&EPA
           United States
           Environmental Protection
           Agency
            Environmental Sciences Research
            Laboratory
            Research Triangle Park NC 27711
EPA-600/2-78-112
June 1978
           Research and Development
Dichotomous
Sampler -- A
Practical  Approach
to Aerosol
Fractionation  and
Collection

-------
                RESEARCH REPORTING SERIES

Research reports of the Office of Research and Development. U S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:

      1.  Environmental Health  Effects Research
      2.  Environmental Protection Technology
      3.  Ecological Research
      4.  Environmental Monitoring
      5.  Socioeconomic Environmental  Studies
      6.  Scientific and Technical Assessment Reports (STAR)
      7.  Interagency Energy-Environment Research and Development
      8.  "Special" Reports
      9.  Miscellaneous Reports

This report has  been assigned  to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.

-------
                                              EPA-600/2-78-112
                                              June 1978
  DICHOTOMOUS SAMPLER - A PRACTICAL APPROACH
    TO AEROSOL FRACTIONATION AND COLLECTION
                      by

    Robert K. Stevens and Thomas G. Dzubay
  Atmospheric Chemistry and Physics Division
   Environmental Sciences Research Laboratory
 Research Triangle Park, North Carolina  27711
   ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
      OFFICE OF RESEARCH AND DEVELOPMENT
     U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA  27711

-------
                                  DISCLAIMER

     This report has been reviewed by the Environmental Sciences Research
Laboratory, U.S. Environmental Protection Agency, and approved for publica-
tion.  Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
                                      ii

-------
                                   ABSTRACT

     Procedures to size fractionate, collect, and analyze ambient concentrations
of particulate matter are described.  Emphasis is placed on the design and
characteristics of the single-stage dichotomous sampler.  A new inlet is de-
scribed that samples aerosol independent of wind speed and direction, and a
discussion of the advantages of a new pneumatic flow control system is included.
Comparative results of the high-volume and dichotomous sampler are presented.

     This report covers a period from January, 1975 to January, 1978, and work
was completed as of January, 1978.

-------
                                ACKNOWLEDGMENTS

     The authors are indebted to Dr. B. W. Loo of Lawrence Berkeley Laboratory
for providing the design shown in Figure 1, and to Dr. Andrew McFarland of Texas
A & M University for providing the data and drawings shown in Figures 2 and 3.
Also the authors acknowledge the assistance of Dr. George Russwurm and Dr.
Dwight Rickle of Northrop Services, Inc. for providing the chemical analyses
cited in this report.  In addition, we thank the West Virginia Air Pollution
Control Board for operating the dichotomous sampler from which some of the
data in this report were derived.  We acknowledge Dr. S. Freeman and Joseph
Nader of System Science, Inc., Chapel Hill, North Carolina for assistance in
preparation of Figures 11, 12, 13, and 14 and interpretation of data related
to measurements made in St. Louis, Missouri.
                                      IV

-------
                                   SECTION 1

                                 INTRODUCTION

     In 1972 the U.S. Environmental Protection Agency launched a program to
develop reliable procedures to separate and collect aerosols in two size
fractions  (<3.5 and >3.5 micrometers  (ym) aerodynamic mean diameter).  The
size fractionation system was designed to collect the aerosols on inert
surfaces so that the mass and chemical composition could be measured with a
minimum of artifact formation.

     Manual (1,2,3) and automated virtual (4,5) dichotomous samplers were
developed and for the past 4 years have undergone extensive field testing and
wind tunnel studies to characterize their aerosol sampling characteristics.
Presently, these samplers are being used in 10 separate field studies across
the United States to obtain mass, sulfate, and elemental composition data;
these data are being compared with aerosol data collected with the high-volume
sampler.  The results of these field and wind tunnel tests, as well as the
advantages of the dichotomous sampler for characterizing the aerosol content
of the atmosphere, are discussed.

-------
                                   SECTION 2

                                  DISCUSSION

VIRTUAL IMPACTION PRINCIPLE

     Impactors that separate aerosol particles depend upon the relative balance
between inertial and aerodynamic forces.  In the conventional aerosol impactor,
an airstream turns abruptly as it approaches a flat plate.  Particles with the
largest inertia tend to maintain a straight trajectory and impact on the plate
while the viscous drag forces of the gasflow carry the smaller particles
along airflow streamlines.  The particles collected on the plate would ideally
consist of all sizes above a well defined cut-off diameter.  To make a con-
ventional impactor a quantitative particle collector, the impaction surface is
coated with a layer of grease to minimize particle bounce errors  (6,7).
Unfortunately, the grease can interfere with some of the required chemical
analyses.  Also, the grease may become completely covered with particles and
thus reduce the effectiveness of the impactor as a quantitative collector.

