Finite-dimensional fate of the Franz-Parisi coupled-replica phase diagram of glass formers: coexistence line, critical endpoint, and the connection to a thermodynamic glass transition at zero coupling

Franz and Parisi (Phys. Rev. Lett. 79, 2486 (1997)) computed, in the mean-field spherical $p$ -spin model, the phase diagram of a glassy system coupled by a field $\epsilon$ conjugate to its overlap with a quenched reference configuration: a first-order coexistence line between a high-overlap ('magnetized') and a low-overlap ('disordered') phase, ending at a finite critical point $(T_{\rm cr}, \epsilon_{\rm cr})$ , with spinodals meeting the $\epsilon = 0$ axis at the dynamic temperature $T_d$ and the static (Kauzmann) temperature $T_s$ , nonzero latent heat for $\epsilon > 0$ , and the complexity $\Sigma(T)$ identifiable with the free-energy difference between the two phases. They proposed that this topology should survive in finite-dimensional glass formers (Lennard-Jones, hard spheres, polymers) and be verified numerically in simulations of such models, with the complexity measured through thermodynamic loops around the critical endpoint. Later work has confirmed large parts: in three dimensions the first-order overlap transition line ending at a critical point (in the random-field Ising universality class) persists in the thermodynamic limit, and the coupled-replica complexity measurement has been realized via umbrella sampling and, in two dimensions, by equilibrium measurements down to $T = 0$ where the Kauzmann transition occurs at zero temperature. The surviving open core is the $\epsilon \to 0$ and deep-supercooled part of the diagram in spatial dimensions $d \ge 3$ : whether the coexistence line connects to a genuine thermodynamic (ideal glass / Kauzmann) transition at finite $T_s$ , whether the complexity $\Sigma(T)$ can be measured in equilibrium down to its vanishing point in three-dimensional glass formers, and how activated finite-dimensional processes round the mean-field spinodal structure at $T_d$ beyond the established numerical characterization of that rounding, for which a quantitative or controlled theoretical understanding remains open.

Background

When a liquid is supercooled below its melting point it may fall out of equilibrium at a laboratory glass temperature $T_g$ , but whether an underlying equilibrium thermodynamic transition exists is the central unresolved question of glass physics. Random first-order transition (RFOT) theory, developed for mean-field models such as the spherical $p$ -spin glass ( $N$ continuous spin variables coupled by a random Gaussian Hamiltonian whose correlation function is a power $p$ of the spin overlap), predicts two characteristic temperatures: a dynamic temperature $T_d$ where an ergodicity-breaking transition would occur (the mean-field analogue of the mode-coupling theory crossover, mode-coupling theory being the microscopic theory that predicts a dynamical arrest from the growth of viscous relaxation times), and a lower static temperature $T_s$ where an ideal glass transition takes place through an entropy crisis, i.e. the vanishing of the configurational entropy or complexity $\Sigma(T)$ , the logarithm of the number of metastable glass states.

Glassy systems have no conventional order parameter, but spin glass theory suggests the overlap $q$ between two independent 'real' replicas (independent copies) of the same system as the natural order parameter. Franz and Parisi introduced the corresponding thermodynamic construction: one thermalizes a reference configuration at temperature $T'$ , and lets a second replica evolve at temperature $T$ under a field $\epsilon$ conjugate to the mutual overlap (the 'Franz-Parisi potential' $V(q)$ is the associated free energy as a function of $q$ ). In the spherical $p$ -spin model they computed the equilibrium phase diagram in the $(\epsilon, T)$ plane: below $T_s$ , for $\epsilon > 0$ , a 'magnetized' (high-overlap, glassy) phase coexists with a 'disordered' (low-overlap, liquid) phase along a first-order line ending at a finite critical point $(T_{\rm cr}, \epsilon_{\rm cr})$ ; the spinodal (the limit of metastability beyond which the magnetized phase ceases to exist) touches the $\epsilon = 0$ axis at $T_d$ while the coexistence line touches it at $T_s$ ; a latent heat is present for $\epsilon > 0$ and vanishes at the critical point and at $T = 0$ ; and the complexity $\Sigma$ appears as the free-energy difference between the coexisting phases, so that closing a thermodynamic loop around the critical endpoint (through regions of stable or metastable equilibrium) and integrating the specific heat should yield $\Sigma(T)$ in a simulation. The authors explicitly proposed numerical tests of these predictions in realistic glass formers such as Lennard-Jones mixtures, hard spheres, and polymer glasses; the small-sample Monte Carlo simulations of their 1998 follow-up and later techniques such as parallel tempering (a Monte Carlo method that swaps temperatures between simultaneous replicas to overcome barriers), umbrella sampling, and swap Monte Carlo are strategies developed in subsequent implementations.

