Reactions of CF3O- core ions with ClONO2 and H2O

Article abstract of DOI:10.1039/b102967n

The reactions of CF₃O^-, CF₃OHF^- and CF₃OH₂O^- with ClONO₂ have been studied in the temperature range 225-325 K. The reactions of CF₃O^- and CF

Full text of DOI:10.1039/b102967n

View Article Online / Journal Homepage / Table of Contents for this issue
Reactions of CF O— core ions with ClONO and H O
3
2
2
C. Stepien,a V. Catoire,*a C. Amelynck,b D. Labonnette,a N. Schoon,b G. Pouleta and E. Arijsb
a L aboratoire de Physique et Chimie de lÏEnvironnement, UMR 6115 CNRS-Universite
dÏOrleans, 45071 Orleans Cedex 2, France. E-mail: V alery.Catoire=univ-orleans.fr
b Belgian Institute for Space Aeronomy, Ringlaan 3, B-1180 Brussels, Belgium
Received 2nd April 2001, Accepted 2nd July 2001
First published as an Advance Article on the web 13th August 2001
The reactions of CF O~, CF OHF~ and CF OH O~ with ClONO have been studied in the temperature
3
3
3
2
2
range 225È325 K. The reactions of CF O~ and CF OH O~ with ClONO proceed at the collision rate
3
3
2
2
whereas the reaction of CF OHF~ with ClONO is slower. No signiÐcant temperature dependence of the rate
3
2
constant has been found for each reaction. The mechanistic study has shown that the reaction of CF O~ with
3
ClONO proceeds by Ñuoride transfer and the product of the reaction between CF OH O~ and ClONO is
2
3
2
2
the cluster ion CF OClONO ~. The equilibrium reaction of CF O~ with H O has been also studied between
3
2
3
2
262 and 295 K. Values of [63.2 ^ 6.7 kJ mol~1 and [114.6 ^ 23.9 J mol~1 K~1 were derived for the
standard reaction enthalpy and entropy, respectively.
of CF O~ core ions with HCl and HNO have been pre-
Introduction
3
3
viously studied.13,14 In the present paper, the study of the
reactions of CF O~, CF OHF~ and CF OH O~ with
Chlorine nitrate (ClONO ) is an important species in control-
ling stratospheric ozone destruction by acting as a temporary
3
3
3
2
2
ClONO has been undertaken under conditions close to those
2
of the stratosphere, i.e. in the temperature range 225È325 K
reservoir for active chlorine. ClONO is formed during
2
nighttime via the reaction of nitrogen dioxide (NO ) and chlo-
2
and pressure range 0.4È16 Torr. The equilibrium reaction of
CF O~ with H O, previously studied by Amelynck et al. at
rine monoxide (ClO) and is photolysed during daytime.1 In
3
2
room temperature,13 has been also investigated in the tem-
perature range 262È295 K in order to estimate the equilibrium
constant at stratospheric temperature, which is needed for the
interpretation of the CIMS Ñight mass spectra.
the presence of polar stratospheric clouds, heterogeneous
chemistry converts the major inactive forms of chlorine
(ClONO , HCl) into active ones leading to the polar ozone
2
loss.2,3
Chemical ionization mass spectrometry (CIMS) is currently
used as an in situ method to measure concentrations of atmo-
spheric trace gases.4v9 In the CIMS method, speciÐc reactant
ions are injected into a Ñow tube coupled to a mass spectro-
meter where they react with trace constituents to form speciÐc
product ions.10 In a joint project of the ““Belgian Institute for
Space AeronomyÏÏ (BISA), the ““Laboratoire de Physique et
Chimie de lÏEnvironnementÏÏ (LPCE), and the ““Physikalisches
Institut of the University of BernÏÏ (PIUB), a new balloon-
borne CIMS apparatus has been developed. Its goal was the
Experimental
Experiments were performed in two Ñowing-afterglow appar-
atuses located at LPCE and at BISA. Both instruments have
been described in detail elsewhere13h15 and are only brieÑy
presented here.
At LPCE, the ionÈmolecule reactions were studied in a 3.5
cm id stainless steel Ñow tube. The Ñow of the carrier gas (N ,
2
purity [99.9999%) was in the range 1.5È30 STP L min~1
simultaneous measurement of stratospheric HCl, HNO ,
(STP \ 273 K, 760 Torr), resulting in Ñow tube pressures
between 0.8 and 16 Torr. The ions were sampled through
three di†erentially pumped chambers and introduced into the
quadrupole mass analyzer, where they were selected according
to mass-to-charge ratio and subsequently detected by a sec-
ondary electron multiplier (SEM).
3
ClONO and N O concentrations by using several ion
2
2 5
sources. Recently, the instrument has been equipped with an
ion source generating CF O~ since this ion appeared to be a
3
good candidate for the detection of stratospheric HCl, HNO
3
and ClONO . In the laboratory, the CF O~ ion has been
shown to react with HCl, HNO and ClONO , mainly by
Ñuoride transfer at room temperature and low pressure,
2
3
The reaction of CF O~ with ClONO was studied at 225,
3
2
3
2
274, 295 and 325 K; the reactions of CF OH O~ and
3
2
leading to the speciÐc ions FHCl~, FHNO ~ and
CF OHF~ with ClONO were studied at 225 and 295 K, and
3
3
2
FClONO ~, respectively.11
the reaction of CF O~ with H O was studied at 262, 276, 285
2
3
2
In addition to CF O~ ions, unexpectedly abundant
and 295 K. To obtain a temperature of 325 K, the outer walls
of the tube were resistively heated. The studies at 262, 274 and
276 K were performed by cooling the tube with cold accumu-
3
amounts of CF OHF~ and CF OH O~ source ions were
3
3
2
observed in the Ñow reactor during two recent balloon Ñights
(June 1997 and June 1999).12 The presence of other ions,
CF OHCl~, CF OHNO ~ and CF OClONO ~, in the Ñight
lator bags, and at 225 K by wrapping the tube with CO
pellets in the pressure range 1.0È1.7 Torr. All gases were
Ñowed through a precooling zone of 60 cm, so that the tem-
perature was homogeneous in the reaction zone (^5 K). Tem-
perature was measured by a Pt resistance probe located at the
center of the Ñow tube. The CF O~ ions were produced in a
2
3
3
3
3
2
spectra, led to the assumptions that CF OHF~ and
3
CF OH O~ react with HCl, HNO and ClONO , respec-
3
2
3
2
tively, by ligand switching. These assumptions required
experimental evidence from laboratory studies. The reactions
3
DOI: 10.1039/b102967n
Phys. Chem. Chem. Phys., 2001, 3, 3683È3689
This journal is ( The Owner Societies 2001
3683

