Kinetic study of the reactions of CF3O2 radicals with Cl and NO

Article abstract of DOI:10.1039/a905933d

Kinetic studies of the reactions CF₃O₂ + Cl and CF₃O₂ + NO were performed at room temperature in the gas phase using the discharge flow mass spectrometric technique (DFMS). The reactions were investigated under pseudo- first-order conditions with Cl or NO in large excess with respect to the CF₃O₂ radicals. The rate constant for the reaction CF₃O₂ + NO was measured at 298 K and the value of (1.6 ± 0.3) x 10^-11 cm³ molecule^-1 s^-1 is in very good agreement with all previous values. For the reaction CF₃O₂ + Cl, we obtain a rate constant at 298 K of (4.2 ± 0.8) x 10^-11 cm³ molecule^-1 s^-1 in excellent agreement with the only published value. Product analysis shows that this reaction occurs via the major reaction pathway CF₃O₂ + Cl → CF₃O + ClO at room temperature. In addition, an ab initio theoretical study was performed to gain insights on the different postulated reaction pathways. There is a significant disagreement between experimental and ab initio values recommended for the formation enthalpies of CF₂O, CF₃O and related molecules produced in this system. Consequently, we provide self-consistent values of enthalpies based on isodesmic reactions for the CF₃O₂ + Cl reaction system using the G2, G2(MP2) and CBS-Q methods. These values are also compared with BAC-MP4 heats of formation calculated in this work.

Full text of DOI:10.1039/a905933d

View Article Online / Journal Homepage / Table of Contents for this issue
Kinetic study of the reactions of CF O radicals with Cl and NO
3
2
Florent Louis,a Donald R. Burgess, Jr.,b Marie-Th e re` se Rayezc and Jean-Pierre Sawerysyn*a
a L aboratoire de Cine
Recherches L asers et Applications, Universite
V illeneuve dÏAscq, France

tique et Chimie de la Combustion UMR-CNRS 8522, Centre dÏEtudes et de

des Sciences et T echnologies de L ille, 59655
b National Institute of Standards and T echnology, Physical and Chemical Properties Division,
Gaithersburg, Maryland 20899, USA
c L aboratoire de Physicochimie Moleculaire UMR-CNRS 5803, Universite de Bordeaux I, 33405
 
T alence, France
Received 22nd July 1999, Accepted 1st October 1999
Kinetic studies of the reactions CF O ] Cl and CF O ] NO were performed at room temperature in the
3
2
3 2
gas phase using the discharge Ñow mass spectrometric technique (DFMS). The reactions were investigated
under pseudo-Ðrst-order conditions with Cl or NO in large excess with respect to the CF O radicals. The rate
3
2
constant for the reaction CF O ] NO was measured at 298 K and the value of (1.6 ^ 0.3) ] 10~11 cm3
molecule~1 s~1 is in very good agreement with all previous values. For the reaction CF O ] Cl, we obtain a
3
2
3
2
rate constant at 298 K of (4.2 ^ 0.8) ] 10~11 cm3 molecule~1 s~1 in excellent agreement with the only
published value. Product analysis shows that this reaction occurs via the major reaction pathway
CF O ] Cl ] CF O ] ClO at room temperature. In addition, an ab initio theoretical study was performed to
3
2
3
gain insights on the di†erent postulated reaction pathways. There is a signiÐcant disagreement between
experimental and ab initio values recommended for the formation enthalpies of CF O, CF O and related
2
3
molecules produced in this system. Consequently, we provide self-consistent values of enthalpies based on
isodesmic reactions for the CF O ] Cl reaction system using the G2, G2(MP2) and CBS-Q methods. These
3
2
values are also compared with BAC-MP4 heats of formation calculated in this work.
CF O ] Cl ] Products
(2)
Introduction
3
2
We have determined the rate coefficient (k ) for the reaction of
Recently, considerable attention has been focused on the fate
of free radicals produced by atmospheric degradation of
hydroÑuorocarbons (HFCs) and hydrochloroÑuorocarbons
2
CF O radicals with Cl atoms at room temperature using
3 2
mass spectrometry to directly monitor the temporal decay of
CF O radicals in the presence of Cl atoms in excess. This
reaction was considered as a possible secondary reaction
occurring under laboratory conditions in the system
(
HCFCs) proposed as environmentally acceptable alternatives
3 2
to replace chloroÑuorocarbons (CFCs). A few of these mol-
ecules, such as CF CH F (HFC-134a) and CF CHF (HFC-
3
2
3
2
CF O /ClO /O (x \ 0, 1) corresponding to the postulated
125), release the CF radical during oxidation in the atmo-
3 x
x
3
3
catalytic cycle for destroying the O layer2,7 in polar strato-
sphere. Once liberated, these CF radicals are quickly con-
verted into CF O and CF O radicals. It has been
3
3
spheric clouds. To the best of our knowledge, this reaction has
3
2
3
only been studied by Canosa-Mas et al.11 using resonance
Ñuorescence of Cl atoms for monitoring the reaction. Utilizing
mass spectrometry, we have also identiÐed and quantiÐed the
reaction products in this system. In addition, we have per-
formed ab initio theoretical studies in order to gain insight on
the possible di†erent reaction pathways. Furthermore, in
order to validate our experimental protocol of generation and
suggested1,2 that these radicals might play an important cata-
lytic role in ozone depletion in the stratosphere. However,
recent studies at room temperature of the CF O ] O and
3
2
3
CF O ] O reactions3h8 have shown that these reactions are
3
3
too slow to be an important cause of stratospheric O deple-
3
tion. Nevertheless, to improve the knowledge on the potential
reactivity of CF O and CF O radicals in the atmosphere,
kinetics of the reactions of CF O and CF O with di†erent
trace species were investigated in a systematic manner by a
few research groups. For example, the reactivity of CF O
3
2
3
monitoring of CF O radicals, we have re-investigated the
3 2
3
2
3
reaction (1) CF O ] NO, the rate constant of which is well
3 2
known at room temperature.
3
2
radicals has recently been probed using kinetic studies by
Biggs et al.9,10 They report data for reactions between CF O
Experimental
3
2
and O(3P) atoms,9 OH and HO . They used laser-induced
2
Method
Ñuorescence of NO (produced from CF O ] NO titration),
2
3 2
resonance Ñuorescence of OH,10 and chemiluminescence
detection of O(3P) atoms9 as detection methods.
In this paper, we report experimental and ab initio data for
the following reactions:
Experiments were carried out using a discharge Ñow system
with mass spectrometer detection (DFMS). The experimental
set-up has been described in detail elsewhere.12,13 Only spe-
ciÐc information is given here. The reactor consisted of a 60
cm long, 2.4 cm diameter Pyrex tube. Typical Ñow velocities
were 2.3È2.9 m s~1 and pressures of 100È400 Pa (0.7È3.0 Torr)
CF O ] NO ] CF O ] NO
(1)
3
2
3
2
Phys. Chem. Chem. Phys., 1999, 1, 5087È5096
This journal is ( The Owner Societies 1999
5087

