PB-248 259
THE FATE OF NITROGEN OXIDES IN THE ATMOSPHERE
Chester W. Splcer
Battelle Columbus Laboratories
Prepared for:
Environmental Protection Agency
13 September 1974
DISTRIBUTED BY:
National Technical Information Service
U. S. DEPARTMENT OF COMMERCE
-------
KEEP UP TO DATE
Between the time you ordered this report—
which is only one of the hundreds of thou-
sands in the NTIS information collection avail-
able to you—and the time you are reading
this message, several new reports relevant to
your interests probably have entered the col-
lection.
Subscribe to the Weekly Government
Abstracts series that will bring you sum-
maries of new reports as soon as they are
received by NTIS from the originators of the
research. The WGA's are an NTIS weekly
newsletter service covering the most recent
research findings in 25 areas of industrial,
technological, and sociological interest—
invaluable information for executives and
professionals who must keep up to date.
The executive and professional informa-
tion service provided by NTIS in the Weekly
Government Abstracts newsletters will give
you thorough and comprehensive coverage
of government-conducted or sponsored re-
search activities. And you'll get this impor-
tant information within two weeks of the time
it's released by originating agencies. '
WGA newsletters are computer produced
and electronically photocomposed to slash
the time gap between the release of a report
and its availability. You can learn about
technical innovations immediately—and use
them in the most meaningful and productive
ways possible for your organization. Please
request NTIS-PR-205/PCW for more infor-
mation.
The weekly newsletter series will keep you
current. But learn what you have missed in
the past by ordering a computer NTISearch
of all the research reports in your area of
interest, dating as far back as 1964, if you
wish. Please request NTIS-PR-186/PCN for
more information.
WRITE: Managing Editor
5285 Port Royal Road
Springfield, VA 22161
Keep Up To Date With SRIM
SRIM (Selected Research in Microfiche)
provides you with regular, automatic distri-
bution of the complete texts of NTIS research
reports only in the subject areas you select.
SRIM covers almost all Government re-
search reports by subject area and/or the
originating Federal or local government
agency. You may subscribe by any category
or subcategory of our WGA (Weekly Govern-
ment Abstracts) or Government Reports
Announcements and Index categories, or to
the reports issued by a particular agency
such as the Department of Defense, Federal
Energy Administration, or Environmental
Protection Agency. Other options that will
give you greater selectivity are available on
request.
The cost of SRIM service is only
domestic (600 foreign) for each complete
microfiched report. Your SRIM service begins
as soon as your order is received and proc-
essed and you will receive biweekly ship-
ments thereafter. If you wish, your service
will be backdated to furnish you microfiche
of reports issued earlier.
Because of contractual arrangements with
several Special Technology Groups, not all
NTIS reports are distributed in the SRIMi
program. You will receive a notice in your
microfiche shipments identifying the excep-
tionally priced reports not available through
SRIM.
A deposit account with NTIS is required
before this service can be initiated. If you
have specific questions concerning this serv-
ice, please call (703) 451-1558, or write NTIS,
attention SRIM Product Manager.
This information product distributed by
U.S. DEPARTMENT OF COMMERCE
National Technical Information Service
5285 Port Royal Road
Springfield, Virginia 22161
-------
&*
I
I
I
September 13, 1974
Final Report
THE FATE OF NITROGEN
OXIDES IN THE ATMOSPHERE
Baffeiie
Columbus Laboratories
A report of research conducted for the
Coordinating Research Council Inc.
and
U.S. Environmental Protection Agency
£
j
-------
BIBLIOGRAPHIC DATA 1- R«-port No. 2-
SHEET CRC-APRAC-CAPA-9-71-1
4. Title ;incj Subtitle
The Fate of Nitrogen Oxides .i n the Atmosphere
7. Author(s) _,. ,_ _ .
Chester W. Spicer
9. Performing Organization Name and Address
Battelle Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
12. Sponsoring Organization Name and Address
Environmental Protection Agency, Research Triangle Par, North
Carolina 27711 and The Coordinating Research Council, Inc.
30 Rockefeller Plaza, New York, New- York 10020
3. Recipient's Accession No.
5. Report Date
Seot 13f 197U
6.
8. Performing Organization Rept.
No.
10. Project/Task/Work Unit
No.
11. Contract/Grant No.
13. Type of Report & Period
Covered
Final Report
14.
15. Supplementary Notes
Not releasable to the general public until December 1975
16. Abstracts ^Q prOgram consisted of three distinct phases involving analytical methods
development, field studies, and analysis and interpretation of the field study results.
The analytical development phase of the program involved developing or refining state-
of-the-art techniques for the determination of ambient levels of PAN,NH3, and NH^. The
field sampling phase of the program consisted of 5 weeks of air monitoring and particu-
late collections in St. Louis, MO and West Covina, CA. In addition, NO, N029 03, MO",
NO" and C, H, N were determined. Meteorological variables including wind speeds wind
direction, temperature, relative humidity and solar intensity were also monitored con-
tinuously. In St Louis, several rainfall samples were collected and analyzed for trace
nitrogen compounds. Composite dust samples were collected in both cities and analyzed
for nitrogen constituents. Silver-membrane filter samples were also taken in West
Covina for analysis by electron spectroscopy chemical analysis (ESCA).
17. Key Words and Document Analysis. 17a. Descriptors
Air Pollution
Nitrogen
Atmospheric
Air Quality
Ozone .
NO-NOj-NO
Ammonia x
Peroxyacetyl Nitrate
Nitric Acid
Aerosol
17b. Identifiers/Open-Ended Terms
St. Louis, Missouri
West Covina, California
Rainfall
Dustfall
17c. COSATI Fie Id/Group
13B; 07B; 07C; 07F); OUB; 11G
18. Availability Statement
Releasable to the general public.
19. Security Class (This
Report)
UNCLASSIFIED
[21. No. of Pages
20. Security Class (This
Page
UNCLASSIFIED
FORM NTIS-38 (REV. 3-72)
THIS FORM MAY. BE REPRODUCED
USCOMM-DC M932-P72
-------
REPORT
on
THE FATE OF NITROGEN OXIDES IN THE ATMOSPHERE
to
COORDINATING RESEARCH COUNCIL, INC.
(CAPA-9-71)
and
U.S. ENVIRONMENTAL PROTECTION AGENCY
(Contract No. 68-02-0799)
September 13, 1974
Prepared by: Chester W. Spicer
BATTELLE
Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
I*/
-------
The Battelle-Columbus staff who made major contributions to this program are listed below
along with their project responsibilities.
James L. Gemma
Darrell W. Joseph
Arthur Levy
Chester W. Spicer
Statistics
Field Study, Data Management
Advisor
Principal Investigator
We would like to acknowledge the help and guidance of the members of the CAPA 9-71
Project Committee:
Dr. Emmett Jacobs - Chairman
Dr. Joseph J. Bufalini
Dr. Emmett Burk
Dr. William Glasson
Dr. Robert Hammerle
Mr. Dale Hutchison (deceased)
Dr. John Pierrard
-------
TABLE OF CONTENTS
Page
INTRODUCTION 1
SCOPE OF SUMMARY AND PROGRAM 1
TECHNICAL BACKGROUND . . . 3
Nitrogen Distribution in Smog Chamber Studies 3
Nitrogen Distribution in the Atmosphere 4
Atmospheric Nitrogen Balance 6
Global Considerations . . . 6
Urban Environments 7
EXPERIMENTAL METHODS 9
Meteorological Measurements 9
Air Quality Measurements 9
Ozone . 11
NO-N02-NOX 11
Ammonia 12
Peroxyacetyl Nitrate (PAN) 12
Nitric Acid 13
Continuous Monitoring 13
Integrated Analysis 14
Aerosol Collections 15
Aerosol Analysis 16
Rainfall Analysis 17
Dustfall Analysis 17
Vertical Measurements 17
SAMPLING SITES 19
St. Louis, Missouri 19
West Covina, California 19
RESULTS 21
St. Louis, Missouri 21
West Covina, California 28
ANALYSIS AND INTERPRETATION 33
St. Louis, Missouri 33
Air Quality Data 33
Aerosol Composition 35
Distribution and Balance of Nitrogen Compounds 38
Statistical Interpretation 38
Nitrogen Balance 41
West Covina, California 48
Air Quality Data 48
Aerosol Composition 49
Distribution and Balance of Nitrogen Compounds 52
Statistical Interpretation 54
Nitrogen Balance 56
REFERENCES 71
-------
LIST OF TABLES
Page
Table 1. Background Levels of Atmospheric Nitrogen Compounds 6
Table 2. Meteorological Measurements 9
Table 3. Air Monitoring Instrumentation 11
Table 4. Summary of Air Quality Data - St. Louis 22
Table 5. Summary of St. Louis Hydrocarbon Data 24
Table 6. St. Louis Vertical Sampling 25
Table 7. St. Louis Aerosol Data 26
Table 8. Analysis of Composite St. Louis Dust Sample 27
Table 9. St. Louis Rainfall Analysis . . . 27
Table 10. Summary of Air Quality Data - West Covina 29
Table 11. West Covina Aerosol Data 31
Table 12. Analysis of Composite West Covina Dust Sample 32
Table 13. St. Louis Aerosol Composition (Weight Percent) 35
Table 14. Size Classification of Various Aerosol Components by City (Weight Percent) . . 36
Table 15. St. Louis Filter Nitrogen Balance 37
Table 16. St. Louis AID Results ...:.... 39
Table 17. St. Louis Regressions 40
Table 18. St. Louis Regressions 40
Table 19. St. Louis Regressions 41
Table 20. Emissions Inventories for St. Louis Region 44
Table 21. St. Louis-West Covina Comparisons 48
Table 22. West Covina Aerosol Composition (Weight Percent) 50
Table 23. Interlaboratory Comparisons 51
Table 24. Aerosol Elemental Ratios for St. Louis and West Covina (Molar Ratios) .... 51
Table 25. West Covina Filter Nitrogen Balance . . 53
Table 26. Aerosol Nitrogen Balance ...'..... 53
Table 27. West Covina AID Results (First Three Predictor Variables) ........ 54
Table 28. West Covina Regressions 55
Table 29. West Covina Regressions 56
Table 30. Results of AID Analysis of "NOX Loss" Dependent Variable 67
Table 31. West Covina "NOX Loss" Regressions •. ... 67
-------
LIST OF FIGURES
Figure 1. Battelle-Columbus'Mobile Air Quality Laboratory 10
Figure 2. Average Diurnal Air Quality and Meteorological Profile, St. Louis 23
Figure 3. Average Diurnal Air Quality and Meteorological Profile, West Covina 30
Figure 4. St. Louis Regression: CO Versus C2H2 43
Figure 5. St. Louis Regression: CO Versus NOX 45
Figure 6. St. Louis Regression: C2H2 Versus NOX 47
Figure 7. Scattergram of L.A.-Data Fate of NOX 58
Figure 8. West Covina NOX Loss 60
Figure 9. West Covina NOX Loss and PAN + HN03 Profiles 69
-------
FIRST YEAR REPORT
on
THE FATE OF NITROGEN OXIDES IN THE ATMOSPHERE
to
COORDINATING RESEARCH COUNCIL. INC.
(CAPA-9-71)
and
U.S. ENVIRONMENTAL PROTECTION AGENCY
(Contract No. 68-02-0779)
from
BATTELLE
Columbus Laboratories
September 13, 1974
INTRODUCTION
The objective of this program is to determine the distribution and ultimate fate of nitrogen oxides in
the atmosphere. A secondary goal of the project is to uncover relationships among the nitrogen compounds
and other atmospheric parameters in order to define the conditions under which nitrogen oxides are removed
and nitrogen reaction products accumulate in the atmosphere. Nitrogen oxides long have created problems
in understanding smog chemistry due to the multifarious nature of their reactions. Nitrogen oxides are in-
volved in virtually every aspect of photochemical smog formation, and although much has been learned about
the mobile and stationary sources of nitrogen oxides, the fate or ultimate disposition of these species in
polluted atmospheres has remained a mystery. Because of the possibility that undetermined nitrogen-
containing reaction products may be biologically hazardous and because of the need for smog modelers and
persons responsible for control strategies to understand the mechanism of photochemical smog formation,
this study was undertaken.
SCOPE AND SUMMARY OF PROGRAM
The current program consisted of three distinct phases involving analytical methods development, field
studies, and analysis and interpretation of the field study results.
i
The analytical development phase of the program involved developing or refining state-of-the-art tech-
niques for the determination of ambient levels of PAN, NH3, HNO-, and NH*
The field sampling phase of the program consisted of 5 weeks of air monitoring and particulate collec-
tions in St. Louis, Missouri, and 5 weeks in West Covina, California. In addition to the chemical measure-
ments mentioned above, NO, N02, 03/ NOj, NOj, and C,H,N were determined. Meteorological variables
including wind speed, wind direction, temperature, relative humidity, and solar intensity were also monitored
continuously. In St. Louis, several rainfall samples were collected and analyzed for trace nitrogen compounds.
Vertical NOX sampling using a fixed-wing aircraft was conducted during several days in St. Louis. Composite
-------
dust samples were collected in both cities and analyzed for nitrogen constituents. Silver-membrane filter
samples were also taken in West Covina for analysis by electron spectroscopy chemical analysis (ESCA).
The analysis and interpretation phase of the program involved statistical analysis of the data from both
cities with the purpose of deriving relationships among chemical and meteorological parameters in order to
help us improve our understanding of the fate of nitrogen oxides.
Although the analysis and interpretation of our results will continue into another year, there are a
number of observations and tentative conclusions which stand out at this time. Several of these are
summarized below.
During the early phases of this program, two techniques for monitoring nitric acid were developed.
The first of these is a continuous procedure which utilizes a modified, acid-detecting coulometric instrument.
The second technique is an integrating method wherein nitric acid in the air is separated from nonvolatile
particulate nitrate and collected in solution. The solution is subsequently analyzed for nitrate. These two
methods have been employed successfully in smog chamber research and in the field studies described in
this report. Based on these two techniques, nitric acid was observed in the atmospheres of both St. Louis
and West Covina. The nitric acid concentration was found to vary from less than 1 ppb to greater than
30 ppb on an hourly average basis.
In order to derive a "balance" between nitrogen oxide reactant and NOX reaction products, it was
necessary to determine the fraction of NOX which was converted to products or removed from the atmos-
phere at any given time. This "NOX Loss" was derived using the equation
..~ .
N°x Loss = . --
chemical smog days. This portion of the "NOX Loss" profile was highly correlated with ozone concentra-
tion. If certain assumptions are accepted, the afternoon or photochemical loss of nitrogen oxides appears
to be largely accounted for by the sum of the PAN and nitric acid concentrations.
There was some evidence that nitric acid is removed from the air by alkaline-surface glass-fiber filters,
but not by high-purity quartz-fiber filters. If true, this would mean that much particulate nitrate data
collected over the years may have been strongly influenced by gaseous nitric acid.
There was also some indication that ozone is transported into the St. Louis region during early morning
hours. While the source of this ozone is not yet clear, there is some evidence that PAN and possibly nitric
acid are associated with the nighttime ozone.
The program is continuing, and the second year effort will be devoted in part to further validation and
documentation of our nitric acid measurements and in part to further analysis and interpretation of our
first-year field results. Special emphasis in the second year will be placed on expanding our data base to
include data collected at the same time by other research groups in the St. Louis and Los Angeles area.
Inclusion of these data in our analyses will increase both the scope of the investigation and the accuracy
. of the results. V
-------
TECHNICAL BACKGROUND
This section contains a discussion of previous laboratory
and atmospheric studies relating to the distribution and fate
of nitrogen oxides in the atmosphere.
-------
3
TECHNICAL BACKGROUND
To give some insight into previous investigations of the chemistry of the nitrogen oxides in laboratory
studies and in the atmosphere, a review of pertinent research is in order. We believe much can be learned
from laboratory simulations of photochemical smog reactions and will therefore start by reviewing attempts
to determine nitrogen balances in smog chamber studies.
Nitrogen Distribution in Smog Chamber Studies
The question of the fate of atmospheric nitrogen oxides probably first arose from the lack of an ade-
quate nitrogen balance in laboratory studies of photochemical smog-type'reactions. Early investigators
attributed the loss of nitrogen during irradiation experiments to adsorption of nitrogen oxides on the reac-
tion vessel walls or to the formation of complex polymers of carbon and nitrogen.
In his study of trans-2-butene photooxidation, Tuesday'1' reported that all of the initial nitrogen in
his system was recoverable as peroxyacetyl nitrate (PAN), methyl nitrate, and residual nitrogen dioxide.
However, Gay and Bufalini'2' have pointed out that the application of Stephens''3' more recent infrared
absorption coefficient to Tuesday's data reveals an actual nitrogen balance of only 78 percent at the end of
his experiment. Altshuller and Cohen'4' reported only a 13 percent recovery of nitrogen as nitrogen dioxide
and methyl nitrate at the conclusion of ethylene photooxidation. In the photooxidation of the propylene-
nitrogen oxide system, Altshuller, et al.'5', reported good nitrogen balances up to the time when nitrogen
dioxide reached a maximum. After that time, the nitrogen balance depended on the initial nitrogen oxide
concentration. When the initial NOX was less than 0.2 ppm, all of the nitrogen could be accounted for as
PAN. Between 0.2 and 0.5 ppm initial NOX, 50 to 90 percent of the nitrogen oxide which reacted was
attributable to PAN. At initial NOX levels between 0.5 and 1.5 ppm, only 35 to 75 percent of the nitrogen
oxide consumed could be recovered as PAN. Methyl nitrate was also detected but it accounted for only a
minute fraction of the missing nitrogen. No other organic nitrogen compounds could be found in their
system. Altshuller, etal.'6' irradiated both toluene-NOx and m-xylene-NOx systems. Peroxyacetyl nitrate
accounted for only 10 to 20 percent of the reacted nitrogen in the former case and 10 to 75 percent in
the latter. However, as found previously, the percentage of recovered nitrogen increased rapidly as the
initial nitrogen oxide concentration was decreased below 1.2 ppm.
Several other organic nitrogen compounds have been investigated in laboratory photooxidations in
the hope that they would improve the poor nitrogen balances. Chief among these are the alkyl nitrates
and various nitroolefins. Methyl and higher molecular weight alkyl nitrates have been identified by
several researchers'1.7-10) jn laboratory photooxidations of olefins and aromatic hydrocarbons with NOX.
Several alkyl nitrates have also been identified as products in irradiated automobile exhaust.'11-12' The
yields of alkyl nitrates are much lower than PAN yields, however, and can account for only a small per-
centage of the missing nitrogen. The possible presence of nitroolefins has been tested in irradiated
propylene-NO2 mixtures, in irradiated auto exhaust, and in the atmosphere.'131 The conclusion was
reached that these compounds could not be present in concentrations greater than one or two parts per
billion.
Our understanding of the fate of nitrogen oxides was increased by the study of Gay and Bufalini'2',
wherein nitrate (presumably derived from nitric acid formed on the reaction vessel walls by hydrolysis of
nitrogen pentoxide) was identified as one of the principal nitrogen-containing products. By determining
nitrate and nitrite in the liquid used to scrub their reaction-chamber walls, they obtained nitrogen balances
of 97 percent for 2-methyl-2-butene photooxidation, 90 percent for m-xylene, 72-80 percent for isopropyl-
benzene, 92-100 percent for 1,3 butadiene, and 100 percent for both propylene and 1-butene In addition
to the excellent nitrogen balances they achieved. Gay and Bufalini also demonstrated that neither molecular
nitrogen nor nitrous oxide is a likely product of photochemical smog reactions.
-------
In the recent study by Spicer and Miller'14' of nitrogen balances in smog chamber systems, both PAN
and HN03 were monitored continuously for the first time. Excellent nitrogen balances were reported
throughout smog chamber irradiations in single hydrocarbon/NOx and synthetic auto exhaust/NOx systems.
The primary nitrogen-containing reaction products were reported to be PAN and HN03. with the ratio of
PAN/HN03 dependent on the amount and nature of the surface available for heterogenous reaction and
the overall reactivity of the hydrocarbon system.
Nitrogen Distribution in the Atmosphere
While much effort has gone into identification and analysis of nitrogen-containing compounds in the
atmosphere, there has never been a concerted effort to determine ell the various nitrogen-containing species
at one place and at one time. The greatest level of effort in the measurement of atmospheric nitrogen com-
pounds has been accorded to nitric oxide and nitrogen dioxide. Almost all of the oxides of nitrogen emitted
from both mobile and point sources is released as NO. The means by which the NO is converted to N02 has
been a subject of controversy. However, the result of the conversion process is approximately equal concen-
trations of NO and N02, on a yearly average basis, at most monitoring locations. Naturally, figures for yearly
averages and maximum daily averages vary from one location to another. Using Los Angeles NOX levels as
examples, the yearly averages have been reported'16' as 0.08 ppm and 0.06 ppm for NO and N02, respec-
tively. Kopczynski.et al.'16', found morning NOX levels in downtown Los Angeles averaging 0.41 ppm in
the fall of 1968. The morning concentrations were found to vary from 0.09 to 1.06 ppm NOX during the
period of the study.
Most of the investigations which have determined atmospheric PAN and PPN levels have been carried
out in the general Los Angeles area. Examples of such studies are those of Darley, et al.'17', Stephens and
Price'13', and Kopczynski, et al.'16'. Darley and co-workers found 50 ppb PAN and 6 ppb PPN during a
day of heavy air pollution in Riverside, California. Samples taken during April and May of 1968 by Stephens
and Price showed between 5 and 40 ppb PAN in Riverside. Kopczynski, et al., studying reactions in Teflon
bags in downtown Los Angeles, reported between 8 and 97 ppb PAN formed in naturally irradiated Los
Angeles morning air.
A search for some atmospheric nitroolefins was carried out by Stephens and Price'13' in the metro-
politan Los Angeles area in the spring of 1964. None of the samples collected showed any trace of
nitroolefins.
There is good reason to suspect alkyl nitrates as products of smog reactions. These compounds are
known decomposition products of PAN and they have been identified in many laboratory smog chamber
studies. However, Stephens and Darley'18' found no evidence of alkyl nitrates in atmospheric samples.
They would have detected as little as 0.5 ppb.