     In the virtual impaction collection method, instead of having the larger
particles collect on grease-coated plates, they are impacted into a slowly
pumped void and collected on a filter.  Since a small constant fraction of the
total flow is being pulled through the void, the portion of fine particles
contained in the large particle fraction is directly related to the ratio of
the inlet flow and the flow rate in the large particle collector.  Figure 1
shows a cross-sectional view of a virtual impactor recently designed by Loo
(private communication, B. W. Loo, 1978).  It operates at an inlet flow rate
of 17 liters/minute and has a 50 percent cut point at 3.5 ym.  The shapes and
sizes of the orifices were chosen to minimize aerosol losses.  Measurements by
Loo have shown that the losses are essentially zero in the 1 to 2.5 and 4.5 to

-------
20 pm ranges, and the loss curve peaks at 5 percent  for particles in the

vicinity of  3.5  ym.   Such low losses were measured for liquid particles and

represent the worst  case.   For solid particles, which have  a tendency to

bounce, the  wall losses are even lower (5).
                                 INLET
                                 171pm
                                TO COARSE
                                PARTICLE
                                 FILTER
                                  1.7 Ipm
 TO FINE
PARTICLE
 FILTER
 15.3 Ipm
             Figure  1.   Cross-sectional view of virtual  impactor.

                                        3

-------
     For the virtual impactor shown in Figure 1, 10 percent of the sampled air
passes through the coarse particle filter.  Thus 10 percent of the fine
particle mass is collected on the coarse particle filter.  A correction of the
coarse particle concentration is made at the time of sample analysis and is
based on the measured fine particle concentration.  Equations for making this
correction are given by Dzubay and Stevens  (3).

     In previous work, investigators operated virtual impactors with only 2 to
5 percent of the sampled air passing through the coarse particle filter
(1,2,3,4,5).  To obtain this lower flow fraction, and yet maintain adequately
low losses, it was necessary to use two virtual impactor stages in series.  An
obvious disadvantage of this approach is that the sampler is more expensive and
complicated to fabricate.  For a virtual impactor operating at 14 liters/minute,
the two jets of the second stage have diameters of only 0.173 cm (0.068 in) and
0.234 cm (0.092 in) and the jets have a tendency to become clogged after a few
months of operation.  In contrast, the jets of the single-stage virtual im-
pactor shown in Figure 1 have jet diameters of 0.386 cm  (0.152 in) and 0.518 cm
(0.204 in) and consequently have a much lower tendency to become clogged.  Thus
the single-stage virtual impactor costs less to build and provides more reli-
able results.  Although single-stage virtual impactors were first described by
Houman and Sherwood (8) in 1965 and by Conner (9) in 1966, these early devices
were not optimized for low particle loss as is Loo's new device (Figure 1).

     During the initial phases of development of the virtual impactor, it was
decided that dividing and collecting aerosol into two size ranges (hence the
name dichotomous sampler) would provide the optimum amount of information to
distinguish primary from secondary pollution sources.  Separating the particles
into more than two fractions increases the complexity of the sampler, reduces
the amount of aerosol per stage, and increases the opportunity for losses of
particles in the sampling train.

     The main advantages of the virtual impactor over conventional aerosol
size-fractionation designs are:

-------
         Particle bounce and reetrainment problems associated with ungreased
         impactors and the chemical interferences caused by the required
         grease coating are eliminated.

         Sufficient quantities of aerosol particles are collected on inert
         surfaces, providing ideal samples for gravimetric, elemental, and
         chemical analysis.

         Aerosols are uniformly deposited on filters, providing ideal samples
         for X-ray fluorescence  (XRF) elemental analyses and 3-gauge mass
         measurements.
COLLECTION SURFACES


     High-porosity Teflon with l-pm pores is the most suitable filter medium

for use in dichotomous samplers.  This filter medium is preferred for the

followed reasons:


     •   Collection efficiency for particles above 0.01 urn greater than 99
         percent  (10).

     •   Extremely stable mass for high gravimetric accuracy.

     •   Negligible tendency to absorb or react with gases.

     •   Minimal impurities to interfere with analyses for chemical and
         elemental species.

     •   Low mass per unit area (desirable for gravimetric, XRF, and 6-gauge
         measurements).


     In addition, Teflon filters are ideal for the collection of sulfuric acid

since there is little if any interaction with the filter medium.  In laboratory

experiments, 1 to 100 micrograms (pg) of 0.3-ym droplets of sulfuric acid

were deposited on Teflon filters, and  > 90 percent of the  sulfate and equi-

valent acidity were recovered.


     For gravimetric determination of collected aerosol mass, Teflon collection

surfaces are significantly less affected by changes in relative humidity than

glass fiber, quartz, and cellulose ester collection surfaces.

-------
     For aerosol sampling, the Teflon filters must be supported or mounted in
a manner to provide adequate structural strength.  In previous studies, we
used 37-urn-diameter Fluoropore   (Millipore Corporation) filters, which consist
of a Teflon membrane bonded to a polyethylene net for support.  However, this
type of filter tends to curl badly after sampling.  Recently, a Teflon membrane
filter supported by a thin annular polyester ring became available from Ghia
Corporation, Pleasonton, California; this filter has superior handling char-
acteristics.  Because the filter is not partly obstructed by a support net, it
has improved flow and loading properties.  The lower mass per unit area enables
better detection limits to be achieved for elemental analysis by XRF.

     Investigations are in progress on the collection properties of a newly
developed high porosity (>70 percent) Teflon filter with pores in the 2 to 5 ym
range.  Such filters have much lower flow resistance and should therefore be
capable of withstanding longer sampling periods and achieving high particulate
loadings without the problem of clogging.

