Algorithm prediction of single particle irradiation effect based on novel TFETs.

Area of Science:

• Semiconductor device physics
• Radiation effects in electronics
• Artificial intelligence in engineering

Background:

• Single particle irradiation poses a risk to the reliability of semiconductor devices like tunnel field-effect transistors (TFETs).
• Predicting transient effects caused by such irradiation is crucial for ensuring device performance and longevity.

Purpose of the Study:

• To develop and validate a deep learning algorithm for predicting single particle transient characteristics in TFETs.
• To assess the performance of the deep learning model against traditional machine learning methods and TCAD simulations.

Main Methods:

• A deep learning network model was constructed to predict key parameters of single particle transients.
• Computer-Aided Design (TCAD) simulations were employed to study the impact of single particle effects on stacked source trench gate TFETs.
• The deep learning model's predictions were compared against five traditional machine learning algorithms.

Main Results:

• Increased drain voltage, heavy ion incident width, and linear energy transfer led to higher drain transient current and collected charge.
• The deep learning model achieved high accuracy, with relative error percentages for drain current pulse peak (I_DMAX) and collected charge (Q_c) below 10%, and most predictions under 1%.
• The deep learning algorithm demonstrated superior performance and lower average error compared to support vector machines, decision trees, K-nearest neighbors, ridge regression, and linear regression.

Conclusions:

• The developed data-driven deep learning model effectively predicts single particle transient data for TFETs under various conditions.
• This model serves as a valuable tool for researchers and digital circuit designers to assess device reliability under irradiation.
• Deep learning shows significant promise for advancing the prediction of irradiation effects in semiconductor devices.