Nonlinear phonon dispersion in disordered solids and non-Debye vibrational spectra

This study demonstrates that nonlinear phonon dispersion in disordered solids arises from a disorder-induced mesoscopic lengthscale and, through analysis and simulations, reveals that both this nonlinear softening and non-phononic vibrations significantly contribute to non-Debye anomalies like the boson peak, with their relative importance depending on the material's disorder strength.

The Gist

Imagine a solid object, like a piece of glass or a block of metal, as a giant, invisible orchestra. When you tap it, it doesn't just sit there; it vibrates. These vibrations travel through the material as waves, much like sound waves travel through air. In physics, we call these vibrations "phonons."

For over a century, scientists have used a classic rulebook called Debye's Model to predict how these vibrations behave. The rulebook says: "If you look at low-frequency (slow) vibrations, they travel in a perfectly straight line, and the number of vibrations increases in a predictable, smooth way."

However, when scientists look at disordered solids (like window glass, which has no crystal structure, unlike a diamond), the music gets messy. The vibrations don't follow the straight line; they curve, and there are way more vibrations than the old rulebook predicted. This extra "noise" creates a famous mystery in physics called the Boson Peak.

For a long time, scientists argued about what causes this mess. Is it because the waves themselves are bending (nonlinear dispersion)? Or is it because the disorder creates entirely new, weird types of vibrations that don't exist in perfect crystals (non-phononic modes)?

This paper acts like a detective story that finally solves the case by measuring the waves directly and separating the two suspects.

The Detective Work: A New Way to Listen

The biggest problem was that in a messy, disordered solid, it's hard to tell exactly how fast a wave is traveling at a specific pitch. It's like trying to hear a single violin in a crowded, chaotic room.

The authors developed a clever new technique called the "Imposed-Wave Method."

• The Analogy: Imagine you are in a crowded room. Instead of waiting for someone to start singing, you gently push every person in the room in a specific wave pattern (like a "wave" in a stadium). You then measure how the room reacts.
• By doing this mathematically on a computer, they could force the material to vibrate in a specific pattern and measure exactly how the speed of that wave changed as the pitch got higher. This allowed them to map out the "curved" path of the waves with high precision.

The Discovery: The "Hidden Ruler"

They found that in disordered solids, the waves don't start bending just because they hit the size of an atom (as they do in perfect crystals). Instead, they start bending because of a mesoscopic lengthscale.

• The Analogy: Think of a perfect crystal as a grid of perfectly spaced tiles. If you walk across it, you trip when you hit the edge of a tile.
• In a disordered solid (like glass), there are no tiles. However, the authors found a "hidden ruler" (let's call it $\xi$) that is much larger than a single atom. This ruler represents the scale at which the material's stiffness starts to fluctuate randomly.
• The Finding: The waves travel smoothly until they get big enough to "feel" this hidden ruler. Once they hit this size, they start to slow down and bend (soften). This hidden ruler also controls how much the waves get scattered and lost (attenuation). The more disordered the glass is, the bigger this hidden ruler becomes.

Solving the Boson Peak Mystery

Once they knew exactly how the waves bend, they could calculate how many vibrations should exist based on that bending alone. Then, they compared this calculation to the actual total number of vibrations observed.

They found that the "Boson Peak" (the extra vibrations) is actually a duet between two different sources:

1. The Bending Waves (Phononic): The waves themselves are bending due to the hidden ruler, creating extra vibrations.
2. The Weird Localized Jiggles (Non-Phononic): Because the material is messy, some parts of it get stuck in a "jiggle" that doesn't travel like a wave at all. These are localized, trapped vibrations.

The Verdict:

• In very disordered glasses (like those made by cooling very quickly), the "weird localized jiggles" are the main culprit for the extra vibrations.
• In stable, realistic laboratory glasses (the kind we actually use), the "bending waves" and the "weird jiggles" contribute almost equally.

Why This Matters

The paper concludes that for a long time, scientists might have been blaming the wrong thing. Some thought the extra vibrations were only from the weird localized jiggles. Others thought they were only from the bending waves.

This study shows that both are essential players. The amount each contributes depends on how "messy" the glass is. In the glasses we make in real labs, you can't ignore either one; they both significantly shape the vibrational spectrum.

In short: The paper didn't just find the Boson Peak; it built a new map of the terrain, showing that the "messiness" of glass creates a hidden scale that bends waves, and that this bending works hand-in-hand with trapped vibrations to create the unique sound of disordered solids.

Technical Summary

Technical Summary: Nonlinear Phonon Dispersion in Disordered Solids and Non-Debye Vibrational Spectra

Problem Statement
While crystalline solids are well-described by Debye's classical model—where low-frequency acoustic phonons follow a linear dispersion relation $\omega(k) \propto k$ and a vibrational density of states (VDoS) scaling as $D(\omega) \propto \omega^2$—disordered solids (glasses) exhibit significant deviations known as non-Debye anomalies. The most prominent of these is the "boson peak" (BP), an excess of vibrational modes over the Debye prediction. The origin of these anomalies remains a subject of debate. Two primary mechanisms are proposed: (1) the softening nonlinearity of the phonon dispersion relation $\omega(k)$ due to structural disorder, and (2) the existence of non-phononic, quasi-localized vibrational modes intrinsic to disordered structures. A critical gap in current understanding is the inability to quantitatively separate and determine the relative contributions of these two mechanisms to the onset of non-Debye behavior and the boson peak. Furthermore, the nature of the disorder-induced lengthscale governing phonon dispersion and wave attenuation in non-crystalline solids is not fully understood.

