Reactions of NH2 with NO2 and of OH with NH2O

Article abstract of DOI:10.1021/jp004570x

The reaction system produced by 193 nm flash photolysis of a mixture of NH₃ and NO₂ was studied experimentally and modeled. The experiment divided roughly into two time scales. On a short time scale, reaction between H and NO₂ was complete in 12 μsec, and reaction between NH₂ and NO₂ was about 90% complete in 100 μsec. On the time scale of several hundred microseconds, OH decays and HNO rose through reaction between NH₂O and OH. The HNO was produced on a slower time scale than that of the NH₂ + NO reaction and formed by the reaction of OH with NH₂O. OH decay rate depended strongly upon the initial OH concentrations. The match between the observed concentrations and the model was excellent when the uncertainties in IR absorption cross sections, on whether the laser is exactly on the IR absorption peaks, and in the model, and the base time variations were considered.

Full text of DOI:10.1021/jp004570x

J. Phys. Chem. A 2001, 105, 6121-6128
6121
Reactions of NH2 with NO2 and of OH with NH2O
†
‡
§
F. Sun, J. D. DeSain, Graham Scott, P. Y. Hung, R. I. Thompson, G. P. Glass,* and
R. F. Curl*
Chemistry Department and Rice Quantum Institute, Rice UniVersity, Houston, Texas 77005
ReceiVed: December 22, 2000; In Final Form: April 2, 2001
The reaction system produced by 193 nm flash photolysis of a mixture of NH
experimentally and modeled. The accepted belief that only two channels are of significance for the reaction
between NH and NO , producing (a) N O and H O and (b) NH O and NO, is confirmed by the absence of
absorption signals and the absence of early HNO, as H and HNO are produced by two of the possible
five NH + NO channels. The fact that the OH concentration extrapolated to the flash is less than the initial
NH concentration indicates that the channel producing two OH molecules is not significant. HNO is observed
to be produced on a slower time scale than that of the NH + NO reaction and is believed to be formed by
the reaction of OH with NH O (OH is formed by the reaction of NO with H produced by the flash photolysis
of NH ). NH O does not appear to react with NO at 296 K on our time scale. Modeling of the reaction
system gives a rate for the reaction between NH
O and OH of 1.8(10) × 10 cm s . An excess continued
decay of OH at long times after NH O has virtually disappeared can be accounted for by reaction of OH with
HNO with a rate in the range (2-8) × 10 cm s .
3 2
and NO has been investigated
2
2
2
2
2
H O
2 2
2 2
O
2
2
2
2
2
2
3
2
2
-
10
3
-1
2
2
-
1
3
-1
Introduction
the most probable route to HNO + HNO involves at least one
1
intermediate that is energetically inaccessible.
The interest in the reaction between NH2 and NO2 arises
primarily from the work of Glarborg et al., who proposed that
this reaction produces a significant yield of NH2O and that this
species plays a key role in the thermal DeNox process (i.e.,
the selective noncatalytic reduction of NO by ammonia). The
reaction is also of some importance in the mechanism of
decomposition of the solid propellant ammonium dinitramide.
1
Since 1995, several measurements of the product branching
7
-11
ratio into channel 1a, k1a/k1, have been made.
The results
2
of these measurements vary between 0.14 and 0.59. However,
the two most recent measurements are in relatively good
10
agreement. In 1997, Park and Lin reported the branching ratio
as 0.19 ( 0.02 and determined that it did not vary significantly
as the temperature was raised from 300 to 910 K. In the same
3
,4
Several possible product channels of the title reaction are
1
1
thermodynamically accessible.1,5 Among these are
year, Lindholm and Hershberger obtained a branching ratio
of 0.24 ( 0.04 for reaction into channel 1a. These latter authors
also measured (by assuming that only channels 1a and 1b are
present) a branching ratio into the NH2O + NO channel, 1b, of
0.76 ( 0.1.
NH + NO f H O + N O (∆H ) -378 kJ/mol)
(1a)
(1b)
(1c)
2
2
2
2
f NH O + NO (∆H ) -67 kJ/mol)
2
Despite this agreement, several problems concerning the
overall mechanism of the reaction of NH2 with NO2 remain.
The only evidence for the dominant channel (1b) is the
appearance of NO as a final product, the theoretical calcula-
f H O + N (∆H ) -355 kJ/mol)
2
2
2
f OH + OH + N2 (∆H ) -140 kJ/mol)
f HNO + HNO (∆H ) -19 kJ/mol)
(
1d)
6
tions, and the chemical reasonableness of the postulated
(1e)
9
reaction. At room temperature, both Quandt and Hershberger
8,10
and Park and Lin observed significantly higher concentrations
of H2O than were predicted by the assumed mechanism. Also,
much of the secondary chemistry remains unexplored. In view
of this, we decided to re-examine the reaction by using infrared
kinetic spectroscopy.
On the basis of previous theoretical and experimental work, it
is generally believed that the dominant product channels of the
reaction are eqs 1a and 1b. An ab initio molecular orbital study
of the NH2 + NO2 potential energy surface concluded that both
channels 1c and 1d are inaccessible because the energies of the
transition states leading to them are higher than the energy of
the reactants. This conclusion is supported by the observation
6
Experimental Section
1
of Glarborg et al. that very little N2 is formed at temperatures
In these experiments, a mixture of NH3 and NO2 in the
presence of excess He buffer gas is flash photolyzed at 193 nm
and the time evolution of products observed by time-resolved
laser-based infrared absorption spectroscopy. A schematic
diagram of the apparatus is shown in Figure 1. A standard
multipass (“white”) infrared absorption cell was used. The IR
probe beam traveled back and forth throughout this cell in the
horizontal plane, while the UV photolysis beam was angled
between 850 and 1350 K in the thermal reaction of NH3 with
NO2. Reaction 1e is also thought to be unimportant because
†
Present address: Combustion Research Facility, Sandia National
Laboratories, Livermore, CA 94551-0969.
