Rate coefficients for the reactions of OH with CF3CH 2CH3 (HFC-263fb), CF3CHFCH2F (HFC-245eb), and CHF2CHFCHF2 (HFC-245ea) between 238 and 375 K

Article abstract of DOI:10.1021/jp056248y

Rate coefficients for reaction of the hydroxyl radical (OH) with three hydrofluorocarbons (HFCs) CF₃CH₂-CH₃, HFC-263fb, (k₁); CF₃CHFCH₂F, HFC-245eb, (k ₂); and CHF₂CHFCHF₂, HFC-245ea, (k ₃); which are suggested as potential substitutes to chlorofluorocarbons (CFCs), were measured using pulsed laser photolysis-laser-induced fluorescence (PLP-LIF) between 235 and 375 K. The Arrhenius expressions obtained are k₁(T) = (4.36 ± 0.72) × 10^-12 exp[-(1290 ± 40)/T] cm³ molecule ^-1 s^-1; k₂(T) = (1.23 ± 0.18) × 10^-12 exp[-(1250 ± 40)/T] cm³molecule^-1 s^-1; k₃(7) = (1.91 ± 0.42) × 10^-12 exp[-(1375 ± 100)/T] cm³ molecule^-1 s ^-1. The quoted uncertainties are 95% confidence limits and include estimated systematic errors. The present results are discussed and compared with rate coefficients available in the literature. Our results are also compared with those calculated using structure activity relationships (SAR) for fluorinated compounds. The IR absorption cross-sections at room temperature for these compounds were measured over the range of 500 to 4000 cm^-1. The global warming potentials (GWPs) of CF₃CH₂CH ₃(HFC-263fb), CF₃CHFCH₂F(HFC-245eb), and CHF₂CHFCHF₂(HFC-245ea) were calculated to be 234, 962, and 723 for a 20-year horizon; 70, 286, and 215 for a 100-year horizon; and 22, 89, and 68 for a 500-year horizon; and the atmospheric lifetimes of these compounds are 0.8, 2.5, and 2.6 years, respectively. It is concluded that these compounds are acceptable substitutes for CFCs in terms of their impact on Earth's climate.

Full text of DOI:10.1021/jp056248y

6724
J. Phys. Chem. A 2006, 110, 6724-6731
Rate Coefficients for the Reactions of OH with CF3CH2CH3 (HFC-263fb), CF3CHFCH2F
(HFC-245eb), and CHF2CHFCHF2 (HFC-245ea) between 238 and 375 K†
B. Rajakumar,|,‡ R. W. Portmann,| James B. Burkholder,| and A. R. Ravishankara*,|,§
National Oceanic and Atmospheric Administration, Earth System Research Laboratory, 325 Broadway,
Boulder, Colorado 80305, and The CooperatiVe Institute for Research in EnVironmental Sciences,
UniVersity of Colorado, Boulder, Colorado 80309
ReceiVed: October 30, 2005; In Final Form: December 25, 2005
Rate coefficients for reaction of the hydroxyl radical (OH) with three hydrofluorocarbons (HFCs) CF3CH2-
CH3, HFC-263fb, (k1); CF3CHFCH2F, HFC-245eb, (k2); and CHF2CHFCHF2, HFC-245ea, (k3); which are
suggested as potential substitutes to chlorofluorocarbons (CFCs), were measured using pulsed laser photolysis-
laser-induced fluorescence (PLP-LIF) between 235 and 375 K. The Arrhenius expressions obtained are
k1(T) ) (4.36 ( 0.72) × 10-12 exp[-(1290 ( 40)/T] cm3 molecule-1 s-1; k2(T) ) (1.23 ( 0.18) × 10-12
exp[-(1250 ( 40)/T] cm3 molecule-1 s-1; k3(T) ) (1.91 ( 0.42) × 10-12 exp[-(1375 ( 100)/T] cm3
molecule-1 s-1. The quoted uncertainties are 95% confidence limits and include estimated systematic errors.
The present results are discussed and compared with rate coefficients available in the literature. Our results
are also compared with those calculated using structure activity relationships (SAR) for fluorinated compounds.
The IR absorption cross-sections at room temperature for these compounds were measured over the range of
500 to 4000 cm-1. The global warming potentials (GWPs) of CF3CH2CH3(HFC-263fb), CF3CHFCH2F(HFC-
245eb), and CHF2CHFCHF2(HFC-245ea) were calculated to be 234, 962, and 723 for a 20-year horizon; 70,
286, and 215 for a 100-year horizon; and 22, 89, and 68 for a 500-year horizon; and the atmospheric lifetimes
of these compounds are 0.8, 2.5, and 2.6 years, respectively. It is concluded that these compounds are acceptable
substitutes for CFCs in terms of their impact on Earth’s climate.
Introduction
are among a class of compounds that are potential substitutes
to CFCs.1 The rate coefficients for the reaction of these
compounds with OH, to the best of our knowledge, have only
been reported by Nelson et al.,3 and only at room temperature.
A theoretical investigation by Percival et al.4 has reported the
temperature dependence of the title reactions. The lack of
experimentally determined rate coefficients at atmospheric
temperatures has led us to the present investigation.
Chlorofluorocarbons (CFCs) have been shown to be detri-
mental to Earth’s stratospheric ozone layer.1 Therefore, the
Montreal protocol2 and its amendments have been accepted
internationally. This decision has prompted industry to look for
alternatives to CFCs in various applications.1 Partially fluori-
nated hydrocarbons (HFCs) are among the leading environ-
mentally acceptable CFC alternatives from the point of view
of ozone depletion. However, HFCs absorb strongly in the IR
region of the Earth’s outgoing radiation. Therefore, for HFCs
to be environmentally acceptable, their contribution to the
greenhouse effect needs to be small. Quantification of the
possible role of HFCs as greenhouse gases requires accurate
information on their lifetimes and IR absorption cross-sections,
which are key parameters in determining their global warming
potentials (GWPs), a relative index of the greenhouse potential
of a gas.
