processes Article Novel Technique for Coating of Fine Particles Using Fluidized Bed and Aerosol Atomizer Rongyi Zhang *, Torsten Hoffmann and Evangelos Tsotsas Thermal Process Engineering, Otto von Guericke University, Universitäts Platz 2, 39106 Magdeburg, Germany; torsten.hoffmann@ovgu.de (T.H.); evangelos.tsotsas@ovgu.de (E.T.) * Correspondence: rongyi.zhang@ovgu.de Received: 29 October 2020; Accepted: 19 November 2020; Published: 24 November 2020 Abstract: Fine particles are widely used in many industrial fields, and there are many techniques applied for these particles, like electroplating, and chemical and physical vapor deposition. However, in the food and pharmaceutical industries, most coating processes conducted with fluidized bed use core particles with a diameter larger than 200 µm, otherwise agglomerates are formed. This study contributes to the development of a new coating process for fine particles with diameters of around 50 µm. The innovation lies in the combined use of a Wurster fluidized bed and a novel aerosol atomizer. The feasibility of the operation is based on the application of the aerosol atomizer, which generates droplets smaller than 1 µm in diameter. A series of experiments with different coating solutions and glass beads in a 150 mm fluidized bed fed with droplet aerosol supplied from the cone chamber bottom is presented. The quality of the coating product is analyzed by scanning electron microscopy and CAMSIZER® . In this way, the influence of different conditions and core material properties on the product quality were determined. Experimental results showed the coating layer quality getting worse as coating solution viscosity became lower, meanwhile moderate process temperature was found to enhance coating layer formation and quality of that. It was also observed that lower aerosol feed rates help improve the yield of the process. Keywords: coating; aerosol generator; fluidized bed; fine particles 1. Introduction The coating of particulate materials is an important process for various industries, including the chemical, food and pharmaceutical industries [1]. Coating is frequently used for protective or functional purpose [2,3]. Additionally, the visual properties of the particles can also be changed [4]. Therefore, through coating processes, various properties can be achieved, such as stabilization of active substances, enhanced mechanical properties, masking of taste and odor, improved optical appearance, and defined drug release profile in the human body [5]. Among different particulate materials, fine particles have come to be of major interest lately. Many industrial sectors are expected to find applications and take advantage of the new functionalities and many desirable properties attributed to ultrafine particles, e.g., their use as drug delivery carriers [6]. Application of the coating layer to the surface of a particle is a complex process [5]. Typically, a suspension or solution is sprayed onto the particles, after that, every single particle is wrapped by the liquid film. New solid layers are formed after the liquid in the film has evaporated, which results in coating of the particles. Specifically, depending on the apparatus used and the coating material conditions, the process may comprise the following steps: droplet formation, wetting, spreading, evaporation, and drying [7]. Fluidization is one of the most important techniques in the particle formation process because it possesses the advantages of high heat and mass transfer rates, good mixing and homogeneity. With the Processes 2020, 8, 1525; doi:10.3390/pr8121525 www.mdpi.com/journal/processes Processes 2020, 8, 1525 2 of 18 increasing progress of experimental methods and mathematical models in recent years, the spray fluidized bed coating of particulate materials is becoming an established industrial process [8,9]. Despite this, fluidized bed coating without severe agglomeration is typically limited to particle sizes larger than 100 µm [5]. On the one hand, the reason for this is that fine powders fluidize poorly due to their strong interparticle cohesive forces, exhibiting channeling, lifting as a plug and forming “rat holes” when aerated [10]. On the other hand, when using ordinary two-fluid nozzles, the spray is composed of droplets with an average diameter of 40 µm or more [11], which can be considered to be large droplets relative to the particle size. Hence, an over-wetted particle surface will result in strong cohesive forces of liquid bridges that hold particles together after the particles collide with each other during the drying process. These are converted to solid bridges upon drying, fixing the structure of agglomerates that will not break in the subsequent steps [12]. To overcome these issues, conventional methods can be optimized by improving the particles or modifying the fluidized bed. Thereby, powders can be granulated with larger particles in order to enhance the fluidization, drying and coating process in the fluidized bed [13]. Otherwise, fluidization can also be assisted by a variety of mechanisms such as vibration, mechanical agitation, sound and magnetic force to improve the flowability of cohesive particles [6,14,15]. Such approaches are effective in the formation of coating layers, but they also have the disadvantage of limited application ranges. When granulating the fine particles with coarse ones, the high concentration of easily fluidized material will finally influence the product purity. Additionally, most assistance techniques require complicated modifications to the traditional fluidized bed configuration that might demand large additional costs. Recently, several researchers have tried to coat fine particles by using different apparatuses for under various operating conditions. For instance, a Wurster fluidized bed was used by Ichikawa et al. [16] along with 12 µm cornstarch as the core particles and composite of polymers as the coating material. The corresponding product only contained 3 wt.% agglomerates. However, there were still some ineluctable problems when using this method: for example, sufficiently prolonged release properties could not be obtained. Watano et al. [6,17] tried the rotating fluidized bed using fine cornstarch particles. In this case, the cohesive cornstarch particles behaved like particles that could be well fluidized because the centrifugal forces on increased the apparent particle weight. This led to cornstarch being successfully coated nearly without agglomeration. However, this method required a complex apparatus and a significant amount of energy caused by the huge pressure drop through the particles. Chen et al. [18,19] proposed a way in which core materials were pretreated by deposition of nano-particles on the surface of fine particles. The addition of a nano-particle surface coating reduces the cohesive behavior so that fine particles can fluidize in a conventional fluidized bed and good quality coating product may be obtained. Nevertheless, the pretreatment process was inefficient and time-consuming, which is a big drawback in large-scale industrial production. The recent work of Hampel et al. [20] developed a new method that combined two particle formulation processes: spray drying and coating. The idea was to spray fine particles to be coated together with coating solution means of a two-fluid nozzle, aiming at the generation of droplets with only one contained particle at the nozzle outlet. After evaporation of the water in such a droplet, a thin solid film is built on the particle surface. This innovative, low-cost and easy-to-implement method could avoid the occurrence of agglomeration to a certain extent, but partly coated or even uncoated products appeared under certain process conditions. As indicated by the previous experiments conducted in a fluidized bed, a difficulty lies in successfully fluidizing the original fine particles. Previous research by Ichikawa et al. demonstrated that the bottom-spray Wurster bed is an effective way of improving the fluidization behavior of cohesive fine particles [16]. However, as mentioned earlier, large spray droplets lead to the formation of agglomerates, so that individual coating of fine particles still be a challenge and this work is focus on the topic. Processes 2020, 8, 1525 3 of 18 A radical change from conventional spray to aerosol for coating process in fluidized bed is Processes 2020, 8, x FOR PEER REVIEW presented in this work. The aerosol, generated by a recently developed method [21,22], was supplied 3 of 18 into the chamber of a Wurster fluidized bed from the bottom. It was recently shown that ultrathin coating can be easily produced on large particles in a traditional fluidized bed by such an aerosol [23]. coating can be easily produced on large particles in a traditional fluidized bed by such an aerosol [23]. Here, the coating of fine particles was in focus. The working hypothesis was that aerosol with Here, the coating of fine particles was in focus. The working hypothesis was that aerosol with diameters diameters 10–100 times smaller than regular sprays would be more suitable for fine particles and 10–100 times smaller than regular sprays would be more suitable for fine particles and enable uniform enable uniform coating layers and without the excessive agglomeration that traditional spray coating layers and without the excessive agglomeration that traditional spray fluidized bed particle fluidized bed particle coating systems show. coating systems show. In this work, investigations will be presented on the effect that operating conditions such as core In this work, investigations will be presented on the effect that operating conditions such as material size, air temperature, coating material concentration and the presence of the additive in core material size, air temperature, coating material concentration and the presence of the additive in coating solution have on efficacy of the presented process to produce coated fine particles. Finally, coating solution have on efficacy of the presented process to produce coated fine particles. Finally, the influence of operating conditions on the resulting product quality will be described and discussed; the influence of operating conditions on the resulting product quality will be described and discussed; the strengths and weaknesses of the novel aerosol coating process to its conventional counterpart will the strengths and weaknesses of the novel aerosol coating process to its conventional counterpart will be summarized. be summarized. 