Formation of nitric acid in the gas-phase HO2 + NO reaction: Effects of temperature and water vapor

Article abstract of DOI:10.1021/jp051534v

A high-pressure turbulent flow reactor coupled with a chemical ionization mass spectrometer was used to investigate the minor channel (Ib) producing nitric acid, HNO₃, in the HO₂ + NO reaction for which only one channel (la) is known so far: HO₂ + NO → OH + NO ₂ (la), HO₂ + NO → HNO₃ (1b). The reaction has been investigated in the temperature range 223-298 K at a pressure of 200 Torr of N₂ carrier gas. The influence of water vapor has been studied at 298 K. The branching ratio, k_1blk_1a, was found to increase from (0.18+_-0.06^+0.04)% at 298 K to (0-87 _-0.08^+0.05)% at 223 K- corresponding to k_1b = (1.6 ±0.5) x 10_-14 and (10.4 ±1.7) ×10 ^-14 cm³ molecule ^-1 s^-l, respectively at 298 and 223 K. The data could be fitted by the Arrhenius expression k_1b = 6.4 × 10-17 exp((1644 ±76)l7) cm ³ molecule^-1 s^-1 at T = 223-298 K. The yield of HNO3 was found to increase in the presence of water vapor (by 90% at about 3 Torr of H₂O). Implications of the obtained results for atmospheric radicals chemistry and chemical amplifiers used to measure peroxy radicals are discussed. The results show in particular that reaction Ib can be a significant loss process for the HO_x (OH, HO₂) radicals in the upper troposphere.

Full text of DOI:10.1021/jp051534v

J. Phys. Chem. A 2005, 109, 6509-6520
6509
Formation of Nitric Acid in the Gas-Phase HO2 + NO Reaction: Effects of Temperature
and Water Vapor
,†
‡
†
†
N. I. Butkovskaya,* A. Kukui, N. Pouvesle, and G. Le Bras
CNRS, Laboratoire de Combustion et Syst e` mes R e´ actifs, 1C AV. de la Recherche Scientifique, 45071 Orl e´ ans
Cedex 2, France, and CNRS, SerVice d’A e´ ronomie, Paris, France
ReceiVed: March 24, 2005; In Final Form: May 30, 2005
A high-pressure turbulent flow reactor coupled with a chemical ionization mass spectrometer was used to
investigate the minor channel (1b) producing nitric acid, HNO
one channel (1a) is known so far: HO + NO f OH + NO (1a), HO
has been investigated in the temperature range 223-298 K at a pressure of 200 Torr of N
3
, in the HO
2
+ NO reaction for which only
(1b). The reaction
carrier gas. The
2
2
2
+ NO f HNO
3
2
influence of water vapor has been studied at 298 K. The branching ratio, k1b/k1a, was found to increase from
+
-
0.04
+0.05
-0.08
-14
(
0.18
)% at 298 K to (0.87
)% at 223 K, corresponding to k1b ) (1.6 ( 0.5) × 10 and (10.4 ( 1.7)
0.06
-
14
3
-1 -1
×
10 cm molecule s , respectively at 298 and 223 K. The data could be fitted by the Arrhenius expression
-
17
3
-1 -1
k
1b ) 6.4 × 10 exp((1644 ( 76)/T) cm molecule
s
at T ) 223-298 K. The yield of HNO
O). Implications of the obtained
results for atmospheric radicals chemistry and chemical amplifiers used to measure peroxy radicals are
discussed. The results show in particular that reaction 1b can be a significant loss process for the HO (OH,
HO ) radicals in the upper troposphere.
3
was found
to increase in the presence of water vapor (by 90% at about 3 Torr of H
2
x
2
1
. Introduction
The HO2 + NO reaction occurs in radical chemical amplifiers
e.g., 1-4] to convert low concentrations of HO2 into measur-
able NO2 concentrations in a chain reaction (1-3):
the presence of water:
HO + NO + H O f HNO (or HOONO) + H O
[
2
2
3
2
The observation of HNO3 in reaction 1 at room temperature, at
very low but measurable yield, even in the absence of H2O, led
us to extend the study to low temperatures. An increase of the
HNO3 yield with decreasing temperature being possible, reaction
1b
(1a)
(2)
2
HO + NO f OH + NO2
OH + CO f H + CO2
H + O + M f HO + M
(3)
2
2
Chain termination takes place mainly through the gas-phase
reaction
HO + NO + M f HNO + M
(1b)
2
3
may become a significant sink of HOx (OH, HO2) in the upper
troposphere.
HO + NO + M f HONO + M
(4)
Reaction of HO2 with NO is very important in atmospheric
chemistry because of its central role in the OH/HO2 recycling
and reduction of ozone depletion by HOx cycles in the
stratosphere and in ozone production in the troposphere. The
importance of this reaction was demonstrated by Howard and
and the HO2 radical wall loss. It was found that the chain length
decreases in the presence of water vapor and that the expected
increase of HO2 wall losses may account for only some part of
the observed water effect.1 It has been suggested that this effect
could be, at least partially, attributed to the reaction of the HO2‚
H2O adduct with NO yielding nitric acid, HNO3, or peroxyni-
,2
5
Evenson who carried out the first direct measurement of its
rate constant. Further studies covering atmospheric conditions6
established a negative temperature dependence of the rate
constant in the 183-403 K range with a recommended value
of (8.8 ( 1.2) × 10
-9
2
trous acid, HOONO. Recently, a detailed study of the phe-
nomena resulting in a chain termination provided additional
evidence that formation of nonradical products (e.g., HNO3) in
the HO2 + NO reaction in the presence of water is likely to be
responsible for the chain shortening.4
The initial aim of the present work was to clarify this problem
by studying the possible formation of HNO3, or its isomer
HOONO, as a minor channel of the HO2 + NO reaction in
-
12
3
-1 -1
10
cm molecule
s
at 298 K. The
reaction is supposed to proceed via the formation of the excited
HOONO intermediate complex followed by rapid decomposition
to OH and NO2:
HO + NO f HOONO** f OH + NO2
(1a)
2
*
Permanent address: Institute of Chemical Physics of the Russian
Academy of Sciences, 117334 Moscow, Russian Federation.
6
†
‡
Howard considered two competing channels, stabilization of
CNRS, Laboratoire de Combustion et Syst e` mes R e´ actifs.
CNRS, Service d’A e´ ronomie.
HOONO complex and formation of HNO + O2:
1
0.1021/jp051534v CCC: $30.25 © 2005 American Chemical Society
Published on Web 07/01/2005

