In recent years, cutting-edge research and engineering experiments in quantum computing have become a hot area of focus for major scientific research institutions and technology giants, and the technology has gradually moved from the laboratory to reality. Among them, the use of superconducting circuits to achieve qubits is one of the mainstream implementation schemes of quantum computing.
Recently, the Alibaba Damo Academy Quantum Lab team published a paper on superconducting quantum computing at arXiv[1] and announced the latest results at APS March Meeting 2022. By designing a fluxonium-based quantum processor that integrates the computing power of initialization, readout, and general-purpose quantum logic gates while implementing highly coherent qubits, the team achieved an average single-bit gate fidelity of 99.97% and a two-bit gate fidelity of 99.72%, achieving the best level in the world for such bits. In the field of superconducting quantum computing, this work is the first time that the new bit platform has achieved a quantum operation accuracy close to that of the mainstream transmon bit platform, which poses a powerful challenge to the mainstream solution.
Calculations are performed using quantum behavior in a large number of electrons in superconducting circuits
The qubits in superconducting quantum computing consist of superconducting circuits. In a superconductor, all electrons are in unison during motion, i.e. in a superconducting state in which all electrons have the same single property, and that state can be described by a single quantum state. The quantum states in such circuits can be used to store quantum information, or they can be manipulated to implement quantum logic gates that enable quantum computing.
The existing mainstream superconducting qubits are transmon. Transmon uses the two states of the circuit with/without electron oscillation excitation (plasmon) as the "0" and "1" of the qubits, and its structure is so simple that it only takes one or two Josephson tunnels to penetrate the knot, and is insensitive to electrical noise. Fluxonium works just as its name contains "fluxon" flux quanta, using the flux quanta in a superconducting ring circuit as the "0" and "1" of the bits.
It is also reported that fluxonium uses the magnetic field as a way to store quantum information, which is not only insensitive to electrical noise, but also its sensitivity to the loss of the dielectric is greatly reduced compared to transmon, so it is a qubit that is more resistant to external noise interference. At the same time , fluxonium is a qubit closer to the dichooneric level , and when manipulating between " 0 " and " 1 " at high speeds , it is less likely to transition to other energy levels other than " 0 " and " 1 " , enabling higher precision quantum operations.
Figure | Schematic diagram of fluxonium superconducting quantum chip (Source: arXiv)
High-fidelity gate operation is a breakthrough
Despite the intuitive advantages of fluxonium, combining high coherence with fast operation to achieve high-fidelity logic gate operations, initialization, and data reading remains challenging. To do this, the team needs to solve the following problems: First, to overcome the bottleneck of high two-bit gate error rate, and find the optimal solution in the case of multiple error superposition to perform comprehensive optimization to achieve the best two-bit gate operation. The second is the first implementation of several of the most important basic quantum computing operations in fluxonium through integration work, including bit initialization, bit reading, and the implementation of general-purpose single-bit gate and two-bit gate operations. The solution to these problems lays a good foundation for continuing to expand the platform and implement more complex quantum computing.
Because this is a systematic effort, the researchers have made a lot of optimizations in the design and preparation of the chip, the implementation and calibration of the operation. In terms of design, starting from the most basic circuit noise model, the parameter space of the chip was optimized as a whole, balancing the effects of different noises and errors on accuracy; in terms of preparation, they developed a new process to uniformly prepare a large number of Josephson junctions on the same wafer, integrating the yield of a large number of Josephson junctions (each fluxonium bit is composed of more than 100 Josephson structure); in the measurement work, they implemented a new initialization scheme, solved Fluxonium itself cannot passively initialize the problem.
At the same time, they implemented a two-bit gate called iSWAP, that is, a virtual swap gate, in fluxonium superconducting qubits for the first time, and made good use of the characteristics of fluxonium-like binary systems to implement gate operations, making it exceed all previous fluxonium operation accuracy. In the verification and compilation of quantum gates, the team subsequently implemented SQiSW (half the time consumption is lower than iSWAP) on the basis of iSWAP gates, and proposed a compilation algorithm that verifies the accuracy and efficiency of the representation of SQiSW gates to compile quantum algorithms into SQiSW gate operations.
Figure | Stochastic benchmark for single qubit fidelity (source: arXiv)
It is reported that the ratio of gate operation time to bit coherence time is the decisive factor in the fidelity of the quantum gate. The bit coherence time refers to the average amount of time that quantum information can be preserved in qubits. This means that the more door operations that can be completed in coherent time, the higher the fidelity of the door in general. Although the coherence time of the bits in this work is not the longest in the fluxonium type bits, the research team achieved a high-fidelity two-bit gate that surpassed the previous high-fidelity by speeding up the operation speed of the door. At present, the DHARMA academy team is working hard to improve the relevant time, and has the confidence to further improve the fidelity of the door operation in the future.
Figure | Two-Qubit Gate Architecture Scheme and Benchmark Results (Source: arXiv)
Quantum computing requires the combination of science and engineering
For the future application potential of quantum computing, Deng Chunqing, a scientist at the Quantum Laboratory of Dharma Academy, summed up the following: Quantum technology can solve some simulations related to quantum itself, such as the simulation of drug synthesis and the simulation of material synthesis. These simulations can help people in related fields better explore new materials and drugs. At present, these tasks are difficult to do on supercomputers, and on quantum computers, such problems can be solved more efficiently due to the quantum properties of the computed object itself. He said that quantum computing can solve these problems, and the main reason behind it is the exponential increase in computing power.
In addition, the team is the only industrial team that makes fluxonium a qubit type, and the reason for choosing fluxonium is based on their thinking about the physical aspects behind the contrast. It is gratifying that after several years of accumulation, they can finally reflect the above advantages.
According to reports, Deng Chunqing is one of the main scientists since the establishment of the Quantum Laboratory, who graduated from the School of Information Science and Technology of Peking University with a bachelor's degree and graduated from the Institute of Quantum Computing at the University of Waterloo, Canada. Talking about his work at the Dharma Institute's Quantum Laboratory, he said that doing research in industry has both purely scientific exploration and engineering projects. Moreover, quantum computing is at an important point in time, and many scientific problems are being transformed into engineering problems, which can take advantage of the interdisciplinary and professional advantages of industry teams.
Talking about follow-up plans, the team said that the paper is just the beginning, and its goal is to achieve fault-tolerant quantum computing systems, that is, quantum computing systems that are both high-precision, multi-bit and compatible with quantum error correction. While this work has demonstrated excellent logic gate accuracy well, it is a long process to achieve quantum error correction in order for all errors to be suppressed. As a result, they will continue to improve gate operation accuracy and number of bits, and will integrate more system operation techniques to enable fault-tolerant quantum computing.
-End-
reference:
1、Bao F, Deng H, Ding D, et al. Fluxonium: an alternative qubit platform for high-fidelity operations.arXivpreprintarXiv:2111.13504 (2021).