Dissertations / Theses on the topic 'Crystallisation processes'

Crystallisation is an important separation and purification technology in the pharmaceutical and fine chemical industry. Crystallisation processes are designed to generate and control supersaturation, nucleate desired polymorphs as well as crystal shapes and growing product crystals to the required particle size distribution and purity. Traditionally crystallisation is carried out in batch mode, due to the necessary process flexibility, although continuous processing can offer advantages of reproducible product quality, more sustainability as well as lesser waste and lower carbon footprint. This work describes a route towards the development of continuous crystallisation processes for small organic molecules. Supersaturation is the ultimate requirement for nucleation and in processes where mixing of two or more solutions is required to generate supersaturation, this step can affect nucleation. This thesis shows that higher mixing intensities yield higher solid recoveries over time and a smaller particle size. However, this phenomenon only applies to low and medium mixing flow rates, whereas results for high mixing flow rates are at the same level as for the medium ones. Furthermore, this effect was only observed in Valine - water:isopropanol (1:1) and at high supersaturations in Glycine - water:isopropanol (1:1) systems. In L-Glutamic acid - water:isopropanol (1:1) or L-Asparagine - water:isopropanol (1:1) systems this effect was not observed at all. Even the reactive precipitation of L-Glutamic acid (H-Glu) from Na-Glutamate and H2SO4 did not show any effects of mixing intensity on solid recovery over time. The mixing insensitive reactive precipitation of H-Glu was used to study the effect of post-mixing flow treatment on solid recovery over time and final polymorphic population. Micromixed samples were exposed to different batch flow units with hydrodynamics of a quiescent crystalliser (QC), stirred tank crystalliser (STC), magnetically stirred crystalliser (MSC), peristaltic pump recirculation loop (PPL) and an oscillatory baffled crystalliser (OBC). Harsh hydrodynamic conditions or mechanical impact like in the STC, MSC or OBC yield the metastable prismatic Alpha H-Glu polymorph and significantly increase solid recovery over time. Milder hydrodynamics like in a QC or PPL yield the stable platelet/needle like Beta H-Glu polymorph, where the PPL shows enhanced solid recovery over the QC. Despite XRPD analysis indicating pure Beta phase, the QC samples also contain about 0.1 % of the Alpha form, which growth kinetics suggest that they must have formed very shortly after mixing. Connecting the continuous mixing setup with a Beta enhancing flow-through PPL unit and a sample collection vessel made a fully continuous Beta H-Glu crystalliser. However, this system never reached steady-state operation, fouling and blockage was a major challenge and an unexpected change in the polymorph population from the stable Beta to the metastable Alpha was observed. This system did not perform satisfactorily and therefore experiments were discontinued. For the mixing insensitive antisolvent crystallisation of H-Glu, a novel rapid continuous antisolvent crystallisation setup was developed to produce crystal suspension of the Beta polymorph with a small size and narrow particle size distribution. The system jetinjects aqueous H-Glu solution into the bulk of isopropanol antisolvent and its performance was characterised with respect to different antisolvent mass fraction, bulk supersaturation, polymorphic population, steady-state operation, solid recovery over time, crystal size, particle size distribution and scale-up capabilities. Results show that increasing the antisolvent mass fraction reduces the final crystal size and particle size distribution, crystal product is of pure Beta form with a high yield, the system rapidly achieves very high supersaturation and reaches steady-state operations after about 20-30 min. Higher total flow rates and scale-up of the system did not show any effect on particle and system properties. The produced crystal slurry exhibits ideal properties of a Beta H-Glu seeding suspension for further crystal growth. Continuously seeded continuous crystal growth cooling crystallisation experiments were carried out in a tubular continuous oscillatory baffled crystalliser (COBC). H-Glu solution of two different concentrations was pumped through the system and Beta H-Glu seeding suspension of two different seed loadings were injected into the saturated solution. Mass balance calculations, supersaturation data, seed loading & solution concentration, crystal morphology information, crystal growth rates and mean residence time in each temperature section along the rig were used to predict solution concentration, desupersaturation behaviour and the temperature profile of the process. Online Focused Beam Reflectance Measurement (FBRM) analysis and offline laser diffraction particle sizing measurements recorded crystal growth in the system. Offline solid recovery analysis over time at various points along the rig successfully confirmed the predicted solution concentrations and desupersaturation profile of the COBC. However, micrographs and Scanning Electron Microscopy (SEM) analysis indicated that final product crystals are agglomerated. The effect of seed loading & solution concentration on the final agglomerated particle size distribution and solid recovery over time was investigated. Higher solution concentration led to larger product crystals, whereas different seed loadings did not show a clear trend. Steady-state crystallisation was demonstrated based on particle size distribution as well as supersaturation data after each crystalliser residence time and was achieved within 20 min after the system was conditioned with crystal slurry. Fouling and secondary Alpha nucleation was not a problem as long as the supersaturation did not exceed a level of 3.7 in the bulk solution.

