United States
 Environmental Protection
 Environmental Monitoring Systems
 Research Triangle Park NC 27711
 Research and Development
EPA-600/S4-82-040  Sept. 1982
 Project Summary
Wide  Range
Aerosol  Classifier

Dale A. Lundgren and David C. Rovell-Rixx
  The purpose of this project was to
 design, construct, calibrate, and field
 test a mobile  ambient particulate
 matter sampler  (Wide Range Aerosol
 Classifier) to collect size-classified
 samples of large aerosol particles. The
 sampler design was based on a similar
 stationary sampling system previously
 constructed  by  the Principal Investi-
 gator, Dr. Dale Lundgren.
  The sampler is fitted into a trailer
 and consists of a large, high flow rate
 inlet from which five isokinetic samples
 are withdrawn.  Four of the samples
 are passed  through single-stage im-
 pactors  with different  cutpoints and
 the fifth is  passed through a  total
 particulate   matter filter. The  four
 impactors were designed  to  collect
 particles  greater than  7.5 mm, 15
 mm, 30  mm and 60 mm diameter.
 Aerosol particles smaller than 7.5//m
 are sized by using separate lower flow
 rate cascade impactors following the
 last single stage impactor.
  An accompanying analysis lab was
 set up in a mobile van. Analysis
 equipment  includes an analytical
 balance  and a  sample  equilibration
  The mobile sampler was briefly field
 tested in Gainesville, Florida.
  The full report was submitted by the
 Department of Environmental  Engin-
 eering, University of Florida, in fulfill-
 ment of U.S. Environmental Protection
 Agency grant number R-806714-
 010.  The report  covers  the total
 project period from October 1, 1979
to November 30, 1980.
  This Project Summary was devel-
 oped by EPA's Environmental Moni-
 toring Systems Laboratory, Research
 Triangle Park, NC. to announce key
 findings of the research project that is
 fully documented in a separate report
 of the same title (see Project Report
 ordering information at back).

  The present air quality standard for
 particulate matter is based upon the
 total amount of suspended particulate
 matter collected by the  reference
 method, a high volume air sampler. This
 sampler has been assumed to collect
 particles less than 100 /urn diameter
 (Stokes equivalent) Tests conducted by
 Wedding and  co-workers indicate,
 however, that the sampling efficiency of
 the high volume air sampler may be as
 low as 7 percent for 50fjm particles and
 18  percent for 30 fjm particles in
 moderate winds The high  volume air
 sampler is also sensitive to orientation,
 showing a 20 percent drop in collection
 efficiency for particles 1 5/urn in diameter
 and larger with a 45 shift in  wind
 direction. Because the mass of a particle
 increases as the cube of its diameters,
 the  mass concentrations measured by
 the  high volume air sampler can also
 vary widely.
  The U.S. Environmental  Protection
Agency (EPA) in 1979 defined Inhalable
 Particles (IP) as particles less than 15
fjm  in diameter and is  currently con-
 sidering an upper size limit of 10 /urn.
The EPA has also considered establish-
ing a fine particle standard consisting of
particles less than ~2 to 3/jm diameter.
Consideration  of an IP standard  has
generated considerable interest in
defining the total  atmospheric particle

size  distribution.  Only then will  it  be
possible to determine  what fraction
of the total atmospheric aerosol is being
collected and what fraction would  be
desirable to collect by existing  or pro-
posed sampling devices for mhalable or
other paniculate measurements
  Lundgren  and  Paulus1  previously
described a stationary sampling  system
which effectively sampled large atmo-
spheric particles ( up to 100 /urn), and
determined  the large particle  size
distribution and total atmospheric mass
concentration; they compared  results
with collections by dust fall plates and
with a standard high volume air sampler.
That study provided the background  for
the present project to design, construct,
and  field test a mobile large particle
sampler  which determines  the mass
distribution of large atmospheric part-
icles. With this sampler it is possible to
characterize the total particle mass size
distribution of ambient aerosol  and to
compare the results with particle mass
data collected simultaneously  by the
Total Suspended  Particulate (TSP) Hi-
Vol, the Size Selective Inlet Hi-Vol (SSI),
small particle cascade impactors, and
the dichotomous samplers.  The sam-
pling system will be especially useful for
evaluating areas  with high fugitive
particle concentration.  Data showing
the relationship between TSP, IP, FP and
other measures of Particle Concentration
can  be obtained while the total atmo-
spheric particle mass distribution is
being measured.  Differences  in the
quantity of particulate mass measured
by various devices canthen be rationally
explained and properly  related to the
aerosol  size  distribution in the atmo-
  The objective of this project  was to
design and construct a mobile particle
sampler  capable  of  collecting size-
separated mass samples providing
mass distribution  of the total  atmos-
pheric particle size range up to 200/jm
aerodynamic diameter.

