Mean-field dynamics of infinite-dimensional particle systems: global shear versus random local forcing

Elisabeth AgoritsasJournal of Statistical Mechanics Theory and Experiment 2021 · DOI · arXiv

Abstract

In infinite dimensions, many-body systems of pairwise interacting particles provide exact analytical benchmarks for the features of amorphous materials, such as the stress–strain curve of glasses under quasistatic shear. Here, instead of global shear, we consider an alternative driving protocol, as recently introduced by Morse et al 2020, which consists of randomly assigning a constant local displacement on each particle, with a finite spatial correlation length. We show that, in the infinite-dimensional limit, the mean-field dynamics under such a random forcing are strictly equivalent to those under global shear, upon a simple rescaling of the accumulated strain. Moreover, the scaling factor is essentially given by the variance of the relative local displacements of interacting pairs of particles, which encodes the presence of a finite spatial correlation. In this framework, global shear is simply a special case of a much broader family of local forcing, which can be explored by tuning its spatial correlations. We discuss the specific implications for the quasistatic driving of glasses—initially prepared at a replica-symmetric equilibrium—and how the corresponding ‘stress–strain’-like curves and elastic moduli can be rescaled onto their quasistatic-shear counterparts. These results hint at a unifying framework for establishing rigorous analogies, at the mean-field level, between different driven disordered systems.

Introduction

Understanding the mechanical response of amorphous materials—such as glasses, foams, and granular piles—remains a central challenge in condensed matter physics. Unlike crystalline solids, these materials possess a complex potential energy landscape due to their disordered atomic or molecular structures. To probe this landscape, physicists typically employ two primary driving protocols: Global Shear (Athermal Quasistatic Shear, or AQS), which involves a uniform deformation of the entire sample, and Random Local Forcing (Athermal Quasistatic Random Displacement, or AQRD), where particles are subjected to small, random individual nudges.

While empirical evidence has suggested a relationship between these two methods, a rigorous theoretical bridge has been missing. This research establishes that in the infinite-dimensional limit, these two protocols are mathematically equivalent. By applying a specific rescaling factor based on spatial correlations, the macroscopic stress-strain behavior of a system under random local forcing becomes identical to one under global shear.

Theoretical Framework: The Infinite-Dimensional Limit

The study utilizes the infinite-dimensional limit ($d \to \infty$) as a mathematical laboratory. In high dimensions, the complexity of many-body interactions simplifies significantly because each particle interacts with an effectively infinite number of neighbors. This allows the system to be described via Dynamical Mean-Field Equations (DMFE), reducing the million-body problem to an effective "one-body" problem: a single particle moving in a self-consistent "bath" or "field" created by all other particles.

Using Langevin dynamics to model particle motion, the research defines local displacements through quenched random fields. These fields are characterized by a Gaussian distribution with a tunable spatial correlation length ($\xi$). The goal is to determine how the "accumulated strain" in a sheared system relates to the "random displacement amplitude" in a locally forced system.

The Equivalence Principle and Rescaling

The core insight of this work is the Equivalence Principle. The author demonstrates that global shear is not a distinct phenomenon but rather a special case of random local forcing. Specifically, global shear represents a scenario where the variance of relative displacements between interacting particles is exactly equal to the square of the interaction length.

To bridge the two protocols, the paper identifies a rescaling factor, $\mathcal{F}$, which represents the variance of relative local displacements. By rescaling the strain ($\gamma$) by $\sqrt{\mathcal{F}}$, the stress-strain curves of various protocols collapse onto a single master curve. This equivalence is not restricted to zero-temperature (athermal) systems; it holds for full dynamics at finite temperatures, suggesting a universal behavior across driven disordered systems.

Key Results

1. Stress-Strain Collapse

The research provides an explicit mapping formula to translate AQRD data into AQS equivalents: $$\sigma_{AQS}(\gamma_{eff}) = \frac{\ell}{\sqrt{\mathcal{F}}} \sigma_{AQRD}(\gamma)$$ This formula allows researchers to predict how a material will respond to uniform shearing based solely on its response to localized random perturbations.

2. Elastic Moduli and Spatial Correlation

The theory predicts that the ratio of elastic moduli (the initial stiffness of the material) between the two protocols is exactly $\mathcal{F}/\ell^2$. Furthermore, for systems with Gaussian correlations, the elastic modulus exhibits a power-law crossover based on the correlation length $\xi$: it scales as $1/\xi$ for short-range correlations and $1/\xi^3$ for long-range correlations.

3. Validation

While derived for infinite dimensions, the theoretical predictions were compared against 2D numerical simulations of particles with Hertzian contact potentials. The results show that the "infinite-D" theory provides a remarkably accurate description of low-dimensional systems, confirming the robustness of the mean-field approach.

Limitations and Future Directions

While the equivalence is robust in the pre-yielding regime, the current derivation assumes "Replica Symmetry" (RS). This describes the material's state before it reaches the yielding transition—the point where it begins to flow or break. Extending this theory to describe the actual point of failure will require incorporating "Replica Symmetry Breaking" (RSB) states.

Additionally, while the theory performs well in 2D, minor quantitative discrepancies appear when the correlation length $\xi$ is very small, suggesting that finite-dimensional fluctuations play a larger role in those specific regimes. Finally, while this work focuses on quasistatic (infinitely slow) driving, the mathematical framework suggests that the equivalence should hold for finite-rate shearing, offering new avenues for research in fluid rheology and the study of non-equilibrium steady states.

Summary updated

Abstract

Abstract In infinite dimensions, many-body systems of pairwise interacting particles provide exact analytical benchmarks for the features of amorphous materials, such as the stress–strain curve of glasses under quasistatic shear. Here, instead of global shear, we consider an alternative driving protocol, as recently introduced by Morse et al 2020 (arXiv: 2009.07706 ), which consists of randomly assigning a constant local displacement on each particle, with a finite spatial correlation length. We show that, in the infinite-dimensional limit, the mean-field dynamics under such a random forcing are strictly equivalent to those under global shear, upon a simple rescaling of the accumulated strain. Moreover, the scaling factor is essentially given by the variance of the relative local displacements of interacting pairs of particles, which encodes the presence of a finite spatial correlation. In this framework, global shear is simply a special case of a much broader family of local forcing, which can be explored by tuning its spatial correlations. We discuss the specific implications for the quasistatic driving of glasses—initially prepared at a replica-symmetric equilibrium—and how the corresponding ‘stress–strain’-like curves and elastic moduli can be rescaled onto their quasistatic-shear counterparts. These results hint at a unifying framework for establishing rigorous analogies, at the mean-field level, between different driven disordered systems.
BibTeX
@article{agoritsas2021mean,
  author = {Elisabeth Agoritsas},
  title = {{Mean-field dynamics of infinite-dimensional particle systems: global shear versus random local forcing}},
  journal = {Journal of Statistical Mechanics Theory and Experiment},
  year = {2021},
  doi = {10.1088/1742-5468/abdd18},
}
Figure