INLET DESIGN

     Conventional aerosol samplers have a variety of aerosol intake designs,
ranging from the gabled roof used in the high-volume sampler to circular "hat"
designs.  In these conventional designs, the intake sampling efficiency varies
with wind speed and direction.  Davies  (11,12) has investigated the effects
that particle inertia and gravitational settling have upon the performance of
aerosol inlets and found that quantitative sampling becomes increasingly more
difficult to achieve as the particle size increases above 10 to 20 urn.  Our
inlet was, therefore, designed to reject particles above 20 um.  Although the
high-volume sampler does collect particles larger than 20 pm, the efficiency
for collecting such particles depends upon the wind speed and wind direction
relative to the orientation of the gabled roof of the sampler.

     Under EPA Grant 804190, Andrew McFarland at Texas ASM University has
developed an inlet, shown in Figure 2, that is compatible with the dichotomous
sampler.  McFarland has shown through wind tunnel tests that the particle

-------
sampling efficiency of this inlet remains relatively constant  at wind speeds
between 4 and  8  kilometers per hour (km/hr).  Figure 3 shows inlet particle
penetration as a function of particle size at a wind speed of  4  km/hr.   Wind
tunnel tests are continuing at Texas A & M to further characterize the aerosol
sampling efficiency of this new inlet.
         SCALE, in  SCALE, cm  :'•
             OurO
                -10
               FLOW
                                                             SHIELD
                                          FLOW TO OICHOTOMOUS SAMPLER
               Figure  2.   Aerosol inlet for dichotomous sampler.

-------
                     100
                   a  80
                   O)
                   CO
                   £  60
                      40
                      20
                   
-------
only a few  inexpensive  but highly reliable components are needed, this approach
provides the  ruggedness,  reliability,  and low cost needed for an aerosol
sampler designed  for  routine field use.   McFarland tested the differential
flow control  system at  temperatures from -20 to +40°C, and less than a 5
percent change  in flow  was observed at a flow rate of 14 liters/minute.
Side-by-side  comparisons  of the differential flow control system with a com-
mercial anemometer flow controller incorporated into two identical dichotomous
samplers were performed at Research Triangle Park, North Carolina.  Both
controllers maintained  constant flow to better than 5 percent for pressure
drops across  the  filter of up to 25 cm Hg.
                  iNLET
                 16.7 Ipm
                                      CONSTANT
                                 PRESSURE-DIFFERENTIAL PRESSURE REGULATOR
                                   FLOW REGULATOR         (OPTIONAL)
                        I  FINE
                       S   15 Ipm
                       k«HMBHB>BB
                                               FINE
                                            FLOW VALVE
                     COARSE 1.71pm
              ROTAMETER
                              FILTER
 COARSE
FLOW VALVE
                                       DUAL HEAD
                                     DIAPHRAMPUMP
      Figure 4.   Schematic view of dichotomous sampler with  constant flow
                 rate system which uses a pressure  regulator to maintain a
                 constant pressure differential across a  fixed orifice.

-------
STACKED FILTER SAMPLER

     Stevens and Dzubay  (1) have previously described an alternative type of
dichotomous aerosol fractionating sampler that consists of two filters in
series.  The first is a Uuclepore filter with 12-ym pore diameters, and the
second is a Teflon filter with 1-ym pores.  The 12-ym Nuclepore filter has
been characterized by Parker et al. (13) to show that when operated at the
appropriate flow rate, the collection efficiency curve for particulates with
a specific gravity of 2 approximates the Atmospheric Conference of Govern-
mental Industrial Hygienists (ACGIH) criteria for respirable sampling (14) and
has a 50 percent cut point at an aerodynamic diameter of 3.5 ym.  As the
specific gravity varies, there is, however, a slight variation in the collec-
tion efficiency curve versus aerodynamic diameter.  Tests in ambient air
indicate that the fractionation curve is not affected by particle loading
(13).  One serious difficulty with this sampler is the tendency for particle
bounce errors.  Recent tests conducted in our laboratory have shown that
liquid particles have significantly different collection efficiencies than
solid particles of the same diameter.   By adding additional chemically treated
stages, as shown in Figure 5, HLS and S0» can be collected.  Such a sampler is
capable of simultaneously collecting both gases and particles to
determination of the relationship between gaseous and particulate sulfur (15).
Because of the particle bounce problem, the gas measuring capability of the
tandem filter sampler may be more useful than the particle fractionating
capability.

TYPICAL RESULTS

     Dichotomous samplers of the virtual impaction design have been operated
in a number of geographical areas, and extensive sampling has been conducted
in St. Louis, Missouri, and Charleston, West Virginia.  Figure 6 shows a
comparison of the total (fine plus coarse) mass concentrations determined from
high-volume and dichotomous samplers operated in St. Louis during the summer
of 1975.  in the determination of the mass values, the Fluoropore filters used
in the dichotomous samplers were weighed to a precision of 10 yg with an
electrobalance in a room adjusted to 40 percent relative humidity.  Immediately
                                      10

-------
before the filters were weighed, they were passed in front of a    Po radio-
active source to remove any electrostatic charge.  The glass fiber filters
used in the high-volume samplers were weighed with a mechanical balance.  The
results of this study, shown in Figure 6, indicate the potential of the dicho-
tomous sampler to be used as a substitute for the high-volume sampler for
determining mass concentrations.  However, in dusty locations when turbulent
winds suspend significant amounts of particles larger than 20 vim in the atmo-
sphere, such good agreement between samplers is not expected since the high-
volume sampler collects significantly more of the larger particles.
r











n
^
^
1
I

u




^••^




IKS.
g




n
^^,
\\v
y//
W
^
u




COARSE
PARTICLES
(12/^mNP) I
FINE PARTICLES
(0.2/umNP) i
H2S(AgN030N
CELLULOSE) •!
S02
-------
                   100
                 i
                 gso
                 z
                                       so
                                FINE + COARSE MAS§,^S/m3
100
    Figure  6.   A comparison of total mass concentrations determined using
               high-volume  and dichotomous samplers operated at the St.  Louis
               Botanical  Gardens  between August 18 and September 7, 1975.