2. Experimental 2. Experimental 2.1. Aerosol Generation 2.1. Aerosol Generation A simple and reliable technique for the generation aerosolwas droplets was recently developed A simple and reliable technique for the generation of aerosolof droplets recently developed and and investigated by Mezhericher [21,22]. Compared to traditional liquid atomization technologies investigated by Mezhericher [21,22]. Compared to traditional liquid atomization technologies [24], [24], this method based on shear-driven disintegration gas jets of thin films by this method is based on isshear-driven disintegration by gas jets by of thin liquid filmsliquid formed by formed gas on a liquid The main part isofanthe device is antube elastic rubber tube which is bubblesgas on bubbles a liquid surface. The surface. main part of the device elastic rubber which is perforated perforated with ostioles along the tube at several axial positions. The tube is horizontally placed and with ostioles along the tube at several axial positions. The tube is horizontally placed and partially partially in with a vessel filledsolution: with coating theoflower part is of immersed the tube is in immersed in submerged in asubmerged vessel filled coating the solution: lower part the tube the the liquid, and the upper portion of the tube is exposed to the air, see Figure 1a. The air inlet tube is liquid, and the upper portion of the tube is exposed to the air, see Figure 1a. The air inlet tube is connected to compressed air, and the stable compressed air is discharged through the perforated connected to compressed air, and the stable compressed air is discharged through the perforated ostioles. many bubbles small bubbles are caused the compressed air through the part lower ostioles. Many small which which are caused by the by compressed air through the lower ofpart the of the tube, then come up to the liquid surface and create a foam of thin spherical liquid films near the upper tube, then come up to the liquid surface and create a foam of thin spherical liquid films near the part of the tube. Soon afterwards, the air discharged through the upper ostioles breaks upper part of the tube. Soon afterwards, the air discharged through the upper ostioles breaks and and disintegrates these into foams into aerosol droplets, as schematically illustrated in Figure 1b. In this study, disintegrates these foams aerosol droplets, as schematically illustrated in Figure 1b. In this study, a self-made aerosol generator developed based on this principle was used in the experiments. a self-made aerosol generator developed based on this principle was used in the experiments. (a) (b) Figure 1. (a) Simultaneous production of bubbles, micro-sprays fine dropletsinina a crossFigure 1. (a) Simultaneous production of bubbles, gas gas jets jets andand micro-sprays of of fine droplets cross-section ofof the atomizing tube [25]; schematic of of novel aerosol generator. section the atomizing tube [25](b) ; (b) schematic novel aerosol generator. The central part ofpart the device was a 220 polyvinyl chloride tube oftube 12 mm The central of the device wasmm a 220 mm polyvinyl chloride of 12outer mm diameter. outer diameter. The 1.5The mm1.5 walls the tube were with 0.6 mm0.6 ostioles along the tube at mmofwalls of the tubeperforated were perforated with mm ostioles along thecircumference tube circumference at four axial positions (38 orifices per line, 144 in total). The tube was horizontally oriented and partially four axial positions (38 orifices per line, 144 in total). The tube was horizontally oriented and partially submerged into a plastic vessel filled tap water coating solution. This vessel dimensions submerged into a plastic vesselwith filled with tapor water or coating solution. This had vessel had dimensions of 16.5 cm (diameter) × 18.5 cm (height) and a circular outlet of 4 cm at its upper horizontal wall. A flexible PVC tube with a length of 165 cm connected the outlet of the container and the fluidized bed, so that droplets and gas could flow out horizontally from the container. Droplet size distributions produced by the aerosol generator were characterized separately by means of a laser diffraction system (Particle Analyzer LA-960, Retsch Technologies, Haan, Germany) Processes 2020, 8, 1525 4 of 18 of 16.5 cm (diameter) × 18.5 cm (height) and a circular outlet of 4 cm at its upper horizontal wall. A flexible PVC tube with a length of 165 cm connected the outlet of the container and the fluidized bed, so that droplets and gas could flow out horizontally from the container. Processes 2020, xxFOR PEER 4 4ofof1818 Processes 2020,8,8, FOR PEERREVIEW REVIEW Droplet size distributions produced by the aerosol generator were characterized separately by means of a laser diffraction system (Particle Analyzer LA-960, Retsch Technologies, Haan, Germany) Measurement results obtained for different coating solutions are shown ininFigure 2a. ensure Measurementresults resultsobtained obtainedfor fordifferent differentcoating coatingsolutions solutionsare areshown shownin Figure2a. 2a.To Toensure ensureaaa Measurement Figure To constant liquid level, solution was pumped totothe aerosol atomizer through peristaltic pump from constantliquid liquidlevel, level,solution solutionwas waspumped pumpedto theaerosol aerosolatomizer atomizerthrough throughaaaperistaltic peristalticpump pumpfrom from constant the another container. Then, the flow rate of the droplets was calculated by measuring the amount of another container. Then, the flow rate of the droplets was calculated by measuring the amount another container. Then, the flow rate of the droplets was calculated by measuring the amount of liquidof liquid left in the storage tank and the corresponding time interval. Figure 2b indicates the flow rate liquid leftstorage in the storage the corresponding time interval. Figure 2b indicates flow rate left in the tank andtank the and corresponding time interval. Figure 2b indicates the flowthe rate of the ofofthe aerosol generator under different operating conditions. The volumetric size distributions of the generator aerosol generator under different conditions. The volumetric size distributions aerosol under different operatingoperating conditions. The volumetric size distributions of Figure 2aof Figure 2a demonstrate that aerosol droplet diameters range from 0.1 μm to 10 μm, and the mean Figure 2a demonstrate aerosol droplet diameters 0.1 and μm the to 10 μm,droplet and the mean demonstrate that aerosol that droplet diameters range from 0.1range µm tofrom 10 µm, mean size of droplet size ofofthe aerosol used ininthe experiments isissmaller than 11μm. Figure 2b shows that the droplet size the aerosol used the experiments smaller than μm. Figure 2b shows that the the aerosol used in the experiments is smaller than 1 µm. Figure 2b shows that the liquid flow rate of liquid flow rate of the generated aerosol increases with atomizing pressure. liquid flow rate of theincreases generated aerosol increases with atomizing pressure. the generated aerosol with atomizing pressure. (a) (a) (b) (b) Figure 2.2.(a) (a) Volumetric droplet size distributions aerosols generated 55bar Figure2. (a)Volumetric Volumetric droplet size distributions indifferent different aerosols generated baratomizing atomizing Figure droplet size distributions inin different aerosols generated at 5atat bar atomizing air air pressure; (b) aerosol flow rate vs. atomizing pressure for different liquid solutions. air pressure; (b) aerosol rateatomizing vs. atomizing pressure for different solutions. pressure; (b) aerosol flow flow rate vs. pressure for different liquidliquid solutions. 2.2. 2.2. 2.2.Materials Materials 2.2.1. Core Material 2.2.1. 2.2.1.Core CoreMaterial Material To compare the effects of different particle sizes on the new coating technology, two different To Tocompare comparethe theeffects effectsofofdifferent differentparticle particlesizes sizeson onthe thenew newcoating coatingtechnology, technology,two twodifferent different sizes of glass particles (Cerablast GmbH, Löchgau, Germany) were used as the core material in the sizes of glass particles (Cerablast GmbH, Löchgau, Germany) were used as the core material sizes of glass particles (Cerablast GmbH, Löchgau, Germany) were used as the core materialininthe the coating experiments. Particle Particle size distributions distributions (PSD) were were measured offline by CAMSIZER-XT coating coatingexperiments. experiments. Particlesize size distributions(PSD) (PSD) weremeasured measuredoffline offlineby by CAMSIZER-XT CAMSIZER-XT (Retsch Technologies, Germany). Measured PSD results for each kind of particles are given inin Figure 3. (Retsch (RetschTechnologies, Technologies,Germany). Germany).Measured MeasuredPSD PSDresults resultsfor foreach eachkind kindofofparticles particlesare aregiven given inFigure Figure The mean diameter was 63 µm and 202 µm, respectively. In both cases, the density of the core particle 3.3.The diameter was 63 μm and 202 μm, respectively. InInboth cases, the density ofofthe core Themean mean diameter was 63 μm and 202 μm, respectively. both cases, the density the core 3 and sphericity is 0.94. is 2500 kg/m 3 and sphericity is 0.94. particle is 2500 kg/m particle is 2500 kg/m3 and sphericity is 0.94. Figure 3. Initial normalized particle size distributions of the two groups of glass particles. Figure Figure3.3.Initial Initialnormalized normalizedparticle particlesize sizedistributions distributionsofofthe thetwo twogroups groupsofofglass glassparticles. particles. 2.2.2. 