Decoupled algorithm for the multicomponent potential theory of adsorption of gas mixtures

TL;DR

This paper introduces a decoupled implementation of the multicomponent potential theory of adsorption, improving parameter interpretability and individual component adjustment, tested on extensive experimental data with modest accuracy improvements.

Contribution

It presents a novel decoupled approach for the multicomponent potential theory, enhancing parameter clarity and individual component calibration.

Findings

4.67% improvement in excess adsorption predictions

Average total excess adsorption error of 6.97%

Average combined mixture prediction error of 15.30%

Abstract

In this paper, we present a new implementation of the Multicomponent Potential Theory of Adsorption model. The proposed interpretation establishes a clear cut between parameters that depends on the adsorbent from those depending on the adsorbate, which leads to a better understanding of the parameters signification. The interdependence between pure isotherms is eliminated, which mean that each component can be individually finely adjusted. This new approach was tested against 14 datasets for a total of 510 experimental mixture adsorption data of CH4, CO2, N2, H2, O2, H2S, C2H6, C3H6 and C3H8 on activated carbons, MOF and zeolites. A slight improvement of 4.67% on excess adsorption predictions was found, leading to an overall average error of 6.97% for total excess adsorption and 15.30% for combined mixtures and components excess adsorption predictions.