Environmental Science* Raaaarch
            Laboratory
            n«Marcti Triangto Park NC 27711
EPA-600/3-79-019
March 1979
   and Davaajpmant
Sources and
Transport of Trace
Metals in  Urban
Aerosols

-------
                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-600/3-79-019
                                          March 1979
       SOURCES AND TRANSPORT OF TRACE
          METALS IN URBAN AEROSOLS
                      by
  John W. Winchester and J. William Nelson
         Department of Oceanography
          Florida State University
         Tallahassee, Florida  32306
                Grant R802132
              Project Officers

  William E. Wilson and Ronald K. Patterson
           Aerosol Research Branch
  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 publication.  Approval does not signify that
the contents necessarily reflect the views and policies of the
U.S. Environmental Protection Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use.
                               11

-------
                            ABSTRACT

     A methodology was developed, tested, and applied to the
identification of trace element aerosol sources and trace element
aerosol transport phenomena.  The method uses lightweight and
portable field sampling equipment and results in the acquisition
of large data sets suitable for statistical analysis and the
testing of aerosol transport models.  Aerosol sampling in the
ambient 'atmosphere was accomplished through a combination of
time-sequenced filter collections with hourly resolution, and
fractionated aerosol collections with 0.25 ym resolution.
Elemental analysis was performed with particle induced X-ray
emission, a sensitive and rapid method.

     Aerosol characterization studies were focussed on St. Louis,
Missouri, during the Regional Air Pollution Study.  Trace element
pollutants were distinguished from their natural components on
the basis of particle size distributions.  Statistical analysis .
of time-sequenced concentration measurements revealed four dis-
tinct pollution sources for lead.  A related analysis demonstrated
that meteorological factors controlling the transport of titanium
aerosol across the city could be identified.

     Extensions of the methodology were applied to (1) determina-
tion of trace element pollutant deposition in the human respira-
tory tract through direct sampling of exhaled breath, (2) the
indoor environment by comparison of time variability in elemental
concentrations with those measured outdoors, and (3) nonurban
continental and marine atmospheres.

     This report was submitted in fulfillment of Grant No. R802132
by Florida State University under the sponsorship of the U.S.
Environmental Protection Agency.  This report covers the period
from April 1, 1973 to January 31, 1976, and work was completed as
of December 1, 1978.
                               111

-------
                          CONTENTS

Abstract	iii
Acknowledgment 	  vi

   1.  Introduction	   1
   2.  Conclusions	   3
   3.  Trace Metal Identification Techniques 	   6
           PIXE principles	   6
           PIXE spectrum resolution  	  10
   4.  Sampling the Urban Aerosol  	  15
   5.  Clues to the Sources of Trace Metals in Urban
       Aerosols	22
           Evidence from particle size distribution  ...  22
           Meteorological analysis of the titanium
             pollution plume 	  30
           Lead, a multiple source pollutant in
             St. Louis	34
   6.  Aerosol Properties and Human Respiratory Exposure  .  40
           Respiratory response in aerosol inhalation  .  .  40
           The fate of lead aerosol along a roadway  ...  44
           Lead aerosol in the indoor residential
             atmosphere	48
   7.  Aerosol Transport out of the Urban Source Region   .  51
           The north Florida coastal zone	51
           Air pollution present in the North Atlantic
             marine atmosphere at Bermuda  	  54

References	58
List of Publications and Reports	  59

-------
                         ACKNOWLEDGMENT

     The support and encouragement of William E. Wilson, Ronald
K. Patterson, and members of the Aerosol Research Branch have
been crucial in bringing the PIXE analytical method and its
associated aerosol sampling methods from a concept to a
demonstrated methodology for the investigation of air pollution;
this is gratefully acknowledged.  We thank the numerous members
of the PIXE group at FSU, whose works are cited in this report,
for their role in bringing about the results described.  Special
thanks are due to Scott Rheingrover and Jean Sasso for their
constructive work in the initial and final stages of preparing
this report, and to Helena Johnston and Janice Szabo for
stenography and typing.
                               VI

-------
                            SECTION 1

                          INTRODUCTION

     The suspended particulate matter of urban atmospheres
contains substantial quantities of trace metals together with
sulfates, organic matter, and other non-metallic constituents.
Trace metals, of course, occur naturally since all of them are
constituents of naturally occurring dust which originates from
materials at the earth's surface.  However, urban activities,
including the combustion of fuels, industrial processes, and the
motions of vehicular traffic which tend to raise surface dust,
are activities which tend to increase the concentrations of
certain trace metals in the atmosphere above their natural
levels.

     Some trace metals play an active role in the metabolism
of man and other living things, and some metals, when present
to excess, may be toxic to man.  Trace metals may also play an
active role in the chemistry of the atmosphere, for example by
acting as catalytic substrates for chemical reactions that
convert gaseous into particulate material.  Such may be the case
for some metal oxides which may accelerate the conversion of
sulfur dioxide to sulfates or may facilitate the condensation of
organic vapors into small liquid droplets.

     For these reasons a thorough investigation of the contents
of trace metals in urban atmospheres is needed for setting
optimum air pollution control strategies.  Such an investigation
should seek to find ways to identify the sources of individual
trace metal constituents of the particulate matter and the
factors which determine the extent of transport of these metals
from their sources to other atmospheric locations and eventual
removal from the atmosphere.

     The atmosphere in most cities contains particulate matter
derived from a multiplicity of sources.  Seldom does one find
an element which is derived predominantly from one source type,
but rather any trace metal may enter the atmosphere from a
number of different processes and types of human activity.
Therefore it is not a straightforward task to identify the
sources of trace metals by measuring their concentrations in the
atmosphere.  Indeed, it is also not straightforward to determine
unambiguously whether the measured trace metal concentrations
in the urban atmosphere are higher than their natural levels.

-------
Consequently special measurement techniques should be devised
for the study of trace metals in urban atmospheres which will
help to identify sources and how trace metals are transported
through the atmosphere.

     New methods of investigation are needed for the collection
of samples of atmospheric particulate matter, for their chemical
analysis so that individual trace metal concentrations may be
determined, and for the study of the concentration data, for
example using methods of statistics or meteorology so that
source and transport information may be obtained.  In addition,
it is important to gain some understanding of whether the
concentrations observed and the characteristics of the particles
containing the trace metals are such as to present a potential
health risk to man, for example to his respiratory tract.
Finally, a study of trace metals in urban atmospheres is not
complete without some parallel investigation of clean air, air
which has not been contaminated to the degree found in urban
areas.  Such a study is important, both to provide an estimate
of the composition of the natural atmosphere, and thus by
comparison the degree to which the urban atmosphere has been
contaminated, and also to determine if urban atmospheric
materials can be transported to cleaner areas sufficiently to
modify the atmosphere of non-urban or marine areas by long
range transport of air pollution.

     In this report we shall review briefly the advances we have
made at Florida State University in the methods of investigation
of trace metals in urban aerosols; the characteristics of
aerosol particles in the city of St. Louis, Missouri  (in many
ways a typical U.S. city); the clues that these characteristics
provide for determining the sources of trace metals; the
relationships between aerosol properties and the exposure
presented to the human respiratory tract upon inhalation; and
finally aerosol properties in cleaner air from which we may
evaluate the extent of long range transport of pollutants
outward from the urban areas.

-------
                            SECTION 2

                           CONCLUSIONS

1.   The method of proton induced X-ray emission analysis, PIXE,
is extremely precise and sensitive for the measurements of trace
metal concentrations in atmospheric aerosol samples.  It has
been successfully calibrated and intercompared with other
analytical methods for absolute accuracy of the order of 10%.

2.   The PIXE method permits the use of small portable aerosol
sampling equipment which can be economically deployed over
networks on the urban scale.  The small area of the analyzed
portion of a sample permits use of miniature vacuum pumps with
modest electrical power requirements.  Thus, large scale aerosol
sampling programs may be carried out with greater versatility
and lower costs than if larger field sampling equipment were
required.

3.   Particle size distribution measurements can be made by
sampling particle size fractions with cascade impactors.
Impactors compatible with the PIXE method provide 6 particle
size fractions, <0.25 ym, 0.25-0.5 ym, 0.5-1 ym, 1-2 ym,  2-4 ym
and >4 ym aerodynamic diameter.  The size range below about 4 ym,
which comprises most of the particle mass for several trace
metals of pollution importance, includes.the respirable.size range
over which it is important to determine the distribution of
elemental composition.  Not only is this information desirable
from a health effects standpoint, but it also assists in the
identification of different pollution source processes for
atmospheric trace metals.

4.   Measurements of the time variability of aerosol trace metal
concentrations with hourly time resolution can be made conve-
niently with the continuous time sequence "streaker" filter
sampler.  This time resolution is matched to the scale of
important variability in the atmospheric motions that control
the transport of aerosols over the urban dimensions.  With time
variability information transport models may be tested, and
specific pollutants may be identified as useful transport
tracers.

5.   In St. Louis, trace metals from low temperature pollution
sources may occur primarily in particles larger than 1 ym
aerodynamic diameter; whereas high temperature sources

-------
preferentially emit submicrometer particles to the atmosphere.
For example, automotive lead aerosol particles are significantly
smaller than those of smelter lead aerosol.  Industrial iron
aerosol appears to have a larger component of submicrometer
particles than iron derived from soil or other dust.

6.   A local titanium pollution source in St. Louis causes
elevated titanium concentrations throughout the St. Louis urban
area.  The concentrations show variations which are strongly
dependent on wind direction, variability in mean wind direction,
wind speed, atmospheric temperature gradient, land and water
differences, and the occurrence of precipitation.  The time
variability pattern of titanium observed at different measurement
stations has been used to evaluate quantitatively the importance
of these factors in governing the titanium transport.  In
principle other trace elements and their associations may also
be used as tracers to test transport mechanisms from stack
emissions in the urban area in a manner similar to that
demonstrated for titanium.

7.   The association of lead with other trace metals in the
St. Louis atmosphere has been analyzed statistically to demon-
strate the existence of four different pollution components of
atmospheric lead.  These are the automotive component with a
strong bromine correlation, a zinc related component possiby
from smelter emissions, a crustal element related component
possibly from fly ash, and an independent component possibly
from secondary lead smelting, showing no other elemental
association.

8.   The high sensitivity of the PIXE method permits the sampling
of exhaled human breath aerosol in normally polluted atmospheres
by means of cascade impactors.  Comparison of breath with ambient
aerosol particle size fractions has enabled us to estimate the
efficiency of particle deposition in the human respiratory tract
as well as a humidification response on inhalation.  The fact
that a minimum deposition efficiency is observed for particles
in the range 0.25-0.5 ym or possibly smaller points to the
importance of particle size resolution in this range when
evaluating potential health effects of pollution aerosols.

9.   Streakers and cascade impactors have been used for
monitoring lead, sulfur, and other elements along a heavily
travelled roadway.  Particle size of lead aerosol is found to
be significantly greater for aged air parcels compared to air
from the traffic lanes, an observation of importance to
estimating human respiratory deposition of lead.  Aged air is
also characterized by a lower Br/Pb ratio than that of fresh
automotive aerosol, suggesting bromine loss from the particles
during atmospheric transport.

-------
10.  Indoor air pollution has been monitored directly with the
streaker sampler and compared with outdoor concentrations.  A
detailed study of the correspondence in time variability patterns
for certain residential atmospheres has established that lead
and sulfur generally penetrate indoors ftfom the outdoor urban
atmosphere but that potassium has significant indoor sources.
Some evidence has also been obtained indicating that a portion
of the lead observed indoors may be due to recirculated c'on-
taminated household dust.

11.  Measurements in nonurban continental and marine atmospheres
have been made to evaluate the natural atmospheric concentrations
with which the urban atmosphere may be compared.  It is note-
worthy that at Bermuda, more than 1000 kilometers from the
eastern U.S. seaboard, fine particle sulfur aerosol was found
to be many times above its level at highly remote sites in the
Southern Hemisphere and indicates substantial long-range
transport of pollution out over the Atlantic from the North
American continent.

-------
                            SECTION 3

              TRACE METAL IDENTIFICATION TECHNIQUES

     The measurement of trace metals in urban air particulate
matter requires highly sensitive detection methods in order
to measure the small quantities of each metal that may occur in
samples collected over short periods of times.  Furthermore,
valuable clues concerning the sources of trace metals may be
•obtained by examining the distribution of concentration over
different particle size ranges.  Since sampling devices for
short sampling intervals and for particle size selection   ,  .
gather only small amounts of suspended particulate matter, a'
high degree of sensitivity is required for a suitable elemental
analysis technique.

     The method of particle induced X-ray emission, using  ;
protons accelerated by a Van de Graaff accelerator up to 5 MeV,
provides sensitivity for accurate measurements of many of the
important trace metals in urban atmospheres down to a nanogram
or less of each metal.  Furthermore, this method is best used
to analyze samples of very small dimensions, a few mm^'or less,
and those samples which are collected conveniently by time
sequence and particle size fractionating devices.  This
method, known as PIXE, is the method of choice for obtaining
the atmospheric chemical information described here.


PIXE PRINCIPLES

     In principle, PIXE involves the bombardment of a sample
for a minute or so in the vacuum chamber of a particle
accelerator so as to generate X-rays characteristic of each
element in the sample.  These X-rays are detected by a solid
state lithium drifted silicon detector, Si(Li).  Electrical
pulses from the detector are transformed electronically into
signals recorded on a magnetic tape.  After a few hundred
analyses have been performed and recorded on the magnetic tape,
the tape records of X-ray spectra of the individual samples are
resolved into the concentrations of the individual elements by
a digital computer.  Most of the techniques for aerosol
sampling, PIXE analysis and computer X-ray spectrum resolution
which have been used in the work reported here have been
developed especially for the purposes of this study of trace
metals in urban aerosol particulate matter.

-------
                                        Plate  1
Table I. Results of Analysis of Thin Homogeneous
    Standard Samples
                               PIXE results
              Table II. Results from an Intel-laboratory
                 Comparison Study

Element
(compound)
Cl (NaCl)
Cl (KC1)
K (KC1)
Ca (CaF)
Ti metal
Cr metal
Fe metal
Cu (CuS)
Ga (GaP)
Br (CsBr)
Cs (CsBr)
Amount
given,
ug
6.8
6.3
7.0
3.7
10.2
7.2
11.5
7.9
4.6
3.1
5.2

Found,
ug
5.6
6.6
7.1
3.6
8.9
7.1
13.7
8.1
4.6
3.0
5.5
Number
of deter-
minations
3
7
7
5
3
5
5
14
2
3
3
« Std dev
per deter -
urination
5.2
6.7
6.5
9.8
9.0
1.0
3.1
7.1
7.8
3.0
6.0
.. 	
Table III. Comparison between Heterogeneous and
    Homogeneous Samples
               Heterogeneous foil
                ("i mm piece)
Zr-foil       5.66 ± 0.53 mg/cm2
Fe in Al-foil 21.3 ±2.0 fig/cm2
Zn in Al-foil  1.02 ± 0.15 (ig/cm2
                                                          Sample on Milllpore
                                                              filter
                                            Sample on Whatman 4t
                                                 filter
                                                  K
                                                  V
                                                  Mn
                                                  Fe
                                                  Cu
                                                  As
                                                  Pb
Amount
given,
ng"
2040
374
1649
1916
300
356
813
Amount found
Amount given
1.032
1.071
1.022
1.093
1.110
1.154
1.064
                                            given,

                                             943
                                             233
                                             693
                                            1269
                                             202
                                             244
                                             558
                                              Mean
                                                                 1.078
Amount given
  0.981
  1.069
  0.915
  0.988
  1.151
  1.120
  1.083

  1.044
               "Given amount of each element within area of proton beam
              striking sample. * Ratio of PIXE result to given amount for 2 to 4
              replicate determinations. The standard deviation for a single de-
              termination by PIXE is approximately 15%.
   Homogeneous foil
 5.01 ± 0.35 mg/cm2
23.3 ±1.7 ng/cm2
 0. 80 * 0.10 Mg/cm2
          Ing    long
                     lOOng    l^g    icyig
                     Weight - *
     5. Results from linearity tests
                                                                                     50  100
                  0-'  0.2    0.5    I   2     5   10   20
                                Thickness (mg/cm2)
                     X-Ray attenuation in a homogeneous sample as a function
              of sample thickness
 Source:   A-2
 Thomas B.  Johansson,  Rene  E.  Van  Grieken,  J. William  Nelson,  and
 John W.  Winchester, Elemental  trace  analysis of  small samples by
 proton induced  X-ray  emission,  Anal.  Chem.,  47,  855-860, 1975.

