6 Future Work

  I suggest four topics, which should be developed further. First, an extension of the two-dimensional analytic ion implantation module to three-dimensional simulation domains is necessary. Therefore, advanced geometry representations, like finite octrees, are needed. Due to the highly nonplanar structures the simulation grid should also be changed for the ion implantation modules from ortho-product grids to unstructured grids in order to resolve the geometry details more accurately.

The second area addresses the diffusion module. From the modeling point of view a model for phosphorus diffusion at high concentrations including precipitation kinetics is needed. Till now experimental data obtained from Yoshida [Yos74] are not consistently reproduced. Additionally, models for recently established materials like silicides are of considerable interest.

The data management of process model parameters becomes increasingly important. Within this work we presented the first attempt to store modeling
parameters in VISTA's Material Data Base. However, the hierarchical structure of this data base might not be generic enough for the requirements arising with process data. Therefore, it is necessary to develop a model and data server, which should manage all process models and modeling data.

At present the diffusion module cannot handle moving boundaries as given
during an oxidation process. Normally, the gridding problem limits the applicability of oxidation programs, because after each transient time step regridding of the simulation domain is required. As efficient grid generator programs are incorporated into the diffusion module, it would offer manifold fields of activities to implement the simulation of the oxidation process. This module should be
able to calculate the oxide growth simultaneously with dopant diffusion including oxidation induced defects.