Atomistic simulations and the requirements of process simulator for novel semiconductor devices

The successful scaling down of novel semiconductor devices requires that corresponding simulation tools should reach atomic resolution, while satisfying the need of the industrial users in term of e.ciency. In this context, we show the potential of multi-scale methodologies based on interconnected approaches ranging from quantum mechanical calculations to Monte Carlo (MC) methods for system kinetics. We prove that one key element for a successful matching of di.erent theoretical methods is the use of low level approaches, not onlyfor parameter extraction, but also for the direct derivation of e.ective interaction models implemented in MC codes. The matching procedure requires on lattice MC model settlement. This gives an added value to the developed stochastic code: it is able to simulate in detail the evolution of nano-structures (impurityaggregates, impurity-defects complex, extended defects) concurring to the overall material modi.cation during processing. The predictivityof this approach is related to the accurate modeling of atomic level phenomena (e.g. di.usion, cluster formation/dissolution, structural transitions) spanning manyorders of magnitude in time. We will report examples of the method application to the simulation of di.erent defective Si system.