Rigidity Percolation in Disordered 3D Rod Systems

Samuel Heroy, Dane Taylor, Feng Shi, M. Gregory Forest, Peter J. MuchaMultiscale Model. Simul 2022 · OpenReview

TL;DR

This paper introduces a novel algorithm, Rigid Graph Compression (RGC), to precisely determine the "rigidity percolation threshold" in 3D systems of randomly oriented rods. This threshold marks the point at which these rods form a rigid, spanning network, which is crucial for understanding the mechanical strength transitions in composite materials. The findings reveal that this rigidity transition, much like electrical conductivity percolation, scales inversely with the excluded volume of the rods.

What Problem Does This Paper Address?

Many composite materials, like polymers reinforced with stiff fibers or particles, exhibit a sudden increase in mechanical strength (rheological percolation) when the concentration of reinforcing particles reaches a critical point. This is analogous to electrical conductivity percolation, where a connected network of conductive particles allows current to flow. While conductivity percolation is well understood and linked to particle contact, the underlying mechanism for rheological percolation has been less clear. A leading hypothesis is that the material becomes mechanically stable when the rigid reinforcing elements form a spanning, rigid structure – a phenomenon termed "rigidity percolation." This paper aims to provide a theoretical and computational framework to identify and quantify this rigidity percolation threshold in a realistic 3D system of randomly dispersed rods.

Methods

The core of this work is the extension of the Rigid Graph Compression (RGC) algorithm from 2D to 3D systems. RGC operates by identifying and "compressing" small, locally rigid substructures (rigid motifs) within a network of interacting components. This process is repeated iteratively, progressively identifying larger and larger rigid components until the entire system is analyzed. The algorithm relies on rigidity matroid theory, which uses graph embeddings and linear algebra to determine the rigidity of a structure. The authors developed specific "rigid motifs" for 3D rod systems and implemented a Monte Carlo simulation protocol to generate random rod configurations. Finite-size scaling analysis was then used to estimate the rigidity percolation threshold and the correlation length exponent.

Datasets

Not discussed in the paper. The study relies on simulated data generated through Monte Carlo simulations of randomly placed and oriented rods.

Key Insights and Ideas

  • Rigidity Percolation as the Driver of Mechanical Strength: The paper strongly supports the hypothesis that rigidity percolation, not just simple contact percolation, is the key to understanding the mechanical strength transitions in rod-based composites.
  • 3D RGC Algorithm: The successful extension of the RGC algorithm to 3D is a significant methodological contribution, enabling the study of rigidity in more complex, real-world systems.
  • Inverse Scaling with Excluded Volume: Similar to contact percolation, the rigidity percolation threshold scales inversely with the excluded volume of the rods. This suggests a fundamental geometric constraint governs both phenomena.
  • Lower Critical Contact Number than Maxwell's Prediction: The critical number of contacts per rod at the rigidity percolation threshold is found to be lower than predicted by Maxwell's isostatic condition, indicating that the system can become rigid with fewer constraints than previously thought.
  • Aspect Ratio Dependence: While the inverse scaling with excluded volume holds, the correlation length exponent shows some dependence on rod aspect ratio, suggesting subtle differences in how rod shape influences the percolation process.

Key Results

  • The authors successfully implemented a 3D RGC algorithm and used it to estimate the rigidity percolation threshold ($p_{min,r}$) for systems of monodisperse spherocylindrical rods with varying aspect ratios ($\gamma \in \{23, 48, 98\}$).
  • They found that $p_{min,r}$ scales inversely with the excluded volume of the rods, $V_{ex}(\gamma)$, with a relationship $p_{min,r}(\gamma) = b/V_{ex}(\gamma)$, where $b \approx 3.17$.
  • The critical contact number at the rigidity percolation threshold was found to be approximately constant for aspect ratios above a certain low value, and consistently lower than Maxwell's prediction of $10/3$.
  • The correlation length exponent ($\nu(\gamma)$) was estimated to be around 0.7-0.8 and showed a slight, though not statistically significant, dependence on aspect ratio.
  • The ratio of the rigidity percolation threshold to the contact percolation threshold ($p_{min,r}/p_{min,c}$) was found to increase with aspect ratio and approach a constant value in the slender body limit.

Limitations and Future Work

  • Sufficiency, Not Necessity: The RGC algorithm is "sufficient but not necessary" for identifying rigidity. This means it can definitively identify rigid components, but it might miss some rigid components. The authors address this by using multiple implementations and exploring a larger set of candidate rigid motifs to improve accuracy.
  • Computational Cost: Extending RGC to 3D and performing extensive simulations is computationally intensive, limiting the range of system sizes and the number of implementations explored.
  • Idealized Rods: The model assumes "soft-core" rods that intersect at contact points, which is an idealization. Future work could explore "hard-core" (impenetrable) rods.
  • Random and Isotropic Systems: The current study focuses on randomly placed and oriented rods. Future research could investigate the effects of spatial heterogeneity (clustering) and particle alignment, which are common in real composites.
  • Order of Motif Compression: While the order of motif compression was found to have minimal impact on the overall threshold estimation, it can lead to different identifications of individual rigid components, suggesting a need for careful consideration in complex systems.
  • Universality Class: The authors suggest that future work could explore whether their RGC framework can help establish a rigorous connection between rigidity matroid theory and universality classes for different types of network transitions.

Summary updated

Abstract

In composite materials composed of a soft polymer matrix and rigid, high aspect-ratio particles, the composite undergoes a transition in mechanical strength when the incorporated particle phase surpasses a critical density. This phenomenon, termed rheological percolation, is well known to occur in many polymer-rod and polymer-platelet composites at a critical density that exceeds the conductivity percolation threshold (which occurs when the conducting particles form a large connected component that spans the composite). Contact percolation in rod-like composites has been routinely exploited to engineer thermal or electrical conductivity in otherwise nonconducting polymers, and the characterization of contact percolation is well established. Mechanical or rheological percolation, however, has evaded a complete theoretical explanation and predictive description. A natural hypothesis is that rheological percolation is due to a rigidity phenomenon, whereby a large rigid component of inclusions spans the composite. Here we build an algorithm to detect the rigidity percolation threshold in rod-polymer composites. We model the composites as systems of randomly distributed, soft-core (intersecting at contact) rods and study the emergence of a giant (i.e., spanning) rigid component. Building on our previous results for two-dimensional composites, we develop an approximate algorithm that identifies spanning rigid components through hierarchically identifying and compressing provably rigid motifs---equivalently, decomposing a giant rigid component into rigid assemblies of a hierarchy of successively smaller rigid components. We apply this algorithm to random monodisperse systems that are generated in Monte Carlo simulations to estimate a rigidity percolation threshold (critical density) and explore its dependence on rod aspect ratio. We show that this transition point---like its contact percolation analogue---scales inversely with the excluded volume of a rod. Moreover, this implies that the critical contact number (i.e., the number of contacts per rod at the rigidity percolation threshold) is constant for aspect ratios above some relatively low value and is lower than the prediction from Maxwell's isostatic condition.
BibTeX
@inproceedings{heroy2022rigidity,
  author = {Samuel Heroy and Dane Taylor and Feng Shi and M. Gregory Forest and Peter J. Mucha},
  title = {{Rigidity Percolation in Disordered 3D Rod Systems}},
  booktitle = {Multiscale Model. Simul},
  year = {2022},
  url = {https://openreview.net/forum?id=W7mrVmAPOQY},
}
Figure