Ultracold Atom Quantum Simulation

Status: bosonic and fermionic Hubbard-model realizations, single-site quantum-gas microscopy, antiferromagnetic correlations, artificial magnetic fields, topological single-particle bands, one-dimensional Abelian gauge-theory mappings, correlation light cones, and local thermalization through entanglement are established. Low-temperature doped Hubbard physics, interacting topological matter, higher-dimensional gauge theories, strongly interacting synthetic dimensions, and controlled exceptions to thermalization are active. A settled phase diagram of the repulsive doped two-dimensional Hubbard model, an experimentally scalable non-Abelian lattice gauge theory, and an end-to-end analog quantum advantage are not established.

Last reviewed: 26 July 2026. Temperatures, system sizes, model mappings, and publication status are date-sensitive. “Classically difficult” and “classically impossible to validate” are not interchangeable claims.

Core Question

When does a controllable atomic system answer a question about another quantum model, rather than merely display complicated dynamics of its own?

What Is Quantum Simulation? owns the general digital–analog–hybrid taxonomy and simulation contract. This page applies that contract to the date-sensitive evidence for ultracold-atom platforms.

Analog Quantum Simulation owns the platform-independent Hamiltonian correspondence, effective-model reduction, observable-error, resource, and validation workflow used here.

An ultracold-atom simulator has a compelling starting point: its microscopic constituents are well understood, its geometry and interactions are tunable, and many particles can evolve coherently. The scientific conclusion still depends on a chain:

$$\begin{aligned} \text{laboratory controls} &\longrightarrow \text{effective Hamiltonian} \\ &\longrightarrow \text{prepared state} \\ &\longrightarrow \text{measurement channel} \\ &\longrightarrow \text{observable with uncertainty} \\ &\longrightarrow \text{claim about the target model}. \end{aligned}$$

Every arrow can fail independently. A precisely calibrated lattice can hold a state that is too hot. A very cold state can be measured through a parity-only detector that hides doublons. A measured correlator can agree with one numerical method while remaining compatible with several effective Hamiltonians.

The frontier is therefore not control alone. It is validated inference in a regime where useful classical calculations become incomplete.

Why It Matters

Quantum matter can be assembled from known ingredients

Optical lattices supply periodic potentials without chemical disorder. Contact interactions, magnetic Feshbach resonances, lattice geometry, confinement, spin state, and artificial gauge fields can be varied independently over useful ranges. This permits controlled paths through model parameter space rather than comparison among different materials.

That cleanliness is not literal perfection. Beam profiles create inhomogeneity, spontaneous scattering heats the gas, higher bands modify the single-band model, and preparation ramps can freeze nonequilibrium defects. The advantage is that these effects can often be measured and changed systematically.

Quantum-gas microscopes expose the microscopic state

Site-resolved imaging turns an atomic lattice into a source of many-body snapshots. From repeated shots one can estimate:

• density and spin distributions;
• two-point and higher-order correlations;
• string and parity observables;
• defect-conditioned correlators;
• currents after basis rotations;
• full counting statistics in selected regions; and
• Rényi entropies using interference between copies.

These observables are closer to the model variables than most probes of solids. They are not automatically unbiased. Imaging loss, hopping during detection, light-assisted collisions, imperfect spin removal, and finite optical resolution define a measurement channel that must be calibrated.

The platform joins equilibrium and dynamical questions

A lattice can be loaded slowly to target equilibrium or quenched rapidly to study transport, correlation spreading, prethermalization, and thermalization. The same apparatus can compare a state-preparation ramp with its reverse, prepare local defects, or measure the response to a calibrated force.

This dual use is important because many strongly correlated states are hard to cool into but have distinctive dynamics. It is also dangerous: a long-lived transient is not necessarily an equilibrium phase.

Atomic simulators connect condensed matter and gauge theory

Hubbard and spin models reproduce central structures of correlated-electron physics. Superlattices and constrained occupation sectors can also encode matter fields, electric fields, and Gauss-law constraints. Synthetic dimensions and Floquet phases create magnetic fluxes and topological bands for neutral particles.

The common language is Hamiltonian engineering. The standards of evidence differ. A gauge-theory simulator must measure gauge violation; a topological simulator must connect observables to a bulk invariant or boundary response; a Hubbard simulator must establish filling, temperature, and interaction scale.

Quantum simulation can become a two-way benchmark

In a classically controlled regime, numerical calculations validate the experiment. At lower temperature, stronger frustration, higher dimension, or longer time, experimental data can instead expose where approximate numerical methods disagree.

This is not a declaration that the device has replaced computation. Experiments and numerics can form a loop:

$$\text{calibrate} \longrightarrow \text{predict} \longrightarrow \text{measure} \longrightarrow \text{diagnose discrepancy} \longrightarrow \text{refine}.$$

The strongest frontier result improves both the physical understanding and the methods used to test it.

Canonical Boundary

This page evaluates current capabilities and open research questions. Stable derivations live elsewhere.

• Ultracold Atoms owns quantum degeneracy, low-energy collisions, Feshbach tuning, dimensionality, equilibration, and loss.
• Optical Lattices owns recoil scales, Bloch bands, Wannier functions, tunnelling, Hubbard parameters, lattice calibration, heating, and model validation.
• Hubbard Model owns the fermionic Hamiltonian, symmetries, exact limits, Mott physics, and model diagnostics.
• Bose–Hubbard Model owns bosonic hopping, onsite interactions, number fluctuations, and the superfluid–Mott transition.
• Chern Numbers owns the geometric invariant behind the topological-band discussion.
• Eigenstate Thermalization Hypothesis owns the formal ETH statement and its scope.

Here the canonical content is the evidence ledger: which mappings have been implemented, which observables support each interpretation, where classical benchmarks end, and what remains uncertain as of the review date.

Minimal Background

Analog simulation is a controlled model correspondence

An analog simulator implements a Hamiltonian $H_{\mathrm{lab}}$ whose low-energy or constrained dynamics approximate a target $H_{\mathrm{tar}}$:

$$P H_{\mathrm{lab}} P = \alpha H_{\mathrm{tar}} +\delta H,$$

where $P$ projects onto the intended subspace, $\alpha$ fixes units, and $\delta H$ contains residual terms. A useful model claim bounds both leakage outside $P$ and the effect of $\delta H$ on the reported observable.

The relevant error is not always an operator norm. A small term can be important near a phase boundary or over a long evolution time. Conversely, some large microscopic corrections may only renormalize a fitted parameter. Validation should be observable and regime specific.

The Hubbard simulator has a short parameter list and a long error budget

A common two-component Fermi–Hubbard realization is

$$\begin{aligned} H =& -\sum_{\langle i,j\rangle,\sigma} t_{ij} \left( c_{i\sigma}^{\dagger}c_{j\sigma} +c_{j\sigma}^{\dagger}c_{i\sigma} \right) \\ &+ U\sum_i n_{i\uparrow}n_{i\downarrow} + \sum_{i,\sigma}V_i n_{i\sigma}. \end{aligned}$$

The model looks simple because many laboratory details have been compressed into $t_{ij}$, $U$, and $V_i$. A credible realization also bounds:

$$\left( \frac{t'}{t}, \frac{t_{\mathrm{anis}}}{t}, \frac{U_{\mathrm{nn}}}{U}, p_{\mathrm{band}>0}, \Gamma_{\mathrm{heat}}, \Gamma_{\mathrm{loss}} \right).$$

At large repulsive $U/t$ and near half filling, virtual hopping produces the superexchange scale

$$J=\frac{4t^2}{U}.$$

Cooling below $t$ does not guarantee magnetic order; the experiment often needs $k_{\mathrm B}T$ comparable to or below $J$. Doped phases can require still lower scales.

