You could be forgiven for thinking that the vision of port wine crawling up the inside of a glass was simply the result of having drunk the experiment. Yet James Thomson investigated such a phenomenon 150 years ago, and attributed the cause to gradients in surface tension. Writing in Physical Review Letters, Yutaka Sumino et al. now report that this so-called Marangoni effect makes oil droplets perform even more stunning tricks. When placed on a glass substrate submerged in water, the droplets spontaneously climb stairs or loop-the-loop on the inside of a vertical glass hoop (see pictures, in clockwise sequence from top left).
The spectacle is driven by the action of surfactant molecules, which attach uniformly to the surface of glass substrates placed in solution, thereby creating a hydrophobic coating. But when the surfactants are next to a droplet, they move from the surface into the oil phase. The resulting local perturbation of the coating makes the surface less hydrophobic, and creates a gradient in the surface tension between the front and rear of the droplet that induces motion.
Soon after a droplet has passed by, the surfactant molecules that it has removed are replaced from the solution, returning the substrate to its hydrophobic state. Unlike in other self-running droplet systems, trajectories can therefore cross each other repeatedly. It is this feature that makes the droplet's loop-the-loop possible. It also means that, by using narrow glass strips, the droplets can be forced from random movement into regular back-and-forth motion.
All movement stops after a few tens of seconds, however. Sumino et al. attribute this to the depletion of charged complexes in the oil droplets. As long as they are present, the complexes pair up with surfactant molecules and accelerate them across the oil–water interface. Without this effect, the surface-tension gradient falls below the value needed to drive droplet motion against viscous damping.
About this article
Cite this article
Helmer, M. Oiled acrobatics. Nature 433, 693 (2005). https://doi.org/10.1038/433693a