EPA-650/3-75-011
June 1975 Ecological Research Series
OUTDOOR
SMOG CHAMBER STUDIES
EFFECT OF HYDROCARBON REDUCTION ON NITROGEN DIOXIDE
U.S. Environmental Protection Agency
Office of Research and Development
Washington, D. C. 20460
-------
EPA-650/3-75-011
OUTDOOR
SMOG CHAMBER STUDIES
EFFECT OF HYDROCARBON REDUCTION ON NITROGEN DIOXIDE
by
Harvey Jeffries, Donald Fox, and Richard Kamens
University of North Carolina
Chapel Hill, North Carolina 27514
Grant No. 800916
ROAP No. 21AZJ-008
Program Element No. 1AA008
EPA Project Officer: Basil Dimitriades
Chemistry and Physics Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, D. C. 20460
June 1975
-------
EPA REVIEW NOTICE
Tliis report has been reviewed by the National Environmental Research
Center - Research Triangle Park, Office of Research and Development,
EPA, and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environ-
mental Protection Agency, have been grouped into series. These broad
categories were established to facilitate further development and applica-
tion of environmental technology. Elimination of traditional grouping was
consciously planned to foster technology transfer and maximum interface
in related fields. These 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
9. MISCELLANEOUS
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to
develop and demonstrate instrumentation, equipment and methodology
to repair or prevent environmental degradation from point and non-
point sources of pollution. This work provides the new or improved
technology required for the control and treatment of pollution sources
to meet environmental quality standards.
This document is available to the public for sale through the National
Technical Information Service, Springfield, Virginia 22161.
Publication No. EPA-650/3-75-011
11
-------
ABSTRACT
o o
A 312 m (11,018 ft ) Teflon film outdoor smog chamber was constructed
in rural North Carolina. The chamber was operated with natural cond-
itions of solar radiation, temperature and relative humidity which
existed at the time of a run. Ninety-two 12-hour runs using propylene
and oxides of nitrogen were conducted to assess the performance of the
system. A photochemical model, in which only the light intensity
magnitude and pattern and the rate of heterogeneous surface reactions
were changed, was used to compare the outdoor results with those of
three indoor chambers. Good agreement was found in all cases. One-
hundred-thirty 12-hour runs were conducted using a simulated urban
hydrocarbon mix and oxides of nitrogen. Reduction of the hydrocarbon
concentration resulted in reductions of nitrogen dioxide (NO^) maximum
concentration and, for large reductions, the daytime N02 dosage. Other
factors investigated included N02 to NO ratio at constant NO , effect
of slow dilution, and results of extended 24-hour and 36-hour runs.
-------
CONTENTS
Page
Abstract iii
Table of Contents v
List of Figures vi
List of Tables xi
Acknowledgments xiii
Sections
I Summary and Conclusions 1
II Recommendations 5
III Introduction 7
IV Design and Construction of Smog Chamber Facility 14
V Experimental Methods and Procedures 25
VI Results and Discussion, Part I: Facility
Performance 53
VII Results and Discussion, Part II: Effects of
Hydrocarbon Reduction on Nitrogen Dioxide
Concentrations 121
VIII Other Results and Observations 165
IX References 177
X Glossary 182
XI Appendices 186
-------
LIST OF FIGURES
No. Page
1 Dependence of nitrogen dioxide maximum concentra- 9
tion on initial nitrogen oxides and nonmethane
hydrocarbon in the Bureau of Mines study by
Dimitriades
2 Idealized representation of NO, NOp. 03,aerosol 10
and PAN profiles for various initial conditions
3 The University of North Carolina outdoor smog chamber 18
4 Schematic of UNC outdoor smog chamber 19
5 Orientation of UNC outdoor smog chamber with respect 21
to seasonal sunrise and sunset positions
6 Solar altitude and zenith angle at noon at the 21
UNC outdoor smog chamber for each month
7 Flowchart of on-site data treatment procedures 30
8 Flowchart of preliminary off-site data treatment 31
procedures
9 Flowchart of final off-site data treatment procedures 32
10 CAP 1-6 definition of N02 formation rate applied to 41
typical indoor and outdoor smog chamber profiles
11 Average daily solar radiation on a horizontal surface 49
in the USA for January
12 Average daily solar radiation on a horizontal surface 49
in the USA for July
13 Theoretical solar radiation on a surface normal to 51
the solar beam as a function of month
14 Theoretical solar radiation on a horizontal surface 51
as a function of month
15 Specific photolysis rate for nitrogen dioxide and 58
incident total solar radiation and ultraviolet
radiation as a function of time for September 18, 1974
16 Specific photolysis rate for nitrogen dioxide and 59
incident total solar radiation and ultraviolet
radiation as a function of time for September 19, 1974
VI
-------
LIST OF FIGURES (CONTINUED)
No. Page
17 Selected air temperature profiles in outdoor 61
smog chamber
18 Selected dewpoint profiles in outdoor smog chamber 61
19 Comparison of three different N02 measurement methods 64
in a propylene/NO system in the outdoor smog chamber
A
20 Calibration and correction technique for environmental 67
chromatograph
21 Sample gas chromatogram of hydrocarbon mix from 73
chamber air immediately after injection
22 Example of NO oxidation run in background air 76
23 Example of reproducibility between two chamber 80
halves for propylene
24 Example of reproducibility between two chamber 80
halves for urban hydrocarbon mix
25 Example of reproducibility between two chamber 81
halves for urban hydrocarbon mix
26 Effect of hydrocarbon reduction and solar radiation 82
27 Effect of hydrocarbon and solar radiation 82
28 Comparison of total solar radiation for a totally 83
clear day and a partly cloudy day
29 Comparison of mix (olefins and paraffins) and 85
mix with toluene
30 Comparison of mix (olefins and paraffins) and 86
mix with toluene
31 Comparison of (mix with toluene) with (mix, 87
toluene and m-xylene)
32 Comparison of mix and mix with m-xylene 87
as nonmethane hydrocarbon
33 Comparison of mix and propylene as nonmethane 89
hydrocarbon
vii
-------
LIST OF FIGURES (CONTINUED)
No. Page
34 Comparison of mix and propylene as nonmethane 89
hydrocarbon
35 Variable light intensity model of May 25, 1974 94
36 Dual matched propylene run in UNC outdoor smog chamber 94
37 Actual solar radiation for May 25, 1974 95
38 Variable light intensity model of May 7, 1974 96
39 Dual matched propylene run in UNC outdoor chamber 96
40 Actual solar radiation for May 7, 1974 97
41 Constant light intensity model 99
42 Ramping light intensity model 99
43 Fast HN02 reaction model 100
44 Fast HN02 reaction model 100
45 Actual profiles for EPA-325 104
46 UNC model of EPA-325 105
47 Actual profiles of run EC-17 in SAPRC chamber at 106
Riverside, California
48 UNC model of EC-17 107
49 Actual profiles of run EC-17 in SAPRC chamber at 108
Riverside, California
50 UNC model of run EC-17 109
51 Actual profiles of EC-18 in SAPRC chamber at 110
Riverside, California
52 UNC model of EC-18 111
53 Actual profiles of EC-18 in SAPRC chamber at 112
Riverside, California.
54 UNC model of run EC-18 113
vm
-------
LIST OF FIGURES(CONTINUED)
No. Page
55 Actual profiles of a Lockheed run 116
56 UNC model of a Lockheed run 117
57 Actual profiles of Bureau of Mines run 119
58 Mix run for comparison with BOM run 120
59 Mean nitrogen dioxide concentration as a function 125
of mean initial oxides of nitrogen for urban
hydrocarbon mix in UNC outdoor smog chamber
60 Nitrogen dioxide maximum concentration as a function 126
of initial oxides of nitrogen for individual runs
for urban hydrocarbon mix in UNC outdoor chamber
61 Effect of single light pulse on nitrogen oxides 133
conversion rate in a system initially containing
NO /propylene
A
62 Example of rapid changes in solar radiation 133
accelerating occurrence of nitrogen dioxide maximum
63 Nitrogen dioxide profiles at constant initial NO 136
of 0.51 ppm
64 Nitrogen dioxide profiles at constant initial NO 137
of 0.36 ppm
65 Nitrogen dioxide profiles at constant initial NO 138
of 0.24 ppm
66 Nitrogen dioxide profiles for October 5 and 6, 1974 139
67 Total solar radiation for a clear summer and a clear 140
fall day
68 Profiles for dual run 141
69 Profiles for dual run 142
70 Mean nitrogen dioxide 10-hour average concentra- 144
tion as a function of mean initial oxides of nitrogen
for urban hydrocarbon mix in UNC outdoor chamber
71 Extended run of June 16-17, 1974 154
72 Total solar radiation for June 16, 1974 155
ix
-------
LIST OF FIGURES(CONTINUED)
No. Page
73 Extended dual run of June 12-13, 1974 156
74 Total solar radiation for June 12, 1974 157
75 Extended dual run of September 20-21, 1974 158
76 Total solar radiation for September 20, 1974 159
77 Differential percentage N0? run under clear sky 166
78 Differential percentage N0? run under clear sky 166
79 Differential percentage NOp run under erratic solar 167
radiation
80 Differential percentage NOp run under erratic solar 167
radiation
81 Example of dilution in a reactive system 169
82 Example of dilution in a less reactive system 169
83 Extended 36-hour run of June 12-13, 1974 172
84 Total solar radiation for June 12-13, 1974 173
85 Effect of air temperature at maximum rate on 175
maximum nitric oxide disappearance for clear
days for November 1973 and NO /mix in UNC outdoor
smog chamber
86 Formaldehyde photolysis chemistry 196
87 Flow diagram of one liter quartz globe CSTR 205
for nitrogen dioxide photolysis rate measurements
-------
LIST OF TABLES
No. Page
1 Instrument Characteristics 26
2 Calibration Sources for Gases 27
3 Simulated Urban Hydrocarbon Mixture 44
4 Selected Climatological Data for Piedmont
North Carolina 46
5 Number of Days in which Percentage Maximum
Possible Sunshine was Greater than or
Equal to a Given value. Tuesdays and
Thursdays in 1969-1971 (313 Days) bymonth 48
6 Actual Physical and Chemical Characteristics
of the Chamber 54
7 Specific Photolysis Rate for Nitrogen Dioxide
in Outdoor Smog Chamber During September,
1973 57
8 Relative Distribution of Individual Compounds
of the Hydrocarbon Mix in Tanks and in
Chamber Air After Injection 72
9 Rates of Ozone Decay in UNC Outdoor Chamber
Using Background Air 75
10 Rate of NO Loss and N02 Formation in UNC
Outdoor Chamber Using Background Air.
(Initial) 77
11 Mix Performance With and Without Aromatics
(Side to Side Comparison) 84
12 Mean Values of Nitrogen Dioxide Maximum and
Average Concentration, Average Daily Solar
Radiation as a Function of Mean Initial
Nitrogen Oxides and Mean Initial Hydrocarbons 122
13 Nitrogen Dioxide Maximum Concentrations, Time
to NO-NO? Crossover, Net N02 Formed, Average
Solar Radiation at Initial NOX Average Con-
centration of 0.361 ppm Broken Down by Initial
Hydrocarbon Average Concentrations Average
Solar Radiation Values 128
XI
-------
LIST OF TABLES (cont)
14 Pairwise Comparison of Effects of Reducing
Initial Hydrocarbon Concentration on
Nitrogen Dioxide Maximum Concentrations
for Selected Dual Runs with Urban Hydro-
carbon Mix in UNC Outdoor^jChamber 135
15 Mean Nitrogen Dioxide 10 Hour Average Con-
centration and Mean Average to Maximum
Nitrogen Dioxide Ratio Broken Down by
Mean Initial Oxides of Nitrogen and Mean
Initial Hydrocarbon for Urban Hydrocarbon
Mix in UNC Outdoor Chamber 146
16 Pairwise Comparison of Effects of Reducing
Initial Hydrocarbon Concentration on
Nitrogen Dioxide Average Concentration
and Dosage for Selected Dual Runs with
Urban Hydrocarbon Mix in UNC Outdoor
Chamber 147
17 Magnitudes of Terms in Nitrogen Dioxide Max-
imum Regression Equation 150
18 Pairwise Comparison of Effects of Reducing
Initial Hydrocarbon Concentration on
Final Nitrogen Dioxide Concentration for
Selected Dual Runs with Urban Hydrocarbon
Mix in UNC Outdoor Chamber 152
19 Comparison of Daytime and Nighttime Nitrogen
Dioxide Dosages for Extended Dual Runs with
Urban Hydrocarbon Mix in UNC Outdoor Chamber 160
20 Number of Hours Exceeding Various M02 Concen-
trations 163
21 Air Temperature Statistics at Various Chemical
Events in Outdoor Chamber 175
22 Organic Symbols Used in Mechanism 197
23 Confidence Intervals on Accuracy of Measurement
of Ozone at 8 PPHM 203
xn
-------
ACKNOWLEDGEMENTS
Several people at the Department of Environmental Sciences and Engineer-
ing, School of Public Health, University of North Carolina (UNC), and
at the Environmental Protection Agency, National Environmental Research
Center, Research Triangle Park (EPA), contributed substantially to the
work performed on this project. Laura Alexander (UNC) served as data
management assistant and masterly handled the many details of day-to-
day operation. Dr. Robert Baker (UNC) was primarily responsible for
experimental and data processing program development and also performed
statistical analysis of data. The experimental facility was constructed
by Jack Brown and Wallace Pendergraft of the UNC School of Public
Health Shop under the management of Frank Malcolm. Many of the experi-
ments were performed by UNC graduate students George Si pie, Dennis
Stotts, Surendra Joshi, and Charles Feigley. Charles Feigley also
performed the detailed statistical analysis of instrument calibration
and the majority of the statistical analysis of results. Much of the
detail of the data processing was performed by UNC graduate students
Chingman Kwan, Glen Sims, and Mary Witul. Robert Jones implemented
most of the computer-instrument interface and the automatic control
systems for the chamber. Joseph Sickles developed the theory and
design of the continuous NCL chemical actinometer used to make chemical-
ly meaningful light intensity measurements.
In addition, Joseph Sickles and Bruce Gantner were instrumental in
developing the photochemical models used. Messrs Jones, Sickles,
and Gantner were also graduate students of UNC.
Drs. Marcia Dodge and John Overton of EPA's Chemistry and Physics
Laboratory supplied a copy of EPA's kinetics simulation program and
smog chamber data from several of the chambers modeled in this report.
William Lonneman, also of EPA, implemented the gas chromatographic
system used in this study.
xm
-------
Dr. Basil Dimitriades and other members of the Chemistry and Physics
Laboratory of EPA spent many hours in discussion and planning meetings
with the UNC staff throughout the project. Dr. Dimitriades1 comments
were especially appreciated.
Dr. Lyman Ripperton, presently associated with the Research Triangle
Institute provided valuable guidance during the initial phases of the
project.
xiv
-------
SECTION I
SUMMARY AND CONCLUSIONS
One hundred thirty 12-hour runs were conducted in a large dual, outdoor
smog chamber utilizing ambient conditions of solar radiation, tempera-
ture and relative humidity, and using a chemical system consisting of a
simulated urban hydrocarbon mix and oxides of nitrogen (NO ). These
y\
experiments were performed over an initial hydrocarbon range of 0.22 to
4.2 parts per million carbon (ppmC) and an initial oxides of nitrogen
concentration range of 0.007 to 0.7 parts per million (ppm). The initial
NO concentration consisted of 80% nitric oxide (NO) and 20% nitrogen
A
dioxide (NO-) in all but a few experiments. Data were obtained by
performing runs in pairs, with the same initial NO on both sides of the
A
dual chamber, but different initial nonmethane hydrocarbon (NMHC). This
allowed a direct determination of the effect of reducing NMHC on NOp
maximum concentrations and dosages. Relevant evidence was also obtained
from comparison of runs conducted on different days, with different
ambient atmospheric conditions.
A non-linear multiple regression equation for N02 maximum concentration
accounted for 92% of the total variance in N02 maximum. The independ-
ent variables were initial NO , initial NMHC, average solar radiation
X
between sunrise and the NO-NOo crossover, the average solar radiation
near the NO- peak. In the equation, N02 maximum was primarily dependent
upon initial NO ; the functional dependence upon initial hydrocarbon
+
concentration could cause a - 20% variation in the initial NO dependence;
y\
the functional dependence upon solar radiation measures could also cause
a further -5% non-linear variation in the initial N0v dependence.
Comparison of results from dual runs conducted on the same day, under
the same solar radiation and initial NO conditions, clearly showed a
A
decrease in N0? maximum concentrations with decreasing initial NMHC for
all cases. For example, at an initial NO concentration of 0.51 ppm, a
.A
72% reduction of NMHC, from 2.58 ppmC to 0.70 ppmC, reduced both the
-------
N02 maximum and the NOp dosage by 35%.
Comparison of results from runs conducted on different days, under
different solar radiation and constant initial NO conditions, showed
A
a decrease in NCL maximum concentrations with decreasing initial NMHC,
but the decrease was also a function of exact solar radiation conditions
during each run. For example, at an initial NO concentration of 0.36
X
ppm on different days, a 56% reduction in NMHC, from 3.2 ppmC to 1.4
ppmC, reduced the NOp maximum by 20%.
Overall, for initial NMHC and NO levels similar to those typically
A
present in current urban atmospheres, reduction of the NMHC reactant
by 60-90% caused a 20-35% reduction in both NOp dosage and maximum NOp
concentrations.
These results are from runs which were started before sunrise and
terminated at 1700 hours EDT and were conducted under static conditions.
To explore the possible effects of conditions more closely simulating
the real atmosphere, several runs were extended to 24 or 36 hours.
Other runs were diluted at rates of 7-10% per hour from early morning
until 1700 hours.
In the 24-hour extended runs, high concentrations of NOp sometimes
occurred throughout the nighttime period. These high nighttime NOp
concentrations were the result of high NOp and low 0., at the end of
the solar day. Those runs which resulted in high (> 0.1 ppm) ozone
concentrations at sunset had much lower nighttime NOp concentrations
because the NOp was consumed either during the day or during the
nighttime period by reaction with 03- Because various solar radiation
conditions could lead to different Oo concentrations at the end of the
solar day from the same initial NO and NMHC, the nighttime occurrence
A
of NOp was a function of the exact solar radiation conditions during
the daylight period. For those runs that had high nighttime NOp
concentrations the nighttime contribution to the total 22-hour NOp
-------
dosage was as high as 60%. For those runs in which NCL disappeared
during the night, the nighttime contribution to total N0? dosage was
only 25-28%.
One extended experiment was performed which indicated,, that although
hydrocarbon reduction would be capable of reducing both 03 and NOo
maximum concentrations during the first daylight period, it might have
little effect on 0., and N02 maximum concentrations in the second
daylight period. For example, in a 36-hour static experiment, a 50%
reduction in NMHC (from 2.9 ppmC to 1.3 ppmC at initial NO of 0.36
J\
ppm) resulted in a 10% reduction in N0? maximum concentration and a
74% reduction (from 0.392 to 0.103 ppm) in 03 maximum concentration in
the first daylight period. In the nighttime period, N02 in both sides
decreased to approximately 0.05 ppm while the ozone decreased to 0.24
and 0.04 ppm; the NO concentration was expected to be essentially
zero. In the second daylight period, the N02 remained relatively
constant at 0.045 ppm while the ozone increased in both sides with np_
additional material added to a value near that of the higher NMHC run
on the first day (0.40 and 0.37 ppm 0.0. This occurred even though
the N02 concentrations remained below 0.05 ppm and the "reactive"
hydrocarbons in the mix had presumably been consumed the previous day.
Slow dilution of relatively reactive propylene/NO systems led to
X
lower N02 and 03 maximum concentrations than occurred in the same
undiluted chemical system. Dilution of lower reactivity hydrocarbon
mix/NO systems led to lower N0? concentrations, but higher 0., concen-
s\ £- J
trations than occurred in the same system without dilution. This was
primarily due to more rapid achievement of higher N02/N0 ratios in the
diluted system at the most optimum time in the solar day for 03
generation.
The concept of a large outdoor chamber, which used ambient conditions
of solar radiation, temperature, and relative humidity, was very
useful in studying photochemical systems representative of urban
-------
environments. The data from such a chamber were more difficult to
interpret than that from more traditional (artifically illuminated)
smog chambers because of the day-to-day variations in solar radiation
and other meteorological factors. The advantages of performing a
large number of experiments under essentially "real" conditions,
however, proved to be significant. The results under these real
conditions differed from those which had been previously obtained in
indoor chambers with constant conditions of light, temperature, and
humidity. The experimental approach adopted in this study represents
somewhat of a half-way point between the traditional smog chamber
approach and that based on direct observation of the real atmosphere.
-------
SECTION II
RECOMMENDATIONS
These experiments were conducted under essentially static conditions.
Work reported in this document would suggest that the static situation
may not always be the "worst case" condition because of interaction
between the chemical system and the natural diurnal light-intensity
and physical operating conditions. There was a solar radiation effect
in the results reported. Any factor which changes the speed of NO to
N0? conversion will probably affect the N0? maximum concentration and
most definitely will affect the 0^ formation. It is recommended that
such factors as rapid dilution and continued injection of fresh
materials just after the early morning peak be examined.
The information content of the present data set has not been exhausted.
It is recommended that further data analysis be done, especially in
relating solar radiation variations to outcomes in experiments. This
analysis will also require further specialized experimental runs to
support findings. The formation of 0~ under outdoor conditions is
very much a function of the photochemically weighted light-history.
These factors need to be determined and have not been addressed in
this report.
Factors relating rate of NO- photolysis to total solar radiation and
ultraviolet radiation need further exploration, especially on days
which exhibit rapid increases and decreases in light intensity, and on
thinly overcast days.
The roles of nitrous acid and other photoacceptors such as aldehydes
at the very beginning of irradiation require clarification since the
presence of these photoacceptors determines the initial speed of NO
to N0? conversion and therefore affects the subsequent formation of 0~
under outdoor light conditions. A similar situation exists with
respect to the initial N0~/N0 ratio. It is recommended that direct
C. A
-------
comparison experiments be conducted in which runs with added nitrous
acid anH aldehydes are compared to runs with only hydrocarbon/NO .
}\
Further experimental runs should measure specific N0?, nitric acid,
PAN, specific hydrocarbons and aldehydes. This data would be needed
to validate future photochemical models and aid in determining the
factors leading to CL generation.
The modeling program should be modified to allow variable temperature
and dewpoint profiles to continuously adjust those rate constants that
are known to be temperature dependent or include water concentration.
The second and third day irradiation of so-called "spent" reaction
systems requires a detailed exploration. Results reported herein would
suggest that although hydrocarbon reduction might be beneficial, in
terms of controlling N0? and ozone concentrations on the first day of
irradiation, it may not be successful on the second and subsequent
days. The ozone concentration in the downwind air parcel on these
days may be primarily a function of dilution and two or three day
weather patterns.
The precursor-pollutant relationship found in this study was differ-
ent than that which had been found under constant light-intensity con-
ditions in a previous study. This has been attributed to the diurnal
variation in solar radiation and other natural phenomena associated
with solar radiation. Because of this outcome, it is recommended that
other precursor-pollutant relationships that have been or will be de-
termined in smog chamber studies should be examined to determine if
real solar radiation conditions will affect the relationship.
-------
SECTION III
INTRODUCTION
BACKGROUND
On April 30, 1971, the Environmental Protection Agency promulgated na-
tional ambient air quality standards (NAAQS) for photochemical oxidants
(0 ), nitrogen dioxide (N09), and nonmethane hydrocarbons (NMHC) . As
A L,
put forth in the Federal Register, NMHC at an ambient level of 160
o
ug/m (0.24 ppm as methane, 6-9 a.m. average not to be exceeded more
than once per year) would serve as a guide to achieve the one hour
oxidant standard of 0.08 ppm (not to be exceeded more than once per
year). Reduction in nitrogen oxide emissions was employed as the most
effective way of controlling NO^. The N02 standard, which is an
annual average of 0.05 ppm, was developed from N0~ health effects
23
studies, independently of 0 control strategy.
A
At the time the standards were set, it was not possible to fully
assess the complex short-term photochemical relationships between
oxides of nitrogen (NO ), NMHC, and 0 . Hence, the national control
A A
strategy for ambient ozone consisted of reducing only the hydrocarbon
precursor based on an upper limit relationship between nonmethane
hydrocarbon and ozone. This relationship was developed from aerometric
data only. Such a relationship did not take into account the role
played by the oxides of nitrogen and the potential effects that NMHC
reduction may have on NO^ concentrations.
4
Early chamber findings of Korth, et al. , were able to show, by using
dilute auto exhaust, that a decrease in initial hydrocarbons resulted
in a proportional decrease in the rate of N09 formation (at constant
r *-
initial NO ). Dimitriades in the Bureau of Mines (BOM) smog chamber
A
using NO (90% NO, 10% N09) and auto exhaust, irradiated at constant
A C.
light intensity, demonstrated that maximum N02 concentrations (N02M)
-------
were essentially independent of initial NMHC concentrations in the
range of 0.0 to 1 ppm initial NO (INOX) and 0 to 4.6 ppmC initial NMHC
A
(IHC). Dimitriades' data are shown in Figure 1; from these data he
estimated a value of 0.63 for the ratio of N02M to INOX. Possible dif-
ficulties with various control strategies involving both NO and NMHC
fi
were later succintly illustrated by Stephens . Although Stephens in-
tentionally does not quantify his axes (Figure 2), his diagram suggests
that excessive hydrocarbon control compared to NO control could have
J\
the tendency of shifting high N02 concentrations to a later time in the
day with a possibility of greater N02 dosages.
Aitshuller has stated in summary that "the formation of nitrogen di-
oxide can be slowed up by a high degree of control of hydrocarbons".
However he also points out that "the way irradiation chambers are
operated starting at near midday light intensities and temperatures
accelerates the early stages of the reaction unrealistically rapidly
compared to atmospheric conditions". For this reason it is important
to determine if a relationship like the one in Figure 1 would hold
under more realistic conditions of natural solar radiation, temperature,
and relative humidity. Large reductions in NMHC could lead to higher
N02 dosages with potentially, worse N02 health effects, while achieving
the oxidant standard. This would require more rigid emission standards
for NO .
X
Thus, while the effect of reducing NMHC on oxidant formation is con-
sidered beneficial, the effect of this reduction on NO,, concentrations
in the ambient atmosphere is not clear. Will N02 peaks be lower,
higher, or the same? Will NOo dosages increase? Will it be more
difficult to meet the annual standard for N0?? Is the formation of N0?
slowed so significantly that large N02 concentrations occur at the end
of the day? If so, what might be the fate of this remaining NO and
N02?
-------
en
o
CO
o
r-
-------
OJ
I/)
c >>
O i—
•r- QJ
tO
— K CC — O OX— Q
Q.
-o
C
(O
^—
o
to
o
s-
s_
o
w
o>
r—
•r*
'[
O
S-
Q-
s — s
cu
03
i-
4_>
*r~
C
>^
o
(X3
>^
><
o
O)
CL
^ — •
i.
Q)
J3
•<-
'oi
•o
rO
C
•r-*
-o
0
o
o
(1)
-C
H-
•
•"»
CU
o
c:
-------
There are three investigative methods for finding answers to these
questions, each with inherent advantages and disadvantages. These are:
1) examine existing ambient air monitoring data (aerometric data) and
collect new aerometric data, 2) use models of chemical and physical am-
bient air processes, and 3) perform new smog chamber experiments.
While the aerometric data approach has potential for giving answers to
the questions posed above, the existing data is inadequate. Collection
of new data is expensive and long term. Since we cannot control the
conditions observed in the atmosphere, the most interesting conditions
might not actually exist in the present atmosphere. Any relationship
that might be determined would probably be a correlative or statistical
one and not a causative one. Aerometric data could, perhaps, provide
validation of results derived from the other two approaches.
The modeling approach could give a detailed mechanistic or physical un-
derstanding of the effect of hydrocarbon reduction on NCL behavior. Any
combination of conditions and strategies could be tested and the condi-
tions would be easily manipulated. The difficulty with this approach,
at the present time, is that vital information is missing or untested.
The model must be validated.
The traditional smog chamber approach suffers from the fact that con-
ditions in the chamber are not necessarily very representative of those
in ambient air. Factors such as surface effects, constant light inten-
sity, unrealistic spectral distribution, constant temperature and rela-
tive humidity, static operating conditions, and reactant identity
usually differ from actual urban air and thus would make it very diffi-
cult to extrapolate results obtained in smog chambers to ambient air
behavior. On the other hand, within the limits of the operating con-
ditions, a properly designed smog chamber system could provide answers
to the questions asked above. Compared to the other two approaches
these could be obtained relatively inexpensively and quickly. Like the
11
-------
modeling approach, conditions not found in the ambient air could be
tested and although a mechanistic understanding would not necessarily
result, cause and effect relationships could be found.
PURPOSE
The purpose of the work performed in this study was to determine, in a
smog chamber operated under conditions typical of ambient air, the ef-
fects of hydrocarbon reduction on nitrogen dioxide concentrations in
chemical systems reasonably representative of urban conditions. In
addition, because the chamber would be located outdoors, an assessment
was to be made of operating characteristics and performance of the
chamber. Results were to be compared with those from indoor chambers.
APPROACH
To obtain physical operating conditions in a smog chamber that are
typical of ambient atmospheric conditions, a smog chamber can be
built outdoors. The daily variations in outdoor conditions present
difficulties not found with indoor chambers. On the other hand, these
variations exist in nature and are therefore useful in the experiment.
The effects of these variations can be treated by proper experimental
design and analysis. Using these concepts, an overall approach was
devised that consisted of four major steps. These were:
1) The design and construction of a dual outdoor smog chamber
which was capable of operating under conditions of light,
temperature, and humidity typical of urban ambient situations.
12
-------
2) The testing of the chamber and associated facility. This
consisted of determining if the chamber design was physi-
cally sound and durable, and if it exhibited reasonable
performance characteristics with respect to generally
accepted smog chamber measures.
3) The performance of a series of 12 hour experimental runs in
which the initial starting conditions of NO and NMHC were
X
varied over a range typical of morning urban atmospheric
concentrations, before and after hydrocarbon control.
4) The limited exploration of the effects of variations in
operating conditions on the outcome of the experiments in 3)
above. These consisted of:
a) different initial percentage of NC^.
b) extended run times (24 and 36 hours durations),
without additional material injected, and
c) slow dilution of the reactive system with clean
air.
In addition, numerous supportive tasks that could affect data quality
or understanding were performed. Some of these tasks were instrument
performance testing, development of a chemically meaningful light
intensity measurement system, development of computer programs and
hardware needed to operate the project effectively, analysis of any
existing data that would be of benefit to the overall project goals,
and the investigation of current modeling technology.
13
-------
SECTION IV
DESIGN AND CONSTRUCTION OF SMOG CHAMBER FACILITY
DESIGN CRITERIA
Having made the decision to build a chamber outdoors for the purpose
of achieving a more realistic simulation of urban ambient conditions,
the choice of location, size, construction materials and methods, air
handling or air movement, pollutant introduction, orientation, and
type of supporting facilities remained. Boundaries on these choices
were introduced by identifying desired characteristics and physical,
chemical, or financial limitations. Several of these factors affected
more than one of the areas of choices listed above; therefore, these
factors will be discussed first.
Besides being located outdoors, it was desired that the chamber exhibit
low surface effects. This implied a very large volume and choice of
inert materials (few metal surfaces). This large volume requirement
introduced other design requirements, such as: 1) finding a location
that had background air that would be suitable for use without clean
up because it would be prohibitively expensive to clean 350,000 £ for
every run, 2) providing for relatively short purging times so that
runs could be performed every day if necessary, 3) assuring that the
volume was well mixed to avoid the problems associated with inhomogen-
eous chemistry, and 4) selecting arrangements so as to provide adequate
laboratory space near chamber without shadowing but still allow easy
access to chamber contents.
Because we recognized that outdoor conditions can vary from day to
day, the chamber design had to provide some method of experimental
control over these effects. This requirement suggested a dual chamber
arrangement or dual compartment system. This meant doubling the air
handling system hardware and required either two or each instru-
14
-------
ment type or 4 .nesharing of each instrument between the two chamber
compartments.
Since the purpose of operating outdoors was to use atmospheric light
and temperature conditions, the chamber design u-d to allow for full
solar intensity and reasonable solar spectrum fidelity, but had to
avoid a large "greenhouse effect" for air temperature. These desired
characteristics affected chamber orientation, wall materials, and the
physical arrangements of the chamber.
The varying ambient conditions, the long run times, the number of
variables to be measured, and the need to timeshare the instruments
suggested that it would be necessary to fully automate the data collec-
tion and processing. More than 10,000 numbers would be required for
each dual run, primarily because of varying ambient conditions. A
minicomputer data acquisition and control system would greatly increase
the project productivity.
Finally, the entire chamber facility had to be constructed at a reason-
able cost (less than $20,000). This meant that, where possible, wood
was used as opposed to metal. Techniques that could be implemented by
project staff were used extensively as opposed to those that required
commercial hardware or personnel.
FACILITY DESCRIPTION AND CONSTRUCTION METHODS
Location
A site was chosen approximately 32 kilomete, "• from the University of
North Carolina at Chapel Hill. It is in Chatham County, North Carolina
and is approximately 10 kilometers from the small town of Pittsboro.
Chatham County is one of the most rural, least industrialized counties
in North Carolina and is heavily wooded. The background concentrations
15
-------
of NO and nonmethane hydrocarbons were usually less than 0.025 ppro
A
and less than 0.20 ppmC. More Importantly, the air exhibited very low
reactivity in the chamber.
Materials
The most important factors that influenced the choice of materials
were that the chamber surfaces had to be chemically inert, durable,
and have a very high transmission for the entire solar spectrum.
Glass was unsatisfactory because of its spectral properties in the
infrared (IR) and short wave length ultraviolet (UV) regions. Most
films exhibit undesirable surface effects at low concentrations of
trace gases. Fluorinated ethylene propylene (FEP) Teflon film however,
had been used in many laboratories for years to make bags for photo-
chemical experiments and seemed to be the most acceptable material.