Ammonia concentrations in nonurban atmospheres rarely exceed 0.015 ppm.'19' For example,
Hodgeson, et al.'20', report ammonia concentrations ranging between 0.002 and 0.010 ppm in the
Research Triangle Park area. Urban NH3 concentrations reported by the National Air Surveillance
Network (NASN)'19' have been considerably higher than the nonurban concentrations. Ammonia con-
centrations between 0.10 and 0.22 ppm were found by Morgan and co-workers'21' in several urban and
nonurban locations. However, the method of chemical analysis for NH3 used by NASN is subject
to some uncertainty and has been discontinued as a routine measurement.
A recent study of ammonia concentrations at Harwell in Great Britain'22-23' found the average
ammonia concentration to be 1-3 ppb. The ammonia profile at Harwell is almost constant with the
highest ammonia levels reaching only 7 ppb. Lodge, et al.'24', have reported on ammonia concentrations
in the St. Louis, Missouri, area. They generally found ammonia present at 0-10 ppb and could find no
firm evidence that ammonia was associated with the urban plume of St. Louis. Their results indicate that,
within experimental error, the concentration of NH3 within the urban plume is probably the same as that
outside the plume.
-------
All reported attempts to identify nitric acid in the gas phase in urban atmospheres have met with failure.
Scott, et al.'25', tested for HNO3 by an infrared technique which should have detected as little as 0.1 ppm. Re-
cent long-path Fourier transform infrared measurements made by Hanst and co-workers'261 have also failed to
uncover nitric acid. The possibility remains, however, that HN03 is present in urban air but at levels below the
detection limits of current instruments; or it may be present but has gone undetected either due to its reaction
with other atmospheric constituents or its adsorption on the walls of the infrared sampling system.
Nitric acid has been identified in the lower stratosphere by Murcray, et al.'27', using infrared absorp-
tion techniques during balloon flights. The nitric acid appears to be concentrated in a layer between 22
and 30 km altitude and is apparently associated with the ozone layer. Rhine, et al.'28', have attempted to
quantitate the stratospheric nitric acid reported by Murcray, et al.'27', and report approximately 3 ppb
HN03 associated with the ozone layer. They suggest that the reaction of MO with N02 may be the most
important producer of nitric acid in the ozone layer. However, the smog chamber evidence from the study
of Spicer and Miller'14' implicates the heterogenous reaction of N205 with H20 as the primary source of
nitric acid in photochemical smog, with only a minor contribution from the HO + N02 reaction. Reaction
conditions in the lower stratosphere are quite different from those occurring in urban atmospheres, however,
so the mechanism of nitric acid formation may be quite different at the two altitudes. Crutzen'29' has
calculated production rates and concentrations of ozone, nitric acid, and several other species in the strat-
osphere, and reports that the nitric acid values found by Mulcray, et al,'27', in the ozone layer are much
higher than the levels which are predicted using current rate constants.
Based on an 8-year study'30' of the mean concentration of selected particulate contaminants in the
atmosphere of the United States, it appears that nitrate on the average contributes somewhat less than
2 percent of the total suspended particulate weight. The figures vary depending on location. Several rep-
resentative urban areas were Atlanta at 2.0 Mg/m3, Chicago at 2.5 jug/m3, Boston at 2.3 pg/m3, and
Pittsburgh showing 3.0 M9/m3.
Air samples from Cincinnati, Chicago, Philadelphia, and Fairfax, Ohio, were collected and analyzed
for phosphate, nitrate, chloride, and ammonium by Lee and Patterson.'31' These investigators obtained
size distributions on the nitrate aerosol which showed that nitrate is primarily associated with submicron
particles. They concluded that suspended nitrate originates from a gaseous source rather than a wind-erosion
source. Reaction of HN03 with their filter materials was not discussed. Average nitrate concentrations
reported by Lee and Patterson were 2.96 /;g/m3 in Cincinnati and 2.83 jzg/m3 in Fairfax, Ohio.
Recent work by Miller, et al.'32', in Columbus, Ohio, and the New York City area differs from the
Lee and Patterson results in that up to 50 percent of the nitrate was found in particles greater than 2 /um
in diameter. Gordon and Bryan'33' have reported the presence of NH4N03 in the Los Angeles aerosol
based on infrared spectra of high-volume filter extracts and a molar ratio of NO3/NH4 very close to one
over the past several years. However, reactions and interactions of NH3 and HN03 with glass fiber filters
employed in their study were not considered.*
The size distribution and concentration of ammonium aerosols has been investigated by Lee and
Patterson'31' in Chicago, Philadelphia, and Fairfax, Ohio. Ammonium concentrations were found to be
4.00/jg/m3, 9.45 Aig/m3, and 5.74 pg/m3, respectively. They found that ammonium particulate is pre-
dominantly submicron in size. A similar size distribution was also reported.by Miller, et al.'32'. The size
distributions suggest that ammonium originates from gaseous materials rather than from wind erosion.
Evidence was also obtained that at least some of the ammonium present in the atmosphere is in the form
of (NH4)2SO4. Junge'34' has reported a high correlation between ammonium and sulfate in ratios close
to that for (NH^jSO^, and Dubois and co-workers'35' in a recent presentation have also shown correla-
tion between atmospheric ammonium and sulfate using NHt specific ion electrodes.
A somewhat different view of nitrogen species in aerosol samples has been reported by Novakov,
et al.'3^', who employed electron spectroscopy chemical analysis (ESCA) to determine the oxidation
states of elements found in Pasadena aerosol samples. They report nitrate to be associated primarily with
larger (2-5 /urn) particles. However, they find most of the nitrogen in the aerosol in the 0.6-2.0 nm size
range, this nitrogen reportedly consisting of pyridino, amino, and to a lesser extent ammonium compounds.
The ESCA technique was also employed for several samples collected for this program.
*We believe such reactions may be important on typically alkaline glass filters and will discuss such interactions later in the
report.
-------
6
Atmospheric Nitrogen Balance
Global Consideration!
A study of the global nitrogen cycle, reported by Robinson and Robbins'371, estimates the background
concentrations of the major trace atmospheric constituents as shown in Table 1.
TABLE 1. BACKGROUND LEVELS OF ATMOSPHERIC
NITROGEN COMPOUNDS
Compound Ambient Concentration
N2O 0.25 ppm
NO 2 ppb over land
NO2 4 ppb over land
NH3 6 ppb
NOj 0.2 jug/m3
NHj 1.0jug/m3
Nitrous oxide accounts for most of the mass of trace nitrogen compounds. Its cycle is independent
of the other nitrogen compounds in the troposphere, consisting largely of a balanced system of biological
production and biological removal from the atmosphere. Another minor sink exists above 30 km, where
N20 can participate in photochemical reactions. However, in the troposphere N20 appears to play no part
in photochemical smog reactions. Gay and Bufalini121 have also established that N20 is not a product of
smog reactions.
Ammonia and ammonium aerosols are reportedly produced primarily in soil with an estimated at-
mospheric residence time for NH3 of less than one week. The major scavenging processes on a global basis
are gaseous deposition and aerosol formation. Ammonium aerosols are subsequently removed from the
atmosphere by precipitation and dry deposition. An earlier report by Robinson and Bobbins'3"' that NHg
is oxidized in the atmosphere to oxides of nitrogen is discredited in their most recent work.
The NO and NO2 cycle involves both natural and anthropogenic sources at an emissions ratio of
approximately ten to one. On a global basis the NO is scavenged largely by oxidation to NO2, with the
NO2 removed by gaseous deposition and aerosol formation. Nitrate aerosol is removed by precipitation
and dry deposition.
The sources and sinks for atmospheric nitrogen compounds reported by Robinson and Robbins are
of utmost importance when considering global nitrogen balance. However, the sources and removal paths
for the same nitrogen species are quite different in urban polluted atmospheres. While the global sinks for
the minor nitrogen compounds are sufficient to prevent long-term buildup of their concentrations, still the
natural scavenging processes are not rapid enough to affect nitrogen compound concentrations over short
time intervals (hours) in urban atmospheres. Therefore, short-term losses of nitrogen compounds in polluted
atmospheres must depend on other atmospheric processes.
l_CVy139) nas formulated a steady-state model of the chemistry in the lower troposphere and used it
to calculate the hourly concentrations of a number of species including NOj, N2Og, HNOg, and HNO2.
It must be emphasized that this model was developed for the troposphere in general and is not meant to
-------
simulate the rather unique chemistry of urban polluted atmospheres. The model predicts the major nitrogen
products of daytime reactions to be HN03 and HN02, with HN03 predominating. Levy suggests that nitric
acid never approaches its full potential concentration (~100 ppb) due to the formation of nitrate aerosols
and their subsequent scavenging by precipitation. However, the nitrate rain-out rate which follows from the
model is almost two orders of magnitude higher than the measured rate reported by Erickson.'40' Levy
concludes that ". . . the question of nitrogen balance in the troposphere and the role of HN03 are far from
settled".
Junge'41' has examined the distribution of ammonium and nitrate in rainwater over the United States
and reports ammonium levels of 0.1-0.2 mg/liter in the St. Louis area during the summer and about 1.0 mg/
liter in the Los Angeles vicinity during early summer. Rainwater nitrate levels were approximately 0.3-0.4
mg/liter for St. Louis and 0.5 mg/liter for Los Angeles during early summer. These results are for the years
1956-1957. The distribution pattern for nitrate in rain samples seems to eliminate the photochemical pro-
duction of N02 in the stratosphere as a source of tropospheric nitrate. There was also no correlation of
NO^ in rainwater with thunderstorm activity, thus ruling out the argument that a large fraction of the nitrate
is formed by lightning. The conclusion reached by Junge was that the major fractions of both NH^ and N03
in rainwater have as their source the earth's soil. Alkaline soils especially appear to generate large quantities
of ammonia. The formation of small amounts of NO and N02 in soil was also postulated.
Georgii'42' has also studied oxides of nitrogen and ammonia in the atmosphere, primarily by examining
the concentrations of these species in precipitation. He reports nitrite present in rainwater at up to 10 per-
cent of the nitrate concentration. He also suggests and gives evidence for the formation of HN03 and HNO2
by simple absorption of gaseous NO2 in rainwater.
The results of these investigations into the global nitrogen budget have a considerable bearing on the
study of the fate of nitrogen oxides in urban atmospheres. However, the processes for forming and scavenging
nitrogen compounds on a global scale are generally too slow to have a significant influence over the short-
term balance of nitrogen species in an urban area. The following section discusses the investigations which
have attempted to examine this short-term balance.
Urban Environments
Perhaps the first report of urban nitrogen balance resulted from the study of Gordon, et al.'43', in
Los Angeles. In that investigation, hourly bag samples were collected in downtown Los Angeles (DOLA)
and Azusa from 5:00 a.m. through 5:00 p.m. for 46 weekdays. The bag samples were subsequently
analyzed for C2-C5 hydrocarbons. Hourly averages for oxidant, nitrogen oxides, and carbon monoxide
were obtained from the DOLA and Azusa monitoring stations. From the integrated bag sample values for
acetylene and the hourly average NOX concentrations from the monitoring stations, ratios of acetylene to
NOX were computed for several hours during the day. The average ratios vary between about 0.30 and
1.00, considerably higher than the acetylene to nitrogen oxide ratio reported'44' for integrated bag sam-
ples collected during the California test driving cycle. This discrepancy could result from a nonrepresen-
tative driving cycle'43', from a loss of atmospheric NOX, or from a nonautomotive source of atmospheric
acetylene. Since acetylene has been found to originate almost entirely from the automobile in urban at-
mospheres, we are left with the possibilities of a nonrepresentative driving cycle or a rapid atmospheric
loss process for NOX. For most of the hourly average ratios, the magnitude of the daily oxidant level
appeared to have little influence on the difference between the driving cycle ratio and the atmospheric
ratio. In other words, the discrepancy between the two ratios was similar on both high and low oxidant
days. This is an important fact, since many processes which may be important in removing nitrogen ox-
ides from the atmosphere involve reactions of ozone and would be expected to increase on high oxidant
days. If this were the case, then a difference between the acetylene/NOx ratios would be expected for
low and high oxidant days.
Another investigation which employed the ratio approach to study atmospheric nitrogen balance has
been reported by Eschenroeder and Martinez.'45' In their analysis of Los Angeles atmospheric data from
1968 to 1969, these researchers employed both CO and C2H2 as tracers in order to study the nitrogen
-------
8
balance. Both CO and C2H2 are derived almost exclusively from auto exhaust and both are relatively inert to
photochemical smog reactions. Since reasonably reliable emissions data are available for CO and NOX and
since the C^h^/NO,, auto exhaust ratio has been determined in the California driving cycle, one has available
the CO/NOX and C^Hj/NO,, ratios from the emissions sources. Two types of ratios are pertinent to our dis-
cussion. The "vehicular" ratio of CO/NOX and C2H2/NOX is the ratio of these pollutants present in auto
exhaust. The "all sources" ratio is somewhat lower than the vehicular ratio, due to NOX emission from
sources other than the automobile. This ratio includes the emissions input from all known sources of CO,
CjHj, and NOX. Once the emissions ratio for CO/NOX or C2H2/NOX is determined for a given region, the
discrepancy between the theoretical or emissions inventory ratio and the measured ratio can be used as an
indication of nitrogen oxide losses, assuming CO and C2H2 are inert over short time intervals.
In the report of Eschenroeder and Martinez, C2H2/NOX ratios were averaged over all the days in 1969
in Commerce, California, and the composite hourly averages plotted as a function of time. Similar plots
were made of the CO/NOX ratios from Huntington Park, California, for a limited number of days in 1968
and for Commerce throughout 1969. The CO/NOX ratios ware also separated into high oxidant O0.2 ppm)
and low oxidant (<0.2 ppm) days and plotted accordingly.
The 1968 CO/NOX ratios from Huntington Park differed considerably from the emissions ratios, with
high oxidant days showing much higher ratios (presumably greater NOX losses) than the low oxidant days.
At its maximum (approximately 7:00 a.m.) the CO/NOX curve for high oxidant days exceeded the "all
sources" emissions ratio by a factor of four, indicating very large and rapid nitrogen oxide losses.
The ratios of both CO/NOX and C2H2/NOX for all days in 1969 at Commerce showed rather different
trends than the limited 1968 data. The C2H2/NOX ratios varied between the vehicular and all sources ratios
with a rise toward late morning. The CO/NOX ratios for low oxidant days showed the same behavior. On
high oxidant days, the trend was similar, but the curve was generally higher (greater NOX losses) than on low
oxidant days. The indication from the 1969 data is that losses of nitrogen oxides are rather small on low
oxidant days, with slightly greater losses occurring on days of higher oxidant. In general, the maximum
losses appeared between 10:00 a.m. and 12:00 a.m.
The difference in IMOX losses between the 1968 and 1969 data may reflect the relatively small number
of sampling days from 1968 as opposed to the full year's data for 1969. In this regard, one would expect
the 1969 results to be the more reliable because of the much greater number and wider range of sample days,
It should be pointed out that two interpretations can be made of the greater NOX loss on high oxi-
dant days. One view is that reactions of ozone lead to NOX loss, with the loss consequently higher on
high oxidant days. Another interpretation holds that, on certain days, some undefined process removes
NOX from the atmosphere leading to a higher hydrocarbon/NOx ratio and a more rapid formation of
oxidant. From this point of view, the lower NOX causes the high oxidant day.
Another study which reported on the fate of nitrogen oxides in an urban area was carried out by
Lodge, et al.'24', in St. Louis, Missouri. A number of pollutants were investigated in the program.
Although the results of the study are very interesting, little was learned about the fate of nitrogen oxides
because the experimental design did not take into account NOX input to the urban plume from localized
NOX sources along the plume's path.
-------
EXPERIMENTAL METHODS
This section contains a discussion of the experimental
methods employed during the field sampling phase of the
program.
-------
9
EXPERIMENTAL METHODS
Battelle-Columbus' Mobile Air Quality Laboratory, pictured in Figure 1, was used in St. Louis, Mo.,
and West Covina, Calif., to monitor meteorological conditions, solar irradiation intensity, and gas-phase
composition, while simultaneously collecting aerosol samples for subsequent chemical analysis. Data from
all instruments were fed to a Digi-Tem data acquisition system where the analog input is converted to
digital form, serialized by bit, and presented in ASC II code on paper tape. Each instrumental channel
was interrogated every 10 minutes for 23 hours each day, from 11:30 p.m. to 10:30 p.m. The informa-
tion on paper tape was read into Battelle's CDC-6400 computer, conditioned, and permanently stored on
magnetic tape. A computer program was used to average all analyses (except wind direction) and plot all
analyses as 23-hour profiles.
Instrumentation and experimental methods employed in the Mobile Laboratory are discussed in the
following paragraphs.
Meteorological Measurements
The Battelle Mobile Laboratory makes use of an automated MRI, Inc., Model 1071 weather station.
The instruments listed in Table 2 provide a continuous readout of temperature, relative humidity, wind
speed, and wind direction. The laboratory also employs an Eppley Laboratory, Inc., 180° pyrheliometer
to determine global radiation (total sun and sky).
Laboratory and newspaper records were kept of general weather conditions such as cloud cover and
intervals during which rain occurred.
TABLE 2. METEOROLOGICAL MEASUREMENTS
Analysis Instrument
Wind Speed and Direction MRI, Model 1074-2 sensor
Temperature MRI, Model 802 sensor
Relative Humidity MRI, Model 907 sensor
Global Radiation Intensity Eppley Lab, 180° pyrheliometer
Air Quality Measurements
The instruments used for air quality monitoring are listed in Table 3. Air samples are pulled into the
Mobile Laboratory through an aluminum stack used for high-volume aerosol sampling. The top of the
stack is about 15 feet above the trailer roof or roughly 25 feet above the ground. Flow rate through the
stack is at least 30 cubic feet per minute, so that the residence time is no more than about 6 seconds in
the stack. From the stack the samples are transported through short lengths of 1/4-inch Teflon tubing to
the appropriate instrument. A brief description of the instrumental methods employed in this study follows.
-------
10
FIGURE 1. BATTELLE COLUMBUS' MOBILE AIR QUALITY LABORATORY
-------
11
TABLE 3. AIR MONITORING INSTRUMENTATION
Analysis
03
NO T
Instrument
REM, Model 612 chemiluminescence monitor
"i
I
F
-*
.. * ,. .... . I analyzer, carbon catalytic converter
N02 (by difference) J *
NH3 Bendix, Model 8101-B chemiluminescence
analyzer, dual stainless steel and carbon
catalytic converters
Peroxyacetyl Nitrate (PAN) Varian Series 1200 gas chromatograph with
electron-capture detection
HNO3-continuous Modified acid-detecting Mast coulometric
analyzer
HN03-integrated Modified chromatropic acid procedure
Ozone
V
Ozone determination is based on the detection of the chemiluminescence from the reaction of 03 with
excess ethylene. Peak intensity from this reaction is in the 4000-4500 A region. The technique is thought
to be free from major interferences. The sensitivity of the instrument used for this study is about 0.001 ppm.
The instrument was zeroed daily and was calibrated at the beginning and end of the field study using a cali-
brated McMillan Electronics Corporation ozone generator. A heated Hopcalite catalyst bed was used to
destroy the ethylene effluent from this instrument.
NO-N02-NOX
The instrument used for nitrogen oxide determination employs the reaction of nitric oxide with excess
ozone for the generation of infrared (1.2 /u peak) chemiluminescence. The infrared radiation passes through
a filter and is detected by a sensitive photomultiplier tube. Noise from the photomultiplier tube is reduced
by cooling the tube to subambient temperatures. The instrument used here has a sensitivity of 0.005-
0.010 ppm and is linear over a range of 0.005-5.00 ppm. Activated carbon is used to remove ozone from
the instrument's exhaust gas.
The chemiluminescent technique works quite well for nitric oxide, with no interferences having been
documented. The determination of nitrogen dioxide by chemiluminescence requires initially a reduction of
N02 to NO with subsequent NO determination as described above. A carbon catalyst operated at 260 C
has been employed during this study to reduce N02 to NO. The efficiency of this converter is excellent -
i.e., greater than 99 percent as measured before and after the field program.
Basic nitrogen compounds such as NH3 and amines are known to interfere if high-temperature cata-
lytic converters are used, but the use of low-temperature carbon converters eliminates this interference.
However, we found in the initial stages of this program that several more highly oxidized nitrogen-containing
molecules are reduced to and detected as nitric oxide, even by the relatively low-temperature carbon con-
verter. Any specie which is reduced to NO by the converter will give a positive N02 response. The inter-
ference caused by nitric acid, while showing considerable variation, was often close to quantitative. There
-------
12
was some indication that the response was initially quantitative but slowly decreased as if the catalyst were
subject to fatigue or poisoning. Nearly 100 percent response was found for ethyl nitrate at low concentrations.
Peroxyacetyl nitrate was also found to Interfere, although we have not yet determined the response factor.
The possible effects of these interferences on the current program will be discussed later in this report.
The nitrogen oxide chemiluminescent instrument was zeroed and spanned each day. The concentration
of the nitric oxide (in nitrogen) calibration gas was determined at the beginning and end of the field program
by an ozone titration technique.
Ammonia
The continuous determination of gaseous ammonia was accomplished through the use of a dual catalyst
chemiluminescent technique which has been described by Hodgeson and co-workers.*46"481 The procedure
capitalizes on the fact that basic nitrogen compounds such as ammonia are oxidized to NO by high-
temperature (>650 C) catalytic converters but not by low-temperature catalysts. Higher oxides of nitrogen
are reduced to NO by either high- or low-temperature converters. Passing the sample air stream through the
high-temperature converter yields total NOX + NH3, while passing the sample gas through the lower tempera-
ture converter yields only total NOX. The difference between the two outputs is the ammonia gas concentration.
The instrument used for this study was a Bendix Model 8101 -B NO-NO2-NOX chemiluminescent analyzer
which was modified by the addition of a Thermo Electron, Inc., high-temperature stainless steel converter
operated at 700 C. The air stream was alternately passed through the high-temperature and low-temperature
converters with the difference in output equaling the ammonia concentration. Because of the strong tendency
toward ammonia adsorption, the instrument was further modified by moving the flow-restricting capillary,
which is normally positioned just prior to the detector cell, to the inlet of the instrument. This modification
permitted the entire instrument to operate under a partial vacuum of 460 torr rather than at atmospheric
pressure and therefore helped reduce NH3 adsorption. Teflon tubing was used throughout the system where
possible to further minimize adsorption effects. Even after these modifications, however, the instrument's
NH3 signal, while ultimately quantitative, exhibited a 10 to 20-minute delay in responding to a step change
in ammonia concentration. Thus, short-term fluctuations in ammonia concentration in the atmosphere are
not discerned by this instrument.