View Article Online
Results and discussion
(
1) CF O + NO Ç CF O + NO (k )
3 2 3 2 1
In order to validate our experimental protocol, we have re-
investigated the rate coefficient k at 298 K and 133 Pa (1
1
Torr). Fig. 2 shows typical plots of ln[I(CF O `)] as a func-
2
2
tion of time for several initial concentrations of NO ([NO] ).
Fig. 3 displays the plot of the pseudo-Ðrst-order rate constant
Fig. 1 Schematic diagram of the discharge Ñow tube.
0
k
as a function of [NO] . Analysis of these data using
obs
0
linear least squares regression yields a bimolecular rate con-
stant:
with He as carrier gas. The reactor walls were coated with
halocarbon wax in order to reduce the heterogeneous loss of
reactive species. Fig. 1 shows the introduction of various
radical precursors and the production of radicals of interest.
CF O radicals were produced in a double sliding injector
k (298 K) \ (1.5 ^ 0.1) ] 10~11 cm3 molecule~1 s~1
1
Here we have taken into account corrections due to radial
and axial di†usion e†ects (\10%) on the concentration dis-
3
2
tribution of CF O radicals. The di†usion coefficient for
(
Fig. 1) from the reaction sequence:
F ] CHF ] HF ] CF
3 2
CF O in the reactor has been estimated to be 335 cm2 s~1 at
3
2
(3)
(4)
133 Pa (1 Torr) and 300 K using Lennard-Jones parameters of
3
3
CF Cl as suggested by Caralp et al.20 The stated errors are
CF ] O ] He ] CF O ] He
3
only statistical uncertainties at the 95% conÐdence range for
3
2
3 2
the Ðt of experimental data. However, there are additional
uncertainties in the experimental measurements. In the
absence of NO, we observed a decrease (\5%) in the intensity
of CF O ` ions through the reactor. This decrease is likely
F atoms were generated by passing a mixture F ÈHe through
a microwave discharge formed in an alumina tube by an
2
Evenson cavity before reacting with the CHF ÈO mixture.
3
2
Attack of the uncoated Pyrex wall just after the alumina dis-
2
2
due to recombination reactions of CF O radicals. We have
charge tube by F atoms produces SiF , which is detected at
3
2
4
found that heterogeneous losses are unimportant on this time
scale when this reactor is coated with halocarbon wax. Table
m/z 85 (SiF `). Initial concentrations of the precursors were:
3
[
F] \ (3.8È6.7) ] 1012 cm~3, [CHF ] \ (0.9È1.3) ] 1015
3
1
shows the recombination reactions responsible for consump-
cm~3 and [O ] \ (0.9È11.2) ] 1015 cm~3. Concentrations of
2
tion of CF O along the reaction zone and secondary reac-
Ñuorine atoms were determined by monitoring the disap-
3
2
tions involving NO and CF O. Numerical modeling of the
pearance of Cl2 with the mass spectrometer using the fast
3
chemical mechanism including secondary reactions (Table 1)
titration reaction F ] Cl ] FCl ] Cl as proposed by Clyne
2
and reaction (1) shows that the consumption of NO due to its
et al.14 CF O radicals were titrated with excess NO using the
3
2
reaction with CF O is less than 17%. This consumption of
reaction CF O ] NO ] CF O ] NO and monitoring the
amount of NO produced. Cl atoms were also generated by a
microwave discharge in a mixture of Cl ÈHe through a Ðxed
sidearm located in the upstream part of the Ñow tube. The
3
3
2
2
3
2
NO is responsible of the X-intercept (1.6 ] 1012 cm~3)
obtained on the plot k \ f([NO] ) displayed in Fig. 3. In
obs
0
2
concentrations of chlorine atoms were determined by mass
spectrometric measurements by monitoring the production of
C H Cl formed in the reaction Cl ] CH CHBr ] Br
2
3
2
]
CH CHCl as proposed by Park et al.15 In order to quan-
2
tify the concentration of ClO radicals produced by the reac-
tion CF O ] Cl, we generated ClO radicals by the reaction
3
2
Cl ] OClO ] 2 ClO16 and quantiÐed their concentrations by
the fast reaction ClO ] NO ] Cl ] NO using a large excess
2
of NO and measuring the NO product.
2
Detection
All experiments were performed under pseudo-Ðrst-order con-
ditions with Cl atoms (or NO) in excess with respect to the
initial concentration of CF O radicals. Kinetic measurements
Fig. 2 Plots of ln[I(CF O `)] vs. reaction time at 298 K [NO]
2
2
0
3
2
(1013 cm~3): A \ 1.41; B \ 2.05; C \ 2.70; D \ 3.35; E \ 3.99;
F \ 4.65; G \ 5.30.
were carried out by monitoring the temporal decay of the
intensity of the fragment ion (CF O `) detected at the mass/
2
2
charge ratio m/z 82. This ion (detected at 40 eV) is speciÐc to
CF O radicals as previously shown by di†erent authors
3
2
using a similar technique.17h19.
Materials
CHF ([99.995%) was obtained from Air Liquide. The
3
source of F (or Cl ) was a commercial mixture (Alphagaz)
2
2
constituted of 5% of F ([99.3%) or 2% of Cl ([99.99%) in
2
2
helium ([99.9995%). As an additional diluent, helium (Air
Liquide, purity [99.995%) was puriÐed by circulating
through a liquid nitrogen trap and molecular sieves with indi-
cator. NO ([99%) was also puriÐed by slowly circulating
through a cold trap at [110 ¡C to eliminate traces of NO .
2
NO ([98%) was degassed by repetitive freeze-pump-thaw
2
cycles and highly diluted with helium in a darkened 20 L glass
Fig. 3 Pseudo-Ðrst-order rate constants k plotted vs. [NO]
0 for
obs
bulb.
the reaction CF O ] NO at 298 K.
3
2
5088
Phys. Chem. Chem. Phys., 1999, 1, 5087È5096

View Article Online
Table 1 Secondary reactions considered for modeling the temporal
Table 2 Comparison of our kinetic results with the previous liter-
ature data concerning the reaction CF O ] NO ] CF O ] NO
proÐles of NO along the reaction distance for the system CF O
2 at
3
2
3
2
3
]
NO
room temperature
k (298 K)/
cm3 molecule~1 s~1
k ] 1011/
Reactions
Ref.
cm3 molecule~1 s~1
Techniquea
Ref.
1
.7 ^ 0.3
DFMS
LPMS
DFMS
17
18
19
CF O ] CF O ] CF O ] CF O ] O
1.2 ] 10~12
2.5 ] 10~11
2.5 ] 10~11
4.7 ] 10~11
21
22
22
23
8
6
0
5
0
0
6
8
1
5
3
2
3
2
3
3
2
1.4 ^ 0.2
CF O ] CF O ] CF OOOCF
5
3
2
3
3
3
1.5 ^ 0.3
CF O ] CF O ] CF OOCF
4
3
3
3
3
1.5 ^ 0.4
CF O ] NO ] CF O ] FNO
1
3
2
1.5 ^ 0.2
FTCIMS
PRUV
LPLIF
FPMS
DFLIF
DFMS
24
25
6
26
23
3
1
1
1
1
.6 ^ 0.2
8
.5 ^ 0.3
7
.5 ^ 0.3
summary, taking into account corrections due to the recombi-
nation reactions of CF O radicals and consumption of NO,
7
.7 ^ 0.3
6
3
2
1.6 ^ 0.3
This work
the kinetic data leads to a corrected rate constant k
, c at 298
1
a DFMS \ discharge Ñow mass spectrometry; LPMS \ laser pho-
tolysis mass spectrometry; FTCIMS \ Ñow tube chemical ionization
mass spectrometry; PRUV \ pulse radiolysis ultra violet spectros-
copy; LPLIF \ laser photolysis laser induced Ñuorescence;
FPMS \ Ñash photolysis mass spectrometry; DFLIF \ discharge
Ñow laser induced Ñuorescence.
K:
k
(298 K) \ (1.6 ^ 0.3) ] 10~11 cm3 molecule~1 s~1
1, c
This value is in very good agreement with other reported
values6,17h19,23h26 (Table 2) using a number of di†erent
methods. Consequently, we have conÐdence in our method of
generation and monitoring of CF O radicals.
3
2
exothermic); the reactions producing CF Cl, CF O and
(
2) CF O + Cl Ç products (k )
3
2
3 2 2
Products studies. By analogy to the reaction pathways pro-
CF OOCl/CF OClO are strongly exothermic; and the reac-
3 3
tions producing CF and CF O are strongly endothermic.
3
2 2
posed by Jungkamp et al.27 for the reaction CH O ] Cl, we
propose the following potential channels for reactions between
CF O radicals and Cl atoms:
It is important to note here that there is a complication in
this system in that the formation enthalpy for one of the
3
2
important reference molecules, CF O, is the subject of signiÐ-
3
2
2
cant disagreement between the experimental and ab initio
CF O ] Cl ] CF O ] ClO
(2a)
(2b)
3
2
3
values. The recommended experimental formation enthalpy
*
H \ [11 kJ mol~1
for CF O is [641.0 kJ mol~1,28 while the recommended ab
r
2
initio value is [608.0 kJ mol~1.29,30 That is, the ab initio-
]
1CF O ] FCl
2
2
derived values are approximately 30 kJ mol~1 higher than the
experimental values and this di†erence is outside the reported
uncertainties in the values. Furthermore, another one of the
*
H \ ]179 kJ mol~1
r
E \ ]267 kJ mol~1
reference molecules in this system, CF O, has a formation
a
3
enthalpy that is referenced through experimental measure-
]
]
]
]
3CF O ] FCl
2
2
ments to the formation enthalpy for CF O. It has been noted
2
*
H \ ]294 kJ mol~1
in analysis of the reaction data for CF O, CF O, CF OF and
r
2
3
3
related molecules, that side reactions may complicate accurate
determinations of thermochemical quantities in these systems
(see references and more discussion in the Appendix). This
current work does not attempt to resolve the disagreement
between experimental and ab initio values for the enthalpies of
formation of CF O, CF O and related molecules. Conse-
quently, the energetics reported here are dependent upon the
values chosen for the reference molecules. Towards this
purpose, we have attempted to employ values that are self-
consistent. We note that recent work strongly suggests that
CF Cl ] O (3&)
(2c)
3
2
*
H \ [190 kJ mol~1
r
CF Cl ] O (1*)
3
2
*
H \ [96 kJ mol~1
r
2
3
CF O ] FOCl
(2d)
(2e)
2
*
H \ [42 kJ mol~1
r
E \ ]78 kJ mol~1
the experimental value for CF O is an error.29h32
a
2
In contrast to the reaction CH O ] Cl, only the bimolecu-
]
]
]
]
CF ] ClOO
3 2
3
lar channels forming CF O (2a), CF Cl (2c) and CF O (2d)
3
3
2
*
H \ ]143 kJ mol~1
are thermodynamically favored at room temperature. These
are in addition to the combination reactions (2f) forming
CF OOCl and CF OOClO complexes. Channel (2b), involv-
ing the formation of CF O (both singlet and triplet states), is
strongly endothermic and can be neglected at room tem-
r
CF ] OClO
3
3
3
*
H \ ]148 kJ mol~1
r
2 2
CF OOCl
(2f)
perature. Similarly, the highly endothermic channel (2e),
3
involving the formation of CF (and the isomers OClO or
*
H \ [151 kJ mol~1
3
r
ClOO), can also be neglected at room temperature. In order
CF OClO
to evaluate the relative importance of these di†erent channels,
a systematic study by mass spectrometry was performed to
detect the di†erent potential products formed by these reac-
tions.
3
*
H \ [112 kJ mol~1
r
The reaction enthalpies at 298 K (* H) and the reaction bar-
r
riers or activation energies (E ) for the CF O ] Cl reaction
Reaction (2c) producing CF Cl and O (1* or 3&) is
a
3 2
3
2
system were calculated using the formation enthalpies [see
Table 5] given in the literature and from G2 calculations [see
Appendix for further discussion]. We can see that the reaction
strongly exothermic at room temperature. However, yields of
O cannot be determined under our experimental conditions
2
because very high concentrations (B1015 molecules cm~3) of
producing CF O is nearly thermoneutral (slightly
O
are used to generate CF O radicals. The potential pro-
3
2
3 2
Phys. Chem. Chem. Phys., 1999, 1, 5087È5096
5089

