Investigation of N2O production from 266 and 532 nm laser flash photolysis of O3/N2/O2 mixtures

Article abstract of DOI:10.1021/jp014242c

Tunable diode laser absorption spectroscopy has been employed to measure the amount of N₂O produced from laser flash photolysis of O₃/N₂/O₂ mixtures at 266 and 532 nm. In the 532 nm photolysis experiments very little N₂O is observed, thus allowing an upper limit yield of 7 × 10^-8 to be established for the process O₃? + N₂ → N₂O + O₂, where O₃? is nascent O₃ that is newly formed via O(³P_J) + O₂ recombination (with vibrational excitation near the dissociation energy of O₃). The measured upper limit yield is a factor of ～600 smaller than a previous literature value and is approximately a factor of 10 below the threshold for atmospheric importance. In the 266 nm photolysis experiments, significant N₂O production is observed and the N₂O quantum yield is found to increase linearly with pressure over the range 100-900 Torr in air bath gas. The source of N₂O in the 266 nm photolysis experiments is believed to be the addition reaction O(¹D₂) + N₂ + M ^k₆ → N₂O + M, although reaction of (very short-lived) electronically excited O₃ with N₂ cannot be ruled out by the available data. Assuming that all observed N₂O comes from the O(¹D₂) + N₂ + M reaction, the following expression describes the temperature dependence of k₆ (in its third-order low-pressure limit) that is consistent with the N₂O yield data: k₆ = (2.8 ± 0.1) × 10^-36(T/300)^{-(0.88±0.36)} cm⁶ molecule^-2s^-1, where the uncertainties are 2σ and represent precision only. The accuracy of the reported rate coefficients at the 95% confidence level is estimated to be 30-40% depending on the temperature. Model calculations suggest that gas phase processes initiated by ozone absorption of a UV photon represent about 1.4% of the currently estimated global source strength of atmospheric N₂O. However, these processes could account for a significant fraction of the oxygen mass-independent enrichment observed in atmospheric N₂O, and they appear to be the first suggested photochemical mechanism that is capable of explaining the altitude dependence of the observed mass-independent isotopic signature.

Full text of DOI:10.1021/jp014242c

5880
J. Phys. Chem. A 2002, 106, 5880-5890
Investigation of N2O Production from 266 and 532 nm Laser Flash Photolysis of O3/N2/O2
Mixtures
E. G. Estupin˜a´n,† J. M. Nicovich,‡ J. Li,† D. M. Cunnold,† and P. H. Wine*,†,‡
School of Earth and Atmospheric Sciences and School of Chemistry and Biochemistry,
Georgia Institute of Technology, Atlanta, Georgia 30332
ReceiVed: NoVember 15, 2001; In Final Form: April 1, 2002
Tunable diode laser absorption spectroscopy has been employed to measure the amount of N2O produced
from laser flash photolysis of O3/N2/O2 mixtures at 266 and 532 nm. In the 532 nm photolysis experiments
very little N2O is observed, thus allowing an upper limit yield of 7 × 10-8 to be established for the process
†
†
O3 + N2 f N2O + O2, where O3 is nascent O3 that is newly formed via O(3PJ) + O2 recombination (with
vibrational excitation near the dissociation energy of O3). The measured upper limit yield is a factor of ∼600
smaller than a previous literature value and is approximately a factor of 10 below the threshold for atmospheric
importance. In the 266 nm photolysis experiments, significant N2O production is observed and the N2O quantum
yield is found to increase linearly with pressure over the range 100-900 Torr in air bath gas. The source of
k6
N2O in the 266 nm photolysis experiments is believed to be the addition reaction O(1D2) + N2 + M
98 N2O
+ M, although reaction of (very short-lived) electronically excited O3 with N2 cannot be ruled out by the
available data. Assuming that all observed N2O comes from the O(1D2) + N2 + M reaction, the following
expression describes the temperature dependence of k6 (in its third-order low-pressure limit) that is consistent
with the N2O yield data: k6 ) (2.8 ( 0.1) × 10-36(T/300)-(0.88(0.36) cm6 molecule-2 s-1, where the uncertainties
are 2σ and represent precision only. The accuracy of the reported rate coefficients at the 95% confidence
level is estimated to be 30-40% depending on the temperature. Model calculations suggest that gas phase
processes initiated by ozone absorption of a UV photon represent about 1.4% of the currently estimated
global source strength of atmospheric N2O. However, these processes could account for a significant fraction
of the oxygen mass-independent enrichment observed in atmospheric N2O, and they appear to be the first
suggested photochemical mechanism that is capable of explaining the altitude dependence of the observed
mass-independent isotopic signature.
1. Introduction
2a and about 60% via channel 2b.4 Reaction 2b represents the
dominant source of total reactive nitrogen (NOy) in the
Nitrous oxide (N2O) is a climatically important species
affecting the Earth’s radiation budget. Its contribution to the
greenhouse effect is considerable due to its long residence time
of 120 ( 30 years1 and its relatively large energy absorption
capacity per molecule.2 On a per molecule basis, N2O is
estimated to be 296 times more powerful than CO2 as a
greenhouse gas (based on 100 year global warming potentials).3
Nitrous oxide is inert in the troposphere. However, in the
stratosphere, especially in the middle and upper stratosphere,
N2O is destroyed by photolysis (λ ∼ 180-215 nm):
stratosphere.5,6
The global atmospheric N2O budget is the least well-
constrained of the major greenhouse gas budgets. Well-
documented major sources all introduce N2O into the atmo-
sphere near the Earth’s surface, and include soils under natural
vegetation, oceans, agricultural activities, combustion, and
biomass burning.3 Even though considerable progress has been
made in recent years in the identification of new sources, the
uncertainties associated with individual sources have not been
reduced (i.e., (50% or larger). Recent estimates of the global
N2O source strength range from 7 to 37 Tg of N2O per year,3,7-9
with a “best guess” value of ∼16 Tg of N2O per year. The
possible existence of in situ atmospheric sources of N2O is still
a controversial subject, although laboratory studies are reported
in the literature that provide evidence for the existence of such
sources.10-12 In addition, recent isotopic studies have cast doubts
on the current understanding of the global N2O budget. One
important aspect that these studies have revealed is that
atmospheric N2O samples show a mass-independent heavy
oxygen isotope effect, i.e., an anomalous ratio of N217O-to-N218O
concentration ratio, which increases with altitude (or distance
from known sources).13,14 Calculations by Miller and Yung15,16
predict that N2O UV photolysis in the stratosphere selectively
destroys “light N2O”, thus leaving behind N2O that is enriched
in the heavier isotopes of both N and O. Recent laboratory
N2O + hV f N2 + O(1D2)
(1)
Process 1 accounts for ∼90% of photochemical N2O destruc-
tion.1 The other 10% of N2O photochemical loss is via reaction
with O(1D2):4
N2O + O(1D2) f N2 + O2
f NO + NO
(2a)
(2b)
About 40% of the N2O + O(1D2) reaction proceeds via channel
* To whom correspondence should be addressed at School of Chemistry
and Biochemistry. E-mail: pw7@prism.gatech.edu.
† School of Earth and Atmospheric Sciences.
‡ School of Chemistry and Biochemistry.
10.1021/jp014242c CCC: $22.00 © 2002 American Chemical Society
Published on Web 05/22/2002

experiments17-21 support the predictions of Miller and Yung,
and recent field observations21-23 are also consistent with the
magnitude of the enrichments in 14N14N18O, 14N15N16O, and
15N14N16O predicted by the Miller and Yung theory. However,
N2O photolysis is a mass-dependent process14,16 and, therefore,
does not account for the mass-independent fractionation in N2O.
The research described in this paper was initiated to evaluate
the possible formation of N2O from the interaction of nascent
(highly vibrationally excited) O3 (O3†) with N2 in N2 + O2 buffer
gas:
As described by Cliff and Thiemens,13 O3 is an ideal candidate
as the source of mass-independently isotopically enriched N2O
in the atmosphere. Laboratory studies of reaction 3 have shown
that O3 formation is accompanied by mass-independent enrich-
ment, i.e., δ17O ≈ δ18O ≈ 85‰.24 In addition, measurements
of both stratospheric O325,26 and tropospheric O327,28 have shown
extraordinarily large O217O excesses (compared to the mass-
dependent O217O/O218O ratio), ∆17O, ranging from about 22‰
to 35‰.26-28 Reaction 4b would result in a direct transfer of
mass-independently enriched oxygen to N2O, and thus, it could
provide a possible explanation for the observed atmospheric
isotopic signature of N2O.13,14
Our study of reaction 4b has also been motivated by the work
of Zipf and Prasad12 and Prasad and Zipf,29 who report an N2O
yield, k4b[N2]/{(k4a + k4b)[N2] + k5[O2]}, of 4 × 10-5 in air at
pressures of 1-

5882 J. Phys. Chem. A, Vol. 106, No. 24, 2002
Estupin˜a´n et al.