The instrument used for NH3 determination was zeroed daily and spanned at least every 2 weeks by
successively diluting gas from a known high-concentration cylinder of NH3 in nitrogen. The concentration
of NH3 in the cylinder was confirmed by an extended range chemiluminescent instrument equipped with a
high-temperature stainless steel converter.
Peroxyacetyl Nitrate (PAN)
Determination of PAN was carried out with a Varian Series 1200 gas chromatograph fitted with a
tritium-source electron-capture detector. A 19-inch column of 1/8-inch Teflon packed with 10 percent
Carbowax 400 on Anakrome ABS was operated at 30 C with oxygen-free nitrogen at 30 cc/minute as the
carrier gas. The gas chromatograph was calibrated with authentic samples of PAN which had been analyzed
prior to dilution using the infrared absorption coefficients reported by Stephens.'3' Known concentrations
of ethyl nitrate were included with the dilute PAN samples so the ratio of sensitivities of PAN to C2H5ONQ2
could be determined. This ratio was very similar to that reported by Dimitriades.'49' The lower limit of
detection for PAN was approximately 0.001 ppm.
During the field program the instrument was occasionally checked for sensitivity by preparing and
analyzing low concentrations of ethyl nitrate in Teflon bags. PAN sensitivity can then be determined using
the ratio of PAN/C2H5ON02 sensitivities determined earlier. A daily check of detector standing current
was used to correct short-term sensitivity fluctuations.
-------
13
Another check of the accuracy of the gas chromatograph technique was made during the St. Louis
phase of the field study. Comparisons were run between an EPA-operated gas chromatograph and the
Battelle instrument using PAN samples diluted down from a high-concentration cylinder of PAN provided
by EPA. The EPA and Battelle instruments agreed within 10 percent on the PAN concentration in the
cylinder.
Occasionally, during the field program a 3-foot chromatographic column of the same composition as
described earlier was operated at 0 C in an attempt to detect methyl and ethyl nitrate. There was some
indication that both nitrates were present; however, the concentrations never exceeded about 0.001 ppm.
Exact determination of these alkyl nitrates was complicated by peaks, perhaps Freons, which eluted from
the chromatograph with retention times very similar to the nitrates.
Nitric Acid
Two techniques, one continuous and one integrated, have been developed at Battelle-Columbus for
nitric acid analysis. The continuous technique employs coulometry for detection, in effect detecting nitric
acid by its acid properties. The integrated method involves a colorimetric procedure which makes use of
the nitrate moiety for quantitative detection.
Continuous Monitoring. The technique used for continuously monitoring nitric acid has been described
in detail by Miller and Spicer'50' and will be discussed only briefly here. The instrument used was a Mast
microcoulomb meter (Model 724-21) which was adapted for sensing acids rather than oxidants. Details con-
cerning the configuration and theory of operation of the Mast meter in monitoring oxidants have been
reported by Mast and Saunders.'51' Adaptation of the meter to permit detection of acid gases was reported
by Miller, et al.<52' Since many atmospheric gases, among them O3, NO2, and SO2, are detected by the
acid-sensing coulometric cell, sample pretreatment is necessary to provide specificity for nitric acid. The
apparatus for sample conditioning consists of a gas-titration cell with an ethylene source to remove inter-
ference due to ozone. The sample stream, after ozone titration, is passed alternately (1) directly into the
detection cell and (2) through a trap containing loosely packed nylon fiber (Atlas Electronic Devices
Company) enroute to the detector. The manner of packing the nylon trap is critical. It must be sufficiently
loaded to remove all the HNO3 yet not so heavily that it starts to adsorb SO2 or NO2. The reading ob-
tained from direct operation is an indication of total acid content; the reading after passing the sample through
the nylon trap indicates essentially all acid gases except nitric. The nitric acid concentration is thus obtained
by difference. A timer is set to take the instrument through a complete sample/zero cycle every 15 minutes,
thus generating a series of square waves on a recorder, with the difference between each peak and valley being
attributed to nitric acid. In practice the data are reduced manually by drawing a continuous curve through
the level portion of all the "sample" intervals and a second curve through the level portion of all the "zero"
intervals. The average difference between the two curves is the average nitric acid concentration. Large
changes in S02 or NO2 concentrations manifest themselves as changes in both the sample and zero curves
and have generally been noted to have a time constant which is much longer than one instrument cycle. If
the SO2 or NO2 level were to increase or decrease markedly and then return to the original value in the
course of 2-3 minutes, that is, completely within the sample mode or zero mode of the instrument, then a
positive or negative interference could result, depending on the direction of the change and the instrument
mode. Such sudden changes would be very obvious but have not been observed in our sampling programs.
They might be expected to occur much more frequently in the vicinity of major S02 or NO2 sources.
The theoretical response of the detector cell to a strong acid can be determined from Faraday's Law
and the parameters of the Mast titration cell to be 10 ju ampere/ppm acid. The instrument was calibrated
by preparing and purifying nitric acid on a high-vacuum line and isolating known (pvt) volumes of acid in
transfer flasks. Acid was subsequently injected into 50-cu-ft Teflon bags. Within an estimated accuracy of
5 percent, the instrument's response to nitric acid was quantitative; i.e., 10 /ia/ppm HN03 (using a con-
ventional 500-ohm resistor in series with the meter, the equivalent response in millivolts dc is 5/ppm HNO3).
Additional verification of the calibration was obtained from time-integrated analyses of nitric acid using a
colorimetric method described in the next section. The sensitivity of the instrument at a signal-to-noise
ratio of 2/1 is about 2 ppb.
-------
14
While calibrating, several important observations were made which are worth noting here: (1) at low to
moderate humidities nitric acid persists in the gas phase. When pure HN03 at the ppm level was injected
into bags at somewhat higher humidities, the acid seemed to cluster as aerosols with water, as indicated by a
condensation nuclei counter (Environment One). Simultaneous measurements of the condensation nuclei
concentrations and nitric acid concentrations Indicated further that such clusters may be labile; (2) the acid
will significantly absorb then desorb from Teflon sampling tubing; and (3) chemiluminescence instruments
commonly used for determining N02 as NO by catalytic reduction also respond to nitric acid.
Considerable effort has been devoted to investigating potential interfering species which might be
encountered under the conditions for which the monitor was designed to operate. No detectable inter-
ference has been observed for S02 (0.5 ppm), N02 (1 ppm), PAN (0.2 ppm), H2S04 (0.2 ppm), and
CH20 (0.5 ppm). Minor interference has been observed for HC02H. Interference testing is still con-
tinuing at Battelle-Columbus, particularly for N205 and organic acids. At this time, however, we believe
strongly in the specificity of the technique under conditions likely to be encountered in atmospheric and
laboratory situations. This confidence has been gained through examining nitric acid data from numerous
smog-chamber experiments conducted under a variety of common and exaggerated atmospheric condition?
Integrated Analysis. Because of the possibility, discussed earlier in this report, that up to 100 ppb of
nitric acid might be found in the troposphere (and presumably even higher concentrations in urban polluted
atmospheres), it was thought that a second check on the nitric acid concentration could be carried out by
colorimetry. The procedure described below can be used to measure moderate to high levels of nitric acid
(>20 ppb) with reasonable accuracy. At low levels of HN03, less than 10-20 ppb daily average, interfer-
ence by NO2 becomes a major problem, affecting the accuracy of the method. At these lower HN03 levels
it is best to consider the procedure only as an upper limit method.
The vapor pressure of pure nitric acid at 20 C is close to 47 torr'63-54', so that several tons of thou-
sands of parts per million of pure nitric acid could be present in ambient air without reaching saturation.
The vapor pressure drops quite drastically when nitric acid is combined with water, however, so the possi-
bility of nitric acid-water mixtures forming as aerosols or adsorbed on the surface of existing aerosols should
be considered. The integrated method developed for nitric acid analysis makes use of the highly volatile
nature of nitric acid to separate the acid, whether in the gas or aerosol phase, from the nitrate particulates
present in the atmosphere. Since separation depends on passing HNO3 through a filter, it became necessary
to demonstrate that nitric acid is indeed volatile enough to pass through a filter under simulated atmospheric
conditions. Separation is accomplished by passing the sample air stream through a Teflon mat filter element
(Millipore). Extensive testing prior to use of this method in the field has shown that nitric acid passes
through such filters quantitatively under all conditions of relative humidity and aerosol loading studied. Even
at 90 percent relative humidity and a particulate count of 107 (latex beads and NiO; nuclei count just after
addition to the chamber) the nitric acid passed quantitatively through the Teflon filter. Several interesting
observations were made during these tests of nitric acid filtration:
(1) Teflon or stainless steel filter holders are a necessity; nylon or other plastic
filter holders remove nitric acid from the sample stream. This is consistent
with our earlier work, since nylon fiber was found to efficiently remove
nitric acid when used in conjunction with the coulometric technique de-
scribed earlier.
(2) Excellent agreement was observed between the coulometric and colorimetric
procedures for synthetic mixtures of nitric acid in air. Both in turn agreed with
the theoretical concentration of nitric acid added to the system.
(3) Both Teflon fiber filters and high-purity quartz filters passed nitric acid quanti-
tatively. Polycarbonate membrane filters and glass fiber filters partially remove
nitric acid from the air stream.
-------
15
Operationally, the procedure employed for integrated nitric acid analysis consisted of drawing the
ambient air sample through a Teflon mat filter in a Teflon filter holder, then through a fritted bubbler in
30 cc of 0.03 N NaOH, and finally through a moisture trap and pump. A jeweled orifice at the pump
exit was used to accurately maintain the desired flow. During the field study, sampling was carried out
at approximately 1 liter per minute for 23 hours a day, from 11:30 p.m. to 10:30 p.m., to cover the
same time interval as the other gas-phase and aerosol analyses.
Nitrate in the NaOH solution was subsequently determined by a modification of the chromotropic
acid method.'55-56' Of the many colorimetric methods available for nitrate determination, the chroma-
tropic acid technique was chosen because of its sensitivity, its fairly wide acceptance among analysts, and
its ready adaptability to field use.
The basic analytical procedure is that given by West and Ramachandran'56'. However, several modi-
fications were incorporated to make the method suitable for air analysis and to simplify it for field use.
These changes include
(1) Addition of five drops of sodium sulfite-urea solution to each 2.5 ml sample
to eliminate interferences from nitrite and other oxidizing agents
(2) Elimination of an acidic antimony solution recommended'56' to remove chloride
interference in water samples
(3) Nitrate interference is expected from the absorption of NO2 in the basic solu-
tion according to the reaction
2N02 + H2O -* HN03 + HN02 .
The fraction of N02 yielding N0§ in solution has been reported'57' to be
28 percent. However, several laboratory experiments with our apparatus
yielded a nitrate value of 20 percent. This value was used, along with our
chemiluminescent data for N02 concentrations, to correct the integrated HN03
results. As stated earlier, at high average levels of nitric acid (>20 ppb) this
correction for N02 interference will be small at typical ambient NO2 concen-
trations. However, at low levels of nitric acid (less than 10-20 ppb on a daily
average basis) the N02 interference becomes a major fraction of the nitrate
signal, thus affecting the accuracy of the measurement. Under these condi-
tions the integrated procedure is useful as an upper limit estimate of the nitric
acid concentration.
High concentrations of formaldehyde were found to give a positive nitrate interference, even though the
analytical wavelength for nitrate determination (X - 410 nui) is at a minimum in the absorbance curve for
the formaldehyde-chromatropic acid complex (X max = 580 mju). Attempts to eliminate this interference
by complexing or reducing the formaldehyde using sodium bisulfite, 3-methyl-2-benzothiazolone hydrazone
hydrochloride (MBTH), and phenylhydrazine were unsuccessful. In theory, a correction factor can be
derived from the ratio of absorbances of the chromatropic acid complex at 410 and 580 m/u, combined
with the sample absorbance at X = 580 mju, as long as the nitrate complex does not interfere at 580 mp.
AbsX = 410 nut
While nitrate interference at 580 nut was not a problem, the ratio — for the formaldehyde
AbsX = 580 mn
complex was neither constant nor linear in formaldehyde concentration. Indeed, the absorbance of the
formaldehyde complex at X - 410 mju reached a limiting value beyond which it remained constant regard-
less of formaldehyde concentration. The reasons for this are unclear; however, it renders any formaldehyde
correction based on the absorbance ratio ambiguous.
One technique which has been successfully used to minimize formaldehyde interference makes use of
the absorbance limit reached by the formaldehyde complex. By adding an excess of formaldehyde to both
sample solution and blank, the formaldehyde interference essentially is cancelled. Any difference in absorb-
ance between sample and blank should be related only to the NO^ concentration. As will be discussed
shortly, this technique was unnecessary in our atmospheric work because of the relatively little formaldehyde
present. However, it has been successfully employed in our smog chamber work to eliminate the interference
due to high concentrations of formaldehyde (0.2-0.6 ppm).
-------
16
The technique of adding excess formaldehyde to the sample solutions was unnecessary in our atmo-
spheric work due to the low average formaldehyde levels in the atmosphere. Average formaldehyde levels
of 0.04 ppm have been reported by Altshuller and McPherson<58) for 7 a.m. to 4 p.m. during the smog
season in Los Angeles. During the highest oxidant levels in Los Angeles in 1973, Hanst, et al.(26), found
no formaldehyde but up to 0.04-0.06 ppm formic acid, which will also likely interfere with the nitrate
analysis. These concentrations are for peak daylight periods. The levels of formaldehyde have been shown
to approach 0.02 ppm at night in Los Angeles. Therefore, a reasonable daily average formaldehyde or formic
acid concentration might be 0.03-0.04 ppm. Based on the work done in this laboratory on the absorbance
of the formaldehyde chromophore at 410 nr>M, these average levels of formaldehyde would yield a 0.001-
0.002 ppm interference with respect to nitric acid.
Although we correct for the positive interference due to N02 and believe the formaldehyde inter-
ference is minimal, still it is wise to consider the nitric acid results from the integrated procedure as
upper limits, especially when the concentration of HN03 is low.
Aerosol Collections
Aerosol samples for nitrogen compound analyses were usually collected on a 23-hour-per-day basis,
although some day/night collections were also made. Samples were collected from 25 feet above ground by
two 6-inch-diameter aluminum stacks. High-volume blowers were used to maintain an average flow rate of
40 cfm. Pressure-drop measurements were made at the beginning and end of each day's sampling in order
to correct for day-to-day fluctuations in flow rate. Initial calibration of the blowers was done with a cali-
brated venturi.
Aerosol samples were collected on 6-inch-diameter high-purity quartz mat filters (Pallflex Product
Corporation) backed by a stainless steel fritted disk. Each filter was cut and preweighed in the laboratory
at 40 percent relative humidity and then stored in an individual glassine envelope enclosed in a sealed poly-
ethylene bag. After sampling, the filter was returned to its glassine envelope in its plastic bag, purged with
and then sealed under argon. On return to the Columbus Laboratories the filters were equilibrated at
40 percent relative humidity and re weighed. The filters were then partitioned for analysis.
A third high-volume sampler incorporated a specially designed cyclone which served as a collector for
particles >2.5 Mm in diameter. The cyclone was operated for 23 hours per day for the entire sampling
period in each city. The total participate collected by the cyclone was analyzed as a representative average
"dustfdll" sample for each city.
Ten aerosol samples were collected in Los Angeles on 1-inch silver filters for analysis by ESCA
(electron spectroscopy chemical analysis). The preweighed silver filters were held in a stainless steel filter
holder and operated at a flow rate of either 10 liters/minute or 1 cfm. The filters were then stored in
clean glass vials under argon prior to weighing and analysis.
Aerosol Analysis
Filters from the field study were analyzed at Battelle's Columbus Laboratories for NHj, NO^, NO3~,
and total carbon, hydrogen, and nitrogen. In addition, selected filters were analyzed for nitrogen com-
pounds by several other laboratories to confirm and extend the determinations carried out at BCL. A
brief description of the various methods follows:
• Ammonium was determined by dissolving soluble ammonium salts in water,
adding NaOH, and measuring the resulting NH3 with an ammonia gas sensing
electrode.
• Nitrite was determined using an aliquot of the water extract of the filter and
a diazotization-colorimetric method, ASTM 1254.
-------
17
• Nitrate analysis made use of an aliquot of the water extract from the filter and
and the brucine sulfate method, ASTM D992. Some samples were also analyzed
by the chromotropic acid method described earlier.
• Total C, H, N were determined with a Perkin Elmer Model 240 Elemental
Analyzer which employs pyrolysis and thermal conductivity detection for
elemental analysis.
Several other research groups aided this study by undertaking collaborative analyses.
The following laboratories* performed the analyses indicated:
• California AIHL — ammonium determination by the indophenol blue method
and nitrate by the 2,4 xylenol method
• Rockwell International — nitrate determination by a recently developed
polarographic technique
• Lawrence Berkeley Laboratory, University of California - reduced and
oxidized forms of nitrogen by electron spectroscopy chemical analysis (ESCA).
Samples were cooled to liquid nitrogen temperature to avoid loss of volatiles
under the vacuum conditions necessary for analysis.
Rainfall Analysis
Rainfall samples were collected in two stainless steel rainfall jars mounted on the roof of the Mobile
Laboratory. After a rain, the sample was sealed in a clean glass vial with a Teflon stoppered cap and returned
to the Columbus Laboratories,for NH^, NO^, and NO^ analysis. A log was kept of the time during which
rain occurred and the amount of rain which fell. Newspaper records were used for the latter determination.
Dustfall Analysis
Particles larger than 2.5 Aim were collected 23 hours per day by a special cyclone sampler connected
to a high-volume blower. Air was delivered to the cyclone from 25 feet above ground through a 6-inch-
diameter aluminum stack. At the end of sample collections in each city, the composite "dustfall" sample
was returned to Battelle's Columbus Laboratories, weighed, and analyzed for NH^, NO^, NO^, and total
C, H, N. The results of these analyses are reported in percent by weight of the total dust sample.
Vertical Measurements
Two days of vertical sampling for NO and NOX were carried out in St. Louis, Missouri. A twin-
engined aircraft belonging to Battelle's Northwest Laboratories was stationed in St. Louis for another
program and was made available for verticle nitrogen oxide measurements.** Nitrogen oxides measure-
ments were made with a REM, Inc., NO-NOX instrument within the airplane. Temperature readings were
taken simultaneously with the NOX data.
The aircraft was flown initially to the Mobile Laboratory site at St. Louis University. Data on
temperature and nitrogen oxide concentration were then collected as the aircraft spiralled upward from
1000 feet to 5000 feet.
"We mould like to express our appreciation for these analyses to Dr. Bruce Appel of the California AIHL, Dr. T.
Novakov of the University of California, and Dr. Ed Parry of Rockwell International.
"We would like to thank Drs. J. M. Hales and A. J. Alkezweeny for their help in this effort.
-------
SAMPLING SITES
The two field sampling sites are described in
this section of the report.
-------
19 i
SAMPLING SITES
St. Louis, Missouri
The topography of the St. Louis region is gently rolling, with elevations from 480 feet above sea level
in the downtown area to 550 feet at Lambert Field 12 miles to the northwest, with a slight ridge rising to
about 600 feet in between. The Mississippi River marks the eastern boundary of the city and separates the
downtown area from highly industrialized East St. Louis, Illinois. The elevation of the Mississippi River at
this point is about 400 feet above sea level. The area is generally free from major land surface features which
could strongly influence airflow characteristics over the area.
The Battelle Mobile Air Quality Laboratory was situated at St. Louis University approximately 2 miles
due west from the St. Louis Arch. The laboratory was bounded by a small University faculty parking lot
along the edge of a soccer field. The area immediately surrounding the University campus can best be char-
acterized as nonindustrial urban. Our laboratory was located next to an EPA-CPL mobile laboratory which
conducted detailed hydrocarbon analyses and CO determinations on integrated bag samples collected through-
out St. Louis for several days simultaneously with our sampling program.
West Covina, California
Sampling in the Los Angeles basin was carried out in West Covina, a suburban community located
approximately 25 miles east of "downtown" Los Angeles. The city of West Covina is flanked by the San
Gabriel mountains to the north and the Puente Hills to the south. The area is therefore centered in the
corridor through which the prevailing easterly winds funnel the urban Los Angeles air mass.
The mobile laboratory was situated on West Covina school district property next to a seldom used
athletic field approximately 1-1/2 miles south of the San Bernardino Freeway. The surrounding area can
be described as suburban, consisting largely of single-family dwellings with no major industry.
-------
RESULTS
The results of the field sampling phase of the
program are presented in this section of the report.
-------
21
RESULTS
Summaries of the field monitoring data collected in St. Louis, Missouri, and West Covina, California,
w|ll be presented in this section. To facilitate the presentation of the results, St. Louis and West Covina
will be treated separately here and in the subsequent data analysis discussions.
St. Louit, Missouri
A detailed summary of the air quality data for the 26 days of field monitoring in St. Louis in July
and August, 1973, is given in Table 4. The gas-phase data are shown as both 23-hour averages and as
maximum 1-hour averages. The weather during the field program can be generally categorized as hot and
humid.
Figure 2 profiles the average diurnal air quality and meterology during the field program. The indi-
vidual daily profiles are included as Appendix A in a separate volume of this report. Also included in
Appendix A are the detailed hydrocarbon and CO data collected by the Environmental Protection Agency
mobile laboratory situated next to our mobile laboratory in St. Louis. These data are provided as 2-hour
averages from 6:00 a.m. to 7:00 p.m. for 5 simultaneous EPA-Battelle sampling days. All EPA hydro-
carbon and carbon monoxide data were taken by gas chromatography. A summary of these EPA data is
given in Table 5.
Three days of vertical NOX sampling were carried out in St. Louis using the Battelle-Northwest
airplane and chemiluminescent NO-NOX analyzer. The results of these airborne studies are given in
Table 6.
Table 7 presents the aerosol results from the St. Louis field program. Detailed inorganic nitrogen
compositions are shown in the table along with total carbon, hydrogen and nitrogen, and total mass loading
values. It must be emphasized that these aerosol samples were collected and analyzed on high-purity quartz
fiber filters as opposed to the traditional glass fiber filters.
The analysis of the composite dust sample for the 5-week St. Louis monitoring program is shown in
Table 8. The composite dust sample was collected over the 5-week program with a cyclone sampler
operating to collect particles of mean diameter greater than 2.5 micrometers. The results are given as
weight percent of the total dust sample.