In this paper, we present the rate coefficient for the reaction
of OH radicals with CF3CH2CH3, CF3CHFCH2F and
CHF2CHFCHF2 between 238 and 375 K.
k1
OH + CF3CH2CH3
9
8 products
(1)
(2)
(3)
k2
9
OH + CF3CHFCH2F
8 products
k3
Because of the presence of one or more C-H bonds, HFCs
react with tropospheric hydroxyl radicals, resulting in atmo-
spheric lifetimes shorter than those of CFCs. Yet, their reactivity
with OH is sufficiently slow that their atmospheric lifetimes
can be years.
OH + CHF2CHFCHF2
98 products
In addition, we report the IR absorption cross-sections, atmo-
spheric lifetimes and GWPs for these compounds.
The three HFCs studied here, CF3CH2CH3 (HFC-263fb),
CF3CHFCH2F (HFC-245eb), and CHF2CHFCHF2 (HFC-245ea),
Experimental Section
In this work, the rate coefficients for the title reactions,
k1(T), k2(T), and k3(T) were determined in the temperature range
238-375 K. Also, infrared absorption cross-sections for
CF3CH2CH3, CF3CHFCH2F, and CHF2CHFCHF2 were deter-
mined at 298 K. We have described the apparatus, data
acquisition methods, and data analysis procedures for measuring
rate coefficients for reactions with OH5 in previous publications.
† Part of the special issue “David M. Golden Festschrift”.
* Correspondence to A. R. Ravishankara (A. R.Ravishankara@noaa.gov);
NOAA, Earth System Research Laboratory, Chemical Sciences Division,
R/CSD 2, 325 Broadway, Boulder, CO 80305.
| National Oceanic and Atmospheric Administration.
‡ CIRES, University of Colorado.
§ Also affiliated with the Department of Chemistry and Biochemistry,
University of Colorado, Boulder, CO 80309.
10.1021/jp056248y CCC: $33.50 © 2006 American Chemical Society
Published on Web 03/04/2006

Rate Coefficients for Reaction of OH with HFCs
J. Phys. Chem. A, Vol. 110, No. 21, 2006 6725
Therefore, we only briefly describe here the essentials needed
to understand the present investigation.
calibrated electronic mass flow meters. Flow meters were
calibrated independently for each reactant. The reactant con-
centration in the gas flow was also measured by infrared
absorption at 298 K as described below. The reactant concentra-
tion in the reactor was corrected for the differences in temper-
ature and pressure and for dilution. The concentrations of
CF3CH2CH2, CF3CHFCH2F, and CHF2CHFCHF2 in the reactor
were varied in the ranges (2.8-44) × 1014, (2.2-54) × 1014,
and (2.3-66) × 1014 molecule cm-3, respectively. The con-
centration determined using flow rates agreed with that from
infrared absorption measurements to within (5% under all
experimental conditions.
Rate Coefficient Measurements. The rate coefficients k1-
k3 were determined under pseudo-first-order conditions in OH
concentration using pulsed laser photolysis (PLP) production
of OH and its laser-induced fluorescence (LIF) detection. The
LIF reactor consisted of a 15-cm-long jacketed Pyrex cell with
an internal volume of ∼200 cm3. Orthogonal ports on the reactor
were used to propagate the laser beams. The photomultiplier
tube (PMT) detector was mounted on a port orthogonal to the
laser beams. The LIF reactor temperature was regulated to (1
K by circulating heated/cooled silicon oil through its jacket.
The gas mixture entering the reactor attained the temperature
of the reactor before reaching the reaction zone, defined by the
intersection of the photolysis beam and probe laser beam. The
temperature of the gases in the reaction zone was directly
measured with a calibrated thermocouple inserted into the gas
flow. The thermocouple was withdrawn from the detection
region while measuring the OH temporal profiles. Pressures in
the LIF reactor were measured with a 10, 100, or 1000 Torr
capacitance manometer.
Infrared Absorption Cross-Section Measurements. Infra-
red (IR) absorption cross-sections of the HFCs were measured
using a Fourier transform infrared spectrometer (FTIR) equipped
with a 15-cm-long, 2.5-cm-diameter Pyrex absorption cell with
KBr windows. An incandescent light source, KBr beam splitter,
and liquid N2 cooled HgCdTe (MCT) detector were used.
Spectra were recorded at a spectral resolution of 1 cm-1 over
the range 500 to 4000 cm-1 with 100 coadded scans. The IR
cross-sections were measured by filling the absorption cell with
1-3 Torr of the compound measured using a 10 Torr capaci-
tance manometer. A minimum of 6 concentrations were used
for each compound. The peak absorbance for each band varied
linearly with concentration, i.e., obeyed Beer-Lambert’s law.
Peak absorption cross-sections were obtained using linear least-
squares fits of absorbance versus concentration. The absorption
spectra over the entire wavelength region were converted to
cross-sections using the cross-sections derived at the peaks. The
cross-sections derived from different peak cross-sections agreed
within 5%. Spectrum measurements were also made with the
absorption cell pressurized with 100 Torr of N2 to investigate
the broadening of the HFC bands; the bands were not broadened
at this pressure. Therefore, the IR cross-sections obtained using
the samples alone are applicable under atmospheric pressures.
Rate coefficients were measured at total pressures in the range
of 50 to 60 Torr (He) with linear gas flow velocities in the LIF
reactor of 10 to 20 cm s-1. This flow provided a fresh gas
mixture for each photolysis laser pulse (10 Hz).
OH radicals were produced by 248 nm excimer (KrF) laser
photolysis of H2O2
H2O2 + hν f OH + OH
(4)
where the quantum yield for OH production in reaction 4 is 2.
OH radicals were excited in the A2Σ+ r X2Π band (282
nm) using the frequency-doubled output of a pulsed Nd:YAG
pumped dye laser. The laser-induced fluorescence was detected
by a PMT after it passed through a band-pass filter (peak
transmission at 310 nm with a band-pass of (20 nm, fwhm).
The PMT signal was averaged at various reaction times, ranging
from 10 µs to 50 ms, with a gated charge integrator and recorded
for subsequent analysis.
