Content uploaded by Daryl L Roberts
Author content
All content in this area was uploaded by Daryl L Roberts on Jan 13, 2019
Content may be subject to copyright.
Inhalation DECEMBER 2011 7
Inspecting used impactors—Practice
and principles
Cascade impactors are the backbone of quality control
for inhalable drug products. Consequently, ensuring that
used impactors are suitable for continued use is a basic
need for developers and manufacturers of inhalable
drug products. Industrial standards, such as the United
States Pharmacopeia or the European Pharmacopoeia,
describe periodic optical inspection of the nozzles of
each stage of an impactor as the means of ensuring that
the impactor is suitable for continued use. The notion
of “periodic” is left to the user to validate, but it is most
common for users to seek independent, third-party
inspection of impactors every six months or 12 months.
Most of the available literature on inspecting impactors
focuses on new impactors. For example, Chambers et
al.1have described available optical inspection tools for
measuring the diameter of the nozzles in cascade
impactors, particularly as applied to representative new
impactor nozzles. Mitchell2summarized some of the
existing practices for determining whether used
impactors are suitable for continued use, including leak
checking, cleaning and optical inspection, but there is
little evaluation of preferred methods.
In used impactors, more questions arise during inspec-
tion than for new impactors. For example, used
impactor nozzles have more irregular edges than do
new impactor nozzles. Consequently, the focus of an
optical inspection machine at the nozzle exit plane can
become less defined, sometimes as a result of corrosion
(mostly of the aluminum impactors) or because of
inappropriate cleaning attempts. In addition, nozzles are
three-dimensional and from time to time, extraneous
debris can appear inside the nozzle of a used impactor.
Questions arise about when and if and how to clean
nozzles. Overall, the assessment of suitability of used
impactors, therefore, requires attention to a number of
factors not present with new impactors.
This article describes ways to approach these issues and
suggests practical methods for inspecting used impactors
in a manner that accounts for their various non-ideal
conditions.
Shipping and getting started
Surprisingly, users ship impactors for inspection using a
wide range of physical packaging arrangements, which
are not always favorable. Frequently, used NGIs are
shipped without cups that could protect the nozzle
pieces. Used Andersen impactors are often shipped with
their nozzle pieces placed separately in the packaging
rather than being stacked and clamped together.
Transporting a completely assembled impactor is the
safest arrangement because the nozzles are protected
from damage. Figures 1a and 1b show sample best prac-
tices for shipping the NGI and the Andersen impactor.
When a used impactor arrives at its destination, after
visual inspection for dents, scratches and debris, the
first question that must be determined is the degree of
Daryl L. Roberts and Christian A. Lavarreda
MSP Corporation
Being confident about
quantifying the imperfect
c-MSPart_7-11_Jasper Art5.0 #11 11/17/11 1:14 PM Page 7
cleaning the impactor stages should undergo—if any—
before optical inspection. Impactor users need to know
if their routine laboratory work practices keep their
impactors suitable for use during the six months or 12
months between inspections. Obviously, if the optical
inspector were to clean the nozzles of the impactor
stages before inspecting, the user cannot answer this key
question. Therefore, it is important no t to clean the
impactor before inspection, but rather to lightly blow
the impactor stages with clean compressed air to
remove dust particles or packaging debris that may have
attached themselves to the impactor during transit. This
practice gives the optical inspector the ability to tell the
user the information he or she needs, namely the “as
found” results, without surprises from spurious dust,
hair or packaging materials.
Optical inspection
In concept, the optical inspection, sometimes called
“mensuration,” determines if the nozzles are the proper
diameter, as in USP section 601 and EP section 2.9.18.
In Figures 2a, 2b and 2c, we see three types of nozzles
—a perfect nozzle, a typical nozzle and a rare and diffi-
cult nozzle. Inspection methods for used impactors
must anticipate all types, not just the clean, round noz-
zles typical of new impactors.
Figure 2c emphasizes not simply the potential for a noz-
zle to be dirty. Rather, this figure shows the three-
dimensional aspect of impactor nozzles, which are typi-
cally as deep, in the direction of air flow, as the nozzle
exit-plane diameter. In Figure 2c, the focus is on the
nozzle exit plane and the out-of-focus contaminants are
upstream of the nozzle exit plane.
