next up previous contents
Next: Comparison with ab initio Up: Results for damaged diamond Previous: Results for damaged diamond

   
General description and justification of the method of calculation

The samples used for the calculations contained 5120 carbon atoms, initially arranged as a perfect diamond crystal of size $10\times
8\times 8$ unit cells. The effects of the sample size on the damage area structure and on its annealing has been checked and addressed below. Various boundary conditions were applied according to the specific problem addressed, including periodic (for the annealing process) and fixed, energy dissipative boundaries (for the energetic bombardment). For the latter, the procedure suggested by Berendsen [106] was applied to the three outermost layers only, i.e. velocity rescaling was performed at each step of the computation for these layers, minimizing the effects caused by spurious reflected energy from the boundaries, and returning the crystal temperature to $\sim$ 0K within 2.5 ps from the local energy deposition event. Any attempt to use another thermostat that affects directly the velocities of the atoms during the bombardment process changed the final structure and led to erroneous results.

In the MD calculations, Newton's equations of motion were solved using a leapfrog algorithm and the Predictor-Corrector algorithm [107] for comparison. Throughout this study, interactive visualization [108] with color coding for different atomic bonding and differently coordinated atoms was implemented, thus aiding the identification of particular defect structures. The Tersoff potential [109] was used to describe the interactions between the C atoms.



 
next up previous contents
Next: Comparison with ab initio Up: Results for damaged diamond Previous: Results for damaged diamond
David Saada
2000-06-22