HNO3 forming channel of the HO2 + NO reaction as a function of pressure and temperature in the ranges of 72-600 torr and 223-323 K

Article abstract of DOI:10.1021/jp074117m

A high-pressure turbulent flow reactor coupled with a chemical ionization mass-spectrometer was used to determine the branching ratio of the HO ₂ + NO reaction: HO₂ + NO -OH + NO₂ (1a), HO₂ + N-O → HNO₃ (1b). The branching ratio, β-k_1b/k_1a, was derived from the measurements of chemically amplified concentrations of the NO₂ and HNO₃ products in the presence of O₂ and CO. The pressure and temperature dependence of βwas determined in the pressure range of 72-600 Torr of N₂ carrier gas between 323 and 223 K. At each pressure, the branching ratio was found to increase with the decrease of temperature, the increase becoming less pronounced with the increase of pressure. In the 298-223 K range, the data could be fitted by the expression: β(T,P) = (530 ±10)/T(K) + (6.4 ±1.3) x 10-⁴P(Torr- (1.73 ±0.07), giving β 0.5% near the Earth's surface (298 K, 760 Torr) and 0.8% in the tropopause region (220 K, 200 Torr). The atmospheric implication of these results is briefly discussed.

Full text of DOI:10.1021/jp074117m

J. Phys. Chem. A 2007, 111, 9047-9053
9047
HNO3 Forming Channel of the HO2 + NO Reaction as a Function of Pressure and
Temperature in the Ranges of 72-600 Torr and 223-323 K
†
Nadezhda Butkovskaya,* Alexandre Kukui, and Georges Le Bras
CNRS, Institut de Combustion, A e´ rothermique, R e´ actiVit e´ et EnVironnement (ICARE),
1
C AV. de la Recherche Scientifique, 45071 Orl e´ ans Cedex 2, France
ReceiVed: May 28, 2007; In Final Form: July 16, 2007
A high-pressure turbulent flow reactor coupled with a chemical ionization mass-spectrometer was used to
determine the branching ratio of the HO + NO reaction: HO + NO f OH + NO (1a), HO + NO f
HNO (1b). The branching ratio, â ) k1b/k1a, was derived from the measurements of “chemically amplified”
concentrations of the NO and HNO products in the presence of O and CO. The pressure and temperature
dependence of â was determined in the pressure range of 72-600 Torr of N carrier gas between 323 and
23 K. At each pressure, the branching ratio was found to increase with the decrease of temperature, the
increase becoming less pronounced with the increase of pressure. In the 298-223 K range, the data could be
2
2
2
2
3
2
3
2
2
2
-
4
fitted by the expression: â(T,P) ) (530 ( 10)/T(K) + (6.4 ( 1.3) × 10 P(Torr) - (1.73 ( 0.07), giving
â ≈ 0.5% near the Earth’s surface (298 K, 760 Torr) and 0.8% in the tropopause region (220 K, 200 Torr).
The atmospheric implication of these results is briefly discussed.
1
. Introduction
Reaction 1b is another chain termination reaction suggested to
be significant in the upper troposphere. Regarding the potential
importance of reaction 1b in the whole atmosphere, it is
necessary to determine the branching k1b/k1a over the whole
ranges of atmospheric pressures and temperatures.
1
We have recently reported the observation of a minor channel
forming HNO3 (1b) in the gas-phase reaction of the HO2 radical
with NO:
1
HO + NO f OH + NO2
(1a)
(1b)
In the present work, measurements of the branching ratio for
reaction 1 was extended to the pressure range of 72-600 Torr
in the temperature range 223-323 K to provide a parametriza-
tion equation to be used to assess the role of reaction 1b in
model calculations of the atmospheric composition. To increase
the HNO3 signal-to-noise ratio, “chemical amplification” ac-
cording to reactions 1-3 was utilized in the TFR with the chain
length of the order of ten. A linear pressure dependence for the
branching ratio k1b/k1a was found, and a negative temperature
dependence, previously observed at P ) 200 Torr,1 was
confirmed for the whole range of the indicated pressures
between 323 and 223 K.
2
2
HO + NO f HNO3
The study was carried out in a turbulent flow reactor (TFR)
coupled with a chemical ionization mass-spectrometer (CIMS).
The channel (1b) was quantified by direct detection of HNO3
at pressure P ) 200 Torr in the reactor, and a value of 0.18%
was obtained for the branching ratio k1b/k1a at T ) 298 K. A
negative temperature dependence was found for this branching
ratio with a value of 0.9% at 223 K.
