LAUSR

The basic requirements on process design of extremely scaled devices involve appropriate work function and tight doping control due to their significant effect on the threshold voltage as well as other critical electrical parameters such as drive current and leakage. This paper presents a simulation study of 22-nm fin field-effect transistor (FinFET) performance based on various process design considerations including metal gate work function (WF), halo doping (Nhalo), source/drain doping (Nsd), and substrate doping (Nsub). The simulations suggest that the n-type FinFET (nFinFET) operates effectively with lower metal gate WF while the p-type FinFET (pFinFET) operates effectively with high metal gate WF in 22-nm strained technology. Further investigation shows that the leakage reduces with increasing Nhalo, decreasing Nsd, and increasing Nsub. Taguchi and Pareto analysis-of-variance approaches are applied using an L27 orthogonal array combined with signal-to-noise ratio analysis to determine the best doping concentration combination for 22-nm FinFETs in terms of threshold voltage (Vt), saturation current (Ion), and off-state current (Ioff). Since there is a tradeoff between Ion and Ioff, the design with the nominal-is-best Vt characteristic is proposed, achieving nominal Vt of 0.259 V for the nFinFET and −0.528 V for the pFinFET. Pareto analysis revealed Nhalo and Nsub to be the dominant factor for nFinFET and pFinFET performance, respectively.