Quantum Thermalization

Quantum thermalization is the emergence of thermal predictions for a declared class of local observables or subsystems from the unitary dynamics of an isolated many-body system. The global pure state does not literally become a mixed Gibbs state. Information remains in nonlocal correlations, exact conserved quantities remain fixed, and a finite system can recur.

This page is a quantum-matter evidence bridge. It owns the experimental and numerical workflow for deciding whether a material model or quantum simulator has thermalized. Relaxation and Thermalization is the canonical home for dephasing, equilibration bounds, diagonal ensembles, constrained thermal candidates, recurrences, and error decompositions. Eigenstate Thermalization Hypothesis owns the ETH ansatz. Integrability and Generalized Gibbs Ensembles Preview and Prethermalization Preview own their respective alternatives.

The operational claim has two logically separate parts:

$$\begin{gathered} \text{small late-time fluctuations} \quad \text{around an equilibrated value}, \\[4pt] \text{equilibrated value} \quad = \quad \text{prediction of a justified ensemble}. \end{gathered}$$

The first is equilibration. The second is thermalization. A smooth plateau can satisfy the first and fail the second.

Status at a Glance

Statement	Status	Boundary
Local observables can equilibrate under closed unitary dynamics	Standard and widely demonstrated	The global state remains pure and reversible
Generic nonintegrable lattice models often satisfy ETH in bulk spectral windows	Strong numerical and experimental support	ETH is not a theorem for every local Hamiltonian
Integrable systems can equilibrate to generalized ensembles	Standard in controlled model classes	The charge family and observable class must be complete enough
Weakly broken constraints can create prethermal plateaus	Standard mechanism	Eventual drift is part of the claim
A Markovian open system relaxes to an attractor	Standard open-system physics	The attractor is not evidence for closed-system ETH
Every isolated many-body system thermalizes	False	Integrability, localization, fragmentation, scars, and symmetry sectors provide exceptions
A recent high-temperature theorem settles generic thermalization	False	Its qubit, locality, translation, state-ensemble, and no-perfect-resonance assumptions matter

Begin with a Protocol Ledger

A thermalization claim is incomplete until it specifies:

Item	Required information
Hamiltonian	Local terms, boundaries, dimension, interaction range, and any drive
Initial state	Preparation, energy density, charge densities, and entanglement structure
Exact sectors	Particle number, magnetization, momentum, parity, gauge constraints, and other symmetries
Observable class	Local densities, few-point correlators, currents, or a stated subsystem
Candidate ensemble	Microcanonical, canonical, grand canonical, generalized, or prethermal
Parameter fixing	How temperature and chemical potentials are obtained without fitting the tested observable
Time window	Microscopic, local-relaxation, transport, boundary-reflection, bath, and recurrence scales
Error budget	Preparation, calibration, sampling, readout, truncation, finite-size, and model error

For a closed system prepared in $|\psi_0\rangle$,

$$|\psi(t)\rangle = e^{-iHt/\hbar}|\psi_0\rangle, \qquad \rho_A(t) = \operatorname{Tr}_{\bar A} |\psi(t)\rangle\langle\psi(t)|.$$

The reduced state $\rho_A(t)$ can approach thermal predictions even though

$$S\!\left( |\psi(t)\rangle\langle\psi(t)| \right) = 0$$

for all $t$. Entanglement transfers locally accessible information into correlations with $\bar A$; it does not destroy global information.

Closed Quantum Systems

Choose an operational target

For observables $O_a$ in a declared set $\mathcal O$, define a normalized residual

$$R_a(t) = \frac{ \langle O_a(t)\rangle - \langle O_a\rangle_{\mathrm{ens}} }{ \sigma_a },$$

where $\sigma_a$ is a stated experimental or numerical uncertainty scale. A thermalization claim should show that:

1. $R_a(t)$ becomes statistically compatible with zero for all tested $a$;
2. late-time fluctuations remain controlled over a nontrivial window;
3. the result survives changes of initial state at fixed macroscopic constraints;
4. no omitted exact or quasilocal charge explains the residuals;
5. boundaries and external baths have not set the observed value.