‡
Present address: Department of Molecular and Human Genetics, Baylor
College of Medicine, One Baylor Plaza, Houston, TX 77030.
§
Present address: Department of Physics and Astronomy, University
of Calgary, Calgary, Alberta T2N 1N4, Canada.
1
0.1021/jp004570x CCC: $20.00 © 2001 American Chemical Society
Published on Web 06/06/2001

6122 J. Phys. Chem. A, Vol. 105, No. 25, 2001
Sun et al.
(2)
densities by using the equation
Ae ) σNL
where Ae is the base e absorbance, σ is the peak infrared
absorption cross section, N is the species concentration in
3
molecules/cm , and L is the path length. Because of the way
that the infrared beams only partially overlap the excimer beam
in the multipass arrangement of Figure 1, the path length, L, is
not easy to calculate accurately and instead was measured in
the scheme described here.
An effective value for L was obtained by measuring the
recombination rate of methyl radicals and the observed CH3
absorbance. This scheme can be outlined as follows. The
reaction is
2
CH f C H
6
(3)
3
2
Figure 1. Infrared kinetic spectroscopy apparatus using a DFG source.
A second apparatus employed in this work uses a color center laser
source.
The concentration of CH3 as a function of time is given by the
standard second-order rate law expression.
TABLE 1: Transitions, Frequencies, and Peak Infrared
Absorption Cross Sections of Species Monitored
[
CH3]0
[
CH ] )
(4)
3
1
+ 2k [CH ] t
frequency (cm-1) σ
(×10 cm2)
18
3
3 0
species
transition
(7)a
0
p
0.37b
NH
NO
NH
OH
3
ν
ν
ν
3
1
1
s P
+ ν
6
3345.621
2915.854
3428.786
3407.989
2576.5415
3607.263
3693.4792
2632.8088
2658.7
However, the absorbance given by eq 4 is measured so that the
time behavior of the methyl signal is given by
c
2
3
164,13 r 154,12
F
1
0.61b
d
2
c
6
4,2 r 53,3
F
1
0.37
P(4,5)1+
4.37
4.69
0.75
0.087
c
c
A0
N
H
2
O
2ν
ν
1
R(16)
A(t) )
(5)
2
O
1
5
41 r 634
1
+ κA0t
c
NO
v 2 r 0 P22(e8.5)f
f
HNO
ν
ν
ν
1
2
3
5
14 r 523
1.81
g
r
with κ ) 2k3/(σ0L). Thus, by measuring κ, the path length can
H O
2 2
+ ν
6
Q
0
unknown
r
(2)h
i
be calculated from
CH
3
Q
0
3154.7468
3.00
a
From ref 12. b From ref 13. c From ref 14. All ref 14 cross sections
2
k3
d
assume a line width determined by Doppler broadening. From ref 15.
From ref 16. From ref 17. From ref 18. From ref 19. This work.
L )
(6)
e
f
g
h
i
κσ0
if k3 and σ0 are known. Under the experimental conditions, the
upward as it traversed the cell, so as to intersect the IR beam
over the approximately 35-cm section of the cell that is shaded.
20
-11
second-order rate constant for this reaction is k3 ) 3.7 × 10
3
-1
cm s . However, σ0 for CH3 is not known under the exact
conditions of these experiments and must be measured.
The peak absorption cross section, σ0, of the CH3 line at
Two different tunable infrared laser sources were used in this
investigation, a CW color center laser (Burleigh Instruments,
FCL-20, Li:RbCl crystal, 2.6-3.2 µm), pumped by a krypton
ion laser, and a newly developed laser based on difference
frequency generation (DFG) in periodically poled LiNbO3. The
DFG apparatus is diagrammed in Figure 1; the color center laser
apparatus is similar except that the DFG IR generation system
-
1
3
154.7468 cm was measured in the presence of 9 Torr of He
by flash photolyzing N2O (272 mTorr) at 193 nm in the presence
of CH4 (1.82 Torr), initiating the next series of reactions.
1
N O + hν (193 nm) f N + O( D)
(7)
+
2
2
is replaced by a color center laser pumped by a Kr laser.
Reaction mixtures typically contained 3-10 mTorr of NH3,
1
O( D) + CH f OH + CH
(8)
(9)
4
3
3
0 mTorr of NO2, and 9 Torr of He. Under normal operating
conditions, 10-15% of the ammonia in the excimer beam was
dissociated by the laser pulse.
2
CH f C H
3 2 6
To avoid depletion of reagents and buildup of reaction
products in the photolysis cell, the transit time through the cell
was decreased by increasing the pumping speed and by
decreasing the diameter of the glass tube forming the cell wall.
This ensured that the reaction mixture was exposed to only one
photolysis pulse during transit through the cell. The depletion
by the first shot and the subsequent chemistry was 10-20%
for NH3 and 5-10% for NO2, and the buildup of NO per shot
was about 5-10% of the original NO2 concentration.
CH + OH f products
(10)
3
Because [CH3]0 ) [OH]0, the CH3 peak absorption cross section
can be obtained by comparison of the CH3 signal with the OH
signal at t ) 0, using the known Doppler cross section of OH.
The 1/e half width of OH found in these experiments is 190
14
(
MHz, which is quite close to the Doppler width of 183 MHz.)
