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

Wenting Cheng, Tiago Pernambuco, Thomas A. Witten, Haim Diamant, Justin C. Burton · 2026

Read STL-to-Stokeslet Computation of Mobility Tensors and Sedimentation Dynamics for Shaped Particles by Wenting Cheng, Tiago Pernambuco, Thomas A. Witten, Haim Diamant, Justin C. Burton (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-00158.txt.

Introduction

Cheng and colleagues present a numerical framework that computes a particle's mobility tensor from a 3D-printer STL file, splitting the surface into point drag forces called stokeslets. The tensor fixes how a particle moves and spins under given forces and torques at low Reynolds number. Validated against analytic solutions and recent experiments, it shows shifting the centre of mass off a symmetry centre collapses all starting orientations into one, two or three final motions. The tool aids steerable-particle design for colloidal transport, locomotion, diffusion and self-assembly.

Catalog record

CREATORWenting Cheng, Tiago Pernambuco, Thomas A. Witten, Haim Diamant, Justin C. Burton
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE40470
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.00158

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

STL-to-Stokeslet Computation of Mobility Tensors and Sedimentation Dynamics for Shaped Particles Wenting Cheng, Tiago Pernambuco, Thomas A. Witten, Haim Diamant, Justin C. Burton Abstract Sedimentation is extremely common in nature, occurring throughout the atmosphere and oceans, and in every laboratory centrifuge. The shape and mass distribution of a particle uniquely determines its motion at low Reynolds number, and complex dynamics can emerge from even simple particle shapes. The dynamics are governed by the particle’s hydrodynamic mobility tensor, which dictates the translational and rotational velocities given the forces and torques . However, to date the inference of the mobility tensor from the object shape has been cumbersome and tricky. Starting with an input file representing an object for a 3D printer, such as an STL file, here we present an efficient numerical framework to compute the mobility tensor by discretizing the particle surface into distributed point drag forces called stokeslets. We validate our results against analytical solutions of simple geometries and recent experimental measurements. With our calculated mobility tensors in hand, using standard transformation laws, we demonstrate the dramatic effect of shifting the center of mass from a center of symmetry: all initial orientations evolve into one, two, or three particular final motions dictated by the object. By providing a user-friendly and efficient framework to compute the mobility tensor and resulting particle dynamics, this work offers a broadly applicable tool for the soft-matter, fluid-mechanics, and biophysics communities, and facilitates the design of steerable particles under diverse external forces, with relevance to colloidal transport, biological locomotion, diffusion, and self-assembly. Significance statement: Sedimentation is a fundamental process across many natural systems, from atmospheric aerosols and marine snow to the locomotion of microorganisms against gravity. The shape of a particle crucially determines its sedimentation behavior, often leading to complex rotational and orbital trajectories. However, predicting these dynamics for arbitrary particle shapes remains a challenge. Here we demonstrate an efficient numerical framework to compute sedimentation dynamics in viscous fluids directly from a three-dimensional model. The framework employs a user-friendly computational interface, can handle highly complex particle shapes and density inhomogeneities, and serves as a predictive tool for tuning locomotive properties of particles in viscous fluids and their responses to various external forces. The motion of particles in viscous fluids is governed not only by external forces but also by particle shape and mass distribution ; these can induce nontrivial coupling between translation and rotation 45 . A simple example is marine snowfall, where complex particle shapes form through aggregation and sediment over years 5 . Moreover , shape can directly influence interactions, leading to both attractive and repulsive effective forces 18 ; 33 ; 23 . Yet, the dynamics of even a single shaped particle is rich, exhibiting fixed points, limit cycles, and slow transients manifested as twists, tumbles, and orbital motion. Predicting these dynamics requires accurate evaluation of the particle’s 6\times 6 mobility tensor 27 ; 21 ; 45 ; 19 ; 29 , which fully determines how forces and torques are converted into translational and rotational motion in