THE EFFECT OF THE NATURE OF THE INTERACTION POTENTIAL ON CLUSTER REACTION-RATES

The effect of two different interaction potentials, a two-body potenti al, on thermal cluster reaction rates was studied for 2-13 atom nickel clusters using the classical trajectory method. The reaction rates we re computed for cluster-monomer and cluster-cluster collisions at T = 1200 K, using the bulk and dimer parametrized Lennard-Jones (LJ) poten tials and were compared with the rates previously obtained for these c ollisional events by using a more realistic many-body tight-binding se cond moment approximation (TB-SMA) potential. For cluster-monomer coll isions, close agreement exists between the reaction cross section resu lts for dimer fitted LJ (LJD) potential and TB-SMA potential suggestin g that the cluster-monomer collisions may be dominated by pairwise int eractions. The bulk fitted LJ potential (LJB) underestimates the stick ing cross section results of the other two potentials for most cluster sizes. This discrepancy however appears to be due to the relatively s maller cluster binding energies obtained for this potential as a resul t of which a larger cross section for dissociation is observed. For cl uster-cluster collisions, for most cluster sizes, no agreement exists between the reaction cross section results for the three potentials. T he discrepancy between the cross section results for the LJ potentials and the TB-SMA potential appears to lie in the difference in the scal ing of cluster energy with cluster coordination for these two types of potentials (i.e., linear for LJ vs square root dependence for TB-SMA) . Some characteristics of the cross section results of both LJB and LJ D potentials correlate with the relative cluster stability pattern for the LJ clusters. For TB-SMA case, no such correlation exists, which h owever is consistent with the smooth and featureless size distribution s observed experimentally for nickel and other transition metals. The cut-off used in the TB-SMA potential appears to lead to a significant underestimation of the total reaction cross section for N = 13, in the case of the cluster-cluster collisions. The results of this study ind icate that the rate calculations may be sensitive to both the nature a nd parameterization of the simulation potential depending on the tempe rature range considered and cluster growth process simulated. (C) 1996 American Institute of Physics.