Single droplet drying and particle formation during spray drying are difficult to investigate, because of the complex two-phase flow in spray dryers and other difficulties . A technique to investigate single droplet drying is acoustic levitation, where single droplets can be placed close to the pressure nodes of a standing acoustic wave. The first drying stage is known as the constant-rate period, since the surface of the droplet decreases at a constant rate due to evaporation of the liquid from the surface. This drying stage therefore follows the so-called d2 law . Suspension droplets exhibit this behavior as long as the surface is wetted . At high ratios of droplet-shrinkage-rate-to-diffusivity in the liquid phase, a non-uniform radial distribution of the solid may arise, leading to crust formation . The local porosity of the solid, surface roughness and solids distribution inside the microsphere may result in a highly complex evolution of the second drying stage .
In the present work we studied the evaporation kinetics of suspension droplets, with solids dissolved in various carrier liquids, and the resulting microsphere morphology of pharmaceutical excipients. Furthermore, the mechanical properties of individual microspheres were evaluated by hardness measurements and related to the solid’s solubility in the liquid. Knowing the influencing factors during spray drying of suspensions will make it possible to choose process parameters during drying which avoid change of particle properties.
Materials and methods
Lactose GranuLac® 230 (Molkerei Meggle, Wasserburg, Germany) was used as solid substance, and water of analytical grade (Ultrapure water system TKA, Niederelbert, Germany) and isopropanol in high purity ≥ 99.8 % (Carl Roth, Karlsruhe, Germany) as liquid mixture to form the suspension. The ratio of water-isopropanol was varied according to the solubility curve in Figure 1, resulting in different amounts of dissolved lactose. Lactose exhibits a high solubility in water and a low solubility in isopropanol. The following definitions were chosen, i.e., the mass of the liquid mixture mliq consisting of H2O and solvent I (isopropanol) in Eq. 1, the water mass fractionwH2O in Eq. 2 and the solubility of lactose XL in Eq. 3 with md as the mass of the dissolved excipient (alpha-lactose monohydrate).
Figure 1. Solubility XL,s of alpha-lactose monohydrate in isopropanol-water mixtures at 25°C.
An acoustic levitator was used to assess the drying kinetics of lactose suspensions with different lactose mass fractions and solvent compositions. This technique enables drying without contact between the levitated object and a solid body, and therefore heat and mass transfer are influenced only from the acoustic streaming which can be controlled by additional air streams around the object. The levitation setup used can be seen in Figure 2, consisting of a transducer and reflector to generate the acoustic sound wave for droplet positioning, and a macro lens with a light source to record the drying progression.
Figure 2. Ultrasonic levitator and imaging equipment for measuring individual droplet drying kinetics.
The composition of a droplet with the mass m consists of the dissolved mass md and the suspended mass ms of the solid material in a mixture of the liquid masses mH2O and mI of water and isopropanol. Depending on the water mass fraction wH2O (Eq. 2), different mass of solids msolid (Eq. 4) were present in one suspension droplet. The solids mass msolid consists of dissolved md and suspended solids ms.
In terms of definition, the mass m was used as reference to define the dissolved solid mass fraction wd (Eq. 5), the suspended solid mass fraction ws (Eq. 6) and the total solid mass fraction wsolid (Eq. 7).
Table 1 shows the investigated suspension compositions with variations in the suspended lactose mass fraction ws from 10 % to 50 %. Afterwards, the measurement series was continued with the same suspended lactose mass fraction ws of 10 % and varying dissolved lactose mass fractions wd , ranging from 0.0072 % to 17 %. These variations are due to different water mass fractions wH2O, ranging from 0 % to 100%.
Table 1. Compositions of suspensions for drying experiments.
The hardness of individual microspheres was measured with the side-crushing-strength (SCS) test in a slightly modified version. The procedure for this test method is described in ASTM Standard D4179-01 for crushing of single catalyst pellets . The agglomeration strength of individual dried microspheres was measured using a rheometer MCR 300 from Anton Paar (Graz, Austria) in the plate-plate configuration. For this measurement, individual microspheres dried in the acoustic levitator were placed in the center of the bottom plate, and the upper plate was moved downward at the constant velocity of 5 x 10‑6 ms-1 (see Figure 3). The normal forces were recorded while the upper plate was moving.
