Formation of CF3O- in the gas phase

Article abstract of DOI:10.1063/1.478753

We report experimental studies of the formation of CF₃O^- by ion-molecule and electron attachment reactions, and theoretical investigations of the structure and energetics of CF₃O^- and its neutral counterpart CF₃O. The anion CF₃O^- is formed from the rapid attachment of free electrons to its neutral dimer, (CF₃O)₂. Potential sources of CF₃O^- through ion-molecule reactions of CF₃^- and F^- were surveyed. CF₃O^- is formed in the bimolecular ion-molecule reaction of CF₃^- with SO₂ and the third-order association reaction of F^- with CF₂O. In addition, rate constants for the reactions of CF₃^- with a variety of neutral compounds were measured. A number of cases were found in which formation of CF₃O^- was energetically allowed but was not observed. The potential energy surfaces of CF₃O and CF₃O^- have been investigated using a variety of density functional theory (DFT) techniques. The ground-state minimum energy structure of CF₃O was found to be a ²A′ Jahn-Teller distorted C_s-symmetry structure, while for the anion the ground state is ¹A₁ with a C_3v-symmetry minimum. A search for other low-energy minima for CF₃O^- was unsuccessful. The DFT methods support a value for the adiabatic electron affinity of CF₃O near 4.1 eV.

Full text of DOI:10.1063/1.478753

Formation of CF 3 O − in the gas phase
Robert A. Morris, Thomas M. Miller, John F. Paulson, A. A. Viggiano, Michael T. Feldmann, Rollin A. King, and
Henry F. Schaefer III
Citation: The Journal of Chemical Physics 110, 8436 (1999); doi: 10.1063/1.478753
View online: http://dx.doi.org/10.1063/1.478753
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/110/17?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Experimental and ab initio studies of the reactive processes in gas phase i-C3H7Br and i-C3H7OH collisions with
potassium ions
J. Chem. Phys. 141, 164310 (2014); 10.1063/1.4898377
Experimental study of the reactive processes in the gas phase K+ + i-C3H7Cl collisions: A comparison with Li
and Na ions
J. Chem. Phys. 138, 184310 (2013); 10.1063/1.4804188
Observation of dihalide elimination upon electron attachment to oxalyl chloride and oxalyl bromide, 300 – 550 K
J. Chem. Phys. 124, 184313 (2006); 10.1063/1.2196409
Trajectory studies of S N 2 nucleophilic substitution. IX. Microscopic reaction pathways and kinetics for Cl − +CH
3 Br
J. Chem. Phys. 118, 2688 (2003); 10.1063/1.1535890
Reaction on the O − +CH 4 potential energy surface: Dependence on translational and internal energy and on
isotopic composition, 93–1313 K
J. Chem. Phys. 106, 8455 (1997); 10.1063/1.473904
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.42.202.150 On: Fri, 21 Nov 2014 20:10:00

JOURNAL OF CHEMICAL PHYSICS
VOLUME 110, NUMBER 17
1 MAY 1999
Formation of CF3O؊ in the gas phase
Robert A. Morris,a) Thomas M. Miller,b) John F. Paulson, and A. A. Viggiano
Air Force Research Laboratory, Space Vehicles Directorate (AFRL/VSBP), 29 Randolph Road,
Hanscom Air Force Base, Massachusetts 01731-3010
Michael T. Feldmann, Rollin A. King, and Henry F. Schaefer III
Center for Computational Quantum Chemistry, Department of Chemistry, University of Georgia, Athens,
Georgia 30602-2525
͑Received 16 November 1998; accepted 2 February 1999͒
We report experimental studies of the formation of CF3OϪ by ion-molecule and electron attachment
reactions, and theoretical investigations of the structure and energetics of CF3OϪ and its neutral
counterpart CF3O. The anion CF3OϪ is formed from the rapid attachment of free electrons to its
neutral dimer, (CF3O)2 . Potential sources of CF3OϪ through ion-molecule reactions of CF3Ϫ and FϪ
were surveyed. CF3OϪ is formed in the bimolecular ion-molecule reaction of CF3Ϫ with SO2 and the
third-order association reaction of FϪ with CF2O. In addition, rate constants for the reactions of CF3Ϫ
with a variety of neutral compounds were measured. A number of cases were found in which
formation of CF3OϪ was energetically allowed but was not observed. The potential energy surfaces
of CF3O and CF3OϪ have been investigated using a variety of density functional theory ͑DFT͒
techniques. The ground-state minimum energy structure of CF O was found to be a 2A Jahn–Teller
Ј
3
1
distorted Cs-symmetry structure, while for the anion the ground state is A1 with a C3v-symmetry
minimum. A search for other low-energy minima for CF3OϪ was unsuccessful. The DFT methods
support a value for the adiabatic electron affinity of CF3O near 4.1 eV. © 1999 American Institute
of Physics. ͓S0021-9606͑99͒01117-4͔
I. INTRODUCTION
II. EXPERIMENT
A. Selected ion flow tube
The CF3OϪ anion has been used as a chemical ionization
agent for the detection of trace gas phase species in the at-
mosphere. It reacts with a variety of compounds by fluoride
transfer, and Huey et al.1 have reported laboratory kinetics
results for such chemical ionization reactions of CF3OϪ with
trace gases of interest in atmospheric chemistry. Arijs and
Kopp2 have used CF3OϪ as a chemical ionization agent in
the stratosphere for detection of nitric acid. Thus there is
practical interest in determining efficient and convenient
means of forming the CF3OϪ anion and for understanding
the energetics of CF3OϪ and related species. The neutral
CF3O radical has been suggested to play a role in catalytic
ozone destruction cycles in the stratosphere.3,4
There is also fundamental interest in the ability of com-
putational chemistry techniques to faithfully describe the
structure and energetics of anions, their neutral counterparts,
and the corresponding potential energy surfaces. In particu-
lar, the computation of electron affinities and the structural
changes involved when an electron is added to a molecule is
an important capability in chemical physics. Recently den-
sity functional theory has been used to investigate these
questions for a growing number of systems and has provided
results which are generally in good agreement with
experiment.5,6
Gas phase kinetics measurements of ion-molecule reac-
tions were made at 300 K under thermal conditions using the
selected ion flow tube ͑SIFT͒ instrument at the Air Force
Research Laboratory ͑AFRL͒. The apparatus and technique
are standard and have been described previously.7 The CF3Ϫ
and FϪ reactant ions were formed in a remote electron im-
pact ion source from C3F6O ͑hexafluoropropene oxide͒ and
SF6, respectively. The reactant ion of interest was mass se-
lected in a quadrupole mass filter and injected through a
Venturi inlet into a fast flow of He buffer gas in a 1-m-long
stainless steel flow tube. Nitrogen buffer was used in the
experiments on the association reaction of FϪ with CF2O.