The number of O(3PJ) atoms generated per laser pulse was
calculated from the incoming laser power (corrected for window
losses), the fraction of laser photons absorbed by the O3
molecules in the reaction cell, and the estimated quantum yield
of O(3PJ) from the chemistry taking place in the reaction cell
(i.e., 2.0 ( 0.3). The fraction of laser photons absorbed by the
O3 molecules was computed from the measured O3 concentra-
tion, the laser path length through the photolysis cell, and the
known room-temperature absorption cross section of O3 at 266
nm (i.e., 9.49 × 10-18 cm2 molecule-1).39 Measurement of the
incoming and exiting radiation (“dual beam” absorption mea-
surement) permitted a second, more direct, evaluation of the
fraction of laser photons absorbed by the O3 molecules in the
reaction cell.
Nitrous oxide was monitored at 2207 cm-1 using highly
monochromatic infrared radiation from a lead salt tunable diode
laser housed in a liquid nitrogen Dewar. The pressure in the
photolysis cell, the Hg lamp intensity, and the ratio of the two
digitized first harmonic signals from the IR detection system
were all digitized and fed into an MS-DOS compatible
microcomputer, where information was stored for later analysis.
Figure 1. Plot of number of N2O molecules detected per O(3PJ) atom
generated (left axis) and number of N2O molecules detected per O(1D2)
atom generated (right axis) as a function of air pressure. Filled circles
indicate experiments performed in a ∼580 cm3 volume reaction cell,
and empty circles indicate experiments performed in a ∼95 cm3 volume
reaction cell.
2.2. 532 nm Irradiation Experiments. The experimental
system used was similar to the one employed in the preliminary
266 nm experiments. Only the differences are discussed below.
3. Results and Discussion
3.1. Preliminary Experiments. All the experiments were
carried out under static-fill conditions. The first set of experi-
ments was designed to test the possibility that N2O could be
generated by background sources (“blank” experiments). Three
1 h irradiations at 266 nm were performed at 100, 500, and
900 Torr total pressure of air (no O3 present). In addition, two
O3/air mixtures (i.e., 0.1 and 0.3 Torr of O3) at total pressures
of 500 and 900 Torr, respectively, were allowed to remain in
the photolysis cell for periods of 1 h (i.e., typical irradiation
times) with the Hg lamp on but the laser blocked. In all five
cases, when the photolysis products were expanded into the
infrared cell, negligible, if any, N2O was detected.
The second set of experiments was designed to determine
the number of N2O molecules produced per O(3PJ) atom
generated. Figure 1 shows the results of these experiments.
Readily detectable N2O yields were measured, and the yields
were observed to increase linearly with pressure. However, the
N2O yields depicted in Figure 1 are about 2 orders of magnitude
lower than the N2O yield reported by Zipf and Prasad.12 Three
possible interpretations of the results of these preliminary
experiments are possible. The first interpretation is that the N2O
yield from the reaction of nascent O3 with N2 is about a factor
of 100 smaller than the one measured by Zipf and Prasad.12
The second interpretation is that the observed N2O results from
some other process (for example, the addition of electronically
excited oxygen atoms (O(1D2)) to N2) and, therefore, the yield
of N2O from the reaction of nascent O3 with N2 is negligibly
small. It is important to point out that if the yield of N2O from
the reaction of nascent O3 with N2 had turned out to be of the
order reported by Zipf and Prasad,12 then the N2O resulting from
other sources would have made a negligible contribution to the
total number of N2O molecules detected. The third interpretation
is, of course, that two or more processes are contributing a
significant fraction of the total number of N2O molecules
detected at the end of the experiment.
O(3PJ) atoms were produced by 532 nm laser flash photolysis
of mixtures containing typically 1 Torr of O3 in synthetic air
buffer gas in a static cell equipped with quartz windows. The
reaction cell was a Pyrex cylinder 15 mm in diameter and 50
cm in length with O-ring joints for attaching windows (internal
volume ∼95 cm3). A Quanta Ray Nd:YAG laser (Model DCR-
2A, pulse width ∼6 ns) operating at a frequency of 10 Hz served
as the photolytic source. The laser power exiting the reaction
cell (typically 95 mJ/pulse) was monitored by a Scientech Model
214 disk calorimeter. The corrected (for window losses) exiting
laser power could be used as a measure of the incoming laser
photons since at 532 nm and 1 Torr of O3 less than 0.5% of the
laser radiation is absorbed by the O3 molecules in the reaction
cell. The number of O(3PJ) atoms generated per laser pulse was
calculated from the incoming laser power, the fraction of laser
photons absorbed by the O3 molecules in the reaction cell, and
the known O(3PJ) yield of unity at 532 nm.43
2.3. Temperature-Dependent 266 nm Irradiation Experi-
ments. The experimental apparatus used to study reaction 6 was
almost identical to the one used in the preliminary experiments.
Only a slightly different reaction cell was employed. The cell
was a Pyrex cylinder 15 mm in diameter and 50 cm in length
with O-ring joints for attaching antireflection (AR) coated quartz
windows (248-355 nm) (internal volume ∼100 cm3). The cell
was maintained at constant temperature by circulating ethylene
glycol (T > 298) or methanol (T < 298) from a thermostated
bath through the outer jacket. A copper-constantan thermo-
couple with a stainless steel jacket was inserted into the center
of the reaction cell to measure the gas temperature under the
precise pressure conditions of the experiment. The temperature
measurement was made after all the kinetic experiments were
performed simulating the same bath and pressure conditions as
during the actual experiments.
The O2 used in this study was Ultra Pure Carrier Grade with
a minimum purity of 99.996%. Synthetic air was UHP/Zero
Grade with stated total hydrocarbon and water contents of less
than 0.5 and 3.5 ppm, respectively. The N2O calibration gas
was a certified standard containing 0.969 ppmv N2O in UHP
N2. All three gases were used as supplied.
In an attempt to decide if the second interpretation was
feasible, the number of O(1D2) atoms generated in each
experiment were integrated and a second N2O yield was
calculated (i.e., the number of N2O molecules detected per
O(1D2) atom generated; right axis of Figure 1). Preliminary
calculations indicated that, in this scenario, the room-temperature

N2O Production by Laser Flash Photolysis
J. Phys. Chem. A, Vol. 106, No. 24, 2002 5883
Figure 3. Plot of number of O(3PJ) atoms generated per Nd:YAG laser
pulse (empty circles, bottom curve) and [O3] (empty squares, top curve)
versus laser irradiation time in the same experiment shown in Figure
2. Each data point represents an 18-s average.
Figure 2. Plot showing the time history of N2O absorption measure-
ments made after the irradiation in experiment 2 (Table 1). Details of
the experimental procedure and data interpretation are given in the text.
rate coefficient for reaction 6 would be about 25 × 10-37 cm6
molecule-2 s-1. In fact, some agreement exists between this
preliminary value and the three previously published studies of
reaction 6. Our preliminary room-temperature rate coefficient
is a factor of 3-7 larger than two literature values32,33 but within
10% of the room-temperature rate coefficient reported by
Gaedtke et al.31 In addition, the observed linear dependence of
the N2O yield on pressure seems consistent with the expected
behavior of reaction 6 in the low-pressure third-order regime.
Both the magnitude of the observed N2O yield and the pressure
dependence of the observed N2O yield are, therefore, consistent
with the possibility that the observed N2O is coming from
reaction 6.
3.2. 532 nm Irradiation Experiments. To shed some light
on the first possible interpretation discussed above, the reaction
of nascent O3 with N2 was studied by irradiating O3/air mixtures
with radiation that only yields O(3PJ) atoms from the photolysis
of O3 (i.e., 532 nm). This set of experiments was designed to
completely eliminate the possible interference from reaction 6
to the observed N2O yield and, thus, definitively assess the role
of reaction 4b in the N2O budget. However, to detect an N2O
yield of the order presented in Figure 1, or even lower, we were
forced to slightly change our experimental apparatus and initial
conditions. At 532 nm, the yield of O(3PJ) from the photolysis
of O3 is unity43 (compared to the estimated O(3PJ) yield of 2.0
( 0.3 in the 266 nm irradiations) and, more critically, the O3
absorption cross section is only 2.65 × 10-21 cm2 molecule-1
(about a factor of 3600 lower than at 266 nm).44 As a result, to
generate a detectable level of N2O, this new set of experiments
were carried out, as briefly pointed out in the Experimental
Section, in a smaller volume reaction cell (∼95 cm3) and with
a higher initial O3 concentration (∼1 Torr). Furthermore,
irradiations were allowed to proceed for 10 h.