2.2.2.Coating CoatingSolution Solution Sodium Sodiumbenzoate benzoatesolution solutionwas wasused usedasasaacoating coatingagent agentininthese theseexperiments. experiments.Sodium Sodiumbenzoate benzoate (NaB, C 7H5NaO2, Trigon Chemie GmbH, Schlüchtern, Germany) is a water-soluble material which is (NaB, C7H5NaO2, Trigon Chemie GmbH, Schlüchtern, Germany) is a water-soluble material which is used widely as a preservative in the food industry, but also in academic coating research [26]. To Processes 2020, 8, 1525 5 of 18 2.2.2. Coating Solution Sodium benzoate solution was used as a coating agent in these experiments. Sodium benzoate (NaB, C7 H5 NaO2 , Trigon Chemie GmbH, Schlüchtern, Germany) is a water-soluble material which is Processes 2020, 8, x FOR PEER REVIEW 5 of 18 used widely as a preservative in the food industry, but also in academic coating research [26]. To analyze the influence of viscosity of the coating solution, different amounts of hydroxy-propyl-methyl-cellulose methyl-cellulose (HPMC, trade name Pharmacoat 606, from Shin-Etsu, Tokyo, Japan) were added (HPMC, trade name Pharmacoat 606, from Shin-Etsu, Tokyo, Japan) were added into the solution. into the solution. The coating solution density at 20 °C was measured by using a density measuring The coating solution density at 20 ◦ C was measured by using a density measuring device DMA 58 device DMA 58 (Anton Paar GmbH, Graz, Austria). A rheometer MCR 302 (Anton Paar GmbH, Graz, (Anton Paar GmbH, Graz, Austria). A rheometer MCR 302 (Anton Paar GmbH, Graz, Austria) was Austria) was used to measure the dynamic viscosity of different coating solution with various NaB used to measure the dynamic viscosity of different coating solution with various NaB content. Both of content. Both of the devices were calibrated by using distilled water. The solution composition and the devices were calibrated by using distilled water. The solution composition and measured properties measured properties are summarized in Table 1. are summarized in Table 1. Table1.1. Material Material properties propertiesof ofcoating coatingsolution. solution. Table Solution Composition 1 Density, ρcs(kg/m3)3 Viscosity, ηcs(Pa·S) Viscosity, ηcs (Pa·S) Solution Composition 1 Density, ρcs (kg/m ) 30% NaB 1128.14 0.0039 1128.14 0.0039 30%30% NaBNaB + 0.5% HPMC 1129.98 0.0062 30%30% NaBNaB + 0.5% HPMC 1129.98 0.0062 + 1.0% HPMC 1130.01 0.0111 30%30% NaBNaB + 1.0% HPMC 1130.01 0.0111 + 1.5% HPMC 1131.90 0.0618 30% NaB + 1.5% HPMC 1131.90 0.0618 1 The composition of solution is based on mass fraction (wt%). 1 The composition of solution is based on mass fraction (wt%). 2.3. Fluidized Bed Coating Process 2.3. Fluidized Bed Coating Process The coating process of fine particles was conducted in a laboratory-scale fluidized bed (GPCG 1 The coating process of fine particles was conducted in a laboratory-scale fluidized bed (GPCG 1 in in Wurster version, Glatt GmbH, Weimar, Germany). The lab-scale set up use the distributer plate Wurster version, Glatt GmbH, Weimar, Germany). The lab-scale set up use the distributer plate which which has the diameter of 150 mm and a 550 mm cone fluidized chamber. This setup was modified has the diameter of 150 mm and a 550 mm cone fluidized chamber. This setup was modified to to incorporate a bottom inlet into chamberproviding providingthe theinflow inflowofofaerosol aerosoland andmixing mixing with with the the incorporate a bottom inlet into thethe chamber fluidized particles. particles. The The aerosol aerosol atomizer atomizer in in the the experiments experiments can can well well assume assume the the role role of of the the nozzle nozzle in in fluidized the ordinary ordinary Wurster Wurster fluidized fluidized bed bed (see (see Figure Figure 4). 4). In In the the course course of of the the study, study, the the distance distance between between the Wurster rises tube and distributer plate was set as 5cm. Different length of copper tubes (inner Wurster rises tube and distributer plate was set as 5cm. Different length of copper tubes (inner diameter: diameter: 2.2installed cm) wereinside installed therises Wurster tubeplace and take place of the to introduce 2.2 cm) were the inside Wurster tube rises and take of the nozzle to nozzle introduce solution in the fluidized bed. solution in the fluidized bed. Figure 4. 4. Scheme Scheme of of the the experimental experimental setup setup using using aerosols aerosols for for particle particle coating coating in in fluidized fluidizedbed. bed. Figure During the the experiments, experiments, the the aerosol aerosol and and compressed compressed air air coming coming from from the the aerosol aerosol generator generator were were During introduced introduced into into the the fluidized fluidized bed bed system system from from the the bottom bottom through through aa copper copper inlet inlet tube. tube. To To prevent prevent particles from falling back into the copper tube, its top of was closed. Ten openings with a diameter of particles from falling back into the copper tube, its top of was closed. Ten openings with a diameter 0.5 cmcm each were evenly arranged at at a distance of 0.5 each were evenly arranged a distanceofof1 1cm cmfrom fromthe thetop. top.To Tostudy studythe theeffect effectof of different different opening heights on the experimental results, different lengths of tubes were used in the experiments. Namely 6, 7, and 8 cm. In the beginning, 1 kg of core particles were placed in the fluidized bed and fluidized by hot air at a mass flow rate of 15 kg/h for 63 μm particles, 23 kg/h for 202 μm particles. After the temperature in the product vessel was stabilized, the coating solution was sprayed into the fluidized bed. The Processes 2020, 8, 1525 6 of 18 opening heights on the experimental results, different lengths of tubes were used in the experiments. Namely 6, 7, and 8 cm. In the beginning, 1 kg of core particles were placed in the fluidized bed and fluidized by hot air at Processes 2020, 8, x of FOR of 18 a mass flow rate 15PEER kg/hREVIEW for 63 µm particles, 23 kg/h for 202 µm particles. After the temperature6 in the product vessel was stabilized, the coating solution was sprayed into the fluidized bed. The aerosol are listed in operated Table 2. Every 10 compressed min, a sample takentemperature. from the fluidized bed. Every was atomizer was by 5 bar airwas at room A peristaltic pumpsample adjusted kept in a container for further analysis. After 3 h of operation, the experiment was finished. the liquid level in the coating solution vessel. The operating parameters of each experiment are listed in Table 2. Every 10 min, a sample was taken from the fluidized bed. Every sample was kept in a 2. Experimental plan. container for further analysis. After 3 h Table of operation, the experiment was finished. Coating NaB HPMC Water Table 2. Experimental plan. Material Percentage Percentage Percentage 1 Glass (63 μm) 30% 0% 70% NaB HPMC Water Coating 2 Serial Glass (202 μm)Material 30% Percentage 0%Percentage 70% Percentage 3 1Glass (63Glass μm)(63 µm) 30% 0% 70%70% 30% 0% 4 2Glass (63 μm) 30% 0% 70%70% Glass (202 µm) 30% 0% 5 3Glass (63Glass μm)(63 µm) 30% 0% 70%70% 30% 0% 30% 0% 6 4Glass (63Glass μm)(63 µm) 30% 0% 70%70% 30% 70% 7 5Glass (63Glass μm)(63 µm) 30% 0.5% 0% 69.5% 6 Glass (63 µm) 30% 0% 70% 8 7Glass (63Glass μm)(63 µm) 30% 1% 0.5% 69% 30% 69.5% 9 8Glass (63Glass μm)(63 µm) 30% 1.5% 1% 68.5% 30% 69% 30% 68.5% 10 9Glass (63Glass μm)(63 µm) 30% 0% 1.5% 70% 30% 0% 11 10Glass (63Glass μm)(63 µm) 30% 0% 70%70% Serial 11 Glass (63 µm) 30% 0% 70% Air Inlet Copper Tube Temperature Length 70 °C 8 cm Air Inlet Copper Tube 70 °C 8 Temperature Length cm 50 °C 8 cm ◦ 70 C 8 cm ◦ 60 °C 70 C 8 cm 8 cm 80◦ C °C 50 8 cm 8 cm ◦ 60 8 cm 8 cm 90 C °C 80 ◦ C 8 cm 70◦ °C 8 cm 90 C 8 cm 70 °C 70 ◦ C 8 cm 8 cm ◦ 70 °C 70 C 8 cm 8 cm 70 8 cm 7 cm 70◦ C °C ◦ 70 7 cm 6 cm 70 C °C 70 ◦ C 6 cm 2.4. Characterization Methods 2.4. Characterization Methods The PSD of each sample taken during the experiment was measured offline by using an optical The PSD of device, each sample taken during the experiment was measured offline using opticalan measurement the CAMSIZER-XT (Retsch Technologies, Germany). Theby PSD is inan general measurement device, of theproduct CAMSIZER-XT (Retsch Technologies, Germany). The PSD is in general an important measure quality. Moreover, the accumulated differences in volume frequency important measure of product quality. Moreover, the accumulated differences in volume frequency between product PSD, fi,pr, and raw material PSD, fi,raw, above the point, m, of intersection of those between product PSD, fi,prto, and raw material PSD, fi,raw , above theThis point, m, of intersection of those two curves can be used evaluate the degree of agglomeration. corresponds to the blue area of two curves can be used to evaluate the degree of agglomeration. This corresponds to the blue area Figure 5 and has been proposed in order to quantify agglomeration by Chen et al. [27] as well asofby Figure 5 and has[20]. beenIt proposed in order be to quantify agglomeration by Chen et al. [27] as well as byto Hample et al. should, however, noted that size enlargement by coating is also captured Hample et al. [20].by It ϕ should, however, be noted (1). that size enlargement by coating is also captured to a a certain extent obtained from Equation certain extent by φ obtained from Equation (1). 𝜙X = (𝑓 , − 𝑓, ), (1) φ= fi,pr − fi,raw , (1) i≥m Figure 5. Principle for calculating the agglomeration ratio. Figure 5. Principle for calculating the agglomeration ratio. Moreover, the extent of particle surface coating, which is named coating coverage here, was determined from scanning electron microscopy (SEM) images, taken with a Phenom G2 Pro microscope. Each image used for this purpose contained 35–40 small particles and had a resolution of 1024 × 1024 pixels sufficient to identify the coated area. Three images with, in total, 100 particles Processes 2020, 8, 1525 7 of 18 Moreover, the extent of particle surface coating, which is named coating coverage here, was determined from scanning electron microscopy (SEM) images, taken with a Phenom G2 Pro microscope. Each image used for this purpose contained 35–40 small particles and had a resolution of 1024 × 1024 pixels sufficient to identify the coated area. Three images with, in total, 100 particles were Processes 2020, 8, x FOR PEER REVIEW 7 of 18 typically used for each sample. An analysis was conducted with image identification tools in Python. The image analysis procedure shown Figure a first step, image preprocessing was The image analysis procedure is is shown in in Figure 6. 