Reaction 1 is a very important atmospheric reaction. It plays
a key role in controlling the interconversion between OH and
HO2 radicals in the troposphere through the cycle
2
. Experimental Section
.1. Chemical Reactor. The experimental setup consisting
HO + NO f OH + NO2
(1a)
(2)
2
2
OH + CO f H + CO2
of a high-pressure turbulent flow reactor coupled to a chemical
ionization mass spectrometer was similar to that used in our
first study of the HO2 + NO reaction. A scheme of the
1
H + O + M f HO + M
(3)
2
2
apparatus is presented in Figure 1. The flow in the TFR was
Reaction 1a is also a major source of tropospheric ozone through
the conversion of NO to NO2 followed by NO2 photolysis to
NO and O-atoms, these latter combining with O2 to produce
ozone. The efficiency of the above cycle is decreased by chain
termination reactions, one of the most important being
created by N carrier gas evaporating from a liquid nitrogen
2
tank. The pressure in the TFR was controlled by regulating the
throttling valve of the reactor pump and varying the main N2
flow. The Reynolds numbers were in the range from 4900 to
1
1000 at flow rates from 64 to 145 SLPM corresponding to
pressures from 70 to 600 Torr. Reactor temperature was
controlled by combined effects of cooling via passing the main
flow through the metal coil immersed into a Dewar with liquid
N2 and heating using a CB100 digital controller (RKC Instru-
ment). This controller also allowed heating the reactor up to
50 °C.
OH + NO + M f HNO + M
(4)
2
3
*
Corresponding author. E-mail: bout@cnrs-orleans.fr. Permanent
address: Institute of Chemical Physics of the Russian Academy of Sciences,
17334 Moscow, Russian Federation.
1
†
CNRS, Service d’A e´ ronomie, IPSL, Paris.
1
0.1021/jp074117m CCC: $37.00 © 2007 American Chemical Society
Published on Web 08/24/2007

9048 J. Phys. Chem. A, Vol. 111, No. 37, 2007
Butkovskaya et al.
-
HNO3 was detected using the reaction with SF6 , giving the
4
peak at m/e 82:
SF6- + HNO f NO ‚(HF) + SF
-
(4i)
3
3
5
The branching ratio of reaction 1, â ) k1b/k1a, was obtained by
measuring the concentration ratio of the HNO3 and NO2
products from channels 1b and 1a, respectively. To convert the
measured intensity ratio of the products, ∆I82/∆I46, to the
concentration ratio, it is necessary to know the ratio of the
apparatus sensitivities to HNO3 and NO2, SHNO3/SNO2. The SHNO3/
SNO2 ratio and the absolute sensitivities themselves are deter-
mined by a number of parameters which vary with changing
the pressure in the TFR. The most important parameters include
pressure and flow velocity in the IMR, sampling conditions at
the interface between the TFR and IMR, optimum potentials
applied to the sampling cone, and the ion optics elements behind
it. To account for the dependence of the sensitivities on the
pressure in the TFR, the following approach was adopted. At
P ) 200 Torr, the absolute and relative sensitivities were
determined using a calibration method described in our previous
Figure 1. Experimental setup: 1, ion source; 2, ion-molecule reactor;
3
7
, temperature controller; 4, ,turbulizer.; 5, injector; 6, resistance;
, cooling bath; 8, discharge tube; 9, microwave discharge; 10, sampling
II
cones; 11, temperature sensor; 12, Fe (SO
4
) filter; 13, cool bath; 14,
NO cylinder.
HO2 radicals were produced in the TFR by the reaction
H + O + M f HO + M
(3)
2
2
1
study. In brief, the three-step procedure consisted of (1) NO2
calibration using a standard NO2 gas mixture, (2) OH calibration
using reaction 5 at short reaction times and low NO2 concentra-
tions, and (3) calibration of nitric acid by measuring the kinetics
of the OH decay and appearance of HNO3 in reaction 4:
with H-atoms generated by a microwave discharge in H2/He
gas mixtures flowing through a quartz tube concentrically
connected to the movable injector. The flows of N2 in the
injector and of He in the discharge tube were optimized
for maximum H-atoms production. He (AlphaGaz 2) was
purified by passing through molecular sieves cooled by
liquid N2. Tank grade H2 (AlphaGaz 2) was used without
further purification. NO was introduced into the TFR upstream
of the tip of the injector. The tank grade NO (AlphaGaz N20)
passed successively through ethanol/liquid nitrogen cooled
H + NO f OH + NO
(5)
(4)
2
OH + NO + M f HNO + M
2
3
The advantages of such “chemical” calibration are (i) in situ
production of HNO3 that eliminates the problems connected with
the introduction of HNO3 into the reactor and (ii) absence of
the errors connected with the determination of NO2 concentra-
tion, as these errors vanish in the sensitivity ratio SHNO3/SNO2.
This procedure was also used in some experiments at P ) 100,
II
traps and Fe (SO4) filter to remove NO2 and heavier nitrogen
oxides. NO flow rate of about 0.6 SCCM was maintained
using a TYLAN flow controller. O2 (AlphaGaz 2) and CO
(
AlphaGaz N47) were added into the main N2 stream using
CELERITY flow controllers. The maximum distance from the
injector tip to the orifice of the inlet cone of the ion-molecule
reactor was L ) 50 cm, which corresponded to a reaction
time in the TFR of about t ) 30 ms at P ) 200 Torr and T )
3
00, 400, and 500 Torr. However, to avoid potential problems
related to increasing formation of peroxynitrous acid, HOONO,
5
with pressure in reaction 4 and to simplify the calibration
2
98 K.
.2. CIMS Detection and Sensitivity. Gas mixtures from
procedure for other pressures in the TFR, the SHNO3/SNO2 ratios
were derived from the changes of NO2 and HNO3 sensitivities
relative to those at 200 Torr determined from direct introduction
of HNO3 and NO2 into the TFR.
