Computational Study of Temperature Effects on MOSFET Channel Material Benchmarking

Abstract

Temperature ( ${T}$ ) effects on MOSFET channel material benchmarking are explored using atomistic quantum and Monte Carlo simulation. Performance metrics such as OFF-current, maximum supply voltage ( ${V} _{\text {DD,max}}$ ), and effective inverter drive current ( ${I} _{{eff}}$ ) for novel channel materials (InGaAs and Ge nMOS) are compared with those of Si nMOS at 27 °C vs. 101 °C. For novel channel materials, leakages from source-to-drain or band-to-band tunneling may become significant, while their ${T}$ -dependence is much lower than in thermal leakage. This may have significant impact on the benchmarking analysis results, such as the increased ${V} _{\text {DD,max}}$ of novel channel materials at high ${T}$ , which allows one to exploit the boosted ${I} _{{eff}}$ in a wider range of supply voltage with leakage power still matching Si. This illustrates the importance of considering operating ${T}$ ’s of actual circuits when assessing individual transistors, especially for novel channel materials.