Predictive simulation of non-steady-state transport of gases through rubbery polymer membranes

Abstract

15We develop a multiscale, physically-based reaction-diffusion kinetics model for non-steady-state 16 transport of simple gases through a rubbery polymer. The rubbery polymer case applies to 17 membrane applications where high permeability is desired or has developed due exposure to 18 plasticizers such as CO2. To construct a model that has no adjustable parameters, we utilize 19 experimental data from the literature as well as new measurements of non-steady-state 20 permeation. We have also performed molecular dynamics simulations of collisions of CO2 with 21 the surface of PDMS to obtain an estimate of the gas-polymer sticking probability. The model 22 successfully reproduces time-dependent experimental data for two distinct systems: (1) O2 23 quenching of a phosphorescent dye embedded in poly(n-butyl(amino) thionylphosphazene), and 24 (2) O2, N2, CH4 and CO2 transport through poly(dimethyl siloxane). The modeling study 25 provides new insights to microscopic aspects of permeant transport through rubbery polymers 26 and is used to predict selectivity targets for two gas separation applications. 27 28 2

Permeation of CO2 and N2 through glassy poly(dimethyl phenylene) oxide under steady‐ and presteady‐state conditions

“…Further details of the MD simulations can be found in our previous publication. 12 Additionally, the free energy of CO 2 and N 2 within the polymer was determined using molecular metadynamics simulations. Five gas molecules were inserted into the simulation box, to produce a gas pressure of 2 atm; when all 5 gas molecules are sorbed into the polymer, this produces a concentration of 0.0404 mol/L.…”

“…62 For this reason, we begin with the model framework from a study of permeation of gases through a rubbery polymer. 12 We then expand upon that model by including dynamic changes in volume and the pressure- 3B. Numerical Procedure The reaction-diffusion scheme is solved using a stochastic method, [63][64] a type of kinetic Monte Carlo (kMC), implemented in the open access package Kinetiscope.…”

“…The reaction-diffusion models are set up to allow for direct comparison of simulation predictions to experimental data on downstream pressure rise, as in our previous work. 12 A general schematic of the reaction-diffusion system is shown in Figure 2. to the membrane permeation system.…”

“…The standard application of the solution-diffusion model is as a steadystate model, intended to describe situations where the membrane properties and external conditions are constant. In cases where the polymer properties change upon initial exposure to a permeant or where the external permeation concentration is changing (e.g., produced in systems driven by intermittent, renewable energy sources such as sunlight or wind [4][5][11][12] ) such non-steady state permeation cannot be predicted in a mechanistic way by the solution-diffusion model. In contrast, physically-based, mechanistic descriptions of permeation that capture time-dependent physical and chemical processes will provide computational frameworks that are predictive, afford greater scientific insight across length and timescales, and apply to a wider range of permeation conditions.…”

“…In contrast, physically-based, mechanistic descriptions of permeation that capture time-dependent physical and chemical processes will provide computational frameworks that are predictive, afford greater scientific insight across length and timescales, and apply to a wider range of permeation conditions. In previous studies, [11][12] we have reported such descriptions for gas sorption and permeation involving rubbery polymer membranes and sorption of aqueous solutions of methanol by Nafion. Those works allowed a multiscale simulation framework for transport of weakly and strongly interacting permeants to be developed.…”