In a joint effort between Prof. Nunes and Prof. Szekely, we collaborated on a project to push the boundaries of membrane selectivity beyond what is currently available. Our groups combined expertise in membrane fabrication and performance analysis with the shared goal of developing highly selective macrocycle membranes that could handle complex separations, especially in pharmaceutical and chemical applications.

We focused on crown ethers, a class of macrocycles known for their ability to differentiate between ions and molecules. In tailoring thin-film composite membranes, crown ethers have been used as additives to enhance permselectivity in water purification and ion separations. In our approach, we integrated crown ether building blocks directly into the polymer backbone to fully utilize their unique structure. These highly selective membranes were tested in the challenging organic solvent nanofiltration (OSN) and organic solvent reverse osmosis (OSRO) applications to effectively separate drug molecules from impurities and solvents.

We employed amino-functionalized crown ethers in the aqueous phase of the interfacial polymerization reaction, forming a fully crosslinked polymer network on top of a porous support (Fig. 1). This strategy increased the macrocycle density within the film, allowing us to fully harness their selective properties. Dibenzo-18-crown-6, in particular, has a cavity size of 0.26–0.32 nm, ideal for separations in the region between OSN and OSRO.

To better understand the impact of crown ethers on the separation performance, we followed a comprehensive framework (Fig. 2). We thoroughly assessed various factors, including:

• Material Perspective: We compared crown ether-based membranes with in-house fabricated control membranes and commercial OSN membranes. The comparison included testing membranes made from analogous polyether-based monomers (ethoxyaniline and isosorbide), which share similar chemistry with crown ethers but lack their intrinsic cavity. We also evaluated the performance of classical polyamide thin-film composite membranes and commercial OSN membranes, DuraMem^®150 and GMT-oNF-2, under the same testing conditions.
• Diverging from Standard Testing: To better simulate the complex conditions in pharmaceutical applications, we tested the membranes using a mixture containing 13 solutes with molecular weights below 850 g mol^–1 and analyzed the rejection using HPLC.
• Rejection Selectivity Analysis: We mapped size-normalized separation factors for all solute pairs using a selectivity matrix, offering valuable insight into the membranes’ selectivity beyond molecular size. We also examined the correlation between rejection selectivity and the rejection of less permeable solutes to compare the performance of different membranes. The rejection selectivity data exhibited diagonal gridlines that decreased as log(1–R_B) increased. A wider distribution of data points in the x–y plane indicated a greater capacity to differentiate solutes with similar characteristics, thereby increasing the overall rejection selectivity of the membrane.

Key Findings and Implications

Comparative analysis with benchmark membranes highlighted the unique performance attributed to the macrocyclic structure of crown ethers. These membranes demonstrated both higher solvent permeance and enhanced selectivity. The acetonitrile permeance improved by 93% compared to conventional counterparts. Additionally, they showed a particularly sharp rejection profile, outperforming commercially available nanofiltration membranes by 23–90% selectivity margins.

Mapping the rejection selectivity using a mixture of structurally diverse solutes revealed that crown ether-based membranes are highly effective at retaining pharmaceutical ingredients with molecular weights around 800 g mol^–1. This capability enables the efficient concentration of molecules in this size range or their purification from smaller impurities. Our research also highlighted the efficacy of membranes incorporating crown ethers in separating solute pairs with molecular weights between 100 and 370 g mol^–1. The broad distribution observed on the rejection selectivity grid suggested these membranes excel at separating solutes with similar physicochemical properties. In contrast, DuraMem^®150, while effective at rejecting small solutes, struggled to fractionate the mixture.

In summary, our study presents the potential of crown ether-based membranes to improve molecular separations in OSN and OSRO applications. The membranes demonstrated notable stability and selectivity for pharmaceutical ingredients. While further optimization is needed for large-scale implementation, the findings contribute to the development of selective membranes and present an interesting approach to characterizing membrane performance.