One observable can relax accidentally. Several observables constrain the candidate ensemble and make a coincidental match less likely.

Fix the ensemble independently

For a canonical candidate,

$$\rho_\beta = \frac{e^{-\beta H}}{Z(\beta)}, \qquad Z(\beta) = \operatorname{Tr}e^{-\beta H},$$

the inverse temperature should be fixed by energy matching,

$$\langle\psi_0|H|\psi_0\rangle = \operatorname{Tr}(\rho_\beta H),$$

after projecting into every exact sector used by the dynamics. If particle number fluctuates because the preparation samples sectors, the predicted average must reproduce the same sector weights or use a justified grand-canonical approximation.

Fitting $\beta$ separately to every observable makes the test circular. The strong design uses one independently fixed parameter set and reserves multiple observables as out-of-sample checks.

Separate local and global statements

A finite fraction of the system need not be thermal merely because a fixed small region is. The appropriate scaling limit declares how the subsystem size $\ell_A$ behaves relative to total size $L$:

$$\ell_A \ \text{fixed as}\ L\to\infty \qquad \text{or} \qquad \frac{\ell_A}{L} \to f>0.$$

Those limits answer different questions. Local statistical mechanics is usually associated with the first. Tomography of half a small quantum processor probes the second and requires stronger finite-size caution.

Eigenstate Thermalization Interface

ETH connects the structure of energy eigenstates to thermal behavior. The practical diagonal test asks whether eigenstate expectation values of a local operator become a smooth function of energy within one exact symmetry sector:

$$O_{nn} = \langle E_n|O|E_n\rangle \approx O_{\mathrm{th}}(E_n).$$

For quantum matter, four controls are non-negotiable:

1. resolve every exact sector before plotting levels or matrix elements;
2. use bulk energy-density windows unless the claim concerns an edge;
3. subtract the smooth energy dependence before measuring fluctuations;
4. track extreme outliers as well as typical variance.

Small variance supports weak ETH. Absence of outliers in accessible sizes does not prove strong ETH in the thermodynamic limit. Conversely, one atypical scarred eigenstate does not prevent broad classes of initial states from thermalizing if they have negligible overlap with it.

ETH is sufficient under appropriate initial-state and spectral assumptions, but an experiment can establish local thermalization without reconstructing all eigenstates. Time-domain tests and eigenstate tests are complementary, not interchangeable.

Integrability

An integrable model carries an extensive family of local or quasilocal conserved quantities $Q_j$. An ordinary Gibbs state constrained only by energy and particle number can therefore discard information that local observables retain. The generalized candidate is

$$\rho_{\mathrm{GGE}} = \frac{1}{Z_{\mathrm{GGE}}} \exp\!\left( - \sum_j\lambda_jQ_j \right).$$

This is still a maximum-entropy construction, but with more constraints. A valid GGE test must:

• specify and justify the charge family;
• determine $\lambda_j$ from the initial state;
• compare out-of-sample observables;
• check whether truncating the charge set changes predictions;
• break integrability deliberately and track the resulting drift.

The quantum Newton’s cradle is a canonical warning: many collisions and strong interactions do not guarantee ordinary thermalization when one-dimensional kinematics and near-integrability constrain scattering.

Prethermalization

A prethermal state is a long-lived intermediate regime controlled by approximate conservation laws, a large energy separation, weak integrability breaking, or a high-frequency drive. Its defining hierarchy is

$$\tau_{\mathrm{local}} \ll t \ll \tau_{\mathrm{escape}},$$

where local observables have relaxed but the slow constraint has not.

The plateau must be tested against an effective or generalized ensemble appropriate to the approximate charges. Eventual drift is not an experimental nuisance to erase; it distinguishes prethermalization from exact stationarity. Parameter scaling is especially valuable. For example, if an integrability-breaking perturbation has strength $g$, a measured escape time should be compared with a theoretically motivated $g$ dependence rather than reported at one value.

