Methodology to Predict Random Telegraph Noise Induced Threshold Voltage Shift Using Machine Learning

Abstract

We suggest the methodology to predict the distribution of threshold voltage (V-t) shift caused by random telegraph noise (RTN). Poly-silicon channels were randomized with a single trap and the neural network was modeled to predict RTN trap-induced Vt fluctuation in 3D NAND Flash Memory. 3D Technology Computer-Aided Design (TCAD) simulations were performed in a unit cell to calculate the Vt shift in a 3D vertical channel. Finally, we extract the distribution of Vt fluctuation using machine learning.