2
the TFR were sampled through a Teflon cone with orifice
diameter of 0.5 mm into the ion-molecule reactor (IMR) located
perpendicular to the TFR. The flow rate of the Ar carrier gas
in the IMR was 3.6 SLPM at the typical pressure of 0.7 Torr.
The change of SHNO3 with pressure was determined by flowing
gaseous HNO3 from the mixture of HNO3 (Aldrich, 69%) and
H2SO4 (Sigma-Aldrich, 90%) aqueous solutions (10:1 volume
ratio) into the reactor. Gaseous HNO3 was transported by a He
flow bubbling through the solution mixture in a glass trap. The
trap was kept below 16 °C to avoid saturation of the HNO3
signal. The helium flow was varied by means of a CELERITY
mass flow controller with 10 SCCM maximum flow rate. A
dilution by larger He flow regulated using a 250 SCCM TYLAN
flow controller took place shortly downstream of the trap. The
trap was connected to the reactor via a PFA tube. At each
pressure SHNO3 was determined from the linear increase of the
signal intensity at m/e 82 with the flow rate of He passing
through the bubbler. The HNO3 partial pressure over the
solution, PHNO3, was not known, but this was not critical because
only the sensitivities relative to those at 200 Torr, SP/S200 (HNO3)
+
The primary Ar ions and electrons were generated in the ion
source by a heated filament. The emission current from the
filament was always stabilized during the measurements. SF6
was continuously introduced into the IMR downstream of the
ion source. The primary SF6- negative ions were produced by
attachment of thermalized electrons to SF6.
-
OH radicals and NO2 were detected as OH (m/e 17) and
-
- 2
NO2 (m/e 46) ions formed by electron transfer from SF6 :
SF6- + OH f OH + SF6
-
(1i)
(2i)
SF6- + NO f NO2 + SF6
-
2
HO2 radicals were detected at m/e 140 using the reaction3
SF6- + HO f [SF ‚O ] + other products
)
SHNO3(P)/SHNO3(200), were used. A rough estimation using
the average ratio SHNO3/SNO2 ≈ 4 from the chemical calibration
-
(3i)
2
4
2
at 200 Torr gave PHNO3 ≈ 0.03 Torr at 15 °C.

HNO3 Forming Channel of the HO2 + NO Reaction
J. Phys. Chem. A, Vol. 111, No. 37, 2007 9049
Figure 3. Pressure dependence of â ) k1b/k1a at 298 K. Open circles
denote measurements with chemical HNO calibration; star is the
3
extrapolation to atmospheric pressure (760 Torr).
of produced nitric acid that was a critical factor in reducing
experimental uncertainties, and confirmation of the previously
1
measured branching ratio k1a/k1b. Amplification of the signal
was achieved by adding into the TFR high concentrations of
CO and O2, creating a chain mechanism that includes termina-
tion reaction 1b:
HO + NO f OH + NO
(1a)
(1b)
(2)
2
2
2 2 2
Figure 2. Chemical amplification in the HO /NO/CO/O (a) and HO /
HO + NO f HNO
3
NO/CO/O (b) systems at 200 Torr and 298 K: (a) [HO
2
2
] ) 4.8 ×
2
1
1
13
16
17
1
0 , [NO] ) 5.9 × 10 , [O
2
] ) 1.8 × 10 , and [CO] ) 2.2 × 10
-
3
11
molecule cm (chain length 5.6); (b) [DO
2
] ) 8 × 10 , [NO] ) 3 ×
OH + CO f H + CO2
13
16
17
-3
10 , [O
.6).
2
] ) 1.2 × 10 , [CO] ) 3 × 10 molecule cm (chain length
4
H + O + M f HO + M
(3)
2
2
Calibration of the NO2 signal was made by flowing a
commercial NO2 gas mixture (Alpha Gaz, 5000 ppm in N2)
into the TFR. The mixture flow rate was measured by a
TYLAN-2900 mass flow controller. The NO2 sensitivities, SNO2,
were obtained from the slopes of the linear dependence of the
signal intensity at m/e 46 on the concentration calculated from
the flow rate. The signal intensities were measured in the range
At the same time, CO in reaction (2) served as a scavenger of
OH radicals preventing formation of nitric acid in the secondary
reaction (4):
OH + NO + M f HNO + M
(4)
2
3
11
13
-3
of [NO2] ) 3 × 10 to 2 × 10 molecule cm for which a
good linearity was observed at each pressure. Similarly to HNO3,
the pressure dependence was expressed relative to that at P )
Figure 2a illustrates the increase of the NO2 and HNO3
concentrations with the increase of the distance L between the
injector tip and the interface cone in the HO2/NO/CO/O2 system.