Fouling involves the unwanted deposition and build-up of solid material on surfaces within a process. This problem is widely encountered in multiphase and solid phase processing in many industries including oil and gas, pharmaceutical and fine chemical manufacturing sectors. Although it is acknowledged to impact both batch and continuous processing methods it poses a particular challenge to the controlled operation of continuous crystallisation processes where extended operation under non-equilibrium conditions is required. Whilst the factors impacting on fouling have been proposed, there have been only a relatively limited number of studies into fouling mechanisms to date. With increased interest in deploying continuous crystallisation processes for pharmaceutical manufacturing, the motivation for this work was to develop an improved understanding of the influence of material properties and process conditions on fouling processes. In this work, a number of studies were conducted in which key materials and process parameters were investigated. These have included different materials of construction (MOCs), process conditions (flow, supersaturation, temperature gradients (ΔT)) and crystallising solutions (solute and solvent). Primary fouling studies were conducted using a small scale batch crystallisation setup to explore the influence on MOCs, supersaturation and agitation rate upon both bulk crystal nucleation and surface fouling of paracetamol. The prominent fouling mechanism was found to be particle deposition which was influenced by supersaturation, agitation rate, different MOCs and exposure time. Fouling is known to occur on heat exchange interfaces due to the localised supersaturation that can be generated e.g. in a plug flow continuous cooling crystalliser. A novel surface induced continuous crystallisation fouling assessment platform (C-FAP) was developed in conjunction with Cambridge Reactor Design (CRD). The C-FAP was evaluated as an assessment tool by exploring different MOCs and process conditions upon fouling and fouling mechanisms via in situ imaging and temperature measurement. The platform was characterised and used to explore surface induction mechanisms in which initiation and growth was strongly influenced by different MOCs, with stainless steel showing a greater tendency than PTFE, in addition to the degree of supersaturation. The temperature difference across the MOC interface (ΔTMOC) was demonstrated to influence nucleation and growth to varying extents. An ideal scenario would be to be able to predict or rule out unfavourable combinations of solute, solvent and MOC properties early in process design to avoid late stage problems. A screen was carried out to assess the potential to develop a multivariate predictive model for fouling propensity and fouling behaviour. The models provide insight into the most influential parameters comprising MOC, solute, solvent and process descriptors to steer subsequent experiments. The importance of MOC properties and process conditions was highlighted for all models. A variety of assessment tools were demonstrated within this work in which recommendations for fouling evaluation were provided in addition to methods to further develop fouling understanding.

The development of a Semi-Open Refrigeration Cycle had been shown to be of economic value to the process of crystallisation. The experimentation took the form of several bench scale experiments to develop the designs of the individual vessels that were to constitute the final pilot plant. In view of the safety hazards involved with a plant contacting Sulphuric Acid and Liquid Butane, a substantial Hazop and HAZAN study has been carried out. From this data the pilot plant was constructed, with all necessary safety features. Computer monitoring and partial computer control was installed. The pilot plant was operated as a closed loop for extended periods, to enable the production of both Iron (2) and Copper (2) Sulphates to be produced on a continuous basis. The results showed that the crystals produced are of a small size, indicative of the instantaneous chilling of the mother liquor and are not contaminated to a significant degree with the refrigerant. The thermal efficiency of the cycle was lower than expected at 80%, but improvements in insulation coupled with the redesign of the condenser, eliminating the necessity for a separator vessel, should increase this to the expected 90%.

The air/water interface and crystallisation have long been challenging to study despite their crucial importance in many areas of modern science. They can be difficult to study without interfering in some way with the process being observed and once the work it can be very challenging to interpret the results which have been gathered. This work attempts to understand the air/water interface and other air/solvent interfaces using variable angle spectroscopic ellipsometry. The values of the complex refractive index which are measured give insight into the nature of these interfaces and the apparent differences between water and other solvents. Work has also been performed to study the process of crystallisation through second harmonic imaging microscopy and single crystal x ray diffraction. To bridge these two areas, p-Nitrophenol has been studied in depth, with a literature review of the surrounding work on the compound, as well as studies on the effects of irradiation on the α crystal polymorph. Finally, the growth of a p-Nitrophenol layer at the air/water interface has been studied using Langmuir Blodgettry and tensiometry to understand the rate and mechanism of the growth.

Membrane technologies are considered a promising solution for water scarcity in arid regions. However, fouling is a major challenge facing the application of membrane technologies. Fouling limits the economic viability and reduces the overall efficiency of membrane processes. Therefore, fouling mitigation is a crucial factor in spreading the use of membrane technologies for new applications. The first step in fouling mitigation is to predict the propensity of fouling. Unfortunately, there are immense limitations in current industrial practises for fouling propensity prediction. These limitations come from using outdated and inapplicable approaches, in which crucial assumptions are made. For example, in the case of crystallisation fouling or “scaling” one of the major simplifications is the use of pure scaling salt data to predict the propensity of scaling when, in reality, co-precipitation is present. This research work aims to introduce a new approach to systematic assessment of the fouling problem under real and complex conditions and to enhance understanding of the importance of including interactive effects and co-precipitation in the prediction of scaling propensity. In this research work a novel procedure accounting for the local variation of thermodynamic properties along a long membrane channel is proposed. A new approach considering ion interaction and process hydrodynamics for the prediction of the scaling propensity is then introduced. This new approach provides for the first time a completely theoretical assessment for pure salt scaling propensity along a full scale filtration channel without the use of any empirical constants. A new procedure for including the effect of co-precipitation on scaling propensity prediction is developed. The effect of process pressure on solubility products is included theoretically for the first time to enhance the accuracy of scaling propensity prediction during the full scale RO process. This research work helps to produce more reliable and accurate prediction of the onset of scaling which will help strategies to mitigate scaling and increase the overall efficiency of RO/NF processes. The new approach can be applied in practical situations and could be developed to a user-friendly programme able to give an accurate prediction of the fouling propensity in full scale processes allowing the optimisation of membrane processes accordingly. Moreover, comprehensive experimental work has been carried out during this PhD research work to enhance understanding of crystallisation fouling and coprecipitation. The effect of salinity and dissolved organics (DO) in CaSO4 and SrSO4 precipitation and co-precipitation are studied and discussed. Quantitative and qualitative thermodynamic and kinetic analyses combined with structural analyses of deposits are carried out to investigate the effect of salinity, DO presence and coprecipitation on SrSO4 and CaSO4 precipitation. The observations in this experimental study are very important for a deeper understanding of the effect of scaling salts’ coexistence, salinity and DO presence on the behaviour of the scaling salts. This is crucial to reaching a reliable prediction of the scaling propensity within RO/NF processes. Finally, the new developed approaches in this thesis have been validated using set of hydrodynamic tests. This set of tests has been carried out using a newly installed laboratory membrane rig. Moreover, a new technique to simulate full scale membrane processes is proposed using a laboratory membrane rig combined with the programs previously developed in this thesis. This new technique can be used to study the effect of process hydrodynamics on scaling and process performance of full scale membrane processes using a laboratory membrane rig. The outcomes of this research work can be used to investigate the optimal operating conditions and to guide design criteria for different RO/NF practical scenarios.

