Technologies for Optimisation and Control of Nucleation and Growth for New Generations of Industrial Crystallizers

(1)

Technologies for Optimisation and Control

of Nucleation and Growth for New

Generations of Industrial Crystallizers

(2)

(3)

Technologies for Optimisation and Control

of Nucleation and Growth for New

Generations of Industrial Crystallizers

Proefschrift

ter verkrijging van de graad van doctor

aan de Technische Universiteit Delft,

op gezag van de Rector Magnificus prof. ir. K.C.A.M. Luyben,

voorzitter van het College voor Promoties,

in het openbaar te verdedigen op maandag 24 november 2014 om 15:00 uur

door

Anamaria SOARE

Chemical Engineer, University Politehnica of Bucharest

geboren te Boekarest, Roemenië

(4)

Copromotor:

Dr.ir. H. J. M. Kramer

Samenstelling promotiecommissie:

Rector Magnificus, voorzitter

Prof. dr. ir. A. I. Stankiewicz, Technische Universiteit Delft, promotor Dr.ir. H. J. M. Kramer, Technische Universiteit Delft, copromotor Prof. dr. ir. G. Schembecker Technische Universität Dortmund

Prof. dr. ir. J. ter Horst University of Strathclyde Prof. dr. E. Vlieg Radboud University Nijmegen Prof. dr. R. F. Mudde Technische Universiteit Delft

Dr. R. Geertman DSM

ISBN 978-94-6259-398-5

Cover design by Anamaria Soare & Christy Renard

Front Cover Image: Represents the Andromeda spiral galaxy with explosive stars in its interior, and cooler, dusty stars forming in its many rings. The image is a combination of observations from the Herschel Space Observatory taken in infrared light (seen in orange hues), and the XMM-Newton telescope captured in X-rays (seen in blues).

Image Credit: ESA/Herschel/PACS/SPIRE/J. Fritz, U. Gent; X-ray: ESA/XMM Newton/EPIC/W. Pietsch, MPE

All rights reserved. No part of the material protected by this copyright notice may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording or by any information storage and retrieval system, without written prior permission from the author.

(5)

(6)

(7)

Chapter 1 Introduction

(10)

Crystallization is the operation by which one or more components are separated in form of crystals from a homogeneous system, solution or melt. It is one of the oldest and economically most important separation and purification technology in chemical industry1. The purpose of crystallization can be purification of a substance, to recover a component from a solution or to get a product into a marketable form, which requires meeting the specifications imposed on the shape, appearance, purity and size of crystals. Large quantities of crystalline substances are manufactured every year, like sodium chloride and sucrose, with production rates of over 100 million tons per year. Also the production of fine and functional chemicals and pharmaceutical compounds relies heavily on crystallization as the major separation and purification technique. Approximately 70% of the products sold by the process industry and the pharmaceutical industry, as bulk products, intermediates, fine chemicals, biochemicals, food additives and pharmaceutical products are solids.

The important position of crystallization as a separation and purification process is due to the strong separation capacity of crystallization, which in most cases allows for one stage purification determined by the high selectivity by which molecules, atoms and ions are built in a crystal lattice. The building blocks of a crystal, the constituent molecules or ions, are arranged in an ordered pattern extending in all three spatial dimensions, its form being well defined and specific for the compound. One of the effects of the internal organization of the crystalline structure is the anisotropic nature of the mechanical, electrical, magnetic and optical properties. Some compounds are capable of crystallizing in different crystalline but chemically identical2 forms. These substances exhibit polymorphism. Due to the different internal organization, different polymorphs will exhibit different physical properties. The different crystalline forms exhibited by one substance may result from a variation in the crystallization conditions, for example by changing the temperature, pressure or the type of solvent.

In order to assure that products meet their specifications, the control of the crystallization processes is very important. During crystallization many physical phenomena occur, as supersaturation generation, nucleation, growth, attrition, agglomeration and classification of crystals. All these phenomena complexly interact between them.

(11)

Introduction

1.1. Fundamentals of crystallization processes

This subchapter highlights the fundamental aspects of crystallization by relating the crystal product quality to the crystallization methods, equipment types, driving forces and physical processes that takes place during crystallization.

1.1.1. Supersaturation

A necessary condition for crystallization to occur is the generation of a driving force for the process. The driving force can be expressed in thermodynamic terms as the difference in chemical potential of the crystallization compound in the solution and that of the crystalline phase at the same temperature and pressure:

∆ ln ln (1.1)

where Δµ is the difference in chemical potential, R is the gas constant, T is the temperature, a is the actual solute activity, asat is the solute activity at saturation, c is the actual

concentration, csat is the saturation concentration, γ and γsat are the corresponding activity

coefficients to c and csat and S is the supersaturation.

At relative low concentrations the activity coefficients become unity and the supersaturation can be written in terms of concentration ratio as:

(1.2) or as relative supersaturation, which is defined as:

(1.3)

The saturation concentration depends on the system and is a function of temperature. Supersaturation is a prerequisite for nucleation and growth, which are decisive not only for the formation of a solid phase but also for the size distribution of crystals and their shape. The degree of supersaturation is determined both by the flows of materials and energies and by crystallization kinetics, such as nucleation and growth.

1.1.2. The solubility diagram

There are several ways to represent the phase diagram, showing the relation between temperature T, pressure P, concentration c and the crystal structure of solid phases. As the solubility of most of the compounds is only slightly dependent on the system pressure, solubility diagrams are mostly used in crystallization, as shown in Figure 1.1, in which the solubility of the compound is expressed as a function of temperature.

(12)

Figure 1.1. The solubility diagram

A solution whose concentration lies below the solubility curve is undersaturated and existing crystals will dissolve. A solution with a concentration above the solubility curve is called supersaturated, since the amount of dissolved solute is larger than the equilibrium value. At low supersaturations, although existing crystals will grow, it is difficult to create new crystals. In a supersaturated solution, solute molecules form dynamical clusters whose size and structure fluctuate with time due to attachment and detachment of molecules. The corresponding change in the overall excess free energy will depend on the excess free energy of the formation of an interface between the nucleus and the surrounding solution and on the volume excess free energy. The surface excess free energy is positive (destabilizing) and the volume excess free energy is negative (stabilizing). The classical theory of nucleation postulates that there is a critical nucleus size, with a maximum in free energy. Molecular clusters smaller than this critical size will dissolve and the bigger ones will grow, thus resulting in both cases in a decrease of the free energy of the clusters. The higher the supersaturation is, the smaller the critical nucleus size is. Once a certain level of supersaturation is exceeded, new crystals form spontaneously and the solution is called unstable. The region between equilibrium line and unstable zone is called the metastable zone, a region of limited supersaturation where no primary nucleation is observed within a reasonable time. The width of this zone depends on the solution purity, temperature and experimental conditions. The metastable zone forms the operation window in which primary nucleation can be minimal.

1.1.3. Crystallization methods and product quality

There are different methods to generate the supersaturation which leads to different crystallization methods. For melt crystallization, the supersaturation is generated by cooling the melt below its melting point. For crystallization from solution four different methods are used:

- Cooling - Decrease solubility by decreasing temperature

C o nc entr at io n, c Temperature, T Solubility curve Supersaturated region Undersaturated region Ev aporation Cooling

(13)

Introduction

- Evaporative - Increase concentration by evaporating solvent

- Precipitation - Mix two well-soluble reactants to give a poorly soluble product

- Anti-solvent - The solute is well-soluble in the original solvent, but is slightly soluble in the solvent/anti-solvent mixture1-3.

