Abstract
Hydrodynamics accurately describe relativistic heavyion collision experiments well before local thermal equilibrium is established^{1}. This unexpectedly rapid onset of hydrodynamics—which takes place on the fastest available timescale—is called hydrodynamization^{2,3,4}. It occurs when an interacting quantum system is quenched with an energy density that is much greater than its groundstate energy density^{5,6}. During hydrodynamization, energy gets redistributed across very different energy scales. Hydrodynamization precedes local equilibration among momentum modes^{5}, which is local prethermalization to a generalized Gibbs ensemble^{7,8} in nearly integrable systems or local thermalization in nonintegrable systems^{9}. Although many theories of quantum dynamics postulate local prethermalization^{10,11}, the associated timescale has not been studied experimentally. Here we use an array of onedimensional Bose gases to directly observe both hydrodynamization and local prethermalization. After we apply a Bragg scattering pulse, hydrodynamization is evident in the fast redistribution of energy among distant momentum modes, which occurs on timescales associated with the Bragg peak energies. Local prethermalization can be seen in the slower redistribution of occupation among nearby momentum modes. We find that the timescale for local prethermalization in our system is inversely proportional to the momenta involved. During hydrodynamization and local prethermalization, existing theories cannot quantitatively model our experiment. Exact theoretical calculations in the Tonks–Girardeau limit^{12} show qualitatively similar features.
This is a preview of subscription content, access via your institution
Access options
Access Nature and 54 other Nature Portfolio journals
Get Nature+, our bestvalue onlineaccess subscription
$29.99 / 30 days
cancel any time
Subscribe to this journal
Receive 51 print issues and online access
$199.00 per year
only $3.90 per issue
Buy this article
 Purchase on Springer Link
 Instant access to full article PDF
Prices may be subject to local taxes which are calculated during checkout
Similar content being viewed by others
Data availability
The data for all the figures can be found at https://doi.org/10.7910/DVN/KFGNRH. Source data are provided with this paper.
Change history
01 August 2023
A Correction to this paper has been published: https://doi.org/10.1038/s41586023064514
References
Rapp, R. Theory highlights of quark matter 2004. J. Phys. G Nucl. Part. Phys. 30, S951–S961 (2004).
Florkowski, W., Heller, M. P. & Spaliński, M. New theories of relativistic hydrodynamics in the LHC era. Rep. Prog. Phys. 81, 046001 (2018).
Schenke, B. The smallest fluid on earth. Rep. Prog. Phys. 84, 082301 (2021).
Berges, J., Heller, M. P., Mazeliauskas, A. & Venugopalan, R. Qcd thermalization: ab initio approaches and interdisciplinary connections. Rev. Mod. Phys. 93, 035003 (2021).
Berges, J., Borsányi, S. & Wetterich, C. Prethermalization. Phys. Rev. Lett. 93, 142002 (2004).
Borsanyi, S. Prethermalization and the first 1 fm/c of a heavy ion collision. Acta Phys. Hung. A 22, 317–323 (2005).
Rigol, M., Dunjko, V., Yurovsky, V. & Olshanii, M. Relaxation in a completely integrable manybody quantum system: an Ab Initio study of the dynamics of the highly excited states of 1D lattice hardcore bosons. Phys. Rev. Lett. 98, 050405 (2007).
Gring, M. et al. Relaxation and prethermalization in an isolated quantum system. Science 337, 1318–1322 (2012).
D’Alessio, L., Kafri, Y., Polkovnikov, A. & Rigol, M. From quantum chaos and eigenstate thermalization to statistical mechanics and thermodynamics. Adv. Phys. 65, 239–362 (2016).
Mallayya, K., Rigol, M. & De Roeck, W. Prethermalization and thermalization in isolated quantum systems. Phys. Rev. X 9, 021027 (2019).
Doyon, B. Lecture notes on generalised hydrodynamics. SciPost Phys. Lect. Notes https://doi.org/10.21468/SciPostPhysLectNotes.18 (2020).
