THE PERFORMANCE of thermoelectric materials and devices strongly depends on the thermal environment as well as given geometric constraints. In order to exhaust the available thermal energy at the best, thermoelectric generators have to be optimized for each given situation. Thereby, the design of custom devices for given environments depends on a physically based simulation framework. In mainstream microelectronics, Technology Computer Aided Design (TCAD) has been successfully applied for years to maintain the development of semiconductor devices. Besides its role in process development and device optimization, insights to the internal behavior of devices improve the physical understanding of internal quantities. This work implies a fruitful extension to the field of semiconductor device simulation of thermoelectric devices.
A transport model suitable for thermoelectric devices has been systematically derived by the method of moments, which is compatible to an approach based on phenomenological irreversible thermodynamics. Thereby, special attention has been paid to the Seebeck coefficient by a comparison of measurement data with the theoretical formulation.
In addition to the Silicon-Germanium material system which is used for high temperature thermoelectric applications, lead telluride plays an important role for the intermediate temperature range. The material database of MINIMOS-NT has been extended by all relevant models and parameters for lead telluride in order to enable predictive simulation of such devices.
The behavior of both classical thermoelectric devices and novel structures incorporating a large scale pn-junction has been investigated with MINIMOS-NT. Thereby the influence of geometry, material composition, and forced carrier generation on the device performance has been assessed. Non-ideal thermal environments have been discussed and considered using mixed-mode simulations. While the number of possible device variations of conventional thermoelectric devices is relatively small, the numerous design parameters of the novel structures allows good adaption to given environments.
M. Wagner: Simulation of Thermoelectric Devices