The initial OH radical concentration, [OH]0, was estimated
from the measured laser fluence (varied over the range 1.0-
6.8 mJ cm-2 pulse-1), the absorption cross-section of the
precursor at the photolysis wavelength, its OH quantum yield,
and the precursor concentration. [OH]0 values in the range 1.2
× 1011 to 7.5 × 1011 molecule cm-3 were used over the course
of the kinetic measurements.
Materials. He (99.999%) was used as supplied as the buffer
gas in all the kinetic measurements. Nitrogen (>99.99%) was
used as supplied in the measurement of IR absorption cross-
sections. Concentrated hydrogen peroxide (>95%, by mole
fraction, as determined by titration with a standard solution of
KMnO4) was prepared by bubbling N2 for several days through
a H2O2 sample which was initially ∼60% by mole fraction. A
small flow of He, approximately 1% of the total gas flowing
through the reactor, was passed through a bubbler containing
the >95% pure liquid H2O2. This mixture from the bubbler was
added to the main gas flow before entering the LIF reactor.
During low-temperature kinetic measurements, the H2O2 res-
ervoir was maintained at a temperature lower than that of the
LIF reactor to avoid possible condensation of H2O2 in the cold
reactor. Ozone was prepared by passing O2 through a com-
mercial ozonizer and stored on a silica gel at 195 K. A dilute
mixture of ozone in He was prepared in a darkened Pyrex bulb
from this sample.
The concentration of H2O2 in the LIF reactor was estimated
from the first-order rate coefficient for loss of OH measured in
the absence of the reactant and attributed to the reaction of OH
with H2O2
OH + H2O2 f HO2 + H2O
(5)
where k5(T) ) 2.9 × 10-12 exp(-160/T) cm3 molecule-1 s-1 6
.
Impurities in Excess Reactant. The rate coefficients for
reactions 1-3 are relatively small, <2 × 10-13 cm3 molecule-1
s-1, and hence, reactive impurities can lead to systematic errors
in the measured values. Unsaturated hydrocarbon impurities,
which can react with OH with rate coefficients of up to 3 orders
of magnitude faster than the HFCs, are therefore of concern.
The HFC samples were analyzed by gas chromatography-mass
spectroscopy (GCMS) using a capillary GS-Al column. No
measurable impurities were observed in the CF3CH2CH3 sam-
ple. The CF3CHFCH2F sample was found to have trace
amounts, <100 ppmv, of CF3CHFCHF2, CF3CH2CH2CF3, and
The concentration of H2O2 in the LIF reactor was varied between
1.6 × 1013 and 1.2 × 1014 molecule cm-3 over the course of
the kinetic measurements but was maintained constant during
each individual rate coefficient determination. The ozone
concentration in the LIF reactor, typically 1 × 1012 molecule
cm-3, was determined from measured flow rates and pressures.
The concentrations of the HFC reactants were determined
by two independent methods. It should be noted that the
concentrations of the reactant in the reactor were not directly
measured. The reactant concentration prior to entering the
reactor was determined from the flow rate measured using

6726 J. Phys. Chem. A, Vol. 110, No. 21, 2006
Rajakumar et al.
CF3CHFCH3. The CHF2CHFCHF2 sample was found to have
CF3CHdCF2. We have estimated the CF3CHdCF2 abundance
to be ∼50 ppmv by assuming a mass detector response similar
to that of CH3CHdCH2. This level of impurity would increase
the observed k3 value by ∼5% at 298 K and ∼20% at 238 K.
We have also checked for the presence of unsaturated hydro-
carbons using a purification test, which consisted of passing
the reactant through a column containing sulfuric acid (20 wt
% deposited on Chromosorb WHP). Sulfuric acid would remove
unsaturated compounds from the sample. For the rate coefficient
measurements, the reactant samples were introduced into the
gas flow after passing through the calibrated mass flow meters.
The small impurity levels of the samples do not contribute
significant error to the measured IR absorption cross-sections.
Results and Discussions
All rate coefficients were measured under pseudo-first-order
conditions in OH with [reactant] g 1000[OH]0. Under these
conditions, the OH temporal profiles followed the pseudo-first-
order rate law
ln[OH]t ) ln[OH]0 - k′t
(6a)
or
ln(St) ) ln(S0) - k′t
(6b)
where k′ () k1[reactant] + kd′) is the pseudo-first-order rate
coefficient for the reaction of OH with the reactant, and St and
S0 are the LIF signals from OH at time t and time zero,
respectively. kd′ is the pseudo-first-order rate coefficient for loss
of OH in the absence of the reactant. kd′ is attributed to the
sum of the first-order rate coefficients for diffusion of OH out
of the detection zone and its reaction with the photochemical
precursor and possible impurities in the bath gas. Typical
Figure 1. Plots of the measured first-order rate coefficient, k′, against
the concentration of the reactants (a) CF3CH2CH3, (b) CF3CHFCH2F,
and (c) CHF2CHFCHF2 at 238 K (circles) and 373 K (triangles). The
lines are the linear least-squares fit of the data to eq 6b. The figure
shows that k1 is larger than k2 while k2 and k3 are nearly equal.
measured values of kd′ were in the range 80-300 s-1
.
Values of k′ were obtained from an unweighted linear least-
squares fit of the measured values of St as a function of reaction
time to eq 6b. In all cases, the measured OH profiles were
exponential, confirming that the loss rates of OH were pseudo-
first order in [OH]. At each temperature, k1(T) was obtained
from a linear least-squares fit of the measured values of k′
obtained at various concentrations of the reactant to the
equation: k′ ) k1[reactant] + kd. Figure 1 shows k1(T), k2(T),
and k3(T) values that were obtained from data collected at the
temperature extremes of the current study. Under all conditions,
k′ varied linearly with the concentration of HFC, and the linear
least-squares fits to the data were very precise. Tables 1-3
give a summary of the experimental conditions used in
the determination of k1-k3 along with the obtained rate
coefficients.
