EPA-600/3-77-025
August 1977
Ecological Research Series
APPLICATION OF FOURIER TRANSFORM
SPECTROSCOPY TO AIR POLLUTION PROBLEMS
Interim Report - 1976
Environmental Sciences Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ECOLOGICAL RESEARCH series. This series
describes research on the effects of pollution on humans, plant and animal
species, and materials. Problems are assessed for their long- and short-term
influences. Investigations include formation, transport, and pathway studies to
determine the fate of pollutants and their effects. This work provides the technical
basis for setting standards to minimize undesirable changes in living organisms
in the aquatic, terrestrial, and atmospheric environments.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/3-77-025
August 1977
APPLICATION OF FOURIER TRANSFORM SPECTROSCOPY
TO AIR POLLUTION PROBLEMS
Interim Report - 1976
by
J. G. Calvert
W. H. Chan
E. Niple
R. J. Nordstrom
J. H. Shaw
W. R. Skinner
W. M. Uselman
The Ohio State University
Research Foundation
Columbus, Ohio 43212
Grant Number R803868-1
Project Officer
John Spence
Atmospheric Chemistry and Physics Division
Environmental Sciences Research Laboratory
Research Triangle Park, North Carolina 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences Research
Laboratory, U. S. Environmental Protection Agency and approved.for
publication. Approval does not signify that the contents necessarily
reflect the views and policies of the U. S. Environmental Protection
Agency, nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
ii
-------
ABSTRACT
A Fourier Transform Spectrometer has been used to explore problems
in air pollution. Spectra of air samples at ground level of approximately
10m, 100m, and 1km, and solar spectra obtained for solar zenith angles
between 40 ..and 87 have been obtained in the spectral region from 700
to 3500 cm with a spectral resolution of better than 0.2 cm . Examples
of these spectra in the region from 1100 to 1200 cm are presented to-
gether with synthetic spectra calculated from the compilation of in-
formation concerning the spectral lines of atmospheric gases prepared by
the Air Force Geophysical Laboratories. The procedure used to generate
these synthetic spectra is described briefly. These atmospheric spectra
have been searched for the presence of absorption features of a number
of gases including HNO_, HNO,- HCOOH. From the absorption features of
fluorocarbon-12 near 1160 cm in solar spectra, a mean tropospheric
abundance of 0.34 ppb has been estimated.
A photochemical cell capable of approximating the solar noon ir-
radiance at ground level between 300 and 400 nm and in which path lengths
in excess of 200m can be obtained is described. This cell has been used
to study the photoysis of HNCL and a value of 0.070 min has been ob-
tained for the rate constant.
Spectra of more than 20 gases of importance to air pollution prob-
lems, obtained under controlled concentration and pressure conditions
and covering the region from 700 to 1500 cm~ , are presented.
The progress in the construction of an absorption cell designed to
simulate the environment of the stratosphere is described. This cell
can be cooled to -60 C and can be irradiated with short wave-length UV
radiation down to 170nm.
This report was submitted in fulfillment of Grant Number R803868-1
by The Ohio State University Research Foundation under the sponsorship
of the U.S. Environmental Protection Agency. This report covers the
period July 15, 1975, to July 14, 1976.
111
-------
CONTENTS
Abstract iii
Figures vi
Tables ix
Acknowledgment A x
1. Introduction 1
2. Identification of Trace Atmospheric Constituents 2
Identification of fluorocarbon-12 in solar spectra 2
Exploration of techniques for spectral data reduction
and analysis-calculation of synthetic spectra 13
Atlas of ground level air spectra and solar spectra 23
3. Design of Absorption Cell to Simulate the Environment of the...
Stratosphere 35
4. Study of Some Key Reactions of Probable Importance in Photo-...
chemical Smog Formation 41
Construction and performance of photochemical cell 41
Photolysis experiments 56
5. Library of Spectra 66
6. Concerning Problems Identified as a Result of the Present
Investigation 94
Solar spectra and ground level air spectra analysis 94
nitrous acid band analysis 94
References 96
Appendices
A. Summary of obj ectives from 1975 proposal 99
B. Papers published or accepted for publication 101
C. List of papers presented at meetings 102
-------
FIGURES
Number
1 Apparatus used to Collect Solar Spectra ^
2 Low-Resolution Solar Spectrum from 500 to ^200. cm'1 6
3 High-Resolution Spectra of Fluorocarbon-12 (CC12F2) from
800 to 1200 cm"1 . 7
k Comparison of Solar Spectra, Laboratory Spectra of
Fluorocarbon-12, and Computer Synthesized Solar Spectra
(with no FC-12 Absorptance) near 923 and Il6l cm'1 8
5 Transmittance of 5.0 x 10~3 Torr FC-12 and 0, 200 and
700 Torr N2. Total path length = 170 m 10
6 Peak Absorbance of FC-12 at 1160.9 cm'1 as a Function of
Optical Density (pi) 12
7 Effect of the Width of the Instrument Function a (cm1) on
Resolution 18
8 Atmospheric Spectra between 1100 and 1128 cm"1 (see Table V) 26
8a Synthetic Atmospheric Spectra between 1100 and 1128 cm'1
(see Table VI ) 25
9 Atmospheric Spectra between 112 U and 1152 cm'1 28
9a Synthetic Spectra between 112l* and 1152 cm"1 27
10 Atmospheric Spectra between 1150 and 1178 cm"1 30
lOa Synthetic Spectra between 1150 and 1178 cm"1 29
11 Atmospheric Spectra between 117^- and 1202 cm"1 32
lla Synthetic Spectra between 117^ and 1202 cm"1 31
12 Schematic Diagrams of the Cell and Bath . 37
13 Schematic Diagram of Typical Bellows Connection from Cell
to Bath 38
ll<- Radiation Field Inside Cell ^0
15 Optical Transfer System from Interferometer to Photochemical
Cell , te
16 Gas Inlet System to 6 m Cell . UU
vl.
-------
FIGURES (Continued)
Number . Page
1? Gas Handling System for 6 m Cell ^5
l8a Spectral Distribution of Fluorescent Lamp Output after
las sing through the Glass Walls of the Cell k7
l8b Relative Spectral Distribution of Solar Energy at Ground
Level U7
19 Radiant Flux Detector Readings (in mW/cm2- sr) Measured in
a Plane Containing the Cell Axis kQ
20 The Theoretical Magnitude of the Rate Constant k, for the
Formation of 0(3p) Atoms through N02 Photolysis at Different
Solar Zenith Angles as Estimated for the Atmosphere of Los
Angeles 50
21 A Plot of 0rc(Ao/At) vs Time for N03 Photolysis at Low
Pressure [N0?]0 = 30.*J- ppm 52
\
22 Plot of [N02] versus Time of Irradiation in the Photochemical
Cell 5^
23 Plot of [N02] versus Time of Irradiation in the Photochemical
Cell 55
2k The Time Dependence of [HONO] and [C02] in the Photolysis
of an HONO-NOX-CO Mixture 60
25 The Time Dependence of [HONO] and [C02] in the Photolysis
of an HONO-NO -CO Mixture 6l
X
26 The Time Dependence of [HONO] and [C02] in the Photolysis
of an HONO-NOX-CO Mixture 62
27 The Time Dependence of [HONOl and [CO?] in the Photolysis
of an HONO-NO -H20 Mixture " 63
' ji
28 Chart of Chlorine and Fluorine Addition to Methane 66
29 High-Resolution Spectra of Methane 67
30 High-Resolution Spectra of Methyl-fluoride 68
31 High-Resolution Spectra of Dichloro-methane 69
32 High-Resolution Spectra of Chloroform 70
vii
-------
FIGURES (Continued)
Number
33 High-Resolution Spectra of Fluorocarbon-21 71
3k High-Resolution Spectra of Fluorocarbon-22 72
35 High-Resolution Spectra of Fluoroform -> 73
36 High-Resolution Spectra of Carbon Tetrachloride fk
37 High-Resolution Spectra of Fluorocarbon-11 75
38 High-Resolution Spectra of Fluorocarbon-12 76
39 High-Resolution Spectra of Carbon Tetrafluoride 77
kO Infrared Spectrum of Hydrogen Peroxide 79
^1 Infrared Spectrum of Nitrous Acid 80
k2. Infrared Spectrum of Methyl Nitrate 8l
U3 Infrared Spectrum of O^one 82
M* Infrared Spectrum of Carbon monoxide 83
^5 Infrared Spectrum of Methanol 8*4-
1+6 Infrared Spectrum of t-Butyl-Alcohol 85
U7 Infrared Spectrum of di-t-Butyl-Peroxide 86
1+8 Infrared Spectrum of t-Butyl-Hydroperoxide 87
1+9 Infrared Spectrum of 2, 2' - Azoisobutane 88
50 Infrared Spectrum of Azomethane 89
51 Infrared Spectrum of Formic Acid 90
52 Infrared Spectrum of Formaldehyde 91
53 Infrared Spectrum of Acetone 92
viii
-------
TABLES
No. Page
I Equivalent Air Masses for Zenith Angles 0° to 90°
(from Ref. 16) 16
II Surface Number Densities at T = 29*1 K and P = 1013 mbar
for Atmospheric Gases with Constant Mixing Ratios 19
III Midlatitude Summer Atmospheric Model (from Ref. 8) 21
IV Thickness, Mean Pressure, and Temperature of Layers Used
in Slant Path Computation 22
V Information Concerning Spectra Shown in Figures 8-11 2k
VI Parameters for Synthetic Spectra in Figures 8a-lla 2^
VII Rate Data on the Photolysis of N02 at Low Pressure 53
VIII Theoretical Estimates of First-Order Decay Constants for
the Photodecomposition of Some Light-Sensitive Compounds
of Interest in this Work (Photochemical Chamber) 58
IX Absorption Coefficients Determined in this Work (Pressure
N£ = 700 Torr, Temperature = 23 + 2°, Resolution = 1 cm"1 59
X Initial Conditions of the HONO Photolysis Experiments 59
ix.
-------
ACKNOWLEDGMENTS
We would like to thank Dr. S. Z. Levine and Mr. M. Whitbeck for carrying
out computer simulations and for preparing compounds for some of the photolysis
experiments.
-------
INTRODUCTION
This report summarizes progress made under EPA Grant No. R803868-1. This
grant began in July, 1975 > and was a successor to EPA Grant No. R803075-01
which extended from April, 197**, to July, 1975• The work completed under the
initial grant is described in the Final Report dated August, 1975-
During the first year an Infrared Fourier Transform Spectrometer (IFTS)
was made operational and used in conjunction with a 21 m multiple traversal
cell to study infrared spectra of atmospheric gases and several pollutant gases
including: NO, N02, HN02. Some preliminary solar spectra between 3 and 20 ^m
were also obtained.
The kinetics of the formation and decay of nitrous acid were followed
by monitoring bands of NO, W?, and cis- and trans-HN02. Rate constants for
the reactions were derived and a mechanism accounting for the formation of HN02
in the atmosphere was proposed.
In addition to these studies a six-meter multiple traversal cell in which
path lengths up to 300 m can be obtained was constructed.
This cell is surrounded with fluorescent tubes with output in the region
from 300 to ^50 nm. The cell irradiance in this region closely simulates the
solar irradiance at ground level in spectral distribution and intensity.
The work performed under the present grant has built on the experience
gained during the first year of work, and the progress made in several areas
is described in this report.
-------
IDENTIFICATION OF TRACE ATMOSPHERIC CONSTITUENTS
Among the objectives of this work is the search for evidence of trace gas
constituents in the atmosphere and the estimation of their abundances and
variabilities. Infrared techniques are among the most unambiguous and sensi-
tive methods for identifying traces of gases, especially if high-resolution
spectra of long atmospheric paths are available. We have obtained relatively
high resolution (0.125 cm"1)spectra of various path lengths of ground level
air and also solar spectra. Analyses of these spectra are being made and the
present status of these investigations is described below.
IDENTIFICATION OF FLUOROCARBON-12 IN SOLAR SPECTRA
Concern over the atmospheric concentrations and fates of halogen-substituted
methanes has grown since the first report of the potential influence of these
compounds in the stratosphere. A variety of techniques has been employed to
measure trace pollutants such as fluorocarbon-11 (FC-11, CC13F) and
fluorocarbon-12 (FC-12, CCl2Fp). Gas chromatography has been used extensively
by several research groups throughout the world to record in situ concentra-
tions of these compounds at a variety of altitudes.2'5
Although gas chromatography is a popular method for investigating atmospheric
pollutants, several forms of infrared spectroscopy are also being used to
identify trace gas.es. By using infrared Fourier transform spectroscopy and a
cryogenic procedure for concentrating trace pollutants, Hanst ejt al.6 detected
background levels of several pollutants including FC-11 and FC-12. Murcray
et al.7 performed in situ infrared measurements in the stratosphere, and de-
tected FC-11 and FC-12 as well as HN03 with a grating instrument on a balloon
flight at 30 km.
We have initiated a project to record high-resolution spectra of the infrared
transmission of solar radiation through the atmosphere to detect trace
pollutants. The solar spectra were recorded using Fourier transform spectros-
copy from a ground-based station in Columbus, Ohio. Nominal resolution for
this instrument is 0.125 cm'1.
Data were recorded over a two-month period in September and October, 1975.
Preliminary analysis of these solar spectra indicates the presence of FC-12
in the atmosphere. This analysis was made by comparing these solar spectra
with laboratory spectra of FC-12 and also by comparing the observed solar
spectra with computer synthesized solar spectra which exclude absorption fea-
tures due to trace pollutants. These synthetic spectra were generated from
the Air Force Cambridge Research Laboratories (AFCRL) Atmospheric Absorption
Line Parameters Compilation.8 This compilation is available on magnetic tape.
Results of these comparisons indicate that there are absorption features in
our collected solar spectra which are not present in the computer-generated
spectra. Furthermore, these absorption peaks agree in both position and re-
lative intensity with several absorption peaks of FC-12.
-------
Equipment
A sketch of the apparatus used to collect the solar spectra is shown in Fig0 1.
Solar radiation incident on the first coelostat mirror (CMx) is reflected to
the second coelostat mirror (CM2)0 This second mirror reflects the radiation
down through an opening in the roof and into the laboratory. Both coelostat
mirrors are 30 cm diameter flats.
The first coelostat mirror is equipped with a clock drive which rotates the
mirror to track the motion of the sun. The coelostat thus provides a fixed
"beam of solar radiation into the laboratory. Mirror M3 reflects the vertical
beam to an optical table where mirrors NLj and M5 are positioned to direct the
radiation into the Michelson interferometer. Both of these mirrors are rec-
tangular flats.
