Environmental Protection Technology Series
FACTORS AFFECTING THE COLLECTION
EFFICIENCY OF ATMOSPHERIC
SULFATE
Environmental Sciences Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
Environmental Health Effects Research
Environmental Protection Technology
Ecological Research
Environmental Monitoring
Socioeconomic Environmental Studies
Scientific and Technical Assessment Reports (STAR)
Interagency Energy-Environment Research and Development
“Special” Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en
• vironmentat degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of potlut,on Sources to meet environmental quality standards
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
1.
2.
3,
4.
5.
6.
7.
8.
-------
EPA-600/2-77-076
May 1977
FACTORS AFFECTING THE COLLECTION
EFFICIENCY OF ATMOSPHERIC SULFATE
R.W. Coutant
BATTELLE
Columbus Laboratories
Columbus, Ohio 43201
Contract No. 68-02-1784
Project Officer
Eva Wittgenstein
Sampling and Analysis Methods Branch
Atmospheric Chemistry and Physics Division
Research Triangle Park, North Carolina 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences Research
Laboratory, U.S. Environmental Protection Agency, and approved for publication.
Approval does not signify that the contents necessarily reflect the views and
Policies of the U.S. Environmental Protection Agency, nor does mention of trade
names or commercial products constitute endorsement or recommendation for use.
11
-------
ABSTRACT
[ Lactors that influence the collection and measurement of atmospheric
sulfate were investigated. Special emphasis was given to those factors that
cause the formation of extraneous sulfate during the sampling process. The
factors considered were filter type and composition, ambient SO 2 concentra-
tion, temperatUres relative humidity, ambient gas composition, sampling time
and rate, storage time and conditions, and the presence of potential oxidation
catalysts in the particulate cat
The approach of the investigation was twofold. After conducting an exten-
sive laboratory program to identify and quantify significant sulfate generating
interactions, a brief field study was performed to test the significance of the
laboratory observations under typical field operating conditions.
The results of this investigation indicated that the most significant cause
of sulfate sampling error is the interaction of basic filter components with
ambient SO 2 . This interaction is affected by ambient atmospheric conditions.
A relationship based on established chemistry was developed for the prediction
of sulfate error caused by this interaction. Recommendations of filter media
appropriate for ambient sulfate monitoring and further development of a method-
ology are included.
This report was submitted in fulfillment of Contract No. 68-02-1784 by
Battelle’s Columbus Laboratories under the sponsorship of the U.S. Environmental
Protection Agency. This report covers the period April 15, 1975 to November 30,
1976, and work was completed as of November 30, 1976.
111
-------
C ONT E NTS
1 . Introduction
Objective and General Approach.
2. Laboratory Studies
Apparatus
Results
Filter Composition
Comparative Filter Responses
Storage Effects
Catalyst Effects
Sampling Velocity
Sampling Time
SO 2 Absorption
Discussion
General Observations .
Oxidation Catalysts
Chemistry of SO 2 Sorption by
Oxidation of Absorbed SO 2 .
Sulfate Sampling Error
Conclusions
3. Field Sampling Program
Objective
Approach
Siting
Fire Station No. 5.
Pymatuning State Park
Columbus
Experimental Plan
Problem Areas
Analyses
Results
Discussion
Conclusions
4. RecommendationS.
References.
Appendices
A. Gas Composition Data
3. Field Sampling Data
C. Los Angeles Field Sampling Data.
D. Filter Identifications
111
vi
vii
1
2
3
3
6
6
9
9
12
15
15
16
20
20
20
Filters 21
27
27
30
31
31
31
31
32
32
32
34
35
35
36
36
42
44
46
48
58
85
90
Abstract
Figures
Tables
v
-------
FIGURES
Number Page
1 Laboratory Sampling Rig . 4
2 Laboratory Instrumentation 5
3 Typical SO 2 Response Curve 17
4 Effect of Filter and Atmospheric Variables on SO 2
Sorption . . . . 26
5 Sulfate Error for Basic Filters. . . . . . 29
6 Youngstown Field Sampling Area . . . 33
7 Correlation Between Calculated and Measured Sulfur
Differences 37
vi
-------
Number
1 Filter Analyses (OES) .
2 Filter Alkalinities
3 Filter Response Comparison
4 Effect of Storage on Apparent Sulfate.
5 Particulate Composition - Run 4A
6 Catalyst Effect on Sulfate Collection.
7 Results of Velocity Effect Experiment.
8 Sampi i ng Time Effects
9 SO 2 Sorption by Various Filters
10 Sorption of SO 2 by Basic Filters
11 Experimental and Predicted Excess Sulfates
12 Field Sample Summary
13 Field Sample Data
14 Statistical Comparison of ADL Filter Performance
With Other [ ii ters
Gas Composition Data
Elemental Analysis of Field Samples,
25 Field Atmospheric Condi tionc
Los Anqel es Samp ii nq Data
Los Anqel es Ambi cut Condi tions (Hourly Averaqes
D—l Filter Identifications
TABLES
Page
8
9
10
11
13
14
15
16
18
19
28
38
40
A-1-A-9
B-i
B- 2-B-
C-i
C-?- C- 6
Weiqht Percent.
• . . . 41
49-57
• . 59
60-84
• . . . 86
• . 87-89
90
vii
-------
SECTION 1
INTRODIJCTI ON
The need for accurate determination of ambient sulfate levels has been
intensified in recent years because of the identification of sulfates as signif-
icant health hazards. While progress has been made in reducing atmospheric levels
of sulfur dioxide, corresponding reduction in ambient sulfate levels has not, in
general, been well correlated with the reductions in sulfur dioxideW. This
fact along with both direct and indirect criticisms of the standard hi-vol sampl-
ing techniques for ambient sulfate collection have made obvious the need for a
thorough investigation of sulfate sampling and analysis procedures.
Investigations such as those conducted by Lee and Wagman and more recently
by Radian personnel (3) and workers at Ford 4 have shown that filters commonly
used for sulfate sampling can absorb SO 2 at ambient levels, and that this can
lead to generation of extraneous sulfate during the sampling process. It can
(5)
be inferred from other investigations that the presence of potential SO 2
oxidation catalysts in ambient aerosols also may affect ambient sulfate measure-
ments. Still other questions have been raised concerning the mechanical arid
dynamical factors involved in hi—vol sampling.
These potential sources of error during ambient sulfate sampling have been
recognized by the U.S. Environmental Protection Agency, with the result that this
program was developed as a part of EPA s overall efforts towards refinement of
ambient air sampling and analysis methodology. The main thrust of this program
has been examination of the effects of sampling and environmental factors on the
formation of artifact sulfate. The present work included consideration of the
effects of (1) filter type and composition, (2) ambient SO 2 concentration, (3)
temperature, (4) relative humidity, (5) ambient gas composition, (6) sampling
time and velocity, (7) sample storage time and conditions, and (8) particulate
composition (SO 2 oxidation catalysts) on the formation of extraneous sulfate,
in order to provide a quantitative base for the assessment of sulfate sampling
methodology.
-------
OBJECTIVE AND GENERAL APPROACH
The objective of this program was a thorough investigation of the factors
that influence the collection and measurements of atmospheric sulfate. Special
emphasis was given to those factors giving rise to extraneous sulfate formation
as a consequence of sampling and measurement procedures.
The approach of this program has been twofold. The initial phase of the
program was devoted to study of various sulfate collection parameters in the
laboratory. Although this phase of the program involved controls possible only
within the laboratory setting, ambient air served as a base for all measurements.
The second part of the program involved a field sampling study in an attempt to
verify the observations of the laboratory study under field operations conditions.
2
-------
SECTION 2
LABORATORY STUDIES
APPARATUS
Ambient air was used as a base for all of the laboratory experiments. This
was drawn into the laboratory at about 520 cfrn through a stainless steel duct
from a point some 20 feet above the building (ca. 50 ft about ground level). In
the laboratory, the air stream was divided through two parallel stainless ducts
(see Figures 1 and 2). At the top of each inside duct, injection ports and mix-
ing orifices were added to enable spiking of the ambient air with various gases
as needed for specific experiments. Sampling nozzles were fitted near the bottom
of the parallel ducts, with provision for four identical sampling points in each
duct. This arrangement permitted simultaneous isokinetic collection of as many
as eight filter samples in any given run. Sample volume for each filter was
about 1.2 percent of the total air flow. Standard 142-mm Millipore filter holders
were fitted with modified conical inlets to achieve uniform distribution of
particles over the filter surface. Outlets from the filter holders were fitted
with calibrated orifices, and pumping across the orifices was achieved via a
vacuum manifold using a Stokes Microvac pump with a capacity of 80 cfm. Normal
24-hour sampling volumes were of the order 1 n1 3 /cni 2 .
Routine measurements were made of the concentration (by flame photometry),
NO (by chemiluminescence), the ozone concentration (by chemiluminescence), temper-
ature, relative humidity, filter pressure drops, wind speed, and wind direction.
Analyses for other gaseous components such as H 2 5 and NH 3 were made for specific
experiments. Monitoring during the course of a run was automatic, with both strip-
chart readouts and paper tape ar mHlation for computer arlal’/sis.
PROCEDURES
The laboratory apparatus was used for basically two types of experiments;
24-hour sampling runs, and shorter term examination of SO 2 sorption by filter
media and by filters preloaded with various potentially active SO 2 oxidation
3
-------
FIGURE 1. LABORATORY SAMPLING RIG
-------
—
C - ,,
FIGURE 2.
LABORATORY INSTRUMENTATION
-------
catalysts. In the 24-hour runs, preweighed filters were mounted in the holders
and were exposed to ambient air for a period of 24 hours. The filters were then
removed, weighed in a constant temperature/humidity room, cut in half, reweighed,
and submitted for analysis. In most cases, one half of each filter was analyzed
as quickly as possible, and the other half was stored for possible analysis at a
later date. Vaiations between samples collected in similar filters at the eight
different samling ports were random and were usually 1-3 percent based on either
the total masses or on the sulfate catches.
In shorter term SO 2 sorption runs, the air flow in one branch of the parallel
duct system was spiked with SO 2 to some appropriate level between ambient and 300
ppb. This level was preset without a filter in place. The filter was then placed
in the system; the SO 2 was turned on; and SO 2 concentrations before and after the
filter were monitored as a function of time. Inasmuch as fluctuations in ambient
temperature, relative humidity, and SO 2 concentration were prone to occur during
these experiments, data displaying variations greater than 2 percent in any one
variable were rejected.
In examination of potential catalyst effects, dispersions of the agents in
alcohol were sprayed onto the filters. In most cases, reagent grade or pigment
grade agents were used. However, the vanadium oxide employed was a presized
sample (
-------
Listings of OES results for the various filter types studied are shown in
Table 1. Alkalinities and pH’s for these same filter types are shown in Table 2.
It is notable in these results that the glass fiber filters display high concen-
trations of sodium, potassium, calcium, and other basic components, and also show
relatively high alkalinities and pH’s. The high purity silica filters display
only slightly basic character, and the cellulose acetate (Celotate) and Pallflex
E 70/2075 W filters are slightly acidic. Samples of the Gelman AE filters showed
a slight variation in alkalinity with examples from different positions in the
package; a filter from the top of the package displayed the lowest alkalinity.
While a hot-leach procedure (ASIM D 202) was used with most of the filters, one
filter was leached at room temperature with the result of exhibiting only about
half as much available alkalinity as the hot-leach samples.
Comparative Filter Responses
Three series of 24-hour sampling runs were made to evaluate differences
between the responses of different filter media. Results of these measurements
are shown in Table 3. While the absolute quantities of sulfate collected varied
considerably, like pairs of filters consistently yielded similar results. Sulfate
variations were therefore primarily due to differences between the filter media.
Storage Effects
Data given in Table 4 are the results of an experiment designed to reveal
the effects of storage environment on apparent sulfate collection. In this experi-
ment, two filters of each of four different types were exposed to identical ambient
sampling conditions for a 24-hour period. Halves of the filters were then treated
as follows:
(a) No storage - immediate analysis
(b) Storage for 4 weeks in dry air
(c) Storage for 4 weeks in wet air
(d) Storage for 4 weeks in argon.
Results of the sulfate analysis shown in Table 4 indicate large differences between
apparent sulfate collections by the various filter types. An analysis of variance
of this data reveals no significant variation with respect to storage condition,
but differences between the responses of different filter types are significant
at the 95 percent confidence level.
7
-------
TABLE 1. FILTER ANALYSES (OES)
Gelman
Element MSA-flO6BH AE
Gelman
AA
Elemental
Spectro-
grade
Content,
AOL
weight
Mitex
percent
Celotate QAST E 70/2075w
Al 3 3 2-4 3 0.1 1 0.0007 0.02 0.07
B 1 1 1 1 0.007 <0.001 0.0007 <0.002
Ba 0.01 0.2 0.1 <0.002 0.025 <0.0002
Ca 8 8 1-2 8 8 0.03 0.003 0.02 0.5
Cd <0.01 <0.01 <0.01 <0.01 <0.01 <0.0002
Co <0.001 <0.001 <0.001 <0.001 <0.001 <0.00007
Cr <0.002 <0.002 0.002 <0.002 <0.002 0.003 0.003 <0.002 0.002
Cu <0.001 <0.001 <0.001 <0.001 <0.001 0.001
Fe 0.04 0.04 0.1 0.04 0.02 0.05 0.0007 0.02 0.02
K 1 1 0.5 1 <0.02 0.8 0.01 0.01 0.4
Mg 2 2 1 2 0.01 0.02 0.002 0.01 0.2
Mo <0.002 0.002 <0.002
Na 10 10 2-3 10 0.1 0.3 0.002 0.1 4-7
Ni <0.001 <0.001 <0.001 <0.001 <0.001 0.002 <0.00007 0.001 <0.001
Pb <0.002 <0.002 0.002
Sr 0.005 0.01 0.0C5 0.003 <0.002 0.002 <0.0002 <0.005 0.05
Ti 0.003 0.005 0.005 0.035 0.01 0.03 0.00007 0.003 0.003
V <0.001 <0.001 <0.001 <0.001 0.002 <0.00007
Zn <0.01 0.1 <0.01 <0.01 <0.01 <0.01 0.001 <0.01 0.1
Zr 0.005 0.01 0.01 0.003 <0.001 0.00007
8
-------
TABLE 2. FILTER ALKALINITIES
Alkalinity,
Filter Type pH nieq/g
MSA-1 1O6BH 9.2 3.1 x io2
AOL Quartz 8.1 1.0 x 10
Mitex 7.0 0.0
Spectrograde 7.2 1.0 x
Celotate 6.65 (1.8 x
Paliflex QAST 8.1 3.8 x l0
Pailfiex 6 2 (7 6 10 -3 (a)
E 70/2075W . . X /
Gelman A -- 3.32 x 102
Gelman AA 8.9 3.24 x io2
Gelman AE(b) 94 4.0 x io2
Gelman AE ’ 9.2 5.0 x io2
Gelman AE ’ 9.3 4.89 x io2
Gelman AE ’ 93 4.90 x io2
Gelman AE ’ 9.2 4.34 x io2
Gelman AE(c e) 9.4 2.24 io2
(a) Acidic.
