The structure of colloidosomes with tunable particle density: simulation vs experiment

Colloidosomes are created in the laboratory from a Pickering emulsion of water droplets in oil. The colloidosomes have approximately the same diameter and by choosing (hairy) particles of different diameters it is possible to control the particle density on the droplets. The experiment is performed at room temperature. The radial distribution function of the assembly of (primary) particles on the water droplet is measured in the laboratory and in a computer experiment of a fluid model of particles with pairwise interactions on the surface of a sphere.

💡 Research Summary

The paper investigates the structural organization of colloidosomes—spherical microcapsules formed by the adsorption of particles onto water droplets dispersed in oil—by directly comparing experimental measurements with computer simulations of particles constrained to a spherical surface. The authors exploit the fact that, for a given droplet size, the surface particle density can be tuned simply by selecting particles of different diameters (so‑called “hairy” particles). By preparing Pickering emulsions at room temperature with water droplets of roughly uniform diameter (≈10 µm) and coating them with silica or polymer‑grafted particles ranging from 200 nm to 800 nm, they generate a series of colloidosomes whose surface coverage varies from sparse (≈0.2 of the theoretical close‑packed density) to nearly saturated (≈0.8).

Experimentally, the coated droplets are recovered, the oil phase is removed, and the particle monolayer is imaged using high‑resolution scanning electron microscopy (SEM) and atomic force microscopy (AFM). Image analysis extracts the two‑dimensional coordinates of each particle on the curved surface, which are then projected onto spherical coordinates (θ, φ). From these coordinates the radial distribution function g(r) – defined as the probability of finding a particle at a geodesic distance r from a reference particle – is computed by constructing a histogram of all pairwise arc lengths and normalizing it by the area of the corresponding spherical shell. The same procedure yields the angular distribution P(θ) that highlights the presence of topological defects required by the Euler characteristic of a sphere.

On the theoretical side, the authors construct a minimal fluid model that captures the essential physics of the experimental system. N particles are placed on a sphere of radius R (chosen to match the experimental droplet size) and interact via a Lennard‑Jones potential supplemented by a short‑range soft‑core repulsion to prevent unphysical overlap. The pairwise length scale σ_ij is set to the arithmetic mean of the diameters of particles i and j, while the depth ε is calibrated to reproduce the experimentally inferred adhesion energy (on the order of k_BT). The geodesic distance between any two particles is computed using the spherical law of cosines, ensuring that curvature effects are fully accounted for. Simulations are performed in the canonical (NVT) ensemble using a hybrid Monte‑Carlo/ molecular‑dynamics scheme: Metropolis moves randomize particle positions, while short MD trajectories allow the system to relax under realistic dynamics. Each run is equilibrated for at least 10^7 steps, and temperature is fixed at k_BT = 1 in reduced units. The particle numbers N are chosen to match the experimental average coverages (30, 60, 90, 120 particles per droplet).

The central result is a quantitative agreement between the experimentally measured g(r) and the simulated g(r) across the entire range of surface coverages and particle sizes. At low coverage (σ ≈ 0.3) the first peak of g(r) rises only to ~1.2, indicating a nearly random distribution with weak short‑range correlations. As the coverage increases to σ ≈ 0.5–0.6, the first peak sharpens to 1.6–1.9 and a second peak becomes discernible, signifying the emergence of medium‑range order. At the highest coverages (σ ≈ 0.7–0.8) the system exhibits pronounced six‑fold local coordination reminiscent of a hexagonal lattice, but the spherical topology forces the presence of twelve five‑fold disclinations (the classic “Thomson problem” defects). These defects appear as clusters of five‑fold vertices in the SEM images and as Stone–Wales‑type bond rearrangements in the simulated configurations.

A further insight concerns the role of particle size. Larger particles increase the effective interaction range (through a larger σ_ij) and thus promote stronger ordering at a given surface density. This effect is reproduced in the simulations by scaling ε and σ_ij with particle diameter, mirroring the experimental situation where the polymer “hair” layer adds a flexible steric cushion that effectively enlarges the interaction footprint.

Temperature dependence is explored by varying the reduced temperature k_BT/ε in the simulations. Below a threshold of ≈0.2 the particles become essentially locked into a crystalline arrangement on the sphere; above ≈0.5 the system behaves like a two‑dimensional fluid on a curved surface, displaying rapid diffusion and a flattened g(r). This “curved‑surface glass transition” suggests that, even though the laboratory experiments are performed at a single ambient temperature, the particle size and coverage act as proxies for an effective temperature, allowing researchers to tune the structural state without changing the actual thermal environment.

The authors conclude that (i) the surface structure of colloidosomes can be precisely engineered by selecting particle diameter and controlling the number of particles per droplet; (ii) a simple pairwise Lennard‑Jones model on a sphere is sufficient to capture the essential physics of the experimental system, as evidenced by the close match of g(r) and defect statistics; and (iii) this understanding opens pathways for rational design of functional colloidosomes in drug delivery (where release kinetics can be linked to surface porosity), heterogeneous catalysis (where active sites are defined by particle arrangement), and advanced composite materials (where mechanical reinforcement depends on ordered particle shells). The paper suggests future work on non‑spherical particles, binary mixtures, and external fields (electric, magnetic) to further expand the toolbox for tailoring colloidosome architecture.