O2, was added to the photolysis cell and the transmitted light
intensity (I) at 253.7 nm was again measured. The photolysis
cell was subsequently filled to a final pressure of 904 Torr by
adding synthetic air from its high-pressure tank. For the next
45 min, the O3/air mixture was allowed to mix thoroughly until
the light intensity of the 253.7 nm Hg lamp radiation had
stabilized to the same level measured before the addition of
synthetic air to the photolysis cell (i.e., I). At this time, the 532
nm irradiation was started and allowed to proceed for 10 h. As
depicted in Figure 3, the concentration of O3 decreased over
the irradiation period by a factor of about 1.8. This small drop
in the O3 concentration is the result of the reaction of O(3PJ)
with O3 (even though most of the O3 photolyzed at 532 nm is
regenerated via the recombination reaction O(3PJ) + O2 + M
f O3 + M). The calculated number of O(3PJ) atoms generated
per laser shot is also shown in Figure 3, and as expected, it
decreases over the irradiation period since after each laser pulse
slightly less O3 is available to generate O(3PJ) by photolysis.
About 1 min after the irradiation period was completed (t ) 1
min in Figure 2), the multipass cell was filled to 33 Torr with
a 0.969 ppmv N2O standard mixture and its absorption was
measured. About 2 min later, the multipass cell was pumped
out and the background absorption was obtained from t ) 6
min to t ) 8 min by pumping on the cell. At t ) 8 min, the
pump-out valve was closed and the valve between the cells was
opened so that the reaction products could expand into the
infrared cell. The total pressure in the infrared cell was then
reduced to about 33 Torr by pumping, and from t ) 9 min to
t ) 10 min the N2O content of the reaction products was
measured. As observed from Figure 2, only a trace of N2O, if
any, was found to be produced photochemically from this
experiment.
Table 1 summarizes the conditions and results from the seven
experiments that were performed to study the N2O yield from
the reaction of nascent O3 with N2. The experiments can be
grouped into three main categories:
(I) Experiments 1 and 2 were carried out under “normal”
conditions.
(II) Experiments 3 and 4 were “blank” experiments.
(II) Experiments 5-7 had a known amount of N2O added to
the initial photolysis mixture.
Figures 2 and 3 show typical experimental results (experiment
2 in Table 1). Figure 2 shows how the N2O absorption signal
(as measured in the multipass infrared absorption cell) varied
during the measurement cycle after the Nd:YAG laser irradiation
at 532 nm was completed. Figure 3 shows how the concentration
of O3 and the number of O(3PJ) atoms generated per laser shot
(as measured in the photolysis cell) varied as a function of the
laser irradiation time. The experimental procedure employed
to obtain the data shown in Figures 2 and 3 is discussed below.
Before the experiment was started, both cells were pumped
out, and the transmitted light intensity (I0) at 253.7 nm was
measured. About 2 min later, 1.1 Torr of O3, with some residual
In Table 1, photochemically generated N2O was taken to be
the difference between the number of N2O molecules detected
in a given experiment (category I experiments) and the number

5884 J. Phys. Chem. A, Vol. 106, No. 24, 2002
TABLE 1: Summary of Experimental Results from O3 + N2 Experiments (532 nm Irradiations)
Estupin˜a´n et al.
†
av Nd:YAG
photolytic
initial
air
irradiation laser powera
initial N2O
N2O detected
N2O detected O(3PJ) generated
net N2O
expt no. O3 (Torr) press. (Torr) time (min) (mJ/pulse) (1012 molecules) (1012 molecules) (1012 molecules)
(1020 atoms)
yield (10-9
)
1
2
3
4
5
6
7
1.0
1.1
0.0
1.0
1.0
1.0
1.0
904
904
900
905
892
690
890
600
615
605
600b
500
960b
480
92
95
93
0
95
0
0.0
0.0
0.0
0.0
222
127
86
2.8
-2.9
2.8
-2.8
227
133
83
2.8
-2.9
2.9
3.3
9.8
-8.9
2.5
2.1
95
a 532 nm. b Total time of initial mixture in reaction cell.
of N2O molecules detected in a “blank” experiment (category
II experiments). The net N2O yield was taken to be the number
of photolytically generated N2O molecules divided by the total
number of O(3PJ) atoms produced in the same experiment. The
number of O(3PJ) atoms generated per laser shot was calculated
as
laser turned off. In all three cases, the amount of N2O detected
at the end of the experiment was (within the precision of the
measurement) equal to the amount of N2O added to the initial
photolysis mixture. The observations from experiments 1-7
suggest that very little, if any, of the N2O detected was generated
photochemically from reaction 4b.
Based on our experimental results, the upper limit quantum
yield for production of N2O from the reaction of nascent O3
with N2 is conservatively estimated to be 7 × 10-8. This value
is about a factor of 600 smaller than the value reported by Zipf
and Prasad12 and it is more than an order of magnitude below
the threshold for atmospheric importance. In addition, our results
suggest that the dominant source of N2O molecules observed
in the preliminary 266 nm experiments is something other than
the reaction of nascent O3 with N2 (probably the addition of
O(1D2) to N2).
N ) (E)(P)(Φ)(f)
(I)
where E is the Nd:YAG laser pulse energy, P is the number of
photons per millijoule at 532 nm, Φ is the O(3PJ) quantum yield
from O3 photodissociation at 532 nm, and f is the fraction of
photons absorbed by the O3 molecules in the path of the laser
beam. The parameter f was calculated as
f ) 1 - exp{-[O3](σO )532 nm(l1)}
(II)
3
In the experiment of Zipf and Prasad,12 atomic oxygen was
produced by the photodissociation of O2 in ultrapure synthetic
air using radiation from a filtered argon flash lamp. All the
experiments were performed in a stainless steel photolysis
chamber under flow conditions, and N2O was detected using a
gas chromatograph fitted with an electron capture detector.
Typically, irradiations lasted for 10-15 min and the N2O formed
in each experiment was cryogenically trapped and concentrated
in a small loop before analysis. Zipf and Prasad12 studied N2O
production over a large pressure range (i.e., 1-1000 Torr), and
they employed two reaction vessels of different surface-to-
volume ratios in an attempt to discriminate between surface and
gas phase reactions. Zipf and Prasad12 explain the nature of the
sources of N2O in their experiments in terms of two different
types of processes (i.e., type I and type II processes). Type I
processes, which the authors attribute to surface reactions, appear
dominant below 100 Torr, and type II processes, which the
authors attribute to gas phase reactions, emerge above 100 Torr.
The authors explain that type II processes must be due to gas
phase reactions (since surface processes are expected to decrease
with increasing pressure), and they suggest that the most likely
species responsible for type II production of N2O is nascent
O3. To obtain the N2O yield value of 4 × 10-5 from the nascent
O3 + N2 interaction, Zipf and Prasad12 were forced to use a
complex numerical simulation coupling the chemistries of O,
O3, and nascent O3 as well as the transport of these species to
the reactor walls. By comparison, our study minimizes artifact
production of N2O on various surfaces and, therefore, avoids
the use of complex numerical simulations of likely interfering
reactions as a tool to extract the N2O yield.
where (σO )532 nm is the O3 absorption cross section at 532 nm
3
and l1 is the laser beam path length (50 cm). The absolute O3
concentration is derived from the observed 253.7 nm light
intensity using Beer’s law
[O3] ) {ln(I0/It)}/{(σO )253.7 nm(l2)}
(III)
3
where I0 is the light level when the reaction cell is empty, It is
the light level at a time t during the irradiation, (σO )253.7 nm is
3
the 253.7 nm O3 absorption cross section, and l2 is the path
length for O3 detection (i.e., perpendicular to the Nd:YAG laser
beam; 5.4 cm).
As mentioned above, category I experiments were carried
out under “normal” conditions which were selected to optimize
the chances of observing a very small photochemical yield of
N2O (i.e., 900 Torr of total air pressure, 1 Torr of O3, maximum
Nd:YAG laser power at 532 nm, and 10 h irradiation times). In
these experiments, the observed N2O yields were +9.8 × 10-9
and -8.9 × 10-9, respectively; the average net yield was 4.4
× 10-10
.
In category II experiments, one experiment (experiment 3)
was carried out by irradiating 900 Torr of air only and checking
the N2O content of the products after 10 h of irradiation. A
second experiment (experiment 4) was carried out by leaving
an O3/air mixture in the reaction cell for 10 h (i.e., laser turned
off) and then measuring its N2O content. The number of N2O
molecules observed in these experiments is not statistically
different from the N2O levels observed in category I experi-
ments. Therefore, the number of N2O molecules from experi-
ments 3 and 4 were averaged, and that number (essentially zero)
was used to compute the number of photolytically generated
N2O molecules in category I experiments.