The ions were transported by the buffer past an inlet through
which the reactant neutral was added. The reactant and prod-
uct ions were sampled through an aperture by a second quad-
rupole mass spectrometer and detected by pulse counting in a
particle multiplier. Reaction rate constants were calculated
from the pseudo-first-order attenuation in reactant ion signal
as a function of reactant neutral concentration in combina-
tion with the measured reaction time. Neutral reactant com-
pounds were obtained commercially and used without further
purification. For the reaction of FϪ with CF2O,
FϪϩCF2OϩN2→CF3OϪϩN2 ⌬HϭϪ41.4 kcal molϪ1
͑1͒
rate constants were measured as a function of pressure ͑in an
N2 buffer͒ over the range 0.22–0.60 Torr by throttling a
valve in the main flow tube pumping line. All of the reported
rate constants are accurate to within Ϯ30%.
a͒
Author to whom correspondence should be addressed; Electronic mail:
morris@plh.af.mil
b͒
Under contract F19628-95-C-0020 to Visidyne, Inc., Burlington, MA.
0021-9606/99/110(17)/8436/7/$15.00
8436
© 1999 American Institute of Physics
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.42.202.150 On: Fri, 21 Nov 2014 20:10:00

J. Chem. Phys., Vol. 110, No. 17, 1 May 1999
Morris et al.
8437
TABLE I. Second-order rate constants (k2) for reactions of CFϪ3 with se-
lected neutrals measured in ϳ0.4 Torr helium gas. The estimated accuracy is
Ϯ30%.
Reaction
⌬H ͑kJ molϪ1
͒
k2 ͑cm3 sϪ1
͒
CFϪ3 ϩSO2→CF3OϪϩSO
CFϪ3 ϩNO2→NOϪ2 ϩCF3
CFϪ3 ϩCO2→CF3COϪ2
Ϫ112
Ϫ189
Ϫ168
Ϫ182
Ϫ224
Ϫ246
Ϫ276
Ϫ213
Ϫ57
?
7.5ϫ10Ϫ10
4.2ϫ10Ϫ10
1ϫ10Ϫ11
1.1ϫ10Ϫ9
4.3ϫ10Ϫ10
6.5ϫ10Ϫ11
4.5ϫ10Ϫ10
1.6ϫ10Ϫ11
6.1ϫ10Ϫ10
5.0ϫ10Ϫ10
CFϪ3 ϩHCl→ClϪϩCF3H
CFϪ3 ϩHBr→BrϪϩCF3H
CFϪ3 ϩCH3Cl→ClϪϩCF3CH3
CFϪ3 ϩCH3Br→BrϪϩCF3CH3
CFϪ3 ϩCF3Br→BrϪϩC2F6
CFϪ3 ϩSiF4→SiF5ϪϩCF2
CFϪ3 ϩOCS→CF3SϪϩCO(15%)
→CF3OCSϪ(85%)
?
FIG. 1. Second-order rate constants for the association reaction of FϪ with
CF2O plotted as a function of N2 buffer gas pressure. The error bars repre-
sent estimated relative error limits.
B. Flowing afterglow Langmuir probe
Measurements of rate constants for attachment of free
electrons to (CF3O)2 ,
approaching a limiting value. Passivation required nearly an
hour of flowing the 1% (CF3O)2 mixture, or a few minutes
of flowing neat SF4 gas.
