&EPA
United States
Environmental Protection
Agency
Environmental Sciences Research
Laboratory
Research Triangle Park NC 27711
EPA-600/3-79-078a
August 1979
Research and Development
Oxidant-Precursor
Relationships Under
Pollutant Transport
Conditions
Outdoor Smog
Chamber Study
Volume 1.
-------
RESEARCH REPORTING SERIES
Research reports of the Oftice of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7 Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ECOLOGICAL RESEARCH series. This series
describes research on the effects of pollution on humans, plant and animal spe-
cies, and materials. Problems are assessed for their long- and short-term influ-
ences. Investigations include formation, transport, and pathway studies to deter-
mine the fate of pollutants and their effects. This work provides the technical basis
for setting standards to minimize undesirable changes in living organisms in the
aquatic, terrestrial, and atmospheric environments.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/3-79-078a
August 1979
OXIDANT-PRECURSOR RELATIONSHIPS UNDER POLLUTANT TRANSPORT CONDITIONS
Outdoor Smog Chamber Study
Volume 1
by
J. E. Sickles, II
L.A. Ripperton
W. C. Eaton
R. S. Wright
Research Triangle Institute
Research Triangle Park, North Carolina 27709
Contract No. 68-02-2207
Project Officers:
J. J. Bufalini
B. W. Gay, Jr.
Atmospheric Chemistry and Physics Division
Environmental Sciences Research Laboratory
Research Triangle Park, North Carolina 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences Research
Laboratory, U.S. Environmental Protection Agency, and approved for publica-
tion. Approval does not signify that the contents necessarily reflect the
views and policies of the U.S. Environmental Protection Agency, nor does
mention of trade names or commercial products constitute endorsement or
recommendation for use.
-------
ABSTRACT
A total of 146 multiple-day experiments was conducted in the four
outdoor smog chambers that comprise the RTI smog chamber facility. These
experiments were conducted to investigate the influence of simulated
transport on ozone generation by various combinations of a surrogate urban
hydrocarbon mixture and nitrogen oxides. The simulation of transport was
accomplished by progressively diluting the contents of the chambers with
purified air.
Maximum first-day ozone concentrations were reduced under dilution
conditions. First-day ozone maxima were found to be sensitive both to the
dilution rate and the time in the experiment that dilution was initiated.
Second and third-day ozone maxima were reduced by increasing dilution rates
but the reduction was less than proportional to the extent of dilution.
Second and third-day ozone maxima were reduced by dilution but the ozone-
generative potential, as measured by net ozone concentrations, was not.
The ozone-generative potential of an aged photochemical system contained
in the RTI smog chambers generally exceeded 0.08 ppm.
Additional experiments were conducted to examine the ozone-generative
potential of low reactivity hydrocarbons, to provide data for testing and
validation of a computer-based photochemical simulation model, and to
compare the behavior of two types of outdoor smog chambers.
Volume 1 contains all textual material. Volume 2 (Appendixes) contains
individual hydrocarbon analyses from each experiment as well as concentration
(0_, oxidant, nitrogen oxides, THC, CH,, NMHC, HCHO, CN, and Freon-12) and
J *+
environmental (solar radiation and temperature) data from each experiment.
This report was submitted in fulfillment of Contract No. 68-02-2207
by the Research Triangle Institute under the sponsorship of the U. S.
Environmental Protection Agency. This report covers the period June 6,
1975 to June 23, 1978.
iii
-------
-------
CONTENTS
Abstract iii
Figures vii
Tables x
Acknowledgments xii
1. Introduction 1
2. Conclusions ....... 3
3. Recommendations 5
4. Experimental 15
Overview 15
RTI Smog Chamber Facility 18
Air purification unit 22
Reactant injection system 22
Sampling system 25
Smog chamber operating characteristics 25
Mixing 27
Performance of air purification system .... 27
Chamber tightness 27
Dilution 31
Sample line losses 32
Characterization experiments 33
Reagents 41
Measurement methods 44
Ozone 46
Oxidant 46
Nitrogen oxides (NO, N02, and N0x) 47
Nitrogen dioxide 48
Total hydrocarbons, methane, nonmethane
hydrocarbons, and carbon monoxide 48
Individual hydrocarbons 49
Formaldehyde 51
Condensation nuclei . 52
Freon-12 52
Solar radiation 53
Environmental variables 53
Data reduction and handling 53
Procedure 55
-------
CONTENTS
5. Results and Discussion 57
Overview '. . 57
Basic research program 57
Analysis of overall data set 58
Analysis of individual experiments 65
Experiments conducted to provide data for model
testing 78
Ozone-generative potential of low reactivity hydro-
carbons 80
Comparison between two outdoor smog chamber facilities:
matched experiments 85
Cross calibration comparisons 87
Background conditions 90
The matched experiment of 20-21 September 1976 . . 90
The matched experiment of 5 November 1976 100
Discussion 107
References 109
vi
-------
FIGURES
Number Page
Shape of growth function of mixed layer for St. Louis
based on 1976 RAPS data
2 Fractional distribution of NO as N02 along a south-to-
north cross-section of St. Louis from 0500 to 0700 CST and
from 1100 to 1500 CST on June 8, 1976 ......... . . 9
3 Fractional distribution of NO as N0» along a west-to-east
cross-section of St. Louis from 0500 to 0700 CST and from
1100 to 1500 CST on June 8, 1976 ............. 10
4 ' Ozone distribution along a south- to-north cross-section of
St. Louis from 0500 to 0700 CST, from 1100 to 1500 CST and
for the maximum hourly concentration on June 8, 1976 ... H
5 Ozone distribution along a west-to-east cross-section of
St. Louis from 0500 to 0700 CST, from 1100 to 1500 CST and
for the maximum hourly concentration on June 8, 1976 ... 12
6 Ozone concentrations aloft and on the ground during the
flight of Da Vinci II on June 8-9, 1976 .......... 13
7 General design of RTI outdoor smog chambers ........ 20
8. Overall system design of RTI Smog Chamber Facility .... 21
9. Air purification unit for RTI Smog Chamber Facility .... 23
10. Reactant injection system for RTI Smog Chamber Facility . . 24
11. Sampling system for RTI Smog Chamber Facility ....... 26
12 First-day maximum ozone concentrations versus initial
hydrocarbon concentrations for all experiments and for
static experiments .................... 59
13 First-day maximum ozone concentrations versus initial NO
concentrations for all experiments and for static
experiments ........................ 60
vii
-------
FIGURES
Number PaSe
14 First-day maximum ozone concentrations versus initial
hydrocarbon-to-NO ratios for all experiments and for
static experiments °
15 Third-day net ozone concentrations versus sunrise hydro-
carbon concentrations for all experiments and for static
experiments
16 Third-day net ozone concentrations versus sunrise N0x
concentrations for all experiments and for static
experiments
17 Maximum and net ozone concentrations for static experiments
and for dilution experiments initiated at sunrise 66
18 Maximum and net ozone concentrations for static experiments
and for dilution experiments initiated at crossover .... 67
19 Maximum and net ozone concentrations for static experiments
and for dilution experiments initiated at 1700 68
20 Maximum ozone concentrations plotted by case number and day
for dilution beginning at sunrise, 1700 and crossover ... 71
21 Net ozone concentrations plotted by case number and day for
dilution beginning at sunrise, 1700 and crossover 72
22 Maximum ozone concentrations under dilution conditions
relative to the maximum ozone concentration obtained
under static conditions ..... 77
23 Comparison of the designs and sizes of the RTI and UNC
outdoor smog chambers 88
24 Total solar radiation profiles for UNC and RTI on
September 20, 1976 95
25 Nitric oxide, nitrogen dioxide and ozone concentration
profiles for RTI Chamber 1 and RTI Chamber 2 on
September 20, 1976 96
26 Nitric oxide, nitrogen dioxide and ozone concentration
profiles for RTI Chamber 3 and RTI Chamber 4 on
September 20, 1976 97
27 Nitric oxide, nitrogen dioxide and ozone concentration
profiles for UNC Red Chamber and UNC Blue Chamber on
September 20, 1976 98
28' Ozone concentration profiles for the two UNC chambers
and the four RTI chambers on September 20, 1976 99
viii
-------
FIGURES
Number
29 Total solar radiation profiles for UNC and RTI on
November 5, 1976 103
30 Nitric oxide, nitrogen dioxide and ozone concentration
profiles for UNC Blue Chamber and RTI Chamber 2 on
November 5, 1976 104
31 Nitric oxide, nitrogen dioxide and ozone concentration
profiles for UNC Red Chamber and RTI Chamber 2 on
November 5, 1976 105
ix
-------
TABLES
Table Page
1 Summertime Mixing Heights and Apparent Dilution Factors in
Selected Urban Areas 7
2 Summary of Experimental Conditions for Urban Mix Runs 16
3 Summary of Experimental Conditions for Runs not Involving
Urban Mix as Reactant 17
4 Basic Research Program 19
5 Hydrocarbon Concentrations in the RTI Smog Chambers
Following Air Cleanup Operations 28
6 Dilution in RTI Smog Chambers 30
7 Maximum Ozone Concentration Achieved in RTI Smog Chambers
During Purified Air Irradiation Experiments 35
8 Ozone Half-Lives in RTI Smog Chambers 36
9 NO Oxidation in RTI Smog Chambers 37
10 Summary of Ozone Half-Lives for Various Smog Chambers 39
11 Reagents 42
12 Analyses of Hydrocarbon Mixtures Used in Urban Mix Experiments . 43
13 Measurement Methods , ,.,,... 45
14 Summary of Results from Experiments Conducted to Provide
Data For Model Testing , 79
15 Comparison of RTI Outdoor Smog Chamber Results with Results
From SAPRC Indoor Chamber. ,..,.,,.., , , , 81
16 Summary of Results for Low Reactivity Experiments QA
17 Maximum Ozone Concentrations Generated in Various Smog
Chambers Using Low Reactivity Hydrocarbons Q,
* * * * • OD
X
-------
TABLES
Table Page
18 Comparison of Physical and Chemical Characteristics
of Research Triangle Institute and University of
North Carolina Outdoor Smog Chambers 89
19 Measurement of Preinjection Chamber Contents and
Background Ambient Air 91
20 Summary of Results for Matched Experiments on 20
and 21 September 1976 92
21 Summary of Results for Matched Experiment on 5 November
1976 101
-------
ACKNOWLEDGMENTS
This project was conducted by the Research Triangle Institute under
Contract Number 68-02-2207 for the U.S. Environmental Protection Agency.
The support of this agency is gratefully acknowledged as is the advice and
guidance of the EPA personnel who contributed to the project: Dr. Basil
Dimitriades, who initiated the project, and Dr. J. J. Bufalini, and Mr. B.
W. Gay, Jr. who served as Project Officers.
Several persons in the Environmental Measurements Department of the
Research Triangle Institute contributed substantially to this project. Mr.
C. E. Decker was Laboratory Supervisor for the project. Mr. D. L. Ewald and
Mr. D. P. Dayton conducted day-to-day chamber operations, data reduction,
and data verification. Messrs. R. B. Denyszyn, L. T. Hackworth, and P. M.
Grohse developed the gas chromatographic procedures used and, along with Mr.
D. L. Hardison, conducted all the GC analyses. Mrs. A. L. Turner conducted
the wet chemical determinations. Mrs. S. K. Burt transferred the data into
a computer-capatible format. Dr. L. M. Worsham served as a consultant to
RTI during analysis of selected data.
The matched smog chamber experiments described in this report were
conducted and analyzed jointly by the authors and Mr. R. M. Kamens and Dr.
H. E. Jeffries of the University of North Carolina at Chapel Hill.
We gratefully acknowledge these individuals for their efforts in bring-
ing this project to a successful conclusion.
xii
-------
SECTION 1
INTRODUCTION
Anthropogenic activity accounts for the high concentrations of ozone
precursors that accumulate in urban areas during the early morning hours.
Frequently by mid-morning the high concentrations of hydrocarbons and
nitrogen oxides that were trapped within the lowest 200 to 400 meters over
an urban center become diluted by atmospheric convective mixing processes
that are promoted by solar heating. Thus, these emissions experience
significant dilution on the day of their release. They also experience
approximately 12 hours of irradiation. As an urban plume is transported
downwind of a city, it may experience several diurnal cycles of irradia-
tion and continued dilution with nonurban air.
Until recently most smog chamber simulations of photochemical smog
employed initial reactant concentrations that are typical of those in urban
areas, usually Los Angeles, during the early morning hours. Most experi-
ments were conducted with indoor chambers that used artifical irradiation.
In addition, the duration of these irradiations was usually short, e.g.
2 to 6 hours.
The current study was designed to investigate various aspects of photo-
chemical ozone production under conditions that simulate the ambient atmos-
phere more closely than had been accomplished in previous smog chamber
investigations. Multiple-day experiments were conducted in outdoor smog
chambers that are designed to permit the simulation of atmospheric dilution
conditions. The purpose was to investigate the influence of dilution on
ozone generation by various combinations of a surrogate urban hydrocarbon
mixture and nitrogen oxides. The relationship between ozone and ozone
precursors (hydrocarbons and nitrogen oxides) were examined under various
dilution conditions and repeated diurnal cycles of temperature and solar
radiation.
-------
-------
SECTION 2
CONCLUSIONS
Smog chamber experiments were employed to investigate the influence
of simulated transport on ozone generation by various combinations of a
surrogate urban hydrocarbon mixture and nitrogen oxides. The basic
research plan called for three-day smog chamber experiments to be conducted
in the four outdoor smog chambers that comprise the RTI Smog Chamber
Facility. Chamber contents were diluted to simulate the dispersion of an
air parcel during transport. Experiments were conducted at four target
initial reactant concentrations and at four dilution rates with initiation
of dilution occurring at three different times on the first day of an
experiment.
Findings from this study are summarized as follows:
1. For those experiments where 'dilution was initiated before the
first-day maximum ozone concentrations occurred, the maxima
were reduced in comparison to the maxima obtained under static
conditions. In no case examined in the current study did dilu-
tion cause first-day ozone levels to exceed those of correspond-
ing static experiments.
2. First-day ozone maxima were found to be sensitive both to the
dilution rate and to the time at which dilution was initiated.
Maximum ozone concentrations were reduced approximately in
proportion to the extent of dilution when it was initiated at
sunrise. When dilution was begun at NO-N02 crossover, however,
maximum ozone concentrations were reduced, but the reduction
was less than proportional to the extent of dilution.
3. The time required to achieve the first-day [0-j]max concentration
was reduced under dilution conditions relative to the times
observed under static conditions.
-------
4. Under both static and dilution conditions, first-day
ozone maxima were greater than second-day maxima, which
in turn, were greater than third-day maxima.
5. Minimum ozone concentrations on both the second and third
days of an experiment were reduced with increasing dilu-
tion rates.
6. Second and third-day ozone maxima were reduced by dilution,
but the ozone-generative potential as measured by net ozone
concentrations was not.
7. Second and third-day ozone maxima were reduced with increas-
ing dilution rates; however, the decrease was less than
proportional to the extent of dilution.
8. Second-day net ozone concentrations at 39 or 77 percent
dilution were greater than the corresponding levels under
static conditions.
9. Second and third-day net ozone concentrations generally
ranged from 0.08 to 0.30 ppra. Thus, in smog chambers, aged
photochemical systems have relatively high ozone-generative
potential.
Ancillary experiments were conducted to address additional points.
Major findings from these studies are listed below:
1. Maximum first-day ozone concentrations generated by low
reactivity hydrocarbons, such as propane, in the presence
of NOX are highly temperature sensitive and decrease sharply
with decreasing temperature.
2. If the classification of organics as low reactivity species
is based on whether or not these species can produce 0.08
ppm of ozone during smog chamber investigations, then the
results of the current study indicate that ethane, acetylene
and propane must be considered to be reactive hydrocarbons.
3. Comparison of results from two same-day matched experiments
conducted in the RTI and UNC outdoor smog chambers revealed
good agreement both for propane/NOx and for propene/NOx
experiments. This should increase confidence in the relia-
bility of data obtained in outdoor smog chambers.
-------
SECTION 3
RECOMMENDATIONS
Based on findings in the current study, first-day-ozone generative
potential is sensitive not only to the dilution rate experienced by a
reacting photochemical system but also to the time at which this dilution
is initiated. It is recommended that efforts in the next phase of this
research program be directed at improving the understanding of these
effects. The following questions should be addressed:
1. Is the effect of dilution one of simply reducing HC and
NOX at a constant HC/NOX ratio to lower absolute concen-
trations?
2. Is the sensitivity of the photochemical smog system to
dilution a manifestation of nonlinear effects resulting
from the photochemical mechanism?
3. At what point in the sequence of events leading to the
photochemical generation of ozone is the sensitivity to
dilution the greatest, and does it occur during the
time period leading to NO-N02 crossover?
4. How is the point of maximum sensitivity related to
dilution rate?
Experiments designed to elucidate these issues can be conducted in smog
chambers.
A topic currently of interest to those individuals who are responsible
for planning and implementing oxidant control strategies is the impact of
transported ozone on urban air quality. This issue can also be addressed
by smog chamber studies. Chamber operation, however, must be carefully
controlled to simulate both urban chemistry and meteorology.
To define the operating conditions required to simulate the meteorology,
the extent of dilution that an urban air parcel experiences must first be
defined.
-------
Nocturnal surface-based inversions exist in most urban areas during
evening and early morning hours. Urban emissions accumulate within this
layer until shortly after sunrise when the convective activity resulting
from solar heating causes this layer to erode and mixing to occur to the
daily maximum mixing height. Mean summertime early morning and afternoon
mixing heights are tabulated for selected cities in Table I1, If it is
assumed that the increase in mixing height during a day roughly reflects
the extent of dilution experienced by a typical early morning urban air
parcel, then the average extent of dilution for these urban areas is 78
percent.
The next point that must be addressed is the manner in which dilution
occurs—the shape function for growth of the mixed layer. This function
2
has been defined empirically from data collected in the RAPS Study and
is shown in Figure 1. During the first hour after sunrise the inver-
sion begins to erode, and 2 percent of the total volumetric exchange occurs.
During the next hour, 10 percent occurs. From 0830 to 0930 GST the inver-
sion is dissipated, so that by 1000 GST, 58 percent of the total volumetric
exchange has occurred. Most of the exchange, 85 percent, has occurred by
noon, and the maximum mixing height is established by 1500 GST. This suggests
that to a first approximation a uniform dilution rate of 30% hr should
be employed in smog chamber simulation of these events. The duration of
dilution should be 5 to 6 hours; it should begin at two hours past sunrise
and terminate at noon.
i
With the major meteorological parameters defined, the chemical para-
meters should be defined as well. There is evidence to suggest that the
maximum ozone concentration is determined by the total mass of ozone pre-
cursors injected into a smog chamber divided by the total mixing volume at
3
the time of the maximum-. Thus, the procedure commonly employed in smog
chamber studies of introducing the HC and NOX reactants as a batch injection
just prior to sunrise is a reasonable approach.
It is recommended that a surrogate urban mix be designed and employed.
Such a mix could be as simple as the two-component EKMA mix recently employed
in EPA's computer modeling efforts4"6 or as complex as to contain not only
alkane, alkene, and aromatic hydrocarbons, but aldehydes as well.
-------
Table 1. Summertime Mixing Heights and Apparent Dilution Factors
in Selected Urban Areas-'-
Urban Area
Dayton, OH
Nashville, TO
New York, NY
Peoria, IL
Pittsburgh, PA
Washington, DC
Mixing
Morning
349
417
662
272
333
378
Heights, m
Afternoon
1661
1845
1512
1498
1794
1884
Dilution Factor
0.21
0.23
0.44
0.18
0.19
0.20
t.
Dilution factor is the ratio of the morning to afternoon mixing heights;
extent of dilution = 1 - dilution factor.
2000 r-
CO
(4
-------
In previous smog chamber studies the initial fraction of NOX as N02
has been 10 to 25 percent. Examination of data collected in St. Louis
during the flight of the DaVinci II has shown this fraction to be much
greater . The fractional distribution of NO as NO- is presented along
A £*
South to North and West to East axes in Figures 2 and 3. The corresponding
ozone distribution is shown in Figures 4 and 5. On 8 June 1976, for the
0500 to 0700 CST hours, approximately 65 percent of N0x was as NO,,. For the
1100-1500 CST hours this value had increased to 80 percent. Although the
mean early morning N02/NOX percentage on this day was 65 percent, within
the urban area where the ozone concentrations were insufficient to oxidize
all of the NO to N02, approximately half of the total NOX existed as N02.
o
This finding has also been noted by others . Thus, it is recommended that
the fraction of NO as NO to be employed in future smog chamber experiments
A ^
be increased to 50 percent.
If experiments conducted to simulate the urban atmosphere require
dilution, then the chemical composition of the diluent air must be defined
as well. To a first approximation this air is normally very low in ozone
precursors ' . It may contain considerable amounts of ozone, however, as
is indicated by data collected in the DaVinci II experiment .
Hourly ozone concentration data collected on and beneath the DaVinci
II are presented in Figure 6. These data suggest that between 1000 and
1700 CST, on this day, the atmosphere was reasonably well nixed from
ground level to the altitude of DaVinci II. Near sunset, surface cooling
resulted in the establishment and growth of a surface-based radiation
inversion. This permitted the accumulation of ozone-destructive agents
near the surface. Ozone within this layer was depleted rapidly by surface
deposition and by reaction with emitted destructive agents. The ozone
above the nocturnal radiation inversion was insulated from the ground and
the destructive agents released at the ground. On the evening of 8 June and
the morning of 9 June the ozone aloft decayed at a very slow rate with a
half-life of 116 hours. Thus, during a single nighttime period, ozone was
transported aloft essentially undiminished over a distance of 300 km from
St. Louis, MO to Evansville, IN.
-------
100
80
LU
LU
Q-
60
CNJ
o
20
0
•I'M
8 JUNE 1976
I
I
1
I
I
I
I
I
100
80
60
20
50 HO 30
SOUTH
20 10 0 10 20
FROM RRMS 101
30 10
NORTH
50
0
Figure 2, Fractional Distribution of NOX as N02 Along A South-To-North Cross-Section of St. Louis
from 0500 to 0700 CST (o) and from 1100 to 1500 CST (A) on June 8, 1976.
-------
100
80
LU
Q_
60
CXI
O
20
0
1 I ' I r
8 JUNE 1976
50
1
I I I I I I I I I . I
i L I
30 20 10 0 10 20
WEST KM FROM RflMS 101
30 HO
ERST
100
80
60
20
50
0
Figure 3. Fractional Distribution of NOX as NC>2 Along A West-To-East Cross-Section of St. Louis
from 0500 To 0700 CST (0) and from 1100 to 1500 CST on June 8, 1976.
-------
0.25
0.20
ol 0.15
a.
UJ
Z ,»
M 0.10
o .