-1
To do this, frequency scans of the OH line at 3407.989 cm
-1
and of the CH3 line at 3154.7468 cm at a series of points
after the flash were obtained and fitted with a Gaussian profile
as illustrated in Figure 2, giving both the peak intensities and
the line widths. The resulting peak intensities are then plotted
versus time and extrapolated to t ) 0 as shown in Figure 3.
The peak absorption cross section for CH3 is calculated from
Table 1 lists the identity of the species monitored during this
investigation, the particular infrared transitions that were
monitored,1
2-19
the frequencies (cm ) of these transitions, and
-1
the peak infrared absorption cross sections (excepting H2O2).
Infrared absorbance measurements were converted into column

Reactions of NH2 with NO2 and of OH with NH2O
J. Phys. Chem. A, Vol. 105, No. 25, 2001 6123
Figure 4. Reciprocals of the CH
3
absorbances plotted versus time
during the decay of CH by recombination. Ptot ) 10 Torr.
3
a Gaussian profile. The reciprocals of the peak intensities were
plotted versus time to obtain κ as shown in Figure 4. The
effective path length L obtained from eq 6 was 1050 cm, or 33
cm per pass in 32 passes in a 1 m white cell with partial overlap
of the excimer and IR beams. An initial CH3 concentration of
1
3
-3
about 5 × 10 cm was observed.
To confirm that NH2 is formed quantitatively by photolysis
of ammonia, an initial series of experiments was performed in
the absence of NO2. NH2 was produced with a yield of 1.0 (
-
1
Figure 2. Frequency scans of the OH line at 3407.989 cm and the
-
1
CH
at 193 nm of a mixture of CH
mTorr), and He (9 Torr). The times are 10-20, 30-40, 50-60, 100-
10, 200-210, and 400-410 µs.
3
line at 3154.7468 cm at various times after the photolysis flash
4
(1.82 Torr), N O (272 mTorr), NO (32
2
0
.1 when NH3 (∼10 mTorr) was photolyzed, and it decayed
very slowly, exhibiting a half-life of about 10 ms.
1
Methodology
NH2 radicals were produced by photolysis of ammonia at
1
93 nm using an ArF excimer laser operating at pulse energies
of 50-100 mJ
NH + hν f NH + H
(12)
3
2
while NO2 is simultaneously photolyzed in a much smaller yield.
1
NO + hν f ΝÃ + Ã( D)
(13a)
(13b)
2
3
f ΝÃ + Ã( P)
The high-resolution transient infrared absorption peaks of a
number of reactants, intermediates, and products were monitored
throughout the ensuing reaction using tunable infrared laser
sources. A new difference frequency laser source was used that
allowed all of the following species to be monitored (although
NH, H2O2, and HONO were not in fact observed in this
system): NH3, NH2, NH, NO2, NO, N2O, HNO, H2O, OH,
H2O2, and HONO.
The H atoms produced by the NH3 photolysis react very
rapidly (1/e ) 6 µs) with NO2 to produce NO and OH.
Figure 3. Extrapolations of the peak signals for OH and CH
3
from
NO + H f NO + OH
(14)
2
Figure 2 plotted versus time and extrapolated to t ) 0. Total IR power
signal 2.5 V for OH and 2.6 V for CH
3
.
The stoichiometric production of OH proved to be quite
beneficial to the study, because the reaction of OH with NH2O
to produce HNO and H2O provided indirect but powerful
evidence that NH2O is in fact being produced in high yield.
Because the 193 nm UV absorption cross section of NO2 is
the equation
∆
V(OH)Doppler A (CH )
0 3
σ (CH ) ) σ (OH)
Doppler
(11)
0
3
0
-
19
2 21,22
∆
V(OH)obs A0(OH)
relatively small (3 × 10 cm )
and the absorption cross
-17
2 23,24
section of NH3 at 193 nm is large (10 cm ),
it is possible
-
18
2
to be 3.0 × 10
25 MHz).
Methyl radicals were produced in 9 Torr of He by the 193
cm at 9 Torr of He (CH3 1/e halfwidth )
to produce NH2 + H without substantial interference from the
photolysis of NO2, even with a severalfold excess of NO2 over
NH3. The photolysis products of NO2, NO, O( P), and O( D),
do not interfere significantly with the study of the NH2 + NO2
reaction. NO reacts rapidly with NH2 but cannot compete with
2
3
1
nm flash photolysis of acetone (∼100 mTorr), and similar
frequency scans of the CH3 line were obtained and fitted with

6124 J. Phys. Chem. A, Vol. 105, No. 25, 2001
Sun et al.
sequence responsible for the slow decay of OH through the
reaction
HNO + OH f ΝÃ + Η O
(17)
2
1
8
Glarborg et al. and Park and Lin use a value for k17 of 6 ×
-
11
3
-1
1
0
cm s in their modeling. H2O is produced by reaction
1a and by both steps of the two step secondary reaction sequence
(16 and 17) responsible for the slow decay of OH. Much of
this report is concerned with the interpretation and analysis of
the slow processes.
The aim of this work was twofold: to obtain more direct
evidence for the production of NH2O in high yield and to obtain
experimental evidence about the existence of channels 1c, 1d,
and 1e, which had been found theoretically not to be open.6
The observation of a vibrational transition of NH2O combined
with an analysis of its rotational structure would be convincing
evidence for the presence of this species. However, a scan of
the region 2800-3500 cm did not reveal any spectrum that
could be ascribed to NH2O. This was somewhat surprising, as
a quantum chemical calculation of the infrared intensities (and
harmonic frequencies) at the B3LYP 3-311++g(2df,2pd) level
indicated that the B2 symmetry NH stretch has an intensity of
Figure 5. Time dependence of the concentrations of all the species
observed. Conditions are pNO2 ) 30 mTorr, pNH3 ) 10 mTorr, and ptot
)
9 Torr using He buffer gas. Excimer energy ) 90 mJ.
the large excess of NO2 as long as flow rates are large enough
and repetition rates are small enough to avoid product buildup.
