Porosity Local Analysis (PoLA): A New Approach to Describe the Porous Volume Distribution in Amorphous Carbons.

A new procedure, named PoLA (Porous Local Analysis), is presented to describe the porosity of amorphous carbons accurately. Unlike models based on predefined geometrical pores, PoLA is based on a point-by-point description of the inner void, and it is particularly suitable for amorphous materials. The porous volume is partitioned into small elements (blocks) of user-defined size, and each block is assigned a micro-, meso-, or macroporous nature according to its minimum distance from the material walls. This method is very fast and characterizes any porous volume uniquely: most importantly, this distribution of volume allows one to predict the gas adsorption behavior of the material. To show this, a number of carbon models have been defined, spanning a large range of porosities, and the adsorption isotherm of nitrogen at 77 K has been accurately simulated with Grand Canonical Monte Carlo in each model. We show that PoLA porous volume distributions and adsorption isotherms are strongly correlated so that N isotherms at 77 K can be accurately predicted by a machine learning procedure on the basis of PoLA results. We expect that this approach will be of great help in the design of new adsorbents and in the interpretation of experimental gas adsorption.