Ecological Research Series
SPECTROSCOPIC STUDIES OF
PHOTOCHEMICAL SMOG FORMATION AND
TRACE POLLUTANT DETECTIOI
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-600/3-76-084
July 1976
SPECTROSCOPIC STUDIES OF PHOTOCHEMICAL SMOG FORMATION
AND TRACE POLLUTANT DETECTION
by
Jack G. Calvert
Walter H. Chan
Robert J. Nordstrom
John H. Shaw
The Ohio State University
Research Foundation
Columbus, Ohio 43212
Grant No. R-803075
ROAP No. 21AKC
Task 31
Program Element No. 1AA008
Project Officer
Philip Hanst
Atmospheric Chemistry and Physics Division
Environmental Sciences Research Laboratory
Research Triangle Park, North Carolina 27711
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND DEVELOPMENT
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
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 pub-
lication. Approval does not signify that the contents necessarily re-
flect the views and policies of the U.S. Environmental Protection Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
11
-------
ABSTRACT
An infrared Fourier transform spectrometer has been used with a long
path length multiple traversal cell to study the infrared spectra of
atmospheric gases and several pollutant gases. Solar spectra have also
been obtained between 3 and 20 ym wavelength.
The kinetics of the formation and decay of nitrous acid have been
followed by monitoring bands of nitric oxide, nitrogen dioxide, and both
cis- and trans-nitrous acid. Rate constants and the equilibrium con-
stant for the reactions have been derived. A mechanism accounting for
the formation of nitrous acid in the atmosphere is proposed. These rate
data have been used to speculate on the potential importance of nitrous
acid formation in power plant and auto exhaust plumes.
A new technique involving the use of infrared spectroscopy and two cells
of unequal length to study a two-component system in equilibrium is
described.
A six-meter multiple traversal cell in which path lengths of up to 700 m
can be obtained has been constructed. The cell is surrounded with
fluorescent tubes with output in the region from 300 to 400 nm. The
cell irradiance closely simulates the solar irradiance at ground level
in spectral distribution and intensity.
ILL
-------
CONTENTS
ABSTRACT iii
FIGURES vii
TABLES x
I SUMMARY AND CONCLUSIONS 1
II BACKGROUND AND INTRODUCTION 3
III INSTRUMENTATION AND EXPERIMENTAL PROCEDURES 5
A. Instrumentation 5
1. Introduction 5
2. Twenty-One Meter Stainless Steel Absorption Cell 5
3. Optical System 7
4. Six Meter Glass Cell 10
5. Fourier Transform Spectrometer and Transfer Optics 23
6. Coelostat 28
7. Gas Handling System 28
B. Experimental Procedures 31
1. Spectroscopic Studies of Chemical Reactions 31
2. Calibration of NO and NO- Gases and HONO 33
3. Preparation of Nitrous Acid 35
4. Study of HNO Decay 38
5. Study of HNO^ Formation 38
IV RESULTS AND DISCUSSION 39
A. Collection of Gas Spectra 39
B. Study of the N02 - N^ System 41
1. Introduction 41
2. Experimental 42
3. Theory 44
4. Results 45
5. Conclusions 47
C. Study of the Kinetics and the Mechanism of the
Reactions in the Gaseous System, HONO, NO, N02, H20 51
1. Introduction 51
2. Experimental Methods and Techniques 55
3. Experimental Results and Discussion 60
REFERENCES 77
-------
CONTENTS - (Continued)
APPENDIX A. OPTICAL ALIGNMENT PROCEDURES
Twenty-One Meter Absorption Cell
Six Meter Cell
Transfer System from the Interferometer to the
Absorption Cell
Interferometer and Source Optics
APPENDIX B. FOURIER TRANSFORM SPECTROMETER
Description of Operation
Undersampling to Obtain High Resolution
APPENDIX C. UNITS FOR MEASURING GAS CONCENTRATIONS
79
79
79
79
80
81
81
81
87
VI
-------
FIGURES
Number Page
1 Optical arrangement of the 21 m cell, (a) optical
paths; (b) mirror shape 6
2 Image pattern on the field mirror Mp 8
3 Six-meter glass cell construction details. Stainless
steel spacers separate the tube and stainless steel end
plates, E, are shown 11
4 Support system for 6 m cell 12
5 Cradle supports for 6 m cell 13
6 Gas inlet system to 6 m cell 14
7 Support system for fluorescent lamps around 6 m cell 15
8 (a) Fluorescent lamps arrangement for 6 m cell;
(b) Wiring diagram for lamps. The ballasts A , A , A
and A. supply power for the lamps in Bank A of Fig. 8a.
Switch S1 controls the power to the lamps as shown in
Fig. 8c 17
9 Reflector supports for 6 m cell 18
10 (a) Spectral distribution of fluorescent lamp output;
(b) Spectral distribution of fluorescent lamp output
after passing through glass walls of cell; (c) Relative
spectral distribution of solar energy at ground level 19
11 Spectral transmittance of glass tubes of 6 m cell 20
12 Mirror system for 6 m cell. The numbers refer to
consecutive reflections by the mirrors. The number
of traversals is altered by turning D. 21
13 Optical diagram of interferometer 24
14 Optical transfer system from interferometer to
absorption cell 26 .
15 Optical transfer system from absorption cell to InSb
detector 27
16 Gas handling system 29
VII
-------
FIGURES - (Continued)
Number
17 Spectral ratioing, (a) Spectrum of detector profile
and evacuated cell; (b) Spectrum with sample of
N0_ and N_0. in cell; (c) Ratio of spectrum (b)/
spectrum faj. Spectral resolution 0.5 cm 32
18 Dependence of absorbance of NO Q-branch (1876
cm ) on the path length x NO concentration
product. Total pressure 700 mmHg. Spectral
resolution, 0.5 cm 34
19 Dependence of the absorbance of NO- Q-branch (823
cm ) on the path length x NO- concentration product.
Total pressure 700 mmHg. Spectral resolution, 0.5
cm' . Temperature 23 +_ 1°C 36
20 Apparatus for preparation of HNO- 37
21 (a) Spectrum of approximately 10 m of air at
ambient pressure and temperature, spectral res-
olution 0.125 cm" ; (b) Spectrum of approximately
100 m of air; (c) Spectrum of approximately 1000
m of air; (d) Solar spectrum, solar zenith angle
- 60° 40
22 The experimental apparatus. Either of the two
absorption cells can be aligned at the indicated
position 43
23 Spectrum of NO. + N-0 from 700 cm" to 900 cm" ,
See text for experimental details 46
24 Plots of absorbance of 791 cm Q-branch of NO- as
a function of total sample gas pressure in each
absorption cell 48
25 Plot of absorbance of 791 cm" Q-branch of NO- as
a function of partial pressure of NO times path
length. The squares represent data taken through
the 39.5 cm cell 50
/J\J*.^/\-±\Jtl *2 p V^ W I, J. WUII W J- J.VU-^U-1-11£ 111.1. ,/\ I. WtJ. V^ W.L. 1 1VX11W y
), N0_, and H-0; note in (a) decrease in the v.
mds of trans-MONO (791 cm" ) and cis_-HONO (853
26 Absorption spectrum of reacting mixture of MONO,
NO,
bane
cm ) and buildup of NO- transitions as time pro-
gresses in an originally HONO-rich mixture; note in
(b) the growth of MONO peaks as time progresses in
an originally HONO-poor mixture 58-59
Vlll
-------
FIGURES - (Continued)
Number
27 Absorbance versus time plot for the trans-HONO
C1264 cm) and cis-HONO C853 cm" ) in a HONO-
N0-N02-H20 mixture originally rich in HONO 61
28 Plot of the reciprocal of the absorbance of trans -
.HONO (1264 cm" ) and cis_-HONO (853 cm ) versus
time for a HONO-NO-NO_-H_0 mixture originally rich
in HONO 62
29 Plot of total HONO concentration versus time for
an extended run of 1648 min duration using an
initially HONO-rich mixture; initial concentra-
tions, ppm: HONO, 2.14; NO, 2.37; N02, 3.53;
H-0, 4200; temperature, 23 C. Curves shown
have been calculated relation 48 and the rate
constant values indicated on the figure. 64
30 Plot of cis-HONO concentration versus time using
an initially HONO-poor mixture; initial concen-
trations, ppm: HONO, 0.141; NO, 10.6; N02> 9.72;
HO, 2240; temperature, 23°C. Curves shown have
been calculated using relation 48 and the rate constant
values indicated on the figure. 68
31 Plot of [HONO] and temperature within the reaction
cell versus time for an HONO-NO-N02-H 0 mixture near
equilibrium; typical data used to determine the
equilibrium constant for the reaction, NO + N02 +
HO = 2HONO, and to provide a qualitative test of
tne temperature dependence to the rate constants. 71
32 Plot of J,nK(atm~ ) versus 1/T for literature data
and present results for the equilibrium, NO + NCL
+ HO = 2HONO; open circles, Ashmore and.Tyler ;
square average of data of Wayne and Yost ; tri-
angles, Waldorf and Babb ; closed circles, this
work. JANAF suggested values are shown along with
a squares fit of published data excluding present
work 73
IX
-------
TABLES
Number
1 Data for Determination of the Equilibrium Constant
for the NCL - NO. System
48
2 Summary of the Rate Constant Estimates for the
Reactions 66
3 Equilibrium Data Derived from the Temperature Drift
Experiments 72
4 Theoretical Development of HONO as a Function of
Time for Mixtures of NO, NO-, and FLO for Compositions
Typical of Stack Gas Emissions and Ambient Conditions 75
5 IFTS Parameters to Achieve 0.125 cnf Resolution 84
6 Characteristics of Optical Filters and Spectral
Limits to Achieve 0.125 cm Resolution 85
7 Relations Between Units for Amount of Absorbing
Gas in a Given Path Length 88
-------
SECTION I
SUMMARY AND CONCLUSIONS
This study has shown that an infrared Fourier transform spectrometer
(IFTS), when used to observe infrared spectra of gases in long path
length, multiple traversal cells, can detect trace amounts of some gases
at the ppb to ppm level. This instrumentation can be used to determine
accurate information of the kinetics of chemical reactions involving
concentrations of gas samples similar to those of their atmospheric
abundances.
In addition to adapting an IFTS for use with the multiple traversal cell
and also with a coelostat to observe solar spectra, a 6 m glass-walled
cell, suitable for photochemical studies, has been constructed.
With these systems, spectra of air samples ranging in length from 10 m
to 1 km have been obtained between UOO cm"1 and 3500 cm"1 together with
solar spectra at several zenith distances. All of these spectra were
obtained at 0.125 cm"1 resolution. Most of the lines observed in these
spectra can be identified with transitions of H20, 03, COs, N20 and CH4
in the AFCRL line listing.1 A program for collecting high resolution
spectra of the principal bands of pure samples of the permanent atmo-
spheric gases and of known and suspected pollutant gases has been
initiated.
Among the advantages of the IFTS system over the more conventional
grating spectrometers is its ability to acquire high resolution spectral
information over wide spectral intervals in short time intervals. Since
the information is generated in digital form it is suitable for storage
on magnetic tape and for further computer processing. Thus, the spectra
can be ratioed to remove spectral variations in detector sensitivity and
source radiance, and absorption bands of other molecular species can be
removed. The system has been shown to be stable, linear in output, and
to give highly reproducible spectra, essential requirements for quanti-
tative analysis of gas concentrations.
The IFTS system has been used to determine the equilibrium constant of
the reaction
s 2N02
A value of O.lU ± 0.02 atm was obtained.
The system has also been used to measure the extinction coefficients of
bands of NO, NO?, and cis- and trans-HONO and to make a detailed study
of the kinetics, equilibrium constant, and the temperature dependence
of the thermal reactions.
-------
H20 + NO + N02 -2HONO (l)
k2 , ,
2HONO -> H20 + NO + N02 (2)
Using infrared spectroscopy, it is possible to observe cis-HONO and
trans-HONO separately. We have found that the kinetics of the decay
(second order in nitrous acid) are the same for both isomers. This
result supports the fact that
cis-HONO * trans-HONO
proceeds much faster than either reactions in Eqs. (l) and (2)
The rate k2 in Eq. (2) has been determined to be k2 = l.U ± O.U x 10~3
ppm'-hnin"1 at 23°C. This rate is strongly temperature dependent. A
value of 2.2 ± 0.7 x 10~9 ppnr^min-1 was obtained for rate constant ki.
Based on estimates of the typical concentrations of NO, N02 and H20 in
car exhausts and the effluent from electrical generator stacks when
fossil fuels are consumed, the measured reaction rates, and typical
values for the rate of dilution and cooling of these emissions as they
enter the atmosphere, it is estimated that there is sufficient HONO
present in urban atmospheres during the early morning hours to be an
important source of the OH radical.
-------
SECTION II
BACKGROUND AND INTRODUCTION
The concentrations of gaseous pollutants in the free atmosphere are
typically in the range 10~6 to 10~12 by volume. Even at these low con-
centrations many pollutants affect the environment and the air quality.
They can also undergo reactions with other atmospheric molecular species
to form new products. Many of these reactions which occur at ground
level are initiated by solar radiation at wavelengths between 300 and
500 nm.
To understand the nature of these reactions and to determine which are
of importance to air pollution, it is necessary to study gaseous systems
which contain the reactants at concentrations which approximate those
found in ambient air and which can be irradiated at levels corresponding
to daytime solar radiant energy levels.
Although many techniques have been proposed for monitoring low levels of
gases in gas samples and some have been useful in particular situations,
infrared spectroscopic studies of long optical paths through the samples
remain among the more specific, sensitive, and accurate methods for the
detection of many molecular species.
The detectivity of infrared spectroscopic techniques increases with the
path length and spectral resolution available. It may be decreased if
there is interference caused by more than one constituent absorbing in
the spectral region of interest.
The instrumentation available for the present study include a stainless
steel multiple traversal absorption cell in which path lengths in excess
of 1 km can be obtained. A glass-walled cell has also been constructed
in which path lengths in excess of 500 m are possible. The cell can be
irradiated to give an irradiance which is similar in spectral distribu-
tion and magnitude to ground level solar radiant energy between 300 and
UOO nm.
The length of time required to collect an infrared spectrum typically
increases as the power of the incoming signal decreases, as the spectral
resolution is increased and as the width of the spectral interval sur-
veyed increases. The slow response typical of other types of infrared
systems has been largely overcome by IFTS, which can collect spectral
information over wide spectral regions in relatively short times. The
system in use has obtained a solar spectrum of the region from 2UOO to
3500 cm"1 with a spectral resolution of 0.125 cm'1 in less than
2 minutes. This rapid collective speed allows chemical reactions to be
monitored rapidly so that rate constants can be determined.
From a large number of possible reactions which could be studied we
have chosen to investigate the reaction between H20, NO, and NOp which
-------
leads to the formation of HONO. The presence of this transient species
in the atmosphere has been suggested although it has not so far been
observed. The present investigation has shown that this molecule will
not only be formed in power plant plumes and in car exhausts, but that
it may exist in the ambient atmosphere under suitable conditions.
-------
SECTION III
INSTRUMENTATION AND EXPERIMENTAL PROCEDURES
A. INSTRUMENTATION
1. Introduction
One of the objectives of this study was to investigate the application of
infrared Fourier transform spectrometers (IFTS) to air pollution problems.
We have been concerned with the ability of such a spectrometer, used in
conjunction with long path length absorption cells, to identify and
monitor trace gases in the atmosphere and to study reactions among gases
at concentrations in the ppm range by measuring their infrared bands.
The intensities of the infrared bands of most gases are such that, if a
gas is present in ppm quantities in air at atmospheric pressure, path
lengths in the range of 100 m to 1 km are required for accurate measure-
ment.
We have carried out investigations with a 21 m stainless steel cell
already constructed. During the past year a 6 m glass cell suitable for
photochemical studies has been built. The design of these cells and
their adaptation for use with the IFTS are described.
2. Twenty-one Meter Stainless Steel Absorption Cell
Most of the investigations made under this grant employed a 21 m stain-
less steel absorption cell containing a. White multiple traversal optical
arrangement.2 With this system useful absorption measurements can be
carried out with path lengths in excess of 1 km.