6510 J. Phys. Chem. A, Vol. 109, No. 29, 2005
Butkovskaya et al.
HO + NO + M f HOONO + M
2
HO + NO f HNO + O2
2
The upper limit for each of these channels was estimated to be
lower than 0.13% from the measurement of OH and HO2
concentrations as a function of reaction time at P ) 1-3 Torr
and T ) 271 and 303 K using a laser magnetic resonance
method. The predominance of channel 1a was confirmed by
8
Bohn and Zetzsch, who obtained for the HO2 + NO reaction
a quantum yield of OH g 95% using time-resolved detection
of OH by CW UV-laser absorption at P ) 75, 375, and 750
Torr. They argued that stabilization of HOONO with possible
further rearrangement to HNO3
Figure 1. Experimental setup: (1) ion source; (2) ion molecule reactor
(
IMR); (3) temperature controller; (4) “turbulizer”; (5) injector; (6)
resistance; (7) liquid N cooling bath; (8) discharge tube; (9) microwave
discharge; (10) sampling cones; (11) temperature sensors; (12) Fe -
SO ) sieves; (13) liquid N /ethanol cooling bath; (14) NO cylinder.
HO + NO + M f HOONO** + M f
2
2
II
HOONO + M f HNO + M
3
(
4
2
seemed to be unlikely at atmospheric pressures but could not
be excluded totally. In the recent work of Bardwell et al. a
9
A T-type thermocouple inserted in the middle of the reactor
served as a temperature input of the controller. Two K-type
thermocouples were used to measure the temperature at both
ends of the reactor.
The NO flow rate was varied from 0.05 to 120 SCCM using
two TYLAN flow controllers; the corresponding concentrations
(100 ( 5)% yield of NO2 was obtained using the chemical
ionization mass spectrometry (CIMS) technique. No stabilized
adducts or secondary product channels were observed within
the error limits of their study over the 183-300 K temperature
range. Because of its atmospheric importance, the formation of
organic nitrates in RO2 + NO reactions is a topic of intense
interest (see Discussion). It is interesting to mention that if the
HO2 + NO reaction produces HNO3, it becomes a part of the
family of peroxy radical reactions with NO that produce nitrates.
In the present work, the branching fraction and the kinetics
of nitric acid formation in the HO2 + NO reaction were
investigated by direct detection of HNO3 by CIMS at 298 K in
the absence and presence of H2O and in the temperature range
1
2
16
in the reactor were from 5.0 × 10 to 1.2 × 10 molecule
-3
cm . The tank grade NO (AlphaGaz N20) passed successively
through three ethanol/liquid N2 cooled traps and an iron sulfate
filter to remove NO2 and heavier nitrogen oxides. At P ) 200
Torr, detection of HNO3 from reaction 1 was complicated by
the secondary reaction 5:
OH + NO + M f HNO + M
(5)
2
3
2
98-223 K without H2O. All the experiments were carried out
To suppress this reaction, OH radicals were scavenged by high
concentrations of NO or cyclohexane.
at P ) 200 Torr of N2, except a few ones in the range of 100-
4
00 Torr as a test of the HNO3 formation mechanism.
Cyclohexane entered the reactor with the He flow passing
through the trap with liquid C6H12 (Riedel de Ha e¨ n, > 99.5%),
which was preliminarily purified in several freeze-pump-thaw
cycles. Water vapor in experiments at 298 K was introduced
into the reactor together with a part of N2 flow bubbling through
distilled H2O in the stainless steel reservoir of 25 cm height.
The reservoir was placed in the cryostat maintained at a constant
temperature of 19 °C. Assuming that the partial water pressure
in the flow was equal to the saturated vapor pressure, the
estimated water concentration in the reactor varied from 1 ×
2
. Experimental Method
2
.1. Main Reactor. The title reaction was studied in a high-
pressure turbulent flow reactor (HPTFR) coupled with an ion-
molecule reactor and a quadrupole mass spectrometer (Figure
1
2
The total flow through the reactor was up to 96 SLPM. Mixing
and flow conditions were determined by a Reynolds number of
Re ≈ 7300, which ensured fast turbulent mixing of reagents
and a flat radial velocity profile. A detailed description of the
experimental setup is presented elsewhere.11 NO was introduced
into the reactor upstream of the tip of the movable injector of
). The pressure in the reactor (Pyrex tube of 2.4 cm i.d.) was
00 Torr; flow velocity of N2 carrier gas was about 17 m/s.
16
17
-3
1
0
to 1.17 × 10 molecule cm .
2
.2. Chemical Ionization. The ion-molecule reactor (IMR)
consisted of a stainless steel tube of 4 cm i.d. and 40 cm length
situated perpendicular to the chemical reactor. Ar (Alpha Gaz
1
.1 cm i.d. which served as a prereactor for producing HO2
2
) was used as a carrier gas in the IMR at a flow rate of 2.2
radicals from the H + O2 reaction. H-atoms were generated by
microwave discharge in a H2/He gas mixture in a quartz tube
of 0.6 cm i.d. concentrically connected with the injector. He
SLPM and pressure of about 1 Torr. This gave an average flow
velocity of 44 m/s and a reaction time of the order of 4 ms.
The carrier gas was purified by passing through liquid N2 cooled
traps. Electrons were generated by electrical heating of a
(Alpha Gaz 2) was purified by passing through the molecular
sieves cooled by liquid N2. Concentration of molecular oxygen
+
thoriated irridium filament. Primary Ar ions were produced
1
6
-3
(
Alpha Gaz 2) in the injector (1.5 × 10 molecule cm ) was
by electron bombardment. SF6 (Alpha Gaz N37) was introduced
sufficient to consume the H-atoms entirely during the ∼2 ms
-
into the IMR downstream of the filament. The SF6 negative
residence time in the injector. Typical HO2 concentrations in
ions were produced by attachment of thermalized electrons to
SF6.
1
1
-3
the main reactor were (2-5) × 10 molecule cm . The
maximal distance from the injector tip to the orifice of the inlet
cone of the ion-molecule reactor was L ) 50 cm corresponding
to a reaction time of t ≈ 30 ms. The cooling of the reactor was
achieved by immersing a metal coil with the carrier gas into a
Dewar vessel filled with liquid nitrogen. The temperature was
regulated using an RKC Instrument CB100 digital controller.
Two chemical ionization methods were used to detect
HNO3: at m/e 82 using the fast ion-molecule reaction 1i as
12
suggested by Huey et al.,
SF6- + HNO f [NO ‚HF] + SF
-
(1i)
3
3
5

Nitric Acid Formation in Gas-Phase HO2 + NO Reaction
J. Phys. Chem. A, Vol. 109, No. 29, 2005 6511
and at m/e 186 using reaction 2i originally described by Huey
and Lovejoy,1
3a
SiF5- + HNO f [SiF ‚HNO ]
-
(2i)
3
5
3
1
3b
the latter being used for HNO3 atmospheric measurements.
-
SiF5 ions were produced by fast transfer reaction 3i when SiF4
(Air Products 99,94%) was added to SF6 flow:
SF6- + SiF f SiF5 + SF5
-
(3i)
4
However, at low pressure in the ion molecule reactor, the
sensitivity to HNO3 using SiF4 is substantially lower than that
1
2,13a
using SF6,
and this detection method was employed in our
study only in order to prove that the signal observed as a product
of 1b corresponds to HNO3.
OH (m/e 17), NO2 (m/e 46), and O3 (used to test the system,
m/e 48) were ionized by the charge-transfer reactions with
-
12,14
SF6 :
SF6- + OH f OH + SF6
-
(4i)
(5i)
(6i)
SF6- + NO f NO2 + SF6
-
2
SF6- + O f O3 + SF6
-
3
Figure 2. Calibration of OH radicals from H + NO
NO consumption (a) and calibration of HNO from the OH + NO
reaction measuring OH consumption (b).
2
reaction measuring
HO2 radicals could be detected at m/e 140 using reaction15
2
3
2
SF6- + HO f [SF ‚O ] + other products
-
(7i)
2
4
2
the injector when L was changed from 3 to 50 cm, confirming
the constant NO2 concentration over the reactor cross section.11
In this work we also show that HONO can be detected at m/e
2
.3.2. HO2. HO2 signal was quantified during the HO2 +
6
6 and m/e 170 using reactions (see also ref 11, Figure 8)
NO reaction at sufficiently high NO concentration, when all
HO2 radicals were converted to NO2. The reaction channel with
HNO3 formation may be neglected for the purpose of HO2
calibration.
SF6- + HONO f [HF‚NO ] + SF
-
(8i)
(9i)
2
5
SiF5- + HONO f [SiF ‚HONO]
-
5
2
.3.3. OH. Determination of the concentrations of other
species involved in the study was based on NO2 calibration using
the fast reaction 7 followed by reactions 4 and 5
The relative sensitivities were approximately S(NO2):S(OH):
-
S(HNO3):S(HONO) ) 1.0:5.5:4.5:3.0 with SF6 . The 2σ detec-
9
tion limits for integration time of 20 s were 3.2 × 10 (NO2),
8
8
8
H + NO f OH + NO
(7)
2
.5 × 10 (OH), 2.5 × 10 (HNO3), and 2.3 × 10 (HONO)
2
-
3
-
8
molecule cm using SF6 and 5.6 × 10 (HNO3) molecule
cm using SiF5 .
.3. Calibration of Signals. 2.3.1. NO2. The NO2 signal from
HO + NO + M f HONO + M
(4)
(5)
-
3
-
OH + NO + M f HNO + M
2
2
3
reaction 1 was calibrated by introducing known concentrations
of NO2 into the reactor. The NO2 flow rate was determined
from measurement of the pressure drop in a known volume
when a preprepared NO2/He mixture was passing into the reactor
via a regulating valve. The mixtures were prepared by mixing
either NO or gaseous NO2 (MATHESON) (60-100 Torr) with
O2 (1 atm) in a glass flask which was kept for 1 day and night
to allow NO to react with oxygen. Then, NO2 was purified by
condensation at liquid N2 temperature and pumping away the
excess O2 at a pressure of about 50 Torr. He, purified by passing
through the liquid N2 cooled trap with molecular sieves, was
added into the flask to get a total pressure slightly over 1 atm
to obtain a 1% NO2/He mixture in one flask and 7-12% mixture
in another one. The 1% mixture was used to calibrate OH
radicals. The presence of N2O4 dimer in the NO2 mixture was
taken into account in the calculation of NO2 concentration in
the reactor. The signal intensity at m/e 46 showed linear
dependence on the measured NO2 flow in the [NO2] ) (0.02-
-
12 16
with the rate constants k4 ) (3.0 ( 0.5) × 10
and k5 )
-
12
3
-1 -1
(6.4 ( 0.4) × 10
cm molecule at 298 K and 200
s
1
7
Torr. NO2 was introduced into the main reactor upstream of
the H-atoms injector. At low NO2 concentrations and short
reaction times, reactions 4 and 5 can be ignored and calibration
of OH radicals can be done by measuring NO2 consumption at
different NO2 flow rates. In this case ∆[OH] ) ∆[NO2]. At t
≈ 5 ms the linearity between the OH signal intensity and NO2
1
2
-3
consumption holds up to [NO2] ≈ 2 × 10 molecule cm
(Figure 2a).
2.3.4. HNO3. A direct calibration by preparing a standard
synthesized sample of HNO3 was not successful because of the
1
8
difficulties to introduce HNO3 into the reactor. Linearity of
the mass spectrometric signal from the 5% mixture of HNO3
vapor in He introduced into the reactor was not reached even
for the largest available flow rate corresponding to [HNO3] )
1
3
-3
5 × 10 molecule cm . That is why calibration of HNO3 was
made by producing HNO3 in situ via reactions 7 and 5. Time
profiles of NO2 and reaction product concentrations are pre-
sented in Figure 3. In this example the initial concentration of
13
-3
3
) × 10 molecule cm range in the main reactor. When the
NO2/He mixture was introduced through the movable injector,
the mass spectral response did not depend on the position of

