Reaching New Heights: Achieving Record-Breaking Single- and Two-Qubit Gates with Fluxonium Qubits
Publication in Physical Review X pushes the fluxonium qubit into the limelight.
Authored by Leon Ding, Atlantic Quantum Co-Founder & Head of Calibration.
Recent research performed by Leon Ding, Max Hays, et. al. demonstrates a new qubit coupling architecture that reduces errors in quantum computing operations (Physical Review X). This architecture uses the fluxonium superconducting qubit, which differs from the more conventional transmon in construction, operation, and underlying physics.
A Time for Change
One of the most promising approaches to building a quantum computer is to use superconducting qubits, which are microscopic circuits printed on silicon chips. At a low enough temperature, the circuits become superconducting and exhibit quantum properties. In contrast to other quantum systems such as individual atoms trapped in electromagnetic fields, the properties of superconducting qubits can be engineered by changing the circuit layout. Despite the tremendous design flexibility available to this quantum computing platform, the community has so far relied on the simplest possible circuit to build qubits, commonly known as the transmon qubit.
Since the invention of the transmon in 2007, each new advancement toward building a superconducting quantum computer at scale has been made with the same circuit. These advancements include those in single-qubit gate accuracies, two-qubit gate accuracies, and the number of connected qubits. Notably, major players such as Google Quantum AI, IBM Quantum, Rigetti, and IQM, who are actively engaged in building quantum computers with superconducting qubits, have chosen the transmon as their preferred qubit. Meanwhile, other superconducting qubit design variants have been pigeon-holed into a more academic role. They have offered interesting characteristics to study, but none more promising or easier to scale up than the transmon qubit. In many cases, these alternative qubits theoretically support higher quantum coherence times, but create other serious challenges in fabrication, layout, and control.
In many respects, the fluxonium qubit was the poster-child of this dilemma. It makes better use of the large design space of superconducting circuits by adding an additional elements to the circuit compared to the transmon. However, all demonstrations showing improved performance of individual qubits have come with additional complications with no clear path forward to retain an advantage in a large system of many qubits — thus far.
A New Take on Superconducting Quantum Computing
At Atlantic Quantum, we find the fluxonium qubit to be the ideal balance between circuit complexity and performance, with substantial evidence emerging recently. In 2019, it was first demonstrated that fluxonium could enhance individual qubit performance, with rapid advancements following. Our recent study in Physical Review X, co-authored by Atlantic Quantum co-founders, reveals fluxonium’s superior performance in multi-qubit devices compared to even the best transmon-based quantum computers. This progress in gate fidelities results from a blend of circuit design, physics, and machine learning. Unlike many superconducting qubit structures where one qubit’s operation can interfere with others, our research outlines a viable scaling route for large quantum processors. This is achieved by allowing flexible qubit operation, which prevents frequency space collisions between neighboring qubits. Importantly, our findings unveil a unique synergy: the circuit facilitates both improved qubit operations and better scalability.
In fact, there are many aspects of the fluxonium that make the qubit control even easier than with a transmon. The lower qubit frequency allows for cheaper control electronics and makes it easier to synthesize precise control signals. Furthermore, the fluxonium qubit allows for faster single-qubit gates without incurring unwanted control errors (due to the order-of-magnitude higher anharmonicity for any physicists reading). These errors are similar in nature to the jaggedness experienced when swinging on a swing too high, which can be remedied by a larger swing.
Now, the fluxonium qubit has decidedly surpassed the incumbent transmon in nearly every measure. We believe it heralds a new era of scalable, high-quality quantum computers, and we are committed to realizing this potential.
About the author
Leon Ding is the Co-Founder & Head of Calibration at Atlantic Quantum, where he leads the characterization and calibration of quantum processors. During his PhD at MIT, Leon pioneered the tunable coupling architecture for fluxonium qubits. This innovative approach not only set a new world record for two-qubit gate fidelity in superconducting quantum computers — previously held by fellow Atlantic Quantum co-founder Youngkyu Sung — but also achieved the lowest error rates across any quantum computing hardware modality. Leon holds a PhD in Electrical Engineering and Computer Science from MIT and a BS in Physics from the California Institute of Technology (Caltech). Learn more about Atlantic Quantum at: http://www.atlantic-quantum.com.
