Experimental and theoretical study of the carbon-13 and deuterium kinetic isotope effects in the Cl and OH reactions of CH3F

Article abstract of DOI:10.1021/jp807609d

A laser flash photolysis-resonance fluorescence technique has been employed to determine absolute rate coefficients for the CH₃F + Cl reaction in N₂ bath gas in the temperature range of 200-700 K and pressure range of 33-133 hPa. The data were fitted to a modified Arrhenius expression k(T) = 1.14 × 10^-12 x (T/ 298)²²⁶ exp{-313/T}. The OH and Cl reaction rates of ¹³CH₃F and CD₃F have been measured by long-path FTIR spectroscopy relative to CH₃F at 298 ± 2 K and 1013 ± 10 hPa in purified air. The FTIR spectra were fitted using a nonlinear least-squares spectral fitting method including line data from the HITRAN database and measured infrared spectra as references. The relative reaction rates defined by α = k_light/k _heavy were determined to be k_OH+CH3F/k_OH+CD3F = 4.067 ± 0.018, k_OH+CH3F/kOH+¹³CH₃F = 1-067 ± 0.006, k_Cl+CH3F/k_Cl+CD3F = 5.11 ± 0.07, and k_Cl+CH3F/k_Cl+13CH3F = 1-016 ± 0.006. The carbon-13 and deuterium kinetic isotope effects in the OH and Cl reactions of CH₃F have been further investigated by quantum chemistry methods and variational transition state theory.

Full text of DOI:10.1021/jp807609d

1
2416
J. Phys. Chem. A 2008, 112, 12416–12429
Experimental and Theoretical Study of the Carbon-13 and Deuterium Kinetic Isotope
Effects in the Cl and OH Reactions of CH3F
†
†,‡
§
|
Marina Marinkovic, Margret Gruber-Stadler, J. Michael Nicovich, Raenell Soller,
‡
§,|
†
,†
Max M u¨ lh a¨ user, Paul H. Wine, Lihn Bache-Andreassen, and Claus J. Nielsen*
Centre for Theoretical and Computational Chemistry, Department of Chemistry, UniVersity of Oslo,
P.O. Box. 1033, Blindern 0315 Oslo, Norway, Studiengang Umwelt-, Verfahrens- & Biotechnik,
MCIsManagement Center Innsbruck Internationale Fachhochschulgesellschaft mbH, Egger-Lienz-Strasse 120,
A-6020 Innsbruck, Austria, and School of Chemistry and Biochemistry and School of Earth and Atmospheric
Sciences, Georgia Institute of Technology, Atlanta, Georgia 30332
ReceiVed: August 26, 2008; ReVised Manuscript ReceiVed: September 20, 2008
A laser flash photolysis-resonance fluorescence technique has been employed to determine absolute rate
coefficients for the CH
3
F + Cl reaction in N bath gas in the temperature range of 200-700 K and pressure
2
-
12
range of 33-133 hPa. The data were fitted to a modified Arrhenius expression k(T) ) 1.14 × 10 × (T/
2
.26
13
2
98) exp{-313/T}. The OH and Cl reaction rates of CH
3 3
F and CD F have been measured by long-path
FTIR spectroscopy relative to CH
3
F at 298 ( 2 K and 1013 ( 10 hPa in purified air. The FTIR spectra were
fitted using a nonlinear least-squares spectral fitting method including line data from the HITRAN database
and measured infrared spectra as references. The relative reaction rates defined by R ) klight/kheavy were
determined to be kOH+CH3F/kOH+CD3F ) 4.067 ( 0.018, kOH+CH3F/kOH+13CH F ) 1.067 ( 0.006, kCl+CH3F/kCl+CD3F
3
)
5.11 ( 0.07, and kCl+CH3F/kCl+13CH3F ) 1.016 ( 0.006. The carbon-13 and deuterium kinetic isotope effects
in the OH and Cl reactions of CH
variational transition state theory.
3
F have been further investigated by quantum chemistry methods and
2
7-29
1
. Introduction
Isotopes are frequently employed in studies of chemical
published for the fluoromethane reactions with Cl atoms
and OH radicals.
3
0-35
reactions to distinguish between alternative mechanisms. In
chemical kinetic studies, the isotope effects (KIEs) give ad-
ditional information about the potential energy surface of the
reaction and serve as benchmarks in quantum chemistry
descriptions. Experimental KIEs are available for the radical
2
. Experimental and Computational Methods
.1. Reference Spectra and Relative Rate Experiments.
2
The experiments were carried out in a 250 L electropolished
stainless steel smog chamber equipped with a White type
multiple reflection mirror system with a 120 m optical path
length for rovibrationally resolved infrared spectroscopy. The
infrared spectra were recorded with a Bruker IFS 66v FTIR
instrument employing an InSb detector. The reaction chamber
is equipped with UV photolysis lamps mounted in a quartz tube
inside the chamber, and all experiments were carried out in
synthetic air (AGA 99.99%; CH4, CO, and NOx < 100 ppbv)
at 298 ( 2 K and 1013 ( 10 hPa. Each of the Cl and OH
experiments was carried out several times for different mixtures
1-3
4,5
reactions of the key atmospheric compounds CH4, HCHO,
6
,7
and CO (and references therein). Data are also available for
8
,9
10
a few halogenated C1-compounds such as CH3Cl, CHCl3,
CH3I, and CH3OH.
1
1
12
Fluoromethane (CH3F, Freon 41, HFC 41, F 41, R 41) is
presently not of environmental concern, but, being the simplest
HFC, it has served as both a standard and a mimic of larger
HFCs and has been the subject of several studies reporting
1
3-19
kinetic results from the reaction with OH radicals
and Cl
13
of CH3F + CD3F and CH3F + CH3F. Initial partial pressures
2
0-24
atoms.
The data have been critically reviewed and the
of the fluoromethane isotopologues were in the range of 0.1-1.5
Pa and the cell was filled to 1013 hPa with synthetic air (99.99%,
AGA). As a standard, 128 scans were coadded at a nominal
2
5
-12
current JPL recommendations are k(T)OH+CH F ) 2.5 × 10
3
-
1430/T
-11
-1200/T
×
e
and k(T)Cl+CH F ) 1.96 × 10
× e
. The
3
2
6
current IUPAC recommendations are k(T)Cl+CH F ) 4.0 ×
-1
3
resolution of 0.125 cm and Fourier transformed using boxcar
-
12
-730/T
1
0
× e
in the temperature range of 240-370 K and
apodization. The kinetic experiments were programmed with
periods of data collection (ca. 4 min) alternated with UV
photolysis and a waiting period of 2 min.
-12
-1350/T
k(T)OH+CH F ) 1.9 × 10
× e
in the temperature range
3
of 240-300 K or a three-parameter fit k(T)OH+CH F ) 3.66 ×
3
-
18
2
-818/T
1
0
× T × e
in the temperature range of 243-393 K.
CH3F (99%, Fluorochem), CD3F (>99.9 atom% D, CDN
Likewise, several quantum chemical modeling studies have been
13
Isotopes), and CH3F (99% and 90%, ICON) were used as
received without further purification. Hydroxyl radicals were
generated by photolysis of O3 in the presence of H2 (99%,
AGA). Ozone was produced from oxygen (99.995%, AGA)
using a MK II ozone generator from BOC, which has a
conversion efficiency of approximately 5%, and was collected
in a trap filled with silica beads at 195 K. Typical partial
pressures of ozone and hydrogen were 50 and 200 Pa,
*
Corresponding author. Tel: +47 2285 5680. Fax: +47 2285 5441.
E-mail: claus.nielsen@kjemi.uio.no.
†
‡
University of Oslo.
MCIsManagement Center Innsbruck Internationale Fachhochschulge-
sellschaft mbH.
§
School of Chemistry and Biochemistry, Georgia Institute of Technology.
|
School of Earth and Atmospheric Sciences, Georgia Institute of
Technology.
1
0.1021/jp807609d CCC: $40.75  2008 American Chemical Society
Published on Web 11/07/2008

Kinetic Isotope Effects in the Reactions of CH3F
J. Phys. Chem. A, Vol. 112, No. 48, 2008 12417
respectively. Photolysis of ozone was carried out at intervals
of 10-15 min using a Philips TUV 30W lamp (λmax ∼ 254
nm). This OH production method produces not only OH radicals
in the ground-state but also in excited vibrational states.
The collision quenching rate coefficients of OH by N2 and O2
CO, CO2, O3, and CH4 were taken from the HITRAN database.
1
3
For CH3F, CH3F, CD3F, HC(O)F, DC(O)F, DCl, and the
-
1
overtone/combinations bands of O3 in the 3100-2700 cm
3
6-38
range (for which HITRAN does not yet provide all isotopic
data), experimental FTIR spectra obtained in our laboratory were
-
15
-13
3
-1 -1
13
is on the order of 10
and 10
cm molecule s ,
used. The infrared spectrum of a trace impurity in the CH3F
3
9
respectively. As the mixing ratios of O2 and N2 are 4 orders
of magnitude larger than those of the fluoromethane isotopo-
logues, one may safely assume that these react exclusively with
OH in the vibrational ground state.
sample was identical to one previously obtained from the
1
3
spectrum of the commercial CH3Br sample by carrying out a
short photolysis experiment with Cl2 present. The impurity
1
3
13
reacted much faster than CH3F and CH3Br did and by
spectral subtraction employing spectra of the irradiated sample,
Chlorine atoms were generated by photolysis of molecular
chlorine (2-3 Pa; 99.8%, AGA) at intervals of 1-10 min by
employing a Philips TUV 30 W lamp (λmax ∼ 254 nm). The
use of this lamp was preferred, as it will also photolyze HC(O)F
and CH2FOOH produced in the degradation of CH3F; see later
discussion. At these short wavelengths, Cl atoms can be formed
1
3
HCl, and CH2O, we obtained a spectrum of the C-H
13
stretching region of what we tentatively suggest is di- C-labeled
CH3OCH3. Table S1 (Supporting Information) sums up the
wavenumber regions employed and the chemical components
included in the analyses.
2
-1
in the excited spin-orbit state P1/2, which is 882 cm above
The data from the independent experiments were analyzed
according to eq 1 using a weighted least-squares procedure that
2
the ground state P3/2. The quenching rate coefficients of
Cl( P1/2) by O2 and N2 are on the order of 10
molecule s . The mixing ratios of O2 and N2 are 4 orders
of magnitude larger than that of CH3F, and we assume that only
Cl( P3/2) will react with CH3F, which is supported by a recent
2
-13
-15
3
43
and 10 cm
includes uncertainties in both reactant concentrations. The
-
1
-1 40
uncertainty in the concentration determination was taken as 1%
of the initial concentration.
2
2.2. LFP-RF Experimental Technique. The laser flash
photolysis (LFP)-resonance fluorescence (RF) technique was
employed in a series of experiments to measure temperature-
dependent rate coefficients for the reaction of Cl atoms with
CH F in N bath gas. The LFP-RF apparatus was similar to
study of the reactivity of thermalized Cl atoms toward CH3F
2
1
and other organic compounds.
The kinetic study was carried out by the relative rate method
in a static gas mixture, in which the removals of the reacting
species are measured simultaneously as a function of reaction
time. Assuming that the reactants react solely with the same
radical and that none of the reactants are reformed in any side
reactions, the relative rate coefficient, krel, is given according
to the following expression:
3
2
those employed in several previous studies of chlorine atom
4
4-49
kinetics carried out at Georgia Tech.
A schematic diagram
5
0
of the apparatus is published elsewhere. Important features
of the apparatus as well as experimental details that are specific
to this study are summarized below.
Two different reaction cells were employed in the LFP-RF
study. All experiments at T e 428 K employed a jacketed Pyrex
[
P]0
[I]0
kP
ln
) krel ln
; k )
(1)
rel
50
3
{
}
{
}
reaction cell with an internal volume of ∼160 cm . The cell
was maintained at a constant temperature by circulating ethylene
glycol (for 428 K g T g 300 K) or a 2:1 ethanol-methanol
mixture (for T e 276 K) from thermostated baths through the
outer jacket. All experiments at T > 428 K employed a quartz
[
P]t
[I]t
kI
where [P]0, [I]0, [P]t, and [I]t are concentrations of the parent
and the isotopologue at the start and at the time t, respectively,
and kP and kI are the corresponding rate coefficients. A plot of
ln{[P]0/[P]t} vs ln{[I]0/[I]t} will thus give the relative reaction
rate coefficient krel ) kP/kI as the slope.
Control experiments were performed to check for loss of
fluoromethane via photolysis, dark chemistry, and heterogeneous
reactions in the reactor. The lifetime of fluoromethane in the
reaction chamber was investigated with purified air as diluent
and with the relevant radical precursor mixtures in purified air
in experiments lasting from 8 to 15 h. The photostability of
fluoromethane toward the radiation used in generating the
radicals (see below) was studied in a separate experiment with
purified air as diluent: no direct photolysis in the reactor was
detected.
3
reaction cell with an internal volume of ∼250 cm . This cell
was maintained at a constant temperature by passing current
through high-resistance-wire heaters that were wrapped around
it. Copper-constantan (for low temperature studies) or
chromel-alumel (for high temperature studies) thermocouples
could be inserted into the reaction zone through a vacuum seal,
thus allowing measurement of the gas temperature in the reaction
volume (i.e., the volume from which fluorescence could be
detected) under the precise pressure and flow rate conditions
of the experiment. Temperature variation within the reaction
volumes of these cells have been found to be <0.5 K at T )
2
70 K and <1.5 K at T ) 570 K.
The experimental FTIR spectra were analyzed using a global
Chlorine atoms were produced by 266 nm laser flash
4
1
nonlinear least-squares spectral fitting procedure. In this
method, the spectrum of the mixture of absorbing species is
first simulated by calculation from initial estimates of the
absorber concentrations. The calculation is then iterated to
minimize the residual between the measured and simulated
spectrum using the Levenberg-Marquardt algorithm to adjust
the absorber concentrations, continuum level, and instrument
line shape parameters. The absorption coefficients are normally
photolysis of phosgene:
Cl CO + hν (λ ) 266 nm) f 2 Cl + CO
(2)
2
Fourth harmonic radiation from a Quanta Ray model DCR-
2A Nd:YAG laser served as the 266 nm light source; the pulse
width was ∼6 ns and fluences employed in this study ranged
-
2
-1
from 1.2 to 10 mJ cm pulse .
4
2
calculated from the HITRAN database. When HITRAN line
parameter data are not available, high-resolution FTIR spectra
are used to approximate the absorption coefficients.
An atomic resonance lamp situated perpendicular to the
photolysis laser excited resonance fluorescence in the photolyti-
cally generated chlorine atoms. The resonance lamp consisted
of an electrodeless microwave discharge through about 1 Torr
of a flowing mixture containing a trace of Cl2 in helium.
Radiation was coupled out of the lamp through a magnesium
The spectral features used in the analyses were the C-H and
-
1
C-D stretching bands in the 3150-2700 and 2300-1900 cm
regions, respectively. The spectral data for HCl, H2O, HDO,

