Replicated liquid theory in $1+\infty$ dimensions

Basic Information

• Paper ID: 2508.21639
• Title: Replicated liquid theory in $1+\infty$ dimensions
• Authors: Yukihiro Tomita (Osaka University), Hajime Yoshino (Osaka University)
• Categories: cond-mat.soft (soft matter), cond-mat.dis-nn (disordered systems and neural networks), cond-mat.stat-mech (statistical mechanics)
• Publication Date: August 2025 (arXiv v2: November 13, 2025)
• Paper Link: https://arxiv.org/abs/2508.21639

Abstract

This paper develops a replicated liquid theory for structural glasses that describes spatial variations of physical quantities along one axis (e.g., the z-axis). The theory becomes exact when the transverse dimension approaches infinity ($d-1 \to \infty$). It provides an exact free energy functional containing a spatially-dependent glass order parameter $\Delta_{ab}(z)$. As the first application of this framework, the paper investigates diverging length scales associated with kinetic and static glass transitions in hard-sphere systems with and without confinement. The obtained critical exponents are consistent with previous mean-field model studies. Furthermore, the theory predicts non-trivial spatial distributions of the glass order parameter $\Delta_{ab}(z)$ within cavities, exhibiting scaling characteristics near the kinetic glass transition.

Research Background and Motivation

Problems to be Addressed

The theoretical description of glassy matter has been a central challenge in statistical physics. While exact mean-field theories in the large-dimension limit ($d \to \infty$) have been established, such theories suffer from a fundamental deficiency: they cannot describe any spatial variations or fluctuations. However, glass-forming liquids and glassy states exhibit many important spatial heterogeneity phenomena:

1. Dynamical heterogeneity: Observed in supercooled liquid states
2. Isostatic length: A characteristic length scale that diverges near jamming
3. Shear bands: Formation near yielding or fracture

Importance of the Problem

Understanding spatial heterogeneity and diverging length scales in glass transitions is crucial for:

• Elucidating the essential mechanisms of glass transitions
• Understanding kinetic glass transitions and the thermodynamic Kauzmann transition
• Predicting material behavior under external perturbations such as compression and shear
• Connecting microscopic particle-scale phenomena to macroscopic observations

Limitations of Existing Approaches

1. Standard $d \to \infty$ theory: Completely uniform, unable to describe spatial variations
2. Ginzburg-Landau type field theories: Valid only at sufficiently long wavelengths, lacking microscopic precision
3. Inhomogeneous MCT (mode coupling theory): While starting from first principles, limited to dynamical descriptions

Motivation for This Work

Develop a $1+\infty$ dimensional theory that can:

• Describe spatial variations along one direction (z-axis)
• Remain exact when the transverse dimension $d-1 \to \infty$
• Maintain exactness at the particle scale, starting from microscopic principles
• Provide theoretical tools for phenomena like jamming and yielding that require precise microscopic descriptions

Core Contributions

1. Established a $1+(d-1)$ dimensional inhomogeneous replicated liquid theory framework: For the first time, extended the exact mean-field replicated liquid theory to describe one-dimensional spatial variations, with the theory being exact in the $d-1 \to \infty$ limit
2. Derived spatially-dependent free energy functional: Obtained exact expressions for the free energy functional containing spatially-dependent glass order parameter $\Delta_{ab}(z)$ and self-consistent equations
3. Systematically studied diverging length scales related to glass transitions:
   • Obtained correlation length $\xi_d^{\text{hessian}} \propto \delta\phi^{-1/4}$ through Hessian matrix analysis
   • Verified the same scaling behavior through point-to-set length analysis in cavity systems
   • Found that the spatial distribution of order parameters reflects the same correlation length
4. Predicted non-trivial spatial distributions of glass order parameter in cavities: Revealed scaling characteristics similar to surface critical phenomena
5. Confirmed universality of critical exponents: Kinetic transition exponent $\nu=1/4$, Kauzmann transition exponent $\nu=1$, consistent with previous mean-field spin glass models, inhomogeneous MCT, and Kac glass models

Detailed Methodology

Task Definition

The research objective is to construct a theoretical framework describing spatially inhomogeneous glassy states, specifically:

