Coating of finely dispersed particles by two-fluid nozzle

G Model PARTIC-1050; No. of Pages 14

ARTICLE IN PRESS Particuology xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

Particuology journal homepage: www.elsevier.com/locate/partic

Coating of finely dispersed particles by two-fluid nozzle N. Hampel ∗ , E. Roydeva, A. Bück, E. Tsotsas Thermal Process Engineering, Otto von Guericke University, Universitätsplatz 2, 39106 Magdeburg, Germany

a r t i c l e

i n f o

Article history: Received 29 September 2016 Received in revised form 25 April 2017 Accepted 10 May 2017 Available online xxx Keywords: Coating Spray drying Fine dispersed particle Two-fluid nozzle

a b s t r a c t Particle coatings are used extensively to generate dispersed solids with well-defined properties, e.g., to protect active ingredients, with most coating processes using core particles of a diameter larger than 200 ␮m. This work contributes to the development of a coating process for fine dispersed particles (diameter less than 50 ␮m) by combining two particle-formulation processes, namely, coating and spray drying. The feasibility of the operation is based on and demonstrated by the innovative application of a two-fluid nozzle. Experiments were conducted by using glass particles as core particles and sodium benzoate as the coating agent. The coating of finely dispersed particles is achieved by the spraying of particles and coating solution as a homogeneous suspension. The aim is to create droplets with only one contained particle at the nozzle outlet. After evaporation of the water in the droplet, a thin solid film is built on the particle surface. The suspension viscosity was measured and compared with empirical equations from the literature. The liquid-film thickness on the particle surface was calculated to predict the building of a uniform coating layer or agglomerates. In this study, the feasibility of pneumatic transport through the nozzle and an investigation of the process were illustrated. The agglomeration fraction and degree of coating of the particle surface were analyzed optically by scanning electron microscopy. In this way, the influence of different processes and suspension parameters on the product quality were determined. © 2017 Published by Elsevier B.V. on behalf of Chinese Society of Particuology and Institute of Process Engineering, Chinese Academy of Sciences.

Introduction The coating of fine dispersed particles is becoming increasingly important in the food and pharmaceutical industries, and in chemistry. Coating processes are motivated by the fact that different functional properties can be achieved, the release of active ingredients can be manipulated, adhesion and compatibility between various components of the composite materials can be improved. A fluidized bed coating without severe agglomeration is typically limited to particle sizes that are appreciably larger than 100 ␮m (Dewettinck & Huyghebaert, 1998). In this case, however, fine particles are extremely difficult to fluidize because of their high cohesion (Geldart, 1973). In general, the fluidization of such powders leads to the formation of channels or ratholes. To overcome these processing issues, fine particles are granulated with larger, easily fluidized powders to enhance processing (Ehlers et al., 2009; Takano, Nishii, & Horio, 2003; Thiel & Nguyen, 1982; Shen, 1996; To & Davé, 2016). Such approaches can produce coated products, but they have the disadvantage for limited application ranges, because of a need for high concentration of well fluidized excipient, which is

∗ Corresponding author. Fax: +49 0049391 67 18265. E-mail address: [email protected] (N. Hampel).

problematic for products that require high active-ingredient loadings. Several methods exist for fluidized bed coatings or the agglomeration of fine cohesive particles without the addition of significant amounts of excipient materials. However, most methods require complicated and potentially expensive modifications to the typical fluidized-bed configuration. Kawaguchi et al. (2000) coated cohesive active-pharmaceutical-ingredient particles in a rotating fluidized bed and prepared reproducible acetaminophen granules (∼500 ␮m). The production of this product in a conventional fluidized bed is difficult. Later, Watano et al. (2003, 2004) reported that 15-␮m cornstarch particles could be granulated or coated individually in a rotating fluidized bed. In this case, the centrifugal forces exerted onto normally cohesive cornstarch particles increased the apparent particle weight and made them behave like well fluidized particles. In other studies, Hamashita et al. (Hamashita, Nakagawa, Aketo, & Watano, 2007; Hamashita et al., 2008, 2009) presented that micronized core particles that contain ibuprofen could be granulated using an impeller-agitated fluidized bed. Miyadai, Higashi, Moribe, and Yamamoto (2012) used a similar apparatus to produce microgranules of ibuprofen particles. These methods could produce agglomerates with high loadings of active materials reproducibly. Ichikawa and Fukumori showed in a series of studies that micronized and cohesive powders could be coated and agglom-

http://dx.doi.org/10.1016/j.partic.2017.05.015 1674-2001/© 2017 Published by Elsevier B.V. on behalf of Chinese Society of Particuology and Institute of Process Engineering, Chinese Academy of Sciences.

Please cite this article in press as: Hampel, N., et al. Coating of finely dispersed particles by two-fluid nozzle. Particuology (2017), http://dx.doi.org/10.1016/j.partic.2017.05.015

G Model PARTIC-1050; No. of Pages 14

ARTICLE IN PRESS N. Hampel et al. / Particuology xxx (2017) xxx–xxx

2

Nomenclature Latin symbols A Surface area (m2 ) d Diameter (m) d32 Sauter mean diameter (m) Force (kg m/s2 ) F h Calculated liquid-film thickness (m) M Mass (kg) ˙ M Mass flow rate (kg/s) q Normalized density distribution (m−1 ) Q Normalized cumulative distribution t Time (s) Temperature (◦ C) T V Volume (m3) X Loss of product/solid (%) Greek symbols ˛ Particle fraction in suspension Viscosity (Pa s)  Total moment (mi )  Density (kg/m3 )  2  Variation (m2 ) Solute mass fraction in coating solution  Subscripts 1 Single particle ad Adhesion Calculated calc CS Coating solution dr Droplet dry After drying in oven Experimental exp g Gas Gravitational gr inlet Inlet gas Mass fraction M max Maximum N Nozzle Particle p s Solid V Volume fraction

erated in a bottom-spray Wurster-type spouted bed (Ichikawa, Shin, & Kang, 1999; Jono, Ichikawa, Miyamoto, & Fukumori, 2000; Fukumori et al., 1991). The particles could not be fluidized in a strict sense; hence, high superficial velocities were required to make them spout and a draft tube was used to enhance their circulation. Watano et al. (2003, 2004) compared cohesive forces to other forces such as buoyancy or a resultant drag force, indicating that to counteract cohesion one may need to increase the body weight of the powders. A cohesion reduction by introducing flow additives to fine powders could be used to promote fluidization as was shown in US patent 6,833,185 (Zhu & Zhang, 2004). Yang, Sliva, Banerjee, Dave, and Pfeffer (2005) showed that surface modification through dry coating with nanosized flow additives is a simple and practical method to provide more significant and reliable results compared with additive blending. Cohesive powders were made to flow and be potentially fluidizable. In dry coating, a discrete, fairly uniform layer of nanosilica particles is applied onto the surface of the cohesive host particles. Yang et al. derived a simple equation that showed that a reduction in cohesion because of the introduction of nanosized surface asperities as a

