Prediction of Single Event Effect in Inverter Circuit Based on Deep Learning

Fully Depleted Silicon on Insulator (FDSOI) technology can solve the short channel effect very effectively, with low power consumption, and low voltage, and can improve the subthreshold characteristics of the device. In addition, FDSOI devices have good radiation resistance, which has become an important research object in the field of device research. Single event effect (SEE) is an important index of radiation resistance of FDSOI devices. At present, the research on SEE of FDSOI devices typically employs heavy-ion irradiation experiments and TCAD software simulations. Taking FDSOI technology as an example, this paper presents a research method of device modeling and performance prediction based on deep learning. The accuracy of the peak of transient current $(I_{peak})$ predicted by this method is 96.45%, the accuracy of total collected charge $(Q_{total})$ is 97.86%, and the determination coefficient of drain transient current pulse (It) is 0.97717. This method can obviously improve the simulation speed and reduce the calculation cost, and provide a new feasible method for the research of FDSOI devices.