• Input: System parameters (density/temperature), boundary conditions (cavity size)
• Output: Spatially-dependent density distribution $\rho(z)$ and glass order parameter $\Delta_{ab}(z)$
• Constraints: Thermodynamic consistency, normalization conditions

Theoretical Architecture

1. Basic Model Setup

Consider a system of $N$ particles in a cylindrical container with cross-sectional area $S$, with coordinates $\mathbf{x}_i = (x_i^1, ..., x_i^{d-1}, z_i)$. The Hamiltonian is:

$H = \sum_{i=1}^N \frac{|\mathbf{p}_i|^2}{2m} + \sum_{i<j} v(r_{ij}) + \sum_{i=1}^N U(z_i)$

where $v(r_{ij})$ is the two-body interaction potential and $U(z_i)$ is an external confining potential.

2. Key Scaling in the Large-Dimension Limit

In the $d-1 \to \infty$ limit, the key insight is to introduce rescaled coordinates:

$z = \frac{D}{\sqrt{d}}\hat{z}$

where $D$ is the microscopic length scale (e.g., hard-sphere diameter). This reflects that the characteristic scale of inter-particle distance fluctuations is $O(1/\sqrt{d})$.

3. Inhomogeneous Liquid Theory

First, develop density functional theory. The free energy functional (to first order in Mayer expansion) is:

$-\beta F[\rho] = \int d\hat{z}\rho(\hat{z})[1-\ln(\lambda_{th}^d\rho(\hat{z}))] + \int d\hat{z}\rho(\hat{z})(-\beta U)(\hat{z})$$+ \frac{1}{2}\int d\hat{z}_1 d\hat{z}_2 \rho(\hat{z}_1)\rho(\hat{z}_2) \frac{e^{-(\hat{z}_1-\hat{z}_2)^2/2}}{\sqrt{2\pi}}(-\mathcal{F})$

where $\mathcal{F}$ contains contributions from interactions.

4. Replicated Liquid Theory

To describe the glassy state, introduce $m$ replicated systems. The key is to decompose coordinates:

$\mathbf{x}_i^a = \mathbf{x}_i^c + \mathbf{u}_i^a$

where $\mathbf{x}_i^c$ is the "molecular" center of mass and $\mathbf{u}_i^a$ is the intra-replica fluctuation. Define the spatially-dependent glass order parameter:

$\Delta_{ab}(z) = \alpha_{aa}(z) + \alpha_{bb}(z) - 2\alpha_{ab}(z)$

where:

$\alpha_{ab}(z)\rho(z) = \frac{d}{D^2}\frac{1}{S}\sum_{i=1}^N \langle \mathbf{u}_i^a \cdot \mathbf{u}_i^b \delta(z-(z_i)_c)\rangle$

5. Free Energy Functional

In the $d-1 \to \infty$ limit, the free energy functional for the replicated system is exactly:

$-\beta F_m[\rho, \alpha_{ab}] = \int d\hat{z}\rho(\hat{z})\left\{1-\ln(\rho(\hat{z})\lambda_{th}^d) + d\ln m + \frac{(m-1)d}{2}\ln\frac{2\pi e D^2}{d^2\lambda_{th}^2}\right\}$$+ \int d\hat{z}\rho(\hat{z})\left[(-\beta U)(\hat{z}) + \frac{d}{2}\ln\det(\hat{\alpha}^{m,m}(\hat{z}))\right]$$+ \frac{d}{2}\frac{\Omega_d D^d}{d}\int d\hat{z}_1 d\hat{z}_2 \rho(\hat{z}_1)\rho(\hat{z}_2)\frac{e^{-(\hat{z}_1-\hat{z}_2)^2/2}}{\sqrt{2\pi}}(-F_{\text{int}}(\Delta_{ab}(\hat{z}_1,\hat{z}_2)))$

where the replicated Mayer function is defined as:

$-F_{\text{int}}(\Delta_{ab}) = \int_{-\infty}^\infty d\xi e^\xi e^{-\frac{1}{2}\sum_{ab}\Delta_{ab}\partial_{\xi_a}\partial_{\xi_b}}\left[\prod_a e^{-\beta v(D^2(1+\xi_a/d)^2)} - 1\right]_{\{\xi_a=\xi\}}$

6. One-Step RSB Solution

Adopt the simplest symmetry-breaking ansatz:

$\Delta_{ab}(\hat{z}) = \Delta(\hat{z})(1-\delta_{ab})$

The self-consistent equation is:

