Experimental Measurement of Pure and Mixed-Gas Kinetics of Ethylene and Ethane Adsorption onto Mordenite, Zeolite 13X, and ZJU-74a: Insights into the Value of Single- vs Mixed-Gas Kinetics, Validity of IAST and Sips Isotherm Models, and Temperature/Pressure Relationships for Process Modeling

Abstract

Adsorption processes can be applied to the separation of alkane/alkene mixtures. Process modeling is a key tool used to assess if the operating and capital cost of these processes can compete with the industry standard (e.g., cryogenic distillation for ethylene/ethane). These process models rely on experimental adsorbent data, which, due to simplicity, is typically gathered using only pure gases at equilibrium conditions. Gas separations are fundamentally mixed-gas processes, raising the following question: is equilibrium pure gas adsorption data suitable to predict mixed-gas separation performance? To answer this, our work provides process modeling data sets for ethylene and ethane on Mordenite, Zeolite 13X, and ZJU-74a and illustrates that pure gas kinetic measurements are sufficient to predict mixed-gas behavior and provides caution on relying on equilibrium conditions predicted by the ideal adsorbed solution theory (IAST) and extended-Sips methods. Using a modification of the volumetric method for measuring adsorption, we report pure gas kinetics as a function of temperature (20–80 °C) and pressure (<100 kPa) for the adsorption of ethylene and ethane onto Mordenite, Zeolite 13X, and ZJU-74a to supplement isotherms and aid in process modeling. At 293 K, the pseudo-second-order adsorption rate constants of ethylene and ethane, respectively, on Mordenite (4.2 and 2.0 kg·mol–1·s–1) and Zeolite 13X (6.4 and 10 kg·mol–1·s–1) increase with increasing pore size, whereas ZJU-74a (4.9 and 6.6 kg·mol–1·s–1) does not. Only Mordenite exhibited kinetic selectivity (3.2), with the ratio of rate constants for Zeolite 13X and ZJU-74a near unity, suggesting that process models of Zeolite 13X and ZJU-74a will not benefit from considering kinetics. Rate constants of all materials follow an Arrhenius trend with temperature, and the effect of pressure was within measurement error and considered negligible, allowing for simple implementation of these results into process models. Additionally, we show that the extended-Sips isotherm and IAST methods for predicting binary equilibria from pure isotherms perform poorly at predicting the equilibrium condition observed in mixed-gas kinetic experiments.