Since its development in 1938 by Brunauer, Emmett and Teller (J. Am. Chem. Soc., 1938, 60, 309), BET theory has become one of the most widely used theories in the characterization of porous materials. The seemingly straightforward BET isotherm equation describes multilayer adsorption on homogenous, unbounded surfaces and it is the primary method for estimating specific surface areas of porous solids. However, recent applications of the BET equation to microporous materials, in particular metal-organic frameworks, have raised questions concerning its ability to accurately characterize materials where confinement effects and structural heterogeneities cannot be simply neglected.
With the aid of molecular simulation methods, we have undertaken a detailed and systematic study to examine the range of applicability of the BET method in determining the surface area of microporous materials. Using grand canonical Monte Carlo simulations, we have calculated adsorption isotherms for argon and nitrogen using a number of models, including simple slit- and cylindrically-shaped pores, single-walled carbon nanotube bundles, amorphous carbons, metal-organic frameworks and zeolites. We compare estimates of the specific surface areas obtained using the BET equation to analyze the simulated isotherm data with exact measures of the surface area calculated from the geometry of the materials. Our results show that in some cases the BET method provides accurate estimates of geometric surface areas, while in other cases discrepancies arise. We shed light on the successes and failures of the BET equation and discuss the role of confinement effects, surface curvature and energetic heterogeneities in determining its range of validity. Finally, we discuss the application of the BET method to real microporous materials and suggest strategies for consistently implementing the method and interpreting the resulting surface areas.