0.05
0.00
I • I • I ' I ' I • I ' I ' I ' I '
SOUTH-NORTH RflMS OZONE 8 JUNE 1976
50
I
I ,-- I XI .! i I , I , I ,
0.25
0:20
0,15
0.10
0.05
30 20 10 0 10 20 30'
SOUTH KM FROM RflMS 101 NORTH
50
0.00
Figure 4. Ozone Distribution Along A South-To-North Cross-Section of St. Louis From 0500 to 0700
CST (+), from 1100 to 1500 GST (o) and for the Maximum Hourly Concentration (A) on
June 8, 1976.
-------
0.25
0.20
a.
0.15
M 0.10
o
0.05
^.00
I ' I ' I ' I ' I ' I ' I ' I ' I '
WEST-EflST RflMS OZONE 8 JUNE 1976
" • - + •
I.I.I
I
50 10 30
WEST
20 10 0 10 20
KM FROM RflMS 101
- -I-
1
1
0.25
0.20
0.15
0.10
0.05
30 HO
EflST
50
0.00
Figure 5. Ozone Distribution Along A West-To-East Cross-Section of St. Louis from 0500 to 0700
CST (+), from 1100 to 1500 CST (o) and for the Maximum Hourly Concentration (A) on
June 8, 1976.
-------
0.25
a! 0.15
Q_
I ' I ' I
OZONE RLOFT & flT THE GROUND
0.00
0600
0.25
1200
1800 0000
TIME, CST
0600
- 0.20
- 0.15
- 0.10
- 0.05
0.00
1200
Figure 6. Ozone Concentrations Aloft and on the Ground During the Flight of DaVinci II on June
8-9, 1976. (A = DaVinci II, + = Mobile Van Underneath DaVinci II and o = Nearest
RAMS Station).
-------
On the morning of 9 June, the breakup of the radiation inversion resulted
in the downward mixing of transported ozone to the surface. The solar
radiation that causes the breakup of the inversion is also instrumental in
photochemical processes that generate ozone. Since precursors are emitted
at the ground, it is likely that the ozone-generative potential was greater
near the ground than aloft. The relative contributions of transport and
synthesis to increasing ozone concentrations observed at the ground on the
morning of 9 June cannot be separated in this case.
Transported ozone plays several roles in defining air quality within a
region. It may be mixed to the ground, enhancing local concentrations; it
may hasten additional ozone synthesis by early conversion of NO to N0£; or
it may modify the ozone-producing potential of a reacting photochemical
system in some as yet undefined way.
Modeling results have suggested that when ozone is added to an alkane-
N02 photochemical system it can act as a net radical source or radical sink
q
depending on the initial HC/NO^ ratio . These roles are reflected by
differences in net ozone production. The impact of "so called" transported
ozone on the ozone-generative potential of a photochemically reacting urban
mix is an issue that should be addressed. The following questions are
particularly relevant and amenable to experimental investigation:
1. What are the effects of adding ozone directly to a nondiluted
reacting photochemical HC/NOX systems?
2. Is the time of introduction of this ozone important (e.g. at
sunrise, at time of NO-N02 crossover, at time of [N0x]max)?
3. Are the effects of added ozone sensitive to initial HC and
NOX concentrations and their ratios?
4. Under dilution conditions, how does adding ozone to the diluent
air influence the maximum ozone concentrations achieved in exper-
iments designed to simulate urban conditions?
conditions?
It is recommended that the next phase of this research program be
directed toward extending the current understanding of the effects of
dilution on ozone production and toward defining the effects of transported
ozone on ozone production.
14
-------
SECTION 4
EXPERIMENTAL
OVERVIEW
This subsection describes the experimental design and provides a gen-
eral overview of the experiments that were conducted in this study. De-
tailed descriptions of the smog chamber facility, reagents, measurement
methods, data reduction and handling, and procedures are provided in subse-
quent subsections.
A total of 170 experiments was conducted in the four outdoor smog
chambers that comprise the RTI smog chamber facility. Two types of experi-
ments were conducted: those that involved the irradiation of a surrogate
urban hydrocarbon mix in the presence of NO and those that dealt with other
chemical systems. The urban mix experiments are identified in Table 2. The
current investigation is a continuation of an earlier contract, EPA 68-02-
1296 . To provide a complete summary of the research program, the 73
urban mix experiments conducted under the current study, EPA 68-02-2207, as
well as the 24 urban mix experiments conducted in the earlier study, are
*
identified in Table 2. The data collected during these experiments are
tabulated in Appendixes A and B (printed under a separate cover as Volume 2
of this report).
The 73 experiments that did not involve the urban mix are identified in
Table 3. These smog chamber runs were comprised of chamber characterization
experiments, experiments with low reactivity hydrocarbons, runs designed for
comparison with predictions of EPA's photochemical kinetics simulation
model, and experiments designed to permit comparison of the behavior of the
RTI and UNC outdoor smog chambers. The data collected during these experi-
ments , except for the chamber characterization runs which are summarized in
the text, are also listed in Appendixes A and B (Volume 2).
The basic research plan involved the urban mix and called for 3-day
smog chamber runs with four different target initial reactant concentra-
tions, four different dilution rates, and three different times at which
15
-------
TABLE 2. SUMMARY OF EXPERIMENTAL CONDITIONS FOR URBAN MIX RUNS
Starting
Date
7/17/75
7/22/75
7/28/75
8/4/75
8/8/75
8/12/75
8/17/75
8/21/75
8/27/75
9/25/75
9/29/75
10/8/75
10/14/75
10/28/75
11/4/75
H/18/75
11/18/75
11/18/75
11/18/75
1/9/76
4/14/76
4/28/76
5/6/76
8/4/76
8/24/76
9/13/76
9/23/76
11/11/76
Time of Initiation
of Dilution
Sunrise
1700
N0-N02 Crossover
N0-N02 Crossover
NO-NOZ Crossover
Static
1700
Sunrise
N0-N02 Crossover
Sunrise
Sunrise
Sunrise
NO-NO., Crossover
NO-N02 Crossover
1700
Static
N0-N02 Crossover
N0-N02 Crossover
N0-N02 Crossover
Static
Sunrise
Static
Static
Static
Static
Static
Static
Static
Percent
Di hit ion
in 24 Hrs
95
95
95
77
77
0
77
39
39
95
95
39
95
95
39
0
77
39
95
0
95
0
0
0
0
0
0
0
Measured Initial Conditions
HC (ppmC)/NO.. (pp.,,)
1
6.6/0.44f
11.5/1.0
7.0/1.0
7.4/1.0
S.9/.78
7.7/1.0
7.S/.96
8.7/.91
9.2/1.0
6.4/1.1
10.0/1.0
10.9/.98
9.7/1.1
S.9/.94
8.9/1.0
5.S/.25
5.7/.S4
1.0/.10
1.1/.40
3.5/1.1
24/1.0
4.0/1.1
7.2/1.1
12/1.1
Chamber
2
3.6/.13
4.5/.2S
4 . 2/ . 25
4 . ',/ . 24
3.S/.37
4.3/.24
4.S/.23
4.3/.22
4.9/.26
3.9/.2S
3.7/.24
3.6/.22
S.7/.26
S.7/.23
3.8/.2S
S.2/.24
S.4/.72
0.88/.14
1.5/.47
3.7/1.1
6.2/.74
3. 1/. 72
4.6/.77
6.5/.81
No.
3
3.4/.3S
4. I/. 73
4.0/.77
4.4/.71
3. I/. 51
3.8/.71
4.S/.69
4.S/.66
4. I/. 73
3.7/.80
5. I/. 73
5.3/.70
5.4/.80
S.3/.69
S.4/.74
S.3/.24
S.6/.34
4.7/.63
0.92/.10
1.6/.16
3.3/1.1
S.6/.25
3.9/.25
3.7/.26
6.4/.27
4
1.2/.06
1.8/.11
1 . 2/ . 1 1
I.O/. 11
0.93/. 15
1.6/. 11
0.99/. 10
0.98/. 10
1.1/.15
1.3/.13
1.21 . 10
0.58/.09
0.73/.12 Pli
0.99/.10 Ph
I.I/. 11
4.0/.26
Wii
3.S/.63 Pli
0.57/.11 Pli
1.8/.05 Pli
4.0/1.1 Pn
2.5/.11
~/.U
0.71/. 11
0.99/.12 Wii
CoiunitMils
and HC on Second Day
on Second Day
Winter Run
NO and HC on Second Day
in Chambers 1 and 3
NO on Second Day
NOX on Second Day
Preceded by NO Oxidation Ruuc
Winter Run
Conducted under EPA Contract 68-02-2207, unless noted otherwise; see Volume 2 (Appendixes A and B) for data.
Defined as target percent of initial chamber contents replaced by clean air in the 24-hour dilution period.
Entries denoted by blanks (--) indicate that analyses are not available; unless noted otherwise, HC was determined
.as NMHC by Beckman 6800.
Conducted under EPA 68-02-1296.
JJMHC not available, reported HC is the summation of the individual HC species.
"0 concentration is estimated.
-------
TABLE 3. SUMMARY OF EXPERIMENTAL CONDITIONS FOR RUNS NOT INVOLVING URBAN MIX AS REACTANT
Starting
Date
7/10/75
7/11/75
8/1/75
1/9/76
1/9/76
1/9/76
4/6/76
4/23/76
4/25/76
6/24/76
6/24/76
6/25/76
6/25/76
8/3/76
8/28/76
9/8/76
9/8/76
9/8/76
9/8/76
9/17/76
9/20/76
10/1/76
10/6/76
10/6/76
10/11/76
10/27/76
10/27/76
11/5/76
11/5/76
Time of Initiation
of Dilution
Static
Static
Static
Static
Static
Static
Static
Static
Static
Static
Sunrise
Static
Sunrise
Static
Static
Static
Static
Static
Static
Static
Static
Static
Static
Static
Static
Static
Static
Static
Static
Percent
Dilution
in 24 Hrs
•
0
0
0
0
0
0
0
0
0
0
95
0
95
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Measured Initial Conditions
HC (ppmC)/NOx (ppm)
Chamber No.
1
0.10/.001
0.06/.63
--/O.O
.35/.010
.056/.001
— f-.
.16/-
— / —
.02/1.0
--/.02
6.9/.17
— /.82
10.2/.21
1.3/.14
.68/1.6
5.5/.10
1.7/.55
2
0 . 04/ . 00 1
O.O/. 67
— /O.O
.96/.S3
O.O/. 006
— /.001
--/ —
.039/ —
--/-.
.03/1.0
-/.02
12/.18
— /.82
11. 1/. 22
.64/.S2
2.4/1.7
10/.22
3.3/.S6
3
0. 16/0.0
0.02/.62
--/.001
.01/.006
.065/.001
— /_-
.073/ —
../-.
.01/1.0
--/.02
5.6/.16
~/.»7
11.3/.21
1.2/.53
1.3/.55
3. I/. 056
2.6/.05
4
0.24/0.0
0.02/.68
— /.001
.34/.69
.007/.001
— /__
.063/ —
--/-.
.02/1.1
— /.02
-07/.17
— /.80
10.3/.20
.023/.10
1.2/.50
.01/.056
2.8/.047
Comments
d e f
Purified Air .Irradiation ' '
d c •
NO Oxidation ' '
03 Decay" '*
Purified Air Irradiation '
Propene Run
NO Oxidatione> e f
Purified Air Irradiation,'
Purified Air Irradiation
03 Decay ,
Purified Air Irradiation,
Purified Air Irradiation
Purified Air Irradiation,
Purified Air, Irradiation
NO Oxidation f
Purified Air Irradiation
Ethane
Propane
Acetylene
Control (NO ,added)
NO Oxidation
KTI/UNC Comparison, Propane
03 Decay
Runs for EPA Model, Propene
Runs for EPA Model, Control
Runs for EPA Model, Propene
Propane
Control (NO added)
RTI/UNC Comparison, Propene
Propane
Conducted under EPA Contract 68-02-2207, unless specified otherwise; see Volume 2 (Appendixes A and B) for data.
Defined as percent of initial chamber volume replaced by clean air in the 24-hour dilution period.
Entries denoted by blanks (—) indicate that analyses are not available; unless noted otherwise, HC was determined
as individual HC species or sum of these species.
Conducted under EPA 68-02-1296.
^Determined as NMHC by Beckman 6800.
Chamber characterization experiment; results of these experiments are summarized in text and are NOT listed in Appendixes.
-------
dilution was to have been initiated. Dilution of chamber contents with
purified air was employed to simulate atmospheric transport conditions.
Static operation (no dilution) was used to simulate stagnation conditions.
Following reactant injection, each experiment was begun in the static mode
and, if the experimental design called for simulated transport, dilution was
initiated on the first day at one of three times: sunrise, NO-NC^ cross-
over, or 1700 EST. One of four target dilution rates was employed such that
95, 77, 39, or 0 percent of a tracer present at the start of dilution would
have been removed by the end of the 24-hour dilution period. Dilution was
terminated 24 hours after it had been initiated. The chamber was then
operated in the static mode until the experiment was terminated at 1700 EST
on the third day. Target initial conditions called for initial HC/NO
A.
ratios between 7 and 20 with the initial [NMHC] of 1 to 10 ppmC and the
initial [NO ] (20 percent NO ) of 0.1 to 1.0 ppm. The target experimental
A £*
conditions and the dates on which the corresponding experiments were con-
ducted are identified in Table 4.
RTI SMOG CHAMBER FACILITY
The Research Triangle Institute Smog Chamber Facility consists of four
cylindrically shaped smog chambers. Each chamber has a surface area of 51
23 -1
m , a volume of 27 m , and a surface-to-volume ratio of 1.9 m . Figure 7
illustrates the general design. The chambers are located outdoors, and
irradiation is provided by natural sunlight. The walls are fabricated from
®
0-13-mm thick (5 mil) fluorinated ethylene propylene (FEP) Type A Teflon
film. The Teflon walls are supported by an interior aluminum framework.
The floors are 0.25 mm (10 mil) thick FEP Teflon film laid over a reflective
layer of aluminum foil, which serves to raise the light intensity within the
chambers and thus compensate for transmission losses through the walls.
Mixing in each chamber is provided by a 0.45-m diameter aluminum fan blade
on a shaft that is driven at 345 RPM by a 185-W (1/4 hp) motor using a
belt-pulley system.
In addition to the chambers proper, provisions were made for air puri-
fication, reactant injection, and sample collection with subsequent instru-
mental and wet chemical analyses. The overall system is illustrated in
Figure 8.
18
-------
TABLE 4. BASIC RESEARCH PROGRAM
VO
Time of
Initiation
of Dilution
Static
Sunrise
Crossover
1700
o
Experiments
Target Initital Reactant
Concentrations
Target HC/NO (ppmC/ppm)
Dilution
(%) Chamber 1 Chamber 2
0
0
0
0
0
0
39
39
95
95
95
39
39
77
77
77
95
95
39
77
95
considered
10/1.0
5.0/0.24
10/1.0
10/1.0
10/1.0
10/1.0
10/1.0
10/1.0
10/1.0
10/1.0
10/1.0
10/1.0
10/1.0
10/1.0
10/1.0
10/1.0
10/1.0
10/1.0
in the analysis
5.0/0.24
5.0/0.24
5.0/0.24
5.0/0.24
5.0/0.24
5.0/0.24
5.0/0.24
5.0/0.24
5.0/0.24
5.0/0.24
5.0/0.24
5.0/0.24
5.0/0.24
5.0/0.24
5.0/0.24
5.0/0.24
5.0/0.24
5.0/0.24
Chamber 3
5.0/0.71
5.0/0.71
5.0/0.71
5.0/0.71
5.0/0.71
5.0/0.71
5.0/0.71
5.0/0.71
5.0/0.71
5.0/0.71
5.0/0.71
5.0/0.24
5.0/0.71
5.0/0.71
5.0/0.71
5.0/0.71
5.0/0.71
5.0/0.71
Chamber 4
1.0/0.10
1.0/0.10
1.0/0.10
1.0/0.10
1.0/0.10
1.0/0.10
1.0/0.10
1.0/0.10
1.0/0.10
1.0/0.10
1.0/0.10
1.0/0.10
1.0/0.10
1.0/0.10
5.0/0.24
1.0/0.10
1.0/0.10
1.0/0.10
Starting
Date of
Experiment
8-12-753
11-18-75
8-4-76
8-24-76
9-13-76
9-23-76
8-21-75a
10-8-75
7-17-75
9-25-75
9-29-75a
8-27-75\
11-18-75
8-4-753
8-8-75
11-18-75
7-28-75\
11-18-75
ll-4-75a
8-17-753
7-22-7S3
of individual experiments.
Multiple dilution experiment.
-------
3.0
METERS
ALUMINUM
FRAME
Figure 7. General Design of RTI Outdoor Smog Chambers.
20
-------
AMBIENT AIR PURGE
N>
MANUAL
OSAMP
LING
PORT
INSTRUMENTATION
LABORATORY
MANUAL
SAMP-
LING
PORT
Figure 8. Overall System Design of RTI Smog Chamber Facility.
-------
Air Purification Unit
Details of the air purification unit are shown in Figure 9. This unit
has three modes of operation: purge, cleanup, and dilution.
During the purge mode, air is supplied by a blower from a 10-m tower.
By opening a manway in the floor and allowing the tower blower to force air
3 -1
through each chamber, purge flow rates of up to 2.3 m min are attained.
After purging with ambient air, the chambers are sealed, and air is
recirculated through the purification unit in the cleanup mode. The purifi-
cation unit contains the following equipment:
1. Desiccant column (6.5 kg of 4A molecular sieves);
2. Two HEPA particle filters;
3. Heated catalyst column (5 kg of 0.5 percent Pd on alumina catalyst;
operating temperature: 200-475° C);
4. Air cooler;
5. Purafil column (6.5 kg of Purafil for NO and 0- removal); and
A J
6. Humidifier.
Solenoid-driven valving allows the inclusion or exclusion of this equipment
as may be appropriate in achieving desired experimental conditions. In this
study, items 2, 3, 4, and 5 were included for the cleanup and dilution
operations. The purification or "cleanup" operation requires 8 to 12 hours
3 -1
at a flow rate of approximately 0.28 m rain . Pollutant removal efficiency
of the purification unit is discussed in a subsequent subsection.
To effect dilution, the chamber contents are recirculated through the
purification unit at flow rates corresponding to the desired dilution rate.
3 -1
Flow rates for this operation mode between 0.0085 and 0.058 m rain are
employed for 24 hours to simulate between 39 and 95 percent dilution. Of
course, the purification unit was not employed (zero flow rate) to simulate
static conditions.
Reactant Injection System
A schematic of the reactant injection system is seen in Figure 10.
There are three injection manifolds from cylinders of compressed gases. The
flow rates are controlled by calibrated fine metering valves, and the quan-
tity of each injection is controlled by timed, manual operation of the
appropriate solenoid valves. The hydrocarbon mix, toluene, and carbon
22
-------
u>
1 y 1 • Mi~rr*l~
D-4{ | ! STAGE
HUMIDIFIER
DILUTION FLOWRATE
CONTROL VALVE
COOLING WATER
Figure 9. Air Purification Unit for RTI Smog Chamber Facility.
-------
NJ
-P-
r
LAMP 1RANSFORMEH
U.V. OZONE GENERA! OH
"X
CHAMBER
VENT
1 10 VAC
VENT
Figure 10. Reactant Injection System for RTI Smog Chamber Facility.
-------
monoxide (CO) are injected sequentially from a copper manifold. Nitric
oxide and NO. are injected sequentially from a Teflon manifold. Ozone may
be added by injecting 0- from a copper manifold through an 0_ generator,
located at each chamber; this feature was employed in 0_ decay experiments.
After the reactants have been injected, each of the manifolds is flushed
with nitrogen.
Sampling System
The sampling system is illustrated in Figure 11. An automatic timer
activates the appropriate sampling solenoid valves and provides for a 10-
minute sample from each chamber once per hour. During the remaining 20
minutes of each hour, the instruments sample ambient air from the 10-m
tower, they are used to analyze bag samples, or they are calibrated.
The sampled air must pass from a chamber to the laboratory through
lengths of 4.8-mm ID TFE Teflon tubing that range from 26 to 48 m, depending
3 -1
on the chamber. The sample is drawn at a flow rate of 0.005 m min (5
1pm) by a Metal-Bellows pump located in the laboratory. A Metal-Bellows
MB-21 was used until 11 September 1975, and an MB-41 was used thereafter.
The sample is delivered to a glass manifold from which the instruments draw
their samples. The residence time in the sampling line is less than 10
seconds. The instruments include an 0_ analyzer, a NO-NO--NO analyzer, and
O £• A
an environmental chromatograph and have a total volumetric flow requirement
of 0.003 m3 min"1 (3 1pm).
In addition to the automated sampling system described above, a 1-m
long, 4.8-mm ID FEP Teflon tube is located under each chamber for periodic
manual grab sampling. Wet bubbler samples for oxidant, NO-, and formal-
dehyde (HCHO) determinations are collected at this location. Samples for
individual hydrocarbon analyses are also collected at this point in 10-liter
(R)
Tedlar bags. These samples are drawn through the 1-m long sampling tube
with a Metal-Bellows MB-41 pump and exhausted into the Tedlar bags. Typi-
cally, grab samples were collected manually twice a day from each chamber
(see Appendixes A and B for exact times).
Smog Chamber Operating Characteristics
Operating characteristics of the chambers comprising the RTI Smog
Chamber Facility are documented in the following paragraphs. This informa-
25
-------
AMBIENI
AIR
SAMPl E
I
TFE TEFLON f^j
TUBING
3/16" ID.
1/4" OD
\
NO/NO
ANALYZER
r
OZONE 1
ANALYZER I
x — x _
1
n
Environmental
Chromatograph
GLASS MANIFOLD
VENT
METAL BELLOWS PUMP
5 LITERS/MINUTE
"^ "—
1
1
i
i
1
K^
M
K/I ;
Xl
1
1
1
'
K
1 1 1 1 a
TIMER
o 115 VOLTS AC
+ 12 VOLT
DC POWER
SUPPLY
OUTDOORS f/J INSTRUMENTATION LABORATORY
Figure 11. Sampling System for RTI Smog Chamber Facility
-------
tion is reported to provide a basis for assessing the performance of the RTI
chambers and to permit comparison with other chambers.
Mixing—
As noted earlier, mixing in each chamber is provided by a fan designed
for that purpose. Unless specified otherwise, the fan operated continually
during each experiment.
Air velocity measurements have been conducted within each chamber. The
minimum air velocity was measured to be greater than 0.05 m sec within
0.02 m of the floor. Air velocities increased with distance from the walls
to a maximum that was greater than 4.0 m sec near the moving fan blade.
If a smog chamber can be considered to be an agitator-stirred tank,
then a published relationship can be used to estimate the time required for
complete mixing. This procedure indicates that the mixing of an injected
gas should be 90 percent complete within 24 seconds and 99 percent complete
within 43 seconds after injection.
