We extend the scattering theory of the Josephson effect to include a coupling of the Josephson junction to a gapless electron reservoir in the normal state. By opening up the system with a quasiparticle escape rate , the supercurrent carried at zero temperature by an Andreev level at energy is reduced by a factor . We make contact with recent work on “non-Hermitian Josephson junctions,” by comparing this result to different proposed generalizations of the Josephson effect to non-Hermitian Hamiltonians.
The thermodynamic properties of a quantum dot are affected by the coupling to a superconductor. The Josephson effect is a striking example, and a current will flow through the quantum dot in equilibrium if the quantum dot forms a weak link between two superconductors.1 This supercurrent depends periodically on the phase difference of the pair potential in the superconductors. A dissipative electromagnetic environment may degrade the supercurrent via phase fluctuations.2–4 Here, we investigate an altogether different decay mechanism, the coupling of the quantum dot to a gapless electron reservoir (see Fig. 1).
Josephson junction, formed by a quantum dot that is tunnel coupled with rates and to two superconductors (gap and phase difference ). The quantum dot has an additional weak coupling (rate ) to a gapless electron reservoir in the normal state. We seek the -dependence of the current–phase relationship .
Josephson junction, formed by a quantum dot that is tunnel coupled with rates and to two superconductors (gap and phase difference ). The quantum dot has an additional weak coupling (rate ) to a gapless electron reservoir in the normal state. We seek the -dependence of the current–phase relationship .
The reservoir is a source of dephasing because quasiparticles that enter it from the quantum dot return without any phase coherence. Such a dephasing mechanism for persistent currents was introduced by Büttiker5 and applied to superconducting circuits by several authors.6–13 Our main advance is that we obtain closed-form expressions for the supercurrent–phase relationship, for the case of a spatially uniform coupling rate to the electron reservoir.
Because the effective Hamiltonian of the quantum dot becomes non-Hermitian when we open it up to the reservoir, such a system is a simple example of a “non-Hermitian Josephson junction,” a topic of current interest.14–17 Our explicit expressions support Ref. 17 in the debate over the proper generalization of the current–phase relationship to complex eigenvalues E of the effective Hamiltonian.
We open up the Josephson junction by weakly coupling it to a gapless electron reservoir in the normal state. Quasiparticles enter the reservoir at a rate , assumed to be spatially uniform in the junction region. This can be modeled by coupling to the reservoir via a spatially extended tunnel barrier.20,21 The reservoir is closed, and it does not drain any current, but particles that enter it return to the junction without any phase coherence, so they no longer contribute to the supercurrent.
The full scattering matrix of the normal region, including the scattering channels to the electron reservoir, is unitary, with a sub-unitary submatrix for the scattering channels to the superconductors. Similarly, the Andreev reflection matrix is now sub-unitary also for .
As a first application, we consider the case that the two superconductors are coupled via a quantum dot containing a single resonant level, at energy relative to the Fermi level. We ignore Coulomb blockade effects, which will be less significant in the open system.
Zero-temperature current–phase relation of the quantum dot Josephson junction of Fig. 1, with and without the coupling to the electron reservoir. The plot is computed from Eq. (18) for the case of unit transmission probability at the Fermi level through the quantum dot.
The second application is a point contact junction of length L short compared to the superconducting coherence length . In this short-junction regime, the energy dependence of can be neglected relative to the energy dependence of , and we may evaluate at the Fermi level. We assume that time reversal symmetry is preserved; hence, .
For the third and final application, we consider a superconductor–normal-metal–superconductor (SNS) junction of length . Both states above and below then contribute to the supercurrent.
The same as Fig. 2, but now for a single-mode ballistic SNS junction in the long junction limit (length L large compared to the superconducting coherence length ). The plot is computed from Eq. (29).
The coupling of the Josephson junction to the electron reservoir pushes the poles of its scattering matrix into the lower half of the complex plane, down to . These poles can be considered as the complex eigenvalues of an effective non-Hermitian Hamiltonian . The open Josephson junction thus provides a physical realization of the non-Hermitian Josephson effect studied in Refs. 14–17. Different generalizations have been proposed for the relation when E is complex. Let us compare these with our findings.
In summary, we have calculated how the coupling to a gapless electron reservoir in the normal state reduces the supercurrent through a Josephson junction. A simple answer is obtained for the model where the escape rate of quasiparticles into the reservoir is spatially uniform. At zero temperature, the reduction factor for a given Andreev level equals . This applies to a weakly coupled quantum dot Josephson junction, or to a short point contact (length L smaller than the coherence length ). A more complicated -dependence is obtained in a long junction ( ), when also states above the gap contribute to the supercurrent.
I have benefited from discussions with A. R. Akhmerov, Yu. V. Nazarov, P.-X. Shen, and T. Vakhtel. This project has received funding from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme (Advanced Grant No. 832256).
AUTHOR DECLARATIONS
Conflict of Interest
The author has no conflicts to disclose.
Author Contributions
C. W. J. Beenakker: Conceptualization (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Writing – original draft (equal); Writing – review & editing (equal).
DATA AVAILABILITY
Data sharing is not applicable to this article as no new data were created or analyzed in this study.