Its transmission in the UV and visible regions of the solar spectrum is
8 9
excellent ' and it has only a few absorption bands in the IR, a
property necessary to reduce the "greenhouse effect". It has a very
9 10
low permeability for most chemical species and can be heat sealed
to form large durable panels. For this application its worst property
was its ability to hold a static charge for long periods of time.
Type A film, 0.13 mm thick was chosen since film suppliers had indicated
that Type C film would not hold a heat seal when exposed to sunlight .
This film is available in widths of 117 cm and in rolls of 34.4 m. To
facilitate the replacement of a damaged section, the technique for en-
closing the volume had to be one that would accommodate individual 117
cm wide Teflon panels for the sides. Other large areas that would not
be as subject to wear could be fabricated from 117 cm wide panels com-
mercially heat sealed together to form one piece, leakfree panels.
16
-------
Physical Design ana Construction Techniques
Because inlet and outlet doors, stirring fans, manifolds and other
fittings had to enter the film in an airtight manner, it was decided
early in the design that these fittings would be inserted through a
solid floor. This left the sides free for light _ntry. The floor of
the chamber was elevated approximately 1.2 m to allow for easy access.
Several shapes were considered for the chamber: spherical, half
cylinder, box, dome, and A-frame. All but the A-frame were eliminated
because of either difficulties in achieving a watertight structure or
prohibitive costs. The final design was an A-frame 9.14 m wide, 12.19
m long and 6.10 m high at the peak on a plywood floor 1.22 m above the
ground. Wooden beams, 5.08 cm by 20.32 cm, located on edge at 99.1 cm
centers formed an exterior framework on top of the plywood floor.
Continuous 16.46 m lengths of film were attached to the inside of the
beams with a 0.61 m overlap on the floor at each side. The film was
held in place from the inside by aluminum u-channels, nailed firmly to
the wooden beams compressing the film against the external support.
See Figures 3 and 4.
One piece, heat sealed Teflon film panels were used in such places as
the triangular end panels and floors. The floor to side seals were
achieved by a 0.61 m overlap of side panel film over floor panel film.
A rubber strip under the film and an aluminum strip on top of the film
completed the seal. All other seals were Teflon to Teflon under pres-
sure of the aluminum u channel. A single unsupported heat sealed Tef-
lon panel similar to the end panels was used to separate the chamber
into two halves of equal volume. It was sealed to the floor and side
panels in the same manner as described above. The film was stretched
as tight as possible during hanging, but enough flexing occurred to
compensate for small volume changes due to temperature changes from
cloud passage or wind pressure changes.
17
-------
o
JD
CO
c
O)
(L>
ts
0)
«*-
o
CO
0)
-p
s-
3
O
(U
-Q
o
i-
o
o
-o
4J
(O
o
s-
o
-p
o
q-
o
10
S- CO
=> _>>
-------
UJ
Q
Q
19
-------
Aluminum foil was placed under the film on the floor to reflect the
light and heat backup through the chamber. This was necessary to re-
duce solar heating of the air to a value that was within normal urban
environments and to compensate for transmission losses through the
Teflon film.
Orientation
The chamber was oriented with its long axis approximately north to
south. The actual long axis orientation is along a 27° - 207° true
heading. The orientation with respect to the sunrise and sunset at
different times of the year and the altitude of the sun at noon for
different times of the year are illustrated in Figures 5 and 6.
It was expected that the reflective chamber floor would interact with
the sun to increase the photon flux in the chamber and thus compensate
for transmission losses through the film. Multiple reflections in the
two chamber sides would probably help to equalize overall solar expo-
sure.
Air Handling Systems
There were three air handling systems in each half of the chamber, one
for exhausting, one for sampling, and one for mixing. The exhaust
system consisted of two intake stacks, 0.61 m x 0.91 m intake doors,
0.61 m x 0.61 m exhaust doors, and an exhaust blower. The exhaust
blower was a dual blower on a single shaft driven by a 1.5 horsepower
motor. Air entered the system through the two 5.49 m high by 30.5 cm
diameter stacks. This system was designed to permit rapid exhausting
of chamber contents and replacement with ambient air. The filling
rate was 7190 1/min. The chamber could be flushed to a 99.6% decay of
initial contents in 2 hours.
20
-------
JUNE 22
JULY 23.MAY2I
AOO 24, APRIL ZO
SEPT 23, MAR 21
OCT 23, FBB 20
NOV 23, JAN 21
DEC 22
207*
/ s
Figure 5. Orientation of UNC outdoor smog chamber with respect to
seasonal sunrise and sunset positions.
Outdoor Dual
Smog Chamber
Figure 6. Solar altitude and zenith angle at noon at the UNC outdoor
smog chamber for each month.
21
-------
The second air handling or manifold system was for sampling and in-
jection of pollutant materials into the chamber halves. To insure
representative sampling, a 3.17 cm I.D. glass manifold ran from a
point 1.83 m above the floor in the center of each chamber half down
through the floor and over to a sampling laboratory. The sampling
volumes required by all the instruments did not exceed 5 1/min but to
reduce losses due to long resident times it was necessary to have
higher flows in the manifold. The flow rate in the manifold was 60
1pm. Over the duration of a 12 hour experiment, this could have
exhausted 28% of the chamber volume. To avoid the necessity of makeup
air, the sampling manifold was made as a closed loop. Squirrel cage
blowers with housings and fans that were Teflon coated were used to
circulate air through the manifolds. The unused sample air was then
returned through a 3.17 cm I.D. glass manifold to the chamber. These
return manifolds provided a convenient method for injecting the initial
reactants.
Inside each chamber half were two mixing fans located in opposite cor-
ners. These provided circulation and mixing of the chamber contents.
The fans were 50.8 cm diameter cast aluminum units that had been FEP
Teflon coated. They operated in a horizontal position, 0.76 m above
the floor on 2.54 cm diameter Teflon coated steel shafts that extended
through the chamber floor. Under the floor, 1/4 horsepower, 1750 rpm
motors provided power through a belt and pulley system to each fan.
3
Each fan operated at approximately 31.15 m /min.
Laboratory
A laboratory was built adjacent to the chamber. It was a 3.66 m W x
15.24 m L x 3.05 m H wooden structure oriented perpendicular to the
chamber and 3.66 m away from it to avoid any shadowing. The first
5.49 m nearest the chamber contained the instrumentation, manifolds,
and calibration systems. The next 5.49 m contained the data acquisi-
tion computer system and the operations area. The last 4.27 m of the
22
-------
laboratory was a utility area with running water and storage facilities.
Gas tanks necessary to operate the instruments and perform calibration
were located in a 1.22 ra x 1.83 m room completely closed-off from, but
attached to the laboratory. The injection system gas tanks and valves
were housed in a second 1.52 m x 1.52 m well-insulated, heated room
adjacent to the end of the laboratory nearest the chamber.
Injection System
Pollutants were injected into the chamber sides via the return side of
the sampling manifolds. The return manifolds entered the chamber
sides under one of the mixing fans. The injection process used gas
cylinders containing pollutants at high concentrations (1000-10,000
ppm range), two stage, stainless steel diaphragm regulators, on-off
solenoid valves, and capillary flow restrictors or precision needle
valves with vernier handles. The flow rate of injected material into
the manifolds could be well established for a given pressure and
length of capillary tubing or needle valve setting; the total injection
volume could be accurately controlled as a function of the time the
solenoid valve was open. Conditions could be varied sufficiently to
have the injection time range from a few minutes for each component to
1-2 hours for a programmed injection used to simulate the buildup of
pollutants in urban areas.
Data Acquisition System
A computer based data acquisition and control system (DAS) was used to
acquire, process, and record data for the chamber instrument system.
The DAS consisted of: 1} a PDF 11/40 computer with 32 kilobytes of
core memory, 2) a high speed papertape reader/punch, 3) a 2.4 million
byte cartridge disk system, 4} thirty character per second keyboard/-
printer terminal, 5] time-of-day clock, 6) 5-1/2 digit, digital
voltmeter, and 7) a 200-channel, 3-wire crossbar scanner.
23
-------
Output signals from each instrument were wired to the crossbar scanner
(or analog signal multiplexer). Under control of programs in the com-
puter, the scanner connected the selected signal leads to the input of
the digital voltmeter. The digital voltmeter, which had excellent
noise and spurious signal suppression and could measure a 1 volt sig-
nal with a resolution of 10 microvolts, was triggered to acquire a
reading and to supply the binary coded decimal, 5-1/2 digit number to
the computer. The computer processed the information and then com-
manded the scanner to move to the next channel and repeat the process.
The data processing will be described in a later chapter. The timing
of the computer processes was under the control of a digital date and
time-of-day clock which signaled event times to the computer. Another
digital timer operated Teflon solenoid valves on the inlet of each in-
strument to connect the instruments to one chamber half during a given
cycle of data acquisition.
This system provided fully automatic acquisitions and processing of
data during a run and provided the operator with immediate, full data
in physical units. Given this information about what was happening,
the operator could then concentrate on what he wanted to do. It also
allowed the massive amount of information generated during a run to be
processed in a more effectual manner after the run was over.
24
-------
SECTION V
EXPERIMENTAL METHODS AND PROCEDURES
MEASUREMENT METHODS
Instrumentation
The instrumentation was located in the temperature-controlled labora-
tory adjacent to the chamber. Gas instrumentation included equipment
for measuring total hydrocarbons, methane, carbon monoxide, nitric
oxide, nitrogen dioxide, and ozone. Standard meteorological instru-
ments were used to measure solar radiation, ul^a-violet radiation,
air temperature, and dew point. A temperature-programable, dual,
flame ionization gas chromatograph was used for detailed hydrocarbon
analysis. Table 1 lists the instrument models and their characteristics.
At the sample inlet of each gas instrument was a 3-way Teflon ac sole-
noid valve. Since there were two intake manifolds (one for each cham-
ber half), air from either manifold could be drawn through a 3-way
valve and into the instrument. In this manner timesharing of one in-
strument between the two chamber halves was possible.
Calibration
The techniques used to calibrate the continuous monitoring instruments
closely followed the procedures set forth by EPA , although other
methods were used as needed (see Table 2). An attempt was made to
interlock the NO and Oo calibration apparatus so that each specific
A O
method could be used as a check against the others. That is, ozone,
from a neutral buffered potassium iodide calibrated ozone generator
was used in combination with a standard tank source of NO to produce
12
an equivalent amount of N0~ by a N0-03 titration. This calibration
25
-------
s-
ID
Ol
>-
O
•o
O
o
c
o
o
o u
i- O) £
i— »/> W
U rO
co
I/I
3
o
n
c
o
o
c
o
o
o
c
I/I
o
f~ ID
«
c
31 i
a °-
Oil .-
I.
D.
o
o
o
CO
o
o
CM
o
»s-
I
*io f-
u E
CM
o
1O
u
ID
O
O
O
in
O x:
O u
• ID
O OJ
o JT
• o
o *
v o>
o
o
ID
U
r— C
in CM CM
r~ u o
O »«
CM r—
JD
IO
O)
o
re
c\r
o
X
•r*
•o
O)
CO
CO
I
o
00
X
•o
c
o>
oo
0>
•— .C -O
ID O. O
«C +•> ID X
CO C t-
m a; o 01
CM E o cne
>> > i-
CO C J=
LU (->
JD
E
to
o
u
ID
•— OJ
ca *J
Q.T3
O. C
LU (0
00 Q»
TJ
o
•P
0)
0)
U
1/1
0)
01
u
i/i
OJ
ID ID
t— N
O
o
•^
•a
u
ID
4->
V)
00
C
1C
o
O
4->
l/l
Ol
CM X
o o o
«»• c
X O i-
<_J O O
ex
Ol
o
L.
ID
"o o
r— ID ID
IO T- Qi
•*-> T3
O ID •>
t— Q£ 3
O
s- ox
3 0£ J=
+j HH a. T>
>a i o o)
1- OO L. i—
OJ ^i j-> M-
CL > U 10
E i a; -u
ID
g
U
o
u
a>
jz
as
u
O)
a>
•o
i
26
-------
CM
OO
cr>
CO
CU
O
E
cu
cu
^~
cu
I—
en
cr>
cu
0)
r-.
cr>
(O
(O
CL)
-C
QJ
o:
r-
cr>
en
cu
c.
a
2
3
CO
ai
ercia
Co
"O
cu
s-
cu
o
CO
CO
3
o
a
rc
Co
cu
S-
cu
o
CO
CO
c
-------
factor for N09 was compared with those obtained from a volumetrically
13 14
calibrated permeation tube , standard bubbler method and from
certified N02 commercial cylinders. The total hydrocarbon (THC),
methane (CH.) and carbon monoxide (CO) modes of the environmental
chromatograph were calibrated using commercially prepared gases in the
1-15 ppmC range in HC-free air (< 0.1 ppmC). The response factor for
the THC mode of the instrument for nonmethane hydrocarbon (NMHC) was
determined by additional calibration with commercially prepared cylin-
ders of butane and hexane. These calibrations were also compared with
a volumetrically calibrated propane permeation tube. Thus, multiple
methods were employed to insure as accurate a calibration as possible.
Initially, a complete calibration of all instruments was performed
prior to each run until instrument-calibration familiarity and general
drift patterns were established. Subsequently, it was found that the
environmental chromatograph required daily calibration; the NO and
J\
03 meters, however, only required calibration approximately once a
week.
STANDARD OPERATING PROCEDURE
Prior to an experimental run, the chamber was purged overnight with
background ambient air. At precisely 0500 hours EOT the DAS began
recording measurements of the background concentrations of THC, CH.,
CO, NO, NO- and 0.,. Single instruments were timeshared on an alterna-
ting four minute cycle between the two sides. The chamber intake and
exhaust doors were sealed before sunrise. The hydrocarbon analyzer
was calibrated. The background concentrations of NMHC and NO were
J\
then subtracted from the desired pollutant concentrations to determine
the injection concentrations. The starting concentrations of NO, NOp
and NMHC were injected, in the order given, as a slug injection by the
use of high concentration gas tanks, electric on-off solenoid valves
and precision needle valves. Depending on the pollutant to be injected,
the on-off solenoid valves were actuated from inside the laboratory
28
-------
for predetermined periods of 1-7 minutes. Time-flow relationships and
instrument readings verified the injection concentrations. Automatic
monitoring of the pollutants occurred until 1700 hours EOT when the
run was terminated. The chamber doors were then opened and the chamber
contents were purged with ambient air. Data was printed out and
punched from the computer disk and the computer was programmed to ini-
tiate another run at 0500 hours the following day.
DATA TREATMENT PROCEDURES
General
Figures 7, 8, and 9 are flowcharts of the processes used to treat the
data collected. The processes are divided into those that occurred
during a run, those performed immediately after a run, and those that
were done later. The operations were carried out on three different
computers and a programmable calculator with plotter. The steps in
Figure 7 occurred on the POP 11/40 at the chamber site and on a Hew-
lett-Packard 9810 programmable calculator with plotter at UNC. The
steps in Figure 8 occurred on a POP 11/40 at EPA which was equipped
with a compatible disk drive and nine track magnetic tape unit. The
major data manipulation steps illustrated in Figure 9 were performed
on an IBM S370/165 at the Triangle Universities Computational Center
(TUCC) which supports a wide range of statistical programs, telepro-
cessing operations for graphical processing, and large on-line and
off-line storage facilities.
During Run
The operator began data acquisition operations by answering questions
from the keyboard (Step 1A, Figure 7). These included supplying the
starting day and time and the finishing day and time. With this
information, the acquisition program would initiate and cease data
29
-------
-------
(C
Ul
o
>
to
ns
+->
(O
O)
-P
c
'i
O)
Q.
o
s-
JC
u
o
u.
cd
O)
CD
31
-------
TEKTRONICS
4010
TERMINAL
WRITES OVER OLD
FILE ON TSO DISK
WRITE OVER OLD DAT FILE
[CORRECTIONS HOLE]
FILLING CHANGES
[DELETIONS J
UNC, ES. F279C. BAKER. MMM. ODD. DAT
4 MINUTE CLEAN RAW DATA
i :
PHYSICAL EXTRACTION
CREATE A NEW FILE IN ADDITION TO DAT
UNC. ES. F279C. BAKER. MMM.DDD. SMTH
SMOOTHED DATA, DERIVATIVES INTEGRALS
}
-------
collection automatically at the appropriate times. This allowed the
operator to have considerable background level information available
when he arrived at the site to make the injections. The operator was
responsible for entering information about the run on a calendar-type
log which served as one type of key to data retrieval since all runs
had the run date as part of their names.
A 12-hour run was divided into four minute intervals. A digital timer
operated Teflon solenoid valves on the inlet of each instrument to
connect the instruments to one chamber half during each four minute
period. The DAS acquired the signal outputs every minute and converted
the voltage readings to physical units using calibration equations
stored in memory. At the end of each four minute period, the valves
were switched and the physical data was written onto the disk system
for subsequent data processing.
Several options were available to the operator concurrent with data
acquisition. He could 1) request a minute by minute data listing to
be generated, 2) request that a line printer type plot be generated of
any data on the disk, including old runs, 3) request that a comment be
entered and stored along with the current time-of-day in a comment
file on the disk, 4) request a listing of system status information,
5) request that data from an instrument not be stored on the disk if
the instrument was defective or being calibrated, and 6) alter the
calibration equations for each substance. Each substance was assigned
a default calibration equation (see Appendix B). Instrument performance
was susceptible to change from one day to the next; altering the cali-
bration equations provided a means of compensating for such a change
rather than having to make daily adjustments to the instrument.
After Run
Because the instrument response time was not immediate at the valve
switching, the readings made in the first minute, and for some instru-
33
-------
ments the second minute, of a four minute cycle could have been based
on a mixture of air from both chambers. In addition, certain instru-
ments provided an update in their readings only once every minute or
once every four minutes. Consequently, the two processing programs
(steps 2 and 3, Figure 7) that operated on these minute-by-minute data
files were designed to use only the one reading out of four (usually
from the third minute of the cycle) that was appropriate for each
substance. The first of these programs (step 2, Figure 7) could gen-
erate three different outputs: 1) a formated, four minute, summary
listing of the data, 2) a line printer plot of the data with a resolu-
tion of 0.8% full scale, and 3) a file in which the data and comments
were combined and stored as though they were punched on cards. The
second processing program (step 3, Figure 7) generated a paper tape
which could be read by a Hewlett-Packard programmable calculator with
plotter (HP). Thus, after each run, plots and listings of all the
substances were immediately available. The strip charts from each
instrument, the one minute listing with operator comments that was
generated during the run, the formated, four minute summary listing,
the paper tape, and the HP plot were filed together in folders and
indexed by run date. These became the basic information source for a
run. The card image file described above was stored on magnetic tape
so that it could be read by an IBM computer for later analysis.
The creation of the card image tape files was delayed until the UNC
disk cartridge could be transferred to another POP 11/40 computer at
EPA which had a magnetic tape unit and a disk cartridge unit among its
peripheral devices (Figure 8). This was usually done after 20 runs
had been accumulated.
Other than the plots and listings, no other data treatment procedures
were normally attempted on the site computer. There were several
reasons for this, among them: limitations on program size, the pre-
sence of extensive statistical packages on other computers, and pri-
marily a shortage of available time when the computer would not be
34
-------
collecting data.
After the tape files were created, the tape was carried to TUCC. This
installation has an IBM S/370 computer with extensive programming
support and statistical packages available. The tape files were input
to a project-generated program called CONVERT which created files on a
project-owned IBM disk pack resident at TUCC (Step 6, Figure 9).
Besides converting from ASCII (the character code used by POP computers)
to EBCDIC (the character code used by IBM computers), CONVERT formated
the data so that each record of the output file contained the last
time of a four minute cycle, the reading collected in that cycle for
each of the 12 substances, the chamber side from which the readings in
that cycle were taken, and the year, month, an^ day of the run. These
files had names composed of the month, year, and day on which the run
was started followed by the extension ".DAT". For purposes of uniform-
ity, each file contained 180 data records, sufficient to hold 12 hours
of data between 0500 and 1700 hours. If a run was less than 12 hours,
a missing data code was inserted in each record where necessary. The
data records were followed by a varying number of comment records
containing the date and time of the comment as well as the comment
itself.
The raw data in these files could contain various errors: the readings
might have come from a calibration period when the suppress option was
forgotten, the sensitivity of an instrument was changed without an
appropriate calibration equation change, or data were missing due to
hardware failure or operation suppression.
To correct these errors, personnel used a project-generated program
called CRCTSCP (Step 8, Figure 9). The .DAT tile was copied to an
online system pack so that it would be accessible under an IBM supplied
interactive timesharing system called TSO. This is labeled in Figure
9 as the "CORCOPY" step. Personnel then entered the TSO system using
a Tektronix 4010 graphics display terminal as the input/output device.
35
-------
Using the 4010 offered several advantages: first, data could be
rapidly plotted on the screen; second, the interactive nature of the
system allowed the user to select a section of the plot and have it
expanded to full screen size so that fine detail could be examined;
third, the 4010 has adjustable crosshairs that allow the user to
select a point on the screen and have its coordinates supplied to the
computer. Thus the data could be graphically edited to correct spurious
values and to remove gaps. Removal of gaps was necessary to make the
data continuous across time so that subsequent numerical calculation
of integrals and derivatives would be possible. Data were plotted on
the screen and the crosshairs were used to edit the data in two ways:
either to supply a specific data value at a certain time, or to supply
the starting and finishing values over a period of time and have the
intervening values computed by the program using linear interpolation.
When all necessary editing had been performed, the edited data file on
the on-line disk was copied over the old .DAT file on the project-
owned, off-line disk. The .DAT extension was retained. The original
raw data (that is, unedited data) remained available on the magnetic
tape. This is labeled in Figure 9 as the "FINCOPY" step. The edited
.DAT files were then passed through a high-frequency cut-off filter
program called SMOOTH. The function of this program was to attenuate
any high frequency noise spikes in the data so that meaningful deriva-
tives could be calculated. The output of this program was 1) a formated
2
listing of the raw data, 2) the r resulting from a moving, least
squares parabola fitted to five sequential points, 3) the value of
the mid-point of the fitted parabola taken as the smooth data, 4) the
derivative of the curve at each point, computed by a least squares,
numerical procedure, based on five points, and 5) the integral of the
curve from the start of data collection to the point under consideration.
In general, the filtering was not necessary for the 0- and NO data,
O J\
but a fair amount of high frequency noise was present in the THC,
NMHC, and CO data. The smoothed data along with the computed deriva-
tives and integrals for each variable were written onto the project-
36
-------
owned disk with a name composed of the month, year, and day, as with
the original raw data file, but with the extension ".SMTH". These
.SMTH files became the final working files for a run.
A project-generated program called SMTHPLT was us j to plot data in
the .SMTH files on a Broomall plotter using the TUCC computer. These
are the plots used as figures in this report. The .SMTH file listings
were examined by project personnel and the actual (rather than the
intended) initial conditions, peak concentrations, maximum rates,
meteorological conditions, and dosages were extracted and key punched
on cards. A total of 135 measures were extracted or computed from each
side of each dual run. These data were used to generate plots for
graphical analysis and as input to statistical analyses for what was
called initial condition analysis. The key punched data were input to
a project-generated PRINT program which listed the data from all runs
using various groupings such as the same substance measured at various
times in a run, and the values of different substances at the same
time in a run. Times of importance included crossover times and times
at which peak concentrations and peak rates were achieved. The key-
punched data were also used as input to the non-linear, least-squares,
multiple regression program (NLREG) available in the TUCC library.
The TUCC installation has several comprehensive statistical packages
available, including the Statistical Analysis System (SAS) and the
Statistical Package for the Social Sciences (SPSS). These programs
were used to create an SPSS data file and an SAS data file from the
cards. SAS was used to generate scattergrams from variable pairs and
to perform selected regression analyses. SPSS analyses included
descriptive statistics (means, standard deviations, etc.) of the
variables, factor analyses, and joint frequency distributions for
groups as well as pairs of variables.
All the runs over the two year data period occupy approximately 8
37
-------
million bytes (characters) of storage on the project disk pack. In
addition, there is one magnetic tape of raw data holding 4.9 million
bytes, and more than 500 computer listings of data in various forms.
STANDARD NOMENCLATURE AND CONVENTIONS
During the grant period a system of nomenclature developed. Thus,
certain terms used in this report have meanings specifically related
to this project. The purpose of this section is to present and define
the most important conventions used in this report. A more complete
list, defining names and abbreviations, is contained in the glossary
and a detailed description of the nomemclature conventions is given in
Appendix A.
During the construction and operation of the smog chamber, conventions
were adopted for naming parts of the facility and for naming experimen-
tal procedures. The chamber itself was divided into 169.9 m (6000
o
ft ) halves, each half was designated as a "side". For purposes of
identification, one side was called the "BLUE" side and the other was
called the "RED" side (see Figure 5). Manifolds and instrument
sample lines were color coded and the color designation was used in
data handling, from acquisition through final analysis. An experiment
in a single chamber side was designated as a "run". A "dual run" was
an experiment using both sides of the chamber to compare the effects
of differences in controlled variables (such as initial concentrations)
under identical solar radiation and temperature conditions. Data was
collected by computer for each run from 0500 to 1700 Eastern Daylight
Time. For purposes of defining starting conditions, however, the
initial time was considered to be the time at which the hydrocarbon
injection was completed.
In analyzing data from the smog chamber it was necessary to establish
names for measures of reactivity and experimental variables which were
33
-------
consistant with requirements of the computer programs used. In general,
these names had to consist of only alphanumeric characters, to begin
with a letter of the alphabet, and to be no longer than eight characters.
The names were composed of the abbreviation for the chemical or physical
quantity to which the name applies, plus prefixes ind suffixes describ-
ing the type of measure and the time of measurement. The types of
measures considered included concentrations ( ) (no special prefix
or suffix), dosage (D ), rate of change of concentration ( R),
differences ( DIP ), maximum concentrations ( M) and averages
( AV ). The times of measurement correspond to various uniquely
defined events in the course of a run. The events which were of
interest included time at which initial conditions were achieved
(I ), crossover (X ), the maximum rate of NC disappearance ( NOR),
the maximum rate of N02 formation ( N02R), the maximum rate of 03
formation ( 03R), the maximum Cu concentration ( 03M), and the end
of run ( FIN). To name measures computed between the above events,
the periods of time between events were given a numeric designation.
The time between achieving initial conditions and crossover is "1",
between crossover and N02R is "2", between N02R and N02M is "3",
between N02M and 03R is "4", and between 03R and 03M is "5". For
example, "HCAV3" is the average nonmethane hydrocarbon concentration
between the times of the maximum rate of NCL formation and the maximum
NO- concentration. "03AV35" is the average ozone concentration between
the times of N02M and the ozone maximum concentration.
Some of the measures used in data analysis are different from the
measures considered when analyzing indoor chamber data. Measures of
reactivity as defined by the Coordinating Research Council project
15
CAPI-6 often are not clearly applicable to outdoor chamber data. An
example of an ambiguous definition is that of NCL formation rate.
CAPI-6 defines NCL formation rate as
Rate = [N0]i / 2T]/2
39
-------
That is, half the initial NO divided by the time to form an amount of
N02 equal to 1/2 the initial NO. The broken line in Figure 10(a)
shows such a rate on data from a typical indoor chamber. The slope of
the N09 concentration line is quite close to the straight-line slope
at any instant. Examining the data from a hypothetical run in an
outdoor smog chamber in Figure 10(b), it will be noted that the
instantaneous rate of NO^ formation differs markedly from the straight-
line slope. This is mainly a result of constantly increasing light
intensity in the morning. The length of the initial period with zero
light intensity is a function of injection time and sun rise, not a
characteristic of system reactivity as would be desirable.
One means of avoiding this is to consider the maximum instantaneous
rate instead of some average rate. The maximum rate is uniquely
defined for nearly all N0? curves. It may be calculated by hand, or a
computer can easily calculate and output the maximum rates. Because
of the difficulties in applying CAPI-6 definitions to outdoor chamber
data, other measures were adopted. The principal measures of reactivity
used in the analysis were the following:
1. Time to crossover, when the NOp concentration equals
the NO concentration (TIMEX).
2. Maximum rate of NOp formation (N02R).
3. Maximum rate of NO disappearance (NOR).
4. Maximum rate of 03 formation (03R).
5. Maximum N02 concentration (N02M).
6. Maximum 0., concentration (03M).
7. Average rates of concentration change between events in
the system, e.g., N02RAV3.
40
-------
'1/2
8
10
Figure 10.
456 7
HOURS
(b)
CAPI-6 definition of NCL formation rate applied to typical
indoor(a) and outdoor(b) smog chamber profiles.
41
-------
EXPERIMENTAL DESIGN
General
There were five factors considered in the experimental design: the
reactant identity, the reactant concentrations, the purpose of the
run, the climatology and the distribution of illumination, and the
time period available. The first three factors were subject to choice;
the last two imposed limitations.
Basic Program Plan
Two different types of hydrocarbons were used. Because of the large
amount of previous smog chamber and modeling work that had been per-
formed with propylene, experiments used to test the chamber, those in-
tended for comparison with previous indoor work, and those to be used
for modeling purposes were designed with propylene as the only NMHC.
To determine the effect of NMHC control on the behavior of NCL, a
hydrocarbon mixture that was reasonably representative of urban condi-
tions was needed. The design and composition of this mixture will be
discussed in detail in a subsequent section.
Since the primary effect under investigation was hydrocarbon reduction,
the primary variable between the two chamber halves was the initial
NMHC concentration, while the INOX concentrations and the percentage
of INOX that was NOg (PCTN02) were to be the same in the two chamber
sides. Runs were always performed in pairs.
A relatively few INOX values were planned while a larger number of IHC
values were to be explored. Although a large range of concentrations
was used, no attempt was made to obtain a balanced design in terms of
number of runs at each concentration because of the variation in
meteorological factors expected (see discussion below). With respect
to the choices for INOX, a large number of runs was planned at a value
42
-------
which, according to Dimitriades1 work ' , would result in a maximum N0?
value that would achieve the annual air quality standard for N02 using
Larsen's technique to transform a one hour average value to a.n annual
mean value17'18.
Fewer runs were planned at INOX values above and below this standard
value with the number of runs at each value decreasing as the value
became further removed from the standard. The range of INOX was
chosen from background { ^ 0.007 ppm) to ^1.0 ppm. The
choice of IHC values was more arbitrary than the INOX; because of the
background air concentrations, it was difficult to do experiments at
or below 0.22 ppmC and no runs were performed above approximately 4.2
ppmC. Within this range IHC were spaced at approximately 0.5 ppmC
intervals.
Some runs were planned in a 2 x 2 pattern. That is, all combinations
of two sides of the chamber and two IHC with constant INOX and PCTN02
were designated, yielding four runs at each IHC level on three different
days. Runs and run pairs were planned in a somewhat random fashion
within sets as were INOX concentrations where possible; runs were not
performed in random order if certain matches of weather conditions
were needed.
In all basic experiments, the initial percentage N0? was chosen to be
10% of the INOX for the propylene runs and 20% of the INOX for the
hydrocarbon mix runs. Some experiments were planned to investigate
the effects of different PCTN02 levels. In these paired runs, the IHC
and INOX was to be the same in each chamber half with the only difference
between sides being the PCTN02 value.
43
-------
Table 3. SIMULATED URBAN HYDROCARBON MIXTURE
Relative Concentrations Mole Fraction
Class/Compound ppm ppmC Total Without
Mix Aromatics
1. Acetylenic
acetylene 265 530 0.131 0.171
Subtotal 265 530 0.131 0.171
2. Paraffins
isopentane 172 860 0.085 0.111
n-pentane 286 1430 0.141 0.184
2-methyl-pentane 85 510 0.042 0.055
2,4-dimethyl pentane 69 483 0.034 0.044
2,2,4-trimethylene pentane 76 608 0.038 0.049
Subtotal 688 3891 0.340 0.444
Average carbon number =5.7
3. Olefins
butene-1 40 160 0.020 0.026
cis-2-butene 43 172 0.021 0.028
2-inethyl-l-butene 26 130 0.013 0.017
2-methyl-2-butene 32 160 0.016 0.021
ethylene 360 720 0.178 0.232
propylene 97 291 0.048 0.062
Subtotal 598 1633 0.296 0.385
Average carbon number =2.7
4. Aromatics
toluene '. 140 980 0.069
m-xylene 90 720 0.044
n-propyl-benzene 60 540 0.030
sec-butyl-benzene 60 600 0.030
l,2,4,trimethyl-benzene 123 1107 0.061
Subtotal 473 3947 0.234
Average carbon number =8.3
TOTAL with aromatics 2024 10,001 1.000
without aromatics 1551 6,054 1.000
44
-------
Urban Hydrocarbon Mixture
The chemical composition of the NMHC had to be reasonably representa-
tive of urban conditions. The Chemistry and Physic Laboratory of EPA
had developed a mix based on field sampling by the ilnvironmental
Protection Agency in Los Angeles and Cincinnati which was selected for
19
use in this study . This mix was a compromise between actual urban
early morning analysis and what could be easily analyzed with a gas
chromatograph. Percentage distributions by mole fraction and average
carbon number were balanced in the makeup of the mix (see Table 3).
As originally proposed the mix consisted of 13% acetylene, 34% paraffins
(average carbon number 5.7), 30% olefins (average carbon number 2.7)
and 23% arornatics (average carbon number 8.3). It was very difficult
to operate with the full aromatic content of the mix because of the
retention of the heavier compounds from one day's run to the next
either in the chamber or the sample lines. Experiments were performed
with only toluene as 23% of the mix, with 23% of the mix as toluene
and m-xylene, and with no aromatics at all. These experiments are
described in the discussion of Chamber Chemical Performance in this
report. Essentially no difference was found in NO, N0?, and 0-
profiles in dual matched runs in which the same NMHC-ppmC concentration
was used but the relative percentage of the toluene was 0% and 23%.