Performance of Air Purification System—
The air purification system routinely reduces the NO content of the
^L
purified air to a measured zero (minimum detectable concentration [MDC]:
0.005 ppm). The catalytic hydrocarbon oxidation system typically reduces C«
to C 0 hydrocarbons measured by gas chromatography to less than 20 ppbC
(MDC: 0.1 ppb [v/v]). Typical "postcleanup" hydrocarbon levels minus the
ethane + ethene and the toluene components are shown in Table 5 for each
chamber. The correction was made because of possible sampling line and bag
contamination. This point is discussed in detail in a subsequent subsection
(see "Individual Hydrocarbons").
Chamber Tightness—
Exchange of chamber contents with the ambient atmosphere is expected.
Chamber leaks may be attributed to replacement of the volume required by
sampling and to chamber "breathing" caused by diurnal temperature variations
3 -1
and buffeting by winds. The sampling flow rate of 0.005 m min for 10
minutes per hour corresponds to a loss coefficient of 0.002 hr or dilution
of 5 percent in 24 hours.
27
-------
TABLE 5. HYDROCARBON CONCENTRATIONS IN THE RTI SMOG CHAMBERS
FOLLOWING AIR CLEANUP OPERATIONS
Date Chamber No. 1
8-17-75
8-27-75
9-4-75
9-29-75
11-18-75
4-23-76
6-24-76
8-3-76
8-24-76
9-23-76
10-6-76
10-11-76
10-14-76
10-27-76
11-1-76
11-5-76
11-11-76
0.012
0.009,
b
0.007
0.038
0.023
0.039
0.008
0.009
0.002
0.010
0.006
0.013
0.043
0.010
0.003
0.008
[HC]a,
Chamber No- 2
0.002
0.008
0.005
0.009
0.057
0.011
0.011
0.009
0.017
0.018
0.017
0.035
0.009
0.034
0.018
0.030
ppmC
Chamber No. 3
0.008
0.013
0.006
0.016
0.008
0.015
0.018
0.004
0.010
0.004
0.003
0.011
0.006
0.004
0.023
0.015
0.031
Chamber No. 4
0.014
0.050
*
0.074
0.002
0-016
0.009
0-008
0.020
0.001
0.023
0.010
0.008
0.019
0.011
0.030
[HC] is summation of concentrations of individual C- to C,,. hydrocarbon species
less ethane + ethene and toluene.
bBlanks indicate that no analysis is available.
28
-------
To quantify chamber leakage, first order leak rate coefficients were
estimated by least squares regressions of In [tracer] versus time data.
During the course of the current experimental program CO, NO , and Freon-12
X
were used as tracers to assess chamber leakage. Results from these experi-
ments are summarized in the portion of Table 6 that corresponds to a target
24-hour dilution of zero.
Most of the experiments in the basic research program shown in Table 4
had been completed between June 1975 and the end of April of 1976. The mean
leak rate coefficient across all four chambers during this period is 0.008
±0.004 hr"1.
Based on the results in Table 6, chamber 2 had the lowest leak rate,
chambers 4 and 1 had slightly higher leak rates, and chamber 3 had the
highest value. The leak rate coefficient in each chamber generally in-
creased between June 1975 and January 1977. They ranged from approximately
0.005 hr to greater than 0.015 hr which correspond to dilutions of 11 to
30 percent in 24 hours. The increased leakage over the intervening 19-month
period is due to weathering caused by continuous exposure of the chambers to
"the elements."
As noted earlier, sample replacement corresponds to a leak rate coeffi-
cient of 0.002 hr" . If 0.010 hr is taken to be typical for the current
program, then environmental factors account for most of the leakage. Many
of the leak coefficients listed in Table 6 for the 0 dilution condition were
determined from data collected over one or more 24-hour periods. Although
the data reflecting this behavior are not presented in this report, leak
rate coefficients exhibit diurnal cycles that correspond qualitatively with
wind speed. Daytime leak rate coefficients exceeded nighttime values by
from 50 to as much as 200 percent.
Initially, the leak rates for the RTI chambers were approximately
0.005 hr . The Teflon chamber walls were replaced in May of 1977, and leak
rate coefficients were redetermined with Freon-12 as the tracer. Values of
0.004, 0.005, 0.004 and 0.004 hr were again found. These leak rate coeffi-
cients as well as those listed in Table 6 agree well with the 0.01 hr
12
value reported for the newly constructed UNC outdoor chambers. Concentra-
tion data in the present study were not adjusted for dilution.
29
-------
TABLE 6. DILUTION IN RTI SMOG CHAMBERS
Target
Date Dilution
7/17/75
7/22/75
7/28/75
9/25/75
9/29/75
10/14/75
10/28/75
11/18/75
4/14/76
6/24/76
8/4/75
8/8/75
8/17/75
11/18/75
8/21/75
8/27/75
10/8/75
11/4/75
11/18/75
7/11/75
8/13,14/75
8/17/75
11/4/75
11/18,19,20/75
1/9,10,11/76
1/9,10,11/76
4/28,29,30/76
5/6,7,8/76
6/24/76
8/3/76
9/17,18/76
12/16,17,18,19/76
12/20,21,22/76
1/13/77
95b
95
95
95
95
95
95
95
95
95
77°
77
77
77
39d
39
39
39
39
oe
0
0
0
0
0
oe
0
0
0
oe
oef
°f
°f
or
Dilution Coefficients
Chamber 1
0.1437
0.1428
0.1467
0. 1334
0.1333
0.1285
0.1040
—
0.1251
--
0.0868
0.0829
0.0793
—
0.0539
0.0214
0.0439
0.0221
--
--
0.0022
--
0.0113
0.0037
0.0056
—
0.0080
0.0114
0.0170
0.0164
0.0143
0.0131
0.0205
0.0227
Chamber 2
0.1687
0.1655
0.1797
0.1462
0.1460
0.1500
0.1156
—
0.1205
0.1327
0.0969
0.0965
0.0859
0.0957
0.0547
0.0446
0.0360
0.0334
--
--
0.0025
--
—
—
0.0031
—
0.0063
0.0085
--
0.0076
0.0134
0.0072
0.0160
0.0089
Chamber 3
0.1635
0.1530
0.1695
0.1542
0.1527
0.1511
0.1190
—
0.1344
0.1518
0.103
0.0960
0.0848
--
0.0506
0.0235
0.0748
0.0300
0.0309
0.0082
0.0072
--
0.0153
—
0.0058
—
0.0134
0.0228
—
0.0213
0.0201
0.0191
0.0270
0.0183
Chamber 4
0.1747
0.1724
0.1839
0.1583
0.1543
0.1922
0.1190
0.1527
0.1354
0.1612
0.0933
0.0899
0.0775
--
0.0347
0.0413
0.0320
0.0296
--
0.0089
0.0072
0.0079
0.0137
—
0.0065
0.0109
0.0089
0.0104
—
0.0123
0.0129
0.0082
0.0141
0.0110
First order loss coefficient as determined from the slope of least squares
regression of In [tracer] versus time data; unless noted otherwise the tracer
is CO; units: hr
[j
The target dilution coefficient required to achieve 95 percent dilution in
24 hours is 0.125 hr" .
The target dilution coefficient required to achieve 77 percent dilution in
24 hours is 0.061 hr" .
3
The target dilution coefficient required to achieve 39 percent dilution in
24 hours is 0.021 hr" .
^NO is the tracer.
f X
Freon-12 is the tracer.
30
-------
Dilution—
Target recirculation flow rates through the purification unit that were
employed in this study are 0.058, 0.029, 0.010, and 0.0 m3 min"1. These
rates correspond to target 24-hour dilution rates of 95, 77, 39, and 0
percent and may be interpreted to mean that, after 24 hours of operation,
95, 77, 39, or 0 percent of a tracer present at the start of dilution would
have been removed. A manual control valve is used to set the recirculation
flow rate through the purification unit. The valve setting is established
manually prior to each run in which dilution is to be employed.
To quantify the dilution for each experiment, first order dilution co-
efficients were estimated, using CO as a tracer, by least squares regres-
sions of In [CO] versus time data. These estimates are summarized in Table 6,
and should be considered by those who desire to model individual experiments
by computer simulation.
The actual dilution rates are subject to losses due to sample replace-
ment, to chamber leakage, and to leakage by the air purification system dur-
ing dilution operations. Since the target recirculation flow rates were set
without regard for these factors, the actual dilution rates are generally
greater than the target values.
Freon-12 tracer experiments conducted at the highest dilution rate (95
percent) in December 1976 and January 1977 suggest that a significant por-
tion, perhaps half of the dilution air volume, was ambient air rather than
air that had been completely recirculated through the air purification sys-
tem. This is considered to be a worst case estimate, and the ambient air
fraction is expected to be much lower at the lower dilution rates.
A target 24-hour dilution of 95 percent corresponds to a dilution coef-
ficient of 0.125 hr" . The mean of the measured dilution coefficients
across all four chambers is 0.148 ±0.002 hr and corresponds to a 24-hour
dilution of 97 percent.
At a target 24-hour dilution of 77 percent the dilution coefficient is
0.061 hr"1, while the mean of the measured values is 0.090 ±0.008 hr .
This corresponds to an actual 24-hour dilution of 88 percent.
The target 24-hour dilution of 39 percent corresponds to a dilution
coefficient of 0.021 hr . The mean of the measured dilution coefficients
is 0.039 ±0.014 hr" and corresponds to a 24-hour dilution of 61 percent.
31
-------
As discussed in the previous subsection, chamber leakage and sample
replacement account for dilution of the chamber contents under zero dilution
conditions. A dilution coefficient under these conditions of 0.010 hr is
considered typical for the current program and corresponds to an actual
24-hour dilution of 21 percent.
The target dilution rates were arbitrarily chosen to permit the inves-
tigation of the photochemical ozone production by HC-NO systems under a
X,
variety of simulated transport conditions. The average measured 24-hour
dilution rates of 97, 88, 61, and 21 percent offer a reasonable spread of
experimental conditions for examining the effects of simulated transport.
The discrepancies between the target and measured dilution rates are there-
fore of little consequence in achieving this objective.
Sample Line Losses--
The most distant chamber is 48 m from the instruments in the laboratory.
When NO, NO , and 0 are present in a chamber during periods of irradiation,
a small reduction of NO and 0. and a slight increase in NO- may occur in the
13
dark sample line due to the dark-phase reaction of NO and 0-. In view
of the short residence time (10 seconds), this contribution should be small
in most cases, and the data were not corrected for these effects.
The sampled air volume must pass through a considerable length of
sampling line (26 to 48 m) and a pump before it is delivered to the instru-
ments for analysis. Sample modification is therefore expected. A 20 percent
loss of ozone between the chambers and laboratory was reported on contract
EPA 68-02-1296. A thorough investigation of sampling line loss rates
was conducted on the current contract and has shown the initially reported
loss rate to be in error. In this study, single-component mixtures of air
and 0 , NO, and NO were prepared in Teflon bags at 3 to 5 concentrations
ranging between 0.08 and 1.0 ppm. Concentrations were determined by sampling
each bag directly at the instrument in the laboratory. Next, concentrations
were determined by connecting the bag to the sampling line located within
each chamber and by sampling in the usual manner. Sampling line losses for
0 , NO, and N02 were found to be less than 1 percent and to be independent
of concentration. The 03, NO, and NO data reported in Appendix B for EPA
68-02-1296 and EPA 68-02-2207 were therefore not corrected for sampling line
losses.
32
-------
Grab samples for individual hydrocarbon analyses were collected at each
chamber, were drawn through a 1-m long, 4.8-mm ID FEP Teflon tube and a MnO_
scrubber with a Metal-Bellows MB-41 pump, and were exhausted into 10-liter
Tedlar bags. A recent evaluation of sample collection containers for hydro-
carbon sampling has demonstrated that C_ to C_ hydrocarbons are stable for
7
up to 18 days when stored in Tedlar bags. However, high background
levels of FID-responsive species in the Cfi to C range were also found in
Tedlar bags. In the current study, the high background may prevent the
accurate determination of compounds such as toluene that fall within the C,
o
to CIQ range.
A MnO- scrubber located in-line at the intake of the HC sampling pump
was employed to remove 0« from the. sampled air. A recent evaluation of this
•^ 7
technique has shown the MnO. to pass C~ to C_ hydrocarbons quantitatively.
The scrubber, however, was shown to be much less than 100 percent efficient
at removing ozone. It is therefore anticipated that the ozone passed by the
MnO« scrubber would have reacted with 0,,-reactive hydrocarbons, primarily
the olefins. This could have occurred in most of the HC samples except
those collected on the morning of the first day of a run and would be re-
flected as a low bias for the reported olefin concentrations. Recently, FEP
Teflon tubing has been found to release considerable quantities of ethene to
a passing gas stream. The use of FEP Teflon tubing in the collection or
transfer of samples may have prevented the accurate determination of ethane
plus ethene.
Characterization Experiments--
The role of surface-mediated reactions in smog chamber investigations
is unclear. Contamination or "dirty chamber" effects have been observed in
14 15
glass, aluminum, and Teflon chambers. ' The levels of background FID-
responsive contaminants that are released by Teflon film vary from batch
to batch. ' Teflon film is the currently accepted material of choice
for the fabrication of smog chamber walls. However, the experimental condi-
tions for which the influence of chamber-associated contaminants may be
safely neglected are yet to be defined.
Three types of chamber characterization experiments were conducted to
document the behavior of the RTI smog chambers with respect to contaminant-
33
-------
associated effects: purified air irradiations, 0 decays, and NO oxidations.
The results of these experiments are presented in Tables 7, 8, and 9-
The purpose of the purified air irradiations was to document the concen-
trations of ozone, [Oj , that accumulated in the RTI smog chambers when
3 max
purified air was irradiated. The ozone results from photochemical processes
involving trace levels of nitrogen oxides and organics. These contaminants
either remain in the air after purification or desorb from the chamber
walls. Results from these experiments are summarized in Table 7. Ozone
levels generated in the RTI chambers on the first day of purified air irra-
diations ranged from 0.03 to 0.17 ppm. A seasonal effect is apparent: the
largest [0 ] occurred on July 10, 1975, whereas the smallest [03]max
occurred on January 9, 1976. The results of multiple-day experiments indi-
cate that, on the first day, ozone levels near 0.08 ppm could be achieved.
Also, second- and third-day maximum ozone concentrations could exceed the
first-day levels. The [0_] values presented in Table 7 compare favorably
J IQdX 1 r\
with the value of 0-14 ppm reported for the outdoor UNC facility and with
the values of 0.04 ,0.05 ,0.10 , and 0.22 ppm reported for indoor
chambers.
It is unclear how the results from purified air irradiations may be
related to results from HC/NO experiments. While purified air irradiations
A
may indicate the level of chamber-associated contaminants relative to other
purified air irradiations they may not be accurate indicators of the effects
of these contaminants in HC/NO experiments. These effects may be over-
X
shadowed completely in photochemically reactive HC/NO systems. It is ex-
X
pected that as the absolute reactant concentrations and their ratio change,
the impact of chamber effects on the overall behavior of the system will
also change. Chamber influences are anticipated to be nonlinear with chang-
ing reactant concentration, and the experimental conditions at which chamber-
related influences begin to dominate the behavior of chemical systems remain
to be defined.
It should be noted that the [0 ] that accumulates during a single
3 max
irradiation period is the net result of both ozone-formation and destruction
reactions that occur within a smog chamber. It is possible for a given
chamber to have a low light-phase ozone-destructive component, low absolute
concentrations of background contaminants (ozone precursors) and yet achieve
34
-------
TABLE 7. MAXIMUM OZONE CONCENTRATION ACHIEVED IN RTI SMOG
CHAMBERS DURING PURIFIED AIR IRRADIATION EXPERIMENTS
co
Cn
Date
7/10/75
1/9/76
4/6/76
4/7/76
4/8/76
4/23/76
4/24/76
6/24/76
6/25/76
6/26/76
6/27/76
6/28/76
8/28/76
8/29/76
8/30/76
%ssa
76
100
89
93
77
97
92
85
86
41
38
91
52
66
72
T ,
max
°C
30.0
0.0
22.2
22.8
22.2
31.1
32.2
28.8
32,2
30.0
30.6
31.7
31.1
33.9
26.1
[03J , ppra
•* max
Chamber 1
.155
.032
--
—
--
.108
.149
.155
.142
.134
.095
.152°
.089
.062
.093°
Chamber 2
.149C
.070
.101
.086
.043
.079
.08le>f
.117
.06le
.050
.102°
.111
.049
.062C
Chamber 3
.154
--
.080
.123
.117
.078
.121
.063e>8
.099
.050e
.053
.124°
.099
.099
.152C
Chamber 4
.170C
--
--
--
--
.078
.124
.048e>h
.102
.049e
.064
.137°
.147
.188
.235
Comments
Run followed a propene/NO
X
experiment
Three-day experiment
Two-day experiment
12-hour dilution from dawn
until dusk
Four-day experiment with 24-
hour dilution from dawn of
6/26/76 until dawn of 6/27/76
Three-day experiment
Duration of solar radiation reported as percent of possible minutes of direct sunshine (see text).
Unless noted otherwise, each experiment was preceded by a cleanup and was conducted" in the static mode
(no dilution).
Q
Last measurement of experiment.
Blanks denote that no experiment was conducted.
Experiment was conducted under dilution conditions.
Dilution rate based on CO loss was 13.3 percent hr
^Dilution rate based on CO loss was 15.2 percent hr
Dilution rate based on CO loss was 16.1 percent hr
-------
TABLE 8. OZONE HALF-LIVES IN RTI SMOG CHAMBERS
co
a
Date
8/l/75d
4/25,26/76e
10/l,2/76f
10/2,3/76e
8/1/758
4/26/76h
10/1/761
10/2/761
10/3/76J
%ssa
Dark
Dark
Dark
Dark
65
78
0
0
8
Chamber
[03]ib tl
0.915
0.830
0.442
0.363
0.755
0.635
0.690
0.266
0.245
1
c
i
21.0
26.2
16.8
18.7
10.1
14.0
17.6
11.4
15.1
Chamber
[03]i tJ
0.905
0.803
0.944
0.468
0.776
0.668
0.688
0.759
0.351
2
&
27.4
38.9
31.6
30.1
10.7
18.0
17.4
16.3
16.5
Chamber
[03]i tJ
0.907
0.663
0.937
0.382
0.760
0.459
0.682
0.700
0.260
3
&
23.0
18.9
23.7
23.1
11.1
12.8
16.6
13.5
17.4
Chamber
[03]i tk
0.910
0.603
0.507
0.368
0.798
0.453
0.687
0.411
0.276
4
31.2
24.8
40.0
27.0
10.4
11.2
25.5
19.8
19.6
Duration of solar radiation reported as percent of possible minutes of direct sunshine (see text).
Initial ozone concentration in ppm.
p
Half-life in hours; calculated from the slope of least squares regression of In [(}„] versus
time data.
Experiment was conducted from 0000 until 0400 EST.
Experiment was conducted from 1900 EST until 0500 EST.
Experiment was conducted from 1800 EST until 0500 EST.
Experiment was conducted from 0600 EST until 1500 EST.
^Experiment was conducted from 0600 EST until 1100 EST.
Experiment was conducted from 0600 EST until 1700 EST.
JExperiment was conducted from 0600 EST until 1000 EST.
-------
TABLE 9. NO OXIDATION IN RTI SMOG CHAMBERS
OJ
Chamber 1
Date
1/9/76
1/10/76
8/3/76
9/17/76
9/18/76
7/11/75
1/9/76
1/10/76
1/11/76
8/3/76
9/17/76
%ssa
Dark
Dark
Dark
Dark
Dark
10
WO
88
18
22
65
[N0]ib
--
—
0.63
0.81
0.53
0.58
—
—
—
0.80
0.74
Chamber 2
R° [N0]i
--
--
1.6(1.2)
1.2d
1.8(0.9)
2.9
—
—
—
1.2
3.2
--
—
0.72
0.80
0.48
0.61
—
--
--
0.82
0.73
Chamber 3
R [N0]i
—
--
1.3(1.1)
l.ld
1.7(0.7)
0.9
—
__
--
0.5
4.0
--
--
0.56
0.79
0.45
0.58
—
--
--
0.80
0.71
Chamber 4
R [NOji
—
—
2.2(1.3)
1.6(0.8)
2.4(1.0)
2.6
—
--
—
2.1
4.6
0.48
0.32
0.67
0.82
0.51
0.62
0.61
0.36
0.27
0.86
0.73
R
3.0(1.6)
2.6(1.5)
1.6(1.0)
1.4(0.9)
1.5(0.7)
2.0
2.2
2.5
4.7
1.4
3.4
Duration of solar radiation reported as percent of possible minutes of direct sunshine (see text).
Initial NO concentration in ppm.
*>
"Ratio of experimentally determined second order rate constant for oxidation of NO to the established
(literature) value k t^^l't* ratios in parentheses have been corrected for chamber leakage; ratios
not in parentheses have not Been corrected for chamber leakage.
Variability of NO data did not permit the determination of a leak rate coefficient.
X
-------
a higher [00] than another more contaminated chamber with a higher light-
j max
phase ozone-destructive component. Thus, the ozone-destructive component
under irradiation should be considered when comparing [0,] values achieved
j ID3X
in clean air irradiation experiments conducted in various chambers.
Ozone can disappear inside a chamber by interacting heterogeneously
with the walls or by reacting homogeneously with contaminants present inside
the chamber. Ozone decay rates reported as half-lives under both dark
conditions and irradiation have been used as measures of smog chamber reac-
tivity. For fixed-volume smog chambers, the apparent CL half-life is influ-
enced by chamber leakage. Leak rates were not determined for the Q~ decay
experiments in the current study. The 03 half-lives for the RTI chambers
listed in Table 8 have not been corrected for leakage. These tabulated
results indicate consistent behavior from chamber to chamber. Apparent
dark-phase half-lives range from 16.8 to 40.0 hours with an overall mean of
26.4 ±6.7 hours. The corresponding values under exposure to natural irradia-
tion range from 10.1 to 25.5 hours. Based on these results, 10 hours may be
considered to be a reasonable estimate of the ozone half-life in the RTI
chambers on a sunny summer day. The reduced half-life for irradiated condi-
tions has been attributed mainly to secondary reactions following ozone
1 9
photolysis (such as 0( D) + H.O, 0~ + OH, and 0_ + Hop .
Ozone half-lives reported for various smog chambers are summarized in
Table 10. Comparison of the ozone half-lives for the RTI chambers with
those reported for other chambers indicates that the RTI chamber surfaces
are relatively unreactive with 0«. In addition, the dark-phase 0_ half-lives
listed in Table 8 are in accord with a recently reported value of 116 hours
that was estimated for an elevated air parcel that was traveling downwind of
the St. Louis area.