(b) Batch A.
(c) Batch B.
(ci) Different positions in same package.
(e) Five-minute room temperature leach.
9
-------
IABLE 3. FILTER RESPO1 SE COMPARISO
Run
No. Eli
S02
ter ppb
Apparent
Sulfate,
pg/rn 3
1A1 Mitex 180 25.8
1A2 Mitex 180 31.8
1A3 Spectrograde 180 (33•5)(
1A4 Spectrograde 180 (32.l)
lAS Gelman AE 180 46.8
1A6 Gelman AE 180 46.3
1A7 Celotate 180 26.6
1A8 Celotate 180 24.4
2A1 Mjtex 225 11.9
2A2 Mitex 225 11.6
2A3 MSA 11068R 225 25.2
2A4 NSA 11O6BH 225 24.9
2A5 ADL-Quartz 225 12.9
2A6 AOL-Quartz 225 12.7
2A7 Spectrograde 225 16.8
2A8 Spectrograde 225 — (b)
IOB1 Pallflex 107 2.2
E 70/2075—W
10B2 Paliflex 107 6.0
E 70/2075-W
1083 Gelman A.A 107 24.8
1034 Gelman AA 107 26.0
10B5 Paliflex QAST 107 14.4
10B6 Paliflex QAST 107 11.8
1087 Gelinan AE 107 28.1
1038 ADL 107 11.8
Average deviations: (sulfate/flow values)
Between filter pairs - 2.9
Between filter types - Run 1 - 30.7%
Run 2 - 36.4 ’ ;.
(a) Torn filter. Run 10 - 39 ’ ;
(b) Sample lost.
10
-------
TABLE 4. EFFECT OF STORAGE ON APPARENT SULFATE
Storage Sulfate,
Filter Type Condition pg/rn 3
ADL none 6.30
ADL argon (6.OO)
ADL dry air 8.24
ADL wet air 6.85
Celotate none l4.l7
Celotate argon 5.32
Celotate dry air 989 (b)
Celotate wet air 8.16
Gelman AE none 16.88
Gelman AE argon 16.69
Gelnian AE dry air 17.19
Gelman AE wet air 17.66
MSA 11063H none 12.55
NSA 11O6BH argon 15.49
MSA 11O6BH dry air 11.88
NSA 11O6BH wet air 14.32
(a) Sample lost, value estimated by Yates’
Method [ Kirk, Roger E., “Experimental
Design: Procedures for the Behavioral
Sciences”, Brooks/Cole Publishing Co.,
Belmont, California, p 146 147 (1968)].
(b) Questionable data.
11
-------
Catalyst Effects
Table 5 shows the results obtained for the analysis of a typical particulate
sample obtained in the laboratory sampling rig. The principal component of this
ambient particulate sample is carbon; of those elements likely to exhibit activity
as SO 2 oxidation catalysts, iron and manganese are the principal components. Both
the ADL microquartz filters and the MSA 1106-BH filters were used. Analyses for
sulfate, sulfite, and total nitrogen were carried out on both types of filters.
The results reflect the differences in basicities of these two filter types; the
more basic MSA filter shows higher sulfate and sulfite contents and lower nitrogen
content. The oxygen content of the sample was estimated by requiring that the
sample be electrically neutral, assuming that the carbon is present as elemental
carbon. The deficit in negative species indicated by this procedure could, of
course, also be logically represented in terms of carbonate. If the latter
hypothesis is applied, the total mass balance is 119 percent rather than the 102
percent shown in Table 5. Either way, the mass balance seems reasonable in light
of the number of individual analyses included in the sumation.
A series of experiments was conducted to determine the possible significance
of the presence of potentially active oxidation catalysts on the overall collect-
ion of ambient sulfate. These experiments included both short-term measurements
of SO 2 absorption by filters that were preloaded with catalysts, and 24-hour
sampling runs. Potential catalysts included Fe 2 0 3 (from three sources including
a basic oxygen furnace fly ash), Mn0 2 , carbon (acetylene black), vanadium oxide
(high vanadium fly ash), ferric sulfate, manganese sulfate, and a mixture of
carbon, and the oxides of iron, vanadium, and manganese. None of these materials
was found to increase the level of SO 2 absorption exhibited by the filter. In
fact, the more acidic iron and manganese sulfate-loaded filters tended to display
less SO 2 absorption than the blank filters. An analysis of variance for the 24—
hour sampling runs (Table 6) showed no significant effect (at the 90 percent
confidence level) due to the presence of the oxides or carbon. Even with the
vanadium fly ash sample, where the sulfate level appears to be relatively high,
the result is not statistically different from the other samples.
-------
TABLE 5. PARTICULATE COMPOSITION
RUN 4A
Total Mass
Fe
B
Mg
Mn
Pb
Al
Na
Ti
K
Ca
Zr
S 04
sO
S
Organic S
C
NH
NO
NO 2
N
0 (b)
8.717
0.25
0.00
0.13
0.03
0.13
0.00
1.25
0.00
0.13
0.00
0.00
2.05
N.D.
N.D.
x
3.8
0.48
0.14
<0.01
0.53
0.54
2.87
0.00
1.49
0.34
1.49
0.00
14.34
0.00
0.49
0.00
0.00
23.52
43.59
5.51
1.61
6.08
6.19
Total ii Y2.4
Particulate
Coniposi tion,
Species Mass, mg weight %
(a)
(3.24)
(0.11) (a)
(a)
(N.D.)
(0.35) (a)
(a) MSA filter, all other data on ADL filter.
(b) By charge balance.
13
-------
TABLE 6. CATALYST EFFECT ON SULFATE
COLLECTION
Loading,
Additive my
Sulfate,
pg/rn 3
Blank 7.36
Blank —- 8.28
Fe 2 0 3 0.9 7.08
Fe 2 0 3 2.8 7.07
Mn0 2 2.85 11.09
C 2.1 7.30
C 6.4 8.23
Blank 19.6
Blank 19.3
Mn0 2 0.35 20.5
Mn0 2 1.95 13.9
Mn0 2 3.80 22.3
BOF 1.46 19.6
V-ash 235 24.96
C 1.84
BOF 0.5
V-ash Mixture 0.5 21.5
Mn0 2 0.5
14
-------
Sampling Velocity
In this series of experiments, the flow rates in the two branches of the
laboratory duct system were modified such that flow through one branch was
approximately three times that in the other branch. Rate-controlling orifices
in the filter holders were modified accordingly to provide approximately iso-
kinetic sampling in each branch, and a 24—hour sampling run was conducted.
Results of this experiment are shown in Table 7.
TABLE 7. RESULTS OF VELOCITY EFFECT EXPERIMENT
Filter
Number
Total Gas
Volume, m 3
ig
3
S0 4 /m
1
56
3.80
2
142
5.44
3
54
5.00
4
150
3.48
5
54
3.76
6
143
3.38
7
56
2.77
8
137
1.97
Average
for low
rate
3.83
±
0.91
Average
for high
rate
3.57
±
1.43
Grand average
3.70
±
1.1
(a) Gelman AE filters.
The results of this experiment show somewhat larger variations between individual
samples than normally experienced with the laboratory rig. Three out of the four
“pairs” show higher sulfate values for the lower sampling rates. However, these
variations are not associated with sampling rate differences but appear to be
randomly distributed amongst all samples.
Sampling Time
In this series, a set of simultaneous samples was taken in which one filter
type was used to sample ambient air for 48 hours. During this time period, the
same filter was used to sample each of the 24-hour halves of overall time. Two
15
-------
filter types were employed; the ADL filter and the Gelman AE filter. Filter
samples were analyzed for both sulfate and sulfite contents, with the results
shown in Table 8.
TABLE 8. SAMPLING TIME EFFECTS
Sample
Number
Filter
Mass
Loading,
j.ig/rn
Sulfat ,
pg/rn
Sulfite,
pg/rn 3
5A
ADL
97.8
13.3
1.4
5B
ADL
112.0
20.3
0.8
5AB
ADL
105.1
15.5
0.6
5A
AE
114.6
13.8
5.2
5B
AE
159.3
24.8
8.0
5AB
AE
107.6
20.3
0.2
It is noteworthy in the results of this run that the ambient SO 2 levels for
the two 24-hour periods are appreciably different (see Appendix A). This fact
appears to be reflected in the relative amounts of sulfate collected on each day,
especially with the AE filter. The principal effect of sampling over the longer
time period appears to be the obvious one of averaging the high and low fluctu-
ations. The 48-hour sulfate values are close to the averages of the 24-hour
samples for each filter. Sulfite values cited in Table 8 also seem to reflect
some differences in basicity between the two filter types, but, in view of
difficulties encountered in analysis of sulfite in these samples, the data must
be considered semiquantitative at best.
SO 2 Absorption
Measurements of the absorption of SO 2 by various filter media were made under
conditions similar to those encountered in ambient air sampling. Although spiking
of the ambient air with SO 2 was employed in many of these runs, spiking levels
were kept within the bounds of ambient levels.
A typical response curve obtained in these experiments is shown in Figure 3.
Initially, the SO 2 concentration on the downstream side of the filter was close
to zero. This low reading persisted for varying times depending on the nature
16
-------
I.0
O.9
O.8
cL
30.7
ON
0
20
30
Time, mm
FIGURE 3. TYPICP L SO 2 RESPONSE CURVE
40 50 60
70
17
-------
of the filter, the SO 2 concentration, and the relative humidity. The independent
effect of temperature was not readily obvious however. After some period of time,
usually of the order of 5—15 minutes for the more basic filters, breakthrough
of the SO 2 was observed, and the response curve followed an S-shaped approach to
the input level of SO 2 . The total amount of SO 2 absorbed was calculated from
knowledge of the flow rates, concentrations, and numerical integration of the
response curve. Blank runs, i.e., runs without a filter in the system, also were
made, and data for the filter responses were corrected for the iiherent system
lag. Typical blank breakthrough times were of the order of 0.1 mm. ., and blank
corrections were of the order of 16 ig 502. Data for the total SO 2 sorption by
various filter types are given in Table 9 for comparison of different filter
behaviors. It is obvious from this table that the extent of SO 2 sorption by the
more basic filters (MSA l106—BH, Gelman AE, and Gelnian AA) is significantly
different from that of the other filters studied, Indeed, the SO 2 responses of
the acidic filters (Celotate and Paliflex E 70/2075 W) and the polymeric filters
(Mitex, Fluoropore, and Polyimide) were indistinguishable from the blank. The
more basic filters were therefore examined in more detail, with the results
shown in Table 10.
TABLE 9. SO 2 SORPTION BY VARIOUS FILTERS
_____ so 2
RH, Input, Capaci,ty,
Filter Type T,GC Percent ppb pg/cn1
tISA 11OGBH 25 51 230 6.04
Gelrnan AE 25 46 236 8.03
Gelmari AA 18 49 181 4.65
Spectrograde 25 45 236 LOB
ADL Microquartz 25 44 236 0.36
Paliflex QAST 14 60 286 0.27
Duralon 25 69 236 0.45
Pa llflex 20 44 150 0
E 70/2075W
Celotate 25 65 248 O.OO
Polyiniide 25 50 150 O.OO
Mitex 26 50 260 0.OO
Fluoropore IA 26 50 260 O.OO
Fluoropore FU 26 50 260
I . -. _____
(a) Indistinguishable front blank.
18
-------
TABLE 10. SORPTION OF 502 BY BASIC FILTERS
So 2
RH Input, Capacity,
Filter Type 1, °C Percent ppb pcj/cnl 2
IiSA-11O60H 25 40 150 5.63
ft 25 40 298 6 99
a 23 8 65 154 5•13
25 71 310 8.57
25 51 230 6.04
ii a 18 81 315 7.83
‘ 18 81 150 6.44
18 81 305 7.98
a 18 78 150 5.70
22 61 150 6.00
22 61 298 6.47
(b)
Geirnan AE 15 93.7 298 11.55
( i Sa tc h-A)
‘ ‘ 22.8 41 150 6.05
23.3 34.5 300
I a 25 45 326 8.03
II 27.8 48 146 4.85
(c)
Celman AE 2t .b 44 20! 6.58
(Batch-B)
26.8 39 300 5.48
II II 27.3 38 405 6.48
Ge1rnanAE 16 51 180 4.88
II 11.5 87.2 199 5.04
II 12.3 79.4 269 4.46
11.3 92.7 199 4.79
Gelman AA 1) 17 49 181 4.69
(a) Alkalinity = 0 . 215 u eq/cm 2 , BlarJ sulfate = 9.05 q/cn !.
(b) Alkalinity = 0,30?;. eq/cm’, Blank sulfate 1.1 q/cm 2 .
(c) Alkalinity = O.3?9 cq/cn 2 , Blank sulfate = 4.2 q/cm 2 .
(d) Alkalinity = 0.254 eq/cm 2 , Blank sulfate 0.39 g/cui 2 .
19
-------
DISCUSSION
General Observations
Significant variations in apparent sulfate collection were observed for
different filter media. These differences appear to reflect primarily chemical
differences between the various filters. Mechanical factors related to sampling
conditions, such as, sampling rate, and sampling time do not appear to affect
the apparent sulfate collection, although, an effect due to sampling rate might
well be expected. This latter parameter is discussed later in this report.
Storage of filter samples under various conditions also was not observed to
measurably affect the apparent sulfate collection by any given filter type. Nor
was there any appreciable effect on apparent sulfate collection due to the presence
of presumably active SO 2 oxidation catalysts on the filter surface. The only effect
noted was caused by a possible modification of the pH of the filter medium due
to the presence of some of the additives Indeed, the principal effect leading
to excess sulfur on the filter catch was due to the interaction between the filter
medium and SO 2 . This interaction is dependent on the SO 2 concentration, relative
humidity, temperature, and the chemical nature of the filter medium.
