Modeling of type II high-resolution sorption isotherms: Evaluation of different approaches

Several classical sorption isotherm equations were fitted to experimental high-resolution data for type II multilayer isotherms. Both the models and the experimental data (sorption of Ar (87 K), N2 (77 K), H2O (303 K), n-pentane (298 K), neopentane (273 K) on macroporous silica are taken from the literature. Further data include N2 sorption on nonporous functionalized silica and alpha-alumina. The following equations are tested: BrunauerEmmett-Teller (BET), its modification by Anderson, Aranovich, the zeta-isotherm, Harkins-Jura, Frenkel-HalseyHill (FHH), that of de Boer and Zwikker manifesting itself in later work of Fuller, Condon, and Adolphs, the dualsite Langmuir equation with monolayer clustering recently presented by Calzaferri et al., and at last the general cluster sorption isotherm (GCSI, Micropor. Mesopor. Mater. 316 (2021) 110909). The GCSI describes all experimental data with high accuracy, not only for homogeneous but also for heterogeneous surfaces with a Gaussian energy distribution. All other models fail more or less, depending on the initial slope of the experimental isotherm. Besides theoretical considerations, the quality of each isotherm modeling is evaluated by the pure statistical Akaike criterion.

Automated summary

This study fitted several classical sorption isotherm equations to experimental high-resolution data for type II multilayer isotherms. The results showed that the general cluster sorption isotherm (GCSI) accurately describes the experimental data for both homogeneous and heterogeneous surfaces. Other models fail to varying degrees, depending on the initial slope of the experimental isotherm. Additionally, the quality of each isotherm modeling was evaluated using the pure statistical Akaike criterion.