6. In In a first step, thethe image preprocessing was performed by Hough circle detection, which can identify and locate the position of particles in performed by Hough circle detection, which can identify and locate the position of particles in thethe original image. Particles that were partially outside image were discarded. Figure shows original image. Particles that were partially outside of of thethe image were discarded. Figure 6a6a shows thethe identified particles in the original SEM picture. Applying the particle areas obtained by the pretreatment identified particles in the original SEM picture. Applying the particle areas obtained by the of the original image, the particles were isolatedwere and the background cleared. Circle areas were then pretreatment of the original image, the particles isolated and thewas background was cleared. Circle measured, images were converted to 8-unit images,to and a threshold to binary) was applied. areas were then measured, images were converted 8-unit images,(greyscale and a threshold (greyscale to The above step successfully removed the noise caused by reflection. Subsequently, the coating coverage binary) was applied. The above step successfully removed the noise caused by reflection. was calculated the proportion black pixels within theproportion circular image areaspixels that correspond Subsequently, theascoating coverageofwas calculated as the of black within the to identified particles, seecorrespond Figure 6b. To eliminate particles, the impact ofFigure the inability accurately circular image areas that to identified see 6b. To to eliminate the identify impact ofthe edge pixels of the particles on the results, only the central part of each particle in the original image the inability to accurately identify the edge pixels of the particles on the results, only the central part in the calculation thatused is, the region whose radius was 80% initial radius. ofwas eachused particle in the original process, image was in the calculation process, that is, of thethe region whose Please note that SEM images are planar, so there would be errors in the directly calculated coverage radius was 80% of the initial radius. Please note that SEM images are planar, so there would be errors the curvature of coverage the particles. eliminate this of effect, a geometrical correction was applied in due the to directly calculated due To to the curvature the particles. To eliminate this effect, a to the pictures. The respective change of S (surface area on sphere) to S (visible area) with 3D 2D geometrical correction was applied to the pictures. The respective change of S3D (surface area onthe distance particle center can be seen in Figure 6c. In this study, to eliminate the blurring of the sphere) to Sfrom 2D (visible area) with the distance from particle center can be seen in Figure 6c. In this circular border, evaluation hasof been by dividing the central area (80% based diameter) study, to eliminate the blurring the conducted circular border, evaluation has been conducted byon dividing the of each particle into ten parts, see Figure 6d. central area (80% based on diameter) of each particle into ten parts, see Figure 6d. (a) (b) (c) (d) Figure 6. (a) Pretreatment of SEM pictures byby Hough circle detection. (b)(b) Processed binary picture forfor Figure 6. (a) Pretreatment of SEM pictures Hough circle detection. Processed binary picture to Sto2DSwith respect to the distance to particle center. thethe calculation of of coating coverage. (c)(c) Ratio of of S3DS3D calculation coating coverage. Ratio respect to the distance to particle center. 2D with Area subdivision image error elimination method. (d)(d) Area subdivision of of thethe image error elimination method. For further analysis, the yield of the coating process Y was also calculated. For this purpose, at first, each of the samples taken from the product was dried in an oven to get the dry mass M1 of coated particles accurately. Subsequently, the coating layers were totally removed by soaking and stirring the sample in distilled water for 24 h. The second desired mass M2 was obtained after the core Processes 2020, 8, 1525 8 of 18 For further analysis, the yield of the coating process Y was also calculated. For this purpose, at first, each of the samples taken from the product was dried in an oven to get the dry mass M1 of coated particles accurately. Subsequently, the coating layers were totally removed by soaking and stirring the sample in distilled water for 24 h. The second desired mass M2 was obtained after the core particles dried again. Combined the total sprayed amount Ms and raw material mass Mr process yield can then be calculated as: (M1 − M2 )/M2 Y= , (2) Rcoating,max Rcoating,max = Ms , Mr (3) 3. Results and Discussion The coating process in a fluidized bed is complex. It involves an ingenious balance between mixing, wetting and drying, which depends on several parameters such as air temperature and flow rate, sprayed solution rate, coating solution composition and many other factors [1,28–30]. All the above variables have a decisive effect on the quality of the product. Comprehensive experimental and numerical research has been done in the past few years for large particles coated in fluidized beds with conventional nozzles to explain the effect of these variables. From the past studies it is clear that: (1) the rise of inlet air temperature enhances the drying of wet particles during the coating; (2) more uniform dispersion of coating solution droplets in the equipment results in more uniform layers on the particle surfaces; (3) growth of coating layer thickness depends on the flow rate of the coating solution. Despite the many results on the coating of large particles (diameter > 200 µm) in the literature, less attention has been paid to the investigation of the coating parameters of fine particles. However, with the help of the novel aerosol atomizer, the coating of fine particles is easier, so it is possible to make a systematic investigation of the effects of important operating parameters on the coating performance. 3.1. Effect of Particle Size As mentioned earlier, large particles and small particles usually behave differently in fluidized bed coating. This difference is usually caused by the imperfect fluidization of small particles. In this study, in order to enhance the fluidization of small particles, the Wurster fluidized bed was used. At the same time, the new aerosol generator was also applied in the coating process for both kinds of particles. The experimental results show that both large and small particles could be coated under the usage of aerosol, but there are still some differences. Particle size distributions of uncoated and coated glass particles are plotted in Figure 7a. Average particle sizes of small and large coated particles are 66 µm and 215 µm respectively, with the original particles at 63 µm and 202 µm, as mentioned before. This implied that most of the particles were individually coated. In addition, the agglomeration rate φ were 5.4% and 6.8%, respectively, when using the calculation method from Section 2.4. These results all prove that both large particles and small particles could be successfully coated without severe agglomeration by the novel method. Figure 7b shows the evolution of the average particle diameter for the product samples. The absolute growth value of large particles is higher than that of small particles. On the other hand, the growth interval of small particles is mainly concentrated in the first 40 min and particle diameter is stabilized after that, while large particles can continue to grow for nearly 90 min. the original particles at 63 μm and 202 μm, as mentioned before. This implied that most of the particles were individually coated. In addition, the agglomeration rate ϕ were 5.4% and 6.8%, respectively, when using the calculation method from Section 2.4. These results all prove that both large particles and small particles could be successfully coated without severe agglomeration by the Processes 2020, 8, 1525 9 of 18 novel method. Processes 2020, 8, x FOR PEER REVIEW 9 of 18 Figure 7b shows the evolution of the average particle diameter for the product samples. The (a)of large particles is higher than that of small particles. (b) On the other hand, the absolute growth value growth interval of small is mainly concentrated in theparticle first 40diameter min andd1,3 particle diameter is Figure (a)Particle Particle sizeparticles distributions, and(b) (b) evolutionofofaverage average particle diameter d 1,3for forparticles particles Figure 7.7.(a) size distributions, and evolution stabilized after that, while large particles can continue to grow for nearly 90 min. different size. ofofdifferent size. SEM images with different magnification of product (shown in Figure A1 of Appendix A) reveal SEM images different of product in proving Figure A1the of Appendix revealof the larger and with smaller glass magnification beads were both coated(shown by NaB, successful A) coating the larger and smaller glass beads were both byThe NaB, proving the successful coating of particles by the novel coating process withcoated aerosol. corresponding coating coverage asparticles shown in byFigure the novel coating process with and aerosol. The corresponding coating coveragecovered, as shown in Figure 8 8 reveals that both large small particles were almost completely with the mean reveals that both large and small particles were almost completely covered, with the mean coverage coverage value of large particles being 99.2% and that of small particles being 99%. Since this is close value of large and that of small particles 99%. Since this is close to unity, to unity, the particles coveragebeing had a99.2% very small standard deviation inbeing both cases. the coverage had a very small standard deviation in both cases. Figure 8. Coverage comparison based on particle size. Figure 8. Coverage comparison based on particle size. Smaller particle size results in a relatively lower yield, as shown in Table 3; 13.59% for glass Smaller size results a relatively lower yield, as shown in Table glass cores cores of 63 µmparticle as compared with in 21.51% for glass cores of 202 µm. The loss 3; of13.59% coatingfor material is of 63 μm as compared with 21.51% for glass cores of 202 μm. The loss of coating material is mainly mainly attributed to the collected fines in the filter above the fluidized bed. Because the droplet size collectedsmall, fines itinwas the easy filter to above the fluidized thetodroplet size the ofattributed the aerosoltoisthe extremely be dried by hot airbed. andBecause entrained the filter byof the aerosol is extremely small, it was easy to be dried by hot air and entrained to the filter by the fluidization gas. fluidization gas. 1 1 Table 3. Yield of aerosol coating process with different particles 1 . Table 3. Yield of aerosol coating process with different particles 1. Particle Type Mass of Cores (g) Coating Mass (g) Yield (%) 2 Particle Type Mass of Cores (g) Coating Mass (g) Yield (%) 2 Small particle 2.5010 ± 0.55 0.0849 ± 0.017 13.59 Small particle 2.5010 ± 0.55 0.0849 ± 0.017 13.59 Large particle 2.0592 ± 0.16 0.1122 ± 0.012 21.51 Large particle 2.0592 ± 0.16 0.1122 ± 0.012 21.51 Total mass of