People use many chemical products in their day-to-day lives. Medicines, makeup, fuel, glues, and plastics are all chemicals. Making these products requires separating mixtures of many chemicals into single streams of pure chemicals. About one tenth of the world's energy use goes to purifying chemicals. Purifying chemicals with membranes could lower this energy use by ninety percent. Membranes are very thin materials with very small pores that separate chemicals by size and type. Unfortunately, current membrane materials do not work for many purification processes. Plastic membranes are easy to make in large amounts but dissolve in many chemical mixtures. Inorganic membranes are hard to make in large amounts but are stable in most chemical mixtures. This project explores a new way to add inorganics to plastic membranes. These new membranes are easy to make in large amounts and more stable in chemical mixtures. Because these inorganics can be added to plastics in many ways, trial-and-error testing is impractical. Instead, this project uses computational simulations and data analytics to speed up discovery of the best membranes. Students on this project will be trained in membrane science, computational simulations, big data science, and materials testing. This project also supports computational materials science in Georgia Tech's open-access, student-run Materials Innovation and Learning Lab (The MILL). The MILL provides students with free access to materials research tools. Hundreds of students get trained on these research tools each year.
This research project will expand a recently discovered class of hybrid membrane materials created via vapor phase infiltration (VPI), a gas-phase process that infuses polymers with inorganic constituents intermixed at the atomic level. These hybrid membranes show dramatically enhanced stability in organic solvents while retaining salient membrane properties of high permeance and discerning selectivity. Because the design space for such membranes - including polymer chemistry, inorganic chemistry, and hybrid microstructure - is enormous, traditional Edisonian-based materials development methods are impractical. To address this challenge, the research team combines expertise in: (1) phenomenological theory of VPI materials synthesis, (2) membrane and separation science, (3) materials simulations and data-driven design, and (4) advanced statistical algorithms that incorporate known phenomenological physics with limited initial data. Efforts in each of these areas will rapidly steer the search towards chemical, morphological, and processing spaces of opportunity. Specifically, this project will focus on the design of materials synthesis processes for targeted membrane chemistries and microstructures. The outcomes of this research will be (1) the creation of tangible hybrid membranes based on polymers of intrinsic micro-porosity with superior performance and stability, (2) the identification of key physiochemical descriptors for controlling structure and performance in these hybrid membranes, and (3) the development of new strategies for handling data sparsity and physical phenomena integration into materials informatics-based design.
This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.