Zipf and Prasad12 carried out their study motivated by several
laboratory experiments suggesting the production, or lack of
production, of N2O from various mixtures of N2, O2, and
O3.45-50 Nevertheless, Zipf and Prasad12 have suggested in their
analysis that the small N2O yield obtained in most of these other
literature studies could be explained by a loss process destroying
Category III experiments were designed to verify that the
photolysis, transfer, and detection processes could be carried
out without destroying any photochemically generated N2O. One
experiment (experiment 6) was carried out with the Nd:YAG

N2O Production by Laser Flash Photolysis
J. Phys. Chem. A, Vol. 106, No. 24, 2002 5885
TABLE 2: Summary of Kinetic Data for the Reaction
to be the number of N2O molecules detected in a given 266 nm
irradiation. Note that the 295 K rate coefficient measured in
this set of experiments is within 5% of the value determined in
the preliminary experiments.
O(1D2) + N2 + M f N2O + M
O(1D2)
air press. N2O yield generated no. of k6d (10-36 cm6
T (K) (Torr)a
324 401-805
(10-6
0.9-1.7
)
(1020 atoms) exptsc molecule-2 s-1
)
Unlike the 532 nm experiments, readily measurable levels
of N2O were observed at all pressures and temperatures
investigated. Irradiation times ranged from 10 to 110 min and,
in most cases, irradiations were stopped when approximately
50 ppbv N2O was expected to have been generated in the
photolysis cell (about 30 ppbv in those experiments performed
at the lowest pressures and highest temperatures). As observed
from Table 2, the yield of N2O was obtained at four or more
different pressures per temperature investigated. The number
of O(1D2) atoms generated per laser shot was calculated as in
eq I with the difference that Φ ) 0.88 36 and P is the number
of photons per millijoule at 266 nm. The fraction of photons
absorbed by the O3 molecules in the photolysis cell was also
computed as in eq II, but in this case, the O3 absorption cross
section at 266 nm was used instead of the O3 absorption cross
section at 532 nm. As mentioned in the Experimental Section,
a direct measurement of the fraction of laser photons absorbed
by the O3 molecules in the reaction cell was also performed. If
I0 is taken to be the number of transmitted 266 nm photons
when the reaction cell is empty and I is taken to be the number
of transmitted photons at any time during the 266 nm irradiation,
the difference (I0 - I) directly gives the number of O(1D2) atoms
produced. On average, the two determinations of the number
of O(1D2) atoms generated over the course of a given irradiation
agreed to within (4%.
0.3-0.6
0.3-1.1
0.4-2.2
0.9-3.2
0.5-1.2
4
6
7
4
4
2.62 ( 0.28
2.68 ( 0.43
3.13 ( 0.46
3.56 ( 0.51
3.48 ( 0.65
295 192-793b 0.5-1.9
270 193-802
243 178-778
220 152-760
0.4-2.1
0.5-2.4
0.4-2.3
a Typical initial [O3] ranged from 9.5 × 1015 to 9.8 × 1015 molecules
cm-3 unless otherwise noted. b Two experiments, at 198 and 793 Torr,
were performed with an initial [O3] of 4.5 × 1015 molecules cm-3
.
c Experiment ≡ determination of a single N2O yield at a given air
pressure. d Individual rate coefficients shown in the table have been
corrected for the effect of reaction 2. Uncertainties are 2σ and represent
precision only.
any N2O formed via reaction 4b when O3 is present initially in
the reaction cell. However, category III experiments from our
study tested this hypothesis, and they conclusively rule out an
efficient N2O loss process under our experimental conditions.
Another potential point of concern is the possibility that
ground state O3 could have rapidly deactivated vibrationally
excited O3 and, thus, suppressed N2O formation. It is interesting
to note that our experiments were designed to greatly increase
the N2-to-O3 ratio by using only 1 Torr of O3 in 900 Torr of air
(compared to 5-100 Torr of O3 in typical mixtures employed
in previous experiments). In addition, the N2-to-O3 ratio was
even smaller in the preliminary 266 nm experiments (where
only 0.1-0.3 Torr of O3 was initially present in the reaction
cell). Nonetheless, the O3:air ratio employed in our experiments
is several orders of magnitude larger than the O3:air ratio in
the atmosphere. Even though there is no direct information in
the literature on the average energy removed from internally
excited O3 per collision with diatomics and triatomics, informa-
tion is available on the average energy removed per collision
from internally excited SO2, CS2, and NO2 by several mona-
tomic, diatomic, and triatomic species.51-53 Focusing our
attention on internally excited NO2, which is the most thoroughly
studied molecule of the three, it has been observed that at high
internal excitation energies the ratio of the energy removed by
a collision with N2 or O2 is only about a factor of 5 smaller
than the energy removed per collision by triatomics such as
CO2, N2O, and NO2.52 Furthermore, the available evidence
suggests that there is no significant enhancement in the average
energy removed per collision when the collider and the excited
species are chemically identical, i.e., the average energy removed
Rate coefficients for reaction 6 were computed from the net
N2O yields. Figure 4 shows two sample plots of the pressure
dependence of the net N2O yields obtained at T ) 220 and 295
K. At all five temperatures investigated, the plots were linear,
which is consistent with N2O production via a termolecular
process that occurs in competition with a bimolecular process,
i.e., deactivation of O(1D2) to O(3PJ) by N2 and O2:
O(1D2) + N2 f O(3PJ) + N2
(8)
-
O(1D2) + O2 f O(3PJ) + O2(X3Σg ) Φ ) 0.2
(9a)
+
f O(3PJ) + O2(b1Σg ) Φ ) 0.8 54-57 (9b)
As a result, the net N2O yield is the ratio of the rate of reaction
6 divided by the sum of the rates of reactions 6, 8, and 9. Since
k6[N2][M] , k8[N2] + k9[O2], it then follows that the association
rate coefficient for reaction 6 at each temperature can be
computed from the following expression:
†
in NO2 -NO2 collisions is about the same as the average energy
†
removed in collisions of NO2 with other triatomic molecules.
Assuming that O3 behaves similarly to the molecules studied
by Hartland et al.51,52 and Chimbayo et al.,53 it seems highly
unlikely that O3 could have competed with N2 + O2 at
deactivating vibrationally excited O3 under our experimental
conditions.
3.3. Temperature-Dependent 266 nm Irradiation Experi-
ments. 3.3.1. Kinetic Results. Under the assumption that all N2O
observed in the 266 nm photolysis experiments is produced via
reaction 6, the N2O yield data can be employed to obtain values
for k6(T). Rate coefficients were evaluated at five different
temperatures ranging from 220 to 324 K, and the results are
summarized in Table 2. The net N2O yield was taken to be the
number of photolytically generated N2O molecules divided by
the number of O(1D2) atoms produced in the same experiment.
Since the “blank” runs yielded negligible N2O in the preliminary
experiments, photolytically generated N2O was simply taken
k6 ) Φ(P){k8[N2] + k9[O2]}/[N2][M]
(IV)
where Φ(P) is the pressure-dependent N2O yield and [M] is
the total gas concentration. Although x-intercepts in Figure 4
do vary systematically with temperature, they are always within
the reported 2σ uncertainty (precision only) of zero.
The temperature-dependent rate coefficients for reaction 9
were obtained from the current NASA panel recommendation.58
On the other hand, the rate coefficients for reaction 8 were
derived from expression V:
k8(T) ) 2.1 × 10-11 exp(115/T) cm3 molecule-1 s-1 (V)
Expression V is based on recent results from three laboratories
which suggest that k8(T) is somewhat faster than previously
thought.59 It is interesting to note that using the 2000 NASA

5886 J. Phys. Chem. A, Vol. 106, No. 24, 2002
Estupin˜a´n et al.
Figure 4. Plots of net N2O yield versus air pressure for data obtained
in the study of reaction 6 at T ) 220 and 295 K. Solid lines are obtained
from least-squares analyses. Their slopes give the following values in
units of 10-9 Torr-1: 3.24 ( 0.60 at 220 K and 2.10 ( 0.34 at 295 K.
Their intercepts give the following values (where uncertainties are 2σ
and represent precision only): (-1.65 ( 3.10) × 10-7 at 220 K and
(0.76 ( 1.88) × 10-7 at 295 K.
Figure 5. Arrhenius plot for reaction 6. The solid line is obtained
from a least-squares analysis which weights each data point equally; it
represents the Arrhenius expression given in the text. Individual rate
coefficients have been corrected for the effect of reaction 2.