This crystallization method is mainly chosen on the basis of the thermodynamic and physical properties of the compound and the solvent, as well as on the required purity of the product and economic aspects. For compounds with high solubility and a large slope of the solubility – temperature curve cooling crystallization is usually chosen. For compounds with a small slope of the solubility – temperature curve evaporative crystallization is used in industrial practice. For soluble substances cooling and evaporating the solution in the same time could also be possible, by combining them by flash evaporation.

The requirements of the product properties rather than the method of creating supersaturation are decisive when selecting a crystallizer. The main product quality characteristics are:

- Crystal Form & Shape - The external appearance of a crystal is described in terms of morphology (the periodic structure of the crystal lattice) and habit (general shape of a crystal given by the relative length to width of the crystal faces). It is extremely difficult to predict the shape of crystals.

- Crystal Size - The crystal size and the width of the crystal size distribution. Crystal size influences performance characteristics such as filterability, settling behavior, caking behavior, bulk density and dissolution rates.

- Purity - Mechanisms affecting purity are inclusion of mother liquor in the lattice, entrapment of mother liquor in cracks, agglomerates, incorporation of impurities in the crystal lattice and adsorption of impurities or solvent on crystal surfaces. Impurities increase the caking tendency, may lead to further reactions and undesired chemical composition. The quality components of the product are strongly influenced by the geometry and type of the crystallizer, the operating conditions, and the properties of the liquid and solid phases.

1.1.4. Types of crystallizers

The most commonly used crystallizers in industrial practice are the agitated-tank crystallizers, forced-circulation crystallizers, draft-tube-baffle crystallizers and the fluidized-bed crystallizers.

The simplest crystallizers are agitated-tanks provided with a mixer for internal circulation. Usually they are used for cooling crystallization and the temperature decrease is obtained by heat exchange with an intermediate fluid circulating in the jacket of the tank. These simple tanks are normally used in batch processes, as in processing of pharmaceuticals and are prone to scaling. Batch processes normally provide a relatively variable product quality between different batches4.

(14)

The forced-circulation crystallizers are usually used for evaporative crystallization. A circulation pump keeps the crystal slurry in homogeneous suspension throughout the tank and assures reasonable velocities at the surface of the external heat exchanger. The fluidized-bed crystallizer, mentioned above, is a refining of the forced circulation crystallizer, equipped with a large crystals settling zone to increase the retention time and to roughly separate heavy slurry zones from clear liquid. The recirculated solution is freed of crystals by sedimentation so that only a substantially crystal-free solution passes the recirculation pomp. The disadvantage of this type of equipment is the difficulty in operation. Severe problems in operating behavior can be encountered, such as encrustations on the wall surfaces and irreversible collapse of the fluidized bed.

The draft-tube-baffle crystallizer basically consists of a crystallization body where crystals suspended in their mother liquor are agitated by an upward draft tube propeller, surrounded by an annular baffle from which a stream of mother liquor with fine crystals in suspension is extracted and sent to a heat exchanger in case of evaporative systems.

The above configurations are usually used in industrial crystallization, but several modification and combinations may be encountered as well.

1.1.5. Physical processes during crystallization

In order to have a good control over crystallization processes, ideally, one must be able to control the physical processes occurring in the crystallizer. Below a short overview of the most important physical processes occurring in crystallization processes is presented.

Primary nucleation

Primary nucleation is the formation of a new solid phase from a clear liquid. This type of nucleation can be further subdivided into homogeneous and heterogeneous nucleation. The heterogeneous nucleation starts on foreign substrates of microscopic particles, as dust or dirt particles, or on the surfaces of the equipment. If such substrates are absent, new phase formation takes place by statistical fluctuations of solute entities clustering together, a mechanism referred to as homogeneous primary nucleation. Homogeneous nucleation rarely occurs in practice, as the presence of small quantities of microscopic particles is usually unavoidable.

Crystal nucleation from a clear solution does not occur immediately as the solution becomes supersaturated. An energy barrier has to be passed in order for a stable nucleus to be formed. According to the classical nucleation theory, a nucleus is formed by a stochastic process of attachments and detachments of molecules in clusters. The probability of a cluster to grow to a stable nucleus increases with supersaturation. A foreign substance present in a supersaturated solution is generally known to reduce the energy required for nucleation. Nucleation in a heterogeneous system generally occurs at a lower supersaturation than a homogeneous system.

(15)

Introduction

Different alternative methods for inducing nucleation often involve mechanical shock, agitation, friction and extreme pressures2. The effects of external influences such as sonic and ultrasonic5 irradiations, electric and magnetic fields6, X-rays, γ-light, ultraviolet light have been studied in the past years, but the mechanism by which they induced nucleation is not yet understood, their effect cannot be always controlled or predicted and because of that none of these methods have significant application in industrial crystallization processes.

Secondary nucleation

Nucleation of new crystals induced by the presence of crystals of the material being crystallized is called secondary nucleation. These parent crystals have a catalyzing effect on the nucleation phenomena, and thus, nucleation occurs at a lower supersaturation than needed for primary nucleation. There are various mechanisms by which secondary nucleation can occur7-8: initial breeding9 (crystalline dust swept off a newly introduced seed crystal), polycrystalline breeding10 (the fragmentation of weak polycrystalline mass), dendritic1 (the

detachment of weak out-growths), fluid shear11-12_{(fragmentation of fragile crystals under the}

influence of a fluid shear) and contact13 (a complex process resulting from the interaction of crystals with one another or with parts of the vessel).

Contact nucleation has been found to be the most effective and common method for secondary nucleation. Contact nucleation, also referred to as attrition, occurs as a result of crystal-impeller, crystal-pump, crystal-vessel wall or crystal-crystal collisions. There are several factors that influence the secondary nucleation, as the supersaturation, the cooling rate, the agitation and the presence of impurities.

In industrial crystallizers, secondary nucleation is mainly influenced by the supersaturation level and mechanical stress of the crystals by the action of a stirrer or pump. Research activities in secondary nucleation aim at a better understanding of the mechanism, which would allow for the use of the nucleation rate data obtained in one crystallizer to predict the nucleation rate in another one with different dimensions, design and operation conditions. This extrapolation of the nucleation rate data can only be possible if a deeper understanding of the secondary nucleation phenomena is achieved. Most models which consider mechanical stress assume that nucleation is proportional to the energy transferred to the crystals during collisions14. This is not based on any physical relationship and, in addition, the models do not satisfactorily reproduce the experimental results. Recent studies aim at developing kinetic equations which can model the secondary nucleation rate function of the operating conditions and crystallizer geometry1, 4. However, these models usually include empirical correlations, which are unique to the system investigated. Thus, accurate prediction of the secondary nucleation is still difficult. Also the control of secondary nucleation in impeller mixed crystallizers is very difficult or even impossible.