eϩ͑CF3O͒2→CF3OϪϩCF3O ⌬HϭϪ54.7 kcal molϪ1
͑2͒
were performed in a variable-temperature flowing afterglow
Langmuir probe ͑FALP͒ apparatus8 at AFRL. The reaction
enthalpy was calculated from the dimer bond strength ͑46.8
kcal molϪ1͒9 and the electron affinity of CF3O ͑present cal-
culations͒. The FALP technique10 has been well described in
the literature. The decay in electron density down the axis of
a flow tube due to the coupled effects of electron attachment
and ambipolar diffusion is measured using the FALP
method. This information, together with measurements of the
plasma velocity and reactant gas concentration, allows us to
deduce the electron attachment rate constant. Mass spectro-
metric sampling at the downstream end of the flow tube
showed that the only product ion was CF3OϪ over the entire
temperature range 297–553 K, in 1 Torr helium gas. The
(CF3O)2 was synthesized and provided by Professor Darryl
DesMarteau. Infrared ͑IR͒ analysis of the (CF3O)2 sample at
Clemson University showed that it contained a small amount
of COF2 and possibly CO2 ͑neither of which attach electrons
in our temperature range͒. The (CF3O)2 sample was de-
gassed at 77 K prior to use. Mixtures of (CF3O)2 in He gas
were prepared at fractions between 0.6 and 1.0% to facilitate
accurate flow measurements of the reactant gas into the
FALP apparatus. Daily passivation of the flowmeter and
feedline was required, as evidenced by an increase in the
apparent attachment rate constant with time, asymptotically
III. EXPERIMENTAL RESULTS
A. Ion-molecule reactions
Reaction rate constants and products of the systems in-
volving CFϪ3 which were found to be reactive are given in
Table I, and nonreactive cases are listed in Table II along
with upper limits to the rate constants. Reported thermo-
chemistry ͑other than the present computational results͒ is
based on data from NIST.11 In this survey, the only reaction
of CFϪ3 which was found to produce CF3OϪ was an O atom
transfer reaction between CFϪ3 and SO2, a reaction with a
large rate constant. Exothermic formation of CF3OϪ was
possible for all the other oxygen-containing reactant neutrals
studied ͑except for CO͒ but was not observed.
Other reactions of CFϪ3 involved charge transfer, atom
transfer, displacement, fluoride transfer, and association, as
seen in Table I. The fast charge transfer reaction to NO2
confirms that the adiabatic electron affinity of the CF3 radical
is less than the 2.273 eV value known12 for NO2. The reac-
tion of CFϪ3 with CO2 proceeds entirely by association, de-
spite the fact that formation of CF3OϪ is exothermic by 31.3
kcal molϪ1. CFϪ3 reacts by halide displacement with the acids
HCl and HBr and with the methyl halides CH3Cl and CH3Br,
with faster rates for the bromides. The reaction with CF3Br is
inefficient and proceeds by bromide displacement. CFϪ3 effi-
TABLE II. Rate constant upper limits for neutrals unreactive with CFϪ3 in
ϳ0.4 Torr helium gas.
TABLE III. Rate constants (ka) for electron attachment to (CF3O)2 as a
function of temperature. The estimated accuracy is Ϯ30%.
ka (10Ϫ8 cm3 sϪ1
)
System
k2 ͑cm3 sϪ1
͒
T ͑K͒
CFϪ3 ϩO2
Ͻ6ϫ10Ϫ13
Ͻ1ϫ10Ϫ12
Ͻ8ϫ10Ϫ13
Ͻ1ϫ10Ϫ12
Ͻ2ϫ10Ϫ12
297
363
424
494
553
0.92
1.2
1.3
1.5
1.8
CFϪ3 ϩCO
CFϪ3 ϩN2O
CFϪ3 ϩSF6
CFϪ3 ϩCF3Cl
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.42.202.150 On: Fri, 21 Nov 2014 20:10:00

8438
J. Chem. Phys., Vol. 110, No. 17, 1 May 1999
Morris et al.
FIG. 2. Arrhenius plot of the rate constants for electron attachment to
(CF3O)2 . The error bars represent estimated relative error limits.
ciently transfers fluoride to SiF4. The reaction with OCS is
fast and displays both atom transfer and association reaction
channels, while the energetically allowed formation of
CF3OϪ was not observed.
The association reaction of FϪ with CF2O, forming
CF3OϪ ͓Reaction ͑1͔͒, was studied as a function of pressure
in N2 buffer from 0.22 to 0.6 Torr at 300 K. The second-
order rate constants are plotted as a function of pressure in
Fig. 1. The data fall on a straight line within the relative error
bars ͑shown as Ϯ7.5%͒. However, the intercept is distinctly
nonzero. This implies that there is either a radiative compo-
nent to the association reaction or that fall-off behavior is
applicable. The limited pressure range is not sufficient to
distinguish between these two possibilities. If radiative asso-
ciation is the reason for the nonzero intercept, the radiative
component of the second-order rate constant is derived from
the intercept and is 4.1ϫ10Ϫ11 cm3 sϪ1. In this scenario the
slope is the third-order rate constant, equal to 1.5
ϫ10Ϫ27 cm6 sϪ1
.
B. Electron attachment
FIG. 3. Optimized geometries for the Cs-symmetry 2A minimum and the
Ј
Results from the electron attachment experiments are
given in Table III. Each datum is the average of 2–4 runs at
different (CF3O)2 concentrations. At room temperature the
electron attachment rate constant for Reaction 2 is 9.2
ϫ10Ϫ9 cm3 sϪ1, which corresponds to one attachment event
in roughly every 30 collisions.13 An Arrhenius-type plot of
the data is shown in Fig. 2. The rate constants can be fit with
the expression k2ϭ3.5ϫ10Ϫ8 cm3 sϪ1 exp(Ϫ0.036 eV/kT)
over this limited temperature range, where k2 is the attach-
ment rate constant. The 36 meV activation energy in this
expression is likely related to energy given up by the incom-
ing electron to induce vibrations in (CF3O)2 which enhance
bond scission and promote electron attachment. Observation
of CF3OϪ as the only ionic product of attachment at thermal
energies is consistent with the earlier electron-beam work of
MacNeil and Thynne.14 FϪ was also seen at zero electron
energy in the earlier work, but was attributed to interaction
with a hot filament in the apparatus.
