Crystallization is voodoo. I have not changed my opinion one bit since this 2015 post – if anything, I’m more convinced than ever. We chemists tend to talk about such fields with some pride, because we have to deal with them by our hard work and our wits, but when you get right down to it, it’s kind of an embarrassing situation. Not being able to predict melting points is a related embarrassment, and what this tells you is that there are things about the transition to an ordered solid phase (or out of it) that we do not understand very well at all. A lot of very tricky, counterbalancing, inter-related thermodynamics are in play as you move in or out of a liquid state, and we’re not up to handling all of them yet.
So what we rely on, until that far-off day when we can just calculate our way out of such difficulties, are empirical tricks and techniques. The number of these in the crystallization/solubility field are beyond counting (a sure voodootropic indicator if ever there was). And I just read about a new one, which seems to have now been put on a bit firmer theoretical framework.
It was discovered in 1996 that shining a small, bright laser beam into a concentrated solution can induce crystallization, and no one has been quite sure how that works. Or maybe it’s better to say that several attempts have been made to be sure about that, without general agreement. There have been many reports since then of this effect, even to the point of producing different polymorphs in a controlled fashion. That area has had some reproducibility problems, but the general laser-induced crystallization phenomenon does not seem to be in doubt.
Several proposals have been made for a mechanism, but this new paper has an interesting one: it appears that the intense laser flux promotes movement of those molecules with a higher refractive index into the beam. This “laser tweezer” effect concentrates a given species in one zone, then, and depletes it in the surrounding one. This works for liquid-liquid separations (as shown in this paper, using nitrobenzene and decane), which is an easier test system. But it should apply to all sorts of phase transitions, including crystallization:
Because the laser-tweezing potential depth scales with the refractive index of the new phase and because nearly all solids have a higher refractive index than their corresponding liquid or solution phase, this effect can in principle explain all known laser-induced crystal-nucleation results, with the exception of nucleation induced by pulsed-laser-induced vapour or plasma bubbles.
. . .This is a generic effect that does not only apply to poorly mixing liquids but to any mixture or solution. However, the ease with which the laser-tweezing potential can initiate phase separation is enhanced near a liquid–liquid demixing critical point or binodal line. Phase manipulation and nucleation can be induced with a straightforward low-power laser diode. This suggests that this effect can be used to control matter in a range of practical applications.
There will need to be some careful work done to control for things like local heating and viscosity (the present authors have thought about this as well), but overall, this looks like a potentially very interesting way to induce crystallization, to separate liquid phases, and more. Worth keeping an eye on!