betaFeedback

WorksMathematics

DLR variational principle for the one-dimensional log-gas and another proof of existence of Sine-beta

Abstract

We show that every stationary solution of the Dobrushin–Lanford–Ruelle (DLR) equations for the one-dimensional log-gas with a quantitative control on its number variance minimizes the free energy functional. This yields a DLR variational principle for this continuous long-range system. Moreover, we show that any limit point of the finite-volume log-gases is stationary and satisfies this condition. This allows us to give a new, physical proof of the existence of a microscopic limit for the Gaussian Beta Ensembles, known as the Sine-beta point process.

Introduction

The Gaussian β\beta-ensembles

The Gaussian β\beta-ensembles form a family, indexed by β∈(0,+∞)\beta\in(0,+\infty), of probability densities given by:

dPN,β(x1,…,xN):=1ZN,β∏1≤i<j≤N∣xi−xj∣βexp⁡(−β4N∑i=1Nxi2)dx1…dxN.(1)\mathrm{d}\mathbb{P}_{N,\beta}(x_1,\ldots,x_N) := \frac{1}{Z_{N,\beta}} \prod_{1\leq i<j\leq N} |x_i-x_j|^\beta\exp\left(-\frac{\beta}{4N}\sum_{i=1}^{N}x_i^2\right)\mathrm{d}x_1\ldots\mathrm{d}x_N . \tag*{(1)}

where ZN,βZ_{N,\beta} is a normalizing constant and dx1…dxN\mathrm{d}x_1\ldots\mathrm{d}x_N is the Lebesgue measure on RN\mathbb{R}^N. They describe the joint law of eigenvalues in important models of random matrix theory. Historically speaking, the three classical values were β∈{1,2,4}\beta\in\{1,2,4\} [19], but in a celebrated work, Dumitriu and Edelman [14] constructed, for every β>0\beta>0, tridiagonal symmetric matrices with independent entries whose eigenvalues have joint law (1.1). This constitutes the “random matrix interpretation” of the measure PN,β\mathbb{P}_{N,\beta}.

One major achievement of the theory of β\beta-ensembles is the proof by Valkó–Virág [48] of the existence of a limiting point process for all β>0\beta>0, called the Sine⁡β\operatorname{Sine}_{\beta} process, see § 1.2.

Log-gas interpretation. Dyson [18] was the first to suggest rewriting (1.1) as a Gibbs measure

dPN,β(XN)=1ZN,βexp⁡(−βHN(XN))dXN,HN(XN):=12∑1≤i≠j≤N−log⁡∣xi−xj∣+∑i=1Nxi24N,(2)\mathrm{d}\mathbb{P}_{N,\beta}(X_N)=\frac{1}{Z_{N,\beta}}\exp\left(-\beta H_N(X_N)\right)\mathrm{d}X_N,\qquad H_N(X_N):=\frac{1}{2}\sum_{1\leq i\ne j\leq N}-\log|x_i-x_j|+\sum_{i=1}^{N}\frac{x_i^2}{4N}, \tag*{(2)}

where XN=(x1,…,xN)∈RNX_N=(x_1,\ldots,x_N)\in\mathbb{R}^N. In this “statistical physics interpretation”, PN,β\mathbb{P}_{N,\beta} describes NN particles on the line, with equal charges, repelling each other through the logarithmic interaction potential, and confined by a quadratic external field whose strength depends on NN. The parameter β\beta plays the role of an inverse temperature, HNH_N is the energy and ZN,βZ_{N,\beta} the partition function.

Since the seminal paper [46], this point of view has been used to study the microscopic scale. The goal of the present paper is to give a new proof of the existence of a limiting point process, using only the statistical physics perspective on β\beta-ensembles. This approach could be implemented in other Gibbsian models for which no random matrix theory interpretation is available.

Scaling. With our scaling convention, in which the quadratic potential has strength 1N\frac{1}{N}, the typical distance between neighboring eigenvalues is of order 11, and the total size of the system is proportional to NN. A classical consequence of Wigner’s semicircle law [51] (see e.g. [2]) is that the rescaled empirical measure 1N∑i=1Nδxi/N\frac{1}{N}\sum_{i=1}^{N}\delta_{x_i/N} converges a.s. to the semicircle law, whose density is given by:

ρsc(v):=12π(4−v2)+.(3)\rho_{\mathrm{sc}}(v) := \frac{1}{2\pi}\sqrt{(4-v^2)_+}. \tag*{(3)}

Near Nv∈RNv \in\mathbb{R} with v∈(−2,2)v \in(-2,2), there are approximately ρsc(v)\rho_{\mathrm{sc}}(v) eigenvalues per unit length. We write:

ϱ:=ρsc(0)=1π.\varrho:= \rho_{\mathrm{sc}}(0) = \frac{1}{\pi}.