-
1
3
1
O( P) reacts with NO2, producing O2 and NO. O( D) also reacts
in this way and, in addition, reacts with NH3, producing NH2
and OH.
1
O( D) + NH f OH + NH
(15)
3
2
7
.2 km/mol (actually, the calculation produced a nonplanar
25
structure, but the microwave spectrum analysis found a planar
C2V structure). This is weak, but we believe it might be
detectable with our sensitivity. The A1 symmetric NH stretch
was predicted to have an intensity of 0 km/mol within reasonable
error limits. The harmonic frequencies were, for A , 3424 cm
and, for B , 3556 cm .
This reaction has the same net products as the photolysis of
NH3.
Two kinds of data were obtained in this study: (1) time scans
in which the probe frequency was tuned at or near the absorption
maximum and (2) frequency scans in which the probe frequency
was scanned over the line and data were acquired at six time
intervals after the flash. The time scans provide more data, more
rapid data collection, and better S/N for a given collection time.
The frequency scans once analyzed remove any time variation
in the baseline, remove any uncertainty about whether the probe
is tuned to the peak of the absorption line, and simultaneously
provide line widths. Each time window in a frequency scan is
fitted with a Gaussian to obtain the peak absorbance and the
line width. Figure 2 already illustrates this process.
-1
1
-
1
2
We made efforts to check on the existence of channels 1c,
1
1d, and 1e. Glarborg et al. report that N as determined by a
2
nitrogen balance is less than 10% at 900 K, increasing to about
30% at 1350 K. This suggests that 1c and 1d should be
negligible at room temperature. We searched for but did not
observe H O . This is only mild evidence that the 1c channel
2
2
is not present, because the infrared absorption cross section for
the H O line searched for is unknown (see Table 1). Because
2
2
so much OH is produced so rapidly in our system, we are unable
to verify independently that 1d is unimportant. Regarding
channel 1e, we did observe the formation of HNO. However,
HNO forms much more slowly than NH2 disappears, actually
displaying an induction period. Channel 1e is approximately
Observations and Results
This reaction provides a unique opportunity for monitoring
reaction products. It was possible to observe the NH2 and NO2
reactants, H2O, N2O, and NO products, and OH produced by
reactions 14 and 15. In addition, HNO, which we believe is
produced by a secondary reaction
-
1
1
600 cm exoergic, and the lowest vibrational mode of HNO
26
-1
is at 1500 cm . This channel involves a complete rearrange-
ment of the atoms so that there is reason to expect a statistical
distribution of the reaction energy between the two HNOs
produced. Thus, the slowly rising HNO signal must be from
other sources (to be discussed later), and any major contribution
from channel 1e is ruled out. Because the absorption cross
section for the HNO line monitored is one of the largest of the
species we monitored, this conclusion is on rather firm ground.
The slow decay of the OH signal and the slow rise of the
HNO signal provide strong indirect evidence for the formation
NH O + OH f ΗΝÃ + Η O
(16)
2
2
could also be observed. Figure 5 shows the time behavior of
the various species observed. It must be kept in mind that
products can be produced in excited vibrational states, which
have to be quenched to the ground vibrational level to be
monitored. This is clearly the case for H2O which exhibits an
initial induction period.
of large amounts of NH O in this system. We are convinced
2
that these time variations arise from the rapid reaction (16)
This experiment divides roughly into two time scales. On
the short time scale, reaction 14 between H and NO2 is complete
in 12 µs, and reaction 1 between NH2 and NO2 is about 90%
complete in 100 µs. Thus, by 100 µs, H has disappeared, and
the decay of NH2 and the rise of N2O are substantially complete.
Then on a time scale of several hundred microseconds, OH
decays and HNO rises through reaction 16. The time behavior
of NO and H2O is mixed. NO is produced by the photolysis of
NO2, by reactions 1b and 14, and, we believe, by the reaction
between NH2O and OH and by a slower reaction (17) between
OH and HNO. We shall present evidence that NH2O does not
react rapidly with NO2, and it is obvious from Figure 5 that
HNO does not react with NO2 or any reaction products. By
analyzing these decay profiles, rates for reactions 15 and 16
can be estimated. However, reasonably accurate OH time
profiles are required in order to obtain information on these
rates. First, as accurate data as possible were collected for the
earlier times when reaction 16 is dominant. For this purpose,

Reactions of NH2 with NO2 and of OH with NH2O
J. Phys. Chem. A, Vol. 105, No. 25, 2001 6125
TABLE 2: Reactions and Rate Constants Used in the
Kinetic Modeling
reaction
+ NO f H
+ NO f ΝΗ
rate (cm3 s-1
)
-
-
-
-
-
12
11
10
10
11
1
1
1
1
1
1
1
2
2
2
2
2
2
a. NH
2
2
2
O + N
2
2
O
4.80 × 10
1.52 × 10
1.20 × 10
1.80 × 10
5.00 × 10
4.60 × 10
1.00 × 10
1.60 × 10
6.40 × 10
9.70 × 10
1.44 × 10
1.60 × 10
1.50 × 10
b. NH
4. NO
6. NH
2
2
2
2
O + NO
+ H f ΝÃ + ÃΗ
2
O + OH f ΗΝÃ + Η O
7. HNO + OH f ΝÃ + Η O
2
a
-13 a
8. OH + NO
9a. OH + NH
2
f ΗΝÃ
f à + NH
f H O + NH
f ΗΝÃ + Η
f ΝÃ + Ã
O + N
3
-
-
13
13
2
3
0. OH + NH
3
2
2
3
-11
-12
1. O( P) + NH
2. O( P) + NO
2
3
2
2
-
-
11
12
3a. NH
3b. NH
2
+ NO f Η
2
2
Figure 6. Concentrations as a function of time of four OH frequency
scan analyses. These were obtained with pNH3 ) 10 mTorr, pNO2 ) 30
mTorr, and ptot ) 9 Torr using He buffer gas.