  The sampler is fitted into a trailer and
consists of a large, high flow rate (1380
CFM) inlet from which  five isokinetic
samples  are withdrawn.  Four  of the
samples are passed  through single-
stage impactors with different cutpoints
and  the fifth  is passed through  a  total
particulate matter filter.  The  four
impactors are  designed to collect
particles greater than 7.5 mm, 15 mm,
30 mm, and 60 mm diameter, respec-
tively.  Smaller aerosol particles are
 sized by attaching cascade impactors
 following the single-stage impactors.
   An accompanying analysis lab is set
 up in a  mobile van. Analysis equipment
 includes a  precision balance, optical
 sizing microscope, and  a sample equi-
 libration chamber.

 Results and Discussion

 Inlet Design.

  The major task  in measuring large
atmospheric particles is transporting a
true or  representative sample of  the
particles into a  measurement device.
This difficulty is particularly serious for
the collection  of particles larger than
30 /um diameter because of their great
inertia and high settling rate.
  Particle  settling  velocity  (Vs) is
determined by a balance between the
force of gravity acting on a particle  and
the  fluid  drag force exerted by  the
medium through  which the  particle
falls. Simply put, settling velocity
increases rapidly with increasing particle
size. Sampling tube inlets which  are
pointed  upward will  capture a greater
porportion of large particles than are in
the  actual  distribution due to  the
settling  of large particles into the inlet.
The increase in the relative number of
large particles sampled is equal to 1 +
[Vs/Vo], where Vo equals the sampling
tube inlet velocity. The error can be kept
reasonably small if the inlet velocity (Vo)
is made several times greater than the
settling velocity  (Vs)  of the largest
particle desired. Therefore, a Vo of 25
times Vs would give a permissible error
of only  4 percent  overestimation for a
tube pointed upward. The Wide Range
Aerosol Classifier  meets this criterion
for a 55 /L/m diameter particle, and an 11
percent overestimation  of a  100-yum
diameter particle is predicted.
  Particle inertia is a function of particle
mass. From Newton's first law, where
force equals mass times acceleration, a
larger force is required to accelerate (or
decelerate) a larger (heavier) particle as
quickly  as a smaller (lighter) particle.
Because the  size of a  particle  also
effects the drag it experiences from the
air, the relaxation time (T) is used as an
indication of its ability to accelerate or
decelerate. Particle relaxation  time (T)
has  been defined  as:  T = Dp2p/18r;,
where  Dp  is  the diameter of  the
particles, p is its density, 77 is the viscos-
ity of air, and 7 has units of time.
  It is also convenient to define particle
stopping distance  /, as the distance a
particle  will travel  when decelerated in
a  fluid medium  (air)  from an  initial
velocity (V) to rest. It has been shown
   Davies2 uses/to evaluate the effect of
particle inertia on sampling efficiency
for two situations. When sampling at a
known  inlet  flow rate,  Q, a suction
velocity (Vx)  will  be produced; Vx ;s a
function of the distance  from  the inlet
orifice To examine the case where Vx is
evaluated at a distance /from the center
of the orifice, Davies uses the equation
Vx = Q/477-/2,  where Q is the flow rate of
the sampler inlet and 4nP is the surface
area of a sphere of radius /
  More appropriate might  be an equation
determined by Dalle Valle3from measured
velocity contours  of exhaust hoods For
a point at a distance / along the center
line from the hood face, Vx =Q/10/2+A,
where A  equals  the  inlet face  area
Davies reasons that if the radius of the
inlet (R) is several times larger than /,
the effects of inertia will  be negligible.
  Davies also examined the situation of
sampling  in  a crosswind  with  some
velocity (Vs) and reasons that the inlet
radius (R) should  be much greater than
the stopping  distance  associated with
the wind (i.e., /= Vw T). He suggeststhat
R  should be  at least 5/. For the mobile
sampler,  the condition  that R  = 51
requires that / be less  than 6 cm. At a
distance of 6 cm  from  the inlet orifice,
the maximum velocity  using the latter
equation would be 2 m/s (~ 4.5 mph).
The stopping distance for a  100-/um
diameter particle at this velocity is 6 15
cm. This nearly meets Davies' criteria
and suggests that the  mobile  sampler
inlet  is dimensionally  adequate to
efficiently sample 100  //m diameter
particles in winds up  to ~2 m/s (4.5
  Several researchers  have noted that
Davies' theoretical criteria seem overly
restrictive for efficient sampling and are
not met by several commercially avail-
able sampling instruments. A  theoret-
ical study of sampling efficiencies by
Agarwal and  Liu4 took into account both
particle  inertia  and  settling.  They
determined  the  flow  field around a
vertical  thin-walled  inlet from the
Navier-Stokes equations  and then cal-
culated particle trajectories. They deter-
mined a critical particle trajectory which
is  a distance, Re, from the inlet axis,
such  that those  particles inside this
radius enter the  inlet.  Sampling ef-
ficiency was calculated  by comparing
the concentration of particles entering
the inlet versus the actual concentra-
tion of the air sampled.

