United States
Environmental Protection
Agency
Environmental Monitoring
Systems Laboratory
P.O. Box 15027
Las Vegas NV 89114-15027
EPA-600/4-84-059
June 1984
Research and Development
vvEPA
Western Particulate
Characterization
Study
-------
WESTERN PARTICIPATE CHARACTERIZATION
STUDY
T. A. Cahill, R. G. Flocchini, R. A. Eldred
and P. J. Feeney
Crocker Nuclear Laboratory Department of Physics
University of California, Davis, CA 95616
^ Cooperative Agreement Number R808563
Jo
sS Project Officer
do
^ Marc Pitchford
t/) Advanced Monitoring Systems Division
r^ Environmental Monitoring Systems Laboratory
A) Office of Research and Development
^ U.S. Environmental Protection Agency
Las Vegas, Nevada 89114
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, NEVADA 89114
-------
NOTICE
The information in this document has been funded wholly or in part by
the United States Environmental Protection Agency under cooperative agreements
R-806459 and R808563 with the University of California, Davis. It has been
subject to the Agency's peer and administrative review, and it has been ap-
proved for publication as an EPA document. Mention of trade names or commer-
cial products does not constitute endorsement or recommendation for use.
-------
ABSTRACT
The design and operation of a 40-station network to sample atmospheric
particles resulted in the collection, over a 2-year period from July 1979
to September 1981, of the data summarized in this report. The sampling sta-
tions were located in eight western states: Arizona, New Mexico, Utah,
Colorado, Wyoming, Montana, North Dakota, and South Dakota. The 40 remote
sites are regionally representative and are located on a nearly regular grid
over the eight states.
Stacked Filter Unit samplers were chosen for their ease of operation,
reliabilty, low-cost, and capability to provide samples segregated by size and
compatible with elemental and gravimetric analyses techniques. The samplers
collected both course (diameter between 15 and 2.5 microns) and fine (diameter
less than 2.5 microns) particles. Collected twice weekly at each site, the
samples covered 72-hour periods. Particle-induced x-ray emission (PIXE)
analyses examined samples for elements greater in atomic weight than sodium.
Samples were also analyzed gravimetrically for mass. Quality control proce-
dures and an externally operated quality assurance audit program established
the precision and accuracy of the resulting data.
Coarse particle concentrations are generally greater than the fine particle
concentrations. The former are dominated by soil-related elements. Fine par-
ticles, representing about 30% of the total measured particle concentration,
are of interest because of their impact on visibility and their capacity for
long-range transport. Unlike coarse particle parameters, fine mass and fine
sulfur concentrations are present in regional patterns on a scale that is
easily seen by the monitoring network. The fine particles are comprised pri-
marily of sulfates, soil-related materials, smoke, and materials too light in
atomic weight to be seen by PIXE analysis (e.g. carbonaceous material and
nitrates). Meteorological data implicate copper smelters and urban and
industrial areas as the source of much of the fine particulate sulfur in the
study area.
-------
CONTENTS
Abstract ill
Figures vi
Tables vii
1. Introduction 1
2. Sample Collection and Analysis 6
Sampling sites 6
Site location 6
Site criteria 6
General site criteria 6
Specific site criteria 6
Sampling Equipment 7
Stacked Filter Unit 7
Other sampling instruments 18
Sample Analysis 18
Methods 18
Validation of PIXE analyses 27
X-Ray matrix corrections 29
Minimum sensitivities .... 31
Validation of gravimetric analyses 31
Quality Assurance Procedures 33
Internal quality assurance 33
External quality assurance 34
3. Statistical Summaries of Data 35
General Considerations 35
Coarse Particles 35
Fine Particles 45
Size Distribution 51
Soil-Derived Elements 51
IV
-------
CONTENTS (Continued)
Page
Other Elements 69
Meteorology 72
Annotated Bibliography 78
Techniques of Particle Collection and Analysis 78
Samplers 78
Gravimetric analysis 78
PIXE analyses 79
Analysis of very light elements (hydrogen to flourine) .... 79
Design of the Network ..... 79
Ancillary Studies 80
VISTTA study 80
Particle sampling in Utah 80
Mt. St. Helens 81
Dust storms 81
Early Interpretive Efforts 81
Literature Cited 83
Appendices
A. Personnel and project organization 86
B. Nonstandard air sample 95
C. Quality assurance procedures 105
-------
FIGURES -
Number Page
1.1 Western particle characterization study sampling sites 3
2.1 Sulfur/sulfate results from Charleston study 8
2.2 Stacked filter unit 9
2.3 Aerosol collection efficiency 10
2.4 Variation of aerodynamic diameter as a function of face
velocity 12
2.5 Variation of aerodynamic diameter as a function of pore diameter. . 13
2.6 Effect of coatings on soil bounce-through to fine filters 14
2.7 Samples in final data set by site and start date 7/27/79
to 9/28/81 19
2.8 UCD elemental analysis system 22
2.9 Analyses of EPA polymer films 25
2.10 Forward alpha spectrum 26
2.11 Production cross sections 28
3.1 Probability plot (lognormal) Carlsbad Cavern 7/79 - 9/81 36
3.2 Scattergram comparison of coarse elements-Rocky Mountain 10/79 -
9/81 39
3.3 Coarse silicon time plot 41
3.4 Coarse soil maps 8/79 - 4/80 42
3.5 Coarse soil maps 5/80 - 1/81 43
3.6 Coarse soil maps 2/81 - 9/81 44
3.7 Network average composition of fine mass 48
3.8 Scattergram comparison of fine elements Rocky Mountain 10/79 -
9/81 49
3.9 Fine particle time plot 50
3.10 Fine soil maps 8/79 - 4/80 52
3.11 Fine soil maps 5/80 - 1/81 53
3.12 Fine soil maps 2/81 - 9/81 54
3.13 Fine sulfur maps 8/79 - 4/80 55
3.14 Fine sulfur maps 5/80 - 1/81 56
3.15 Fine sulfur maps 2/81 - 9/81. 57
3.16 Scattergram comparison of coarse and fine sulfur 61
3.17 Scattergram of coarse and fine iron for various sites 64
3.18 Coarse calcium/iron ratio map 65
3.19 Scattergram comparison of aerosol and soil calcium/iron ratios. . . 66
3.20 Scattergrams of potassium and iron for various sites 67
3.21 Distribution of potassium/iron ratios for coarse and fine
particles ......... 68
3.22 Excess fine potassium time plot 70
3.23 Excess fine potassium maps 71
3.24 Three-day mean wind vector maps 8/79 - 4/80 73
3.25 Three-day mean wind vector maps 5/80 - 1/81 74
3.26 Three-day mean wind vector maps 2/81 - 7/81 75
3.27 Three-day mean wind vector 8/79 - 7/81 76
vi
-------
TABLES
Number
1.1 Western Particle Characterization Study Sampling Sites 4
2.1 Percent of Mass on Fine Stage 15
2.2 Calibration Curve Regression Constants Including Altitude
Correction 17
2.3 Comparison of Analytical Techniques. Results are Mean and Standard
Deviation of Ratio of each Group's Results to the Reference
Results 24
2.4 Self-Absorption Corrections for PIXE Analysis 30
2.5 Elemental Analysis Minimum Sensitivities and Fraction of Time
Found 32
3.1 Coarse Stage Geometric Mean Concentrations 7/79 - 9/81 37
3.2 Coarse Stage Percentage Found by PIXE Analysis 7/79 - 9/81. ... 38
3.3 Fine Stage Geometric Mean Concentrations 7/79 - 9/81 46
3.4 Fine Stage Percentage Found by PIXE Analysis 7/79 - 9/81 47
3.5 Fine Fraction of Total Gravimetric Mass 58
3.6 Fine Fraction of Total Sulfur Mass 59
3.7 Fine Fraction of Total Iron Mass. . 60
3.8 Soil Composition - Comparison of Network Average to Average
Rocks 62
vn
-------
CHAPTER 1
INTRODUCTION
The region of the western United States centered on the Great Basin and
Rocky Mountain states includes about a quarter of the area of the contiguous 48
states but only about 8% of their population. It is an area rich in grandeur,
with remarkable vistas, diverse and fragile ecological communities, and,
consequently, a large proportion of U.S. national parks and monuments. The
region also contains enormous energy and mineralogical resources, including
deposits of coal, oil shale, and uranium, as well as ores of many nonferrous
metals. The utilization of these resources may result in impacts upon air
quality that threaten the natural condition of the region. In particular,
visibility adequate to appreciate the scenery can be harmed even by air quality
degradation that is modest in terms of existing health and welfare standards.
Thus, visibility itself is now protected in many parks and monuments in the region.
Data on air quality taken routinely in the study region are largely
irrelevant to the problem. Most existing air monitoring sites are in urban
areas subject to local pollution sources. Little information is available from
such sources on particulate size or composition, factors important both for
evaluating natural and man-made contributions to particulate matter and their
potential impacts on visibility. Existing monitoring data for particles are
generated on a one-day-in-six protocol that hinders both statistical studies
and correlations with meteorology. Finally, relatively few particulate
monitoring sites exist in the area. There have been intensive research studies
of air quality in the region, most notably the Visibility Impairment due to
Sulfur Transport and Transformation (VISTTA) study funded by the U.S. Environ-
mental Protection Agency (EPA). The valuable and detailed information from
such studies is limited, however, to relatively short periods of time (about
two weeks for VISTTA) and limited areas. Thus, they can provide little infor-
mation on the entire area, seasonal factors, or baseline values for future
trend analysis.
The three main purposes of a cooperative agreement between EPA and the
University of California at Davis (UCD) were to: 1) develop sampling and
analytical techniques suitable to the problem of nonurban air quality in the
study region, 2) use these techniques to provide statistically sound data over
the entire region for 27 months, and 3) begin the process of interpreting these
results in terms of natural and man-made particles of local or transported
origins. Each of these topics is described in this report.
Chapter 2 describes the sampling and analytical techniques developed and
1
-------
used in this program. Traditional air monitoring procedures for sampling and
analysis could not be used in most cases due to the requirements for more
detailed information than that available from such techniques, the very low
values of most atmospheric contaminants, and the costs required to establish a
large network at remote sites. Within the study area, anthropogenic sources
tend to be isolated from each other or clustered in small subregions. Distances
from such sources to remote receptor sites such as Class I air quality areas
(international parks, national wilderness areas, and national memorial parks
exceeding 5,000 acres; and national parks exceeding 6,000 acres) tend to be
great, so that primary gaseous pollutants are usually immeasureably low by
urban monitoring techniques. Thus, the study design included no routine
measurements of primary gaseous pollutants. Particles, however, are known to
be important throughout the region, both in terms of absolute concentrations,
which could on occasion exceed all federal and state standards, and in terms of
impacts such as visibility degradation. Particulate matter includes both
natural and anthropogenic components, requiring both particulate size and
chemical composition data to achieve source attributions. Particles also act
as a major sink for gaseous pollutants and dominate transport of potentially
toxic metals. Hence, the study was designed to emphasize particulate measure-
ment on a regional scale.
Sampling sites were chosen so that as many sites as possible were located
in Class I air quality areas. Sampling sites are shown in Figure 1.1 and
listed in Table 1.1. Each site was chosen so as to minimize local sources of
dust and avoid terrain effects such as atmospheric inversions associated with
basins and valleys. Local operators were recruited and trained when possible,
with the assistance of public agencies, especially the National Park Service.
Exchange of samples between field locations and UCD was accomplished via the
U.S. Postal Service and United Parcel Service using UCD-designed sealed filter
cassettes that eliminated handling of filters by sampling station personnel.
Samples were collected in over 91% of all possible sampling periods, often
under very difficult conditions of access and weather. In order to obtain
accurate mean values for particulate concentration, the network operated on a
cycle of two 72-hour periods per week, 52 weeks per year, over three summers
and two winters in the study. This allowed coverage of 86% of all time periods,
versus the 17% in traditional one-day-in-six particulate sampling. It also
allowed the program to maintain a weekend-weekday separation of time periods,
while the 72-hour duration still allowed for analysis of the effect of synopic
meteorological changes on particulate concentrations.
Since the role of particulate size is of great importance in pollutant
transport, source identification, and visibility impacts, samples were collected
in two particulate size categories: fine particles below 2.5 micrometer (ym) in
aerodynamic diameter and coarse particles between 2.5 ym and 15 pm. A 15-ym
size selective inlet baffled to reduce the effect of wind on the upper cut-off
points was developed and deployed in this study.
Analysis of the samples collected on the coarse and fine filters of the
dichotomous stacked filter unit (SFU) and virtual impactor (VI) samplers was
carried out at UCD. Gravimetric analysis of mass was accomplished on all
samples and new techniques were developed and published to allow us to handle
the extremely low mass levels encountered in the study. All samples were
-------
SCALE 100 KM/CM
Figure 1.1. Western particulate characterization study sampling sites,
-------
TABLE 1.1. WESTERN PARTICULATE CHARACTERIZATION STUDY SAMPLING SITES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
NWR
NMP
NP
=========== ============= =====
Murphy Lake, MT
Malta Airport, MT
Medicine Lake NWR, MT
Upper Souris NWR, ND
Belt Creek Ranger Station, MT
Jordan Airport, MT
Theodore Roosevelt NMP, ND
Bald Hill Dam, ND
Big Hole Valley, MT
Bluewater Fish Hatchery, MT
Charlie Odell's Ranch, MT
Lake Hiddenwood State Park, Sd
Yellowstone NP, WY
Buffalo Airport, WY
Mount Rushmore NM, SD
Lake Andes NWR, SD
Lander Airport, WY
Fort Laramie NHS, WY
Fossil Butte NM, WY
Saratoga, WY
Fish Springs NWR, UT
Brown's Park NWR, CO
Rocky Mountain NP, CO
Cedar Mountain, UT
Delta County airport, CO
1st Combat Evaluation Group,
Strategic Air Command, CO
Bryce Canyon NP, UT
Canyon! ands NP, UT
Great Sand dunes NM, CO
Grand Canyon NP, AZ
Chaco Canyon NM, NM
Fort Union NM, NM
Montezuma Castel NM, AZ
Petrified Forest NP, AZ
Grand Quivira NM, NM
Organ Pipe Cactus NM, AZ
Tonto NM, AZ
Gila Cliff Dwelling NM, NM
Carlsbad Caverns NP, NM
Fort Bowie NHS, AZ
-- National Wildlife Refuge
-- National Memorial Park
National Park
Longitude Latitude
Deg. Min. Deg. Min. Elevation
114
107
104
101
110
106
103
98
113
ins
104
100
110
106
103
98
108
104
110
106
113
108
105
110
108
103
112
109
105
112
107
105
111
109 .
106
112
111
108
104
109
NM
NHS
53
54
27
35
49
58
30
05
24
42
04
00
25
43
27
35
45
30
45
47
21
58
34
37
04
25
11
50
31
10
55
01
50
47
06
48
08
14
26
29
National
National
48
48
48
48
47
47
46
47
45
45
45
45
44
44
43
43
42
42
41
41
39
40
40
39
38
38
37
38
37
36
36
35
34
35
34
31
33
33
32
32
45
21
29
27
02
20
56
02
22
21
30
34
55
23
53
20
50
07
51
28
50
48
22
10
47
04
34
27
44
04
02
55
36
05
16
57
38
14
10
08
3100
2280
2000
1700
5800
2638
2340
1300
6060
3700
3500
17R4
6270
4950
4820
1442
5400
4260
6900
7000
4300
5500
7900
7665
5166
4238
8000
5925
8225
6800
6385
6700
3200
5820
6620
1650
2600
5700
4400
4500
Monument
Histori
c Site
-------
analyzed by particle induced x-ray emission (PIXE) for elements sodium and
heavier.
Over 17,000 multi-elemental analyses were made of the 14,920 valid, analyz-
able samples included in the final data set, for about 440,000 observed values
or upper detection limits of individual elements. Other techniques such as
x-ray fluorescence (XRF) and forward alpha scattering techniques (FAST) were
used to provide additional information on selected samples. The entire sampling
and analysis program was under an independent quality assurance program designed
to document both accuracy and precision via flow and system audits, and analytical
intercomparisons.
Chapter 3 presents statistical summaries of the data from the network.
These data are broken down into coarse and fine particle size fractions, separated
by the 2.5 ym cut point of the samplers. Coarse particle concentrations vary
from site to site, and no consistency can be found across the network. Since
these particles have little ability to travel long distances, they originate
from local sources in most cases. High intercorrelations among the traditionally
classified soil-derived particle elements Al, Si, K, Ca, Ti, Mn, and Fe, and
high correlations between these soil tracers and coarse mass, establish the
dominant role of soils in coarse particles. This is confirmed by the composi-
tional similarity between resuspended soil samples taken at all network sites
and the coarse particles present at the same site. Coarse mass dominates the
total mass of all particles between 0 and 15 microns diameter.
Fine particle concentrations are distributed more uniformly across the
network, as might be expected due to the ability of particles smaller than
2.5 ym to travel long distances. Fine particulate mass is usually dominated by
sulfur containing particles, but a large fraction of the mass is composed of
elements lighter than sodium and thus unobservable by normal x-ray based tech-
niques such as PIXE or XRF. . Strong seasonal and regional patterns are observed,
aiding in allocation of fine particles to sources both local and distant. A
major impact of smoke upon fine particle concentration was discovered, through
use of the fine particle K tracer in the smoke. The 1980 copper smelter strike
provided data on the role that smelters play in fine particles in the southern
4 states of the network.
An annotated bibliography shows the progress made on interpretation of the
network data. A list of all publications resulting from the program, in terms
of development of sampling and analytical techniques, network design, ancillary
studies and understanding of particles and their sources, is given. A fuller
accounting of interpretitive efforts is beyond the scope of this report; however,
data (on magnetic tape and as printed summaries) from the network are available
to and being used by the research and regulatory communities.
-------
CHAPTER 2
SAMPLE COLLECTION AND ANALYSIS
SAMPLING SITES
Site Location
The first step in analyzing ambient air quality over the 8-state region
was the establishment of a nonurban sampling network. Forty sampling sites
were chosen for the Western Participate Characterization Study and placed in a
relatively uniform grid over the region with as many sites as possible in Class
I areas. (Site locations are shown in Figure 1.1 and listed in Table 1.1.) A
local station operator was responsible for changing the filtering media supplied
by the UCD.
Site Criteria
Each site was visited by a member of the site selection team to determine
suitability and accessibility. Possible local sources, local climatological
conditions, and power requirements were evaluated. Each location had to meet a
set of general site criteria, and the sampler position had to meet a set of
specific site criteria.
General Site Criteria--
The site had to be:
1. Remote and regionally representative (for a distance of approximately
100 KM). Valleys were avoided if possible as was proximity to local
sources.
2. Located with existing pollutant and meteorological monitoring equipment
when possible without compromising .the first criteria.
3. Easily accessible throughout the year and maintained by responsible
persons.
4. Reasonably secure.
5. Able to supply 110 V 60 cycle power at not less than 600 watts.
Specific Site Criteria
1. The intake of the sampler was from 2 meters to 15 meters above the
-------
ground. The intake of the sampler is about 0.5 meters above its base.
In cases where the roof of a building was a possible site, the intake
was located approximately two meters above the roof. No roof sites
were considered when furnace or incinerator flues were nearby.
2. The distance from the site to buildings, trees, or other obstructions
was at least twice the height of the obstruction above the sampler.
3. The sampler was positioned away from parking lots, paved roads where
traffic loads were greater than 30 vehicles per hour, and dirt roads
that were a possible source of resuspended dust.
A detailed description of the sites selected for this network, including a
site description, a map, and a set of photographs, is available in Flocchini
et al. (1981).
SAMPLING EQUIPMENT
Stacked Filter Unit
The stacked filter unit (SFU) is a two-stage sampler in which size
segregation is provided by the collection efficiency of 8-ym pore size Nuclepore
pre-coated membrane filter. The sampler was designed so that the fine stage
collected respirable particles (less than 2.5 pro diameter), while the sum of
the two stages collected inhalable particles (less than 15 urn diameter). The
SFU was used at all forty sites.
Early versions of the SFU were tested in the EPA-DOE Charleston Sampler
Intercomparison (Camp et al., 1978) by both the EPA (Research Triangle Park) and
UCD. Results for an important pollutant species, sulfur/sulfate, are shown
in Figure 2.1. In general SFU's were equivalent in accuracy and precision to
other units for predominantly fine particles. These tests did show some enhance-
ment of fine soil particles in the SFU's, an effect corrected in later units by
coating the first filter. Successful performance of SFU's at Charleston led to
their further development and their use in this study as the primary particulate
sampler. This choice was based upon the simplicity and reliability of the SFU
design, their ability to withstand severe meteorological conditions, without
any shelter whatsoever, and their suitability for use of filter cassettes that
eliminated any direct filter handling at the sites. Forty units of the revised
UCD design were built by AeroVironment Inc.
