Ga2O3 TCAD Mobility Parameter Calibration Using Simulation Augmented Machine Learning With Physics-Informed Neural Network

Publication Date

1-5-2026

Document Type

Article

Publication Title

IEEE Transactions on Electron Devices

Volume

73

Issue

2

DOI

10.1109/TED.2025.3648986

First Page

775

Last Page

781

Abstract

In this article, we demonstrate the feasibility of performing automatic technology computer-aideddesign (TCAD) parameter calibration and extraction using machine learning (ML), with the machine trained solely by TCAD-simulation data. The methodology is validated using experimental data. Schottky barrier diodes (SBDs) with different effective anode workfunctions (WF) are fabricated with emerging ultrawide bandgap material, Gallium Oxide (Ga2O3), and are measured at various temperatures (T). Their current-voltage (I–V) curves are used for automatic Ga2O3 Philips unified mobility (PhuMob) model parameters calibration. Five critical PhuMob parameters (µmax, µmin, Nref, α, and θ) were calibrated. The machine consists of an autoencoder (AE) and a neural network (NN) and is trained solely by TCAD simulation data with variations in WF, T, and the five PhuMob parameters (seven variations in total). Then, Ga2O3 PhuMob parameters are extracted from the noisy experimental curves. Subsequent TCAD simulation using the extracted parameters shows that the quality of the parameters is as good as an expert’s calibration at the preturned-on regime, but not in the ON-state regime. By using a simple physics-informed NN (PINN), the machine performs as well as the human expert in all regimes.

Funding Number

2046220

Funding Sponsor

National Science Foundation

Keywords

Gallium oxide (Ga₂O₃), machine learning (ML), physics-informed neural network (PINN), simulation, technology computer-aided design (TCAD)

Department

Electrical Engineering

Link to Full Text

Share

COinS