Methodology
To address these challenges, the authors employed large-scale computer simulations of three distinct classes of 3D disordered solids:

1. Disordered Elastic Networks: Composed of relaxed Hookean springs with varying average coordination numbers ($z$), ranging from the critical Maxwell threshold ($z_c=6$) to well above it.
2. Polydisperse Inverse-Power-Law (PIPL) Glasses: Repulsive systems subjected to varying thermal histories, quantified by the parent temperature $T_p$ at which the system falls out of equilibrium.
3. Binary Glasses: Including both repulsive binary mixtures and "sticky-sphere" glasses with tunable attractive forces controlled by a parameter $r_c$.

The study utilized a novel computational technique termed the "imposed-wave method" to compute the phonon dispersion relation $\omega(k)$ in disordered systems. Unlike crystals, where phonons are exact eigenfunctions of the Hessian, disorder prevents a unique identification of $\omega$ for a given $k$. The imposed-wave method assigns a unique frequency to each admissible wavevector by imposing a wave-like force field on the particles and calculating the linear response displacement. This allows for the extraction of $\omega(k)$ and the associated group velocity.

The authors also employed the Kernel Polynomial Method (KPM) to compute the VDoS $D(\omega)$ for very large system sizes ($N \sim 10^6$), enabling the resolution of low-frequency tails and the separation of phononic and non-phononic contributions. A dimensionless disorder quantifier, $\chi$, defined as the ratio of shear modulus fluctuations to its mean over mesoscopic scales, was calculated for each system to unify the description of disorder across different models.

Key Contributions and Results

1. Discovery of a Mesoscopic Disorder-Induced Lengthscale ($\xi$):
   The authors demonstrated that the nonlinear softening of the phonon dispersion relation in disordered solids is not governed by the microscopic interatomic distance $a_0$ (as in crystals), but by a much larger, emergent mesoscopic lengthscale $\xi \gg a_0$. The dispersion relation is characterized by the form $\omega(k) = f(k\xi)ck$, where the leading-order nonlinearity scales as $\omega(k) \approx ck - c\Upsilon \xi^2 k^3$ (with $\Upsilon = 1/96$).

   • Scaling Relations: $\xi$ was found to scale with the degree of disorder: $\xi \sim 1/\sqrt{z-z_c}$ for elastic networks near the unjamming transition, and $\xi$ increases with $T_p$ and $r_c$ for glasses.
   • Unified Disorder Measure: Crucially, the authors showed that $\xi/a_0$ scales linearly with the dimensionless disorder quantifier $\chi$ ($\xi/a_0 \sim \chi$), unifying the effects of diverse disorder control parameters.

2. Connection to Wave Attenuation:
   The same lengthscale $\xi$ was shown to control the Rayleigh scattering of wave attenuation rates $\Gamma(k)$. The authors derived a relation $\Gamma(k) \sim \xi^2 k^4$ for small $k$, consistent with the Heterogeneous Elasticity Theory (HET) and the Rayleigh-Klemens relation. This establishes a direct link between the nonlinear dispersion and wave attenuation, both governed by the same mesoscopic disorder-induced lengthscale.

3. Quantitative Decomposition of Non-Debye Spectra:
   The paper quantitatively separates the contributions of nonlinear phonon dispersion (phononic) and non-phononic vibrations to the VDoS.

   • Phononic Contribution: Using the measured $\omega(k)$, the authors calculated the phononic VDoS $D_{ph}(\omega)$, which deviates from Debye's law as $D_{ph}(\omega) \approx A_D \omega^2 + A_w \omega^4 + \dots$. The coefficient $A_w$ depends on $\xi$.
   • Non-Phononic Contribution: For glasses, a gapless distribution of non-phononic modes exists with a universal tail $D_G(\omega) \approx A_g \omega^4$.
   • Relative Importance: The ratio $A_g/A_w$ determines which mechanism dominates the onset of non-Debye behavior. The results show that this ratio is highly sensitive to the degree of disorder. In highly disordered (unstable) glasses, non-phononic modes dominate ($A_g/A_w > 1$). However, in stable, realistic laboratory glasses (low disorder), the phononic contribution from nonlinear dispersion often dominates ($A_g/A_w < 1$). This challenges previous interpretations attributing the $\omega^4$ tail solely to non-phononic modes.

4. The Boson Peak Composition:
   By analyzing the cumulative VDoS at the boson peak frequency $\omega_{BP}$, the authors defined a ratio $R(\omega_{BP})$ representing the fraction of non-phononic vibrations in the total excess modes.

   • Result: $R(\omega_{BP})$ was found to be a sizable fraction of unity (ranging from $\sim 0.2$ to $\sim 0.5$) across a broad range of glassy disorder.
   • Conclusion: The boson peak is not populated exclusively by softened phonons (as suggested by some recent theories) nor exclusively by non-phononic modes. Instead, it arises from a significant hybridization of both nonlinear phonon softening and non-phononic quasi-localized excitations.

Significance
The paper claims to provide a "basic progress in understanding disordered solids" by establishing a comprehensive, quantitative framework that links wave dispersion, wave attenuation, and vibrational spectra through a single mesoscopic lengthscale $\xi$. By resolving the relative contributions of phononic and non-phononic mechanisms, the work clarifies the physical origins of the boson peak and non-Debye anomalies. The findings suggest that the composition of the boson peak is not universal but depends strongly on the specific disorder strength and thermal history of the glass. The developed analysis framework and the imposed-wave method offer testable predictions for experimental vibrational spectra, potentially resolving long-standing controversies regarding the nature of low-frequency excitations in amorphous materials.