Original Research Paper
Experimental kinetics studies of seeded batch crystallisation of mono-ammonium phosphate

https://doi.org/10.1016/j.apt.2010.02.017Get rights and content

Abstract

A seeded batch cooling crystalliser in a laboratory setup was utilised to produce mono-ammonium phosphate (NH4H2PO4) crystals. The effects of different initial saturation, seed size and seed loading ratio on crystal size distribution were thoroughly studied to gather more insight into mono-ammonium phosphate batch crystallization, its kinetics, and for model validation purposes. The seed loading ratio shows the most significant effect on the crystal size distribution profiles followed by the supersaturation factor and seed size. The supersaturation level not only influences the growth rate but the crystal shape as well. A simultaneous method has been applied in the procedure for the estimation of crystallisation kinetics. The nucleation constant (kb) was found to be in the range of 1.0–6.9 × 108 [no/kg s (kg/kg)b+1] and the growth constant (kg) measured was in the range of 3.0–9.4 × 10−5 [m/s (kg/kg)g] with the nucleation (b) and growth (g) exponents determined as 2.2 and 1.4, respectively. Wavelet orthogonal collocation method was utilised to solve the seeded batch crystalliser model and the solution was compared with the experimental data.

Introduction

Previous research into mono-ammonium phosphate (MAP) crystallisation dates back to 1967 and 1970 by Mullin and co-workers [1], [2]. A single seed crystal with a constant supersaturation in an isothermal batch crystalliser was used and only growth kinetics was reported. The use of seeds in batch crystallisation processes improves the product quality, specifically the crystal size distribution (CSD) and the purity [3], [4], [5]. The known seeding effects are qualitative and highly dependent on the operating conditions and the solute thermodynamics. Quantitative information is limited, for example it is not possible to determine the seed amount and sizes that should be introduced in order to obtain a desired size distribution. A seeding technique is adopted on the basis of a trial-and-error approach in industry and is not guaranteed to perform optimally. Certain important aspects such as the seed loading ratio, the seed sizes, and the initial supersaturation effect have been studied to obtain more information on the process and to obtain the crystallisation kinetics.

There are two general methods used to determine crystallisation kinetics in batch crystallisers: the method of isolation, and the simultaneous method. The classical approach is to isolate the growth and nucleation processes, and then determine their kinetics separately by direct and/or indirect methods under different hydrodynamic conditions. The isolation method allows calculation of the rates of the elementary processes from separate experimental results. The simultaneous method requires the experimental observation of several output variables during experiment. Kinetics data are then calculated after appropriate mathematical treatment which is intended to separate the contribution of the observed variables on the growth, nucleation, and other mechanisms. This mathematical treatment of data may become complex if all the rates are to be determined accurately.

There is a significant number of publications describing the development of wavelet theory [6], [7], [8], [9], [10], [11], and the wavelet method is now one of the accepted applied numerical methods used to solve an integral-partial differential population balance equation. Wavelet Orthogonal Collocation (WOC) method was developed by Bertoluzza and Naldi [9] and uses trial functions generated by autocorrelation of the usual compactly supported Daubechies scaling functions. Liu and Cameron [12] applied Wavelet Orthogonal Collocation (WOC) methods for a chromatographic [13] and population balance application [14].

Our studies on seeded crystallisation in a batch mode have been conducted in order to obtain essential information on the kinetics, to provide insight into the process characteristics and for the MAP seeded batch crystallisation model validation based on wavelet method.

Section snippets

Materials and methods

The fundamental phenomena controlling the crystallisation process is the saturation concentration of a substance in a solvent, i.e. solubility. Solubility data are generally obtained experimentally by determining the maximum amount that is soluble at a given temperature. Buchanan and Winner [15] undertook work on mono-ammonium phosphate (MAP) and di-ammonium phosphate (DAP) solubilities. The range of temperature investigated was 5–90 °C. Their work provided fundamental studies on MAP and DAP

Effect of seed size (constant seed loading)

The seed loading ratio or seed concentration (Cs) is defined as the ratio of the actual mass of seeds loaded to the theoretical yield of crystal mass due to maximum supersaturation [20]. For the same seed loading ratio, i.e. Cs = 5%, the different seed sizes will result in different average sizes of the final crystal product. The seeds with a single peak distribution grew and formed a dual peak product size distribution. The 137.5 μm (set-77) and 165.0 μm (set-72) mean size seeds grew to sizes of

Conclusions

  • (1)

    In the MAP seeded cooling crystallisation system studied here, the seed loading ratio has the most significant effect on the CSD profiles, followed by the supersaturation factor and the seed size. Similarly, as also observed in a non-seeded system, the cooling rate, stirring speed and crystal yield do not significantly affect the CSD product if enough seeds are loaded under well-mixed conditions. In addition, the degree of supersaturation influences not only the kinetics but also the crystal

Recommendation

This kinetics data can be used for the map reactive crystallisation system which was performed in a separate study. The crystallisation and reactive kinetics obtained can be used for crystalliser design, performance testing, and as a tool in simulating the same or similar system with a limited research to be performed.

Acknowledgements

The first author acknowledges the support from the Australian government, the University of Hyogo, and Curtin University of Technology during his studies from February 2005 to February 2009.

References (21)

There are more references available in the full text version of this article.

Cited by (5)

View full text