glass beads: 1000 g; Mean aerosol spray rate: 282 g/h. 2 Average value from three samples. Total mass of glass beads: 1000 g; Mean aerosol spray rate: 282 g/h. 2 Average value from three samples. From the above results, it can be seen that both large and small particles can be properly From the above results, it can be seen that both large and small particles can be properly fluidized in the modified Wurster equipment, and uniform coating can be obtained when the aerosol fluidized in the modified Wurster equipment, and uniform coating can be obtained when the aerosol generator replaces the conventional two-fluid nozzle. However, it is worth noting that due to the generator replaces the conventional two-fluid nozzle. However, it is worth noting that due to the extremely small size of the aerosol, it takes a longer time for the core particles to obtain an almost extremely small size of the aerosol, it takes a longer time for the core particles to obtain an almost complete coating compare to the conventional coating process. Moreover, according to experimental observations, small particles would rather stick to the wall of the fluidized bed due to electrostatic and other micro-scale forces, which might explain why the coverage rate is slightly lower than that of large particles. Processes 2020, 8, 1525 10 of 18 complete coating compare to the conventional coating process. Moreover, according to experimental observations, small particles would rather stick to the wall of the fluidized bed due to electrostatic and other micro-scale forces, which might explain why the coverage rate is slightly lower than that of large particles. 3.2. Effect of Air Inlet Temperature Hot fluidized gas passing through the distributor plate and the fluidization chamber is responsible for drying of the particles by heat and mass transfer to the fluidized bed. Based on previous experience, higher temperature during the process favors drying, leading to rapid evaporation of droplets. Processes 2020, 8, x FOR PEER REVIEW 10 of 18 In comparison, when temperature is lower, the evaporation rate of the coating solution decreased, leading to slight agglomeration. A moderate coating process in the modified fluidized bed is only coating solution decreased, leading to slight agglomeration. A moderate coating process in the obtained when wetting and drying reached a proper balance. The above conclusions are also applicable modified fluidized bed is only obtained when wetting and drying reached a proper balance. The to the use of traditional nozzles for large particles. However, the balance is more difficult to maintain above conclusions are also applicable to the use of traditional nozzles for large particles. However, when using fine particles, and the temperature control the coating of fine fluidized particles using the the balance is more difficult to maintain when using fine particles, and the temperature control the aerosol generator is more challenging. coating of fine fluidized particles using the aerosol generator is more challenging. The particle size distribution results for small glass particles when air inlet temperature increased The particle size distribution results for small glass particles when air inlet temperature from 50 ◦ C to 90 ◦ C are plotted in Figure 9a. Product size does not have obvious differences as increased from 50 °C to 90 °C are plotted in Figure 9a. Product size does not have obvious differences temperature changes; mean particle size is 67.5 µm, 66.8 µm, 66 µm, 65.7 µm and 65.8 µm respectively. as temperature changes; mean particle size is 67.5 μm, 66.8 μm, 66 μm, 65.7 μm and 65.8 μm Figure 9a also indicates the slight difference of agglomeration ratios (13.1%, 8.6%, 5.4%, 3.6% and respectively. Figure 9a also indicates the slight difference of agglomeration ratios (13.1%, 8.6%, 5.4%, 3.4% based on the PSD results). It can be seen that the overall agglomeration rate is at a relatively 3.6% and 3.4% based on the PSD results). It can be seen that the overall agglomeration rate is at a low level, but there are still more agglomerated particles under low temperature conditions than at relatively low level, but there are still more agglomerated particles under low temperature conditions high temperature. than at high temperature. (a) (b) Figure (a)Particle Particlesize size distributions, (b) evolution of average particle diameter d1,3 at Figure9.9. (a) distributions, andand (b) evolution of average particle diameter d1,3 at different different temperatures. temperatures. As in Figure 9b, in each is achieved at the beginning the experiment. Asdepicted depicted in Figure 9b, in case, each faster case, growth faster growth is achieved at the of beginning of the After 60 min, average size is still increasing. The increasing. process is stopped at 180 is min with fluctuating experiment. After 60particle min, average particle size is still The process stopped at 180 min average particle diameters of approximately µm in the product. with fluctuating average particle diameters 66 of approximately 66 μm in the product. SEM SEMimages imagesshown shownininFigure FigureA2 A2ofofAppendix AppendixAAthat thatclearly clearlyillustrate illustratethat thatduring duringthe thecoating coating process in the modified fluidized bed, agglomeration hardly happened. At the lowest temperature process in the modified fluidized bed, agglomeration hardly happened. At the lowest temperature of ◦ C, coating quality was though worse than at higher temperature. When particles are coated of5050°C, coating quality was though worse than at higher temperature. When particles are coated at a athigher a higher temperature, images showa abetter better coating coating structure. structure. To To differentiate temperature, thethe images show differentiate the the coating coating performance at the above five temperatures, coating coverage calculation based on SEM images performance at the above five temperatures, coating coverage calculation based on SEM imageshas has been beencarried carriedout outand andthe theresults resultsare areshown shownininFigure Figure10. 10.ItItisisseen seenthat thatthe thecoating coatingcoverage coveragetends tendstoto ◦ (with coating coverage increase increasewith withthe thefluidized fluidizedbed bedtemperature, temperature,though thoughwith withaamaximum maximumatat70 70 C °C (with coating coverage ◦ ◦ ◦ ◦ ◦ C). ofof86% for 50 C, 88% for 60 C, 99% for 70 C, 94% for 80 C and 92% for 90 86% for 50 °C, 88% for 60 °C, 99% for 70 °C, 94% for 80 °C and 92% for 90 °C). higher temperature, the images show a better coating structure. To differentiate the coating performance at the above five temperatures, coating coverage calculation based on SEM images has been carried out and the results are shown in Figure 10. It is seen that the coating coverage tends to increase with the fluidized bed temperature, though with a maximum at 70 °C (with coating coverage Processes 2020, 1525 11 of 18 of 86% for8,50 °C, 88% for 60 °C, 99% for 70 °C, 94% for 80 °C and 92% for 90 °C). Figure 10. Coverage comparison based on temperature. Figure 10. Coverage comparison based on temperature. The influence of temperature on the yield of the aerosol coating process is shown in Table 4. The results indicate that when the temperature is too high, the yield may be much lower than at moderate temperature; 13.59% for the temperature of 70 ◦ C as compared with 8.83% for the temperature of 90 ◦ C. It is known that temperature may be a critical factor that can affect process yield even when using a two-fluid nozzle. In fact, when the aerosol generator is used, because the aerosol has smaller droplet size, the temperature causes a more noticeable effect on the yield. Higher temperature results in the loss of coating material due to rapidly evaporation before deposition and coating can take place on the surface of the cohesive fluidized particles, which allows the aerosol droplets to be directly dried into fine powder and then quickly blown out of the fluidization chamber with the airflow. Table 4. Yield of aerosol coating process at different temperatures 1 . 1 Temperature (◦ C) Mass of Cores (g) Coating Mass (g) Yield (%) 2 50 60 70 80 90 2.0113 ± 0.07 2.2572 ± 0.37 2.5010 ± 0.55 1.8998 ± 0.16 1.3141 ± 0.05 0.0772 ± 0.015 0.0779 ± 0.002 0.0849 ± 0.017 0.0435 ± 0.012 0.0293 ± 0.041 14.48 14.22 13.59 9.08 8.83 Total mass of glass beads: 1000 g; Mean aerosol spray rate: 282 g/h. 2 Average value from three samples. In summary, it can be seen that the evaluation indicators of coated products are susceptible to temperature changes. At low temperature, some aerosol droplets cannot be dried in time. So even though the droplets are very small when the aerosol generator is applied, liquid bridges between the particles may still lead to agglomeration. On the contrary, agglomeration is negligible at high temperature, but the loss of coating solution by overspray may cause lower solution yield. Therefore, in the case of various process parameters being adjusted, it is particularly important to first determine an appropriate range of mild temperatures to obtain high-quality products. 3.3. Effect of Coating Solution Viscosity To investigate the influence of coating solution viscosity on coating speed and process time with the same core particles as before, an additive was mixed into the coating solution. As mentioned above, the additive used was HPMC, which is suitable for the agglomeration of excipients and also plays an important role as a coating material in the pharmaceutical industry. The concentration of HPMC in the solution should be carefully selected. Due to its extremely high viscosity, HPMC is usually used as an agglomerating binder with more than 4 wt% in the solution [31]. In this study of coatings, less HPMC was added to the NaB solution, namely zero to 1.5%. The influence of coating solution viscosity on coating performance was experimentally investigated by adding different mass fraction of HPMC (with a relationship between HPMC mass fraction and Processes 2020, 8, 1525 12 of 18 solution viscosity, as can be seen in Table 1), and the corresponding results are discussed in the following. Figure 11a indicates that more HPMC in the solvent makes the size distribution of coated glass particles slightly different and somewhat narrows. The average size of the final coated particles decreased monotonically from 66 µm to 64.8 µm when the HPMC mass fraction in the NaB solution was increased from 0% to 1.5%. Figure 11b shows the evolution of average particle size with different viscosity of the coating solution. In these experiments, a much more quicker particle growth is achieved when the coating solution has a higher viscosity. A fluctuating steady state is reached in only 20 min for 1.5 wt% HPMC, compared to 40 min for 0% HPMC. On the other hand, higher solution viscosity results smaller final particle from Processesin 2020, 8, x FOR PEER REVIEWsize. The process stops after 180 min with particle size growth 12 of 18 1.8 µm to 3 µm. (a) (b) Figure Figure11. 