-------
     In the development of the PIXE method a considerable amount
of testing has been carried out to assure accuracy and reli-
ability of the method.  For example, Plate 1.shows in three
tables that standard samples can be analyzed to an accuracy of
1 - 10% on an absolute basis.  An interlaboratory comparison of
PIXE method with the results obtained using other analytical
methods at other institutions (Camp et al., 1974, 1975) shows an
overall agreement for several trace metals of better than 10% on
the average.  We likewise have verified that the PIXE method is
capable of analysis of a very small spot of sample as well as an
evenly distributed wide area sample.  This advantage is of
special significance because it permits the analysis of indi-
vidual particle size fractions collected by cascade impactors of
the single orifice type which are described further below.  This
impactor collects samples as circular deposits of approximately
1 mm2 diameter and requires the analysis of the entire sample
spot.  In the PIXE method the proton beam strikes a larger
area which completely contains the spot and measures each
trace metal contained within it.
                      i

     Also shown in Plate 1, Figure 5, is a demonstration that
the PIXE method permits a linear response of X-ray signal over
at least a factor of 10,000 in amount of sample irradiated by
the proton beam.  A potential limitation to linearity is
samples which are very thick and attentuate X-rays as they
leave the sample after their proton induction.  In Plate 1,
Figure 6, a calculated absorption of X-rays in the sample shows
that samples thinner than about 1 mg/cm^ should give less than
about 10% absorption for X-rays as soft as those for sulfur.
More energetic X-rays are absorbed to a lesser extent so that
lead X-ray attentuation reaches 10% only for thicknesses
substantially greater than 10 mg/cm^.  Samples collected by
the time sequence and particle size fractionation devices
described below produce sample deposits thinner than these
values for convenient sampling times in urban atmospheres.

     For specialized applications of PIXE to the analysis of
iron or other materials composed primarily of heavy elements,
additional sources of error should be considered in the
determination of trace elements with atomic number lower than
that of the matrix, although these appear not to be serious
in general for atmospheric particulate matter.  Plate 2
illustrates the effect known as "enhancement,"  In the bombard-
ment of a sample composed principally of iron, for example,
considerable amounts of iron X-rays are generated.  These
X-rays 'may be critically absorbed by the trace metals also
present in the sample and cause a secondary generation of their
characteristic X-rays.  Therefore the trace metal X-rays
detected will be those not only due to the proton induction but
also the secondary X-ray induction.  As a result, a signal
enhancement occurs which will cause an erroneously high
analysis unless enhancement is corrected for.  As indicated in

-------
                                       Plate  2
                      TABLE 1
                      Enhancement in % from the Fe Ka and Fe tig X-rays on Ka
                      X-rays from different elements. Calculated for 2.5 MeV pro-
                      tons in a 100% iron sample.
z
25
24
23
22
21
20
19
17
Element
Mn
Cr
V
Ti
Sc
Ca
K.
0
Exciting
FeK,,
0
41
24
14
8.6
4.9
2.7
0.83
X-ray
FeK,,
7.8
4.7
2.8
1.6
0.95
0.54
0.30
0.09
                       TABLE 2
                       Analysis of NBS low alloy steel SRM 1261 (95.6% Fe), with
                       5 MeV protons.

                         Element              Concentration (%)
                                        No        With     NBS value
                                    enhancement  enhancement
                                   •   correction     correction
                           Mn

                           Cr
     0.80

     1.24
0.71

0.71
0.66

0.69
                          70

                          60
                        - 50
                        §40
                        o>
                        £ 30
                        o
                        c20
                       LU
                          10
                                   CrKa
CaKfl
                            012345
                                  Ep(MeV)
                       Fig. 2. Enhancement from Fe Ka X-rayi on Cr K, and C« K,
                       X-rays as a function of proton energy in • 100% iron sample.

Source:   A-U
Mats.  S. Ahlberg,  Enhancement  in  PIXE  analysis,  Nuclear  Instru-
ments  and  Methods, 142_, 61-65,  1977.

-------
Table 1 of Plate 2 the enhancement of chromium X-rays in an
iron sample may be more than 40%.  In a typical steel sample
analyzed without enhancement correction the chromium analysis
was found to be high, but if the enhancement correction is made,
the analysis is in good agreement with the standard value.  As
seen in Figure 2, Plate 2, the enhancement problem becomes more
severe at higher proton energies but is less severe for elements
of considerably lower atomic number than the matrix, such as
calcium in comparison with the matrix element, iron.

     In aerosol analysis most major elements are of lower atomic
number than the trace elements measured by PIXE.  For example,
more than half of the total particulate mass typically found in
urban atmospheres is composed of the elements C, H, 0, N, S,
with additional Na, Mg, Al and Si, all of these being elements
of atomic number lower than the trace metals whose concentra-
tions are discussed further in this report.  Even for the
extreme case of measuring trace amounts of sulfur in pure rock
salt, NaCl, calculations show that enhancement of sulfur X-rays
is not a source of significant analytical error.


PIXE SPECTRUM RESOLUTION

     A crucial part of the PIXE analytical procedure, and one
without which the method would not succeed, is the translation of
the accumulated electronic signals from the Si(Li) X-ray detector
into a list of elemental concentrations in the sample.  This
feat is accomplished by the unraveling of the X-ray spectrum,
consisting of a number of X-ray counts recorded in each of 600
or more individual energy channels, into the contribution from
each elemental constituent of the sample.  Plate 3, Figure 2
shows a typical X-ray spectrum for an aerosol sample collected
on a filter.  The three parts of the figure show the effect of
interposing a thin absorbing material in front of the Si(Li)
detector window causing some of the X-rays, especially those of
lowest energy (lowest channel number), to be absorbed.  In
both cases individual elements present in the sample account
for peaks in the spectrum, but a thin Mylar plastic absorber
skews the entire spectrum in the low energy end so that the
elements Al, Si, and S are completely suppressed.  In samples
where these elements may be so abundant that their X-rays are
intense enough to limit the permissible bombardment intensity of
the sample by the proton beam (because of limitations on the
speed of electronic circuitry), and thus lengthen the total
analysis time required to detect the heaviest elements in the
sample, this technique of using an X-ray absorber is useful.  By
suppressing these troublesome light elements, a more intensive
proton bombardment is permitted, and shorter analysis time is
required for accurate measurement of heavier metals, for
example lead.


                               10

-------
IOS[ Nelson Streaker


IOM
c«, Absorber: None
'"' Hole:-
si § A
A(/\ ,, ; 1
GJ 1 ! v \ A' \COJ
1 ! /•~"'V-.\T, s

COUNTS/
5,
10


,0«
"1 ^1" i i
/ N&! -r
''\
s^lzn) A Br IP61
'"^L ^l y. r
% WB/ L' V »' s, «FW z.
T^— ; — ^^f-^-f'f^g
> 100 200 300 MOO 500 600
io5


10
UJ

cr
5 „
COUNTS/
5<
10


• io'o
Nelson Streaker
'• Absorber. -350 urn Mylar
jj, (l Hole: None
A i i
HIM „ 1 (F"
1 ^ A'^V/)*" ' •
1 UP0"^.,;/ i
1'. ' v\ A A IPW
/ V1t3ri: M
-* Hyi/bsr*
*^l ' \ i\jV( *.
^V^"—--~i/ V Vl* *
	 ^r

100 200 300 400 500 600








!


'
CHANNEL NUMBER CHANNEL NUMBER
T1
105


U
-1 l°
UJ
z
i
5 ,
£io3
^
g
U
io2
..i
Nelson Streaker
Absorber :350/Ltm Mylar
Sfi r. Hole: 9%
s A A
5; 5 i i .
A A .. ' to"
'D VP ^ <\Tiv ' '"*'
1 T"\l VWivlJln; . <\
;,- ~~' "^t/ •.
"'.' '•t'o.Z" »
/ vy\ A
' v«^, , " ^ A
•Ozn), , ^ ;;
' ' 1 ' * ' 1
^•V ! : U l»-> & HPtH
'-. \(Rl)i U -i » '. Zr
^^/VL
^~~^-.--v ^



v
,-)-
CD









CO
(1)





'

                         100   200    300    MOO    500   600
                                   CHANNEL NUMBER
                 Fig. 2. Fits to proton induced X-ray emission spectra obtained from bombarding an aerosol sample obtained with the Nelson
                 Streaker. For quantitative comparison see table 1.
Source:   B-8.
Henry C.  Kauf mann,  K.  Roland Akselsson,  and William J.  Courtney,  REX-a compu-ter
programme for PIXE  analysis, Nuclear Instruments  and Methods, 142, 251-257, 1977.

-------
                                                                               rt
                                                                               (D

                                                                               CO
                         zoo
                          CMJkNNIL
Figure 4.
                     Display of relative contributions of S, Pb,  and background
                     X-ray components for equal masses of S and Pb in typical
                     streaker sample subjected to routine PIXE analysis.  Calcu-
                     lations with (	)  and without  (	) self-absorption
                     corrections and a relative background position for Nucle-
                     pore filters are shown as separate components.

Source:   A-18.
K.R. Akselsson, K.A.  Hardy, G.G.  Desaedeleer, J.W.  Winchester,  W.W. Berg,  T.B. Vander
Wood,  and J.W. Nelson,  X-ray techniques for aerosol sulfur baseline assessment along
an urban freeway, Advances In .X-Ray Analysis, vol.  19, R.W. Gould, C.S.  Barrett, J.B.
Newkirk, and C.O. Ruud,  eds., pp.  415-M-25, Kendall  Hunt, Dubuque, Iowa,  1976.

-------
     A more usual case in aerosol analysis is where both the
light and heavy elements must be determined, and therefore a
complete suppression of the light elements is not desired.  In
that case the absorber is used as before but in a modified form,
such that a small hole is cut in the center of the absorber in
order to transmit without absorption a small fraction of the
unattenuated X-rays; for example the spectrum using a 9% hole
.absorber is shown in Figure 2 of Plate 3.  The light elements,
which are completely suppressed by the remaining 91% of the
absorber area, are seen as unattenuated X-ray peaks at 9%
of the original intensity.  The heaviest elements are
effectively unattenuated by the thin absorber, as before.  A
few elements of intermediate character, in this case K and Ca,
are partially attentuated by the absorber, but from calibration
measurements their overall transmissions may be calculated.
With this X-ray absorber the entire spectrum of elements from
Al to Pb and beyond may be measured in a single bombardment
lasting only a minute or so depending on sample size.

     A practical case of potential X-ray interference between
elements is illustrated for the case of measuring sulfur in
the presence of lead.  Along roadways where leaded gasoline is
burned the concentration of atmospheric lead may be highest
relative to sulfur in comparison with air anywhere in the urban
.environment.  The lead X-ray spectrum contains many peaks, both
in the higher energy region where the peaks are used for element
identification and in the low energy region of the most important
sulfur peaks.  These low energy M X-rays of lead pose a potential
interference in the accurate resolution of sulfur from the
spectrum.  The computer X-ray resolution procedure developed
for this research removes the interference by accurately
calculating the amount of lead present, estimating the low
energy lead X-ray component, and determining the net sulfur
X-ray present.  Under normal circumstances the extent of lead
X-ray interference in sulfur determination does not exceed a few
percent, and even in a roadway situation the interference,
whereas perhaps ten times greater, is still accurately
measurable.

     This example illustrates that the PIXE X-ray spectrum
resolution problem requires a detailed evaluation of all physical
relationships among the constituent elements in the sample.  An
especially important part of the problem is the accurate deter-
mination of the X-ray background contributed by the sample mass
and the sample substrate, for example filter paper or particle
impaction surface.  This background is produced by atomic
interactions during proton bombardment distinct from those which
produce the characteristic X-rays of the elements.  In the light
element region of sulfur and neighboring elements the background
X-ray spectrum exhibits considerable curvature, and in the
case of the absorber with hole the curvature is of complex
shape.  If the concentrations of light elements are relatively

                               13

-------
low, the estimation of their amounts above background requires
an accurate measurement of the background.  In the computer
procedure we do not use separate measurements of substrate
materials without sample, e.g. blank filters, nor direct
measurement of the background since this would not include the
mass effect of the sample itself in causing the background
X-radiation.  Instead, in our procedure the principal physical
processes causing the background have been identified, and the
background shape has been successfully modeled.  Comparison
with standard samples indicates that this modeling of the
background permits accurate measurements of the important light
elements.  Thus the computer procedure for spectrum resolution
requires only the data contained in the sample spectrum alone,
and offers a great advantage in simplicity and speed of the
computer data processing (normally about one minute per
spectrum) and in overall accuracy.

-------
                            SECTION 4

                   SAMPLING THE URBAN AEROSOL

     A comprehensive investigation of trace metals in urban
aerosols requires techniques for measuring their concentrations
in air which will:

1.   Provide a determination of trace metal distributions among
     particles of different size over the stable aerosol size
     range <0.25 ym to >4 ym aerodynamic diameter.

2.   Provide an indication of the time variability of trace
     metal concentration on a time scale comparable to the
     3-hour meteorological data provided by the National
     Weather Service.

3.   Provide for the simultaneous measurement in each sample
     of several trace metals of air pollution interest so that
     inter-element correlations may be studied as indicators of
     sources and transport pathways.

4.   Provide data sets of sufficient size for statistical
     methods of interpretation to be employed.

5.   Permit the ready deployment of large numbers of aerosol
     sampling devices simultaneously over an urban area, with
     minimum restrictions on site selection due to electric
     power, mounting platforms, or other logistic requirements.

6.   Permit reasonable speed and cost for carrying out both
     sampling and elemental analysis of aerosols collected in
     field programs.

     These conditions require the development and validation
of aerosol particle sampling devices and methods which have not
previously been employed on a large scale in studies of trace
metals in the urban atmosphere.

     The measurement of elemental composition as a function
of particle size is accomplished by sampling with a cascade
impactor.  Plate 4 is a schematic diagram of the impactor used
in connection with PIXE analysis.  The basic design is that of
the Battelle 5-stage impactor (Mitchell and Pilcher, 1959) and
requires a flow rate of 1 liter/minute.  The impactor may be

                                15

-------
SMALL JET^SMA^
     MTIAI  HIM
       wilk
    STICKY COATIMG
    I	r
                                    Plate  U

                              AIR    FLOW- fliier/«in
CASCADE  IMPACTOR5

OPERATED IN THE INVERTED POSITION

                              6
                                                         B
                                                    DISC SUPPORTS  (3)
                                                    DELRIN  WASHER
CRITICAL-FLOW
   ORIFICE


FILTER STAGE
(for 25mm, 0.4p
  Nuclepore)
                            i
                        «W -H
                       X X  k
                       QJ H  Q)
                       fc CU  g
                       QJ   <
                       PH "O
                         C  «
                        • (d  CO
                       1-3   -H
                        • bO CO
                       « c  >
                         •H H
                        "H  (d
                       fc OH C
                       QJ g <
                       •P rd
                       co co -p
                       QJ    a
                       ,C k  r0
                       O QJ -P
                                                                           5 0) H
                                                                             & o  •
                                                                           . 4J PL, r-
                                                                           s co   r-
                                                                               h en
                                                                           3 -a -H H
                                                                             c <
                                                                           •> id   '~-
                                                                               C co
                                                                               H CO
                       c
                       o
                       CO fc
                       H O CO
                       QJ -P QJ
                       S O O
                         Id C
                        • a< (d •
                       ^ e >•
QJ    CO
•H  C QJ >M
H  O O
CO -H QJ 5
QJ -P PH QJ
J  (d  S
   N *
 ••H g «
O  ^ 3 CO
 •  QJ -H J^
CJ -P CO QJ
 •  O O QJ
<  rd ft C
   SU E-H
 « rd >i bC
C ^ CO C
QJ  O  W
CO   U-|
d H O H
QJ  O   rd
1-3  CO bO O
   O C-H
 •  ^-H g
CQ  Q) T3 QJ
  < QJ ^
 "   QJ O
 .  »> O
fc  C O4H
"-3  QJ 5n O
                                                                        CO

                                                                        <
                        "CO   QJ
                       'a w a -p
                       fc   -H 3
                       rd C   -P
                       TD rd  "-H
                       O > CO -p
                       O   -H CO
                       3-  • co c
                                  TO VACUUM PUMP
                                                                        O
                                                                        CO
                         -a (d
                         c c
                                                                                 O
                                        16

-------
operated in any position so that air may be drawn vertically
downward into the device from the top, or the impactor may
be inverted face down so that air is drawn upward from below.
The air is passed successively through a series of circular
orifices of decreasing diameter which causes the linear flow
rate of air to speed up as it goes through successive stages.
Directly downstream of each orifice is an obstruction in the
form of a sticky Mylar film backed by a solid support surface.
such that the largest particles passing through the orifices
at each stage are caused to impact and stick to the Mylar
surface.  Smaller particles pass around the impaction surface
but at the next stage, owing to increased air velocity, the
largest of them will be collected by impaction.  The impactor
has been designed to give 50% collection efficiency for stages
0 through 5 of 8, 4, 2, 1, 0.5 and 0.25 ym aerodynamic diameter,
so that the particle size range of the stages is the range
of diameters between these values.  Particles smaller than
those collected by impaction stage 5 are collected by filter
stage 6.  The Nuclepore filter used is of small enough pore
diameter to assure at least 60% collection of any particle
size (Liu and Lee, 1976); calculations show that the minimum
efficiency is in the range of the few hundredths of a
micrometer diameter where diffusion and impaction mechanisms
of particle uptake by the filter are comparable.  Each sample
is collected conveniently in a small area suitable for analysis
by the proton beam in PIXE.