   Agarwal and  Liu  noted that the
 sampling  efficiency  was  a  function
 independently  of both  the Stokes
 number (STK =pCDp2Vo/9^D) and the
 relative settling velocity (Vs' = Vs/Vo).
 The characteristic length  and velocity
 are the inlet diameters, D,  and the inlet
 velocity, Vo. Their research  indicates
 that if the product (STK) (Vs') is less than
 0.10, then sampling efficiency will be
 greater than 90 percent. Agarwal and
 Liu caution, however, that their criterion
 for efficiency regards all particles which
 enter the inlet as sampled, whether or
 not they impact a nd possibly stick on the
 inside wall of the inlet. The value of (STK)
 (Vs')for sampling  100/urn particles with
 the mobile sampler is 0.025. Solving for
 particle diameter with (Stk) (Vs'( equal to
 0.10 indicates that the mobile sampler
 can sample particles as large as 130^m
 diameter  with  an  accuracy  of 90
 percent These values assume, however,
 that Stokes law holds for particles this
  The shroud  of the  Wide  Range
Aerosol  Classifier was designed to
provide a calm air space around the inlet
orifice so that the sampler would be less
sensitive to cross winds. To be effective,
it  was  made  large  enough  so that
particles would not impact on it but would
flaw around and  over the  shroud  To
insure this, it was designed  large in
comparison to /. Using the same ratio of
1:5, this criterion was met for a particle
with a stopping distance less than or
equal to 1 /5 the diameter of the shroud,
or 30 cm.  This condition was met for
100-/um diameter particle in a 9 m/s (20
mph) wind.
  A rain shield designed for use when
the sampler is operating consists  of a
90-cm flat disk supported 30 cm above
the top of the shroud  and centered
above the inlet. The shield prevents rain
from falling directly into the inlet under
light wind conditions. A winddriven ram
should fall at an angle and impact on the
side of the tube From there it should run
down the side and into the air plenum
box and not into the samplers

