Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Article
  • Published:

Spin transport in a tunable Heisenberg model realized with ultracold atoms


Simple models of interacting spins have an important role in physics. They capture the properties of many magnetic materials, but also extend to other systems, such as bosons and fermions in a lattice, gauge theories, high-temperature superconductors, quantum spin liquids, and systems with exotic particles such as anyons and Majorana fermions1,2. To study and compare these models, a versatile platform is needed. Realizing such systems has been a long-standing goal in the field of ultracold atoms. So far, spin transport has only been studied in systems with isotropic spin–spin interactions3,4,5,6,7,8,9,10,11,12. Here we realize the Heisenberg model describing spins on a lattice, with fully adjustable anisotropy of the nearest-neighbour spin–spin couplings (called the XXZ model). In this model we study spin transport far from equilibrium after quantum quenches from imprinted spin-helix patterns. When spins are coupled only along two of three possible orientations (the XX model), we find ballistic behaviour of spin dynamics, whereas for isotropic interactions (the XXX model), we find diffusive behaviour. More generally, for positive anisotropies, the dynamics ranges from anomalous superdiffusion to subdiffusion, whereas for negative anisotropies, we observe a crossover in the time domain from ballistic to diffusive transport. This behaviour is in contrast with expectations from the linear-response regime and raises new questions in understanding quantum many-body dynamics far away from equilibrium.

This is a preview of subscription content, access via your institution

Access options

Buy this article

Prices may be subject to local taxes which are calculated during checkout

Fig. 1: Tunability of the XXZ model with 7Li and implemenation of a spin helix.
Fig. 2: Ballistic and diffusive spin transport.
Fig. 3: Power-law scalings.
Fig. 4: Temporal crossover from ballistic to diffusive transport for negative anisotropies Δ < 0.

Similar content being viewed by others

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.


  1. Lee, P. A. From high temperature superconductivity to quantum spin liquid: progress in strong correlation physics. Rep. Prog. Phys. 71, 012501 (2007).

    ADS  Google Scholar 

  2. Broholm, C. et al. Quantum spin liquids. Science 367, eaay0668 (2020).

    CAS  PubMed  Google Scholar 

  3. Nichols, M. A. et al. Spin transport in a Mott insulator of ultracold fermions. Science 363, 383–387 (2019).

    ADS  CAS  PubMed  Google Scholar 

  4. Brown, R. C. et al. Two-dimensional superexchange-mediated magnetization dynamics in an optical lattice. Science 348, 540–544 (2015).

    ADS  MathSciNet  CAS  PubMed  MATH  Google Scholar 

  5. Fukuhara, T. et al. Quantum dynamics of a mobile spin impurity. Nat. Phys. 9, 235–241 (2013).

    CAS  Google Scholar 

  6. Fukuhara, T. et al. Microscopic observation of magnon bound states and their dynamics. Nature 502, 76–79 (2013).

    ADS  CAS  PubMed  Google Scholar 

  7. Hild, S. et al. Far-from-equilibrium spin transport in Heisenberg quantum magnets. Phys. Rev. Lett. 113, 147205 (2014).

    ADS  PubMed  Google Scholar 

  8. Koschorreck, M., Pertot, D., Vogt, E. & Köhl, M. Universal spin dynamics in two-dimensional Fermi gases. Nat. Phys. 9, 405–409 (2013); corrigendum 10, 170 (2014).

    CAS  Google Scholar 

  9. Trotzky, S. et al. Observation of the Leggett–Rice effect in a unitary Fermi gas. Phys. Rev. Lett. 114, 015301 (2015).

    ADS  CAS  PubMed  Google Scholar 

  10. Palzer, S., Zipkes, C., Sias, C. & Köhl, M. Quantum transport through a Tonks–Girardeau gas. Phys. Rev. Lett. 103, 150601 (2009).

    ADS  PubMed  Google Scholar 

  11. Sommer, A., Ku, M., Roati, G. & Zwierlein, M. W. Universal spin transport in a strongly interacting Fermi gas. Nature 472, 201–204 (2011).