We distinguish two lengthscales: the microscopic one, of order 11, and the macroscopic one, of order NN.

The Sine-beta point process

Local configurations. The microscopic behavior of the system near a given x∈Rx \in\mathbb{R} is encoded by the local point configuration seen from this point. For xx in the bulk (−2,2)(-2,2), we let

XN,x:=∑i=1Nδρsc(x)(xi−Nx),\mathbf{X}_{N,x} := \sum_{i=1}^{N}\delta_{\rho_{\mathrm{sc}}(x)(x_i-Nx)},

where (x1,…,xN)(x_1,\ldots,x_N) is distributed according to PN,β\mathbb{P}_{N,\beta}. This is the configuration seen from NxNx, rescaled by ρsc(x)\rho_{\mathrm{sc}}(x) to have intensity close to one near 00. For simplicity, we will mostly work with x=0x=0 (our arguments extend to any xx in the bulk), namely with:

XN,0=∑i=1Nδϱxi.(4)\mathbf{X}_{N,0} = \sum_{i=1}^{N}\delta_{\varrho x_i}. \tag*{(4)}

Let Conf be the space of locally finite point configurations on R\mathbb{R}, endowed with the vague topology, for which two configurations are close if they “look the same on a large interval”. A finite configuration can thus be close to an infinite one. The law of XN,x\mathbf{X}_{N,x} is a probability measure on Conf, i.e. a point process.

Existence of a limit. For β∈{1,2,4}\beta\in\{1,2,4\}, proving the existence of a limit is a fairly simple analytic task due to a very specific integrable structure. For general β\beta, there is no such structure. In fact, even the existence of limit points, i.e. of tightness for the law of (XN,x)N(\mathbf{X}_{N,x})_N, is a delicate question: the configuration XN,x\mathbf{X}_{N,x} could a priori have anywhere between 00 and NN points in a fixed interval, and controlling this requires understanding the system down to the microscopic scale. In a major breakthrough [48], Valkó and Virág proved that this limit exists and is the same, up to a scaling, at every point of the bulk.

Theorem. For every x∈(−2,2)x \in(-2,2), the law of XN,x\mathbf{X}_{N,x} converges, as N→∞N \to\infty, to the law of a stationary point process of intensity one, called Sineβ_\beta, which does not depend on xx.

The approach of [48] starts from the tridiagonal matrix model of [14], and analyzes its eigenvalue equation through a phase function, in the spirit of Sturm–Liouville oscillation theory. The work of Killip and Stoiciu [33], based on a “pentadiagonal” matrix model [32] and on a similar analysis of phases, proved the existence of a limit for a very closely related model, the Circular β\beta-ensemble, whose physical interpretation is a log-gas with periodic boundary conditions. Those limits are actually the same [39].

Several other descriptions of Sineβ_\beta have been given since: as the spectrum of a random Dirac operator [49], as the zero set of a random analytic function [50], among others. They have been used to study Sineβ_\beta itself, beyond its mere existence. None of them uses the interpretation of PN,β\mathbb{P}_{N,\beta} as the Gibbs measure of a system of particles with logarithmic interaction.

In recent works [DHLM21, EHL21], the Sineβ_\beta process itself was given a statistical physics description, but the following natural question remained:

Can we prove the existence of a limit for XN,x\mathbf{X}_{N,x}, using only the Gibbsian structure of (1.2)?

As sketched in [37 Section 1.4], the following chain of results would lead to a positive answer:

  1. Prove local laws for the finite-volume model, valid down to the microscopic scale. This ensures the existence of limit points for the law of (XN,x)N(X_{N,x})_N, with good controls on the number of points. This was done by Peilen [41].

  1. Show that the limit points satisfy DLR equations, i.e. are infinite-volume Gibbs measures. This was essentially done in [DHLM21].

  1. Show that the limit points are stationary and retain good properties from the finite-volume model.

  1. Show that every stationary solution (with good properties) of the DLR equations minimizes the free energy functional. This is called a “DLR variational principle”.

  1. Use the fact, proven in [20], that this free energy functional has a unique stationary minimizer.

  1. Conclude: the sequence (XN,x)N(X_{N,x})_N has only one limit point, and thus converges!

The goal of our paper is to complete this program by implementing steps 3 and 4.