     In addition to being  suitable  for  gravimetric  analyses, the Teflon col-
lection surfaces of the dichotomous  sampler are well  suited  for making nonde-
structive elemental and chemical analyses.  An XRF spectrometer is used to
determine elements with atomic numbers  above 13  (2,3,4,16).  The Oionex Ion
Chromatograph  (17) is used to determine a wide variety  of ionic species,
including sulfate, sulfite, and  nitrate.  Using the Brosset  thorin titration,
the Gran titration, and an ion selective electrode, one can  determine SO.,
H , and NH  concentrations (18).  Such  analyses are important  in determining
sulfuric acid,  ammonium sulfate, and ammonium bisulfate in the sample.
     Table 1 shows the average concentrations deduced  from  20  sampling periods
in St. Louis using a dichotomous sampler and an XRF analyzer.   In the fine
particle range, sulfur was the predominant species, and analysis of  the
samples using ESCA  (photoelectron spectroscopy)  (19) revealed  the sulfur  to  be
in the form of sulfate.  The paucity of metals in the  sample,  especially  at
the rural site, indicates that metal sulfate compounds are  of  minor  importance.
The measurements of ammonium and hydrogen ions indicate that the sulfate  is
usually in the form of ammonium sulfate.
                                      12

-------
        TABLE  1.  MASS AND PERCENTAGE COMPOSITION OF  SIZE-FRACTIONATED
                           ST.  LOUIS AEROSOL  SAMPLES
                      FROM AUGUST  18 TO  SEPTEMBER 7,  1975



Si
S
K
Ca
Ti
Fe
Zn
Br
Pb
Urban
Fine, %
29 yg/m3
1.00
12.50
0.40
0.70
1.10
1.40
0.35
0.33
2.20
a
Coarse, %
22 yg/m
8.00
1.40
1.20
8.20
2.00
4.80
0.20
0.16
0.60
Rural
Fine, %
26 yg/m
0.50
12.60
0.30
0.50
<0.10
0.30
0.13
0.06
0.51

Coarse , %
15 yg/m
4.00
0.90
0.90
4.20
0.20
1.30
0.15
0.04
0.11

 Located at the Missouri Botanical Garden in St. Louis.
 Located in an agricultural area in Illinois, 40 km south of St. Louis.

     Table 1 also shows that the sulfur in St. Louis occurs predominantly in
the fine particle fraction.  Figures 7 and 8, which are plots of fine and
coarse particle sulfur and mass fractions collected in Charleston, West
Virginia, also show that the sulfur occurs predominantly in the fine particle
fraction.  This same pattern has also been observed for samples collected in
Los Angeles, Denver, New York, Philadelphia, Milford (Michigan), and Durham
(North Carolina).  The only exception to this pattern was observed near a
fertilizer plant in East St. Louis, Illinois, where the sulfur in the coarse
particle fraction sometimes exceeded that in the fine fraction.  A high cal-
cium content in the sample indicated that the particles consisted of calcium
sulfate  (CaSO ); this was confirmed by microscopic examination.

                                      13

-------
24 -
22 -

20-
18-
10
E ,6-
3 14 _
uf
g
D 10 -
c/>
8-
6 -
4 -

2 -
0-

	 FINE FILTERS
	 COARSE FILTERS













II l











1 ,
i






ii
1








1 ,
i
1






Ii I
T i
















I !
II i








i







1
III
i








|





















i i

















i !•
i








,
n !















in








ii




















!
T 1

<~'~'^<
2 2 Q 0 Q ^A-rc-2? 8 °* 2 ffi °


 Figure  7.   Fine  and coarse particle sulfur measurements for aerosols collected
            in Charleston,  West Va., during 1976 using a dichotomous sampler.
160
150 -
140 -

130 -
120 -
to
-^ MO -
2 100-
o" 90 •
z:
Q 8O -
0 70 -
i
_, 60 •
^f
p 50-
° 40 -

30 -
20 -

10 -
0 -









_ FINE FILTERS
	 COARSE FILTERS
































i
{£ 3-
< <
O O
co co









1.1
HE


ll
III1

i

i















1









1
i
l
1
l
l
l
i
I
i
>_
§
g















|

l!
it










I
1
































i




i

!

I
>- >- z z
11^^
8 S °bATE-












1





It

HI
i
INI
ill!