Do not call every two-stage trace prethermal. Inhomogeneous dephasing, transport from distant regions, detector bandwidth, or a weak external bath can also produce separated timescales.

Relationship to Open Systems

Closed and open thermalization answer different dynamical questions.

For an isolated system,

$$\dot\rho = -\frac{i}{\hbar}[H,\rho],$$

and local thermal appearance emerges despite globally reversible dynamics. For a Markovian open system,

$$\dot\rho = -\frac{i}{\hbar}[H,\rho] + \sum_\mu \left( L_\mu\rho L_\mu^\dagger - \frac12 \left\{ L_\mu^\dagger L_\mu,\rho \right\} \right),$$

and the environment can select an attractor by removing information and energy.

Observation	Closed-system interpretation	Open-system alternative
Local Gibbs-like values	Internal dephasing plus ensemble equivalence	Bath-imposed detailed balance
Entropy growth of a subsystem	Entanglement with the rest of the isolated system	Entanglement and information loss to an environment
Exponential relaxation	Emergent many-body rate over a window	Liouvillian decay mode
Initial-state independence	Local forgetting within conserved sectors	Attraction to a unique steady state
Energy drift	Inconsistent with a time-independent isolated Hamiltonian	Heating, cooling, loss, or drive

An isolation audit should compare the observed relaxation time $\tau_{\mathrm{rel}}$ with independently measured environmental times:

$$\tau_{\mathrm{rel}} \ll \min \left( \Gamma_{\mathrm{loss}}^{-1}, \Gamma_\phi^{-1}, \Gamma_{\mathrm{heat}}^{-1} \right).$$

If this hierarchy fails, an open-system model is required. Steady States and Relaxation owns Liouvillian attractors, while this page owns the experimental boundary between the two interpretations.

Platform Evidence

Platform	Strong thermalization probes	Characteristic limitations
Optical-lattice atoms	Site occupations, correlations, interaction quenches, subsystem Rényi entropy	Traps, atom loss, finite entropy calibration, boundary propagation
One-dimensional quantum gases	Momentum distributions, correlations, tunable integrability breaking	Tube averaging, residual confinement, limited local access
Trapped-ion spins	Programmable Hamiltonians, state-resolved readout, correlation spreading	Long-range couplings, modest size, inhomogeneity, decoherence
Superconducting qubits	Local tomography, randomized measurements, entanglement spectra	Gate/Trotter error, calibration drift, short coherent depth
Pump–probe solids	Ultrafast order parameters, spectra, quasiparticle relaxation	Strong bath coupling, nonthermal distributions, depth averaging, heating

In a solid, “electrons thermalize in $50\ \mathrm{fs}$” often means that a measured distribution is well fit by a Fermi–Dirac form over a limited energy window. A complete claim should test other observables, conservation constraints, electron–phonon energy transfer, matrix-element effects, and whether one temperature describes the whole electronic subsystem.

A Thermalization Claim Ladder

Evidence level	Defensible statement
One observable becomes smooth	Relaxation of that observable
Several observables become stationary	Local equilibration over the measured window
Independently fixed constrained ensemble predicts them	Thermalization for the tested observable class
Matched initial states converge to the same local state	Initial-state independence within the declared constraints
Size, time, and subsystem scaling remain consistent	Evidence toward a thermodynamic local statement
ETH matrix elements or eigenstates also pass sector-resolved tests	A microscopic mechanism consistent with thermalization

No row licenses the claim that the global wavefunction became Gibbsian.

Current Research Boundary

Most experimentally relevant thermalization claims remain model- and observable-specific. Rigorous progress is correspondingly assumption-indexed.

Pilatowsky-Cameo and Choi proved a high-temperature local-thermalization result for local, translation-invariant qubit Hamiltonians without perfect spectral resonances and for typical low-complexity initial states drawn from a high-entropy ensemble. Published online in December 2025 and in Nature Communications volume 17 (2026), it is a substantial theorem, not a proof that every local Hamiltonian thermalizes. Disorder, exact integrability, localization, special state families, low-temperature sectors, and perfect resonances lie outside or can violate its assumptions.