Oxidation Catalysts
Examination of Table 6 indicates that the addition of these potentially active
materials to the filter had little or no effect on the removal of SO 2 . However,
other investigators ‘ used similar procedures to demonstrate that iron and
manganese are catalytically active in the atmosphere. Comparison of the work of
these authors with that shown above reveals three important differences:
(1) The current work was done under realistic atmospheric
conditions, i.e., ambient level SO 2 .
(2) The catalyst forms used in the current work were not
identical to those studied previously. (It is well
recognized that catalytic activity of a given sub-
stance can vary greatly depending on preparation,
treatment, aging, etc. Furthermore, Cheng, et
clearly demonstrate that different chemical forms of
manganese have different activities.)
20
-------
(3) With the levels of SO 2 used in the above exper ments,
minimum rates expected from the work of Cheng, et
and Chun 6), might riot be distinguishable in the short-
term SO 2 sorption experiment.
Items 1 and 3 are related, and really concern the mechanism of SO 2 retention
by the particles. Urone, et al. 8 measured apparent rates of the order of 4
percent/mm but they do not present sufficient data for interpretation of
reaction order. Cheng, et cite a first—order mechanism, but revert to
the use of gross average rates in their extrapolation to ambient conditions.
There too data are insufficient for a good test of reaction order. Churi and
Quon 6 suggest a capacity limited transport mechanism with no dependence on
SO concentration, but all of their measurements appear to lie outside of any
kirietically controlled regime. Theoretical considerations by Freiberg
indicate a second order reaction. Problems caused by this unfortunate con-
fusion and the lack of good kinetics data impinge not only on the sulfate
sampling question but also on the general question of the fate of SO 2 in the
atmosphere.
The question of reaction order is very important to any comparison of
the current “catalyst” results with the earlier data mentioned above. Without
a clear definition of reaction order, the earlier work cannot be extrapolated
to an ient conditions. The current results do suggest that the overall reaction
of SO 2 with potential catalysts is concentration dependent, and that this type
of reaction does not contribute significantly to the artifact sulfate problem.
Chemistry of S02 Sorption by Filters
Many of the glass fiber filters commonly used for air pollution samping are
known to be hygroscopic and quite basic. For example, pH’s, as measured by the
ASTM D 202 procedure, are generally in the range of 8-10. It therefore seems
appropriate to examine the utility of the concept of absorption of SO 2 on the
moist-basic surface of these filters in terms of the well-characterized chemistry
of absorption of SO 2 by basic solutions.
The sorption of SO 2 in basic solutions has been studied in some detail with
regard to stack-gas cleanup systems. Much of the early work traces to a series
of papers published by H.F. Johnstone and co-workers during the late 1930’s and
early l94O’s 9 . Because of the relatively unsophisticated procedures used by
21
-------
Johnstone, his work has been subjected to numerous reexaminations, including a
program conducted by Coutant and Levy at BCL in 1968(10). The result is that the
sorption of SO 2 by various alkaline solutions has been reasonably well character-
ized over a fairly wide range of the variables.
For the purpose of the discussion, the chemistry of sorption of SO 2 in basic
solution can be considered in terms of the following equilibria
SO 2 + 0H HS0 3
HSO + OW so
with tacit recognition of the presence of other species such as M+ (sodium,
potassium, etc.) and S0 in the solution.
The equilibrium partial pressure of SO 2 in contact with such a solution is
given by 2
[ HSO ]
P E P = K (1)
S 2 [ SOs]
where K is the overall equilibrium constant.
While Equation 1 accurately depicts the relationship between the equilibrium
partial pressure of SO 2 and the concentrations of bisulfite and sulfite in
solution, these latter parameters are not normally easily measured, especially
with respect to the case of sorption on an alkaline filter. Therefore, follow-
ing Johnstone’s approach, we choose to make the following substitutions:
S [ s ] = [ HSO] + [ SO ]
and, by requiring electroneutrality,
C [ M ] = [ OW] - [ Hf] + 2 [ S0 3 ] + [ HSO ] + 2 [ SO ].
Recognizing that, at equilibrium, [ OW] and [ H ] will be relatively small com-
pared to C and S. Equation 1 can be arranged to yield
( 2S/C - 1 + 2 [ S0 4 ]/C) 2
P=KC — (2)
(1 - S/C - 2 [ SO ]/C)
Equation 2 thus gives the partial pressure of SO 2 in terms of the concentration
of metal ions and the ratio of absorbed sulfur to metal ion, S/C. The term
involving the ratio of sulfate to metal ion is relatively insignificant for
most filter media inasmuch as blank sulfates are usually quite low. However,
22
-------
this term is significant for at least one of the common filter media, and is
retained here for the sake of completeness.
Equation 2 was quite satisfactory for the type of results obtained by
Johnstone, and was used by him in correlation of data for a number of different
kinds of alkaline solutions. However, the more precise work of Coutant and Levy
showed that this type of equation fitted SO 2 equilibrium data over a wider range
of variables if the C-term was replaced by C 213 . Presumably, this form of the
equation allows for variation in activity coefficients over a wider range of
concentration.
Equation 2 can be rearranged and solved for S/C, the molar ratio of absorbed
sulfur to metal ion in solution, with the result
= ( 4-q-8z) + (q 2 + Sq - l6zq) ’ 2 (3)
where q = P/KC 2 ” 3 and z = [ SO ]/C. The concentration of the solution on the
surface of the filter depends on the relative humidity; the filter will tend to
absorb water until the equilibrium partial pressure of water of the resulting
solution is equal to the partial pressure of water in the gas phase. For many
salt solutions, the relationship between the water partial pressure and salt
concentration is nearly linear, and the following expression is a useful
approximation:
AC = (1 - RH) (4)
where A is a characteristic constant for a particular salt, and RH is expressed
as a decimal fraction. Examination of the handbook tabulations of A and values
of K given by either Johnstone or Coutant and Levy indicates that, at SO 2 partial
pressures of the order of 102 ppb, q<
-------
l/2 1/3
S/A 1.0 + 0.708 1’2 1’3 1’2 . (6)
K ‘ (1-RH) ‘ (l—2z) ‘
Equation 6 thus indicates that the SO 2 sorption capacity of a filter should be
directly proportional to the alkalinity of the filter, and should vary directly
with the square root of the SO 2 partial pressure, and inversely with the cube
root of (1-RH). Furthermore, the temperature dependence is implied by K 2 ,
and a dependence on the blank sulfate content of the filter also is indicated.
As suggested previously, this dependence on blank sulfate is insignificant for
most filter media.
The nature of the base might be expected to influence the temperature co-
efficient of the equilibrium constant. However, Johnstone found that various
bases, including sodium, potassium, ammonia, and methylamine had similar
temperature coefficients. His value for the potassium system was verified by
the BCL work, with K being given as
K = exp [ 6.111 - 4232/T°K] (torr/mole/l00 moles H 2 0). (7)
Preliminary analysis of the data in Table 4 consisted of a statistical exam-
ination with reference to the form of Equation 6. Because of the general con-
sistency between temperature coefficients found by Johnstone and that found in
the earlier BCL work, the function given in Equation 7 was used to normalize
the experimental results to 25° C. Regression techniques were then used to
test the significance of the equation
S/A = a + b P 2 /(l-RH) 3 (1-2z) 2 . (8)
Similarly, examination was made using the square root RH function that would
result from Johnstone’s original formulation of the chemistry. Correlation
with the cube root function was slightly better, and an overall correlation
coefficient of 0.87 was determined. The results of the analysis also indicated
that the presence of the constant a was statistically significant, i.e., the
filters tend to absorb some minimum amount of SO 2 regardless of the exposure
conditions.
A more detailed analysis of the data was then conducted using the “least-
squares-cubic” procedure developed by York 12 . In applying this method,
weighting factors were derived primarily from the results of the alkalinity and
blank sulfate measurements which suggest uncertainties of ±7 percent in the
24
-------
left- and right-hand side variables in Equation 8. A plot of the data along
with the resulting least-squares values of a and b are shown in Figure 4.
These experimentally determined values of a and b can be compared with
corresponding values based on the Johnstone work and that of Coutant and Levy
on vapor pressures of SO 2 over potassium sulfite solutions. Using the handbook
value of A for sodium sulfate, and maintaining consistent units, the calculated
value of b is 7 x l0 . This calculated value is only slightly more than 2
removed from the experimental value obtained in this work. The theoretical
value of the intercept is obviously considerably different from the experimental
value. However, alkalinities used in reducing these data were derived by the
ASTM procedure involving hot leach of the filters. As noted in Table 2, these
‘high temperature” alkalinjtjes are considerably greater than alkalinities derived
from ambient temperature leaching of the filters. If it is assumed that the low
temperature value of the alkalinity more correctly represents the alkalinity
available for SO 2 sorption during ambient sampling, the corrected value of the
experimental intercept is 0.55 ± 0.13, and the theoretical value is only slightly
more than 2o removed from the experimental value. In view of the very long
extrapolation required between the previous work and the current results, this
“approximate agreement” seems quite good, and it seems reasonable to apply the
chemical model in discussing the sorption of SO 2 on the strongly basic filter
me di a.
Some caution is due in interpretation and application of the model at very
low SO 2 concentrations. Obviously, if no so 2 is present, there can be no SO 2
absorption. The model however implies that some finite minimum capacity for
SO 2 absorption exists even at near-zero SO 2 levels, and it is assumed that the
total °2 exposure is sufficient to satisfy that minimum capacity.
In reviewing °2 sorption data for other filter types, as shown in Table 5,
it is obvious that the chemical model presented above does not adequately repre-
sent the performances of the weakly basic filters. With these materials, the
absolute amount of SO 2 absorbed is much less than that absorbed by the strongly
basic filters, but still greatly exceeds that predicted from consideration of
the alkalinities. In these cases, it may be that a small amount of water vapor
is absorbed by capillary condensation and that this water serves as a sorbent
for the SO 2 . However, direct evidence for such a phenomenon is not seen in the
current work.
25
-------
1.3
MSA 1 106-BH
I.’
Gelman
AE
0-
Gelman
I.0
AA
0.9
x
0.8
C
4)
a
>
£1
0.7
x
0.6
0.5
0.4
Slope =0.019 ± 0.005
Intercept0.25 ± 0.13
0.3
0.l
0
0 10 20 30 40 50 60
f (PRH,T,S0 4 ) [ Equation 8]
FIGURE 4. EFFECT ON FILTER
ON SO 2 SORPTION
AND ATMOSPHERIC VARIABLES
26
-------
Oxidation of Absorbed SO 2
The oxidation of sulfites has been considered in depth by a number of workers,
e.g., ei e, et al. The reaction is highly dependent on pH, with half-lives
as short as 4-5 minutes being indicated by Schroeter for solutions of sodium
sulfite under conditions where the rate of oxygen supply is not a limiting
factor. It is, therefore, expected that the rate of oxidation of SO 2 absorbed
on basic filter media is quite high, and that nearly all of the absorbed SO 2
would be oxidized within a normal 24—hour ambient sampling period. Workers
at Radian found this to be true, and renorted conversion of about 90 percent
of the absorbed SO 2 within a period of 2 hours. Sulfite contents of filter
samples collected in the present work were generally low relative to the total
sulfur contents of the samples, confirming the notion of rapid conversion to
sul fate.
Sulfate Sampling Error
It is obvious from the above discussion that, under any given set of atmo-
spheric sampling conditions, the absolute sulfate error per unit filter area is
fixed by the available alkalinity. However, ambient sulfate levels are usually
reported as mass per unit volume of air. Thus, the relative error in the mea-
sured sulfate loading is inversely proportional to the flow per unit area. For
practical purposes then Equations 7 and 8 can be combined and rewritten as
I 7 P 1 ! 2
SO 4 ( g/m ) = 0.25 + 1.57 x lO e’ wf l - 1’3 1/2 )
F L (l-RH) ‘ (l-2z) JJ
where F is the air volume per unit area (m 3 /cm 2 ) in a 24-hour sample and the
units of A are peq/cm 2 .
Table 11 shows a comparison of some measured excess sulfate levels with
corresponding values calculated from Equation 9. These are generally in good
agreement and indicate that the model can be useful in estimating excess levels
of sulfate due to °2 absorption.
Standard Hi-Vol procedures call for use of 24-hour sampling volumes of
3—4.5 m /cm 2 . Figure 5 shows predicted levels of sulfate sampling errors,
based on Equation 9, for these sampling rates and filter alkalinities of 0.1
and 0.3 .ieq/cm 2 . Thus in the range of normal sampling conditions, the expected
error due to SO 2 sorption on basic filters is between 0.3 and 3 pg/rn 3 depend-
ing on specific conditions and sampling variables. Excursions beyond this level
of error are easily conceivable however, and it is obvious that the lower the
27
-------
TABLE 11. EXPERIMENTAL AND PREDICTED EXCESS SULFATES
Run Filter
No. Comparison
3A Gelman AE ADL
3A MSA-ADL
lOB Gelman AA-QAST
(or ADL)
lOB Gelman AE-QAST
(or ADL)
ii MSA_QAST(d)
Sul fate
Difference, ug/m 3
Exp. ai.(b)
10.7±0.8 11.3
7.2±2 8.6
13.6 12.8
Corresponding hourly averages selected to yield highest value for
Equation 9.
Calculated by Equation 9.
Laboratory samples at F 1 m 3 /cm 2 .
Standard Hi-Vol sample at F = 3 m 3 /cm 2 .
502 (a)
ppb
RH,(a)
Percent
T,(a)
°K
90
80
286
90
80
286
125 98+ 298 16.3 17.9
135 88 298 2.1 3.0
(a)
(b)
(c)
(d)
28
-------
0 10 ao 30 40
6
5
4
‘0
E
0
C,)
Normal ambient Possible ambient
-s—-range — — excursions —
F 3, 4=0.3
F=4.5, 4=0.3
F=3, 4=0.1
F=4.5,A=0. I
I I I
2—
I —
1/2 1/3
P 2 /(1-RH)
FIGURE 5. SULFATE ERROR FOR BASIC FILTERS
29
-------
sampling rate, the more sensitive the error is to changing environmental con-
ditions.
CON CL US IONS
It is concluded that the principal error involved in sampling of ambient
air for total sulfate particulate content is caused by the absorption and sub-
sequent oxidation of ambient SO 2 in the presence of basic components of the
filter medium. A model of this absorption process has been developed based on
well-characterized chemistry of the sorption of SO 2 in alkaline solutions.