The cell body is a cylinder, 21 m long, with an internal diameter of
76 cm, made of 6 mm stainless steel, type 30k. It is designed for a
pressure range of 0 to k atm and for a temperature range of -60 to
250°C. The temperature control system has not yet been installed.
A diagram of the optical system is shown in Fig. 1. The mirrors, DL
and D2, were cut from ko cm diameter blanks so that they could be placed
side by side inside the cell. Their horizontal width is 33 cm. The
field mirror, M-p, has the shape shown in Fig. 1.
The mounts for the mirrors are made from stainless steel and mechanical
feedthroughs allow mirrors DL and D2 to be adjusted from outside the
cell. Thus, the number of traversals and the optical path in the cell
can be altered without disturbing the gas sample. The cell is evacu-
ated with a Kinney high vacuum booster pump, Series KMBD. Pump down
time from atmospheric pressure to a few microns Hg is approximately one
hour. Leakage rates of less than 1 mmHg/day are typically observed.
-------
33cm
D,
Figure 1. Optical arrangement of the 21 m cell, (a) optical
paths; (b) mirror shape
-------
3. Optical System
All of the three mirrors in Fig. 1 have the same radius of curvature
(20.5 m) and are placed a distance apart of 20.5 m. Radiant energy
enters the cell through a window near Mp and forms an image of the
source in the plane of Mp close to one of its vertical edges. The
energy is reflected by DI back to Mp and from there to T)2. The number
of traversals of the cell made by the radiant energy before it passes
the opposite side of Mp and leaves the cell is altered by adjusting the
mirrors DL and DS-
Two rows of images of the source are formed on Mp, as shown in Fig. 2,
when the numbers identify successive images. As each image is formed,
the radiant energy makes two additional traversals of the cell. The
maximum number of images in the top row of images
Nmax^l (3)
where
D = width of the upper half of Mp (23 cm), and
d = width of the entering beam in the plane of Mp.
The total number of images on Mp is then (2Nmax + l), and the corre-
sponding number of traversals of the cell is ^Nmax. Thus the maximum
path length
where R = 20.5 m is the distance between the mirrors.
If d = 10 mm, then the maximum path length Ifoax is approximately 1.9 km.
If the entering beam has a spectral radiance BV [v = wavenumber (cm"1)],
forms an image of area Ac of the source in the plane of Mp, and com-
pletely fills the mirror DI the spectral radiant power in the frequency
interval v to v + dv leaving the cell is
Pvdv = BvAcwcTcdv (5)
where
uc = solid angle subtended by Dj. at Mp, and
Tc = transmission coefficient of the cell.
-------
0
2468
10
Figure 2. Image pattern on the field mirror
-------
The solid angle
«c -
where
ADi = area of DI « nrj^ ,
rj) ~ 1? cm, and
R = distance between Mp and DX
= 20.5 m
Thus uc = 2.2 x 10~4 sr. The shape of Mp, shown in Fig. 2, is well
suited for use with a grating spectrometer. The narrow entrance and
exit slits of such instruments are filled by forming an image of a line
source such as a Nernst glower on them. An IFTS typically has a much
larger "throughput" since there are no slits and radiant beams with
circular symmetry are used. Thus, the area Ac of the beam, matching an
IFTS, in the plane of Mp is typically given by
Ac = «rj (7)
where 2rjj = d, the diameter of the beam.
To obtain long path lengths (Eq. (U)) 2r^, ~ 10 mm thus, Ac « 0.79 cm?
and the product for this cell,
Acuc * l.?U x KT4cm2sr. (8)
The transmission coefficient TC in Eq. (5) depends on the reflection
coefficient of the mirrors as well as on the absorption and scattering
effects of the windows, aerosols, and gases in the optical path. More
specifically, each traversal of the cell requires a reflection at a
mirror. If the reflection coefficients, r, are the same for each mirror
and the total number of traversals is n, the transmission coefficient
due to the mirror reflectivity alone
Tm = rn (9)
The mirrors are gold coated and if r = 0.98, a reasonable value for the
infrared, Tm =s 0.20 for n = 80 and an optical path of 1.5 km. The
observed value of Tce-j_-j_ for n = 80 is smaller than this and is attrib-
uted to aberrations caused by the mirrors . Excellent spectra of path
lengths of 1 km of air are obtained with the IFTS using this cell.
-------
k. Six-Meter Glass Cell
Although the stainless steel cell allows long absorbing paths to be
obtained, it is not readily adapted for photochemical studies since the
walls are opaque. During the past year, a 6 m multiple traversal cell
has been constructed with glass walls. As shown in Fig. 3? "the main
body of the cell consists of four 1.5 m lengths of Corning low expansion
borosilicate glass with an internal diameter of approximately 29 cm.
The ends of these tubes are ground flat and butted against stainless
steel spacers, 25 mm thick, with teflon gaskets. Vacuum tight joints
are made by bolting the glass tubes together by means of flanges on each
side of the joints. Stainless steel end plates E are attached to the
ends of the cell in a similar manner. Windows to allow infrared radi-
ation to enter and leave the cell have been placed in one end plate.
The other end plate has an outlet to the vacuum pump and mechanical feed
throughs to adjust the mirrors.
Tables, 2.8 m long, were constructed from 5 x 5 cm angle iron to support
each pair of tubes. For ease in moving, castors were attached to each
table. When the tube was in its final position the two tables were
rigidly bolted together and the castor wheels were lifted off the floor
by levelling screws attached to the base of each table.
The tubes were mounted on the tables by attaching 2.5 cm diameter steel
rods, which extend horizontally, to spacers Sj^ and 83, as shown in
Fig. U. These rods pass through holes in vertical blocks of aluminum
mounted on bases with levelling screws which rest on the horizontal
lengths of angle iron forming the table top. These blocks carry the
main weight of the glass tubes. To prevent rotation of tubes about the
rod axis, angle iron cradles were placed about 1 m from the spacers Sx
and S3. Details of these cradles are shown in Fig. 5- The cradles are
also supported on the same angle iron as the spacers and can be adjusted
by levelling screws.
The pairs of tubes on each table were adjusted to be in the same hori-
zontal plane before the entire assembly was bolted together at 83.
The system has been evacuated to a pressure of less than 1 Torr and a
leak rate of less than 1 Torr/wk has been measured. With the system
evacuated a strain gauge consisting of a pair of crossed polaroids
was used to investigate the strain in the glass tubes. Although some
strain was observed it is fairly evenly distributed around the tubes.
A gas inlet has been provided in each of the spacers and also in the
endplates, as shown in Fig. 6, to help ensure a uniform distribution of
gases throughout the cell.
The entire length of the cell can be irradiated by banks of fluorescent
lamps. The holders for these lamps are attached to lengths of 2 cm wide
aluminum channel mounted to metal frames surrounding the tube as shown
in Fig. 7. These frames also rest on the horizontal angle iron of the
mounting table.
10
-------
Figure 3. Six-meter glass cell construction details. Stainless steel spacers
separate the tube and stainless steel end plates, E, are shown.
-------
BLOCK
''LEVELLING
SCREWS
[— GAS INLET
BOLT HOLES
ANGLE IRON
5X5 cm
Figure U. Support system for 6 m cell
-------
y?
I~.L
OO
?T
Figure 5. Cradle supports for 6 m cell
-------
TO GAS
HANDLING
SYSTEM
TO PUMP
® BELLOWS VALVE
|X1 BUTTERFLY VALVE
Figure 6. Gas inlet system to 6 m cell
-------
J L
o
o
o.
o
I I
Figure 7. Support system for fluorescent lamps around 6 m cell
-------
Four banks of lamps have been installed along the length of the tubes
as shown in Fig. 8a. Three of the banks consist of eight 1.8 m long
tubes arranged symmetrically around the glass cell. The individual
banks are overlapped by about 15 cm to improve the uniformity of irradi-
ation. A bank of ten lamps 1.2 m long was placed at one end of the
cell. The two extra lamps were included since the spectral radiance of
these shorter tubes is about 20/0 lower than that of the 1.8 m tubes.
The electrical wiring diagram is shown in Figs. 8b and c. The irradi-
ation of the cell can be varied by switching 2, U, 6, or 8 lamps in each
bank. Based on the manufacturers literature it is estimated that, with
all the lamps on, the irradiation in the cell is approximately equal to
the solar irradiance of the lower atmosphere.
The fluorescent lamp assembly is surrounded by an aluminum foil reflec-
tor. The foil is supported by a flexible plastic backing as shown in
Fig. 9-
The lower reflector is held in position by cords attached to the hori-
zontal angle iron of the table top. The upper reflector is held in
position by a 2 x 2 cm aluminum angle iron frame, which also rests on
the horizontal angle iron of the table top. The foil is attached to the
plastic backing by rubber cement.
The fluorescent lamps installed in the system are GE FT2T12/BL/HO and
F^OBL black lights. The spectral radiance of these lamps, as described
by the manufacturer, shows that the greater part of the output lies in
the spectral region from 3000-UOOO A. The emission of a 1.2 m lamp was
measured with a Turner spectrophotometer and the spectrum obtained is
shown in Fig. lOa. A similar spectrum, obtained after the radiant
energy had passed through the wall of the glass cell is also shown
together with the estimated relative spectral distribution of solar
radiation3 for a solar altitude of U5°. The spectral transmittance of
the glass tube is shown in Fig. 11.
The uniformity of the irradiance inside the cell due to the fluorescent
lamps has been measured with a Hewlett Packard 8330A radiant flux meter.
The measurements indicate uniform irradiance along the length of the
glass cell and variations of less than a factor of three radially from
the cell center to the walls.
It is planned to observe changes in the concentrations of the gases in
the cell by monitoring their infrared absorption bands. The multiple
traversal optical system is a modified White arrangement and has been
described by Hanst.4 Figure 12 shows that the single mirror Mp in
Fig. 1 has been replaced by four rectangular mirrors, Ma, M^j Mc, and
M
-------
(o)
1.8m
Si
(b)
«i
S4
2 7 7 VAC
B
1.8m
1.8m
6m
Bi
lili
B.
O
0
mi
c,
C,
o
02
o
\.2rn
I/O
o o
Figure 8. (a) Fluorescent lamp arrangement for 6 m cell; (b) Wiring
diagram for lamps. The ballasts A!, A2, A3, and A4 supply
power for the lamps in Bank A of Fig. 8a. Switch Sx
trols the power to the lamps as shown in Fig. 8c
17
-------
CO
— Al ANGLE
UPPER REFLECTOR
LOWER REFLECTOR
Figure 9. Reflector supports for 6 m cell
-------
300
320
340 360
WAVELENGTH (nm)
380
400
Figure 10. (a) Spectral distribution of fluorescent lamp output;
(b) Spectral distribution of fluorescent lamp output
after passing through glass walls of cell;
(c) Relative spectral distribution of solar energy
at ground level
19
-------
LU
O
\-
^
O)
(T
H
300
340
380
420
WAVELENGTH (nm)
Figure 11. Spectral transmittance of glass tubes of 6 m cell
20
-------
Ma
Mb
Mc
Md
4 12
0
8 16
14
»
6
10
16.5cm
19.7cm
2
7 15
D4
D2
3 II
Figure 12. Mirror system for 6 m cell. The numbers refer to consecutive reflections
by the mirros. The number of traversals is altered by turning D4
-------
D2, Da, and D4. The radiant energy leaves the cell on the opposite side
of Mu,. In order to form each image on the mirrors M the radiant energy
has to make two traversals of the cell.
The maximum number of images which can be formed on M^
N
max d
where
Dj. = length of M)., (l6.5 cm), and
d = width of entering beam at M^.
Wnen used with an IFTS instrument this beam will typically have a
circular cross section of area A where
A = nr2 = «(£
Since there are four rows of images, one on each of the mirrors Ma, M^,
MC, and M(j, the total number of images is =* ^Nmax, and since each image
results from two traversals of the cell, the maximum path length
1^(meters) = 8 RNmax = UU Nmax
where the distance between the mirrors R = 5-55 m.
Assuming d a. 10 mm then, from Eq. (9)
Nmax = 16
and
Lmax » 700 m.
It is now possible to calculate the AgUg product for this cell as in
Eq. (7)
where
d = 10 mm
r - 5 cm
R = 5-5 m
Agtog = 2.0 x 10"4 cm?sr . (13)
22
-------
This AgCJg product is approximately the same as that for the stainless
steel cell.
5. Fourier Transform Spectrometer
and Transfer Optics
The IFTS available for this study is a Digilab Model FTS 20. It is
capable of achieving 0.125 cm"1 resolution. The principle of operation
is based on the Michelson interferometer and is well known. An essen-
tially parallel beam of radiant energy falls on the germanium beam
splitter B-S, shown in Fig. 13. Here part of the beam is reflected to
the fixed mirror M*. and some fraction of the beam reflected by M4 is
transmitted through the beam splitter. Part of the incident beam is
also transmitted by the beam splitter to M5 when it is reflected back
to the beam splitter and is then reflected. The emerging beam is a
combination of energy from M* and M5. This beam then passes through the
absorption cell to a detector.
In operation, the detector output is sampled as the mirror M5 moves from
a position of zero path difference between the beam splitter and NU and
M5 through a maximum distance of U cm. This output is computer pro-
cessed to provide a spectrum of the source modified by absorption in
the path.
It is well known that the maximum resolution in this spectrum is given
by
Av = i- (HO
when d = distance travelled by MS from the position of zero retardation.
If d - k cm, Av = 0.125 cm"1. In order to achieve this resolution,
which is independent of frequency, the divergence of the beam falling
on the beam splitter (and which consequently reaches the detector) must
be such that
R = f<^ (15)
AV — W-r
when R is the resolving power of the instrument at frequency v, and co-j-
is the solid angle containing the radiation falling on the beam
splitter.
When a germanium beam splitter is used the maximum practical spectral
frequency which can be observed is ~ UOOO cm"1. Thus, for Av equal to
0.125 cm"1
U < 1.96 x 10"4 sr . (16)
23
-------
OUTPUT
B-S
INPUT
ro
\ /
7 V
M 4
M5
i i
4 cm
Figure 13. Optical diagram of interferometer
-------
According to the manufacturer's specifications, the maximum allowable
diameter of the beam falling on the beam splitter to obtain this resolu-
tion is 1 inch. Thus the cross- sectional area of the beam
^ 5.0 cm?
and
Aj-cjj < 10" 3 sr cur
This value is that required to obtain 0.125 cm'1 resolution at
UOOO cm"1. Larger values are allowed if measurements are made at lower
wavenumbers or lower resolution. The minimum value of AjWy is, however,
approximately five times larger than the Aw products of the multiple
traversal cells previously described when they are aligned for the
maximum usable path lengths.
Since the absorption cells are the limiting factors in determining the
power transmitted by the system, the optical transfer systems from the
source through the interferometer to the cells must be designed to make
the actual Aw products for the beams passing through the cells as close
to those which have been calculated. This requires that the optical
systems form: (l) images of the source approximately 10 mm in diameter
at the field mirrors of the cells. If a Nernst glower, with a diameter
of about 1.6 mm and a length of 25 mm is the source, a magnification of
a factor of 6 is required to produce images of this width; (2) the
aperture stop of the interferometer system and the aperture stop of the
absorption cell should be conjugate points. The former is the 1 inch
diameter beam allowed at the beam splitter and the latter are the
mirror S.DI in Figs. 1 and 12.
Based on these considerations, the optical transfer system shown in
Fig. lU has been used with the 21 m cell. This system can also be used
with the 6 m cell with minor modifications since the Aw products are
similar.
Fjiergy from the source N is reflected by the spherical mirror ML
(fi = 15 cm) to product an enlarged image N]_ on the plane mirror MS-
Since most of the energy from the elongated image of the source cannot
be usefully collected if high resolution is required, all of M? except
a. horizontal strip approximately 3 nm wide was covered so that NX was
nearly square. NI is at the focal point of the off axis paraboloid M3
(f3 = 2? cm) so that collimated radiation falls on the interferometer.