The oxidation of NO should proceed in the dark by the third-order
thermal reaction, NO + NO + 0 •* 2 N0_. In the dark, in the absence of re-
active organic species, the above thermal reaction should be the major path-
way for NO disappearance. Apparent second order rate constants for NO dis-
appearance were determined from the slope of least squares regressions of
[NO] versus time data. For fixed-volume smog chambers, the apparent
second order rate constant must be corrected for chamber leakage. Leak rate
coefficients based on the dark-phase loss of NO were estimated for each
X
38
-------
TABLE 10. SUMMARY OF OZONE HALF-LIFES FOR VARIOUS SMOG CHAMBERS
OJ
vo
Chamber
Identity
Bureau of Mines
Bureau of Mines
Battelle
Shell
Exxon
SRI
CSARB
SAPRC
--
Lockheed
General Motors
--
UNC
UNC
--
RTI
Construction
Material
Al, Teflon film
Al , glass
Al, Teflon film
Stainless steel
Al, Teflon film
Teflon-coated Al
Glass
Teflon-coated Al
Teflon bag
Glass
Stainless steel
Tedlar Bag
Al, Teflon film
Al, Teflon film
Teflon bags
Al, Teflon film
Volume,
1
1
2
1
4
4
7
3
5
1
8
4
1
2
1
2
.8
.8
.7
.0
.3
.7
. 1
.8
1
.9
.4
.4
.6
.0
.0
.7
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
103
103
10"
10*
103
103
10"
103
102
103
103
102
10s
10s
102
10"
a
Half-lives, hr
Dark
12
14
8
1.7-6 (24° C)
13.9
20
6.8
9.5
7.2 (35° C)
10.0
11.9
49-75C
21.2
45-150d
(25° C)
17-40e
Light
1 (54° C)
:t.6 (34° C)
6
1.5 (27° C)
1.3-2.5 (30° C)
2.7
7.8
5.8
4.2 (29° C)
--
3.0 (35° C)
2.0
--
16-23°
(Nov. & Dec. 1973)
9.4
(June 1974)
9-l6d
(June 1975)
10-26e
11 lumir
nation
A
A
A
A
A
A
A
A
A
A
A
A
N
N
N
N
Reference
18
18
19
20
21
22
23
17
24
25
26
27
12
28
9
Current
Study
Measured at approximately 1.0 ppm ozone.
Artificial: A; natural sunlight: N, taken under sunny conditions unless stated otherwise.
Range presented for simultaneous experiments in two chambers.
Half-lives varied over this range with water vapor concentration.
"Range of half-lives reported in Table 8.
-------
experiment from the slope of least squares regressions of In [NO ] versus
X
time data. The apparent second order rate constants for NO disappearance
were corrected for chamber leakage. Ratios of both the apparent and
-2 -1
leak-corrected rate constants to the established value of 1.77 x 10 ppm
hr"1 29 are presented in Table 9. The dark-phase loss rate coefficients
for NO range from 0.003 to 0-011 hr"1 with a mean of 0.0072 ±0.0025 hr .
X
The uncorrected ratios of rate constants range from 1.1 to 3.0 and the
corrected ratios range from 0.7 to 1.6. These ratios suggest good agreement
between the established value of the dark phase NO oxidation rate constant
and the value as determined in the RTI chambers.
Loss rates of NO under irradiation in excess of the thermal rate may be
attributed to participation of organic contaminants in the normal photo-
chemical NO-oxidation reactions. Dimitriades has suggested that the rate of
NO loss under irradiated conditions provides a measure of chamber contamina-
18
tion levels. Apparent second order rate constants for NO disappearance
were determined by the technique noted above using data collected under
irradiation. For these conditions NO loss was much greater than under dark
X -1
conditions—the NO loss rate coefficients range from 0.008 to 0.032 hr
x _j
with a mean of 0.019 ±0.007 hr . Although increased daytime wind speeds
may increase the daytime leak rate, the contribution of leakage to the
current NO loss rates is unclear. Gas-phase or surface-mediated reactions
^^
that lead to the formation of N_0,_ and HNO« with the subsequent loss of NO
^ J -j X
may also contribute to the loss of NO . The data therefore do not permit
£^
correction of the apparent light-phase rate constants for leakage.
The uncorrected ratios of rate constants determined for irradiated con-
ditions compare favorably with those determined under dark conditions and
range from 0.5 to 4.7. These results for the RTI chambers also compare
line
12
18
favorably with the value of 4.0 for the Bureau of Mines chamber and to
values of 4.8 and 3.0 for the outdoor UNC facility.
In addition to the above characterization experiments, matched experi-
ments conducted during the experimental program show consistent behavior
from chamber to chamber. For example, a matched urban mix experiment was
conducted on 4 August 1976. Target initial conditions were 5 ppmC hydrocar-
bon and 1 ppm NO (20 percent NO ). The NO-NO crossovers for the four
X *f £m
chambers occurred within 6 minutes of each other at approximately 0655 EST,
40
-------
and the maximum N02 concentrations ranged from 0.79 to 0.86 ppm. The 0
maxima on the first day of the experiment occurred during the 1500 hour and
were in good agreement: 1.05, 1-07, 1.04, and 1.06 ppm.
A second matched experiment was conducted on 20 September 1976 with
target initial conditions of 12.0 ppmC propane and 0.2 ppm NO (20% N00).
X ^
The NO-NO crossovers occurred within 26 minutes of each other, and the
maximum NO^ concentrations ranged from 0.150 to 0.158 ppm. The first day
^°3^max values were also in reasonable agreement: 0.20, 0.26, 0.18, and
0.21 ppm. These results are discussed in detail in a subsequent section of
this report. The good agreement demonstrated in these and other matched
experiments increases confidence in the reliability of data obtained in the
RTI smog chambers.
REAGENTS
Air used in the smog chamber experiments was supplied from the air
purification systems that are located under each chamber. The gaseous
reagents that were employed are identified in Table 11. These gases were
used as received from the suppliers without further purification.
The 00 that was used for 0_ injections was zero (Z-2) grade. The N9
jLm -J fm
that was used to purge the injection lines was Matheson "Oxygen Free" grade.
Nitrogen oxides were introduced from separate tanks that contained NO in N~
and NO in N2-
Several different hydrocarbons were used in runs that did not involve
the urban mix (see Table 3). Ethane, acetylene, and propane were injected
as pure gases. The propene-air mixture that was used in several experiments
was blended at RTI from 99 percent propene (Phillips Petroleum Company) and
hydrocarbon-free air (Airco).
The hydrocarbon mixtures that were used in the urban mix runs (see
Table 2) were supplied by Matheson Gas Products. Two different mix tanks
were employed: one was used until it was depleted in late November of 1975;
the second was used for all the urban mix runs that were conducted in 1976.
The individual hydrocarbon analyses for each tank were supplied by Matheson
and are listed in Table 12. It should be noted that the first mix did not
contain acetylene or toluene—normal constituents of polluted urban air. In
the 1975 urban mix experiments, acetylene was introduced by syringe injec-
41
-------
TABLE 11. REAGENTS
Reagents
b.p.'
Purity
Supplier
Air
Oxygen -183.0
Nitrogen -195.8
(Oxygen Free)
Nitric Oxide -151.7
Nitrogen Dioxide 21.1
Ethane - 88.2
Acetylene - 75-0
Propane - 42.1
Propene - 47.7
Propene - 47.7
Carbon Monoxide -191.5
Toluene 110.6
Hydrocarbon
Mix Ib
Hydrocarbon
Mix IIC
99.95%
99.998%b
220 ppm in N2
115 ppm in N-
>99%
99.6%
99.5%
99%
244 ppm (v/v) in air
20,300 ppm in N~
28 ppm (v/v) in N-
212 ppm (v/v) in N0
Airco
Scientific Gas Products
Matheson Gas Products
Matheson Gas Products
Matheson Gas Products
Phillips Petroleum Company
Matheson Gas Products
Linde
Phillips Petroleum Company
RTI Blend
Matheson Gas Products
Matheson Gas Products
Matheson Gas Products
563 ppm (v/v) in N2 Matheson Gas Products
3Boiling Point, °C.3°
Contains less than 0.5 ppraC total hydrocarbons.
Mix used in 1975 urban mix runs, see Table 12 for analysis.
Mix used in 1976 urban mix runs, see Table 12 for analysis.
42
-------
TABLE 12. ANALYSES OF HYDROCARBON MIXTURES USED IN URBAN MIX EXPERIMENTS'
Hydrocarbon Mix I
Hydrocarbon. Mix II
Compound
Acetylene
Propane
N-Butane
Isopentane
Cyclopentane
Ethene
Propene
1-Butene
Trans-2-Butene
2-Methyl-2-Butene
ppm (v/v)
—
26
66
33
1
51
11
10
8
6
ppmC
.
78
264
165
5
102
33
40
32
24
ppm (v/v)
99
50
154
77
4.4
105
27.9
17.6
18.8
9.4
ppmC
198
150
616
385
22
210
83.7
70.4
75.2
37.6
aThese analyses were provided by gas supplier, Matheson Gas Products.
Mix used in 1975 urban mix runs, see Table 2.
CMix used in 1976 urban mix runs, see Table 2.
43
-------
tion as a pure gas; whereas toluene was injected from a separate tank that
contained toluene in N_. Acetylene was included in the second hydrocarbon
mixture; therefore, only toluene injections were required in the 1976 urban
mix experiments.
The hydrocarbon mixture was blended so as to permit the injection, in
combination with separate injections of acetylene and toluene, of a reason-
able surrogate of urban air. The composition of the surrogate urban mix was
specified by EPA such that on a carbon basis, the target relative fractions
of alkanes, alkenes, aromatics, and acetylene were 0.49, 0.22, 0.20, and
0.09- Propane, n-butane, isopentane, and cyclopentane comprised the alkane
fraction with propane representing itself; n-butane the straight chain
alkanes; isopentane the branched chain alkanes; and cyclopentane the cyclic
alkanes. Ethene, propene, 1-butene, trans-2-butene, and 2-methyl-2-butene
comprised the alkene fraction with ethene and propene representing themselves;
1-butene the terminally double-bonded olefins, trans-2-butene the internally
double-bonded olefins, and 2-methyl-2-butene the branched chain olefins.
Toluene represented aromatic compounds, and acetylene represented itself.
The relative amounts of each model species within its class were chosen
31
based on data reported for Los Angeles air.
In addition to the hydrocarbon mix, acetylene, and toluene injections,
CO was injected from a separate tank that contained CO in N_. Although CO
is a normal constituent of polluted urban air, CO was injected as a tracer
in the urban mix experiments at a constant initial target concentration of
10 ppm.
MEASUREMENT METHODS
The measurement methods employed in the chamber studies are described
below. A summary of these methods is presented in Table 13.
Both instrumental and manual analytical techniques were employed.
Ozone, NO, NCL, total hydrocarbons (THC), methane (CH,), and CO concentra-
tion data from automated instruments were recorded from each chamber once
per hour during the corresponding 10-minute sampling period. The reported
0 , NO, and N02 concentrations were reduced from strip chart records at the
eighth minute of each 10-minute sampling period. The THC, CH,, and CO
concentrations were reduced from the strip chart records of the second
44
-------
TABLE 13. MEASUREMENT METHODS
Ul
Measured Quantity
°3
Oxidant
NO, NO NO
N02
THC, CH4, NMHC, CO
Individual HC
HCHO
Method
Chemi luminescent
NBKI (Wet Bubbler)
Chemi luminescent
Saltzman
(Wet Bubbler)
Gas Chrornatography/
Flame lonization
Gas Chroma tography
Flame lonization
Chromotropic Acid
Manufacturer
Bendix Model 8002
Thermo-Electron Model
Beckman Model 6800
Perkin Elmer Model 900
Range
0-1 ppm
0- 1 ppm
HB 0-1 ppm
0-5 ppm
0-10 ppmC
>0.1 ppb(v/v)
>50 ppb(v/v)
0-1 ppm
MI)Ca
0.001 ppm
0.001 ppm
0.001 ppm
0.005 ppm
0.01 ppmC
0.1 ppb(v/v)b
50 ppb(v/v)
0.015 ppm
Condensation Nuclei (CN)
Freon-12
Solar Radiation
% Possible Minutes
Sunshine
Ambient Temperature
(Wet Bubbler)
Photoelectric
Nucleus Counter
Long Path Infrared
Spectroscopy
Pyranometer
Photoelectric Cells
Gardner Associates,
Type CN
Wilks Scientific
Mi ran I
Eppley Model 2
Foster Sunshine Switch
0-107 CN/cm3 100 CN/cm3
0-700 ppm 0.08 ppm
(0-1 AU)
0-2 Langleys
.Minimum detectable concentration as reported by the manufacturer.
This value is for cryogenic sample concentration.
This value is for direct injection of a 1 ml volume of sample.
Data collected by the U.S. EPA, Division of Meteorology, Research Triangle Park, North Carolina.
CData collected by NWS Forecast Office at RDU32.
-------
5-minute cycle of each 10-minute sampling period. The sampling frequency of
the data reported in Volume 2 (Appendix B) is once per hour for each chamber.
Oxidant, NCL, and HCHO were determined by wet chemical techniques.
Samples were collected twice daily from each chamber at midmorning and
midafternoon (see Volume 2 [Appendix B]) for exact times). Individual
hydrocarbon concentrations were determined by gas chromatographic analyses
of grab samples collected in Tedlar bags twice daily at sunrise and late
afternoon. Hydrocarbon analyses from the chamber runs, as well as the exact
times of sample collection, are reported in Volume 2 (Appendix A).
Ozone
Ozone was monitored with a Bendix Model 8002 Ozone Analyzer. The
principle of operation employs the chemiluminescent gas-phase reaction
between ozone and ethene. The instrument operates in the continuous mode
with a range of 0 to 1 ppm and an MDC of 0.001 ppm. Calibration was per-
formed prior to each experiment using a stable ultraviolet light ozone
generator. The output of the 0^ generator was determined by gas-phase titra-
tion of 0., with known NO concentrations that were blended from air and cer-
33
tified standard mixtures of NO in nitrogen. A recent intercomparison of
the NO calibration cylinder with four other certified calibration cylinders
indicated that the [NO] in the original cylinder was 7.0 percent less than
the label value. The reported ozone data have been corrected, and a correc-
tion factor (multiplier) of 1.070 has been applied to the raw data to arrive
at the ozone concentrations reported in Volume 2 (Appendix B).
Oxidant
For chamber studies, photochemical oxidant concentrations were measured
33 34
by the NBKI technique. ' This involved passing a known volume of
chamber air through two all-glass midget impingers in series; each contained
a 1% neutral-buffered potassium iodide (NBKI) solution. These solutions
were subsequently analyzed with a Bausch and Lomb Spectronic 100 spectropho-
tometer. Calibration curves and blanks were prepared periodically according
to the referenced procedures. The sampling duration was normally 10 minutes,
and the flow rates ranged from 600 to 700 ml min~ . Flow rate was controlled
by a calibrated hypodermic needle protected from overspray by a dessicant
46
-------
cartridge and glass wool trap. The reported data have not been corrected
for interferences from nitrogen dioxide, peroxyacyl nitrates or other oxi-
dizing species. In the reported data (Volume 2 [Appendix B]), the time
assigned to a wet chemical measurement for a chamber is the automated instru-
ment sampling period for that chamber which is closest to the beginning of
the bubbler sampling period.
Nitrogen Oxides (NO. NCL. and NO )
a * 2-* x-
Nitrogen oxides were monitored with a Thermo-Electron (TECO) Model 14B
NO-NO Analyzer. The principle of operation employs the chemiluminescent
A
gas-phase reaction between NO and 0». Two modes of operation are required
to determine NO, NO-, and NO . Nitric oxide is measured first using the
^ X
reaction of NO and Ov The determination of NO- and NO , however, requires
*J £» X
catalytic reduction of NO- to NO prior to the reaction of NO with ozone.
After reduction of NO to NO, the signal from the total NO in the sample is
taken to be the NO concentration. Electronic subtraction of the original
A
NO signal from the NO signal yields the N09 concentration. The instrument
A £m
operates with a 90-second cycle time: NO concentration is updated at the
end of the first portion of the cycle; and NO- and NO concentrations are
^ A
updated at the end of the cycle.
The instrument was calibrated prior to each experiment. Calibration of
the NO and NO channels was performed by dilution of a known concentration
of NO from a certified cylinder of NO in nitrogen. Calibration of NO- was
performed by using the NO,, produced from the gas-phase titration of known NO
concentrations with 0, from the calibrated ozone generator. As was noted in
the earlier discussion of the ozone calibration procedure, recently the [NO]
in the calibration cylinder was found to be 7-0 percent less than the label
value. The reported NO, NO-, and NO data have been corrected, and a
^ X
correction factor (multiplier) of 1.070 has been applied to the raw data to
arrive at the NO, N0_, and NO concentrations reported in Volume 2 (Appendix B)
fc A
The NO analyzer was usually operated on the 0 to 1 ppm full scale (FS)
A
range, although both the 0 to 0.2 and the 0 to 0.5 ppm FS ranges were used
occasionally. In contrast to the MDC of 0.001 ppm as specified by the
manufacturer, the effective MDC in the current study was approximately
47
-------
0-005 ppm on the 0 to 1 ppra FS range. This is determined by instrument
noise and the width of the strip chart trace. In addition, the zero base-
line for the NO and the NO channels displayed a strong temperature depend-
A
ence. Diurnal variations of the room temperature could produce as much as
10 percent FS shift in the zero baseline over the course of an experiment.
Frequent zero checks reduced the impact of this behavior. The data in
Appendix B that are suspected of having residual error due to zero drift
have been identified. The NO channel did not exhibit this behavior because
NO is the difference between the NO and NO channels which were both shift-
^ X
ing by equal amounts.
It has been demonstrated that nitric acid, PAN, and ethyl nitrate
interfere with NO and NO determinations in instruments of the type employed
35 X
in this study. The interfering species were not determined in this
study. Therefore, the reported NO- and NO data have not been corrected for
£• **
interferences.
Nitrogen Dioxide
In addition to the chemiluminescent measurements, the Saltzman method
36
was used to determine N0? levels in the chambers. Chamber air was drawn
through a glass, fritted Mae West bubbler containing Griess-Saltzman reagent.
The sampling duration was normally 15 minutes, and the flow rates ranged
from 600 to 800 ml min . Flow rate was controlled by a calibrated hypoder-
mic needle protected from overspray by a dessicant cartridge and glass wool
trap. Samples were analyzed with a Bausch and Lomb Spectronic 100 spectro-
photometer. Calibration curves and blanks were prepared periodically accord-
ing to the referenced procedure. The reported data have not been corrected
for interferences from ozone or peroxyacyl nitrates. In the reported data
(Volume 2 [Appendix B]), the time assigned to a wet chemical measurement for
a chamber is the automated instrument sampling period for that chamber which
is closest to the beginning of the bubbler sampling period.
Total Hydrocarbons, Methane, Nonmethane
Hydrocarbons, and Carbon Monoxide
Total hydrocarbon, CH,, NMHC, and CO were determined by a Beckman Model
6800 Air Quality Chromatograph that employs a flame ionization detector
(FID). The instrument operates with a 5-minute cycle time and provides THC,
48
-------
CH^, and CO concentrations once per cycle. Thus, once per hour during each
10-minute sampling period two sets of measurements were performed on the air
sampled from each chamber. Only the second set of measurements were reduced
and reported.
The NMHC concentration is a calculated quantity that is computed by
subtraction of the CH^ from the THC concentration. In the instrument, THC
is determined by injection of a portion of the sampled air directly to the
FID. A second portion of the sampled air is injected into a Porapak Q
stripper column that separates CH, and CO from C02, H.O, and the C2-and-
heavier hydrocarbons. As the CH, and CO emerge from the stripper column in
a single peak, they are diverted to a molecular sieve column. Here CH, and
CO are separated into two peaks with CH, eluting first. They are then
passed through a nickel catalyst methanator before introduction to the FID.
Methane passes through the methanator unaffected and is detected as methane
by the FID. Carbon monoxide is converted to methane in the methanator, and
is detected as methane by the FID.
Calibration was performed daily. Certified mixtures of CH, and CO in
hydrocarbon-free air that were supplied by Scott Research Laboratories were
employed as calibration gases. The calibration procedure was similar to the
37
present Federal Reference Method. This approach should provide accurate
determinations of CH, and CO. It should be noted, however, that the sub-
sequent determination of NMHC may be subject to a substantial low bias.
Most NMHC species yield a lower FID response on a per-atom-carbon basis than
methane. Results from a recent study that employed a Beckman 6800 have
indicated that the effective carbon numbers for typical urban NMHC species
oo
range from 0.42 to 0.60 relative to methane. Thus, to correct NMHC data
for this reduced response efficiency, a correction factor (multiplier) of
1.7 to 2.4 may be necessary. To maintain consistency with the Federal Refer-
ence Method, however, the NMHC data in Appendix B have NOT been corrected.
Individual Hydrocarbons >
The concentrations of individual hydrocarbon species were determined by
gas chromatographic separation with flame ionization detection. The results
of these 'determinations are listed in Volume 2 (Appendix A).
49
-------
Samples were collected from each chamber at sunrise and sunset on each
day of an experiment. Each sample was drawn through a 1-ra long 4.8-mm ID
Teflon tube and a MnO. scrubber by a Metal-Bellows MB-41 pump and was
exhausted into a 10-liter Tedlar bag. A specific volume of sample from this
bag was cryogenically trapped in a 3.2-mm OD stainless steel loop that had
been immersed in liquid oxygen. The trap was connected to the modified
Perkin-Elmer Model 900 gas chromatograph that was employed in the current
study. Next, the trap was heated, and its contents passed into the column.
Low molecular weight hydrocarbons (C.-C^ were separated on a 1.8-m x
3.2-mm OD stainless steel column packed with n-octane on Porasil. The
column temperature was 23° C, and the helium carrier flow rate was 12 ml
min
High molecular weight hydrocarbons (including aromatics) were separated
on a 1.8-m x 3.2-mm OD stainless steel column packed with GP 5% SP-1200/5%
Bentone 34 on 100/120 Supelcoport. The column temperature was 75° C, and
the helium flow rate was 20 ml min
Normally, the samples were analyzed within 8 hours after their collec-
tion. In some cases, however, longer storage periods, up to two days, were
required.
A Hewlett-Packard Model HP-3352 data system acquired the output signal,
integrated peak areas, and converted the areas into concentration values
which were printed by a teletype. Strip chart records of output signals
were maintained to supplement data system records. The MDC from the data
system using the liquid oxygen trapping injection technique is less than
0.1 ppb (v/v). Occasionally HC measurements were performed by direct injec-
tion of the sample into the GC column from a 1-ml sampling loop without
cryogenic concentration of the sample. This technique was generally used
for experiments involving high concentrations of single hydrocarbons and is
noted in Volume 2 (Appendix A). The MDC for this technique is 50 ppb (v/v).
Identification and quantification of compounds was based on comparison
of retention times and peak areas with those of calibration mixtures. Cali-
*If a bag initially contained 0.1 ppm 0 and 0.1 ppmV ethene, approximately
11 percent of the initially present etnene would be consumed in 8 hours
by reaction with ozone. This percentage would increase for more reactive
olefins. For example, under the same initial conditions 43 percent of the
initially present propene would be consumed in 8 hours.