2A transition state of CF O.
Љ
3
C. Theoretical methods
The quantum chemical computations have been executed
using the PSI 2.0 program package15 to complete spin-
restricted closed-shell Hartree Fock ͑RHF͒ and spin-
restricted open-shell Hartree Fock ͑ROHF͒ computations and
the GAUSSIAN 94 program system16 to complete the density
functional theory ͑DFT͒ computations. The DFT methods
used were the BLYP, B3LYP, BP86, and BHLYP methods.
The BLYP method uses Becke’s 1988 exchange functional
17
͑B͒ with the Lee, Yang, and Parr correlation functional
͑LYP͒.18 The B3LYP method is a DFT-HF hybrid method
19
that uses Becke’s exchange functional ͑B3͒ with the LYP
correlation functional. The BP86 method uses the B ex-
change functional with the correlation correction of Perdew
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.42.202.150 On: Fri, 21 Nov 2014 20:10:00

J. Chem. Phys., Vol. 110, No. 17, 1 May 1999
Morris et al.
8439
TABLE IV. Relative energies of the 2A and 2A stationary points of CF O
2
Ј
Љ
3
attempted, the A structure is lower in energy, though the
Ј
in cmϪ1 a
.
DFT methods give an energy difference of only 26–57
cmϪ1
.
BLYP
BP86
B3LYP
0.0͑0.0͒
BHLYP
SCF
0.0͑0.0͒
2
The DFT harmonic vibrational frequencies for the A
and A stationary points are included in Table V. The DFT
Ј
2A CF O
0.0 ͑0.0͒
0.0͑0.0͒
0.0͑0.0͒
2
Ј
3
Љ
2A CF O 50.7͑53.1͒ 57.1͑60.8͒ 40.2͑42.8͒ 26.1͑31.4͒ 10.8͑16.1͒
Љ
3
aAll results were obtained with the DZPϩϩ basis set. Values are corrected
for ZPVE via the harmonic approximation and the omission of the low
frequency CF2 twist-rock mode from the computation of the ZPVE of each
structure. Noncorrected values in parentheses.
TABLE V. Harmonic vibrational frequencies in cmϪ1 computed with the
DZPϩϩ basis set.a
SCF BLYP BP86 B3LYP
BHLYP
Experimental
CF O 2A
͑P86͒.20 Finally, the BHLYP method uses Becke’s half-and-
half exchange functional ͑BH͒ along with the LYP corre-
lation functional.
Ј
3
a (a )
1455 1157 1192 1252 1356 ͑353͒
1421 1081 1115 1200 1316 ͑433͒
1415 1010 1042 1164 1303 ͑371͒
Ј
21
1
C–O stretch
a (e)
Љ
1221.6b
907c
The basis set used was the standard double-␨ plus polar-
ization ͑DZP͒ basis set augmented with diffuse orbitals,
called the DZPϩϩ basis set. The DZ part of the basis was
constructed from the Huzinaga–Dunning22,23 set of con-
tracted Gaussian functions. The DZP basis was formed by
the addition of a set of d-type polarization functions con-
structed using the five pure d components ͕␣d(C)ϭ0.75,
␣d(F)ϭ1.00, ␣d(O)ϭ0.85͖. The DZPϩϩ basis was con-
structed by the addition of diffuse functions to the DZP ba-
sis. Each atom received one additional s type and one addi-
tional set of p-type functions ͕␣s(C)ϭ0.04302, ␣p(C)
ϭ0.03629, ␣s(F)ϭ0.10490, ␣p(F)ϭ0.08260, ␣s(O)
ϭ0.08227, ␣p(O)ϭ0.06508͖. The diffuse function orbital
exponents were determined in an ‘‘even tempered sense’’ as
a mathematical extension of the primitive set, according to
the formula of Lee and Schaefer.24 The contraction scheme
for each atom is 10s6p1d/5s3p1d.
C–F stretch
a (e)
Ј
C–F stretch
a (a )
991
687
658
648
465
233
835
573
551
532
366
234
852
582
560
538
367
245
886
608
585
566
396
252
944 ͑3͒
646 ͑14͒
621 ͑3͒
605 ͑12͒
429 ͑1͒
253 ͑2͒
Ј
1
C–F stretch
a (a )
Ј
1
C–F3 bend
a (e)
Љ
CF2 twist
a (e)
Ј
CF2 scissor
a (e)
483c
Ј
CF2 wag
a (e)
Љ
CF2 rock
CF O 2A
Љ
3
a (a )
1454 1154 1168 1247 1352 ͑327͒
1423 1062 1093 1189 1309 ͑410͒
1413 1027 1061 1176 1310 ͑404͒
Ј
1
C–O stretch
a (e)
Ј
C–F stretch
a (e)
Љ
C–F stretch
a (a )
C–F stretch
991
687
672
632
457
835
574
563
519
368
852
582
571
525
371
887
608
597
553
395
283i
943 ͑3͒
646 ͑13͒
634 ͑5͒
590 ͑8͒
424 ͑Ͻ1͒
284i
Ј
D. Computational results
1
The structures of CF3O and CF3OϪ have previously been
studied computationally with ROHF and RHF wave func-
tions and a 3-21G basis set.25 A more recent study of the
a (a )
Ј
1
C–F3 bend
a (e)
Ј
CF2 scissor
CF3OϪ anion reported an optimization within C3v-symmetry
a (e)
Љ
26
**
constraints at the 6-311ϩϩG RHF level. This work,
CF2 twist
while adding the effects of electron correlation to these pre-
vious computations, includes an investigation of a larger por-
tion of the potential energy hypersurface, in search of here-
tofore undiscovered low-energy minima.
