A Computational Analysis of a Combined Cooling Anti-Solvent Continuous Crystallization Cascade: A Paracetamol/Methanol/Water Case Study

With increased attention being given to the development of continuous crystallization systems to replace existing batch crystallizations, the need for a deeper understanding of their behaviour has arisen and as such much research has been conducted into this field [1] – [7]. Experimental determinations and characterisations of continuous processes are notoriously labour, time and material intensive as compared to batch investigations. To combat this, complex computational models of continuous crystallizations are needed to compliment the experimental investigations.

Developed here, in the MATLAB computing environment is a numerical model of continuous combined cooling and anti-solvent crystallization of Paracetamol in methanol and water. The model incorporates the effects of particle growth, nucleation (primary and secondary), aggregation and breakage as well as convection through the vessel, assumed to be perfectly mixed. The kinetic parameters used in the model were obtained and validated from batch cooling and anti-solvent experimental investigations [8] [9]. The methods and technologies developed and implemented here are very valuable in that the model can be applied to or adapted for any crystallization once kinetics are known.

Due to the highly non-linear behaviour and physical inter-dependence of such processes, it is necessary to be able to determine the underlying driving mechanisms for particle dynamics. For example, is the system growth, aggregation or nucleation dominated? Analysis techniques for simple growth/nucleation systems have existed for some time and are well understood, however techniques to elucidate this complex cascade behaviour is lacking if non-existent. In this work, methods have been developed to analyse the underlying particle dynamics for cascaded anti-solvent/cooling crystallizations. The algorithms developed in this work allow for the determination of the relative contributions of each of the physical phenomena to the underlying dynamics. For example, it may be determined what the dominant source of particle size increase or surface area generation is. This decoupling of the physics affords the capacity to design/modify continuous systems to maximise or minimise any one or more desirable or undesirable mechanisms of crystallization.

Optimisation of crystallization systems’ operation is vital for their continued efficient operation. Outlined in this text are methods for determining the operating conditions which will give either a maximum particle size or maximum recovery for a given feed mass flow and tank configuration. It must be noted that this is not limited to the maximum particle size but may be used to target a specific particle size.

The input parameters of interest here are the feed mass flow rate, relative sizes of the two tanks and the distribution of anti-solvent and feed across the tanks. Preliminary findings indicate that the distribution of feed and anti-solvent across the cascade has some impact on the solute recovery but the largest response is seen in the particle size distribution. The maximisation of recovery is generally seen where all feed is directed to the first tank and a portion of the anti-solvent is directed to the first tanks with the remainder directed to the second tank. The optimum distribution of anti-solvent is seen to be a function of the total feed mass flow rate and the relative tank volumes. For each configuration (flow rate and volume) there exists an optimum distribution of streams to maximise recovery and/or to target a particular particle size.

The model has the capability of fully analysing the overall behaviour of the system as well as affording a near complete understanding into the underlying physics and dynamics of such a system. It must be noted that the results shown here are for a single feed composition with no recycle and at a fixed stage temperature. Further analysis of the effects of composition, stage temperature and recycle is possible without further development.

In conclusion, a powerful computational model has been developed which, if coupled with experimental campaigns and data will afford a very deep understanding of the dynamics of continuous crystallization. This depth of understanding can feed directly into the design of new processes or the optimisation of the operation of existing processes.

References

[1] H. Zhang, J. Quon, A.J. Alvarez, J. Evans, A.S. Myerson, B. Trout, Development of Continuous Anti-Solvent/Cooling Crystallization Process using Cascaded Mixed Suspension, Mixed Product Removal Crystallizers. Organic Process Reseach & Development  16  (2012) 915-924.

 [2] C.J. Callahan, X.W. Ni, Probing into Nucleation Mechanisms of Cooling Crystallization of Sodium Chlorate in a Stirred Tank Crystallizer and an Oscillatory Baffled Crystallizer. Crystal Growth & Design12 (5) (2012) 2525–2532.

[3] C.J. Brown and X. Ni, Online evaluation of paracetamol antisolvent crystallisation growth rate with video imaging in an oscillatory baffled crystalliser. Crystal Growth & Design 11 (2011) 719-725.

[4] S. Lawson, G. Steele, P. Shering, I. Laird, L. Zhao and X. Ni, Continuous crystallisation of pharmaceuticals using a continuous oscillatory baffled crystalliser. Organic Process Research & Development (13) (2009) 1357-1363.

[5] Z.K. Nagy, G. Fevotte, H. Kramer, L.L. Simon, Recent advances in the monitoring, modelling and control of crystallization systems. Chemical Engineering Research and Design, In Press, Corrected Proof, Available online 27 July 2013.

[6] S.Y. Wong, Y. Cui, A.S. Myerson, Contact Secondary Nucleation as a Means of Creating Seeds for Continuous Tubular Crystallizers. Crystal Growth & Design (2013).

[7] S. Qamar, M.P. Elsner, I. Hussain, A. Seidel-Morgenstern,  Seeding strategies and residence time characteristics of continuous preferential crystallization. Chemical Engineering Science (71) (2012) 5-17.

[8] C. Ó'Ciardhá, K. Hutton, N. Mitchell, P. Frawley, Simultaneous parameter estimation and optimisation of a seeded anti-solvent crystallization. Crystal Growth and Design, 12 (11) (2012) 5247–5261.

[9] C.T. Ó’Ciardhá, P.J. Frawley, N.A. Mitchell, Estimation of the nucleation kinetics for the anti-solvent crystallisation of paracetamol in methanol/water solutions. Journal of Crystal Growth 328 (1) (2011) 50-57.