    ADS  CAS  PubMed  Google Scholar 

  12. Krinner, S. et al. Mapping out spin and particle conductances in a quantum point contact. Proc. Natl. Acad. Sci. USA 113, 8144–8149 (2016).

    ADS  CAS  PubMed  Google Scholar 

  13. Vasseur, R. & Moore, J. E. Nonequilibrium quantum dynamics and transport: from integrability to many-body localization. J. Stat. Mech. 2016, 064010 (2016).

    MathSciNet  MATH  Google Scholar 

  14. Bertini, B. et al. Finite-temperature transport in one-dimensional quantum lattice models. Preprint at (2020).

  15. Ljubotina, M., Žnidarič, M. & Prosen, T. Spin diffusion from an inhomogeneous quench in an integrable system. Nat. Commun. 8, 16117 (2017).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  16. Gopalakrishnan, S. & Vasseur, R. Kinetic theory of spin diffusion and superdiffusion in XXZ spin chains. Phys. Rev. Lett. 122, 127202 (2019).

    ADS  CAS  PubMed  Google Scholar 

  17. Ilievski, E., De Nardis, J., Medenjak, M. & Prosen, T. Superdiffusion in one-dimensional quantum lattice models. Phys. Rev. Lett. 121, 230602 (2018).

    ADS  PubMed  Google Scholar 

  18. Else, D. V., Monroe, C., Nayak, C. & Yao, N. Y. Discrete time crystals. Annu. Rev. Condens. Matter Phys. 11, 467–499 (2020).

    Google Scholar 

  19. Basov, D. N., Averitt, R. D. & Hsieh, D. Towards properties on demand in quantum materials. Nat. Mater. 16, 1077–1088 (2017).

    ADS  CAS  PubMed  Google Scholar 

  20. Langen, T., Gasenzer, T. & Schmiedmayer, J. Prethermalization and universal dynamics in near-integrable quantum systems. J. Stat. Mech. 2016, 064009 (2016).

    MathSciNet  MATH  Google Scholar 

  21. Zhang, J. et al. Observation of a many-body dynamical phase transition with a 53-qubit quantum simulator. Nature 551, 601–604 (2017).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  22. Bernien, H. et al. Probing many-body dynamics on a 51-atom quantum simulator. Nature 551, 579–584 (2017).

    ADS  CAS  PubMed  Google Scholar 

  23. Barends, R. et al. Digital quantum simulation of fermionic models with a superconducting circuit. Nat. Commun. 6, 7654 (2015).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  24. Davis, E. J. et al. Protecting spin coherence in a tunable Heisenberg model. Phys. Rev. Lett. 125, 060402 (2020).

    ADS  CAS  PubMed  Google Scholar 

  25. Signoles, A. et al. Glassy dynamics in a disordered Heisenberg quantum spin system. Preprint at (2019).

  26. Trotzky, S. et al. Time-resolved observation and control of superexchange interactions with ultracold atoms in optical lattices. Science 319, 295–299 (2008).

    ADS  CAS  PubMed  Google Scholar 

  27. Gross, C. & Bloch, I. Quantum simulations with ultracold atoms in optical lattices. Science 357, 995–1001 (2017).

    ADS  CAS  PubMed  Google Scholar 

  28. Jaksch, D., Bruder, C., Cirac, J. I., Gardiner, C. W. & Zoller, P. Cold bosonic atoms in optical lattices. Phys. Rev. Lett. 81, 3108–3111 (1998).

    ADS  CAS  Google Scholar 

  29. Kuklov, A. B. & Svistunov, B. V. Counterflow superfluidity of two-species ultracold atoms in a commensurate optical lattice. Phys. Rev. Lett. 90, 100401 (2003).

    ADS  CAS  PubMed  Google Scholar 

  30. Duan, L.-M., Demler, E. & Lukin, M. D. Controlling spin exchange interactions of ultracold atoms in optical lattices. Phys. Rev. Lett. 91, 090402 (2003).

