Hydrogen is an essential ingredient in semiconductor technology. It is known to enhance the electronic behavior of silicon in regimes with a large amount of dislocations, as is the case for amorphous silicon or the Si-SiO2 interface. It is usually introduced in various processing steps, e.g. through wet oxidation or the formation of a gas anneal.
The Negative Bias Temperature Instability (NBTI) of MOSFETs has become a severe reliability issue in modern semiconductor technology. It is generally attributed to the depassivation of hydrogen-passivated silicon dangling bonds at the Si-SiO₂ interface and the subsequent diffusion of this hydrogen through the oxide.
Within the Reaction-Dispersive-Diffusion (RDD) model, this diffusion is described using the Multiple Trapping (MT) approach, which was initially developed to model the dispersive transport of electrons within amorphous semiconductors. The assumption of the MT model is that the diffusing particles exist either in the conduction state, where they are able to move freely, or in a trap state, where no movement is possible. The release of a particle from a trap state is only possible through thermal activation. The trap states are assumed to be statistically distributed on the energy scale.
The mathematical properties of the approach have been analyzed by means of analytic approximations. Here we focus particularly on the behavior of traps in the case of saturation, which is an accelerated equilibration due to trap filling. In order to test the applicability of MT, the diffusion of deuterium in disordered silicon was calculated using a newly implemented simulator. Simulation output and measurement data show good agreement regarding temperature and saturation behavior.