Powernet: SOI Lateral Power Device Breakdown Prediction With Deep Neural Networks

Jing Chen,Yufeng Guo,Maolin Zhang,Yibo Lin,Jun Zhang,David Z Pan,Mohamed Baker Alawieh

doi:10.1109/access.2020.2970966

Abstract

The breakdown performance is a critical metric for power device design. This paper explores the feasibility of efficiently predicting the breakdown performance of silicon on insulator (SOI) lateral power device using multi-layer neural networks as an alternative to expensive technology computer-aided design (TCAD) simulation. In this work, we propose the first breakdown performance prediction framework, PowerNet, for SOI lateral power devices, based on deep learning methods. The framework can provide breakdown location prediction and breakdown voltage (BV) prediction by utilizing a two-stage machine learning method. In addition, it demonstrates 97.67% accuracy on breakdown location prediction and less than 4% average error on the BV prediction compared with TCAD simulation. The proposed method can be used to measure changes in performance caused by random variability in structural parameters during manufacturing process, allowing designers to avoid unstable structural parameters and enhance design robustness. More importantly, it can significantly reduce the computational cost when compared with the TCAD simulation. We believe the proposed machine learning technique can significantly speedup the design space exploration for power devices, eventually reducing the overall product-to-market time.

Highlights

Silicon on insulator (SOI) lateral power devices have been widely adopted in power integrated circuits due to their high breakdown voltage (BV), excellent isolation, and low power consumption [1]–[3]
The breakdown location prediction accuracy using deep neural networks (DNN) which comprises the first stage in PowerNet is shown, and the accuracy is compared with different classification algorithms
5) RESULTS ON BV PREDICTION COMPARED WITH technology computer-aided design (TCAD) SIMULATOR Fig. gives an example showing that the BV changes sharply as the doping concentration changes between ×1015cm−3 to ×1015cm−3, which tends to happen in the manufacturing process

Summary

INTRODUCTIONS

Silicon on insulator (SOI) lateral power devices have been widely adopted in power integrated circuits due to their high breakdown voltage (BV), excellent isolation, and low power consumption [1]–[3]. Zeng et al [5] adopted Silvaco to obtain the breakdown voltage of lateral MOSFET These simulation tools acquire the breakdown performance by solving the physical equations such as Poisson equation and the current continuity of electrons and holes at the set mesh point. CarrilloNuñez et al [21] proposed the prediction of the effect of statistical variability in Si junctionless nanowire transistors using a multi-layer neural network (NN), which can greatly save computational cost These results have demonstrated the potential of machine learning approaches in device performance predictions. We propose a first stage classification model to predict the breakdown location followed by a second stage location-specific regression to predict breakdown voltage for SOI lateral power device.

PRELIMINARIES

GAUSSIAN PROCESS REGRESSION FOR BV PREDICTION

EXPERIMENTAL RESULTS

CONCLUSION