The downscaling of modern devices is increasing the power density to be dissipated via thermal conducting and, in some cases, self-heating effects may degrade the device performance. There is, therefore, the need to properly model heating and dissipation due to the Joule’s effect. Thanks to its deep connection with Drift-Diffusion, Thermal has established itself as a powerful and flexible tool to compute power balance in realistic devices. Effects going beyond the standard diffusive model, such as the energy relaxation of hot electrons, can be added by a constant value heat source. Furthermore, a number of thermal boundary conditions allows to model the environment around the device in a realistic way. For example, the heat dissipated by the substrate can be modeled as an effective thermal surface resistance. Thermal insulating and conducting surfaces can be easily added, as well.

The key feature of Thermal is its integration with Drift-Diffusion. The interplay of the two modules allows to easily compute the power balance at the diffusive level. The special module Selfconsistent handles the variable exchange between them, such as heat source, temperature and thermal flux. The temperature computed by Thermal is used to update the temperature dependent physical quantities used by Drift-Diffusion (e.g. electron mobility, Seebeck coefficient and so on).

Main characteristics of the model:

• Heat transport according to the Fourier law
• Heat generated by the charge carriers (Joule’s effect)
• Effective heat source to take into account non equilibrium effects
• Thermal boundary resistance
• Insulating and conducting thermal surfaces
• Clamp boundary condition
• Peltier effect