50
-------
bration was performed at three-week intervals and showed a precision of
±5 percent. The hydrocarbon calibration mixtures were supplied by Scott
Research Laboratories and were certified to ±1 percent accuracy.
Several recent findings in our laboratory may have a significant
impact on the reliability of the individual EC data reported in Appendix A.
The concentration of C_ to C hydrocarbons has been shown to be
stable for up to 18 days in Tedlar bags.
High background levels of FID-responsive species in the Cfi to C n
range have been found in Tedlar bags.
Samples of FEP Teflon tubing have been found to release substan-
tial quantities of ethene (or a material with a similar retention
time) to a passing gas stream.
The MnO_ scrubber that was employed to remove 0_ from the HC
samples has been shown to pass C_ to C_ hydrocarbons quantita-
tively. However, the scrubber has also been found to be less than
100 percent efficient at removing ozone.
In view of these findings, reported C« to C_ alkane concentrations should be
reliable. Ethane plus ethene concentrations may be erratic due to the use
of FEP Teflon tubing during sample collection. The unstabilized or residual
ozone could react with sampled olefins and provide a low bias.* The first-
day samples that were collected on the mornings of each run should not have
been subject to this bias because there was no ozone present at the time of
sample collection. Toluene determinations may have been obscured by the
high C, to C-Q background of the Tedlar bags.
Formaldehyde
In the chamber studies, formaldehyde was determined by the chromotropic
acid method. ' A 1% sodium bisulfite solution was employed as the
collection medium. Chamber air was drawn through two glass midget impingers
in series. The sampling duration was normally 30 minutes, and the flow
*If a bag initially contained 0.1 ppm 03 and 0.1 ppmV ethene, approxi-
mately 11 percent of the initially present ethene would be consumed in
8 hours by reaction with ozone. This percentage would increase for more
reactive olefins. For example, under the same initial conditions 43
percent of the initially present propene would be consumed in 8 hours.
51
-------
rates ranged from 600 to 700 ml rain . Flow rate was controlled by a cali-
brated hypodermic needle which was protected from overspray by a dessicant
cartridge and glass wool trap. Samples were analyzed with a Bausch and Lomb
Spectronic 100 spectrophotometer after treatment with chromotropic and
sulfuric acids. Calibration curves and blanks were prepared periodically
according to the referenced procedures. The reported data have not been
corrected for interferences from ethene, propene, or other hydrocarbons. In
the reported data (Volume 2 [Appendix B]), the time that has been assigned
to a wet chemical measurement for a chamber is the automated instrument
sampling period for that chamber which is closest to the beginning of the
bubbler sampling periods.
Condensation Nuclei
The number of condensation nuclei per cubic centimeter of sampled air
(CN) was determined with a Gardner Associates Type CN Small Particle Detector.
The sampling frequency varied from experiment to experiment. Normally CN
determinations were performed from one to four times per day. The samples
were taken directly from each chamber through a 1-m length of 4.8-mm ID
Teflon tubing. The manufacturer's calibration curve was used.
Freon-12
.—i
Freon-12 (difluorodichloromethane) was used at initial concentrations
of approximately 300 ppm as a tracer in several experiments which were
designed to quantify chamber leakage under various modes of operation. It
was measured by a Wilks Miran I Variable Filter Infrared Analyzer. This
instrument uses a single-beam spectrophotometer in conjunction with a 20-
meter variable pathlength White cell. When Freon-12 is present in the cell,
the spectrum of infrared light will show a strong absorption band (44 atm~ cm"
for 1 cm resolution) at a spectral frequency of 930 cm . Typical instru-
mental resolution in this spectral region is 20 cm . The instrument is
equipped with a linear absorbance accessory which produces an output signal
proportional to the Freon-12 concentration. This signal was recorded on a
strip chart recorder.
The Metal Bellows pump of the sampling system was used to flush air
continuously through the White cell. This was accomplished by connecting
the cell's intake port to the vent of the glass sampling manifold (see
Figure 11).
52
-------
Bags also containing mixtures of approximately 300 ppm Freon-12 in air
were prepared by injecting pure Freon-12 via a syringe into 100-liter bags
which contained breathing air. Periodic span checks were made using these
mixtures. The instrument's zero was established during the ambient air
portion of the sampling cycle.
Solar Radiation
Total solar radiation data reported in this study were collected by the
U.S. Environmental Protection Agency Division of Meteorology. The solar
radiometer, an Eppley Precision Spectral Pyranometer was located at a point
approximately 0.5 km from the RTI Smog Chamber Facility. This instrument
employs a thermopile sensing element and determines light intensity at
wavelengths longer than 295 nm. The hourly average values reported in
Appendix B were reduced from continuous strip chart records.
Environmental Variables
Other environmental variables reported in this study (see Appendix B)
are ambient air temperature at 3-hour intervals, the daily maximum tempera-
ture (T ), and the percent of possible minutes of direct sunshine (% SS).
IHcl3C '.
The % SS is determined by a Foster Sunshine Switch, which consists of two
photoelectric cells and a recorder. One cell is shaded from direct sun-
light; the other is not. These cells are connected such that the recorder
is actuated when the intensity of direct sunshine is sufficient to produce a
shadow. The temperature and the sunshine data are collected by the National
32
Weather Service (NWS) Forecast Office at the Raleigh-Durham Airport (RDU).
RDU is located at a distance approximately 10 km from the RTI Campus.
DATA REDUCTION AND HANDLING
Data have been handled in several forms in the current study. They
have been transformed from continuous or discrete values into discrete
computer-compatible values. Listings of these data are presented in Appen-
dixes A and B.
Strip chart records of 03> NO, N02> THC, CH4> and CO were manually
reduced into concentration units and entered onto coding forms. The 0^, NO,
and NO concentrations were reduced at the eighth minute of each 10-minute
sampling period. Strip chart traces of the THC, CH^, and CO chromatographic
53
-------
peaks were manually reduced into concentration units from the records of the
second 5-minute chromatographic cycle of each 10-minute sampling period.
These data were also entered at times corresponding to the eighth minute of
the sampling period.
Concentrations of oxidant, NO-, and HCHO that were determined manually
by wet chemical techniques were entered onto coding forms. The time assigned
to a wet chemical measurement for a chamber is that corresponding to the
automated instrument sampling period for that chamber which is closest to
the beginning of the bubbler sampling period. The manually determined CN
data were treated similarly.
Strip chart records of continuous solar radiation data were obtained
from the EPA Division of Meteorology. From these records, hourly average
values of solar radiation intensity were manually calculated and entered
onto coding forms.
Ambient temperature at 3-hour intervals, the maximum daily temperature,
and the percent of possible minutes of direct sunshine were recorded at the
32
NWS Forecast Office, RDU airport. Tabulated data were obtained and trans-
ferred to coding forms.
The coded data noted above were keypunched and processed into the
format shown in Volume 2 (Appendix B). The data processing program, in
addition to providing a convenient format for listing the coded data, also
calculated and listed time, NO , NMHC, and the cumulative solar radiation
X
(CUM-SR). The following formulae were used in these computations:
NO = NO + NO
A ^
NMHC = THC - CH4
h-1
CUM-SR = Z SR(i) + m SR (h)
i=0 60
Where SR(i) is the average solar radiation for the i
complete hour;
h is the number of the indicated hour; and
m is the number of minutes from the top of the indi-
cated hour until the indicated time.
It should be noted that the SR data are listed in Volume 2 (Appendix B) as
-2 -1
hourly averages in units of Langleys per minute (cal cm rain ). The
CUM-SR is a measure of the cumulative solar radiation that had occurred
_2
to the indicated time and is expressed in Langleys (cal cm ).
54
up
-------
Individual hydrocarbon data were handled differently than the data
noted above. Grab samples were collected manually and these samples were
determined with a gas chromatograph. The analog signals from the gas chro-
matograph were identified and transformed by an automatic data acquisition
system to species identifiers and concentration units which were printed by
a teletype. These data were subsequently entered onto coding forms. The
coded individual EC data were keypunched and processed into the format shown
in Volume 2 (Appendix A).
Time is listed in fractional hours (EST) in Volume 2. The times of
measurements were coded and keypunched as time of day in hours and minutes
(EST). These data were converted by the processing programs to fractional
hours (EST).
PROCEDURE
Two types of experiments were conducted in the present study: those
that involved the irradiation of a surrogate urban hydrocarbon mix in the
presence of NO and those that dealt with other chemical systems (see Tables 2
A
and 3). The basic research plan involved the urban mix and called for 3-day
smog chamber runs with four different target initial reactant concentrations,
four different dilution rates, and three different times at which dilution
was initiated (see Table 4). The procedure employed for urban mix experi-
ments is described in this subsection. The procedures employed in those
experiments that did not involve the urban mix were similar and therefore
are not addressed in this discussion.
A three-day smog chamber run requires four days of chamber activities.
On the day before a run is to start, the chambers are operated in the purge
mode. Starting at 0900 EST and lasting for six to eight hours each chamber
3 -1
is flushed with ambient air at a flow rate of up to 2.3 m min . At approxi-
mately 1500 EST the purging operation is terminated, each chamber is sealed,
and the cleanup operation is begun. Hydrocarbon and NO contaminants in the
chamber air are removed by recirculation of the chamber contents through the
3 -1
air purification unit for 8 to 12 hours at a flow rate of 0.28 m min . In
the current experimental programs, the humidity of the chamber contents was
not altered during the cleanup. This was accomplished by "by-passing" both
the humidifier and the desiccant columns during the cleanup operation. The
cleanup operation is terminated two hours before sunrise.
55
-------
The next step in the procedure is reactant injection. Appropriate
amounts of NO and N0? are sequentially introduced into each chamber from
separate gas cylinders using the reactant injection system (see Figure 10).
After the NO injection, the hydrocarbon mix, toluene, and CO are injected
j\
from separate cylinders using the same reactant injection system. The
injection procedure is completed 1 hour before sunrise. This allows time
for mixing and initial reactant sampling prior to sunrise.
In the current study, target initial [NO ] (20 percent N0») ranged
X «
between 0.1 and 1.0 ppm and target initial [NMHC] ranged from 1 to 10 ppmC,
and the target initial [CO] was 10 ppm. Specific target [NMHC]: [NO ]
A
combinations were 10 ppmC: 1.0 ppm, 5 ppmC: 0.71 ppm, 5 ppmC: 0.24 ppm and
1.0 ppmC: 0.1 ppm. In the 1975 urban mix experiments, acetylene had to be
introduced into the chambers by syringe injection of the pure gas, because
it had been omitted from the hydrocarbon injection mix. However, this was
not the case in the 1976 urban mix experiments, since acetylene had been
included in the 1976 hydrocarbon injection mix (see Table 12).
The contents of the chambers were sampled and monitored for the next
three days. Ozone, NO, N0_, THC, CH,, and CO were monitored once per hour
for each chamber. Typically, wet bubbler samples for oxidant, NO-, and HCHO
analyses were collected from each chamber twice daily at midmorning and at
mid-afternoon (see Volume 2 [Appendix B]) for exact times). Samples for
hydrocarbon analyses were collected in 10-liter Tedlar bags twice daily at
sunrise and late afternoon (see Volume 2 [Appendix A]) for exact times).
Dilution of chamber contents with purified air for 24 hours was employed
to simulate atmospheric transport conditions. Static operation (no dilution)
was used to simulate stagnation conditions. Following reactant injection,
each experiment was begun in the static mode and, if the experimental design
called for simulated transport, dilution was initiated on the first day at
one of three times: sunrise, NO-NO- crossover, or 1700 EST. One of four
dilution rates was employed such that 95, 77, 39, or 0 percent of a tracer
present at the start of dilution would have been removed by the end of the
24-hour dilution period. Dilution was terminated 24 hours after it had been
initiated. The chamber was then operated in the static mode until the
experiment was terminated at 1700 EST on the third day. The target experi-
mental conditions and the dates on which the corresponding experiments were
conducted are identified in Table 4.
56
-------
SECTION 5
RESULTS AND DISCUSSION
OVERVIEW
The results of smog chamber experiments conducted under EPA contracts
68-02-1296 and 68-02-2207 are discussed in this section. Data collected
during these experiments are tabulated in Volume 2. The experimental condi-
tions were identified previously in Tables 2 and 3. During these 60-hour
experiments a surrogate urban mix in the presence of NO was exposed to
A.
three daylight periods of natural irradiation and two nighttime periods of
darkness. Ancillary multi-day experiments involving other chemical systems
were conducted as well. The basic research program was designed to permit
the investigation of the influence of simulated transport on the ozone
*
generation by various combinations of the surrogate urban hydrocarbon mix
and NO from experiments conducted in smog chambers. The ancillary experi-
X
ments did not involve the urban mix, but instead,'were directed toward
exploring the ozone-generative potential of low reactivity hydrocarbons, at
conducting experiments for comparison (by EPA personnel) with results pre-
dicted by photochemical kinetics simulation model, and toward permitting
comparison of the behavior of the RTI and UNC outdoor smog chambers.
BASIC RESEARCH PROGRAM
/ i
The experiments considered in this study are identified in Table 4.
Eighteen four-chamber sets of experiments were considered: five sets of
static experiments, twelve sets of dilution experiments, .and one set of
multiple dilution experiments. Dilution was initiated at sunrise in five
sets, at the time of N0-N02 crossover in four sets, and at 1700 in three
sets. The parameters considered in this study were selected from the follow-
ing: the time at which dilution was initiated; the extent of dilution; the
initial NMHC and NO concentrations as well as their ratio; the [NMHC],
X
[NO J, and [NHMCj/[NO ] at sunrise of the second and third days; the maximum
X X
0- concentration on the first day; and the maximum, minimum, and net 0,
57
-------
concentrations on the second and third days.* The parameters may be deter-
mined for each experiment from the data tabulated in Volume 2 (Appendix B).
Analysis of Overall Data Set
The data on each of the three days of an experiment were stratified by
precursor concentration. The NO and NMHC concentration ranges used were
X
chosen arbitrarily such that the concentrations did not overlap and that
approximately an equal number of data would fall within each range. Since
the experiments were conducted under a variety of conditions, i.e., at
different temperatures, solar intensities, and times of the year, the stand-
ard deviations for the 0 concentrations within each range are large. Data
41
were rejected when permitted by Chauvenet's criterion.
Initially, the entire data set was considered. This data set is identi-
fied in Table 4 and is comprised of both static and dilution experiments.
These data were used to determine the effects of ozone precursor concentra-
tion and their ratio on the maximum concentrations of ozone produced on the
first day of an experiment. Results are presented in Figures 12, 13, and
14. Figures 12 and 13 suggest that in these experiments the maximum 0^
concentration generated on the first day generally increases with increasing
precursor concentration. The average CL concentration at 0.37 ppm NO in
O X
*Net ozone concentration on any day is the difference between the
maximum and minimum concentrations: A0~ = [0-] - [0^] . . The accumula-
tion of ozone in smog chambers is the net result of synthesis and destructive
processes. Ozone is synthesized by photochemical processes. At night in
the absence of sunlight, ozone-destructive processes prevail. During the
night following the first day of an experiment, the ozone concentration
declines as the ozone that was generated during the previous daylight period
is destroyed by homogeneous reactions with various reactants and products
and by heterogeneous interactions with the chamber walls. On the morning of
the second and third days, ozone synthesis begins with the reintroduction of
sunlight to the system. When the ozone synthesis rate begins to exceed the
destruction rate, the concentration profile passes through a minimum and the
ozone concentration increases. In the current investigation, the difference
between this minimum and the maximum concentration that accumulates on any
given day is known as the net ozone concentration. Since the initial (mini-
mum) ozone concentration is always zero on the first day, the first-day
maximum and net ozone concentrations are identical. This is not the case on
the second and third days. The first-day net or maximum ozone concentration
presumeably reflects the ozone production potential of a "fresh" chemical
system; whereas, the second-day net ozone represents the ozone production
potential after one diurnal cycle and the third-day net ozone represents the
ozone production potential after two diurnal cycles.
58
-------
1.2 r
1.0
a
a.
eTO.8
o
«8
e
o
u
o
M
O
0.6
0.4
0.2 •
X
X
0 2 4 6 8 10
Initial Hydrocarbon Concentration, ppmC
Figure 12. First-Day Maximum Ozone Concentrations Versus
Initial Hydrocarbon Concentrations for All
Experiments ( ) and for Static Experiments
( ) . Points represent mean values; error
bars represent ± one standard deviation; and
the absence of error bars indicates that only
one data point was available.
12
59
-------
1.2r
1.0
c
o
0.8
JJ
4-1
C
01
o
g
c
o
N
O
§
!
(0
<>•*
0.2
1
Figure 13.
0.2 0.4 0.6 0.8
Initial NO Concentration, ppm
1,0
First-Day Maximum Ozone Concentrations Versus
Initial NO Concentrations for All Experiments
(: ) and for Static Experiments ( ) .
Points represent mean values; error bars repre-
sent ± one standard deviation; and the absence
of error bars indicates that only one data point
was available.
60
-------
1.2 ,-
1.0
c
O.
c 0.8
o
C
OJ
o
§0.6
o
o
N
O
§ 0.4
0.2
0
Figure 14.
5 10
Initial Hydrocarbon-to-NO
15 20
Ratio, ppmC/ppm
25
First-Day Maximum Ozone Concentrations Versus Initial
Hydrocarbon-to-NO Ratios for All Experiments ( --- )
and for Static Experiments (-
-). Points represent
mean values; error bars represent ± one standard
deviation.
61
-------
Figure 13 is lower than might be anticipated. Three of the five data compris-
ing this set are from experiments in which 95 percent dilution was initiated
at sunrise and may be responsible for the indicated behavior. The results
in Figure 14 are highly variable and permit few generalizations concerning
the influence of initial NMHC/NO ratio on the resulting maximum 03 concen-
tration.
Next, the data were further stratified to consider CU production under
static conditions. First-day data from both static experiments and experi-
ments in which dilution was initiated at 1700 were used, since for these
dilution experiments, dilution was not commenced until after the maximum 0»
concentration had been achieved. The results for static conditions were
treated similarly to those from the entire data set. Results are presented
in Figures 12, 13, and 14 where they may be compared with those of the
entire data set to permit a qualitative assessment of the effects of dilution.
As shown in Figures 12 and 13, for these experiments the first-day [0-]
j in&x
increases with increasing precursor concentration. Although the results in
Figure 14 are highly variable, the maximum first-day 0,, concentration occurs
at an initial NMHC/NO ratio of approximately 9. In general, the behavior is
*•>
similar for both sets of data. The effects of dilution on first-day ozone
production are apparent in Figures 12, 13, and 14. In each case higher
maximum 0« concentrations were achieved for the static (no dilution) experi-
ments than for the entire set of static and dilution experiments. More
detailed assessments of the effects of dilution on 0~ production are pre-
sented in a subsequent subsection in which the results of individual experi-
ments are analyzed.
To determine whether the net 0_ concentrations that were generated on
subsequent days were affected by dilution, third-day results from all dilution
experiments were considered. The resulting net 0,. concentrations are compared
in Figures 15 and 16 with those from the static experiments. Similar trends
are suggested by both figures—third-day net 0« concentrations may increase
slightly with increasing precursor concentration. In general, however, the
variability of the data does not permit distinction between static and
diluted conditions. Such a distinction is sought in subsequent analyses of
individual experiments.
62
-------
ON
O.Ar
0.3
g
•H
4-1
0}
4-1
a
o °-2
g
o
01
a
o
N
0 0.1
0)
o—' cr-;
_L
JL
JL
0.5
1.0 1.5 2.0 2.5
Sunrise Hydrocarbon Concentration, ppmC
3.0
3.5
Figure 15.
Third-Day Net Ozone Concentrations Versus Sunrise Hydrocarbon Concentrations for Dilution
Experiments ( ) and for Static Experiments ( ). Points represent mean values;
error bars represent ± one standard deviation; and the absence of error bars indicates
that only one data point was available.
-------
0.4r
a
a.
c
o
0.3
03
S-i
4J
§ 0.2
c
o
C!
O
N
o 0.1
.u
0)
Z
4-
10 20 30 40
Sunrise NO Concentration, ppb
50
60
Figure 16. Third-Day Net Ozone Concentration Versus Sunrise NO
Concentrations for Static Experiments ( ) and for
Dilution Experiments ( ). Points represent mean
values; error bars represent ± one standard deviation;
and the absence of error bars indicates that only one
data point was available.
64
-------
Analysis of Individual Experiments
The nine sets of four-chamber experiments that were considered in this
analysis are also identified in Table 4. Data from these experiments were
examined to determine the effects of such independent variables as initial
[EC] and [NO ], dilution rate, and time of initiation of dilution on the
X
maximum and net concentrations of ozone produced during each of the three
days of an experiment.
The bar graphs of Figures 17, 18, and 19 may be used to compare maximum
and net ozone production during static experiments with those from experiments
conducted at various dilution rates. Experiments conducted at each of the
four target initial [EC] and [NO J conditions were considered as separate
cases in each of these figures (case 1: 10 ppmc of HC/1 ppm of NO ; case 2:
&L
5 ppmC/0.24 ppm; case 3: 5 ppmC/0.71 ppm; and case 4: 1 ppmC/0.1 ppm).
Each of the figures is similar except for the time of initiation of dilution
(Figure 17 at sunrise, Figure 18 at NO-NO? crossover, and Figure 19 at
1700). Also, the first bar for each day represents the results of the
corresponding static experiment for that particular case.
Net 0_ concentrations are represented by the shaded areas and the
minimum 0- concentrations by the nonshaded areas. On the first day, the
maximum and net 0« concentrations are identical and are shaded. On the
second and third days, the sum of the shaded and nonshaded areas represents
the maximum 0» concentration.
For the experiments in which dilution was initiated at sunrise (see
Figure 17), the 0 maxima decrease with increasing dilution rate on the
first and second days, and in three of four cases on the third day. First-
day 0 maxima are greater than second-day values, which in turn exceed
third-day values with only a single exception. Second- and third-day minimum
0 levels decrease with increasing dilution rate. On the second day, the
largest net 0_ occurs at 39 percent dilution—exceeding levels under both
static and 95 percent dilution conditions. On the second day, at 95 percent
dilution, the net [0_] exceeds the static levels in two of four cases.
Third-day net (O.J , compared to the second-day levels, increase for the
static experiments but decrease for the dilution experiments with one excep-
tion.
65
-------
0.
o.
c
•u
o
c
o
OJ
c
o
1.2 -
_.
1.0-
0.8-
0.6-
0.4-
0.2-
n
•m Case 1
vv
rft
HC/NO
X
\
I
•MMMI
%
i
0 39 95 0 39
Day 1 Day
1.2-
1.0-
0.8-
0.6-
0.4-
0.2.
n
Case 2
HC/NO
X
7^
I
1
I
1
I
>x
>x
1
/^
P7
I
t
••••••
i
P
0 39 95 0 39
Day 1 Day
-
= 10/1.0
'Tj^l
s^l
vv^
•H^B
I
^^^H
I
Case 3
HC/NO - 5/0.71
5C
1
1
^•^^
| || i || g
95 0 39 95 0 39 95 0 39 95 0 39 95
2 Day 3 Day 1 Day 2 Day 3
= 5/0.24
n
^^v
1
« .
r%
'yy
Y/
Case 4
HC/NO = 1/0.1
X
I
1
%
%
'//
VS
V/
1
K, [fi ^ ffl M
95 0 39 95 0 39 95 0 39 95 0 39 95
2 Day 3 Day 1 Day 2 Day 3
Figure 17. Maximum and Net Ozone Concentrations for Static Experiments
and for Dilution Experiments Initiated at Sunrise (Shaded
Areas are Net [03] and Open Areas are Minimum [03]).