1
2418 J. Phys. Chem. A, Vol. 112, No. 48, 2008
Marinkovic et al.
fluoride window and into the reaction cell through a magnesium
fluoride lens. Before entering the reaction cell, the lamp output
passed through a flowing gas filter containing 3 Torr cm of N2O
in N2; this filter blocked virtually all O atom impurity emissions
at 130-131 nm while transmitting the Cl atom emissions in
the 135-140 nm wavelength range. Fluorescence was collected
by a magnesium fluoride lens on an axis orthogonal to both the
photolysis beam and the resonance lamp beam and was imaged
onto the photocathode of a solar blind photomultiplier. The
region between the reaction cell and the photomultiplier was
purged with N2; in addition, a calcium fluoride window was
inserted into this region to prevent detection of lamp emissions
at wavelengths shorter than 125 nm (Lyman-R H atom emission,
for example). Fluorescence intensities were found to vary
linearly with atom concentration up to levels several times higher
cc-pVTZ//MP2/aug-cc-pVTZ and CCSD(T)/aug-cc-pVQZ//
MP2/aug-cc-pVQZ single-point correlation energies were ex-
trapolated toward the basis-set limit using the extrapolation
6
2
scheme of Halkier et al.:
3
3
X E - Y E
Y
∞
XY
X
E
)
(3)
3
3
X - Y
where EX is the correlation energy obtained with the highest
cardinal number X, and EY is the correlation energy obtained
with cardinal number Y equal to X - 1 (the cardinal numbers
of the aug-cc-pVTZ and aug-cc-pVQZ basis sets are 3 and 4,
respectively). In the following, such extrapolations will be
denoted CCSD(T)/EB43 with “EB” being short for extrapolated
basis.
1
1
The minimum energy path (MEP) connecting reactants and
products were computed at the MP2/aug-cc-pVXZ (X ) D, T)
level of theory using the intrinsic reaction coordinate (IRC)
than any used in the kinetics experiments ([Cl]0 e 4 × 10
-
3
atoms cm in all experiments). Signals were processed using
photon-counting techniques in conjunction with multichannel
scaling. For each chlorine atom decay measured, signals from
a large number of laser shots were averaged to obtain a well-
defined temporal profile over (typically) a factor of 30 variation
in the Cl atom concentration. The multichannel scalar sweep
was triggered prior to the photolysis laser in order to allow a
pretrigger baseline to be obtained.
63
method of Gonzales and Schlegel. The IRC calculations were
carried out in mass-weighted Cartesian coordinates using a step
1
/2
size of 0.02 u bohr.
.4. Calculations of Rate Coefficients. Calculations of rate
coefficients were carried out using interpolated variational
2
64
transition state theory by mapping (IVTST-M) using the sparse
grid of geometries, gradients, and Hessians from the ab initio
IRC calculations as input. The MEP was followed in mass-
scaled (scaling mass equal to 1 u) curvilinear coordinates using
In order to avoid accumulation of photochemically generated
reactive species, all experiments were carried out under “slow
flow” conditions. The linear flow rate through the reactor was
6
5
the RODS algorithm. In both reactions a step size of 0.002
bohr was used. The electronic energy of the MEP was corrected
-1
in the range of 2.0-3.1 cm s , while the laser repetition rate
was varied over the range of 2-10 Hz (it was 5 Hz in most
experiments). Since the direction of flow was perpendicular to
the photolysis laser beam, no volume element of the reaction
mixture was subjected to more than a few laser shots. Phosgene
and CH3F were flowed into the reaction cell from 12-L Pyrex
bulbs containing dilute mixtures in N2, while N2 was flowed
directly from its high-pressure storage tank. The partial pressure
of each component in the reaction mixture was evaluated from
mass flow rate and total pressure measurements. The fractional
concentration of the CH3F in the CH3F/N2 mixture was
periodically determined via UV photometry at 184.9 nm using
a Hg pen-ray lamp as the light source and a 1/4 m monochro-
mator to isolate the 184.9 nm Hg line. The molecular absorption
cross-section for CH3F at 184.9 nm was measured during this
6
6
by the interpolated single-point energy method (ISPE) using
the extrapolated CCSD(T) energies computed at all stationary
points as input. In a second procedure, the interpolated optimized
6
7
energies (IOE) method, in which the extrapolated CCSD(T)
energies and scaled frequencies at the stationary points were
used, was applied to the reaction of CH3F with chlorine. The
electronic energies of the MEP were corrected by the SECKART
method, in which a single Eckart potential is fitted to the
differences between the higher level CCSD(T)/EB43//MP2/aug-
cc-pVXZ (X ) T, Q) and the lower level MP2/aug-cc-pVTZ
energies at all stationary points. The frequencies were corrected
by the ICL method (interpolated corrections based on the
logarithm of ratios). From this information the ground-state
G
vibrationally adiabatic potential curve, Va , was obtained.
-
21
2
study and determined to be (4.9 ( 0.4) × 10 cm , where the
uncertainty represents an estimated accuracy at the 95%
confidence level.
68
The POLYRATE program was used to calculate the rate
coefficients over the temperature range of 200-1000 K using
69
improved canonical variational theory (ICVT) for the chlorine
7
0
The pure gases used in this study had the following stated
minimum purities: N2, 99.999%; Cl2CO, 99.0%; CH3F, 99.0%;
for Cl2CO and CH3F, the stated purities refer to the liquid phase
in the high-pressure storage cylinders. Nitrogen was used as
supplied, while Cl2CO and CH3F were degassed repeatedly at
reaction and canonical variational theory (CVT) for the OH
reaction. Semiclassical tunneling corrections have been included
using the centrifugal-dominant small-curvature adiabatic ground-
state tunneling (called small-curvature tunneling or SCT)
7
1
approximation. The transmission coefficient also includes the
72
7
7 K before use.
classical adiabatic ground-state (CAG) transmission coefficient,
5
1
2
.3. Electronic Structure Calculations. MP2, CCSD(T),52
which adjusts the quantal corrections for the difference between
5
3,54
G
and CBS-QB3
calculations were carried out with the
Va at its maximum and at the CVT transition state. The
55
Gaussian 03 program. Unrestricted wave functions were used
to describe open shell systems and bond-breaking processes;
singlet ground-state structures were calculated using a restricted
wave function. Additional RHF-CCSD(T) coupled cluster
geometry optimizations were carried out by employing the
Molpro 2002.6 program package, in which open shell systems
are described by the RHF-UCCSD(T) model, whereas closed
shell systems are described by the RHF-RCCSD(T) model.
calculations of the kinetic isotope effects were based on a single
reaction path using the methods outlined in refs 65 and 73.
2
In the chlorine reaction, the two spin-orbit states P3/2
5
2
2
(lowest) and P1/2 of Cl having degeneracies of 4 and 2,
-
1
respectively, and separated by 882 cm were included in the
calculation of the electronic partition function. Similarly, in the
56
2
2
OH reaction the Π and Π spin-orbit states of the OH
3
/2
1/2
57-59
radical were included; both having degeneracy equal to 2 and
-
1
The core electrons were kept frozen in the calculations.
separated by 140 cm . To estimate the contribution of SO
(spin-orbit) coupling to the potential energy surfaces of
reactions, SO matrix elements were calculated for the stationary
Dunning’s correlation-consistent aug-cc-pVXZ (X ) D, T, Q)
60,61
basis sets
were employed in all calculations. CCSD(T)/aug-