$\frac{1}{\Delta(\hat{z})} = \frac{m}{2}\int \frac{d\hat{z}'}{\sqrt{2\pi}}\hat{\phi}(\hat{z}')e^{-(\hat{z}-\hat{z}')^2/2}\int_{-\infty}^\infty d\xi e^{\xi-\frac{1}{2}\Delta(\hat{z},\hat{z}')}g^m(\xi,\Delta(\hat{z},\hat{z}'))(f'(\xi,\Delta(\hat{z},\hat{z}')))^2$

where $\hat{\phi}$ is the rescaled volume fraction, and $g(\xi,\Delta)$ and $f(\xi,\Delta)$ are determined by the interaction potential.

Technical Innovations

1. Dimension decomposition strategy: Cleverly decompose space into one "active" dimension (z-axis) and $d-1$ "frozen" dimensions, allowing longitudinal variations while maintaining infinite transverse dimension
2. Exact large-dimension limit treatment: Through appropriate coordinate rescaling and density rescaling, ensure the theory is exact in the $d-1 \to \infty$ limit, avoiding approximations
3. Cavity system design: By setting $\Delta=0$ (frozen region) outside the cavity, cleverly implement measurement of point-to-set length, a standard method for studying correlation lengths in glass transitions
4. Multi-scale analysis: Simultaneously study correlation length from three perspectives:
   • Fluctuation analysis of Hessian matrix (longitudinal modes)
   • Cavity size dependence of transition point
   • Characteristic length of spatial order parameter distribution

Experimental Setup

Research System

Hard-sphere system: The simplest glass-forming system with interaction potential:

$v(r) = \begin{cases} \infty & r \leq D \\ 0 & r > D \end{cases}$

The corresponding Boltzmann factor is $e^{-\beta v(r)} = \theta(r-D)$.

Two Configurations

Configuration 1: Infinite System ($-\infty < z < \infty$)

• Uniform density $\hat{\phi}(\hat{z}) = \hat{\phi}$
• Uniform glass order parameter $\Delta(\hat{z}) = \Delta$
• Analyze spatial correlations of Hessian matrix

Configuration 2: Cavity System

• Cavity region: $0 < \hat{z} < \hat{L}_{\text{cav}}$
• Outside cavity: $\Delta(\hat{z}) = 0$ (frozen)
• Inside cavity: Solve self-consistent equations for $\Delta(\hat{z})$
• Uniform density $\hat{\phi}$

Numerical Methods

1. Discretize cavity region with grid spacing $d\hat{z} = 0.1$
2. Use simple Euler method for integration with step size $d\hat{z}' = d\xi = 0.1$
3. Iterate self-consistent equations until convergence
4. Initialize with bulk system solution $\Delta_{\text{bulk}}$

Evaluation Metrics

1. Kinetic transition density $\hat{\phi}_d$: Density where saddle-point solution disappears (spinodal in spin-glass sense)
2. Correlation lengths:
   • $\xi_d^{\text{hessian}}$: Derived from Hessian matrix
   • $\xi_d^{\text{PS}}$: Point-to-set length, defined as $\xi_d^{\text{PS}}(\hat{\phi}_d(\hat{L}_{\text{cav}})) = \hat{L}_{\text{cav}}/2$
   • $\xi_d^{\text{profile}}$: Fitted from spatial order parameter distribution
3. Kauzmann transition density $\hat{\phi}_K$: Density where complexity $\Sigma=0$
4. Critical exponent $\nu$: $\xi \propto \delta\phi^{-\nu}$, where $\delta\phi = (\hat{\phi}-\hat{\phi}_c)/\hat{\phi}_c$

Experimental Results

Main Results

1. Kinetic Transition of Bulk System

For hard-sphere system ($m=1$), numerical solution of self-consistent equations yields:

$\hat{\phi}_d = 4.8067787037$

This is the critical density where the solution of self-consistent equations disappears.

