In recent years, superconducting quantum computing has developed rapidly with the potential to scale to thousands of qubits in the next few years and explore practical applications of Noisy Intermediate-Scale Quantum (NISQ) devices based on this. However, superconducting qubits have large sizes, and each qubit requires dedicated RF-control lines, making it increasingly challenging to integrate more qubits on a single chip as the number of qubits increases. Modularity offers a viable approach for scaling up to large numbers of qubits in the near term; however, this relies on very high-performance interconnects between the computing modules.

Recently, researchers resolved this bottleneck and realized ultralow-loss interconnects with quality factors as high as 8.1 x 10⁵, a tenfold improvement over previous experiments, comparable to the native coherence of transmon qubits. They improved the inter-module quantum state transfer fidelities up to single-chip level (99%), entangling up to 12 qubits across three modules in a Greenberger-Horne-Zeilinger (GHZ) state. The results demonstrate a viable modular approach to building large-scale superconducting quantum processors.

Chair Professor Dapeng Yu’s research team from the Shenzhen Institute for Quantum Science and Engineering (SIQSE) at the Southern University of Science and Technology (SUSTech) has made new progress in ultralow-loss interconnects for scalable distributed quantum networks.

Their research findings, entitled “Low-loss interconnects for modular superconducting quantum processors,” have been published in the journal Nature Electronics.  

After nearly two years of technological innovation, Prof. Yu’s research team has developed a superconducting coaxial cable with ultralow-loss and easy bonding connection, and integrated impedance converters on quantum chips to reduce the loss at the connection interfaces. Through a series of technological innovations, high-performance superconducting quantum chip interconnection has been achieved, with a channel single-photon quality factor of 8.1 x 10⁵ and a coherence time equivalent to that of a transmon qubit on a single chip. Based on this, the fidelity of inter-module quantum state transfer has reached 99%, leading the international community in this field and laying the foundation for the scale-up of superconducting quantum processors.

Figure 1. Modular quantum processor design

Using these ultralow-loss interconnects, the research team linked five quantum modules together, each of which has four qubits, forming a 20-qubit modular quantum processor. Based on this modular processor, they demonstrated the generation of multi-qubit inter-module GHZ states, achieving a 4-qubit GHZ state distributed across two chips with a fidelity of 92%, comparable to the performance of similar multi-qubit entangled states on a single chip. This is the first time that a modular superconducting quantum processor has achieved single-chip performance in the preparation of multi-qubit entangled states, which is a milestone achievement. Through multiple inter-module quantum state transfers and single-chip logical gate operations, the researchers eventually entangled up to 12 qubits in a GHZ state with 55.8% fidelity. These results demonstrate a viable modular approach to building large-scale superconducting quantum processors.

Figure 2. High fidelity inter-module quantum state transfer and multi-qubit entanglement generation

Assistant Researcher Jingjing Niu is the first author of this paper. Profs. Song Liu and Youpeng Zhong are the corresponding authors, and Chair Prof. Dapeng Yu is the last author.

The research work was supported by the National Natural Science Foundation of China (NSFC), Department of Science and Technology of Guangdong Province, Science, Technology and Innovation Commission of Shenzhen Municipality, and SUSTech.

Paper link: https://www.nature.com/articles/s41928-023-00925-z