After passing through the interferometer the energy is reflected by the
plane mirror Me to the spherical mirror M7 (f7 = 60 cm) which produces
a further enlarged image N? between M7 and M^. An image BL of the beam
splitter is also produced between M7 and N2. After striking the
spherical mirror Me (fs = 91 cm) the beam passes through the entrance
window of the absorption cell and forms an image N3, approximately
25
-------
rv>
M
Co :Ge
DETECTOR
MIRROR
M
N-
LONG PATH CELL
INTERFEROMETER
Figure lU. Optical transfer system from interferometer to
absorption cell
-------
10 mm wide, in the plane of the field mirror Mp (in Fig. 1) and an
image ?>•? of the beam splitter on the mirror DL in Fig. 1.
The beam energing from the exit window of the cell is then collected by
a detector. The two detectors available include a liquid nitrogen
cooled InSb detector for use in the region from l800 to UOOO cm"1 and a
liquid helium-cooled, copper-doped germanium detector whose useful
range is from 300 to 3500 cm"1. Both detectors have sensitive elements
approximately 1 mm x 1 mm and can accept radiant energy over a solid
angle of about one steradian.
The optical transfer system from the cell to the InSb detector is shown
in Fig. 15. This system was designed to produce a 1 mm diameter image
N" of the 10 mm diameter image N at M at the detector. The beam from
the absorption cell strikes the spherical mirror M7 (fg = 28 cm) and
forms an intermediate image of N at N' and of the mirror Ii2 at D1.
After reflection from the spherical mirror M]_o (flo = 15 cm) the beam
strikes a plane mirror (not shown in the diagram) which directs the beam
upwards to the detector at the base of the LN5 dewar. It was found
that, when the detector was placed at N", the vibrations of the cell
picked up from the surroundings were magnified by the long optical
lever arm (equal to the entire optical path in the absorption cell) and
caused the image N" to move with an amplitude of 1-2 mm around a fixed
position. These excursions seriously perturbed the signal from the
detector. A considerable improvement in the signal-to-noise ratio was
obtained by placing the detector at D". This image, about 2 mm
diameter, overfills the detector but has much less jitter than N" be-
cause the optical path DMg ~ 22 m is smaller than the optical path in
the cell. Because the detector is now overfilled and is at a conjugate
point to both D and the beam splitter B, it acts as the effective aper-
ture stop of the entire system.
N
Figure 15. Optical transfer system from absorption cell to
InSb detector
27
-------
The Cu-.Ge detector has an associated optical system for use with a
conventional Digilab IFTS spectrometer. It is designed to accept a
collimated beam and condense it to the detector. This instrument has
been placed to receive energy emerging from the cell after reflection
by a plane mirror as shown in Fig. lU. This mirror can be adjusted to
give maximum detector output. At this point an image of the mirror D
is formed on the detector element. As with the InSb detector, the
element is overfilled and acts as the aperture stop of the system.
When this detector was received the vacuum in the dewar was poor. The
system was modified so that the vacuum chamber is continuously pumped
with a cryopump line. After transferring liquid helium to the 0.75 £
capacity dewar the detector stays cold for 8-10 hours.
6. Coelostat
A coelostat consisting of two plane, 25 cm diameter mirrors is mounted
on the roof of the laboratory. One mirror rotates at a constant angular
speed about an axis lying in the plane of the mirror and parallel to the
earth's polar axis. This mirror reflects solar energy to the second
mirror which directs the beam into the laboratory. This instrument has
been used with the interferometer to obtain solar spectra.
The solid angle subtended by the sun at the earth, u>_ = 6.8 x 10~5 sr,
s
is smaller than the minimum solid angle required to fill the interfer-
ometer, dif m^n = 1.96 x 10"4 sr. Although a higher energy throughput
through the interferometer could be achieved by using a mirror tele-
scope, the spectral radiance of the sun is so great that the improvement
in signal-to-noise ratio gained by its use is not significant, and
excellent spectra are obtained by directing the beam from the coelostat
directly onto the interferometer. The angular diameter of the sun,
a.: = 9-^ x 10~3 rad, requires a 10 cm focal length mirror to condense
J ~
the energy onto a 1 x 1 mm detector.
7. Gas Handling System
The gas handling system shown in Fig. 16 was constructed to inject pre-
measured amounts of gas into the 21 m stainless steel absorption cell.
This gas handling system was made entirely of glass and stainless steel
and was fitted with teflon stopcocks. Samples of up to five gases can
be stored in this system at one time.
Before the system is used, the entire volume is evacuated by a Sargent-
Welsh duo-seal pump, Model 1^05 H. Ultimate vacuum in the gas handling
system is about 10 nmHg. After pumping, the stopcock to the pump is
closed and a sample gas is expanded into the glassware. One of the five
storage bulbs is opened and the pressure of the sample gas is measured
on Wallace-Tiernan gauges which are isolated from the sample gas by a
quartz spiral gauge. One Wallace-Tiernan gauge has a range of
0-20 Torr, the second gauge covers the range 0-50 Torr, and the third
28
-------
Hq MANOMETER
SPIRAL GAUGE
LASER TARGET
TO SAMPLE CELLS
PUMP
Figure 16. Gas handling system
29
-------
gauge covers the range 0-800 Torr. These gauges were calibrated against
a mercury manometer. It is estimated that pressures above 100 mm can
be measured to a precision of better than ± 0.1<$> and an absolute
accuracy of ± 0.2f0 standard deviation.
When the desired amount of gas has been injected into the gas handling
system, the stopcock to the storage bulb is closed, and the excess
sample gas is pumped out of the glassware by the pump. The gas is
exhausted out of the laboratory through a ventilator. Nitrogen is
flushed through the glassware to remove any residual sample gas, and
the glassware is again evacuated.
The entire filling process is repeated if it is necessary to measure
out another gas. The gas handling system has five gas storage bulbs.
Each storage bulb is approximately 1 {, in volume. The volumes of these
cells were obtained from the observed pressure changes when a gas sample
contained in a known calibrated volume was successively expanded into
the gas lines and then into each storage bulb. A series of independent
measurements of the individual volumes gave a reproducibility of
± 0.3% (standard deviation). When the gas handling system is connected
to the 21 m stainless steel cell, which has a volume of 9-53 x 103 £,
the pressure of sample gas in the cell can be computed from the pressure
of sample gas in the 1 I bulb. Assuming the ideal gas law
= PpV2
9-53 x 103
Pressure and volume in cell and bulb are designated respectively by
subscripts 1 and 2. Thus, there is a reduction in pressure by a factor
~ 104 when the gas is expanded into the cell. If a storage bulb is
filled to one atmosphere with a sample gas, the concentration of gas in
the cell will be ~ 100 ppm when the gas is expanded into the cell.
Because of the large difference in volumes between the cell and the
bulb it can be assumed that a complete transfer of gas in its sample
bulb to the cell occurs provided the cell is initially evacuated so
that P2 « PX. To ensure complete injection, the sample system is first
filled with nitrogen to an atmosphere and then opened to the cell and
flushed with nitrogen. The procedure is repeated four to six times to
transfer the gas sample into the cell. The concentrations of gases
introduced into the absorption are obtained from measurements of the
volumes of the storage bulbs, the volume of the cell and sample pres-
sure measurement. It is estimated that the maximum error of the amount
of gas transferred to the cell is on the order of ± 2%. The actual
amount of gas in the absorbing path depends on the degree of mixing in
the cell and adsorption/desorption effects of the walls.
30
-------
B. EXPERIMENTAL PROCEDURES
1. Spectroscopic Studies of Chemical Reactions
In order to follow chemical kinetics of reactions, an accurate knowledge
of the concentrations of reactants and products as functions of time are
required. Not infrequently, direct measures of these quantities are
impossible and they .can only be inferred from other related properties.
In this work, absorbance measurements of selected transitions of various
molecules are employed to relate to their corresponding concentrations
according to the Beer-Lambert law:
a = ecg (17)
where
a = absorbance
e = absorption coefficient
c = concentration (see Appendix C) and
0 = path length.
In the regime where the above relationship holds, monitoring the absor-
bance of an infrared band of the molecule of which e and i are known
gives the corresponding concentration of that compound. Although
integrated band absorbance data were not measured in this study, never-
theless, the monitoring of peak absorbances of selected transitions at
constant spectral resolution can give an unambiguous measure of the
concentrations.
The spectrometer system employed is a single beam instrument and, in
order to obtain the absorbance of a sample of a given compound, a back-
ground spectrum of the detector profile and the evacuated cell must
first be recorded to be ratioed against the sample spectrum taken at a
later time. An illustration of the ratioing process is shown in
Fig. I?. The top spectrum is a single beam spectrum of the detector
profile and an evacuated cell; the middle one is a single beam spectrum
obtained of a sample mixture of NO? and N-sQ* in the same cell. The
profile of the background makes it difficult to determine accurately the
concentration of the gases. In addition, absorbance caused by atmo-
spheric C02 and H20 also appears in the spectrum which overlaps the
N02-N204 absorption bands. The bottom spectrum corresponds to a ratioed
spectrum of the sample spectrum against that of the background. The
unwanted C02 and H20 absorptions are almost completely ratioed out and
a well-defined baseline is obtained which makes the determination of
absorbance unambiguous.
Since the optical system is not purged, atmospheric C02 and H20 absor-
bance may overlap infrared bands of other compounds of interest.
31
-------
<
z
CD
(b)
UJ
o
§
r :
650 850 (c) 1050 . 1250
WAVENUMBERS(CM'')
Figure 17. Spectral ratioing, (c) Spectrum of detector profile and evacuated
cell; (b) Spectrum with sample of N02 and N204 in cell; (c) Ratio
of spectrum (b)/spectrum (a). Spectral resolution 0.5 cm"1
32
-------
Although there are many bands of NO, N02, and HN02 in the mid-infrared
region, they are not all used for concentration monitoring purposes
because they are either interfered by C02 and H20 or by mutual inter-
ference. The bands that are monitored to obtain concentration informa-
tion are for N02, the v2 band, Q-branch transition at 823 cm"1; for NO,
the fundamental band Q-branch transition at l8?6 cm"1; for cis-HONO; the
v^ band Q-branch transition at 853 cm"1, and for trans-HONO, the v3 band
Q-branch transition at 126U cm"1. All absorbance measurements were made
at a total pressure of 700 mm.
2. Calibration of NO and N02 Gases and HO NO
To apply the concept discussed in Section B.I a number of correlation
curves of absorbance and concentration was generated. Both NO and N02
gases are stable at room temperature and their concentration can be
easily measured. However, the HN02 molecules are rather unstable and
correlation curves cannot be generated directly. The method used to
determine the HN02 concentration will be described later.
a. Calibration of NO and N05
The calibration of NO and N02 gases has been carried out in three cells
of 7.5 cm, 39.5 cm and 21 m path lengths, respectively. Known pressures
(mmHg) of NO (99$j Matheson Co.) were measured and injected into these
cells . High, purity nitrogen gas was used to bring the total pressure
to 700 mmHg. Then the spectrum of the sample gas was taken and ratioed
against the proper background spectrum of the cell filled with high-
purity nitrogen to a total pressure of 700 mm. A series of spectra
taken at room temperature (23 ± 1°C) with varied path length concentra-
tions products allowed the absorbance (a) vs. path concentration (P-.0)
plot shown in Fig. 18 to be obtained. The band used is the fundamental
rotation vibration band of NO which has an isolated and distinct
Q-branch absorption at 1876 cm"1.
Because N02 is always in dynamic equilibrium with N204, the total
pressure of N02 measured corresponds to the sum of the partial pressures
of N02 and N204. N02 (+N204) with a purity of 99-5fo was obtained from
Matheson Co. and before using, it was trapped with a LN2 cold trap and
pumped to remove volatile residual impurities such as NO, N203, N20,
etc. The equilibrium constant for the reaction:
N204 * 2N02
is defined by
K-
PT -
where
PN02 = Partial pressure of N02
PT =
33
-------
o
00
cr
o
CO
CD
< .5-
NO
=0. 513*0.006 m'torr
I
2.0
1.0 2.0 3.0
P*L (TORR-METER)
4.0
Figure 18. Dependence of absorbance of NO Q-branch (1876 cm'1)
on the path length x NO concentration product.
Total pressure 700 mmHg. Spectral resolution,
0.5 cm
-i
-------
Hence
= 0
and by solving this quadratic equation, partial pressures of N02 and
can be determined according to
PN02 = (-1 + ^1 + UPT/K)K/2
(2Pr + K) - N/K? + 1|VT>- ' ' '
PN204 = 2
By using the reported equilibrium constant at room temperature for the
N02 - N20* system,5 and the above equations, partial pressures of N02
and N2CU were determined. Once the N02 partial pressure is known and
the absorbance determined, the calibration curve for NOp is generated
essentially the same way as that for NO gas. A correlation curve for
the N02 Q-branch transition at 823 cm'1 at room temperature (23 ± 1°C)
is shown in Fig. 19- From these curves, the absorption coefficients for
NO at 1876 cm"1 and N02 at 823 cm'1 are found to be 0.513 ± 0.006 mm"^"1
and O.OSO? ± 0.0012 mm'-'-m"1, respectively, from least square fits of the
data. These results are for a spectral estimation of 0.5 cm"1.
b. Calibration of HN02
Because HN02 is unstable at room temperature, it was not possible to
generate known amounts of this gas. However, absorption coefficients
of both cis- and trans-HN02 can be obtained by either comparing the
decrease of HN02 absorbance to the formation of total NOX (NO and N02)
in the study of HN02 decay and formation reactions, or comparing the
equilibrium absorbances and concentrations of this system with those
expected from the equilibrium constant quoted in the literature. Based
on the results of the HONO study reported in Section IV.C, absorption
extinction coefficients for the cis-HONO Q-branch at 853 cm'1 and the
trans-HONO Q-branch at 126U cm"1 are found to be 9.36 ± 0.9^ mm~1m~1 and
5.20 ± 0.52 mm~1m"1, respectively for a spectral resolution of 0.5 cm"1.
3. Preparation of Nitrous Acid
Nitrous acid (HONO) was produced by reacting sodium nitrite with dilute
sulfuric acid at room temperature. Three methods were attempted:
(i) A dilute H2S04 solution was added to a dilute solution of NaWOa:
(ii) A dilute solution of HpSO* was added to NaNOp solids, and (iii) A
dilute NaNOp solution was added to a dilute H3S04 solution. It was
found that the first two methods, besides generating HNOp, also generate
rather large amounts of NO and N02. The second method gave a slightly
better yield of HONO than the first. The third method gave the highest
HNOp to NOX(NO + N02) ratio and was used in our study. Cooling of the
apparatus for the preparation of HN02 gave no significant improvement in
yield of HNOp. The apparatus used is shown in Fig. 20.
35
-------
UJ 0.2-
O
2!
<
CD
o:
o
CO
CD
< 0.1-
6 NO = 0.0807* 0.0012 m'torr
i'o
z.o 3.0
PxL (TORR-METER)
4.0
Figure 19.
Dependence of the absorbance of N02 Q-branch (823 cm-1) on the
path length x N02 concentration product. Total pressure 700 mmHg.
Spectral resolution, 0.5 cm'
-i
Temperature 23 ± 1°C
-------
CLAMP
DILUTE
NaNO,
CONCENTRATED
NaOH
DILUTE H2S04
Figure 20. Apparatus for preparation of HN02
37
-------
k. Study of HN02 Decay
Before injection of HN02 into the reaction cell, a known amount of
water was evaporated into the evacuated cell and the cell was then
filled to a pressure of UOO-500 mm with high-purity nitrogen. The
3-neck flask in Fig. 20 was first flushed and filled with nitrogen. In
the process of producing HN02, the passage to the concentrated NaOH
trap is blocked off and that to the absorption cell is open. Dilute
NaNC>2 solution (~ 0.3 M) was added dropwise to the dilute HpSO* solution
(~ 2.5 M) in the 3-neck flash. The system was flushed with nitrogen to
aid the HN02 injection into the absorption cell. At the end of the
preparation (or 5 minutes after start of reaction), the passage to the
cell was closed and the remainder of HN02-NO-N02 was trapped in the
concentrated NaOH solution. Through another valve with a much bigger
opening, high-purity nitrogen gas was injected (in a time of ~ 5-10 min)
into the cell to a final pressure of 700 mm. Spectra, each of which
took h minutes to collect and compute were taken at regular time inter-
vals varying from U minutes to 10 minutes and they were ratioed against
the background of the evacuated cell recorded before the run of the
experiment to obtain absorbances of NO, N02 and HN02. H20 absorbance
was not monitored because of interference from atmospheric H20 in the
path external to the cell. All the runs were carried out with a path
length in the cell of 250 m. Temperature inside the cell was recorded
during the run period by a chromel-alumel thermocouple installed inside
the cell.