nanosilica coating is inversely proportional to the particle size, and hence, the coating of nanosilica would improve the flow properties. Chen et al. (Chen, Jallo, Quintanilla, & Dave, 2010; Chen, Yang, Dave, & Pfeffer, 2008; Chen, Yang, Mujumdar, & Dave, 2009) showed in a few studies that surface asperities, the level of coating, and the surface energy play a major role in cohesion reduction. The addition of a nanosilica surface coating can reduce the granular bond number (the ratio of cohesive forces to inertial forces) by more than an order of magnitude for aluminum particles below 5 ␮m. This cohesion reduction can allow for fine cohesive particles to be fluidized in a conventional fluidized bed and has led to a fluidized-bed-coated product without significant agglomeration. Work on the dry coating of drug powders indicates that their cohesion can be reduced, albeit less significantly than the cohesion of materials such as cornstarch via dry coating with silica; however, their fluidization behavior has not been investigated (Jallo, Ghoroi, Gurumurthy, Patel, & Dave, 2012). Recent work of To and Davé (2016) shows that silica nanoparticles can be dry coated onto the surface of as-received and micronized ibuprofen particles with a median size of 41–74 ␮m and a corresponding Sauter mean diameter of 21–41 ␮m. Ibuprofen was selected as a model drug because of its poor flowability and hence poor fluidizability. Dry coating with nanosilica particles allows previously cohesive particles to be fluidized-bed coated with minimal agglomeration. The main objective of this contribution is to investigate the application of an innovative process for the coating of fine dispersed particles (dp < 50 ␮m). The newly developed method combines the advantages of two particle formulation processes, namely, spray drying and coating. Spray-drying processes are applied widely in numerous industrial fields, particularly in dairy and food production. They produce a dry powder by the rapid evaporation of water from droplets that are atomized in a stream of hot air. The main advantage of this method is the short residence time of particles inside the drying chamber. Thus, spray drying is a suitable drying technique for thermally sensitive products (Tran, Jaskulski, & Tsotsas, 2016). All spray dryers use some type of atomizer or spray nozzle to disperse liquid into a controlled-drop-size spray. Most common are rotary disks and single or multiple fluid pressure nozzles. Depending on the process needs, drop sizes from 10 to 500 ␮m can be achieved with appropriate choices (Kröll, 1978). In wet coating, particles are contacted with droplets from the liquid, which deposit on the particle surface. After the evaporation of water from the droplets, a coating layer is built. The liquid can be a solution or a suspension (Hampel, 2015). The most-used equipment for wet coating includes the pan coater, drum coater, and the fluidized-bed coater (Hampel, 2015). In this study, the change in suspension viscosity with particle mass and solute was determined experimentally and calculated. Experimental studies were conducted with a variation in operating parameters, such as nozzle pressure, mass fraction of core particles in suspension, and mass fraction of solute in the coating solution. The aim in this study was to establish the influence of process and suspension parameters on the individual processes, namely, coating (partial or complete), agglomeration, and spray drying. It is desirable to maintain a maximal amount of completely coated particles, with a minimal amount of agglomerated particles and spray-dried droplets. For this reason, dependency of the formation of agglomerated particles on process parameters and properties of the suspension was analyzed. An outline of this study is as follows. The general principles of the proposed process are discussed. The suspension viscosity was measured and the dependency of the particle and coating solute mass fraction were determined and the results were compared. An experimental plant and the feasibility of the coating process in this

Please cite this article in press as: Hampel, N., et al. Coating of finely dispersed particles by two-fluid nozzle. Particuology (2017), http://dx.doi.org/10.1016/j.partic.2017.05.015

G Model PARTIC-1050; No. of Pages 14

ARTICLE IN PRESS N. Hampel et al. / Particuology xxx (2017) xxx–xxx

Fig. 1. Aim of innovative application of a two-fluid nozzle for fine-particle coating.

plant are described. The influence of process parameters on the product quality is described and discussed. New approach for coating fine dispersed particles The main objective of this contribution was to investigate the application of an innovative process to coat fine dispersed particles (dp < 50 ␮m). The newly developed method combines the advantages of two classical particle-formulation processes, namely, spray drying and coating. The main equipment that was used is an externally mixing two-fluid nozzle, in which two fluid phases were conveyed through two nozzle channels as shown in Fig. 1. The gas phase is transported independently through one nozzle channel. A suspension that consists of a coating solution (water as a solvent and a dissolved solid) and the particles to be coated, was introduced through the second nozzle channel into the system. Pre-mixing and external atomization yields a spray, in which every core particle is covered with a liquid-coating-solution film at the nozzle outlet. Because of the ensuing drying process, this liquid film is solidified and forms a thin coating layer on the particle surface, see Fig. 1. Essential prerequisite material properties must be considered to form a homogeneous suspension: • particles must be insoluble in the coating solution, • no chemical reaction must occur between the coating solution and the particles, • for aqueous coating solutions, the particles should not be hydrophobic, • suitable mass fractions of particles and coating solutions should exist in the suspension. In general, the gas pressure in the nozzle has an inverse relationship with the droplet size. An increase in pressure will typically reduce the droplet size, whereas a reduction in pressure will increase the droplet size. Similarly, an increase in spray angle will reduce the droplet size, whereas a reduction in nozzle outlet diameter will decrease the droplet size and increase the number of droplets (Hampel, 2010; Schick, 2006; Lefebvre, 1989). The liquid flow rate has a directly proportional influence on droplet size. An increase in flow rate will increase the droplet size (Schick, 2006; Spray Drying Systems Co., 2000). The liquid properties of importance with respect to two-fluid atomization are: viscosity, liquid density, and surface tension. In general, an increase

3

in viscosity will increase the droplet size (Hampel, 2010; Rizkalla & Lefebvre, 1975). Of all factors that influence the mean droplet size, the gas velocity is the most important. This is well known for low-viscosity liquids, where the mean droplet size is approximately inversely proportional to the air velocity (Hede, Bach, & Jensen, 2008). Because of the heterogeneous nature of the atomization process in a two-fluid nozzle, liquid threads and ligaments that are formed by the various mechanisms of the jet and liquid-sheet disintegration vary extensively in diameter. For this reason, droplets vary in size. Practical nozzles do not produce sprays of uniformly sized droplets at any given operating conditions. Instead, a spectrum of droplet sizes is obtained (Hampel, 2010; Lefebvre, 1989). These researches show that the distribution of droplet size depends on the nozzle type, construction, nozzle diameter, and liquid properties. The experimental results have not been correlated. The coating of fine dispersed particles with the innovative application of a two-fluid nozzle involves complex microscale interactions between particles and droplets of the coating solution, as depicted in Fig. 2. Here, the gas flow conditions and their influence on the microprocesses are not yet considered. After spraying the homogeneous suspension, four cases are possible: Case I. If the particle size dp is much smaller than the droplet size ddr , two or more particles can be deposited within a droplet. Therefore, after solvent evaporation from the droplet, an agglomerate of two or more particles can form. For post-processing breakage of the agglomerate, the coating layer will be damaged partially. Case II. If the particle size dp is less than or equal to the droplet size ddr , only one particle is present within a droplet. In this case, after the evaporation of solvent from the droplet, a thin solid film is formed on the surface of the particle. Here, two scenarios are possible. In the first scenario, the coating layer dries rapidly in the flow of a heated gas. This scenario is the best for the coating of fine dispersed particles with the described technique. In the second scenario, the coating layer dries slowly, i.e., it remains wet for an extended period. Therefore, particles that come in contact because of particle collisions may form agglomerates because of the absorption of collision energy by a viscous film. Case III. If the particle size dp is larger than the droplet size ddr , droplet wetting occurs only on a part of the particle surface. In this case, after solvent evaporation from the droplet, a thin solid film is formed only on a section of the particle surface, which results in partial coating. Case IV. If no particle is present in the droplet, a spray-dried particle that consists of the solute is formed after solvent evaporation. This undesired product is termed overspray. It may be separated mechanically from the coated particles, and recycled to prepare new coating solutions. As it can be inferred from Fig. 2, a significant number of parameters affect the path of a single entity as it moves along the network to produce coated particles. In practice, particles to be coated are usually not monodisperse, accordingly, the size of the sprayed droplets would be distributed. For this reason, it is not possible to produce only coated particles. Therefore, an undesired fraction, such as overspray particles, agglomerates or only partially coated particles will always exist in the product of a coating process with the innovative application of a two-fluid nozzle. The heated gas mass can move in co-current flow, countercurrent flow, or mixed flow relative to the liquid (Schmidt, Rieck, Bück, & Tsotsas, 2015). The selection of air inlet temperature depends on the temperature sensitivity of the material, and the desired particle size. The temperature can be varied between 30 and 800 ◦ C (Kunii & Levenspiel, 1991). The flow rate of drying air

Please cite this article in press as: Hampel, N., et al. Coating of finely dispersed particles by two-fluid nozzle. Particuology (2017), http://dx.doi.org/10.1016/j.partic.2017.05.015

G Model PARTIC-1050; No. of Pages 14 4

ARTICLE IN PRESS N. Hampel et al. / Particuology xxx (2017) xxx–xxx

Fig. 2. Microprocesses related to wetting and drying with an influence on the outcome of the proposed innovative coating process of fine dispersed particles with a two-fluid nozzle.