1

t









1
j
! i


























i




I












1



i
n
jl





































1
H _1 _J J C
A \ * ^ *
cvj — M °°
Figure 8.  Fine and coarse particle mass measurements for aerosols during  1976.
                                      14

-------
     Figures 9 and 10 show a comparison between XRF measurements for sulfur
 (expressed as sulfate) and measurements for sulfate on the same set of fine
particles collected in Charleston, West Virginia using the ion chromatograph
and the thorin titration.  The solid lines illustrate the case of perfect
agreement, and 'the dashed line represents the results of linear regression
analyses.  The closeness of the two lines in Figures 9 and 10 illustrates the
excellent agreement between methods.  For example, in Figure 9 the expression
describing the linear regression is y = 0.983 x - 323.1 with a goodness of
              2
fit value of R  = 0.962.  The corresponding expression for Figure 10 is
y = 0.980 x + 220.6 and R  = 0.978.  This comparison of the sulfur concen-
tration by XRF with sulfate measurements by ion chromatography strongly sug-
gests that 85 percent or more of the sulfur in particles less than 3.5 pm in
aerodynamic diameter is in the form of sulfate.  Since the ammonia concentra-
tion in the same samples was also found to be present in the same equivalent
concentrations as the sulfate, the sulfate is concluded to be in the form of
ammonium sulfate.  The chemical form of sulfur in the fine particle fraction
in Charleston, West Virginia, appears (see above) to be the same as that
observed in aerosols collected and analyzed in St. Louis, Missouri.

     Ten automated dichotomous samplers (ADS) located in the St. Louis,
Missouri, area and part of EPA's Regional Air Monitoring System (RAMS) (20)
have been in continuous operation since March of 1974.  At these same sites
from March 1975 to March 1977, high-volume samplers were operated approxi-
mately every third day.  The fine and coarse fraction of aerosols collected
with the ADS were analyzed for mass and elemental composition (by XRF), and
aerosols collected with the high-volume samplers were analyzed for mass and
sulfate content  (methyl-thymol blue method for sulfate).

     Figure 11 presents time series plots for September through December 1975
for two RAMS stations, 106 (located in Busch Botanical Gardens) and 122
 (located 45 km North of 106), of total sulfate concentrations determined for
aerosols collected by the high-volume and dichotomous samplers.  Although
dichotomous sampler data are available on a daily basis, only those points are
plotted which correspond to days when high-volume data were taken.  The data
are typical of those obtained in St. Louis during this period.

                                      15

-------
       CO
        o
        -  6
        <  5
        u.  -
   DICHOTOMOUS FILTERS  FINE
—     CHARLESTON, W. VA.
      APRIL TO AUGUST, 1976
        t/J
       O
       oc
           1 —
                                    345
                                      -EQSULFATE)X10-3
Figure 9.   A comparison of fine particle  sulfur analyses using an ion  exchange
           liquid chromatograph for sulfate and an XRF spectrometer.
                DiCHOTOfVJQyS FILTERS- F8NE
             _    CHARLESTON. W. VA.
                  APRIL TO AUGUST, 1976
                            2       3       4        5    „   6
                     BHOSSETTHORiN (NANO-EQ SULFATE) X JO 3
    Figure  10.  A comparison of fine particle sulfur 'analyses'using XRF and
               thorin titration methods.
                                    16

-------
     ,§
         S04i Hl-VOLUHE SflnPLES VERSUS DICHOTOdOUS SflMPLES FOR SIHT10N 106
         	 HI VOl IN8I«UnlHI
         — oitHOIonout tunnel
                                                                              8-
      ^48.002^7.00ZJe.OO275.00  2(4.00293.00  302.00  3U.OO320.00  319.00339.00  347.00  SMTOO  385.00°
                                       1979 JUU»» OAT
      g
      5-

      •8
     o •
     »*

     4
         S04i Hl-VOLUHE SflMPLES VERSUS D1CHOTOMOUS SRMPLES FOR STHT10N 122
        	 m-voi INSTItunCM
            2fe7.00  ike. 00  276.00  284.00  293.00  30Z.OO  311.00  32oToO  328.00  338.00  347.00  WB.OO  JM-OO'
                                       1975 JtlLIIMI OUT
     Figure 11.   Comparison of total  sulfate concentrations measured by
                  high-volume and dichotomous samplers at two  St.  Louis sites.
      Inspection of the plots  reveals three salient qualitative  features of  the
data:
          The high-volume sulfate data are  characterized by spikes  in concen-
          tration that are local and show no apparent  correspondence from
          station to station.

          The high-volume sulfate concentrations are on  the average higher than
          those obtained with the dichotomous sampler.   Preliminary analysis
          indicates  that this conclusion still holds even if outliers during
          spike periods are not  included.   This observation is made more ap-
          parent by  inspection of Figure 12 in which plots for the  ratios of
          high-volume sulfate to total dichotomous sulfate are presented for
          the same two stations.   The average ratio, 1.25, at station 106 is
          lower than that of 1.44 obtained  at station  122.

          The data at a single station for  both methods  are fairly  well cor-
          related.
                                         17

-------
     O
     i
         SULFHTE RHT10: HI-VOL OVER DICHOTOtlOUS 106
        9.00  ?S7 .00  1'66.00  215-00  ZU* -00  Z93-00  302 -00  311 .00  320.00  329-00  338.00  347.00  356-00  365 -00 °
                                      1975 JUL.IRN HfiT
    2
    I
              RRtlOi HI-VOL OVCR UICHOTUMUUS
                              \
        e.oo  ST.oo  ?eT.To  pTsToo  STToo  zfo.oosbz.oo311.00  320.00  329.00  339.00  347.00  3&B.00  365.oo"
                                      1975 JUL1HM OUT
     Figure 12.   Ratio of high-volume sulfate to total dichotomous sulfate
                  at two St.  Louis sites.

