Thermodynamics typically assumes that a system in contact with a thermal reservoir relaxes to the Gibbs state, or to the grand canonical Gibbs state when it also exchanges particles. However, this only holds when the coupling is very weak. Beyond this, the expected equilibrium state is the mean force Gibbs (MFG) state, obtained by tracing the reservoir out of the thermal state of the combined system. Despite its relevance, general analytic results for the MFG state have so far been obtained almost exclusively for bosonic reservoirs. Here we derive the grand canonical MFG state to second order in the coupling for a generic fermionic system coupled to a thermal reservoir at arbitrary chemical potential, using the concept of the mode reduced state in order for the fermionic MFG state to be well defined. The correction to the bare Gibbs state separates into particle and hole channels weighted by the empty and occupied reservoir states, related through the Kubo-Martin-Schwinger condition. We apply the result to the resonant level model, two modes coupled to a shared reservoir, and the single impurity Anderson model. For the two modes, the reservoir mediates coherence between modes that do not interact directly, something the bare Gibbs state cannot capture. We verify our results against numerically exact calculations, including real-time dynamics for the interacting model, and find agreement between the static prediction and the dynamical relaxation.
AI Summary
The paper addresses finite-coupling corrections to thermal equilibrium for fermionic systems exchanging particles with a reservoir. It derives the grand-canonical mean force Gibbs state—the system reduction of the joint thermal state—for an interacting, number-conserving system coupled by particle-exchange terms to a Gaussian fermionic reservoir at chemical potential $\mu$.
An imaginary-time Kubo expansion gives $\tau_{\mathrm{MF}}=\tau_S+\lambda^2(\Delta_{\mathrm{diag}}+\Delta_{\mathrm{off}})+O(\lambda^4)$. The expression uses system Bohr-frequency eigenoperators and particle/hole reservoir correlations; the KMS condition determines all coefficients from one reservoir function and its derivative. Fermionic reduction is handled through the mode-reduced state, using a system-first ordering compatible with conventional partial tracing. The paper also derives the second-order Hamiltonian of mean force. At high temperature, state corrections scale as $O(\lambda^2\beta^2)$, improving to $O(\lambda^2\beta^3)$ for coupling operators linear in annihilation operators. A rewritten expression permits a zero-temperature limit with persistent energy-basis coherences.
Benchmarks cover the resonant level model, two noninteracting modes sharing a reservoir, and the single impurity Anderson model. Correlation-matrix calculations provide exact quadratic-model comparisons; thermofield-chain matrix-product-state dynamics benchmark the interacting case. Agreement is reported at weak-to-intermediate coupling, including reservoir-mediated intermode coherence.
The result remains perturbative. The truncated Hamiltonian’s Gibbs state additionally requires population corrections small relative to bare populations, which can fail at low temperature. The zero-temperature derivation assumes a nondegenerate system grand-canonical ground state. Dynamical relaxation to the mean force state is tested numerically, not proved generally; dark modes can prevent it.
Little sign of AI-generated textIncremental
Error found: The claimed low-temperature boundedness of the auxiliary functions B_± is not unconditional: at zero transition frequency and with a smooth reservoir spectral density satisfying J(μ)>0, they diverge logarithmically with inverse temperature, although this does not invalidate the central finite-temperature perturbative result.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
Thermodynamics typically assumes that a system in contact with a thermal reservoir relaxes to the Gibbs state, or to the grand canonical Gibbs state when it also exchanges particles. However, this only holds when the coupling is very weak. Beyond this, the expected equilibrium state is the mean force Gibbs (MFG) state, obtained by tracing the reservoir out of the thermal state of the combined system. Despite its relevance, general analytic results for the MFG state have so far been obtained almost exclusively for bosonic reservoirs. Here we derive the grand canonical MFG state to second order in the coupling for a generic fermionic system coupled to a thermal reservoir at arbitrary chemical potential, using the concept of the mode reduced state in order for the fermionic MFG state to be well defined. The correction to the bare Gibbs state separates into particle and hole channels weighted by the empty and occupied reservoir states, related through the Kubo-Martin-Schwinger condition. We apply the result to the resonant level model, two modes coupled to a shared reservoir, and the single impurity Anderson model. For the two modes, the reservoir mediates coherence between modes that do not interact directly, something the bare Gibbs state cannot capture. We verify our results against numerically exact calculations, including real-time dynamics for the interacting model, and find agreement between the static prediction and the dynamical relaxation.
