Methodology for the Simulation of the Variability of MOSFETs With Polycrystalline High-k Dielectrics Using CAFM Input Data

Abstract

In this work, a simulation methodology, whose inputs are Conductive Atomic Force Microscope (CAFM) experimental data, is proposed to evaluate the impact of nanoscale variability sources related to the polycrystallization of high-k dielectrics (i.e., oxide thickness, tox, and charge density, ρox, fluctuations in the nanometer range) on the Metal-Oxide-Semiconductor Field Effect Transistor (MOSFET) variability. To simulate this variability, a Thickness And Charge MAp Generator (TACMAG) has been developed and used in combination with an in-house-built 3D device simulator (VENDES). From CAFM experimental data (topography and current) obtained on a small area of a given polycrystalline dielectric, the TACMAG generates a high amount of tox and ρox configurations of the gate dielectric, with identical statistical characteristics to those experimentally measured. These dielectrics are then introduced into the device simulator, with which the impact of the tox and ρox fluctuations in the dielectric on the variability of MOSFETs (i.e., threshold voltage) is analyzed. The proposed methodology has been verified by comparing the Vth variability of MOSFETs whose dielectric properties were generated with TACMAG to other devices whose gate oxide characteristics have been determined experimentally with CAFM. Finally, it has been used to analyze the impact of different nanoscale parameters, such as the Grain size and Grain Boundaries depth (of polycrystalline dielectrics) on such variability.

A General Toolkit for Advanced Semiconductor Transistors: From Simulation to Machine Learning

“…Previously, we have applied VENDES to model MOS-FETs [12], FinFETs [13], gate-all-around (GAA) nanowire (NW) FETs [14] and nanosheet (NS) FETs [15], among others. As an example, Fig.…”