This model relates the extent of SO 2 sorption to environmental conditions,
including the relative humidity, temperatures and SO 2 concentration, and to
filter characteristics, such as alkalinity and blank sulfate content. An
effect due to the rate of sampling per unit area of filter is implicit in the
model. The results indicate that sulfate loading errors of the order of 0.3—3
g/m 3 can be expected with the use of common glass fiber filters, and that
larger sulfate errors are possible under extreme sampling conditions. Other
variables such as sampling time, storage conditions, and particulate composition
appear to have no significant effect on the sulfate content of the filter catch.
30
-------
SECTION 3
FIELD SAMPLING PROGRAM
OBJECTIVE
The objective of the field sampling phase of this program was an attempt to
verify laboratory observations under the constraints of field sampling conditions.
Specifically, experiments included comparison of representative filter types,
responses of these filters to high SO 2 loads, and examination of the effects of
iron- and manganese-rich particulate matter on sulfate collection.
APPROACH
The concept of simultaneous sampling was maintained in the field through the
use of two standard hi-vol units joined via a common inlet duct. This modification
ensured that physically and chemically similar input to each of the hi-vol units
was maintained during the experiments. Stainless steel ductwork for the modified
inlet was derived from the laboratory rig, with mixing orifices included for
experiments involving spiking of the inlet air with SO 2 .
Instrumentation for auxillary measurements included a flame photometric SO 2
detector, an ozone chemiluminescence detector, and a chemiluminescence NO instru-
ment. Measurements of temperature and relative humidity were made using wet
and dry thermocouples mounted on the dual hi-vol unit. The instruments for gas-
eous species measurements were located with the BCL Mobile Air Sampling Van, a
self-contained air-conditioned (or heated) mobile laboratory. Although this
unit includes a generator for remote site operation, power for this program was
obtained through the cooperation of local state, and city officials to minimize
procedures needed for preserving the identity of the local air sample.
Siting
Arrangements for sampling sites used were made through the cooperation of Mr.
Doug Dean, Environmental Engineer, Ohio EPA; Mr. Robert Ramhoff, Director, tlahoning—
Trumbell Air Pollution Control Agency; Mr. Lou Newman, Park Manager, Pymatuning
31
-------
State Park, Ohio Department of Natural Resources; the mayor of the City of
Youngstown; the Youngstown Fire Department; and the Youngstown City Maintenance
Department.
Fire Station No. 5--
The first site used was located at Fire Station No.5 in Youngstown. This
site is routinely used for sampling of SO 2 , sulfates, and total particulate.
As indicated on the map shown as Figure 6, this station is located centrally
with respect to the power plant and steel mill complexes of northwestern
Youngstown. Historically, average SO 2 , sulfate, and particulate loadings have
been relatively high in this area. For example, the 1975 annual average sulfate
level was approximately 20 ig/m 3 .
Hi-vol sampling at this site is conducted on the roof of the fire station.
At the time of our visit to this site, sampling was being conducted on every
sixth day. However, during our stay, Mahoning-Trumbull Air Pollution Control
Agency agreed to perform several extra hi-vol runs using their filters. Data
from these runs is not available for this report, but will ultimately be avail-
able for comparison with our results. Because of space limitations at the
fire station, gaseous pollutants at this site were measured in the parking lot
of the Youngstown City Maintenance Department located across the street from
and approximately upwind of the fire station.
Pymatuning State Park--
During the second week of the field study, hi-vol samples were collected at
Pymatuning State Park, which is located approximately 35-40 miles northeast of
Youngstown. Particulate measurements made by the Ohio EPA at this site tradi-
tionally show very low particulate loadings. This site is located about 25 miles
from the nearest urban area.
Col umbus--
During the initial stages of setup and checkout of equipment and procedures
for use in the Youngstown area, several runs were made in the Columbus area, at
a site removed by several miles from our laboratory facilities. This was an
urban residential site obtained through the cooperation of Mr. G. William
Keigley of BCL.
32
-------
UNITED STATES
FIGURE 6 YOUNGSTOWN FIELD SAMPLING AREA
bEPARTMENT OF THE INTERIOR
33
-------
Experimental Plan
Four filter media were chosen for use in the field study. This are
(1) ADL quartz - This is a high-purity quartz fiber
filter developed by the Arthur D. Little Co.
Samples of this material were obtained through
the cooperation of ADL.
(2) Pallflex QAST - This high—purity quartz fiber
filter is similar in many respects to the ADL
quartz, but is available commercially.
(3) MSA llO6-BH - This is a glass fiber filter that
has been used extensively in both source and
ambient sampling.
(4) Gelrnan A - The Gelman A glass fiber filter is
similar to MSA 1l06-BH.
The ADL filter was used as the reference filter in ll experiments except those
involving SO 2 spiking.
Experiments consisted of three basic types:
(1) Filter comparisons — Each of the above filters was
used in one side of the dual hi-vol rig, with the
ADL filter serving as the reference material on
the other leg of the system. Comparisons were
made during both daytime and nighttime periods
for each of the filter combinations to take
advantage of normal temperature and humidity trends.
(2) Catalyst effects - Several samples of the ADL
filters were preloaded with either ferric oxide
or manganese dioxide. These were then run in
day and night experiments in comparison with
clean AOL filters.
(3) SO 2 effects - Each of the filter types was run
in comparison with another sample of the same
type, but with one leg of the dual hi-vol spiked
with additional SO 2 . Again, day and night runs
were made.
Experiments 1 and 2 were conducted at the Youngstown site, and all three types
of experiments were conducted at the Pymatuning site.
34
-------
Problem Areas
As with almost any set of experiments conducted outside of the laboratory,
several problems associated with uncontrolled variables arose during the field
study. The most serious of these problems was the weather. Several periods of
heavy rainfall occurred during the 2-week field program. While it was possible
to shield the apparatus sufficiently during a light rain to enable changing of
filters, such was not possible during a heavy rain. Thus, when over 3 inches
of rain fell in Youngstown during the afternoon and early evening of September
17, no experiments could be started during that period.
Although it had been expected that we might obtain a good comparison between
high temperature/low humidity daytime periods and low temperature/high humidity
nighttime conditions, relative humidities were generally high throughout the
sampling period. This was especially true at Pymatuning where fogs or mists
persisted well into the daylight hours most of the time.
Some instrumentation problems also were encountered that could not be
corrected readily in the field. An electronic failure in the N0 instrument
could not be corrected, except that some periodic measurements were managed
during the second week of the study. The ozone instrument also did not yield
satisfactory readings when operated in the continuous mode. Thus, only periodic
readings with this instrument were possible.
Analyses
All filter samples were analyzed for sulfate, sulfite, total sulfur, ammonium
ion, nitrate, and total nitrogen. The ADL filters also were analyzed for metal
ion content by OES in order to provide some characterization of the background
particulate. Results of the nitrogen species analyses are not cited in this
report, but will be included in future reports for EPA Contract No. 68-02-2213.
A corresponding exchange of sulfate data from the current Los Angeles field study
under the latter program has been agreed upon. Not all of the data from that
program are yet available. However, results of comparison Gelman AA and QAST
filters are given in Appendix C.
Sulfate analyses were performed using the Dionex ion chromatograph. Attempts
at sulfite analyses by this method have not yet proven successful in our laboratory.
Instead we chose to use the iodine titration procedure. This method of course
yields total oxidizable content and is not specific for sulfite. Total sulfur
values were obtained by micro-combustion.
35
-------
RESULTS
Complete data sheets for the Columbus, Youngstown, and Pymatuning field work
are given in Appendix B. These data are condensed in Table 11.
In general, the results shown in Table 11 are less than ideal; sulfite and
sulfate values are consistently less than the corresponding total sulfur values
(if expressed as sulfate). Apparent scatter in the data is much greater than
that experienced in the laboratory measurements, and the apparent effect of the
consistently wet weather is evident in the very low sulfate and particulate
values on some days. Ffowever, some conclusions based on these data are possible.
DISCUSS ION
The particulate loadings in this series of experiments range from a low of
about 5 pg/rn 3 to a high of over 300 pg/rn 3 . The former being observed on a rainy
day at Pymatuning, and the latter being observed in Youngstown on a day when the
wind was directly from a steel mill stack that appeared to be operating without
particulate emission controls.
Particulate compositions at the Pymaturiing and Youngstown sites are appreciably
different. Those at Pymatuning represent essentially background particulate
matter, while the Youngstown data show a typical steel mill emission type of
particulate composition.
Inasmuch as the iodine titration procedure used for sulfite should have
measured essentially all oxidizable sulfur in the filter catches, it is disturb-
ing that the sulfate values are very much lower than the total sulfur levels.
This would appear to be due to an error in the sulfate analysis procedure used
for the field samples. Further evidence for this apparent deficiency is seen if
a comparison is made between the calculated sulfur error (Equation 9) and the
differences in sulfate, sulfite, and total sulfur levels between pairs of filters
shown in Table 11 . An attempt at correlation between calculated sulfur error
nd measured sulfite or sulfate yields correlation coefficients of 0.35 and
‘.20 respectively. On the other hand, as shown in Figure 7, a correlation cc-
fficient of 0.89 is obtained with the total sulfur values.
The data from Table 12 are further condensed and reorganized for purposes
of statistical analysis in Table 13. In this reorganization the data for the
sulfite measurements have been omitted. Table 14 shows a corresponding summary
of the mean differences in measurements and the statistical significance of the
36
-------
w
10
9
8
7
6
5
4
3
2
1
o
w
-C
4)
U
(t
U
C)
x
x
x
x
x
x
x
x
x
-2 -1
r = 0.89
1 2 3 4 5 6 7 8 9 10 11 12 13 14
S (expressed as sulfate)
FIGURE 7. CORRELATION BETWEEN CALCULATED AND MEASURED SULFUR DIFFERENCES
15
-------
TABLE 12. FIELD SAMPLE SUMMARY
Run
Filter No.
- -
SO ,
pg/rn 3
SO 3 ,
pg/rn 3
S,
pg/rn 3
Relatively High
Particulate, Particulate Metals
pg/rn 3 Fe Mg Pb K Ca Zn
x x
x
x
x
x
x x x x x
xx x x x x
xx x x x x x
/kDL
V-la
5.8
<0.5
7.28
129.6
X
X
Gelman A
V—la
13.7
0.98
11.2
129.9
AOL
V—lb
10.1
<0.2
6.3
165.2
XX
X
Geirnan A
V-lb
13.2
0,4
8.1
159.9
AOL
V-ic
11.39
<0.2
8.29
175.4
X
X
QAST
V-id
11.73
<0.2
7.76
170.0
ADL
Y-ld
2.1
<0.2
2.5
51.5
X
X
QAST
V—id
3.1
1.5
2.8
47.9
ADL
V-ie
0.35
<0.4
1.3
23.9
X
MSA
V-ie
0.22
<0.4
3.1
30.4
ADL
V-if
9.53
1.0
1.7
27.6
X
MSA
V-if
6.68
0.4
3.88
31.4
ADL
Y-ig
3.35
0.4
3.02
55.0
ADL + Mn0 2
ADL
V-ig
V-lh
1.91
5.54
0.2
1.44
3.95
5.77
60.0
188.7
ADL + Mn0 2
ADL
Y—lh
V-li
7.50
6.75
<0.2
0.5
7.40
6.41
175.7
331.0
AOL + Fe 2 0 3
V-li
5.48
<0.2
6.78
273.7
ADL
P-la
5.42
0.4
6.71
56.4
Geirnan A
P-ia
5.14
<0.4
6.71
45.9
ADL
P-lb
4.75
2.4
4.64
41.8
Gelman A
P-lb
5.64
1.0
7.44
39.2
ADL
P-ic
0.1
0.45
1.13
11.8
QAST
P-ic
0.88
1.0
0.9
11.6
ADL
P—id
0.33
1.01
0.68
13.9
QAST
P-id
0.20
0.45
0.90
11.9
AOL
P-ie
(-0.3)
1.5
0.34
6.1
MSA
P-le
1.88
1.0
4.4
9.3
ADL
P-if
0.0
0.45
0.56
15.4
MSA
P-if
4.69
1.90
5.28
18.5
-------
TABLE 12. CONTINUED
Run
Filter No.
SO 4 ,
pg/rn 3
SO 3 ,
g/rn 3
S,
pg/ni 3
Relatively High
Particulate, Particulate Metals
g/rn 3 Fe Mg Pb K Ca Zn
AOL P-lg 1.09 0.4 1.76 28.2
AOL + Mn0 2 P-i9 1.86 0.4 2.20 29.1
ADL P-lh 1.02 1.49 1.60 30.3
ADL + MnO., P-ih 2.05 0.45 2.29 30.9
AOL P-li (—0.1) 1.44 0.33 12.1
AOL + Fe 2 0 3 P-li (—0.2) 1.01 0.56 12.3
ADL P-lj (—0.35) 0.45 0.00 5.5
AOL + Fe 2 0 3 P-1j (-0.35) 0.45 0.11 4.8
MSA P-2a (-1.1) <0.5 2.86 17.1
MSA P-2a 0.0 1.0 2.06 14.6
Gelman A P-2b 4.0 1.93 4.08 36.0
Gelman A P-2b 2.89 2.40 3.65 33.3
QAST P-2c 2.11 2.36 2.70 34.2
Q1 ST P-2c 1.44 2.36 1.80 30.7
ADL P-2d 2.24 1.92 3.16 34.1
AOL P-2d 2.35 1.91 3.71 33.8
ADL C-l 0.66 0.11 1.46 54.9
MSA C-i 0.57 0.55 3.74 56.4
AOL C-2 3.50 1.13 3.96 83.0
MSA C-2 3.52 0.56 5.07 87.1
QAST C-3 10.0 — — 101.6
MSA C-3 7.86 75.2
-------
CD
TABLE 13. FIELD SAMPLE DATA
Variable
Site
Comparison
Group
AOL
vs
Geirnan
A
AOL
vs
QAST
AOL
vs
MSA
AOL
vs
AOL +
Mn0 2
AOL
vs
AOL +
Fe 2 0 3
so 4 , iig/rn 3
Youngstown
5.80
10.10
13.70
13.20
11.39
2.10
11.73
3.10
0.35
9.53
0.22
6.68
3.35
5.54
1.91
7.50
5.75
5.48
Pymatuning
5.42
4.75
5.14
5.64
0.10
0.33
0.88
0.20
-0.30
0.00
1.88
4.69
1.09
1.02
1.86
2.05
-0.10
-0.35
-0.20
-0.35
Columbus
0.66
3.50
0.57
3.52
S, 1 19/rn 3
Youngstown
7.28
6.30
11.20
8.10
8.29
2.50
7.76
2.80
1.30
1.70
3.10
5.77
3.02
7.40
3.95
6.41
6.78
Pyrnatuning
6.71
4.64
6.71
7.44
1.13
0.68
0.90
0.90
0.34
0.56
4.40
5.28
1.76
1.60
2.20
2.29
0.33
0.00
0.56
0.11
Columbus
.