5. Study of HN02 Formation
A known amount of H20 was measured and evaporated into the evacuated
cell. This was followed by injection of NO gas before the pressure
inside the cell was brought up to ~ UOO-500 mmHg. Then N02 of measured
quantity was injected and the cell pressure was increased to 700 mmHg
filling it with high-purity nitrogen gas. Spectra were then recorded
at regular intervals (~ 10 min) and they were ratioed against a pre-
viously recorded background to obtain absorbance data on NO, N02 and
HNOp. All experiments were carried out with a path length of
meters. Temperature inside the cell was recorded at intervals.
38
-------
SECTION IV
RESULTS AND DISCUSSION
A. COLLECTION OF GAS SPECTRA
The investigations which have been undertaken during the past year show
the complexity of the spectra of apparently simple gas systems. For
example, spectra of N02, even when well diluted, show evidence of NpO*
bands, spectra of mixtures of NO and N02 show evidence of additional
bands which may be due to N203; spectra of long air paths show a complex
structure due to the permanent atmospheric gases H20, C02, CH-i CO, and
N20. In order to make qualitative analyses of any of these spectra it
is desirable to have spectra of the individual gases which may be present
in a given sample.
The IFTS system available has the ability to collect spectra with
0.125 cm"1 resolution over the spectral region from UOO to UOOO cm"1.
While this resolution is not as high as can be achieved by some instru-
ments, it is nevertheless capable of resolving most of the fine struc-
tures of bands of atmospheric gases at ambient temperature and pressure
since the half width at half height of the absorption coefficients of
lines of most gases is about 0.12 to 0.16 cm"1 under these conditions.
One of the advantages of the IFTS over more conventional spectrometers
is its ability to store the observed spectra in digital form on magnetic
tape; these spectra can be recalled and further processed by the com-
puter to produce spectra with a degraded resolution. At the present
time spectra of NO, N02, N204, H20, HNOp, and CO have been obtained for
a limited range of concentrations and spectral regions.
In addition to obtaining spectra of individual gases we have also
initiated a program to collect typical air spectra for a number of
different path lengths. These include paths of approximately 10, 100,
and 1000 m. The 10 m path includes the air in the optical system from
the source to the detector excluding the absorption cell. Four tra-
versals of the absorption cell increases the path to approximately 90 ni
and U8 traversals of the cell gives a total path length of about 990 m.
Typical spectra between 700 and 725 cm"1 are shown in Fig. 21 for these
paths. Most of the absorption in this spectra region is due to C03.
Also shown in this figure is a solar spectrum obtained with a solar
altitude of about U0°. The effective path traversed by the solar radi-
ation is equivalent to about 10 km at surface pressure. By making
observations with a solar altitude of less than 5° path lengths equiva-
lent to 50 km at surface pressure have been obtained in some spectral
regions. At the present time, spectra of 10, 100, and 100 m have been
generated for the region from UOO to 2500 cm"1 and solar spectra from
UOO to UOOO cm"1 for high sun. Solar spectra from 2500 to 3500 cm"1
have been obtained for solar altitudes of 2-3°.
39
-------
CO
L« 10 METERS
-o-
100 METERS
t
t
L « I KMETER
L« IOKMETERS
(SOLAR SPECTRUM )
[700 , .
725 CM
Figure 21. (a) Spectrum of approximately 10 m of air at
ambient pressure and temperature, spectral
resolution 0.125 cm"1; (b) Spectrum of approxi-
mately 100 m of air; (c) Spectrum of approxi-
mately 1000 m of air; (d) Solar spectrum, solar
zenith angle ~ 60°
UO
-------
It is planned to continue adding to this collection and to prepare
reports describing the results as appropriate.
B. STUDY OF THE N02 - N204 SYSTEM
1. Introduction
During the past few years, nitrogen dioxide has become the subject of
an increasing number of studies. This interest stems largely from
attempts to monitor N02 effluents in the troposphere and from attempts
to follow photochemical processes involving NOp in the stratosphere.6
Laboratory studies of pure NOp, are often compromised by the presence
of dinitrogen tetroxide in the sample. This Np04 interference can be
eliminated either by heating the sample chamber7>8 to dissociate the
N204, or by using extremely low partial pressures of N02 to reduce the
quantity of N204 to a negligible amount. Both of these techniques
suffer from possible shortcomings. They force the experimenter to
investigate properties of NOp in a temperature or pressure region which
might not be feasible or relevant. For studies of NOp at temperatures
and pressures at which N204 is present, it is necessary to know the
equilibrium constant for the association-dissociation reactions
2 N02 ^ N204.
The work reported here is the result of preliminary steps of an investi-
gation of gas kinetics involving N02 and other gases at room tempera-
ture, using Fourier transform spectroscopy. The concentrations of the
components in a reaction can be monitored by observing their various
infrared absorption bands if the spectral information as a function of
concentration is known for each reactant and product. For our work,
the spectral information which was used to monitor gas concentrations
was the peak absorbance of an infrared band of the gas. Measurement of
peak absorbance provides an adequate technique for determining concen-
tration if all spectra are recorded at the same resolution.
To obtain a meaningful calibration curve of absorbance versus a wide
range of concentration for NOp, it was necessary to record spectra of
known concentrations of NOp under conditions similar to those of an
actual kinetic experimental run. Elimination of N204 from the sample
gas was not feasible. For these conditions of temperature and pressure
(296 K and 2 to 600 Torr) it became apparent that N204 would cause
appreciable interference to the calibration of N02. Thus, it was neces-
sary to know the equilibrium constant for the N02 - N204 system.
This work reports our results on the determination of this equilibrium
constant. Our value for the equilibrium constant is in good agreement
with values previously reported, and the emphasis here is on the tech-
nique used to investigate the equilibrium. A method has been developed
using infrared spectroscopy and two absorption cells of different
lengths to study this system at a single temperature. This method is
-------
significantly different from the method of Vosper8 who reported the use
of two absorption cells to study the equilibrium constant. This differs
from other methods10 which rely on data taken at higher temperatures
for the evaluation of the equilibrium constant at room temperature.
2. Experimental
The apparatus used to measure the equilibrium constant of the N02 + ^0*
system is shown in Fig. 22. Radiant energy from the Nernst glower is
collimated and passes through the interferometer. The beam exits the
interferometer and passes through one of the two absorption cells. The
radiation is then focused on the Cu:Ge detector.
The output from the detector is called the interferogram and is the
auto-correlation function of the electric field. This interferogram
signal is digitized by the analog-to-digital converter and is stored by
the computer. The computer used is a Nova 1200 mini-computer with kk
(^096) words of core and 128k words additional data storage on a fixed
head disc. The entire apparatus, including the interferometer, was
built by Digilab Inc.
Once the interferogram has been stored in the computer, it is trans-
formed to produce the spectrum. The usable spectral region is defined
by the response of the Cu:Ge detector which is from about UOO cm"1 to
3500 cm'1.
Both absorption cells were constructed of stainless steel and were
fitted with sodium chloride windows. The lengths of the cells were
LI = 39-5 cm and L2 = 7-5 cm, which gives a ratio LX/L2 = 5.2?.
The N02 (+ N20-t) (nominal purity, 99-5%) was obtained from Matheson Gas
Company. The gas was expanded into a gas handling system which was con-
nected to the absorption cells and pressure gauges. The gas handling
system was built entirely of glass and stainless steel and is fitted
with teflon stopcocks. A line sketch of this apparatus is shown in
Fig. 20. The gas was first collected in a liquid nitrogen cooled trap
and the volatile gases pumped off. When the cold trap was warmed to
room temperature, the sample gas was expanded into an evacuated storage
bulb. From this storage bulb, the gas was injected into the two sample
cells simultaneously. The pressure of N02 + N204 was measured on a
calibrated Wallace-Tiernan gauge and a mercury manometer which were
isolated from, the reactive gas by a glass spiral gauge. Typical sample
gas pressures ranged from 2 to 600 Torr.
After the absorption cells were filled to the desired pressure of
sample gas, the valves on the cells were closed. The N02 + N204 in the
glass tubing of the gas handling system was pumped out, and the system
was flushed with high purity nitrogen gas. The valves on the absorption
cells were again opened and nitrogen was injected into the cells until
the total pressure in both cells was 7^0 Torr. The valves on the cells
were finally closed and the two absorption cells were removed from the
-------
Cir-Ge
DETECTOR
SAMPLE CELL
INTERFEROMETER
PLOTTER
CONTROLLER
COMPUTER
Figure 22. The experimental apparatus. Either of the two absorption
cells can be aligned at the indicated position
-------
gas handling system, and each cell in turn was positioned in the
optical path.
Single beam spectra were recorded through each cell. All spectra taken
through the 39-5 cm cell were ratioed against a background spectrum
through the same cell. Similarly, all spectra taken through the 7-5 cm
cell were ratioed against a background spectrum recorded through that
cell. These two background spectra were prepared by filling each
absorption cell to 7^0 Torr with nitrogen and recording a spectrum
through each cell. These background spectra were stored in the computer
and could be recalled when needed. This ratioing procedure gave trans-
mission spectra of N02 + N2C>4 from which absorbance data could be
obtained.
3. Theory
The equilibrium constant for the reactions Na04 - 2N02
(19)
where P-^Q and Pj^ Q are the partial pressures of the two gases.
Expressing the equilibrium constant in terms of the total sample gas
pressure
P = PN02 + PN204 '
the equilibrium constant can be written
K = , (21)
P ~ PN02
The only directly measurable quantity in this expression is the total
gas pressure P. The partial pressure of N02 can be monitored indirectly
by infrared absorption measurements .
When infrared radiation passes through a cell of length L which is
filled with N02 to a partial pressure P^ , the absorbance can be
written
(22)
This equation is referred to as the Beer-Lambert law of absorption of
radiation. Here, &^Q (v) is the spectral absorbance due to the presence
of N02, %o2(v)Ao(v) is tne spectral transmittance at frequency v, and
(v) is the absorption coefficient of N02.
-------
The use of the Beer-Lambert law is valid when the absorbance, which is
the negative of the natural logarithm of the spectral transmittance is
proportional to the product of pressure times length. When the product
of pressure times length becomes too large, the resulting absorbance is
no longer linear with this product. For cells of fixed length, there
is an upper limit to the pressure which can be used if the Beer-Lambert
law is to apply.
Using two cells of different lengths, the N02 absorbances at a given
frequency VQ are equal in the two cells when
(23)
where the subscript 1 refers to the first cell and subscript 2 refers
to the second cell. From Eq. (23)
When this condition applies, the absorbances in the two cells are equal-.
If the slight correction due to pressure is ignored,11 the equilibrium
constant for the two cells can be written,
where PL and P2 are the total pressures of sample gas in cell 1 and
cell 2, respectively. Combining Eqs. (2U) and (25),
where a = L]_/L2. By substituting this equation into the expression
for K in Eq. (25),
_ (Pi(q)2 - P2)g
_
=
-------
FREQUENCY
)
Figure 23.
Spectrum of N02 + N204 from 700 cm
See text for experimental details
-1
to 900 cm
-1
-------
a spectrum of this region recorded through the 7-5 cm cell at a nominal
resolution of 0.5 cm'1. The pressure of NO2 + N20.i was 12 Torr and the
cell was filled with nitrogen to 7^0 Torr. The regularly spaced absorp-
tion spikes are due to N02 Q-branch transitions, while the broad central
feature is mostly due to N204 and partially due to N02.
All spectra were recorded at 0.5 cm"1 resolution. The absorbances at
several N02 Q-branch center frequencies were chosen as monitors of the
N02 concentration. Although peak absorbance in a spectrum is dependent
on the resolution used to record the spectrum, nevertheless, the peak
absorbance provides an unambiguous measure of concentration if all
spectra are recorded at the same resolution.
From the recorded spectra, the absorbances of the NOp Q-branches near
791 cm"1, 807 cm"1, and 823 cm"1 were measured. Plots of the absor-
bances as a function of sample gas pressure in both cells were made.
Figure 2U shows absorbance plots for the 791 cm"1 Q-branch obtained by
using the 7-5 cm cell and the 39-5 cm cell. Using these plots and
similar plots for the other Q-branch absorbances, a. pair of pressures,
P! and P2, can be determined for any chosen absorbance on a particular
Q-branch. From these pressures, values of the equilibrium constant
were calculated using Eq. (27). A region of linear dependence of
absorbance on pressure is not apparent in these plots because the inde-
pendent variable is the total sample gas pressure, not the N02 partial
pressure. The absorbance values used to calculate the equilibrium con-
stant were chosen to be sufficiently small to assure that we were well
within the pressure limits defined by the validity of the Beer-Lambert
law for both cells.
Table 1 shows the results of these calculations for the three N02
Q-branches which were studied. The average of the calculated equilib-
rium constants is O.lUO atm at 296 K ± 1 K and is in good agreement
with values reported previously.
5. Conclusions
The results in Table 1 show that the two cell technique provides a
reasonable method for the determination of the equilibrium constant for
the N02 + N204 system. Using the average value of the equilibrium con-
stant determined by this technique, the partial pressure of N02 can be
evaluated for all data points on Fig. 23. Plotting the absorbance
versus PNQ2'L f°r both cells shows a linear dependence of absorbance
on the product P^OP'^ as shown in Fig. 25. This linear dependence
extends to about 0.3 on the absorbance scale for all Q-branch absorp-
tions which indicates that the data were recorded within the limits of
the validity of the Beer-Lambert law.
The major source of experimental error comes from inaccuracies in
measuring the pressures P! and Pp. Differentiating Eq. (27) gives
-------
UJ
o
CD
(T
O
c/>
CD
0.6-
0.0
0
L = 7.5cm
25 50 75
TOTAL PRESSURE
(TORR)
100
125
Figure 2k. Plots of absorbance of 791 on-1 Q-branch of N02 as a
function of total sample gas pressure in each absorp-
tion cell
-------
Table 1. Data for Determination of the Equilibrium
Constant for the N02 - NgO* System
Q-Branch
(cm'1)
791
791
791-
807
807
807
823
823
823-
Absorbance
0.05
0.10
0.15
0.05
0-1Q
OJ.5.
0.05
0.10
0.15
Pi
(mmHg)
1.00
2.10
3-15
1.22
2...50.
3-727
l.UU
2.70
3.9^
P2
(mmHg)
5-^7
12.00
18.65
6.75
i^vUo
22. kO
8.00
15.70
23.90
K
(mmHg)
111.5
102.2
102.7
L02..0
109-9
iok.a .
110.6
106.2
103-6
105.68
(O.lUO ± 0.020 atm)
-------
2.5 -
2.5 •
1.5 -
LU
O
CD
OH
O
m i.o -
0.5 -
0.0
10
15 20 25
(TORR -METER)
30
35
Figure 25.
Plot of absorbance of 791 cm-1 Q-branch of N02 as a
function of partial pressure of N02 times path length.
The squares represent data taken through the 7.5 cm
cell and the circles represent data taken through the
39.5 cm cell
50
-------
dK_ dKP^dP, oft PS urg r?o>,
K ~ cipT K PL ^ K P2 ^ '
where dK/K, dPi/Pi, and dP2/P2 represent fractional errors in K, PL,
and PS, respectively. Using Eq. (27), it is found that
o-K PL = q(q - l)PiPa
c^PY K (P2 - aP1)(P1ir£> - P2)
and
'(* ~ V*& (29b)
_
K (P2 - aP1)(Picr? - P2)
For the pressures measured in this experiment and cr = 5-27, these
expressions are approximately equal to 10. Thus
* = 10 fL - 10
K. r\
Our error in measuring the pressure was estimated to be about 1%.
