---
title: Kolmogorov-type Non-Thermal Fixed Points
url: https://www.emergentmind.com/topics/kolmogorov-type-non-thermal-fixed-points
type: topic
---

# Kolmogorov-type Non-Thermal Fixed Points

Searching arXiv for relevant papers on Kolmogorov-type non-thermal fixed points.
Kolmogorov-type non-thermal fixed points are far-from-equilibrium scaling states in which a many-body system exhibits universal, non-thermal power laws together with directed transport of conserved quantities across scales. In dilute Bose gases, the concept was made concrete by relating non-thermal fixed points to superfluid turbulence, quantized vortices, and acoustic excitations, with the single-particle momentum distribution $n(k)\sim k^{-\zeta}$ serving as a central diagnostic [1012.4437]. In this framework, the fixed point is “non-thermal” because the distribution is not of Rayleigh–Jeans or Bose–Einstein form, yet it is “fixed” in the sense of stationary or quasi-stationary scaling under the nonequilibrium evolution equations. The qualifier “Kolmogorov-type” refers to the structural analogy with turbulence theory: universal scaling laws emerge from conservation laws, dimensionality, and scale-to-scale transport, and in suitable observables the incompressible kinetic-energy spectrum exhibits the Kolmogorov law $E(k)\sim k^{-5/3}$ [1012.4437].

## 1. Definition and scaling characterization

A non-thermal fixed point is a metastable far-from-equilibrium configuration of an isolated many-body system characterized by stationary scaling of correlation functions and momentum distributions in some infrared or ultraviolet regime, together with self-similarity of the evolution near the fixed point [1302.1448]. For a Bose gas, the central observable is the single-particle momentum distribution
\[
n(\mathbf{k}, t) = \langle \psi^\dagger(\mathbf{k}, t)\psi(\mathbf{k}, t)\rangle ,
\]
and its radial average $n(k,t)$ with $k=|\mathbf{k}|$ [1012.4437]. The fixed-point condition is expressed by the scaling law
\[
n(sk) = s^{-\zeta} n(k),
\]
equivalently $n(k)\sim k^{-\zeta}$ over a finite range of momenta [1012.4437].

In a broader dynamical formulation, non-thermal fixed points are associated with self-similar spatio-temporal scaling,
\[
\bar n(k,t)=\left(\frac{t}{t_0}\right)^\alpha f\!\left[\left(\frac{t}{t_0}\right)^\beta k\right],
\]
or, equivalently in related formulations,
\[
n(k,t)=t^\alpha f(t^\beta k),
\]
with universal exponents $\alpha$ and $\beta$ constrained by conservation laws [2203.14752]. In Bose-gas experiments and simulations, this scaling is interpreted as an attractor solution in the space of nonequilibrium states, closely analogous to renormalization-group fixed points in equilibrium critical phenomena, but reached from strongly out-of-equilibrium initial conditions rather than by fine tuning control parameters [2203.14752].

Kolmogorov-type behavior enters when the scaling is understood as the manifestation of a cascade. In this sense, a Kolmogorov-type non-thermal fixed point is a stationary or quasi-stationary scaling regime maintained by a constant directional flux of a conserved quantity, typically particle number or energy, across scales [2203.14752]. This definition encompasses both classical and quantum turbulence, while distinguishing non-thermal fixed points from thermal fixed points, where detailed balance eliminates net fluxes and the distribution reduces to a Gibbsian form [1302.1448].

## 2. Field-theoretic origin and universality

The modern formulation of non-thermal fixed points was developed in nonequilibrium quantum field theory using the two-particle irreducible effective action, Kadanoff–Baym equations, and later functional renormalization group methods [0809.5208]. In this language, non-thermal fixed points are stationary solutions of the full dynamical equations for statistical and spectral correlators that do not satisfy a fluctuation–dissipation relation and therefore lie outside thermal equilibrium [0809.5208]. For an $O(N)$ scalar theory, the functional renormalization group reveals a hierarchy of fixed-point solutions that includes vacuum, thermal equilibrium, and nonequilibrium fixed points, with the nonequilibrium solutions admitting scaling exponents that explicitly depend on spatial dimension $d$ [0809.5208].