Kinetic Isotope Effects in the Reactions of CH3F
J. Phys. Chem. A, Vol. 112, No. 48, 2008 12419
The vibrational-rotational partition functions were assumed
to be separable and the rotational partition functions were
approximated by their classical limit. The vibrational partition
functions were calculated within the harmonic oscillator ap-
proximation for all modes, except for the C · · ·H· · ·OH torsional
mode in the transition state of CH3F + OH, which was handled
7
5
by the hindered rotor RW scheme. The potential of the
torsional mode, V(φ), was calculated at the MP2/aug-cc-pVTZ
1
level of theory and fitted to V(φ) ) /2W(1 - cos φ); the barrier
-1
height was found to be W ) 454 cm . For the reaction of CH3F
with chlorine, all frequencies were scaled in the calculation
based on the MP2/aug-cc-pVTZ MEP. The scaling factor c )
76
0.9552 applied to the frequencies was calculated according to
exp theo
theo 2
c )
∑( ν˜ i
ν˜ i )/
∑( ν˜ i
)
(4)
exp
where ν˜ i are the experimental frequencies of vibration, and
theo
ν˜ i are the theoretically calculated frequencies of vibration for
the modes of CH3F, CH2F, and HCl.
3
. Results and Discussion
3
.1. Experimental Study. 3.1.1. RelatiWe Rate Experiments.
The kinetic isotope effects were determined in relative rate
experiments carried out under low NOx conditions (NOx < 100
24
ppbv). According to a previous mechanistic study, the Cl (and
presumably also the OH) initiated oxidation of CH3F will, under
the present reaction conditions, essentially proceed via reactions
5
-9:
X + CH F f CH F + HX; X ) Cl, OH
(5)
(6)
3
2
CH F + O f CH FO
2
2
2
2
CH FO + CH FO f CH FO + CH FO + O
(7a)
(7b)
(8)
2
2
2
2
2
2
2
fHC(O)F + CH FOH + O
2
2
CH FO + O f HC(O)F + HO
2
2
2
CH FO +HO f HC(O)F + H O + O
(9a)
(9b)
2
2
2
2
2
fCH FOOH + O
2
2
Wallington et al.24 found that reaction 7 proceeds predomi-
nantly, if not exclusively, via reaction 7a. Further, reaction 9a
is much slower than the reaction of other alkyl peroxy radicals
with HO2 and they only observed a spectral feature around 3650
-
1
cm , which was tentatively attributed to CH2FOOH, in
experiments with initial CH3F concentrations around 100 ppm.
In addition, a later mechanistic study of the gas phase reaction
of CH2FO2 radicals with HO2 shows that reaction 9a dominates
7
7
over reaction 9b.
During the reaction of Cl atoms with fluoromethane, OH
radicals will eventually be generated from the HO2 radicals and
thereby influence the kinetics. We therefore modeled the
degradation reactions of CH3F in the static reactor using
3 3 2
Figure 1. Experimental FTIR spectra of CH F/CD F/Cl reaction
7
8
FACSIMILE to find the optimum starting conditions. The
basic model of our reactor consists of around 200 equations
describing Ox, HOx, NOx, FOx, ClOx, BrOx, and CH4 related
mixtures after 12 min photolysis and of the major chemical components
retrieved by spectral fitting. (a) Spectrum of the CH-stretching region
-
1
3
150-2750 cm . Bands marked with * are due to HCl. Compounds
7
9
included in the spectral analysis: CH F, HCOF, H O, HDO, and HCl.
3
2
“background” reactions in the chamber; the rate coefficient
-1
(
b) Spectrum of the CD-stretching region 2185-1900 cm . Compounds
included in the spectral analysis: CH
CO, and DCl.
25
80
data have been taken from the latest JPL and IUPAC kinetic
data evaluations (the photolysis rates of photolabile constituents
were determined experimentally in our laboratory). In addition,
we included the reactions describing the Cl- and OH-initiated
3 3 2
F, CD F, HCOF, DCOF, H O,
points obtained at the MP2/aug-cc-pVTZ level of theory using
the Breit-Pauli Hamiltonian as implemented in the Molpro
oxidation of CH3F as indicated by reactions 5-9 and the
7
4
81-83
2
002.6 program package. Wave functions for the SO states
subsequent degradation reactions of HC(O)F.
Using 5 ppm
were generated using MRCI with reference CASSCF configura-
tions only and the 6-311++G(3df, p) basis set.
CH3F and 10-20 ppm Cl2 as starting conditions, the model
indicated that the loss of CH3F due to reactions with OH radicals

1
2420 J. Phys. Chem. A, Vol. 112, No. 48, 2008
Marinkovic et al.
in the Cl atom experiments was less than 1% of the total loss.
The model also predicted that the concentration of HC(O)F
formed could become as high as 0.5 ppm; the absorption lines
of this molecule could therefore interfere with the analyses of
the C-H and C-D stretching regions of the spectrum. To
prevent/reduce such a buildup, we employed the Philips TUV
Figures S2 and S3 (Supporting Information) show experi-
1
3
mental FTIR spectra of the reaction mixtures CH F/ CH F/
3
3
O /H and CH F/CD F/O /H , respectively. The ozone band at
3
2
3
3
3
2
-1
3046 cm (3ν ) partly obscures the perpendicular band of CH F
3
3
(ν (E)) and overlaps the parallel band of CH F (ν (E)). This
4
3
1
complicated the spectral analysis substantially. The relative
concentrations of CD3F during the reactions were determined
from spectral analysis of the C-D stretching region (2230-2150
3
0W lamp with λmax ∼ 254 nm, which will photolyze HC(O)F:
HC(O)F + hν f CO + HF (10)
Examples of the FTIR spectra obtained from the kinetic
-
1
cm ), as this region is not obscured significantly by the CO2
fundamental (ν3) and the 2ν1/3ν2 ozone bands. From the decay
data of the fluoromethane isotopologues in the presence of OH
studies of the CH3F/CD3F reaction with Cl atoms and the
resulting residuals from the nonlinear least-squares spectral
13
radicals,shownintheformofln{[CH3F]0/[CH3F]t}vsln{[ CH3F]0/
13
analyses are shown in Figure 1. Spectra from the CH3F/ CH3F
Cl experiments are shown in Figure S1 (Supporting Informa-
1
3
[
CH3F]t} and ln{[CH3F]0/[CH3F]t} vs ln{[CD3F]0/[CD3F]t} in
+
1
3
Figure 2c,d, the following kinetic isotope effects were extracted:
tion). The decays of CH3F, CH3F, and CD3F in the presence
kOH+CH F/kOH+13CH CH F ) 1.067 ( 0.006 and kOH+CH F/kOH+CD F
3
3
3
3
3
of Cl atoms are shown as plots of ln{[CH3F]0/[CH3F]t} vs
13
13
) 4.067 ( 0.018. Again, the uncertainties represent 2σ from
the least-squares analysis and refer to precision only.
ln{[ CH3F]0/[ CH3F]t} and ln{[CH3F]0/[CH3F]t} vs ln{[CD3F]0/
CD3F]t} in Figure 2a,b. Least-squares fitting of the data resulted
in the following kinetic isotope effects: kCl+CH F/kCl+13CH F )
[
The “natural” removal of CH3F from the gas phase (purified
air) in the reactor is shown in Figure S4 (Supporting Informa-
tion). Assuming the (small) decay to be first-order in CH3F,
3
3
1
.016 ( 0.006 and kCl+CH F/kCl+CD F ) 5.11 ( 0.07. The
3
3
uncertainties represent 2σ from the least-squares analysis and
-
8
refer to precision only.
the natural loss rate coefficient was determined to be 4 × 10
3 3 3
Figure 2. Relative rate plots showing the decays of CH F, and CD F at 1013 hPa and 298 K in the presence of Cl atoms and OH radicals.
F, 13CH
The slopes correspond to the relative reaction rate R ) klight/kheavy; the errors on the slopes correspond to 2σ from the statistical analyses and do not
include possible systematic errors. Error bars assigned to the individual data points correspond to either 1% of the starting concentration or the 1σ
13
error in the spectral fitting procedure, whichever is largest. (a) Decay of CH
3
F vs CH
F vs CD F in the presence of Cl atoms; R ) 5.11 ( 0.07 (26 data points
F in the presence of OH radicals; R ) 1.067 ( 0.006 (56 data points from three
3 3
F vs CD F in the presence of OH radicals; R ) 4.067 ( 0.018 (55 data points from five independent
3
F in the presence of Cl atoms; R ) 1.016 ( 0.006 (determined
from 14 data points from three independent experiments). (b) Decay of CH
3
3
1
3
3 3
from four independent experiments). (c) Decay of CH F vs CH
independent experiments). (d) Decay of CH
experiments).

Kinetic Isotope Effects in the Reactions of CH3F
J. Phys. Chem. A, Vol. 112, No. 48, 2008 12421
s- (lifetime ∼ 240 days). For a realistic reaction mixture in
the OH radical experiments, CH3F/H2/O3 5:2000:200 ppm in
purified air, the loss rate coefficients of CH3F and O3 were
1
TABLE 1: Experimental 13C and H (D) Kinetic Isotope
2
3 3
Effects, r ) klight/kheavy, in the Reactions of CH F, CH Cl,
CH
4
, and CH
3
OH with Cl Atoms and OH Radicals at 298 K
-
7
-6 -1
simultaneously determined to be 5.6 × 10 and 1 × 10 s ,
respectively. The lifetime of CH3F in the reaction chamber under
these conditions thus decreases to around 20 days. Assuming
the difference between the two CH3F loss rates to be due to
kinetic isotope effect
isotopologues
Cl reaction
OH reaction
12CH
13
a
a
a
8
3
F/ CH
3
F
1.016 ( 0.006
1.067 ( 0.006
4.067 ( 0.018
1.059 ( 0.008
a
CH
3
F/CD
3
3
F
5.11 ( 0.07
gas-phase reaction with O3 places an upper limit of kO +CH F <
12
13
8
3
3
CH Cl/ CH Cl
1.070 ( 0.010
3
-
22
3
-1 -1
8
8
1
0
cm molecule s . The kinetic analysis according to
CH Cl/CD Cl
4.91 ( 0.07
3.9 ( 0.4
3
3
1
2
13
3
2
eq. 1 assumes that only one loss process is taking place: other
loss processes, such as gas-phase or surface reactions with the
radical precursors and natural wall loss, will give rise to
systematic errors in the kinetic isotope effects, especially in the
OH reaction, which is slow. The OH experiments generally
lasted for 3-4 h, including spectral recording and waiting time
between the photolysis, during which around 15% of the initial
fluoromethane reacted. Taking the rate coefficient for the OH
CH
4
/ CH
4
1.058 ( 0.002
1.0039 ( 0.0004
3
1
CH
CH
4
/CHD
3
3
13
4.73 ( 0.04
3.30 ( 0.50
1
2
12
12
OH/ CH
3
OH
1.055 ( 0.016
3.011 ( 0.059
1.031 ( 0.020
12
1
2
CH
3
OH/CD
3
OH
2.57 ( 0.11
a
This work.
All experiments were carried out under pseudo-first-order
conditions with CH3F in large excess over Cl atoms. Hence, in
the absence of side reactions that remove or produce Cl atoms,
the Cl atom temporal profile following the laser flash would be
described by the relationship
-
14
3
-1
reaction with fluoromethane as 2.1 × 10
at 298 K, a rough estimate of the average OH concentration
during the entire experiment (including waiting periods and time
cm molecule
-
1
25
s
8
-3
for spectrum recording) is 5 × 10 molecules cm , which gives
a ratio >20 between the loss due to reaction with OH radicals
and the “dark” loss of fluoromethane. Taking the numbers given
above, a numerical simulation shows that the systematic error
in kOH+CH F/kOH+CD F is estimated to be -0.014. For kOH+CH F/
ln{[Cl] ⁄ [Cl] } ) (k
[CH F] + k )t ) k′t (11)
3 12
0
t
Cl+CH3F
where k12 is the rate coefficient for the reaction
3
3
3
Cl f first-order loss by diffusion from the
detector field of view and ⁄
kOH+13CH F, the corresponding systematic error is estimated to
3
(12)
be even smaller. That is, the known systematic errors in the
OH radical experiments due to dark reactions of fluoromethane
result in apparent KIEs that are too small, but only by the same
amount as the 2σ statistical error of the data analysis. For a
realistic reaction mixture in the Cl atom experiments, CH3F/
Cl2 5:20 ppm in purified air, the loss rate coefficient of CH3F
or reaction with background impurities
The bimolecular rate coefficients of interest, kCl+CH F(p,T),
3
are determined from the slopes of k′ vs [CH3F] plots for data
obtained at constant T and p. Observation of Cl temporal profiles
that are exponential, i.e., obey eq 11, a linear dependence of k′
on [CH3F], and invariance of k′ to variation in laser photon
fluence and photolyte (i.e., phosgene) concentration, strongly
suggests that reactions 5 and 12 are, indeed, the only processes
that significantly influence the Cl time history.
-
7 -1
was determined to be 3.7 × 10
s
(lifetime ∼ 31 days). The
Cl atom experiments generally also lasted 3-4 h; the conversion
of the fluoromethanes, however, was around 30%. The recom-
mended value of the CH3F reaction rate coefficient with Cl
-13
3
-1 -1
25
atoms at 298 K is 3.5 × 10
cm molecule
s
at 298 K,
Typical data are shown in Figures 3 and 4, and measured
which places the average Cl atom concentration during the entire
bimolecular rate coefficients, kCl+CH F(p,T), along with relevant
3
7
-3
experiment around 6 × 10 atoms cm , which gives a ratio
experimental conditions, are summarized in Table 2. The results
from Table 2 are shown in the form of an Arrhenius plot in
Figure 5. The dashed line in Figure 5 is a fit to the data using
the modified Arrhenius expression:
>
60 between the loss due to reaction with Cl atoms and the
dark loss of fluoromethane. The dark reactions will therefore
not constitute an important systematic error in this system.
Experimental kinetic isotope effects for the Cl and OH
reactions of CH4, CH3Cl, and CH3F are summarized in Table
n
k
(T))A × (T/298) exp{-(E /RT)}
(13)
Cl+CH3F
a
1
2
13
1
. The C/ C KIEs in the Cl and OH reactions of CH3OH
A nonlinear least-squares analysis leads to the following best-
1
2
recently determined by Feilberg et al. have been included for
comparison. As can be seen, the C/ C KIE in the Cl reaction
with CH3OH is of similar magnitude as in CH3Cl, CH4, and
-12
3
-1 -1
fit parameters: A ) 1.14 × 10 cm molecule s , n ) 2.26,
1
2
13
and Ea/R ) 313 K.
As can be seen in Table 2 and Figure 5, the random errors
encountered in the determination of a rate coefficient at a given
temperature and pressure are relatively small. Reactive impuri-
ties contained in the CH3F could be a possible source of
systematic error and, in part, could be an explanation for the
non-Arrhenius temperature behavior of kCl+CH F(T) observed in
this study. In a previous study of this reaction, Manning and
reported an analysis of a CH3F sample from the same
supplier as used in the present study. Methane and C2H5F were
found at levels of 10 and 35 ppm, respectively. Using these
12
13
12
CH3F; for the OH reaction, the CH3OH/ CH3OH and CH3Cl/
1
3
CH3Cl KIEs are of similar magnitude as for CH3F but, again,
an order of magnitude larger than for CH4.
.1.2. LFP-RF Results. Photodissociation of Cl2CO at λ
235 nm has been shown to be a concerted process; i.e., both
chlorine atoms are produced on a time scale that is short
3
≈
3
84
compared to a rotational period. The fraction of chlorine atoms
generated in the spin-orbit excited state, Cl( P1/2), is thought
to be significant. Rate coefficients for the relaxation of
Cl( P1/2) in equilibrium with Cl( P3/2) are reported in the
literature for each of the components (N2, Cl2CO, CH3F) of the
experimental gas mixtures; given the values of those quenching
rate coefficients, it can be safely assumed that all Cl + CH3F
kinetic data reported in this study are representative of an
20,25
Kurylo
2
8
5
2
2
25
concentrations and the evaluated literature values for the
temperature-dependent rate coefficients for the reaction of Cl
atoms with each of the above impurities, it can be concluded
that neither impurity reaction made a significant contribution
to our reported values for kCl+CH3F(T). It is worth noting that
85
2
2
20
equilibrium mixture of Cl( P1/2) and ground-state Cl( P3/2) atoms.
the study of Manning and Kurylo was carried out using a

