The Extraction of Charge Matrices from Two-Dimensional Device Simulations of MOS Transistors

Quasi-static models of n-terminal semiconductor devices usually express terminal current as the sum of steady-state and transient components. In order to satisfy charge conservation it is convenient to represent such transient components as the time derivative of charge. This has an additional advantage for table models of reduced memory requirements when compared with formulations using capacitance. However, two fundamental questions must be answered. Firstly, is it valid to express the transient component of terminal current as the time derivative of a "terminal charge" and secondly, how can these terminal charges be extracted? These issues are discussed in this paper. Experimental justification is given for the validity of charge based models for four-terminal MOS transistors and a method for extracting terminal charges is described along with an example.