Report dimensionless coordinates of the state

Temperature in kelvin is rarely the useful headline. A state in a lattice should be located by quantities such as

$$\mathbf x = \left( \frac{U}{t}, \frac{k_{\mathrm B}T}{t}, n, \delta, \frac{t'}{t}, \frac{V_{\mathrm{trap}}}{t}, \frac{\Gamma_{\mathrm{heat}}\hbar}{t} \right),$$

where $n$ is filling and $\delta=1-n$ is hole doping for a one-particle half-filled convention. If the effective model is Heisenberg-like, $T/J$ may be more informative than $T/t$.

Uncertainties and spatial distributions matter. A harmonically trapped cloud can contain a half-filled center, doped shells, and dilute edges in the same shot. Box potentials and digital micromirror devices reduce this inhomogeneity but do not remove calibration.

Thermometry is an inference problem

There is no universal thermometer that can be inserted into an isolated lattice gas without changing it. Experiments infer temperature by comparing one or more observables with:

• exact results in a limiting regime;
• determinant quantum Monte Carlo where sign problems are absent;
• auxiliary-field methods with stated constraints;
• numerical linked-cluster expansions;
• quantum Monte Carlo for effective spin models;
• fluctuation–dissipation relations; or
• an independently calibrated reservoir.

In schematic form,

$$p(T,\boldsymbol\theta\mid D) \propto p(D\mid T,\boldsymbol\theta) p(T,\boldsymbol\theta),$$

where $D$ is the measured data and $\boldsymbol\theta$ includes Hamiltonian, trap, and detector nuisance parameters. A quoted temperature inherits the model assumptions in the likelihood.

Far from equilibrium, a fitted “effective temperature” may summarize one observable without defining a Gibbs state. Different observables should agree before the equilibrium label is trusted.

A microscope measures through a channel

For a binary site occupancy, let $\alpha$ be the false-negative probability and $\beta$ the false-positive probability. If the true occupation probability is $p$, then

$$p_{\mathrm{obs}} =(1-\alpha)p +\beta(1-p).$$

Correlators require the multi-site version of this channel, including correlated loss and parity projection. Correcting each mean independently can misestimate connected correlations.

Spin-resolved measurements add state-removal or state-transfer errors. Doublon-sensitive measurements must distinguish genuine pair occupation from light-assisted loss. A frontier result should publish the calibration matrix or a forward model sufficient to reproduce the inference.

Observables form a hierarchy of claim strength

A simulator can measure increasingly discriminating quantities:

$$\begin{aligned} \text{density} &\longrightarrow \text{two-point correlations} \longrightarrow \text{higher-order and conditional correlations} \\ &\longrightarrow \text{response and currents} \longrightarrow \text{entanglement-sensitive observables}. \end{aligned}$$

One observable rarely identifies a phase uniquely. For example, suppressed double occupancy supports Mott physics but does not establish antiferromagnetic order. A staggered spin peak supports antiferromagnetism but must be distinguished from finite-size and trap effects. Stripe precursors need joint charge and spin evidence.

Gauge simulation requires a local constraint

For a target lattice gauge theory, physical states satisfy a local Gauss law

$$G_j|\psi\rangle =g_j|\psi\rangle$$

at every site $j$. A useful gauge-violation diagnostic is

$$\epsilon_G = \frac{1}{L} \sum_{j=1}^{L} \left\langle \left( G_j-g_j \right)^2 \right\rangle.$$

A conserved global charge is not enough: positive and negative local violations can cancel. Experiments should measure local constraint fulfilment, its time dependence, and sensitivity to protection strength.

Current cold-atom gauge simulations use finite-dimensional quantum-link representations or constrained Bose–Hubbard sectors. They reproduce selected gauge-theory dynamics; they do not yet reproduce the full continuum Standard Model.

A synthetic dimension promotes an internal label to a coordinate

Let $m$ label nuclear-spin, hyperfine, momentum, vibrational, or Rydberg states. Coherent transitions can act as hopping along a synthetic direction:

$$H_{\mathrm{syn}} = -\sum_{x,m} \left[ t_x c_{x+1,m}^{\dagger}c_{x,m} + \Omega_m e^{i\phi x} c_{x,m+1}^{\dagger}c_{x,m} +\mathrm{h.c.} \right].$$

The position-dependent phase produces flux through a real–synthetic plaquette. Internal-state populations can then resolve synthetic edges directly.

Synthetic dimensions provide precise couplings and unusual boundary conditions. They also create atypical interactions: particles at different synthetic sites may occupy the same real-space location and interact with a range or symmetry unlike a local higher-dimensional material.

Topology needs an operational diagnostic

For an isolated filled band, the Chern number is

$$C = \frac{1}{2\pi} \int_{\mathrm{BZ}} \Omega(\mathbf q)\,d^2q.$$

Cold-atom experiments have inferred topology through transverse drift, Berry-curvature mapping, circular dichroism, pumping, edge motion, and Hall response. Each probe has assumptions about band occupation, adiabaticity, interactions, and confinement.

In an interacting few-body system, agreement with a Laughlin-state overlap, gap, correlations, and fractional response can establish a controlled minimal realization. It does not by itself establish a thermodynamic topological phase with macroscopic anyonic excitations.

Nonequilibrium evolution separates several notions of relaxation

After a quench,

$$|\psi(t)\rangle =e^{-iHt/\hbar}|\psi(0)\rangle.$$

Several distinct behaviors can follow:

• dephasing: observables settle because many phases cancel;
• prethermalization: a long-lived quasistationary state appears before final heating or relaxation;
• thermalization: local observables approach an ensemble fixed by conserved quantities;
• integrable relaxation: a generalized ensemble retains extra charges;
• localization or fragmentation: memory persists because dynamics is constrained; and
• open-system relaxation: an external bath, loss, or noise sets the stationary state.

These labels require tests. A density imbalance that remains nonzero for the experimental duration establishes slow memory, not an infinite-time phase.

Local thermal entropy can emerge from global purity

For a pure state on $A\cup B$,

$$\rho_A(t) = \operatorname{Tr}_B |\psi(t)\rangle\langle\psi(t)|,$$

and the subsystem entropy is

$$S_A =-\operatorname{Tr} \left( \rho_A\ln\rho_A \right).$$

The full state can remain pure while $S_A$ grows through entanglement. If $\rho_A$ approaches the reduced state predicted by a thermal ensemble, local statistical mechanics emerges without a fundamental loss of unitarity.

This is the operational content behind many atomic tests of thermalization. Entanglement growth is necessary for generic local equilibration but does not guarantee that every model or initial state obeys ETH.

Validation should have overlapping regimes

A mature simulator is tested in a ladder:

1. one and two particles;
2. small many-body systems with exact diagonalization;
3. sign-problem-free or tensor-network regimes;
4. controlled limiting cases;
5. intermediate regimes where methods overlap; and
6. target regimes where no single classical method is trusted.

The final step inherits credibility from the earlier ones only if the same calibration and measurement pipeline is used. Extrapolating across a phase transition, a change in dimensionality, or a new heating regime needs fresh validation.

What Is Known

Bose–Hubbard quantum simulation is established

Jaksch et al. showed how cold bosons in optical lattices realize the Bose–Hubbard model and its superfluid–Mott transition Physical Review Letters 81, 3108–3111 (1998). Greiner et al. subsequently observed the transition by varying lattice depth Nature 415, 39–44 (2002).

These results established the central analog-simulation method: optical potential depth tunes the ratio $U/t$, while coherence and number fluctuations diagnose distinct regimes. Later microscopes resolved Mott defects and correlations site by site.