View Article Online
high-pressure ion source by dissociative attachment of elec-
stored in a bubbler at 213 K in the dark and was cooled to
194 K for the experiments.
trons to CF OOCF , as described by Amelynck et al.13,14
3
3
Electrons were produced by a 500 V discharge in an argon
Ñow (purity [99.9999%, 150 STP cm3 min~1). A small Ñow
(2 STP cm3 min~1) of a CF OOCF mixture (2500 ppm in
At BISA, the reactions of CF O~ and CF OH O~ with
3
3
2
ClONO were studied in a 6.5 cm id Ñow tube at room tem-
2
perature and at pressures ranging from 0.4 to 1 Torr. A Ñow
3
3
Ar) was added downstream of the discharge zone. The CF O~
of He (purity [99.9999%) of 4.5 STP L min~1 was used as a
3
ions were introduced into the Ñow tube through a capillary
carrier gas and pumped by a Roots blower. ClONO was
2
tube (6 mm id) located 108.6 cm upstream of the ion analysis
introduced through two electrically insulated Ðxed ring inlets
zone. Above 273 K, only CF O~ and CF OHF~ ions were
located at 43.4 and 77.8 cm from the mass spectrometer inlet.
3
3
observed in the mass spectra, with CF O~ as the major ion.
CF O~ core ions were produced in the same way as at LPCE.
3
The presence of CF OHF~ was attributed to the reaction of
3
3
To study the reaction of CF OH O~ with ClONO at room
3
2
2
CF O~ with trace impurities in the CF OOCF ] Ar
temperature, the CF OH O~ ions were generated by the
3
3
3
3
2
3
mixture. The ratio of CF OHF~ to CF O~ ion count rates in
three-body reaction of CF O~ with residual water in a cooled
3
3
the mass spectra increased with the Ñow tube pressure and
ranged from 0.3 to 10% between 0.8 and 16 Torr. At 225 K,
CF OH O~ and CF O(H O) ~ were formed from the reac-
(195 K) high-pressure section of the Ñow tube upstream of the
reaction zone.13,14 A large fraction of the CF OH O~ ions
3
2
exiting the cooled high pressure zone dissociated into CF O~
3
2
3
2
2
3
and H O in the section of the Ñow tube at room temperature.
2
tion between CF O~ and H O, which is much faster at 225 K
3
2
than at room temperature (see below). The presence of H O
The formation of CF OH O~ at low pressure in this section
2
3
2
was attributed to the presence of impurities in the bu†er gas
was negligible, as veriÐed by using a second CF O~ ion
3
(H O \ 0.5 ppm), to desorption of water from the Ñow tube
source located upstream of the reaction zone. The
2
walls or/and to possible leaks at low temperature (due to con-
[CF OH O~]/[CF O~] ratio at the end of the Ñow tube was
3
2
3
traction of the TeÑon ring used as a seal between the ion
source and the reactor). This enabled a study of the reaction
of CF OH O~ with ClONO at low temperature.
found to be around 5%.
Ions were sampled through an oriÐce of 0.2 mm diameter in
a Ðrst chamber pumped by a 1600 L s~1 turbo-molecular
pump. They were guided by an octopole from the gas expan-
sion zone, behind the inlet oriÐce, to the mass Ðlter located in
a second, separately pumped, chamber. To minimize collision-
induced dissociation of ions in the gas expansion zone, the
di†erence in bias voltage between the inlet plate and the octo-
pole rods was kept at a value of only 0.5 V. After analysis by a
quadrupole mass Ðlter, ions were detected by a SEM.
The determination of the absolute rate constants for the
3
2
2
The neutral reactant ClONO was introduced by a Ðnger
2
inlet located 46 cm upstream of the ion source in order to get
a homogeneous mixture of ClONO and the carrier gas before
2
introduction of the reactant ion. Thus, the distance between
the ion source and the analysis zone (108.6 cm) was taken as
the reaction distance. Electron attachment to ClONO has
2
been previously studied by Van Doren et al.,16 who found
that this process led to the production of NO ~ (D50%),
2
NO ~ (D30%), and ClO~ (D20%) with a rate constant of
3
reactions involving ClONO required the measurement of the
2
(1.1 ^ 0.5) ] 10~7 cm3 molecule~1 s~1. During the study of
ClONO concentration in the Ñow tube. In both apparatuses,
2
the reaction of CF O~ with ClONO , the NO ~ and ClO~
this concentration was measured by UV absorption spectrom-
3
2
2
ions were never observed, showing that the presence of elec-
trons in the Ñow tube was negligible. The ion residence time in
the Ñow tube was typically 18 ms. The latter was measured by
pulsing a biased grid and by synchronously recording the
arrival time of the ion swarm on the SEM detector.17 The
etry in a cell connected to the Ñow tube. The UV absorption
system consisted of a 50 cm long ] 0.7 cm id absorption cell
Ðtted with Suprasil windows and connected to the Ñow tube
by a TeÑon tube. A diluted mixture of ClONO and argon
2
(purity [99.9999%) was Ñowed at 2È15 STP cm3 min~1 with
ClONO concentration range in the Ñow tube was typically
(0.2È2.0) ] 1011 molecule cm~3.