Girardeau, M. Relationship between systems of impenetrable bosons and fermions in one dimension. J. Math. Phys. 1, 516–523 (1960).
Heller, M. P. & Spaliński, M. Hydrodynamics beyond the gradient expansion: resurgence and resummation. Phys. Rev. Lett. 115, 072501 (2015).
Giacalone, G., Mazeliauskas, A. & Schlichting, S. Hydrodynamic attractors, initial state energy, and particle production in relativistic nuclear collisions. Phys. Rev. Lett. 123, 262301 (2019).
Cooper, F., Habib, S., Kluger, Y. & Mottola, E. Nonequilibrium dynamics of symmetry breaking in λΦ^{4} theory. Phys. Rev. D 55, 6471–6503 (1997).
Alexander, M. & Kollar, M. Photoinduced prethermalization phenomena in correlated metals. Phys. Status Solidi B 259, 2100280 (2022).
Lieb, E. H. & Liniger, W. Exact analysis of an interacting Bose gas. I. The general solution and the ground state. Phys. Rev. 130, 1605–1616 (1963).
van den Berg, R. et al. Separation of time scales in a quantum Newton’s cradle. Phys. Rev. Lett. 116, 225302 (2016).
Gross, C. & Bloch, I. Quantum simulations with ultracold atoms in optical lattices. Science 357, 995–1001 (2017).
Kinoshita, T., Wenger, T. & Weiss, D. S. A quantum Newton’s cradle. Nature 440, 900–903 (2006).
Langen, T., Gasenzer, T. and Schmiedmayer, J. Prethermalization and universal dynamics in nearintegrable quantum systems. J. Stat. Mech.https://doi.org/10.1088/17425468/2016/06/064009 (2016).
Wilson, J. M. et al. Observation of dynamical fermionization. Science 367, 1461–1464 (2020).
Malvania, N. et al. Generalized hydrodynamics in strongly interacting 1D bose gases. Science 373, 1129–1132 (2021).
Tang, Y. et al. Thermalization near integrability in a dipolar quantum newton’s cradle. Phys. Rev. X 8, 021030 (2018).
Li, C. et al. Relaxation of bosons in one dimension and the onset of dimensional crossover. SciPost Phys. 9, 058 (2020).
Fabbri, N. et al. Dynamical structure factor of onedimensional bose gases: experimental signatures of beyondluttingerliquid physics. Phys. Rev. A 91, 043617 (2015).
Meinert, F. et al. Probing the excitations of a liebliniger gas from weak to strong coupling. Phys. Rev. Lett. 115, 085301 (2015).
CastroAlvaredo, O. A., Doyon, B. & Yoshimura, T. Emergent hydrodynamics in integrable quantum systems out of equilibrium. Phys. Rev. X 6, 041065 (2016).
Bertini, B., Collura, M., De Nardis, J. & Fagotti, M. Transport in outofequilibrium XXZ chains: exact profiles of charges and currents. Phys. Rev. Lett. 117, 207201 (2016).
Schemmer, M., Bouchoule, I., Doyon, B. & Dubail, J. Generalized hydrodynamics on an atom chip. Phys. Rev. Lett. 122, 090601 (2019).
Eigen, C. et al. Universal prethermal dynamics of bose gases quenched to unitarity. Nature 563, 221–224 (2018).
Yin, X. & Radzihovsky, L. Quench dynamics of a strongly interacting resonant bose gas. Phys. Rev. A 88, 063611 (2013).
Gramsch, C. & Rigol, M. Quenches in a quasidisordered integrable lattice system: dynamics and statistical description of observables after relaxation. Phys. Rev. A 86, 053615 (2012).
Murthy, C. & Srednicki, M. Relaxation to Gaussian and generalized Gibbs states in systems of particles with quadratic Hamiltonians. Phys. Rev. E 100, 012146 (2019).
Kinoshita, T., Wenger, T. & Weiss, D. S. Alloptical boseeinstein condensation using a compressible crossed dipole trap. Phys. Rev. A 71, 011602 (2005).