Slight differences were found in dual runs with mix verses mix, toluene,
and m-xylene. The observed differences could have been due to at
least two factors: 1) different reactivity of the aromatic that
replaces some of the paraffins and olefins, and 2) fewer molecules
being present in the mix and aromatic runs, because the same ppmC was
used on each side and the average carbon number of the aromatics was
higher than that of the mix. Because there were no significant differ-
ences in the mix and toluene runs and mix only runs, and because the
mole fraction weighted reactivity of the five aromatic compounds was
very close to that of toluene, the aromatics were omitted from the
full mixture entirely. Therefore, the hydrocarbon mix adopted for
45
-------
Table 4. SELECTED CLIMATOLOGICAL DATA FOR
PIEDMONT NORTH CAROLINA
Maximum Temperature
Average Temperature
Minimum Temperature
Average Dewpoint
Average Wind Speed
Hours Sunshine
% Possible Sunshine0
Sky Cover
Mean
69°F
58°F
47°F
46°F
9.2 mph
7.86 hours
64.2%
0.59
20°-95°F
14°-82°F
0°-72°F
-3°-73°F
3.5-73°F
0-14.2 hrs.
31.5%=SDC
0.34=SDC
aThe (daylight hours which had full direct sun)/(total daylight
hours) x 100%.
Fraction of the total sky hemisphere which had clouds.
Standard deviation.
46
-------
the experimental runs was 17% acetylene, 44% paraffins, and 39% olefins,
as listed in the last column of Table 3, although there were some runs
with aromatics added to the mix. In the latter case, the mixture
composition was as in column 3 of Table 3 with 23% aromatics (i.e..
toluene).
Climatology and Solar Factors
The general climate of piedmont North Carolina is described by
20
Petterssen as a warm temperate, rainy climate, without dry seasons,
and with hot summers. The piedmont has an average rainfall of 45
inches. Selected climatological data are given in Table 4. These
21
data were calculated from data obtained from U,3. Commerce Department
for Tuesdays and Thursdays for 1969-1971 (313 days). From Table 4 one
can see that, although on the average 60% of the sky was covered by
clouds, 64% of the possible sunshine hours were hours of direct, full
sun.
In Table 5, the number of Tuesdays and Thursdays in the period 1969-
1971 which had various percentages of possible direct sunshine values
are broken down by month. This table suggests that the periods of
greatest sunshine would be in the spring and fall and that the middle
summer months would be characterized by cumulus cloud formation that
reduced the total time of direct sunshine. Figures 11 and 12 report
the average daily solar radiation on a horizontal surface in the
22
United States for January and July respectively . These figures show
that, although the chamber site and Los Angeles are approximately on
the same latitude and should receive essentially the same solar radia-
tion, the chamber site receives, on the average, approximately 10-12%
less solar radiation in July and approximately the same solar radiation
in the winter. Considering the information in Table 5, this result is
most probably due to more partly cloudy days in July at the chamber
site while Los Angeles has more clear days. In the winter, both
locations must have approximately the same sky cover.
47
-------
Table 5. NUMBER OF DAYS IN WHICH PERCENTAGE MAXIMUM POSSIBLE
SUNSHINE WAS GREATER THAN OR EQUAL TO A GIVEN VALUE.
TUESDAYS AND THURSDAYS IN 1969-1971 (313 DAYS) BY MONTH.
(number of days)
Percentage Maximum Possible Sunshine
100%
<90%
<75%
<50%
January
February
March
Apri 1
May
June
July
August
September
October
November
December
2
8
6
3
3
0
0
2
2
2
5
2
6
15
12
11
6
3
1
5
8
4
11
7
11
17
13
17
15
9
7
9
15
7
15
12
13
19
19
23
20
19
14
20
21
14
21
19
Total
35
89
147
222
Summation
Percentage
Actual
Percentage
11.2%
11.2%
28.4%
17.2%
47%
18.6%
71%
24%
48
-------
350
Figure 11. Average daily solar radiation on a horizontal surface in the USA for January
-2
The units are cal-cm -day. (Sellers ).
22,
100'
Figure 12. Average daily solar radiation on a horizontal surface in the USA for July. The
units are cal-cnf 2-day. (Sellers )
49
-------
It is possible by making assumptions about the average dust concentra-
tion, water vapor concentration, and ozone concentration aloft, to
calculate the theoretical total solar radiation reaching any point on
the earth's surface. Figures 13 and 14 illustrate the variation in
solar radiation as a function of time and month on a surface normal to
the beam and horizontal to the ground. These data are also corrected
for atmospheric path length absorption. The cosine view factor is the
principal difference between the data in Figure 13 and Figure 14. The
major effect of season on intensity normal to the beam is on duration
of daylight which can vary from a minimum in December of 9 hours, 34
minutes to a maximum in June of 14 hours, 25 minutes. The seasonal
effects on intensity on a horizontal surface include not only the
duration of the daylight but also a variation in maximum daily intensity
-2 -1
which varies from 0.6 to 1.4 cal-cm -min . The data of Figure 14
can be compared with actual solar radiation profiles presented in
subsequent sections of this report. The agreement was found to be
very good for totally clear days (the time axis in Figure 14 is Eastern
Standard Time).
The above discussion would suggest that there would be essentially un-
predictable variations in solar radiation and temperature in any
outdoor experimental design. The obvious method of dealing with these
variations would be to run sufficient replicates at the critical
locations in the experimental design so as to assure a wide range of
solar radiation conditions including totally clear skies. The dual
chamber design would assist in achieving a degree of control over
these variations. Also to the extent possible, runs should be performed
across seasons.
Time Period Available
Weather conditions in piedmont North Carolina do not become satisfactory
for outdoor runs until approximately the third week in April. In this
50
-------
2.00
1.80
1.60
o
1 1.00
a
°, 0.80
£ 0.60
5 o.
-------
period the last freeze, the last high winds, and the last cool days
occur. Piedmont North Carolina usually has a rather long "Indian
summer" in the fall months of late September and October, sometimes
extending through the second week of November. These periods are
characterized by cool to cold nights and totally clear days with
temperatures reaching 30°C (85°F). Thus, the period from May through
early November (approximately six and one half months) is available
for outdoor runs. Under ideal conditions, approximately twenty dual
runs per month could be performed, giving an upper IJmit of 260 runs
or 130 dual runs per season. Unexpected maintenance or bad weather
can reduce this number substantially.
52
-------
SECTION VI
RESULTS AND DISCUSSION
PART I: FACILITY PERFORMANCE
Because of the uniqueness of this chamber, extensive testing was per-
formed to assess its operating characteristics. These tests were to
determine if the chamber design was physically sound and durable, and
if it exhibited reasonable performance characteristics with respect to
generally accepted smog chamber measures. The actual testing con-
sisted of performing standard runs, comparing results with other
chambers, demonstrating that generally accepted photochemical mod-
eling approaches could be applied to the chamber results, and deter-
mining if the chamber facility produced data that was consistent and
interpretable, even though it was sometimes subject to fluctuations in
physical factors such as solar radiation, temperature, and relative
humidity.
CHAMBER PHYSICAL PERFORMANCE
Design
The wooden A-frame and Teflon wall chamber was extremely durable. It
withstood high winds, snow, freezing rain, and below freezing tempera-
tures without any obvious weakness or stress. One panel had to be re-
placed in April 1974 after a tornado passed through the southern pied-
mont area of North Carolina; this replacement was accomplished in two
days. The actual chamber volume and surface area that resulted and
other physical and chemical data of interest are given in Table 6.
Of concern in any study of this nature are the mixing characteristics
53
-------
Table 6. ACTUAL PHYSICAL AND CHEMICAL CHARACTERISTICS
OF THE CHAMBER
Item
Dimensions
Volume
Surface Area
Surface/Volume
Moles Air @25°C (77°F)
Volume of gas from 1% tank
to yield 1 ppm in chamber
@ 25°C
Volume increase caused by 14°C
temperature increase (base
temperature - 25°C)
Exhaust and filling rate
Residence time on exhausting
Time to 99.6% turnover
Sampling manifold flow rate
Sampling manifold resident time
Injection manifold resident time
Mixing time in chamber
Sampling instrument consumption
of volume for 12 hours operation
Data for each side of chamber
9.14 m W x 6.10 m L x m L x 6.10 m W
A-Frame (30 ft.) x (20 ft.) x (20 ft.)
156 m3 (5509 ft3)
204.4 m2 (2200 ft2)
1.31 m"1 (0.40 ft."1)
6380 Moles
15.6 i
4.49 m3
(4.7% total volume)
7208 i/min (255 ft3-™™"1)
21.6 mins
2.0 hrs
60 £/nrin
<14 sec
<5 sec
<3 min to stable reading
1588 a (1.0% total volume)
54
-------
of the chamber. This pertains to the rapid ability to achieve a
steady concentration immediately after injection and ability to
maintain a homogeneous mixture in the chamber for the duration of the
run. Mixing in each chamber side was accomplished in three ways: two
mixing fans, small constant movement of the Teflon walls from outside
breezes, and thermal currents from slight temperature gradients which
developed from floor to peak of chamber. No discernible changes in
concentrations could be observed in either NO, NCL, 0~ or NMHC when
the mixing fans were turned on for one hour, off for two hours, and
then on once more. Injection information over the past two years has
consistantly shown that no more than 2 1/2 minutes were required to
achieve a constant concentration (within the precision of the NO
J\
analyzer +_ 0.002 ppm) after the completion of the injection process.
In order to insure that the sample air from the center of the chamber
was representative of other air in various parts of the chamber, the
sample and return manifolds were switched during an experiment. This
permitted sampling from the corner of the chamber (0.5 m from the
floor) under one of the mixing fans; again no difference could be seen
in the measurements just prior to and after the change in sampling
location.
Air Handling Systems
Tracer experiments using ozone demonstrated the chamber could be
purged to 0.1% of its original concentration with clean background air
in approximately 1.5 hours. The decay profile followed that of an
exponential dilution system. The purge rate was adjusted downward
however, to prevent excessive stress on the Teflon walls and floor.
This meant that a minimum of 3 hours of chamber purging was required
before the startup of a new experiment.
Flow measurements taken inside the sampling manifold with a hot wire
anemometer were 60 1/min; with an average manifold length of 9.14 m,
thus less than 14 seconds were required for air in the chamber to
55
-------
reach the instruments. Since the chamber air on summer days was
warmer than the manifolds inside the laboratory, heating tapes had to
be installed on the entire manifold length to prevent condensation.
Chamber Tightness
3
It was impossible to expect that a 12,000 ft Teflon film structure
would be entirely leak free. The rate of dilution by exchange with
outside air tended to be small according to CO tracer experiments.
Material expanded out of the chamber when the temperature was rising
and this did not result in any change in mixing ratio. When the
temperature decreased, the volume decreased and outside air was drawn
into the chamber, which caused dilution. Generally, air temperatures
tended to rise throughout the course of the experiment, causing little
dilution by this mechanism. Another mechanism for exchange was the
compression and expansion of the film due to wind gusts which "pumped"
the chamber walls. This was the predominant mechanism for exchange,
since the looseness in the walls absorbed the small volume changes due
to solar radiation changes which caused air heating and cooling. The
overall mass loss during the course of a 12 hour run was never more
than 12%, according to CO data mostly on windy days (see ozone and NO
decay tests in a later discussion for other species loss rates). No
correction was made for any mass loss by exchange.
Solar Radiation, Temperature, and Humidity
It has been stressed throughout this report that natural outdoor
conditions made these smog experiments unique and more realistic.
Materials were carefully selected to insure that these natural condi-
tions would exist inside the chamber. Upon completion of the chamber,
various tests were run. A comparison of the light in the chamber with
that outside showed that approximately 85% of the total solar and UV
radiation was transmitted inward through the Teflon. These data were
56
-------
Table 7. SPECIFIC PHOTOLYSIS RATE FOR NITROGEN DIOXIDE
IN OUTDOOR SMOG CHAMBER DURING SEPTEMBER, 1973.
Incident total solar radiation, $ka inside chamber
-2 -1
cal-cm -min 2 ft. above floor,
min"
0.63 0.22
0.78 0.28
1.09 0.37
recorded by Eppley total solar radiation and UV sensors which do not
measure through a full 360° volume, but only 180°. Further comparisons
were made by measuring the specific rate of photodissociation ($k ) of
N02 (in nitrogen in Teflon bags) inside the chamber. These measurements
were made approximately 2 feet above chamber floor and are reported in
Table 7. At all of the outdoor solar conditions in which these com-
_o
parisons were performed (solar radiation between 0.7 and 1.1 cal-cm -
min" outside of chamber), the $k for N00 was always higher inside
a <-
the chamber than outside. The reflective aluminum floor was responsible
for this difference.
In September 1974, a continuous quartz globe N09 actinometer (QGA) was
23
designed and built by Sickles and Jeffries . This device is described
in Appendix E. Two days of comparison data for TSR by Eppley sensor
and QGA were generated and are reported in Figures 15 and 16. Essen-
tially the same ratio of k /TSR (that is 0.36 to 0.37 min'Vcal-cnf 2-
-1
min ) was found with the QGA as had been determined in 1973 with N02
bag experiments inside the chamber.
In addition to raising the light flux (and thereby compensating for
the light losses through the Teflon), the reflective aluminum floor
helped prevent large temperature increases inside the chamber. Before
57
-------
mi ns
-i
CM
v£>
CO
co
LO
r~
LO
r-
CD
Qd
LJ
CO
LU
I—
Q_
UJ
CO
oo
CO
LO
o
CO
en oj
OJ LO
OJ «-«
vD
r-
o
o
o
o
n i r
I I I I I I I
1,1,1,1,1,1
CD
CO
LO
CO
OJ
UJ
CO
CC
CO
OCDOO
O
•
OJ
O CD CD O CD
CD CD CD CD CD
c
0
to
•1—
•a
s-
S-
(0
r™~
o
to
,_~.
r0
O
•P
-P
SZ
CU
-a
•r—
O
c
•r™
-a
c
oj
cu
-o
X
o
•^™
-o
c
cu
en
o
i-
-p
•r—
c
S-
o
M-
cu
rcJ
S-
to
to
5^
r~"
o
+J
o
a-
o
•r—
<*-
•1—
o
0)
a.
oo
T-,
r^
o
oo
•
^J-
r^
CTt
A
00
r—
t.
cu
JD
E
0)
Q.
cu
oo
o
14—
O)
•r-
-p
tl_
o
£=
o
•r-
4^
u
c:
3
£
•p
r^
3
to
•1—
cu
c:
r-
•o
cu
c~
in
03
-o
•»
o
•r*
-p
OJ
•1—
-a
ro
S-
S-
, —
o
t/)
r*-
fa
o
-p
to
•r—
cu
c.
•r—
1 —
.
r—
cu
s-
-------
mins
-i
r-
01
CD
UJ
en
z:
UJ
\-
Q_
UJ
CO
_Lj_Lj_Lj_L_L
CO
OJ
CO
r-
LO
O
O
OJ
O
CO
CD
=1-
CD
OJ
O
O
O
CO
CD
CD
=J-
CD
OJ
CD
CD
CD CD. CD
yuios
1
r^
Q
UJ
•
CO
a:
ID
O
1C
c
o
1 \
+-*
CO
•i—
-o
CO
s-
i.
cO
'o
a>
o
•»->
c
s-
o
4-
0)
xJ
T1*
CO
S-
(/)
•r—
to
>>
o
+j
o
_c
a.
o
-a
•r™"
"o
oo
•d-
r-.
01
Ol
s_
CD
J3
E
CU
-(->
Q.
a>
oo
s_
o
«*-
0)
,. —
-(->
*«-
o
c
o
+J
o
c
3
<*-
CO
to
CO
C
o
+->
CO
-o
cO
-!->
O)
"o
•r™
>
CO
S-
cu
rrt
iO
O
t/)
r— •
•^
«t-
^.^
c
o
-M
CO
•o
cO
4->
CU
"o
•i—
>
CO
S-
-»->
r— •
3
(/)
•r-
(L>
C
*^"
TD
OJ
JZ
to
CO
T3
• A
C
O
.f—
•M
CO
•r-
•^
CO
t-
S-
cO
o
to
i_—
CO
_L \
T-^
0
1 <
•P"'
•
p~"
^
O)
CO
<
-o
c
CO
X — ^
c
•r—
E
CM
r— ~
CO
U
ro
i
O
r—
X
LT)
O
a) -a
Q. c
00 CO
0)
CT)
cu
c
59
-------
the aluminum floor was installed, a 2Q°F air temperature rise in the
chamber occurred on sunny days with outside temperatures of 85° to
90°F. With the aluminum floor, the air temperature inside the chamber
was only 4.5° to 5.6°C (8-10°F) higher than the outside ambient
temperature with full summertime solar radiation conditions. Typical
air temperature profiles for several days are shown in Figure 17.
No attempt was made to control the relative humidity in the chamber.
During the evening and early morning purging procedure (1700 to 0500
hours), condensation would sometimes form on the inside walls of the
chamber. In experiments run on mornings after this had happened it
could take until approximately 0900 hours for this condensation to
evaporate. Typical dew point profiles are shown in Figure 18.
INSTRUMENT PERFORMANCE
Ozone Meter
The chemiluminescent ozone meter used in this study has performed
without any failure for the past 24 months. The Teflon particulate
filter in the sample line had to be replaced every 2 months and the
rotometer setting on the sample and ethylene adjusted weekly to the
correct value. The zero and span drifts over the course of a week
were usually less than 3 ppb and 20 ppb on the 1 ppm scale. The ozone
calibration was studied in detail and this study is reported in Appendix
D. This study placed a 95% confidence limit on the overall calibration
of the ozone meter of less than + 0.008 ppm at the NAAQS of 0.08 ppm.
Oxides of Nitrogen Meter
The chemiluminescent NO-NO monitor did not have major maintenance or
X
mechanical problems. During calibration, the efficiency of the N0?
converter was always greater than 99%. After 2 months of constant
use, the vacuum pump was routinely rebuilt, the 03 generator cleaned
60
-------
100.
90.
80.
u- 70.
*
« 60.
Of
50.
cc
ce
30.
20.
10.
0.
' I ' I ' I ' \ ' I ' I ' I '
I . I
I . I . I , I . I . I . I . I . I
100.
SO.
80.
10.
60.
SO.
40.
30.
20.
10.
5 6 7 8 9 10 U 12 13 14 15 16 17
HOURS, EOT
Figure 17. Selected air temperature profiles in outdoor smog chamber. Curve 1. June 9, 1974;
curve 2. Hay 14, 1974; curve 3, October 12, 1974.
O
Q_
100.
90.
80.
70.
60.
50.
10.
30.
20.
10.
0.
Figure 18.
T ' I ' I ' I ' I ' I ' I ' I T I
I ' .
i I i I i I . i . I . I . I . I . I . i
I
100.
90.
80.
70.
60.
SO.
40.
30.
20.
10.
Selected
curve 2,
7 8 9 10 11 12 13 14 15 16 17
HOURS. EOT
dewpoint profiles in outdoor smog chamber. Curve 1, October 4, 1973;
Hay 16, 1974.
61
-------
and the Teflon filter in the sample line replaced. Because of long
continuous operation the NCL catalyst exhibited "converter fatigue".
This problem manifested itself in a decaying response to high N0~
concentrations when the meter was used to measure short term high N0?
(^4 ppm) concentrations. The problem was solved by operating the NCL
reactor at 325°C while purging with helium overnight and then returning
the reactor temperature to 265°C.
Two very important interference characteristics were exhibited by the
NO monitor used in this study. Initially, this instrument was operated
24
with a gold wire catalyst to reduce N02 to NO. It was observed that
in propylene runs which generated high oxidant (0.2 to 1 ppm), an
increasing positive interference to N0? coincided with oxidant buildup;
also inordinately high NO concentrations were observed at the end of
these smog runs. Neither interference was observed in specially
prepared clean air systems of O^/NOp/NO. Generally, it was found that
until the NOp peak was reached, NOp concentrations reported by the
chemiluminescent meter were in good agreement with Saltzman N0?
bubbler measurements. The interference became most pronounced after
the N0? peak was reached and oxidant formation began to exceed 0.2
ppm.
A new carbon catalyst, furnished by the manufacturer, was installed in
the NO instrument in November 1973. Subsequent smog profiles from
X
the chamber still exhibited' high NO values after the peak, but they
J\
did not exhibit the high residual of NO that previously occurred in
instrument readings from the spent smog system. Nevertheless, it is
strongly suspected that although NO levels in the middle of the run
are in closer agreement with those predicted by photostationary equili-
25
brium for coexisting NO, N02, and 03 concentrations , near the end of
the run there is a minimum concentration (0.01-0.03 ppm) below which the
instrument will not read correctly. Therefore it is difficult to make
photostationary equilibrium calculations near the end of the run when
NO concentrations should be approximately .003 ppm. The N0~ interfer-
62
-------
ence in the NO mode of the chemiluminescent analyzer was probably
A
caused by reduction of organic and inorganic nitrogen compounds such
as PAN and nitric acid (HONCL) to nitric oxide in the NCL converter.
The catalytic reactor was run at lower temperatures to reduce the
possibility of reducing interfering compounds. Lower converter
temperatures (^200°C) did not yield a linear NCL conversion efficiency.
Shortly after these tests had been conducted at the UNC chamber, two
pi-
other smog chamber groups at the EPA Smog Chamber Conference ,
described more detailed experiments relating to the interference
characteristics of the Bendix NO analyzer. The Riverside group
A
presented data to show essentially a 100% response to PAN and methyl
nitrate, but they found an erratic response to HON09. This data was
27
subsequently published. The Battelle group presented data which
indicated a 100% response to HON09, but they found an erratic response
28
to PAN . It was not clear how the experiments were performed or if the
usual Teflon in-line sample filter was used in either experiment.
It is clear from NO? data produced by the UNC chemiluminescent NO
C. X
analyzer that either the UNC analyzer was not responding with 100%
efficiency to both HONOp and PAN or that one of these compounds was
not present in the sampling lines, because in high concentration (4-5
ppmC) propylene runs, 60% of the initial NO was not measured near the
A
end of the run (see Figure 19). The Battelle group has developed a
technique for measuring HON09 based on its selective removal by fine
29
nylon fibers. A propylene/MO experiment was conducted using two
A
Bendix chemiluminescent NO analyzers. The project analyzer was
A
operated in its normal manner with a Teflon in-line sample filter; the
second NO analyzer had a nylon fiber HON09 scrubber installed in
J\ £
front of the Teflon in-line filter so that it would not respond to
HONO?. Manual Saltzman bubbler data were also taken. The results of
this test run are shown in Figure li/.
All three measurements agreed well until 30-60 minutes after the
63
-------
CD
LO
in
LO
CO
CD
00
LO
C\J
CD
CM
LO
LO
CD
CD
CD
CD CD CD O O CD
CD
s-
O)
N
c
0
1 —
Q
1 1 i
_
CO
>,
CO
X
o
2t
•"•s^.
0)
c
cu
r—
>>
o.
o
S-
Q.
rO
C
•r~
CO
•o
O
_C
-p
cu
E
-P
c
cu
E
cu
S-
, —
(Q
C
fO
X
CD
z
X
•I—
•o
c:
cu
ca
i
OJ
cu
>
s_
3
0
•
s_
cu
-p
•r™
M-
c
o
<0
s-
-p
c
cu
o
c.
o
0
o
:n
2:
•
CO
•p
c
cu
E
cu
3
CO
fl3
CU
5-
CU
JD
ja
3
JD
ID
CD Q
CO CO M-
(V* "O >
CU i—
|4- «
«4- C
•r- tO
LO
CD
LO
LO
O
LO
CO
CD
CO
LO
OJ
CD
OJ
LO
CD CD
uudd
CD
T—I
CD
LO
CD
CD
CD
CD
CD
N
-P
to
CU
CU
cu
X
•P •!-
•o
<4- C
o cu
ea
c
O I
CO
r- n-
fO O)
a_ >
E S-
O 3
c_> <_>
0)
JQ
X 3
O
CO
ro
o
C
O O
">> Q.
C Q.
<2ON 'ON
64
-------
peak. This was the time of maximum formation of PAN and HONCL. The
two chemiluminescent instruments agreed almost perfectly the entire
run, indicating that the project analyzer was not measuring HONCL.
Subsequent measurements of MONO,, using the Battelle method have shown
that HONCL was present in the sampling manifold at approximately the
expected concentrations during the propylene/NO <._Aperiments. We have
A
concluded that the particular sampling arrangement probably influences
whether the Bendix NO analyzer can measure HON09, and that in our
X L.
arrangement, the analyzer does not detect HON02, but does respond to
PAN, probably at 100% efficiency. Therefore, N02 data from the project
NO analyzer after the N02 peak was actually N02 + PAN.
Since PAN does not begin to form in appreciable quantities until after
the N00 has reached its peak (or more properly, the N0/N0? ratio has
become very small), the N02 peak measurements reported in this study
are valid measurements of the N02 maximum concentration. This is
supported by Saltzman bubbler data taken on a number of runs at various
concentrations (an example is shown in Figure 19). In propylene
experiments, where the most interference would be expected, the Saltzman
and chemiluminescent N02 data agreed until 30-60 minutes after the N02
peak. The mix contained only a small amount of compounds capable of
forming the peroxyacyl radical needed for PAN formation, so therefore
the interference in mix experiments would have been very small. In
5
the auto exhaust experiments of Dimitriades , the PAN formation ranged
from .001 to .08 ppm over a hydrocarbon range of . 1 to 4.8 ppmC.
Saltzman bubbler data in the mix runs were also in good agreement with
the NO analyzer near the peak.
/\
Only runs employing the hydrocarbon mix were used in assessing the
effects of hydrocarbon reduction on N0? concentrations. No corrections
were applied to the N0? data in these experiments. Nitrogen dioxide
maximum concentrations were conside^d to be valid based on the studies
described above. Nitrogen dioxide dosages in the high hydrocarbon
65
-------
runs could have been high by 3-5% due to PAN interference in the
afternoon.
Runs employing propylene as the hydrocarbon were used for testing and
modeling purposes. The NCL data in these runs were also not corrected
for the PAN interference and in the figures in this report which show
propylene data, the N02 data after the peak should be taken as NO,,
plus PAN.
ENVIRONMENTAL CHROMATOGRAPH
The environmental chromatograph required more attention than the other
instruments. Columns were replaced or reactivated monthly. Daily
calibrations were required; the reproducibility on a 3 ppm calibration
source was - 3%. Thus, for a single measurement, it was possible for
the CH. to be low by 0.05 ppmC and the total hydrocarbon (THC) high by
0.05 ppmC. For this reason, the NMHC curves (difference between THC
and CH.) for a given run appear to be more noisy than the 0.,, NO, or
N02 profiles.
The environmental chromatograph is not subject to interference, per
se, but can exhibit different responses to different hydrocarbon
compounds. The Federal Register method indicates that both the CH.
and THC modes of the instrument be calibrated with CH. gas mixtures,
as in Figure 20a. This calibration method will work only when the
detector response to CH. and to higher hydrocarbon compounds is the
same; that is, each carbon giving the same increment in signal whether
it is CH. (increment of 1 unit) or the carbons in C^Hg (increment of 2
units) or C3Hg (increment of 3 units). This was found not to be true
for this environmental chromatograph.
When both modes, THC and CH., were calibrated with gas mixtures contain-
ing only CH., as in Figure 20a, the response to a nonmethane hydrocarbon,
such as butane, hexane, or toluene (when present alone or in mixtures
66
-------
oo
z
O
a.
oo
UJ
CXL
THC
CH4
a) THC and CH. outputs c?,ibrated
with CH. only
TIME
oo
o
a.
00
LU
a:
THC
CH NMHC=0.62 TRUE NMHC
LU
t/0
O
a.
oo
LU
a:
C
THC
NMHC =0.62 TRUE NMHC
CH.
4
TIME
b) NMHC standard only
THC
TRUE THC
RESPONSE
r
'€
>ading „**
X
X
(
1
>
^e
V
ading
f t -
T
CH4
TIME
c) CH4 + NMHC standard
TRUE NMHC = (NMHC reading)/eff factor
TIME
d) calibration or efficiency factor applied
to (THC - CH4) reading
Figure 20. Calibration and correction of environmental chromatograph.
67
-------
with CH4) was only 62% of the expected value, as in Figure 20b. When
NMHC was added to CH4 mixtures, the THC mode responded 100% to the CH^
and only 62% to the NMHC as in Figure 20c. The response factor to
NMHC could be improved by increasing the hydrogen to air ratio in the
detector, but this was at a sacrifice to sensitivity and precision.
The response difference was only between CH, and hydrocarbon compounds
with more than one carbon, that is, CH» was detected with one efficiency
and other hydrocarbon compounds were detected with another efficiency.
It is not unusual for the first member of a homologous series to be
different from higher members. Since CH, was measured by itself, it
was possible to correct the NMHC after calibrating both modes with CH,
only, as shown in Figure 20d.
Each week an average efficiency factor was determined by measuring the
response to individual known concentrations of butane, hexane, and
toluene. This efficiency was always 62% - 2% of the response to CH,
only. As the data were acquired during an experiment, the difference
between the THC and the Cti. readings was computed. This gave a value
for NMHC that was 62% of the true concentration. It was therefore
multiplied by 1.6129 to arrive at the true NMHC concentration. These
calculations were performed by the computer for each reading. This
method is the only way valid data can be obtained from instruments
which exhibit the above characteristics because the THC data is a
mixture of material with 100% response and materials with 62% response.
To determine if the response characteristics described above were
peculiar to the environmental chromatograph used in this study, two
other instruments , one produced by a different manufacturer (Beckman),
were tested. The Beckman chromatograph had a 74% response to NMHC
when both the THC and CH, modes were calibrated with CH». The other
instrument also had a 60-64% response to NMHC. We would therefore,
caution other users of environmental chromatographs to determine the
actual response efficiency to NMHC rather than assume it is the same
as CH4.
68
-------
GAS CHROMATOGRAPH
A gas chromatographic system was designed and developed by William
Lonneman of the U.S. Environmental Protection Agency using a Perkin
Elmer 900. The original intent of this system was to obtain a qualita-
tive spectrum analysis of the hydrocarbon mixture in the chamber after
the completion of the injection process. Chamber air was introduced
into a 25 cc sample loop and then concentrated in an associated liquid
nitrogen freeze out trap. Hydrocarbon separation and detection were
accomplished with a 200 ft. S.C.O.T. column and a flame ionization
detector. The analysis was temperature programmed from 30°C to 110°C
at 16°/min, with an initial time of 2 minutes and a final temperature
of 110°C for 10 minutes.
This system did not permit the separation of C,-C3 compounds and 2-
methyl-butene-1 from 2-methyl-butene-2. Although this method was used
as a qualitative tool, it was possible to obtain some quantitative
information on the compounds above by use of hexane as an internal standard.
INJECTION SYSTEM
Of the many different methods initially considered to introduce pollu-
tants into each half of the 312 m smog chamber, concentrated
commercially prepared tank and associated valving proved to be most
satisfactory. Nitric oxide and NOp injections were accomplished in
less than one minute with an additional 8 to 12 minutes required to
bring these pollutants up to the desired concentrations. Injection
rates from the concentrated NO and N02 tanks did not vary by more than
+_ 5% from day to day.
Vapor pressure considerations placed an upper limit on the individual
concentrations in the hydrocarbon mix tank(s). The original mix was
prepared in two cylinders. Experience in preparation by the manufactur-
69
-------
er and storage and injection at the UNC smog chamber permitted a
subsequent blend of the same mix to be prepared in one cylinder. A
comparison of typical relative concentrations at the beginning of runs
on July 28, 1973 (two cylinder injection) and September 18, 1974 (one
cylinder injection) is given in Table 8 and a typical chromatogram is
shown in Figure 21. These data demonstrated the ability to inject the
same general distribution of hydrocarbon species over the course of
two smog data periods and at different initial hydrocarbon concentra-
tions. Although there were differences in the actual butene concentra-
tions between the two injections (tanks), the total butenes were 8.89
and 8.66% carbon. The chromatograms in Figure 21 show that the areas
for these peaks are quite small and measurement errors were possible.
CHAMBER CHEMICAL PERFORMANCE
Natural Background Air
The natural background air used to purge the chamber was not entirely
free from NMHC and oxides of nitrogen. In the spring and early summer
months the background NMHC ranged from 0.10 to 0.30 ppmC. During the
fall this background could increase from 0.30 to 0.55 ppmC. Oxides of
nitrogen in the spring were usually <0.01 ppm. Most of the time this
NO background was equally divided between NO and N09. During a two
X <—
week inversion period in July 1974, values as high as 0.035 ppm NO
A
were measured in the early morning air. To minimize the effects of
this background, experiments with high IHC and INOX were performed on
these days. Carbon monoxide background varied between 0.3 and 0.6
ppm. Early morning ambient 0, concentrations were usually below 0.02
ppm with 1 hour daytime maximums falling between 0.03 and 0.05 ppm. A
few times during the summer ambient 03 maximum concentrations were
above 0.08 ppm.
To test the oxidant potential of the background in the chamber, experi-
ments were performed by closing the chamber with only background air
inside. On April 30, 1974, the maximum 1 hour ambient ozone measured
70
-------
outside the chamber was 0.082 ppm. At this time, 1440 EOT, the oxidant
inside the chamber was 0.137 ppm. Data for NMHC end NO were not
y\
available on this day. A similar run on July 7, 1974, produced 0.167
ppm in the chamber while ambient ozone just outside the chamber was
0.066 ppm. During the course of this experiment, NO in the chamber
J\
increased from 0.009 ppm to 0.020 ppm (NO was 0.09 ppm and N02 was
0.011 ppm), and NMHC increased from 0.1 to 0.2 ppmC. This increase in
NO and NMHC was apparently due to the degassing of materials from the
X
Teflon film after an experiment the previous day with mix and NOV.
yv
Ozone Decay in Background Air
The average nighttime rate of destruction of 0, in rural North Carolina
30
air is approximately 0.11 ppb/min . UNC chamber tests with starting
03 concentrations of 0.85 and 0.54 ppm in background air show daytime
destruction rates of approximately 0.33 to 0.50 ppb/min depending upon
the time of day. The average nighttime rate (initial Og = 0.6 ppm)
was approximately 0.13 ppb/min, very close to the rate reported for
rural air.
