Batch Crystallization Process Research Articles

Real-time optimal control of an anti-solvent isothermal semi-batch crystallization process

This study presents a novel approach for the implementation of a real-time dynamic optimal control methodology on a seeded, anti-solvent batch crystallization process. The novel features of the approach include online estimation of the nucleation, growth rates and moments of paracetamol in isopropanol–water solution using the FBRM ® probe, and the real-time single and multi-objective optimization. In addition, the supersaturation was measured by ATR-FTIR spectroscopy. The crystallization model, which involved population and mass balance, was solved based on the finite difference method. The results of the multi-objective optimization showed a substantial improvement of the end of batch properties expressed in terms of the volume-weighted mean size and yield. The effect of mean size of seeds was studied as well, and it was found that the seeds with smaller mean volume-weighted mean size promoted the growth rate. The multi-objective optimal control resulted in an increase in the volume-weighted mean size by 120.8 μ m compared to the seeds and 51% of theoretical yield, which showed better performance compared to other optimal control policies.

Product quality improvement of batch crystallizers by a batch-to-batch optimization and nonlinear control approach

Batch crystallization is one of the widely used processes for separation and purification in many chemical industries. Dynamic optimization of such a process has recently shown the improvement of final product quality in term of a crystal size distribution (CSD) by determining an optimal operating policy. However, under the presence of unknown or uncertain model parameters, the desired product quality may not be achieved when the calculated optimal control profile is implemented. In this study, a batch-to-batch optimization strategy is proposed for the estimation of uncertain kinetic parameters in the batch crystallization process, choosing the seeded batch crystallizer of potassium sulfate as a case study. The information of the CSD obtained at the end of batch run is employed in such an optimization-based estimation. The updated kinetic parameters are used to modify an optimal operating temperature policy of a crystallizer for a subsequent operation. This optimal temperature policy is then employed as new reference for a temperature controller which is based on a generic model control algorithm to control the crystallizer in a new batch run.

Dynamic modeling of a batch crystallization process: A stochastic approach for agglomeration and attrition process

This paper deals with the dynamic modeling of a batch crystallizer. A complete model taking into account primary and secondary nucleations, crystal growth, agglomeration and attrition mechanisms is established. The proposed model is not restricted to binary agglomeration and breakage phenomena. From markovian considerations, continuous kernel functions are built and the basic balance equations are then presented. The complete model is solved using a finite difference method for the discretization of the size variable. As to distinguish agglomeration and breakage parameters from the others, on line measurement of the Crystal Size Distribution is necessary, a new on line measurement strategy is proposed. Finally, simulations of the crystal size distribution are compared with experimental results at different times. It appears that simulated curves are in good agreement with the experimental data.

ESTIMATION OF CRYSTAL SIZE DISTRIBUTION OF A BATCH CRYSTALLIZATION PROCESS USING A GROWTH SIZE DEPENDENT MODEL

A high gain observer was designed to estimate the Crystal Size Distribution in batch crystallization processes. This observer is based on a model where growth kinetics depend on crystal size. The observer is based on the discretization of population balance equations solved by the finite difference method. A reduction of the number of state variables was done in order to reduce the computation time. The observer's output allows to estimate the moments of the CSD, which are interesting for control purposes.

Optimal control of a batch crystallization process

A dynamic model of an aluminium trihydroxide batch crystallization is considered in this work. The process model, which takes into account kinetics of nucleation, growth and agglomeration, is based on the mass balance of the process and the population balance of the dispersed crystals. Assuming that the temperature of the solution and the seeding policy are variable, optimal control techniques are applied to the model to optimize various performance criteria. Some interesting numerical results for the behaviour of the model are presented.

Process Modelling Tools for Continuous and Batch Organic Crystallization Processes Including Application to Scale-Up

Computational fluid dynamics (CFD) is a powerful simulation tool that was successfully used to investigate mixing, turbulence, and shear in a laboratory-scale MSMPR and batch cooling crystallizer for an organic fine chemical. CFD gives a qualitative engineering insight into the effects of the impeller configuration on the crystallization rates and particle size distribution. A process-modelling tool, gPROMS (Process Systems Enterprise), was used to model particle size and size distribution in both batch and continuous laboratory-scale crystallization processes with predictive simulations in good agreement with experimental results. CFD simulations of large-scale crystallizations using constant specific power input per unit mass, predict an increase in macromixing and decrease in micromixing and turbulence. This effect should improve process performance of batch cooling crystallizers on scale-up including the product quality of the final solid form in terms of the particle size and crystal habit. This is d...