ClONO was synthesized at LPCE, using the method of
an additional Ar Ñow to obtain a constant total Ñow of 100
STP cm3 min~1 through the absorption cell before being
introduced into the main reactor. In order to check the purity
of the chlorine nitrate introduced in the Ñowing afterglow
apparatus, the reaction of SF ~ with ClONO was used.
2
2
Schmeisser18 based on the reaction of Cl O with N O . Cl O
2
2
5
2
and N O were synthesized by the methods of Cady19 and
2
5
6
2
Davidson,20 respectively. During each step of the synthesis,
special care was taken to minimize water traces and hydro-
lysis reactions that would produce HOCl and HNO . Cl O
SF ~ is known to react with ClONO and with all possible
6
2
impurities in ClONO with a comparable rate constant.23,24
2
SF ~ was produced in the same ion source as CF O~, by
3
2
6
3
was obtained by condensing Cl on freshly crystallized and
dried mercuric oxide (HgO) in excess at 77 K. The trap was
electron attachment to SF (purity [ 99.9%) Ñowing at 2È5
2
6
2
STP cm3 min~1. ClONO ~ and mainly FClONO ~ were
2
slowly warmed up to 233 K. After 3 days at this temperature,
Cl O was separated from HgO by distillation. N O was syn-
observed as products of the reaction of SF ~ with ClONO ,
6
2
in agreement with previous studies.23,24 However, FHNO ~
2
2
5
3
thesized by adding a Ñow of NO (purity [99%) to a Ñow of
produced by the reaction of SF ~ with HNO was also
6
3
O
(diluted in O ) at atmospheric pressure. NO and NO ,
observed in the mass spectra. This HNO was assumed to
3
2
2
3
3
resulting from the oxidation of NO by O , reacted to form
result from the ClONO hydrolysis by the trace water
adsorbed on the walls of the tube between the ClONO
bubbler and the reactor. This was conÐrmed by doing experi-
3
2
N O which was collected in a trap cooled at 195 K. ClONO
2
5
2
2
was synthesized by adding Cl O to an excess of N O in a 77
2
2 5
K trap. The trap was slowly warmed up to 233 K and the
ments in which no FHNO ~ was observed when ClONO
3
2
reaction was complete after 3 days. ClONO was puriÐed by
was introduced directly into the Ñow tube without passing
through the UV-absorption cell. The residence time of
2
trap-to-trap distillations to remove traces of N O , Cl O and
2
5
2
Cl . ClONO purity was checked by measuring its vapor
pressure P in the temperature range 186 to 216 K. The plot
ClONO in the UV cell was reduced by minimizing its
2
2
2
volume and by increasing the total Ñow in the cell. By heating
vap
of ln P \ f (1/T ) yielded the values of the vaporization enth-
and passivating the feedlines, the HNO concentration could
vap
vap
3
alpy * H¡ \ 29.4 kJ mol~1, and of the vaporization entropy,
be signiÐcantly decreased, but not totally reduced in the
*
S¡ \ 155.4 J mol~1 K~1 in the above temperature range.
system. Using the rate constants of the reaction of SF ~ with
vap
6
These values are in very good agreement with literature
ClONO and HNO 23 and the ratio of the products of these
2
3
values: * H¡ \ 29.3 kJ mol~1 and * S¡ \ 153.7 J mol~1
reactions, the [HNO ]/[ClONO ] ratio was evaluated to be 2
vap
vap
3
2
K~1 in the temperature range 185È238 K21 and * H¡ \
to 10% in the Ñow tube.
vap
28.8 kJ mol~1 and * S¡ \ 152.7 J mol~1 K~1 in the tem-
perature range 193È299 K.22 This pure ClONO was then
The ClONO absorption was measured at 216 nm, where
vap
2
the absorption cross section p is maximum (p \ 3.45 ] 10~18
2
3684
Phys. Chem. Chem. Phys., 2001, 3, 3683È3689

View Article Online
cm2 molecule~1).25 The concentration of ClONO in the Ñow
the low ratio [HNO ]/[ClONO ] in the present experiments,
2
3
2
tube, was determined using the following equation:
the observation of NO ~ could not be assigned to the pres-
3
ence of HNO in the Ñow tube. A possible explanation is that
3
A
I0
I
lpQ
B
NO ~ is formed by decomposition of FClONO ~ in the sam-
ln
Q
P
T
3
2
AC FT AC
pling zone. Changing the electrostatic conditions of the ion
guiding elements indeed led to changes in the ratio
[NO ~]/[FClONO ~], with a minimum ratio of 5% in the
[ClONO ] \
(I)
2
P
T
FT AC FT
3
2
where l is the length of the absorption cell (cm), Q is the total
standard Ñow, P the pressure, T the temperature (AC and FT
denoting the absorption cell and the Ñow tube, respectively),
and I and I¡ the measured light intensities with and without
LPCE spectra. In the BISA spectra, this ratio was only 3%.
The FHNO ~ ions were the main product ions of the reac-
3
tion of CF O~ with the HNO impurity:11,14
3
3
ClONO present in the absorption cell, respectively.
CF O~ ] HNO ] FHNO ~ ] CF O
(2)
2
3
3
3
2
The rate constants of the reactions of the ions with
As reaction (1) progresses further, secondary product ions,
ClONO were inferred from the slopes of the logarithmic
2
NO ClONO ~, NO HNO ~ and NO (HNO ) ~, were also
decays of the reactant ion count rates as a function of the
3
2
3
3
3
3 2
observed, as shown in Fig. 1. The mechanism of formation of
these ions is discussed hereafter.
product of the ClONO concentration and the reaction time,
2
given that ClONO was always in large excess with respect to
2
First, NO ClONO ~, which was the most abundant of the
the ions used. Overall uncertainties for the rate constants were
3
2
secondary products, was formed in the reaction of
estimated to be 30%, including two standard deviations, the
uncertainty in the reaction time measurement and the uncer-
FClONO ~ with ClONO . This suggests that FClONO ~
2
2
2
has the structure of a FClÉONO ~ cluster, reacting with
tainty in the determination of ClONO concentrations.
