Improvement of TCAD Augmented Machine Learning Using Autoencoder for Semiconductor Variation Identification and Inverse Design

Abstract

A machine learning (ML) model by combing two autoencoders and one linear regression model is proposed to avoid overfitting and to improve the accuracy of Technology Computer-Aided Design (TCAD)-augmented ML for semiconductor structural variation identification and inverse design, without using domain expertise. TCAD-augmented ML utilizes TCAD simulations to generate sufficient data for ML model development when experimental data are inadequate. The ML model can then be used to identify semiconductor structural variation for given experimental electrical measurements. In this study, the variation of layer thicknesses in the p-i-n diode is used as a demonstration. An ML model is developed to predict the diode layer thicknesses based on a given Current-Voltage (IV) curve. Although the variations of interest can be incorporated easily in TCAD simulations to generate ML training data, the TCAD-augmented ML model generally is overfitted and cannot predict the variations in experiment well due to hidden variables which also alters the IV curves. We show that by using an autoencoder, this problem can be solved. To verify the effectiveness, another set of TCAD simulation data is generated with hidden variables (dopant concentration variation) to emulate experimental data. Testing on the second set of data shows that the proposed model can avoid overfitting and has up to 15 times improvement in accuracy in thickness prediction. Moreover, this model is used successfully to perform inverse design and can capture an underlying physics that cannot be described by a simple physical parameter.

Multilayer perceptron–random forest based hybrid machine learning–neural network model for GaN high electron mobility transistor's parameter estimations

“…The direct feature engineering in ML and NN algorithms enables straightforward interpretation and adaptability of such techniques to varying device parameters, which gives an edge over other methods and is the focus of this work. As reported in the literature, TCAD simulation based data is one of the widely used approaches for testing and validating the model results 24,72–74 . This work presents an inclusive study covering the automation of device's gain, width and power metrics for AlGaN/GaN HEMTs based on TCAD data.…”

“…Moreover, it is one of the best alternatives for data generation when working on ML based problems. [24][25][26] The simulation's speed, convergence, and accuracy are governed by the compact model of the device used to describe the device's physics and various performance parameters. There exist several models for AlGaN/GaN HEMTs such as charge based analytical models, surface voltage based empirical models, and physics based compact models.…”

“…As reported in the literature, TCAD simulation based data is one of the widely used approaches for testing and validating the model results. 24,[72][73][74] This work presents an inclusive study covering the automation of device's gain, width and power metrics for AlGaN/GaN HEMTs based on TCAD data. Physical insights into different variables affecting the device's performance are necessary to understand and are also covered under this work.…”

Prediction of Static Characteristic Parameters of an Insulated Gate Bipolar Transistor Using Artificial Neural Network

“…Recently, machine learning techniques for predicting the electrical characteristic parameters of the semiconductor device have been booming due to their ability to learn the relationship between structural parameters and characteristics efficiently [8][9][10][11][12][13]. However, most work is limited to providing only one characteristic parameter, such as the threshold voltage of a junctionless nanowire transistor [11] or the breakdown voltage of a lateral power device [12].…”

Variation Model Abstraction and Adaptive Control Based on Element Description Method Toward Smart Factory

“…An effective method for reducing this variability is to model the variability and compensate based on the model. The above method has been studied in various fields such as chemical-mechanical polishing processes [10], roll-to-roll manufacturing [11], and semiconductor fabrication [12].…”