Issue link: https://www.e-digitaleditions.com/i/1032386
20 OctOber 2018 Inhalation However, the blending conditions for the two instru- ments were very different. The total mixing time of 15 minutes required for the Turbula blender is about 30 times longer than the blending time needed in the Pico- mix mixer. is is due to the efficient mixing principle used in the Picomix instrument. e powder changes position in the mixing vessel not only horizontally but also vertically (Figure 4). is so-called "Cyclomix prin- ciple" leads to high shear forces acting on the powder. erefore, a decreased blending time of 30 seconds was sufficient to prepare homogenous powder blends. Furthermore, it was assumed that the Picomix mixer could affect the particle size distribution of the carrier, due to its blending tool and high mechanical stress. is hypothesis was verified via laser light diffraction, mea- suring particle size distribution after 1, 2 and 5 minutes at 500 rpm in the Picomix instrument. Figure 5 depicts the x 10 , x 50 and x 90 values as well as the percentage of par- ticles < 15 μm from Parteck M DPI, unmixed and after blending. A blending time up to 120 seconds at a rota- tion speed of 500 rpm showed no significant influence on all of these values (p > 0.435). For the shorter mixing time of 1 minute, a significantly (p = 0.003) lower per- centage of particles < 15 μm could be detected. is may be due to the fact that fines adhered to the mixer's walls and blending tool, or were pressed to the carrier and thus were not detected by laser diffraction. After 5 minutes, a significantly (p = 0.0005) higher percent- age of particles < 15 μm could be observed, caused by the effect of abrasion of the carrier. Increasing blending times led to further particle size reduction and a shift of x 90 to lower sizes. erefore, diminution during blend- ing should be observed carefully and the shortest blend- ing time possible should be used in a high shear mixer. For the high shear mixer, a twelve-times-faster rotation speed was used, which was the lowest possible rpm set- ting. Nevertheless, the total number of rotations was 1.5 times greater than in the tumble blender. e results of aerodynamic characterization of the SBS blends prepared with the two blenders did not show significant differences; fine particle fraction (FPF): p Novolizer > 0.388; p Cyclohaler > 0.059; fine particle dose (FPD): p Novolizer > 0.256; p Cyclohaler > 0.241 (Figure 6). In theory, a more intensive blending of the powder (with ments, blending conditions were explored to achieve a homogeneous blend with high recovery. Homogenous blends with a maximum RSD < 2.87% and a mini- mum recovery of 96.32% for the Picomix mixer and a maximum RSD < 1.44% and a minimum recovery of 98.43% for the Turbula blender were prepared. Figure 5 Overview of the x 10 , x 50 and x 90 values as well as the percentage of particles < 15 μm from Parteck M DPI unmixed and after blending in the Picomix mixer; mean values with standard deviation; n = 6 Parteck M DPI unmixed 500 rpm/1 min 500 rpm/2 min 500 400 300 200 100 0 Particle size (μm) Percentage of particles < 15 μm (%) 500 rpm/5 min x10 x50 x90 Parteck M DPI unmixed 500 rpm 1 min 500 rpm 2 min 500 rpm 5 min 20 10 0 100 80 60 40 20 0 102 101.5 101 100.5 100 0 1200 2400 Time (minutes) dm Target RH Mass (%) Target RH (%) Figure 3 DVS measurement of Parteck M DPI; mass in percent is depicted in gray a) b) Figure 4 Schematic drawings of the blender principles used in a) the Turbula low shear tumble blender and b) the Picomix high shear mixer