Olshanii, M. Atomic scattering in the presence of an external confinement and a gas of impenetrable bosons. Phys. Rev. Lett. 81, 938–941 (1998).
Yang, C. N. & Yang, C. P. Thermodynamics of a onedimensional system of bosons with repulsive deltafunction interaction. J. Math. Phys. 10, 1115–1122 (1969).
Tolra, B. L. et al. Observation of reduced threebody recombination in a correlated 1D degenerate bose gas. Phys. Rev. Lett. 92, 190401 (2004).
Rigol, M. & Muramatsu, A. Groundstate properties of hardcore bosons confined on onedimensional optical lattices. Phys. Rev. A 72, 013604 (2005).
Cazalilla, M. A., Citro, R., Giamarchi, T., Orignac, E. & Rigol, M. One dimensional bosons: from condensed matter systems to ultracold gases. Rev. Mod. Phys. 83, 1405–1466 (2011).
Xu, W. & Rigol, M. Expansion of onedimensional lattice hardcore bosons at finite temperature. Phys. Rev. A 95, 033617 (2017).
Acknowledgements
We acknowledge N. Malvania for preliminary experimental work related to this paper. This work was supported by NSF grants PHY2012039 (D.S.W., Y.L. and Y.Z.), PHY2012145 (Y.Z. and M.R.) and DMR1653271 (S.G.). The computations were carried out at the Institute for Computational and Data Sciences at Penn State University.
Author information
Authors and Affiliations
Contributions
Y.L. carried out the experiments and experimental analysis. Y.Z. carried out the theoretical calculations. D.S.W. oversaw the experimental work. M.R. and D.S.W. oversaw the theoretical work. Y.L, Y.Z, S.G, M.R. and D.S.W were all involved in the analysis and interpretation of the results, and all contributed to writing the paper.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature thanks the anonymous reviewers for their contribution to the peer review of this work.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Extended data figures and tables
Extended Data Fig. 1 Theoretical momentum distributions.
Calculations are done for a single 1D tube in the TG limit at zero temperature. We use the same trap and quench parameters as in the \({\bar{\gamma }}_{0}=2.3\) experiment, and choose the atom number in the tube to be N = 32 in order to match the experimental average energy density. These curves match the experimental evolution times in Fig. 1.
Extended Data Fig. 2 Experimental rapidity distributions for \({\bar{{\boldsymbol{\gamma }}}}_{{\bf{0}}}={\bf{2.3}}\) at a sequence of times after the Bragg scattering quench.
The central part of the rapidity distribution does not appreciably change over 1.5 ms, which is less than 10% of the trap oscillation period. The rapidities measurement for the side peaks are slightly distorted by the fact that the flat potential does not extend far enough for atoms moving that fast (after they have moved approximately 20 μm, they are accelerated slightly). Still, since the distortion is approximately the same for all times within the first approximately 0.5 ms, the fact that the measured distribution does not change implies that that part of the rapidity distribution also remains constant. After approximately1 ms, the side peaks start to be noticeably slowed as they climb up the potential of the Gaussian axial trap.
Extended Data Fig. 3 Theoretical rapidity distributions.
These simulations are done for a single 1D tube in the TG limit at zero temperature with same trap and quench parameters as in the \({\bar{\gamma }}_{0}=2.3\) experiment. The atom number in the tube is chosen to be N = 32 in order to match the experimental average energy density. These curves match the experimental evolution times of Extended Data Fig. 2.
Extended Data Fig. 4 Integrated experimental rapidity energy.
The curves are extracted from the rapidity profiles for \({\bar{\gamma }}_{0}=2.3\) integrating over different 0.2 ħk wide rapidity groups. The different colours denote different rapidity groups (as in Fig. 3), defined in the key. For the average rapidity \(\bar{\theta }=0.5\,\hbar k1.3\,\hbar k\) curves, we have horizontally shifted the points for times longer than t_{ev} = 20 μs in order to better resolve the different rapidity group energies at the same time. The dashed lines show the average energy for each rapidity group. There is no detectable change in the energy of each rapidity group in the first approximately 300 μs, in stark contrast to the 3 dBscale rapid changes in the energies associated with momentum groups in Fig. 3b.
