A Method for Automatically Predicting the Radiation-Induced Vulnerability of Unit Integrated Circuits

Area of Science:

• Semiconductor technology
• Integrated circuit reliability
• Radiation effects in electronics

Background:

• Modern semiconductor technology faces increased susceptibility to particle radiation due to reduced operating and threshold voltages.
• Shrinking process sizes exacerbate charge sharing effects, making single event effects a primary cause of integrated circuit failures.
• Effective sensitivity evaluation methods are crucial for optimizing integrated circuit design and enhancing irradiation reliability.

Purpose of the Study:

• To develop an accurate method for evaluating the sensitivity of integrated circuits to particle radiation.
• To address the growing challenge of soft errors caused by charge sharing effects in advanced semiconductor devices.
• To improve the prediction and mitigation of single event effects in integrated circuits.

Main Methods:

• Established a device model for charge sharing effects and performed simulations.
• Developed a neural network model to predict charge amounts in primary and secondary devices.
• Proposed and validated a comprehensive automated method for calculating soft errors in unit circuits using TCAD simulations.

Main Results:

• The proposed automated method for soft error calculation achieved an error margin of 2.8-4.3% when validated against TCAD simulations.
• The study successfully modeled charge sharing effects and predicted charge amounts using a neural network.
• Demonstrated the accuracy and effectiveness of the developed soft error calculation method.

Conclusions:

• The developed automated method provides an accurate and effective approach for calculating soft errors in integrated circuits.
• The findings contribute to optimizing integrated circuit design and improving experimental methods for single event effects.
• This research enhances the understanding and mitigation of radiation-induced failures in semiconductor devices.