View Article Online
duction of CF Cl can be experimentally observed by mass
3
spectrometry using both isotopic peaks at m/z 85 and 87
(
CF 35Cl` and CF 37Cl`). In the absence of CHF , a signiÐ-
2
2
3
cant contribution to the peak at m/z 85 is from the SiF `
fragment ion produced by F attack on uncoated Pyrex walls.
3
So, only the intensity of the peak at m/z 87 (CF 37Cl`) was
2
investigated to highlight the formation of CF Cl. No contri-
3
bution, however, to this peak was observed under our experi-
mental conditions. Hence, reaction (2c) can be considered
negligible at room temperature within our detection limits.
Finally, the reaction between CF O radicals and Cl atoms
3
2
is characterized by the appearance of a mass peak at m/z 53
Fig. 4 Plot of consumed CF O vs. formed ClO at 298 K, P \ 150
(
37ClO`) the intensity of which increases as a function of reac-
3
2
Pa (1.14 Torr), v \ 2506 cm s~1, [CF O ] \ 3.15 ] 1012 cm~3,
tion time. The intensity of this peak can have two contribu-
3 2
[
Cl] \ 1.31 ] 1013 cm~3. [CF O ] and [ClO] in cm~3.
tions: 37ClO` fragment ion of FO37Cl [produced by reaction
0
3 2
(
2d)] and 37ClO` ion resulting from the ionization of ClO
radicals [produced by reaction (2a)]. That is, I \ I
(
5
3
53
Determination of the rate constant at 298 K. Typical tempo-
ral decay plots of ln[I(CF O `)] are displayed in Fig. 5 for
FO37Cl) ] I (37ClO). The production of FOCl could also
53
be monitored using the intensity of the molecular peak
FO35Cl or FO37Cl at m/z 70 or 72. Again, the contribution of
this product to these peaks cannot be distinguished from the
at m/z 70 (35Cl ) and 72 (35Cl37Cl) due to the
2
2
several initial concentrations of Cl atoms. Least squares
analysis provides values of the pseudo-Ðrst-rate constant k
obs
and plots of k including di†usion corrections vs. [Cl] gives
ones of Cl2
2
obs
0
the values of the bimolecular rate constant k (Fig. 6):
relatively high Cl concentrations (B1013 molecules cm~3)
2
2
used to generate Cl atoms. To assess the relative importance
k (298 K) \ (3.9 ^ 0.2) ] 10~11 cm3 molecule~1 s~1
2
of (2d), it was necessary to undertake a theoretical approach.
As shown in the second section of this paper, the potential
energy surface describing this channel is characterized by
several reaction intermediates (two complexes and one tran-
sition state). The G2(MP2) calculated activation barrier
with statistical uncertainties at the 95% conÐdence range for
the Ðt of the experimental data. Table 3 shows the secondary
reactions considered for evaluating their impact on the con-
centration proÐles of CF O radicals and Cl atoms along the
3
2
investigated reaction zone. The numerical treatment of this
mechanism indicates that these reactions can contribute
between the complex CF OClO and the cyclic transition state
3
is high enough to make channel (2d) very unlikely (80 kJ
\
21% to the consumption of Cl atoms. The consumption of
Cl atoms along the reaction zone is highlighted by the X-
mr e ol al t~i v1 e to the reactants CF O ] Cl and 190È230 kJ mol~1
3
2
relative to the CF OClO and CF OOCl complexes). Conse-
3
3
intercept (1.2 ] 1012 cm~3) obtained on the plot k
\
quently, the variation of the intensity of peak m/z 53 is con-
obs
f([Cl] ) of Fig. 6. Taking a mean variation of 10% for Cl
sidered as mainly due to ClO radicals, i.e. I \ I (37ClO).
0
5
3
53
atoms through the reactor, we derive a corrected value k
In order to evaluate the branching ratio k /k , we have
2, c
2
a 2
(including corrections due to di†usion e†ects and impact of
quantiÐed the production of ClO radicals and the consump-
tion of CF O radicals. In Fig. 4 is represented the plot of
ClO formed vs. CF O destroyed. The linear squares regres-
3
2
3
2
sion treatment of this plot gives a slope of 1.01 ^ 0.08 in 95%
conÐdence intervals indicating that the ClO production is
largely major under our experimental conditions. Therefore
we conclude that k /k B 1 within experimental uncertainties.
2a 2
This experimental study shows that the reaction between
CF O radicals and Cl atoms occurs predominantly via the
3
2
following reaction pathway:
CF O ] Cl ] CF O ] ClO
(2a)
3
2
3
Using the titration reaction ClO ] NO ] Cl ] NO ,
2
Canosa-Mas et al.11 recovered the initial value of the RF
signal which was obtained for Cl atoms before the addition of
CF O radicals, and also concluded that ClO formation
Fig. 5 Plots of ln[I(CF O `)] vs. reaction time at 298 K [Cl] (1013
2 2
atoms cm~3): A \ 1.38; B \ 1.54; C \ 1.95; D \ 2.12; E \ 2.31.
3
2
0
occurred very efficiently.
Table 3 Secondary reactions considered for modeling the temporal proÐles of Cl atoms along the reaction distance for the system CF O ] Cl
3
2
k (298 K)/
cm3 molecule~1 s~1
Reactions
Ref.
CF O ] CF O ] CF O ] CF O ] O
1.2 ] 10~12
2.5 ] 10~11
2.5 ] 10~11
2.2 ] 10~12
1.9 ] 10~11
4.3 ] 10~16
2.7 ] 10~33 [M]
2.3 ] 10~10
1.2 ] 10~11
21
22
22
Analogya
Analogyb
35
36
36
36
3
2
3
2
3
3
2
CF O ] CF O ] CF OOOCF
3
2
3
3
3
CF O ] CF O ] CF OOCF
3
3
3
3
CF O ] ClO ] products
3
2
CF O ] Cl ] products
3
Cl ] CHF ] HCl ] CF
3
3
Cl ] O ] M ] ClOO ] M
2
Cl ] ClOO ] Cl ] O
2
2
Cl ] ClOO ] ClO ] ClO
a Rate constant assimilate to the rate constant found for the reaction CH O ] ClO from ref. 33. b Rate constant assimilate to the rate constant
3
2
found for the reaction CH O ] Cl from ref. 34.
3
5090
Phys. Chem. Chem. Phys., 1999, 1, 5087È5096

