On the Self-Purification Cascade during Crystal Growth from Solution

We report on the failure of the Cabrera–Vermilyea (CV) step pinning model to reproduce the elementary step kinetics for the case of tetragonal lysozyme crystals growing from contaminated solutions. We measured the supersaturation dependency of the step velocity using confocal microscopy for three different commercially available lysozyme batches with varying levels of impurity content, that is, Seikagaku, Fluka, and Sigma. Strong nonlinear dependencies are obtained in the high to intermediate supersaturation range and near-linear dependencies at lower driving forces. The clear absence of a dead zone for the Fluka and Seikagaku data is in direct contradiction to the CV model. As such, we developed a time-dependent impurity model based on Bliznakov kinetics assuming Langmuir adsorption. Admissible fits are obtained for Fluka and Seikagaku lysozyme corroborating the self-purification interpretation due to the diminishing terrace exposure times at higher supersaturation levels. The steeper recovery toward pure kinetics for Sigma lysozyme than predicted by Langmuir adsorption prompted us to expand the model to allow for impurity–impurity interaction. The resultant kinetic model, which assumes a Kisliuk-like mode of impurity adsorption, did yield acceptable fits with Sigma step kinetics. This Bliznakov–Kisliuk model also predicts clustering of impurity molecules on the surface, which is corroborated by our in situ experimental atomic force microscopy observations.