9  Outlook

It has been shown to be insufficient to determine the characteristic step-heights caused by single defects in electrostatic and drift-diffusion-based simulations. Consequently, variability models for RDD and first order quantum correction models have been integrated for future investigations into the SHE-based simulator ViennaSHE. The advantage of directly solving the BTE is that changes in the mobility and charge carrier energy in the presence of RDD can be directly investigated. This is necessary since in the following  [98] one needs to reproduce the low-field mobility-doping relation found by Caughy and Thomas  [191], when using RDD in a large resistor with a, at least one nanometer, fine mesh. To calibrate the RDD model in ViennaSHE one should undertake a statistical significant number of simulations with RDD to find the same resistance in the limit of an infinite large slab of silicon with RDD as one would find using continuous doping under low-field conditions. In all simulations with RDD fine meshes are required to correctly resolve the granularity of the doping and leads to a high computitional demand. Thus, for this investigation the simulator ViennaSHE needs to be parallelized using MPI  [192] and an algebraic multi-grid  [193] solver such as Prometheus  [194] is required.