View Article Online
of the elements and the thermal contributions ** H¡ (298.15
f
K) to the enthalpies of the reference states are also given in
Table 4. ** H¡ (298.15 K) is the di†erence between * H¡ at
f
f
the reference temperature 298.15 K and 0 K.
As indicated previously, the reaction enthalpies for the
CF O ] Cl reaction system were calculated using the forma-
3
2
tion enthalpies given in the literature and from G2 calcu-
lations (see Table 5). For the largest molecules and the
transition states, where full G2 calculations were impractical,
we used the less computationally expensive G2(MP2) method.
We also calculated formation enthalpies using the CBS-Q
method to aid in assigning uncertainties to the ab initio-
derived formation enthalpies. The ab initio values were cor-
rected using isodesmic reactions, where the formation
enthalpies of the reference molecules have well-deÐned experi-
mental values.
Fig. 6 Pseudo-Ðrst-order rate constants k plotted vs. [Cl] for the
obs
reaction CF O ] Cl at 298 K in the pressure range 100È400 Pa (0.7È
0
3
2
3
.0 Torr).
secondary reactions):
(298 K) \ (4.2 ^ 0.8) ] 10~11 cm3 molecule~1 s~1
As mentioned previously, there is a disagreement between
experimental and ab initio values for the enthalpies of forma-
tion of CF O, CF O and related molecules. Toward this
k
2, c
2
3
This value is identical to that determined by Canosa-Mas et
al.11 They used resonance Ñuorescence to probe Cl atoms
reacting with CF O radicals in excess.
purpose, we have attempted to employ values that are self-
consistent. Consequently, the energetics reported here are
dependent upon the values chosen for the reference molecules.
We refer the reader to the Appendix for the rationale for
our selection of self-consistent values for the formation enth-
alpies of molecules in the CF O ] Cl reaction system.
3
2
Theoretical
Computational methods
3
2
Structures and energies for intermediate radical recombination
Ab initio molecular orbital calculations have been carried out
using the program package Gaussian 9437 to compute opti-
mized geometries, vibrational frequencies and relative stabil-
ities of possible intermediates produced along the reaction
pathways. For G2 calculations,38,39 geometries were initially
fully optimized at the Hartree-Fock level using the 6-31G(d)
basis set. Harmonic vibrational frequencies were calculated at
this level using analytical second-order derivatives. Structures
were reoptimized at the MP2(Full)/6-31G(d) level of theory.
For CBS-Q calculations,40 the 6-31G(d@) basis set was used
for geometry optimization and calculation of vibrational
frequencies. A series of higher level single-point energy
calculations were then performed as deÐned by the G2,38
G2(MP2),39 and CBS-Q40 methods. Total energies for
each molecule were then determined from the composite of
the single-point energy calculations for each method.
products of stoichiometry CF O Cl
3 2
Two stable isomers were characterized by ab initio methods
for the reaction CF O ] Cl: CF OOCl and CF OClO. The
3
2
3
3
optimized geometric parameters and vibrational frequencies
calculated at di†erent levels of theory are given as supplemen-
tary information.¤ The CF OOCl isomer represents the inter-
3
mediate complex directly formed via reaction (2a). The
CF OClO isomer could be produced as a secondary interme-
3
diate complex from encounters between CF O and ClO as
3
products of (2a). The heats of formation of these two complex-
es calculated by the G2 and CBS-Q methods are given in
Table 5. Relative to CF OOCl, the CF OClO isomer was
3
3
found to be about 40 kJ mol~1 higher in energy. No tran-
sition state connecting the two complexes on the potential
energy singlet surface was found for both UHF/6-31G(d) and
MP2(Full)/6-31G(d). The G2(MP2) results indicate that the
The G2 and CBS energies were also compared with those
obtained by the BAC-MP4 method.41 In the latter method,
bond additivity corrections are added to MP4(SDTQ)/6-
recombination product CF OOCl is estimated to be formed
3
in reaction (2a) with about 150 kJ mol~1 excess energy with
31G(d,p) energies in order to account for the basis set trunca-
respect to its reactants.
tion. Additional corrections are made for spin contamination
in open shell systems.
Reaction (2a). The overall reaction is slightly exothermic
(
[11 kJ mol~1) with no barrier in the reaction path. This
Results and discussion
Calculations of formation enthalpies of studied species
reaction can be considered to proceed as the barrierless com-
bination of the two radicals CF O and Cl forming the chemi-
3 2
cally activated intermediate CF OOCl (with 150 kJ mol~1)
3
that decomposes by direct Ðssion of the OÈO bond leading to
The formation enthalpies of all species were calculated from
the products CF O and ClO.
atomization energies. Table
4 lists calculated ab initio
3
G2(MP2), G2 and CBS-Q total energies for the reference ele-
ments C, H, O, F and Cl. Recommended literature values for
¤ Available as electronic supplementary information. See http://
www.rsc.org/suppdata/cp/1999/5087.
the enthalpies of formation * H¡ (298.15 K) at standard state
f
Table 4 Total energies and formation enthalpies for reference atoms
Total Energy/E
Enthalpy/kJ mol~1
h
G2(MP2)
G2
CBS-Q
** H¡
* H¡
Uncert.
Ref.
f
f
C
[37.783 887
[0.500 000
[37.784 301
[37.785 151
[0.499 818
5.48
1.96
2.38
2.11
1.68
716.67
218.00
249.17
79.39
0.46
0.01
0.10
0.30
0.01
28
28
28
28
28
H
O
F
[0.500 000
[74.982 030
[99.632 814
[459.676 627
[74.978 678
[99.628 941
[459.666 717
[74.987 050
[99.642 238
[459.682 894
Cl
121.30
Phys. Chem. Chem. Phys., 1999, 1, 5087È5096
5091

View Article Online
Table 5 Electronic dissociation energies, zero point energies and formation enthalpies for CF O ] Cl system
3
2
D /kJ mol~1 a
Enthalpy/kJ mol~1
e
Lit/Calcb
G2(MP2)c
G2d
494.2
380.5
972.9
1123.4
729.6
CBS-Qe
ZPEf
Ref. 1g
** H¡ h
Ref. 2i
* H¡ j
Uncert.
Ref. 3k
Datal
f
f
O
O
H O
H O
HO
(3&)
(1*)
503.0
408.2
971.1
1122.3
729.6
495.7
384.1
976.3
1129.3
733.7
506.3
383.2
974.0
1127.4
731.9
9.6
8.9
53.3
66.7
35.5
42
43
42
43
43
0.0
0.0
[2.9
[6.0
[2.9
42
44
42
45
45
0.0
0.00
0.5
0.04
0.2
42
43
42
45
46
a
a
a
a
a
2
2
94.3
[241.8
[135.9
12.6
2
2
2
1.7
2
CH
4
1755.4
1285.6
2976.6
2354.9
1562.4
1671.8
1942.3
1755.0
1280.4
2973.1
2351.0
1574.3
1672.2
1946.4
1757.2
1282.4
2976.1
2353.1
1571.8
1672.9
1943.4
1756.0
1283.3
2975.8
2352.8
1568.8
1673.4
1945.8
113.4
76.1
189.7
129.3
67.5
90.4
108.4
47
43
47
47
47
43
50
[8.0
[3.4
[15.6
[8.5
[3.8
[7.9
[9.8
28
28
28
28
28
49
51
[74.6
146.4
[84.0
52.4
0.6
0.4
0.4
0.5
0.5
4.2
5.0
28
48
28
28
28
49
52
a
a
a
a
a
a
a
CH
3
CH ÈCH
3
3
CH xCH
2
CH O
2
2
[108.7
CH O
3
16.7
CH O
9.2
3
2
CH F
1826.6
1914.1
1444.1
3083.0
3169.7
3297.6
2431.4
1841.6
1937.1
1456.7
3099.2
3198.5
3348.3
2457.1
1838.4
1932.3
1452.1
3097.4
3195.3
3339.2
2454.7
1834.3
1926.6
1450.1
3094.6
3190.8
3328.0
2453.4
84.2
66.3
31.9
155.0
134.5
77.3
53
47
43
54
54
54
28
[7.6
[7.0
[2.8
[13.9
[12.8
[8.3
[6.9
28
28
28
54
54
54
28
[452.2
[696.7
[472.0
[500.8
[745.6
[1342.6
[336.4
1.7
2.2
5.2
6.3
1.7
6.3
4.0
28
28
28
54
54
54
28
a
a
a
a
a
a
a
2
CHF
3
2
CF
3
CH ÈCHF
3
2
CH ÈCF
3
3
CF ÈCF
3
3
CH xCF
92.9
2
2
CF O
1785.0
1857.3
2122.6
1773.7
1879.9
2148.4
1768.3
1874.9
2140.1
1764.6
1867.9
2135.8
36.6
42.7
53.8
47
55
55
[3.1
[5.0
[6.9
42
51
51
[602.8
8.0
5.0
12.0
6.0
56
28
56
57
56
58
59
59
c
2
[
640.2
a
d
a
d
a
e
CF O
[629.8
655.6
[639.1
674.9
3
[
CF O2
12.0
3
[
1
CF O
3
1686.0
1569.4
1700.5
1584.0
1686.6
1578.1
40.5
40.6
55
55
[2.7
[3.9
51
51
[288.8
[173.4
14.0
14.0
2
CF O
2
2
2
e
HOF
HOCl
ClO
OClO
ClOO
FCl
670.3
687.3
270.3
524.0
1239.7
251.8
411.0
1805.1
667.4
696.4
263.8
521.5
662.9
690.1
260.5
510.6
664.9
694.7
268.6
532.1
34.8
33.3
5.2
43
43
42
42
[2.9
[3.0
[0.3
[2.5
42
42
42
42
[98.3
[74.5
101.2
101.7
97.5
4.2
2.1
2.2
6.3
1.7
0.4
10.0
3.3
42
42
42
60
61
42
56
42
a
a
a
a
b
a
d
a
15.0
263.5
411.5
1837.2
260.2
404.7
259.7
406.0
1827.0
4.7
13.1
39.6
42
62
42
[0.1
[2.3
[5.1
42
62
42
[50.3
FOCl
43.5
CF Cl
[707.9
3
CF ÈOOCl
2274.6
2233.6
1848.3
2037.3
2307.4
2266.4
1874.5
2062.6
2292.7
2246.5
60.2
57.1
48.3
47.9
55
55
55
55
[8.1
[7.3
[7.0
[6.9
51
51
51
51
[666.2
[627.6
[250.6
[440.0
12.0
15.0
14.0
14.0
59
59
59
59
e
e
e
e
3
CF ÈOClO
3
1
CF OOÉ É ÉFCl
2
CF OÉ É ÉFOCl
2041.7
2
a D , Electronic dissociation energy or atomization energy excluding ZPE. b Lit/Calc, D calculated using * H¡ from literature or calculated in this work. c G2(MP2),
e
e
f
D calculated using G2(MP2) total energies. d G2, D calculated using G2 total energies. e CBS-Q, D calculated using CBS-Q total energies. f ZPE, zero point energy
e
e
e
correction. g Ref. 1, Reference for vibrational frequencies used to compute ZPE and/or ** H¡. h ** H¡, Thermal correction to the enthalpy or [* H¡ (298.15
f
f
f
K) [ * H¡ (0 K)]. i Ref. 2, Reference for thermal correction to enthalpy or ** H¡. j * H¡, Enthalpy of formation at standard state or [* H¡ (298.15 K)]. k Ref. 3,
f
f
f
f
Reference for enthalpy of formation or * H¡ (298.15 K). l Data, Datatype; a, Experimental enthalpies. Experimental geometries and frequencies; b, experimental
f
enthalpies. MP2 geometries and HF scaled frequencies; c, G2/isodesmic reaction corrected enthalpies. Experimental geometries and frequencies; d, G2/isodesmic
reaction corrected enthalpies. MP2 geometries and HF scaled frequencies; e, G2(MP2)/isodesmic reaction corrected enthalpies. MP2 geometries and HF scaled
frequencies.
Reaction (2b). We note that the isomer used here is diÑuo-
the reactants. Obviously, thermodynamics at 298 K is not in
favor to these reaction pathways, conÐrming the experimental
non-observation of postulated products.
rocarbonyloxide É CF OO É, rather than the cyclic diÑuoro-
2
dioxirane È(OCF O)È, which is more stable by about 180 kJ
2
mol~1.63 No direct path to the latter was found. Two di†erent
channels are possible for this reaction: one on the singlet
potential energy surface, and one on the triplet surface. As
suggested by ab initio calculations performed for CH O ] Cl
Reaction (2c). Again, two pathways are potentially possible
for this reaction: one on the potential energy singlet surface,
one on the triplet surface. A four-member ring cyclic tran-
sition state involving a forming CÉ É ÉCl bond could be
assumed between the reactants and the formation of products
CF Cl ] O (1* or 3&). Although these reactions are very
3
2
reaction,27 a Ðve-member ring cyclic transition state including
a forming FÉ É ÉCl bond connects the primary CF OOCl
3
complex to the products 1CF O and FCl on the singlet
surface, whereas a transition state corresponding to the direct
attack of any Ñuorine atom of the CF O radical by a Cl
2
2
3
2
exothermic, no intermediate species were successfully charac-
terized on the two potential surfaces at both UHF/6-31G(d)
3
2
atom corresponds to the transition state occurring on the
triplet surface. The vibrational frequencies and optimized geo-
metric parameters calculated at di†erent levels of theory for
the singlet transition state and for the 1CF O and 3CF O
and MP2(Full)/6-31G(d) levels of theory. Furthermore, CF Cl
3
was not experimentally observed.
Reaction (2d). The production of CF O and FOCl from the
2
2
2 2
2
products are given as supplementary information.¤ Compared
reaction between CF O and Cl requires numerous interme-
3
2
to the Criegee intermediates assumed for the CH O ] Cl
diate species (two complexes and one Ðve-member ring cyclic
transition state), which make this pathway very unlikely. The
structural parameters of this cyclic transition state optimized
3
2
reaction, the formation of the biradicals 1CF O and 3CF O
2
2
2 2
require much more energy with respect to the energy level of
5092
Phys. Chem. Chem. Phys., 1999, 1, 5087È5096

