Dynamics of viscous liquids and the Random Barrier Model

This study demonstrates that the parameter-free Random Barrier Model outperforms the von Schweidler law in fitting the inherent dynamics of a viscous ternary Lennard-Jones liquid and accurately predicting its diffusion coefficient, despite the model's unrealistic assumption of identical energy minima.

The Gist

Here is an explanation of the paper using simple language and everyday analogies.

The Big Picture: The "Frozen" Liquid Mystery

Imagine a glass of water. It flows easily. Now, imagine that same water getting colder and colder until it turns into ice. But what happens in between? When you cool a liquid like honey or syrup, it gets thicker and thicker until it stops flowing and becomes a solid-like "glass."

Scientists have been trying to understand the rules of how molecules move in this "thick, sticky" stage. It's like watching a crowded dance floor where everyone is moving very slowly, bumping into each other, and getting stuck.

This paper asks a simple question: Is there a single, universal rule that explains how these molecules move, no matter what kind of liquid they are?

The Two Competing Theories

To answer this, the researchers compared two different "maps" or theories that try to predict how these molecules move:

1. The "Von Schweidler" Map (The Old School Rule):
   Think of this like a standard GPS route. It has a few knobs you can turn (parameters) to adjust the route based on traffic. It's flexible, but because you have to tweak the knobs, it might just be "fitting" the data rather than finding the true underlying rule.

2. The "Random Barrier Model" (RBM) (The Universal Rule):
   Imagine a giant, chaotic maze made of walls. The walls have different heights (barriers). The rule here is simple: "Particles are like hikers trying to jump over these walls."

   • The Twist: This model assumes all the "floors" (energy levels) in the maze are exactly the same height. In real life, liquids have floors of different heights, so this seems unrealistic.
   • The Magic: Despite this unrealistic assumption, this model has zero adjustable knobs. It's a rigid, mathematical prediction. If it works, it means the universe is following a very specific, simple law.

The Experiment: The "Super-Computer" Dance

The researchers didn't just guess; they ran massive computer simulations.

• The Setup: They created a digital liquid made of 12,000 tiny particles (like a digital crowd).
• The Problem: At very low temperatures, these particles move so slowly that a normal computer simulation would take thousands of years to see them move even a tiny bit.
• The Solution: They used two super-smart tricks:
  1. The "Swap" Trick: To get the liquid ready (equilibrated) at low temperatures, they allowed particles to swap places instantly, like a dance partner switch, to speed things up.
  2. The "Inherent State" Filter: Once the simulation was running, they used a special filter. Imagine taking a photo of a dancer mid-jump, then instantly dropping them to the floor to see where they would land if they stopped vibrating. This removes the "jitter" (thermal vibrations) and shows the true path the particle took to get from A to B.

The Results: The Surprise Winner

They compared the computer data against the two maps (Von Schweidler vs. RBM).

• The Result: The Random Barrier Model (RBM) won. It fit the data much better than the flexible Von Schweidler map, even though the RBM had no adjustable knobs to "cheat" and fit the curve.
• The Analogy: It's like trying to predict the path of a ball rolling down a hill.
  • The Von Schweidler method is like drawing a line that you can bend to match the ball's path perfectly.
  • The RBM method is like a rigid, pre-drawn line based on a simple law of physics.
  • The Shock: The rigid, pre-drawn line (RBM) actually matched the ball's path better than the bendable one, even though the ball was rolling on a bumpy, complex hill, not a smooth, idealized one.

Why Does This Matter?

1. Prediction Power: Because the RBM is so accurate, scientists can run a short simulation (a few hours) and use the RBM to predict exactly how the liquid will behave over a million years. It's like looking at a few seconds of a movie and knowing exactly how the ending will play out.
2. The Mystery Remains: The biggest puzzle is why the RBM works. The RBM assumes all the "floors" in the maze are the same height. But in real liquids, the energy "floors" are all different heights.
   • The Metaphor: It's like predicting how a car drives through a city by assuming every street is perfectly flat. In reality, the city has huge hills and valleys. Yet, the "flat street" model predicts the car's speed perfectly.
   • The Conclusion: The researchers don't know why this works yet. It suggests that deep down, the messy, complex world of glassy liquids might be governed by a surprisingly simple, universal law that we haven't fully understood yet.

Summary in One Sentence

By using super-computers to filter out the "noise" of vibrating atoms, scientists discovered that a simple, rigid mathematical model (which assumes a perfectly uniform world) predicts the movement of messy, real-world liquids better than complex, adjustable models, hinting at a hidden universal law of nature.

Technical Summary

Here is a detailed technical summary of the paper "Dynamics of viscous liquids and the Random Barrier Model" by Schrøder, Dyre, and Scalliet.

1. Problem Statement

The paper addresses two central questions regarding the dynamics of highly viscous, glass-forming liquids:

1. The Universality Paradox: Experimental and simulation data suggest that the motion of atoms in viscous liquids follows a "universal" shape described by the Random Barrier Model (RBM) in the extreme disorder limit. However, the RBM relies on unrealistic assumptions for liquids, specifically that particles jump between sites of identical energy (implying zero inherent specific heat). Real liquids possess a broad distribution of inherent state energies (potential energy minima). The authors ask: Why does a model assuming identical energy minima accurately describe the dynamics of liquids with broad energy distributions?
2. Historical Oversight: Given that viscous liquid dynamics have been studied for decades, why has this apparent RBM universality not been observed previously?