    ADS  PubMed  Google Scholar 

  31. García-Ripoll, J. J. & Cirac, J. I. Spin dynamics for bosons in an optical lattice. New J. Phys. 5, 76 (2003).

    Google Scholar 

  32. Altman, E., Hofstetter, W., Demler, E. & Lukin, M. D. Phase diagram of two-component bosons on an optical lattice. New J. Phys. 5, 113 (2003).

    ADS  Google Scholar 

  33. Amato-Grill, J., Jepsen, N., Dimitrova, I., Lunden, W. & Ketterle, W. Interaction spectroscopy of a two-component Mott insulator. Phys. Rev. A 99, 033612 (2019).

    ADS  Google Scholar 

  34. Jordan, P. & Wigner, E. Über das Paulische Äquivalenzverbot. Z. Phys. 47, 631–651 (1928).

    ADS  CAS  MATH  Google Scholar 

  35. Ioffe, A. F. & Regel, A. R. Non-crystalline, amorphous and liquid electronic semiconductors. Prog. Semiconduct. 4, 237–291 (1960).

    Google Scholar 

  36. Mott, N. F. Conduction in non-crystalline systems IX. the minimum metallic conductivity. Philos. Mag. A 26, 1015–1026 (1972).

    ADS  CAS  Google Scholar 

  37. Brown, P. T. et al. Bad metallic transport in a cold atom Fermi–Hubbard system. Science 363, 379–382 (2019).

    ADS  CAS  PubMed  Google Scholar 

  38. Dubkov, A. A., Spagnolo, B. & Uchaikin, V. V. Lévy flight superdiffusion: an introduction. Int. J. Bifurc. Chaos 18, 2649–2672 (2008).

    MATH  Google Scholar 

  39. Andreev, A. È. et al. Correlation theory of processes with stationary random increments of order n. Am. Math. Soc. Transl. 8, 87 (1958).

    Google Scholar 

  40. Agarwal, K., Gopalakrishnan, S., Knap, M., Müller, M. & Demler, E. Anomalous diffusion and Griffiths effects near the many-body localization transition. Phys. Rev. Lett. 114, 160401 (2015).

    ADS  PubMed  Google Scholar 

  41. Vosk, R., Huse, D. A. & Altman, E. Theory of the many-body localization transition in one-dimensional systems. Phys. Rev. X 5, 031032 (2015).

    Google Scholar 

  42. Guardado-Sanchez, E. et al. Subdiffusion and heat transport in a tilted two-dimensional Fermi–Hubbard system. Phys. Rev. X 10, 011042 (2020).

    CAS  Google Scholar 

  43. Antal, T., Rácz, Z., Rákos, A. & Schütz, G. M. Transport in the XX chain at zero temperature: emergence of flat magnetization profiles. Phys. Rev. E 59, 4912–4918 (1999).

    ADS  CAS  Google Scholar 

  44. Gobert, D., Kollath, C., Schollwöck, U. & Schütz, G. Real-time dynamics in spin-\(\tfrac{1}{2}\) chains with adaptive time-dependent density matrix renormalization group. Phys. Rev. E 71, 036102 (2005).

    ADS  MathSciNet  Google Scholar 

  45. Misguich, G., Pavloff, N. & Pasquier, V. Domain wall problem in the quantum XXZ chain and semiclassical behavior close to the isotropic point. SciPost Phys. 7, 025 (2019).

    ADS  MathSciNet  CAS  Google Scholar 

  46. Ljubotina, M., Žnidarič, M. & Prosen, T. A class of states supporting diffusive spin dynamics in the isotropic Heisenberg model. J. Phys. A Math. Theor. 50, 475002 (2017).

    ADS  MathSciNet  MATH  Google Scholar 

  47. Castro-Alvaredo, O. A., Doyon, B. & Yoshimura, T. Emergent hydrodynamics in integrable quantum systems out of equilibrium. Phys. Rev. X 6, 041065 (2016).

