Atomistic Simulations in Materials Processing

Abstract

The traditional continuum approach, based on solving partial differential equations (PDEs), cannot simultaneously handle the increasingly high number of different phenomena that take place during deep sub-micron device fabrication. Although ab initio calculations can provide a parameter-free description of small atomic structures and classical molecular dynamics (MD) can be used to assess the behavior of extended defects, the time scale that can be reached by these techniques (nanoseconds) precludes their direct use in materials processing simulators. In this chapter, we show that the Kinetic Monte Carlo (KMC) technique is particularly apt to fill in the gap between the continuum and truly atomistic simulations. KMC is capable of taking the fundamental knowledge database obtained from ab initio/MD calculations up to the space and time scales involved in typical silicon IC processing conditions. After reviewing experimental and theoretical knowledge about diffusion and defects in Si, we introduce the fundamentals of a particular KMC scheme for the simulation of point and extended defects. We describe the operation and different components of a specific implementation of this scheme. Finally, we give examples to illustrate some of the capabilities of this approach.