🔬 condensed matter

Rheology of dense suspensions of granular spherocylinders by particle-based simulation

This paper presents a particle-based simulation model that successfully predicts the viscosity and microstructure of dense suspensions of granular spherocylinders under shear flow, revealing how these properties evolve with volume fraction and aspect ratio while corroborating limited experimental data.

Original authors: Alex Dixon, John Hone, Gavin Melaugh, Christopher Ness

Published 2026-07-31

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Alex Dixon, John Hone, Gavin Melaugh, Christopher Ness

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a world where the stuff you pour, pump, or mix isn't just water or sand, but a chaotic soup of tiny, rigid sticks floating in liquid. This is the realm of rheology, the science of how things flow. You've likely seen this in action: think of a river clogged with fallen trees forming a massive "log jam," or the thick, sticky slurry used to make paper. In these situations, the particles are so crowded that they constantly bump into each other, and their shape matters just as much as how many of them there are. If you try to stir a bucket of marbles, they roll past each other easily. But if you try to stir a bucket of long, thin rods, they get tangled, lock together, and suddenly the whole mixture acts like a solid. Scientists have long understood how round balls behave in these crowds, but the rules for long, rod-shaped particles have been a bit of a mystery, especially when they are packed tight. Understanding this is crucial because these "sticky stick" mixtures appear everywhere, from industrial manufacturing to the flow of molten rock inside volcanoes.

In this paper, a team of researchers built a super-smart computer simulation to act as a virtual laboratory for these crowded rod suspensions. Instead of mixing real chemicals in a lab, they created a digital world filled with thousands of "spherocylinders"—particles that look like capsules or pills (a cylinder with a hemisphere on each end). They set these capsules floating in a virtual fluid and then started "shearing" them, which is a fancy way of saying they dragged the top layer of the fluid sideways while holding the bottom still, just like spreading butter on toast. The goal was to see how the mixture's thickness (viscosity) changed as they packed more rods into the space and as they made the rods longer.

The simulation revealed some fascinating behaviors. First, when they started the shear, the mixture got incredibly thick for a split second—a "viscosity spike"—before settling down. This happened because the rods were initially jumbled in random directions, bumping into each other like a mosh pit. Once the flow started, the rods began to line up, like soldiers marching in formation, and the mixture became easier to stir. The researchers found that the more rods they packed in (the volume fraction, $\phi$ ), the thicker the mixture got, eventually reaching a point where it jammed and stopped flowing entirely. Interestingly, the longer the rods were (the aspect ratio, $A$ , ranging from 1 for a sphere up to 20 for very long sticks), the sooner this jamming happened. For example, with rods of length 5, the jamming point was around a volume fraction of 0.57, but for rods of length 20, the mixture jammed at a much lower density of about 0.41.

The team also discovered that the rods didn't just line up; they lined up specifically with the flow. They measured an "order parameter" ( $S$ ) to see how well they were aligned. At low densities, the rods wobbled and rotated in loops (known as Jeffery orbits), but as the crowd got denser, they locked into a straight line along the direction of the flow. However, if the crowd got too dense, the rods actually started to get frustrated and couldn't stay perfectly aligned anymore, causing the order to drop slightly. The researchers also broke down the forces at play, showing that when the mixture was thin, the liquid's own resistance mattered most, but as it got crowded, the physical bumps and friction between the rods became the dominant force, taking over the job of making the mixture thick.

By creating a model that balances the physics of contact, friction, and fluid drag, the authors provided a new tool to predict how these complex mixtures behave. They didn't just guess; they ran thousands of simulations with different numbers of particles, different rod lengths, and different packing densities. Their results suggest that while the basic idea of "more stuff equals thicker fluid" holds true, the shape of the particles dramatically changes the point at which the mixture locks up. This work doesn't claim to solve every mystery of fluid mechanics, but it offers a solid, simulated foundation for understanding how to manage these tricky, rod-filled fluids in the real world, from preventing log jams in rivers to designing better industrial slurries.

Technical Summary: Rheology of Dense Suspensions of Granular Spherocylinders by Particle-Based Simulation

Problem Statement
Dense suspensions of rod-shaped granular particles are prevalent in natural and industrial contexts, ranging from log-jams in rivers to crystalline suspensions in magma and fiber networks in paper manufacturing. While the rheology of spherical suspensions is well-characterized, particularly through the unification of non-Brownian suspensions and dry granular flows by Boyer et al., the behavior of elongated particles remains incompletely understood. The additional degree of freedom associated with particle orientation introduces complex microstructural phenomena, such as Jeffery orbits and shear-induced alignment, which significantly alter the effective viscosity and stress response. Existing experimental data for dense regimes (high solid volume fractions, $\phi$ ) is limited due to preparation and measurement difficulties. Furthermore, previous simulation approaches face trade-offs: Stokesian Dynamics (SD) is computationally expensive for dense systems due to matrix inversions required for many-body hydrodynamics, while Discrete Element Method (DEM) models often oversimplify particle geometry (e.g., chains of spheres) or neglect necessary hydrodynamic forces. There is a need for a computationally tractable particle-based model that faithfully captures the dominant microscopic physics of dense, non-Brownian spherocylindrical suspensions under shear.

