Researchers from the University of Minnesota Twin Cities have developed a new superconducting diode that can enhance the performance of artificial intelligence (AI) systems and scale up quantum computers for industrial applications. This device surpasses existing diodes in terms of energy efficiency, simultaneous processing of multiple electrical signals, and the integration of gate control for energy flow. Diodes are crucial components that allow current to flow in one direction in an electrical circuit, and the researchers’ superconducting diode offers advantages over traditional semiconducting diodes.

The team created the diode using three Josephson junctions, which consist of superconducting material sandwiched between non-superconducting layers. By connecting the superconductors with semiconductor layers, they achieved a unique design allowing voltage control of the device’s behavior. Unlike regular diodes that handle one input and output, this superconducting diode can process multiple signal inputs, making it potentially useful for neuromorphic computing, a method that mimics brain neuron functions to improve AI performance.

The researchers’ device boasts high energy efficiency and the ability to add gates and apply electric fields to tune its effects. Moreover, it utilizes industry-friendly materials, making it compatible for wider industrial applications. The scalability of quantum computers is essential for tackling complex real-world problems, and this superconducting diode contributes to the development of hardware necessary for quantum computers to implement advanced algorithms. The researchers believe that their work demonstrates the power of universities in generating ideas that eventually find practical integration in the industry.

Reference:

Gupta, M., Graziano, G. V., Pendharkar, M., Dong, J. T., Dempsey, C. P., Palmstrøm, C., & Pribiag, V. S. (2023). Gate-tunable superconducting diode effect in a three-terminal Josephson device. Nature Communications.