Variability Characterisation of Nanoscale Si and InGaAs Fin Field-Effect-Transistors at Subthreshold

Device characteristic variability induced by Line Edge Roughness (LER) and Metal Gate Work function (MGW) is simulated and compared in two nanoscaled fin field-effect-transistors (FinFETs) with Si and InGaAs channel materials. The study is carried out using a workload manager developed to deploy the simulations and to collect the data across different computational infrastructures. We have found that for the most of the figures of merit, the InGaAs channel device is more tolerant to the LER and MGW variability sources. The MGW has a larger impact (14% more for InGaAs and 33% more for Si) than the LER variability. We have also found a larger correlation between the threshold voltages at low and high drain biases of the InGaAs FinFET than of the Si counterpart.