Crystal growth

Crystal growth is the addition of solute molecules from a supersaturated solution to the crystal lattice. Besides increasing crystal size, crystal growth also largely determines crystal

(16)

morphology, surface structure and purity of the crystal. The growth of a crystal is often described by a linear growth rate which represents the change in a characteristic dimension of the crystal with time. However, crystals are made up of a number of faces that can grow at different rates. Therefore, a fundamental expression of the growth rate is the linear growth rate of a particular face. This refers to the rate of growth of the face in the direction normal to the face. When a linear growth rate is used to describe the growth of an entire crystal, it is describing the increase in some characteristic dimension of the crystal. If the crystal was a sphere, the characteristic dimension would be the diameter and we would express the crystal growth rate by the increase of the diameter with time. This characteristic dimension can be related to the volume and surface area of the crystal through the shape factor.

Crystal growth is a three-step process consisting of mass transfer, surface integration and heat transfer. Mass transfer and surface integration occur sequentially and in parallel with heat transfer. Mass transfer involves the diffusion of growth units (molecules, atoms or ions) to the crystal surface. Also the removal of impurities away from the surface can at times be important. Surface integration consists of surface diffusion, orientation and the actual incorporation into the lattice2_{. The relative importance of these two steps has a relation to}

solubility of the compound; in well soluble systems mass transfer is more important than surface integration, for poorly soluble crystals it is the other way around. Heat transfer is often a rate-limiting step in melt crystallization, but this is usually not the case in solution crystallization.

If the concentration falls below the saturation level, crystals will dissolve. This can happen due to dilution or temperature changes. The mechanism of crystal dissolution is not the exact opposite of growth. The release of solute molecules from the crystal lattice is generally speaking not a limiting step in the dissolution process. Therefore, the dissolution rate is typically governed by mass transport limitations only. Much faster dissolution rates compared to growth rates are measured in practice. The factors that influence the crystal growth are the same as the ones that influence nucleation, supersaturation, cooling rate, agitation and presence of impurities.

Mixing and suspension

Crystallization is usually carried out in a suspension and thus the study of crystallization requires knowledge of mixing. The effect of mixing on crystallization is mainly considered on two scales of mixing, the macro mixing and the micro mixing. The macro mixing refers to the residence time distribution, which defines retention times of the elementary volumes. The micro mixing describes communication between elementary volumes15_{. Usually the}

crystallizers are equipped with rotors, such as stirrers or pump impellers, which induce mixing.

During crystallization processes a high turbulent mixing16 is needed to obtain a good crystal suspension17 and reduce the degree of settling. However, such a high turbulence may cause entrainment of air from the headspace, evoke a high level of attrition and cause shear damage to the crystals18. Furthermore, vessel mixing conditions have been shown to affect

(17)

Introduction

the nucleation kinetics of the system19. Therefore an optimum flow conditions are needed in all crystallizers to avoid settling and maintain a uniform suspension, minimizing attrition and shear damage of the crystals. Similar to attrition, breakage can occur as a result of crystal impeller, crystal-pump, crystal-vessel wall or crystal-crystal collisions. The difference between breakage and attrition is not a distinct one. The fracture of a particle into one slightly smaller particle and many much smaller fragments is defined as attrition. Breakage involves the fracture of a particle into two or more pieces. The total fracture of a particle requires considerably more energy than attrition1.

There are different techniques available to analyze the mixing and flow patterns such as the classical pressure or velocity measurements with pitot tubes, venturi tubes or hot-wire anemometers or the novel ones like laser-induced fluorescence, Doppler velocimetry and particle image velocimetry, which will be further reviewed in the beginning of chapter 5.

Agglomeration

An agglomerate is defined as the enlargement of particles by the cementation of individual particles, probably by chemical forces20. Agglomerates are usually undesirable because they contain mother liquor between the primary crystals that form the agglomerate. This liquor is hard to remove during drying, and promotes caking of the product during storage.

Agglomeration is a multi-step process. The first step requires the collision of two or more crystals. This collision frequency depends on the particle concentration, the turbulence and the sizes of the crystals involved. In the next step, the colliding crystals must form an aggregate as a result of inter-particle forces. Finally, cementation of these crystals as a result of growth before the aggregate is disrupted is required to create a stable agglomerate. However, the mechanisms and rates of collision and disruption are related to the type of liquid flow that the mother crystals and the agglomerate experience, which in turn are dependent on their respective sizes. In particular, the influences of the absolute and relative sizes of mother particles, of the local energy dissipation and of the fluid viscosity differ according to the three types of motions, i.e. Brownian, laminar, turbulent. Besides this, the rapidity of the crystal growth, which in turn is a function of the supersaturation, plays a major role in the strengthening rate. In general the rate of agglomeration is function of the suspension density, crystal growth rate, particle size and shape, mechanical and fluid dynamics processes21_.

1.2. State-of-art in crystallization processes

Batch crystallization from solution is generally started by a decrease in temperature or an increase in concentration by evaporation, until primary nucleation starts and proceeds over a period of time which allows the nuclei to grow. Unseeded batch crystallization processes sometimes have reproducibility problems due to difficulties in controlling the primary nucleation in the initial phase leading to large variations in the product quality. Often primary

(18)

nucleation is avoided by seeding, using seeds produced by grinding crystals, by anti-solvent crystallization or by using part of the crystal product from a previous batch. Using grinded seeds, the CSD (crystal size distribution) of the product can be manipulated, but physical grinding processes require careful preparation procedures. The grinded seed have a tendency to dissolve caused by the internal lattice strain which means that a high seed load is required in order to preserve reproducibility of the product CSD. Using part of the product from previous batch has the advantages that no dissolution is observed even for very low seed load and no preprocessing is required4. However, these crystals are relatively large and have a small specific surface area which can result in buildup of a high supersaturation levels and thus to excessive nucleation during the batch. It is difficult to manipulate the CSD of these seeds and to control the CSD of the product. Another alternative is to produce seeds by anti-solvent crystallization. Anti-anti-solvent crystallization is a crystallization technique in which large supersaturation can be achieved by the mixing of solution and anti-solvent. This leads to high nucleation rates and thus small average crystal size. These seeds have the advantage that there is no dissolution observed even for very low seed loads. However at the anti-solvent feed point, high local supersaturation can be created which is difficult to control and leads to agglomeration of crystals. Another alternative for producing seeds is ultrasound irradiation of a solution, which is claimed to produce nuclei at low supersaturation, reducing the induction time, decreasing the mean crystal size and reducing agglomeration compared to conventional nucleation. However, the mechanism behind the ultrasound induced nucleation has not been understood and the nucleation rate cannot be always manipulated or predicted. The position of the sonotrode horn was shown to have an effect on the number of crystals, as was the use of a different ultrasound probe with identical power output. These findings show that the flow pattern and degree of mixing in the vessel play an important role on the number of crystals produced. The optimal sonotrode horn submergence depends on the sonotrode horn, vessel geometry and properties of the solution. The flow pattern in the vessel is strongly influenced by the position of the sonotrode horn and this has an effect on the nucleation rate observed. It is generally assumed that crystal nucleation is associated with cavitational collapses, but the mechanistic pathway by which cavitation and collapse lead to crystal nucleation has not been satisfactorily explained.