TABLE 3: Reaction Mechanism Used To Simulate the 266
nm Kinetics Experiments
reaction
ka
panel recommended expression for k8(T) in eq IV to compute
k6(T) would lead to rate coefficients about 12% smaller than
those obtained using expression V. These changes are well
within our reported uncertainties in k6(T) (see error analysis
discussion below).
An Arrhenius plot for reaction 6 is shown in Figure 5. The
temperature dependence of reaction 6 is characterized by a small
negative activation energy. A linear least-squares analysis of
the ln k6 vs 1/T data gives the following expression:
O3 + hV f O(1D) + O2(1∆)
0.15b,c
f O(3P) + O2
0.02b,c
O(1D) + O2 f O(3P) + O2
f O(3P) + O2(1Σ)
8.0 × 10-12
3.2 × 10-11
1.2 × 10-10
1.2 × 10-10
4.9 × 10-11
6.7 × 10-11
3.0 × 10-11
2.6 × 10-36 d
3.8 × 10-15
1.7 × 10-18
1.0 × 10-20
0-20b
O(1D) + O3 f O2 + O2
f 2O(3P) + O2
O(1D) + N2O f N2 + O2
f NO + NO
O(1D) + N2 f O(3P) + N2
O(1D) + N2 + M f N2O
O2(1∆) + O3 f O(3P) + 2O2
O2(1∆) + O2 f products
O2(1∆) + N2 f products
O2(1∆) f O2
k6(T) ) (1.3 ( 0.5) × 10-36
×
exp{(230 ( 110)/T} cm6 molecule-2 s-1 (VI)
O2(1∆) + O f products
O2(1Σ) + N2 f products
O2(1Σ) + O2 f products
O2(1Σ) + O3 f O(3P) + 2O2
f products
2.0 × 10-16
2.1 × 10-15
3.9 × 10-17
1.5 × 10-11
6.6 × 10-12
8.0 × 10-14
6.2 × 10-34 d
8.0 × 10-15
Uncertainties in the above expression are 2σ and represent
precision only. These uncertainties refer to the Arrhenius
parameters only. Error estimates for individual rate coefficients
are derived below.
The NASA panel for chemical kinetics and photochemistry
data evaluation58 typically approximates the temperature de-
pendence of rate coefficients for association reactions in their
low-pressure regime with an expression of the form
O2(1Σ) + O f products
O + O2 + M f O3
O + O3 f 2O2
a Rate coefficients used to simulate the room-temperature experi-
ments. Units are cm3 molecule-1 s-1 unless otherwise noted. Temper-
ature-dependent rate coefficients for each reaction in the simulation
were primarily obtained from the expressions recommended by the 2000
NASA panel.58 Exceptions were the temperature-dependent rate coef-
ficients for reaction 8, where expression V was employed instead (see
discussion in text), and the temperature-dependent rate coefficients for
reaction 6, where the rate coefficients derived from this study were
used instead. The third-order rate coefficient entered for reaction 6 was
obtained after running the model at each pressure studied at 295 K
and correcting the individual N2O yields at each pressure investigated
(i.e., the calculation of the final temperature-dependent rate coefficients
for reaction 6 shown in Table 2 was an iterative process). b Units are
k0(T) ) k0(300 K)(T/300)-n
(VII)
Fitting our measured values to eq VII gives the following
expression:
k6,0(T) ) (2.8 ( 0.1) × 10-36(T/300)-(0.88(0.36)
×
cm6 molecule-2 s-1 (VIII)
s-1 c Typical first-order O3 decay rate used to model the O3 concentra-
.
tion. The O3 decay rate shown in the table was used to model one
Once again, the stated uncertainties refer to the fit parameters
only (i.e., they represent 2σ precision). It is worth noting that
this is the first time that a temperature-dependent kinetics study
of reaction 6 has been reported.
At each temperature and pressure investigated, it becomes
necessary to consider the effect of reaction 2 on the observed
net N2O yields. Thus, to perform the appropriate corrections to
our kinetic data, we decided to simulate our experiments using
ACUCHEM, a numerical integration routine written at the
National Institute of Standards and Technology. Table 3 shows
the chemical mechanism used in our simulations including the
experiment carried out under the following conditions: P ) 192 Torr;
T ) 295 K; [O3]0 ) 9.8 × 1015 molecules cm-3
.
d Units are cm6
molecule-2 s-1
.
rate coefficients employed to simulate the room-temperature
experiments.
Initially, we attempted to simulate the total number of N2O
molecules produced and destroyed in each experiment by
matching the ACUCHEM-simulated O3 profile to the experi-
mentally observed O3 profile (ACUCHEM most conveniently

N2O Production by Laser Flash Photolysis
J. Phys. Chem. A, Vol. 106, No. 24, 2002 5887
simulates the photolysis of O3 with a first-order decay rate).
Trial-and-error inputs of different first-order O3 decay rates were
made until a suitable match between both profiles was obtained.
However, it was found that, in every case, the model overes-
timated the total number of O(1D2) atoms generated during the
irradiation period (typically by factors of 1.4-2.0), and thus it
overestimated the total number of N2O molecules detected (by
approximately the same factor). This motivated us to modify
our modeling approach and change the model input of the first-
order O3 decay rate (once again by trial and error) to one that
would lead to a close match between the calculated and
simulated total number of O(1D2) atoms generated in each
experiment. With this new procedure, good matches were readily
obtained and the total number of N2O molecules generated in
each experiment was properly simulated (within the precision
of the N2O detection measurement). Nonetheless, this second
approach required inputting a smaller value for the O3 first-
order decay rate than the one used when attempting to match
the O3 profiles. A combination of two possible factors can
perhaps explain the mismatch between O3 profiles. The first
one is the fact that we are unable to model properly the
experimental O3 profile with ACUCHEM because we assumed
a first-order photolysis rate when, in reality, the effective first-
order O3 decay rate changed as a function of irradiation time
(due to optical thickness effects). Alternatively, the disagreement
in the O3 profiles could be suggestive of the possible presence
of one or more additional O3 loss process(es) not considered in
our chemical mechanism (see Table 3). It is interesting to note
that an additional unidentified O3 loss process is not present
when the laser is turned off or when 532 nm radiation is used
instead of 266 nm irradiation (concluded after a similar
ACUCHEM simulation of the 532 nm irradiation experiments).
Even though the exact nature of the additional O3 loss process
(if any) is not clear at this time, it is important to point out that
its possible presence would not affect the value of our net N2O
yields, or our calculated rate coefficients for reaction 6. This is
the case because the number of O(1D2) atoms generated per
laser pulse was calculated from the measured O3 profile and
not from the simulated O3 profile. Moreover, it seems unlikely
that a process other than photolysis would lead to the production
of O(1D2) atoms.
(266 nm) is production of two O(3PJ). Since the only allowed
photolysis channel is spin-forbidden, one would expect the O2-
(a1∆g) absorption cross section at 266 nm to be very small. To
our knowledge, no measurements of O2(a1∆g) absorption cross
sections have been reported at wavelengths longer than 200 nm.
However, theoretical calculations of long-wavelength absorption
cross sections have been reported60 that confirm the expected
very small absorption cross section, i.e., σ(266 nm) ≈ 4 × 10-24
cm2 molecule-1
.
3.3.2. Estimated Accuracy of Reported Rate Coefficients. The
three major contributors to the overall accuracy of the rate
coefficients in this temperature-dependent kinetic study of the
addition of O(1D2) to N2 are the precision of the net N2O yields,
the accuracy of the total number of O(1D2) atoms generated
per laser pulse, and the uncertainty in the literature values for
reactions 8 and 9. These sources of error are discussed below.
The accuracy of the total number of O(1D2) atoms generated
per laser pulse is a function of the uncertainty in the following
factors: the measurement of the total number of incoming laser
photons, the measurement of the fraction of photons absorbed
by the O3 molecules in the reaction cell, and the literature value
for the quantum yield for production of O(1D2) in the photolysis
of O3 at 266 nm. We estimate that the accuracy of our
measurement of the total number of incoming laser photons is
(5%. The quantum yield for production of O(1D2) is known to
(3%.36 The accuracy of the fraction of photons absorbed by
the O3 molecules in the reaction cell is itself a function of the
accuracy in the measurement of the O3 concentration and the
uncertainty in the literature value for the O3 absorption cross
section at 266 nm. We estimate that our measurement of the
O3 concentration has a (3% uncertainty (which itself includes
a (1% error in both I and I0, and a (2% error in the O3
absorption cross section at 253.7 nm58). On the other hand, the
O3 absorption cross section at 266 nm is known within (3%.58
Therefore, the estimated accuracy of the fraction of photons
absorbed by O3 is (4%. Combining all the above factors
together, the error in the total number of O(1D2) atoms generated
in each irradiation is estimated to be about (7%.
