Influence of droplet size on the crystallization behaviour of aqueous D-mannitol solutions during...

ILASS – Europe 2011, 24th European Conference on Liquid Atomization and Spray Systems, Estoril, Portugal, September 2011

1

Influence of droplet size on the crystallization behaviour of aqueous D-mannitol solutions during spray drying

E. M. Littringer*1, A. Mescher2, H. Schroettner3, S. G. Maas4, P. Walzel,2 N.A. Urbanetz1

1: Research Center Pharmaceutical Engineering, Austria 2: Department of Biochemical and Chemical Engineering, TU Dortmund, Germany

3: Austrian Centre for Electron Microscopy and Nanoanalysis, TU Graz, Austria 4: Institute of Pharmaceutics and Biopharmaceutics, Heinrich-Heine-Universitaet Duessel-

dorf, Germany

Abstract As the morphology of spray dried products may be strongly impacted by the droplet size of the atomizing spray, the aim of this work is to study the influence of droplet size of aqueous D-mannitol solutions on the crystalliza-tion behaviour during spray drying. Therefore drying experiments of 15 % [w/w] aqueous mannitol solutions were performed on a lab- and pilot-scale spray dryer at different outlet temperatures. The obtained spray dried products are intended to be used as carrier particles for pulmonary drug delivery. At lab scale, where droplets of approximately 30 µm are generated, crystallization from a melt is observed, leading to rough particles at high outlet temperatures and smooth particles at low ones. At pilot scale, where droplets of approximately 140 µm are dried, crystallization from a solution is found which leads to rougher surfaces, consisting of larger single crystals at lower outlet temperatures than at higher ones.

Introduction In pharmaceutical technology there are many processes, such as tabletting, mixing or dosing, where powder

properties play a crucial role. Often excipients of defined properties are added to improve for example the com-pactability in tabletting or the flowability in multi-dose dry powder inhalers (DPIs). Due to the fact that powder flow, which is important for many processes, largely depends on particle shape spray drying is a favourable technology as generally spherical particles of good flowability are obtained. However the most prominent ex-cipient, lactose, is found amorphous or partly amorphous upon spray drying [1-3]. Therefore this technology is not suitable for the preparation of spherical lactose particles because of recrystallization which may take place during storage. However mannitol, another well known excipient, is found crystalline after spray drying.

Maas et al. [4] showed and described the advantages of spray dried mannitol as carrier particles for the ac-tive pharmaceutical ingredient (API) in dry powder inhalers. In such powders the reproducible delivery of the API to the lower lung is a prerequisite. To reach the targeted site, API particles should have aerodynamic diame-ters of 1 µm to 5 µm. Powders consisting of such particles are highly cohesive and exhibit poor flowing proper-ties. The solution to cope with cohesivity and therefore flowabiltiy is overcome by the addition of excipients of larger size. In a mixing process the fine API particles are mixed with so called carrier particles. During this proc-ess the API particles are attached to the surface of the larger excipient particles. Interparticle forces are highly important in such mixtures. They must be high enough to guarantee mixing homogeneity and stability of the mixture during transport and dosing and low enough to enable the detachment of the fine API particles upon inhalation. If the API particles do not detach from the surface of the carrier they will impact, due to inertia, in the throat and will not have access to the lung. Interparticle forces in such mixtures are influenced by various factors such as particle shape, roughness, hygroscopicity, mixing rate and many more. There are several studies which show the importance of surface roughness on the performance of DPI lactose carrier particles [5-7].

Maas et al. [8] showed that spray drying of aqueous mannitol solutions at different outlet temperatures at lab scale lead to carrier particles of different surface topography. Further he showed that the amount of API that reaches the lung can be successfully targeted by surface roughness [4]. However one problem encountered by Maas was the lack of reproducibility of dosing due to the particle size of the carriers, which were not large enough to improve the flowability sufficiently. The particle size was restricted to approximately 12 µm due to the size of the spray tower. Larger droplets, leading to larger particles, require larger drying times, hence larger spray towers. Therefore the aim of Littringer et al. [9] was the preparation of larger mannitol carrier particles of adjustable surface roughness and sufficient flowability on a pilot scale spray dryer.