In the chemical industry severe problems are still frequently encountered during the design and operation of crystallization processes. These problems may be related to product quality requirements such as filterability, caking behavior, purity, form of polymorph or hydrates formation, and process requirements, such as production capacity and plant availability. A number of the above-mentioned problems are related to decisions taken during the design stage22. What is surprising, however, is that despite the important role that crystallization has in process control and in determining solid-phase outcomes, crystallization phenomena are often neglected in industry until a problem is encountered. During the design stage the crystallization equipment is usually selected from a limited number of existing types of industrial crystallizers, following an optimisation of the chosen equipment, which leads23, among other, to limited flexibility, limited controllability and process instability and dynamics.

(19)

Introduction

- Limited flexibility - Optimisation is limited due to the fact that the conventional equipment

lacks sufficient methods of actuation. Multiple physical phenomena as mixing, supersaturation generation, mass and heat transfer, primary and secondary crystal nucleation, crystal growth and dissolution take place simultaneously. This makes the optimisation with respect to individual phenomena practically impossible due to the strong entanglement between them4. Therefore there is limited flexibility to produce different grades of product.

- Limited controllability - The control of batch crystallization processes is a challenging task

due to their highly nonlinear behavior, irreproducible start-up, lack of reliable measurements for the system state and most important due to the limited possibilities for actuation24. The options for actuation are limited to what is offered by the equipment. There is limited control over physical phenomena and for example, a change in temperature affects the supersaturation which influences in the same time the nucleation, growth, agglomeration, impurity inclusions etc.

- Process instability and dynamics - The complexity and the interrelation of all the physical

phenomena results in highly nonlinear process behavior, which might give rise to long start up times due to oscillations. This also reduces the window of available combinations of operational parameters for stable operation. As a result of these instabilities and response dynamics, a significant amount of off-spec product can arise. In practice, the product may not meet the target values of the quality characteristics because the process design simply cannot deliver the process characteristic at the desired level. Also, in literature there are many cases of unwanted or previously unknown nucleation events. Dunitz and Bernstein25_presented

several cases where it was difficult to obtain a given polymorphic form. These cases provide evidence for the consequences of poor process control and process instabilities in crystallization of polymorphic systems.

In order to improve the design of crystallization processes a shift in thinking is necessary. Process intensification is employed in order to make significant reduction in the size of chemical plants, amount of off-spec products and the utilization of resources. To facilitate such principles, a new design approach, which considers the important phenomena as starting point for design rather than the equipment itself, was recently adopted. The hierarchical design procedure proposed by Bermingham26 and Westhoff27 has been extended by Menon28 and Lackerveld29_{by adopting a task-based design approach. The task-based design approach}

is based on the assumption that the process unit under design may be considered to be a collection of physical tasks.

The physical phenomena are of key importance as in the end they determine the properties of the crystalline product. This fact has been the starting point of a task-based design strategy for industrial crystallization which is a novel approach for the conceptual design of crystallization process units. Task based design uses physical phenomena to construct tasks, which are used as building blocks for design. The order of task-based design proposed by Menon28 is as follows:

(20)

2. setting up a sequence of tasks going from initial to final state; 3. identifying the proper physical phenomena to perform such tasks;

4. establishing the internal rate processes to enhance the tasks and grouping these in operational units by selecting compatible operating conditions and space-time requirements; 5. finding spatial arrangements for selected, feasible combinations of tasks in equipment. The feasibility of the task based design approach and the potential gain in flexibility and controllability in operation was shown by Lakerveld29. However, for crystallization processes it is difficult to develop specific physical devices that will execute primarily a single crystallization task, as current crystallizers facilitate many of those tasks and the control of individual tasks is not possible.

1.3. Scope, objectives and outline of the thesis

The scope of this thesis is to study new methods of isolation and actuation of the most important phenomena in crystallization processes: nucleation and crystal growth. In order to achieve this, novel process units should be developed. The process units should utilize new methods with alternative driving forces that specifically target a certain task in an attempt to disentangle the physical phenomena, which makes individual optimisation of driving forces possible.

The specific research questions addressed in this thesis are:

- Is it possible to adequately isolate, control and predict crystal nucleation using an alternative driving force?

- Is it possible to understand the mechanism of crystal nucleation induced by the chosen driving force?

- Can crystal growth be isolated and well controlled?

Small scale experiments were done with newly built dedicated equipment that made it possible to isolate and optimise single crystallization tasks. These are isolation of the task primary nucleation using laser irradiation and the task crystal growth by using an airlift crystallizer.

In Chapter 2, the feasibility to control primary nucleation for solution crystallization was investigated using laser irradiation. The relationship between creation, expansion and collapse of a vapor cavity induced by a 6 ns laser pulse and the subsequent nucleation of crystals was investigated. The number of crystals formed per laser pulse was reproducible for each event.

Chapter 3 focuses on the theoretical study of the enhancement of crystal nucleation generated by cavities induced by laser irradiation, as the mechanism behind it was still poorly

(21)

Introduction

understood. A mechanism of crystal nucleation induced by laser cavitation was proposed and the bubble dynamics and the crystal formation were modeled.

Chapter 4 shows that crystal growth can be isolated in an airlift crystallizer, which is a strong alternative to conventional crystallizers especially for systems in which attrition has to be minimized. For optimized process conditions both primary nucleation and attrition can be avoided.

In Chapter 5 the hydrodynamic and the growth behaviors in the airlift crystallizer were studied, varying the airflow rates, sparger types and seeds loads. The temperature gradients and thus the supersaturation gradients were also studied for different cooling rates. It was proved that the crystallizer is approaching ideal mixing with very small temperature gradients for relatively low airflow rates even for high cooling rates.

The major contributions of this thesis along with suggestions regarding key directions for the future are summarized in chapter 6.

References

1. Mersmann, A., Crystallisation Technology Handbook. 1995; p 215-325.

2. Mullin, J. W., Crystallization, 4th Ed. Butterworth-Heinemann: Oxford 2001.

3. Jancic, S. J.; Grootscholten, P. A. M., Industrial Crystallization. Delft University Press ;: Delft,

1984.

4. Kalbasenka, A. N. Model-Based Control of Industrial Batch Crystallizer Experiments on Enhanced Controllability by Seeding. TU Delft, Delft, 2009.

5. Virone, C.; Kramer, H. J. M.; van Rosmalen, G. M.; Stoop, A. H.; Bakker, T. W., Primary Nucleation Induced by Ultrasonic Cavitation. Journal of Crystal Growth 2006, 294 (1), 9-15.

6. Evans, G. J., Influence of External Fields on Nucleation and Crystal Growth. Crystal Growth of N-Octylbiphenyl from Solution in the Presence of Magnetic and Electromagnetic Fields. Journal of

the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases 1985, 81

(3), 673-678.

7. Strickland-Constable, R. F., Kinetics and Mechanism of Crystallization from the Fluid Phase and of the Condensation and Evaporation of Liquids Academic Press: London, New York 1968; p

347.

8. Nyvlt, J., Nucleation and Growth-Rate in Mass Crystallization. Progress in Crystal Growth and

Characterization of Materials 1984, 9 (3-4), 335-370.

