Modeling of implantation and mixing damage during etching of SiO2 over Si in fluorocarbon plasmas

Abstract

Energetic ion bombardment during plasma etching of microelectronics devices is necessary to activate chemical process and define features through the ions’ anisotropic trajectories. These energetic fluxes can also cause damage and mixing of the constituents of crystalline lattices. These properties are likely best modeled using molecular dynamics (MD) simulations. The computational expense of these techniques makes feature scale simulations difficult, and so motivates development of approximate methods that can be used to model full features. In this regard, an implantation and mixing model has been developed and implemented into a Monte Carlo feature profile model to simulate the mixing and damage to the underlying Si during high aspect ratio (HAR) etching of SiO2 trenches. Fluxes to the surface were provided by a reactor scale model. The feature scale model was validated by comparison to the mixing produced by Ar+ bombardment of Si with and without F and CF fluxes as predicted by MD simulations. Scaling of mixing damage of underlying Si during HAR of SiO2 etching in Ar/C4F8/O2 plasmas for rf bias powers of 1–4 kW was investigated. The authors found that mixing damage at the bottom of HAR features, though increasing in magnitude with increasing ion energy, does not scale as dramatically as on flat surfaces. This is due to the reflection of ions off of sidewalls which moderate the ion energies.