Bi-Directional Long Short-Term Memory Neural Network Modeling of Data Retention Characterization in 3-D Triple-Level Cell NAND Flash Memory

Abstract

— Data retention (a time-variant characteristic of 3-D-NAND flash memory) is predicted through a bi-directional long short-term memory (LSTM) neural network (NN) model that learns sequential data obtained from chip measurements of a triple-level cell (TLC). The predicted results for all time points of each program (PGM) state are accurately predicted by the threshold voltage (Vth) distribution. Thus, the predicted Vth can be used to analyze the cause of retention failure. When the Vth of the target cell is high or when that of the adjacent cell is small, the Vth loss of the target cell is large. In addition, the Vth loss increases as the Vth of the adjacent cell decreases. Using a fully calibrated TCAD simulation, we verify the NN-based Vth prediction by checking the change in the electron concentration in the nitride layer. Furthermore, the NN model predicts the Vth for cells existing in other blocks, showing that they are consistent with the measured Vth. The prediction times were 5 × 105 s, 5 × 106 s, and 2 × 106 s, but using machine learning (ML), we reduced the time required to predict the Vth to only 2 s. Therefore, the proposed ML method enables fast, accurate, and effective predictive modeling of the time-variant Vth of 3-D TLC NAND flash memory. Finally, the predicted Vth can be included in the read retry table or included in the lookup table of the compensation circuit in NAND solutions. This can save a significant amount of time that would otherwise be spent on actual long-term measurements.