9. Tsokanis, E. A.; Thompson, R. W., Further Investigations of Nucleation by Initial Breeding in the Ai-Free Nh4-Zsm-5 System. Zeolites 1992, 12 (4), 369-373.

10. van der Heijden, A. E. D. M.; van der Eerden, J. P.; van Rosmalen, G. M., The Secondary Nucleation Rate: A Physical Model. Chemical Engineering Science 1994, 49 (18), 3103-3113.

11. Kubota, N.; Kubota, K., Secondary Nucleation of Magnesium-Sulfate from Single Seed Crystal by Fluid Shear in Agitated Supersaturated Aqueous-Solution. Journal of Crystal Growth 1986, 76 (1), 69-74.

12. Sung, C. Y.; Estrin, J.; Youngqui.Gr, Secondary Nucleation of Magnesium Sulfate by Fluid Shear. Aiche Journal 1973, 19 (5), 957-962.

(22)

13. Youngqui.Gr; Randolph, A. D., Secondary Nucleation in a Class-Ii System - Ammonium Sulfate-Water. Aiche Journal 1972, 18 (2), 521-&.

14. Ploß, R.; Mersmann, A., A New Model of the Effect of Stirring Intensity on the Rate of Secondary Nucleation. Chemical Engineering & Technology 1989, 12 (1), 137-146.

15. Sha, Z.; Palosaari, S., Mixing and Crystallization in Suspensions. Chemical Engineering Science

2000, 55 (10), 1797-1806.

16. Barrett, M.; O'Grady, D.; Casey, E.; Glennon, B., The Role of Meso-Mixing in Anti-Solvent Crystallization Processes. Chemical Engineering Science 2011, 66 (12), 2523-2534.

17. Ochieng, A.; Lewis, A. E., Cfd Simulation of Solids Off-Bottom Suspension and Cloud Height.

Hydrometallurgy 2006, 82 (1-2), 1-12.

18. Kougoulos, E.; Jones, A. G.; Wood-Kaczmar, M., Cfd Modelling of Mixing and Heat Transfer in Batch Cooling Crystallizers Aiding the Development of a Hybrid Predictive Compartmental Model.

Chemical Engineering Research and Design 2005, 83 (1 A), 30-39.

19. O'Grady, D.; Barrett, M.; Casey, E.; Glennon, B., The Effect of Mixing on the Metastable Zone Width and Nucleation Kinetics in the Anti-Solvent Crystallization of Benzoic Acid. Chemical

Engineering Research and Design 2007, 85 (7 A), 945-952.

20. Randolph, A. D., Larson M. A., Theory of Particulate Processes: Analysis and Techniques of

Continuous Crystallization. 2nd ed.; Academic Press: 1988; p 369.

21. David, R. E., F.; Cameirao, A.; Rouleau L. , Developments in the Understanding and Modeling of the Agglomeration of Suspended Crystals in Crystallization from Solutions. KONA 2003, (21), 40 - 46.

22. Lakerveld, R. Development of a Task-Based Design Approach for Solution Crystallization Processes. Dissertation, Delft University of Technology, Delft, 2010.

23. Lakerveld, R.; Kramer, H. J. M.; Stankiewicz, A. I.; Grievink, J., Application of Generic Principles of Process Intensification to Solution Crystallization Enabled by a Task-Based Design Approach. Chemical Engineering and Processing: Process Intensification 2010, 49 (9), 979-991. 24. Mesbah, A.; Kalbasenka, A.; Huesman, A.; Kramer, H.; Van den Hof, P. M. J. In Real-Time

Dynamic Optimization of Batch Crystallization Processes, 17th IFAC World Congress, Korea,

Chung, Ed. Korea, 2008; pp 3246-3251.

25. Dunitz, J. D.; Bernstein, J., Disappearing Polymorphs. Accounts of Chemical Research 1995, 28 (4), 193-200.

26. Bermingham, S. K. A Design Procedure and Predictive Models for Solution Crystallization Processess Development and Application. Delft University of Technology, Delft, The Netherlands,

2003.

27. Westhoff, G. M. Design and Analysis of Suspension Crystallizers, Aspect of Crystallization Kinetics and Product Quality. Delft, 2002.

28. Menon, A. R.; Pande, A. A.; Kramer, H. J. M.; Jansens, P. J.; Grievink, J., A Task-Based Synthesis Approach toward the Design of Industrial Crystallization Process Units. Industrial &

Engineering Chemistry Research 2007, 46 (12), 3979-3996.

29. Lakerveld, R.; Kramer, H. J. M.; Jansens, P. J.; Grievink, J., The Application of a Task-Based Concept for the Design of Innovative Industrial Crystallizers. Computers & Chemical Engineering

(23)

Chapter 2 Crystal Nucleation by Laser

Induced Cavitation

This chapter is published as:

Soare, A.; Dijkink, R.; Pascual, M. R.; Sun, C.; Cains, P. W.; Lohse, D.; Stankiewicz, A. I.; Kramer, H. J. M., Crystal Nucleation by Laser-Induced Cavitation, Cryst. Growth Des., 2011, 11 (6), 2311-2316.

(24)

High-speed and high-resolution photography have been used to investigate the relationship between creation, expansion and collapse of a vapor cavity induced by a 6 ns laser pulse and the subsequent nucleation of crystals. A thin layer of supersaturated aqueous solutions of (NH4)2SO4 and KMnO4 was confined between two glass plates with a separation of 50 and

100 µm. The expansion and collapse of the laser-induced vapor bubble occurred over a total timescale of 200 µs, while the crystals are detected one second after the laser pulse. Crystals were observed to form on a ring with a diameter of 70 µm centred in the focal point of the laser. The ring is preceded by an optical disturbance observed through the cavity around 30 - 50 µs after the laser pulse and vapor cavity formation. This ring shaped optical disturbance originates from changes in refractive index induced by crystal nuclei formation. The formation of the nuclei most probably coincides with the formation of the bubble, when the rate of evaporation and the supersaturation are at their maxima. After 30 – 50 µs the bubble interface velocity is relatively low and the optical disturbance generated by the particles become clearly visible.

(25)

Crystal Nucleation by Laser Induced Cavitation

2.1. Introduction

The effectiveness and reproducibility of crystallization is often limited by crystal nucleation1, a stochastic process that is generally not well understood and that can often show considerable variability even under constant and well-controlled conditions2. A number of expedients can be employed to overcome these difficulties, the most common of which are seeding and the application of ultrasound energy3_{. Ultrasound in the so-called ‘power’}

frequency range 20 – 100 kHz has been shown to induce nucleation in a wide range of crystallization processes4, but its effectiveness is by no means universal. Most of the observed physical and chemical effects of power ultrasound have been attributed to sonically induced cavitation (bubble or void formation)5, and in particular to the collapse of cavitational voids during the compression phase of the sonic cycle6_{. It is generally assumed}

that crystal nucleation is associated with such cavitational collapses, but the mechanistic pathway by which cavitation and collapse lead to crystal nucleation has not been satisfactorily explained. An alternative explanation of ultrasonically induced nucleation7 assumes that the bubble surface created in cavitation acts as a foreign particle, so that nucleation is heterogeneous. However, results of experiments in which gas bubbles are introduced in place of ultrasound show that this nucleation is not sensitive to the application time of the gas.

