BET surface area of carbonaceous adsorbents—Verification using geometric considerations and GCMC simulations on virtual porous carbon models Piotr A. Gauden a , Artur P. Terzyk a, *, Sylwester Furmaniak a , Peter J.F. Harris b , Piotr Kowalczyk c a N. Copernicus University, Department of Chemistry, Physicochemistry of Carbon Materials Research Group, Gagarin St. 7, 87-100 Torun´, Poland b Centre for Advanced Microscopy, University of Reading, Whiteknights, Reading RG6 6AF, UK c Applied Physics, RMIT University, GPO Box 2476 V, Victoria 3001, Australia 1. Introduction In 2008 we celebrated the 70th Anniversary of the publication of the BET equation-the model applied mainly for calculation of the surface area of solids from gas sorption data. Although more sophisticated methods have been developed, the BET model is still a standard (recommended by IUPAC [1,2]) and the most widely used procedure to characterize various adsorbents [3–7]. Of course, the concept of the surface area for a microporous carbon is very useful unless one is operating at low pressures during adsorption. At high pressures where pore filling dominates it is the pore volume that is a more important characteristic. On the other hand, the BET surface area is provided in almost each paper where activated carbons are applied, since it is a standard parameter. The main advantage of the BET equation is its simplicity and this is the reason why, in spite of its success, this equation is also one of the most criticized theories since its origin [3–10]. In the light of the problematic assumptions of the original BET equation, the key question about this relationship is whether the surface area it provides is the ‘‘true’’ and ‘‘correct’’ value. Therefore, in our opinion information about its accuracy can be quite useful. Recent developments in the field of computer modelling and molecular simulations make it possible to check the validity of different assumptions and concepts applied in the field of physical adsorption. The assumption of the atomistic structure of hetero- geneous adsorbents offers, for example, the possibility to test the BET method and the limitations of its application reported in literature [8–18]. From the analysis of published results it can be seen that for different types of adsorbents the overestimation, the underestimation and/or the similarity of the surface area calculated by the BET method (in comparison with the geometrical considerations) can be observed [8–18]. Therefore, it is very difficult to correlate the differences between both types of surface areas with that of an adsorbent, the structural heterogeneity, and porosity. Considering activated carbons a crucial role in checking of different adsorption concepts and models has been played by the so called Virtual Porous Carbon models (VPC). One of them i.e., the VPC model proposed by Harris and developed in our previous reports [19–27] is considered in this study. It was obtained from the analysis of HRTEM images of carbon samples; however, as recently shown by Powles et al. [28] the quenched MD simulations of carbon structures using sophisticated carbon–carbon potential provided a very similar structure, so the model seems to be realistic. In this study, known molecular structure of a series of VPCs having various average pore sizes and porosity is used for calculations of both the BET (from the low-temperature simulated adsorption isotherms of Ar) and the accessible surface areas (the Applied Surface Science 256 (2010) 5204–5209 ARTICLE INFO Article history: Available online 28 December 2009 Keywords: Computer simulation GCMC Virtual Porous Carbon Surface area BET ABSTRACT The applicability of BET model for calculation of surface area of activated carbons is checked by using molecular simulations. By calculation of geometric surface areas for the simple model carbon slit-like pore with the increasing width, and by comparison of the obtained values with those for the same systems from the VEGA ZZ package (adsorbate-accessible molecular surface), it is shown that the latter methods provide correct values. For the system where a monolayer inside a pore is created the ASA approach (GCMC, Ar, T = 87 K) underestimates the value of surface area for micropores (especially, where only one layer is observed and/or two layers of adsorbed Ar are formed). Therefore, we propose the modification of this method based on searching the relationship between the pore diameter and the number of layers in a pore. Finally BET; original and modified ASA; and A, B and C-point surface areas are calculated for a series of virtual porous carbons using simulated Ar adsorption isotherms (GCMC and T = 87 K). The comparison of results shows that the BET method underestimates and not, as it was usually postulated, overestimates the surface areas of microporous carbons. ß 2009 Elsevier B.V. All rights reserved. * Corresponding author. Tel.: +48 056 611 4371; fax: +48 056 654 2477. E-mail address: aterzyk@chem.uni.torun.pl (A.P. Terzyk). Contents lists available at ScienceDirect Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc 0169-4332/$ – see front matter ß 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.apsusc.2009.12.097