An aperture A reduces the beam diameter to approximately 2.5 cm before the
radiation enters the interferometer. This aperture aids in alignment of the
beam into the interferometer in the following way. The beam splitter is re-
moved from the interferometer and mirrors M4 and M5 are adjusted until the
2.5 cm diameter beam falls in the center of the 5 cm diameter interferometer
mirror, and the beam which is reflected from the interferometer mirror passes
back through the optical system and out the aperture. In this fashion, the
beam is made reasonably perpendicular to the mirror surface. The beam splitter
is then replaced.
Mirror MS directs the radiation which exits the interferometer into the
detector. For this study, a copper-doped germanium photoconductor operated
at liquid helium temperature was used. The spectral range of this detector
extends from 300 cm"1 to 3500 cm"1. However, because of sampling limitations
involved in high-resolution Fourier transform spectroscopy, the spectral range
was limited by an optical filter to the region 300 cm"1 to 1500 cm"1. The
detector assembly is equipped with optics which focus parallel radiation onto
the 1 mm square detecting crystal.
The output from the detector is digitized by a 15-bit analog-to-digital con-
verter, and the interferogram is computed into a spectrum by a Nova 1200 mini-
computer. This computer has 8K of core and has a 128K, fixed-head disc and a
9-track, high-density magnetic tape unit as peripheral equipment. The computed
spectra are stored on the magnetic tape and can be recalled at any future time
for plotting on a digital plotter. The entire system was supplied by Digilab,
Inc.
High-resolution spectra of FC-12 were recorded with the same Fourier transform
instrument. A Nernst glower was used as the infrared source. Radiation from
the glower was collimated into the interferometer. The radiation leaving the
interferometer passed through a multiple traversal cell and was collected by
the copper-doped germanium detector.
The absorption cell was constructed entirely of glass and stainless steel.
The mirror system in the cell was designed by Hanst6 and the base path length
between the mirrors is approximately 5°3 meters. The spectra of FC-12 were
recorded by using a variety of path lengths within the cell.
-------
CLOCK
DRIVE
LOW SUN
SOLAR .RADIATION
NTERFEROMETER
Cir.Ge DETECTOR
Figure 1. Apparatus used to collect solar spectra
-------
Detection of Fluorocarbon-12
Figure 2 shows a low-resolution survey spectrum of the atmospheric transmission
of infrared solar radiation from 500 cm"1 to U200 cm"1. This plot shows that
the resident atmospheric gases, most notably H20 and C02 are totally absorbing
in several regions, dividing the spectrum into limited infrared "windows."
If pollutants are to be detected in the atmosphere by observing infrared solar
spectra, it is necessary that their infrared absorptions fall within an atmo-
spheric transmission window. Even within a window, however, interference from
partially absorbing lines can obscure the detection of trace gases.
A high-resolution solar spectrum is rich with a complexity of absorbing lines.
In order to sort out which of these lines, if any, are due to an atmospheric
contaminant, it is necessary to make extensive comparisons of the solar spec-
trum with various laboratory spectra. Many gases have been selected for com-
parison, however at this time only FC-12 has been identified in the solar
spectra.
Figure 3 shows two high-resolution transmission spectra of FC-12 in the region
of the 1000 cm"1 solar window. Nominal resolution for these spectra is 0.125
cm"1. Several features stand out as being ideal signatures for possible iden-
tification of FC-12 in solar spectra. These features are the very strong,
narrow absorptions near 923 cm"1 and Il6l cm"1. which are possibly due to Q-
branch transitions. Narrow absorption spikes such as these might be separable
at high resolution from absorption features of the resident atmospheric gases.
Figure U shows several comparisons which were made to verify that FC-12 is
indeed present in our solar spectra. The two spectral regions where the nar-
row FC-12 absorptions occur are plotted. The top spectrum in each region is
FC-12 at low concentration. Below this spectrum is one of the many solar
spectra which were collected. The solar angle to the zenith was approximately
^5° for this spectrum. In the region near 923 cm"1 there is a strong water
absorption which partially masks the position of the FC-12 absorption features.
Nevertheless, three distinctive absorption characteristics are evident in the
solar spectrum which might be attributed to FC-12. Murcray et al.7 have pre-
viously identified these absorption features as belonging to FC-12.
The region near Il6l cm"1 appears to be a much better spectral region in which
to observe atmospheric FC-12. From the laboratory spectrum, it is clear that
the FC-12 absorption in this region is approximately twice as strong as the
strongest feature near 923 cm"1 which leads to a more sensitive detection.
Also, the region around Il6l cm"1 is free from water vapor interference. The
moderately strong and regular absorption lines in this region of the solar
spectrum are due to N?0. Weaker absorption lines are due to 03. Figure k
shows an absorption line in the solar spectrum which coincides with the strong
absorption of FC-12.
Although these four absorption lines in the solar spectrum coincide with the
positions of FC-12 absorptions, it would be most informative to compare the
solar spectrum with a similar spectrum which definitely had no FC-12 present.
This comparison was accomplished by first generating synthetic solar spectra.
Results of this computer synthesis are shown in Fig. h. The spectrum below
-------
cr\
I
1000
I
2000
3000
T
4000
WAVENUMBER (cmH)
Figure 2. Low-resolution solar spectrum from 500 to U200 cm
-------
o
CO
CO
Cr>
s*'
cr
- 0
_ . i
800
•0
900
1000
1100
1200
WAVENUMBER (cm"1)
Figure 3. High-resolution spectra of fluorocarbon-12 (CC12F2) from 800 to
1200 cm'1
-------
915
920
FC-12
925
SOLAR
SPECTRUM
COMPUTED
SPECTRUM
\
/j-VVi^A,1,1-
'ii V ! if ',"
!! i 1 l!
• 1 : ;
. ;i ! ' t i
1 i '
;. ; ij
'MM1'
r
!:', A' • s N
r::!-''1."1!!:.
!. ... M! .,
II "<•] {1 :i; .
1 i ! '•
\ i i t
" \
1;
ii
1
t
\
i
: i i ^ |
i
1155
1160
1165
Figure U.
Comparison of solar spectra, laboratory spectra of fluorocarbon-12,
and computer synthesized solar spectra (with no FC-12 absorptance)
near 923 and Il6l cm'1
-------
the collected solar spectrum is a computer generated solar spectrum; the
method for generating this spectrum is discussed below.
In the computerized spectrum it was found that the water vapor absorption line
marked with the asterisk (*) did not coincide with the same line in our col-
lected spectrum. The collected spectra were examined to determine if the
position of this line was consistent within the collected data. It was found
that the collected data, recorded periodically over two months, were self-
consistent and therefore it was decided to move this line in the synthetic
spectrum to lower frequencies by approximately 0.25 cm"1 to obtain the agree-
ment shown in the figure. In the 10 wavenumber region around Il6l cm"1 there
is good correlation between the collected and computed spectra for over 70
lines.
In both spectral regions the absorption features which have been tentatively
identified as belonging to FC-12 are missing in the synthetic spectrum. The
fact that these features are absent in the computed spectrum and the fact that
all other absorption features in the synthetic spectrum show good agreement
with the collected spectrum places much credibility to the assignment of these
four features as being due to FC-12 absorption.
Analysis
In order to estimate the atmospheric concentration of FC-12, laboratory spectra
of known concentrations of FC-12 were recorded. However, several features of
the absorption of solar radiation by the atmosphere cannot be reproduced in a
laboratory absorption cell. One such feature is the vertical profile of tem-
perature and pressure of the atmosphere as the radiation approaches the surface
of the Earth. The resulting solar transmission spectrum is a composite of
transmissions of strata of air, each at a different temperature and pressure.
Another feature of solar spectra which cannot be duplicated in a laboratory
environment is the fluctuation of absorber concentration along the optical
path.
No temperature variation studies could be performed with our absorption cell.
However, since the bands of FC-12 which appear in Fig. U are fundamental bands,9
large changes in the appearance of the transmission spectrum with temperature
are not expected.10
A careful study of the dependence of the absorption bands of FC-12 as a func-
tion of total cell pressure was performed. A known amount of FC-12 was intro-
duced into the cell and a spectrum was recorded. High-purity nitrogen gas was
then injected into the cell to bring the total pressure up to some desired
amount and another spectrum was recorded. The total pressure was increased in
this manner from a few Torr up to atmospheric pressure.
Results of this pressure study are shown in Fig. 5- As this figure indicates,
there is no dramatic change in the appearance of the transmission spectrum as
the total pressure within the cell is varied over this wide range. The fact
that these spectra show very little broadening with pressure is fortunate when
it comes to analyzing the solar spectra data to determine the atmospheric FC-12
-------
0.0 -
0.0 .
0.0
700
200
0
.
915 920 925 1155 1160 1165
WAVENUMBER (cm-)
Figure 5. Transmittance of 5.0 x 10~3 Torr FC-12 and 0, 200, and ?00 Torr N2.
Total path length = 1?0 m
10
-------
concentration. Essentially, the entire atmosphere can be considered to "be a
homogeneous host for the trace pollutant since all strata of the atmosphere
have the same effect on the FC-12 spectra. Furthermore, if FC-12 is assumed
to be uniformly mixed in the troposphere,4 and assuming that the small amount
of FC-12 in the stratosphere contributes negligibly to the absorption in the
solar spectrum, the problem of calibrating the amount of this fluorocarbon
contaminant in the atmosphere is greatly simplified.
Several methods for determining the atmospheric concentration of FC-12 from
solar spectra are being investigated. Perhaps the most obvious technique is
to compare the center frequency transmissions of each of the four suspected
FC-12 peaks in the solar spectrum with laboratory spectra of known amounts of
the compound. This technique requires that all spectra be recorded at the
same resolution and does not take account of possible overlap of spectral
features.
The spectral transmittance of a gas can be-expressed as
T(v) = exp [-K(v) o-l], . (1)
where K(v) is the absorption coefficient at frequency v, p is the number den-
sity of absorber molecules in molecules/cm3, and 1 is the path length (in cm)
of the radiation through the absorber. The absorbance A(v) is defined as
^[T(v)l. Therefore,
A(v) = K(v) • P • 1. (2)
A plot of absorbance at a given frequency vo vs. the product p • 1 for an
absorber should give a straight line with an intercept at the origin and a
slope K(VQ).
Since the absorption features of FC-12 near 923 cm"1 are masked by water vapor,
a good calibration of this pollutant cannot be made in this region. The
strong, narrow spike at 1160.9 cm'1 was used instead. Absorbance data were
extracted from the FC-12 laboratory spectra, and plotted against the p • 1
product; this plot is shown in Fig. 6. The optical density p • 1 is in
molecules/cm2. It is clear from this plot that the data can be represented
by a straight line. The line chosen in the plot is an estimated fit based on
the confidence placed on the individual data points. This fit yields an
intercept which, within experimental error, passes through the origin. The
slope of the line is 8.9 x 1C)-18 (molecule/cm2)-1, or 2UO (atm • cm)-1.
In measuring the absorbance from the solar spectrum a background level must
be assumed, and the overlap of the FC-12 absorption feature with N^O and 03
makes this difficult. Nevertheless, from the solar spectrum in Fig. U, an
atmospheric optical density of 1.3 x 1016 molecules/cm2 can be estimated. As
indicated earlier, the solar zenith angle was approximately U5° for this
spectrum. Assuming the approximate height of the troposphere to be 10 km, a
partial pressure of FC-12 at 3.1). x 10"10 atm can be calculated.
11
-------
UJ
O
Z
<
CD
ct:
o
c/)
CD
1.2
1.0
0.8
0.6
0.4
0.2
0.0
o
o
10
12
14
02468
X I016
OPTICAL DENSITY
(MOL/CM2)
Figure 6. Peak observance of FC-12 at 1160.9 cm"1 as a function of optical
density (pi)
12
-------
Examination of other solar spectra recorded over this two-month interval shows
similar results. Hanst, et al.6 measured a partial pressure of 2.3 x 10~10
atm for FC-12. Hester et alT* used gas chromatography to measure ambient
levels of FC-12. Their results show a wide fluctuation in concentration over
a period of 1^ days, but the average amount of FC-12 appears from their data
to be close to U.O x 10~10 atm. Thus the quantity of atmospheric FC-12 which
is measured from our solar spectra is in agreement with other reported values.
"EXPLORATION OF TECHNIQUES FOR SPECTRAL DATA REDUCTION AND ANALYSIS CALCULATION
OF SYNTHETIC SPECTRA
The spectrum of molecular absorbers along a given path can be calculated if the
amounts of the absorbing gases, the temperature and pressure variations along
the path, and the requisite molecular parameters of the absorbing gases are
known. McClatchey et al.8 have compiled the available information on the posi-
tions, intensities, and widths (for pressure broadening by air), together with
the lower state transition energies for many absorption lines of seven per-
manent atmospheric gases; i.e., 03, CO, N50, C02, CH4, 03, and H20. This line
listing has been used to calculate absorption spectra of horizontal homogeneous
paths or slant paths through the atmosphere by a number of workers (see, for
example, Refs. 11, 12, 13). We have used this compilation to obtain synthetic
spectra for comparison with experimental spectra of ground level air and with
solar spectra. Since all these spectra were recorded at ground level we have
assumed that the spectral lines are pressure broadened and can be represented
by the Lorentz line shape
K(v) = Sa/7t[(v-v0)2 + a2], (3)
where K(v) = absorption coefficient at frequency v(cm~1) due to a line of
intensity S(cm"1/molecules cm~2) and half width ( at half height) a(cm~1).
This line shape is a good approximation of the shape of most lines of atmo-
spheric gases at pressures greater than about 10 Torr. At lower pressures
the Voigt profile should be used.14 However, with the exception of 03, the
contribution to the absorption in ground level solar spectra by atmospheric
molecules at pressures below 10 Torr is small compared with that due to the
molecules at higher pressures. The attenuation due to molecular scattering,
and to aerosol absorption and scattering, has also been neglected in the com-
putations since these factors are not strongly frequency dependent8?15 and the
experimental data show relative rather than absolute transmittances.
Absorption By A Gas
The monochromatic absorbance of a homogeneous layer of thickness Al due to a
single absorption line with a Lorentz shape is
13
-------
where Po = power incident on the layer, and
P = power transmitted by the layer.
Since each gas gives ri.se to many absorption lines the total absorbance
at frequency v due to the kth gas in the layer is found by summing Eq. (h]
over the Nk lines which contribute to the absorption at that frequency;
sik
(v
i= 1
where i represents the ith line of the kth gas.