a (e)
Ј
CF2 wag
a (e)
262i 262i 273i
Љ
CF2 rock
CF OF 2A
Ј
2
Francisco and Williams25 observed that the E state of
2
a
1767 1793 1888 2061 ͑648͒
1219 1256 1291 1363 ͑471͒
Ј
C3v-symmetry CF3O is Jahn–Teller distorted, causing CF3O
to have two low-lying Cs-symmetry stationary points corre-
sponding to 2A and 2A electronic states. The computed
C–O stretch
a
Ј
C–F stretch
Ј
Љ
a
930
714
592
548
250
123
57
950
723
598
554
259
128
58
984
760
622
576
206
104
50
1041 ͑63͒
812 ͑49͒
647 ͑14͒
605 ͑7͒
Ј
2
2
C -symmetry optimized geometries for the A and the A
Ј
Љ
s
C–F stretch
F C–O
2
stationary points are given in Fig. 3. For the A stationary
Ј
a
Љ
2
point, the pure DFT methods, BLYP and BP86, give an in-
plane C–F bond which is larger than the out-of-plane C–F
bonds by 0.01 Å, while the BHLYP and self consistent field
͑SCF͒ predict the out-of-plane C–F bonds are slightly
longer. The reverse situation occurs for the A state where
only the pure DFT methods predict that the out-of-plane C–F
bonds are longer. The Jahn–Teller distortion is clearly very
sensitive to differences in the exchange and correlation treat-
ments. The relative energies of the optimized A and A
stationary points are provided in Table IV. For all methods
out-of-plane
a
Ј
CF2 rock
a
Ј
CF2 scissor
2
Љ
a
101 ͑2͒
Ј
O–F stretch
a
33 ͑Ͻ1͒
27 ͑Ͻ1͒
Ј
C–O–F bend
F–C–O–F
a
Љ
2
2
Ј
Љ
torsion
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.42.202.150 On: Fri, 21 Nov 2014 20:10:00

8440
J. Chem. Phys., Vol. 110, No. 17, 1 May 1999
Morris et al.
TABLE V. ͑Continued.͒
Matrixd Crystale
CF3OϪ 1A1
a1
1685 1614 1629 1648 1680 ͑881͒
1514
1560
960
813
595
C–O stretch
e
C–F stretch
a1
C–F stretch
a1
1079 805 832 883 983 ͑508͒ 919, 1039
923 721 743 784 859 ͑107͒
670 524 538 570 621 ͑Ͻ1͒
808
C–F3 bend
e bend
e bend
655 484 505 541
465 369 377 398 431 ͑Ͻ1͒
603 ͑15͒
555
576
422
aExperimental frequencies are fundamentals. Intensities in km molϪ1 are re-
ported in parentheses for BHLYP. Irreducible representations in parenthe-
ses are those of the vibrational modes in the C3v-symmetry point group.
bReferences 30 and 31.
cReference 29.
dReference 35.
eReference 34.
FIG. 4. Optimized geometry for the C3v-symmetry structure of CF3OϪ.
methods consistently show the 2A stationary point is a mini-
Ј
mum energy structure, and the 2A stationary point, possess-
Љ
minimum, with a O–F bond distance of 2.156 Å and a
C–O–F bond angle of 108.5°. This loosely bound structure,
relative to the FϩCF2O fragments, has energy Ϫ8.89
͑BLYP͒, Ϫ8.51 ͑BP86͒, Ϫ3.78 ͑B3LYP͒, and Ϫ1.19
kcal molϪ1 ͑BHLYP͒. The structure is unbound at the ROHF
level. The computed vibrational frequencies for this weakly
bound structure are given in Table V.
ing one imaginary vibrational frequency, is a transition state.
A detailed description of the analogous situation for CH3O
has been given by Colwell, Amos, and Handy.27 There are
2
actually three equivalent A minima and three equivalent
Ј
2A transition states for CF O, each set arising from a dis-
Љ
3
tinct plane of symmetry in the C3v point group. Each of the
2A stationary points is a transition state for the interconver-
Љ
CF3OϪ has been experimentally produced in several
ways. Electron attachment to (CF3O)2 probably produces the
C3v-symmetry structure, which is expected to be the most
stable isomer of its composition. However, the rapid addition
of FϪ to CF2O leaves the type of addition more in doubt. Is
the C3v structure of CF3OϪ formed directly or is a relatively
weak electrostatically bound FϪϩCF2O structure formed
with a barrier to the C3v structure? A third possibility is a
stable F2COFϪ structure.
2
sion of equivalent A minima.
