7.4 Onsager relations

Fundamentals of Onsager relations

Onsager relations describe how systems behave when they're slightly pushed out of equilibrium. They connect a deep microscopic fact (that the laws of physics look the same if you reverse time) to the macroscopic irreversible processes we actually observe, like heat conduction and diffusion. The payoff is a set of symmetry constraints on transport coefficients that reduce the number of independent quantities you need to measure and reveal unexpected connections between seemingly different phenomena.

Reciprocity theorem

The reciprocity theorem states that the matrix of transport coefficients linking thermodynamic forces to fluxes is symmetric:

$L_{ij} = L_{ji}$

where $L_{ij}$ are the Onsager coefficients (also called kinetic coefficients or phenomenological coefficients). This means the influence of force $j$ on flux $i$ equals the influence of force $i$ on flux $j$.

• Applies only in the linear regime, i.e., for systems close to equilibrium where higher-order terms in the force-flux expansion are negligible.
• Derived from the principle of microscopic reversibility and time-reversal symmetry of the underlying dynamics.
• Predicts cross-phenomena: for example, the Seebeck coefficient (voltage from a temperature gradient) and the Peltier coefficient (heat flow from an electric current) are not independent but related through $L_{ij} = L_{ji}$.

Linear response theory

Linear response theory provides the statistical-mechanical foundation for the Onsager framework. It asks: if you apply a small external perturbation to an equilibrium system, how does the system respond?

• The response is assumed to be linear in the applied perturbation. This is valid when the perturbation is small compared to thermal energy scales.
• The response of a macroscopic observable is expressed through time-correlation functions of microscopic fluctuations evaluated at equilibrium.
• This leads directly to the Green-Kubo relations, which express transport coefficients as time integrals of equilibrium autocorrelation functions. For example, the electrical conductivity $\sigma$ can be written as an integral of the current-current correlation function.
• The theory applies broadly: electrical conductivity, thermal conductivity, viscosity, and diffusion coefficients can all be obtained this way.

Microscopic reversibility principle

At the microscopic level, the equations of motion (Newton's laws, Schrödinger's equation without explicit time-dependent potentials) are invariant under time reversal: replacing $t \to -t$ and reversing all momenta yields equally valid trajectories.

• In equilibrium, the probability of observing a particular microscopic trajectory equals the probability of observing its time-reversed counterpart. This is the detailed balance condition.
• This does not contradict macroscopic irreversibility. The second law emerges statistically because the overwhelming majority of microstates consistent with a non-equilibrium macrostate evolve toward equilibrium.
• Microscopic reversibility is the physical ingredient that makes the Onsager matrix symmetric. Without it, there would be no reason to expect $L_{ij} = L_{ji}$.

Thermodynamic forces and fluxes

Non-equilibrium thermodynamics organizes transport phenomena into pairs: a thermodynamic force (the "cause") and a thermodynamic flux (the "effect"). Identifying the correct conjugate pairs is essential because the Onsager formalism only works when forces and fluxes are chosen consistently with the entropy production.

Generalized forces

Generalized forces $X_i$ represent the thermodynamic driving factors that push a system away from equilibrium.

• They are typically gradients (or differences) of intensive thermodynamic variables. For heat conduction, the force is $\nabla(1/T)$; for diffusion, it involves $\nabla(\mu/T)$, where $\mu$ is the chemical potential.
• More precisely, they are derived from the entropy production expression: you write $\sigma = \sum_i J_i X_i$ and read off the forces as the quantities conjugate to each flux.
• Forces can be scalar (chemical affinity driving a reaction) or vectorial (a temperature gradient driving heat flow).

Thermodynamic fluxes

Thermodynamic fluxes $J_i$ describe the rates of flow of extensive quantities in response to the generalized forces.

• Examples: heat flux (energy flow per unit area per unit time), particle flux (number of particles crossing a surface per unit time), electric current density.
• Fluxes are the quantities you measure experimentally to extract transport coefficients.
• Each flux is conjugate to a specific force, and the pairing must be consistent with the entropy production formula.

Force-flux relationships

In the linear regime near equilibrium, each flux depends linearly on all the forces:

$J_i = \sum_j L_{ij} X_j$

• Diagonal coefficients $L_{ii}$ describe direct effects: a temperature gradient drives heat flow (Fourier's law), a concentration gradient drives diffusion (Fick's law).
• Off-diagonal coefficients $L_{ij}$ (with $i \neq j$) describe cross-effects: a temperature gradient driving particle flow (Soret effect), or an electric field driving heat flow (Peltier effect).
• The Onsager reciprocal relation $L_{ij} = L_{ji}$ cuts the number of independent off-diagonal coefficients roughly in half, which is a significant simplification for multicomponent systems.

