Nucleation, growth and dissolution of faceted single crystals

Abstract

The crystallisation and dissolution (non-sink) behaviour from the solution phase is studied for some selected pharmaceutical representative materials, notably urea and paracetamol, and is interrelated to an assessment of surface chemistry of the crystal habit facets. The inclusion of a small amount of biuret, a known decomposition impurity of urea, is found to increase the solution metastable zone. Polythermal analysis is consistent with a concentration dependence of the nucleation behaviour of both the pure and the doped systems associated with the mechanism changing from progressive to instantaneous with increasing concentration, with a concomitant decrease in the interfacial tension and a significant increase in nucleation rate of the doped system, from 9.22-20.48 to 9.25-67.73 nm-3. s-1, and decrease in the critical nucleus size. The crystal habit of urea in solution is found to be dominated by the {110} and smaller polar {111} capping faces resulting in the {-1-1-1} not being observed. The mean crystal growth rates of the {110} and {111} faces are found to increase with respect to supersaturation with {111} exhibiting a greater level of increase than {110}. The addition of biuret to the solution is found to have a greater effect on retarding the growth of {111} compared to {110}, resulting in a more compact morphology. Rationalising this behaviour with computational molecular modelling studies reveals stronger additive binding on {111} compared to {110}. The mean crystal dissolution rates of {110} and {111} faces of urea in ethanolic solutions are found to increase with respect to the degree of undersaturation, with the dissolution behaviour being mechanistically consistent with the growth behaviour. The mean crystal dissolution rates of both faces in acetonitrile are very similar to each other, and to the dissolution rate of the {110} face in ethanol. The dissolution rate of the {111} face in ethanol is found to be faster than that of the other faces. Rationalising this behaviour with computational molecular modelling reveals a higher wetting energy of {111} compared to {110}. Dissolution modelling based on the experimental data, were consistent with boundary layer thicknesses of 0.5μm and 0.3μm for the same undercoolings for ethanol and acetonitrile, respectively. These values are smaller than expected but are consistent with modelling data. The relative solubilities of paracetamol are higher in acetonitrile than in fed state simulated intestinal fluid (FeSSIF). The crystal habit of paracetamol in solution is ii found to exhibit five equivalent morphologically significant faces, giving rise to a prismatic crystal. The mean crystal dissolution rates of these faces are found to increase with respect to degree of undersaturation in acetonitrile, with the dissolution rates of all faces being very similar. The mean crystal dissolution rate of these faces is found to increase with respect to temperature in FeSSIF, with the dissolution rates of the faces being similar. The dissolution rates in acetonitrile and FeSSIF are rationalised through prediction of the intermolecular interactions. Dissolution modelling revealed the boundary layer thicknesses to be 0.3μm and 0.1μm for acetonitrile and FeSSIF, respectively. This might reflect the greater number of binding sites of water compared to acetonitrile, as well as the assumption that water is a representative probe for FeSSIF. The importance of this work in enhancing the quality of dissolution testing is also highlighted, notably, the utility of relating dissolution properties at the single particle level to the same material as it progresses throughout the drug product processing cycle.