Abstract
Glass is a microscopically disordered, solid form of matter that results when a fluid is cooled or compressed in such a fashion that it does not crystallise. Almost all types of materials are capable of glass formation -- polymers, metal alloys, and molten salts, to name a few. Given such diversity, organising principles which systematise data concerning glass formation are invaluable. One such principle is the classification of glass formers according to their fragility. Fragility measures the rapidity with which a liquid's properties such as viscosity change as the glassy state is approached. Although the relationship between features of the energy landscape of a glass former, its configurational entropy and fragility have been analysed previously, an understanding of the origins of fragility in these features is far from being well established. Results for a model liquid, whose fragility depends on its bulk density, are presented in this letter. Analysis of the relationship between fragility and quantitative measures of the energy landscape reveal that the fragility depends on changes in the vibrational properties of individual energy basins, in addition to the total number of such basins present, and their spread in energy. A thermodynamic expression for fragility is derived, which is in quantitative agreement with kinetic fragilities obtained from the liquid's diffusivity.
The Fragility Problem
When liquids cool toward a glass transition, their viscosity increases dramatically. However, the rate of this increase varies significantly between materials. "Strong" liquids, such as silica, exhibit a steady, Arrhenius-like increase in viscosity. In contrast, "fragile" liquids, like many polymers, maintain low viscosity until they are very close to the glass transition, at which point it increases sharply.
Physicists use the Potential Energy Landscape (PEL)—a high-dimensional map of all possible particle configurations and their associated energies—to understand this behavior. While previous theories suggested fragility was a simple function of the total number of energy minima (basins), these models often failed to match simulation data. This study addresses the missing link: the specific features of the energy landscape that dictate whether a liquid is strong or fragile.
Methods: Mapping the Landscape
The study employs Molecular Dynamics (MD) simulations of a binary Lennard-Jones mixture, a standard model for glass-forming liquids. By using the Inherent Structure (IS) formalism, liquid configurations are mathematically "quenched" to their nearest local energy minimum. This allows the partitioning of the complex landscape into individual basins.
The author calculates Configurational Entropy ($S_c$)—the entropy related to the number of available basins—using both indirect Thermodynamic Integration (TI) and a direct PEL-based counting method. A key innovation is the explicit calculation of basin entropy ($S_{vib}$), which represents the "wiggle room" particles have within a single energy minimum, and how this entropy varies between basins.
Key Insights: The Vibrational Correction
A major finding is that higher-energy basins, visited at higher temperatures, have different vibrational properties than lower-energy ones. In this constant-density model, higher-energy basins are "stiffer," possessing lower vibrational entropy.
Previous models predicted that fragility should depend solely on the total number of states. However, this study found that as density increases, the number of states decreases while the liquid becomes more fragile. This apparent contradiction is resolved by the basin entropy change ($\delta S$). The rate at which vibrational "wiggle room" changes as the liquid moves between basins of different energies is the crucial factor determining the net rate of entropy loss and, consequently, the fragility.
Results: Reconciling Thermodynamics and Kinetics
The author derives a new thermodynamic fragility index ($K_{AG}^{PEL}$) that incorporates both the energy spread of basins and the vibrational shift. This index shows remarkable quantitative agreement with the kinetic fragility ($K_{VFT}$) calculated directly from diffusivity data.
The results demonstrate that increasing the density of the Lennard-Jones liquid makes it significantly more fragile. Furthermore, the study provides strong validation for the Adam-Gibbs relationship, which links the dynamics of a liquid (diffusivity) directly to its configurational entropy.
Conclusion
By accounting for the vibrational properties of energy basins, this work provides a robust theoretical framework for understanding fragility. It demonstrates that the "stiffness" of energy basins and how it changes with energy is as important as the number of basins themselves. While the harmonic approximation used here has limits at high temperatures, the model successfully bridges the gap between microscopic landscape features and macroscopic glass-forming behavior.
Summary updated