Observing dissipationless flow of an impurity in a strongly repulsive quantum fluid

TL;DR

The study investigates dissipationless transport of a microscopic quantum impurity in a strongly interacting 1D Bose gas. By injecting a finite-momentum impurity into an array of 1D tubes and quenching to strong repulsion, the authors observe rapid relaxation to a stationary state with a finite residual velocity, accompanied by shock waves for supersonic initial motion. The long-time impurity dynamics are well described by a moving polaron picture, supported by matrix-product-state simulations that reproduce the relaxation times and final momentum. This work demonstrates dissipationless impurity propagation in a 1D quantum fluid and provides insights into polaron dynamics and quantum transport in strongly correlated systems.

Abstract

The frictionless motion of an object through a fluid medium is commonly viewed as a hallmark of superfluidity. According to Landau, kinematic constraints prohibit superfluid behavior in one-dimensional (1D) bosonic systems. Here, using ultracold atoms, we show how a microscopic impurity can propagate through a strongly interacting 1D Bose gas without any friction, at odds with conventional expectations. We inject the impurity with initial velocities ranging from the subsonic to supersonic regime, and subsequently track its dynamics. For supersonic initial velocities, we observe the formation of a shock wave and a remarkably fast relaxation to a stationary regime, on a time scale that increases with decreasing impurity velocity. After reaching the stationary state, the impurity continues its motion through the system with a finite velocity. Our findings demonstrate how quantum effects can conspire to eliminate dissipation of a microscopic object immersed in a quantum fluid, thereby bringing novel insights into the propagation of matter and information in the quantum realm.

Figures (15)

• Figure 1: Experimental procedure.a, Experimental realization of an array of 1D Bose gases in tubes formed by two retro-reflected laser beams. b, Pictorial representation of the experimental sequence. Top panel: interaction strength $\gamma_i$ throughout the experimental sequence. Middle panel: on average one impurity (blue sphere) per tube is created from a strongly correlated host gas (brown spheres) via a radio-frequency pulse. After its creation the impurity, which is initially not interacting with the host gas, is accelerated by gravity to the desired momentum $Q$. Subsequently the entire system is dropped and interactions between the impurity and the host are switched on. c, Magnetic-field dependence of the scattering length $a$ for collisions between atoms in the host gas (dashed line) and $a_i$ between impurity and host atoms (solid line). d, Left panel: Edge of the excitation spectrum of a 1D Bose gas without impurity -- the plasmon branch (dashed line). Middle panel: The blue parabola corresponds to the dispersion of a single non-interacting impurity. Right panel: The solid line shows the edge of the excitation spectrum with one impurity interacting with the host gas -- the polaron branch. The quench induces excitations (blue shading) in the continuous many-body spectrum above the polaron branch. For both the plasmon and the polaron branch, the excitation spectrum is a $2\,k_\text{F}$-periodic function of the total momentum of the system.
• Figure 2: Subsonic versus supersonic impurity dynamics. Absorption images for initially a, subsonic ($Q\!=\!0.40(1)\, k_\text{F}$) and b, supersonic impurities ($Q\!=\!2.30(8)\, k_\text{F}$), showing the host (top panels, red) and impurity (bottom panels, blue) distributions during the relaxation process for the times indicated. The interactions are set to $(\gamma_i,\gamma)\!=\!(2.6(3),5.9(3))$. Each image is an average of 10 experimental realizations. The vertical axis in each image is converted to a momentum scale. The horizontal axis is not calibrated as this direction is perpendicular to the tubes and no relevant dynamics occur along this direction.
• Figure 3: Time evolution of the impurity momentum distribution $n(k)$.a, b, c, Experimental $n(k)$ for $(\gamma_i,\gamma)\!=\!(11.0(7),18.5(12))$ and selected values of the initial impurity wavevector $Q/k_\text{F} \!=\! 0.2, 0.9, 1.4$, respectively, for times up to $t = 7$$t_\text{F}$ in time steps of $0.6$$t_\text{F}$. d, e, f, Numerical results for $(\gamma_i,\gamma)\!=\!(9.9, 18.7)$ and $Q/k_\text{F} \!=\! 0.3, 0.8, 1.3$, respectively, as discussed in the text. g, h, i, Experimental distributions for from $t\!=\!0$ to $t\!=\!7 t_\text{F}$ in steps of $0.9 t_\text{F}$ (from green to purple). Each experimental dataset is the average of $10$ realizations.
• Figure 4: Ultrafast relaxation dynamics of the impurity. The time evolution of $n_Q$ is measured for $(\gamma_i,\gamma)\!=\!(11.0(7),18.5(12))$ and distinct momenta that, from slowest to fastest decay (yellow to blue), are $Q/k_\text{F}\!=\! 0.2, 0.4, 0.6, 0.9, 1.0, 1.1, 1.4$. The solid curves correspond to exponential fits including an offset. Each data point is the average of 10 repetitions. The error is the standard error of the mean. Inset: Time constant $\tau$ extracted from the exponential fits as a function of $Q$. The solid curve is a phenomenological fit with the function $a(k_\text{F}/Q) + c$, where $a\!=\!0.41(3)$, and $c\!=\!0.64(4)$.
• Figure 5: Dissipationless motion of the impurity after relaxation. The mean steady-state momentum of the impurity $Q_\text{f}/k_\text{F}$ as a function of its initial momentum $Q/k_\text{F}$, for different values of interaction strength $\gamma_i$. Pink, red, purple and blue datasets correspond to $(\gamma_i, \gamma)\!=\! (0, 2.6 )$, $(1.9,5.1)$, $(3.5,7.4)$, and $(11.5, 19.4)$, respectively. Each data point is an average of 10 repetitions except for the non-interacting case ($\gamma_i\!=\!0$), where 3 repetitions were taken. The error is the standard error of the mean. The dashed line is the predicted momentum of a non-interacting particle. The solid curves are results from numerical simulations.