The effect of primary particle surface energy on agglomeration rate in fluidised bed wet granulation

The effect of primary particle surface wettability by a binder solution on the rate of agglomeration in a fluid-bed top-spray granulation process was investigated. A model system consisting of hydrophilic and hydrophobic spherical primary particles with a narrow size distribution, and an aqueous solution of hydroxy propyl-cellulose (HPC) as binder, was used. The surface energy of the primary particles was measured by inverse gas chromatography (IGC) and their wettability was characterised by static and dynamic contact angle. Granulation was carried out in a desktop fluid-bed granulator and the resulting granule size distribution and granule microstructure were analysed. The hydrophobic particles gave a wider granule size distribution (larger maximum granule size) than hydrophilic ones under otherwise identical conditions, and the granules were notably rounder and more compact. However, the fraction of un-granulated fines was also higher in the case of hydrophobic primary particles. SEM analysis of granule microstructure revealed that the hydrophilic particles were coated by the binder solution, which left a smaller amount of binder available to form bonds at particle contacts. On the other hand, all of the binder was found to form solid bridges in the case of hydrophobic primary particles. A population balance model was used to explain the observed granulation behaviour.     

The effect of granule microstructure on dissolution rate

The relationship between the microstructure of granules and their dissolution rate has been investigated. Granules consisting of mannitol primary particles and PVP aqueous binder have been prepared by top-spray fluid-bed granulation, and granules consisting of sucrose primary particles and PEG binder by in-situ melt fluid-bed granulation. Granule microstructure has been systematically varied by manipulating the primary particle size distribution and the binder content in each case. In both cases granule porosity was found to be a decreasing function of binder content and a minimum of porosity as function of the fine/coarse primary particle mixing ratio has been observed, in line with theoretical expectations. Granule microstructures have been analysed using X-ray computed micro-tomography and compared with three-dimensional “virtual granules” generated by a computer simulation of the agglomeration process. The dissolution rate of granules has then been measured. While porosity was found to have a strong effect on the dissolution rate of mannitol granules, the dissolution rate was found to be practically independent of porosity in the case of sucrose granules. The formulation-microstructure and microstructure-dissolution correlations established in course of this work are in line with previous computer simulation results and form part of a computer-aided granule design methodology.     

The effect of powder size on induction behaviour and binder distribution during high shear melt agglomeration of calcium carbonate

Growth mechanisms in high shear mixer granulation were observed over a wide range of particle size and liquid-to-solid (L/S) ratio. The materials used were calcium carbonate (CaCO₃; size fractions in the range 1.5 to 85 μm) with a binder of polyethylene glycol 6000 (PEG 6k). The binder, solid at room temperature, was added by the “melt-in” method. A 10 L vertical-axis granulator was used, with a chopper and a four-bladed impeller.

The mean granule size and granule size distribution were measured at regular intervals during the agglomeration process by careful sampling and sieving. The uniformity of binder distribution among the granules was also measured.

The growth behaviours of each grade of primary particles were classified and compared. An induction type mechanism was observed with an initial period of slow growth in mean particle size that lasted 2 to 3 min (the induction period). This was followed by a short rapid growth phase lasting 1 to 2 min. The final stage was a plateau of more or less zero growth. Interestingly, the end of the induction period and the onset of rapid growth corresponded to a change in the granule size distribution from bimodal to monomodal and a similar change in the distribution of binder. Induction period growth rate tended to be lower for granules of finer particles, but these grew more rapidly during the rapid growth stage and produced larger granules than the coarser primary particles.

The liquid-to-solid (L/S) ratio had a significant effect on the growth rate during the rapid growth stage but a minor effect on the granule size distribution and binder distribution. Primary particle size had a significant effect on the final average size of granules, the growth rate during the rapid growth stage and the distribution of granule size and binder.     