As illustrated in Figure 2.2, the air enters the SFU through an intake
manifold with a 50% capture of 15-ym aerodynamic diameter particles. The intake
characteristics were validated by wind tunnel tests performed in June 1980 at
Texas A M. The first stage uses an 8-ym Nuclepore filter. The collection
efficiency for this filter at a flow rate of 10 liters per minute (1pm) is
shown in Figure 2.3 (Cahill et.al, 1978 b). This curve will shift to the right
as the pore size increases or as the flow rate decreases. Shown for comparison
is the collection efficiency of the upper respiratory tract and the bimodal
distribution of particles in typical urban aerosols.
The 50% capture diameter, or cutpoint, is the particle diameter at which
50% of the particles are collected by the filter. The variation of the cutpoint
as a function of flow rate (expressed in terms of face velocity) is shown in
-------
2.0
1.5
VO
0.5
I I I I I
Total
S04
UL U.
V) (A
RPTAEOMNRCLSGHI JK
Sampler Identification
1.5
1.0
0.6
1 1 1 1 1 1 1 1 1 1
Small
T *
I^L IT '
jr if *
r^ *
| ji-i,
D 3
u. u.
i i i i i i i i i I
i i i i i I i
55lT
11
D 0.38
fc 0.36
I I i 1 I I 1
ROBUPDMNRCLSGI(O)
Sampler Identification
Figure 2.1. Sulfur/sulfate results from Charleston study. Average and
standard deviations of the ratio of a sampler's result to the
average of all results for 16 sample periods for total (all
sizes) and fine particle sulfur/ sulfate.
8
-------
O
> 0 o
i
Figure 2.2. Stacked filter unit.
-------
o/
/o
100
80
60
40
20
0
0.4/xm Nuclepore
Nasopharynx
(20jB/min)
8/im Nuclepore
(5cm/sec)
I
Pasadena Volume
Distribution
0.01
O.I
I 10
Dp (/im)
100
Figure 2.3. Aerosol collection efficiency.
-------
Figure 2.4 for several filter pore sizes. These curves indicate that the 50%
capture diameter does not change significantly at low flow rates.
Nuclepore filters are produced by irradiating a polycarbonate film with
charged particles from a nuclear reactor. The tracks left by the particles are
preferentially etched into uniform, cylindrical pores. By controlling the time
length of the etching process, a specific pore diameter is produced. The rated
pore size specifies the maximum pore diameter. The variation in the 50% capture
diameter as a function of pore diameter is shown in Figure 2.5. At 10 1pm the
variation is nearly linear. The variation in pore size produces a significant
change in the cutpoint, whereas a change in flow rate does not. This problem
was eliminated by using a single batch of filters for the entire program. The
filters were selected and tested for a cutpoint of 2.5 pm, and were pre-analyzed
for elemental purity.
In order to minimize passage of dry particles onto the fine filter via
particle bounce off, the 8-pm filters were coated with a thin, 5 microgram per
square meter (pg/m2), layer of Apiezon L grease. The coating is hydrophobic,
and mass values change little even after exposure to vacuum. Extensive tests
of coated and uncoated filter at UCD (Figure 2.6) and by the California Depart-
ment of Public Health with laboratory aerosols indicate a negligible change in
capture diameter for wet aerosols and greatly improved collection efficiently
for dry aerosols (John, 1980).
Because the collection efficiency of the 8-pm Nuclepore filter does not
provide an infinitely sharp cut-off at 2.5 pm, particles near this diameter may
be collected on the wrong stage. The magnitude of this effect was calculated from
the collection efficiency curve and an assumed ambient distribution of particle
sizes for the elemental species. For this calculation, a series of ambient size
distributions were used, each had an accumulation mode curve with a peak at 0.2 pm
and dropped to zero at 6 pm, and a coarse mode curve beginning at 0.4 pm and
reaching a maximum at 12 pm. Table 2.1 shows the distribution of masses (percent
of mass on fine stage) for the SFU sampler and the corresponding "true" distri-
bution as would be realized by an idealized sampler with infinitely sharp
cut-off at 2.5 pm. For predominately coarse particles, such as soil-derived
particles, the SFU coarse concentrations are smaller and the fine concentrations
are larger than the corresponding true concentrations. For the entire network,
19% of the mass of soil-derived elements was on the fine stage. From the table,
this indicates that 17% of the soil mass was truly smaller than 2.5 pm. Thus
coarse concentrations were 2% too small (81/83=.98), while the fine concentrations
were 12% too large (19/17=1.12). For predominately fine particles, such as
sulfate, the SFU fine concentrations are smaller and the coarse concentrations
are larger than the corresponding true concentrations. For the network, 88% of
the sulfur mass was on the fine stage, corresponding to a true distribution for
sulfur of 90% smaller than 2.5 pm. Thus 17% of the sulfur on the coarse stage
was actually smaller than 2.5 pm. Supporting this analysis, the network data
indicate that the coarse-stage sulfur concentrations are not dominated by
incorrectly sized fine sulfur, as there is a low correlation between coarse and
fine sulfur concentrations.
The purpose of the fine-stage filter is to collect all particles that pass
through the coarse filter. Any highly retentive filter that delivers a sample
11
-------
o> £
« ^
^~ -^^^
4
3
o -
o
Eo 2
o
IS
o>
<
n W. John
A Parker et al
o Liu et al
iii
5 10 20
Face Velocity cm/s
50
100
Figure 2.4. Variation of aerodynamic diameter as a function of face velocity.
-------
CO
5
k.
£
5? 4
05 3
*-
o CL
0
\\iers
67 8 9 10
Max Pore Diameter (/xm)
12
Figure 2.5. Variation of aerodynamic diameter as a function of pore diameter.
-------
o FINE SOIL-NO COATING
18
iii |^
CO IO
o> l4
uj 12
^*9
fio
0 8
c/> 6
_ i
0 4
" *
0 o
1
-0"°-
-
6K
O
JJNCOATED^
1 ,
W 1 1 IV l_ OV-
13.5
-1 1
* t
COATED
1 1
M L. V/\-»f-\ 1 l_ L/
T=40°C, Low RH
i - 4 -
T T
i i
Tests
i
i
l
I
i
NONE | 0.15% 0.3% 0.6% 1.2% 2.5%
TOLUENE SOLUTION STRENGTH
WASH APIEZON-L in TOLUENE
Figure 2.6. Effect of coatings on soil bounce-through to fine filter.
-------
TABLE 2.1. PERCENT OF MASS ON FINE STAGE
Infinitely Sharp
SFU 2.5 urn
15% 12%
19% 17% (soils)
22% 20%
32% 30%
41% 40%
50% 50%
60% 60%
69% 70%
79% 80%
88% 90% (sulfur)
97% 99%
15
-------
suitable for analysis of mass and composition may be used. The filter chosen
for this study was a 0.3-um Nuclepore custom-made, high pore density filter.
This filter has an areal density of 0.8 mg/cm2, is highly efficient (99%) in the
region of interest, and is easily analyzed by particle induced x-ray emission
(PIXE). All filter batches, both coarse and fine, were analyzed for elemental
contamination. A trace amount of bromine was detected on about 6% of the coarse
filters.
Air flow was maintained by a 1/8 horsepower diaphragm pump with a volu-
metric flow controller. A flow rate of 10 1pm was regulated within ±5 percent
to a pressure drop of 35 cm Hg across the filters and was monitored by the
system's rotameter. The sample flow rate was calibrated using an orifice meter
with magnehelic gauge when the samplers were purchased and again any time the
rotameter or pump was replaced. The orifice meter was calibrated at UCD with
a spirometer. A calibration curve for the orifice was obtained in the form
log(Q) = ai + a0log(P)
where Q is the flow rate as measured by the spirometer and P is the pressure
drop across the orifice in inches of water.
When the units were first placed in the field, initial calibration curves
for the samplers' rotometers were made. A spirometer-calibrated orifice was
attached to the sampler inlet so that the sample flow passed through the orifice.
Measurements were taken for several flow rates, a plot of the rotameter reading
versus flow rate was constructed and a least squares straight line determined
for it. This regression line was used as the calibration curve for each SFU
tested after adjusting it for site elevation.
The flow rate 'was corrected for site elevation using the equation
Q' = Q (P0/Ps)°'5
where Q1 is the indicated flow rate, obtained from the calibration curve, P0 is
the pressure at Davis, CA (assumed to be 760 mmHg), and Ps is the pressure at
the site. Table 2.2 lists the calibration curve constants corrected for site
elevation for each site. In addition to the initial calibration, each sampler
was spot checked periodically with an orifice meter by the site operator. Most
of the sites were recalibrated by UCD personnel once each year.
A 7-day timer/controller turned the pump on and off at preset times, while
the elapsed timer provided an accurate account of sampling duration. The
sampler's elapsed time meter was checked in the laboratory and in the field
every twelve months against, a timepiece of known accuracy. A gain or loss of
more than five minutes in twenty-four hours warranted replacement. The cali-
bration for the on-off timer was monitored routinely by checking the elapsed
time reported from the field. If it varied more than 180 minutes from 4320
with no explanation from the operator, the operator was requested to check it.
Before the sampler? were taken to the sites each was tested for pump
reliability and time accuracy. They were flow calibrated and tested for
16
-------
TABLE 2.2. CALIBRATION CURVE REGRESSION CONSTANTS* INCLUDING
ALTITUDE CORRECTION (October 20, 1980)
Site a b Site a
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
0.059
1.056
-0.036
0.087
-0.054
0.160
-0.066
-0.590
-0.260
0.851
0.427
0.644
0.248
-1.183
0.968
0.209
0.058
-0.280
1.721
0.745
0.957
0.714
0.859
0.893
1.016
0.878
0.827
0.917
0.970
0.967
0.914
0.745
1.023
0.814
0.964
0.765
1.026
0.958
0.916
0.890
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
-0.299
0.841
0.524
-0.114
0.756
0.559
-0.087
-0.254
0.000
-0.007
0.393
-0.061
-0.428
0.220
-0.061
0.000
0.700
0.059
0.000
0.000
0.998
0.820
1.014
0.962
0.902
1.004
1.022
1.044
1.058
1.000
0.961
1.110
0.970
0.895
1.108
0.935
0.907
1.032
0.983
O.P86
*QC = a + b Qm Qc = Correct flow Rate Qm = Measured Flow Rate (rotameter)
17
-------
intercomparability. Results of side-by-side tests for all units using ambient
aerosols gave the following precision (one standard deviation) values:
Stage 1 (coarse) 10%
Stage 2 (fine) 7%
Total (sum of 1 + 2) 7%
Units were shipped to field sites, and installed by UCD personnel. Local
operators were taught proper operating procedures.
Sampling began at some sites in August 1979, and the network was fully
operational by October 1979. Sampling ended on October 1, 1981. There were
approximately 8200 sampling periods for all the sites or approximately 16,400
individual samples possible. Nine percent of these were not delivered to UCD
due to a malfunction of the unit or to the inability of the operator to get to
the sampler to change the filter. Two percent of the samples were discarded
after analysis. Eighty-nine percent were received, analyzed and included in
data. The distribution of valid analyzed samples is shown for each site in
Figure 2.7.
Other Sampling Instruments
While the SFU's provided the bulk of the data for the network, a number of
other sampling instruments were employed. The reasons for this were varied;
the need for side-by-side intercomparisons for quality assurance purposes, the
need for more size information, the need for additional time resolutions, and
support for intensive studies carried on within the network region during this
program such as the EPA VISTTA program. Details of these units and their use
are given in Appendix B.
SAMPLE ANALYSIS
Methods
Particle-Induced X-ray Emission (PIXE)--
A diagram of the analytical system developed by the UCD is given in
Figure 2.8. A 4.5 MeV proton beam from the 76" isochronous cyclotron passes
through remotely readable graphite collimators and impinges on a thin target
which is mounted at an angle of 45 degrees to the incoming beam. The target
changing mechanism operates under real-time computer control.
The beam spot is made uniform over the target and is collected by a Faraday
cup which integrates to a precision of approximately 2% to give the total
charge Q that passes through the sample. X-rays pass through an active filter
and a 25 ym Be window and are converted into electrical pulses by a 30 mm2 x 3 mm
thick liquid nitrogen cooled Si(Li) detector and associated pulsed optical
feedback circuitry. Data are accumulated in a POP 15/40 computer.
The number of characteristic x-rays, NT, corresponding to some transition
of element Z was correlated with the areal density (pt) of the element present
in the sample by N^ = AQ(pt) where A is determined through the use of gravimetric
18
-------
111111111 I
il 1 ,' i '! -) h I « » 1
SI It
N JM'«F K
1
1
t
1
1
|
1
1
1
1
\
1
1
1
1
1
1
1
1
1
1
1
rt
rt
j
')
U
1
1
1
1
1
1
>
c
r.
i
c
c
i
5
5
u
4
J"
4
in
i 4
1 7
if'\
f i
10
lu
1 1
\
*
IS
IV
3 J
£()
*'
le
1 V
r >
c o
7
10
17
Si
cM
CO
41
4
7
11
\ >
c 1
CT>
>*
1 1
1 S
«
4
U
17
il
It
1 *J
7V
7V
7v
7V
7V
7V
7V
7V
/V
7V
7V
7V
7V
7V
7V
7V
7V
/ **
70
7^
7Q
7 Q
I*4
~t u
7V
7V
7V
7V
So
RO
HO
HO
MO
pvrxifj.x.1. i
recce, c.0- c.c :
7V
H'.l
MW
j ik
HO
8 O
*
4
* *
* *
4
*
* * *
* * *
* *
* * *
*
*
*
*
*
* * *
* * *
**
* *
*
* *
*« *
* * *
»*
* *
4*4
* * *
.*.
** *
* * *
* *
* * *
* * *
* ** *
* * *
» * *
** **
* * *
* * *
* * * * * * *
* * * * * *
* ******** * * **
*** ** * * ***»
* **** **** t * * ****
** **** ** * ****
* **** **«* * ** ***
* 4*4* 4 4 4 * ** **
* * ****** ** *
***** ***** ** ***
* ***** ***** ** ****
* ***** ** * ** ** *
4 * ********** ** ****
»*******4****** ** ****
****** ******* *** ***
************ ******* * *
4******* ***** ****** *
»*** «.«« **************
**************** * * ***
« A****************** ***
****** ***** ******* * *
***** ***** ********* *
**** ** ********* **
*************** ** **
*********** *** *** ***
*********** ******* * *
********** ***** ** ***
***********************
******4 ***** **»4*4*
* * * * * 4 4 * 4 4 4 « 4 4 **** ***
***** ** ******** * *
********* *** ** *« ***
********* **** * *'* ***
*** ****** ** *» ** ***
***************** ** *
A************* * ** ***
*»*** 4 4 4 4 * 4 » 4 ** ***
***** *********** ******
»***» 4 4 ** 4 4 ** ******
* ************ * **«
* *************** ******
* * *
* * * * *
* *****
* * * * * *
* * * 4 * * *
* * * * * 4 *
* * * t * * *
* * * * * * *
* * * * *
4 * * *
4 ***** *
** *****
« 4 * * * * *
* ***** ««
* ***** **
******** **
«* ***** **
***********
***********
***********
« ***
* *** ***
* * * * * *
***********
*»* ******
**** ******
** ********
***********
******* ***
*** ***** *
* *********
**********
**** ******
***********
***********
***********
* * * * * * * * * *
******** **
******** **
*** **** **
****** ***
****** ***
****** ***
4*4*4*4*444
******** **
***********
***********
** *****4**
***********
* *********
***********
***********
***********
***********
***********
1 M F ] N « L
,14 1 & St I
Figure 2.7.
Samples in final data set by site and start
date 7/27/79 to 9/28/81.
19
-------
111111111
SITt
4
^
U
S
s
h
7
7
7
-i
A
^
0
0
0
J
0
'J
1
Ij
I
1*
J 1
c 1
9
19
^t>
L ||
I I
i!
7
11
IP
«-s
U
is
H
f.e
\
S
19
~*
10
1 I
17
cd
51
7
10
7
1
S
1 »
f?
MU
MO
HO
MO
MO
MO
MU
MO
A ll
MO
«0
My
MO
Mtl
AH
Mu
MU
MO
MU
MU
AH
MO
HO
Mu
MO
80
MO
An
MO
MO
MO
MO
Ail
MO
Ctf CKtCX* OL 3
AO
MO
MO
MO
Ml
ft*************************** ******** * *
***** *** ******** ****** ** ******** *
**************************** ***** *****
A************ ************** ***** *****
a*************************** ***** *****
*** A*********************** «*** *****
******************* ******** ********* *
****************** ********* ******* **
** **** **** **** ********* ******** **
******** ******************* ******* **
******** * * ************ ** ***********
***** ********* ** ****** ** ********* *
***** ****** ** ***********.* *********
************* ******* ****** « ******* *
************* ************** ***** ***
* ** ******* ************** ********
* *** ********************** ***** * *
***** ************ ********* * ***** * *
* ************************** *** *****
A*************************** *********
************* ***** ******** ****** **
****************** *** *** ******* ***
***** ********************* ******* ***
************************* * ******* ***
************* ************* ******* ***
*** ******»«*»**»*»»*»*** ***********
**************** ***** ***************
******»***** ******** **************
******* »**«****«*»****«****»**»»**
******* * ****************** **********
**** *************** *********
******* * *****************************
******* * * ************************ **
****** **************************** **
************************************** *
******************************** *******
******* *** A************************ *
****************************************
A********************** * **************
ft***********************************
*«****** ******* *********** ******* *
* *********«** ****** *********** *
ft*********** ***** ****** ***» ****** *
***** ******************* * *********
*** A***************** * * ******* *
**** * ********** ***** *** * **** *
******* ********** ***** *** ******* **
** * ********** ***** *********** **
******* *****«* ***** ********** **
*** ** ********* *** ***«***« »t
*** »* *** ***** ***** *********** **
***** * *** *********** *********** **
* ****** ************** ****** *«*»***
IN r
iM 1 A
M A L
St T
Figure 2.7. (Continued).
20
-------
si u
1
5
5
r>w^^.^v««
^
U
5
3
s
)
»
s
b
o
O
7
7
7
7
A
14
n
H
H
H
I
J
«J
V
' ^
1 *>
\ .
CO
J
'
13
U.
j i
h
1*
5
3
/
<-j
i
M
11
IS
H
CO
1 i
14
1 7
31
7
I'1
1 t
17
*M
41
.4
1 1
M
1 M
«!
* i
M !
U 1
HI
Ml
HI
Ml
HI
Ml
Hi
Hi
81
H
M
HI
Ml
HI
Ml
HI
Ml
Hi
HI
Ht
Ml
Ml
HI
1
Hi
H 1
Ml
Ml
HI
M 1
Ml
Ml
HI
Ml
HI
Ml
H»
Ml
Ml
Ml
Ml
HI
Ml
Ml
Ml
Ml
******* *****************
******* ****************
******* ********* ******
******* *************** *
*********************** *
************ *********
*************** *********
******** ********* **** *
******** ********* ******
*** ** ********** *****
*** ** ************** *
«** *** ********* * * «.
*** ************* ****
****************** ****
********** **** ** ** *
********** * ***** *****
A**********************
a**********************
******************** **
***********************
******** ********* ****
******** *********** **
******** *********** ***
******************** **
A******************* *
**«******* ****** **
******** *** **********
** **** **********
****** *** ************
************ ********
A************ * ********
************* * *******
««**«**« *** *********
a**********************
*** A*************
* *«* ********
* **** *********
A*********************
******************* * *'*
***** ***** ****** ***
* ***** ****** ***
************* ****** ***
A***************** ***
A**************** **
** ** * * *
* * * ** **
**** ** *«* **
» ** *****
«* * ** **
* * * * * *
************ *
**************
**************
**************
*********«« *
************ *
**************
**************
*** ******** *
«** ******* **
*** ***********
************* *
************ *
************
**************
********»***
**************
******* *****
*** * *****
*** ** ***
**** ** * **
******* * **
***** ** **
***** ******
************
***********
************
************
***********
***** ******
***********
**** *******
*** *******
*** ******
******* ***
«*** *******
** * * ****
* ***** ***
************
*** *** * *
*** ******
* ** ***
** * * *
** *****
* * * * * * *
«** *****
**** ******
**** ******
**** *«**»
* * * * * * *
* ** ******
SA-.PLF
IN F I H 1 1
)»I4 StT
Figure 2.7. (Continued).
21
-------
TO CONTROL ROOM
ro
ro
SWEEPING
MAGNET
(PROPOSED')
* TO S.VEEPE?
- > ro PDP-/5
Figure 2.8. UCD elemental analysis system.
-------
standards. Major effort has been expended in design of the electronics and
beam handling in order to insure accuracy in the acquisition of the events
comprising Nz.
The most recent formal analytical intercomparison published is from the
EPA/DOE Charleston study (Camp et al., 1978) shown in Table 2.3. Extensive
analytical intercomparisons were also included as part of the present program,
and their results are given in the Quality Assurance Procedures section and
Appendix B.