This run was done at P ) 200 Torr, T ) 298 K, and initial
2
00 Torr, SP/S200 (NO2) ) SNO2(P)/SNO2(200). The sensitivities
1
1
13
for NO2 and HNO3 exhibited different pressure dependences,
which very likely resulted from different formation mechanisms
of product ions. Both were found to decrease with increasing
pressure in the TFR keeping all the other parameters unchanged
as it was the case in these experiments. Finally, having made
calibrations for NO2 and HNO3, a pressure factor fP ) [SP/S200
concentrations [HO2] ) 4.8 × 10 ; [NO] ) 5.9 × 10 ; [O2]
16 17 -3
) 1.8 × 10 ; and [CO] ) 2.2 × 10 molecule cm . The
initial concentration of HO2 was determined by measuring the
NO2 concentration formed in the reaction in the absence of CO.
The maximum NO2 concentration formed in the presence of
1
2
-3
CO was 2.8 × 10 molecule cm , corresponding to a chain
length of 5.8. The chain length was regulated by NO concentra-
tion, the limiting step being reaction 1a. It was necessary to
limit the chain length in order to make the rate of reaction (4)
negligible and thus to control the rate ratio of reactions 4 and
2, which is the probability of HNO3 formation in reaction 4. In
the given example, for maximum distance L ) 50 this ratio
(HNO3)]/[SP/S200 (NO2)] was determined for the calculation of
the HNO3 to NO2 concentration ratio.
Change of the temperature in the TFR at a constant pressure
has no noticeable influence on the signal intensities of NO2 and
HNO3 as was shown in test experiments at 200 Torr. In these
experiments, NO2-He and HNO3-He mixtures were continu-
ously introduced into the reactor where the temperature was
lowered from 300 to 239 K while constant pressure was
maintained by regulating the reactor pumping.
-
4
was R ) k4[NO2]tot/k2[CO] ) 1.8 10 . [NO2]tot is the mean
total NO2 concentration in the reactor equal to [NO2]react/2 +
[NO2]bgr, where [NO2]react is the NO2 concentration produced
in the reaction at maximum L and [NO2]bgr is the background
concentration originating mainly from the NO inlet system. The
rate constants k2 and k4 were taken from ref 6. A linear
amplification was observed for both NO2 and HNO3 products
described by the equations I46 ) (814 ( 11)L + 2282 and I82
3
. Results
.1. Chemical Amplification of HNO3 formation in the
HO2/NO/CO/O2 System. Use of the chemical amplification
1a-3) in the TFR allowed an increase of the signal intensities
3
(

9050 J. Phys. Chem. A, Vol. 111, No. 37, 2007
Butkovskaya et al.
TABLE 1: Determination of â ) k1b/k1a at Different Pressures and 298 K Using Chemical Calibration of HNO a
3
e
f
g
P
Torr
∆I82/∆I46
b
82/b46d
10
â
int
â
slope
R
Nb
S
c
S
OH
c
S
c
10
-4
-4
%
%
%
NO2
HNO3
8
0- cps/molecule cm-3
1
2
00
1
2
3
4
1.23
1.74
1.98
0.93
2.32
3.31
2.90
1.67
3.79
51.9 ( 2.8
75.5 ( 4.3
61.7 ( 5.5
113 ( 5.4
51.5 ( 1.8
72.9 ( 5.3
63.6 ( 5.6
111 ( 4.9
0.169 ( 0.023
0.172 ( 0.024
0.180 ( 0.028
0.169 ( 0.023
0.173 ( 0.010
0.167 ( 0.022
0.166 ( 0.024
0.185 ( 0.030
0.166 ( 0.022
0.171 ( 0.016
0.018
0.019
0.028
0.016
7.63
6.80
6.67
Average:
100
300
400
500
1
1
1
1
1.75
1.49
1.97
1.07
4.49
1.66
2.08
1.32
12.0
4.57
2.10
1.16
74.2 ( 4.5
73.1 ( 3.5
28.7 ( 3.9
37.4 ( 4.0
73.1 ( 4.8
0.108 ( 0.015
0.238 ( 0.037
0.269 ( 0.043
0.345 ( 0.077
0.107 ( 0.016
0.012
0.012
0.013
0.008
35.0 ( 3.8
0.345 ( 0.074
a
reactions. b Experiment number. c
d
Calibration using H + NO
2
and OH + NO
2
S
X
denotes sensitivity to species X.
b
M
is the slope of the linear
]tot/k
2 3
e
f
g
dependence of signal intensity at m/e M versus reaction time. Values from the intensities. From the slopes. R ) k
limit for the probability of HNO
5
[NO [CO] is the upper
3
formation in the side OH + NO
2
reaction.
