Engineers need new materials to make computer chips—and the devices they power—even smaller and more efficient.
Faculty members from the University of Dallas, collaborators, and industry partners have teamed up to design and test indium-based materials to manufacture the next generation of computer chips.
The researchers have received a $1.9m, three-year grant through the National Science Foundation’s Future of Semiconductors (FuSe2) programme to support their work.
The UTD funding is part of $42.4m in FuSe2 grants announced in September to support the goals of the federal CHIPS (Creating Helpful Incentives to Produce Semiconductors) and Science Act of 2022.
This aims to make computer chips more energy-efficient and facilitate the domestic production of integrated circuits.
Improving the performance of computer chips
By introducing indium-based materials, the researchers aim to facilitate patterning in the extreme ultraviolet (EUV) range. Patterning, or lithography, is a key step in the semiconductor fabrication process in which patterns are created on the surface of a wafer to serve as pathways for transistors and other components.
Moving from deep UV to EUV range makes it possible to produce smaller, more precise features on computer chips for better performance and energy efficiency.
During the traditional patterning process in the manufacturing of computer chips, silicon wafers are coated with a removable layer of material called a photoresist before being exposed to UV photons.
The next generation of lithography uses very high-energy photons — 92 electronvolts — in the EUV region. Due to the high energy of these photons, conventional photoresist materials will not work.
The researchers’ new materials also could enable the production of 3D circuits, which are designed by stacking layers of chips like high-rises in a crowded city.
New materials are needed to build added layers on a 3D chip without disturbing the existing circuits.
Better semiconducting properties without overheating
“If you are making a layer of devices on top of another layer of devices, you cannot heat it to a high temperature. Otherwise, you will destroy the existing layers,” explained Dr Julia Hsu, professor of materials science and engineering and leader of the study.
Using indium-containing materials for the EUV photoresist and the transistors should lead to more efficiency in computer chips by eliminating a step in integrated circuit manufacturing that involves solvents.
The researchers are testing a technique called photonic curing to convert EUV-patterned structures to nanoscale devices.
Photonic curing uses pulses of light at high intensity but low energy to complete the chemical reactions that allow the indium oxide to achieve better semiconducting properties without overheating the underlying devices.
Incorporating machine learning into the process
Hsu’s preliminary work on indium-containing materials as an EUV photoresist has been supported by a Semiconductor Research Corporation (SRC) grant to investigate new materials for computer chips.
Hsu also plans to incorporate machine learning — a method she learned with support from a 2023 Simons Foundation Pivot Fellowship — into the project’s design and testing methodologies.
“The FuSe2 project will enable us to take our preliminary results from the SRC project to a much higher level and bigger impact,” Hsu said.
“We will bring computation and synthetic chemistry to expand beyond currently commercially available materials.”
Key collaborations in the project
Hsu’s co-principal investigators include Dr Cormac Toher, an assistant professor of materials science and engineering and computational materials scientist, and Dr Kevin Brenner, an assistant professor of materials science and engineering.
Toher will design the indium-containing molecules, and Brenner will fabricate and test the devices.
The project also includes semiconductor industry workforce training for community college students through UTD’s North Texas Semiconductor Institute and a class that Hsu will teach as an immersive experience in the semiconductor industry.