X-Ray Fluorescence (XRF)--
UCD operates a modern energy dispersive x-ray fluorescence system fully
compatible with samples used in the PIXE system. The XRF system participated
in the Charleston study (Table 2.3). While most of the SFU samples were done
by PIXE, periodic re-analysis by XRF for elements calcium and heavier confirmed
quoted PIXE values and achieved improved sensitivities for some transition
metals. Figure 2.9 shows performance of the UCD PIXE and XRF groups on polymer
films developed by the EPA-Research Triangle Park. The mean ratios to quoted
values of the PIXE system were 0.985 ± 0.05, while those of the XRF group were
1.00 ± 0.06.
Forward Alpha Scattering Techniques (FAST)
The need to gain a better knowledge of the very light elements led to the
further refinement of the forward alpha scattering techniques which have been
used occasionally at UCD since 1970. In this method, samples on teflon filters
are exposed to a 30 MeV alpha beam, and the number and energy of the scattered
alphas (and protons from hydrogen) give the amount and mass of elements hydrogen
through sodium. An example of a spectrum is shown in Figure 2.10. This tech-
nique when coupled with PIXE permits a complete elemental characterization of a
sample on a single substrate.
Gravimetric Analysis--
All network samples were analyzed gravimetrically for total mass using a
Cahn model 25 microbalance. Comparisons between pre- and post-sampling weights
gave accurate mass values. A detailed procedure was developed to handle
subtle problems due to electrostatic and filter-filter holder interactions
(Engelbrecht et al., 1981) as described in the Quality Assurance section.
23
-------
TABLE 2.3. COMPARISON OF ANALYTICAL TECHNIQUES. RESULTS ARE MEAN AND STANDARD DEVIATION OF
RATIO OF EACH GROUP'S RESULTS TO THE REFERENCE RESULTS
Results expressed as the mean value of the ratio
of the results obtained by each laboratory to
that obtained by R. Glauque, LBL, with the
uncertainty representing one standard deviation.
Group
UCD
EPA 1
LBL
FSU
EPA 2
BNL
WU.
Method
PIXE
XRF
XRF
XRF
PIXE
Chem.
AA
Chem.
Gamma
Rays
S
1.01
±0.07
0.93
±0.04
0.98
±0.04
0.89
+0.05
1.06*
+p.io
1.04*
+0.13
0.54**
+0.15
CA
1.41
+0.26
0.89
+0.06
1.27
±0.13
1.06
+0.16
Ti
1.04
^0.26
1.14
+0.13
1.33
+0.19
1.01
+0.05
Fe
1.01
+0.05
0.96
+0.13
0.98
±0.05
0.94
+0.04
Cu
0.90
+0.05
1.04
+0.13
1.02
±0.07
1.33
+0.04
Zn
1.14
±0.05
0.95
+0.06
1.10
±0.06
1.12
+0.20
Se
0.93
+0.05
0.96
+0.14
0.87
±0.14
0.49
Br
0.91
+0.05
0.92
+0.05
1.16
±0.06
0.92
+0.07
Pb
0.92
±0.03
0.93
+0.05
1.00
±0.06
0.84
+0.05
1.03
+0.02
All Values All -Worst
1.03 ± 0.16 0.98 ± 0.08
0.97 + 0.08 0.95 + 0.04
1.08 + 0.15 1.05 ±0.13
0.96 + 0.23 1.01 +_ 0.16
*Results obtained for sulfate, divided by 3.
**Results withdrawn during meeting.
LBL - Lawrence Berkeley Laboratory
FSU - Florida State University
BNL - Brookhaven National Laboratory
WU - Washington University, St. Louis, MO.
-------
Polymer film comparison
x
A=UCD PIXE
=UCD XRF
I I i I i i
Ti V Cr Mn Fe Co Ni Cu Zn Ge As Rb Sr Zr
Element
Figure 2.9. Analysis of EPA polymer films.
25
-------
RUN 01008
INITIAL DATA
ro
CO
Tf
T-l
o
X
CD
O
O
O
KAPTON
1.23 mg/cm2
CHANNEL NUMBER
Figure 2.10. Forward alpha spectrum.
500
-------
Validation of PIXE Analyses
Validation of the analytical system presented a number of problems, since
the amount of any given element (Na to U) recorded in an aerosol sample varied
from a few nanograms to a few micrograms. The following techniques were used
to insure accuracy.
1. In order to determine calibration, 28 gravimetric standards for
elements between sodium and uranium were purchased. The values for
each standard were quoted in pg/cm^ to an uncertainty of ±5% by the
manufacturer. These foils were run in the same configuration as the
aerosol samples, and provided primary calibration for the system.
Total error was about ±7% in operational conditions.
2. In order to check the standards, as well as to provide calibration
for elements for which no standard was purchased, atomic cross
sections for ion excitation at a number of ion energies were
extracted. These were plotted and fit with a polynomial expansion.
This fit resulted in values superior to those provided by individual
foils, as shown in Figure 2.11.
3. Calibration of the PIXE system was verified through reanalysis of
selected samples by an independently calibrated XRF (x-ray flour-
escence) analysis system. The XRF system, also located at UCD, was
used for studies of losses of volatiles, since no vacuum was required
for the x-ray system.
4. Important procedures in the verification of analytical systems are
interlaboratory comparisons. Formal and informal interlaboratory
comparisons involving UCD PIXE and XRF systems have been conducted
regularly since 1973. During 1981 an interlaboratory evaluation of
thin polymer films provided data directly relating the use of x-ray
emission analysis for determining elemental composition of aerosol
particles collected on membrane filters (Dzubay et al., 1981). The
data from this study are presented in Figure 2.9. These comparisons
confirmed that the quoted standard error of ±10% reported for PIXE
analyses was a conservative estimate of system accuracy.
5. Considerable effort was expended at the beginning of every
analytical run to verify system operation analysis of six to eight
gravimetric standards. In addition., a program of regular reanalysis
of aerosol samples was instituted that resulted in about 5% of all
samples being checked during each year. The analytical precision
can be estimated from the ratio of the standard deviation to the
mean for samples which were reanalyzed. The precision of the major
soil elements (silicon, potassium, calcium, and iron) was 10-12% for
all coarse stage samples and for most fine stage samples. For
samples with low soil loadings the precision dropped to 20-25%. The
precision for sulfur was 10% for most fine stage samples. For low
loadings, the precision for sulfur was around 20%. For moderate
loadings of copper and zinc the precision was 15-20%, and 30% for
lead. For a given sample the variation of the ratios of elements from
run to run was considerably less, with a precision of around 2%.
27
-------
10'
10'
10
K X-RAY
4ir(d
-------
6. Tests have been made to detect changes that occur in samples during
exposure to vacuum and beam irradiations. In samples run to very
high excitations (65 times normal integrated beam flux), some loss
of halogens has been detected (25ft)- This effect appeared to be
linear with excitation, and thus was negligible for normal analysis.
Likewise, in exposure to vacuum for 24 hours, 8 times normal exposure,
no loss of any element on a sample was detected (< 5%).
7. Other effects such as beam energy loss through the aerosol samples
and filter backing have been calculated. The resulting change in
cross section is negligible (< 5%), so no change in calibration is
needed versus sample loading. Beam scattering by the sample was also
calculated and measured, as it could reduce the beam seen by the
Faraday cup. This too was found to be negligible.
X-ray Matrix Corrections
Matrix corrections are necessary to compensate for the absorption of
x-rays before reaching the detector. These involve the product of the mass
absorption coefficient and the path length through the material. There are
two types of corrections: that due to the finite size of the particle from
which the x-ray was emitted, and that due to a thin layer of very small
particles.
When the material is deposited as separate particles of finite size, the
correction is given by an integral over the particle size range of the samples
collected. This integral includes the ambient mass distribution for the particle.
It is assumed that the particles for each element follow a bimodal mass distri-
bution as found in the California Aerosol Characterization Experiment (Hidy et
al., 1975). The relative contribution of each element by particle size was
selected on the basis of three years of data collected by UCD for the California
Air Resources Board in California urban areas. Each particle is assumed to
contain the standard oxide of the elements found, except for NaCl and PbBr.
The size dependent collection efficiency for the SFU is incorporated into the
size corrections shown in Table 2.4. These values are the inverse of the
fraction lost by absorption.
When the material is deposited as a thin layer of small particles, the
correction depends only on the thickness and the composition of the material.
The thickness is estimated from the sum of all x-ray masses multiplied by 7
to account for low z elements. The composition was estimated from a large
number of samples from California that were analyzed for all elements, including
H to F. The mass absorption coefficients are similar to those for cellulose.
These layer corrections are much smaller than those for size, except for very
heavily loaded samples. They are only applied for fine stages. Typical
corrections for average and 10 times average mass loadings 10 yg/cm^ and 100
are given in Table 2.4.
29
-------
TABLE 2.4. SELF-ABSORPTION CORRECTIONS FOR PIXE ANALYSIS
Elements
Na
Mg
AT
Si
P
S
Cl
K
Ca
Ti
V
Cr
Mn
Fe
Ni
Cu
Zn
Br
Pb
Size
Coarse
2.16
2.12
1.83
1.45
1.33
1.16
1.25
1.10
1.10
1.07
1.05
1.06
1.05
1.05
1.04
1.04
1.04
1.04
1.06
Correction
Fine
1.08
1.30
1.22
1.12
1.03
1.01
1.02
1.03
1.03
1.02
1.00
1.00
1.00
1.01
1.00
1.00
1.00
1.00
1.00
Layer
10 yg/cm2
1.02
1.01
1.01
1.01
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
Correction
100 pg/cm2
1.21
1.12
1.08
1.05
1.04
1.03
1.03
1.02
1.01
1.01
1.01
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
30
-------
Minimum Sensitivities
The minimum detection limit for each element is calculated from the back-
ground in the x-ray spectrum at the energy of the primary peak. Table 2.5
lists the detection limits assuming that the minimum number of counts in the
peak equals 3.29 times the square root of the background counts under the peak.
The table gives the average minimum sensitivities for standard 3-day samples,
including matrix corrections. The limits for lightly loaded samples are slightly
lower than these average values, since the beam current is increased and there
is no background from the loading. The sensitivities changed significantly on
three occasions. First, a new detector system with improved resolution, installed
in early 1981, was used to analyze fine-stage filters collected after November
1, 1980, and coarse-stage filters after January 1, 1981. Second, the analysis
time for the coarse-stage filters was decreased for samples collected after
April 1, 1981, in order to reduce analysis costs. Third, the SFU samplers were
modified in June 1982 to use 25 mm teflon fine-stage filters with active areas
of 27% of those used previously. (The field tests for this modification were
performed at selected sites during the summer of 1981.) Since the backgrounds
produced by the two substrates are similar, the minimum sensitivities were
decreased by a factor of 4. Though not applicable for the data covered by
this report, the average minimum sensitivity values for samples taken between
June and November 1982 are included for comparison.
In the data set provided on magnetic tapes, the minimum sensitivities are
based on a more conservative definition of minimum detection limits, with values
approximately twice those in Table 2.5 for Na through Zn and 5 times those for
Se, Br, and Pb. No measured values were eliminated because they were below the
calculated limit. When an element was not found, the negative of the minimum
detection limit was stored instead of a measured concentration.
Table 2.5 also includes the percentage of time that each element was
detected. Fine S and coarse and fine major soil-derived elements were found on
almost all samples. The ambient concentrations of most other elements were
near the minimum sensitivities for samples through 1981 and were found on less
than 1/3 of the samples. With the more sensitive fine stages of 1982, several
important elements were also found on a majority of the samples: Cu, Zn, Pb,
Na, and Ti.
Validation of Gravimetric Analyses
Weighing of the Nuclepore filters was performed on a Cahn 25 electro-
balance with a printer. Filters were initially passed over a 210p0 source to
remove any static charge. The filters were then stored between aluminum sheets
for three weeks to allow any remaining charge to leak off. Each filter was
then passed over the 210Po source again, then it was preweighed, mounted in a
filter holder and sent to a site. Upon returning, the filters were removed
from the filter holders, checked for damage, passed over the 210p0 source, post-
weighed, and stored for PIXE analysis. These procedures were checked with
calibration weights and regular reweighings of control filters. The balance
was calibrated 1 to 3 times a day.
31
-------
TABLE 2.5. ELEMENTAL ANALYSIS MINIMUM SENSITIVITIES AND FRACTION OF TIME FOUND
Minimum Sensitivities
Na
Mg
AT
Si
P
S
Cl
K
Ca
Ti
V
Cr
Mn
Fe
Ni
Cu
Zn
Se
Br
Pb
7/79
to
12/80
70
43
35
27
22
17
16
10
7
6
6
6
5
5
4
4
4
4
5
10
Coarse
1/81
to
3/81
37
23
19
16
14
10
10
7
5
4
4
4
4
3
3
2
3
2
2
4
4/81
to
10/81
90
61
50
39
33
27
26
19
14
11
11
10
9
8
6
6
7
7
9
18
(ng/m3)
7/79
to
10/80
28
20
19
15
13
11
11
7
5
4
5
4
4
4
3
3
3
3
4
8
Percentage Found
Fine
11/80
to
10/81
18
15
13
12
10
9
8
6
4
4
4
4
3
3
2
2
2
2
2
4
6/82
to
11/82
5.2
4.7
3.2
2.8
2.5
2.1
2.0
1.8
1.4
1.2
1.2
1.2
1.1
1.0
0.9
0.8
0.7
0.7
0.7
1.5
Coarse
7/79
to
10/81
11
15
67
97
5
43
15
97
99
67
10
7
35
99
5
14
11
3
16
11
Fine
7/79
to
10/81
36
21
56
93
9
98
6
92
92
24
6
12
11
91
5
13
21
2
8
27
6/82
to
11/82
62
46
85
99
8
ion
1
100
92
60
7
8
38
100
14
54
82
--
40
67
32
-------
There were 950 reweights for preweighing and 950 reweights for post-
weighings throughout the study. Filters were reweighed by a different operator
on the same day as the initial weighing. The average difference was 0.0 ± 6.0 pg/
filter (equivalent to 0.0 ± 0.14 pg/m3) for the preweights and 0.0 _+ 5.8 pg/fliter
(0.0 +; 0.13 pg/m3) on the postweights.
Five hundred control filter cartridges were kept over the 26-month project.
Each control cartridge was loaded in the same way as the field cartridges and
kept at UCD for an average of 6 weeks before postweighing. The average weight
change was 38.65 +_ 6.7 pg/filter (0.89 + 0.16 pg/m3) for the fine filters and
3.7 +_ 11.1 pg/filter (0.09 + 0.26 pg/m37 for the coarse filters. Approximately
100 filter cartridges were returned to Davis unused. The average weight gain
was 43.7 yg/filter (1.01 pg/rn3) for the fine stage and 24.1 pg/filter (0.56
pg/m3) for the coarse stage. All fine gravimetric values were corrected for a
weight gain of 40 pg (0.92 pg/m3).
QUALITY ASSURANCE PROCEDURES
All aspects of the program have been subject to extensive quality assurance
procedures in order to establish both the accuracy and uncertainties of the
data. Detailed procedures for quality assurance are documented in the project's
Quality Assurance Manual. A summary of these procedures can be found in
Appendix C, while the major findings are given here. The quality assurance
procedures fall into two catagories: the internal procedures developed, used,
and published by the UCD since 1972, and procedures used by an external quality
assurance auditing group designed and implemented for the present program by
the Rockwell International Environmental Monitoring and Services Center for the
EPA. The UCD quality assurance coordinator for this program, Lowell L. Ashbaugh,
was responsible for maintaining, implementing, and updating quality assurance
procedures and coordinating with the external audit group.
jnternal Quality Assurance
All internal quality assurance activities were documented in the 135-page
UCD Quality Assurance Manual that governed all aspects of the program. The
internal quality assurance procedures were designed to provide verifications of
accuracy and precision of results independent from the methods described earlier
for analysis validation. In the case of gravimetric analysis results, this
involved parallel handling of loaded filter cassettes sent to fixed sites and
returned without being exposed to an air flow. In some cases, such samples
arose accidently (e.g., from pump failure). . One hundred filter cartridges were
available as "field" blanks. Pre- and post-gravimetric analysis of field
blanks indicated a weight gain for fine filters equivalent to 1.0 pg/m3. This
compared well with the weight gain value of 0.89 _+ 0.16 pg/m3 measured using
the 500 control filters maintained at UCD. The reason for this gain (isolated
only during the last year of the program), was transfer of material from the
0-ring in the Nuclepore cassette to the filter. Good agreement between labora-
tory measurements, field blank results, and mass estimates made from the PIXE
data supported the correction of the fine-stage gravimetric results by 0.92
pg/m3. The coarse-stage gravimtetric values, however, differed between labora-
tory control values, 0.09 ± 0.26 pg/m3, and field blank values, 0.56 pg/m3.
Reasons for this discrepancy are being examined through study of external
contaminations, filter nonuniformity, and other aspects, but this difference
has prevented any corrections for coarse mass gravimetric data.
33
-------
Internal quality assurance procedures for PIXE and XRF data were handled by
a program of reanalysis. Typically, PIXE analysis is conducted in one multi-day
run per month. During each such run, selected samples from each analysis run
during the past 12 months are reanalyzed. Originally, samples were selected
randomly, but better coverage is achieved if samples are chosen to represent
both extreme and average loading conditions and different sites in the network.
Agreement between the values generated in each analysis of these samples
guarantees that no drift in calibration has occurred in the past year for
typical samples. These data also provide evidence on the ability of samples to
stand repeat exposure to vacuum and radiation without detectable changes in
various elemental species (i.e., sulfur, halogens, etc.). During the entire
period of this work, mean accuracy from the reanalysis procedure agreed very
well with the calibration standards, confirming absolute accuracies to ± 5%.
External Quality Assurance
External quality assurance audits were handled by Rockwell International
for this program. The areas covered by this program included audits of gravi-
metric and elemental analysis of particle samples and sample flow rates, and an
overall systems audit which reviewed and critiqued all pertinent procedures
and systems. Weight audits using NBS traceable, Class M weights, showed that
absolute mass determinations were in agreement to a few micrograms. More
relevant to the determination of sample mass by the difference of pre- and
postweights is the weight differences for the audit weights which were often
accurate to better than 1 microgram, equivalent to an uncertainty of ± 0.03
yg/nP for this program. Flow audits were designed to check operation of field
units, and they proved very useful in helping UCD track down problems in the
flow regulating systems. Two major flow audits gave mean differences of -0.9%
and -1.2% between unit readings and calibration devices, but a scatter of 14%
occurred between units. The problem in regulator drift was traced to the
erosion of a teflon seal and was corrected late in the program. Field audits
at the sites confirmed proper operation of the unit, maintenance, and operator
training. Analytical audits confirmed that the "UCD PIXE Technique is funda-
mentally sound and is capable of measuring elemental concentration on a filter."
Details are found in Appendix C.
In summary, the quality assurance procedures used in this study were more
detailed and formal than any others known to the investigators. They were a
major factor in our confidence in these data and the success of the program.
34
-------
CHAPTER 3
STATISTICAL SUMMARIES OF DATA
GENERAL CONSIDERATIONS
This statistical summary is based on the SFU measurements from July 1979
through September 1981, excluding samples which had heavy fallout from the
eruption of Mt. St. Helens on May 18, 1980. Statistics involving gravimetric
mass are from October 1979 through September 1981.
The concentrations of the various elements for both coarse and fine parti-
cles followed the usual lognormal distribution. Figure 3.1 shows a set of
probability distributions for a typical site. The cumulative frequency of the
number of cases falling below a given concentration are plotted versus the log
of the concentration. The axes are arranged so that a lognormal distribution
will follow a straight line, while a normal distribution will curve upward.
Since these data are more closely represented by a lognormal distribution, then
a normal distribution, the median concentration is best characterized by the
geometrical mean, rather than by the arithmetic mean.
COARSE PARTICLES
The geometric mean concentrations of the major elements on the coarse
stage are given in Table 3.1. When the concentration of an element was below
the detection limit of the PIXE analysis system, a value of 1/2 of the
minimum sensitivity was used in calculating the mean. The table should be used
in conjunction with Table 3.2 which shows the fraction of time that each element
was detected. This fraction depends on the analytical minimum sensitivity and
the ambient concentration. When the fraction was significantly below 50%, then
the average concentration was dominated by the minimum sensitivities, and does
not reflect the average ambient concentration.
The two tables both show that the coarse particles were dominated by six
soil-derived elements, Si, Al, Ca, Fe, K and Ti. The coarse concentrations of
the major soil elements were highly intercorrelated, with bivariate correlation
cofficients at a given site generally exceeding 0.90. Figure 3.2 shows the
correlation of iron with coarse gravimetric mass and three of the soil elements
for a typical site. The small intercept for the plot of iron and gravimetric
mass shows that there was no systematic offset in either analytic technique,
while the good correlation indicate no problems with analytical nonlinearity.