Examination of Table 2 shows that on average the precision
of the net N2O yields is (15%. Conservatively estimating that
the ACUCHEM simulations lead to an additional (5% error
in the net N2O yields, we estimate that the accuracy of the total
number of N2O molecules measured in our study is (16%.
Combining the error in the determination of the total number
of N2O molecules with that of the total number of O(1D2) atoms,
we obtain a preliminary overall error in our reported rate
coefficients of (17%. As indicated by eq IV, the temperature-
dependent rate coefficients for reaction 6 depend, in our analysis,
on the temperature-dependent rate coefficients for reactions 8
and 9. Both of those rate coefficients are known to (20% at
room temperature and to (35% at 220 K.58 Combining the
accuracy in our reported rate coefficients (i.e., (17%) with the
uncertainty in the literature values for reactions 8 and 9 leads
to an overall accuracy of (27% in our reported room-
temperature rate coefficient and to an overall accuracy of (39%
in our 220 K rate coefficient. The following format is widely
used to express uncertainties in rate parameters that are used in
modeling atmospheric chemistry:58
Based on the 266 nm ACUCHEM simulations (using the
second approach described above), it was calculated that, on
average, about 10% of the total number of N2O molecules
produced photolytically were destroyed by reactions 2a and 2b
over the course of each irradiation experiment (i.e., the
percentage of N2O destroyed actually ranged from 4% to 23%
depending on conditions and irradiation time). Changes to
individual net N2O yields led to changes of typically less than
10% in k6(T). Therefore, even if the ACUCHEM simulations
were somewhat uncertain, the rate coefficients would not change
much.
One potential complication in the 266 nm photolysis experi-
ments concerns the long-lived species O2(a1∆g), which is
generated directly from O3 photolysis and also from the O(1D2)
+ O2 interaction via the intermediate O2(b1Σg+) (see Table 3).
O2(a1∆g) is sufficiently stable that it is not expected to
completely decay away during the 0.1 s between laser flashes.
Hence it is necessary to consider the possibility that 266 nm
photolysis of O2(a1∆g) could be an additional source of O(1D2)
that must be accounted for in the data analysis. This possibility
can, however, be ruled out on energetic grounds since, based
on the best available thermochemical information,58 the only
energetically allowed photolysis channel for O2(a1∆g) + hν-
f(T) ) f(298 K)exp{(∆Ea/R)(T-1 - 298-1)}
(IX)
Applying expression IX to our results leads to the following
estimated parameters: f(298 K) ) 1.27 and (∆Ea/R) ) 75.
3.3.3. Comparison of Reported Rate Coefficients with Lit-
erature Values. As mentioned above, there are three literature

5888 J. Phys. Chem. A, Vol. 106, No. 24, 2002
Estupin˜a´n et al.
TABLE 4: Summary of Literature Values for the O(1D2) + N2 + M Rate Coefficient
N2O
detection
technique
k6(298 K)
Ea
reference
P (atm)
(10-36 cm6 molecule-2 s-1
)
(kJ mol-1
)
Gaedtke et al. (1972)31
Kajimoto and Cvetanovic (1976)32
Maric and Burrows (1992)33
this work
GCa
GC
1-200
20-120
1
2.8 ( 1.4b
0.35 ( 0.30c
0.88 ( 0.33
2.8 ( 0.8
GC
TDLASd
0.1-1.2
-1.9 ( 0.9
a Gas chromatography. b Upward revision appears necessary (see text). c Error bar estimated by NASA panel.58 d Tunable diode laser absorption
spectroscopy.
reports of the room-temperature rate coefficient for reaction 6,
i.e., Gaedtke et al.,31 Kajimoto and Cvetanovic32 (which is the
basis of the current NASA panel recommendation58), and Maric
and Burrows.33 Each one of them is discussed below. Table 4
summarizes the literature values for k6.
as a function of the square of the nitrogen pressure. Extrapolation
of their results to 1 atm leads to a quantum yield for N2O
production of 3.1 × 10-7, which translates into a room-
temperature rate coefficient of 3.5 × 10-37 cm6 molecule-2 s-1
(i.e., about a factor of 7.5 slower than the one derived from
this study).
The first reported study of reaction 6 was performed by
Gaedtke et al.31 These investigators introduced 20 Torr of O2,
0.2 Torr of O3, and 1-200 atm of N2 buffer gas into a steel
reaction tube. O3 photolysis was carried out with a 200 W Xe-
Hg lamp and irradiation times lasted between 1 and 20 min.
After the irradiation period, N2O was collected in a trap and
analyzed by gas chromatography. Both the quantum yield for
O3 photolysis at 260 nm and the number of N2O molecules
measured per O3 molecule destroyed (defined by the authors
as the “N2O yield”) were measured as functions of N2 pressure.
Gaedtke et al.31 observed that the N2O yield increased with
increasing pressure up to a value close to unity at very high
pressure. To obtain a value for the room-temperature rate
coefficient for reaction 6, the authors approximated their
observed N2O yield by a ratio of the rate of reaction 6 divided
by the sum of the rates of reaction 6 and twice the rate of
reaction 10:
Like Gaedtke et al.,31 Kajimoto and Cvetanovic32 were forced
to study reaction 6 at very high N2 pressures (20-120 atm)
due to their relatively poor N2O detection sensitivity. As a result,
Kajimoto and Cvetanovic32 performed a long extrapolation in
order to derive the N2O yield at a pressure of 1 atm. The
disadvantage of such an approach is that a small systematic error
in the N2O yield measured at high pressures can result in a large
error in the extrapolated N2O yield at 1 atm (even though their
N2O yield data appear to have excellent precision). The better
N2O detection sensitivity in our study permitted kinetic mea-
surements at lower, more realistic atmospheric pressures. It is
also interesting to note that Kajimoto and Cvetanovic32 observed
a quadratic dependence of the N2O quantum yield on pressure
in the high-pressure regime, whereas our lower pressure results
show a linear dependence of the N2O yield on pressure.
As discussed by Kajimoto and Cvetanovic,32 a nonlinear
dependence of N2O yield on pressure can be rationalized if it
is assumed that multiple collisions with the bath gas are required
to deactivate newly formed N2O to an energy sufficiently low
that crossing to a triplet surface (and subsequent rapid dissocia-
tion to O(3PJ) + N2) cannot occur. When O(1D2) interacts with
N2, it is thought that some energy is rapidly shunted into N2
vibration, yielding an energized complex with a rather long
lifetime (estimated to be 1-10 ps; i.e., hundreds of vibrational
periods) toward dissociation back to reactants.61-63 Hence, the
lifetime of energized N2O is limited by the rate of crossing to
the triplet surface which, in the absence of collisions, occurs
with essentially unit probability in competition with dissociation
back to reactants.61,62 For our experimental conditions, where
the mean time between collisions ranged from 100 to 1000 ps,
it seems reasonable that most N2O is generated when a single
collision with the bath gas deactivates the energized complex
to an internal energy where crossing to the triplet surface cannot
occur (i.e., the probability of an energized complex experiencing
multiple collisions during its lifetime is very small). On the other
hand, a multiple collision mechanism may have been operative
under the high-pressure conditions employed by Kajimoto and
Cvetanovic.32
O(1D2) + O3 f 2O2
(10)
using a value of 2.5 × 10-10 cm3 molecule-1 s-1 for k10. As a
result, Gaedtke et al.31 obtained a room-temperature rate
coefficient for reaction 6 of (2.8 ( 1.4) × 10-36 cm6 molecule-2
s-1. Interestingly, this value is virtually identical to our room-
temperature rate coefficient derived from expression VI!
Nonetheless, it is our opinion that the expression that Gaedtke
et al.31 used to derive their third-order room-temperature rate
coefficient should be modified. First, it is currently believed
that the rate coefficient for reaction 10 is 1.2 × 10-10 cm3
molecule-1 s-1 58
O(3PJ) atoms with O3
.
In addition, it appears that the reaction of
O(3PJ) + O3 f 2O2
(11)
contributed as an important ozone loss process under the
experimental conditions employed by Gaedtke et al.31 Quantita-
tive reanalysis of Gaedtke et al.'s data is not possible because,
at the high pressures employed in their study, a complex energy
transfer scheme may be operative32 (see below) that results in
a nonlinear dependence of N2O yield on pressure.
In the study performed by Maric and Burrows,33 a mixture
of 3350 ppm O3 in synthetic air at 1 atm pressure was irradiated
with a filtered (λ g 254 nm) low-pressure mercury lamp. The
reaction cell was made of Pyrex and it was equipped with
Suprasil windows. The concentration of O3 was monitored
photometrically at 311.6 nm using radiation from a D2 lamp.