View Article Online
at UHF/6-31G(d) and MP2(Full)/6-31G(d) are given as sup-
plementary information.¤ The saddle point connects the reac-
tant CF OClO and the products CF O and FOCl. The
rections (ZPE) and thermal corrections [** H¡] that are
f
based on experimental vibrational frequencies rather than
scaled ab initio frequencies. Using the following isodesmic
reactions to correct the ab initio energies
3
2
activation barrier, calculated with the G2(MP2) energies, was
found to be about 190 kJ mol~1 relative to the energy of the
CF O ] CH ] CH O ] CH F
intermediate CF OClO and still about 80 kJ mol~1 above the
2
4
2
2 2
3
original reactants CF O ] Cl. This channel is not likely to
CF O ] C H ] CH O ] CH CHF
3
2
2
2 6
2
3
2
be important at room temperature.
CF O ] C H ] CH O ] CH CF
2
2 4 2
2
2
Conclusion
we calculate G2/isodesmic formation enthalpies of [602.8,
[
597.9 and [587.7 kJ mol~1, respectively. In order to
The rate constants for the CF O ] Cl and CF O ] NO
3
2
3 2
provide an estimate of the uncertainty in the ab initio calcu-
lations, we also report a set of individual and composite calcu-
lations based on the G2 and CBS-Q methods.
reactions in the gas phase have been determined at room tem-
perature using the discharge Ñow mass spectrometric tech-
nique (DFMS). The reactions were investigated under
pseudo-Ðrst-order conditions with Cl or NO in large excess
with respect to the CF O radicals. The value measured at
Using the Ðrst reaction (CH F reference), we Ðnd iso-
2
2
desmic reaction-corrected formation enthalpies of [600.2,
3
2
[
594.8, [605.4, [600.5 and [602.8 kJ mol~1 using MP4/
2
98 K for the rate constant of CF O ] NO reaction was in a
very good agreement with all previous values. For the CF O
3
2
TZ ] P, QCI/(TZ ] P), G1, G2(MP2) and G2 energies,
respectively. Similarly, we Ðnd [600.9, [602.1 and [604.9
kJ mol~1 using MP4/(CbsB3), MP4/(CbsB3 ] CBS) and
CBS-Q energies, respectively. These data correspond to a
range of about 8 kJ mol~1 in the ab initio calculations.
3
2
]
Cl reaction, the rate constant obtained at 298 K was in
excellent agreement with the only published value. For this
reaction, an experimental analysis of the reaction products
was also performed. ClO was identiÐed as a product showing
that the reaction between the CF O radicals and Cl atoms
DeÐning G2 as the base calculation and using the CH F
isodesmic reaction as the most elementary correction, we
2
2
3
2
occurs at room temperature via the following reaction channel
select the G2/iso(CH F ) value for CF O of * H¡ \ [602.8
CF O ] Cl ] CF O ] ClO. Furthermore, an ab initio theo-
2
2
2
f
3
2
3
^
8.0 kJ mol~1 for use in this work. We assigned an uncer-
retical study was performed to gain insight on the di†erent
postulated reaction channels. Self-consistent values of enth-
alpies based on isodesmic reactions are proposed for the
CF O ] Cl reactions using the G2, G2(MP2) and CBS-Q
tainty of 8 kJ mol~1 as the sum of an estimated 4 kJ mol~1
for the ab initio calculations and a contribution of 4 kJ mol~1
from the reported uncertainties in the reference molecules (see
Table 5). We note that this value is 37 kJ mol~1 higher than
the ““recommended experimentalÏÏ value28 and also slightly
higher (5 kJ mol~1) than the previously ““recommended ab
initioÏÏ value.29,30 We emphasize that because there is a dis-
agreement between experimental and ab initio values for the
enthalpies of formation of CF O, CF O and related mol-
3
2
methods. BAC-MP4 formation heats of di†erent species of
interest were calculated and compared with the values
obtained using the previous methods.
Acknowledgements
The authors acknowledge the IDRIS Computer Center
2
3
ecules, the absolute energetics reported here are dependent
upon the values chosen for the reference molecules. We have
however attempted to employ values that are self-consistent
(correct relative energies).
(
Orsay, France) for providing computing time for the theoreti-
cal calculations under projects 97.0792 and 98.0792. We are
also grateful to the Ministere de la Recherche et de
lÏEnseignement Superieur, the Region Nord/Pas de Calais and
the Fonds Europeen de Developpement Economique des Re
`





