GLCT manufactures high-performance synthetic diamonds for electronic and quantum applications using Chemical Vapor Deposition (CVD). Diamond’s thermal conductivity, wide bandgap, and defect tolerance make it ideal for power electronics, quantum sensing, and photonics. However, optimizing CVD growth remains difficult due to nonlinear interactions among gas composition, plasma dynamics, substrate geometry, and thermal gradients. Traditional experimentation is slow, resource-intensive, and difficult to scale.
This project proposes a machine learning framework for systematic optimization of diamond growth recipes. The first component applies Gaussian Process Regression with Bayesian Active Learning to recommend parameters that maximize growth rate. The second introduces a Vision Transformer to predict strain-related optical quality from birefringence images and cluster labels. Controlled next-frame baselines validate pixel-level prediction and assess when temporal extrapolation succeeds or fails, informing future monitoring. Structured process variables, plasma conditions, substrate parameters, and optical metrics are sourced from GLCT’s dataset. This framework reduces experimentation time, improves yield consistency, and enables scalable, data-driven optimization of diamond CVD processes.
Watch the team present this project at 01:49:40 in the session recording here.
Keywords: Chemical Vapor Deposition (CVD), single-crystal diamond, birefringence imaging, Gaussian Process Regression (GPR), Bayesian optimization, active learning, Vision Transformers (ViT), growth recipe optimization, quantum materials, semiconductor process modeling, machine learning for materials science, strain classification, optical quality prediction, Bayesian neural networks (BNN), PyMC, self-driving laboratories
Faculty Advisor
Batu Gundogdu holds a Ph.D. in Electrical Engineering and brings over 15 years of experience developing deep-learning models across the fields of speech, language, and medical imaging. Most recently, he served as Research Faculty in the Department of Radiology at the University of Chicago for four years, where he led AI efforts at the MRI Research Center. Dr. Gundogdu is currently a Senior AI Engineer at Eva, a startup building next-generation AI compute platforms. In the UChicago MS in Applied Data Science program, he shares his practical insights and experiences with students on being an AI engineer and scientist across diverse venues—including research institutions (UChicago MRI Research Center), corporate companies (Analog Devices – ADI), government (Navy and NATO), and startups (Eva). He teaches courses on Bayesian Machine Learning for Generative AI and Advanced Linear Algebra for Machine Learning.