The high intercorrelation between the soil-derived elements makes it
reasonable to define a soil parameter as the sum of the concentrations of the
soil individual elements plus the related concentration of oxygen, if the
elements are assumed to be in their standard oxide forms: Si02, A^Oj, CaO,
35
-------
aes
60S
ZS
98 X
60S
2S
98 X
60S
zs
98 X
60S
ZS
98 X
60S
ZS
98 X
60S
ZS
98 X
60S
ZS
FINE POTASSIUM
FINE IRON
FINE SILICON
FINE MASS/10
COARSE POTASSIUM
60S
ZS
60S
ZS
COARSE CALCIUM
COARSE SILICON
90S
ZS
COARSE MASS/10
10°
10
102 103
NG/M3
10 «
Figure 3.1. Probability plot (lognormal) Carlsbad Cavern 7/79 - 9/81.
36
-------
TABLE 3.1. COARSE-STAGE GEOMETRIC MEAN CONCENTRATIONS 7/79 - 9/81
Nanograms/M**3
Site S AT Si Ca Fe Cl K Ti Mn Cu Mass
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
26
27
30
35
20
32
30
29
23
38
24
27
20
23
24
33
24
27
24
25
31
23
21
23
34
25
23
24
20
24
23
21
31
25
26
38
34
29
35
37
76
174
150
208
69
458
168
114
69
119
179
130
57
141
85
196
190
171
114
78
80
100
93
90
253
179
92
96
141
88
193
103
193
114
122
313
206
84
136
204
372
1092
915
1210
238
2246
934
1040
215
615
1027
1019
185
769
410
1478
954
910
609
382
343
472
355
470
1152
793
345
441
648
308
746
429
901
495
540
1435
831
348
702
922
76
166
199
315
57
703
196
303
42
322
183
196
46
614
237
550
421
355
225
140
313
162
75
190
441
250
191
165
143
119
143
121
661
189
222
434
231
-105
308
338
63
158
147
189
47
710
185
143
32
100
164
152
31
131
71
294
152
166
162
60
55
74
68
76
190
153
70
72
124
64
115
72
196
92
91
256
177
83
109
169
19
22
22
23
35
22
22
23
22
21
20
21
19
19
21
21
21
20
18
21
44
21
18
16
17
20
22
25
16
26
18
19
25
20
20
28
25
19
20
23
58
108
93
121
41
220
100
95
37
74
127
100
33
79
60
144
109
105
62
61
54
69
57
56
121
91
52
92
93
48
68
58
. 106
60
71
185
121
60
101
125
10
18
16
21
10
34
17
16
10
13
20
17
9
11
10
18
16
15
11
11
10
11
11
10
19
15
12
11
17
11
17
11
16
12
12
23
18
13
13
15
6
7
7
9
6
18
8
16
6
7
8
10
6
6
6
37
7
7
6
6
7
6
5
6
6
6
6
7
7
6
6
7
6
6
8
7
6
5
6
4
5
4
5
4
5
5
5
4
5
4
4
4
4
4
5
5
4
4
4
5
4
4
3
4
4
4
4
4
5
4
4
5
4
4
7
9
6
4
9
2900
5800
5200
'8200
2400
15500
5700
7400
1900
4200
6300
5900
1600
5800
3200
10300
5700
6300
3900
3200
3400
3100
2300
3600
6900
5500
2800
3400
4300
2300
4000
2900
7000
3100
4400
8600
5500
2900
5100
6400
All 27 146 732 253 137 23 88 15 8 5 5000
37
-------
TABLE 3.2. COARSE-STAGE PERCENTAGE FOUND BY PIXE ANALYSIS 7/79 - 9/81
Site
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
S
43
28
46
50
17
47
49
41
20
68
23
36
17
27
42
49
26
49
47
42
40
35
33
60
62
45
37
39
34
30
39
35
52
49
52
54
62
60
66
69
Si
95
99
100
99
93
100
99
98
86
100
98
99
89
100
98
100
97
99
100
96
93
95
99
99
100
99
95
96
100
96
98
95
99
98
100
100
98
97
95
100
Fe
98
99
100
99
97
100
99
97
88
99
100
100
94
100
98
100
98
99
100
96
95
98
99
100
100
99
99
99
100
96
99
98
100
99
100
100
100
100
97
100
K
95
97
96
99
94
100
98
97
93
98
99
100
90
99
97
99
97
99
97
96
95
96
95
99
90
98
93
99
100
95
98
95
100
97
98
100
99
98
97
100
Ca
97
99
99
100
95
100
99
99
92
100
98
99
90
100
99
100
99
100
100
98
98
97
97
99
100
100
100
98
100
96
99
98
100
99
100
100
100
99
99
100
Ti
54
76
71
73
53
94
71
68
48
63
79
73
42
63
53
76
75
76
64
52
40
59
59
64
81
78
63
61
78
51
78
64
75
67
63
86
79
64
68
78
AT
52
71
71
81
50
85
71
50
41
66
73
66
37
73
61
71
73
77
68
52
51
62
65
66
78
75
64
65
84
62
83
65
78
68
68
84
79
60
62
81
Mn
29
34
39
48
23
86
44
74
22
26
48
66
18
35
25
96
34
33
30
28
16
30
22
31
33
30
22
23
42
22
30
26
28
25
26
40
37
27
30
31
Cl
11
5
9
6
64
8
5
11
15
8
7
5
9
9
17
3
8
5
10
12
52
10
8
12
6
17
14
28
1
17
3
16
20
15
14
62
23
14
16
21
P
7
3
3
2
9
3
4
3
12
4
5
4
12
3
6
1
6
5
8
7
7
6
8
3
1
2
5
4
3
7
2
7
1
6
4
2
3
6
2
0
Cu
7
7
10
7
8
9
11
4
12
11
5
9
7
9
10
7
10
8
14
8
8
11
13
7
7
24
10
9
1
12
14
10
18
18
15
31
54
43
18
62
Zn
10
8
10
21
7
25
10
20
6
7
8
12
2
5
9
21
11
10
16
8
5
5
6
7
20
25
5
4
6
6
10
4
18
9
9
12
28
12
19
20
Pb
10
10
11
3
7
12
8
12
11
8
9
9
6
12
10
12
17
10
10
3
7
9
12
3
16
25
9
8
0
9
16
8
23
16
11
16
17
12
17
13
Br
13
11
10
7
11
13
15
11
18
13
16
12
13
17
17
17
22
17
16
14
14
15
14
16
25
22
16
20
5
16
17
17
19
18
15
20
17
16
22
15
Ni
7
5
1
3
3
2
4
3
7
6
6
3
4
4
6
3
4
4
7
5
7
5
5
4
2
7
7
3
3
6
7
5
3
7
6
4
2
5
9
6
V
9
9
10
6
19
5
11
9
17
8
10
9
21
8
10
4
6
8
6
11
13
12
13
8
5
7
7
12
9
8
14
14
7
9
13
5
10
10
5
8
Cr
9
6
5
5
11
2
6
3
14
7
5
4
13
5
9
3
7
6
7
8
11
3
10
5
6
7
17
9
9
7
5
6
6
9
6
1
6
7
7
2
All 43 97 99 97 99 67 67 35 15 5 14 11 11 16 5 10 7
38
-------
23.6
0.32
2.31
§
u
y>
w
ce
0,94
COARSE IRON
0.62
COARSE IRON
0.62
1.05
R - 0,92
0.47
R 0,36
COARSE IRON
0.62
COARSE IRON
0.62
Figure 3.2 Scattergram comparison of coarse elements-Rocky
Mountain 10/79-9/81.
39
-------
(2FeO + Fe203), 1^0 and Ti02« The correlation coefficient between this
coarse soil parameter and the coarse gravimetric mass at each site averaged
0.90. However, the coarse soil mass was consistently 1/2 the coarse gravimetric
mass. For the 40 sites, the ratio averaged 0.54 with a standard deviation
between sites of 0.04.
Figure 3.3 shows the time variation of coarse silicon averaged over various
regions of the network. The other coarse elements and the coarse mass follow
the same pattern. It is apparent that the coarse soil concentrations were much
larger from April to October than they were during the winter months.
The maps in Figures 3.4 to 3.6 show the spatial variation of the coarse
soil concentrations for each month of sampling. The diameters of the circles
are proportional to the monthly arithmetic mean soil concentrations. The last
map shows the average concentrations for the entire time period. The most
noticeable feature was the sudden increase in soil concentrations in the north-
western Great Plains during April and May.
An additional soil element, manganese, was measured more than 40% of the
time at seven sites, all in the northwestern Great Plains. At these sites,
manganese correlated strongly with the other soil elements, especially calcium
and iron.
Coarse sulfur was found a large portion of the time at one-half of the
sites. It correlated with the soil elements at two sites, 21 and 25, both of
which are near deserts. Large concentrations of coarse sulfur have also been
measured near Owens Dry Lake in California (Barone et al., 1981).
40
-------
o
: 7.0
_ NORTHERN 20 SlVES
. SOUTHERNZO SI
ONOJFMAMJJA50NDJFMAMJJA8
OCT 1973 TO SCP 1381
Figure 3.3. Coarse silicon time plot.
41
-------
AUG 1978
SEP 1373
OCT 1373
NOV 1373
DEC 1373
JAN 1380
FEE 1380
MAR 1980
APR 1380
r1^ «r*5 "\ f' o 2
\ o . o\ \ . O
\ » * \ S v^CU-J
.^-r « I >*-r is
CONCENTRATlOfJ CM1CROCRAMS/M««33 PROPORTIONAL TC DIAMETER
Figure 3.4. Coarse soil maps 8/79 - 4/80.
42
-------
MAY 1380
JUN1980
JUL 1380
NOV 1980
DEC 1980
JAN 1981
_ __ _
p ! \ P - ' T3 \ p
\ O A \ 0 . \ \ O
S _. , « - ^"^ S._n - "^ \ _ >
VH' I >-f M »
- CU1CROGRAMS/M--33 PROPORTIONAL TO DIAMETER
Figure 3.5. Coarse soil maps 5/80 - 1/81.
43
-------
FEB 1961
MAR 1981
APR 1301
MAY 1981
JUN 1981
JUL 19B1
AUG 1981
SEP 1981
JUL 1973 - SEP 1981
\ V5
COT4CENTRATION CMICROGRAMS/M««'o: Pf{Of'C.RTia?i.AL TO DIAMETER
Figure 3.6. Coarse soil maps 2/81 - 9/81.
44
-------
FINE PARTICLES
The geometrical mean concentrations of the major elements on the fine
stage are given in Table 3.3. When the concentration of an element was below
the detection limit of the PIXE analysis system, a value of 1/2 of the minimum
sensitivity was used in calculating the mean. Table 3.4 gives the fraction of
time that each element was detected. These tables show that of the PIXE measured
elements sulfur and soil related elements dominate the fine-stage mass. Samples
collected in northern Arizona in the VISTTA study project were analyzed for sulfur
via PIXE and for sulfate via ion chromatography. Comparison showed that all of
the particulate sulfur was present as sulfate (Macias et al., 1981 a). If the
sulfur is assumed to be ammonium sulfate, then the fine sulfate concentration
averaged over the network was 1.6 times the fine soil concentration. Figure
3.7 shows the apportionment of the fine gravimetric mass between ammonium
sulfate, soil related elements, and smoke for all sites in the network. Ammonium
sulfate accounted for 38% of the fine mass, while fine soil contributed 23$. The
smoke concentration, as discussed below, is based on the concentration of
potassium in excess of that calculated from the soil composition, and on the
assumption that potassium in smoke accounts for 5% of the smoke mass.
The correlations between fine gravimetric mass, fine soil and fine
sulfur are shown in Figure 3.8, for a typical site. Over the entire network
the correlation coefficient between fine mass and fine sulfur averaged 0.66,
with typical coefficients between 0.7 and 0.8, but never exceeding 0.8. The
correlation between fine gravimetric mass and fine soil was generally lower,
with an average value of 0.53. Only at site 11 did the correlation between
mass and soil (0.85) significantly exceed that between mass and sulfur (0.35).
The correlation between fine mass and the sum of fine sulfate and fine soil was
somewhat better than the correlations of each separately, with an average
coefficient of 0.78.
Figure 3.9 shows the time variation of fine sulfur, soil, potassium and
gravimetric mass for the northern and southern halves of the network. The
fine soil tended to follow the coarse soil pattern of higher concentrations
in spring and summer. In the north, there were three high-sulfur episodes,
in March and May 1980 and August 1981. Generally, the sulfur concentrations
were higher in summer than in winter, especially in the south. The smaller
increase during the summer of 1980 compared to the summer of 1981 in the
south was probably related to the copper smelter strike from July 1980 until
October 1980. Fine potassium could be produced by either soil or smoke.
Generally it follows the time pattern of fine soil, with the exception of
June 1980 and 1981 in the southern states, when the relative concentration of
potassium increased. The fine gravimetric mass showed less seasonal variation
than did the soil and sulfur.
45
-------
TABLE 3.3. FINE-STAGE GEOMETRIC MEAN CONCENTRATIONS 7/79 - 9/81
Site
All
233
Nanograms/M**3
Al Si Ca Fe K Tl Cu Zn
38
134
43
26
29
Pb
13
Mass
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
215
153
262
225
130
214
273
178
98
220
162
241
86
169
211
282
158
264
230
140
144
191
189
237
142
214
247
232
187
183
215
204
313
299
305
475
460
357
258
568
23
37
39
44
26
54
40
29
28
32
40
34
24
36
30
43
38
40
36
31
31
36
32
35
42
41
39
33
42
34
46
34
49
36
37
66
44
34
44
45
82
135
137
147
61
213
161
122
77
111
125
128
46
151
100
199
134
159
120
94
92
100
83
136
167
145
117
97
139
90
174
119
190
126
129
281
175
100
171
202
16
25
41
57
14
76
35
42
11
59
30
36
11
81
38
90
52
50
53
28
55
31
19
34
63
48
44
31
32
25
29
26
119
33
40
80
49
26
50
60
16
23
27
27
11
88
39
21
10
18
23
25
8
23
17
48
23
33
45
14
15
18
16
22
26
31
23
18
24
16
25
19
42
24
24
49
37
19
26
37
41
28
34
33
19
38
30
25
17
22
25
27
14
23
23
38
27
31
22
23
20
23
21
23
33
38
27
33
27
21
23
24
40
24
28
54
41
27
41
41
4
6
5
6
4
6
6
5
5
5
5
5
4
5
5
5
5
5
4
5
5
5
5
4
4
5
5
5
5
5
5
5
5
5
5
6
5
5
5
5
2
3
3
3
2
3
3
3
3
3
3
3
2
2
2
3
3
3
3
3
3
3
2
2
2
4
3
3
2
3
2
2
3
3
3
4
5
4
3
4
3
3
3
3
3
3
3
4
3
3
3
3
2
3
3
4
3
3
3
3
3
3
2
2
3
5
3
3
2
3
3
3
3
3
3
4
8
5
4
6
11
15
12
13
11
16
12
15
12
12
11
13
10
11
13
15
20
13
13
11
13
12
12
9
19
16
13
12
8
14
11
11
21
13
12
18
20
13
12
19
3800
2200
3300
3300
1800
4000
3300
2500
1400
2100
2000
2800
900
2300
2100
4200
2700
3000
2700
1600
1900
2100
1800
2400
3000
3000
2100
2000
2100
1700
2000
2000
3700
2600
2700
4500
4000
2400
2500
4100
2600
46
-------
TABLE 3.4. FINE-STAGE PERCENTAGE FOUND BY PIXE ANALYSIS 7/79 - 9/81
= = = s =
Site
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
All
S
99
97
100
98
97
99
99
99
93
100
97
99
96
97
100
98
98
99
99
95
99
98
98
99
100
99
99
98
99
98
98
97
100
100
100
99
100
98
97
100
98
Si
93
93
94
92
80
99
96
90
85
92
91
95
77
97
98
97
93
97
97
85
89
90
90
96
98
92
91
92
100
87
97
96
98
96
98
100
95
90
93
98
93
Fe
89
89
93
90
78
99
98
87
73
92
92
95
70
96
93
99
92
98
98
84
88
89
92
94
94
95
93
87
99
81
91
89
98
97
95
98
95
89
91
98
91
K
97
91
96
94
87
97
95
89
75
89
88
93
77
89
92
97
93
96
87
87
87
88
89
93
98
93
91
90
98
82
88
90
97
94
97
99
96
94
92
98
92
Ca
76
85
96
97
76
97
92
93
69
99
90
95
72
98
98
100
93
99
99
87
90
93
86
94
95
94
96
89
98
83
89
91
99
94
96
100
93
92
92
98
92
Tl
15
26
26
24
17
36
25
21
20
20
26
23
17
19
16
28
22
26
20
28
19
26
26
23
23
20
23
21
40
19
31
26
30
24
27
31
29
23
32
31
24
Al
46
53
54
64
43
72
56
43
33
48
56
49
40
58
50
61
58
64
57
46
40
56
52
62
66
57
56
51
77
48
64
59
69
60
61
73
60
57
51
67
56
Mn
5
9
12
11
7
27
14
31
12
7
11
17
10
9
11
60
11
6
7
8
7
12
10
6
6
15
5
8
13
8
9
10
9
6
14
8
10
5
8
8
11
Cl
7
10
4
4
12
4
5
6
15
3
4
6
10
6
5
4
4
2
0
10
8
4
6
3
4
7
4
4
9
7
3
7
5
3
6
10
4
1
2
4
6
P
10
11
8
10
14
6
11
8
19
9
9
5
18
6
10
7
9
10
22
15
13
14
10
5
7
4
7
6
11
10
6
13
7
6
9
3
5
12
7
2
9
Cu
6
5
7
10
3
10
8
14
5
3
5
7
7
5
5
6
4
10
3
8
7
6
10
5
5
35
6
8
4
7
14
10
20
15
12
25
57
47
13
57
13
Zn
34
14
26
16
11
25
20
26
14
9
8
17
8
10
9
33
10
9
31
6
11
4
6
14
11
46
12
6
8
12
15
11
24
28
26
31
75
48
44
67
21
Pb
18
20
15
14
9
43
17
32
19
15
9
20
9
18
25
30
49
30
30
11
17
15
27
22
46
44
19
19
5
24
24
16
62
31
23
39
61
27
28
60
27
Br
2
7
2
1
2
13
3
8
2
0
3
3
2
5
4
11
27
6
7
8
8
3
8
7
23
19
7
9
1
5
3
6
22
5
8
14
14
3
5
9
8
Ni
4
4
3
3
6
2
4
4
5
3
5
6
3
3
3
6
3
5
4
3
4
7
5
3
5
6
7
7
2
6
7
4
4
7
7
4
5
6
7
4
5
V
7
6
6
4
5
4
3
9
8
6
9
7
9
4
8
3
8
6
4
7
4
4
5
5
3
7
7
6
3
9
6
7
5
4
5
4
3
4
9
5
6
Cr
14
14
13
12
18
7
11
17
16
14
11
15
21
8
13
9
10
13
13
15
13
15
15
10
11
7
13
14
18
14
9
9
9
17
11
10
9
10
11
8
12
47
-------
W
Figure 3.7. Network average composition of fine mass.
48
-------
a 97
R 0.7Z
a 40
FINE GRAV MASS 9.5
FINE GRAV MASS 9.5
0.19
a
a.
R O.S8
. V ' '
i ' *
0.97
R ate
FINE f.KAV MASS 9.5
FINS
0.40
Figure 3.8. Scattergram comparison of fine elements Rocky
Mountain 10/79 - 9/81.
49
-------
o
NORTHERN 20 SITES FINE GRAY MASS
: soo
o
: 3.0
i t.s
120
eo
. NORTHERN 20 SITES * ' ' Fink SULFUR*
. NORTltERN 2b SlVES '
? on fc-SOUTHERN 20 SITES FINE &RAV*MASS
?
7 , ..-l SOUTHERN 20 SITES
l.c
FINE SULFUR
SOUTHERN 20 SITES
14Q
70
SOUTHERN 20 SITES FINE POTASSIUM
ONDJFMAMJJASONDJFMAMJJAS
OCT 1979 TO SEP 1981
Figure 3.9. Fine particle time plot.
50
-------
The maps of Figures 3.10 to 3.15 show the spatial variation of the fine
soil and fine sulfur concentrations for each month of sampling. The diameters
of the circles are proportional to the monthly arithmetic mean soil concentra-
tions. The last map in each set shows the average concentrations for the
entire time period. The fine soil concentrations, like the coarse, increased
rapidly in April and May for the northern Great Plains. In Arizona, the fine
soil concentrations were greatest in May and June, while fine sulfur was largest
from June to October, as long as the smelters were operating. In the northern
Great Plains, sulfur was highest during February, March, and May of 1980, and
May of 1981.