Assuming that the photolysis of 0, produces 14% 0 D and 86% 03P
-2
oxygen atoms at a specific rate of 5.1 x 10 times the N09 photolysis
1 <-
specific rate ($ka), and that the 0 D oxygen atoms react primarily
3
with water to yield 22% hydroxyl radicals and 89% 0 P atomic oxygen,
then only 3% of the ozone photolyzed is lost by this route because the
3
primary fate of 0 P oxygen atoms is to react with 0^ to form Q3-
Using a maximum value of 0.55 min for $ka, the net Oo loss rate due
to photolysis would be a maximum of 0.7 ppb/min. For the data in
Table 9, the maximum $ka for N02 would be approximately Q.28 min"
using the relationship for $ka and solar radia^'on given in Figure 16
and the maximum solar radiation for November given in Figure 14. A $ka of
0.28 min" corresponds to an 0- instantaneous photolysis rate of 12.4
ppb/min at [0.,] = 0.85 ppm. If the net loss of Oo by photolysis is
71
-------
Table 8. RELATIVE DISTRIBUTION OF INDIVIDUAL COMPOUNDS OF
THE HYDROCARBON MIX IN TANKS AND IN CHAMBER AIR
AFTER INJECTION3
(percent carbon)
Compound
acetylene
ethylene
propylene
butene-1
cis-2-butene
2-methyl -butene-1
2-methyl -butene-2
isopentane
n-pentane
2-methyl -pentane
2,4-dimethyl-pentane
i methyl-
specified
HC
mix
3.
3.
2.
3
57
81
91
57
TOTAL
19.20
31.96
11.41
10.73
13.65
100.81
Two tank
injection of
July 28. 1973
4.01
2.01
21.20
30.60
13.18
12.32
13.75
One tank
injection of
Sept. 18. 1974
2.93
3.20
18.66
37.60
12.00
9.60
13.50
99.94
100.02
aData for injections based on area under peak from chromatograms gene-
rated by freeze-out of multiple 25 cc air samples and analysis on
Perkin-Elmer 900 Gas chroma tograph with 200' SCOT column coated with
OV-101. See Figure 21 for chromatograph example.
Methane from background air and acetylene, ethylene, propylene from
mix were inseparable under the conditions of analysis, therefore the
relative distribution was computed omitting these components.
Total non-methane hydrocarbon concentration was 0.7 ppmC injected from
two tanks each contain only part of the mixture.
Total non-methane hydrocarbon concentration was 2.0 ppmC injected from
one tank containing the entire mixture.
eThese compounds were not resolvable from each other.
Total butenes in each injection was 8.89% and 8.66% for two tank and one
tank injections.
72
-------
tO
fO
T3
s_
re
0)
.a
E
re
^:
o
o
2-auaq.nq-sp
l-auainq
O
O
-o
e
(O
re
o
.c
o
re c
O) O
r— O
CL
-------
only 3% of the instantaneous rate as described above, the 0^ disappear-
ance rate would be 0.4 ppb/min. Assuming that the nighttime rate of
03 loss in the chamber (0.13 ppb/min) was due to thermal reactions
with the walls and other materials and that it was of the same magnitude
during the daytime, then the net theoretical loss rate of 0-, at 1300
J
hours (solar noon) would be 0.53 ppb/min, in very close agreement with
the actual rate measured (0.533 ppb/min). This rate excludes any
reactions of 03 with photochemically generated reactants.
The 03 loss rate in the chamber decreases with decreasing light intensity.
Of course, any 0^ generation due to photochemical processes would
decrease the net 03 loss rate. The 0, data in Table 9 agree well
enough with theory, that any 0^ generation would have to be less than
0.1 ppb/min. This rate of 0^ generation would create 0.06 ppm 0^ in
the chamber in 10 hours which is approximately what occurred when
background air was irradiated in the chamber. The nighttime 03 loss
rates yield a half-life of 48 to 70 hours and the daytime loss rates
yield a half-life of approximately 18 hours.
Oxidation of NO in Background Air
Nitric oxide oxidation tests were conducted by injecting 90% NO and
10% NO- into the chamber with background air. During the daytime the
average rate of NO disappearance was approximately 0.2 to 0.3 ppb/min
for starting NO concentrations ranging from 0.45 to 0.65 ppm. A
typical NO/Background dual run is shown in Figure 22. The diurnal
trends for two NO oxidation runs are given in Table 10. On November
3, 1973 an initial rate of 0.09 ppb/min was computed for the early
morning (the theoretical thermal rate would be 0.06 ppb/min); this was
followed by a midday rate of 0.31 ppb/min (thermal rate = 0.051 ppb/min)
and an afternoon value of 0.24 ppb/min (thermal rate = 0.040 ppb/min),
thus these rates appear to be a function of light intensity. The
rates of N0? formation, also included in Table 10, show that the NO
decay rate is almost twice the NO,, formation rate, indicating that the
74
-------
cc.
CO
Ol
CO
O
Q.
a.
ID O i—
«D «=i- co
«3- CO i—
O O O
vo 10 co LO
o co «3- o
10 LO co i—
O O CD O
O
C£
CD
ii
O
<
CO
cs
Q
LU
(U
o:
CO
O
Q.
a.
CO •— r—
co f^ co
LO CO r—
O O O
r-v r^. co er>
r-v co r^ CM
LO LO CM r—
CD O CD CD
r-x LO
O CM
CM CO
O CD
rj
UJ
CQ
e^
=c
0
C£
O
O
Q
h-
r>
o
A
cu
CD
c
5
o
co
0
-O
« 1 1 1
00
f"s^
•
E 0
Q.
Q.
CU
Z3 4_>
i— S-
CQ ft]
4.)
00
LO
00
*3-
CO
o
o
0
•
o
(—
00
r-
•
0
co
LO
•
o
o
<£>
O
VO
*
0
cr»
f***.
•
o
^t*
o
^o
00
*
o
LD
co
f*^
•
0
fN^
Ol
0V
1^
*
o
CO
o^
^o
*
o
«^~
LO
oo
p»x
*
o
LO
CM
l£>
1C
•
O
1^
co
CFi
vo
.
o
o
LU
Q
o
o
u_
o
oo
T3
« C
CD •> UJ
cn E
c= a.
ra a.
x:
O Q -(->
UJ t-
CO C£ 03
O 4->
OO
co CM r^
^O r— CO
CO 00 r—
f^ ID LD
CD O CD
«d- CO CM
r— U3 i—
LD CO CO
co r-- LO
O O CD
«* o r- •*
CO «3- O CM
CTl CO O IX)
r-.. r~. r- ^o
CD CD CD CD
VO
«*
co
cr»
o
o
00
CD CD CD CD
IO CO
r^. r~-
cr> LD
*a- «d-
CD CD
CM
VO
CD CD
. CT» CO
r-i— i— CM
I I I I
r— r— r— r—
O O O O
CO LD (-^ O\
co ^*
CO LD
r*~ r*"
I I
CO CO
LO I—
co co
CM i—
co
r--
O)
co
co
co
r^.
cr>
CU
o
O
cu
o
o
cu
o
75
-------
OO)OOr»«jOlOst*OQCM«-iO
r-tOOOOOOOOOO
I l I ' T' I ' I ' I ' I
0}
T-*
to
LU
z
ID
10
i — en
CO
Of
o
._ o>
1,1,1.1.1.1.1.1,1
j
iO
10
cr>
oo
CM
uJdd *' '
'ON
l/>
c
0
•r~
+->
TD
C
O
o
r~—
-------
o
~^**
ID
z
1 — 1
C£.
O <
1— Q
< Z
o; o
o o:
LJ- CD
^
0 <
2: CQ
CO
Q CD r~-
z.z. en
00 ID
OO CO
O C£
_l UJ
CO >
o s: o
Z < 2=
u. o
o
C£
oo o
LU O
\— 0
•=C (—
C£ O
O
.
O
1 —
QJ
JD
rD
h-
O> C
fO £
cc \
CMQ
O Q.
Z CL
O) sr
-P •!—
ro E
o: \
_a
o a.
Z Q.
f-T
ro
C S-
o a»
•1- -p
-P C
i-
•p ^~
c o
ai
U -P
c s-
Orrt
ro
O -P
OO
CM
0 -P E
•z. ra Q.
a.
,_r
ro
>
s»_
E O)
O -P
•p- C
•P ^-~<
ro
S- 4-
-P O
c~
CO -P
0 S-
0 -P
O 00
E
o -P a.
2: ro a.
LO LO
O^ ^S" P* ~" **& c i
0 r— r- 0 r—
• • • • *
O O O O O
+ + ~h ~f" -f"
O 1^ LO LO CTl
CT) LO O *^t* O*l
O i — co oj i —
• • » • *
O O 0 O O
1 1 1 1 1
l£> CO i — O CM «3"
LO i-x en i — CM i r~~
o o o r— i — i cr»
o o o o o i
r\
r—
~>~.
ro
S
•=3- co co co co
LO CO i — UD CM 1
«=!• ^j- ^f co co i
O CD O O O
OOLO<£>OUDOCOO i
CvJLOi— <-DO.
o
r^.
o
CM
r—
1
O
o
o
1 —
^-~
LO
•d-
1—
1
f>v^
o
CM
1 —
LO
LO
f — s
1—
1
LO
^J-
i —
LO
If;
r--
i —
CO
CD
ro
r
"
;>
<;
oooooooooo
oooooooooo
•^-LOLDr-^cocnoi— CMCO ai
r— i— i— i— i— i— CMCMCMCM CT>
I I I I I I I I I I ro
CMOOOOOOOOO S-
LOOOOOOOOOO CO
CM^fLOLDr-COCTlOr— CM >
r— r— r— t— •— r— r— CMCMCMea;
77
-------
NOp was being converted into other nitrogen compounds, probably HONOp,
by either OH, 0, or 03-
Side-to-Side Performance
Before useful comparisons could be made between sides, it was important
to demonstrate that each side could perform almost identically with
the same initial chemical conditions. Figures 22 to 25 show the side-
to-side behavior at various IHC and INOX. As can be seen, almost no
difference exists between the chamber halves over the variety of
initial conditions tested. Figures 23 and 24 are propylene runs and
Figure 25 is a mix run.
Day-to-Day Comparisons
The effect of some possible day-to-day variation in physical conditions
on the same chemical system at two different hydrocarbon levels (4 and
2 ppmC) and 0.5 ppm initial NO was illustrated by the propylene
X
experiments of October 15 and 16, 1973. (Figures 26 and 27). The
initial conditions on October 15 were the same as in the runs on
October 16 except that the chamber halves were used in reverse. The
solar radiation on these two days is given in Figure 28. There was
more ozone formed in the high IHC side on October 16 than the high IHC
side on October 15. This corresponded, however, to a higher morning
temperature (the rate of reaction between NO and 03 is slightly tempera-
ture dependent) and also to choppiness in solar radiation (see detailed
discussion of the effects of temperature and solar radiation in subse-
quent sections).
After 1230 hours, the solar radiation on October 16 decreased signifi-
cantly due to cumulus cloud formation. The high IHC side, which had
almost twice as much propylene as the low IHC side on October 16, was
essentially unaffected by the reduction in solar radiation; its NO
had converted to N0? and the maximum 0^ had already been reached
78
-------
before the solar radiation reduction. In contrast, the low IHC side
on October 16 was only at the initial stage of ozone generation at
1230 hours. Hence, the peak 03 concentration for this side was less
than would normally be expected on this day (see Figures 24 and 26)
and corresponds to the drop off in solar radiation.
Comparison Experiments with Aromatics
Because there were certain questions as to the use of aromatics in the
mix, experiments were designed to show the importance of including or
excluding aromatics on the overall hydrocarbon mix behavior. Table 11
summarizes some of the critical data from these comparison experiments.
The NO, N02 and 03 profiles are given in Figures 29, 30, 31 and 32.
As long as the same carbon concentration was injected in both sides
and toluene did not make up more than 27 to 40% of the total mixture,
little difference in the NO, NO- and 0, behavior could be observed.
31
According to the Jackson reactivity scale, the UNC mix has a mole
fraction weighted reactivity of 2.61 and toluene alone has a reactivity
of 2.2. When injections of equal ppmC were used, the side with toluene
had 12 to 18% fewer hydrocarbon molecules initially available for
reaction. Also the Jackson molar reactivity of the mix plus 40%
toluene by carbon is slightly less than that of the mix without toluene.
Hence, it is difficult to completely explain the excellent side-to-
side agreement between the mix runs with and without toluene that were
conducted on the same day with the same initial carbon concentrations.
The mix/toluene/m-xylene experiments on September 13, 1974, and Sep-
tember 14, 1974 (Figures 31 and 32), on the other hand, did indicate
that if m-xylene was substituted for a portion of the toluene, the
smog reactions proceeded at a slightly faster rate. The shift in
rate, however, was small and would not greatly affect the trends
presented in this report.
79
-------
e
o.
o.
I ' I ' I ' I ' I ' I ' I '
OCTOBER 12. 1973
5 6 7 8 9 10 11 12 13 H 15 16 17
HOURS. EOT
Figure 23. Example of reproducibility between two chamber halves for propylene. Initial conditions:
NOX(—) 0.503, (—) 0.503 ppm; N02 (—) 0.052. (—) 0.054 ppm; NHHC (—) 4.05,
(—} 4.09 ppmC propylene In UNC outdoor smog chamber.
E
D.
O.
1.0
0.9
0.8
0.7
0.6
0.5
O.f
0.3
0.2
0.1
0.0
Figure 24.
1 I ' I ' I '
1 1 ' I ' l
JULY 5.
8
1.0
0.9
0.8
0.7
0.6
0.5
O.f
0.3
0.2
0.1
H 15 -16 17
0.0
9 10 11 12 13
HOURS. EOT
Example of reproduceblUty between two chamber halves for urban hydrocarbon mix.
Initial conditions: NOX(—) 0.500. (—) 0.500 ppm; N02 (—) 0.120. (—) 0.122 ppm;
NMHC (—) 0.877. (--) 0.876 ppmC urban mix 1n UNC outdoor smog chamber.
-------
1 —
o
III
*
CO
cr
ZD
O
n:
.
X
'i
c
0
.a
s-
fO
o
0
S-
•o
>>
--* —
c
r—
(C
_c
s_
OJ
jQ
(O
^:
o
o
+j
c
r1™
a.
E
n3
X
a.
a.
o
r*^.
0
CD
x — •,
1
I
i
«• — f
t\
<\J
r^x
o
d
i*^.
1
^
CM
O
•z.
• rt
E
ex
ex
LO
LO
CO
0
^-^
1
1
1
A
CO
LO
CO
•
o
JL
X
o
. .
10
c
o
•r*
4->
•r*
-a
c
o
o
1 — •
*XJ
+J
•r-*
C
i.
0)
JO
E
(O
.c
o
CD
O
tO
S-
o
0
T3
+J
3
O
0
•z^
rj
c
X
•1—
E
C
o
JD
s-
rO
O
O
S-
-a
>,
^
c
(O
.a
s-
3
C_)
ex
ex
-------
E
Q.
O.
o
CD"
10 11 12
HOURS. EOT
Figure 26. Effect of hydrocarbon reduction and solar radiation. Initial conditions: NOX(—) 0.505,
(—) 0.512 ppm; M>2 (—) 0.058, (—) 0.064 ppm; NMHC (—) 1.98, (—) 3.88 ppnC
propylene In UNC outdoor smog chamber.
E
o.
o.
•o
1.0
0.9
0.8
0.7
0.6
0.5
1 0.^
i 0.3
0.2
0.1
0.0
Figure 27.
111111
10 11 12 13 H
HOURS. EOT
15 16 17
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
Effect of hydrocarbon and solar radiation. Initial conditions: NO (—) 0.527, (—)
0.528 ppm; N02(—) 0.059. (—) 0.059 ppm; NMHC (—) 3.89. (—) 2.05 ppmC propylene
in UNC outdoor smog chamber.
82
-------
o
•
C\J
CO
CD CO UO sf OJ CD
--I CD CD CD T- CD
O
•
C\J
CO
o
C\J
T_u.iuJ-2_ujo
CO
•
O
UD
OJ
oo
O
Do
83
-------
oo
o
i— i
i
s.
cu
tl»*
t^_
•o
o
vi c
•)->
c
o>
£T
o
0
+J
I/)
O
E
<0
CM
co
s: E
CO Q,
O Q.
0
o
y i^j
CM E
O Q-
'ZZ Q,
^_
r—
•
l/>
CU
r—
•^
«t-
o
s_
CL
C
•r-
a>
o
c
OJ
s_
UD
co
o
•
o
U7
r—
i^
•
CU
o
c
cu
s-
cu
1-
M-
•1—
TD
O
C.
T^-
CM
CM
•
O
in
^_
i — •
•
w
cu
^~
•r—
«4-
O
S-
CL
c:
"r~
LO
^^
CM
•
O
^.
1—
p-.
*
a>
O)
^>
jg^
4-*
*r-
00
(O
«*-
to
o
CO
^J-
•
CM
co
•
£
•o
0)
o
(O
"a.
•r-
-O
to
a>
•i-j
3
C
•r—
E
in
^_
N
C
(!)
C
cu
£
-C
+->
•r~
3:
CU
-o
•1—
in
o
co
00
CTl
0
•
O
(/)
r—
(O
1 *
23
o
5^
CU
5
O
r—
to
CM
CO
CO
CTl
O
•
O
1—4
^.
^^
^
§
O
u_
C£.
Q.
X
1 — 1
:>
r—
i—
CU
r—
(D
^
C
0
(/)
,r-
s-
n3
Q.
0
cu
•o
C/)
o
4J
ai
•r—
00
•^.^
C
cu
r—
>^
>^
|
2T
CU
c
cu
3
r—
O
I—
„
(O
X
•r—
2:
X
o
1 — 1
n
O
T~
1— 1
C
o
JQ
£_
03
O
^S
o
•
o
c
o
JD
(O
O
^•^
o
CJ
^
O
E
Q.
a.
•cj"
CO
•
o
t
0
E
a.
a.
CM
•
o
E
CL
a.
^-
00
•
o
CO
to
cu
4->
(T3
Q
r—
•^.
>~r
o
•
o
CM
•
I""**
OJ
c^f"
^o
•
o
o
o
CM
•
CO
CO
*
o
CO
^0
r—
^•s^
r—
r^
O 0
• •
O 0
o r-»»
• •
o c^
CO
GO r- ,
CVJ ^f
• •
CM r—
CO CTl
0 0
CM CM
• «
CO «d-
cvj co
• •
CM CVJ
CO CO
O^ CTl
f__ p«.
**^,i "v^
i-™ p—
r— r—
r— O
• ,
co o
^~
r>. CM
1 — . ^O
i— co
CM CM
CTl CTl
• •
r_ r-*
00 CO
00 CTl
• «
O O
^h ^f
00 O
• •
CM CO
^ ^
CO CO
r— r—
*>^ ^^
CTl CTl
O
•
O
o
0
00
CTl
•
^~
IT)
UO
r~~
•
00
•
r—
*
^r
^~
^^
CTl
ir>
•
«^-
CM
O
0
co
•
t—
co
10
•
r—
tn
•
•—
*
^
r—
^x^
CTl
>v
C
O
to
•r-
M—
CU
r*1™
o
-o
rtj
to
cr
•r-
<4-
fO
i.
fO
D_
(Q
84
-------
LT)
O
O
c?o
CD
OJ
CD
CD
CD
CD
CO
cn
T-H
v£)
r-4
LU
CO
UJ
O
Q
1_U
ID
O
to
o
•o
C
O
O
03
OJ
c
0)
-o
c
to
00 £
ra
s-
(C
Q.
O
D.
Q.
00
O
o
Q.
Q.
CO
O
O
CO
O
O
X 3
CM r—
O O
O)
"o
X
•i—
s-
OJ
o
cn
o
to
s-
o
o
-a
3
O
O
•^
c.
•r~
X
'i
o
o.
Q-
00
O
Q.
Q.
to r-
C O
•i- CM
C
o
(O
o
-p
c
o cu
o
T cu
CM
o o c
CM T-
C . T3
O O 3
to i—
•r- ^T» O
i- C
LO
O
O
X X
LO
o
=t-
CD
CO
CD
C\'
o
CD
CD
ujdd -B0 ''ON 'ON
cn
CM
cu
3
cn
85
-------
o
86
-------
0.5 r
E
o.
o.
o
c*
o
0.1
0.3
0.2 -
0.1 -
NO
i • i ' r' i • i • i ' i • i '
SEPTEMBER 13._JL971__
0.5
0.1
0.3
0.2
0.1
0.0
5 6 7 8 9 10 11 12 13 11 15 .16 17
HOURS. EOT
Figure 31. Comparison of mix and toluene with mix,toluene,and m-xylene. Initial conditions;
NOX(—) 0.198, ( —) 0.188 ppm, N02(—} 0.057, {—) 0.054 ppm; NMHC(—) 3.04 ppmC
mix including 36.2 percent carbon as toluene, (—) 2.84 ppmC mix including 17.7
percent carbon as toluene and 13.1 percent carbon m-xylene in the UNC outdoor chamber.
0.5
o.
a.
o
N
O
z
o
0.3
0.2
0.1
1 I ' I ' I ' I ' I ' I ' I '
SEPTEMBER 11. 1971 _
NO
0.5
0.1
0.3
0.2
0.1
Q Q I^Tf^jy I I ! I 1 r-'-r 1 I I 1 I I 1 1 1 1 1 1 1 1 1 0 Q
'5 6 7 8 9 10 11 12 13 .11 15 .16 17
r.jURS. EOT
Figure 32. Comparison of mix and mix with m-xylene. Initial conditions: NO (—) 0.150, (—)
0.150 ppm; N02(—) 0.032. {---) 0.034 ppm; NMHC(—) 1.98 ppmC mix, (—) 1.51 ppmC
mix including 24.5 percent m-xylene in UNC outdoor smog chamber.
87
-------
Comparison Experiments Between Propylene and Mix
To assess the relative reactivity of the mix and propylene, two dual
runs in which the mix was used as the NMHC in one chamber side and the
same ppmC of propylene was used as the NMHC in the other side were
run. In one dual run, 0.5 ppm NO and 4.05 ppmC NMHC was used in both
.A
sides and in the other run 0.21 ppm NO and 1.02 ppmC NMHC was used in
.A
both sides (see Figure 33 and 34).
In the propylene and mix sides of the 0.51 ppm NO runs, the time to
.A
N02M was 224 mins and 376 mins and the N02M values were .406 ppm and
0.365 ppm. The times to 03M were 328 mins and 560 mins and the values
were 0.888 ppm and 0.274 ppm. The maximum N02 rates were 3.83 ppb/min
and 1.70 ppb/min and the 03 maximum rates were 12.87 ppb/min and 1.36
ppb/min for propylene and mix.
Thus, propylene was 225% faster than the mix in forming N02, 946%
faster in forming 03 and propylene formed 11% more N02 and 224% more
ozone. Overall, it appeared that the reactivity of the mix, relative
to propylene (5.5) on the Jackson scale, was 2.4.
COMPARISON WITH OTHER CHAMBERS: MODELING
The only meaningful technique to compare the performance of the UNC
outdoor chamber with that of artificially illuminated (indoor) chambers
was by means of photochemical kinetics modeling. We felt that if a
chemical mechanism based on fundamental chemical principles, could be
developed, which offered a reasonable explanation of the chemical
events in the UNC chamber, it could be used to explore the differences
in results between the UNC chamber and other chambers that are more
familiar to smog chamber experimenters and modelers. This model would
be specific for propylene and would use "generally acceptable" reactions
and rate constants. It would be "validated" against several propylene
runs from the UNC chamber. This in itself would lead to an increased
88
-------
E
a.
a.
o
ft
o
o
TOBER 27. 1973 _
10 11 12
HOURS. EOT
13
Figure 33.
Comparison of mix and propylene as nonmethane hydrocssbon. Initial conditions: NOX(
— )
0.505, ( — ) 0.513 ppm; N02(— ) 0.052. ( — } 0.057 ppm; NHHC(— ) 4,03 ppmC mix. ( — )
4.05 ppmC propylene in UNC outdoor song chamber.
o.
a.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0
I ' I
I ' I ' I ' I '
OCTOBER 4. 1974 _
8
13 H 15 16 17
9 10 11 12
HOURS. EOT
Figure 34. Comparison of mix and propylene as nonmethane hydrocarbon. Initial conditions:
NOX(—) 0.208. (—) 0.204 ppm; N02(—) 0.043,(—) 0.042 ppm; NMHC(—) 0.93 pproC
mix, (---) 1.02 ppmC propylene 1n UNC outdoor smog chamber.
0.5
0.4
0.4
0.3
0.3
0.3
0.2
0.1
0.1
0.0
0.0
89
-------
understanding of the events taking place in the chamber. The same
model would then be used to model runs from several indoor constant
light intensity chambers. If the agreement was reasonably good in the
models of these indoor chambers, then we could feel reasonably assured
that nothing anomalous was happening in the UNC chamber. If the
models of the indoor chambers were poor, then no real conclusion could
be drawn, since we would not know whether it was the model or some
factor in one or more of the chambers that was responsible for the
poor results.
Propylene was chosen as the test hydrocarbon because it is the simplest
hydrocarbon that exhibits essentially all the characteristics of
photochemical smog and it is the most studied hydrocarbon in smog
chambers. Propylene smog chamber data have been obtained from the EPA
Cincinnati smog chamber, the Lockheed smog chamber, and the Statewide
Air Pollution Research Center (SAPRC) smog chamber at Riverside,
California. Auto exhaust smog chamber data from the BOM study of
Dimitriades were also obtained for comparison with mix runs. All of
these have been replotted on the same time scale axis (and where
feasible on the same concentration axis); runs from the UNC outdoor
chamber and from the modeling results have also been plotted on this
scale to facilitate comparison.
Photochemical Model for Propylene
A thirty-five to fifty-five step mechanism (model) for the photochemis-
try of propylene (C^Hg), NO and N02 was developed by Jeffries, Sickles
and Gantner at UNC. The derivation and details of this mechanism are
given in Appendix C. A modified version of the kinetic simulation
program (which "solves" the chemical mechanism using the Gear differen-
tial equation solving subroutine) developed by Overton and Dodge at
EPA was used to simulate the behavior of various photochemical mechan-
isms. The modifications were to increase the utility of the program
by allowing for varying light intensity, for any reaction's rate
90
-------
constant to be changed at selected times during the run to simulate
temperature changes or flow systems, for example, and for improved
analysis of the contribution of each reaction in *he output. The
model used in this section was the ninth version written and tested.
It had been reduced to the most important elements necessary to give
an adequate representation of the events in several actual runs.
Selection of rate constants is discussed in Appendix C. In general,
either values recommended or adopted by the National Bureau of
32 33 34
Standards ' or by Demerjian, Kerr, and Calvert were used; except
for reactions which were suspected of having major heterogeneous
character these rate constants were adjusted only within the range of
uncertainity recommended by the above investigators.
$ka for Nitrogen Dioxide
To model the outdoor runs, it was necessary to specify actinic light
intensity as a function of time. The continuous quartz chemical acti-
nometer for NCL (see Appendix D) was used to generate two days of $ka (k,)
data which was compared to standard solar radiation sensor data (Eppley
total solar radiation sensor (TSR) and Eppley ultraviolet radiation
sensor (UV) data). These data were presented in Figures 16 and 17
earlier in this report. It appeared reasonable to assume that there
was a general linear relationship between TSR and k-, and that the
1 _i
slope of this relationship would be approximately 0.3575 min /cal-
-2 -1
cm -min . The data also suggested that this slope might vary some-
what from day-to-day depending upon upper atmospheric conditions of
water vapor content, CL concentration, and thin, high cirrus clouds.
For purposes of the calculations in this section however, the numerical
relationship stated above was used to scale TSR readings for the
individual day that was to be modeled and this k, profile was then
used as the NCL-specific rate in the modeling program. Other light-
dependent rate constants were expressed as sub-multiples of k,.
91
-------
Model Validation
The dual matched propylene run of May 25, 1974, was one of the valida-
tion runs. The actual profiles of NO, NCL, 0-, and TSR and the model
L- *3
profiles are given in Figures 35, 36, and 37. The agreement among
actual and model profiles was quite good except for the more rapid
conversion of NO to NOp by the model from the point of NO-NOp crossover
to the NOp peak. This acceleration of conversion led to an increased
03 production in the model in contrast with the actual data, but the
later stages of 0, formation were in closer agreement. It is not
known if this was caused by some unknown physical behavior in the
chamber (such as light intensity differences between the chamber and
the model), or by some misrepresentation of the chemistry in the
model.
It was not necessary to include the small variations in the actual TSR
profile in the model because NO was very low at the time of occurrence
and the duration of the cloud cover was not sufficient to allow 03/N02
dark phase reactions to predominate.
A second day, May 7, 1974, with different initial conditions, and its
model are given in Figures 38, 39 and 40. There were significant
variations in the TSR for this day which occurred at critical times in
the course of the run; it was necessary to approximate these variations
in the model k, values as shown in Figure 38. The model profiles show
rapid response to these abrupt light intensity changes which are dis-
cernible in the actual data profiles for 03 even after the high-fre-
quency filtering that occurred in data processing. The actual NO and
N09 profiles do not appear to exhibit these rapid changes. The NO
c. X
analyzer required a 30 second integration period to take a sample and
each side was only sampled every eight minutes. It is quite possible
that the combination of integration and low sampling rate missed these
events.
92
-------
The model of May 7, 1974, suffers from the same type of deviation from
the experimental NO, N02 and 03 profiles that characterized the May 25,
1974 model. In addition, the fact that the N(L readings of the chemi-
luminescent NO analyzer also included PAN in the actual data is
/\
clearly shown by comparison of the post-peak N0? model with actual
data profiles for both of the runs modeled. The higher propylene
concentration of May 25, 1974 led to a more rapid consumption of the
NO^ which, in turn led to a longer flat portion on the N0~ post-peak
profile due to the quasi-conservative nature of N09 + PAN and the NO
L. A
chemiluminescent meter response.
Other outdoor runs were modeled with approximately the same degree of
success. The k,/TSR relationship used for the May 7 and 25 runs was
too large for October 1974 runs, but was somewhat more satisfactory
for October 1973 runs. This was apparently due to increased haze and
high cirrus clouds that occurred more frequently during October 1974;
no k, data was taken during this period.
The propylene model can apparently provide an adequate representation
of the most significant chemical events that occurred in the outdoor
chamber propylene runs and can thus serve as a tool for exploring some
of the behavioral characteristics of the outdoor chamber in relation
to the other indoor chamber work with propylene. Before examining
these indoor chamber runs and their models, some of the detailed
characteristics of the outdoor chamber model should be discussed.
Relative Rate of Outdoor Chamber Reactions
Relative to indoor smog chamber profiles, the UNC outdoor chamber
appears to be slow. That is, the time required to convert the initial
NO to N02 appears to be long. Some of this long time interval can be
accounted for by the fact that the initial light intensity is zero,
followed by a sine-wave-type increase. Figures 41 and 42 show the
behavior of the same propylene model and initial conditions used in
93
-------
I ' I ' I ' I ' I ' I ' I
VRRIfiBLE LIGHT INTENSITY
01234-56
HOURS
Figure 35. Variable light intensity model of May 25, 1974. Initial conditions:
N02 0.045 ppm, NMHC 1.10 ppmC propylene. The k, profile was scaled
radiation shown in Figure 37.
10
NOX 0.345
from total
PP».
solar
E
Q.
Q.
o
o
1.0
0.9
10 11 12
HOURS. EOT
13 H 15 16 17
Figure 36.
Dual matched propylene run in UNC outdoor smog chamber. Initial conditions: NO (—)
0.351, (—) 0.348 ppm; N02(—} 0.044, (—) 0.047 ppm; NMHC (—) 1.10, (—) 1.10
ppmC propylene.
' 94
-------
2.80
2.40
7c 2.00
i 1.60
3 1.20
0.80
£
0.00
1 I ' I ' I ' I ' I ' I ' I ' I ' I ' I '
MflY 25. 1974 •
5 6 7 8 9 10 11 12 13 14 15 16 17
HOURS. EOT
Figure 37. Actual solar radiation for May 25, 1974.
2.80
2.40
2.00
1.60
1.20
0.80
0.40
0.00
95
-------
I ' I ' I • I ' I
M(?DEL OF MflY 7, 1971
°-00 .::--«*
HOURS
Figure 38. Variable light Intensity model of May 7, 1974. Initial conditions: NO 0.502
ppro, N02 0.049 ppm, NMHC 2.09 pproC propylene. The k, profile was scaled for
total solar radiation shown in Figure 40.
0.00
E
o.
o.
O
c5
o
1 I ' I ' .
MRY 7. 1971 _
8
13 H 15 .16
Figure 39.
9 10 11 12
HOURS. EOT
Dual matched propylene run in UNC outdoor smog chamber. Initial conditions: NOXC
0.503, (---) 0.503 ppm; N02(—) 0.048 (—) 0.047 ppm; NMHC(—) 2.09, (—) 2.05
propylene in UNC outdoor smog chamber.
-)
ppmC
96
-------
2.80
2.40
rc 2.00
7f
i
3 1.20
o
i 0.80
a:
«E
o 0.40
0.00
I ' I
I ' I ' I '
MflY 7. 1974
I I I . I I
6 7 8 9 10 11 12 13 . H 15 . 16 17
HOURS. EOT
Figure 40. Actual solar radiation for Hay 7, 1974.
2.80
2.fO
2.00
1.60
1.20
0.80
0.40
0.00
97
-------
the May 25 simulation but under a) constant light intensity conditions
typical of indoor chambers, and b) an increasing ramp of light intensity
such that the average k, = 0.300 min .
The constant light intensity model (Figure 41) still required over 1.3
hours to reach NO-NQ2 crossover and 2 hours to reach the N02 maximum
concentration, but these times are considerably shorter that those in
the model of May 25. The final 03 concentration is somewhat lower
than the variable light intensity model of May 25 (Figure 35). The
ramping k, simulation required 4 hours to reach crossover, longer than
the model of May 25, but the final 0- was approximately the same as
the constant k, model, even though the average value for k, was less.
