Machine learning teaches membranes to sort by chemical affinity
Ultrafiltration membranes, a cornerstone of pharmaceutical manufacturing and various industrial processes, have traditionally relied on separating molecules based on their size. However, a groundbreaking study by Cornell researchers introduces a paradigm shift in this field. The study, published in Nature Communications, unveils a novel approach to creating membranes that can filter molecules based on their chemical makeup, not just size.
The key to this innovation lies in the use of block copolymer micelles, tiny self-assembling polymer spheres. These micelles, when blended with chemically distinct block copolymers, can be engineered to create membranes with chemically diverse pore surfaces. This is a significant advancement, as it allows for the separation of molecules with identical sizes but different chemical structures, such as antibodies with distinct molecular structures.
The research, led by Ulrich Wiesner, the Spencer T. Olin Professor of Materials Science and Engineering, demonstrates how neutral and repulsive interactions among micelles influence their self-assembly within the top separation layer. By combining up to three distinct block copolymers, the team showed that these interactions control the distribution of different chemistries in the pores of the membrane's surface.
Despite the simplicity of the concept, its experimental realization is challenging. Lilly Tsaur, the lead author and a Ph.D. candidate in the Wiesner group, employed scanning electron microscopy to image hundreds of samples, studying the arrangement of different micelles. However, due to the difficulty in identifying chemistries through imaging, she utilized machine learning to detect subtle differences in pore patterns, pinpointing the locations of each micelle type.
Co-author Fernando A. Escobedo, the Samuel W. and M. Diane Bodman Professor of Chemical and Biomolecular Engineering, conducted molecular simulations to reveal the rules governing micelle self-organization. This was a complex task due to the large number of micelles and their tendency to assemble in non-equilibrium states, requiring highly coarse-grained models and numerous calibrations.
This research builds upon the Wiesner group's previous advancements in block copolymer self-assembly, which led to the founding of Terapore Technologies, a startup focused on cost-effective UF membranes for virus filtration. The new study opens up exciting possibilities for companies to adapt the same manufacturing process to produce membranes capable of affinity separations based on programmed pore surface chemistry.
Wiesner envisions a paradigm shift in UF-based operations, enabling the creation of membranes with chemically diverse pore surfaces by simply adjusting the recipe. This innovation has the potential to revolutionize filtration processes and open up new avenues for the application of UF membranes, including smart coatings and biosensors.
The research was supported by the National Science Foundation and facilitated by the Cornell Materials Research Science and Engineering Center and the Cornell Nuclear Magnetic Resonance Facility. The study's findings were published in Nature Communications, marking a significant step forward in the field of membrane technology.