the creeping-flow regime where inertial effects are negligible . Classical analytical results for rigid-body hydrodynamics are largely restricted to highly symmetric shapes such as spheres 21 . Numerical methods often represent particles as assemblies of point forces (stokeslets) acting on the fluid 28 , and the resulting flow fields and particle motions can be obtained through linear algebraic operations 48 . However, their validation and practical application have remained largely confined to relatively simple geometries such as spheroids and cylinders. This limitation motivates the development of efficient and generalizable approaches capable of treating particles with arbitrary shape. Modern methods for structure-based hydrodynamic modeling are now well developed 4 . A popular computational approach treats particles as assemblies of beads with varying degrees of complexity in their interactions 9 ; 10 ; 34 ; 37 ; 40 ; 3 ; 39 ; 16 ; 50 ; 7 . Many of these methods are designed to efficiently predict single-valued transport properties, for which fine geometric features are largely irrelevant 11 . As a result, they do not provide direct access to the mobility tensor. High-accuracy methods based on multipole expansions 14 ; 6 ; 15 or boundary element formulations 2 ; 1 can in principle recover the exact mobility tensor, however, their substantial computational cost and limited scalability have largely restricted their use to benchmark studies or simplified geometries. Recent studies have sought to access the mobility tensor using bead-based simulations combined with experimental trajectory fitting 22 , direct inference from sedimentation experiments 31 , or theoretical calculations based on boundary integral methods 35 ; 27 for highly symmetric particles 24 . While successful for particular geometries, these approaches remain either computationally intensive or specialized, underscoring the need for a general, efficient framework that directly links arbitrary particle geometry to its full rigid-body hydrodynamic response. To overcome these limitations, we introduce an STL-to-stokeslet framework that computes the rigid-body mobility tensor and the resulting dynamics directly from an STL file of an arbitrarily shaped particle. These files are commonly used for 3D printing, and the method requires no geometric decomposition, symmetry assumptions, or experimental calibration. Using the stokeslet discretization implicit in the STL primitives and prescribing independent rigid-body motions, the mobility tensor is efficiently extracted and used to predict responses to external forces and torques. We validate our results using analytical solutions of simple particle geometries 21 and recent sedimentation experiments using helical ribbons 22 and bent disks (i.e., disks bent out of plane, which gyrate despite being nonchiral, sometimes called “U-shaped” disks in the literature) 31 . The framework further incorporates translations of the center of mass (CoM) through a simple transformation of the mobility tensor. We find that during sedimentation under gravity, the dynamics of certain particle shapes display a pronounced sensitivity to shifts in the CoM smaller than 1%. Such density variations are common in natural settings and in industry where particles are fabricated. The framework is highly efficient: on a standard personal computer, the mobility matrix for N\simeq 10^{4} stokeslets can be computed in approximately 1\,\mathrm{min} . By directly linking geometric shape to hydrodynamic response, our approach also provides a versatile framework to rationally design and control steerable particles under external forces. This enables a wide range of applications, from engineering locomotion at low Reynolds number, such as bacterial flagella 36 ; 26 ; 13 ; 41 , to predictive modeling of the sedimentation and swimming of natural particulate matter, including plankton 20 ; 43 . More broadly, it offers new opportunities for hydrodynamic trapping and flow-directed assembly of particles in solution 44 ; 17 , and establishes a geometry-based framework for synthesizing microscale swimmers in applications such as water decontamination, chemical mixing, targeted drug delivery, and noninvasive microsurgery 47 ; 38 ; 46 ; 12 . Figure 1: Sedimentation of arbitrarily shaped particles. (a–d) Input geometries from STL files (left) and their stokeslet discretizations (middle) for a DNA-shaped helicoid, a chiral snowflake, a dolphin-like body, and a fractal ice grain 32 , respectively. The right panel shows a zoomed-in view of the highlighted region, illustrating the local distribution of the stokeslets on the discretized surface. (e–h) Predicted CoM sedimentation trajectories under gravity for

Read the complete text