Letting the Repulsion Grow: Watching Colloids Order as Their Charge Builds

Almost every experiment on colloidal crystallisation sets the interparticle potential once — by choosing a salt concentration or a solvent — and then watches the system relax toward whatever that potential prefers. Jung, Revadekar and co-workers, writing in Nature Communications, built a system where the potential itself changes while you watch, and used it to catch the ordering transition in the act.

Close-up of soap bubbles packed into a dense raft
Repulsive discs pushed together find six neighbours each. Photograph by John Thomas (Unsplash); illustrative, not from the study.

What they did

The platform is an emulsion nested inside an emulsion. Using co-flow microfluidics the authors produced droplets of cyclohexyl bromide — a weakly ionic oil, abbreviated CHB — each one carrying many aqueous alginate sub-droplets, and dispersed the whole thing in an aqueous poly(vinyl alcohol) solution. Alginate is the gelling polysaccharide from brown seaweed; it crosslinks when it meets divalent cations.

The outer aqueous phase supplies those cations. Barium ions accumulate at the water–CHB interface, diffuse through the oil, and progressively crosslink the alginate sub-droplets. Crosslinking raises each particle's surface charge, so the repulsion between neighbours strengthens steadily over hours rather than being fixed at the start. Meanwhile buoyancy presses the particles into locally quasi-two-dimensional layers against the inside of the CHB droplet, which is what makes them trackable one by one under a microscope.

The result is a time-lapse of ordering. Over twelve hours in 40 mM barium acetate the top layer runs from a loosely correlated arrangement to a clearly hexagonal lattice, with Voronoi analysis showing the mean coordination number tightening toward six — 5.6 ± 1.2 early on, 5.9 by the end, with the spread narrowing as it goes.

Why it matters

The appeal is that the control parameter is a clock. In a conventional experiment you compare several samples prepared at several salt concentrations and infer the path between them. Here a single droplet traverses the path, with every particle's position known at every step, so the structural signatures of the transition are measured on one system rather than stitched together across samples.

That matters beyond this particular chemistry. Biological and industrial colloids rarely sit at a fixed interaction strength — ionic conditions drift as reactions proceed, as solvent evaporates, as a gel sets. A platform where the drift is deliberate and resolvable is a way to ask what ordering looks like when it never reaches equilibrium.

The key physics

Two numbers carry the mechanism.

The first is the Debye screening length, κ−1, which sets how far a particle's electrostatic influence reaches before the surrounding ions blur it out. The authors could not measure it directly in the oil phase, so they inferred it mechanically: they extracted an effective lattice spring constant from the particles' cage-relative thermal displacements — how far each one rattles inside the cage its neighbours make — then ran Brownian dynamics simulations with a screened-Coulomb (Yukawa) potential, tuning the zeta potential and the screening length until the simulated lattice had the same stiffness. That calibration puts κ−1 at roughly 2.5 to 3 micrometres, which is enormous by aqueous standards and is exactly what makes CHB useful: in a weakly ionic oil, charges talk to each other across several particle diameters.

The second is a dimensionless interaction parameter, Γ, the ratio of the electrostatic interaction energy between neighbours to the thermal energy kBT. It is the same bookkeeping as the electrostatic–thermal competition that decides whether a suspension stays dispersed: one organising influence divided by the randomising one. Γ climbs as the crosslinking proceeds, and hexagonal order emerges once it reaches roughly 117 to 149. Below that, thermal motion wins and the arrangement stays liquid-like.

The supporting controls are clean. Ordering switches on sharply between 0 and 0.5 mM barium acetate, and monovalent potassium or sodium at 40 mM leaves the sub-droplets essentially random — consistent with divalent crosslinking, rather than ionic strength as such, being what builds the charge. There is also a depth gradient: the top layer is both the most ordered and the one with the largest particles (mean diameter 4.4 ± 0.7 µm, against 2.7 ± 0.3 µm three layers down), a reminder that buoyant sorting and ordering are happening together here rather than independently.

Open questions

The honest limitation is in the paper's own framing: because the interaction strengthens continuously, no single pair of values for the zeta potential and the screening length describes any particular instant, and the reported figures rest on a representative choice. Pinning the two independently, in situ, would turn Γ from a well-grounded estimate into a measurement.

The reversibility result points somewhere interesting too. Doping the alginate with iron oxide nanoparticles made the assemblies magnetically responsive, and the ordered state recovered after both mechanical and magnetic perturbation. A system whose repulsion is still growing while it is being shaken is a natural testbed for how far an out-of-equilibrium lattice can be driven and still find its way back.

Where it sits

This work lands next to a question we return to often: what decides whether a suspension organises or stays disordered, and which competition is doing the deciding. In the classical account that competition is between electrostatic repulsion and van der Waals attraction, with the screening length as the dial. Here the dial is turned slowly and on purpose, and the answer arrives as a threshold in a dimensionless group rather than as a phase boundary inferred from endpoints. It is a good demonstration that a well-chosen dimensionless number is not a summary of an experiment — it is the thing the experiment was built to measure.

Jung, I. H., Revadekar, C. C., Lee, H. S., Hwang, H., An, H., et al. & Park, B. J. “Nonequilibrium ordering dynamics of confined soft alginate hydrogel colloids driven by time-evolving electrostatic interactions.” Nature Communications, 17, 3662 (2026). DOI: 10.1038/s41467-026-70266-w (open access)

Next
Next

Give the Particles a Third Dimension and They Build Something Else