Measurement of the OH reaction rate constants for CF3CH2OH, CF3CF2CH2OH, and CF3CH(OH)CF3

Article abstract of DOI:10.1021/jp983961x

The OH reaction rate constants have been measured for CF₃CH₂OH, CF₃CF₂CH₂OH, and CF₃CH(OH)CF₃ over the temperature range 250-430K. Kinetic measurements have been carried out using the discharge flow, laser photolysis, and flash photolysis methods, combined respectively with the laser-induced fluorescence technique to monitor the OH radical concentrations. The influence of impurities contained in the sample of CF₃CF₂CH₂OH has been investigated by means of sample purification using gas chromatography. No sizable effect of impurities was found on the measured rate constants of these three fluorinated alcohols. The Arrhenius rate constants have been determined from the respective kinetic data as k(CF₃CH₂OH) = (2.00 ± 0.37) × 10^-12 exp[-(890 ± 60)/T], k(CF₃CF₂CH₂OH) = (1.40 ± 0.27) × 10-¹² exp[-(780 ± 60)/T], and k(CF₃CH-(OH)CF₃) = (6.99 ± 1.56) × 10^-13 exp[-(990 ± 70)/T] cm³ molecule^-1 s^-1. A method of predicting the OH reaction rate constants for fluorinated alcohols, hydrofluorocarbons, alkanes, and alcohols has been proposed.

