Direct Measurement of Back-Tunneling Current during Program/Erase Operation of Metal–Oxide–Nitride–Oxide–Semiconductor Memories and Its Dependence on Gate Work Function

Abstract

The diffusive propagation of electromagnetic waves in an absorbing random medium is studied and the localization corrections to the diffusion coefficient are calculated. The differences between absorption and inelastic scattering are discussed. Unlike inelastic scattering which introduces a cut-off length into the diffusion coefficient without affecting the total intensity, absorption reduces the intensity but cancels out of the diffusion coefficient. It is predicted that a sharp mobility edge can exist in a sufficiently strongly disordered medium even in the presence of a significant degree of absorption. Novel scaling properties of the transmitted waves at the mobility edge are predicted.

Re-examination of Performance and Reliability Degradation in MONOS Memory with Ultra-thin (~2nm) SiN Charge Trap Layers

“…2. The flatband voltage shift (∆V fb ) is extracted from C-V curves and injected charge (Q inj ) during programming or erasing is measured with a Coulomb meter by correcting the displacement charge with C-V integration [3,4]. Charge loss during data retention is evaluated at 85ºC by analyzing change of V fb under different gate applied bias conditions [5]. Fig.…”

“…This result means that the density of trapped charges (Q trap ) in MONOS never changes with thinning the thickness of SiN from 5 nm to 2 nm, under an assumption that trapped electrons are located at the charge/block interface. It was experimentally confirmed that the trapped electrons are localized at the charge/block interface in the case of SiN ~5 nm [3,4]. Since injected electrons from Si substrate have a finite energy relaxation length necessary for Fowler-Nordheim (FN) tunneling [3][4][5][6][7], it is reasonable that the position of charge centroid in SiN does not move toward the tunnel oxide side by thinning its thickness from 5 nm to 2 nm.…”

Charge Trapping and Reliability Properties of MONOS Memory with High-k Blocking Layer

“…One of the successful methods of charge centroid evaluation is to measure the injected charge during P/E, in addition to the flatband voltage shift [22,23]. In the case of 100% capture of injected charge into a charge trapping layer, the charge centroid is extracted as H H (1) where z eff is the electric (SiO 2 equivalent) distance of the charge centroid from the gate, H ox is the dielectric constant of SiO 2 , 'V fb is the flatband voltage shift, Q trap is the trapped charge density, and Q inj is the injected carrier density.…”

Metallic doping: a new strategy for suppressing shallow-trap centers in vertically stacked charge-trapping flash memories

“…It should be noted that different curves' shifts result from different charging centroid in the CT layer. 35) Owing to higher electrons charging efficiency for traps close to the tunneling oxide, V th shifts' difference is enlarged from 5e-3s to 5e-2s. However, when most traps are charged and the charge-trapping process goes to a saturation region, the difference decreases and arrives to a certain value, which is determined by the charge centroid difference.…”