arXiv:2609.19260
Hysteresis in Atomic Josephson Junctions
Vijay Pal Singh, Ludwig Mathey, Luigi Amico
Abstract
Hysteresis and retrapping are hallmarks of underdamped Josephson dynamics, yet they have remained elusive in atomic Josephson junctions. Here, we introduce a velocity-sweep protocol to demonstrate these phenomena in the underdamped regime of an atomic Josephson junction. Specifically, using classical-field simulations of a two-dimensional bosonic superfluid, we show hysteresis in the velocity-imbalance characteristics and determine the retrapping current. The dynamics are quantitatively captured by the resistively and capacitively shunted junction (RCSJ) model. We show that hysteresis originates from persistent phase slips below the critical current, with the associated dissipation mediated by vortex-antivortex pair nucleation at the weak link. The retrapping current follows the Stewart-McCumber scaling of underdamped junctions, establishing a common phase-dynamical framework for atomic and superconducting Josephson junctions.
AI Summary
The paper asks whether a single atomic Josephson junction exhibits underdamped hysteresis and retrapping, and how these phenomena relate to microscopic dissipation. It studies a homogeneous two-dimensional Bose gas using classical-field simulations with thermally sampled initial conditions. A Gaussian barrier separates two reservoirs; smoothly increasing its velocity above the critical velocity and then decreasing it provides a current-sweep protocol. Population imbalance measures the chemical-potential difference. The simulations show distinct switching and retrapping velocities, with wider hysteresis at larger Stewart–McCumber parameter $\beta_C$. Effective junction parameters are extracted independently from constant-velocity simulations and used in deterministic and stochastic resistively and capacitively shunted junction (RCSJ) models. These reproduce the principal hysteresis trends. Phase and density trajectories identify persistent phase slips below the critical current, accompanied by vortex–antivortex pair nucleation at the weak link and propagation into the reservoirs. For simulated $\beta_C$ values extending from approximately $4$ to $59$, the retrapping current approaches the strongly underdamped prediction $I_r/I_c\simeq4/(\pi\sqrt{\beta_C})$. The hysteresis remains sweep-time dependent rather than stationary: its width decays approximately as $T_s^{-\eta}$, with $\eta\sim0.3$–$0.6$ in the atomic simulations versus $0.05$–$0.07$ in the stochastic RCSJ model. This difference is attributed to additional many-body relaxation. Reported retrapping currents are therefore finite-time, quasiequilibrium estimates, with residual sweep-time dependence used for uncertainties. Finite-size boundary reflections also leave residual imbalance, requiring threshold-based retrapping extraction. The results are numerical, not an experimental demonstration.
Error found: The stochastic RCSJ model does not quantitatively reproduce the reported sweep-time relaxation exponents, a limitation explicitly acknowledged by the paper. The retrapping scaling is supported by finite-time, threshold-defined estimates rather than a demonstrated stationary, protocol-independent retrapping curve. The End Matter's claim that longer sweeps lower the retrapping current while leaving the switching current unchanged contradicts the main text's decreasing hysteresis width. The supplied parameters imply E_c/ℏ ≈ 0.156 s^−1, not the printed 0.155 ms^−1. Sweep durations are inconsistently labeled in seconds and milliseconds across captions, leaving the absolute dynamical calibration ambiguous.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 7+2 pages, 4+3 figures