Researchers at the University of Hong Kong have developed a programmable cryogenic neuromorphic platform based on silicon carbide transistors. It simulates the spiking activity of neurons at temperatures as low as 10 mK.
The core of this development is the controlled negative differential resistance in SiC MOSFET transistors. This effect occurs when cooled below 2 K, enabling the creation of energy-efficient neuromorphic circuits, including sensory, logical, and integrating neurons.
The authors view the platform as a means to reduce heat dissipation and wiring volume in cryogenic qubit control systems. Potential applications include quantum error correction, real-time qubit management, and electronics for deep space.