(b) via direct IR absorption. The reactant concentration deter-
mined by the two independent methods agreed to better than
5% under all experimental conditions. On this basis, we estimate
the systematic uncertainty in the reactant concentration in the
LIF reactor to be (8% at the 95% confidence level.
Another possible source of systematic error in the measured
values of k(T) is the contributions from reactions of OH with
impurities in the reactants. As mentioned earlier, CF3CH2CH3
did not have any measurable impurities. Hence, the contribution
due to the impurities in the measurement of k1 is negligible.
This conclusion was further checked by measuring k1 after
passing the sample through a column containing sulfuric acid
(20 wt % deposited on Chromosorb WHP). This column
selectively removed olefins and, hence, the reactive impurities.
The measured value of k1 was the same (see Table 1) at the
lowest temperature of the study, where the impurities are
expected to contribute the most. Therefore, we are confident
that the measured value of k1 is not influenced by reactive
impurities. In the case of CF3CHFCH2F, trace amounts of
CF3CHFCHF2, CF3CH2CH2CF3, and CF3CHFCH3 were ob-
served in the sample. However, the rate coefficients for the
reactions of OH with these impurities are roughly the same as
that with CF3CHFCH2F; therefore, their presence did not
Uncertainties in the Measured Values of k′. We estimate
the systematic uncertainties in the measured parameters to be
relatively small: (1% in total pressure, (2% in flow rate, and
<1% in temperature. The precision of the first-order rate
coefficient, k′, was on the order of a few percent, and the
uncertainty in the slopes of the k′ versus the concentration of
HFC were also small. The largest source of uncertainty in the
present study lies in the determination of reactant concentration
in the LIF reactor. The concentration of the reactant was
determined using two independent methods: (a) using flow rates
(measured using calibrated mass flow meters) and pressure and

Rate Coefficients for Reaction of OH with HFCs
J. Phys. Chem. A, Vol. 110, No. 21, 2006 6727
TABLE 1: Summary of the Experimental Conditions and Rate Coefficients for the Reaction OH + CF3CH2CH3 (HFC-263fb)
temperature
K
pressure
Torr
flow rate
cm s-1
[H2O2],1013
[OH]0, 1011
laser fluence
[CF3CH2CH3], 1014
molecule cm-3
k1, 10-14
molecule cm-3
molecule cm-3
mJ cm-2 pulse-1
cm3 molecule-1 s-1
238
238a
256
274
297
297
297
297
323
348
373
50
50
50
52
52
52
210
52
50
50
50
6.0
6.2
6.5
7.3
7.3
17.6
6.3
7.3
8.2
8.7
9.3
3.19
6.45
3.68
3.05
3.18
3.35
1.60
5.25
3.58
3.77
3.89
2.75
5.57
3.35
2.78
2.32
3.35
1.60
1.19
3.25
3.42
3.54
4.8
4.8
5.1
5.1
4.1
5.6
5.6
1.3
5.1
5.1
5.1
6.53-43.6
6.62-41.3
8.91-35.0
5.18-27.0
3.97-32.9
2.80-16.6
4.83-30.4
9.46-41.8
3.68-27.5
5.28-20.1
4.05-27.8
1.96 ( 0.04b
1.97 ( 0.02
2.71 ( 0.06
3.96 ( 0.08
5.42 ( 0.04
5.55 ( 0.14
5.54 ( 0.04
5.48 ( 0.08
8.03 ( 0.12
10.7 ( 0.10
14.3 ( 0.10
a Experiments carried out by passing the sample through a purification column containing sulfuric acid (20 wt %) deposited on Chromosorb
WHP. b The quoted uncertainties are the 2σ (95% confidence limit) precision from the least-squares fit of the measured k′ data to eq 6.
TABLE 2: Summary of the Experimental Conditions and Rate Coefficients for the Reaction OH + CF3CHFCH2F (HFC-245eb)
temperature
K
pressure
Torr
flow rate
cm s-1
[H2O2], 1013
[OH]0, 1011
laser fluence
[CF3CHFCH2F], 1014
k2, 10-14
molecule cm-3
molecule cm-3
mJ cm-2 pulse-1
molecule cm-3
molecule cm-3 s-1
238
238a
254
272
297
297
297
297
324
348
349
374
50
52
50
51
51
49
210
50
51
50
50
50
6.2
6.3
6.8
7.0
7.4
18.7
6.4
7.5
8.3
8.5
8.8
9.6
2.35
5.2
2.50
4.4
2.83
4.22
1.52
11.7
4.33
5.0
2.14
4.26
2.28
3.20
2.7
4.03
1.38
3.72
3.35
3.87
5.45
4.49
5.1
4.6
5.1
4.1
5.3
5.3
5.1
1.8
4.3
4.3
4.3
4.3
6.57-53.5
3.90-51.8
5.64-42.6
5.40-42.7
10.2-40.4
2.91-13.5
4.69-36.6
3.47-27.3
5.75-35.3
2.15-30.1
5.15-30.3
2.86-28.6
0.63 ( 0.01b
0.63 ( 0.01
0.94 ( 0.01
1.22 ( 0.02
1.81 ( 0.04
1.79 ( 0.08
1.79 ( 0.06
1.80 ( 0.08
2.50 ( 0.06
3.43 ( 0.08
3.39 ( 0.12
4.58 ( 0.08
7.05
5.8
a Experiments carried out by passing the sample through a purification column containing sulfuric acid (20 wt %) deposited on Chromosorb
WHP. b The quoted uncertainties are the 2σ (95% confidence limit) precision from the least-squares fit of the measured k′ data to eq 6.
