Presented in this thesis are the results from an integrated experimental and modeling study on damage cascade formation in ion bombarded solids. The molecular dynamics (MD) simulations were performed by using LAMMPS (Large-scale Atomic/Molecular Massively Parallel Simulator). In one subtask, we studied damage cascade interactions caused by two 2 keV Si atoms simultaneously bombarding a crystalline Si substrate. We found that the enhanced displacement creation appears primarily in the thermal spike stage with all atoms at energies less than the displacement threshold. The study lead to the conclusion that the cascade interactions increased local melting by increasing energy deposition density, thus promoting defect creation. In another subtask, we studied radiation damage in Si0.8Ge2 layer caused by Agn clusters with number of atoms in a cluster, n, taking values from 1 to 4. It showed that strained SiGe, a material known to have poor radiation tolerance, still follows the overlap model, rather than the direct amorphization model. In the third subtask, MD simulation has shown that crowdion defects formed in bcc Fe are propagating along <111> directions. Crowdion defect starts to form when damage cascade reaches the maximum volume and contributes a second peak in defect buildups with increasing times. Upon defect recombination, crowdion defects shrink and form <111> oriented dumbbell defects at the crowdion end. In subsequent structural relaxation, <111> dumbbell defects rotate and finally align themselves with <110> directions. The surviving dumbbell defects represent a significant contribution to the final defect distribution after thermal spike formation. The overall research reveals atomic scale details of damage buildups at early stages of defect developments. Although the target systems cover both semiconductor materials and metal, these results show that MD simulation is a powerful tool to show the details at a spatial and time scale beyond experiments. These details are very important to develop understanding the precursor formation in defect clustering in such a case.
