On the Kinetics of Palm Oil Crystallisation
Abstract
In this work the crystallisation of palm oil (a typical melt) is used as a model process and kinetic parameters at different end temperatures and cooling rates were evaluated using three model approaches: two from the traditional melt fractionation, one from the classic solution crystallisation. The objective was to establish critical understanding on the various model approaches and their applications. Differential Scanning Calorimetry (DSC) was used to obtain the melting temperatures, isothermal induction times and enthalpy of the crystallisation for the Fisher-Turnbull and Avrami model evaluations, while turbidity and temperature probes were utilized to generate metastable zone width as a function of cooling rates for the classical Nyvlt model analysis. Our results show that the Fisher-Turnbull and the Nyvlt models are useful in estimating the nucleation rate constants (kn) with reasonable agreement: this unites the model approaches and allows comparison between fat fractionation and solution crystallisation of organics. While the Avrami model is capable of evaluating the growth mechanism of the formed crystals (n) and the overall crystallisation rate constant (k), none of the parameters can be compared with other models because of the different definitions of the growth mechanism as well as the dependence of k value on the growth mechanism when dealing different fats of varying compositions. This is the first contribution of this work. In addition, our results indicate that the growth of nuclei to a stable size was generally slower in melt crystallisation compared to a typical organic solution system due to the relatively lower rate constant (kn) and the high viscosity and multicomponent properties of the melt system used in this study. This is the second contribution of this work.
Keywords
Download Options
Introduction
Crystallisation by cooling is a well-known unit operation used in a wide range of industries from pharmaceuticals to food for product purification, separation and/or recovery. For pharmaceutical ingredients, for example, the purity of the solute out of the solution is a critical product parameter. For edible oils/melts in food industry, on the other hand, it is the purity or the clearness of the solution that matters. Although the terminology of crystallisation is used for all inorganic and organic compounds, while fractionation is applied to all melts, the essence of the process is the same, i.e. by cooling for all. In spite of the similarities in the processes, different model approaches have been used for crystallisation from solution (e.g. APIs) than from melts (e.g. fats). Taking palm oil as the model compound for melt, it consists of a wide range of triglyceride (TAGs) fractions based on their physical state at different temperatures[1] – olein (liquid), mid-fractions (soft solid) and stearin (solid). These are significantly different to crystallisation of paracetamol for instance. The aims of this work are to evaluate and compare kinetics from a melt crystallisation extracted using models for both melts and organics; and to establish critical understanding on the various model approaches and their applications.
Conclusion
In this work we use the crystallisation of palm oil (a typical melt) as the model process and evaluated the kinetic parameters using three model approaches: two from the traditional melt fractionation, one from the classic solution crystallisation. The objective was to establish critical understanding on the various model approaches and their applications. Our results show that the Fisher-Turnbull and the Nyvlt models are useful in estimating the nucleation rate constants (kn) with reasonable agreement: this unites the model approaches and allows comparison between fat fractionation and solution crystallisation. The lower rate constants predicted from the melt crystallisation in comparison to that from the solution crystallisation of organics are expected and resulted from the high viscosity that impedes the growth of nuclei to a stable size.
While the Avrami model is capable of evaluating the growth mechanism of the formed crystals (n) and the overall crystallisation rate constant (k), none of the parameters can be compared with other models because of the different definition of the growth mechanism as well as the dependence of k value on the growth mechanism when dealing different fats of varying compositions.
Finally, the end temperatures and cooling profiles are the essential operational parameters that can be explored in order to obtain crystal product of desired physical properties.