Strained Si material has emerged as a strong contender for developing transistors for the next generation electronics. Strain lifts the degeneracy of the valence and conduction bands which can be used to deliver superior transport properties in comparison to bulk Si. The effect of strain on device characteristics can be studied using the Monte Carlo method. Based on Monte Carlo simulations, a comprehensive set of models for the strained Si/SiGe material system can be obtained. These models are to be implemented in a device simulator and then used to investigate and design different strained Si device structures.
An analytical model has been developed to describe the anisotropy of the low-field electron mobility in strained Si on arbitrarily oriented SiGe substrate. The model includes valley splitting for a given strain tensor, the effect of reduced inter-valley scattering with increasing splitting, and doping and temperature dependence. In order to validate the model, Monte Carlo simulations were performed and the results obtained were fit to the experimental data, available mainly in the form of piezo-resistance coefficients. It was observed that changing the deformation potential gave good agreement with piezo-resistance-based mobility for low strain levels, whereas adjusting the inter- and intra-valley coupling constants delivered the desired mobility enhancement.
The electron high-field transport in strained Si has also been studied using Full Band Monte Carlo simulations. From Monte Carlo simulations, it was observed that the valley velocities decrease with an increase in strain, whereas the total velocity increases. This phenomenon can be explained by the repopulation of valleys induced by the field. The total velocity also shows a region of small negative differential resistance (Gunn effect). Two different modeling approaches capturing these velocity-field characteristics have been developed. The first one is based on calculating the total velocity from the valley-specific electron velocities and populations, while the second is a direct fit to the total velocity. Future work will concentrate on modeling of the various parameters for this material system based on Monte Carlo results, with special focus on the surface mobility and obtaining device characteristics.