A Quantum Coherence Microscope in the Hubbard Regime

AI summary

Overview

Research area: Quantum simulation with ultracold atoms in optical lattices — specifically, the development of a new measurement technique ("quantum coherence microscope") that combines Talbot-effect matter-wave evolution with single-atom-resolved imaging to map off-diagonal correlations onto density signals.

Technical level: Intermediate. The core idea is conceptually simple (let the wavefunction evolve, then recapture and look at density modulations), but full engagement with the paper requires familiarity with the Bose-Hubbard model, the single-particle density matrix, and quantum gas microscopy.

Scope (one sentence): The paper introduces and demonstrates a spatially resolved probe of off-diagonal coherence in a Bose-Hubbard simulator of erbium atoms, applying it to the superfluid–Mott insulator transition in a harmonically trapped cloud and to next-nearest-neighbor coherence in an engineered two-leg ladder.

What This Paper Is About

Quantum gas microscopes routinely deliver site-resolved snapshots of density, but off-diagonal coherence — the phase relationship between atoms on different lattice sites — has remained accessible only through time-of-flight (ToF) imaging, which averages over the whole cloud and washes out spatial structure. The authors solve this by letting atoms freely evolve for short, controlled times (at multiples of the Talbot time), recapturing them in a phase-tunable lattice, and reading out how the recaptured filling oscillates as the recapture phase is scanned. This converts coherence between sites separated by distance r into a measurable density modulation with a spatial resolution of roughly r lattice sites.

Key Contributions

1. A quantum coherence microscope with near-single-site resolution. By combining short Talbot evolution with single-atom-resolved imaging, the authors measure local off-diagonal single-particle density matrix elements, rather than a cloud-averaged coherence signal.
2. Separation-tunable coherence probes. Choosing the free-evolution time selects which lattice separation r = |i − j| is probed; r = 1 gives single-site spatial resolution, while larger r trades spatial resolution for access to longer-range coherence.
3. A spatially resolved measurement of the superfluid–Mott transition in a single cloud. The same harmonically trapped sample is shown to contain coexisting regions with sharply different coherence behavior versus U/t.
4. Isolation of coherence beyond nearest neighbors in an engineered region. Using a low-entropy sample region separated from a reservoir, the authors measure next-nearest-neighbor (r = 2) coherence in a two-leg ladder and use it to distinguish superfluid from insulating regimes.

Main Findings

• Coherence maps onto recapture filling. For coherent initial states, the recaptured filling oscillates sinusoidally with the recapture lattice phase, with amplitude A, offset O, and relative modulation R = |A|/O. Incoherent states show no such revivals or modulations.

• A single global calibration suffices. The measured modulation is related to the normalized coherence by R = η g⁽¹⁾ᵢ,ⱼ, with a global value η = 0.47. Within uncertainties, η showed no statistically significant dependence on probed separation r, ToF geometry (2D versus 1D), initial lattice depth, filling, interaction strength, or s-wave scattering length.

• The center of the trapped cloud loses coherence. In the central region (roughly unit filling), the recapture modulation amplitude decreased substantially as U/t was increased from 10 to 30, consistent with the superfluid-to-Mott-insulator transition.

• The edge of the cloud stays superfluid. In the edge region (roughly one-third filling), the modulation amplitude remained nearly unchanged up to U/t ≈ 40, indicating retained superfluid coherence. The recapture filling there is lower than in the center because the initial density is lower.

• Theory matches with one scaling parameter. DMRG calculations (maximum bond dimension 600) reproduced both center and edge data using the same global scaling parameter; images in the density panel were averaged over 77 shots.

• Ladder coherence depends strongly on interaction regime. In the two-leg ladder (~12 × 2 sites) with V ≈ 3t, the next-nearest-neighbor modulation grew immediately with interchain tunneling t⊥ at U/t ≈ 3 (superfluid regime), while at U/t ≈ 12 the modulation stayed weak across the explored range, consistent with an insulating tendency. TEBD simulations including number fluctuations, scattering-length uncertainty (±3 Hz in U), and non-adiabatic ramps reproduced the data with the same η.

• The efficiency deficit is explained qualitatively. Ideal conditions would give vanishing recapture filling; the authors attribute η < 1 to finite lattice rise/fall times (~1 μs fall for the initial lattice, ~2 μs rise for the recapture lattice, estimated 5% energy shift), shot-to-shot phase fluctuations of the accordion lattice (standard deviation 0.065 waves, estimated 4% shift), and a lattice-depth mismatch between the 20–30 recoil-energy initial lattice and the 25 recoil-energy recapture lattice (estimated 1% shift) — together giving an estimated average final energy of 91% of the final lattice depth.

Methodology in Plain English

The experiment uses up to a few hundred erbium-164 atoms held in a single layer of a three-dimensional optical lattice, described by an extended Bose-Hubbard Hamiltonian with tunneling t, on-site interaction U, a spatially varying chemical potential μᵢ, and long-range dipolar interactions Vᵢ,ⱼ.