SIZE DISTRIBUTIONS
Tables 3.5 to 3.7 give the fraction of total mass on the fine stage by
site for gravimetric mass, sulfur and iron. Shown are the correlation coeffi-
cients between the concentrations in the two size fractions and three estimates
of the fine fraction. The first estimate is from the slope of the best-fitting
line through the origin in a plot of fine versus coarse. The second estimate is
the average ratio of fine to the total, while the third is the ratio of the mean
fine to the mean total. A plot of coarse versus fine sulfur at one site is
shown in Figure 3.16. The correlations between the coarse and fine concentrations
were low for gravimetric mass and sulfur, but somewhat better for iron. The
good correlation between the coarse and fine iron indicates that there is
probably a common mechanism for producing the iron on both stages. That is,
the airborne soil may have a fairly uniform size distribution which overlaps
the 2.5 ym separation of the two stages.
Approximately 1/3 of the gravimetric mass below 15 ym was smaller than
2.5 pm. Eighty-eight percent of the sulfur and 18% of the iron were collected
on the fine stage.
The size distribution for the less common elements can be estimated from
the fraction of times that the element was detected on each stage (Tables 3.2
and 3.4). Manganese, chlorine, and vanadium were predominately on the coarse
stage, while lead and zinc were primarily on the fine stage. Copper was
equally divided between the coarse and fine stages.
SOIL-DERIVED ELEMENTS
The composition of the soil-related elements in the atmosphere was similar
to that of average rocks and sediment, as defined by the CRC Handbook of Chemistry
and Physics. Table 3.8 compares the composition relative to iron for the two
SFU stages with that for various classifications of rock and sediment. At each
site, the ratio was determined by the slope of the best-fitting line through
the origin. The table lists the arithmetic mean and standard deviation of the
ratios determined for each site. Coarse sulfur is included to show that possibly
1/2 of the coarse sulfur could be soil-derived.
51
-------
AUG 1379
SEP 1373
OCT 1373
NOV 1379
DEC 1373
JAN 1980
Bf" o^T \ v T1
\ O » °\ \ ' °\
*» .. r-i l» . r-t
n^-U n^> al
\ v . « \
\ . V
CONCENTRATION CMICROGRAMS/M"«33 PROPORTIONAL TO DIAMETER
Figure 3.10. Fine soil maps 8/79 - 4/80.
52
-------
JUN 1980
JUL 1980
o\ \
V 0T" I V °
\ O jp o\ \ O
ONCENTRATION CM1CROGRAMS/M--CO P^OPr.:?7 1 OfoL TO DIAMETER
Figure 3.11. Fine soil naps 5/80 - 1/81.
53
-------
i9Bi
MAR 1981
APR 1S81
MAY 1981
JUN19B1
JUL 1981
AUG 1981
SEP 1981
JUL 1979 - SEP 1981
CONCENTRATION CMICROCRAMS/M--3D PROPORTIONAL TO DIAMETER
Figure 3.12. Fine soil maps 2/81 - 9/81.
54
-------
AUG 1979
SEP 1979
OCT 1979
NOV 1979
DEC 1979
JAN 1980
CONCENTRATION CMICROGRAMS/M--32 PROPORTIONAL TO DIAMETER
Figure 3.13. Fine sulfur maps 8/79 - 4/80.
55
-------
MAY I960
JUN I960
JUL I960
CONCENTRATION CM1C*OG.*AMS/M"3} PROF'ORTIONAL TO DIAMETER
Figure 3.14. Fine sulfur maps 5/80 - 1/81.
56
-------
FCB 1961
MAR 1381
APR 1981
MAY 1981
JUN 1981
JUL 1981
C^ <> 4 ° \ C *
\ o O > .\ \ .
S _/a ct * H ^ > "
V-P* a i v-«
°o *
AUG 1981
SEP 1981
JUL 1979 SEP 1981
5 \ f =~^
\ N °
"*
SCALE
O
1.0
o o
SCALE
CONCENTRATION CMICROGRAMS/M-«33 PROPORTIONAL TO DIAMETER
Figure 3.15. Fine sulfur maps 2/81 - 9/81.
57
-------
TABLE 3.5. FINE FRACTION OF TOTAL GRAVIMETRIC MASS
Site
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Average
Cases
166
175
194
140
164
173
182
156
145
196
175
195
157
159
194
194
186
197
159
167
168
186
195
157
130
122
176
195
98
181
162
185
199
187
198
197
187
150
132
192
t
Corr
0.20
0.53
0.43
0.31
0.60
0.24
0.43
0.48
0.67
0.56
0.43
0.41
0.50
0.28
0.51
0.20
0.23
0.29
0.59
0.73
0.30
0.47
0.64
0.52
0.43
0.18
0.39
0.45
0.66
0.30
0.52
0.60
0.48
0.51
0.44
0.35
0.49
0.64
0.40
0.53
0.45
Slope thru 0
0.42 ± 0.04
0.22 ± 0.01
0.22 ± 0.01
0.17 ± 0.01
0.38 ± 0.01
0.15 ± 0.01
0.24 ± 0.01
0.18 ± 0.01
0.38 ± 0.01
0.27 ± 0.01
0.15 ± 0.01
0.22 ± 0.01
0.37 ± 0.02
0.08 ± 0.01
0.27 ± 0.01
0.20 ± 0.01
0.23 ± 0.02
0.24 ± 0.01
0.35 ± 0.01
0.31 ± 0.01
0.11 ± 0.01
0.32 ± 0.01
0.37 ± 0.01
0.31 ± 0.01
0.18 ± 0.01
0.28 ± 0.03
0.28 ± 0.02
0.32 ± 0.01
0.24 ± 0.01
0.33 ± 0.02
0.31 ± 0.01
0.35 ± 0.01
0.30 ± 0.01
0.40 ± 0.01
0.25 ± 0.01
0.28 ± 0.01
0.36 ± 0.01
0.41 ± 0.01
0.23 ± 0.02
0.33 ± 0.01
0.28 ± 0.01
Average Ratio
0.56 ± 0.01
0.30 ± 0.01
0.40 ± 0.01
0.32 ± 0.01
0.44 ± 0.01
0.23 ± 0.01
0.38 ± 0.01
0.30 ± 0.01
0.47 ± 0.01
0.36 ± 0.01
0.26 ± 0.01
0.35 ± 0.01
0.44 ± 0.02
0.31 ± 0.01
0.41 ± 0.01
0.31 ± 0.01
0.34 ± 0.01
0.34 ± 0.01
0.42 ± 0.01
0.38 ± 0.01
0.37 ± 0.01
0.41 ± 0.01
0.46 ± 0.01
0.41 ± 0.01
0.33 ± 0.01
0.38 ± 0.02
0.46 ± 0.01
0.40 ± 0.01
0.34 ± 0.01
0.41 ± 0.01
0.36 ± 0.01
0.43 ± 0.01
0.36 ± 0.01
0.46 ± 0.01
0.40 ± 0.01
. 0.35 ± 0.01
0.43 ± 0.01
0.49 ± 0.01
0.36 ± 0.01
0.40 ± 0.01
0.38 ± 0.01
Ratio of Means
0.53 ± 0.03
0.26 ± 0.02
0.32 ± 0.02
0.26 ± 0.02
0.42 ± 0.02
0.19 ± 0.01
0.32 ± 0.02
0.24 ± 0.02
0.43 ± 0.02
0.32 ± 0.02
0.21 ± 0.02
0.30 ± 0.02
0.41 ± 0.02
0.25 ± 0.03
0.36 ± 0.02
0.27 ± 0.02
0.30 ± 0.02
0.30 ± 0.01
0.39 ± 0.02
0.34 ± 0.02
0.28 ± 0.03
0.37 ± 0.02
0.42 ± 0.02
0.37 ± 0.02
0.26 ± 0.02
0.35 ± 0.03
0.40 ± 0.02
0.37 ± 0.02
0.29 ± 0.02
0.39 ± 0.02
0.34 ± 0.02
0.39 ± 0.02
0.33 ± 0.01
0.43 ± 0.02
0.44 ± 0.02
0.33 ± 0.01
0.39 ± 0.01
0.46 ± 0.02
0.31 ± 0.02
0.37 ± 0.01
0.34 ± 0.02
Fraction Fine = Fine / (Fine + Coarse).
Element must be found on both stages to be included.
58
-------
TABLE 3.6. FINE FRACTION OF TOTAL SULFUR MASS
Site
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Average
Cases
78
51
92
74
29
87
91
67
32
140
43
71
31
51
87
96
52
101
79
76
69
69
69
96
05
56
69
81
34
62
74
70
115
101
113
117
126
100
103
144
Corr
0.73
0.48
0.55
0.51
-0.08
0.32
0.43
0.50
0.60
0.37
0.49
0.40
0.43
0.36
0.26
0.61
0.44
0.34
0.59
0.31
0.02
0.47
0.36
0.54
0.08
0.73
0.31
0.39
-0.03
0.29
0.30
0.26
0.19
0.18
0.35
0.32
0.27
0.30
0.16
0.55
0.37
Slope thru 0
0.87 ± 0.01
0.89 ± 0.01
0.89 ± 0.01
0.88 ± 0.01
0.89 ± 0.02
0.89 ± 0.01
0.91 ± 0.01
0.86 ± 0.02
0.89 ± 0.01
0.87 ± 0.01
0.88 ± 0.01
0.91 ± 0.01
0.85 ± 0.02
0.90 ± 0.01
0.92 ± 0.01
0.89 ± 0.01
0.88 ± 0.01
0.91 ± 0.01
0.89 ± 0.01
0.89 ± 0.01
0.73 ± 0.14
0.92 ± 0.01
0.93 ± 0.01
0.91 ± 0.01
0.80 ± 0.03
0.90 ± 0.01
0.93 ± 0.01
0.92 ± 0.01
0.92 ± 0.01
0.92 ± 0.01
0.92 ± 0.01
0.93 ± 0.00
0.90 ± 0.01
0.93 ± 0.01
0.92 ± 0.00
0.88 ± 0.02
0.92 ± 0.01
0.92 ± 0.01
0.91 ± 0.01
0.93 ± 0.00
0.89 ± 0.01
Average Ratio
0.87 ± 0.01
0.83 ± 0.02
0.86 ± 0.01
0.86 ± 0.01
0.86 ± 0.02
0.83 ± 0.01
0.89 ± 0.01
0.89 ± 0.01
0.81 ± 0.02
0.82 ± 0.01
0.84 ± 0.02
0.87 ± 0.01
0.83 ± 0.02
0.88 ± 0.01
0.88 ± 0.01
0.87 ± 0.01
0.83 ± 0.02
0.89 ± 0.01
0.88 ± 0.01
0.84 ± 0.01
0.75 ± 0.03
0.89 ± 0.01
0.90 ± 0.01
0.89 ± 0.02
0.74 ± 0.01
0.89 ± 0.01
0.89 ± 0.01
0.89 ± 0.01
0.91 ± 0.01
0.86 ± 0.01
0.88 ± 0.01
0.90 ± 0.01
0.87 ± 0.01
0.90 ± 0.01
0.90 ± 0.01
. 0.89 ± 0.01
0.90 ± 0.01
0.91 ± 0.01
0.85 ± 0.01
0.92 ± 0.00
0.97 ± 0.01
Ratio of Means
0.87 ± 0.02
0.86 ± 0.02
0.88 ± 0.01
0.87 ± 0.02
0.87 ± 0.01
0.86 ± 0.02
0.90 ± 0.01
0.88 ± 0.03
0.86 ± 0.03
0.85 ± 0.01
0.87 ± 0.02
0.89 ± 0.01
0.85 ± 0.02
0.89 ± 0.01
0.90 ± 0.01
0.88 ± 0.01
0.86 ± 0.02
0.90 ± 0.01
0.89 ± 0.01
0.87 ± 0.01
0.73 ± 0.05
0.91 ± 0.01
0.91 ± 0.01
0.90 ± 0.01
0.75 ± 0.03
0.90 ± 0.01
0.91 ± 0.01
0.91 ± 0.01
0.92 ± 0.01
0.89 ± 0.01
0.90 ± 0.01
0.92 ± 0.01
0.88 ± 0.01
0.92 ± 0.01
0.91 ± 0.01
0.88 ± 0.02
0.91 ± 0.01
0.91 ± 0.01
0.88 ± 0.01
0.93 ± 0.00
0.88 ± 0.01
Fraction Fine = Fine / (Fine + Coarse).
Element must be found on both stages to be included.
59
-------
TABLE 3.7. FINE FRACTION OF TOTAL IRON MASS
Site
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Average
Cases
149
165
186
131
130
172
180
143
107
189
170
189
128
162
189
195
184
197
159
151
145
170
187
148
124
119
173
179
97
165
164
176
211
197
204
209
192
147
139
202
Corr
0.72
0.81
0.79
0.54
0.58
0.54
0.75
0.85
0.64
0.76
0.45
0.84
0.72
0.68
0.78
0.53
0.81
0.80
0.69
0.78
0.44
0.74
0.75
0.74
0.60
0.65
0.68
0.77
0.65
0.79
0.76
0.68
0.78
0.70
0.85
0.62
0.81
0.70
0.60
0.69
0.70
Slope thru 0
0.16 ± 0.01
0.12 ± 0.00
0.13 ± 0.00
0.12 ± 0.01
0.19 ± 0.01
0.10 ± 0.01
0.15 ± 0.00
0.10 ± 0.00
0.18 ± 0.01
0.15 ± 0.01
0.11 ± 0.01
0.12 ± 0.00
0.16 ± 0.01
0.13 ± 0.01
0.14 ± 0.00
0.12 ± 0.01
0.11 ± 0.00
0.14 ± 0.00
0.20 ± 0.01
0.18 ± 0.01
0.15 ± 0.01
0.17 ± 0.01
0.21 ± 0.01
0.21 ± 0.01
0.09 ± 0.01
0.15 ± 0.01
0.20 ± 0.01
0.20 ± 0.01
0.13 ± 0.01
0.19 ± 0.01
0.19 ± 0.01
0.18 ± 0.00
0.17 ± 0.00
0.19 ± 0.01
0.18 ± 0.00
0.14 ± 0.00
0.18 ± 0.00
0.21 ± 0.01
0.18 ± 0.01
0.19 ± 0.01
0.16 ± 0.01
Average Ratio
0.22 ± 0.01
0.14 ± 0.00
0.17 ± 0.00
0.14 ± 0.01
0.22 ± 0.01
0.12 ± 0.00
0.19 ± 0.01
0.14 ± 0.01
0.26 ± 0.01
0.16 ± 0.01
0.14 ± 0.01
0.16 ± 0.01
0.23 ± 0.01
0.16 ± 0.01
0.20 ± 0.01
0.15 ± 0.01
0.14 ± 0.01
0.18 ± 0.00
0.23 ± 0.01
0.21 ± 0.01
0.24 ± 0.01
0.21 ± 0.01
0.21 ± 0.01
0.25 ± 0.01
0.14 ± 0.01
0.18 ± 0.01
0.27 ± 0.01
0.22 ± 0.01
0.18 ± 0.01
0.21 ± 0.01
0.19 ± 0.01
0.22 ± 0.01
0.18 ± 0.00
0.22 ± 0.01
0.22 ± 0.01
. 0.17 ± 0.00
0.19 ± 0.00
0.22 ± 0.01
0.21 ± 0.01
0.19 ± 0.01
0.19 ± 0.01
Ratio of Means
0.19 ± 0.02
0.13 ± 0.01
0.16 ± 0.02
0.13 ± 0.02
0.20 ± 0.02
0.12 ± 0.01
0.17 ± 0.01
0.11 ± 0.01
0.21 ± 0.03
0.15 ± 0.01
0.13 ± 0.02
0.14 ± 0.01
0.17 ± 0.02
0.15 ± 0.01
0.17 ± 0.02
0.14 ± 0.01
0.12 ± 0.01
0.16 ± 0.01
0.22 ± 0.01
0.20 ± 0.02
0.20 ± 0.02
0.18 ± 0.01
0.20 ± 0.02
0.23 ± 0.02
0.12 ± 0.01
0.17 ± 0.02
0.24 ± 0.02
0.21 ± 0.01
0.16 ± 0.02
0.20 ± 0.01
0.19 ± 0.01
0.20 ± 0.01
0.17 ± 0.01
0.21 ± 0.01
0.20 ± 0.01
0.16 ± 0.01
0.18 ± 0.01
0.21 ± 0.02
0.21 ± 0.02
0.19 ± 0.01
0.18 ± 0.01
Fraction Fine = Fine / (Fine + Coarse).
Element must be found on both stages to be included.
60
-------
0.10
cr
u_
to
u
C/)
o
o
*
FINE SULFUR
0.37
Figure 3.16. Scattergram comparison of coarse and fine sulfur.
61
-------
TABLE 3.8. SOIL COMPOSITION - COMPARISON OF NETWORK AVERAGE TO AVERAGE ROCKS
Ratios to Iron
Si/Fe Al/Fe Ca/Fe K/Fe Ti/Fe S/Fe
Average Rock
igneous
shale
sandstone
limestone
sediment
Network Average
coarse
st. dev.
fine
st. dev.
5.4
5.8
37.
6.4
6.7
6.9
±1.5
7.2
±2.7
1.6
1.7
2.6
1.1
1.8
1.8
±0.4
2.2
±0.5
0.7
0.5
4.0
80.
1.0
1.9
±0.9
1.8
±0.9
0.5
0.6
1.1
0.7
0.6
0.6
±0.1
0.9
±0.3
0.12
0.08
0.15
0.10
0.09
0.09
±0.02
0.09
±0.02
-
0.11
0.06
0.11
0.11
0.19
±0.13
-
_
62
-------
Calcium differed from the other elements in two ways. First, the network's
Ca/Fe ratio was the only ratio that differed from that of sediment, with the
network showing increased calcium. Second, the variation between sites, as
measured by the value of the standard deviation relative to the mean, was 2.5
times greater for the calcium ratio than for the other ratios. The average
ratios at each site were always much less than the 80 for limestone, rarely
exceeding 5. A ratio of 5 would result if the soil were characterized by
sediment with a 5% admixture of limestone. Figure 3.17 shows the correlation
between calcium and iron at 6 sites. It is evident that the correlation at any
given site was good, but that the slopes significantly differed even between
nearby sites. This indicates that the soil particles are predominantly from a
region smaller than the grid pattern of the network (about 100 km). The plots
also show that the ratios for the fine stage were similar to those for the
coarse stage. The map in Figure 3.18 shows the spatial distribution of the
Ca/Fe ratio for the coarse stage. Most of the calcium-rich sites were clustered
in the center of the network. Since this region does not have significant
limestone deposits, this may reflect a calcium enriched sediment. The highest
Ca/Fe ratio was at site 21, near the Great Salt Lake Desert.
Soil samples were collected at most sites, resuspended in the laboratory,
and analyzed for elemental content (Cahill et al., 1981). The correlation
between the Ca/Fe ratio measured in the air and the value measured in the
resuspended soil sample is shown in Figure 3.19. The only anomaly occurred at
Carlsbad Caverns, where the site is located on a limestone ridge, but the
surrounding region is not limestone. The soil sample, from the limestone
ridge, had considerably more calcium than the average aerosol. This indicates
that the airborne soil particles tend to be from a region of at least several
kilometers.
The composition of the coarse and fine soil-related elements at a given
site were similar for all elements except potassium. From Table 3.8, the
K/Fe ratio for the fine stage was 1.5 times that for the coarse stage. The
differences between coarse and fine potassium can also be seen in the
correlation plots of Figure 3.20. For the fine stage, there were many cases
when the potassium values were well above the best-fitting line for the coarse
stage, indicating enriched fine potassium. Figure 3.21 shows the distribution
of ratios for each stage. For the coarse stage, the distribution was compact,
while for the fine stage, the distribution was broader and shifted.
The excess fine potassium is probably due to smoke. Previous studies in
the Willamette and Sacramento valleys, Lake Tahoe, and northern Arizona have
shown elevated fine potassium concentrations during periods of known smoke
episodes. However, the fraction of potassium in the smoke depends on the nature
of the material being burned, ranging from 1% for nearby pine fires to 7% for
hardwood fires. A fraction of 5% was obtained from field burning in the
Willamette Valley (Lyons et al., 1980) and in a northern Arizona forest fire
(Macias et al., 1981b). With a fraction of 5%, the smoke mass would account
for approximately 10% of the total mass for the entire network, as shown in
Figure 3.7.
63
-------
z>
II
u
_J
u
Id
CO
Of
O
U
u
_J
<
o
LU
COARSE IRON
FINE IRON
Figure 3.17. Scattergrams of coarse and fine iron for various sites.
-------
Figure 3.18. Coarse calcium/Iron ratio
map.
65
-------
8
7
O 4
O
'
X
X
CARLSBAD,
O
00
I 2345678
Soils
Figure 3.19. Scattergram comparison of aerosol and soil
calcium/iron ratios.
66
-------
CO
CO
o
Q_
UJ
O
(J
CO
CO
o
Q.
UJ
COARSE IRON
FINE IRON
Figure 3.20. Scattergrams of potassium and iron for various sites.