Figure 3. Microsphere on the rheometer plate (a) before and (b) after hardness test.
Results and discussion
Individual droplet drying experiments were performed for the fixed suspended lactose mass fraction ws of 10 % and different compositions of the carrier liquid of the suspension. The varying liquid compositions (water mass fractions wH2O from 0 % to 70 %) exhibit different solubility of lactose, and therefore, resulted in different dissolved lactose mass loadings Xd from 0.072 % to 34.4 %. Figure 4 depicts the drying of individual droplets with the same suspended lactose mass fraction ws of 10 % and different compositions of the carrier liquid. The drying air flow with 0.66 m/s (Re = 170) was applied to control the vapor content of the acoustic streaming vortices. The drying of an individual droplet consisting of lactose and isopropanol without water revealed the fastest evaporation rate in the first drying stage. Adding water to the droplet liquid reduced the evaporation rate, resulting in a longer first drying stage. These drying kinetics of binary liquid mixtures are due to the different activities of the two liquid components, which furthermore depend on the composition of the liquid phase.
Figure 4. Drying of lactose suspensions with the same suspended lactose mass fraction ws of 10 % and different dissolved lactose mass loadings Xd between 0.072 % and 34.4 % (water mass fractions wH2O between 0 % and 70 %).
Figure 5. Scanning electron micrographs of microspheres collected from the levitator after drying at 25°C with 0.66 m/s air flow. Drying of lactose isopropanol-water mixtures with the suspended lactose mass fraction ws of 0.1. Left column (i) shows the surface, and right column (ii) the inside structure (area enclosed by dotted line) and the surface (indicated by dashed line) of the microspheres. The dissolved lactose mass loadings Xd and the water mass fractions wH2O in brackets are (a) 5.3 % (20 %), (b) 6.1 % (30 %), (c) 14.4 % (50 %) and (d) 34.4 % (70 %).
Another effect that reduces the drying rate at higher amounts of dissolved solid is the formation of bridges and shells. The amount of dissolved solids increased with the water content of the liquid, resulting in a higher total solid loading in the droplet. The dissolved solids precipitated during the drying process and formed bridges between primary particles or a skin on the surface, which reduced the rate of evaporation of the volatile components . At low concentration, the dissolved solids formed a porous structure allowing for fast solvent evaporation. In contrast to this, high amounts of dissolved solids precipitated at the surface and sealed the pores, hindering the evaporation of the liquids.
The electron micrographs in Figure 5 depict the surface and bulk morphologies of microspheres dried in the levitator with the constant suspended lactose mass fractionws of 10 % and different dissolved lactose mass loadings Xd between 5.3 % and 34.4 %. For low levels of Xd no significant effects of the dissolved solid could be observed, and primary particles had no visible connections by precipitated material. Raising the water mass fraction wH2O to 30 % and, therefore, the dissolved lactose mass loading Xd to 6.1 %, resulted in visible connections of the primary particles with precipitated solids on the surface (Figure 5b(i)). The precipitated mass acts as a binder and “glues” the primary particles together, resulting in a more densely packed network structure. The particle packing at the droplet surface becomes more intense as the dissolved lactose mass loading Xd increases to 14.4 % (Figure 5c(i)). An additional increase of Xd can even lead to full covering of the microsphere surface with precipitated solids (Figure 5d(i)). At the dissolved lactose mass loading Xd of 34.4 %, no primary particles were visible any more.
The mechanical strength of microspheres produced by acoustic levitation was measured with the SCS test in a slightly modified version. Figure 6 depicts characteristic profiles of the normal force on individual particles, produced and measured by the rheometric device. The first change in the normal force is observed as the upper plate takes up contact with the microsphere. This contact point indicates a typical dimension of the microsphere of 685 µm. Thereafter, the normal force increases rapidly until a maximum force is reached, which indicates the force needed for breaking the microsphere and, therefore, the mechanical strength of the object against normal stress. As the microsphere breaks, the recorded force decreases immediately. Further linear motion of the upper plate leads to normal forces due to the compression of the fragmented microsphere.