8DECEMBER 2011 Inhalation
Figure 1
Proper packaging for shipping used impactors
a) NGI
b) Andersen impactor
Figure 2
Examples of nozzles found in used impactors
a) Clean, round nozzle
b) Typical, used nozzle
c) Nozzle with contaminants at the exit
plane and the interior of the nozzle
c-MSPart_7-11_Jasper Art5.0 #11 11/17/11 1:14 PM Page 8
Inhalation DECEMBER 2011 9
Because of the various imperfections encountered in
the inspection of used impactors, it is critical that the
optical inspection eliminate human focus and other
judgment factors. Repeatability, especially in the pres-
ence of human bias, is a necessary but insufficient
requirement for optical inspection. Optical inspections
must also be accurate. For this reason, optical inspec-
tions should be conducted with equipment that has an
automated, operator-independent means of defining the
edge of the nozzle and of determining the lighted area
of the nozzle. The equipment used at MSP Corporation
is the Mitutoyo QV404 optical inspection tool, which is
known to have one-micron or smaller repeatability and
accuracy for measuring nozzles of the most common
inhaler testing impactors.3For the NGI and the
Andersen impactor, the nozzle locations have been pro-
grammed for automatic operation. More importantly,
the system has been qualified with proper settings for
focus, brightness and the intensity definition of an edge
pixel, with NIST-traceable reticles, eliminating operator
biases.
Handling optical data
Optical systems like the Mitutoyo report two diameters
from two fundamentally different means of looking at
the nozzles. In one means of assessing the lighted pix-
els, the Mitutoyo system determines the Area-Based
Diameter (ABD) by counting the number of lighted
pixels. The ABD is calculated from the total area of
lighted pixels by the equation
AB D ⫽4* are a 1/2
Essentially, the ABD is the diameter of a perfect circle
that would have the same area of lighted pixels as the
nozzle being examined.
The Mitutoyo system also reports the Edge-Contrast
Diameter or ECD. The edge is located first by finding
all pixel pairs that consist of a lighted pixel and an
unlighted pixel oriented with the lighted pixel toward
the inside of the circle. The Mitutoyo software then cal-
culates the length of a contiguous line that connects the
intersections of these pixel pairs. The length of this line
is the circumference of the edge of the nozzle, which
has been given the symbol “C.” The ECD is calculated
from the formula
ECD ⫽C
Essentially, the ECD is the diameter of a perfect circle
that has the same circumference as the nozzle being
examined.
For various reasons of imperfection in the nozzles, the
ABD and the ECD are not equal in used impactors (nor
are these measures precisely the same for new impactor
nozzles, providing a measure of roundness). Both diame-
ters are important to the aerodynamic performance of
the nozzle. Roberts4has shown that the most meaningful
nozzle diameter is the “effective diameter” of each noz-
zle (and of the stage) calculated by the simple formula
Def f ⫽(AB D )2/3•(ECD)1/3
A rational acceptance criterion for used impactors is
therefore whether the effective diameter of a stage is in
the range of nozzle diameters allowed by the USP and
the EP.4
What if some nozzles need cleaning?
There are times when a stage of an impactor has little
or no “in-use” margin [that is, it is right on the edge of
the allowed range of effective diameter]. It is important
to seek the user’s input at this point, because it is impor-
tant not to change the impactor without the user’s con-
sent. It is ideal to tell the user exactly which nozzles
need cleaning, but in some cases, the entire stage needs
to be thoroughly cleaned. Suitable cleaning methods
include ultrasonification and manual cleaning with small
dental brushes (e.g. Kleen Teeth; http://www.kleen-
teeth.com/mycategory.php?do=cat&cname=End-
Tufted_Brushes.html) or pins that are smaller than the
nozzle diameter.
A typical biodegradable, solvent-based cleaner can be
useful for the tough cases where material is not easily
dislodged (e.g. Krud Kutter; www.krudkutter.com).