11.(a) (a)Particle Particlesize sizedistributions distributionsand and(b) (b)evolution evolutionofofaverage averageparticle particlediameter diameterd1,3 d1,3for fordifferent different HPMC concentrations. HPMC concentrations. With traditional two-fluid nozzles, higher viscosity of the coating solution can increase the rate With traditional two-fluid nozzles, higher viscosity of the coating solution can increase the rate of undesired agglomeration in the course of coating processes. However, the novel aerosol atomizer of undesired agglomeration in the course of coating processes. However, the novel aerosol atomizer can generate small aerosol droplets, which makes it difficult to produce solid bridges at a suitable can generate small aerosol droplets, which makes it difficult to produce solid bridges at a suitable process temperature even when the solution viscosity is high. The present results confirm that the process temperature even when the solution viscosity is high. The present results confirm that the agglomeration rate remains small even at high viscosity of the coating solution (agglomeration rate of agglomeration rate remains small even at high viscosity of the coating solution (agglomeration rate 2.99% for 0.5 wt% HPMC, 2.32% for 1 wt% HPMC and 1.24% for 1.5 wt% HPMC based on the PSD). of 2.99% for 0.5 wt% HPMC, 2.32% for 1 wt% HPMC and 1.24% for 1.5 wt% HPMC based on the SEM images of coated particles are shown in Figure A3 of the Appendix A. It appears that the PSD). increase in viscosity of the solution decreases significantly the overall average coverage of particles, SEM images of coated particles are shown in Figure A3 of the Appendix A. It appears that the with increasing differences among individual particles of the population. In general, island growth is increase in viscosity of the solution decreases significantly the overall average coverage of particles, observed, and patchy particles are obtained. In the case of 1.5 wt% HPMC, the overall coverage is only with increasing differences among individual particles of the population. In general, island growth 21%. In addition to the decrease in coverage, the coating layers also gradually become compact and is observed, and patchy particles are obtained. In the case of 1.5 wt% HPMC, the overall coverage is brittle as the viscosity increases. Figure 12 shows the coverage results evaluated from the SEM images only 21%. In addition to the decrease in coverage, the coating layers also gradually become compact at different HPMC concentrations. and brittle as the viscosity increases. Figure 12 shows the coverage results evaluated from the SEM As the solution viscosity increases, the coverage shows a downward trend. One explanation is that images at different HPMC concentrations. after the high-viscosity liquid dries on the surface of the particles, the dense and compact coating layers break off due to collisions between particles. Finally, the growth and breakage of the coating layer reach an equilibrium at a certain surface coverage, which may be quite low. To verify this conjecture, an additional experiment has been conducted with 5 wt% of HPMC. During this experiment, the yield of the solution has been evaluated every 10 min. Results are plotted in Figure 13. It can be seen that the yield decreases to zero 60 min, that is, the rate of coating layer formation becomes equal to the rate of dried coating layer breakage. This trend is stronger when introducing HPMC to increase the viscosity of the solution; it is, therefore, not easy to obtain a coating product with 100% coverage in such cases. Figure 12. Coverage comparison based on HPMC concentration As the solution viscosity increases, the coverage shows a downward trend. One explanation is increase in viscosity of the solution decreases significantly the overall average coverage of particles, with increasing differences among individual particles of the population. In general, island growth is observed, and patchy particles are obtained. In the case of 1.5 wt% HPMC, the overall coverage is only 21%. In addition to the decrease in coverage, the coating layers also gradually become compact and brittle as8,the SEM Processes 2020, 1525viscosity increases. Figure 12 shows the coverage results evaluated from the 13 of 18 images at different HPMC concentrations. Processes 2020, 8, x FOR PEER REVIEW 13 of 18 HPMC to increase the viscosity of the solution; it is, therefore, not easy to obtain a coating product with 100% coverageFigure in such 12.cases. Coverage comparison based on HPMC concentration Figure 12. Coverage comparison based on HPMC concentration As the solution viscosity increases, the coverage shows a downward trend. One explanation is that after the high-viscosity liquid dries on the surface of the particles, the dense and compact coating layers break off due to collisions between particles. Finally, the growth and breakage of the coating layer reach an equilibrium at a certain surface coverage, which may be quite low. To verify this conjecture, an additional experiment has been conducted with 5 wt% of HPMC. During this experiment, the yield of the solution has been evaluated every 10 min. Results are plotted in Figure 13. It can be seen that the yield decreases to zero 60 min, that is, the rate of coating layer formation becomes equal to the rate of dried coating layer breakage. This trend is stronger when introducing Figure 13. Coating solution yield versus process time. Figure 13. Coating solution yield versus process time. In summary, agglomeration can be avoided by the innovative coating system even when the In are summary, can beisavoided by thesolution innovative systemthe even when the particles fine andagglomeration the coating solution sticky. Sticky may coating even increase coating rate particles are fine and the coating solution is sticky. Sticky solution may even increase the coating rate by promoting the deposition of aerosol droplets on the surface of core particles. At the same time, by promoting deposition of aerosol droplets onthan the surface of core particles. At coating the same time, however, coatingthe layers are denser and more compact at low viscosity. Such dense layers however, coating layers are denser and more compact than at low viscosity. Such dense coating layers seem to be brittle, and are more likely to break off after collisions between particles and with walls in seem to be fluidized brittle, and arewhich moreresults likely to off after collisions between and with walls in the Wurster bed, in break low product coverage and unevenparticles overall coating of patchy the Wurster fluidized bed, which results in low product coverage and uneven overall coating of product particles. patchy product particles. 3.4. Effect of Aerosol Inlet Position 3.4. Effect of Aerosol Inlet Position In this study, the aerosol was introduced into the fluidized bed from the bottom by means of a thistube study, the aerosol was introduced into the fluidized bedfluidized from thebed bottom means of a copperIninlet through the distributor plate. When using a Wurster withby a traditional copper inlet tubethe through the between distributor When using Wurster fluidized bed with traditional two-fluid nozzle, distance theplate. nozzle outlet anda the bottom distributor platea can affect two-fluid nozzle, the distance between the nozzle outlet and the bottom distributor plate can affect the coating quality. Similarly, in the novel coating system, aerosol and fluidized particles move the coating in quality. Similarly, the novel coating may system, aerosol and fluidized move concurrently riser, so the inletinposition of aerosol affect its contact time withparticles the particles. concurrently in riser, so the inlet position of aerosol may affect its contact time with the particles. Therefore, the influence of aerosol inlet position on the experimental results was explored using three Therefore, the influence of lengths aerosol of inlet position on the experimental results was explored using three different copper tubes with 6 cm, 7 cm and 8 cm, respectively. different tubes lengths size of 6 distributions cm, 7 cm andin8 Figure cm, respectively. As cancopper be seen fromwith the particle 14, the inlet tube of 8 cm has a minor As can be seen from the particle size distributions in Figure 14, the inlet tube change of 8 cm in hasaverage a minor effect in altering the particle size, showing scarce agglomeration and only a slight effect insize. altering particle size, showing scarce agglomeration and only a of slight change in average particle Withthe a decrease of the inlet tube length to 7 cm, the amount larger agglomerates particle size. With a decrease of the inlet tube 7 cm,that thedecreasing amount ofthe larger increased slightly. Results based on calculated fromlength graphstoverify tubeagglomerates length from slightly. basedincreased on calculated from verify that decreasing theproportion tube length 8 increased cm to 6 cm results Results in a slightly particle sizegraphs from 66 µm to 67.5 µm, and the 8 cm to 6 increases cm results in a slightly increased sizereason from for 66 μm to 67.5 μm, andthe the offrom agglomeration significantly from 5.4% to particle 13.2%. The this behavior is that proportion of agglomeration increases significantly from 5.4% to 13.2%. The reason for this behavior is that the shorter tube increases the residence time of droplets in the Wurster riser, but it also increases the humidity in regions with many particles, so agglomeration is more likely to occur. As depicted in Figure 14b, the experiment reaches steady state after around 60 min with average particle size of approximately 67.5 μm in the product when the copper tube is 6 cm. Compared to the 7 cm and 8 cm experiments, the growth rate is faster at beginning, and the average particle size is Processes 2020, 8, x FOR PEER REVIEW 14 of 18 Processes 2020, 8, 1525 14 of 18 shorter increases theREVIEW residence time of droplets in the Wurster riser, but it also increases14the Processestube 2020, 8, x FOR PEER of 18 humidity in regions with many particles, so agglomeration is more likely to occur. (b) (a) Figure 14. (a) Particle size distributions and (b) evolution of average particle diameter d1,3 for different aerosol inlet tube lengths. SEM images of the final product (Appendix A, Figure A4) show good coating of the glass particles for all investigated aerosol inlet tube lengths. The related coating coverage shown in Figure 15 ranges from 99.8% to 99%. All three conditions have a small standard deviation, which means the (b) (a) particle coverage is relatively uniform. Table 5 indicates that a longer aerosol tube results in lower yield; 13.59% for a copper tube of 8 Figure 14. size and (b) evolution ofofaverage particle diameter d1,3 different Figure 14.