Analysis of fluctuations in semiconductor devices

Abstract

The random nature of ion implantation and diffusion processes as well as inevitable tolerances in fabrication result in random fluctuations of doping concentrations and oxide thickness in semiconductor devices. These fluctuations are especially pronounced in ultrasmall (nanoscale) semiconductor devices when the spatial scale of doping and oxide thickness variations become comparable with the geometric dimensions of devices. In the disseration, the effects of these fluctuations on device characteristics are analyzed by using a new technique for the analysis of random doping and oxide thickness induced fluctuations. This technique is universal in nature in the sense that it is applicable to any transport model (drift-diffusion, semiclassical transport, quantum transport etc.) and it can be naturally extended to take into account random fluctuations of the oxide (trapped) charges and channel length.

The technique is based on linearization of the transport equations with respect to the fluctuating quantities. It is computationally much (a few orders of magnitude) more efficient than the traditional Monte-Carlo approach and it yields information on the sensitivity of fluctuations of parameters of interest (e.g. threshold voltage, small-signal parameters, cut-off frequencies, etc.) to the locations of doping and oxide thickness fluctuations. For this reason, it can be very instrumental in the design of fluctuation-resistant structures of semiconductor devices.

Quantum mechanical effects are taken into account by using the density-gradient model as well as through self-consistent Poisson-Schrödinger computations. Special attention is paid to the presenting of the technique in a form that is suitable for implementation on commercial device simulators. The numerical implementation of the technique is discussed in detail and numerous computational results are presented and compared with those previously published in literature.