(167i) Performance of Thermally Rearranged Polymers for Olefin/Paraffin Separation from All-Atom Molecular Dynamics Simulations

The separation of olefins and paraffins is an important process in the petrochemical industry. This process is currently driven by cryogenic distillation, which requires a massive energy input. Combining membranes with distillation for olefin purification has the potential to significantly reduce energy consumption. Thermally rearranged polymers (TRP) are high performing, amorphous polymer membranes and have shown great potential in separating hydrocarbons with small differences in their molecular sizes such as olefin/paraffin mixtures. Free volume elements (FVEs) – the void spaces that arise from the inefficient packing of bulky groups on the polymer chain – are an important feature of membranes, and their distribution can be tuned in TRP to improve overall performance. When polymer membranes operate at high pressures, penetrant molecules adsorb on the polymer matrix, making the chains more mobile and easily deformable. This phenomenon is known as plasticization and has prevented the widespread use of polymer membranes in the industry. TRPs are inherently rigid, and their rigidity can be enhanced using crosslinking to mitigate the effect of plasticization. Membrane performance is usually governed by several design parameters that are challenging to control during synthesis. While current membrane transport theories explain experimental results, they often adopt macro-scale assumptions to describe molecular-level phenomena. Obtaining a deeper molecular understanding will enable us to design superior TRP membranes to use for olefin/paraffin mixtures.

In this study, we use molecular dynamics (MD) simulations to predict the permeability of different TRP structures with ethane/ethylene and propane/propylene mixtures. Starting with thermally rearranged HAB-6FDA, we apply crosslinking and chemical functionalization to increase the rigidity and microporosity respectively. Crosslinking density is increased from 0%-50% at 10% increments. Fluorination and spiro-center groups are added to the HAB-6FDA TR backbone to increase the microporosity. Two non-equilibrium molecular dynamics (NEMD) methods are used to simulate the permeation process, namely moving wall and concentration gradient methods. In the moving wall method, two impermeable graphene walls are constructed on the two ends of the simulation box. A large and small force are applied on the two walls to create a pressure gradient. In the concentration gradient, the force is applied on the penetrant molecules located in a small defined region (slab) of the simulation box. The results of the two approaches are compared against experimental techniques. This work elucidates the effect of crosslinking and chemical functionalization on the overall performance of TRP membranes and will guide the design and synthesis of more efficient olefin/paraffin membranes.