Physically based modelling of damage, amorphization, and recrystallization for predictive device-size process simulation

摘要

Current advanced CMOS source/drain engineering involves the use of amorphizing implants with 3D geometry. Upon annealing, the induced transient enhanced diffusion (TED) can only be accurately predicted if the amorphized region is correctly modeled, as well as the formation and evolution of extended defects, particularly 3 1 1's and dislocation loops. In addition to the extended defects, already modeled in the atomistic kinetic Monte-Carlo simulator DADOS, we have developed a physically based modeling approach for the implant-induced damage build-up, amorphization and recrystallization, suitable to handle device-size process simulation. It is based on amorphous pockets (3D, irregular shape agglomerates of an arbitrary number of interstitials and vacancies, plus trapped impurities) with a size-dependent activation energy for recombination. The model is able to reproduce experimental aspects like the crystal-amorphous transition temperature and the super linear increase of damage with dose. We describe the model and present simulation examples. The efficiency of the model, in terms of CPU time and memory requirements, will also be discussed.