The mixing ratio of N2O was determined by cryogenically
trapping the reaction products and transferring a 2 mL sample
(with a gastight syringe) to a gas chromatograph fitted with an
electron capture detector. Maric and Burrows33 calculated the
In the study of Kajimoto and Cvetanovic,32 excited oxygen
atoms were generated by photolyzing O3 at wavelengths between
220 and 280 nm using filtered radiation from a 500 W Hanovia
medium-pressure Hg lamp. Experiments were performed in a
quartz reaction cell containing typically 10 Torr of O3, 100 Torr
of O2, and 20-120 atm of N2. After a period of irradiation
ranging from 13 to 48 h, the resulting N2O was cryogenically
trapped and measured by gas chromatography. Kajimoto and
Cvetanovic32 observed that the N2O quantum yield increased

N2O Production by Laser Flash Photolysis
J. Phys. Chem. A, Vol. 106, No. 24, 2002 5889
Introduction, since O(1D2) is generated from the photolysis of
O3, and since O3 has been observed to be isotopically mass-
independently enriched in the atmosphere (with an extraordinar-
ily large 17O excess of ∆17O ≈ 30‰),26-28 reaction 6 could
contribute to the observed mass-independent enrichment of N2O.
To explain the entire 17O excess of ∆17O ≈ 1‰ in atmospheric
N2O,13,14,30 about 3% of the total N2O source strength must
originate from an N2O source stemming from O3. Since our
results suggest that reaction 6 constitutes about 1.4% of the
currently estimated total N2O source strength, the contribution
of reaction 6 to the N2O budget is of the right magnitude to
account for a significant fraction of the oxygen mass-
independent enrichment observed in atmospheric N2O. As noted
above, about two-thirds of the total atmospheric N2O production
from reaction 6 takes place in the stratosphere. This finding is
consistent with the observation by Cliff et al.14 that stratospheric
N2O displays larger oxygen mass-independent enrichments than
tropospheric N2O. Although an additional N2O source may be
necessary to explain the observed 17O excess of ∆17O ≈ 1‰,
this is the first time that a mechanism that generates N2O
photochemically in the atmosphere has been reported that may
explain the altitude dependence of the N2O isotopic signature.
Even though the mechanism proposed by Rockmann et al.30
(i.e., transfer of mass independently enriched oxygen from O3
to NO2, followed by a second transfer to N2O via the reaction
NO2 + NH2 f N2O + H2O) can partially explain the N2O mass-
independent fractionation in the troposphere, the coupling of
the proposed mechanism to the ammonia cycle means that it
cannot account for the increase in mass-independent oxygen
enrichment in stratospheric N2O. It does appear, however, that
reaction 6 together with the mechanism proposed by Rockmann
et al.30 could potentially fully explain the mass-independent
oxygen isotope fractionation of tropospheric and stratospheric
N2O.
Figure 6. Altitude dependence of annual N2O production rate from
processes initiated by absorption of an ultraviolet photon by O3. Yields
of N2O reported in this study are used to evaluate the production rates.
Each data point represents a globally averaged integration over the next
1 km.
room-temperature rate coefficient for reaction 6 from the
measured rate of formation of N2O, the measured decay rate of
O3, and the mean O3 concentration during the irradiation period.
Their value of (8.8 ( 3.3) × 10-37 cm6 molecule-2 s-1 is a
factor of 2.5 faster than the value derived by Kajimoto and
Cvetanovic,32 although the values agree to within combined error
limits; their value is approximately a factor of 3 lower than the
value reported in this study. The study of Maric and Burrows33
employed total pressures similar to those used in our study.
Unfortunately, it is difficult to comment on the accuracy of the
data analysis and determination of N2O quantum yields in the
study of Maric and Burrows33 due to the lack of detail given in
their paper.
3.3.5 Alternate Interpretation of the 266 nm Photolysis
Experiments. In a recent conference presentation,67 Prasad has
suggested that the N2O observed in our 266 nm photolysis
experiments may not result from the O(1D2) + N2 association
reaction, but rather from the reaction of undissociated electroni-
cally excited O3 (lifetime ∼ 10 fs) with N2. To justify his
somewhat more exotic interpretation of our results, Prasad cites
the faster value for k6 obtained from our data (compared to the
value reported by Kajimoto and Cvetanovic32) and the linear
dependence of the N2O yield on pressure observed in our
experiment (compared to the quadratic dependence observed
in the higher pressure study of Kajimoto and Cvetanovic32). At
this time there does not appear to be any theoretical or
experimental information available to definitively rule out
Prasad’s suggestion. On the other hand, there is no experimental
or fundamental theoretical information available to support
Prasad’s suggestion, either. Pending further experimental and/
or theoretical research on this issue, we consider O(1D2) + N2
association to be the probable source of the N2O observed in
our 266 nm photolysis experiments.
3.3.4. Implications for Atmospheric Chemistry. To assess the
potential importance of reaction 6 as an atmospheric source of
N2O, a total yearly N2O production rate was calculated using
the kinetic data obtained in our study of reaction 6. O3
concentrations above 100 mbar were obtained from the middle
atmospheric model of Wang et al.64 Below 100 mbar, O3
concentrations were obtained from observed O3 distributions
in the troposphere.65 O3 photolysis rates were computed from
the chemical model developed by Ramaroson et al.66 together
with updated O(1D2) quantum yield data from Sander et al.58
The total number of N2O molecules generated per year in the
troposphere and the stratosphere was calculated to be 3.8 ×
1033. Figure 6 shows the altitude dependence of the annual N2O
production rate from reaction 6, where each data point represents
a globally averaged integration over the next 1 km. The
troposphere accounts for only about one-third of the total N2O
production rate, while the stratosphere accounts for about two-
thirds of the total N2O production rate. The most recent study
of the atmospheric N2O budget by Kroeze et al.8 suggests that
the input flux of N2O into the atmosphere is approximately 3.9
× 1035 molecules per year, even though recent estimates range
from 2.1 × 1035 to 3.9 × 1035 molecules per year.3,7-9 Based
on the above analysis, and considering the uncertainty in the
global flux of N2O into the atmosphere, reaction 6 represents a
source that is about 1.4 ( 0.4% of the estimated total N2O
source strength.
It is worth noting that the atmospheric implications discussed
in section 3.3.4 are essentially independent of whether N2O is
produced from O(1D2) + N2 or from electronically excited O3
+ N2. The model calculation of annual N2O production would
be identical in the two cases as long as it is assumed that only
the excited O3 electronic state that dissociates to singlet products
can interact with N2 to produce N2O. Similarly, since O3
provides the O atom to N2O in both cases, the potential
contribution to mass-independent isotope effects is also inde-
pendent of which of the two possible pathways is operative.
An approximate 1.4% contribution from reaction 6 to the
global yearly N2O source strength, although seemingly small,
may have important implications for the observed isotopic
composition of N2O in the atmosphere. As mentioned in the

5890 J. Phys. Chem. A, Vol. 106, No. 24, 2002
Estupin˜a´n et al.
(26) Krankowsky, D.; Lammerzahl, P.; Mauersberger, K. Geophys. Res.
Lett. 2000, 27, 2593 and references therein.
(27) Krankowsky, D.; Bartecki, F.; Klees, G. G.; Mauersberger, K.;
Schellenbach, K.; Stehr, J. Geophys. Res. Lett. 1995, 22, 1713.
(28) Johnston, J. C.; Thiemens M. H. J. Geophys. Res. 1997, 102, 25395.
(29) Prasad, S. S.; Zipf, E. C. Chemosphere: Global Change Sci. 2000,
2, 235.
(30) Rockmann, T.; Kaiser, J.; Crowley, J. N.; Brenninkmeijer, C. A.
M.; Crutzen, P. J. Geophys. Res. Lett. 2001, 28, 503.
(31) Gaedtke, H.; Glanzer, K.; Hippler, H.; Luther, K.; Troe, J.
Fourteenth Symposium (International) on Combustion; 1972; p 295.
(32) Kajimoto, O.; Cvetanovic, R J. J. Chem. Phys. 1976, 64, 1005.
(33) Maric, D.; Burrows, J. P. J. Photochem. Photobiol. A: Chem. 1992,
66, 291.
(34) Estupin˜a´n, E. G.; Stickel, R. E.; Wine, P. H. Chemosphere: Global
Change Sci. 2000, 2, 247.
(35) Estupin˜a´n, E. G.; Stickel, R. E.; Wine, P. H Chem. Phys. Lett. 2001,
336, 109.
(36) Brock, J. C.; Watson, R. T. Chem. Phys. Lett. 1980, 71, 371.
(37) Yoshino, K.; Freeman, D. E.; Esmond, J. R.; Parkinson, W. H.
Planet. Space Sci. 1988, 36, 395.
(38) Barnes, J.; Mauersberger, K. J. Geophys. Res. 1987, 92, 14861.