Ј
A laser induced fluorescence study of CF3O in a super-
sonic jet has produced a high resolution spectrum of the
28
˜ ˜
progressions were concluded to indicate a C3v-symmetry E
origin band of the A←X transition. The rotational band
2
ground state structure. Another laser induced fluorescence
experiment29 has assigned a value of 907 cmϪ1 to the C–F
symmetric stretching mode and 483 cmϪ1 to a F–C–O bend-
ing mode. The former is in good agreement with the B3LYP
value of 886 cmϪ1 while the latter is significantly larger than
the usually reliable B3LYP frequency of 396 cmϪ1. Also,
matrix infrared spectroscopy experiments30,31 have provided
an experimental antisymmetric C–F stretching frequency of
1221.6 cmϪ1 which is consistent with the B3LYP prediction
of 1200 cmϪ1. Although the computed vibrational frequen-
cies are harmonic, experience has shown that DFT ͑and es-
pecially B3LYP͒ harmonic frequencies tend to be close in
value to experimental fundamentals.32,33
The computed C3v-symmetry optimized geometries of
CF3OϪ are given in Fig. 4. Wiberg26 optimized the
C3v-symmetry CF3OϪ structure at the 6-311ϩϩG RHF
**
level and obtained a geometry similar to the RHF results
reported here, except for the C–O bond length for which he
obtained a shorter length of 1.214 Å. The harmonic vibra-
tional frequencies of CF3OϪ are provided in Table V.
TABLE VI. Adiabatic and vertical electron affinities computed with the
DZPϩϩ basis set in eV.a
Other portions of the CF3O potential energy surface
were investigated using the DZPϩϩB3LYP level of theory. A
C2v-symmetry optimization of CF2OϩF, with F on the O
end of CF2O, obtained a stationary point with one b1-type
(16i cmϪ1) and one b2-type (46i cmϪ1) imaginary vibra-
tional frequency. Displacing and optimizing in the plane of
symmetry containing only one fluorine atom resulted in a
BLYP BP86 B3LYP BHLYP SCF
4.60 4.69 4.60 4.13
Expt.
EA (2A CF O)
2.75 4.26Ϯ0.22b
Ј
ad
3
͑4.59͒ ͑4.68͒ ͑4.58͒
͑4.11͒ ͑2.73͒ 3.70Ϯ0.5c
3.25Ϯ0.14d
EAvert(2A CF O) ͑3.66͒ ͑3.77͒ ͑3.61͒
͑3.15͒ ͑1.81͒
Ј
3
transition state with one a -type (50i cmϪ1) imaginary fre-
Љ
aCorrected for zero-point vibrational energy. Non-corrected values in paren-
quency and an energy that was 1.5 kcal molϪ1 below the
FϩCF2O fragments. This transition state had an O–F bond
distance of 2.303 Å and a C–O–F bond angle of 90.5°. Dis-
placing into C1 symmetry and optimizing leads to a planar
theses.
bReference 38.
cReference 39.
dReference 40.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.42.202.150 On: Fri, 21 Nov 2014 20:10:00

J. Chem. Phys., Vol. 110, No. 17, 1 May 1999
TABLE VII. Bond dissociation energies of CF3O and CF3OϪ in kcal molϪ1 a
Morris et al.
8441
.
G2 Free
Energy
͑298 K͒
b
b
BLYP
BP86
B3LYP
BHLYP
SCF
G2 ͑0 K͒
Expt.
2A CF O→CF OϩF
25.1
͑26.0͒
46.4
27.8
͑28.8͒
49.3
25.0
͑26.1͒
49.2
25.0
͑26.4͒
52.5
16.3
͑17.8͒
48.5
24.1
16.6
41.1
22.9Ϯ2.6c
42.6Ϯ2.0d
Ј
3
2
1A1CF3OϪ→CF2OϩFϪ
49.3
͑47.0͒
͑50.0͒
͑50.0͒
͑53.5͒
͑49.5͒
aResults were obtained with the DZPϩϩ basis set and are corrected for ZPVE via the harmonic approximation. Values in parentheses are the uncorrected
values.
bReference 43.
cD298 K , Ref. 41.
dReference 38.
IV. CONCLUSIONS
Experimental fundamental frequencies are available
from crystal34 and from matrix35 investigations. Ion pairing
in the latter experiment caused a distortion of CF3OϪ away
from the C3v structure, but the correspondence between the
observed frequencies and the computed C3v frequencies is
still clear. The BHLYP and B3LYP methods give the best
agreement with these experimental frequencies.
Several methods for generating CF3OϪ were examined.
A number of reactions of CFϪ3 with oxygenated neutrals
were surveyed, but only the reaction of CFϪ3 with SO2
formed CF3OϪ despite the availability of exothermic path-
ways for most of the other reactant neutrals studied. The
association reaction of FϪ with CF2O either has an radiative
channel or is in the fall-off region in 0.22–0.6 Torr N2 gas.
The attachment of thermal electrons to (CF3O)2 formed
CF3OϪ exclusively over the temperature range from 297 to
553 K, and the activation energy derived from this tempera-
ture dependence most likely reflects an energy barrier to this
process. Thus, several methods of generating CF3OϪ were
found.
An investigation of other possible structures was carried
out using the DZPϩϩB3LYP level of theory. In C2v symme-
try, optimization with the FϪ on the F2 side of CF2O led to a
transition state with one imaginary b1 frequency (59i cmϪ1),
while optimization with FϪ on the O side of CF2O led to
dissociation. Displacements out of the molecular plane into
Cs symmetry, followed by optimization, led to the
C3v-symmetry structure. No barrier to addition in this plane
was found, and no change of orbital occupation for
FϪϩCF2O→CF3OϪ in this plane is required. Optimizations
with all atoms in a plane, however, led only to dissociation
or to very high-lying transition states, displacing from which
into C1 symmetry led to the C3v-symmetry minimum.