The total absorbance due to n different gases is
k=n i=%
^ X (vT "^ Ta2 ' (6)
k=l
If the temperature and pressure or the mixing ratios of the gases vary along
the path, the total absorbance can be calculated by dividing the path into M
layers, each of which is considered homogeneous and summing over all the layers
M n %
„ , , V V pkm V" Sikm «ikm Alm , v
AM(V) = ) > -^ ) 7 ^ T~ • 7
U U « ^ (v-voik)2 + o
m=l k=l
Although the positions of the lines are essentially independent of the pressure
and temperature of the sample, the line intensity S is a function of temperature
and the line width a depends on both temperature and pressure8 such that
S = So (—Y1 exp[E" (To-T)/0.69^6 T0T], (8)
\1 I ^
and x '
a = ob (TO/T)I/! P/PO, (9)
When S0 is the line intensity at temperature T0, OQ is the half width at
temperature To and pressure Po, E" is the energy of the lower state of the
transition, and n is a parameter which depends on the molecule.8 For H20, 03,
CH4, n= 1.5, and for CO, N20, C02> and 02, n= 1.
lU
-------
Since most measurements of linear parameters have been made for room tempera-
ture and pressure, the AFCRL line compilation has chosen To = 296 K and Po =
1013 mbar.
In our applications the atmospheric paths to be simulated were either homo-
geneous (M= l) or slant paths corresponding to solar spectra. Provided the
zenith angle z of the sun is less than about 80° the plane parallel approxima-
tion can be used to describe the atmosphere and the path length through a
given layer can be written as
Alm = ^ sec z>
when Ahm is the vertical thickness of the layer. For larger zenith angles the
effects of refraction and departures from the plane parallel approximation be-
increasingly important and corrections must be obtained from suitable
tables.16 Table I shows equivalent air masses for zenith angles between 0°
and 90C.
As indicated by Eq. (3)5 the absorption coefficient for a Lorentzian line
shape is never completely zero. Therefore every line in a spectrum contributes
to the absorbance at all frequencies, and these contributions should all be
summed to find the total absorbance at a given frequency. However, the amount
of computation time to perform this task would be prohibitive. Therefore,
several approximations have been incorporated into the programming to reduce
the computation time. The first approximation is the elimination of lines
which have a negligible contribution to the total absorbance at their center
frequencies. This cancellation is performed by calculating the absorbance of
a line at center frequency. From Eq. (^) this absorbance is A(vo) =
S-o-I/art. If this quantity is less than 0.001 the line is ignored.
The second approximation employed to produce synthetic spectra is the elimina-
tion of the wings of each absorption line. It was assumed that any absorption
line had negligible contribution to the total absorbance at frequencies greater
than 20 cm"1 away from the center frequency of the line. This is a conserva-
tive estimate of the range of influence for many of the absorption lines, and
use of this cut-off value has resulted in computed spectra which show good
agreement with observed spectra. Thus the summations over the index i over
N^ lines in Eq. (5>6,7) are subject to these limitations.
A spectrum with an effective resolution &v cm"1 is obtained by calculating
the monochromatic transmittances at frequency intervals ^v cm"1. As ?v goes
to zero, the resolution becomes infinite. In order to synthesize a spectrum
the size of the interval 5v between calculated data points should be equal to
a/m where a is the halfwidth of a computed line and m is an integer greater
than 5. For ground level homogeneous path spectra with line halfwidths typically
about 0.05 cm"1, a good value of 5v is about 0.01 cm"1. However, in order to
obtain a good representation of ozone lines in solar spectra whese halfwidths
are less than 0.01 cm"1 a better value for 5v is 0.002 cm"1 or less.
15
-------
TABLE I. EQUIVALENT AIR MASSES FOR ZENITH ANGLES 0°'to 90° (FROM REF. l6)
0° 1° 2° 3° 4° 5° ' 6° 7° 8° 9°
0° 1.00 1.00 1.00 1.00 1.00 1.00 1.01 1.01 1.01 1.01
10° 1.02 1.02 1.02 1.03 1.03 1.04 1.04 1.05 1.05 1.06
20° 1.06 1.07 1.08 1.09 1.09 1.10 1.11 1.12 1.13 1.14
30° 1.15 1.17 1.18 1.19 1.20 1.22 1.24 1.25 1.27 1.28
40° 1.30 1.32 1.34 1.37 1.39 l.4l 1.44 1.46 1.49 1.52
50° 1.55 1.59 1.62 1.66 1.70 1.74 1.78 1.83 1.88 1.94
60° 2.00 2.06 2.12 2.20 2.27 2.36 2.45 2.55 2.65 2.77
70° 2.90 3.05 3.21 3.39 3.59 3-82 4.08 4.37 4.72 5.12
80° 5.60 6.18 6.88 7-77 8.90 10.40 12.44 15.36 19-79 26.96
16
-------
In order to compare synthetic spectra with actual spectra whose spectral re-
solutions are less than &v the resolution of the synthetic spectra must be
degraded by convolution with an appropriate instrument function. In the work
reported here a triangular function was chosen with a halfwidth a(cm'1).
The transmittance T(v) of the degraded spectrum is then obtained from the
expression
T.(v) / ^Vj.v') T(v') dv', (1Q)
/ cr(v15v') dv'
where cr(v1}v') is the instrument response function. This integration is done
numerically by the trapezoid rule
j(a-|Vi-v|)T(v)} - (a-|vv_a-v|)T(vv_a) - (a-|
(11)
where the summation extends over all computed frequencies from v - a to v+ a.
It was found that the accuracy of the convolution was strongly dependent on
the precision with which the calculation is performed. For this reason, the
convolution was performed in double precision.
The effect of degrading the spectral resolution on the appearance of a spectrum
is illustrated in Fig. 7? which shows a spectrum of a 1 km path of air at 296 K
and 760 Torr. The spectral resolution increases from the top to the bottom of
figure. In the upper curve a = 0.5 cm"1 and then successively changes to
0.35 0.1, 0.05, and 0.01 cm'1 in the lowest curve. In addition to degrading
its resolution it is also noticeable that the spectrum approaches a more con-
tinuous curve as the value of a/Av increases.
Specification of Parameters for Synthetic Spectra
Horizontal Baths—
For homogeneous, horizontal paths it is necessary to specify the temperature,
pressure, and path length of the absorbing medium as well as the mixing ratios
of the absorbing gases. Values for the number densities of CH4, 02, C05, CO,
and N20 which were used to compute ground-level spectra at 29^ K and 1013 mbar
ars shown in Table II. These values are adjusted as described below when
spectra at different temperatures and pressures are computed. There is no
absorption by oxygen in the spectral region investigated here and the amount
of 03 at ground level is not sufficient to produce significant absorption in
this spectral region for the path lengths used.
Local and temporal variations in CO and C02 are known to occur and slight
changes in the number densities of these molecules were made to obtain the
best agreement between observed and computed spectra. The large variability
17
-------
.
fsi
V 'v
T I r r i i
iS^.CiC; liTi-r.l'O ;i!r~.i:? 1L~O.L''J .i7^.[!0 117^.00 l\."i
Figure 7. Effect of the width of the instrument f unct ion a (cm1) on resolution
for values of a of 0.5, 0.3, 0.1, 0.05, and 0.01 cm'1, respectively
18
-------
TABLE II. SURACE NUMBER DENSITIES AT T = 2$k K and P = 1013 MBAR
FOR ATMOSPHERIC GASES WITH CONSTAOT MIXING RATIOS
Density
Gas (molecules cm"3)
CH4 U.30 x 1013
02 5.63 x 1019
C02 8.87 x 1015
CO 2.03 x 1012
N20 7.28 x 1012
19
-------
of atmospheric water vapor on the other hand, makes its concentration difficult
to determine. For a collected air spectrum the water vapor content was esti-
mated from the air temperature and relative humidity. Thus each new synthetic
spectrum required a different estimate of the water vapor concentration in
order to obtain a good spectral fit.
Slant Path (Solar Spectra) —
In order to calculate the transmittance of a slant path through the atmosphere
it is necessary to assume vertical profiles through the atmosphere showing the
dependence of temperature, pressure, and mixing ratios of the absorbing gases
as functions of height. Since most of the solar spectra were obtained in the
summer and autumn of 1975 the midlatitude summer atmospheric model described
by McClatchey et al.8 has been used. This model, which is partially reproduced
in Table III shows pressure, temperature, and the mass mixing ratios of water
vapor and ozone as functions of height. It was assumed that the other ab-
sorbing gases have constant volume mixing ratios as shown in Table II. Thus
the number density p^ at height h, pressure P, and temperature T, is related
to the number density PQ^ at the surface (pressure Po and temperature To) by
ok .= Pok p T0/P0T.
The number densities of H.,0 and 03 can be calculated from the mass mixing
ratios given in Table III by finding the partial pressures of these gases Pp
by
Pp =
where c = Mg/Mair is the ratio of the molecular mass Mg of either H20 or 03 to
the mean molecular mass of air, Maj_r, and then by using the perfect gas
equation
Pp=pkR'T
to give
Pk = P mr/c R T.
Synthetic solar spectra such as the one shown in Fig. 8 were calculated by a
computer program which divided the atmosphere between 0 and 30 km into the 13
layers shown in Table IV. In each layer the average temperature, pressure,
and number density were assumed to be those at the arithmetic mean height of
the layer. These values were used to calculate the absorbance of each layer.
The total absorbance of the atmosphere is the sum of the individual absorbances
and from this total value the atmospheric transmittance was calculated [Eq. (?)]•
Then transmittances were convolved with a triangular instrument function by
using Eq. (ll) to give synthetic spectra.
20
-------
TABLE III. MIDLATITUDE SUMMER ATMOSPHERIC MODEL (FROM REF. 8)
Mixing Ratio
Height Pressure Temperature Water Ozone
(ion) (mbar) (K) (kg/kg) (kg/kg)
0 1013.00 29!* 1.2E-2
l 902.00 290 8.6E-3 5.56E-8
2 802.00 285 6.0E-3 6.15E-8
3 710.00 279 3.7E-3 7.01E-8
h 628.00 273 2.UE-3 8.00E-8
5 . 55^.00 267 l.lffi-3 9.15E-8
6 U87.00 261 9.UE-U 1.06E-7
7 k26.00 255 6.3E-U 1.29E-7
8 372.00 2U8 U.OE-U 1.52E-7
9 32^.00 21*2 2.6E-4 1.8UE-7
10 281.00 235 1.5E-4 2.16E-7
n 2^3.00 229 5-9E-5 2.98E-7
12 209.00 222 1.8E-5 3.67E-7
13 179.00 216 6.2E-6 5.20E-7
Ik 153.00 216 U.OE-6 7-33JE-7
15 130.00 216 3.6E-6 9.03E-7
16 111.00 216 3.6E-6 1.17E-6
17 95.00 216 3-6E-6 1.56E-6
18 81.20 216 3.8E-6 2.15E-6
19 69.50 217 U.UE-6 2.88E-6
20 59.50 218 ^.8E-6 3.60E-6
21 51.00 219 6.3E-6 U.U7E-6
22 ^3-70 220 7.^E-6 5.21+E-6
23 37.60 222 9.2E-6 5.80E-6
2h 32.20 223 1.2E-5 6.38E-6
25 22.70 22^ 1.6E-5 7.00E-6
30 13.20 231+ 2.7E-5 1.5IE-5
35 6.52 2^5 1.7E-5 l.UUE-5
l»0 3.33 258 1.3E-5 1.23E-5
21
-------
TABLE IV. THICKNESS, MEAN PRESSURE, AND TEMPERATURE OF LAYERS
USED IN SLAM1 PATH COMPUTATION
Mean Height Mean p Mean T
km (km) (mbar) (K)
0-1 0.5 957.5 292.0
1-2 1.5 852.0 287.5
2-3 2.5 756.0 282.0
3-1+ 3-5 669.0 276.0
U-5 U.5 591.0 270.0
5-7 . 6.0 U87.0 261.0
7-9 8.0 372.0 21+8.0
9-11 10.0 281.0 235.0
11-15 13-0 179-0 216.0
15-19 17.0 95.0 216.0
19-23 21.0 51.0 219.0
23-27 25.0 27.7 22)4.0
27-31 29.0 16.1 232.0
22
-------
ATLAS OF GROIM) LEVEL AIR SPECTRA AND SOLAR SPECTRA
During the past two years a series of spectra from MDO cm"1 to 3500 cm"1 of
path lengths of approximately 10, 100, and 1000 m of air at ground level tem-
perature and pressure have been obtained by using a 21 m stainless steel
multiple traversal cell. The short air path spectra were obtained by ob-
serving the absorption due to the approximately 8.5 m air path between the
source, interferometer, and detector. Four traversals of the cell, when
filled with air, increased the air paths to 92 m and, by using kQ traversals
of the cell, a total path length of 993 m was obtained. In addition to these
ground level, homogeneous path spectra, solar spectra have been obtained between
^4-00 and 3500 cm"1 for solar zenith angles ranging from kO° to more than 87° •
The effective atmospheric path length for small zenith angles is equivalent to
about 8 km at surface temperature and pressure; this increases to more than 50
km effective path length as the solar zenith angle approaches 90°•
Examples of these spectra are shown in Figs. 8-11, each of which covers a
28 oa"1 interval. Tables V and VI give some information for these figures.
The upper three curves are spectra of 8.5 m, 92 m, and 993 m of air at ambient
temperature and pressure obtained using the 21 m multiple traversal cell. The
two lower curves are solar spectra recorded for solar zenith distance of appro-
ximately 1+0° and 85°. At the present time we are compiling an abbreviated
atlas of spectra which will cover the region from 700 cm"1 to 1500 cm"1, each
page of which will "be similar to these figures.
The principal atmospheric absorption features in this spectral region are
bands of H20, C02, 03, CIU, and N20. These figures show how individual absorp-
tion features increase in intensity as the absorbing path is increased. They
also show that even for the longest atmospheric paths, taken with low sun,
there are spectral regions with high transmittance in which features of other
absorbing gases can be sought.
We have made comparisons of the appearances of the several thousand features
in the spectral regions investigated, between different spectra. The spectra
obtained for a given path length with our instrument on different days and at
different seasons are very similar. There are differences in the intensities
of 03 lines (in solar spectra) and H20 lines on spectra taken on different
days and at different seasons because of the variabilities in the abundances
of these gases.
One of the reasons for undertaking this study was to search for absorption
features of atmospheric pollutants. This can be done by comparing spectra of
a suspected pollutant with the atmospheric spectra and searching for character-
istic absorption features as described earlier in this chapter. It is often
found that the spectral regions when the strongest and most characteristic
features occur also contain absorption lines of other gases. These must be
identified if the existence or otherwise of the pollutant is to be verified.