     It should be emphasized that the particle size discrimina-
tion of the cascade impactor is effective for particles as small
as 0.25 ym aerodynamic diameter.  (Physical diameter differs
from aerodynamic diameter if the density of the particles is
different from 1 g/cm^ and the shape of the particles is
sufficiently different from spherical to cause shape-dependent
air drag effects.)  Cascade impactors may be designed to
operate at higher flow rates but generally with the smallest
size cut at larger diameters, e.g. 0.5 ym or greater.  The
work described in this report will demonstrate that the
information obtained by small particle size resolution in the
0.25 ym range is extremely valuable for interpreting sources
:and transformations of trace elements in atmospheric particulate
matter.  Moreover, the response of the human respiratory tract
to inhalation of fine particles varies significantly with
particle diameter in the 0.25 ym region so that health effects
can better be evaluated if particle size distribution and
composition for particles this small can be determined.

     The time variability of trace metal concentrations in the
air can be obtained easily and automatically if sampling is
carried out with a filter sampler whose exposed surface is
changed continuously.  In Plate 5a a diagram of a special sampler
developed at Florida State University for this program is
shown.  This time sequence filter sampler consists of a single

                               17

-------
CO
              BRASS  BARS

             TO  ACTIVATE
            MICRO SWITCHES
              OluNUCLEPORE
                                                                           SYNCHRONOUS
                                                                                   MOTOR
                                                                                   (I r. p. h. )
                                                                                      rt
                                                                                      Ct>
                                                                        I I 0 V
                                          2 x 5 m m
                                      TEFLON  ORIFICE
Source:  A-8.
A.P. Woodard, Jr., B. Jensen, A.C.D. Leslie, J.W. Nelson, J.W. Winchester,  R.J.  Ferek,
and P. Van Espen, Aerosol characterization by impactors and streaker  sampling and PIXE
analysis, in Proceeding of Symposium, Recent Advances in Air Pollutant Analysis, Ameri-
can Institute of Chemical Engineers, New York.j (in press) 1977.

-------
                           Plate 5b
                              Wmr
                              •PUT ^(PWiWp^^^^^^
  Figure 2.1  View of the continuous filter sampler.  The 2- X 5- mm
  air intake (upper right center) is driven to the left at a rate of
  1 mm per hour by the synchronous motor.  A Nuclepore filter strip is
  placed over the intake to produce an 84- mm long strip sample in
  one week.
Source:   B-ll.
J.  William Nelson, Proton-induced aerosol analyses:  methods and
samplers, in X-Ray Fluorescence Analysis of Environmental
Samples, Thomas G. Dzubay, editor, pp. 19-34, Ann Arbor Science
Publishers, 1977.
                               19

-------
strip of Nuclepore filter stretched on a frame and mounted in a
device which causes a sucking orifice on Teflon or other smooth
plastic to be drawn gradually from one end of the filter to
the other.  The sucking orifice is attached to a vacuum pump,
and as the orifice is drawn along the length of the filter, the
suction seals the orifice to the smooth back of the Nuclepore
while at the same time permitting air to be drawn through the
filter at a rate controlled by its porosity.  The device is
designed to be driven by a clock motor connected to a precision
threaded screw so that 7 days (168 hours) is required for a full
transit along the length of the filter.  In urban atmospheres
a feasible streak of aerosol particle deposit is produced over
the one week sampling period.  Thus the sampler is known as a
"streaker."  If desired the clock mechanism can be used to
control other sampling devices in a sampling program by means
of micro-switches which are attached to the moving orifice
mounting bracket.  These micro-switches can be activated by
small brass bars mounted in positions as desired so as to close
electrical circuits at specific times.  These circuits may be
used to automatically operate other samplers, for instance
cascade impactors, during the streaker sampling period.

     The Nuclepore filter strip can be analyzed sequentially
by the proton beam in time .steps corresponding to sampling time
intervals of two hours or less.  In our accelerator bombard-
ment facility we carry out these stepwise bombardments semi-
automatically so that 90-100 separate bombardments, representing
a one-week streaker scan in 2-hour time steps plus blank
filter analyses, may be performed in as short a time as two
hours.  Although individual samples may require longer PIXE
analysis times owing to low concentrations or other special
considerations, the procedure does permit rapid scans of
relatively large numbers of streaker samples so that data sets
of many hundreds of individually analyzed time steps over a
sampling network may be carried out.  An illustration of the
•streaker sampler used in this program is shown in Plate 5b.
The appearance of a typical streaker sample is shown in Plate 6
as it is being loaded into the vacuum chamber of the Van de
Graaff accelerator for PIXE analysis.
                               20

-------
                             Plate  6
   Figure 2.4  A streaker sample being  inserted into the sample holding
   chamber in  preparation for proton bombardmemt.
Source:  B-ll.
J. William  Nelson, Proton-induced  aerosol analyses:   methods and
samplers, in X-Ray Fluorescence  Analysis of Environmental
Samples, Thomas  G. Dzubay, editor,  pp.  19-34, Ann Arbor Science
Publishers,  1977.
                                 21

-------
                            SECTION 5

     CLUES TO THE SOURCES OF TRACE METALS IN URBAN AEROSOLS

     Direct measurement of concentrations of trace metals in
the atmosphere of a city does not in itself demonstrate the
presence of air pollution.  Trace metals all occur naturally
as well as the result of emissions from polluting processes.
In order to demonstrate the presence of a pollution component
we must make use of characteristics such as particle size
distribution and time and'location variability of the concentra-
tions of individual trace metals.  In addition, meteorological
and statistical analysis of the concentration data needs to
be carried out so that associations with measures of atmospheric
transport dynamics and relationships among the various trace
metals may be used as clues to the kinds and locations of
pollution sources which may be present.


EVIDENCE FROM PARTICLE SIZE DISTRIBUTION

     The city of St. Louis has been used as a case study area
in this investigation, conducted during the Regional Air
Pollution Study, RAPS.  Plates 7 and 8 present the concentra-
tions of a number of trace metals as a function of their
particle size in St. Louis during August 1973.  In these
detailed plots most of the data points represent concentrations
at an industrial area site "BH" (Broadway and Hurck Streets
intersection) in the south of the city, and a central city site
"HI" (Holiday Inn Downtown) during two meteorological episodes
of about 1% days each — I during southerly air flow and
II during generally northerly air flow.  For comparison,
additional data are shown for a north St. Louis site "WW"
(Water Works), North Florida nonurban locations "NFL," and the
marine atmosphere at Bermuda "BDA."  For the element titanium,
which has a large and distinct pollution source a short distance
south of the BH site, individual samples for successive 12-hour
periods beginning in the forenoon and evening of each day,
August 16 to 21, 1973, are shown.

     Titanium at the BH site appears to be principally derived
from the pollution source when sampling was carried out during
meteorological period I.  During this time the generally
southerly air flow transported pollution titanium over the BH
sampling site and caused the concentrations to be generally in


                               22

-------
         K 3
         O  CD
         "-f  £
             X1
H- PJ S CD S3 CO
3 3 H- 4  •   O
   PJ 3  H-    C
X M O  CD  >
to
CO
- H- 1 «< V X1 X" 0
h-1 X* PJ H- cn 3 CD '••
-j CD cn
cnpj>rt4fficn>
• 3 3 O - • 01
0, pi O 3 H
M rt > • » co
O^< V\) '
• cn CD •rd X O
OH- H1 PJ •

• cn c/> H- C 1000
v rt 0 H) 0 IWW
*J • PI 3 4
C < rt PJ cn
c: O f H- 3 H-
CL M 0 0 3 3
- • C 3 «• H-
CD M Cn O X
Q,OO H) • O
cn « >d ?d • -, 100
' /i\ L i > t
CD W • t^ e
" S(K? 4 • "
• H- O O •
•o t- o rt y s e
T) • 3 O PI CD
PJ 3 *d H-
hd H 1 3 3
cn H- H- PJ CD
oo o > 3 3 4 ,„
T M ^' d " -
cn CD O CH
co cn TI CD • i-3
-j- o a s; •
\i |— ' • QJ
O M X
^o • c i s;
M CO rt 4 CD C^
' i ' « -f««Til 1 ' ' ' :
• 1 O -IMM '
' i * -IT* i '
I f *• "**••• k.
• x! • -*> a •
; • - IMte /
1 * D -XMMM! &
• ~ tOft Y . >% ]x

• A**I ** ; I,'' ° !

• ». . ... tS
•TL§/7i , x i
/ * • T •
, ' " f •
v' ^
A j 5
^
^

i ,/ f * ;
* T / 4^
• 4 	 X jl^L
T C '^^
' L f V

• * J / J
• Bx ' * '
/ \ , £
i \f/T Ti-BH
fA
I
• > i


IOOO








100





E
^
c 10






|




/\ m
_. 22^ • T S •
o •*•• M •
« NI-I
4 Nl-l* i
— <*>MM« 1 —
A NI-I I ;
» • WW I .
- i NFL
• « BOA

. m.vn . A * -
j

»T 1
1 T
T I A
^ 1 ! ^
* A A 1 "
. ^ T 1
' I 0 "
i T '+
'it * A
. * i ^ j
^ i ' ; f 4
: 9 -A. i /
J JL
T, 1, ^^' -
; tr v -j ° *i
• jT f

— * —
~ • ~
• '.
* •

1 1 t 1 1 1


IOOO












100


10
E
gi
e



10


-'•





• BN-1
O W«-I
- A HI-I
. A HI-B
• WW
" i NFL
• BOA

«TL«/T»

•




%
_ T
1 1
T <
TT 1
• M J;
Ti

^

_





•
*
3 ' o'<< cn'B' ' 6543ZI w-*«S4S2l z- 	 j 	 	 	



Fe * -

.
i -T -

T j •

A. . •

T I
4 ^M
j^L
• • •
ri x
jlr j •
Mr >
11 '
•

i
•
«
_
.

























TJ
I-1
pJ
rt
CD
^j
PJ







4321
gS?3 g « cn
IT) 4 rt H- cn


m rto.cn 3 § Fl'°- 2- Concentration of titanium and iron as a function of particle
cn rt<< H-D.- size in Gt. Louis aerosol, 16-22 August 1973, at BH and HI sites.
™ " * § r_, ># Means of individual measurements for periods I and II are shown as
2^ a s: M dotted lines for Ti at BH in the left diagram. Other diagrams show
CD ro < • • means with standard deviations for periods I and II at BH and
HI, with
s ' § < HI n and d series means also plotted separately for Ti . Averages of
£ E} samples taken at locations WW, NFL, and BDA are also shown.
cn



-------
H X1 X
CD H- I
-J ^ »
cn X1 PJ
  PJ >
  3 3
  CL (U
  cn s; cn
  H- H- 4
  cn 3  H-
    o  n>
  r+ 3J X1
  O CD  CD cn CD
    cn  3 CD
  H- H--  M
    CD
             f-•
           > 4
           x- o
       CD
  • cn
  O H-
  • cn
         cn
       ffi cn
    •*  •  o
       O 3
    J> •  »

           X O
       tr1 M PJ •
       O H- C
                                             K>
                                              i
    o e
       Hi O
       3 4
  CD  H
  CL oo
  d  t-
     Tl
  cn H«
  OD O
  oo X"
  I  H
  Cn (D
  co cn
    o
    •
    C/>
  CD
       r+ P) cn
       H. 3 H-
       O 3 3
       3 -  H-
    H-0 X
    OH)-  O
    3   X) -
    PJ >d •  tr1
    M ^
       o o
    > H- 3J 3
    H- O PI CD
    H. 3 T3 H-
       I  B 3
    T) H- PJ CD
    033
          •- r+ ^
03
  CD
  3
  C
  3

  T) 4
  4 fD
  CD H-
  cn r+
  cn *
    £ o
    H- (D
    H. a
    o
    3 X
             oaE
             CO
    H- PJ
       SJ
       CD
         O
         3"
    CD
  CD 00
    > cn
    D- cn
    < H-
    PJ O
    3 3
    O
    CD PJ
    cn 3
cn
O
3  3
»  cn
   cn
pi  o
3  3
                            Ti/Fe
                 m
                                -
                                     t
                     694321
                         8TA«C







u*w%
moo
too

10


1
1


















»
I
-
•












• MM
o wi-a
* Ml-l
A Mi-n
• ww
s NFL
* it>^

rn. S/TS

'}
1
*1
•
•

6 5



















*
i
;



4
STi
& 1 <







i
T
Co

«_
4
A
_ W -
* " !
T
§
1 ' * : •
1 ' '
t 1 .
I -
b
A
—



^321-
LOP
4Ow







1




(



i
i

!
i
! ,


S
9



01
• 1

D A9
w*w C











1 1
[
'' f
^-* i i
1 ~





_ '
•
: Ca/Fe
i i
6 5











I i
H
H?
i

JL 4 >
,







4 3
STA4C



-







i II
if
ft
•





• Ml-l
OWI-«
AHI-I
4HI-I
• WW
>nn,
• MA
ITL »/n
2 1











&'
«
"


.






«












i
1
i
Plate
-*J
(__j
Cr









O CD
4 S  H-
X1  I  3
             PJ
             3
                      Fig.  3.  Mean  values  for concentrations of calcium and weight ratios
                      of  titanium  and calcium to iron as a  function  of particle size.  Average
                      composition  of soil(s)  and earth's crust(c) are shown for Ti/Fe and  Ca/Fe,
                      Upper limits to means  are indicated by^curved  marks  above points.

-------



1000

E
^
?
100


20
•
•
• i
f t J '

: f } . i

•

r
I |
! s 1
.
1 1 1 1
6543
STAGE


-5-

•
L:.
STL


i
.4
I
A

,
2



IM-I
Ml -X
HI-B
WW
NFL
• 0*
•/rs

tr
1
i
i 6


,
i


Plate 8a


-
•
-
•
_
-
-





10

9
N,
,



0.2


•*]
• 1 '
•
• M-l
0 «*-•
A Ml -I
a Ht-l
• WW
• nn.
• §M
I ITL »/n
. f

T
I
1 * •
J

• S/ Fe a H 4


5

:

•
-

•

-
.
•
*
•


100




10
a
•}
K
- T 4i
'• i 1
; . • "
T ] ,

' 1 l 1 *
T . f *
- I : ,
'.BH-ll
u c BM-B'
, A Ml -I
. . i ^ Hi-n
i > ww
* | • NFL
STU/TS
	 1 	 1 	 1 	 1 	 1 	 1 	
.

-




-
-
            543
               STAGE
                                     654321
                                          STAGE
                               O.I
                                                 WW :
                                                 NFL 1
H    il
                                                 H
                                 r K/Fe
     654321
        STA6C
                                                            I
                           CO
                           O •
                           •f™ Q^
                           •»-> N
                           (O -r-
                           i. CO

                           +-) 0)

                           cr> o


                           'it
                             n)
                           TJ Q.


                             O
                           CO
                           c c
                           O O

                           •M -M
                           (O O

                           4-> 3
                           C «»-
                           
                                                            •— O
                                                             (O Q.
                                                             C C
                                                             
-------
                          Plate 8b





100

10


1



•


•
B





• BN-I
0 SH-B
* HI-I
A Hl-n
• WW
i Hri
• 80*
ru. «/73
r

•








•
<





Zn


hH
iiM
f ( t ..

k

>
[ -
• :
A
« i
i
A





1

9
O.I

0.02


• IM-I
0 M-l
L * Ml -I
4 A HI ••
Iff • WW

Tl 1 *TL */r>
Till
Pi
Y A U ^
T ^ O
" \ *r i
- Zn/Fe *
I
- /s=aoo2 I 1
— — i 	 ± 	 i 	 i 	 1 	 H





-
:

•*
•
-
     654  321
          STAGE
                                      654
                                           STAGE
  100
m
 £
  10
                 Pb
                STL e/TS
       6
                             10
                            O.I
                                            STL8/73
                                  Pb/Fe
                                                $
                                 6
 4   3
STAGE
                                                          c
                                                          o
                                                          if}
                                                          o
                                                          (O
                    O)
                    •r-
                    
                     1-
                    n) (d
                    S- CL
                    4J
                    C 4-
                    0) O
                    O
                    C C
                    o o
                    U T-
                     -p
                    OJ
                    => (O

                    (O (/)
                    > <0

                    C T3
                    (O  (O

                     • o
                    en c
                    •r™ •»—
                    U- N
              54321
               STAGE
 Source:  A-13.
 R.  Akselsson, C.  Orsini, D.L. Meinert, T.B. Johansson, R.E.
'Grieken, H.C. Kaufmann, K.R. Chapman, J.W. Nelson,  and J.W.
 Winchester, Application of proton-induced X-ray  emission
 analysis to the  St.  Louis Regional  Air Pollution Study, Advances
 in X-Ray Analysis,  vol. 18, W.L.  Pickles, C.S.  Barrett, J.B.
 Newkirk, and  C.O.  Ruud, eds., pp.  588-597, Plenum Press, New
 York, 1975.
                              26

-------
the range of 1,000 ng/m3 or above.  Particle sizes containing
most of the titanium were greater than 1 ym aerodynamic diameter,
i.e., impactor stages 3, 2, and 1.  During meteorological period
II, generally northerly air flowj the titanium concentrations
decreased to 100 ng/m^ or less and were principally found in
particles of diameter greater than 1 ym, impactor stage 1.  The
•latter values may represent a mixture of natural dust and other
pollution titanium, e.g. from fly ash, whereas the former repre-
sent primarily contributions from the unique titanium pollution
source.  This source, which emitted Ti02 particles from a paint
manufacturing process, apparently added smaller sized particles
to the atmosphere than did the other dust sources.

