Monte Carlo Simulator for Threshold Voltage Distribution of 3-D nand Flash Memory Using Machine Learning
Jang-Kyu Lee, Eunseok Oh, Hyungcheol Shin
Abstract
In this article, we propose a machine learning model-based simulator and method for predicting the threshold voltage ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${V}_{\text {t}}$ </tex-math></inline-formula> ) distribution of 3-D NAND flash memory. The proposed machine learning modeling method aims to predict each incremental step pulse program (ISPP) slope after ensuring the model’s accuracy through training and test using only a small subset of the data from numerous devices that require prediction. As a result of model verification during the test phase of this model after training, the maximum error rate was 2.82%, confirming that high accuracy for prediction was achieved. Using the verified machine learning model, Monte Carlo simulations of random strings are performed, taking into account the factors that influence the formation of the <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${V}_{\text {t}}$ </tex-math></inline-formula> distribution. The completed simulator demonstrates the ability to predict <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${V}_{\text {t}}$ </tex-math></inline-formula> distribution in various environments, such as quad-level cell (QLC) and penta-level cell (PLC) operations.