Skip to Main Content
Your organization might have access to this article on the publisher's site. To check, click on this link:http://dx.doi.org/+10.1116/1.580838
Sputtering, although often optimized by trial and error, increasingly needs sound modeling techniques because of the high capital investment. Usually the simulation problem is broken down into three stages: emission at the target, transport through the host gas, and deposition at the substrate. It turns out that these stages are necessarily linked and must be properly decoupled. Here we are mainly interested in the transport through the vapor phase, which may be simulated by our Monte Carlo model, SIMSPUD. However, at high operating pressures, significant deposition takes place onto the target. To modify our Monte Carlo routine, rather than emitting particles in proportion to the erosion profile of the target and waiting for some of these to completely escape, all particles which land on the target are re-emitted until they ultimately escape to the substrate or chamber walls. This modification makes the net emission rate equal to the target erosion rate. At sufficiently high pressures, the Monte Carlo routine becomes very slow. In fact, the problem then becomes a diffusion problem best handled by numerical solution of the Laplace equation. At any rate, it is shown that when SIMSPUD has been modified in this way, good agreement with experiment is obtained over a wide range of working pressures. © 1997 American Vacuum Society.