🔬 condensed matter

STL-to-Stokeslet Computation of Mobility Tensors and Sedimentation Dynamics for Shaped Particles

This paper presents an efficient numerical framework that converts 3D STL files into mobility tensors using a Stokeslet-based discretization method, enabling the prediction of complex sedimentation dynamics and the design of steerable particles for applications in soft matter and biophysics.

Original authors: Wenting Cheng, Tiago Pernambuco, Thomas A. Witten, Haim Diamant, Justin C. Burton

Published 2026-03-03

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Wenting Cheng, Tiago Pernambuco, Thomas A. Witten, Haim Diamant, Justin C. Burton

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 you are dropping a tiny, oddly shaped toy into a thick jar of honey. You might expect it to just fall straight down. But in the world of tiny particles moving through thick fluids (like water or honey), things get weird. A toy shaped like a dolphin might spin, wobble, or even spiral as it falls. A toy shaped like a twisted ribbon might dance in a circle.

This paper is about building a digital crystal ball that can predict exactly how any weirdly shaped object will move as it sinks through a fluid, without needing to build the object in real life first.

Here is the breakdown of their discovery using simple analogies:

1. The Problem: The "Shape" Mystery

In the microscopic world, inertia (the tendency to keep moving) doesn't matter much. Instead, the fluid acts like thick syrup. The way a particle moves depends entirely on its shape and where its weight is balanced.

For simple shapes like perfect balls or flat disks, scientists have known the math for a long time. But for complex shapes—like a jagged ice crystal, a twisted DNA strand, or a dolphin-shaped toy—predicting the motion has been incredibly difficult. It's like trying to guess how a crumpled piece of paper will flutter down without actually dropping it. Previous methods were either too slow, too expensive, or only worked for simple shapes.

2. The Solution: The "Sticky Dot" Trick (STL-to-Stokeslet)

The authors created a new, fast, and easy-to-use computer program. Here is how it works, step-by-step:

The Input (The Blueprint): You start with a standard 3D printer file (called an STL file). Think of this as the digital blueprint for a 3D object. It's just a list of triangles that make up the surface of the object.
The Magic (The Stokeslets): The program takes that blueprint and covers the surface of the object with thousands of tiny, invisible "sticky dots." In physics, these are called Stokeslets.
- The Analogy: Imagine the object is a boat. Instead of the whole boat pushing against the water, imagine the boat is covered in thousands of tiny, independent oars. Each oar pushes against the water a little bit.
The Calculation: The computer calculates how all these thousands of tiny oars interact with each other and the water. By solving a giant math puzzle (a matrix), it figures out exactly how much force and spin the whole object experiences.
The Output: The result is a Mobility Tensor. Think of this as a "movement instruction manual" for that specific shape. It tells you: "If you push this object down, it will also spin to the left. If you push it sideways, it will tilt forward."

3. The Big Discovery: The "Weight Shift" Effect

The most surprising part of the paper is what happens when the center of gravity isn't perfectly in the middle of the shape.

The Analogy: Imagine a spinning top. If the weight is perfectly centered, it spins smoothly. But if you tape a tiny coin to the side (shifting the center of mass by less than 1%), the top starts to wobble, spiral, or change its path entirely.
The Finding: The researchers found that for tiny particles, even a tiny shift in weight (like a tiny air bubble inside a plastic bead or a slightly denser spot in a mineral) completely changes the motion.
- A particle that was supposed to spin in a perfect circle might suddenly spiral into a straight line.
- A particle that was falling straight down might start to tumble wildly.
- It's like a tiny nudge that sends a falling leaf on a completely different journey.

4. Why This Matters (The "So What?")

This tool is a game-changer for several fields:

Nature: It helps us understand how "marine snow" (dead plankton and poop falling in the ocean) moves, which affects how carbon is stored in the deep sea.
Medicine: It helps design microscopic robots (microrobots) that can swim through blood to deliver drugs. If you know exactly how a specific shape moves, you can design a robot that swims straight to a tumor instead of spinning in circles.
Industry: It helps engineers design better filters or mixers where particle shape matters.

Summary

The authors built a universal translator that turns a 3D shape file into a prediction of how that shape will move in thick fluid. They proved that even the tiniest imbalance in weight can turn a simple fall into a complex dance. This tool is now open for anyone to use, allowing scientists and engineers to design better particles for everything from cleaning water to curing diseases, all by simply looking at a 3D model on a computer screen.

1. Problem Statement

Sedimentation and low-Reynolds-number hydrodynamics are governed by a particle's hydrodynamic mobility tensor ( $6 \times 6$ matrix), which relates external forces and torques to translational and rotational velocities.

The Challenge: While analytical solutions exist for highly symmetric shapes (e.g., spheres, spheroids), predicting the mobility tensor and resulting dynamics for arbitrary, complex 3D shapes remains difficult.
Limitations of Current Methods:
- Bead models: Often lack direct access to the full mobility tensor and ignore fine geometric features.
- Boundary Element Methods (BEM) / Multipole expansions: Highly accurate but computationally expensive and difficult to scale for complex geometries.
- Experimental inference: Requires specific setups and is not a general predictive tool.
The Gap: There is a need for an efficient, generalizable framework that can directly convert a standard 3D geometric model (like an STL file used for 3D printing) into a full rigid-body hydrodynamic response, including the effects of mass distribution (Center of Mass shifts).

2. Methodology: The STL-to-Stokeslet Framework

The authors propose a numerical framework that discretizes a particle's surface into a set of point forces called Stokeslets to compute the mobility tensor.

