For years, the characterization of the surface area and pore size distribution (PSD) of porous materials has been dominated by the analysis of adsorption isotherms of nitrogen (N2) measured at its boiling point (77 K). Because N2 is easily available, not expensive, and routinely used in laboratories, it will continue to be the most used adsorbate for solid surface characterization.
The modern approach to characterize the texture of porous materials is based on using molecular models such as the density functional theory (DFT) or Monte Carlo simulations. The molecular interaction potentials are essential in the development of models for the PSD calculations. Models based on inert gases are best for pore size characterization because their adsorption is most sensitive to the solid geometry like pore width and not to specific interactions with chemical sites on the surface. Therefore, Ar is better than N2, which has a strong quadrupole moment that may interact with polar surface sites and thus influence the adsorption isotherm. The preference of using Ar to N2 for surface characterization was officially recommended by the IUPAC Technical Report (2015).
Carbon dioxide, CO2, another gas often used to characterize microporous carbons, exhibits an even higher quadrupole moment than N2. In the recent study, we substituted N2 and CO2 with O2 and H2 gases with much lower quadrupole moments. The PSD calculations were performed using molecular models based on classical and quantum corrected two-dimensional nonlocal density functional theory (2D-NLDFT). We demonstrated a quantitative agreement between the PSD results derived from the adsorption isotherms of O2 and N2 measured at 77 K and Ar at 87 K on several representative carbon samples. These gases may be used as individual probes for the PSD calculation or combined with the corresponding kernels to fit corresponding isotherms measured on a given sample simultaneously. Such comprehensive analysis of the adsorption data provides more robust results than those derived from a single isotherm. An additional benefit from the models providing consistent PSD results is the possibility of prediction of one isotherm from the other measured on the same sample.