Summing the squares of these errors and taking the square root yields
a theoretical error of about 15%. Thus, by this technique, we estimate
the value of the equilibrium constant for the reactions W204^2N02 at
room temperature to be O.lU ± 0.02 atm.
This result is to be compared with the results of Verhock and Daniels5
who found K = 0.1^26 atm at room temperature using a nonspectroscopic
technique which assumed the validity of the ideal gas law. Dunn, Wark,
and Agnew10 used an infrared spectroscopic technique to find
K = 0.130 atm at room temperature. Harris and Churney9 used ultraviolet
spectroscopy to study the equilibrium constant at several temperatures.
They found K = 0.162U atm at 299.71 K, which was the closest tempera-
ture to our 296 K study.
C. STUDY OF THE KINETICS AND THE MECHANISM OF THE
REACTIONS IN THE GASEOUS SYSTEMS, HONO, NO, N02, H20
1. Introduction
Many of the general aspects of the mechanism of photochemical smog
formation are reasonably well understood today.11 The chemical trans-
formation of the atmospheric mixture of hydrocarbons, carbon monoxide,
and the oxides of nitrogen to the notorious blend of irritants present
in photochemical smog, ozone, peroxyacylnitrates, acids, aldehydes, and
the other oxidation products of the hydrocarbons, is believed to occur
by way of a series of chain reactions which involve various active free
radical species as chain carriers. In theory, one of the most important
of the active species is the HO-radical. The nitrous acid molecule is
one of several potential primary sources of this radical. Nitrous acid
may build up in the atmosphere through the occurrence of reactions 1
and 2:
51
-------
. NO + NO2 + H20 -* 2HONO (l)
2HONO -f NO + N02 + H20 (2)
In turn the photolysis of nitrous acid in sunlight may generate
HO-radicals:
HONO + hv(A < UOOO A) -» HO + NO (31)
It is believed that the HO-radical reacts readily with the hydrocarbons
(RH), aldehydes (RCHO), carbon monoxide (CO), and other species, to
initiate the sequence of reactions which result in the oxidation of NO
to N02 and the ultimate generation of ozone. The HO-radical reaction
with each of these types of compounds in the atmosphere leads to the
formation of alkyl peroxy (R02), acyl peroxy (RC002) and/or hydroperoxy
(H02) radicals:
HO + RH-^ H20 + R (32)
R + 02 -» R02 (33)
HO + RCHO -» H20 + RCO (3*0
RCO + 02 -* RC002 (35)
HO + CO-> H + C02 (36)
H + 02 (+M) -» H02 (+M) (37)
These peroxy radicals are thought to oxidize NO to NOp and, subsequently
regenerate additional HO-radicals:
R02 + NO -» RO + N02 (38)
RO + 02 -> H02 + R'O (39)
(e.g., CH;£> + 02 -* H02 + CH20)
RCOOp + NO -» RC02 + N02 (UO)
RC02 -* R + C02
H20 + NO -» HO + N02 (U2)
Thus, a single primary HO-radical may initiate a sequence of reactions
involving many cycles which may result in the oxidation of several
molecules of hydrocarbon, CO, and NO. The oxidation of NO to N02 is a
particularly significant event in that it controls the ozone level
through the following reactions:
-------
N02 + hv(7\ < IGOO A) -» 0(3P) + NO (U3)
0(3P) + 02 (+M) -» 03 (+M) (UU)
03 + NO -* 02 + N02
The rates of reactions ^3-^5 are sufficiently fast in a sunlight
irradiated atmosphere that a photostationary state of 63 develops which
is determined largely by the [N02]/[NO] ratio and the solar irradiance
in the ulxraviolet region effective for reaction U3- Thus, it can be
seen that the primary rate of generation of the HO-radical in a polluted
atmosphere is an important key to the concentration of ozone, the rate
of the generation of the other manifestations of photochemical smog,
and the severity of the smog episode.
There has been much speculation on the nature of the primary sources of
the HO-radical in the sunlight-irradiated, polluted atmosphere, and it
is not clear even today to what extent reactions 1, 2, and 31 generate
HO. Although Calvert and McQuigg1? concluded from smog simulation
studies that the nitrous acid molecule is a major primary source of HO
through nitrous acid photolysis in sunlight, in their reaction mechanism
the major source of HONO was the reaction of H02 with N02 :
H02 + N02 -» HONO + 02 (U6)
Although the reactions 1 and 2 may in theory also control the nitrous
acid levels in the atmosphere, their importance in the atmosphere is
uncertain. The existing kinetic information on these reactions appar-
ently does not apply to the homogeneous reactions . In the first de-
tailed study of the NO-N02-H20-HONO system, Wayne and Yost13 followed
N02 disappearance in a mixture rich in NO and HaO . The rates were
reasonably fast. Leighton14 derived from these data kL == U.3 x 10"6
ppm"2min~1, assuming that the kinetics of the reaction followed those
of the elementary reaction 1. However, in the subsequent study by
Graham and Tyler,15 much smaller values for k], were observed:
kL s (1.2 ± 0.6) x 10""9 ppm'^in"1. As in the Wayne and Yost results,
the kinetics at high water concentration did not follow the order with
respect to the reactants predicted by the elementary step 1. The large
difference between the rate constants observed in the two studies was
attributed to the large^dLfference in surface-to-volume ratio of the
reaction vessels used "(factor of UO). Hence, the observed reaction was
thought to be/controlled entirely by heterogeneous wall reactions .
Recently England and Corcoran16 have invoked the occurrence of
reactions 1 and 2 to rationalize their observed kinetic effects of water
vapor on the rate of oxidation of NO. to N02 in the NO, N02, 02, HpO
system. They concluded from very indirect evidence that reaction 1 did
occur homogeneously at temperatures of UO°C and above with
kL = 1.5 x 10~e
Reaction 2 has not been observed in kinetic studies, but if the homo-
geneous reaction 1 is slow, as the present evidence suggests, then one
53
-------
expects the reverse reaction 2 to be slow as well, since the ratio of
the two rate constants is equal to the equilibrium constant for re-
action 1. Indeed qualitative evidence for the slowness of reaction 2
has been suggested in the studies of Asquith and Tyler,17 Nash,18 and
Cos and co-workers.19 They have observed that MONO vapor is seemingly
quite stable at concentrations well above its equilibrium level.
All present evidence suggests that the rate constants for the homoge-
neous reactions 1 and 2 are small, and there is some question as to
whether these reactions occur measurably at all in the gas phase.
Demerjian, Kerr, and Calvertlla also came to this conclusion in their
review of the mechanism of photochemical smog formation. However, they
noted that if HONO was present in an auto exhaust polluted atmosphere
of typical composition at a concentration near those for equilibrium
with the NO, N02, and H20 present in the early morning hours, the pre-
dicted rate of conversion of NO to N02 would be enhanced by a factor of
two over that expected in the absence of HONO. Then it is clearly
important to evaluate the rate constants for the homogeneous reactions 1
and 2 in order to define the initial rates of smog formation. Since the
heterogeneity of the system was evident in previous studies of the
H20-NO-N02-HONO system, existing data can give no information as to the
desired homogeneous rate constants.
In all of the previous kinetic studies of the nitrous acid system no
direct measurement of nitrous acid itself was made. Previous workers
depended upon either the determination of changes in visible light
absorption due to the reactant N02 alone, or on chemiluminescence de-
tection of NO and NOX coupled with the selective chemical removal of
HONO vapors. Obviously, rate information based on the direct measure-
ment of HONO vapor is necessary to test the hypotheses formulated from
the indirect experiments and to derive meaningful values of kx and k2.
This has been possible in the work described here in which an infrared
Fourier transform spectroscopic analysis of the NO, N02, H20, HONO
system allowed essentially continuous monitoring of all the reactants
and products. The relatively large reaction vessel employed in this
work ensured a low surface-to-volume ratio and the minimization of wall
effects. The cell size and the pressure of added nitrogen gas
(700 Torr) favored the homogeneous reaction paths for reactants. The
multiple pass White optical system allowed path lengths from 82 m to
1.5 km and measurements of very low reactant and product concentrations
in the ppm range of major interest in the atmospheric reactions. We
believe that this work provides the first direct measurement of the rate
constants of the homogeneous reactions 1 and 2. In terms of these re-
sults reasonable speculation is made about the potential role of these
reactions in the atmosphere.
-------
2. Experimental Methods and Techniques
a. Reaction vessel and the associated optical system
The reactions between NO, N02, H20, HONO mixtures were studied in a
large stainless steel (Type 30*0 tank, 21 m in length and 76 cm in
internal diameter. It could be evacuated using its associated Kinney
high vacuum booster pump (KMBD) to a pressure of a few microns in about
one hour when starting at atmospheric pressure within the cell. The
cell housed a White optical system composed of three mirrors (20.5 m
radius of curvature) with a base path length of 20.5 m. Two of the
mirrors could be adjusted through external controls to allow alignment
and variation in the number of optical traversals employed. Path
lengths from 82 m up to 1.5 km could be employed. A Digilab FTS 20
system, a Nernst glower, and associated transfer optics provided the
infrared beam which entered and exited the cell through Nad windows.
The exit beam was collected by a detector. In this study a liquid
helium cooled, copper-doped germanium detector was employed; its useful
range extended from 300 to 3500 cm"1.
b. Gas handling system
The gas handling system was constructed from bulbs and tubing made of
glass and stainless steel, and all stopcocks were made of Teflon. The
introduction of measured quantities of the standard reactant gases, NO
and N02, was accomplished by filling calibrated 1 I bulbs to some
measured pressure and expanding these into the tank. About a 104 to 1
dilution of reactant concentrations occurred on expansion. Pressures
were measured on one of three calibrated Wallace and Tiernan gauges
(0-20, 0-50, 0-800 Torr) appropriate for the measurement. The gauges
were isolated from the sample by a quartz spiral gauge which was used
as a null instrument.
c. Calibration of reactant gas absorptance data
Although there are several infrared absorption bands characteristic of
NO, N02, and the isomers of HONO in the mid-infrared region, all of
them could not be employed for concentration monitoring purposes since
some are either overlapping those of atmospheric C02 and H20 or they
mutually interfere. These interferences were reduced by obtaining all
spectra with a spectral resolution of 0.5 cm"1. Those unambiguous
absorption peaks which were employed in this work are: N02, v2 band,
Q-branch transition at 823 cm"1; NO, fundamental Q-branch at l8?6 cm"1;
cis-HONO, V4 band, Q-branch transition at 853 cm"1; trans-HONO, v3 band,
Q-branch transition at 126U cm"1. All absorbances were measured after
adjusting the total pressure of the gaseous mixture to 700 Torr with
added nitrogen gas. The single beam spectrometric system employed
required that a background spectrum of the detector profile and the cell
filled with high purity nitrogen (700 Torr) be recorded first and then
ratioed against the sample spectrum. Through the use of this procedure
the unwanted interferences from the absorption due to atmospheric water
55
-------
vapor and COc; were minimized. Calibration was carried out in three
cells of varied path length, 7-5 cm, 39-5 cm, and 20.5 m, respectively.
In the case of the reactant nitric oxide (Matheson Co., 99fo NO), a
measured pressure of the gas was introduced into one of the three cells,
and then the mixture was pressurized to 700 Torr with high purity nitro-
gen. The absorbance at the distinctive Q-branch of NO at 1876 cm"1 was
then measured. Plots of absorbance versus path length-concentration
product were linear over the wide range of concentrations investigated
here. The calibration procedure with nitrogen dioxide gas was neces-
sarily more complicated because of the dynamic equilibrium, 2N02 ~ N20*.
The NOp-NpO* sample (Matheson Co., 99.5% Nx02x) was purified further
before use by condensation at liquid N2 temperature and degassed to re-
move any volatile impurities (NO, NaO, N2, 02j etc.). The degassed
sample was then allowed to expand into the desired cell for calibration
and the pressure of the N02-N204 mixture was measured. The correspond-
ing partial pressure of the N02 gas was then calculated using the known
equilibrium constant for the temperature employed (23°C). Again the mix-
ture was pressurized to 700 Torr with added high purity nitrogen gas.
The absorbance of the N02 was measured at the characteristic Q-branch
transition at 823 cm-1.
The calibration of nitrous acid vapor spectra is complicated because
pure samples of HONO vapor cannot be prepared at some desired pressure
since it is relatively unstable at concentrations above those corre-
sponding to its rather low equilibrium pressure in NO, N02, H20 mix-
tures. Thus, measurements were made on equilibrium mixtures at 23°C to
establish our HONO calibrations. We have calculated the extinction
coefficient of the HONO isomers using the known equilibrium ratio of
[trans-HONO]/[cis-HONO] = 2.29, and KI = 1.51 x 10~6 ppm'1 at 25°C
(least squares fit of published data) ,13'2>o's>1 together with measured
absorbances .of trans- and cis-HONO, NO, and N02 in mixtures of known
pressure of water vapor. The extinction coefficients
(e = [^(I0/l)]/pl) which we have determined and employed in this study
(spectral resolution of 0.5 cm"1, temp., 23°C) are as follows:
NO (1876 cm"1), e = 0.513 ± 0.006 Torr"^"1; N02 (823 cm"1),
0.0807 ± 0.0012 Torr"1^"1; cis-HONO (853 cm"1), 9.36 ± 0.9U Torr^m"1;
trans-HONO (126U cm"1), 5-20 ± 0.52 Torr"1!!!"1.
In kinetic studies of the HONO decomposition reaction, samples of
nitrous acid much in excess of the equilibrium pressure were introduced
into the large reaction cell following a procedure modified somewhat
from that used in the studies of Nash18 and Cox and co-workers.13
Three methods were attempted: (l) dilute H2S04 was added to dilute
NaN02 solution; (2) dilute H2S04 was added to solid NaN02; and (3) di-
lute NaN02 was added to dilute H2S04 solution. The third method proved
to be the best for our conditions because it gave the maximum ratio of
[HONO]/([NO] + [N02]). The gaseous nitrous acid was produced in the
apparatus shown in Fig. 20. A dilute NaNOp solution (~ 0.3 M) was added
slowly to a dilute H2S04 solution (~ 2.5 M) by means of the dropping
funnel shown. Nitrogen gas was allowed to flow through the entering
tube which was directed upward at the solution-gas interface. The
56
-------
vapors containing HONO were swept into the tubing leading to the cell
by the flow of nitrogen gas. The entire apparatus was flushed with
nitrogen before use. The trap containing concentrated NaOH solution
prevented the entry of air during the preparation and served as a sink
for the excess of reactants after HONO preparation. The passage to the
NaOH trap was closed by the clamp as HONO was prepared and flushed into
the tank.
d. Experimental procedure in absorption
(l) HONO decomposition reaction. Before each run a carefully measured
quantity of liquid water was allowed to evaporate and enter the evacu-
ated reaction cell. The transfer was facilitated through the use of a
nitrogen carrier gas. Water sufficient to provide in the tank
600-U200 ppm in H20 concentration was used so that H20 which was formed
in the reactions was always a negligible fraction of the total water
present. Then the cell was filled to a pressure of UOO-500 Torr of high
purity nitrogen gas. During a five-minute period the HONO was added by
means of the nitrogen carrier and the apparatus shown in Fig. 20.
Finally the passage to the cell was blocked off, and the remainder of
the HONO-NO-NOp mixture was trapped in the concentrated NaOH solution.
During a 5-10 minute period, a very fast flow of nitrogen gas was used
to mix the gaseous reactants turbulently and to achieve a final pressure
of 700 Torr. Spectra were taken (2 min collection period) at regular
time intervals (k~10 min), stored on magnetic tape, and the absorption
spectra were computed automatically by ratioing against the previously
recorded background cell spectrum. In these runs a 2U6 m path was
normally employed. Typical spectra for several time periods can be
seen in Fig. 26a. The absorption due to the vibrational transitions of
trans-HONO (791 cm"1) and cis-HONO (853 cm"1) are seen to decrease as
the decomposition of HONO occurs, and absorption due to N02 transitions
appear at 823 cm'1 and at several other wavelengths between the two
HONO v4 bands. One N02 absorption peak overlaps that of trans-HONO so
that the 791 cm"1 band was not used as a quantitative measure of trans-
HONO. The 12.6U cm"1 absorption band of this isomer showed no such
interference and was used for quantitative trans-HONO estimates.
