EPA-910/9-75-010
May 1975
A STUDY
OF THE SUSPENDED
PARTICULATE PROBLEM
IN THE DUWAMISH BASIN
U.S. ENVIRONMENTAL PROTECTION AGENCY
REGION X, Surveillance and Analysis Division
Seattle, Washington 93101
-------
This report has been reviewed by the Environmental Protection Agency
and approved for publication. Approval does not signify that contents
necessarily reflect the views and policies of the Agency.
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
EPA Laboratory, 1555 Alaskan Kay South, Seattle, V7A 98134
DP-,',:
SUBJI-.C1: Analytical Approach for Identifying Pa
Source"Contributions
t ROM: Jim Littlejohn
Air Quality Assurance Officer
TO; Sea Addressees
THRU: Arnold Gahler
Chief, Laboratory Branch
y t , * T k4 r ^ 7 r _
r ' 9^JW'75
<:/?s
Enclosed is a report of a study recently completed by Region X.
The objective of the study was to determine the impact of various
particulate sources in Seattle's main industrial area. The study
included the identification of major particulate sources plus an
evaluation of analytical techniques.
A wide range of analytical techniques were used to identify
jarticulate source contributions on hi-vol filters. Gravimetric
md organic analyses -were performed on glass fiber filters. Elemental
md compound (Optical microscopy and X-Ray diffraction) analyses
/ere performed on membrane filters.
The analytical data was then correlated with the chemical
;osposition of specific sources. Cocb ining this information with
aeteorological data allowed a quantitative estimation of particulate
sources. The data showed the following contribution of sources:
% Contribution to Total
Natural Sources 27
Pollen and spores, wind erosion,
open burning and biological
materials
Transportation 39
Transportation, vehicles
& road dust
Industry 34
Point and multiple source areas
This study is an example of how advanced analytical techniques
can be used to determine the source contribution on the air quality
of a specific area. If you would like to discuss this report in any
greater detail, please contact me.
Addressees: f? Ei(^/Q ,y>/f [_ /?//'. Qoft9UI~Y c tf
€<:•«<»>* ITE? $
-------
EPA-68-02-1499
May 1975
A STUDY OF THE SUSPENDED PARTICULATE PROBLEM
IN THE DUWAMISH BASIN
by
Ray H. Olsen, Marcia Y. Almassy,
and A. Lewis Wingert
The Boeing Company
Seattle, Washington 98124
Contract No. 68-02-1499
Project Officer
James Littlejohn
Prepared for
Region 10
Surveillance and Analysis Division
Environmental Protection Agency
Seattle, Washington 98101
-------
ABSTRACT
Air quality data accumulated since 1965 indicate the primary and secondary national air
quality standards have been exceeded in the Duwamish Basin area of Seattle, Washington.
The objective of this study was to determine the nature of suspended particulate and
subsequently to quantify the impact of particulate sources. Ambient and source particulate
was collected on fiberglass and membrane filters. Results from gravimetric, elemental, and
compound analyses were combined with meteorological data for correlation and analyses.
The contribution of sources for the basin was found to be 27% from natural sources, 39%
transportation, and 34% point industry and multiple-area sources. Results from this study
show the complexity of suspended particulate. Source tests showed a multitude of elements
and compounds present and, to complicate matters, surrounding soil has been contaminated
by industrial and area sources. About 35% of the particulate is directly related to road
dust-type emissions.
This report was submitted in fulfillment of contract number 68-02-1499 by The Boeing
Company under the sponsorship of the Environmental Protection Agency. Work was
completed as of April 16, 1975.
ii
-------
CONTENTS
Page
Abstract ii
List of Figures iv
List of Tables v
Acknowledgments vi
Sections
I Conclusions 1
II Recommendations 4
III Introduction 5
IV Study Objectives and Approach 7
V Procedures for Sampling and Analysis 10
VI Results and Discussion 19
VII References 63
VIII Glossary of Abbreviations 65
Appendix 66
iii
-------
FIGURES
No. Page
1 Duwamish Basin Air Monitoring Sites 8
2 Sample Flow-Through Analyses 9
3 Representative Photographs of Some Particle Classes 18
4 Particulate Concentration, Daily Arithmetic Average With Three-Point
Smoothing Function Applied 24
5 Particulate Concentration, Arithmetic Averages by Site and Wind Direction . . 25
6 Geographic Locations of Particulate Sources 26
7 Concentration of Particulates and Ashable Organics, Arithmetic Averages
by Site and Day of Week (Northerly Winds Only) 27
8 Infrared Spectrograms of Chloroform Extracts 34
9 Infrared Spectrogram of Xylene Extract From K.99 (8-21-74) 35
10 TLC Overlay of Chloroform Extracts 36
1 1 TLC Overlays of Xylene Extract From K99 (8-21-74) . 37
12 Normalized Bar Graphs of Mass Spectra Obtained From Extracts of
K99 (9-21-74) 38
13 Normalized Bar Graphs of Mass Spectra Obtained From Total
Chloroform Extracts 39
14 Normalized Bar Graphs of Mass Spectra Obtained From Diethylether/n-Heptane
TLC Separations of Chloroform Extract of K59 (7-25-74) Total Extract and
Zone R1 40
15 Normalized Bar Graphs of Mass Spectra Obtained From Diethylether/n-Heptane
TLC Separations of Chloroform Extract of K59 (7-25-74)—Zones R2 and R3 . . 41
16 Normalized Bar Graphs of Mass Spectra Obtained From Diethylether/n-Heptane
TLC Separations of Chloroform Extract of K59 (7-25-74)—Zones R4 and R5 . . 42
17 Normalized Bar Graphs of Mass Spectra Obtained From Diethylether/n-Heptane
TLC Separations of Chloroform Extract of Gasoline Engine Exhaust 43
19 Mean densities of Samples by Site 51
20 Representative SEM Photographs (K60, 8-21-74) 57
21 Average Concentration of Aluminum by Site (Northerly Winds) 67
22 Average Concentration of Silicon by Site (Northerly Winds) ........ 68
23 Average Concentration of Sulphur by Site (Northerly Winds) 69
24 Average Concentration of Chlorine by Site (Northerly Winds) 70
25 Average Concentration of Potassium by Site (Northerly Winds) 71
26 Average Concentration of Calcium by Site (Northerly Winds) 72
27 Average Concentration of Titanium by Site (Northerly Winds) 73
28 Average Concentration of Vanadium by Site (Northerly Winds) ....... 74
29 Average Concentration of Manganese by Site (Northerly Winds) ....... 75
30 Average Concentration of Iron by Site (Northerly Winds) 76
31 Average Concentration of Nickel by Site (Northerly Winds) 77
32 Average Concentration of Copper by Site (Northerly Winds) 78
33 Average Concentration of Zinc by Site (Northerly Winds) . 79
34 Average Concentration of Lead by Site (Northerly Winds) 80
35 Average Concentration of Niobium by Site (Northerly Winds) 81
iv
-------
No
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
TABLES
Page
Suspended Particulate Levels in the Duwamish Basin 6
Site Location Information 11
Source Samples 12
Particle Class Criteria 16
Meteorological and Gravimetric Data for Generally Northerly Wind Days ... 20
Meteorological and Gravimetric Data for Generally Southerly Wind Days ... 22
Meteorological and Gravimetric Data for Days Having Variable Wind Conditions 23
Total Ashable Organics: Ambient Air Test Filters 29
Total Ashable Organics; Source Test Filters 30
Concentration of Ashable Organics -Northerly Winds 31
Organic Content by Soxhlet Extraction 32
Precise Mass Values and Element List for TLC Zones of K59 (7-25-74) .... 44
Analysis of Source Tests 49
Particle Size Distribution by Site 50
Semiquantitative Particulate Classification by Site 53
X-Ray Diffraction Analysis of Air and Source Tests 58
Statistical F Values Calculated From 75 Northerly Wind Samples 59
Average Concentration of Variables by Site 61
Symmetric Correlation Matrix Calculated From 75 Northerly Wind Samples , . 62
v
-------
ACKNOWLEDGMENTS
Many individuals and organizations have contributed generously of their time, expertise and
equipment. The authors are especially grateful to the following organizations for the use of
their equipment and stations located in the Duwamish Basin: the Puget Sound Air Pollution
Control Agency (PSAPCA) (K55, K59, K60 and K6 1), Seattle City Light (K64), and the
Washington State Department of Ecology (K99) which also furnished additional high
volume (Hi-Vol) samplers and calibrated all of the air flow meters.
The following individuals made contributions through their highly professional advice,
encouragement, and assistance:
1. G. C. Hofer, Environmental Protection Agency
1. J. Littlejohn, Environmental Protection Agency
3. A. Dammkoehler, PSAPCA
4. A. Kellogg, PSAPCA
5. M. Svoboda, PSAPCA
6. G. R. Freeman, Washington State Department of Ecology
vi
-------
SECTION I
CONCLUSIONS
The Duwamish Basin area is a heavily concentrated industrial region of South Seattle,
Washington. Air quality data accumulated since 1965 indicate the primary and secondary
national air quality standards have been exceeded. Computer modeling studies predict only
borderline compliance with 1975 secondary standards. The objective of this study was to
determine the nature of suspended particulate in the Duwamish Basin and subsequently
quantify the impact of specific suspended particulate sources. A network, of six sites each
equipped with two high volume air samplers (Hi-Vol) for simultaneous collection of
particulate on fiberglass and membrane filters was operated from July through November of
1974. Results from gravimetric, elemental, and compound analyses were combined with
meteorological data for correlation and analysis.
The analytical approach toward identifying particulate types is a positive step from simply
making mass measurements. This approach leads toward identification of airborne
particulate, giving environmental agencies an opportunity to make assessments of size
distribution, chemical elements, and compound effects on public health and well being.
Additionally, through positive identification of particulate, positive corrective action can be
taken.
Based on the results of this rather limited study, several conclusions can be drawn:
1. The particulate loading for northerly winds was higher than southerly winds.
N-I02.3^g/m3, S -52.7/ig/m3.
2. The highest particulate loading was found in the northerly sites and generally decreased
in southerly sites, regardless of wind direction.
3. Particulate loading was highest during the work week and lower on weekends and
holidays.
4. There was a measurable seasonal variation in particulate loading, with the highest
loading occurring in September and early October.
5. The ashable organic concentration ranged from 8.2-44.6/ig/m3 for northerly winds.
The highest concentrations were found during the work week.
6. Chloroform extracts indicated the same types of organic compounds were found from
day to day and from site to site.
7. The major portion of chloroform extractable organics was identified as gasoline engine
exhaust. Chloroform extracts represented less than half of the total ashable organics,
the remainder being extractable with xylene, and contained highly unsaturated
polyaromatic hydrocarbons.
1
-------
8. Iron, manganese, zinc, and to a lesser extent, lead, correlated with each other and
decreased from north to south under northerly wind conditions. Iron as an example,
averaged 14*ig/m3 at the norther site and 4jutg/m3 at the southernmost site. Iron, lead,
and zinc were identified as major constituents in steel mill emissions.
9. Calcium, aluminum, silicon, sulfur, potassium, and titanium correlated with each other.
These elements were identified in road dust, cement, and gypsum emissions. Particle
size analysis indicated a homogeneous particle size distribution with a slight trend
toward decreasing size from north to south.
10. The average physical particle size of the particulate is less than 3pi. Northern sites
showed 66-69% of particles 0.5-3/n, whereas southern sites showed 71-75% 0.5-3ju, There
were many extremely small particles (less than 0.1 jut) found in the ambient air samples.
11. The Duwamish road dust appears to be contaminated from industrial and area sources.
12. The particle classes showing the highest concentration by weight include furnace slags
and spheres, fly ash, pollens and spores, concrete minerals, road dust minerals,
and soot.
Particle classes Maximum Concentration observed, wt%
Furnace slags 10
Furnace spheres 20
Furnace ash 8
Fly ash 8
Concrete minerals 10
Road dust minerals 45
Soot 20
Starch 10
Pollens and spores 20
13. Semiquantitative analysis of particulate by sources resulted in the following averages:
Natural (Pollen and spores, wind erosion, open burning
biological materials) 27%
Transportation (Transportation vehicle and road dust) 39%
Industry (Point and multiple source areas) 34%
14. Particulate loading increased with the intensity and duration of temperature inversion,
and also increased for approximately 200 hours after rainfall, then appeared to level
off.
