The Brunauer–Emmett–Teller model on alumino-silicate mesoporous materials. How far is it from the true surface area?

Abstract

Determining the surface area of porous materials through the Brunauer–Emmett–Teller (BET) model is a common practice. The method is generally applied in commercial software packages, where the assumptions are sometimes accepted by the experimenter whilst they may sometimes require a deeper analysis. One element of debate is the molecular cross-sectional area of the adsorptive. There is not yet agreement about the correctness of the BET model using a certain value for cross-sectional area of N2; the conventionally-used parameter seems to overestimate the surface areas. In this work, a preliminary study of a modified method is presented, which introduces an ‘apparent’ cross-sectional area for N2, which is smaller to the typically-used value. This value was obtained after measuring a number of relevant mesoporous materials in N2 and Ar, using a model that considers an apparent value for the cross-sectional area. The model predicts outcomes very close to the Ar-based measurements in terms of low relative error. Then, we went one step further and looked into the geometrical surface areas, also referred to as true surface areas. By combining prior studies with our work, it was found that the surface area, using N2 and the conventionally-used cross section, can be ca. 50% higher than the geometrical surface area. Therefore, the significance of the BET surface area seems to be far from well understood, though it is widely applied. This approach also allowed to define an ‘effective’ cross section for N2, that relates it to the geometrical surface area. Its value agrees with prior considerations for an epitaxial orientation of the N2 molecule with a hydroxylated silica surface. As a final recommendation, critical thinking is needed about the default settings in standardised calculations, which may not represent a reliable measure of the true surface area.