(360ax) Understanding the Normal Bicontinuous Cubic Phase in Gemini Lyotropic Liquid Crystals in Order to Design Selective Separations | AIChE

(360ax) Understanding the Normal Bicontinuous Cubic Phase in Gemini Lyotropic Liquid Crystals in Order to Design Selective Separations

Authors 

Schwindt, N. - Presenter, Rose-Hulman Institute of Technology
Sahu, S., University of Colorado
Shirts, M., University of Colorado Boulder
Gin, D. L., University of Colorado
Noble, R., University of Colorado
Lyotropic liquid crystal (LLC) monomers are promising materials for membrane separations due to their uniform, nanoporous structures and robust crosslinked networks. LLCs show significant variation in phase morphology, self-assembling into a number of highly ordered phases. The normal bicontinuous cubic (QI) phase is of particular interest for membrane separations due to its tortuous, annular pores that percolate in three dimensions. However, it is difficult to study the nanoscale structure and transport of LLC membranes experimentally.

Molecular simulation provides a route to fully explore these new membrane materials. The complexity of the QI phase and the long timescales required for self-assembly prevent direct atomistic simulation. However, we have developed a procedure to locate stable simulations of the QI phase, which includes knowledge of the ideal geometry, atomistic simulations of the lamellar bilayer, and coarse-grained models of the monomer. Studying how the atomistic model of the lamellar bilayer fits within the mathematical representations of the QI phase reveals the most likely structure is the gyroid Ia3d space group. We achieve a metastable coarse-grained model of the gyroid structure by tuning the model to match properties of the atomistic bilayer. We locate a stable atomistic gyroid structure by backmapping the coarse-grained model to an atomistic model. From the atomistic model, we characterize the local pore environment within the QI phase. Our simulations reveal a dynamic pore, ranging in size from 0.25 nm to 0.75 nm, which is consistent with molecular-size rejection seen experimentally. The pore consists of a percolating hydrophilic region without well-defined pore boundaries. The LLC monomers penetrate the solvent-filled region, which restricts the hydrophilic region to smaller pore diameters. Preliminary transport simulations of the crosslinked membrane show hindered water diffusion in the membrane similar to experiment.