15. The variation in particulate concentration was largely accounted for by the variation in
ashable organics and iron,
16. Results from this study indicate that suspended particulate is extremely complex.
Source tests showed a multitude of elements and compounds present in each source,
and to complicate matters, the surrounding soil had become contaminated from
2
-------
industrial and area sources. Thus, control strategies must be more encompassing than
simply controlling industrial and vehicular emissions. Consideration must also be given
to cleaning up roads and parking lots and seeding barren areas. About 35 wt% of
particulate is directly related to these sources.
3
-------
SECTION II
RECOMMENDATIONS
To make an evaluation of what control strategies would be most effective, further work
must be done in understanding the contributions of the various sources. Specific
recommendations include:
1. Pinpoint specific sources more closely using directional sampling coupled with a
thorough understanding of emissions.
2. Locate new site(s) north of K60 to determine background level of air reaching
Duwamish on north wind days.
3. Conduct further analyses on the organic portion.
4. Through further sampling, particle sizing, chemistry and multiple correlation analyses
using select meteorological criteria, determine the contribution of sources and
re-entrainment.
5. Future studies of industrialized areas should be planned to cover a longer period of
time so that a sufficient number of samples could be collected, analyzed and
interpreted during the first phase of the study. Secondary phases would follow with
redirection of sampling and/or analytical techniques to obtain the most meaningful
data. Additionally, it should be recognized that meteorological conditions are an
extremely important variable in a study of this type and that this is a completely
uncontrollable parameter. Therefore, sufficient program time must be allowed to pick
and choose the sampling days with appropriate conditions.
4
-------
SECTION III
INTRODUCTION
The Duwamish Basin-Harbor Island area is a heavily concentrated industrial region of South
Seattle, Washington. It is bordered on the eastern and western sides by rather abrupt valley
walls rising to an average ridge height of 350 feet above mean sea level; it opens to Puget
Sound through Elliott Bay on the north and to a flat agricultural area on the south.
Air quality data compiled by the Puget Sound Air Pollution Control Agency (PSAPCA)
since 1965 indicate that both the primary and secondary national air quality standards have
been exceeded in the Duwamish Basin-Harbor Island industrial area of South Seattle shown
in table 1.
Although it appears that the air quality is improving, the area still exhibits visibility
problems, soiling, and other evidence of poor air quality. The reductions in 24-hour
gravimetric accumulations reported in table 1 may not be a good indicator of improved air
quality since the larger particulate was the first to be removed by industrial clean-up
methods. Boeing studies carried out in 19731 indicated that more than 50% of the airborne
particulate was of respirable size of three microns or smaller.
When existing industry is in compliance with present regulations, PSAPCA estimates that
1200 tons of particulate will be emitted per year. Other sources such as vessels, trains,
automobiles, home heating (excluding dust entrained by wind or traffic) presently emit
1000 tons of particulate per year. PSAPCA personnel have applied the urban diffusion
model AQDM to the area, resulting in the prediction of only borderline compliance with the
1975 secondary standards." Recent studies have shown that the gravel roads in the
Duwamish Basin contribute 800 tons of particulate per year measuring less than
10 microns. This is a significant contribution to the particulate problem which was not
considered in the model above.
Several governmental and private agencies have been emphasizing the need to develop the
industrial potential of the Duwamish Basin in order to ensure economic growth and
long-range economic stability. Based on 1973 data (geometric mean = 68pg/m3)} with
virtually all existing point sources in compliance with Regulation 1,^ the study of the
suspended particulate problem is essential to provide background data for additional control
strategies.
5
-------
Table 1. SUSPENDED PARTICULATE LEVELS IN THE DUWAM1SH BASIN
Geometric
No. of 24-hr
mean
periods exceeding
Maximum,
Federal regulations for
Year
jug/m3
150
^jg/m3
suspended particulate
1965a
97
15
399
Primary standard: annual
1966a
121
42
394
average = 75/jg/m3; 1 day
maximum = 260 /;g/'m3
1967a
99
19
241
1968a
101
20
531
Secondary standard: annual
1969a
108
25
472
average = 60 ^ig/rn3; 1 day
maximum =150jzg/m3
3/70-12/70^
78
9
238
8/71-12/71c
68
1
164
1972c
81
5
306
1973°
68
1
190
aMeasured at Fire Station No. 14, 3224 4th Ave. South,
''Measured at 4600 E. Marginal Way South
cMeasured at 4500 E. Marginal Way South
6
-------
SECTION IV
STUDY OBJECTIVES AND APPROACH
OBJECTIVES AND SCOPE
The objectives of this study were to determine the nature and distribution of suspended
particulate in the Duwamish Basin and to subsequently quantify the impact of specific
suspended particulate sources. The scope of the program included the identification of
major particulate components and sources plus an evaluation of analytical techniques.
Emphasis was placed on identification of specific inorganic compounds as distinguished
from organic components since this was an initial study; however, the percent total organics
and some preliminary "fingerprinting" of organics were also accomplished.
APPROACH
A network of six air monitoring sites located in the Duwamish Basin was used (fig. 1). Each
site was equipped with two Hi-Vol samplers for simultaneous collection of particulate on
Gelman Type A, low pH, glass fiber filters and Acropor AN-800 membrane filters. The analy-
sis scheme for the two filters is shown in figure 2.
Gravimetric and organic (low-temperature ashing and initial fingerprinting) analyses were
performed on glass fiber filters. Elemental and compound (optical and X-ray diffraction)
analyses were performed on membrane filters.
Results from the gravimetric analysis, elemental analysis and low-temperature ashing
analysis for total amounts of organics present were combined with meteorological data in a
computer program so that a multiple regression analysis could be performed for correlation
and prediction of site-to-site pollution levels as a function of climatic conditions.
7
-------
Interstate 90"
WEST SEATTLE %
BOEINGX^
K61
V
LAKE WASHINGTON
RAINIER
VALLEY
BEACON HILL
MERCER
ISLAND
[K59
%
PUGETSOUND
\
%
BURIEN
' State Highw,y
%
'"iiiim
kTUKWILA'
Interstate 40b
SEATTLE-TACOMA
INTERNATIONAL
AIRPORT//m
Figure 1. Duwamish Basin air monitoring sites
8
-------
SAMPLE = FILTER PAIR
vo
RESULTS PARTICLE SIZE
~ DISTRIBUTION
RESULTS
TOTAL SUSPENDED PART.
PARTICULATE CONC. (/ug/m3)
RESULTS
*-W% ELEMENT, SEMIQUANTITATIVE
RESULTS
W%TOTAL
ASHABLE ORGANICS
QUALITATIVE
IDENTIFICATION
OF
INORGANIC
COMPOUNDS
RESULTS
RESULTS
CHARACTERIZE <
ORGANIC COMPOUNDS
SOME COMPOUND
IDENTIFICATION
RESULTS
SEMIQUANTITATIVE
IDENTIFICATION
BY WEIGHT OF PARTICULATE
XRDFOR
COMPOUNDS
MEMBRANE
STORAGE
GLASS
FIBER
LOW TEMP ASHING
GRAVIMETRIC
ANALYSIS
DETAILED
ORGANIC
ANALYSIS
OPTICAL ANALYSIS
DENSITY SEPARATION
OPTICAL
PARTICLE
SIZING
ELEMENTAL ANALYSIS
EDXRA
SEM PARTICULATE
MORPHOLOGY
ELECTRON MICROPROBE
ANALYSIS
EDXRA—ENERGY DISPERSIVE
X-RAY ANALYSIS
XRD—X-RAY DIFFRACTION
SEM-SCANNING ELECTRON MICROSCOPE
W—WEIGHT
Figure 2. Sample flow-through analyses
-------
SECTION V
PROCEDURES FOR SAMPLING AND ANALYSIS
SAMPLING
A network of six air monitoring sites located in the Duwamish Basin (fig, 1 and table 2) was
used to collect airborne particulate. Each site was equipped with two Hi-Vol samplers for
simultaneous collection of particulate on Gelman Type A, low pH glass fiber filters and
Acropor AN-800 membrane filters. Each Hi-Vol was calibrated and operated in accordance
with standard .procedures.5
Air sampling was conducted every third day from July 10, 1974 through July 18, 1974.
July 31, 1974 was the first day of predicted inversion conditions so three 8-hr sampling
periods were chosen for that one day. From August 3, 1974 through November 26, 1974
sampling was conducted on every sixth day with the exceptions of October 10th, chosen for
its southerly wind conditions; and October 16th, chosen because of the air stagnation alert.
With the exception of July 31, all sampling periods were planned to last for twenty-four
hours (00.00 to 24.00 Pacific Standard Time) so that enough particulate could be collected
for all of the analyses.
Stack samples from specific point sources were collected on glass and membrane filters.
Other point source samples were collected; these took the form of scrapings from towers,
baghouses, etc. See table 3 for a full listing of the source samples collected.
Source test samples were treated the same, analytically, as air test samples.
GRAVIMETRIC ANALYSIS
Gravimetric analysis of the glass fiber filters was performed in accordance with standard
procedures.5 Membrane filters were not analyzed gravimctrically due to their high
hygroscopic nature.
ORGANIC ASHING ANALYSIS
The weight percent of total ashable organics was determined using a low-temperature asher
for all of the fiberglass samples. Low-temperature ashing (LTA) completely oxidizes organic
substances from an inorganic matrix at relatively low temperatures (100°-150°C),
DETAILED ORGANIC ANALYSIS
Characterization and identification of organic compounds were performed on the glass fiber
filters. The organic material was removed from the filter by Soxhlet extraction in a suitable
solvent. Initial mass spectra of the total unknown indicated that further characterization
10
-------
Table 2. SITE LOCATION INFORMATION
Site code
Street address
aUTM (zone 10)
coordinates,
km
Elevation
above mean
sea level,
ft
Sampling height
above ground,
ft
K60
Weyerhaeuser Seattle Laboratory
3400 13th Avenue S.W.
Seattle, Washington
x = 548.89
y = 5269.00
20
18
K55
Metro Duwamish Pumping Station
4500 Block E. Marginal Way S,
Seattle, Washington
x= 549,77
y = 5267.74
14
34
K99
Department of Ecology Trailer
6770 E. Marginal Way S.
Seattle, Washington
x= 550,63
y = 5265.75
14
16
K61
Concord Elementary School
723 S. Concord Street
Seattle, Washington
x = 550.98
y = 5263.37
165
23
K64
Seattle City Light Trailer No. 1
10000 W, Marginal Way S.W,
Seattle, Washington
x= 552.32
y = 5262.34
20
18
K59
Duwamish Fire Station
King County Fire District No. 1
12026 42nd Avenue S.
Seattle, Washington
x = 554,32
y = 5260,65
25
28
aUTM—Universal Transverse Mereater
-------
Table 3. SOURCE SAMPLES
Source
Sampler
Remarks
Wood products
Hi-Vol
Outlet of cyclones used to control emissions from a
sander and other woodworking machines
Gypsum products
Hi-Vol
Fugitive emissions from conveyor belt which occurred
during ship unloading of gypsum rock
Harbor Island road dust
Hi-Vol
Road dust which occurred from driving a one-ton van
on a gravel road in a parking lot on Harbor Island
Gasoline engine exhaust
Hi-Vol
Tailpipe emissions from a 1968 one-ton van with a
gasoline engine
Diesel switch engine exhaust
Hi-Vol
Emissions from the stack of a railroad switchyard
engine,
Flour Mill
Hi-Vol
Emissions from cyclone on five-mill headhouse for
processing flour
Flour Mill
Hi-Vol
Emissions from a baghouse outlet used to treat the air
used in wheat protein concentrate mixing and processing
Steel plant
Hi-Vol
Emissions from the high-temperature baghouse used to
treat air exhausted from furnace hoods and two electric
arc furnaces
Steel plant
Hi-Vol
Emissions from low-temperature baghouse used to treat
air exhausted from roof of electric arc furnace building
Diesel truck exhaust
Hi-Vol
Exhaust emissions collected with a radar Hi-Vol sampling
system from a diesel truck engine, 318 HP with turbo
and supercharge
Diesel oil boiler
Hi-Vol
Emissions from a small (2 gallons/hour) diesel oil boiter
Cement plant
Hi-Voi
Emissions from stack
Cement plant
_
Fugitive dust collected from inside of cooling tower near
south boundary of cement plant
Cement plant
Dust taken from electrostatic precipitator screw conveyor.
This dust also goes out of the stack.