2. Correlation Length from Hessian Analysis

The Fourier transform of the longitudinal Hessian matrix at small wave vector:

$\tilde{M}(k) = \frac{d}{2}(m-1)\hat{\phi}\left[M_0 + \frac{k^2}{2}M_2 + O(k^4)\right]$

where:

• $M_0 = \frac{1}{\Delta^2} - \frac{m}{2}\hat{\phi}X(\Delta)$
• $M_2 = \frac{m}{4}\hat{\phi}X(\Delta)$

Correlation length:

$\xi_d^{\text{hessian}} = \sqrt{\frac{M_2}{2M_0}}$

Numerical results show:

• $M_0 = a\sqrt{\delta\phi}$, where $a \approx 0.635$
• $M_2 = \frac{1}{2\Delta_d^2} \approx 0.376$

Therefore:

$\xi_d^{\text{hessian}} \propto \delta\phi^{-1/4}$

Critical exponent $\nu = 1/4$.

3. Point-to-Set Length in Cavity System

By varying cavity size $\hat{L}_{\text{cav}}$, study the change in kinetic transition density $\hat{\phi}_d(\hat{L}_{\text{cav}})$:

• In smaller cavities, transition occurs at lower density
• Define $\xi_d^{\text{PS}}(\hat{\phi}_d(\hat{L}_{\text{cav}})) = \hat{L}_{\text{cav}}/2$

Numerical results perfectly verify:

$\xi_d^{\text{PS}} \propto \delta\phi^{-1/4}$

Consistent with Hessian analysis results.

4. Spatial Distribution of Order Parameter

Define deviation:

$\delta\Delta(\hat{z}) = \Delta_{\text{bulk}} - \Delta(\hat{z})$

In sufficiently large cavities, exhibits exponential decay:

$\delta\Delta(\hat{z}) = A \exp\left(-\frac{\hat{z}}{\xi_d^{\text{profile}}}\right)$

The fitted $\xi_d^{\text{profile}}$ perfectly matches $\xi_d^{\text{hessian}}$ (see Figure 3(c)), both following $\delta\phi^{-1/4}$ scaling.

5. Scaling Functions

Near the critical point, the order parameter distribution exhibits universal scaling form. Define:

$f(\hat{z}, \delta\phi) = \frac{\Delta_d - \Delta(\hat{z}, \delta\phi)}{\Delta_d - \Delta_{\text{bulk}}(\delta\phi)} - 1$

Data shows $f$ is a universal function of $\hat{z}/\xi_d^{\text{profile}}$ (Figure 3(b)), satisfying $\lim_{\hat{z}\to\infty}f(\hat{z},\delta\phi)=0$.

6. Kauzmann Transition

Complexity of cavity system:

$\Sigma^* = \frac{d}{2}\left(\ln d - \hat{\phi}\left[2-f(\hat{L}_{\text{cav}})\right]\right)$

where:

$f(\hat{L}_{\text{cav}}) = 1 - \sqrt{\frac{2}{\pi}}\hat{L}_{\text{cav}}^{-1} + O(e^{-\hat{L}_{\text{cav}}^2/2})$

Kauzmann transition density:

$\hat{\phi}_K(\hat{L}_{\text{cav}}) = \frac{\hat{\phi}_{K,\text{bulk}}}{2-f(\hat{L}_{\text{cav}})} = \hat{\phi}_{K,\text{bulk}}\left(1 + \sqrt{\frac{2}{\pi}}\hat{L}_{\text{cav}}^{-1} + O(e^{-\hat{L}_{\text{cav}}^2/2})\right)$

Point-to-set length:

$\xi_K^{\text{PS}} \propto (\hat{\phi} - \hat{\phi}_{K,\text{bulk}})^{-1}$

Critical exponent $\nu = 1$.

Key Findings

1. Consistency of three length scales: $\xi_d^{\text{hessian}} = \xi_d^{\text{profile}} \propto \xi_d^{\text{PS}}$, all following the same scaling law
2. Universality of critical exponents:
   • Kinetic transition: $\nu = 1/4$
   • Kauzmann transition: $\nu = 1$
   
   Perfectly consistent with previous theoretical predictions (TR Kirkpatrick & PG Wolynes 1987, inhomogeneous MCT, Kac glass models, etc.)
3. Analogy with surface critical phenomena: The spatial distribution of order parameter $\Delta(\hat{z})$ reflects correlation length in a manner analogous to surface critical phenomena in ferromagnets
4. Cavity confinement effects: The frozen region outside the cavity constrains the cavity interior, causing glass transition to occur at lower density in smaller cavities