Soaking the nozzles in this type of solvent is a benign
cleaning method that does not introduce defects in the
edges of the nozzles.
If one or more nozzles need cleaning, the “as left” data
will differ from the “as found” data. The nature of the
change in the data can be instructive to the user as to
steps to take to improve his or her own methods of
keeping nozzles clean between periodic inspections.
Flow resistance
As an additional means of finding unexpected, non-ideal
conditions, it is valuable to examine the flow resistance of
each impactor stage. Measuring flow resistance requires
measuring the pressure drop at a given, controlled flow
rate. Milhomme et al.5and Shelton et al.6have described
this method of examining impactor nozzles. Clean, per-
fect nozzles have a flow resistance somewhat greater than
that predicted by the Bernoulli relationship:
⌬P⫽1
V2
2
The more imperfect the nozzles are, the greater the devi-
ation from this simple relationship. In particular, defects
that are not in the nozzle exit plane will cause greater
deviations since the optical systems focus on the nozzle
(
(
c-MSPart_7-11_Jasper Art5.0 #11 11/17/11 1:14 PM Page 9
exit plane. The type of imperfection shown in Figure 2c
is particularly difficult to find because much of the conta-
mination is not in the plane of optical focus. Flow resis-
tance measurements assist the inspector in detecting this
kind of contamination.
Examples of impactor histories
Little data have been presented in the open literature on
the experience of inspecting used impactors. Two
examples are presented here to indicate the types of
behavior that can be observed, if the approach
described above is implemented by the inspector.
In the first example, we follow the time history of a par-
ticular NGI stage 6. The impactor stage starts new with
the effective diameter exactly nominal (0.323 mm; Tables
1 and 2). At its first inspection, the nozzles had become
reduced by some form of contaminant build up.
It was found that the practices of the user were contami-
nating the nozzles with a material that was not easy to
remove. Whereas the nozzles on stage 6 were relatively
nominal at the outset, when the impactor came for its first
inspection, the average nozzle diameter was, in fact, too
small to be within acceptable limits. Upon ultrasonic
cleaning, the nozzles were opened to bring the average
diameter back to 0.318 mm. This experience has been
repeating itself. The user has been informed and is not
able to identify anything specific that he can change to
keep the nozzles more open. One option for this user is
more frequent inspection and cleaning. However, during
the most recent cleaning, the solvent soaking method and
ultrasonification were used and these steps resulted in a
cleaner ‘as left’ condition. This user is repeatedly coached
about keeping the nozzles within specifications because it
can take time and effort to reach a satisfactory, long-term
performance.
In a second example, even with the best of information,
users do not necessarily take appropriate steps. In this
example of a stage 4 of an NGI, the nozzles start rather
close to the nominal size but then consistently and slowly
decrease in size (Tables 3 and 4). A prudent step would
have been to clean the nozzles after the ‘as found’ data in
October 2010 showed that the nozzles were near the
lower limit of 1.197 mm. The user appreciated the warn-
ing and chose to do the cleaning himself. It is hoped that
when the impactor returns for inspection in late 2011, the
nozzles are in satisfactory condition.
A few simple rules
Observing a few simple rules in the periodic inspection
of used impactors can result in a high degree of confi-
dence when the impactor is placed back into service:
1) Use an optical inspection tool whose focus, calibra-
tion and edge-definition are independent of human
bias.