The long times reflect the slow initiation processes in the propylene
model that was validated on the outdoor chamber runs. It was necessary
in modeling the UNC chamber to set the rate constant for Reaction 25
(see Appendix D) at the upper limit of the homogeneous rate suggested
32
by Garvin , a very low value.
25: N02 + NO H£° 2HN02 K25 = 3.2 x 10"5ppm"1 min'1
rate = K[N02] [NO] [H20]2 ; K = 9 x 10"14 ; [H20] = 18,000 ppm
This reaction apparently has a major heterogeneous component, and its
rate constant is used by most modelers to adjust the initial speed of
the photochemical system based on the individual surface characteristics
of each chamber. The photolysis of HN02 provides an early source of
OH radicals which can initiate the OH/C0HC/H00/NO/OH chain. Without
JO d.
this source of OH, initiation in the model must depend upon products
of the Oo + CgHg reaction or upon the photolysis of aldehydes. The 03
+ CoHg reaction is slow initially because of the high NO to N02
ratio. Aldehydes are not present initially in the smog systems modeled
but do increase as the reaction proceeds.
98
-------
1.00
0.90
0.80
6 0.70
o.
tt 0.60
1 0.50
i O.fO
c5 0.30
0.20
0.10
0.00
T
T
T
T
CONSTflNT LIGHT INTENSITY MODEL -
10
1.00
0.90
0.80
0.70
0.60 r"
0.50 |
10
Q.W ~
0.30
0.20
0.10
0.00
Figure 41. Constant light intensity model. Same chemical conditions as Figure 35. Constant k,
* 0.38 min"1.
Figure 42. Ramping light intensity model. Same chemical conditions as Figures 35 and 41.
The average k « 0.30 m1n .
99
-------
1 I I I I
FflST HNO, REflCTION MODEL
0.00
Figure 43. Fast HN02 reaction model. Same chemical conditions as 35,41,42.
Constant k,» 0.38 mln .
FflST HNO, REflCTION MODEL
0.00
Figure. 44. Fast HN00 reaction model. Same chemical conditions as 35, 41-43.
' i
fnno ^•.H.4' 1< V ft ^ft _1 H *
Constant k1 " 0.30 min
100
-------
To examine the effect o1 using rates for Reaction 25 that have been
commonly used in indoor smog chamber simulations, the model used to
produce the profiles in Figure 41 was re-run with a rate constant for
-2 -1 -1
Reaction 25 of 1.5 x 10 ppm min , the value used by Hecht, Seinfeld,
35
and Dodge in modeling the EPA chamber data and it includes both the
heterogeneous and homogeneous reactions. The results are shown in
-1 -1
Figures 43 and 44 for k, = 0.3835 min ' and k, = J.300 min . Other
conditions were the same as in Figures 35, 41, and 42.
The reaction systems in these simulations are much faster than those
using the rate constant that fits the UNC outdoor chamber. The higher
light intensity (Figure 43) produced a higher ozone concentration.
2 3
The UNC outdoor chamber had a surface to volume ratio of 1.31 m/m
(0.40 ft2/ft3) which was less than that in the EPA chamber (1 ft2/ft3)
modeled by Hecht, et al., but this does not seem sufficient to account
o/r
for the differences above. In the comparisons performed by Lockheed ,
Teflon surfaces exhibited the slowest behavior of the four surfaces
tested (Teflon, glass, aluminum, and stainless steel) and perhaps the
fact that more than 98% of the UNC chamber's surface was Teflon could
partially explain why it was slow relative to indoor chambers.
There was ample water available for reaction since at the cool morning
starting conditions, the dewpoint temperature often approached the air
temperature, and frequently surface water was available from condensa-
tion. Perhaps the HNO- formed in this condensed water, but did not
evaporate back into the gas phase until later in the reaction sequence
when its impact would not be noticeable. The NO/Background test runs
however, showed no significant loss of NOV to the walls at night nor
X
did relative humidity show any significant correlation across a wide
range of experimental conditions.
Comparison of Outdoor Chambers with Indoor Chambers
Smog chamber data for propylene from three indoor smog chambers were
101
-------
collected and replotted on the same scale and time axis as the UNC
outdoor chamber profiles and the previous model profiles presented
above. Dr. Marcia Dodge (Chemistry and Physics Laboratory, USEPA,
Research Triangle Park, N.C.) kindly supplied the numerical data for
35
EPA run 325 which was published by Hecht, et al. Data from the
SAPRC Smog Chamber operated by Pitts at Riverside and the Lockheed
Smog Chamber were digitized from 8.5 x 11 in. plots in publications by
Hecht37 and by Jaffe38.
In the simulations of these indoor chambers, presented below, only the
light intensity (k,) and the rate of Reaction 25 were changed from the
rate constant values used in the modeling of UNC chamber runs of May 7
and 25, 1974. The rate of Reaction 25 was used to account for the
individual surface characteristics of each chamber. The potential
surface chemistry of NpOc was ignored.
Model of EPA-325
Nitric oxide, N02> and 03 profiles of EPA-325 and the UNC model of
this run are given in Figures 45 and 46. The k, value used in this
simulation was 0.4 min" and the rate constant for Reaction 25 was 1.5
-2 -1 -1
x 10 ppm min . These values were the same as those used by Hecht
et al. in their model of this run. Although Hecht, et al. stated that
the potassium iodide oxidant data, which was 50% higher than the Mast
data, was probably more accurate than the Mast oxidant data, both
their model and the UNC model predicted the Mast data quite reasonably.
This suggests that the potassium iodide data had a positive interference
that was not accounted for in the corrections applied by the chamber
operators.
The UNC model of EPA-325 has the same deviation of N02> NO, and 03
after N0-N0? crossover that was exhibited by the models of UNC runs on
May 7 and 25, 1974. The slowness of the UNC chamber makes this
102
-------
deviation more obvious than the much faster indoor chamber runs. The
fact that the models of both chambers show this difference implies
that the difficulties are within the chemical representation in the
model and not in some unrecognized physical behavior in the chambers.
One possible cause of the problem discussed above may be Reaction 27.
32 -1
Recent work has suggested a much higher rate constant (100-400 ppm
min~ ) for Reaction 27 than was used in this work. This rate constant
is still subject to review and validation by NBS.
H02 + N02 -> HN02 + 02 (27)
A large rate constant for Reaction 27 followed by
HN02 ^U NO + OH (30)
would compete effectively with
H02 + NO -»• N02 + OH (3)
since the chain carrier, OH, is regenerated in each sequence. Larger
values for Reaction 27 would bring the model N02 into closer agreement
near the peak NO,,. Large ratios of k27 to k3 however, can make it
difficult to sustain large ratios of N02 to NO needed for high 0.,
production after the N02 peak. Also a large rate constant for Reaction
27 followed by Reaction 30 would make the N02 maximum concentration
light intensity dependent. Further research will be required to
clarify the role of Reaction 27.
Models of SAPRC Chamber
The UNC model of run EC-17 from the Riverside Chamber again shows re-
markable agreement for NO, N02, 0,, PAN, and C3Hg profiles at a very
low concentration of starting reactants. The original data for alde-
103
-------
CD O
O CO
« •
CM r-H
pprnc
CD O O CD O CD
CM CD CO vO st~ CM
• t • » • •
*—« T-I O CD CD CD
O
o
' I '
in
- CVJ
en
or
Q_
- UJ
' '
I ' I ' I ' I ' I ' I
CO
01
r>
o
OJ
o
Q.
Q.
o
o
O
Q.
Q.
CO
c
o
-o
c
o
o
(O
1,1,1,1,1,1,1,
CD
O
O
O)
•
0
CD CD O
co r^* *-Q
• • •
CD CD CD
uudd
CD
LO
CD
CO
CD
CD
CD O CD O
Q.
O
Q.
a. in
cr>
(O r—
-P T3
O C
-------
ppfflc
CD
o
OJ
CD
CO
CD
OJ
CD
CD
O
CO
•
O
CD
OJ
CD
CD
O
O
CD
CD
CT)
O
CD
O
r-
CD
O
LO
O
=}-
O
CO
o
C\J
CD CD CD O O CD
ujdd 'eg 'SON 'ON
CD
O
•
O
105
-------
PflN. ppm
O CO vD
CM «-H »-<
O O O
OJ
O
CO
O
OJ
o
o
o
O
(O
0
-
". 1 , 1 , 1 , 1 , 1 , I
^ CO vO sf OJ 0 CO
"\J T— 1 T— 4 V— t «— 1 T-l O
^« — 1
o
-
•I.I.I."
J —
o
to
O)
• pw
g
Q.
1
tJ
*•=
vO sfr- CN1 O r^
0 O O O **
. • O)
OOCOOCDOC3CD
uidd '.eg («ON 'ON
o
CD
106
-------
PRN, HN03. ppm
CD
CO
CD
00
CD
CD
VJD
CD
CD
«*-
CD
CO
CD
CD
CD
•
CD
CD
CVJ
CD
OO
,—r
CD
r^.
i
3
S-
co
i!
o
CO
OJ
3
CD
^~
«—i
CD
CO CD CO iQ
«—i «—i CD CD
CD CD CD CD
'C0 <2ON 'ON
=fr-
CD
•
CD
CO
CD
*
CD
CD
CD
107
-------
CM
CO
CD
iO
O
CM
O
O
O
CO
C\J
CO
O
vD
O
sa-
cs
CM
O
O
O
O O
'aid lfH*o
to
•r-
c
s_
o
as
O
d)
T3
01
(O
s_
0)
J3
I
O
r^
i—
o
LU
C
s_
M-
O
d)
M-
o
o.
cr>
O)
rs
01
108
-------
OJ
1—I
•
CD
CD
•
CD
CO
CD
CD
*
CD
CD
•
O
OJ
o
CD
CD
o
UJ
s-
<+-
o
—
ce. CD
—> ~°
—; o
O E
o
IT)
OJ
OJ
»—<
•
CD
CD
»—i
CD
CO
O
O
CD
O
•
CD
OJ
CD
•
O
CD
O
'02HO <9HeO
109
-------
o
CO
r-
c\i
OJ
•
o
OJ
•
o
PflN, ppm
00 LO C\J
01
O
CD
O
CO
CD
CD
CD
CD
CO
CO
o
o
m
ai
3
CD
'ON
no
-------
PflN, HN03, ppm
CD
CO
o
r-
oj
•
CD
OJ
CD
OJ
•
CD
co in
CD
OJ
0)
CD
CD
O
CO
CD
CD
CD
CD
CO
o
GO
cc:
ZD
O
31
oo
i
o
c
3
S-
O)
•a
o
CVJ
LO
a>
3
CO
CD
OJ
CD
LO OJ
01
CD
OJ «—< «—• »—i
CD CD O CD CD
uudd 'eg <2ON 'ON
v£»
CD
CO
CD
•
CD
CD
CD
CD
in
-------
CD
o
o
CD
•
O
o
00
o
r-
o
vD
•
o
o
if)
o
sf
o
CO
o
CM
O
O
O
. I . , . I . I . ,
oo
I.I.I.!
CD
O
O
CD
•
CD
CD
CO
•
O
o o
r- VD
• •
o o
uidd
O
LO
o
=i-
o
O
CO
CD
CD
CM
O
O
CO
c
s-
o
t-
•r-
to
O
M
O)
(O
S-
(U
a:
(O
i.
-------
o
CD
CD
CD
CD
o
oo
CD
in
o
CO
CD
CM
CD
CD
CD
CD
CO
I
o
LU
c
3
s-
CO
O
LD
rs
O)
CD
CD
CD
CJ)
CD
CD
00
CD
CD
r-
CD
v£)
CD
CD
LO
CD
CD
CD
CO
CD
CD
O
CD
113
-------
hydes was labeled only "Aldehyde" (represented by Din Figure 49) and
the simple sum of formaldehyde and acetaldehyde predicted by the model
would be somewhat less than the original "Aldehyde" data. The N02
data for this chamber was taken by a chemiluminescent NOV analyzer and
37
subsequently corrected for PAN and HN03 interference . The model
predicts slightly higher NCL concentrations beginning at approximately
2.5 hours. Since the PAN data in the model agrees with the actual PAN
data quite well, the difference between the model N0? and the actual
N02 data would presumably be due to HNO~ formation or this difference
might have resulted from the correction applied to the raw N09 data.
-1
The k-, value used in EC-17 and EC-18 was 0.223 min . The rate constant
for Reaction 25 that produced the best results in the model was the
3he
-1
-2 -1
highest of any of the four chambers modeled, being 8.6 x 10 pprn
min , implying a very high heterogeneous rate of formation of HN02
and therefore very rapid initiation of the photochemical conversion
process. Since Reaction 35 also has a heterogeneous component its
rate probably should have been higher which would have produced more
HMO- in the model, thus leading to a better N0? fit.
In the UNC model of EC-18, the agreement for N02 and 0., is poorer than
in EC-17, but the model profiles of C3Hg, CH20 + RCHO ("Aldehydes"),
and NO, are as good as or better than those for EC-17. The observed
N02 profile for EC-18 is somewhat strange, since it decays to a plateau
level of 0.02 ppm for approximately 4 hours, even though it was supposed-
ly corrected for PAN and HNO- interference. Because the model maintains
a higher N02 concentration during the 1.5 to 2 hour period, the Oo
concentration in the model is higher than that in the actual data.
From 3.5 hours to the end of the simulation, the model 0^ decay is due
almost exclusively to ozone reacting with the remaining propylene.
Model of Lockheed Chamber
The last chamber/reaction system modeled was a very high concentration
37
propylene run in the Lockheed chamber which was mostly Pyrex glass
114
-------
for this run. Although no actual data points were shown in the original
plot of the Lockheed data, solid lines were drawn which exhibited
rapid changes in slope. These were reproduced in the digitizing of
the original plot and are present in Figure 55, the replotting of the
Lockheed data on the standard axis box for this report. The NO data
in the Lockheed study was from a chemiluminescent instrument and the
oo
N00 data from an automated Saltzman instrument . There may be some
calibration or response problem associated with this combination,
since, as shown in Figure 55, there was a gain in NO of 0.5 ppm from
A
start to crossover (NO=NO?=1.00 ppm, therefore NO =2.00 ppm) and the
£ -X
N0? maximum concentration exceeded the initial NO concentration. The
Lockheed data was taken at a k , (pseudo-first-order rate constant for
1
NOp photolysis) of 0.40 min which for modeling must be translated
into k,, the actual first order rate constant.
Three model simulations were run in which the rate constant for Reaction
25 and k-, were varied to obtain a reasonable fit to the actual data.
-4
Figure 56 is the last run of these three and used a value of 8 x 10
ppm -min for Reaction 25, well below that of EPA-325 and EC-17 and
18 but larger than the UNC chamber rate. The k-, value used was 0.256
_1 '
min which is similar to that in the SAPRC chamber, but much lower
than the value of 0.40 min~ in the EPA-325 simulation and the peak
value of approximately 0.55 min" which occurs in the UNC outdoor
chamber. Since 0- concentration is a function of the light intensity
under which it forms, both the SAPRC and Lockheed chambers would
probably yield 0., concentrations which are less than those that would
be found for the same concentrations of starting materials in the
older EPA chamber or the UNC chamber on clear days.
The model and Lockheed data show reasonable agreement as to time of
events and agree well in profiles for NO and C~Hg, but the post-peak
N02 behavior differs substantially and no clear ozone peak and decay
were exhibited by the model data. This difference might be attributed
115
-------
CD
o
O
CO
CD
in
o
OJ
o
en
CD
CO
Q
O
CO
C\J C\J
CNJ
T
Q
Ld
LU
O
O
ro
o
CVJ
O _
I , I I I I I , I , I I I I I I I
LO
CO
C\J
CC
ZD
O
O
c
=j
s-
-o
O)
u
o
-------
o
o
•
CO
r-
OJ
CO
LO
OJ
C\J OJ
o
en
o
o
CO
o
o
o
o
I ' I ' I ' I ' I ' I ' I ' I ' I
Q
LU
UJ
X
^
o
o
UJ
O
O
co
o
CM
O
I . I . I . I , I . I . I . I . I
LO
-o
Ol
O)
a
o
CO
OJ
o
n:
O
ai
-a
o
o
(£3
in
a>
3
C7)
O
o
oooooo
=j- «-• co LO c\i en
o
CO
o
o
CO OJ OJ OJ
•—* «—< «—• O CD O
mdd ''HC3
-------
to a poor representation of the Lockheed N205 surface chemistry in the
UNC model. Although the Q3 profiles differ substantially, the maximum
03 concentrations agree reasonably well.
Considering that the model consists of only 35 steps, that only two
parameters (the light intensity and the rate constant for HN02 formation
from NO and N02 which represents individual surface chemistry of each
chamber) were changed from chamber to chamber, and that a wide range
of physical and chemical conditions were modeled, the results above
speak well for both the model and the chemical performance of the UNC
outdoor chamber on which the model was validated. The differences in
behavior of propylene runs between the UNC chamber and more conventional
indoor chambers can be simply and readily explained on the basis of
differences in selected surface chemistry (which primarily affect the
speed of the system) and light intensity magnitudes and time distribu-
tion (which primarily affect the detailed profile shape and the
maximum concentration of ozone).
Bureau of Mines and UNC Chamber Comparison
One final comparison was made between a NO /auto exhaust run in the
J\
BOM smog chamber from data supplied by Dimitriades and a NO /simulated
}\
urban mix run in the UNC outdoor chamber. Figures 57 and 58 show
these two runs. The BOM run had 0.05 ppm more NO and approximately
the same IHC. The UNC run was slower than the BOM run, but the BOM
run was at constant light intensity. Accounting for the light differ-
ence effects, these two runs are in general agreement. The UNC run
formed 0.2 ppm more ozone, but the light intensity under which this
formation occurred was higher than that in the BOM Chamber- An
adequate hydrocarbon mix model has not yet been developed to allow a
more detailed comparison.
118
-------
CD O
O O)
CD
CO
*
o
o
UD
o
LO
o
CO
CM
O
CD
CD
i • i ' i •
• i
3
<0
s-
3
CO
CO
3
O X
CD
co I
•i- 3
M- rt3
O
S- O
a. e
Q-
i— OL
CO
3 OO
-P
O CO
LO
s-
3
cn
119
-------
cn
•
o
CO
•
o
r-
•
o
LO
•
CD
CO
•
o
OJ
•
o
«-i O
* •
O O
Q.
00
I ,... ,1,1,1,, ... I , ...
cn
UD
co
Q
LU
CO
O
O
s_
-o
(O
jQ
S-
X E
O Q.
z: Q.
•• o
(/) UD
C
O CO
•i- I
-o i
C I
o •—
o
n
i— O
(C UD
•r- •
-!-> CO
o E
CQ a.
•r- r-.
X o
S- I
ea I
CL s—'
o
U CO
00
S- O
o
t- o
o
CO
o
S-
o
o
+->
3
o
X CM
•i- O C
LO
t • •
»-H O O
• • » » • t
O O O O O O
'ON
CO
in
O)
cn
120
-------
SECTION VII
RESULTS AND DISCUSSION, PART II:
EFFECTS OF HYDROCARBON REDUCTION ON NITROGEN DIOXIDE CONCENTRATIONS
GENERAL
Of primary concern in this project was the effect of reducing NMHC
concentrations (for the purpose of reducing oxidant) on the behavior
of N02- The data generated by 130 mix runs in the outdoor chamber
were examined to determine the effect of hydrocarbon reduction on: N02
maximum concentrations (N02M), NOp ten-hour daytime dosages (DN02DIFT),
N02 ten-hour daytime average concentrations (N02AVT), and the nighttime
disappearance of any residual NOp at the end of runs. A nonlinear
regression model that expressed N02M as a function of initial NO
J\
(INOX), initial NMHC concentration (IHC), and average measures of
solar radiation at various intervals (SRAV1 and SRAV3) was generated
and explored. Data were also obtained and examined for possible
effects on N02M by factors such as initial N09/N0 , dilution, and
c. x
repeated irradiation.
RESULTS
Computer listings of data profiles and the summary listings of the
significant measures for all runs are not included in this report, but
are available as separate appendices.
For the purpose of discussion, mean values of clusters of data and
plots of significant variables will be used. Selected individual
profiles or exemplary runs will be used to illustrate specific points
in the discussion.
Table 12 gives the mean values, standard deviations, and number of
121
-------
LU
^1
_
LU O
<=> g
:z ca x:
LU =z o
<_> LU
^i *£ z:
o o LU
u ce CD
ca cz>
si ce.
=3 = \—
S 2 —
XI— ~~
00 3
S^ =
in _J ca
o — zr
i— ca
:== LU
LD
u_ «
o S
LU
oo >
cxi :z
—I LU
l_>
Ul ~^~
-1 O
CQ ' _ *
-..S 3SS
**?t*°b° z
eg O »
to ry w us £4
o t-- m cv ro
^< d <: d -H CO
cjcco
Ait
d d
CM ^H
O -(
o o
41 44
^t r*
d c
oco
r- »r
o -^
A
§1
O -1
si
o o
1n^
o •-<
o ^
o o
\p cy
0 -1
CM n-
O O
o a
44 +4
r-( I/N
rn r\j
IH in
O* rH
H
ITS r-l
. . w
0 ° S
41 44 41 «
CO rH F*
odd
«
0
SSS-'
0 0 O
IH.
o o o «
44 44 44 O
ON^ IfNr^
O O O
»-« m O
•H Q CM
°°rH. .
000 •
-* O CU fO
rH rHNO
odd
'•nvp m
O o cu
O O O O
4-1 +i +i tf
S tr\ (\J cy
d d c
d S
M S
122
-------
*R ~
o o
•ft -H
**" £
O H
d .-H
Q\ o
O iH,
O *-l
O O
Sf
O (M
CO OJ
CM IT*
0 Oj
0 0
•M -W
ry c*-
C\j \O
c\j cr
0 rH
CM co
C\J M
o -<
d o
(V C\J
o" .-<
&
d
i-j
m f \j H
O O OJ
0,0 o *>
•H •« -M OJ
(\J i— I Q"v
rn mj* t*-
m oj ^
o' d d
Ct
B)
c
u
OJ CM CO
odd
rH OJ O
O O -*
o o o »
u
OJ VO OJ f-
OJ ^ t--
o o o
IA m rH
rH H CO
O O i-H
O O 0 0
w
•H -ft -M OS
VO vD O
ir\ tH rH m
r-H fH CO
o o o
"R§
0 0 ft ^
d o o v
odd
(8
OJ
OJ
£2
0 0
^ OJ
IA J
d oj
ii
0 C
VN If
I/NCC
m^
o c
0 0
o o
-H -H
•H J-
o cy
o
d
IA
t-
CO ITN-*
u\^ ^p
o o o B
o o o «
« 4< •» S
^ irvoo
CO rH CO tf\
« CM(-
o o o
CM CO CO
> O O O 0)
c
) r-t I/N OtCO
r CU-H IA
jl odd
•H O rH
0 60
O O O u
IACO ru
odd
e5
IA
K »
JE X o
^f O-- lfv t g
do! d ' d
•
J| m
do d d d •>
mvjo oj •-< \n
d oj d d o
d »
I I
t- »A
rO
fK
IAOO
O OJ
si
o o
IA f-
vO o H t^-
H 000
o o o o o t)
O rf) O O O
d .
4-1 m
^ IA
m
sj »§§
do odd*
p^.ey w jA»g ^_
IA o m oj IA
d .* d o o
do d d d g
•H+* ^* S
J^S "w^.1^
d ^* odd
E o
OJ O O>O C£
OJ O © ^ CO H
r— ( C\) O O C"1
o •& o o o
d S
m
o
OJ O
*-4 rH
J-
123
-------
cases for the variables of most interest at each initial condition
cluster. The standard deviation of initial conditions represents the
spread of the data around each of the target initial conditions designa-
ted in the experimental design. Table 12 does not present the pairing
of runs that occurred in the experiments; this will be taken up in
other tables and plots.
EFFECTS ON NITROGEN DIOXIDE MAXIMUM CONCENTRATIONS
The N02M data were examined by several types of analysis, beginning at
mean values for the entire data set and proceeding to an examination
of individual paired runs (dual runs with different hydrocarbon concen-
trations). IHC concentration was found to have a significant effect
on N02M concentrations, but other factors such as solar radiation
conditions also caused variation in N02M values at constant IHC and
INOX. These other factors were not related in any simple manner to
N02M concentrations and required a detailed examination. This was
performed on the 51 runs at INOX = 0.36 ppm. Nitrogen dioxide profiles
for different IHC from individual runs (different days) at three
constant INOX values were compared. Nitric oxide, NO^, and 0- profiles
were compared for several dual runs and the percentage reduction in
N02M for a reduction in IHC at different INOX was computed for several
dual runs.
Mean Values
If the mean N02M (N02M) for each INOX cluster is computed without
regard for the conditions under which the value occurred (i.e., ignoring
effects of IHC, SR, Temp, and PCTN02; across rows in Table 12) and it
is plotted as a function of the mean value in each INOX cluster (INOX),
(Figure 59), the values of N02M for cases INOX <0.7 ppm can be fitted
by
= 0.5788 INUX + 0.0071 (1)
124
-------
Q.
CL
00
UJ
° o
o £
r- *
CJ3
O
o
(/) E
CU •!-
-O •—
O O
r— C
to o
•r- *
C 3
•r- CT
UJ
C
fO
cu
E £
CD
a o
C E
3 to
M-
fO O
o
to -o
fd +J
4J
C
CO -r- l/l
S C •r--
11
X
fO
LO
00
O
CO
O
LO
CM
C\J
LO
LO
o
CD
00
O
O
'wnwixyw 2ON
Ol
LO
CU
rs
125
-------
o-
cx
.
c5
^-
1
cr
i — i
1—
K^H
I.IM^
•r-
C
•r"
H-
O
C
o
•r™
CJ
C
"*-
X
en
o
s-
^^
•r-
C
O
•r—
4J
(O
3
0"
UJ
0)
JQ
E
ai
s-
3
CO
'XUN
126
-------
2
with an r of 0.9988. Equation 1 implies that over the range of
conditions that occurred in the experimental series, the N02M was
primarily dependent upon INOX.
A plot of individual N02M values as a function of individual INOX
(Figure 60) however, reveals not only the scatter around each target
INOX, but also a large range of N02M values which tend to cluster near
the upper end of each range. It is also evident that the INOX cluster
at INOX = 0.7 ppm might also be consistant with the line from Equation
1 if sufficient runs had been made to give a representative mean.
Individual Values at Constant INQX
To investigate the factors which might be responsible for the scatter
of cases around the N02M, the 51 cases for INOX = 0.361 ppm were
examined in detail. In Table 13, the individual N02M values and asso-
ciated information have been classified according to the IHC and
ranked in ascending order. In Table 13, N02M values increase with
increasing IHC, with two exceptions (IHC = 2.48 and 2.70), but there
was sufficient scatter at each IHC to cause an overlap in N02M values
from one IHC value to another.
Several factors were investigated to determine if they were associated
with the scatter in N02M at each IHC. The most important of these are
also listed in Table 13. A number of observations relevant to the
scatter in N02M can be made from the data in Table 13:
1) The total average solar radiation (SRAVT) for a day is not a
strong predictor of N02M given constant INOX and IHC and neither is
the average solar radiation from the start until the occurrence of
N02M (SRAV13),
2) Although some values of IHC exhibit a trend in time to N02M, time
to N02M is also not a good predictor of N02M,
127
-------
cn
C
CD
p5
o
o
to
fQ
CD
CL
O
(—
4J
ro
0)
rO
•o
j^
(I)
&•. U\L
| -r- JC ••-
P. CL
| t/1 C\( t/1
*--.
S-
01
+->
c^_ •
(O CO
S.
S- -C
ro
CD LO
•— •
Q.
CL
O
_C"
U
QL
O
JC
0
to
CD
*^
•I—
Q.
(/>
CO r— VD O r—
00
•
o
CM
CO VO
o o
i— O r— r- i—
I I I I I
i— O
«d- LO LO LO
CM 10 CO
t~- OO «*•
ttf- LO LO
CO «d"
"=C
„
r~\
LU
^~
Cr*
o
u_
CM
o
i .-,
LU
"Z.
CQ
•z.
o
--'
LU
o
CO
E
O-
Q.
r—
vo
ro
CD
U-
o
zz
o
I — I
H-
O£
1—
UJ
o
0
»-H
1—
1— 1
Q
C£
a:
e^
_L. J
o
oo
LU
CD
(V
LU
<-f
****
00
•z.
o
I— 1
h-
C£
h-
•z.
LU
o
o
o
^
~o •>
h-
> 0
00
U «
CM
O
•z.
*y
CM
O
•z.
O «%
2:
CM
O
•z.
fO
•z.
C£.
1
C
•r~
E
vl
1
o
1
r—
03
O
gz
Q.
/^
LJ.
E
Q.
CL
O
VD
OO
•^
CD
VD
CM
•—
CD
LO
O
CM
CD
o:
cn
CM
VD
o
fv.
CO
cn
o
CM
CO
o
o
JC
f—
o
Q.;
CM
CM
cn
o
^
CO
o
f—
VO
o
o
.C
"!d"
o
CQ
r—
0
o
^
l-s.
LO
CO
0
r—
CO
o
o
CO
LO
o
,;£
oo
CM
cn
o
IV.
LO
LO
o
CM
CO
0
o
r_
vo
o
CQ
CM
CM
0
l-s.
oo
o
o
r—
o
CM
oo
o
CQ
o
p—
o
l-s.
oo
cn
0
1 —
CM
*—
o
CM
cn
o
£•£
VD
o
o
fs.
cn
CO
LO
o
oo
oo
"—
o
vo
o
CM
o
CQ
CO
o
o
^
l-s.
o
o
o
vo
LO
o
Q
^^
CM
r—
o
CQ
f"^
CM
cn
o
(^
o
o
r^
r~-
o
o
CM
^~
1—
0
CQ
^~
CM
cn
o
1-s,
CO
0
LO
0
o
VD
r—
o
CO
cn
r—
VD
o
l-s.
o
o
LO
o
r—
o
^-
cn
CD
CQ
o
CM
cn
o
^
o
0
oo
oo
•—
o
1— -
f—f
CM
o
C£L
1—
f~~
^o
o
fv.
CO
LO
o
0
vo
CD
^M
oo
CM
CD
C£
CM
!»
o
^
vo
f*1^
CD
vo
LO
•—
CD
VD
oo
CM
•
o
CQ
CM
r—
vo
o
!<
« U
O E
cn
o
128
-------
•o
c.
o
CJ
co
CD
(O
<4-
UJ
CS
o; CM
i i
O
I
to
u
0)
CO
i— CO
u_ c
S 'i
E
i
CM
O
to
o
o "
CM
o
CM
O
CM
O
d
Q.
" CJ
'±
1-
ta
Ol
o
i
o
l~
CO
o
«d-
oo
00
o
CM
O
o
r—
CM
O
CD
cn
r—
cn
o
^"
'"-
cn
cn
to
13
CO
o
o
i
CTl
O
CO
VO
co
LO
CO
o
in
o
CM
CM
0
or
LO
o
o
r—
^-
^
>»
Q.
Q. (
O •'
.C 7
O
^> -
0)
> 1-
co
cn
i
LO
O
O
vo
co
O
vo
o
LO
'
o
o
co
CM
o
CQ
LO
CM
O
^_
«^
*"-
£
DL Q.
- O
3 -C
0
O
- E
CM
O
1
O
VQ
o
vo
CO
00
r^
o
LO
o
co
CM
o
c«:
CM
CM
o
r—
*^"
^
CO
CU CO
V? O
•r—
VO 2
1
o
•*
o
o
CO
CO
r~-
o
r---
co
o
VO
CM
o
co
cn
CM
VO
o
^j"
^
CO
-o
3
O -t-> «7
r— CO CU
O (C ^
O •!-
>, S- Q.
C CD CO
res >
E O CO
1
CO CM *3-
• • | •
1 1 1 1
CO f— Lf>
i *d-
0
OO ^- CM O
^O t-Q r— O
CO ^^ LO ^J"
^O CO *l}" CO
00 O^ ^^ ^O
co i — «3- r^
O 0 0 O
r— co o cn
r— O LO «3-
o o o o
O CO r- CO
CO CO CM CM
r— i— CM CM
O O O O
a: co or a:
r^ CM «d- co
CM CM CM CM
cn cn cn cn
o o o o
sf ^f ^t- ^>
r-- r^. r~- r^
CO
a.
o
0
t.
ro CO
cu a.
13 T3
•* CO
•— O
1 1
CO CO
cn vo
o o
CO CM
O CO
j^j C^^
00 i,^
o o
«=J- 00
«^- LO
o o
(O O
CM CO
CM CM
O O
X *
r- 00
O O
0 0
"^ '^
r- r^
S-
ro
0)
13
0
1
CO
CO
o
CO
vo
CO
CO
r^
o
VO
o
CO
CO
CM
O
a-
o
r— "
o
*^F
^
>>
4J
CO
•> Q.
co a.
Q. 0
•r- JCZ
-o o
CO
o
1
!"*•
o
^»
co
CO
r^
0
CO
o
cn
co
CM
0
CQ
VO
o
o
^S"
^
i-
ta
0)
"o
LO
"
i
cn
o'
CM
LO
CO
ro
^J*
CO
0
**
'
o
o
co
CM
0*
or
p— •
CTl
o
IN*
o
CO
Q.
•r*
T3
O
CM
O
*
1
CM
VO
O
CM
cn
co
CO
d
00
vo
o
CO
CM
O
03
CM
r—
O
r—
^"
r^
$-
(O
CD
13
(0
a.
LO
CO
Q
i
co
i**
O
CM
CO
VO
LO
d
co
VO
1
d
cn
CM
o
Q£
cn
CM
o
^-
^
!>
•p—
D.
CO
S-
0)
CO
CO
1
LO
CO
d
CO
vo
co
co
co
vo
d
CO
LO
*
d
o
LO
CM
d
o:
0
CM
O
^
CM
CM
129
-------
.