2. Methodology

To investigate these questions, the authors employed state-of-the-art simulation techniques to access the "extremely viscous regime" (temperatures significantly lower than those previously studied), where dynamics are roughly 30 times slower than in prior studies.

• Model System:

  • A ternary Lennard-Jones (LJ) mixture (referred to as the KA2 model) consisting of three particle types (A, B, C) in a 4:1:1 ratio.
  • This model extends the standard Kob-Andersen (KA) binary model by adding a third particle type to prevent crystallization and enhance equilibration efficiency via particle-swap dynamics.
  • Simulations were performed at a density $\rho = 1.35$ to ensure liquid stability and avoid gas-glass instabilities.
  • Temperatures ranged from $T \approx 0.65$ down to $T = 0.509$ (well below the Mode-Coupling Theory transition temperature $T_{mct} \approx 0.62$).

• Simulation Techniques:

  • Equilibration: A hybrid Molecular Dynamics (MD) / Particle-Swap Monte Carlo algorithm was used to generate equilibrium configurations at very low temperatures where standard MD fails due to ergodicity breaking.
  • Dynamics: Long-duration GPU-accelerated MD simulations (using the RUMD package) were run on the equilibrated configurations to track equilibrium relaxation dynamics over unprecedented time scales (up to $4.3 \times 10^7$ time units).
  • Observables: The primary observable was the Mean-Squared Displacement (MSD), $\langle \Delta r^2(t) \rangle$. Crucially, the authors analyzed the Inherent State (IS) MSD, obtained by quenching configurations along the trajectory to their nearest potential energy minima. This removes the contribution of thermal vibrations (the short-time plateau), isolating the structural relaxation dynamics.

• Comparative Analysis:
  The authors compared the IS MSD data against two functional forms:

  1. Inherent von Schweidler (IvS) Law: An empirical expression derived from Mode-Coupling Theory (MCT) containing one dimensionless free parameter (the exponent $b$).
  2. Random Barrier Model (RBM) Prediction: A theoretical solution for diffusion in the extreme disorder limit containing no dimensionless free parameters (shape parameters).

3. Key Contributions

• Extension to Lower Temperatures: The study successfully pushed simulations into a regime where the alpha-relaxation time is four orders of magnitude larger than at $T_{mct}$, allowing for the observation of dynamics closer to the experimental glass transition.
• Validation of IS MSD Universality: The authors demonstrated that the Inherent MSD collapses onto a single master curve when scaled by the RBM prediction, confirming the universality of the RBM for the structural component of liquid dynamics.
• Superior Predictive Power: The paper establishes that the RBM (with zero free parameters) outperforms the von Schweidler law (with one free parameter) in both fitting accuracy and, more importantly, in extrapolating long-time behavior from short-time data.

4. Results

• Fit Quality:

  • Both the IvS law and the RBM fit the data reasonably well at higher temperatures (near $T_{mct}$).
  • As temperature decreases, the RBM fit remains robust, while the IvS fit deteriorates. The residuals (log-ratio of data to fit) show that the IvS law increasingly overestimates or underestimates the MSD at long times and low temperatures.
  • The RBM fits the data within 10% error across the entire temperature range, despite having no adjustable shape parameters.

• Predictive Extrapolation:

  • When fitting data only up to a short time window ($t < 3 \times 10^5$) and extrapolating to longer times, the IvS law significantly overestimates the diffusion coefficient ($D$). The shape parameter $b$ in the IvS law varies significantly (by >30%) depending on the fitting window and does not converge even over nine decades of time.
  • In contrast, the RBM extrapolation is highly accurate. The estimated diffusion coefficient varies by less than 10% even when the fitting window is extended by two decades. This allows for the reliable extraction of $D$ in regimes where the diffusive limit is not directly reachable by brute-force simulation.

• Thermal vs. Inherent Dynamics:

  • While the Inherent MSD collapses perfectly onto the RBM master curve, the Thermal (True) MSD does not. The thermal MSD includes an additive constant representing "cage rattling" (vibrations), which is non-universal and depends on the specific particle type and temperature. This explains why RBM universality was not observed in earlier studies that focused on total MSD.

• Generalizability:

  • The authors extended the analysis to a polydisperse LJ model (from previous literature) and found similar agreement between the IS MSD and the RBM, even at temperatures corresponding to an extrapolated relaxation time 11 decades longer than the onset of glassy dynamics.

5. Significance and Discussion

• Resolution of the "Identical Energy" Paradox: The results suggest that while the energies of inherent states in real liquids are broadly distributed, the dynamics of structural relaxation are dominated by the percolation of the lowest energy barriers. The RBM captures this percolation physics, which is robust against the details of the energy distribution, provided the distribution is continuous at the percolation threshold.
• Theoretical Challenge: The paper highlights a major open question: Why does a model assuming zero inherent specific heat (identical minima) so accurately reproduce the dynamics of systems with high specific heat? This challenges current theoretical frameworks and suggests that the "inherent" dynamics are governed by a universal mechanism (percolation on the lowest energy links) that is insensitive to the specific energy landscape topology.
• Practical Utility: The RBM provides a powerful tool for predicting diffusion coefficients in supercooled liquids from short-time simulations, bypassing the computational cost of waiting for the system to reach the diffusive regime. This is crucial for testing theories like the Stokes-Einstein breakdown in the extremely viscous regime.

In conclusion, the paper provides strong evidence that the Random Barrier Model describes the inherent structural dynamics of glass-forming liquids with remarkable universality, outperforming traditional MCT-based fits, and offering a robust method for extrapolating transport properties in the deeply supercooled regime.