* Corresponding author: [email protected]

24th ILASS – Europe 2011 Influence of droplet size on the crystallization behaviour of aqueous D-mannitol solutions during spray drying

2

Beside surface roughness Littringer et al. [10] studied the influence of spray drying process parameters on particle size, polymorphism and breaking strength of the mannitol powders. One of the main results concerning surface roughness, mentioned in [9] and [10], was that low outlet temperatures lead to rough, coarse crystalline particles and higher temperatures to smoother ones. These results are in contrast to those of Maas et al. [8] who showed smooth surfaces at low outlet temperatures and rough surfaces at higher ones. Maas et al. as well as Lit-tringer et al. mention that the different roughness is a result of the underlying crystallization processes. Therefore the aim of this work is to study the influence of process scale, thus droplet size - and particle size respectively - on the crystallization behaviour of aqueous mannitol solutions during spray drying.

Materials and Methods

Material Mannitol (Pearlitol® 200SD) was kindly provided by Roquette Frères (Lestrem, France).

Droplet size analysis at pilot scale The particle size distribution of the spray generated by a laminar operated rotary atomizer [11] of a 15 %

[w/w] aqueous mannitol solution at room temperature with a feed rate of 10 l/h was measured by a laser diffrac-tion system (Malvern 2600, Malvern Instruments, Malvern, United Kingdom). The distance between the laser beam and the atomizer wheel was at 30 cm radially from the atomizer outer circumference. Parameters used for spray characterization were volume mean droplet size d50,3 and droplet span [(d90,3-d10,3)/d50,3] .

Spray drying at pilot scale A laminar operated rotary atomizer [11] with a diameter of 100 mm and containing 60 bores of 3 mm was

used in order to produce particles of the desired mean size, but with a narrow particle size distribution. The at-omizer was running at a speed of 7200 min-1. The dimensions of the pilot scale spray dryer were: diameter 2.7 m, total height 3.7 m. The spray was produced from a solution of mannitol dissolved in water (15 % [w/w]) at room temperature with a feed rate of 10 l/h. Two products at 67 °C and 102 °C outlet temperature, termed M67 and M102, were prepared.

Spray drying at lab scale Spray dried particles were prepared on a Niro Mobile Minor (Niro Atomizer, Niro, DK-Kopenhagen) with a

50 mm rotary atomizer having 24 openings of 6 mm height and 3 mm width. The atomizer revolution rate was 27500 min-1. The tower dimensions were: 0.800 m inner diameter, total height 1.315 m. The spray was also pro-duced from a solution of mannitol dissolved in water (15 % [w/w]), again at room temperature but with a feed rate of 0.84 l/h. Three products at 60 °C, 90 °C and 120 °C outlet temperature were prepared and termed M60, M90 and M120.

Particle surface investigations The powder samples were examined using a scanning electron microscope (SEM) (Zeiss Ultra 55, Zeiss,

Oberkochen, Germany; Particles were sputtered with gold-palladium prior to analysis) operating at 5kV and a SEM (Hitachi H-S4500 FEG, Hitachi High-Technologies Europe, Krefeld, Germany; Particles were unsputtered) operating at 1kV.

Particle size distribution Particle size distribution of pilot scale spray dried products was determined by analytical sieving for 15 min

(amplitude 20 %) on a sieving machine (Analysette Type 3010, Fritsch GmbH, Idar-Oberstein, Germany). Laser light diffraction (Helos/KF-Magic, Sympatec, Clausthal-Zellerfeld, Germany) including a dry dispers-

ing system (Rodos, Sympatec, Clausthal-Zellerfeld, Germany) was used to determine particle size distributions of powders dried at lab scale. The powder was fed to the diperser via a vibrating chute (Vibri, Sympatec, Claust-hal-Zellerfeld, Germany). The measurements were carried out at the dispersing pressure of 2.1 bar and the nega-tive pressure of 71 mbar. Evaluation of the data was performed using the software Windox 4 (Sympatec, Claust-hal-Zellerfeld, Germany).


3

Particle structure investigations Pilot scale spray dried particles were embedded in epoxy resin (Spezifix 40, Struers, Willich, Germany) and

after polymerization cross sections were prepared by a microtom (Ultracut UCT, Leica). Cross sections were investigated by SEM (Zeiss Ultra 55, Zeiss, Oberkochen, Germany; Particles were sputtered with gold-palladium prior to analysis) at 5 kV.

