During the antisolvent crystallization of L-histidine from aqueous solution with the antisolvent ethanol, the polymorphic fraction of the metastable polymorph B increased with increasing supersaturation ratio. For supersaturation ratio S A e 2.0, a mixture of the stable polymorph A and the metastable polymorph B was obtained, while for high supersaturation, S A g 2.3, only form B was observed. The supersaturation ratio was increased either by increasing the L-histidine concentration in the aqueous solution or by increasing the ethanol volume fraction. The induction time for the antisolvent crystallization was one to several orders of magnitude smaller than the time to transform the metastable polymorph B into stable polymorph A in solution, indicating that both polymorphs formed simultaneously. Process simulation of the batch antisolvent crystallization of the polymorphs of L-histidine was applied to predict the outcome of the competition between heterogeneous nucleation and polynuclear growth rates for both polymorphs. When the interfacial energy was calculated as a function of the solubility, only the formation of the stable polymorph A was erroneously predicted. By modification of the relative interfacial energy and the relative step free energy, the predicted polymorphic fraction could be directed toward the observed polymorphic fraction.

For precipitation processes, measurement of the nucleation rate as a function of supersaturation is not so easy because the formation of the precipitate is fast and often hard to control due to the high supersaturation. For a number of studies on both ionic and molecular precipitating compounds, the experimental methods are discussed. From the experimentally determined nucleation rates, the probable nucleation mechanism, homogeneous or heterogeneous, is derived and compared to the theoretically expected mechanism. For all ionic compounds and for one molecular compound considered in this work, theoretically homogeneous nucleation is achievable. However, only for two ionic compounds, barium sulfate and boehmite, the measurements possibly indicated a homogeneous nucleation mechanism. For the other studies, including all molecular compounds, a heterogeneous mechanism was more likely. The difference between a theoretically expected and experimentally observed mechanism may be explained by not only the use of less suitable experimental methods and measurement techniques but also by secondary processes, such as agglomeration and transformation, taking place in the precipitated suspension. On the basis of these findings, an experimental procedure is suggested to standardize and to improve nucleation rate measurements in precipitation.

This paper describes the onset of a new approach for the prediction of the nucleation rates of polymorphs crystallized from solution. The first step in the development of the new method is to come to an accurate prediction of the interfacial energy between the nucleus and solution, one of the key factors in the nucleation process. This prediction method is described in the paper and is tested for a simple binary Lennard-Jones mixture. The interfacial energy term shows a decreasing trend with increasing cluster size at cluster sizes between 300 and 750 atoms.

Measurement of the stationary nucleation rate as a function of supersaturation in reactive precipitation requires that nucleation and growth of the crystals to an observable size take place under well-controlled hydrodynamic conditions and at a constant supersaturation. An experimental method was chosen that is based on the measurement of the increase of the particle concentration with the increase of the residence time in a precipitation tube. This method was applied to measure the nucleation rate for the pH shift precipitation of the molecular compound H4EDTA. The parameters A and B in the classical nucleation theory equation were derived from the measured nucleation rates. A value of A = (5.7 ± 1.0) × 1015 m-3 s-1 was found for the intercept with the ordinate and a value of B = (3.3 ± 2.0) × 102 was found for the slope. The estimated value for parameter A is lower than the theoretical value of A HON = 1033 m-3 s-1 for homogeneous nucleation. Using the experimentally determined value for B the interfacial energy is found to be γ = 21 ± 9 mJ m-2 compared to a theoretical value of γHON = 38 mJ m-2 based on bulk properties. The experimentally obtained values for A and γ indicate heterogeneous nucleation.