TD
-U
c
o
o
•
oo
,-_
a;
_Q
"3
\—
CD
4->
E
c
E
O
O
«4- 1
LU E
CD T-
Z E
S '
CY* pv i
Ui. VNI
1 1
rv gz
oo o
to
o
cu
CO
i — tO
U. E
t— f •!—
Q E
2:
1 — 1
1—
f—
1
-O « E
h- T-
> E
DC CM
00 1
(j
r—
n3
0
U •>
CM
O
•z.
>- E
1 O
1 1—1.
^"T" r~i
*£. 1-J-
C\J
O
•2.
o •»
2! E
CM a.
f~\ O1
i a
(O
2: o
n z
a:
J 'i
S-
>
r—
+->
S-
(0
0.
LO
CO
o
1
co
r^
o
0
00
CM
co
00
o
00
LO
o
o
CM
o
CQ
01
CM
O
«^-
r^-
a.
•r-
T£J
CU
cn
S-
(O
r—
p—
00
CTt
o
1
LO
LO
0
CM
H~
co
cn
co
LO
o
r*^
LO
o
^"
^*
CVJ
o
rv^
LO
CM
O
^^
1^
>> s-
O- to
Q- CU
Oi—
JC O
u
o
r—
,_
|
O
CM
0
CM
p*^
CM
LO
LO
00
0
LO
!*-•
0
r-^.
^>
CM
O
CQ
LO
O
O
^~
f~»
cr
cu
1-
4- to
Q.
r— "r-
P"™ TO
10
E
CO
CM
r—
|
LO
cn
o
LO
CTi
CM
cn
i^
r^
o
CO
CO
o
^J-
LO
CM
O
o;
CM
LO
0
^"
f^
_
cu
s-
0)
x:
3
oo
a
a
p
^^
;,_
I/)
(O
E
O)
>
•r-
D>
tO
*!»•
$-
cu
JQ
E
3
E
E
3
S-
CU
<7"
H-
E
3
i_
(0
CD
E
•r™
s-
3
-o
f—
0
•r—
4J
(O
&-
^_>
E
CU
O
E
Q
O
CM
O
E
3
_E
X
to
E
CU
.E
1—
.
•a
cu
+j
o
0)
•o
•r-*
CM
O
,— .
fQ
•r-
4-^
*(-•
C
•r—
O)
c-
•^
(/)
3
E
•r^
E
£~
3
E
X
§
CM
0
z:
•
•o
cu
E
C
o
CM
0
E
a.
a.
4->
CU
E
CU
J!^
r-
,
O
O
rC.
i—
|~
+3
E
3
E
O
• r—
4J
O
CU
tr™3
E
•r—
E
O
X)
5-
(O
0
O
s-
•a
>^
-C
cu
CT"
4_j
E
o
Cf_
c
O
-M
fO
-a
to
s-
t«
(O
r—
o
to
r_
fO
4^
O
-t_^
CU
cn
fO
s-
cu
>
H-
CU
0
E
CU
$-
3
O
O
O
cu
JC.
•M
TD
E
tO
E
O
-4_)
0
cu
•r~5
c
•r-
g;
O
_Q
s-
n3
O
O
S-
TJ
C"
cu
. ^
E
CU
cu
+J
cu
Lr^
cu
5J
E
CU
S-
0)
M-
M-
"O
cu
E
•r-
•fj
CU
C"
H-
2:
CM
0
•Z.
CU
-C
4J
O
-(->
s-
cu
o
to
to
0
S-
o
CM
O
O
^^
E
o
s-
tf_
E
O
•r~
to
T3
(O
S-
t.
ro
r~
0
to
o
E
•r—
CU
cn
E
to
i-
cu
C*
i—
•
E
CU
^
•r-
C7)
cu
cn
E
to
S-
cu
^J
E
•r-
E
O
4-}
(0
•1 —
•o
(O
X-
^
•r—
4->
(O
4J
••-
to
3
cr
•
j_
cu
>
o
(/)
to
o
i.
o
•M'
O
•o
•r—
Q
CM
co
co xi
a>
cn
130
-------
3) At every IHC value which contained a June run, a Oune run exhibited
the maximum N02M for the category,
4) The September and October 1974 runs tend to c^cur near the lower
values of N02M at each IHC value and the two IHC which have only runs
from these months have lower maximum N02M values than would be ex-
pected from the trend discussed above,
5) The October runs of 1974 with lower SRAVT than a typical clear
_o
October day (^0.8 cal-cm -m
October and September days.
-? —l
October day (^0.8 cal-cm~-min~ ) have higher N02M values than clear
6) The maximum N02M at a IHC value is never associated with the
maximum SRAVT or SRAV13 for that IHC value.
The explanation for all of these observations can be related to the
effects of variable, sometimes "choppy", solar radiation and to the
method of reporting solar radiation measurements used in this study.
The values of SRAVT are computed from the total integrated solar
radiation for the day (ISRDIFT) and the total time difference (TIMEDIFT)
and represent the true average over the period. If a dip or "hole"
occurs in solar radiation due to clouds, the ISRDIFT reflects this,
but it does not indicate when it occurred. The same is true for
SRAV13 (which is computed from the integrated solar radiation from
start until the time the N02M occurs, i.e., ISRDIF13/TIMEDIF13) except
that if SRAV13 is low then the dip or hole occurred before the N02M.
Thus, very low values of SRAVT indicate a fairly cloudy or overcast
day and very high values of SRAVT indicate a very clear day. Inter-
mediate values probably indicate some choppiness but not when this
choppiness occurred.
From Table 13, and observation (6) above it appears that the maximum
N02M in each HC category occurred on choppy days. Since a clear day
-2 1
in June yields a SRAVT value ^0.2 cal-cm -min higher than a clear
October day, the SRAVT for a choppy June day can be nearly equal to
the value for a clear October da./ and furthermore, the June day will
131
-------
have peak values that greatly exceed the values for the clear October
day. In addition, the actual solar radiation peak values for a choppy
day exceed the peak values for a total clear day as was shown in
Figure 28. Since this phenomenon occurs in the readings from both the
Eppley total solar radiation radiometer and the Eppley ultraviolet
radiometer, which operate on different sensing principles, it is
apparently real and probably results from the cloud increasing the
forward scatter of radiation into the direct solar beam just before
and just after it passes over the sun.
Furthermore, comparisons of closely spaced runs, for clear days and
choppy days with otherwise identical chemical and temperature conditions
indicates that choppy days are faster, have larger maximum NO oxidation
rates, higher NO,, maximum values, larger 0- formation rates and fre-
quently higher final 0., concentrations depending upon the degree of
choppiness (see for example Figures 23, 26, and 27).
The above discussion suggests that the effects of solar radiation on
photochemical systems is not solely dependent on the dosage or average
value, but also on the maximum value that occurs during an interval.
The effect of rapidly changing solar radiation depends upon the state
of the photochemical system at the time the change occurs. Two examples
of this are shown in Figure 61 and 62. In Figure 61, the short, small
burst of solar radiation at 0910 hours was able to initiate the system
sufficiently so that a large change in the NO and N0? rate occurred,
which was sustained for the rest of the run even though the solar
radiation never again reached the peak value during the run. The dual
run shown in Figure 62 was a test of the effect of changing the initial
percentage N09 for constant NO and HC. The expected effect of differ-
L. A
ent N09/N0 did not occur in this dual run after 1030 hours due to the
L- X
difference in response of each chemical system to the rapid decrease
and then increase in solar radiation at 0900 and 1000 hours.
132
-------
E
O.
0.5
0.4
0.4
0.3
0.3
0-3
0.2
o.i
0.1
0.0
0.0
Figure 61.
II T .'[ I
I ' I ' I ' I ' I '
MflY 5. 1974 _
2.0
1.8 g
1.6 £
1 4
1 •"
3>
O
8
17
o
1.0 T
o
0.8 *i
3
0.6 ^
Q.I
0.2
0.0
9 10 11 12 13 14 15 .16
HOURS. EOT
Effect of single light pulse on nitrogen oxides conversion rate in a system initially
containing 0.511 ppm NOX> 0.114 ppm NO^ and 1.89 ppmC propylene in UNC outdoor smog chamber.
1 I ' I ' I '
MflY 19. 1974
14 15 16 17
0.0
9 10 11 12 13
HOURS. EOT
Figure 62. Example of rapid changes in solar radiation accelerating occurrence of nitrogen dioxide
maximum. Compare with clear day behavior shown in Figure 77. Initial hydrocarbon
concentration 1.95 ppmC propylene in both sides.
133
-------
Individual Runs on Different Days and at Different INOX
Comparison of runs conducted on different, essentially totally clear
days which had the same INOX but different IHC does show a decrease in
N02M with a decrease in IHC. Figures 63, 64, and 65 depict only the
N02 time profiles at different IHC for INOX - 0.5, 0.35, and 0.22 ppm
respectively. In Figure 66, the N02 profiles for two dual October
runs of INOX = 0.36 and four different IHC are shown. Figure 67 shows
typical solar radiation profiles for the runs in summer and the October
runs.
From these figures, the effect of HC reduction on N0? maximum concentra-
tion, N02 dosage, and dosage distribution, is clearly discernible.
Thus, 52% and 56% reduction in IHC at INOX <= 0.36 ppm, cause reductions
in maximum concentrations of approximately 7% and 20%. The October
runs in Figure 66 exhibit much longer times to N02 maximum due to the
lower total solar radiation available and the shorter solar day (Figure
67). October 1974 also had more haze and thinly overcast days than
October 1973 reducing further the amount of actinic radiation available
during this period. The effects of reducing IHC, however, are clearly
evident in Figure 66.
Dual Runs at Different IHC
Pairwise determination of the effects of reducing IHC was accomplished
by running differential runs in the dual chamber with all conditions
identical except IHC on the two sides. Thus, unlike the data in
Figure 63, a comparison of N02 behavior can be made for the same solar
radiation or temperature conditions. The N02> NO, and 03 profiles
from two of these dual runs are shown in Figures 68 and 69.
In Table 14, the differential percentage reduction in N02M is given as
a function of IHC differential reduction. The N02M reduction also
appears to be a function of the INOX; higher INOX yields a greater per-
centage N02M reduction for approximately the same IHC reduction.
134
-------
Table 14. PAIRWISE COMPARISON OF EFFECTS OF
REDUCING INITIAL HYDROCARBON CONCENTRATION
ON NITROGEN DIOXIDE MAXIMUM CONCENTRATIONS
FOR SELECTED DUAL RUNS WITH URBAN HYDROCARBON
MIX IN UNC OUTDOOR CHAMBER
Initial
NOX,
ppm
0.261
0.257
0.358
0.369
0.373
0.513
0.520
0.508
Initial
Run
No.
741013
740622
740611
740612
740629
740616
740620
740702
High
Side
0.71
1.87
2.80
2.90
2.11
2.58
2.50
4.00
HC,
ppmC
Low S
Side
0.
0.
1.
1.
0.
0.
0.
2.
32
85
40
30
72
70
80
29
; Reduc-
tion
54.
54.
50.
55.
65.
72.
68.
42.
9
5
0
2
9
9
0
8
N02 maximum, ppm
High
Side
0.
0.
0.
0.
0.
0.
0.
0.
1781
1731
2385
2638
2669
3427
2979
3713
Low
Side
0.
0.
0.
0.
0.
0.
0.
0.
1680
1577
2110
2362
2053
2213
2152
3129
% Reduc-
tion
5.7
8.9
11.5
10.5
23.1
35.4
27.8
15.7
135
-------
in
o
CO
•
o
CO
CD
CO
CJ
C\J
o
CD CD O O
0)
>
3
O
:r r^
I £
CM
rtJ V£>
•i- O
C E
•r- Q.
Q.
T3
C O
(O 00
E CD
O.
o. -o
c
i— rC
LO
O CM
i|_ Qj
O >
i-
X =3 i.
O O OJ
z .. -i
Q *
UJ +5
CM
CO
O
31
c
(O
^J
in
c
o
0)
r»> o
to
o
§. fe
Q. O
en
O CM
O
CM
+J
rtf «
=3
O
O
Q.
QJ
X CM
O ~^
LO
If)
O
=t-
o
zt-
•
CD
CO
CD
CO
O
CO
CD
OJ
*
CD
CD
CD
CD
CD
C
O)
0>
o
-(->
CO
d)
s_
o>
o
X5
(O
O
O
-a
Q.
a.
o
o
ujdd
136
-------
LO
=*-
•
o
CO
•
CD
OJ
»
O
cu
s-
3
O
I ' I ' I ' I
X
o
e
D.
D-
CO
•
O
O
X
h-
o
LU
- UJ
CM
O
1,1,1,1,1.1,1.1
o
o
03
CO
CM
00
r-
LO
*
o
o
CO
o
CM
I
o
Q.
Q.
h-
Q
UJ
CO
rv*
7^**
^1H
o
TL
-o
c
re
E
Q.
Q.
UD
CO
.
O
if-
o
X
0
z:
,
re
•i —
+•>
c
•r- •
^J
c
re
1/5
c
o
o
+J
re
w
cu
4!
o
S-
Q.
CU
-o
*^™
X
o
*l«
"O
c
cu
en
o
S-
^_>
• 1—
2T
O
^~
•
*""
"O
c
re
^— s
CM
0)
s-
3
O
A
r--.
IT)
' —
*^»
o
E
Q.
Q.
CO
LO
•
'~
^
"^
CU
i.
3
0
«t
"^f"
r^
CO
-^
LO
^**
o
E
Q.
Q.
CM
•
CO
•
3
0
o
^yt
^
c
•>-
X
•r"
E
c
o
fl
c^
re
o
0
T3
>^
c**
c:
re
-Q
s-
3
S-
o
4-
CU
3
cn
137
-------
to
o
CO
O
CO
O
CO
O
CM «-i
O CD
O
O
CD
O
oo
• i • i
X
O
e
Q.
Q.
CO
CD
o
31
O
h-
o
LU
LU
I i_L .1 .1.1.1
03
o
.£3
S_
•4-> O
••- o
c $-
•i- "O
re
E
Q.
Q.
CM
C
(0
CO
0)
LU .re
co .;=
O
DC
CM
LO
CD
=1-
•
CD
CO
CD
CO
CD
CO
CD
CM
CD
CD
CD
CD
CD
CD
.fj
c
re
t i
H— '
c
o
0
re
O)
•r~
O
S-
Q.
CU
•o
*r~"
X
o
•5
c
cu
o
s-
•M
•r-
Z
•
vo
cu
s-
CJ
E
Q.
Q.
CO
o
•
•~
c
re
A
^-^>»
CM
CU
s_
3
O
^p
r**^
^»^
CM
CM
^^
if^
x»x
O
E
Q-
Q.
c
CU
•i
re
x:
0
CD
O
W
S-
o
o
T3
-P
3
O
o
^^
r)
iudd
138
-------
to
O
O
CD
CD
LO
1 —
Q
LoJ
•
CO
ttl
• < \
ID
O
1 i '
'
£1
•i—
-a
c:
03
S
CL
Q.
10
co
•
o
0
X
O
2:
r— •
•2
^_>
•r—
c
•r—
-f~>
(O
^.
CM
|—
UD
T3
C
(0
ir>
S-
a>
-Q
o
4_>
0
o
s-
o
M-
l/)
at
*—
ijl
O
Q.
O)
•a
•r-
X
o
•5
a>
en
0
S-
4-)
•r—
ZZZ
f^
••^^
ir>
O
i—
s^_x
0
E
a.
a.
00
r~
*\
— — x
CM
CD
3
O
A
«s}"
1*^
"^x^^
UD
•^v^.
o
—
C_J
CL
Q.
00
"d"
Cvj
„
^— ^»
r~
ai
J.
3
U
•t
S
^x,^
Lf>
0
•CS
c_>
E
a.
a.
o
o
*
CO
t+_
o
o
ac
s:
2:
en
o
E
CO
$_
o
o
T3
4_>
3
0
O
^
c
•I—
X
•r™
E
c
o
J3
c
(O
O
0
i.
•a
^>
c
(0
.a
s-
3
^~*
<&
0)
^
3
O
n
^J-
px.
•*s^
*^
•^x^
O
CI-
o
a.
a.
o
CM
•
r~
-a
c
|
o
139
-------
ta
C\J
OOOOO
CU
-o
3
\y\y\ ' i ' i
IO CD
i- C
3 O
O _l
O
O
O t-
O (U
E jQ
s- re
re _c
t— O
o
GO CO
-o
re
o
re
-a
c
re
cu
s-
o
o
-o
3
O
o
0)
re
cu
£
3
10
s_
re
cu
"o
re
S-
o
M—
c
o
•(—
4_>
«
•1 —
•o
re
i_
5-
(O
•M
~a>
•i-.
•—
re
a
c
s-
cu
4^
to
re
UJ
CO
S-
3
0
.C
o
CM
CO
.
.^-^
z:
o
CM
p^.
•
in
CO
cu
•o
3
-P
•r*
•4-^
fO
o
tO r— «
fO ^
re -r- o
4-> E
o o en
I— c r-.
LO
CM
O
0)
3
CD
140
-------
o
CD
•
O
CO
•
O
r-
*
o
vO
t
o
LO
•
O
CO
•
O
CM
*
O
O
o
•
CD
1 —
o
LU
•
CO
a:
ID
O
m
* A
E
a.
O.
0
l__
in
o
— -
i
i
i
i-».
o
un
o
,_x
1
J^
X
o
•z.
in
c
o
+J
»r~
•a
c
O
o
^
>
x:
c
n3
J3
i.
3
O
£
Q.
a.
o>
CM
•
C\J
T*
1
Ci
o
o
•
^d-
_
0
3:
s:
• A
E
Q.
Q.
CM
O
•
O
x-^S
|
1
1
x_x
rv
o
o
p—
,
o
<^-^t
1
1
CM
O
2=
S-
O)
JO
E
(O
JZ
o
en
O
£
3
O
O
z
ZD
C
fr~
X
•r-
E
LO
o
o
o
o
o
o
00
O)
UJdd -eg '*ON 'ON
141
-------
o o> co
• • •
«-« CD O
in
CO CM
1 I M
o>
— O)
CM
LU
- Z
1,1.1.1,1.1.1.1.1
Q-
Q.
CO
.
o
CO
CO
o
c:
o
XI
s-
(O
o
o
s-
(O
Q-
a.
(NJ
LU £
o
-* CO -
8
oo
o
r— •«• S-
ro
E
Q.
Q.
O
00
O
O)
•
o
00
•
o
ujdd
LO «»- CO CM
O CD O O
D ''ON 'ON
«-H O
(O
T3
S-
O
to
OJ
O
S-
Q.
01
O)
i-
cn
o
CO
o
|
o
CD
o
S-
o
o
-o
3
O
CM
X
1
142
-------
EFFECTS OF HYDROCARBON CONTROL ON N02 DOSAGE AND AVERAGE CONCENTRATION
Dosage may be treated as dosage or, in equivalent terms, as the average
concentration that would be required to achieve the same dosage over
the same time period. The latter is somewhat easier to deal with
conceptually and is easier to relate to maximum values. The time
period over which total dosage was computed in the^e outdoor runs was
approximately 10-11 hours. Both measures, the total N02 dosage
(DN02DIFT) and the 10 hour average N02 concentration (N02AVT), will be
used, somewhat interchangably, in the following discussion.
If the mean N02AVT (N02AVT) is computed for each INOX cluster in Table
12 without regard to the conditions under which the value occurred
(i.e., ignoring the effects of IHC, SR, Temp, and PCTN02; across rows
in Table 12) and plotted as a function of the mean INOX in each class
(Figure 70), the values of N02AVT, except for INOX =0.7 and the 4
cases in which INOX = 0.433, can be fitted by
N02AVT = 0.4246 INOX + 0.0042 (2)
2
with an r = 0.9989 for 114 cases. Equation 2 is similar to Equation
1 and would suggest that, in the mean, the 10 hour average N02 concen-
tration was primarily dependent upon INOX. Since both the N02M and
N02AVT can be reasonably fitted by equations in which the independent
variable is INOX only, the ratio between average and maximum N02 can
be found by a ratio of the slopes of Equations 1 and 2, neglecting the
small difference in intercepts.
N02AVT 0.4246
= =0.7337 (3)
N02M 0.5788
This value is to be compared with the mean value of the ratio of
N02AVT to N02M for each INOX, IHC cluster which is 0.7252.
143
-------
CO
CD
•o
X
O
to
E
Q.
Q.
CO
LU
Q
I-H
X
O
o
o
LO
CO
CD
CD
CO
CD
LO
CM
CD
CD
CM
CD
LO
CD
LO
CD
•
CD
CD CD
CD
CD
to
O)
c
c
o
Q.
Q.
O
"O
tO
0
c
3
|j
*-H-
to
CO
to
C
o
•r—
4-5
to
i.
± \
+•*
C
0)
o
c
o
o
OJ
CD
fO
s-
O)
>
(0
S_
3
O
J=
O
r™
(1)
T3
•r-
X
o
•1—
•o
c
O)
CT>
O
s_
4->
•r-
C
C
CO
to
3
o-
LU
•
i-
O)
-Q
E
CO
J=
0
O)
o
E
CO
S-
o
o
•o
4->
3
0
o
•z.
c
•r-
X
•r—
E
C
0
JD
S_
CO
O
o
5-
-o
>,
c
to
-Q
l_
3
s_
o
<+-
c
O)
Dl
0
s-
4->
C
tf-
o
J=.
4->
o
.a
s_
o
«t-
co
0)
to
(0
CJ
td-
>>
i —
E
O
•o
CO
J=
>,
^ff
T3
QJ
4-»
CO
C
cn
•r-
CO
0)
•a
CO
4J
C
•r—
O
0_
B
<^-
0
o
• o
+
X
o
•z.
II
UJ
•^.^
•<
CM
O
Z
c!H 0! 8ON NU3W
144
-------
In Table 15, the mean N00(WT and mean N02AVT to N02M ratio across all
other conditions of SR, TEMP, PCTN02 are broken down by INOX and IHC
cluster mean values. There appears to be an upward trend, a maximum
and then a decline in N02AVT with increasing TflC in each INOX cluster.
The trend is relatively small and the data exhibits noise apparently
due to solar radiation effects and the small number of runs in a given
class. The IHC at which the maximum N02AVT occur' appears to be an
increasing function of INOX. Overall, the N02AVT constituted a relative-
ly fixed percentage of the N02M (72%), but it does appear that the
ratio is somewhat smaller at lower INOX and higher IHC.
Examination of Figures 63, 64, 65 and 66 in terms of N02 dosage indi-
cates that even though the total dosage (DN02DIFT) might have been
fairly close over a 10 hour period, the distribution of dosage was
not. In Figure 64, curve 1 (IHC = 3.2 ppm) resulted in a higher
dosage before solar noon and curve 3 (IHC = 1.40 ppmc) resulted in
essentially the same dosage at the end of the run, but had a much
smaller dosage before noon. The behavior of N0'2 concentration and
dosage after 1700 hours will be discussed later.
The pairwise comparison of runs conducted on the same day with the
same INOX, but different IHC is given in Table 16. There was a strong
effect of reducing IHC on the total N02 dosage and 10 hr. average N02
concentration which parallels the reduction in N02M. This is not
unexpected given the near constant ratio of 10 hr. average N02 (N02AVT)
to N02M discussed above. A 72% reduction in IHC from 2.58 ppmC to
0.70 ppmC resulted in a 35% reduction in both N02M and DN02DIFT and
thus in N02AVT by approximately the same amount. It should also be
noted that reducing IHC generally leads to a higher final N02 concentra-
tion, if the N0? maximum actually occurs (_i.e., dNOp/dt = 0), This
means that the evening and nighttime periods will begin with a sometimes
much larger (up to 100% higher) NO- concentration in the case of
hydrocarbon reduction versus no reduction. The implications of this
condition will be discussed in detail in a subsequent section.
145
-------
"^
2:-
2:
p-H
X
^£
^^
o
i—
LU
CD
C£
LU
>
«=c
z:
<£.
o;
Z3
o
o
1 —
LU
Q
£2
O
HH
f 1
z:
LU
CI3
O
C£
I — -
1 — *
z:
z:
<
^ t j
s:
to
•—
0)
^~
-d
H_
z
<:
LU
s:
Q
z:
<£.
z:
LU
CD
O
cc
\—
1 — 1
z:
U_
O
co
LU
Q
I 1
X
O
_l
1 — 1
1—
z:
i—1
z:
c^
LU
s:
fY>
z:
3
o
Q
2H
LU
sx
o
Q;
CQ
o
i — i
\—
<
/•Y*
L-L.
LU
/ ^
i — i
X
o
h-H
t ^
^y
LU
CD
O
C£
1—
z:
(O •
OL
LU
CQ
^"
^^
~T~
O
CD
O
s:
CO
o;
o
o
Q
I—
*T")
o
0
z:
^3
z:
t-H
X
1— 1
^^
o
CQ
Cd
O
o:
Q
>-
31
z:
^^
CQ
rv^
ID
o:
0
Ll_
z:
CQ
C£
-
zn
i
e£
i — i
1—
1 1
p-H
E
Q.
Q.
O
r^
to
.
O
II
X
o
z:
i — i
E
Q.
Q.
J—
CO
II
X
O
z:
h- H
E
Q.
Q.
CM
CM
CM
II
X
O
z:
i — i
. £
o to
z:
z: o.
CJ C_?
5 %
i CM i r- r^ LO i ,— i
ii i i r--
co <* o co en co
i i ... i . i
1 to 1 to 1 to 1 LO
p^- f^«. to i>*» r*^ to
!-•» i — r — LO «* LO
to »* co r— co CM
1 i — 1 CM CM CM 1 CM 1 CM
1 1 ... | .|
O O O O O 0
ir— r^.i — r*~co^-CT>^i' i
i f— i
LOi — OCTtLOLOlOLO
ICO^J-COOCMCM'^r^ 1
p^i^i-^r^r^r^r^Lo
OCM^^CTlLOOOCO
LocM^voLor-^ooLO
oooooooo
LO (Tl LO CO CO CM 1 i— 1 1
1 1 1
LO ^}- ^ ^J- i — O> CO
*3- O CTl CD «d- CM 1 i — 1 1
to i^ to to r*«* to to
tO Cn CTi ^J" to OO CM
^^ CTJ CT^ CO t~~ ' r^ 1 OO 1 I
O O O c— !— O 1 O 1 1
o o o o o o o
i — CTitOi — CT»OOOO^3-CM
C\J f^* CD *Hf CT> ^" ^^ CTi ^i~ r^
O O i — i — i — CMCMCMOO^
CO
CM
o
r-~
f*^
f"—
CSJ
CVJ
o
LO
00
,_!
r~»
00
LO
1
o
00
CM
LO
CO
to
LO
CD
o
o
o
1—1
r—
(O
C
(O
CU
E
o
I '
(O
r—
a.
o
Q.
CU
i.
'£
c.
ai
CU
.C
-------
z:
O OO
P rD
«=c cc
cc
z: <
LU rD
O Q
z:
O Q
<_> UJ
1—
z: o
o LU
CQ _J
QC UJ
«=C OO
0
o ce:
o: o
O U-
>-
rc uj
c n
1 ef
<: 00
"-" O
1— Q
i — r
"5* f"\
i— i "Z
e^
C_D
z: z.
>-> o
0 >-H
m f—
Q ec
1 1 1 ft*
ry | —
^y
U_ UJ
o o
— ; ff
OO O
h- O
UJ UJ
U_ r_9
U- <
1 1 1 ry'
UJ
U_ >
O <
z: uj
0 Q
OO I— I
I— I X
DC O
< I-"
Q- a
2:
o z:
o uJ
C5
LU C?
oo a:
I-H |—
CV*
UJ
CQ
^^
^£
rc
o
C£
O
O
Q
h-
O
O
z:
~5
z:
i — i
X
i — t
^~
z:
o
co
rv
ec^
O
O
a:
0
>-
*T"
z:
e^
CQ
a:
rD
E
a.
a.
*
CM
O
z:
^
f—
• r-
rC
a
0)
01
(0
s-
O)
>
^^
I
U
3
^3
CD
Q:
^
^
0
1
C"
o>
•r—
31
•
c:
•r—
E
E
QL
Q.
A
0)
ro
t/)
O
Q
CM
O
1
O
13
-a
OJ
o:
Cl~^
2
O
_J
J^~
a>
•r-
ac
(_)
E
a.
a.
A
0
^
1—
fO
•r"
-f_)
•r—
c*
l— l
i
u
33
-o
cu
Q^
c^
2
o
1
r"
CD
•i —
rc
c
o
•r—
4J
CM
*
r—
^_
, —
OO
^o
r~
•
O
a>
P«^
r—
•
o
LO
•
LO
cn
•
C7^
CO
CM
•
ID
cn
o
•
0
LO
0
•
r-~
o
OO
•
CM
CO
•
t_J*
0
VO
f—
•
o
00
10
r—
•
o
p>^
•
*st"
CM
•
IjO
CJ^
cn
•
0
o
r—
CM
•
LO
LO
O
oo
•
^—
o
cn
•
CM
oo
•
CM
CM
O
LO
^—
•
O
oo
cn
l~—
•
o
f—
•
CM
CM
r-x.
•
o
CO
«3
•
oo
o
r—
cn
•
LO
vo
CM
|>^.
•
o
r—
i —
•
CM
,-_
•
LO
OO
O
|*«x.
r—
•
o
CM
to
CM
•
O
r—
•
LO
00
r*^.
•
o
o
1 —
CM
•
LO
Lh
r—
CJ^
•
CM
1*^,
o
•
o
CO
LO
•
CM
CM LO
• •
1-^ ID
CM
OO O
to oo
i— CM
• •
O O
*^" ^O
CM <3-
CM CM
• •
O O
•— o
• •
to co
CM
r-~ co
• •
r— t^o
O CM
r— r~~
r*^ oo
• •
r*** r**.
oo oo
r~* i—
0 CO
CO CM
o cn
CO CM
O CM
O 0
LO O
• •
CM «*
r- r— CM
LO to to
o o o
IO O CM
r- CM O
to to r^
o o o
•r- X
-MO
•r- z:
O.
a.
r— fv.
VO LO
CM CM
CD CD
co cn oo
LO to r~x
oo oo oo
O CD O
00 O CO
r- CM O
LO LO LO
000
147
-------
FUNCTION RELATIONSHIPS
A search was made for meaningful functional relationships among nitrogen
dioxide measures, initial conditions and meteorological factors by
multivariate regression analysis. A combination of preselection (a
variable is forced into equation on a particular step) and stepwise
2
(program chooses the variable that will contribute the most to the r
from a list of variables available) techniques were used to build the
equations. Testing was performed on various model equations by linear
least squares curve fitting techniques. The variables that were
treated as dependent were: N02M, DN02AVT, and N02AVT. The independent
variables were INOX, IHC, PCTN02, all solar radiation averages (SVRAV1-
T), various combinations of SR averages, and all temperatures and
temperature averages.
Various sets and subsets of the entire data set (130 cases) were
analyzed as well as the entire set. Given the information presented
in Figures 59 and 60, and the fitting of Equation 1 to the mean condi-
tions of only the 0.51, 0.36, 0.222, 0.122, and 0.007 ppm INOX classes
it seemed reasonable to form a subset consisting only of data in these
classes (the limited number of runs in the other classes probably do
not form a representative data base for those conditions). In addition,
dilution runs, runs in which the percentage IN02 fell outside of the
range 18 to 24% and of course, runs with one or more of the desired
independent variable missing were also excluded. This resulted in a
data set with 110 runs.
Models consisting of simple linear combinations of initial conditions
and meteorological variables gave relatively poor fits to data. Since
most chemical systems exhibit multiplicative behavior (i.e., rate =
k[NO] [HC]), models involving the logarithmic transformation of the
chemical variables were explored.
In the simple case of the mean values of each INOX class (Figure 59),
148
-------
the logarithmic transformed variables gave an equation with a slightly
greater r than the simple linear Equation 1.
0.8864 ?
N02M = 0.5336 (INOX) r = 0.9997 (4)
When the same functional form was fitted to the individual points of
Figure 60, slightly different coefficients were found (Equation 5): a
much
line.
2
much lower r indicated a large amount of scatter around the fitted
0.9199 9
N02M = 0.5631 (INOX) / = 0.846 (5)
Equation 6 resulted from a linear least squares fit of the logarithmi-
cally transformed chemical variables and untransformed meteorological
variables for 110 cases.
0.844 n ,..
N02M = 0.4927 (INOX) (IHC)U''^ EXP. (0.144 SRAV3 - 0.270
SRAV1)
r2 = 0.9244 (6)
The magnitude of various components of Equation 6 are given in Table
17. As suggested by Equation 4 and 5, the N02M is dependent primarily
on INOX. The values of IHC ranged from 0.2 to 4.2 ppmC in these runs
and, as indicated in Table 17-B, this range can introduce a +20%
variation in the N02M predicted by Equation 6 at constant INOX and
constant SRAV3/ SRAV1. At constant INOX and IHC, the typical variations
in solar radiation can lead to a +_5% variation in predicted N02M.
There was unavoidable confoundment between IHC and solar radiation and
solar radiation averages, with higher IHC leading to faster reaction.
Thus, events occur earlier in the solar day. Furthermore, the average
values of solar radiation used v.; this study are computed between
event? in the chemical system (such as between maximum N02 rate and
149
-------
Table 17. MAGNITUDES OF TERMS IN NITROGEN DIOXIDE
MAXIMUM REGRESSION EQUATION
= 0.4927 ([NOxJ0)°-844([HC]o)°-143 EXP(0.
Part Aa Part Bb
[N
0.
0.
0.
0.
0.
Table
10 )Q
007
122
222
361
510
17-A
Part A
ppm
0.006
0.0837
0.1383
0.2085
0.2791
Table
[HCJ
ppmC
0.31
0.79
1.06
1.41
1.99
2.48
2.70
2.90
3.44
4.00
17-B
PART B
0.846
0.967
1.01
1.05
1.10
1.13
1.15
1.16
1.19
1.21
Part Cc
17-C
SR3/SRy°PART C
1.804 0.9316
1.8575 1.000
1.928 1.0556
Part A expresses the_dependence of [N09] on [NO ] for [HC] = 1.0
ppmC and a ratio of SR./SR = 1.8575. ^U?ix[NO ] to°look up Part A
in Table 17-A J ' x °
Part B is the multiplier factor for [HC] other than 1.0 ppmC. Use
Tab]^ 17-B and [HCl to find multiplier. [N0?] = (Part A) x (Part B)
at SR-/SR = 1.8575° x
cPart C is the multiplier factor for solar radiation variation. Use
ratio in Table 17-C to find multiplier factor.
