Hot-Carrier Injection Degradation of 1.8V nMOSFET and a New Predication Model Based on LSTM

The hot-carrier injection (HCI) degradation mechanism of 1.8V n-channel metal-oxide semiconductor field-effect transistor (nMOSFET) is studied, and the degradation tendency of the saturated drain current (
$I$
dsat
) under different gate bias stresses is shown. Combining with the experimental data, a neural network model based on the long short-term memory (LSTM) for predicting the behavior of the HCI degradation is trained and tested. The model reaches a fairly good predictive ability (R2 = 0.974 and mean absolute error = 0.06686). The degradation rate δI
dsat
(t) from 0% to 6% is successfully used to predict the remaining degradation process until 10% shift of
$I$
dsat
(t) based on the model, which can obviously decrease the testing time and labor cost.