     At the HI site it was found that the alternating 12-hour
sampling during meteorological period I showed more than 10
times higher concentrations during the AM ("n," night and
morning) than the PM samples ("d," afternoon and evening).  The
former high concentrations we attribute to fumigation of an
overlying polluted layer during early morning break-up of an
atmospheric temperature inversion.  This interesting case.of
vertical pollution transport in the atmosphere of St. Louis is
discussed further below.

     The particle size distribution of iron in the St. Louis
samples is shown in Plate 7.  At the HI site concentrations are
not strictly dependent on wind direction as indicated by
similarity between meteorological periods I and II.  During
period I at the BH site iron concentrations were higher, but
during period II they were similar to those found at HI.  Iron
is known to be emitted by steel mills and other sources in the
urban area of St. Louis.  No single unique source for iron is
identified as is the case for titanium.  However, one of the
pollution sources may be locally very important in the south
St. Louis industrial region and account for the high BH values
during period I.

     It is significant that iron in the smaller particle size
fractions, <1 ym diameter, stages 4, 5, and 6, is found at
relatively much higher concentrations than is found in the
North Florida aerosol, NFL, or the presumably terrestrially
derived iron at Bermuda, BDA.  The coarse particle iron
concentrations in St. Louis are generally only somewhat higher
than in the relatively clean North Florida atmosphere; however,
the fine particle iron appears to be present in much greater
abundance.  This result suggests that the impact of polluting
activities on the content of iron in the atmosphere of St. Louis
is greatest in the sub-micrometer respirable size range.

     Plate 7 includes a plot of the ratio Ti/Fe for the
different sampling locations and periods.  In clean atmospheres,
as exemplified by North Florida and Bermuda, the ratio Ti/Fe
is very close to 0.1, the average composition of soil and the

                               27

-------
 earth's crust.  During the northerly air  flow period  II  in
 St.  Louis, the ratio  is  significantly below this  value
 suggesting the presence  of additional iron, such  as from
 steel mills, which does  not contain relatively as much titanium
 as average earth crust materials.  The depression of  Ti/Fe
 below 0.1 is more pronounced for sub-micrometer particles
 than for larger sizes, in keeping with our observation earlier
 that the pollution impact on iron concentrations  was  greater
 for  smaller particle  sizes.  Also shown in this figure are  the
 very much higher Ti/Fe ratios, often in excess of unity, when
 the  localized titanium pollution source made its  greatest
 contribution.  The particle size distribution of  Ti/Fe indicates
 that the pollution titanium is present in larger  particles  on
 the  average than is the  case for pollution iron,  although
,both elements' pollution contribution may be mainly in smaller
 particle sizes than that of soil dust.

     In Plate 8 the particle size distributions of zinc  and
 lead contrast significantly from those of iron.   Both elements
 are  enriched strongly above earth's crust values  relative to
 iron, and zinc to a greater degree than is seen in North Florida,
 which suggests that pollution sources predominate over natural
 sources in St. Louis.  Both elements are  found to a much greater
 extent on smaller particle sizes in the sub-micrometer
 respirable range than is the case for iron.  These results
 suggest the importance of pollution emissions at  high
 temperatures, which may  favor the formation of smaller particle
 sizes and more efficient transfer to the  atmosphere through
 stack emission control devices than would be the  case for more
 refractory and coarse particulate ash.  Sulfur likewise  shows a
 tendency to be found  on  smaller particle  sizes, although a
 secondary mode of sulfur is found on coarse particles.   The
 elements potassium and calcium contrast with each other, in
 that potassium appears to have a greater  abundance in fine
 particles relative to coarse than is the  case for iron,  whereas
 the  particle size distributions of calcium and iron are  similar.
 .•These results suggest that some potassium may enter the
 atmosphere with combustion gases or fine  particles formed from
 them, whereas calcium may be essentially  nonvolatile  even at
 high temperatures.

     From these particle size distributions we have been able
 to formulate lines of further inquiry concerning  possible
 sources of these trace metals in the atmosphere of St. Louis.
 The  ideas presented for  possible sources  should be used  as  the
 basis for more detailed  investigation and verification of
 the  hypotheses stated.

     Plate 9 is a more detailed presentation of the time
 variability of titanium  and other elements at station HI
 mentioned above.  We  have plotted, as a function  of time every
 12 hours from August  17  to 22, 1973, the  concentrations  of

                               28

-------
.coo
to
     o  -••
     S.
       o ro
       <-h £U
     t—   o
     O O O
     C -h 3
     -"•   O
     «/) l/> O>
     >.  Q> 3
       3 c+
     —••a -5
     ~~l —' O>
     i  n> n-
     ro   ->•
     ro o o
       o 3
     in n> o
     C O -h
     «/) rt-
     rt -•• O
     »•  O O*
       3 '—•
     —•   O
     003 3
       d- -5
         O •

       O »

       -i. 01
       a. 3 -
       f\t ^^
       3 <-t
       3 DJ
         3
        I 0>
         in
 100
               :o
                 H.i. .
          .8/73
           Cox 0.2
6(7?,^. ,STAP^ . ,e/,7^,Hf . jSTAftEI. .  .9/73 HI. .?TAgc?
                       Co
                                                       Cox 5
                                                          Cox 10
                                                                 .  O-  -CX
dndndndndnd  dndndndndnd  dndndndndnd  dndndndndnd
17 A  19 20 21    17  18 19  20 21    17 18  19 20 21    17  0 19 20 21
FRI      MON      FRI       MON     FRI      MON     FRI     MON
                                                                                         CoxlO
                                                                                         Fex2
                                                                   dndndndndft
                                                                    17 It  19 2O SI
                                                                   FRI      MON
                                                                                                 , H-
     Source:   A-12.
     John  W.  Winchester, Dennis  L.  Meinert,  J.  William Nelson, Thomas B. Johansson,  Rene E.
     Van Grieken,  Celso Orsini,  Henry C.  Kaufmann, and Roland Akselsson, Trace metals  in the
     St. Louis aerosol, Proceedings of the  Second International  Conference  on Nuclear
     Methods  in Environmental  Research, July 29-31,  1974, University of Missouri, Columbia,

     J.R.  Vogt and W.  Meyer, eds.,  USERDA CONF-740701, pp. 385-394, 1974.

-------
titanium, iron and calcium for each individual particle size
fraction collected by impactor stages 1 through 5.  (Impactor
stage 6, the particles smaller than 0.25 ym collected by a
filter did not yield consistently measurable amounts of these
elements and therefore is not plotted.)  It is seen that the
alternation of concentration for titanium for meteorological
period I extends over a factor of 30 for stage 3, 1-2 ym
particles, and nearly that much for larger and smaller sizes.
For iron and calcium a similar, though not as drastic,
alternation is seen.

     The nature of the summer St. Louis atmosphere is such that
stratified pollutant laden air layers could develop over the
city at night.  After sunrise these may mix with surface air
as the result of ground heating and vertical turbulence.
This may lead to a highly concentrated pollution layer, to dip
suddenly downward over the sampling site, a process termed
"fumigation," and increase the concentrations momentarily to
very high levels.  Apparently during meteorological
period I night time emissions of titanium and certain other
metallic pollutants were carried over the sampling site and
caused the average concentration during the period "n"
(midnight to noon) to become very high due to the influx toward
the surface in the morning hours.  Titanium therefore has
proved to be a convenient "natural" tracer of air-pollution
transport in the city.
METEOROLOGICAL ANALYSIS OF THE TITANIUM POLLUTION PLUME

     The element titanium emitted as a fine particluate
pollutant in the atmosphere of St. Louis is a useful tracer
for pollution transport over the entire city.  From a detailed
analysis of the correlation between its concentrations at various
sampling locations and measurements of average wind direction,
directional variability, wind speed, vertical temperature
gradient, and distance from the pollution source, a good
understanding can be obtained concerning the factors which
control transport of the titanium pollution plume.   If this
understanding is sufficiently detailed, the transport of any
other aerosol pollutant generated in the city could be
predicted.

     Using the streaker sampler we have measured the time
variability of titanium concentrations with a 2-hour time
resolution at a number of locations around the city, utilizing
the Regional Air Monitoring Stations of the Regional Air
Pollution Study as sampling points.  This time resolution is
comparable to that of the meteorological parameters obtained
at these stations.
                               30

-------
     Plate 10 shows as examples the trend of titanium concentra-
tion seen at two of the Regional Air Monitoring Stations.
Station 105, located 11 kilometers to the NNE of the titanium
pollution sources across the Mississippi River in Illinois, shows
concentrations reaching almost 1,500 ng/m3 although most of
the time the concentrations are only a few tens of ng/m3.
At station 122 some 60 kilometers to the north in Illinois,
the titanium fluctuations are also seen although the peak
values are 10 time.s lower.  At other stations similar fine
structure to the titanium time variability is observed.  The
sharp maxima seen over short intervals of time are presumed
to represent transport of the titanium pollution plume on the
sampling site.

     It would be useful to know if the plume transport can be
understood in terms of routinely measured meteorological
variables.  If so, we could develop an ability to forecast
pollution transport of any trace metal in plumes.  Four
stations were selected for a meteorological analysis of the
time variability of titanium.  These stations are arranged
in a line generally running from the titanium source northward
and at a distance of 5, 11, 21 and 35 kilometers for stations
111, 105, 113 and 121 respectively.  Since the titanium
pollutant is emitted from stacks which are about 35 m in
altitude, whereas the samplers were placed at about 10 m
height, both vertical and horizontal transport must occur for
the titanium to reach the sampler.

     The meteorological analysis employed consisted of first
sectoring the concentrations measured over the one-week study
period into groups according to the local wind direction
observed at each sampling station.  The sectoring scheme
differed somewhat among the four stations so as to provide a
suitable number of observations within each sector for the
statistical analysis.  These sectoring schemes are shown on
Plate 11.

     Treating the observations within each sector as individual
subsets, it was generally found that concentrations were
highest in the directly downwind sector (5 for stations 111 and
105 and M- for stations 113 and 121).  However, the concentra-
tions were not uniformly high in these sectors and sometimes
high concentrations were observed for the nearby wind direction
sectors.  Therefore interactions between the sectors and the
meteorological measurements of standard deviation of wind
direction (i.e., directional variability), wind speed, and
the temperature differences between 30 m and 5 m (a measure of
temperature gradient) were examined.  In Plate 11 a summary of
these correlations is presented.  The sign, positive or
negative, of the parameter b indicates qualitatively whether
the correlation is direct or inverse.
                                31

-------
                               Plate 10
              i tatata-
              2S0
         STHT I DM
           t B3ST
    IB   17   IB   19   ZB   II   22  23  2H   2f

                
-------
              Station III
                                        station DS
CO
CO
                                                                    TABLE IX.  Final form of the Interaction Dummy
                                                                             Variable Multiple Regression Model
Station 111
Variable b 3ig
SECTORS
S3BYSPD
SECTOR4
S4BYSPD
S4BYGRAD
SECTORS
S5BYSTD
SECTORS
S6BYSTD
S6BYGRAD
S7BYSTD
Constant
+ 1.86
-0.45
+ 4. 54
-0.85
-0.75
+ 2.12
-0.03
+ 4.70
+ 0.06
-2.39
+ 0.05
+ 1.55
. 000
.006
.001
.014
.105
.000
.055
.023
.111
.046
.000
.000
Station 105
Variable b
SECTORS
S3BYSTD
S3BYGRAD
SECTOR4
S4BYSPD
S4BYSTD
S4BYGRAD
SECTORS
SECTOR?
S7BYSPD
Constant

+ 1.
-0.
-0.
+ 5.
-0.
-0.
-1.
+ 0.
+1.
-0.
+1.

96
OS
77
06
35
05
70
99
43
25
69

sig
.000
.008
.002
.009
.089
.035
.023
.000
.052
.150
.000

                                                                    Station 113
                                                               Variable
                                                                               .721
                                                                                sig
                                                                                                                    r+
                                                                                                                    fD
                                                                                                       = .487
                                                                                              Station 121
                                                                                         Variable
                                                                                                            sig
SECTORS
S3BYSPD
SECTOR4
S4BYSPD
S4BYSTD .
SMBYGRAD
Constant

+ 1. 46
-0.12
+56.73
-3.48
-0.08
-20.96
+ 1.64
R2 =
.001
.111
.012
.009
.001
.017
.000
.797
S4BYSPD
S4BYSTD
SUBYGRAD
SECTORS
S5BYSPD
S5BYGRAD
Constant

+ 0.18
-0.06
+ 0.85
-1.04
+ 0.26
+ 0.40
+ 1.55
R2 =
.000
.000
.000
.075
.006
.069
.000
.903
      Fig. 10.  Sectoring schemes for stations 111, 105, 113, and      ;
              121 with the theoretical source to station plume
              centerline angle represented by the dashed line,
              the sectors numbered clockwise starting with the
              180 degree upwind sector numbered 1, and the degree
              width of the remaining downwind sectors labled on
              the perimeter.


      Source:   D-2.
      Scott Rheingrover,  A Statistical  Model  for Titanium  Pollution  Transport  in  the Atmos-

      phere of St.  Louis, M.S.  Thesis,  Dept.  of Meteorology,  Florida State  University,

      June  1977.

-------
     The calculations indicate that variability in wind
direction is associated with higher concentrations of titanium
in sectors adjacent to the directly downwind sector, implying
horizontal dispersion during plume travel.  However, the
reverse is found for the directly downwind sector, suggesting
plume dilution during variable winds which reduces concentra-
tions .  Higher wind speeds tend to be associated with lower
concentrations at nearby stations suggesting turbulent mixing,
'but at more distant stations wind speed is associated with
higher concentrations suggesting the importance of less time
available for particle settling en route.  The temperature
gradient information indicates that vertical mixing downward
enhances concentration at nearby stations but enhances removal
and lowers concentrations by the time the plume reaches more
distant stations.  The overall variability accounted for by
the statistical analysis is 70-90% for three of the stations
but is less for station 105.  This station is separated from
the titanium source by the waters of the Mississippi River,
compared with the other stations where plume transport
occurs only over land.  The perturbing effect of the water
thus could account for greater variability observed for
station 105 than is seen for the other three stations.  However,
the overall degree of variability at the four stations accounted
for by this analysis is impressive.

     In Plate 12 a schematic diagram is presented for the
apparent plume movements for both vertical and horizontal
directions for the four stations downwind of the titanium
'source.  The diagram, which has been constructed purely on
the basis of the statistical analysis, is in good agreement
with conventional ideas concerning plume transport.  The
titanium "natural" tracer is thus a means of quantifying the
description of a plume as it is transported over a complex
urban area.  The quantitative calibration of a plume model may
permit prediction of the transport of any plume carrying
pollutants across the city.


LEAD, A MULTIPLE SOURCE POLLUTANT IN ST. LOUIS

     Conventional wisdom says that urban lead is predominantly
derived from the combustion of leaded gasoline.  The attention
given to this idea by air pollution investigators has tended
to obscure the fact that lead in the urban atmosphere may be
derived from other pollution sources as well.  In the city of
St. Louis other potential lead aerosol sources include the
smelting of ores, the combustion of fuels, and metallurgical
..refining processes.  The contributions of these sources to
the overall atmospheric inventory of lead in St. Louis must
be determined before we can accurately predict the adequacy of
different emission control strategies on air quality.


                                34

-------
                                   Plate 12
        3
        X
                                             AZ
                                             AZ
                                                  SMALL
                                                 LARGE
            80
            60
            40
            20
                    station ill
                               station 105
                                                station 113
                                                                      station 121
               Ti -»
               samplers —• /,'
                                                21
                                                                       39
                                         KILOMETERS
               Fig. 18.  Vertical (XZ) plume concentration profiles suggested by the AT+2
                      associations.  (Adapted from Strom, 1976)               AZ
            5-
            5.
            10
                                           cr(9) SMALL

                                           ff(e) LARGE
                                                                      3
                   station III      slation 105
                                              station 113
                                                                    station 121
              Fig. 17.  Horizontal (XY) plume concentration profiles suggested by the
                      o(9) associations.
Source:    D-2.
Scott  Rheingrover, A  Statistical  Model for Titanium  Pollution
Transport in  the  Atmosphere of St.  Louis,  M.S.  Thesis,  Dept. of
Meteorology,  Florida  State  University, June  1977.
                                       35

-------
     The time sequence filter sampler has a capability for
Collecting a large enough number of samples at many sites
around the city so that a data set of sufficient size for
statistical analysis may be assembled in a reasonable length
of time.  To test the possibility of resolving lead into
its different components, a weeklong record of lead concentra-
tion was obtained with 2-hour time resolution for 11 of the
Regional Air Monitoring Stations within and immediately
outside the city.  In general 10 qr more elements could be
detected in each of the time step samples, opening the
possibility of using correlations with other elements as the
basis for identifying different sources of lead.