SR-. is the average solar radiation from the hydrocarbon injection
until the NO-NO2 crossover. SR3 is the average solar radiation from
the occurrence of the N0~ maximum formation rate to the occurrence of
the N0? maximum concentration. Since the latter events usually occur
close together, SR3 is a good measure of solar radiation conditions
near the peak
eThe 0.931 value is typical of October solar radiation conditions.
The 1.055 value is typical of June solar radiation conditions.
150
-------
N02 maximum concentration] and not at fixed time intervals. The
average solar radiation values, therefore, have a time confoundment
that is related to the IHC value. Other methods of expressing solar
radiation are under investigation, but no entirely suitable method has
yet been found.
NIGHTTIME BEHAVIOR OF RESIDUAL NITROGEN DIOXIDE
Previous discussion has indicated that some runs with lower IHC have
final N02 concentrations (FN02), that is, N02 concentrations at 1700
hours EOT, that are higher than those with higher IHC (see Figures 68
and 69). Table 18 is a pairwise (side-to-side) comparison of FN02
values for various INOX and IHC and is a companion table to Table 14
(the same dual runs were used). As indicated in Table 18, the FN02 on
the low IHC side was usually higher than the high IHC side and sometimes
was slightly lower. In those cases where FN02 was lower, the N02 did
not reach a true peak, that is, there was incomplete conversion of NO
to N02 by 1700 hours.
Daytime Conditions and Nighttime NOp
In a few runs, the chamber was not exhausted after 1700 hours and the
instruments continued to record data on strip chart recorders (the
computer stopped at 1700 hours). Figure 71 shows the NO and 0.,
A O
profiles from the June 16, 1974 dual run in which N00 was recorded
during the night. Ink problems with the 03 strip chart prevented the
recovery of 0., data for the nighttime period. Figure 72 shows the
solar radiation profile for this day, and although it was not a very
clear day, there are periods of very high SR, approximately equivalent
to ka for N02 of 0.57 roin . At 1630 hours, there was a rapid decrease
in SR due to a summer squall (intense, short m'nstorm) which rapidly
lowered temperatures and light.
The N02 behavior exhibited by the dual run of June 16, 1974 was not
151
-------
0 —1
CO LU
DC OO
e^
o Di
o o
fy i |
0
>- z
DZ O
1— (
1 J—
e^ <^
t— I fV^
1— I—
Z LU
i— i O
•z:
C3 O
Zi <_>
1 — 1
O LU
•^ 0
O i — i
LU X
o: o
Lu 0
o
OO LU
1— C3
0 O
LU Ct:
U- h-
Lu i— i
LU Z
LU _l
0 <
"^
•Z. i— i
O Lu
OO
i — i ~z*
0£ 0
e^
Cu "Z.
2: o
O 1-1
0 h-
LU Qi
OO 1 —
t— 1 ^3^
3 LU
rV o
i— i Z:
< 0
Cu 0
•
co
i —
CD
f—
_Q
(X3
rv*
LU
CO
2:
rc
0
fV
o
o
o
t—
~~1
o
re
o E
z: o.
ZD D-
3^ **
i— i CM
o
X Z
1 — 1
"re
2: c
O -r-
CQ Lu
a:
^^
o
o
CD
in
re
cu
t-
o
c
1—*
3*
o
_J
c*
en
IE
o:
Q
>-
~T*
"^
c^
03
a:
~~i
o
3Z E
I— a.
1— 1 Q.
3
«
OO C_)
py "T"
m>
a: i—
n3
r \ «r-
< +J
a c
i — i
1
0
•™i
-a
0)
o:
^-^
^
o
— I
J^"
CD
•r-
HI
c
3
^"
^5 •
^"
CM
Lf>
Qj ^
"O r~~
•r— •
oo o
CO
CO
LO
T3 c—
CO CD
c en
o
•r" ^f
-t-> LO
•*,
^J-
^^
CM
f*^
f~m
•
o
en
vQ
f—.
I—
•
o
o
CD
LO
o
•
r_
o
CO
•
CM
,—
r-*
**o
0
^J-
o
•
LO
CO
UD
o
en
r—
•
o
o
CO
o
r—
•
o
CM
•
LO
LO
0
CO
•
l-~
o
<7l
•
CM
CM
r-~
(£>
o
«^
r--.
•
CM,
oo
oo
o
CM
•
o
o
CO
en
^_
•
o
en
LO
CM
•
0
^_
r—
•
CM
en
CM
to
o
«3-
p-%
•
CM
1
co
r—
CM
CM
•
o
LO
oo
LO
CM
•
o
en
CM
o
•
o
CO
LO
•
CM
vo
r—
VO
o
*d"
CO
r—
CM
1
CM
LO
^_
CM
•
o
CM
LO
p^
CM
• •
o
o
CO
VO
o
00
0
o
LO
•
CM
o
CM
o
-------
unique. Another dual differential IHC run on June 12, 1974, Figure 73
showed the expected high FNOp in the lower IHC run. Unlike the low
IHC run on June 16, however, the June 12 low IHC run exhibited a
significant deterioration in NOp after 1700 hours. See Figure 74 for
solar radiation for June 12.
Analysis of the two lower hydrocarbon runs for June 12 and 16 indicated
that the major difference was the concentrations of 03 and NO in the
afternoon. The lower IHC side on June 16 produced only 0.02 ppm ozone
and had elevated evening N0? concentrations while the lower IHC side
on June 12 generated about 0.10 ppm of ozone and experienced a decline
in N02 after 1700 hours.
Several other runs of this type were made and these followed the same
trends as the June 12 and 16 runs. Data for these runs are given in
Table 19. Two of the runs listed in Table 19 had very similar initial
conditions (June 12 BLUE and September 20 BLUE, Figures 73 and 75) and
yet one of these (June 12) had relatively low nighttime NOp while the
other (September 20) had high evening N0?. Examination of the solar
radiation conditions for these days reveals that much more favorable
reaction conditions were present on June 12. The total integrated
solar radiatii
September 20.
p
solar radiation values were 466 and 364 cal-cm for June 12 and
It seems apparent that different combinations of meteorological and
chemical conditions can result in high nighttime NOp in outdoor smog
chambers. These conditions have in common low afternoon concentrations
of ozone and fairly high afternoon NOp and NO. The contribution of
nighttime NOp to the total NOp dosage can be very substantial in
certain cases. In most of these experiments nighttime NOp constitutes
from 25 to 35% of the total 22 hour NOp dosages. In the runs with
high evening NOp, the contribution of the nighttime NOp to the total
22 hour dosage was of the order of 60%. These days also had signifi-
cantly higher overall NOp dosages.
153
-------
I I I I I I I
•«• X
M
o
J3
(O
CD U
— °
I -O
: £
oo
LO
*s rO
CO JD
r— 1-
un 3
o o
T I
X O
o r-.
CD
LO
LO
*
CD
0
CD
if)
00
CD
00
O CD
uJdd
LO CD LO CD
CXI OJ »-H '"•
CD CD CD CD
D 'eON 'ON
LO
CD
CD
CD
CD
in ^~
c i
O I
•i- I
o 8
•!—
co £
§"
§^
r>.
CO
10
O)
c
<+-
o
c
3
&_
•a
(D
T3
c
CD
-P
X
LU
cu
ZJ
en
CO
IT)
CVJ
a.
a.
i.
LO O)
O JD
*~I »0
o .c:
o
oo
o
o
a>
trt
s-
o
o
•o
4J
rs
o
CM
o
154
-------
O
*
C\J
CD
OJ
CD
O
OD
•
O
sfr-
•
O
OJ
CD
O
*
CD
L o)
I I I I I I I I l
i I i I LO
O
OJ
T_u!Uj-2_ajo-|Do '
LO
CO
CM
C\J
(U
3
•-D
S-
o
c
O
- o 42
LU "S
s-
P- - $_
«-n CO
-------
un
ro
c
o
s-
-------
CD
CD
CM
o
oo
o
CM
O
O
CD
OO
O
UD
CD
cu
CD
CD
I ' I ' I ' I ' I ' I ' I ' I
L 5f
_ r-
H- &
— CM
»—i
LU
- 2
ID
1,1,1,1,1,1,1,1,1
CD
CD
•
C\J
o
00
CD
(M
CD
CD
CD
00
•
O
CD CD CD
^D =J- CM
LO
CO
CO
CM
CM
CD
in
CO
cn
LO
CD
CD
•
O
en
CM
cu
I"!
-J o
O H-
•M
(O
•5
«J
i.
ta
cu
3
157
-------
CM
CO
CO
c
ra
Q
Q LO
LO =*-
• •
O O
O
ct-
LO
CO
O
CO
LO
OJ
O
OJ
LO
CD
LO
O
O O O O O
-«o <2ON 'ON
o
CD
O
•^ r--
X O
CO I
c i
o i
c.
O
O
CO
c: o
i—i re
cu
CO
fy*
ID
0
31
—
A
C\J
0
C\J
i.
-
X
.,»
e
c
o
S-
(O
o
o
S-
•o
>^
^
158
-------
O
C3C3
OJ
CD
h1 I ' I '
L =t-
L_ r-
en
o
c\j
QQ
z:
LU
f—
Q_
LU
CD
I ' I ' I ' I ' I ' I '-
I i I i I i I . I . I . I . I . I
L/}
CO
CO
C\J
cr>
o
CM
+->
CL
-------
0)
Q
LU
0
LU
1—
X
LU
Ct:
0
Ll-
CO
LU
CD
CO
0
Q
LU
Q
i— I
X
0
1 — 1
0
LU
CD
O
Ct:
\—
1— 1
^y
UJ
i — i
I ,
n^
IE
CD
i — i
3^
f"i
Z.
LU
^^
^_
\ \
1—
ec£
Q
U_
Q;
LU
CO
^
IE
0
CD
O
CO
O
~zr
ID
2^
K — 1
X
1^
^^
•^
o
CQ
(Tf
e^
0
o
o;
Q
X
~y
^—
CO
~~"}
h-
cu
•r—
•P
-P
-C
Ol
•r™
'1
O
O
•r-
-P
(0
o:
cu
(O E
V)
O (O
o
LO
CM
0 -P
Z. (O
O)
cn
Q.
fO
•
E
Q- E
un Q.
to
CM E
O/-~)
1~L
O CO CO
CTl 00 LO
CM CM CM
CO CO CTi
**• o co
<3- «* CM
CM CM CM
«5j- CO CTl
CO CM CO
r»» r^ CD
CM CM CTi
r-* r^H co
o o o
o i— «d-
^*« r^ co
CM CM CM
o o o
co
CM
co
r—
f—
CO
CM
CTl
LO
f^
CM
CM
o
o
LO
CM
o
CTl
co
co
co
LO
LO
CM
co
^— •
co
o
«=!-
o
LO
CO
CM
0
o
CO
co
,_
t —
LO
co
CM
CM
CO
CM
o
o
^-
CTl
o
CM
co
CM
,_
o
LO
CO
«*•
00
LO
CM
^*
CM
o
CM
^^
co
o
CTl
^
LO
LO
CM
o
^
^»
-------
NOp Removal Process
t "
It seems reasonable that a partial explanation for the consumption or
disappearance in nighttime NCL on June 12 (in the lower IHC side)
could be the reaction between 0- and KCL,
In order to assess the importance of the 03 and NCL reaction in a
competitive situation, a photochemical model was run using the residual
concentrations of all products from two variable sunlight simulations
as starting conditions in two nighttime simulations. The two daylight
simulations were designed to yield different concentrations of end
products so as to demonstrate the variation typically occurring in the
ambient atmosphere at the end of the day. In the first sunlight
simulation the final (600 minutes) concentrations were NO,, = 0.209 ppm
and 03 = 0.468 ppm with a minimal amount of NO, 0.006 ppm. In the
second sunlight simulation the concentrations were N02 = 0.331 ppm, NO
= 0.081 ppm and 03= 0.059.
The only reactions having any significant effect in the second night-
time simulation were:
2: NO + 03 + N02 + Q2 k = 24.0 ppm'^min"1
31: N02 + 03 -> N03 + 02 k = 0.046 ppm'1-min"1
For NO- and NO to have a similar demand for 03, the NO^/NO ratio must
be greater than 500 : 1. The model N02/N0 ratio was only 4.09, indica-
ting that most of the available 03 would react with NO to form N02-
In this simulation, all of the 03 and NO combined to form N02 by 6
minutes into the dark phase run with virtually no reaction between NOp
and 03>
The dark phase reaction using the first set of concentrations showed
161
-------
somewhat different results. Since very little NO was available, a
substantial concentration of 0, remained after all the NO had reacted.
The remaining 03 therefore, reacted with N02 as shown below.
31: N02 + 03 + N03 + 02 K31 = .046
NO, t NO, + N,0, Kf = 55QO PP™"!
3 2-25 K = 17.1 min'1
H20 r
N205 ^ 2HN03 K35 = .01 min '
Within 100 minutes into the run 95% of-the N02 disappeared to form
N03, N205, and HN03- The remaining amount went to N20g (0.12%) and
N03 (5.62%). During the entire 600 minute dark phase run, a total of
169 ppm N20r was formed. Most of this decomposed by the back reaction,
but ample opportunity was available to form HN03.
Hydrocarbon Reduction and N0? Dosages
r --£
Hydrocarbon reduction had a tendency to lower, broaden and extend N0?
peak concentrations. The June 12 run (Fig. 73) had N02 peak values
above 0.24 ppm for 2 hours on the higher IHC side while no N02
concentrations on the lower IHC side exceeded this level. The lower
IHC side had 8 hours in excess of 0.15 ppm compared to about 5 1/2
hours on the higher IHC side. The number of hours in these runs
exceeding selected concentrations are given in Table 20. The September
20 BLUE run had high evening N02 and the highest overall dosage.
Although there were not any N02 concentrations above 0.24 ppm in this
run, there were almost twice the number of hours above 0.19 ppm and
0.15 ppm N02 than in the others.
Differences can also be observed between runs which had approximately
the same overall N02 dosage. Due to day-to-day variations in solar
radiation, however, it is useful to look at hydrocarbon reduction from
162
-------
i
o
co
00
§
in
o
IO
co
CO
co
f.
10
1
£?
S
£
UJ
s
o
JD
ic
tn
10
X
CM
O
JC
u
i
e
M-
VI
J^
J
O
z
i
o
in
,
0
r- ro 10 CM o
o en «• o »
in * in oo f-.
i
o
r<_
•
o
CM r* o r* r^
in »»• oo w i—
• • • • •
^^ ^^ <^ tft if*
i
a
en
.
o
SCO CO IO *O
^ "~. ^ ^
co co *& ^ in
U
a
CM
CM
1*^ CO C\l O
CO 00 r- 1 9
. . 1
CM CM CM CO
^^
0
WD
^
o
*0
•
^f
V
IO
*
00
1
1
CM
I
SE
£
t—t
a
o
x
CM "
CM
oo oo
<*> co
in
CO
en
-------
dual runs (i.e., performed on the same day) as opposed to a comparison
across all days based on initial hydrocarbon values. The dual run on
June 12 is suitable for such a comparison. Reduction of initial NMHC
from 2.9 to 1.3 ppmC (under the same solar radiation and final NCL
dosage) was associated with a 28% increase in concentrations above
0.17 ppm and a 50% increase in NO above 0.13 ppm.
164
-------
SECTION VIII
OTHER RESULTS AND OBSERVATIONS
EFFECT OF INITIAL PERCENTAGE NITROGEN DIOXIDE
The ratio of N09 to NO (PCTN02) at the initiation of an experiment is
L. X
an important parameter to investigate in terms of the effects of NMHC
on NO to N0~ conversion. Most experiments discussed in this report
have a constant ratio of N09 to NOV equal to 0.20 (that is, 20%). In
C, J\
a few selected runs however, the PCTN02 was varied to determine its
effect on NO to NOp conversion. In these differential runs, each side
had the same IHC and INOX but had a different initial PCTN02 (e.g.,
20% and 40%). Experiments on May 8, 1974 (Figure 77) and July 15,
1974 (Figure 78) were differential PCTN02 runs that occurred on clear
days; runs on May 4, 1974 (Figure 79) and October 14, 1974 (Figure 80)
occurred on days with erratic solar radiation.
Higher PCTN02s in differential runs on clear days appear to provide
more optimum conditions for 0, formation; chemical events occurred
earlier in the day and 0- formation therefore occurred under higher
light conditions in the higher PCTN02 run, thus leading to higher
ozone concentrations.
In differential PCTN02 runs conducted on days that had choppy solar
radiation conditions, the outcomes were very much a function of the
detailed light-history (see discussion of Figure 62 in N02M discussion).
Figure 79 shows a PCTN02 run in which the more typical 0- behavior of
May 8 and July 15 was reversed and the lower PCTN02 run formed approxi-
mately 2.5 times more ozone than the higher PCTN02 run.
In the October 14 experiment, the higher PCTN02 run was slower than
would be expected based on the clear day run results. Both October 14
and May 4 had erratic solar radiation conditions; the exact effect of
165
-------
E
o.
o.
O
I ' I ' I ' I ' I
MRY 8.
T-r-l ..1,1,1,1
8
15 .16 17
0.0
9 10 11 12 13 H
HOURS. EOT
Figure 77. Differential percentage N02 run under clear sky- Initial conditions: NO (—) 0.364,
(---) 0.366 ppm; N02(—) 0.151, (---) 0.079 ppm; NHHC(—) 3.40. (—) 3.20 ppmC urban
hydrocarbon mix in UNC outdoor smog chamber.
E
O.
o.
o
tS
8
15 16 17
Figure 78.
9 10 11 12 13 H
HOURS. EOT
Differential percentage N02 run under clear sky. Initial conditions: NOX(—) 0.363, (—)
0.360 ppm; N02(—) 0.130. (—) 0.076 ppro; NMHC(—) 1.49, {—) 1.52 ppmC urban hydro-
carbon mix in UNC smog chamber.
166
-------
E
o.
o.
o
8
13 H 15 16
9 10 11 12
HOURS. EOT
Figure 79. Differential percentage N02 run under erratic solar radiation conditions. Initial
conditions: NOX(—) 0.372, (---) 0.372 ppm; N02(—) 0.151, (—) 0.076 ppm; NMHC
(—) 2.82, (—) 2.82 ppraC urban hydrocarbon mix 1n UNC outdoor smog chamber.
E
O.
o.
o
0.5
O.f
OCTOBER H, 19/f _
8
13 H 15 16
Figure 80.
9 10 11 12
HOURS. EOT
Differential percentage NO, run under erratic solar radiation conditions. Initial
conditions: NOX(—) 0.241, (—) 0.244 ppm; N02(—) 0.049, { —) 0.086 ppm; NMHC(—)
1.05, (---) 0.92 ppmC urban hydrocarbon mix in UNC outdoor smog chamber.
167
-------
these conditions appears to be a function of when in the chemical
system the light intensity changes and the direction in which it
changes. Additional investigation of PCTN02 will further clarify its
participation in NO to NCL conversion under different physical condi-
tions.
EFFECT OF DILUTION ON N0g AND OZONE FORMATION
Selected experiments were conducted to investigate the effect of
dilution on NO to N02 conversion and ozone formation. These experiments
were designed such that the same initial concentrations of nonmethane
hydrocarbons (NMHC) and oxides of nitrogen (NO ) were introduced into
A
both sides of the outdoor chamber. One side remained sealed off from
the ambient atmosphere while the other side was slowly diluted with
background air starting in the morning. Two different dilution vs
static smog runs will be discussed: a) using propylene as the hydrocarbon
which was relatively fast, and b) another system which used the urban
hydrocarbon mix.
In the propylene experiments, both chamber sides had identical reactant
profiles until 0705, when dilution began in the BLUE side. Propylene
N0? and NO were then slowly removed from the BLUE chamber by the
dilution system. At the same time however, N0? was being generated at
a faster rate than it was being removed by dilution, and actually
appeared almost as fast as N02 in the RED (undiluted) chamber. This
created a situation in which the ratio of NO^/NO was higher in the
BLUE chamber than the RED. If one assumes that pseudo photostationary
equilibrium can be used to approximate the instantaneous Q., concentra-
tion (Equation 1) then the ratio of NO^/NO governs 03 concentration at
a given light intensity.
[NOJ _ k
[NO] k1
168
-------
I I I I I I—I—1—I—I—1—I—1
0.00
6 7 8 9 10 11 12 13 11 15 16 17
HOURS. EOT
Figure 81. Example of dilution In a reactive system. Initial conditions: NO (—) 0.450, (—)
0.460 ppm; N02(—) 0.043, (-"--) 0.050 ppn; NMHC(—} 2.76, (—) 2.63 ppmC propylene
In UNC outdoor smog chamber.
0.5
0.1 -
0.3 -
e
o.
o.
o
c5
z 0.2 -
0.1
0.0
OCTOBER 7. 1971 HC MIX - NOx —
0.5
o
- 0.3
— 0.2
XI
3
= 0.1
0.0
6 7 8 9 10 11 12 13 11 15 16 17
HOURS. EOT
Figure 82. Example of dilution in a less reactive system. Initial conditions: NOX(—) 0.504,
(- —) 0.512 ppm; N02(—) 0.106, (—) 0.111 ppm; NMHCf—) 2.39, (---) 2.43 ppmC urban
hydrocarbon mix 1n UNC outdoor smog chamber.
169
-------
Where k-, is the photolysis rate constant for NO,, and k~ is the rate
constant for the reaction of NO and Oo.
In the reactive system (Figure 81), ozone was initially generated more
quickly in the diluted side. However, NO was being removed by dilution
A
and the potential NO,, peak was therefore greater in the undiluted side
during a period of the day when the solar intensity was greatest
(i.e., greater value for k,) and was accompanied by a faster ozone
generation rate and higher ozone concentrations.
Analysis of the less reactive smog system can be made in a similar
manner. Both sides initially contained 2.4 ppmC of urban hydrocarbon
mix and 0.5 ppm NO (20% NO,,). Dilution began in the RED side about
X £-
0900 hours (Figure 82) and soon after, this side was generating ozone
faster than the static side. The major difference between this less
reactive system and the reactive propylene system was, for the less
reactive system, a higher ratio of N0?/N0 was not reached in the
undiluted side until the sun was beginning to go down. The net effect
was lower and lower values of k, during the greatest period of potential
ozone generation for the undiluted side.
In summary, the effect of different dilution rates and time at which
dilution begins on 0- cannot be presently quantified. From the two
runs discussed however, extrapolation to the general case may be
possible. Reactive systems, which would normally achieve maximum
oxidant values during the early afternoon, may initially go faster
under dilution conditions but probably will not reach the oxidant
levels generated by static systems. By contrast, less reactive systems
under certain dilution conditions may actually generate more oxidant.
In the experiment considered (Figure 82), the diluted side actually
exceeded the National Ambient Air Quality Standard for oxidant while
the undiluted side did not. Greater hydrocarbon reduction compared to
NO reduction will have the tendency to shift the urban atmosphere to
X
one containing less reactive mixtures. Dilution in smog chamber
170
-------
experiments with these less reactive mixtures may prove to be important
in assessing the worst case "condition".
SECOND DAY IRRADIATION
The dual run of June 13, 1974, discussed previously under nighttime
behavior of N02, actually was a 36 hour run which included two daylight
periods. The results on the second day were not expected. The 36-
hour profiles for N02 and 03 are given in Figure 83. The NO profiles
are also given, but only for the first 12 hours and the last 7 hours.
The values for NOp and 0^ between 1700 hours June 12 and 1000 hours
June 13 were from digitized strip chart ; NO was not recorded on
strip charts.
The experiment of June 12 represented a situation in which hydrocarbon
control of ^50% occurred (from 2.9 ppmC to 1.3 ppmC) at constant INOX
of 0.36 ppm. This hydrocarbon control affected both the N02 and 03
maximum concentrations; approximately a 10% reduction in N02M and 74%
reduction in 03M resulted during the first daylight period. Because
both sides had 0, > 0.1 ppm , the N02 concentrations decayed throughout
the nighttime period to values slightly below the NAAQS of 0.05 ppm.
During the second daylight period, the 03 increased in both sides
while the N0? remained relatively constant at 0.045 to 0.050 ppm. The
peak Oo values in this second daylight period, with no additional
material added, were 0.40 and 0.37 ppm. Thus, the reduced NMHC run,
which barely exceeded the NAAQS for 0, in the first daylight period,
exceeded the standard for ozone in the second daylight period by a
factor of four. The higher NMHC run produced essentially the same 0-
in both daylight periods, also exceeding the NAAQS for ozone by a
factor of four. This occurred even though the N02 concentration was
below the NAAQS for N02 and the "reactive" hydrocarbon compounds had
presumably been consumed the previous day.
171
-------
LO CD
CD CD
CD
LO
LO
cf
CD
O
LO
CO
CO
LO
CM
CD
C\J
LO
LO
CD
CD
CD
CD CD O
tudd '*[)
O
O
00
CO
fl3
o
Q.
Q.
O
co
X i—
in
e:
o
o
o
(O
Q ~
o
CM
0
•
^^
r^
2
•»
CO
i—
i
CM
r—
cu
c
3
'"^
o
E J3
Q. t-
Q. (O
O
r^ o
in f-
co -a
o ^r
'ON
172
-------
CD
O
OJ
O
CO
UD
CD
OJ
CD
CO
CD
OJ
CO
OJ
UJ
CD
CD
CD
CD
O
CO
CD
OJ
CD
O
O
00
CD
it-
r-
c—t
LO
»—I
CO
«—I
«—I
«—I
en
r-
LO
CO
^ o
UJ ^
CO - o»
OJ CO ""
- §5 "
^ g s
en 7
3
'-3
5-
O
M-
C
O
•r—
-P
(TJ
T3
03
OJ «-«
CD
OJ
*
CD
CD
O
rtJ
4->
O
00
s_
yuios
173
-------
This system might represent the behavior of a .stagnant air mass which
could give rise to the so called "Sunday effect" in which 03 concentra-
tions on Sundays are often as high as those on weekdays. It would be
interesting to determine if fresh material were injected into the
system that existed on the second day, would the CL concentration have
been any higher than what occurred?
This system might also give insight into transport effects on 03
concentrations. If one assumes that dilution would have only linear
effects on the concentrations, the results shown in Figure 84 indicate
that the air mass, even on the reduced NMHC side, might be capable of
generating 0., in excess of the NAAQS, downwind on the second day.
j. -3
TEMPERATURE AND HUMIDITY EFFECTS
Table 21 contains average air temperature information at various
events during a run. The mean values across the table provide some
indication of the degree of temperature variation that occurs during a
run in the outdoor chamber. The maximum and minimum values illustrate
the wide seasonal range that occurs from early spring to late fall.
There is some confoundment between solar radiation and temperature, in
that, as solar radiation increases during a day, so does air temperature,
and although solar radiation decreases fairly rapidly toward sunset,
the back radiation of the earth tends to maintain the air temperature.
The coolest part of a 24 hour day is usually just before sunrise and
the initial temperatures in Table 21 are near the minimum daily air
temperature.
Wide variations in initial temperature have no clearly discernible ef-
fects on the reaction profiles, provided the midday temperatures are
above ^ 70°F. Cases in which the temperature at the NO and NCL
maximum rate (usually near the NQ-NCL crossover and always before the
N02M) fs less than 70°F do show a temperature effect on the rate of NO
174
-------
Table 21. Air Temperature Statistics at Various Chenlcal Events in Outdoor Chamber
Air Temperature, *F
Statistic
MEAN
Std.Dev.
Nln
Max
1 Cases
Initial
55.9
15.0
17.3
81.5
122
NO-N02
Crossover
76.4
8.9
51.0
92.6
114
N02
max Rate
74.8
8.9
48.6
93.3
118
N02
79.3
9.1
49.9
91.9
120
03
maximum
80.4
8.4
53.8
92.7
120
Final
79.4
9.2
49.9
92.5
120
Dally
Average
74.3
9.7
44.4
88.8
120
0.6 1 1 1 1 j 5 1 1 1 1 1 1 - i 0.6
c 0.5 |- /r+ -\ O-5
E
UJ
§ 0.3
ac
§ 0.2
i 0.1
I
I
0.3
0.2
0.1
0.0
0 0 ' 1 ' ' 1 « 1 1 1 1 1 i 1 [
50.0 55.0 60.0 65.0 70.0 75.0 80.0
DEGREES FflHRENHEIT
Figure 85. Effect of air temperature at maximum rate on maximum nitric oxide disappearance for clear
days for November 1973 and initial conditions of 0.20 ppm N0x> 0.04 ppm N02 and 0.80
ppmC urban hydrocarbon mix in UNC outdoor smog chamber.
175
-------
disappearance and on the CL maximum. Figure 85 shows a 50% reduction
in the rate of NO oxidation for a 14°F drop in temperature. These
cases are from closely spaced replicate runs under clear skies in
November, 1973. The coldest run is approaching the NO oxidation rate
of hydrocarbon background runs (0.2-0.3 ppb/min). A 0.24 ppm
NO /background hydrocarbon (<0.2 ppmc) run in June had a NO oxidation
A
rate of 0.31 ppb/min and made 0.007 ppm 03 at a temperature of 89.9°F.
As the NO oxidation rate decreased in the runs shown in Figure 85, a
reduction in the 03 maximum occurred. The same chemical conditions in
June produced 0., maximum values of 0.17, 0.10, and 0.06 ppm 03 at
temperatures of 90, 84, and 89°F respectively. The runs in Figure 85
produced 0.06, 0.05, 0.03, 0.03, 0.014, and 0.012 ppm 03 at 03 maximum
temperatures of 80, 80, 70, 69, 53.8, and 53.4°F respectively. These
data suggest that there might be a lower temperature limit (^55°F) at
which the 03 formation rate is so low as to preclude ambient concentra-
tions from reaching the National Ambient Air Quality Standard of 0.08
ppm. This could be due to the very slow oxidation rate of NO at lower
temperatures. These observations have been supported by an exami-
nation of field data collected by the Research Triangle Institute.
No effects of relative humidity have been observed in the range of 40%
to ^ 85% relative humidity. The data exhibits no perceptible trends
with relative humidity including those runs that have essentially the
same chemical, temperature and solar radiation conditions, but different
relative humidity.
176
-------
REFERENCES
1. "National Primary and Secondary Ambient Air Quality Standards,"
Environmental Protection Agency, Fed. Register, 36, No. 84,
p. 8187 (April 30, 1971).
2. Shy, C., et al., "The Chattanooga School Study, Effects of Community
Exposure to Nitrogen Dioxide," J^. Air Pollution Contr. Assoc.,
20, p. 539 (1970).
3. "Air Quality Criteria for Nitrogen Oxides," Air Pollution
Control Office, Publication No. AP-84. Environmental Protection
Agency, U. S. Government Printing Office, Washington D. C., 1971.
4. Korth, W., Rose, A., and Stahman R., "Effects of Hydrocarbon to
Oxides of Nitrogen Ratios on Irradiation Auto Exhaust," J_. Air
Pollution Contr. Assoc.. 14, p. 168 (1964).
5. Dimitriades, B., "Oxidant Control Strategies. Part I. An Urban
Oxidant Control Strategy Derived from Existing Smog Chamber Data".
Paper submitted for publication in Environ. Sci. and Techno!.
6. "Proceedings of the Conference on Health Effects of Air Pollution,"
Committee on Public Works, United States Senate, U. S. Government
Printing Office, Washington, D. C. Stock No. 5270-02105 (1973).
7. Altshuller, A. P., "Relationship of HC and NO Concentrations
X
and Reactants Ratios to Federal Air Quality Standards for
Photochemical Oxidant and Nitrogen Dioxide," Presented to
Committee on Motor Vehicle Emissions, Panel 1, National Academy
of Sciences. (1972).
177
-------
8. Dupont Technical Information Bulletin T-5, "Optical Properties."
9. Dimitriades B., "Methodology in Air Pollution Studies Using
Irradiation Chambers," J_. Air Pollution Contr. Assoc., 17.,
pp. 460-6 (1967).
10. Dupont Technical Information Bulletin T-3D, "Chemical Properties."
11. Dupont Technical Information Bulletin T-6B, "Weatherability
Performance."
12. Relme, K. A., Martin, B. F., and Hodgeson, J. A., "Tentative
Method for the Calibration of Nitric Oxide, Nitrogen Dioxide and
Ozone Analyzers by Gas Phase Titration." EPA, No. 2-73-246
(March, 1974).
13. Saltzman, B. F., Burg, W. R., and Ramswany, G., "Performance of
Permeation Tubes as Standard Gas Sources," Environ. Sci. and
Techno].., 5., pp. 1121-1128 (1971).
14. Saltzman, B. F., "Determination of Nitrogen Dioxide and Nitric Oxide,"
In: Selected Methods for the Measurement of Air Pollutants,
U. S. Dept. of H.E.W., Public Health Service Publication
No. 999-AP-ll (1967).
15. Coordinating Research Council, CAPI-6 Committee, New York,
CRC Report No. 398, "Individual Hydrocarbon Reactivity Measure-
ments: State-of-the-Art" (June, 1966).
16. Dimitriades, B., "Effects of Hydrocarbon and Nitrogen Oxides on
Photochemical Smog Formation," Environ. Sci. and Techno!., 6_,
p. 253 (1972).
178
-------
17. "A Critique of the 1975-1976 Federal Automobile Emissions Stan-
dards for Hydrocarbons and Oxides of Nitrogen," Report of Panel 1
on Standards to Committee on Motor Vehicle Emissions, National
Academy of Sciences, Washington D, C., pp. 11-12 (1973).
18. Larsen, R. I., "A Mathematical Model for Relating Air Quality
Measurements to Air Quality Standards," Preliminary Publication,
Environmental Protection Agency, Office of Air Programs, No. AP-89,
Research Triangle Park, N. C. (1971).