2
+ NO f ÃΗ + Η + Ν
2
a
-13
4. OH + NO f HONO
a
The second-order rate constant of this termolecular reaction under
the assumed conditions.
frequency scans were obtained and analyzed. Then a very crude
estimate of the rate of reaction 17 is obtained by fitting the
long time behavior of time scans.
effect on the OH decay. In as far as we have been able to
ascertain, the rate of reaction 19b has never been measured,
and we have omitted it. The OH decay is not very sensitive to
this rate, as NH2 reacts primarily with the large excess of NO2
and disappears early in the reaction. The reaction
The first step illustrates the real difficulties of rate measure-
ments of reactions between reaction products. Data of the kind
illustrated in Figure 2 must be converted into OH concentrations
at each of the times given. The observed line widths (1/e HW)
for all of the data to be analyzed here range from 182 to 197
MHz, and the calculated Doppler width is 183.4 MHz at 296
K. Thus, line broadening from pressure broadening and unre-
solved hyperfine structure is small and is neglected in converting
the tabulated16 integrated absorption line cross section into a
peak cross section. Using this peak absorption cross section,
the OH data from such frequency scans can be directly converted
into column density and, by the use of the measured path length,
into concentration. Figure 6 shows the results of such an analysis
for four OH time-resolved frequency scans taken sequentially.
In this figure, there seems to be a general decrease in the
OH concentration with time, as might be expected if the excimer
power had drifted down or the NH3 flow rate had drifted down.
However, the decay rate of OH does not share this trend, and
we shall see that the OH decay rate depends strongly upon the
initial OH concentration. It seems likely that the scatter in Figure
OH + NH f H O + NH
2
(20)
3
2
-
13
between NH3 and OH was included as its rate is 1.6 × 10
cm s at room temperature.
3
-1
5
A very minor addition to the model was the inclusion of the
effect of 193 nm photolysis of NO2. The ratio of the 193 nm
2
2
absorption cross section of NO2 to that of NH3 is 0.031. The
photolysis of NO2 produces NO and O atoms with slightly more
1
22
than half of the O atoms in the O( D) excited state, with R )
.55 ( 0.03.
0
1
3
NO + hν (193 nm) f NO + RO( D) + (1-R)O( P) (13)
2
1
O( D) will disappear from the system in, at most, a few
3
microseconds either by reaction or quenching. In contrast, O( P)
reacts away quite slowly by two reactions
6
is random. Therefore, the four data sets were averaged for
further treatment of the time decay. On the same day that these
OH data were collected, the NH2 signal in the absence of NO2
3
O( P) + NH f HNO + H
(21)
2
1
3
was also measured from three such graph sets to be 5.0 × 10
3
-
3
O( P) + NO
2
f NO + O
2
(22)
cm . Subsequent modeling calculations indicate that this
number is about 14% too high to match the OH data. The fact
that the NH2 initial concentration is higher than that needed to
model the OH data indicates that reaction channel 1d is not
significant. The measurement of [NH2] is more difficult than
29,5
-11
Reactions 21 and 22 have rate constants of k21 ) 6.4 × 10
-
11
3
-1
1
and k22 ) 1 × 10
cm s . As remarked earlier, O( D) can
react rapidly with either NH3 or NO2. It can also be quenched
by collisions with either of these gases or the He buffer gas.
The extent of reaction versus quenching is difficult to estimate.
Fortunately, it does not matter, as the results we obtain modeling
the system assuming that all O( D) is quenched to O( P) are
imperceptibly different from those omitting reactions 13, 21,
and 22.
1
3
that of [OH] because the signal is weaker, the cross sections
are probably less accurate, and the measurement was done in
the absence of NO2; therefore, only the OH signal measurements
were used in modeling.
For modeling the kinetics in addition to the reactions given
here, the reactions
1
3
However, the reaction
OH + NO + M f HNO + M
(18)
(19a)
(19b)
2
3
NH + NO f H O + N
(23a)
(23b)
2
2
2
2
OH + NH f à + ΝΗ3
f ÃΗ + Η + Ν2
with28 k23a ) 1.44 × 10
-11
and k23b ) 1.6 × 10
-12
cm s
3
-1
OH + NH f H O + NH
2
2
does have a small but perceptible effect on the modeling. There
is some controversy about whether the products of 23b should
be as written above or should be OH + HN2. The advantage of
the choice made is that the rate constants for the reactions of H
are known, while those of HN2 are more uncertain. Because
need to be included. Reaction 18 has been extensively studied.
2
7
-13
3
-1
The value of 4.6 × 10
cm s at 9 Torr of He was used
28
-13
3
for its rate (Table 2). The rate of reaction 19a is 10 cm
at room temperature, which has a completely negligible
-
1
s

6126 J. Phys. Chem. A, Vol. 105, No. 25, 2001
Sun et al.