TABLE 3: Summary of the Experimental Conditions and Rate Coefficients for the Reaction OH + CHF2CHFCHF2
(HFC-245ea)
temperature
K
pressure
Torr
flow rate
cm s-1
[H2O2], 1013
[OH]0, 1011
laser fluence
[CHF2CHFCHF2], 1014
molecule cm-3
k3, 10-14
molecule cm-3
molecule cm-3
mJ cm-2 pulse-1
molecule cm-3 s-1
238
238a
254
254a
272
272a
297
297
297
297
324
348
374
51
52
52
52
51
52
63
205
55
50
51
52
50
6.2
6.1
6.6
6.6
7.0
6.8
6.2
6.6
7.5
18.2
8.2
8.7
9.50
2.47
4.65
2.74
9.40
2.75
8.10
2.59
1.80
8.35
5.30
3.56
3.86
6.05
1.79
3.81
2.37
7.27
2.50
6.63
2.94
1.64
7.60
4.58
3.08
3.68
5.23
4.1
4.6
4.8
4.3
5.1
4.6
6.4
5.1
5.1
4.8
4.8
5.3
4.8
9.53-46.5
2.34-41.7
4.38-41.8
3.07-43.9
8.56-41.1
4.81-26.7
8.39-65.6
6.87-64.5
9.58-52.1
3.36-23.9
3.56-27.8
5.60-35.5
3.73-28.9
0.80 ( 0.01b
0.60 ( 0.01
1.03 ( 0.02
0.79 ( 0.01
1.36 ( 0.04
1.17 ( 0.06
2.02 ( 0.04
1.90 ( 0.04
1.93 ( 0.02
1.87 ( 0.06
2.68 ( 0.08
3.70 ( 0.08
4.75 ( 0.06
a Experiments carried out by passing the sample through a purification column containing sulfuric acid (20 wt %) deposited on Chromosorb
WHP. b The quoted uncertainties are the 2σ (95% confidence limit) precision from the least-squares fit of the measured k′ data to eq 6.
influence the determination of k2. Kinetic measurements made
using the purification method noted above also indicated that
olefinic impurities were not a problem. The CHF2CHFCHF2
sample was contaminated with ∼50 ppmv of CF3CHdCF2. To
our knowledge, the rate coefficient for the reaction OH +
CF3CHdCF2 has not been reported. However, using the rate
coefficient7 for the reaction of OH with CH3CHdCH2 (3 ×
10-11 cm3 molecule-1 s-1) and the measured abundance of this
impurity, we compute the contribution of this unwanted reaction
to be roughly 5% at 298 K increasing to 20% at 238 K. This
analysis was verified when k3 was measured using the purifica-
tion method noted above. The k3 value measured at the lowest
temperature, 238 K, was lower by 20% compared with that
measured using the sample without purification. Yet another
possible indication of the reactive impurity was the slight
curvature in the Arrhenius plot for k3 shown in Figure 3. The
influence of the impurity was also verified at 238, 254, and
272 K, and the results of these rate coefficient measurements
are given in Table 3 and shown in Figure 3. The Arrhenius
plot for k3 measured using purified samples is almost linear.
This analysis also shows (Figure 3) that the impurity contribution
is negligible at 298 K and above. Therefore, the low temperature
(<297 K) k3 values obtained using the sample without purifica-
tion were not used in the final Arrhenius fitting.

6728 J. Phys. Chem. A, Vol. 110, No. 21, 2006
Rajakumar et al.
Figure 3. Arrhenius plot for the rate coefficient data obtained for the
OH radical reaction with CHF2CHFCHF2 (k3) over the temperature
range 238-374 K. The triangles correspond to the data obtained with
the unpurified sample and the circles to the data obtained with a purified
sample. The solid line is the linear least-squares fit to the data obtained
by using the purified samples (see text for details). The dashed lines
represent the 2σ uncertainty limits, f(T), calculated using f(298 K) )
1.10 and g ) 40 (see text for definition of terms). The square
corresponds to the room-temperature value reported by Nelson et al.3
An unweighted linear least-squares fit of the k1-3(T) data
given in Tables 1, 2, and 3 to the equation, ln(k) ) ln(A) -
(E/R)(1/T) yielded
k1(T) ) (4.36 ( 0.63) × 10-12 exp[-(1293 (
42)/T] cm3 molecule-1 s-1
k2(T) ) (1.23 ( 0.15) × 10-12 exp[-(1250 (
36)/T] cm3 molecule-1 s-1
Figure 2. Arrhenius plot for the rate coefficient data obtained for the
OH radical reaction with (a) CF3CH2CH3 (k1) and (b) CF3CHFCH2F
(k2) over the temperature range 238-373 K. The triangles correspond
to data obtained with the unpurified sample and the circles to the data
obtained with a purified sample. The solid lines are the linear least-
squares fit to all the data (see text for details). The dashed lines represent
the 2σ uncertainty limits, f(T), calculated using f(298 K) ) 1.10 and g
) 20 (see text for definition of terms). The open squares are the room-
temperature values reported by Nelson et al.3
k3(T) ) (1.91 ( 0.39) × 10-12 exp[-(1377 (
60)/T] cm3 molecule-1 s-1
where the errors are 2σ from the precision of the fit and σA )
AσlnA. As discussed earlier, we estimate the uncertainty in the
determination of the concentration of the reactant in the reactor
to be 8% at the 95% confidence level. Propagation of this
uncertainty into the above Arrhenius expressions and rounding
off the uncertainties in E/R yields
The measured rate coefficients for the reaction of OH with
CF3CH2CH3, CF3CHFCH2F, and CHF2CHFCHF2 as a function
of temperature are shown in Figures 2 and 3. As shown in Figure
2, the Arrhenius plots for reactions 1 and 2 are linear even when
the
unpurified
samples
of
CF3CH2CH3
and
k1(T) ) (4.36 ( 0.72) × 10-12 exp[-(1290 (
40)/T] cm3 molecule-1 s-1
CF3CHFCH2F were used.
Other sources of systematic error in the rate coefficient
determination include the possible influence of unwanted and
unrecognized secondary reactions. The measured values of
k1-3(T) were independent of the photolysis laser fluence, total
pressure, flow velocity, and [OH]0, as outlined in Tables 1-3.
If the reactants absorbed the 248 nm photolysis radiation and
generated free radicals that could react with OH, the measured
rate coefficient would be dependent on photolysis laser fluence.