Recent quantum-simulation experiments also access early-time chaos through entanglement-spectrum statistics. Such indicators probe a mechanism associated with thermalization, but early level repulsion in an entanglement Hamiltonian is not by itself a late-time Gibbs-state test.

Evidence Workflow

1. Declare the closed-system model and preparation.
2. Resolve exact sectors and conserved densities.
3. Choose local observables or subsystem sizes before inspecting late-time data.
4. Fix ensemble parameters from conserved quantities, not from the test observables.
5. Separate stationarity error from ensemble-prediction error.
6. Repeat with several initial states at matched energy and charge densities.
7. Compare integrable, weakly broken, and generic parameter points.
8. Measure the external loss, dephasing, and heating rates independently.
9. Vary size, time window, boundary distance, and subsystem fraction.
10. Report the narrowest justified claim: relaxation, equilibration, prethermalization, generalized thermalization, or ordinary thermalization.

Common Mistakes

• Saying that the global pure state becomes a Gibbs density operator.
• Equating a time average with pointwise convergence.
• Calling a stationary value thermal without comparing an ensemble.
• Fitting a separate temperature to every observable.
• Omitting exact symmetry sectors or conserved charges.
• Treating ETH as synonymous with random-matrix level statistics.
• Using one product state to establish initial-state independence.
• Confusing integrable GGE relaxation with failure to equilibrate.
• Calling a long plateau prethermal without observing or bounding escape.
• Ignoring hydrodynamic tails and boundary reflections.
• Interpreting bath-driven relaxation as closed-system thermalization.
• Using subsystem entropy alone without separating thermal mixing from entanglement.

Exercises

1. Global purity and local entropy

A bipartite system starts in a pure product state and evolves unitarily into an entangled state. What happens to the global von Neumann entropy and to the entropy of subsystem $A$?

Solution

Unitary evolution preserves the eigenvalues of the global density operator, so

$$S(\rho_{AB}(t)) = 0$$

at all times. The reduced state

$$\rho_A(t) = \operatorname{Tr}_B\rho_{AB}(t)$$

can become mixed as entanglement develops, so $S(\rho_A)$ can increase. Local entropy growth therefore does not require global information loss.

2. Detect circular temperature fitting

An experiment measures three observables and chooses three different temperatures so that a Gibbs model matches each one. Has thermalization been tested?

Solution

No. The temperatures are fit parameters that absorb disagreements among observables. For a closed system with fixed energy, one should determine a single $\beta$ from

$$\langle H\rangle_0 = \operatorname{Tr}(\rho_\beta H),$$

including all exact sector constraints, and then predict the three observables without refitting. Their residuals are the test.

3. Choose the ensemble

Two initial states have equal energy and particle number but different values of an additional exactly conserved local charge $Q$. They equilibrate to different local densities. Does this refute equilibration or ETH?

Solution

It first refutes the incomplete ensemble that omitted $Q$. The appropriate candidate must retain its constraint, for example

$$\rho \propto \exp(-\beta H-\lambda_Q Q).$$

If both states equilibrate locally to predictions fixed by their respective charge values, equilibration and generalized thermalization are compatible with the observations. ETH must likewise be formulated within the correct charge data or integrable structure.

4. Identify a prethermal window

A local correlator relaxes by $\tau_{\mathrm{local}}=4\hbar/J$, remains near a plateau until $\tau_{\mathrm{escape}}=800\hbar/J$, and then drifts to the Gibbs value. At $t=100\hbar/J$, which description is appropriate?

Solution

The time lies in the hierarchy

$$4 \ll 100 \ll 800,$$

in units of $\hbar/J$. The system is in a prethermal window: local relaxation has occurred, but the slow constraint has not yet decayed. The plateau should be compared with an effective or generalized prethermal ensemble. Its eventual escape supports, rather than contradicts, the prethermal interpretation.

