Abstract
We study the influence of the softness of the interparticle interactions on the fragility of a glass former by considering three model binary mixture glass formers. The interaction potential between particles is a modified Lennard-Jones type potential, with the repulsive part of the potential varying with an inverse power q of the interparticle distance, and the attractive part varying with an inverse power p. We consider the combinations (12,11) (model I), (12,6) (model II), and (8,5) (model III) for (q,p) such that the interaction potential becomes softer from model I to III. We evaluate the kinetic fragilities from the temperature variation of diffusion coefficients and relaxation times, and a thermodynamic fragility from the temperature variation of the configurational entropy. We find that the kinetic fragility increases with increasing softness of the potential, consistent with previous results for these model systems, but at variance with the thermodynamic fragility, which decreases with increasing softness of the interactions, as well as expectations from earlier results. We rationalize our results by considering the full form of the Adam-Gibbs relation, which requires, in addition to the temperature dependence of the configurational entropy, knowledge of the high temperature activation energies in order to determine fragility. We show that consideration of the scaling of the high temperature activation energy with the liquid density, analyzed in recent studies, provides a partial rationalization of the observed behavior.
Introduction
In the study of glass-forming liquids, "fragility" is a critical parameter describing how abruptly a liquid's relaxation time or viscosity increases as it approaches the glass transition temperature ($T_g$). "Strong" liquids, like silica, follow an Arrhenius-like slow-down, whereas "fragile" liquids exhibit a sudden, super-Arrhenius plunge into rigidity.
A long-standing hypothesis in condensed matter physics suggests that the "softness" of the interparticle potential is a primary driver of fragility. Intuitively, one might expect softer particles to pack more easily, leading to a smoother energy landscape and lower fragility. However, this paper reveals a surprising paradox: making interactions softer actually increases kinetic fragility while decreasing thermodynamic fragility.
Methods and Models
The researchers employed Molecular Dynamics (MD) simulations of a modified Kob-Andersen binary mixture. By varying the exponents $(q, p)$ of the repulsive and attractive components of the potential, they created three distinct models representing a gradient of softness:
- Model I (12, 11): The "hardest" interaction.
- Model II (12, 6): The standard Lennard-Jones potential.
- Model III (8, 5): The "softest" interaction.
Kinetic fragility ($K_{VFT}$) was derived from the temperature dependence of diffusion and relaxation times, while thermodynamic fragility ($K_T$) was calculated from the rate at which configurational entropy ($S_c$) vanishes during cooling.
The Fragility Paradox and the Adam-Gibbs Relation
The most striking result is the divergence between kinetic and thermodynamic measures. As the potential softens (Model I $\rightarrow$ III), the liquid becomes kinetically more fragile but thermodynamically less fragile.
To resolve this, the authors turned to the full Adam-Gibbs (AG) relation. Standard interpretations often assume the high-temperature activation energy term ($A$) is constant. This study demonstrates that $A$ varies significantly with softness. In softer systems, the high-temperature activation energy ($E_0$) drops much faster than the entropy-related terms. This drop effectively "stretches" the dynamics, making the liquid appear more fragile kinetically even if its entropy landscape suggests otherwise.
Key Results
- Kinetic Fragility ($K_{VFT}$): Rose from 0.20 in the hardest model (I) to 0.26 in the softest model (III).
- Thermodynamic Fragility ($K_T$): Dropped significantly from 0.551 (Model I) to 0.211 (Model III).
- Activation Energy ($E_0$): The high-temperature activation energy for Model I (5.67) was more than four times higher than for Model III (1.28).
Conclusion
By accounting for the model-dependent activation energy in the full Adam-Gibbs relation, the authors successfully rationalized why kinetic data shows a trend opposite to thermodynamic expectations. This work highlights that fragility is not just a low-temperature phenomenon but is heavily influenced by high-temperature dynamics and the specific geometry of the potential energy landscape. Future research will look to refine the understanding of the activation energy parameter and explore how simultaneous changes in density and softness affect these transitions.
Summary updated