2
2
ClONO by ligand switching:
The reaction of CF O~ with H O was studied at LPCE.
2
3
2
Water vapor was introduced into the Ñow tube by two Ðnger
inlets located 56 and 76 cm upstream of the sampling cone.
FClÉONO ~ ] ClONO ] NO ClONO ~ ] FCl (3)
2
2
3
2
H O concentrations in the Ñow tube were in the range
This is in agreement with the results reported by Huey et al.11
The reaction
2
1 ] 1012È1 ] 1014 molecule cm~3. Three methods were used
to introduce controlled Ñows of H O into the reactor. In the
2
Ðrst method, a mass Ñowmeter controlled the Ñow of mixtures
FHNO ~ ] ClONO ] NO ClONO ~ ] HF
(4)
3
2
3
2
of H O vapor and N (purity [99.9999%) prepared in a bulb.
2
2
also occurred as a minor channel for the production of
The second method consisted in determining the rate of pres-
NO ClONO ~, because FHNO ~ has the structure of a
sure drop in the bulb containing known mixtures of H O
3
2
3
2
FHÉONO ~ cluster, allowing switching between HF and
vapor and nitrogen that were introduced into the Ñow tube. In
2
ClONO . This structure was deduced from previous
studies11,14 which reported that NO HNO ~ was produced
in reaction (5):
the third method, a Ñow of nitrogen, controlled by a Ñow-
meter, was sent through a bubbler containing distilled water.
2
3
3
This bubbler was designed to avoid H O droplets being
2
carried by N with H O vapor, which would lead to an
2
2
FHÉONO ~ ] HNO ] NO HNO ~ ] HF
(5)
underestimation of the H O concentration in the Ñow tube. In
this method, the Ñow was assumed to be saturated with water
2
3
3
3
2
vapor. The three methods agreed within experimental error.
Results and discussion
1. Reaction of CF O— with ClONO
3
2
1.1. Mechanism. At room temperature, the reaction of
CF O~ with ClONO was studied at pressures ranging from
3
2
0.4 to 16 Torr. The major product ion observed in the mass
spectra was FClONO ~, obtained by Ñuoride transfer, as pre-
2
viously reported by Huey et al:11
CF O~ ] ClONO ] FClONO ~ ] CF O
(1)
3
2
2
2
NO ~ (D5% of the product ions), FHNO ~ (4È15%) and
3
3
traces of CF OClONO ~ (\1%) were also observed.
3
2
CF OClONO ~ could not be unambiguously assigned as
3
2
formed in a minor channel of reaction (1) since CF OHF~,
3
which was also present in the Ñow tube, may lead to the for-
mation of CF OClONO ~ by ligand switching.
3
2
For NO ~, Huey et al.11 found that this ion was not pro-
3
duced by the reaction of CF O~ with ClONO , but was the
3
2
main product of the reaction of CF O~ with N O . In the
3
2 5
present experiments, the temperature of the ClONO -
2
containing bubbler was too low to allow for the presence of
N O in the Ñow tube. Moreover, the ClONO concentration
2
5
2
was determined at three wavelengths (200, 216, 230 nm), for
which the ClONO to N O cross-section ratios vary from
2
2 5
0.31 to 2.10. At all wavelengths, the same value of ClONO
concentration was inferred, indicating that N O could only
be present at insigniÐcant levels. Another potential source of
2
Fig. 1 Example of the ion count rates vs. [ClONO ] in the study of
2
the reaction CF O~ ] ClONO ()) [CF O~]; (K) [CF OHF~];
2
5
3
2
3
3
(>) [FClONO ~]; (=) [FHNO ~]; (+) [NO ~]; (…)
2
3
3
NO ~ is the reaction of CF O~ with HNO which leads to
[NO ClONO ~]; (Â) [NO HNO ~]; (L) [CF OClONO ~]; (|)
3
3
3
3
3
2
3
3
3
2
NO ~ as a minor product ion (5È8%).11,14 However, due to
[NO (HNO ) ~]).
3 2
3
Phys. Chem. Chem. Phys., 2001, 3, 3683È3689
3685

View Article Online
Reaction (5) also occurred in the present study as a minor
with ClONO ,27 do not occur in the present system. In addi-
2
channel for the NO HNO ~ production, in addition to reac-
tion, using the total energies of FClÉONO ~ ([2 206 271.168
3
3
2
kJ mol~1) and FHÉONO ~ ([999 785.601 kJ mol~1) with
2
tion (6) and reaction (7):26
zero-point energy corrections of 46.1 and 69.9 kJ mol~1,
FClÉONO ~ ] HNO ] NO HNO ~ ] FCl
(6)
respectively, and considering the results of Mebel and Moro-
kuma,27 exothermicities of 17.3, 40.0, 51.4, 28.7 and 11.4 kJ
mol~1 have been found for reactions (3), (4), (5), (6) and (7),
respectively. These results suggest that all these reactions may
be favored in the present study.