Figure 7. Average [OH] values from the data of Figure 6 are plotted
as six points. The dotted curve near the points represents the expected
time behavior from the model used and the fitted rate constants k16
)
-
10
3
-1
-11
3 -1
1
.8 × 10 cm s and k17 ) 5 × 10 cm s . The upper dash-dot
Figure 8. Time scan of [OH] concentration under the same conditions
of initial concentrations. The vertical scale of this figure was adjusted
to give the best fit because the probe laser may not have been set on
the exact peak of the OH line. The rescaled experimental data are the
heavy solid line and the dotted curve almost exactly superimposed upon
-10 3 -1 3 -1
line corresponds to k16 ) 1.8 × 10 cm s and k17 ) 0 cm s , the
-
10
3
-1
lower dashed line corresponds to k16 ) 1.8 × 10 cm s and k17
)
-
11
3
-1
9
1
× 10
cm s . The long dash line corresponds to k16 ) 2.5 ×
-1 -11 3 -1
-
10
3
0
cm s and k17 ) 2 × 10 cm s . The upper solid line far
from the points corresponds to k16 ) 0 (and k17 ) 0).
-10
3
-1
it uses the fitted rate constants k16 ) 1.8 × 10 cm s and k17 ) 5
-
11
3
-1
×
10 cm s . The upper dash-dot line corresponds to k16 ) 1.8 ×
the dominant channel is 23a, we believe that the choice of
products of 23a cannot affect the results. We have also included
the minor reaction
-10
3 -1 3 -1
1
0
cm s and k17 ) 0 cm s . The long dash line directly below
3 3
-10
-1
-11
it corresponds to k16 ) 2.5 × 10 cm s and k17 ) 2 × 10 cm
-
1
-10
3
-1
s
. The lower dashed line corresponds to k16 ) 1.8 × 10 cm s
-11 3 -1
and k17 ) 9 × 10 cm s . The upper solid line far from the points
OH + NO + M f HONO + M
(24)
cm3
corresponds to k16 ) 0 (and k17 ) 0).
-
13
with an effective two body rate constant of 1.5 × 10
-1
30
s . Anderson, Margitan, and Kaufman reported the rates of
reactions 18 and 24 in the same paper, finding that reaction 24
has a rate constant approximately one-third that of reaction 18.
We therefore divided the rate of reaction 18 by 3 to obtain the
rate of reaction 24.
Figures 7-9 show the results of modeling the OH decay by
including reactions 16 and 17. Figure 7 shows the fit of the
average of the OH data points shown in Figure 6. We regard
these data as the most reliable, because they cannot be affected
by baseline problems and because the center of the OH
absorption peak is determined precisely. Figure 8 shows a fit
of an OH time decay under similar conditions to Figure 7. These
data were acquired several months before the data of Figure 7;
the gas flow conditions should have been identical, and the
excimer laser pulse energy was similar. We do not have a
satisfactory explanation for the difference in the peak OH
concentrations of Figures 7 and 8; possibly the alignment
between the excimer and IR probe beams was different. Figure
9
shows the OH decay taken within a few minutes of that of
Figure 9. Time scan of [OH] with the initial partial pressure of NH
3
Figure 8, but with an NH3 concentration only about 10% of
that of Figure 8. (The ratio of the two NH3 concentrations could
not be determined very precisely, because the flow controller
was operating near its lower limit for the low NH3 flow of Figure
reduced by about a factor of 10. The initial partial pressures of NO2
and He were the same as in Figure 7, and the excimer pulse energy
and probe laser frequency were the same as Figure 8. The vertical scale
is the same as in Figure 8. The experimental data are recognizable by
the noise. The dotted curve uses the fitted rate constants k16 ) 1.8 ×
9
.) The OH concentration scale of Figures 8 and 9 could not be
-10
3
-1
-11
3
-1
1
0
cm s and k17 ) 5 × 10 cm s . The upper dash-dot line
determined directly from the observed signal with complete
reliability because of the uncertainty in setting the lasers to the
OH peak. Instead, it was adjusted in the fitting process.
However, Figures 8 and 9 are expected to share the same vertical
scale, as the lasers were not adjusted during the data collection.
Therefore, the factor used to convert absorbance to OH
concentration is the same for both Figures 8 and 9.
-10 3 -1 3 -1
corresponds to k16 ) 1.8 × 10 cm s and k17 ) 0 cm s , and the
-
10
3
-1
lower long dash line corresponds to k ) 2.5 × 10 cm s and k17
1
6
-
11
3
-1
) 2 × 10 cm s . The dashed line between the long dash line and
-
10
3
-1
the dotted line corresponds to k16 ) 1.8 × 10 cm s and k17 ) 9
-
11
3
-1
×
10 cm s . The upper solid line far from the points corresponds
to k16 ) 0 (and k17 ) 0).
valuable to examine the time scans (see Methodology). Because
of the uncertainty about whether the laser was tuned to the peak
of the OH absorption line, the vertical scale of Figure 8 was
The data of Figure 7 were used to determine k16 and limit
k17. To obtain a complete overview of the time behavior, it is

Reactions of NH2 with NO2 and of OH with NH2O
J. Phys. Chem. A, Vol. 105, No. 25, 2001 6127
determined by adjusting [NH2]0 ) [OH]0 to get the best match
of its decay for the first few hundred microseconds using the
k16 and k17 values determined from the data of Figure 7. In
-
10
3
Figure 7, the dotted line corresponds to k16 ) 1.8 × 10
cm
-1
-11
3
-1
s
and k17 ) 5 × 10 cm s . We believe it represents the
-
11
best compromise fit of the data. This value for k17 (5 × 10
3
-1
-11
3 -1
cm s ) is close to the value of 6 × 10
cm s previously
1
8
assumed by Glarborg et al. and Park and Lin. The dashed
-
10
3
-1
line in Figure 7 has k16 ) 1.8 × 10
cm s . The dash-dot line has k16 ) 1.8 × 10
and k17 ) 0. The solid line at the top, far from the data,
assumes k16 ) 0. Clearly k16 is needed. The long dash line has
cm s and k17 ) 9 ×
-
11
3
-1
-10
3
1
s
0
cm
-
1
-
10
3
-1
-11
3
-1
k16 ) 2 × 10
cm s and k17 ) 2 × 10
cm s . It fits
the data of Figure 7 slightly better than the dotted curve but
fits poorly to the data of Figure 8.
