383475 Crystallization Kinetics in an Airlift Crystallizer: A Parameter Estimation Study

Friday, November 21, 2014: 9:14 AM
301 (Hilton Atlanta)
Fatemeh Anisi1, Jurian Jansen van Gaalen1, Jeroen van Krochten1, Pieter Vonk2, Richard Lakerveld1 and Herman J. M. Kramer1, (1)Process & Energy Department, Delft University of Technology, Delft, Netherlands, (2)DSM Chemtech Center, Geleen, Netherlands

Batch solution crystallization is one of the oldest downstream processes, which involves the contribution of several phenomena such as growth, nucleation, agglomeration, dissolution etc. Among those kinetic processes the growth of seeds is often the only desirable one, whereas primary and secondary nucleation and agglomeration give rise to problems in both process and product control leading to variability in the product quality and unwanted crystal size distributions (CSD), shapes and impurity uptakes.

Crystallization in traditional stirred-vessel crystallizers, in which especially secondary nucleation is difficult to avoid, often leads to a wide multimodal CSD. Recently it has been shown [1,2] however that when crystallization is performed in a recently developed air lift crystallizer a strong suppression in the amount of fine crystals can be obtained compared to a stirred crystallizer; probably due to the suppression of the rate of secondary nucleation by attrition.  

To prove this suppression in the rate of secondary nucleation and to get more insight in the kinetics and thus in the performance in the airlift crystallizer a more quantitative study is necessary in which the nucleation and growth phenomena are modelled and analyzed in both type of crystallizers.

The mixing strategy is the major distinction between an air lift and a stirred vessel crystallizer.  In an airlift crystallizer, injected air bubbles are responsible for the mixing and the suspension of the crystals in an airlift, whereas the impeller fulfils these tasks in a stirred vessel crystallizer. It is most likely that the differences in hydrodynamics between the two types of crystallizer could have a strong impact on the crystallization kinetics. Especially secondary nucleation might be affected due to the absence of the high energetic collisions between the impeller and the crystallizer which are dominant in a stirred vessel crystallizer. [3-6].

In this contribution, the growth and nucleation behavior during batch operation in an air lift crystallizer for L-ascorbic acid/water system has been analyzed. Model verification and parameter estimation was performed based on experiments under a range of process conditions. A comparison is made between the results of the eighteen liter air lift crystallizer and that of a five liter stirred tank crystallizer to get more insight in dominating crystallization mechanisms in the airlift; allowing a further optimization of this novel type of crystallizer.

Crystallization experiments, described earlier [1], have been used with a number of different supersaturation profiles, batch times and seed loadings to estimate the parameters of the crystallization models in a model verification and parameter estimation procedure. The measured supersaturation profile and CSD of seeds and final product from these experiments were used to select the most appropriate models for nucleation and growth and to estimate the relevant kinetic parameters using a variety of models implemented in gPROMS Model Builder.

 Both size dependent and size independent growth models are used along with different models for primary and secondary nucleation. Different secondary nucleation mechanisms were taken into account ranging from attrition based to activated nucleation mechanism [2,3,5,6,7].

 The results show indeed a strong suppression of the secondary nucleation by attrition in the airlift crystallizer compared to stirred vessel crystallizer under all studied process conditions, leading to larger crystals and a reduced number of fines. However, the supersaturation profile and CSD in the airlift crystallizer cannot be simply described by a reduction in attrition rate. The parameter estimation study revealed that a different, activated nucleation model is needed to describe the experimental results properly. Apparently in the absence of the impeller in the airlift crystallizer, which avoids the attrition caused by the crystal-impeller and crystal-baffle collisions, an alternative secondary nucleation mechanism becomes dominant. An activated surface nucleation model, in which it is assumed that preordered clusters or pre-nuclei are formed in the immediate vicinity of the crystal surface or on the crystal surface [7] gives the best fit for the experimental data of the air lift crystallizer.

This means that although secondary nucleation by attrition is suppressed in airlift crystallizer, secondary nucleation is not completely absent in this type of crystallizer for the studied process conditions. The indication  of the presence of an alternative nucleation mechanism is important as it opens possibilities to optimize the production of high quality crystals in this type of crystallizer.

Acknowledgments: This work was supported by the European Commissions Framework 7 program through the OPTICO consortium

References:

  1. R. Lakerveld, J.J.H van Krochten, H.J.M.. Kramer, An airlift crystallizer can suppress secondary nucleation at a higher supersaturation compared to a stirred crystallizer, submitted (2013)
  2. Soare, A., Lakerveld, R., van Royen, J., Zocchi, G., Stankiewicz, A.I., Kramer, H.J.M., 2012. Minimization of Attrition and Breakage in an Airlift Crystallizer. Ind Eng Chem Res 51, 10895-10909.
  3. J. W. Mullin, M. Chakraborty, K. Mehta,  Nucleation and growth of ammonium sulphate crystals from aqueous solution, Journal of Applied Chemistry, 20, 12 (1970)
  4. A. SoareS. A. Pérez Escobar, A. I. Stankiewicz , M. Rodriguez Pascual , H. J. M. Kramer, 2-D Flow and Temperature Measurements in a Multiphase Airlift Crystallizer, Ind. Eng. Chem. Res., 52, 34 (2013)
  5. S.K. Bermingham, H.J.M. Kramer, and G.M. van Rosmalen, Towards on-scale crystalliser design using compartmental models. Computers & Chemical Engineering, 22 (1998)
  6. Kramer, H.J.M., S.K. Bermingham, and G.M. van Rosmalen, Design of industrial crystallisers for a given product quality. Journal of Crystal Growth,. 198 (1999)
  7. A. Mersmann, B. Braun, M. Loffelmann, Prediction of crystallization coefficients of the population balance, Chemical Engineering Science, 57, 20 (2002)

Extended Abstract: File Not Uploaded
See more of this Session: Modeling and Control of Crystallization
See more of this Group/Topical: Separations Division