1.46
3.96
3.74
5.07
Particulates,
ug/m 3
Columbus
129.6
165.2
129.9
159.9
175.4
51.5
170.0
47.9
23.9
27.6
30.4
31.4
55.0
188.7
60.0
175.7
331.0
273.7
Pymatuning
56.4
41.7
45.9
39.2
11.8
13.9
11.6
11.9
6.1
15.4
9.3
18.5
28.2
30.3
29.1
30.9
12.1
5.5
12.3
4.8
Columbus
54.9
83.0
56.4
87.1
-------
TABLE 14. STATISTICAL COMPARISO OF ADL FILTER PERFORMANCE WITH OTHEF FILTERS
ADL Filter
Versus
ADL+
ADL+
Site
Variable
Gelman A
QAST
MSA
Mn0 2
Fe 2 0 3
All 5
Youngstown
Number of Pairs
SO 4 , pg/rn 3
S, pg/rn 3
Particulate, pg/ni 3
2
-5.50
-2.86
2.50
2
-0.67
0.12
4.50
2
1.49
1 • 99 (a)
-5.15
2
-0.26
-1.28
4.00
1
1.27
-0.37
57.30
9
-0.96
138 (b)
7.67
Pyrnatuning
Number of Pairs
so 4 ,
S, pg/rn 3
Particulate, pg/rn 3
2
-0.30
-1.40
6.55
2
-0.32
0.01
1.10
2
-3.44
-4.39
2
-0.90w
-0.57
-0.75
2
-0.05
-0.17
0.25
10
-0.98w
130 (a)
0.80
All Field
Number of Pairs
4
4
6
4
3
21
Samples
pg/ms
S, pg/rn 3
Particulate, pg/rn 3
-2.90
-2.l3
4.52
-0.50
0.60
280 (a)
-0.64
269 (b)
3 • 68 (b)
-0.58
-0.92
1.62
0.46
024 (a)
19.27
087 (a)
137 (b)
3.40
(a) Statistically significant at the l0 significance level (90 confidence level).
(b) Statistically significant at the 5% significance level (95% confidence level).
-------
data is presented. The data were analyzed using Student’s 1-test where the data
were considered as paired observations. That is, comparisonsare made between
the ADL reference filter and the test filter for each experiment. This pairing
reduces extraneous influences on the subsequent measurements, and results in a
more powerful test. The T-test tests the hypothesis that the mean difference is
zero.
Separate statistical tests were run for each comparison group in Table 12.
The data were tested for the two sites individually and collectively. In examin-
ing the mean differences for statistical significance, no strong patterns emerge
from consideration of the individual sites. At the Youngstown site, significantly
higher sulfur values were obtained with the MSA filter than with the ADL filter,
and this significance also was noted when the ADL filter was compared with the
aggregate of all of the other filters. At Pymatuning, the sulfate measurements
were significantly higher with the Mn0 2 -treated ADL filter, and this significance
is repeated in comparing the ADL filter with the aggregate of samples. Total
particulate measurements at Pymatuning were significantly higher for the MSA
filter than for the ADL.
When the data were combined across sites, the ADL filters consistently yielded
significantly smaller measures of sulfur than the other filter types, with the
exception of the QAST filter. This observation is the strongest pattern to emerge
from the data.
Because of the limited number of samples and variety of experiments conducted
on this program, the number of observations for a statistical test is quite small.
This, of course, limits the power of the test to detect significant trends. From
a statistical viewpoint, the trends suggested in Table 13 should be considered
indicative, but not necessarily conclusive.
CONCLUS IONS
The objective of the field program phase of this work was to attempt to
verify laboratory observations under the constraints of field operating con-
ditions. In a limited sense, this has been accomplished. The relatively clean
AOL and QAST quartz filters do yield lower sulfur levels than the more alkaline
glass fiber filters that display a tendency for sorption of SO 2 . Furthermore,
differences in apparent sulfur collections correlate well with values predicted
by the results of the laboratory measurements. The field data do suggest that
the presence of Mn0 2 and possibly Fe 2 0 3 can influence the apparent ambient
42
-------
sulfur loading. This is in contrast with the laboratory measurements indicating
no significant effect. This may be indicative of an interaction that is sensitive
to other parameters that have not been recognized. It would seem that this
question should be investigated further. In any case, it seems probable that
the filter itself would not be involved in that interaction.
43
-------
SECTION 4
RECOMMENDATIONS
The observations made on this program indicate that the most serious cause
of ambient sulfate error, absorption of S0 2 ’can be avoided through proper choice
of filter media. The best filter medium is one that is free of alkaline
impurities and has a low blank sulfate. From the standpoint of good analytical
practice, blank contents of all impurities should be as low as possible. J\dher—
ence to such a criterion would permit more accurate analysis of the general
composition of the filter catch. Several filters meet these general criteria:
ADL Microquartz, Pallflex QAST, Pallflex E 70/2075-W, Mitex, Fluoropore, and
Celotate. These are not the only criteria however for filter media that are to
be used in a routine fashion for remote (nonlaboratory) sampling, in an air
pollution surveillance network.
At very least, the following additional criteria need to be considered:
(1) The collection efficiency for submicron particles
should be 99÷ percent.
(2) The pressure-drop across the filter should not
change markedly with normal particulate loadings.
(3) Mechanical strength must be sufficient to enable
handling, mounting, demounting, and possibly fold-
ing of the filter without appreciable loss of filter
weight due to flaking, shedding, or adhesion to con-
tacted surfaces.
(4) Unit cost must not prohibit large volume usage.
To the best of our knowledge, all of the above filter media except iitex(15) can
be obtained in types that meet the first criterion. The second criterion is prob-
ably most important with respect to the use of the filter in conventional 24-hour
hi-vol sampling. With membrane filters such as the Mitex, Fluoropore, and Celo—
tate, pressure-drop changes tend to be high and nonlinear with accumulation of
large quantities of particulate material on the filter surface. Accurate
44
-------
definition of flow rate changes becomes difficult under these conditions, and
samples can become biased with respect to the early portion of the sampling
period. In special cases, where shorter sampling periods are employed, the
second criterion is not as important.
The third criterion deals with a problem normally inherent in fibrous filters
such as the glass and quartz materials. In general, these filters do tend to
shed fiber fragments especially when the fiber length is very small.
The unit cost of the Teflon filters is very high compared to that of the
other filter media. Because of the high cost and pressure-drop characteristics
of these filters, their use is probably best suited to short-term sampling with
sampling devices using filter sizes smaller than the conventional 8 x 10 hi-vol
filters.
At this point, the quartz fiber filters would seem to be the best choices
for conventional hi—vol sampling. If and when other high purity silica filters
become commercially available, they also should be considered.
While the current work suggests that these filter media would be suit-
able for general hi-vol sampling of sulfates, additional data should be obtained
under extended field operational conditions. It is recommended that simultaneous
sampling with these filter media be conducted at selected air pollution monitoring
sites. These sites should be chosen with a view towards representing as wide
a range of typical ambient conditions as possible. Design of the experiment
should include statistical considerations to ensure adequate recognition of the
variables. Furthermore, inasmuch as sampling methods other than the conventional
hi-vol system are being used by EPA personnel for special purpose sampling, it
seems advisable to include the use of those sampling systems in the overall
program. For example, the dichotomous sampler and the tape sampler used for
aerial sulfate sampling should be included in this field methodology evaluation
program. With the current program serving as a base, such an extended-term
field program would contribute significantly towards validation of ambient sulfate
monitoring practices. Of course, other pollutants, such as, nitrates should be
included in the overall scheme.
45
-------
REFERENCES
1. Frank, N.H., Temporal and Spatial Relationships of Sulfates, Total Suspended
Particles, and Sulfur Dioxide. In: Proceedings of the 67th Annual Meeting
of the Air Pollution Control Association, Denver, June, 1974.
2. Lee, R.E., Jr. and Wagman, 3., A Sampling Anomaly in the Determination of
Atmospheric Sulfate Concentration, Am. md. Hyg. 3., 27:266, 1966.
3. Radian Corporation Final Report to Electric Power Research Institute,
Sulfur Dioxide Interferences in the Measurement of Ambient Particulate
Sulfate, January, 1976 (EPRI 262).
4. Pierson, W.R., Harnmerle, R.H., and Brachaczek, W., Sulfate Formed by
Interaction of Sulfur Dioxide with Filters and Aerosol Deposits, Anal.
Chem., 48:1808, 1976.
5. Novakov, 1., Chang, S.G., and Harker, A.B., Sulfates as Pollution Particulates:
Catalytic Formation on Carbon (Soot) Particles, Science, 186:259, 1974.
6. Chun, K.C. and Quon, J.E., Capacity of Ferric Oxide Particles to Oxidize
Sulfur Dioxide in Air, ES&T, 7:532-538, 1973.
7. Cheng, R.T., Corn, M., and Frohliger, B.L., Contribution to the Reaction Kinetics
of Water Soluble Aerosols and SO 2 in Air at ppm Concentrations, Atmos.
Environ., 5:987-1008, 1971.
8. Urone, P., Lutspe, H., Noyes, C.M., and Parker, J.E., Static Studies of
Sulfur Dioxide Reactions in Air, ES&T 2:611-618, 1968.
9. Freiberg, 3., Effects of Relative Humidity and Temperature on Iron-Catalyzed
Oxidation of SO 2 in Atmospheric Aerosols, ES&T, 8:731-734, 1974.
10. Johnstone, H.F., Recovery of Sulfur Dioxide From Waste Gases-Equilibrium
Particle Pressures Over Solutions of the Ammonia-Sulfur Dioxide-Water
System, md. Eng. Chem., 27:587, 1935; ibid 30:101, 1938.
11. Coutant, R.W. and Levy, A., Study of Phase Equilibrium in Selected SO?
Sorbent Systems, unpublished report, data available on request.
46
-------
12. York, D., Least-Squares Fitting of a Straight Line, Can. 3. Phys., 44:1079,
1 966.
13. Beilke, S., Lamb, D., and Muller, 3., On the Uncatalyzed Oxidation of
Atmospheric °2 by Oxygen in Aqueous Systems, Atmos. Environ., 9:1083, 1975.
14. Schroeter, L.C., 3. Pharm. Sd., 52:559, 1963.
15. Liu, B.Y.N. and Lee, K.W., Efficiency of Membrane and Nuclepore Filters for
Submicrometer Aerosols, ES&T, 10:345-350, 1976.
47
-------
APPENDIX A
GAS COMPOSITION DATA
Tables A-i through A-9 show hourly averages of various ambient gaseous
pollutant levels for selected 24-hour laboratory experiments.
48
-------
TABLE A-i. GAS COMPOSITION DATA(a)
RUN 2B
24-hour
average
240 214±14 80±14
27 47 75 298
66
ppb unless otherwise noted.
Variables other than SO 2 are the same in both branches.
Left and right system branches.
Hour
SO 2
L(C) R(C) NO NO 2
T
I,
° RH,%
NO
1
140
199
47
3
32
35
302
51
2
163
219
58
14
44
58
303
50
3
186
213
63
4
39
43
303
51
4
225
212
72
7
37
44
302
55
5
252
201
86
5
46
51
301
59
6
270
209
89
4
47
51
301
64
7
325
213
95
5
47
52
300
69
8
338
208
95
5
55
61
300
69
9
362
206
95
4
30
34
292
69
10
359
210
92
5
29
34
291
68
11
348
205
91
5
24
29
291
76
12
348
203
91
2
24
26
291
76
13
339
205
93
5
32
37
291
78
14
323
203
93
7
53
60
291
85
15
316
206
89
17
59
76
291
87
16
289
206
89
23
66
89
290
81
17
176
203
68
103
74
177
290
77
18
121
209
60
225
95
330
290
73
19
110
245
78
160
75
235
289
66
20
116
250
83
32
68
100
290
60
21
136
236
85
9
45
54
290
57
22
147
232
72
6
37
43
291
54
23
170
225
69
2
37
39
291
53
24
180
232
66
2
29
31
291
52
(a)
(b)
(c)
49
-------
TABLE A-2. GAS
RUN
COMPOSITION DATA(a)
3A
Hour 03 SO 2 NO NO 2 NO T, °K RH, %
1
88.0
86.1
73.6
2.0
77.6
290.0
55.6
2
115.0
83.9
84.5
3.0
87.5
290.0
55.7
3
130.0
92.4
92.0
3.5
95.5
290.3
55.9
4
155.0
93.6
100.5
5.5
106.0
290.5
55.9
5
157.5
97.0
106.0
7.3
113.3
289.9
59.1
6
160.0
95.9
110.5
7.0
117.5
289.3
62.3
7
160.0
94.8
119.0
5.3
124.3
289.0
63.2
8
160.0
95.9
125.0
6.5
131.5
288.6
65.5
9
157.5
91.2
128.5
6.0
134.5
287.6
70.2
10
150.0
91.2
135.0
5.0
140.0
287.0
72.0
11
147.5
88.9
140.0
7.3
147.3
286.9
72.3
12
150.0
88.8
147.0
7.5
154.5
286.5
74.8
13
147.5
90.1
153.5
7.5
161.0
286.5
73.9
14
140.0
91.2
159.0
7.5
166.5
286.0
77.7
15
137.5
90.1
162.5
8.3
170.8
286.0
77.9
16
130.0
86.4
167.0
8.5
175.5
285.8
80.1
17
120.0
85.1
173.0
6.5
179.5
285.6
81.8
18
110.0
86.4
180.0
10.5
190.5
286.1
74.8
19
110.0
78.5
182.0
7.5
189.5
286.6
74.9
20
115.0
83.9
184.5
8.0
192.5
287.4
69.6
21
135.0
86.4
187.5
8.5
196.0
288.5
64.2
22
140.0
86.4
193.5
9.0
202.5
289.4
61.2
23
165.0
87.6
196.3
8.3
204.5
289.7
60.2
24
175.0
88.9
197.5
9.5
207.0
280.1
58.4
138.5
89.1
143.8
6.7
150.5
288.1
67.3
24—hour
average
(a) ppb unless otherwise noted.
