Current Approaches to APSD Measurements of OIPs Based on Inertial Impaction - Good Cascade Impactor Practices, AIM and EDA for Orally Inhaled Products - page 39

Fig. 2.12 ACI system: mass fraction on first AIM collection stage or ratio of AIM system com-
pared with corresponding summed stages—MMAD varied in the range 1-5 μ m with GSD fixed at

2.0 ( From [ 43 ] , courtesy of D L Roberts and J P Mitchell )

of APSD profile. Roberts and Mitchell concluded that such a small discrepancy
would likely not be observable experimentally. In addition, for an aerosol with
MMAD of 4.0

m, the fraction on the first stage of the abbreviated NGI was close
to 0.5 and nearly independent of the dispersity of the input aerosol (Fig. 2.15 ). This
was not the case for the corresponding ACI data where the mass fraction on the first
stage (stage 2) reduced as the dispersity of the incoming aerosol decreased
(Fig. 2.12 ). The improved robustness of the NGI system derives from the relative
sharpness of the individual NGI stage efficiency curves compared with those of the
ACI and the consequent minimal overlap between neighboring stages [ 21 ]. The
increase in the mass fraction collected by the first stage (stage 2) of the abbreviated
ACI with the model aerosols having larger MMAD values (Fig. 2.12 ) and also for
the corresponding situation with the NGI system (Fig. 2.14 ) would be a linear func-
tion of MMAD if there is no bias associated with the nonideal nature of the stage
collection efficiency curves.
From this analysis of both CI systems, Roberts and Mitchell observed that bias
in measures of the ratio of “AIM” to “summed stages” is systematically in one
direction but can sometimes expand and on other occasions reduce the difference
between the input aerosol and the abbreviated CI results in terms of mass fractions.
Importantly, no cases were observed in which the abbreviated ACI deviated from

μ