Influence of liquid binder dispersion on agglomeration in an intensive mixer

In industrial scale mixer granulation, liquid binder is usually sprayed onto the agitated powder bed by means of a nozzle in order to enhance the agglomeration process. The early stage of this process, where granule nuclei are formed and grow, is not well understood. As it is desirable to model the agglomeration state right from the beginning of the process for the purposes of control and modeling, this nucleation step is therefore an important field of interest.To investigate the influence of binder droplet size on the nucleation stage of the agglomeration process, experiments were carried out with lactose and water in an intensive mixer. Water was sprayed in to the mixer with different nozzles to vary the size of the produced droplets. As a comparison, water was also directly poured into the turning mixer. Samples of the produced granules were taken at specific time intervals and analysed for size and water content. As the experiments were focused on examining short granulation times, the first samples were taken after only half of the water was added.Particle size distribution and liquid distribution in the wet granule samples were analyzed. It was found, that the droplet size of the binder liquid has great influence on agglomerate size and binder distribution at short mixing times, with increasing time, the mechanical stresses acting in the mixer becomes more and more dominating in the process. Preliminary comparisons are also carried out with single drop penetration tests in an attempt to correlate drop size to penetration time and also to produced granule size.In conclusion this paper studies the effect of different drop size conditions and subsequent spray flux on the behaviour of the nucleation and the early stages of the agglomeration process. The context of these findings for agglomeration in an intensive mixer is examined.     

Particle growth mechanisms in fluidised bed granulation—I: The effect of process variables

Two types of product granule have been identified in an experimental study of batch fluidised bed granulation; agglomerates which consist of two or more, and usually several, initial particles; and layered granules, which consist of single primary particles with dried feed material adhering to the surface. Increasing the excess fluidising gas velocity, in the range U-U_mf = 0.15-0.525 m s^?1, resulted in decreased particle growth rates and, depending upon the binder material, a change in product morphology from agglomerates to layered granules. Similar changes resulted from increasing the mean size of the starting material. Thus, a mechanism of particle growth is proposed in which the strength of inter-particle bridges and the extent of fluid drag and inertial forces on particles determine the equilibrium granule form and size. However, if the initial particles are porous this mechanism may break down since liquid may enter intra-particle pores and be unavailable for the initial formation of liquid bonds.     

Computer simulation of granule microstructure formation

The diagenesis (porous microstructure evolution) of granules formed by a layering growth mechanism in a wet granulation process has been modelled. The model includes the packing of primary particles with a given size and shape distribution, and the deposition, spreading, and solidification of binder droplets within the growing granule. The dependence of granule porosity on the binder/solids ratio, primary particle size and morphology, and the rates of binder spreading and solidification has been investigated. The results are presented in the form of structure maps relating volume-averaged microstructure parameters with dimensionless groups including the ratio of droplet spreading and solidification times and the mean time between particle collisions. These graphs can guide the selection of process operating conditions or formulation ingredient properties required to obtain a particular granule microstructure.     

Hot melt granulation of coarse pharmaceutical powders in a spouted bed

The hot melt granulation of a coarse pharmaceutical powder in a top spray spouted bed is described. The substrate was lactose-polyvinylpyrrolidone particles containing or not acetaminophen as a drug model. Polyethylene glycol (MW, 4000) used as binder was atomized onto the bed by a two-fluid spray nozzle. The granulation experiments followed a 2³ factorial design with triplicates at the center point and were carried out by varying the spray nozzle vertical position, the atomizing air flow rate and the binder feed rate. Granules were evaluated by their pharmacotechnical properties like size distribution, bulk and tapped densities, Carr index, Hausner ratio and tableting characteristics. Analysis of variance showed that granule sizes were affected by the PEG feed rate and atomizing air pressure at the significance levels of 1.0 and 5.0%, respectively, but spray nozzle distance to the substrate bed was not significant. The spray conditions also affected granule flow and consolidation properties, measured by the Carr index and Hausner ratio. Measured densities, Carr indexes and Hausner ratios proved that granules flowability and consolidation properties are adequate for pharmaceutical processing and tableting. Tablets prepared with acetaminophen-containing granules showed good properties and adequate release profiles in in vitro dissolution tests. The results indicate the suitability of spouted beds for the hot melt granulation of pharmaceutical coarse powders.     

Effects of process parameters on granules properties produced in a high shear granulator

Results of a study on the influence of process parameters such as impeller speed, granulation time and binder viscosity on granule strength and properties are reported. A high shear granulator (Cyclomix manufactured by Hosokawa Micron B.V., The Netherlands) has been used to produce granules. Calcium carbonate (Durcal) was used as feed powder and aqueous polyethylene glycol (PEG) as the binder. The dried granules have been analysed for their strength, density and size distribution. The results show that increasing the granulation time has a great affect on granules strength, until an optimum time has been reached. The underlying cause is an increase in granule density. Granules are consolidated more at higher impeller speeds. Moreover, the granule size distribution seems not to be affected significantly by an increase in impeller speed. Granules produced with high binder viscosity have a considerably lower strength, wide strength distribution due to poor dispersion of binder on the powder bed. Binder addition methods have showed no considerable effect on granule strength or on granule size distribution.     