Related Work

Development of Glass Theory

1. Mean-field spin glass theory:
   • TR Kirkpatrick & PG Wolynes (1987): First predicted diverging length scales and $\nu=1/4$ exponent
   • Used mean-field spin glass models with phenomenology similar to structural glass phenomena
2. Large-dimension liquid theory:
   • Kurchan, Parisi, Zamponi et al. (2012-2014): Established exact mean-field theory for $d\to\infty$
   • Unified framework for kinetic transition, Kauzmann transition, jamming, and yielding
   • Limitation: Completely uniform, unable to describe spatial variations
3. Inhomogeneous MCT:
   • Biroli, Bouchaud, Miyazaki, Reichman (2006): Study dynamical length scales from first principles
   • Dynamical theory, limited to kinetic transition
4. Other related theories:
   • Kac glass model (Franz & Montanari 2007)
   • One-dimensional hard-sphere chain model (Ikeda & Ikeda 2015)
   • Replicated theory for deep neural networks (Yoshino 2020): Revealed non-trivial spatial order parameter distributions

Positioning of This Work

This paper can be viewed as:

• Thermodynamic (static) counterpart of inhomogeneous MCT
• Continuous limit of hard-sphere chain model
• Extension of large-dimension theory to spatially inhomogeneous cases

Advantages of this work:

• Starting from microscopic first principles
• Simultaneously treating kinetic and static transitions
• Exact at particle scale
• Generalizable to 2D and 3D inhomogeneities

Conclusions and Discussion

Main Conclusions

1. Successfully constructed $1+\infty$ dimensional inhomogeneous replicated liquid theory: Exact when transverse dimension is infinite, capable of describing longitudinal spatial variations
2. Verified critical behavior of correlation length in glass transitions:
   • Kinetic transition: $\xi_d \propto \delta\phi^{-1/4}$
   • Kauzmann transition: $\xi_K \propto (\phi-\phi_K)^{-1}$
3. Revealed spatial distribution characteristics of glass order parameter:
   • Exhibits exponential decay in cavities
   • Characteristic length equals thermal fluctuation correlation length
   • Shows universal scaling near critical point
4. Provided new tools for studying spatial heterogeneity in glassy states: Generalizable to phenomena like jamming and yielding

Limitations

1. Still a strong mean-field theory:
   • Transverse dimension must be infinite
   • Misses many subtle features of finite-dimensional liquids
   • Cannot address existence of kinetic/static glass transitions in real systems
2. Only considers longitudinal modes:
   • Complete analysis requires including replicon and anomalous modes
   • These modes contribute to dynamical heterogeneity quantities like $\chi_4$
   • Present work limited to longitudinal fluctuations
3. Limited to one-step RSB:
   • Above Gardner transition at high density requires full-RSB scheme
   • Results limited to below Gardner transition
4. Dynamical relevance of static length scales:
   • Static length scales remain small when dynamics are slow
   • Actual glass dynamics may involve other mechanisms like avalanches
   • However, point-to-set length in deeply annealed glasses does describe dynamical characteristic length (Shiraishi & Berthier 2024)

Future Directions

1. Generalization to 2D and 3D inhomogeneities: Direct extension while maintaining infinite transverse dimension
2. Dynamical version: Extend dynamical mean-field theory (Maimbourg, Kurchan, Zamponi 2016)
3. Applications to yielding and jamming:
   • Study spatial heterogeneities like shear bands and isostatic length
   • Requires precise microscopic description, particularly suited for this theory
4. Glass state following: Adapt Franz-Parisi scheme to study glass state evolution under compression and shear
5. Above Gardner transition: Develop full-RSB version

In-Depth Evaluation

Strengths

1. Theoretical Innovation

• High originality: First successful extension of exact mean-field theory to spatially inhomogeneous cases
• Clever methodology: Dimension decomposition strategy maintains exactness while allowing spatial variations
• Complete theory: Rigorously derived from microscopic Hamiltonian without phenomenological parameters

2. Technical Rigor

• Detailed mathematical derivations (38 pages main text + extensive appendix)
• Exact treatment of large-dimension limit, clear scaling relationships
• Explicit numerical methods, reproducible

3. Convincing Results

• Three independent methods (Hessian, point-to-set, spatial distribution) yield consistent results
• Critical exponents perfectly match multiple previous theories
• Scaling behavior verified over wide parameter range