A similar choice of line types is used in Figures 8 and 9 as
was used in Figure 7. In Figure 8, the model curve (a dotted
-
10
3
-1
line) corresponding to k16 ) 1.8 × 10
cm s and k17 ) 5
-11
3 -1
×
10 cm s is almost exactly superimposed on the rescaled
experimental data (a thick dashed line). Figure 8 indicates that
the long time behavior is most affected by the value k17. The
data for Figure 9 were taken without adjusting the lasers, and
thus Figure 9 provides a test of the scaling of Figure 8 and the
value of k16. Because the concentration of the radical pool is so
low that even NH2O is reacting away slowly, reaction 17 plays
almost no role in the OH decay observed in Figure 9. All models
give an acceptable fit to the data in Figure 9 with the preferred
Figure 10. Effect of changing the NO
2
concentration by a factor of 2
upon the rate of OH decay. The dashed line corresponds to an NO
2
14
3
concentration of 5.9 × 10 molecules/cm and the solid line to [NO
2
]
15
3
)
1.17 × 10 molecules/cm .
-
10
3
-1
-11
3
model (k16 ) 1.8 × 10
cm s and k17 ) 5 × 10
cm
-1
s ), perhaps paralleling the decay curve slightly more ac-
curately.
-
10
3
-1
We conclude that k16 is about 1.8(10) × 10
cm s and
that reaction 17 proceeds at an appreciable rate with k17
-
11
-11
3
-1
somewhere between 2 × 10
our best guess of 5 × 10
and 8 × 10
cm s .
cm s , with
-
11
3
-1
In all this, it is assumed that there is no reaction at room
temperature between NH2O and NO2. Yet such a reaction
seemed likely on the basis of chemical intuition. In this case,
intuition proved unreliable. No effects which could be ascribable
to such a reaction were observed when the system was
investigated at different NO2 concentrations. The results of
varying the NO2 concentration upon the OH decay are shown
in Figure 10. Two effects are observed: the peak of the OH
signal moves to shorter times when [NO2] is increased, and the
Figure 11. Concentrations of the various species resulting from the
-
10
3
-1
-11
3
-1
model (k16 ) 1.8 × 10 cm s and k17 ) 5 × 10 cm s ) cast
in terms of percent yield, where 100% corresponds to [NH . In this
and OH produced by the reaction of O( D) with NH
was assumed to be produced almost instantaneously and lumped into
[OH] signal at a very long time decreases with increasing [NO2].
2 0
]
Both effects can be explained without bringing in a reaction
between NH2O and NO2. The time of the maximum [OH] occurs
when the rate of production of OH equals the rate of its loss.
Increased NO2 shortens the time required for OH production
by reaction 14, causing the OH production rate to fall off earlier,
and it shortens the time required to produce NH2O by reaction
1
model, any NH
2
3
the NH
are [NH
2
produced by photolysis. The initial concentrations assumed
13
14
2
] ) [H] ) 4.1 × 10 , [NO
2
-3
] ) 9.79 × 10 , [NH
3
] ) 4.4 ×
14
3
12
1
0 , and [O( P)] ) 3.47 × 10 cm
.
which removes HNO. At larger NO2 concentrations, OH at long
times is smaller because of the three body reaction (18).
Figure 11 shows the time variations of the concentrations of
the various species observed and depicted in Figure 5. We
believe the match between the observed concentrations and the
model is excellent when the uncertainties in infrared absorption
cross sections, the baseline time variations, the uncertainty about
whether the laser is exactly on the IR absorption peaks, and
the uncertainties in the model are considered.
1
b, causing the OH loss rate to peak earlier. The [OH] signal
falls off more rapidly at long times with increasing [NO2]
because a higher [NO2] increases the rate of reaction 18. Our
modeling of the reaction predicts these changes in the OH decay
curve when the NO2 concentration is changed. Figures 9 and
1
0 demonstrate that, if NH2O reacts with NO2, the products of
that reaction must react with OH with a near gas kinetic rate
constant. Moreover, the yield of HNO actually increases very
slightly with more NO2. To explain this, the hypothetical
products of the reaction of OH with the products of the reaction
of NO2 with NH2O must react with OH to produce HNO with
unit efficiency. The small increase of HNO with increased NO2
is explainable by a reduction in the reaction of HNO with OH
The yield of N O from reaction 1a has been measured
2
7
-11
previously several times.
Under the conditions of our
-
18
experiment, we obtain an N O cross section value of 3 × 10
2
2
-18
cm , which agrees fairly well with that calculated (3.3 × 10
2
13
cm ) from the HITRAN value for the integrated line strength

6128 J. Phys. Chem. A, Vol. 105, No. 25, 2001
Sun et al.
when our measured 1/e half line widths of 214 MHz are used.
The Doppler 1/e halfwidth is 143.5 MHz, which would yield a
(8) Park, J.; Lin, M. C. Int. J. Chem. Kinet. 1996, 28, 879.
9) Quandt, R. W.; Hershberger, J. F. J. Phys. Chem. 1996, 100, 9407.
(10) Park, J.; Lin, M. C. J. Phys. Chem. A 1997, 101, 2643.
(
-
18
2
-18
2
σ0 of 4.69 × 10
cm . Our cross section (3 × 10
cm )
(11) Lindholm, N.; Hershberger, J. F. J. Phys. Chem. A 1997, 101, 4991.
gives a yield for N2O of 0.27(5), and the use of HITRAN with
(12) Angstl, R.; Finsterh o¨ zl, H.; Funder, H.; Illig, D.; Papousek, D.;
-
18
2
our observed line widths (3.3 × 10
cm ) gives a yield of
Pracna, P.; Rao, K. N.; Schr o¨ tter, H. W.; Urban, S. J. Mol. Spectrosc. 1985,
114, 454.