6512 J. Phys. Chem. A, Vol. 109, No. 29, 2005
Butkovskaya et al.
uncertainty was within 10% including the experimental errors
in the measurement of signal intensities.
Additional possible uncertainty is connected with a branching
in reaction 5. According to recent studies of reaction 5, at 200
Torr and 298 K, formation of the HOONO isomer must be
considered since the branching ratio of channel 5b might be up
1
9-21
to 22%:
OH + NO + M f HOONO + M
(5b)
2
Although these data are in disagreement with those from other
studies [e.g., 22, 23] where HOONO was not observed by IR
2
4
spectroscopy, or with a LIF study confirmed by a priori
2
5
calculation where extremely high pressures were needed for
detection of HOONO, this possibility cannot be ignored. It was
21,24,26
suggested
that the isomer is a long-lived molecule, which,
then, might be detected in our experiments. The rate constant
of reaction 1i is close to the calculated ADO (average dipole
Figure 3. Concentration-time profiles of the species in the H + NO
2
reaction. Solid curves are the calculated profiles with initial concentra-
1
3
11
-3
12
tions [NO
2
]
0
) 1.5 × 10 and [H]
0
) 9.0 × 10 molecule cm
.
orientation) collisional rate constant. So, the sensitivity for
-
HOONO with detection using SF6 is expected to be either
13
-3
NO2 (1.5 × 10 molecule cm ) was sufficient to consume
most of the OH radicals. Figure 2b shows the dependence of
the HNO3 signal intensity on the change of OH concentration
at reaction times between 4 and 22 ms obtained from the data
presented in Figure 3. First, the OH signal was calibrated as
described above. The dependence is linear and allows us to
determine the sensitivity of the mass-spectrometric detection
for HNO3. In principle, to determine the branching ratio of
reaction 1 we do not need absolute sensitivities for NO2 and
HNO3 but only its ratio, which is directly provided by reaction
similar or less than that for HNO3. If HOONO is detected with
similar sensitivity as HNO3, the HNO3 calibration would not
be perturbed. If it does not give a signal at m/e 82, the HNO3
concentration would be overestimated. Hence, the uncertainty
in the knowledge of whether the ion-molecule reaction of
-
-
HOONO with SF6 can produce the NO3 ‚HF ion increases
the lower uncertainty limit in the HNO3 calibration to ap-
proximately 30%.
2
.3.5. Detection of HNO3 Using SiF4. The sensitivities for
HNO3 provided by both reactions 1i and 2i were determined
using the H + NO2 reaction system (see Figure 3). Concentra-
5. Yet, it is difficult to measure accurately small differences in
-
-
tions of SF6 and SiF5 ions in the IMR were monitored at
NO2 concentration (Figure 3), and calibration by OH signal is
preferable.
their isotopic peaks SF6- (m/e 148; 4.4% of the major isotope
34
32
-
29
-
28
-
SF6 ) and SiF5 (m/e 124; 5.1% of the major isotope SiF5 )
The errors inherent to the HNO3 calibration using reactions
because of the possible saturation of the signals of the main
isotopes. First, absolute sensitivities were measured with SF6
in the IMR as described above. They were (in units of molecule
7
-5 have been considered in detail. Figure 3 illustrates the
inaccuracies resulting from the incompleteness of reaction 5 and
unaccounted OH losses. At sufficiently high NO2 concentration
and long reaction time, the OH radicals produced in reaction 7
are totally converted into HNO3, and the HNO3 concentration
could be calculated as [HNO3] ) ∆[NO2]/2. In such a case
HNO3 can be calibrated without doing OH calibration. The
measured NO2 consumption at t ) 28 ms was ∆[NO2] ) 1.57
-
3
7
7
7
cm /cps) 6.6 × 10 (NO2), 1.2 × 10 (OH), and 1.4 × 10
(
HNO3). The measured signal intensities are presented in the
first line of Table 2. NO2 intensity corresponds to the absence
of reaction (discharge off), whereas HNO3 intensity corresponds
to the discharge on-off difference at maximal reaction time.
The second line in Table 2 shows the change of intensities when
SF4 was added to SF6 flow to give approximately equal amounts
1
2
-3
×
10 molecule cm . Rough estimation of formed HNO3 gives
11
-3
[
HNO3] ) ∆[NO2]/2 ) 7.9 × 10 molecule cm . Account
-
-
of the primary SF6 and SiF5 ions; the third line gives the
for the nonreacted OH (9%) gives [HNO3] ) ∆[NO2]‚(1-0.09)/
-
-
intensities measured in excess of SiF5 over SF6 . As the
concentrations in the main reactor were not changed, we obtain
a relative sensitivity for HNO3 detection with SF6 with respect
to SiF5 of 22.9 ( 1.0. This ratio somewhat exceeds the one
expected from the corresponding ion molecule reaction rate
11
-3
(
2-0.09) ) (7.5 ( 1.3) × 10 molecule cm , where the error
limits are defined by the noise of the NO2 signal. Using
calibration from Figure 2b, we obtain for the HNO3 final
-
-
11
concentration a very close value of (7.3 ( 0.4) × 10 molecule
-
3
cm , where the error limits are the 2σ of the slope of the
calibration curve in Figure 2b. Simulation of the kinetic profiles
of the products shows that the errors in HNO3 calibration using
OH signal is due mainly to the OH wall loss (∼5%), whereas
the OH + OH and OH + NO reactions account for another 4%
and 2%, respectively. The calculated time profiles with [H]0 )
-9
3
-1 -1 12
constants, k1i ) (2 ( 0.7) × 10 cm molecule
s
and k2i
-10
3
-1 -1 13a
)
(3.0 ( 1) × 10
cm molecule s , which can be
explained by the mass discrimination inherent to the instrument.
The obtained sensitivity ratio was compared with that obtained
in the HO2 + NO system (see section 3.1).
2
.3.6. Effect of H2O on the Signal Intensities of HNO3 (m/e
11
-3
9
× 10 molecule cm are shown in Figure 3. The reaction
82) and NO (m/e 46). A change of signal intensity at several
2
scheme used to simulate the H + NO2 chemical system is
masses including m/e 46 and 82 was observed when water vapor
was introduced into the reactor. This effect was examined for
-
1
presented in Table 1. The wall loss kw ) 14 ( 1 s was
measured in independent experiments. As a rule, calibration of
HNO3 was made using higher NO2 concentrations ([NO2] )
different conditions: (1) N carrier gas (background signals);
2
(2) adding NO and producing NO from O + NO reaction;
2
2
3
13
-3
(
1.6-2.5) × 10 molecule cm ), to get a faster completion of
(3) producing HNO from OH + NO reaction. Figure 4 shows
3
2
reaction 5 and thus to decrease wall loss and, with lower H-atom
the influence of different water concentrations on the intensity
of mass spectral peaks at m/e 46 and 82. The results can be
summarized as follows:
11
-3
concentration ([H] < 5 × 10 molecule cm ), to minimize
OH loss by self-reaction. The resulting HNO3 calibration

Nitric Acid Formation in Gas-Phase HO2 + NO Reaction
TABLE 1: Reaction Scheme Used to Simulate the H+ NO
J. Phys. Chem. A, Vol. 109, No. 29, 2005 6513
+ NO Chemical Systems
2
and HO
2
k (298 K, 200 Torr)
3
-1 -1
reaction
number in text
cm molecule
s
ref
I. H + NO
2
System
-
10
12
12
12
12
H + NO
2
f OH + NO
2
(7)
(5)
(4)
1.3 × 10
16
17
16
16
16
-
OH + NO
+ M f HNO
3
+ M
6.4 × 10
-
OH + NO + M f HONO + M
3.0 × 10
-
OH + OH + M f H
2
O
2
+ M
2.7 × 10
-
OH + OH f O + H
OH f wall
2
O
1.9 × 10
-
1
14 s
this work
II. HO
2
+ NO (+C
(1a)
(1b)
(5)
(4)
6
H12) System
-
12
14
12
12
12
10
HO
HO
OH + NO
2
+ NO f OH + NO
2
8.8 × 10
1.6 × 10
6.4 × 10
3.0 × 10
1.7 × 10
1.1 × 10
10
this work
17
16
16
16
this work
this work
51
29
29
29
29
30
-
-
-
-
-
2
+ NO + M f HNO
+ M f HNO
3
3
+ M
+ M
2
OH + NO + M f HONO + M
HO
OH + HO
OH f wall
HO f wall
OH + C
2
+ HO
2
f H
2
O
2
+ O
2
2
2
f H O + O
2
-
-
1
1
14 s
10 s
2
-
-
-
-
-
4
12
11
12
12
12
6
H
12 f C
+ M f C
6
H
11 + H
2
O
(6)
(8)
(9)
7.0 × 10
1.3 × 10
5.6 × 10
1.1 × 10
9.5 × 10
6
C H
6
C H
6
C H
6
C H
6
C H
11 + O
2
6
H
H
H
11
O
2
+ M
11O + NO
11ONO
11
11
11
O
O
O
2
2
2
+ NO f C
6
2
+ NO f C
6
2
+ NO
2
f C
6
H
11
O
2
NO
2
-1
11O f HC(O)(CH
2
)
4
CH
2
(10)
6.3 × 10 s
reactions 11-17: see text
-
-
2
TABLE 2: Detection of NO , HNO
3
and HONO Using Ion-Molecule Reactions with SF
6
and SiF
4
intensity (counts per second)
-
-
-
-
-
4
conditions
in IMR
SF
m/e 148
6
SiF
5
NO
2
HF‚NO
m/e 82
3
HNO
m/e 186
3
‚SiF
m/e 124
m/e 46
] ) 8.0 × 10 molecule cm-3
12
I. H + NO
2
Reaction System; [NO
219
2
only SF
6
442 843
122 846
30 179 ( 522
n.o.a
7
7
(
6.6 × 10 )
(1.4 × 10 )
-
-
[
[
SiF
SiF
5
] ≈ [SF
6
]
]
230 812
27 344
225 594
418 094
65 188
8949
14 677
1402
617
-
-
5
] . [SF
6
1316 ( 47
8
b
(
2
3.43 × 10 )
c
2.9 ( 1.0
] ) 1.4 × 10 molecule cm-3
13
2 2
II. H + NO Reaction System; [NO
only SF
6
722 651
140 223 6
b
n.o.
7
b
(
2.0 × 10 )
(3.9 × 10 )
-
-]
[
SiF
5
] ≈ [SF
6
423 893
81 542
3064
14
III. HO
2
+ NO Reaction System; [NO] ) 9.3 × 10 molecule cm-3
only SF
6
21 184
13 088
2096
321 ( 35
n.o.
-
-
[
[
SiF
SiF
5
] ≈ [SF
6
]
211
28.2
7.7 ( 2.5
-
-
5
] . [SF
6
]
15.5 ( 3.3
c
c
2
0.7 ( 5.0
IV. HO
2
+ NO Reaction System; [NO] ) 9.3 × 10 molecule cm-3
15
only SF
6
19 975
1986
129 ( 10
11.6
n.o.
[
SiF
5
-] . [SF -
6
]
7.0 ( 2.3
1
8.4 ( 6.1
a
n.o.” refers to “not observable”. Values in parentheses below intensities are reciprocal sensitivities in molecule cm-3/cps. c Ratio of sensitivities
b
“
-
-
5
and SiF .
for detection of HNO
3
with SF
6
(
1) NO2 and background signal intensities at m/e 46 decrease
In principle, formation of nitric acid from H2O and background
NO2 is possible on the walls of the reactor, IMR, or inlet tubes.
1
6
28
with increasing water concentration in the range (0-8) × 10
-
3
molecule cm . This decrease is explained by the general drop
of sensitivity when large concentrations of water are introduced
into the main reactor. Inspection of spectra showed that
introduction of water results in a decrease of SF6- ion
However, it was shown that HNO3 does not leave the surface
2
8
and, hence, cannot be detected in the gas phase. It is more
-
likely that the NO3 ‚HF ion is formed from the background
-
NO3 ion and HF molecules (or complexes) produced in the
1
7
concentration (nearly 30% decrease with [H2O] ) 1.1 × 10
IMR in the presence of H2O.
-3
-
molecule cm ). Although it is known that SF6 ions react with
H2O,27 the existing mechanism cannot explain such a large
3
. Results
-
decrease of SF6 concentration, and this phenomenon needs
further investigation.
3.1. Detection of HNO3 Using OH Scavenging by NO.
(
2) The signal of HNO3 decreases in the presence of H2O
similarly to that of NO2.
3) The background signal at m/e 82 increases in the presence
of H2O. There is no unambiguous explanation of this increase.
Figure 5 shows the changes of NO2 and HNO3 concentrations
13
with the increase of the concentration of NO from 1 × 10 to
1
5
-3
(
9.4 × 10 molecule cm . The measurements were done at L
) 45 cm (t ) 29 ms). As expected, NO2 concentration increases