Fig. 3. Innovative coating of fine dispersed particles with two-fluid nozzle in consideration of local flow conditions, i.e., acting forces in microprocess network.

determines the powder residence time in the spray dryer and the moisture content in the product. The influence of drying on the agglomeration of wet particles, which should be avoided (Fig. 2), is discussed in Dernedde, Peglow, and Tsotsas (2013) and Dadkhah (2014). Even conditions that would lead to a desired product (“coating”) according to Fig. 2 are subject to and can be jeopardized by acting forces and local gas-flow conditions. This is illustrated in Fig. 3 by comparing the adhesion forces Fad to the sum of the gravitational force Fgr and flow (drag) force Fflow that act on the liquid coating film. The adhesion force is assumed to maintain a liquid coating film on the particle to be coated, whereas the other two forces are considered to assist towards partial or total film detachment. In an ideal scenario, when the particle size dp is less than or equal to the droplet size ddr (see Fig. 2, Case II), three main cases are possible (Fig. 3). Depending on the ratio of the adhesion forces Fad to the sum of the flow forces Fflow and gravitational force Fgr , the droplet may retain its position, be partially destroyed or may be blown away. Stronger adhesion forces yield no change in droplet position. After evaporating water from the droplet, the formation of a coating layer on the particle surface or the formation of agglomerates is possible. If the adhesion forces cannot withstand the forces that are exerted by the gas flow and by gravity, the droplet will be destroyed partially or it will be removed completely from the par-

Table 1 Material properties of glass particles. Sauter diameter, d32 (␮m) Particle density, p (kg/m3 ) Sphericity

45 2500 0.94

ticle surface. Subsequently, the droplet or part of the droplet will be dried. Depending on the amount of liquid on the particle surface, either a partial coating layer or no coating is formed. Materials and properties Core material Glass particles (Cerablast GmbH) were used as the core material in the coating experiments. The particle size distribution (PSD) was measured offline by using an optical measurement device CAMSIZER-XT (Retsch Technologies, Germany). The material possesses a narrow PSD between 10 and 70 ␮m (Fig. 4); the main properties of the material are listed in Table 1. The Sauter mean diameter of the solid particles is described by the ratio of particle volume to surface area: d32 = 6

Vp , Ap

(1)

Please cite this article in press as: Hampel, N., et al. Coating of finely dispersed particles by two-fluid nozzle. Particuology (2017), http://dx.doi.org/10.1016/j.partic.2017.05.015

G Model

ARTICLE IN PRESS

PARTIC-1050; No. of Pages 14

N. Hampel et al. / Particuology xxx (2017) xxx–xxx

5

Fig. 4. Initial normalized particle size distribution of glass particles. Fig. 5. Calculated and measured suspension viscosities sus as a function of solid volume fraction ˛V . The suspension contains 30 wt.% sodium benzoate in water and particles.

Table 2 Material properties of coating solution. Properties of the coating solution

Solute mass fraction in coating solution,  (wt%)

3

Coating solution density, CS (kg/m ) Coating solution viscosity, CS (Pa s)

10

20

30

1038.33 0.0014

1081.93 0.0023

1126.23 0.0039

and will be used to represent the variation of the entire PSD. The evolution of this parameter can be considered a principal quantitative criterion for the coating. Coating solution Aqueous sodium benzoate solution was used as a coating agent (Tigon Chemie GmbH). Sodium benzoate (C7 H5 NaO2 ) has a good solubility in water and is used widely as a preservative in the food industry having the E number E211. The solution mass fraction was varied as 10, 20, and 30 wt%. Sodium benzoate is a salt solution and a Newtonian fluid. Its influence on the quality of coated particles will be discussed in Section Solute mass fraction in the coating solution. The sodium-benzoate solution density was determined at 20 ◦ C by using a density measuring device DMA 58 (Anton Paar GmbH, Austria). The device was calibrated by determining the density of the distilled water. The densities of the 10, 20, and 30 wt.% sodium benzoate solution are summarized in Table 2. A Höppler viscometer (Medingen, 1966) was used to determine the dynamic viscosity of the coating solution at different solids concentrations. The device was calibrated by using calibration liquids. The results of the viscosity measurements are listed in Table 2 As expected, both coating solution parameters (density and viscosity) increase with an increase in solute mass fraction in the solution. Suspension Suspension viscosity Extensive literature exists on the viscosities of suspensions. Large scatter exists in the experimental results, presumably because of the different particle diameters, liquid viscosities, and the apparatus used. For example, information is available on the dependence of dynamic viscosity of a suspension sus on solid fraction in classical studies (Einstein, 1911; Jogwich, 1957; Vand, 1948; Eyring, Henderson, Stover, & Eyring, 1964; Hatschek, 1920; Orr & Dalla Valle, 1954; Kurpiers, 1984; Mooney, 1951; Krieger & Dougherty, 1959). The equations for suspension viscosity from these studies are summarized in Appendix A, Table A1. A wide range of values is reported for the maximum solid volume fraction ˛V,max , even for suspensions with a monodis-

perse particle diameter distribution (for example, in Orr and Dalla Valle (1954) ˛V,max = 0.54; in Kurpiers (1984), Mooney (1951), and Krieger and Dougherty (1959) ˛V,max = 0.74). The viscosity sus of five different suspensions with different volume fractions of particles (30%, 40%, and 50%) and mass fraction of sodium benzoate (10, 20, and 30 wt.%) was measured by a concentric-cylinder rheometer (RheoPlus, model MCR301, Anton Paar GmbH). The rheometer was equipped with a temperaturecontrol unit that surrounded the test section. This unit kept the suspension at a constant, specified temperature. An inner cylinder with an outer radius of 19 mm was used. The height of the inner cylinder was 55 mm. The rheometer measured the angular speed and the corresponding exerted motor torque. The shear rate and shear stress were calculated at the inner cylinder. Before starting the measurement, the rheometer was calibrated and checked by measuring pure water. Viscosity measurements were made at 20 ◦ C and a shear rate of 400 s−1 . Within this range of shear rates, particles do not segregate and air bubbles do not form in the suspension, which can distort the measurements. Fig. 5 gives the measured and calculated viscosities according to the equations that are summarized in Table A1 of Appendix A, for particle volume fractions of 16%, 23%, and 31%. The suspension viscosity increases with an increase in solid volume fraction ␣v . The extent of this increase differs between the various equations used and is minimal for the equation of Orr and Dalla Valle (1954). A comparison between the measured and calculated results of the suspension viscosity (Newtonian fluid) shows that the experimental viscosity is higher than the calculated viscosity. The differences increase with increasing particle volume fraction in suspension ˛V . This can be explained by the PSD of the glass particles. A Sauter diameter of 45 ␮m was assumed, but the particle size ranges from 10 to 70 ␮m, see Fig. 4. This influences the viscosity; larger particles yield a higher suspension viscosity, which results in the observed difference between the calculated and measured suspension viscosity. Similar results were achieved in the theoretical and experimental investigations of suspension viscosity sus as a function of solute mass fraction in the coating solution , see Fig. 6. In this case, the change in suspension viscosity results from a change in coating solution viscosity CS , which increases with an increase in solute mass fraction in the coating solution . A comparison between the measured and calculated results shows that the experimental viscosity is higher than the calculated viscosity. This can again be attributed to the PSD.

Please cite this article in press as: Hampel, N., et al. Coating of finely dispersed particles by two-fluid nozzle. Particuology (2017), http://dx.doi.org/10.1016/j.partic.2017.05.015

G Model PARTIC-1050; No. of Pages 14

ARTICLE IN PRESS N. Hampel et al. / Particuology xxx (2017) xxx–xxx

6

Fig. 6. Calculated and measured suspension viscosities sus as a function of sodium benzoate mass fraction in the coating solution , with a particle volume fraction of ˛V = 0.31.