Cement plant
-
Type-One cement baghouse emission from finish mills
Road dust
Hi-Vol
Road dust generated by a truck at the Port of Seattle,
Terminal 106 W,
Battery processor
Hi-Vol
Fugitive emissions from reverb furnace baghouse
Battery processor
—
Dust caught in the baghouse on the reverb furnace
12
-------
and separation of the Soxhlet extracts were required for the positive identification of even a
few organic compounds; thin layer chromatography (TLC) fulfilled this need and also
provided a means of easily comparing the organic contents of different test filters.
A Soxhlet extraction using distilled chloroform was performed on ten glass fiber filters and
a subsequent Soxhlet extraction using distilled xylene was performed on one of these filters.
TLC separation was performed on each total extract; the resultant zones were compared and
collected for further characterization by mass spectroscopy. The total extract was subjected
to analysis by both mass spectroscopy and infrared spectroscopy.
ELEMENTAL ANALYSIS
Elemental analysis of the collected particulate was accomplished through the use of energy
dispersive X-ray (EDX) fluorescence.6 Specimens for EDX were concentrated and separated
from the Acropor through an acetone extraction. A reference blank from an unused, clean
membrane filter was prepared using both methods so that X-ray intensity data could be
corrected for the presence of residual filter mat.
Standardization for the semiquantitative analysis was accomplished through a combination
of reagent grade chemicals, National Bureau of Standards (NBS) mixtures and independent
analyses of three filter specimens using atomic absorption techniques.
The raw data (X-ray intensities per element per specimen) were corrected for the presence of
filter mat in the specimen. These intensities were then corrected for interelement inter-
ferences using mathematical relationships determined experimentally during calibration. The
corrected intensities were converted to micrograms of element present and a weight percent
calculated from the actual specimen weight. All of these calculations were computerized.
Significant errors in the semiquantitative analysis arose from large variations in the specimen
weight, differences in specimen geometry, and large variations in the concentrations of the
major element. Errors due to interelement interferences were minimal and mathematically
corrected for most elements.
METEOROLOGICAL ANALYSIS
The meteorological analysis and summary for each sampling day were compiled by PSAPCA
using the following sources for data;
1. The National Weather Service (NWS)
2. Boeing Field (King County Airport) hourly surface weather observation
3. The Environmental Meteorological Support Unit (EMSU) upper air sounding taken at
Portage Bay, Seattle
13
-------
4. The NWS air stagnation forecasts
5. Wind data from the PSAPCA data acquisition system
Additional air quality data used was the soiling index (coefficient of haze, COH) measured
by PSAPCA at site K55.
These meteorological data were considered to be the averages for the entire Duwamish Basin
area and therefore were applied to all samples taken on the same day regardless of site
location.
COMPUTER ANALYSIS
Several existing computerized statistical models were investigated to determine the most
valid model, or set of models, capable of predicting site to site polution levels and elemental
composition of these levels as a function of meteorological conditions (prediction being the
first step toward determination of point sources of ambient particulate). The programs
selected for testing were BMDX64 (General Linear Hypothesis, July 27, 1965)7 and
BMD02R (Stepwise Regression, June 2, 1965).7 Both programs were modified slightly to
accommodate air particulate data since both were originally coded for biomedical analyses.
A third computer program was developed for data reduction; i.e., to convert the EDX raw
data (X-ray intensities) to a usable form; weight percent of element; actual weight of
element expected to be present in a normal 24-hour sample; or micrograms of element per
cubic meter of air sampled. All three forms of data were input to BMDX64 and BMD02R
for evaluation.
OPTICAL ANALYSIS
Optical analysis of the collected particulate using standard techniques for microscopic
examination was used for semiquantitative classification of both organic and inorganic types
of particulate. Two different methods of optical analysis were performed; a particulate size
distribution on the filter surface, and a detailed semiquantitative compound identification-
characterization of the particulate mounted on microscopic slides.
Particle size was determined using a standard filar eyepiece (10X) and Feret's diameter. The
number of particles in three size ranges (0.5^j to 3/u to 7/u; and larger than 7/u) were
determined. These size ranges were chosen to correlate with results obtained in an earlier
study, 1 which used an Anderson Particle Size Impactor. A single field from each membrane
filter was counted to determine the size distribution with a precision of ±20% of the
reported value as determined by duplicate counts.
Particulate Identification-Characterization
The particulate was separated from the membrane filter using acetone. The particulate was
suspended in reagent grade monobromobenzene and pipetted into a prepared density
gradient column. The density gradient was established through the use of pure liquids and
14
-------
mixtures of monobromobenzene, bromoform, and di-idomethane producing an eight-layer
density column. Each density gradient column was monitored by the addition of one crystal
each of potassium dichromate (specific gravity = 2.61) and ammonium dichromate (specific
gravity =2.15).
The density columns chosen for detailed optical examination (particulate characterization)
were drained by density level into small, clean watch glasses, labeled and evaporated to
dryness in a class 100 clean bench. The dry particulate was resuspended in chlorobenzene
(specific gravity = 1.1) to concentrate it in the center of the watch glass, and examined
under a stereoscopic microscope to evaluate the particulate distribution. A portion of the
particulate was removed with a disposable pipet and deposited on a clean microscope slide;
both slide and watch glass particulate were then reexamined to insure that a representative
sample had been taken. The representative specimen (on the slide) was mounted in Aroclor
5442 for examination.
Particle identification and characterization were performed using a Nikon Apophot
microscope with polarizing attachments and a Carl Zeiss standard microscope with episcopic
darkfield attachments. The semiquantitative analysis was based on the number and size (a
real analysis) of particles of known density conforming to given optical criteria. The total
particulate was divided into thirteen classes (see table 4 and fig, 3).
Electron Microprobe Analysis
Several individual particles were removed from slides for individual qualitative and
semiquantitative elemental analysis on electron microprobes. An Applied Research
Laboratories—EMX electron microprobe was used with wavelength dispersive X-ray analysis
to determine the elemental nature of these individual particles.
SCANNING ELECTRON MICROSCOPY ANALYSIS
Representative particulates were selected at random from several ambient air membrane
Filters for high magnification microscopy on an Ultra Scan scanning electron microscope.
This analysis was used to supplement the optical analysis in characterizing the submicron
particles which were too small for optical resolution.
X-RAY DIFFRACTION ANALYSIS
Approximately 30 ambient air samples and source samples were selected for X-ray
diffraction analysis. This analysis was to determine the nature of any inorganic compounds
found. Macropowder samples were run on a diffractometer while micropowder samples were
run on a Debye-Scherrer camera.
15
-------
Table 4. PARTICLE CLASS CRITERIA
Class
Class name
Criteria for classification8
Error, % of
reported value
1
Pollens and spores
Morphology; waif and surface structure; size
1
2
Starch
Predominantly oval or circular, transparent, smooth, colorless particle 10-20 /u
in diameter. Crossed polarized light causes it to appear white with a black cross
through of the grain.
1
3
Plant parts
Morphology; cell structure or organization; low density. Includes charred plant
materials from open burning, plant fibers, material from food processing except
starch, and naturally occurring materials.
10
4
Furnace slags
Glassy material of variable color, transparency, density, and index of refraction;
no birefringence (except from strain); perfect conchoidal fracture. Two distinct
morphologies: brown to opaque flakes of nearly uniform thickness; striated,
multicolored clear to opaque flakes of variable thickness.
1
5
Furnace spheres
Metal and magnetite spheres; thin-shelled glass bubbles. Occurs in every density
range with a wide variety of colors from transparent to opaque; 5 y to 20 ju
diameter; no birefringence. Black spheres have same optical characteristics as
soot so classified with respect to density and size.
20
6
Furnace ash
Irregular morphology; generally opaque; variable color; no birefringence. Surface
often appears to be composed of numerous small particles. (If specific gravity
2.2, then distinguished from charred plant parts and tire rubber by glassy
formations.)
10
aSee references 8 through 18 for a more detailed discussion of particulate classification criteria.
-------
Table 4 (concluded). PARTICLE CLASS CRITERIA
Class
Class name
Criteria for classification3
Error, % of
reported value
7
Fly ash
Large particles (1 0/4; spheroidal; restricted to transparent yellow to orange
"bulbs" with pockmarked surface.
1
8
Abrasive dusts
Restricted to positive identification of garnet abrasives and silicon carbide. (Many
types of industrial abrasives are difficult to distinguish from locally common
natural minerals.)
0
9
Concrete
minerals
Restricted to cement cinders (specific gravity 3.0) and hydrated calcium silicates,
ferrites, aluminates, sulfates, and oxides (specific gravity 2.4) includes gypsum and
calcite. Hydrated particles may resemble clay minerals.
20
10
Road dust
Birefringence, refractive index and density of common minerals (e.g., quartz, horn-
blende, olivine, feldspars, and common clay minerals).
10
11
Soot (oil)
Black spheres (specific gravity 2.2), size (1 ji to 25 jj).
30
12
Tire rubber
Morphology (black coned cylinders with poorly defined edges at 500X and no
predominant ordering).
10
13
Other
All items not fitting the criteria for classes 1 through 1 2 and not representing a
significant portion of the total specimen. Examples: diatoms, metal turnings,
moth scale, glass, and nylon fibers not morphologically attributable to the
sampling filters used.
aSee references 8 through 18 for a more detailed discussion of particulate classification criteria.
-------
wmfiM.
5. FURNACE SPHERES, 100X
3. PLANT PARTS, 120X
2, STARCH, 500X
CINDER HYDRATED
9. CONCRETE MINERALS, 500X
*
* #
s
J* • FLAKE
STRIATED
4. FURNACE SLAGS, 500X
» %
7. FLY ASH, 500X 6. FURNACE ASH + Fe304, 200X
Figure 3, Representative photographs of some particle classes (Numbers refer to particle class)
-------
SECTION VI
RESULTS AND DISCUSSION
SAMPLING
A total of 198 glass fiber and 202 membrane filters were collected.
GRAVIMETRIC ANALYSIS
All gravimetric data are tabulated by date, meteorological conditions and geographic
location in tables 5 through 7,
Figure 4 reveals the seasonal variation of the average particulate concentration, tJ-gjm . A
three-point smoothing function has been applied to the data from tables 5 through 7 to
reduce the scatter caused by the other variables (degree of vertical ventilation, wind
direction, and type of day). This figure clearly indicates a gradual increase of airborne
particulate beginning in early July and extending through late September. There is then a
gradual decrease in particulate loading until mid-October; the decrease becomes quite
pronounced with the onset of the autumn rainy season.
Figure 5 illustrates the average particulate concentration as a simultaneous function of wind
direction and geographic location (sites are listed north, K60, to south, K59). These curves
indicate that the highest particulate loading occurs in the north end of the Duwamish Basin.
This may be attributed to the high average density (specific gravity) of the collected
particulate (see Optical Analysis) and/or the location of major sources of particulate in the
vicinity of the northernmost site, K60.
The particulate loading for northerly winds is consistently higher than that for southerly
winds; this may be attributed to the degree of vertical ventilation and/or the average wind
speed. The southerly winds were more frequently associated with rain (implying good
vertical ventilation) than the northerly winds which were frequently associated with
inversion conditions. The southerly wind speeds were generally higher than the northerly
wind speeds, resulting in a much greater mixing depth, and more instability providing better
dispersion of particulate throughout the sampling network.
The individual cujves for predominant wind direction yield some information concerning
general locations of major sources of airborne particulate. The northerly winds indicate that
particulate sources exist north of K60 and between K61 and K64. The southerly winds
indicate major particulate source(s) between K61 and K99 and minor source(s) between
K55 and K60. The mixed winds generally fall between the particulate loading for the north
and south winds, displaying characteristics of both; there is no obvious explanation for the
high average particulate concentration (PC) at K99 and K64. These observations are
summarized in figure 6 which shows the general geographic locations of particulate sources.
Figure 7 illustrates the effect of the day of the week on the average particulate
19
-------
Table 5. METEOROLOGICAL AND GRAVIMETRIC DATA FOR GENERALLY NORTHERLY WIND DAYS
Date
Meterological data
Soiling3
index
COH
Gravimetric data „
Particulate concentration, ^g/m
Sites listed north to south
Daily
arith.
b
ave.
