Fourier Transform Measurement of NO2 Absorption Cross-
sections in the Visible Range at Room Temperature.
A.C. Vandaele,, C. Hermans, P.C. Simon, M. Van Roozendael
J.M. Guilmot, M. Carleer, and R. Colin
Abstract
New laboratory measurement of NO
2
absorption cross-sections were performed
using a Fourier transform spectrometer at 2 and 16 cm
-1
( 0.03 and 0.26 nm at 400
nm) in the visible range (380-830 nm) and at room temperature. The use of a
Fourier transform spectrometer leads to a very accurate wavenumber scale (0.005
cm
-1
, 8x10-5 nm at 400 nm). The uncertainty on the new measurements is better
than 4%. Absolute and differential cross-sections are compared with published
data, giving an agreement ranging from 2% to 5% for the absolute values. The
discrepancies in the differential cross-section can however reach 18%. The
influence of the cross-sections on the ground-based measurement of the
stratospheric NO
2
total amount is also investigated.
Key words
Fourier Transform Spectroscopy, NO
2
, absorption cross-sections, differential
absorption cross-sections, visible, stratospheric and tropospheric measurements
Introduction
Nitrogen dioxide plays an important role in the chemistry of the troposphere and
the stratosphere. It is produced from the oxidation of NO in the troposphere, where
it acts as the main source of tropospheric ozone, and is a precursor to species,
such as nitric acid, which play a role in the acidification of the environment. Its role
in stratospheric photochemistry has been pointed out by Crutzen (1970). Beside its
catalytic interaction in the control of ozone, it regulates the amounts of ClO, which
in turn controls the ozone loss due to the chlorine catalytic cycle, and of ClONO
2
,
which is an important stratospheric reservoir of chlorine. It plays thus an important
role in the coupling of the NO
x
and ClO
x
families.
Molecular absorption in the UV-Visible region has been widely used to measure
the concentrations of gases in the atmosphere, either in the troposphere or in the
stratosphere. The instruments used range from ground-based spectrometers
measuring tropospheric or stratospheric concentration by the Differential Optical
Absorption Spectroscopy (DOAS) technique ( see for example Platt and Perner,
1980, Solomon et al., 1987, Edner et al., 1993, Vandaele et al.,1992, Evangelisti et
al., 1995, Camy-Peyret et al., 1996), to spectrometers on board satellites such as
the GOME ( Global Ozone Monitoring Experiment) launched in April 1995 on
board ERS-2 satellite, and to SCIAMACHY and GOMOS instruments to be
launched in 1999 on board ENVISAT-1.
All these instruments require absorption cross-sections of the observed
molecules, measured at a resolution of 0.02 nm or better and with an
accuracy better than 5% (Chance et al, 1990). Accurate cross-sections are
also needed for the chemical-dynamical-radioactive modeling of the
atmosphere.
The measurement of the NO
2
absorption cross-section is complicated by the
presence of its dimer N
2
O
4
. Several studies have attempted to measure NO
2
cross-sections. Hall and Blacet(1952) measured absorption spectra of NO
2
-
N
2
O
4
mixtures at 298K and deduced the contribution of N
2
O
4
. Johnston and
Graham (1974) measured NO
2
cross-sections in the 185-420 nm spectral
region at room temperature. Bass et al.(1976) investigated the 185-410 nm
range at 298 K. They corrected their data for the presence of N
2
O
4
. Leroy et
al.(1987) reported values from 427 to 450 at 298K. Schneider et al.(1987)
obtained NO2 absorption cross-sections between 200 nm and 700 nm, at
298 K and determined the absorption cross-sections of N2O4 between 200
and 255 nm. Koffend et al.(1987) used a pulsed dye laser to perform high
resolution measurement of NO
2
absorption structures in the 392-395 nm and
411-414 nm. Davidson et al.(1988) investigated the dependence of the NO
2
cross-sections on temperature and the influence of this dependence on the
determination of the photolysis rate of NO
2
in the atmosphere. Harwood and
Jones (1994) studied the temperature dependence of the ultraviolet-visible
absorption cross section of NO
2
. The cross sections of N
2
O
4
. were also
derived by the latter, as well as new values for the equilibrium constant.
Mérienne et al.(1995) measured NO
2
absorption cross-sections in the 300-
500 nm region at 293K. The use of an absorption path length of 61m allowed
them to work at very low pressure (<0.04 Torr) and to minimize the influence
of the dimer. Discrepancies of the order of 20% or sometimes more are
found between all the measurements.
This work presents new absolute absorption cross-sections of NO
2
between
12000 and 26000 cm
-1
(380-830 nm) at 294K. The absorption cross-sections
have been measured at two resolutions : 2 and 16 cm
-1
( 0.03 and 0.26 nm at
400 nm). Cross-sections obtained in this work have been measured with a
Fourier Transform Spectrometer (FTS), which combines the advantages of a
great sensitivity and a built-in wavenumber calibration. This last advantage is
achieved by the presence of a He-Ne laser, which allows the interferogram to
be digitised at equal intervals. This leads to a highly accurate and
reproducible wavenumber calibration. With the Fourier Transform
Spectrometer described below, an accuracy of about 0.005 cm
-1
( 8x10-5 nm
at 400 nm ) on the wavenumbers is achieved.
These new cross-sections will then be compared with published data sets,
measured at similar resolutions. A comparison of cross-sections smoothed to
a resolution of 1.0 nm will also be shown. Finally, the importance to dispose
of cross-sections of high quality will be stressed, by investigating their
influence on ground-based stratospheric NO
2
measurements in the visible
range.