Similarly lab scale spray dried products were embedded in epoxy-resin (Epofix, Struers, Willich, Germany), cut via microtom (Jung Rm2055, Leica, Wetzlar, Germany) and analyzed by SEM (Hitachi H-S4500 FEG, Hi-tachi High-Technologies Europe, Krefeld, Germany; Particles unsputtered) operating at 1kV

Hot stage microscopy The investigation of the drying process was carried out using an Olympus BH2 polarization microscope

(Olympus Optical, Vienna, Austria) equipped with a Kofler hot stage (Reichert Thermovar, Vienna, Austria) at 60°C and 120°C. Aqueous mannitol solutions of 15 % (w/w) were studied. Photomicrographs were acquired using the Olympus BH2 and a stereomicroscope (Olympus SZX-12, Olympus Optical, Vienna, Austria) equipped with a ColorViewIII CCD camera using the software Cell-D (Olympus Optical, Vienna, Austria).

Results and Discussion

Spray drying at pilot scale Analysis of the spray, generated by the laminar operated rotary atomizer, shows a volume mean droplet di-

ameter of 137 µm and a span of 0.28. Considering the true density of mannitol of 1490 kg/m3 [12] and a density of 1050 kg/m3 of the 15 % [w/w] aqueous mannitol solution at room temperature the mean particle size after removal of the solvent should be 65 µm. However the mean particle size of the spray dried powders was in the range of 80 µm (Tab. 1), thus suggesting hollow or porous spray dried particles with a porosity of 47 %. Cross sections of the spray dried products (Fig. 2) verify this by clearly showing a void space inside the particles.

M67 M102


4

Fig. 1: SEM micrographs (Hitachi H-S4500 FEG) of products dried at pilot scale at 67 °C and 102 °C outlet

temperature. Figure 1 shows the different morphologies that are obtained at pilot scale at 67 °C and 102 °C outlet temper-

ature. It can be clearly seen that at low outlet temperature (67 °C) particles with a rough surface are obtained (Fig. 1, left column). The surface of those particles is covered by crystals of rod-like shape. The single crystals, which are approximately 3 µm, can be clearly distinguished. However those crystals have no preferred orienta-tion. At 102 °C the surface of the particles is still rough, due to single crystals at the surface, but they are smaller than those at 67 °C outlet temperature.

When talking about particle shape spray drying is always referred as to form spherical particles, which is, especially for larger particles not always true. For M67 it was possible to get perfectly round particles. However the increase in outlet temperature to 102 °C results in the emergence of shriveled particles of irregular shape (Fig. 1).

M67 M102


5

Fig. 2: SEM micrographs (Zeiss Ultra 55) of cross sections of pilot scale spray dried particles which were

embedded into epoxy resin. The SEM micrographs of cross sections, which are shown in Fig. 2 reveal that all spray dried products, in-

dependent on the outlet temperature, consist of an outer shell and a porous inside. This can be explained by the fact that droplets, which have an initial mean droplet diameter of approximately 137 µm, get smaller due to the evaporation of water. Consequently the concentration of mannitol, especially at the surface of the droplet rises until the solubility concentration or even a certain supersaturation is exceeded and the crystallization process starts resulting in the formation of a shell. With the formation of an outer shell the evaporation rate declines and due to the lower drying rate larger crystals inside the particles are formed giving the particle a porous appear-ance.

Table 1. Particle size distribution of mannitol spray dried at pilot scale determined by analytical sieving.

x10,3 / µm x50,3 / µm x90,3 / µm Span

M67 55 84 124 0,82

M102 53 77 113 0,78

Analytical sieving was used to determine the particle size distribution of the powders. As the atomization conditions were similar, the samples show a comparable mean particle size of approximately 80 µm, as well as a comparable width of the size distribution, characterized by the span values.

Spray drying at lab scale Due to the low volumetric flow rate a pulsating spray generation was observed and no reliable information

on droplet sizes could be gained by laser diffraction analysis. In order to estimate the droplet size of the spray, and afterwards to relate it to the particle size of the final product, the drop size was calculated according to Maas et al. [8] for the case of rotary atomization by dripping. A force balance between centrifugal acceleration and the capillarity of the liquid led to an averaged droplet diameter of approximately 28 µm. An investigation of the spray generation by usage of a stroboscopic light source could verify the droplet formation by dripping.