The superfluid–Mott transition is now a calibration benchmark, not the end of the frontier. Current questions involve frustrated geometry, artificial flux, long-range interactions, entropy redistribution, dimensional crossover, and dynamics beyond equilibrium.

Fermionic Mott physics and antiferromagnetism are established

Fermionic Mott insulators were observed in optical lattices in 2008. Quantum gas microscopes then brought spin and charge correlations into real space. Mazurenko et al. realized a two-dimensional Fermi–Hubbard antiferromagnet of about 80 sites, with a correlation length reaching the system size and magnetic correlations surviving to roughly $15\%$ hole doping Nature 545, 462–466 (2017).

Established: long-range antiferromagnetic order in a finite, approximately half-filled cold-atom Hubbard system, with site-resolved diagnostics.

Not established by that result: the low-temperature phase diagram at finite doping or superconductivity in the repulsive square-lattice Hubbard model.

Hubbard simulators entered a substantially lower temperature regime

Xu et al. prepared a low-entropy band insulator and transformed it into strongly correlated states by dynamically changing lattice geometry and confinement. At half filling they reported nearly saturated long-range antiferromagnetic order and inferred

$$\frac{k_{\mathrm B}T}{t} = 0.05_{-0.05}^{+0.06}$$

through comparison with numerically controlled calculations Nature 642, 909–915 (2025). Doped samples showed short-range correlations consistent with state-of-the-art constrained-path auxiliary-field calculations.

Established: a several-fold cooling advance and a programmable state-transformation route into a new Hubbard temperature regime.

Active: determining which doped phases occur there. Away from half filling, the comparison methods are approximate and the thermometer is more model dependent. The reported low temperature is an enabling condition, not an observation of a pseudogap, stripe phase, or superconducting state.

Doped carriers can be imaged as dressed many-body objects

Koepsell et al. imaged magnetic polarons: mobile doublons locally weakened and reversed antiferromagnetic correlations in a doped square-lattice Fermi–Hubbard system Nature 572, 358–362 (2019).

Lebrat et al. observed Nagaoka polarons in a triangular-lattice simulator, with ferromagnetic bubbles around particle dopants and frustration-dependent antiferromagnetic response around holes Nature 629, 317–322 (2024).

These are microscopic quasiparticle results. They establish how an isolated or dilute dopant reorganizes its spin background. They do not determine the collective phase formed by a finite density of interacting polarons.

Stripe precursors have been observed in an engineered geometry

Bourgund et al. used a mixed-dimensional Fermi–Hubbard model to raise the energy scale associated with hole attraction. They measured extended hole–hole correlations, higher-order spin correlations, and signatures consistent with individual fluctuating stripes Nature 637, 57–62 (2025).

The careful interpretation is stripe precursors in a mixed-dimensional finite system. The experiment did not establish long-range stripe order in the isotropic two-dimensional square-lattice Hubbard model, nor settle whether stripes compete or coexist with superconductivity.

Artificial magnetic fields and topological bands are established

Jotzu et al. implemented the Haldane model with ultracold fermions in a periodically driven honeycomb lattice, mapped the topological transition, and measured anomalous transverse drift Nature 515, 237–240 (2014).

Aidelsburger et al. measured the Chern number of Hofstadter bands with ultracold bosons through the transverse response to a force Nature Physics 11, 162–166 (2015).

These experiments established band-topology engineering and response-based diagnostics for neutral atoms. The harder frontier is strongly interacting topological matter, where a single-particle band invariant is insufficient.

A minimal fractional quantum Hall state has been realized

Léonard et al. prepared a lattice analogue of the bosonic $\nu=1/2$ Laughlin state using two particles on 16 sites. They measured a many-body gap, suppressed double occupation, vortex-like density correlations, and a fractional Hall response

$$\frac{\sigma_H}{\sigma_0} =0.6(2),$$

consistent with the expected value $1/2$ Nature 619, 495–499 (2023).

Established: controlled preparation and multiple diagnostics of a minimal interacting fractional-Hall state.

Not established: a large thermodynamic fractional quantum Hall fluid, direct braiding of its quasiparticles, or topological protection at scale.

Interacting flux ladders now support local current measurements

Impertro et al. realized 48-site bosonic flux ladders at half filling with a synthetic magnetic flux and strong interactions. Quantum-gas microscopy combined with local basis rotations resolved bond currents and supported an interacting Mott–Meissner phase Nature Physics 21, 895–901 (2025).

The experiment benchmarked density correlations against numerical calculations and inferred an effective temperature on the order of the tunnelling scale. It established a large interacting chiral ladder, not a two-dimensional fractional Chern insulator.

Synthetic dimensions support chiral edges and Hall response

Mancini et al. used nuclear-spin states as sites along a synthetic dimension and observed chiral edge motion in a fermionic Hall ribbon Science 349, 1510–1513 (2015).

Zhou et al. later built two- and three-leg synthetic Hall bars from $^{173}\mathrm{Yb}$ nuclear-spin states and directly measured a compensating Hall voltage and reactive Hall resistance Nature Communications 16, 10247 (2025).

These results establish that internal states can act as experimentally addressable spatial legs with engineered flux. Extending to many synthetic sites while retaining controllable interactions and low heating remains active.

One-dimensional Abelian gauge constraints are experimentally established

Yang et al. encoded an extended $U(1)$ quantum-link model in a 71-site Bose–Hubbard superlattice containing matter and gauge-field degrees of freedom. They measured local gauge-invariant configurations, tuned a quantum phase transition, and quantified deviations from Gauss’s law Nature 587, 392–396 (2020).

Earlier building-block experiments had implemented local $U(1)$ or $\mathbb Z_2$ structures. The 71-site result established a many-body, locally diagnosed Abelian gauge mapping.

As summarized in a 2025 review Nature Physics 21, 25–36 (2025), most cold-atom experiments remain in $(1+1)$ dimensions and use Abelian, truncated gauge fields. A proposed $(2+1)$D $U(1)$ simulator in a two-dimensional Lieb superlattice Communications Physics 8, 273 (2025) is a theoretical proposal, not an experimental realization.

Correlation light cones have been observed

Cheneau et al. quenched a one-dimensional Bose–Hubbard system and measured the spreading of parity correlations. The resulting front propagated at a finite velocity consistent with a quasiparticle picture and an effective Lieb–Robinson light cone Nature 481, 484–487 (2012).

This established time-resolved many-body correlation transport. It did not measure a relativistic causal speed; the velocity depends on the lattice Hamiltonian and quench.

Local thermalization through entanglement is established

Kaufman et al. evolved an isolated Bose–Hubbard system from a pure product state and measured subsystem purity and entanglement entropy. Local subsystems approached thermal predictions while the full measured system remained pure Science 353, 794–800 (2016).

The experiment gave direct operational evidence for entanglement as the source of local thermodynamic entropy in a closed quantum system. It supports ETH for the tested nonintegrable setting and initial states; it is not a proof that all isolated many-body systems thermalize.

Long-lived localization has been observed in one-dimensional quasiperiodic systems

Schreiber et al. prepared an interacting fermionic charge-density wave in a one-dimensional quasirandom lattice and observed persistent imbalance above a disorder threshold Science 349, 842–845 (2015).

The result established interaction-dependent, many-body-localization-like dynamics over experimental times in a one-dimensional quasiperiodic system. The infinite-time stability of many-body localization in higher dimensions, and the role of rare thermal regions, remain debated. Loss, residual coupling, finite size, and finite observation time must be separated from intrinsic delocalization.