Fig. 7. Schematic view to estimate liquid-film thickness on particle surface.

Calculated liquid-film thickness The liquid-film thickness on the particle surface is important in terms of process efficiency. A thick film requires a longer evaporation time, which means that particle collisions may enhance the formation of agglomerates. For a low liquid-film thickness, the risk of forming a non-uniform coating layer exists. To estimate the liquid-film thickness (see Fig. 7), the following is assumed: • ideal particles • ideal coated particles • local flow conditions are not considered The volume of coating solution VCS can be calculated as the difference between the suspension volume Vsus and the particle volume Vp , and is expressed by the difference between the mass of suspension Msus and the particle mass Mp (Mp = Vp p ). VCS = Vsus − Vp =

Msus − Vp p . CS ()

Msus ˛M . p

np =

(3)

Another way to determine the volume of the coating solution VCS is on the basis of the total number of particles np in a single droplet of suspension and the volume of liquid on only one particle surface VCS,1 : VCS = np VCS,1 .

(4)

The quantity VCS,1 also can be written as: VCS,1 =

 ␲ 3 (dp +hCS ) − dp3 , 6

where hCS is the liquid-film thickness. The total number of particles np is calculated from the ratio of total particle volumes Vp to the volume of a single particle Vp,1 :

(2)

The total particle volume Vp can be calculated from the suspension mass Msus , particle mass fraction in the suspension ˛M , and particle density p : Vp =

Fig. 8. Influences of particle properties on film thickness hCS : (a) particle diameter dp , (b) particle density p , and (c) particle mass fraction in suspension ˛M .

(5)

Vp . Vp,1

(6)

From Eqs. (2) to (6), the following relationship for the liquid-film thickness can be derived:



hCS = dp

3

␲(1 − ˛M )P . 6cs ()˛M

(7)

Eq. (7) shows that the calculated liquid-film thickness over the particle depends on particle parameters such as particle diameter dp , particle density p , and particle mass fraction in the suspension ˛M . Only one parameter of the coating solution, which is the solute mass fraction in the coating solution, influences the liquid-film thickness. The resulting influence of particle diameter dp , particle density p , and particle mass fraction in the suspension ˛M on the film thickness hCS is depicted in Fig. 8.

Please cite this article in press as: Hampel, N., et al. Coating of finely dispersed particles by two-fluid nozzle. Particuology (2017), http://dx.doi.org/10.1016/j.partic.2017.05.015

G Model PARTIC-1050; No. of Pages 14

ARTICLE IN PRESS N. Hampel et al. / Particuology xxx (2017) xxx–xxx

7

Fig. 8(a) shows a linear dependence between the particle diameter dp and the film thickness hCS . In this case, however, the total particle volume Vp remains constant, so the change in particle diameter corresponds to a change in the number of particles in the suspension np . Therefore, the film thickness increases with an increase in particle diameter and a decrease in the number of particles np . The influence of particle density p on the liquid-film thickness hCS is shown in Fig. 8(b). With increasing the particle density p , the particle total volume Vp , and the number of particles in the suspension np decrease. The effect of variation in particle mass fraction in the suspension ˛M on the liquid-film thickness hCS is shown in Fig. 8(c). To make the process more effective, the particle mass fraction needs to be increased. This is limited by the pneumatic transport of particles through the nozzle. A larger particle mass fraction ˛M yields thinner liquid films and increases the possibility of building a non-uniform layer during drying. By increasing the particle mass fraction, the particle total volume Vp , and the number of particles in suspension, np , also increases. According to Eq. (7), the relationship between the solute mass fraction in the liquid  and the film thickness hCS can also be analyzed. In this case, Eq. (8) can be applied because the particle properties remain constant.



hCS ∼

3

Fig. 9. Flow chart of experimental setup.

1 . CS ()

(8)

The solute mass fraction  influences the coating solution density CS . For the three applied solutions (see Table 2 only a low change in coating solution density exists (CS () is approximately constant) and, thus, only a small change in film thickness results. Experimental Experimental setup A simple laboratory-scale spray dryer was used to carry out the coating experiments. The process chamber was cylindrical (1.5-m length, 0.4-m diameter) and was constructed from non-insulated Plexiglas. A two-fluid-nozzle (type 970/0 S4, Düsen-Schlick GmbH, Germany) was positioned concentrically at the top of the apparatus, (see Hampel, 2010; Schmidt, Bück, & Tsotsas, 2015; Hoffmann, Rieck, Bück, Peglow, & Tsotsas, 2015). The nozzle was located 20 mm below the cover of the setup. A flow chart of the experimental setup is given in Fig. 9. Pre-dried air that was conditioned before entering the cover of the setup was used as the drying gas. To achieve the desired process temperatures at the inlet and in the apparatus, drying air was passed via an electrical heater before entering the process chamber at the top of the plant. At the plant outlet, the air and the product were separated in a cyclone. The gas temperatures were measured via thermocouples at the inlet, outlet, and inside the spray tower. The mass flow rates of the drying gas and the nozzle air were controlled by thermal gas flow measurements. All sensor data were registered by a data-acquisition system. Transport of suspension to the nozzle The uniformity in pneumatic transport of the suspension to the nozzle was checked by using the setup shown in Fig. 10. A defined amount of suspension, which consisted of 50 wt.% glass beads and 50 wt.% sodium benzoate solution (30 wt.% sodium benzoate mass fraction in the solution), was pumped out of tank 1 at 8.5 g/min and was collected in a beaker (tank 2). This suspension has the largest viscosity, see Section Core material. The pumped suspension was weighed and dried in a drying oven (Fa. Memmert) at 105 ◦ C. The dried sample was cooled in a desiccator and weighed again. The

Fig. 10. Schematic illustration to test suspension transport.

Table 3 Experimental data from suspension transport tests. M (tank 1) (g)

Msus

MH2 O

Mp + Ms

Mp

Ms

100 50

35 17.5

65 32.5

50 25

15 7.5

Msus (tank 2) (g)

Ms,dry + Mp,dry (g)

X (%)

98.48 48.65

64.25 31.82

1.01 1.02

particle mass Mp and sodium benzoate mass Ms in the solutions in the two tanks were determined and compared. Solids loss by suspension transport was calculated as: X=

Ms (tank1)+Mp (tank1) × 100%. Ms (tank2)+Mp (tank2)

(9)

Table 3 provides an overview of the experimentally obtained results. Two measurements were carried out with different suspension masses. The total solids loss was approximately 1%. The loss can be explained by particle sedimentation at the bottom of tank 1 or inside the hose. It can be assumed that the suspension transport proceeds nearly completely.

Please cite this article in press as: Hampel, N., et al. Coating of finely dispersed particles by two-fluid nozzle. Particuology (2017), http://dx.doi.org/10.1016/j.partic.2017.05.015

G Model PARTIC-1050; No. of Pages 14 8

ARTICLE IN PRESS N. Hampel et al. / Particuology xxx (2017) xxx–xxx

Fig. 11. Process and product variables during continuous coating: variations with coating time of (a) flow rates of drying and nozzle air, (b) gas temperatures in system, (c) accumulated sprayed suspension mass, and (d) particle size distributions of core and coated particles.

