Suppression of Statistical Variability in Stacked Nanosheet Using Floating Fin Structure

A novel floating fin-structured nanosheet (FNS) is proposed to improve the work function (WF) variation by the fluctuation of the space between channels ( ${T}_{{\text {sp}}}$ ) in vertically stacked nanosheet devices. Through computer aided design (TCAD) simulations with calibrated models, it is found that the WF variation induces the serious variations of electrical characteristics such as threshold voltage and current drivability, and FNS is more robust against the ${T}_{{\text {sp}}}$ fluctuation than laterally long nanosheet (LNS) thanks to its unique structure where the channel adjacent to the ${T}_{{\text {sp}}}$ occupies a small portion of the entire channel. Furthermore, based on figures of merit (FoMs) analysis, it is revealed that the slow corner margin of FNS is significantly improved without sacrificing the device performances, which is advantageous for near-threshold-voltage (NTV) designs.