Compared to the “random matrix theory” approach, this has arguably two upsides:

  • It can be implemented for other, non-RMT-related systems, as was done recently in [9] for the two-dimensional log-gas at inverse temperature 0<β≤20 < \beta\le2 — existence of a limiting point process was previously only known for β=2\beta= 2.

  • It “directly” yields universality of the limiting point process, retrieving some of the results of [7, 8]. Indeed, our argument does not use the specific form of the Gaussian potential: it relies on the local laws of [41], which hold for quite general confining potentials.

DLR equations

Heuristics

In finite volume, PN,β\mathbb{P}_{N,\beta} is a Gibbs measure: its density is of the form exp⁡(−β×energy)\exp(-\beta\times\text{energy}) with respect to a reference “volume” measure. It is natural to ask whether the limit points of (XN,x)N(X_{N,x})_N admit a similar representation. The right answer goes through the Dobrushin–Lanford–Ruelle (DLR) formalism: instead of looking for a global density, one asks for a Gibbsian description of the local, conditional distributions: for every bounded set Λ\Lambda, the law of the configuration in Λ\Lambda, given the configuration XΛcX_{\Lambda^c} outside Λ\Lambda, should be of the form

dP(XΛ′∣XΛc)∝exp⁡(−β(Energy⁡(XΛ′)+Energy⁡(XΛ′,XΛc)))dPoi⁡Λ(XΛ′).\mathrm{d}\mathbb{P}\left(X'_{\Lambda}\mid X_{\Lambda^c}\right)\propto\exp\left(-\beta\left(\operatorname{Energy}\left(X'_{\Lambda}\right)+\operatorname{Energy}\left(X'_{\Lambda},X_{\Lambda^c}\right)\right)\right)\mathrm{d}\operatorname{Poi}_{\Lambda}\left(X'_{\Lambda}\right).

where Poi⁡Λ\operatorname{Poi}_{\Lambda} is a Poisson point process in Λ\Lambda. The self-energy of XΛ′X'_{\Lambda} is not mysterious, since Λ\Lambda is bounded. The puzzling term is the interaction between the points in Λ\Lambda and the infinitely many points outside Λ\Lambda:

Energy⁡(XΛ′,XΛc):=??∬Λ×Λc−log⁡∣x−y∣ d(XΛ′−Leb⁡)(x) d(XΛc−Leb⁡)(y),(5)\operatorname{Energy}\left(X'_{\Lambda},X_{\Lambda^c}\right)\mathrel{\overset{??}{:=}}\iint_{\Lambda\times\Lambda^c}-\log|x-y|\,\mathrm{d}\left(X'_{\Lambda}-\operatorname{Leb}\right)(x)\,\mathrm{d}\left(X_{\Lambda^c}-\operatorname{Leb}\right)(y), \tag*{(5)}

where the Lebesgue measure Leb⁡\operatorname{Leb} plays the role of the neutralizing background. Because the interaction is long-range, and Λc\Lambda^c is unbounded, there is no reason for (1.5) to converge.

Move functions, DLR equations

As explained in [DHLM21], it is enough to make sense of the difference between the interactions of two configurations X′,XX',X with the exterior, namely of the move function

MΛ(X′,X)≔∬Λ×Λc−log⁡∣x−y∣ d(XΛ′−XΛ)(x) d(XΛc−Leb⁡)(y),(6)M_{\Lambda}(X',X)\coloneqq\iint_{\Lambda\times\Lambda^c}-\log|x-y|\,\mathrm{d}\left(X'_{\Lambda}-X_{\Lambda}\right)(x)\,\mathrm{d}\left(X_{\Lambda^c}-\operatorname{Leb}\right)(y), \tag*{(6)}

which represents the energy cost of moving the points of XX inside Λ\Lambda to the positions given by X′X', while feeling the effect of the (fixed) exterior configuration and of the background. If X′X' and XΛX_{\Lambda} have the same number of points, the signed measure X′−XΛ\mathbf{X}'-\mathbf{X}_{\Lambda} has total mass zero, and this neutrality provides an additional cancellation which makes (1.6) converge. This leads to the canonical DLR equations, which are stated as follows: conditionally on XΛc\mathbf{X}_{\Lambda^c} and on the number nn of points in Λ\Lambda,

