Analysis of Fluctuation Sensitivity Map Algorithms Applied to a 10nm GAA NW FET

Abstract

Industry and academia present an increasing interest in the spacial distribution of variability in nanodevices in order to minimise its effects. The fluctuation sensitivity map is a novel technique which not only provides with information about the spacial effect of different sources of variability in certain figures of merit, but also enables its estimation. Three algorithms have been developed and compared in order to find the optimal method to compute the sensitivity. Such algorithms were applied to the study of metal grain granularity work-function variations in a 10 nm gate length Si gate-all-around (GAA) nanowire field-effect transistor (FET). Having analysed both the predictive power and spacial determination of the sensitivity for each of the algorithms, it can be thereby concluded that the slope of the linear fit between the figure of merit and the work function for each discretised point in the gate is the best of the methods to determine the fluctuation sensitivity map.