Wind
dir. and
speed,
knots
Inversion
K60
K55
K99
K61
K64
K59
7-13h
N at 11
Yesd
0.4
Abort
Abort
55.6
40.2
60.4
43.2
49.8
7-18
NW at 5
No
0.5
94.7
61.8
58.2
26.3
30.4
32.6
50.7
7-25
NWat 12
Yes
0.5
135.3
103.6
98.8
49.6
96.1
63.7
91.2
7-28h
N at 10
Yesd
0.2
49.7
58.0
52.0
52.9
78.7
58.6
58.3
7-31 m
N at 2
Yes
Abort
185.2
105.6
118.7
100.9
91.2
77.7
113.2
7-31 n
N at 8
Yes
Abort
222.2
106.4
127.7
117.9
239.2
105.6
128.0
7-31 e
N at 6
No
Abort
213.8
101.5
95.8
84.0
103.2
107.0
117.6
8-3h
NW at 8
Yesd
Abort
56.5
61.7
64.0
58.4
77.3
27.3
57.5
8-9
NWat 8
Yes
Abort
129.8
84.9
137.6
61.0
65,2C
67.5C
91.0
8-15
NW at 7
Yes
0.6
91.2
76.1
94.8
77.5
81.8e
82.0
83.9
8-21
NW at 5
Nod
0.7
130.3
92.4
65.0
61.7
97.7
68.9
86.0
8-27
N at 6
Yes
1.3
186.7
176.4
Abort
93.0
113.5
100.4
134.0
9-2h
NWat 10
Yesd
0.8
92.1
114.7
Abort
86.8
96.8
92.1
96.5
9-14h
NWat 6
Yesd
1.5
118.4
132.8
115.2
92.0
105.0
89.5
108.8
9-20
NW at 9
Yes
1.1
257.4
175.6
185.9
137.9
198.6
123.3
179.8
9-24
NW at 9
Yes
1,3
195.2
171.3
180.0
118.9
155.3
115.8
156.1
10-16
NW at 5
Yes
2.6
243.7
207.6
Abort
151.9
159.6
120.9
176.7
11-1
NWat 5
No
1.1
86.0
55.8
54.6
77.6
47.5
55.2
62.8
11-13
NW at 4
Yes
0.8
117.1
79.5
76.0
89.9
48.1
Abort
70.8
-------
Table 5 (concluded). METEOROLOGICAL AND GRAVIMETRIC DATA
FOR GENERALLY NORTHERLY Wl ND DAYS
Arithmetic averages
Gravimetric data
particulate concentration, fig/m"
Sites listed north to south
K60
K55
K99
K61
K64
K59
All days^
144.7
109.2
98.7
83.1
102.4
75.8
(number of data points)
(18)
(18)
(16)
(19)
(19)
(19)
Workdays (Monday through Friday)9
163.5
114.2
107.8
89.2
109.1
80.7
(number of data points)
(14)
(14)
(12)
(14)
(14)
(14)
Nonworkdays (Sat., Sun., and holidays)9
79.2
91.8
71.7
66.1
83.6
62.1
(number of data points)
(4)
(4)
(4)
(4)
(4)
(4)
Difference: workdays minus nonworkdays
84.3
22.4
36.1
23.1
25.5
18.6
% difference (workdays)
51%
20%
33%
26%
23%
23%
Note: 7-31 is one day consisting of three 8-hour sampling periods: Morning (m), noon (n), and evening (e).
Measured by tape sampler at site K55
^Plotted in figure 4; aborts ignored
cUnintentional sample period of 37 hours
^No soundings were taken
e Unintentional sample period of 48 hours
f
Plotted in figures 15-17
^ Plotted in figure 8
Nonworkday
-------
Table 6. METEOROLOGICAL AND GRAVIMETRIC DATA FOR GENERALLY SOUTHERLY WIND DAYS
Meteorological data
Wind
dir. and
speed,
Inversion
Soiling
index
COH
ui dvimeu iu udio
Particulate concentration, [iqim
Sites listed north to south
Daily
arith,
ave.
Date
knots
K60
K55
K99
K61
K64
K59
7-4
SW at 7
Noc
0.2
27.7
18.8
Abort
18.5
Abort
24.9
22.5
7-16
SE at 7
Yes
0.9
89.6
65.6
72.4
33.6
32,2
Abort
58.7
7-22
SE, SW at 10
Yes
0.4
67.8
49.2
100.7
27.2
24.4
26.9
49.4
9-8
SE, SW at 14
;z
o
CL
0.3
63.1
68.4
65.0
Abort
Abort
Abort
65.5
10-2
S at 10
No
0.5
68.7
71.8
115.6
37.0
54.4
38.1
64.3
10-10
SW, SE at 8
No
0.9
81.9
70.0
114.6
49.5
47.8
49.8
68.9
10-26
SE at 4
Yesc
1.7
124.7
93,8
Abort
108.8
76.0
69.8
94.6
11-7
SE at 5
No
1.1
44.4
43.4
36.4
26.5
19.2
26.5
32.7
11-19
SE at 10
No
0.6
34.6
36.3
23.8
18.1
18.6
26.0
26.2
11-25
SW at 7
No
0.6
79.1
47,1
77.0
40.7
32.3
41.4
52.9
j
Arithmetic average
68.2
56.4
75.7
40.0
38.1
37.9
(Number of readings)
(10)
(10)
(8)
(9)
(8)
(8)
aMeasured by tape sampler at site K55.
Plotted in figure 4, aborts ignored.
cINlo soundings were taken.
^Plotted in figure 5.
-------
Table 7. METEOROLOGICAL AND GRAVIMETRIC DATA FOR DAYS HAVING VARIABLE WIND CONDITIONS
Soiling3
3
Particulate concentration, jjg/m ; sites listed north to south
index
Daily
Date
1
COH
Inversion
K60
K55
K99
K61
K64
K59
arith. ave.
7-10
0.3
No
50.3
31.3
43.0
17,7
Abort
20.0
32.5
9-26
0.5
No
138.2
82.0
163.4
65.4
83.7
62.0
99.0
10-8
1.8
Yes
180.1
140.2
Abort
130.3
234.9
109.5
159.0
10-14
1.6
Yesc
Abort
118.8
158.4
107.1
144.7
87.2
123.2
10-20
0.5
Noc
33.7
40.0
Abort
28.1
21.7
27.5
25.2
*3
Arithmetic average, p,q!m
d
100.6
82.5
121.6
69.7
121.2
61.2
{Number of readings)
(4)
(5)
(3)
(5)
(4)
(5)
a Measured by tape sampler at site K55
^ Platted in figure 4, aborts ignored,
c No soundings were taken.
d Plotted in figure 5,
-------
C*}
as
a.
160
140-
120
to
¦P*
O
1-
<
tr
H
2
UJ
o
2
O
O
LU
t—
<
_1
3
O
i—
ac
<
o.
100*
80
60
40
20 , , 1 f—
9 15 21 27
JULY
—, , 1—
14 20 26
AUGUST
-T 1 1 1—
7 13 19 25
SEPTEMBER
—I 1 1 r—
7 13 19 23
OCTOBER
_ymm
9
1 ¦ ""T"'-
15 21 27
NOVEMBER
Figure 4. Particulate concentration, daily arithmetic average with three-point smoothing function applied
-------
~ NORTHERLY WINDS
x SOUTHERLY WINDS
® VARIABLE WINDS
Figure 5. Particulate concentration, arithmetic averages by site and wind direction
25
-------
Interstate 90 \£:«*£*...... . .,
ijvi;.;Mjiiji.iw>f,
¥:£•£
lIili|PV%
»»7 %
MERCER
ISLAND
:;^ WEST SEATTLE
giiililllll
RAINIER
VALLEY
O MAJOR SOURCE
0Pillli|O Jj MINOR SOURCE^
BEACON HILL
0
IliSIt
**K61
BURIEN
tukwila:
SEATTLE-TACOMA
INTERNATIONAL
AIRPORT t///ft
Figure 6. Geographic locations of particulate sources
26
-------
150
100
50
30
25
20
15
-L
±
_L
X
X
K60
-T-
K55
T"
K99
T~
K61
K64
"T"
K59
¦ AVERAGE OF WORKDAYS
• AVERAGE OF IMONWORKDAYS
Figure 7. Concentration of particulates and ashable organics, arithmetic averages
by site and day of week (northerly winds only)
27
-------
concentration for northerly winds. Except for K60 the differences between the average PC
on workdays and nonworkdays are fairly constant, suggesting that the differences could be
attributed to nonnatural sources, i.e.. industry and/or transportation. Some suppositions
that may be gleaned from close scrutiny of the PC curves in this figure are: (1) there may be
industrial sources north of K60 that emit particulate only five days a week (also supported
by optical analysis data from density separations); (2) the area between K55 and K99 may
contain a minor source of particulates; (3) there may be sources between K61 and K64 that
emit particulate seven days a week.
Figure 7 also illustrates the effect of the day of the week on the average concentration of
ashable organics. The average workday concentration is higher than the average nonworkday
concentration for ashable organics and for particulate loading.
ORGANIC ASHING ANALYSIS
Results cover all sampling days through October 10, 1974, plus twelve source test filters.
Ambient air sample results are tabulated by weight percent in table 8, while source test filter
results are listed in table 9. The relationship between the particulate concentrations
observed on workdays and nonworkdays (for generally northerly wind conditions)
prompted a similar interpretation for the concentration of ashable organics in ambient air
test filters (see table 10 and fig. 7). The similarities between the two sets of curves indicate
that ashable organics play a large role in the level of airborne particulate. In addition, the
low averages at K61 may be attributed to its geographic location away from areas of large
traffic volume; and the positive slope of the workday curve between K55 and K99 may
indicate that transportation is a major contributor to the ashable particulate.
DETAILED ORGANIC ANALYSIS
Soxhlet Extraction
Seven ambient air test filters (six sites sampled on July 25, 1974, and site K99 sampled on
August 21, 1974) were extracted using the Soxhlet method with the solvent chloroform;
K99 (8-21-74) was subsequently Soxhlet extracted with the solvent xylene. These results are
summarized in table 11 where it can be seen that less than half of the total ashable organics
(as determined by LTA) are soluble in chloroform while the remainder are soluble in xylene.
Two source test filters were Soxhlet extracted using chloroform. These were the gasoline
engine exhaust (yielding 3.3 mg of organic material) and the diesel oil boiler (yielding
0.3 mg of organics, which is less than the yield of the reference blank, 0.9 mg; less than the
accuracy, ±0.4 mg, of the analytical balance used; and less than the minimum amount of
material required for any of the subsequent analyses). The gasoline engine exhaust was not
analyzed gravimetrically so no reliable LTA data are available; therefore the amount of
chloroform extractable orpnics in this source test cannot be related to the amount of total
ashable organics present. The amount of chloroform extractable organics (1 % by weight) in
the diesel oil boiler is much less than LTA data (70% by weight) indicate. This discrepancy
28
-------
Table 8. TOTAL ASHABLE ORGANICS; AMBIENT AIR TEST FILTERS
(percent by weight)
Date
K60
K55
K99
K61
K64
K59
7-4
26.3
33.1
Abort
30.4
Abort
34,2
7-10
21.4
26.7
27.0
27.4
Abort
32.7
7-13
Abort
Abort
25.5
27.7
27.5
32.8
7-16
24.6
29,9
31.3
35.0
40.1
Abort
7-18
15.0
21.5
28.0
31.3
32.7
36.7
7-22
18.0
21.1
25.4
28.2
30.9
32,2
7-25
10.8
12.8
20.9
16.9
17.8
21.2
7-28
28.4
21.9
25.2
29.3
19.1
27.4
7-31m
19.6
28.0
29.4
26.6
25.5
31,0
7-31 n
17.4
29.3
28.7
26.0
18.0
33.2
7-31e
14.2
29.6
25.3
15.7
17,2
20.1
8-3
21.5
22.9
16.0
23.2
20.2
27.2
8-9
16.6
24.8
20.9
19.1
22.7
27,0
8-15
19.1
27.9
22.3
23,4
19.8
24.1
8-21
13.3
22.3
24,2
27.6
23.6
29.8
8-27
21.4
22.4
22.3
25.2
23.4
27.6
9-2
25.4
21.7
Abort
19.6
21.4
24.1
9-8
26.8
20.1
26.0
5.8
14.9
Abort
9-14
26.5
28.0
32.3
28.4
27.2
30.3
9-20
16.0
19,1
23.5
23.3
18.1
26.5
9-24
18.3
19.6
24.8
26.8
22.8
29.2
9-26
18.9
19.8
19.4
23.8
19.5
27.9
10-2
20.0
19.5
No data
24.4
17.0
25.6
10-8
26.5
35.2
Abort
34.2
30.4
37.8
10-10
28.8
29.6
No data
32.5
34.3
35.8
29
-------
Table 9. TOTAL ASHABLE ORGANICS: SOURCE TEST FILTERS
Total
Total
ashable
suspended
organics,
Source
particulate, gm
wt %
Gypsum ship unloading
2,4788
7
Gypsum ship unloading
2.0291
7
Diesel switch engine exhaust
0.2080
2
Harbor Island road dust
3.9781
2
Gasoline engine exhaust
(0.2421 )a
<9}a
Steel plant hi-temp baghouse
0.0525
43
Steel plant low-temp baghouse
0.0120
58
Diesel oil boiler
0.0292
70
Flour Mill five-mill cyclone
0.1732
72
Flour mill baghouse
0.0712
66
Cement plant stack
0.1665
11
Road dust
1.87
7
aNo gravimetric analysis was performed on this filter so total suspended particulate
and total ashable organics are estimates.