References

  1. [1]Michael Aizenman, Sheldon Goldstein, and Joel L. Lebowitz. Bounded fluctuations and translation symmetry breaking in one-dimensional particle systems. Journal of Statistical Physics, 103(3–4):601–618, 2001.
  2. [2]Greg W. Anderson, Alice Guionnet, and Ofer Zeitouni. An Introduction to Random Matrices, volume 118 of Cambridge Studies in Advanced Mathematics. Cambridge University Press, 2010.DOI
  3. [3]Michael Aizenman and Philippe A. Martin. Structure of Gibbs states of one dimensional Coulomb systems. Communications in Mathematical Physics, 78(1):99–116, 1980.DOI
  4. [4]Theodoros Assiotis and Kohei Suzuki. Sineβ is tail trivial, 2026. arXiv:2610.10849.
  5. [5]Theodoros Assiotis. Uniqueness for DLR equations of Sineβ, 2026. arXiv:2609.06832.
  6. [6]Gérard Ben Arous and Alice Guionnet. Large deviations for Wigner’s law and Voiculescu’s non-commutative entropy. Probability Theory and Related Fields, 108(4):517–542, 1997.DOI
  7. [7]Paul Bourgade, László Erdős, and Horng-Tzer Yau. Universality of general β-ensembles. Duke Mathematical Journal, 163(6):1127–1190, 2014.
  8. [8]Florent Bekerman, Alessio Figalli, and Alice Guionnet. Transport maps for β-matrix models and universality. Communications in Mathematical Physics, 338(2):589–619, 2015.
  9. [9]Jeanne Boursier and Thomas Leblé. Existence of a fluid phase for the two-dimensional one-component plasma, 2026. arXiv:2611.17174.arxiv.org/abs/2611.17174
  10. [10]Jean Bricmont, Joel L. Lebowitz, and Charles-Édouard Pfister. On the equivalence of boundary conditions. Journal of Statistical Physics, 21(5):573–582, 1979.DOI
  11. [11]Florent Bekerman, Thomas Leblé, and Sylvia Serfaty. CLT for fluctuations of β-ensembles with general potential. Electronic Journal of Probability, 23:Paper No. 115, 2018.DOI
  12. [12]Paul Bourgade, Krishnan Mody, and Michel Pain. Optimal local law and central limit theorem for β-ensembles. Communications in Mathematical Physics, 390(3):1017–1079, 2022.DOI
  13. [14]Ioana Dumitriu and Alan Edelman. Matrix models for beta ensembles. Journal of Mathematical Physics, 43(11):5830–5847, 2002.
  14. [15]David Dereudre. Introduction to the theory of Gibbs point processes. In David Coupier, editor, Stochastic Geometry, volume 2237 of Lecture Notes in Mathematics, pages 181–229. Springer, Cham, 2019.arxiv.org/abs/1701.08105
  15. [17]David Dereudre and Thibaut Vasseur. Number-rigidity and β-circular Riesz gas. The Annals of Probability, 51(3):1025–1065, 2023.DOI
  16. [18]Freeman J. Dyson. Statistical theory of the energy levels of complex systems. I. Journal of Mathematical Physics, 3(1):140–156, 1962.DOI
  17. [19]Freeman J. Dyson. The threefold way. Algebraic structure of symmetry groups and ensembles in quantum mechanics. Journal of Mathematical Physics, 3(6):1199–1215, 1962.DOI
  18. [20]Matthias Erbar, Martin Huesmann, and Thomas Leblé. The one-dimensional log-gas free energy has a unique minimizer. Communications on Pure and Applied Mathematics, 74(3):615–675, 2021.
  19. [21]László Erdős and Horng-Tzer Yau. A Dynamical Approach to Random Matrix Theory, volume 28 of Courant Lecture Notes in Mathematics. American Mathematical Society, 2017.DOI
  20. [22]Jürg Fröhlich and Charles-Édouard Pfister. On the absence of spontaneous symmetry breaking and of crystalline ordering in two-dimensional systems. Communications in Mathematical Physics, 81(2):277–298, 1981.DOI
  21. [23]Sacha Friedli and Yvan Velenik. Statistical Mechanics of Lattice Systems: a Concrete Mathematical Introduction. Cambridge University Press, 2017.
  22. [24]Hans-Otto Georgii. Canonical and grand canonical Gibbs states for continuum systems. Communications in Mathematical Physics, 48(1):31–51, 1976.DOI
  23. [25]Hans-Otto Georgii. Translation invariance and continuous symmetries in two-dimensional continuum systems. In Mathematical Results in Statistical Mechanics (Marseille, 1998), pages 53–69. World Scientific, River Edge, NJ, 1999.
  24. [26]Hans-Otto Georgii. Gibbs Measures and Phase Transitions, volume 9 of De Gruyter Studies in Mathematics. Walter de Gruyter, Berlin, second edition, 2011.