     Plate 13 shows a detail of the time trends observed for
a one-day period, Thursday, July 17, 1975, at station 111 in
an industrial section of St. Louis.  During this day all
elements were found to exhibit large concentration fluctuations,
although not at precisely the same times.  Thus iron and zinc
show a maximum at time step 14 but calcium shows a maximum
at step 15, two hours later.  Of special interest is lead
which shows a step 14 maximum.  Bromine, with which automotive
lead is invariabily associated, shows a maximum at step 13,
two hours earlier than the maximum for lead.  Similarly the
earlier maximum of lead at step 9, corresponding to a zinc
maximum at step 9, does not correspond to the maximum of iron
at step 8 or that of bromine, also at step 8.  Unless auto-
motive lead and bromine can become decoupled in the atmosphere
to such an extent as to cause these time shifts, the data
suggest that lead has other sources as well as automotive.

     A multiple linear regression analysis of lead has been
•carried out to examine the extent of correlation with other
elements, in which each of the elements of the data set is
taken in turn as independent variables.  A summary of the
results is shown in Plate 14.  At all stations in the 11
station network, bromine is the element most highly correlated
with lead.  Its coefficient of about 5 is in keeping with the
ratio of bromine/lead of about 0.2 often seen in urban environ-
ments.  However, the regression analysis continues by examining
.the correlation between residual lead, i.e., the difference
between the lead measured in each time step sample and the
smooth regression curve through all of the data, and each of
the other elements.  At 7 of the stations zinc is seen to be a
significantly correlated element with this residual lead.
The coefficients suggest that the zinc-related lead is about
1/10 as abundant as the apparently automotive bromine-related
lead.  The zinc-related lead could be derived from the smelting
of lead zinc ores.

     As the regression analysis is continued, residuals of
lead over and above the first two smooth regression relationships
exhibit frequent correlations with other elements, including

                                36

-------
CO
-J
        eooor
        4000-
         2000-
6OOOr
                                                  6OOOr
                                                         150
    10
 E
\

 c
                                     RAPSStn.lll

                                     Time Key
          I50r
          100 -
5   IO   15
5   10  15
                                              5   10  15

                                           TIME   STEP
                          10  15
                                                                            10   15
    Source:  A-16.

    James 0. Pilotte,  John, W. Winchester, and J.  William Nelson,  Components  of lead in  the

    atmosphere of St.  Louis, Missouri, J. Applied Meteorology,  17,  627-635,  1978.

-------
106
108
108**
(31-81)
111
113
628
504

307
478
540
428
844

223
336
361
6.40 Br
8.04 Br

7.59 Br
4.37 Br
6.15 Br
-

0.72 Zn
2.34 Zn
0.41 Zn
:

1.33 Ti
2.48 V
0.14 Fe
9.45
-35.03

-57.44
-18.09
113.27
0.987
0.107

0.481
0.658
0.878
-

0.260
0.241
0.028
-

0.146
0.092
0.074
                         TABLE 5.   Multiple Linear Regression Analysis of Lead.*


              PbaV                                                     »?An?AD?9
  Station     ncj/ml   _Pb       Vl_V2       V3       Int       ^V     ^Y    ^V      ^

    102        442    287     5.52  Br   0.64 Zn      -         41.28    0.978    0.003      -       0.981
    104        640    436     5.35  Br   0.32 Zn    0.40 K     -64.48    0.862    0.055    0.034     0.951
    105        464    330     4.76  Br   1.09 Zn      -         91.76    0.692    0.189      -       0.881
                                                                                                 0.987
                                                                                                 0.107

                                                                                                 0.887 3
                                                                                                 0.991 &
£   113        540    361     6J5  Br   6".41 Zn    6."l4 Fe    113^27    6!s78    o".628    6!o74     o!980 'Jo
    121        381    501     7.22  Br   0.67 Zn    0.67 Fe    191.24    0.441    0.074    0.250     0.765 M
    122        133    142    11.79  Br   0.84 K       -        -86.72    0.917    0.032      -       0.949 -P
    123        115    108     7.69  Br     -          -        -16.64    1.000      -        -       1.000
    125         82     63     6.51  Br                -        -12.92    0.9555     -        -       0.955
         *The columns list station number, arithmetic  mean  Pb concentration over all  time steps,
          standard deviation of  the distribution of measured concentrations around the mean, first,
          second, and third independent variables in the  regression equation with their  coefficients,
          zero intercept on Pb coordinate, and contributions of first, second, and third independent
          variables to overall R2 given in last column.

        **Anomaly of time steps  1-30 excluded.

     Source:  A-16.
     James 0. Pilotte,  John W.  Winchester,  and J.  William Nelson,  Components of  lead  in the
     atmosphere  of St.  Louis,  Missouri,  J.  Applied Meteorology, 17,  627-635, 1978.

-------
potassium and iron, which may be typical of fly ash or other
earth crust materials.   Such a correlation may be expected if
combustion of coal is a significant contributor of lead to
the St.  Louis atmosphere.

     Station 108 is anomalous in showing little correlation
with any element other than bromine, and in fact the bromine
correlation is poor.  During the first 30 time steps (60 hours),
lead concentrations were singularly high, although concentra-
tions of other elements were not especially high.  When the
regression analysis was repeated for the following steps only,
a regression result similar to many of the other stations was
obtained.  The anomalously high lead concentrations were
subsequently found to be due to emissions from a secondary
lead smelter which did not carry other associated elements
into the atmosphere.  Thus a new fourth atmospheric lead
component in St. Louis was identified at this station.

     Examining the values of R^ in the last column and the
partial  AR2 values of the first, second and third regression
results  in the previous columns shows that on the whole a very
high degree of variability in lead is accounted for.  In the
most nonurban stations the bromine correlation is found to
account  for nearly all of the variability.  In such nonurban
areas automotive emissions predominate over industrial sources
of lead.  In contrast,  in the more urban locations the bromine
correlation usually accounts for less of the overall variability
of lead, and additional elemental correlations are found in
general  to account for most of the rest.  Even for station
108 this is true during times when the secondary lead smelter
emissions did not blow toward the sampling site.

     By  carrying out a multiple linear regression analysis a
trace element fingerprinting of the lead from different sources
has been obtained.  If this result is compared with the
transport modeling result of the titanium plume study described
above, we see that it may also be possible to trace the
transport of lead plumes across the city.  Thus the complexity
of the atmospheric trace metal content of the city of St. Louis
may be unraveled and understood by judicious choice of sampling
methods  and data analysis techniques.
                               39

-------
                            SECTION 6

        AEROSOL PROPERTIES AND HUMAN RESPIRATORY EXPOSURE

     Much of the interest in trace metals contained in the
suspended particulate matter of the urban atmosphere stems from
the potential for transfer of toxic materials to the human
respiratory tract.  We have already stressed the importance
of particle size as providing clues about the sources of
airborne trace metal pollutants in the city.  The particle
size is also critically important in regulating the transfer
of trace metals to different parts of the human respiratory
tract.


RESPIRATORY RESPONSE IN AEROSOL INHALATION

     When aerosol particles are inhaled, two distinct responses
may occur:   the first is the effect of the high humidity of
the respiratory tract on the incoming particles, possibly
causing a sudden particle size increase if the inhaled air
is initially dry and if the particles themselves are
hygroscopic.  The second is the deposition of incoming aerosol
particles in different parts of the respiratory tract, with
an efficiency which depends on particle size after humidifica-
tion.  The largest particles tend to be removed from the
inhaled air by impaction in the nasal passages and the ciliated
regions of the upper respiratory tract.  However, particles
smaller than about 1 ym diameter may bypass these regions
and enter the pulmonary region.  There they may be deposited
on the lung walls, but some particles are not deposited
and may be exhaled again.  Particles that are substantially
smaller than 0.25 ym are deposited by their Brownian diffusion
to the walls of the pulmonary sacs and retained in the lung;
the efficiency of deposition by diffusion increases with
decreasing particle size.  Therefore viewing the lung as
a whole the largest particles and the smallest particles
may be scavenged from the inhaled air in the upper and the
lower regions, respectively, by these two mechanisms, but
intermediate sized particles, in the region of a few tenths
of a micrometer diameter, may be retained by the lung to
a lesser extent and exhaled with the next breath.  Thus
the efficiency of particle deposition in the respiratory
tract exhibits a minimum in the 0.5 ym size range.  To determine
the position of the minimum experimentally for a specific aerosol


                               40'

-------
type requires that the effect of humidification be clearly
distinguished from the effect of deposition so that the
effective size of the particles within the lung can be
determined.

     The PIXE technique is sufficiently sensitive so that a
human being breathing normally polluted air for a matter of
minutes into a cascade impactor can exhale enough trace metal to
be collected by the impactor stages and be detected by PIXE.
This makes it practical to measure simultaneously the concentra-
tions of trace metals in particle size fractions collected by
identical cascade impactors sampling the ambient atmosphere
and the exhaled human breath.  The ambient atmosphere should
preferably be brought to a high relative humidity corresponding
to that in the lung before the sampling.

     On Plate 15 the results of an experiment are shown comparing
the particle size distributions of 8 elements in welding fume
aerosol sampled in the unhumidified inhaled air and the
humidified exhaled air.  The prominent feature is a twofold
increase in average particle diameter owing to the humidifica-
tion effect.  In addition, careful numerical analysis of 9
similar tests shows that particle deposition in the respiratory
tract also occurred to the extent of 20 to 50% depending on the
element.  The lower part of Plate 15 shows the results of
another of the 9 tests, this being an anomalous run in which
the particle size shift was not observed.  The anomaly is
probably associated with a slight temperature difference
between the exhaled breath temperature and that at which the
cascade impactor was maintained, which would cause a difference
in relative humidity in the sampling device and possible
dehydration of exhaled particles.  This result emphasizes
the necessity of exact humidity and temperature control during
the sampling of exhaled human breath.

     Plate 16 illustrates the sampling arrangement used for
experiments of this kind.  Cascade impactors inside a heated
area are maintained such that there is no humidity change
during the passage of exhaled breath through them.  Outside
the heated area the ambient aerosol impactors may be used
without humidity adjustment if the ambient humidity is high;
otherwise prehumidification of the incoming airstream is
required.

     Plate 16 shows the results of the sampling of automotive
lead aerosol under conditions of high outdoor humidity (during
rain).  The lead concentration data, as well as data from a
similar chalk dust experiment, can be used to calculate the
concentration ratio exhaled/ambient.  This ratio is a measure
of the fractional deposition on inhalation for each particle
size range.   Minimum deposition is clearly seen for impactor


                               41

-------
                              Plate  15
                ;
4 i ' t i i  i ' t i  j
   Catcodt Impact or Stagt
                                         t t
        Fig.6(a).  Particle size distributions of trace elements
           in  inhaled and exhaled air for Run 5.   Metallic
           elements generally show a shift toward larger particle
           sizes in exhaled air.
                                    t t  < i ' >  > 4 > ' 1 t i  i ' t I  J
         Fig.6(b).  Particle size distributions of trace  elements
            in inhaled and exhaled air for Run 8.  No systematic
            shift  toward larger sizes is found, and exhaled
            concentrations are generally lower than inhaled.
Source:   A-25.
K. Roland Akselsson, Georges G. Desaedeleer,  Thomas  B.  Johansson,
and  John W. Winchester,  Particle  size distribution and  human
respiratory deposition of trace metals in  indoor work environ-
ments,  Annals of Occupational Hygiene, 19,  225-238,  1976.

-------
-p
CO
               CD CO
           CO  CD  O
        S H- O  C
        .
        fu
        3
              oq  O
           O  CD  fl>
           H) en  ••
        CD fu  CD
        CD H-  •   cd
        rt 3     I
        ^ O  O M
        >••  CO  CD  OO
           *rj  CO  •
        M ^  fb
        P CD  CD

        H- H-  CD
        4 O  M
        O     CD
        3 CD  CD
        •  CD  I

        CO O  CD
        O W  3
        H- O  d.
 CD (U ^J1
 O43
 ^ H-

 O O •
 M I-1
 •  CD s:
       H.

I to p. O
-  N S^
    CD CD
 ID    CO
 ^J Hi r+
 H 4 CD

 l£> O -
 ^J rr
 N5 H- i-3

    § CD
 H CO O
 tD    CD
 •—1 H*

       CD
    CD rr


    (D
    M PJ
    CD 3
    a fu
           _F
                                                       FLOW
                                                      METER
                                                          EXHALED
                                                           BREATH
                                                AMBIENT
                                               --AIR IN

                                                    AL
           Large and Small Cascade Impactors

Figure 1. Diagram for simultaneous sampling of exhaled and ambi-
ent aerosols, showing air ballast arrangement for constant sampling
flow rate, temperature regulator for humidity control, and flow meter
for breathing rate regulation
                                              Coscade  Impactor   Stage
                         Flgur* 3. Concentration ratios as a function of particle size for si-
                         multaneous exhaled and ambient samples
                         Sold points are for separate automotive exhaust and chalk dust experiments;
                         open points are for a mixed experiment. Pb, Q, Br data for stages 6 and S by
                         •mat Impactor, stages 4 ami 3 by targe Impactor; Ca data by smal impactor
                         •xoept where marked by *. Error bars represent absolute standard devto-
                         flons In the ratios
                                                                                                                                                           Ambient

                                                                                                                                                           Enhaled
                                                                                                                                             Stage
                                                                                                       Figure 2. Concentrations as a function of particle size for separate
                                                                                                       automotive exhaust and chatk dust experiments
                                                                                                       Small cascade tmpactors (1 l./mln air flow) were used for aD data poJma ex-
                                                                                                       cept for Pb, Br. Cl on stages 4 and 3 where large Impactor data (12 L/mto air
                                                                                                       ftow) were used. Exhaled air (O) deficit below ambient air (X) represents par-
                                                                                                       tKto loss during breathing. Analytical errors of relative <
                                                                                                       4a£8 than the sizes of the plotted points
                                                                                                                                                  CD
                                                                                                                                                  rt-
                                                                                                                                                  CD
                                                                                                                                                  O>

-------
stages 5 and 4 encompassing the diameter range 0.25-1 ym, a
result which is in close agreement with model calculations
and experiments with synthetic aerosols.  The beauty of being
able to measure this curve directly as described here is
that it can easily be accomplished while breathing ordinary
polluted air.  Possible differences between different polluting
elements can be examined, and a variety of human respiratory
parameters as well as environmental parameters can be tested
for actual atmospheric aerosol pollutants.
THE FATE OF LEAD AEROSOL ALONG A ROADWAY

     Along a roadway lead aerosol is generated primarily by
automobiles which burn leaded gasoline.  As we saw for the
city of St. Louis, most of the automotive lead is found in
the very smallest particles, typically smaller than 0.25 ym
aerodynamic diameter.  The studies of lead inhalation described
in the previous paragraphs indicate that particles of this size
are more likely to be deposited in the lower respiratory tract
than particles somewhat larger in size.  By comparing measure-
ments of the particle size distribution of lead with those of
the efficiency of deposition during inhalation we may estimate,
as a function of particle size, the overall deposition of lead
which occurs in the atmosphere as it enters the respiratory
tract.

     In Plate 17 are presented the results of measurements at
three locations along a heavily traveled roadway in the city
'of Los Angeles.  During the typical afternoon crosswinds
blowing in that region, Site A was upwind and Sites C and D were
immediately downwind of the traffic lanes.  The overall
appearance of the particle size distributions at the upwind
and downwind sites are similar.  However, close examination
shows that the two downwind sites have a measurably higher
relative concentration of <0.25 ym diameter lead than is found
at the upwind site.  For the downwind sites the aging time is
shorter between emission from the exhaust pipes of the nearby
automobiles and sampling at the sites a few meters away.  On
the other hand, lead at the upwind site is more likely to be
due to the general lead aerosol in the city which has aged for
a much longer time between its automotive sources and sampling
at the site.