Methodology
The authors developed a particle-based simulation model for dense suspensions of non-bendy, non-Brownian spherocylinders (cylinders with hemispherical caps) under simple shear flow. The model operates in a regime where particle inertia is present but negligible.

Equations of Motion: Particle trajectories are updated using the Velocity-Verlet algorithm, integrating translational and rotational equations of motion. These equations account for pairwise contact forces, pairwise lubrication forces, and single-body hydrodynamic forces (drag, lift, and torque).
Hydrodynamics: Single-body forces are modeled using analytical solutions for prolate spheroids as an approximation for spherocylinders. This includes orientation-dependent drag and shear-induced lift. The hydrodynamic torque reproduces Jeffery orbits.
Contact and Lubrication:
- Contact: Modeled as Hookean with Coulombic friction, including normal and tangential spring-damper components.
- Lubrication: A novel scheme was developed to compute lubrication forces between spherocylinders. To address the discontinuity in curvature at the end-shaft junction, the model calculates three interaction components (shaft-shaft, end-shaft, and end-end) simultaneously. These are weighted based on the relative position of the particles to ensure smooth transitions between interaction types, resolving discontinuities present in previous models (e.g., Yamane et al., Butler and Shaqfeh).
Dynamic Timestep: A key methodological innovation is a dynamic timestep calculation. The timestep is determined by the smallest characteristic timescale among contact interactions, lubrication forces, and lubrication torques. This ensures numerical stability while avoiding unnecessarily small timesteps, improving computational efficiency.
Simulation Setup: The system uses periodic boundaries in the flow and gradient directions and Lees-Edwards boundaries for shear. A bidisperse mixture of radii (1:1.4) is used to prevent crystallization. Simulations cover aspect ratios ( $A$ ) from 1 to 20 and volume fractions ( $\phi$ ) from dilute to near-jamming.

Key Results

Viscosity Behavior: The model predicts a systematic increase in steady-state viscosity ( $\eta_s$ $η_{s}$ ) with both volume fraction ( $\phi$ $ϕ$ ) and aspect ratio ( $A$ $A$ ).
- Transient Response: Upon shear start-up, a viscosity spike is observed, attributed to the initial random distribution of particle orientations. This spike subsides as particles align, leading to a steady state. The magnitude of the spike relative to the steady state increases with $\phi$ .
- Steady State: The steady-state viscosity follows a power-law divergence $\eta_s \sim (1 - \phi/\phi_m)^{-\varsigma}$ . The jamming volume fraction $\phi_m$ decreases as $A$ increases (from $\approx 0.57$ for $A=5$ to lower values for higher $A$ ). The model confirms rate-independence of the steady-state viscosity, consistent with the framework for granular suspensions.
- Force Contributions: At low $\phi$ , viscosity is dominated by fluid stress and lubrication forces. As $\phi$ increases, contact forces (both normal and tangential) become dominant, eventually exceeding lubrication contributions around $\phi \approx 0.5$ .
Microstructure and Alignment:
- Order Parameter: The orientational order parameter ( $S$ ) increases with $\phi$ up to a critical point, after which it decreases slightly due to geometric frustration preventing perfect alignment.
- Alignment Direction: Particles align preferentially along the flow direction ( $x$ -axis), with a small inclination angle ( $\phi_v$ ) relative to the flow-gradient plane. This inclination angle increases with $\phi$ up to a threshold and then decreases.
- Aspect Ratio Effect: Higher aspect ratios lead to stronger alignment at lower volume fractions but also result in a lower jamming point. For very high $A$ , the inclination angle becomes negligible.
- Transient vs. Steady State: The initial viscosity spike correlates with the time required for the system to transition from a random state to a sheared-aligned state.

Significance and Claims
The paper claims to provide a fundamental tool for exploring the fluid mechanics, microstructure, and rheology of dense granular rod suspensions. By successfully modeling the transition from dilute to dense regimes and capturing the interplay between hydrodynamic forces, frictional contacts, and particle geometry, the model corroborates limited experimental rheology data available for granular rods.

The authors state that their work extends the understanding of granular suspension rheology (previously established for spheres) to the broad class of rod-shaped particles. The model offers a platform for future studies to map the dependence of the jamming point on microstructural details, investigate normal stresses, and explore behavior under inhomogeneous conditions. The paper modestly frames its contribution as a step toward a full fluid mechanical description of these materials, deferring comprehensive quantification of the jamming point's relation to alignment and friction to future work. It does not claim to solve specific industrial processing problems directly but rather provides the simulation infrastructure to investigate the underlying physics of such systems.

More like this