Simulating current flow in a nanoscopic thermoelectric-gated field effect transistor via non-equilibrium Green's functions

Existing data centers consume large amounts of power to cool processors, and the power needs of data centers will only increase as this sector grows. However, if the temperature of processors and boards can be measured more locally and efficiently with a low circuit footprint sensor, existing robust cooling solutions can be applied to hot spots more precisely, in turn lowering the energy needs of data centers. The aim of our research is to model such a temperature sensor at the atomic level using ab initio atomistic methods. We investigate a nanoscopic field-effect transistor, with the gate voltage applied to the channel region via a thermal difference across a thermoelectric gate electrode material. We model the interfacial stability of the thermoelectric-gated transistor materials, the coefficients of thermal expansion of all constituent materials, and the electron transport from the gate thermoelectric to the transistor junction channel region. Transport properties are investigated by calculating electron transmission and conductivity as a function of potential difference across the thermoelectric-oxide-silicon junction through a Keldysh formulation of Green’s functions. These properties will additionally inform the temperature-dependent Seebeck coefficients, which can be used to assess the thermoelectric figure of merit for the chosen material. These ab initio modeling efforts will demonstrate such a device’s efficacy and describe the thermoelectric features required to operate it.