Product engineering for crystal size distribution

AbstractProduct engineering problems are considered that arise in the use of batch crystallization processes. Specifically forward and reverse process functions are identified that in the forward direction enable the calculation of product crystal size distributions (CSDs) from specified operating conditions, and in particular a specified temperature–time (T–t) profile. In the reverse direction the process function calculates the required processing conditions—the T–t profile—needed to make a specified product CSD. A series of example calculations is reported, based on the kinetics of the pentaerythritol–water system, that show that in the presence of nucleation and growth only, the relationship between desired, or target, CSD and required T–t profile is, in practice, unique. It is observed that some such profiles are either impractical or impossible, from which it is possible to conclude that the corresponding target CSD is itself impossible. A further example is considered in which the property function relating the product property (CSD) to the end‐use properties (cake porosity and filtration time) enables the specification of a target CSD and of the process conditions necessary for its production. © 2006 American Institute of Chemical Engineers AIChE J, 2006

Multi-objective optimization of seeded batch crystallization processes

The optimization of a batch cooling crystallizer has been traditionally sought with respect to the cooling profile and seeding characteristics that keep supersaturation at an optimum level throughout the operation. Crystallization processes typically have multiple performance objectives and optimization using different objective functions leads to significantly different optimal operating conditions. Thus different temperature profiles and seeding characteristics impose a complex interplay on the crystallizer behavior and there is a trade-off between the performance objectives. Therefore, a multi-objective approach should be adopted for optimization of a batch crystallizer for best process operation. This study presents the solution of various optimal control problems for a seeded batch crystallizer within a multi-objective framework. A well known multi-objective evolutionary algorithm, the elitist Nondominated Sorting Genetic Algorithm, has been adapted here to illustrate the potential for the multi-objective optimization approach.

Nonlinear observer of crystal‐size distribution during batch crystallization

AbstractA high‐gain observer was designed to estimate the crystal‐size distribution (CSD) in batch crystallization processes. The observer is based on the discretization of population balance equations describing the evolution of the CSD using finite difference method. Due to process impurities and other batch‐to‐batch variations, the kinetic parameters involved in the dynamic model of the crystallization, relating primary and secondary nucleation in particular, are subject to significant variations. In order to avoid any estimation divergence, an on‐line parameter identification algorithm was added to the observer. Assuming that measurements of the nuclei particles are available, the observer is shown to provide a discretized reconstruction of the entire CSD which can be used for control purposes or process supervision. © 2006 American Institute of Chemical Engineers AIChE J, 2006

Direct Design of Pharmaceutical Antisolvent Crystallization through Concentration Control

Recent advances in in situ measurement technology and automation of batch crystallizers have enabled the development of batch crystallization recipes in which the desired supersaturation profile is followed by feedback control. This paper describes a new approach for following supersaturation setpoints for antisolvent crystallizations that is easy to implement for the tried crystallization. Simulations and application to a proprietary drug compound demonstrate how this combination of automation and in process measurements enables the rapid development of batch crystallization processes in the pharmaceutical industry.

Calcite-seeded crystallization of calcium phosphate for phosphorus recovery

For phosphorus recovery from wastewater, the present paper aims at understanding the crystallization of calcium phosphate by using calcites (Juraperle and Coccolith) as seeds from hard water. Synthetic hard water with 60 mg C l −1 carbonate and 10 mg P l −1 phosphate was prepared and used in the batch experiments of calcite-seeded crystallization. The solution composition was measured throughout the batch crystallization process, and the corresponding saturation indices with respect to the minerals were calculated with PHREEQC Program. The surface of the calcite seeds was observed and measured with environmental scanning electron microscopy (ESEM), Fourier transform infrared (FTIR) and BET method. The studies show that both calcites are effective seeds for the crystallization of calcium phosphate from hard water; the used calcites are more efficient than the original ones because newly formed crystals with calcium phosphate have covered their surfaces. The studies show that the calcite seeds can be used for phosphorus recovery from hard waters.