Conventional coherence measurements switch the lattice off and let atoms expand for about 10 ms, producing interference peaks whose sharpness reflects coherence averaged over the entire cloud. This work instead performs a "short" expansion: with the vertical lattice still on, atoms fly freely for integer multiples of half the Talbot time (T_T/2 = 29 μs, from a 266 nm lattice spacing and the erbium mass; T_T ≈ 58 μs), and are then recaptured in a phase-tunable "accordion" lattice with the original spacing. Scanning the recapture lattice phase produces a sinusoidal modulation in the recaptured filling whose amplitude reports the coherence between sites at a chosen separation. Because the whole procedure ends in a site-resolved image, the coherence signal is obtained region by region rather than as one cloud-wide number.

For the transition study, U is tuned with a magnetic Feshbach resonance (scattering lengths 40 to 80 a₀), U is kept larger than the nearest-neighbor dipolar repulsion V, and the harmonic confinement (k/t ≈ 0.17) naturally provides regions of different filling in one sample. For the ladder study, tunneling is t = h × 6 Hz (transition study) or h × 8 Hz (ladder), the dipole moment is fixed at 7 μ_B with the field perpendicular to the 2D plane, and nearest-neighbor repulsion is roughly h × 23 Hz, decaying as a/r³ + b/r⁴ + c/r⁵ + d/r⁶ with (a, b, c, d) = h × (33.63, 0.13, −19.69, 9.08) Hz. A low-entropy sample region is defined next to a reservoir using repulsive walls, and coherence is measured only inside the sample.

Why This Matters

Impact on research. The work adds off-diagonal, spatially resolved observables to the standard toolbox of quantum gas microscopy, which has so far been dominated by density-basis measurements. Because the measurement is local, it can distinguish coexisting phases and domain walls inside a single inhomogeneous cloud — precisely the situations where cloud-averaged ToF signals can mischaracterize a state. The authors frame the result more broadly as a template: coherent manipulation before projective imaging extends what a quantum gas microscope can measure.

Real-world applications:

• Benchmarking quantum simulators and annealers: a way to verify that a simulated many-body state actually possesses the coherence its programmer intended.
• Materials-relevant physics: distinguishing superfluid, insulating, and modulated phases, and measuring the spatial decay of correlations, is directly relevant to models of strongly correlated electrons and dipolar quantum matter.
• Quantum sensing and metrology: coherence is the resource behind interferometric sensing, and a site-resolved coherence probe offers a local diagnostic of decoherence in atom-based devices.
• Validation of numerical methods: the reported agreement with DMRG and TEBD at a single global scaling parameter gives theory groups an experimental target for finite-temperature, finite-ramp-rate simulations.

Industry relevance. The technique is most directly relevant to cold-atom quantum simulation and computing platforms, where verifying coherence and entanglement structure across a lattice is a bottleneck. The authors' competing-interests statement notes that co-author M. Greiner is a co-founder, shareholder, and consultant of QuEra Computing.

Future Directions

• Probing coherence at even larger separations. The authors state the technique can be extended to larger r (SM), enabling measurements of the spatial decay of off-diagonal correlations, quasi-long-range order, and power-law correlations in strongly interacting matter.
• Fermionic systems. The protocol is proposed as extendable to fermions, where programmable pre-imaging evolution has been proposed to reveal momentum distributions, Fermi surface structure, and signatures of unconventional superfluidity.
• Quantitative modeling of the recapture contrast. A first-principles prediction of η would require modeling the full interacting recapture dynamics — including dipolar interactions at short separations — which the authors explicitly identify as beyond the current scope.
• Broader classes of pre-imaging coherent transformations. The conclusion frames this as one instance of a family of protocols in which tailored matter-wave evolution before imaging expands the measurement basis of atomic and molecular quantum systems.

Target Audience

Experimentalists and theorists in ultracold atoms and quantum simulation, particularly groups running quantum gas microscopes and those studying the Bose-Hubbard and extended Hubbard models. It is also valuable for researchers in quantum information and metrology interested in measuring coherence in many-body systems, and for numerical physicists (DMRG/TEBD) seeking experimental benchmarks. Readers without a background in optical lattices will find the conceptual idea accessible but the details demanding.

Authors’ abstract

Quantum coherence underlies collective quantum phenomena and emerging quantum technologies. Quantum gas microscopes have transformed quantum simulation by providing projective snapshots of many-body states with single-atom resolution, but spatially resolved measurements of off-diagonal correlations have remained elusive. Here, using the Talbot effect, we introduce a quantum coherence microscope that maps off-diagonal correlations onto site-resolved density signals with near-single-site resolution. We use this technique to locally probe the superfluid-Mott transition in a layer of a three-dimensional optical lattice and to measure coherence beyond nearest neighbors in an engineered potential landscape. By mapping off-diagonal correlations onto density signals through controlled Talbot evolution, this work opens new possibilities for accessing observables beyond the density basis through tailored matter-wave evolution and recapture.