30
-------
Table 10. CONCENTRATION OF ASHABLE ORGANICS3—NORTHERLY WINDS
(micrograms/cubic meter)
Date
K60
K55
K99
K61
K64
K59
Workdays
7-18
14,2
13,6
18.3
8.2
9.9
12.0
7-25
14,6
13.3
20.6
8.4
17.1
13,6
7-31 m
36.3
29,6
34.9
26.8
23.2
24.0
7-31n
38,6
31.2
36.6
30.6
43,0
35.1
7-31e
30,4
30,0
24.2
13.1
17.8
21.5
8-9
21.5
21,0
28.8
11.6
14.8
18.2
8-15
17.4
21.2
21.1
18.1
16.0
23.0
8-21
17.3
20.6
15.7
17,0
23.0
20.5
8-27
39.9
39.5
8.8
23,4
26,6
27.7
9-20
41,2
33.5
43.7
32.1
35.9
32.7
9-24
35.8
33,6
44,6
31.9
35.4
33.8
Nonworkdays
713
Abort
Abort
14.2
11.1
16,6
14.2
7-28
14.1
12.7
13.1
15,5
15,0
16.0
8-3
12.1
14,1
10.2
13,5
15.6
13,6
9-2
23.4
24.9
Abort
17,0
20,8
22.2
9-14
31,4
37.1
37.2
26,1
28,6
27,1
Arithmetic
averages
Workdays'3
27.9
26.1
26.9
20.1
23.9
23.8
Nonworkdays'3
20.3
22.2
18.7
16.7
19.3
18,6
Alt days
25.9
25,1
24.7
19.0
22,5
22.2
Concentration of ashable organics = (Particulate concentration x % ashable organics)/100
L_
Plotted in figure 7
31
-------
Table 11. ORGANIC CONTENT BY SOXHLET EXTRACTION
Site
Date
sampled
TSP,a
mg
TAO,13
wt %
Chloroform extractable organics
gm
% TSPa
% TAOb
K60
7-25-74
133
11
4.2
3.2
29
K55
7-25-74
97
13
2.5
2.6
20
K99
7-25-74
103
21
4.6
4.5
21
K99
8-21-74
64
24
9.7
10.7
45
K99
8-21-74
64
24
5.5C
15.3C
55c
K61
7-25-74
50
17
2.7
5.4
32
K64
7-25-74
108
18
6.6
6.1
34
K59
7-25-74
66
21
4.4
6.7
32
aTSP = 0,75 x total suspended particulate as determined from gravimetric analysis
TAO = Total ashable organics as determined from low-temperature ashing
cXylene extraction which followed a chloroform extraction of the same filter sample.
Note: Chloroform extraction followed by xylene extraction of the same filter sample
accounts for 100% of TAO as determined by low temperature ashing.
32
-------
may be explained by the presence of a large amount of xylene soluble organics or by the
presence of a large amount of elemental carbon which is not soluble in chloroform but is
oxidizable (LTA).
Infrared Spectroscopy
All of the chloroform extracts produce infrared spectrograms indicating varying amounts of
a few types of molecular bonds which are predominantly associated with aliphatic
hydrocarbons and aliphatic carbonyl-containing compounds (esters and acids). The xylene
extract (K99, 8-21-74) contains some aliphatic hydrocarbons and esters but is primarily
composed of aromatic hydrocarbons (substituted and/or polynuclear). See figures 8 and 9
for absorption band assignments. The chloroform-extracted reference blank produced an
absorption band associated with silicones; this band is not called out in the spectrograms as
it is felt that contributions to it can be attributed entirely to the glass filter.
Thin Layer Chromatography (TLC)
Figure 10 shows that TLC is a more sensitive method for characterizing trace quantities of
organic materials than infrared spectroscopy. Close scrutiny of the TLC overlays of the
chloroform extracts reveals (1) that all of the ambient air test samples contain many of the
same types of organic materials as the source test sample of the gasoline engine exhaust;
(2) that the TLC patterns of K99 (8-21-74), K99 (7-25-74) and K59 (7-25-74) are identical;
and (3) that the TLC patterns from K64, K61 and K55 (all from 7-25-74) are also identical.
These results indicate that the same organic materials are present from day to day (from (2)
above); and site to site (from (2) and (3) above) on the same day. Therefore, a
representative filter (such as K59, 7-25-74) was chosen for intensive study by mass
spectroscopy.
The xylene extract of K99 (8-21-74) was chromatographed using various solvents and
combinations of solvents. Figure 11 illustrates these results and the various degrees of
separation obtainable.
Mass Spectroscopy
Mass spectra were obtained from all of the Soxhlet extracts listed in table 11, plus the
chloroform extract of the source test of the gasoline engine exhaust (figs. 12 and 13). In
addition, mass spectra were obtained from TLC separations of the chloroform extracts from
K59 (7-25-74) and the gasoline engine exhaust (figs. 14 through 17 and table 12).
The outstanding feature in all of the low-resolution mass spectra of the chloroform extracts
is the typical fragmentation pattern of high molecular weight aliphatic hydrocarbons. These
are the peaks between 40 and about 150 mass units corresponding to ion formulas
CnH2n + 1 and CnH2n " 1 with smaller peaks at CnH2n- The peak intensities become
progressively smaller with increasing mass for this series. The parent peaks of the many
possible hydrocarbons that contribute to these peaks are either very small or do not appear
at all. The similarities between the chloroform extracts of the gasoline engine exhaust and
33
-------
00
90
80
70
60
50
40
30
20
00
90
80
70
60
50
40
30
20
00
90
80
70
60
50
40
30
20
K99 (7-25-74)
CH
CH.
FINGERPRINT
OF ESTER TYPE
COMPOUND
ESTER
CH
CH
K60 (7-25-74)
2000
500
3000
VIBRATION FREQUENCY, cm"1
Figure 8. Infrared spectrograms of chloroform extracts
34
-------
100
90
60
50
40- AROMATIC CH
ESTER TYPE
ABSORPTIONS
30 -
CH
CH
20
CH
CH
10 -
AROMATIC
CH
AROMATIC
500
2000
3000
4000
VIBRATION FREQUENCY, cm""1
Figure 9. Infrared spectrogram of xylene extract from l<99 (8-21-74)
-------
SOLVENT SYSTEM: DIETHYLETHER/n-HEPTANE
K99
8-21-74
ORIGIN
K99
7-25-74
K59
7-25-74
K64
7-25-74
K61
7-25-74
K55
7-25-74
K60
o?<
7-25-74 iJOl
< Z X
(J LU LU
SOLVENT FRONT
VISIBLE ONLY AFTER SPRAYING
O O o
cd CD CD CD
Q CD CD CD
O CD CD
CD CD CD CD
CD CD
CD
\
GINA
ZONES VIEWED WITH ULTRAVIOLET LIGHT (365 NANOMETERS)
Figure 10. TLC overlay of chloroform extracts
36
-------
SOLVENT SYSTEM
ISOPROPYL ALCOHOL
CHLOROFORM/BENZENE CHLOROFORM BENZENE ACETONE
(2/1) (3/1)
FRONT
I ^
ORIGIN
FRONT
FRONT
FRONT
ZONES VIEWED WITH ULTRAVIOLET LIGHT (365 NANOMETERS)
Figure 11. TLC overlays of xylene extract from K99 (8-21-74)
37
-------
100
80
60
40
20
0
100
80
60
40
20
0
JL
XYLENE
JU L_
50
100
150
200
250
MASS UNITS
300
350
400
¦0.
tl llltllllllllllll
CHLOROFORM
iilwM llfal Hilt lllil ill
300
"1—
350
50
100
150
200
250
MASS UNITS
400
Figure 12. Normalized bar graphs of mass spectra obtained from extracts of K99 (8-21 74)
-------
K64 (7-25-74)
oSA-
—I
400
50
MASS UNITS
100
GASOLINE ENGINE EXHAUST
50
MASS UNITS
400
Figure 13. Normalized bar graphs of mass spectra obtained from total chloroform extracts
-------
100
>
H
w
z
LU
LU
>
<
-J
LU
OC
TOTAL EXTRACT
50
MASS UNITS
1
400
[X.
o
100
co
z
LU
<
DC
qL^.
ZONE R1
50
MASS UNITS
400
Figure 14, Normalized bar graphs of mass spectra obtained from diethylether/n-heptane
TLC separations of chloroform extract of K59 (7-25-74)—total extract and
zone R1
-------
lOOr
o^
50
100r
ZONE R2
1ASS UNITS
ZONE R3
I
300
OSH
50
MASS UNITS
300
Figure 15, Normalized bar graphs of mass spectra obtained from diethylether/n-heptane
TLC separations of chloroform extract of K59 (7-25 74)-zones R2 and R3
-------
100
>
h;
CO
2
LU
I-
2
LU
>
I-
<
LU
QC
Ohh
ZONE R4
—,
400
50
K)
100
>
H
ZONE R5
>
LU
cc
gUfJ
ULli
llliti ll 1*1 Mm nnl- lint u_
50
MASS UNITS
400
Figure 16. Normalized bar graphs of mass spectra obtained from diethylether/n-heptane
TLC separations of chloroform extract of K59 (7-25-74)—zones R4 and R5
-------
Ui
100>
>
t
C/3
LU
>
h-
<
_J
LU
cc
oKa-
100
50
>
t
to
LU
>
LU
DC
Q
50
ZONE R1
MASS UNITS
ZONE R2
150
MASS UNITS
150
100
w
LU
>
H
<
_l
LU
DC
0 *-\A-
100
ZONE R3
50
>
t
to
H
<
OhA"
1
MASS UNITS
150
ZONE R4
50
MASS UNITS
150
Figure 17. Normalized bar graphs of mass spectra obtained from diethylether/n-heptane
TLC separations of chloroform extract of gasoline engine exhaust
-------
Table 12. PRECISE MASS VALUES AND ELEMENT LIST FOR TLC ZONES OF
K59 (7-25-74)
TLC,a
M
meas'
Empirical
Mcalc'
3 ^
M x 10J,
zone
mass units
formula
mass units
mass units
R1
129.0892
c3H9N6
129.0889
0.3
C5HhN30
129.0902
1.0
C7H13°2
129.0915
2.3
149,0248
C8H5°3
149.0239
0.9
167.0345
C8H7°4
167.0344
0.1
R2
No peaks intense enough for measurement
R3
276.0970
C8H14N5°6
276.0944
2.6
C13H14N3°4
276.0984
1.4
C11H12N6°3
276.0971
0.1
C15H16°5
276.0998
2.8
C22H12
276.0939
3.1
R4
252.0976
C9H12N6°3
252.0971
0.5
C11H14N3O4
252.0984
0.8
C13H16°5
252.0998
2.2
C20H12
252.0939
3.7
253.0963
C10H13N4°4
253.0937
2,6
C12H15N5°
253.0950
1.3
C13H11N5°
253.0964
0.1
C15H13N2°2
253.0977
1.4
C19C13H12*
253.0973
1.0
aZones are numbered R1 to R5 with R1 nearest to the origin and R5 taken from the solvent front
U
Absolute value of the difference between Mmegs and Mcg|c
*Ci3 Isotope of carbon
44
-------
Table 12 (concluded), PRECISE MASS VALUES AND ELEMENT LIST FOR TLC ZONES
OF K59 (7-25-74)
TLC,a
^meas'
Empirical
M , ,
calc'
3 k
M x 10,
zone
mass units
formula
mass units
mass units
R5
149.0248
C8H5°3
149.0239
0.9
149.1349
C11H17
149.1330
1.9
255.2126
C19H27
255.2113
1.3
255.9928
_
-
_
256.2172
C14H28N2°2
256.2151
2.1
C19H28
256.2191
1.9
279.1640
C10H23N4°5
279.1668
2.8
C17H19N4
279.1610
3.0
C19H21NO
279.1623
1.7
386.3897
C23H50N2°2
386.3872
2.5
C28H50
386.3912
1.5
400.4072
C24H52N2°2
400.4029
4.3
C29H52
400.4069
0,3
aZones are numbered R1 to R5 with R1 nearest to the origin and R5 taken from the solvent front
^ Absolute value of the difference between and M.„
rncdb Caiu
45
-------
the ambient air samples indicate that the major contributor to the chloroform extract is
gasoline engine exhaust.