Predictive control of crystal size distribution in protein crystallization

This work focuses on the modelling, simulation and control of a batch proteincrystallization process that is used to produce the crystals of tetragonal hen egg-white(HEW) lysozyme. First, a model is presented that describes the formation of proteincrystals via nucleation and growth. Existing experimental data are used to developempirical models of the nucleation and growth mechanisms of the tetragonal HEWlysozyme crystal. The developed growth and nucleation rate expressions are used within apopulation balance model to simulate the batch crystallization process. Then, modelreduction techniques are used to derive a reduced-order moments model for the purpose ofcontroller design. Online measurements of the solute concentration and reactortemperature are assumed to be available, and a Luenberger-type observer is usedto estimate the moments of the crystal size distribution based on the availablemeasurements. A predictive controller, which uses the available state estimates, is designedto achieve the objective of maximizing the volume-averaged crystal size whilerespecting constraints on the manipulated input variables (which reflect physicallimitations of control actuators) and on the process state variables (which reflectperformance considerations). Simulation results demonstrate that the proposed predictivecontroller is able to increase the volume-averaged crystal size by 30% and 8.5%compared to constant temperature control (CTC) and constant supersaturationcontrol (CSC) strategies, respectively, while reducing the number of fine crystalsproduced. Furthermore, a comparison of the crystal size distributions (CSDs)indicates that the product achieved by the proposed predictive control strategy haslarger total volume and lower polydispersity compared to the CTC and CSCstrategies. Finally, the robustness of the proposed method (achieved due to thepresence of feedback) with respect to plant-model mismatch is demonstrated. Theproposed method is demonstrated to successfully achieve the task of maximizingthe volume-averaged crystal size in the presence of plant-model mismatch, andis found to be robust in comparison to open-loop optimal control strategies.

Platinum-group element systematics in Mid-Oceanic Ridge basaltic glasses from the Pacific, Atlantic, and Indian Oceans

The concentrations of Ir, Ru, Pt and Pd have been determined in 29 Mid-Oceanic Ridge basaltic (MORB) glasses from the Pacific ( N = 7), the Atlantic ( N = 10) and the Indian ( N = 11) oceanic ridges and the Red Sea ( N = 1) spreading centers. The effect of sulfide segregation during magmatic differentiation has been discussed with sample suites deriving from parental melts produced by high (16%) and low (6%) degrees of partial melting, respectively. Both sample suites define positive and distinct covariation trends in platinum-group elements (PGE) vs. Ni binary plots. The high-degree melting suite displays, for a given Ni content, systematically higher PGE contents relative to the low-degree melting suite. The mass fraction of sulfide segregated during crystallization (X sulf ), the achievement of equilibrium between sulfide melt and silicate melts (R eff ), and the respective proportions between fractional and batch crystallization processes (S b ) are key parameters for modeling the PGE partitioning behavior during S-saturated MORB differentiation. Regardless of the model chosen, similar sulfide melt/silicate melt partition coefficients for Ir, Ru, Pt and Pd are needed to model the sulfide segregation process, in agreement with experimental data. When corrected for the effect of magmatic differentiation, the PGE data display coherent variations with partial melting degrees. Iridium, Ru and Pt are found to be compatible in nonsulfide minerals whereas the Pd behaves as a purely chalcophile element. The calculated partition coefficients between mantle sulfides and silicate melts (assuming a PGE concentration in the oceanic mantle at ∼0.007 × CI-chondritic abundances) increase from Pd (∼10 3) to Ir (∼10 5). This contrasting behavior of PGE during S-saturated magmatic differentiation and mantle melting processes can be accounted for by assuming that Monosufide Solid Solution (Mss) controls the PGE budget in MORB melting residues whereas MORB differentiation processes involve Cu-Ni-rich sulfide melt segregation.

A Review of the Use of Process Analytical Technology for the Understanding and Optimization of Production Batch Crystallization Processes

Crystallization at production scale is typically a poorly understood unit operation, with little implementation of the first principles aspect of crystallization in its design, optimization, and control. Problems with production crystallizers include the following: (1) inconsistencies of batch-to-batch in terms of the size and number of crystals produced and (2) the purity profile (residual impurities in crystals, or wrong polymorph or chiral purity). This can have a significant impact both on product quality and downstream process unit operations including filtration, drying, milling, and product formulation. This contribution reviews typical problems encountered in production crystallization, with case studies, advice, and strategies to understand and avoid these problems through the use of in situ crystallization characterization tools.

Determination of crystallization kinetics in solution

A mathematical model for the calculation of the linear crystal growth rate in solution crystallization was developed, using transient crystal size distribution (CSD) data, the solution concentration curve, and the suspension density of a batch crystallization process without taking in to the secondary processes. Although this method is developed based on a cooling crystallization system, it can be easily extended to other systems e.g., dilution crystallization, precipitation crystallization, etc. An advantage of this method is that crystal shape factors do not affect the linear crystal growth rate. The example of ciprofloxacin hydrochloride crystallization is employed to illustrate this method. The results from this study shows that the model gives more reliable kinetics than alternative methods and the convergence of the model can be easily achieved.