Several rainfall samples were also collected in St. Louis and analyzed for nitrogen-containing species.
The results of these analyses are shown in Table 9. The quantity of rain which fell over the city is
reported in the table in inches.
-------
TABLE 4. SUMMARY OF AIR QUALITY DATA - ST. LOUIS
Pollutants
Weather Conditions
Date
7-18
7-19
7-20
7-22
7-23
7-24
7-25
7-26
7-27
7-30
7-31
8-1
8-2
8-3
8-4
8-5
8-6
8-7
8-8
8-9
8-10
8-12
8-13
8-14
8-15
8-16
Day
W
Th
F
Sun
M
T
W
Th
F
M
T
W
Th
F
Sat
Sun
M
T
W
Th
F
Sun
M
T
W
Th
General(a)
S
R.S
S
S
S.R
S
S
R.C
C
R
PC.R
C
C
S
S
S
PC
S
S
R
S.R
S.R
R
R.S
PC
R.S
Temp. C
28
27
29
27
26
27
27
26
27
25
24
21
22
23
25
26
25
27
30
24
24
26
23
21
23
23
RWfc
75
92
79
82
84
86
80
69
67
76
72
72
66
61
68
76
72
84
82
95
82
85
93
82
76
83
Aerotol Mass
Loading
Mg/mi
114.1
80.7
51.5
32.6
42.8
37.0
53.9
33.4
50.9
68.2
39.2
41.9
43.5
53.7
79.9
63.0
74.9
61.4
51.5
47.9
66.4
40.6
63.3
42.0
90.5
79.6
23 -Hour Average
ppm
0.049
0.038
0.027
0.072
0.041
0.037
0.044
0.032
0.028
0.023
0.032
0.022
0.022
0.034
0.052
0.045
0.042
0.038
0.023
0.004
0.041
0.037
0.016
0.035
0.041
0.029
NO.
ppm
0.012
0.016
0.013
0.019
0.005
0.011
0.021
0.013
0.013
0.026
0.015
0.016
0.014
0.011
0.007
0.009
0.011
0.012
0.014
0.024
0.015
0.026
0.026
0.013
0.034
0.016
NOX.
ppm
0.061
0.061
0.047
0.037
0.034
0.048
0.056
0.035
0.038
0.057
0.034
0.028
0.034
0.035
0.033
0.039
0.038
0.037
0.036
0.051
0.041
0.047
O.OS4
0.029
0.062
0.044
MH3.
ppm
0.010
0.008
0.008
0.002
0.004
0.006
0.005
0.004
0.004
—
0.004
0.002
0.003
0.002
0.003
0.003
0.002
0.002
0.003
0.003
0.002
..
0.005
0.002
0.003
0.002
PAN.
ppm
0.001
0.002
0.002
0.001
0.001
0.001
0.001
0.001
0.001
—
0.001
0.001
0.001
0.002
0.003
0.002
0.001
0.002
0.002
0.002
0.005
0.003
0.002
0.002
0.003
0.003
ppm
0.012
0.005
0.002
0.001
0.006
0.004
0.002
0.000
0.000
0.001
--
--
--
0.000
0.004
0.003
0.002
0.005
0.003
0.001
0.004
0.001
0.008
0.000
0.009
0.003
ppm
0.007
0.004
0.000
0.007
0.009
0.011
0.002
0.004
0.004
0.006
0.002
0.003
0.005
0.004
0.008
0.001
0.002
0.004
0.004
0.000
0.002
0.007
0.000
0.000
0.001
0.000
ppm
0.110
0.107
0.083
0.146
0.086
0.113
0.128
0.066
0.067
0.050
0.068
0.033
0.051
0.087
0.124
0.092
0.078
0.076
0.043
0.011
0. 163
0.138
0.058
-0.063
0.096
0.065
NO.
ppm
0.043
0.061
0.034
0.026
0.020
0.039
0.070
0. 0'J4
0.042
0.054
0.047
0.030
0.033
0.037
0.051
0.034
0.035
0.030
0.031
0.060
0.083
0.101
0.051
0.021
0.103
0.042
1-Hour Maximum
NOX.
ppm
0.114
0.121
0.070
0.050
0.067
0.081
0.108
0.060
0.069
0.089
0.084
0.049
0.061
0.077
0.0%
0.091
0.075
0.067
0.059
0.089
0.106
0.133
0.094
0.059
0.147
0.078
NH3.
ppm
0.014
0.011
0.011
0.016
—
0.011
0.009
0.005
0.006
--
0.005
0.003
0.005
0.004
0.006
0.006
0.004
0.005
0.006
0.007
0.004
—
0.008
0.004
0.009
0.005
PAN.
ppm
0.003
0.007
0.006
0.003
0.002
0.004
0.004
0.001
0.004
—
0.002
0.002
0.002
0.004
0.006
0.003
0.003
0.004
0.004
0.004
0.019
0.006
0.004
0.004
0.009
0.008
HN03,
ppm
0.044
0.022
0.012
0.003
0.055
0.016
0.017
0.000
0.000
0.006
--
—
10
0.000 *°
0.015
0.018
0.029
0.020
0.021
0.011
0.024
0.005
0.042
0.004
0.080
0.017
(a) S = Sunny, R = Rain, C = Clear, PC = Partly Cloudy, Cl = Cloudy.
(b) Continuous Coulometric.
(c) Integrated Colorimetric.
-------
St Louis
23
I I I , I .' I , I i I • :| I I i I , I
c
0>
i
c
PAN 0.004 ppm
HN03 0.009 ppm
Nitrogen oxides Q070 ppm
Ammonia 0.006 ppm
100
90
80
70
60
50
40
30
20
10
0
St Louis
Legend
Ozone 0.080 ppm
Nitric oxide 0.040 ppm
Nitrogen dioxide 0.040 ppm
Nitrogen oxides 0.070 ppm
100
90
80
70
60
50
40
30
20
10
St Louis
2 4 6 8 10 12 14 16 18 20 22
Hour of Day
Legend
Global irradiance 0.900 cal/sq cm/min
Wind speed 7.000 mph
Temperature 30.000 Centigrade
Relative humidity 100.000 percent
FIGURE 2. AVERAGE DIURNAL AIR QUALITY AND METEOROLOGICAL PROFILE, ST. LOUIS
-------
TABLE 5. SUMMARY OF ST. LOUIS HYDROCARBON DATA<«>
Hydrocarbon values in ppbC; carbon monoxide
values in ppb.
Date
Sample Number"*1
Component
Methane
CO
C2H2
°2H4
Olefins
Aromatics
Nonme thane
Hydrocarfaons
C2H4/C2H2
7-18-73
1
2420
»674
21
29
80
-
570
1.38
2
2541
2008
24
29
88
-
454
122
3
2390
1266
18
17
54
-
452
033
4
2053
578
5
7
34
-
140
1.42
1
2057
1405
17
30
77
-
295
1.79
7-19-73
2
2326
1746
26
28
92
-
435
1.06
3
2271
2134
38
40
103
-
555
1.03
4
2235
1634
27
26
73
-
424
OJ96
1
2310
1536
27
24
75
140
513
038
7-20-73
2
2418
2047
23
30
100
273
818
1.32
3
1984
594
12
15
60
166
530
122
4
1S70
390
10
11
39
109
376
148
1
2445
882
12
20
73
129
448
142
7-23-73
2
2235
1018
14
18
53
353
657
1.31
3
2094
817
4
8
35
220
448
1*1
4
2079
528
4
9
36
217
420
249
1
2107
1010
15
20
66
155
486
1.30
7-24-73
2
2129
1423
21
26
74
198
579
125
3
2285
1113
13
17
62
78
387
1.31
4
2117
930
12
15
41
46
284
126
(a) Courtesy of the U.S. Environmental Protection Agency.
-------
25
TABLE 6. ST. LOUIS VERTICAL SAMPLING
(Above the St. Louis University Site)
Date
7/30/73
7/31/73
8/1/73
Altitude, ft
Ground Level
1500
2100
3500
4500
Ground Level
1000
2500
4000
5500
Ground Level
1000
Temperature, C
32.0
27.7
25.3
22.5
19.2
26.5
23.6
12.1
15.7
15.9
21.8
—
NOX, ppm
0.060
0.100
0.080
0.060
0.045
0.022
0.041
0.040
0.030
0.022
0.020
0.045
N02, ppm
0.036
0.030
0.020
0.010
0.010
0.014
0.009
0.014
0.007
0.003
0.011
0.007
NO, ppm
0.024
0.070
0.060
0.050
0.035
0.008
0.032
0.026
0.023
0.019
0.010
0.038
-------
26
TABLE 7. ST. LOUIS AEROSOL DATA
Date
7/18/73
7/19/73
7/20/73
7/22/73
7/23/73
7/24/73
7/25/73
7/26/73
7/27/73
7/30/73
7/31/73
8/1/73
8/2/73
8/3/73
8/4/73
8/5/73
8/6/73
8/7/73
8/8/73
8/9/73
8/10/73
8/12/73
8/13/73
8/14/73
8/15/73
8/16/73
Average
N>
mg
11.6
7.4
1.2
3.9
4.4
2.8
2.8
1.3
1.9
2.8
1.7
2.6
2.1
1.7
6.1
7.5
7.4
8.2
1.6
0.2
2.0
1.6
4.8
4.1
8.9
11.2
4.3
3
Aig/m3
7.0
4.5
.77
2.3
2.7
1.7
1.7
.77
1.2
1.8
1.1
1.6
1.3
1.1
3.8
4.8
4.6
5.2
1.0
.13
1.3
1.0
3.1
2.6
5.6
6.8
2.7
NOj.
mg
.012
.009
.004
.010
.007
.011
.011
.010
.014
.006
.009
.006
.002
.009
.007
.002
.004
.005
.005
.001
.002
.000
.000
.000
.003
.000
.006
Nl
mg
.60
.75
.70
.40
.45
.62
.45
.40
.75
.55
.50
.43
.38
.80
1.00
.45
.60
.90
.60
.83
.50
.28
.40
.30
.70
.78
.58
Tr
MQ/m3
.35
.45
.45
•24
.27
.38
.28
.24
.46
.34
.31
.27
.24
.51
.63
.28
.37
.56
.39
.53
.31
.18
.26
.19
.43
.48
.36
c,
mfl
24.8
19.7
16.6
11.7
14.0
15.5
16.1
14.0
18.5
18.2
11.6
11.6
13.9
14.6
22.3
11.5
13.7
18.1
13.0
14.5
22.6
12.7
17.9
13.9
29.2
25.3
16.8
H.
mg
7.1
5.8
3.0
0.3
3.3
3.2
2.8
2.1
2.9
2.1
0.0
2.2
2.6
2.4
4.6
5.1
3.1:
5.8
1.9
2.1
3.6
1.4
2.5
1.8
5.7
6.6
3.2
N,
mg
9.3
6.7
1.6
3.6
4.5
2.5
2.7
1.7
2.2
2.8
1.9
2.6
2.1
1.7
5.0
5.3
6.1
7.1
1.3
.7
2.4
1.6
4.1
4.1
7.9
10.8
4.0
Mass
Loading,
WJ/m3
114.1
80.7
51.5
32.6
42.8
37.0
53.9
33.4
50.9
68.2
39.2
41.9
43.5
53.7
79.9
63.0
74.9
61.4
51.5
47.9
66.4
40.6
63.3
42.0
90.5
79 j6
57.8
-------
27
TABLE 8. ANALYSIS OF COMPOSITE ST. LOUIS
DUST SAMPLE
Particle diameters >2.5 /im.
Specie
NH;
NO-
NO'
Total Carbon
Total Hydrogen
Total Nitrogen
Concentration,
weight percent
0.55
0.001
2.65
14.6
1.8
1.5
TABLE 9. ST. LOUIS RAINFALL ANALYSIS
Date
7/9/73
7/24/73
7/30/73
8/9/73
8/10/73
8/13/73
8/16/73
Inches of Rain
0.10
0.21
0.18
0.97
Trace
1.18
0.03
NHj,
ppm
<1
2
<1
1
<1
<1
1
NO;.
ppm
0.08
0.13
0.13
0.36
0.23
0.23
0.37
NOj,
ppm
2.0
2.0
3.6
1.0
3.8
2.4
2.3
-------
28
West Covina, California
A detailed summary of the air-quality data for the 29 sampling days in West Covina, California, is
given in Table 10. The field program in West Covina was carried out in August and September of 1973.
The gas-phase data are shown in the table both as daily (23 hour) averages and as 1-hour maximum
averages. The weather during most of the West Covina field sampling program was qvercast in the morn-
ing until about noon, after which clear or partly cloudy conditions prevailed. An exception to this
pattern occurred on the last 3 days of the study when desert or "Santa Anna" winds reversed the wind
flow over the Los Angeles basin, bringing high temperatures and very low relative humidities to the area.
Figure 3 profiles the average diurnal air quality and meteorology during the field sampling in West
Covina. The individual daily profiles are included as Appendix B in a separate volume of this report.
Also included in Appendix B are Los Angeles Air Pollution Control District (LAAPCD) data from Azusa
for carbon monoxide, nitrogen oxides, and total hydrocarbons covering the perjod of our West Covina
monitoring program.
The aerosol data from the West Covina sampling program are shown in Table 11. The concentra-
tions of ammonium, nitrite, nitrate, and total carbon, hydrogen, and nitrogen are presented in micro-
grams per cubic meter along with the total daily mass loading data.
A composite dust sample (particle diameters >2.5 jum) was also collected by cyclone in West
Covina. The results of the composite dust sample analysis are shown as weight percent in Table 12.
-------
TABLE 10. SUMMARY OF AIR QUALITY DATA - WEST COVINA
Weather Conditions
Date
8-24
8-26
8-27
8-28
8-29
8-30
8-31
9-3
9-4
9-5
9-6
9-7
9-8
9-9
9-10
9-11
9-12
9-13
9-14
9-17
9-18
9-19
0-20
9-21
9-24
9-25
9-26
9-27
9-28
Day
F
Sun
M
T
W
Th
f
M-
T
W
Th
F
Sat
Sun
M
T
W
Th
F
M
T
W
Th
F
M
T
W
Th
F
GeneralW
S
C
S
PC
PC
PC
C
S
Cl
PC.S
S
S
Cl.S
Cl.S
Cl
Cl.S
Cl.S
Cl.S
Cl.S
--
S
PC
' --
--
c
S
S
S
c
Temp, C
21
18
20
19
20
19
18
18
17
18
20
18
18
19
17
18
17
17
17
16
16
17
16
18
16
19
22
23
23
Rffib
89
93
89
—
—
—
--
—
—
—
66
87
79
83
90
79
88
85
80
77
82
82
85
76
71
S3
50
37
44
Aerosol Mass
Loading
Vig/m3
115.8
54.9
65.1
81.7
98.9
96.1
87.0
64.5
79.7
77.6
118.1
135.8
76.6
63.6
64.7
81.8
105.5
94.1
123.7
127.4
145.8
137.3
116.5
122.7
73.1
106.3
198.2
123.7
105.6
Pollutants
24-Hour Average
ppm
0.054
0.035
0.039
0.050
0.060
0.044
0.043
0.042
0.021
0.048
0.083
0.047
0.055
0.036
0.017
0.041
0.031
0.048
0.064
0.070
0.058
0.052
0.048
0.050
0.032
0.051
0.040
0.031
0.052
NO.
ppm
0.071
0.046
0.029
0.061
0.060
0.020
0.021
0.007
0.013
0.020
0.080
0.114
0.045
0.055
0.056
0.'025
0.058
0.020
0.013
0.020
0.046
0.044
0.041
—
0.121
0.037
0.177
0.147
0.175
NOX.
ppm
0.180
0.081
0.103
0.152
0.171
0.120
0.113
0.062
0.082
0.090
0.204
0.239
0.093
0.105
0.122
0.096
0.151
0.071
0.075
0.109
0.152
0.140
0.112
—
0.163
0.103
0.252
0.221
0.256
NH3.
ppm
0.000
--
0.008
0.007
0.005
0.005
0.004
0.002
0.002
0.002
0.002
0.004
0.002
0.002
0.002
0.002
0.002
—
—
--
--
—
—
—
—
--
--
--
PAN.
ppm
0.003
--
--
—
—
—
0.001
0.002
0.003
0. 007
0.013
0.008
0.007
0.005
0.003
0.007
0.006
0.010
0.018
0.018
0.020
0.016
0.012
0.010
0.007
0.012
0.008
0.006
0.009
HN03.
ppm
0.009
0.001
0.000
0.002
0.003
0.002
0.000
0.002
0.001
0.000
0.012
0.011
0.004
0.001
0.002
0.002
0.002
0.001
0.005
0.007
0.010
0.004
0.004
—
—
0.002
0.000
0.000
0.002
HNO3.
ppm
0.010
0.002
0.000
0.003
0.000
0.000
0.001
0.005
—
0.000
0.018
0.003
0.009
--
—
0.006
0.000
0.010
0.026
0.016
0.018
0.019
0.008
0.000
0.007
0.011
0.007
0.009
0.014
ppm
0. 173
0.068
0.142
0.210
0.264
0.171
0.157
0.132
0.088
0.171
0.271
0.180
0.147
0.103
0.040
0.184
0.151
0.154
0.209
0.234
0.213
0.227
0.210
0.170
0.132
0.176
0.150
0.152
0.173
NO
ppm
0.202
0.089
0.065
0.274
0.205
0.045
0.051
0.042
0.032
0.052
0.324
0.346
0.065
0.068
0.090
0.048
0.176
0.062
0.022
0.071
0.092
0.133
0.110
--
0.439
0.277
0.407
0.453
0.557
1-Hour Maximum
NOX.
ppm
0.296
0.166
0.143
0.360
0.272
0.173
0.161
0.099
0.125
0.127
0.445
0.487
0.143
0.127
0.160
0.129
0.253
0.121
0.126
0.135
0.227
0.202
0.199
0.409
0.403
0.470
0.706
0.691
NH3.
ppm
0.002
—
0.013
0.013
0.008
0.007
0.006
0.003
0.003
0.003
0.005
0.005
0.003
0.003
0.003
0.003
0.003
--
—
—
—
--
--
—
--
--
—
--
PAN,
ppm
0.004
—
—
~
«
—
0.003
0.006
0.006
0.015
0.036
0.027
0.017
0.010
0.008
0.023
0.027
0.025
0.040
0.042
0.046
0.044
0.040
0.025
0.019
0.029
0.013
0.008
0.025
HNO3.
ppm
0.020
0.003
0.000
0.011
0.015
0.009
0.000
0.015
0.005
0.000
0.040
0.022
0.015
0.005
0.013
0.010
0.019
0.007
0.024
0.031
0.034
0.015
0.016
--
--
0.010
0.000
0.000
0.009
-------
30
"5
*-
c
0>
o
Los Angeles
Legend
PAN 0.020 ppm
HN03 0.010 ppm
Nitrogen oxides 0.200 ppm
Ammonia 0.006 ppm
Los Angeles
Legend
Ozone 0.200 ppm
Nitric oxide 0.200 ppm
Nitrogen dioxide 0.100 ppm
Nitrogen oxides 0.200 ppm
Los Angeles
8 10 12 14 16 18 20 22
Hour of Day
Global irradiance 0.900 cal/sq cm/min
Wind speed 6.000 mph
Temperature 30.000 Centigrade
Relative humidity 100.000 percent
FIGURE 3. AVERAGE DIURNAL AIR QUALITY AND METEOROLOGICAL PROFILE, WEST COVINA
-------
31
TABLE 11. WEST COVINA AEROSOL DATA
NHj
Date
8/24/73
8/26/73
8/27/73
8/28/73
8/29/73
8/30/73
8/31/73
9/3/73
9/4/73
9/5/73
9/6/73
9/7/73
9/8/73
9/9/73
9/10/73
9/11/73
9/12/73
9/13/73
9/14/73
9/17/73
9/18/73
9/19/73
9/20/73
9/21/73
9/24/73
9/25/73
9/26/73
9/27/73
9/28/73
Average
mg
3.5
1.3
2.5
2.5
4.8
8.2
8.0
4.8
10.5
8.1
4.9
10.5
8.1
4.3
4.7
3.9
6.3
11.0
15.0
16.1
17.5
14.5
12.6
8.9
2.4
3.8
2.3
1.2
2.5
2.5
jug/m3
2.4
.99
1.7
1.6
3.2
5.4
5.2
3.2
6.9
5.6
3.3
7.2
5.6
2.9
3.2
2.6
4.4
7.7
10.4
11.1
12.1
10.0
8.7
6.1
1.7
2.6
1.4
0.7
1.6
4.7
NOj,
mg
.012
.004
.008
.008
.007
.009
.055
.027
.016
.003
.047
.063
.011
.004
.007
.015
.009
.019
.043
.027
.018
.023
.018
.014
.076
.027
.045
.027
.050
.024
N
mg
5.7
3.3
1.8
2.6
2.2
2.2
1.0
1.3
0.6
0.6
1.5
2.0
0.5
0.9
1.2
1.0
1.6
1.3
0.5
1.1
1.2
1.3
1.2
2.0
2.3
5.9
6.3
4.2
5.2
2.2
Oj
Atg/m3
3.9
2.5
1.2
1.7
1.5
1.4
0.7
0.8
0.4
0.4
1.0
1.3
0.3
0.6
0.8
0.7
1.1
0.9
0.4
0.8
0.8
0.8
0.8
1.4
1.6
4.0
4.0
2.7
3.2
1.7
c,
mg
29.8
10.6
16.5
26.8
35.4
28.1
24.7
17.5
19.7
22.3
39.7
41.8
18.8
14.7
16.5
26.1
29.2
25.2
36.0
36.3
43.4
39.4
32.6
34.8
21.4
31.7
50.2
40.0
35.4
28.9
H,
mg
5.5
2.9
1.3
3.5
6.1
5.2
4.9
3.8
5.0
5.0
6.3
9.3
6.1
3.5
3.0
5.1
6.4
6.1
8.7
8.8
12.0
9.7
7.3
7.9
2.3
5.0
7.2
5.8
3.8
5.7
N.
mg
5.8
2.5
2.8
3.8
6.6
8.0
8.0
5.7
9.3
7.0
7.1
14.0
8.0
5.5
6.3
5.3
9.1
10.8
13.3
13.9
17.2
14.3
11.3
10.3
3.7
7.5
6.5
4.6
4.5
7.9
Mass
Loading,
115.8
54.9
65.1
81.7
98.9
96.1
87.0
64.5
79.7
77.6
118.1
135.8
76.6
63.6
64.7
81.8
105.5
94.1
123.7
127.4
145.8
137.3
116.5
122.7
73.1
106.3
198.2
123.7
105.6
101.4
-------
32
TABLE 12. ANALYSIS OF COMPOSITE WEST
COVINA DUST SAMPLE
Concentration,
Specie weight percent
NHj 0.63
NOj 0.001
N0$ 4.8
Total Carbon 12.6
Total Hydrogen 1.8
Total Nitrogen 2.2
-------
ANALYSIS AND INTERPRETATION
This section contains a statistical analysis and interpretation
of the field sampling results.
