX and Raman mapping. With positron annihilation lifetime spectroscopy the authors show that the pore size of the membrane matches well the void size of the cages, making a selective separation feasible.
Having a membrane with a smooth particle-matrix interface but discrete pores at hand, the authors investigate the membrane separation performance and demonstrate drastic increase of CO2 permeability by 340 % compared to the pristine Matrimid membrane.
While H2, N2 and CO2 pass the membrane almost equally well, a N2/SF6 gas mixture was successfully separated with a 2-3 fold increased selectivity because the large kinetic diameter of SF6 surpasses the pore limiting envelope of the POCs and prevents SF6 from passing through the membrane.
However, using cyclohexane abundant CC11_35 cages, in which only one out of six cyclohexane linkers is exchanged for an ethylene linker, as well as increasing the filler loading results in decreased selectivity, supposedly because of interfacial defects as result of irregular cage packing.
As proof of principle, the authors applied the membrane in organic solvent nanofiltration (OSN), a technique commonly found in industrial processes, which allows separation of smaller molecules from larger ones based on a size sieving effect. Successfully, polystyrene oligomers in a size range of 200 – 2000 Da could be separated from a methanol solution with a rejection rate of 95%. Using the composite membrane in a thin film configuration could even increase the rejection rate to essentially 100% in methanol, ethanol and toluene solutions.
The original article can be found here:
Zhu, G., Zhang, F., Rivera, M. P., Hu, X., Zhang, G., Jones, C. W., & Lively, R. P. Molecularly Mixed Composite Membranes for Advanced Separation. Angew. Chem. Int. Ed. 58, 2638-2643 (2019)
 Tozawa, T. et al. Porous Organic Cages. Nat. Mater. 8, 973-978 (2009)
 Jiang, S., Jones, J., Hasell, T., Blythe, C. E., Adams, D. J., Trewin, A., Cooper, A. I. Porous organic molecular solids by dynamic covalent scrambling. Nat. Commun. 2, 207 (2011)