Numerical investigation of the equilibrium Kauzmann transition in a two-dimensional atomistic glass
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Condensed Matter. The day in condensed matter: materials, phases and the measurements under them.
Kai: We're looking at the paper: "Numerical investigation of the equilibrium Kauzmann transition in a two-dimensional atomistic glass".
Mira: Here's the gist of the paper. Dense liquids gradually transform into non-equilibrium amorphous solids as they pass through the experimental glass transition.
Kai: This work overcomes a longstanding barrier by developing a computational strategy that combines three advanced Monte Carlo methods to access the equilibrium thermodynamic properties of a two-dimensional atomistic glass-former down to zero temperature across a range of system sizes.
Mira: This enables direct measurement of thermodynamic and structural observables providing evidence that the system undergoes a Kauzmann transition at a temperature that vanishes in the thermodynamic limit.
Kai: Computational Strategy for Equilibrium Sampling The numerical strategy developed integrates the complementary strengths of three Monte Carlo techniques to achieve ergodic sampling of the configuration space down to T = 0.
Title and authors: Kai: So we were just discussing the core idea behind "Numerical investigation of the equilibrium Kauzmann transition in a two-dimensional atomistic glass," which means they are checking for an equilibrium liquid-to-glass transition at zero temperature in two dimensions. We touched on how they managed to sample that low temperature space using swap Monte Carlo parallel tempering and population annealing, and their specific heat test confirmed equilibration.
Mira: And the results from those simulations strongly point toward a random first order transition towards a low-temperature equilibrium glass phase, which is characterized by this hierarchical organization of states two <ref:2507.03590#pg1>. They found that the transition temperature TK approaches zero as the system size gets larger.
Lev: That scaling behavior of TK approaching zero with increasing system size is what connects their numerical findings to the idea that there should be a zero temperature critical point in the thermodynamic limit twenty-five <ref:2507.03590#pg2>. It means if you could make a system infinitely large, this transition happens exactly at zero Kelvin.
Kai: What does that mean for us when we look at real systems? It suggests that for any finite system, you’re getting closer and closer to a state where an equilibrium glass phase might form right at absolute zero.
Mira: And they characterized the overlap between two independently sampled configurations using what they call the quenched Franz-Parisi potential, which shows a sharp change as temperature drops below TK, marking when the glass free energy becomes lower than that of the liquid three <ref:2507.03590#pg1>.
Lev: That potential measurement is key because it directly measures that free energy difference delta V between the liquid and glass states, which is what drives that transition three <ref:2507.03590#pg1>.
The paper's summary: Kai: To summarize what this paper does, they took a realistic off-lattice glass model for two dimensions and used their advanced Monte Carlo sampling to fully characterize the thermodynamics of the system down to zero temperature. This lets them look at both the ground state and how low energy excited states are organized.
Mira: The main implication is that they found evidence for a random first order transition towards an equilibrium glass phase, which they characterize by this hierarchical organization of low-temperature states two <ref:2507.03590#pg1>. They showed that above this Kauzmann temperature TK, the configurational entropy T Sc exists and vanishes at zero temperature.
Lev: So what's the practical meaning here? It’s providing the first glimpse of what a liquid-glass transition looks like in two dimensions, which is often much more complex than what we see in higher dimensions twenty-five <ref:2507.03590#pg2>.
Kai: And they also looked at how this organization happens by analyzing the low-lying minima in their potential energy landscape. They found that these states are organized into families based on mutual overlap greater than zero point eight.
Mira: That hierarchical structure is crucial because it suggests that a non trivial Kauzmann transition actually occurs right at zero temperature in this system two <ref:2507.03590#pg1>. It’s not just some random fluctuation; there’s an underlying organization.
Lev: From a hardware perspective, if you were designing a system where you wanted to study these glassy dynamics, knowing that the low-energy states are structured like that helps you narrow down where to look for those critical points three <ref:2507.03590#pg1>.
The paper's improvements: Kai: The authors suggest several improvements to how we approach this problem. They’re emphasizing the development of their original computational strategy, which is the combination of swap Monte Carlo parallel tempering and population annealing.
Mira: That method is the key improvement because it overcomes a longstanding sampling challenge in realistic off-lattice glass models, allowing them to achieve ergodic sampling down to T equals zero across a range of system sizes twenty-five <ref:2507.03590#pg2,across a range of system sizes>.
Lev: It’s important because traditional methods just can't get you there; you need that integrated approach to push the limits of what you can sample in this field three <ref:2507.03590#pg1>.
