Figure

Description

The phase qubit is a superconducting qubit based on a current-biased Josephson junction operating in the phase regime (). The qubit states are the two lowest energy levels in one metastable well of the tilted-washboard potential, which becomes approximately cubic when the bias approaches the junction critical current.

When biased near the critical current (), the washboard potential has shallow metastable wells containing only a few quantized levels. The two lowest levels serve as and , with transition frequency tunable by adjusting . The cubic asymmetry makes higher levels more closely spaced and gives excited states exponentially larger escape rates.

Readout exploits this escape-rate contrast. A fast bias pulse lowers the barrier so that the excited-state population escapes with much greater probability than ; the switched state is then amplified into a classical circulating-current or voltage signal. The original 2002 device first drove to enhance escape, whereas later capacitively shunted devices used direct state-selective tunneling.

The phase qubit was historically important—the Martinis group used it extensively from 2002 through the early 2010s—but it was superseded in quantum processors by transmon-family circuits with much longer coherence and nondestructive dispersive readout. Current-biased Josephson junctions remain active as macroscopic quantum systems and threshold detectors, not as a competitive processor-qubit architecture.

Hamiltonian

For a current-biased junction, with , , and , let denote the total capacitance seen by the junction phase mode, including the junction and intentional shunt capacitances,

Writing , the local minimum is at . For , expansion about that minimum gives

with plasma frequency

The exact barrier from the local minimum to the adjacent maximum is

where the approximation holds for . In the cubic-well limit the ground-state escape rate has the leading WKB dependence ; excited levels escape much faster.

Motivation

The phase qubit provided early demonstrations of quantum coherence and entanglement in superconducting circuits. Its straightforward readout mechanism (tunneling → voltage) was simpler than dispersive readout, making it an important stepping stone. However, its sensitivity to current-bias noise and the destructive nature of the tunneling measurement motivated the transition to transmon-based architectures.

Experimental Status

First coherent phase qubit — Martinis et al. (2002):

  • Rabi oscillations observed in a large-area current-biased Josephson junction at GHz
  • Spectroscopic quality factor and an inferred coherence time of about 10 ns in the reported device
  • State-selective tunneling readout with 85% excited-state fidelity and greater than 99% ground-state preparation/readout fidelity

Capacitively shunted redesign — Steffen et al. (2006):

  • Separated the shunt capacitor from the tunnel junction, reducing strongly coupled two-level defects by roughly an order of magnitude
  • Demonstrated single-shot state tomography with about 90% measurement fidelity

Two-qubit entanglement — Steffen et al. (2006):

  • Demonstrated entanglement between two capacitively coupled phase qubits
  • Reconstructed a Bell state with 87% fidelity after correcting measurement error
  • Reported ns and ns for the coupled devices

Three-qubit entanglement — Neeley et al. (2010):

  • Generated both three-qubit W and GHZ states
  • State tomography gave and , both above their respective genuine-three-party-entanglement witness thresholds

Materials endpoint — Patel et al. (2013):

  • A phase qubit with a single-crystal silicon shunt capacitor reached at 5.093 GHz
  • The same device had ns because the architecture has no flux sweet spot

Recent status — 2024–2026:

  • No phase-qubit processor result or coherence/gate record superseding the early-2010s endpoint was found
  • Ouyang et al. (2024) experimentally revisited the bias-controlled crossover of a current-biased junction from a nearly harmonic oscillator to a nonlinear two-level artificial atom
  • Badarne et al. (2025) theoretically proposed a memory-integrated phase qubit based on a superconducting-ferroelectric Josephson junction whose history-dependent barrier modulates the critical current; it is a device concept, not an experimental phase-qubit revival or processor benchmark
  • Chai et al. (2025, 2026) developed current-biased Josephson-junction threshold-detector protocols from few-photon sensitivity toward simulated single-photon sensitivity using fast nonadiabatic bias sweeps; these are detector results, not a revival of the phase-qubit processor architecture

Key Metrics

MetricValueNotesFidelity reference
1.6 μsSingle-crystal-Si shunt; best value in the cited devicePatel et al. 2013
110 nsNo flux sweet spot; low-frequency flux noise dominatedPatel et al. 2013
AnharmonicityRepresentative shallow well with Martinis et al. 2002
Transition frequency5.093 GHzOperating point of the 2013 crystalline-Si devicePatel et al. 2013
Readout fidelity85% (2002); about 90% (2006)Destructive, tunneling-based single-shot readoutMartinis et al. 2002; Steffen et al. 2006
Two-qubit Bell-state fidelity87%Tomographic fidelity after measurement-error correctionSteffen et al. 2006
Three-qubit state fidelity; Both exceeded the relevant entanglement-witness thresholdNeeley et al. 2010
Operating temperature25 mK (2002); 35 mK (2013)Dilution refrigeratorMartinis et al. 2002; Patel et al. 2013

References

Original proposal / demonstration

Experimental demonstrations

Recent current-biased-junction context

Reviews

Linked Papers

Evergreen context

  • josephson-junction-as-nonlinear-element — the tilted cosine potential and plasma-mode anharmonicity that the current-biased junction turns into a qubit
  • coherence-time-hierarchy — how to interpret the short and numbers that ultimately made the phase qubit a transitional architecture
  • dispersive-readout-mechanism — the non-destructive resonator readout paradigm that replaced tunneling-based measurement in mainstream superconducting processors