Abstract
Adsorption data of aromatic molecules adsorbed in silicalite show highly unusual characteristics which were attributed to structural effects caused by the comparable size of molecules and pores. In this study, the interaction of aromatic compounds with silicalite are examined on the molecular level. The interactions are calculated by atom-atom approximation using Lennard-Jones potentials. The constants are calculated, without fitting, from Kirkwood-Muller formulas. Benzene and p-xylene are represented as a rigid structure of 12 and 18 atom centers. The model is anisotropic. The diffusional behavior of molecules is examined by minimizing the potential energy in the channels which requires less computational time than Molecular Dynamics. The activation energy for the diffusion of benzene, 27.6 kJ/mol, is in excellent agreement with data, 28.8 kJ/mol. The results indicate that both molecules can enter the smaller zig-zag channels. The energetically most favorable location in the main channels is the mid-point between intersections. All rotations are restricted in the channels but the molecules can rotate in any direction (with some movement of the center) at intersections. The Henry's law constant and internal energy of adsorption at zero coverage are calculated by direct integration of the configurational integral. Direct integration is more efficient than Monte Carlo and Molecular Dynamics simulations since the molecules are highly restricted in the pores. The predicted internal energy of adsorption, − 54.86 and − 75.30 kJ/mol for benzene and p-xylene is in good agreement with data of − 50.92 and − 62.15 kJ/mol respectively. There is appreciable difference between the predicted and experimental Henry's law constants. The agreement can be improved by fitting the Lennard-Jones constants which has not been attempted. Although the calculations are performed at infinite dilution and entropy effects are not included, the results bring insight to the behavior of molecules in highly restricted environments such as in tight pores. Similar simplified calculations can be used to close the gap between highly idealized molecular simulations and complicated systems common in real applications.