Kai: They also point out that the way they calculated the configurational entropy using different formalisms—like the harmonic approximation around inherent states and the Frenkel-Ladd method—actually agreed well enough for N equals seventy-seven.
Mira: That agreement is a big deal because it means they are confident that their estimate of the configurational entropy is physically meaningful, especially since it vanishes at zero temperature two <ref:2507.03590#pg1>. It settles some doubt about the entropy calculation itself.
Lev: So they're not just proving a result; they’re also validating the tools they used to measure things, which makes the whole picture much more solid for future work three <ref:2507.03590#pg1>.
Conclusion: Kai: So wrapping up this paper "Numerical investigation of the equilibrium Kauzmann transition in a two-dimensional atomistic glass," we see that their combined simulation techniques successfully probed the equilibrium liquid-to-glass transition down to zero temperature.
Mira: The main point is that they observed a random first order transition towards an equilibrium glass phase, defined by this hierarchical organization of low-temperature states two <ref:2507.03590#pg1>. Crucially, they showed that the Kauzmann temperature TK approaches zero as the system size increases.
Lev: This means for large systems, we have compelling evidence for a zero temperature critical point at TK equals zero in the thermodynamic limit twenty-five <ref:2507.03590#pg2>. It’s a very concrete prediction based on their numerical analysis.
Kai: And they directly addressed the Kauzmann paradox by showing that the configurational entropy vanishes at zero temperature, which excludes the idea of a finite temperature entropy crisis three <ref:2507.03590#pg1>. They are confident this solution is physically meaningful.
Mira: It provides a direct measurement of thermodynamic observables down to T equals zero, giving us strong evidence for a transition at absolute zero in two dimensions. We should be looking at how they test this picture in more complex models and three-dimensional systems next.
Lev: I think the real challenge moving forward is taking these findings from two dimensions and seeing if we can replicate that hierarchical organization in three-dimensional materials three <ref:2507.03590#pg1>.
Kai: That sounds like a good plan. Thanks to Kai, Mira, and Lev for joining us today on this one.
Gerhard Jung, Misaki Ozawa, Giulio Biroli, Ludovic Berthier
Laboratoire Interdisciplinaire de Physique (LIPhy) · Laboratoire de Physique de l’Ecole Normale Sup´erieure, ENS · Gulliver, CNRS UMR 7083, ESPCI Paris, PSL Research University
cond-mat.soft, cond-mat.dis-nn, cond-mat.mtrl-sci
Submitted: 2025-07-04
Updated: 2026-11-06
Journal ref: Proc. Natl. Acad. Sci USA 123, e2612355123 (2026)
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Importance score: 78/100
The gist: Dense liquids gradually transform into non-equilibrium amorphous solids as they pass through the experimental glass transition.
Key concepts
- Kauzmann Transition
- This is a theoretical transition where the configurational entropy of a system drops to zero. The paper investigates whether this happens at a finite temperature or if it occurs only in the limit where temperature approaches zero, indicating a critical point at absolute zero.
- Configurational Entropy ($s_{harm}$)
- This measures the number of ways the atoms can arrange themselves in a disordered state. The study shows that this entropy vanishes exactly at T=0, which strongly suggests that the proposed estimate for configurational entropy is physically meaningful and resolves the paradox.
- Franz-Parisi Potential ($eta V(Q)$)
- This potential quantifies the overlap ($Q$) between two independent atomic configurations. A sharp change in this potential signals a Kauzmann transition, marking the point where the free energy of a glass state becomes lower than that of a liquid state at temperature TK.
