Straightforward considerations suggest that chemical reactions including fragmentation processes at or near the front of a chemically sustained shock wave can be important in regulating the shock velocity. This picture is supported by results of recent computer simulations of a model A-B energetic material.The hydrodynamic theory of detonation has enjoyed constant development and apparent success over the past 40 years (for a recent review, see [ 1 ] ). On the other hand, theories which couple the atomic-scale chemical reactions to the shock wave have only recently begun to appear [ 2-10]. For example, after a short initial period, detonations always attain constant velocities that depend only on the properties of the detonating material, and not on the conditions used to initiate the detonation or on the amount of reacted material. It has been argued that this behavior can be explained on the basis of microscopic chemical kinetics [ 11,121, but how such effects couple to the shock propagation has not been clearly established.Detonations travel through solids with supersonic velocities in the range of 3-7 km/s, or 30-70 A/ps. A simple microscopic analysis of this velocity range reveals that the detonation front traverses a typical molecular length in roughly a vibrational period. This rapid chemical kinetics coupled with the violent nature of the process make direct experimental studies of the atomic-scale dynamics of detonations in solids difficult. This time and spatial regime, however, is ideal for study using molecular dynamics computer simulation. To the extent that systems consisting of up to a few thousand atoms mimic the properties of real detonations, molecular dynamics should be capable of providing considerable insight into the atomic-scale chemical dynamics associated with detonating solids.The energy that maintains the shock wave in a detonation comes from solidstate chemical reactions. So far, most previous simulations of detonating solids [ 6-81 have relied on so called "predissociative" pair potentials to model the shockinduced release of chemical energy on an atomic scale. While these types of interactions allow energy release in a straightforward manner, their use also implies that energy is liberated as chemical bonds are broken.