The evidence is observable specific

Research claim	Strong supporting observables	Common blind spot
Hubbard-model realization	band spectroscopy, tunnelling, doublon energy, density profile	unmeasured longer-range terms
low temperature	several thermometers in an overlap regime	model-dependent extrapolation into doped sign-problem regime
antiferromagnetism	structure factor, correlation length, staggered magnetization	finite size and spatially varying filling
stripe precursor	joint charge, spin, and higher-order correlations	mistaking local strings for long-range order
band topology	transverse drift, Berry curvature, pumping, edge response	nonuniform band occupation
interacting topology	gap, response, nonlocal correlations, size scaling	relying on overlap or one marker alone
gauge theory	local Gauss-law compliance and target dynamics	checking only global charge
thermalization	observable agreement, subsystem state, entropy growth	confusing dephasing or prethermal plateaus with Gibbs equilibrium
localization	memory, transport, entanglement growth, size and time scaling	finite observation window

What Is Actively Developing

Doped Hubbard matter below the exchange scale

The central condensed-matter target is the low-temperature, finite-doping regime of the repulsive two-dimensional Hubbard model. Candidate phenomena include:

• a pseudogap with momentum-selective spectral suppression;
• fluctuating or ordered charge and spin stripes;
• magnetic and geometric-string polarons;
• unconventional pairing correlations;
• competition among stripes, pairing, and charge order; and
• dependence on next-nearest-neighbor hopping and lattice geometry.

The experimental difficulty is not only cooling. Doping changes entropy redistribution, compressibility, transport during loading, and the available classical thermometers. A state can be cold in the central region while remaining spatially nonuniform or out of equilibrium.

The decisive evidence for a pairing phase would combine a pairing susceptibility, long-distance correlations, response to a phase twist or pair field, finite-size scaling, and temperature dependence. Local attractive correlations are not enough.

Thermometry where unbiased numerics are unavailable

At half filling, determinant quantum Monte Carlo and spin-model calculations provide controlled temperature references in useful regimes. Finite doping introduces a sign problem, so constrained-path or tensor-network results carry approximation choices.

Active strategies include:

• preparing the same state through multiple adiabatic paths;
• using an attached reservoir with independently known equation of state;
• comparing several local observables with distinct parameter sensitivity;
• fluctuation–dissipation thermometry;
• entropy accounting from the initial band insulator;
• cross-method numerical consensus with blind predictions; and
• learning thermometers only in regimes with controlled labels, followed by explicit out-of-distribution tests.

A thermometer trained on one Hamiltonian family can silently identify the nearest training example rather than the true temperature. Its uncertainty must include model mismatch.

Higher-dimensional and non-Abelian gauge theories

Moving beyond $(1+1)$D Abelian quantum-link models requires more local degrees of freedom and interactions that implement plaquette or magnetic terms. Higher dimensions also enlarge the gauge-invariant Hilbert space and make state preparation more demanding.

Active targets include:

• $(2+1)$D $U(1)$ dynamics with matter and electric fields;
• $\mathbb Z_N$ and non-Abelian gauge groups;
• string breaking and confinement with controlled continuum extrapolations;
• finite-density regimes that are difficult for classical Monte Carlo;
• topological terms and anomaly-related dynamics; and
• real-time processes inaccessible to Euclidean lattice methods.

An atomic mapping is useful even when it uses a truncated link space, but the truncation must be varied or theoretically controlled. A finite-dimensional quantum link is not automatically close to a continuum gauge field.

Gauge protection without freezing the target dynamics

Energy penalties, emergent symmetries, Floquet selection rules, and angular-momentum conservation can suppress gauge-violating transitions. If the protection scale is too small, the state leaves the physical sector. If it is too large, target couplings can become slow and vulnerable to heating.

A useful optimization balances

$$\epsilon_G(t) \quad\text{against}\quad \frac{t_{\mathrm{target}}}{t_{\mathrm{coh}}},$$

where $t_{\mathrm{target}}$ is the time needed to observe the desired process and $t_{\mathrm{coh}}$ is the usable coherence time. Reporting only the initial gauge fidelity misses coherent leakage during evolution.

Interacting topological matter at larger scale

Single-particle Chern bands are mature. The frontier is many-body topology: fractional Chern insulators, chiral spin liquids, topological pumps with interactions, and edge dynamics in the presence of disorder and heating.

Larger systems need both low entropy and a preparation path that avoids closing a many-body gap. Floquet engineering adds a second tension: the drive creates complex tunnelling but can heat an interacting gas.

Evidence should combine:

$$\left( \Delta_{\mathrm{mb}}, \sigma_H, \mathcal C_{\mathrm{nonlocal}}, S_{\mathrm{topo}}, \text{edge response}, \text{size scaling} \right),$$

where $\Delta_{\mathrm{mb}}$ is the many-body gap and $\mathcal C_{\mathrm{nonlocal}}$ denotes suitable nonlocal correlations. Not every experiment can measure every entry, but relying on one proxy makes alternative explanations hard to exclude.

Strongly interacting synthetic dimensions

Synthetic dimensions naturally produce flux ladders and direct access to edge populations. The next challenges are:

• more synthetic sites with uniform coherent couplings;
• interactions whose range along the synthetic axis is characterized;
• controlled open, periodic, or twisted boundaries;
• state-dependent loss and dephasing across many internal levels;
• local observables that mix real and synthetic coordinates; and
• fractional or symmetry-protected phases beyond band topology.

The interaction matrix

$$U_{m_1m_2m_3m_4}$$

is part of the model, not a nuisance to be hidden. Internal states that are well separated in the synthetic coordinate can still collide at the same real-space point.

Local spectroscopy of strongly correlated excitations

Microscopes increasingly combine static snapshots with Raman, modulation, and quench spectroscopy. This gives access to momentum-dependent magnons, polarons, spectral weight, current response, and local susceptibilities.

For example, magnon–polaron spectroscopy in a doped Fermi–Hubbard simulator has resolved momentum-dependent shifts and spectral-weight changes Nature Physics 21, 1548–1554 (2025). Such probes move atomic simulators closer to the question set of neutron scattering and angle-resolved spectroscopy while retaining microscopic control.

The inverse problem remains difficult. A broadened spectrum can result from intrinsic lifetime, inhomogeneity, finite pulse duration, detector response, or heating. Resolution functions should be published and forward modelled.

Three-dimensional and magnified microscopy

Conventional fluorescence microscopes are naturally optimized for a two-dimensional plane. Matter-wave magnification and related techniques can expand a lattice distribution before imaging, giving sublattice information for three-dimensional or short-spacing systems.

Active work asks whether these tools can supply:

• number and spin resolution in three dimensions;
• controlled reconstruction of overlapping planes;
• low distortion of high-order correlations;
• compatibility with fermions and species lacking convenient imaging cycles; and
• enough repetition rate for rare-event statistics.

Three-dimensional access is important for finite-temperature phase transitions and models whose geometry cannot be represented faithfully in a single plane.

Nonequilibrium transport and hydrodynamics

Local preparation and imaging permit measurements of charge, spin, energy, and correlation transport. Active questions include diffusion versus superdiffusion, bad-metal transport, bounds on relaxation, operator spreading, and universal coarsening.

The hydrodynamic limit requires scale separation:

$$a \ll \ell_{\mathrm{micro}} \ll \ell_{\mathrm{obs}} \ll L,$$

where $a$ is the lattice spacing, $\ell_{\mathrm{micro}}$ a local relaxation scale, $\ell_{\mathrm{obs}}$ the fitted transport scale, and $L$ the system size. Many microscope experiments operate in the crossover where ballistic, diffusive, and finite-size behavior all compete.

Transport coefficients should be extracted from several initial profiles or response protocols. One density expansion can fit more than one effective model.

Prethermalization, Floquet control, and heating

Periodic drives generate complex hopping, synthetic flux, and effective spin interactions. At high drive frequency, a prethermal effective Hamiltonian can describe the system for a long but finite time.

