Current Approaches to APSD Measurements of OIPs Based on Inertial Impaction - Good Cascade Impactor Practices, AIM and EDA for Orally Inhaled Products - page 35

Table 2.12 Comparison of input MMAD and GSD with CITDAS ® -generated
values for both ACI and NGI ( From [ 42 ] , courtesy of DL Roberts and JP Mitchell )
Input APSD ACI: Q = 28.3 L/min NGI: Q = 30.0 L/min

MMAD ( μ m) GSD MMAD ( μ m) GSD MMAD ( μ m) GSD

1.10 2.0 1.16 1.98 1.10 2.00

2.50 2.0 2.53 1.99 2.57 2.08

4.00 2.0 3.96 1.93 4.12 2.07

5.00 2.0 4.77 1.96 5.02 2.06

4.00 1.5 4.00 1.50 4.17 1.57

4.00 1.2 3.98 1.26 4.13 1.33

Fig. 2.9 Input and calculated APSDs from a hypothetical OIP-generated unimodal, lognormal
distributed aerosol having MMAD of 5 μ m and GSD of 2.0 ( From [ 42 ] , courtesy of D L Roberts

and J P Mitchell )

GSD values (Table 2.12 ). The larger discrepancies arose as the result not so much
because of the nonideal nature of stage collection efficiency curves but because of
the inaccuracies of the two-point linearization of the lognormal curve fit. These
divergences were more evident when the MMAD of the input aerosol was nearly
identical with one of the stage d 50 values.
The difference between the input and the calculated aerosol APSDs is illustrated
in the example shown in Fig. 2.9 where the input MMAD and GSD were chosen to
be 5

m and 2.0, respectively. This example was chosen because the differences
between input and calculated APSDs arose mostly as the result of overlap between
neighboring CI stages. The calculated APSD from the NGI data was very close to
that of the input aerosol, whereas the calculated APSD for the ACI was shifted only
slightly to finer sizes, mostly as a result of the significant overlap between stage 2
and stages 1 and 3 with this CI. This finding most probably arises because of the

μ