Full text of DOI:10.1021/jp983961x

2664
J. Phys. Chem. A 1999, 103, 2664-2672
Measurement of the OH Reaction Rate Constants for CF3CH2OH, CF3CF2CH2OH, and
CF3CH(OH)CF3
,
†
†
†
†
†
Kazuaki Tokuhashi,* Hidekazu Nagai, Akifumi Takahashi, Masahiro Kaise, Shigeo Kondo,
†
‡
†
‡
Akira Sekiya, Mitsuru Takahashi, Yoshihiko Gotoh, and Atsuo Suga
National Institute of Materials and Chemical Research, 1-1 Higashi, Tsukuba, Ibaraki 305, Japan, and
Research Institute of InnoVatiVe Technology for the Earth, Hongo Wakai Building,
2
-40-17, Hongo, Bunkyo-ku, Tokyo 113, Japan
ReceiVed: October 5, 1998; In Final Form: December 30, 1998
3 2 3 2 2 3 3
The OH reaction rate constants have been measured for CF CH OH, CF CF CH OH, and CF CH(OH)CF
over the temperature range 250-430K. Kinetic measurements have been carried out using the discharge
flow, laser photolysis, and flash photolysis methods, combined respectively with the laser-induced fluorescence
technique to monitor the OH radical concentrations. The influence of impurities contained in the sample of
CF
effect of impurities was found on the measured rate constants of these three fluorinated alcohols. The Arrhenius
rate constants have been determined from the respective kinetic data as k(CF CH
OH) ) (2.00 ( 0.37) ×
OH) ) (1.40 ( 0.27) × 10 exp[-(780 ( 60)/T], and k(CF CH-
3 2 2
CF CH OH has been investigated by means of sample purification using gas chromatography. No sizable
3
2
-
12
-12
-1 -1
1
(
0
exp[-(890 ( 60)/T], k(CF
3
3
-
CF
2
CH
2
3
13
3
OH)CF
) ) (6.99 ( 1.56) × 10 exp[-(990 ( 70)/T] cm molecule s . A method of predicting the OH
reaction rate constants for fluorinated alcohols, hydrofluorocarbons, alkanes, and alcohols has been proposed.
Introduction
temperature range 250-430 K by the discharge flow (DF), laser
photolysis (LP), and flash photolysis (FP) techniques. In all
Fully halogenated chlorofluorocarbons (CFCs) cause deple-
tion of the stratospheric ozone layer as well as global warming.
Partially fluorinated alcohols are important in relevance with
potential substitutes of CFCs. Since, these molecules do not
contain Cl atoms, they do not contribute to the ozone depletion,
but may potentially cause the global warming effect. In general,
the molecules of these substances contain hydrogen atoms, and
cases, the experiments have been carried out under a large excess
of reactant (fluorinated alcohols) over the initial OH radical
concentrations. The concentration of OH radicals has been
measured by the laser-induced fluorescence (LIF) method.
Discharge Flow (DF) Method. The schematic diagram of
the discharge flow reactor is shown in Figure 1. Two flow tubes
about 700 and 900 mm long were used in the kinetic study.
They are made of Pyrex glass and are 20 mm in inner diameter.
The flow tubes are accommodated with a movable double
1
are expected to be oxidized by OH radicals in the atmosphere.
Therefore, study of the reactivity against OH radicals is
indispensable for the evaluation of atmospheric lifetime of these
molecules. Although there are numerous data of rate constants
for various compounds, there is little information regarding the
OH reaction rate constants for fluorinated alcohols. In this paper,
we report the results of the kinetic measurements for the
reactions of OH radicals with CF3CH2OH, CF3CF2CH2OH, and
CF3CH(OH)CF3 over the temperature range 250-430 K. Effects
of impurities on the measured rate constants have been
investigated by the sample purification with gas chromatogra-
phy. On the other hand, to obtain the reaction kinetic data from
the experiment, considerable labor is required. It is desirable to
develop the empirical prediction method of the reaction rate
constants. In this paper, we have reexamined the prediction
method of OH reaction rate constants which considers the
contributions from the substituent groups. The prediction method
is applicable to the OH reaction rate constants of fluorinated
alcohols, hydrofluorocarbons, alkanes, and alcohols.
2
sliding injector made of stainless steel. Helium was used as
the primary carrier gas. OH radicals are produced by the reac-
tion
H + NO f OH + NO
2
where hydrogen atoms are generated by microwave discharge
of Ar gas containing a trace amount of hydrogen. NO2 di-
luted with Ar is supplied in excess to the flow tube through the
top of the outer tube (6 mm o.d.) of the sliding injector. The
reactant vapor is added to the gas stream through the top of
the inner tube (3 mm o.d.) of the sliding injector. The top of
the inner tube is located 400 mm (for 700 mm tube) or 500
mm (for 900 mm tube) downstream from the top of the outer
tube. Both the flow tube and the sliding injector are coated with
Teflon to minimize wall loss of OH radicals. To establish a
stable profile of the reactant concentration in a short time,
helium gas is always slowly supplied to the inner tube of the
sliding injector together with the reactant vapor. The position
of the sliding injector was controlled by a pulse stage connected
with a pulse controller. A typical experimental condition is as
follows: total pressure is 4-6 Torr, the linear flow velocity of
Experimental Section
The OH reaction rate constants have been measured for
CF3CH2OH, CF3CF2CH2OH, and CF3CH(OH)CF3 over the
*
Author to whom correspondence should be addressed at Department
of Physical Chemistry, National Institute of Materials and Chemical
Research, 1-1 Higashi, Tsukuba, Ibaraki 305, Japan.
-1
gas mixture is 5-14 m s , and the NO2 concentration is (1.7-
1
3
-3
†
14) × 10 molecule cm . Initial concentration of OH radi-
cals, which is a function of the hydrogen flow rate, is (0.2-
National Institute of Materials and Chemical Research.
Research Institute of Innovative Technology for the Earth.
‡
1
0.1021/jp983961x CCC: $18.00 © 1999 American Chemical Society
Published on Web 03/19/1999

OH Reaction Rate Constants for Fluorinated Alcohols
J. Phys. Chem. A, Vol. 103, No. 15, 1999 2665
Figure 1. Schematic diagram of the discharge flow reactor.
Figure 2. Schematic diagram of the laser photolysis/laser-induced fluorescence apparatus.
1
1
-3
1
(
.0) × 10 molecule cm . The primary carrier gas, He
with an ArF excimer laser (Japan Storage Battery, EXL-200)
in the presence of He bath gas at a total pressure of 15-70
Torr. The power density of the excimer laser at the exit window
99.995%), was purified by heated copper oxide followed
by a liquid nitrogen trap. H2 diluted with a large excess of Ar
was prepared manometrically in a stainless steel vessel, or
purchased as a mixture with a H2 concentration of around
-
2
-1
of the reaction cell is around 2-5 mJ cm pulse . To ensure
1
rapid conversion of O( D) atoms to OH radicals, the ratio of
0
.005%. The samples of H2 (99.99%) and Ar (99.995%) were
H2O to N2O is kept larger than about 20:1. N2O concentration
14
-3
used without further purification. NO2 (99.5% up) diluted with
Ar was purchased as a mixture with an NO2 concentration of
1
is about (1.2-4.0) × 10 molecule cm . The photolyzing light
is directed to the reaction cell along the axial line of the cell
through a quartz window.
-2%.
Laser Photolysis (LP) and Flash Photolysis (FP) Methods.
The apparatus used in the FP method is almost the same as
that for the LP method shown in Figure 2 except for the
photolysis light source and the gases used. In the case of the
FP method, H2O (100-200 mTorr) is directly photolyzed with
pulsed light of Xe flash lamp (EG & G, FX-193U, 600-1300 V
A schematic diagram of the LP-LIF apparatus is shown in Figure
. The Pyrex glass reactor with inner diameter of 25 mm is
2
used for the LP and FP experiments. The length of the reactor
is 40 cm. In the case of LP method, OH radicals are produced
by the reaction
-1
-1
and 2 µF, 0.36-1.69 J pulse , typically 0.64 J pulse , pulse
width is 10-20 µs, λ g 180 nm, quartz cutoff). The photolyzing
light is weakly focused by quartz lenses, and is directed to the
reaction cell along the axial line through the quartz window.
Ar is used as carrier gas in most cases (some experiments have
1
O( D) + H O f 2OH
2
1
where O( D) atoms are generated by photodissociation of N2O