2
3
3
3
NO ClONO ~ ] HNO ] NO HNO ~ ] ClONO (7)
3
2
3
3
3
2
Finally, the least abundant of the secondary product ions,
NO (HNO ) ~, was produced by the reaction of
3
3
3 2
NO HNO ~ with HNO :
3
3
1.2. Kinetics. Since the rate constant for the reaction of
CF O~ with HNO is twice as large as that with ClONO
NO HNO ~ ] HNO ] M ] NO (HNO ) ~ ] M (8)
3
3
3
3
3 2
3
3
2
In ab initio calculations, Mebel and Morokuma27 have
determined the structures and energies of the species
ClONO , HNO , NO ~, NO ClONO ~ and NO HNO ~
presently determined,14 the presence of HNO had to be
3
taken into account to correct the rate constant of CF O~
3
with ClONO . The value of the rate constant k was derived
2
3
3
3
2
3
3
2
1
using the density functional B3LYP/6-31G(d) approach. Using
the same method with the GAUSSIAN 98 program,28 the
energies and structures of FClONO ~ and FHNO ~ have
by multiplying the apparent rate constant by the ratio of the
count rates of primary product ions resulting from reaction
(1), i.e. FClONO ~ and NO ~, to the count rates of all the
2
3
2
3
been determined and are shown in Fig. 2. In agreement with
the experimental results, the structure of the former ion was
product ions, representing a yield of 85 to 96%, obtained at
low ClONO concentration. The dependence of the room-
2
found to be that of the FClÉONO ~ cluster, with atomic
temperature rate constant of the CF O~ ] ClONO reaction
2
charges of [0.28e on FCl and of [0.72e on NO , with a
3
3
2
on pressure is presented in Fig. 3. There is a slight di†erence
(less than 10%) between the values determined at LPCE and
at BISA, but this is not signiÐcant with respect to the reported
statistical uncertainties (two standard deviations). The average
geometry of NO ~ similar to that of free NO ~ (1.264 A),27
3
3
with a FCl bond length close to that of the free FCl (1.664 A)
and a ClO bond length much longer than in free ClONO
2
(1.708 A).27 Similarly, FHNO ~ was found to be
a
of 35 series of experiments led to k \ (1.2 ^ 0.4) ] 10~9 cm3
3
1
FHÉONO ~ cluster (Fig. 2). These theoretical results conÐrm
molecule~1 s~1 at 295 K, where the quoted uncertainties
2
the mechanisms presented above. The existence of the cluster
include estimated systematic errors. This value is in very good
agreement with that of Huey et al.,11 who obtained
(1.1 ^ 0.3) ] 10~9 cm3 molecule~1 s~1, and with the collision
ions F~ÉClONO and F~ÉHONO can thus be excluded and
2
2
other mechanisms involving these clusters, which lead to the
same Ðnal products as those observed in the present study,
can be rejected. For instance, ““attachmentÈdetachmentÏÏ
rate constant k calculated from the parametrized theory of Su
c
and Chesnavich,29,30 1.1 ] 10~9 cm3 molecule~1 s~1, using a
mechanisms, similar to that of the reaction of Cl~ÉHONO
value of 0.77 D31 for the dipole moment k and a value of
2
D
6.28 A
Ž
3 for the polarizability a of ClONO .27 This theory is
2
based on a combined variational transition state theory/
classical trajectory study of thermal energy ionÈpolar mol-
ecule capture collisions. Su and Chesnavich have shown that
the ratio of k to the Langevin rate constant depends on k , a
c
D
and the temperature.29,30
Kinetic results have been obtained at three other tem-
peratures, with average values of k of (1.3 ^ 0.4) ] 10~9 cm3
1
molecule~1 s~1 from 2 series of experiments at 225 K,
(1.2 ^ 0.4) 10~9 cm3 molecule~1 s~1 from 2 series of experi-
ments at 274 K and (1.1 ^ 0.4) ] 10~9 cm3 molecule~1 s~1
from 4 series of experiments at 325 K. Given the reported
statistical uncertainties (10%), no signiÐcant temperature
Fig. 3 Rate constant of the reaction CF O~ ] ClONO as a func-
3
2
Fig. 2 Geometries (in
A
and degrees) of FHÉONO ~ (a) and
tion of pressure. Circles and triangles correspond to values obtained
at LPCE and at BISA, respectively. Error bars represent two standard
deviations.
2
FClÉONO ~ (b) optimized at the B3LYP/6-31G(d) level. Bold
2
numbers are the natural atomic charges (in e).
3686
Phys. Chem. Chem. Phys., 2001, 3, 3683È3689

View Article Online
dependence of the rate constant k can be derived, even if a
slight increase seems to be observed as the temperature
was CF OClONO ~, formed by ligand switching. At further
1
3 2
extent of the reaction, the secondary reaction of
decreases.
CF OClONO ~ with ClONO
occurred, leading to
3
2
2
From the measurement of the rate constant k , and using
NO ClONO ~:
1
the rate constant value k (2.3 ] 10~9 cm3 molecule~1 s~1),14
2
3
2
CF OClONO ~ ] ClONO ] NO ClONO ~ ] products
the ratio [HNO ]/[ClONO ] could be inferred for the low
3
2
2
3
2
3
2
ClONO concentrations:
(12)
2
[HNO ]
k P
1 2
2 1
This suggests that CF OClONO ~ has the structure of a
3
\
(II)
3
2
CF OClÉONO ~ cluster. CF OH O~ also reacted with
[ClONO ] k P
3
2
3
2
2
HNO with a rate constant k \ 2.1 ] 10~9 cm3 molecule~1
3
13
where P and P are the ion count rates of the products of
s~1:14
1
2
reaction (1) and (2), respectively. The ratio varied between 2%
and 9%, in excellent agreement with that obtained using the
reactions of SF ~ with ClONO and HNO , as reported in
CF OH O~ ] HNO ] products
(13)
3
2
3
The measured constant rate k
formula
was corrected by using the
6
2
3
meas
the Experimental section. Additionally, the rate constant of
the reaction of FClONO ~ ] ClONO has been estimated by
2
2
analyzing the linear part of the FClONO ~ decay shown in
k
\ k
[ k [HNO ]/[ClONO ]
(IV)
2
11
meas
13
3
2
Fig. 1. A lower limit of (8.2 ^ 0.4) ] 10~10 cm3 molecule~1
The average value of 5 series of experiments yielded a rate
s~1 at 1 Torr of He and at room temperature has been found,
constant k \ (1.2 ^ 0.4) ] 10~9 cm3 molecule~1 s~1 at
using the rate constant of the reaction of FClONO ~ with
11
2
3
room temperature. This value is close to the calculated colli-
HNO to take into account the presence of HNO in the
3
sion limit of 1.0 ] 10~9 cm3 molecule~1 s~1.29,30 At 225 K, a
rate constant of (1.3 ^ 0.4) ] 10~9 cm3 molecule~1 s~1 was
found. All the rate constants determined in this work are sum-
marized in Table 1.