The computed electron affinities of CF3O are given in
Table VI. The DFT adiabatic electron affinities for CF3O
range from 4.13 to 4.69 eV. Past experience32,36,37 suggests
that the lower end of this range is probably the most accu-
rate. The BHLYP value of 4.13 eV is, in fact, within the
error bars of the experimental electron affinities of Larson
and McMahon38 as well as Spyrou and co-workers.39 The
much lower value of 3.25 eV from Taft et al.40 is clearly too
low. The vertical electron affinities are about 1 eV lower
than the adiabatic electron affinities, the adiabatic correction
involving an increase in all of the C–F bonds in CF3O by
0.08–0.10 Å and a contraction of the C–O bond distance by
0.12–0.14 Å.
In order to examine whether the three sources of CF3OϪ
produced different isomers, DFT calculations for both CF3O
and CF3OϪ were performed. The DFT geometries and fre-
quencies predict a Jahn–Teller distorted CF3O, giving rise to
2
2
C -symmetry A minima and A transition states, a pair
Ј
Љ
s
corresponding to each plane of symmetry in the C3v point
group. For the CF3OϪ anion the ground state is A1 with a
1
C3v-symmetry minimum. The examination of possible struc-
tural isomers of CF3OϪ located several stationary points, but
none were shown to be minima by evaluation of their har-
monic vibrational frequencies. No transition states were
found, or barrier predicted, for addition of FϪ in the symme-
try plane perpendicular to the molecular plane in
C2v-symmetry CF2O. Therefore, the observation of rapid FϪ
addition to CF2O appears to be directly to the C3v-symmetry
structure of CF3OϪ.
Thus, CF3OϪ may be synthesized by any of the routes
studied here for use as a chemical ionization agent in trace
neutral detection. The choice is driven by source gases: SO2
is readily available and easily handled, but the CFϪ3 primary
ion is not so convenient to produce without creating possibly
troublesome species in a chemical ionization environment.
Electron attachment to (CF3O)2 is a clean process, forming
only CF3OϪ, but the starting compound is not commercially
available. CF2O gas is not so pleasant to use, but is commer-
cially available,42 and as we have shown, the adduct with FϪ
is efficiently formed.
Bond dissociation energies for CF3O and CF3OϪ are
given in Table VII. Schneider and Wallington41 recom-
mended a fluorine dissociation energy from CF3O of 22.9
Ϯ2.6 kcal molϪ1 based on experimental and theoretical data.
The DFT computed values agree with the high end of this
estimate and range from 25.0 to 27.8 kcal molϪ1. The DFT
fluoride dissociation energies from CF3OϪ are more scat-
tered with a range of 46.4–52.5 kcal molϪ1 and lie above the
38
experimental value of 42.6Ϯ2.0 kcal molϪ1
.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.42.202.150 On: Fri, 21 Nov 2014 20:10:00

8442
J. Chem. Phys., Vol. 110, No. 17, 1 May 1999
Morris et al.
ACKNOWLEDGMENTS
J. A. Montgomery, K. Raghavachari, M. A. Al-Laham, V. G. Zakrzewski,
J. V. Oritz, J. B. Foresman, J. Cioslowski, B. B. Stefanov, A. Nanay-
akkara, 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. J. DeFrees, J. Baker, J. P. Stewart, M. Head-Gordon, C.
Gonzalez, and J. A. Pople ͑Gaussian, Inc., Pittsburgh, PA, 1995͒.
17 A. D. Becke, Phys. Rev. A 38, 3098 ͑1988͒.
Helpful discussions with Dr. Yukio Yamaguchi, Dr.
Wesley Allen, Dr. Amy Stevens Miller, and Dr. John Seeley
are gratefully acknowledged. We thank Darryl DesMarteau
for supplying the (CF3O)2 . This research was supported by
the United States Air Force Office of Scientific Research
under Grant AFOSR 95-1-0057 and Task 2303EP4.
18 C. Lee, W. Yang, and R. G. Parr, Phys. Rev. B 37, 785 ͑1988͒.
19 A. D. Becke, J. Chem. Phys. 98, 1372 ͑1993͒.
20 J. P. Perdew, Phys. Rev. B 33, 8822 ͑1986͒.
21 A. D. Becke, J. Chem. Phys. 98, 5648 ͑1993͒.
1 L. G. Huey, P. W. Villalta, E. J. Dunlea, D. R. Hanson, and C. J. Howard,
22 S. Huzinaga, J. Chem. Phys. 42, 1293 ͑1965͒.
J. Phys. Chem. 100, 190 ͑1996͒.
23 T. H. Dunning, J. Chem. Phys. 53, 2823 ͑1970͒.
24 T. J. Lee and H. F. Schaefer, J. Chem. Phys. 83, 1784 ͑1985͒.
25 J. S. Francisco and I. H. Williams, Chem. Phys. Lett. 110, 240 ͑1984͒.
26 K. B. Wiberg, J. Am. Chem. Soc. 112, 3379 ͑1990͒.
27 S. M. Colwell, R. D. Amos, and N. C. Handy, Chem. Phys. Lett. 109, 525
͑1984͒.