(2) HONO formation reaction. Water was introduced to the evacuated cell
as before, then NO gas, followed by N2 gas to bring the total pressure
to UOO-500 Torr. N02 was then added and the cell pressurized to
700 Torr with a very rapid flow of nitrogen gas. Spectra were recorded
as before, but, in this case, a path of U92 m was employed. Absorbances
of NO, N02, cis-HONO, and trans-HONO were then calculated from plots of
the ratioed absorption spectra for each measured time period. Typical
absorptance changes in one spectral region can be seen in Fig. 26b. The
spike absorptances due to the N02 Q-branch transitions at 823 cm'1 and
several other spectral positions occur as prominant features while
growth of the trans-HONO and cis-HONO bands is observed at the longer
run times.
57
-------
LU
O
LU
O
o:
TIME* 236 min
irw*^^
TIME =602 min
750
800 850
WAVENUMBERS(CM-')
(a)
900
Figure 26. Absorption spectrum of reacting mixture of HONO, NO, N02, and
H30; note in (a) decreae in the v4 bands of trans-HOHO (791 cm-1)
and cis-HONO (853 cm-1) and buildup of NQ2 transitions as time
progresses in an originally HONO-poor mixture
-------
LJ
O
CO
z
<
cr
TIME = 0 min
LJ
O
CO
<
o:
TIME = 301 min
LJ
O
I
CO
tr
TIME= 602 min
750
800 850
WAVENUMBERS(CMn)
(b)
Figure 26. Continued
900
59
-------
3. Experimental Results and Discussion
a. The kinetics of HONO decomposition
in the gas phase
In the present work we have introduced gaseous HONO into the reaction
cell at concentrations well in excess of its equilibrium values and
observed by infrared Fourier transform spectroscopy the kinetics of its
return to equilibrium with NO, N02, and HpO vapors presumably through
the occurrence of reactions 1 and 2.
2HONO -> NO + N02 + H20 (2)
NO + N02 + H20 -* 2HONO (l)
The data given in Fig. 27 show the time dependence of the absorbance
measured at the maximum in the Q-branch of the trans-HONO v3 band at
12614 cm"1 and the cjls-HONO v4 band at 853 on'1. Note that the two
isomers decay essentially in tandem, and, that the rate is very slow;
for the concentrations employed here, somewhat less than one-half of
the HONO has disappeared in about 200 minutes. The ratio of the two
absorbances (acis/atrans) stays nearly constant although it may increase
very slightly with time. For example, in the data shown in Fig, 27, the
linear least squares fit of the dependence of the absorbance ratio on
time follows the equation, (acis/atrans) = 0.72U ± 0.012 + (7.2 ± 3.8)
x 10~5t(min), through. 575 min of the reaction. The absorbance ratios
noted during the reaction appear to be somewhat lower than those
observed for mixtures near equilibrium at room temperature. From 172
different determinations in runs with temperatures from 21.k to 30.0°C,
the average value of acis/atrans = 0.790 ± O.OUO; no significant trend
with temperature over this relatively small temperature interval was
obtained. Thus there is some suggestion that in the decomposition
reaction 2 the cis-HONO isomer is depleted more rapidly than the trans-
HONO isomer. However, our data provides no convincing evidence of this;
in fact, one would not expect such an effect in theory since the rela-
tively small barrier to isomerization of the HONO isomers,
8.7 ± 1 kcal/mole,?2 should allow equilibration of the isomers to occur
very rapidly even if one of the isomers were removed preferentially.
Thus, if the isomerization reaction is in the first order region at
700 Torr and it has a "normal" pre-exponential factor (~ 1013 s'1),
then T1/2 = 1.6 x 10"7 s at 23°C.
A clue as to the nature of the kinetics of the decay of nitrous acid
can be obtained from the plot of l/acis and 1/a-trans versus time data
for the early times in a typical run shown in Fig. 28. Note that the
data follow well a second order decay for both isomers over the time
range employed. The least squares slopes of the lines, (5-03 ± O.OU)
x 10"3 and (3.86 ± 0.08) x 10~3 min"1 for the cis- and trans-HONO data,
respectively, may be used to derive an estimate of the second order rate
constant for total HONO loss.
60
-------
1.5-
LU
O
00
cr
o
CO
GO 0.5
trans-HONO (1264cm"1)
cis-HONO (853 cm'1)
100
200
300
400
500
TIME (MIN)
Figure 27. Absorbance versus time plot for the trans-HONO (126U cm"1) and
cis-HONO (853 cm'1) in a HONO-NO-N02-H20 mixture originally
rich in HONO
-------
LU
O
CD
cr
o
O)
CD
4.CH
3.5
3.0
2.5
2.0
1.5
1.0
0.5
cis-HONO
trans- MONO
100 200 300 400 500
TIME (MIN)
Figure 28. Plot of the reciprocal of the absorbance of
trans-HONO (126U cm"1) and cis-HONO (853 cm'1)
versus time for a HONO-NO-N02-H20 mixture
originally rich in HONO
62
-------
Here kc^s and k^.rans refer to the apparent second order rate constants
observed from the slopes of the cis- and trans -HONO decays, respec-
tively, and R is the equilibrium ratio of [trans-HON01/[ cis -HONO ] s 2.29
(23°C). From this procedure we estimate k2 = (1.28 ± oToS) x 10~3
As the reaction proceeds to much longer times the rate of nitrous acid
loss decreases until equilibrium levels of HONO, N02, NO, and H20 are
achieved. This may be noted in the extended run shown in Fig. 29. The
total nitrous acid concentration is plotted for a run which extended for
over 28 hours. Obviously a nitrous acid forming reaction becomes impor-
tant at long times. It is reasonable to assume this effect to be the
result of the occurrence of reaction 1, the reverse of reaction 2.
NO + N02 + H20 -» 2HONO (l)
If reactions 1 and 2 are the only reactions involving HONO in our
system, then the simple rate law which should describe HONO decay over
the entire time scale is Eq.
- —dl~ = (PHONO)?2k2 - PNoPN02PH2o2k1
For the conditions employed in our study there is a large excess of
water vapor present so that PHPQ remains essentially constant through-
out the run. The integrated form of the rate Eq. (Uy) for the condi-
tions employed may be given by Eq. (U8).
(PHOHO)t - (FMO). - ! <«»
where
a = (PHONO )Q 2k2 - (PHpO^PNOsJo (PNO)02ki
b - - (^k2( PHONO )0 + (PH20)0[(PN02)0 + (PNO)o]k
c = 2k2 -
q = kac - b?
and PHONO refers "to the total pressure of nitrous acid present,
^trans-HONO + ^cis-HONO' Subscripts of t and o on the reactant pres-
sures refer to time t and 0, respectively. A test of the fit of the
data of an extended run of 16U8 min duration to the form of Eq. (U8) is
63
-------
NO*NOp+ H90 ^ 2 MONO
. u
K2
= |.5l*IO-6ppm-' -- Keq
k, - 1.4 « I03ppm"lmin"'
k, = 2.0«I03ppm"'min
0.4
500
IOOO
1500
TIME (MIN)
Figure 29. Plot of total HONO concentration versus time for an extended run of 16U8 min
duration using an initially HONO-rich mixture; initial concentrations, ppin:
HONO, 2.1k; NO, 2.37; N02, 3.53; H20, ^200; temperature, 23°C. Curves shown
have been calculated using relation kQ and the rate constant values indicated
on the figure.
-------
shown in Fig. 29. In attempting to fit the theoretical time dependence
curve to these data and derive rate constant information, we varied the
value of k2 chosen for the calculation, but then the choice of kL was
fixed through the known equilibrium constant and its relation to kx and
k2: kx = k2Ki ; KI corresponds to the equilibrium: NO + N02 + H20 -
2HONO. The sensitivity of the fit of the theoretical curve to the
choice of k2 may be seen in the figure. For the particular conditions
employed in this run a choice of k2 = l.U x 10~3 ppnT^-min"1, and hence,
kL = (1.51 x 10~6)(1.U x 10"3) = 2.1 x 10"9 ppm^min'1, fits well the
data through the Eq. (U8) over the entire range of the experiment. This
choice of k2 also is in good accord with that derived from the l/acis
and 1/a-trans versus time plots of the initial rate data as outlined
above: k2 = (1.3 ± 0.1) x 10~3 ppm'-'-min"1. Similar kinetic constants
were obtained from other experiments in which the initial pressures of
HONO, N02, NO, and H20 were varied and the data were fitted again as
outlined. The estimates of k2 and kx derived from these experiments
are summarized in Table 2. There is reasonably good agreement between
the estimated constants from the different experiments. Note that the
rate constant estimates obtained from runs at 20 Torr and 300 Torr of
added nitrogen gas pressure are, within the experimental error, the same
as those found in the usual experiments which were pressurized to
TOO Torr of nitrogen gas. This supports the conclusion that the rate
of HONO disappearance in these runs is controlled by homogeneous reac-
tions and is not influenced significantly by reactions which occur at
the cell wall.
This study provides the first measurements of the HONO decomposition
reaction. The data seem to suggest that HONO decay occurs by the
stoichiometry required by the elementary reaction 2 with a rate con-
stant k2 = (l.U ± O.U) x 10~3 ppnT^-min'1 at 23 ± 1°C. The fit of these
rate data to the decay at long times gives kL = (2.1 ± 0.7) x 10"9
ppm^min'1 at 23 ± 1°C.
b. The kinetics of the HONO formation
Reaction in the gas phase
Evidence of the kinetics which control reaction H20 + NO + N02 -» 2HONO,
is derived from the dependence of the rate of HONO decay over the long
times during which equilibrium is approached; values of kL were obtained
from the type of kinetic data in the previous section. An attempt was
made to determine kj. in more direct experiments in which NO, N02 and
HpO were added to the cell and the rate of HONO formation was observed.
This reaction proved to be much more difficult to observe quantitatively
in our apparatus than the study of reaction 2. The difficulty arose
primarily in making up the original N02, NO, and H20 dilute mixture in
the cell without allowing the reactants to encounter one another pre-
maturely at elevated concentrations. In general, we were successful in
preparing mixtures which were only a factor of two or three from the
equilibrium level of HONO, so the data are less useful than we had
hoped. One set of data from an experiment of this type is shown in
65
-------
Table 2. Summary of the Rate Constant Estimates3 for the Reactions:
NO + N02 + H^ -» 2HONO
2HONO -* NO + N02 + H20
(1)
(2)
Initial concentrations, ppm
[NO!
a) 2HONO
5.U2
5.65
U.6U
5.18
2.37
U.23
1.91
b) NO +
15.8
10.6
[N02]
-4 NO + N02
7.8l
7.01
5.60
U.89
3-53
8.50
10.7
N02 + H20 -»
8.16
9.72
[H20] [HONO]
+ HgO Experiments
560 6.73
1120 6.0U
22UO 5.21
2800 3.95
U200 2.1U
190 9.01
190 U . 77
2HONO Experiments
lUoo 0.130
22UO O.lUl
PN2, Torr
700
700
700
700
700
300
20
700
700
Best Estimates:
k2 5 ppm min
1.6 x IO-3
1.6 x lO'3
l.U x 10"3
0.90 x 10"3
l.U x IO-3
1.5 x 10-3C
2.3 x 10-3C
0.80 x IO-3
l.U x 10~3
(l.U ± O.U) x IO-3
kx, ppm^min-1
2.U x 10"9
2.U x 10"3
2.1 x 10"9
l.U x 10"9
2.1 x IO-9
2.3 x 10-nc
3.5 x 10-9C
1.2 x IO-9
2.1 x 10'9
(2.2 ± 0.7) x lO'9
Temperature in all runs, 23 ± 1°C.
bThe unit ppm used here is defined as [pressure(Torr)'/?60] x 10s at 23°C.
Q
Calculated assuming there is no pressure dependence in the extinction coefficient of the Q-branch of
the vibrational bands used for concentration estimation; this appears to be a good approximation from
a limited study of the effect of pressure on the absorbance of the species involved here.
-------
Fig. 30. Again an attempt was made to match the concentration versus
time profile for HONO formation using the rate equation given by
Eq. (U8). The fits using various values for kj. are shown. In this case
k2 was adjusted so that k2 = ki/K^.. The estimated rate constants from
two such runs are given in Table 2, Section B. Although our precision
is low for these experiments, we conclude that the same rate constants
derived in the study of reaction 2 fit the reverse reaction kinetics
reasonably well; the best estimate from all of the data gives
kL = (2.2 ± 0.7) x 10~9 ppnT^min"1 . Reaction 1 has been studied pre-
viously and the results of these studies should be reviewed here in
light of our new findings.
The Graham and Tyler1"5 kinetic data for reaction 1 were derived from
experiments over a wide range of reactant concentrations:
PNO = ^-3-200 Torr; PNQ^ = 1.7-7-5 Torr; F^Q = U.7-11* Torr. Although
these results seem to show an enhanced rate of reaction for runs at
PJJ Q > 9 Torr, which suggests an order for H20 higher than one, the cal-
culated third order rate constants for the reaction 1 for P^Q < 9 Torr
appear to be reasonable constant over the wide range of reactant concen-
trations employed. Indeed it is interesting to observe that the average
value of the third order rate constant kj. derived from the Graham and
Tyler data, ki = (1.2 ± 0.6) x 10~9 ppm"?min~1, is equal, within the
experimental uncertainties of the two very different measurements, to
that observed here: kL = (2.2 ± 0.7) x 10~9 ppn^min"1. Since the
surface-to-volume ratio of the reaction vessel in our study was much
lower than, that of Graham and Tyler (0.052 cm"1 compared to 1 cm"1) and
the concentration ranges of all of the reactants were several orders of
magnitude lower in our work, the degree of agreement observed from the
two very different systems provides strong evidence that the rate con-
stant observed here is indeed that for a homogeneous reaction. Appar-
ently the data of Graham and Tyler also may apply to the homogeneous
system, at least at the lower H20 pressures.
c. Mechanism of the nitrous acid
formation and decay reactions
In most previous work the mechanism of the nitrous acid formation
reaction has been considered to be the following:
N02 + NO * N203 (U9)
N203 + H20-* 2HONO (50)
Since the equilibrium ^9 is established rather rapidly,5'23'24 the occur-
rence of this mechanism would be consistent with the overall kinetics
of nitrous acid formation observed here and in the work of Graham and
Tyler. The rate of nitrous acid formation would be given by Eq. (51)
^HONO ovtxTXtx (^ }
dt ~ 22
67
-------
oo
CL
CL
CO
o
I—I
o
z
O
X
0.08
0.06
0.04
k, = 2.5 x | 0-9ppm-2min-'
k, = 2-1 x I 0-9ppm-2min->
k, = 1.7 x
•^ = 1.51 xl 0-6pprrr' = K
kg
eq
50
100
150
TIME (MIN)
200
Figure 30. Plot of cis-HONO concentration versus time using an initially HONO-poor
mixture; initial concentrations, ppm: HONO, O.lUl; NO, 10.6; N02, 9.72;
H20, 22UO; temperature, 23°C. Curves shown have been calculated using
relation kQ and the rate constant values indicated on the figure.
-------
The value for K2i is 5.2 x 10"7 ppm"1 at 25°C.25 Thus in terms of the
mechanism ^9, 50 for HONO formation and the present value for kx ,
k22 = ^.2 x 10~3 ppm~1min~1 . The transition state for this hypotheti-
cal route of reaction would, in theory, involve a rather complex
transition state:
N203 + K20 * * 2HONO (52)
o'V
The pre- exponential A-factor for such a highly ordered transition state
required for the hypothetical reaction 52 should be much lower than the
bimolecular collision number. If the activation energy for reaction 50
is near zero (£$.293 = 0.6 kcal), then, in a sense, the N^a involvement
could be justified in terms of the present kinetic data. However, if
one considers the reverse reaction 2 and the requirement that the same
transition state be involved as in reaction 1, then the kinetic data
definitely favor an alternative reaction scheme.