These identifications can be carried out in a number of ways. For example,
we have compared our spectra with atmospheric spectra of similar resolution
obtained by other workers.17'18'19 The agreement in general is good although
some differences have been noted. We have also attempted to compare the
23
-------
TABLE V. INFORMATION CONCERNING SPECTRA SHOWN IN FIGURES 8-11
a
b
c
d
e
Path
Length
(m)
8.5
92.0
993.0
Solar
Zenith Angle
87
Date
(1975)
Nov Ik
Nov 13
Nov 6
Sept 30
Oct 3
.TABLE VI. PARAMETERS FOR SYNTHETIC SPECTRA IN FIGURES 8A-UA
Molecular Number Density
(No./cm3)
H20
(xlO17)
a 3.85
b 3.85
c 3-85
C02
(xl0ls)
8.87
8.87
8.87
CO
(xlO12)
2.03
2.03
2.03
N20
(xlO12)
7.28
7.28
7.28
k.30
k.30
0
Halfwidth of
Triangular
Convolution
.A 3
(xlO13) (xlO11) Function
6.70
6.70
6.70
.025
.025
.025
(K)
293
293
293
P
(mbar)
1013
1013
1013
d See
c Table
III
Surface Molecular Number
Density
(No./cm3)
8.87
8.87
2.03
2.03
7.28
7-28
U.30
See
Table
III
.060
.120
2k
-------
T
T
T
f\
/^r^ pr*rT'
^r^
f
i i I i l I : l
110C.SC 1104.00 1103.00 1112.00 1115.00 1120.00 1124.CO 1125.CO
WflVENUMBER
Figva'e 8a. Synthetic atmospheric spectra "between 1100 and 1128 cm"1 (see Table VI)
25
-------
Figvire 8. Atmospheric spectra "between 1100 and 1128 cm'1 (see Table V)
26
-------
I \l
A
1121J.OO 1129.CC 1132. DO 1130.00
i • i
1140.00 1144. CO 1145.00 115?. OG
WRVENUMBER
Figure 9a. Synthetic spectra between 112U and 1152 cm
-i
27
-------
Y
Figure 9- Atmospheric spectra between 112^ and 1152 cm"1
28
-------
I"
J
f
r, A A,
150.00 1154.00 liS3.Cn 1152.00 1156.00 1170.00 1)74.00 117S.CG
NflVENUMBER
Figure lOa. Synthetic spectra between 1150 and H?8 cm'
-i
29
-------
-^V^yV. ^A>^*~V~^WW^V
—y
Figure 10. Atmospheric spectra between 1150 and 1178 cm"-
30
-------
T
T
i—^^ r
71]. I'D 117S.CIO
118? . CO 1135. CC ; 190. 00 1 194 . DO \ 1.0*. PC 1 20? . 00
WflVENUMBER
Figure Ha. Synthetic spectra between 1171* and 1202 csT1
31
-------
, ^/*/**V*~V-»VA^VO*AV-**A~^^
A
Figure 11. Atmospheric spectra between 11?U and 1202 cm
n-1
32
-------
atmospheric spectra with spectra of individual gases obtained with our instru-
ment. These comparisons have confirmed that essentially all the features in
the atmospheric spectra are due to known permanent atmospheric constituents.
However, such direct comparisons cannot be made to identify water vapor absorp-
tion features in low sun solar spectra because the amount of water vapor in the
path traversed by solar radiation is much greater than can be obtained in
laboratory spectra. Although some of the H20 lines have been identified in
spectra taken by other observers20 and most can be predicted from the actual
models21 of water vapor molecular parameters such methods of identification
of absorption features in our spectra can be ambiguous and other methods of
identification such as those described below must be used.
Each atmospheric gas gives rise to absorption features characterized by their
spectral positions and intensity. The intensity in a spectrum depends on the
path length-mixing ratio product. For the ground-level atmospheric spectra,
when the sample is at a uniform temperature and pressure it is possible to
estimate the abundance of a given constituent by comparing the intensities of
its bands with other laboratory spectra of the gas when the path length and
concentration are known.
The abundances of C02, CO, N20, CH4, and H20 in the ground-level air samples
whose spectra were recorded can be obtained in this way. However, it is not
possible to estimate the abundances of these gases by comparing the intensities
of these bands in solar spectra with laboratory spectra .because of the varying
temperature and pressure along the path traversed by solar radiation. More
sophisticated techniques are required to estimate abundances of atmospheric
gases from these spectra, such as the calculation of synthetic spectra as
described earlier in this chapter.
Spectra similar to those shown in Figs. 8-11 can be analyzed to search for
the appearance of features of suspected trace gases. We have analyzed these
spectra for evidence of various contaminants such as HN02, HN03, HCOOH,
fluorocarbon-11, fluorocarbon-12, and fluorocarbon-22. At this time only
fluorocarbon-12 (CC12F2) has definitely been identified. While it is possible
to set upper limits on the amounts of these gases in ground-level air by di-
rect comparison of such spectra with laboratory spectra, it is not possible to
use such direct comparisons to set upper limits to the concentrations of un-
detected gases in solar spectra since the atmospheric profiles of pressure and
temperature are not reproducible in laboratory spectra.
Although the spectra shown in Figs. 8-11 are for a limited number of path
lengths and spectral resolutions they were recorded digitally on magnetic tape.
Such spectra can be manipulated with a digital computer to produce spectra
with degraded resolution or with different apparent absorber concentrations.
These artifical spectra can be used to predict the transmittance of atmospheric
paths of different lengths and at the same or lower spectral resolutions than
those originally recorded. These manipulations do not allow improved resolu-
tion to be obtained, neither can the absorptance of a slant path through the
atmosphere be simulated.
However, if the line parameters are available as in the AFCRL line listing,8
synthetic spectra which simulate the actual transmittance can be calculated,
33
-------
as described earlier. We have used these techniques to calculate synthetic
spectra of both homogeneous ground-level atmospheric paths and also slant
paths through the atmosphere. Examples of these synthetic spectra are shown
in the overlays to Figs. 8-11 which are labelled Figs. 8a-lla, respectively.
These synthetic spectra have a spectral resolution similar to the actual
spectra and the amounts of the absorbing gases have been chosen to give an
absorptance similar to that observed. It is seen that the general agreement
between the two sets of spectra for all five spectra on each page, including
the solar spectra, is excellent. This range of path lengths is from 10 m to
the equivalent of 105 m, a ratio of 104. One set of line parameters was used
for all the calculations. It is also seen that there are some disagreements,
of which some must be attributed to inaccurate line parameters. Improved
line parameters can be obtained from the experimental measurements.
Thus an important use of the air atlas is to allow improved line parameters
to be obtained. Once accurate line parameters an. available it is possible to
calculate synthetic spectra for any type of path and for any spectral
resolution. These synthetic spectra can then be used for comparison with
actual spectra to estimate abundances of any atmospheric constituent from any
spectrum, including low sun spectra. Such synthetic spectra will allow unpre-
dicted features to be easily identified.
The experience gained so far in comparing synthetic and actual spectra indicates
that some adjustments to the AFCRL compilation are required to obtain better
agreement with experimental data. It has also been found that the compilation
contains only information about seven permanent gases. No line parameters for
possible pollutant gases are included. Our investigations into the identifi-
cation of halocarbon and other features in solar spectra have shown the urgent
need to be able to create synthetic spectra of these gases. This information
is not at present available and must be collected.
-------
DESIGN OF ABSORPTION CELL TO SIMULATE THE ENVIRONMENT OF THE STRATOSPHERE
Another objective proposed for this work is to test the chemistry of the Freon-
03 - air system under conditions which simulate as many features of the strato-
sphere as possible. This involves four basic components; i.e., temperature,
pressure, gas composition, and radiation field. Since our laboratory does not
now possess a chamber or source suitable for these studies, it has been neces-
sary to design a new photochemical reaction chamber and short wavelength source
with which to observe the direct photolysis of the Freons in air at strato-
spheric temperatures and pressures.
During the initial phases of the cell design, it was found that Professor R.
K. Long, in the Department of Electrical Engineering at The Ohio State Univer-
sity, was interested in building a cell with characteristics similar to those
required in this study since he wished to measure the transmittance of infra-
red laser radiation through ozone under simulated stratospheric conditions.
It was therefore decided to build a single cell which can be used for both
investigations.
Although both groups are interested in obtaining long path lengths by using
multiple traversal optics it was recognized that the problems associated with
cooling the system increased rapidly as the size of the system increased. An
initial decision was therefore made to use a set of White22 multiple traversal
system mirrors with a 2 m radius of curvature which was already available in
the laboratory. With these mirrors path lengths in excess of 100 m can easily
be obtained in a cell approximately 2 m long and 0.5 m in diameter. This
fixed the rough dimensions of the cell.
The corrosive properties of many of the reactant gases and their expected pro-
ducts fixed the allowable construction materials. The only material with ex-
cellent resistance to 03, HF, C12, CF20, etc. is iridium. Unfortunately the
high cost and extreme difficulty of working with this material necessitated a
compromise. Stainless steel was determined to be the best substitute material
for the main body of the chamber and the mirror mounts and other structures
inside the chamber. Many of the experimenters who have worked with ozone have
reported the formation of CO and C03 inside their cells as the ozone
decays.23}2"*}25 In hopes of reducing the sources of carbon inside our cell
and for improved weld quality, an extra-low-carbon composition, type 3l6
stainless steel, was finally chosen.
There are two main approaches to controlling the temperature of a cell. One
is to wrap coils around the cell and circulate a coolant through the coils.
The coolant is then passed through heat exchangers to control the temperature.
This method uses a relatively small mass of coolant to control the temperature
of the much larger cell mass. The cost and ease of operation of such systems
are relatively independent of the size of the cell to be cooled. Consequently
this approach has been used to cool large cells such as the 10 m multiple
traversal cell designed by Horn and Pimentel26 For this cell a temperature
uniformity of 0.3°C over the length of the cell was reported although tempera-
tures were measured several degrees higher at the end plates. This temperature
difference could be reduced by installing additional cooling coils at the end
plates.
35
-------
The second approach is to place the cell in a bath of coolant which can be
stirred. This approach requires a larger mass of coolant than the first
approach and, since the cell is immersed in the coolant, severe sealing pro-
blems may be encountered. On the other hand, a temperature bath is better
suited to phase change cooling than the first approach. In this process the
energy removed by the coolant produces a phase change (e.g., the melting of
ice or the sublimation of C03) without a corresponding temperature change. .
This is to be contrasted with kinetic cooling in which the removed energy in-
creases the kinetic energy of the coolant molecules and hence its temperature.
The bath approach has been used, for example, by Burch, Gryvnak, and Patty27
to cool aim cell and by Blickensderfer, Ewing, and Leonard28 for a liquid
nitrogen cooled 3.8 m cell.
The need for precise temperature control in chemical reaction studies has long
been recognized. Indeed in a recent study29 of the reaction
H,0 + NO + NO, «i 2HONO
conducted in this laboratory it was found that a 60$ change in the equilibrium
concentration of HONO occurred for a change in temperature from 23° to 30°C.
Although both cooling approaches have merits and disadvantages, it has been
decided to adopt the cooling bath approach in this application. We concluded
that, although this approach was more bulky, it was simpler to operate and
would cause less vibration (which could affect the mirror stability) than a
flowing coolant system. It also appeared that a uniform temperature could be
maintained throughout the chamber more cheaply than with a flowing coolant
system.
A diagram of the proposed photochemical cell and cooling bath is shown in
Fig. 12. The cell will be made from k.fS mm (3/l6") thick plate rolled into a
tube. One elliptical shaped end cap will be welded onto one end and another
will be bolted onto the other end through the flanges shown. One of the flanges
has three concentric grooves cut in it. It is planned that the inner groove will
contain a silicone 0-ring with good resistance to ozone at low temperatures and
that the outer groove will contain a spring-loaded Teflon 0-ring with excellent
resistance to the coolant and the ability to seal at temperatures up to 260°C
and down to cryogenic temperatures. The middle groove will be continuously
pumped to prevent mixing of a sufficient quantity of 03 and coolant vapor to .
produce an explosion. Infrared radiation will be introduced into the cell
through windows at the right end (Fig. 12A). These windows are shown in
Fig. 12B. Ultraviolet radiation will be introduced through the windows shown
in the middle of Fig. 12A.
The cell will be mounted inside the bath shown in Fig. 12C. This is constructed
of 6.35 mm (1/V1) thick steel plate. Ports at the end and in the middle of the
bath will connect to the cell through the stainless steel bellows and teflon
0-ring seals shown in Fig. 13. These are designed to be leak free as far as
the coolant is concerned and 'flexible enough to allow for misalignment between
cell and bath as well as temperature-induced structural deformation. The bath
will be covered with 12.7 mm (1/2") thick plate glass resting on teflon seals
36
-------
A
i'i
252 CM
U)
'TT
Figure 12. Schematic diagrams of the cell and bath. (A) sideview of cell
showing window for ultraviolet radiation, (B) end view of cell
showing windows for infrared radiation, (c) side view of "bath,
(D) end view of bath. Bath is shown without insulation.
-------
HELIARC
WELD
Figure 13. Schematic diagram of typical bellows connection from cell to bath.
(l) cell wall, (2) bath wall, (3) coolant space, (U) bellows,
(5) bellows length-adjustment rod, (6) window mounting plate,
(7) mating flanges for port (6), (8) and (9), teflon 0-rings
38
-------
to contain the coolant vapors and prevent leakage of warm room air into the
bath. Styrofoam insulation 15.21* cm (6 inches) thick will cover the bath on
all sides.
The bath-cell arrangement is estimated to weigh 653.12 kg (1751 Ib) dry and
1137.65 kg (3050 Ib) when full of ethanol. It is estimated that it will re-
quire about 1.12 kg (3 Ib) of dry ice per 0.56 K (°F) of temperature change
and about 1.12 kg (3 Ib) per hour to maintain a temperature of 188 K (-85°C).
Different operating temperatures may be selected by using different coolants
(e.g., H20) at their ice points or sublimation boiling points (e.g., C02). It
is planned to obtain intermediate temperatures by dripping liquid nitrogen
into the coolant and stirring with several motor-driven propellers- immersed
in the liquid. It is hoped that an overall temperature uniformity of 1 K
or better can be achieved.
