OK, let’s completely nerd out today on crystallization. That’s a topic dear to organic chemists everywhere, because crystals are obviously terrific per se, but they have a great deal of practical importance. Crystallization has always been a frontline technique for purifying compounds – done under the right conditions, you can form extremely pure crystals of your desired substance while leaving almost all of the junk floating around in the solution to be rinsed off. Filtering off a pile of crystals and watching them turn white as you pour solvent over them (and watching the colorful impurities pour through into the vessel below) never fails to bring joy, whether you’re an undergraduate in your first course of organic chemistry or an experienced process chemist scaling up a new drug intermediate.
And the type of crystals that you form can be very important indeed. As those process chemists know (generally to their sorrow) most compounds are capable of forming polymorphs, different crystalline forms of the same substance. These things can range widely in their properties, with totally different melting points, ease of dissolution in various solvents, ability to soak up water from the atmosphere or tendency to take solvent along with them while forming their crystal lattices – a whole range of things that you care very much about if you’re trying to formulate a new drug in a reproducible manner. Patent disputes have erupted, and how, many times over the years about the properties of these things. That link above will take you to some famous examples of desirable polymorphs that suddenly became difficult or impossible to produce when a more stable one happened to form, a process that in extreme cases is indistinguishable from the effects of an evil wizard throwing a curse on you. Solving polymorph problems has cost a great deal of money and effort over the years, and we are not very good at all about predicting when something like this is going to happen. Here’s a single compound with twelve well-characterized crystal forms, and it’s a proving ground for attempts to predict how many more might be waiting out there.
This earlier post on the crystallization of glycine will illustrate the difficulties. I mean, You Would Think That the crystallization behavior of the simplest amino acid would be completely worked out by now, but you would be wrong. It isn’t just glycine, either; phenylalanine is a notoriously difficult crystallography problem. As you see from reading these, there are extremely subtle processes at work on the nanoscale with large consequences for all of us up here. There are all sorts of tricks to try to get an overview of the polymorph landscape for any individual compound, some of them quite odd, and a lot of ways to try to produce particular ones on demand. That last process generally is a case-by-case thing which is occasionally done (or very nearly) to the sound of alarm bells and frantic cursing.
One of these odd tricks, known for about 25 years, is to take a concentrated solution of a compound and shine pulses of laser light into it. That seems to nucleate crystal formation – something happens under the laser illumination that aligns molecules in ways that are more likely to produce crystals, but what exactly that is has been the subject of much debate. This new paper is a good summary of the area, and has some new data that might tell us what’s really happening. As it references, there are many pieces of evidence for some unusual nanoparticle forms that lead to crystallization – glycine, for example, may well have nM-sized droplets of a more glycine-rich liquid phase (shades of biomolecular condensates!) that are the precursors to actual crystals, and both it and other compounds show evidence of larger amorphous clusters that can lead to crystal formation, too. We know very little about these species, since they’re so small and exist so transiently.
An oddity of the laser-nucleation method is that it seems to work much better when the solutions involved are aged, as opposed to freshly formed. That’s been seen with all sorts of substances, and it has always suggested that there’s some sort of organization that gradually takes place as the solutions sit around. The laser pulse then can take advantage of this, while in a fresh solution it has nothing to work with. This paper shows that aged glycine solutions accumulate tiny amorphous particles, up to one micrometer in size, but finds that these apparently are not themselves intermediates for crystallization. These particles are characterized by Raman spectroscopy, and are not just glycine hydrates. But they are indeed what reacts to laser irradiation, being somehow activated and producing a crystal nucleation site.
The effect of the laser might be heating, causing reorganization within the particle, which leads to a speedup of classical nucleation. On the other hand, the electric field of the laser may facilitate the crystallization from the amorphous phase by inducing net alignment and reducing the free energy barrier through the optical Kerr effect, which may now operate on a micrometersized structure (providing sufficient high local concentration as opposed to individual molecules), allowing the laser polarization to influence the polymorph selection.
That Kerr effect hypothesis is supported by the observation that the alpha polymorph of glycine is what you get by spontaneous nucleation (or concentrating a glycine solution down to dryness), while laser irradiation produces significant amounts of the gamma polymorph. In this work, the authors saw that in about 38% of the laser-formed crystals. There’s an unstable glycine polymorph between those (the beta form, naturally) which is thought to be an intermediate for both the alpha and gamma forms, but this team was unable to detect it. That might be because it’s too short-lived for their Raman spectroscopy experiments to see it, or it might be that the electric-field-driven Kerr effect is allowing the beta form to be bypassed entirely. Finer-grained experiments will be needed to resolve that question. But this work does strongly suggest that laser-induced crystal nucleation is indeed working on amorphous clusters in solution, which take time to form, which solves some of the “aged solution” mystery.