A seemingly more realistic mechanism for the reactions 1 and 2 involves
a much simpler transition state:
°\
N
NO + N02 + H20 * 0 - 2HONO (53)
H' NH
Here the transition state involves formation of only one common H-atom
bond between the reactants . Note the similarity between this reaction
pathway and that proposed for the ternary reaction of NO oxidation and
its reverse bimolecular reaction:26"27
NO + NO + 02 * Q' * 2N02 (5U)
0
<0
The entropy of activation for reaction 5h should be very similar to that
for 53 so one might expect very similar pre -exponential factors for the
two reactions. The activation energy of reaction 2 is not known so we
cannot estimate directly the experimental A-factor for this reaction;
£#29 a f°r the reaction is about -9-5 kcal. The measured rate constant
for the forward reaction 5^> k54 = 1.^3 x 10~9 ppnT^min"1 ,27 is very
nearly equal to that observed for the reaction 1; kL - (2.2 ± 0.7) x
10~9 ppm'^min"1. The similarity in transition states of reactions 53
69
-------
and 5U, and hence pre-exponential factors, coupled with the similarity
in rate constants, leads one to conclude that EI may be near zero as
with £54. Thus, we may estimate E2 = 9-5 kcal and A2 = 1.3 x 10"4
ppm~1min~1. This value is the approximate magnitude of the pre-
exponential factor for the bimolecular reaction given by the reverse of
reaction 5^ which has a very similar transition state; A_54 = 1.0 x 10~4
pprn'-'min"1,28
All of the present data seem consistent with the simple mechanism out-
lined in Eq. (5*0 as a description of reactions 1 and 2. The involve-
ment of N203 as an intermediate in the reaction seems unnecessary. The
activation energy assignments made here may be tested with the limited
temperature variation studies described in the following section.
d. Equilibrium studies and the temperature dependence of the
nitrous acid formation and decay reactions
The effect of temperature on the rate of the HONO decay reaction could
not be determined accurately with the large reaction vessel employed in
this work since the tank had no special temperature control other than
the air conditioning system for the laboratory in which it was housed.
This air control system was used in a series of runs designed to de-
termine the equilibrium constant and rates of [HONO] change with tem-
perature. Typical data for one of the mixtures studied is shown in
Fig. 31. The tank contained a mixture of NO, N02, H20, and HONO at
equilibrium. The air conditioner was turned off in the morning hours,
and the temperatures of the room and the tank housed therein were
allowed to rise during the day as the outside temperature climbed. The
infrared spectra of the reactants and the temperature at the center of
the tank were monitored at regular intervals. Note in Fig. 31 that the
nitrous acid pressure versus time plot mirrors the temperature-time
variation within the cell; the nitrous acid is depleted as the tempera-
ture rises and the system attempts to reestablish equilibrium as a shift
away from the exothermic direction of the equilibrium position occurs:
2HONO -» NO + N02 + H20. After the system had reached a desired tempera-
ture, the air conditioner was turned on again, and the temperatures of
the room and the cell were lowered while the monitoring continued. At
about 626 minutes after the start of the experiment shown in Fig. 31 >
the rate of change of nitrous acid became equal to zero; at this time
the values of PNQ, %02J ?ti.20> pcis-HONO> and ptrans-HONO correspond to
those for equilibrium at the particular cell temperature recorded for
this time (29.8°C in Fig. 31). From a series of such temperature drift
data in which the d[HONO]/dt = 0 was caused to occur at a series of
different temperatures, we were able to estimate the equilibrium con-
stant for the NO, N02, H20, HONO system at several temperatures. These
data are given in Table 3 and refer to the equilibrium, NO + N02 + H^ =
2HONO. The data are reasonably consistent with equilibrium data derived
from previous indirect measurements in which only N02 concentrations
were monitored and a mass balance assumption was made. All of the data
are plotted in Fig. 32. A least squares fit of the previous equilibrium
data for the HONO, H20, NO, N02 system13*2"0'5"1 gives the equation,
70
-------
o
o
o
200 400
TIME (MIN)
600
800
Figure 31. Plot of [HONO] and temperature within the reaction cell versus
time for an HONO-NO-N02-H20 mixture near equilibrium; typical
data used to determine the equilibrium constant for the reaction,
NO + N02 + H20 ^ 2HONO, and to provide a qualitative test of the
temperature dependence to the rate constants
-------
Table 3- Equilibrium Data Derived from the
Temperature Drift Experiments:
NO + N02 + H20 *; 2HONO (l)
Conditions at d[HONO]/dt = 0; concn., ppma
Temp., K
295.3
295.6
295.6
296.9
297.0
297.9
299.7
300.1
303.0
303.1
NO
9.26
9.26
2.7U
lU.O
2.57
9-30
9.^9
7.96
7.^8
9.28
N02
1 |l C
1^4 *5
7.00
23.3
6.21
1U.5U
12.6
1U.6
15.2
12.7
HONO
0.539
0.5^2
0.350
0.63^
0.308
O.U70
O.U05
O.U06
0.3^7
0.355
H20
1323
1323
U200
915
U225
1335
921
1301
131U
931
KI , atm"1
1.6U
1.65
1.52
1.35
l.Ul
1.22
1.U9
1.09
0.806
1.15
aThe unit ppm used here is defined as [pressure (Torr)/760] x 10s at
the temperature shown for the particular experiments.
-------
0.5-
o.o-
T -0.5-
£
-i.o-
-1.5-
-2.0-
JANAF Data
2.8
3.0
3.2
o3
3.4
Figure 32. Plot of ^KCatnT1) versus 1/T for literature data and
present results for the equilibrium, NO + N02 +
2HONO; open circles, Ashmore and Tyler21; square,
average of data of Wayne and Yost1 ; triangles,
Waldorf and Babb20; closed circles, this work. JANAF
suggested values are shown along with a squares fit
of published data excluding present work.
73
-------
) = -15.56 ± 0.62 + (U.73 ± 0.19) x 103/T; the average of the
data of Wayne and Yost was included only as a single point in this
treatment since the scatter is greatest for this early work.- From this
relation we would estimate AH298 = 9-^ ± O.U kcal. Our present data
over the very limited temperature range available to us give,
^KCatm"1) = -15.09 ± 3-^5 + (^.59 ± 1.03) x 103/T and £H298 = 9-1 ±
2.0 kcal. Within the very limited accuracy of our data which extend
over only a 7.8°C temperature range, the agreement of all the data is
considered to be satisfactory.
The temperature drift data can be used to test, approximately, the
suggested temperature dependence of the rate constants, kx and k2. Here
the differential Eq. (U?) was employed with not only the reactant con-
centrations but the rate constants themselves being time dependent func-
tions related to the temperature changes with time. For one case,
shown in Fig. 31 by the dotted curve, we have tried to match well the
rate of change of P^QNO a* ^e early times; we have taken
ksCppm^min-1) = 6.7 x io4e-10(kcal/mole)/RT with kx = KLk2. The data
appear to be qualitatively consistent with the expectations of rela-
tion U8 and EL =0, E5 = 10 kcal/mole. The temperature drift data are
not good enough to attempt further refinement of values of EX and E2
since the nonuniformityof the cell temperature along its length could
not be avoided completely. However, we may conclude that these data
are consistent with a low activation energy for reaction 1 and
E2 s 10 kcal/mole. These data gives further credence to the alternative
reaction scheme 5^ which supports this choice of activation energies
indirectly.
e. Significance of the nitrous acid formation and
decay reactions in the polluted atmosphere
The present rate data for reactions 1 and 2 provide a new basis for the
evaluation of the significance of these reactions in the atmosphere.
Using the present rate data we have estimated the time dependence of
the HONO pressure for various levels of N02, NO, and H^ which are
typical of those encountered near pollution sources and in ambient air.
The results of these calculations are shown in Table h. Note that for
very high NOX levels, which correspond to those encountered near power
plant stacks and automotive exhaust pipes, the rate of HONO generation
is very significant. After only 10 seconds of elapsed time the PHQNQ
has achieved the 1.7 ppm level for the high NOX case. Obviously dilu-
tion of the reactants occurs as the exhaust gases mix turbulently with
the air, and no attempt has been made here to simulate that process.
The results of Table k show that the generation of nitrous acid slows
significantly as the NOx-HaO mixture reaches ambient levels of these
pollutants. Although after only five minutes of contact between reac-
tants in a mixture of NOX-H20 of composition equivalent to that of power
plant stack emissions, the PHQNO nas reached 79% of its equilibrium
value. However, the NOx-HaO mixture at ambient levels has reached only
6% of its equilibrium levels of HONO after 60 hours. The present data
-------
Table 14. Theoretical Development of HOMO as a Fujiclion of Time for Mixtures of NO, N02, and H30
for Compositions Typical of Stack Gas Kmiasions and Ambient Conditions
Initi
NO,
ppn
500
100
50
10
5
1
0.5
0.1
0.05
500
100
50
10
5
1
0.5
0.1
0.05
al Concentrations
Relative
H02, Humidity,
ppn % (25'C)
50
15
9
2
1.15
0.25
0.125
0.025
0.0125
50
15
9
2
1.15
0.25
0.125
0.025
0.0125
100
100
100
100
100
100
100
100
100
50
50
50
50
50
50
50
50
50
[NOal/
[NO]
0.10
0.15
0.18
0.20
0.23
0.25
0.25
0.25
0.25
0.10
0.15
0.18
0.20
0.23
0.25
0.25
0.25
0.25
Concentration of MONO at
1 s
0.057!*
0.003U5
0.00103
1*. 6xlO-s
1.3x10'*
5. 8 xlO-7
1.1* x 10"T
5.8x10-''
1.1* x 10":'
0.0287
1.7xlO'3
5-2 xlO-*
2.3x10""'
6.6x10-"
2.9x10-''
7.2 xlO""1
2.9x10- '
7.2X1Q-1'
1O s
0.572
0.031*5
0.0103
l».6xlO-"
1.3x10'"
5.8x10"'
l.U xlO-''1
5.8x10'"
l.lt xlO"'
0.286
1.7X10-"
5.2x10'-'
2.3x10-'
6.6x10""'
2.9x10-'-
7.2x10-'
2. 9x10-"
7.2x10" '
30 s
1.70
0.103
0.0310
l.U xlO-3
I*. Ox ID'*
1.7x10-"'
l*.3x ID'"
1.7x 10"'
1* . 3 x 10-"
0.85I.
0.0511*
1.5x10-'
6.9x 10""
2.CxKT"
8.6x10"
2.2xlO'r
8.6x10'"
2.2x10-"
5 min
11*. 6
1.008
0.306
l.UxlO-?
U.Ox 10'3
1.7x10-*
li.3xlO'*
1.7 xlO'7
It . 3 x 10'7
7.89
0.509
0.15I*
6.9xlO-n
2.0xlO'3
8.6x10-"'
2.1x 10-"
8.6x10-'
2 .2 x 10-7
10 rain
22.9
1.95
0.6oi*
2.7xlO-?
7.9x 10-3
3. It x 10-*
8. 6x10-*
3. It xlO"5
8.6x10-'
13-7
1.00
0.305
0.0137
0.0039
1.7x10-'
I* . 3 x 10-"'
1.7x10""
l*.3xlO-7
time shown,
1 hr
29.1
6.67
2.81
0.160
U. 7x10-*
2.1xlO-3
5.2x10"*
2.1X10"5
5.2x10'°
21.7
It .28
1.59
0.0810
0.0235
0.00103
2.6x10-"
l.OxlO-*
2.6x10-°
ppm
6 hr
29.1
7.33
U.06
0.677
0.2U3
.1.2x10-*
3.1x10-°
1.2x10-*
3.1x10-*
21.7
5-Ul
2.98
o.Uoo
0.131
0.0061
0.0015U
6.2x10-*
1.5x10-*
60 hr
29.1
7.33
U.06
0.860
O.U6U
0.080
0.026
1.2 xlO"3
3.1x10-*
21.7
5.U1
2.99
0.632
0.3UO
O.OU8U
0.01142
6.1x10"*
1.5x10'*
Equilibrium
29.1
7.33
U.06
0.860
O.U6U
0.0971
O.OU85
9.7 x 10"3
U.gxlO-3
21.7
5-Ul
2.99
0.632
0.31*1
0.072
0.0356
7.1xlO-3
3.6xlO"3
BThe unit ppm used here is defined as [pressure (Torr)/7oO] x 101"1.
-------
suggest that MONO may be generated at a significant rate during the
early stages of dilution of N0x-rich exhaust gases, but they also point
to the unimportance of reactions 1 and 2 in forming or removing HONO
when ambient levels of NOX are present.
It is believed that the oxidation of about 25% of the original NO to
NC>2 in auto exhaust gases occurs during the dilution process by reac-
tion 5^-5 "the nearest equivalent to reaction 1 noted. Thus the signifi-
cant occurrence of reaction 1 as power plant and automotive exhaust
gases mix with air seems highly likely. The present data suggest that
HONO vapor will be generated in the evening and the early morning hours
as NOx-rich exhaust gases are pumped into the air. It is reasonable to
assume that HONO pressures in the early morning hours may be a signifi-
cant fraction of the HONO equilibrium pressures expected from the NO,
N02, H20 ambient pressures present in the morning hours. Thus it
appears likely that the photolysis of HONO can provide a reasonable
boost to the primary HO-radical reactant which in consort with CO and
hydrocarbons drives the NO to N02 and generates the products of smog.
Once this initial burst of HO-production has begun, the generation of
HONO through the chain transfer reaction k6 will likely be the major
source of this species and will maintain a significant level of HONO as
the smog development occurs.
H02 + N02 -» HONO + 02 (^6)
In view of this work it is recommended that those atmospheric scientists
concerned with the chemistry within stack plumes, include the reactions
1 and 2 in their considerations and simulations; it appears to be an
important step in the triggering of the chemical changes which occur.
The present theory appears to offer a more reasonable mechanism for
HO-radical generation in power plant plumes than the sunlight photolysis
of ozone:
03 + hv(A < 3100 A) -* O^D) + 02 (55)
O^D) + H20 -» 2HO (56)
Ozone levels must remain quite low until considerable conversions of NO
to N02 have occurred; [03] = 0.02 [N02]/[NO] ppm for a solar zenith
angle of kO° . At [N02]/[NO] ratios near unity the rates of 0(1D) and
HO-radical formation from the sequence 55 and 56 will be very much
lower than the HO-radical formation rate from HONO photolysis.
76
-------
REFERENCES
R. A. McClatchey, W. S. Benedict, S. A. dough, D. E. Burch,
R. F. Calfee, K. Fox, L. S. Rothman, and J. S. Caring, "AFCRL
Atmospheric Absorption Line Parameters Compilation," AFCRL-TR-73-
0096, Air Force Cambridge Research Laboratories Environmental
Research Papers, No. k^h (1973).
2. J. U. White, J Opt. Soc. Amer., §3, 285
3. Radiation in the Atmosphere, K. Ya, Kondratzev, Academic Press,
New York (1969) p. 225.
h. "Spectroscopic Methods for Air Pollution Measurement," P. L. Hanst,
in Advances in Environmental Science and Technology, Vol II,
J. N. Pitts and R. L. Metcalf, (Eds.) (Wiley, New York, 1971)
P. 91.
5. F. H. Verhoek and F. Daniels, J. Am. Chem. Soc., 53, 1250 (1931).
6. J. C. Fontanella, A. Girard, L. Gramont, and N. Louishard, Appl.
Opt. lU, 825 (1975).
7. S. Hurlock, K. Narahari Rao, L. A. Weller, and P. K. L. Yin,
J. Mol. Spectrosc. U8, 372 (1973).
8. A. J Vosper, J. Chem. Soc. (A), 1970, 625.
9. L. Harris and K. L. Churney, J. Chem. Phys . *£7, 1703 (1967).
10. M. G. Dunn, K. Wark, Jr., and J. T. Agnew, J. Chem. Phys. 37,
(1962).
11. For a general review and evaluation of the potential reactions in
photochemical smog formation and references to the earlier litera-
ture see: a) K. L. Demerjian, J. A. Kerr, and J. G. Calvert, Adv.