Experimentally, it is very difficult to locate and monitor a single cavitation event in an ultrasonic field8. Instead, we have examined in detail the relationship between the growth / collapse of a cavity and crystal nucleation using a focused laser pulse to induce a single cavity in a supersaturated solution at a specific, pre-determined location. However, we note that for laser induced cavitation the bubble dynamics9 and the subsequent physical effects differ from those for ultrasonic cavitation. Whereas in the latter case the bubble mainly consists of gas, laser induced bubbles mainly consist of vapor, and growth and shrinkage are controlled by evaporation, condensation and thermal diffusion which is much faster than mass diffusion, so that is inappropriate to speak of some equilibrium bubble size. We will discuss these differences between gas and vapor bubbles in more detail below.

Laser irradiation has been reported to induce crystal nucleation on its own by a number of mechanisms, some of which involve cavitation10 and some of which do not11. The crystallization of anthracene from solution in cyclohexane was triggered by a single femtosecond laser pulse of energy above the threshold for cavity formation (3.1 µJ), and occurred at the bubble surface10. A similar femtosecond pulse was used to induce the nucleation of hen egg-white lysozyme (HEWL)12, but here nucleation was only observed 24 – 48 h after irradiation and it is not known whether cavitation occurred. In so-called ‘nonphotochemical’ laser-induced nucleation or NPLIN, an expanded laser beam is employed11, and presumably cavitation is avoided. Here, the induction of nucleation is ascribed to the optical Kerr effect, in which the polarized electromagnetic vectors of the beam align polar functions within the crystallizing molecule and thus induce ordering in the nucleating clusters. To support this hypothesis, a switch in the polymorphs of glycine that crystallize from solution was observed depending on whether the beam is linearly or

(26)

nucleation in supersaturated solutions of KCl according to this mechanism14, and this method was also used to induce spatially-controlled crystallization of KCl in an agarose gel15. It has also been shown that ice nucleation in supercooled water can be initiated by the optical breakdown induced by a focused laser pulse of 1064 nm wavelength16_{. Even if the detailed}

mechanism is not proven yet it is argued that homogeneous nucleation in the compressed liquid phase is a plausible explanation of the effect. The laser generated bubbles emit pressure waves, which could trigger the actual crystal nucleation in the compressed liquid around the bubble.

Here we report the direct observation of nuclei formation around of a single forming, expanding and collapsing cavity, created by a focused laser pulse. By using a thin layer of supersaturated solution between two glass slides, it was possible to obtain well-focused 2-D digital images of the bubble growth, collapse and the resulting nucleation.

2.2. Experimental

Experiments were carried out using aqueous solutions of two simple inorganic salts – (NH4)2SO4 and KMnO4. Attempts to extend the range of case studies to organic solutes and

solvents (methanol, ethanol) were hampered by the solvent volatility and possible thermal degradation. With (NH4)2SO4 solutions, the operating temperature was 22.5 oC and relative

supersaturation levels of 0.2 and 0.4 % were employed; with KMnO4 the operating

temperature was 29 o_{C and the relative supersaturations were 7, 14 and 21 % (see results}

section). Magenta ink was added to the (NH4)2SO4 solutions to facilitate the absorption of the

laser energy in the solution, and hence the creation of the cavitation bubble.

Figure 2.1. Experimental setup for observing laser-induced cavitation and nucleation

The liquid was placed in between two glass plates with a 50 µm or 100 µm gap, see Fig. 2.1. A Nd-YAG 6 ns laser pulse (wavelength 532 nm) of energy 0.05 – 0.5 mJ (Solo PIV, New Wave, Fremont, CA, USA) was then focused by a 20x objective lens to the centre of the liquid to create a cavitation bubble. The bubble grows explosively, rapidly exceeding the distance between the glass plates, thus appearing two dimensional. A single frame high-sensitive camera (PCO) was also employed, in combination with a 1 µs light pulse from a high intensity LED (Seoul P7), to image the bubble and the crystals that subsequently

(27)

formed. A digital delay generator (Model 555, Berkeley Nucleonics Corp., CA, USA) was used to synchronize these cameras and the laser. The motion of the bubble was recorded by a high-speed camera (HPV-1, Shimadzu Corp., Japan). The time lapse between pictures was 4 μs. Illumination for the camera was provided by a fiber lamp (Olympus ILP-1) emitting white light, which was redirected by a fiber optic arm, passing through the dichroitic mirror filter to the camera.

2.3. Results

Fig. 2.2 shows the formation, growth and collapse of the cavitation bubble for a typical experiment with (NH4)2SO4 solution. In (a) at 4 µs after the laser pulse, the vapor bubble is

already formed at the focal point of the pulse. The bubble is initially a small sphere that expands spherically until it occupies almost the entire gap between the two slides (a), after which preferential growth in the radial direction begins and the shape of the bubble changes to that of a cylindrical slice (b), (c) and (d). Two very thin liquid films most probably remained on the walls, indicating wetting of the hydrophilic surface. Then the bubble collapse (e), (f) and (g), leaving small gas bubbles behind (black circles in (g) and (h)). These are non-condensable air bubbles. In the order of minutes later crystals were observed as seen in (h).

Figure 2.2. Evolution of cavitational bubble with time, (NH4)2SO4 solution, 0.4%

supersaturation, 100 µm gap size. After the laser pulse the bubble forms rapidly by evaporation, grows and collapses, over a time scale of 200 µs, leaving behind crystals which in time grow to sizes of tens of µm.

Fig. 2.3(a) shows the evolution of the bubble radius during formation and collapse, taken from analysis of the photographic images. The bubble expands quickly, and shrinks more slowly after reaching its maximal size, which means that the (absolute) velocity of the interface, Fig. 2.3(b), is greater during the expansion. The absorbed laser pulse superheats the liquid which finally evaporates explosively, leading to a shock wave emission. The volume of

500 μm

(a)

(b)

(c)

(d)

(e)

(f)

(g)

4 μs 24 μs 48 μs 84 μs 184 μs 204 μs 308 μs 3 min

(h)

Vapor bubble Gas bubbles Bubble

colapse _Gas Crystals

bubbles Gas

(28)

the vapour bubble then expands with a velocity smaller than the shock wave. This interface velocity is given in Fig. 2.3(b).

Figure 2.3. Evolution of (a) the bubble radius and (b) the bubble interface velocity with time. Experimental conditions as in Fig. 2.2. The focused laser beam width is 20 µm

Images taken with the high-resolution camera at key points in the sequence of cavity evolution and crystal nucleation in Fig. 2.2 are shown in Fig. 2.4.

Figure 2.4. High-resolution images during cavity formation and nucleation of (NH4)2SO4

solution, 0.2% supersaturation, 50 µm gap size

Fig. 2.4(a) shows the clear solution prior to firing the laser; the small round black dots are air

0 100 200 300 400 500 600 700 800 0 100 200 Bu bb le radi us, μ m Time, μs ‐30 ‐20 ‐10 0 10 20 30 40 50 60 0 100 200 Veloc ity of th e bub ble in te rf ac e, m/s Time, μs

63 µm

- 1 s

50 μs

1 s

30 s

Optical

disturbance

ring

Crystals ring

(a)

(b)

(c)

(d)

Crystals ring

(a) (b)

(29)

2.2(c), ~50 µs. Close examination reveals a thin ring-shaped optical disturbance, indicated by the arrow, surrounding the focal spot of the laser and visible through the bubble. One second after the collapse of the bubble, Fig. 2.4(c), a ring of small crystals may be seen in exactly the same place as the disturbance in Fig. 2.4(b). After further 30 s the crystals can be seen to have grown, Fig. 2.4(d). A similar experiment is depicted in the annex 2-A.