6514 J. Phys. Chem. A, Vol. 109, No. 29, 2005
Butkovskaya et al.
-Forming
TABLE 3: Branching Ratio, â, for the HNO
3
Channel in the HO + NO Reaction at 298 K
2
a
I. OH scavenging by NO; R ) k
5
‚[NO
2
]
tot/k ‚[NO]
4
a
[
1
NO] [NO
2
]
bgr ∆[NO
2
]
[NO
10
2
]
tot ∆[HNO
3
]
â
%
I
82/I46
%
R
15
11
11
11
8
0
10
10
10
%
3
.6
.1
.9
.4
.0
0.94
2.7
6.0
4.9
6.0
1.3
1.1
1.8
1.4
5.0
3.1
5.4
1.3
1.2
3.3
0.7
3.3
5.0
4.8
3.9
6.0
2.2
2.3
5.2
5.7
8.5
14.4
4.9
3.9
4.6
1.5
1.3
4.2
7.7
10.8
8.8
12.0
3.5
3.3
7.0
7.1
13.5
17.5
10.3
5.2
6.9
10.5
8.5
8.0
11.6
3.9
4.9
5.4
7.5
12.2
16.1
8.5
5.1
6.9
2.9
1.7
0.208 1.34 0.025
0.213 1.75 0.053
0.180 1.07 0.059
0.205 1.54 0.020
0.194 1.33 0.128
0.176 0.59 0.006
0.216 0.70 0.008
0.104 0.40 0.017
0.133 0.50 0.024
0.143 0.60 0.046
0.112 0.52 0.053
0.171 0.65 0.013
0.129 0.57 0.012
0.150 0.70 0.013
0.197 0.92 0.011
0.137 0.60 0.004
3
3
9
2
1
2.3
9
9
6
6
7
7
9
9
9
.4
.3
.2
.2
.0
1
1
.3
.3
.3
5.8
4.8
2.0
Figure 4. Water effect on mass spectrometric signal intensities: added
NO (2), NO produced in the O + NO reaction (*), HNO produced
in the OH + NO reaction (9) and background at m/e 82 (0).
2
2
3
3
0.8
2
average: 0.166 ( 0.037
a
II. OH scavenging by C
6
H
12; R ) k
5
‚[NO
2
]tot/k
6
6 12
‚[C H ]
b
[
C
1
6
H
0
12] [NO
2
]
bgr ∆[NO
2
] [NO
2
]
tot ∆[HNO
3
]
â
%
I
82/I46
%
R
1
5
11
11
11
8
10
10
10
10
%
1
3.2
.6
4.2
25
1.1
2.1
10.4
2.2
1.4
35
3.3
4.8
20.0
3.1
2.9
0.191 1.22 0.056
0.162 0.81 0.008
0.203 0.95 0.0062
3
1
average: 0.185 ( 0.017
a
R is the upper limit of the probability of HNO
3
formation in the
reaction; concentrations are in units of molecule cm .
-3
side OH + NO
2
determined from the yield of HNO3 in reaction 5, as discussed
above. In this way, systematic errors in the determination of
NO2 concentration do not affect the obtained branching ratio.
At [NO] > 5 × 10 molecule cm-3 the relative rate of HNO3
15
formation from OH in reaction 5 is less than R ) k5‚[NO2]/k4‚
Figure 5. Production of NO
a function of NO concentration at 298 K. Dotted curves are separately
calculated for HNO from reactions 5 and 1b; solid curve is their sum.
2 3 2
and HNO in the HO + NO reaction as
-
4
[
NO] ) 1.6 × 10 , which is less than 10% of the measured
HNO3 yield, and was neglected. The range of the measured
3
8
HNO3 concentrations from reaction 1 was (3-16) × 10
-
3
until the complete conversion of HO2 to NO2 in reaction 1a,
molecule cm , well above the HNO detection limit discussed
3
1
1
reaching a constant value of [NO2]f ) 3.9 × 10 molecule
above.
-
3
cm . The increase of HNO3 concentration due to reaction 5
A strong indication that the observed product is effectively
HNO3 was obtained from comparison of its detection sensitivi-
OH + NO + M f HNO + M (5)
2
3
-
-
ties using SF6 and SiF5 , with that for HNO3. As the detection
-
-
limit increases strongly with switching from SF6 to SiF5 ,
detection of the product from 1b with SiF5 was made after a
is followed by its decrease because of more important consump-
tion of OH in reaction 4 at higher [NO]
-
careful optimization of the parameters of the ion source and
ion optics. This allowed us to attain the sensitivities of 2.0 ×
OH + NO + M f HONO + M
(4)
7
-3
6
-3
1
0 molecule cm /cps (NO2) and 3.9 × 10 molecule cm /
8
-
down to a constant nonzero level of [HNO3]f ) 8 × 10
cps (HNO3) when they were detected by reaction with SF6 .
-
3
molecule cm . The branching ratio for the HNO3-forming
channel of the primary HO2 + NO reaction was determined as
a ratio of concentrations of the formed HNO3 to the formed
These values were obtained from the calibration using H + NO2
1
3
-3
reaction with [NO2] ) 1.4 × 10 molecule cm (see part II
-
-
in Table 2). When half of SF6 ions were converted to SiF5 ,
the signals changed as in the previous experiment (part I). The
signal intensities at m/e 46, 82, and 186 measured in the course
of the HO2 + NO reaction (discharge on-off difference) are
presented in parts III and IV of Table 2.
-
3
NO2, â ) [HNO3]f/[NO2]f ) 2.05 × 10
at high NO
concentrations corresponding to the total consumption of OH
in the reaction 4. Table 3 summarizes the measurements at 298
K in the absence of water. The experiments which have been
performed during a long period are placed in chronological
order. A change of the relative sensitivity (I82/I46) is clearly seen
after a change of the operating regime of the ion source. The
average value obtained from 16 experiments with excess NO
14
Part III of Table 2 corresponds to [NO] ) 9.3 × 10
-3
molecule cm with incomplete scavenging of OH radicals. Here
the HNO3 signal consists of the contributions from both
reactions 1b and 5. The sensitivity ratio of this combined signal
+
0.40
-3
-
-
is â ) (1.66- ) × 10 . The indicated error is a combination
with respect to detection by SF6 to SF5 is 20.7 ( 5.0. Part
0.60
-
3
15
-3
of the mean standard deviation (σ ) 0.32 × 10 ) and the
IV of Table 2 corresponds to [NO] ) 9.3 × 10 molecule cm ,
systematic uncertainty of the ratio of HNO3 to NO2 sensitivities
and the sensitivity ratio becomes 18.4 ( 6.1. The latter two

