Accueil > English > News

The Ouzo effect : nanosynthesis made simple

Nature is fond of variety, but it also knows that a uniform population is easier to control than all different individuals. In the context of emulsions, a natural phenomenon is known from ages to lead spontaneously to a uniform population of tiny droplets : the Ouzo effect, that appears when water is poured into anise-flavored aperitif, such as the pastis. Mixing water, alcohol and the oil from the anise seed, gives a lot of droplets, each of about same submicrometric size, without any adjusting parameter… in the new world of the nanotechnology, one could speak of a quasi miraculous synthesis, for the simplicity and robustness of the process ! (who ever failed to prepare a pastis ?)

To understand in details this dramatic effect, experiments have been performed at ESPCI Paris on a model system, with smell not as pleasant as pastis, but easier to control : mix of water, acetone and a polymer that is soluble into acetone, and not in water ; kinetics (over 1 second, since emulsification is quite fast) was followed by ultrafast small angle X-ray scattering at ESRF Grenoble to measure the droplets radii ; numerical simulations were done at LPS Orsay to analyze the data, and to study the mechanisms.


Which scenario came out from these analyses ?... well, droplets acquire electrical charges because of the hydroxide ions HO<sup–, and coalescence must then stop when charges on the droplets are too large ; radius distributions become narrow because droplets of different charges are more likely to merge than droplets of similar charges. All right. Then, we had to check these mechanisms, that is to compare evolutions of the averaged droplet radius and the standard-deviation σ, to mathematical results of the kinetics taking into account electrostatic repulsion. Strangely (since the equations are complicated), the exact result is known, as found by Dammer and Wolf in 2004. However, this result said that σ and should be proportional each other, that is not consistent with our experimental data : width of the distributions, σ, increased experimentally at a much slower rate with . Conclusion should have been : either some ingredient was missing in the equations, or the analytical result was wrong. Actually, mischievous nature imagined an alternative way : the equations were correct as well as the exact result, but another solution existed…
At this point, numerical simulations of kinetics of coalescence have been useful. The experimental data were recovered in the simulations for the first second of physical time. Then, beyond that time, distribution widths increased still and always as the square root of , the same as in the experiments… Being patient, we left the simulation running, and, around a billion years of the physical time, things changed : the system settled on the exact Dammer and Wolf solution ! The pieces of the puzzle set up : the exact solution is mathematical correct, but appears at immeasurably long times.
Then, was the exact solution a kind of anecdote ? Actually, not really so, since it allowed us to understand that this change in the scaling laws was the expression of a general behavior – the theory of the universal fluctuations – and that the machinery of that effect has things in common with other, totally different, systems. However, this aspect will need another section to be developed…

Contacts :
Robert Botet
Roger Kevin
Bernard Cabane

References :
Coalescence of Repelling Colloidal Droplets : A Route to Monodisperse Populations
R. Kevin, R. Botet and B. Cabane
Langmuir 29, 5689 (2013).

Universal Fluctuations
R. Botet and M. Ploszajczak
World Sci. ed, Singapore (2002).