Logarithmic behavior of degradation dynamics in metal--oxide semiconductor devices

TL;DR

This paper presents a theoretical model showing that degradation in metal-oxide semiconductor devices follows a universal logarithmic relaxation law, with amplitude influenced by temperature and Fermi level, supported by simulations and literature data.

Contribution

The authors introduce a simple statistical model that predicts a universal logarithmic degradation law in MOS devices, validated by simulations and experimental data.

Findings

Degradation density follows a logarithmic law over time.

Amplitude of degradation depends on temperature and Fermi level.

Model is supported by Monte Carlo simulations and literature experiments.

Abstract

In this paper the authors describe a theoretical simple statistical modelling of relaxation process in metal-oxide semiconductor devices that governs its degradation. Basically, starting from an initial state where a given number of traps are occupied, the dynamics of the relaxation process is measured calculating the density of occupied traps and its fluctuations (second moment) as function of time. Our theoretical results show a universal logarithmic law for the density of occupied traps $\bar{<n(t)>} \sim \phi (T,E_{F}) (A+B \ln t)$, i.e., the degradation is logarithmic and its amplitude depends on the temperature and Fermi Level of device. Our approach reduces the work to the averages determined by simple binomial sums that are corroborated by our Monte Carlo simulations and by experimental results from literature, which bear in mind enlightening elucidations about the physics of degradation of semiconductor devices of our modern life.