     These results may be used together with the measured
deposition efficiency curve described previously to estimate
an overall lead fraction deposited in the respiratory tract.
At downwind sites C and D fifty percent more of the stage 6
(particle size class 6) lead is deposited in thr respiratory
tract than for the upwind Site A.  Differences also exist for
the larger particle size classes.  Thus the total concentration
of lead in the atmosphere is only approximately indicative of

                               44

-------
                                     Plate  17
    1.0
    0.3.-
  5 0.11-
  •&
  c
  o

  • aos- -
   0.01

    0.8
                   I    I
           •i
                         KEY
                          • S|TE  A
                          • SITE  C
                          • SITE  D
                                              003 -
                                               O.OI- -
•TJ

"5


o

I  o.r

o
CD
Ik
9

• Q03- -
   Q01
                                              OOOI--
                                                                                      I
                                                                             Sample Sil«
                                                                     Exposition     &  C.
                                                                      Pulmonary    O  •
                                                                      Tracheobronch.  A  A
                                                                      Nosopharyng,   D  •
                                                           Particle  Size  Class
                                            Fig. 3. Predicted deposition efficiencies for lead aerosol at 35 m
                                            upwind (A) and 8 m downwind (C) sites along a Los Angeles
                                            freeway, using particle size distributions  for lead at the two
                                            sites measured by PIXE and published deposition curves for
                                            three regions of the human respiratory tract.
           6
              5432
            Particle Size Class
 Fig. 2. (a) Particle size distribution of lead aerosol, normalized
 .to 1 us, Pb m~3, at the upwind site (site A) and downwind sites
 (sites C and D).
 (b) Fraction of the atmospheric lead at the upwind and down-
 wind site deposited  into the respiratory tract as a function of
 particle size.
Source:    B-20.
Georges  G.  Desaedeleer,  John  W.  Winchester,  and K.  Roland
Akselsson,  Monitoring aerosol elemental composition  in  particle
size  fractions  for predicting human  respiratory uptake,  Nuclear
Instruments  and Methods,  142, 97-99,  1977.

-------
the amount of deposition one can expect in the respiratory
tract, since deposition must also depend upon particle size,
and this will depend upon aging time in the atmosphere.

     Plate 17 also shows the results of combining the measured
particle size distribution of lead at two of the sampling
locations, A and C, with the theoretical predictions for deposi-
tion in three different regions of the respiratory tract:
pulmonary, tracheo-bronchial, and naso-pharyngeal (Task Group
on Lung Dynamics, 1966).  These theoretical curves have not
been verified by experimental methods for lead such as described
above but are assumed to be reasonably good predictions.
The results show that particles of size classes 1 and 2 (>2 ym
diameter) are most likely to be deposited in the naso-pharyngeal
region whereas particles of size class 6 (<0.25 ym) are most
likely to be deposited in the pulmonary region.  A much smaller
fraction of these smallest particles are deposited in the
tracheo-bronchial region.  Therefore the particle size of
the lead in the atmosphere determines not only the overall
efficiency of deposition in the human respiratory tract but
the site of the deposition as well.

     Another feature of the lead aerosol along a roadway which
can be used as an indicator of lead aerosol aging time, and
therefore be of interest in the prediction of respiratory
deposition, is the content of bromine relative to lead in the
aerosol.  Plate 18 shows the time variation in concentration of
lead and in the bromine/lead ratio, every two hours from Friday
until Monday, September 13-16, 1974, at Site A along the same
Los Angeles freeway discussed previously.  At this site,
which was upwind during afternoon hours, the concentration of
lead on a 24-hour basis exhibited a diurnal variation.  In
the afternoon "upwind" condition the lead concentrations were
very low and represented lead, not just from the freeway, but
mostly from other parts of the city.  In the morning and night
times a crosswind in the opposite direction developed, and
Site A became "downwind."  At these times the lead concentration
reached values more than ten times higher, representing
primarily relatively young lead aerosol generated along the
freeway.  The bromine to lead ratio also exhibited a diurnal
fluctuation.  If bromine and lead were equally stable elemental
constituents of the aerosol particles generated by automotive
emissions, the ratio should be invariant with time and total
lead concentration.  However, during the times when lead
concentrations were highest, and therefore the lead aerosol
aging time was shortest, the bromine content per unit of lead
is seen to be highest.  At other times, when the lead aerosol
had been more aged, the bromine content relative to the lead
was much less, suggesting that bromine is released from the
aerosol during aging.

-------
                                      Plate  18
                         10 r
                           -  0.5
                           -  0.2
                       a
                       Q.
                        0.5
                        0.2
                           -0.05
"-0.02
                       Fig. 1.
     1200  2400  1200 2400  1200 2400  1200
    Fri.9/13   Sot.9/14   Sun.9/15    Won.9/16

   Time-dependent  Pb concentration  and  Br/Pb
   weight ratio relationships in the Los Angeles
   aerosol, freeway site A, 1974.
                                         TABLE I

                 Lead and Bromine Concentrations in Cascade Impactor Particle Size Fractions*


Total
Stage 6
5
4
3
2
1
Pb, ng/m3
A
253
139
41.6
33.2
28.9
<28
<28
C
2185
1395
157
179
221
199
<34.5
D
2207
1243
280
177
246
199
61.6
Pb, % mass
A

46.2
13.8
11.0
9.6
<9.5
<9.5
C

63.9
7.2
8.2
10.1
9.1
<1.6
D

56.3
12.7
8.0
11.2
9.0
2.8
Br, ng/m3
A
91
37.8
11.9
<9.2
8.6
10.5
7.4
C
791
510
30.5
51.9
92.7
82.1
23.9
D
707
454
48.7
38.3
77.1
64.3
25.3
Br, % mass
A

41.5
13.1
<10.1
9.5
11.5
8.1
C

64.4
3.8
6.6
11.7
10.4
3.0
D

64.1
6.9
5.4
10.9
9.1
3.6
Br/Pb
A

0.27
0.29
<0.28
0.30
>0.37
>0.26
C
0.36
0.36
0.19
0.29
0.42
0.41
>0.69
D
0.36
0.36
0.17
0.22
0.31
0.32
0.41
•Aerodynamic diameters, Mm, for stages:  >4(1), 4-2(2), 2-1(3), 1-0.5(4), 0.5-0.25(5), <0.25<6, filter). % mass for each stag*
 relative to sum of stages.  Samples collected at Los Angeles freeway sites A (upwind); C, D (downwind), 1530-1730 Thursday,
 12 September 1974.

 Source:   A-17.

 G.G.^Desaedeleer,  J.W.  Winchester,  R.  Akselsson, and K.A.  Hardy,
 Bromine  and  lead relationships with particle size  and time along
 an urban freeway,  Trans.  Am. Nucl.  Soc. ,  21, Suppl.3,  36-37,1975.

-------
     These results are consistent with the suggestion already
made that, during times when Site A was upwind with air flow
from the sampling site to the roadway, the measured lead was of
significantly older age.  This lead also had significantly
lower relative concentration of finest particle lead.  The aging
process then results both in the liberation of bromine to the
atmosphere as a gas and also the coagulation of finest particles
into larger sizes.  The aging of particles and the increasing of
their average particle size is of direct bearing on estimating
the efficiency and location of lead deposition in the respiratory
tract.  The bromine content relative to lead provides a useful
indicator of the extent of aging that has occurred.
LEAD AEROSOL IN THE INDOOR RESIDENTIAL ATMOSPHERE

     Lead aerosol from automotive emissions is seemingly
ubiquitous and occurs both near and far from automobile traffic
lanes.  Whereas there may be other sources of lead in an urban
atmosphere, automotive lead aerosol, even in the city of
St. Louis where smelting and other industrial activities are
carried out intensively, is still the major constituent of lead
It is of some interest to know to what extent the lead aerosol
in the indoor residential atmosphere is automotively derived
and how rapidly the air exchanges between indoors and outdoors
to bring lead from the outdoor atmosphere to the indoor
environment.  In other words, does the house act as a barrier
and cleanse the incoming air of lead?  Or, is there an indoor
source of lead, distinct from automotive, which should be taken
into account when estimating average human exposures over a
normal day of human activity?

     Plate 19 summarizes the results of measurements by
streaker samplers made simultaneously indoors and outdoors in
a residential neighborhood in Denver, Colorado.  During the
one week of sampling that was carried out a meteorological
change by chance occurred outdoors which provides us with the
opportunity of estimating whether indoor sources of lead may
exist.  On the left hand side of Plate 19 the outdoor concentra-
tions of lead, bromine, and the bromine/lead ratio are
presented.  During the first and last days of the week, lead
concentrations outdoors were at values probably typical for
the city.  The same was also true for bromine and the
bromine/lead ratio.  During the middle of the week, however, a
several day period was encountered when lead levels at this
particular sampling site dropped to very low values.  The same
was true for bromine, but, owing to the inability to detect
bromine much of the time, the bromine/lead ratio was usually
indeterminant.

     At an indoor location the lead concentrations during the
initial and final periods of the week were similar to those

                               48

-------
                                          Plate  19
        Figure 3. Time variation of Pb and Br concentrations (in relative unto) and Br/Pb ratio in Denver single-family dwelling. Time steps represent 2-hour sampling
        intervals, starting and ending at 1600 hours October S-12, 1976. Mean value! FD, Br, and TSr/Fo for each of three time periods are indicated by horizontal ban,
                                    ,_  \»trti rtnsto scanted dmrtottara a* vctttcil ban.
Source:   A-23.
D.J.  Moschandreas,  W.  J. Courtney,  J.O. Pilotte,  J.W.  Winchester, H.C. Kaufmann,  J.W.
Nelson,  and  R.M.  Burton, Indoor  and outdoor sources of particulate air pollution  in a
residential  environment, in proceeding, Fourth Joint Conference  on Sensing of
Environmental Pollutants,  1977.

-------
measured outdoors (actually somewhat lower) and the
bromine/lead ratios were also similar.  During the middle part
of the week, however, when lead outdoors was scarcely
detectable, indoor lead concentrations were significantly
higher.  Although lower than the indoor concentrations during
the initial and final parts of the week, they were sufficiently
higher than the simultaneous measurements outdoors to suggest
that an indoor source of lead existed in this house.  The
bromine also was only somewhat lower during the middle of the
week than early and late in the week, in contrast with outdoors
when the bromine level was much less, and the indoor
bromine/lead ratio was actually greater during this period than
that measured either outdoors or indoors during the beginning
and end of the week.

     These results suggest that lead aerosol during normal
times infiltrates into the house with only a small degree of
particle removal so that the indoor lead concentration is
similar to that outdoors.  However, during the middle of
the week the outdoor lead was so low that a secondary indoor
lead source, characterized by a somewhat higher than average
bromine/lead ratio, prevailed.  This source may be secondary
entrained dust from furnishings in the house which had become
inoculated with, one may assume, automotive lead over many
years of occupancy of the house.  Normally, this component of
lead is a small fraction of the total concentration of lead in
the indoor atmosphere.  By chance, however, we encountered
a time when, because of special meteorological circumstances,
the secondary indoor source became dominant and could be
discerned.  Consequently the indoor living environment can not
be assumed to have a simple relationship with the outdoor
atmosphere.  Some pollutants may be infiltrated from outdoors
and some may be generated from within.  In the case of lead
in the Denver house, apparently both source processes are
operative.
                               50

-------
                            SECTION 7

        AEROSOL TRANSPORT OUT OF THE URBAN SOURCE REGION

     In studies of trace metals in the urban atmosphere it is
of considerable importance to have a general understanding of
the composition of cleaner air.  A study of clean air serves
both to provide a baseline or reference for comparison of the
urban atmosphere, in order to determine to what degree the
urban atmmosphere has been polluted beyond natural levels,
and also to facilitate the detection of pollutants transported
outward from the urban region into cleaner areas.  Two case
studies are of a special interest in the investigation of urban
trace metals in North American cities.


THE NORTH FLORIDA COASTAL ZONE

     North Florida is a region containing small cities and
towns and large areas of forested land.  Heavy industry is
virtually absent, and the quality of air is considered good.
Within the eastern United States, north Florida may have an
atmospheric composition closer to natural than any other
area.  Furthermore, the coastal zone is a zone of interaction
of maritime and terrestrial influences and provides an
opportunity to compare the chemical relationships in these
two types of air.

     As in the city of St. Louis, we have sought in Florida
to determine the relative trace metal composition of the
aerosol as a function of particle size.  In Plates 20 and 21
trace element concentrations and their ratios, as determined
by cascade impactor sampling and PIXE analysis, are presented.
These several elements divide naturally into at least three
groups:  elements associated primarily with soil dust; elements
which are anomalously enriched in the atmosphere, including
copper and zinc; and the element sulfur, a class by itself,
which may be largely due to gas-to-particle conversion into
sulfates.  The soil dust elements iron, potassium, calcium,
titanium, and manganese are very nearly in the ratios in which
they occur in natural soils, and all of them occur
preferentially on particles greater than 1 ym aerodynamic
diameter (impactor stages 3, 2 and 1).  This observation is
consistent with the dispersion nature of the soil dust
source, forming large particles preferentially to smaller


                                51

-------
                                         Plate 20

1000

too
en
O)
c 10
1
0.1
100
10

o
1 1
0.1
o.ot
• •• 1 1 1 I.I
Fe

fr
|T | I '
- \ ' •
1 *
1
1 I I 1 1
K/F.


T
Soil ratio ™
12, 12 10. 12. 12,
k ,-» ,"• ,n» ,io» y
i i i i 	 — i —
, S/Fe
T
IT
- *
*' i !
T ' f-
* *

12, 12, 10, 12, 12,
' 95 »6 '°5 95 \~
1 t 1 1 1
Ca/Fe

\
j iM -
% I Soil ratio
10. 12, 10. 12 12
' 35 96 "5 '°5 V
i 1 i I i

100

10
o
IK
0.1
0.01





                    5(321       5(321
                                       IMPACTOR STAGE

    Fig. 3.  Arithmetic mean values of Fe concentrations and of weight ratios for other elements to Fe plotted versus impac-
   tor stage  number, equivalent to a logarithmic scale of aerodynamic particle diameter. The vertical bars represent one
   standard deviation of the mean of the data set. The numbers at the bottom of each graph show the number of nonzero val-
   ues included in the averages. Where this number is less than the corresponding number of samples analyzed (Table I), the
   element was below detection limit in the remaining analyses. If the element was detected in more than half the number of
   analyses, a value has been plotted; otherwise, an upper limit value is indicated. Points for city site A are indicated by open
   circles, for city site B by closed circles, for the forest by open squares, and for the coast by crosses.

Source:   B-25.
Thomas  B.   Johansson,  Rene  E.  Van  Grieken,  and  John W. Winchester,
Elemental  abundance  variation  with  particle  size in  north
Florida  aerosols,  J.  Geophys.   Res.,  81,   1039-1046, 1976.
                                            52

-------
                           Plate 21
10
1

0.1
0.01
0.001
10
1
0.1
0.01
0.001
1 III
Ti/F,
-
**" *
^\8 * <$ *} o*i -
Soil ratio
7i "•> 10i I2i 12->
— ' •*.*» "* H "* A *«•
5I '°4 105 '°5 85
1 lit
Cu/Fe
'I
> .
x . s
1 * °s I'
1 * i r
t r
Soil ratio
/72 63 63 53 53
'• 2, 33 \ \ \'
. . \ I 1 1
I 1 1 1 I
Mn/Fe
-
£ i
-
i i6
-x r / /i i*
Soil ratio °
1 1 1 1 1
| | j | j
Z"/F,
• jj f '
^ ' ir
i li L
YfF
Soi ratio
• — '». 10, 10, 10 12
' 83 65 83 8A V
1 ' i i i
54321 54321
                             IMPACTOR STAGE
                           Fig. 3. (continued)
Source:  B-25.
Thomas B. Johansson,  Rene E. Van Grieken,  and John W. Winchester,
Elemental abundance variation with particle  size in north
Florida aerosols,  J.  Geophys. Res.,  81,  1039-1046, 1976.
                                53

-------
ones by the mechanical action of wind on rock and soil
material.

     The elements copper and zinc have concentrations relative
to iron significantly greate-r than the soil ratios.  The
ratios observed here are typical of those found elsewhere in
clean atmospheres and may represent a natural state of the
atmosphere in which certain elements are "anomalously enriched."
In remote areas the anomalous enrichment is believed to be due
to a natural process not associated with long range transport
of air pollution.  However, to detect one of these elements as
a pollutant above the natural level is complicated by our lack
of clear understanding of the cause of the natural anomalous
enrichment in the first place.  The important task of developing
techniques for monitoring the long range transport of heavy
metal air pollution has been begun by this study of the North
Florida aerosol.

     The element sulfur is in a class by itself.  Present
indications are that all of the eastern United States, including
north Florida, is covered by a large cloud of air pollution
derived sulfate.  This sulfate migrates into the north Florida
area and diminishes in concentration gradually during transport
over longer distances.  The particles are in general of very
fine size resulting from the gas-to-particle conversion
nature of the source process.  Thus, as observed, the ratio of
sulfur to iron increases strongly with decreasing particle
size.  Little evidence for local sources of sulfur are observed
in the north Florida area when compared to the overwhelming
amount of sulfate aerosol transported into the region.

     In conclusion, we may say that the north Florida aerosol
provides a useful reference of relatively clean air for
comparison with more polluted urban atmospheres.  However,
at least sulfur is already a significant contaminant.
AIR POLLUTION PRESENT IN THE NORTH ATLANTIC MARINE ATMOSPHERE
AT BERMUDA

     As with the study of the north Florida aerosol, the
particle size distribution of trace elements in the atmosphere
of Bermuda provides a useful indicator of the presence of long
range transport of air pollution.  Bermuda lies some 1500
kilometers southeast of New York and its atmosphere is
representative of much of that over the North Atlantic ocean
far from the North American continent.  Air pollution is not
visibly present.

