Scientists develop semiconductor material for energy-efficient AI computing
A team of scientists from the University of Michigan unveiled a 5nm ferroelectric semiconductor material, paving the way for more efficient computation and quantum computing development. The researchers believe this semiconductor material can potentially expand the performance of artificial intelligence and sensing solutions. The researchers say it can also enable batteryless IoT devices that are increasingly becoming integral to the smart ecosystem.
“This will allow the realization of ultra-efficient, ultra-low-power, fully integrated devices with mainstream semiconductors,” said Zetian Mi, a University of Michigan professor of electrical and computer engineering and co-author of the study. “This will be very important for future AI and IoT-related devices.”
Previously, Mi’s group had demonstrated ferroelectric behavior in a semiconductor composed of aluminum nitride spiked with scandium, an element often utilized to strengthen aluminum in performance bikes and fighter jets. However, they realized that the material could only be used for advanced computing devices if made into films thinner than 10 nanometers.
To achieve a 5-nanometer ferroelectric semiconductor material, the researchers used a technique called molecular beam epitaxy. The same method produces the semiconductor crystals that power the lasers in CD and DVD players.
“With this thinness, we can really explore the minuscule physics interactions,” said Ping Wang, University of Michigan, a research scientist in electrical and computer engineering. “This will help us to develop future quantum systems and quantum devices.”
The properties of ferroelectricity have been previously explored through several studies, all of which concluded that the complexity of these materials limits their usage in integrated circuits. However, this new category of material demonstrates electrical, mechanical and thermal properties that may help develop next-gen energy-efficient devices when integrated into semiconductor processes.
Ferroelectric materials can change their polarity and thus switch positive and negative ends. According to the press release from the University of Michigan notes, this feature enables the material to be used in various applications, including sensing light and acoustic vibration.
The study was funded by the US Department of Defense Advanced Research Projects. Future plans involve manufacturing nanoscale ferroelectric materials to leverage their optical and acoustic properties for quantum technologies.
Article Topics
chip | edge AI | energy efficiency | manufacturing | University of Michigan
Comments