1
2422 J. Phys. Chem. A, Vol. 112, No. 48, 2008
Marinkovic et al.
theory. In addition to the saddle point of the hydrogen
abstraction reaction, we have located a prereaction van der
Waals adduct on the MEP for both reactions. On the product
side of the MEP there is a postreaction van der Waals adduct
between CH2F and HCl. These adducts have also been reported
2
8,33,34
in previous theoretical studies.
We were not able to
reproduce the postreaction adduct of the OH reaction with CH3F,
3
3
which is reported by El-Taher at the UHF/6-31G(d) level of
theory.
The stationary points on the MEP of the Cl and OH reactions
with CH3F were optimized at several levels of theory: MP2/
aug-cc-pVXZ (X ) D, T, Q) and CCSD(T)/aug-cc-pVDZ. The
structures of the stationary points of the Cl reaction with CH3F
are given in Figure S5 (Supporting Information), while the
structural parameters obtained at the different levels of theory
are listed in Table S2 (Supporting Information). Structures and
structural parameters for the stationary points of the OH reaction
are given in Figure S6 and Table S3 (Supporting Information),
respectively. The structural results indicate that the aug-cc-pVTZ
basis set provides a reasonable description of the stationary
points when compared to the larger aug-cc-pVQZ basis set; the
largest difference is 2.4 pm for the C-F bond in CH2F, which
Figure 3. Typical Cl atom temporal profiles observed at T ) 298 K
1
4
in 133 hPa of N
2
. Experimental conditions: [Cl
2
CO] ) 7.4 × 10
-3
3
-
3
11
molecules cm ; [Cl]
0
) 3.5 × 10 atoms cm ; [CH
molecules cm ) (A) 0, (B) 0.839, (C) 2.12, (D) 4.08, and (E)
.57. The lines are obtained from linear least-squares analyses of the
F] in units of
1
5
-3
1
7
0
data; their slopes give the following pseudo-first-order decay rates in
86
is within what is expected for the two levels. The structures
-
1
units of s : (A) 54, (B) 400, (C) 869, (D) 1650, and (E) 2930. For the
sake of clarity, traces B, C, D, and E are scaled by factors of 2, 1.75,
of the stationary points were also optimized at the CCSD(T)/
aug-cc-pVDZ level of theory to test the quality of the MP2
2
, and 0.75, respectively.
8
6
model, the CCSD(T) model being more robust. For the
reaction of OH with CH3F, the structures predicted by the MP2/
aug-cc-pVQZ and CCSD(T)/aug-cc-pVDZ model show only
minor differences (see Table S3, Supporting Information), the
largest difference concerning the C-H(4) bond (3.4 pm) of the
saddle point, leading to a slightly more product-like structure
of the saddle point in the CCSD(T) model compared to the MP2
model. The differences between the two models are more
pronounced for the Cl reaction with CH3F. As can be seen from
Table S2 and Figure S5 (Supporting Information), the CCSD(T)
model predicts the saddle point of the system to be much more
product-like than what is predicted by the MP2 model. It is
difficult to estimate the full effect of this as CCSD(T) calcula-
tions currently are too time-consuming to offer a realistic
alternative in computations of the potential energy surfaces
(PES) of the reactions investigated in this work. When compar-
ing the structural results for CH3F and OH as well as the reaction
products CH2F, HCl, and H2O with the experimental values
Figure 4. Plots of k′, the pseudo-first-order Cl atom decay rate, vs
CH F] over a range of temperatures. All data shown were obtained at
a total pressure of 133 hPa. The solid lines are obtained from linear
least-squares analyses, and the resulting bimolecular rate coefficients,
i.e., the slopes of the plots, are listed in Table 2.
(
Table S2 and S3, Supporting Information), it is seen that all
[
3
bond lengths are calculated within a range of less than (3.0
pm from the experimental values for the MP2/aug-cc-pVXZ
(X ) T, Q) and CCSD(T)/aug-cc-pVDZ levels of theory, in
87
agreement with the established error margins. The bond angles
of CH3F and H2O are described in close agreement with
experimental data (largest deviation -0.5°); the bond angle of
the CH2F radical deviates around -2.5° in the MP2 and
CCSD(T) calculations. The applied levels of theory thus provide
reasonable descriptions of the reactants and products.
technique very similar to the LFP-RF portion of this study,
although over a much smaller temperature range (216-296 K),
and that their results are almost identical with those reported in
this study. Taking into account the small amount of random
errors, the possibility of unidentified impurities (unlikely to be
important based on the above discussion) and other systematic
errors such as the determination of the temperature and of the
concentration of the reactant, we estimate that the overall
The energetics of the stationary points involved in the Cl and
OH reactions of CH3F are summarized in Table 3 and 4 and
compared to previous results in Tables S4 and S5 (Supporting
Information), respectively. The agreement with literature data
is generally good. The difference between the CCSD(T)/aug-
cc-pVQZ//MP2/aug-cc-pVQZ and the CCSD(T)/EB43//MP2/
accuracy of our measured values of kCl+CH F(T) is (20% (95%
3
confidence level) over the entire temperature range of the
LFP-RF study.
-
1
3.2. Computational Study. 3.2.1. Structures and Energetics
aug-cc-pVXZ (X ) T, Q) energies are less than 2 kJ mol for
both reactions, indicating that the results are close to the basis
set limit of the respective level of theory. The calculated
enthalpies of reaction for the OH and Cl reactions of CH3F are
in reasonable agreement with the experimental values. The
of Stationary Points. The minimum energy path (MEP) con-
necting reactants and products of the Cl and OH reaction with
CH3F was computed using the intrinsic reaction coordinate
6
3
(
IRC) method at the MP2/aug-cc-pVXZ (X ) D, T) levels of