TABLE 2: Determination of the Pressure Dependence of â
)
(4.19 ( 0.13)L + 14.8. The top panel of Table 1 presents
a
) k1b/k1a at 298 K Using Sensitivity Pressure Factor
the branching ratios calculated from this experiment (first line)
and three other similar experiments carried out at 200 Torr with
chemical calibration. In these experiments, the product intensity
ratios, ∆I82/∆I46, were determined both from the slopes of the
amplification kinetics and by averaging the ∆I82/∆I46 values
for different positions of the injector. The average value is â )
d
P
[NO
10
2
]
bgr ∆[NO
2
]
[CO]
∆I82/∆I46
â
%
R
%
Nb Torr
11
12
17
a
-4
â/â200c
10
10
f
P
10
1
2
100
50
200
1.6
2.4
2.7
2.0
2.5
2.5
1.9 0.62
2.2 0.73
47.7
49.7
51.9
0.57 0.097 0.012
0.70 0.118 0.020
1
2.5
1
1
0.169 0.017
200
250
4.0
3.9
3.6
4.2
4.2
3.0
2.3
0.54
2.5
1
61.7
57.0
55.5
47.1
1
0.180 0.028
(
0.173 ( 0.008)%. Taking into account the systematic errors
2.8 1.24
3.1 1.61
3.3 2.12
1.15 0.206 0.020
1.45 0.261 0.016
1.62 0.280 0.008
3
3
00
50
in calibration, the final result can be expressed as â ) (0.17 (
0
.03)%.
3
200
00
500
00
2.0
4.2
4.2
4.8
2.4
1.6
1.1
0.43
2.1
1
75.5
47.4
46.1
42.1
1
0.173 0.018
Chemical amplification of the deuterated nitric acid, DNO3,
4
3.2 2.55
3.5 3.60
3.4 4.59
1.64 0.282 0.012
2.19 0.378 0.010
2.57 0.441 0.008
was obtained in the isotopic DO2/NO/CO/O2 system, when the
discharge in D2/He mixture was used to produce D-atoms.
Figure 2b shows the results of the kinetic measurements for
NO2, DNO3 and HNO3 at P ) 200 Torr, T ) 298 K, and initial
6
4
5
200
00
1.9
4.7
2.4
0.72
2.4
6.0 4.59
1
57.0
30.5
1
0.173 0.016
6
2.45 0.425 0.005
11
172
2.3
2.4
1.8
2.4
3.0
3.6
3.6
4.3
3.3
2.8
1.4
1.6
1.4
1.7
1.1
0.78
2.7 0.73
2.7
40.8
34.3
32.0
26.5
23.5
21.7
20.1
17.8
0.95 0.164 0.018
1
concentrations [DO2] + [HO2] ) 9.5 × 10 ; [NO] ) 2.9 ×
2
2
3
00
50
00
1
0.173 0.017
13
16
17
-3
1
0 ; [O2] ) 1.2 × 10 ; [CO] ) 2.5 × 10 molecule cm .
2.3 1.24
2.8 1.61
3.3 2.16
3.7 2.61
3.5 3.71
4.2 4.59
1.16 0.201 0.013
1.25 0.216 0.012
1.48 0.256 0.012
1.65 0.285 0.012
2.18 0.377 0.010
2.39 0.413 0.008
Appearance of HNO3 (m/e ) 82) is explained by the formation
of OH and H-atoms from the dissociation of H2O traces in He
carrier gas. The initial ratio of [H]/[D] ) 0.14 was obtained by
the titration of the discharge products with NO2. Isotopic DNO3
also exhibits linear kinetics proving that the nitric acid observed
in the HO2 + NO reaction is not a product of the surface reaction
of NO2 with H2O. Assuming similar sensitivities to HNO3 and
DNO3, the observed DNO3 signal intensity indicates that the
branching ratio for the isotopic DO2 + NO reaction would be
a factor of ∼2 lower than for reaction 1.
350
4
5
6
00
00
00
6
72
00
150
1.3
1.6
2.2
3.5
2.9
2.7
1.3
2.3
3.0
4.6
4.1
1.5
2.5 0.51
2.5 0.62
2.5 0.73
48.4
48.1
43.8
43.6
41.0
27.7
0.57 0.098 0.005
0.68 0.118 0.010
0.76 0.132 0.017
1
2
2
4
00
50
50
3.7
1
1
0.173 0.020
4.0 1.24
7.2 1.81
1.10 0.189 0.019
1.95 0.337 0.006
a
See Experimental Section for determination of pressure correction
3.2. Pressure Dependence of the Branching Ratio at Room
b
c
factor f
ing ratio with respect to that at 200 Torr.
probability of HNO formation in the OH + NO
Concentrations are in molecule cm
P
using external source of HNO
3
.
Experiment number. Branch-
Upper limit for the
side reaction.
Temperature. At room temperature, the kinetics of the forma-
tion of NO2 and HNO3 products were measured in the presence
of CO as described above at different pressures ranging from
d
3
2
-
3
.
