Crystallization
Controlled Batch and Continuous Crystallizations
With the Balanced Nucleation and Growth (BNG) Model
For controlled crystallization of precision Crystal Sizes the Balanced Nucleation and Growth (BNG) model was developed in an industrial setting by Dr. Leubner (see book). This model is a new and powerful Crystallization Theory which links crystal nucleation and growth with experimental reaction control parameters.
The model fills the Practical Need to relate the results of crystallization processes to experimental reaction control parameters. Until now supersaturation was considered the main control parameter for crystallizations; however the experimental determination of supersaturation and its control frequently has been tedious and has been unusable for highly insoluble crystals. During the nucleation phase (see below), supersaturation goes through wide swings based on the reactant addition rate and other factors. The BNG method eliminates the need to use supersaturation as a control parameter. Supersaturation during crystal growth (following nucleation) is quantitatively determined the from the reaction results.
The BNG approach models the Nucleation Phase based on known practical crystallization reaction control factors: temperature, reactant addition rate, crystal solubility, ripeners and restrainers. Further, the model and experiments allowed correcting erroneous ideas about the importance of initial reactor content and geometry, and the control of mixing and stirring.
The effects of Growth Restrainers for nano-sizing and of Ripening Agents for macro-sizing during crystal nucleation were first described, quantitatively modeled, and experimentally confirmed by Dr. Leubner. Inadvertent presence of restrainers and ripeners in the reactor content, reactants, and protective colloids will lead to non-reproducibility of crystal size and distribution.
For the Nucleation Phase presently untreatable processes, like supersaturation, nucleation rate, crystal formation and size distribution were modeled and related to crystallization reaction control factors. The nucleation phase model translated seamlessly to the analytical quantitative model of relating the number of crystals formed and thus their size as a function of the experimental reaction control parameters.
For controlled Batch Crystallizations, the BNG model quantitatively correlated the crystal number with crystallization reaction control factors and gave guidance for the most effective and practical experimental design to quantify and plot the results.
For Continuous Crystallizations, the BNG model related the crystal size to the reaction parameters and the reactor residence (reaction turn-over) time. Previously unknown theoretical predictions were confirmed for the crystal size as a function of reaction volume, residence time, reactant addition rate, suspension density, and solubility. The steady-state of crystal population distribution was found to occur between eight and twelve residence times, Not the four residence times for achieving mass-steady-state.
For Seeded Continuous Precipitations, the crystal size distribution was modeled for reactor residence time, controlled crystal growth rate, and reactant addition rate. This model also allows calculating the size distribution in non-seeded continuous crystallizations.
Renucleation is the formation of new crystals during crystal growth where the reactant addition rate exceeds the maximum growth rate of the seed crystal population. It is suspected that in many crystallization processes renucleation mimics secondary nucleation, i.e., breaking up of crystals by stirring.
The number of newly formed crystals due to renucleation was quantitatively related to the number and geometry of the seed crystals and to the reactant addition rate. Experimental results confirmed the model predictions. They allowed the precision determination of the maximum growth rate of the seed crystals.
Precipitations based on Temperature Changes or Non-Solvent Addition are frequently used. These processes can be modeled by the BNG model. This work is in progress.
