Theoretical Proposal Turns Josephson Junction Tunneling into a Memory Probe

A quantum tunneling event is usually treated as a strictly local affair: a particle either escapes a barrier right now, or it does not. A new theoretical study argues that in superconducting circuits, the environment can leave a fingerprint that spans time itself. The work, published in the journal Discover Quantum Science, proposes that bath-mediated correlations in a current-biased Josephson junction could carry information from before a suppressed tunneling window to the switching statistics measured after it, producing signatures that no purely local-in-time rate model can mimic.

The paper, authored by Mahgoub A. Salih and Mohamed Y. Shirgawi, is purely theoretical. No experimental data were generated or analyzed; the authors state this explicitly in their data availability statement. What the study offers instead is a controlled effective theory, a set of falsifiable predictions, and a concrete experimental protocol. The central object is a cross-time memory kernel that couples pre-window dynamics to post-window switching, with an amplitude that decays exponentially with the length of the suppressed interval and depends on a history-weighted functional of what happened before the window opened.

The physical setting is macroscopic quantum tunneling in a Josephson junction. When a superconducting junction is biased below its critical current, the phase difference across it behaves like a quantum particle trapped in a metastable well of the tilted washboard potential. Escape from that well can occur through quantum tunneling, with a rate given semiclassically by the barrier height and the plasma frequency. Both quantities depend on the bias current, which modern superconducting circuits can modulate on nanosecond timescales. That controllability is what makes the junction an attractive platform for the proposed test.

The protocol imagines lowering the bias current during a finite interval, from time t1 to time t2. During this window the instantaneous escape rate is strongly suppressed, by a factor epsilon far below unity. In the conventional Markovian picture, tunneling at any moment depends only on the instantaneous conditions, and the survival probability satisfies a simple rate equation. When the full rate is restored at t2, the switching-time distribution shows a sharp clustering peak: the release of population that accumulated during the suppressed interval. The authors are careful to stress that this peak is not itself a non-Markovian effect. Any local rate gate produces it.

The distinctive prediction lies elsewhere. Starting from the Caldeira-Leggett influence functional, which describes how an environmental bath affects the quantum system, the authors identify the approximations under which bath-mediated correlations between tunneling events, or instantons, can be coarse-grained into a Volterra equation for the survival probability. Under conditions of dilute instantons, weak inter-bounce coupling, irreversible switching to the voltage state, and fast intrawell relaxation, the resulting equation contains a memory term. Its cross-window part vanishes unless the earlier time lies before t1 and the later time lies after t2, isolating exactly the bath-mediated connection between pre-window and post-window dynamics.

The amplitude of this cross-time channel carries a factor that decays exponentially with the barrier duration, scaled by the inverse memory time of the bath. That factor doubles as a spectroscopic handle: measuring how the tail amplitude falls as the window lengthens would directly probe the bath correlation time. The switching distribution then contains both the Markovian clustering peak at t2 and a non-Markovian post-window tail, whose amplitude depends on a pre-window memory functional built from the survival probability before the barrier. The coexistence of the peak with a history-dependent tail, the authors argue, constitutes a definitive experimental discriminant against purely local rate-gating, because no local model with a fixed post-window rate can reproduce the protocol dependence.

The authors are equally explicit about what the theory is not. The kernel is described as a controlled phenomenological reduction, not an exact microscopic identity. The derivation controls the two-time correlation structure of the bath, while the final survival equation is a coarse-grained population model valid only in its stated regime. Positivity of the switching distribution is guaranteed for monotonic non-negative kernels, but the experimentally attractive resonator case produces an oscillatory, sign-changing kernel, so admissibility must be checked numerically: the survival probability must remain between zero and one, the switching density must stay non-negative, and it must integrate to unity over the fitted window. Where those tests fail, the full real-time path integral or a master equation treatment is required. The authors also emphasize that no causality violation is involved; the pre-window dynamics excite bath degrees of freedom that evolve forward in time and later modify escape statistics. They call the phenomenon temporal tunneling, but note that the mechanism differs from spatial tunneling, and all dynamics respect causal ordering.

Feasibility analysis separates viable platforms from hopeless ones. In standard resistively shunted junctions with Drude-cutoff environments, the memory time is on the order of ten picoseconds. For nanosecond-scale barriers, the cross-time amplitude is then suppressed by roughly e to the minus one hundred, effectively zero. The proposed observable regime instead uses a structured electromagnetic environment: a high-Q microwave resonator with a frequency near five gigahertz and a quality factor around one hundred thousand. Such a resonator has a damping time of roughly six microseconds, extending the effective memory into a range where both the peak-plus-tail signature and its scaling laws become experimentally accessible.

Temperature and artifacts impose further constraints. To ensure switching is genuinely quantum rather than thermally activated, the device must operate below the thermal-to-quantum crossover temperature, which for plasma frequencies of five to ten gigahertz lies around forty to eighty millikelvin, well within reach of a dilution refrigerator. The MQT regime should be verified by the saturation of escape rates as temperature is lowered. The proposed falsification protocol has four steps: calibrate the junction and detector, confirm the Markovian clustering ratio after accounting for finite bin width and pulse rise time, repeat the measurement for several barrier durations while holding the post-window waveform fixed, and vary only the pre-window preparation to test the predicted history dependence. Thermal activation, pulse ringing, detector afterpulsing, and heating each receive dedicated controls.

Using illustrative parameters, the authors estimate a cross-time signal roughly nine percent of the local post-window peak, large enough to motivate an experimental search but small enough that systematic controls are essential. The decisive tests are the exponential scaling of the tail amplitude with barrier duration and its dependence on the pre-window protocol at fixed post-window dynamics. The authors note that existing switching measurements in engineered environments have not, to their knowledge, reported this specific pair of tests, and that published null observations would constrain the effective signal strength rather than falsify non-Markovian Josephson dynamics in general. Within its stated limits, the work offers a concrete route to a kind of time-domain spectroscopy of environmental memory in superconducting circuits.

Subject of Research: A theoretical model of non-Markovian cross-time memory effects in macroscopic quantum tunneling of Josephson junctions

Article Title: Cross-time memory kernels in macroscopic quantum tunneling non-Markovian signatures in Josephson junctions

Article References: Salih, M. A., & Shirgawi, M. Y. (2026). Cross-time memory kernels in macroscopic quantum tunneling non-Markovian signatures in Josephson junctions. Discover Quantum Science, 2(1), Article 25. https://doi.org/10.1007/s44464-026-00029-5

Image Credits: AI Generated

DOI: 10.1007/s44464-026-00029-5

Keywords: Josephson junctions, macroscopic quantum tunneling, non-Markovian dynamics, memory kernels, superconducting circuits, Caldeira-Leggett model, switching statistics, temporal barriers, high-Q resonators, quantum escape, open quantum systems, theoretical physics