For dilute Bose gases, the nonperturbative 2PI analysis yields a central infrared prediction: for constant radial quasiparticle flux, the occupation spectrum scales as
\[
\zeta^{\mathrm{IR}}_Q=d+2,
\]
where $d$ is the spatial dimension [1012.4437]. In contrast, thermal Rayleigh–Jeans scaling for quadratic dispersion gives $\zeta=2$ in the ultraviolet [1012.4437]. The infrared exponent $d+2$ is therefore the characteristic signature of a non-thermal rather than thermal fixed point. The same exponent appears in relativistic scalar field models, which supports universality across relativistic and nonrelativistic systems [1012.4437].

A complementary low-energy effective field theory for multicomponent $\mathrm{U}(N)$ Bose gases integrates out density fluctuations and describes interacting Goldstone modes of total and relative phases. In the large-$N$ limit, this theory predicts the same self-similar scaling exponents associated with the non-thermal fixed point and identifies a Gaussian fixed point dominated by relative phases, while anticipating a non-Gaussian fixed point for evolution with a distinctly lower power of time [1807.10228]. This suggests that Kolmogorov-type non-thermal fixed points are not tied to a single microscopic derivation, but recur across kinetic, effective-theory, and renormalization-group descriptions.

## 3. Cascades, spectra, and Kolmogorov-type laws

The canonical turbulent interpretation starts from continuity equations in momentum space for radial particle and energy densities. In weak wave turbulence, constant particle and energy fluxes define Kolmogorov–Zakharov spectra; for a $d$-dimensional Bose gas with quadratic dispersion, the weak-turbulence ultraviolet exponents are
\[
\zeta_Q^{\mathrm{UV}}=d-\frac{2}{3},\qquad \zeta_P^{\mathrm{UV}}=d,
\]
while the strong-turbulence infrared exponents become
\[
\zeta_Q^{\mathrm{IR}}=d+2,\qquad \zeta_P^{\mathrm{IR}}=d+2+z,
\]
with dynamical exponent $z$ [1302.1448]. For the nonrelativistic Bose gas, $z=2$, so the infrared particle-cascade exponents are $\zeta_Q^{\mathrm{IR}}=4$ in $d=2$ and $\zeta_Q^{\mathrm{IR}}=5$ in $d=3$ [1302.1448].

The concrete Bose-gas simulations that established the link to superfluid turbulence observed precisely these values: in $d=2$, an infrared law $n(k)\sim k^{-4}$ and in $d=3$, $n(k)\sim k^{-5}$, together with a thermalized high-$k$ tail $n(k)\sim k^{-2}$ [1012.4437]. The ultraviolet sector can additionally display a weak-wave-turbulence scaling with $\zeta^{\mathrm{UV}}=d$ in suitable parameter regimes [1012.4437].

The Kolmogorov analogy becomes explicit upon decomposing the kinetic energy into incompressible, compressible, and quantum-pressure parts. With $\psi=\sqrt{n}\,e^{i\phi}$ and superfluid velocity $\mathbf{v}=\nabla\phi$, one defines generalized velocities
\[
\mathbf{w}_{\mathrm v}=\sqrt{n}\,\mathbf{v},\qquad \mathbf{w}_{\mathrm q}=\nabla\sqrt{n},
\]
and decomposes $\mathbf{w}_{\mathrm v}$ into incompressible and compressible components, $\mathbf{w}_{\mathrm i}$ and $\mathbf{w}_{\mathrm c}$ [1012.4437]. The associated radial energy spectra are related to occupation-like spectra through
\[
n_\delta(k)=k^{-d-1}E_\delta(k),\qquad \delta\in\{\mathrm i,\mathrm c,\mathrm q\}.
\]
In the turbulent regime, the incompressible component develops a kinetic-energy spectrum consistent with
\[
E_{\mathrm i}(k)\sim k^{-5/3},
\]
which is the direct quantum analogue of Kolmogorov’s inertial-range law [1012.4437].

