UIC researchers chosen for feature article in Journal of Vacuum Science & Technology A
block
After collecting articles focused on Atomic Layer Disposition (ALD) and AI, Department of Chemistry PhD student Pouyan Navabi and UIC Richard and Loan Hill Department of Biomedical Engineering and Department of Chemical Engineering Professor Christos Takoudis identified and explained how to fill these gaps. From there, their article, published in the Journal of Vacuum Science & Technology A, was chosen as a featured article by the journal editors.
After their review article, Takoudis and Navabi are also expanding their research to use machine learning for process optimization of ALD in collaboration with Kurt J. Lesker, a well-known vacuum systems company. They also reached out to collaborators at Kansas State University who helped to explore different possibilities of AI and ALD.
Takoudis shared that this research process began slowly, but once Navabi joined, it took off.
“We see that many companies are working on AI and hardware and technology, such as NVIDIA, are growing so fast, but there is a problem now,” Takoudis said. “All of these companies are depending on the Taiwan Semiconductor Manufacturing Company (TSMC), the world’s largest dedicated chip foundry.”
“If you look at the hierarchy of the AI companies, they are all dependent on TSMC and Samsung,” he said. “Building is a core part of making semiconductor materials and the U.S. market is behind in the race. If these companies cannot collaborate with TSMC, they may fall and crash, so it is very important to get ahead in this field, and our effort is going to help a lot.”
Their current research involves creating a model to optimize the system. Specifically, the saturation time and precursor, which they hope will add better accuracy in predicting the saturation time and has a higher efficiency, compared with traditional machine learning methods.
Navabi hopes that this review paper helps the ALD community as a starting point to think more about how to implement AI and machine learning in research.
He also hopes their research reaches the point where someone approaches them to request a material that should have specific properties. From there, they can determine what is needed to be done in terms of conditions, time, and economics to deliver the product as promised.
Their research involves being able to observe in real time, how the system and how the instrument behaves and performs.
They are also collaborating with UIC BME Visiting Professor Urmila Diwekar, CHE former PhD graduate Harshdeep Bhatia, University of Pittsburgh School of Dental Medicine Professor Cortino Sukotjo, Kansas State University Professor of Mathematics and Data Science Majid Jaberi-Douraki, and KSU Graduate Faculty Associate/Fellow Remya Ampadi Ramachandran.