Floquet–Magnus Expansion organizes the effective model. The frontier question is experimental: whether the useful observation window opens before micromotion, drive-induced loss, and many-body heating corrupt the target state.

A complete report distinguishes:

$$\text{micromotion} \neq \text{slow effective dynamics} \neq \text{ultimate heating}.$$

Stroboscopic agreement over a few periods does not establish a stable Floquet phase.

Controlled failures of thermalization

Integrability, disorder, kinetic constraints, Hilbert-space fragmentation, scars, and gauge symmetries can preserve memory. Atomic simulators can vary the perturbation that breaks each protection mechanism.

The strongest experiment does more than observe a plateau. It measures how the relaxation time scales with perturbation, size, disorder, and conserved quantities. It also distinguishes a prethermal lifetime

$$\tau_{\mathrm{pre}} <\infty$$

from an asymptotically localized phase.

Quantum–classical co-validation

Rather than searching for one classical method that validates the whole experiment, current work combines methods with different failure modes:

• determinant quantum Monte Carlo at half filling;
• constrained-path auxiliary-field methods when doped;
• tensor networks for cylinders or low-entanglement dynamics;
• numerical linked-cluster expansions at sufficiently high temperature;
• exact diagonalization for small clusters;
• effective spin models at large $U/t$; and
• experimentally measured sum rules and symmetries.

Disagreement is information. If several calculations match the calibrated regime but diverge in the target regime, the experiment can discriminate them. If they already disagree with calibration observables, the simulator has not yet reached the intended frontier.

What Is Debated or Uncertain

Has the doped Hubbard problem been “solved” experimentally?

No. The simulator can produce states in regimes where classical methods are challenging and can measure observables that sharpen the debate. A solution of the low-temperature phase diagram would require reproducible state preparation, trustworthy thermometry, multiple order diagnostics, finite-size and boundary analysis, and consistency across parameter paths.

The 2025 cooling advance changes feasibility, not the logical standard. It opens the regime in which pseudogap, stripe, and pairing questions can be asked more sharply.

What does a very low inferred temperature mean?

A posterior estimate such as $T/t=0.05_{-0.05}^{+0.06}$ is strong evidence when based on observables and numerics that overlap reliably. It is not a direct thermometer reading, and its asymmetric interval reaches the physical boundary $T=0$.

At finite doping, sign-problem-free reference calculations are generally unavailable. A temperature estimate can therefore be conditional on an approximate many-body method. Independent thermometers and entropy accounting are especially important.

Does a mapping preserve the physics of interest?

Two Hamiltonians can agree in a restricted subspace while differing in symmetry, locality, excitation spectrum, or continuum limit. The mapping must preserve the observable and process under study.

For gauge theories, a finite link representation may capture string breaking but not weak-coupling photons. For synthetic dimensions, nonlocal interactions along the internal-state axis may create physics unlike a short-range real-space lattice. Such differences can be features if stated explicitly.

How much gauge violation is acceptable?

There is no universal threshold for $\epsilon_G$. A small average violation can be concentrated at boundaries or rare defects that dominate a string observable. A larger violation can be harmless for a short-time local quantity.

The acceptable level depends on sensitivity:

$$\delta O \approx \sum_j \frac{\partial O}{\partial \epsilon_{G,j}} \delta\epsilon_{G,j}.$$

Experiments should vary the protection strength and test whether the inferred physics is stable, rather than selecting one arbitrary cutoff.

When is a finite system topological?

A finite system can have a well-defined many-body Chern response, protected edge structure, or high overlap with a topological target state. It need not display all thermodynamic signatures.

The phrase “realization of a state” can be justified in a minimal system when several diagnostic predictions are verified. Claims of a topological phase, topological order, or protected quasiparticle statistics require stronger scaling and nonlocal evidence.

Is many-body localization stable beyond one dimension?

One-dimensional quasiperiodic experiments provide compelling evidence for long-lived nonergodic dynamics. In higher dimensions, rare thermal regions may destabilize localization at sufficiently long times. Finite systems and finite experimental windows cannot directly resolve that asymptotic limit.

The scientifically useful output is a relaxation-time map and its scaling, not a binary label detached from observation time.

Does local thermalization imply chaos?

ETH connects thermal behavior of observables with the structure of energy eigenstates in nonintegrable systems. Local relaxation can also occur through mechanisms that do not imply random-matrix statistics over the entire spectrum.

A claim of quantum chaos is strengthened by combining:

• level statistics in accessible small systems;
• ETH-like smooth diagonal matrix elements;
• off-diagonal fluctuation scaling;
• operator or entanglement spreading; and
• loss of initial-state memory for multiple states.

One thermal-looking density profile is insufficient.

Can an analog simulator establish quantum advantage?

The simulator may enter a regime where exact classical prediction is impractical. Advantage requires more:

1. a well-defined task and accuracy target;
2. end-to-end resource accounting;
3. comparison with the best relevant classical methods;
4. evidence that the measured answer is correct enough;
5. favorable scaling beyond a finite-size crossover; and
6. reproducible data and calibration.

Analog sampling can be scientifically valuable before any advantage threshold. Avoiding the word “advantage” does not diminish the physics.

How directly do Hubbard results transfer to cuprates?

The single-band Hubbard model captures important competition among kinetic energy, onsite repulsion, magnetism, and doping. Real cuprates also contain material-specific orbitals, longer-range hopping and interactions, electron–phonon coupling, disorder, and three-dimensional structure.

Atomic results can settle questions about the model. Their relevance to a material is a second inference that needs a model of the material. An atomic temperature converted into kelvin by matching one energy scale is an analogy, not a literal material temperature.

Can machine learning validate a simulator?

Machine learning can classify snapshots, reconstruct images, infer parameters, and identify anomalous patterns. It can also learn artifacts, simulation priors, or the geometry of the training set.

Trustworthy use requires:

• physically interpretable baselines;
• held-out parameter regions;
• synthetic measurement-channel stress tests;
• calibration uncertainty in the training labels;
• comparison with simple sufficient statistics; and
• release of code and weights where possible.

A classifier confidence is not a phase probability unless the probabilistic model supports that interpretation.

Key Experimental Platforms

Platform	Principal control	Principal observable	Natural frontier	Main limitation
bosonic optical lattices	$U/t$, superlattices, quenches, flux	parity, density, currents, interference	gauge mappings, Floquet ladders, nonequilibrium dynamics	parity projection and drive heating
fermionic quantum-gas microscopes	spin, doping, geometry, local potentials	spin- and density-resolved snapshots	doped Hubbard matter, polarons, transport	entropy and model-dependent thermometry
alkaline-earth-like gases	many nuclear-spin states and narrow transitions	spin populations, momentum, spectroscopy	synthetic dimensions, SU($N$) matter, Hall response	multilevel calibration and state-dependent loss
box-trapped bulk gases	homogeneous density and large atom number	equation of state, momentum, response	thermodynamics and transport	less direct microscopic access
programmable superlattices	site offsets and bond-dependent tunnelling	occupation patterns and local correlations	gauge constraints, dimers, state transformation	calibration complexity and slower effective couplings
Floquet-engineered lattices	complex tunnelling and artificial flux	drift, current, band population	topological bands and chiral matter	micromotion and many-body heating
matter-wave magnifiers	expansion before imaging	enlarged density and correlations	three-dimensional and short-spacing systems	reconstruction distortion and spin resolution

Different platforms can realize the same nominal Hamiltonian with different boundary conditions, detector channels, and preparation paths. Cross-platform agreement is therefore a powerful validation tool.

Key Theoretical Tools

Wannier reduction and multiband calculations

The first calculation maps laser fields and scattering parameters to $t_{ij}$, $U$, density-assisted tunnelling, longer-range interactions, and higher-band coupling. Wannier functions provide the standard basis, while few-body calculations are needed near strong confinement or resonances.