Kinetic Isotope Effects in the Reactions of CH3F
J. Phys. Chem. A, Vol. 112, No. 48, 2008 12423
3
TABLE 2: Results of the LFP-RF Study of the Reaction of Cl Atoms with CH F
T/K p/hPa no. of expta [CH
F]max/1015 cm-3 [Cl
CO]/1014 cm-3 [Cl]
/10 cm
11
-3
fluence/mJ cm-2 k′max/s-1 (k ( 2σ) /10 cm s
b
13
3
-1
3
2
0
200
221
232
244
264
276
298
298
298
298
320
344
348
378
396
426
428
428
429
555
699
133
133
133
133
133
133
73
133
133
133
133
133
133
133
133
133
33
6
6
6
6
6
6
6
6
11
6
6
6
5
6
6
6
10
6
6
7.86
11.7
9.29
9.24
8.40
6.39
2.69
4.85
5.03
7.57
3.86
5.02
4.76
2.06
4.42
4.41
3.13
3.81
1.47
0.815
0.718
10
10
9.6
9.3
9.0
7.7
6.4
6.3
7.0
7.4
7.2-14
6.0
6.7
5.4
5.1
4.9
1.5
0.7-1.5
1.2
1.1
0.8-1.3
1
2.5
1
0.7-2.5
3.5
1.1
1.3
0.9
0.6
1
1.9
1.8
0.8
0.7
0.3-0.5
1.2
2.5
1.2-2.4
2.1
2.0
1.5-2.4
2.1
6.8
2.6
1.6-6.6
7.4
1.2-2.5
3.6
2.2
1.9
3.2
835
1698
1538
1905
2324
2149
1060
1841
2047
2929
2012
2982
2198
2047
4537
5205
3969
5270
2281
2160
3388
1.03 ( 0.08
1.40 ( 0.02
1.63 ( 0.14
2.02 ( 0.13
2.64 ( 0.16
3.26 ( 0.09
3.85 ( 0.10
3.53 ( 0.34
3.73 ( 0.21
3.75 ( 0.11
5.00 ( 0.33
5.80 ( 0.14
6.63 ( 0.22
9.62 ( 0.19
10.0 ( 0.4
11.5 ( 0.3
11.7 ( 0.8
13.6 ( 0.4
14.7 ( 1.0
26.2 ( 1.1
46.9 ( 2.6
6.4
2.4-10
2.5
2.2
1.7-3.1
7.1
3.2-12
5.4
4.9
2.7
133
133
133
133
7
6
2.7
a
Experiment defined as measurement of one pseudo-first-order decay rate. b Uncertainties represent precision only.
Cl reaction calculated by the ISPE method using single-point
energy corrections from the CCSD(T)/EB43//MP2/aug-cc-pVXZ
(
X ) T, Q) calculations and including the effect of spin-orbit
coupling. The generalized normal-mode frequencies of the
system as a function of the reaction coordinate are shown in
Figure 6b. The curves result from an interpolation of 470 ab
initio points with gradients and where the Hessians have been
computed at 160 of those points. The ∆VaG curve has two
maxima: one at -0.65 bohr (global) and the other at +0.32
bohr with a shallow minimum between the two maxima at ca.
+
0.19 bohr. One could assume that two dynamical bottlenecks
exist in the reaction system. Recrossing between such two
bottlenecks could occur and the assumption of one dividing
surface in CVT no longer would be valid. Examination of the
situation in the present reaction system shows that, at all
temperatures considered, the energy difference between the
second-highest maximum and the minimum is less than kBT,
kB being the Boltzmann constant. Therefore, recrossing is not
expected and the assumption of only a single dividing surface
holds.
3
Figure 5. Arrhenius plot for the reaction of Cl atoms with CH F: (b)
this work (LFP-RF technique), (s) Tschuikow-Roux et al. using
9
5
updated temperature-dependent rate coefficients for the Cl + CH
4
2
5
20
reference reaction, (- -) Manning and Kurylo, (0) Wallington et
4
4
23
21
al., (O) Tuazon et al.; and (4) Hitsuda et al. As described in the
text, a good fit to our data is obtained using the modified Arrhenius
-12
2.26
3
-1
expression k(T) ) 1.14 × 10 (T/298) exp(-313/T) cm molecule
Figure 7 shows an Arrhenius plot of the calculated rate
coefficients for the Cl reaction with CH3F. For comparison, the
figure includes plots of the experimental LFP-RF results of
-
1
s
(dotted line).
2
5,26
results obtained at the CCSD(T)/EB43//MP2/aug-cc-pVXZ (X
T, Q) and CBS-QB3 levels of theory are within the
this study and of the recommended rate coefficients,
)
including their uncertainty limits. Since spin-orbit coupling
present in the Cl atom becomes smaller during the reaction, it
will contribute to the potential energy surface by effectively
increasing the barrier height. This was taken into account in
the calculation of the rate coefficients. The calculations predict
the SO coupling to lower the asymptotic potential energy of
uncertainty of experimental values (Table S6, Supporting
Information). The experimental values show large uncertainty
limits due to the large uncertainties in the heat of formation of
CH3F and CH2F.
For the loosely bound reaction adducts, one might expect the
basis set superposition error (BSSE) to be significant. To
estimate the magnitude of this effect, the BSSE in the MP2/
aug-cc-pVQZ model of the pre- and postreaction adduct in the
chlorine reaction and the prereaction adduct in the OH reaction
-
1
the reactants on the MP2/aug-cc-pVTZ MEP by 3.3 kJ mol ,
in good agreement with the a priori assumption of 1/3 of the
-
1
SO coupling constant of Cl (3.5 kJ mol ) and previous
8
9
calculations on similar systems. For the prereaction adduct, a
8
8
-1
of CH3F was approximated by the counterpoise correction.
SO coupling of 1.1 kJ mol was found. At the transition state,
-1
-1
The BSSE was found to be around 1 kJ mol , which we
consider negligible in the present context.
the effect of SO coupling is almost quenched (0.1 kJ mol ).
The excited spin-orbit states were only considered in the
3
.2.2. Absolute Rate Coefficients. Figure 6a shows the
electronic partition function of the chlorine atom. In the ISPE
-
1
potential energy curve of the minimum energy path, VMEP, and
model, the barrier height was increased by 3.2 kJ mol
the ground-state vibrationally adiabatic potential energy curve,
according to the SO coupling calculated for the MP2/aug-cc-
pVTZ geometries. The reaction enthalpy and the energy of the
G
∆
Va , as a function of the reaction coordinate for the CH3F +

1
2424 J. Phys. Chem. A, Vol. 112, No. 48, 2008
Marinkovic et al.
TABLE 3: Born-Oppenheimer and Zero-Point Corrected Energies (kJ mol-1) of Stationary Points Involved in the Reaction
a
CH
3
F + Cl f CH
2
F + HCl
∆E‡
∆(E +ZPE)
‡
∆EPC
∆(EPC+ZPE)
∆
rxn
E
∆rxn(E+ZPE)
computational level
∆ERC
∆(ERC+ZPE)
this workb
MP2/aDZ
MP2/aTZ
MP2/aQZ
MP2/EB43
CCSD(T)/aDZ//MP2/aDZ
CCSD(T)/aTZ//MP2/aTZ
CCSD(T)/aQZ//MP2/aQZ
CCSD(T)/EB43//MP2/aXZ (X ) T, Q)
CCSD(T)/aDZ
c
-8.9
-8.4
-2.9
0.8
-9.2
-8.6
-8.5
-8.7
-9.2
-7.7
-7.4
-1.9
1.8
-8.0
-7.5
-7.5
-7.7
-9.2
29.3
24.7
28.2
31.5
28.6
24.7
22.3
21.3
28.7
12.1
7.8
11.4
14.6
11.4
7.9
5.5
4.5
10.1
5.2
-1.8
0.6
3.4
8.7
3.3
0.2
-1.2
8.6
-9.2
-17.0
-14.8
-12.6
-6.5
-11.8
-15.2
-16.6
-7.9
15.0
7.5
9.3
11.6
18.6
12.6
8.6
-4.8
-12.7
-11.1
-8.7
-1.2
-7.6
-11.7
-13.5
-1.4
6.8
18.5
a
RC is the energy of the prereaction complex, ∆E‡ is the barrier height evaluated at the saddle point, ∆EPC is the energy of the
∆
E
b
postreaction complex, and ∆rxnE is the energy of reaction. All energies are relative to that of the reactants. Structures of the stationary points
are displayed in Figure S5 (Supporting Information). Abbreviations: aDZ, aug-cc-pVDZ; aTZ, aug-cc-pVTZ; aQZ, aug-cc-pVQZ; EB,
c
extrapolated basis limit (see eq 3 and the following text).
TABLE 4: Born-Oppenheimer and Zero-Point Corrected Energies (kJ mol-1) of Stationary Points Involved in the Reaction
a
CH
3
F + OHf CH
computational level
this workb
2
2
F + H O
∆ERC
∆(ERC+ZPE)
∆E‡
∆(E +ZPE)
‡
∆
rxn
E
∆rxn(E+ZPE)
c
MP2/aDZ
-18.1
-12.4
-17.1
-20.8
-18.5
-18.0
-17.4
-17.2
-16.9
-12.1
-6.4
31.9
35.8
30.3
26.0
21.7
20.4
20.2
19.7
21.3
25.2
29.4
24.0
19.7
15.0
14.0
13.8
13.3
12.5
-78.7
-78.8
-87.5
-93.0
-59.6
-65.4
-69.3
-71.3
-59.6
-83.1
-83.8
-92.4
-97.9
-64.0
-70.0
-74.2
-76.2
-63.3
MP2/aTZ
MP2/aQZ
MP2/EB43
CCSD(T)/aDZ//MP2/aDZ
CCSD(T)/aTZ//MP2/aTZ
CCSD(T)/aQZ//MP2/aQZ
CCSD(T)/EB43//MP2/aXZ (X ) T, Q)
CCSD(T)/aDZ
-11.1
-14.8
-12.4
-12.0
-11.4
-11.2
-11.4
a
‡
∆ERC is the energy of the prereaction complex, ∆E is the Born-Oppenheimer barrier height evaluated at the saddle point, and ∆rxnE is the
b
Born-Oppenheimer energy of reaction. All energies are relative to that of the reactants. Structures of the stationary points are displayed in
Figure S6 (Supporting Information). Abbreviations: aDZ, aug-cc-pVDZ; aTZ, aug-cc-pVTZ; aQZ, aug-cc-pVQZ; EB, extrapolated basis (see
c
eq 3 and the following text).
postreaction adduct were also corrected for the effect of
spin-orbit coupling. This method of accounting for SO coupling
in dual-level calculations has proved to give reasonable results
higher temperature range. Thus, our results support a nonlinear
temperature dependence of the Arrhenius expression over the
whole temperature range of 200-700 K. The recommended
values of Sander et al. should only be used in the range of
approximately 290-370 K. An extrapolation to lower temper-
atures leads to significantly lower rate coefficients than those
measured in our study and those recommended by Atkinson et
al. (see Figure 7). The results of our theoretical study as well
8
9,90
for other reaction systems.
We see that our ICVT/SCT
calculation significantly underestimates the rate coefficients for
the reaction and does not reproduce the temperature dependence
correctly at low temperatures. However, there are some dis-
crepancies in the evaluation of experimental data leading to
different recommended rate coefficients and slopes of the
2
7
as that of the theoretical studies of Xiao et al. and Rosenman
2
6
29
Arrhenius plot. Atkinson et al. in their evaluation took only
and McKee predict a larger temperature dependence in the
2
0
the results of Manning and Kurylo into account, covering a
low-temperature range than was found experimentally in our
2
5
20
temperature range of 216-296 K, whereas Sander et al.
study and that by Manning and Kurylo.
included in their evaluation only the study of Tschuikow-Roux
et al., who found a steeper temperature dependence between
73 and 368 K. Atkinson et al. commented in their evaluation
There are two reasons for the somewhat short fallen calculated
rate coefficients compared with the experimental measurements:
First, the vibrational frequencies have been calculated within
the harmonic oscillator approximation, which is known to
overestimate the true frequencies. Second, the tunneling ap-
proximation we have usedsthe SCT modelsunderestimates the
true tunneling probability. To account for the deficiency in
frequencies, a scaling of all modes was applied within the IOE
model. The scaling of the frequencies results in almost no
changes of the rate constants and the temperature dependence.
For the TST model, the calculated rate constants reproduce the
experimental values within the uncertainty limits. A comparison
2
2
2
that Tschuikow-Roux et al. did not report individual data points
but only an Arrhenius expression that meets the other data at
2
98 K but has a greater temperature dependence in the
temperature range above 298 K. Sander et al. in their evaluation
slightly modified the results of Tschuikow-Roux et al. based
on comparisons of the Ea/R value of the CH3F + Cl reaction
with the Cl and OH reactions with CH3Cl and CH3Br. They
supposed that the results of Manning and Kurylo are affected
by some systematic error at lower temperatures, leading to a
less steep slope of the Arrhenius expression. Unfortunately, they
do not state the temperature range covered by their recom-
mended values. The results of our LFP-RF study covering a
temperature range of 200-700 K show good agreement with
both studies, that of Manning and Kurylo in the lower
temperature range and that of Tschuikow-Roux et al. in the
2
7
to the theoretical work of Xiao and co-workers shows the
sensitivity of the rate constants to the barrier height. In their
rate constants calculations, they applied the ICVT/SCT model
based on ab initio calculations of the PES at the QCISD(T)//
‡
BH&HLYP level of theory. An electronic barrier height, ∆E ,
-
1
of 20.6 kJ mol results in an underestimation of the rate