Linear irreversible thermodynamics

This framework takes the force-flux structure and the Onsager symmetry as its starting point to describe irreversible processes systematically. It's the workhorse theory for transport near equilibrium.

Near-equilibrium systems

• The system must be close enough to equilibrium that the linear approximation $J_i = \sum_j L_{ij} X_j$ holds. In practice, this means small gradients in temperature, chemical potential, etc.
• Local equilibrium is assumed: at each point in space, thermodynamic variables ($T$, $\mu$, $P$) are well-defined, even though they vary from point to point. This breaks down for very steep gradients or very fast processes.
• The Onsager relations are most accurate and most useful in this regime.

Entropy production

The rate of entropy production per unit volume is:

$\sigma = \sum_i J_i X_i$

• This quantity is always non-negative ($\sigma \geq 0$), consistent with the second law of thermodynamics.
• It serves as a bilinear form in forces and fluxes, and the correct conjugate pairs $(J_i, X_i)$ are defined so that this expression gives the total dissipation.
• In a steady state, the system often arranges itself to minimize entropy production subject to external constraints (Prigogine's minimum entropy production principle, valid in the linear regime).

Onsager coefficients

The coefficients $L_{ij}$ encode all the transport physics in the linear regime.

• Symmetry: $L_{ij} = L_{ji}$ (Onsager reciprocal relations).
• Positive semi-definiteness: The matrix $L$ must be positive semi-definite to guarantee $\sigma \geq 0$. This means all eigenvalues are non-negative, and the diagonal elements satisfy $L_{ii} \geq 0$.
• They can be determined experimentally (by measuring fluxes under controlled forces) or computed from equilibrium correlation functions via the Green-Kubo relations.

Symmetry in transport coefficients

The symmetry $L_{ij} = L_{ji}$ is not just a mathematical convenience. It reflects a deep physical symmetry and has practical consequences for predicting and measuring cross-effects.

Cross-phenomena effects

Cross-phenomena arise from the off-diagonal Onsager coefficients and represent some of the most interesting predictions of the theory.

• Seebeck effect: A temperature gradient across a conductor produces a voltage. The relevant coefficient is $L_{12}$ (coupling thermal force to charge flux).
• Peltier effect: An electric current through a junction of two materials causes heating or cooling. The relevant coefficient is $L_{21}$ (coupling electrical force to heat flux).
• Onsager reciprocity tells you these are related: $L_{12} = L_{21}$, which connects the Seebeck and Peltier coefficients through $\Pi = ST$, where $\Pi$ is the Peltier coefficient, $S$ is the Seebeck coefficient, and $T$ is the absolute temperature.
• Other examples include the Soret effect (thermal diffusion), the Dufour effect (diffusion-driven heat flow), and the Nernst effect (thermomagnetic voltage).

Curie principle

The Curie principle imposes additional constraints based on the tensorial character of forces and fluxes.

• In an isotropic system (no preferred direction), forces and fluxes of different tensorial rank do not couple. A scalar force (like a chemical affinity) cannot drive a vectorial flux (like heat flow) in an isotropic medium.
• This reduces the number of independent Onsager coefficients further, simplifying the transport description considerably.
• The principle applies when there are no external fields (magnetic field, rotation) that break isotropy.

Onsager-Casimir relations

When an external magnetic field $\mathbf{B}$ or system rotation is present, time-reversal symmetry is modified. Reversing time also reverses $\mathbf{B}$ (and angular velocity), so the standard Onsager relation becomes:

$L_{ij}(\mathbf{B}) = L_{ji}(-\mathbf{B})$

• The transport matrix is no longer symmetric at a given field; instead, symmetry relates coefficients at opposite fields.
• This is essential for understanding magneto-transport phenomena like the Hall effect, where a transverse voltage appears in a current-carrying conductor placed in a magnetic field.
• These relations also apply to transport in superconductors and other systems where time-reversal symmetry is explicitly broken.

Applications of Onsager relations

Thermoelectric effects

Thermoelectric effects are the most classic application of Onsager relations because they involve two coupled fluxes (heat and charge) driven by two forces (temperature gradient and electric field).