Uncertainty propagation should preserve correlations. An intensity calibration error can shift both $t$ and $U$, so treating their errors as independent can exaggerate or understate the uncertainty in $U/t$.

Determinant quantum Monte Carlo

For the repulsive bipartite Hubbard model at half filling, symmetry removes the fermion sign problem and determinant quantum Monte Carlo can provide controlled finite-temperature benchmarks. This makes half filling an especially strong calibration regime.

Doping, frustration, or certain hoppings restore the sign problem. Failure there is a property of the algorithmic representation, not proof that the physical model is intrinsically unknowable.

Auxiliary-field and constrained-path methods

Auxiliary-field quantum Monte Carlo can reach ground-state and finite-temperature properties, but constrained-path or phaseless conditions introduce trial-state bias when the sign or phase problem is present.

Comparisons should vary the constraint and report benchmark errors in nearby controlled regimes. Agreement with experiment can validate both methods, but shared assumptions can also create shared bias.

Tensor networks

Matrix product states give highly controlled results in one dimension and useful cylinder calculations in two dimensions. Their cost grows with entanglement, cylinder width, and evolution time.

Tensor networks are particularly valuable for gauge models and synthetic ladders because constraints can reduce the Hilbert space. A low bond dimension that reproduces local observables may still miss long-range entanglement or late-time dynamics.

Quantum Monte Carlo for bosons and spin models

World-line and stochastic-series-expansion methods can benchmark many sign-problem-free bosonic and spin systems. They support thermometry, finite-size scaling, and dimensional-crossover studies.

The method’s applicability must be checked after adding flux, frustration, or complex hopping. A model that is easy at zero flux can acquire a sign problem under the very gauge field being simulated.

Numerical linked-cluster expansions

Linked-cluster methods directly estimate thermodynamic-limit observables by summing connected clusters. They are powerful at sufficiently high temperature and avoid trap-size extrapolation.

Their breakdown as temperature decreases should be diagnosed by order convergence. Extrapolating past that point can produce smooth but unreliable thermometry.

Exact diagonalization and Krylov evolution

Exact methods provide complete benchmarks on small clusters and are essential for pulse design, detector validation, and few-body topological states. Finite-size and boundary effects are often larger than numerical error.

A small-system match tests implementation. It does not guarantee that the same observable remains insensitive to omitted terms at larger size.

Linear response, Kubo formulas, and sum rules

Hall response, conductivity, compressibility, and spectral functions connect measured dynamics to response coefficients. Sum rules constrain integrated spectral weight and can expose normalization or resolution errors.

Finite pulses and trapped geometries rarely realize the ideal infinite-time, homogeneous Kubo limit exactly. The experimental protocol should be simulated directly when possible.

Geometric and topological diagnostics

Berry curvature, Chern numbers, Wilson loops, charge pumping, many-body polarization, and entanglement spectra provide complementary diagnostics. Their canonical mathematical definitions belong to Geometric Phases and Topology.

Theoretical work should specify which invariant is defined for the finite, interacting, or driven system actually realized.

Gauge-theory mappings and constraint-preserving numerics

Quantum-link models truncate gauge fields to finite local Hilbert spaces. Schrieffer–Wolff and degenerate perturbation theory derive constrained effective interactions from superlattice processes. Tensor networks can enforce Gauss’s law exactly in the simulated basis.

Comparing an exactly constrained numerical model with an experiment requires adding measured gauge violations. Otherwise the numerical benchmark may be more symmetric than the hardware.

Statistical inference and model comparison

Bayesian parameter estimation, likelihood-free inference, bootstrap resampling, and posterior predictive checks can combine calibration and many-body observables.

Model selection should compare alternatives such as:

$$H_0 \quad\text{versus}\quad H_0+t'H' \quad\text{versus}\quad H_0+\mathcal L_{\mathrm{heat}}.$$

Goodness of fit to one preferred model is weaker than evidence that plausible alternatives fail.

Representative Papers and Reviews

Platforms and Hubbard models

1. D. Jaksch et al., “Cold bosonic atoms in optical lattices,” Physical Review Letters 81, 3108–3111 (1998), doi:10.1103/PhysRevLett.81.3108.
2. M. Greiner et al., “Quantum phase transition from a superfluid to a Mott insulator in a gas of ultracold atoms,” Nature 415, 39–44 (2002), doi:10.1038/415039a.
3. I. Bloch, J. Dalibard, and W. Zwerger, “Many-body physics with ultracold gases,” Reviews of Modern Physics 80, 885–964 (2008), doi:10.1103/RevModPhys.80.885.
4. W. S. Bakr et al., “A quantum gas microscope for detecting single atoms in a Hubbard-regime optical lattice,” Nature 462, 74–77 (2009), doi:10.1038/nature08482.
5. C. Gross and W. S. Bakr, “Quantum gas microscopy for single atom and spin detection,” Nature Physics 17, 1316–1323 (2021), doi:10.1038/s41567-021-01370-5.
6. F. Schäfer et al., “Tools for quantum simulation with ultracold atoms in optical lattices,” Nature Reviews Physics 2, 411–425 (2020), doi:10.1038/s42254-020-0195-3.

Correlated fermions

7. A. Mazurenko et al., “A cold-atom Fermi–Hubbard antiferromagnet,” Nature 545, 462–466 (2017), doi:10.1038/nature22362.
8. J. Koepsell et al., “Imaging magnetic polarons in the doped Fermi–Hubbard model,” Nature 572, 358–362 (2019), doi:10.1038/s41586-019-1463-1.
9. M. Lebrat et al., “Observation of Nagaoka polarons in a Fermi–Hubbard quantum simulator,” Nature 629, 317–322 (2024), doi:10.1038/s41586-024-07272-9.
10. D. Bourgund et al., “Formation of individual stripes in a mixed-dimensional cold-atom Fermi–Hubbard system,” Nature 637, 57–62 (2025), doi:10.1038/s41586-024-08270-7.
11. M. Xu et al., “A neutral-atom Hubbard quantum simulator in the cryogenic regime,” Nature 642, 909–915 (2025), doi:10.1038/s41586-025-09112-w.

Topology and synthetic dimensions

12. G. Jotzu et al., “Experimental realization of the topological Haldane model with ultracold fermions,” Nature 515, 237–240 (2014), doi:10.1038/nature13915.
13. M. Aidelsburger et al., “Measuring the Chern number of Hofstadter bands with ultracold bosonic atoms,” Nature Physics 11, 162–166 (2015), doi:10.1038/nphys3171.
14. M. Mancini et al., “Observation of chiral edge states with neutral fermions in synthetic Hall ribbons,” Science 349, 1510–1513 (2015), doi:10.1126/science.aaa8736.
15. N. R. Cooper, J. Dalibard, and I. B. Spielman, “Topological bands for ultracold atoms,” Reviews of Modern Physics 91, 015005 (2019), doi:10.1103/RevModPhys.91.015005.
16. J. Léonard et al., “Realization of a fractional quantum Hall state with ultracold atoms,” Nature 619, 495–499 (2023), doi:10.1038/s41586-023-06122-4.
17. A. Impertro et al., “Strongly interacting Meissner phases in large bosonic flux ladders,” Nature Physics 21, 895–901 (2025), doi:10.1038/s41567-025-02890-0.
18. T.-W. Zhou et al., “Measuring Hall voltage and Hall resistance in an atom-based quantum simulator,” Nature Communications 16, 10247 (2025), doi:10.1038/s41467-025-65083-6.