Nitric Acid Formation in Gas-Phase HO2 + NO Reaction
J. Phys. Chem. A, Vol. 109, No. 29, 2005 6515
values can be compared with the sensitivity ratio of 22.9 ( 1.0
obtained for HNO3. Although the error limits of the obtained
sensitivity ratios are rather high, it is likely that only one and
the same compound is detected in the three cases.
3
.2. Detection of HNO3 Using OH Scavenging by C6H12.
Kinetic behavior of HO2 radicals and the OH, NO2, and HNO3
reaction products in the HO2 + NO chemical system with [NO]
1
3
11
-3
)
2.5 × 10 and [HO2] ) 2.8 × 10 molecule cm is shown
in Figure 6a. The measured concentrations well agree with the
results of the numerical integration presented by solid curves.
Reactions used in the calculation are listed in Table 1. The
changes in the kinetic curves shown in Figure 6b are a result of
the addition of about 4 × 10 molecule cm of cyclohexane
to the reaction mixture. The most important change is the
increase of NO2 concentration by more than a factor of 3.
Formation of extra NO2 is explained by the subsequent reactions
of the cyclohexyl radical, C6H11, in the presence of NO and
15
-3
1
5
-3
large concentrations of O2 (3.3 × 10 molecule cm ):
OH + C H f C H + H O
(6)
(8)
(9)
6
12
6
11
2
C H + O + M f C H O + M
6
11
2
6
11
2
C H O + NO f C H O + NO
2
6
11
2
6
11
Platz et al.29 have shown that the cyclic C6H11O radical under
our conditions isomerizes to the linear HC(O)(CH2)4 C˙ H2 radical:
C H O f HC(O)(CH ) C˙ H
2
(10)
6
11
2 4
Figure 6. Concentration-time profiles of the species in the HO +
2
NO reaction. (Bottom panel) In the presence of C
correspond to the absence of the chain (see text).
6
H12. Dotted curves
The further oxidation mechanism of this radical in excess of
NO over NO2 is the following:30
were that of reaction 9, and a 16% branching ratio for formation
of nitrates in reactions 9, 12, and 15 was accounted in the
HC(O)(CH ) C˙ H + O f HC(O)(CH ) CH O (11)
2
4
2
2
2 4
2
2
2
9
-11
3
-1
calculation. A value of k17 ) 2.6 × 10
cm molecule
HC(O)(CH ) CH O + NO f HC(O)(CH ) CH O + NO
2
4
2
2
2 4
2
2
-1
s
was taken for reaction 17 by analogy with the reaction C2H5-
(12)
3
1
CHOH + O2 f C2H5CHO + HO2. Dotted curves in Figure
6 were calculated with k17 ) 0 and show concentrations in the
absence of the chain, with HNO3 only from the primary step.
The branching ratio of reaction 1b can be determined as a
ratio of the HNO3 to NO2 produced in the initial period of
reaction 1 (Figure 6b). The four initial points from this
HC(O)(CH ) CH O f HC(O)CH C˙ H(CH ) CH OH (13)
2
4
2
2
2 2
2
HC(O)CH C˙ H(CH ) CH OH + O f
2
2 2
2
2
HC(O)CH CH(OO)(CH ) CH OH (14)
2
2 2
2
HC(O)CH CH(OO)(CH ) CH OH + NO f
-3
2
2 2
2
experiment give an average value of (1.62 ( 0.22) × 10 .
HC(O)CH CH(O)(CH ) CH OH + NO (15)
Formation of HNO3 in OH + NO2 reaction even for the highest
NO2 concentration formed in this experiment was R ) k5‚[NO2]/
2
2 2
2
2
HC(O)CH CH(O)(CH ) CH OH f
-5
2
2 2
2
k6‚[C6H12] ) 7.8 × 10 , which is less than 5% of the obtained
HC(O)CH CH(OH)(CH ) C˙ HOH (16)
branching ratio. Table 3 presents results of the three experiments
2
2 2
with cyclohexane. The average branching ratio is (1.85 ( 0.17)
HC(O)CH CH(OH)(CH ) C˙ HOH + O f
-3
2
2 2
2
× 10 , in good agreement with that obtained from OH
+
0.04
HC(O)CH CH(OH)(CH ) CHO + HO (17)
scavenging by NO. A mean value of (0.18
)% from both
-0.06
2
2 2
2
methods with error limits including uncertainty in the calibration
is recommended.
Reactions with O2 producing HO2 compete with the isomer-
ization reactions 13 and 16. Termination reactions with NO and
NO2 are not presented for simplicity. HO2 produced in reaction
3.3. Detection of HNO3 in the Presence of Water Vapor.
Water vapor was introduced in the reactor upstream of the
injector position fixed at a distance of 50 cm. In one of the
experiments, mass spectral response at m/e 46 and 82 at different
water concentrations was measured with OH scavenging by
cyclohexane. The initial concentrations in this experiment were
1
7 will regenerate OH through reaction 1 and, thus, creates a
chain, in which one HO2 molecule is produced per four NO2
molecules (this ratio can somewhat differ due to omitted
reactions). The increase of HNO3 signal with time agrees with
its supposed formation in the primary reaction 1. Solid curves
in Figure 6b were calculated with the rate constants given in
Table 1. As many rate constants in the cyclohexane oxidation
1
5
13
[C6H12] ) 3.9 × 10 , [NO] ) 2.5 × 10 , and [HO2] ) 2.2 ×
1
1
-3
8
10 molecule cm . Bohm and Zetzsch determined that the
rate constant of reaction 1 does not change in the presence of
water which means that the concentration of the formed NO2
in the predominant channel also does not change in the presence
of water. The observed NO2 signal decreased and background
at m/e 82 increased in agreement with the calibration tests. The
dependence of the branching ratio k1b/k1a on [H2O] was
-
12
are not known, assumptions were made. A value of 7 × 10
3
-1 -1
cm molecule was used for the radical + O2 addition
s
reactions 11 and 14. The rate constants of the isomerization
reactions 13 and 16 were assumed to be the same as that of
reaction 10. The rate constants taken for reactions 12 and 15

6516 J. Phys. Chem. A, Vol. 109, No. 29, 2005
Butkovskaya et al.
TABLE 4: Branching Ratio, â, for Formation of HNO
3
in the HO
+ NO Reaction at 298 K in the Presence of Watera
2
1
5
13
12
8
I. [C
6
H
12] ) 3.9 × 10 ; [NO] ) 2.5 × 10 ; ∆[NO
2
] ) 1.3 × 10 ; ∆[HNO
3
] )8.0 × 10
[
H
2
O]
0
1.4
2.8
4.3
5.9
8.1
0.281
â (%)
0.162
0.205
0.238
0.247
] ) 5.0 × 10
4.55
0.225
] ) 8.5 × 10
3.71
0.166
] ) 3.3 × 10
9.35
0.310
] ) 4.9 × 10
0.273
1
5
11
II. [NO] ) 3.1 × 10 ; ∆[NO
2
[H
2
O]
0
1.09
0.197
2.33
5.95
0.263
â (%)
0.213
0.207
1
5
11
III. [NO] ) 6.2 × 10 ; ∆[NO
2
[H
2
O]
0
1.07
0.147
2.26
5.39
0.176
6.86
0.195
7.62
0.216
â (%)
0.143
0.157
1
5
11
IV. [NO] ) 4.5 × 10 ; ∆[NO
2
[H
2
O]
0
5.96
0.295
9.13
9.77
0.328
11.7
0.393
â (%)
0.208
0.343
15
11
V. [NO] ) 17 × 10 ; ∆[NO
2
[H
2
O]
0
6.0
11.0
â (%)
0.171
0.215
0.271
15
11
VI. [NO] ) 9.3 × 10 ; ∆[NO ] ) 4.6 × 10
2
[
H
2
O]
0
5.07
0.212
â (%)
0.150
a
Concentrations are in units of molecule cm-3; water concentration is in units of 1016 molecule cm-3
.
Figure 7. Water effect on the branching ratio of the HNO
3
-forming
channel of the HO
experiments.
2
+ NO reaction. Different symbols denote separate
determined from the ratio of the HNO3 to NO2 signals assuming
independence of the ratio of NO2 to HNO3 sensitivities on the
concentration of H2O. The results of this experiment are given
in the first entry of Table 4. All other measurements presented
in Table 4 were done employing NO as an OH quencher. Figure
7
is a plot of the measured values of â against [H2O]. Each
Figure 8. Temperature dependence of the branching ratio of the HNO
forming channel of the HO + NO reaction (a) and Arrhenius plot for
1b (b).
3
-
experiment from Table 4 is presented in different symbols.
Figure 7 shows that â linearly increases with the increase of
water concentration with the maximal increase by 90% at [H2O]
2
k
TABLE 5: Temperature Dependence of the Branching
17
-3
)
1.17 × 10 molecule cm .
.4. Temperature Dependence of the Branching Ratio. The
branching ratio of reaction 1 at temperatures between 298 and
23 K was measured in order to determine the rate constant of
Ratio, â ) k /k , and the Rate Constant k for HO + NO
1b
1a
1b
2
Reactiona
3
(10-12)b
1b (10-14
)
T (K)
â (%)
k
1
k
2
2
2
2
98
83
73
0.18 ( 0.06
0.22 ( 0.06
0.30 ( 0.06
0.35 ( 0.07
0.41 ( 0.07
0.52 ( 0.07
0.71 ( 0.08
0.87 ( 0.08
8.8 ( 1.2
9.3 ( 1.3
9.6 ( 1.3
9.9 ( 1.4
10.3 ( 1.4
10.8 ( 1.5
11.3 ( 1.6
11.9 ( 1.7
1.6 ( 0.5
2.0 ( 0.6
2.9 ( 0.7
3.5 ( 0.8
4.2 ( 0.9
5.6 ( 1.0
8.0 ( 1.4
the HNO3-forming channel for conditions relevant to the upper
troposphere. Three experiments were performed at a high NO
15
-3
concentration of [NO] ) 9.3 × 10 molecule cm and reaction
time from 15 to 30 ms. A constant pressure of 200 Torr was
kept in the reactor when the main flow was cooled. Figure 8a
represents the obtained dependence â(T), while Figure 8b shows
the data in Arrhenius form. Different symbols denote separate
experiments. The cumulative results of these experiments are
given in Table 5. Nearly a 5-fold increase of the branching ratio
â was found when the temperature was lowered across this
range. It was found that the rate constant increases from
263
2
2
2
2
53
43
33
23
10.4 ( 1.7
a
Rate constants are in units of cm molecule s-1. b Reference 10.
3
-1
(10.4 ( 1.7) × 10 cm molecule s-1 at 223 K. In the 223-
300 K range the rate constant of the HNO3-forming channel
can be expressed as
-
14
3
-1
-
14
3
-1 -1
k1b ) (1.6 ( 0.5) × 10
cm molecule
s
at 298 K to

