Modeling of transient enhanced dopant diffusion by using a moment-based model describing point-defect clustering

The accurate simulation of the transient enhanced diffusion (TED) of dopants during the post-implantation anneal is still a big challenge for process simulation. As a point-defect mediated process, the kinetics of TED is mainly determined by the interaction of intrinsic point defects with extended defects formed by point-defect agglomeration. Using a continuum model for the simulation, the reaction-diffusion equations describing the behavior of the species considered have to be solved numerically. In such an approach, a partial differential equation has to be solved for each kind and each size of extended defects. Because extended defects may consist of some hundred point defects, such models lead to inadmissibly high CPU times and CPU resources needed. Therefore, we have implemented an approach where the concentration of extended defects as a function of their size is described by moments. It was found that the moment-based approach leads to a significant reduction of the number of coupled differential equations to be solved as well as of computation time, with a good agreement between the results of the full model and those of the moment-based approach.