50
-------
TABLE A-3. GAS COMPOSITION DATA(a)
RUN 5A
Hour 03 SO 2 NO NO 2 NO T, °K RH, %
1 0 60 22 11 3 292 30
2 0 65 11 15 24 295 26
3 15 66 5 12 16 296 26
4 80 69 7 11 17 297 23
5 80 76 13 13 23 297 23
6 80 68 11 19 29 297 22
7 65 60 18 25 44 297 22
8 80 71 53 20 74 297 22
9 95 70 75 17 97 296 24
10 97 81 184 17 204 296 26
11 102 78 289 14 315 294 26
12 102 78 275 15 298 293 28
13 100 81 274 24 309 293 26
14 107 89 223 20 254 292 30
15 105 84 233 23 265 291 29
16 117 84 308 18 340 290 34
17 120 86 300 17 328 289 33
18 122 85 265 25 287 290 32
19 120 84 252 15 276 289 36
20 105 78 240 16 264 291 30
21 100 78 311 16 338 290 36
22 95 81 352 17 387 292 30
23 65 86 175 25 207 292 31
24 85 85 48 25 74 296 25
87 77 165 18 188 293 28
(a) ppb unless otherwise noted.
(b) 10:30 a.rn.
51
-------
TABLE A-4. GAS COMPOSITION DATA(a)
RUN 5B
Hour 03 °2 NO NO 2 NO I, °K RH, %
1 (b) 20 63 15 19 36 299 20
2 40 59 13 17 37 299 19
3 25 60 14 17 32 298 21
4 47 86 39 24 65 297 21
5 85 63 34 23 60 297 22
6 72 89 72 21 102 297 23
7 100 93 126 17 149 295 24
8 90 105 64 21 88 295 26
9 115 144 84 26 115 294 27
10 132 154 77 16 98 294 28
11 125 93 29 19 52 294 27
12 112 156 30 18 51 293 29
13 170 124 19 17 37 291 35
14 150 122 16 16 34 291 34
15 160 125 21 17 39 291 34
16 132 123 36 15 53 292 31
17 135 89 52 15 64 292 34
18 42 64 63 15 84 292 33
19 0 147 40 18 61 291 33
20 95 225 31 15 51 291 35
21 80 202 32 18 53 292 78
22 80 144 22 12 33 293 73
23 80 382 32 18 53 292 75
24 82 211 27 17 45 291 46
90 130 41 18 62 294 34
(a) ppb unless otherwise noted.
(b) 11:00 a.m.
52
-------
TABLE A-5. GAS COMPOSITION DATA(a)
RUN 6A
Hour 03 SO 2 NO NO 2 NO I, °K RH, %
1 (b) 55 86 10 9 13 279 36
2 77 96 11 4 13 279 38
3 90 86 8 6 11 279 40
4 102 84 9 12 12 279 39
5 100 96 8 13 11 280 31
6 100 91 9 15 13 281 29
7 95 84 9 13 13 281 29
8 105 86 9 13 14 280 28
9 120 89 13 15 17 279 29
10 132 88 8 15 13 279 29
11 125 91 9 14 13 279 29
12 127 89 7 10 10 279 28
13 120 89 8 10 10 279 30
14 120 88 7 9 10 279 30
15 120 84 7 8 10 278 30
16 120 85 7 8 9 278 31
17 115 84 7 5 9 278 31
18 122 85 7 6 9 277 33
19 120 86 6 4 8 277 34
20 120 85 7 5 8 277 32
21 120 78 7 7 9 277 34
22 117 86 9 13 12 277 33
23 90 81 14 12 20 277 33
24 87 81 21 16 24 278 32
110 87 9 10 12 279 32
(a) ppb unless otherwise noted.
(b) 2:30 p.m.
53
-------
TABLE A 6. GAS COMPOSITION DATA(a)
RUN 7
Hour 03 SO 2 NO NO 2 NO T, °K RH, %
(a) ppb unless otherwise noted.
1
72.5
66.5
6.3
16.8
17.5
288.4
81.4
2
90.0
36.2
0
9.0
6.0
289.6
76.3
3
110.0
43.9
0
9.8
6.5
289.9
78.8
4
130.0
50.9
0
5.5
0
290.3
74.5
5
135.0
55.4
0
6.0
0
290.1
78.3
6
145.0
59.1
0
4.0
0
288.6
79.2
7
137.5
59.1
0
10.8
8.0
286.8
84.3
8
140.0
51.7
0
5.5
1.5
286.8
81.8
9
140.0
53.3
0
8.3
3.3
285.6
79,1
10
140.0
55.6
0
10.5
6.0
284.4
76.2
11
142.5
54.3
0
12.5
6.5
283.5
75.5
12
140.0
63.0
0
7.5
2.5
282.9
72.2
13
137.5
59.5
0
4.5
0.5
282.5
68.3
14
125.0
63.0
0
4.0
0
281.9
77.1
15
130.0
57.6
0
4.5
0
282.4
73.1
16
125.0
51.1
0
2.0
0
282.1
72.9
17
102.5
66.3
0
6.5
0.5
281.3
72.1
18
105.0
63.8
0
8.5
2.5
280.5
74.4
19
90.0
62.9
0
11.5
10.8
280.3
74.2
20
105.0
63.1
4.0
11.0
16.5
280.1
75.5
21
100.0
84.8
24.3
13.0
38.5
280.3
74.9
22
65.0
100.1
59.5
23.0
89.0
280.5
75.9
23
60.0
115.8
28.8
11.5
44.5
281.1
75.7
24
60.0
44.8
15.5
6.0
24.0
283.0
75.2
12.3
284.3
76.2
24-hour
average
113.5 62.9 5.6 9.1
54
-------
TABLE A-7. GAS COMPOSITION DATA(a)
(a) ppb unless otherwise noted.
RUN 8
Hour 03 SO 2 NO NO 2 NO 1, °K RH, %
1
157.5
55.6
8.3
7.5
16.0
289.9
64.5
2
185.0
69.6
7.5
6.0
14.0
290.9
65.4
3
227.5
67.8
7.5
13.8
23.0
291.4
64.8
4
245.0
81.2
9.0
15.0
25.5
291.5
64.9
5
245.0
79.7
6.8
18.0
26.3
290.3
68.6
6
245.0
78.5
7.0
16.0
24.0
289.3
68.2
7
262.5
79.8
7.0
17.8
27.3
288.3
69.5
8
260.0
69.6
10.5
23.5
36.5
287.6
69.0
9
247.5
77.8
8.5
21.3
32.0
287.3
71.0
10
240.0
69.6
9.5
23.0
35.5
286.4
73.9
11
247.5
72.7
10.6
21.5
34.0
285.7
75.3
12
240.0
75.6
4.0
18.5
24.5
285.5
73.2
13
245.0
77.0
3.8
15.0
20.8
284.8
75.9
14
240.0
93.2
4.0
15.0
20.5
284.5
81.8
15
235.0
236.1
9.3
19.0
32.0
236.1
82.2
16
225.0
230.7
14.5
23.0
40.0
284.0
80.1
17
227.5
199.5
17.5
25.5
43.5
284.1
79.5
18
235.0
178.5
27.0
23.0
52.5
284.3
76.1
19
117.5
262.7
57.8
24.8
88.0
284.1
79.5
20
10.0
105.2
36.5
15.0
61.5
285.1
76.8
21
30.0
78.3
16.8
14.8
33.3
287.3
74.7
22
40.0
78.3
14.0
14.0
35.0
286.3
75.8
23
55.0
80.3
10.5
9.5
24.5
290.3
71.2
24
65.0
83.7
9.0
8.0
18.0
290.6
72.6
24-hour
average
189.0 110.0
13.4 17.1 33.0
287.3 73.1
55
-------
TABLE A-8.
GAS COMPOSITION DATACa)
RUN lOB
24-hour
average
219.4 106.1
17.4 299.2
96.2
(a) ppb unless otherwise noted.
(b) Questionable data.
Hour
03
SO 2
NO(t NO 2 )
N0
T °V flU 0’
, I’ F\i1 /0
1
42.0
60.7
22.3
301.0
92.1
2
65.0
69.6
15.5
299.5
97.1
3
97.5
79.8
10.8
299.6
98.5
4
125.0
83.9
12.5
300.4
97.1
5
175.0
102.1
12.3
299.9
99.0
6
225.0
112.3
8.0
299.5
100.0
7
235.0
126.6
17.5
299.5
99.0
8
270.0
114.4
2.5
296.3
100.0
9
287.5
119.9
8.3
296.9
100.0
10
290.0
119.9
15.5
296.9
100.0
11
300.0
120.8
22.5
297.0
100.0
12
300.0
118.1
9.0
296.8
100.0
13
305.0
119.9
5.3
297.3
97.5
14
320.0
123.4
7.5
297.5
98.0
15
327.5
125.1
3.8
297.5
95.9
16
335.0
126.9
7.5
297.4
94.9
17
315.0
128.2
38.3
296.6
100.0
18
230.0
95.9
50.0
296.5
99.0
19
165.0
92.0
49.8
297.8
98.0
20
160.0
148.5
42.0
299.8
92.3
21
180.0
92.4
15.5
301.9
90.0
22
200.0
98.1
13.0
304.0
86.1
23
217.5
99.2
11.3
304.6
87.4
24
215.0
95.9
8.5
305.3
56
-------
24- hour
average
TABLE A-9.
353.2 107.8
GAS COMPOSITION DATA(a)
RUN hA
40.4 300.5
86. 3
(a) ppb unless otherwise noted.
(b) Questionable data.
Hour 03 SO 2 NO NO 2 (b) N0 — 1, °K RH, %
1
195.0
85.6
24.8
304.4
79.3
2
230.0
86.4
36.5
304.9
82.2
3
250.0
87.6
32.0
303.3
84.1
4
285.0
104.3
28.0
305.3
80.7
5
345.0
110.6
24.8
301.3
91.4
6
355.0
112.5
38.0
303.5
84.3
7
375.0
115.3
45.8
302.7
84.4
8
400.0
119.9
49.5
302.1
86.4
9
427.5
125.1
37.0
301.3
86.7
10
445.0
125.2
48.0
300.5
86.9
11
450.0
129.4
51.0
300.2
86.8
12
470.0
131.9
86.0
299.9
89.5
13
507.5
135.1
28.3
298.8
88.8
14
520.0
135.1
14.5
298.8
88.3
15
525.0
142.5
26.0
298.5
88.3
16
535.0
136.6
28.0
298.1
89.2
17
415.0
108.3
54.0
297.6
87.1
18
290.0
88.9
57.0
298.4
86.4
19
245.0
88.9
83.0
298.3
87.8
20
250.0
88.9
25.5
297.6
89.0
21
240.0
85.1
39.0
297.5
87.6
22
240.0
83.9
71.0
298.4
88.3
23
247.5
85.1
30.3
300.3
84.5
57
-------
APPENDIX B
FIELD SAMPLING DATA
Table B-i shows particulate compositions for samples collected on the
field program. In most cases, only the AOL numbers of the filter pair was
analyzed. However, in experiments P-2a and P-2b where the ADL filter was not
employed, the analyses were performed on the Gelman A and MSA-l106-BH filters.
Data for those cases are not corrected for filter blanks.
Tables B-2, etc., show sampling conditions for the various field experi-
ments.
58
-------
TI\bLL -1. ELEME!JAL ANALYSIS OF FIELD SAMPLES, !EIGHT PERCENT
V-i aA
V-i bA
Y-lcA
V-idA
V-i eA
Y- lfA
Y- lgA
Y-lhA
V-i iA
P-1aP
P- ibA
P- lcA
P-idA
P-leA
P-lfP .
P—lqA
P—lhA
P—i iA
P-li A
P-2aA
P-2bA
P—2cA
P—2dA
.003 .01
.003 .01
.003 .01
.002 .01
.002 .01
.003 .01
.003 .01
.003 .01
.005 .01
.002 .01
.003 .01
.002 .01
.002 .01
.002 .01
.002 .01
.002 .01
.002 .01
.003 .01
.002 .01
.005 .003
.005 .005
<.002 .003
.003 .01
<.01 .003
.02 .003
.04 .003
.02 .003
.01 .003
.02 .003
.01 .003
.02 .003
.02 .003
<.01 .003
<.01 .003
<.01 .003
<.01 .003
<.01 .003
<.01 .003
.01 .003
<.01 .003
<.01 .003
<.01 .003
<.01 .005
<.01 .005
<.01 <.003
<.01 <.003
Sample Al B Ba Ca Cr Cu Fe K Mg Mn Mo Na Ni Pb Sr Ti Zn Zr
.1 .01 <.005 .05 .002 .001 .1 .02 .05 .005 <.001 .1 <.001 .02
.2 .01 <.005 .1 .095 .003 .4 .05 .05 .02 <.001 .1 <.001 .05
.2 .01 <.005 .1 .005 .003 .2 .05 .05 .02 <.001 .1 <.001 .05
.03 .01 <.005 .04 .002 .005 .07 .02 .03 .005 <.001 .1 <.001 .02
.01 .01 <.005 .005 <.002 .003 .03 <.02 .03 .001 <.001 .1 <.001 .01
.1 .01 <.005 .05 <.002 .005 .04 .02 .02 .002 <.001 .1 .001 .01
.1 .01 <.005 .1 .003 .005 .1 .05 .05 .01 <.001 .1 .002 .05
.2 .01 <.005 .2 .002 .005 .7 .06 .1 .02 <.001 .1 <.001 .04
.2 .01 <.005 .4 .003 .005 1.0 .1 .1 .03 .001 .1 <.001 .05
.02 .01 <.005 .03 .002 .003 .05 .02 .02 .002 <.001 .1 <.001 .01
.1 .01 <.005 .03 <.002 .002 .03 <.02 .02 .001 <.001 .1 <.001 .005
.1 .01 <.005 .03 <.002 .002 .02 <.02 .02 .001 <.001 .1 <.001 .003
.1 .01 <.005 .03 <.002 .002 .03 <.02 .02 .001 <.001 .1 <.001 .003
.03 .01 <.005 .03 <.002 .002 .03 <.02 .02 .001 <.001 .1 <.001 .003
.03 .01 <.005 .03 <.002 .001 .03 <.02 .02 .001 <.001 .1 <.001 .003
.1 .01 <.005 .04 .005 .003 .05 .02 .02 .002 <.001 .1 .002 .005
.1 .01 <.005 .03 .002 .003 .03 <.02 .02 .003 <.001 .1 <.001 .005
.1 .01 <.005 .05 <.002 .003 .02 <.02 .02 .001 <.001 .1 <.001 .005
.1 .01 <.005 .03 <.002 .003 .02 <.02 .02 .001 <.001 .1 <.001 .003
2.0 1.0 <.005 7.0 .003 .003 .03 1.0 1.0 .005 <.001 5.0 <.001 .02
2.0 1.0 .01 7.0 .003 .003 .05 1.0 1.0 .005 <.001 5.0 .001 .02
.01 <.002 .05 .04 .002 .005 .04 .02 .02 .005 .003 .1 .001 .01
.1 .01 <.005 .05 .002 .003 .03 <.02 .02 .002 <.001 .1 <.001 .005
-------
Table B-2.
