The strength of simulation tools is that after modeling and calibration the effects of changing process parameters, materials, and geometries can be predicted in a fast and simple way. The key for accurate simulation results and all-purpose simulation tools are physically based models. All important process steps, as listed in Section 1.1, which influence the topology and characteristics of a device significantly are worth for modeling and simulation. One of these process steps is thermal oxidation.
Modeling of thermal oxidation has a long tradition. Already in the middle of the 60's the Deal-Grove model has been developed which is still used in modern oxidation simulators. The model is based on two parameters, the so-called linear and parabolic rate constant, in which all the physics of the oxidation process is included. The rate constants must be determined by experiments for the respective oxidant species. Later in the 80's, the Deal-Grove concept has been extended with additional fitting parameters, in order to describe thin oxide films.
The modeling of stress sources and simulation of its effects in semiconductor devices and micro-electro-mechanical systems becomes more and more important. Stress in a material or structure can lead to various negative or undesirable effects. During the fabrication process it can influence the physics of a process in a unpredictable way. Stress can also impair the electrical characteristics of a device and even reduce the life-time of an integrated circuit. In micro-electro-mechanical systems, which are mainly used as sensors, stress can not only change the electrical and magnetic characteristics, it can also cause unwanted deformation in a free standing structure.
The continuously shrinking device dimensions in the state of the art ultra large scale integration (ULSI) technology brings up three-dimensional effects which can not be investigated with two-dimensional simulations. However, the industry is still often confined to use two-dimensional process simulation tools, because of missing three-dimensional alternatives. Therefore, the development of universal three-dimensional models is the actual challenge in process simulation.