Janus Nanomotor: Catalytic Self-Propulsion

Pt-capped Janus microswimmers in a periodic 44 um bath under a Blender-built microscope: ballistic motion crossing into an enhanced random walk, with live MSD, Deff and orientation-memory graphs.

Switch to 2D

Presets

Playback

Parameters

Camera

Layers

Quality

Model check

The science

What is modelled

A Janus particle has two faces: a silica body and a thin platinum cap (gold-grey shell in the model). In hydrogen peroxide the cap decomposes the fuel, 2 H2O2 -> 2 H2O + O2. The product gradient around the particle drives a slip flow along its surface (self-diffusiophoresis), so the particle swims along its own axis at speed v0, away from the cap here. Thermal kicks from the water also move it and tumble its axis. Every particle is an active Brownian particle (ABP): persistent self-propulsion competing with rotational diffusion. The scene shows a 44 um cube of bath with periodic edges, looked at under a microscope stage; the pipette adds fuel for show.

Equations

  • Viscosity of water: eta(T) = 2.414e-5 * 10^(247.8 / (T - 140)) Pa s (T in kelvin), times the viscosity factor.
  • Stokes-Einstein: translation Dt = kB T / (6 pi eta r) and rotation Dr = kB T / (8 pi eta r^3). Rotation is very sensitive to size: halving the radius makes the axis diffuse eight times faster.
  • Speed law: v0 = v_max c / (K + c) with K = 2.2 %. A phenomenological saturation curve, not a measured constant.
  • Langevin step: r += v0 u dt + sqrt(2 Dt dt) N(0,1) per axis; the unit axis u is rotated about a random vector of length sqrt(2 Dr dt) |N| (Rodrigues rotation), which gives <u(t).u(0)> = exp(-2 Dr t) in 3D.
  • Mean-square displacement, exact for an ABP in d dimensions with g = (d-1) Dr: MSD(t) = 2 d Dt t + (2 v0^2 / g^2)(g t - 1 + exp(-g t)). Short times: ballistic v0^2 t^2. Long times: diffusive with Deff = Dt + v0^2 / (d g), in 3D Dt + v0^2 / (6 Dr).
  • Persistence length v0 / g and the Peclet number Pe = v0 r / Dt measure how far the swimmer goes straight.

Checks run in the page (re-run whenever you move a slider)

  • 1500 fresh passive walkers must give MSD / (6 t) = kB T / (6 pi eta r) (independent random seed, no walls).
  • 2500 fresh active walkers must reproduce the exact MSD formula and the orientation memory at a lag of two memory times.
  • The long-time slope of the exact MSD law must equal Deff (consistency of the closed form). The three graphs compare the visible swimmers (64 tracers, time-averaged over many origins) with the same formulas live.

Honest simplifications

The swimmers neither touch each other nor the glass (the 44 um window is periodic; only the display wraps), there is no hydrodynamics beyond Stokes drag, no gravity or tilt (real Pt particles are denser than water and sit near the bottom), and the speed law is a fit rather than derived from the surface chemistry. Real swimming speeds are a few um/s and fall for larger spheres (Howse et al. 2007); v_max is a free slider here so you can explore the regimes. The orange glow and the bubble stream behind the gold tracer are visual aids. Fuel molecules and the microscope are drawn far out of scale (a water molecule is a thousand times smaller than the dots), and the stage is a decorative backdrop.

What to try

  • Turn the cap off: the green MSD curve drops onto the passive 6 Dt t line.
  • Raise the fuel level: the slope-2 ballistic stretch appears at short lags and bends over at the memory time (graph 1), and the dot in graph 2 rides up the Deff curve.
  • Pick the 0.5 um preset: smaller particles lose their heading eight times faster and look like fast random walkers.
  • Pick the thick medium preset: viscosity lowers both Dt and Dr, so the same swimmers travel straighter but slower.
  • Use the Follow tracer camera to ride along with the gold swimmer and watch product bubbles stream from its cap.
Loading model...
Loading Blender model...
Mean-square displacement vs lag (log-log)
Effective diffusion vs fuel level
Orientation memory
Made with 🟠 Blender · Three.js