Coating process with an innovative application of a two-fluid nozzle In this section, the feasibility of the newly developed coating process is presented. The damper position of the two-fluid nozzle was kept constant in each experiment. First, the sodium-benzoate coating solution was prepared and glass particles were added. During the process, the suspension was homogenized using a stirrer. Before starting the experiments, the nozzle was cleaned to avoid deposit build-up and plugging. The inlet air temperature and gas mass flow rate were set to the experimental parameters. Before starting the process, it was necessary to operate the setup under the experimental conditions until a constant temperature was reached. The pump was purged and filled with the suspension and the process was commenced. At the end of the process, the cyclone was emptied and the product mass was weighed and the spray chamber, the nozzle, and the pump were cleaned. The product PSD was analyzed by using an optical device CAMSIZER XT. To demonstrate the feasibility of the coating process at the plant, the measured and derived data for one experiment are presented in Fig. 11. The experimental process parameters are summarized in Table 4. During the experiment, the drying and nozzle air flow rates were constant, see Fig. 11(a). Immediately after starting the experiment, because of the constant spraying of suspension into the chamber, the temperatures in the chamber and at the outlet decreased rapidly and remained constant (Fig. 11(b)). The accumulated mass

Table 4 Process parameters. Parameter

Value

Gas inlet temperature, Tinlet (◦ C) Flow rate of drying air, V˙ g,dry (L/min) Flow rate of nozzle air, V˙ N (L/min) ˙ sus (g/min) Spraying rate, M Solute mass fraction in coating solution (wt.%) Particle volume fraction in suspension ˛V Particle mass fraction in suspension ˛M Process time (min)

85 300 10 8.5 30 0.31 0.5 10

of sprayed suspension with time is shown in Fig. 11(c). The mass increases gradually, according to the spraying rate. As mentioned previously, the PSD was measured offline by using an optical measurement device, the CAMSIZER-XT. The PSD provides an indication of product quality. It is desirable to achieve a narrow distribution and a uniform coating layer. The evolution of product PSD is used to represent the coating process, see Fig. 11(d). Thus, one can evaluate whether agglomerate effects and overspray occur. The PSD of the coated particles shows newly created particles on the left and on the right side of the PSD compared with the core particles. The left side corresponds to dried droplets without contained particles (overspray) and the right side corresponds to agglomerated particles. The calculated variances of the PSD and Sauter mean diameters d32 of the core and coated particles are summarized in Table 5. The increase in Sauter mean diameter can be explained by building the

Please cite this article in press as: Hampel, N., et al. Coating of finely dispersed particles by two-fluid nozzle. Particuology (2017), http://dx.doi.org/10.1016/j.partic.2017.05.015

G Model

ARTICLE IN PRESS

PARTIC-1050; No. of Pages 14

N. Hampel et al. / Particuology xxx (2017) xxx–xxx

9

Fig. 12. Schematic representation for coating-process mass balance.

Table 5 Calculated Sauter diameter d32 and variance  2 of core and coated particles. Particles

d32 (␮m)

 2 (␮m2 )

Core particles Coated particles

44 57

103 312

coating layers of the particle surfaces and by agglomeration. The variation in calculated variance of the PSD shows the formation of overspray and agglomerates. Despite such effects, the average diameter increases and the shift in central part of the PSD to the right in Fig. 11(d) shows the feasibility of the coating process by means of the innovative application of a two-fluid nozzle. The mass balance was determined. It is assumed that all particles are transported through the nozzle, see Section Transport of suspension to the nozzle. The loss in solid/product X results from tendency of particles to cling to the wall surface of spray dryer and because of small spray-dried droplets, which cannot be separated from the outlet gas as a product in the cyclone. A schematic representation for the mass balance calculations is presented in Fig. 12. The suspension consisted of particles, solute material from the coating solution, and water. During spraying, water was evaporated from the suspension. Particles together with coated solute material were separated in the cyclone as a product. The product was dried in the drying oven. In the mass balance, the sum of the mass of particles Mp and solid material Ms before spraying is compared with that of the mass of particles Mp,dry and the solute material Ms,dry after drying in the oven. The loss of solid/product X can be expressed as: X=

(Mp +Ms ) − (Ms,dry +Mp,dry ) (Mp +Ms )

× 100%.

(10)

The calculated product loss was 5%. This value is relatively low and is considered acceptable. The product quality was determined from scanning electron microscopy (SEM) images. An exemplary SEM image of the product is shown in Fig. 13. In addition to glass particles, a large number of spray-dried droplets (overspray) is visible (A). These occur when a drop contains no internal particles, see Fig. 2. Overspray can be separated using mechanical processes (for example, air separation) and can be used to prepare coating solution. Agglomerated glass particles are visible in the SEM image (B). Some glass particles are completely encased by the solid layer (C), whereas others have no solid layer on the surface (D). Most particles are partially coated (E). The extent of particle-surface coating is quantified in four categories, i.e., 0–25%, 26%–50%, 51%–75%, and 76%–100% coated particle surface. The quantification of coated particles was conducted manually according to the following steps: 1. three images were taken (each image contained 99–422 particles); 2. particles that were partially outside of the image were discarded; 3. overspray was excluded; 4. the total number of particles was determined; 5. each unpaired particle was classified into one of four major categories; 6. the fraction of each category was determined;

Fig. 13. An exemplary SEM image of product obtained from innovative application of a two-fluid nozzle with circled letters representing differently featured entities. A: spray-dried droplets, B: agglomerated particles, C: coated particle, D: uncoated particle, E: partially coated particle.

7. the number of agglomerated was defined and the agglomeration extent as a fraction of the total number of particles was calculated. Disadvantages of this evaluation include the fact that the evaluation is based on two-dimensional areas and no information exists on the rear of the particles, and second the particle orientation is not considered. The optical evaluation between the completely coated and partially coated particles is not optimal, because no two-dimensional information exists on the particles. Particles on each image with a 500-times magnification were assigned to different categories and were counted within each category. The percentage in a category was determined from the total number of particles in the image. The agglomeration ratio was determined in a similar manner. Results from a particle analysis of the process are shown in Fig. 14. The coating degree of the particle surface is distributed between 0 and 100%. Most particles (31%) have coated particle surfaces ranging from 76% to 100%. The fraction of agglomerates in this experiment is less than 10%. This experiment contains a larger fraction of coated particles compared with the agglomerate fraction; 90% of the particles were coated, at least partially. The feasibility of coating fine dispersed particles by the innovative application of a two-fluid nozzle is shown. Results The main aim of this study was to design a new process to coat fine dispersed particles (dp < 50 ␮m). To obtain more knowledge

Please cite this article in press as: Hampel, N., et al. Coating of finely dispersed particles by two-fluid nozzle. Particuology (2017), http://dx.doi.org/10.1016/j.partic.2017.05.015

G Model

ARTICLE IN PRESS

PARTIC-1050; No. of Pages 14

N. Hampel et al. / Particuology xxx (2017) xxx–xxx

10

Fig. 15. Influence of flow rate of nozzle air V˙ N on coating degree and agglomeration effects. Fig. 14. Results of particle analysis for experiment with process parameters summarized in Table 6. Table 6 Calculated Sauter mean diameter d32 , variance  2 of the core particles and coated particles, and loss of solid X obtained by various nozzle air flow ratesV˙ N . Particles

d32 (␮m)

 2 (␮m2 )

X (%)

Core particles Coated particles (V˙ N = 5 L/min) Coated particles (V˙ N = 10 L/min) Coated particles (V˙ N = 15 L/min)

44 113 57 50

103 903 312 308

– 6 5 20

Table 7 Calculated Sauter mean diameter d32 , variance  2 of core particles and coated particles, and loss of solid X by various flow rates of drying air V˙ g,dry . Particles

d32 (␮m)

 2 (␮m2 )

X (%)

Core particles Coated particles (V˙ g,dry = 200 L/min) Coated particles (V˙ g,dry = 300 L/min) Coated particles (V˙ g,dry = 400 L/min)