Note that the first-day maximum net [0 ] are identical
and that on subsequent days the maximum [0,,] equals the
of the minimum and the net ozone concentrations.
sum
66
-------
Case 1
HC/NO., = 10/1.0
0 39 77 95
Day 1
0 3977 95
Day 2
0 39 77 95
Day 3
1.2 -
: i.o -
2 °-8
u
e
-------
a
a.
a.
a.
C
II
u
C
o
o
O
N
O
1.2 -
1.0-
0.8 -
0.6 -
0.4 -
0.2 -
•m
o
1
1
i
1
I
^ Case
HC
•«••
^
1
/NC
••••••••i
1
^>
• 0 39 7795 0 39
1
) = 1
X
i
77 95
Day 1 Day 2
1.2 -
1.0 -
0.8 .
0.6 -
0.4 -
0.2 -
r\
Case
HC/NO
1
1
^
1
1
1
1
1
%
y/
1
1
//
1
1
1
1
1
__
^
•^•H
^
2
0/1.0
I
%
0 39
%
77
i
95
Day 3
!
:
Case 3
1
i
1
1
HC/NO = 5/0.71
0 3977 95 0 39 7795 0 39 7795
Day 1 Day 2 Day 3
= 5/0.24
X
i
0 39 77 95 0 39 7795
Day 1 Day 2
I
^•••M
^
i
^
Case 4
HC/NO = 1/0.1
X
I
i
I
1
S—r—\ n__ , (7Z
j j 1 y>>'
rrST/WWfa TrtftZvM'.
039 7795 039 7795 0 3977 95 0 397795
Day 3 Day 1 Day 2 Day 3
Figure 19. Maximum and Net Ozone Concentration for Static Experiments and for
Dilution Experiments Initiated at 1700 (Shaded areas are net [0_]
and open areas are minimum [0»J.) Note that the first-day maximum
and net [0_J are identical ana that on subsequent days the maximum
[0_j equals the sum of the minimum and the net ozone concentrations.
68
-------
Results from the experiments in which dilution was initiated at NO-NO
crossover are presented in Figure 18. If the first-day results from the 27
August 1975 experiment (39 percent dilution) are not considered, then the
first-day 0^ maxima decrease with increasing dilution rate. Across all
these experiments, the second- and third-day maxima also decrease with
increasing dilution rate. First-day 0 maxima are greater than second-day
values, which in turn exceed third-day levels. Ozone minima decrease with
increasing dilution rate on the second day and in two of four cases on the
third day. Second-day net [CL] is greater for dilution than for static
conditions, with the largest increase at 77 percent in three of four cases.
Compared to second-day levels, third-day net [0_] increases for static
o
experiments but decreases for dilution experiments.
Figure 19 depicts the results from experiments in which dilution was
initiated at 1700. Since for these conditions dilution was initiated after
the first-day 0« maxima were achieved; comparison of the first-day maxima
permits an assessment of precision under various environmental and opera-
tional conditions. Except for the 4 November 1975 experiment (39 percent
dilution), which may have been heavily influenced by seasonal effects, the
precision is excellent. With a single exception, the second-day 0~ maxima
decrease with increasing dilution rate and, except at 95 percent dilution,
this occurs on the third day as well. First-day 0_ maxima are greater than
second-day values, which in turn exceed third-day levels except at 95 percent
dilution. Ozone minima decrease with increasing dilution rates on the
second day and, with a single exception, on the third day. On the second
day the largest net [0 ] occurs at 77 percent dilution. Net [03J on the
second day at 39 and 77 percent dilution exceeds the static values with a
single exception. In three of four cases, however, the static levels exceed
those at 95 percent dilution. The second-day net [03] are greater than the
third-day levels for 39 and 77 percent dilution; whereas the third-day
levels are greater than the second-day values for the static and 95 percent
dilution conditions. Third-day net [03] at 95 percent dilution exceeds
static levels in three of four cases and levels at 39 and 77 percent dilu-
tion. Static third-day net [Oj also exceeds levels at 39 and 77 percent
dilution in three of four cases.
69
-------
The maximum and net ozone results that are presented in Figures 17, 18,
and 19 are displayed in an alternate format in Figures 20 and 21. Those
experiments in which dilution was initiated at the same times are presented
in each of three separate graphs within each of these figures. Within each
graph the three run days are treated separately. The ozone concentrations
achieved at each of the four target initial concentrations on a given day
are plotted. It should be noted that Case 1 corresponds to target initial
HC/NO conditions of 10 ppmC/1 ppm, Case 2 to 5/0.24, Case 3 to 5/0-71, and
X
Case 4 to 1/0.1. Thus, within each day the cases are ordered on the graphs
such that they are in increasing order of both hydrocarbons and oxides of
nitrogen concentrations (order: 4, 2, 3, 1). The connected symbols repre-
sent a common target dilution rate. Results from both static and dilution
experiments are included to permit their comparison. It should also be
noted that since first-day maximum and net ozone concentrations are identi-
cal, the first-day results in Figures 20 and 21 are also identical.
Ozone maxima shown in Figure 20 generally increase with increasing
precursor concentration. First-day ozone maxima display an increasing trend
with both hydrocarbon and NO concentrations. The erratic first-day behavior
of Case 3 is due to the low HC/NO ratio for this case. Under these condi-
tions, NO is in excess and the formation of 0~ may be slowed by NO inhibi-
x j
tion. Thus, the system may tend to be light-limited and therefore more
sensitive to environmental variables than the other cases. The erratic
behavior noted on the first day of Case 3 is not present on the second and
third days.
Maximum ozone concentrations on both the second and third days generally
increase with increasing first-day precursor concentration and also with
increasing first-day maximum ozone. Thus, based on maximum daily concentra-
tions , it is difficult to separate the influence of second- and third-day
ozone generation by precursors remaining after the first day's reaction from
the fraction of first-day ozone that remains on subsequent days. This
consideration prompted the development of the net ozone analysis.
Under both static and dilution conditions (except 95 percent at 1700),
first-day ozone maxima are greater than second-day values, which in turn are
greater than third day levels.
70
-------
Dilution Begun at Sunrise
Dilution Begun at Crossover
e
(X
PL.
C
O
(0
M
n
01
1.2
1.0
0.8
E 0.6
g
N
O
1 0.4
0.2
O
X
'A
x—*
D
111
423142314231
Day 1 Day 2 Day 3
423142314231
Day 1 Day 2 Day 3
Dilution Begun at 1700
O
O 95% Dilution
O 77% Dilution
A 39% Dilution
I Static
423142314231
Day 1 Day 2 Day 3
Figure 20. Maximum Ozone Concentrations Plotted by Case Number and Day for Dilution
Beginning at Sunrise, 1700, and Crossover.
-------
Dilution Begun at Sunrise
Dilution Bep.un at Crossover
423142314231
Day 1 Day 2 Day 3
423143214321
Day 1 Day 2 Day 3
Dilution Begun at 1700
O
U 95% Di hit ion
O 77% Dilution
A 39% Dilution
X Static
423142314231
Day 1 Day 2 Day 3
Figure 21. Net Ozone Concentrations Plotted by Case Number and Day for Dilution
Beginning at Sunrise, 1700, and Crossover.
-------
In no case examined in the current study did dilution cause first-day
ozone levels to exceed those of corresponding static experiments.* Although
increased first-day ozone concentrations under dilution conditions have been
42 43
reported ' , no attempt was made in the current study to duplicate the
reported conditions. During 18 November 1975 experiments, however, the
maximum ozone concentrations generated under different dilution conditions
were comparable. In this experiment each chamber had the same target ini-
tial EC and NO conditions (5.0 ppmc and 0.24 ppm). Different dilution
A
rates were initiated in each chamber at NO-NO„ crossover. Under static
conditions the [0.1 was 0.49 ppm; at 39 percent dilution 0.49 ppm; at 77
O filo A
percent 0.39 ppm, and at 95 percent 0.32 ppm. Thus, it seems likely that
under the appropriate conditions, maximum ozone concentrations generated
under dilution conditions could exceed the corresponding static levels.
For those cases in which dilution was begun at sunrise or at crossover,
first-day maximum ozone concentrations are generally reduced by dilution.
This reduction is most apparent at 95 percent dilution. The impact of
dilution on first-day [0_] is more apparent when dilution was initiated
O IQ3X
at sunrise than at crossover.
Second- and third-day static 0~ maxima are greater than the concentra-
tions achieved in experiments in which dilution occurred. At a fixed time
of initiation of dilution, maximum second- and third-day 0. concentrations
decrease with increasing dilution rate (except 95 percent at 1700). Although
dilution acted to reduce second- and third-day [0_] , the reduction was
j ID3X
always less than proportional to the extent of dilution at the time of
occurrence of the maximum concentration.
The second- and third-day net ozone concentrations presented in Figure
21 display a slight trend toward increasing with increasing first-day pre-
A
cursor concentration and also with increasing first-day maximum ozone levels.
If it is assumed that the amount of ozone remaining from the first day does
not influence second- and third-day ozone generation, then these results
suggest that the apparent trend is caused by increased absolute quantities
of precursors remaining on subsequent days after the first day's reaction.
*Note that experiments in which dilution was initiated at 1700 were not
considered, since the late introduction of dilution could not have influenced
the first-day maxima.
73
-------
It should be noted, however, that the net ozone concentrations on second and
third days are frequently of the same magnitude as concentrations observed
in purified air irradiations. Thus, the extent to which the chambers them-
selves affect the second- and third-day results cannot be defined at this
time.
In general, the second-day net ozone concentrations in dilution experi-
ments are greater than the corresponding third-day levels (except for 95
percent dilution at 1700).* For static conditions, however, third-day net
ozone concentrations are greater than second-day levels.
With few exceptions (95 percent dilution at sunrise and at 1700) the
second-day net ozone concentrations in dilution experiments are greater than
or equal to those for static experiments. Second-day net ozone concentra-
tions generated at 39 and 77 percent dilution generally exceed those for
static and 95 percent dilution. In contrast, except for case 4, the third-
day net ozone concentrations in static experiments are greater than those
for dilution experiments. There are no apparent trends for third-day net
results under dilution conditions.
The amount of net ozone produced on the second and third day falls
between 0.08 and 0.30 ppm in 80 percent of the cases. Thus, in the RTI smog
chambers, aged photochemical systems displayed relatively high ozone-
generative potentials.
The following generalizations may be drawn from inspection of Figures
17 through 21.
For those experiments in which dilution was initiated before the
maxima occurred, maximum first-day ozone concentrations are reduced
under dilution in comparison to static conditions. In no case
examined in the current study did dilution cause first-day ozone
levels to exceed those of the corresponding static experiments.
Second- and third-day ozone maxima also decrease with increasing
dilution rate. The decrease is less than proportional to the
extent of dilution.
"'Under these conditions, dilution was occurring at the time of maximum
[0,.] on the second day; whereas dilution was not occurring on the third day.
Thus, the impact of dilution at a high rate (95 percent) on the second day
was sufficient to reduce the observed second-day ozone concentrations below
those achieved during static conditions on the third day.
74
-------
First-day 03 maxima are greater than those produced on the second
day, and second-day maxima are greater than those produced on the
third day under static and dilution conditions.
On both the second and third days, the minimum 0_ concentrations
decrease with increasing dilution rate.
The second-day net ozone concentrations at 39 or 77 percent dilu-
tion are greater than at the corresponding levels under static
conditions.
The second- and third-day net ozone concentrations generated under
dilution conditions may be greater than the corresponding levels
under static conditions, although the maxima are lower. Thus, the
maximum second- and third-day ozone concentrations are reduced by
dilution but the ozone-generative potential is not.
For dilution conditions, the second-day net ozone concentrations
exceed third-day levels, while for static conditions the third-day
levels exceed those for the second day.
Second- and third-day net [0,,] generally range between 0.08 and
0-30 ppm. Thus the ozone-generative potential of aged photochem-
ical systems in the RTI smog chambers is high and usually exceeds
0.08 ppm.
For experiments in which dilution was initiated before the first-day
maximum 0_ concentration occurred, dilution is expected to influence the
behavior of the chemical systems in comparison to those of the corresponding
static cases. On the average for static experiments, N0-N02 crossover was
achieved 2.2 hours after sunrise at approximately 0734 EST. Chamber 2 with
the highest target initial HC/NO ratio (5.0/0.24) was the first to achieve
A
crossover at 1.6 hours after sunrise. The lowest initial target HC/NOx
ratio was in Chamber 3 (5.0/0.71), which was the slowest to achieve cross-
over at 3.1 hours after sunrise. Chambers 1 and 4 required intermediate
times, 1.7 and 2.4 hours. The time to crossover may be reduced under dilu-
tion conditions.42'43 Examination of the current data set for such compari-
sons, however, revealed too few cases to yield conclusive results.
The time required to achieve the first-day maximum 0^ concentration was
reduced under dilution conditions. In comparison to the time required under
static conditions, the time to [03]max was reduced when dilution was initi-
75
-------
ated at NO-NO crossover. The time to [Oj was reduced to a greater
extent when dilution was initiated at sunrise. On the average, t^3Jmax was
achieved at 9.4 hours after sunrise for the static runs. When dilution was
initiated at NO-NO crossover, [0-] was achieved at 8.7 hours after
£ j 01a2C
sunrise, while the corresponding time requirement when dilution was initi-
ated at sunrise was 7.8 hours.
The effect of dilution on the absolute magnitude of the first-day
[0 ] can be explored by examination of Figure 22. The 0» concentrations
3 0)3 X <3
in these bar graphs are represented as percentages of the first-day [0-]
3 ma A
achieved under static conditions. These percentages are given at each of
the tested dilution rates for each of the four target initial HC/NO ratios.
A
Also shown in these illustrations by broken lines is the theoretical percent-
age of the static [0_J that would have resulted if the ozone produced
o max
with dilution were equal to the fraction of the chamber's volume that was
present when the [00] was achieved.
* 3 max
In 13 of 20 dilution experiments, the ozone production exceeded the
quantity expected if first-day ozone production were proportional to the
extent of dilution. Although seasonal effects may account for the dis-
crepancy between theoretical and achieved [0_] for the 95 percent sunrise
r J 3 max r
dilution experiments of 29 September 1975, in only three of the eight sunrise
dilution cases does the measured [0-] exceed the theoretical value.
In 10 of 12 crossover dilution cases the measured [0_] exceeds the
3 max
theoretical value. For many of these experiments the theoretical value is
exceeded by over 30 percent.
These results suggest that in photochemically reacting systems of
surrogate urban hydrocarbons and NO , ozone production is sensitive to the
A
time at which dilution is initiated. When dilution is begun at sunrise,
ozone production may be reduced more nearly in proportion to the extent of
dilution than when dilution is initiated at NO-NO- crossover. This behavior
is somewhat surprising when it is considered that the average delay between
sunrise and NO-NO, crossover is only 2.2 hours — from 0524 to 0734 EST.
When dilution is initiated at sunrise, the photochemical reactions are
also being initiated. Dilution removes precursors that are important to the
establishment and maintenance of various sequences of reactions instrumental
in the generation of ozone. When dilution is initiated after photochemical
76
-------
HC/NO =
x
10/1
Chamber 1
100.
CO
a 60-
« 40-
20-
HC/NO =
X
5/0.24
Chamber 2
m
//s//
II
I
HC/NO =
X
5/0.71
Chamber 3
0 39 77 95
Target
HC/NO =
x
1/0.1
Chamber 4
,0
0 39 77 95 0 39 77 95 0 39 77 95
Dilution (Initiated at Sunrise)
100
*80-
CO
Chamber 1
Chamber 2
Chamber 3
20-
Chamber 4
f///////////
mm,
- /////Y// '//
mm
/v
iill
iill
ny/y/v/
y/wy/v/
0 39 77 95
Target
0 39 77 95 0 39 77 95 0 39 77 95
Dilution (Initiated at Crossover)
Figure 22.
Maximum Ozone Concentrations Under Dilution Conditions
Relative to the Maximum Ozone Concentration Obtained
Under Static Conditions (Shaded area represents the
percentage of the static maximum ozone concentrations
achieved under dilution conditions; the broken line
represents the percentage of the static maximum con-
centration estimated by applying the appropriate dilu-
tion factor to the static maximum ozone concentration.)
77
-------
initiation reactions have occurred, the resulting oxidation of NO to NC^ is
well on its way toward establishing an appropriate ratio to permit ozone
accumulation. Thus the finding that the [0_J is sensitive to the time of
j max
initiation of dilution does not seem unreasonable in terms of the current
understanding of the sequence of events leading to photochemical ozone
generation. It is likely, however, that additional experimental investiga-
tions supplemented with modeling efforts will be required to identify the
detailed mechanistic subtleties responsible for the observed behavior.
The findings that first-day ozone-generative potential is sensitive
both to the dilution rate experienced by the reacting system and to the time
at which dilution is initiated have implications to the real atmosphere.
The descriptive discussion of urban meteorology and chemistry in Section 3
suggests that significant dilution begins at 0830 to 0930—3 to 4 hours
after sunrise. The average time of crossover in the current study was 0734,
2.2 hours past sunrise. In addition, it is suggested that early morning
(0600 to 0900 CST) hourly average NO data indicate that up to 80 percent of
A
the NO may exist as N09. Thus, if the St. Louis atmosphere is representa-
A £i
tive of most urban areas and if the current findings may be extrapolated to
the larger dilution rates that prevail in an urban atmosphere, then the
effects of atmospheric mixing on maximum urban ozone concentrations are less
than proportional to the extent of dilution of the air parcel.
EXPERIMENTS CONDUCTED TO PROVIDE DATA FOR MODEL TESTING
Eight experiments were conducted in October of 1976 to provide EPA with
data for model verification and testing. In these experiments, mixtures of
propene and NO were irradiated. Results from these experiments are summa-
X
rized in Table 14, and individual data are presented in Volume 2 (Appendixes
A and B).
The experiments conducted on 6 October in Chamber 3 and on 11 October
in Chambers 3 and 4 can be compared directly. On 6 October the sunlight
profile was "choppy" after 1015 EST. The solar radiation integrated both to
the time of [0 J and for the complete days was slightly higher on 11
j ID3X
October than on 6 October. In spite of the differences in irradiation, a
higher ozone concentration was observed on 6 October. Inspection of the
maximum ambient temperature levels on these two days suggests that the
78
-------
TABLE 14. SUMMARY OF RESULTS FROM EXPERIMENTS CONDUCTED TO
PROVIDE DATA FOR MODEL TESTING3
T ISR
max
Run °C Langleys
10-6-76 25.0 377
HC
[HC]ib
[NO ]ic
1 X4
HC/NO
d
txo
[OoJ C
10-11-76 18.9 420
HC
[HCji
[NOXH
HC/NOx
t
[OJ
1 3 max
Chamber 1
Propene
1.35 (1.5)e
0.14 (0.1)
9.6 (15.0)
8.39
0.39
Propene
0.68 (0.7)
1.57 (1.5)
0.43 (0.47)
g
0.00
Chamber 2
Propene
0.64 (0.7)
0.52 (0.5)
1.2 (1.4)
10.34
0.08
Propene
2.36 (3.0)
1.68 (1.5)
1.4 (2.0)
10.17
0.02
Chamber 3
Propene
1.21 (1.5)
0.53 (0.5)
2.3 (3.0)
9-45
0.47f
Propene
1.26 (1.5)
0.55 (0.5)
2.3 (3.0)
9.59
0.37f
Chamber 4
Propene
- (o)
0.098 (0.1)
- (-)
14.63
0.009
Propene
1.17 (1.5)
0.50 (0.5)
2.3 (3.0)
9-83
0.25
I
3For individual data see Volume 2 (Appendixes A and B).
bUnits ppmC of propene only; total hydrocarbon concentration slightly higher;
see Appendix A.
c
Units ppm.
dt is time of NO-NO- crossover in hours EST.
xo 2
eNumbers in parentheses represent target values.
fMixing fan turned off shortly after injection.
Crossover did not occur.
79
-------
6.1° C warmer conditions on 6 October may have been partially responsible
for the higher ozone concentration on this day. This is consistent with the
results of propane irradiations discussed in a subsequent section of this
report.
The effects of stirring were examined in Chambers 3 and 4 on 11 October.
Although a higher [0.] was observed in this single experiment with the
j
mixing fan turned off shortly after reactant injection, the initial condi-
tions were not identical, and this difference may be responsible for the
apparent discrepancy. It is recommended that additional experiments be
conducted to address this issue.
A comparison of results from experiments conducted in the RTI outdoor
chambers with results from the evacuable indoor chamber at the Statewide Air
Pollution Research Center (SAPRC) at Riverside, California, is presented
in Table 15- The agreement is generally good: the [0 ] levels in the
j
RTI experiments generally exceed those of the SAPRC experiments . The apparent
agreement may be fortuitous in view of the differences in operating condi-
tions for the two very different types of smog chambers. Ambient diurnal
temperature and solar radiation profiles define the environmental conditions
employed at the RTI facility. The duration of irradiation is normally 12
hours, from sunrise until sunset. In contrast, the SAPRC facility is operated
at fixed temperature and irradiation intensity. The duration of SAPRC
experiments is normally 6 hours.
OZONE-GENERATIVE POTENTIAL OF LOW REACTIVITY HYDROCARBONS
Hydroxyl radical attack on hydrocarbons is thought to be the major
removal mechanism for these species in reacting photochemical HC/NO systems.
X
The rate constant for OH attack on olefinic hydrocarbons is much greater
than that for alkane hydrocarbons. This accounts for the more rapid deple-
tion of olefins in comparison to alkanes that has been observed both in
ambient air and in smog chamber studies. Hydrocarbon data collected in the
urban mix runs of the current study and presented in Volume 2 (Appendix A)
confirm these observations.
The ozone-generative potential of many organics was established based
on studies conducted in smog chambers. In an effort to reduce photochemical
formation of ozone in polluted atmospheres , regulations on solvent emissions
80
-------
TABLE 15. COMPARISON OF RTI OUTDOOR SMOG CHAMBER RESULTS
WITH RESULTS FROM SAPRC INDOOR CHAMBER17
RTI
EC
RTI
RTI
RTI
EC
EC
EC
EC
EC
Run
No. 1 10-6-76
- 11
No. 3 10-6-76
No. 3 10-11-76
No. 4 10-11-76
- 7
- 13
- 57
- 52
- 51
Ta
°C
25
29
25
18
18
30
29
30
30
30
[HC]i
ppmC
.0
.7
.0
.9
.9
.9
.2
.1
.1
.0
1
1
1
1
1
1
1
1
1
1
.35
.34
.21
.26
.17
.41
.50
.59
.65
.66
HC/NOx
f°3]max
ppmC/ppm
9
9
2
2
2
2
2
2
2
2
.6
.9
.3
.3
.3
.4
.6
.7
.8
.9
0
0
0
0
0
0
0
0
0
0
ppm
.39
.23
.47b
.37
.25
.38C
.37
.24
.31
.29
aT is maximum daily temperature for RTI experiments; T is average run tempera-
ture for SAPRC experiments
bMean RTI [Oj _ = 0.36 ± 0.11 ppm.