-------
V)
Ol
'igneous rock
n shale, sediment
~"~ limestone
m
r
sandstone
t i
coarse particles
fine particles
l i
0.5 1.0 1.5
mean potassium'/ iron
2.0
Figure 3.21. Distribution of potassium/iron ratios for
coarse and fine particles.
68
-------
The excess potassium (XSK) for a sample can be as calculated from the K/Fe
ratio for coarse particles and the measured fine Fe concentration according
to:
XSK = Kfine - Fefine x (K/Fe)coarse
The coarse K to Fe ratio was determined from the slope of the best-fitting line
through the origin of coarse K versus coarse Fe. Time plots of excess potassium
(Figure 3.22) show elevated values throughout the Arizona region during June of
both 1980 and 1981. During these times winds were generally from the south in
Arizona. This suggests that there may have been field burning in Mexico at the
time. Figure 3.23 shows the spatial variation for both years combined.
OTHER ELEMENTS
Chlorine was found on a majority of coarse samples at sites 5, 21, and 36.
For the 1981 data, the correlation coefficient between chlorine and sodium was
0.83 at site 36 and 0.94 at site 21, indicating that the chlorine was present
as NaCl. Since site 36 is only 80 km from the Gulf of California, it is probable
that this was due to marine air. Site 21, at the edge of the Great Salt Lake
Desert, was characterized by unusually high concentrations of calcium, chlorine
exceeding 10 yg/nP. The chlorine here is evidently blowing salt. At site 5,
it was not possible to determine the source of chlorine. Although routinely
observed, the levels of chlorine were uniformly low, so that if the chlorine
were present as NaCl, the sodium would have been below detection limits.
The lead concentrations at the remote sites were very low, rarely exceeding
100 ng/rrr*. Lead was found on 40% to 60% of the samples at eight sites (6,17,
25,26,33,36,37,40). The presence of bromine on these samples would indicate an
automotive source. Unfortunately, the bromine mass in automotive exhaust is
around one-third that of lead, so that the bromine levels would fall below the
PIXE detection limit in many cases. At these sites, bromine is found on 14% to
50% of the samples with lead. At sites 36, 37, and 40, the fraction is lowest,
with less than 25% of the samples with lead also having bromine. This indicates
that these three sites are more likely to have a nonautomotive lead source. In
this case the source is the group of copper smelters in southern Arizona (Eldred
et al., 1982). At all eight sites, when lead and bromine were both found on a
sample, the correlation coefficients were around 0.8, and the Br/Pb ratios were
near 0.20. Sites 6, 17 and 25 were at small airports, while site 26 was at a
small Air Force base. (However, sites 2 and 14, also at small airports, produced
no high lead levels.) Site 33 was approximately one-half kilometer from Inter-
state 17. At site 26 there were also intercorrelations between fine zinc, lead
and mass, and between coarse copper, lead, soil, and mass.
At site 39 there were high correlations between fine zinc, bromine and
vanadium, although the number of pairs of samples was low. This was the only
site at which fine vanadium was more common than coarse vanadium.
69
-------
o
MONTEZUMA CASTLE
PETRIFIED FOREST
ASONDJFMAMJJASONDJFMAMJJAS
AUG 2373 TO SEP 2381
Figure 3.22. Excess fine potassium time plot.
70
-------
APR
MAY
JUN
JUL
Cf^CENTRATlflN CM1CROGRAU3/U"33 PftOr'tftTlONAl. TO DIAMETER
Figure 3.23. Excess fine potassium maps.
71
-------
Coarse copper and fine copper, zinc and lead were detected on a larger
fraction of samples at the four sites in the region of the Arizona copper
smelters (36, 37, 38, and 40), than at any other group of sites. At these
sites, the elements intercorrelated with each other and with sulfur with
correlation coefficients in a range of 0.6 to 0.8.
METEOROLOGY
Meteorological factors affecting particle concentrations in the network
are actively being studied. In an effort to explore the nature of particle
transport and to determine possible sources, work has been carried out on
analysis of surface wind and transport fields, long-range trajectories, and
diffusion model calculations.
Surface wind fields give insight into the general features of the flow
patterns in the network states. Figures 3.24 to 3.26 show the monthly resultant
vector average wind fields for the network for the period from August 1979 to
July 1981. These fields were computed by averaging the wind vectors at each
site for each 3-day sample period which began in the month indicated. The
3-day vectors were interpolated from National Weather Service sites. In the
north, an eastward flow dominated in the fall and winter months. In spring and
summer, occasional westward flows reduced or reversed the average eastward
flow. Note in particular the westward flow in March and May, 1980, and in May
and July, 1981. In the south, a northward flow prevailed overall though it
varied from northwest to northeast. An annual variation in the wind field was
not apparent as in the north. In general, however, the highest winds occurred
in spring and summer, and the lowest winds occurred in winter. In March, May,
and June of both 1980 and 1981, the winds were quite high and averaged from the
southwest.
Figure 3.27 shows the vector average wind for August 1979 through July
1981. This map shows a resultant southerly flow at the southern U.S. border
which persisted throughout Arizona and New Mexico. Near the southern border
of Wyoming, the flow became westerly and continued westerly throughout Wyoming
and Montana. In North and South Dakota, the flow varied from westerly to
northerly.
Transport of fine sulfur has been calculated by taking the product of the
sulfur concentration and the resultant wind during a sampling period. Mean north-
ward flow in Arizona and New Mexico results in northward transport of fine sulfur
because of large sources of sulfur at copper smelters in southern Arizona and
New Mexico. At the Northern Great Plains sites occasional westward flow (opposite
in direction to the mean eastward flow) can result in significant sulfur transport
from sources to the east of the network.
A long range transport model (Heffter 1980) has been used for transport and
diffusion calculations. Model calculations of sulfate concentrations have been
made using simplified assumptions of S02 emissions from copper smelters in the
south, S02 to sulfate conversion rates, and dry and wet deposition. These model
calculations show that sulfate concentrations, resulting from copper smelter
emissions, are of the same order of magnitude as the sulfate concentrations
determined by measurements in the network.
72
-------
AUC 1373
SEP 1373
OCT 1379
NOV 1379
FEB I960
DEC 1379
MAR I960
JAN 1980
APR 1980
r;.
i »
MAXIMUM VELOCITY OF 4.0 METERS/SEC
Figure 3.24. Three-day mean wind vector maps 8/79 - 4/80.
73
-------
MAY I960
JUN I960
UUL 1380
AUC 1980
SEP I960
OCT 1980
JAN 1381
MAXIMUM VELOCITY OF 4.0 METERS/SEC
Figure 3.25. Three-day mean wind vector maps 5/80 - 1/81.
74
-------
FEB 1981
MAR 1381
APR 1981
MAY 1981
JUN 1981
JUL 1981
MAXIMUM VELOCITY OF 4.0 METERS/SEC
Figure 3.26. Three-day mean wind vector maps 2/81 - 7/81.
75
-------
Figure 3.27. Three-day mean wind vector R/79 - 7/81,
76
-------
Additional work includes detailed trajectory analyses which indicate that
periods of high sulfur at Grand Canyon National Park tended to be associated
with slow transport and stagnation. Furthermore, during the summer of 1980 when
most of the smelters in southern Arizona and New Mexico were not operating, high
sulfur episodes at Grand Canyon were associated with trajectories coming from
southern California.
77
-------
ANNOTATED BIBLIOGRAPHY
Detailed interpretation of the data set gathered during this study lies
beyond the scope of this report. However, there has been significant initial
interpretation of these data. Considerable effort was expended in development
of techniques of particle collection and elemental analysis and in the overall
design of the network. Finally, several ancillary studies were conducted under
this cooperative agreement. These efforts have resulted in numerous publications
which are presented in this annotated bibliography. The publications are
presented and discussed in the following four main sections:
o Techniques of Particle Collection and Analysis
o Design of the Network
o Ancillary Studies
o Early Interpretive Efforts
TECHNIQUES OF PARTICLE COLLECTION AND ANALYSIS
Early work in the program was centered on insuring accurate data from the
network sites in terms of sampler performance, proton induced x-ray emission (PIXE)
and gravimetric analyses. Later work resulted in extension of analytical
techniques to the very light elements, hydrogen through fluorine, in order to
match the sum of elements to the gravimetric mass. This was accomplished for
the first time from a single sample collected in the Sacramento Valley. The
role of agricultural smoke to visibility as studied in the Sacramento Valley by
full elemental analyses has direct relevance to the role of forest fires in the
intermountain region.
Samplers
Cahill, T. A. Innovative Aerosol Sampling Devices Based Upon PIXE
Capabilities. Nuclear Instruments and Methods 181:473-480 (1981)
Cahill, T. A. Comments on Surface Coatings for Lundgren-Type
Impactors. Aerosol Measurement, Dale A. Lundgren, Editor; University
Presses of Florida, pp. 131-134 (1979).
Cahill, T. A. Size-Chemical Profiles of Environmental Samples by
PIXE. American Nuclear Society. San Francisco (1981).
Gravimetric Analysis
Engelbrecht, D. R., T. A. Cahill and P. J. Feeney. Electrostatic
Effects on Gravimetric Analysis of Membrane Filters. Journal of the
Air Pollution Control Association. 30:391-392 (1980).
78
-------
PIXE Analyses
Cahill, T. A. Sensitivity, Quality Assurance, ana Cost in Automated
Analysis via Ion Induced X-Rays. Nuclear Instruments and Methods
149:431-433 (1978).
Cahill, T. A. Proton Microprobes and Particle Induced X-Ray Analytical
Systems. Annual Reviews Nuclear and Particle Science 30:211-252 (1980).
Cahill, T. A. Ion Beam Analysis of Environmental Samples. Symposium
on Recent Developments in Biological and Chemical Research with Short
Lived Radioisotopes; American Chemical Society Advances in Chemistry
Series, Vol 197, pp 511-522.
Hill, M. W. , L. D. Hansen, N. F. Mangelson, K. J. Faucette, D. J.
Eatough, T. A. Cahill and B. Kusko. Measurement of Sample Temperatures
Reached During Proton and Alpha Particle Irradiation of Thin PIXE
Targets. Nuclear Instruments and Methods 181:69-70 (1981).
Cahill, T. A., R. G. Flocchini. Comments on Inaccuracies Encountered
in Sulfur Determination by Particle Induced X-Ray Emission. Analytical
Chemistry, Vol. 54, pp 1874-1877, (1982).
Analysis of Very Light Elements (hydrogen to flourinej
Cahill, T. A. Analysis of Very Light Elements in Aerosols.
CA. American Nuclear Society (1982).
Los Angeles,
Cahill, T. A., Y. Matsuda, B. H. Kusko. The Nature of Sacramento
Valley Aerosols During Agricultural Burning. American Association for
Aerosol Reseach, Santa Monica Meeting. (1982).
Cahill, T. A., Y. Matsuda, D. Shadon, R. A. Eldred, B. H. Kusko. Forward
Alpha Scattering Techniques (FAST) for Element Hydrogen Through Florine,
Nuclear Instruments and Methods, Accepted for Publication 6/83.
Cahill, T. A., R. A. Eldred, D. Shadon, P. J. Feeney, B. H. Kusko. Com-
plete Elemental Analysis of Aerosals: PIXE, FAST, LIPM, Mass. Nuclear
Instruments and Methods, Accepted for Publication 6/83.
DESIGN OF THE NETWORK
Due to the nature of the problem, the final design of the monitoring
network was unlike that of any previous network in terms of the 6-day-in-7
sequence, the samplers, and the analysis. Two papers discuss the nature of the
physical and financial constraints that led to the design.
Flocchini, R. G. , T. A. Cahill, Lowell L. Ashbaugh, R. A. Eldred,
P. J. Feeney and D. Shadoan. Regional Scale Synoptic Air Monitoring for
Visibility Evaluation Based on PIXE Analyses. Nuclear Instruments and
Methods 101:407-410 (1981).
79
-------
Ashbaugh, L. L., J. B. Barone, T. A. Cahill, R. A. Eldred and R. G.
The Influence of Synoptic Weather on the Distribution of Sulfur Aerosols
in a Portion of the Western United States. (1980).
Ashbaugh, L. L., J. B. Barone, T. A. Cahill, R. A. Eldred and R. G.
Flocchini. Characterization of Particle Episodes in the Arid West.
Presented at the 74th annual meeting of the Air Pollution Control Association,
paper 81-14a (1981).
ANCILLARY STUDIES
Several studies were performed within the umbrella of the cooperative
agreement. The foremost of these was the EPA Visibility Impairment due to
Sulfur Transport and Transformations in the Atmosphere (VISTTA) study from 1979
to 1981. It provided valuable testing of the equipment prior to deployment,
showing equivalence of the stacked filter units, virtual impactors, and physical
impactors. It also generated valuable early data on fine particles and visi-
bility in Arizona and Utah.
The data from the early EPA funded studies of particles in Utah were
a source of information which allowed the design of the network to proceed with
good knowledge of both the levels of fine particles and the temporal variability
of aerosols with synoptic weather. These data led to the 3-day integration
period to maximize sensitivity, retain synoptic variability, generate excellent
mean values, and minimize cost.
The Mt. St. Helen's eruption of May, 1980, provided a valuable by-product
of the program, since the plume drifted across the network.
The visit of Dr. Wang Ming-Xing from Peiking and the presence of Prof.
John Winchester at UCD for three months resulted in a study of dust storm
profiles.
VISTTA Study
Macias, E. S., J. 0. Zwicker, J. R. Ouimette, S. V. Hering, S. K. Friedlander,
T. A. Cahill, G. A. Kuhlmey and L, W. Richards. Regional Haze in the
Southwestern U.S. Symposium on Plumes and Visibility, Grand Canyon,
Arizona 1980. Atmospheric Environment 15:1971-1986 (1981).
Cahill, T. A., B. H. Kusko, L. L. Ashbaugh, J. B. Barone, R. A. Elred, and
E. G. Walther. Regional and Local Determinations of Particulate Matter
and Visibility in the Southwestern United States during June and July,
1979. Symposium on Plumes and Visibility, Grand Canyon 1980. Atmospheric
Environment 15:2011-2016 (1981).
Ouimette, J. R., S. K. Friedlander, T. A. Cahill. Technique for Determining
the Contributions of Aerosol Chemical Species to the Extinction Coefficient.
Air Pollution Control Association Meeting. (1980).
Particle Sampling in Utah
Flocchini, R. G., T. A. Cahill, L. L. Ashbaugh, R. A. Eldred and M.
80
-------
Pitchford. Seasonal Behavior of Particulate Matter at Three Rural Utah
Sites. Atmospheric Environment 15:315-320 (1981).
Pitchford, M., R. G. Flocchini, R. G. Draftz, T. A. Cahill, L. L. Ashbaugh
and R. A. Eldred. Silicon in Submicron Particles in the Southwest.
Atmospheric Environment 15:321-333 (1981).
Cahill, T. A., and L. L. Ashbaugh. Size/Composition Profiles of Resuspended
Fly Ash. Environmental and Climatic Impact of Coal Utilization. J. J.
Singh and A. Deepak, Editors; Academic Press, pp. 569-573 (1980).
Mt. St. Helens
Cahill, T. A., J. B. Barone, B. Kusko, L. L. Ashbaugh. Size-Composition
Profiles of Airborne Fine Particulate Matter from Mt. St. Helens. American
Geophysical Union Fall Meeting. (1980).
Leifer, R., L. Hinchliffe, I. Fisenne, H. Franklin, E. Knutson, M. Olden,
W. Sedlacek, E. Mroz, and T. Cahill. Measurements of the Stratospheric
Plume from the Mt. St. Helens Eruption Radioactivity and Chemical Composition.
Science 214:904-907 (1981).
Dust Storms
Ming-Xing, W., J. W. Winchester, T. A. Cahill, R. Lixin. Chemical
Elemental Composition of Wind Blown Dust Beijing, April 1980. Dexue
Tongbao (Science Bulletin), Vol 27, pp. 1193-1199 (1981).
EARLY INTERPRETIVE EFFORTS
The earliest interpretive efforts were primarily descriptive. They were
hampered by the partial data set available when they were conducted. More
recent work has the benefit of the full data set allowing more robust statistical,
case study and trend analysis techniques.
Cahill, T. A., R. A. Eldred, R. G. Flocchini, M. Pitchford. Size/Composi-
tion Signatures of Fine Soil Aerosols in the Western United States. ACS
Hawaii , April (1979).
Cahill, T. A., L. L. Ashbaugh, R. A. Eldred, P. J. Feeney, B. Kusko and
R. G. Flocchini. Comparisons between Size-Segreated Resuspended Soil
Samples and Ambient Aerosols in the Western United States. AMS Meeting.
Las Vegas, Nevada (1981).
Eldred, R. A., J. B. Barone, L. L. Ashbaugh, M. Pitchford. Characterization
of Atmospheric Particulate Concentrations in the Western Fine Particle
Network (1980).
Ashbaugh, L. L., J. B. Barone, R. A. Eldred, L. 0. Myrup and R. G. Flocchini.
The Influence of Synoptic Weather on the Distribution of Sulfur Aerosols
in a Portion of the Western United States. (1980).
81
-------
Flocchini, R. F., T. A. Cahill, M. L. Pitchford, R. A. Eldred, P. J. Feeney,
and L. L. Ashbaugh. Characterization of Particles in the Arid West.
Atmospheric Environment 15:2017-2030 (1981). Symposium on Plumes and
Visibility, Grand Canyon, Arizona. Atmospheric Environment 15:2017-2030
(1981).
Pitchford, A., M. Pitchford, W. Malm, R. G. Flocchini, T. A. Cahill and E.
Walther. Regional Analysis of Factors Affecting Visual Air Quality.
Symposium on Plumes and Visibility, Grand Canyon, Arizona 1980. Atmospheric
Environment 15:2043-2054 (1981).
Pitchford, M. The Relationship of Regional Visibility to Coarse and Fine
Particle Concentration in the Southwest. J. Air Pollution Control Associa-
tion 32:814-821 (1982).
Eldred, R. A., L. L. Ashbaugh, R. G. Flocchini, L. 0. Myrup and M. C.
Pitchford. Effect of the 1980 Copper Smelter Strike on Southwestern
Ambient Aerosols. American Association for Aerosol Research (AAAR) Santa
Monica, CA (1982).
Eldred, R. A., L. L. Ashbaugh, T. A. Cahill, R. G. Flocchini. Sulfate
Levels in the Southwest During the 1980 Copper Smelter Strike. J. Air
Pollution Control Association 33:110-113 (1982).
82
-------
LITERATURE CITED
Barone, 0. B., L. L. Ashbaugh, B. H. Kusko, and T. A. Cahill. 1980. The
effects of Owens Dry Lake on air quality in the Owens Valley with Impli-
cations for the Mono Lake area. Atmospheric Aerosol: Source/Air Quality
Relationships. E. S. Macias and P. K. Hopke, editors. ACS Symposium
Series. Vol. 167, pp. 327-346 (1981).
Cahill, T. A., R. G. Flocchini, R. A. Eldred, P. J. Feeney, S. Lange, D. Shadoan,
and G. Wolfe. Monitoring of Smog Aerosols with Elemental Analysis by
Accelerator Beams. Proceedings of the 17th Materials Research Symposium.
National Bureau of Standards, Maryland. (1976).
Cahill, T. A., L. L. Ashbaugh, J. B. Barone, R. A. Eldred, P. J. Feeney,
R. G. Flocchini, C. Goodart, D. J. Shadoan, and G. W. Wolfe. Analysis of
Respirable Fractions in Atmospheric Particulates via Sequential Filtration.
APCA Journal 27:675-678. (1978a).
Cahill, T. A., R. A. Eldred, J. Barone, and L. L. Ashbaugh. Ambient Aerosol
Sampling with Stacked Filter Units, Federal Highway Administration,
RD-78-178, 1978. (1978b).
Cahill, T. A., L. L. Ashbaugh, R. A. Eldred, P. J. Feeney, B. H. Kusko and
R. G. Flocchini. Comparisons between Size-Segregrated Resuspended Soil
Samples and Ambient Aerosols in the Western United States, Atmospheric
Aerosol: Source/Air Quality Relationships, E. S. Macias and P. K. Hopke,
Editors; American Chemical Society. Symposium Series, Vol. 167, pp.
269-285. (1981).
Camp, D. C., A. L. Van Lehn, B. W. Loo. Intercomparison of Samplers Used in
the Determination of Aerosol Composition. EPA 600/7-78-118. (1978).
Camp, D. C., A. L., Van Lehn, T. J., Rhodes, A. H. Pradzynski. Intercomparison
of Trace Element Determinations in Symulated and Real Air Particulate Samples
X-Ray Spectrometry. Vol. 4 pp. 123-137 (1975).
Dzubay, T. G., N. Morosoff, G. L. Whitaker, H. Yasuda. Evaluation of Polymer
Films as Standards for X-Ray Fluorescence Spectrometers, Submitted to
Electron Microscopy and X-Ray Applications to Environmental and Occupational
Health Analysis, Vol. 3, Ann Arbor Science Publishers. (1981).