-------
33
ANALYSIS AND INTERPRETATION
St. Louis, Missouri
In this discussion of the St. Louis results, we will first review the trends in the air quality data,
then the aerosol results, and finally comment in some detail on the distribution and balance of nitrogen
compounds in the St. Louis atmosphere. Obviously, the data presented here were collected over only
a limited number of days during the summer of 1973. Any generalization of our results to other
seasons or other years must be regarded with some degree of skepticism.
Air Quality Data
The general behavior of the St. Louis air quality data appears to follow the classical pattern of
urban photochemical smog formation. Referring to the average profiles in Figure 2, a large increase
in both NO and N02 is noted between 6:00 a.m. and 8:00 a.m. After that time, NO drops rapidly
and N02 decreases, although somewhat more slowly, as 03 increases to a peak at 2:00 p.m. During
the late afternoon and evening, ozone drops off as its rate of production from photochemical reactions
falls below its rate of removal. Throughout this late afternoon period, the reaction of 03 with NO,
which is continually emitted in auto exhaust, is an important mechanism for the removal of both
species. Only when 03 has been reduced to a very low level is NO permitted to increase substantially.
The period between 10:30 and 11:30 p.m. is missing from the profiles due to the routine shutdown
of the mobile laboratory for calibration and maintenance at this time each day. However, the per-
formance of the various pollutants can be inferred from the trends prior to the 10:00 p.m. average and
after the midnight average. An interesting feature of the data is the gentle rise in O3 between midnight
and 4:00 a.m. A likely explanation of this behavior is that, with the rapid decrease of NO over this time
period, 03 transported into the area is not scavenged by NO so that the ozone concentration at our sam-
pling site increases. As NO begins to rise again after 4:00 a.m., ozone drops off steadily to its lowest level
at 7:00 a.m. Whether this nighttime ozone is from natural sources (e.g., transport from the stratosphere)
or from photochemical smog processes will be discussed shortly.
The average level of ozone provides ample evidence of photochemical smog activity during the
course of our study. Indeed, ozone exceeded the 1-hour maximum standard of 0.08 ppm on half of
the monitoring days.
The increases in NO and N02 late at night are not surprising when one considers the stable
meteorological conditions existing at that time (note the wind-speed profile in Figure 2) combined
with the fairly heavy traffic occurring in metropolitan areas until late at night during the summer
months. These are conditions under which one would expect the accumulation of primary pollutants.
The interference with the chemiluminescent NO2 determination by PAN and HN03, discussed
in the Experimental Methods section, is not likely to have a major impact on the average N02 con-
centration due to the very low average PAN and HNO3 levels encountered in St. Louis. However,
during the afternoon hours when PAN and HN03 are high, up to one-half of the observed N02 might
actually be PAN and HN03 on the average. Since the profile for NO2 does not fluctuate in the same
manner as the PAN and HN03 curves, however, the interference is probably not this great. The
variability in the HN03 response factor mentioned in the Experimental Methods discussion may ex-
plain the nonquantitative interference.
The average concentrations of PAN, HN03, and NH3 are profiled in Figure 2. The average
concentration of PAN in St. Louis was quite low during the period of our study. PAN reached its
peak during the day at 1:00 p.m., consistent with its association with photochemical smog reactions.
After 1:00 p.m., the concentration of PAN drops rather slowly, falling during the evening and through-
out the night. Since both PAN and 03 are accumulating products of photochemical smog processes,
it is instructive to compare their profiles. As might be expected, both 03 and PAN begin to increase
-------
34
at about the same time in the morning and, generally speaking, peak out together during the early after-
noon. During the late afternoon and evening, the concentration of ozone decreases much faster than
PAN, no doubt reflecting the rapid removal of ozone by the reaction with NO discussed earlier. As was
noted earlier for ozone, PAN also increases slightly between 1:00 and 3:00 a.m. While this increase ap-
pears to be only marginal in the 26 day - average profile, on several individual daily profiles (Appendix A)
very marked rises in early morning PAN and 03 are observed. The association of an increase in PAN with
the increase in ozone indicates a photochemical smog source rather than a natural source for this nighttime
ozone. The coinciding increase in the concentrations of these two photochemically generated contaminants
early in the morning gives very definite indication of the long-distance transport of a reacted air mass which
passes over St. Louis during the early morning hours. The subject of long-distance pollutant transport is
currently under investigation by Battelle-Columbus and several other organizations under the auspices of the
U.S. Environmental Protection Agency.
The average nitric acid profile shown in Figure 2 is, to our knowledge, the first report of nitric
acid in an urban atmosphere. The average concentration of nitric acid was rather low in St. Louis and
was somewhat variable, as noted by the several peaks in the profile. An examination of the profiles
shows little correlation between HN03 and any other parameter. Discussion of HN03 will be reserved
for the statistical analysis of the data treated later in this section.
Ammonia is another compound which has rarely been determined continuously in urban atmo-
spheres. The ammonia profile exhibits a slight but definite diurnal pattern with the average concentra-
tion of NH3 about 5 ppb. The decrease in ammonia between midnight and 3:00 a.m. does not reflect
the true behavior of NH3, but rather is an artifact of the NH3 calibration procedure. Desorption of
NH3 from the instrument after calibration is a lengthy process requiring several hours. For this reason,
the calibration was generally performed on Sunday evenings when data were not being collected. Un-
fortunately, the desorption process continued after midnight on three Sunday nights so that erroneously
high NH3 values resulted. These high ammonia values show up in the average profile of Figure 2 be-
tween midnight and 3:00 a.m. The more correct pattern of NH3 over this time period, as judged from
the individual daily ammonia profiles in Appendix A, is a gradual decrease and leveling off from about
6:00 p.m. in the evening until ammonia begins to rise again at about 9:00 a.m. the next morning. It
must be emphasized, again, that this is the average pattern and that individual profiles may differ sub-
stantially from the average. The pattern of NH3 is similar in a general way with the profiles of wind
speed and temperature. Since both the sources and sinks of ammonia are presumed to be biological in
nature, temperature especially may play a part in the diurnal pattern of ammonia.
It is interesting to note that there appears to be little correlation between the nitric acid and
ammonia profiles. It has often been suggested that these two species may react rapidly with one
another in urban atmospheres to form NH4NO3, thereby maintaining both gaseous species at extremely
low concentrations. While there is little evidence of such rapid removal processes here, the possibilities
will be reviewed later in sections detailing the West Covina results.
On three consecutive days in St. Louis, above-ground concentrations of NO, NO2, and NOX were
monitored using the Battelle-Northwest twin-engined airplane. The results of this vertical sampling
study are given in Table 6. The vertical sampling was conducted approximately above our St. Louis
University'ground site. The ground-level concentrations shown in the table are averages over the air-
plane sampling interval taken from our ground station. Both the ground station and airborne NOX
instruments were spanned with the same cylinder of calibration gas. On July 30, the temperature pro-
file gives no indication of inversion conditions. The NO concentration at 1500 feet is almost three
times the ground level NO; from 1500 to 4500 feet the concentration.decreases at a fairly constant
rate. At 4500 feet the NO concentration is still higher than the ground station level. The concentra-
tion of N02 is slightly lower at 1500 feet than on the ground. Above 1500 feet the NO2 level de-
creases regularly up to 3500 feet and then remains constant at 10 ppb.
On both July 31 and August 1, we observed the same pattern of higher NOX levels aloft. The
temperature profile for the 31st indicates a temperature inversion between 2500 and 4000 feet. The
decrease in temperature between 1000 and 2500 feet was also rather dramatic.
-------
35
The surprising feature of these vertical data is the high level of NOX found at 1000 feet and above.
On the 30th and 31st, the NOX levels did not approach ground concentrations until 3500 feet or higher.
During one flight the plane flew through a partially dispersed stack plume which we observed visually.
At the same time, the NOX measurements Increased beyond the range of the monitoring instrument
(10 ppm). It may well be that these high vertical NOX readings are caused by the dispersing plume
from a stack in the vicinity of our St. Louis University site.
Aerosol Composition
The average particulate mass loading during our St. Louis field study was approximately 58 M9/m3
as determined by high-volume sampling. It was pointed out earlier that high-purity quartz fiber filters
were employed for high-volume sampling. Referring to Table 7, we note that carbon on the average,
makes up about 19 percent of the total aerosol mass, hydrogen about 3.7 percent, and nitrogen roughly
4.7 percent. These three elements alone make up 27 percent of the average aerosol mass. It is
assumed that water, metals, oxygen, and sulfur comprise most of the remaining mass. Of the nitrogen
compounds determined specifically, ammonium at 2.7 /ig/m3 accounts for 4.7 percent of the mass,
nitrate at 0.36 jug/m3 makes up about 0.6 percent, and the contribution of nitrite to the total mass
of aerosol is negligible.
Table 13 gives the comparison between the average total aerosol analysis and the composite dust
(particle diameter > 2.5 nm) analysis. Both ammonium and total nitrogen are seen to be associated
primarily with smaller particles. This is consistent with other reports'31' of NHj as a submicron
aerosol from a gas-phase source. Carbon and hydrogen are also more dominant in the small particles.
Surprisingly, nitrate is much more prevalent in the larger particles. This is contrary to studies'31'
which have found nitrate predominately in the submicron aerosol range, indicating a gaseous source.
Indeed, not only is the distribution of nitrate at variance with earlier studies, but the absolute concen-
tration of nitrate also appears anomalously low.'59' A discussion of a collaborative study to confirm
the accuracy of our nitrate analyses is discussed with the West Covina results. We attribute these
nitrate anomalies to the inefficient removal of gaseous or submicron nitric acid aerosols by quartz
fiber filters as opposed to basic-surfaced glass fiber filters. The average nitric acid we measured in
St. Louis (0.003 ppm), if all removed by the high-volume filter, would account for almost 8 /Ltg/m3 addi-
tional nitrate, bringing our nitrate values into correspondence with earlier reports. Although we would
not expect all of the nitric acid to be removed by high-volume samplers, removal of even a small fraction
would drastically affect the apparent distribution and absolute concentration of nitrate aerosol. A brief
laboratory study seems to confirm the fact that glass fiber filters remove nitric acid from air streams,
while high-purity quartz filters do not. Because of the possible health effects of respirable nitrate
aerosols, this is a rather important subject area, in which research is continuing at Battelle-Columbus.
TABLE 13. ST. LOUIS AEROSOL COMPOSITION (WEIGHT PERCENT)
NHj NO; NOg _C_ _H_ _N_
Average Total Aerosol Composition 4.7 - 0.62 19.0 3.6 4.6
Large Particle (> 2.5 (Jtm) Composition 0.55 0.001 2.6 14.6 1.8 1.5
A recent CRC-EPA program at Battelle-Columbus has reported'32' NH^, NO^, carbon, and hydro-
gen in two particle-size ranges for aerosols collected in Columbus, Ohio, and New York City. These
results are shown in Table 14. The large-particle compositions from both Columbus and New York are
very similar to our St. Louis results shown earlier in Table 13. Comparison of the small-particle
-------
36
TABLE 14. SIZE CLASSIFICATION OF VARIOUS AEROSOL
COMPONENTS BY CITY (WEIGHT PERCENT)
Columbus
Aerosol Diameter (^m)
Constituent
Carbon
Hydrogen
Nitrate
Ammonium
<2
17
6
1
6
>2
17
1
2
0
New
<2
26
6
4
7
York
> 2
16
1
3
0
composition is not straightforward, since the St. Louis data are for total aerosol only, with small-diameter
particles not separated as a class. However, similar trends in the three cities can be inferred from the
large-particle/total mass distribution. In all three cities, ammonium was associated almost totally with
small particles, while in both Columbus and St. Louis nitrate appeared at higher percentages in the
large-particle range. In New York, the nitrate aerosol distribution favors the small particles, but only
slightly. A comparison of the St. Louis and West Covina C, H, N data for the two size ranges is re-
served for the West Covina aerosol discussion.
The objective of this program involves ultimately a determination of the fate of nitrogen oxides
in the atmosphere. One important step toward this goal is a determination of the total nitrogen which
can be accounted for as aerosol. Since "total aerosol nitrogen" analyses were performed on the filter
samples collected for the study, it is possible to determine a filter nitrogen balance to ascertain to what
extent known nitrogen compounds can account for the total aerosol nitrogen. This is important, since
it will enable us to determine whether some unsuspected nitrogen-containing aerosol might be a signifi-
cant sink for the nitrogen oxides.
The filter nitrogen balance results are presented in Table 15. For some days, more than one filter
has been analyzed so that the total number of filter samples shown is greater than the data from earlier
tables. The values in the table are in milligrams of nitrogen per filter. The two important columns are
(1) the sum of the ammonium and nitrate nitrogen and (2) the independently determined total nitrogen
figures. The final column shows the percentage of the total aerosol nitrogen which can be accounted
for as ammonium and nitrate. While the figures fluctuate from day to day, the average is about 82 per-
cent. Thus, only 18 percent of the total aerosol nitrogen is unaccounted for. This amounts to an
average of less than one ppb of aerosol nitrogen unaccounted for. This figure is negligible when the
total nitrogen balance in urban atmospheres is considered. The concentration of nitrite found in bur
aerosol samples was extremely low and is also considered to be negligible in terms of overall nitrogen
balance.
From the filter nitrogen data just discussed there is no indication that any previously unsuspected
nitrogen-containing aerosol is playing a significant role in the fate of nitrogen oxides in the St. Louis
atmosphere. Certainly, over the longer term after an air mass is well beyond the bounds of an urban
area, conversion of both primary and secondary gaseous nitrogen products to nitrogen aerosols is no
doubt very important. However, over the short interval during which an air mass resides in a metro-
politan area, these filter data suggest that we must look for gaseous nitrogen reaction products or
physical removal processes to explain any nitrogen imbalance.
-------
37
TABLE 15. ST. LOUIS FILTER NITROGEN BALANCE
Date
7/18
7/19
7/20
7/22
7/23
7/24
7/25
7/26
7/27
7/30
7/31
8/1
8/2
8/3
8/4
8/5
8/6
8/7
8/8
8/9
8/10
8/12
8/13
8/14
8/15
8/16
Filter
2
1
2
1
2
2
2
2
1
2
2
2
1
1
2
2
2
2
2
1
2
2
2
2
2
2
2
2
2
2
NHj-N,
mg
9.02
4.04
5.76
.93
3.03
3.42
2.18
2.18
.78
1.01
1.48
2.18
1.32
1.71
2.02
1.63
1.32
4.74
5.83
5.76
7.23
6.38
1.24
.16
1.56
1.24
3.73
3.19
6.92
8.71
NOg-N,
mg
.14
.11
.17
.16
.09
.10
.14
.10
.07
.09
.17
.12
.11
.02
.10
.09
.18
.23
.10
.14
.10
.20
.14
.19
.11
.06
.09
.07
.16
.18
SNHi-N + NOg-N, Total N, ZNHj-N + NOg-N „ ,AA
mg
9.16
4.15
5.93
1.09
3.12
3.52
2.32
2.28
.85
1.10
1.65
2.30
1.43
1.73
2.12
1.72
1.50
4.97
5.93
5.90
7.33
6.58
1^38
.35
1.67
1.30
3.82
3.26
7.08
8.89
mg
9.3
6.4
6.7
1.6
3.6
4.5
2.5
2.7
1.9
1.7
2.2
2.8
1.9
2.8
2.6
2.1
1.7
5.0
5.3
6.1
7.2
7.1
1.3
.7
2.4
1.6
4.1
4.1
7.9
10.8
Average
Total N
98.5
64.9
88.5
68.0
86.7
78.2
92.8
84.4
44.6
64.7
75.0
82.3
75.3
61.8
81.4
81.7
88.3
99.3
111.9
96.6
101.8
92.7
106.2
49.6
72.7
81.2
93.2
79.5
89.6
82.3
82.5%
-------
38
Distribution and Balance of Nitrogen Compounds
The ultimate aim of this program is to determine the fate of nitrogen oxides emitted to the
atmosphere. This objective requires that the distribution of trace nitrogen compounds in the atmo-
sphere be determined and that relationships among the nitrogen species and between these species
and other atmospheric parameters be derived.
Essentially, the distribution of gaseous nitrogen compounds has been given as averages in
Table 4, as composited time-dependent profiles in Figure 2, and as individual daily profiles in
Appendix A.
On an average basis, the sum of PAN and nitric acid concentrations represent only a very small
fraction of the total average NOX. Since it was shown in the previous section that total aerosol nitrogen
is also extremely low in comparison to the average level of NOX, we are left with two hypotheses re-
lating to the fate of nitrogen oxides in urban atmospheres. In the first hypothesis, it is assumed that
nitrogen oxides are being-removed from urban atmospheres at a fairly rapid rate and we have not deter-
mined the dominant removal processes. The second hypothesis holds that the reactions removing
nitrogen oxides from the atmosphere are rather slow and are not important over the short time interval
during which an air parcel resides in an urban area. By this argument, we should not expect to find
high concentrations of nitrogen-containing products, since NOX scavenging processes would be too slow
to allow significant quantities of reaction products to accumulate.
In order to test the validity of these two hypotheses, it is necessary to determine the fraction of
NOX removed from the atmosphere and then attempt to "balance" the nitrogen oxides reactants with
their chemical reaction products. The nitrogen balance will define the extent to which chemical reaction
products can account for the removal of nitrogen oxides from an air mass. If poor nitrogen balances
result, we must consider the possibility of unknown chemical reaction products or physical removal
mechanisms as potential sinks for the nitrogen oxides.
Statistical Interpretation. Various statistical techniques will be employed to determine relation-
ships among the many chemical and meteorological parameters measured during the field study. In
particular, relationships will be sought to define the conditions under which loss of nitrogen oxides
occurs. An understanding of the conditions necessary for NOX removal from the atmosphere should
in itself indicate the types of sinks and removal mechanisms which are important.
The initial scanning of the data was accomplished using the Automatic Interaction Detector (AID)
statistical program.'60' In essence, this technique identifies which independent variables are the best
predictors of a given dependent variable. In quantitative terms, the best predictor is that independent
variable which maximizes an F-ratio. In intuitive terms, the F-ratio is the ratio of the statistical varia-
bility of y that is accounted for by the variability of x, to the variability of y that is not accounted for
by the variability of x. The AID program simply computes the F-ratios associated with each inde-
pendent variable and splits the data using the variable that yields the maximum F-ratio.
Brief descriptions of the AID graphic-tree output and the initial series of AID runs are given in
Appendix C. The significance of the AID splits is related to the size of the data base, the extent
(graphically, the width) of the split, the number of times splitting occurs on any one independent vari-
able, and, ultimately, the between sum of squares to total sum of squares ratio. Table 16 lists several
dependent variables and the first three independent or predictor variables for each. The sign of the
predictor variables is also listed; positive indicates that high values of the predictor variable predict high
values of the dependent variables and negative indicates that low values of the independent variable
predict high values of the dependent variable.
While the results of the AID runs are interesting in themselves, the real usefulness of the AID analy-
sis lies in its ability to sort a large volume of data quickly and indicate which relationships appear to
be most important. Of primary concern in this program are the variables NOX, PAN, and HN03.
-------
Dependent Variable
39
TABLE 16. ST. LOUIS AID RESULTS
First Three Predictor Variables
Three Primary Predictor Variables
NOX
PAN
HN03
NH3
NHj
N03
Total Aerosol Nitrogen
Mass Loading
03(-)
03(+)
N02 (+)
Temperature (+)
Total Aerosol Nitrogen (+)
NO 2 (+)
NH| (+)
Total Aerosol Nitrogen (+)
HN03 (+)
Wind Direction (-)
Wind Direction (-)
NOX (+)
-
N02 (-)
Wind Speed (-)
HN03 (+)
Humidity (-)
N02 (+)
03(+)
N02 (-)
-
Total Aerosol Nitrogen (-)
-
N02 (+)
Ozone appears as a primary predictor variable for all three of these compounds. Ozone and NOX appear to
ba negatively correlated, while a positive relationship is maintained between ozone and the two NOX
reaction products, PAN and HN03. These are precisely the types of relationships one would expect from
kinetic mechanisms used for atmospheric modeling. High levels of ozone are expected to increase the
rate of removal of NOX and the rates of production of PAN and HN03 from NOX. The positive rela-
tionship between N02 and both PAN and HN03 is also predictable from kinetic models, since N02 is
vitally important in the generation of 03 and is necessary for the formation of both PAN and HN03.
The negative sign accompanying wind direction as a PAN and HN03 predictor indicates that winds from
the northeastern and southeastern quadrants (0° to 180°) lead to higher levels of PAN and HN03 than
do other wind directions. Since the most heavily industrialized sections of St. Louis and East St. Louis,
Illinois, lie to the east of our monitoring site, this result is not surprising.
The positive relationship between temperature and NH3 was discussed earlier as possibly relating
to the biological nature of the ammonia cycle. The positive NH^-total aerosol nitrogen relationship is
understandable, since NH£ makes up the greatest fraction of the aerosol nitrogen mass. Other relation-
ships indicated in the table are more or less explainable, based on our current understanding of atmo-
spheric chemical processes. Since they are not directly pertinent to the objectives of this project, how-
ever, they will not be discussed here.
It was pointed out earlier that the main utility of the AID program is to indicate potentially
important relationships and thereby serve as a guide for further data analysis. Based on the results of
the AID analyses, several regressions have been run to further define the nature of the relationships
primarily among NOX, 03, PAN, HN03, and relative humidity.
The first set of regression parameters is shown in Table 17 for the hourly average data for 23 hours
each day. In this and subsequent regression tables, the first two columns show the Y and X coordinate
variables, while the third column lists the correlation coefficient (R). In Table 17 the only relationships
which stand out are those between 03 and NO, N02, and NOX. While none of these correlation co-
efficients is particularly high, a negative slope is apparent in all three cases. The statistical significance of
the 03-N02 result is only marginal, however.