Formation of hollow granules from liquid marbles: Small scale experiments

Research into formation of hollow granules from liquid marbles is an emerging field in hydrophobic granulation. A liquid marble is formed by a network of self-assembled hydrophobic powder around a droplet, and this paper investigates the conditions required for forming hollow granules from a liquid marble precursor.Single drops of fluid were produced using a syringe and placed onto loosely packed powder beds of hydrophobic powders. Liquid marbles formed from several powder/liquid combinations were dried at several conditions to investigate the drying conditions required for formation of a stable hollow granule.The formation of stable hollow granules was found to depend on drying temperature and binder concentration. For HPMC and PVP binder, formation of hollow granule is proportional to binder viscosity and for HPC binder, this relationship is constant. Different combinations of powder and binder at both drying temperatures - 60 °C and 100 °C - had mixed success rates in forming hollow granules, but generally the success rate was improved by using higher drying temperatures, smaller particles or higher viscosity binder fluids.     

Behaviour of soft granules under compression: Effect of reactive and non-reactive nature of the binder on granule properties

Granulation is a process where primary powder particles are made to adhere to form multi-particle entities called granules and this is achieved by using a binder. The binders can be broadly classified into two categories viz. reactive (reacts with base powder) and non-reactive (does not react with the base powder). The effect of various parameters related to binder liquid (binder viscosity, addition rate, distribution over the bed etc.) on the mechanism of granulation and physical/mechanical properties of granules is well studied. However, comparison of physical and mechanical properties of granules made via reactive and non-reactive binder using the same base primary particles has not been reported. In this paper, granulation of sodium carbonate primary particles under reactive and non reactive conditions was studied. The mechanical properties of sodium carbonate granules were characterized using single granule compression measurements. The average single granule apparent strength of reactive granules was higher compared to non-reactive granules. It was observed that granules formed using non reactive binder were brittle and showed multiple breakages. However granules made using reactive binder showed single breakage followed by significant plastic flow. In addition, bulk granule compression measurements were also carried out. Known models of Heckel, Kawakita and Ludde, and Adams et al. (developed mainly for pharmaceutical and metal powders) were used to predict mechanical properties of soft detergent granules. The bulk granule compression measurements also showed that reactive granules have higher strength compared to non-reactive granules. However, the absolute values of granule strength obtained from the empirical models were lower than the granule strength obtained from single granule compression measurements.     

Population balance modelling for a flow induced phase inversion based granulation in a two-dimensional rotating agglomerator

A novel two-dimensional rotating agglomerator was developed to carry out the flow induced phase inversion (FIPI) based granulation. The process in this agglomerator shows that a continuous paste flow (mixed with liquid binder and primary particles) is extruded into the interstice of two relatively rotating disks, as the paste becomes solidified due to the loss of heat to the disks, it is then broken into granules by the shearing force imposed by the rotating disk. Experimental measurements have shown that the size of these granules is enlarged along the positive radial direction of the disks. It is also found that these granules contain approximately the same quantity of binder in terms of its volume fraction. The paper thus proposes a population balance (PB) model to describe the growth of the granules by considering a size independent agglomeration kernel. The PB simulated results are found to be well capable of describing the change of the particle size distribution (PSD) of the granules in the radial direction. This study also proposes a velocity profile for the paste flow and attempts to establish a quantitative relationship between the granulation rate and the deformation rate as this would help us understand the mechanism of the agglomeration. It is hoped that this study would be used to improve the design of the agglomerator and to assure the control of the process and the granular product quality.     