Terminology
Summary
Dense liquids gradually transform into non-equilibrium amorphous solids as they pass through the experimental glass transition. p. 0 This work overcomes a longstanding barrier by developing a computational strategy that combines three advanced Monte Carlo methods to access the equilibrium thermodynamic properties of a two-dimensional atomistic glass-former down to zero temperature across a range of system sizes. p. 0 This enables direct measurement of thermodynamic and structural observables providing evidence that the system undergoes a Kauzmann transition at a temperature that vanishes in the thermodynamic limit. p. 0
Computational Strategy for Equilibrium Sampling
The numerical strategy developed integrates the complementary strengths of three Monte Carlo techniques to achieve ergodic sampling of the configuration space down to T = 0. This approach is composed of two main steps: first, combining swap MC with parallel tempering to produce large sets of independent equilibrium configurations down to T = 0. p. 1105, and second, using these configurations as starting points for the population annealing algorithm to push equilibration limits to much lower temperatures. p. 1 The success of this strategy is validated by observing that the time-dependent specific heat cV (t = Ncτ) attains a plateau at all investigated temperatures down to T → 0, indicating equilibration. p. 9
Characterization of Configurational Entropy
The authors employed three different techniques to evaluate the configurational entropy, including an expansion near inherent states, the Frenkel-Ladd method, and a Shannon entropy associated with the distribution of inherent states. The analysis shows that while these distinct methodologies agree remarkably well in their estimates for N = 77, the configurational entropy vanishes at T = 0 exactly. p. 2 Specifically, the observation that s harm c(T → 0) → 0 is a strong indication that the proposed estimate of the configurational entropy is physically meaningful. p. 1
Evidence for a Zero-Temperature Critical Point
The study demonstrates that the Kauzmann transition resembles a conventional first-order phase transition associated with a jump in the order parameter Q, where the free energy difference ∆V = V (Qglass)−V (Qliq) vanishes at TK < Tg < Tmin < Tg < Tsc < 0. The evolution of the critical Kauzmann temperature TK with system size suggests that TK → 0 for large N, which is consistent with the entropy extrapolation for N = 704 in Fig. p. 3 2b and indicates a zero-temperature critical point at TK = 0 in the thermodynamic limit. p. 4
Hierarchical Organization of Low-Energy States
The analysis of low-lying minima reveals a complex, hierarchical structure at the bottom of the energy landscape, where inherent states are grouped into families based on mutual overlap Q > 0.8. This organization is visualized through a tree and matrix representation showing how families cluster into higher-order families, suggesting that a non-trivial Kauzmann transition occurs at TK = 0 in the system. p. 5
The Role of System Size and Model Specifics
The study uses a ternary mixture of Lennard-Jones particles in two dimensions (KA2D) with specific interaction parameters, and the results show that for N = 44 and 77, the transition temperature TK is identified as 0.057. Furthermore, for N = 66, a peculiar two-step increase in average overlap occurs at lower temperatures, suggesting some kind of order or symmetry unlike the completely disordered structures seen for other system sizes. The findings provide a foundation for future simulation work on larger systems and three-dimensional materials. p. 0
Overlap and Franz-Parisi Potential
The overlap Q between two independently sampled configurations is quantified by the quenched Franz-Parisi potential, βV (Q) = − 1/N X Σ m=1 Σ n=1 ln X Ns m n δ(Q − Qmn). This potential reveals a sharp change from small to large overlap at TK, where the glass free energy becomes lower than that of the liquid, and it shows that transitions such as those shown in Fig. p. 4 3 have only been observed for structural glass-forming liquids using biased Monte Carlo simulations. p. 0 The transition temperature TK depends on N, with TK rapidly shifting towards zero as N increases. p. 4
Amorphous Ground States and Hierarchy
The ground states obtained for all investigated N are completely amorphous, suggesting that crystallization is efficiently suppressed. The analysis of low-energy minima reveals that the bottom of the potential energy landscape is populated by minima whose energy differences are much smaller than the energy per particle, yet these correspond to completely different configurations in real space with very small mutual overlap. p. 6 This organization reflects a complex, hierarchical structure at the bottom of the energy landscape. p. 5 The final order of states is iteratively determined by starting with the state with lowest inherent energy and including all those states that belong to families which are similar to the selected family. p. 4
Conclusion on the Kauzmann Paradox
The observation that s harm c(T → 0) → 0 is thus a strong indication that we have correctly sampled the equilibrium configurations at low temperatures, and that the proposed estimate of the configurational entropy is physically meaningful. The observation of this vanishing at T = 0 in this two-dimensional system directly excludes a finite-temperature entropy crisis, providing a solution to the Kauzmann paradox. p. 2 This conclusion is supported by direct equilibrium measurements down to T = 0. p. 6 The testing of this picture in more complex models and three-dimensional systems remains an exciting and important challenge. p. 6
Methods for Free Energy Measurement