Eldred, R. A., L. L. Ashbaugh, T. A. Cahill, R. G. Flocchini, and M. L. Pitchford,
Sulfate Levels in the Southwest During the 1980 Copper Smelter Strike,
Journal of the Air Pollution Control Association. Vol. 33, pp. 110-113.
(1983).
Engelbrecht, D. R., T. A. Cahill and P. J. Feeney. Electrostatic Effects on
Gravimetric Analysis of Membrane Filters. Air Pollution Control Association
Journal 30:391-392. (1981).
83
-------
Flocchini, R. G., T. A. Cahill, D. J. Shadoan, S. J. Lange, R. A. Eldred,
P. J. Feeney, G. W. Wolfe, D. C. Simmeroth, and J. K. Suder. Monitoring
California's Aerosols by Size and Elemental Composition. Environmental
Science and Technology 10:76-82. (1976).
Flocchini, R. G., T. A. Cahill, L. L. Ashbaugh, R. A. Eldred and M. Pitchford.
Seasonal Behavior of Participate Matter at Three Rural Utah sites.
Atmospheric Environment. 15:315-320. (1980).
Flocchini, R. G., T. A. Cahill, P. J. Feeney, and D. J. Shadoan. Western Energy
Resource Development Area Fine Particulate Characterization Study Sites,
EPA Report No. 600-4-81 March 1981.
Harrison, J. and R. A. Eldred. Data Acquisition and Reduction for Elemental
Analysis of Aerosol Samples. Advances in X-Ray Anal. 17, 560. (1973).
Heffter, J. L. (1980) Air Resources Laboratories Atmospheric Transport and
Dispersion Model (ARL-ATAD). NOAA Technical Memorandum ERLARL-81, Air
Resources Laboratories, Silver Spring, Maryland, 20910. 17 pp.
Hidy, G. M., B. R. Appel, R. J. Charlson, W. E. Clark, S. K. Friedlander,
D. H. Hutchinson, T. B. Smith, J. Suder, J. J. Wesolvwski and K. T. Whitby
(1975), Summary of the California Aerosol Characterization Experiment,
J. Air. Poll. Control Assoc. £5, 1106-1114.
John, W. AIHL, Berkeley, Report to the California Air Resources Board, and
Private Communication. (1980).
Loo, B. W., J. M. Jaklevic and F. S. Goulding. Dichotomous Virtual Impactors
for Large Scale Monitoring of Airborne Particulate Matter.
Lyons, C. E., I. Tombach, R. A. Eldred, and J. E. Core. Relating Particulate
Matter Sources and Impacts in the Willamette Valley During Field and Slash
Burning, Procceedings of the 72nd Annual Meeting of the Air Pollution Control
Association, Paper 79-46.3.1. (1980).
Macias, E. S., J. 0. Zwicker, J. R. Ouimette, S. V. Hering, S. K. Friedlander,
T. A. Cahill, G. A. Kuhlmey, L. W. Richards. Regional Haze Case Studies
in the Southwestern U.S.-I Aerosol Chemical Composition, Atmospheric
Environment 15:1971. (1981a).
Macias, E. S., J. 0. Zwicker, W. H. White. Regional Haze Case Studies in the
Southwestern U.S.-II Source Contributions, Atmospheric Environment 15:1987.
(1981b).
Rhodes, J. R. and C. B. Hunter. Particle size Effects in X-Ray Emission
Analysis: Simplified Formulae for Certain Practical Cases. X-Ray Spectro-
metry 1 (1972) 113.
Stevens, R. K. and T. G. Dzubay. Recent Developments in Air Particulate
Monitoring. Proceedings on Instrumental Methods for Air and Water Measure-
ments and Monitoring. IEEE Transactions on Nuclear Science. (1975).
84
-------
Stevens, R. K. and T. G. Dzubay. Dichotomous Sampler - A Practical Approach to
Fractionation and Collection. EPA -600/2-78-112 (1978).
85
-------
APPENDIX A
PERSONNEL AND PROJECT ORGANIZATION
TABLE A-l. PROJECT PERSONNEL
Name
Affiliation
Grant
funded
FTE/yr
(Total/yr)
Responsibilities
Thomas A. Cahill, Ph.D.
Physics Dept.
0.12 Principal Investigator
(0.4) Analytical Techniques
Soils;Special Studies
Robert G. Flocchini, Ph.D.
Land, Air Water 0.12
Resources (0.3)
Principal Investigator
Project Manager;
Aerosol Characterization
Robert A. Eldred, Ph.D.
Air Quality Group 1.0
(AQG); Crocker
Muclear Lab (CNL)
Investigator; Head of
Data Analysis, Reduc-
tion Handling,Stat-
istical Associations
Patrick J. Feeney, M.S.
AQG, CNL
1.0 Manager of Network Oper-
ations; Sampling Instru-
mentation; Selection
Dan J. Shadoan, M.S.
AQG, CNL
. 0.3
X-Ray Fluorescence;
Sampling Instrumen-
tation, Non-
Standard Samples
Lowell L. Ashbaugh, M.S. AQG, CNL
Ph.D.
1.0 Head of Quality
Assurance
Meteorological Analyses,
Trajectories, Sulfur
Fine Particle Transport
86
-------
TABLE A-l. (Continued)
Name
Affiliation
Grant
funded
FTE/yr
(Total/yr)
Responsibilities
John J. Barone, Ph.D
Bruce Kusko, M.S.
(Ph.D. Candidate)
AQG, CNL
0.2
AQG, Physics 0.2
Statistical Methods,
Airborne Techniques,
Remote Sampling.
Sample Analysis; VISTTA
Program.
David Englebrecht, B.S. AQG, CNL
(to 8/1/80)
1.0
Head of Sample Handling;
Gravimetric Analysis
John Olivera, B.S.
AQG, CNL
1.0
Replaces D. Englebrecht
Monica Echaves, Student AQG, CNL
0.5
Sample Handling.
Robert Pimental, Student AQG, CNL
0.5
Sample Handling.
Andrew McFarland, Ph.D. Texas AM
Contract Inlet Tests
Wang Ming Xing, Ph.D.
Academica (2 wks) Dust Sampling in
Sinica, China Remote Areas.
John Winchester, Ph.D.
Florida State (3 mos) Sulfur Chemistry
Yatsuda Matsuda, D. Eng. Osaka Prefecture (4 mos) Sulfur Chemistry;
Aerosol Characteriza-
tion
87
-------
Name
TABLE A-2. COOPERATING SCIENTISTS
Affiliation Project Component
David Johnson
Andre Sarna-Wojicki
Ray Wilcox
Robert Leifer
Peter Hobbs
Paul Hammond
Bob Sidle
Lee Hansen
Mike Darzi
Andy Andreae
Alistaire Leslie
Kin Coulson
Peter McMurray
Will Richards
Berndt Simoneit
Leonard 0. Myrup
USGS, Menlo Park
USGS, Menlo Park
USGS, Denver
DOE, New York
Univ. of Washington
Portland State Univ.
Foundation Sciences
Brigham Young Univ.
Florida State Univ.
Florida State Univ.
Florida State Univ.
NOAA, Mauna Loa
Univ. of Minnesota
MR I
UCLA
UCD
Mount St. Helens, general
Mount St. Helens, general
Mount St. Helens, general
Mount St. Helens, U2 flight
Mount St. Helens, B-23
flights.
Mount St. Helens, ground
based samples
Mount St. Helens, ground
based samples
Sulfur analyses with PIXE,
beam temperature measure-
ments
Remote analyses/volcanoes
Remote C analyses
Streaker sampling/WFPN
Remote S sampling Observa-
tory
VISTTA sampling
VISTTA sampling
(planned) Analyses of
organic material
Transport; wind trajectory
88
-------
TABLE A-3. COOPERATING AGENCIES
National Park Service U.S. Fish and Wildlife Service
U.S. Forest Service Bureau of Land Management
89
-------
PROJECT ORGANIZATION
The organizational chart for the program is shown in Figure A-l.
The job description for each position and its associated.responsibilities are
outlined below.
I. PROJECT MANAGER
A. Basic Function:
The Project Manager is responsible to the principal investigator for all
aspects of the program, with emphasis on maintaining communications between
the EPA and all group members and monitoring financial matters.
B. Responsibilities and Authorities:
1. Maintains communications with EPA project supervisor.
2. Supervises communications with all members of the project, normally
through weekly meetings and agenda items.
3. Monitors all expenditures on the program, working with management and
accounting at Crocker Nuclear Laboratory.
4. Oversees publication of results, working with the principal investi-
gators.
5. Represents the principal investigators at meetings and conferences.
II. NETWORK OPERATIONS SUPERVISOR
A. Basic Function:
The network operations supervisor is responsible for all aspects of the
program bearing upon the collection of samples from network sites, up to the
transfer of the samples to the Analytical Supervisor for analysis.
B. Responsibilities and Authorities:
1. Oversees the acquisition and maintenance of all equipment and supplies
associated with sample collection, including air samplers, flow devices
and sampling substrates.
2. Evaluates all equipment prior to placement in the field, including side
by side tests, flow calibration, etc.
3. Oversees site selection, draws up siting criteria, trains field operators,
maintains communication with field personnel.
4. Coordinates all aspects of sample handling, including handling of
substrates and return of flow data and information about environmental
conditions from field observers.
90
-------
Network Operations
Supervisor
P.J. Feeney
Quality Assurance
Coordinator
L L Ashbaugh
D. Englebracht
Wort Study Student(s)
Principal Investigators
T.A. Cahill
R.O. Flocchini
Project Manager
R.G. Flocchini
Analytical
Supervisor
D.J. Shadoan
B. Kusko
Data
Supervisor
R.A. Eldred
Special Projects
Coordinator
T.A. Cahill
L L Ashbaugh
J.B. Batone
Figure A-l. Organization chart.
-------
5. Evaluates flow readings prior to entering into the analytical system
for conversion constants to ng/m^.
6. Supervises all gravimetric analyses.
7. Coordinates quality assurance field experiments, working with the
Quality Assurance Coodinator and Special Projects Supervisor.
8. Oversees sample retrieval and storage after analyses.
9. Prepares Network Report every three months.
10. Maintains records on equipment changes and maintenance, down-time,
voided samples.
III. QUALITY ASSURANCE COORDINATOR
A. Basic Function:
The Quality Assurance Coordinator is responsible for the conduct of the
quality assurance program and for taking or recommending measures.
B. Responsibilities and Authorities:
1. Develops and carries out quality control programs, including statistical
procedures and techniques, which will help agencies meet authorized
quality standards at minimum cost.
2. Monitors quality assurance activities of the group to determine confor-
mance with policy and procedures and with sound practice; and makes
recommendations for correction and improvement as may be necessary.
3. Seeks out and evaluates new ideas and current developments in the field
of quality assurance and recommends for their application wherever
advisable.
4. Advises management in reviewing technology, methods, and equipment,
with respect to quality assurance aspects.
5. Coordinates schedules for measurement system functional check calibra-
tions, and other checking procedures.
6. Evaluates data quality and maintains records on related quality control
charts, calibration records, and other pertinent information.
7. Coordinates and/or conducts quality-problem investigations. Interfaces
with the external quality assurance program.
92
-------
IV. ANALYTICAL SUPERVISOR
A. Basic Function:
The analytical supervisor (AS) is responsible for the elemental analysis of
all samples associated with the EPA grant. Since the duties of the AS correspond
to those of the director of Analytical Services, it is desirable that the two
positions be held by the same person.
B. Responsibilities and Authorities:
1. Maintains the PIXE system, prepares the system for analysis prior to
each cyclotron run, and is responsible for preparing a schedule for the
run.
2. Maintains a logbook of procedures for pre-run set up, for sample
analysis, and for post-run shutdown.
3. Obtains the written log of flow rates and gravimetric masses for the
SFU's and VI's from the appropriate Network supervisor and prepares
a computer log. Prepares run sheets and tray files.
4. Prepares a special reanalysis tray to be run before and after the
network samples. The AS shall certify the accuracy of the system
before any samples are analyzed. Submits periodical reports on the
quality assurance of the analysis system.
5. Validates the results of all analyses and stores the collated concen-
trations on magnetic tape for use by the Data Supervisor. Is responsible
for the computer reanalysis of samples which were incorrectly analyzed
during the run.
6. Maintains the computer programs for data acquisition, for the analysis
of sample and standard foil spectra, and for the manipulation of the
files of collated concentrations. Maintains a notebook describing the
programs and how they are used.
V. DATA SUPERVISOR
A. Basic Function:
The Data Supervisor (DS) is responsible for the analysis of the network
data delivered by the Analytical Supervisor.
B. Responsibilties and Authorities:
1. The DS maintains the following libraries:
a. All network SFU data by time period;
b. All particle data by site;
c. All special samples;
93
-------
d. All appropriate meteorological parameters.
(Magnetic Tape from Library (b) shall be sent to EPA/LV).
2. Isolates anomolous results, and decides, with the cooperation of the
Analytic Supervisor, which samples need to be reanalyzed.
3. Prepares standard data summaries for each site and for the entire network.
4. Prepares periodic comparisons of the data of coincident samples.
5. Is responsible for characterizing the particulate concentrations by site
and season. Also is responsible for statistical programs to investigate
special effects.
6. Is responsible for the production of appropriate maps and plots.
7. Is responsible for investigating the relationship between the particulate
concentrations and appropriate meteorological parameters. Also is
responsible for studies of visibility reduction at the major sites.
8. Maintains a notebook describing the analysis programs and how they are
used.
VI. SPECIAL PROJECTS COORDINATOR
A. Basic Function:
The Special Projects Coordinator is responsible for all air sampling and
sample reduction programs not associated with network operations.
B. Responsibilities and Authorities;
1. Coordinates all special projects involving non-network air sampling,
analyses of materials, etc., with the EPA project supervisor.
2. Oversees all special project operations in cooperation with the Network
Supervisor (when the network is involved) and the Quality Assurance
Coordinator.
3. Oversees preparation of samples for analysis, including flow rate
calibrations, mass values, etc., prior to transmittal to the Analytical
Supervisor.
4. Coordinates data reduction with Data Reduction Supervisor.
5. Prepares documentation of all special projects, prior to publication.
94
-------
APPENDIX B
NONSTANDARD AIR SAMPLERS
In the course of this study, a number of other instruments were used for
a variety of purposes: Sierra Multiday impactors provided daily (24-hour)
data at Zion, Bryce, and Canyonlands National Parks, supported the VISTTA
study, and aided in sampler intercomparisons. The additional outpoint at 0.5
urn greatly aids studies of visibility, and several publications have appeared
with these data. Virtual impactors were used solely for sampler intercomparison.
Intercomparison between the SFU and VI were done at UCD; Las Vegas, Nevada;
Canyonlands National Park, Utah; Yellowstone National Park, Wyoming; Carlsbad
Caverns National Park, New Mexico; and Cedar Mountain, Utah. Other virtual
impactors were operated as part of the VISTTA study. Preliminary data were
taken with Battelle 8-stage impactors, Florida State streakers, and a high
sensitivity SFU, but findings from these are not included in this report.
SIERRA INSTRUMENTS MULTIDAY IMPACTOR
Particulate samples were collected at Yellowstone National Park, Wyoming;
Canyonlands National Park^ Utah; Zion National Park, Utah; and Cedar Mountain,
Utah, in three size ranges (15 pm to 3.5 pro, 3.5 pm to .50 pm, and 0.50 ym to
0.01 pm) by Lundgren type rotating drum impactors. These devices use two
impaction stages and a filtration stage. The rotary drum impaction surfaces
were covered with 530 pg/cm^ of type S mylar with a coating of approximately 65
pg/cm2 of Apiezon L grease. The final filter used was a 0.4 ym pore size
Nuclepore. The samplers were equipped with constant flow devices which
maintained a flow rate of 20 1pm. This unit delivered daily samples suitable
for optical, beta-gauge, and PIXE analyses. The use of this instrument is
described in Flocchini (1976). Extensive studies describing particle loss,
particle sizing, sample transport and handling are described in Cahill (1976).
VIRTUAL IMPACTOR
Virtual impactor (VI) samples were collected at Carlsbad Caverns, New Mexico;
Cedar Mountain, Utah; Canyonlands National Park, Utah; and Yellowstone National
Park, Wyoming, coincident with the network's SFU operation to provide a comparison
of this studies data to that of other studies which use the VI sampler.
The VI selected was the Sierra Instruments Model 244 Dichotomous Sampler
(Stevens 1975, 1978) (Loo 1975, 1979). This sampler collects two size fraction-
ated samples (2.5 to 15 pm, and less than 2.5 pm), nominally the same as the
SFU sampler. The two-particle fractions are collected on 37mm Teflon membrane
filters. The flow-controlled pump maintains a 16.7 1pm flow rate within ±3%
over a 0 to 40 cm Hg pressure drop across the fine-particle filter by maintaining
a constant pressure differential across a valve on the inlet side of the vacuum
pump.
95
-------
The sampler's inlet is designed to have an upper particle cut-off diameter
of approximately 15 ym. Particles smaller than 15 ym pass through a flow
straightening tube and enter the virtual impactor head (Figure B-l). The aero-
sol then is accelerated through the impactor jet. By virtue of their greater
inertia, particles larger than 2.5 ym impact into the "void" of the receiver
tube, and are then collected on a 37mm filter. Since ten percent of the flow
passes to the coarse stage, ten percent of the fine particles are also
collected on this filter. The remainder of the fine particle laden air passes
around the receiver tube end and is collected on a second 37mm filter at a flow
rate of 15.0 1pm.
Table B-l summarizes a comparison study using SFU, VI and a low-volume
open face filter samplers performed at UCD. The arithmetic mean and standard
deviation of ratios of results from the three types of samplers are shown. For
coarse soil type particles (Si, K, Ca, Fe), the average ratio was 0.93 (VI/SFU).
The ratio was 0.94 for the sum. For sulfur, the ratio was 1.16 (VI/SFU).
Other VI/SFU comparison studies have been conducted. The major EPA/DOE
study referred to in the body of this report included several VI's (units
D,R,C,L, and S in Figure B-2). The original paper describing the development
of SFU's by the EPA-Research Triangle Park (Parker et al. 1977) included a VI
comparison with good agreement, as did a study at Los Alamos (Ferek et al.
1977). Tables B-2 and B-3 summarize the agreememt between SFU and VI samplers
from these studies, none of which, incidently, used the coated filters (designed
to inhibit the problem of solid particle bounce) now used on all units.
The first large scale SFU/VI intercomparison in field conditions was in the
Willamette Valley study done by the Oregon Dept. of Environmental Quality,
June-November 1978. Results for sulfur are shown in Figure B-3 corresponding
to a correlation coefficient of 0.82 and a slope of 1.08 ± 0.12. The corres-
ponding values for fine iron, a soil-derived particle, are 0.78 correlation
coefficient and a slope of 0.97 ± 0.12, showing the value of coated filters in
field conditions. With the development at UCD of a capability to measure
accurate small mass values gravimetrically, a study was done near Las Vegas on
fine mass levels. The results gave an SFU/VI relationship: SFU = 0.917 VI +
1.0 yg/m3.
The co-location of VI with SFU sampler produced a modest data set for com-
paring the results of these systems. Though most of the possible samples to be
collected by the VI were either not collected or invalidated due to instrument
malfunction, a significant data set for comparing the VI and SFU remained. The
resulting comparison set consists of 158 sample periods: 0 from Carlsbad Caverns
National Park, 36 from Canyonlands National Park, 47 from Cedar Mountain and 75
from Yellowstone National Park. The operational problems of the VI resulted
largely from the lack of highly trained field operators to promptly identify
the subtle problems associated with this relatively complex sampler, the lower
priority for repairs and for preliminary data screening for the VI samplers
compared with the SFU and the length of time required for repairs (most repairs
were not performed in the field).
The average concentrations measured by the two samplers were generally in
96
-------
Fine Particles,
Lass Than a.5 Microns
Coarse Particles.
Greater Than a.5 Microns
Filter
Cassette
Fine
Particle
Filter
Inlet Tube
Virtual
Impactor
Nozzle
Virtual
Impactor
Receiver Tube
Filter Cassette
Coarse
Particle
Filter.
37mm Dia.
Filter
Holder
Figure B-l. Virtual Impactor particle sizing mechanism.