Electrical impedance spectroscopy (EIS) is a method used to study the frequency dependence of the dielectric properties of colloidal suspensions by applying an alternating electric field. When an alternating electric field is applied, a dipole moment can be induced on a charged-particle due to the relative motion between the particles and their electric double layer. The macroscopic display of induced dipole moment is usually represented by the impedance parameters, including the impedance real part, imaginary part, phase angle and the relaxation frequency. These quantities are related to the size, shape and surface of the dispersed particles, the nature of the dispersed medium, and also the concentration of the particles. This thesis describes a fundamental study of the EIS method applied to colloidal particles. The relationship between the impedance parameters and the properties of particle suspensions is investigated. The study reveals the effects of particle size, particle concentration and ionic concentration dependence on the detected impedance parameters. Based on the study, new methods, including modelling, signal process, test set-up and data analysis, for characterisation of particles in suspensions are developed through the experimental approach and theoretical analysis. The methods are verified with silica suspensions and applied to crystallisation processes. The online measured electrical impedance spectra associated with L-glutamic acid nucleationgrowth processes and a polymorphic transformation are analysed. It is demonstrated that the methods can be applied for on-line monitoring of the particle size and polymorphs in crystallisation processes. Electrical impedance tomography based on EIS measurement conducted with different materials, including non-conductive plastic bar, banana, and silica suspensions are studied. The responses of electric polarisation of colloidal particles on tomographic images can be observed. However, the difference in particle size cannot be observed in the tomographic images possibly due to the limits of the imaging resolution from an 8-electrode sensor and the signal quality affected by the limits of the common mode voltage rejection ratio of the instrument.

Crystallisation of L-glutamic acid (LGA) and D-mannitol were studied in batch and continuous reactors using non-invasive, wide area illumination Raman spectrometry to identify the nucleation point, monitor crystal growth and identify the polymorphs formed. Non-invasive broadband acoustic emission (AE) spectrometry was also investigated as a means to monitor the crystallisation profile (by integrating the AE intensity over specific frequency ranges) and changes in particle size (by ratioing the intensities at high and low frequency regions). A design of experiments approach evaluated the effects of solution concentration, oscillation frequency and amplitude, and final temperature on crystal formation and polymorph transformation for batch and continuous oscillatory baffled reactors. For both compounds, concentration had the biggest impact on the particle properties while the main interaction was between the oscillation frequency and amplitude. Comparisons were made with results from crystallisations in batch stirred tank and continuous mixed suspension, mixed product removal (MSMPR) reactors. Metastable a-LGA was more easily obtained in both continuous reactors than conventional batch reactors. Narrower size distributions (e.g. spans of 1.5 - 2 and 4 - 8, respectively) and lower yields (e.g. 20 - 50% and 30 - 70%, respectively) were obtained in the continuous OBR than the MSMPR. The continuous OBR exhibited limited operating conditions, outside of which poor nucleation or blocking of the reactor occurred. The MSMPR reactor was more versatile and could be operated for up to 70 hours compared to several hours for the continuous OBR. The study of D-mannitol in stirred tank and batch OBR explained how different operating conditions (especially oscillation frequency and amplitude) affected the polymorph formed, the rate of crystal formation, particle size range, and the degree of polymorph conversion. The metastable a- and the stable ?-forms were the predominant polymorphs obtained; higher solution concentration and mixing intensity gave more alpha and vice versa for the gamma form.