44 61 57 51

103 387 312 304

– 11 5 9

on the process, experimental tests were conducted with various process parameters. The focus is on four significant process parameters, namely the flow rate of nozzle air V˙ N , the flow rate of drying air V˙ g,dry , the particle mass fraction in the suspension ˛M , and the solute mass fraction in the coating solution , to describe and discuss their effects on the product quality. Nozzle air flow rate The flow rate of nozzle air V˙ N (or gas pressure in the nozzle) influences the droplet size directly (Hampel, 2010; Hede et al., 2008; Schick, 2006; Lefebvre, 1989). An increase in flow rate of nozzle air V˙ N leads to a decrease in droplet size and an increase in the number of droplets. Therefore, this parameter influences agglomerate formation, overspray, and the degree of the coated particle surface during the process (see Fig. 2). Three experiments with various flow rates of nozzle air (V˙ N = 5, 10, and 15 L/min) were carried out. All other process parameters remained as summarized in Table 4 By this parameter variation, suspension viscosity sus and the theoretical liquid-film thickness hCS remain constant (sus = 0.0225 Pa s; hCS = 53 ␮m). The parameters examined were the product PSD, the coating degree of the particle surface, and the fraction of agglomerated particles. The mean particle size d32 and the variance  2 in PSD is given in Table 6 along with a solid material loss X that is calculated from mass balance. The increase in flow rate of nozzle air V˙ N leads to a smaller droplet size, which yields an increase in fraction of spray-dried droplets, see Fig. 2, Case IV. The calculated Sauter mean diameter d32 and variance  2 are given in Table 6 The Sauter diameter d32 is highest for V˙ N = 5 L/min. In this case, liquid droplets can be much larger than the particles. This leads to agglomerate formation, see Fig. 2, Case I. The loss of solid X increases with an increase in the fraction of spray-dried droplets. They are not separated in the cyclone and leave the plant together with the exhaust gas. We compare the coating degree and agglomeration in the three experiments, see Fig. 15. The results from the PSD (Sauter diameter d32 and variance  2 ) on the agglomeration effects are confirmed. The highest agglomerate fraction is observed for V˙ N = 5 L/min. The

Fig. 16. Influence of drying air flow rate V˙ g,dry on coating degree and agglomeration effects.

optimum gas flow in the nozzle is V˙ N = 10 L/min, because a coating degree of between 76% and 100% is achieved in this case. The respective agglomerate ratio is below 10%. A variation in flow rate of the nozzle air V˙ N influences mainly the agglomeration effects and the fraction of spray-dried droplets. There is no strong change in coating degree of the particle surface. SEM images of the experiments are shown in Appendix B, Fig. B1, and the effect of nozzle air flow rate is clearly visible. Drying air flow rate Three experiments with various drying air flow rates V˙ g,dry (200, 300, and 400 L/min) were carried out. All other process parameters remained unchanged, with values according to Table 4 The suspension viscosity sus and the theoretical liquid-film thickness hCS remained constant (sus = 0.0225 Pa s; hCS = 53 ␮m). The increase in flow rate of drying air V˙ g,dry leads to the building of a higher amount of spray-dried droplets and to a decrease in the Sauter diameter d32 (Table 7). Less drying air leads to an incomplete drying of the particle surface, and thus enhances agglomerate formation. Increased agglomerates increase the Sauter diameter d32 and the variance in particle distribution  2 . A comparison of the coating degree and agglomeration ratio of the three experiments is shown in Fig. 16. The highest fraction of agglomerates is observed in the experiment with the lowest flow rate of drying air, V˙ g,dry = 200 L/min. The optimum flow rate of dry-

Please cite this article in press as: Hampel, N., et al. Coating of finely dispersed particles by two-fluid nozzle. Particuology (2017), http://dx.doi.org/10.1016/j.partic.2017.05.015

G Model

ARTICLE IN PRESS

PARTIC-1050; No. of Pages 14

N. Hampel et al. / Particuology xxx (2017) xxx–xxx

11

Table 8 Calculated Sauter mean diameter d32 , variance  2 of the core particles and coated particles, and loss of solid X for experiments with various particle mass fractions in suspension ˛M . Particles

d32 (␮m)

 2 (␮m2 )

X (%)

Core particles Coated particles ˛M = 0.3 (˛V = 0.16) Coated particles ˛M = 0.4 (˛V = 0.23) Coated particles ˛M = 0.5 (˛V = 0.31)

44 52 54 57

103 368 323 312

– 12 7 5

Fig. 17. Influence of particle mass fraction in suspension ˛M on coating degree and agglomeration effects.

ing air appears to be V˙ N = 300 L/min. The highest coating degree of 76%–100% was achieved with a low agglomeration of ∼6% of the product particles. In general, the higher flow rate of drying air V˙ g,dry appears to decrease the particle surface coverage with coating. A partial coating may be favored by the influence of flow forces, as indicated in Fig. 3. SEM images of the three experiments are given in Appendix B, Fig. B2. Particle mass fraction ␣ in the suspension Experiments were conducted with various particle volume/mass fractions in the suspension and all other process parameters remained unchanged. With decreasing the particle fraction in the suspension, the fraction of coating solution also increased. Consequently, the solid and water contents in the suspension increased. In this case, the suspension viscosity decreased (see Fig. 4) and the calculated liquid-film thickness increased (see Fig. 8(c)). The increase in particle fraction is limited by the ability to transport the particles pneumatically through the nozzle. Three experiments with different particle mass fractions ˛M (0.3, 0.4, and 0.5) were analyzed. All other process parameters are summarized in Table 4. The core particle size and product d32 , the variance  2 , and the calculated loss of solid for the three experiments are summarized in Table 8. A decrease in particle mass fraction yielded no significant change in Sauter diameter d32 and the variance  2 results. The calculated loss of solid/product X (Eq. (10)) increased with a decrease in the particle mass fraction in the suspension. A larger solute mass fraction in coating solution leads to a larger proportion of spraydried droplets, see Fig. 2, Case IV. The product loss increases as the spray-dried droplets leave the system with the exhaust gas. The coating degree and agglomeration behavior in the three experiments were compared (Fig. 17). The experiment with a particle mass fraction ˛M = 0.3 presents a high fraction of coated particles at 76%–100%. This results from the higher fraction of coating solution in the suspension. The variation in particle fraction ˛ leads to only a small change in the amount of agglomerates in the product. The largest proportion of overspray was observed in the experiment with a particle fraction of ˛M = 0.3. For economic reasons,

Fig. 18. Relationships of fractions of agglomerates and coated particles with coating degree of 76%–100% on properties of suspension: (a) experimental suspension viscosity sus and (b) calculated film thickness hCS at various particle mass fractions ˛M .

a mass ratio of particles to coating solution in the suspension of 0.3:0.7 suspension is not useful. The dependency between the suspension properties (suspension viscosity sus and liquid-film thickness hCS ) and the fractions of agglomerates and coated particles with coating degree of 76%–100% is presented in Fig. 18. An increase in particle fraction in the suspension increases the suspension viscosity sus and decreases the calculated liquid-film thickness hCS . The increase in viscosity sus leads to less agglomerate formation and particle-surface coating. Particle formation processes (agglomeration and coating) yield a similar trend. The gradual increase in liquid-film thickness hCS tends to result in both agglomeration and partial coating. The aim of this work was to establish parameters that support particle surface coating. Changes in particle fraction that increase the coating degree appeared to result in the formation of agglomerates. This effect was unexpected. SEM images of the experiments are presented in Appendix B, Fig. B3. Solute mass fraction in the coating solution Experimental investigations of the coating process using various solute mass fractions in the coating solution  were conducted, with all other process parameters as given in Table 4. A decrease in solute mass fraction in the coating solution  yields an increase in the water fraction in the suspension, whereas the suspension viscosity sus and liquid-film thickness hCS decrease. The experimentally examined parameters are summarized in Table 9 The increase in Sauter diameter d32 and the variance  2 with less solid in the coating solution can be explained by increased agglomeration because of the larger water content in the system.

Please cite this article in press as: Hampel, N., et al. Coating of finely dispersed particles by two-fluid nozzle. Particuology (2017), http://dx.doi.org/10.1016/j.partic.2017.05.015

G Model

ARTICLE IN PRESS

PARTIC-1050; No. of Pages 14

N. Hampel et al. / Particuology xxx (2017) xxx–xxx

12

Table 9 Calculated Sauter mean diameter d32 , variance  2 of core particles and coated particles, and loss of solid X for various solute mass fractions in coating solution . Particles

d32 (␮m)

 2 (␮m2 )

X (%)

Core particles Coated particles ( = 10 wt%) Coated particles ( = 20 wt%) Coated particles ( = 30 wt%)

44 63 58 57

103 412 322 312

– 10 6 5

Fig. 19. Influence of solute mass fraction in coating solution  on coating degree and agglomeration effects.