Kinetics of fluidised bed melt granulation I: The effect of process variables

This paper is the first of a series to study the influence of operating conditions on the kinetics of fluidised bed melt granulation. First, we identify the rate processes responsible for the net growth in granule size in a top-sprayed fluidised bed granulator and propose a sequence of events based on these rate processes. The overall kinetics during the process is identified to be a combination of particle aggregation, binder solidification and granule breakage. By conducting experiments in a small-scale modified commercial fluidised bed granulator, the influence of various operating conditions (binder spray rate, bed temperature, atomising pressure, fluidising air velocity) on the granule growth behaviour was examined. The results indicate the granule growth rate to be directly dependent on the relative amount of binder sprayed into the bed, which essentially determines the speed of the aggregation process. The overall granule growth rate is observed to increase relatively with increased bed temperature for a more viscous PEG4000, while a maximum growth is seen for a lower viscosity PEG1500. A larger droplet size was also seen to have increased the overall growth rate, even though a smaller droplet seems to be able to induce a faster initial growth. The results also reveal the increase in fluidising air velocity to reduce the overall granule growth rate. The final granule size distribution was also observed to become narrower with increased bed temperature and fluidising air velocity. These observations are effectively explained using the proposed sequence of rate events.     

Particle growth mechanisms in fluidised bed granulation—II: Comparison of experimental data with growth models

Growth data from batch granulation experiments have been fitted to simple geometrical models. Layered granule growth approximates to concentric coating of core particles with binder, although the actual mechanism of growth is different. In the case of agglomeration, a relationship is established betwee mean granule size and binder content.     

Controlling the morphology of ceramic and composite powders obtained via spray drying – A review

Spray drying is one of the most convenient methods for drying suspensions (slurries) and for granulation of materials. Spray dried powders have good flowability, narrow size distribution and controllable morphology. Morphology of powder particles (also called granules or microspheres) strongly affects the use and handling of powders. This review discusses the latest research on parameters that affect morphology and size of granules obtained by spray drying: atomization parameters, properties of sprayed slurry, mass transfer etc. The formation of hollow and dense granules is extensively reviewed. Granule size is affected by droplet size, slurry concentration and initial particle size. Morphology mostly depends on size distribution of initial ceramic particles, agglomeration tendency in the slurry and mechanical strength of the shell of a granule during the drying process compared to capillary force of the suspension liquid. Polymer additives (e.g. binders and lubricants) change the properties of granule shell and the evaporation of moisture; thus, polymer additives significantly affect morphology.     

Mechanisms in high-viscosity immersion-granulation

An atypical high-shear granulation process is investigated in which a fine inert powder is bound with a highly viscous surfactant paste. The mechanism comprises adsorption of powder particles onto paste fragments, breakage of powder-coated paste granules, micro-mixing of the granules with incorporation of the powder, granule growth via coalescence, and finally granule consolidation. These stages are supported by micrographic and granule size distribution data. The agglomeration process features two main mechanisms, namely binder distribution followed by granule consolidation and coalescence, the transition between which is shown to be dependent upon the operating parameters.A number of time-dependent consistency regimes can be identified and quantitatively described using bulk tapping compaction tests. Of particular interest is the trend in Hausner ratio, which provides information on the inter-granular friction and cohesivity. Various pseudo-steady state tapping parameters are used to track the agglomeration process, the results of which are consistent with the Iveson et al. (2001a, Powder Technology 117, 83-97) steady state agglomeration regime maps. The effects of paste/powder composition, paste rheology and mixing speed upon the agglomeration rate can be explained physically in terms of adsorption, viscous and mechanical energy dissipation mechanisms. In summary, the work introduces a preliminary analysis of an immersion-granulation mechanism in which a number of key features are identified.     

A combined experimental and computational study of wet granulation in a Wurster fluid bed granulator