reactor. The latter rate constant was derived to be 1.8 ] 10~9
cm3 molecule~1 s~1 as derived from the theory of Su and
Chesnavich29,30 using values of 2.17
D for the dipole
moment32 and 4.5 A
Ž
3 for the polarizability of HNO .11
3
4. Reaction of CF O— with H O
2. Reaction of CF OHF— with ClONO
3
2
3
2
The reaction between CF O~ and H O has been studied
The reaction of CF OHF~ with ClONO
3
2
3
2
between 262 and 295 K. As described previously, known
amounts of water vapor were introduced into the Ñow tube at
CF OHF~ ] ClONO ] products
(9)
3
2
56 and 76 cm from the sampling zone. The decay of CF O~
ions was found to be independent of the length of the reaction
zone (i.e. the reaction time), indicating that the equilibrium
was reached under the experimental conditions used:
3
led principally to the formation of FClONO ~. Traces of
2
CF OClONO ~ were also observed, but could not be unam-
3
2
biguously assigned as product of this reaction because of the
simultaneous presence of CF O~ in the Ñow tube. The
3
FClONO ~ formation was deduced from the unchanged ratio
CF O~ ] H O % CF OH O~
(14)
(V)
2
3
2
3
2
[FClONO ~]/[CF OClONO ~] as [CF OHF~]/[CF O~]
increased from 0.3 to 10%. In order to obtain the rate con-
2
3
2
3
3
The equilibrium constant is
stant k , the rate constant for reaction (10) had to be con-
[CF OH O~]
9
3
2
K \
sidered:
[CF O~][H O]
3
2
CF OHF~ ] HNO ] products
(10)
Since CF OH O~ was the only product ion observed, its con-
3
3
3
2
centration could be taken as: [CF OH O~] \ [CF O~]
A simple correction was obtained from the expression:
3
2
3
0
[ [CF O~] where [CF O~] is the concentration of
3
3
0
k \ k
[ k [HNO ]/[ClONO ]
(III)
CF O~ in the absence of H O in the Ñow tube. Therefore,
9
meas
10
3
2
3
2
K was expressed as a function of CF O~ concentrations
where k
is the measured rate constant and k \ 2.1
3
meas
10
and its value could be deduced from the slope of
] 10~9 cm3 molecule~1 s~1.14 At room temperature, k was
9
([CF O~] /[CF O~] [ 1) vs. [H O]. The determination of
found to be (0.7 ^ 0.2) ] 10~9 cm3 molecule~1 s~1, obtained
3
0
3
2
the equilibrium constant was thus not sensitive to mass dis-
crimination. The overall uncertainty in the value of K was
estimated to be 50%, essentially coming from the experimental
scatter (two standard deviations) and from the uncertainty in
from 15 linear plots in the pressure range 1.6È4 Torr N . This
2
value is lower than the collision rate constant, 1.0 ] 10~9 cm3
molecule~1 s~1, determined from the theory of Su and Ches-
navich.29,30 At 225 K, a rate constant k \ (0.8 ^ 0.2) 10~9
9
the measurements of H O concentrations. At 295 K, the equi-
cm3 molecule~1 s~1 was obtained from 2 series of experi-
2
librium constant was measured at pressures ranging from 1.7
ments. No study could be performed at other temperatures,
to 2.7 Torr. Values of 6.3 ] 10~15, 7.2 ] 10~15 and
6.5 ] 10~15 cm3 molecule~1 were derived from the Ðrst,
since CF OHF~ was present at too low concentration. Thus,
3
no Ðrm conclusion could be drawn for the temperature depen-
dence of the rate constant.
Table 1 Summary of the rate constants obtained in this work
3. Reaction of CF OH O— with ClONO
2
3
2
Reaction
Temperature/K
k/10~9 cm3 molecule~1 s~1
The reaction of CF OH O~ with ClONO was studied at 295
3
2
2
K in 0.4 Torr of He at BISA, and at 225 K in 1.3È1.9 Torr
at LPCE. At room temperature, the ratio
[CF OH O~]/[CF O~] was around 0.05. At 225 K, this
CF O~ ] ClONO
225 ^ 5
274 ^ 5
295 ^ 1
325 ^ 5
225 ^ 5
295 ^ 1
225 ^ 5
295 ^ 1
1.3a
1.2
1.2
1.1
0.8
0.7
1.3
1.2
3
2
N
2
3
2
3
reaction was studied for di†erent [CF OH O~]/[CF O~]
CF OHF~ ] ClONO
3 2 3
3
2
values:
0.20, 0.31 and 6.10. For the highest
CF OH O~ ] ClONO
[CF OH O~]/[CF O~] ratio, CF O(H O) ~ ions were also
3
2
2
3
2
3
3
2
2
observed. The major ion produced by reaction (11),
a Two standard deviations lead to 10% errors.
