Impact of Chemical Heterogeneity on the Accuracy of Pore Size Distributions in Disordered Solids

Disordered porous solids, such as heterogeneous catalysts, possess internal void spaces with high degrees of both chemical and geometric heterogeneity. For pore size distributions (PSDs) to be useful void space descriptors for aiding understanding of phenomena such as catalyst activity, their level of accuracy must be known. Due to differences in specific interactions and wetting behavior, the choice of adsorbate can severely impact the accuracy of PSDs derived for disordered materials using gas sorption. Wetting effects in the capillary condensation region have recently been studied extensively by simulation, but there have been no complementary, definitive experimental studies of sorption mechanisms. However, in this work, the novel integrated gas sorption and mercury porosimetry technique has been used to prepare a particular subset of pores, located within the disordered network of a given amorphous material (eg silica or alumina), that had dead-ends with different chemical properties to the walls, and then definitively compare the individual sorption behavior of two different, common adsorbates (nitrogen and argon) within just these pores. The presence of a heavy metal surface was found to lead to a shift to higher pressure in the positions of adsorption and desorption branches and a widening of the hysteresis for argon isotherms for silica pores, but not for alumina pores. In contrast, no such effect was observed for nitrogen. Hence, using argon adsorption rather than nitrogen adsorption was found likely to lead to pore size inaccuracies of at least similar to 200% in characterizing silica-supported platinum catalysts, for example. The experimental findings on effects of wetting and pore length were found to be consistent with findings from recent Monte Carlo and MFDFT simulations. Further, analysis of hysteresis widths also suggested that the standard approach using NLDFT would be unsuitable for the type of samples studied here. The results highlight particular issues for the accuracy of PSDs for catalysts consisting of metal nanoparticles supported on oxide materials.