• The Seebeck effect generates a voltage from a temperature difference. The Seebeck coefficient $S$ is typically on the order of microvolts per kelvin for metals and much larger for semiconductors.
• The Peltier effect produces heating or cooling at a junction when current flows. Onsager reciprocity predicts $\Pi = ST$, and this has been confirmed experimentally to high precision.
• The thermoelectric figure of merit $ZT = S^2 \sigma T / \kappa$ (where $\sigma$ is electrical conductivity and $\kappa$ is thermal conductivity) determines the efficiency of thermoelectric devices. Optimizing $ZT$ is an active area of materials research.

Diffusion processes

• Fick's law ($J = -D \nabla c$) is a special case of the Onsager formalism with a single force-flux pair.
• In multicomponent systems, cross-diffusion becomes important: the flux of species $i$ depends on the concentration gradients of all species, not just its own. The off-diagonal diffusion coefficients satisfy Onsager symmetry.
• The Soret effect (thermal diffusion) describes particle migration driven by a temperature gradient. Its reciprocal, the Dufour effect, is heat flow driven by a concentration gradient. Both are captured by off-diagonal $L_{ij}$.
• Applications span materials science (alloy solidification), biology (membrane transport, protein separation), and geophysics (mantle convection).

Chemical reactions

• For chemical reactions near equilibrium, the "force" is the chemical affinity $A = -\Delta G / T$ and the "flux" is the reaction rate.
• When multiple reactions are coupled, the Onsager matrix describes how the affinity of one reaction can drive another. This is relevant in biochemical cycles where endergonic reactions are driven by exergonic ones (e.g., ATP hydrolysis driving biosynthesis).
• The linear approximation holds only for small affinities ($A \ll k_B T$), which limits applicability to reactions very close to equilibrium.

Limitations and extensions

Non-linear regimes

When driving forces become large, the linear relation $J_i = \sum_j L_{ij} X_j$ breaks down. Higher-order terms in the force-flux expansion become significant.

• Non-linear effects can produce qualitatively new phenomena: pattern formation (Rayleigh-Bénard convection cells), self-organization (dissipative structures), and oscillations (Belousov-Zhabotinsky reaction).
• There is no general analog of the Onsager reciprocal relations in the non-linear regime. Each system must be analyzed on its own terms.
• Methods from non-linear dynamics, bifurcation theory, and numerical simulation are used instead.

Far-from-equilibrium systems

• Characterized by large gradients, strong driving, or rapid temporal changes where even the concept of local equilibrium may fail.
• Examples: turbulent flows, laser-driven plasmas, strongly driven biochemical networks, active matter (self-propelled particles).
• Theoretical frameworks for these systems include extended irreversible thermodynamics (which promotes fluxes to independent variables) and stochastic thermodynamics (which tracks entropy production along individual fluctuating trajectories).

Fluctuation-dissipation theorem

The fluctuation-dissipation theorem (FDT) generalizes the connection between fluctuations and response that underlies Onsager's work.

• It states that the linear response of a system to an external perturbation is determined by the time-correlation function of spontaneous fluctuations in equilibrium.
• Mathematically, the imaginary part of the response function $\chi''(\omega)$ is related to the spectral density of fluctuations by a factor involving temperature.
• The FDT applies to both classical and quantum systems and is foundational for understanding noise in electrical circuits (Johnson-Nyquist noise), Brownian motion, and many other phenomena.
• It breaks down in non-equilibrium steady states, and characterizing how it breaks down is an active research topic connected to fluctuation theorems.

Experimental verification

Measurement techniques

• Thermoelectric measurements are the most direct test: measure the Seebeck coefficient and the Peltier coefficient independently, then check whether $\Pi = ST$.
• Diffusion coefficients in multicomponent mixtures can be measured using interferometric techniques (e.g., Gouy interferometry) or Taylor dispersion methods, and the symmetry of the diffusion matrix can be verified.
• Neutron scattering and dynamic light scattering probe microscopic correlation functions, providing a route to verify Green-Kubo predictions for transport coefficients.
• Microfluidic devices allow controlled studies of coupled diffusion and thermal diffusion in small volumes.

Case studies

• Thermoelectric materials: The Seebeck-Peltier reciprocity $\Pi = ST$ has been confirmed in metals and semiconductors across a wide temperature range.
• Multicomponent diffusion: Measurements in ternary liquid mixtures (e.g., water-salt-polymer systems) have verified the symmetry of the cross-diffusion coefficient matrix.
• Biological membranes: Onsager symmetry has been tested in ion transport through channels and synthetic membranes, with good agreement in the linear regime.
• Magneto-transport: Hall effect measurements confirm the Onsager-Casimir relation $L_{ij}(B) = L_{ji}(-B)$.

