UH engineers making AI faster, reducing power consumption

(Press-News.org) Addressing the staggering power and energy demands of artificial intelligence, engineers at the University of Houston have developed a revolutionary new thin-film material that promises to make AI devices significantly faster while dramatically cutting energy consumption.  

The breakthrough, detailed in the journal ACS Nano, introduces a specialized two-dimensional (2D) thin film dielectric - or an electric insulator - designed to replace traditional, heat generating components in integrated circuit chips. This new thin film material, which does not store electricity, will help reduce the significant energy cost and heat produced by the high-performance computing necessary for AI. 

“AI has made our energy needs explode,” said Alamgir Karim, Dow Chair and Welch Foundation Professor at the William A. Brookshire Department of Chemical and Biomolecular Engineering at UH. “Many AI data centers employ vast cooling systems that consume large amounts of electricity to keep the thousands of servers with integrated circuit chips running optimally at low temperatures to maintain high data processing speed, have shorter response time and extend chip lifetime.”  

The solution: “Low-k” electronic material 

To keep a lid on power usage while improving performance, Karim and his former doctoral student, Maninderjeet Singh, used Nobel winning organic framework materials to develop these dielectric films.  

“These next-generation materials are expected to boost the performance of AI and conventional electronics devices significantly,” said Singh, a postdoctoral researcher at Columbia University who developed these materials during his doctoral training at UH, in collaboration with Devin Shaffer, a UH professor of civil engineering and doctoral student, Erin Schroeder. 

Not all dielectrics are created equally. Those with high permittivity, or high-k, store more electrical energy and dissipate more of it as heat than those with low-k materials. So, Karim focused on low-k materials made from light elements like carbon, known as lightweight covalent organic frameworks, which speed up signals and reduce delays. 

“Low-k materials are base insulators that support integrated circuit conductors carrying high speed and high frequency electrical signals with low power consumption (i.e. high-efficiency because chips can run cooler and faster!) and also low interference (signal cross talk),” said Karim. 

The team created the new material with carbon and other light elements forming covalently bonded sheetlike films with highly porous crystalline structures. Then, along with another student, Saurabh Tiwary, they studied their electronic properties for next generation low-k applications in devices.  

“Incorporation of low-k materials into integrated circuit devices has the tremendous potential to greatly lower power consumption by the booming AI data centers growth. We discovered that the 2D sheets had an ultralow dielectric constant and ultrahigh electrical breakdown strength needed for high-voltage operation for high power devices, with good thermal stability even at elevated device operating temperatures,” reported Karim and Singh. 

To create the films, Shaffer and Schroeder used a method called synthetic interfacial polymerization, where molecules are dissolved into two liquids that don’t mix and end up stitching molecular building blocks to form the strong crystalline layered sheets. It is a method discovered by 2025 Chemistry Nobel Prize winners Omar M. Yaghi, UC Berkeley professor of chemistry, and other Nobel colleagues. 

The research was funded by the American Chemical Society’s Petroleum Research Foundation New Direction program. 

END