IHclX
CMean SAPRC [0_] = 0.32 ± 0.06 ppm.
81
-------
were established to limit emissions of reactive orgaaics and to be less
restrictive on emissions of organics considered to be nonreactive. Through
substitution of less reactive organics for the more reactive ones, not only
could an increase in organic emissions result, but the substitute compounds
such as alkane hydrocarbons, by virtue of their low reactivity, could persist
in air parcels transported downwind from urban areas.
Results from the current study suggest that on the second and third
days, early morning HC/NO ratios were high, alkane hydrocarbons persisted,
A:
the ozone-generative potential was high, and the net ozone on both days
generally exceeded 0.08 ppm. Other smog chamber studies have shown alkane
hydrocarbons to produce considerable amounts of ozone when irradiated in the
q o£ AA-A7
presence of NO at high HC/NO ratios. ''
X A
Dimitriades has recently suggested a procedure to identify low reactiv-
ity organics as those which would not under any circumstances yield ozone
47
concentrations greater than 0.08 ppm. The results of EPA smog chamber
studies conducted at initial HC/NO conditions of 4 ppm (v/v)/0.2 ppm were
A
reported to suggest that propane is the boundary point between reactive and
nonreactive organics. The experiments discussed in this section of the
report were conducted to examine the ozone-generative potential of the
nonreactive hydrocarbons, propane, ethane, and acetylene in light of the
results of Dimitriades.
On 8 September 1976 four experiments were conducted with low reactivity
hydrocarbons. Ethane, acetylene, and propane were investigated separately
in each of three chambers. Target initial conditions were 4 ppm (v/v) of
the hydrocarbon and 0.2 ppm NO (20% N09). The fourth chamber was employed
X fc
as a control: 0.2 ppm NO was added to purified air in this experiment.
X
First-day maximum 0. concentrations were high in the first three chambers:
ethane produced 0.32 ppm; acetylene 0.45 ppm; and propane 0.65 ppm. A
maximum ozone concentration of only 0.02 ppm was observed in the control
experiment.
The significant differences between these results and those of Dimitri-
47
ades, prompted experiments to determine if the high concentrations were
caused by chamber-related effects. Matched experiments involving the four
RTI chambers and two UNC chambers were conducted on 20 September 1976. The
tThis experiment is discussed in detail in a subsequent section of this
report.
82
-------
same target initial conditions were employed that were used in the 8 September
propane experiment. The average [0 1 produced in the six chambers was
j in
-------
TABLE 16. SUMMARY OF RESULTS FOR LOW REACTIVITY EXPERIMENTS'
Run
9-8-76
HC
[HC]ib
[NO ]ic
'53>*aXC
9-20-76
HC
[HC]i
(N0x]i
t°3]max
10-27-76
HC
[HC]i
[N0x]i
[°3]max
11-5-76
HC
[HC]i
[N0x]i
[°3]max
T ISR
max
°C Langleys Chamber 1
31.7 567
Ethane
6.9(8.0)e
0.17(0.2)
10.36
0.32
29.4 383
Propane
10.2(12.0)
0.21(0.2)
9.07
0.20
9.4 366
Propane
5.5(6.0)
0.10(0.1)
10.65
0.039
11.7 328
Chamber 2
Propane
11.7(12.0)
0.18(0.2)
8.85
0.65
Propane
11.1(12.0)
0.22(0.2)
8.68
0.20
Propane
10.4(12.0)
0.21(0.2)
11.04
0.028
Chamber 3
Acetylene
5.6(8.0)
0.16(0.2)
10.33
0.45
Propane
11.3(12.0)
0.21(0.2)
8.93
0.18
Propane
3.1(3.0)
0.06(0.05)
10.22
0.047
Propane
2.6(3.0)
0.06(0.05)
10.47
0.041
Chamber 4
Control
-(0)
0.17(0.2)
14.10
0.029
Propane
10.3(12.0)
0.21(0.2)
9.12
0.21
Control
-(0)
0.06(0.05)
f
0.002
Propane
2.8(3.0)
0.05(0.05)
g
0-062
g
For individual data see Volume 2 Apppendixes A and B).
Units ppmC of target compound only; total hydrocarbon concentration slightly
higher; see Appendix A.
-»
"Units ppm.
t is the time of NO-NO- crossover in hours EST.
a
"Numbers in parentheses represent target values.
Crossover did not occur.
Crossover did not occur; NO was injected as N0_.
X ^
84
-------
concentrations is presented in Table 17. This comparison of results from
the current study with those reported by other investigators45'47 suggests
that the maximum ozone concentrations generated by low reactivity compounds
are highly temperature sensitive. In addition, they show a favorable compar-
ison between the [03]max values of the 8 September RTI outdoor smog chamber
experiments and those of the NAPCA indoor smog chamber.45 Mechanistic
justification for the extreme temperature sensitivity of ozone production by
the low reactivity organics cannot be offered based on these results.
Isolation of the causal factors from among the tens and perhaps hundreds of
chemical reactions involved in the photooxidation of propane is difficult
without the aid of a computer simulation model. Recent findings however of
the strong temperature dependence of PAN decomposition and possibly of
pernitric acid decomposition as well suggest that these reactions could be
instrumental in accounting for the temperature sensitivity of ozone produc-
tion by low reactivity organics.
Results from experiments conducted in the RTI smog chambers suggest
that the criteria for determining reactivity should be modified. If a low
reactivity hydrocarbon must not yield ozone in excess of 0.08 ppm in smog
chamber investigations, then propane, ethane, and acetylene must be considered
to be reactive organics.
The use of smog chamber data to establish the reactivity of organics is
a controversial topic. Detailed discussion of the reactivity issue is
beyond the scope of this report. A critical review of hydrocarbon reactiv-
ity has been published recently and should be consulted for more information
48
on this subject.
COMPARISON BETWEEN TWO OUTDOOR SMOG CHAMBER FACILITIES: MATCHED EXPERIMENTS
To investigate the extent to which matched initial reactant conditions
would yield consistent results in outdoor smog chambers, matched experiments
were jointly conducted at the University of North Carolina's Ambient Air
Research Facility (UNC) and at the Research Triangle Institute's Outdoor
Smog Chamber Facility (RTI). Each matched experiment was conducted on the
same day at both facilities. Since the two facilities are located within 32
kilometers of each other, similar diurnal profiles of light intensity and
ambient temperature are seen at each facility on any given day. These
85
-------
TABLE 17. MAXIMUM OZONE CONCENTRATIONS GENERATED IN VARIOUS SMOG
CHAMBERS USING LOW REACTIVITY HYDROCARBONS
a
Maximum Ozone Concentration
Chamber
RTI (9-8-76)
RTI/UNC (9-20-76)
RTI (10-27-76)
NAPCA45
EPA47
T
max
°C
31.7
29.4
9.4
33.9
22.2
Ethane
0.32
0.2e
0.08
Acetylene
0.45
0.5e
0.13
Propane
0.65
0.21°
0.028
0.6e
0.11
•,
Blank0
0-02
0.002d
0.06e
0.04
Unless noted otherwise, initial conditions 4.0 ppm(v/v) HC and 0.2 ppm NO
H
Unless noted otherwise, a blank is clean air plus 0.2 ppm NO .
c x
Average maximum ozone concentration from six chambers; see text.
Clean air plus 0-06 ppm NO .
4&
Initial conditions 6.0 ppm(v/v) HC and 0.2 ppm NO .
86
-------
variables are expected to be the major determinants of chemical behavior for
any given set of initial reactant conditions. Performing the matched experi-
ments on the same day, therefore, eliminates the possibility that differ-
ences in these variables would influence the results.
Although the chambers at both facilities are located outdoors and are
fabricated from transparent FEP Teflon film to enable natural sunlight and
ambient temperatures to govern the photochemical reactions, they are quite
different in appearance, design, and manner of operation. Figure 23 permits
comparison of the design and size of the two chamber types. RTI operates
four side-by-side, 27-cubic-meter cylindrical chambers. The UNC gas-phase
dual chamber is a much larger A-frame structure. Table 18 compares the
physical and performance characteristics of the chambers. Major differences
are: location (32 kilometers apart); shape; volume; surface-to-volume
ratio; method of pretreatment of background air; and sample handling.
In these experiments, two types of hydrocarbons were used: an alkane
(propane) which has a low reactivity and an olefin (propene) which has a
high reactivity and which has frequently been used in smog chamber testing.
Any substantial "wall effect" differences would be expected to be more
evident in the propane experiment than in the propene experiment.
Cross Calibration Comparisons
To insure that a valid comparison could be made of the results obtained
at both facilities, a detailed cross-comparison study was undertaken. In
addition to investigating the hydrocarbon and NO standards that were used to
calibrate the instruments, cross-measurements of initial hydrocarbon injec-
tions were made at both facilities. Based on measurements of three UNC- and
five RTI-certified NO calibration standards, a common calibration was achieved
to which all NO measurements were referenced. As the N02 and 0^ measurements
were calibrated via gas phase titration, the N02 and 03 results could also
be validly compared. All hydrocarbon calibrations were referenced to an
NBS-certified propane cylinder.* Immediately following the predawn injection
*It should be noted that while all the hydrocarbon concentrations
reported in this subsection are referenced to a common calibration standard,
the concentration reported in Appendix A are referenced to an RTI calibration
standard which yields 8 percent lower values for propane and 7 percent lower
values for propylene.
87
-------
00
oo
Figure 23. Comparison of the Designs and Sizes of the RTI (Foreground) and UNC (Background) Outdoor
Smog Chambers. The Shaded Area is Hanging Teflon Film Separating the Two UNC Chambers.
-------
oo
TABLE 18. COMPARISON OF PHYSICAL AND CHEMICAL CHARACTERISTICS OF RESEARCH TRIANGLE INSTITUTE
AND UNIVERSITY OF NORTH CAROLINA OUTDOOR SMOG CHAMBERS (a)
Location
Shape
Dimensions, Meters (feet)
3 3
Volume, m (ft )
2 2
Surface Area, m (ft )
S/V, m"1 (ft'1)
Background Air Treatment
Sampling Handling
0- Half-Life, Hours
3
(b)
Background Air 0., Formation,
NO Oxidation Relative to
Thermal Rate (Dark)
Research Triangle Institute
Research Park
4 Adjacent Cylinders
3.1 Dia x 3.7 II
(10 Dia x 12 H)
27 (1000)
51 (580)
1.9 (0.58)
8-12 Hours Recirculation thru
0.5% Pd Catalyst (260°C) +
Purafil
5 mm ID Teflon. Lines
<10 sec Residence Time
10-25
17-40
0.03-0.17
0.7-1.6
University of North Carolina
Rural Area
A-Frame Divided into Two
Compartments
9.1 W x 6.1 L x 6.1 H(e)
(30 W x 20 L x 20 H)
153 (5500)1
204 (2200)
1.3 (0.40)
(e)
(e)
15-20 Turnovers with Rural Air,
Start with 4-5 AM Rural Air
38 mm Glass Manifold
8 sec Residence Time
Excess Air Returned
18
48-70
0.02 - 0.14
1.5 to 4.8
(a)
(b)
(c)
(d)
(e)
Factors in common: material in both is 5 mil Teflon film and aluminum strips; both types use
natural sunlight and ambient temperature (0 to 5°C rise above ambient in a day); both have
reflective floors (aluminum under Teflon) to compensate for light transmission losses; both are
stirred by fans.
Function of temperature and season when measurement was made.
Function of temperature when measurement was made.
Function of temperature, season and background air when measurement was made.
Data for each of the two chambers.
-------
of reactants into the six chambers, two samples of each chamber's contents
were collected in 25-liter Tedlar bags. One of these remained at the origi-
nating facility for hydrocarbon analysis while the second bag was transported
to the other facility for analysis. The exception to this procedure was
UNC's analysis of its chambers in real-time. This approach permitted two
independent measurements of the initial hydrocarbon concentrations in each
chamber.
Background Conditions
To conduct a complete comparison it was important to determine the
amount of preinjection chamber contaminants. This was particularly true in
the case of the UNC facility where the injections are made into chambers
containing rural ambient air. The preinjection measurements of 5 November
1976 were taken as representative of typical conditions (see Table 19).
These indicated that NO levels were below the minimum detectable concentra-
tion (1 ppb) in all six chambers. Five ppb of NO 'were seen in both UNC
chambers while the four RTI chambers ranged between 0 and 5 ppb. The back-
ground air at the UNC facility contained 15 ppbV of hydrocarbons in the C_
through C_ range. Post-cleanup hydrocarbon concentrations in the RTI chambers
were between 1 and 9 ppbV for the same group. In both cases, a significant
fraction of the hydrocarbons were ethane and/or ethene (eluted at one peak)
while the remainder was composed of traces of propane, propene, acetylene,
normal butane, iso-butane, normal pentane, and isopentane.
The Matched Experiment of 20-21 September 1976
The first matched experiment was conducted on 20-21 September 1976 and
involved a propane/NO chemical system. Target initial reactant concentra-
X
tions for all six chambers were 4.00 ppmV propane and 0.20 ppm NO (20%
A
N0_). The propane was injected from common commercially supplied liquid
propane tanks. Gas chromatographic analysis conducted at RTI revealed the
contents to be 85% propane, 13% normal and iso-butane, 1% ethane and ethylene,
and 1% methane on a volume basis. Nitric oxide and NO. were injected from
high concentration gas cylinders.
Ideally, the initial reactant concentrations should be equal in all six
chambers; measured initial concentrations shown in Table 20 indicate that
all six chambers closely matched the target concentrations. After correc-
90
-------
TABLE 19. MEASUREMENT OF PREINJECTION CHAMBER CONTENTS AND BACKGROUND AMBIENT AIR
(DATA COLLECTED BEFORE DAWN ON 5 NOVEMBER 1976)
Nitric Oxide (ppb)
Nitrogen Dioxide (ppb)
Ethane/Ethylene (ppbV)
Propane (ppbV)
Propylene (ppbV)
Acetylene (ppbV)
Isobutane (ppbV)
N-Butane (ppbV)
1-Butene (ppbV)
Trans-2-Butene (ppbV)
Isopentane (ppbV)
Cyclopentane (ppbV)
N-Pentane (ppbV)
Total Measured Hydrocarbons
(C2 thru C5) (ppbV)
UNC
Background
Air
—
—
6
2
1
2
1
1
0
0
1
0
1
-15
Preinjection
UNC UNC RTI
Red Blue #1
000
550
_^__ ~\
— — 0
— — 0
/\
— — o
— — 0
— — 0
f\
f\
— — 0
— — 0
— — ~1
Chamber Contents
RTI
#2
0
5
5
1
0
1
0
1
0
0
1
0
0
~9
RTI
#3
0
4
5
1
0
1
0
1
0
0
1
0
0
~9
RTI
#4
0
0
4
1
0
1
0
0
0
0
1
0
0
~7
-------
TABLE 20. SUMMARY OF RESULTS FOR MATCHED EXPERIMENTS ON 20 AND 21 SEPTEMBER 1976
N>
Target Injections
Ambient Temperatures
(a)
Initial Propane Concentration (ppmV)
- Analysis by UNC
- Analysis by RTI
Initial NO Concentration (ppm)
Initial N(>2 Concentration (ppm)(c)
Initial HC/NO Ratio (ppmV/ppm)(d)
Vv
Time of NO/N02 Crossover (EOT)
Time of Maximum NC>2 Concentration (EOT)
Maximum NC>2 Concentration (ppm)'c'
0-j Concentrations (ppm)'c'
- First Day Maximum
- Overnight Minimum
- Second Day Maximum
- Net Second Day Production
Propane
N02
4.00 ppmV
0.04 ppm
NO 0.16 |>pm
1IC/NO Ral. io 20 |>pinV/p|>m
First Day Near Dawn (0800 EDT) 16° First Day Maximum (1400 KI)T) -!U° C
Second Day Near Dawn (0800 EDT) 19°C Second Day Maximum (1700 HUT) 27° C
RTI #1
RTI #2
RTI #3
RTI
UNC Rfd
UNC Blue
J.73
3.81
0.170
0.040
18.1
1004 (e)
1308
-------
tion for differences in the hydrocarbon calibration standards, the mean of
the six RTI propane analyses was 4 percent higher than the mean of the UNC
propane analyses. Based on the UNC analyses the UNC propane injections were
an average of 13 percent higher than the RTI propane injections. This
average increased to 17 percent for the RTI analyses. All NO injections
were within 6 percent of each other. There was, however, a greater relative
variation in the N02 injections. The initial NO- concentrations at the UNC
facility were about 13 ppb higher than those at the RTI facility.
Two other variables that have major influences on the reactions that
occur in the chambers are the diurnal profiles of light and ambient tempera-
ture. The total solar radiation profiles measured at each facility for 20
September 1976 are shown in Figure 24. Both profiles indicate that the day
was partially cloudy with a fairly clear morning. The integrated light
intensity from sunrise (approximately 0700 EDT) to near the times of the
NO-NO crossovers (1000 EDT) was 49.18 cal-cm at the RTI facility and
-2
46.46 cal-cm at the UNC site. This indicates that the two facilities were
exposed to very similar light conditions. The total solar radiation pro-
files for the second day of the experiment (not presented here) show a
similar pattern of "choppy" sunshine although that sky was more overcast in
the morning. It is assumed that the temperature profiles at the two facili-
ties did not differ greatly from that measured at the nearby Raleigh-Durham
Airport. The dawn and the daily maximum temperatures for both days were
somewhat higher than those normally occurring in central North Carolina at
this time of the year.
After the predawn reactant injections, a complex series of photochemical
reactions began within the six chambers which ultimately led to smog forma-
tion. One indicator of the rate of the photochemical process is the time
from dawn until the concentrations of NO and N02 are equal (crossover time).
Modeling results suggest that this measure is relatively sensitive to any
heterogeneous component of the reactions. If wall-related reactions differ
in two smog chambers, these differences should be more strongly manifested
for low reactivity chemical systems than for chemical systems of higher
reactivity. The crossover times for the six chambers cluster rather closely
around an average value of 184 minutes. RTI Chamber No. 2, however, achieved
crossover in 88 percent of this time. This suggests that there may have
93
-------
been surface-related processes occurring in this chamber. The two UNC
chambers had crossover times about 8 percent longer than the average value,
which indicates that surface-to-volume ratio still may be an influencing
factor in large volume chambers for low reactivity chemical systems. Never-
theless, the close agreement of the six chambers for this variable points
toward similar chemical behavior.
The ordering of the six chambers with respect to crossover times was
roughly duplicated for the times to maximum N0? concentration. The relative
spread of the times about the average time of 348 minutes was approximately
the same as the relative spread about the average crossover time. It should
be noted, however, that the sampling frequency of the RTI measurements was
once per hour; whereas, the UNC measurements were collected at 8-minute
intervals for each of the two chambers. Consequently, the RTI-measured
times to maximum N0_ concentration may be up to 30 minutes away from the
actual times.
The two UNC chambers showed excellent agreement for their maximum NO
concentrations as did the four RTI chambers. Nitrogen balances (NO plus
NO-) at this stage of the reaction process are approximately 85 percent for
RTI and 74 percent for UNC. Ratios of the maximum NO,, concentration to the
initial NO concentration are about 74 percent in the RTI chambers and about
X
65 percent at the UNC facility. This latter value is somewhat larger than
the average value of 58 percent found by the UNC group for lower initial
49
concentrations of urban hydrocarbon mixture.
The most interesting result of the propane experiment is to be found in
the first-day maximum 0., concentrations. The agreement among all six chambers
for this parameter and the time to maximum 0, concentration (see Figures 25
through 28) is impressive. While five of the six chambers closely cluster
about an average value of 0.202 ppm (standard deviation: 0.016 ppm), however,
RTI No. 2 is 28 percent higher. This is significant in the light of the
excellent agreement in the other five chambers. This behavior is consistent
with the earlier divergence of this chamber and also suggests the existence
of unusual suface-related or chamber history-related processes in RTI No. 2.
Since the maximum 0,. concentrations in the other five chambers are in agree-
ment, it is believed that wall-related effects play a much lesser role for
these chambers in determining the magnitude of the maximum. The average
94
-------
VO
en
2.00
1.80
1.60
1.40
1.20
1.00
0.80
0.60
0.40
0.20
0.00
I r I ' I ' I ' I '
SEPTEMBER 20. 1976 - 1.80
- 1.60
8
9 10 11 12 13
HOURS. EOT
15 16 17 18 19
1.20
1.00
0.80
0.60
0.40
0.20
0.00
Figure 24. Total Solar Radiation Profiles for UNC (-
20, 1976.
-) and RTI (- - ) on September
-------
.300
.270
ex
ex
ON
o
.210
.180
.150
.120
.090
.060
.030
.000
I ' I ' I ' I ' I ' I ' I ' I ' I '
SEPT. 20. 1976, RTI CHflMBERS 1 & 2
*
- v
/ x
'' 0,
1 NO
8
9 10 11 12 13
TIME. EOT
.300
.270
15 16 17 18 19
.210
.180
.150
.120
.090
.060
.030
.000
Figure 25, Nitric Oxide, Nitrogen Dioxide and Ozone Concentration Profiles for RTI Chamber 1
(_ __) ana RTJ Chamber 2 (- ^ - - -) on September 20? 1976,
-------
VD
.300
.270
6 .210
5 .180
.150
.120
§ .090
.060
.030
.000
r~rT"r I ' I ' I ' I ' I ' I ' I ' I ' 1 -300
SEPT. 20. 1976. RTI CHflMBERS 3 & * - .270
°
8
10 11 12 13
TIME. EOT
15 16 17 18 19
.2*0
.210
.180
.150
.120
.090
.060
.030
.000
Figure 26, Nitric Oxide, Nitrogen Dioxide and Ozone Concentration Profiles for RTI Chamber
3 '(__ .—) ^nd RTI Chamber 4 ( ) on September 20, 1976,
-------
VO
oo
1 ' ' I
SEPTEMBER 20. 1976
.000
- .030
12 13 . H 15 16 17 18 19
TIME. EOT
.000
Figure 27. Nitjric Oxide, Nitrogen Dioxide and Ozone Concentration Profiles for UNC Red
Chamber (— —-) and UNC Blue Chamber C- ^ ' 1 on September 20, 1976.
-------
vo
VD
^ .300
Q.