This coupled-replica construction became the backbone of the finite-dimensional RFOT program. Franz, Parisi, Ricci-Tersenghi, and Rizzo developed a replica field theory predicting that the critical endpoint of the overlap transition belongs to the universality class of the random-field Ising model (RFIM), because the quenched reference configuration acts as a self-induced random field. Modern algorithms (swap Monte Carlo, which exchanges particle diameters to accelerate relaxation; umbrella sampling, which biases the overlap to sample its rare fluctuations; parallel tempering; population annealing, which evolves and resamples a population of configurations according to their statistical weights) now equilibrate model glass formers far below conventional numerical limits, making the proposed measurements concrete.

Problem Statement

Franz and Parisi (1997), in the mean-field spherical $p$ -spin model, predicted for a glassy system coupled by a field $\epsilon$ to a quenched reference configuration a phase diagram in the $(\epsilon, T)$ plane consisting of a first-order coexistence line between high-overlap (glassy) and low-overlap (liquid) phases that departs the $\epsilon = 0$ axis at the static (Kauzmann) temperature $T_s$ (where the configurational entropy would vanish), has its magnetized-state spinodal (limit of metastability of the high-overlap phase) touching $\epsilon = 0$ at the dynamical temperature $T_d$ (the mode-coupling-like ergodicity-breaking temperature), terminates at a finite critical point $(T_{\rm cr}, \epsilon_{\rm cr})$ , carries a nonzero latent heat for $\epsilon > 0$ that vanishes at the critical point and at $T = 0$ , exhibits Maxwell-constructed (equal-area) discontinuities in the overlap $q(\epsilon)$ , and identifies the complexity $\Sigma(T)$ (the configurational entropy, i.e. the logarithm of the number of metastable glass states) with the free-energy difference between the coexisting phases, measurable in a simulation through closed thermodynamic loops around the critical endpoint with specific-heat integration. The $\epsilon > 0$ portion of this topology (first-order line ending at a finite critical endpoint, random-field Ising universality class) is now established in $d = 3$ and absent in $d = 2$ , so the open core is the $\epsilon \to 0$ and deep-supercooled part of the diagram in spatial dimensions $d \ge 3$ . For finite-dimensional, short-range glass-forming systems (e.g., Lennard-Jones mixtures, hard spheres, polymer glasses), determine the following independently-standing questions, each of which is mandatory; a partial answer to one of them does not resolve the others:

(a) whether, in $d \ge 3$ , the coexistence line connects at $\epsilon = 0$ to a genuine thermodynamic (ideal glass / Kauzmann) transition at finite $T_s$ with vanishing complexity, or terminates otherwise (e.g., at a zero-temperature critical point or not at all);
(b) whether the complexity $\Sigma(T)$ of a three-dimensional glass former can be measured in equilibrium down to its vanishing point via coupled-replica thermodynamics; and
(c) given the established numerical characterization of the activated modification of the mean-field spinodal at $T_d$ (strong softening of the square-root singularity in $d = 3$ , progressively restored toward mean-field scaling for $d \ge 8$ ), how to obtain a quantitative or controlled theoretical understanding of how activated (barrier-hopping) processes round that spinodal in finite dimensions, including the possible role of nonperturbative fluctuations.

Question (a) is resolved by a definite answer to whether the $\epsilon = 0$ transition exists at a finite $T_s$ , with the endpoint located or excluded; question (b), independently of any theoretical account, by an equilibrium measurement reaching the vanishing point of $\Sigma(T)$ or by a demonstrated obstruction to reaching it; question (c) by a controlled account of the activated rounding, as distinct from a reproduction of the established numerics. The common objective is the completion of the coupled-replica phase diagram of finite-dimensional glass formers and its interpretation within (or against) random first-order transition theory.