This structural relation between fixed-point scaling and cascade transport is also borne out experimentally. In a homogeneous three-dimensional Bose gas, the ultraviolet scaling exponents $\alpha_{\rm UV}=-0.70(7)$ and $\beta_{\rm UV}=-0.14(2)$ satisfy $\alpha/\beta\approx 5$, in agreement with the energy-conservation constraint $\alpha/\beta=d+z$ for $d=3$ and $z=2$, while the measured $\beta_{\rm UV}=-0.14(2)$ is consistent with the weak-wave-turbulence prediction $\beta=-1/6$ [2203.14752]. This identifies a direct energy cascade toward high momenta, coexisting with an infrared particle cascade.

## 4. Vortices, defect statistics, and defect-controlled scaling

A defining advance in the Bose-gas literature was the identification of the infrared non-thermal fixed point with a gas of quantized vortices. In the Madelung representation, vortices correspond to phase singularities around points in two dimensions or lines and loops in three dimensions where the density vanishes [1012.4437]. Their velocity profile scales as $|\mathbf v|\sim 1/r$, and after regularization by the factor $\sqrt{n}$ the resulting generalized velocity $\mathbf{w}_{\mathrm v}$ remains finite [1012.4437].

For a finite density of independent vortices and antivortices in $d=2$, the superposition of long-range vortex velocity fields yields
\[
n(k)\sim k^{-4},
\]
while in $d=3$ independent vortex lines and rings yield
\[
n(k)\sim k^{-5}
\]
[1111.6127]. These exponents coincide with the strong-turbulence prediction $\zeta_Q^{\mathrm{IR}}=d+2$ and thereby provide a direct physical interpretation of the infrared fixed point in terms of defect statistics [1111.6127].

The same paper demonstrates that pair correlations alter the scaling. In two dimensions, a vortex–antivortex pair behaves as a dipole whose far-field velocity decays faster, leading to
\[
n(k)\sim k^{-2}
\]
at sufficiently small $k$ [1111.6127]. In three dimensions, small rings or strongly correlated line segments produce analogous weakening of the infrared exponent, such as $n(k)\sim k^{-3}$ in appropriate regimes [1111.6127]. The long-time departure from the non-thermal fixed point is therefore associated with vortex–antivortex pairing and annihilation rather than with a mere loss of scaling [1111.6127].

This defect-based interpretation was generalized in later work. In a two-component Bose gas, universal scaling exponents differ across the miscible–immiscible transition because the dominant defects change from vortices and skyrmions to domain walls and spin textures [1307.7368]. In that setting, vortices generate incompressible spectra $n_i(k)\sim k^{-4}$, while domain walls yield spin-pressure spectra $n_s(k)\sim k^{-3}$, and the total single-particle spectrum reflects the superposition of these defect sectors [1307.7368]. More generally, the geometry of the dominant defects determines the scaling law: vortices in $d$ dimensions produce $n(k)\sim k^{-(d+2)}$, while solitons or domains produce $n(k)\sim k^{-(d+1)}$ [1302.1448].

## 5. Acoustic sectors, multibranch fixed points, and multiple attractors

Kolmogorov-type non-thermal fixed points are not exhausted by incompressible vortex turbulence. The compressible sector can itself exhibit a non-thermal power law. In the ultracold Bose-gas simulations, the compressible contribution follows
\[
n_{\mathrm c}(k)\sim k^{-(d+1)},
\]
which is interpreted as acoustic wave turbulence driven by long-wavelength sound modes on top of the vortical background [1012.4437]. These acoustic excitations can survive even after vortices have annihilated, indicating a distinct but related universality sector of the non-thermal fixed-point dynamics [1012.4437].

