Researchers Develop Brain-Inspired Chip Operating Near Absolute Zero for Quantum Computing
Key Takeaways
- ▸HKU researchers created neuromorphic chips operating at 10mK using SiC MOSFETs, achieving thousands of times greater energy efficiency than conventional cryogenic electronics—potentially eliminating the thermal burden that constrains large-scale quantum computer architecture
- ▸A robust negative differential resistance effect in SiC discovered at cryogenic temperatures enables scalable manufacturing using existing automotive and power grid foundries, dramatically reducing commercialization barriers compared to exotic-material approaches
- ▸Integrated cryogenic control electronics could be placed directly alongside quantum processors, eliminating extensive wiring requirements and reducing heat dissipation challenges that currently limit qubit scale and fidelity
Summary
Researchers at the University of Hong Kong have developed a groundbreaking cryogenic neuromorphic hardware platform capable of operating at temperatures near absolute zero (10 millikelvin). Led by Professor Yuhao Zhang and PhD student Xin Yang, the team created programmable circuits using Silicon Carbide (SiC) MOSFETs that can reproduce the energy-efficient 'spiking' behavior of biological neurons while functioning in extreme cold—a feat never before demonstrated.
The breakthrough directly addresses a critical bottleneck in quantum computing: existing control electronics consume significant power and generate unwanted heat, requiring physical separation from sensitive qubits through extensive wiring that hinders performance and scalability. The new approach exploits a previously unknown strong 'S-shaped' negative differential resistance (NDR) effect in SiC MOSFETs at cryogenic temperatures, driven by electron-donor impact ionization. Because this effect arises from the material's atomic properties rather than device engineering, it is inherently robust and reproducible across manufacturing batches.
The researchers demonstrated that these artificial neurons can be cascaded into larger networks, enabling advanced local data processing at cryogenic temperatures and improving key quantum functions like error correction and real-time quantum control. Because Silicon Carbide is already widely deployed in electric vehicles and power grids, the team can leverage existing global semiconductor foundries to manufacture these cryogenic chips on standard 300-mm wafers. Beyond quantum computing, the technology could support deep space exploration systems designed to operate reliably in the harsh conditions of the Moon's surface or distant regions of the solar system.
- The technology has applications extending beyond quantum computing to deep space exploration systems requiring reliable operation at extreme temperatures
Editorial Opinion
This research represents a significant engineering breakthrough for practical quantum computing by tackling one of the field's most stubborn challenges: how to control sensitive qubits without introducing thermal noise into cryogenic systems. The use of Silicon Carbide—a mature semiconductor with established industrial manufacturing—makes this solution far more commercially viable than competing approaches reliant on exotic materials or custom fabrication. If the HKU team can successfully demonstrate cascaded neuromorphic networks in real quantum control applications, this work could substantially accelerate the path to fault-tolerant quantum computers at scale.
