Determination of the positions and orientations of concentrated rod-like colloids from 3D microscopy data

T H Besseling,B De Nijs,M Hermes,M Dijkstra,T-S Deng,A Imhof,A Van Blaaderen,A Kuijk

doi:10.1088/0953-8984/27/19/194109

T H Besseling, B De Nijs + Show 6 more

Open Access

https://doi.org/10.1088/0953-8984/27/19/194109

Copy DOI

Abstract

Confocal microscopy in combination with real-space particle tracking has proven to be a powerful tool in scientific fields such as soft matter physics, materials science and cell biology. However, 3D tracking of anisotropic particles in concentrated phases remains not as optimized compared to algorithms for spherical particles. To address this problem, we developed a new particle-fitting algorithm that can extract the positions and orientations of fluorescent rod-like particles from three dimensional confocal microscopy data stacks. The algorithm is tailored to work even when the fluorescent signals of the particles overlap considerably and a threshold method and subsequent clusters analysis alone do not suffice. We demonstrate that our algorithm correctly identifies all five coordinates of uniaxial particles in both a concentrated disordered phase and a liquid-crystalline smectic-B phase. Apart from confocal microscopy images, we also demonstrate that the algorithm can be used to identify nanorods in 3D electron tomography reconstructions. Lastly, we determined the accuracy of the algorithm using both simulated and experimental confocal microscopy data-stacks of diffusing silica rods in a dilute suspension. This novel particle-fitting algorithm allows for the study of structure and dynamics in both dilute and dense liquid-crystalline phases (such as nematic, smectic and crystalline phases) as well as the study of the glass transition of rod-like particles in three dimensions on the single particle level.

Highlights

Colloidal particles are applied throughout industry, for example in paints, personal care products, food, ceramics and pharmaceutics [1,2,3]
Apart from confocal microscopy images, we demonstrate that the algorithm can be used to identify nanorods in 3D electron tomography reconstructions
This novel particle-fitting algorithm allows for the study of structure and dynamics in both dilute and dense liquid-crystalline phases as well as the study of the glass transition of rod-like particles in three dimensions on the single particle level

Summary

Introduction

Colloidal particles are applied throughout industry, for example in paints, personal care products, food, ceramics and pharmaceutics [1,2,3]. Hard-sphere colloidal suspensions have proven to serve as a model system to investigate phenomena such as crystallization, the glass transition and flow induced behaviour on the single particle level [5,6,7,8,9,10,11,12] In many of these studies, an image processing technique was applied based on the algorithm described by Crocker and Grier [13]. Mohraz and Solomon, were one of the first to determine the 3D position and orientation of uniaxial ellipsoidal particles, i.e. all five coordinates, using 3D confocal microscopy and a novel anisotropic feature-finding algorithm [23] Their algorithm identifies the points that are located on the central axis (or backbone) of a rod. This algorithm enabled the quantitative determination of the 3D translational and rotational motion of a dilute suspension of ellipsoids [32]

Methods

Results

Discussion

Conclusion