     Figure 13 shows the  dichotomous sampler data in more  detail in that
simultaneous time series  plots for  the  sulfate found in  both fine and coarse
fractions are presented.   It is clear  from  these data that in the particle
size range below 20 ym the  overwhelming proportion of sulfate is contained  in
the fine fraction  (below  3.5 ym).

     A  brief summary of St.  Louis sulfate concentrations is shown in Table  2
for eight RAMS stations monitoring  sulfate  by both the high-volume and dicho-
tomous  sampler methods.   For a composite of eight stations the time averaged
high-volume sulfate concentration was 10.3  yg/m3 while that for the dichoto-
mous sampler (fine plus coarse fractions) was 7.8 yg/m3.   For every station
the time-averaged concentration measured by the high-volume sampler exceeded
that measured by the dichotomous sampler.   The discrepancy between the two
methods ranged from 23 to 53 percent (less  if outliers are excluded).   Al-
though  no firm explanation of this preliminary result is offered,  either
                                         18

-------
aerosol sampling efficiency variations or  sulfate artifact formation on  the
glass fiber  filters of  the high-volume sampler is suspected.   Note that  if the
measured difference were real,  there would be a significant sulfate contribu-
tion from particles of  greater  than 20-pm  aerodynamic size.   (Such particles
might be unimportant  insofar as human health hazards  are concerned.)  The
dichotomous  sampler data show that on 94 percent of all the days sampled, the
percentage of sulfur  in the fine fraction  (below 3.5  pm) was  no less than 70
percent.
     12-
        SULFUKi OHIir WVtRHGES FOK S1HIIUN 106
          Muf msmo»   a NOmiw. OHIO « BMOH until null
        — comic rune lion  i niitmo unlit • niuiw) «
      °t 9.001*16.00  £63.00  290.00297.0030< .00  311.00  318.00  325.00  332.00  339.00  346.00  3^3.00
                                       1975 JU4.1RN OUT
        SULFURi  DRILY RVERRDES FOR STRTION 122
          MNf IRdCIION   Q NOHnlU. 00IB « DCIOU OCICC! Llnlt
        — coonst rudcnoM  i nicsmo ooio • niesiNO t unot'teuo
                 ris.OO290.00  291.00  304.00  311.00  s'ie.00  32b.~00332.00  339.00  348.00  3&3.00
                                       1919 JIM. ion 001
      Figure  13.   Fine  and coarse  dichotomous sulfur fractions  at two
                   St. Louis sites.
      Table 2 also shows  that the  time-averaged high-volume sampler data  cor-
relate well with the dichotomous  sampler data.   A calculation of the Spearman
rank-order correlation coefficient yields a  value of  0.83 at a  significant
level of >0.99  (21).
                                          19

-------
             TABLE 2.  SUMMARY OF ST. LOUIS SULFATE CONCENTRATIONS
   DETERMINED BY TWO METHODS FOR THE PERIOD SEPTEMBER THROUGH DECEMBER 1975
  RAMS      Average      Average total   Ratio high-volume     Percentage of
 station  high-volume     dichotomous    concentration to        days when
 number  concentration,  concentration,  total dichotomous  fine S/total S>0.70
             yg/m
                 3
yg/m;
concentration
*106
108
112
115
118
120
*122
124
11.0
12.7
10.6
9.6
9.9
9.9
8.9
9.8
8.9
10.0
9.2
7.8
6.6
7.3
6.2
6.4
1.25
1.27
1.23
1.23
1.50
1.35
1.44
1.53
97
79
95
99
83
100
94
99
Average of
the 8 RAMS
stations
Spearman
10.3
rank-order
7.8
correlation coefficient =
1.32
0.83, P<0.01
94


*Stations
presented
in time series plots


     Figure 14 shows  (corresponding) time series plots for total mass col-
lected by high-volume and dichotomous samplers.  Many of the same qualitative
features noted above for the sulfate measurements are also apparent in the
mass data:  spikes in the high-volume measurement and a higher  (in some cases
much higher) value obtained by the high-volume sampler.  The time-averaged
value for the ratio of high-volume mass to total dichotomous mass at station
106 was 1.8.  The corresponding ratio at station 122 was 1.2.  Recall that the
sulfate ratios were higher at station 122 than at station 106.  Further work
is needed to clarify these relationships for both sulfate and mass.
                                      20

-------
         HI-VOLUME  iflHPLES VERSUS DICHOTOMOUS SHHPLES FOR STH'ION  IUB
                           £84.00  C93.00   302.00   311.00  3£0.00   K3.00  3J3.03   3*7.OC   356.00
   MISSi HI-VOLUME SRMPLE5 VERSUS OICHOTOMOUS SHMPLES  FOR STflTION 122
   •  HI-VOl INSIDUWHI
   — OICHOTOROUS SflnPLER
g
S-;
TsToo  21.7.00268.00  275.00  2)4.00   293.00  302.00   311.00  310.00  329.00   336.00  3*47.00  jse.oo   SToo°
                                        1976 JULIDN mi
Figure 14.   Comparison  of  total mass concentrations  measured  by  high-
                volume and  dichotomous samplers at two St.  Louis  sites.
                                              21

-------
                                   SECTION 3

                    CURRENT STATUS OF EQUIPMENT DEVELOPMENT

     Prototypes of manual and automated dichotomous samplers have been thoroughly
tested to establish their reliability as aerosol fractionators.  These prototypes
were expensive to fabricate because of the machining precision required and
the expensive flow controllers used.  The flow control problem has been poten-
tially solved by employing a low-cost differential flow controller, but problems
still exist with the fabrication and assembly of the critical flow components.