Extended Data Fig. 5 Effect of finite temperature on momentum distributions and integrated kinetic energy.
a–c, Theoretical momentum distributions for a single 1D tube in the TG limit at 0 nK, 5 nK and 10 nK, respectively (see Methods). We use the same trap and quench parameters as in the \({\bar{\gamma }}_{0}=3.4\) experiment, and choose the atom number in the tube to be N = 29 to match the experimental average energy density. These temperatures are the highest we expect there to be in the experiment. One can see that the effect is mostly to smooth out the sharp peaks. Note also that similar smoothing also results from finite timeofflight in the experiment [24]. d–f, Corresponding theoretical integrated kinetic energy at 0 nK, 5 nK and 10 nK, respectively. Each curve shows the time evolution of the integrated energy in a different 0.2 ħk_{0} wide momentum group. Curves in the same momentum group have very similar shapes at zero and finite temperatures. Since hydrodynamization involves a much larger energy scale than that associated with these temperatures, it is not surprising that temperature does not significantly affect the signatures of hydrodynamization. We conclude that the differences between the experimental and theoretical results in Fig. 3 are not a result of finite temperature effects.
Extended Data Fig. 6 The time evolution of the occupations (O) of different momentum groups within the central peak.
a, Experimental curves for \({\bar{\gamma }}_{0}=2.3\) are plotted on a log scale. Each curve is obtained by integrating the area of the normalized momentum distribution within the designated momentum range. Different colours denote the different momentum groups as shown in the legend. The last three momentum groups, 0.30 − 0.36 ħk, 0.36 − 0.42 ħk, and 0.42 − 0.48 ħk, have twice the integration range, so their occupations are divided by 2. b, Theoretical curves for a single 1D gas in the TG limit with the same average energy as in the experiment with \({\bar{\gamma }}_{0}=2.3\). In both the experiment and the theory, the occupation of higher momentum groups evolves faster than the lower ones, as expected for local prethermalization. The fact that each of these curves has a different shape makes it difficult to quantitatively compare time constants among them. The theoretical curves evolve more than twice as fast as the experimental curves, presumably reflecting the difference between infinite and finite g (see Eq. (1)). k = k_{0} for the experiment and k = 4k_{0} for the theory.
Extended Data Fig. 7 Time evolution of the FWHM.
a, The evolution of the FWHM of the central peak in the momentum distributions after the Bragg scattering quench for different coupling strengths. To a greater degree than for the p_{50}(t_{ev}) curves of Fig. 3, these curves all have different shapes. b, The evolving momentum distribution of the central peak for \({\bar{\gamma }}_{0}=0.94\). These curves correspond to the first five points of the orange curve in a. The momentum distribution clearly evolves during the first 0.05 ms, even though the FWHM does not change. This illustrates that the FWHM is not a reliable marker of the evolution of these momentum distributions.
Extended Data Fig. 8 Relationships among p_{f}, \({{\boldsymbol{\tau }}}_{{{\boldsymbol{p}}}_{{\boldsymbol{f}}}}\), and ϵ.