Field Atmospheric Conditions
Flow Rates:
A 44.1 cfin
B 43.9 cfm
Filters:
A AOL
B Gelman A
Sample Time 5.50
Hourly Averages
Particulate
hr
Comments: Afternoon run, NO instrument not working, Ozone
instrument questi nable.
Site Voungstown OH
Date 9-14-76
Exp. No. Y-la
so 2 ,
Hour ppb T, °C RH, %
03, NO, NON,
ppb ppb ppb
Filter Loading, 1 g/m 3
so 4 so 3 S
60
-------
Table B-3. Field Atmospheric Conditions
Flow Rates:
A 44.9 cfm
B 44.8 cfm
Filters:
A AOL
B Gelman A
Sample Time
12.00 hr
Hourly Averages
So 2 , 03, NO, NOx,
Hour pp LL
i
94
26
44
2
79
26
48
3
112
26
48
.
4
126
26
50
5
133
26
54
6
139
26
55
7
149
26
54
8
147
26
55
9
142
26
57
10
157
26
56
11
153
. 26
56
12
137
26
58
Particulate
Filter Loading, g/m 3
Site Youngstown , OH
Date 9-14/15-76
Exp. No. Y-lb
SO 4 SO 3 S
Comments ; Night run
61
-------
Table B—4. Field Atmospheric Conditions
Site Youngstown , OH
Date 9-15-76
Exp. No. Y-lc
Flow Rates:
A 43.5 cfm
B 43.9 cfm
Filters:
A AOL
B QAST
Sample Time
11.75 hr
Hourly Averages
SO 2 , 03, NO, NOx,
Hour ppb i, °c RH, % ppb ppb ppb
== -
1 155
2
138
3
217
4
150
5
159
6
149
7
109
8
101
9
106
114
11
104
-
12
91
Filter Loading, pg/rn 3
Particulate
Comments;
62
-------
Table B-5. Field Atmospheric Conditions
Site Youngstown , OH
Date 9-15/16-76
Exp. No. Y-ld
Flow Rates:
A 43.8 cfm
B 44.3 cfm
Filters:
A ADL
B QAST
Sample Time 11 .87
Hourly Averages
hr
SO 2 ,
Hour ppb
03,
1, °C RH, % ppb
NO, NOx,
ppb ppb_—
i
67
20
72
2
62
20
74
48
20
82
25
19
82
<15
20
81
6
<15
20
78
<15
20
79
8
<15
19
85
<15
18
90
10
35
18
gi
11
48
•
19
91
12
50
19
91
Filter Loading, ig/m 3
Comments ; Night run, some rain
Particulate
SO 4 So 3 S
63
-------
Table B-6. Field Atmospheric Conditions
Flow Rates:
A 44.1 cfm
B 44.3 cfm
Filters:
A ADL
B MSA
Sample Time 7.30
Hourly Averages
hr
SO 2 , 03, NO, NOx,
Hour ppb 1, °C RH, % ppb ppb ppb
- = = == -
1 33 17 92
2
3
<15
17
95
<15
17
96
4
<15
17
96
<15
17
95
6
I
— <1 5
<15
96
97
—
—
8
—.
9
—
—_____
10
—
—
—____
11
—
—
12
—
—
Filter Loading, pg/rn 3
Site Youngstown , OH
Date 9-16-76
Exp No. Y-le —
5
Comments ; Rain
64
-------
Table B-7. Field Atmospheric Conditions
Site Youngstown , OH
Date 9-16/17/76
Exp. No. Y-lf
Flow Rates:
A 43.9 cfm
B 43.8 cfm
Filters:
A ADL
B MSA
Sample Time 12.08
Hourly Averages
SO 2 , 03, NO, NOx,
= = Ji 2 L
1 18 17 98 10
2
34
17
98
30
45
17
96
51 —
17
94
55
17
97
6
59
17
98
65
17
99
8
68
17
98
9
72
17
98
10
73
17
98
11
77
17
97
12
77
17
98
<10
Filter Loading, pg/rn 3
hr
Parti cul ate
SO 4 SO 3 S
Comments ; Night run -- Rain
65
-------
Table B-8. Field Atmospheric Conditions
Site Youngstown ,
Date 9-17-76
Exp. No. Y-1g
OH Flow Rates:
A 44.6 cfm
B 44.3 cfm
Fi 1 ters:
A ADL
B_ADL +_Mn0 2
Sample Time 11 .78 hr
Hourly Averages
SO 2 ,
Hour ppb 1, °C RH, %
1 52 17 .99
Filter Loading, ug/m 3
Particulate so 4 so 3 s
I T •°T
1 0.2 3.95
Comments:
Ra i n
03,
ppb
NO, NOx,
ppb ppb
<10
2
4
24 —
39
16
17
17
100 <10
100
99
- -
—
5
67
— 17
96
—
6
34
17
97
—
29
17
99
—
8
27
17
98 —________
36
17
96
—
—
10
40
18
94
11
17
95
—
12
16 98
—
66
-------
Table B-9. Field Atmospheric Conditions
Youngstown, OH
Site______________
Date 9-18/19—76
Exp. No. Y-lh
Flow Rates:
A 44.5 cfm
B 44.7 cfm
Filters:
A ADL
B ADL + Mn0 2
12
Sample Time 11.92 hr
Hourly Averages
97
S
Filter Loading, 1 i9/m 3
Particulate
S04
SO 3
S
• 188
.7
5 . 54 J 1.
44
5
.77
B 175
.7
7.50_L
<0.2
7
.40
Comments : Night run
SO 2 , 03, NO, NOR,
Hour ppb T, °C RH, % ppb ppb ppb
-
i 102 18 62 20
2
84
17
70
3
79
15
82
4 76
65
16
81
15
87
.
6
63
16
86
7
83
16
84
8
82
16
85
80
16
89
10
11
87
16
93
97
93
16
67
-------
Table B-1Q Field Atmospheric Conditions
Flow Rates:
A 43.9 cfm
B 43.8 cfm
Filters:
A ADL
B ADL + Fe 2 0 3
Sample Time
11.50 hr
Hourly Averages
SO 2 , 03, NO, NOx,
Hour ppb 1, °C RH, % ppb ppb ppb
70
15
91
2
66
15
90
3
77
17
86
.
4
89
19
84
5
20
82
6
87
21
66
98
23
59
8
26
39
118
25
38
10
121
24
37
11
129
. 22
41
12
12 Y
20
50
20
Particulate
Filter Loading, pg/rn 3
Site Youngstown , OH
Date 9-18-76
Exp. No. Y-li
SO 4 SO 3 S
Comments : Wind shift -- directly from dirtiest stack
68
-------
TableB-il. Field Atmospheric Conditions
Site Columbus, OH
Date
Exp. No. FC-l
Flow Rates:
A 44.6 cfm
B 43.8 cfm
Filters:
AMSA-l 106-BH
BADL
Sample Time 12
Hourly Averages
hr
SO 2 , 03, NO, NOx,
Hour — ppb T, RH, % ppb ppb ppb
i
35.0
14.0
78.8
2
44.6
13.4
83.8
3
46.2
13.0
85.0
4
45.5
12.6
86.6
5
48.7
12.4
87.8
•
6
42.5
12.7
85.4
7
40.4
12.1
88.8
8
12.3
87.5
—
26.6
12.3
90.6
10
18.3
12.7
87.7
11
(9.0)
. 13.2
87.2
12
14.8
79.4
Particul ate
Filter Loading, pg/rn 3
Comments ; Night run, no NO 2 or 03 measured
S04 503 S
69
-------
Table B—la Field Atmospheric Conditions
Site Columbus OH
Date
Exp. No. FC-2
Flow Rates:
A 43.5 cfm
B 43.3 cfm
Fi 1 ters:
A MSA
B ADL
Sample Time 12
Hourly Averages
SO 2 , 03, NO, NOx,
Hour ppb I, °C RH,% ppb ppb ppb
1
39.6
19.1
63.9
2
62.7
21.0
61.2
3
54.1
22.3
59.4
4
41.8
25.1
48.8
5
47.9
24.5
49.8
6
54.5
24.1
51.6
7
64.8
24.0
48.6
8
69.8
23.6
42.6
9
74.1
22.5
48.1
10
75.1
21.6
52.1
11
73.9
20.3
56.3
12
73.9
19.1
60.5
Filter Loading, ig/m 3
Comments ; Day run, no NO or 03
x
mea S ured
hr
Particu1 te SO 4 SO 3 S
A 87.1 3 . 52 J 0.56 5.07
B 83.0
3.50J 1.13 3.96
70
-------
Table B—13. Field Atmospheric Conditions
Site
Pymatuning
Flow
A
B
Rates:
44.5
44.5
cfm
cfm
Filters:
A ADL
Date
9-19/20-76
Exp.
No. P-la
BGelman A
Sample Time
12.42
Hourly Averages
SO 2 , 03, NO, NOx,
I ppb T,°C RH, % ppb ppb ppb
63
15
98
28
2
76
13
99
3
94
12
99
4
105
12
99
5
87
12
97
6
85
12
98
7
80
12
98
8
74
13
100
9
72
14
100
10
71
15
100
11
67
* 15
100
12
66
16_
99
30
Filter Loading, ug/m 3
Comments ; Light rain, NO instrument not working
hr
Parti cul ate
71
-------
Table B—14. Field Atmospheric Conditions
Site Pym tiining
Date 9/20/76
Exp. No. Mb
Flow Rates:
A 44.3 cfm
B 44 9 cfm
Filters;
A/kDL
BGelman A
Sample Time
11.45 hr
Hourly Averages
SO 2 , 03, NO, NOx,
Hour ppb_ 1, °C — RH, % ppb ppb ppb
32
14
98
15
2
54
14
100
12
3
63
15
100
33
4
71
15
100
38
5
73
15
97
43
.
6
75
16
95
37
72
17
91
35
8
70
16
95
28
9
76
16
91
29
10
80
16
92
28
ii -
90
100
15
15
— 95
100
8
25
Filter Loading, j g/m 3
Comments ; Light rain
Particulate
72
-------
Table B-15. Field Atmospheric Conditions
Si tepymatuni ng
Dateg- QJ2i -76
Exp. No. P-ic
Flow Rates:
A 44.5 cfm
B 44L.5 cfm
Filters:
A ADL
B QAST
Sample Time 11.75
Hourly Averages
hr
so 2 , 03, NO, NOx,
Hour ppb T, °C ppb ppbppb
1
75
14
100
25
2
87
13
100
10
3
96
13
100
4
103
13
100
5
108
13
100
6
107
12
100
96
12
100
8
88
12
100
10
g
83
12
100
8
10
79
— 12
100
9
11
74
- 11
99
14
12
69
9
100
10
Filter Loading, itg/m 3
Particulate
Comments ;
73
-------
Table B—16. Field Atmospheric Conditions
Site Pyniatuning
Date 9-21-76
Exp. No. P-id
Flow Rates:
A 449 cfm
B cfm
Sample Time - 11.62
Hourly Averages
hr
SO 2 , 03, NO, NOR,
Hour ppbT,°CRH ,%ppb ppb ppb
1
54
9
100
20
2
58
12
93
18
3
65
14
7 L
.
4
68
15
61
27
5
55
15
60
22
6
79
17
45
125
105
16
48
—
125
8
105
15
52
33
95
g
92
12
71
12
65
60
—a—-----
75
10
92
9
100
<5
11
•
13
12
98
8
96
23
Filter Loading, .ig/m 3
Comments ; Rain, NO instrument partly repaired
Filters:
A__ADL
B QAST
Particulate
A 1339
74
-------
Table B-17. Field Atmospheric Conditions
Site Pymatuning
Date 9-21/22-76
Exp. No. P—le
Flow Rates:
A 44.5 cfm
B 44.6 cfm
Filters:
A ADL
B MSA
Sample Time 11.77
Hourly Averages
hr
So 2 , 03, NO, NOR,
Hour ppb T,°C RH,% ppb ppb ppb
1
57
8
89
20
40
2
69
8
91
43
82
8
94
78
89
8
8
94
100
91
95
•
120
6
4
8
93
130
96
8
95
130
8
97
8
95
135
g
97
8
94
130
10
95
7
97
120
11
89
6
95
113
12
87
— 6
95
14
-
Filter Loading, hg/rn 3
Particul ate
Commen ts :
75
-------
Table B-18. Field Atmospheric Conditions
Site Pymatuning Flow Rates: Filters:
Date 9—22—76 A 44.6 cfni A ADL
Exp. No, P—if B 44.7 cfm B NSA
Sample Time 11.75 hr
Hourly Averages
SO 2 , 03, NO, N0 ,
Hour ppbT,°CR %ppb ppb ppb
1
75
3
95
15
35
70
2
3
83
90
6
8
86
78
70
4
86
10
65
20
45
66
5
26
ii
58
70
6
28
11
53
97
7
38
12
53
80
8
40
12
53
30
45
65
g
31
13
50
47
10
— 42
—
13
50
53
11
12
53 12 54
9 80 15
45
70
95
Filter Loading, i.ig/m 3
Particulate $04 SO 3 S
A 15.4 0.0
8 18.5
4.69
1.90 5.28
Comments ;
76
-------
Table B-19. Field Atmospheric Conditions
Site Pymatuning
Date 9—22/23—76
Exp. No. P—lg
Flow Rates:
A 44.2 cfm
B 44.2 cfm
Filters:
A ADL
BADL + Mn0 2
Sample Time 12.10
Hourly Averages
hr
03, NO, NOx,
Hour ppb —- T °C RH, % ppb ppb —
16
6
100
15
30
30
2
40
5
100
40
46
4
100
55
4
47
5
50
5
47
5
63
6
6
63
7
39
7
60
8
50
8
100
63
9
9
68
10
9
70
11
9
__________
70
Filter Loading, ig/m 3
Comments ; Rain
77
-------
Table B-20, Field Atmospheric Conditions
Pymatuning
9—23—76
No. P-lh
Flow Rates:
A 44.7 cfm
B 44.7 cfm
Filters:
A ADL
BADL + Mn0 2
Sample Time 11.50
Hourly Averages
hr
SO 2 , 03, NO, N0 ,
Hour ppb T, °C RH, % ppb ppb ppb
-
1 8 96 20 40 55
2 9 97
3
4
<15
<15
10
11 —
97
-
20
30
55
-
70
5
<15
12
12
12
90
75
6
16
90
-
25
40
75
7
26
89
90
8
40
13
80
67
9
48
12
35
45
65
lo_
37
13
93
—
40
. 12
96
65
12
40
11
100
30
35
65
Filter Loading, pg/rn 3
Si te
Date
Exp.
