Today's development of semiconductor devices strongly depends on the aid of process and device simulation. The down-scaling of the feature sizes creates needs for advanced simulation models. These models must be able to predict the dopant profiles even for sub-micron technologies. Within this work we present several models to simulate the ion implantation and the diffusion processes in nonplanar arbitrary structures. Both modules are included into the set of process simulators of the VISTA ( Viennese Integrated System of TCAD Applications) project.
We present an analytical ion implantation module, which is an extension of
the numerical range scaling technique and can handle arbitrary
two-dimensional
simulation structures. A library of several
distribution functions is provided
including the four parameter
kappa distribution function, which is used for the first time in
semiconductor engineering to calculate the dopant distributions in the
implantation targets.
The module for diffusion is able to solve a large class of nonlinear,
coupled partial differential equations (PDEs) in an arbitrary
two-dimensional simulation domain. For the numerical solution of the model
equations the box integration discretization method is applied to
unstructured simulation grids. Furthermore, the
module is
equipped with an adaptive grid refinement algorithm and an automatic time
step control for transient simulations.
The major part of this work are the physically based diffusion models used
to simulate several kinds of anomalous diffusion behavior in different
semiconductor materials. We present a diffusion model for the outdiffusion
of dopants from a
polysilicon layer for different process
temperatures and annealing techniques.
Thereby the enhanced
diffusivity of the dopants in polysilicon material is
reproduced. During this outdiffusion process the material interface between
the polysilicon layer and the underlying substrate material strongly affects
the amount of outdiffused dopants. We give a new model for the calculation
of the interface concentrations. Additionally, the complicated phenomenon of
epitaxial realignment is included to the polysilicon interface
and bulk models.
Boron exhibits anomalous diffusivity after implantation into the silicon
substrate. It is known that point defects are responsible for this transient
diffusion enhancement. We give a new modeling approach for transient
enhanced diffusion of
boron in silicon, which is based on
aggregation of point defects during the diffusion
process.
The point defects also play a key role during dopant activation. After implantation the dopants are located at interstitial sites in the lattice. We present a diffusion model which is able to capture the highly non-equilibrium activation process during the early stages of damage annealing.