Selective Sampler Design
  The size-selective sampler inlets
were designed to  extract an isokinetic
sample from the air flowing through the
inlet tube. The samplers were  con-
structed of 0.10  cm  aluminum.  The
sampler inlets are of equal  area so that
an equal  volume of air  is sampled
through each sampler. Each impactor
inlet measures 17.8 cm * 6.55 cm. The
inlet area was determined  by the slot
width required  for the first  impactor;
therefore, the Number 1 impactor is a
straight nozzle (neither converging nor
  To minimize losses and  facilitate
construction, a  rectangular jet design
was used. The impactors were designed
with the jets 17.8 cm long in order to
leave more room for the airflow over the
impactor plate  to turn down into the
filter. This reduces the particle losses on
the walls of the impactor.
  The fifth sampler,  which  collects a
total particle sample, was designed with
a square inlet so that it would fit better
between the impactor inlets.  The sides
of all  sampler  inlets are straight to
reduce any  loss of particles  onto the
walls. Inlet area of the fifth sampler is
the same as for the other samplers.
  An additional  advantage of using high
volume air sampler blowers is that flow
controllers are readily available on the
market. These flow controllers insure a
constant flow rate through the sample
so  that the total air volume for the
sampling  period is known and that a
representative sample is drawn for each
hour over the  normal  24-hr period.
Therefore, if the total sampling time is
known and the  flow rate is known and
constant, the total air volume sampled is
known  Aerosol mass concentrations
can then be computed. A constant flow
rate is additionally critical for  impactors
because the aerosol impaction efficiency
is a function  of the flow rate.  Field
sampling and  evaluation at selected
sites has shown that the sampler can
provide reliable information on the size
distribution  of particle mass including
particles up to 100 /urn  aerodynamic
diameter. The  system is self sufficient
with the accompanying mobile labora-
tory for processing  and analyzing the
collected samples.

Conclusions and
  The  results  of  limited  field testing
indicate that the Wide Range Aerosol
Classifier (WRAC) can be easily set up at
a suitable field location and operated to
collect valid  24-hr samples of  size-
separated  atmospheric aerosols.  The
accompanying  analysis van can be used
to prepare, condition, and weigh the
collected aerosol samples. The samples
can  be used  to determine  an aero-
dynamic particle size distribution based
upon the mass of the ambient aerosol.
  Calibration  tests  indicate that the
classification is adequate to  provide a
good characterization of the atmospheric


1. Lundgren, D.A. andH.J. Paulus. The
   mass distribution  of  large atmo-
   spheric particles. J. Air Poll. Cont.
   Assoc.,  25:1227-1231, 1975.
2. Davies,  C.  N. The entry of aerosols
   into sampling tubes and heads. Brit.
   J. Appl. Phys., Ser 2, 1:921-932,
3. Dalle Valle, J. M.  Exhaust Hoods.
   New York: Industrial Press. 1952.
4. Agarwal, J  K., and B. Y.  H. Liu. A
   criterion for accurate aerosol sam-
   pling in calm air. Am. Indust.  Hyg.
   Assoc. J., 41:191-197  1980.
  Dale A. Lundgren and David C. Rovell-Rixx are with the University of Florida,
     Gainesville, FL 32601.
  Robert M. Burton is the EPA Project Officer (see below).
  The complete report, entitled "Wide Range Aerosol Classifier," (Order No. PB
     82-256 264; Cost: $9.00. subject to change) will be available only from:
          National Technical Information Service
          5285 Port Royal Road
          Springfield, VA 22161
          Telephone: 703-487-4650
  The EPA Project Officer can be contacted at:
          Environmental Monitoring Systems Laboratory
          U.S. Environmental Protection Agency
          Research Triangle Park, NC 27711
                                                                                     * U.S GOVERNMENT PRINTING OFFICE 198Z-559-017/0837

United States
Environmental Protection
Center for Environmental Research
Cincinnati OH 45268
Postage and
Fees Paid
EPA 335
Official Business
Penalty for Private Use S300
                                        PS   0000329
                                        U S  tNVlR PROTbCIIUN
                                        KE&10N  5  LIBRAKY
                                        230  S  DEARBORN  STRtET
                                        CHICAGO  IL 60604