-
CF O
3
gions (FEDER) for partial funding of this work and for sup-
porting CERLA. We also thank Prof. R. P. Wayne (University
of Oxford, England) for the communication of his results
about the rate constant of the reaction CF O ] Cl. We are
As indicated above, the accepted experimental formation
enthalpy for CF O is implicitly referenced to the formation
3
2
enthalpy for CF O (through reaction enthalpies and other
3
2
measured values). Batt and Walsh57 recommended * H¡ \
also grateful to Dr Vladimir Orkin (NIST) for the improve-
ment of the manuscript and fruitful discussions.
f
[
655.6 ^ 6.0 kJ mol~1 based on the formation enthalpy for
CF OOCF and assigning the bond dissociation energy
3
3
D(CF OÈOCF ) \ 195.8 kJ mol~1 based on their evaluation
Appendix
3
3
of kinetic measurements of the CF OOCF decomposition
3 3
rate. The formation enthalpy for CF OOCF is based on
We deÐne here the following notation to describe the compos-
ite energies. MP4/TZ ] P is the MP4(SDTQ)/6-311 ] G(d,p)
energy. QCI/(TZ ] P) is the MP4/TZ ] P energy plus an
approximate QCI level correction [QCISD(T)/6-311G(d,p)È
MP4(SDTQ)/6-311G(d,p)]. MP4/(CbsB3) is the MP2/CbsB3
energy plus an approximate MP4 level correction
3
3
measurements of its equilibrium with CF O ] CF OF, where
2
3
3
the reaction enthalpy * H¡(CF OOCF ) \ 102.5 kJ mol~1
r
3
was determined.64 The accepted value for the formation enth-
alpy of CF OF is, in turn, based on measurements of its equi-
3
librium with CF O ] F , where the reaction enthalpy * H¡
2
2
r
[
MP4(SDQ)/CbsB4ÈMP2/CbsB4)]. MP4/(CbsB3 ] CBS) is
(CF OF) \ 125.9 kJ mol~1 was determined.42 This set of data
3
the MP4/(CbsB3) plus the approximate CBS correction.
leads to the following relationship:
CF O
* H¡ (CF O) \ * H¡ (CF O) ] 1[D(CF O [ OCF )
2
f
3
f
2
2
3
3
As indicated previously in this paper, the ““recommended
experimentalÏÏ formation enthalpy at 298 K for CF O is
[ * H¡ (CF OOCF ) [ * H¡ (CF OF)]
r
3
3
r
3
2
\ * H¡ (CF O) [ 16.3 kJ mol~1
[
641.0 ^ 5.0 kJ mol~1,28 while the ““recommended ab initioÏÏ
f
2
value is [608.0 ^ 7.0 kJ mol~1.29,30 This ab initio value is
based on a set of calculations employing corrections to the ab
initio energies utilizing reference molecules and isodesmic
reactions. We have repeated these calculations and found
Using this relationship, * H¡ (CF O) \ [624.2 kJ mol~1 or
f
3
* H¡ (CF O) \ [619.1 kJ mol~1 may be suggested, depend-
f
3
ing on whether one uses the previously recommended ab initio
value from the literature or the G2/iso(CH F ) value from this
2
2
*
H¡ \ [622.2 kJ mol~1 and * H¡ \ [618.5 kJ mol~1 for
work for * H¡ (CF O). However, we will select yet another
f
f
f
2
CF O using the ab initio G2 and CBS-Q methods, respec-
value, as detailed below, in attempt to provide self-consistent
2
tively. We note that we have employed zero point energy cor-
enthalpies for the CF O ] Cl reaction system.
3
2
Phys. Chem. Chem. Phys., 1999, 1, 5087È5096
5093

View Article Online
The formation enthalpy for CF O can also be determined
from the G2 and CBS-Q ab initio energies employing correc-
tions derived from isodesmic reactions and reference mol-
CF O and CH O, where there exist experimentally derived
enthalpies of formation, using the following isodesmic reac-
tions:
3
3
3
ecules. We calculate values of * H¡ \ [647.3 kJ mol~1 and
f
CF O ] CH ] CH O ] CHF
*
H¡ \ [640.4 kJ mol~1 for CF O using the ab initio G2
3 2
4
3 2
3
f
3
and CBS-Q methods, respectively. We note that we have
CF O ] H O ] CF O ] H O
3
2
2
3
2 2
employed ZPEÏs and thermal corrections for CF O that are
3
CF O ] H O ] CH O ] H O ] (CHF [ CH )
based on scaled (0.8929) HF/6-31G(d,p) vibrational fre-
3
2
2
3
2 2
3
4
quencies. Using the following isodesmic reactions to correct
the ab initio energies,
CF O ] C H ] CH O ] CH CF
3
2
2 6 3 2 3
3
We calculate G2/isodesmic formation enthalpies as [639.1,
CF O ] CH ] CH O ] CHF
[
639.6, [639.5 and [631.4 kJ mol~1, respectively. Using
3
4
3
3
the Ðrst reaction (CHF reference), we Ðnd isodesmic reaction-
CF O ] CH ] CH O ] CF
3
3
3
3
3
corrected formation enthalpies of [633.8, [631.8, [640.5,
CF O ] CH O ] CH O ] CF O ] (CHF [ CH F )
[637.4 and [639.1 kJ mol~1 using MP4/TZ ] P, QCI/
3
2
3
2
3
2 2
(
TZ ] P), G1, G2(MP2) and G2 energies, respectively. Simi-
CF O ] C H ] CH O ] CH CF
3
2 6
3
3
3
larly, we Ðnd [637.0, [635.6 and [636.7 kJ mol~1 using
MP4/(CbsB3), MP4/(CbsB3 ] CBS) and CBS-Q energies,
respectively.
we calculate G2/isodesmic formation enthalpies of [629.8,
[
635.0, [629.8 and [622.1 kJ mol~1. For the reference mol-
ecules, ZPEÏs and thermal corrections were used based on
experimental vibrational frequencies. Using the Ðrst reaction
DeÐning G2 as the base calculation and using the CHF3
isodesmic reaction as the most elementary correction, we
select the G2/iso(CHF ) value for CF O of * H¡ \ [639.1
(
CHF reference) to gauge uncertainties in the ab initio calcu-
3
3
3 2
f
lations, we Ðnd isodesmic reaction-corrected formation enth-
alpies of [627.7, [625.4, [629.6, [628.6 and [629.8 kJ
mol~1 using MP4/TZ ] P, QCI/(TZ ] P), G1, G2(MP2) and
G2 energies, respectively. Similarly, we Ðnd [625.5, [626.1
and [625.1 kJ mol~1 using MP4/(CbsB3), MP4/
^ 12.0 kJ mol~1 for use in this work. We assigned an uncer-
tainty of 12 kJ mol~1 as the sum of an estimated 6 kJ mol~1
for the ab initio calculations and a contribution of 6 kJ mol~1
from the reported uncertainties in the reference molecules (see
Table 5). In the CF O ] Cl reaction system, the relative
3 2
enthalpies of CF O and CF O are important. We note that
(
CbsB3 ] CBS) and CBS-Q energies, respectively. Again as
3 2 3
the Ðrst two isodesmic reactions reference CF O to signiÐ-
above, deÐning G2 as the base calculation and using the
3 2
cantly di†erent species, CH O and CF O, respectively, yet
CHF isodesmic reaction as the most elementary correction,
3
3 2
3
we select the G2/iso(CHF ) value for CF O of * H¡ \
give roughly the same corrected enthalpy. This suggests that
3
3
f
[
629.8 ^ 12.0 kJ mol~1 for use in this work. We note that
this value is slightly lower than, but not inconsistent with the
the decision to utilize a formation enthalpy for CF O based
3
on G2/iso(CHF ) was correct and provides the best self-
3
[
619 to [624 kJ mol~1 that we derived above based on an
consistent values for this system. We note that isodesmic
isodesmic-corrected ab initio value for CF O and experimen-
reaction-corrected ab initio values for * H¡ (CF O ) are
2
f
3 2
tal reaction enthalpies. For * H¡ (CF O), we assigned an
roughly 35È40 kJ mol~1 higher than the Batt and Walsh58
recommendation. However, as indicated previously, all mol-
ecules in this system are implicitly referenced to * H¡(CF O),
f
3
uncertainty of 12 kJ mol~1 as the sum of an estimated 4 kJ
mol~1 for the ab initio calculations and a contribution of 8 kJ
mol~1 from the reported uncertainties in the reference mol-
ecules (see Table 5).
f
2
which may be 30È35 kJ mol~1 higher than literature values
based on experimental measurements alone.
We note that this value for CF O is about 26 kJ mol~1
3
1CF O and 3CF O
2 2
higher than the recommended value kJ mol~1 di†erence (calcÈ
2 2
expt) for CF O. These values should be roughly the same
As indicated earlier, the isomer used here is diÑuorocarbony-
loxide É CF OO É, rather than the cyclic diÑuorodioxirane
2
since the enthalpy of formation of CF O is tied directly to
3
2
that for CF O (see above where we found * H¡ (CF O) \
È(OCF O)È. Although the latter is more stable by about 180
2
f
3
2
*
H¡ (CF O) [ 16.3 kJ mol~1 based on experimental enth-
kJ mol~1,63 no direct reaction path to the latter was found for
f
2
alpies of reaction). We do not attempt to resolve this 11 kJ
the CF OO ] Cl reaction. G2(MP2) and CBS-Q calculations
3
mol~1 uncertainty here. However, we note that this relation-
for 1CF O yield values of * H¡ \ [303.3 and [289.4 kJ
2
2
f
ship was based on
a
bond dissociation energy for
mol~1, respectively. These ab initio data can be corrected by
CF OÈOCF estimated from kinetic measurements, which
referencing to CH O and CH F using the following iso-
3
3
3 2
2 2
could easily be in error by 5È10 kJ mol~1.
desmic reaction:
CF O ] CH ] CH O ] CH F ] (CH [ CH )
CF O2
3
2 2
4
3 2
2 2
3
4
There is no experimental value for the enthalpy of formation
for CF O . The currently accepted literature value for this
We report corrected values for * H¡ of [288.8, [265.5,
f
[270.2 and [281.5 kJ mol~1 using G2(MP2), MP4/(CbsB3),
3
2
molecule, * H¡ \ [674.9 kJ mol~1, is based on * H¡(CF O),
MP4/(CbsB3 ] CBS) and CBS-Q energies. Selecting the
G2(MP2)/iso value for 1CF O , we adopt * H¡ \ [288.8
f
f
3
group additivity and estimated bond strengths by Batt and
Walsh.58 Francisco and Williams65 and Schneider and
Wallington66 have performed semi-empirical and ab initio cal-
culations in order to estimate the heat of formation of CF O
and obtained [656.9 kJ mol~1 and [618.4 kJ mol~1, respec-
tively. The available experimental values67,68 based upon
kinetic studies of the dissociation reaction CF O ] CF
O are consistent with the higher value.
Our G2 and CBS-Q calculations on CF O yield ab initio
values of * H¡ \ [656.5 and [652.3 kJ mol~1, respectively.
ZPEÏs and thermal corrections were calculated from scaled
2
2
f
^ 14.0 kJ mol~1 (assuming 8 and 6 kJ mol~1 for the ab initio
and experimental contributions to the uncertainties,
respectively). G2(MP2) and CBS-Q calculations for the triplet
state of this biradical, 3CF O , yield values of * H¡ \ [188.0
3
2
2
2
f
and [182.1 kJ.mol~1, respectively. Note that this excited
state is about 110È120 kJ mol~1 above the ground state.
3
2
3
]
Assuming the same correction to the enthalpy as for 1CF O ,
2
2 2
we adopt a G2(MP2)/iso corrected value of * H¡ \ [173.4
3
2
f
^ 14.0 kJ mol~1 for 3CF O .
f
2 2
CF OOCl and CF OClO
HF/6-31G(d,p) vibrational frequencies for CF O and using
3
3
3
2
experimental values for reference molecules. The ab initio data
G2(MP2) and CBS-Q calculations for CF OOCl yield values
3
for * H¡ can be corrected by referencing to either CH O ,
of * H¡ \ [699.0 and [684.3 kJ mol~1, respectively. These
f
3 2
f
5094
Phys. Chem. Chem. Phys., 1999, 1, 5087È5096