O
H .200
o:
UJ
o
z
o
o
UJ
z
o
M
O
IOO —
8
SEPTEMBER 20,1976
10 12 14
TIME, EOT
16
18
Figure 28. Ozone Concentration Profiles for the Two UNC Chambers (-
Chambers ( ) on September 20, 1976.
-) and the Four RTI
-------
value of 0.202 ppm is substantially larger than the 0.11 ppm value which was
measured in EPA's 400-liter Pyrex chamber for similar initial reactant
concentrations. This difference could be the result of different tempera-
tures or light intensities, the presence of additional hydrocarbons in this
experiment (see above) or different surface effects as judged by NO behavior
A
in the EPA and the outdoor chambers.
One of the several possible causes of high oxidant concentrations in
rural areas is the long-range transport of "partially spent" photochemical
systems from urban areas. In such systems, low reactivity hydrocarbons
are likely to play a significant role in second- and third-day oxidant for-
mation. To investigate this possibility, the contents of the chambers were
irradiated for a second day on 21 September 1976. As the differences between
the overnight minimum and the second-day maximum 0,, concentrations indicate,
substantial 0,, formation occurred on this second day. The RTI chambers
formed between 95 and 126 percent of the previous day's maximum 0,, concentra-
tion while values of 85 and 95 percent were seen in the UNC chambers.
The Matched Experiment of 5 November 1976
The matched experiment of 5 November 1976 differed in several respects
from the previous matched experiment: a higher reactivity hydrocarbon
(propene) was used; two different sets of initial reactant concentrations
were injected instead of one set for all the chambers and the experiment
lasted only 1 day. Two chambers (RTI No. 1 and UNC Red) had target propene
injections of 0.50 ppmV. Table 21 indicates that UNC Red came rather close
to this target while RTI No. 1 overshot by about 18 percent. The other two
chambers which were operated in this matched experiment (RTI No. 2 and UNC
Blue) were intended to have initial propene concentrations of 1.33 ppmV.
The UNC injection was close to this value. The RTI injection, however, was
13 percent low. Averaged across all four chambers, the RTI hydrocarbon
analyses were 3 percent higher than those of UNC. Both facilities injected
roughly 10 percent more NO and N0_ than the target initial concentrations of
0.40 and 0.10 ppm.
Ambient temperatures for this experiment were rather low (5° C below
normal for this date). Based on the results of previous UNC research on the
effect of temperature on oxidant formation, the maximum ozone concentration
should have been considerably less than that which would be expected on a
100
-------
TABLE 21. SUMMARY OF RESULTS FOR MATCHED EXPERIMENT ON 5 NOVEMBER 1976
Target Injections
Ambient Temperatures
(a)
Initial Propylene Concentration (ppmV)
- Analysis by UNC
- Analysis by RTI , ..
Initial NO Concentration (ppm) ^
Initial N02 Concentration (ppm)^c'
Initial HC/NO Ratio (ppmV/ppm)
-------
warm day. On the other hand, the sky was cloudless throughout the major
portion of the day. This would produce the largest amount of oxidants
possible for this time of year given the observed temperatures. As illus-
trated in Figure 29, both facilities were exposed to nearly perfect light
profiles from about 0745 until 1430 EDT.
In the high concentration chambers, it is seen that NO and NCL crossed
over in UNC Blue in 135 minutes while RTI No. 2 took 20 minutes longer.
This difference in crossover times can be most easily explained by observing
that the initial hydrocarbon-NO ratio for the RTI chamber was 12 percent
X
lower than the value for the UNC chamber. As the difference appears to be
associated with the initial reactant concentrations, it is not necessary to
invoke wall-related effects. Assuming that this is the case, it would
appear that such effects are not as influential in this experiment as in the
previous one. This could be because the chamber wall history was different
for this experiment, because the temperature was lower, or because the high
reactivity propene/NO chemical system may not be as sensitive to chamber
X
contamination as the lower reactivity propane/NO chemical system. There is
JEB>
insufficient evidence to draw any firm conclusions as to the true cause.
The relationship between the two chambers' crossover times was maintained
for the times of maximum NO- concentration (the earlier caution concerning
this parameter is again stressed). Overall, the two chambers showed very
good agreement as is shown in Figure 30.
The times to N0-N0? crossover were approximately equal in the low
concentration propene chambers. RTI No. 1 was 17 minutes faster than UNC's
crossover time of 227 minutes. This difference is somewhat less than what
would be expected considering that the initial hydrocarbon-NO ratio of the
X
RTI chamber was 19 percent larger. Despite this close agreement in the
early stages of the photochemical reactions, Figure 31 shows that the two
chambers began to widely diverge shortly after the crossovers. RTI No. 1
had a N02 peak which occurred at least an hour before the N0~ peak in UNC
Blue. Afterwards, the decline of NO- was more rapid in the RTI chamber than
in the UNC chamber. This behavior is consistent with the higher hydrocar-
bon-NO ratio, it is not obvious, however, why this was not as strongly
^v
manifest in the times to NO-NO^ crossover. It should be noted that there
was nearly as large a relative difference in initial hydrocarbon-NO ratios
A:
102
-------
o
to
2.00
1.80
1.60
1.40
1.20
1.00
* 0.80
0.60
Of
0.1-0
§ 0.20
0.00
I ' I ' I ' I
I ' I ' I ' I ' I ' I ' I2-00
NOVEMBER 5. 1976 - 1.80
1.60
8 9 10 11 12 13 14 15 16 17 18 19
HOURS, EOT
1.20
1.00
0.80
0.60
0.40
0.20
0.00
Figure 29. Total Solar Radiation Profiles for UNC
1976,
-) and RTI (- , ) on November 5,
-------
.750
.675
.600
.525
I .450
o.
; .375
$ .300
! .225
o
.150
.075
.000
. ' I ' I ' I ' I
NO
I ' I ' I ' ! ' I ' I ' I '
NOVEMBER 5. 1976
7 8 9 10 11 12 13 14 15 16 17 18
HOURS. EOT
19
.750
.675
.600
.525
.450
.375
.300
.225
.150
.075
.000
30, Nitric Oxide, Nitrogen Dioxide and Ozone Concentration Profiles for UNC Blue Chamber
(•— ) and RTI Chamber 2 ( ) on November 5, 1976.
-------
.750
.750
1 I • I ' I ' I • I ' I ' I ' I ' I ' I ' ' '
.675 I- NOVEMBER 5, 1976 -| .675
o
8 9 10 11 12 13 14 15 16 17 18 19
HOURS, EOT
.000
Figure 31, Nitric Oxide, Nitrogen Dioxide and Ozone Concentration profiles for UNC Red Chamber
(;"•••) and R,TI Chamber 2 (r *- - -) on November 5, 1976.
-------
for the high concentration propene chambers and yet no large difference was
observed for the N0_ concentration profiles. The implication is that the
effect of variations in the ratio become more pronounced for lower initial
hydrocarbon concentrations.
Nitrogen balances at the maximum NO concentration agreed well within
the sets of matched chambers. In the high concentration propene chambers,
RTI No. 2 showed a 94 percent nitrogen balance while 91 percent of the
initial NO was still present in UNC Blue. For the two chambers having
X
lower initial propene concentrations, the nitrogen balances were 85 percent
for RTI No. 1 and 78 percent for UNC Red. It appears that the nitrogen
balance improves with increasing initial hydrocarbon concentration when the
49
initial NO concentration is held constant as was found by the UNC group.
x»
The major NO,, loss up until the N0? maximum is thought to be due to the
formation of HNO by the reaction of N02 with hydroxyl radicals. This is a
thermal reaction and longer times to the NO- maximum would tend to increase
the NO. loss by this pathway (assuming essentially the same average hydroxyl
concentrations in the chambers).
In the high concentration propene chambers, there was very close agree-
ment for the maximum 0» concentrations. The values measured in RTI No. 2
and UNC Blue were within 2 percent of each other. This result provides
additional support to the belief that the chambers at the two facilities
exhibit similar chemical behavior. Whatever had produced the deviant behav-
ior of RTI No. 2 in the previous matched experiment was apparently not
evident in this case (see above).
In contrast to the results for the high concentration propene chambers,
the maximum 0, concentrations in the lower concentration chambers showed
•3
significant disagreement. RTI No. 1 produced nearly twice the ozone formed
in UNC Red. An explanation for this behavior can be found in the initial
reactant concentrations. The RTI chamber had an initial hydrocarbon-NO
ratio that was 20 percent larger than that of UNC Red. Since the maximum
ozone surface is rather steep in low concentration regions, a small differ-
ence in the initial reactant concentrations can result in a large difference
in the maximum ozone concentration. Since the surface flattens out as one
moves along a constant [NO ] line toward higher hydrocarbon concentrations,
X
the observed differences in the initial concentrations of the high concentra-
106
-------
tion propene chambers would not produce as large a disagreement in the
maximum ozone concentrations as the same relative difference would produce
at the lower initial concentrations. However, wall-related effects cannot
be ruled out as a possible cause of the disagreement.
Discussion
The results of an initial set of matched experiments comparing the
chemical behavior of two outdoor smog chamber facilities have been presented.
While it is not prudent to make definitive statements based on such a small
number of experiments, the data that have been presented here support the
belief that the four RTI and the two UNC outdoor smog chambers display very
similar chemical behavior. It is likely that this is the result of the
similarity of their wall materials and of the near-identical light intensity
and temperature profiles that prevailed during the matched experiments. The
excellent overall agreement of the results should increase confidence in the
reliability of data obtained from outdoor smog chambers.
107
-------
108
-------
REFERENCES
1. Holzworth, G. C., 1972. Mixing Heights, Wind Speeds and Potential for
Urban Air Pollution Throughout the Contiguous United States. Environ-
mental Protection Agnency Publication No. AP-101.
2. Schere, K. L. and Demerjian, K. L. , 1978. A Photochemical Box Model
for Urban Air Quality Simulation. Proceedings, 4th Joint Conference on
Sensing of Environmental Pollutants. American Chemical Society, Washing-
ton, D.C., pp. 427-433.
3. Jeffries, H. E., Kamens, R. , Fox, D. L. , and Dimitriades, B., 1977.
"Outdoor Smog Chamber Studies: Effect of Diurnal Light, Dilution and
Continuous Emission on Oxidant Precursor Relationships" in International
Conference of Photochemical Oxidant Pollution and its Control, Proceed-
ings. Environmental Protection Agency Publication No. EPA-600/3-77-001,
pp. 891-902.
4. Dodge, M. C. , 1977. Effect of Selected Parameters on Predictions of a
Photochemical Model. Environmental Protection Agency Publication No.
EPA-600/3-77-048.
5. Office of Air Quality Planning and Standards, 1977. Uses, Limitations
and Technical Basis of Procedures for Quantifying Relationships Between
Photochemical Oxidants and Precursors. Environmental Protection Agency
Publication No. EPA-450/2-77-021a.
6. Office of Air Quality Planning and Standards, 1978. Procedures for
Quantifying Relationships Between Photochemical Oxidants and Precursors:
Supporting Documentation. Environmental Protection Agency Publication
No. EPA-450/2-77-021b.
7. Decker, C. E., Sickles, J. E., II, Bach, W. D., Vukovich, F. M. , and
Worth, J. J. B., 1978. Project Da Vinci II: Data Analysis and Interpre-
tation. Environmental Protection Agency Publication No. EPA-450/3-78-028
8. Personal Communication, Edwin L. Meyer, Jr., Office of Air Quality
Planning and Standards, Environmental Protection Agency, June 12, 1978.
9. Sickles, J. E., II, 1976. Ozone-Precursor Relationships of Nitrogen
Dioxide, Isopentane and Sunlight Under Selected Conditions. Doctoral
Dissertation. Department of Environmental Sciences and Engineering,
University of North Carolina, Chapel Hill, North Carolina.
10. Ripperton, L. A., Sickles, J. E. , II, and Eaton, W. C., 1976. Oxidant-
Precursor Relationships During Pollutant Transport Conditions. Environ-
mental Protection Agency Publication No. EPA-600/3-76-107.
109
-------
II. Khang, S. J., and Levenspiel, 0., 1976. The Mixing Rate Number for
Agitator-Stirred Tanks. Chemical Engineering, October 11, p. 141.
12. Jeffries, II., Fox, D. , and Kamens, R. , 1975. Outdoor Smog Chamber
Studies: Effect of Hydrocarbon Reduction on Nitrogen Dioxide. Environ-
mental Protection Agency Publication No. EPA-650/3-75-011.
13. Butcher, S. S. , and Ruff., R. E., 1971. Effect of Inlet Residence Time
on Analysis of Atmospheric Nitrogen Oxides and Ozone. Analytical
Chemistry, 43: No. 13, p. 1890.
14. Bufalini, J. J., Kopczynski, S. L., and Dodge, M. C., 1972. Contaminated
Smog Chambers in Air Pollution Research, Environmental Letters, 3_:
No. 2, p. 101.
15. Bufalini, J. J., Walter, T. A., and Bufalini, M., 1977. Contamination
Effects on Ozone Formation in Smog Chambers. Environmental Science
and Technology, 11: No. 13, p. 1181.
16. Jones, A. C., and Mindrup, R. F., Jr., 1976. Regional Air Pollution
Study: Gas Chromatograph Laboratory Operations. Environmental Protec-
tion Agency Publication No. EPA-600/4-76-040.
17. Pitts, J. N., Jr., Darnall, K. R., Winer, A. M., and McAfee, J. M.,
1977. Mechanisms of Photochemical Reactions in Urban Air: Volume II.
Chamber Studies. Environmental Protection Agency Publication No.
EPA-600/3-77-014b.
18. Dimitriades, B. , 1967. Methodology in Air Pollution Studies Using
Irradiation Chambers. Journal of Air Pollution Control Association,
17: No. 7, p. 460.
19. Scofield, F. , Levy, A., and Miller, S. E., 1969- I. Design and Valida-
tion of a Smog Chamber, National Paint, Varnish and Lacquer Association
Publication No. 797.
20. Grasley, M. H., Appel, B. R., Burstain, I. G., Laity, J. L., and
Richards, H. F., 1969. The Relationship of Smog Chamber Methodology
to Hydrocarbon Reactivity in Polluted Air. American Chemical Societyf
Division of Organic Coating and Plastics Chemistry, 29, p. 422.
21. Powers, T. R., 1977. Effect of Hydrocarbon Composition on Oxidant-
Hydrocarbon Relationships. Environmental Protection Agency Publication
No. EPA-600/3-77-109a.
22. Doyle, G. J., 1970. Design of a Facility (Smog Chamber) for Studying
Photochemical Reactions under Simulated Tropospheric Conditions.
Environmental Science and Technology, _4_: No. 11, p. 907.
23. O'Brien, R. J., 1974. Photostationary State in Photochemical Smog
Studies. Environmental Science and Technology, 8i_: No. 6, p. 579.
110
-------
24. Bufalini, J. J., and Altshuller, A. P., 1965. Kinetics of Vapor Phase
Hydrocarbon-Ozone Reactions. Canadian Journal of Chemistry 43
p. 2243. " L —'
25. Jaffe, R. J. , Smith, F. C., Jr., and Last, K. W., 1974. Study of
Factors Affecting Reactions in Environmental Chambers: Final Report
on Phase II. Environmental Protection Agency Publication No. EPA-650/
3-74-004a.
26. Heuss, J. M., February 10-12, 1975. Smog Chamber Simulation of the Los
Angeles Atmosphere. A paper presented at the Environmental Protection
Agency Scientific Seminar on Automotive Pollutants, Washington, B.C.
27. McNelis, D. N. , 1974. Aerosol Formation from Gas-Phase Reactions of
Ozone and Olefin in the Presence of Sulfur Dioxide. Environmental
Protection Agency Publication No. EPA-650/4-74-034.
28. Kuhlman, M. R. , 1974. The Ambient Aerosol Research Facility: Design
Criteria and Validation. Master's Thesis. Department of Environmental
Sciences and Engienering, University of North Carolina, Chapel Hill,
North Carolina.
29- Hampson, R. F., Jr., and Garvin, D. , Eds., 1975. Chemical Kinetics and
Photochemical Data for Modeling Atmospheric Chemistry. NBS Technical
Note 866.
30. Handbook of Chemistry and Physics, 1962. Forty-third Edition, Chemical
Rubber Publishing Company, Cleveland, Ohio.
31. U.S. Department of Health, Education, and Welfare, 1970. Air Quality
Criteria for Hydrocarbons. National Air Pollution Control Administra-
tion Publication No. AP-64.
32. Local Climatological Data: National Weather Service Forecast Office,
Raleigh-Durham Airport. U.S. Department of Commerce, National Climatic
Center, Ashville, North Carolina.
33. Federal Register, 1976. Measurement of Photochemical Oxidants in the
Atmosphere. 41: No. 195, p. 44049.
34. Intersociety Committee, Method 406, 1972. Methods of Air Sampling and
Analysis. American Public Health Association, Washington, D.C.
35. Winer, A. M. , Peters, J. W. , Smith, J. P., and Pitts, J. N., Jr., 1974.
Response of Commercial Chemiluminescent NO-N02 Analyzers to Other
Nitrogen-Containing Compounds. Environmental Science and Technology,
£: No. 13, p. 1118.
36. Intersociety Committee, Method 403, 1972. Methods of Air Sampling and
Analysis. American Public Health Association, Washington, D.C.
Ill
-------
37. Federal Register, 1971. Appendix E—Reference Method for Determination
of Hydrocarbons Corrected for Methane. 36; No. 84, p. 8198.
38. Harrison, J. W., Tiiamons, M. L. , Denyszyn, R. B., and Decker, C. E.,
1977. Evaluation of the EPA Reference Method for Measurement of Non-
methane Hydrocarbons. Environmental Protection Agency Publication
No. EPA 600/4-77-033.
39. Intersociety Committee, Method 110, 1972. Methods of Air Sampling and
Analysis. American Public Health Association, Washington, D.C.
40. Intersociety Committee, Method 111, 1972. Methods of Air Sampling and
Analysis. American Public Health Association, Washington, D.C.
41. Young, H. D., 1962. Statistical Treatment of Experimental Data.
McGraw-Hill Book Co., Inc., New York, N.Y., pp. 76-79.
42. Fox, D. L., Kamens, R., and Jeffries, H. E., 1975. Photochemical'Smog
Systems: Effect of Dilution on Ozone Formation. Science, 188: p. 1113.
43. Sickles, J. E., II, Ripperton, L. A., Eaton, W. C., and Wright, R. S.,
1978. Atmospheric Chemistry of Potential Emissions from Fuel Conversion
Facilities: A Smog Chamber Study. Environmental Protection Agency
Publication No. EPA-600/7-78-029.
44. Bufalini, J. J., Gay, B. W., and Kopczynski, S. L., 1971. Oxidation
of n-Butane by Photolysis of NC^- Environmental Science and Technology,
_5: No. 4, p. 336.
45. Altshuller, A. P. Kopczynski, S. L., Wilson, D., Lonneman, W. A., and
Sutterfield, F. D., 1969- Photochemical Reactivities of n-Butane and
Other Paraffinic Hydrocarbons. Journal of the Air Pollution Control
Association, 19: No. 10, p. 787.
46. Zafonte, L. and Bonamassa, F., 1977. Relative Photochemical Reactivity
of Propane and n-Butane. Environmental Science and Technology, 11:
No. 10, p. 1015.
47. Dimitriades, B. and Joshi, S. B., 1977. "Application of Reactivity
Criteria in Oxidant-Related Emission Control in the USA" in Inter-
national Conference of Photochemical Oxidant Pollution and Its Control:
Proceedings. Environmental Protection Agency Publication No. EPA-600/
3-77-001, pp. 705-710.
48. Calvert, J. G. and Jeffries, H. E., 1977. International Conference on
Oxidants, 1976. Analysis of Evidence and Viewpoints. Part II. The
Issue of Reactivity. Environmental Protection Agency Publication No.
EPA-600/3-77-114.
49- Jeffries, H. E., Fox, D. L., and Kamens, R., "Photochemical Conversion
of NO to NO™ by Hydrocarbons in an Outdoor Chamber", Journal of the
Air Pollution Control Association, 26 (5): 480, 1976.
112
-------
50. Jeffries, H. E., Sickles, J. E., II and Ripperton, L. A., "Ozone
Transport Phenomena: Observed and Simulated." Presented at the
69th Annual Meeting of the Air Pollution Control Association,
Portland, Oregon, June 27-July 1, 1976.
51. Kamens, R. M. and Jeffries, H. E. , "Progress Report on Winter Oxidant
Study in UNC Outdoor Chamber," EPA Grant 800916, Univeristy of
North Carolina, Department of Environmental Science and Engineering,
May 4, 1976.
113
-------
TECHNICAL REPORT DATA
{Please read Instructions on the reverse before completing)
- =ORT N
3-79-078a
3. RECIPIENT'S ACCESSION-NO.
4. T,TLE AND SUBTITLEOXIDANT-PRECURSOR RELATIONSHIPS
UNDER POLLUTANT TRANSPORT CONDITIONS
Outdoor Smog Chamber Study
Volume 1 __
5. REPORT DATE
August 1979
6. PERFORMING ORGANIZATION CODE
/. AO i no
8. PERFORMING ORGANIZATION REPORT NO.
J.E. Sickles, II, L.A. Ripperton, W.C. Eaton,
and R.S. Wright
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Research Triangle Institute
Research Triangle Park
North Carolina 27709
10. PROGRAM ELEMENT NO.
1
11. CONTRACT/GRANT NO.
68-02-2207
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory-RTF, NC
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final 6/75 - 6/78
14. SPONSORING AGENCY CODE
EPA/600/09
15. SUPPLEMENTARY NOTES
Volume 2. Appendixes
16. ABSTRACT
Multiple-day experiments were conducted in outdoor smog chambers to investi-
gate the influence of simulated transport on ozone generation by various com-
binations of a surrogate urban hydrocarbon mixture and nitrogen oxides. The
simulation of transport was accomplished by progressively diluting the contents
of the chambers with purified air.
First day ozone maximum concentrations were reduced under dilution conditions an
were sensitive to both dilution rate and time at which dilution was initiated.
Second and third-day ozone maxima were reduced at increasing dilution rates, but
the reduction was less than proportional to the extent of dilution. The ozone-
generative potential of an aged photochemical system generally exceeded 0.08 ppm.
Additional experiments were conducted to examine the ozone-generative po-
tential of low reactivity hydrocarbons, to provide data for testing and validation
of a computer-based photochemical simulation model, and to compare the behavior of
two types of outdoor smog chambers.
Volume 1 contains all textual material. Volume 2 contains all the smog
chamber analytical data for hydrocarbons, NO , 0 , CO, CH 0, and condensation
nuclei, as well as dilution and meterologicaJ data.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
COSATI Field/Group
*
*
Air pollution
Ozone
Nitrogen oxides
Test chambers
Solar radiation
Photochemical reactions
Chemical analysis
13B
07B
07C
03B
07E
07D
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (THISReport/
UNCLASSIFIED
21. NO. OF PAGES
126
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
Form 2220-1 (9-73)
114
------- |