    Google Scholar 

  48. Bertini, B., Collura, M., De Nardis, J. & Fagotti, M. Transport in out-of-equilibrium XXZ chains: exact profiles of charges and currents. Phys. Rev. Lett. 117, 207201 (2016).

    ADS  PubMed  Google Scholar 

  49. Caux, J.-S. & Mossel, J. Remarks on the notion of quantum integrability. J. Stat. Mech. 2011, P02023 (2011).

    MATH  Google Scholar 

  50. Dimitrova, I. et al. Enhanced superexchange in a tilted Mott insulator. Phys. Rev. Lett. 124, 043204 (2020).

    ADS  CAS  PubMed  Google Scholar 

  51. Greiner, M. Ultracold Quantum Gases in Three-dimensional Optical Lattice Potentials. PhD thesis, Ludwig-Maximilians-Universität München (2003).

  52. Łącki, M., Delande, D. & Zakrzewski, J. Dynamics of cold bosons in optical lattices: effects of higher Bloch bands. New J. Phys. 15, 013062 (2013).

    Google Scholar 

  53. Will, S. et al. Time-resolved observation of coherent multi-body interactions in quantum phase revivals. Nature 465, 197–201 (2010).

    ADS  CAS  PubMed  Google Scholar 

  54. Lühmann, D.-S., Jürgensen, O. & Sengstock, K. Multi-orbital and density-induced tunneling of bosons in optical lattices. New J. Phys. 14, 033021 (2012).

    ADS  Google Scholar 

  55. Stöferle, T., Moritz, H., Schori, C., Köhl, M. & Esslinger, T. Transition from a strongly interacting 1D superfluid to a Mott insulator. Phys. Rev. Lett. 92, 130403 (2004).

    ADS  PubMed  Google Scholar 

  56. Secker, T., Amato-Grill, J., Ketterle, W. & Kokkelmans, S. High-precision analysis of Feshbach resonances in a Mott insulator. Phys. Rev. A 101, 042703 (2020).

    ADS  MathSciNet  CAS  Google Scholar 

  57. Bradley, C. C., Sackett, C. A. & Hulet, R. G. Bose–Einstein condensation of lithium: observation of limited condensate number. Phys. Rev. Lett. 78, 985–989 (1997).

    ADS  CAS  Google Scholar 

  58. Ketterle, W., Durfee, D. S. & Stamper-Kurn, D. M. Making, probing and understanding Bose–Einstein condensates. In Bose–Einstein Condensation in Atomic Gases (eds Inguscio, M. et al.) 67–176 (IOS Press, 1999).

Download references


We thank M. D. Lukin, N. Y. Yao and M. Knap for discussions, Y. K. Lee for experimental assistance, as well as C. M. E. Paus for sharing computing resources, and J. de Hond for comments on the manuscript. We acknowledge support from the NSF through the Center for Ultracold Atoms and grant number 1506369, NSF (grant number OAC-193471), ARO-MURI Non-Equilibrium Many-Body Dynamics (grant number W911NF-14-1-0003), AFOSR-MURI Photonic Quantum Matter (grant number FA9550-16-1-0323), AFOSR-MURI Quantum Phases of Matter (grant number FA9550-14-1-0035), ONR (grant number N00014-17-1-2253), ARO (grant number W911NF-20-1-0163), the Vannevar Bush Faculty Fellowship, and the Gordon and Betty Moore Foundation EPiQS Initiative (grant number GBMF4306).

Author information

Authors and Affiliations



P.N.J., J.A.-G., I.D. and W.K. conceived the experiment. P.N.J., J.A.-G. and I.D. developed the experimental setup and took the data. P.N.J. led the data analysis. W.W.H. and E.D. developed the theoretical analysis. W.W.H. performed the numerical simulations. All authors discussed the results and contributed to the writing of the manuscript.