Experimental details
The experimental set-up consists of a light source, the spectrometer, and the
gas handling system. The choice of the light source, beam splitter and
detector is determined by the spectral region to be investigated.
Combinations of a high pressure O
3
free Xenon source (USHIO, 450 W), a
Tungsten filament lamp, a Quartz Suprasil beam splitter with a Si diode or a
UV vacuum solar blind diode have been used in order to cover the entire
spectral region from 12000 to 26000 cm
-1
. A blue filter provided by BRUKER
or a copper sulphate solution filter have been used. The experimental
conditions are reported in Table 1. A photo feedback system allowed the
lamp intensity to be stabilised within 1% throughout the experiments.
The absorption spectra were recorded using either a BRUKER IFS120HR or
a BRUKER 120M Fourier Transform Spectrometer. The theoretical resolving
power attainable with these spectrometers is 106. However, the resolution is
essentially limited in the UV by the quality of the mirrors and beam splitter,
leading to a maximum resolution of the order of 10-2 cm
-1
in this region.
Following some recommendations concerning the measurements of
absorption cross-sections (Chance, 1990), spectra have been recorded with
resolutions of 2 and 16 cm
-1
, corresponding respectively to 0.03 and 0.26 nm
at 400 nm.
NO2 gas (UCAR, stated purity of 99.5%) used without further purification,
was introduced in a 5.15 cm absorption cell located in the sample
compartment of the Fourier Spectrometer. The partial pressure of the gas
was then monitored with a 100 Torr full scale Baratron gauge. We waited
until the stabilisation of the partial pressure (10 to 30 min, without any lamp
illumination ) before filling with oxygen in order to obtain a total pressure of 1
atm. The presence of oxygen is of great importance as it induces the
reconversion of any NO into NO
2
, maintaining the NO
2
concentration
constant. NO can be present in the cell as an initial impurity or produced by
the photo dissociation of NO
2
during the experiment. The cell has been
described in an earlier publication (Hurtmans et al., 1993). Its characteristics
will be briefly summarised here : the cell is made of anodised aluminum and
has quartz windows; the temperature of the gas is monitored inside the cell
with a temperature transducer characterised by an accuracy of 0.2K in the
temperature range used in the present work.
The temperature in the cell was stabilised with a liquid circulating around the
cell. Finally, the whole experimental set-up was located in a temperature
stabilised room.
Absorption of NO
2
on the inner surface of the cell is inevitable. To limit this
effect on the pressure measurement, we waited before and after filling the
cell with oxygen. This is however not enough to guarantee a stable NO
2
partial pressure throughout the experiment. Each spectrum was therefore the
average of a number of scans (see Table I) which have been recorded by
sequential blocks of either 256 or 1024 scans. Only the blocks of scans
whose absorption did not differ from the first block by more than 1% were
retained. We thus checked that the gradual decrease of the NO
2
pressure
inside the cell was small. Moreover it proved that photochemical degradation
of the sample due to the irradiation by the lamp was not important during the
experiment.
Spectra have been obtained using a double sided recording mode, during the
forward movement only of the mobile mirror and no apodization function was
used. Blank spectra, i.e. with an empty cell, were recorded before and after
each measurement.
Determination of the absorption cross-sections
Absorption cross-sections s(l) are derived from the experimental data using
the Beer-Lambert law:
I(l) = Io(l) exp( - n l s(l) )
where n is the gas concentration in the cell,
l the absorption path length,
Io(l) and I(l) the intensities of the signal with an empty cell and a filled cell.
The NO
2
Û N
2
O
4
equilibrium implies that N
2
O
4
is always present in the cell.
Absorption of N
2
O
4
occurs at wavenumbers greater than 25000 cm
-1
(Hall
and Blacet, 1952, Schneider et al.,1987), but not in the spectral region
investigated in this work. N
2
O
4
should not therefore interfere with the
absorption structures of NO
2
studied here.
However the presence of N
2
O
4
must be considered when determining the
partial pressure of NO
2
. This was done by considering the following
equations :
where P
t
is the total partial pressure in the cell and the partial pressures
and are related through the equilibrium constant K
P
:
where K
P
is the equilibrium constant.
Partial pressures of NO
2
and N
2
O
4
can thus be calculated. Hurtmans et al.
(1993) reviewed the values of the equilibrium constant found in the literature
and determined an empirical relation for its temperature dependence. For
temperatures ranging from 233 K to 403 K, the dependence was expressed
by a fourth degree polynomial expansion. At 294.15K, the value of the
constant KP is 105.72 hPa. Using the above equations the partial pressure of
NO2 can be calculated from the various total pressures used.
The NO
2
absorption cross-sections at the resolution of 16 cm
-1
were obtained
by taking the mean value of all the spectra i.e. all the spectra taken at
16 cm
-1
and the spectrum at 2 cm
-1
degraded to 16 cm
-1
. All measurements
were found to agree within 2% which is lower than the uncertainty stated in
Table II. The results are plotted in Figure 1 for the entire spectral region.
Figure 2 shows a detailed region from 22200 to 23200 cm
-1
of the spectrum
at the resolution of 2 cm
-1
.
Discussion
Error evaluation
The error budget on the absorption cross-sections was carefully evaluated,
taking into account the errors on the pressure and temperature
measurements, on the reaction constant K
P
, the uncertainty on the
absorption path length, the presence of possible impurities in the samples,
the adsorption of NO
2
taking place inside the cell and the absorbance
accuracy.