The low resolution mass spectra of the five zones obtained by TLC separation of K59
(7-25-74), chloroform extract, were dominated by peaks typical of hydrocarbons. These
peaks obscure features of spectra of other organic compounds that may have been
chromatographically separated. Thus positive identification of additional compounds in
ambient air samples was not furthered by TLC separation due to the small quantity of
material available for analysis.
The TLC separation spectra from the gasoline engine exhaust were very weak and show the
peaks expected from hydrocarbons, with the exception of peaks at 149 and 167.
Some features of mass spectra are more readily interpreted when combined with precise
mass measurements of the peaks, using the peak matching technique. If the mass of an ion
can be determined with a precision of ±2 millimass units or better, the combination of
elements in its empirical formula can be reduced to a few possibilities or sometimes a single
possibility. Tables^ consulted for actual compound identification were restricted to the
four most common elements found in organic compounds: C, H, N, and O. Furthermore,
the possible formulae were limited to a minimum of one carbon and a maximum of six each
of nitrogen and oxygen.
A peak at 149 appeared in every chloroform extract, but with varying intensity. The 149
peak is usually the largest peak in phthalic acid esters. Precise mass measurements confirm
that this ion is CgF^C^"1-, and is very likely a fragment of alkyl phthalates or phthalic acid.
A 167 peak, CgL^C^4", is sometimes observed along with 149 in phthalates and was found
in a few of the extracts. This group of esters is widely used as a plasticizer in many common
plastic materials such as vinyl chloride, etc.
The gasoline engine exhaust chloroform extract spectrum exhibits peaks at 276, 300 and
326; precise mass measurements indicate that these are probably C'22H 12; C24H12, and
C'26H 14, respectively. (See fig. 18 for possible isomers corresponding to these formulae.)
These all correspond to well known members of the polyaromatic hydrocarbons, and have
been detected previously in engine exhausts and air pollution studies. 19 The chloroform
extracts of K59, K60, K61, and K99 (all from 7-25-74) had peaks at mass 284. The precise
mass values gave CigH^gO-)4" as one possible formula. Stearic acid is a very likely possibility
whose spectrum has a prominent peak at this mass. Stearic acid and its esters are used in
rubber compounding and could be extracted from rubber particles on the filter or deposited
directly from the air.
The mass spectra of the chloroform and xylene extracts of K99 (8-21-74) indicate the
presence of aliphatic hydrocarbons (alkanes and alkenes) in the chloroform extract and
highly unsaturated polyaromatic hydrocarbon compounds in the xylene extract. This was
verified by precise mass measurements and infrared spectroscopy.
46
-------
MASS PEAK POSSIBLE ISOMER AND CHEMICAL FORMULA
276
300
326
C22H12
C24H12 (CORONENE)
C26H14
Figure 18, Possible organic compounds found in the chloroform extract
of gasoline engine exhaust
47
-------
ELEMENTAL ANALYSIS
The results of the ambient air test specimens that were analyzed For elemental composition
by energy dispersive X-rays show large variations (±200% of the amount present) for
individual elements and specimens. However, the effect of these variations appears to be
minimized when treated statistically.
The EDX results of the source test specimens are listed in table 13. Several elements were
found in the sources with many elements common from source to source. The road dust
samples appear to be contaminated by local sources.
OPTICAL ANALYSIS
Particle Size Distribution
Forty-seven membrane filter samples from ambient air testing were examined to determine
their particle size distribution. At least six samples from each site were included with all sites
represented for August 3, 1974 and August 21, 1974. Table 14 lists the (arithmetic) average
particle size distribution; the sites are arranged from north (K60) to south (K59). This table
indicates a homogeneous particle size distribution in the Duwamish Basin with a slight trend
toward decreasing size from north to south.
Density Separation
Density gradient columns were prepared for all filters. The mean densities of collected
particulate are illustrated in figure 19 for 22 sampling intervals. The general trend is for
mean density of particulate to decrease from north to south with K60 having the heaviest
particles. The mean density of K60 is lower for weekends and holidays than for weekdays;
this implies that a source of heavy particulate is located north of K60 (meteorological data)
and emits pollution at a higher rate on workdays than on nonworkdays. The area north of
K64 appears to contain another source for heavy particulate if the average wind conditions
of the Basin are indeed applicable to K64.
Particulate Identification—Characterization
The results of the semiquantitative optical analysis of eleven ambient air test samples are
listed in table 15 according to the restrictions and particle class criteria described in
table 4. The five samples taken on 10-16-74 were examined to determine the
effect of an air stagnation alert. The other samples were chosen on the basis of density
separation data as being representative collections of particulate for each site. The last three
rows of table 15 represent probable particulate origins; some of the particle classes cannot
be strictly attributed to one type of origin; in these cases the values are distributed among
the probable origins as described below.
Particles having a probable natural origin include those attributable to wind erosion of
exposed surfaces, open burning of biological materials, pollination, etc.
48
-------
Table 13. ANALYSIS OF SOURCE TESTS
Source test
Elements detected by EDXa
Major
Minor
Trace
Gypsum—ship conveyor belt
S, Ca, Sr
K
Al, Fe, Mn, Cu, Br
Harbor Island road dust
Si, Fe, Ca
Al, K, Pb, S
Ti, CI, Mn, Zn, Sr, Cu, Zr, As, Ni, V
Gasolinl? engine exhaust
Pb, Fe
Br, S
CI, Ca, As, Zn, Al, K, V, Ti
Diesel switch engine exhaust
S, CI, Si, Ca
Fe, Al, K
Zn, Cu, Ti, Pb, Mn, Ni, Sr, Zr, V
Flour Mill
5-mill cyclone
S, CI
K, Ca, Fe
Zn, Cu, Ni, Zr, Ti, V, As, Nb
Baghouse
S, CI
Ca, Fe, K
Zn, Sr, Pb, Al, V, As, Zr, Ti
Steel plant
High-temperature baghouse
Fe, Zn
S, Pb
Mn, Ca, K, Cu, CI, Ni, Br, V, Ti, Nb
Diesel truck exhaust
CI, S
Ca, Fe, Zn
Cu, Mn, K, V, Zr, Ti
Diesel oil boiler
s
Fe
Zn, Sr, Zr, V, Ti, K, Pb
Road dust {Port of Seattle)
Fe, Si, Ca
Al, K
S, Ti, Mn, Sr, Zn, Pb, Cu, As, Zr, V, Mo
Battery processor fugitive emissions
Si, CI, Sb
S, Pb
Al, Fe, As, Ti, Cu, Zn, Sr, K, Mo
Cement plant fugitive emissions
Ca, Fe
Al, Si, S,
K, Ti, Mn, As
Cu, Zn, Pb, Sr
aE!ements analyzed by EDX: Mg, Al, Si, P, S, CI, K, Ca, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn, As, Br, Sr, Zn, Nb, Mo and Pb.
-------
Table 14. PARTICLE SIZE DISTRIBUTION BY SITE
(average % by particle count)
0.5-3
3-7
Over 7
Site
microns
microns
microns
K-60
66
27
7
K-55
69
24
7
K-99
68
23
9
K-61
72
21
8
K-64
71
21
8
K-59
75
19
6
50
-------
Xl
V ^
-X
7-4, THUR.* S 7-10, WED. M 7-22, MON. S
-*sV^
7-28, SUN. N 7-31m, WED. N 7-31n, WED. N
7-31e, WED. N 8-3, SAT. N 8-15, THUR. N
8-21, WED. N 8-27, TUE. N 9-2, MON,# N
9-8, SUN. S 9-14, SAT. N 9-28, FRI. M
9-24, TUE. N 9-26, THUR. M 10-2, WED. S
Figure 19, Mean densities of samples by site
51
-------
10-8, TUE. M
10-10, THUR. S
V-
10-14, MON.* M
10-16, WED. N
LEGEND
2.61
>
LL
1.50
K59
K55
K61
K99
K60
K64
m = morning 1 b
n = noon
e = evening
* = HOLIDAY
N = GENERALLY NORTHERLY WINDS
S = GENERALLY SOUTHERLY WINDS
M = MIXED NORTHERLY AND SOUTHERLY WINDS
= NO DATA
Figure 19 (concluded). Mean densities of samples by site
52
-------
Table 15. SEMIQUANTITATIVE PARTICULATE CLASSIFICATION BY SITE
{% by weight)
Site and date sampled
Particle class
K-60
K-55
K-99
K-61
K-64
K-59
(7-13)
(10-16)
(8-15!
(10-16)
(8-27)
(7-10)
(10-16)
(8-3)
(10-16)
(7 31)
(10-16)
1 Pollens and spores
5
10
13
10
12
20
10
20
8
13
10
2 Starch
2
2
10
1
5
1
4
1
2
2
1
3 Plant parts
1
4
5
4
7
2
5
5
5
7
10
4 Furnace slags
10
7
3
1
1
1
1
2
1
2
2
5 Furnace spheres
15
20
16
15
5
2
12
3
18
5
8
6 Furnace ash
8
8
5
5
5
3
8
2
8
4
5
7 Fly ash
8
5
2
2
5
3
2
5
2
5
2
8 Abrasive dusts
1
1
1
1
1
1
1
1
1
1
1
9 Concrete minerals
3
5
2
8
2
1
8
6
10
1
8
10 Road dust minerals
36
20
30
30
35
45
40
40
32
35
35
11 Soot
5
15
6
20
12
13
15
8
10
12
12
12 Tire rubber
1
1
2
2
1
4
2
3
1
5
2
13 Other
5
2
5
2
10
5
2
5
3
8
5
Probable origin3
Natural
10
22
26
24
26
42
21
45
19
31
30
Transportation
45
30
27
44
41
45
39
36
37
44
41
Industry
45
47
47
11
33
11
39
18
44
25
28
aThe sum of these values is 98-100% because classes representing 1 % were not distributed. See text for method of determining these values.
-------
% natural origin = (% pollen and spores) + (% wind attributed
road dust) + (% plant parts - % starch)
Because the aerolian effects that carry the pollen also lift and carry fine exposed minerals,
the percentage of road dust minerals attributed to wind erosion is arbitrarily set equal to the
pollen percentage. This is a simplification, but if the road minerals were excluded or added
in their totality, a much more biased value would result. A similar logic applies to class 3
(plant parts). Many of the plant parts found were wheat related and should therefore be
included with the industrial group of particles. If the "% plant parts minus % starch" value
was less than 1%, it was assigned the value "0."
Particles having a probable transportation-related origin include those emitted and entrained
by moving vehicles.
% transportation origin = (% road dust - % pollen) + (% fly ash)
+ (% soot) + (% tire rubber)
The reduced road mineral value represents that which as a minimum could be attributed to
transportation-related dust. Fly ash is included in this group because of its restricted
definition. (Normally the term fly ash would include the three classifications of fly ash,
furnace ash, and furnace spheres.) The distinction in this study was made on the basis of
standards available for the Duwamish Basin. Fly ash is restricted to bulbous, yellow-orange,
large particulate; this type of material was not found in any of the industrial furnace source
samples, but was seen in some of the oil boiler and diesel engine samples. More source
samples must be evaluated before attributing this material to a specific source, but whatever
the source, the size of the particulate implies a lack of filtering following the materials'
production; therefore the most probable cause of this material is shipping and transporta-
tion. The contribution from stationary boilers was not considered because most stationary
boilers in the Duwamish Basin burn natural gas during the summer.
Soot is restricted in this study to refer to oil soot. The source of oil soot is, by the same
argument, predominantly the result of transportation-related activities. Some interference
would be expected with furnace spheres due to the similar optical and density
characteristics of some of the furnace spheres. Tire rubber has optical characteristics similar
to some of the furnace ash, but the extent of misrepresentation is probably small because of
the accessory formations present on the more brittle black furnace ash.