Corresponding author

Correspondence to Paul Niklas Jepsen.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information Nature thanks Jean-Philippe Brantut, Tomaž Prosen and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

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 Determination of the Heisenberg parameters.

a, Corrections for tunnelling \((\tilde{t}-\tilde{t}{}^{(0)})/\tilde{t}{}^{(0)}\) (green), on-site interactions (U −U(0))/U(0) (blue), superexchange (J − J(0))/J(0) (purple) and off-site interactions −2V/J(0) (orange), where \(\tilde{t}\)(0), U(0) and \({J}^{(0)}=4{(\tilde{t}{}^{(0)})}^{2}/{U}^{(0)}\) are the uncorrected values and \(\tilde{t}\), U and \(J\,=\,4{\tilde{t}}^{2}/U\) include corrections, at a lattice depth of 13ER (solid line) and 9ER (dotted line). b, As a function of magnetic field B we show the scattering lengths a (blue), a (purple) and a (orange) measured in our previous work (points)33. Here a0 is the Bohr radius. c, d, We interpolate the data in b using hyperbolic fits (solid lines) and calculate values for the transverse coupling constant Jxy (purple), the longitudinal coupling constant Jz (yellow) and the anisotropy Δ = Jz/Jxy (black), without corrections (dashed line) and including corrections (solid line) for a lattice depth of 11ER. The excluded region (|a| > 700a0) is around a Feshbach resonance in the |↑ state near 845.4 G.

Extended Data Fig. 2 Contrast measurement.

ad, The distribution of atoms in the |↑ state. Every pixel is a local measurement of the column density (number of atoms per unit area). The y and z axes are displayed in units of the lattice spacing a = 0.532 μm. The images are projected (integrated) along the y direction from y = −30a to +30a to obtain the linear density (number of atoms per unit length). The resulting 1D distributions are fitted with \(f(z)=g(z)[1+{\mathscr{C}}\,\cos (Qz+\theta )]/2\) (solid line), where g(z) is a Gaussian envelope (dashed line), between z = ±54a. The data in ad were measured at different evolution times t = 0 (a), 2.3ħ/Jxy (b), 6.3ħ/Jxy (c) and 12.0ħ/Jxy (d), for anisotropy Δ ≈ 0 and wavelength λ = 10.4a. The obtained contrast \({\mathscr{C}}(t)\) is shown in Fig. 2a. In general, we also normalize by the initial contrast \({\mathscr{C}}(0)\) to correct for finite optical imaging resolution. This is important for shorter wavelengths λ close to the optical resolution of 3 μm, where the measured contrast \({\mathscr{C}}(t)\) is reduced compared to the real contrast \(c(t)={\mathscr{C}}(t)/{\mathscr{C}}(0)\).

Extended Data Fig. 3 Dispersion relations.

For all positive anisotropies Δ ≥ 0, the time evolution of the contrast c(t) shows a damped oscillatory component, in addition to the overall exponential decay. For larger Δ, the oscillations become smaller. ac, Decay and weak oscillation at the isotropic point Δ ≈ 1 measured for different wavelengths λ, at three different lattice depths 9ER (orange), 11ER (blue) and 13ER (yellow). Solid lines are fits c(t) = [a0 + b0cos(ωt)]et/τ + c0 and dashed lines show the overall decay a0et/τ + c0, around which the oscillations take place. The oscillations become more pronounced for short wavelengths λ, because the decay time (τ λ2) decreases with smaller wavelength more strongly than the oscillation period (T λ). d, The oscillation frequencies follow linear dispersion relations ω(Q) = vQ shown for Δ = −0.12 (red), 0.35 (orange), 0.78 (yellow), 1.01 (blue) and 1.27 (light blue). e, The obtained velocities v decrease with increasing anisotropy Δ. For Δ = 1.58 (open symbol), oscillations are small and the measurement was limited to large values of Q, which precluded recording a full dispersion relation. We note that although the oscillations are difficult to discern by eye (for example, in a), especially for large anisotropies Δ and small wavevectors Q, the fitted oscillation frequencies ω all fall very well on linear dispersion relations, which demonstrates that those barely visible oscillations are real. The linear scaling ω(Q) = vQ persists even in the superdiffusive, diffusive and subdiffusive regimes, where the power-law scaling of the decay time constant τ Qα is strongly nonlinear. This small ballistic (oscillatory) component may be related to our initial condition of a spin helix, which in the mapping to lattice fermions is a 100% density modulation, which reduces scattering at early times.