2 E. Arijs and E. Kopp ͑private communication͒.
3
´
M. K. W. Ko, N.-D. Sze, J. M. Rodrıguez, D. K. Weistenstein, C. W.
Heisey, R. P. Wayne, P. Biggs, C. E. Canosa-mas, H. W. Sidebottom, and
J. Treacy, Geophys. Res. Lett. 21, 101 ͑1994͒.
4 J. A. Montgomery, H. H. Michels, and J. S. Francisco, Chem. Phys. Lett.
220, 391 ͑1994͒.
5 G. S. Tschumper and H. F. Schaeffer, J. Chem. Phys. 107, 2529 ͑1997͒.
6 J. C. Rienstra-Kiracofe, D. E. Graham, and H. F. Schaeffer, Mol. Phys. 94,
767 ͑1998͒.
28 X.-Q. Tan, M.-C. Yang, C. C. Carter, J. M. Williamson, T. A. Miller, T.
E. Mlsna, J. D. O. Anderson, and D. D. DesMarteau, J. Phys. Chem. 98,
2732 ͑1994͒.
7 A. A. Viggiano, R. A. Morris, F. Dale, J. F. Paulson, K. Giles, D. Smith,
and T. Su, J. Chem. Phys. 93, 1149 ͑1990͒.
29 Z. Li and J. S. Francisco, Chem. Phys. Lett. 186, 336 ͑1991͒.
30 K. C. Clemitshaw and J. R. Sodeau, J. Phys. Chem. 93, 3552 ͑1989͒.
31 L. Andrews, M. Hawkins, and R. Withnall, Inorg. Chem. 24, 4234 ͑1985͒.
32 R. A. King, V. S. Mastryukov, and H. F. Schaefer, J. Chem. Phys. 105,
6880 ͑1996͒.
8 T. M. Miller, A. E. S. Miller, J. F. Paulson, and X. Liu, J. Chem. Phys.
100, 8841 ͑1994͒.
9 J. B. Levy and R. C. Kennedy, J. Am. Chem. Soc. 94, 3302 ͑1972͒.
10 D. Smith and P. Spanel, Adv. At., Mol., Opt. Phys. 32, 307 ͑1995͒.
11 NIST Standard Reference Database No. 69, Vol., edited by W. G. Mallard
and P. J. Linstrom ͑http://webbook.nist.gov͒ Gaithersburg, 1998 ͑unpub-
lished͒.
33 A. P. Scott and L. Radom, J. Phys. Chem. 100, 16502 ͑1996͒.
34 K. O. Christie, E. C. Curtix, and C. J. Schack, Spectrochim. Acta A 31,
1035 ͑1975͒.
35 B. S. Ault, J. Phys. Chem. 84, 3448 ͑1980͒.
12 K. M. Ervin, J. Ho, and W. C. Lineberger, J. Phys. Chem. 92, 5405
͑1988͒.
36 R. A. King, J. M. Galbraith, and H. F. Schaefer, J. Phys. Chem. 100, 6061
͑1996͒.
13 C. E. Klots, Chem. Phys. Lett. 38, 61 ͑1976͒. A polarizability of 5 Å was
assumed for (CF3O)2 to estimate the collision cross section.
14 K. A. G. MacNeil and J. C. J. Thynne, Int. J. Mass Spectrom. Ion Phys. 9,
135 ͑1972͒.
37 T. J. V. Huis, J. M. Galbraith, and H. F. Schaefer, Mol. Phys. 89, 607
͑1996͒.
38 J. W. Larson and T. B. McMahon, J. Am. Chem. Soc. 105, 2944 ͑1983͒.
39 S. M. Spyrou, S. R. Hunter, and L. G. Christophorou, J. Chem. Phys. 81,
4481 ͑1984͒.
15 PSI 2.0.; C. L. Janssen, E. T. Seidl, G. E. Scuseria, T. P. Hamilton, Y.
Yamaguchi, R. B. Remington, Y. Xie, G. Vacek, C. D. Sherill, T. D.
Crawford, J. T. Fermann, W. D. Allen, B. R. Brooks, G. B. Fitzgerald, D.
J. Fox, J. F. Gaw, N. C. Handy, W. D. Laidig, T. J. Lee, R. M. Pitzer, J.
E. Rice, P. Saxe, A. C. Scheiner, and H. F. Schaefer. This program is
generally available for a handling fee of $100. ͑PSITECH, Inc., Watkins-
ville, GA 30677, 1995͒.
40 R. W. Taft, I. J. Koppel, R. D. Topsom, and F. Anvia, J. Am. Chem. Soc.
112, 2047 ͑1990͒.
41 W. F. Schneider and T. J. Wallington, J. Phys. Chem. 98, 7448 ͑1994͒.
42 Oakwood Products, Inc., West Columbia, SC; Lancaster Synthesis, Inc.,
Windham, NH.
43 L. A. Curtiss, K. Raghavachari, G. W. Trucks, and J. A. Pople, J. Chem.
Phys. 94, 7221 ͑1991͒.
16
GAUSSIAN 94. M. J. Frisch, G. W. Trucks, H. B. Schlegel, P. M. W. Gill, B.
G. Johnson, M. A. Robb, J. R. Cheeseman, T. A. Keith, G. A. Petersson,
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.42.202.150 On: Fri, 21 Nov 2014 20:10:00