The authors used the Franz-Parisi potential to measure the free energy difference ∆V = V (Qglass)−V (Qliq), which can lead to a genuine entropy crisis. This is calculated by measuring the mutual overlap Qmn between two independent equilibrium configurations m and n using Eq. p. 4 (1). The measurement of the Franz-Parisi potential down to T = 0 reveals a Kauzmann transition from the liquid to the ideal glass phase, taking place when the glass free energy becomes lower than that of the liquid. p. 4 This is achieved by measuring Qmn using Ns = 103 independent configurations from independent SMC+PT+PA runs. p. 9 The results show that the transition temperature TK approaches zero in the large-N limit, providing compelling evidence for a zero-temperature critical point at TK = 0 in the thermodynamic limit. p. 4
Sampling Techniques and Convergence Tests
Equilibration is achieved by the unique combination of swap Monte Carlo (SMC), parallel tempering (PT), and population annealing (PA) techniques. The specific heat test cV(t) serves as a reliable measure to test whether an ensemble represents an equilibrium ensemble, and the plateau attainment indicates equilibration. p. 8 The combination of SMC/PT/PA is vital to ensure equilibration down to T → 0, as neither technique alone can achieve this goal. p. 9 The specific heat is a superior measure for equilibration compared to the potential energy since its equilibration time is much longer, consistent with previous findings. p. 9
Entropy Calculation Formalisms
The configurational entropy s harm c is calculated as s harm c = stot − s harm glass − s anh. The vanishing of the configurational entropy at T = 0 results from a cancellation of three large terms in Eq. p. 0 (S14), where stot is obtained from a thermodynamic integration starting at T = ∞, while s harm glass + s anh is based on the harmonic approximation around the T → 0 inherent state. p. 2 The observation that s harm c(T → 0) → 0 is thus a strong indication that we have correctly sampled the equilibrium configurations at low temperatures. p. 9 The results show very good agreement between the glass entropy calculated from the Frenkel-Ladd method, s FL glass, and the harmonic approximation. p. 3 The use of these methods confirms that inherent states are not required to define a configurational entropy. p. 3
Finite-Size Corrections and Model Details
The system is modeled as a modified Kob-Andersen mixture in two dimensions (KA2D) with specific interaction parameters and particle density ρ = 1.19. The linear box sizes Lx(N) are specified for various system sizes N, ranging from N = 44 to N = 704. p. 8 The finite-size corrections in the calculation of s harm c account to ∆stot(N = 77) > 0. p. 403 which is a very relevant contribution, especially at low temperatures. p. 4 The study confirms that the transition temperature TK < 0. p. 001 in the large-N limit, which is substantially lower than the glass transition temperature Tg = 0. p. 415 and conspicuously close to TK = 0. p. 4 The results show that N = 66 features some kind of order or symmetry unlike the completely disordered structures seen for other system sizes.
Specific Case of Sixteenth Particles
For N = 66, the average overlap grows to ⟨Q⟩ ≈ 0.5 at T = 0.035, but the final transition towards ⟨Q⟩ > 0. p. 07 when all configurations correspond to the ground state only happens at lower temperatures, T < 0. p. 402. This behavior can be rationalized by analyzing the configurations themselves, which show that the ground state features a pronounced symmetry along the x = y and x = −y axes. The study provides compelling evidence for a zero-temperature critical point at TK = 0 in two-dimensional glasses. p. 4
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Simulation of equilibrium thermodynamics down to zero temperature for two-dimensional atomistic glass-formers across a range of system sizes is now possible using a combination of swap Monte Carlo parallel tempering and population annealing algorithms. This enables researchers to
directly measure thermodynamic and structural observables that provide unambiguous evidence that the system undergoes a Kauzmann transition at a temperature that vanishes in the thermodynamic limit.
-
The improved AI can directly probe the statistical mechanics of the bulk transition to a non-ergodic glass state, providing
the first demonstration that computer simulations can fully probe the statistical mechanics of the bulk transition to a non-ergodic glass state.
-
The system can now characterize low-lying excited states and the organization of low-lying minima in complex potential energy landscapes. This allows for a detailed analysis of
the structure of the potential energy landscape
and investigation intolocalized excitations
which are found to proliferate with system size. -
The AI can determine the existence of a zero-temperature critical point at Kauzmann temperature TK = 0 in two dimensions by analyzing the evolution of the Franz-Parisi free energy potential V(Q). This transition is characterized by a jump in order parameter Q from Qliq to Qglass as T decreases below TK.
-
The system can provide a direct exclusion of the Kauzmann paradox solution involving a finite temperature entropy crisis, demonstrating that
sc only vanishes at T = 0
in this two-dimensional system.
Sources
- The Ideal Glass and the Ideal Disk Packing in Two Dimensions
- Ideal noncrystals: A possible new class of ordered matter without apparent broken symmetry
- Molecular motion at the experimental glass transition
- Starting from the amorphous ground state: linking landscape thermodynamics to slow dynamics and crossover
Related papers
- Coarse-graining dense, deformable active particles
- Three-body corrections to the Brownian viscosity of hard spheres without hydrodynamic interactions
- Thin active nematohydrodynamic layers: asymptotic theories and instabilities
- Origins of suppressed self-diffusion of nanoscale constituents of a complex liquid
- Soft material mechanical property degradation at high strain rates
- Avalanche criticality emerges by thermal fluctuation in a quiescent glass