Extended Data Fig. 4 Effect of finite hole concentration.

By varying the thermal fraction Nth/N of the Bose–Einstein condensate before it is loaded into the optical lattice, we vary the energy and entropy of the atoms in the spin chain, and therefore the concentration of holes. (For our conditions, doubly occupied sites have higher energies than holes). Measurements are shown here for Δ ≈ 0 and λ = 10.4a, at a lattice depth of 11ER. a, Decay curves c(t) for varying hole concentrations ranging from low (blue) to high (orange) thermal fraction. Solid lines are fits c(t) = [a0 + b0cos(ωt)]et/τ + c0. b, The background contrast c0 increases monotonously with thermal fraction Nth/N. A linear fit (solid line) extrapolates to c0 = 0.01(2), consistent with zero, for Nth/N = 0. This suggests that all of the background contrast is due to hole excitations. c, Higher hole concentrations suppress the oscillating fraction b0/(a0 + b0). d, Holes do not affect the oscillation period T = 2π/ω. e, Holes decrease the decay time τ, albeit slightly. be show that almost all of our measurements are not sensitive to a small thermal fraction, which is usually Nth/N ≤ 0.05 throughout this work. The behaviour shown in c and e is most probably caused by mobile holes in the central part of the Mott insulator. Indeed, numerical simulations of the \(\tilde{t}\)J model reproduce such effects (Fig. 2a). Note though that for the isotropic case Δ ≈ 1, a previous work7 found a ~50% change in decay time when the hole concentration changed from 0 to 5%. Our numerical simulations (Extended Data Fig. 8b) do not show such strong sensitivity (for any anisotropy, even at Δ = 1), possibly owing to asymmetry in the on-site interactions (U ≠ U ≠ U) in our system. On the other hand, a finite background contrast (b) is probably caused by immobile holes located in the outer parts of the atom distribution where first-order tunnelling is suppressed by the gradient of the (harmonic) trapping potential50. Immobile holes disrupt spin transport, and so we expect that the imprinted spin modulation in these regions will not (or only very slowly) decay. f, g, The region with immobile holes is visible as a shell of low atomic density surrounding the Mott insulator in the in situ images for large hole concentration (f) and is absent for low hole concentration (g). The three curves in both f and g show the local contrast as a function of distance r from the centre of the atom cloud for the evolution times t = 0 (top), 2.7ħ/Jxy (middle) and 21.7ħ/Jxy (bottom). The two in situ images in both f and g are for t = 0 (top) and 21.7ħ/Jx (bottom). The dashed lines indicate contours of constant radius, r = 30a (f) and r = 20a (g).

Extended Data Fig. 5 Decay behaviour as a function of anisotropy.

a, b, Decay behaviour ranging from negative (a) to positive (b) anisotropy, for a fixed wavelength λ = 10.4a. Using Δ ≈ 0 as a reference point, we show how the temporal profile of the decay curve c(t) changes when we introduce positive or negative interactions. Every data point is an average of two measurements at lattice depths 11ER and 13ER. In a, from bottom to top: Δ = −0.12 (red), −0.59 (pink), −0.81 (yellow), −1.02 (blue), −1.43 (green) and −1.79 (purple). In b, from bottom to top: Δ = −0.13 (red), 0.08 (purple), 0.35 (pink), 0.55 (orange), 0.78 (yellow), 1.01 (blue), 1.27 (light blue) and 1.58 (green). Regardless of sign, for increasing |Δ| the decay always slows down and the revivals damp more quickly. However, there is a big difference in how this slowdown happens: for increasing positive interactions Δ > 0, the initial rate of decay decreases continuously (b); by contrast, for all negative interactions Δ < 0, the initial rate of decay stays constant (and is ballistic), coinciding with the Δ ≈ 0 case (a). It is only after a critical time t0 that the decay suddenly starts slowing down (and becomes diffusive) for times t > t0. This critical time t0 decreases with increasing negative interaction strength |Δ|.

