New Materials Could Keep Electrons Moving in Tiny Chips

A team led by Northwestern Engineering’s James Rondinelli studied a group of materials that could make future computer chips faster and use less energy. These materials help electricity move smoothly through wires that are just a few nanometers wide, which is thousands of times thinner than a human hair. 

The findings could help make future computer chips more energy-efficient and powerful. As technologies like AI, cloud computing, and large data centers use more electricity, better interconnect materials—the conductive substances used to link different components or devices together so electrical signals and power can flow between them—could reduce energy use and heat in microchips.

Current chips rely on copper wiring, which struggles as circuits get smaller. The new materials maintain strong electrical flow at very small scales, reducing heat and energy loss while supporting the continued shrinking of electronic components.

The researchers developed a computational workflow that simulates how electrons move through ultrathin nanowires. The approach allows hundreds of topological materials to be evaluated quickly while accounting for surface roughness, disorder, and orientation. The study highlights binary conductors such as titanium sulfide, zirconium diboride, molybdenum carbide, tungsten carbide, and several mononitrides including molybdenum, tantalum, and tungsten.

“This work gives us a clearer picture of which of this unique quantum materials could actually work in nanoscale devices,” Rondinelli said. “By incorporating realistic defects and variations into our simulations, we can identify candidates that are likely to retain high conductivity even after imperfections from manufacturing.”

Rondinelli is the Walter Dill Scott Professor of Materials Science and Engineering at the McCormick School of Engineering. He presented his work in the paper “Accelerated Discovery of Topological Conductors for Nanoscale Interconnects” last month in the academic journal Advanced Science.

The leader of the Materials Theory and Design research group, Rondinelli collaborated with IBM to push the project toward practical applications. The partnership, Rondinelli said, provided opportunities to combine computational screening with experimental validation, ensuring that theoretical discoveries can be tested and applied in actual chip manufacturing.

“Working with IBM has led us to better understand device-relevant constraints, ensuring materials we screen for are compatible with semiconductor fabrication steps and back-end-of-line process temperatures,” Rondinelli said.

The researchers also identified the most promising ways to grow and test these materials, including which crystal surfaces to use, how to arrange the structures, and which elements to add. This provides a practical roadmap for turning computer simulations into real devices. 

Keeping electrons moving quickly at such tiny scales is also important for continuing the trend of making chips smaller and more capable.

This work builds on Rondinelli’s earlier studies of individual compounds such as molybdenum nitride and silicide and related topological materials. The new study looks at many more compounds and employs automated methods that account for imperfections and surface effects. Together, these advances turn discovery from a slow, trial-and-error process into a faster, more systematic approach for identifying materials ready for practical use.

The research also leverages machine learning to predict which chemical, crystallographic, and topological features are associated with strong surface conduction.

“The machine learning models allow us to predict which chemical and structural features are most important for maintaining conduction on the surfaces of these materials at the nanoscale,” Rondinelli said. “These models provide an initial framework for designing new interconnect compound with targeted properties.”

The next steps for the research include making and testing the most promising materials, especially titanium sulfide, zirconium diboride, and molybdenum nitride. The team also plans to study more complex compounds, see how the materials behave at normal operating temperatures, and check that they can work with standard chip-making processes, including staying stable with insulating layers and resisting damage from electrical currents over time.