Stable extraction of threshold voltage using transconductance change method for CMOS modeling, simulation and characterization

We proposed a stable extraction algorithm for threshold voltage using transconductance change method by optimizing node interval. With the algorithm, noise-free g(m2) (= dg(m)/dV(GS)) profiles can be extracted within one-percent error, which leads to more physically-meaningful threshold voltage calculation by the transconductance change method. The extracted threshold voltage predicts the gate-to-source voltage at which the surface potential is within kT/q of phi(s) = 2phi(f) + V-SB. Our algorithm makes the transconductance change method more practical by overcoming noise problem. This threshold voltage extraction algorithm yields the threshold roll-off behavior of nanoscale metal oxide semiconductor field effect transistor (MOSFETs) accurately and makes it possible to calculate the surface potential phi(s) at any other point on the drain-to-source current (I-DS) versus gate-to-source voltage (V-GS) curve. It will provide us with a useful analysis tool in the field of device modeling, simulation and characterization.