  25. [28]Hans-Otto Georgii and Hans Zessin. Large deviations and the maximum entropy principle for marked point random fields. Probability Theory and Related Fields, 96(2):177–204, 1993.DOI
  26. [29]Martin Huesmann and Bastian Müller. Gap metrics for stationary point processes and quantitative convexity of the free energy, 2025. arXiv:2509.08659.arxiv.org/abs/2509.08659
  27. [30]Benedikt Jahnel, Jonas Köppl, Yannic Steenbeck, and Alexander Zass. The variational principle for a marked Gibbs point process with infinite-range multibody interactions. Electronic Journal of Probability, 30:1–32, 2025.
  28. [31]Kurt Johansson. On fluctuations of eigenvalues of random Hermitian matrices. Duke Mathematical Journal, 91(1):151–204, 1998.DOI
  29. [32]Rowan Killip and Irina Nenciu. Matrix models for circular ensembles. International Mathematics Research Notices, 2004(50):2665–2701, 2004.
  30. [33]Rowan Killip and Mihai Stoiciu. Eigenvalue statistics for CMV matrices: From Poisson to clock via circular beta ensembles. Duke Mathematical Journal, 146(3):361–399, 2009.DOI
  31. [34]Eugene Kritchevski, Benedek Valkó, and Bálint Virág. The scaling limit of the critical one-dimensional random Schrödinger operator. Communications in Mathematical Physics, 314(3):775–806, 2012.DOI
  32. [35]Thomas Leblé. Logarithmic, Coulomb and Riesz energy of point processes. Journal of Statistical Physics, 162(4):887–923, 2016.
  33. [36]Thomas Leblé. DLR equations, number-rigidity and translation-invariance for infinite-volume limit points of the 2DOCP, 2024. arXiv:2410.04958.arxiv.org/abs/2410.04958
  34. [37]Thomas Leblé. Les log-gas et leurs amis. Habilitation à diriger des recherches, Université Paris Cité, 2026.
  35. [38]Thomas Leblé and Sylvia Serfaty. Large deviation principle for empirical fields of log and Riesz gases. Inventiones mathematicae, 210(3):645–757, 2017.
  36. [39]Fumihiko Nakano. Level statistics for one-dimensional Schrödinger operators and Gaussian beta ensemble. Journal of Statistical Physics, 156(1):66–93, 2014.DOI
  37. [40]Joseph Najnudel and Bálint Virág. Uniform point variance bounds in classical beta ensembles. Random Matrices: Theory and Applications, 10(4):2150033, 2021.DOI
  38. [41]Luke Peilen. Local laws and a mesoscopic CLT for β-ensembles. Communications on Pure and Applied Mathematics, 77(4):2452–2567, 2024.DOI
  39. [42]Christopher Preston. Random Fields, volume 534 of Lecture Notes in Mathematics. Springer, 1976.
  40. [43]Firas Rassoul-Agha and Timo Seppäläinen. A course on large deviations with an introduction to Gibbs measures, volume 162 of Graduate Studies in Mathematics. American Mathematical Society, 2015.
  41. [44]Thomas Richthammer. Translation-invariance of two-dimensional Gibbsian point processes. Communications in Mathematical Physics, 274(1):81–122, 2007.
  42. [45]Sylvia Serfaty. Lectures on Coulomb and Riesz gases, 2024. arXiv:2407.21194.arxiv.org/abs/2407.21194
  43. [46]Etienne Sandier and Sylvia Serfaty. 1D log gases and the renormalized energy: crystallization at vanishing temperature. Probability Theory and Related Fields, 162(3–4):795–846, 2015.arxiv.org/abs/1303.2968
  44. [47]Eric Thoma. Overcrowding and separation estimates for the Coulomb gas. Communications on Pure and Applied Mathematics, 77(7):3227–3276, 2024.arxiv.org/abs/2210.05902
  45. [48]Benedek Valkó and Bálint Virág. Continuum limits of random matrices and the Brownian carousel. Inventiones mathematicae, 177(3):463–508, 2009.DOI
  46. [49]Benedek Valkó and Bálint Virág. The Sineβ operator. Inventiones mathematicae, 209(1):275–327, 2017.
  47. [50]Benedek Valkó and Bálint Virág. The many faces of the stochastic zeta function. Geometric and Functional Analysis, 32:1160–1231, 2022.DOI
  48. [51]Eugene P. Wigner. On the distribution of the roots of certain symmetric matrices. Annals of Mathematics, 67(2):325–327, 1958.DOI

Paper details

vlnc:W_x7hrmrf9
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

OpenAI and Anthropic models

View statement
Looking for related papers…
Contents