Kinetic Isotope Effects in the Reactions of CH3F
J. Phys. Chem. A, Vol. 112, No. 48, 2008 12425
Figure 7. Arrhenius plot of calculated and experimental rate coef-
ficients for the reaction CH
3
F + Cl f CH F + HCl. (4) This work,
2
ICVT/SCT with ISPE including SO coupling. (O) This work, ICVT/
SCT with IOE (scaled vibrational frequencies, scaling factor c )
0
.9552) including SO coupling. (2) This work, TST with ISPE
including SO coupling. (b) This work, TST with IOE (scaled vibrational
frequencies, scaling factor c ) 0.9552) including SO coupling. (Blue
line) This work, LFP-RF experimental measurements with uncertainty
2
5
limits. (Black line) Sander et al., evaluation of reference data with
22
uncertainty limits (based on the data of Tschuikow-Roux et al. ). (Red
2
6
line) Atkinson et al., evaluation of reference data with expanded
uncertainty limits corresponding approximately to a 95% confidence
2
0
level (based on the data of Manning and Kurylo ) and including the
value at 298 K with its uncertainty.
disagreement concerning the curvature of the Arrhenius plot.
1
3
15
Schmoltner et al. and DeMore suggest in their evaluations
Figure 6. (a) Minimum energy path, VMEP, and vibrationally adiabatic
2
6
G
Arrhenius plots without curvature, whereas Atkinson recom-
mends an Arrhenius plot with a small curvature. However, these
studies include only a temperature range of 240-480 K. A view
of previous theoretical studies shows a diverse picture. Most
of the studies fit the barrier to reaction to achieve agreement
between calculated and experimental rate constants. Electronic
ground-state potential energy curve, ∆V
a
, as a function of the reaction
F + HCl. The
coordinate, s, for the reaction CH
3
F + Cl f CH
2
contribution of spin-orbit coupling has been included. All quantities
are with respect to the reactants. (b) Generalized normal-mode
vibrational frequencies as a function of the reaction coordinate, s, for
the reaction CH
3
F + Cl f CH F + HCl.
2
3
2
32
35
32
31
barrier heights of 11.7, 12.8, 15.1, 19.2, 21.8, and 21.5
coefficients, whereas a barrier height of 11.4 kJ mol- gives
rate constants in agreement with experimental values. Both
barriers are lower than the one obtained in the present study
1
-
130
kJ mol
can be found. There is no agreement about the
treatment of the C · · ·H· · ·OH torsional mode as a hindered rotor
as well as about the contribution of tunneling. Albu et al. and
3
5
-
1
(
24.5 kJ mol including SO coupling).
3
2
The VMEP and ∆VaG potential energy curves of the OH
Lien et al. found in their studies small curvature tunneling
dominating the reaction, whereas in the study of Espinosa-Garc ´ı a
et al., large curvature tunnelling is dominating. Schwartz and
co-workers did not only fit the barrier height but also the
imaginary frequency of the saddle point to get results calculated
in agreement with experiments. The imaginary frequency
influences the barrier width and thus tunnelling. By adjustment
of the imaginary frequency, they fit the calculated curvature of
the Arrhenius plot to experimental rate constants.
.2.3. Kinetic Isotope Effects. The calculated C and H
kinetic isotope effects in the Cl reaction with CH F are given
in comparison to experimental data in Table 5. The correspond-
ing calculated KIEs in the temperature range of 200-1000 K
are available in Table S7 (Supporting Information). Our ICVT/
SCT calculations of the KIE( CH3F/ CH3F) and KIE(CH3F/
CD3F) are not able to reproduce the experimental values neither
with unscaled nor with scaled frequencies. The ICVT/SCT
calculations predict for the CH3F/CD3F KIE values that are
significantly too low, whereas the carbon-13 KIE is calculated
to be larger than the experimental value. In the TST calculations,
the obtained CH3F/CD3F KIE is significantly larger than the
reaction with CH3F at the MP2/aug-cc-pVTZ PES and the
corresponding curves corrected by the ISPE method are given
in Figure 8a; the generalized normal-mode frequencies are given
in Figure 8b. The Arrhenius plot of the calculated rate
coefficients in comparison to the experimental values for the
OH reaction with CH3F is shown in Figure 9. Like for the
reaction with Cl, the effect of spin-orbit coupling of the OH
radical was taken into account for the calculation of the rate
coefficients. The electronic barrier height was increased by 0.8
3
1
3
0
1
3
2
3
-1
kJ mol according to the calculated spin-orbit coupling at the
stationary points. In contrast to the calculations of the Cl reaction
with CH3F, where the harmonic oscillator approximation was
applied for all frequencies in the calculation of the rate
coefficients for the OH reaction, the internal C · · ·H· · ·OH
torsional mode was treated as a hindered rotor. As can be seen
from Figure 9, our TST and CVT/SCT calculations within the
ISPE model give rate coefficients in reasonable agreement with
experimental data. However, the temperature dependence of the
calculated rate coefficients is more pronounced compared to the
experimental values. There can be observed experimental
3
1
2
13

1
2426 J. Phys. Chem. A, Vol. 112, No. 48, 2008
Marinkovic et al.
Figure 9. Arrhenius plot of calculated and experimental rate coef-
ficients for the reaction CH
CVT/SCT with ISPE including SO coupling. (b) This work, TST with
ISPE including SO coupling. (Black line) Sander et al., evaluation of
reference data with uncertainty limits. (Blue line) Atkinson et al.,
3
F +OHf CH
2
F + H
2
O. (0) This work,
2
5
2
6
evaluation of reference data with expanded uncertainty limits corre-
sponding approximately to a 95% confidence level and including the
2
6
value at 298 K with its uncertainty. (Red line) Atkinson et al.,
evaluation of reference data using a three-parameter fit.
TABLE 5: Kinetic Isotope Effects, r ) klight/kheavy, in the
a
Reaction of CH
3
F with Cl Atoms at 298 K
ICVT/SCT
isotopologues
TST
10.7
1.013
1.533
2.850
unscaled
scaled
experimentb
CH
CH
3
F/CD
3
F
3.273
1.023
1.367
2.020
3.235
1.033
1.360
2.008
5.11 ( 0.07
1.016 ( 0.006
1
2
13
3
F/ CH
F/CH DF
F/CHD
3
F
CH
CH
3
2
3
2
F
a
The reaction rates are obtained by direct variational transition
state theory with interpolated single-point energies (ISPE) dynamics
calculations based on electronic structure information at the MP2/
aug-cc-pVTZ level of theory, as explained in the text. In addition,
the frequencies are scaled by a factor of 0.9552 within the IOE
model to correct the vibrational frequencies within the harmonic
approximation. KIEs calculated with conventional transition state
theory (TST) and improved canonical variational transition state
theory with small-curvature tunneling (ICVT/SCT) are compared
Figure 8. (a) Minimum energy path, VMEP, and vibrationally adiabatic
G
ground-state potential energy curve, ∆V
a
, as a function of the reaction
F + H O calculated
coordinate, s, for the reaction CH F + OHf CH
3
2
2
at the MP2/aug-cc-pVTZ level of theory without and with dual-level
reaction path dynamics by interpolated single-point energies (ISPE),
where the contribution of spin-orbit coupling has been included. All
quantities are with respect to the reactants. (b) Generalized normal-
mode vibrational frequencies as a function of the reaction coordinate,
b
with experimental measurements. This work.
s, for the reaction CH
3
F + OH f CH
2
2
F + H O.
(
Supporting Information). A large variational effect for all
1
2
13
experimental value, whereas the CH3F/ CH3F KIE is predicted
in good agreement with the experimental value.
temperatures can be found for the CH3F/CD3F KIE in the present
study: ηvar(200 K) ) 0.17, ηvar(298 K) ) 0.34, ηvar(500 K) )
0
To analyze the kinetic isotope effects in more detail, a factor
analysis as described in refs 91 and 92 was performed but
limited to investigations of the variational effect and the
tunneling contribution only:
.53. The tunneling contribution for the CH3F/CD3F KIE is
unexpectedly predicted ηtun < 1 (Table S10, Supporting
Information). The CH3F + Cl reaction is characterized by a
‡
low vibrationally adiabatic barrier height, ∆(E +ZPE) ) 7.6
-
1
kJ mol at the CCSD(T)/EB43//MP2/aug-cc-pVXZ (X ) T,
klight
KIE )
) ηtransη η η η ∝ η η
(10)
Q) level of theory including the effect of SO coupling. An
rot vib tun var
tun var
kheavy
G
inspection of the vibrationally adiabatic ground-state PES, ∆Va ,
The tunneling contribution is given as ηtun ) κSCT(light)/
(heavy), i.e., the ratio of the tunneling factors for the light
and the heavy isotopologue, and the variational effect is given
of the CD3F + Cl reaction shows that for the perdeuterated
reaction the effective vibrationally adiabatic barrier height is
higher compared to the perprotic barrier (see Figure S7,
SCT
κ
as ηvar ) KIE(
I)CVT
/KIE , i.e., the ratio of KIEs calculated
TST
Supporting Information). Additionally, the shape of ∆Va around
G
using variational and conventional transition state theory.
In the case of the CH3F + Cl reaction, the factor analysis
showed that both the variational effect and the tunneling
contribution are of similar magnitude for the C/ C KIE for
the whole temperature range, as can be seen from Table S9
the saddle point changes considerably. As a consequence, the
dynamic bottleneck is also shifted. The increase of ∆VaG for
the perdeuterated reaction is caused by the difference in ZPEs
of the transition state and reactants [∆ZPE ) (ZPE)TS -
(ZPE)reactants], which reduces the barrier for the CH3F + Cl
1
2
13

Kinetic Isotope Effects in the Reactions of CH3F
J. Phys. Chem. A, Vol. 112, No. 48, 2008 12427
Comparable to the Cl reaction for the 12C/ C KIE in the OH
reaction, only a small variational effect can be observed. The
tunneling contribution shows unexpected values of ηtun < 1 up
to 400 K (Table S13, Supporting Information). The variational
effect is large (ηvar(298 K) ) 0.41) for the CH3F/CD3F KIE,
13
TABLE 6: Kinetic Isotope Effects, r ) klight/kheavy, in the
a
3
Reaction of CH F with OH Radicals at 298 K
experimentb
isotopologues TST CVT/SCT
lit. values
c
d
e
CH
CH
3
F/CD
3
F
4.062
F 1.019
1.803
1.029
1.681
2.172
1.45, 1.57, 3.17 4.067 ( 0.018
1.067 ( 0.006
1
2
13
3
F/ CH
3
3
2
which was also found by Lien et al. In contrast to the Cl
reaction, the tunneling contribution for the OH reaction is ηtun
CH
CH
3
F/CH
2
DF 1.355
1.615
3
F/CHD
2
F
>
1 for temperatures up to 500 K (Table S14, Supporting
a
The reaction rates are obtained by direct variational transition state
Information). For the KIE(CH3F/CH2DF) and KIE(CH3F/
CHD2F), weak variational effects are predicted: ηvar(298 K) )
1.07 and 1.05, respectively (Tables S15 and S16, Supporting
Information). The tunneling contribution is predicted to be ηtun
theory with interpolated single-point energies (ISPE) dynamics
calculations based on electronic structure information at the MP2/
aug-cc-pVTZ level of theory, as explained in the text. KIEs calculated
with conventional transition state theory (TST) and canonical
variational transition state theory with small-curvature tunneling (CVT/
SCT) are compared with experimental measurements and literature
data. This work. Lien et al., CVT/µOMT, MP2, barrier height
1.7 kJ mol
kJ mol
mol
>
1 for temperatures up to 400 K, except for one abstraction in
the reaction of CH2DF with OH, where ηtun < 1 can be found
for 200 and 223 K. As discussed above, probably the tunneling
contribution is significantly underestimated for the CH3F
reaction.
b
c
32
-
1
d
32
1
.
Lien et al., CVT/µOMT, MP2, barrier height 12.8
-
1
e
32
.
Lien et al., CVT/µOMT, PM3, barrier height 19.2 kJ
-
1
.
reaction by 5.3 kJ mol- more than for the CD3F + Cl reactionG.
1
4. Conclusions
For comparison, in Figure S7 (Supporting Information) ∆Va
1
3
of the CH3F + Cl reaction is also displayed. As can be seen,
there is no change in the barrier height or in the dynamic
bottleneck. The changes in the ∆VaG and the shifted dynamic
bottleneck for the CD3F reaction result in a different vibrational
force field near the dynamic bottleneck and apparently leads to
larger tunneling probabilities for the C-D abstraction reaction
compared to the C-H abstraction reaction, although normally
one expects tunneling to be more important for the lighter
hydrogen than for deuterium. Obviously, the small curvature
tunneling approximation used in the present study is not able
to describe the tunneling properly; the “corner cutting effect”
for this reaction takes place over a larger part of the reaction
swath than the SCT approximation is able to describe.71,93,94
For the KIE(CH3F/CH2DF), a weak variational effect can be
observed for all temperatures: ηvar(298 K) ) 0.92 (Table S11,
Supporting Information). This effect is slightly increased for
the KIE(CH3F/CHD2F), where ηvar(298 K) ) 0.74 can be found
The results of the experimental study applying the laser flash
photolysis-resonance fluorescence technique to determine the
absolute rate coefficients for the reaction of CH3F with chlorine
in the temperature range of 200-700 K can be fitted to a
modified Arrhenius expression with significant curvature. Thus,
the experimental values are in good agreement with the previous
20
experimental investigations performed by Manning and Kurylo
2
2,95
between 216 and 296 K and by Tschuikow-Roux et al.
between 273 and 368 K, each of which gave an Arrhenius
expression with different temperature dependence. Our theoreti-
cal calculations of the rate coefficients support the observed
curvature of the Arrhenius plot. They show, however, a slightly
larger temperature dependence compared to the experimental
results.
As seen in Table 1, the 12C/ C KIE in the OH reaction of
CH3F is like that for the reactions of CH3Cl and CH3OH, also
an order of magnitude larger than for CH4. This can be attributed
to the difference in the barrier height of the C · · ·H· · ·OH
torsional mode. In the transition state of the OH reaction with
13
(
Table S12, Supporting Information). For both KIEs again an
unexpected tunneling contribution of ηtun < 1 is predicted for
two of the three possible abstraction reactions.
-
1
CH4, the torsional barrier is only 17 cm at the MP2/aug-cc-
pVDZ level of theory, whereas the barrier of the corresponding
mode in the transition states of the OH reaction with CH3F and
1
3
2
Table 6 gives the C and H KIEs in the OH reaction of
CH3F in comparison to experimental values. The corresponding
calculated KIEs in the temperature range of 200-1000 K are
available in Table S8 (Supporting Information). The experi-
mental value of the KIE(CH3F/CD3F) is significantly underes-
timated by our CVT/SCT calculation, whereas the TST calcu-
lation reproduces the experimental value excellently. For the
-1
CH3Cl are 454 and 303 cm , respectively. Since in the hindered
rotor scheme we have used the internal rotational axis going
through the center of mass of the whole system, the internal
12
13
rotational mode contributes to the C/ C kinetic isotope effect.
1
2
13
On the basis of the results from this work, the deuterium
kinetic isotope effects can be explained by variational effects
and tunneling. Our calculations are not able to reproduce the
observed values for the deuterium KIEs in the OH and Cl
reactions of CH3F. However, as we have pointed out already,
the tunneling method used in this work (SCT) is not able to
describe the tunneling effect properly. This is clearly seen in
the factor analysis for the perdeuterated species. We expect that
the use of a more sophisticated tunneling method will result in
better agreement with the experimental observations of the KIEs.
KIE( CH3F/ CH3F), both the CVT/SCT and TST models
predict too low calculated values compared to the experimental
value. In our calculations, the internal C · · ·H· · ·OH torsional
3
2
mode was treated as a hindered rotor. Lien et al., who in their
study also applied a hindered rotor treatment of the torsional
C· · ·H· · ·OH mode, calculated the CH3F/CD3F KIEs by ap-
plying dual-level variational transition state theory with adopted
barrier heights to fit experimental rate constants. In their CVT/
µOMT calculations based on the MP2/6-31+G* PES applying
-
1
a barrier of 11.7 and 12.8 kJ mol they calculated KIEs of
.45 and 1.57, respectively, at 300 K. Their calculations based
1
on a PM3 PES gave a KIE of 3.17 at 300 K. Further, the KIEs
obtained at the PM3 PES show a considerable temperature
dependence in the range of 200-1000 K, whereas for the cal-
culations based on the MP2 PES only a weak temperature
dependence is observed. Our calculations show a moderate
temperature dependence between 200 and 1000 K, as can be
seen from Table S8 (Supporting Information).
Acknowledgment. This work is part of the ACTION project
supported by the Norwegian Research Council under contract
160272/V30. The Norwegian High Performance Computing
Consortium is acknowledged for grants of computing time.
Research carried out at Georgia Tech was supported by the
National Aeronautics and Space Administration throught grants
NAGW-1001 and NNG06GD90G.

