Interpretation of integrated gas sorption and mercury porosimetry studies of adsorption in disordered networks using mean-field DFT

Abstract

The presence of co-operative adsorption behaviour, operating between neighbouring pores within a disordered, void-space network, such as advanced adsorption effects, can significantly complicate the interpretation of gas adsorption data for complex porous solids, such as coked heterogeneous catalysts. The novel integrated gas sorption and mercury porosimetry method can be used to abstract the specific adsorption and desorption behaviour for particular, small sub-sets of similarly-sized pores contained within the complex network of a disordered porous solid. It will be shown in this work how, for ink-bottle geometries, the integrated experiments also allow the deconvolution of the gas sorption behaviour in pore necks, as well as pore bodies, and therefore enable the mechanism of desorption from the pore bodies to be determined. However, proper interpretation of the adsorption data from integrated experiments can be problematic using classical adsorption theories. In this work, it has been demonstrated that the experimental observations can be better understood in the light of meanfield DFT simulations of adsorption in representative pore models. Hence, a better description of the particular physical mechanisms underlying adsorption isotherms in disordered porous solids has been obtained. In addition, the new method allows more detail of the void space geometry to be obtained, such as the ratio of pore neck length relative to pore body length.

Connecting theory and simulation with experiment for the study of diffusion in nanoporous solids

“…Nevertheless, these shortcomings can be considerable, as the interpretation of adsorption measurements using nitrogen or other probes, as well as (mercury) porosimetry, again requires a model, which involves modeling assumptions as well. The interpretation of porosimetry measurements is a mathematical "inverse problem", thus inferring a pore size distribution and the pore connectivity is far from easy for a material with limited additional information on the pore shape [320,[345][346][347][348][349][350][351][352][353][354][355]. The danger in this is conflicting or circular argumentations.…”

“…A problem is to find a relation between the pore network model and the real material. Structural properties of the porous material can be obtained based on various experimental methods, such as gas adsorption (with nitrogen as the standard adsorptive) [351][352][353][354], mercury porosimetry [320, 345-348, 354, 355], small-angle X-ray scattering, electron and X-ray micro-and nanotomography [311-319, 321, 426, 437], and nuclear magnetic resonance [438][439][440][441][442]. The next problem is the pore network reconstruction from measurement data.…”

“…However, if the experimental data come from non-imaging techniques such as mercury porosimetry or gas adsorption, where not all pore space characteristics are readily available, regular pore network construction approaches are usually applied with assumed connectivity. Rigby and Chigada [354,355] have used mean-field DFT [285,286] to interpret data from integrated gas sorption and mercury porosimetry. The authors demonstrated that the experimental observations can be better understood in the light of mean-field DFT simulations of adsorption in representative pore models.…”

Investigation of the problems with using gas adsorption to probe catalyst pore structure evolution during coking

“…This interpretation requires that the aforementioned pore bodies and pore necks in the coked sample C1 fill independently for nitrogen but not for water. As has been shown in past simulations [14], the critical ratio of pore body to pore neck size to permit the onset of advanced adsorption depends upon the properties of the fluid, and fluid-surface interaction strength, and is not fixed at a value of 2, as inferred from the Kelvin equation. Previous studies [21] of coking on HDS catalysts have suggested that the coke formed can be highly polar in nature, suggesting that the coke formed here might interact more strongly with water than nitrogen.…”

“…As suggested by simulation in previous work [14], the aforementioned advanced adsorption phenomena have particular, significant implications for the interpretation of gas adsorption data for samples undergoing coking reactions. For example, the above theory suggests that, if a sample were undergoing pore-mouth blocking, then, while the size of the narrow neck forming (via coke deposits) at the periphery of the pore exceeded half the size of the remaining pore body, advanced condensation would mean that it would appear that the whole pore length was being reduced in diameter, rather than just the mouth.…”