7
2 to 600 Torr. The bottom panel of Table 1 contains the results
The obtained pressure dependence is shown in Figure 3. It is a
linear function of pressure expressed by
obtained using chemical calibration at 100, 300, 400, and 500
Torr and Table 2 contains the results from the six experiments
with variable pressure using direct calibration of HNO3. In most
experiments, the ∆I82/∆I46 ratio was determined from the slopes
of the amplification kinetics. Some experiments were done at a
fixed L making several measurements of NO2 and HNO3 with
the discharge switched on and off (experiment 5 in Table 2).
The measured intensity ratios were corrected by sensitivity
pressure factor fP determined with respect to 200 Torr as
described in the Experimental Section. Then, division of the
corrected intensity ratios by that for 200 Torr gives â/â200, the
reaction branching ratio at a given pressure relative to that at
â(P) )
-
4
-2
(6.4 ( 0.3) × 10 P(Torr) + (4.2 ( 0.8) × 10 (Eq2)
with 2σ uncertainty limits. Extrapolation to atmospheric pressure
gives â(760) ) 0.53%, which is indicated by a star symbol in
Figure 3. It is worth noting that the obtained dependence has a
positive zero intercept. Assuming â f 0 when P f 0, the
measurements indicate that there should be a nonlinear depen-
dence at low pressures.
3
.3. Temperature Dependence of the Branching Ratio at
2
00 Torr which is a known value:
â(P) ) [∆I /∆I (P)/∆I /∆I (200)]f â
P 200
Different Pressures. Temperature dependence of â was mea-
sured by keeping a constant mass flow rate of the carrier gas
and the reactants at a given pressure. The constant pressure
(Eq1)
82
46
82
46

HNO3 Forming Channel of the HO2 + NO Reaction
J. Phys. Chem. A, Vol. 111, No. 37, 2007 9051
pressure. The extrapolation to atmospheric pressure (black dotted
line) was obtained from the power function fit for this decrease.
When the normalized temperature curve for every pressure
is multiplied by the corresponding â coefficient given by eq 1,
we obtain a full picture for the branching ratio, as shown in
Figure 5. In the 298-223 K range temperature dependences
for different pressures present a set of nearly parallel and
equidistant straight lines. The linear fit fails at temperatures
higher than 298 K, where curvatures were observed for P )
1
00, 200, and 400 Torr. Below 298 K the whole set of data can
be described by the simple three-parameter expression of the
general form
â(P, T) ) a/T + bP + c
(Eq3)
Coefficient a was found by averaging the slopes of the observed
temperature dependencies, coefficients b and c were determined
by standard two-parameter least-square fit of the data. The
numerical expression can be written as
Figure 4. Normalized temperature dependence of â ) k1b/k1a at
different pressures. Dotted line is the extrapolation to 760 Torr.
â(P, T) ) (530 ( 20)/T(K) +
during the cooling was maintained by regulating the throttling
valve of the reactor pump. As a rule, the measurements were
done at a fixed position of the injector, corresponding to L )
-
4
(6.4 ( 1.3) × 10 P(Torr) - (1.73 ( 0.07) (Eq4)
with 2σ uncertainties). In Figure 6 the pressure dependence
(
5
0 cm. Table 3 gives the measured signal intensity ratios and
for HNO3 production is reported at four temperatures: 298, 263,
49, and 228 K, with the nearest available experimental points
the ratios normalized to the room-temperature value, which is
identical to the normalized branching ratio, â/â298. The con-
centrations of NO, O2, and CO given in Table 3 correspond to
T ) 298 K. Figure 4 presents a summary of the normalized
branching ratios as a function of 1000/T. The observed plots
could be rather well approximated by a linear fit. These fits are
shown in Figure 4 by solid lines (the data obtained above 298
K were excluded from the linear regression analysis). The slopes
of the linear fit monotonically decrease with the increase of
2
taken in the vicinity of these temperatures. The straight lines
were calculated using eq 4. Extrapolated to zero pressure, they
exhibit larger intercepts for lower temperatures.