Part of the agglomerated particles was deposited on the wall of the spray dryer. Therefore, the increase in product loss by decreasing the solid fraction can be explained. The coating degree of coated particles and extent of agglomeration are compared in Fig. 19. The agglomeration increase with a decrease in solute mass fraction in coating solution can be confirmed. The fraction of coated particles with a coating degree of 76–100% decreases with solute mass fraction in coating solution mainly because of the reduced amount of solid in the suspension to be coated on the particle surface. The two particle formulation processes exhibit reverse trends. The dependency of agglomeration and coating on the suspension properties (suspension viscosity sus and liquid-film thickness hCS ) at various solute mass fractions in the coating solution is presented in Fig. 20. An increase in suspension viscosity sus with more solute in the coating solution leads to a reduction in the agglomerate fraction in the product and to an increase in the fraction of almost completely coated particles, see Fig. 20(a). These effects correspond to the aim in the new process to coat fine dispersed particles. An increase in the solute mass fraction in the coating solutions corresponds to a reduced mass fraction of water in the suspension. For this reason, after spraying the suspension, the water evaporates faster and the agglomeration fraction of the product is reduced. A lower agglomerate fraction yields a higher amount of coated particle surfaces. The liquid-film thickness hCS increases by reducing the solute content in solution . This explains why the agglomeration rate increases and the coating becomes less efficient, see Fig. 20(b). The SEM images of the three experiments are presented in Appendix B, Fig. B4. Summary and conclusions The objective of this study was to realize a new concept for coating fine dispersed particles. The main idea is to implement this process by means of an externally mixing two-fluid nozzle, in which two fluid phases are introduced into two existing nozzle channels. The suspension, which consists of coating solution and particles of core material, is introduced through one of the nozzle channels into the system. The aim is to create droplets that contain one particle only, and by evaporating the water in the droplet, a thin solid film on the particle surface results. The process was implemented in a

Fig. 20. Dependency of fractions of agglomerates and coated particles with coating degree of 76–100% on properties of suspension: (a) experimental suspension viscosity sus and (b) calculated film thickness hCS at various solute mass fractions in coating solution .

spray dryer plant to identify the major parameters that influence the coating and show its feasibility. In the experimental investigations, glass particles were used as a core material. Such particles have a narrow size distribution and are spherical. An aqueous solution of sodium benzoate with different solute mass fractions was used as a coating agent. The viscosities of five different suspensions were measured and calculated by using equations from literature. Even suspensions with a large viscosity can be fed without significant solid losses (∼1%). Liquid-film thickness can be used to calculate the dependency on the properties of the experimental materials. Liquid films that are too thin can result in partially coated particles. Too-viscous liquid films result in undesired agglomeration between single particles. The calculated liquid-film thickness can allow such effects to be assessed. In separate experiments, the nozzle air and drying air flow rates, particle fraction in the suspension, and the solute mass fraction in the coating solution were varied. The air flow rate in the nozzle is an important flow parameter and has the biggest influence on product agglomeration. The particle fraction in the suspension and the solute mass fraction in the coating solution are parameters that affect the suspension properties. An increase in both parameters has an economically positive influence. The analyzed parameters have an intensive impact on product characteristics. The agglomerate fraction of the product increases with an increase in three out of the four investigated parameters. The amount of particles coated to 76%–100% particle surface behaves differently with a variation in different parameters. High gas flow rates in the nozzle do not change the coating degree significantly. An increase in particle fraction leads to more agglomerates

Please cite this article in press as: Hampel, N., et al. Coating of finely dispersed particles by two-fluid nozzle. Particuology (2017), http://dx.doi.org/10.1016/j.partic.2017.05.015

G Model PARTIC-1050; No. of Pages 14

ARTICLE IN PRESS N. Hampel et al. / Particuology xxx (2017) xxx–xxx

in the product. These effects are reduced if the solute mass fraction in the coating solution is increased. This behavior is a consequence of a higher liquid-film thickness on the particle surface, because the particle mass is kept constant. An improvement of the coating degree of the particle surface may be achieved by increasing the solute mass fraction in the coating solution. Agglomeration can be decreased by decreasing the gas flow rate in the dryer. In this way, the fraction of coated particles in the product can be influenced directly and positively.

13

Table A1 Summary of some equations to calculate particle suspension viscosity as a function of solid volume fraction and coating solution viscosity. Author

Equation for suspension viscosity, sus (Pa s)

Einstein (1911) Jogwich (1957) Vand (1948) Eyring et al. (1964) Hatschek (1920)

sus sus sus sus sus

Orr and Dalla Valle (1954)

= CS (1 + 2.5˛V ) = CS (1 + 2.5˛V + 7.54˛2V ) = CS (1 + 2.5˛V + 7.17˛2V + 16.2˛3V ) = CS (1 + 2.5˛V + 10.05˛2V + 0.00273e16.6˛V ) CS = 1/3

sus =

1−˛

V

CS

(1 − ˛V /˛V,max )

1/8

;

˛V,max = 0.54

(A1) (A2) (A3) (A4) (A5) (A6)

2

Kurpiers (1984)

Acknowledgement

Mooney (1951)

The authors gratefully acknowledge the funding of this work by the German Federal Ministry of Science and Education (BMBF) as part of the project WIGRATEC+ (Grant No. 03WKCI3D).

Appendix A. Table A1.

Krieger and Dougherty (1959)

2.5˛V (A7) ] ; ˛V,max = 0.74 2(1 − ˛V /˛V,max ) 2.5˛V sus = CS [exp ]; ˛V,max = 0.74 (A8) 1 − ˛V /˛V,max sus = CS (1 − (˛V /˛V,max )−B˛V,max ); (A9) ˛V,max = 0.74; B = 2.5 for solid spherical particles sus = CS [1 +

Appendix B. Figs. B1–B4.

Fig. B1. SEM images from experiments using various nozzle air flow rates V˙ N : (a) 5, (b) 10, and (c) 15 L/min.

Fig. B2. SEM images from experiments using various drying air flow rates V˙ g,dry : (a) 200, (b) 300, and (c) 400 L/min.

Fig. B3. SEM images from experiments using various particle mass fractions in suspension ˛M : (a) 0.3, (b) 0.4, and (c) 0.5.

Please cite this article in press as: Hampel, N., et al. Coating of finely dispersed particles by two-fluid nozzle. Particuology (2017), http://dx.doi.org/10.1016/j.partic.2017.05.015

G Model PARTIC-1050; No. of Pages 14 14

ARTICLE IN PRESS N. Hampel et al. / Particuology xxx (2017) xxx–xxx

Fig. B4. SEM images from experiments using various solid mass fractions of the coating solution : (a) 10 wt.%, (b) 20 wt.%, and (c) 30 wt.%.