The paper addresses finite-coupling corrections to thermal equilibrium for fermionic systems exchanging particles with a reservoir. It derives the grand-canonical mean force Gibbs state—the system reduction of the joint thermal state—for an interacting, number-conserving system coupled by particle-exchange terms to a Gaussian fermionic reservoir at chemical potential μ.
An imaginary-time Kubo expansion gives τMF=τS+λ2(Δdiag+Δoff)+O(λ4). The expression uses system Bohr-frequency eigenoperators and particle/hole reservoir correlations; the KMS condition determines all coefficients from one reservoir function and its derivative. Fermionic reduction is handled through the mode-reduced state, using a system-first ordering compatible with conventional partial tracing. The paper also derives the second-order Hamiltonian of mean force. At high temperature, state corrections scale as O(λ2β2), improving to O(λ2β3) for coupling operators linear in annihilation operators. A rewritten expression permits a zero-temperature limit with persistent energy-basis coherences.
Benchmarks cover the resonant level model, two noninteracting modes sharing a reservoir, and the single impurity Anderson model. Correlation-matrix calculations provide exact quadratic-model comparisons; thermofield-chain matrix-product-state dynamics benchmark the interacting case. Agreement is reported at weak-to-intermediate coupling, including reservoir-mediated intermode coherence.
The result remains perturbative. The truncated Hamiltonian’s Gibbs state additionally requires population corrections small relative to bare populations, which can fail at low temperature. The zero-temperature derivation assumes a nondegenerate system grand-canonical ground state. Dynamical relaxation to the mean force state is tested numerically, not proved generally; dark modes can prevent it.
The Gibbs state is widely taken to be the equilibrium state of a system in contact with an environment at temperature T. However, non-negligible interactions between system and environment can give rise to an altered state. Here, we derive general expressions for this mean force Gibbs state, valid for any system that interacts with a bosonic reservoir. First, we derive the state in the weak coupling limit and find that, in general, it maintains coherences with respect to the bare system Hamiltonian. Second, we develop a new expansion method suited to investigate the ultrastrong coupling regime. This allows us to derive the explicit form for the mean force Gibbs state, and we find that it becomes diagonal in the basis set by the system-reservoir interaction instead of the system Hamiltonian. Several examples are discussed including a single qubit, a three-level V-system, and two coupled qubits all interacting with bosonic reservoirs. The results shed light on the presence of coherences in the strong coupling regime, and provide key tools for nanoscale thermodynamics investigations.
Two qubits strongly coupled to a common bosonic reservoir can become entangled with each other, despite having no direct interaction. In equilibrium, such coupling-induced coherences can be described by the mean-force Gibbs state. Here we derive approximate, analytic expressions for the two-qubit mean-force Gibbs state, and use these to characterize equilibrium qubit-qubit entanglement mediated by a thermal reservoir. Entanglement, which is highest at lowest temperatures, is a non-monotonic function of the system-reservoir coupling strength. Moreover, we find that broadening the reservoir spectral density beyond a single mode, as is realistic for typical baths, can enhance the qubit entanglement. Our results provide a comprehensive understanding of reservoir-mediated two-qubit entanglement in thermal equilibrium and provide a benchmark to compare with numerical methods, as well as demonstrating the utility of strong system-reservoir coupling as a resource.
An open quantum system interacting with a heat bath at a given temperature is expected to reach the mean force Gibbs (MFG) state as a steady state. The MFG state is given by tracing out the bath degrees of freedom from the equilibrium Gibbs state of the total system plus bath. When the interaction between the system and the bath is not negligible, it is different from the usual system Gibbs state obtained from the system Hamiltonian only. Using the path integral method, we present the exact MFG state for a coupled system of quantum harmonic oscillators in contact with multiple thermal baths at the same temperature. We develop a nonperturbative method to calculate the covariances with respect to the MFG state. By comparing them with those obtained from the system Gibbs state, we find that the effect of coupling to the bath decays exponentially as a function of the distance from the system-bath boundary. This is similar to the skin effect found recently for a quantum spin chain interacting with an environment. Using the exact results, we also investigate the ultrastrong coupling limit where the coupling between the system and the bath gets arbitrarily large and makes a connection with the recent result found for a general quantum system.