Nitric Acid Formation in Gas-Phase HO2 + NO Reaction
J. Phys. Chem. A, Vol. 109, No. 29, 2005 6517
k ) 6.4 × 10 exp((1644 ( 76)/T) cm molecule s-1
-
17
3
-1
1b
The mechanism of HNO3 formation explaining the obtained
strong negative temperature dependence of k1b and its atmo-
spheric impact are discussed below.
4
. Discussion
.1. Other Possible Contributions to the Signal at m/e 82.
4
To assign the signal observed at m/e 82 during reaction 1 entirely
to the products of the gas-phase reaction, we had to verify that
switching on the discharge did not lead to the appearance of
other contributions to this mass peak. The possible causes of
such contributions could be unknown processes in the IMR (ion
source), formation of nitric acid from the trace NO2 in the
injector, and reactions on the surface of the reactor. Different
tests were done which showed that the discharge products do
not interfere with signal at m/e 82. In particular, this was checked
for the discharge in He or H2/He flow, with or without O2 in
the injector. As well, no change of signal at m/e 82 was observed
in the absence of O2 flow through the injector with turning on
the discharge in He or the H2/He mixture when NO was flowing
through the reactor. Although formation of small amounts of
O-atoms and/or OH radicals was sometimes registered in the
discharge of He flow, it means that possible oxygen and water
impurities in He did not lead to formation of HNO3. It also
means that possible formation of HNO3 from the trace back-
ground NO2 (e.g., degassing of viton inlet tubings) was below
the detection limit.
Figure 9. Pressure dependence of the branching ratio of the HNO
3
-
forming channel of the HO + NO reaction at 298 K with [HO ] ) 5
2
2 0
11
-3
× 10 molecule cm . The error limits correspond to systematic errors
in the measurement of the signals and uncertainties in the calibration
3 2
of the HNO to NO signal ratio.
hence, the HNO3 formation cannot be explained by the
possibility for radicals to reach the surface during the initial
mixing process. We have performed a large number of experi-
ments in which the influence of the injector flow rate on the
mixing of injector and main reactor components has been
investigated. As a result of these experiments, we use the flow
rate which corresponds to the most homogeneous distribution
at shortest reaction times. The experiments were performed with
stable species (NO, NO2, O3) as well as with radicals (OH, F,
O) being flowed through the injector. Accounting for the
homogeneous losses in the case of the radicals, the observed
concentration profiles allow us to exclude the possibility of
radical losses during the initial mixing process at the used
injector flow rates.
Two mechanisms can be considered for the possible formation
of HNO3 on the reactor wall: (i) the diffusion of the radicals
(OH or HO2) to the wall and their conversion to HNO3 on the
surface and (ii) the transport of the radicals to the surface during
the initial mixing process.
Another question is a possible contribution to the signal at
m/e 82 from NO2 produced in reaction 1. This was tested using
The former mechanism (i) can be excluded on the following
basis: the NO concentration in our experiments was higher than
1
6
-3
the reaction of ozone with NO yielding NO in situ under
10
molecule cm corresponding to first-order reaction rates
2
higher than 10 s-1 for both OH and HO2 radicals. The
5
conditions similar to that used during the study of reaction 1
including maintenance of the same O2 and H2 flow rates:
maximum wall loss rates observed in our experiments for these
-
1
radicals were less than 15 s . Comparing these rates, the
maximum yield of any product of radicals conversion on the
O + NO f NO + O
2
(18)
3
2
-
4
wall of the reactor cannot be higher than 2 × 10 , and hence,
-
3
this mechanism cannot be responsible for the observed 2 × 10
Reaction 18 was held in a large excess of NO, and different
concentrations of NO2 were obtained by varying the ozone
concentration. No change of the background level at m/e 82
yield of HNO3 at 298 K. Another argument supporting the
unimportance of this mechanism is the pressure dependence of
the HNO3 yield. Figure 9 presents the branching ratio measured
in excess of NO at pressures of 100, 200, 300, and 400 Torr.
The measurements were done at room temperature with [NO]
1
1
12
was observed in the range of [O3] ) 3 × 10 -5 × 10
-3
molecule cm . These experiments proved that formation of NO2
in the reactor does not yield an increase of the signal at m/e 82.
4.2. Mechanism of the HNO3 Formation and Theoretical
Models. The HO2 + NO reaction is the first member of the
RO2 + NO reaction family, proceeding via the ROONO
intermediate with subsequent branching to RO + NO2 or to
RONO2 products. It is interesting to discuss the obtained
branching ratio for HNO3 formation in reaction 1, k1b/k1a )
16
11
-3
)
(0.9-1.2) × 10 and [HO2]0 ) 5 × 10 molecule cm .
The results show positive linear pressure dependence incompat-
ible with the heterogeneous formation of HNO3. Under the
conditions of our turbulent reactor, the limiting process deter-
mining the rate of wall loss of the radicals is the diffusion of
the radicals through the laminar layer near the reactor wall. As
the turbulent conditions in the reaction region are not stationary
3
3-37
0.18% at 298 K, in view of the existing experimental
theoretical
and
3
8-43
(that is the turbulence parameters are not constant along the
studies of the RONO2 yields from alkyl peroxy
reactor), the thickness of this layer cannot be easily estimated
theoretically. However, assuming that the thickness is indepen-
dent of pressure, the wall loss should be slower at higher
pressures, and consequently, the yield of HNO3, in case of its
formation from radical conversion on the surface, would be
lower.
The latter mechanism (ii) can be ruled out taking into account
that in the HO2 + NO + C6H12 system the radicals are
regenerated along the length of the entire reaction zone, and
radical reactions with NO. Table 6 shows experimental results
of the RONO2 formation for small n-alkyl peroxy radicals. Very
small yields were observed for C2 and C3 nitrates. Starting from
C5, the yield increases monotonically approaching a limit of
∼35% at 298 K for large n-alkyl peroxy radicals. As has been
already mentioned, formation of HNO3 in the HO2 + NO
reaction has never been observed previously, and early ab initio
calculations explained this by a high barrier for isomerization
of HOONO to HNO3.
37