97
-------
TABLE B-l. COMPARISON OF VIRTUAL IMPACTOR AND STACKED FILTER UNIT
VI/SFU VI/Total Filter SFU/Total Filter
Si:
Coarse
Fine
Sum
Jil
Coarse
Fine
Sum
Ca:
Coarse
Fine
Sum
Fe:
Coarse
Fine
Sum
Pb:
*Coarse
Fine
*Sum
li
*Coarse
Fine
*Sum
Total Mass:
Coarse
Fine
Sum
0.91 ± 0.11
0.68 ± 0.25
0.91 ± 0.10
0.90 ± 0.12
1.08 ± 0.09
0.92 ± 0.11
0.87 ± 0.12
0.88 ± 0.12
0.84 ± 0.15
1.02 ± 0.15
0.59 ± 0.07
1.20 ± 0.45
0.40 ± 0.04
1.16 ± 0.21
0.86 ± 0.11
0.40 ± 0.04
1.16 ± 0.21
0.86 ± 0.11
0.83 ± 0.16
1.23 ± 0.13
0.93 ± 0.14
0.84 ± 0.10
1.01 ± 0.25
0.80 ± 0.28
1.09 ± 0.34
0.69 ± 0.14
0.96 ± 0.05
0.91 ± 0.11
1.07 ± 0.28
0.97 ± 0.32
1.00 ± 0.41
0.75 ± 0.33
1.06 ± 0.21
*Based on Two Samples
98
-------
2.0
1.5
1.0
0.5
1.5
1.0
0.6
Total
so,
D D
LL LL
(0 (0
I I I I I I I I
RPTAEOMNRCLSGHI JK
Sampler Identification
1 1 1 1 1 1
:'} H
i i i i i i
i i i i i i i i i i i
Small
*JS !}
D 3 D 0 38
w w to ° 36
1 1 1 1 i i i i I 1 1
ROBUPOMNRCLSGI(O)
Sampler Identification
Figure B-2. Sulfur/sulfate results from Charleston study,
99
-------
TABLE B-2. COMPARISONS OF PARTICLE COLLECTION EFFICIENCY - ALL SIZES SUMMED
Reference
Parker
et al
Ferek
et al
Camp
et al
Cahill
et al
Site
Durham
Los Alamos
Charleston
Dav i s
Samplers
Compared
SFU/VI
SFU/VI
SFU/Mean
SFU/Median
SFU/TF
Soils
0.78
1.30
0.87±
0.17
0.89
1.06
Ratio of Observed
Sulfur
1.07
0.85
0.94±
0.02
1.01
1.01
Masses
Lead
1.00
0.85
0.94±
0.04
0.97
0.90
Mass
-
-
0.95±
0.14*
0.98
0.99
* = Sum of Elements
TF = Total Filter
TABLE B-3. COMPARISONS OF PARTICLE SIZING EFFICIENCY, SFU vs. VI.
FRACTION OF MASS ON FINE STAGE
Site
Year
Samplers
Soil Sulfur Lead
Reference
Durham 1967 SFU: (12ym, 12.7 cm/s)* 0.11 0.74
VI 0.12 0.83
Los Alamos 1976
Charleston 1977
SFU: (8ym, 7.4 cm/s) 0.13 0.83
VI 0.15 0.90
SFU: (12ym, 12.7 cm/s
and 8ym, 7.4 cm/s) 0.25** 0.92
VI: 0.10 0.89
0.68
0.74
0.75
0.83
0.81
0.76
Parker et al
(1977)
Ferek et al
(1978)
Camp et al
(1978)
*Nuclepore filter pore size, face velocity
**UCD SFU (8 um, 7.4 cm/s) - 0.15
100
-------
D
LL
Sulfur
VI
Figure B-3. Scattergram comparison of stacked filter unit and virtual
impactor fine sulfur.
101
-------
good agreement. Table B-4 compares the arithmetic mean concentrations of selected
elements and mass for the three locations. For the coarse particles, the SFU
mean concentrations are about 15% larger than the VI concentration- For fine
sulfur and fine mass, the SFU levels are about 20% and 10% larger, respectively.
TABLE B-4. STACKED FILTER UNIT AND VIRTUAL IMPACTOR CONCENTRATIONS, (ng/m3)
Canyonlands Cedar Mountain Yellowstone
Coarse
Fine
Si
Al
K
Ca
Fe
mass
S
K
Ca
Fe
mass
SFU
859
221
147
292
147
5597
310
76
82
50
3353
VI
828
207
139
272
135
5722
307
69
45
28
4747
SFU
VI
1.04
1.07
1.06
1.07
1.09
0.98
1.01
1.10
1.82
1.79
0.70
SFU
587
151
58
216
79
3643
315
25
41
26
2966
VI
481
138
50
216
88
3497
230
21
27
17
2430
SFU
VI
1.22
1.09
1.16
1.00
0.90
1.05
1.37
1.19
1.52
1.53
1.22
:=======:
SFU
360
91
43
72
54
2221
113
20
19
15
1627
VI
254
52
39
67
46
1581
89
20
23
9
1268
SFU
VI
1.42
1.75
1.10
1.07
1.17
1.40
1.27
1.00
0.83
1.67
1.29
For fine soils, the SFU values are around 50% larger.
The VI also verified the measurements of relative composition by the SFU.
For example, the soil elements on each sampler were highly intercorrelated and
the ratios of the elements for the two samplers were virtually identical.
The VI also verified the coarse soil to coarse mass ratio measured throughout
the network. At all six samplers in the comparison, the sum of the coarse soil
elements with their associated oxide accounted for about half of the coarse
mass. The relative composition of the fine particles was also similar. The
ratio of fine sulfur to fine mass were similar at each site (0.07 and 0.09 at
Canyonlands, 0.11 at Cedar Mountain and 0.07 and 0.08 at Yellowstone). At
Cedar Mountain, the correlation between fine mass and fine sulfur was nearly
the same for each sampler (0.70). The correlations were lower at Canyonlands
and Yellowstone, though marginally higher on the SFU's than on the VI's.
While the mean values for sulfur and other elements seen by SFU's and VI's
in the field are reasonably close the correlation coefficients and intersite
variations are disappointing. Problems were encountered siting virtual impactors
at some locations, and mechanical problems resulted in less than 50% of all
possible samples being received. Since the units were able to operate only
towards the end of the program, quality assurance protocols were still being
modified for the VI's when sampling ended in September, 1981. The VI filters
were uncoated and serious sample loss in transport was documented. Photographic
evidence was obtained of nonuniform deposits on the coarse particle VI stage,
perhaps caused by the movement of loose sample material. Thus, the fact that
SFU's saw more mass than VI's is not suprising.
102
-------
SPECIAL STUDIES
In addition to the routine operation of the network several special
studies were conducted to get detailed information on particle size and chemical
composition. The first involved cooperation with the VISTTA study in June-July
1979 at Zilnez Mesa. Manual (early model) SFU's and multiday impactors were
operated side by side with automated (Beckman) virtual impactors run in a
manual mode, Cal Tech low pressure impactors, electrical mobility analyzers,
nephelometers, and other instruments. There was excellent agreement between
diverse instruments as shown in Figures 3 and 4 of Macias et al. (1981a). Some
of the results of the UCD work are shown in Table B-5. Despite differences in
flow rate (5 1pm to 20 1pm for the SFU's), sampler intakes (with and without
decclerators), and integration periods (one to three days), good agreement was
observed for fine particles. Typical agreement for sulfur was within 5%. Even
more impressive was the ±3% agreement between one- and two-day SFU's for
gravimetric mass, since the fine mass averaged only about 5 pg/rn^ and the
coarse mass about 12 ug/m
103
-------
TABLE B-5. VISTTA DATA - ZILNEZ MESA
SULFUR (ng/m3)
Multiday Impactor3
Stacked Filter unit*3
Date
6/26
6/27
6/28
6/29
7/02
7/03
7/04
7/05
7/06
7/07
7/08
7/09
7/10
7/11
Stage 1
6
9
4
47
5
6
4
5
4
6
8
13
12
19
Stage 2
37
54
65
147
63
50
69
43
24
24
49
18
26
25
Stage 3
152
155
715
827
396
185
226
304
127
175
204
183
223
325
Sum
195
218
784
1021
464
242
299
352
156
206
261
213
260
370
Coarse
23
38
55
51
60
Fine
447
792
432
195
312
Sum*
470**
830**
487
246
372
aStage 1 - 15 ym to 3.6 um; Stage 2 - 3.6ym to 0.65 ym; Stage 3 - less than
0.65 pm D
bl day SFU's only run on these 5 days; cut at 2.5 pm
*l-day SFU's ratio to Multiday is 0.97 ± 0.11. SFU were also run for ?.- and
3-day durations giving ratios to the Multiday of 1.04 ± 0.06 and 0.97 ± 0.14
repectively.
**Time periods not identical to Multiday; off by 10 hrs.
104
-------
APPENDIX C
QUALITY ASSURANCE PROCEDURES
INTERNAL QUALITY ASSURANCE
Sample Handling
When filters were received from Nuclepore Corporation, they were first
checked to see that they were as ordered. If not, they were not used and
replacements were obtained. No further checks were made until exposed filters
were returned from the field.
When filters were received from the field, the sample log sheet was
reviewed for gross errors. The sample dates and elapsed time were checked for
consistency. If not consistent, and no explanatory note was received, the
operator was telephoned to resolve the problem. If it was not resolvable, the
sample was placed in a "dead file" called Tray X for storage. The samples in
Tray X were never used for network purposes, but were stored in case they were
needed for special purposes. The flow rates were checked next. If the end flow
rate was less than 7 1pm, the sample was placed in Tray X. If the elapsed time
was less than 1,000 minutes or greater than 10,000 minutes, the sample was
placed in Tray X. Further elapsed time checks were made during computer
processing of data. If any note was received from the operator describing a
problem with the sample, the problem was resolved or the sample was placed in
Tray X.
The filters were checked next. If a filter was torn upon receipt, neither
stage was used, but both were placed in Tray X. If a filter had a finger-
print, only that filter was placed in Tray X. If the holder grid was installed
upside down, a note was made to mount the sample with the wide bars oriented
perpendicular to the beam. If the filter was torn during weighing, it was
salvaged, if possible.
The data was again reviewed when the filter weights were recorded in the
sample weight logs. If two sets of data existed for the same filter number,
the electrebalance printed record and shipping dates were examined to differ-
entiate the two. The second set was marked by a prime (') on the filter number.
If the sample weight was very small or negative, the balance printer record was
examined for transcription errors, and records for other sites and sample
periods near the affected one were examined for errors or inconsistencies. If
any problems were found, they were resolved or the affected samples were placed
in Tray X. Before an analysis run, all trays were checked against the sample
weight log for consistency.
105
-------
Computer Log and Run Sheets
After all entries for a given month were made in the sample weight log
and checked, a computer log was prepared. Entries in the computer log followed
those in the weight log closely. Each entry listed the start date, elapsed
time, initial and final rotameter readings and coarse and fine masses. (Only
the differences between post- and pre-weight were entered). Masses less than
10 ygm were entered as 10 ugm. If a coarse stage was missing, and a fine stage
was present, a zero coarse mass was entered. All entries in the computer log
were checked with those in the weight log. Gross typing errors were usually
identified by misaligned columns.
A printout was made using a specially written computer program. This
included the original entries plus calculated parameters. Sample duration
was printed: if it ran more than 48 hours beyond the normal 72-hour duration
an error message was printed. The flow rates in liters per minute were corrected
for altitude and rotameter bias. The printout listed the conversion constant
(crn^/m^) and the coarse and fine masses in yg/rn. The printout was reviewed for
anomalous data (unusual run times, conversion constants, masses etc.).
Run sheets to accompany the PIXE analysis were also prepared by computer.
These indicated which positions in the trays had valid samples. In order to
check that every sample had an entry in the run sheet, all empty positions in
the trays were compared to corresponding blanks on the run sheets. The actual
instructions to tell the analysis program which slides to analyze were contained
in files which were typed in using these computer generated run sheets. This
was generally a simple task since only missing samples needed to be identified.
After the entries in this file were checked, a note was made on the run sheet.
PIXE Analysis
Selected elemental foils were analyzed at the start of each cyclotron run.
These standards verified that the analysis system was working properly and
provided a normalization constant to account for changes in beam energy, charge
collection and detector solid angle.
Before the network samples were analyzed a reanalysis tray of SFU samples
was run. The concentrations obtained were compared to previously measured
values. If they were predominantly within 10-15% of the past values, the run
proceeded. If there was a bias, then the elemental foils were rerun. Any
problems were identified and corrected before the run proceeded.
Throughout the run the operators periodically checked the analyses to
verify that no problems occurred. The analysis program printed special messages
of problems: low charge (below 750 nanocoulombs), bad blank subtraction or no
peaks found (indicating a bad sample), and unidentified peaks (usually denoting
energy miscalibration). When these messages occurred the operator isolated the
problem, often rerunning the sample. The operator commented on these problems
on the run sheet.
At the end of the network samples the reanalysis tray was run again. If
these values disagreed with the pre-run values, then selected samples throughout
106
-------
the run were reanalyzed to determine when the problems began. If necessary, a
portion of the samples were reanalyzed.
Post Analysis Procedures
At the end of the network analysis, but before the cyclotron was shut
down, the analysis supervisor reviewed the PIXE run sheets and the PIXE analysis
printout log. If any analyses were missing or irretrievably bad, the samples
were reanalyzed at that time. Spectra needing to be reanalyzed via the computer
were noted.
The elemental masses for each run were collated into summary files by the
analysis program. These files were printed and reviewed. Samples in which key
elements were missing or in which amounts were abnormally large or small were
noted. For the coarse stage, the key elements were silicon, calcium and iron,
while for the fine stage it was sulfur. These spectra, along with those noted
from the previous paragraph, were reanalyzed with the computer to validate the
results. If problems in the spectrum were suspected, the sample was reanalyzed.
EXTERNAL QUALITY ASSURANCE PROTOCOLS
Rockwell International Environmental Monitoring and Service Center under
contract to EPA conducted a program of external audits on the operation of this
network. The program consisted of:
o weight audits
o flow audits
o field audits
o filter analysis system audit
o system audit
Weight Audits
The purpose of the weighing audit was to evaluate the accuracy of the UCD
filter weighing procedure and weighing equipment. There were seven such audits
from July 1980 to April 1982.
At Rockwell a set of six aluminum weights (3 pair) was used to perform the
audit. Each weight was altered by filing away some of the mass so that the
weight difference between a pair of weights would simulate the particular
loading of a typical filter sample. The exact weight of each mass prior to
shipping was determined by performing three weighings conducted on three separate
days and averaging the results. The weighings were done on a Cahn Model 26
Electrobalance which had been calibrated with NBS traceable, Class M weights.
The set of weights was then sent to UCD together with instructions and a
data sheet. UCD personnel were instructed to weigh each mass using the same
procedures used for filter weighings. To perform the weighings, UCD used a
Cahn Model 25 Electrobalance calibrated with NBS traceable, Class M weights.
107
-------
The masses and the completed data sheet were then returned to Rockwell.
The masses were then reweighed on the Cahn Electrobalance to Insure that
no significant change had occurred during transit. This was done by weighing
each mass three times, on three separate days. If no significant change was
found, the "official" Rockwell weight of each mass would be the average of six
weighings. A comparison between Rockwell and UCD values was then made. Two
comparisons of the data were made. One compared the weights of the individual
masses obtained by Rockwell and those obtained by UCD. The other was a com-
parison of the weight differences between pairs of masses. All the weight
audits resulted in good agreement between Rockwell and UCD. An example of a
typical weight audit (August 1981) is shown in Table C-l.
Flow Audits
For the purpose of conducting flow rate performance audits of all the
SFU's, Rockwell sent calibrated orifice flow measuring devices together with
audit procedures to the sampler operators in the field. The device was attached
to the inlet of the SFU allowing the entire sample flow to pass through the
orifice. The pressure drop across the orifice, measured in inches of water
with the magnehelic gauge, was recorded and used to calculate audit flow. The
audit flow was then compared to the flow as indicated by the sampler's rotameter
which was calculated by Rockwell using the calibration curves provided by UCD.
Table C-2 show the results of the audits for August 1980 to August 1981,
respectively. The site flow values in parenthesis represent flows modified by
a temperature correction not used by UCD except for extreme temperatures. The
audit flow values in parenthesis represent flows assuming the site operator
misread the magnehelic gauge (e.g., 0.19 instead of 1.19 inches of water).
Misreading of the gauge was eliminated in subsequent audits by having the
operator draw the position of the needle on an illustration of the gauge face.
The values in parenthesis in the percent difference column is appropriate
assuming the parenthetic flow value is correct.
Field Audits
In addition to the flow audits, on-site field audits were performed by
Rockwell at selected sites. The purpose of the audits was to assess the perfor-
mance of the personnel and equipment and to provide technical assistance as
required, in order to control and improve the quality of the data. The system
audits involved interviews with personnel and a visit to the sampling site in
order to evaluate the following elements of the program: organization and staffing,
training, site selection, equipment, procedures, maintenance, documentation and
communication. Field audits were performed in the fall of 1980 and the fall of
1981. In general Rockwell reported a satisfactory field operation though they
did make recommendations having to do with improved or better documented operator
procedures and communications between the field and UCD.
108
-------
TABLE C-l. WEIGHT AUDIT RESULTS (mg)
Mass No.
1
2
3
4
5
6
*Difference
Rockwell
Mass
Mass Difference
9.950
0.894
10.844
13.089
0.573
12.516
17.951
0.707
17.244
= Davis - Rockwell
UCD
Mass
Mass Difference
9.956
0.894
10.850
13.095
0.572
12.523
17.955
0.708
17.247
Difference*
Mass
0.006
0.006
0.006
0.007
0.004
0.003
Mass
Difference
n.ooo
0.001
0.001
109
-------
TABLE C-2. FLOW AUDIT RESULTS
Site No.
1
*2
3
4
5
6
7
8
9
10
11
12
13
14
*15
*16
*17
*18
19
20
*21
*22
23
24
25
26
27
28
29
30
31
*32
33
34
35
36
37
38
39
40
Mean ± Standard
*°L Difference
Site flow
(1pm)
12.5
10.3
9.0
10.4
10.1
10.3
9.4
10.2
10.1
9.6
10.5
Invalid
11.5
9.8
11.1
10.2 (9.5)
11.9 (11.2)
8.8
11.8
9.7
10.9
10.7 (10.1)
10.7
11.2
10.7
10.6
8.6
10.7
8.6
9.0
11.0
12.1
10.7
11.1
11.4
9.0
10.5
10.9
9.3
12.3
Deviation
= Site flow - Audit
Audit Flow
Opm)
11.6
4.9 (9.7)**
9.3
12.5
11.1
11.9
12.6
11.7
11.3
12.1
11.1
I nv a 1 i d
11.9
8.8
4.2 (10.1)
8.9
11.2
5.9 (10.9)
11.1
11.2
4.5 (9.4)
9.0
12.4
10.4
11.3
8.0
10.4
11.6
10.9
9.5
11.5
4.2 (10.3)
9.0
11.7
12.7
7.5
10.8
11.2
10.4
12.2
Flow x 100
% Diff.*
7.8
110.0 (-1.0)
-3.2
-16.8
-9.0
-13.0
-25.0
-12.8
-10.6
-20.7
-5.4
Invalid
-3.4
11.4
164.0 (9.9)
14.6 (6.7)
6.3 0)
49.2 (19.0)
6.3
-13.4
142.0 (16.0)
18.9 (12.2)
-13.7
7.7
-5.3
32.5
-17.3
-7.8
-21.1
-5.3
-4.3
188.0 (17.5)
18.9
-5.1
-10.2
+20.0
-2.8
-2.7
-10.6
0.8
0.9 ± 13.3
Audit Flow
**Parenthetic values are explained in the text.
110
-------
Filter Analysis System Audit
Four filter analysis audits were done between fall 1980 and spring
1982. In the first, 24 identical audit samples were sent to UCD as a check of
internal precision while in the remaining audits equivalent samples were sent
to UCD and the NEA Research Laboratories, Beaverton, Oregon, for XRF analysis.
The results of the first and second audits are summarized in Table C-3 and
Figure C-l respectively. Subsequent audits produced similar results. Rockwell
concluded that "The UCD PIXE technique is fundamentally sound and is capable of
measuring elemental concentration on particulate filters."
System Audits
Rockwell representatives visited UCD in May of 1980 and 1982. The purpose
of these visits was to examine the UCD facilities and evaluate the perfor-
mance in carrying out the Western Particulate Characterization Study. Overall,
Rockwell found that "UCD is doing a very good job".
TABLE C-3. ELEMENTAL ANALYSIS AUDIT RESULTS, 21 SAMPLES ANALYZED
Standard Percent Standard
Mean Deviation Deviation
Element (yg/m3) (yg/m3) (%)
Cu 46.47 3.28 7.06
S 22.19 1.83 8.25
Al 44.44 3.79 8.54
Ca 45.67 3.57 7.81
Ni 36.28 2.74 7.56
Fe 42.93 3.23 7.51
Mn 36.54 3.03 8.30
Cl 13.63 1.33 9.72
K 49.53 4.00 8.08
Br 22.56 1.61 7.13
111
-------
Normalized
Aerosol
Concentration
(ug.nV3.g"M)
60
40
30
20-
10
J
NEA
JUCD
1 Probable
J Range at
ft 10% Uncertainty
o «r
Al S Cl K Ca Mn Fe Ni Cu Br
Figure C-l. Comparison of elemental analysis. Mean and standard deviation
shown for eight samples. Results normalized to typical
aerosol concentration.
112
------- |