Current Approaches to APSD Measurements of OIPs Based on Inertial Impaction - Good Cascade Impactor Practices, AIM and EDA for Orally Inhaled Products - page 45

Fig. 2.17 Calibration data for the USP/Ph.Eur. induction port at 15, 30, and 60 L/min ( From
[ 55 ] —used with permission )

way as the collection surfaces of the CI stages or with up to 10 mL of a suitable
solvent to eliminate particle bounce and re-entrainment. A similar requirement
exists also for use of the NGI under these circumstances.
The preseparators for the ACI and NGI are illustrated in Figs. 2.18 and 2.19 ,
respectively. Ideally, the preseparator should not starve the first stage of the impac-
tor of particles. However, the preseparator used with the ACI has been shown by
calibration at 28.3 L/min to have its cut size close to 9 m in aerodynamic diameter,

almost identical with the d 50 size of the first stage (stage 0 in Table 2.2 ) at this flow
rate [ 58 ]. Its sharpness of cut is relatively poor by comparison to equivalent values
for the CI stages (GSD pre-sep is about 1.55), due primarily to the strong influence of
gravity. The preseparator therefore starves the second as well as the first stage of this
impactor.
Recognizing the limitations of the ACI preseparator, the consortium developing
the NGI designed its preseparator to function as a single component [ 57 ] but with
two distinct steps in the size-separation process (Fig. 2.19 ). The incoming aerosol is
first passed through a so-called scalper impingement stage that removes the coarsest
particles. The remaining aerosol then passes immediately through a more conven-
tional impaction stage before leaving the preseparator. Calibration with monodis-
perse particles has demonstrated that its GSD pre-sep is close to 1.3 at 30, 60, and
100 L/min. This value is only slightly greater than estimates of GSD stage for well-
designed CI stages in which inertial rather than gravitational or turbulent size-
separation predominates. Its measured d 50 values (10.0, 12.7, and 14.9

μ

m at 100,
60, and 30 L/min, respectively) are sufficiently separated from the corresponding
values for stage 1 (6.07, 8.29, and 11.4

μ

μ m) that starvation of the first stage of the

CI does not occur to a significant extent.