19. Kopczynski, S. L., Personal communication to Harvey Jeffries,
National Environmental Research Center, Research Triangle Park,
North Carolina, June, 1972.
20. Petterssen, S., Introduction to Meteorology. McGraw-Hill Book
Company, New York, pp. 286-7 (1958).
21. "Local Climatological Data," U. S. Department of Commerce,
National Climatic Center, Asheville, N. C. (1974).
22. Sellers, W. D., Physical Climatology. University of Chicago
Press, Chicago, pp. 26-7 (1965).
23. Sickles, J. and Jeffries, H. E.,"Development and Operation of a
Device for the Continuous Measurement of Photolysis Rate of
Nitrogen Dioxide," Univ. of North Carolina, Dept. of Environ.
Sci. and Eng. Publication No. 396. (1975).
24. Joshi, S. D., "Comparison of Several Techniques for Determining
NO ," Univ. of North Carolina, Dept. of Environ. Sci. and Eng.,
Master's Thesis (1974).
179
-------
25. Leighton, P. A., Photochemistry of Air Pollution, Academic Press,
New York (1961).
26. Dinritriades, B., Proceedings of the Smog Chamber Conference.
(In press). U.S. Environmental Protection Agency, National
Environmental Research, Research Triangle Park, N. C. (1975)
27. Winer, A. W., et al., "Response of Commercial Chemiluminescent
NO-NCL Analyzers to Other Nitrogen-Containing Compounds,"
Environ. Sci. and Techno!. £, pp. 1116-21 (1974).
28. Miller, D. F., "NCL Measurements in Smog Chambers," Proceedings
of the Smog Chamber Conference, Research Triangle Park, N. C.
(1975).
29. Miller, D. F., and Spicer, C., "A Continuous Analyzer for
Determining Nitric Acid," Presented at Air Pollution Control
Association Metting, Denver (1974).
30. Jeffries, H. E., "An Experimental Method for Measuring the
Rate of synthesis, Destruction and Transport of Ozone in the
Lower Atmosphere," Univ. of North Carolina, Ph.D. Thesis.
Dept. of Environ. Sci. and Eng., Publication No. 285 (1971).
31. Bonamassa, F. and Wong-Woo, H., "Composition and Reactivity of
Exhaust Hydrocarbons from 1966 California Cars," Presented at
American Chemical Society Meeting, New York (1966).
32. Hampson, R. F., ed., "Chemical Kinetics Data Survey VI, Photo-
chemical Data for Twelve Gas Phase Reactions of Interest for
Atmospheric Chemistry," NBS Report, NBSIR 73-207 (1974).
33. Garvin, D., and Hampson, R. F., "Chemical Kinetics Data Survey
VII. Tables of Rate, and Photochemical Data for Modeling the
Stratosphere (revised)," NBS Report, NBSIR 74-430 (1974).
180
-------
34. Demerjian, K., Kerr, J. A., and Calvert, J., "Mechanism
of Photochemical Smog Formation," In: Advances in
Environmental Sciences and Technology, Vol. 4, Pitts, J, N,
and Metcalf, R. L., ed., John Wiley and Sons, Inc. New
York (1974).
35. Hecht, T. A., Seinfeld, J. H., and Dodge, M. C., "Further
Development of Generalized Kinetic Mechanism for Photochemical
Smog," Environ. Sci. and Techno!.. £, pp. 327-39 (1974).
36. Jaffee, R. J., and Smith, F. C., "Factors Affecting Reactions in
Smog Chambers," Presented at 67th Air Pollution Control Associa-
tion Meeting, Denver (1974).
37. Hecht, T. A., Liu, M. K., and Whitney, D. C., "Mathematical Simulation
of Smog Chamber Photochemical Experiments," Environmental Protection
Agency, Publication No. EPA-650/4-74-040 (1974).
38. Jaffe, R. J., "Lighting Effects in Smog Simulations I, Constant
Intensity vs. Diurnal Variation," Presented at 68th Annual Air
Pollution Control Association Meeting, Boston (1975).
39. Shuck, E. A., Stephens, E. R., and Schrock, R. R., "Rate of
Constant Ratios During Nitrogen Dioxide Photolysis," J_. of Air
Pollution Contr. Assoc., 16, p. 1965 (1966).
181
-------
SECTION X
GLOSSARY
Aerometric data - Data resulting from measurements of pollutant
concentrations in the ambient air.
ARB - California Air Resources Board.
Background run - A run in which the chamber was charged with only
rural, ambient air just prior to sunrise.
BLUE or BLUE side - see "side".
BOM - Bureau of Mines
Calibration equations - Equations used to calculate the substance
values in physical units from instrument
voltage output and/or substance values.
Crossover - The time at which the N02 concentration equals the NO
concentration. (Runs for which the initial N0£ exceeded
the initial NO do not have a crossover. Also, crossover
does not occur in extremely slow runs).
DAS - Data acquisition system.
.DAT files - Computer files at TUCC containing measurement data as
collected during a run plus comments of operator made
during the run.
Dilution run - A run in which the contents of the chamber are di-
luted with rural ambient air after the injection at
a constant flowrate. Chamber is run as a continuous
- flow stirred - tank reactor (CFSTR) with very low
reactant concentrations in the incoming dilution air.
Dosage - The area under a concentration versus time graph for some
specific time interval, in units of concentration x time.
Dual run - An experiment using both sides of the chamber to compare
the effects of differences in controlled variables (such
as initial concentrations) under identical solar radiation
and temperature conditions.
Greenhouse effect - (With respect to a smog chamber) the heating of
the gas within a chamber due to the transmission
of shortwave radiation into the chamber and
absorption of long-wave (infra red) radiation by
the chamber walls preventing radiative cooling
of chamber contents.
182
-------
HC/NOX - The ratio of non-methane hydrocarbons to oxides of nitrogen
present in the chamber initially, ppmC/ppm.
HP - Hewlett-Packard 9810A Programmable calculator with plotter.
IHC (initial hydrocarbon) - non-methane hydrocarbon concentration at
the time of completion of the hydrocarbon
injection.
Injection - The process by which the chamber is charged with reactants.
A solenoid valve is opened for a specified time allowing
gas to flow at a known rate from cylinders, through pre-
cision needle valves, and into the return line of the
sampling manifold. The return link transports injected
material to the chamber just below a mixing fan for rapid
dispersal of material.
INO (initial nitric oxide) - The nitric oxide concentration at the
time of completion of the hydrocarbon.
INOX (initial nitrogen oxides) - The concentration of NOX (essentially
NO and NO?) at the time of completion
of the hydrocarbon injection.
k1 - see "$ka"
$ka - The first order rate constant of N02 photo-dissociation.
Molar reactivity - The Jackson reactivity of a hydrocarbon mixture
computed by weighing the reactivities of the
individual compounds by their mole fractions in
the mixture.
NAAQS - National Ambient air quality standard(s).
NMHC - Non-methane hydrocarbons; total hydrocarbons minus methane as
measured by a flame ionization detector times an empirical
efficiency factor.
Ox - Photochemical oxidants
Oxidant standard - National air quality standard for photochemical
oxidants, measured as ozone. The primary and
secondary standards are both 0.08 ppm, a 1-hour
average not to be exceeded more than once per
year.
PCT N02 - see "percentage N02"
183
-------
Percentage N0? or PCT N02 - The initial N0? divided by the initial
NOX, multiplied by 100.
Percentage NC^ run - A dual run in which the only difference between
chamber sides initially is the percentage N0£
-- the initial NOX and hydrocarbons being nearly
identical.
o
pphm - Parts per hundred million; moles of pollutant per 10 moles of
gas.
ppm - Parts per million; moles of pollutant per 10 moles of gas.
QGA - Quartz globe actinometer; device for measuring the rate constant
of N02 photo-dissociation.
RED or RED side - See "side".
Run - An experiment performed in one side of the chamber with data
collection from 0500 to 1700 Eastern Daylight Savings Time.
Run identification - For the purpose of data processing each run has
been assigned a 7-digit alphanumeric word of the
form: YYMMDDS where YY is the last two digits of
the year in which run was performed, MM is the
number of the month in which run was performed,
DD is the day of the month on which run was per-
formed, S is the side of the chamber on which run
was performed, either R for red or B for blue.
SAS - Statistical Analysis System, a collection of statistical computer
programs.
S.C.O.T. - Supported column, open, tubular.
o
Side - The chamber is divided into two 6000 ft compartments, which are
arbitrarily designated as the "BLUE" side and the "RED" side.
.SMTH files - Computer files at TUCC containing smooth estimates of raw
data, the derivatives and integrals of smoothed estimates
plus operator comments.
SPSS - Statistical Package for the Social Sciences, a collection of
statistical computer programs.
Static run or static operation - A run in which the chamber is operated
as a batch reactor: reactants are
charged before sunrise, the chamber is
then closed, and reaction takes place
without further input of material.
184
-------
Substance - (With respect to data acquistion) any of the chemical
quantities or physical properties being measured.
THC - Total hydrocarbons, i.e., methane plus non-methane corrected by
an empirical efficiency factor measured by a flame ionization
detector.
ISO - Time Sharing Option; an interactive system for on-line computa-
tion.
TSR - Total solar radiation.
TUCC - Triangle Universities Computation Center.
UNC - University of North Carolina at Chapel Hill.
185
-------
SECTION XI
APPENDICES
Page
A. Standard Nomenclature Conventions Adopted 187
B. Equations Used in Computer Programs 190
C. Photochemical Model for Propylene 194
D. Statistical Assessment of Ozone Calibration
Procedure 200
E. Device for the Continuous Measurement of
$k for Nitrogen Dioxide 204
186
-------
Appendix A
STANDARD NOMENCLATURE CONVENTIONS ADOPTED
The following conventions were used to name variables considered in data
analysis:
Main Components
NOX - total nitrogen oxides (NO + N02), ppm
NO - nitric oxide, ppm
NG? - nitrogen dioxide, ppm
HC - nonmethane hydrocarbon, ppmC
TIME - mins after 0500 EOT
TEMP - air temperature, °F
2
SR_ - total solar radiation (2800°A-25,000°A), langleys/min = cal/cm /min
2
UV_ - ultraviolet radiation, millilangleys/min = millical/cm /min
(2800°A-3900°A)
I_TP - integrated temp, °F-mins
2
I_SR - integrated total solar radiation, Langley = gmcal/cm
o
IJJV - integrated ultraviolet radiation, millilangleys = milligmcal/cm
D_ccc - dosage of species ccc, concentration • time
Prefixes & Suffixes (main component indicated by ccc)
ICCC - initial, e.g. INOX - initial NO , IN02 - initial NOp
Xccc - crossover where NO = N0~, e.g. XNOX - NO crossover cone, XHC-HC cone
or c x
ccc_X at NOY crossover point
* /\
cccR - rate, ppb/min; with no other modifier, indicates a max value
e.g. NOR - maximum NO rate 03R - maximum 0^ rate
187
-------
cccM - maximum concentration, ppm; e.g. NOgM - maximum concentra-
tion when a main component name is used as a prefix it
indicates "at the time of"
ccc__INT - initial, just after the HC injection is completed
ccc_FIN - final, the time of the last recorded data or 1700 which ever
is earlier
ccc_DIFl - delta or difference between initial state and crossover
condition
ccc_DIF2 - delta or difference between cross condition and NO- max rate
ccc_DIF3 - delta or difference between NO max rate condition and N0?
max cone 2 ^
ccc_DIF4 - delta or difference between N02 max cone condition and 0,
max rate condition
ccc_DIF5 - delta or difference between 0 max rate condition and Oj max
cone condition "*
ccc_DIF_F - delta or difference between init and final
or
ccc_DIF_T
ccc_AVl - same as DTI thru DT5, except average condition between events
or
ccc_AVEl
thru
ccc_AV5
or
ccc_AVE5
ccc_A23 or ccc__AV23 - average across events 2 and 3 as defined in DTI
definitions i.e., from crossover event to max
N02 event
ccc_A45 to ccc_AV45 - average across events 4 and 5 as defined in DT4
and DT5 definition i.e., from N02 max cone to
0^ max concentration
ccc_NOR - at the time of maximum rate of change of the NO concentration
ccc_N02R - at the time of maximum rate of change of the N02 concentration
188
-------
ccc_NM - at the time of the maximum NOg concentration
or
ccc_N02M
ccc_03R - at the time of the maximum rate of change of the 63 concentra-
tion
ccc_03M - at the time of the maximum Oo concentration
189
-------
APPENDIX B
EQUATIONS USED IN COMPUTER PROGRAMS
TRANSFER EQUATIONS:
All underlined constants can be modified during run, and represent de-
fault values. S represents physical units and V represents measured volt-
ages.
Ozone: SQ3 = 1.0 (V03) + 0.0
Nitric Oxide: SNQ = 1.0 (VNQ) + 0.0
Nitrogen Dioxide: SNQ2 = 1.0 (VNQ2) + 0.0
Oxides of Nitrogen: SNQ = 1.0 (VNQX) + 0.0
Humidity (Dew Point): S = 4.724 (V ) + 0.0185 (VDp)2 +
0.000242(VDp)3 - 40.0
Methane: S^ = fLpJS^n-l) + l_.0_(VcH4)+0.0]/n} +
0.0
where n is number of readings; n can
be reset to 0 during the run
Non-Methane Hydrocarbons: SNMHr = 1.613* [1.0(V ) + 0.0 - S,, ,.] +
mini, THC cn4
0.0
Total Hydrocarbons: S = 1.0 [S + S ] + 0.0
I nt Ln'f INMHL
Carbon Monoxide: SCQ = 2.0 (VCQ) + 0.0
Temperature: S = 100.00 (V ) + 0.0
T T
Solar Radiation: Scn = 123.0 (V ) + 0.0
SR SR
UV Radiation: S = 12195.122 (V ) + 0.0
* This is the default value of multiplier which corresponds to an average
sampling efficiency of 62% and was adjusted by operator, if necessary,
based on daily or weekly calibration.
190
-------
COMPUTATIONS USED IN SMOOTH PROGRAM
Given that x.. < x~ < x < x < x^ and
X2 " Xl = X3 " *2 = X4 " X3 = X5 " X4 *
the parabola that fits the points
in a least squares fashion has the following form:
y = a + a x + a x2
where
a = (-6y, + 24y + 34y + 24y - 6y J/70
o ' 2 o 4 5
" y2 " 2y3 " y4 + 2y5)/14
thus, to compute S , a smoothed estimate of D , a data value at time t,
w L
the following equation is used:
S^ = a
t o
where
and
- 60^1/70
{D. 9, D. .., Di5 D. , D. } is a set of data points
l-d t-1 t t+1 t+2 equally spaced in t1me
p
To compute a measure of fit, r , the squared correlation coefficient is
computed between the sets
Dt«}
and
{St-2' St-T St' Vl' St+2}>
191
-------
dS
The derivative with respect to time , of a smoothed estimate at
dt
time t is given by the following equation:
dS.
dT = V1
where
and
- "t-l + % + 2V10
I = Number of time units between time t and time t+1;
in the UNC data, this was 8 minutes
To derive smoothed estimates of the first two points in a data set,
the following equations are used:
where
S2 *
a = (-6D, + 24D0 + 34D + 24D, - 6DJ/70
0 ' £. 3 4 5
a] = (-2D] - D2 + D4 + 2D5)/10
To derive smoothed estimates of the last two points in a data set with
n readings, the following equations are used:
Sn-l = aO + a] + \
where
Sn = 30 + 2ai
a = (6D + 24D + 34D + 24D - 6D )/70
0 n-4 n-3 n-2 n-1 n"
Dn-4 " °n-3 + °n-l + 2IV/10
= (2Dn-4; °n-3 - 2Dn-2 - Vl + 2°n)/14
192
-------
For the first two points and the last two points, the derivative is
not computed.
To compute the integral, /,S , of a set of smooth estimates up to and
including time t, the following equation is used:
where
I = number of time units between time t and time t+1;
in the UNC data, this was 8 minutes.
193
-------
APPENDIX C
PHOTOCHEMICAL MODEL FOR PROPYLENE
INTRODUCTION
The model described in this appendix was based primarily upon a simpli-
fication of the more than 250-step mechanism used by Demerjian, Kerr,
and Calvert and the work of Hecht, et al. Dr. Dodge also supplied
a copy of EPA's kinetic simulation program developed by J. Overton which
uses the Gear differential equation solving subroutine for stiff differ-
ential equations. The Overton program was modified by Dr. Baker of UNC
to allow for a time-varying light intensity profile, for rate constant
changes for any reaction in the course of a simulation, for variable
dilution rates, and for improved output to allow better assessment of
the importance of an individual reaction step.
DEVELOPMENT
The propylene mechanism presented below was the ninth version written
and tested. It has 35 steps, fifteen of which are "lumped" or net reac-
tion steps which represent the results of a sequence of very fast radical
rearrangement or radical-oxygen reactions. Earlier versions had up to
60 steps and 24 species, the present limit of the kinetic simulation
program. Each reaction was tested for its importance to the time profile
shape of the major species during the course of the 10-hour simulation
and was eliminated if its contribution was found not to be significant
in this time period. For example, the model was frequently run without
the NpOg chemistry. These reactions were found to be not very signifi-
cant in the daylight simulations in the UNC chamber, but merely served
as a cyclic process with little net effect in relation to other, more
important, reactions such as OH+N02 •* HNO.,.
Mass balances in the model were not maintained for such species as H?0,
0?, and C0? and were not completely maintained for CO. The concentration
of H^O and Op were lumped into the necessary rate constants and C0?,
194
-------
formic acid, acetic acid, and other low concentration products were not
quantified (that is, they were allowed to "disappear" from the mechanism),
Since 0 + C-Hg resulted in the same reaction products as CL + C3Hg and
the primary fate of atomic oxygen (0) is 0 + Oo ^ 03, atomic oxygen
reactions were lumped into Reactions 1 and 15. Other atomic oxygen
reactions were found not to be competitive and were deleted.
ORGANIC RADICAL SYMBOLS
The mechanism contains seven organic radical symbols which represent the
radicals that result after rearrangement and reaction with oxygen. The
original radicals are the initial products of 0^ + C^Hfi, OH + C.,Hg,
H02 + C3H6, OH + CH20S and OH + RCHO, where C3Hg represents a propylene
molecule, ChLO represents a formaldehyde molecule, and RCHO represents
an acetaldehyde molecule. The seven organic radicals and the stoichi-
ometry used in the mechanism resulted from summarizing the more specific
34
reaction sequences of Demerjian et al. An example of such a summari-
zation is shown in Figure 87. Table 21 gives, for each organic radical
symbol in the mechanism, the proposed molecular configuration suggested
by Demerjian. In addition, Table 21 also gives the molecular structure
assumed for other organic species in the mechanism.
195
-------
*•"""*• «>-'•'"»
o o
.LO O
VCM !•—
V— •' <•*— -
*~** ...
CO
t
JC '
|
O
1 CM
rc
1 O
o
o
LO
CM
O --^
0 O
^d"
"¥" *~~*
CM
O CM
rc o
o
*f* "1"
CO
CM O
0 Z
•z. rc
f t
O CM
z: o
z:
H- "^
CO CO
o o
0 O
CO
• -a
r— C
re re
cu s2
C JO
•i— 3
•*-> c
cu a
E 0
CU -i-
O 4->
re
>> =3
^Ij C7"
LU
CU
to •
0 •«->
Q. S-
O O
i. CL
Q. CU
i. •
to i —
Q. to re
CU T-
4J JC •!->
to 4-> CU
>> c
C-r- C
re re
4J r- ••-
E CU -r-5
CU TJ J-
E O CU
cu E E
r— CU
cu cu o
c
OJ >>
Q-1^)
O Q.T3
1— 0 CU
$- 4J
a. Q.
o
>^ C T3
s- •!- re
4.)
10 "O CU
•i- CU to
E to O
CU 3 JC
JC -f->
0 LO
c cu
LO O i.
•r— »r* re
to •(->
>, re c
•—=30
0 cr-r-
4-* CU +J*
0 i-
C" "^3 ^J
Q- CU Q.
a.
cu E a.
TD 3 0
j>>_i •<->
cu c
-a .. o
re o to
E 4-> CU
S- 4-> -M
o o re
U- CO S-
•
VO
CO
cu
s-
3
O)
•r—
LU
196
-------
Table 22. ORGANIC SYMBOLS USED IN MECHANISM
Mechanism Molecular
Symbol Formula
RH03 CH3CH(02)CH2OH
RH02 CH=CHCH202
RH04 CH3CH(02)CH202H
RC03 CH3002
HC03 HC002
ROO CH300
R03 CH2(02)OH
RCHO CH3CHO
CH20 H2C=0
PAN CH3COOON02
197
-------
PHOTOCHEMICAL MODEL FOR PROPYLENE
No. Reaction Rate constant9'6
Nitrogen Dioxide Photolysis:
1: N02 •" 1.00
Nitric Oxide Oxidation:
2:
3:
4:
5:
6:
7:
8:
9:
10:
11:
Maj
12:
13:
14:
°3
\j
Lir\
RH09
1.
-> 1.
^ 1.
-»• 1.
+ 1.
-> 1.
-> 1.
-»• 1.
-> 1.
•* 2.
Dioxide
•* 1.
-» 1.
-> 1.
00
00
00
00
00
00
00
00
00
00
NO +
N02
N0? +
N00 +
c.
+
N02 +
+
N0? +
N02 +
+
N02 +
N02 +
N02 +
N0? +
1.00
1.00
2.00
0.28
1.00
0.04
1.00
1.00
1.00
1.00
1.00
1.00
1.00
°3
OH
CH90
Z
HC03
R03
CH20
H02
OH
CH20
ROO
H02
H02
ROO
+ 0.72
+ 0.72
+ 0.96
+ 0.04
+ 1.00
+ 1.00
+ 1.00
+ 1.00
H0~
2
CO
RCHO
CO
CO
RCHO
CO
CH20
k'
2.4 x
2.9 x
4.7 x
4.7 x
4.9 x
4.7 x
1.5 x
1.5 x
1.5 x
1.6 x
10]G
102G
102D
102D
103D
102D
103D'
103D'
103D'
T
IO"'D
Losses:
00
00
00
HN03
PAN
HN03 +
1.00
co2
8.5 x
4.9 x
5.0 x
103G
/^
lO^D
102D
Propylene Oxidation:
15:
16:
17:
°3
OH
H00
4* P H
+ C3H
+ C,H
6 -^ 1.
6 ^ °-
c -" 0.
50
61
95
H02 +
+
RH03 +
RHO. +
0.50
0.50
0.39
0.02
RCHO
ROO
RH02
\\ Q
+ 0.50
+ 1.00
+ 0.02
CH20
CO
RH00
1.8 x
2.1 x
1.1 x
10"2G
104G
10°D
198
-------
Aldehyde Chemistry:
18: CH20 ->
19: OH + CH20 +
20: RCHO +
21: OH + RCHO -»
1.71 H02 + 0.29 HC03 + 0.71 CO
0.71 H02
0.71 H02
1.00 RCO
+ 0.29 HC03 + 0.71 CO
+ 0.29 HC03 + 1.00 ROO
0.004k1 b
2.1 x 104G
0.00175k
2.2 x 10
Hydrogen Peroxide Chemistry:
22: H00 + HO,
23:
24: OH
H2°2
H2°2
1.00 H2
2.00 OH
1.00 HO
4.0 x 10JG
0.01k1 b
1.2 x 103G
Nitrous Acid Chemistry:
1.
-> 1.
-*. 1-
-v 1.
-». 1.
Dioxide
^ 1.
^ 2.
-> 1.
-V 1.
-v 2.
uu
00
00
00
00
00
1 IllUi^
N0? + 1
HN09
2
HN02
N02
OH + 1
Dark Phase
00
00
00
00
00
N03
N02
N9°«;
2 5
N09 + 1
2
HNO,
.00 NO
.00 NO
Chemistry
.00 NO,
3
3.2 x 10"5 d'c
1.7 x 10"3D
5.0 x 10"2D
4.9 x 103H
1.0 x 104
O.Ik, b
4.6 x 10"2G
1.3 x 104G
5.6 x 103G
1.7 x 101 bG
1.0 x 10"2b'8!
aunits are ppm -min unless otherwise stated
units are min~
cwater concentration included in rate constant
heterogeneous and homogeneous rate combined
Q
reference for rate constant; G = Garvin, et al.
D' = Dodge, H = Hampson, et al.
D = Demerjian, et al
199
-------
APPENDIX D
STATISTICAL ASSESSMENT OF OZONE
CALIBRATION PROCEDURE
The use of analytical instruments to measure low concentration of air
contaminants often leads to data of uncertain quality. One measure of
data quality is the confidence interval for a measured volume of concen-
tration. The confidence interval is the range of concentrations believed
to contain the true concentration with a certain level of confidence.
The level of confidence is related to the probability that the actual
value is within the interval, but it is not a true probability since the
actual concentration is a constant for any given measurement.
The calibration procedure for the chemiluminescent ozone instrument was
studied in order to calculate the confidence intervals for measurements
of ozone and to determine which factors in the calibration have the
greatest effects on data quality. The ozone instrument itself is very
precise, but its accuracy depends on the calibration procedure. The
calibration technique employs a neutral buffered potassium iodide as a
reference standard to measure a concentration of ozone from an ozone
generator. The instrument is adjusted to agree with the result of the
NBKI determination.
The relationship between absorbance, X, and equivalent ozone collected,
Y1, was determined by a weighted regression using five sets of serial di-
lutions. It was found that the greatest source of error was in the titra-
tion and dilution of standard I2 solutions. Thus, several sets of data
for different I? solutions must be considered to minimize bias. The slope
was calculated by the following equation:
200
-------
b =
i* Y'
n
z X.
1=1 7
This yields the following least squares fit:
Y'(pphm-liters) = 262.806 X (absorbance)
The confidence interval for the biased ozone measurements after calibra
tion may be determined by the following equations:
—
- V
X,
+ S.
n
z X.
1/2
Where Y is the estimated value of ozone concentration in pphm; t is the
Students' -t statistic with a cumulative probability density of (1 - a/2)
and n - 1 degrees of freedom; Y is a new value of ozone collected from
an actual NBKI bubbler run; k is the number of replicate bubbler runs;
X is the absorbance correcponding to the mean of the concentrations de-
termined by NBKI; £ x. is the sum of all absorbances used in the weighted
1 = 1
regression; Y. and Y . are the actual and predicted values of (L collec-
1 P1 O
ted at each point used in the weighted regression; V is the volume of
air from the ozone generator that is sampled in each bubbler run. The
quantity Var(Y ) must be determined empirically and ranges between 0.69
n 22 2
and 6.75 in the procedure studied. The S x /zx. term was 5.2x .
201
-------
The most critical factor is the number of replicates. Table 22 illus-
trates the effect of the number of replicate NBKI measurements on the
confidence interval on bias for measuring ozone at the Federal air
quality standard -- 8 pphra.
The best-case and worst-case figures represent uncertainty in Var(Y ).
22
As k is increased the term S x /rX. becomes more significant; this is
related to the ability of the analyst to consistently produce standard
IP solutions. This term also has increasing importance at higher con-
centrations.
Initial calibrations performed in the previous first year grant period
were with 7 to 8 replicates. Later, weekly calibrations were done
with 4 replicates. The long-term average slope of Op] versus absor-
bance was used in the 0., concentration prediction. Project personnel
feel that the accuracy of CL determinations at the National Ambient
Air Quality standard of 0.08 ppm was valid to within less than
i^ 0.01 ppm (0.95 confidence limit).
Since the ozone meter was used to measure ozone concentration and
change in concentration for the O^-NO gas phase titration used as a
claibration for the NO meter, the NO meter accuracy would probably
X .X
have been less than the 0., meter accuracy because of further transfer
and measurement errors. Therefore the best-case estimate of the NO
X
accuracy would be + 0.01 ppm.
202
-------
Table 23. CONFIDENCE INTERVALS ON ACCURACY OF
MEASUREMENT OF OZONE AT 8 PPHM
Number of Replicates Confidence Intervals
Best..Casea Worst Caseb
1 8+1.4 pphm 8+4.4 pphm
2
4
6
8
a Data taken by skilled operator in constant practice
Data taken by students performing technique for first time
8 + 1.1
8 + 0.83
8 +0.73
8 + Q.68
8 + 3.1
8 + 2.2
8 + 1.8
8 + 1.6
203
-------
APPENDIX E
DEVICE FOR THE CONTINUOUS MEASUREMENT
OF $ka FOR NITROGEN DIOXIDE
a
A quartz globe actinometer (QGA) utilizing N09 in N9 as the sensing
?"\
gas was developed by Sickles and Jeffries in the fall of 1974. The
QGA is illustrated schematically in Figure 87. It is operated as a
continuous stirred tank reactor (CSTR) and can be used in either
steady-state or nonsteady-state light conditions. The solution of a
simple equation utilizing the inlet and outlet N02 concentrations
gives k directly.
a
The kinetics of N09 photolysis in N9 were orginally derived by Shuck
39
et al. and are expressed in Equation 8:
r = -2k1
RJNO]
[N02]2
h W '
[N02]2
RI R2ENO] R3[02] -1
I , i . J i. (gj
where k, = k for N0?
R1 = k6[M]/k5 = 0.268
R2 = k?[M]/k5 = 0.25
R3 = k2[M]/k5 = 1.55 x 10"3
[N09] = N09 concentration in ppm
-1
r has units of ppm-min and
kc is the rate constant for 0 + N00 -> NO + 00
b c c.
kg is the rate constant for 0 + N02 + (M) -*• N03
kj is the rate constant for 0 + NO + (M) + N02
k is the rate constant for 0 + 0 + (M) -»- 0
Several assumptions are embodied in Equation 8. These are
204
-------
QUARTZ GLOBE*
II27CC
N2
50ppm N02
IN N2
0-RINGS*
TEFLON PLUG
LIGHT SHIELD
\- 50 ppm N02
-600 cc /min
3mm TEFLON
TUBING
N02 ANALYZER
600 cc / min
WASTE
Figure 87. Flow Diagram of One Liter OuartzGlobe CSTR for Nitrogen
Dioxide Photolysis Rate Measurements,
205
-------
[NO.,], and [No05J are small, and that pseudo-steady-state exists
for 0, 03, N03, and N205- The most serious assumption is that of
steady-state for NpOg which is not achieved for nearly 40 seconds at
typical conditions.
If the necessary condition for CSTR operation (ideal mixing) is assumed
to hold in the QGA, then a mass balance yields:
VdT = QCo " QC + Vr
where V = volume of globe
Q = flow rate
C = NCL concentration at inlet
C = NCL concentration at outlet1 and inside the globe
t = time
r = reaction net rate = d[N02J/dt = f(N02,light)
Rearrangement gives:
[NO,,]. - [N0?] A[NO?]
-r = 2 ° , 2~ = —-2- (10)
where T = Q/V and steady flow and light intensity are assumed
and t > 7r
206
-------
Substitution of Equation 8 in Equation 10 yields;
ss
R2([NO]Q + A[N02])
[N02J
(ID
which is the basic equation for the QGA.
If [Op] and [NO] are negligible, then Equation 11 may be simplified
to:
(12)
ss
A[N02]
2i[N02]
A[N02]
i + D + n e-
1 Kl K2 [N02]
• *
If the light intensity is not constant, Equation 8 may be substituted
in Equation 9 directly to yield:
\ ~ 2LN07]
nss L 2J
A[N02] d[N02]
dt
R2([NO]Q + A[N02])
(13)
which for [NO] and [Op] nearly zero can also be simplified to:
207
-------
nss
1
2[N02]
*A[N02]
t
d[N02] "
dt
•
R2A[N02]
[NO,]
(14)
The derviati've of the N02 outlet may be approximated numerically,
however, for outdoor conditions, only a small error occurs by assuming
that it is zero and using Equation 11.
The QGA has been extensively evaluated by experiment and by computer
23
simulation studies and these are reported elsewhere . These results
indicate a 95% confidence limit on precision at k, = 0.50 mins of
+ +
3.25 % and on accuracy of - 10.2%. The limit on accuracy is largely
due to uncertainties in the values of the rate constant ratios R,, R2,
and R.,. In actual use in Sri .indoor smog chamber, the standard devia-
-1+1
tion of nine replicates readings at k, = 0.496 min was - 0.01 min
over a three hour period.
208
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-650/3-75-011
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Outdoor Smog Chamber Studies: Effect of Hydrocarbon
Reduction on Nitrogen Dioxide
5. REPORT DATE
June 1975
6. PERFORMING ORGANIZATION CODE
1. AUTHOR(S)
Jeffries, Harvey E.
Fox, Donald L.
Kamens, Richard M.
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Department of Environmental Sciences &
School of Public Health
University of North Carolina
Chapel Hill, NC 27514
Engineering
10. PROGRAM ELEMENT NO.
1AA008
11. CONTRACT/GRANT NO.
800916
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory
Environmental Protection Agency
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Interim Report (6/15/72-1/31/7
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
3 3
A 312 m (11,018 ft ) Teflon film outdoor smog chamber was constructed in rural
North Carolina. The chamber was operated with natural conditions of solar radiation,
termperature and relative humidity which existed at the time of a run. Ninety-two
12-hour runs using propylene and oxides of nitrogen were conducted to assess the
performance of the system. A photochemical model, in which only the light intensity
magnitude and pattern and the rate of heterogeneous surface reactions were changed,
was used to compare the outdoor results with those of three indoor chambers. Good
agreement was found in all cases. One-hundred-thirty 12-hour runs were conducted
using a simulated urban hydrocarbon mix and oxides of nitrogen. Reduction of the
hydrocarbon concentration resulted in reductions of nitrogen dioxide (NO ) maximum
concentration and, for large reductions, the daytime NO dosage. Other factors
investigated included NO^ to NO ratio at constant NO , effect of slow dilution, and
2 y
results of extended 24-hour and 36-hour runs.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
COS AT I Field/Group
Smog
Smog Chambers
Hydrocarbons
NO (Nitrogen Dioxide)
Photochemical Reactions
Model
Photochemica1
13. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (ThisReport)'
Unclassified
21. NO. OF PAGES
223
20 SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
------- |