References

S. Franz and G. Parisi, Phase Diagram of Coupled Glassy Systems: A Mean-Field Study, Phys. Rev. Lett. 79, 2486 (1997), DOI 10.1103/PhysRevLett.79.2486, https://doi.org/10.1103/PhysRevLett.79.2486
S. Franz and G. Parisi, Effective potential in glassy systems: theory and simulations, Physica A 261, 317 (1998), DOI 10.1016/S0378-4371(98)00315-X, https://doi.org/10.1016/S0378-4371(98)00315-X
M. Cardenas, S. Franz, and G. Parisi, Constrained Boltzmann-Gibbs measures and effective potential for glasses in hypernetted chain approximation and numerical simulations, J. Chem. Phys. 110, 1726 (1999), DOI 10.1063/1.478028, https://doi.org/10.1063/1.478028
M. A. Moore and J. Yeo, Thermodynamic glass transition in finite dimensions, Phys. Rev. Lett. 96, 095701 (2006), DOI 10.1103/PhysRevLett.96.095701, https://doi.org/10.1103/PhysRevLett.96.095701
S. Franz, G. Parisi, F. Ricci-Tersenghi, and T. Rizzo, Field theory of fluctuations in glasses, Eur. Phys. J. E 34, 102 (2011), DOI 10.1140/epje/i2011-11102-0, https://doi.org/10.1140/epje/i2011-11102-0
L. Berthier, Overlap fluctuations in glass-forming liquids, Phys. Rev. E 88, 022313 (2013), DOI 10.1103/PhysRevE.88.022313, https://doi.org/10.1103/PhysRevE.88.022313
G. Parisi and B. Seoane, Liquid-glass transition in equilibrium, Phys. Rev. E 89, 022309 (2014), DOI 10.1103/PhysRevE.89.022309, https://doi.org/10.1103/PhysRevE.89.022309
J. P. Garrahan, Transition in coupled replicas may not imply a finite-temperature ideal glass transition in glass-forming systems, Phys. Rev. E 89, 030301(R) (2014), DOI 10.1103/PhysRevE.89.030301, https://doi.org/10.1103/PhysRevE.89.030301
L. Berthier and D. Coslovich, Novel approach to numerical measurements of the configurational entropy in supercooled liquids, Proc. Natl. Acad. Sci. USA 111, 11668 (2014), DOI 10.1073/pnas.1407934111, https://doi.org/10.1073/pnas.1407934111 (arXiv:1401.5260, https://arxiv.org/abs/1401.5260)
L. Berthier, P. Charbonneau, D. Coslovich, A. Ninarello, M. Ozawa, and S. Yaida, Configurational entropy measurements in extremely supercooled liquids that break the glass ceiling, Proc. Natl. Acad. Sci. USA 114, 11356 (2017), DOI 10.1073/pnas.1706860114, https://doi.org/10.1073/pnas.1706860114
L. Berthier, P. Charbonneau, A. Ninarello, M. Ozawa, and S. Yaida, Zero-temperature glass transition in two dimensions, Nat. Commun. 10, 1508 (2019), DOI 10.1038/s41467-019-09512-3, https://doi.org/10.1038/s41467-019-09512-3
L. Berthier, M. Ozawa, and C. Scalliet, Configurational entropy of glass-forming liquids, J. Chem. Phys. 150, 160901 (2019), DOI 10.1063/1.5091961, https://doi.org/10.1063/1.5091961
B. Guiselin, G. Tarjus, and L. Berthier, On the overlap between configurations in glassy liquids, J. Chem. Phys. 153, 224502 (2020), DOI 10.1063/5.0022614, https://doi.org/10.1063/5.0022614
L. Berthier, P. Charbonneau, and J. Kundu, Finite Dimensional Vestige of Spinodal Criticality above the Dynamical Glass Transition, Phys. Rev. Lett. 125, 108001 (2020), DOI 10.1103/PhysRevLett.125.108001, https://doi.org/10.1103/PhysRevLett.125.108001
B. Guiselin, L. Berthier, and G. Tarjus, Statistical mechanics of coupled supercooled liquids in finite dimensions, SciPost Phys. 12, 091 (2022), DOI 10.21468/SciPostPhys.12.3.091, https://doi.org/10.21468/SciPostPhys.12.3.091 (arXiv:2105.08946, https://arxiv.org/abs/2105.08946)
L. Berthier and D. R. Reichman, Modern computational studies of the glass transition, Nat. Rev. Phys. 5, 42 (2023), DOI 10.1038/s42254-022-00548-x, https://doi.org/10.1038/s42254-022-00548-x
G. Jung, M. Ozawa, G. Biroli, and L. Berthier, Numerical investigation of the equilibrium Kauzmann transition in a two-dimensional atomistic glass, Proc. Natl. Acad. Sci. USA 123, e2612355123 (2026), DOI 10.1073/pnas.2612355123, https://doi.org/10.1073/pnas.2612355123 (arXiv:2507.03590, https://arxiv.org/abs/2507.03590)