Environ . Sci. Technol. , U, 1 (197*0; b) H. Niki, E. E. Daby, and
B. Weinstock, Chapter 2 in Photochemical Smog and Ozone Reactions,
F. F. Gould, (Ed.) Advances in Chemistry Series, 113, Amer. Chem.
Soc., Washington, D.C., 1972, p. 16.
12. J. G. Calvert and R. D. McQuigg, Int. J. Chem. Kinet., in press.
13. L. G. Wayne and D. M Yost, J. Chem. Phys., 19, Ul (1951).
lU. P. A. Leighton, Photochemistry of Air Pollution, Academic Press,
N.Y., I960.
15. R. F. Graham and B. J. Tyler, J. Chem. Soc., Faraday I, 68, 683
(1972).
77
-------
16. C. England and W. H. Corcoran, Ind. Eng. Chem. Fundam., lU, 55
(1975).
17. P. L. Asquith and B. J. Tyler, Chem. Comm., 7^ (1970).
18. T. Nash, Ann. Occup. Hyg., 11, 235 (1968).
19. a) R. A. Cox, J. Photochem., 3, 175 (197*0; b) R. A. Cox, J.
Photochem., 3, 291 (197U/75); c) D. F. Atkins and R. A. Cox,
Atm. Environ.,
20. D. M. Waldorf and A. L. Babb, J. Chem. Phys., 39, ^32 (1963);
kO, U65 (196*0.
21. P. G. Ashmore and B. J. Tyler, J. Chem. Soc., (l96l) 1017.
22. R. T. Hall and G. C. Pimentel, J. Chem. Phys., 38, 1889 (1963).
23. I. R. Beattie and S. W. Bell, J. Chem. Soc., (1957) l68l.
2k. I. C. Hisatsune, J. Phys. Chem., 65, 221+9 (1961).
25. JANAF Thermodynamic Tables, Second Edition, U. S. Dept. Commerce,
N.B.S., 1971.
26. D. R. Hirschbach, H. S. Johnston, K. S. Pitzer, and R. E. Powell,
J. Chem. Phys., 25, 736 (1956).
27. D. L. Baulch, D. D. Drysdale, and D. G. Home, "High Temperature
Reaction Rate Data, No. 5"> Dept. Physical Chemistry, Leeds Univ.,
England, July, 1970, p. U2.
28. Ref. 27, p. U5.
78
-------
APPENDIX A
OPTICAL ALIGNMENT PROCEDURES
The initial alignment of the optical system can be made by using a
helium neon laser. The final adjustment involves the proper spacing of
the mirrors and requires a source which gives a divergent beam and pro-
duces extended images. For preliminary adjustments, a flashlight bulb
can be used but the final adjustments require a more precisely located
source, S, e.g., a hole in a piece of cardboard illuminated by a light
source such as a flashlight. It is assumed that such a source is used
in the following discussion.
TWENTY-ONE METER ABSORPTION CELL
The center of curvature of Mp (Fig. l) is arranged to lie between DL
and DS, on the line between their centers, and in the plane of T>i and
D2. This occurs when the light reflected from Mp produces an image
coincident with the source,S. The center of curvature of DI and D2 are
arranged to lie in the plane of Mp and on a horizontal line through the
center of Mp by placing the source at Mp and examining the image formed
in the plane Mp after reflection by each of Dj. and D2. The mirror DI
is then rotated about a vertical axis until the first image formed on Mp
is immediately below the cut-out in Mp (see Fig. 2). The number of
images of the cell is controlled by turning D2 about a vertical axis.
SIX METER CELL
A similar alignment procedure is used for this cell. When this system
is aligned the centers of curvature of the mirrors Ma, M^, Mc and M^
lie at the center of the mirror system D; the center of curvature of DI
falls on a horizontal line through the center of MC; the centers of
curvature of D2 and D* lie between MJ-, and Mc, and the center of curva-
ture of T>3 lies in M^,.
TRANSFER SYSTEM FROM THE INTERFEROMETER
TO THE ABSORPTION CELL
After the cell has been aligned as described above, light from the
source S is introduced into the exit window of the absorption cell and
an image of the source is produced in the plane of M-p. The mirrors M3,
M7, and M5 in Fig. lU are adjusted to reflect light emerging from the
cell into the center of the exit window to the interferometer. The
light is reflected from the beam splitter to the movable mirror M^ in
Fig. 13 and, by adjusting the entire interferometer, the light can be
returned by M5 along the incoming path. This is checked by making the
79
-------
images N2 (in Fig. lU), formed by the forward and returning light,
coincident. When this occurs the light falling on 145 is collimated and
strikes M5 normally.
INTERFEROMETER AND SOURCE OPTICS
The alignment procedure described previously locates the interfer-
ometer with respect to the absorption cell and ensures that radiation
leaving the movable mirror of the interferometer normally will enter the
cell in the correct direction. In order to check that the source optics
produces a collimated beam which strikes the movable mirror M5 normally,
it is necessary to remove the beam splitter and adjust Ma and MS of
Fig. lU so that the light from this mirror produces an image of the
source concurrent with the image already formed on Ma- When this is
accomplished the light from Na is collimated and strikes M^ normally.
The beam splitter is replaced and the alignment procedure is re-
checked by adjusting the beam splitter alone. The final adjustment
consists of adjusting the fixed mirror of the interferometer to obtain
the best interferogram.
80
-------
APPENDIX B
FOURIER TRANSFORM SPECTROMETER
DESCRIPTION OF OPERATION
Spectral information is obtained from the IFTS by sampling the output
from the infrared detector as the mirror M5 moves from a position near
zero retardation through the chosen distance d. By Eq. (lU) this dis-
tance depends on the desired spectral resolution which can be varied
from 0.125 to 16 cm"1. This sampling is controlled by the output from
a second detector which monitors fringes produced by a helium neon laser
source whose output passes through a second interferometer which in-
cludes a mirror mounted to the rear of Mg. The signal from the infrared
detector is digitized and stored in the memory of the processing system.
By co-adding signals obtained from several scans of M5 the signal-to-
noise ratio can be enhanced. Most of the spectra described in this
report were obtained from 20 scans of M5.
The data collection is performed under the control of an on-line
minicomputer (Nova 1200) which receives instructions via a teletype
machine. A 128k fixed head disc not only provides temporary data stor-
age but also provides storage for the Fourier transform programs. The
stored interferogram can be transferred to magnetic tape for permanent
storage or processed by means of the computer program to give a spec-
trum. Either all or selected portions of this spectrum can be displayed
on a digital plotter or stored on magnetic tape. Additional interfer-
ograms may then be obtained and processed. Other programs, stored in
the disc memory unit, allow further manipulation of the spectra.
One of the most commonly used programs allows two spectra to be
"ratioed". In this program the output signal of one spectrum is divided
by the signal from the second spectrum at each spectral frequency
sampled. For example, by ratioing the spectrum of a gas sample in the
absorption cell with a similar spectrum of the evacuated cell a direct
measurement of the gas sample transmittance is obtained with the spec-
tral variations of the detector sensitivity, source emittance, and
optical system transmittance removed.
UNDERSAMPLING TO OBTAIN HIGH RESOLUTION
The interferogram is obtained by sampling the output from the infrared
detector as the mirror MS moves through a distance d. This distance is
determined by the required spectral resolution Av where
Av (cm'1) = l/2d (cm) = 1/x (57)
The infrared detector output is sampled at mirror positions separated
by a distance AX. This distance is controlled by the reference laser
interferometer and is given by
81
-------
AX = mAL = mvL
where v-^ = 15,800 cm"1 and X^ = 0.633 nm are the laser wavenumber and
wavelength, respectively, and in is an integer.
The number of sample points in the interferogram
n = x/Ax = vL/mAv (59)
Thus, if m = 1 and Av = 0.125 cm"1, n » 126,UOO. This is considerably
larger than the maximum number of points, 32,000, which can be trans-
formed into a spectrum with the IFTS system available.
It is, however, possible to collect interferograms with n < 32,000 from
which spectra with a resolution of 0.125 cm"1 can be obtained provided
m > 1. This process, known as undersampling, cannot provide useful
information unless the signal from the infrared detector is limited to
information in the spectral interval
°v = vmax - vmin
Upper and lower bounds to the spectral interval sampled are required
since, when an interferogram consisting of information sampled at dis-
crete points is transformed the resulting spectrum is repeated at inter-
vals 5v such that
6v = 1/2AX (61)
where AX = sampling interval of the interferogram.
If either 6v or Ax is improperly chosen, information from more than one
spectral region can be superposed in the resulting spectrum obtained
from the interferogram; this effect is known as folding.
From Eqs. (57), (58), and (6l) it follows that
n = ^ (62)
Av
and, hence, for the present IFTS system
^ - 15798.0112 (63)
Av
The entire spectral region from 0 to 8000 cm"1 (the high frequency cut-
off of the germanium beam splitter) can be observed provided
Av > 0.5 on'1. However, to obtain a spectrum with Av = 0.125 cm"1 then
6v < 197^.751^ cm"1 (6U)
82
-------
and, since in this case the interferogram must be sampled over the
entire U cm range of the mirror travel it follows from Eq. (59) that
m > U .
It is possible to select the spectral region over which high spectral
resolution is obtained and, at the same time, avoid folding by choosing
vmax anc* vmin> as defined by Eq. (60), such that
v - h 6v
(65)
= h (vmax - vmin)
where h is an integer.
The limiting values of vmax and vm^n depend on the values chosen for
m, h, and Av. A sampling of these limits is given in Table 5 for an
instrument using a helium neon laser to control the sampling interval
Ax and for the case where Av = 0.125 cm"1. The first three columns give
the undersampling parameter m, the sampling interval AX = mA^, and the
number of points in the interferogram at which data are collected,
respectively. The succeeding columns give the values of vmj_n and vmax
for different values of h.
It is seen that, m > H for n < 32,000.
In order to make use of this instrument capability we have obtained
three optical filters whose characteristics are shown in Table 6.
Although the transmittances of the filters are not completely matched
to the IFTS system it is possible to cover almost the entire spectral
region from UOO to 3800 cm"1 by obtaining no more than three interfer-
ograms. The number of interferograms required could be reduced by
increasing the memory capability of the IFTS system to allow the value
of n to exceed 32,000.
83
-------
Table 5- IFTS Parameters to Achieve 0.125 cm"1 Resolution
Sampling
Undersampling Interval
Factor m Ax (urn)
1 .6329911
2 1.265982
3 1.898973
U 2.53196U
5 3.16U955
6 3.7979^6
7 H. U30937
8 5.063929
No. of
Points
Sampled
12638U 0
63192 o
1*2128 0
31596 0
25277 o
2106U 0
18055 o
15798 o
Spectral Limits for Various Values of h
h = 1
. 7899
39^9
2633
197U
1579
1316
1128
987
.0056
.5028
.0019
.751U
.8011
.5009
.U29U
.3757
7899
39^9
2633
197^
1579
1316
1128
. 987
h = 2
.0056
15798
.5028
7899
.0019
5266
39^9
.8011
3159
.5009
2633
2256
.3757
197U
.0112
.0056
.0037
.5028
.6022
.0019
.8587
.751^
15798
7899
5266
59^9
3159
2633
2256
197U
(cm"1)
h = 3
.0112
23697
.0056
118U8
.0037
7899
.5028
592U
.6022
U739
.0019
39^9
.8587
3385
2962
.0168
.508U
.0056
.251*2
.1*031+
.5028
.2881
.1271
Too many points on collection for m < U for processing by the present IFTS system.
-------
Table 6. Characteristics of Optical Filters and Spectral Limits to Achieve 0,125 cm'1 Resolution
oo
VJJ
Interferometer Param ters
Filter Transmittance
Limits (cm"1)
vmin vmax
0 lUOO
(< Uoo)
1150 2600
2200 H150
Under sampling
Parameter
m
5
5
6
7
k
h
1
2
2
2
2
Allowable
Limits
vmin
0
1579.8011
1316.5009
1128. U29H
197^.751^
Spectral
(cm'1)
vmax
1579.8011
3159.6022
2633.0019
2256.8587
39^9-5028
Usable Spectral
Region (cm"1)
vmin
< Uoo
~ 2000
~ 1500
~ 1150
- 2200
vmax
lUOO
- 2600
~ 2600
- 1800
~ 3800
Detector
Cu:Ge
Cu:Ge InSb
Cu:Ge InSb
Cu:Ge
Cu:Ge InSb
-------
APPENDIX C
UNITS FOR MEASURING GAS CONCENTRATIONS
In this work concentrations are typically reported in ppm by volume at
a total gas pressure of 760 mmHg. Thus, the partial pressure P, exerted
by a gas at a concentration of g ppm
P (mmHg) = 7.6 x 10~4g mmHg.
If the optical path length is L (cm) the total amount of gas in the
path
w = g L ppm-cm
Another commonly used unit to measure the amount of gas in the absorbing
path is the atmosphere-cm (atm-cm). If the gas at partial pressure P
(mmHg) and path length L (cm) were brought to atmospheric pressure the
amount of gas present is
u = P-L/760 atm-cm .
The amount of gas u is dependent on the temperature of the cell T (K).
It is customary to reduce this value to UQ, at S.T.P. by assuming the
gas behaves as a perfect gas
u0 = u'273 atm-cm (S.T.P.)
The number of molecules/N per cm2 area of the optical beam traversing
a cell containing UQ atm-cm
N = Lu0
where L = LoschmidtB1 number (2.68? x 1019 atm"1cm"3).
The relations between these units are shown in Table 7-
87
-------
Table 7. Relations Between Units for Amount of Absorbing Gas in a Given Path Length
oo
CO
1 ppm-cm
1 atm-cm
(atm-cm)STp
mol/cm^
gL
ppm- cm
1
106
273 x 10s /T
1.02 x 10'11
u
atm-cm
io-s
1
273/T
1.02 x 10~17
UO
(atm-cm)STp
10-e T/273
T/273
1
3.722 x 10- 20
N
mol/cm?
9.8U x 1010 T
9.8U x 1016 T
2.687 x 1019
1
-------
TECHNICAL REPORT DATA
(Please read /HStnictions on the reverse before completing)
1. REPORT NO.
EPA-600/3-76-084
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
SPECTROSCOPIC STUDIES OF PHOTOCHEMICAL SMOG
FORMATION AND TRACE POLLUTANT DETECTION
5. REPORT DATE
July 1976
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Jack G. Calvert, Walter H. Chan,
Robert J. Nordstrom, and John H. Shaw
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
The Ohio State University Research Foundation
1314 Kinnear Rd.
Columbus, Ohio 43212
10. PROGRAM ELEMENT NO.
1AA008
11. CONTRACT/GRANT NO.
R-803075
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final 74/75
14. SPONSORING AGENCY CODE
EPA-ORD
15. SUPPLEMENTARY NOTES
16. ABSTRACT
An infrared Fourier transform spectrometer has been used with a long path length,
multiple traversal cell to study the infrared spectra of atmospheric gases and
several pollutant gases. Solar spectra have also been obtained between 3 and
20 urn wavelength.
The kinetics of the formation and decay of nitrous acid have been followed by
monitoring bands of nitric oxide, nitrogen dioxide, and both cis- and trans-
nitrous acid. Rate constants and the equilibrium constant for the reactions
have been derived. A mechanism accounting for the formation of nitrous acid
in the atmosphere is proposed. These rate data have been used to speculate
on the potential importance of nitrous acid formation in power plant and auto
exhaust plumes.
A new technique involving the use of infrared spectroscopy and two cells of
unequal length to study a two-component system in equilibrium is described.
A six-meter multiple traversal cell equipped with fluorescent tubes has been
constructed. The cell irradiance closely simulates the solar irradiance at
ground level in spectral distribution and intensity.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air pollution
Infrared spectroscopy
Optical equipment
Solar spectrum
Nitrous acid
Reaction kinetics
Photochemical reactions
Fourier Transform
Spectrometer
13B
14B
20F
03B
07B
07D
07E
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNri.ASSTFTF.D
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
21. NO. OF PAGES
99
22. PRICE
EPA Form 2220-1 (9-73)
89
------- |