Particles having a probable industrial origin include those attributable to stack emissions,
construction, etc,
% industrial products = (% furnace slag) + (% abrasive dusts)
+ (% furnace ash) + (% furnace spheres)
+ (% starch) + (% plant parts, corrected)
+ (% concrete minerals) + (% other)
The % plant parts was included at the starch percentage value because of the relationship
between chaf, wheat tissue, and wheat trichomes that were usually found in association with
54
-------
the starch. The classification of "other" as used in this study includes a wide range of
particle types, the majority of which are of an industrial nature.
Furnace slags were found in this study to be of two different types that are characteristic of
different types of furnace operations. The first (platy and brownish) is typical of rotary
kilns during normal operation; fines liquified by a brief exposure to the flame impact on the
kiln walls, cool and flake off into the exhaust stream. The second type of slag (striated and
colorful) is typical of furnace operations where slag is tapped off the furnace into pots to
cool; striation occurs during cooling. The color variation is the result of different metallic
ions and compositions of the slags. This second type of slag is often used to cover parking
lots, as aggregates in asphalt and concrete, or as land fill. Its presence in the air sample
indicates that it is airborne as construction material, erosion product, or wind-blown refuse
from the slag dump yard of an industry using this type of furnace, and not as a stack
emission from a furnace. It should be mentioned that this type of slag is often used by the
cement industry as a source of iron in their production of cement and may become airborne
during that operation.
Concrete minerals cover a wide variety of particle types. Cement cinders are the easiest
identified. These cinders can give an indication of the types of cement being produced and
idiosyncrasies of the individual plant operations. Of the eleven air samples optically
evaluated, only two did not contain cement cinders in a detectable concentration, K60
(7-13-74) and K59 (7-31-74). The majority of particles counted in this classification were of
the hydrated cement type. Standards for comparison of this very complex group were
collected from actual Duwamish Basin cement industries and from hydrated samples of
other commercially available cements. These hydrated compounds may result from cement
plant emissions contracting water vapor in the atmosphere, the construction or destruction
of concrete buildings or roads, etc. There are references in the literature to the quantity and
composition of hydrated cement forms emanating from the cement industry,!3 but the
quantity resulting from construction is not well documented because of the difficulty
presented in monitoring such an operation.
In some of the samples, metal turnings amounted to as much as 1% and more of the total
sample. This level of metal turnings implies a high level of other types of metal working
particulate. Welding, power grinding and metal torch cutting result in large amounts of
particles in the furnace ash and/or furnace sphere group. Foundries, metal machining
industries, and general steel construction probably contribute to the furnace ash, furnace
sphere and furnace slag classifications. The amount of metal turnings in a sample may be an
index to the amount of furnace spheres and ash attributable to metal working as opposed to
metal production.
Table 15 clearly shows that natural minerals contribute from 10-45 wt % of the total
particulate. These minerals may originate from wind erosion, vehicle movement on dirt or
gravel roads, or construction involving the use of dried aggregates. Concrete minerals are
found throughout the Basin, though cement cinders are not present in detectable
concentrations in every sample. Oil soot is also a large contributor.
An average of 30% of the total particulate loading originates from natural sources when
considering representative ("normal") samples. Those samples taken during the air
55
-------
stagnation alert show an average of 23% for natural particulate with a corresponding
increase in particulate attributable to transportation and industry. Additionally, the
stagnant air samples showed a more uniform distribution of particulate among the sites and
particle classes than the representative samples; this implies quite a bit of atmospheric
mixing at low elevations including re-entrainment of particulate.
SCANNING ELECTRON MICROSCOPE ANALYSIS
Scanning electron microscope secondary electron images of airborne particulate are
displayed in figure 20. Most of these particulate were found to be submicron versions of or
fragments of similar particulate observed in the optical analysis. It was of particular interest,
however, to see at 10,000X that there are many extremely small particulate (< 0.1)
contained in the ambient samples.
X-RAY DIFFRACTION ANALYSIS
The X-ray diffraction results from source and ambient samples are listed in table 16. The
predominant compounds identified were Fe30«4, Si02 and feldspars in the ambient air test
specimens. The attempt to segregate these compounds out by density was not successful due
to the nature of the particulate, i.e., some high density material occurs as fine coating on
thin-shelled glass bubbles (observed optically).
COMPUTER ANALYSIS
Results from the statistical analysis by computer appear to be valid only when a minimum
of 50 data sets is available for analysis. Data sets include results from gravimetric,
meteorological, and elemental analyses; the number available for statistical analysis being
limited by the number of specimens analyzed for elemental composition. The last sampling
date analyzed by EDX was September 26, 1974, so a total of 110 samples (75 north, 35
south and mixed wind conditions) was available for statistical analysis.
The General Linear Hypothesis program, BMDX64, relates the independence of each linear
term of the model (ju i,oei,ak) with respect to the dependent variable. The calculated values
of the F-random variable,20 with appropriate degrees of freedom, provide a test of the
independence or interrelationship of meteorological variables and site locations. Table 17
lists these values of F calculated using the 75 data sets (samples) associated with northerly
winds and based on concentration of variable per cubic meter of air. There is 95%
confidence (F > 3.97) that there is some minimum amount of ashable organics present in
the atmosphere at all times. There is a 90% confidence (F >2.77) that TSP, PC and Cu are
present at all times. It may be stated with 95% confidence that the following variables are
affected by geographic location of the samples: TSP, PC, Ca, Mn, Fe, Zn, Pb, and Zr; while
Al, Si and K (90% confidence) also seem to be related to site location. Wind speed affects
the amount of ashable organics, Si, S, CI, V (95% confidence), K, Zn, and As (90%
confidence). The intensity of the inversion affects the largest number of variables.
56
-------
A. 1.000X
B. 5.000X
C. 20.000X
Figure 20, Representative SEM photographs (K6Q, 8-21-74)
57
-------
Table 16, X-RAY DIFFRACTION ANALYSIS OF AIR AND SOURCE TESTS
Specimen
Inorganic compounds detected
K60 (7-25-74)
FpgO^, Si02, feldspars
K60 (8-21-74)
FegO^, Si02, feldspars, AI2Q3
K59 (7-25-74)
Fe304
K59 (8-21-74)
FegO^, Si02, feldspars
Harbor Island road dust
SiC>2, feldspars
Road dust (Port of Seattle)
SiOj, feldspars
Gypsum plant
CaS04 2H20
Battery processor baghouse
PbS04, PbCI2
Cement plant
Cooling tower
CaCOg, Si02
Electrostatic precipitator
Alite
Baghouse emission
CaC03, CaS04 2H20 and
?(CdPb05, K3Na(S04)2, Ba-jMgSijOg, St02)?
Steel plant, high-
Fe304, ? (SiO2)^
temperature baghouse
58
-------
Table 17. STATISTICAL F VALUES CALCULATED
FROM 75 NORTHERLY WIND SAMPLES
(if F > 3,97 then 95% confidence; if F > 2,77 then 90%
confidence for all columns except a)
Dependent
variable
V
a
WS
II
ID
LR
AQ
TSP
3,31
8.52
1.16
31.14
0.20
0.02
0.98
PC
3.16
6.90
1.35
32.38
0.32
0.08
1.25
AO
4,26
1.56
10.40
54.52
0.17
0.11
0.64
Al
1,01
2.12
1.70
1.39
0,13
0.06
0.04
Si
0,70
1.97
4.21
1.55
1,47
0.33
0.13
S
1,63
1.48
4,50
0.10
0,36
0.13
0.24
CI
0,96
1.18
5.51
0.25
3,68
0,61
2.03
K
0,98
2.02
3.76
1.58
0.58
0.03
0.16
Ca
0.08
3.43
0.27
0.91
1,12
1.05
1.00
Ti
1.40
1.68
0.63
3.96
0.48
0.13
0.001
V
1.66
1.46
12.85
1.12
0.95
2.24
1.02
Cr
0.30
0.37
0.30
16.73
2.38
0.56
0,77
IVln
2.08
13.54
0.39
13.61
0.74
0
0.66
Fe
0.03
10.95
0.33
16.60
0.81
0.66
1.66
Ni
1.09
0.79
0.08
5.99
1.53
0.40
0,14
Cu
3.18
0.68
0.62
5.61
0.87
0.32
0.90
Zn
0.96
10.92
2.95
18,74
0.27
0.08
1.33
Br
0.10
1.79
0
7.44
1.92
0.06
0.43
Pb
0.73
3.39
1,87
8.97
0.86
0.82
0
Sr
0.45
0.97
0
2.31
1.07
7.88
2.67
Zr
0,08
3.50
0.39
17.97
4.69
2.20
0.03
Nb
0.01
0.21
0.12
1,96
0
1.43
1.18
Mo
0.35
0.60
0.01
0.34
1.54
0
0.27
As
0.42
1.36
3.08
6.44
1.38
0.01
0,75
Explanation of symbols and abbreviations
= minimum amount of dependent variable present in atmosphere at all times
a = site contribution to amount of dependent variable; 95% confidence if F > 2.34
ancf 90% confidence if F > 1,93
WS = wind speed
II = intensity of inversion
)D = duration of inversion
LR = time since last rain
AQ = soiling index as measured by tape sampler at K55
TSP = total suspended particulate normalized to a 24-hr period
PC = particulate concentration
AO = ashable organics
59
-------
BMDX64 also calculates the concentration for each dependent variable by site. These values
are listed in table 18 and illustrated in appendix A, because the visualizations are quite
helpful for interpreting the results of both computer programs.
The Stepwise Regression program, BMD02R, calculates the effect of the dependent variables
on the variation of PC and a symmetric correlation matrix for all 30 of the independent and
dependent variables. Table 19 illustrates the percent correlation between each variable; a
positive value indicates that the quantities influence each other in such a manner that if one
quantity increases so does the other; a negative correlation indicates that if one variable
increases the other will decrease. Please note the high correlation between normalized TSP
and PC; this indicates that a constant rate of flow of air was maintained over the three
months of this study for which elemental data are available. Ashable organics and iron
account for 90% of the variation of particulate concentration.
60
-------
Table 18. AVERAGE CONCENTRATION OF VARIABLES BY SITE
(/jg/m^ except TSP, mg)
Variable
Site
K60
K55
K99
K61
K64
K59
TSPa
218.292
151.578
139.710
115.924
155.662
97,492
PCb
145.500
102.500
95.683
77.731
99.407
73.715
A0C
24.910
24.462
23.375
19.469
21.673
21.051
Al
0.715
1.030
0.651
0.556
1.222
0.639
Si
3.337
9.423
5.862
10.959
14.556
6.020
s
2.118
7.128
4.659
5.795
7.257
3.136
CI
2.824
1.768
1.010
3,854
3.640
1,134
K
0.930
1.552
1.256
1.933
2.260
1,009
Ca
4.783
14.064
6.083
7.245
7.519
3.257
Ti
0.381
0.302
0.293
0.256
0.430
0.226
V
0,040
0.055
0.033
0.034
0.036
0.023
Cr
0,062
0.072
0.078
0.651
0,037
0,053
Mn
0.329
0,217
0.132
0.102
0.111
0.113
Fe
14.181
7.860
4.946
3.937
5.131
4.005
Ni
0.033
0.074
0.046
0.045
0.029
0,030
Cu
0.515
0.603
0.456
0.477
0.611
0.606
Zn
1.127
0.580
0.342
0.289
0.318
0,256
Br
0.075
0.041
0.065
0.067
0.113
0,235
Pb
3.936
2.492
1.500
1.656
1.939
2.579
Sr
0.052
0.041
0.107
0.030
0.037
0.037
Zr
0.055
0.033
0.018
0.025
0.028
0.022
Nb
0.145
0.217
0.240
0.156
0.216
0,176
Mo
0,005
0.001
0.003
0
0.005
0
As
0,006
0,005
0.006
0.006
0.001
0.002
No. of readings
(9)
(13)
(12)
(13)
(15)
(13)
aTSP = total suspended particulate normalized to a 24-hr sample period, mg
L_ O
PC = particulate concentration, Mg/m
c 3
AO = ashable organics, jug/m
61
-------
Table 19. SYMMETRIC CORRELATION MATRIX CALCULATED FROM 75 NORTHERLY WIND SAMPLES
(Percent correlation)
WS TH TL !L IH H
ID
LR
AQ
TP
PC
AO
Al
Si
s
CI
K
Ca
Ti
V
Cr
Mn
Fe
Ni
Cu
Zn
Br
Pb
As
25 20 1 27 31 20
32
30
-15
33
33
30
17
13
13
10
12
4
22
14
7
6
13
-10
-16
3
8
20
7
100 12 -20 14 -8 -7
4
2
-8
-10
-11
-25
-13
-24
-21
-22
-20
-9
-7
-38
0
-4
-12
-21
-17
-23
-2
-15
1
100 79 0 17 19
19
24
-26
30
31
30
17
10
13
17
21
6
23
12
10
12
16
5
-5
12
18
16
17
100 -2 6 IB
14
20
-23
29
30
25
26
28
20
37
32
10
36
18
9
17
23
0
-2
18
19
2
3
100 71 -2
64
66
11
7
7
-3
17
6
5
-7
-4
11
15
-3
-10
9
3
-13
-21
-6
-9
-9
-5
100 40
82
50
-9
23
24
28
28
10
22
4
9
14
21
20
0
9
7
4
-19
8
8
5
-3
100
45
13
-25
53
54
71
24
8
14
6
19
17
27
28
49
37
34
34
22
37
34
44
16
100
52
0
26
26
32
19
0
14
-14
3
17
10
17
14
18
13
18
0
12
0
29
15
100
-31
13
14
11
10
9
1
10
6
-6
11
2
S
5
8
4
-5
2
5
8
17
100
-8
-8
-11
-1
-6
6
-26
-12
13
-12
-10
-17
4
0
-5
6
2
-21
-6
3
100
99
81
36
10
14
22
27
26
54
31
23
67
77
15
25
73
18
52
21
100
84
38
11
16
22
28
27
54
33
26
7Q
78
17
27
74
21
54
24
100
28
6
14
8
19
16
35
36
35
49
51
26
24
58
36
52
25
100
72
79
58
72
69
82
55
3
22
28
14
22
18
14
13
25
100
84
75
92
64
75
54
-5
-1
10
18
23
3
9
-2
18
100
61
82
79
68
73
-4
6
10
32
25
5
4
-2
30
Explanation of table heading!