M60 M90 M120


6

Fig. 3: SEM micrographs (Hitachi H-S4500 FEG) of products dried at lab scale at different outlet tempera-

tures. SEM micrographs of spray dried mannitol display spherical particles with a smooth surface (Fig.3, left col-

umn) when spray drying is performed at an outlet temperature of 60 °C (M60). A smooth surface (Fig. 3, middle column) is also obtained when an outlet temperature of 90 °C is used (M90). In contrast, at the outlet tempera-ture of 120 °C, spherical particles (M120) showing a rough surface, are obtained, due to the formation of larger prismatic mannitol crystals (Fig.3, right column). This result is somewhat surprising because high temperatures are expected to cause rapid drying of the droplets, which usually favors the formation of small crystals, whereas the opposite is expected to be true for low drying temperatures.

M60 M120

Fig. 4: SEM micrographs (Hitachi H-S4500 FEG) of cross sections of lab scale spray dried particles which

were embedded into epoxy resin. SEM micrographs of the cross sections of the M60 and M120 (Fig. 4) show, that, like at pilot scale, the par-

ticles consist of a shell and a hollow or porous inside. The mechanism for the formation of a shell and a porous inside is the same like at pilot scale. Due to the evaporation of water the concentration of mannitol at the surface of the droplets rises until a certain concentration is exceeded where the crystallization takes place resulting in the formation of a shell. The remaining solution inside the particles then crystallizes at lower rates. Interestingly the void of the M120 particles is filled with epoxy resin, which entered the particle via openings in the particle shell which can be seen in Fig. 3 both for M90 as well as M120. However M60 particles and particles dried at pilot scale do not contain epoxy resin because the particle shell is closed. The reason for the openings in the lab scale M90 and M120 particle shells might be due to high vapor pressures inside the particles caused by the evapora-tion of the remaining solution, leading to higher vapor pressures at higher temperatures, and the limited permea-bility of the vapor through the shell. For these particles the pressure inside the particles rises until a certain pres-sure is exceeded and the shell breaks up at its weakest point.


7

Table 2. Particle size distribution of mannitol spray dried at lab scale which was determined by laser diffrac-tion.

x10,3 / µm x50,3 / µm x90,3 / µm Span

M60 3,73 13,53 27,77 1,8

M90 4,93 13,59 26,48 1,6

M120 4,89 13,19 23,71 1,4

In order to evaluate the particle size characteristics of the spray dried products prepared at different tempera-tures laser diffraction was applied. The mean size of all products was approximately 13 µm (Tab. 2). In contrast to particles dried at pilot scale, where a span of approximately 0.8 was found, a higher span value of approx-imately 1.6 was calculated for those at lab scale.

Crystallization processes at pilot and lab scale To gain insight into the formation mechanisms of the pilot and lab scale spray dried particles at different out-

let temperatures, simple hot stage microscopy experiments were performed. Droplets of an aqueous mannitol solution on a glass slide were placed on a hot stage, preheated to 60 °C or 120 °C. The changes upon drying were observed using a polarized light microscope.

Fig. 5: Hot stage microscopy experiments at 60 °C (left) and 120 °C (right). At 60 °C recrystallization starts after a few seconds and results in mostly small acicular crystals that grow in

radial manner from several emerging nucleation centers (Fig. 5, left). Droplets placed on the hot stage at 120 °C shrink quickly due to the fast evaporation of water but most of them remain liquid and do not recrystallize, indicating that the nucleation rate is low at these conditions (Fig. 5, right). The highly supersaturated, viscous solution or even almost waterfree melt recrystallizes instantly in the presence of seed crystals or when the temperature is lowered by 10 °C to 20 °C. Because of the fast crystal growth, significantly larger crystals occur than at 60 °C. These hot stage microscopy experiments show clearly, that the differences in surface topography of the spray dried particles at lab scale are based on two different crystallization processes. At low temperatures (around 60 °C) fine needles crystallize from the supersaturated solution resulting in smooth surfaces. In contrast, at high temperatures (around 120 °C) the mannitol solution dries quickly to a highly supersaturated, viscous liquid since the solvent evaporation rate is high and the nucleation rate of crystalline mannitol is low at these conditions. This metastable supersaturated, viscous liquid crystallizes to coarse crystals resulting in particles or spheres with rough surfaces. The crystallization is triggered most likely by a secondary nucleation process (contact with other, already crystalline particles in the spray dryer which act as seeds).