Extended Data Fig. 6 Collapse of decay curves for positive anisotropies.

All decay curves c(t) for wavelengths λ = 15.7a, 13.4a, 11.7a, 10.4a, 9.4a, 8.5a, 7.8a, 7.2a and 6.7a collapse very well into a single curve for all evolution times t, when time units are rescaled by λα, where the exponent α is a function of anisotropy Δ, both for experiment (points) and theory (solid lines). Experimental points were measured for lattice depths 9ER (red), 11ER (blue) and 13ER (yellow). a, b, Ballistic regime (α = 1). c, Superdiffusion (α = 1.5). d, Diffusion (α = 2). e, f, Subdiffusion (α = 2.5, 3 for experiment and α = 3.5, 4.5 for numerical simulations; in f, experiments covered a reduced range λ ≤ 10.4a). For all anisotropies Δ ≥ 0 (af) the experimentally measured oscillation frequencies ω follow linear dispersion relations (Extended Data Fig. 3) and have a scaling behaviour different from the decay rates. However, such oscillations are small outside the ballistic regime α ≈ 1, and therefore only lead to a small deviation from the collapse behaviour. Note also the different timescales in experiments and simulations for Δ > 1.

Extended Data Fig. 7 Collapse at short times for negative anisotropies.

All decay curves c(t) for different wavelengths λ collapse into a single curve at early times, when time units are rescaled by λ (indicating ballistic behaviour). For later times the decay is diffusive with different scaling. ac, Theory (from top to bottom: λ = 31.3a, 23.5a, 18.8a, 15.7a, 13.4a, 11.7a, 10.4a, 9.4a, 8.5a, 7.8a, 7.2a, 6.7a and 6.3a). The dotted lines are exponential fits \({{\rm{e}}}^{-t/{\tau }_{{\rm{II}}}}\) to the diffusive regime and the time constants τII are shown in Extended Data Fig. 8a. df, Experiment (from top to bottom: λ = 18.8a, 13.4a, 10.4a, 8.5a, 7.2a and 6.3a). Every data point is an average of two measurements at lattice depths 11ER and 13ER. The black dashed line indicates the ballistic case Δ ≈ 0 (see Fig. 2c).

Extended Data Fig. 8 Power-law scalings (theory) and diffusion coefficients.