     EPA has recently requested Beckman Instruments, Fullerton, California, to
design and fabricate a manual dichotomous sampler available for less than
$1,900 and an automated sampler available for less than $4,500.  In certain
parts where the components are made in two halves, Beckman proposes to use
metallized molded plastics that are bonded together with adhesive.  Since the
concentricity of inlet nozzles and receptor orifices in the impactor stages is
critical, each impactor stage is made of one piece.  Then the assembly of
parts becomes non-critical.  Nozzles and receptors remain aligned and are
easily removed and replaced as units for varying the cut points and for cleaning.
The manual sampler built by Beckman will be designed for direct installation
in exposed locations with no additional housing required.  The instrument will
be housed in a seamless glass-reenforced plastic lay-up.  The housing will
have an access door for convenient filter changing and virtual impactor cleaning.

     In addition to the work of Beckman, Sierra Corporation, Carmel Valley,
California, is currently selling a dichotomous sampler developed by Environmental
Research Corporation (ERC), St. Paul, Minnesota, under EPA Contract 68-02-
1744.  Sierra couples the ERC sampler with their flow controller and special
inlet system and markets the complete aerosol collection device at about
$3,500.
                                      22

-------
     The automated system is similar to the manual sampler.  Common components
include inlets, virtual impactors, flow controllers and air transport, and
external housing.  The sampler-changing mechanism will be adapted from the
sample-changer mechanism developed for EPA by Lawrence Berkeley Laboratory
(4,5).  Magazines containing up to 36 pairs of filters will be installed
through an easy-access service port.  The changing of the filters will proceed
according to instructions from the control module.  The control module will be
composed of a microprocessor and pressure drop sensing element.  When the
pressure across the filter exceeds a pre-set value, the filters will automati-
cally be changed.  In most cases the filters will be changed before the pres-
sure drop exceeds the pre-set level, i.e., once every 24 hours.  During periods
of air stagnation, particulate levels may exceed several hundred micrograms
per cubic meter.  In these instances when the filters begin to clog, the
filters will be changed before the flow decreases by more than a few percent
of the initial setting.

     This new family of aerosol samplers is a marked departure from previous
approaches of aerosol sampling.  Years of prototype testing, evaluation, and
intercomparisons have been performed.  These tests have culminated in the
current design concepts soon to be implemented by several instrument manu-
facturers and should provide the various air pollution control agencies with
a powerful tool to aid in understanding sources and transport of atmospheric
aerosols.
                                      23

-------
                                  REFERENCES

1.   Stevens, R.K., and T.G. Dzubay.  Recent Development in Air Particulate
     Monitoring.  IEEE Trans, on Nucl. Sci., NS-22:849-855, 1975.

2.   Dzubay, T.G., and R.K. Stevens.  Ambient Air Analysis with Dichotomous
     Sampler and X-Ray Fluorescence Spectrometer.  Environ. Sci. Technol.,
     9:633-668, 1975.

3.   Dzubay, T.G., and R.K. Stevens.  Application of the Dichotomous Sampler
     to the Characterization of Ambient Aerosols.  In:  X-Ray Fluorescence
     Analysis of Environmental Samples, T.G. Dzubay, ed.  Ann Arbor Science,
     Ann Arbor, Michigan, 1977.  pp. 95-105.

4.   Goulding, F.S., J.M. Jaklevic, and B.W. Loo.  Fabrication of Monitoring
     System for Determining Mass and Composition of Aerosols as a Function
     of Time.  EPA-650/2-75-048, U.S. Environmental Protection Agency, Research
     Triangle Park, North Carolina, 1975.

5.   Loo, B.W., J.M. Jaklevic, and F.S. Goulding.  Dichotomous Virtual
     Impactors for Large Scale Monitoring of Airborne Particulate Matter.
     In:  Fine Particles, B.Y.H. Liu, ed.  Academic Press, New York, New York,
     1976.  pp. 311-350.

6.   Wesolowski, J.J., W. John, w. Devor, T.A. Cahill, P.J. Feeney, G. Wolfe,
     and R. Flocchini.  Collection Surfaces of Cascade Impactors.  In:  X-Ray
     Fluorescence Analysis of Environmental Samples, T.G. Dzubay, ed.  Ann
     Arbor Science, Ann Arbor, Michigan, 1977.  pp. 121-131.
                                       24

-------
7.   Dzubay, T.G., L.E. Hines, and R.K. Stevens.  Particle Bounce Errors in
     Cascade Impactors.  Atmos. Environ., 10:229-234, 1976.

8.   Houman, R.F., and R.J. Sherwood.  The Cascade Centripeter:  A Device
     for Determining the Concentration and Size Distribution of Aerosols.
     Am. Ind. Hyg. Assoc. J., 26:122-131, 1965.

9.   Connor, W.D.  An Inertial-Type Particle Separator for Collecting Large
     Samples.  J. Air Poll. Con. Assoc., 16:35-38, 1966.

10.  Liu, B.Y.H., and G.A. Kuhlmey.  Efficiencies of Air Sampling Media.
     In:  X-Ray Fluorescence Analysis of Environmental Samples, T.G. Dzubay,
     ed.  Ann Arbor Science, Ann Arbor, Michigan, 1977.  pp. 107-119.

