Non-Gaussian behavior and anticorrelations in ultrathin gate oxides after soft breakdown

Abstract

The time dependence of the gate voltage VG(t) after soft breakdown of metal-oxide-semiconductor capacitors with a 2.4 nm SiO2 layer has been measured. It is found that the VG(t) fluctuation distributions are non-Gaussian, but can be described by a Lévy stable distribution. The long-range correlations in VG(t) are investigated within the detrended fluctuation analysis. The Hurst exponent is found to be H=0.25±0.04 independent of the value of the stress current density J. It is argued that these are universal features of soft breakdown and are due to trapping–detrapping of electrons in and away from the primary percolation path.

A percolative model of soft breakdown in ultrathin oxides

“…Since a further increase of the stress voltage causes the film to break down, the emerging non-Gaussian distribution can be thought of as a precursor of failure as observed in experiments [1].…”

“…Starting from the intrinsic value of the network [6], the variance is found to increase significantly as a net effect of the stressing voltage which is ultimately responsible of the breakdown at about 0.9 V . Recent observations of SILC measurements reported non-Gaussian current fluctuations in the soft breakdown region of ultra-thin dielectrics [1]. Therefore, we computed the distribution function of the current fluctuations P (I) for stress voltages of 0.5 V and 0.85 V and plotted them in Fig.…”

“…This process manifests itself by a large increase of the stress-induced leakage currents (SILC) and by the occurrence of giant current fluctuations associated with resistance fluctuations. These fluctuations can be related to trapping and detrapping of electrons in the percolation cluster formed between the electrodes of the capacitor at the soft breakdown [1,2]. In this paper, we investigate the resistance in the pre-breakdown region and the associated fluctuations using two percolations which evolve in competition on a random resistor network [3], as briefly summarized below.…”

Statistical properties of the seismic noise field: influence of soil heterogeneities

“…This kind of analysis is based on the random walk theory (Shlesinger 1987) and has been found to be almost unbiased for fractal noises and motion (Witt & Malamud 2013), providing an important tool for the detection of long-range (auto-)correlation properties in time-series with nonstationarities. It has been successfully applied to such diverse fields of interest as DNA analysis (Buldyrev et al 1995;Peng et al 1995), heart rate dynamics Ho et al 1997), long-time weather records (Koscielny-Bunde et al 1998;Talkner & Weber 2000), cloud structure (Ivanova & Ausloos 1999) and even solid state physics (Kantelhardt et al 1999;Vandewalle et al 1999).…”

Magnitude and sign correlations in conductance fluctuations of horizontal oil water two-phase flow

“…They concluded that the long-range correlations of magnitude series indicate nonlinear behavior and the sign time series mainly relate to linear properties of the original series. The time series analysis methods based on long-range correlation has been successfully applied to diverse fields such as DNA sequences [10][11], long-time weather records [14], geology [15], ethnology [16], physiological dynamics [17] and economics time series [18].…”