Gauge theories and nonequilibrium dynamics

19. B. Yang et al., “Observation of gauge invariance in a 71-site Bose–Hubbard quantum simulator,” Nature 587, 392–396 (2020), doi:10.1038/s41586-020-2910-8.
20. J. C. Halimeh et al., “Cold-atom quantum simulators of gauge theories,” Nature Physics 21, 25–36 (2025), doi:10.1038/s41567-024-02721-8.
21. M. Cheneau et al., “Light-cone-like spreading of correlations in a quantum many-body system,” Nature 481, 484–487 (2012), doi:10.1038/nature10748.
22. M. Schreiber et al., “Observation of many-body localization of interacting fermions in a quasirandom optical lattice,” Science 349, 842–845 (2015), doi:10.1126/science.aaa7432.
23. A. M. Kaufman et al., “Quantum thermalization through entanglement in an isolated many-body system,” Science 353, 794–800 (2016), doi:10.1126/science.aaf6725.
24. M. Ueda, “Quantum equilibration, thermalization and prethermalization in ultracold atoms,” Nature Reviews Physics 2, 669–681 (2020), doi:10.1038/s42254-020-0237-x.

For current claims, the article, supplementary methods, corrections, data release, and numerical code should be treated as one evidence package.

Common Misconceptions

“Ultracold means the target many-body state is cold”

Nanokelvin center-of-mass temperature before lattice loading does not fix $T/t$ or $T/J$ after an interacting ramp. Entropy redistribution, nonadiabaticity, and heating determine the final state.

“An optical lattice exactly is a Hubbard model”

The Hubbard Hamiltonian is an effective single-band description. Longer-range hopping, density-assisted terms, higher bands, confinement, and loss must be bounded for the observable and parameter regime.

“Single-site images are direct wavefunction measurements”

A snapshot samples a chosen basis through a detector channel. Reconstructing coherence, entanglement, or a Hamiltonian requires additional rotations, copies, protocols, and assumptions.

“Agreement with one classical curve validates the frontier regime”

Agreement can calibrate the overlap regime. Extrapolation into lower temperature, stronger frustration, or longer time requires uncertainty from both experiment and numerical method.

“A sign problem proves quantum advantage”

A sign problem obstructs a class of Monte Carlo representations. Other classical methods, approximations, special observables, or instance structure may remain useful. Advantage is an end-to-end scaling claim.

“A low temperature implies superconductivity”

Low $T/t$ makes pairing and competing orders accessible. Superconductivity requires phase-coherent or response evidence and scaling, not temperature alone.

“Gauge invariance follows from a conserved total charge”

Gauge invariance imposes local constraints. Opposite local violations can cancel in a global sum, so Gauss’s law must be checked site by site.

“Synthetic dimensions are ordinary spatial dimensions”

They reproduce a hopping graph, but interactions, measurement, boundaries, and locality can differ. The mapping must be stated for the phenomenon under study.

“A Chern band is an interacting topological phase”

A nonzero single-particle Chern number establishes band topology. Fractional or intrinsically interacting order needs many-body evidence.

“Persistent imbalance proves infinite-time localization”

It proves memory over the measured size and time. Prethermalization, slow transport, finite size, and external decoherence can produce similar plateaus.

“Local thermalization makes the global state mixed”

Unitary evolution preserves global purity. A subsystem becomes mixed because it is entangled with the rest of the system.

Exercises

Exercise 1: Hubbard energy scales

A half-filled repulsive Hubbard simulator has

$$\frac{U}{t}=8, \qquad \frac{t}{h}=500\,\mathrm{Hz}, \qquad \frac{k_{\mathrm B}T}{t}=0.05.$$

1. Find $J/h$ for $J=4t^2/U$.
2. Find the superexchange time $h/J$.
3. Express the temperature as $T/J$ and as $k_{\mathrm B}T/h$.

Solution

Since $U=8t$,

$$J =\frac{4t^2}{8t} =\frac{t}{2}.$$

Therefore

$$\frac{J}{h} =250\,\mathrm{Hz}, \qquad \frac{h}{J} =4.0\,\mathrm{ms}.$$

The dimensionless temperature relative to exchange is

$$\frac{k_{\mathrm B}T}{J} = \frac{0.05t}{0.5t} =0.10,$$

and

$$\frac{k_{\mathrm B}T}{h} =0.05(500\,\mathrm{Hz}) =25\,\mathrm{Hz}.$$

The exercise shows why $T/t$ and $T/J$ answer different questions. Charge motion is measured against $t$, while antiferromagnetic ordering is often controlled by $J$.

Exercise 2: Invert a microscope channel

A detector has false-negative probability $\alpha=0.03$ and false-positive probability $\beta=0.01$. It reports occupation probability $p_{\mathrm{obs}}=0.874$.

1. Invert the binary measurement model to find the true occupation $p$.
2. Why is the same scalar correction generally insufficient for a two-site connected correlator?

Solution

From

$$p_{\mathrm{obs}} =(1-\alpha)p+\beta(1-p),$$

we obtain

$$p = \frac{p_{\mathrm{obs}}-\beta} {1-\alpha-\beta} = \frac{0.874-0.01}{0.96} =0.900.$$

For a two-site correlator, the joint detector channel acts on four true outcomes. Correlated hopping or loss can make the two-site channel fail to factorize. Correcting each one-site mean and multiplying does not reconstruct $\langle n_i n_j\rangle$, so the connected quantity

$$\langle n_i n_j\rangle -\langle n_i\rangle\langle n_j\rangle$$

needs a calibrated joint or forward measurement model.

Exercise 3: Local versus global gauge checks

In a 71-site gauge simulator, six sites have $G_j-g_j=\pm1$ and all others satisfy the constraint exactly. Three violations are positive and three are negative.

1. Compute $\epsilon_G$.
2. Compute the mean signed violation.
3. Explain the diagnostic lesson.

Solution

Each violating site contributes one to the squared diagnostic:

$$\epsilon_G =\frac{6}{71} \approx 0.0845.$$

The signed mean is

$$\frac{1}{71} \sum_j(G_j-g_j) =\frac{3-3}{71} =0.$$

A global signed check would report perfect cancellation even though roughly $8.45\%$ of sites carry unit violations. Gauge symmetry is local, so the site-resolved distribution and its correlations are needed.

Exercise 4: Flux in a synthetic plaquette

The hopping along a synthetic direction carries phase $e^{i\phi x}$, while real-direction hopping is real. Show that the phase accumulated around one real–synthetic plaquette is $\phi$. For $\phi=0.32\pi$, express the flux as a fraction of one flux quantum.

Solution

Traverse the plaquette from $(x,m)$ to $(x+1,m)$, then to $(x+1,m+1)$, back to $(x,m+1)$, and finally to $(x,m)$. The real hoppings contribute no phase. The forward synthetic hop contributes $e^{i\phi(x+1)}$ and the reverse synthetic hop contributes $e^{-i\phi x}$. Their product is

$$e^{i\phi(x+1)} e^{-i\phi x} =e^{i\phi}.$$

Thus the loop phase is $\phi$. Since one flux quantum corresponds to $2\pi$ phase,

$$\frac{\Phi}{\Phi_0} =\frac{\phi}{2\pi} =\frac{0.32\pi}{2\pi} =0.16.$$

The result identifies the single-particle Peierls flux. Interactions and finite synthetic width determine the many-body response.

Exercise 5: Filling of the minimal fractional Hall state

Consider two bosons on a $4\times4$ lattice with uniform flux

$$\frac{\phi}{2\pi}=\frac14$$

per plaquette, and use $N_\phi=N_{\mathrm{site}}\phi/(2\pi)$ for this finite counting exercise.