The independence of the measured values of k1-3(T) with
variations in laser fluence confirms that reaction of photolytic
products with OH was not a problem.
k2(T) ) (1.23 ( 0.18) × 10-12 exp[-(1250 (
40)/T] cm3 molecule-1 s-1
k3(T) ) (1.91 ( 0.42) × 10-12 exp[-(1375 (
100)/T] cm3 molecule-1 s-1
The possible impurity contribution to k3, which decreases with
increasing temperature, is assumed to contribute only to the E/R
term. Hence, this uncertainty has been increased slightly. The

Rate Coefficients for Reaction of OH with HFCs
J. Phys. Chem. A, Vol. 110, No. 21, 2006 6729
TABLE 4: Comparison of OH + CF3CH2CH3, CF3CHFCH2F, and CHF2CHFCHF2 Rate Coefficients with Those from Previous
Studies and Calculations
molecule
k(298 K)a,b
(5.55 ( 0.48) × 10-14
Aa
E/R ( ∆E/R, K
1290 ( 40
T range, K
238-375
techniquec
PLP-LIF
reference
this work
CF3CH2CH3
(HFC-263fb)
(4.36 ( 0.72) × 10-12
(4.20 ( 0.43) × 10-14
298
298
DF-LIF
Nelson et al.3
JPL-20026
4.20 × 10-14
3.09 × 10-14
3.16 × 10-12
(1.23 ( 0.18) × 10-12
1385
IPC
SAR
PLP-LIF
Percival et al.4
Tokuhashi et al.8
this work
4.44 × 10-14
CF3CHFCH2F
(HFC-245eb)
(1.81 ( 0.18) × 10-14
1250 ( 40
238-375
(1.48 ( 0.17) × 10-14
298
298
DP-LIF
Nelson et al.3
JPL-20026
1.50 × 10-14
7.41 × 10-15
1.86 × 10-12
(1.91 ( 0.42) × 10-12
1650
IPC
SAR
PLP-LIF
Percival et al.4
Tokuhashi et al.8
this work
1.05 × 10-14
CHF2CHFCHF2
(HFC-245ea)
(1.88 ( 0.20) × 10-14
1375 ( 100
238-375
(1.60 ( 0.23) × 10-14
298
DF-LIF
IPC
SAR
Nelson et al.3
Percival et al.4
Tokuhashi et al.8
Orkin et al.20
1.86 × 10-14
2.63 × 10-12
(4.41 ( 0.80) × 10-13
(1.45 ( 0.22) × 10-12
1475
1.60 × 10-14
CF3CF2CH3
(HFC-245cb)
CF3CH2CF3
(HFC-236fa)
(1.54 ( 0.04) × 10-15
1690 ( 60
2500 ( 150
287-370
283-403
FP-RF
(3.30 ( 0.50) × 10-16
RR
Hsu and DeMore21.
a Units of cm3 molecule-1 s-1 b Quoted uncertainties from this work are at the 2σ level and include estimated systematic errors. Error limits
.
from other studies are as quoted by the authors. c PLP: pulse laser photolysis. LIF: laser-induced fluorescence. DF: discharge flow. FP: flash
photolysis. RF: resonance fluorescence. IPC: ionization potential correlation. RR: relative rate.
uncertainties in the format used in data evaluations6 where the
error at any given temperature, f(T), is given by
to derive the SAR factors for fluorinated compounds. Rate
coefficient data for a series of OH + HFC reactions have also
been reported by Percival et al.4 Percival et al.4 calculated
Arrhenius parameters based on correlations with calculated
ionization potentials of the HFCs. The results for the molecules
studied in this work are given in Table 4 and are found to be in
reasonable agreement with the experimental values.
1
T
1
298
f(T) ) f(298 K) exp g
-
|
(
)
|
are f(298 K) ) 1.10 and g ) 20 for CF3CH2CH3 and
CF3CHFCH2F and g ) 40 for CHF2CHFCHF2 but are given at
the 2σ level. These uncertainty limits are shown in Figures 2
and 3.
The activation energy, E, for these reactions depends on the
number and reactivities of reaction sites available and the
number of F atoms present in the molecule. Out of the three
molecules we have studied, the OH + CF3CH2CH3 reaction is
expected to have a lower E than the other two molecules, as it
has two F free sites, more H atoms, and two secondary
hydrogens. We have used the SAR method 8(derived by
Tokuhashi et al.) to evaluate the contribution of each site in
the molecule toward the rate coefficient. In CF3CH2CH3, the
contribution of the CH2 group is calculated to be only 7%, while
the rest of the reaction is due to the CH3 group. This is a
surprisingly small contribution. The SAR calculations using
DeMore’s group contributions11 indicate that both CH2 and CH3
groups contribute almost equally to the overall reaction. This
conclusion appears to be more reasonable. We attribute the low
value of the calculated contribution by the CH2 group using
the Tokuhashi et al.’s group factors to an error in their group
contributions, which were derived to optimize the predictability
for a different set of functional groups. However, it should be
noted that the calculated overall rate coefficient is reasonably
close to the measured value.
The values of k1, k2, and k3 obtained in this work are
summarized in Table 4 along with those from the only other
experimental determination by Nelson et al.3 Nelson et al.3
measured the rate coefficients k1- k3 at room temperature using
a discharge flow technique with laser-induced fluorescence
detection of the OH radicals. The value of k1(298 K) measured
by Nelson et al.3 is 4.20 × 10-14 cm3 molecule-1 s-1, which is
lower than our value by a factor of 1.3. The value of k2(298 K)
reported by Nelson et al.3 is also lower than our value, by a
factor of 1.2, but overlaps with ours within the combined
uncertainty limits. The value of k3(298 K) reported by the Nelson
et al.3 agrees with our value within the combined uncertainties.
As discussed earlier, many variations of experimental parameters
were used in our study to identify and minimize possible
systematic errors. These tests give us added confidence in our
results. The source of the small but measurable discrepancies
with the rate coefficient data of Nelson et al.3 is currently
unknown.
We can also compare our measured rate coefficients with
those obtained using structure activity relationships (SAR).
Recently, Tokuhashi et al.8 derived a set of SAR for fluorinated
compounds using the methods previously developed in the
works of Atkinson9,10 and DeMore.11 The rate coefficients
calculated using the SAR analyses for CF3CH2CH3,
CF3CHFCH2F, and CHF2CHFCHF2 are included in Table 4.