During recent years crystallisation has found applications in many chemical industries, such as pharmaceutical, petrochemical, micro-electronics and food industries. Crystallisation is a basic step for purification or separation for a large variety of organic, inorganic and pharmaceutical compounds. Most of the product qualities are directly related to the shape of the crystal size distribution (CSD). The main difficulty in batch crystallisation processes is to accomplish a uniform and reproducible CSD. On-line control during the process allows for improved crystalline product quality, shorter process times and reduction or elimination of compromised batches. The actual prediction and estimation of the shape of the distribution at the end of the batch can provide useful information for monitoring or designing the operating curve for the supersaturation controller. Model-based approaches provide consistency of the CSD, can be used for better control and also for product design by reverse engineering the process to achieve the desired CSD and shape. This research presents a novel methodology for solving the population balance equation (PBE) for the estimation of the shape of the crystal size distribution for batch crystallisation processes. The approach combines the quadrature method of moments (QMOM) and the method of characteristics (MOCH), and provides a computationally efficient technique for the reconstruction of the whole crystal size distribution. The technique was used to estimate the kinetic parameters for the size-dependent growth and secondary nucleation, for potash alum-water system using industrial pilot plant data provided by BASF, Chemical Company. The combined technique was also used to estimate the size-dependent dissolution parameters for potash alum-water system, using laboratory scale data. The QMOM-MOCH solution approach is evaluated in a model-based dynamic optimization study, with the aim to obtain the optimal temperature profiles, which drive the system in both the supersaturated and under-saturated region, to achieve desired target CSD. Using growth, dissolution and nucleation parameters the technique was used to optimise the temperature trajectories to obtain bimodal and mono-modal distributions. The technique can serve as a soft sensor for predicting the CSD, or as a computationally efficient algorithm for off-line design or on-line adaptation of operating policies based on knowledge of the full CSD data. Additionally, the PBE model was solved using the method of characteristics under the assumption of constant supersaturation. At constant supersaturation growth is the dominating phenomenon, yielding a simplified analytical expression for the prediction of the CSD. The research presents the new methodology for the systematic design of the setpoint operating curves for supersaturation controlled crystallisation processes, which produces a desired target crystal size distribution (CSD) at the end of the batch. A design parameter, was introduced as a function of the supersaturation and time, and is evaluated for supersaturation controlled processes. Based on the design parameter and the simplified analytical model, the supersaturation setpoint and batch time are determined using an optimisation approach to obtain a target distribution with a desired shape. Two additional methods are also proposed that use the seed in conjunction with the supersaturation setpoint design, and analytical CSD estimator for shaping the product CSD. The first approach designs a seed recipe as a mixture of crystals resulting for example from standard sieve analysis. In this approach the seed was introduced at the beginning of the batch. The second approach introduces the dynamic seeding concept, which allows an easily implementable methodology to achieve complex target CSDs using seed with mono-modal distribution as a process actuator. These methodologies were validated for potassium dichromate-water system. Size-dependent growth kinetic parameters for the potassium dichromate-water system were identified using as experimental setup developed at Loughborough University. The experiments presented in the thesis also illustrates the simultaneous application of in situ Process Analytical Technology (PAT) tools, such as focused beam reflectance measurement (FBRM) for nucleation detection, attenuated total reflection (ATR) UV/Vis spectroscopy for concentration monitoring, as well as the in-line use of a Mastersizer for real-time CSD measurement in the case of the potassium dichromate in water system. The approaches provide a comprehensive framework for model-based dynamic optimisation of crystallisation processes, which combines efficient numerical solution approaches of the PBE with the formulation of novel optimisation problems. The techniques presented include controlled dissolution, simultaneous optimisation of operating policies and seed recipes and dynamic seeding. Simulation and experimental evaluations of the proposed approaches demonstrate the potential of the techniques to provide significant improvement in the current state-of-the-art in crystallisation control.

Pharmaceutical crystallisation operation is often critical because it determines product properties, such as the crystal size distribution (CSD) and polymorphic form, that can influence the subsequent downstream operations and the product therapeutic performance. Driven by the United States Food and Drug Administration s (FDA) Process Analytical Technology (PAT) initiative and the Quality-by-Design (QbD) concept, the development of control approaches, which can improve the manufacturing of products with desired properties, has become of significant interest. This thesis presents the development and application of PAT-based approaches for the monitoring and control of pharmaceutical crystallisation operations that will ensure consistent production of active pharmaceutical ingredients (APIs) with the desired size and polymorphic form. The approaches utilised Lasentec focused beam reflectance measurement (FBRM) and attenuated total reflectance ultraviolet (ATR-UV) spectroscopy as the in situ monitoring and control tools. Crystallisations of the APIs that posses multiple polymorphs are both critical and challenging. This was illustrated in this work by the crystallisations of sulfathiazole polymorphs using literature methods. The processes were monitored using FBRM and ATR-UV spectroscopy to define the design range of the process parameters. The defined range could be used as a recipe to reproduce the same quality of crystals. The obtained crystals were characterised using various techniques (optical microscopy, scanning electron microscopy (SEM), differential scanning calorimetry (DSC), thermogravimetry, hot-stage microscopy (HSM), Fourier Transform infrared spectroscopy and powder X-ray diffractometry) to assess the success of the crystallisation processes. The combined results of the techniques showed that all methods were able to produce the desired pure polymorphs. As a contribution to the technique of investigating polymorphism, a combined approach of DSC-HSM with image analysis, was introduced. Results show the capability of the approach to provide a unique insight into the polymorphic transformations and thermal behaviour exhibited by the model compound. The novel direct nucleation control (DNC) approach was introduced to control the CSD. The approach utilises information on nucleation, provided by FBRM, in a feedback control strategy that adapts the process variables, so that the desired CSD of product is achieved. It also provides in situ fines removal through the operating policy, rather than having additional equipment and external recycle loops. The approach does not require concentration measurement and has the advantage of being a model-free approach, requiring no information on nucleation or growth kinetics in order to design an operating curve; the system automatically and adaptively detects the boundary of the operating curve. Experimental results, using glycine in water-ethanol mixture as a model system, show the benefits of DNC to produce larger crystals with narrower CSD compared to uncontrolled operations. The capability of seeded cooling crystallization with temperature cycling approach to control crystal size uniformity and polymorphic purity was evaluated. Using sulfathiazole in n-propanol and in water as model systems, the method was found to accelerate the growth and enhance the size uniformity of the crystals, in comparison with runs using a linear temperature profile, by promoting Ostwald ripening. Although the approach is conceptually capable of controlling polymorphic purity of a system, the effect of solvent-mediated nucleation/growth can be more dominant, as shown by the results of the experiments. The insights into this behaviour of sulfathiazole crystals were captured very well by the FBRM. The study also demonstrated the successful use of a simple non-linear function as a calibration model to relate temperature and absorbance data, obtained using the ATR-UV spectroscopy, to solute concentration during the crystallisation process. The effect of temperature cycling, performed during seeded cooling crystallisation, on the surface features of sulfathiazole crystals was investigated using FBRM and ex situ optical microscopy, SEM and atomic force microscopy. It was observed during the initial stage of the process, the heating phases produced crystals with smooth surfaces, whilst the cooling phases promoted growth of features on the surfaces. These changes detected by the FBRM as an increase in the number of coarse counts during heating and a drop during cooling. Laser beam spreading caused by the surface roughness, and signal/chord splitting due to sharp edges are offered as an explanation for the FBRM results. This shows the capability of the FBRM to provide useful information about the changes on the surface of the crystalline products. The information can be used to avoid problems in the downstream operations, or in the final product property due to variations in flowability and friability, which are influenced by the surface property.

