Good Cascade Impactor Practices - Good Cascade Impactor Practices, AIM and EDA for Orally Inhaled Products - page 102

Fig. 4.9 Differential
mass-weighted APSD showing
how data for the same aerosol
(MMAD = 5.0 μm; GSD = 2.0)
can be meaningfully
represented from two different
CI systems (ACI and NGI):
the dotted line is the true
APSD ( From [ 78 ]— used by
permission )

In this instance, API mass is related to the aerodynamic diameter scale through
the d 50 values of the size-separating stages, but the non-sizing components (in this
case stage 0, which has no upper bound defined, the induction port and pre-separator)
are represented by thin lines rather than a size range, with the size scale broken
between the data from the two categories to designate the difference in properties
with respect to aerodynamic diameter. This representation is popular if the measure-
ments are all made using the same CI system and/or the OIP aerosol is captured
predominantly in the size-separating portion of the system.
However, Mitchell and Dunbar showed that its weakness is the lack of a more
formal link between API mass and size, making it impossible to compare data from
two different CI systems in a readily accessible way [ 78 ].
To achieve this goal, it is necessary to invoke the formalism of the differential
mass-weighted APSD, in which the mass frequency of API on each stage of the
size-fractionating part of the CI system is obtained by dividing the absolute mass by
the width of the size range associated with that particular stage (Fig. 4.9 ), in accor-
dance with the relationship:

m
Md

1

Fd ( ) =

i

(4.5)

m e , i

∆

EM 50 , i

in which d ae, i represents the aerodynamic diameter of the i th stage, m i is the mass of
API recovered from that stage, M EM is the emitted mass from the inhaler (or spacer/
valved holding chamber patient interface in the case of pMDI testing with these
add-on devices), and Δ d 50, i is the size width associated with the stage of interest.
Under these circumstances, it is possible to represent APSDs from different CI
systems on the same graph for direct comparison, and the midpoint of each bar of
each histogram corresponds to the place at which the curve of the true APSD (i.e.,
size-resolved to an infinitesimally small amount) intersects.