Since the stainless steel cell has a different coefficient of thermal expansion
than that of the glass mirrors a temperature change will tend to change the
optical alignment. A temperature change of 85 K (85°C)' should produce a con-
traction of 3 mm of the cell with respect to the radius of curvature of the
mirrors. Some provision must then be made to realign the mirrors after the
cell has cooled down. This will be accomplished by feedthroughs built from
Veeco style bellows seal valve assemblies. These are similar to the design of
Blickensdeifer et al., (Ref. 28). These feedthroughs will also enable the
mirror system to be realigned without opening the cell. In order to prevent
condensation on the windows, the boil-off from LN2 will flow over the windows
and flush any moisture from the bellows connection areas.
The optical path of the infrared radiation inside a White multiple traversal
system is well known.22 In the three-mirror system shown in Fig. lU the beam
enters the cell through window W-L and forms an image of the source at M-^.
After multiple reflections inside the cell the beam emerges through window
W2. In a conventional system, the mirrors M2 and M3 are placed close together.
However, in order to introduce the ultraviolet radiation into the cell, we
propose to spread these two mirrors apart as shown. The ultraviolet radiation
from a microwave-excited high-pressure xenon lamp30 (which has a continuum from
150-210 nm31) emerges through a CaF2 window and is introduced into the chamber
through a window in the middle of the cell. The beam is focused on a beam
divider which reflects part of the energy to Mx and part to M2 and M3. Since
the beam divider is at the focal point of all three mirrors, after reflection
from the White mirrors, a parallel beam is produced which travels down to flat
mirrors at the ends of the cell. These reflect the radiation back down the
cell to the White mirrors. By spreading the White mirrors apart, the infrared
path is completely isolated from the ultraviolet optics and no shadows are
formed.
At the present time two-thirds of the purchases of material for the cell have
been made. The stainless steel chamber and mirror mounts are about 60$, com-
pleted and the steel bath and connecting bellows are nearly 80$, completed.
Most of the components for the ultraviolet radiation system have been purchased.
It is estimated that the system will be ready for preliminary testing by the
end of April.
39
-------
BEAM DIVIDER
o
WHITE MIRROR
UV MIRROR
REGION OF INFRARED
RADIATION
Figure lU. Radiation field inside cell (note IK clearance of beam divider)
-------
STUDY OF SOME KEY REACTIONS OF PROBABLE IMPORTANCE
IN PHOTOCHEMICAL SMOG FORMATION
CONSTRUCTION AND PERFORMANCE OF PHOTOCHEMICAL CELL
The major details of the 6 m photochemical cell have been described in the
Final Report under EPA Grant No. R-803075.32 This cell has been completed,
tested, and is now being used to study a variety of photochemical reactions.
The cell is constructed of four 1.5 m lengths of Corning low-expansion boro-
silicate glass tubing giving an overall length of 6 m and an internal diameter
of 30 cm. Surrounding this cell are four banks of black light fluorescent
tubes which emit energy in the UV region (3000/Y - UyOOft). The multiple traver-
sal optical system inside the cell has been described by Hanst.33 It consists
of four rectangular-shaped mirrors and four quadrant-shaped mirrors, each of
which has a radius of curvature of 5-31 m. When photolysis experiments are
being performed, the cell is covered with aluminum foil reflectors. The photo-
chemical cell is fitted with NaCl windows. When the cell is evacuated, an
ultimate pressure of 10~3 Torr can be attained. The leak rate is ~ 0.8 Torr/day.
Additional details of the cell are given below.
Transfer Optics to the Photochemical Cell
A schematic of the optical arrangements used in this work is given in Fig. 15-
Radiation from a Nernst glower (N) is reflected by the spherical mirror ML
(fx = 15 cm) to an off-axis paraboloid M2 (f2 = 27 cm) which results in parallel
radiation on the beam splitter (B.S.) inside the interferometer. This beam
splitter is made of germanium on a KBr substrate. The collimated beam is
amplitude-divided at the beam splitter; half of it is reflected to the movable
mirror M4 and half of it is transmitted to a stationary mirror M5. These two
beams then combine at the beam splitter and exit the interferometer. The
energy, after being reflected by two plane mirrors (M6 and M8) and two spheri-
cal mirrors (M7, f7 = 60 cm; M9, fg=U5 cm), is directed into the long-path
absorption cell and an image is formed at N3 in the plane of the field mirror
and an image of the beam splitter is formed on the top left quadrant at the
other end of the cell. Intermediate images of the Nernst and the beam splitter
are also formed at positions NL and N2; and at B1, respectively. If high
spectral resolution is required, most of the energy from the elongated image
of the source cannot be usefully collected. An aperture should be placed at
either position Nt or N2 to reduce the beam-spreading effect.
When the beam exits the absorption cell, it is focused on the Cu:Ge detector
which is operated at liquid helium temperature.
The entire optical system external to the cell is mounted on a slab of aluminum
and enclosed in a transparent plastic box sealed to be essentially air tight.
The box is purged with dry N2 gas during operation and residual C03 and H20
in the gas are removed by placing traps of KOH and NaOH on the floor of the
box.
Ul
-------
Cu:Ge DETECTOR
ro
LONG PATH CELL
Figure 15. Optical transfer system from interferometer to photochemical cell
-------
Path Length Stability and Signal-to-Noise Ratio
The base path of the photochemical cell is 5-31 m. It has been found that the
highest number of traversals to be used without a significant loss of signal-
to-noise .ratio is 1*8 passes. This gives a total path length of 255 m. The flu-
orescent lamps cause the temperature of the cell to rise about 10 K (10°C) in
1/2 hour above ambient temperature. This results in a shift of the optical
alignment, which in turn affects the signal-to-noise ratio. All the photolysis
experiments reported here were carried out with room air forced into the chan-
nel between the glass surface and the reflectors by a fan. This ensures cool
air circulation and maximum change of temperature of less than 2-3 K (2-3°C)
during the run. A temperature gradient of about 2 K (2°C) is found at the
ends of the cell during photolysis experiments. Under favorable conditions,
a signal-to-noise ratio greater than 100 is usually obtained. Except for the
N02 photolysis experiments in which only one scan is collected for the inter-
ferogram, all other experiments involved five scan-averages per interferogram.
At a resolution of 1 cm"1, the minimum time needed for one scan and storage of
the data is 23 seconds.
Gas Handling and Injection System
The gases are injected into the cell through a stainless steel [6.35 mm
O.D.] injection manifold at five inlets (one at each spacer and one at each
endplate (see Fig. l6). The stainless steel bellows values are mounted close
to the inlets in order to keep reactive gases out of the high surface-to-
volume ratio injection manifold.
The grease and mercury free gas handling system is all glass with teflon
valves (see Fig. 17). Connections are made with stainless steel Cajon unions
employing Viton 0-rings. Low background pressures are achieved by a single
stage oil diffusion pump which is separated from the system by a liquid nitro-
gen trap. Pressures are read on a glass spiral gauge which is calibrated with
one of two Wallace and Tiernan absolute pressure gauges (0-20 mm and 0-800 mm).
The gas handling manifold was designed to have small dead volumes and to allow
positive injections of samples with N2'gas; positive injection capability is
essential for forming multicomponent mixtures in the photolysis cell.
The eight volumes built into the vacuum system (2.7 to 92.** ml) allow injection
of samples of less than 0.01 ppm (0.5 mm in 2.7 ml, cell volume Ul+5 liters).
The volumes were designed for making samples in the low ppm range. For in-
jecting very large concentrations or for samples with low working pressures
(e.g., H2CO), larger volumes can be connected to the system. These add-on
volumes are constructed with two valves to allow positive injection.
Ultraviolet Intensity and Distribution Measurements
The 6.3 m long photolysis cell (30.5 cm interval diameter pyrex tubing) is
surrounded by eight rows of black light fluorescent lamps (G.E. F72T12/BL/HO,
85 watts) for 5-3 m and ten rows of less powerful black lights (G.E. Fl*OBL,
i*0 watts) for the remaining one meter. These lamps are approximately 7 cm from
-------
^0 '
TO GAS
HANDLING SYSTEM
BELLOWS VALVE
BUTTERFLY VALVE
Figure 16. Gas inlet system to 6 m cell
TO PUMP
AND TRAP
-------
VJ1
TO INJECTION LINE
FILL
PORTS
OIL
DIFFUSION
PUMP
LIQ. N2TRAP
Figure 17. Gas handling system for 6 m cell
-------
the outer cell wall. The cell and lights are surrounded.by an external con-
centric shell of aluminum foil approximately 23 cm from the cell wall to in-
crease the intensity and improve the distribution uniformity.
The number and types of lamps were chosen to mimic the solar intensity and
distribution in the photochemically interesting region from 300 to hOO run. In
Fig. 18 the lamp distribution through the cell pyrex wall is compared with the
solar distribution in the region of interest.
In order to get some feel for the radiant energy intensity within the cell and
the intensity uniformity at various positions within the cell a Hewlett-Packard
Radiant Flux Detector (model 833^A) was used. This thermopile detector measures
the incident power in absolute units (W/cm2) and has a flat spectral response
from 300 to 3000 nm (a 1 mm thick Infrasil quartz window covers the thermopile
detector). The detector has an effective field of view of 0.1 sr (20.5° coni-
cal angle) and a detector area of 0.18^ cm2.
When the detector is centered in the cell it measured radiance values between
1.0 and 3-5 mW«cm~2'sr~1 (over all wavelengths) depending on what it sees in
its field of view. A polar plot of detector readings is presented in Fig. 19-
For these readings the detector was mounted on the axis of the cell (0.76 m
from one end) and rotated in a plane containing the axis of the cell (they are
an average of the values for a plane along a row of lights and a plane between
two rows of lights). A similar measurement was performed at the other end of
the cell and gave a very similar profile indicating uniformity of intensity
for these two positions.
When a Corning 7-5^ UV pass filter (2.5 mm thick which transmits from less than
300 to hOO nm) was placed over the detector the radiant intensity was reduced
to about kcrfo of the original reading. We thus conclude that approximately 50%
of the total radiant energy within the cell is in the 300 to UlO nm region of'
interest.
It was observed that the radiant energy did not drop significantly (in the 300
to kOO nm range) as the lights warmed up 10 K (10°C) over a half-hour time
range [the average amount of temperature change we observed with the cooling
we use is 2-3 K (2-3°C) over two hours].
The solar spectral irradiance at the Earth's surface in the wavelength range
290 to 390 nm is about 3.6 mW/cm2 (50° Zenith angle).34 Above the Earth's
surface reflection from the surface would increase this value. At one position
in the cell we calculate the irradiance (from 300 to kl.0 nm) to be 9-^- mW/cm2.
This value is for a planar element along the cell axis and includes contribu-
tions from both above and below the plane. This irradiance corresponds to a
.photon flux irradiance of 1.7 x 1016 photons•cm~2«s~1 for 360 nm photons.
This value will allow us to calculate a rate constant for the photodecomposi-
tion of NC>2 to compare with the experimentally observed rate of 0.52 min"1
(see the following section). The general expression for the rate constant is
k= (2.303) (IQ) (a) (*).
-------
300
Figure 18.
32O
340
360
380
4OO
WAVELENGTH (nm)
(a) Spectral distribution of fluorescent lamp output after passing through the
glass vails of the cell; (b) relative spectral distribution of solar energy at
ground level
-------
270
90'
CELL AXIS
Figure 19. Radiant flux detector readings (in EW/cm2-sr) measured in a plane
containing the cell axis. This average curve is the mean of the
readings obtained when a rott of lights is in or out of the field
of view
-------
A weighted average of the absorption coefficient (a) and quantum yield (j>) for
NOo over the light output distribution curve within the cell (see Fig. lo)
gives a value of 127 i, mole'1 cm'1. This gives a value of 0.50 miiT for an
irradiance of 1.7 x 10ie photons cnr2^-1 which agrees favorably with the
experimentally rate constant of 0.52 min'1. It should also be mentioned that
this value is typical for that found in sunlight (see the following section).
Estimation of UV Cell Radiant Intensity from Measurements of the Rate of N0g
The photolysis of nitrogen dioxide in the N0x-polluted atmospheres is the major
source of 0-atoms and its occurrence results in the generation of 03 through
the following reactions:
HO, + hv(AN02 + 02
In air at the usual ambient levels of the common pollutants, the above reactions
are largely responsible for establishing the ozone concentration. To a first
approximation, relation 15 applies;
[N02]k12
[03] = - (15)
[N0]k14
Thus the apparent first-order rate constant kig is a very important parameter in
smog chemistry. It is a function of the intensity of ultraviolet light present
in the lower atmosphere which is within the N02 absorption region and which is
sufficiently energetic to result in the dissociation of N02. A recent theoretical
determination of the constant k12 for the lower atmosphere near sea level is
shown in Fig. 20 as a function of solar zenith angle.
It has become general practice to determine the experimental value of kL2 in
smog chambers to compare the degree of correspondence of the chamber intensity
with that of the solar intensity within the atmosphere. We have carried out this
determination in two types of experiments. The first and the most direct method
involved the photolysis of pure NOp gas at low pressures within the reaction
chamber. For those conditions only the following reactions are important:
N02 + hv ->0 + NO (12)
and
0 4- NOp -4 WO + 0;
(16)
-------
(Jones, Bayes $ data)
(Pitts, Shar^, Chan data)
O.I -
Figure 20. The theoretical magnitude of the rate constant k, for the formation
of OC3?) atoms through N02 photolysis at different solar zenith
angles as estimated for the atmosphere of Los Angeles; values are
calculated for the conditions of no attenuation of the ultraviolet
within the polluted layer and no reflection from the surface of
Los Angeles
-------
The reaction of 0-atoms with N02 is the only fate of 0 for these conditions, so
relation (17) follows:
-d[N02]/dt = 2k12[NOp] (17)
and
-d[N02]/[N02] - 2kl2dt.
Upon integration of (18) between initial [N02]o a^d [N02]t at time t, Eq. 19
results ;
2n([NOp]0/[N02]t) = 2k12t = MAo/At). (19)
If we restrict our measurement to that range of [N02] for which the Beer-Lambert
relation applies, then the absorbance (A) is proportional to concentration, and
a linear relation "between ^(AQ/A^.) and time is expected with a slope equal to
2*12.
Absorbance data for N02 as a function of time were determined in a series of
experiments in this work. The data from one such experiment are shown in Fig. 21.
It can be seen that the plot is linear within the experimental error. Values of
kl2 derived from this and similar experiments are shown in Table VII. From these
data we conclude that ki2 = 0.515 ± 0.07 min"1.
