Vertical GaN Diode BV Maximization through Rapid TCAD Simulation and ML-enabled Surrogate Model

TL;DR

This paper introduces a rapid TCAD simulation method and an ML surrogate model to efficiently maximize the breakdown voltage of vertical GaN diodes, achieving near-ideal BV values with reduced computational effort.

Contribution

It presents a novel combination of accelerated TCAD simulation and ML-based surrogate modeling for optimized diode design, surpassing traditional methods in speed and performance.

Findings

5X faster accurate simulation method developed

50% more high BV designs identified within same time

ML surrogate model achieved 89% of ideal BV in inverse design

Abstract

In this paper, two methodologies are used to speed up the maximization of the breakdown volt-age (BV) of a vertical GaN diode that has a theoretical maximum BV of ~2100V. Firstly, we demonstrated a 5X faster accurate simulation method in Technology Computer-Aided-Design (TCAD). This allows us to find 50% more numbers of high BV (>1400V) designs at a given simulation time. Secondly, a machine learning (ML) model is developed using TCAD-generated data and used as a surrogate model for differential evolution optimization. It can inversely design an out-of-the-training-range structure with BV as high as 1887V (89% of the ideal case) compared to ~1100V designed with human domain expertise.