11.  Davies, C.N.  The Aspiration of Heavy Airborne Particles Into a Point
     Sink.  Royal Soc. London Proc., Ser. A, Vol. 279(1378):413-419, June,
     1964.

12.  Davies, C.N.  The Entry of Aerosols Into Sampling Tubes and Heads.  Brit.
     J. Appl. Phys., Ser. 2, Vol. 1 (7):921-932, 1968.

13.  Parker, R.D., G.H. Buzzard, T.G. Dzubay, and J.P. Bell.  A Two Stage
     Respirable Aerosol Sampler Using Nuclepore Filters in Series.  Atmos.
     Environ., 11:617-621, 1977.

14.  ACGIH.  Threshold Limit Values for Chemical Substances and Physical
     Agents in the Workroom Environment with Intended Changes for 1973.
     American Conference of Governmental Industrial Hygienists, P.O. Box 1937,
     Cincinnati, Ohio  45201.

15.  Lorenzen, J.A.  Environmental Monitoring Device for X-Ray Determination
     of Atmospheric Chlorine, Reactive Sulfur, and Sulfur Dioxide.  In:
     Advances in X-Ray Analysis, Vol. 18, W.L. Pickles, C.S. Barrett, J.B.
     Newkirk, and C.O. Ruud, eds.  Plenum Press, New York, New York, 1975.
     pp. 568-578.

                                      25

-------
16.  Jaklevic, J.M., F.S. Goulding, B.V. Jarrett, and J.D. Meng.  Applications
     of X-Ray Fluorescence Techniques to Measure Elemental Composition of
     Particles in the Atmosphere.  In:  Analytical Methods Applied to Air
     Pollution Measurement, R.K. Stevens and W.F. Merger, eds.  Ann Arbor Science,
     Ann Arbor, Michigan, 1974.  pp. 123-146.

17.  Mulik, J., R. Puckett, D. Williams, and E. Sawicki.  Ion Chromatographic
     Analysis of Sulfate and Nitrate in Ambient Aerosols.  Analytical
     Letters, 9:653-663, 1976.

18.  Brosset, and M. Fern.  Man Made Airborne Acidity and Its Determination.
     IVL Publications B-314A, Gothenburg, Sweden, 1976.

19.  Novakov, T., S. G. Chang, R.L. Dod, H. Rosen.  Chemical Characterization
     of Aerosol Species Produced in Heterogeneous Gas-Particle Reactions.
     Lawrence Berkeley Laboratory Report No. LBL-5215, Lawrence Berkeley
     Laboratories, Berkeley, California, 1976.

20.  Myers, R.L., and J.A. Reagan.  The Regional Air Monitoring System, St.
     Louis, Missouri.  In:  Proceedings of the International Conference on
     Environmental Sensing and Assessment, Las Vegas, Nevada, Sept. 14-19,
     1975.

21.  Nie, N.H., C.H. Hull, J.G. Jenkins, K. Steinbrenner, and D.H. Bent.
     Statistical Package for the Social Sciences, McGraw-Hill, New York,
     New York, 1975.
                                      26

-------
                                    TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing)
                              2.
                      - A PRACTICAL APPROACH
 TO AEROSOL FRACTIONATION AND COLLECTION
                                                             3. RECIPIENT'S ACCESSION NO.
              5. REPORT DATE
                June 1978
                                                             6. PERFORMING ORGANIZATION CODE
                                                            8. PERFORMING ORGANIZATION REPORT NO.
 T.G. Dzubay
9. PERFORMING ORGANIZATION NAMF. AND ADDRESS
 Environmental Sciences Research Laboratory
 Office of Research and Development
 U.S.  Environmental Protection Agency
 Research Triangle Park, N.C. 27711	
              10. PROGRAM ELEMENT NO.
                 1AD712
              11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
                                                             13. TYPE OF REPORT AND PERIOD COVERED
 Environmental Sciences Research Laboratory — RTP, N.C.
 Office of Research and Development
 U.S.  Environmental Protection Agency
 Research Triangle  Park, N.C. 27711	
              14. SPONSORING AGENCY CODE

                 EPA/600/09
15. SUPPLEMENTARY NOTES
16. ABSTRACT
      Procedures to  size fractionate, collect, and analyze  ambient concentrations
of participate matter are described. Emphasis is placed on the design and characteristic 3
of the single-stage dichotomous sampler.  A new inlet is described that samples
aerosol independent of wind speed and direction, and a discussion of the advantages
of a  new pneumatic  flow control system  is included. Comparative results of the
high-volume and dichotomous sampler are presented.
17.

a.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
"''Air pollution
"Aerosols
"''Sampling
"''Particle size distribution
'''Instruments
b.lDENTIFIERS/OPEN ENDED TERMS  C.  COSATI Field/Group
                                 3.3B
                                 07D
                                 IfB
1H. HIST III HUT ION ST AILMENT

     RELEASE TO PUBLIC
19. SECURITY CLASS (Till!! Report)

      UNCLASSIFIED
21. NO.. OF PAGES
   "Si
                                               20. SECURITY CLASS (Thispage)

                                                     UNCLASSIFIED
                                                                           22. PRICE
EPA Form 2220-1 (Rev. 4-77)   PREVIOUS EDI TION is OBSOLETE
                                             27

-------