4
. Discussion
Recently, a theoretical examination of the mechanisms on
the HNO3 potential energy surface (HO2 + NO S HOONO S
TABLE 3: Measurement of the Temperature Dependence of â ) k1b/k1a at Different Pressures
T
K
∆I82/∆I46
T
K
∆I82/∆I46
T
K
∆I82/∆I46
-
4
â/â298a
-4
â/â298a
-4
â/â298a
10
10
10
13
] ) 1.4 × 10 ; [CO] ) 2.4 × 10 molecule cm-3
16
17
P ) 100 Torr; [NO] ) 4.1 × 10 ; [O
2
2
98
93
88
85
67
113
123
172
198
279
1
265
253
249
245
234
311
432
452
536
618
2.74
3.81
4.00
4.73
5.45
226
223
298
308
323
695
678
97.1
71.8
56.5
6.13
5.98
1
0.74
0.58
2
2
2
2
1.08
1.52
1.75
2.46
1
3
] ) 1.2 × 10 ; [CO] ) 2.5 × 10 molecule cm-3
16
17
P ) 200 Torr; [NO] ) 3.8 × 10 ; [O
2
2
2
2
2
2
2
98
81
72
56
38.7
53.0
76.4
1
243
233
223
132
152
164
3.41
3.93
4.24
298
308
323
33.7
24.9
21.3
1
0.74
0.63
1.45
1.98
2.71
105
13
] ) 1.6 × 10 ; [CO] ) 2.9 × 10 molecule cm-3
16
17
P ) 300 Torr; [NO] ) 5.4 × 10 ; [O
2
2
2
2
98
83
81
73
25.0
30.8
35.1
40.3
1
261
257
254
248
51.0
53.5
60.1
65.1
2.03
2.14
2.40
2.53
241
233
229
69.4
75.1
76.3
2.77
2.98
3.05
1.23
1.40
1.61
13
] ) 2.1 × 10 ; [CO] ) 3.5 × 10 molecule cm-3
16
17
P ) 400 Torr; [NO] ) 6.1 × 10 ; [O
2
2
2
2
98
83
79
66
31.5
41.8
48.0
46.6
1
256
250
241
232
66.1
62.5
74.2
84.4
2.10
1.99
2.36
2.68
228
298
308
323
84.7
16.2
14.6
11.6
2.69
1
0.90
0.72
1.33
1.53
1.48
13
] ) 2.7 × 10 ; [CO] ) 4.8 × 10 molecule cm-3
16
17
P ) 500 Torr; [NO] ) 9 × 10 ; [O
2
2
2
2
2
98
89
85
80
14.2
16.0
16.3
20.2
1
272
262
255
250
20.5
24.7
21.5
26.6
1.44
1.74
1.51
1.87
243
238
233
228
29.4
32.4
33.3
33.8
2.07
2.28
2.34
2.37
1.12
1.15
1.42
13
] ) 3.2 × 10 ; [CO] ) 4.0 × 10 molecule cm-3
16
17
P ) 600 Torr; [NO] ) 4.0 × 10 ; [O
2
2
2
2
98
78
75
18.9
23.8
24.0
1
263
249
238
28.9
34.9
38.6
1.53
1.85
2.04
233
229
223
40.0
42.5
42.8
2.12
2.25
2.26
1.26
1.27
a
Branching ratio with respect to that at 298 K.

9052 J. Phys. Chem. A, Vol. 111, No. 37, 2007
Butkovskaya et al.
Figure 7. Representation of the Figure 6 in logarithmic scale for
9
comparison with the calculation from Zhang and Donahue (thick
Figure 5. Pressure and temperature dependences of â ) k1b/k1a. Upper
dotted line represent extrapolation to P ) 760 Torr.
curves).
kcal mol- lower than the free OH and NO2. However, in
contrast to ref 9, where the isomerization path was involved to
explain the observed decay of trans-HOONO, Zhao et al. believe
that the isomerization occurs easier from cis-HOONO than from
trans-HOONO conformation.
1
7
According to theoretical considerations, the pressure de-
pendences in Figure 6 can be interpreted as a falloff corre-
sponding to stabilization of the excited HONO2 formed at the
HO2 + NO energy. In Figure 7 we compare our presen7t
experimental results with the calculated pressure dependence
presented in logarithmic coordinates. As already mentioned,
there is a rather good agreement between the measurements and
theory at 298 K, although the theoretical pressure dependence
at 298 K is somewhat steeper than the observation. As seen in
Figure 7, the linearity of eq 1 breaks when switching to
logarithmic scale because, formally, â * 0 at zero pressure.
Accordingly, the divergence between the measurements and
calculations increases at lower temperatures because of larger
intercepts. To summarize, theory does not exclude formation
of HNO3 in reaction 1; at ambient temperature, the observed
pressure dependence of â can be reproduced by calculations in
a limited pressure range (100-600 Torr). Also, calculations
confirm the observed negative temperature dependence of â.
Figure 6. Pressure dependence of â ) k1b/k1a at different temperatures.
Solid curves correspond to eq 3.
OH + NO2 S HNO3) has been performed by Zhang and
Donahue using a multiple-well master-equation simulation of
7
the system. To constrain the parameters of the simulation, the
5
. Atmospheric Implication
authors used the experimental evidence for HNO3 production
1
in the HO2 + NO reaction, the experimental data on formation
The obtained results confirmed our previous finding of a
and decay of cis- and trans-conformers of the short-lived
minor HNO forming channel in the HO + NO reaction, with
3
2
8
-10
1
HOONO intermediate in the OH + NO2 reaction,
and on
3
an increase of the HNO yield with decreasing temperature.
isotopic scrambling in 18OH + NO2.