6518 J. Phys. Chem. A, Vol. 109, No. 29, 2005
Butkovskaya et al.
TABLE 6: Literature Data on RO
00 Torr
2
+ NO Branching Ratios for RONO
2
Formation, â, from the Product Studies at 100 < P <
2
RO
reactant
2
detected
products
â(T)
T, K
â(T)
T, K
P, Torr
200
100
100-200
100
method
ref
HO
CH
2
0.0018
<0.03
0.1
0.006
0.006
0.038
298
295
300
298
298
327
0.0087
223
HNO
3
HPTFR/CIMS
HPTFR/CIMS
HPTFR/CIMS
HPTFR/CIMS
HPTFR/CIMS
PR/GC FIDa
this work
3
O
2
CH
CH
3
ONO
2
32
33
34
35
36
<
<0.1
193
213
213
284
3 2 2
O , NO
C
2
C
3
C
5
H
H
H
5
O
O
2
0.02
0.02
0.068
C
C
C
2
3
5
H
H
H
5
ONO
ONO
2
7
2
100
155
7
2
11
O
2
11ONO
2
a
Photoreactor and gas chromatography with flame ionization detection.
arises from the cis-HOONO by O-O cleavage and involves
the hydrogen bonded OH-ONO intermediate which recombines
to nitric acid without a significant barrier. The obtained CBS-
-
1
QB3 barrier height is 18-19 kcal mol .
Even lower barriers were postulated for isomerization transi-
tion states in master equation studies of alkyl nitrate formation
4
1,43
in RO2 + NO reaction.
To simulate experimental results
acquired at different temperatures and pressures, the barrier
-
1
height had to be at least 1 kcal mol below the RO + NO2
asymptotic energy. Then, correct pressure and temperature
dependences could be obtained. Specifically, it was obtained
that RONO2 yields increase drastically with decreasing tem-
perature. Though in these works the yields were calculated for
RdCH3 and heavier reactants, the general consideration included
Figure 10. Energy diagram for HO
2
+ NO system. TS2, transition
RdH. The important conclusion from both studies4
1,43
is that
state for isomerization as calculated in refs 38 (a) and 39 (b); TS3,
22,42
the temperature effect is determined by the energy difference
between TS1 and TS3. Calculations of Zhao et al.42 support a
possibility that the isomerization barrier lies below the dissocia-
tion limit, which can explain the negative temperature depen-
dence for the HNO3 yield observed in the present study.
transition state for isomerization.
ref 44.
Other indicated energies are from
Figure 10 presents the energy diagram for the HO2 + NO
system. Reaction starts with formation of the HOONO complex
-
1
bound by 26 kcal mol as calculated by Li and Francisco at
the QCISD(T)cc-pVQZ level of theory.44 It is important that
the complex has two major conformations: cyclic cis-structure
4
.3. Mechanism of HNO3 Formation in the Presence of
H2O and Chemical Amplifiers. To analyze the role of water
in the HO2 + NO reaction, formation of HO2‚H2O complex is
assumed:
-
1
1
and nearly linear trans-structure separated by a 16 kcal mol
barrier (not shown), trans-conformer lying ∼3 kcal mol
-
higher.2
1,22
The dissociation energy is 18 ( 1 kcal mol .
-1 21,44
HO + H O T HO ‚H O
(19)
2
2
2
2
This energy diagram is generic for the RO2 + NO family, the
dissociation energy varying systematically, with H highest, CH3
lowest, and C2 and heavier R between.43 Activated HOONO**
complex can unimolecularly decompose to OH + NO2 (TS1)
or isomerize to HNO3 (TS2). A key parameter for the isomer-
ization/dissociation branching ratio is a barrier of the transition
state for isomerization (TS2). In the early study employing the
MP4SDQ/6-31G* method, the TS2 was found to lie 60 kcal
The geometry of HO2‚H2O complex was determined at a high
46
level of ab initio calculations and presents a ring structure
-1
with two hydrogen bonds of 6.9 kcal mol . The equilibrium
-19
3
-1
constant for reaction 19 is K19 ) 3 × 10
cm molecule at
47
2
98 K. At water partial pressure of 3 Torr, [HO2‚H2O]/[HO2]
K19 [H2O] ) 0.036; i.e., about 4% of HO2 in the reactor are
)
in the form of the complex. It can be calculated that the observed
effect of approximate doubling the HNO3 yield in reaction 1
with 3 Torr of H2O can be achieved if the yield of HNO3 in the
HO2‚H2O + NO reaction is as high as ∼5%. Let us consider
possible mechanisms.
-
1
45
mol higher than HOONO. The calculation by Sumathi and
3
8
40
Peyerimhoff and more recently by Lohr et al., both using
-
1
the DFT B3LYP/6-31G++** method, gave 39.0 kcal mol .
Calculation of Dixon et al. at the CCSD(T) level of theory gave
-1 39
a substantially lower barrier height of 21.4 kcal mol . In all
these calculations the reaction path was a direct migration of
the OH group from oxygen to nitrogen in trans-conformer via
a three-centered transition state. Although according to the latter
calculation this transition state is energetically accessible, it
cannot explain a negative temperature dependence of k1b as
observed in the present work. A possible explanation of this
The NO approach from the side of any oxygen atom
belonging to HO2 will most probably lead to elimination of
water and formation of HOONO complex bound by ∼26 kcal
-1 44
mol
and probably a less energized HOONO complex than
that formed in reaction 1 because H2O molecule can carry away
a part of the released energy:
2
2
dependence can be found in the study of Dransfield et al.
HO ‚H O + NO f [HO NO‚H O]* f HOONO* + H O
2
2
2
2
2
where the isomerization to HNO3 via an OH + NO2 productlike
transition state (TS3) has been suggested. They propose that at
TS3 the barrier to internal rotation lies below that of dissociation,
and the species may rotate from an HOONO conformation to
an HONO2 conformation and relax (dotted arrow in Figure 9).
Quite recently, Zhao et al.42 optimized the minima and transition
states (stationary points) on such reaction path using UCCSD-
(
20)
This case is similar to the catalytic role of water in the HO2
self-reaction, where an H2O molecule helps to produce a less
excited (HO2)2 dimer, increasing the rate of formation of H2O2
48
+ O2 products. Like in reaction 1, low-energized HOONO*
complex in reaction 20 can decompose to OH + NO2 or
(T)/6-31+G* and CBS-QB3 methods. In their study the TS3
isomerize to HNO3. Then, the question is how the isomerization

Nitric Acid Formation in Gas-Phase HO2 + NO Reaction
J. Phys. Chem. A, Vol. 109, No. 29, 2005 6519
a
to decomposition branching ratio depends on the activation
energy of HOONO complex. Recent theoretical studies of the
RO2 + NO reactions41,43 suggest that decomposition dominates
at initial activation energy (HOONO**), but with a decrease
of the excitation, dissociation rate constant will decrease more
rapidly than the isomerization rate constant, so that at low
energies (below the dissociation limit) isomerization dominates.
Evidently, HOONO* activation is somewhere between the
reactant (HO2 + NO) and radical product (OH + NO2) energy
limits, and we observe a change of the branching ratio from
TABLE 7: Measured Concentrations of Species and Main
x
HO Loss Rates in the Upper Troposphere
concentration (pptv)
species
OH
marine air
continental air
0.1
4.7
49
0.1
1.1
HO
NO
2
443
163
130
63
NO2
HNO
11
44
40
3
HO
2
NO
2
0
.18% (dry case) to ∼5% (participation of H2O). Explicit
b
rate constant
loss rate molecule
calculations of the microcanonical rate constants for decomposi-
tion and isomerization as a function of energy for R ) C5H11
indicate that this change is realistic for RdH.
As an alternative mechanism, it was suggested that H2O
promotes production of HNO3 in a peroxy radical amplifier by
lowering the rearrangement barrier:4
cm molecule s-1
3
-1
-3
HO
OH+HO
HO +HO
OH + HO
x
loss
cm /s
41,43
-
10
4
3
2
3
2
f H
f H
NO
2
O+O
2
1.5 × 10
4.2 × 10
6.4 × 10
1.2 × 10 2.8 × 10
-12
-12
4
2
2
2
O
2
1.6 × 10 8.9 × 10
3
2
2
f
2.2 × 10 3.5 × 10
H O+NO +O
2
2
2
-12
2
4
2
OH+NO
2
f HNO
3
7.4 × 10
5.8 × 10
7.1 × 10 1.0 × 10
-
13
2
OH + HNO
3
f
2.2 × 10 6.6 × 10
2
H O + NO
3
HO ‚nH O + NO f HONO + nH O
(21)
2
2
2
2
4
4
Total:
3.1 × 10
1.0 × 10
1.8 × 10
-
13c
3
3
HO
HO
(
2
+ NO f HNO
+ NO f HNO
3
3
2.0 × 10 4.3 × 10
The mechanism with n ) 2 was developed to explain a quadratic
decrease of chain length with the increase of relative humidity
found in that study. However, this disagrees with our results
giving linear dependence of the HNO3 yield on water concentra-
tion. Generally, mechanism of such concerted isomerization with
n ) 1 can be imagined assuming a seven-membered transition
state, and additional ab initio calculations are needed to test
this possibility. Michele et al. also reported a linear decrease of
the chain length in the radical amplifier in the presence of wet
2
6.4
23
% of total loss)
a
From the SONEX aircraft campaign (T ) 223 K and P ) 260
Torr).50 Rate constants are from ref 10. c This work.
b
The situation can be different in the upper troposphere, due
to the much higher values of the [NO]/[NO2] ratios and of the
rate constant of reaction 1b at low temperature. Considering
the UT at an ∼10 km altitude with typical conditions T ) 223
K and P ) 260 Torr, we have used concentrations data of the
2
-13
air. They calculated that a rate coefficient of about 7 × 10
3
-1 -1
cm molecule
s
for the second-order reaction of the HO2‚
50
SONEX campaign to compare reaction 1b with the known
H2O adduct with NO giving a nonradical product would explain
their observations. Since the HO2 + NO reaction rate does not
change in the presence of water, it means that the required
HOx loss processes. Table 7 contains the results of the aircraft
measurements for low-NOx (marine) and high-NOx (continental)
atmospheric regimes as well as calculated rates for the most
important reactions leading to the loss of HOx. We see that
reaction 1b contributes 23% of the total HOx loss for high-NOx
conditions. Even in the low-NOx regime with predominant loss
via the OH + HO2 and HO2 + HO2 reactions, the HO2 + NO
loss pathway is not negligible and amounts to about 6% of the
total loss.
In conclusion, the HNO3-forming channel of the HO2 + NO
reaction observed for the first time in this work could be a
significant HOx loss process in the UT with a potential impact
on ozone production. The impact of this reaction will be less
in the lower troposphere.
-
13
-12
branching ratio k20/k1a should be about 7 × 10 /8.8 × 10
8% which is consistent with our value of ∼5%.
.4. Atmospheric Implication. Channel 1b forming HNO3
)
4
in the HO2 + NO reaction becomes a new potentially significant
loss process for the HOx (OH, HO2) radicals in the atmosphere.
In the troposphere the OH and HO2 radicals are recycled through
the reactions
OH + CO, CH , NMHC + O f HO
2
4
2
HO + NO f OH + NO2
(1a)
2
and reaction 1a leads to ozone formation through photolysis of
NO2 and further reaction of the formed O-atoms with O2. Hence,
the HOx loss processes reduce ozone formation [e.g., 49].
In the lower troposphere the major HOx loss under NOx “rich”
conditions is reaction
Acknowledgment. The study was performed partly in the
frame of the SCOUT-O3 integrated EU project. The authors
are grateful to Dr. Gerard Laverdet for discussions on chemical
amplifier measurements and technical help in experiments.
OH + NO + M f HNO + M
(5)
References and Notes
2
3
(
1) Hastie, D. R.; Weissenmayer, M.; Burrows, J. P.; Harris, G. W.
nitric acid being taken up by water aerosols and mostly washed
out by precipitations. The significance of reaction 1b as the HOx
loss process can be assessed by calculating the relative rate ratio
of reactions 1b and 5: r ) k1b [HO2][NO]/k5 [OH][NO2]. Such
expression is valid considering the OH/HO2 conversion being
Anal. Chem. 1991, 63, 2048.
(2) Mihele, C. M.; Mozurkewich, M.; Hastie, D. R. Int. J. Chem. Kinet.
999, 31, 145.
1
(3) Coll, I.; Pinceloup, S.; Perros, E. P.; Laverdet, G.; Le Bras, G. J.
Atmos. Res. 2005, 74, 477.
(4) Reichert, L.; Hernandez, M. D. A.; St o¨ bener, D.; Burkert, J.;
Burrows, J. P. J. Geophys. Res. 2003, 108, 4017.
-
14
3
3
fast compared to HOx loss. Using k1b ) 1.6 × 10
cm
(
(
5) Howard, C. J.; Evenson, K. M. Geophys. Res. Lett. 1977, 4, 437.
6) Howard, C. J. J. Chem. Phys. 1979, 71, 2352.
-
-
1
1
-1
-11
molecule
s
at 298 K from this work, k5 ) 1.1 × 10 cm
molecule s-1 at 298 K, 1 atm, and typical values for the
lower troposphere [NO]/[NO2] ≈ 0.5 and [HO2]/[OH] ≈ 50,
we obtain r ≈ 2%. Even considering the rather large variability
of [NO]/[NO2] and [HO2]/[OH] ratios, reaction 1b should not
contribute significantly to the HOx loss in the lower troposphere.
17
(7) Seely, J. V.; Meads, R. F.; Elrod, M. J.; Molina, M. J. J. Phys.
Chem. 1996, 100, 4026.
(
8) Bohn, B.; Zetzsch, C. J. Phys. Chem. A 1997, 101, 1488.
(9) Bardwell, M. V.; Bacak, A.; Raventos, M. T.; Percival, C. J.;
Sanchez-Reyna, G.; Shallcross, D. E. Phys. Chem. Chem. Phys. 2003, 5,
2381.