1
2428 J. Phys. Chem. A, Vol. 112, No. 48, 2008
Marinkovic et al.
Supporting Information Available: Experimental FTIR
(22) Tschuikow-Roux, E.; Yano, T.; Niedzielski, J. J. Chem. Phys. 1985,
1
3
82, 65.
spectrum of a CH3F/ CH3F/Cl2 reaction mixture (Figure S1).
(
23) Tuazon, E. C.; Atkinson, R.; Corchnoy, S. B. Int. J. Chem. Kinet.
992, 24, 639.
(24) Wallington, T. J.; Ball, J. C.; Nielsen, O. J.; Bartkiewicz, E. J.
Phys. Chem. 1992, 96, 1241.
25) Sander, S. P.; Friedl, R. R.; Golden, D. M.; Kurylo, M. J.; Moortgat,
13
Experimental FTIR spectrum of a CH3F/ CH3F/H2/O3 reaction
mixture (Figure S2). Experimental FTIR spectrum of a CH3F/
CD3F/Cl2 reaction mixture (Figure S3). The “natural” removal
of CH3F in the reactor (Figure S4). Structures of the stationary
points of the CH3F + Cl f CH2F + HCl reaction (Figure S5)
and the CH3F + OH f CH2F + H2O reaction (Figure S6).
Vibrationally adiabatic ground-state potential energy curves for
1
(
G. K.; Wine, P. H.; Ravishankara, A. R.; Kolb, C. E.; Molina, M. J.;
Finlayson-Pitts, B. J.; Huie, R. E.; Orkin, V. L. Chemical Kinetics and
Photochemical Data for Use in Atmospheric Studies. EValuation Number
1
5; National Aeronautics and Space Administration, Jet Propulsion
Laboratory; California Institute of Technology: Pasadena, CA, 2006.
26) Atkinson, R.; Baulch, D. L.; Cox, R. A.; Crowley, J. N.; Hampson,
1
3
the reactions of CH3F, CD3F and, CH3F with Cl with the
position of the bottlenecks of the reactions at 298 K indicated
(
R. F.; Hynes, R. G.; Jenkin, M. E.; Rossi, M. J.; Troe, J.; Wallington, T. J.
Atmos. Chem. Phys. 2008, 8, 4141.
(
Figure S7). Wavenumber regions and chemical components
included in spectral analyses of the different reaction mixtures
Table S1). Structural parameters of the stationary points of the
CH3F + Cl f CH2F + HCl reaction (Table S2) and the CH3F
OH f CH2F + H2O reaction (Table S3). Energetics of the
stationary points in the CH3F + Cl f CH2F + HCl reaction
Table S4) and the CH3F + OH f CH2F + H2O reaction (Table
(
27) Xiao, J.-F.; Li, Z.-S.; Liu, J.-Y.; Sheng, L.; Sun, C.-C. J. Comput.
Chem. 2002, 23, 1456.
28) Murray, C.; Retail, B.; Orr-Ewing, A. J. Chem. Phys. 2004, 301,
239.
(
(
(
(
29) Rosenman, E.; McKee, M. L. J. Am. Chem. Soc. 1997, 119, 9033.
30) Schwartz, M.; Marshall, P.; Berry, R. J.; Ehlers, C. J.; Petersson,
+
G. A. J. Phys. Chem. A 1998, 102, 10074.
(
(31) Espinosa-Garcia, J.; Coitino, E. L.; Gonzalez-Lafont, A.; Lluch,
J. M. J. Phys. Chem. A 1998, 102, 10715.
S5) in comparison to literature data. Enthalpies for the Cl and
OH reactions of CH3F in comparison to experimental values
(
32) Lien, P.-Y.; You, R.-M.; Hu, W.-P. J. Phys. Chem. A 2001, 105,
2
391.
(
2
Table S6). Kinetic isotope effects in the temperature range of
00-1000 K in the reaction of CH3F with Cl atoms (Table S7)
and OH radicals (Table S8). Factor analysis of the KIEs in the
(
33) El-Taher, S. Int. J. Quantum Chem. 2001, 84, 426.
(34) Korchowiec, J. J. Phys. Org. Chem. 2002, 15, 524.
(35) Albu, T. V.; Swaminathan, S. J. Phys. Chem. A 2006, 110, 7663.
(36) Aker, P. M.; Sloan, J. J. J. Chem. Phys. 1986, 85, 1412.
(37) Huang, Y.; Gu, Y.; Liu, C.; Yang, X.; Tao, Y. Chem. Phys. Lett.
1
2
13
reaction of CH3F with Cl atoms: KIE( CH3F/ CH3F) (Table
S9), KIE(CH3F/CD3F) (Table S10), KIE(CH3F/CH2DF) (Table
S11), and KIE(CH3F/CHD2F) (Table S12). Factor analysis of
1986, 127, 432.
(38) Streit, G. E.; Whitten, G. Z.; Johnston, H. S. Geophys. Res. Lett.
1976, 3, 521.
12
the KIEs in the reaction of CH3F with OH radicals: KIE( CH3F/
(
39) D’Ottone, L.; Bauer, D.; Campuzano-Jost, P.; Fardy, M.; Hynes,
13
CH3F) (Table S13), KIE(CH3F/CD3F) (Table S14), KIE(CH3F/
A. J. Phys. Chem. Chem. Phys. 2004, 6, 4276.
(40) Tyndall, G. S.; Orlando, J. J.; Kegley-Owen, C. S. J. Chem. Soc.,
Faraday Trans. 1995, 91, 3055.
CH2DF) (Table S15), and KIE(CH3F/CHD2F) (Table S16). This
material is available free of charge via the Internet at http://
pubs.acs.org.
(
(
41) Griffith, D. W. T. Appl. Spectrosc. 1996, 50, 59.
42) Rothman, L. S.; Jacquemart, D.; Barbe, A.; Benner, D. C.; Birk,
M.; Brown, L. R.; Carleer, M. R.; Chackerian, C.; Chance, K.; Coudert,
L. H.; Dana, V.; Devi, V. M.; Flaud, J. M.; Gamache, R. R.; Goldman, A.;
Hartmann, J. M.; Jucks, K. W.; Maki, A. G.; Mandin, J. Y.; Massie, S. T.;
Orphal, J.; Perrin, A.; Rinsland, C. P.; Smith, M. A. H.; Tennyson, J.;
Tolchenov, R. N.; Toth, R. A.; Vander Auwera, J.; Varanasi, P.; Wagner,
G. J. Quant. Spectrosc. Radiat. Transfer 2005, 96, 139.
References and Notes
(
1) Gierczak, T.; Talukdar, R. K.; Herndon, S.; Vaghjiani, G. L.;
Ravishankara, A. R. J. Phys. Chem. A 1997, 101, 3125.
2) Saueressig, G.; Crowley, J. N.; Bergamaschi, P.; Bruhl, C.;
Brenninkmeijer, C. A. M.; Fischer, H. J. Geophys. Res., [Atmos.] 2001,
06, 23127.
3) Feilberg, K. L.; Griffith, D. W. T.; Johnson, M. S.; Nielsen, C. J.
Int. J. Chem. Kinet. 2005, 37, 110.
4) D’Anna, B.; Bakken, V.; Beukes, J. A.; Nielsen, C. J.; Brudnik,
K.; Jodkowski, J. T. Phys. Chem. Chem. Phys. 2003, 5, 1790.
5) Feilberg, K. L.; Johnson, M. S.; Nielsen, C. J. J. Phys. Chem. A
004, 108, 7393.
6) Feilberg, K. L.; Sellevag, S. R.; Nielsen, C. J.; Griffith, D. W. T.;
Johnson, M. S. Phys. Chem. Chem. Phys. 2002, 4, 4687.
7) Feilberg, K. L.; Johnson, M. S.; Nielsen, C. J. Phys. Chem. Chem.
Phys. 2005, 7, 2318.
8) Gola, A. A.; D’Anna, B.; Feilberg, K. L.; Sellevag, S. R.; Bache-
Andreassen, L.; Nielsen, C. J. Atmos. Chem. Phys. 2005, 5, 2395.
9) Sellevåg, S. R.; Nyman, G.; Nielsen, C. J. J. Phys. Chem. A 2006,
10, 141.
10) Nilsson, E.; Nielsen, C. J.; Johnson, M. S. J. Phys. Chem. A, in
press.
(
(
(
43) York, D. Can. J. Phys 1966, 44, 1079.
44) Stickel, R. E.; Nicovich, J. M.; Wang, S.; Zhao, Z.; Wine, P. H. J.
1
Phys. Chem. 1992, 96, 9875.
45) Nicovich, J. M.; Wang, S.; Wine, P. H. Int. J. Chem. Kinet. 1995,
7, 359.
46) Nicovich, J. M.; Wang, S.; McKee, M. L.; Wine, P. H. J. Phys.
Chem. 1996, 100, 680.
47) Bilde, M.; Sehested, J.; Nielsen, O. J.; Wallington, T. J.; Meagher,
(
(
2
(
(
(
(
2
R. J.; McIntosh, M. E.; Piety, C. A.; Nicovich, J. M.; Wine, P. H. J. Phys.
Chem. A 1997, 101, 8035.
(
(
48) Ayhens, Y. V.; Nicovich, J. M.; McKee, M. L.; Wine, P. H. J.
Phys. Chem. A 1997, 101, 9382.
49) Piety, C. A.; Soller, R.; Nicovich, J. M.; McKee, M. L.; Wine,
P. H. Chem. Phys. 1998, 231, 155.
50) Finlayson-Pitts, B. J.; Pitts, J. N. J. Chemistry of the Upper and
Lower Atmosphere; Academic Press: San Diego, 2000; p 146.
(
(
(
(
(
1
(
51) Møller, C.; Plesset, M. S. Phys. ReV. 1934, 46, 618.
(
(52) Raghavachari, K.; Trucks, G. W.; Pople, J. A.; Head-Gordon, M.
Chem. Phys. Lett. 1989, 157, 479.
(
(
11) Bilde, M.; Wallington, T. J. J. Phys. Chem. A 1998, 102, 1550.
12) Feilberg, L. K.; Gruber-Stadler, M.; Johnson, M. S.; M u¨ hlh a¨ user,
(
53) Montgomery, J. A., Jr.; Frisch, M. J.; Ochterski, J. W.; Petersson,
G. A. J. Chem. Phys. 1999, 110, 2822.
(54) Montgomery, J. A., Jr.; Frisch, M. J.; Ochterski, J. W.; Petersson,
G. A. J. Chem. Phys. 2000, 112, 6532.
M.; Nielsen, C. J. J. Phys. Chem. A. [Online early access]. DOI: 10.1021/
jp805643x. Published Online: Oct 9, 2008.
(
13) Schmoltner, A. M.; Talukdar, R. K.; Warren, R. F.; Mellouki, A.;
(55) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.;
Burant,J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci,
B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.;
Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.;
Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox,
J. E.; Hratchian, H. P.; Cross, J. B.; Adamo, C.; Jaramillo, J.; Gomperts,
R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.;
Ochterski, J. W.; Ayala, P. Y.; Morokuma, K. ; Voth, G. A.; Salvador, P.;
Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain,
M. C. ; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.;
Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford,S.; Cioslowski,
J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.;
Goldfarb, L.; Gierczak, T.; McKeen, S. A.; Ravishankara, A. R. J. Phys.
Chem. 1993, 97, 8976.
(
14) Nip, W. S.; Singleton, D. L.; Overend, R.; Paraskevopoulos, G. J.
Phys. Chem. 1979, 83, 2440.
(
(
(
(
(
15) DeMore, W. B. J. Phys. Chem. 1996, 100, 5813.
16) Hsu, K. J.; DeMore, W. B. J. Phys. Chem. 1995, 99, 1235.
17) Howard, C. J.; Evenson, K. M. J. Chem. Phys. 1976, 64, 197.
18) Jeong, K. M.; Kaufman, F. J. Phys. Chem. 1982, 86, 1808.
19) Wallington, T. J.; Hurley, M. D. EnViron. Sci. Technol. 1993, 27,
1
448.
(
(
20) Manning, R. G.; Kurylo, M. J. J. Phys. Chem. 1977, 81, 291.
21) Hitsuda, K.; Takahashi, K.; Matsumi, Y.; Wallington, T. J. J. Phys.
Chem. A 2001, 105, 5131.