The possibility of multiple non-thermal fixed points for the same microscopic symmetry was demonstrated experimentally in a quasi-one-dimensional spinor Bose gas. Two distinct initial conditions in the same easy-plane ferromagnetic phase yield two different scaling regimes with constrained exponents
\[
\beta_{\text{polar}}=\alpha_{\text{polar}}=0.54 \pm 0.02_{\rm stat}\pm 0.05_{\rm sys},
\]
and
\[
\beta_{\text{rot}}=\alpha_{\text{rot}}=0.28 \pm 0.03_{\rm stat}\pm 0.04_{\rm sys},
\]
together with different spectral tail exponents and different dominant excitations, namely phase modes in one case and phase-amplitude defects in the other [2306.16497]. This establishes that symmetry alone does not uniquely determine far-from-equilibrium universality; basins of attraction and defect content matter as well [2306.16497]. A plausible implication is that “Kolmogorov-type” should be understood as a family of scaling attractors rather than a single fixed point.

A related, more abstract generalization appears in a 2026 multibranch shell model of turbulent cascades, where a continuum of stationary forward-cascade fixed points exists, including the Kolmogorov solution $r_{l,n}\equiv 1$ with $|u_{l,n}|\propto k_l^{-1/3}$ [2601.04788]. In that model, the Kolmogorov fixed point is one member of a manifold of non-thermal cascade states distinguished by multiplier statistics and intermittency exponents [2601.04788]. While this shell-model analysis is not a Bose-gas result, it reinforces the broader interpretation of Kolmogorov-type non-thermal fixed points as stationary cascade solutions, often organized into a family or manifold rather than an isolated point [2601.04788].

## 6. Dynamics of approach, stability, and experimental realizations

The dynamical route to a non-thermal fixed point typically begins with a far-from-equilibrium state involving strong overpopulation of low-momentum modes or violent phase perturbations. In dilute Bose gases governed semiclassically by the Gross–Pitaevskii equation,
\[
i\partial_t \psi(\mathbf{x},t) = \left[ -\frac{\nabla^2}{2m} + g|\psi(\mathbf{x},t)|^2 \right] \psi(\mathbf{x},t),
\]
such initial conditions produce shock-wave-like fronts, defect nucleation, and the emergence of vortex tangles or soliton trains [1012.4437]. The system then enters a quasi-stationary regime with bimodal scaling, critical slowing down, and long-lived power laws before ultimately thermalizing [1302.1448].

Experimental evidence for this picture is now extensive. A quasi-one-dimensional Bose gas of $^{87}$Rb atoms showed infrared scaling exponents
\[
\alpha=0.09(5),\qquad \beta=0.10(4),
\]
with $\alpha\approx d\beta$ for $d=1$, consistent with particle-conserving infrared scaling [2203.14752]. A homogeneous three-dimensional Bose gas of $^{39}$K exhibited simultaneous infrared and ultraviolet scaling,
\[
\alpha_{\rm IR}=1.15(8),\quad \beta_{\rm IR}=0.34(5),\qquad
\alpha_{\rm UV}=-0.70(7),\quad \beta_{\rm UV}=-0.14(2),
\]
providing a clean experimental realization of bidirectional cascades and a Kolmogorov-type dual-fixed-point structure [2203.14752]. In turbulent harmonically trapped three-dimensional condensates, exponents extracted from the reconstructed three-dimensional distributions were
\[
\alpha_{\rm 3D}=-0.75,\qquad \beta=-0.2,
\]
indicating ultraviolet-dominated transport associated with quantum turbulence [2203.14752].

More recently, a quasi-one-dimensional Bose gas of strongly interacting $^6$Li$_2$ Feshbach molecules, driven far from equilibrium by imprinting a white-noise phase profile, was found to relax with the same exponents as the earlier weakly interacting $^{87}$Rb system, namely
\[
\alpha=0.09\pm 0.01,\qquad \beta=0.10\pm 0.01,
\]
in a common quasi-one-dimensional scaling window [2505.20213]. This points to a single universal fixed point with a large basin of attraction governing the relaxation of quasi-one-dimensional bosonic systems, largely independent of microscopic details [2505.20213].