Challenges in validation

• Keeping the system close enough to equilibrium for linear theory to apply while still generating fluxes large enough to measure accurately. This is a fundamental tension.
• Isolating a specific cross-effect from other competing transport processes in a real material.
• Achieving the precision needed to confirm $L_{ij} = L_{ji}$ when the off-diagonal coefficients are small compared to the diagonal ones.
• Extending verification to quantum systems and mesoscopic devices, where fluctuations are large and the classical framework may need modification.

Mathematical formalism

Matrix representation

The full set of linear force-flux relations is compactly written as:

$\mathbf{J} = \mathbf{L} \cdot \mathbf{X}$

where $\mathbf{J}$ is the flux vector, $\mathbf{X}$ is the force vector, and $\mathbf{L}$ is the Onsager matrix.

• $\mathbf{L}$ is symmetric ($L_{ij} = L_{ji}$) and positive semi-definite.
• The eigenvalues of $\mathbf{L}$ correspond to the principal transport coefficients, and the eigenvectors define the principal modes of transport (linear combinations of fluxes that decouple from each other).
• Diagonalizing $\mathbf{L}$ can simplify the analysis of coupled transport problems considerably.

Time-reversal symmetry

• Under time reversal ($t \to -t$), coordinates are unchanged but momenta (and magnetic fields, angular velocities) reverse sign.
• Onsager's derivation uses the fact that equilibrium time-correlation functions satisfy $\langle a_i(0) a_j(\tau) \rangle = \langle a_i(0) a_j(-\tau) \rangle$ for variables that are even under time reversal. This symmetry of the correlation function directly implies $L_{ij} = L_{ji}$.
• For variables with definite parity under time reversal (e.g., velocity is odd, position is even), the relation picks up a sign: $L_{ij} = \epsilon_i \epsilon_j L_{ji}$, where $\epsilon_i = \pm 1$ is the time-reversal parity of variable $i$. This is the more general form that leads to the Onsager-Casimir relations.

Fluctuation theory

• Green-Kubo relations express each transport coefficient as a time integral of an equilibrium autocorrelation function:

$L_{ij} = \frac{1}{k_B} \int_0^\infty \langle \dot{a}_i(0) \dot{a}_j(t) \rangle_{\text{eq}} \, dt$

where $\dot{a}_i$ is the time derivative of the fluctuating variable conjugate to force $X_i$.

• These relations provide a direct computational route to transport coefficients via molecular dynamics simulations.
• Fluctuation theorems (Jarzynski equality, Crooks relation) extend these ideas beyond the linear regime, relating free energy differences to non-equilibrium work distributions. They represent the modern generalization of Onsager's original insights.

Historical context

Onsager's contributions

Lars Onsager published the reciprocal relations in two papers in 1931. He was awarded the Nobel Prize in Chemistry in 1968, though the citation was for his work on irreversible processes more broadly.

• Onsager built on earlier observations by Thomson (Lord Kelvin) and Helmholtz, who had noticed reciprocal relationships in specific thermoelectric and electrochemical systems but lacked a general derivation.
• His key insight was to use the regression hypothesis: spontaneous fluctuations around equilibrium decay, on average, according to the same macroscopic transport laws that govern the relaxation of externally imposed perturbations. Combined with microscopic reversibility, this yields the symmetry of $\mathbf{L}$.

Development of non-equilibrium thermodynamics

• Prigogine (Nobel Prize 1977) extended the framework to include chemical reactions and formulated the minimum entropy production principle for linear steady states.
• Green (1954) and Kubo (1957) independently developed the fluctuation-dissipation formalism, providing a microscopic derivation of transport coefficients from equilibrium correlation functions.
• Extended irreversible thermodynamics (Jou, Casas-Vázquez, Lebon) was developed from the 1960s onward to handle fast processes and short wavelengths where local equilibrium fails.
• Stochastic thermodynamics (Sekimoto, Seifert, Jarzynski, Crooks) emerged in the 1990s-2000s, extending thermodynamic concepts to individual fluctuating trajectories in small systems.

Modern perspectives

• Onsager relations are now understood as a consequence of the more general fluctuation theorems, which hold arbitrarily far from equilibrium. In the linear limit, fluctuation theorems reduce to the Onsager reciprocal relations and the Green-Kubo formulas.
• Active areas of application include quantum transport in mesoscopic devices, biological molecular motors, and the design of efficient thermoelectric and photovoltaic materials.
• The framework continues to be extended to active matter (systems with internal energy sources, like bacterial suspensions), where standard Onsager relations do not apply directly but modified fluctuation-dissipation relations can sometimes be formulated.