Spectral Decomposition of Liquid Viscosity into Instantaneous Normal Modes

This paper presents a theoretical framework that decomposes liquid viscosity into contributions from individual instantaneous normal modes, revealing that unstable localized modes dominate viscous dynamics above the mode-coupling temperature while stable modes take over below it, thereby establishing a quantitative link between macroscopic viscosity and microscopic atomic excitations.

The Gist

Imagine a pot of honey. When you stir it, it resists your spoon. That resistance is called viscosity. For a long time, scientists have known that honey is thick, but they haven't fully understood why at the level of individual atoms. It's like knowing a car won't start, but not knowing if it's the battery, the fuel, or the spark plugs.

This paper acts like a high-tech mechanic's diagnostic tool. It takes the "thick" liquid and breaks it down into its tiniest, fastest vibrations to see exactly which ones are causing the resistance.

Here is the story of what they found, explained simply:

1. The "Instantaneous Snapshot" (The INM)

Liquids are messy. Unlike a solid crystal where atoms sit in a neat grid, liquid atoms are constantly jiggling and rearranging. Because they are always moving, you can't take a "perfect" photo of their arrangement.

However, the scientists used a clever trick: they took instantaneous snapshots of the liquid. Imagine freezing a chaotic dance floor for a split second. In that frozen moment, they calculated how every single atom would vibrate if it were stuck in that exact position. They call these "Instantaneous Normal Modes" (INMs). Think of these as the specific "notes" or "tunes" the liquid is humming at that exact moment.

2. The Two Types of "Notes"

When they listened to these notes, they found two very different kinds of vibrations, which they separated like sorting red and blue marbles:

• The "Wobbly" Notes (Unstable Modes): These are vibrations where the atoms are in a precarious spot, like a ball balanced on top of a hill. If you nudge them, they roll down.
  • Some of these wobbly notes are Delocalized: The whole crowd of atoms is wiggling together, like a stadium wave.
  • Some are Localized: Only a tiny, specific group of atoms is wiggling wildly, like a single person tripping in a crowd.
• The "Stable" Notes: These are atoms sitting in a valley (a comfortable spot). They vibrate gently but don't roll away.

3. The Big Discovery: Who is the "Traffic Cop"?

The team wanted to know: Which of these notes is actually responsible for the liquid's thickness (viscosity)?

They ran massive computer simulations of three different liquids (two metallic glasses and a standard model liquid) and compared their results against real-world data. Here is what they found:

• The "Wobbly, Localized" Notes are the Culprits: They discovered that the resistance to flow (viscosity) is almost entirely caused by those tiny, chaotic, localized groups of atoms that are balanced on the edge of a hill (the Unstable Localized Instantaneous Normal Modes).
  • Analogy: Imagine a crowded hallway. The "viscosity" isn't caused by everyone walking smoothly together. It's caused by a few people tripping over their own feet in tight spots, creating a bottleneck that slows everyone else down.
• The "Stable" Notes Don't Matter (at high temps): When the liquid is hot, the stable, gentle vibrations don't really contribute to the thickness.
• The "Delocalized" Notes are for Diffusion: The notes where the whole crowd wiggles together are actually responsible for how fast particles can move through the liquid (diffusion), not how thick the liquid is.

4. The Temperature Switch (The Crossover)

The paper found a fascinating "switch" that happens as the liquid cools down:

• Hot Liquid (Above the "Mode-Coupling Temperature"): The liquid is thick because of those localized, wobbly atoms tripping over each other.
• Cold Liquid (Below the Switch): As it gets colder and closer to turning into glass, the physics changes. The "wobbly" atoms disappear or stop being the main problem. Instead, the stable vibrations start to take over and control the thickness.

It's like a traffic jam that starts because of a few bad drivers (hot liquid), but as the road freezes over, the jam is caused by the ice itself making the whole road slippery and slow (cold liquid).

5. Why This Matters

Before this paper, scientists had a formula to calculate viscosity, but it was like trying to predict the weather by looking at a blurry photo. This work provides a spectral decomposition.

Think of it like a music equalizer. Before, we knew the song was loud (high viscosity), but we didn't know which specific frequencies were making it loud. Now, the scientists have turned the knobs and shown that only the "localized unstable" frequencies are turning the volume up on the viscosity.

In a nutshell:
This paper proves that the "thickness" of a liquid is caused by tiny, chaotic clusters of atoms that are teetering on the edge of instability. By identifying these specific atomic "trips," the authors have built a bridge between the microscopic jiggling of atoms and the macroscopic feeling of a thick liquid, finally answering the question of what makes honey, honey.