10,11
The calculations show
In addition, they show an increase of this yield with increasing
pressure. The atmospheric consequences of this new process
has been already considered in our previous paper, predicting
x
that channel 1b can occur by isomerization of trans-HOONO
to HONO2, if a high barrier between cis and trans conformers
and a low barrier for isomerization to HONO2 are assumed. A
good agreement with our branching fraction for HNO3 formation
a noticeable effect of reaction 1b on the concentrations of HO ,
1
NO , and ozone in the upper troposphere. The plots of Figure
x
1
in reaction 1 at 298 K was obtained with the energy of the
5 show a change of â ) k /k roughly from 0.5% near the
1
b
1a
-1
transition state for isomerization of approximately 5 kcal mol
Earth’s surface (298 K, 760 Torr) to 0.8% in the tropopause
region (220 K, 200 Torr). This suggests that reaction 1b could
also have a significant effect in the middle and lower tropo-
sphere. The effect of reaction 1b on the atmospheric composition
can be quantified by including this reaction in atmospheric
models with the rate constant calculated using eq 4. Such model
calculations have already been done using both 2D and 3D
lower than the energy of OH + NO2. The existence of such a
low-lying “rotational” transition state was first proposed by
Dransfield et al.12 A search for the transition states for HOONO
13
rearrangements was performed by Zhao et al. using density
functional theory. The authors suggest that O-O cleavage in
HOONO may lead to hydrogen-bonded OH‚‚‚ONO complexes
which through rotational motion can be reoriented for N-O
bond formation and collapse to HONO2. The complexes and
the transition states for isomerization were located at the B3LYP/
1
4
models. The resulting impact of reaction 1b on the chemical
composition of the troposphere could even be more significant
considering the positive influence of water vapor on the
branching ratio â which has been observed in the previous test
6
-311++G** and CBS-QB3 levels. All these stationary points
1
were found to be very similar in energy, in the range of 1-2
experiments. This humidity effect is going to be investigated

HNO3 Forming Channel of the HO2 + NO Reaction
J. Phys. Chem. A, Vol. 111, No. 37, 2007 9053
in more details by measuring â as a function of H2O concentra-
tion in extended ranges of pressure and temperature.
under Urban Conditions; Final Report to California Air Resources Board,
CARB Contract No. 03-333; California Air Resources Board: Sacramento,
CA, 2005.
(
6) Sander, S. P.; Friedl, R. R.; Golden, D. M; Kurylo, M. J.; Moortgat,
Acknowledgment. The work was carried out within the
SCOUT project of the European Union and the National
Programmes of Atmospheric Chemistry of CNRS (PNCA and
LEFE-CHAT).
G. K.; Wine, P. H.; Ravishankara, A. R.; Kolb, C. E.; Molina, M. J.;
Finlayson-Pitts, B. J.; Huie, R. E.; Orkin, V. L. Chemical Kinetics and
Photochemical Data for Use in Atmospheric Studies; JPL Publication 06-
2
, Evaluation Number 15; Jet Propulsion Laboratory: Pasadena, CA, 2006.
(
(
7) Zhang, J.; Donahue, N. M. J. Phys. Chem. A 2006, 110, 6898.
8) Hippler, H.; Nasterlack, S.; Striebel, F. Phys. Chem. Chem. Phys.
References and Notes
2002, 4, 2959.
(
9) Fry, J. L.; Nizkorodov, S. A.; Okumura, M.; Roehl, C. M.;
(
1) Butkovskaya, N. I.; Kukui, A.; Pouvesle, N.; Le Bras, G. J. Phys.
Chem. A 2005, 109, 6509.
2) Ikezoe, Y.; Matsuoka, S.; Takebe, M.; Viggiano, A. Gas Phase
Francisco, J. S. J. Chem. Phys. 2004, 121, 1432.
(
10) D’Ottone, L.; Bauer, D.; Campuzano-Jost, P.; Fardy, M.; Hynes,
A. J. Faraday Discuss. 2005, 130, 1.
11) Donahue, N. M.; Mohrschladt, R.; Dransfield, T. J.; Anderson, J.
G.; Dubey, M. K. J. Phys. Chem. A 2001, 105, 1515.
12) Dransfield, T. J.; Donahue, N. M.; Anderson, J. G. J. Phys. Chem.
A 2001, 105, 1507.
(13) Zhao, Y.; Houk, K. N.; Olson, L. P. J. Phys. Chem. A 2004, 108,
5864.
(
Ion-Molecule Reaction Rate Constants Through 1986; Ion Reaction Research
Group of the Mass Spectroscopy Society of Japan, Maruzen Company:
Tokyo, Japan, 1987.
(
(
(3) Elrod, M. J.; Ranschaert, D. L.; Schneider, N. J. Int. J. Chem. Kinet.
2
001, 33, 363.
(4) Huey, L. G.; Hanson, D. R.; Howard, C. J. J. Phys. Chem. 1995,
9
9, 5001.
(
5) Okumura, M.; Sander, S. P.; Sivakumaran, V.; Mollner, A. K.; Fry,
(14) Chipperfield, M. P.; Evans, M. J.; Cariolle, D.; Butkovskaya, N.
I.; Kukui, A.; Le Bras, G. Manuscript in preparation.
J. L.; Feng; L. Gas-Phase Formation Rates of Nitric Acid and its Isomers