Modelling of bias sputter deposition processes

Abstract The Monte-Carlo-based “high dose” program t-dyn has been used to simulate bias sputter deposition processes. t-dyn follows the collision cascades caused by energetic ions in a solid. The changes in the target caused by each collision sequence are stored and the target composition modified accordingly for the next incoming ion. The program also allows the addition of five different non-energetic atomic fluxes to the surface. Calculations of the continuous deposition and resputtering make it possible to simulate the substrate-film interface formation as well as the build-up of a coating on the substrate surface. In this presentation we focus on some phenomena which occur during bias sputter deposition in the limit where the resputtering rate is close to the arrival rate of the deposition species. We shall present simulation studies that illustrate that substrate-dependent preferential sputtering effects frequently appear during film-substrate interface formation. The resputtering yield of the deposition species from the interface may differ by an order of magnitude for different substrates. This effect is so pronounced that for identical processing conditions it is possible to obtain zero net film growth at certain substrates but substantial net film growth at other substrate materials. It is well known that because of preferential resputtering the film composition obtained during bias sputter deposition from an binary alloy target may differ significantly from the target composition. What is less known, however, is that for very thin coatings the resulting film composition may also depend on the underlying substrate. We shall describe how the substrate-dependent resputtering rate effect also may explain this substrate “memory” effect in film composition during bias sputter deposition of binary alloys.