Accuracy of conventional inversion charge modeling in scaled down MOS devices

Abstract

The density of surface inversion charge in MOS structures is typically calculated using the relation Q = cox( V G − V T ). In the present article we try to quantitatively assess the goodness of this relation in the case of MOS devices scaled down according to constant field scaling and constant voltage scaling principles by comparing the inversion charge given by this relation to the inversion charge obtained by numerically solving the Poisson equation in one dimension. It turns out that while in the case of constant field scaling the conventional relation for inversion charge becomes progressively erroneous, the same is not true for devices scaled down according to constant voltage scaling.