A methodology combining theoretical and experimental techniques for analyzing the growth of granules in a fluidized bed granulator was developed. The methodology combines several key features of this complicated process: (i) population balancing (PB) of growth of different granules; (ii) hydrodynamic modeling of the gas-solid mixture flow using Computational Flow Dynamics (CFD); (iii) modeling of contact mechanics and granule formation; (iv) the Stokes number analysis for calculation of successful collisions; (v) well-controlled experimental study of the wet granulation. First, a detailed CFD model of the gas-solid flow and agglomeration (Model CFD-PB) within the Wurster type granulator was developed. Second, a simplified PB model of agglomeration in a homogeneous system (Model H-PB) was developed. The quadrature method of moments (QMOM) was used for solution of PB equations in both models. The kinetic theory of granular flow (KTGF) was employed in both models for calculation of the number of collisions between solid particles of different classes. Success factors, based on the Stokes number analysis, were calculated using results of extensive mesoscale simulations of the formation of realistic three-dimensional virtual granules. Comparison of simulation results of CFD-PB vs. H-PB models allowed evaluation of the KTGF kernel functionality to be used in H-PB model. Next, fluid bed granulation experiments were conducted for typical pharmaceutical excipients (microcrystalline cellulose, mannitol and dicalcium phosphate) with 15% HPC binder solution in a pilot plant Wurster granulator. The observed granule growth was a function of the surface roughness of excipients. Finally, the H-PB model was fitted to the experimental data. The only adjustable parameter of the H-PB model was an effective agglomeration rate constant, which we expect to be mostly related to the binder wetness on the surface of colliding granules.     

Comparison between two types of Artificial Neural Networks used for validation of pharmaceutical processes

Two types of Artificial Neural Networks (ANNs), a Multi-Layer Perceptron (MLP) and a Generalized Regression Neural Network (GRNN), have been used for the validation of a fluid bed granulation process. The training capacity and the accuracy of these two types of networks were compared. The variations of the ratio of binder solution to feed material, product bed temperature, atomizing air pressure, binder spray rate, air velocity and batch size were taken as input variables for training the MLP and GRNN. The properties of size, size distribution, flow rate, angle of repose and Hausner's ratio of granules produced, were measured and used as output variables. Qualitatively, the two networks gave comparable results, as both pointed out the importance of the binder spray rate and the atomizing air pressure to the granulation process. However, the averaged absolute error of the MLP was higher than the averaged absolute error of the GRNN. Furthermore, the correlation coefficients between the experimentally determined and the calculated output values, the corresponding prediction accuracy for the different granule properties as well as the overall prediction accuracy using GRNN were better than using MLP. In conclusion, the comparison of two different networks (MLP, a so-called feed-forward back-propagation network and GRNN, a so-called Bayesian Neural Network) showed the higher capacity of the latter for validation of such granulation processes.     

Wet granulation: the effect of shear on granule properties

This paper presents a study of the wet granulation of fine cosmetic particles using a high-shear mixer granulator on a given particle and binder system. The shear effect on granule properties is highlighted. The granules formed under different impeller speeds are divided into size classes and further examined in terms of porosity, friability and binder content.

The main result of this study is that, depending on operating conditions, the granulation of a fine powder with a given binding liquid can result in the formation of granules of very different characteristics in terms of size, porosity and friability. Mechanical energy brought to the granulation system is as important as the physicochemical characteristics of the powder–binder pair.     

Influence of Slurry Parameters on the Characteristics of Spray-Dried Granules

A systematic study has been performed to determine how the characteristics of granules prepared by spray drying aqueous alumina slurries are influenced by processing parameters: binder type (PEG Compound 20M or PEG-8000), solids loading (30 or 40 vol%), ammonium polyacrylate deflocculant level (0.35-1.00 wt%), and spray-dryer type. Correlations between slurry rheology and granule characteristics have been made, and a model for granule formation is presented. The packing density of the primary particles within the granules is lower for slurries with higher yield stress and is dependent on the slurry solids loading. Granules prepared using 0.35 wt% deflocculant (0.14 mg/m²), which correspond to high slurry yield stress, are of solid morphology, whereas higher deflocculant levels result in hollow granules that contain a single large open pore or crater. The degree to which particles are able to rearrange during drying influences the final granule density and is determined by the strength of the floc structure, as indicated by the slurry yield stress. When the yield stress is low, a crater may form from the inward collapse of the surface of a forming granule when the particle packing density in a droplet continues to increase after the droplet size becomes fixed by the formation of a rigid shell, leaving an internal void with internal pressure lower than that of the surrounding atmosphere.     

Non-uniformity of binder distribution in high-shear granulation

Further experimental investigation based on a microscopic, or single granule, scale has been conducted to investigate the uniformity of binder composition within a given size class for high shear melt granulation. This work assesses whether there is significant non-uniformity of binder composition within size classes to warrant considering this level of detail to improving population balance modelling of high shear granulation. It is concluded that at early times in a batch granulation process there is a broad variation in binder content between individual granules and that this variation persists in granules of small size.