Investigation of stochastic implementation of Reaction Diffusion (RD) models for NBTI related interface trap generation

Abstract

Conventional H/H₂ and poly H/H₂ Reaction-Diffusion (RD) models are compared and the poly version is explored as a more physically likely model for predicting interface trap (N_IT) generation during Negative Bias Temperature Instability (NBTI) in p-MOSFETs. Stochastic implementations of the conventional H/H₂ RD model and the poly H/H₂ RD model are realized and their equivalence to continuum implementations are investigated for large area devices. Impact of dimensionality (1D, 2D, 3D) and device size (W, L) are explored for stochastic implementation. Stochastic simulations for small area devices using the poly H/H₂ RD model show long term 1/6 power law time exponent during stress. A comprehensive framework consisting of H/H₂ RD model for N_IT along with empirical models for hole trapping (N_HT) and bulk trap generation (N_OT) is able to predict experimental data for a wide variety of large area devices for different experimental conditions. Variation of small area device degradation has been simulated and compared to experimental results.