No
The present study demonstrates the application of decision tree algorithms to the co-crystallization process. Fifty four (54) batches of carbamazepine–salicylic acid co-crystals embedded in poly(ethylene oxide) were manufactured via hot melt extrusion and characterized by powder X-ray diffraction, differnetial scanning calorimetry, and near-infrared spectroscopy. This dataset was then applied in WEKA, which is an open-sourced machine learning software to study the effect of processing temperature, screw speed, screw configuration, and poly(ethylene oxide) concentration on the percentage of co-crystal conversion. The decision trees obtained provided statistically meaningful and easy-to-interpret rules, demonstrating the potential to use the method to make rational decisions during the development of co-crystallization processes.

Includes bibliographies.
The MSMPR was designed according to chemical engineering principles and tests showed that it conforms to the assumptions necessary for conventional MSMPR analysis. Various crystalliser offtakes and six system designs were developed and tested to allow measurements using the thermostatted flow cell. The flow cell allows for continuous measurement of the sample dispersion using a noninvasive monitoring instrument - a Malvern Particle Size Analyser. In this technique, measurement of the steady state crystal size distribution is based on a diffraction pattern produced by the particles in a parallel beam of monochromatic light. Comparative measurements were obtained by sampling directly from the crystalliser and analysing the product in a Coulter Counter. In this technique, the steady state crystal size distribution is measured by the change in resistance as particles, in an electrically conductive liquid, pass through the Coulter aperture. Due to the large volumes required by the crystalliser, it was necessary to pool urines. Aliquots of volume 1. 5 dm3 from the 24 hour urines of each of 8 male controls were pooled. The first series of experiments were performed to investigate the effect on calcium oxalate crystallisation of varying the concentration of sodium oxalate solutions used to initiate crystallisation, and also to investigate the effect of varying the temperature. A further series of experiments was performed to determine the effect of adding inhibitors individually (citrate, magnesium, chondroitin sulphate A) and in combination (citrate and magnesium). The inhibitor experiments were performed using two different approaches. In the first, the pooled urine was pre-treated with the inhibitor prior to being introduced into the crystalliser. In the second set of experiments, the inhibitor was introduced into the crystalliser via a separate, independent feed. It is suggested that these approaches represent crude models respectively of (a) the endogenous presence of inhibitors and (b) the oral administration of such inhibitors. Each individual experiment was performed using three different urine pools. The steady state crystal size distributions obtained from both the Coulter and the Malvern instruments were used to determine nucleation and growth rates in each urine. Scanning electron microscopy was used to obtain qualitative and semi-quantitative data related to crystal morphology, number, size and degree of aggregation. Both nucleation and growth rates increased as the concentrations of the initiating sodium oxalate solutions increased. Nucleation and growth rates also increased with increasing temperatures. In the latter series of experiments, different calcium oxalate crystal phases precipitated at the various temperatures. Citrate inhibited nucleation and growth of calcium oxalate, irrespective of how it was admitted to the crystalliser. Magnesium was found to inhibit growth only and, like citrate, this role was independent of how it was introduced into the crystalliser. Although both inhibited growth, the effect of the citrate was greater. When both components were added in combination, inhibition was observed to be additive. In all cases, inhibitory effects increased with increasing concentration. Chondroitin sulphate A was found to inhibit nucleation; inhibition of growth was minimal at the concentrations tested. Aggregation of calcium oxalate crystals was found to be inhibited by chondroitin sulphate A. The results of the calcium oxalate crystallisation kinetic studies, obtained by using Malvern particle sizing and Coulter Counter techniques independently of each other, showed general agreement. These results, in turn, were in agreement with the published results of others, thereby lending confidence to the findings and highlighting the potential application of the non-invasive continuous crystalliser in urolithiasis research.

In this thesis novel configurations and operating strategies in the mixed suspension mixed product removal (MSMPR) crystalliser are investigated, aided by integrated process analytical technologies (PAT) and crystallisation informatics system (CryPRINS) tools. The MSMPR is an idealised crystalliser model that assumes: steady-state operation; well mixed suspension with no product classification, such that all volume elements contain a mixture of particles (small and large) and crystal size distribution (CSD) that is independent of location in the crystalliser and is identical of the product withdrawn; and uniform supersaturation thought, leading to constant nucleation and growth rates. Single-stage MSMPR designs with continuous recycle/recirculation and modified heat exchanger were investigated and found to minimise fouling, encrustation and transfer line blockages. In particular, a modified MSMPR with baffled heat exchanger was found to significantly reduce the temperature between incoming feed hot feed solution and the cooled crystalliser, leading to a significant reduction in fouling, encrustation and blockages. In addition, the concept of the periodic mixed suspension mixed product removal (PMSMPR) crystallisation process is demonstrated for the first time viz single- and multi-stage cascaded operations. This method of operation involves the periodic transfer of slurry (addition and withdrawal) at high flow rates from either a single stirred vessel or between a number of stirred vessels arranged in series. The PMSMPR is therefore characterised by periodic withdrawals of product slurry. Similar to the MSMPR, the product withdrawn from a PMSMPR has exactly the same composition as the vessel at the time of removal. The rapid withdrawal of slurry at high flow rates in PMSMPR operation leads to the prevention of particle sedimentation and blockage of transfer lines. The transfer of slurry (to/from) the PMSMPR is followed by a holding (or pause) period when no addition or withdrawal of slurry takes place. The holding period extends the mean residence time of the PMSMPR relative to a typical MSMPR, thereby increasing the yield and productivity of crystallisation as more time is allowed for consumption of available supersaturation viz crystal growth and nucleation. A state of controlled operation (SCO) in the periodic flow process, defined as a state of the system that maintains itself despite regular, but controlled disruptions was characterised using the PAT tools and CryPRINS within an intelligent decision support (IDS) framework. The crystallisation of paracetamol (PCM) from isopropyl alcohol (IPA) using different configurations of a single-stage continuous MSMPR crystalliser that incorporated continuous recycle and recirculation loop, and a novel design with baffled heat exchanger was investigated. Crystallisations of PCM-IPA carried out in the MSMPR without heat exchanger suffered from severe fouling, encrustation and blockage problems due to the high level of supersaturation (S = 1.39) in the crystalliser, which was required for the initial burst of nucleation to generate enough particles for later growth, as well as the large temperature difference between the incoming feed (45 oC) and the crystalliser (10 oC). Using the modified MSMPR design with baffled heat exchanger, the challenges of fouling, encrustation and blockage were significantly reduced due to the rapid lowering of the feed stream temperature prior to entering the crystalliser. In addition, the closed loop system led to conservation of material, which is a great benefit since large amounts of materials would otherwise be required if the MSMPR was operated with continuous product removal. This design is great for research purposes, in particular, to investigate process design and optimisation. Continuous crystallisation of PCM in the presence of hydroxyl propyl methyl cellulose (HPMC) additive was investigated in the modified MSMPR design with heat exchanger. HPMC was found to improve the crystallisation performance, leading to complete avoidance of fouling, encrustation and blockages at a concentration of 0.05 wt%. However, the yield of crystallisation was significantly reduced (28.0 %) compared to a control experiment (98.8 %, biased due to fouling/encrustation) performed without additive addition. Regardless, the productivity of crystallisation was more than four times that achieved in batch linear cooling (LC) (0.62 0.86 g/L-min) and batch automated dynamic nucleation control (ADNC) (0.24 0.25 g/L-min) runs. Aspects of the periodic flow crystallisation of single- and multi-component (co-crystals) molecular systems have also been examined to demonstrate the concept of state of controlled operation . The single component systems studied were PCM and glycine (GLY), each representative of compounds with slow and fast growth kinetics, respectively. The co-crystal systems investigated were urea-barbituric acid (UBA) and p Toluenesulfonamide-Triphenylphosphine oxide (p-TSA-TPPO). UBA is a polymorphic co-crystal system with three known forms (I, II and III). Form I UBA was successfully isolated in a three-stage periodic flow PMSMPR crystalliser. This study demonstrates the capability of periodic flow crystallisation for isolation of a desired polymorph from a mixture. p-TSA-TPPO exists in two known stoichiometric co-crystal forms, 1:1 and 3:2 mole ratio p-TSA-TPPO, respectively. The two crystalline forms exhibit solution mediated transformation, which proves to be a difficulty for separation. For this study, the implementation of temperature cycles in batch and flow control in semi-batch and periodic PMSMPR crystallisers were investigated to isolate pure 1:1 and 3:2 p-TSA-TPPO, respectively. Different regions of the ternary diagram of p-TSA, TPPO and acetonitrile (MeCN) were investigated. The desired co-crystal form was isolated all crystallisation platforms investigated. However, greater consistency was observed in the semi-batch and PMSMPR operations respectively. Periodic flow crystallisation in PMSMPR is a promising alternative to conventional continuous MSMPR operation, affording greater degrees of freedom operation, slightly narrower RTD profiles, consistent product crystal quality (size, shape and distribution), longer mean residence times, higher yield and productivity and significant reduction in fouling, encrustation and transfer line blockages over prolonged operating periods. Furthermore, the PMSMPR is a versatile platform that can be used to investigate a range of different molecular systems. Relative to batch operation, the PMSMPR can operate close to equilibrium, however, this is dependent on the system kinetics. In addition, retrofitting of batch crystallisers to operate as PMSMPRS fairly simple and require only subtle changes to the existing design space. The integrated array of PAT sensors consisted of attenuated total reflectance ultra violet/visible spectroscopy (ATR-UV/vis), attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR), focused beam reflectance measurement (FBRM), particle vision microscopy (PVM) and Raman spectroscopy. The results from the studies reported here illustrate very well the use of PAT and information system tools together to determine when the continuous and periodic MSMPR operations reaches a steady-state or state of controlled operation (i.e. periodic steady-state). These tools provided a better understanding of the variables and operating procedures that influence the two types of operations.

Innovations and process understanding is the current focus in pharmaceutical industry. The objective of this research was to explore application of high power ultrasound in the slurry crystallisation and application of low power ultrasound (3.5 MHz) as process analytical technology (PAT) tool to understand pharmaceutical processing such as hot melt extrusion. The effect of high power ultrasound (20 kHz) on slurry co-crystallisation of caffeine / maleic acid and carbamazepine / saccharin was investigated. To validate low power ultrasound monitoring technique, it was compared with the other techniques (PAT tools) such as in-line rheology and in-line NIR spectroscopy. In-line rheological measurements were used to understand melt flow behaviour of theophylline / Kollidon VA 64 system in the slit die attached to the hot melt extruder. In-line NIR spectroscopic measurements were carried out for monitoring any molecular interactions occurring during extrusion. Physical mixtures and the processed samples obtained from all experiments were characterised using powder X-ray diffraction, thermogravimetry analysis, differential scanning calorimetry, scanning Electron Microscopy, dielectric spectroscopy and high performance liquid chromatography, rotational rheology, fourier transform infrared spectroscopy and near infrared spectroscopy. The application of high power ultrasound in slurry co-crystallisation of caffeine / maleic acid helped in reducing equilibrium time required for co-crystal formation. During carbamazepine / saccharin co-crystallisation high power ultrasound induced degradation of carbamazepine was negligible. Low power ultrasound can be used as a PAT tool as it was found to be highly sensitive to the changes in processing temperatures and drug concentration.

Now when the interest is increasing to reach a sustainable infrastructure, one possibility SCA is experimenting with is the possibility to produce renewable hydrocarbons from black liquor which can be extracted from a Kraft process. However, when extracting the black liquor, a lot of sodium-based compounds are removed from the recovery process and when hydrocarbons are produced in SCA’s biorefinery, these compounds are caught in an aqueous solution. The aqueous solution is received at 50°C, and the sodium-based compounds are mainly sodium sulfate and sodium carbonate, where the solution do also contain organic compounds and a solvent that is used in the biorefinery.  This thesis focused on building a concept to extract sodium sulfate from the aqueous solution. The thesis did also include if any additional preparatory work needs to be done to the solution before extracting sodium sulfate. Finally, a flow chart that maps the energy needed for the process was created.  The method that was used was crystallisation by cooling the solution. By cooling the solution, sodium sulfates solubility decreases which will result in that sodium sulfate falls out of the solution as crystals. It was determined that the solvent that the solution contains should be extracted if the solvents boiling temperature is below 100°C. Further, by cooling the solution under stirring to 15°C with a residence time of 3 hours, unwanted compounds can be extracted. By later cooling the solution under stirring to 5°C with a residence time of 1 hour, it gave sodium sulfate decahydrate (Na2SO4·10H2O) with small amounts of organic compounds. By removing the water, the dry product reached a purity of 94wt% sodium sulfate with a yield of 12% (mass of dry product/mass of aqueous solution). This result reached the specific objectives that were set at the start of this thesis, which was to reach a purity of 90wt% sodium sulfate with a yield of 5%.  The energy intensity for evaporating the solvent is expected to be high. It highly depends on which solvent is used. However, this process can use the lowest quality of steam that is available from the pulp mill. It is expected that the cooling will require high amounts of cooling water and a high investment cost for the heat exchanger. Yet, this is a vital part of the process to reduce the need for coolers which is powered by electricity.
Nu när intresset ökar, för att nå en hållbar infrastruktur, så experimenterat SCA med möjligheten att producera förnybara kolväten från svartlut som kan extraheras från en sulfatprocess. Vid extrahering av svartluten tas dock mycket natriumbaserade föreningar bort från återvinningsprocessen och när kolväten produceras i SCA:s bioraffinaderi fastnar dessa föreningar i en vattenlösning. Den lösningen tas emot vid 50°C och de natriumbaserade föreningarna är huvudsakligen natriumsulfat och natriumkarbonat, där lösningen också innehåller organiska föreningar och ett lösningsmedel som används i bioraffinaderiet. Denna avhandling fokuserade på att bygga ett koncept för att extrahera natriumsulfat från vattenlösningen. Avhandlingen omfattade också om ytterligare förberedande arbete måste göras av lösningen innan man extraherar natriumsulfat. Slutligen skapades ett flödesschema som kartlägger den energi som behövs för processen. Metoden som bestämde sig för att användas var kristallisering genom kylning av lösningen. Genom att kyla lösningen minskar lösligheten av natriumsulfater vilket leder till att natriumsulfat faller ut ur lösningen som kristaller. Det bestämdes att lösningsmedlet som lösningen innehåller skulle extraheras om lösningsmedlets koktemperatur är under 100°C. Vidare, genom att kyla lösningen under omrörning till 15°C med en uppehållstid på 3 timmar, kan oönskade ämnen extraheras. Genom att senare kyla lösningen under omrörning till 5°C med en uppehållstid på 1 timme gav natriumsulfatdekahydrat (Na2SO4·10H2O) med små mängder organiska föreningar. Genom att avlägsna vattnet nådde den torra produkten en renhet av 94 vikt% natriumsulfat med ett utbyte av 12% (massa torr produkt/massa vattenlösning). Detta resultat nådde de specifika mål som sattes i början av denna avhandling, vilket var att nå en renhet av 90 vikt% natriumsulfat med ett utbyte på 5%. Energiintensiteten för att förånga lösningsmedlet förväntas vara hög. Det beror mycket på vilket lösningsmedel som används. Denna process kan dock använda den lägsta ångkvaliteten som finns tillgänglig från massafabriken. Det förväntas att kylningen kommer att kräva stora mängder kylvatten och höga investeringskostnader för värmeväxlaren. Ändå är detta en viktig del av processen för att minska behovet av kylare som drivs av elektricitet.

Fedmis Pty (Ltd) situated in Palaborwa, South Africa produces phosphoric acid using the wet process production process. For this study, two main areas of concern in the wet process phosphoric acid production were investigated. The first area is the formation of sludge in the system due to impurities that reduces the grade of the acid produced, thereby lowering the selling price. The second area is the crystallisation of the gypsum that influences filtration and thereby affects plant productivity. These two aspects were investigated separately as they occur in different steps of the production process at different acid concentrations. A major component of the acid sludge is known as x-compound, ((Fe,Al)3KH)14 (PO4)8.4H2O). The purpose of the investigation of x-compound is to determine what effects different ionic impurities have on its precipitation and to determine if these effects could be used to decrease the amount of sludge formation. Due to the complexity of the system and the wide variety of impurities only the major impurities were considered in this study. These impurities included potassium (K+), sodium (Na+), magnesium (Mg2+), aluminium (Al3+) and iron (Fe3+). For all the experiments investigating the effect of impurities, analytical reagents were used on laboratory scale. For the silica experiments, commercially available samples were used. For the experiments investigating the impurity effects on the precipitation of x-compound it was found that:

• Agitation increases x-compound precipitation and can be used commercially to increase the precipitation rate to a point where sludge can be removed before transportation.
• Adding x-compound seeding crystals or magnesium ions also increases precipitation.
• Adding gypsum, sodium, hexafluorosilicates or fluorosilic acid reduces the precipitation, with sodium ions producing the lowest yield. This reduction is however not sufficient to be used commercially.

From the Raman study it became clear why x-compound precipitation is such a slow process. E At low acid concentrations, more H2PO4 - ions are present that form a complex with iron and aluminium. E As the acid concentration increases the concentration of H2PO4- ions decrease as the degree of dissociation of phosphoric acid decreases. The ferric- H2PO4- and aluminium- H2PO4- complexes become less stable and ultimately precipitation of the x-compound is favoured above solvation. E Addition of potassium impurities to the solutions had no visible effect on the Raman spectra and is suspected not to form a complex with the acid. From the silica sources investigated namely Dicalite, Serina Kaolin, Foskor silica and Aerosil 200 it can be concluded that none of the sources will be useful for the removal of potassium through formation of potassium hexafluorosilicates. For the determination of the concentration of impurities present in the production of phosphoric acid, the Fedmis monitoring program was initiated. It included the monitoring of Foskor rock analyses on a daily basis, and the monitoring of the 27%, 39% and 54% P2O5 phosphoric acid and precipitate, from these acid solutions on a weekly basis. From the investigation of the effect of these impurities on the solubility of potassium hexafluorosilicates, it was found that magnesium causes K2SiF6 to be the most soluble and fluoride the least. Unfortunately, the impurities did not help to reduce the potassium concentrations in the acid to below the required amount for sludge formation. For the calcium sulphate dihydrate surfactant experiments, the purpose of the investigation was to determine whether higher crystallisation qualities could be obtained to improve plant productivity. The investigation was limited to using surfactants with sulphate or phosphate functionalities and experiments were done on laboratory scale using analytical reagents. Atphos E3205, Atpol E3202 and Atpol E1231 are polyethoxylated alkyl phenol phosphate esters that had no visible effect on the crystal structure of the precipitated gypsum, but differences in the crystal sizes were observed. Smaller crystal structures with relatively equal masses compared to reference experiments are an indication of a growth inhibitor and a nucleation promoter as seen with Atphos E3205 and Atpol E3202. Increased crystal sizes were obtained using Atpol E1231. The use of Calsoline Oil caused a wider crystal size distribution in the precipitated crystals as thin and broad crystals with approximately the same length are found. The crystal mass obtained is also approximately the same as that of the reference experiment. Thus, it can be concluded that the surfactant affects the growth of the crystals and not the nucleation. Arlatone 1489, calcium gluconate monohydrate, Dowfax Hydrotrope and Tamol NN 8906 had no visible effect on the structure or size of the precipitated gypsum crystals. With the use of Nansa SS30, drastic effects were seen on the crystallisation of the calcium sulphate as small hexagonal rods were found. With an increase in surfactant concentration, there is a clear decrease in the mass of crystals obtained as well as crystal size. An increase in the crystal size distribution and a decrease in crystal size reduced the filtration rate dramatically. Experiments carried out for 24 hours exhibited the same trends where there is a decrease in yield with an increase in surfactant concentration. Higher yields were however obtained proving that mass transfer barriers were overcome. The results from these experiments again indicate that the surfactant affects crystal growth and nucleation. With the use of Dowfax 3B2 there is definite reduction in yield with an increase in surfactant concentration reaching a minimum at approximately 70% yield. Due to the presence of large amounts of smaller crystals and the almost constant yield obtained compared to the reference experiment, it can be concluded that this surfactant is a growth and not a nucleation inhibitor. As with Nansa SS30, experiments where Empicol LZ/D was used show a continuous decrease in the yield obtained with an increase in the surfactant concentration. At higher concentration of Empicol LZ/D, it seems as if this surfactant changed from a growth promoter to a growth inhibitor because although broad longer crystals are present, there are now also much smaller crystals formed. The crystal size distribution also broadens considerably. Overall, very high yields were obtained using Empimin KSN70 and the observed crystal size distributions were very narrow. The only difference was that the crystals appeared to be more porous or fibrous compared to the reference experiment. It is recommended that the experiments showing promise as crystal habit modifiers like Nansa SS30 and Empicol LZ/D be investigated in more detail as well as combinations of surfactants. Both the areas of concern in the process were investigated successfully. For the sludge formation problem, it is now clear what effects the precipitation of x-compound as well as what affect the impurities and operating conditions have. For the crystallisation of gypsum using surfactants, it was proven that surfactants could be used to affect crystal growth, shape and distribution and in this way influence filtration.
Dissertation (MSc(Chemical Engineering))--University of Pretoria, 2005.
Chemical Engineering
unrestricted