It could be argued that the photolysis of N02 at low pressures may not reflect
accurately the rate of 0-atom production from reaction (12) in air at 1 atm.
pressure, it is possible that collisional deactivation of N02 excited at the
lower wavelengths, near the dissociation limit of N02, may occur, and hence the
use of k12 derived from low-pressure experiments may be unjustified.38 To test
this hypothesis and to derive a meaningful value for k12 applicable to the system
of greatest interest in our work, air at atmospheric pressure, additional experi-
ments were carried out photolyzing small amounts of N02 in a synthetic air mixture
containing Iko Torr of 02 and 560 Torr of N2. The results of two such experiments
are shown in Figs. 22 and 23. In this case there are a number of reactions which
occur and the treatment of the results to derive values for ki2 is considerably
more complicated. The following reactions must be considered:
NOp. + hv -*0 + NO (12)
0+02 + M-*03 + M (13)
03 + NO ->N02 + 02 (1*0
03 + N02 ->N03 + 02 (20)
N03 + NO ->2N02 (21)
N03 + N02 -*N205 (22)
N205 ->N02 + N03 (23)
51
-------
TIME (sec)
Figure 21. A plot of 0rc[A0/At] vs time for NO? photolysis at low pressure
[N03]0 = 30.U ppm
-------
TABLE VII. RATE DATA ON THE PHOTOLYSIS OF NOp AT LOW PRESSURE
[N02]n
(ppm)
13. ^L
17.7k
28.70
30.^5
30.1*5
36.85
(min11)
1.16
0.913
0.882
0.923
1.261*
1.065
1.113
0.919
fo linear correlation
99.2
92.0
98.0
97.7
100
99.3
100
98.0
Best Estimate: 2ki2 = 1.03 + 0.1^ min'
-i
53
-------
70
60
E 50
CL
O.
-------
70
60
50
Q.
Q.
-40
' cvi
O
'30
20
10
0
4 6
TIME.min
8
10
Figure 23. Plot of [NOa] versus time of irradiation in the. photochemical cell;
[N02]0 = W3.1 ppm, P0 = ihO Torr, PK;> = 560 Torr; solid line is
computer simulation of the [N02] "time dependence using k12 = 0.515
min"1
55
-------
N03 + N02 -»02 + N02 + NO (2k)
N03 + N03 -4 02 + N02 N02 (25)
N205 + H20 ->HOW02 + HON02 (26)
0 + N02 -»NO + 02 (2?)
0 + N02 + M ->N03 + M (28)
NO + NO + 02 ->N02 + N02 (29)
All of the rate constants for the reactions in the above sequence are reasonably
well known39 and a computer simulation of the experimental systems can be made
using these constants and the value of k12 = 0.515 min"1, derived directly from
the low-pressure 'experiments.
The results of such calculations are shown in Figs. 22 and 23 as the solid lines.
The computer -generated variation of [N02] with time fits the data well within the
uncertainty of these experiments.
Tfinus we conclude that the value of kl2 = 0.515 + 0.02 min"1 is a reliable estimate
of the N0a photolysis rate constant in the reaction vessel for our usual condi-
tions (700 Torr air, 23 + 2°C). This estimate is gratifying in that it compares
well with the theoretical value for k12 in sunlight at z = kO° (see Fig. 20).
The intensity of ultraviolet light within the photochemical cell designed in
this work is very similar to that of the sunlight within the lower atmosphere
and both are significantly higher than that found in most of the other smog
chambers constructed in recent years. In these cases ki2 values vary from about
0.22 to O.kO min"1.
PHOTOLYSIS EXPERIMENTS
Theoretical Estimation of Photochemical Decomposition Rates from Existing
Absorption and Quantum Yield Data
In the variety of experiments which are planned for this work a number of light-
sensitive compounds will be employed. Estimates of the apparent first-order
rate constants have been made for several of these species utilizing published
extinction data, quantum yield data, and relative intensity versus wavelength
measurements made utilizing our cells and lamps (see Fig. 18) . The theoretical
first-order decomposition rate constants, k^, for all systems were compared
with that for N02 taking our measured value of kl2 = 0.51 min"1 and the ratios
of the integrals FI = J^2 e(A) (?0 I°(?v)d7v over the entire non-zero range of
values for the ith species and for N02 j
56
-------
These data are given in Table VIII with the corresponding reaction to which they
apply.
These data are useful in experiment planning and selection of reactant concen-
tration necessary for optimum rate measurements in our system. In those
cases where experimental direct measurements of these quantities have been made
they appear to be in reasonable accord with the theoretical estimates of
Table VIII. However, the one case of nitrous acid appears to be significantly
different from that anticipated from the existing absorption data as derived
in Table VIII (see the following section).
Nitrous Acid Photolysis
Nitrous acid has been chosen as the most convenient source of HO radicals to
be employed in the various phases of this work which are planned. As a con-
sequence we have photolyzed HONO mixtures in an attempt to establish the first-
order rate constant for reaction 30 in our photochemical system;
HOE) + hv ^ HO + NO (30)
Mixtures of HONO (and necessary small impurity levels of N02, NO, and H20
introduced in the HONO preparation) with varied amounts of CO in N2 at 700 Torr
were photolyzed to determine k3o. The results for four experiments at differ-
ent CO levels are shown in Figs. 2^-27. The characteristic frequencies and the
absorption coefficients determined for the various reactants and products are
summarized in Table IX. (Here1 base e lagarithms are employed in defining the
absorption coefficients.) The experimentally measured points of nitrous acid
loss and carbon dioxide formation can be observed from these data. The solid
lines drawn represent the computer-simulated rates of HONO loss and C02 for-
mation using k30 = 0.070 min"1 and the reaction scheme (12) - (30) together
with the following additional reactions which are required for this more com-
plicated system:
HO + NO + M -> HONO + M (31)
HO + N02 + M -> HON02 + M (32)
HO + HONO -> H20 + N02 (33)
HO + CO -> H + C02 (31*-)
H + 02 + M -» H02 + M (35)
H02 + NO -> HO + N02 (36)
H02 + N02 -» HONO + 02 (37)
2H02 -> H202 + 02 (38)
57
-------
TABLE VIII. THEORETICAL ESTIMATES OF FIRST-ORDER DECAY CONSTANTS•FOR THE
FHOTODECOMFOSITION OF SOME LIGHT-SENSITIVE COMPOUNDS OF
INTEREST IN THIS WORK (PHOTOCHEMICAL CHAMBER)
Compound
Reaction
QO
N02
HONO
(CH3)2N2
CH20
CH2
CH3
H20
0
ONO
2
N02 + -hv -4 0 + NO
HONO + hv -» HO + NO
CH3
N2CH3 + hv -» 2CH3 + N2
CH20 +
CH2
CH3
H20
0 +
ONO
2 +
hv ->
hv -4
+ hv
hv -4
H + CHO
H2 + CO
-» CH30 + NO
2HO
(arbitrary
51693
897
111
290
8007
86
units)
• 9
.9
.7
(min-1)
0.51
0
0
0
0
0
.0088
.0011
.0029
.079
.00086
-------
TABLE IX. ABSORPTION COEFFICIENTS DETERMINED IN THIS WORK (PRESSURE N2
700 TORRJ TEMPERATURE, 23 + 2°; RESOLUTION, 1 CM'1
Compound Peak Location Absorption Coefficient
(cm-1)
C02
NO
N02
HONO(cis)
HONO( trans)
720
1876
826
853
126U
1.70 x 10-4
2.7^ x 10-*
3.HO x 10-5
6.07 x lO'3
3.8U x 10-3
TABLE X. INITIAL CONCENTRATIONS OF THE HONO PHOTOLYSIS EXPERIMENTS
Initial Concentration (ppm)
[co] [HONO] [NO] [N02] [o2]* [N2]
520 3-26 7.H 10.78 300 9.21 x 10E
1039 3.02 2.32 5.63 300 9.21 x 105
1559 1.81 h.h3 6.68 225 9.21 x 105
2078 3.70 2.H5 U.18 700 9.21 -A 1C5
•^estimated
59
-------
3.0
Q.
Q.
O
2.0
O
X
1.0
12.0
8.0
E
Q.
0.
rs
CJ
O
O
4.0
0
10 20 30 40 50 60
TIME,min.
Figure 2k. The time dependence of [HONO] and [C02] in the photolysis of an
HONO-NOX-CO mixture; [CO] = 520 ppm and other reactants are
given in Table X
60
-------
0
Figure 25.
20 40
TIME,min -
30
20 g-
OJ
o
o
0
60
The time dependence of [HONO] and [C02] in the photolysis of an
HONO-NOX-CO mixture; [C0]o = 1039 PP111 an(i other reactants are
given in Table X
61
-------
TIME, min
Figure 26.
The time dependence of [HOHO] and [C02] in the photolysis of an
HONO-NOX-CO mixture; [C0]o = 1559 PP™ and other reactants are
given in Table X
62
-------
3.8
3-0
• 2.0
CL
Q.
O
O
X
1.0
30,
20.
OJ
O
O
10.
30
60
TIME,min.
Figure 2?. The time dependence of
HONO-NOX-H20 mixture; [C0]o
given in Table X
and [C02] in the photolysis of an
20?8 ppm and other reactants are
63
-------
H + CO + M -> HCO + M (39)
HONO + 0 -> HO + N02 (40)
H202 + hv -> 2HO (41)
HCO + 02 -4 H02 + CO (1(2)
H+NO+M-»HNO+M (43)
HNO + 02 -» H02 + NO (44)
H + N02 -» HO + N02 (45)
HNO + hv -* H + NO (46)
0 + CO + M -» C02 + M (47)
0 + NO + M -» N02 + M (48)
HNO + HNO -> N20 + H20 (49)
HCO + 02 + M -» HC002 + M (50)
HC002 + NO -» HC02 + N02 (51) .
HC02 + 02 -> H02 + C02 (52)
HCOa -» H + C02 (53)
H + HONO -» H2 + N02 (54)
It can be seen from the data of Figs. 24-27 that the rate of C02 formation in
these systems is much greater than the primary rate of HO formation from
HONO decay. It became evident to us that small amounts of 02 are inadvert-
ently added in preparing the mixtures for photolysis and reactions of H-atoms
formed in (34) with 02, NO, N02, and CO must occur to generate ultimately more
HO radicals through (36), (45), and the other reactions shown. The amount of
02 assumed to be present affected the Rco2 estimated rather strongly, while it
had little effect in the HONO decomposition reaction rate for (30). By curve
matching of the C02 formed we could estimate the amount of impurity 02 present
at the start of each run. These data are summarized in Table X which also
gives the initial concentration measured for each of the other reactants.
The value of kso which was necessary to match the HONO loss rate was not very
sensitive to the 02 level chosen or the fit of the C02 rate data. The number
which is derived from the best fit of all the data was k30 = 0.070 ± 0.004 min"1.
It can be seen that this is a factor of two bigger than the value anticipated
from the theoretical treatment of the extinction data described earlier and
summarized in Table VIII. It is somewhat surprising that the new extinction data
for HONO presented by Johnston and Graham40 show an integrated absorption for
HONO which is a factor of two lower than that observed for CH3ONO, the next
64
-------
higher homolog in the series. It is possible that the absolute value of the
extinctions for HOWO are in error by a factor of two. Rote our measured k30
is about equal to that expected for CH3ORO photolysis (see Table VIII). In any
case our experimental estimate of k30 is considered reliable for our conditions
and will be employed in several phases of our studies of the following years.
Preliminary Photochemical Studies of Special Interest in The Formulation of
Smog Mechanisms
The photodecomposition of CH30-air mixtures, CH20-HONO-NO-N02-air mixtures,
(CH3)2N2-NO-W02-air mixtures [(CH3)30)2N2-WO-N02-air mixtures, and HONO, H202,
NO-air mixtures were carried out to test the feasibility of several experiments
which are planned in this continuing study. In those cases we searched the
products for evidence of metastable transients, peroxyformylnitrate, CH302N02,
(CH3)3C02N02, and H02IK)2 which are postulated as possible products of the
reactions of N02 with HC002, CH302, (CH3)3C02, and H02 radicals (see Ref. 38).
Transient absorptions which may correspond to some of those species have been
seen to decay in the dark, but the conclusions which can be formed concerning
the nature of these species remains uncertain now. Further analysis of those
spectra through peak ratioing and subtraction of known compound spectra are
in progress.
-------
LIBRARY OF SPECTRA
One of the objectives of this program has been the collection of a library of
spectra of molecules of importance in atmospheric processes. Although spectra
of most of the gases investigated have been reported elsewhere the results
have been of limited value to this investigation because the spectral resolu-
tion is usually less than employed here and because quantitative information
concerning the physical conditions of the sample were not given. We have
attempted to collect spectra which can be used to give quantitative estimates
of the amount of gas in other samples. Most of the spectra reported.here were .
obtained with a spectral resolution of 1.0 cm'1, since this has proved most
useful in the photochemical studies. We plan to repeat the measurements at
higher resolution when necessary. Two spectra are shown for each gas so that
estimates of the growth of features with increasing gas concentration can be
obtained. Most of the spectra cover the region from about 700 to 1500 cm"1.
The increase in the noise level near 700 cm"1 and between 1*+00 and 1800 cm"1
is caused by absorption by residual C03 and H?0 in the path.
Some of these spectra have been used to obtain quantitative estimates of the
amounts of gas in other mixtures. These estimates were made by measuring the
transmittances of particularly sharp features in the spectra. The corresponding
absorbance is then plotted against the amount of gas in the path to give a
curve of growth.
These curves of growth are dependent on the spectral resolution of the spectra
and cannot be used directly by other observers. We have attempted to measure
the integrated band absorptance of some features by using the arithmetic pack-
age associated with the IFTS system. The integrated band absorption is inde-
pendent of the spectral resolution but uncertainties in estimating the zero
absorptance line and overlapping of other bands was often found to limit the
usefulness of this method.
Most of the spectra are for samples pressure broadened with N2 to 700 Torr.
For most infrared bands the integrated band absorptance and the spectral ab-
sorptance at a particular frequency both increase as the total pressure
increases. Some of the halocarbon spectra appear to show very little pressure
broadening. However, care must be taken when using the data presented to
estimate the amounts of a given gas in a sample at a temperature or pressure
different to that measured. A series of spectra of halogenated methanes has
been collected. Fig 28 shows a chart of progressive chlorination and fluorina-
tion of methane.