At pilot scale supersaturation and viscosity of the liquid is lower due to lower evaporation rates. These lower evaporation rates can be explained be several factors. On the one hand the atomizer at lab scale rotates at a higher circumferential speed, leading to higher relative velocities between the emitted droplets and the surrounding air. As a matter of fact the boundary layer is smaller and therefore higher evaporation rates are obtained. On the other hand for larger droplets the ratio of surface to volume is lower than that at lab scale


8

resulting in lower drying rates at pilot scale than at lab scale. Because of the lower drying rates at pilot scale no highly supersaturated, viscous liquid may be formed.

The surfaces observed at pilot scale can be explained by an ordinary crystallization process, where drying at low rates favours crystal growth, whereas drying at higher rates usually favors nucleation resulting in the formation of numerous smaller crystalls. Therefore the surfaces of pilot scale dried particles at 67 °C consist of larger crystalls than those at 102 °C.

Conclusion and outlook This study shows that particle morphology, especially surface roughness, largely depends on the prior drop

size and thereby on the scale of the used spray dryer. The scale-up of the particle size is difficult without impacting morphology. Depending on the drying rate, which is largly influenced by the droplet size, two different crystallization process of aqueous mannitol solutions take place. Either there is crystallization from a solution or crystallization from a supersaturated, viscous liquid or even waterfree melt. At lab scale and at low outlet temperatures crystallization from a solution is observed, leading to smooth surfaces, probably composed of many very small crystals at the surface. At higher outlet temperatures crystallization from a supersturated, viscous liquid takes place and crystals at the surface are larger giving the surface a rougher appearance. At pilot scale, where larger droplets are dried, no viscous liquid is formed due to lower drying rates. That is why for all temperatures, studied at pilot scale, crystallization from a solution occurs, where usually drying at low temperatures results in the formation of larger crystals due to low nucleation rates. In contrast, drying at higher temperatures leads to higher nucleation rates and hereby to smoother surfaces.

Acknowledgements The authors wish to thank Roquette Frères (Lestrem, France) for providing D-mannitol and the DFG-SPP

1423 for financial support.

References [1] Briggner, L.E., Buckton, G., Bystrom, P., International Journal of Pharmaceutics 105:125-135 (1994). [2] Gombas, A., Antal, I., Szabo-Revesz, P.,Marton, S., Eros, I., International Journal of Pharmaceutics

256:25-32 (2003). [3] Lehto, V.P., Tenho, M., Vaha-Heikkila, K., Harjunen, P., Paallysaho, M., Valissaari, J., Niemela, P.,

Jarvinen, K., Powder Technology 167:85-93 (2006). [4] Maas, S.G., Schaldach, G., Walzel, P., Urbanetz N.A., Atomization and Sprays, 20(9):763–774 (2010). [5] Young, P.M., Roberts, D., Chiou, H., Rae, Herbert, Chan, H.-K., Traini, D., Journal of Aerosol Science

39:82-93 (2008). [6] Dickhoff, B.H.J, de Boer, A.H., Lambregts, D., Frijlink, H.W., International Journal of Pharmaceutics

327:17-25 (2006). [7] Zeng, X. M., Martin, G. P., Marriott, C., Pritchard, J., Journal of Pharmacy and Pharmacology 52:1211-

1221 (2000). [8] Maas, S.G., Schaldach, G., Littringer, E.M., Mescher, A., Griesser, U., Braun, D.E., Walzel, P., Urbanetz,

N.A., Powder technology, submitted. [9] Littringer, E.M., Mescher, A., Eckhard, S., Schröttner, H., Langes, C., Fries, M., Griesser, U., Walzel, P.,

Urbanetz, N.A., Drying Technology Journal, submitted. [10] Littringer, E.M., Mescher, A., Schröttner, H., Walzel, P., Urbanetz, N.A., 23rd International Conference on

Liquid Atomization and Spray Systems, Brno, Czech Republic, September 2010. [11] Schröder, T. and P. Walzel, Chem. Eng. Technol. 21(4): 349 – 354 (1998). [12] Burger, A., Henck, J.-O., Hetz, S., Rollinger, J. M., Weissnicht, A. A., Stöttner, H., Journal of

Pharmaceutical Sciences 89:457-468 (2000).

Date post:	01-Dec-2023
Category:	Documents
Upload:	independent
View:	0 times
Download:	0 times

Influence of droplet size on the crystallization behaviour of aqueous D-mannitol solutions during...

Documents