Fig. 2.5 shows a zoom on the area close to the focal point of the laser. The disturbance through the bubble can clearly be seen in Fig. 2.5(a). The diameter of the ring is about 70 µm. The subsequent ring of crystals in the same place in Fig. 2.5(b) has the same diameter. In the centre of Fig. 2.5(b), a minute crack in the lower glass slide can be seen. This crack was caused by the thermal shock arising from the absorption of laser energy by the glass. Even with the addition of magenta ink, it was difficult to obtain a cavitation bubble without creating such small cracks in the support slide.

Figure 2.5. Zoom view of (a) the optical disturbance around the laser focal spot (as Fig. 2.4(b)), (b) the subsequent ring of crystal nuclei. A minute crack may be seen in the lower glass slide, shown by the arrow. The scale bar in a) also refers to part b) which is shown at the same scale).

Two questions arise: Why are the crystals formed in a ring structure and what physical process determines the final radius of the ring? We did not observe any clear relation between the absorbed laser energy and the radius of the ring of crystals. The crystals are most probably formed at the liquid-vapor interface immediately after the bubble formation, and are then – thanks to inertia – shifted outwards with the expanding vapor-liquid interface. Crystals grow, accumulate mass and inertia so that their movement more and more decouples from that of the less dense liquid. Upon collapse the crystals do not follow the vapor liquid interface any more (similar to gas-liquid interface-driven particles17_{) but remain at some}

radius.

Further experiments with KMnO4 solution were carried out to ensure that we were observing

nucleation caused by laser irradiation, and not secondary nucleation due to small shards of glass detached by cracking or to foreign ions possibly formed by photochemically

(a)

_(b)

Crack

20 µm

(30)

degradation of the ink. The intense coloration of this solution and the consequent high absorption at the laser frequency ensured that no damage to the glass slides occurred, as we established by careful examination. The photographic images show the same sequence of events as those presented above but they are generally less clear because of the coloration of the solution. Fig. 2.6 shows a sequence of cavitation and nucleation events with KMnO4 that

are similar to those recorded above with (NH4)2SO4.

Fig. 2.6. KMnO4 nucleation and growth, 14% initial supersaturation, 50 µm gap size, 0.06 mJ

energy input

In these experiments no ink was added, there were no problems in obtaining cavitation bubbles without any cracking in the glass slides, and the crystals were again observed one second after the laser pulse. However, with KMnO4 solutions, crystals were obtained after

cavity formation only in solutions where the initial supersaturation was higher than 7%. For solutions with supersaturation of 7% or lower, crystals were not obtained after a single laser pulse, but could be obtained after several sequential pulses. The number of pulses required in these cases was between 5 and 20, with 1s time lapses between shots. If the repetition rate of the laser pulses increased 2 - 3 times the crystals appeared after fewer pulses; if the repetition rate decreased 2 - 3 times, more pulses were required to create crystals. This is believed to have occurred because the evaporation arising from a single pulse was insufficient to create nucleation, given that a corresponding local temperature rise will decrease the supersaturation. After multiple pulses, eventually the evaporation will be sufficient for crystal nucleation to take place. Each pulse induces cavitation, and after each collapses the local temperature of the liquid increases, thus less energy is used for heating the liquid and more energy is used for evaporation. With (NH4)2SO4, because the temperature - solubility curve is

very flat, the effect of the small amount of water evaporated by the pulse will not be compensated to the same extent by a solubility increase that results from a rise in temperature.

63 µm

- 1 s

50 μs

_{1 s}

30 s

(a)

(b)

(c)

(d)

(e)

(f)

7 s

1 min

Crystals

(31)

With (NH4)2SO4, 10 - 30 crystals were formed per laser pulse and with KMnO4 only 1 - 5

crystals were obtained for each event. This may also be due to the balancing effects of evaporation and temperature increase as described above, but it may also reflect the narrower metastable zone of (NH4)2SO4.

At very high supersaturation (21%), very small particles formed at the glass surface immediately after the solution was introduced into the gap and prior to laser pulsing. Without laser pulses these grew very slowly, with no increase in size after 15 min. However, these particles differed in appearance and in colour to those of KMnO4, and they were difficult to

remove from the glass after the experiment. We believe they were MnO2 deposits arising

from decomposition rather than crystals of KMnO4 and that they don’t affect the

crystallization of KMnO4. A few large KMnO4 crystals were observed around 1s after a

single pulse, and more appeared after subsequent pulses. In this case, around 3 to 7 more KMnO4 crystals were obtained per laser pulse. This increase in the number of crystals is a

consequence of the higher initial supersaturation of the solution.

2.4. Discussions

We have shown by direct observation that the creation of a laser-induced cavity in a supersaturated solution results in the nucleation of crystals in a ring around the laser focus, where the creation and collapse of the cavity occurs. In the cases investigated, the growth and collapse of the cavity takes place over timescales of ~200 µs, while the detection of visible and identifiable crystals requires in the order of one second, although observable optical disturbances were already present after 30 – 50 µs. The formation of nuclei occurs on a ring centred on the centre of the cavity, in the same position as the optical disturbance occurs during the development of the cavity. Fig. 2.7 shows a time line diagram representing bubble and nuclei evolution.

Fig. 2.7. Time line diagram for crystal nucleation by laser-induced cavitation experiments. The bubble and crystals nucleation are in a dashed rectangle because it is not known when during the first 6 ns they actually take place.

Essentially the same results were observed with (NH4)2SO4 and KMnO4 solutions, but with

(NH4)2SO4 the nucleation effects are much more clearly exhibited. With KMnO4, higher

supersaturations were needed for nucleation to occur, a smaller number of nuclei were created, and in some cases multiple laser pulses were required to create nuclei. The metastable zone width for the crystallization of KMnO4 from aqueous solution was not

(32)

coloration, but the measurements did show that the zone was much wider than that for (NH4)2SO4 solutions18. This can explain why higher supersaturation levels had to be

employed; no nucleation could be induced below 7% supersaturation.

Recent investigations of the formation and collapse of a vapor cavity under laser irradiation in a microtube filled with liquid (water with an added red dye)9,19 reported similar vapour bubble dynamics to those in Figs. 2.2 and 2.3 of this work, namely rapid expansion and somewhat slower contraction and collapse, and evidence of surface wetting as the cavity expanded. In those papers the expansion and contraction of the cavities were successfully modelled, taking both hydrodynamic and thermal effects into account. The calculations indicated that temperature transients of 180 oC occurred on impulse at the start of cavity formation, falling rapidly to 70 – 80 o_{C at the point of maximum expansion and to 50 – 60}o_C

at the point of collapse.