a, b, Decay time constants τ for different anisotropies Δ ranging from negative (a) to positive (b). Numerical results are shown in a for Δ = −1 (blue), −1.5 (green) and −2 (purple) and in b for Δ = 0 (red), 0.5 (orange), 0.85 (yellow), 1 (blue) and 1.5 (green). Solid lines are power-law fits (to the filled symbols). Open symbols are excluded from the fit owing to finite-size effects. Crossed symbols are results from \(\tilde{t}\)J model simulations including 5% hole fraction. Fitted power-law exponents are shown in Fig. 3c. For positive anisotropies Δ ≥ 0 the decay time τ is defined as τ = τ′/ln(1/0.60) with c(τ′) = 0.60. For negative anisotropies Δ < 0, the decay time τI for short times (I) is defined as \({\tau }_{{\rm{I}}}=10{\tau }_{{\rm{I}}}^{{\prime} }\) with \(c({\tau }_{{\rm{I}}}^{{\prime} })=0.90\). For longer times (II), the decay time τII is obtained from exponential fits \({{\rm{e}}}^{-t/{\tau }_{{\rm{II}}}}\) to the diffusive long-time tail (see dotted curves in Extended Data Fig. 7a–c). c, Diffusion coefficients for the diffusive long-time regime (II) obtained from theory (open symbols) and experiment (filled symbols). For negative anisotropies Δ < 0, values were determined from quadratic power-law fits 1/τ = DQ2 to the data points in a (theory) and Fig. 3a (experiment) for the diffusive regime (II). Note that for Δ ≥ 0 the system is only diffusive for Δ = +1, as shown in b (theory) and Fig. 3b (experiment). From the experimental diffusion coefficients, we estimate mean free paths δx using the velocity v = 0.76(1)vF from the ballistic short-time regime (I), and obtain δx = 3.35(15)a, 1.07(7)a and 0.66(4)a for Δ = −1.02, −1.43 and −1.79, respectively. d, Short-time (t ħ/Jxy) decay constant τ = |Γ|−1 obtained from Taylor expansion of the contrast c(t) = 1 + Γ2t2 + … as a function of Q, for Δ = 0 (red), 0.55 (orange), 0.85 (yellow), 0.95 (purple) and 1 (blue). For Δ < 1 all curves in the log–log plot asymptote to the same slope as Q → 0 (continuum limit), whereas there are deviations for larger wavevectors Q. For Δ = 1 the slope is instead different. This indicates that the power-law exponent α in τ Qα depends on the range of wavevectors Q used to determine it. e, Power-law exponents α determined for the short-wavelength regime between λ = 6a and 20a (filled symbols) as in experiments and numerics, and for the long-wavelength regime between λ = 150a and 200a (open symbols) approaching the continuum limit. In the former case, the exponents show a smooth crossover from superballistic to diffusive as Δ → 1 similar to that in the experiments and numerics, whereas in the latter case the exponents show a sharp jump from ballistic to diffusive occurring exactly at Δ = 1.

Extended Data Fig. 9 Finite-size effects from the initial phase of the spin helix.

a, The time evolution of the contrast c(t) depends strongly on the initial phase θ, illustrated here by simulations for Δ = 0 and λ = 10.4a. b, c, The dynamics of the local magnetization \(\langle {S}_{i}^{z}(t)\rangle \) for phases θ = 0 (b) and π/2 (c) reveals that this arises owing to the reflection of ballistically propagating magnetization off the ends of the chain. Depending on the initial phase of the spin helix, the reflected magnetization interferes constructively or destructively with the pattern of the bulk magnetization.

Extended Data Fig. 10 Finite-size effects from the chain length.

a, Contrast c(t) obtained after a weighted average over all different chain lengths between L = 0 and 44a (shown in b), for Δ = 0 and λ = 10.4a. The averaged dynamics (orange, yellow, blue) shows almost no dependence on the phase θ, in contrast to the dynamics determined from a single chain length L = 40a (Extended Data Fig. 9a). Also overlaid are the contrasts for a fixed chain length (L = 40a) averaged over all initial phases 0 ≤ θ < 2π (black solid line), and averaged over only the two phases θ = 0 and π/2 (black dashed line). The close agreement implies that averaging over either chain lengths or phases suppresses the dependence on initial or boundary conditions. b, A cut through the spherical Mott insulator with diameter Lmax = 44a (as in the experiment) illustrates the distribution of different chain lengths (oriented along the z direction). Averaging the local magnetization Sz over the x and y directions provides a 1D magnetization profile (bottom panel), which is an average over all chains. c, The number of chains with length L is given by (π/2)(L/a). The total number of chains is πLmax/(2a)2 ≈ 1,500. d, The number of atoms in chains with length L is given by (π/2)(L/a)2. The contribution of each chain to the imaging signal is proportional to the atom number in the chain, and so the relevant average over chain lengths is weighted by the atom number and is L = (3/4)Lmax = 33a.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Jepsen, P.N., Amato-Grill, J., Dimitrova, I. et al. Spin transport in a tunable Heisenberg model realized with ultracold atoms. Nature 588, 403–407 (2020).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:

This article is cited by


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.


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing