Search

Microplastics have become a highly discussed topic: Whether through the decomposition of larger plastic waste or direct usage on a small particle scale, microplastics find their way into the natural ecosystem with uncertain consequences: In particular, the ingestion and therefore the introduction of these particles into the food web (often carrying adsorbed chemicals) poses still not fully understood hazards for mankind. This work focuses on the separation of such particles from aquatic suspensions and evaluates the viability of magnetic seeded filtration in this field. In this study it could be shown that it is generally possible to separate fine polymeric particles from dilute suspensions in a highly effective manner. A broad range of process conditions were investigated with separation efficiencies reaching up to 95 % . The pH and ultimately the surface charge of the particles were identified to be the crucial process relevant parameters. More exactly, the surface potentials should be either small or oppositely charged leading to a diminishing or even attractive electrostatic interaction. However, when both agglomeration partners exhibit an electrostatic repulsion, a thorough parameter optimization nevertheless yielded high separation efficiencies. In particular, the ionic strength and magnetic seed concentration proved to be determining factors in this case. Lastly, the process kinetics were investigated and a limited exponential growth character was observed which could be modeled with basic equations from agglomeration theory under simplifying assumptions. This knowledge further assists the selection of an optimal parameter set for a process engineering application.

The crystallization of active pharmaceutical ingredients (API) is an important, but complex unit operation in drug manufacturing industry, due to the diverse chemical nature of medicinal formulation compounds, several physicochemical phenomena that take place during particle formation process, and multi-phase governing resistances. This research paper presents a combined experimental–modelling approach to understanding fesoterodine fumarate crystallization and dissolution in 2-butanone. A rigorous in-house mechanistic model algorithm that includes calibration, thermodynamic equilibrium description, energy, mass and population balance equations, heat transfer limitations, and ordering, agglomeration and detachment step kinetics was developed. On the basis of experimentally-determined saturated solubility, a specified initial number of designed dissolution and cooling crystallization experiments was performed, monitored by attenuated total reflection Fourier-transform infrared (ATR–FTIR) spectroscopy, focused beam reflectance measurement (FBRM) technique and microscopy. The estimated kinetic parameters of transport, nucleation, crystal lattice growth, agglomeration and dissolution were evaluated by fitting the measured/simulated concentration curves with crystallite size distributions. Correlation was validated with additional trial results, observing a rather good agreement. The behaviour of the component during validations’ temperature cycling was thoroughly examined by simulations to achieve a solid final insight into mechanisms, which may aid further production development, optimization, as well as scale-up.

Fundamental research on developing a facile, industrially viable approach for producing concentrated single-layered graphene oxide (GO) aqueous dispersions is crucial for applications in related industrial fields. In this innovative work, cross-flow membrane filtration is reported for the first time for preparing stable GO dispersions with concentrations controlled from 1 to 16 g·L−1. The flux behavior with different process conditions (membrane pore diameter, transmembrane pressure, and tangential velocity), the relevant hydraulic resistances, and fouling mechanism have been systematically studied for optimization of the process parameters. The physicochemical properties of GO products are characterized by various methods, and the results show that all the concentrated dispersions possess good fluidity, with hydrolysis of functional groups resulting in a pH decline, and no re-stacking or agglomeration of the GO sheets occurring during the concentration process. The concentrated GO dispersions, presenting distinctive liquid crystal phases, offer tremendous potential in the fabrication of highly ordered macroscopic structures as well as the development of novel graphene-based devices.

In this work, a continuum theory for cohesive particles developed recently ( Kellogg et al., 2017 ) is applied to lightly-cohesive particles in an unbounded, gas-solid riser. The novelty of this recent theory is its fundamental incorporation of the effects of the granular temperature (i.e., continuum measure of impact velocity) and the material and cohesion properties on the rates of aggregation and breakage of agglomerates. Here, we extend this theory to multiphase systems and place particular emphasis on its ability to predict entrainment rate, as past empirical correlations vary by orders of magnitude when applied to the same system ( Chew et al., 2015 ). Specifically, continuum predictions of entrainment rates and agglomerate fraction are compared with Discrete Element Method (DEM) simulations of lightly-cohesive particles in a gas-solid flow. The agreement obtained for these quantities provides preliminary validation for the use of the continuum theory for cohesive particles in gas-solid flows.

Acoustic agglomeration is an aerosol preconditioning technology in which intense sound field is applied to promote relative motions and rapid agglomeration among particles. In order to improve the acoustic agglomeration efficiency and reduce its energy consumption, the method of adding liquid droplets into the aerosol was proposed and demonstrated to be effective. However, modeling and theoretical prediction of this process has been still lacking until now. In this paper, a three-dimensional discrete element modeling of aerosol acoustic agglomeration with addition of sprayed liquid droplets is established, which includes the agglomeration mechanisms of orthokinetic interaction, acoustic wake effect, gravity force and Brownian random force. In particular, the immersion mechanism is employed to describe the microscopic formation of aggregates. In addition, the multiple-time-step algorithm and the contact detection method based on a search grid are employed to speed up the simulation. Meanwhile, the corresponding experiments using coal-fired fly ash particles were performed to validate the simulation. It is found that the results of the discrete element modeling simulation are in reasonable good agreement with the experimental results, for both cases with and without the addition of water droplets. However, when the agglomeration efficiency exceeds 80% at the later stage of the process, some discrepancy is observed due to the breakage of large aggregates which is not taken into consideration in the modeling.

In contrast to materials properties, particle characteristics such as size, specific surface area, degree of agglomeration or crystallinity can be directly controlled by the synthesis process and also be determined experimentally without or with minimal further processing. Although particle characteristics only indirectly influence the (extrinsic) materials properties such as processability or charge carrier mobility and lifetime, they are key parameters to improve materials or device performance in applications. Process parameters – especially the time-temperature-profile in gas phase synthesis – control these nanoparticle characteristics. In this contribution a novel method is presented to discover paths to optimized nanoparticle characteristics in chemical vapor synthesis. As examples, physically realistic time-temperature profiles are predicted for a minimized degree of agglomeration at a desired primary particle size and a maximized degree of crystallinity at a desired specific surface area for the case of titania. The method relies on the integration of a simple model describing particle formation and growth into a Monte Carlo optimization algorithm. The results are surprising but can be rationalized using our understanding of particle formation and growth.

An adaptable direct simulation Monte Carlo (DSMC) method that settles the challenge of coexistence of agglomeration and rebound as a consequence of inter-particle collisions was developed to simulate acoustic agglomeration of solid particles. The method was validated against experimental data under a range of operational conditions. Based on this, the evolution of particle size distribution, agglomeration efficiency and agglomeration behavior with time was numerically investigated. The effect of restitution coefficient on the performance of acoustic agglomeration was also examined. The results show that the occurrence of acoustic agglomeration yields a significant decrease in the number concentration of small particles and a small increase in the number concentration of large particles. Moreover, a worse performance of acoustic agglomeration at a higher restitution coefficient is observed. This study demonstrated the power of the proposed method in simulating the acoustic agglomeration and provided a fundamental approach for investigating the dispersed gas-solid two-phase flows.

The kinetics of hydrophobic particle recovery by flotation declines with particle sizes below 20 µm. This work investigated a much-faster novel process, agglomeration, which utilises a high-internal-phase, water-in-oil emulsion as the binder. The effects of changes in the binder composition were studied by varying the emulsifier:oil ratio from 0.2:1 to 100:1, while maximizing the concentration of the dispersed aqueous phase within the binder. The oil consumption was assessed as an inferred film thickness surrounding the individual hydrophobic particles captured by the agglomeration. This film thickness decreased from 156 nm to as little as 4.5 nm by increasing the emulsifier concentration. This result is consistent with an increase in the emulsifier concentration causing an increase in the internal oil–water interfacial area within the binder. Water permeation into the binder, which reduces the viscous resistance to particle capture, decreased to an asymptotic level beyond a 5 wt% emulsifier concentration.

In the present work, the effect of batch high shear mixer (HSM) geometries is studied on nanoparticles deagglomeration and power draw in water and aqueous glycerol solutions. The kinetics and mechanism of cluster break-up are investigated for disintegrated (less number of larger stator holes), and mesh (higher number of smaller stator holes) heads at different rotor speeds. Mesh head geometry is found to be more efficient for fines generation than disintegrated head. A bi-modal particle size distribution and erosion as dominant break-up mechanism is found. The size of the smallest fines is found to be ≈30 nm for mesh head and ≈50 nm for the disintegrated head. Z-average ranges from 180 to 310 nm for given operating and process conditions. The power draw is more in mesh head and increased with an increase in viscosity. Further, power draw increased with an increase in viscosity of continuous media and rotor speed.

When the liquid is injected into a gas-solid fluidized bed, particles are easily wetted and prone to form liquid bridges among them, further leading to particle agglomeration. Wet particle agglomeration induced by liquid bridge is dynamic and unstable, and its evolution behavior affects the stable operation of fluidized bed reactor. Therefore, based on the formation-fragmentation competition mechanism, a wet agglomeration evolution model is proposed to predict the evolution characteristics of wet agglomerates under different conditions. Furthermore, a modified drag force model which attributed the effect of wet agglomerates into the drag force was established and applied to the CFD simulation of a fluidized bed with wet agglomerates under the liquid spray-evaporation equilibrium state. Meanwhile, the particle image velocimetry (PIV) technology is applied in this work, and it was found that the vertical solid velocity simulated by CFD is in good agreement with the PIV measurement results.

Particle size distribution is of utmost importance in spray dryers. Controlling the degree of agglomeration and coalescence is vital for manufacturing products that conform to the requirements. To predict the particle size accurately, this study presents a CFD model of particle agglomeration and droplet coalescence in a counter-current spray dryer. The stochastic agglomeration model was modified to incorporate more realistic droplet-droplet, droplet-particle and particle-particle interactions. Binary droplet collisions in the atomization zone of a spray dryer can result in either coalescence or separation. The boundaries between collision regimes were experimentally determined, and existing binary collision models were validated for water and milk concentrate to include the effect of viscous forces on collision outcomes. A high-speed camera was used to record the kinematic and physical properties of the droplets to identify various collision regimes. A suitable model was then incorporated in CFD simulations of spray drying. Using existing experimental data from a counter-current spray dryer with strong agglomeration, it is shown that the model can accurately predict product particle properties under different process conditions.

Acoustic wave separation (AWS) is gradually emerging as an effective method for continuous clarification of Chinese hamster ovary (CHO) cells. In this paper, we present a mechanistic model for AWS based on population balance model (PBM) along with acoustic agglomeration kernel to predict cell size distribution as a function of time. Results show that more than 90% CSE was achieved for flowrates 12.5 W. Experimental results confirm that the mechanistic model offers an accurate prediction of the CHO agglomeration process. Furthermore, a generalized correlation has been developed using the Buckingham Pi theorem between the dimensionless numbers composed of relevant process and geometrical variables to predict the agglomeration efficiency. To our knowledge, this is the first such study to model the agglomeration dynamics of CHO cells in acoustic standing waves.

Aided by self-agglomeration, a two-stage reduction process conducted at a higher temperature (600 °C) than the single-stage process resulted in an enhancement of the k constant to more than twice that of the single-stage process. A force balance model coupled with the reduction kinetics of Fe2O3 is first proposed to explain the self-agglomeration mechanism of iron nanoparticles (NPs) during the reduction. This force balance model successfully elucidates the reason for the prevention of defluidization via a two-stage fluidized bed method. At lower temperatures, there is a long stationary phase, which is of great importance in overcoming the sintering of agglomerates and in promoting the reduction reaction. At lower temperatures, the initial NPs first self-agglomerate into particles that are tens of microns in size; they then gradually agglomerate into larger particles (>100 μm) at higher temperatures. By contrast, the sudden growth of iron NP agglomerates causes sintering and defluidization in the single-stage fluidized bed method.

Gas-solid fluidized beds, by their nature, are associated with intense and frequent collisions of solid particles with each other and with the vessel wall, causing tribo-electrification. Accumulation of electrostatic charges in fluidized bed reactors can result in severe problems such as agglomeration, wall fouling, nuisance and hazardous discharge, all reducing the process performance and raising significant safety concerns. Tribo-charging of particles in fluidized beds has also been exploited in a number of useful applications. In this review, the characterization methods of electrostatics and the mechanisms of charge generation and distribution in fluidized beds are presented, followed by an account of the interplay between the hydrodynamics and electrostatic phenomena. Furthermore, techniques of electrostatic charge control in fluidized beds are reviewed, and applications of tribo-electrostatic fluidization systems are summarized. Finally, computational fluid dynamics simulations of the electrostatic effects on the hydrodynamic characteristics of fluidized beds are outlined.

Particle agglomeration can occur in heavy oil upgrading hydroprocessors due to the formation of carbonaceous mesophase, a secondary liquid phase which results from an increased rate of thermal cracking relative to the hydrogenation rate. A cold-flow slurry bubble column operated at atmospheric pressure with an internal diameter of 0.152 m was used to examine particle agglomeration behavior and the overall fluid dynamics in a gas-liquid-liquid-solid system. The experimental system consisted of biodiesel (organic continuous phase), aqueous glycerol solutions (immiscible secondary phase), glass beads and nitrogen. The impacts of the secondary liquid loading, secondary liquid viscosity, and particle diameter on the fluid dynamics were established. Particle agglomeration was studied by measuring the axial solid holdup profiles while varying the superficial gas velocity and secondary liquid loading. Enhanced particle agglomeration was observed when increasing the secondary liquid loading, based on the increased solid holdups at the bottom of the column. Three sparger designs (six-legged spider sparger, perforated plate and conical perforate plate) were compared. Following the glycerol addition, the spider sparger was less effective based on the particle sedimentation at lower liquid/solid ratios when compared to the other sparger designs. A Revolution Powder Analyzer (RPA) was also used for complementary measurements to examine whether operational difficulties due to particle agglomeration can be anticipated using a simplified system.

The population balance model is the common approach to simulation and prediction of the size distribution and other properties of particulate systems. Population balance models include any nucleation, growth, breakage and agglomeration mechanisms that are relevant to all industrial particulate processes. However, there are some limitations to many of the previous population balance model formulations for systems with agglomeration. Limitations include physically irrelevant and/or empirically based agglomeration kernels, difficulties in assessing the influence of process conditions (e.g. hydrodynamics, particulate physical properties), solution method efficiency for optimization and control applications, and loss of information on constituent particles. These limitations have prevented the use of population balance models to accurately predict and simulate agglomeration in suspension techniques such as spherical crystallization. To overcome these limitations, an extension of the concept of a coupled population balance model is presented for application in the simulation and optimization of a spherical crystallization system. A coupled population balance model formulation has been developed for a semi-batch, reverse addition, anti-solvent crystallization system with agglomeration. The system includes nucleation and growth of the primary crystals and subsequent agglomeration. The advantages presented by a coupled population balance model formulation include the ability to optimize for specific primary and agglomerate sizes. This presents an opportunity to find optimal operating conditions that meet both bioavailability and manufacturability demands.

In crystallization a high effort for optimization and process control is made to produce crystalline batches with required specifications, like purity or crystal size distribution (CSD). However, the final product is affected by solid-liquid separation and drying also so that for an efficient process optimization all unit operations have to be considered. Especially if a high temperature dependency of solubility exists, a special attention should be paid to the drying process. Previous studies show that classical static drying is not the best choice in this case, since long contact time between crystals lead to uncontrolled agglomeration. This event need not necessarily result in different characteristics of the crystal size distribution (CSD), but reduced purity. Therefore we investigate systematically two different drying methods – fluid bed and rotary tube drying – concerning behavior of CSD and agglomeration degree in dependence of drying parameters used. Additionally we show that with the aid of the so-called agglomeration degree distribution, which we developed before, a deeper understanding of crystal agglomeration within the CSD is gained. As model system l -alanine/water is used. The results show that the product quality designed by cooling crystallization cannot be entirely maintained, but the formation of agglomerates is reduced in case of both methods.

The micromechanical interactions between hydrate particles and water droplets play an important role in determining hydrate agglomeration, which is a crucial cause of hydrate blockages/bedding in deepwater gas-oil flowlines. The amount of hydrate converted from water droplet during the interaction between hydrate particles and water droplets led to the reduction in liquid volume, which would significantly affect the interaction forces. The existing classic pendular liquid bridge model with fixed liquid volume is not adequate for this unique case. In this study, a modified pendular liquid bridge model has been developed by considering the conversion of hydrate from liquid droplet. Parabolic approximation and a rupture criterion are proposed to determine the liquid bridge profile and rupture distance, respectively. On the basis of experimental observations, the capillary force model between two plates is applied to predict the interaction forces. It is found that the evolution of the profile of bridge (including the hydrate coating section and the pure liquid section) can be predicted with a satisfactory accuracy. Subsequently, it ensures the good agreement between the measured and predicted interaction forces. The profile of the liquid bridge becomes unstable just before the rupture of the liquid bridge, resulting in a large deviation between the experimental results and the model predictions. The proposed model can be successfully used for investigating the effects of different factors, such as contact area and hydrate formation rate, on the interaction behavior/forces, which could help to provide new and critical insights into the hydrate agglomeration process. The developed models contribute a significant progress in existing models development on hydrate agglomeration, and thus could provide a more accurate evaluation of hydrate formation risk in gas-oil flowlines.

Agglomeration of ultrafine hydrophobic particles can be performed using high-internal-phase (HIP), water-in-oil emulsions as the binder. The ultrafast particle recovery achieved using these emulsions can be attributed to the presence of thin, permeable oil films as its organic phase. The internal aqueous phase of these emulsions contains salt, which drives water absorption through these permeable oil films during agglomeration, greatly reducing the effects of the lubrication resistance to particle collision, adhesion and hence agglomeration. In this study, the water permeation was quantified by studying the growth of cylindrical rivulets placed in fresh water. The rivulet diameter increased approximately with the square root of time, suggesting a diffusion-limited process. Water transport rates increased about three fold with increasing internal salt concentration of the emulsion from 0.5 wt% to 10 wt%.

In froth flotation, fine hydrophobic particles selectively attach to the surface of air bubbles, in turn rising through the liquid and then more slowly as part of the foam that emerges from the system. This paper proposes a paradigm shift from the traditional use of air bubbles as the hydrophobic separation medium to potentially a far more powerful version that utilises a novel hydrophobic binder, a concentrated water in oil emulsion. We show empirically the oil consumption increases linearly with the specific surface area of the particles, with an average oil film thickness of 178 nm for fine coal particles, well below the level required for conventional oil agglomeration. The new approach is ultrafast, achieving agglomeration within seconds under batch conditions, with strong selectivity. We also show for the first time continuous steady state agglomeration can be achieved by simply passing the feed and binder suspensions through a partially closed ball valve. Here the residence time through the valve is less than 0.1 s. And, by subjecting the agglomerated product to further shear, the emulsion is inverted, releasing bound water. Pressure driven filtration then delivers remarkably low product moistures. Operational aspects of this new technology are discussed.

Potential flow assurance problems in oil and gas pipelines related to gas hydrates have traditionally been resolved by implementing avoidance strategies, such as water removal, insulation, and injection of thermodynamic inhibitors. As a means of lowering development and operational costs in the industry, hydrate management is becoming a more viable approach, a strategy which seeks to minimize the risk of plugging conditions using methods that allow transportability of hydrate slurries with the hydrocarbon production fluids. A thorough understanding of how hydrates form and behave in different multiphase systems is essential for safe implementation of various hydrate management methods. In order to increase insight into the different processes leading to hydrate plug conditions, a number of experiments have been performed using a visual rocking cell to measure and observe the various stages of hydrate formation, deposition and accumulation during continuous mixing and motion induced by the oscillation of the rocking cell. The fluid combinations studied in these experiments considered oil+water+gas and condensate+water+gas. The effects of added monoethylene glycol (MEG) and a model anti-agglomerant (AA) were also studied in some of the experiments. The experiments showed that hydrate deposition is prevalent in all systems studies, but it was observed that there was a higher tendency to deposit on surfaces exposed to condensate or the gas phase than on oil wetted surfaces. Hydrate growth, agglomeration and bedding were observed in the experiments. Highly porous hydrate deposits formed in conditions with a large temperature gradient between the bulk and the surface, and high subcooling conditions, then suffering from sloughing due to the shear of the fluids on the deposit. Hydrate formation in an experiment with mineral oil, 30% water cut and anti-agglomerant resulted in transportable hydrate slurry. Both the condensate and mineral oil tested were non-emulsifying, but shear-stabilized dispersion of the liquid phases was created prior to hydrate formation by mixing induced by the motion of the cell. Interestingly, the dispersion appeared to completely phase-separate, momentarily, upon hydrate formation onset.

Hydrates of light hydrocarbons are frequently formed during the subsea petroleum production. These crystalline ice-like solids may accumulate at concentrations sensitive from the flow assurance point of view, increasing the overall pumping costs and imposing sufficient risk of the pipe blockage. Modern trend in the assessment of hydrate-related risks is the development of numerical models of multiphase flows laden by hydrates. The present paper describes a computational fluid dynamic (CFD) model capable to simulate turbulent slurry of oil, water and gas hydrates. The population balance technique (PBM) coupled with CFD enables to predict such details of the process as the formation of hydrate phase, agglomeration of formed solids and granular interactions within the hydrate phase. The simulation results, validated with experimental data in terms of the slurry rheology, highlight flow patterns for a pipe system typical in oil industry. The model is in addition compared to the hydrate kinetics model from Colorado School of Mines (CSMHyK).

Continuous crystallization of proteins at short residence time and high supersaturation level is attractive in terms of the space-time yield and production efficiency. Nevertheless, it is rarely pursued due to its less-than-desirable crystal size distribution (CSD) characterized by the abundance of small crystals due to high nucleation rate. The small crystals were prone to random agglomeration, resulting in poor crystals’ residence time distribution, hence low CSD's reproducibility. Herein we developed a segmented slug flow crystallizer (SFC) design operated at short residence time (

Pore resolved simulations are performed to study the mean flow characteristics in porous media formed by random distribution of impermeable and permeable particles. Permeability of the medium and reactive surface areas are analyzed for a range of solid fractions, packing patterns and flow rates. The results are compared with models presented in the literature and with simulations of flow through ordered packings. Permeability of a randomly packed porous medium is found to depend on the heterogeneous distribution of porosity which results from particle agglomeration and is typically ignored by simpler models. The reactive surface area is dependent on the flow field at pore scale in addition to geometry. Moreover, the wake formed downstream of the particles at moderately high flow rates also affects the reactive surface area. Similar analysis is also performed by assuming particles to be permeable to account for dual scale porosity relevant for various engineering applications.

Particle agglomeration is a common phenomenon in a large number of industries. The presence of a liquid bridge between colliding wet particles is the principal source of agglomeration in such processes as wet granulation and thermal conversion of fuels in a high temperature gas–solid fluidized bed reactor. A comprehensive agglomeration model for wet particles must incorporate both liquid capillary and viscous contributions, as well as the liquid bridge volume effect. Darabi et al. (Chem. Eng. Sci. 64 (2009) 1868–1876) developed a coalescence model for a binary collision of identical wet particles. The model provides reasonable results for wet particles with a low viscosity liquid layer, i.e., under capillary limiting conditions. However, application of this model for collisions of particles coated with a high viscosity liquid layer, i.e., under viscous limiting conditions, results in unreliable outcomes. The current study presents an improved version of Darabi et al.’s model, which (i) corrects viscous energy loss equations in the original model, thus providing reasonable outcomes under both capillary and viscous limiting conditions with highly wetting systems, (ii) accounts for the contribution of capillary force in the approach stage when estimating the initial particle speed for the separation stage, (iii) accounts for the collision of particles having different initial speeds with respect to the laboratory reference frame, (iv) employs a more reliable correlation to estimate the liquid bridge rupture distance, and (v) evaluates the collision outcome in three steps based on the values of kinetic energy parameters. We compared experimental particle-plate collision data with the predictions from agglomeration models and it showed that considering the liquid bridge volume effect and the liquid bridge rupture distance can be essential for the accurate prediction of collision outcomes. The results of sensitivity analyses with the improved agglomeration model revealed that the agglomeration outcome under capillary limiting conditions is most sensitive to the thickness of the coating layer, particle inertia, liquid surface tension, and dry restitution coefficient. Under viscous limiting conditions, proper measurement or estimation of the thickness of the coating layer, liquid viscosity, asperity height, and particle inertia are critical for evaluating the particle collision outcome.

A strategy for rationally selecting solvent systems of spherical agglomeration is proposed for the first time. It emphasizes that followed by crystallization and phase separation, spherical agglomeration can be achieved only when the solvent system can induce wetting and adhesion which are estimated by the Lifshitz-van der Waals acid-base approach. Especially for wettability of bridging liquid, solvent and anti-solvent, the strategy indicates that it is of great importance for a successful solvent system. Valid systems therefore are selected in the first place while invalid ones are removed effectively according to the wettability difference. Based on the strategy, valid solvent systems for cefotaxime sodium and benzoic acid spherical agglomeration were selected successfully from 720 and 2184 systems respectively without missing the reported systems. The strategy also predicted no valid systems for potassium chloride, maltitol and trisodium phosphate in the given solvent combinations, which agreed with the experiment results.

An understanding of agglomeration in multiphase flows by means of capillary bridges is important for various technical applications. One of the key parameters that describe the process is the probability of agglomeration, usually termed collision efficiency. In the present paper we study the collision efficiency by using Eulerian–Lagrangian computational fluid dynamics (CFD) modelling. The analysis evaluates the dependence of the collision efficiency on various dimensionless parameters that describe the process, including, among others: Reynolds number, bridge capillary number and wetting angle. In addition we further develop a theoretical expression for a priori engineering estimates of the collision efficiency. This expression is validated against numerical, theoretical and experimental results.

Beside purity, crystalline products are characterized by median crystal size and crystal size distribution (CSD) mainly, but usually no information about agglomerates is given. To overcome this lack the agglomeration degree distribution (AgD) of crystalline product batches was introduced by using image analysis and Discriminant Factorial Analysis (DFA). As material systems l -alanine/water and adipic acid/water were chosen. We showed that a so-called reference discriminant function can be used to characterize the AgD of crystalline particles of the same material system, but produced under different conditions and in different batches. The AgD determined, as well as the corresponding characteristic values, allow a quantitative characterization and comparison of crystalline products with different morphology also. Moreover, the AgD allows general statements whether crystals tend to agglomerate and whether agglomeration takes place during crystallization or further downstream processes.

The present paper proposes a model for agglomeration of solid particles covered with a liquid flocculant. Rather then being derived by time-consuming numerical simulations of collision dynamics as the particles are in contact, the model is based on analytic relations: thus the computational cost is significantly reduced. The model was implemented into a computer code simulating a set of particles using the Eulerian–Lagrangian approach, that is, it forms an attractive alternative to Distinct Element Modelling (DEM). Moreover numerical simulations in a shear-flow system were done and the results were compared with a theoretical model.

An online measurement setup is presented to measure the agglomeration processes occurring during batch cooling crystallization of adipic acid from aqueous solution. Quantification of agglomeration is enabled by image analysis and particle classification using a highly accurate artificial neural network. This model system has a strong tendency to agglomerate during crystallization and therefore a strong tendency to aggregate, too. Aggregates and agglomerates are visually indistinguishable though and, therefore, falsify the measurement by aggregates superimposing agglomerates. To counteract this phenomenon during measurement, it is tested whether ultrasonic irradiation can be used to disaggregate the particles before being measured. It was found that ultrasonic treatment of the sample at various frequencies and input powers resulted in particle breakage (low frequency, high input power) or disaggregation of very large particles that would lead to clogging of the setup. Complete disaggregation of the sample was not feasible as the frequency and input power variation of the ultrasonic setup has shown. However, it was concluded that continuous treatment with ultrasonic irradiation might be a way to reduce impurity inclusions within the agglomerates due to the continuous destruction of the particles.

Initial wetting of crystals by binder droplets is a key rate process in spherical agglomeration, however there are no models to predict the kinetics and formation of agglomerate nuclei. Two new mathematical models are introduced for agglomerate nucleation by an immersion mechanism; immersion rate limited model and collision rate limited model. The agglomerate nucleation number developed in this work predicts different regimes; immersion rate limited, collision rate limited and intermediate. In an immersion rate limited regime, agglomerate size increases with square root of time. In a collision rate limited regime, size increases linearly with time if the bulk crystal volume fraction, φPb, is constant, or with an exponential decay rate for batch crystallisation with decreasing φPb. The timescale for nucleation is less than ten minutes for a broad range of conditions, significantly less than most crystallisation timescales. These models have great promise for population balance modelling and spherical agglomeration optimisation.

A surfactant is used to reduce the agglomeration of iron catalysts during the synthesis of carbon nanotubes (CNTs) and their assembly by chemical vapor deposition. The agglomeration of iron catalysts during synthesis affects the CNT diameter and number of CNT walls. Also, the agglomerated catalysts within the CNT assembly indicate an inefficient reaction of catalysts during the CNT synthesis. In this study, the non-ionic surfactant, polysorbate, is added to the CNT synthesis solution in varying amounts from 0 to 3.0 wt%. The effects of the surfactant are closely related to the crystalline perfection, diameter, and number of walls of CNTs synthesized under different concentrations of the surfactant. Here, 1.0 wt% surfactant is the most favorable concentration to achieve crystalline perfection. When the surfactant is added at more than 1.0 wt%, amorphous carbons increase, although the agglomeration of iron catalysts continues to decrease. The simple addition of this surfactant can facilitate the development of the CNT assemblies that have high mechanical and electrical properties.

The ability to coat particles with a uniform film of a desired thickness is a key requirement in many industries such as pharmaceuticals, food processing and chemical manufacture. Controlling the distribution of the coating material during coating operations is essential for ensuring factors such as homogeneity in the coating thickness and prevention of agglomeration, which are common requirements in most types of coating applications. Despite the significant industrial and commercial importance of spray coating, few numerical models have been developed to model this process. Here, we detail the first computational model to incorporate particles, coating spray droplets and gas flow, dynamically interacting over an entire coating system. Particles and gas flow are modelled using a coupled DEM-CFD method, and spray droplets are modelled as individual Stokesian particles within the gas flow field. The coating model uses a newly developed method for mapping the coating coverage over each particle, based on a spherical harmonic formulation. This allows the coating to be evaluated at an intra-particle level on each individual particle. Distribution functions, such as coating quality and volume deposition, can be easily determined for the entire system at the inter-particle level. The method is applied to a test case of an fluidised bed spray coater, and the effects of varying geometry and system operating conditions are evaluated in terms of the coating coverage and quality. The method enables factors affecting coating distribution in such systems to be investigated and understood, and potential design optimisations to be evaluated.

Good fluidization was generally established when fine and light catalyst particles were fluidized. However, a large decrease in the fluidization quality was observed when two types of particles composed of different materials were mixed and fluidized, even though good fluidization was established with each type of particle alones. Formation of gas channeling was observed for binary mixtures of particles in a two-dimensional (2D) fluidized bed. In a three-dimensional (3D) bed, measurement of the pressure fluctuation showed that abnormal fluidization occurred due to mixing of the particles. In addition, measurements of the minimum fluidization velocity and emulsion phase expansion indicated that particle agglomeration occurred. The phenomena observed in the fluidized bed with the mixture of particles are quite similar to those observed for Group C particles. On the basis of the direction of the electrostatic charge, the particles were separated using electrodes installed above the fluidized bed surface. It was confirmed that one type of particle was positively charged and the other was negatively charged. Thus, it was concluded from these results that the bipolar charging caused by the difference in chemical composition of the particles induced the agglomeration phenomena and poor fluidization.

The present work describes the study of a continuous fluidized bed granulation process involving a solution spray as binder. The study entails the employment of a flux number as an extension to an existing flowsheet model (SolidSim). The agglomeration kernel used here is based on the Equi-Kinetic Energy principle and has been derived and proven for fluidized bed granulation processes using reactive and melt binders, but not for solution binders. This kernel has been correlated in the past to the flux number, which contains the main characteristics of the fluidized solids and the spray-on, such as the binder flux, the particle density and the fluidization gas velocity, in combination with population balances. This correlation was derived for a few liquid/solids systems only, notably small scale batch processes with melt binders, and a first attempt to the generalization of this correlation is shown here by its application to a completely different system, notably a continuous large pilot using a solution binder. The validated SolidSim flowsheet is useful for parameter variations and sensitivity analyses, showing a strong effect of the flux number exponent and the cut size of the screen on the product size distribution and recycle rate. Such quantitative relations between process conditions and product quality allow for significant savings in cost and time for process and product development and optimization.

The fluid, particle and collision dynamics inside a fluidised bed granulator are described in detail using coupled DEM-CFD simulations. Three different granulator configurations are compared in terms of their particle–particle interactions, using criteria such as the average particle velocity, angular velocity, average particle–particle and particle–wall collision velocity and the collision frequency. From these measures the agglomeration probability, the breakage and growth rate and the agglomerate strength can be assessed. Depending on the required product properties the results highlight advantages of specific granulator geometries. The top-spray granulator is the simplest device and is often used for large-scale granulation processes in the food and fertilizer industries, but wetting intensity and growth rate are relatively low compared to the other configurations investigated in this work. The spouted bed has the most intensive gas–liquid–solid contact and produces compact and dense agglomerates. Both the agglomeration probability and the breakage rate are very high. The Wurster-coater achieves the fastest growth rate for the applied process conditions and is characterised by homogeneous cyclic particle wetting and a low breakage rate. The simulation results are confirmed by lab scale batch agglomeration experiments using maltodextrin DE21, a model substance for amorphous food powders. It can be concluded that the DEM model offers large potential for process intensification by investigating the effect of geometry modifications on the fluid and particle dynamics and on the homogeneity of particle wetting. This is shown exemplarily for fluidised bed agglomeration.

Ceria nanoparticles are of interest owing to their interesting properties like photo-absorption and ability to absorb oxygen. Several batch synthesis routes exist however these methods often lack particle size control, result in agglomeration and typically require high temperature processing. In this paper, a continuous flow microchannel synthesis method based on precipitation is demonstrated with a static T-mixer, showing improved control over agglomeration through the manipulation of flow parameters. Three flow rates were chosen representing different mixing modes and computational fluid dynamics simulations were used to visualize flow conditions. The nanoparticles were characterized for composition, size and size distribution using transmission electron microscopy, wave dispersive spectroscopy and x-ray diffraction. The engulfment flow condition, corresponding to the highest Reynolds number in this study, gave the narrowest particle size distribution (15.0±4.7 nm) and best compositional uniformity.

Unwanted agglomeration, especially in highly concentrated submicron and nano- particle suspensions, can rapidly change the physical product properties of produced particles. Knowledge of the agglomeration kinetics and mechanisms is crucial to achieve control over this secondary sub-process of particle formation. To obtain this kinetics a new measurement method is presented to monitor the shear induced agglomeration behavior of submicron barium sulfate particles while passing a narrow pipe. The residence time, i.e. agglomeration time was varied with the pipe length at constant suspension flow rate. Before measurement, the particle suspension was stabilized by rapid mixing with a dispersing agent. Based on measured agglomerate size distribution, the agglomeration behavior is discussed in terms of non-turbulent flow in a narrow pipe. Barium sulfate agglomeration in a simple shear flow could be monitored on submicron scale. The effect of steric stabilization to reduce or prevent agglomeration was shown and quantified. The controlling mechanism for agglomeration in a pipe is identified based on the hydrodynamic characteristics. The adhesion probability was found to decrease with agglomerate size.

Pulverized coal is important for gasification to convert carbonaceous fuel to gaseous products with a usable heating value; it is cohesive with a poor flowability, typical of Group C powders. The fluidization behaviors of pulverized coal in a fluidization column and in a Perspex hopper were investigated. The pulverized coal was extremely hard to be fluidized and its experimental incipient fluidizing velocity was much larger than expected. The discharge of pulverized coal from a carbon steel hopper under several aeration rates was studied. Four flow regions were developed, and the major phenomena that occurred during the discharge were analyzed. The mechanism of aerated discharge was probably the integration of various effects, all of which were confirmed by the experimental data.

Fluid flows with cohesive particles are present in oil industry (e.g. natural gas/oil with hydrates, wax or asphaltenes), medicine (e.g. blood cells), nano- and ferro-fluidic applications (e.g. fluids with nanoparticles subject to the van der Waals and electrostatic interactions) and even in astrophysics (e.g. grains in planetary rings). Such flows may lead to formation of agglomerates that, for example in pipelines, may result in unwanted phenomena such as formation of deposits. The main process parameter governing this is the “collision efficiency”, which is the ratio of the number of collisions resulting in agglomeration to the total number of collisions. This is commonly considered to depend on the relativemagnitudes of attractive and repulsive interactions during a collision. The effect of the particles' mechanical properties on the agglomeration efficiency has, however, not yet been studied. In this paper the agglomeration efficiency is studied as a function of inter-particle friction, stiffness, density and volume fraction by numerical simulation. By running direct numerical simulations (DNS) with Lagrangian particle tracking of a shear flow laden with solid particles, the parameters influencing the agglomeration efficiency are demonstrated and their effects quantified. Finally, an expression that relates the collision efficiency to the salient dimensionless physical parameters is proposed.

There have been significant advances in the understanding of wet granulation processes. Foam granulation is the latest development and an emerging area of interest for pharmaceutical manufacturing. Single foam penetration experiments were carried out on static powder beds, followed by short-nucleation experiments (where nuclei are formed by a nucleation-only mechanism) and full foam granulation experiments (where nucleation, growth and breakage are occurring simultaneously). All experiments were performed with lactose monohydrate powder using a 5 L high shear mixer–granulator. The foam penetration/dispersion behaviour was examined and the granule size distributions were investigated as a function of foam quality (83–97% FQ), impeller speed (105–515 rpm) and wet massing period (0–4 min). Nucleation in foam granulation is postulated to undergo either “foam drainage” or “mechanical dispersion” controlled mechanisms. For “foam drainage” mechanism, the foam penetrates the powder bed to form coarse and broad granule size distributions. For “mechanical dispersion” mechanism, the wetting and nucleation conditions are governed by the powder mixing conditions and similar granule size distributions are produced. Regardless of the mechanism, the initial wetting and nucleation behaviour controls the initial nuclei size distribution, and this initial distribution is retained in the final granule size distribution. This work demonstrated the critical importance of the nucleation and binder distribution in determining the granule size distributions for foam granulation process.

The effect of the primary particle porosity during the formation of agglomerates in spray fluidized beds is presented in this study. The method is based on the single micro-interactions occurring within the fluidized bed such as inter-particle collisions, droplet spread on the particle surface, aging of the deposited droplets and particle coalescence. The porous character of the particles is expected to directly affect the aging process of the deposited binder layer by penetration into the pores of the substrate. The droplet penetration process is experimentally analyzed by single droplet deposition on spherical, porous alumina particles. The results indicate that the penetration process is mainly governed by the viscosity of the liquid and that at relatively low viscosities, droplet penetration is fast. For highly viscous liquids, the penetration velocity slows down and an additional mechanism, namely drying becomes important. A combined imbibition–drying model is developed and included into a comprehensive stochastic agglomeration model that allows the simulation of agglomerate formation in a batch process. Lab-scale agglomeration experiments with porous and non-porous particles are carried out in an attempt to validate the general tendencies predicted by the main agglomeration model. The results show that the agglomeration rate for porous particles is significantly reduced due to the losses of deposited droplets into the pores of the primary particles; this tendency is much more pronounced at low binder viscosities.

Agglomeration occurring during fluidized bed incineration not only causes the fluidization characteristics to change and the system to unexpectedly shut down, but it also leads to the generation of secondary pollutants such as heavy metals. Accordingly, the aim of this study is to estimate the mechanisms of particle agglomeration in fluidized bed incineration. Experiments are carried out, including a pilot-scale fluidized bed test and a laboratory furnace test. Thermodynamic equilibrium simulations are also considered. In addition, the speciation of agglomerates is determined from the analysis of the laboratory furnace using X-ray diffraction (XRD), electron spectroscopy for chemical analysis (ESCA), and field emission scanning electron microscopy/energy dispersive spectrometry (FESEM–EDS). The experimental results reveal that higher concentrations of metals are emitted when defluidization occurs. When we compare the results from two different agglomeration inhibitors, the Al-based additive is shown to achieve higher levels of inhibition versus a Ca-based additive. In addition, the results from our simulation illustrate that the heavy metals Pb, Cr, and Cd have different affinities with Na, Si, Al, and Ca, which causes different emission behaviors during the agglomeration inhibition processes. Good agreement is observed among the experimental data, the agglomerate characterization, and the thermodynamic equilibrium simulations. The primary mechanisms of both particle agglomeration and the emission of heavy metals are also established.

The crystallization kinetics of aluminium hydroxide from the sodium aluminate solution reacted with sodium bicarbonate were systematically investigated in a steady-state MSMPR (mixed-suspension mixed-product removal) crystallizer for the first time, and the expressions of the nucleation rate, growth rate and the agglomeration kernel of aluminium hydroxide were successfully regressed. The aluminium hydroxide particles precipitated from the reactive system are identified as gibbsite by XRD and SEM examinations. The volume growth rate order of gibbsite with respect to the relative supersaturation of the solution is above the linear growth rate order, and the spiral growth mechanism for the growth of the basal face of gibbsite in the reactive system was further identified by the growth rate, morphology analysis as well as the calculated surface entropy factor. The secondary nucleation rate of gibbsite from the reactive system is three to four orders of magnitude larger than that from seeded process reported in the only available literature reference. The agglomeration kernel of gibbsite in the reactive system increases linearly with growth rate and residence time, and the positive order about 0.55 of magma density is thoroughly different from the negative order of magma density for gibbsite agglomeration in seeded process presented in the literature.

When performing Eulerian-Lagrangian simulations of particle-fluid flows collisions between the particles need to be accounted for. One of the methods used for this is the hard-sphere model. This model, however, does not take into account cohesive forces between the particles, and for this reason it is not able to simulate many aspects of real flows, such as the formation of agglomerates. There have been some attempts in literature to treat cohesive forces in simulations of particulate flows but none of these methods were actually implemented directly into the hard-sphere model but rather have been solved separately as a part of the numerical scheme. In this paper we show how the standard hard-sphere model may be extended to include these important interactions in an efficient and proper way. The extended model is presented in detail and some numerical results are shown.

A stochastic model that describes agglomerate growth during fluidized bed spray agglomeration is presented and numerically solved by constant volume Monte Carlo method. The methodology overcomes the difficulties of solving multivariate population balance equations and includes continuous binder addition and drying. Agglomerate formation is treated as a complex combination of consecutive and parallel micro-mechanisms. Due to the discrete nature of the approach, the individual role of the micro-mechanisms on the agglomeration behavior can be analyzed. The results suggest that the droplet capture mechanism governs the agglomeration speed while the maximum agglomerate diameter is ruled by the equilibrium reached between coalescence, rebound and breakage. The mechanism of deposited binder drying is found to play a negligible role on agglomerate formation because of an extremely rapid droplet consumption. The main process variables affecting each micro-mechanism have been identified showing that the liquid spraying rate affects directly the droplet capture mechanism whereas binder properties influence mainly the agglomeration and rebound interactions. The model presented in this study is able to predict qualitatively the experimentally observed response of the system as well as the general shape of the agglomerate size distribution under the variation of several process parameters, demonstrating the potential of the discrete micro-level approach.

Fluidized bed agglomeration is used to stabilize particulate mixtures and reduce dust emissions. This technology is applied to a variety of production processes for the pharmaceutical, chemical, fertilizer and food industries. In most of these applications, agglomerate stability is an essential criterion. Agglomerates and granules that do not conform to size and shape specifications may create problems in downstream processes, such as tableting, thus compromising process efficiency and product quality. When an agglomerate is formed in a fluidized bed, it can grow by incorporating other bed particles, split into smaller fragments, or be eroded by fluidized bed solids. The objective of the present study is to determine the critical agglomerate liquid content at which the rates of agglomerate growth and shrinkage are balanced when artificial agglomerates made from glass beads and water are introduced into a fluidized bed. This study examined the effects of agglomerate size, agglomerate density, liquid viscosity, binder concentration, and fluidizing gas velocity on the critical initial liquid content. This study found that small agglomerates and low density agglomerates displayed higher critical initial moisture contents. When the viscosity was increased by using sugar solutions, agglomerates were very stable and had very low critical initial moisture contents. The study also found that as the superficial gas velocity increased, the agglomerates started to fragment, rather than erode.

In fluid bed processing in the chemical, food or pharmaceutical industries, pneumatic nozzles are typically used to convert binder or coating liquids into droplets. Producing fine droplets from liquids in a gas phase is termed atomisation, and it involves complex phenomena which are not yet fully understood. This paper provides a systematic and up-to-date review of two-fluid nozzle designs and principles together with a presentation of nozzle fundamentals introducing basic nozzle theory and thermodynamics. Correlations for the prediction of mean droplet diameters are reviewed, compared and accompanied by a discussion of their use.

In this paper a multidimensional model for binder granulation is presented. The particles undergo different transformations such as coalescence, compaction, and breakage. Further chemical reaction in the granules is taken into account in order to incorporate binder solidification which is observed to be a significant transformation in many industrial applications. The equations of the model framework are solved numerically with a direct simulation Monte Carlo (DSMC) algorithm. In addition to the comparison between experiment and simulation, the model framework also enables the study of critical parameters in binder granulation such as reaction rate (solidification of binder) and size of the added binder droplets, which demonstrates its promising potential.