     If the aerosol particles are carefully sampled as a
function of particle size under a wide variety of wind direc-
tions and weather conditions varying from stormy to calm, it

-------
is possible by a regression procedure to resolve the ever-
present sea spray component of the aerosol from that of aerosol
particles of other sources.  In an investigation of this
possibility a series of ten cascade impactor samples collected
over a period of three weeks in the fall of 1973 were analyzed
using PIXE for their elemental composition as a function of
particle size.  Plate 22 presents a summary of the findings.

     It was noted that under stormy conditions the large
particle chlorine content of the samples collected was signifi-
cantly higher than that of samples collected under calmer
conditions.  However, for other elemental constituents of the
aerosol, that was not necessarily true.  A resolution of two
aerosol components was made, assuming one to be due to sea
spray containing all of the measured large particle chlorine
and the other consisting of tropospheric, ultimately terrestrial,
components typified by iron, which was assumed to come solely
from this source.  By a graphical regression technique the
ratios of each element to iron, the indicator of the tropospheric
component, were determined for this component as a function of
particle size.  In this way the sea spray contribution to each
of the elements was resolved.  In Plate 22 are plotted the
ratios of the tropospheric component of each element relative
to iron as a function of particle size.  As with north Florida,
the sulfur to iron ratio decreased sharply with increasing
particle size, suggesting that the sulfur may be derived
primarily by gas-to-particle conversion in the atmosphere
whereas the iron is a terrestrial dust component found
predominantly in larger particle sizes.

     The particle size trend of zinc showed similar features,
owing to its being found primarily in smaller particle sizes
than is typical of terrestrial dust.  The origin of the zinc
may be natural or pollution, but in any event is of a size
distribution suggesting that its source process produces fine
particles.  One such source process may be the condensation
of high temperature vapors, but other source processes could
include biogenic evolution of zinc containing aerosol, such
as from organic vapor condensation.

     Calcium is found as a rule on larger particle sizes than
is iron, and the Bermuda data in Plate 22 are typical of many
areas including north Florida (cf. Plate 20).  On the other
hand potassium is found as a rule on smaller particle sizes
than is iron (cf. Plate 20).  Since the contribution of sea
spray related aerosol production has been removed, we examine
in Plate 22 only the net tropospheric or terrestrial component
of these elements.  The large particle size preference for
calcium over iron may be due to the nature of the dispersion
mechanism which generates calcium-bearing aerosol, such as the
dispersion of calcite, from that of iron which may be derived
from different clay mineralic constituents of soil.  In the

                               55

-------
                         Plate 22
   200
    100
O  40
or
I-
i
o
     20
     10
        54321
                           5  4 3  2 I    ""  54321
                       IMPACTOR STAGE
Source:  B-28.
Dennis L. Meinert  and John W.  Winchester,  Chemical relationships
in the North Atlantic marine aerosol, J.  Geophys. Res.,  82,
1778-1782, 1977.
                            56

-------
case of potassium, the significant enrichment over iron in the
smallest particle sizes is as yet an unexplained observation.
A biogenic source of fine potassium containing aerosol is
suggested.

     Studies elsewhere in the world indicate that the sulfur
content observed in fine particles in Bermuda is by no means
typical of the marine atmosphere.  Sulfur in the aerosol
sampled at Bermuda is predominantly in the sub-micron particle
sizes, not the coarse particles greater than 1 ym and typical of
sea spray, and it now appears that the fine particle sulfur
is mainly due to long range transport of air pollution from the
continental areas 1000 km or more distant.  The comparative
study of urban with the maritime atmosphere suggests that
particulate materials may be transported over considerable
distances and contaminate atmospheres far from their sources.
In the case of sulfur the strength of pollution sources is
sufficiently great so that the pollution effect can be seen as
far from the North American continent as Bermuda.  Whether this
is also true for trace metals has not yet been determined.
However, by extending investigations of this kind we may gain a
sharper understanding of the potential for the remote atmosphere
to be contaminated with any air pollutant.  Studies of apparently
clean air serve as an important complement to studies of the
urban atmosphere in understanding the impact of man's activities
on the atmospheric environment as a whole.
                               57

-------
                           REFERENCES

1.   Camp, B.C., J.A.  Cooper,  and J.R.  Rhodes.   X-Ray
     Fluorescence Analysis —  Results of a First Round Inter-
     comparison Study.   X-Ray  Spectrometry,  3:47-50,  1974.

2.   Camp, D.C., A.L.  Van Lehn,  J.R.  Rhodes,  and A.H. Pradzynski.
     Intercomparison of Trace  Element Determinations  in
     Simulated and Real Air Particulate Samples.  X-Ray
     Spectrometry, 4:123-137,  1975.

3.   Liu, B.Y.H. and K.W. Lee.  Efficiency of Membrane and
     Nuclepore Filters  for Submicrometer Aerosols.  Environmental
     Science Technology, 10:345-350,  1976.

4.   Mitchell, R.I. and J.M. Pilcher.  Improved Cascade Impactor
     for Measuring Aerosol Particles  Sizes.   Industrial and
     Engineering Chemistry, 51:1039-1042, 1959.

5.   Task Group on Lung Dynamics Committee II.   Health Physics,
     12:173-208, 1966.
                               58

-------
                LIST OF PUBLICATIONS AND REPORTS


A.   STUDIES WITH MAJOR EPA FUNDING SUPPORT


     Elemental Analysis and Aerosol Sampling Techniques

               Proton Induced X-Ray Emission, PIXE

     A-l    J.W. Nelson, J.W. Winchester, and R. Akselsson,
            Aerosol composition studies using accelerator proton
            bombardment, Proc.  3rd Conf. on Appl. of Small
            Accelerators, North Texas State Univ., Denton,
            Texas, 21-23 October 1974, Vol. 1, The Use of
            Small Accelerators  in Research and Teaching,
            J.L. Duggan and I.L. Morgan, eds., USERDA
            CONF-741040- PI, pp. 139-147, 1974.

     A-2    Thomas B. Johansson, Rene E. Van Grieken, J. William
            Nelson, and John W. Winchester, Elemental trace
            analysis of small samples by proton induced X-ray
            emission, Anal. Chem., 47, 855-860, 1975.

     A-3    M.S. Ahlberg and F.C. Adams, Experimental comparison
            of photon and particle induced X-ray emission
            analysis of air particulate matter, X-Ray
            Spectrometry, 7_, No. 2, 73-80, 1977.

                   Thick Target PIXE Analysis

     A-4    Mats S. Ahlberg, Enhancement in PIXE analysis,
            Nuclear Instruments and Methods, 142, 61-65, 1977.

     A-5    Mats S. Ahlberg, Comparison of yield versus depth
            for particle induced and photon induced X-ray
            emission analysis,  Nuclear Instruments and Methods,
            146, 465-467, 1977.

     A-6    Mats S. Ahlberg, Matrix effects in trace analysis
            by proton induced X-ray emisiion, in Proceedings of
            Fourth Conference on the Scientific and Industrial
            Applications of Small Accelerators, IEEE, 76CH
            1175-9 NPS, pp. 106-109, 1976.
                               59

-------
       Filter and Impactor Sampling of Aerosols

A-7    R.E. Van Grieken, T.B.  Johansson, K.R.  Akselsson,
       J.W. Winchester, J.W. Nelson, and K.R.  Chapman,
       Geophysical applicability of aerosol size
       distribution measurements using cascade impactors
       and proton induced X-ray emission, Atmospheric
       Environment, 10, 571-576, 1976.

A-8    A.P. Woodard, Jr., B. Jensen, A.C.D. Leslie,
       J.W. Nelson, J.W. Winchester, R.J. Ferek, and
       P. Van Espen, Aerosol characterization  by impactors
       and streaker sampling and PIXE analysis, in
       Proceedings of Symposium, Recent Advances in Air
       Pollutant Analysis, American Institute  of Chemical
       Engineers, New York, (in press) 1977.

A-9    Walter W. Berg, A directional air sampling control
       system, Dept. of Oceanography, Florida  State
       University, Tallahassee, FL 32306, July 1974
       (unpublished report).

Urban Aerosol Trace Metal Studies

      Urban Aerosol Trace Metal Characterization

A-10   Thomas B. Johansson, Rene E. Van Grieken, and
       John W. Winchester, Interpretation of aerosol trace
       metal particle size distributions, Proceedings of
       the Second International Conference on  Nuclear
       Methods in Environmental Research, July 29-31,
       1974, University of Missouri, Columbia, USERDA
       CONF-740701, pp. 356-365, 1974.

A-ll   Kenneth A. Hardy, Roland Akselsson, John W. Nelson,
       and John W. Winchester, Elemental constituents of
       Miami aerosol as a function of particle size,
       Environmental Science and Technology, 10, 176-182,
       1976.

A-12   John W. Winchester, Dennis L. Meinert,  J. William
       Nelson, Thomas B. Johansson, Rene E. Van Grieken,
       Celso Orsini, Henry C.  Kaufmann, and Roland
       Akselsson, Trace metals in the St. Louis aerosol,
       Proceedings of the Second International Conference
       on Nuclear Methods in Environmental Research,
       July 29-31, 1974, University of Missouri, Columbia,
       J.R. Vogt and W. Meyer, eds. , USERDA CONF-740701,
       pp. 385-394, 1974.
                          60

-------
A-13   R. Akselsson, C. Orsini, D.L.  Meinert,
       T.B. Johansson, R.E. Van Grieken,  H.C.  Kaufmann,
       K.R. Chapman, J.W. Nelson, and J.W.  Winchester,
       Application of proton-induced  X-ray  emission
       analysis to the St. Louis Regional Air  Pollution
       Study, Advances in X-Ray Analysis, vol.  18,
       W.L. Pickles, C.S. Barrett, J.B.  Newkirk,  and
       C.O. Ruud, eds. , pp. 588-597,  Plenum Press,
       New York, 1975.

A-14   Celso Q. Orsini, Henry C. Kaufmann,  K.  Roland
       Akselsson, John W. Winchester, and J. William
       Nelson, Variation of elemental composition with
       particle size in the St. Louis aerosol,  Nuclear
       Instruments and Methods, 142,  91-96, 1977.

A-15   James 0. Pilotte, J. William Nelson, and John W.
       Winchester, Application of multi-station time
       sequence aerosol sampling and  proton induced X-ray
       emission analysis techniques to the  St.  Louis
       Regional Air Pollution Study for investigating
       sulfur-trace metal relationships,  in Proceedings,
       ERDA X and Gamma-Ray Symposium, Ann  Arbor, MI,
       pp. 161-163 (CONF 760539), 1976.

A-16   James 0. Pilotte, John W. Winchester, and J. William
       Nelson, Components of lead in  the  atmosphere of
       St. Louis, Missouri, J. Applied Meteorology, 17,
       627-635, 1978.

A-17   G.G. Desaedeleer, J.W. Winchester, R. Akselsson,
       and K.A. Hardy, Bromine and lead relationships
       with particle size and time along  an urban freeway,
       Trans. Am. Nucl. Soc. , 21, Suppl.  3, 36-37,  1975.

A-18   K.R. Akselsson, K.A. Hardy, G.G.  Desaedeleer,
       J.W. Winchester, W.W. Berg, T.B.  Vander Wood, and
       J.W. Nelson, X-ray techniques  for  aerosol sulfur
       baseline assessment along an urban freeway,
       Advances in X-Ray Analysis, vol.  19, R.W.  Gould,
       C.S. Barrett, J.B. Newkirk, and C.O. Ruud, eds.,
       pp. 415-425, Kendall Hunt, Dubuque,  Iowa,  1976.

A-19   John W. Winchester, Georges G. Desaedeleer,  and
       K. Roland Akselsson, Time variations in elemental
       composition of atmospheric particulate  matter along
       a Los Angeles freeway, (to be  published).

A-20   Georges G. Desaedeleer, K. Roland  Akselsson, and
       John W. Winchester, Aerosol particle size relation-
       ships along a Los Angeles freeway, (to  be published)
                          61

-------
A-21   William'J.  Courtney, Scott Rheingrover,  James
       Pilotte,  Henry C. Kaufmann, Thomas A.  Cahill, and
       John W. Nelson, Continuous observation of
       particulates during the General Motors Sulfate
       Dispersion Experiment, Journal of the  Air Pollution
       Control Association, 28, 224-228, 1978.

A-22   G. Desaedeleer, J.W. Winchester, J.O.  Pilotte,
       J.W. Nelson, and H.A. Moffitt, Proton  induced X-ray
       emission  analysis of roadway aerosol time sequence
       filter samples for pollution control strategy, in
       Measurement, Detection and Control of  Environmental
       Pollutants,International Atomic Energy Agency,
       Vienna, pp. 233-244, 1976.

A-23   D.J. Moschandreas, W.J. Courtney, J.O. Pilotte,
       J.W. Winchester, H.C. Kaufmann, J.W. Nelson, and
       R.M. Burton, Indoor and outdoor sources of
       particulate air pollution in a residential environ-
       ment, in  Proceedings, Fourth Joint Conference on
       Sensing of Environmental Pollutants, 1977.

A-24   D.J. Moschandreas, J.W. Winchester, J.W. Nelson, and
       R.M. Burton, Fine particle residential air pollu-
       tion, Atm.  Environ., submitted, August 1978.

       Human Aerosol Inhalation and Hair Studies

A-25   K. Roland Akselsson, Georges G. Desaedeleer,
       Thomas B. Johansson, and John W. Winchester,
       Particle  size distribution and human respiratory
       deposition of trace metals in indoor work environ-
       ments, Annals of Occupational Hygiene, 19, 225-238,
       1976.

Marine Aerosol Studies

    Coastal Zone Aerosol and Trace Metal Transport

A-26   Jonathan  Sheline, Roland Akselsson, and John W.
       Winchester, Trace element similarity groups in north
       Florida Spanish moss:  evidence for direct uptake of
       aerosol particles, J. Geophys. Res., 81, 1047-1050,
       1976.

              Air-Water Interface Studies

A-27   Walter W. Berg, Jr., and John W. Winchester, Aerosol
       chemistry of the marine atmosphere, in Chemical
       Oceanography, 2nd edition, vol. 7, J.P. Riley and
       R. Chester, editors, Academic Press, New York,
       In press  1977.   (87 pages + 5 tables + 11 figures).

                          62

-------
B.   STUDIES WITH CONTRIBUTING EPA FUNDING SUPPORT


     Elemental Analysis and Aerosol Sampling Techniques

               Proton Induced X-Ray Emission, PIXE

     B-l    T.B.  Johansson, J.W.  Nelson,  R.E. Van Grieken,  and
            J.W.  Winchester, Elemental analysis  of aerosol  size
            fractions by proton-induced x-ray emission,  Trans.
            Am. Nucl. Soc., 17, 103, 1973.

            Proton Elastic Scattering Analysis,  PESA

     B-2    J.W.  Nelson, I. Williams, T.B. Johansson,  R.E.
            Van Grieken, K.R. Chapman, and J.W.  Winchester,
            Elemental analysis of aerosols using proton
            scattering, IEEE Trans., Nucl. Sci., vol.  NS-21,
            618-621, Feb. 1974.

     B-3    J. William Nelson and Dennis L. Meinert, Proton
            elastic scattering analysis - a complement to
            proton-induced X-ray emission analysis of
            aerosols, Advances in X-Ray Analysis, vol. 18,
            W.L.  Pickles, C.S. Barrett, J.B. Newkirk,  and
            C.O.  Ruud, eds. , pp.  598-605, Plenum Press,  New
            York, 1975.

     B-4    R. Akselsson, J.W. Nelson, and J.W.  Winchester,
            Proton scattering for analysis of atmospheric
            particulate matter, Proceedings of Conference on
            Nuclear Cross Sections and Technology, National
            Bureau of Standards Special Publication 425,
            vol.  2, pp. 484-487, U.S. Depart, of Commerce,
            Oct.  1975.  (Cf. abstract of paper FB-5, Bull.  Am.
            Phys. Soc., Ser. II, 20_, 155, 1975).

     B-5    J.W.  Nelson and W.J.  Courtney, Light element
            analysis by proton scattering, Nuclear Instruments
            and Methods, 142, 127-132, 1977.

                  Computer Spectrum Resolution

     B-6    H.C.  Kaufmann and R.  Akselsson, Non-linear least
            squares analysis of proton-induced X-ray emission
            data, Advances in X-Ray Analysis, vol. 18, W.L.
            Pickles, C.S. Barrett, J.B. Newkirk, and C.O. Ruud,
            eds., pp. 353-361, Plenum Press, New York, 1975.
                               63

-------
B-7    Henry C.  Kaufmann,  K.  Roland Akselsson,  and
       William J.  Courtney,  REX:   a computer program for
       PIXE spectrum resolution of aerosols, Advances in
       X-Ray Analysis,  vol.  19, R.W.  Gould,  C.S.  Barrett,
       J.B. Newkirk, and C.O. Ruud, eds. ,  pp.  355-366,
       Kendall Hunt, Dubuque, Iowa, 1976.