-------
40
TABLE 17. ST. LOUIS REGRESSIONS
All Hourly Averages.
y
NOX
NOX
NOX
°3
03
03
X
R.H.
03
HN03
NO
N02
HNO3
R
0.35
-0.54
0.21
-0.51
-0.33
0.04
An interesting comparison, and in some cases a strengthened statistical relationship, is obtained
when only the daytime data are used in deriving the regression equations. Table 18 shows the same
correlation terms given in Table 17, along with some additional regressions for the daily 6:00 a.m. to
6:00 p.m. time interval. Comparison of the three most important NOX predictor variables (from the
AID analysis) in Table 17 for all hours of the day with the same variables in Table 18 for daylight
TABLE 18. ST. LOUIS REGRESSIONS
Data From 6 a.m. to 6 p.m.
y
NOX
NOX
NOX
°3
03
°3
03
HN03
HN03
HN03
X
R.H.
03
HN03
NO
NO 2
PAN
HN03
Global
NO 2
PAN
R
0.51
-0.58
0.22
-0.61
-0.32
' 0.48
0.00
-0.02
0.41
0.26
hours only shows that only the correlation coefficient for N0x-relative humidity changed substantially.
This may only reflect the fact that there was very little variation in relative humidity during the night-
time hours. The highest level of correlation, again, is between O3 and NO with a coefficient of -0.61.
The inverse relationship between O3 and NO is expected because of the very rapid reaction between
these two trace pollutants. If either O3 or NO is present in air at a high level, the other must neces-
sarily be present at a rather low level. The often used photostationary-state assumption defines the
inverse relationship between 03 and NO as follows:
NO2
-------
41
where K is a function of ultraviolet light intensity. Obviously, the assumption collapses at night. The
validity of the assumption is also questionable under low light intensities and at low concentrations due
to the slower reaction rates.
An Important relationship appears to exist between NOX and 03, since this regression equation
yields the second highest correlation coefficient. However, because of the opposite effects NO and N02
have on ozone concentration as seen in the photostationary-state equation, an explanation of the relation-
ship between 03 and the sum of NO and N02 is not straightforward. Possibly, a negative relationship
exists because 03 is normally quite low in the morning when NO and, consequently, NOX are high; in
the afternoon when 03 accumulates, the high morning levels of NOX have dispersed and the NOX pres-
ent is largely N02. The problem of interpreting these 03-NO-N02-NOX relationships is further compli-
cated by the fact that hydrocarbon data are not included here. Indeed, we know from smog chamber
and atmospheric data that the hydrocarbon to NOX ratio (HC/NOX) is an important determinant of the
potential 03 concentration. To further analyze the 03-IMOX relationship at this point would require
the addition of hydrocarbon data and a discussion of kinetic models. Such a discussion would shed very
little light on the fate of nitrogen oxides and is not warranted here.
A third set of regression!) has been run on the St. Louis data and is shown in Table 19. The first
four regressions include the maximum hourly averages for O3, PAN, and HN03 versus the 6:00-9:00 a.m.
averages for either NO or NOX, The most interesting feature of this series of regressions is the reason-
able correlation found between maximum HN03 and morning NOX.
TABLE 19. ST. LOUIS REGRESSIONS
Maximum Hourly 03, PAN, HN03
vs.
6:00-9:00 a.m. Average NO or NOX
NO 0.16
NOX 0.25
NOX PAN 0.31
NOX HN03 0.62
PAN 0.50
HN03 0.06
Two additional regression; also shown in Table 19 are for the maximum hourly average 03 versus
maximum hourly average PAN $nd HNO3. The 03-HN03 correlation is virtually nonexistent; however,
the O3-PAN regression yields a correlation coefficient of 0.5. The slope of the O3-PAN regression is 5.1
compared with a value of approximately 3 found for many U.S. cities in studies by Lonneman.'61'
Nitrogen Balance. In keeping with the objectives of the program, we have determined the atmo-
spheric distribution of trace nitrogen compounds which are thought to be important to the fate of
nitrogen oxides and investigated the interrelationships and time dependencies of the various compounds
for clues as to the ultimate disposition of the nitrogen oxides. From the distributions it was clear that
the sum of all the measured products of nitrogen oxide reactions could account for only a fraction of
the average NOX concentration so that either (1) we were not observing all of the important NOX
-------
42
removal mechanisms or (2) the NOX removal mechanisms are relatively slow, so that the loss of NOX
over the short term is only a small fraction of the total NOX concentration. Subsequently, AID and
correlation analyses have indicated that 03, relative humidity, and HNO3 are the three most important
predictors of NOX concentration of all the parameters determined in this study. From the correlation
studies, 03 appeared to be an especially important predictor of both the NOX and PAN concentrations.
Two cautionary notes must be sounded here, however. First, the statistical relationships cannot be
translated into chemical relationships; they can only indicate the importance of trends in the data.
Inferring chemical meaning or cause-and-effect relationships from the statistical analysis requires extreme
caution. Second, it must be kept in mind that the major factors affecting the NOX concentration are
related to the intensity of source emissions, inversion conditions, convection, and advection. In our
examination of the trends in NOX concentration for clues as to the NOX removal processes, these gross
determinants of NOX concentration are no doubt masking some of the relationships we seek. What is
needed is a means of computing the fraction of NOX which has been removed from the air mass at any
point in time. Such "NOX loss" values would yield much more meaningful relationships since the
masking effect of emission rates and dispersion would be eliminated. In addition, the "NOX loss" val-
ues can be compared with the sum of the measured NOX reaction products to determine a nitrogen
balance. The closeness of the balance will reveal to what extent we are accounting for the NOX which
has been removed from an air mass by either chemical or physical processes.
In deriving "NOX loss" values, we will make use of the inert tracer technique described in an
earlier section of this report. The tracers we plan to use are carbon monoxide and acetylene, both of
which were measured gas chromatographically by the Environmental Protection Agency mobile labora-
tory situated alongside our mobile lab in St. Louis. Values for CO and C2H2, integrated over 2-hour
periods from 6:00 a.m. to 2:00 p.m., are available for 5 days during which our mobile laboratory took
samples in St. Louis. These data were presented earlier in Table 5. Detailed hydrocarbon data pro-
vided by EPA are given in Appendix A.
With CO as the tracer, the equation used to calculate "NOX loss" is
(CO) measured
(CO/NOx)emjssjon inventory
measured "N0x "oss , (1)
where the first term represents a theoretical or predicted NOX concentration based on emission ratios
and the measured CO at any given time. The difference between the predicted NOX and the measured
NOX is defined as NOX loss. The equation can be used in this form to calculate "NOX loss" from
individual sets of measured CO and NOX data. However, in using the equation as it stands, we over-
estimate the NOX loss due to the presence of background CO, which is in no way associated with
St. Louis emissions sources. For the most accurate determination of "NOX loss", we must eliminate
any CO which is not associated with the emission inventory estimates. A convenient way of doing this
involves rearrangement of Equation (1) as follows:
[(COTNoTi1!. - 1j
-------
43
'I'OVM' CO
C2H2
8.10
2.13 +
I
I
I
I
I
I
I
1
1.75 +
I
I
I
1.55 +
I
CO, ppm I
I
I
1.36 +
I
I
I
I
1.17 +
I
I
I
0.97 +
I
I «•
I «
1 +
0.73 -»•
I
I
I
I
0.58 + «
I «•
I
I - -
i IGHIFICRNCE P -
INTERCEPT -
TICMIFICPMCE fl -
m- ERROR CF P -
15.70 23.30 30.90 38.50
+ .
•
•
•
*
w
y^
* 3
yj
• • g
'
z
O
• w
§
LU
K
W
* 0
A t=
* . • . ^
trf*
^^
LU
s
3
(i
'0 19.50 27-10 34.70
C2H2, ppbC
0.88552 R IPUPPET - 0.78415
0.00001 STD ERR OF E?T - 0.2543?
0.30975 ?TP ERROR OF R - 0. 1279€
0.01314 iLOPE -'.P> - 0.05375
0.00665 riGWIFICRNCE F - 0.00001
-------
44
very good correlation between these two pollutants. The intercept of the curve — at 0.31 ppm CO — will
be important to our subsequent discussion, since it is this fraction of CO which Is not associated with
auto exhaust and which we shall consider as background CO. It is Interesting to note that Kopczynski,
et al.'621, have measured nonurban CO in the St. Louis area at 0.295 ppm. It is very likely this non-
urban CO which comprises our background or intercept CO. The geologic background from natural
processes alone is thought'63'to contribute 0.1-0.2 ppm CO.
The important plot of CO versus NOX is shown as Figure 5. The CO intercept of 0.41 ± 0.20,
again, is quite similar to the intercept from Figure 4 and the nonurban CO reported by Kopczynski,
et al. '®2) The most important feature of this curve is the slope, however. It is this slope which will
be used in the calculation of the average "NOX loss". Also necessary for the calculation are the most
recent emissions inventory values* for CO and NOX, which are given in Table 20 for the city of St. Louis
and the St. Louis Air Quality Control Region (AQCR).
TABLE 20. EMISSIONS INVENTORIES FOR ST. LOUIS REGION
Tons per Day.
CO
NOX
St. Louis City
299,088
35,545
AQCR
3,852,746
433,547
Substituting the measured CO/NOX ratio from the slope of Figure 5, the molar CO/NOX ratio from
the St. Louis AQCR, and the average measured NOX into Equation (2),
NOX loss =— '' ' 1 (0-056)
NOX loss = 0.000 ± 0.012 ppm.
If we use the St. Louis city emission inventory ratio, we obtain
NOX loss =;14'6;1»'2 - 11 (0.056)
" I
NOX loss = 0.003 ±0.012 ppm.
It is apparent that the average NOX loss is very small in comparison with the average concentration
of NOX. The average sum of PAN and HN03 over the same 5-day period was 0.007 ppm, well within
the 0.012 ppm deviation. Thus, the nitrogen balance is quite good.
As a secondary check on NOX balance, we can use C2H2 as a tracer and recalculate the "NOX
loss". The pertinent equation is
f (C2H2/NOx
m
(C2H2/NOX)E...
* Courtesy of C. Masters, U.S. Environmental Protection Agency.
-------
cc
1.94
1.75
1.55
CO, ppm
1.36
1.17
0.97
0.78
0.58
ft OQ
v • «-• •
45
0.03
0.05
.-4. — —
0.07
—4. 4..
0.09 0.11
.—+ 4. 4..
£.13 «•
I
I
I
I
•f
I
I
I
I
•f
I
I
I
I
I
I
I
*
1
I
I
I
4-
; i
• \
0.03 0.04 0.06 0.08 0.10
NOX, ppm
CORRELPTIOK - 0.73454 R JOURREP
?IGNIFICRMCE P - 0.00011 ?TP ERR CF EST -
INTERCEPT ''FO 0.41178 ?TP EPPCP OF P -
CRNCE P - O.OSeC'O ?LOPE '.B>
ERROR OF B - 3.1845£ SiehlFICPUCE B -
0.53955
0.37155
0.19789
14.62534
0.00011
FIGURE 5. ST. LOUIS REGRESSION: CO VERSUS NOX
-------
46
It must be remembered that acetylene is present in the atmosphere at much lower concentrations than
CO, so the error in its measurement is correspondingly larger. The measured C2H2/NOX ratio for the
5 days of available data is obtained from the slope of Figure 6 and found to be 0.20 ± 0.06. The emis-
sion inventory ratio needed for our calculation presents us with a problem, however, since inventories
are not broken down for individual hydrocarbons. We can circumvent this problem by using the
C2H2/NOX ratio for auto exhaust after correction for the nonexhaust fraction of NOX for St. Louis.
Using Kopczynski, et al.'s'621, value of 28 for the CO/NOX ratio in St. Louis auto exhaust, we calculate
that about 60 percent of our measured NOX is attributable on the average to auto exhaust. The back-
ground CO again has been subtracted to improve the accuracy of the calculation.
The C2H2/NOX ratio for St. Louis auto exhaust has been determined by Kopczynski, et al. *62' ,
to be 0.38. Making the usual assumption that auto exhaust is the only source of C2H2, we can correct
the C2H2/NOX exhaust ratio for nonexhaust sources of NOX, so the ratio will be representative of all
St. Louis sources. The corrected ratio is 0.38 x 0.60 or 0.23. Substituting this value for the emissions
inventory ratio in Equation (3) yields
[0.20 ± Q.06
0.23
NOX loss = | "^Voo' - 1 I (0.056)
NOX loss =-0.007 ± 0.015
This calculated average "NOX loss" is in essential agreement with the earlier CO-tracer calculation in that
the average loss of NOX is found to be negligible.
Prior to our determination of the absolute value of NOX loss, AID analyses were performed on
relative "NOX loss" values to determine what factors might be influencing the removal of NOX from the
St. Louis atmosphere. These AID runs incorporated the EPA hydrocarbon data given earlier in Table 5.
The results of these runs were contradictory and somewhat confusing. It is obvious now that, with no
true NOX loss, the AID program was splitting the dependent variable (relative NOX loss) primarily on
noise. Consequently, these runs have no significance and are not presented here.
In summarizing the St. Louis results, we have found that the loss of nitrogen oxides from the
St. Louis atmosphere was minimal on the average and that PAN and nitric acid, which were present at
low levels, can account for what little NOX loss was observed. Nitrite and nitrate aerosol were found to
be unimportant in terms of nitrogen balance.
These findings support the contention that the removal of NOX from the atmosphere is a relatively
slow process requiring considerable residence or reaction time. The fact that our mobile laboratory was
situated within the City of St. Louis meant that the air mass we were sampling normally experienced
only a very short residence time over St. Louis under average wind conditions. If, indeed, the removal
of NOX from the atmosphere is a slow process, a more favorable location for observing NOX loss and
NOX reaction products should be downwind of an urban area where a well-aged air mass is encountered.
This was, in fact, our siting arrangement in the Los Angeles field study, the results of which are dis-
cussed in the following section.
-------
C8H2
47
0.03 0.05 0.07 0.09 0.11
38.50 * *
34.70
11.90
8.10
4.30
30.90
27.10
,» PPbC
23.30
19.50
15.70
I
I
4
I
I
I
I
+
I
I
I
I
•f
I
I
I
I
*•
I
I
I
I
+
I
I
I
I
0.0£
•
0. 04
0.0*
NOX , ppm
0.08
0. 10
CCRPELPTICN f.9.>-
?ieMIFICfiMCE R -
INTERCEPT <:P:> -
S-ieMIFICRMCE P -
TTD ERRCP CF F -
0.61289 P SCll'RPEI'
0.00203 STD ERR CF EST -
5.90584 ?TP EPRPP DF P -
0.0c861 SLCPE <*>
61.09468 SIGMIFICPMCE. B -
0.375*3
7.12815
3.79653
201.04901
0.00203
FIGURE 6. ST. LOUIS REGRESSION: C2H2 VERSUS NO,
-------
48
West Covina, California
In discussing the results of the West Covina field sampling effort, we will first review the trends in
the air quality data, then the aerosol results, and, finally, comment in detail on the distribution and
balance of nitrogen compounds in the West Covina atmosphere.
Air Quality Data
During the 29 days of monitoring in the Los Angeles basin, the symptoms of photochemical smog
were extensive, as evidenced by Table 10. Even though a great many days were heavily overcast until
almost noon, the Federal standard of 0.08 ppm 1-hour average ozone was exceeded on 27 of the 29
sampling days. Of the two substandard days, one was a heavily overcast Monday and the other a clear
Sunday.
In general, the concentrations of NO, NOX, O3, and PAN were considerably higher in West Covina
than in St. Louis, even though the average solar intensity was greater for St. Louis. A comparison of
the overall averages and the' average 1-hour maxima for the two cities is given in Table 21.
TABLE 21. ST. LOUIS-WEST COVINA COMPARISONS
Solar Intensity (cal./cm2/min.)
Wind Speed (m.p.h.)
Temperature (°C)
Relative Humidity (%)
O3 (ppm)
NO (ppm)
NO 2 (ppm)
N0x(ppm)
NH3 (ppm)
PAN (ppm)
HN03 (ppm)
St. Louis
Overall Average
0.314
3.84
25.19
78.39
0.035
0.016
0.029
0.053
0.005
0.002
0.003
Average
1-Hour
Maximum
—
—
— .
—
0.073
0.035
0.037
0.069
0.006
0.003
0.008
West Covina
Overall Average
0.242
2.13
18.46
77.15
0.046
0.058
0.078
0.136
0.004
0.009
0.003
Average
1-Hour
Maximum
—
—
—
—
0.153
0.128
0.097
0.185
0.005
0.020
0.010
The behavior of the West Covina air quality data follows the same general pattern as was noted in
St. Louis. Referring to the average air quality trends in Figure 3, a large increase in NO and NOX is
observed between 5:00 and 7:00 a.m., followed by a rapid decrease in NO and corresponding increase
in NO2. Nitrogen dioxide peaks at about 10:00 a.m. and remains fairly close to this peak level until
late in the afternoon. As in St. Louis, the average N02 concentration is high enough that the sum of
the PAN and HNO3 interferences can make up only a small fraction of the total N02. Again, however,
when PAN and HNO3 are high, as during the afternoon, up to 30 percent of the average measured NO2
-------
49
could actually be PAN and HN03. Elimination of this interference would cause an afternoon dip in the
average N02 profile rather than the almost constant afternoon concentration observed. Since these N02
and NOX values will not be used in the subsequent nitrogen balance discussions and since correction
of the data would be difficult due to the variable interference response, we have elected not to attempt
a correction of the chemiluminescent N02-NOX data.
During the afternoon period, NO drops to a very low level and remains there throughout the after-
noon, as ozone rises rapidly to a peak at about 3:00 p.m. It is noted in Figure 3 that 03 follows the
solar intensity profile quite closely with 1 to 2 hours' delay. The concentrations of PAN and HN03
appear to start increasing during the conversion of NO to N02 and follow the same pattern as the ozone
profile. Both PAN and HN03 reach their maximum levels at about 2:00 p.m. and thereafter drop off
through the evening. The concentration of PAN seems to decrease at about the same rate as 03; how-
ever, HN03 drops off faster than either PAN or 03. This implies either that the processes forming
HN03 terminate more quickly than trie O3 or PAN formation processes or that HN03 removal mech-
anisms are more rapid.
Late in the afternoon (starting at approximately 5:00 p.m.), 03 decreases rapidly as NO from
evening rush-hour exhaust reacts with it to form 02 and N02. During this period, N02 increases to its
highest level at 7:00 p.m. Once O3 has been depleted, NO rises rapidly starting at 7:00 p.m. Wind
speed, also shown in Figure 3, peaks at about 5:00 p.m. on the average, and the mixing thus created
prevents the concentrations of exhaust products from the evening rush hour from building up to the
same high levels noted during the morning rush hour.
The average nitric acid profile in Figure 3 exhibits much less variability than the St. Louis pattern.
Within the limits just discussed, it appears to follow the same trend as the 03 and PAN concentrations,
which strongly implicates photochemical smog reactions in the formation process. The profile is con-
sistent with either of the two mechanisms shown below:
OH + NOj^HNOg
or
03 + N02 -»-02 + NO3
N205 + H20 -* 2HN03 .
However, the results of some of our recent smog-chamber experiments'141 indicate the latter mechanism
to be dominant.
The average ammonia profile, as shown in Figure 3, is similar to that of St. Louis both in concen-
tration and pattern. The levels and patterns of ammonia in both St. Louis and West Covina are com-
pletely consistent with the Great Britain'231 and St. Louis'24' ammonia results discussed in an earlier
section of this report. It is interesting to note that during the late morning and early afternoon hours,
the average behavior of NH3 and HN03 is similar. This does not tend to support the argument for a
very rapid reaction between these two substances; however, it cannot rule out such a reaction since
very rapid emissions or formation of NH3 and HNO3 could lead to the type of behavior observed. The
facts that the concentration of HN03 drops off more rapidly than other smog products and that NH3
is decreasing at the same time, further complicate the picture. At the present time, all one can say is
that NH3 and HN03 do coexist in the atmosphere and that the general time dependence of their con-
centrations is similar in the West Covina atmosphere.
Aerosol Composition
The average paniculate mass loading during the West Covina field sampling study was 101 M9/m3,
as determined by high-volume sampling of air through high-purity quartz fiber filters. Of this total
-------
50
aerosol mass, carbon comprises 19.4 percent, hydrogen 3.8 percent, and nitrogen 5.3 percent. These
three elements alone make up 28.5 percent of the average aerosol mass; water, metals, oxygen, and
sulfur presumably make up most of the remaining mass. The carbon and nitrogen percentages are in
good agreement with the values reported by Hidy(64)of 19 percent carbon and 5 percent nitrogen. Of
the nitrogen compounds determined specifically, ammonium accounts for 4.7 percent of the mass, ni-
trate about 1.7 percent, and the contribution of nitrite to the total mass is negligible, as in St. Louis.
Table 22 gives the comparison between the average total aerosol analysis and the composite dust
(particle diameter > 2.5 jum) analysis for West Covina. The aerosol composition from West Covina can
be compared with the composition of St. Louis aerosol shown earlier in Table 13. The weight per-
centages and relative distribution of size ranges is very similar for all components from the two cities
with the exception of nitrate. The fraction of aerosol mass accounted for as nitrate rose sharply from
St. Louis to West Covina. The NOs fraction increased by a factor of 2.7 for the total aerosol mass,
while an increase of a factor of 1.8 occurred in the large-particle size range. High nitrate measurements
in the Los Angeles area are not surprising; indeed, NASM(59) data have shown high nitrate in the Los
Angeles basin for years. What is surprising at first glance is the relatively low nitrate concentration found
in this study relative to Los Angeles nitrate values reported elsewhere. For example, in the study of
Gordon and Bryan(33), nitrate concentrations on the order of 10-15 M9/m3 were found from 1970
through 1972, compared with our average 1.7^g/m3 concentration. To check the accuracy of our
nitrate analyses to insure that the analytical technique was not responsible for the low nitrate values,
several filter sections from St. Louis and West Covina were sent to the California Air Industrial Hygiene
Laboratory* (AIHL) and to Rockwell International* for comparative analysis. Collaborative ammonium
analyses were also carried out at the same time by the AIHL group. The results of this collaborative
TABLE 22. WEST COVINA AEROSOL COMPOSITION (WEIGHT PERCENT)
NHj NOj NOg C H N
Average Total Aerosol Composition
Large Particle (> 2.5 pm) Composition
4.7
0.63 0.001
1.7
4.8
19.4
12.6
3.8
1.8
5.3
2.2
study are shown in Table 23. The ammonium results agree quite well, especially considering that the
samples were analyzed several months apart. Three different nitrate procedures were employed for selected
, aerosol samples, two by Battelle and one by the AIHL. There is considerable scatter in the results
from both St. Louis and West Covina, but the values determined by the three independent methods are
definitely of similar magnitude. Rockwell International employed a newly developed polaragraphic tech-
nique for the nitrate analysis, and, although they were unable to report quantitative results, their deter-
minations indicated nitrate values similar to or even lower than the Battelle values. These collaborative
nitrate analyses allow us to eliminate the analytical procedure as a factor in the low nitrate results.