10 DECEMBER 2011 Inhalation
Tabl e 1
History of an NGI stage 6, NGI-0098
New Apr 2010 Nov 2010 Jun 2011
As found N/A 0.311 mm 0.312 mm 0.311 mm
As left 0.323 mm 0.318 mm 0.318 mm 0.321 mm
Tabl e 2
Distribution of nozzle sizes for NGI stage 6 (as left),
NGI-0098
Size of nozzles New Apr 2010 Nov 2010 Jun 2011
Smaller than 3 17 10 2
0.313 mm
0.313 mm to 77 260 272 91
0.320 mm
0.321 mm to 187 119 114 303
0.325 mm
0.326 mm to 129 0 0 0
0.333 mm
Larger than 0 0 0 0
0.333 mm
Total 396 396 396 396
Tabl e 4
Distribution of nozzle sizes for NGI Stage 4 (as left),
NGI-0407
Size of nozzles New Sep 2008 Sep 2009 Oct 2010
Smaller than 0 0 3 14
1.197 mm
1.197 mm to 0 35 46 36
1.204 mm
1.205 mm to 50 16 3 2
1.209 mm
1.210 mm 2 1 0 0
to 1.217 mm
Larger than 0 0 0 0
1.217 mm
Total 52 52 52 52
Tabl e 3
History of an NGI stage 4, NGI-0407
New Sep 2008 Sep 2009 Oct 2010
As found N/A 1.204 mm 1.201 mm 1.198 mm
As feft 1.208 mm 1.204 mm 1.201 mm 1.198 mm
c-MSPart_7-11_Jasper Art5.0 #11 11/17/11 1:14 PM Page 10
Inhalation DECEMBER 2011 11
2) Determine the effective diameter, consisting of the
edge-based diameter and the area-based diameter, when
assessing suitable aerodynamic performance.
3) Determine the ‘as-found’ condition of the impactor
nozzles and, if cleaning is required, the ‘as-left’ condi-
tion, enabling the user to assess the effectiveness of his
or her routine impactor maintenance practices.
4) Measure flow resistance to determine if material outside
of the optical focus plane is affecting nozzle performance.
By following these rules, which are based on the funda-
mentals of the impactor performance, one can proceed
with confidence.
Refer ences
1. Chambers, F., J. Mitchell, C. Shelton, A. Ali,
“Assessment of the Accuracy and Precision of Optical
Measurement Systems for Mensuration of Cascade
Impactors,” Drug Delivery to the Lungs 20, The
Aerosol Society, December 9-11, 2009, Edinburgh,
Scotland, pg. 17-20.
2. Mitchell, J. P., “Good Practices of Qualifying Cascade
Impactors (CIs): A Survey of Members of the Euro -
pean Pharmaceutical Aerosol Group (EPAG),” Drug
Delivery to the Lungs 16, The Aerosol Society,
December 8-9, 2005, Edinburgh, Scotland, pg. 189-92.
3. Chambers, F., A. Ali, J. Mitchell, C. Shelton, S.
Nichols, “Cascade Impactor (CI) Mensuration,” AAPS
Pharm. Sci. Tech., 11, 472-84 (2010).
4. Roberts, D. L., “Theory of Multi-Nozzle Impactor
Stages and the Interpretation of Stage Mensuration
Data,” Aerosol Sci. Tech., 43, 1119-29 (2009).
5. Milhomme, K., C. Dunbar, C. Lavarreda, D. L.
Roberts, F. J. Romay, “Measuring Changes in the
Effective Jet Diameter of Cascade Impactor Stages with
the Flow Resistance Monitor,” Respiratory Drug
Delivery 2006, page 405-8, Boca Raton, FL, April 23-
27, 2006.
6. Shelton, C. M., D. L., Roberts, “Application of Flow
Resistance Measurements to Cascade Impactor Quality
Control,” Drug Delivery to the Lungs 16, Edinburgh,
UK, December 7–9, 2005.
Dar yl L. Ro b e rts, Ph .D. is Vic e Pre side n t and Ch ristian
A. Lavarre d a is Pro du c t Man ag e r, Ph arm a an d Bio te c h,
MSP Co rpo ratio n, 5910 Ric e Cre ek Parkw ay, Suite 300,
Sho re v iew, MN, 55126, US, Tel: +1 651 287-8110,
dro be rts@m s p c orp.c o m. Web s ite: w w w.m sp c orp.c o m
www.rddonline.com
JW MARRIOTT DESERT RIDGE
RESORT & SPA
PHOENIX, ARIZONA
MAY 13 - 17, 2012
2012 ARIZONA
CALL FOR PAPERS
Poster Deadline
U January 13, 2012
MARK YOUR CALENDARS AND PLAN TO ATTEND
c-MSPart_7-11_Jasper Art5.0 #11 11/17/11 1:14 PM Page 11