For CF3CH2CH3 and CHF2CHFCHF2, the SAR rate coefficients
are in good agreement (within 20%) with the experimental
values. However, for CF3CHFCH2F, the SAR rate coefficient
is lower than our experimental value by ∼60%. The larger
discrepancy is most likely due to the limited database available
When CH2 is replaced with CF2 in CF3CF2CH3, the reaction
must be exclusively due to H-atom abstraction from the CH3
group; accordingly, there should be an increase in the activation
energy as is observed (see Table 4). However, when the CH3
group is replaced with CF3 in CF3CH2CF3, the E/R is raised
further by ∼1000 K, which is again reasonable, because the
reaction is solely due to the CH2 group which is influenced by
two CF3 groups. In the case of CF3CHFCH2F, the contribution
of both CHF and CH2F groups is evaluated to be ∼50% of the
total reaction, using the methods of both Tokuhashi et al.8 and
DeMore.11 In CHF2CHFCHF2, using Tokuhashi’s8 method, each

6730 J. Phys. Chem. A, Vol. 110, No. 21, 2006
Rajakumar et al.
CHF2 group was found to contribute 25% and the CHF group
contributed 50% to the total reaction. DeMore’s11 method
suggests the contributions of each CH2F group to be 32% and
that of the CHF group to be 36%. The E/R for these two
molecules turn out to be similar. Therefore, the atmospheric
degradation of CF3CHFCH2F and CHF2CHFCHF2 will proceed
via multiple pathways with different possible end products.
Atmospheric Implications
Lifetimes. The atmospheric lifetime of any compound that
is released into the atmosphere depends on the rates of all
processes that remove it. It may be lost by photolysis, wet and
dry deposition, rain out, and reaction with free radicals
(especially OH radicals). The UV absorption cross-sections of
the HFCs included in the current investigation, at wavelengths
greater than 300 nm, are expected to be very small. Therefore,
photolytic loss of these compounds is negligible. Because of
the presence of C-H bonds in HFCs, they are lost in the
atmosphere mostly via their reactions with OH radicals.
Atmospheric lifetimes for gases destroyed primarily by reaction
with OH can be best evaluated using constraints on global OH
densities derived from methyl chloroform measurements.12,13
The analysis of Prinn et al.,12 using measured global methyl
chloroform abundance during the past 25 years, together with
detailed information on production, release, and transport of this
solely industrial gas provides the most complete assessment of
the globally averaged OH density. We use the OH values and
the general approach given by Prinn et al.12 to derive OH-driven
lifetimes for CF3CH2CH3, CF3CHFCH2F, and CHF2CHFCHF2
using a two-box (stratosphere and troposphere) assumption. With
this method, the OH-driven lifetimes of CF3CH2CH3,
CF3CHFCH2F, and CHF2CHFCHF2 were estimated to be 0.8,
2.6, and 2.5 years, respectively.
Global Warming Potentials. Global warming potentials
(GWP) are a measure of the time-integrated radiative forcing
of the climate system due to a pulse release of a particular gas,
relative to a reference molecule (usually taken to be carbon
dioxide).14,15 As such, they can be very useful in evaluating the
relative climatic implications of the use of one halocarbon
substitute compared to another. To compute the GWP, one needs
the radiative forcing, the lifetime of the gas, and the time-
integrated radiative forcing of the reference gas (we use those
given for CO2 in WMO-20031). We compute the adjusted all-
sky radiative forcing using a line-by-line (LBL) radiative transfer
model developed by Portmann et al.16 (see also Forster et al.17).
The infrared absorption cross-sections of CF3CH2CH3, CF3-
CHFCHF2, and CHF2CHFCHF2 measured in this work and
shown in Figure 4 were used. Pinnock et al.18 have presented a
detailed analysis of the sensitivity of GWPs to other factors,
such as band overlaps with gases such as H2O, CO2, and so
forth, global averaging potentials, and background atmospheres.
We have accurately included all of these effects, with the
exception of the stratospheric falloff of the molecule (which
could reduce these estimates by up to 10%). The line-by-line
code used here numerically integrates the radiative transfer
equation to evaluate the upward, downward, and net irradiance
at each level in the atmosphere. We employ the ISCCP monthly
mean zonally averaged profiles of pressure, temperature, H2O,
CO2, and O3, and fields of CH4 and N2O derived from the
NOCAR model. The absorption spectra for these gases have
been taken from the HITRAN-2000 database.19 Spectral overlap
between these gases and CF3CH2CH3, CF3CHFCHF2, and
CHF2CHFCHF2 are considered. The globally and annually
averaged radiative forcing for CF3CH2CH3, CF3CHFCHF2, and
Figure 4. Infrared absorption spectra of (a) CF3CH2CH3, (b)
CF3CHFCH2F, and (c) CHF2CHFCHF2 in the region 500-4000 cm-1
recorded at 1 cm-1 resolution measured using pure samples, i.e., with
no bath gas added.
TABLE 5: Global Warming Potential (GWP) (mass basis)
Referenced to the Decay Response for CO2
global warming potential
time horizons (years)
lifetime radiative forcing
molecule
(years) (W m-2 ppbv-1
)
20
100
500
CF3CH2CH3
0.8
2.6
2.5
0.13
0.23
0.18
234
69.6
21.7
(HFC-263fb)
CF3CHFCH2F
(HFC-245eb)
CHF2CHFCHF2
(HFC-245ea)
962
723
286
215
89.2
67.8
CHF2CHFCHF2 were found to be 0.13, 0.23, and 0.18 W m-2
ppbv-1, respectively. For comparison, the radiative forcing by
CFC-11 is 0.25 W m-2 ppbv-1 1
.
Table 5 presents the GWPs for CF3CH2CH3, CF3CHFCHF2,
and CHF2CHFCHF2 for the same time periods given in WMO-
20031 (to which they can be directly compared, since the same
time-integrated radiative forcing of CO2 has been used). The
information in this table provides a basis for evaluating the
relative climatic impacts of choices between a large number of
short- and long-lived halocarbon substituents.