View Article Online
ab initio data can be corrected by referencing to CF O
value above) using the following isodesmic reaction:
2 (see
Table 6 BAC-MP4 heats of formation (kJ mol~1) at 298 K as calcu-
3
lated in this work
Species
* H¡
CF OOCl ] H O ] CF OO ] HOCl ] (H O [ HO )
f
3
2
3
2 2
2
We report corrected values for * H¡ of [666.2, [660.1,
CF O
2
[602.8
[628.4
[627.2
[236.1
[166.1
51.9
[656.1
[593.3
[443.5
[192.5
f
CF O
3
[
658.7 and [663.7 kJ mol~1 using G2(MP2), MP4/(CbsB3),
CF O
MP4/(CbsB3 ] CBS) and CBS-Q energies. Selecting the
G2(MP2)/iso value for CF OOCl, we adopt * H¡ \ [666.2
3 2
CF O
1
2
CF O
2
2
2
3
3
f
^
12.0 kJ mol~1 (using an estimated uncertainty from
FOCl
CF O ). G2(MP2) and CBS-Q calculations for CF OClO
CF OOCl
3
2
3
3
yield values of * H¡ \ [660.4 and [640.5 kJ mol~1, respec-
CF OClO
f
3
TS CF OÉ É ÉFOCl
tively. Note that this isomer is about 40È45 kJ mol~1 less
2
TS 1CF OOÉ É ÉFCl
stable than CF OOCl. Assuming the same correction to the
2
3
enthalpy as for CF OOCl, we adopt a G2(MP2)/iso corrected
3
value of * H¡ \ [627.6 ^ 15.0 kJ mol~1 for CF OClO. A
slightly higher uncertainty was chosen to reÑect the lack of a
suitable isodesmic reaction scheme.
f
3
BAC-MP4 calculations
The heats of formation proposed and discussed previously can
be compared to BAC-MP4 values that we have also calcu-
lated and which are presented in Table 6. They agree within
FOCl
G2 and CBS-Q calculations for FOCl yield values of * H¡ \
f
1
0 kJ mol~1, except for the1 CF O singlet radical (and the
4
9.8 and 48.5 kJ mol~1, respectively. These ab initio data can
2
2
associated 1CF OOÉ É ÉFCl transition state) for which the
be corrected by referencing to HOCl and HOF using the fol-
lowing isodesmic reaction:
2
BAC-MP4 energy is higher by 53 kJ mol~1 than the proposed
calculated values in spite of the empirical correction in the
BAC-MP4 values which is usually done to account for the
UHF instability present in these singlet species.
FOCl ] H O ] HOCl ] HOF
2
Using this reaction, we report corrected values for * H¡ of
f
4
0.2, 41.8, 42.6, 44.0 and 43.5 kJ mol~1 using MP4/TZ ] P,
QCI/(TZ ] P),G1, G2(MP2) and G2 energies, respectively.
Similarly, we Ðnd 49.5, 52.6 and 47.6 kJ mol~1 using MP4/
References
1
P. Biggs, C. E. Canosa-Mas, D. E. Shallcross and R. P. Wayne,
C. Kelly and H. W. Sidebottom, STEP-HALOCSIDE/AFEAS
Workshop, 23È25 March 1993, Dublin, Ireland.
M. K. W. Ho, N.-D. Sze, J. M. Rodriguez, D. K. Weisenstein,
C. W. Heisey, R. P. Wayne, P. Biggs, C. E. Canosa-Mas, H. W.
Sidebottom and J. Treacy, Geophys. Res. L ett., 1994, 21, 101.
O. J. Nielsen and J. Sehested, Chem. Phys. L ett., 1993, 213, 433.
T. J. Wallington, M. D. Hurley and W. F. Schneider, Chem. Phys.
L ett., 1993, 213, 442.
(
CbsB3), MP4/(CbsB3 ] CBS) and CBS-Q energies, respec-
tively. We adopt * H¡ \ 43.5 ^ 10.0 kJ mol~1 (assuming 4
f
and 6 kJ mol~1 for the ab initio and experimental contribu-
2
tions to the uncertainties, respectively).
3
4
CF OÆ Æ ÆFOCl transition state
2
5
6
M. M. Maricq and J. J. Szente, Chem. Phys. L ett., 1993, 213, 449.
A. A. Turnipseed, S. B. Barone and A. R. Ravishankara, J. Phys.
Chem., 1994, 98, 4594.
G2(MP2) and CBS-Q calculations for the transition state
CF OÉ É ÉFOCl yield values of * H¡ \ [465.3 and [444.4 kJ
2
f
mol~1, respectively. These ab initio data can be corrected by
7
8
9
Ch. Fockenberg, H. Saatho† and R. Zellner, Chem. Phys. L ett.,
1994, 218, 21.
referencing to the products CF O and FOCl. We report cor-
2
rected values for * H¡ of [444.0, [396.2, [441.5 and
C. Bourbon, M. Brioukov and P. Devolder, C. R. Acad. Sci.
Paris, 1995, 322, 181.
f
[
433.8 kJ mol~1 using G2(MP2), MP4/(CbsB3), MP4/
P. Biggs, C. E. Canosa-Mas, D. E. Shallcross, A. Vipond and
R. P. Wayne, J. Chem. Soc., Faraday T rans., 1997, 93, 2477.
(
CbsB3 ] CBS) and CBS-Q energies. Selecting the G2(MP2)/
iso value for CF OÉ É ÉFOCl, we adopt * H¡ \ [440.0 ^ 14.0
2
f
10 P. Biggs, C. E. Canosa-Mas, D. E. Shallcross, A. Vipond and
R. P. Wayne, J. Chem. Soc., Faraday T rans., 1997, 93, 2701.
11 C. E. Canosa-Mas, A. Vipond and R. P. Wayne, Phys. Chem.
Chem. Phys., 1999, 1, 761.
kJ mol~1 (assuming 8 and 6 kJ mol~1 for the ab initio and
experimental contributions to the uncertainties, respectively).
We note that this transition state is above the reactants
CF O ] Cl by about 80 kJ mol~1, above the products CF O
1
2
J.-P Sawerysyn, C. Lafage, B. Me
Phys., 1987, 84, 1187.
F. Louis, PhD Thesis, University of Lille, 1997.
 riaux and A. Tighezza, J. Chim.
3
2
2
]
FOCl by about 120 kJ mol~1, and above the intermediate
1
3
CF OClO by about 190 kJ mol~1.
3
14 M. A. A. Clyne, D. J. McKenney and R. F. Walker, Can. J.
Chem., 1973, 51, 3596.
1
5
6
J. Y. Park, I. R. Slage and D. Gutman, J. Phys. Chem., 1987, 84,
1187.
CF OOÆ Æ ÆFCl transition state
2
1
P. P. Bemand, M. A. A. Clyne and R. T. Watson, J. Chem. Soc.
G2(MP2) calculations for the transition state CF OOÉ É ÉFCl
Faraday T rans., 1973, 69, 1356.
2
yield a value of * H¡ \ [276.9 kJ mol~1. This ab initio data
f
17 I. C. Plumb and K. R. Ryan, Chem. Phys. L ett., 1982, 92, 236.
1
8 A. M. Dognon, F. Caralp and R. Lesclaux, J. Chim. Phys., 1985,
can be corrected by referencing to the products CF O and
2
2
82, 349.
FCl. We report a corrected value of * H¡ \ [250.6 ^ 14.0 kJ
f
19 J. Peeters, J. Vertommen and I. Langhans, Ber. Bunsen-Ges. Phys.
Chem., 1992, 96, 431.
mol~1 using the G2(MP2) energy. CBS-Q calculations for
CF OOÉ É ÉFCl were not obtained, nor both G2(MP2) and
2
0
F. Caralp, R. Lesclaux and A. M. Dognon, Chem. Phys. L ett.,
1986, 129, 433.
2
CBS-Q for the triplet state of CF OOÉ É ÉFCl. In these cases,
2
signiÐcant difficulties were encountered either in Ðnding stable
21 P. Biggs, C. E. Canosa-Mas, J.-M. Fracheboud, C. J. Percival,
R. P. Wayne and D. E. Shallcross, J. Chem. Soc., Faraday T rans.,
transition state structures (both HF and MP2 levels) and/or
convergence of computed energies. Since the triplet state of
CF O is 110È120 kJ mol~1 above the singlet ground state
1
997, 93, 379.
2
2
O. J. Nielsen, T. Ellermann, J. Sehested, E. Bartkiewicz, T. J.
2
2
Wallington and M. D. Hurley, Int. J. Chem. Kinet., 1992, 24,
and, consequently, the products 3CF O ] FCl would be
1
009.
2
2
nearly 300 kJ mol~1 above the reactants CF O ] Cl, we can
23 C. Bourbon, M. Brioukov, B. Hanoune, J.-P. Sawerysyn and P.
3
2
readily ignore the possibility of this channel.
Devolder, Chem. Phys. L ett., 1996, 254, 203.
Phys. Chem. Chem. Phys., 1999, 1, 5087È5096
5095