5. Closed or open?

A simulator relaxes in $\tau_{\mathrm{rel}}=12\ \mathrm{ms}$, while independent calibrations give loss, dephasing, and heating times of $2.0\ \mathrm{s}$, $0.30\ \mathrm{s}$, and $0.80\ \mathrm{s}$. Evaluate the isolation hierarchy.

Solution

The shortest environmental time is $0.30\ \mathrm{s}=300\ \mathrm{ms}$, so

$$\frac{\tau_{\mathrm{rel}}} {\min(\Gamma_{\mathrm{loss}}^{-1}, \Gamma_\phi^{-1}, \Gamma_{\mathrm{heat}}^{-1})} = \frac{12}{300} = 0.04.$$

Relaxation is much faster than the calibrated environmental processes, leaving a plausible closed-system window. The experiment should still propagate calibration uncertainty and check for state-dependent decoherence, but the timescale ordering supports an intrinsic interpretation.

6. Design an initial-state-independence test

Describe a minimal experiment that distinguishes local equilibration from thermalization.

Solution

Prepare at least three microscopically different states with matched energy density, particle density, exact symmetry sector, and other relevant charges. Evolve each under the same Hamiltonian. Measure several local observables and their temporal fluctuations over a window before boundary return or bath coupling.

First test whether each trace becomes stationary. That establishes equilibration. Then fix one constrained ensemble from the common conserved data and compare all late-time observables with its predictions without refitting. Agreement across initial states supports thermalization for the tested observable class.

Connections

• Relaxation and Thermalization is the canonical home for dephasing, equilibration bounds, ensemble selection, finite-size limits, and detailed evidence standards.
• Eigenstate Thermalization Hypothesis owns diagonal, off-diagonal, weak, strong, and subsystem ETH.
• Integrability and Generalized Gibbs Ensembles Preview owns charge completeness, generalized ensembles, and Bethe-ansatz interfaces.
• Prethermalization Preview owns approximate charges, scale separation, effective ensembles, and escape-time diagnostics.
• Many-Body Localization gives the platform-facing evidence workflow for a disorder-driven failure of thermalization.
• Quantum Quenches owns sudden-switch preparation, final-energy distributions, light cones, and Loschmidt amplitudes.
• Quenches gives the platform-facing switch, energy-injection, and pump–probe protocol ledger.
• Time-Dependent Correlations owns stationarity, spectra, and two-time response conventions.
• Statistical Ensembles Overview maps microcanonical, canonical, grand-canonical, and generalized choices.
• Steady States and Relaxation owns open-system attractors and Liouvillian relaxation.

