(225d) Inferring Diffusion Coefficients from Break-through-Curve Measurements Under Uncertainty

With the goal of reducing risk in scale-up of adsorption processes, we infer diffusion coefficients from break-through curve measurements by leveraging techniques from mathematical modeling, numerical optimization, and uncertainty quantification. Considering CO₂ diluted in He and adsorbing onto Zeolite 13X, we model the dynamics by isothermal plug flow with adsorption at a linear-driving-force rate. With thermodynamically consistent interpolation of equilibrium data, we do not restrict the isotherm to a specific functional form. In solving a numerical optimization problem dependent on multiple measurements and constrained by partial differential equations, we calculate diffusion coefficients, and additional challenging-to-measure parameters, often estimated from correlations. By quantifying parameter uncertainties and propagating their influence on the measurement, we determine whether the linear driving-force has a statistically significant change with partial pressure.