Semiconductor Device Simulation Software with Monte Carlo Based Thermal Transport Modeling

Current semiconductor device simulation software adopts Fourierĺs law to perform the thermal transport modeling. However, Fourierĺs law will become invalid at micro/nano-scale, which can cause a large discrepancy in the thermal and electrical performance between the simulation results and the actual application. To improve the current state-of-the-art, a new semiconductor device simulation software that adopts the advanced phonon frequency-dependent Monte Carlo simulation technique to replace the conventional Fourierĺs law for the thermal transport modeling is proposed. Monte Carlo simulations track the phonon movement and scattering so that the thermal properties of the nanomaterial can be extracted stochastically. The simulation accuracy will be improved because the proposed Monte-Carlo-based software can incorporate detailed phonon transport physics. The Monte Carlo simulation will be coupled with Charon and Trilinos software packages, which are funded by the Department of Energy and its Office of Advanced Scientific Computing and Research, form a complete semiconductor device simulation software. In Phase I, the coupling between the frequency-dependent phonon Monte Carlo code, Charon, and Trilinos was demonstrated. Simple one- and two-dimensional test cases were performed to validate the coupling scheme. In addition, four next generation semiconductor materials was added to the material database. If Phase II of the project is funded, a three-dimensional device simulation software with an enhanced coupling and user interface will be developed. Such device simulation software will offer high fidelity simulation results in which will benefit both the semiconductor industries and academia. The current semiconductor device simulation software adopts an oversimplified thermal modeling technique that can give inaccurate results. This proposal aims to use an advanced particle tracking technique to replace the current approach to largely increase the simulation accuracy.