References

1. J. M. Deutsch, “Quantum Statistical Mechanics in a Closed System,” Physical Review A 43, 2046–2049 (1991), doi:10.1103/PhysRevA.43.2046. Introduces the eigenstate-based thermalization mechanism.
2. M. Srednicki, “Chaos and Quantum Thermalization,” Physical Review E 50, 888–901 (1994), doi:10.1103/PhysRevE.50.888. Develops the ETH perspective for chaotic systems.
3. M. Rigol, V. Dunjko, and M. Olshanii, “Thermalization and Its Mechanism for Generic Isolated Quantum Systems,” Nature 452, 854–858 (2008), doi:10.1038/nature06838. Connects ETH to quench thermalization in lattice models.
4. P. Reimann, “Foundation of Statistical Mechanics under Experimentally Realistic Conditions,” Physical Review Letters 101, 190403 (2008), doi:10.1103/PhysRevLett.101.190403. Gives finite-system equilibration bounds under spectral conditions.
5. N. Linden, S. Popescu, A. J. Short, and A. Winter, “Quantum Mechanical Evolution towards Thermal Equilibrium,” Physical Review E 79, 061103 (2009), doi:10.1103/PhysRevE.79.061103. Establishes subsystem equilibration bounds.
6. L. D’Alessio, Y. Kafri, A. Polkovnikov, and M. Rigol, “From Quantum Chaos and Eigenstate Thermalization to Statistical Mechanics and Thermodynamics,” Advances in Physics 65, 239–362 (2016), doi:10.1080/00018732.2016.1198134. Reviews ETH, chaos, and numerical diagnostics.
7. C. Gogolin and J. Eisert, “Equilibration, Thermalisation, and the Emergence of Statistical Mechanics in Closed Quantum Systems,” Reports on Progress in Physics 79, 056001 (2016), doi:10.1088/0034-4885/79/5/056001. Reviews rigorous and operational frameworks.
8. T. Mori, T. N. Ikeda, E. Kaminishi, and M. Ueda, “Thermalization and Prethermalization in Isolated Quantum Systems: A Theoretical Overview,” Journal of Physics B 51, 112001 (2018), doi:10.1088/1361-6455/aabcdf. Reviews ETH, integrability breaking, and prethermal mechanisms.
9. M. Rigol, V. Dunjko, V. Yurovsky, and M. Olshanii, “Relaxation in a Completely Integrable Many-Body Quantum System,” Physical Review Letters 98, 050405 (2007), doi:10.1103/PhysRevLett.98.050405. Establishes generalized-ensemble logic for integrable dynamics.
10. T. Kinoshita, T. Wenger, and D. S. Weiss, “A Quantum Newton’s Cradle,” Nature 440, 900–903 (2006), doi:10.1038/nature04693. Demonstrates striking suppression of ordinary thermalization near one-dimensional integrability.
11. S. Trotzky et al., “Probing the Relaxation towards Equilibrium in an Isolated Strongly Correlated One-Dimensional Bose Gas,” Nature Physics 8, 325–330 (2012), doi:10.1038/nphys2232. Compares lattice-gas relaxation with controlled many-body calculations.
12. M. Gring et al., “Relaxation and Prethermalization in an Isolated Quantum System,” Science 337, 1318–1322 (2012), doi:10.1126/science.1224953. Observes a prethermal state in split one-dimensional Bose gases.
13. 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. Measures subsystem thermal behavior and entanglement in an isolated atom array.
14. C. Neill et al., “Ergodic Dynamics and Thermalization in an Isolated Quantum System,” Nature Physics 12, 1037–1041 (2016), doi:10.1038/nphys3830. Probes ergodic evolution in a superconducting-qubit system.
15. C. Bertoni et al., “Typical Thermalization of Low-Entanglement States,” Communications Physics 8, 301 (2025), doi:10.1038/s42005-025-02161-7. Proves an assumption-indexed thermalization result for operationally accessible state classes under weak spectral smoothing.
16. S. Pilatowsky-Cameo and S. Choi, “Quantum Thermalization Must Occur in Translation-Invariant Systems at High Temperature,” Nature Communications 17, 75 (2026), doi:10.1038/s41467-025-66777-7. Proves local thermalization under explicit qubit, locality, translation, temperature, state-ensemble, and resonance assumptions.
17. N. Mueller et al., “Quantum Computing Universal Thermalization Dynamics in a $(2+1)$D Lattice Gauge Theory,” Nature Communications 16, 5492 (2025), doi:10.1038/s41467-025-60177-7. Uses entanglement-spectrum statistics to probe early-time chaos on a trapped-ion processor.

Further Reading

• A. Polkovnikov, K. Sengupta, A. Silva, and M. Vengalattore, “Colloquium: Nonequilibrium Dynamics of Closed Interacting Quantum Systems,” Reviews of Modern Physics 83, 863–883 (2011), doi:10.1103/RevModPhys.83.863. A broad entry to quenches, relaxation, and nonequilibrium many-body methods.
• J. Eisert, M. Friesdorf, and C. Gogolin, “Quantum Many-Body Systems out of Equilibrium,” Nature Physics 11, 124–130 (2015), doi:10.1038/nphys3215. A concise overview of equilibration and its exceptions.