References Chen, Y., Jallo, L., Quintanilla, M. A., & Dave, R. (2010). Characterization of particle and bulk level cohesion reduction of surface modified fine aluminum powders. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 361(1), 66–80. Chen, Y., Yang, J., Dave, R. N., & Pfeffer, R. (2008). Fluidization of coated group C powders. AIChE Journal, 54(1), 104–121. Chen, Y., Yang, J., Mujumdar, A., & Dave, R. (2009). Fluidized bed film coating of cohesive Geldart group C powders. Powder Technology, 189(3), 466–480. Dadkhah, M. (2014). Morphological characterization of agglomerates produced in a spray fluidized bed by X-ray tomography. Doctoral dissertation. Germany: Otto von Guericke University. Dernedde, M., Peglow, M., & Tsotsas, E. (2013). Stochastic modeling of fluidized bed agglomeration: Determination of particle moisture content. Drying Technology, 31(15), 1764–1771. Dewettinck, K., & Huyghebaert, A. (1998). Top-spray fluidized bed coating: Effect of process variables on coating efficiency. LWT-Food Science and Technology, 31(6), 568–575. Ehlers, H., Liu, A., Räikkönen, H., Hatara, J., Antikainen, O., Airaksinen, S., et al. (2009). Granule size control and targeting in pulsed spray fluid bed granulation. International Journal of Pharmaceutics, 377(1), 9–15. Einstein, A. (1911). Berichtigung zu meiner Arbeit: “Eine neue Bestimmung der Moleküldimensionen”. Annalen der Physik, 339(3), 591–592. Eyring, H., Henderson, D., Stover, B. J., & Eyring, E. M. (1964). Statistical mechanics and dynamics. pp. 96–98. New York: Wiley & Sons. Fukumori, Y., Ichikawa, H., Yamaoka, Y., Akaho, E., Takeuchi, Y., Fukuda, T., et al. (1991). Microgranulation and encapsulation of pulverized pharmaceutical powders with ethyl cellulose by the Wurster process. Chemical and Pharmaceutical Bulletin, 39(7), 1806–1812. Geldart, D. (1973). Types of gas fluidization. Powder Technology, 7(5), 285–292. Hamashita, T., Matsuzaki, M., Ono, T., Ono, M., Tsunenari, Y., Aketo, T., et al. (2008). Granulation of core particles suitable for film coating by agitation fluidized bed II. A proposal of a rapid dissolution test for evaluation of bitter taste of ibuprofen. Chemical and Pharmaceutical Bulletin, 56(7), 883–887. Hamashita, T., Nakagawa, Y., Aketo, T., & Watano, S. (2007). Granulation of core particles suitable for film coating by agitation fluidized bed I. Optimum formulation for core particles and development of a novel friability test method. Chemical and Pharmaceutical Bulletin, 55(8), 1169–1174. Hamashita, T., Ono, T., Ono, M., Tsunenari, Y., Aketo, T., & Watano, S. (2009). Granulation of core particles suitable for film coating by agitation fluidized bed III. Effect of scale, agitator rotational speed and blade shape on granule properties and development of a high accuracy scale-up theory. Chemical and Pharmaceutical Bulletin, 57(4), 325–331. Hampel, N. A. (2015). Diskontinuierliches und kontinuierliches Pelletcoating im Wurster-Wirbelschichtprozess. Doctoral dissertation. Germany: Otto von Guericke University. Hampel, R. (2010). Beitrag zur Analyse von kinetischen Einfl¨ ussen auf die Wirbelschicht-Spr¨ uhagglomeration. Doctoral dissertation. Germany: Otto von Guericke University. Hatschek, E. (1920). The viscosity of liquids. Kolloid-Zeitschrift, 27, 163–166. Hede, P. D., Bach, P., & Jensen, A. D. (2008). Two-fluid spray atomisation and pneumatic nozzles for fluid bed coating/agglomeration purposes: A review. Chemical Engineering Science, 63(14), 3821–3842. Hoffmann, T., Rieck, C., Bück, A., Peglow, M., & Tsotsas, E. (2015). Influence of granule porosity during fluidized bed spray granulation. Procedia Engineering, 102, 458–467. Jallo, L. J., Ghoroi, C., Gurumurthy, L., Patel, U., & Davé, R. N. (2012). Improvement of flow and bulk density of pharmaceutical powders using surface modification. International Journal of Pharmaceutics, 423(2), 213–225. Jogwich, A. (1957). Das Fließverhalten von Suspensionen im turbulenten Bereich. Forschung auf dem Gebiet des Ingenieurwesens, 23(3), 81–90. Jono, K., Ichikawa, H., Miyamoto, M., & Fukumori, Y. (2000). A review of particulate design for pharmaceutical powders and their production by spouted bed coating. Powder Technology, 113(3), 269–277.

Ichikawa, M., Shin, T., & Kang, M. S. (1999). Fine structure of the vomeronasal sensory epithelium of Korean goats (Capra hircus). Journal of Reproduction and Development, 45, 81–89. Kawaguchi, T., Sunada, H., Yonezawa, Y., Danjo, K., Hasegawa, M., Makino, T., et al. (2000). Granulation of acetaminophen by a rotating fluidized-bed granulator. Pharmaceutical Development and Technology, 5(2), 141–151. Krieger, I. M., & Dougherty, T. J. (1959). A mechanism for non-Newtonian flow in suspensions of rigid spheres. Transactions of the Society of Rheology, 3(1), 137–152. Kröll, K. (1978). Trockner und Trocknungsverfahren (Trocknungstechnik Band 2). Berlin–Heidelberg–New York: Springer-Verlag. Kunii, D., & Levenspiel, O. (1991). Fluidization engineering. Boston: ButterworthHeinemann. Kurpiers, P. (1984). Wärmeübergang in Ein-und Mehrphasenreaktoren mit ein-und zweistufigen Rührern. Doctoral dissertation. Germany: Dortmund University. Lefebvre, A. H. (1989). Properties of sprays. Particle & Particle Systems Characterization, 6(1–4), 176–186. Medingen. (1966). Prüfgeräte. Gebrauchsanleitung des Höppler-Viskositmeters. Miyadai, N., Higashi, K., Moribe, K., & Yamamoto, K. (2012). Optimization and characterization of direct coating for ibuprofen particles using a composite fluidized bed. Advanced Powder Technology, 23(1), 40–45. Mooney, M. (1951). The viscosity of a concentrated suspension of spherical particles. Journal of Colloid Science, 6(2), 162–170. Orr, C., & Dalla Valle, J. M. (1954). Heat transfer properties of liquid-solid suspensions. Chemical Engineering Progress, Symposium Series, 50, 29–45. Rizkalla, A. A., & Lefebvre, A. H. (1975). The Influence of air and liquid properties on airblast atomization. Journal of Fluids Engineering, 97(3), 316–320. Schick, R. J. (2006). Spray technology reference guide: Understanding drop size. Spraying Systems Co. Bulletin B, 459, 8–16. Schmidt, M., Bück, A., & Tsotsas, E. (2015). Experimental investigation of process stability of continuous spray fluidized bed layering with internal separation. Chemical Engineering Science, 126, 55–66. Schmidt, M., Rieck, C., Bück, A., & Tsotsas, E. (2015). Experimental investigation of process stability of continuous spray fluidized bed layering with external product separation. Chemical Engineering Science, 137, 466–475. Shen, R. W. (1996). Taste masking of ibuprofen by fluid bed coating. U.S. Patent No. 5,552,152. Spray Drying Systems Co. (2000). Engineer’s guide to spray drying technology. Takano, K., Nishii, K., & Horio, M. (2003). Binderless granulation of pharmaceutical fine powders with coarse lactose for dry powder inhalation. Powder Technology, 131(2), 129–138. Thiel, W. J., & Nguyen, L. T. (1982). Fluidized bed granulation of an ordered powder mixture. Journal of Pharmacy and Pharmacology, 34(11), 692–699. To, D., & Davé, R. N. (2016). Fluid bed film coating of fine ibuprofen particles. Powder Technology, 290, 102–113. Tran, T. T. H., Jaskulski, M., & Tsotsas, E. (2016). CFD model considering whey protein inactivation during the skimmed milk spray drying process. 20th international drying symposium. Vand, V. (1948). Viscosity of solutions and suspensions. II. Experimental determination of the viscosity–concentration function of spherical suspensions. The Journal of Physical Chemistry, 52(2), 300–314. Watano, S., Imada, Y., Hamada, K., Wakamatsu, Y., Tanabe, Y., Dave, R. N., et al. (2003). Microgranulation of fine powders by a novel rotating fluidized bed granulator. Powder Technology, 131(2), 250–255. Watano, S., Nakamura, H., Hamada, K., Wakamatsu, Y., Tanabe, Y., Dave, R. N., et al. (2004). Fine particle coating by a novel rotating fluidized bed coater. Powder Technology, 141(3), 172–176. Yang, J., Sliva, A., Banerjee, A., Dave, R. N., & Pfeffer, R. (2005). Dry particle coating for improving the flowability of cohesive powders. Powder Technology, 158(1), 21–33. Zhu, J., & Zhang, H. (2004). Fluidization additives to fine powders. U.S. Patent No. 6,833,185.

Please cite this article in press as: Hampel, N., et al. Coating of finely dispersed particles by two-fluid nozzle. Particuology (2017), http://dx.doi.org/10.1016/j.partic.2017.05.015