6520 J. Phys. Chem. A, Vol. 109, No. 29, 2005
Butkovskaya et al.
(
10) Atkinson, R.; Baulch, D. L.; Cox, R. A.; Crowley, J. N.; Hampson,
(29) Platz, J.; Sehested, J.; Nielsen, O. J.; Wallington, T. J. J. Phys.
Chem. A 1999, 103, 2688.
(30) Aschmann, S. M.; Chew, A. A.; Arey, J.; Atkinson, R. J. Phys.
Chem. A 1997, 101, 8042.
(31) Miyoshi, A.; Matsui, H.; Washida, N. J. Phys. Chem. 1990, 94,
3016.
(32) Scholtens, K. W.; Messner, B. M.; Cappa, C. D.; Elrod, M. J. J.
Phys. Chem. A 1999, 103, 4378.
R. F.; Hynes, R. G.; Jenkin, M. E.; Rossi, M. J.; Troe, J. Atmos. Chem.
Phys. 2004, 4, 1461.
(
11) Kukui, A.; Borissenko, D.; Laverdet, G.; LeBras, G. J. Phys. Chem.
A 2003, 107, 5732.
12) Huey, L. G.; Hanson, D. R.; Howard, C. J. J. Phys. Chem. 1995,
9, 5001.
(
9
(13) (a) Huey, L. G.; Lovejoy, E. R. Int. J. Mass Spectrom. Ion Processes
1
996, 155, 133. (b) Huey, L. G.; Dunlea, E. J.; Lovejoy, E. R.; Hanson, D.
(33) Bacak, A.; Bardwell, M. W.; Raventos, M. T.; Percival, C. J.;
Sanchez-Reyna, G.; Schallcross, D. E. J. Phys. Chem. A 2004, 108, 10681.
(34) Ranschaert, D. L.; Schneider, N. J.; Elrod, M. J. J. Phys. Chem. A
2000, 104, 5758.
R.; Norton, R. B.; Fehsenfeld, F. C.; Howard, C. J. J. Geophys. Res. 1998,
03, D3, 3355.
14) Ikezoe, Y.; Matsuoka, S.; Takebe, M.; Viggiano, A. Gas-Phase
1
(
Ion-Molecule Reaction Rate Constants Through 1986; Ion Reaction
Research Group of the Mass Spectroscopy Society of Japan, Maruzen
Company: Tokyo.
(35) Chow, J. M.; Miller, A. M.; Elrod, M. J. J. Phys. Chem. A 2003,
107, 3040.
(36) Atkinson, R.; Carter, W. P. L.; Winer, A. M. J. Phys. Chem. 1983,
87, 2012.
(37) Atkinson, R.; Aschmann, S. M.; Carter, W. P. L.; Winer, A. M.;
Pitts, J. N., Jr. J. Phys. Chem. 1982, 86, 4563.
(38) Sumathi, R.; Peyerimhoff, S. D. J. Chem. Phys. 1997, 107, 1872.
(39) Dixon, D. A.; Feller, D.; Zhan, C.-G.; Francisco, J. S. J. Phys. Chem.
A 2002, 106, 3191.
(
15) Elrod, M. J.; Ranschaert, D. L.; Schneider, N. J. Int. J. Chem. Kinet.
001, 33, 363.
16) Sander, S. P.; Friedl, R. R.; Golden, D. M.; Hampson, R. F.; Kurylo,
2
(
M. J.; Huie, R. E.; Orkin, V. L.; Moortgat, G. K.; Ravishankara, A. R.;
Kolb, C. E.; Molina, M. J.; Finlayson-Pitts, B. J. EValuation Number 14,
JPL Publication 02-25; Jet Propulsion Laboratory: Pasadena, CA, 2003.
(
17) D’Ottone, L.; Campuzano-Jost, P.; Bauer, D.; Hynes, A. J. J. Phys.
Chem. A 2001, 105, 10538.
18) Neuman, J. A.; Huey, L. J.; Ryerson, T. B.; Fahey, D. W. EnViron.
Sci. Technol. 1999, 33, 1133.
19) Donahue, N. M.; Mohrschladt, R.; Dransfield, T. J.; Anderson, J.
G.; Dubey, M. K. J. Phys. Chem. A 2001, 105, 1515.
(40) Lohr, L. L.; Barker, J. R.; Shroll, R. M. J. Phys. Chem. A 2003,
107, 7429.
(41) Barker, J. R.; Lohr, L. L.; Shroll, R. M.; Reading, S. J. Phys. Chem.
A 2003, 107, 7434.
(42) Zhao, Y.; Houk, K. N.; Olson, L. P. J. Phys. Chem. A 2004, 108,
5864.
(
(
(
20) Golden, D. M.; Smith, G. P. J. Phys. Chem. A 2000, 104, 3991.
(43) Zhang, J.; Dransfield, T.; Donahue, N. M. J. Phys. Chem. A 2004,
108, 9082.
(21) Bean, B. D.; Mollner, A. K.; Nizkorodov, S. A.; Nair, G.; Okumura,
M.; Sander, S. P.; Peterson, K. A.; Francisco, J. S. J. Phys. Chem. A 2003,
07, 6974.
22) Dransfield, T. J.; Donahue, N. M.; Anderson, J. G. J. Phys. Chem.
A 2001, 105, 1507.
23) Burkholder, J. D.; Hammer, P. D.; Howard, C. J. J. Phys. Chem.
987, 91, 2136.
24) Hippler, H.; Nasterlack, S.; Striebel, F. Phys. Chem. Chem. Phys.
002, 4, 2959.
25) Matheu, D. M.; Green, W. H., Jr., Int. J. Chem. Kinet. 2000, 32,
45.
(44) Li, Y.; Francisco, J. J. Chem. Phys. 2000, 113, 7976.
(45) Cameron, D. R.; Borrajo, A. M. P.; Bennett, B. M.; Thatcher, G.
R. J. Can. J. Chem. 1995, 73, 1627.
(46) Aloisio, S.; Francisco, J. S. J. Phys. Chem. A 1998, 102, 1899.
(47) Sander, S. P.; Peterson, M. E. J. Phys. Chem. 1984, 88, 1571.
(48) (a) Mozurkewich, M.; Benson, S. Int. J. Chem. Kinet. 1985, 17,
787. (b) Zhu, R.; Lin, M. C. Chem. Phys. Lett. 2002, 354, 217.
(49) Ehhalt, D. H. Gas-Phase Chemistry of the Troposphere, in: Global
Aspects of Atmospheric Chemistry; Zellner, R., Ed.; Springer-Verlag: Berlin,
1999.
(50) Jaegl e´ , L.; Jacob, D. J.; Brune, W. H.; Faloona, I.; Tan, D.; Heikes,
B. G.; Kondo, Y.; Sachse G. W.; Anderson, B.; Gregory, G. L.; Singh, H.
B.; Pueschel, R.; Ferry, G.; Blake, D. R.; Shetter, R. E. J. Geophys. Res.
2000, 105, D3, 3877.
1
(
(
1
2
2
8
(
(
(26) Nizkorodov, S. A.; Wennberg, P. O. J. Phys. Chem. A 2002, 106,
55.
(27) Arnold, S. T.; Viggiano, A. A. J. Phys. Chem. A. 2001, 105, 3527.
28) Finlayson-Pitts, B. J.; Wingen, L. M.; Summer, A. L.; Syomin,
(
D.; Ramazan, K. A. Phys. Chem. Chem. Phys. 2003, 5, 223.
(51) Atkinson, R. Atmos. Chem. Phys. 2003, 3, 2233.