4. Physical Insights

• Reveals deep connection between spatial order parameter distribution and fluctuation correlations
• Establishes analogy with surface critical phenomena
• Provides new perspective for understanding spatial heterogeneity in glassy states

5. Writing Quality

• Clear structure, rigorous logic
• Detailed appendix facilitating understanding and reproduction
• Beautiful figures with rich information

Weaknesses

1. Theoretical Limitations

• Dimension restriction: Transverse dimension still requires infinity, far from realistic 3D systems
• Mean-field nature: Cannot capture finite-dimensional fluctuation effects
• RSB limitation: Only handles one-step RSB, inapplicable to high-density region above Gardner transition

2. Application Scope

• Only hard-sphere system studied; behavior for other interaction potentials unexplored
• Only equilibrium states; non-equilibrium processes (e.g., shear) not addressed
• Dynamical theory version not yet developed

3. Numerical Analysis

• Numerical precision depends on grid resolution (0.1); convergence analysis insufficient
• Extrapolation to large cavity limit requires more systematic finite-size scaling analysis
• Error estimates not provided

4. Experimental Verification

• Lacks direct comparison with simulations or experiments
• Static length scale measurements experimentally challenging
• Experimental realization of $1+\infty$ dimensional systems unclear

Impact

Contributions to the Field

• Theoretical methodology: Provides new paradigm for studying spatially inhomogeneous glassy states
• Conceptual deepening: Enhances understanding of correlation length in glass transitions
• Bridge role: Connects mean-field theory, MCT, and surface critical phenomena

Practical Value

• Short-term: Provides theoretical benchmark for understanding spatial heterogeneity in glassy states
• Mid-term: Potential applications to theoretical studies of jamming and yielding
• Long-term: Provides starting point for developing finite-dimensional corrections

Reproducibility

• High: Derivations detailed, numerical methods explicit
• Requires substantial theoretical physics background
• Code not publicly available; complete reproduction requires significant effort

Applicable Scenarios

Ideal Application Scenarios

1. Theoretical research:
   • Mean-field theory of glass transitions
   • Correlation length and critical phenomena
   • Theoretical modeling of spatial heterogeneity
2. Specific systems:
   • Quasi-one-dimensional geometries (nanotubes, slits)
   • Glassy states near interfaces
   • Systems under gradient compression/shear
3. Method development:
   • Benchmark for more complex theories
   • Testing numerical methods
   • Inspiring phenomenological models

Inapplicable Scenarios

1. Quantitative predictions for low-dimensional systems (2D, 3D)
2. Phenomena requiring consideration of fluctuation effects
3. High-density region above Gardner transition
4. Strong non-equilibrium processes

Potential Extensions

1. Methodological extensions:
   • Include replicon/anomalous modes
   • Develop full-RSB version
   • Couple with dynamical theory
2. Application extensions:
   • Soft spheres, Lennard-Jones and other potentials
   • Multi-component systems
   • Glassification of active matter
3. Dimensional extensions:
   • $2+\infty$ and $3+\infty$ dimensional theories
   • Systematic expansion in finite-dimensional corrections

Selected References

7 Giorgio Parisi, Pierfrancesco Urbani, Francesco Zamponi. Theory of simple glasses: exact solutions in infinite dimensions. Cambridge University Press, 2020.

11 TR Kirkpatrick, PG Wolynes. Stable and metastable states in mean-field Potts and structural glasses. Phys. Rev. B, 36:8552, 1987.

22 Giulio Biroli et al. Inhomogeneous mode-coupling theory and growing dynamic length in supercooled liquids. Phys. Rev. Lett., 97:195701, 2006.

26 Harukuni Ikeda, Atsushi Ikeda. One-dimensional Kac model of dense amorphous hard spheres. EPL, 111:40007, 2015.

Overall Assessment: This is a high-quality theoretical physics work that successfully extends exact mean-field theory to spatially inhomogeneous cases, providing important theoretical tools for understanding spatial heterogeneity in glassy states. Despite inherent limitations of mean-field theory, its theoretical rigor, result consistency, and methodological innovation make it a significant contribution to the field. For theoretical physicists, this is essential reading for understanding spatial heterogeneity in glass transitions; for the broader soft matter community, it provides profound insights into correlation length and critical behavior.