CF OH O~ ] ClONO ] CF OClONO ~ ] H O (11)
3
2
2
3
2
2
Phys. Chem. Chem. Phys., 2001, 3, 3683È3689
3687

View Article Online
second and third methods of introduction of H O (as
described in the Experimental section), respectively. The
5. Implications for the interpretation of the CIMS
measurements
2
average value was derived from the linear regression of 120
data, shown in Fig. 4, leading to K \ (6.5 ^ 3.3) ] 10~15 cm3
molecule~1. This value is somewhat higher than the value of
(3.8 ^ 1.9) ] 10~15 cm3 molecule~1 determined in a recent
study13 in the pressure range 0.35È1.2 Torr of He. However,
the values overlap within the combined error bars. The stan-
dard Gibbs energy of the reaction CF O~ ] H O at 295 K
The data obtained for the equilibrium reaction forming
CF OH O~ can be applied to the balloon-borne CIMS Ñight
3
2
conditions. At a temperature comparable with that of the
highest measurement altitude of the balloon-borne CIMS
instrument (31 km, 234 K, 7.5 Torr), a value of the equilibrium
constant K \ 4.3 ] 10~12 cm3 molecule~1 can be extrapo-
lated from the above thermochemical data. Considering a
water vapor mixing ratio of 5 ppm at 31 km,33 the ratio
[CF OH O~]/[CF O~] is expected to be around 7, which is
3
2
was derived as *G¡ \ [29.4 ^ 1.7 kJ mol~1. The equilibrium
constant K was also determined at 262, 276 and 285 K at 1.7
Torr, by using the Ðrst method for H O determination, with
3
2
3
2
in disagreement with the ratio of 1 observed in the Ñight data
at this altitude. Several arguments may be invoked to explain
this discrepancy. One is that the temperature in the Ñow tube
might be higher than that outside the balloon gondola (which
was actually measured) by a few degrees, and would result in a
lower [CF OH O~]/[CF O~] ratio. Another explanation is
values of (1.4 ^ 0.7) ] 10~13, (4.1 ^ 2.1) ] 10~14 and
(1.4 ^ 0.7) ] 10~14 cm3 molecule~1, respectively. Thermoche-
mical data could be deduced from the VanÏt Ho† plot shown
in Fig. 5. The standard reaction enthalpy *H¡ obtained from
this plot equals [63.2 ^ 6.7 kJ mol~1 and the standard reac-
tion entropy *S¡ equals [114.6 ^ 23.9 J mol~1 K~1, neglect-
ing their temperature dependences in the range 262È295 K
with respect to the experimental uncertainties. No value of the
equilibrium constant could be obtained at lower temperature
because CF OH O~ was already present in the Ñow tube
3
2
3
the possible decomposition of CF OH O~ and/or CF OHF~
3
2
3
in the sampling zone of the Ñight instrument, which would
lead to an overestimation of the above calculated ratio.
Finally, the existence of unexpected or unknown chemistry of
the ions produced in the source may also have an impact on
the [CF OH O~]/[CF O~] ratio.
3
2
without introduction of water (see Experimental section) and
water could be condensed on the introduction inlet of the Ñow
tube.
3
2
3
The rate constants and the product distributions of all reac-
tions determined at 225 K in the present study can also be
used to analyze the Ñight data. A preliminary analysis already
shows that the product ions resulting from the reactions of the
primary ions (CF O~, CF OH O~, CF OHF~) with
3
3
2
3
ClONO are not observed at the expected levels in the
Ñight spectra. For example, using the measured ratio
2
[CF OH O~]/[CF O~], the signal of FClONO ~, produced
3
2
3
2
in the reactions of CF O~ and CF OHF~ with ClONO , is
3
3
2
too low in contrast with that of CF OClONO ~, produced in
3
2
the reaction of CF OH O~ with ClONO . An over-
3
2
2
estimation of [CF O~] in the Ñight reactor, which could be
3
due to the dissociation of CF OH O~ and/or CF OHF~ in
3
2
3
the sampling zone, might explain the low level of signal of
FClONO ~ in the Ñight spectra. However, this over-
2
estimation of [CF O~] cannot be quantiÐed at present, partly
3
due to the lack of kinetic data for the reaction of CF O~ with
3
HF.
Conclusion
The reaction of CF O~ with ClONO has been studied over a
3
2
large range of pressure and temperature in two Ñowing after-
glow apparatuses. No pressure and temperature dependence
Fig. 4 Plot of [CF O~] /[CF O~] [ 1 as a function of [H O]
3
0
3
2
giving the equilibrium constant for the reaction CF O~ ] H O %
3
2
of the nature of the ions produced (mainly FClONO ~) has
been noticed. No signiÐcant temperature dependence has been
CF OH O~ at 295 K (see text).
2
3
2
inferred for the rate constant. The CF OH O~ ion has been
3
2
observed to react with ClONO with a rate constant close to
2
the collision limit to form the cluster CF OClONO ~,
3
2
whereas the reaction between CF OHF~ and ClONO pro-
3
2
ceeds at a lower rate with FClONO ~ being the major
2
product. The rate constants and the product distributions of
all these reactions can be used to analyze the Ñight data
obtained with the balloon-borne CIMS instrument using a
CF O~ ion source. A preliminary analysis shows that the
3
product ions (such as FClONO ~ and CF OClONO ~) are
2
3
2
not present at the expected concentrations given the rate con-
stants and mechanisms determined in this work. The present
laboratory study indicates that this cannot be explained by a
change in mechanism between room temperature and low
temperature for the reaction CF O~ ] ClONO . The study
3
2
of the equilibrium between CF O~ and H O also shows a
discrepancy between the expected and observed ratios
3
2
[CF OH O~]/[CF O~] at the highest altitude of the balloon
3
2 3
A laboratory kinetic study of the formation of
Ñight.
Fig. 5 VanÏt Ho† plot of the equilibrium constant K for the reaction
CF O~ ] H O % CF OH O~.
CF OHF~, which is also an abundant ion in the Ñight
3
2
3
2
3
3688
Phys. Chem. Chem. Phys., 2001, 3, 3683È3689

View Article Online
12 N. Schoon, C. Stepien, C. Amelynck, E. Arijs, E. Neefs, D. Neve-
jans, V. Catoire, D. Labonnette, G. Poulet, H.-P. Fink, E. Kopp
and W. Luithardt, Proceedings of the Fifth European symposium,
Air Pollution Research Report 73, Stratospheric Ozone, 1999,
p. 348.
spectra, is clearly required in order to improve the interpreta-
tion of the in situ data. In addition, further Ñight data analysis
together with laboratory studies using the Ñight instrument
are needed in order to derive an accurate stratospheric
ClONO concentration proÐle.
