Tw. Randolph et al., DOES CRITICAL CLUSTERING AFFECT REACTION-RATE CONSTANTS - MOLECULAR-DYNAMICS STUDIES IN PURE SUPERCRITICAL FLUIDS, Journal of physical chemistry, 98(15), 1994, pp. 4173-4179
We describe the results of molecular simulation approaches to the stud
y of reactions in near- and supercritical fluids. For simplicity, we c
arry out the simulations in a pure Lennard-Jones fluid but with the mo
lecules initially tagged as ''unreacted''. We quantify the reaction ra
te in terms of the rate constant, measured either (a) by counting all
collisions between pairs of reactants or (b) by tagging each colliding
pair as ''reacted'', monitoring the concentration of ''unreacted'' mo
lecules with time, and carrying out standard kinetic analyses. At low
densities, both rate constants are identical to the kinetic theory res
ult; at high density, only the reaction-based rate constant gives the
physical result that the rate constant varies inversely with density (
diffusion limitation). At intermediate density, there is a crossover b
etween the two behaviors. We also describe the dependence of each type
of rate constant on the cybotactic diameter (represented by the colli
sion diameter in this work) of the reacting molecules. At low density,
where the two rate constants are the same, their dependence on the co
llision diameter is identical to that of the equilibrium radial distri
bution function; i.e. locally high solute-solute correlations can infl
uence the rate constant dramatically. At high density, where diffusion
limitation sets in, the dependence on the collision diameter can be p
redicted from Smoluchowski theory and the influence of high peaks in t
he solute-solute radial distribution function all but disappears. This
explains why the high solute-solute pair correlations measured in pre
vious simulations do not appear to be manifested in most reaction stud
ies made to date in supercritical fluids.