Mistery of Crystallization

For many years scientists have scratched surfaces to promote crystallization. Crystals are observed to nucleate along the scratches. Despite the popularity of this technique, our understanding of why crystals nucleate in scratches is very limited. There have been no quantitative measurements of the nucleation rate in a groove, and there is no data on the microscopic details of how the nucleation occurs here.

In a recent report in JACS, Researcher from U.K. studied the heterogeneous nucleation of a crystal in a wedge-shaped groove via computer simulation.They found that nucleation in these grooves is indeed many orders of magnitude faster than on a flat surface. It is also been found that there is a competition between the angle of the wedge, and the angles that are intrinsic to the lattice of the crystal. This competition results in a wedge angle at which the nucleation rate goes through a maximum, a phenomenon that is not seen in the nucleation of liquids.

Therefore, generically we expect crystallization to start in grooves or pits in surfaces, not on the flat parts of surfaces. This expectation is consistent with the common observation that crystallization readily occurs on a scratched surface. It been found that nucleation is highly sensitive to the wedge angle β. This is because there are angles that are intrinsic to crystals, such as the 70.5° angle between the close-packed planes in an fcc lattice. As this effect is a direct consequence of the crystal lattice it has been expected to be a general feature of crystallization, even with more complex molecules, for example proteins. Nucleation is faster when the wedge angle is such that a defect-free unstrained piece of the crystal fits perfectly into the wedge. As different crystal polymorphs have different intrinsic angles, this may provide a way to control the polymorph that nucleates. A wedge into which the desired polymorph fits perfectly, but which has the wrong angle for other polymorphs, will favor nucleation only of the desired polymorph.

Reference : J. Am. Chem. Soc., 2009, 131 (48), pp 17550–17551