Other meanings of Superconducting quantum computing
Quantum computing
Superconducting quantum computing is a leading approach to building a quantum computer that uses superconducting electronic circuits to form qubits, the fundamental units of quantum information. These circuits are operated at cryogenic temperatures (typically below 20 mK) to exploit quantum effects such as superposition and entanglement. The most widely used qubit design is the transmon, a form of charge qubit that employs a Josephson junction — a thin insulating barrier between two superconductors — to introduce an anharmonic energy spectrum, enabling selective addressing of the two lowest energy levels. The field has progressed rapidly since the first demonstrations of superconducting qubits in the late 1990s, achieving milestones such as quantum supremacy (2019) and the realization of logical qubits with error correction. Major technology companies including Google, IBM, and Intel, along with academic institutions, are actively developing superconducting quantum processors with increasing numbers of qubits and improved coherence times.
Superconducting qubits are macroscopic electrical circuits that exhibit quantum behavior because the superconducting state allows currents to flow without resistance, enabling coherent quantum superpositions of charge, flux, or phase degrees of freedom. The essential nonlinear element is the Josephson junction, which replaces a linear inductor in an LC resonator and provides an anharmonic energy-level ladder. The junction’s Josephson inductance combined with the capacitance of the circuit forms a weakly anharmonic oscillator whose lowest two energy levels can be used as the qubit basis states |0⟩ and |1⟩. The most prevalent variant, the transmon qubit, is designed to be insensitive to charge noise by operating in a regime where the Josephson energy dominates over the charging energy. Other designs include flux qubits, which encode information in persistent current states, and phase qubits, which use the junction’s phase difference. All superconducting qubits require cryogenic environments — typically dilution refrigerators — to suppress thermal fluctuations below the energy splitting between qubit states, which is in the microwave frequency range (roughly 4–8 GHz).
Current superconducting quantum processors arrange qubits in a two-dimensional lattice on a silicon chip, with each qubit coupled to its nearest neighbors via fixed capacitive or inductive links. Readout is performed using microwave resonators coupled to each qubit, and control is achieved through shaped microwave pulses sent down coaxial lines to the chip. The dominant error-correction strategy is the surface code, which requires a two-dimensional array of physical qubits with high-fidelity gates and a classical decoder to diagnose and correct errors. Scalability challenges include the need for a large number of control wires (currently one per qubit), the limited cooling power of dilution refrigerators (typically a few hundred microwatts at 20 mK), and the requirement that all qubits have coherence times long enough to perform error-correction cycles. Advances in multiplexed readout, cryogenic control electronics, and three-dimensional integration (e.g., using through-silicon vias) are being pursued to address these bottlenecks. The largest processors as of 2025 contain over 1,000 physical qubits, though the number of logical qubits remains in the single digits.
The first experimental demonstration of a superconducting qubit was reported in 1999 by a group at NEC and the University of Tokyo, using a charge qubit design. A key early figure was Michel Devoret, who pioneered the theoretical description of the Josephson junction as a qubit. In 2007, the transmon qubit was introduced by Jens Koch and colleagues at Yale, dramatically improving coherence times. John Martinis’s group at UC Santa Barbara achieved the first demonstration of a two-qubit gate with fidelity above 99% in 2014. In 2019, Google’s Sycamore processor solved a random circuit sampling problem in 200 seconds that would take a classical supercomputer thousands of years, claiming quantum supremacy. IBM has steadily increased its quantum volume, releasing the 127-qubit Eagle processor in 2021 and the 1,121-qubit Condor chip in 2023. A lesser-known but important contribution came from the work of Jaw-Shen Tsai and colleagues, who demonstrated the first coherent oscillations in a superconducting qubit in 2000. The field has also been shaped by theoretical advances in optimal control and noise modeling, such as the development of the randomized benchmarking protocol.
Beyond the mainstream narrative, several niche details are critical to the field’s progress. Quasiparticle poisoning — the tunneling of broken Cooper pairs across the Josephson junction — is a major source of relaxation and dephasing, and its mitigation has led to the use of gap-engineering, such as the incorporation of aluminum oxide layers. Material defects in the substrate and interface, particularly amorphous oxides, create two-level-system (TLS) fluctuators that absorb microwave energy and cause decoherence. Researchers have found that reducing the volume of TLS-hosting materials, using tantalum instead of aluminum for the junction’s superconductor, can improve coherence times. Another obscure but important point is the role of the Purcell effect: qubits coupled to readout resonators can experience spontaneous emission into the resonator line, limiting the qubit lifetime; this is suppressed by using Purcell filters. The first theoretical proposal for a superconducting qubit, by John Clarke and others in 1988, was based on the current-biased Josephson junction and was considered a “macroscopic quantum coherence” experiment. Even today, the exact mechanisms of decoherence in these circuits remain an active area of research, with noise spectroscopy techniques revealing 1/f noise from charge and flux fluctuations that still lack a complete microscopic explanation.
This article covers the specific sense of superconducting quantum computing that uses superconducting circuits as qubits, distinct from other quantum computing modalities such as trapped ions or photonic quantum computing.
Help improve the encyclopedia. Reports go straight to the site manager.