arXiv:2610.28621
Characterizing supersolid phase transitions in trapped quantum gases with tilted dipoles at finite temperature
Julian Kusch, Lauriane Chomaz, Wyatt Kirkby
Abstract
Motivated by a recent experimental work arXiv:2608.20327, we theoretically investigate the nature of phase transitions in a trapped dipolar quantum gas upon tuning the contact interaction strength and dipole orientation. Using extended Gross-Pitaevskii simulations including thermal and quantum effects, we map the phase diagram comprising unmodulated superfluids, triangular droplet arrays, and stripe states. We further characterize the order of the transitions separating them through imaginary-time hysteresis analysis. The unmodulated-to-modulated transition evolves from discontinuous to continuous with increasing dipole tilt before becoming discontinuous again at large tilts. The degree of continuity is further affected by particle number, temperature, and confinement. While continuous behavior generally correlates with the emergence of stripe states, this correspondence is not systematic, and finite-size effects also influence the continuity of the transition. The triangular-to-stripe transition is found to be discontinuous at small contact interaction strengths and becomes increasingly continuous upon approaching the unmodulated-to-modulated phase boundary. Our results reveal a tunable transition order in finite dipolar quantum gases and provide guidance for its experimental investigation.
AI Summary
How do dipole tilt, temperature, atom number, and confinement affect transitions between unmodulated superfluids, triangular droplet arrays, and stripes in trapped dipolar gases? This paper maps these regimes using a finite-temperature extended Gross–Pitaevskii equation with local quantum and thermal corrections. The reference dysprosium system has $N=2\times10^5$, $T=60\,\mathrm{nK}$, and trap frequencies $(40,20,150)\,\mathrm{Hz}$. Transition character is inferred from bidirectional imaginary-time parameter sweeps: density-contrast hysteresis measures metastability across the unmodulated-to-modulated boundary, while Fourier-peak ratios distinguish triangular and stripe structures. The unmodulated-to-modulated transition is discontinuous at small tilt, becomes continuous at intermediate tilt, and develops hysteresis again for $\alpha\gtrsim60^\circ$, where finite-system stripe rearrangements produce multistage transitions. Increasing temperature or atom number, or weakening vertical confinement, narrows the bistability interval and lowers the tilt needed for continuous behavior. At $\nu_z=100\,\mathrm{Hz}$, the transition appears continuous even without tilt. Continuous behavior often accompanies stripes, but not systematically. Triangular-to-stripe hysteresis spans a few degrees at low contact interaction strength and nearly disappears near the unmodulated boundary. Sweep resolutions are approximately $0.12a_0$ and $2^\circ$. The thermal correction is phenomenologically extrapolated from the mean-field-stable regime into the unstable regime. Finite size, trap inhomogeneity, competing metastable arrangements, and ambiguous structural classification limit transition identification, particularly near phase boundaries. Thermodynamic-limit connections and experimentally realistic real-time hysteresis, including stochastic and thermal dynamics, remain unresolved.
Central claim may not hold: The central identification of continuous transitions relies on absence of detected imaginary-time hysteresis at approximately 0.12 a_0 resolution, which cannot exclude narrower bistability intervals or metastable branches missed by continuation. The finite-temperature predictions use a thermal potential extrapolated from the mean-field-stable regime into the unstable regime without independent validation there, making those predictions conditional on the extrapolation.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 12 pages, 10 figures