13 C. Amelynck, A.-M. Van Bavel, N. Schoon and E. Arijs, Int. J.
2
Mass Spectrom., 2000, 202, 207.
14 C. Amelynck, N. Schoon and E. Arijs, Int. J. Mass Spectrom.,
2000, 203, 165.
Acknowledgements
15 V. Catoire, C. Stepien, D. Labonnette, J.-C. Rayez, M.-T. Rayez
and G. Poulet, Phys. Chem. Chem. Phys., 2001, 3, 193.
16 J. M. Van Doren, J. McClellan, T. M. Miller, J. F. Paulson and
A. A. Viggiano, J. Phys. Chem., 1996, 105, 104.
A. Kukui (LCSR-CNRS, Orleans) is gratefully acknowledged
for the ab initio calculations. We also thank the Commission
of the European Communities (Environment and Climate
Program, Contract N¡ ENV4-CT95-0042), the ““Departement
des Sciences de lÏUnivers (CNRS)ÏÏ and the ““Centre National
dÏEtudes SpatialesÏÏ for Ðnancial support.
17 C. Amelynck, E. Arijs, N. Schoon and A.-M. Van Bavel, Int. J.
Mass Spectrom. Ion Processes, 1998, 181, 113.
18 M. Schmeisser, Inorg. Syn., 1967, 9, 127.
19 G. H. Cady, Inorg. Syn., 1957, 5, 156.
20 J. A. Davidson, A. A. Viggiano, C. J. Howard, I. Dotan, F. C.
Fehsenfeld, D. L. Albritton and E. E. Ferguson, J. Chem. Phys.,
1978, 68, 2085.
21 J. Orphal, These de doctorat de lÏUniversite de Paris-Sud, Orsay,
1995.
22 C. J. Schack, Inorg. Chem., 1967, 6, 1938.
23 L. G. Huey, D. R. Hanson and C. J. Howard, J. Phys. Chem.,
1995, 99, 5001.
24 C. Amelynck, C. Stepien, N. Schoon, V. Catoire, D. Labonnette,
E. Arijs and G. Poulet, Int. J. Mass Spectrom., 2001, 207, 205.
25 J. B. Burkholder, R. K. Talukdar and A. R. Ravishankara,
Geophys. Res. L ett., 1994, 21, 585.
26 A. A. Viggiano, R. A. Morris and J. M. Van Doren, J. Geophys.
Res., 1994, 99, 8221.
27 A. M. Mebel and K. Morokuma, J. Phys. Chem., 1996, 100, 2985.
28 M. J. Frisch et al., GAUSSIAN 98, Revision A.6, Gaussian, Inc.,
Pittsburgh, PA, 1998.
29 T. Su and W. J. Chesnavich, J. Chem. Phys., 1982, 76, 5183.
30 W. J. Chesnavich, T. Su and M. T. Bowers, J. Chem. Phys., 1980,
72, 2641.
31 R. D. Surenam and D. R. Johnson, J. Mol. Spectros., 1977, 65,
239.
References
1
F. S. Rowland, J. E. Spencer and M. J. Molina, J. Phys. Chem.,
1976, 80, 2711.
2
3
S. Solomon, Nature, 1990, 347, 347.
W. H. Brune, J. G. Anderson, D. W. Toohey, D. W. Fahey, S. R.
Kawa, R. L. Jones, D. S. McKenna and L. R. Poole, Science,
1991, 252, 1260.
4
5
6
A. A. Viggiano, Mass Spectrom. Rev., 1993, 12, 115.
F. Arnold and S. Spreng, Geophys. Res. L ett., 1994, 21, 1251.
J. Schneider, F. Arnold, V. Burger, B. Droste-Franke, F. Grimm,
G. Kirchner, M. Klemm, T. Stilp, K.-H. Wohlfrom, P. Siegmund
and P. F. J. Van Velthoven, J. Geophys. Res., 1998, 103, 25337.
L. G. Huey, E. J. Dunlea, E. R. Lovejoy, D. R. Hanson, R. B.
Norton, F. C. Fehsenfeld and C. J. Howard, J. Geophys. Res.,
1998, 103, 3355.
E. Arijs, A. Barassin, E. Kopp, C. Amelynck, V. Catoire, H.-P.
Fink, C. Guimbaud, U. Jenzer, D. Labonnette, W. Luithardt, E.
Neefs, D. Nevejans, N. Schoon and A.-M. Van Bavel, Int. J.
Mass Spectrom., 1998, 181, 99.
T. Reiner, O. Mohler and F. Arnold, J. Geophys. Res., 1999, 104,
13943.
7
8
32 R. D. Nelson, D. R. Lide and A. A. Maryott, Natl. Stand. Ref.
9
Data Ser., U.S. Natl. Bur. Stand., 1967, 10.
33 M. M. Abbas, H. A. Michelsen, M. R. Gunson, M. C. Abrams,
M. J. Newchurch, R. J. Salawitch, A. Y. Chang, A. Goldman,
F. W. Irion, G. L. Manney, E. J. Moyer, R. Nagaraju, C. P. Rins-
land, G. P. Stiller and R. Zander, Geophys. Res. L ett., 1996, 23,
2401.
10 F. Arnold and G. Knop, Int. J. Mass Spectrom. Ion Processes,
1987, 81, 33.
11 L. G. Huey, P. W. Villalta, E. J. Dunlea, D. R. Hanson and C. J.
Howard, J. Phys. Chem., 1996, 100, 190.
Phys. Chem. Chem. Phys., 2001, 3, 3683È3689
3689