5704
J. Chem. Phys., Vol. 115, No. 13, 1 October 2001
E. C. Zipf and S. S. Prasad
͑Ref. 9͒ at 1000 torr for O3. There would therefore be a delay
of nearly 3 min before NOX would be detected at the center
of the cell if the source was the heterogeneous chemistry on
the walls. However, if the source is a volume process, then
the NOX would form immediately along the axis of the pho-
tolysis cell. The initial axial concentration of NOX will then
decrease exponentially with time in a manner consistent with
the well-known continuity equation governing diffusion until
the NOX was uniformly distributed. In the absence of a sur-
face source, the ratio of the initial NOX concentration at the
center of the cell to the final fully mixed NOX concentration
should equal the ratio of the irradiated to the total cell vol-
umes. If there is a surface source in addition the volume
process, the former ratio will deviate from the latter. Thus the
existence of surface chemistry can be verified and it strength
estimated. Using this approach, we confirmed that the NOX
formed immediately and that there was no evidence for sig-
nificant wall production in agreement with the implications
of the linearity tests.
FIG. 1. Plot showing the O2•N2 interaction constant, B(,200 K) and the
O3 absorption cross sections that are important in atmospheric modeling.
The UV flux from the D2 lamp used in the quantum yield measurement is
shown for reference.
pressure when the pressure of the other gas was held con-
stant. The experiments at constant O2 pressure are particu-
larly important because they effectively rule out surface
chemistry as a significant NOX source in this experiment.
The linear and quadratic pressure dependencies of P͑NO͒ can
not be accounted for by the photochemistry of any excited
O2 monomers ͓e.g., O2(A3⌺)͔ that may be created in our
experiment. In contrast, the observed pressure dependencies
are readily explained by the photo-excitation of the O2•N2
cluster regardless of whether it is a bound dimer or a colli-
sion complex. The quantum yield was, therefore, determined
in the quadratic pressure domain ͓c.f. curve C in Ref. 3͔
using the expression
III. EXPERIMENTAL DETAILS AND RESULT FOR THE
ISOPTOPIC COMPOSITION
In addition to nitric oxide significant amounts of O3 were
also formed in the O2•N2 experiment so that the primary NO
molecules from process ͑1͒ were rapidly converted into NO2
via reaction ͑3͒,
NOϩO3→NO2ϩO.
This NO2 exhibits
͑3͒
L
a
large 15N enhancement (␦15N
P͑NO͒ϭ
¯
B ,T͒•n O ͒•n N ͒•⌽ ͒
͑
͑
͑
͑
͵ ͵
2
2
0
0
0
ϳ150 ‰). Mass analysis was done using the Extrel quadru-
pole mass spectrometer ͑QMS͒ that had 1.59 cm diameter
poles and was fitted with a high efficiency electron impact
ionizer. This procedure is essentially the same as those used
by others ͑see Anderson et al.10͒. As explained previously,3
this ionizer had other special design feature so that it can
effectively trap and energy analyze both the primary ions and
their dissociative ionization fragments. These data permit a
self-consistent determination of the ␦15N, ␦17O, and ␦18O
from an analysis of the complete fragmentation pattern. In
the study of the wavelength dependence of the isotope ef-
fects produced when NO2 is photodissociated by UV radia-
tion, the NO and O2 were cryogenically separated from the
NO2 before mass analysis. In order to determine whether the
enhancement took place in step ͑1͒ or ͑3͒, we performed the
following experiment. Ozone was synthesized in the pho-
tolysis cell by irradiating research grade O2 (250 torr) with
Hg 184.9 nm photons from a penlite source. Complemen-
tary 253.7 nm absorption measurements were used to moni-
tor the O3 buildup. When the O3 partial pressure reached ϳ1
torr, the Hg lamp was turned off and the isotope signatures of
the O3 was determined. The ozone was found to be enriched
in 17O and 18O in a mass independent manner (␦17Oϳ␦18O
ϳ80 ‰) in agreement with the findings of Thiemens and
Jackson.11 The isotope analysis used the standard technique12
of converting a small amount of the O3 to O2 pyrolytically. A
5% mixture of NO/N2 with the isotope fractionation pattern
of nitrogen in air was then titrated into the O2 /O3 mixture
•exp ϪB ,T͒•n O ͒•n N ͒͒ץ
ץ
l,
͑2͒
͑
͑
͑
͑
2
2
where n͑N2͒ and n͑O2͒ are the gas densities, B(,T) is the
interaction constant measured by Shardanand5 ͑equivalently,
the binary absorption coefficient in the terminology used in
Refs. 6 and 7͒ and ⌽0() is the total incident UV flux in the
interval ␦. The wavelength-averaged NOX yield was ap-
proximately (3.3Ϯ0.8)ϫ10Ϫ2NO/absorbed photon. This es-
timate is based on the B(,298 K) values published by Shar-
danand. There are no other independent measurements of
this quantity. However, it is worth noting that Shardanand’s
B(,T) values for the O2uO2 cluster are in excellent agree-
ment with those measured by others.7 The formation of O3 in
the photolysis of O2•O2 dimer ͑or collision complex͒ by
Herzberg band radiation observed by Brown and Vaida8 ap-
pears to be a similar photochemical process.
To verify that the NOX production was a volume pro-
cess, a simple test was done in which the UV beam irradiated
the gas for a short time ͑20 s͒. A small capillary tube was
inserted radially into the center of the large photolysis cell.
This tube could be valved off at the cell and was connected
to an Extrel quadrupole mass spectrometer/GC. Because of
the axial geometry of the UV beam, all of the initial photo-
dissociation and chemistry involving NOX and O3 takes in a
small cylindrical volume along the central axis of the cell.
The time for any of the by-products of photolysis to diffuse
to the wall is long. For example, it is approximately 135 s