In most cases including both (NH4)2SO4 and KMnO4, temperature transients are unlikely to

result directly in crystal nucleation because solubility increases with temperature and the supersaturation driving force to nucleate will consequently reduce. Similarly, also in ultrasonically induced nucleation thermal effects in the liquid are limited due to the large thermal capacity of water20. For this reason, alternative explanations of ultrasonically induced nucleation have been sought in terms of the pressure transient or shock wave that accompanies collapse21, but these have not yet been examined rigorously.

The position at which crystals are eventually detected is directly related with the optical disturbance observed through the vapor bubble around 30 - 50 µs after the laser pulse, Fig. 2.5. This disturbance has not been characterized completely, but firmly suggests changes in refractive index induced by nuclei formation at the point of maximum rate of vaporization at the start of bubble formation during the laser pulse. The high resolution camera, that captured the optical disturbances, was used in combination with a 1 μs light pulse from a high intensity LED. In the first 30-50 μs the velocity of the bubble interface and of the liquid film around it were very high, 8 to 50 m/s, as can be seen in Fig. 2.3b. This means a movement of 8 to 50 μm in the 1 μs time when the picture was captured. In this condition the optical disturbances given by the movement will cover the optical disturbances given by the presence of the crystals. In Fig. 2.3b it can be seen that for the period between 50 and 150 μs the bubble interface velocity is close to 0 m/s and in the images captured in this period the optical disturbances are clearly detected.

Experiments in a 3D cuvette were also performed, but those experiments didn’t bring new insights regarding the mechanism of cavitation induced crystal nucleation, as the crystals were observed only at the end of the experiments at the bottom of the cuvette.

This investigation was undertaken primarily to understand the mechanism of ultrasonic nucleation, by relating cavitational events and collapse to the phenomenon of crystal nucleation. While there are parallels in the characteristics and behaviour of laser- and ultrasonically-induced cavitation, the sequence of events differs. With laser induction, a shock wave accompanies the formation and growth of the cavity, driven by vaporization and

(33)

expansion at the point of pulse impact. The collapse phase is relatively slow, Fig. 2.3, and the rate of the energy released in the collapse phase is low. With ultrasonic cavitation, the cavity forms more slowly due to tensile stress on the liquid in the expansion phase, and energy release occurs as it becomes unstable and collapses under the pressure force of the compression phase of the sonic wave6. The shock wave is emitted on collapse6, 22, 23 and thermal effects in the liquid do not play an important role, due to the large thermal heat capacity of water20. We note that even if a transient temperature rise existed, the crystal nucleation still could take place at some distance of the imploding cavity. A time delay between the original excitation and nuclei formation would therefore occur with ultrasound.

2.5. Conclusions

Using high-speed photography, the formation of crystal nuclei in supersaturated solutions of (NH4)2SO4 and KMnO4 was observed in the wake of a single forming, expanding and

collapsing cavity created by irradiation with a focused laser pulse. The formation and subsequent collapse of the bubble takes about 200 µs, while the detection of small crystals in the vicinity of the cavity occurs on a timescale of seconds following the pulse, although already after 30 – 50 µs small optical disturbances were observed at the same locations where the crystals appeared in the later stage.

The crystals nucleated in these experiments appeared on a ring of diameter ~70 µm that was centred at the point of impact of the laser pulse. At its largest extent, the diameter of the disk-shaped cavity was 1400 µm. At the point where the cavity approaches its maximum diameter, 50 µs after the pulse, optical disturbances are detected around the centre, and the position of the disturbance corresponds to the ring on which crystals are later seen. The disturbance suggests differences in refractive index and is directly related with the stage at which crystal nucleation takes place.

The course of the cavity dynamics observed corresponds closely with the formation and collapse of vapor cavities in microtubes under similar conditions of laser irradiation. Modelling studies9,19_{for the latter indicate that thermal effects are important in determining}

the cavity dynamics, with the implication that temperature transients up to 180 oC occur on impulse at the start of cavity formation. These transients set up the vaporization that leads to the cavity, and heat dissipation determines the dynamics of subsequent collapse. The nature and sequence of these events differ from the dynamics associated with the formation and collapse of ultrasonically-induced cavities.

(34)

Appendix 2-A

Another example of laser induced crystal nucleation experiment

Figure 2.8. High-resolution images during cavity formation and nucleation of (NH4)2SO4

solution, 0.2% supersaturation, 50 µm gap size

References

1. Mullin, J. W. Crystallization 4th ed. 2001. Butterworth-Heinemann.

2. Kashchiev, D. Nucleation: Basic Theory with Applications 2000, Butterworth-Heinemann,

Oxford; Kashchiev, D.; van Rosmalen, G. M. Cryst. Res. Technol. 2003, 38(7 – 8), 555 – 74.

3. Ruecroft, G.; Hipkiss, D.; Ly, T.; Maxted, N.; Cains, P. W. Org. Proc. Res. Dev. 2005, 9(6), 923 – 32.

4. McCausland, L. M.; Cains, P. W. Biotech. & Genet. Engg. Revs. 2004, 21, 3 -10; Chem. &

Ind. 2003, (5), 15 – 18; McCausland, L. M.; Cains, P. W.; Martin, P. D. Chem. Eng. Progr. 2001, 97(7), 56 – 9; McCausland, L. M.; Cains, P. W.; Maxwell, M. Chem.-Ing.-Tech. 2001, 73(6), 717 – 8.

5. Young, F. R. Cavitation 2nd_{Ed. 1999, Imperial College Press.}

6. Brenner, M. P.; Hilgenfeldt, S.; Lohse, D. Rev. Mod. Phys. 2002, 74(2), 435 – 84; Flannigan, D. J.; Suslick, K. S. Nature 2005, 434(7029), 52 – 5; Didenko, Y. T.; Suslick, K. S. Nature 2002,

418(6896), 394 – 6.

7. Wohlgemuth, K.; Kordylla, A.; Ruether, F.; Schembecker, G. Chem. Eng. Sci. 2009, 64, 4155

Technologies for Optimisation and Control of Nucleation and Growth for New Generations of Industrial Crystallizers

Technologies for Optimisation and Control

of Nucleation and Growth for New

Generations of Industrial Crystallizers

Technologies for Optimisation and Control

of Nucleation and Growth for New

Generations of Industrial Crystallizers

Proefschrift

ter verkrijging van de graad van doctor

aan de Technische Universiteit Delft,

op gezag van de Rector Magnificus prof. ir. K.C.A.M. Luyben,

voorzitter van het College voor Promoties,

in het openbaar te verdedigen op maandag 24 november 2014 om 15:00 uur

door

Anamaria SOARE

Chemical Engineer, University Politehnica of Bucharest

geboren te Boekarest, Roemenië

Contents

Chapter 1

Introduction

Chapter 2

Crystal Nucleation by Laser

Induced Cavitation

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

63 µm

- 1 s

50 μs

1 s

30 s

Optical

disturbance

ring

Crystals ring

(a)

(b)

(c)

(d)

Crystals ring

(a)

(b)

Crack

20 µm

63 µm

- 1 s

50 μs

1 s

30 s

(a)

(b)

(c)

(d)

(e)

(f)

7 s

1 min

Crystals

(a)

(b)

(c)

(d)

_(b)

_{1 s}

_(d)