Memory formation in cyclically deformed amorphous solids and sphere assemblies

Monoj Adhikari, Srikanth SastryThe European Physical Journal E 2018 · DOI · arXiv

Abstract

This study investigates how disordered materials, such as glasses and colloidal suspensions, encode information about past physical deformations. Using molecular dynamics simulations of structural glasses and soft sphere mixtures, we demonstrate that memory is stored through complex "loop-reversible" transitions between stable states. This mechanism is more robust than the "point-reversibility" observed in simpler non-Brownian fluids. Our findings reveal that memory is encoded in the organization of potential particle paths rather than static structural arrangements.

Introduction

In materials science, "memory" refers to the ability of a system to retain information about its history of deformation. Disordered systems—ranging from crumpled paper to concentrated suspensions—exhibit such history-dependent behavior. A fundamental question in condensed matter physics is how these materials encode information about their past physical environment. Specifically, when a glass or suspension is cyclically sheared at a certain amplitude, it eventually reaches a "steady state." If subsequently tested with different amplitudes, the material's response changes sharply at the original "training" amplitude.

Methodology

The researchers employed molecular dynamics simulations to study two distinct model systems under Athermal Quasi-Static (AQS) deformation: the Kob-Andersen Binary Mixture (BMLJ) as a model for structural glasses, and a Soft Sphere Binary Mixture (SS) for non-Brownian colloidal suspensions. To retrieve memories, they used two protocols: "Parallel Read," where multiple copies of a trained system are tested at different amplitudes, and "Sequential Read," which mimics experimental conditions by subjecting a single system to increasing amplitudes.

Mechanisms of Memory: Loop vs. Point Reversibility

A central finding of this research is the distinction between "point reversibility" and "loop reversibility." In simple suspensions, memory is often point reversible: particles eventually find positions where they no longer collide. In contrast, glasses exhibit loop reversibility. In this regime, particles continue to collide and the system jumps between different energy minima during a cycle, but it follows a closed path that returns exactly to the starting configuration.

Particle trajectories (Y coordinate) during cyclic shear. At the training amplitude, the particle follows a closed loop, returning to its initial position, illustrating loop reversibility.

Key Results

The MSD Signature

Memory is primarily revealed through the Mean Squared Displacement (MSD). In a trained glass, the MSD—the average distance particles move from their starting positions after a full cycle—drops to zero precisely at the training amplitude ($\gamma_{train}$). This signature is robust across different training amplitudes and is particularly sharp in parallel reading protocols.

Mean Squared Displacement (MSD) as a function of read amplitude for different training cycles and amplitudes. The MSD reaches zero at the training amplitude, marking the memory signature.

Structural Stealth

Surprisingly, despite the clear dynamic signature in the MSD, the authors found no significant change in the pair correlation function $g(r)$. This suggests that memory is "structurally stealthy," encoded in the organization of the particles' potential paths rather than a simple change in their static spatial arrangement.

Memory Erasure and Multiplicity

Subjecting the material to a deformation larger than the training amplitude acts as a "factory reset," erasing the stored memory. However, smaller deformations leave the memory intact. Furthermore, it is possible to encode multiple memories within a glass, with the largest amplitude being the most robust.

Conclusion

This study clarifies the microscopic mechanisms of memory in amorphous solids, highlighting the role of loop reversibility in complex energy landscapes. While the lack of a clear structural signature remains puzzling, the identification of loop-reversible paths provides a roadmap for future experimental validation in real-world disordered materials.

Summary updated

BibTeX
@article{adhikari2018memory,
  author = {Monoj Adhikari and Srikanth Sastry},
  title = {{Memory formation in cyclically deformed amorphous solids and sphere assemblies}},
  journal = {The European Physical Journal E},
  year = {2018},
  doi = {10.1140/epje/i2018-11717-5},
}

Loading related content...

Figure