Current Approaches to APSD Measurements of OIPs Based on Inertial Impaction - Good Cascade Impactor Practices, AIM and EDA for Orally Inhaled Products - page 34

Table 2.11 Parameters for
ACI stage collection
efficiency curves at
Q = 28.3 L/min

Stage A ( μ m −1 ) B (dimensionless)

0 −1.324 12.51

1 −1.874 11.39

2 −1.928 9.604

3 −2.808 9.329

4 −4.494 9.668

5 −8.258 8.787

6 −14.82 9.483

7 −17.76 7.725

σ g is the geometric standard deviation (GSD). The quantities E 0 … E N are
the fractional efficiency curves of stages 0- N of the CI.
They then developed analytical forms of the real stage efficiency curves of both
CIs. For the NGI, these were based on the hyperbolic tangent functional form first
described by Rader et al. [ 40 ] to compute how this model aerosol distributed itself
on each stage of the CI under consideration:

Here,

Z

d
Y

(2.13)

i

pc

Ed ()tanh =

i c

i

in which d pc is the modified particle diameter corrected for slip in accordance with
the expression

r
r 0

()

p

dd = CdCd =

(2.14)

pc p c ae c ae

and Y i and Z i are best fit parameters for each impaction stage. The values of the
parameters in (2.13) that fit the archival calibration data for the NGI are in the sup-
plementary information associated with Roberts [ 14 ] and available on-line at http://
www.tandfonline.com/doi/suppl/10.1080/02786820903204060
For the ACI at 28.3 L/min, Roberts and Mitchell used the exponential function
developed by Gulak et al. [ 44 ] for the stage efficiency curves:

1

Ed () =

(2.15)

i e

1 + exp( Ad + B )

i e i

Table 2.11 lists the values of the parameters A i and B i that best fitted the adjusted
calibration data of Vaughan for this CI.
When calculated MMAD and GSD parameters for selected APSDs were com-
pared with those of hypothetical aerosols likely to be encountered with OIPs that
entered the NGI or ACI (using CITDAS ® software; version 3.00, Copley Scientific,
Nottingham, UK), the differences were at most 5% for the MMAD and 11% for the