Particul ate
Comments ; Rain
78
-------
Table B-21. Field Atmospheric Conditions
Site Pymatuning Flow Rates: Filters:
Date 9—23/24—76 A 45.0 cfrn A ADL
Exp. No. p—ij B 44.7 cfm BADL + Fe 2 0 3
Sample Time 11.77 hr
Hourly Averages
SO 2 , 03, NO, NOx,
PPb T,°C RH,% ppb ppb
i
8
100
30
30
40
2
<15
6
100
65
18
5
100
75
4
18
4
100
80
5
17
3
100
60
6
<15
2
100
40
7
1
100
48
8
1
100
50
9
1
100
——
10 —
0
100
——
11
0
100
——
12
0
97
——
Filter Loading, g/m 3
Particulate
SO 4
SO 3
S
A
B
12.1
±2.3
j
-0.1
-0.2
1.44
1.01
0.33
0.57
Comments ; Pens stopped on N0 O 3
79
-------
Table B—21. Field Atmospheric Conditions
Site Pvmatuninp Flow Rates: Filters:
Date 9—24—76 A 44.3 cfm A ADL
Exp. No. P—lj B 44.5 cfm B ADL + Fe 2 0 3
Sample Time 11.75 hr
Hourly Averages
SO 2 , 03, NO, NOR,
Hour ppb — 1, O RH,% ppb ppb ppb
1
2
96
0
35
35
2
7
90
3
10
77
4
13
57
20
35
5
6
<15
14
15
51
48
21
28
7
44
14
48
44
8
53
13
54
— 30
40
40
9
29
13
57
30
10
17
12
51
25
11
12
53
-—
33
12
8
89
30
48
Filter Loading, ug/m 3
• Particulate SO 4 SO 3 S —
A —0.35 0.45 0.0
—
4.8
—0.35
0.45
0.11
Comments ;
80
-------
Table B-22. Field Atmospheric Conditions
Site Pymatuning
Date 9—24/25—76
Exp. No. P—2a
Flow Rates:
A 44.7 cfm
B 44 6 cfm
Sample Time
11.50 hr
Hourly Averages
SO 2 , 03, NO, NOx,
Hour ppb - T, °C RH, % ppb ppb ppb
i 97 6 86 30 53
107
3
100
50
3
107
2
100
35
4
111
2
100
35
5
106
1
100
37
6
104
0
100
44
7
107
0
100
51
8
114
0
100
42
g
114
0
100
39
10
111
0
100
36
11
12
117
109
—1 100
—1 100
40
43
40
Filter Loading, .ig/m 3
so 2 spike
Ambient SO 2 <20 ppb
Filters:
AMSA
BMSA
Comments ;
81
-------
TableB-23. Field Atmospheric Conditions
Site Pymatuning
Date 9—25—76
Exp. No. P—2b
Filters:
A Gelman A
B Gelman
so 2 , 03, — NO, NOx,
Hour J L ppb ppb
1
105
1
100
40
40
33
43
62
57
76
2
6
95
3
13
61
4
152
16
57
30
40
5
154
18
45
6
154
17
51
7
147
17
53
— 40
45
97
8
149
17
46
100
9
154
17
52
103
10
161
16
68
115
1]
143
. 13
84
115
12
127
11
97 I
35
100
Filter Loading, ug/m 3
Particulate SO 4 so 3
36.0
33.3
4.00
2.89
Comments; SO 2 spike
Ambient SO 2 30 ppb
A 44.1 cfm
B 43.8 cfm
Flow Rates:
Sample Time 11.77 hr
Hourly Averages
A
S
1.93
2.40
82
-------
Table B-24. Field Atmospheric Conditions
Si te
Date
Exp.
Pymatuning
9—2 5/26—76
No. P-2c
Flow Rates:
A 44.3 cfrn
B 44.1 cfm
Sample Time 11.80
Hourly Averages
hr
So 2 , 03, NO, NOx,
Hour ppb T, °C RH,% ppb ppb ppb
1
137
9
100
40
60
2
134
8
100
70
3
139
8
100
70
4
140
9
-
100
70
5
138
9
—
100
70
6
133
9
100
88
7
140
8
100
88
8
145
8
100
80
9
151
8
100
75
10
151
8
100
65
ii
154
. 9
ioo
55
12
153
9
100
30
55
Particul ate
Filter Loading, pg/rn 3
SO 4 SO 3
Comments ; so 2 spike
Light rain
Ambient SO 2 ‘ 60 ppb
Filters:
A QAST
B QAST
S
83
-------
Table B-25. Field Atmospheric Conditions
Site Pymatunjng
Date 9—26—76
Exp. Nop—2d
Flow Rates:
A 44.3 cfm
B 44.5 cfni
Filters;
A AIJL
B ADL
Sample Time 11.78
Hourly Averages
hr
SO 2 , 03,NO,N Ox,
Hour ppb T, °C RH, % ppb ppb ppb
— — Lrt = ====-=-- - = -=zr=
125 10 98 “ 10 30 34
2
132
11
96
44
3
142
12
95
41
4
147
12
98
30
5
141
12
97
“1O
49
6
150
13
99
53
7
141
13
100
72
8
141
14
100
10
70
106
134
14
100
85
10
137
14
100
75
1]
142
14
100
89
98
12
141
13
100
25
60
Filter Loading, pg/rn 3
Comments :
SO 2 spike
Rain
Ambient SO 2 ‘t 60 ppb
84
-------
APPENDIX C
LOS ANGELES FIELD SAMPLING DATA
As part of a cooperative effort between this work and that being conducted
under Contract Nos. 68-02-2439 and 68-02-2213, additional sulfate field sampling
data was collected in the Los Angeles area. In that program, approximately 250
filter samples were collected using various types of filters at three different
sampling sites. Inasmuch as the data from the Los Angeles work is dedicated
primarily for application on the above programs, analysis of that data has not
yet been completed. However, sulfate data for one site at Upland, California
(Ca. 30 miles east of Los Angeles) is presented in Table C-i. Corresponding
hourly average readings for temperature, relative humidity, and SO 2 are shown
in Tables C—2—C-6.
The comparatively small differences between the responses of the Gelman AA
and QAST filters shown in Table C-i are due primarily to two factors:
(1) The sample volumes used in the Los Angeles work were
of the order of 5.6 m 3 /cm 2 which is somewhat higher
than that employed in normal hi-vol sampling. As
noted on page 27, the sulfate error due to SO 2 absorp-
tion is inversely proportional to the flow per unit
area.
(2) SO 2 levels in the Los Angeles area are generally much
lower than those found in the Youngstown study, thus
further minimizing the contribution of SO 2 sorption
to sulfate error.
In general, these small differences and the fact that ambient conditions were
relatively constant make it difficult to attempt any detailed correlation of
measured sulfate levels with values estimated from Equation 9. However, it
can be seen from Table C-i that the average difference in filter response for
either sulfate or total sulfur compares well with the range calculated from
Equation 9.
85
-------
TABLE C-i. LOS ANGELES SAMPLING DATA
Run
Date Filter No.
S0
pg/rn 3
S0
pg/rn 3
S
pg/rn 3
10/18 Gelman AA LA-i 31.5 1.0 11.5
QAST LA-i 36.1 -0.5 12.7
10/19 Gelman AA LA-2 33.3 0 11.8
QAST LA-2 30.5 0.6 11.3
10/20 Geiman AA LA-3 18.0 0.6 7.2
QAST LA-3 9.2 0 3.8
10/21 Geirnan AA LA-5 18.3 0.6 6.7
QAST LA-5 18.3 1.4 6.2
10/22 Geirnan AA LA-6 12.0 0.6 4.9
QAST L/\-6 11.8 1.2 4.3
A50 4 = 1.55 jg/rn 3
t S = 2.3 5 pq/n1 3 (expressed as sulfate)
ASO 4 (Equation 9) 1.20.1 1 q/ui 3
86
-------
TABLE C-2. LOS ANGELES AMBIENT CONDITIONS
(HOURLY AVERAGES)
Date
Hour,
PST
T,°C
RH, %
SO 2 ,
ppb
10/18
9-10
10-11
11-12
12-1
1-2
2-3
3-4
4-5
5-6
17.1
18.0
19.2
20.6
22.4
22.8
22.0
20.3
18.0
87.9
81.8
75.5
68.0
61.0
57.3
58.0
61.2
73.8
-
-
-
-
—
-
-
-
(a) Instrument not working.
TABLE C-3. LOS ANGELES AMBIENT CONDITIONS
(HOURLY AVERAGES)
Date
Hour,
PSI
T, °C
RH, %
SO 2 ,
ppb
10/19
9-10
19.8
74.2
10-11
23.4
61.6
-
11-12
26.5
47.6
-
12-1
26.3
46.3
-
1-2
25.2
47.6
-
2-3
25.6
43.6
-
3-4
24.2
48.0
-
4-5
22.4
55.1
48.2
(a) Instrument not working.
87
-------
TABLE C-4. LOS ANGELES AMBIENT CONDITIONS
(HOURLY AVERAGES)
Date
Hour,
PST
1, °C
RH, %
SO 2 ,
ppb
10/20
9-10
20.1
63.3
74.7
10-11
11-12
12—1
1-2
2-3
3—4
4-5
23.7
25.3
24.5
22.5
24.6
25.8
24.4
49.5
42.5
44.6
56.4
42.3
34.2
37.1
79.1
60.5
57.4
49.5
51.7
51.0
50.0
TABLE C-5. LOS ANGELES AMBIENT CONDITIONS
(HOURLY AVERAGES)
Hour, SO?,
Date PST T, °C RH, % pp5
10/21 9-10 18.9 87.8 42.2
10-11 20.2 81.3 44.4
11-12 22.4 69.4 45.3
12-1 24.3 59.7 47.5
1-2 24.8 57.1 49.4
2—3 24.3 59.2 55.5
3-4 22.5 63.3 53.6
4-5 20.4 68.2 51.0
88
-------
TABLE C-6. LOS ANGELES AMBIENT CONDITIONS
(HOURLY AVERAGES)
Date
Hour,
I’ST
T, °C
RH, %
SO 2 ,
ppb
10/22
9-10
10-11
11-12
12-1
1-2
2-3
3-4
4-5
18.8
19.5
20.8
22.0
20.5
19.7
17.5
16.4
82.0
78.2
70.8
66.9
68.2
65.0
70.9
75.9
52.5
52.6
52.5
53.6
53.4
50.5
48.8
53.6
89
-------
APPENDIX D
FILTER IDENTIFICATIONS
TABLE D-l. FILTER IDENTIFICATIONS
Filter
Chemical Type
Manufacturer
MSA-fl06-BH
glass fiber
Mine Safety Applicance
Co.
Gelman AE
glass fiber
Gelman instrument Co.
Gelman AA
glass fiber
Gelman Instrument Co.
Spectrograde
ADL
treated glass fiber
quartz
Gelman Instrument Co.
Arthur 0. Little, Inc.
QAST
quartz
Paliflex Products Co.
E 70/2075-W
glass fiber-EVA binder
Paliflex Products Co.
Mitex
Teflon
Millipore Corp.
Fluoropore
Teflon/polyethylene
Millipore Corp.
Celotate
cellulose acetate
Millipore Corp.
Duralon
nylon
Milhipore Corp.
Polyimide
polyimide
Gelman Instrument Co.
(a) Obtained from R. Thompson, EPA.
(b) Similar material recently made available through Gelman Instrument
Co.
90
-------
TECHNICAL REPORT DATA
(Please read I uctions on the reverse before completing)
1. REPORT NO. 2.
EPA-600/2—77—076
3. RECIPIENT’S ACCESSIOINO.
4. TITLE AND SUBTITLE
FACTORS AFFECTING THE COLLECTION EFFICIENCY
OF ATMOSPHERIC SULFATE
5. REPORT DATE
May 1977
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Coutant, R. W.
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Battel le Col umbus Laboratories
Columbus, Ohio 43201
10. PROGRAM ELEMENT NO.
68-02-1784
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory - RTP, N. C.
Office of Research and Development
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
Fin 1 4/75 - 11/76
14.SPONSdRII 1GAGENCYcODE
EPA/600/09
15. SUPPLEMENTARY NOTES
lb. A S IFiA [
Factors that influence the collection and measurement of atmospheric sulfate
were investigated. Special emphasis was given to those factors that cause the
formation of extraneous sulfate during the sampling process. The factors considered
were filter type and composition, ambient SO 2 concentration, temperature, relative
humidity, ambient gas composition, sampling time and rate, storage time and conditions
and the presence of potential oxidation catalysts in the particular catch.
The approach of the investigation was twofold. After conducting an extensive
laboratory program to identify and quantify significant sulfate generating inter-
actions, a brief field study was performed to test the significance of the laboratory
observations under typical field operating conditions.
The results of this investigation indicated that the most significant cause
of sulfate sampling error is the interaction of basic filter components with ambient
SO 2 . This interaction is affected by ambient atmospheric conditions. A relationship
based on established chemistry was developed for the prediction of sulfate error
caused by this interaction. Recommendations of filter media appropriate for ambient
sulfate monitoring and further development of a methodology are included.
17. KEY WORDS AND DOCUMEN ANALYSIS
a. DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
C. COSATI Field/Group
*Air pollution materials
*Sulfates *Interactions
Particles
*sampling
*Effi ci ency
*Sulfur dioxide
13B
07B
l4B
13K
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (misReport)
UNCLASSIFIED
21. NO. OF PAGES
20. SECURITY CLASS (This page)
UNCLASSIFI ED
22. PRICE
EPA Form 2220-1 (9-73)
91
------- |