View Article Online
2
4
T. J. Bevilacqua, D. R. Hanson and C. J. Howard, J. Phys. Chem.,
993, 97, 3750.
44 ** H¡ calculated using bond distances, bond angles and vibra-
f
1
tional frequencies given in references for this molecule in Table 5.
2
2
2
5
6
7
J. Sehested and O. J. Nielsen, Chem. Phys. L ett., 1993, 206, 369.
A. Bhatnagar and R. W. Carr, Chem. Phys. L ett., 1994, 231, 454.
T. P. W. Jungkamp, A. Kukui and R. N. Schindler, Ber. Bunsen-
Ges. Phys. Chem., 1995, 99, 1057.
L. V. Gurvich, I. V. Veyts and C. B. Alcock, T hermodynamic
Properties of Individual Substances, V olume 2, 4th edn., Hemi-
sphere Publishing, NY (1991).
45 L. V. Gurvich, I. V. Veyts and C. B. Alcock, T hermodynamic
Properties of Individual Substances, V olume 1, 4th edn., Hemi-
sphere Publishing, NY, 1989.
46 A. J. Hills and C. J. Howard, J. Chem. Phys., 1984, 81, 4458.
47 T. Shimanouchi, Molecular V ibrational Frequencies, NSRDS-
NBS 39, 1972.
2
8
48 J. Berkowitz, G. B. Ellison and D. Gutman, J. Phys. Chem., 1994,
2
9
W. F. Schneider and T. J. Wallington, J. Phys. Chem., 1994, 98,
98, 2744.
7
448.
49 B. Ruscic and J. Berkowitz, J. Chem. Phys., 1991, 95, 4033.
50 Vibrational frequencies taken from ref. 41. Only 9 modes are
given in this reference. These data were supplemented with scaled
(0.8929)HF/6-31G(d) torsion and CH stretch frequencies of 161,
2896 and 2987 cm~1.
3
3
0
1
D. A. Dixon and D. Feller, J. Phys. Chem. A, 1998, 102, 8209.
T. J. Buckley, R. D. Johnson and R. E. Huie, J. Phys. Chem.,
1
995, 99, 4879.
V. D. Knyazev, A. Bencsura and I. R. Slagle, J. Phys. Chem. A,
997, 101, 849.
F. Helleis, J. N. Crowley and G. K. Moorgat, J. Phys. Chem.,
993, 97, 11464.
V. Dae
3
2
3
1
51 ** H¡ calculated using bond distances, bond angles and vibra-
f
3
tional frequencies given as electronic supplementary information.
1
52 V. D. Knyazev and I. R. Slagle, J. Phys. Chem., 1998, 102, 1770.
53 A. S. Rodgers, J. Chao, R. C. Wilhoit and B. J. Zwolinski, J.
Phys. Chem. Ref. Data, 1974, 3, 117.
54 S. S. Chen, A. S. Rodgers, J. Chao, R. C. Wilhoit and B. J. Zwol-
inski, J. Phys. Chem. Ref. Data, 1975, 4, 441.
55 HF/6-31G(d) vibrational frequencies scaled by 0.8929.
3
3
4
5
Ž le and G. Poulet, J. Chim. Phys., 1996, 93, 10841.
J. L. Jourdain, G. Poulet, J. Barassin, G. Lebras and J. Com-
bourieu, Pollut. Atmos., 1977, 75, 256.
3
6
W. B. DeMore, S. P. Sander, D. M. Golden, R. F. Hampson,
M. J. Kurylo, C. J. Howard, A. R. Ravishankara, C. E. Kolb and
M. J. Molina, in Chemical Kinetics and Photochemical Data for
Use in Stratospheric Modeling, evaluation number 11, publication
56 * H¡ calculated from ab initio G2 energy plus a correction based
f
on an isodesmic reaction and reference molecules. See text for
9
7-4, Jet Propulsion Laboratory, Pasadena, California, 1997.
further discussion.
3
7
GAUSSIAN 94 (Revision E.2), M. J. Frisch, G. W. Trucks, H. B.
Schlegel, P. M. W. Gill, B. G. Johnson, M. A. Robb, J. R. Cheese-
man, T. A. Keith, G. A. Petersson, J. A. Montgomery, K. Ragha-
vachari, M. A. Al-Laham, V. G. Zakrzewski, J. V. Ortiz, J. B.
Foresman, J. Cioslowski, B. B. Stefanov, A. Nanayakkara, M.
Challacombe, C. Y. Peng, P. Y. Ayala, W. Chen, M. W. Wong,
J. L. Andres, E. S. Replogle, R. Gomperts, R. L. Martin, D. J.
Fox, J. S. Binkley, D. F. Defrees, J. Baker, J. P. Stewart, M.
Head-Gordon, C. Gonzalez and J. A. Pople, Gaussian, Inc., Pitts-
burgh PA, 1995.
57 L. Batt and R. Walsh, Int. J. Chem. Kinet., 1982, 14, 933.
58 L. Batt and R. Walsh, Int. J. Chem. Kinet., 1983, 15, 605.
59 * H¡ calculated from ab initio G2(MP2) energy plus a correction
f
based on an isodesmic reaction and reference molecules. See text
for further discussion.
60 M. A. A. Clyne and R. T. Watson, J. Chem. Soc. Faraday T rans.,
1977, 73, 1169.
61 S. Baer, H. Hippler, R. Rahn, M. Siefke, N. Seitzinger and J.
Troe, J. Chem. Phys., 1991, 95, 6463.
62 ZTE and ** H¡ calculated from HF/6-31G(d) geometries and
f
3
3
4
4
4
8
9
0
1
2
L. A. Curtis, K. Raghavachari, G. W. Trucks and J. A. Pople, J.
scaled (0.8929) frequencies. Bond distances: r(OÈF) \ 1.452 Ó,
Chem. Phys., 1991, 94, 7221.
r(OÈCl) \ 1.704 Ó, bond angle: h(FOCl) \ 105.2¡, vibrational fre-
quencies: l \ 393, 771, 1027 cm~1.
L. A. Curtis, K. Raghavachari and J. A. Pople, J. Chem. Phys.,
1
993, 98, 1293.
63 E. Kraka, Z. Konkoli, D. Cremer, J. Fowler and H. F. Schaefer
III, J. Am. Chem. Soc., 1996, 118, 10595.
J. W. Ochterski, G. A. Petersson and J. A. Montgomery, J. Chem.
Phys., 1996, 104, 2598; and references therein.
64 J. B. Levy and R. C. Kennedy, J. Am. Chem. Soc., 1972, 94, 3302.
65 J. S. Francisco and I. H. Williams, Int. J. Chem. Kinet., 1988, 20,
455.
C. F. Melius, in Chemistry and Physics of Energetic Materials, ed.
S. N. Bulusu, Kluwer Academic publishers, Dordrecht, 1990.
M. W. Chase, Jr., C. A. Davies, J. R. Downey, Jr., D. J. Frurip,
R. A. McDonald and A. N. Syverud, ““JANAF Thermochemical
Tables, Third Edition,ÏÏ J. Phys. Chem. Ref. Data 14 Suppl. 1,
66 W. F. Schneider and T. J. Wallington, J. Phys. Chem., 1993, 97,
12783.
67 S. Antonik, Bull. Soc. Chim. Fr., 1983, 117.
68 M. A. TeitelÏboim, M. Ya. GoldÏenberg and V. I. Vedeneev, Oxid.
Commun., 1983, 3, 315.
1
985.
4
3
M. E. Jacox, in NIST Chemistry W ebBook, NIST Standard Ref-
erence Database 69, ed. W. G. Mallard and P. J. Linstrom,
National Institute of Standards and Technology, Gaithersburg,
MD, 1998.
Paper 9/05933D
5096
Phys. Chem. Chem. Phys., 1999, 1, 5087È5096