HKU team makes brain-like silicon carbide chip for cryogenic computing
The device can mimic neuron-like electrical spikes at 10 millikelvin, pointing to lower-heat electronics for quantum processors.
By Lucas Ferreira · Science & Environment Writer
3 min read
Researchers at the University of Hong Kong have developed a brain-inspired chip platform that can operate near absolute zero, the university said. The work matters because quantum computers need control electronics that can function close to extremely cold qubits without adding too much heat.
The team, led by Professor Yuhao Zhang and PhD student Xin Yang, built the cryogenic neuromorphic platform using industry-standard silicon carbide MOSFETs, according to HKU. In tests, a single transistor produced energy-efficient electrical spiking behavior similar to the signaling of biological neurons at temperatures as low as 10 millikelvin.
The findings were published in Nature Communications in a paper titled “Cryogenic neuromorphic circuits using gate-controlled negative differential resistance in silicon carbide.” The authors include Yang, Zhang and collaborators from HKU’s Department of Electrical and Computer Engineering and the Centre for Advanced Semiconductors and Integrated Circuits.
Why quantum computers need colder electronics
Quantum computers rely on qubits that must be held at millikelvin temperatures. HKU said current silicon-based control electronics use substantial power and generate heat, so they are often placed away from the qubits.
That separation adds wiring between the control systems and the quantum processor. According to the university, the extra wiring can hurt performance and complicate efforts to build larger quantum computers.
Zhang said the platform could be placed alongside quantum processors. He said silicon carbide carrier dynamics allow circuits that are “thousands of times more energy-efficient than conventional electronics,” reducing the heat burden on cryogenic systems.
A transistor used in a new way
The HKU team reported a method for producing and controlling negative differential resistance, or NDR, in silicon carbide MOSFETs. The researchers found a strong S-shaped NDR effect when the devices were cooled below 2 kelvin.
HKU said the effect is driven by electron-donor impact ionization. Unlike approaches that rely on heat generated inside a device, the university said this mechanism comes from the material’s atomic properties, making it stable and reproducible across different manufacturing batches.
Yang said the approach is robust and scalable. Because silicon carbide is already used in electric vehicles and power grids, he said existing industrial foundries could be used to make the cryogenic chips on 300-millimeter wafers.
Possible uses beyond quantum processors
The researchers also showed that the artificial neuron devices can be cascaded into larger networks, according to HKU. That could allow local processing inside cold systems and support tasks such as real-time quantum control and quantum error correction.
HKU said the technology may also have uses in deep-space missions because the circuits are designed for extremely cold environments. The university cited possible future systems operating on the Moon’s surface or in distant parts of the solar system.
The work remains a research demonstration, but it points to a way of putting low-power, brain-inspired control hardware much closer to cryogenic computing systems. For quantum machines, that could reduce both wiring complexity and thermal load, two barriers to scaling the technology.
This story draws on original reporting from ScienceDaily.