Technical Summary

Technical Summary: Spectral Decomposition of Liquid Viscosity into Instantaneous Normal Modes

Problem Statement
Viscosity, the resistance of a liquid to flow, is a fundamental transport coefficient governed by atomic-scale friction. However, its microscopic origin remains poorly understood. While macroscopic definitions exist (e.g., Green-Kubo formalism relating stress autocorrelation to viscosity), these methods do not elucidate the specific atomic motions or elementary excitations responsible for momentum transport. A major barrier is the lack of a well-defined normal mode formalism for liquids, which possess no equilibrium configuration, unlike crystalline solids. Previous attempts to link microscopic dynamics to macroscopic transport, such as the connection between unstable delocalized Instantaneous Normal Modes (INMs) and self-diffusion, have left the role of other modes, particularly unstable localized modes, unclear. This work addresses the gap in understanding the microscopic carriers of liquid viscosity and seeks to decompose viscosity into contributions from individual atomic modes.

Methodology
The authors combine extensive Molecular Dynamics (MD) simulations with a theoretical framework based on nonaffine linear viscoelastic response.

1. Theoretical Framework: The study utilizes a derived expression for shear viscosity ($\eta$) based on the imaginary part of the complex shear modulus. The viscosity is expressed as an integral over the density of states (DOS), $g(\omega)$, weighted by an affine force field correlator, $\Gamma(\omega)$, and a memory kernel, $\tilde{\nu}(0)$:
   $$\eta = 3\rho \tilde{\nu}(0) \int \frac{g(\omega) \Gamma(\omega)}{m^2 \omega^4} d\omega$$
   Crucially, the authors define $g(\omega)$ as the INM density of states derived from the diagonalization of the Hessian matrix of instantaneous liquid configurations. The memory kernel is derived from first principles using the projection operator formalism, avoiding free fitting parameters.

2. Simulation Systems: Three distinct systems were analyzed to ensure universality:

   • A Cu$_{50}$Zr$_{50}$ metallic liquid.
   • A covalently bonded Fe$_{80}$P$_{20}$ metallic liquid.
   • A Kob-Andersen (KA) binary Lennard-Jones mixture.

3. Mode Classification: INMs were classified based on eigenvalue signs (stable vs. unstable) and spatial extent. The spatial extent was quantified using the participation ratio ($P_\omega$). A finite-size scaling analysis identified a "mobility edge" (a critical participation ratio threshold, e.g., $P_\omega \approx 0.36$ for CuZr) separating unstable localized INMs (ULINMs) from unstable delocalized INMs. Spectral statistics (nearest-neighbor spacing distributions) confirmed that localized modes follow Poisson statistics while delocalized modes follow Wigner statistics.

Key Results

• Spectral Decomposition Validity: The authors demonstrate that integrating over all INMs (both stable and unstable) in the theoretical formula yields viscosity values that deviate by several orders of magnitude from simulation data (Green-Kubo and non-equilibrium shear MD).
• Dominance of ULINMs at High Temperatures: Above the mode-coupling temperature ($T_{MC}$), the viscosity is accurately predicted by the theoretical formula when the integration is restricted solely to unstable localized INMs (ULINMs). This finding identifies ULINMs as the primary microscopic excitations governing viscous transport in the high-temperature regime.
• Dynamical Crossover at $T_{MC}$: A distinct dynamical crossover is observed near the mode-coupling temperature ($T_{MC}$).
  • For $T > T_{MC}$, viscosity is controlled by ULINMs.
  • For $T < T_{MC}$, the control shifts to stable modes. This transition aligns with the topological change in the potential energy landscape (PEL) from a regime dominated by saddle points (saddles-dominated) to one dominated by minima (minima-dominated).
• Quantitative Model: The study establishes a quantitative link between viscosity, configurational entropy ($S_{conf}$), and the fraction of ULINMs ($f^L_u$). The data reveals a two-stage linear dependence between $S_{conf}$ and $f^L_u$, and a two-stage Adam-Gibbs-like behavior for viscosity. This leads to a proposed model:
  $$\eta \propto A \exp\left[ \frac{B}{T(a f^L_u + b)} \right]$$
  This model successfully connects the macroscopic viscosity to the population of specific microscopic excitations in both Arrhenius and non-Arrhenius regimes.

Significance and Claims
The paper claims to provide the first spectral decomposition of liquid viscosity into individual instantaneous normal modes. Its primary contributions are:

1. Identification of Microscopic Carriers: It identifies unstable localized INMs as the fundamental blocks responsible for liquid viscosity at temperatures above $T_{MC}$, distinguishing them from unstable delocalized modes which govern self-diffusion.
2. Resolution of the "What INMs Are" Debate: The work clarifies the physical nature of unstable modes, proposing that ULINMs act as facilitators for viscous dynamics, likely located at the boundaries between mobile and immobile clusters in dynamically heterogeneous liquids.
3. Predictive Framework: By linking viscosity directly to the population of ULINMs and configurational entropy, the authors propose a path to predict liquid viscosity from elementary excitations without relying on phenomenological parameters.
4. Universality: The observed crossover between ULINM-dominated and stable-mode-dominated viscosity is shown to be universal across metallic liquids and model glass-formers, reinforcing the connection between transport properties and the topological features of the potential energy landscape.

The authors conclude that this decomposition offers a new microscopic theory of liquid viscosity, bridging the gap between atomic-scale vibrations and macroscopic fluid dynamics.