B-8    Henry C.  Kaufmann,  K.  Roland Akselsson,  and
       William J.  Courtney,  REX -  a computer programme
       for PIXE analysis,  Nuclear  Instruments and Methods,
       142, 251-257, 1977.

              Thick Target PIXE Analysis

B-9    R.E. Van Grieken, T.B. Johansson,  J.W.  Winchester,
       and A.L.  Odom, Micro-determination of zirconium-
       hafnium ratios in zircons by proton induced  X-ray
       emission, Z. Anal.  Chem., 275, 343-348,  1975.

       Filter and  Impactor Sampling of Aerosols

B-10   B. Jensen and J.W.  Nelson,  Novel aerosol sampling
       apparatus for elemental analysis,  Proceedings of
       the Second  International Conference on Nuclear
       Methods in  Environmental Research,  July 29-31,
       1974, University of Missouri, Columbia,  USERDA
       CONF-740701, pp. 366-375, 1974.

B-ll   J. William  Nelson,  Proton-induced  aerosol analyses:
       methods and samplers,  in X-Ray Fluorescence  Analysis
       of Environmental Samples, Thomas G. Dzubay,  editor,
       pp. 19-34,  Ann Arbor  Science Publishers, 1977.

B-12   W.J. Courtney, H.C.  Kaufmann, and  J.W.  Nelson,
       Continuous  time sequential  analysis of air
       particulate matter with proton beams, in Proceedings
       of Fourth Conference  on the Scientific and Indus-
       trial Applications  of Small Accelerators, IEEE,
       76CH 1175-9 NPS, pp.  120-123, 1976.

B-13   J.W. Nelson, G.G. Desaedeleer, K.R. Akselsson,
       and J.W.  Winchester,  Automatic time sequence filter
       sampling of aerosols  for rapid multi-element
       analysis by proton induced  X-ray emission, Advances
       in X-Ray Analysis,  vol. 19, R.W. Gould,  C.S. Barrett,
       J.B. Newkirk, and C.O. Ruud, eds.,  pp.  403-413,
       Kendall Hunt, Dubuque, Iowa, 1976.
                          64

-------
B-14   M.S. Ahlberg, A.C.D. Leslie, and J.W. Winchester,
       Environmental and occupational health analyses
       using proton induced X-ray emissions, in Electron
       Microscopy and X-Ray Applications to Environmental
       and Occupational Health Analysis, Philip A. Russell
       and Alan E. Hutchings , eds. , Ann Arbor Science
       Publishers, 1978, pp. 41-46.

B-15   Ronald J. Ferek, Lawrence Wangen, and Thomas A.
       Cahill, Ambient air comparison tests of stacked
       filter units with virtual, multiday, and Battelle
       impactors (to be published).

Urban Aerosol Trace Metal Studies

    Statement of urban aerosol research objectives

B-16   John W. Winchester and Georges G. Desaedeleer,
       Applications of trace element analysis to studies of
       the atmospheric environment, in Nondestructive
       Activation Analysis, Saadia Amiel, Ed., Elsevier,
       Amsterdam, in press, 1977,  (iii + 54 pages,
       7 tables and 10 figures).

B-17   T.B. Johansson, R.E. Van Grieken, J.W. Winchester,
       and J. Bogen, Trace elements in atmospheric aerosol
       in the Heidelberg area, measured by instrumental
       neutron activation analysis (discussion), Atm.
       Environ., 8_, 297-299, 1974.

      Human Aerosol Inhalation and Hair Studies

B-19   Georges G. Desaedeleer and John W. Winchester,
       Trace metal analysis of atmospheric aerosol particle
       size fractions in exhaled human breath, Environ.
       Sci. Technol., 9_, 971-972, 1975.

B-19   A.C.D. Leslie, J.W. Winchester, F.W. Leysieffer,
       and M.S. Ahlberg, Prediction of health effects of
       pollution aerosols, in Trace Substances in
       Environmental Health-X, Delbert D.Hemphill,
       ed., University of Missouri, Columbia, pp. 497-504,
       1976.

B-20   Georges G. Desaedeleer, John W. Winchester, and
       K. Roland Akselsson, Monitoring aerosol elemental
       composition in particle size fractions for
       predicting human respiratory uptake, Nuclear
       Instruments and Methods, 142, 97-99, 1977.
                          65

-------
B-21   E.G. Henly, J.W. Nelson, and M.E. Kassouny, Single
       human hair root analysis by nuclear particle
       accelerator techniques,  in Trace Substances in
       Environmental Health-X,  D.D. Hemphill, eds.,
       University of Missouri,  pp. 333-341, 1976.

B-22   E.G. Henly, M.E. Kassouny, and J.W. Nelson, Proton-
       induced X-ray emission analysis of single human
       hair roots, Science, 197, 277-278, July 15, 1977.

B-23   Mats S. Ahlberg, Harold  J. Annegarn, Bjorn 0.
       Dahlback, Alistair C.D.  Leslie, and John W.
       Winchester, Particle size dependent aerosol
       deposition in the human  respiratory tract, in Trace
       Substances in Environmental Health-XII, Delbert D.
       Hemphill,ed., University of Missouri, Columbia,
       in press, 1978.

Marine Aerosol Studies

    Coastal Zone Aerosol and Trace Metal Transport

B-24   T.B. Johansson, R.E. Van Grieken, and J.W.
       Winchester, Marine influences on aerosol composition
       in the coastal zone, Journal de Recherches  •
       Atmospherique, 8_, No. 3/4, 761-776, 1974.

B-25   Thomas B. Johansson, Rene E. Van Grieken, and
       John W. Winchester, Elemental abundance variation
       with particle size in north Florida aerosols,
       J. Geophys. Res., 81, 1039-1046, 1976.

B-26   J.W. Winchester, Approaches to evaluating dry
       deposition of atmospheric aerosol pollutants onto
       lake surfaces, Journal of Great Lakes Research, 2_,
       Suppl. 1, 33-38, 1976.

B-27   John W. Winchester, Assessing air pollution
       particulate fallout potential for water pollution
       in Lake Michigan, in Proceedings of the Second
       Federal Conference of the Great Lakes, Argonne
       Natl. Lab., March 25-27, 1975, (issued August 1976),
       pp. 330-337.

           Marine Aerosol Characterization

B-28   Dennis L. Meinert and John W. Winchester, Chemical
       relationships in the North Atlantic marine aerosol,
       J. Geophys. Res., 82, 1778-1782, 1977.
                          66

-------
                  Air-Water Interface Studies

     B-29   Walter W.  Berg and John W.  Winchester, Organic and
            inorganic  gaseous chlorine  concentrations in
            relation to the particle size distribution of
            chloride in the marine aerosol,  in' preprints of
            Symposium, the Non-Urban Tropospheric Composition,
            Am.  Geophys.  Union and Am.  Meteorol.  Soc., Miami,
            Nov. 10-12j 1976.

     B-30   Walter W.  Berg and John W.  Winchester, Organic and
            inorganic  gaseous chlorine  concentrations in
            relation to the particle size distribution of
            chloride in the marine aerosol,  J. Geophys. Res.,
            £2,  No. 37, 5945-5953, 1977.


C.   STUDIES RELATED TO EPA PROGRAM BUT WITHOUT EPA SUPPORT


     Elemental Analysis and Aerosol Sampling Techniques

               Proton  Induced X-Ray Emission, PIXE

     C-l    T.B. Johansson, R. Akselsson, M. Ahlberg,
            G. Johansson, and K. Malmqvist,  Analytical use of
            proton-induced X-ray emission, Trans. Am. Nucl.
            Soc., 21,  Suppl. 3, 1975.

     Urban Aerosol Trace Metal Studies

        Statement of urban aerosol research  objectives

     C-2    John W. Winchester, Trace metal  associations in
            urban airborne particluates,  Bull. Am. Meteorol.
            Soc., 5^,  94-97, 1973.

     C-3    John W. Winchester, Application  of neutron
            activation analysis to the  investigation of natural
            and pollution aerosols, J.  Radioanalytical Chemistry,
            19_,   311-317, 1974.

     C-4    John W. Winchester, Natural and  pollution sources of
            trace elements in atmospheric particulates, in
            Trace Contaminants in the Environment, Andre F.
            LeRoy, ed., Am. Inst. Chem.,  Eng. Symp. Ser. No. 149,
            pp.   16-18, 1975.
                               67

-------
C-5    P.R. Harrison, K.A. Rahn, R. Dams, J.A. Robbins,
       and J.W. Winchester, Areawide trace metal distribu-
       tions in northwest Indiana aerosols, in Trace
       Contaminants in the Environment, Andre F. LeRoy,
       ed., Am. Inst. Chem. Eng. Symp.  Ser. No. 149,
       pp. 19-25, 1975.

C-6    John W. Winchester, Atmospheric  sciences and problems
       of society:  atmospheric chemistry, Bull. Am.
       Meteorol. Soc. , 57_, 1447-1449, 1976.

C-7    John W. Winchester, Urban air pollution impacts
       on remote regions, in Earthcare:  Global Protection
       of Natural Areas, E.A. Schofield, ed. , Westview
       Press, Boulder, Colorado, 1977.

C-8    John W. Winchester, Current environmental threats
       to the atmosphere, in Earth Program Journal,
       14th Biennial Wilderness Conference, National
       Audubon Society and Sierra Club, New York,
       pp. 108-111, June 5-8, 1975.

      Urban Aerosol Trace Metal Characterization

C-9    G.G. Desaedeleer, J.W. Winchester, M..  Demuynck,
       and R. Dams, One-year study of total suspended
       particulate matter at 14 locations in  Belgium:
       discussion, Atmospheric Environment, 10, 563-564,
       1976.

Marine Aerosol Studies

   Coastal Zone Aerosol and Trace Metal Transport

C-10   D. Skibin, Raymond Staley and John W.  Winchester,
       Comment on water pollution in Lake Michigan from
       pollution aerosol fallout, Water, Air, and Soil
       Pollution, 2_, 405-407, 1973.

C-ll   John W. Winchester, Sources, sinks, and budgets of
       atmospheric trace constituents,  in Atmospheric
       Chemistry, Problems and Scope, Report  of the
       Panel on Atmospheric Chemistry,  Richard A. Craig,
       chairman, Committee on Atmospheric Sciences,
       National Academy of Sciences, Washington, D.C.,
       pp. 59-69, 1975.

C-12   James 0. Pilotte, John W. Winchester,  and Robert C.
       Glassen, Detection of heavy metal pollution in
       estuarine sediments, Water, Air, and Soil Pollution,
       9 , No. 3, 363-368, 1977.
                          68

-------
C-13   Robert C. Classen and John W.  Winchester,  Review
       of application for site certification,  Arvah B.
       Hopkins Generating Station, Tallahassee,  Florida,
       Department of Oceanography and Florida  Resources
       and Environmental Analysis Center,  Florida State
       University, Tallahassee, November 1974.

C-14   Robert C. Classen and John W.  Winchester,  Review
       of application for site certification,  Caryville
       Steam Plant, Caryville, Florida,  Department of
       Oceanography and Florida Resources  and  Environmental
       Analysis Center, Florida State University,
       Tallahassee, September 1975.

             Air-Water Interface Studies

C-15   John W. Winchester, Atmospheric Chemistry of
       Halogens, McGraw-Hill Encyclopedia of Science and
       Technology, 4th edition, in press.

C-16   Rene E. Van Grieken, Thomas B. Johansson, and
       John W. Winchester, Trace metal fractionation
       effects between sea water and aerosols  from bubble
       bursting, Journal de Recherches Atmospheriques,
       8_, No. 3/4, 611-621, 1974.

C-17   Rene E. Van Grieken, Thomas B. Johansson, John W.
       Winchester, Dennis L. Meinert, and Walter W. Berg,
       Jr., Laboratory tracer experiments on fractionation
       during bubble bursting, Technical Report 8-74,
       Department of Oceanography, Florida State University,
       Tallahassee, August 1974.

C-18   John W. Winchester, The ocean as a source of
       particulate matter, in Background Papers for a
       Workshop on the Tropospheric Transport  of Pollutants
       to the Ocean, Miami, December 1975, Ocean Sciences
       Board,National Research Council (Issued May 1977),
       pp. 270-275.

C-19   John W. Winchester and Robert A. Duce,  The air-water
       interface:  particulate matter exchange across the
       air-water interface, in Fate of Pollutants in the
       Air and Water Environments, J.H. Suffet, ed.,
       Part 1, pp. 27-47, New York, Wiley, 1977.
                          69

-------
D.   THESES AND DISSERTATIONS


     Urban Aerosol Trace Metal Studies

           Urban Aerosol Trace Metal Characterization

     D-l    James 0.  Pilotte, Aerosol composition relationships
            in time sequence samples of an urban atmosphere,
            M.S. Thesis, Department of  Oceanography,  Florida
            State University, June 1977.

     D-2    Scott Rheingrover, A statistical model for titanium
            pollution transport in the  atmosphere of  St.  Louis,
            M.S. Thesis, Department of  Meteorology, Florida
            State University, June 1977.

     Marine Aerosol Studies

                Marine Aerosol Characterization

     D-3    Dennis L. Meinert, Chemical relationships in  the
            North Atlantic marine aerosol, M.S. Thesis, Depart-
            ment of Oceanography, Florida State University,
            June 1974.

     D-4    Michael Darzi, Terrestrial  influences on  aerosol
            composition over the western North Atlantic Ocean,
            M.S. Thesis, Department of  Oceanography,  Florida
            State University, August 1977.

                   Air-Water Interface  Studies

     D-5    Walter W. Berg, Jr., Chlorine chemistry in the
            marine atmosphere, Ph.D. Dissertation, Department
            of Oceanography, Florida State University, August
            1976.
                               70

-------
                                  TECHNICAL REPORT DATA
                           /Please read Iiutnictiom on the n>iNO.
4. TITLE AND SUBTITLE

  SOURCES AND TRANSPORT OF TRACE METALS  IN URBAN
  AEROSOLS
                                       5. REPORT DATE
                                        March 1979
                                       6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)

  John W. Winchester and J. William Nelson
                                       8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
  Department  of Oceanography
  Florida State University
  Tallahassee,  Florida  32306
                                       10. PROGRAM ELEMENT NO.

                                         1AD605   BB-04  (FY-76)
                                       11. CONTRACT/GRANT NO.
                                                             R802132
12. SPONSORING AGENCY NAME AND ADDRESS
  Environmental Sciences  Research  Laboratory - RTF, NC
  Office of Research and  Development
  U.S.  Environmental Protection Agency
  Research Triangle Park, North Carolina  27711
                                       13. TYPE OF REPORT AND PERIOD COVERED
                                          Final  4/73-1/76
                                       14. SPONSORING AGENCY CODE
                                          EPA/600/09
15. SUPPLEMENTARY NOTES
16. ABSTRACT
      A methodology  was  developed, tested, and applied to the identification of  trace
 element aerosol  sources and trace element aerosol  transport phenomena.  The method
 uses light weight and portable field sampling equipment, and results in the acquisi-
 tion of large data  sets suitable for statistical analysis and the testing of aerosol
 transport models.   Aerosol sampling in the ambient atmosphere was accomplished
 through a combination of time-sequenced filter  collections with hourly resolution,
 and fractionated aerosol collections with 0.25  ym  resolution.  Elemental analysis
 was performed with  particle induced X-ray emission,  a sensitive and rapid method.

      Aerosol characterization studies were focussed on St. Louis, Missouri, during
 the Regional Air Pollution Study.  Trace element pollutants were distinguished  from
 their natural components on the basis of particle  size distributions.  Statistical
 analysis of time-sequenced concentration measurements revealed four distinct pollu-
 tion sources for lead.   A related analysis demonstrated that meteorological factors
 controlling the  transport of titanium aerosol across the city could be identified.
 Extensions of the meteorology were applied to  (1)  the determination of trace element
 pollution deposition in the human respiratory tract through direct sampling of
 exhaled breath,  (2) indoor environments, and  (3) nonurban continental and marine
 atmospheres.
17.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                          b.IDENTIFIERS/OPEN ENDED TERMS  C. COSATI Field/Group
   Air pollution    *-Particle size distribu-
                           St.  Louis, MO
   Field tests
   Sampling
 * Aerosols
   Urban areas
 * Metals
 * Chemical analysis
  tion
* Sources
* Identifying
* Meteorology
     13B
     14B
     07D
     05K
     07B
     04B
18. DISTRIBUTION STATEMENT

 RELEASE  TO PUBLIC
                          19. SECURITY CLASS (ThisReport)
                            UNCLASSIFIED
21. NO. OF PAGES

  77
                                              20. SECURITY CLASS (This page)
                                               UNCLASSIFIED
                                                                        22. PRICE
EPA Form 2220-1 (9-73)
                                             71

-------