(13) Stephens, J. W.; Morter, C. L.; Farhat, S. K.; Glass, G. P.; Curl, R.
F. J. Phys. Chem. 1993, 97, 8944.
0
.24(5). These numbers are to be compared with the results of
Park and Lin of 0.19 and with the last measurements11 from
1
0
Hershberger’s group of 0.24. We cannot claim to have improved
on previous measurements of this quantity. Our results agree
well with the last from Hershberger.
(14) Rothman, L. S.; Rinsland, C. P.; Goldman, A.; Massie, S. T.;
Edwards, D. P.; Flaud, J. M.; Perrin, A.; Camy-Peyret, C.; Dana, V.; Mandin,
J. Y.; Schroeder, J.; Mccann, A.; Gamache, R. R.; Wattson, R. B.; Yoshino,
K.; Chance, K. V.; Jucks, K. W.; Brown, L. R.; Nemtchinov, V.; Varanasi,
P. J. Quant. Spectrosc. Radiat. Transfer 1998, 60, 665.
Discussion
(
15) Amano, T.; Bernath, P. F.; McKellar, A. R. W. J. Mol. Spectrosc.
982, 94, 100.
16) Johns, J. W. C.; McKellar, A. R. W.; Weinberger, E. Can. J. Phys.
983, 61, 1106.
17) Unfried, K. G.; Glass, G. P.; Curl, R. F. Chem. Phys. Lett. 1990,
173, 337.
(18) Reddington, R. L.; Olson, W. B.; Cross, P. C. J. Chem. Phys. 1962,
6, 1311.
19) Amano, T.; Bernath, P. F.; Yamada, C.; Endo, Y.; Hirota, E. J.
Chem. Phys. 1982, 77, 5284.
20) Slagle, I. R.; Gutman, D.;. Davies, J. W.; Pilling, M. J. J. Phys.
1
1
We cannot find any other way to explain either the rapid
decay of OH, the appearance of HNO, or the high yield of H2O
in this system other than by the rapid reaction (16) of OH with
NH2O to produce HNO and H2O. The necessity for the reaction
(
(
(17) between HNO and OH is on less firm footing. It was
3
introduced into the model to explain the continued excessive
decay of OH at long times after the [NH2O] becomes too low
for reaction 16 to remove OH at a significant rate. The fact
that reaction 17 is only marginally needed to fit the OH decay
is reflected in the large uncertainty in its rate constant, k17.
(
(
Chem. 1988, 92, 2455.
(21) Schneider, W.; Moortgat, G.; Tyndall, G. S.; Burrows, J. P. J.
Photochem. Photobiol. 1987, A40, 195.
(
22) Sun, F.; Hung, P.-Y.; Glass, G. P.; Curl, R. F. Chem. Phys. Lett.
Acknowledgment. This work was supported by grants from
the Department of Energy and the Robert A. Welch Foundation.
2
001, 337, 72.
(23) Suto, M.; Lee, L. C. J. Chem. Phys. 1983, 78, 4515.
(
24) Davidson, D. F.; Chang, A. Y.; Kohse-Hoinghaus, K.; Hansen, R.
References and Notes
K. J. Quant. Spectrosc. Radiat. Transfer 1989, 42, 267.
25) Mikami, H.; Saito, S.; Yamamoto, S. J. Chem. Phys. 1991, 94,
415.
(
(
1) Glarborg, P.; Dam-Johansen, K.; Miller, J. A. Int. J. Chem. Kinet.
3
1
1
995, 27, 1207.
(26) Johns, J. W. C.; McKellar, A. R. W. J. Chem. Phys. 1977, 66,
(
2) Lyon, R. K. U.S. Patent 3, 900, 559, 1975.
217.
(3) Brill, T. B.; Brush, P. J.; Patil, D. G. Combust. Flame 1993, 92,
(
27) Forster, R.; Frost, M.; Fulle, D.; Hamann, H. F.; Hippler, H.;
Schlepegrell, A.; Troe, J. J. Chem. Phys. 1995, 103, 2949.
28) Baulch, D. L.; Cobos, C. J.; Cox, R. A.; Esser, C.; Frank, P.; Just,
1
78.
(
4) Mebel, A. M.; Lin, M. C.; Morokuma, K.; Melius, C. F. J. Phys.
(
Chem. 1995, 99, 6842.
5) Atkinson, R.; Baulch, D. L.; Cox, R. A.; Hampson, R. F., Jr.; Kerr,
J. A.; Rossi, M. J.; Troe, J. J. Phys. Chem. Ref. Data 1997, 26, 1329.
6) Mebel, A. M.; Hsu, C. C.; Lin, M. C.; Morokuma, K. J. Chem.
Phys. 1995, 103, 5640.
7) Meurier, H.; Pagsberg, P.; Sillesen, A. Chem. Phys. Lett. 1996,
61, 277.
T.; Kerr, J. A.; Pilling, M. J.; Troe, J.; Walker, R. W.; Warnatz, J. J. Phys.
Chem. Ref. Data 1992, 21, 411.
(29) Adamson, J. D.; Farhat, S. K.; Morter, C. L.; Glass, G. P.; Curl, R.
F.; Phillips, L. F. J. Phys. Chem. 1994, 98, 5665.
(30) Anderson, J. G.; Margitan, J. J.; Kaufman, F. J. Chem. Phys. 1974,
60, 3310.
(
(
(
2