(a) (a)Particle Particle sizedistributions distributions average particle diameter d1,3for for cm as compared with 23.13% for 6 cm. Itand can(b)beevolution seen that the closer the aerosol outlet is different to the bottom aerosol inlet tube lengths. aerosol inlet tube lengths. distributor plate, the less coating solution is lost; therefore, the use rate is the highest in this case. As depicted in Figure 14b, the experiment reaches steady state after around 60 min with average SEM images of the final product (Appendix A, Figure A4) show good coating of the glass 1. Table 5. Yield67.5 of aerosol process withthe different aerosol tubesCompared particle size of approximately µm in coating the product when copper tube inlet is 6 cm. to the particles for all investigated aerosol inlet tube lengths. The related coating coverage shown in Figure 7 cm and 8 cm experiments, the growth rate is at beginning, and (g) the average particle size is 2 Inlet Tube Length Mass of faster Cores (g) Coating Mass Yield (%) 15 ranges from 99.8% to 99%. All (cm) three conditions have a small standard deviation, which means the slightly larger at steady state. 6 uniform. 1.6187 ± 0.51 0.0697 ± 0.015 23.13 particle coverage is relatively SEM images of the final7 product (Appendix A,± Figure show good of the glass particles 5.1553 0.21 A4) 0.0848 ± 0.009coating 17.47 Table 5 indicates that a longer aerosol tube results in lower yield; 13.59% for a copper tube of 8 for all investigated aerosol inlet tube lengths. The ±related coverage in Figure 15 ranges 8 2.5010 0.55 coating 0.0849 ± 0.017 shown13.59 cm as compared with 23.13% for 6 cm. It can be seen that the closer the aerosol outlet is to the bottom 2 Average from199.8% to 99%. All beads: three conditions a small deviation, which means thesamples. particle Total mass of glass 1000 g; Meanhave aerosol spray standard rate: 282 g/h. value from three distributor plate, the less coating solution is lost; therefore, the use rate is the highest in this case. coverage is relatively uniform. Table 5. Yield of aerosol coating process with different aerosol inlet tubes 1. Inlet Tube Length (cm) 6 7 8 1 Mass of Cores (g) 1.6187 ± 0.51 5.1553 ± 0.21 2.5010 ± 0.55 Coating Mass (g) 0.0697 ± 0.015 0.0848 ± 0.009 0.0849 ± 0.017 Yield (%) 2 23.13 17.47 13.59 Total mass of glass beads: 1000 g; Mean aerosol spray rate: 282 g/h. 2 Average value from three samples. Figure 15. Coverage comparison based on length of aerosol inlet tube. Figure 15. Coverage comparison based on length of aerosol inlet tube. Table 5 indicates that a longer aerosol tube results in lower yield; 13.59% for a copper tube of 8 cm The above to seen the conventional bottom spray coating when the as compared withresults 23.13%show for 6that, cm. similar It can be that the closer the aerosol outlet issystem, to the bottom coating process is conducted with the novel aerosol generator, the position at which the generated distributor plate, the less coating solution is lost; therefore, the use rate is the highest in this case. aerosol is introduced into the fluidized bed influences the results. Due to higher particle 1. Yield ofofaerosol coating fluidized process with different aerosol inlet tubes concentration Table at the5.bottom the Wurster bed [32], when aerosol enters the fluidized bed at a lower position, the frequency of contacts with particles and the probability of particles capturing Inlet Tube Length (cm) Mass of Cores (g) Coating Mass (g) Yield (%) 2 aerosol droplets rise. Eventually, the process yield will also increase. However, it is worth noting that Figure 15. Coverage comparison aerosol inlet tube. 6 1.6187 ± 0.51based on length 0.0697 of ± 0.015 23.13 7 5.1553 ± 0.21 0.0848 ± 0.009 17.47 8 show that, similar 2.5010 ± 0.55 0.0849 ± 0.017 13.59system, when the The above results to the conventional bottom spray coating mass is of glass beads: 1000 g; Mean aerosolaerosol spray rate: 282 g/h. 2 Average value from samples. coating1 Total process conducted with the novel generator, the position at three which the generated aerosol is introduced into the fluidized bed influences the results. Due to higher particle concentration at the bottom of the Wurster fluidized bed [32], when aerosol enters the fluidized bed at a lower position, the frequency of contacts with particles and the probability of particles capturing aerosol droplets rise. Eventually, the process yield will also increase. However, it is worth noting that Processes 2020, 8, 1525 15 of 18 The above results show that, similar to the conventional bottom spray coating system, when the coating process is conducted with the novel aerosol generator, the position at which the generated aerosol is introduced into the fluidized bed influences the results. Due to higher particle concentration at the bottom of the Wurster fluidized bed [32], when aerosol enters the fluidized bed at a lower position, the frequency of contacts with particles and the probability of particles capturing aerosol droplets rise. Eventually, the process yield will also increase. However, it is worth noting that a too-low inlet position will also cause the humidity of the dense zone of bed to increase, with a potential increase in the agglomeration. 4. Conclusions A novel coating process in a Wurster fluidized bed combined with an aerosol atomizer was presented in this study. It was shown by a series of experiments that different operation parameters influence the particle growth behavior. The experiments have shown the feasibility of using aerosol to produce a coating layer on fluidized particles. It should be noted that submicron aerosol droplets have been used rather than conventional two-fluid nozzle spray with droplet diameters of tens of microns. Due to the extremely small diameter of aerosol droplets, even very small core particles could be individually coated and without severe agglomeration. In addition, in the industrial field, due to the extremely small aerosol droplets, the resulting coating layer is ultrathin, and the yield of the new coating process was determined to be 15–20% under different conditions, which is considered to be improvable. The yield in this study may be enhanced by the recycling of entrained solids or better equipment design, especially when scaling up the experiment plant. Coating performance was characterized from both microscale (SEM) and macroscale (PSD, yield) points of view. Different operation parameters such as particle size, air temperature, coating solution viscosity and aerosol inlet position in the Wurster riser have an influence on product quality. The main conclusions are as follows: • • • • Both conventionally large particles and smaller particles that are difficult to coat in conventional equipment can be coated without agglomeration in a Wurster fluidized bed with the help of an aerosol generator. The highest growth rate is observed at the beginning of the novel process. Over time, the growth rate decreases gradually until a quasi-steady state has been reached. The temperature of the fluidizing gas has a significant impact on the quality of the coated product. A decrease in temperature may increase the agglomeration rate, while excessively high temperature will lower the use of coating solution and the coating rate. Moderate temperature can maintain reasonable yield and low agglomeration ratio. Particle growth rate was observed to increase with the increase of coating solution viscosity. Higher viscosity aerosol is easier to adhere to the particles. Unfortunately, this is accompanied by the creation of compact and dense coating layers, which are more likely to break down due to collisions with other particles and walls in the Wurster fluidized bed. Equilibrium between coating layer growth and breakage may result in lower coverage when the solution viscosity is high, and in patchy product particles. The position where coating aerosol is introduced into the fluidized bed has an impact on the use rate of the coating solution. When aerosol is fed close to the bottom distributor plate, it interacts with the particles in a dense zone of the equipment, hence the use rate is higher. Meanwhile, the agglomeration rate also slightly increases. In summary, the use of very small aerosol droplets enables the coating of fine particles without severe agglomeration in a Wurster fluidized bed. This novel approach of potential interest for pharmaceutical, food, catalyst and other industries, where it may contribute to the development of new products based on coated fine particles. Processes 2020, 8, 1525 16 of 18 Author Contributions: R.Z. carried out the experiment. R.Z. wrote the manuscript with support from T.H. and E.T., T.H. and E.T. helped supervise the project. E.T. and R.Z. conceived the original idea. All authors have read and agreed to the published version of the manuscript. Funding: The first author gratefully acknowledges the funding of this work by the China Scholarship Council (NO. 201708080202). Acknowledgments: We also thank Jing Yang for his assistance in the conduction of the experiments and measurements of samples in the frame of his M.Sc. thesis. Processes2020, 2020, 8,xxFOR FORPEER PEER REVIEWdeclare no conflict of interest. Processes Conflicts of 8, Interest: The REVIEW authors 16ofof18 18 16 AppendixAA Appendix (b) (b) (a) (a) FigureA1. SEMimages imagesfrom fromexperiments experimentsusing using(a) (a)large largeparticles; Figure A1.SEM particles;(b) (b)small smallparticles. particles. (b) (b) (a) (a) (c) (c) (e) (e) (d) (d) ◦ C; (b) 60 ◦ C; FigureA2. A2.SEM SEMimages images from fluidization temperatures (a)(a) 5050 imagesfrom fromexperiments experimentsunder underdifferent different fluidization temperatures (a) 50°C; °C;(b) (b)60 60 Figure under different fluidization temperatures ◦ ◦ ◦ (c) 70 C;°C; (d) 80 C;°C; (e)(e) 90 C.