a, Experimental time constants associated with p_{40} (red circles), p_{50} (black squares), and p_{60} (blue triangles) as functions of \(1/k\sqrt{{\epsilon }}\). As in all the experiments presented in this paper, k is fixed at k_{0}. The time constants are extracted from curves like those in Fig. 4a (see Methods). The dashed lines are leastsquares linear fits; the intercepts are −2.5 ± 5.6, −5.8 ± 4.9, −6.0 ± 3.5 for p_{40}, p_{50}, and p_{60}, respectively. The data are consistent with linear relationships between the time constant associated with each p_{f} and the inverse of the square root of average energy per particle. For a given momentum distribution, the actual values of p_{40} and p_{60} span a range of approximately ±30% around the steepest part of the distribution (at approximately p_{50}). b, Experimental p_{50} vs \(\sqrt{{\epsilon }}\). The green triangles, magenta squares, and orange circles correspond to the initial (\({p}_{50}^{i}\)), middle (\({p}_{50}^{m}\)), and final (\({p}_{50}^{e}\)) values of p_{50} for each experimental condition, extracted from Fig. 4a. The dashed lines are leastsquares linear fits; the intercepts are (1.7 ± 0.5) × 10^{−2}, (8.4 ± 3) × 10^{−3}, and (1.9 ± 2.2) × 10^{−3} for \({p}_{50}^{i}\), \({p}_{50}^{m}\), and \({p}_{50}^{e}\), respectively. The \({p}_{50}^{i}\) points at the lowest ϵ (highest \({\bar{\gamma }}_{0}\)) conditions are more likely to be affected by finitesize corrections to their momentum distributions. The data show a linear relationship between each measured value and \(\sqrt{{\epsilon }}\). c, Theoretical time constants obtained from the p_{40}, p_{50}, and p_{60} curves simulated with k = 4k_{0}. The time constants are obtained from curves like those in Fig. 4b. The error bars are smaller than the marker size. The solid lines are leastsquares linear fits; the intercepts are −0.065 ± 0.047, − 0.47 ± 0.13, and −0.67 ± 0.10 for p_{40}, p_{50}, and p_{60}, respectively. d, Theoretical p_{50} vs \(\sqrt{{\epsilon }}\), simulated with k = 4k_{0}. The intercepts are (14 ± 1.3) × 10^{−4}, (2.7 ± 0.53) × 10^{−4}, and (19 ± 1.2) × 10^{−4}, for \({p}_{50}^{i}\), \({p}_{50}^{m}\), and \({p}_{50}^{e}\), respectively. The momentum feature that is most clearly proportional to \(\sqrt{{\epsilon }}\) is \({p}_{50}^{m}\). Since that is the midpoint value of p_{50} during the evolution, it is likely to be close to \({\bar{p}}_{50}\), the effective momentum to which the p_{50} measurement is sensitive. Taken all together, this figure shows that, for p_{50}, the time constants are proportional to \(1/\sqrt{{\epsilon }}\), which is in turn proportional to the characteristic momentum being measured. We have repeated the entire analysis for p_{40} and p_{60} and the conclusions are the same.
Extended Data Fig. 9 Effect of the average over 1D gases and of finite temperature on p_{50}.
We plot theoretical simulations of p_{50} in the TG limit with the same trap and quench parameters as in the \({\bar{\gamma }}_{0}=3.4\) experiment. For the average over 1D gases, we use the ThomasFermi distribution with experimentally measured R_{TF} = 23 μm and a total particle number of N_{tot} = 2.75 × 10^{5} (see Methods). To simplify the calculations, we round the particle number in each tube in steps of 5. The circles show the results of ground state simulations for a single tube with N = 29 particles (which matches the experimental average energy density). The squares show the results of ground state simulations after averaging over all the 1D gases, as occurs in the experimental setup. The triangles (stars) show simulations for a single tube with N = 29 particles at a temperature of T = 5 nK (10 nK). We did the finite temperature simulations with a larger discretization a = 3.2 × 10^{−8} m (a = 4 × 10^{−8} m) due to numerical limitations (see Methods). The results show no significant changes in the time constant due to either the average over tubes or finite temperature.
Source data
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author selfarchiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Le, Y., Zhang, Y., Gopalakrishnan, S. et al. Observation of hydrodynamization and local prethermalization in 1D Bose gases. Nature 618, 494–499 (2023). https://doi.org/10.1038/s41586023059799
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/s41586023059799
This article is cited by

The theory of generalised hydrodynamics for the onedimensional Bose gas
AAPPS Bulletin (2023)
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.