1. Find $N_\phi$.
2. Find $\nu=N/N_\phi$.
3. Why does the value $\nu=1/2$ not by itself prove a Laughlin state?

Solution

There are 16 sites, so

$$N_\phi =16\left(\frac14\right) =4.$$

With $N=2$ bosons,

$$\nu =\frac{N}{N_\phi} =\frac{2}{4} =\frac12.$$

The filling is necessary for the target bosonic Laughlin state but does not identify the wavefunction. A noninteracting or excited state can have the same particle and flux count. Gap spectroscopy, interaction suppression, correlations, response, and comparison with the target state provide the additional evidence.

Exercise 6: Thermalization without global entropy production

Let a bipartite system begin in a product state and evolve unitarily into

$$|\psi\rangle = \frac{ |0\rangle_A|0\rangle_B + |1\rangle_A|1\rangle_B }{\sqrt2}.$$

1. Find $\rho_A$.
2. Find $S_A$ using natural logarithms.
3. Compare the subsystem entropy with the entropy of the global state.

Solution

Tracing over $B$ removes the off-diagonal terms because $\langle0|1\rangle_B=0$:

$$\rho_A = \frac12 \left( |0\rangle\langle0| + |1\rangle\langle1| \right).$$

Its two eigenvalues are $1/2$, so

$$S_A = -2\left(\frac12\ln\frac12\right) =\ln2.$$

The global density operator $|\psi\rangle\langle\psi|$ is pure and has entropy zero. Entanglement makes a subsystem mixed even though the isolated total system produces no von Neumann entropy. A many-body subsystem can similarly approach a thermal reduced state.

Exercise 7: Distinguish a plateau from localization

An experiment observes a nonzero density imbalance until the maximum time $t_{\max}=100\,\hbar/t$. Design four additional tests that would distinguish many-body localization from slow thermalization or a prethermal plateau.

Solution

A useful campaign includes:

1. Time scaling: extend the observation window and fit several functional forms rather than assuming a constant asymptote.
2. Size scaling: repeat for increasing system lengths to distinguish a boundary-limited plateau from a bulk memory effect.
3. Perturbation tests: vary disorder, interactions, transverse coupling, and integrability-breaking terms; measure how the relaxation time changes.
4. Independent observables: measure transport, local autocorrelations, number fluctuations, and entanglement-sensitive quantities. A density imbalance alone is incomplete.
5. Open-system controls: vary photon scattering, loss, and coupling to residual tubes or planes to bound environmental freezing or relaxation.
6. Initial-state dependence: test several charge and spin patterns at comparable energy density.

No finite experiment proves an infinite-time phase directly. The goal is a scaling account that rules out simpler finite-time explanations.

Exercise 8: Design a quantum-simulation evidence package

A group claims that its doped two-dimensional Hubbard simulator has entered a regime beyond reliable classical calculation and has observed a new phase. List the minimum evidence needed to make the claim scientifically actionable.

Solution

A defensible package includes:

1. independent calibration of $t$, $U$, $t'$, trap shape, filling, and detector response with propagated covariance;
2. state-preparation checks, including ramp-time scans, reverse ramps, heating, loss, and spatial homogeneity;
3. at least two temperature or entropy diagnostics, with explicit numerical assumptions in the doped regime;
4. several observables that discriminate the proposed phase from competing orders and from a crossover;
5. system-size, boundary, and analysis-window tests;
6. agreement with controlled classical methods in overlapping regimes and blind predictions where possible;
7. comparisons against multiple approximate methods in the target regime, including their disagreement;
8. raw or minimally processed snapshots, calibration data, analysis code, and preregistered exclusions;
9. repeated preparations through more than one path to test metastability; and
10. wording that separates observation, inference, and extrapolation.

Being beyond one classical method raises the value of validation; it does not lower the standard.

Connections to Other Volumes

• Hubbard Model develops the canonical fermionic model and its exact limits.
• Bose–Hubbard Model develops the bosonic model behind Mott, quench, and gauge-mapping experiments.
• Lattice Models Overview compares Hubbard, spin, and constrained models.
• Relaxation and Thermalization distinguishes equilibration, thermalization, conserved quantities, and open-system effects.
• Eigenstate Thermalization Hypothesis gives the eigenstate-level criterion used to interpret local thermalization.
• Many-Body Localization Preview explains localization diagnostics and unresolved stability questions.
• Floquet Theorem in Quantum Mechanics develops quasienergies, micromotion, and stroboscopic evolution.
• Chern Numbers gives the geometric invariant behind topological-band experiments.
• Berry Curvature connects parameter-space geometry to anomalous velocity.
• Field Operators supplies the second-quantized language for atomic gases.
• Low-Energy Scattering develops the collision input to lattice interaction parameters.
• AMO Physics Roadmap connects cooling, trapping, light–matter interaction, and many-body platforms.

What Changed This Year

This section covers peer-reviewed work published from January through 26 July 2026 and changes in the interpretation of the 2025 milestones.

Dimensional crossover was mapped with temperature as a control axis

An interacting $^{87}\mathrm{Rb}$ simulator in anisotropic triangular optical lattices mapped quantum regimes from three to zero effective dimensions over controlled temperature and anisotropy ranges. The experiment identified thermal regimes between integer-dimensional quantum regimes and compared the phase maps with quantum Monte Carlo Nature Communications (2026), doi:10.1038/s41467-026-71881-3.

Established in 2026: a broad, experimentally controlled finite-temperature dimensional-crossover map, including an intermediate low-dimensional route to thermal behavior in selected parameter regions.

Qualification: the phase boundaries are inferred from zero-momentum fractions, correlation lengths, fitting procedures, and comparison with a homogeneous numerical model. Trap and atom-number differences contribute quantitative discrepancies.

Bound-cluster tunnelling produced a massive spatial superposition

Zhang et al. used an optical-lattice Bose–Hubbard setting to coherently tunnel bound atomic clusters, forming a composite object with reported mass $608\,\mathrm{AMU}$. An interferometric protocol certified spatial entanglement and supported quantum-enhanced sensing Nature Physics (2026), doi:10.1038/s41567-026-03281-9.

This result expands the nonequilibrium and few-body use of lattice simulators: interaction ordinarily suppresses composite tunnelling, but model-parameter control can engineer a scalable path. Its main interpretation is a many-particle superposition and metrology result, not a solution of a thermodynamic many-body phase.

The 2025 Hubbard cooling result is now the central benchmark

No peer-reviewed 2026 result reviewed here supersedes the 2025 $T/t\approx0.05$ half-filled Hubbard benchmark. The frontier has shifted from whether long-range antiferromagnetism can be made very cold to whether the same preparation quality can be retained under controlled doping, altered geometry, and measurements that discriminate pseudogap, stripe, and pairing physics.

The key uncertainty is thermometry away from half filling, where controlled classical reference calculations become less available.

Gauge-theory progress is separating experiment from proposal more clearly

The 2025 field review consolidated experimental progress in $(1+1)$D Abelian models. The proposed large-scale $(2+1)$D $U(1)$ implementation with dynamical matter remains theory as of this review. It should not be listed as an experimental higher-dimensional gauge simulation until the mapping, local constraint diagnostics, and target dynamics are realized in hardware.

Assessment

The frontier is moving from spectacular isolated demonstrations toward quantitative model discrimination. The next review should check for:

• independent thermometry and pairing diagnostics in the cryogenic doped Hubbard regime;
• larger interacting topological states with nonlocal and response diagnostics;
• an experimental $(2+1)$D gauge simulator with measured local Gauss-law compliance;
• stronger interactions and more legs in synthetic dimensions;
• transport coefficients extracted over a genuine hydrodynamic window;
• long-time and size scaling of purported nonergodic regimes; and
• public, end-to-end comparisons that could support an analog quantum advantage claim.