Power law relationship between diffusion coefficients in multi-component\n glass forming liquids

Anshul D. S. Parmar, S. Sengupta, Srikanth SastryarXiv 2018 · DOI · arXiv

Abstract

In multi-component glass-forming liquids, the dynamics of different species often decouple as the system approaches the glass transition. This study demonstrates a robust power-law relationship between the diffusion coefficients of fast and slow species ($D_{fast} \propto D_{slow}^\gamma$) across a wide range of temperatures and densities. This relationship, confirmed through extensive molecular dynamics simulations and experimental data, provides a critical bridge for theoretical models like the Adam-Gibbs relation, ensuring their internal consistency even when individual component dynamics diverge.

Introduction

As a liquid cools toward its glass transition, its viscosity increases by many orders of magnitude. The traditional Stokes-Einstein (SE) relation, which posits that diffusion is inversely proportional to viscosity ($D \propto T/\eta$), typically breaks down in this supercooled regime—a phenomenon known as decoupling. In multi-component systems, such as metallic glasses, this complexity is compounded by species-species decoupling, where different atoms diffuse at significantly different rates. This raises a fundamental question for the Adam-Gibbs (AG) relation: if the dynamics of components diverge, can a single framework still describe the entire mixture?

Methodology and Datasets

Researchers conducted a meta-analysis and simulation study of 22 distinct glass-forming systems. These included 15 computer-simulated models (spanning 2D, 3D, and 4D) and experimental data from five-component bulk metallic glasses like Vitreloy 1 and Vitreloy 4. The models covered a broad spectrum of interactions, from soft spheres and Lennard-Jones particles to network formers like silica.

The Universal Power Law

The central finding of this research is the existence of a robust power-law relationship between the diffusion coefficients of different species: $D_f \propto D_s^\gamma$. This relationship holds remarkably well across all systems studied, regardless of dimensionality or specific interatomic potentials.

The exponent $\gamma$ is found to be system-specific, typically ranging from 0.5 to 1.0. It appears sensitive to the "softness" of the potential and the mixture's composition. When $\gamma < 1$, the "fast" species becomes increasingly faster relative to the "slow" species as the liquid cools, providing a quantitative measure of species-species decoupling.

Implications for Theoretical Models

This power-law relationship effectively "saves" the Adam-Gibbs relation for multi-component mixtures. If the slow component's dynamics are governed by the AG relation (linking relaxation to configurational entropy), the power law ensures that the fast component follows the same functional form, albeit with a scaled activation energy. This mathematical consistency allows the AG framework to remain a viable universal description of the glass transition in complex mixtures.

Conclusion

The discovery of a near-universal power law between diffusion coefficients simplifies our understanding of transport in multi-component liquids. While the microscopic origin of the specific exponent $\gamma$ remains an area for future research, its empirical robustness across diverse systems—from simple hard spheres to complex metallic alloys—marks a significant step forward in glass physics. Future work will likely focus on deriving $\gamma$ from first principles and applying these insights to industrial glass-forming materials.

Summary updated

Abstract

The slow down of dynamics in glass forming liquids as the glass transition is\napproached has been characterised through the Adam-Gibbs relation, which\nrelates relaxation time scales to the configurational entropy. The Adam-Gibbs\nrelation cannot apply simultaneously to all relaxation times scales unless they\nare coupled, and exhibit closely related temperature dependences. The breakdown\nof the Stokes-Einstein relation presents an interesting situation to the\ncontrary, and in analysing it, it has recently been shown that the Adam-Gibbs\nrelation applies to diffusion coefficients rather than to viscosity or\nstructural relaxation times related to the decay of density fluctuations.\nHowever, for multi-component liquids -- the typical cases considered in\ncomputer simulations, metallic glass formers, etc. -- such a statement raises\nthe question of which diffusion coefficient is described by the Adam-Gibbs\nrelation. All diffusion coefficients can be consistently described by the\nAdam-Gibbs relation if they bear a power law relationship with each other.\nRemarkably, we find that for a wide range of glass formers, and for a wide\nrange of temperatures spanning the normal and the slow relaxation regimes, such\na relationship holds. We briefly discuss possible rationalisations of the\nobserved behaviour.\n
BibTeX
@misc{parmar2018power,
  author = {Anshul D. S. Parmar and S. Sengupta and Srikanth Sastry},
  title = {{Power law relationship between diffusion coefficients in multi-component\textbackslash\{\}n glass forming liquids}},
  year = {2018},
  eprint = {1805.04671},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1805.04671},
}

Loading related content...

Figure