The Gordon and Bryan'33* study mentioned earlier employed the usual fiber glass filters for aerosol
collections. As discussed in the St. Louis section, nitric acid may well interfere with aerosol nitrate
determination on glass fiber filters due to removal of HNOg from the air by basic-surfaced glass fiber
filters. Brief laboratory studies indicate that the neutral quartz filters employed in this study do not
scrub nitric acid from the air. If the average level of nitric acid we determined in West Covina were
removed by the filters, the nitrate values reported here would increase by about 8 Mg/m3- This would
put our nitrate results within the range of values reported by other researchers. This possibility of HNO3
interference with aerosol NO^ on fiber glass filters is an important finding of this study and will be
pursued in subsequent research.
* Our thanks to Dr. Bruce Appel at AIHL and Dr. Ed Parry at Rockwell for performing these analyses.
-------
51
TABLE 23. INTERLABORATORY COMPARISONS
Mg on Filter.
NO;
Filter No.
NHJ
BCL . AIHL
(NH3 Electrode) (Indophenol Blue)
BCL#1
/BrucineX
\Sulfate /
BCL if 2
/Chromatropic\
\ Acid /
AIHL
(2,4-Xylenol)
3
17
25
35
102
124
130
158
5.2
1.0
2.2
7.4
3.8
4.9
8.1
2.3
5.5
1.2
2.1
6.0
1.7
4.7
7.8
1.9
St. Louis
0.50
0.30
0.10(0.43)
0.60
West Covina
<12.4 8.5
1.5 3.3
0.5 1.7
6.3 7.8
0.06
0.20
0.34
0.26
7.9
2.4
0.68
1.7
The average ammonium values reported here are quite similar to NH4 concentrations reported'33)
elsewhere for Los Angeles. Average gaseous ammonia concentrations are considerably above the average
ammonium aerosol levels, so atmospheric conversion of NH3 to NH^ could well account for all the NH^
aerosol mass.
Comparison of the elemental ratios hydrogen/carbon (H/C) and carbon/nitrogen (C/N) is shown for
different aerosol size fractions in Table 24. The total aerosol H/C ratio is very similar for both cities at
about 2.4. This is the same H/C ratio as pentane; obviously, if saturated hydrocarbons are making up
any large fraction of the aerosol, they must be of higher molecular weight than pentane due to vapor
pressure considerations. Higher molecular weight saturates must have a lower H/C ratio than 2.4, and
TABLE 24. AEROSOL ELEMENTAL RATIOS FOR
ST. LOUIS AND WEST COVINA (MO-
LAR RATIOS)
Aerosol Size Range (/Ltm)
Elemental Ratio
Hydrogen/Carbon
Carbon/Nitrogen
St. Louis
West Covina
Total > 2.5 Total > 2.5
2.3
4.8
1.5
11.4
2.4
4.3
1.7
6.7
-------
52
oxygenated hydrocarbons of corresponding molecular weight will have even lower ratios. Thus, we are
left in the position of being unable to explain the H/C ratio on the basis of hydrocarbons alone; addi-
tional hydrogen must be contributed by another specie. The four hydrogens on the ammonium group
provide a ready explanation of the H/C ratio. If we subtract the average NH^-hydrogen contribution
from the total aerosol, the average H/C ratio becomes approximately 1.7. This is very close to the ratio
of 1.6 found for primary auto exhaust aerosols in a Battelle-Columbus study.(65) Because of the pos-
sibility of carbon and hydrogen-containing aerosols from other nonautomotive sources influencing this
ratio, however, agreement may only be fortuitous.
The carbon/nitrogen (C/N) ratio for the total St. Louis aerosol is 4.8. The West Covina ratio of
4.3 is somewhat lower than the St. Louis ratio. It is probable that this ratio is lower due to the higher
NO^-nitrogen in the West Covina atmosphere, although an alternative explanation would be that primary
carbonaceous exhaust aerosol comprises a larger fraction of the total carbonaceous aerosol in West
Covina than in St. Louis. In view of our downwind (well-aged air mass) sampling site in West Covina,
this latter explanation appears unlikely.
The higher C/N ratios in the large particle-size range is undoubtedly caused by the much lower
contribution of NH^-nitrogen in this size fraction.
An important step in understanding the fate of nitrogen oxides in the atmosphere involves a
determination of the fraction of NOX which may result in aerosol. Since "total aerosol nitrogen"
analyses were performed on the filter samples, we can undertake a filter nitrogen balance to ascer-
tain the extent to which known nitrogen compounds can account for the aerosol nitrogen. Daily West
Covina filter nitrogen balances are shown in Table 25. The fraction of total aerosol nitrogen that can
be accounted for by NH^ and NO^ averages 73.6 percent. Nitrite values are not shown in the table
because of their negligible impact oh the final nitrogen balance. The 26 percent total aerosol nitrogen
that cannot be accounted for as NH^ or NO3 amounts to roughly 0.002 ppm average unexplained aerosol
nitrogen in the West Covina atmosphere. If all of this unexplained aerosol nitrogen results from NOX
reactions, then this figure is not negligible in terms of the fate of nitrogen oxides in West Covina. A
comparison of the aerosol nitrogen balances for St. Louis, West Covina, Columbus, New York, and
Pomona is shown in Table 26. The Columbus, New York, and Pomona data were taken from a previous
Battelle-Columbus study.'32' The striking feature of these aerosol nitrogen balances is their similarity
from city to city. The question of the identity of the unaccounted-for nitrogen is unresolved at this
time. Possibly, the amine and pyridine-type compounds reportedly'36'found in atmospheric aerosol
samples by electron spectroscopy chemical analysis (ESCA) will account for the remaining aerosol nitro-
gen. In an effort to determine whether these basic nitrogen species might be present in our samples, we
collected six silver membrane filters in West Covina suitable for ESCA analysis. The filters have been
sent to the Lawrence Berkeley Laboratory for analysis; however, at this writing the results are not yet
available. For this reason, discussion of the ESCA results must be left for the second year of this study.
Distribution and Balance of Nitrogen Compounds
The distribution of gaseous nitrogen compounds in West Covina has been given as averages in
Table 10, as composited time-dependent profiles in Figure 3, and as individual daily profiles in
Appendix B.
-------
53
TABLE 25. WEST CO VINA FILTER NITROGEN BALANCE
Date
8/23
8/24
8/26
8/27
8/28
8/29
8/30
8/31
9/3
9/4
9/5
9/6
9/7
9/8
9/9
9/10
9/11
9/12
9/13
9/14
9/17
9/18
9/19
9/20
9/21
9/24
9/25
9/26
9/27
9/28
Filter
2
2
2
2
2
2
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
mg
2.96
2.72
1.01
1.94
1.94
3.73
6.38
6.22
3.73
8.17
6.30
3.81
8.17
6.30
3.34
3.66
3.03
4.90
8.56
11.67
12.53
13.62
11.28
9.80
6.92
1.87
2.97
1.79
0.93
1.94
NOg-N,
mg
2.80
1.29
.74
.41
.59
.50
.50
.23
.29
.14
.14
.34
.45
.11
.20
.27
.23
.36
.29
.11
.25
.27
.29
.27
.45
.52
1.33
1.42
.95
1.17
ZNHj-N + N03-N,
mg
5.76
4.01
1.75
2.35
2.53
4.23
6.88
6.45
4.02
8.31
6.44
4.15
8.62
6.41
3.54
3.93
3.26
5.26
8.85
11.78
12.78
13.89
11.57
10.07
7.37
2.39
4.30
3.21
1.88
3.11
Total N,
mg
4.6
5.8
2.5
2.8
3.8
6.6
8.0
8.0
5.7
9.3
7.0
7.1
14.0
8.0
5.5
6.3
5.3
9.1
10.8
13.3
13.9
17.2
14.3
11.3
10.3
3.7
. 7.5
6.5
4.6
4.5
ZNHj-N + N03-N
Total X 1°°
125.2
69.1
70.2
83.8
66.5
64.0
86.0
80.6
70.5
89.4
92.0
58.4
61.6
80.1
64.4
62.4
61.5
57.8
81.9
88.6
91.9
80.8
80.9
89.1
71.6
64.6
57.3
49.4
40.9
69.1
Average 73.6%
TABLE 26. AEROSOL NITROGEN BALANCE
City
Average Aerosol Nitrogen
Accounted for as NH^ and
percent
St. Louis, Mo.
West Covina, Calif.
Columbus, 0.
New York, N.Y.
Pomona, Calif.
82.5
73.6
73.6
75.8
70.7
-------
54
In reviewing the West Covina results presented thus far, we note that, as in St. Louis, the sum of
the measured nitrogen oxide reaction products makes up only a small fraction of the average NOX. To
understand the fate of nitrogen oxides we must first determine whether we are measuring all the products
and observing all the processes which account for NOX removal from the atmosphere. Such a determina-
tion can only be made by ascertaining the balance between nitrogen oxides and their reaction products.
The method used for deriving the nitrogen balance will be the same as that employed earlier for the
St. Louis results. Before turning to the nitrogen balance, however, it is important to first examine the
data for general trends and relationships which may .improve our understanding of the conditions under
which nitrogen oxides are removed from the atmosphere.
Statistical Interpretation. The initial scanning of the data was again accomplished using the Auto-
matic Interaction Detector (AID) statistical routine. The graphic-tree outputs for the initial AID analyses
of our West Covina data bank are given in Appendix C. A list of the dependent variables employed in
the AID analyses and the three primary predictor variables for each is shown in Table 27. Several im-
portant similarities are apparent from comparison of this table with the corresponding St. Louis AID
TABLE 27. WEST COVINA AID RESULTS (FIRST
THREE PREDICTOR VARIABLES)
Three Primary Predictor Variables
Dependent Variable
NOX
PAN
HN03
NH3
NH^ (Excluding Total Aerosol N)
N03
Total Aerosol Nitrogen
Mass Loading
1
03(~)
03(+)
PAN (+)
03 (+)
NOX (+)
Wind Speed (-)
NHj (+)
Wind Speed (-)
2
Humidity (— )
NO (-)
N02 (+)
N02 (+)
NH3 (-)
NO (+)
N02 (+)
HNO3 (+)
3
HN03 (+)
HN03 (+)
03(+)
PAN (-)
HN03 (-)
Temperature (— )
NOX (+)
results in Table 16. First, the three primary predictor variables for NOX are the same in both cities —
namely, ozone, relative humidity, and nitric acid. The relationships of these variables to NOX are also
in the same direction for both cities. While recognizing the inherent limitations of these statistical rela-
tionships, the similarity between the two cities still is so striking that it strongly implies similar pro-
cesses occurring in both cities. Continuing with Table 27, we note that 03 is one of the primary pre-
dictors of PAN and nitric acid concentrations, as it was in St. Louis. The relationship is also positive
in both cities; that is, high ozone levels predict high concentrations of PAN and HN03. The negative
correlation between O3 and NOX, along with the positive correlation between 03 and the NOX reaction
products, PAN and HN03, is completely consistent with current kinetic mechanisms of photochemical
smog processes. As might be expected, PAN and HNO3 are strong positive predictors of one another.
The remainder of Table 27 is interesting in itself and in comparison with the St. Louis results;
however, these results are not directly pertinent to the fate of nitrogen oxides and will not be discussed
further here. It must be emphasized that the aerosol results in the table are derived from daily aerosol
collections and, consequently, are based on a much more limited number of observations than the gas-
phase results.
-------
55
Using the AID results as a guide toward potentially important relationships, a series of regressions
was undertaken for further documentation of some of the relationships just discussed. Results from
these regression analyses are shown in Table 28. Only 6:00 a.m. to 6:00 p.m. data were used in the
first series of tabulated regressions in order to examine relationships during the irradiated portion of the
day. The relationships between NOX and its three most important predictor variables are in the same
directions as the corresponding St. Louis regressions given in Table 18, although the West Covina corre-
lation coefficients are not nearly as strong as those for St. Louis, In fact, the correlation coefficients
are so low as to be almost meaningless. This no doubt reflects the fact that the major determinants of
NOX concentration are variables such as traffic density and meteorology.
TABLE 28. WEST COVINA REGRESSIONS
6 a.m. - 6 p.m. Data
y
NOX
NOX
NOX
03
03
°3
03
HN03
HN03
HN03
X
R.H.
03
HN03
NO
N02
PAN
HN03
Global
N02
PAN
R
0.13
-0.26
0.10
-0.41
0.30
0.78
0.63
0.40
0.47
0.71
The correlations between O3 and PAN, 03 and HNO3, and PAN and HNO3 are relatively strong,
reinforcing the contention that PAN and HN03 formation are related to ozone and therefore that ozone
must play a role in determining the fate of nitrogen oxides. It is also interesting to note in the table
that the O3-N02 coefficient is positive in West Covina, while it was negative in St. Louis. This sign re-
versal may reflect some interaction effect not yet fully understood.
A second series of regressions is shown in the upper portion of Table 29 where the maximum
hourly average concentrations of O3, PAN, and HNO3 have been run against the 6:00-9:00 a.m.
averages of either NO or N02. These correlation coefficients are very low, undoubtedly because the
interaction of hydrocarbons has not been taken into account. Hydrocarbon data may well be in-
corporated into the data analysis during the second year of this program.
Two regressions shown in the lower portion of Tablu 29 are for maximum daily 1-hour PAN and
HN03. The 03-PAN coefficient is relatively high, as in St. Louis. In contrast to St. Louis, the 03-HNO3
coefficient is also rather high. The slope (m) of the regression lines is also given in this table. The slope
found for the O3-PAN curve is 2.7, in good agreement with the value of 3 found by Lonneman161' for
several U.S. cities. The 03-HN03 slope is also very close to this value of 3. The fact that the 03-PAN
and 03-HN03 slopes are so similar but the average PAN concentration is considerably higher than HN03
-------
56
TABLE 29. WEST COVINA REGRESSIONS
Maximum Hourly 03, PAN, HN03
6-9 A.M. NO, NOX
m
o3
03
NOX
NOX
03
03
NO
NOX
PAN
HN03
PAN
HN03
0.13
0.27
-0.03
0;15
0.76
0.58
0.06
0.12
-0.24
1.65
2.73
2.92
further confirms the view that PAN persists longer in an air mass than HN03 and that the scavenging
processes for HN03 are probably more rapid than those for PAN.
Nitrogen Balance. From the statistical treatment of the West Covina data just discussed, it is clear
that a determination of the amount of NOX removed from the atmosphere is necessary in order to re-
duce the masking effect which the major determinants of NOX concentration are having on the subtle pro-
cesses which remove NOX from the atmosphere. The method which will be used to calculate "NOX
loss" from the West Covina atmosphere is the same procedure used in St. Louis. However, in West
Covina, accurate gas chromatographic measurements of the tracers CO and C2H2 are not available at
this time. These measurements, taken simultaneously with our monitoring data by the General Motors
mobile laboratory situated next to our laboratory in West Covina, will be made available shortly and will
be incorporated in our data base for use in our continuing data analysis effort. Meanwhile, with the
consent of the CAPA-9 Project Committee, we have proceeded with the West Covina data analysis
using tracer data (CO) from the nearest available monitoring station, the LAAPCD site in Azusa,
California. The Azusa data on CO, NOX, and hydrocarbons is included in Appendix B of this report.
Initially, two options were available to us for analyzing the CO and NOX data for "NOX loss";
(1) we could use the CO data from Azusa in conjunction with our own NOX data from West Covina or
(2) we could use Azusa data for both CO and NOX. The first option was ruled out by correlational
analysis of the NOX data from the two sites. The correlation coefficient for the data from the two sites
was low enough (R = 0.4) that it indicated time lags and, possibly, different local source intensities for
the two sites. The use of CO data from one site as a tracer for NOX concentrations at the other station
appeared dubious. This is not too surprising, since the sites are roughly 5 miles apart. We decided,
therefore, to use the CO and NOX data from Azusa .to calculate "NOX loss". This "NOX loss" could
then be compared with the NOX reaction products' concentrations which we had determined in West
Covina in order to derive a nitrogen balance. The rationale for making this comparison is as follows:
CO and NOX at the two different sites do not correlate well because they are primary pollutants and are
strongly influenced by local source fluctuations (traffic patterns, etc.). However, factors such as PAN
concentration, HNO3 concentration, and extent of NOX loss are properties characteristic of large-scale
air masses and as such should be intercomparable over spatial distances as small as 5 miles. One potential
problem with these comparisons may be time lag; there is no assurance that air masses will reach the two
monitoring stations at the same time. We must remain aware of the possible time lag problem in our
subsequent discussions of the data.
One possibly beneficial side effect accruing from our use of the Azusa NOX data results from the
fact that the LAAPCD employs colorimetric rather than chemiluminescent methods for NOX analysis.
This eliminates the tricky problem of correcting our chemiluminescent NOX data for variable PAN and
HNO3 interferences. It is interesting to note that the average NOX concentration at Azusa during the
-------
57
course of this study was 0.122 ppm, while the average West Covina NOX, as determined by chemilumi-
nescence, was 0.136 ppm. While the agreement is already quite good, if we assume quantitative inter-
ference with the N02 chemiluminescent determination by PAN and nitric acid, then the corrected West
Covina NOX average is 0.124 ppm, in even better agreement with the Azusa data.
One negative aspect of our use of the Azusa data involves the nondispersive infrared (NDIR) pro-
cedure used by LAAPCD for carbon monoxide. This technique has been shown'66' to be less sensitive,
less accurate, and more prone to interference than gas chromatographic techniques. Indeed, the sensi-
tivity of the technique is such that LAAPCD reports CO concentrations in ppm as whole numbers only.
This is certainly sufficient for most monitoring purposes; however, our calculation of "NOX loss" is so
sensitive to the CO concentration that we must again turn to regression analysis for the most precise
determination of the crucial CO/NOX ratio.
The regression plot of CO versus NOX is shown as Figure 7. The use of integer CO values is ob-
vious from the stratification of the plot. The numbers within the coordinate system refer to the number
of points falling on the same spot. Nine or more coinciding points are represented by the number nine.
The total number of points in the plot is close to 700.
The correlation coefficient of 0.73 shown in the statistical tabulation in the figure is quite reason-
able for two primary pollutants which are both associated to some degree with auto emissions. Interest-
ingly, this is the same coefficient found for the St. Louis CO-NOX regression. The most important
information to be derived from this plot is its slope, 18.4 ± 0.6 in consistent units, and its intercept,
1.43 ± 0.06 ppm CO. Fining this slope and the most recent CO/NOX emissions inventory ratio for
Los Angeles'67' into Equation (2) yields:
(C0/N0)
- -1 (0.122)
NOX loss = 0.035 ± 0.006 ppm .
Thus, 0.035 ± 0.006 ppm is the average NOX loss during our monitoring effort in West Covina. The
6-ppb deviation is merely the statistical deviation about the slope calculation. There are several sources
of error which may be having a much greater impact on the accuracy of the "NOX loss" value. First,
the calculation depends on the accuracy of emission inventories which are over 2 years old. Second, it
depends on the elimination of a CO intercept value which is no longer just a geologic background CO
term, but which now presumably incorporates insensitivity and inaccuracy of the analytical technique
along with the background CO. An important argument in favor of eliminating the CO intercept re-
quires thinking of the extrapolation to zero NOX (the CO intercept) in physical terms. Since virtually
all CO sources included in the emissions inventory are also NOX sources (over 90 percent of Los Angeles
CO comes from auto exhaust, while less than 70 percent of the NOX is from automobiles), the CO re-
maining when all NOX sources are eliminated (extrapolation to zero NOX) must not be part of the
emissions inventory ratio and therefore must not be included in the (CO/NOx)m ratio regardless of its
origin.
In terms of a nitrogen balance, we can account for an average of 0.012 ppm of the missing NOX
by the sum of PAN and nitric acid; NO^ along with the remainder of the unidentified aerosol nitrogen
can account for an additional few parts per billion. At best, however, the measured NOX reaction
products can account for less than 20 ppb of the missing NOX on the average. Because of the many
factors which may be influencing the accuracy of the "NOX loss" calculation, we cannot at this time
make an unqualified judgment as to the "goodness" of the nitrogen balance. If the assumptions made
here are justified, then the balance must be close (certainly, three-fourths of the NOX is not lost im-
mediately after emission, as has been suggested), but a more exact balance must await the inclusion of
the more accurate CO and C2H2 data in the second year of this investigation.
-------
58
CO,
ppm
8.00 »
1 . . .
I
T
I
I
T
T
.*.*» »
I
I
I
T
5.93 *
I
T • -
I
1
"5.20" 4 " -" '" ' "
1
1 231.
I
I
"•.SO 4
I
T
I
3.80 4
3.19
22 fl 899999
7.1.0 4
T
1.73 4
I
I
I
I
l.CO »«, ' i, fc • ? »
2.10 U.60 7.2". 9.15
1
T
- _. I
I
T
T
• * * * I
T
T
I
T
I
•3 • « 2 <• 2 5 * • •»
* j
I
I
I
7 * 7 T IT"" u 6 3 ? fc » • • ? » • T »
I
4
I
I
I
4
T
I
------ -• • i
T
I
T
I
I
.1
T
I
•4
-4- ---4 4 4 4- 4 4 * 4 4 4 4 4 4.
12.<»C 15.00 17. 60 2C.?3 27.80 25.i»0 28.00
, pphtn
STATISTICS..