To conclude, though the atmospheric lifetimes of
CF3CH2CH3, CF3CHFCHF2, and CHF2CHFCHF2 are reason-
ably long, they may be considered short-lived when compared

Rate Coefficients for Reaction of OH with HFCs
J. Phys. Chem. A, Vol. 110, No. 21, 2006 6731
(8) Tokuhashi, K.; Nagai, H.; Takahashi, A.; Kaise, M.; Kondo, S.;
Sekiya, A.; Takahashi, M.; Gotoh, Y.; Suga, A. J. Phys. Chem. A 1999,
103, 2664.
(9) Atkinson, R. Int. J. Chem. Kinet. 1987, 19, 799.
(10) Kwok, E. S. C.; Atkinson, R. Atmos. EnViron. 1995, 29, 1685.
(11) DeMore, W. B. J. Phys. Chem. 1996, 100, 5813.
(12) Prinn, R.; Weiss, R.; Millar, B.; Jaung, J.; Alyea, F.; Cunnold, D.;
Fraser, P.; Hartley, D.; Simmonds, P. Science 1995, 269, 187.
(13) Prinn, R.; Cunnold, D.; Simmonds, P.; Alyea, F.; Boldi, R.;
Crawford, A.; Fraser, P.; Gutzler, D.; Hartley, D.; Rosen, R.; Rasmussen,
R. J. Geophys. Res., [Atmos.] 1992, 97, 2445.
(14) RadiatiVe forcing of climate change and an eValuation of the IPCC
IS92 Emission Scenarios; Houghton, J. T., Filho, L. G. M., Bruce, B., Lee,
H., Callander, B. A., Haites, E., Harris, N., Maskell, K., Eds.; Intergov-
ernmental Panel on Climate Change (IPCC), climate change, 1994;
Cambridge University Press: New York, 1995.
to CFCs. Their GWPs are also relatively smaller than those of
CFCs (e.g., CFC-11 with a GWP of 6330 for 20-year time
horizon) primarily because of their shorter lifetimes. They do
not contribute to stratospheric ozone depletion. Hence, these
compounds may be suggested as acceptable substitutes to CFCs.
Acknowledgment. We thank Stephen Montzka for the help
in analyzing the HFC samples by GCMS. We thank David
Nelson and Charles Kolb of Aerodyne research for providing
the samples used in this study. This work was funded in part
by NOAA’s climate program. A.R.R. thanks Dave Golden for
his friendship, encouragement, and scientific interchanges over
the past three decades.
(15) Climate Change, The IPCC Scientific assessment; Houghton, J. T.,
Jenkins, G. J., Ephraums, J. J., Eds.; Cambridge University Press: New
York, 1990.
References and Notes
(16) Portmann, R. W.; Solomon, S.; Fishman, J.; Olson, J. R.; Kiehl, J.
T.; Briegleb, B. J. Geophys. Res., [Atmos.] 1997, 102, 9409.
(17) Forster, P. M. D.; Burkholder, J. B.; Clerbaux, C.; Coheur, P. F.;
Dutta, M.; Gohar, L. K.; Hurley, M. D.; Myhre, G.; Portmann, R. W.; Shine,
K. P.; Wallington, T. J.; Wuebbles, D. J. Quant. Spectrosc. Radiat. Transfer
2005, 93, 447.
(1) WMO Scientific Assessment of Ozone Depletion 2002; Global Ozone
research monitoring project no. 47, Geneva, 2003.
(2) Montreal Protocol on Substances that Deplete the Ozone Layer;
United Nations Environmental Program (UNEP), Geneva, Switzerland,
1987.
(3) Nelson, D. D.; Zahniser, M. S.; Kolb, C. E.; Magid, H. J. Phys.
Chem. A 1995, 99, 16301.
(4) Percival, C. J.; Marston, G.; Wayne, R. P. Atmos. EnViron. 1995,
29, 305.
(5) Vaghjiani, G. L.; Ravishankara, A. R. J. Phys. Chem. 1989, 93,
1948.
(6) Sander, S. P.; Friedl, R. R.; Golden, D. M.; Huie, R. E.; Kolb, C.
E.; Kurylo, M. J.; Molina, M. J.; Moortgat, G. K.; Orkin, V. L.;
Ravishankara, A. R.; Finlayson-Pitts, B. J. Chemical kinetics and photo-
chemical data for use in atmospheric studies; JPL Publication 02-25; Jet
Propulsion Laboratory: Pasadena, 2002; Evaluation Number 14.
(7) Nielsen, O. J.; Jorgensen, O.; Donlon, M.; Sidebottom, H. W.;
O’Farrell, D. J.; Treacy, J. Chem. Phys. Lett. 1990, 168, 319.
(18) Pinnock, S.; Hurley, M. D.; Shine, K. P.; Wallington, T. J.; Smyth,
T. J. J. Geophys. Res., [Atmos.] 1995, 100, 23227.
(19) Rothman, L. S.; Barbe, A.; Benner, D. C.; Brown, L. R.; Camy-
Peyret, C.; Carleer, M. R.; Chance, K.; Clerbaux, C.; Dana, V.; Devi, V.
M.; Fayt, A.; Flaud, J. M.; Gamache, R. R.; Goldman, A.; Jacquemart, D.;
Jucks, K. W.; Lafferty, W. J.; Mandin, J. Y.; Massie, S. T.; Nemtchinov,
V.; Newnham, D. A.; Perrin, A.; Rinsland, C. P.; Schroeder, J.; Smith, K.
M.; Smith, M. A. H.; Tang, K.; Toth, R. A.; Vander Auwera, J.; Varanasi,
P.; Yoshino, K. J. Quant. Spectrosc. Radiat. Transfer 2003, 82, 5.
(20) Orkin, V. L.; Huie, R. E.; Kurylo, M. J. J. Phys. Chem. A 1997,
101, 9118.
(21) Hsu, K. J.; DeMore, W. B. J. Phys. Chem. 1995, 99, 1235.