Spectra have been recorded for most of the compounds on this chart and are
shown in Figs. 29-39. At the time that these spectra were collected, the gas-
handling system was not yet operational and the amount of gas injected into
the glass cell was unknown. These spectra are useful as a survey of band
positions and relative intensities. Spectral resolution for the chloro-
66
-------
as
CH
Cl addition / / \ \ F addition
c,
CHOI CHF
'Zv
CH2CI2 CH2CIF
'3v
CHCL CHCLF CHCIF
^
F
CCL CCLF
2v
^3v
CCIF
Figure 28. Chart of chlorine and fluorine addition to methane
-------
IUU
O
• CO
CO
1
1—
O
100
ON
CO
O
CO
CO
s
cr
h-
o
C H4 1
i i i i > i , i i i
i
v-r - ril^n|
f[|
>, i
\
i , i i i i i i i
'l|f IF
rrffp
1 '. ' • V | . '
1 ; i '
1
WWW/
1 :'l
i
111 1 11
11
If
T^I
TT
w
i '
'^1
1
r rf .»
! 1
Tiif
;
I
i
700
800 900 1000 1100 1200 1300
WAVENUMBER (cm")
Figure 29." High-resolution spectra of methane
1400
1500
-------
100
vo
o
CO
CO
CO
ad
\—
\°
too
O
CO
CO
CO
tr
h-
-vp
0
CH3F
_L
rTn^'ViV",
1 'lib I1'!'1
700 800 900 1000 1100 1200 I30O
WAVENUMBER (cm-)
Figure 30. High-resolution spectra of methyl-fluoride
1400
1500
-------
100
g
CO
CO
2
CO
cr
h-
0
100
<>V£
700
800
900
1000
1100
1200
1300
1400
1500
WAVE NUMBER (cm-)
Figxire 31. High-resolution spectra of dichloro-methane
-------
too
g
CO
CO
2
CO
cr
i—
CHCI-
0 L_.
700
800 900 1000 MOO 1200
WAVENUMBER (cm-')
Figure 32. High-resolution spectra of chloroform
1300
1400
1500
-------
ro
100
o
CO
CO
05
or
h-
o
100
0
CHCI2F
700
800
900
1000 1100 1200
WAVENUMBER (cm-')
1300
i L
I
rvnryf,
1400
1500
Figure 33. High-resolution spectra of fluorocarbon-21
-------
CO
IUO
2"
0
CO
«>
zt
tr
0
100
o
CO
CO
CO
z:
o:
i—
0
^rf
jvf ""x i\y' — ' ' "f-' '""' i' ' "'''i""''i
1 1 M / 1
CHCIF2 ' .
1
•
•
1,1,1,1,1,1,1,
""^\(^ " \ ' f^ \(\/^^P^'T|?
^1
'
i
\/ V ;
W
v
1
'
.
1,1,1,1,1,1,.
'00 800 900 1000 MOO 1200 1300 I4OO 150
WAVENUMBER (cm-)
Figure 3U. High-resolution spectra of fluorocarbon-22
-------
100
o
I
0 L
100 fc
CF3H
o
cr
h- ;
^P
o
I
fr!
TOO
80O
900
1000 1100 1200
WAVENUMBER (cm'1)
1300
1400
1500
Figure 35- High-resolution spectra of fluoroform
-------
100
| ---- r-
0
CO
CO
i
z
<
-------
100
O
CO
CO
. CO
a:
i—
0
100
-
CO
CO
CO
7L
or
\-
0
i
A
i
CCI3F
i
i^Hr^v
\
I
i
i
00 800
1
I
w
1
1
r V
i
i , i i i , i .1
^*nMl-'~ X^" " — ' ' ~:i - A /^" ^ '^t ^
\ [
I1'
. ! '
i . i . • i , i , i
900 1000 1100 1200 1300
1 1
I'''
I
1
1
ft
q
ff\
m\
IP
I
¥
\
i
i
i
'i
i
_i_
i
i
1
n
f
!;
1400 I50(
WAVENUMBER (cm'1)
Figure 37. High-resolution spectra of fluorocarbon-11
-------
100 f
O
CO
CO
5
CO
cr
CCI2F2
«4«i,*WTWS|#"~ l| ' (;' l(| I '
'1 • •
4
1
1 i 1
1200 1300
f1
h
i '
1 •»
f
1
'
1
\
I '
I4OO
I1
Mi ' ',
;r j
150
WAVENUMBER (cm")
Figvire 38. High-resolution spectra of fluorocarbon-12
-------
100 f •
1 •
i
X* .
o
CO
CO;
oo; CF4
o:
i —
i
i
o1
100
^^
0
CO
CO
*=;
CO
Jjj
cr
i
i
o^
o
1
700
• — • -n >A
i
1,1,1,1.1,
^
.
1,1,1,1,1,
800 900 1000 1100 1200
r ' IT ' i
f
1
\
\ \ •
\
\
\
f;
i J
i
'
1
ff$^Vf?F
1 i,
'j - i
M
.
| !
j ]
1
i 1
j
I
\
J
i
f '
.I ,
1300
"•V*
1
TT
I. 1
F^
n
!
i
j
1
1
1400
T
1
If
1
i
,
r
i
i
t
I
"ii "•
:•
i ' i
i i
'1
•
!
,
; 1
, i
.
i
i
i
•
i
i
i
i
.
•
i
i
.
I5CK
Figure 39.
WAVENUMBER (cm-')
High-resolution spectra of carbon tetrafluoride
-------
fluorocarbons is 0.125 cm"1, and the spectral region extends from 700 cm"1 to
1500 cm""1 for all spectra.
Figures hd through 53 show spectra of several compounds of interest in air
pollution studies. These spectra were recorded at one wave number resolution
over the spectral region from 600 to 3500 cm~1j vith the exception of a few
compounds; only the region from 700 to 1500 cm"1 is shown in this library
tabulation.
79
-------
o
CO
COj
H202
^•L
or
t-
0
i - ' ' ' '
i , i . i .1 .1 , I , i ,
CO
CO
o:
0
700
800
900
1000 1100 1200
WAVENUMBEFUcnr1)
1300
1400
1500
Figure kO. Infrared spectrum of hydrogen peroxide; Resolution = 1 cm"1, top
spectrum = ?.U ppm, bottom spectrum =20.8 ppm, total pressure is
less than 20 mm N2) path length is 170 m
-------
700
800
Figure Ul.
900
1000 1100 1200
WAVENUMBER (cm-').
1300
1400
1500
Infrared spectrum of nitrous acid; Resolution = 1 cm'1, top
spectrum = 0.9 ppm, bottom spectrum = 3.0 ppm, total pressure is
700. mm N3, path length is 170 m
-------
CO
CO
CO
2
<
cc
CH3OISI02
00
ro
700
800
900
1000
1100
1200
1300
1400
WAVENUMBER (cm-)
Figure k2. Infrared spectrum of methyl nitrate; .Resolution = 1 cm"1, top
spectrum = 3.1 ppm, bottom spectrum = 19-0 ppm, total pressure is
700 mm N2, path length is 85 m
1500
-------
00
-------
lOOi
o
if)
co!
2
CO
2!
<
o:
CO
A »4l«A
100
CO
CO
S
cr
1500
1600
Figure
M, •
1700
1800
1900
2000
WAVENUMBER (cm-)
i 1
2100
2200
Infrared spectrum of carbon monoxide; Resolution = 1 cm"1, top
spectrum = 2.2 ppm, bottom spectrum = 20.2 ppm, total pressure is
700 mm N2, .path length is 255 m
2300
-------
CO
I'ft^C'M^/tf VM^*1UW|
! I
,
700
800
900
1000 1100 1200
WAVENUMBER (cm-)
1300
1400
Figure
Infrared spectrum of methanol; Resolution = 1 cm"1, top spectrum =
3.5 ppm, bottom spectrum = 19.6 ppm, total pressure is ?00 mm N2,
path length is 255 m . . •
1500
-------
100
700
800
900
1000 . 1100 1200
WAVENUMBER (cm'1)
1300
1400
Figure U6.
Infrared spectrum of t-butyl-alcohol; Resolution = 1 cm"1, top
spectrum = 1.9 ppm, bottom spectrum = 8.2 ppm, total pressure is
700 mm Na, path length is 170 m
1500
-------
100 A
co
CO
"->
CO
z:
<
cr
0
lOOr
(t-Buty|-0)2
co!
CO
or
h-
•xP
"i ffwf^rr,
700
800
900
1000 MOO 1200
WAVENUMBER (cm")
1300
1400
1500
Figure ^7. Infrared spectrxira of di-t-butyl-peroxi4e; Resolution = 1 cm'1, top
spectrum = 1 ppm, tottom specjtrum = U,3 ppnij total pressure is
700 ram N2, path length is 170 m
-------
100
^wr4
-------
100
o
CO
CO
CO
(t-C4HJM);
cr
0
700
800
900
1000
1100
!200
1300
1400
WAVENUMBER (crrT1)
Figure ^9.
Infrared spectrum of 2, 2f - azoisobutane; Resolution = 1 cm'1, top
spectrum = 0.15 ppm, bottom spectrum = 15.^ ppm, total pressure is
700 mm N2, path length is 1?0 m .
1500
-------
lOOr
ol
in
to
or
(CH3N)2
TOO
800
900
1000
IIOO 1200 1300
WAVENUMBER(cm-)
1400
1500
1600
1700
Figure 50.
Infrared spectrum of azomethane; Resolution = 1 cm'1, top spectrum =
3 ppm> bottom spectrum = 15 ppm, total pressure is 700 mm N2, path
length is 170 m; impurity band due to methane appears around 1300 cm"1
-------
VO
H
TOO
eoo
900
1000
1100
1200
1300
1400
1500
1600
1700
1800
WAVENUMBER(cnrr')
1900
2000
Figure 51. Infrared spectrum of formic acid; Resolution = 1 cm"1, top. spectrum =
O.52 ppm, bottom spectrum = 5.5 ppm, total pressure is 700 mm N2,
path length is 1?0 m
-------
100 ^VvV~^
ill
o
CO
cn
s£i
o
CO
LO
CO
<
cr
700
600
900
"^wf-yM
1000
1100
_L
1200
1300
I4OO
WAVENUMBER (cm-)
1500
1600
I
, I I
m
i, •• •'
in
1700
1800
Figure 52. Infrared spectrum of formaldehyde; Resolution = 1 cm"1, top
spectrum = k.6 ppm, bottom spectrum = 9-2 ppnij total pressure is
700 mm N2, path length is 1?0 m
1900
-------
u>
TOO
1
800
1
900
1
1000
1
MOO
1
1200
1
1300
1
1400
1
1500
1
1600
M
1
1700
1
1800
1
1900
1
2000
WAVENUMBER(cm-')
Figure 53. Infrared spectrum of acetone; Resolution = 1 cm"1, top spectrum =
2.2 ppm, bottom spectrum = l6.k ppm, total pressure is 700 mm N2,
path length is 170 m
-------
CONCERNING PROBLEMS IDENTIFIED AS A RESULT OF THE PRESENT INVESTIGATION
SOLAR SPECTRA AND GROUND LEVEL AIR SPECTRA ANALYSIS
The comparison between these spectra and synthetic spectra calculated from the
AFCRL line parameter testing has revealed that some of the line parameters
need to be altered in order to improve the agreement between the spectra. We
do not plan to improve the parameters throughout the entire spectral region
investigated but it may be desirable to correct parameters associated with the
lines of permanent gases which occur in spectral regions occupied by strong
bands of possible atmospheric pollutants, in order to help in their identifi-
cation.
We have attempted to estimate the abundance of fluorocarbon-12 (FC-12) from
the appearance of its characteristic absorption features in solar spectra.
We are currently attempting to derive line parameters for this molecule from
laboratory spectra which will enable us to include the absorptance by this
molecule in the synthetic spectra. Estimates of the abundances or upper limits
to the abundances of these gases in the atmosphere can then be determined.
Such a program requires first the generation of suitable quantitative spectra
of known amounts of each gas and then the subsequent spectral analysis of the
absorption bands. Such procedures are time consuming and will not be attempted
unless they are of immediate concern to the program. It has already been found
that the identification of FC-12 absorption features in solar spectra has been
facilitated by the appropriate parameterization of its Q-branches and the sub-
sequent use of these' parameters to create synthetic solar spectra with varying
amounts of FC-12 absorption.
In view of the possible importance of such studies we intend to cooperate with
other workers in this area where possible. As part of this work analysis of
the rotational structure of HN02 bands near 7.7-13-3 Mm is being carried out.
This work is described below.
In addition to these studies we also plan to investigate the low sun solar
spectra which we have obtained and to identify as many H20 lines as possible
so that other remaining unidentified features can be located.
We also plan to use these synthetic spectra to monitor variations in the
abundances of the permanent atmospheric gases, in particular of C02, N20, CO,
and CH4. We plan to investigate several spectral regions to locate those
which show the greatest change in absorptance with the amount of gas present.
NITROUS ACID BAND ANALYSIS
It has been pointed out in a recent study on the chemical kinetics of HONO"41
that, contrary to previous belief, HONO formation in the free atmosphere can
be of significance in environments which are rich in NOX-H20. This is typical
of conditions of power plant stacks and automotive exhaust pipes. Upon sun-
light irradiation, HONO gives rise to HO radicals which play an important role
.in photochemical smog formation.
-------
In order to facilitate the identification and detection of this molecule (HONO)
in the atmosphere, moderately high resolution (0.125 cm"1) spectra were ob-
tained in the mid-IR region by Fourier transform spectroscopy using a path
length of k2. m and a nitrous acid concentration of less than 5 x 10~5 atm.
The v4 bands of both the trans- and cis- isomers (at 790 and 852 cm"1, respec-
tively) and the v3 band of the trans-isomer (at 1263 cm"1) are A-type transi-
tions and are mildly degraded to higher energies. A partial rotational analysis
of each of these bands has been carried out using Watson's Hamiltonian42 and
the lower state constants given by Cox et al."43 Tentative assignments of
transitions with quantum numbers up to K =~6" and J = 20 have been made. The
accuracy of the assignment is compromised by the fact that the present resolu-
tion does not allow us to resolve all the overlapping features to eliminate
ambiguities. Intensity calculation of these bands will be carried out and
comparisons between the observed spectra and the contours calculated will be
made. With the determined upper state rotational constants, transition fre-
quencies can be computed and they should facilitate detection with higher
resolution tools such as laser spectroscopy.
A much weaker band of HONO has also been observed near 1068 cm~1. This band
exhibits B-type structure and is rather strongly degraded to lower energies.
A preliminary analysis of the PQ and rQ heads suggests that it is an overtone
band due to 2v6 of the trans-isomer. Rotational analysis of this band is
intended to be carried out in the future. However, because of the very weak
band intensity, unless HONO is present in a very large quantity, this band
will be of little use. for identification and detection purposes.