Kinetic Isotope Effects in the Reactions of CH3F
J. Phys. Chem. A, Vol. 112, No. 48, 2008 12429
Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen,
W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, Revision B.03;
Gaussian, Inc.: Pittsburgh, PA, 2003.
(75) Chuang, Y.-Y.; Truhlar, D. G. J. Chem. Phys. 2006, 124, 179903/
1.
(76) Computational Chemistry Comparison and Benchmark DataBase
(CCCBDB), NIST Standard Reference Database 101; National Institute of
Standards and Technology, 2005.
(77) Wallington, T. J.; Hurley, M. D.; Schneider, W. F.; Sehested, J.;
Nielsen, O. J. Chem. Phys. Lett. 1994, 218, 34.
(78) MCPA Software Ltd. FACSIMILE for Windows, Version 4.0.31;
MCPA Software Ltd.
(79) Gola, A. A.; D’Anna, B.; Feilberg, K. L.; Sellevåg, S. R.; Bache-
Andreassen, L.; Nielsen, C. J. Atmos. Chem. Phys. Discuss. 2005, 5, 3873.
(80) Atkinson, R.; Baulch, D. L.; Cox, R. A.; Crowley, J. N.; Hampson,
R. F.; Hynes, R. G.; Jenkin, M. E.; Rossi, M. J.; Troe, J. Atmos. Chem.
Phys. 2006, 6, 3625.
(81) Meagher, R. J.; McIntosh, M. E.; Hurley, M. D.; Wallington, T. J.
Int. J. Chem. Kinet. 1997, 29, 619.
(
56) Amos, R. D.; Bernhardsson, A.; Berning, A.; Celani, P.; Cooper,
D. L.; Deegan, M. J. O.; Dobbyn, A. J.; Eckert, F.; Hampel, C.; Hetzer,
G.; Knowles, P. J.; Korona, T.; Lindh, R.; Lloyd, A. W.; McNicholas, S. J.;
Manby, F. R.; Meyer, W.; Mura, M. E.; Nicklass,A.; Palmieri, P.; Pitzer,
R.; Rauhut, G.; Sch u¨ tz, M.; Schumann, U.; Stoll, H.; Stone, A. J.; Tarroni,
R.; Thorsteinsson, T.; Werner, H.-J. MOLPRO, a package of ab initio
programs designed by H.-J. Werner and P. J. Knowles, version 2002.6;
2
9
1
8
002.
(
57) Knowles, P. J.; Hampel, C.; Werner, H. J. J. Chem. Phys. 1993,
9, 5219.
(
58) Knowles, P. J.; Hampel, C.; Werner, H.-J. J. Chem. Phys. 2000,
12, 3106.
(
59) Watts, J. D.; Gauss, J.; Bartlett, R. J. J. Chem. Phys. 1993, 98,
718.
(82) Bednarek, G.; Arguello, G. A.; Zellner, R. Ber. Der Bunsen-Ges.-
Phys. Chem. Chem. Phys. 1996, 100, 445.
(
60) Dunning, T.H., Jr. J. Chem. Phys. 1989, 90, 1007.
61) Kendall, R. A.; Dunning, T. H., Jr.; Harrison, R. J. J. Chem. Phys.
(
83) Bednarek, G.; Breil, M.; Hoffmann, A.; Kohlmann, J. P.; Mors,
(
V.; Zellner, R. Ber. Der Bunsen-Ges.-Phys. Chem. Chem. Phys. 1996, 100,
1
992, 96, 6796.
62) Halkier, A.; Helgaker, T.; Jorgensen, P.; Klopper, W.; Koch, H.;
5
28.
(
(
84) Maul, C.; Haas, T.; Gericke, K.-H.; Comes, F. J. J. Chem. Phys.
Olsen, J.; Wilson, A. K. Chem. Phys. Lett. 1998, 286, 243.
1
995, 102, 3238.
(
(
63) Gonzalez, C.; Schlegel, H. B. J. Phys. Chem. 1990, 94, 5523.
64) Corchado, J. C.; Coitino, E. L.; Chuang, Y.-Y.; Fast, P. L.; Truhlar,
(
85) Chichinin, A. I. J. Phys. Chem. Ref. Data 2006, 35, 869, and
references therein.
86) Helgaker, T.; Jørgensen, P.; Olsen, J. Molecular Electronic-Structure
Theory; John Wiley & Sons Ltd: Chichester, 2000.
87) Helgaker, T.; Gauss, J.; Joergensen, P.; Olsen, J. J. Chem. Phys.
997, 106, 6430.
88) Simon, S.; Duran, M.; Dannenberg, J. J. J. Chem. Phys. 1996, 105,
1024.
D. G. J. Phys. Chem. A 1998, 102, 2424.
65) Gonzalez-Lafont, A.; Villa, J.; Lluch, J. M.; Bertran, J.; Steckler,
R.; Truhlar, D. G. J. Phys. Chem. A 1998, 102, 3420.
66) Chuang, Y.-Y.; Corchado, J. C.; Truhlar, D. G. J. Phys. Chem. A
(
(
(
(
1
1
4
2
7
2
8
1
999, 103, 1140.
(
(
(
67) Chuang, Y.-Y.; Truhlar, D. G. J. Phys. Chem. A 1997, 101, 3808.
68) Corchado, J. C.; Chuang, Y.-Y.; Fast, P. L.; Vill a` , J.; Hu, W.-P.;
(
89) Capecehi, G.; Werner, H. J. Phys. Chem. Chem. Phys. 2004, 6,
Liu, Y.-P.; Lynch, G. C.; Nguyen, K. A.; Jackels, C. F.; Melissas, V. S.;
Lynch, B. J.; Rossi, I.; Coiti n˜ o, E. L.; Fernandez-Ramos, A.; Ellingson,
B. A.; Pu, J.; Albu, T. V.; Zheng, J.; Steckler, R.; Garrett, B. C.; Isaacson,
A. D.; Truhlar, D. G. POLYRATE-Version 9.6; University of Minnesota:
Minneapolis, MN, 2007.
975.
(90) Manthe, U.; Capecchi, G.; Werner, H. J. Phys. Chem. Chem. Phys.
004, 6, 5026.
(
91) Joseph, T.; Steckler, R.; Truhlar, D. G. J. Chem. Phys. 1987, 87,
036.
(
69) Isaacson, A. D.; Sund, M. T.; Rai, S. N.; Truhlar, D. G. J. Chem.
(92) Corchado, J. C.; Truhlar, D. G.; Espinosa-Garcia, J. J. Chem. Phys.
Phys. 1985, 82, 1338.
000, 112, 9375.
(
(
70) Garrett, B. C.; Truhlar, D. G. J. Phys. Chem. 1979, 83, 1052.
71) Liu, Y.-P.; Lynch, G. C.; Truong, T. N.; Lu, D.-H.; Truhlar, D. C.;
(
93) Skodje, R. T.; Truhlar, D. G.; Garrett, B. C. J. Phys. Chem. 1981,
5, 3019.
94) Truhlar, D. G.; Isaacson, A. D.; Garrett, B. C. Generalized transition
Garrett, B. C. J. Am. Chem. Soc. 1993, 115, 2408.
72) Garrett, B. C.; Truhlar, D. G.; Grev, R. S.; Magnuson, A. W. J.
Phys. Chem. 1980, 84, 1730.
73) Fast, P. L.; Corchado, J. C.; Truhlar, D. G. J. Chem. Phys. 1998,
09, 6237.
74) Berning, A.; Schweizer, M.; Werner, H. J.; Knowles, P. J.; Palmieri,
P. Mol. Phys. 2000, 98, 1823.
(
(
state theory In The Theory of Chemical Reaction Dynamics; Baer, M. ,
Ed.; CRC Press: Boca Raton, FL, 1985;Vol 4, pp 65.
(
(
95) Tschuikow-Roux, E.; Faraji, F.; Paddison, S.; Niedzielski, J.;
1
Miyokawa, K. J. Phys. Chem. 1988, 92, 1488.
(
JP807609D