When the coupling of a quantum system to its environment is nonnegligible, its steady state is known to deviate from the textbook Gibbs state. The Bloch-Redfield quantum master equation, one of the most widely adopted equations for solving the open quantum dynamics, cannot predict all the deviations of the steady state of a quantum system from the Gibbs state. In this paper, for a generic spin-boson model, we use a higher-order quantum master equation (in system-environment coupling strength) to analytically calculate all the deviations of the steady state of the quantum system up to second order in coupling strength. We also show that this steady state is exactly identical to the corresponding generalized Gibbs state, the so-called quantum mean force Gibbs state, at arbitrary temperature. All these calculations are highly general, making them immediately applicable to a wide class of systems well modeled by the spin-Boson model, spanning a diverse range of topics, from nanomaterials to various condensed-phase processes, and quantum computing (e.g., environment-induced corrections to the steady state of a superconducting qubit). As an example, we use our results to study the dynamics and the steady state of a solid-state double-quantum-dot system under physically relevant choices of parameters.
The dynamical convergence of a system to the thermal distribution, or Gibbs state, is a standard assumption across all of the physical sciences. The Gibbs state is determined just by temperature and energies of the system. However, at decreasing system sizes, i.e., for nanoscale and quantum systems, the interaction with their environments is not negligible. The question then arises: Is the system's steady state still the Gibbs state? If not, how may the steady state depend on the interaction details? Here, we provide an overview of recent progress on answering these questions. We expand on the state of the art along two general avenues: First, we take the static point-of-view, which postulates the so-called mean force Gibbs state. This view is commonly adopted in the field of strong coupling thermodynamics, where modified laws of thermodynamics and nonequilibrium fluctuation relations are established on the basis of this modified state. Second, we take the dynamical point of view, originating from the field of open quantum systems, which examines the time-asymptotic steady state within two paradigms. We describe the mathematical paradigm, which proves return to equilibrium, i.e., convergence to the mean force Gibbs state, and then discuss a number of microscopic physical methods, particularly master equations. We conclude with a summary of established links between statics and equilibration dynamics and provide an extensive list of open problems. This comprehensive overview will be of interest to researchers in the wider fields of quantum thermodynamics, open quantum systems, mesoscopic physics, statistical physics, and quantum optics and will find applications whenever energy is exchanged on the nanoscale, from quantum chemistry and biology to magnetism and nanoscale heat management.
By solving the exact master equation of open quantum systems, we formulate the quantum thermodynamics from weak to strong couplings. The open quantum systems exchange matters, energies, and information with their reservoirs through quantum particle tunnelings that are described by the generalized Fano-Anderson Hamiltonians. We find that the exact solution of the reduced density matrix of these systems approaches a Gibbs-type state in the steady-state limit for the systems in arbitrary initial states as well as for both the weak and strong system-reservoir coupling strengths. When the couplings become strong, thermodynamic quantities of the system must be renormalized. The renormalization effects are obtained nonperturbatively after exactly tracing over all reservoir states through the coherent state path integrals. The renormalized system Hamiltonian is characterized by the renormalized system energy levels and interactions, corresponding to the quantum work done by the system. The renormalized temperature is introduced to characterize the entropy production counting the heat transfer between the system and the reservoir. We further find that only with the renormalized system Hamiltonian and other renormalized thermodynamic quantities can the exact steady state of the system be expressed as the standard Gibbs state. Consequently, the corresponding exact steady-state particle occupations in the renormalized system energy levels obey the Bose-Einstein and the Fermi-Dirac distributions for bosonic and fermionic systems, respectively. In the very weak system-reservoir coupling limit, the renormalized system Hamiltonian and the renormalized temperature are reduced to the original bare Hamiltonian of the system and the initial temperature of the reservoir. Thus, the conventional statistical mechanics and thermodynamics are thereby rigorously deduced from quantum dynamical evolution. In the last, this nonperturbative renormalization method is also extended to general interacting open quantum systems.