95
-------
REFERENCES
1. .M. J. Molina and F. S. Rowland, Nature 2*t9, 8lO (197*0-
2. J. E. Lovelock, Nature 230, 379 (1971).
3. J. E. Lovelock, R. J. Maggs, and R. J. Wade, Nature 2*+l, 19** (1973).
H. N. E. Hester, E. R. Stephens, and 0. C. Taylor, J. Air Pollution Control
Association 2U, 591 (197*0.
5- N. E. Hester, E. R. Stephens, and 0. C. Taylor, Environ. Sci. Technol. 9_,
875 (1975).
6. P. L. Hanst, L. L. Spillers, D. M. Watts, J. W. Spence and M. F. Miller, J.
Air Poll. Control Assoc. 2j>, 1220 (1975).
7. D. G. Murcray, F. S. Bonomo, J. N. Brooks A. Goldman, F. H. Murcray, and
W. J. Williams, Geophys. Res. Lett. 2, 109 (1975).
8. R. A. McClatchey, W. S. Benedict, S. A. Clough, D. E. Burch, R. F. Calfee,
K. Fox, L. S. Rothman, and J. S. Garing, "AFCRL Atmospheric Absorption Line
Parameters Compilation," AFCRL-TR-73-0096, Air Force Cambridge Research
Laboratories Environmental Research Papers, No. **3** (1973).
9. E. K. Plyler and W. S. Benedict, J. Research NBS ^7_, 202 (1951).
10. J. C. Breeze, C. C. Ferriso, C. B. Ludwig, and W. Malkmus, J. Chem. Phys.
U2, 1*02 (1965).
11. T. C. Kyle, Atlas of Computer Infrared Atmospheric Absorption Spectra, NCAR-
TN/STR-112 (1975).
12. J. C. Fontanella, A. Girard, L. Gramont, and N. Louishard, Appl. Opt. lU
825 (1975).
13. M. T. Stier and W. A. Traub, Appl. Opt. 15_, 36U (1976)..
lit. R. A. McClatchey, R. W. Fenn, J. E. A. Selby, J. S. Garing, and F. E. Volz,
"Optical Properties of the Atmosphere," AFCRL-70-0527, (1970).
15. R. A. McClatchey and J. E. A. Selby, Atmospheric Attenuation of Laser Radia-
tion from 0.76 pm to 31.25 ym, AFCRL-TR-7U-0003,' (197*0-
16. W. L. Wolfe, Handbook of Military Infrared Technology, Government Print-
ing Office (1965), pp. 228.
17. J. H. Shaw, M. L. Oxholm, and H. H. Claassen, Astrophys. J. Il63 55*+ (1952),
18. C. B. Farmer and P. S. Key, "A Study of the Solar Spectrum from 7 to 1*50
microns," EMI Report No. BMP 2160, (1965).
96
-------
19. M. Migeotte, L. Neven, and J. Swensson, "The Solar Spectrum from 2.8 to
' 23-7 microns," Final Report under contract AF 6l(51^)-^32, AFCRL Bedford
Mass., (1965).
20. D. Murcray, private communication (1976).
21. W. S. Benedict and R. F. Calfee, "Line Parameters for the. 1.9 and 6.3 micron
Water Vapor Bands," ESSA Project Paper No. 2, Washington, B.C., (1967).
22. J. U. White, J. Opt. Soc. Amer., 32_, 285 (19^2).
23. D. J. McCaa, J. H. Shaw, J. Mol . Spec. 25_, 3fk (1968).
2k. A. G. Maki, J. Mol. Spec. j>J_, kl6 (1975).
25. W. E. Nexsen, Thesis, The Ohio State University, University Microfilms,
Publication No. 16090, 53 pp.
26. D. Horn, and G. C. Pimentel, Appl. Optics, 10, 1892, (1971).
27. D. E. Burch, D. A. Gryvnak, R. R. Patty, J. Opt. Soc. Am. 57, 885 (1967).
28. R. Blickensderfer, G. Eving, and R. Leonard, App. Opt. J_, 21lk (1968).
29. W. H. Chan, R. J. Nordstrom, J. G. Calvert and J. H. Shaw, Environ. Sci.
Technol, 10, (1976), to appear June, 1976
30. J. P. S. Haarsma, G. J. deJong, J. Agterdenbos, Spect . Acta. 29_B_, 1 (1974).
31. Opthos Instrument Co., Rockville, Maryland.
32. Spectroscopic Studies of Photochemical Smog Formation and Trace Pollutant
Detection, J. G. Calvert, W. H. Chan, R. J. Nordstrom, and J. H. Shaw,
Final Report EPA Grant No. R803075, August 1975.
33. P. L. Hanst, "Spectroscopic Methods for Air Pollution Measurements," in
Advances in Environmental Science and Technology, Vol. II, J. N. Pitts, Jr.
and R. L. Metcalf Eds. (Wiley, New York, 1971), P- 91-
3k. J. G. Calvert, Environ. Sci. and Technol., 10. 2k8 (1976).
35. J. R. Holmes, R. .J. O'Brien, J. H. Crabtree, T. A. Hecht, J. H. Seinfeld,
Environ. Sci. Techol., 7, 519 (1973).
36. D. H. Stedman and H. Niki, Envir. Sci. and Techol. 7, 735 (1973).
37. C. H. Wu and H. Niki, Environ. Sci. Techol., 9, k6 (1975).
38. K. L. Demerjian, J. A. Kerr, and J. G. Calvert, Adv. Environ. Sci. and
Technol., k, I (197*0 5 see pp.
39. D. Garvin, "Chemical Kinetics Data Surveys" IV (NBSIR 73-203), V (NBSIR
73-206), VII (NBSIR 7k-k30).
97
-------
ho. H.S. Johnston and Graham, Can. J., Chem., 52, ll*15 (197^).
1*1. W. H. Chan, R. J. Nordstrom, J. G. Calvert and J. H. Shaw, Chem. Phys.
Letts., 31, hkl (1976).
1*2. J. K. G. Watson, J. Chem. Phys., k_6_, 1935 (1967).
1*3- A. P. Cox, A. H. Brittain and D. J. Finnigan, Trans. Faraday Soc. , 67, .
2179 (1971).
1*1*. H. S. Johnston and L. Zafonte, quoted by A. C.Lloyd, Int. J. Chem. Kinet.
in press.
1*5. D. D. Davis, W. A. Payne, and L. J. Stief, Sci., 179, 280 (1973).
1*6. R. Simonaites and J. Heicklen, J. Phys. Chem., 77, 1096 (1973).
1*7. J. G. Calvert and E. W. R. Steacie, J. Chem. Phys., 1£, 176 (1951).
1*8. P. L. Hanst and J. G. Calvert, J. Phys. Chem., 63. 71 (1969).
1*9. B. A. DeGraff and J. G. Calvert, J. Am. Chem. Soc., 8£, 221*7 (1967).
50. R. D. McQuigg and J. G. Calvert, J. ibid.,
-------
APPENDIX I
SUMMARY OF OBJECTIVES FROM 1975 PROPOSAL
I. SUMMARY OF OBJECTIVES
The primary objective of the research outlined in this proposal is
to develop infrared fourier transform spectroscopic (FTS) techniques to
characterize certain key air pollutants, their precursors and reaction
products, and to establish quantitative kinetic and mechanistic details
of the interrelationships between these pollutants both in simulated .
and real atmospheres. We have chosen for study during the three year
period of this research effort the following important, yet unresolved,
problems related to the transformations of contaminants in both the
troposphere and stratosphere. These systems are particularly well suited
for study with the FTS system and associated equipment now in use at
the Ohio State University laboratories.
1. Study of the natural-removal mechanisms of the Freons and their
influence on the ozone concentration in the stratosphere:
a) The chlorine atom sensitized decomposition of ozone. The
chemical kinetics and mechanism of interactions in the
irradiated 03, NO, N02, C12 system.
b) The fate of the organic free radicals formed by -phcto-
dissociation of the Freons in the stratosphere. The
chemical kinetics and mechanisms of the reactions of the
CF2C1 and CC12F free radicals in air at stratospheric pressures,
c) The alteration of the stratospheric ozone concentration
through Freon addition. The laboratory simulation of the
photochemistry of the stratosphere perturbed by Freon addi-
tion.
2. Study of certain key reactions, seemingly important in photo-
• chemical smog systems:
a) The role of nitrous acid in the development of the impor-
tant free radical chain carrier, HO, and its relation to
smog formation in the urban atmosphere. A kinetic study of
the NO, N02, MONO, CO system. The quantum efficiency of
HO radical formation in the sunlight photolysis of MONO
in air. The determination of important rate constants for
the HO reactions with NO and N02. and H02 with NO will be
made.
99
-------
b) The role of the aldehydes in smog development in the urban
atmosphere. The nature and mechanism of the formation of
the products formed in the photooxidation of formaldehyde.
A kine,tic study of the CHaO, NO, N02, air, and the CH20,
NO, N02, Cl^, air systems. A search for the ellusive per-
o^yformyl nitrate.
c) A study of some potentially important S02 removal mechanisms
in the lower atmosphere. The rates of the primary reac-
tions of HO and HO, with S02 from the FTS study of the CH20,
S02, CO, air and HOMO, S02, CO, air systems. The nature
of the S02-oxidizing species formed in the 03-olefin dark
reaction. '
d) The determination of high resolution infrared spectra of
the common atmospheric contaminants and their products.
Cataloging of.the absorption coefficient-wavelength data
for use in this work and by other investigators in the field.
e) Direct observation of the infrared absorption spectra of
air samples in Columbus, Ohio. We will employ both in situ
samples, observed using solar infrared transmission spectra
in long atmospheric paths near sunset and sunrise, and con-
fined samples in the laboratory long-path system (up to 1.5 km);
spectra will be taken at regular intervals throughout the
period.
100
-------
APPENDIX II
PAPERS PUBLISHED OR ACCEPTED FOR PUBLICATION
1. "An IRFTS Spectroscopic Study of the Kinetics and the Mechanism of the
Reactions in the Gaseous System, HONO, NO, N02, H20." Walter H. Chan,
Robert J. Nordstrom, Jack G. Calvert, and John H. Shaw, Chemical Phys.
Letters, 37., kkl-kk6 (1976).
2. "A Kinetics Study of HONO Formation and Decay Reactions in Gaseous Mix-
tures of HONO, NOp, H20 and N2." Walter H. Chan, Robert J. Nordstrom,
Jack G. Calvert and John H. Shaw, to appear in Environmental Science and
Technology, July, 1976.
3. "A Spectroscopic Study of the N02-N20it System by the Infrared Absorption
Technique," Robert J. Nordstrom and Walter H. Chan, J. Phys. Chem., 80,
847-850 (1976). ~
101
-------
APPENDIX III
LIST OF PAPERS PRESENTED AT MEETINGS
1. "An IRFTS Spectroscopic Study of the Kinetics and the.Mechanism of the
Reactions in the Gaseous System, HONO, NO, N02, H20." Walter H. Chan,
Robert J. Nordstrom, Jack G. Calvert and John H. Shaw, presented at the
Centennial ACS meeting, New York City, April, 1976.
2. "A Partial Analysis of HONO Bands Between 750 and 1300 cm-1"- Walter
H. Chan, C. Weldon Mathews and Robert J. Nordstrom, presented at the
31st Symposium of Molecular Structure and Spectroscopy, Columbus, Ohio,
June, 1976.
3. "Infrared Spectroscopic Detection of Fluorocarbon-12 in the Atmosphere."
Robert J. Nordstrom, W. R. Skinner, J. H. Shaw, W. H. Chan, W. M. Uselman,
and J. G. Calvert, presented at the 31st Symposium of Molecular Struc-
ture and Spectroscopy, Columbus, Ohio, June, 1976.
h. "-Comparison of Measured and Computer Spectra of Air Samples." R. J. Nord-
strom, W. R. Skinner, J. H. Shaw, W. H. Chan, W. M. Uselman, and J. G.
Calvert, presented at the 31st Sympsoium of Molecular Structure and
Spectroscopy, Columbus, Ohio, June, 1976.
102
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/3-77-025
I. RECIPIENT'S ACCESSIOI*NO.
4. TITLE ANDSUBTITLE
APPLICATION OF FOURIER TRANSFORM SPECTROSCOPY TO AIR
POLLUTION PROBLEMS
Interim Report - 1976
5. REPORT DATE
August 1977
6. PERFORMING ORGANIZATION CODE
7. AUTHOR1S)
J. G. Calvert, W. H. Chan, E. Niple, R.J. Nordstrom,
J.H. Shaw, W.R. Skinner, and W.M. Uselman
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
The Ohio State University Research Foundation
1314 Kinnear Road
Columbus, Ohio 43212
10. PROGRAM ELEMENT NO.
1AA603
11. CONTRACT/GRANT NO.
R803868-1
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory <- RTP, NC
Office of Research § Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
Interim
14. SPONSORING AGENCY CODE
EPA/600/09
15. SUPPLEMENTARY NOTES
See companion reports EPA-600/3-77-026, Computer-Generated Long-Path Air Spectra,
and EPA-600/3-77-040, Comparison of Observed and Computed Air Spectra Between
16. ABSTRACT /UU dilQ
Spectra of air samples at ground level of approximately 10m, 100m, and 1km, and
solar spectra obtained for solar zenith distances between 40 and 87 have been ob-
tained. Examples of these spectra in the region from 1100 to 1200 cm are pre-
sented together with spectra calculated from the atmospheric line parameter listing
prepared by the Air Force Geophysical Laboratories. From the absorption features
of fluorocarbon-12 near 1160 cm in solar spectra, a mean tropospheric abundance .
of 0.34 ppb has been estimated.
A photochemical cell capable of approximating the solar noon irradiance at
ground level between 300 and 400nm and in which path lengths in excess of 200m
can be obtained is described. This cell has been used to study the photolysis of
HNO_ and a.rate constant of 0.070 min has been measured.
Spectra of more than 20 gases of importance to air pollution problems, obtained
under controlled conditions covering the region from 700 to 1500 cm , are presented.
Progress in the construction of a cell, coolable to -60 C, and capable of
being irradiated at wavelengths down to 170nm, is described.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Air Pollution
Infrared spectroscopy
Optical equipment
Solar spectrum
Reaction kinetics
Photochemical reactions
Fourier transform
spectrometer
13B
14B
20F
03B
07D
07E
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)
UNCLASSIFIED
21. NO. OF PAGES
113
20. SECURITY CLASS (This page)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
103
------- |