°C. °C;(c) (c)70 70 °C;(d) (d)80 80 °C; (e)90 90 °C. °C; (a) (a) (b) (b) (e) (d) Figure A2. SEM images from experiments under different fluidization temperatures (a) 50 °C; (b) 60 Processes 2020, 8, 1525 °C; (c) 70 °C; (d) 80 °C; (e) 90 °C. Processes 2020, 8, x FOR PEER REVIEW Processes 2020, 8, x FOR PEER REVIEW (a) (b) (c) (c) (d) (d) 17 of 18 17 of 18 17 of 18 Figure A3. SEM SEM images from experimentsusing usingvarious variousHPMC HPMCmass mass fractions coating solution images from experiments using various fractions ofof the coating solution Figure A3. SEM images from experiments HPMC mass fractions ofthe the coating solution (a) 0 wt%; (b) 0.5 wt%; (c) 1.0 wt%; (d) 1.5 wt%. (b) wt%; 0.5 wt%; wt%; (d)1.5 1.5wt%. wt%. (a) 0 wt%; (b) 0.5 (c) (c) 1.01.0 wt%; (d) (a) (a) (b) (b) (c)(c) Figure SEM images from experiments lengths ofofthe inlet tube (a)(a) 6 cm; (b)(b) Figure A4. SEM images fromexperiments experimentsusing using various lengths the aerosol inlet tube 6 cm; Figure A4. A4. SEM images from usingvarious various lengths ofaerosol the aerosol inlet tube (a) 6 cm; 7 cm; (c) 88 cm. 7 cm; (c) cm. (b) 7 cm; (c) 8 cm. Author Contributions: support from T.H. and Author Contributions: R.Z. R.Z. carried carriedout outthe theexperiment. experiment.R.Z. R.Z.wrote wrotethe themanuscript manuscriptwith with support from T.H. and References 1. 2. 3. 4. 5. 6. 7. 8. 9. E.T., T.H. T.H. and idea. All authors have read E.T., and E.T. E.T. helped helped supervise supervisethe theproject. project.E.T. E.T.and andR.Z. R.Z.conceived conceivedthe theoriginal original idea. All authors have read and agreed agreed to published version the Kleinbach, E.;the Riede, T. Coating ofof solids. Chem. Eng. Process. Process. Intensif. 1995, 34, 329–337. [CrossRef] and to the published version of themanuscript. manuscript. Felton, L.; Timmins, G. A nondestructive technique to determinework the rate of China oxygen permeation into solid Funding: The Scholarship Council Funding: The first first author author gratefully gratefullyacknowledges acknowledgesthe thefunding fundingofofthis this workbybythe the China Scholarship Council dosage forms. Pharm. Dev. Technol. 2006, 11, 141–147. [CrossRef] [PubMed] (NO. 201708080202). 201708080202). (NO. Porter, S.C.; Felton, L.A. Techniques to assess film coatings and evaluate film-coated products. Drug Dev. Acknowledgments: We also thank Jing Yang for his assistance in the conduction of the experiments and Acknowledgments: also thank Jing Yang for his assistance in the conduction of the experiments and Ind. Pharm. 2010, 36,We 128–142. [CrossRef] [PubMed] measurements of samples in the frame of his M.Sc. thesis. measurements samples in frame of his thesis. U.; Smirnova, I.; Heinrich, S. Novel technique for Goslinska, M.;of Selmer, I.; the Kleemann, C.;M.Sc. Kulozik, Conflicts of Interest: The authors declare no conflict of interest. measurement of coating layer thickness fine and porous particles using focused ion beam. Particuology Conflicts of Interest: The authors declare noofconflict of interest. 2019, 42, 190–198. [CrossRef] References References K.; Huyghebaert, A. Fluidized bed coating in food technology. Trends Food Sci. Technol. 1999, 10, Dewettinck, 1. Kleinbach, E.; Riede, T. Coating of solids. Chem. Eng. Process. Process. Intensif. 1995, 34, 329–337. 163–168. [CrossRef] 1. Kleinbach, E.; Riede, T. Coating of solids. Chem. Eng. Process. Process. Intensif. 1995, 34, 329–337. 2. Felton, L.; Timmins, G. A nondestructive technique to determine the rate of oxygen permeation into solid Watano, S.; Imada, Y.; Hamada, K.; Wakamatsu, Y.; Tanabe, Y.; Dave, R.N.; R. Microgranulation 2. Felton, L.; Timmins, G. A nondestructive technique to determine the rate of Pfeffer, oxygen permeation into solid of dosage forms. Pharm. Dev. Technol. 2006, 11, 141–147. fine powders by a Pharm. novel rotating fluidized bed141–147. granulator. Powder Technol. 2003, 131, 250–255. [CrossRef] dosage forms. Dev. Technol. 2006, 11, 3. Porter, S.C.; Felton, L.A. Techniques to assess film coatings and evaluate film-coated products. Drug Dev. Jones, D.M. S.C.; Factors to consider in fluid-bed processing. Pharm. 9, 50–55. 3. Porter, Felton, L.A. Techniques to assess film coatings andTechnol. evaluate1985, film-coated products. Drug Dev. Ind. Pharm. 2010, 36, 128–142. Rieck, C.;Pharm. Bück,2010, A.; 36, Tsotsas, E. Monte Carlo modeling of fluidized bed coating and layering processes. Ind. 128–142. 4. Goslinska, M.; Selmer, I.; Kleemann, C.; Kulozik, U.; Smirnova, I.; Heinrich, S. Novel technique for 4. Goslinska, M.;2670–2680. Selmer, I.;[CrossRef] Kleemann, C.; Kulozik, U.; Smirnova, I.; Heinrich, S. Novel technique for AIChE J. 2016, 62, measurement of coating layer thickness of fine and porous particles using focused ion beam. Particuology measurement of coating layer thickness of fine and particlesofusing focused ion beam. Particuology Pietsch, S.;42, Peter, A.; Wahl, P.; Khinast, J.; Heinrich, S. porous Measurement granule layer thickness in a spouted 2019, 190–198. 2019, 42, 190–198. bed process via optical coherence tomography. Technol. 2019, 356,Food 139–147. [CrossRef] 5. coating Dewettinck, K.; Huyghebaert, A. Fluidized bed coating Powder in food technology. Trends Sci. Technol. 1999, 5. 6. 6. 7. 7. 8. 8. Dewettinck, 10, 163–168. K.; Huyghebaert, A. Fluidized bed coating in food technology. Trends Food Sci. Technol. 1999, 10, 163–168. Watano, S.; Imada, Y.; Hamada, K.; Wakamatsu, Y.; Tanabe, Y.; Dave, R.N.; Pfeffer, R. Microgranulation of Y.; Dave, R.N.; Pfeffer, R. Microgranulation of Watano, S.; Imada, Y.; Hamada, Wakamatsu, Y.; Tanabe, fine powders by a novel rotating K.; fluidized bed granulator. Powder Technol. 2003, 131, 250–255. fine powders by a novel rotating fluidized bed granulator. Powder Technol. 2003, 131, Jones, D.M. Factors to consider in fluid-bed processing. Pharm. Technol. 1985, 9, 50–55. 250–255. Jones, Factors consider fluid-bed Pharm. Technol. 9, 50–55. Rieck, D.M. C.; Bück, A.; to Tsotsas, E. in Monte Carloprocessing. modeling of fluidized bed1985, coating and layering processes. Rieck, C.; Bück, A.; Tsotsas, E. Monte Carlo modeling of fluidized bed coating and layering processes. AIChE J. 2016, 62, 2670–2680. Processes 2020, 8, 1525 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 18 of 18 Geldart, D. Types of gas fluidization. Powder Technol. 1973, 7, 285–292. [CrossRef] Rieck, C.; Hoffmann, T.; Bück, A.; Peglow, M.; Tsotsas, E. Influence of drying conditions on layer porosity in fluidized bed spray granulation. Powder Technol. 2015, 272, 120–131. [CrossRef] Ennis, B.J.; Tardos, G.; Pfeffer, R. A microlevel-based characterization of granulation phenomena. Powder Technol. 1991, 65, 257–272. [CrossRef] To, D.; Dave, R.N. Fluid bed film coating of fine ibuprofen particles. Powder Technol. 2016, 290, 102–113. [CrossRef] Mawatari, Y.; Koide, T.; Tatemoto, Y.; Takeshita, T.; Noda, K. Comparison of three vibrational modes (twist, vertical and horizontal) for fluidization of fine particles. Adv. Powder Technol. 2001, 12, 157–168. [CrossRef] Zhu, Q.; Li, H. Study on magnetic fluidization of group C powders. Powder Technol. 1996, 86, 179–185. Ichikawa, H.; Tokumitsu, H.; Jono, K.; HUKUDA, T.; Osako, Y.; Fukumori, Y. Coating of pharmaceutical powders by fluidized bed process, Part 6: Microencapsulation using blend and composite latex of copoly (ethyl acrylate-methyl methacrylate-2-hydroxyethyl methacrylate). Chem. Pharm. Bull. 1994, 42, 1308–1314. [CrossRef] Watano, S.; Nakamura, H.; Hamada, K.; Wakamatsu, Y.; Tanabe, Y.; Dave, R.N.; Pfeffer, R. Fine particle coating by a novel rotating fluidized bed coater. Powder Technol. 2004, 141, 172–176. [CrossRef] Chen, Y.; Jallo, L.; Quintanilla, M.A.; Dave, R.N. Characterization of particle and bulk level cohesion reduction of surface modified fine aluminum powders. Colloids Surf. A Physicochem. Eng. Asp. 2010, 361, 66–80. [CrossRef] Chen, Y.; Yang, J.; Mujumdar, A.; Dave, R.N. Fluidized bed film coating of cohesive Geldart group C powders. Powder Technol. 2009, 189, 466–480. [CrossRef] Hampel, N.; Roydeva, E.; Bück, A.; Tsotsas, E. Coating of finely dispersed particles by two-fluid nozzle. Particuology 2018, 38, 80–93. [CrossRef] Mezhericher, M.; Ladizhensky, I.; Etlin, I. Atomization of liquids by disintegrating thin liquid films using gas jets. Int. J. Multiph. Flow 2017, 88, 99–115. [CrossRef] Mezhericher, M.; Nunes, J.K.; Guzowski, J.J.; Stone, H.A. Aerosol-assisted synthesis of submicron particles at room temperature using ultra-fine liquid atomization. Chem. Eng. J. 2018, 346, 606–620. [CrossRef] Mezhericher, M.; Rieck, C.; Razorenov, N.; Tsotsas, E. Ultrathin coating of particles in fluidized bed using submicron droplet aerosol. Particuology 2020. [CrossRef] Lefebvre, A.H.; McDonell, V.G. Atomization and Sprays; CRC Press: Boca Raton, FL, USA, 2017. Mezhericher, M.; Razorenov, N.; Mazor, G.; Ju, Y.; Stone, H.A. Submicron aerosols of liquid fuels: Method of production, experimental characterization and a semi-empirical model. Appl. Energy 2019, 235, 1651–1663. [CrossRef] Sondej, F.; Bück, A.; Tsotsas, E. Comparative analysis of the coating thickness on single particles using X-ray micro-computed tomography and confocal laser-scanning microscopy. Powder Technol. 2016, 287, 330–340. [CrossRef] Chen, Y.; Yang, J.; Dave, R.N.; Pfeffer, R. Fluidization of coated group C powders. AIChE J. 2008, 54, 104–121. [CrossRef] Dewettinck, K.; Huyghebaert, A. Top-spray fluidized bed coating: Effect of Process variables on coating efficiency. Lebensm. Wiss. Und Technol. 1998, 31, 568–575. [CrossRef] Liu, L.X.; Litster, J.D. Spouted bed seed coating: The effect of process variables on maximum coating rate and elutriation. Powder Technol. 1993, 74, 215–230. [CrossRef] Takei, N.; Unosawa, K.; Matsumoto, S. Effect of the spray-drying process on the properties of coated films in fluidized bed granular coaters. Adv. Powder Technol. 2002, 13, 333–342. [CrossRef] Du, J.; Bück, A.; Tsotsas, E. Influence of process variables on spray agglomeration process in a continuously operated horizontal fluidized bed. Powder Technol. 2020, 363, 195–206. [CrossRef] Jiang, Z.; Bück, A.; Tsotsas, E. CFD–DEM study of residence time, droplet deposition, and collision velocity for a binary particle mixture in a Wurster fluidized bed coater. Dry. Technol. 2018, 36, 638–650. [CrossRef] Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).