(39) Molina, L. T.; Molina, M. J. J. Geophys. Res. 1986, 91, 14501.
(40) Mauersberger, K.; Barnes, K.; Hanson, D.; Morton, J. Geophys.
Res. Lett. 1986, 13, 671.
(41) Hearn, A. G. Proc. Phys. Soc. (London) 1961, 78, 932.
(42) Inn, E. C. Y.; Tanaka, Y. J. Opt. Soc. Am. 1953, 43, 870.
(43) Wayne, R. P. Atmos. EnViron. 1987, 21, 1683.
(44) Ackerman, M. In Mesospheric Models and Related Experiments;
Fiocco, G., Ed.; Reidel: Dordrecht, Holland, 1971.
(45) Goody, R. M.; Walshaw, C. D. Q. J. R. Meteorol. Soc. 1953, 79,
496.
(46) Harteck, P.; Dondes, S. J. Chem. Phys. 1954, 22, 758.
(47) DeMore, W.; Raper, O. F. J. Chem. Phys. 1962, 37, 2048.
(48) Norrish, R. G. W.; Wayne, R. P. Proc. R. Soc. London, Ser. A
1965, 288, 200.
(49) Katakis, D.; Taube, H. J. Chem. Phys. 1962, 36, 416.
(50) Simonaitis, R.; Lissi, E., Heicklen, J. J. Geophys. Res. 1972, 77,
4248.
Acknowledgment. The experimental component of this
research was supported by the National Aeronautics and Space
Administration through Grant NAG5-8931 and by the National
Science Foundation through Grant ATM-99-10912, while the
modeling component was supported by the National Aeronautics
and Space Administration through Grant NAG1-2202. We thank
Dr. Robert E. Stickel for technical assistance with the TDLAS
system. We also thank Professor John Barker for helpful
discussions concerning energy transfer from highly vibrationally
excited molecules and Dr. Tom Slanger for helpful discussions
concerning O2(a1∆g) photochemistry. Finally, we thank Dr. Sheo
Prasad for communicating his most recent results to us prior to
publication.
References and Notes
(1) Minschwaner, K.; Salawitch, R. J.; McElroy, M. B. J. Geophys.
Res. 1993, 98, 10543.
(2) Dickinson, R. E.; Cicerone, R. J. Nature 1986, 319, 109.
(3) Prather, M.; Ehhalt, D.; Dentener, F.; Derwent, R.; Dlugokencky,
E.; Holland, E.; Isaksen, I.; Katima, J.; Kirchoff, V.; Matson, P. Midgley,
P.; Wang, M. Climate Change 2001: The Scientific Basis, Chapter 4;
Contribution of working group I to the third assessment report of the
Intergovernmental Panel on Climate Change; Houghton, J. T., Ding, Y.,
Griggs, D. J., Noguer, M., van der Linden, P. J., Dai, X., Maskell, K.,
Johnson, C. A., Eds.; Cambridge University Press: Cambridge, UK, 2001.
(4) Cantrell, C. A.; Shetter, R. E.; Calvert, J. G. J. Geophys. Res. 1994,
99, 3739.
(5) Garcia, R. R.; Solomon, S. J. Geophys. Res. 1994, 99, 12937.
(6) World Meteorological Organization (WMO). Scientific Assessment
of Ozone Depletion: 1998, Report No. 44; Global Ozone Research and
Monitoring Project: Geneva, Switzerland, 1999, and references therein.
(7) Mosier, A.; Kroeze, C. IGBP Global Change Newslett. 1998, June
(No. 34), 8-13.
(8) Kroeze, C.; Mosier, A.; Bouwman, L. Global Biogeochem. Cycles
1999, 13, 1.
(9) Olivier, J. G. J.; Bouwman, A. F.; van der Hoek, K. W.; Berdowski,
J. J. M. EnViron. Pollut. 1998, 102, 135.
(10) Zellner, R.; Hartmann, D.; Rosner, I. Ber. Bunsen-Ges. Phys. Chem.
1992, 96, 385.
(11) Delmdahl, R. F.; Gericke, K.-H. Chem. Phys. Lett. 1997, 281, 407.
(12) Zipf, E. C.; Prasad, S. S. Geophys. Res. Lett. 1998, 25, 4333.
(13) Cliff, S. S.; Thiemens, M. H. Science 1997, 278, 1774.
(14) Cliff, S. S.; Brenninkmeijer, C. A. M.; Thiemens, M. H. J. Geophys.
Res. 1999, 104, 16171.
(15) Yung, Y. L.; Miller, C. E. Science 1997, 278, 1778.
(16) Miller, C. E.; Yung, Y. L. Chemosphere: Global Change Sci. 2000,
2, 255.
(17) Rahn, T.; Zhang, H.; Wahlen, H.; Blake, G. A. Geophys. Res. Lett.
1998, 25, 4489.
(18) Rockmann, T.; Brenninkmeijer, C. A. M.; Wollenhaupt, M.;
Crowley, J. N.; Crutzen, P. J. Geophys. Res. Lett. 2000, 27, 1399.
(19) Turatti, F.; Griffith, D. W. T.; Wilson, S. R.; Rahn, T.; Zhang, H.;
Blake, G. A. Geophys. Res. Lett. 2000, 27, 2489.
(51) Hartland, G. V.; Qin, D.; Dai, H.-L. J. Chem. Phys. 1995, 102,
8677.
(52) Hartland, G. V.; Qin, D.; Dai, H.-L.; Chen, C. J. Chem. Phys. 1997,
107, 2890.
(53) Chimbayo, A.; Toselli, B. M.; Barker, J. R. J. Chem. Phys. 1998,
108, 2383.
(54) Noxon, J. F. J. Chem. Phys. 1970, 52, 1852.
(55) Biedenkapp, D.; Bair, E. J. J. Chem. Phys. 1970, 52, 6119.
(56) Snelling, D. R. Can J. Chem. 1974, 52, 257.
(57) Lee, L. C.; Slanger, T. G. J. Chem. Phys. 1978, 69, 4053.
(58) Sander, S. P.; Friedl, R. R.; DeMore, W. B.; Golden, D. M.; Kurylo,
M. J.; Hampson, R. F.; Huie, R. E.; Moortgat, G. K.; Ravishankara, A. R.;
Kolb, C. E.; Molina, M. J. Chemical Kinetics and Photochemical Data for
Use in Stratospheric Modeling, EValuation No. 13; Jet Propulsion Labora-
tory: Pasadena, CA, 2000.
(59) Ravishankara, A. R.; Dunlea, E. J.; Blitz, M. A.; Dillon, T. J.; Heard,
D. E.; Pilling, M. J.; Strekowski, R. S.; Nicovich, J. M.; Wine, P. H.
Geophys. Res. Lett., in press.
(60) Saxon, R. P.; Slanger, T. G. J. Geophys. Res. 1986, 91, 9877.
(61) Tully, J. C. J. Chem. Phys. 1974, 61, 61.
(20) Zhang, H.; Wennberg; P. O.; Wu, V. H.; Blake, G. A. Geophys.
Res. Lett. 2000, 27, 2481.
(21) Rockmann, T.; Kaiser, J.; Brenninkmeijer, C. A. M.; Crowley, J.
N.; Borchers, R.; Brand, W. A.; Crutzen, P. J. J. Geophys. Res. 2001, 106,
10403.
(22) Rahn, T.; Wahlen, M. Science 1997, 278, 1776.
(23) Griffith, D. W. T.; Toon, G. C.; Sen, B.; Blavier, J. F.; Toth, R. A.
Geophys. Res. Lett. 2000, 27, 2485.
(62) Zahr, G. E.; Preston, R. K.; Miller W. H. J. Chem. Phys. 1975, 62,
1127.
(63) Tachikawa, H.; Hamabayashi, T.; Yoshida, H. J. Phys. Chem. 1995,
99, 16630.
(64) Wang, H. J.; Cunnold, D. M.; Froidevaux, L.; Russell, J. M. J.
Geophys. Res. 1999, 104, 21629.
(65) Logan, J. A. J. Geophys. Res. 1999, 104, 16115.
(66) Ramaroson, R.; Pirre, M.; Cariolle, D. Ann. Geophys. 1992, 10,
416.
(67) Prasad, S. S. EOS Trans. AGU 2001, 82 (47), Fall Mtg. Suppl.,
Abstract A32B-0057.
(24) Anderson, S. M.; Mauersberger, K.; Morton, J. In Progress and
Problems in Atmospheric Chemistry; Barker, J. R., Ed.; World Scientific:
Singapore, 1995.
(25) Anderson, S. M.; Mauersberger, K.; Morton, J.; Schueler, B. ACS
Symp. Ser. 1992, 502, 155.