Stability analysis in other nonequilibrium quantum field theories supports the same attractor interpretation. In longitudinally expanding kinetic theory for a non-Abelian plasma, linearization about the BMSS non-thermal fixed point yields negative eigenvalues
\[
\lambda_1=-2,\qquad \lambda_2=-\frac{4}{3},
\]
demonstrating linear stability and algebraic relaxation toward the fixed point [2203.02299]. Although this result concerns an expanding plasma rather than a Bose gas, it supplies a technical example of how non-thermal fixed points function as dynamical attractors with calculable relaxation rates [2203.02299].

## 7. Broader extensions, anomalous cases, and open problems

The concept of Kolmogorov-type non-thermal fixed points now spans nonrelativistic Bose gases, relativistic scalar field theories, non-Abelian plasmas, holographic superfluids, shell models, and even dilute Fermi gases [1607.02160]. In a relativistic $O(N)$ scalar theory solved beyond the weak-coupling classical-statistical limit, non-thermal fixed points persist up to couplings of order unity, with an infrared inverse particle cascade
\[
f(p)\sim p^{-\kappa_{\rm N}},\qquad \kappa_{\rm N}\approx 5,
\]
and, in the symmetric case, a direct energy cascade
\[
f(p)\sim p^{-\kappa_{\rm E}},\qquad \kappa_{\rm E}=\frac{5}{3},
\]
demonstrating that Kolmogorov-type scaling survives beyond weak coupling [1607.02160].

A holographic superfluid in $2+1$ dimensions supplies a strongly coupled realization. Starting from various vortex ensembles, the system exhibits intermediate-time Kolmogorov scaling whose emergence depends on initial conditions, followed by a universal late-time regime where the occupation spectrum scales as $n(k,t)\sim k^{-\zeta}$ with $4.1\lesssim \zeta \lesssim 4.3$ and characteristic length scales grow as $t^{1/2}$ [1410.3472]. The late-time universal regime is interpreted as a non-thermal fixed point of the dynamical evolution, and, in the gravity dual, as a stationary point of the classical bulk equations [1410.3472].

Recent work on decaying two-dimensional superfluid turbulence near an anomalous non-thermal fixed point sharpens the distinction between spatial and temporal universality. During a universal interval with inter-defect distance
\[
\ell_{\mathrm v}(t)\sim t^\beta,\qquad \beta\approx \frac15,
\]
moments of the superfluid velocity circulation obey
\[
\Gamma^p(r)\sim r^{4p/3},
\]
with intermittency corrections for higher $p$ consistent with values measured in fully developed classical turbulence [2509.21285]. The spatial structure is therefore Kraichnan–Kolmogorov-like, while the temporal coarsening exponent is anomalously slow and distinctly quantum [2509.21285]. This suggests that Kolmogorov-type non-thermal fixed points need not share classical coarsening exponents even when their spatial turbulence statistics do.

Several open issues recur across the literature. A complete classification of non-thermal universality classes remains unavailable [2203.14752]. The role of dimensionality, internal symmetry, and defect content remains only partially understood, especially in multicomponent and low-dimensional systems [2306.16497]. The crossover from the vicinity of a non-thermal fixed point to final thermalization is also an active problem, as is the precise relation between defect-mediated coarsening, wave turbulence, and hydrodynamic attractors [1302.1448]. Inference beyond the cited results suggests that the encyclopedia entry “Kolmogorov-type non-thermal fixed point” properly denotes a unifying concept rather than a single model-specific phenomenon: stationary or quasi-stationary cascade states, often defect-mediated, characterized by universal spatio-temporal scaling and organized by conservation laws rather than by thermal detailed balance.

Source: https://www.emergentmind.com/topics/kolmogorov-type-non-thermal-fixed-points