CrossNucleation.html

Understanding and controlling in which structure (or polymorph) a molecule crystallizes are complex and long-standing issues. In many cases, a process yields crystals of more than one polymorph at the same time, making the control of polymorphism even more challenging. Several mechanisms have been proposed to account for this phenomenon, known as concomitant polymorphism. It has been attributed either to competing processes of homogeneous nucleation of different polymorphs or to solvent-mediated conversion of one polymorph into another or, more recently, to the heterogeneous nucleation (or cross-nucleation) of one polymorph on another. We use molecular dynamics simulations to study the early stages of crystallization in a supercooled liquid of spherical particles. We observe the onset of concomitant polymorphism and demonstrate that this phenomenon results from the cross-nucleation of a metastable polymorph on the stable polymorph. We also show that cross-nucleation is selective since it only takes place between polymorphs of almost equivalent free energy. Our simulations provide detailed insights into the molecular mechanism underlying concomitant polymorphism and cross-nucleation.

L. Yu C. Desgranges and J. Delhommelle
J. Am. Chem. Soc. 125, 6380 (2003) J. Am. Chem. Soc. 128, 15104 (2006)

Cross nucleation between polymorphs in D-mannitol (left) and Lennard Jones (right) systems.

Molecular Mechanism for the Cross-Nucleation between Polymorphs

We show that cross-nucleation is mostly governed by kinetics. First, we observe the nucleation of a metastable polymorph (hcp) on a crystal of the stable polymorph. Moreover, cross-nucleation takes place because the metastable polymorph (hcp) grows at a faster rate than the stable polymorph (fcc), as shown in Figure 1 for t = 57. These two findings are supported by the experimental results from Yu et al. In their study of cross-nucleation, they show that metastable polymorphs may nucleate on stable polymorphs. Besides, they conclude that the new polymorph formed grows faster than (or at least as fast as) the one initially present. However, while cross-nucleation is essentially controlled by kinetics, our results also show that the relative stability of the polymorphs plays an important role. We only observe cross-nucleation between polymorphs of almost equivalent free energies. If the free energy of a polymorph is notably larger, as it is the case for bcc here, particles of this polymorph may appear, but they quickly convert into one of the more stable polymorphs before a large cluster forms. This shows that cross-nucleation is selective.

C. Desgranges and J. Delhommelle J. Am. Chem. Soc. 128, 15104 (2006)
C. Desgranges and J. Delhommelle, J. Phys. Chem. B. 111, 1465 (2007)

Controlling Cross-Nucleation

We examine the results obtained for the growth step. We present in Fig. 4 the averages for the number of hcp particles, calculated over all the MD trajectories, obtained for the highest and lowest degrees of supercooling, i.e., 22% and 10%, against the number of particles contained in the crystallites. Our results demonstrate that we succeed in controlling cross nucleation. Figure 4 clearly shows that, by increasing the temperature, we manage to reduce by a factor of 2 the increase in the number of hcp particles with the size of the nucleus, which suggests that cross nucleation has been prevented for a supercooling of 10%. This is conﬁrmed by the snapshots presented in Fig. 5. For the highest supercooling (22%), we observe the formation of layers of hcp particles and thus of large hcp domains within the crystallite. This corresponds to the cross nucleation of the hcp form on the fcc form. For the lowest supercooling (10%), we do not observe the formation of such large hcp domains. hcp particles form on the surface of the fcc core of the crystallite and their number increases with the surface of the nucleus. Therefore, at low supercooling, we do not observe any cross nucleation and obtain an essentially pure fcc crystallite. We interpret this result as follows. The cross nucleation of the hcp form on the fcc form is a kinetic phenomenon, which, like any other heterogeneous nucleation process, is associated with a free energy barrier of activation. As for homogeneous nucleation, increasing the temperature (or decreasing the supercooling) results in an increase in the height of the free energy barrier. This, in turn, prevents cross nucleation and allows us to form pure fcc crystallites at low supercooling.

C. Desgranges and J. Delhommelle, Phys. Rev. Lett. 98, 235502 (2007).