COPRFLATION (f)-
STO~E R»~0~"Lf? "i~~
STGMTFTCA1C.. A -
.'52861
SIG^IFICANCC
.77171
.0 >0'i'l
.o'oai
IMTforppi (ft) -
I.d327~>
"."i837"7"
STO EPRO-? OF A
~STn ERPO" OF B"
, 000
.06C
.035
FIGURE 7. SCATTERGRAM OF LA-DATA FATE OF NOX
-------
59
In the foregoing discussion, we have computed an average "NOX loss" for our 5 weeks of monitor-
ing in the Los Angeles basin and have used this average loss term to determine the average nitrogen
balance in the West Covina atmosphere. In order to elucidate the factors which are influencing atmo-
spheric NOX loss, it would be quite instructive to examine the time dependence of the removal of NOX
from the atmosphere. Presumably, a knowledge of this time dependence of the removal processes might
indicate the nature of the removal mechanisms. Once the temporal variation in "NOX loss" is available
it can be compared statistically with the fluctuation in our other measured chemical and meteorological
parameters. Plots of the calculated "NOX loss" for the entire West Covina monitoring program are shown
in Figure 8. These plots were made with the raw CO and NOX data and must be corrected to account
for the error introduced by the CO analytical problems mentioned earlier. Unfortunately, we have no
way of correcting each individual hourly average; we must apply a general correction derived from the CO
intercept of Figure 7. To compute the average error caused by the CO intercept, we rearrange Equa-
tion (1) to split the CO term into slope and intercept contributions:
NOX loss = [CO]slope + "^intercept
°x IOSS (CO/NOx)Emission lnventory
. . (^intercept 1.43
Intercept corrects - [CQmo = 1^3
Intercept correction -0.100 ppm .
This CO intercept correction factor may now be subtracted from each hourly "NOX loss" value to
account for the average inaccuracy in the CO concentration. Obviously, the CO inaccuracy is going to
fluctuate from hour to hour and day to day; applying this correction will account for the average in-
accuracy, but we must still expect to see positive and negative variation in the corrected "NOX loss"
profile. The correction has been made in Figure 8 merely by drawing a new baseline at 0.100 ppm NOX
loss. Variation in NOX loss should be judged starting from this corrected baseline. It is interesting to
note that for many of the days shown in Figure 8 the new 0.100 ppm baseline seems to correspond
quite well with what might be described as the natural baseline that the curves themselves indicate. On
days of milder photochemical smog (as judged by the lower 03 levels in Table 10), such as September 4,
9, and 10, the natural baseline appears to be very similar to the artificial 0.100 ppm baseline we have
drawn. Further evidence of the appropriateness of this correction factor will be discussed shortly.
Using the corrected hourly average "NOX loss" values just computed as the dependent variable,
we can run an AID analysis on our entire West Covina data bank to indicate which variables are corre-
lated well with NOX removal from the atmosphere. The graphic-tree outputs from these AID runs are
included in Appendix C. A tabulation of the AID results is given in Table 30. In the first AID run, PAN
was unquestionably the best predictor of "NOX loss", with nitric acid and ozone also important vari-
ables. Since PAN, nitric acid, and ozone were shown earlier to be highly correlated in West Covina, the
first split, occurring on PAN, removes some of the potency from the HNOs and 03 predictor capacities.
However, PAN is such a strong predictor that it does appear again in split number three. The signifi-
cance of the split on temperature, which is negative in sign, is unclear at this time.
In the second AID run, PAN was omitted as an independent variable in order to determine what
other factors were correlated with NOX loss. Nitric acid and ozone were the most important predictor
variables, with the inverted temperature correlation appearing again. Solar intensity, which is well corre-
lated with 03, also appeared as an important splitter. In the third AID analysis both PAN and HN03
were removed from the data set. Ozone appeared as the most important predictor of NOX loss in this
run, as expected from the two earlier analyses.
Using the AID results for guidance, several regression analyses have been carried out to further
document the statistical relationship between "NOX loss" and the important predictor variables. The
regression results are shown in Table 31. Slopes (m) and intercepts (b) are included in the table. All of
the correlation coefficients given in the table are reasonably high, with the exception of those for
NOX avg, temperature, relative humidity, and NO3 av_.
-------
035
0 . 30
0 . 25
0 . 20
8-29
n 0.15
ppm
0.10
0 00
-005
8-30
8-31
9-1
9-2
120 130 MO. 150. 160. 170
180.
TIME ,
GRAPH 2
190
hours
200. 210. 220. 230. 2*0
FIGURE 8A. WEST COVINA NOX LOSS
-------
0 . 3S
0 30
0 . ?•>
0 ?0
9-3
n 0.15
ppm
•3,!0
J\.f
0 0'.
-0 OS
9-4
9-5
9-6
9-7
710
750
?*>0 770
780
7 <> 0 300. 310.
T IMF, hours
r, n a P H ?
3?0 . 330
3HO. 3SO.
FIGURE 88. WEST CO VINA NOX LOSS
-------
0 . 3S
0.30
0 . 25
0.20
9-8
n 0.1?
Ppm
\
0 1 0
0 00
9-9
9-10
-DOS
9-11
9-12
O
K>
360 170 •>* 0 . 390. MOO. "MO.
<« 20 . M 30
T T n F , hours
GRAPH ?
MHO. H50. "»60. HTO. <«80
FIGURE 8C. WEST COVINA NOX LOSS
-------
0 . 35
0 . 30
0 75
0.20
ppni
0.10
•j no
-o
9-13
1 • III
J~\ ! - ! ^ \
9-1A
9-15
9-16
9-17
«••»•) SOO SIO 5?0. 530
? •• 0 . ? S 0 .
T t n F, hours
GRAPH ?
560. 5TO. 580. 590. 600
RGURE 8D. WEST COVINA Aft3D PJOa LOSS
-------
0 . 35
0.30
(• 1 0
70. A 1 0 6"4 0 . 650.
660 . 670.
T i nF i hours
R n o P M ?
6PO. 690. TOO. T10. 720
FIGURE 8E. WEST COVINA NOK LOSS
-------
0 35
0 . 30
0 . 25
0 . 20
ppm
o i
0 00
-0 05
9-23
rwi
9-24
9-25
9-26
9-27
J TO
T5 0
760 T TO
T 90 790
T i n r, hours
GRAPH 1
800. 810. 820. 830
8HO
FIGURE 8F. WEST CO VINA NOX LOSS
-------
0.35
0.30
OS I 0
861.0
871.0 881.0
T I n F , hours
GRAPH ?
891.0
901.0
911.0
FIGURE 8G. WEST COVINA NOX LOSS
-------
67
TABLE 30. RESULTS*8* OF AID ANALYSIS OP "NOK LOSS" DEPENDENT VARIABLE
Run Number Variable Omitted
Splitting Order and BSS/TSS of Split
None
PAN
PAN, HN03
Split Number
1 PAN(.18), HN03(.11). 03(.08)
2 Temp(.09>. WS(.08), WD(.07)
3 PAN(.IO), WD(.08), RH(.07)
Split Number
1 HN03(.13), 03(.08), RH(.03)
2 Temp(.Q9), WS(.OB), GK.04)
3 GK.03), WD(.02),Temp.(.02)
Split Number
1 Ozone(.07). Temp(.05), WSI.03)
2 Temp(.08), WS(.06), WD(.03)
3 RH(.04), 03(.03), WS(.02)
(a) WS it wind speed.
WO it wind direction.
RH it relative humidity.
Gl it solar intensity.
TABLE 31. WEST COVINA "NOX LOSS" REGRESSIONS
Daily Average NOX Loss
versus
Daily Average or 1-Hr Maximum Data
y
NOX loss
NOX Loss
NOX Loss
NOX Loss
NOX Loss
NOX Loss
NOX Loss
NOX Loss
NOX Loss
X
03 Avg
NOX Avg
SPAN + HN03 (coulometric)
Max O3
Max SPAN + HN03
Mass Loading
Temp
R.H. (%}
N03 (fig/m3)
R
0.63
0.34
0.74
0.68
0.73
0.76
•0.08
•0.15
0.07
m
1.57
0.20
3.07
0.45
1.03
0.00084
-0.0014
-0.00035
0.002
b
0.059
0.101
0.096
0.057
0.097
0.046
0.158
0.166
0.128
-------
68
The perhaps unexpectedly high correlation between average NOX loss and aerosol mass loading is
probably best understood as a coincidental result; both mass loading and NOX loss are expected to be
highest on photochemical smog days. It is interesting to note that the average "NOX loss" intercept, i.e.,
the apparent NOX loss when all the x-variables are extrapolated to zero, is 0.101 ppm "NOX loss" -
in extremely good agreement with the artificial "NOX loss" baseline which we calculated earlier. This
average intercept indicates an apparent NOX loss of 0.1 ppm, which is an average invariant to changes in
the independent variables shown in the table. Obviously, an NOX loss invariant to all external changes is
physically untenable and indicates some constant error in the calculation. We believe that this apparent
NOX loss is merely a further manifestation of the CO intercept error and that the agreement between
the two values is further confirmation of the validity of the calculated correction factor.
Speculation as to the significance of the slopes of these regression lines in intriguing; however, we
do not believe such a discussion is warranted at this time due to the many sources of error which may
be influencing these results. We would like to reserve such a discussion until after the data analysis and
interpretation phases of the second-year program have been completed.
Some very important progress in understanding the fate of nitrogen oxides has now been made.
We have separated out of the overall NOX concentration profile the gross factors influencing NOX varia-
tion, leaving us with time-dependent profiles of NOX that has been chemically or physically removed
from the atmosphere. The variations in "NOX loss" have subsequently been analyzed, with the result
that PAN, HNO3, and 03 are all highly correlated with the loss fluctuations. Along these lines, we have
determined that the average NOX loss during our 5-week stay in West Covina was only a small fraction
of the average total NOX concentration, on the order of 30-40 ppb. Although the error limits around
this loss value could be as large as 50 percent because of the many factors affecting the calculation,
the important point is that the average NOX loss is low. Another important finding is that the sum of
PAN, nitric acid, and aerosol nitrogen can account for a major fraction, if not all, of the NOX loss.
One final view of the problem can be gained by examining the time dependence of the composited
NOX loss profile. This 5-week averaged plot is shown in Figure 9. It should be observed first of all
that the data form their own baseline at approximately 0.120 ppm NOX loss. This is 20 ppb higher
than our artificial baseline derived from the CO intercept correction. This baseline may well be the more
accurate, however, since it includes any contribution from dry deposition processes and any inaccuracies
due to the emissions inventories; in our previous calculations, we had no way of estimating the effect of
this latter source of error. If the actual average correction factor is indeed 0.120 ppm, then the average
NOX loss is 0.035-0.020 or 0.015 ppm. This is almost exactly the average sum of PAN, HN03, and
aerosol nitrate. Further substantiation of the actual nitrogen balance must await the more accurate
tracer data to be available for the second-year investigation.
Of the two prominent humps displayed in the "NOX loss" curve, the second one can be accounted
for entirely by the sum of PAN and HNO3. This is illustrated by the profile in the lower portion of the
figure. We can say, therefore, that the mechanism of afternoon loss of NOX is photochemically related
and that the magnitude of the loss can be completely accounted for by measured NOX reaction products.
The explanation for the apparent early morning loss of NOX is unclear at this time. Obviously,
for the "NOX loss" to increase as shown in the figure, the (CO/NOx)m ratio must have increased. Since
it is during the period when this early morning hump appears that the CO and NOX emissions sources
are undergoing their most dramatic change of the day, the possibility exists that the hump is artificial.
The rationale is as follows: A major increase in the auto exhaust contribution to the air mass occurs
between 5:00-8:00 a.m., judging from the average NOX profile in Figure 3. Since auto exhaust has a
considerably higher CO/NOX ratio than the normal Los Angeles basin mixture (approximately 24 versus
14.3), our morning "NOX loss" peak may only be a result of a different emissions mix at that time of
day and not truly reflect removal of NOX from the air mass. Indeed, if the additional morning auto ex-
haust burden (at a CO/NOX of 24) raises the normal CO/NOX ratio from 14.3 to 16.5, then the 6:00 a.m.
morning peak would be completely eliminated. At this time there is no sure way to incorporate a varia-
ble emission inventory ratio into our calculations, although it seems very likely that the emissions ratio
must vary during the day due to traffic patterns. At this point we can only suggest the variation in the
CO/NOX emissions ratio as a probable cause of the apparent morning NOX loss.
-------
69c
0.17
0.16
c
o
§ 0-07
u
c.
<-> 0.06
0.05
0.04
0.03
0.02
0.01
PAN + HN03
8 10 12 14 16 18 20 22 24
Hour of The Day
FIGURE 9. WEST COVINA NOX LOSS AND PAN + HNO3 PROFILES
-------
71
REFERENCES
Preceding page blank
(1) Tuesday, C. S., "Chemical Reactions in the Lower and Upper Atmosphere", Cadle, R. D., Editor,
Interscience, New York (1961).
(2) Gay, B. W., and Bufalini, J. J., Environ. Set. and Tech., 5, 422 (1971).
(3) Stephens, E. R..Anal. Chem., 36, 928 (1964).
(4) Altshuller, A. P., and Cohen, I. R., Intern J. Air Water Poll., 8. 611 (1964).
(5) Altshuller, A. P., Kopczynski, S. L., Lonneman, W. A., Becker, T. L., and Slater, R., Environ. Sci. and
Tech., 1. 899 (1967).
(6) Altshuller, A. P., Kopczynski, S. L., Lonneman, W. A., Sutterfield, F. D., and Wilson, D. L., Environ.
Sci. and Tech., 4. 44 (1970).
(7) Stephens, E. R., "Chemical Reactions in the Lower and Upper Atmosphere", Cadle, R. D., editor,
Interscience, New York (1961).
(8) Schuck, E. A., Doyle, G. J., and Endow, N., Air Pollution Found. (Los Angeles), Rept. 31 (1960).
(9) Tuesday, C. S., Arch. Environ. Health, 7, 188 (1963).
(10) Kopczynski, S. L., Intern J. Air Water Poll., 8, 107 (1964).
(11) Leach, P. W., Leng, L. J., Bellar, T. A., Sigsby, J. E., Jr., and Altshuller, A. P., J. Air Poll. Control
Assoc., 14, 176 (1964).
(12) Spicer, C. W., unpublished results, Battelle-Columbus Laboratories (1972).
(13) Stephens, E. R., and Price, M. A., J. Air Poll. Control Assoc., 15, 320 (1965).'
(14) Spicer, C. W., and Miller, D. F., "Nitrogen Balance in Smog Chamber Studies", presented at the 67th
Annual Meeting of the Air Pollution Control Association, Denver, Colorado (1974).
(15) Stern, A. C., "Air Pollution", Vol. 1, Academic Press, New York, pp 33 (1968).
(16) Kopczynski, S. L., Lonneman, W. A., Sutterfield, F. D., and Darley, P. E., Environ. Sci. and Tech.,
6, 342 (1972).
(17) Darley, E. F., Kettner, K. A., and Stephens, E. R..Anal. Chem., 35, 589 (1963).
(18) Stephens, E. R., and Darley, E. f.,Proc. 6th Conf. Methods Air Poll. Studies, 1964, Calif. Dept. of
Public Health, Berkeley, California (1964).
(19) Air Quality Data for 1967 from the National Air Surveillance Networks. Revised 1971, Environmental
Protection Agency, Research Triangle Park, N.C., Publ No. APTD 0741 (1971).
(20) Hodgeson, J. A., Rehme, K. A., Martin, B. E., and Stevens, R. K., "Measurements for Atmospheric
Oxides of Nitrogen and Ammonia by Chemiluminescence", presented at 65th Annual Meeting of the
Air Pollution Control Association, June, 1972.
(21) Morgan, G. B., Golden, C., and Tabor, E. C.,./. Air Poll. Control Assoc., 17, 300 (1967).
(22) Healy, T. V., McKay, H.A.C., and Pilbeam, A., "Ammonia and Related Atmospheric Pollutants at
Harwell", U.K.A.E.A. Research Group Report A.E.R.E. - R6231 (1972).
(23) Healy, T. \I..Atmos. Environ., 8, 81 (1973).
(24) Lodge, J. P., Anderson, J. A. Breeding, R. J., Englert, T., Fogle, B., Haagenson, P., Klonis, H., Morris,
A. L., Pate, J. B., Pogue, R., Pritchett, J., Roper, N., Sheesley, D. C., and Wartburg, A., "Fate of
Atmospheric Pollutants Study - Results of the October 1972 Field Test in the St. Louis Area",
NCAR draft report (May 1973).
-------
72
(25) Scott. W. E., Stephens, E. R., Hanst, P. L., and Doerr, R. C., Proc. Am. Petrol. Inst., 37, 171 (1957).
(26) Hanst, P. L., private communication (March 1974).
(27) Murcray, D. G., Kyle, T. G., Murcray, F. H., and Williams, W. J., Nature, 218, 78 (1968).
(28) Rhine, P. E., Tubbs, L. D., and Williams, D., Applied Optics, 8 (7), 1500 (1969).
(29) Crutzen, P. J., J. Geophys. Res., 76 (30), 7311 (1971).
(30) Public Health Service Publ. No. 978, U.S. Dept. of Health, Education, and Welfare, Washington, D.C.
(1962).
(31) Lee, R. E., and Patterson, R. K.,Atm. Environ., 3, 249 (1969).
(32) Miller, D. F., Schwartz, W. E., Jones, P. E., Joseph, D. W., Spicer, C. W., Riggle, C. J., and Levy, A.,
"Haze Formation: Its Nature and Origin - 1973", Battelle-Columbus Laboratories report to EPA
and CRC, June 1973.
(33) Gordon, R. J., and Bryan, R. J., Environ. Sci. and Tech., 7 (7), 645 (1973).
(34) Junge, C. E., "Air Chemistry and Radioactivity", Academic Press, New York (1963).
(35) Dubois, L., Eagan, L., Teichman, T., and Monkman, J. L., "The Relation Between Concentrations of
Sulfate and Ammonium in Air Samples", presented at 164th National Meeting of the American Chemical
Society, New York (1972).
(36) Novakov, T., Mueller, P. K., Alcocer, A. E., and Otvos, J. W.,./. Colloid and Interface Sci., 39 (1), 225
(1972).
(37) Robinson, E., and Robbins, R. C., J. Air Poll. Control Assoc., 20 (5), 303 (1970).
(38) Robinson, E., and Robbins, R. C., "Sources, Abundance, and Fate of Gaseous Atmospheric Pollutants",
Final Report SRI Project PR-6755 (1968).
(39) Levy, H., Planet. Space Sci., 20, 919 (1972).
(40) Erickson, E., Tellus, 4,'215 (1952).
(41) Junge, C. E., Trans. Am. Geophy. Union, 39, 241 (1958).
(42) Georgii, H. W., J. Geophys. Res., 68 (13), 3963 (1963).
(43) Gordon, R. J., Mayrosohn, H., and Ingels, R. M., Environ. Sci. and Tech., 2 (12), 1117 (1968).
(44) Bonamassa, F., Gordon, R. J., and Mayrsohn, H., paper presented at 155th Meeting, American Chemical
Society, San Francisco, California (1968).
(45) Eschenroeder, A. Q., and Martinez, J. R., "Analysis of Los Angeles Atmospheric Reaction Data From
1968 and 1969", Final Report CRC-APRAC Project No. CAPA-7-68, General Research Corp. (1970).
(46) Hodgeson, J. A., Bell, J. P., Rehme, K. A., Krost, K. J., and Stevens, R. K., AIAA Paper No. 71-1067,
American Institute of Aeronautics and Astronautics, New York (1971).
(47) Martin, B. E., Hodgeson, J. A., and Stevens, R. K., presented at 164th National Meeting of the
American Chemical Society, New York (1972).
(48) Stevens, R. K., Clark, T. A., Baumgardner, R., and Hodgeson, J. A., presented at 164th National
Meeting of the American Chemical Society, New York (1972).
(49) Dimitriades, B.,J. Air Poll. Control Assoc., 17 (7), 460 (1967).
(50) Miller, D. F., and Spicer, C. W., "A Continuous Analyzer for Detecting Nitric Acid", Paper No. 74-17,
presented at the 67th Annual APCA Meeting, Denver, Colorado (1974).
-------
73
...•••• \
(51) Mast, G. M., and Saunders, H. E., Instr. Soc. Am. Trans., 1, 325 (1962).
(52) Miller, 0. F., Wilson, W. E., and Kling, R. G..J. Air Poll. Control Assoc., 21, 414 (1971).
(53) Lloyd, L., and Wyatt, P.A.H., J. Chem. Soc., 2248 (1955).
(54) Duisman, J. A., and Stern, S. A.,./. Chem. Eng. Data, 14, 457 (1969).
(55) West, P. W., and Lyles, G. L., Anal. Chim. Acts, 23. 227 (1960). /
;
(56) West, P. W., and Ramachandran, T. P., Anal. Chim. Acta, 35, 317 (1966).
/
(57) Saltzman, B. E..Anal. Chem., 26, 1949 (1954).
(58) Altshuller, A. P., and McPherson, S. ?.,J. Air Poll. Control Assoc., 13, 109 (1963).
(59) "Air Quality Data for 1967 from the National Air Surveillance Networks (Revised 1971)", U.S.
Environmental Protection Agency, August 1971.
(60) Sonquist, J. A., and Morgan, J. N., "The Detection of Interaction Effects", Monograph No. 35,
University of Michigan (1964).
(61) Lonneman, W. A., private communication (1974).
(62) Kopczynski, S. L., Lonneman, W. A., Winfield, T., and Seila, R., "Gaseous Pollutants in St. Louis
and Other Cities", draft report,. U.S. Environmental Protection Agency (1973).
(63) Junge, C. E., "Air Chemistry and Radioactivity", Academic Press, New York (1963).
(64) Hidy, G. M., "Theory of Formation and Properties of Photochemical Aerosols", presented at Battelle
School on the Fundamental Chemical Basis of Reactions in the Polluted Atmosphere, Seattle,
Washington (1973).
(65) Wilson, W. E., Miller, D. F., Trayser, D. A., and Levy, A., "A Study of Motor-Fuel Composition
Effects on Aerosol Formation - Part III", Battelle-Columbus report to API, Project EF-2 (1972).
(66) Stevens, R. K., Clark, T. A.,. Decker, C. E., and Ballard, L. F., "Field Performance Characteristics
of Advanced Monitors for Oxides of Nitrogen, Ozone, Sulfur Dioxide, Carbon Monoxide, Methane,
and Nonmethane Hydrocarbons", presented at the 1972 APCA Meeting, Miami, Florida (1972).
(67) "Profile of Air Pollution Control", Air Pollution Control District, County of Los Angeles (1971).
------- |