Figure 6. Normal force on individual microspheres dried in the acoustic levitator as a function of plate distance. The normal force is applied and recorded by a rheometer. The suspended lactose mass fraction ws was 10 %. The dissolved lactose mass loading Xd varied between 5.5 % and 8.7 %.
The first measurable breaking force was obtained for the dissolved lactose mass loading Xd of 5.3 %, as depicted in Figure 7a. Small changes in the dissolved lactose mass loadings Xd from 5.5 % to 6.1 % resulted in an increase of the breaking force. More dissolved solids connected the primary particles, resulting in stronger bonds and a significant increase of the breaking force within this transition zone. Increasing dissolved lactose mass loading Xd up to 6.1 % resulted in the formation of stronger agglomerates between primary particles, revealing a stronger increase of the breaking force. The same relation is observed by depicting the breaking force as a function of the water mass fraction wH2O in the suspension, as in Figure 7b.
Based on these data, by extrapolation the point was calculated where no force is needed to break the microsphere. This point reflects a suspension composition with a dissolved lactose mass loading Xd of 5.2 % or, equivalently, a water mass fraction wH2O of 19 %. This corresponds to our observation that quite loose microspheres are formed below these values. These results demonstrate the strong influence of the suspension composition on the mechanical strength of spray-dried particles. There exists a threshold value of the dissolved lactose loading Xd below which loose microspheres are formed. Primary particles are then loosely bonded by precipitated solids, so that the microspheres easily break up, even by weak forces.
Figure 7. The breaking force as a function of (a) the dissolved lactose mass loading Xd and (b) the water mass fraction wH2O.
Conclusions and Outlook
The present work investigates the influence of initial suspension droplet composition on the drying kinetics via acoustic levitation and the resulting microsphere properties in relation to spray drying. It is found that the dissolved lactose mass loading Xd determines the drying kinetics and has a strong influence on the dried microsphere morphology and mechanical strength. The individual microsphere structure changes from loosely packed primary particles with isopropanol as the solvent to denser packing with increasing water mass fractionwH2O. The dissolved solids precipitate during drying and bond the primary particles. The hardness of individual microspheres was measured by a compression strength test. Loosely agglomerated microspheres were formed below the threshold Xd = 5.2 % for the investigated model substance. Based on these findings, drying should be carried out below this threshold to enable drying either of primary particles (i.e. tailored particle properties after crystallization remain unaltered during drying), or above this threshold to produce larger, agglomerated particles (i.e. improved powder flowability as favored in direct compression).
Since this study was focused on only one suspension composition with an excipient as model substance, further research will be conducted with various APIs and solvent combinations. In the end, the sum of experimental data could be used to predict the microsphere strength based on the suspension composition.
 R. Vehring, “Pharmaceutical particle engineering via spray drying,” Pharm. Res., vol. 25, no. 5, pp. 999–1022, 2008.
 A. L. Yarin, G. Brenn, O. Kastner, D. Rensink, and C. Tropea, “Evaporation of acoustically levitated droplets,” J. Fluid Mech., vol. 399, pp. 151–204, 1999.
 A. L. Yarin, G. Brenn, O. Kastner, and C. Tropea, “Drying of acoustically levitated droplets of liquid-solid suspensions: Evaporation and crust formation,” Phys. Fluids, vol. 14, no. 7, pp. 2289–2298, 2002.
 O. Kastner, G. Brenn, D. Rensink, and C. Tropea, “The acoustic tube levitator - A novel device for determining the drying kinetics of single droplets,” Chem. Eng. Technol., vol. 24, no. 4, pp. 335–339, 2001.
 C. Gauge and T. S. Anvils, “Standard Test Method for Single Pellet Crush Strength of Formed Catalysts and Catalyst Carriers,” ASTM D4179-11, 2011.
 A. Mujumdar, Handbook of Industrial Drying, Fourth Edition, vol. 4. CRC Press, 2015.