Input: The method accepts standard STL files (triangulated surface meshes) without requiring geometric decomposition or symmetry assumptions.
Discretization:
- The nodes of the STL triangulation serve as the positions $\{r_1, \dots, r_N\}$ of $N$ Stokeslets distributed on the particle surface.
- The particle is treated as a rigid body; Stokeslets maintain relative positions during motion.
Mathematical Formulation:
- Velocity-Force Relation: The velocity field generated by a Stokeslet is described by the Oseen tensor. The total velocity at any Stokeslet $m$ is the sum of contributions from all other Stokeslets plus a self-drag term:
  $u_m = \sum_{n} O(r_m - r_n) f_n + \frac{1}{6\pi\mu a_{eff}} f_m$
- System Assembly: This forms a linear system $\mathbf{u} = \mathbf{G}\mathbf{f}$ , where $\mathbf{G}$ is a $3N \times 3N$ dense matrix.
- Rigid Body Constraint: The velocities $\mathbf{u}$ are constrained by rigid body kinematics ( $\mathbf{u}_n = \mathbf{V} + \mathbf{\Omega} \times \mathbf{r}_n$ ). By solving for six independent rigid-body motions (3 translations, 3 rotations), the authors extract the forces and torques required to drive them.
- Mobility Tensor Extraction: By inverting the force-torque relationship, the $6 \times 6$ mobility tensor $\mathbf{M}$ is obtained:
  $\begin{pmatrix} \mathbf{V} \\ \mathbf{\Omega} \end{pmatrix} = \mathbf{M} \begin{pmatrix} \mathbf{F} \\ \mathbf{\tau} \end{pmatrix} = \begin{pmatrix} \mathbf{A} & \mathbf{B} \\ \mathbf{B}^T & \mathbf{C} \end{pmatrix} \begin{pmatrix} \mathbf{F} \\ \mathbf{\tau} \end{pmatrix}$
  Where $\mathbf{A}$ is translation-translation, $\mathbf{C}$ is rotation-rotation, and $\mathbf{B}$ is the translation-rotation coupling tensor.
Center of Mass (CoM) Shifts: The framework incorporates density inhomogeneities by applying translation theorems to the mobility tensor. A shift in the Center of Mass ( $\mathbf{d}$ ) modifies the coupling tensor $\mathbf{B}$ :
$\mathbf{B}' = \mathbf{B} + \mathbf{C}[\mathbf{d}]_\times$
This allows the prediction of dynamics for particles with non-uniform density without re-meshing.
Optimization: The authors determine an optimal hydrodynamic radius factor for the Stokeslets ( $a_{eff} = \beta \ell$ , where $\ell$ is the median nearest-neighbor distance). They find $\beta \approx 0.25$ minimizes error across various discretizations.

3. Key Contributions

Direct STL-to-Mobility Pipeline: A user-friendly, efficient algorithm that computes the full $6 \times 6$ mobility tensor directly from STL files, bypassing complex geometric modeling.
Validation: The method is rigorously validated against:
- Analytical solutions: For spheres and thin disks, showing convergence as $N$ increases.
- Experimental data: Reproducing sedimentation trajectories of helical ribbons and U-shaped disks from recent literature.
Discovery of CoM Sensitivity: The study demonstrates that sub-1% shifts in the Center of Mass (relative to the geometric center) can qualitatively alter sedimentation dynamics, breaking symmetries and inducing fixed points in the orientation dynamics.
Open Source Tools: The authors provide implementations in both Mathematica and Python, making the tool accessible to the soft matter and biophysics communities.

4. Key Results

Convergence: For simple geometries (sphere, disk), the numerical error decreases logarithmically with the number of Stokeslets ( $N$ ). The optimal parameter $\beta \approx 0.25$ yields results largely independent of the specific discretization density.
Complex Geometries:
- Chiral Particles (e.g., DNA helicoids, snowflakes): Exhibit closed orbits on the unit sphere of force directions, leading to complex helical or spirograph-like sedimentation paths.
- Non-Chiral/Asymmetric Particles (e.g., Dolphin shape, Ice grains): Converge to fixed points or limit cycles. The "Ice grain" (fractal) showed a large antisymmetric component in $\mathbf{B}$ , leading to helical motion despite being non-chiral in shape.
Impact of CoM Shifts (The "Double-Paddle" Experiment):
- Symmetric CoM: Results in periodic, spirograph-like trajectories.
- Shift along Principal Axis: Modifies the trajectory quantitatively but maintains periodicity.
- Shift off Principal Axis (even 0.2%): Introduces a large antisymmetric component to $\mathbf{B}$ , breaking the symmetry. This causes the particle to spiral toward a fixed point (stable orientation), fundamentally changing the motion from periodic to a settling alignment behavior.
Efficiency: Computing the mobility matrix for $N \approx 10^4$ Stokeslets takes approximately 1 minute on a standard personal computer.

5. Significance and Applications

Predictive Design: The framework enables the rational design of "steerable particles" for applications in colloidal transport, targeted drug delivery, and micro-surgery by tuning shape and mass distribution.
Biological Relevance: It provides tools to model the locomotion of microorganisms (e.g., bacteria with flagella) and the sedimentation of plankton or marine snow, where shape and density inhomogeneities are critical.
Generalizability: Unlike previous methods restricted to simple shapes, this approach handles arbitrary 3D geometries, making it a versatile tool for fluid mechanics, soft matter physics, and biophysics.
Limitations: The method assumes a rigid body with a no-slip boundary (approximated by discrete Stokeslets). It is not intended for scenarios requiring extreme numerical precision (convergence is logarithmic) or for highly deformable bodies, though it can be extended to weak deformations as perturbations.

In summary, this work bridges the gap between 3D geometric modeling and hydrodynamic prediction, offering a robust, efficient, and accessible tool to understand and engineer the motion of complex particles in viscous fluids.