Application of In-Process X-ray Powder Diffraction for the Identification of Polymorphic Forms during Batch Crystallization Reactions

An on-line X-ray diffraction system capable of monitoring and identifying the polymorphic form of organic materials crystallized during batch crystallization processes is described. The overall system comprises a conventional X-ray source, monochromator, collimator, slits, curved position sensitive detector, and a flow-through analysis cell together with the electronics and the data acquisition software. The cell enables X-ray diffraction data to be recorded from crystals suspended within a slurry pumped to and from a batch crystallizer via an external temperature controlled sampling loop. The system is validated via benchmark crystallization studies of urea, l-glutamic acid and monosodium glutamate. The system was found to have a performance consistent with a lower detection limit of ca. 1% wt % of crystals and to be able to follow polymorphic transformation during representative crystallization processes. A forward look to the further development of the system is provided.

Nonlinear kinetic parameter estimation for batch cooling seeded crystallization

AbstractKinetic parameter estimation for most batch crystallization processes is necessary because nucleation and crystal growth kinetic parameters are often not available. The existing identification methods are generally based on simplified population balance models such as moment equations, which contain insufficient information on the crystal size distribution (CSD). To deal with these problems, a new optimization‐based identification approach for general batch cooling seeded crystallization is proposed in this study. The final‐time CSD is directly used for identification. A novel effective method for solving the population balance equation is developed and used to identify nucleation and growth kinetic parameters. Cooling crystallization of ammonium sulfate in water was experimentally investigated, where the concentration was measured by an on‐line density meter and the final‐time CSD was analyzed by a Malvern Mastersizer 2000. Kinetics for ammonium sulfate are determined based on cooling crystallization experiments. Applying these kinetics in simulation provides a good prediction of the product CSD. © 2004 American Institute of Chemical Engineers AIChE J, 50: 1786–1794, 2004

Quantitative Measurement of the Polymorphic Transformation of l-Glutamic Acid Using In-Situ Raman Spectroscopy

Control and optimization of the polymorph formation in batch crystallization processes require in-situ quantitative measurement of the concentration of different polymorphs. In this study, Raman spectroscopy was chosen to measure the polymorphic fraction of l-glutamic acid (l-Glu) in suspension. A sufficiently accurate calibration line was constructed using mechanical mixtures of the two forms, α and β. The calibration line was also valid for the suspension system. The polymorphic behavior in batch cooling crystallizations of l-Glu was monitored using the calibration correlation. It was clearly observed by Raman spectroscopy that at 25 °C the nucleation of the α-form from a clear aqueous solution occurred dominantly and nearly 100% α-form crystals were obtained. During the subsequent solvent-mediated transformation of the α-form crystals to the stable β-form, the α-form fraction gradually decreased. The transformation rate has a strong temperature dependency so that the temperature increase promotes the t...

DYNAMIC OPTIMAL CONTROL OF BATCH CRYSTALLIZATION PROCESSES

This work applies an on-line optimal control strategy developed by Zhang (2001) to two cooling batch crystallization processes. The algorithm initially finds the optimal crystallizer temperature and subsequently uses a feedback control system in order to achieve the desired final product quality of the crystals expressed in terms of the final crystal size distribution. In both batch processes, it is shown that the on-line optimal control approach provides better final product quality as compared with a simplified optimal cooling policy. The improvement is especially noticeable in the presence of plant/model mismatch or errors in the initial conditions.

Model-Based Optimization of Particle Size Distribution in Batch-Cooling Crystallization of Paracetamol

The control objectives of batch crystallization processes are often defined in terms of particle size distribution (PSD), or properties related to the PSD, viz. average particle size, product filterability, dry solids flow properties, etc. To achieve these control objectives, a constrained nonlinear model-based optimization strategy has been adopted. This involves the detailed modeling of batch crystallization including model validation and parameter estimation, on-line monitoring of supersaturation and PSD, and the application of optimization strategies. A deterministic population balance model accounting for solution thermodynamics, crystal growth, and nucleation has been developed. State estimation is achieved by the on-line monitoring of temperature, concentrations in the liquid phase, particle density, and PSD. For this purpose, the focused beam reflectance measurement (FBRM) provides an on-line, in-situ information of crystal size and particle concentration in the form of a chord length distribution...