100
80
38
62
43
—4
0
21
9
19
14
9
-2
15
WD = wind direction
100
65
83
55
5
9
23
27
35
17
13
11
24
WS = wind speed
100
60
63
4
31
31
36
31
26
-6
11
27
TH = ground temperature, high
100
46
12
40
49
12
29
35
15
25
24
TL = ground temperature,
ow
100
6
27
29
56
36
34
5
27
23
IL = inversion base (lower height]
100
46
33
28
30
24
5
41
14
IH = inversion height
II » inversion intensity
100
88
32
42
75
2
61
22
ID * inversion duration
100
29
44
89
12
68
27
LR - time since last rain
100
48
36
3
37
37
AQ = COH index as measured by tape sampler
at K55
100
44
16
56
35
TP = total suspended particulate normalized to 24 hours
100
12
68
17
PC - particulate concentration
100
25
10
AO = ash able organic*
Al—As are the chemical symbols for the elements found by EDX
100
26
100
a*,
to
WD 100
WS
TH
TL
1L
IH
II
ID
LR
AQ
TP
PC
AO
Al
Si
S
a
K
Ca
Ti
V
Cr
Mn
Fe
Ni
Cu
Zn
Br
Pb
As
-------
SECTION VII
REFERENCES
1. Olsen, R. H. The Suspended Particulate Problem in Seattle's Duvvamish Basin,
PNWIS-APCA Paper No. 73-AP-32. (Presented at the Pacific Northwest International
Section - Air Pollution Control Association Annual Meeting. Seattle, Washington.
November 28-30, 1973.) 10 p.
2. Knechtel, K. B, Application of an Urban Diffusion Model to Modeling Suspended
Particulates in the Puget Sound Air Quality Control Region. PNWIS-APCA Paper No.
71-AP-21. (Presented at the 1971 Meeting of PNWIS-APCA.)
3. Roberts, J. W., A. T. Rossano, Jr., P. B. Bosserman, G. C. Hofer, and H. A.
Watters. The Measurement, Cost and Control of Traffic Dust in Seattle's Duwamish
Valley. PNWIS-APCA Paper No. 72-AP-5. (Presented at the 1972 Meeting of
PNWIS-APCA.)
4. Regulation I of the Puget Sound Air Pollution Control Agency. Washington Clean Air
Act, RCW 70.94. 410 West Harrison St., Seattle, Washington. Dec. 12, 1973. 49 p.
5. Reference Method for the Determination of Suspended Particulate in the Atmosphere
(High Volume Method). Federal Regulations Vol. 36, No. 84. Friday, April 30, 1971.
6. Rhodes, J. R., A. H. Pradzynski, C. B. Hunter, J, S. Payne, J. L. Lindgren. Energy
Dispersive X-Ray Fluorescence Analysis of Air Particulates in Texas. Environmental
Science and Technology. 6:922-27, October 1972.
7. Dixon, W. J. Biomedical Computer Programs. University of California Press, 1973.
8. McCrone, W. C. and J. C. Delly. The Particle Atlas, Edition Two. Ann Arbor Science
Publications, Inc., 1973. Four volumes, 1183 p.
9. Seely W. Mudd Series. Industrial Minerals and Rocks. The American Institute of
Mining, Metallurgical and Petroleum Engineers Publication, 1960.
10. Crutcher, E. R. Forensic Applications of Pollen Analysis. (Unpublished paper
presented at the 1973 meeting of the Northwest Forensic Scientists Association,
Missoula, Montana.)
11. Heinrich, E. W. Microscopic Identification of Minerals. McGraw-Hill, 1965.
12. Herdan, G. Small Particle Statistics. Academic Press, Inc., 1960.
13. Insley. H. and V. D. Frechette. Microscopy of Ceramics and Cements; Including
Glasses, Slags, and Foundry Sands. Academic Press, Inc., 1955.
63
-------
14. Kirk, P. L. Density and Refractive Index: Their Application in Criminal Identification.
Charles C. Thomas, 1951.
15. Leigh-Dugmore, C. H. Microscopy of Rubber. W. Heffer & Sons, (England), 1961.
16. Rogers, A. F. and P. F. Kerr. Optical Mineralogy. McGraw-Hill, 1942.
17. Winchell, A. N. and H. Winchell. Optical Properties of Artificial Minerals. Academic
Press, Inc., 1964.
18. Beyon, J. H. and A. E. Williams. Mass and Abundance Tables for Use in Mass
Spectroscopy, Elsevier Publishing Company, New York, 1963.
19. Gordon, R. J. and R. J. Bryan. Environmental Science and Technology. 7:1051-1054,
November 1973.
20. Bowker, A. H. and G, J. Lieberman. Engineering Statistics. Prentice-Hall, Inc.,
Englewood Cliffs, N, J., 1959, pp. 84-87.
64
-------
SECTION VIII
GLOSSARY OF ABBREVIATIONS
COH - Coefficient of haze; an air quality index measured by the amount of light passing
through a sample collected on a continuous tape.
DOE - Washington State Department of Ecology.
EDX - Energy dispersive X-ray (analysis); a method of determination of elemental
composition through characteristic X-ray fluorescence.
Hi-Vol - High-volume air sampler.
LTA - Low-temperature ashing; a method for determination of the total amount of
oxidizable organics.
NA - Numerical aperture.
NBS - National Bureau of Standards.
PC - Particulate concentration, ft g/m3; gravimetric data.
PSAPCA - Puget Sound Air Pollution Control Agency.
TLC - Thin layer chromatography; a microchemical method for characterization and
separation of organic compounds.
TSP or TP - Total suspended particulate, gm or mg; gravimetric data.
£- Micron.
/u g/m3 - Gravimetric data.
65
-------
APPENDIX
Figures 21 through 35 are helpful in visualizing the extent of correlation between variables
and the variation of chemical elements by geographic location. These figures represent
arithmetic averages for 75 samples from days having predominantly northerly winds.
66
-------
—J 1 I I 1 I
K60 K55 K99 K61 K64 K59
SITE
Figure 21, Average concentration of aluminum by site (northerly winds)
67
-------
16
Figure 22. Average concentration of silicon by site {northerly winds)
68
-------
Figure 23. Average concentration of sulphur by site (northerly winds)
69
-------
I I 1 I I I—
K60 K55 K99 K61 K64 K59
SITE
Figure 24, Average concentration of chlorine by site (northerly winds)
70
-------
2,4
2,2
2.0
1.8
1.6
1.4
1.2
1.0
_L
K6Q
K55
K99
SITE
K61
K64
K59
Figure 25. Average concentration of potassium by site (northerly winds)
71
-------
16
14
12
10
K60
K55
K99
K61
K64
K59
SITE
Figure 26, Average concentration of calcium by site (northerly winds)
72
-------
<
t
H
_J L_ JL_ J I L_
K60 K55 K99 K61 K64 K59
SITE
Figure 27. Average concentration of titanium by site (northerly winds)
73
-------
.06
.05
.04
.03
Q2 ,
K60 K55 K99 K61 K64 K59
SITE
Figure 28. Average concentration of vanadium by site (northerly winds)
74
-------
.4
,1
K60
K55
K99
K61
K64
K59
SITE
Figure 29, Average concentration of manganese by site (northerly winds)
75
-------
16
14
12
10
8
6
4
2
0
K60
K55
K99
SITE
K61
K64
K59
igure 30. Average concentration of iron by site (northerly winds)
76
-------
.02
I 1 1, 1 I L.
K60 K55 K99 K61 K64 K59
SITE
Figure 31, Average concentration of nickel by site (northerly winds)
77
-------
,7 "
.4
K60
K55
K99
K61
K64
K59
SITE
Figure 32, Average concentration of copper by site (northerly winds)
78
-------
I I I I I 1—
K60 K55 K99 K61 K64 K59
SITE
Figure 33. Average concentration of zinc by site (northerly winds)
79
-------
I I 1 I 1 i—
K6Q K55 K99 K61 K64 K59
SITE
cigure 34, Average concentration of lead by site (northerly winds)
80
-------
.4
.1
I I I I I I
K60 K55 K99 K61 K64 K59
SITE
Figure 35. Average concentration of niobium by site (northerly winds)
81
-------
? TECHNICAL REPORT DATA
(Please read Instructions on (he reverse before completingj
REPORT NO, 2,
EPA 910/9-75-010 '
3. RECIPIENT'S ACCESSIOWNO,
t. TITLE AND SUBTITLE
i A study of the suspended particulate problem
' in the Duwamish basin
5, REPORT DATE
May 1975
6. PERFORMING ORGANIZATION CODE
7, AUTHORiS)
Ray Olseti, Marcia Y. Almassy, and Lewis Wingert
8. PERFORMING ORGANIZATION REPORT NO,
9, PERFORMING ORGANIZATION NAME AND ADDRESS
Boeing Company
Seattle, WA 98124
10. PROGRAM ELEMENT NO,
11. CONTRACT/GRANT NO.
68-02-1499 ;
12. SPONSORING AGENCY NAME AND ADDRESS
U. S. Environmental Protection Agency
Surveillance and Analysis Division
1200 6th Avenue
Seattle, WA 98101
13. TYPE OF REPORT AND PERIOD COVERED
Final, July-November 1974
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Air quality data accumulated since 1965 indicate the primary and secondary national
air quality standards have been exceeded in the Duwamish Basin area of Seattle,
Washington. The objective of this study was to determine the nature of suspended
particulate and subsequently to quantify the impact of particulate sources. Ambient
and source particulate was collected on fiberglass and membrane filters. -Results
from gravimetric, elemental, and compound analyses were combined with meteorological
data for correlation and analyses. The contribution of sources for the basin was
found to be 27% from natural sources, 39% transportation, and 34% point industry
and multiple-area sources. Results from this study show the complexity of suspended
particulate. Source tests showed' a multitude of elements and compounds present and,
to complicate matters, surrounding soil has been contaminated by industrial and area
sources. About 35% of the particulate is directly related to road dust-type emissions.
\ :
t
/
17. KEY WORDS AND DOCUMENT ANALYSIS
a, DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution
Air Pollution Chemistry
Measurement Methods (Air Pollution)
18. DISTRIBUTION STATEMENT
Release unlimited
19. SECURITY CLASS (This Report/
Unclassified
21. NO. OF PAGES
81
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 C9-73)
------- |