2
666 J. Phys. Chem. A, Vol. 103, No. 15, 1999
Tokuhashi et al.
been carried out using a He carrier). The region between the
flash lamp (or excimer laser for the LP method) and the reaction
cell is purged with dry N2.
reactor was maintained either by electric heater (298-430 K),
or by circulating heated water (298-339 K) and cooled
ethanol-water mixture (250-273 K) in the outer jacket of the
reaction cell from a thermostated bath. It was measured by CA
thermocouple at the top of the inner tube of the sliding injector
Water vapor was supplied by bubbling a certain part of carrier
gas through a vessel filled with water at room temperature. The
total pressure of carrier gas containing water vapor was
measured by a capacitance manometer, and was kept constant
by a control valve. The amount of water vapor supplied to the
reaction cell was estimated from the flow rate of the carrier
gas, water temperature measured by CA thermocouple (type
K), and total pressure of carrier gas containing water vapor. To
prevent accumulation of photofragments or reaction products
in the cell, all photolysis experiments were carried out under
slow flow conditions. For both LP and FP methods, initial
concentration of OH radicals, which is estimated from com-
parison with fluorescence intensities obtained with the DF
(for the DF method) or at the spot around 1-2 cm downstream
from the probe laser beam (for LP and FP methods). During
the experiments, the temperature across the reaction volume was
maintained better than (2 K over the temperature range
examined. The gas flow rates, the total pressure, and the reaction
temperature were monitored by the second computer through a
digital recorder or data acquisition controller, and stored in a
computer via RS-232C circuit. To ensure that the experimental
data are free from any systematic errors, the experiments were
repeated at intervals from several days to several months under
a variety of flow conditions.
method, is always kept smaller than 1011 molecule cm-3. The
repetition rate of the photolysis light was set as 10 Hz.
Sample Analysis and Purification. The samples of fluori-
nated alcohols were analyzed by using gas chromatography with
an FID detector, where the area ratio of main peak against total
area was taken as the purity of the sample. The analytical
columns used in the present work are G-205 (1.2 mm i.d., 40
m long, film thickness 5 µm, Chemicals Inspection & Testing
Institute, Japan, He carrier gas), or stainless steel column (3
mm i.d., 10 m long) packed with mainly Silicone DC 702
A primary carrier gas, He (99.995%) or Ar (99.995%) was
used without further purification. N2O (99.999%) diluted with
He was purchased as a mixture with N2O concentration of
around 1%.
Laser-Induced Fluorescence Method. The concentration of
OH radicals has been measured by the LIF method. In the case
of DF and FP methods, the apparatus is the same as that for the
LP method shown in Figure 2. The excitation light is from a
frequency-doubled tunable dye laser (Spectron, SL4000B), and
the wavelength was tuned at about 308 nm. The dye laser was
pumped by pulsed light of a frequency-doubled Nd:YAG laser
(Shimadzu, N2 carrier gas). The lower value obtained using two
columns was taken as a purity of the sample. The purities of
CF3CH2OH and CF3CH(OH)CF3 samples were found to be
9
9.994 and 99.999%, respectively, and these samples were used
without further purification. The sample of CF3CF2CH2OH
99.56%) was purified by gas chromatography. In the purifica-
(
(Spectron, SL803). The repetition rate of the laser was set as
tion process, the vapor of CF3CF2CH2OH was charged in an
10 Hz. Fluorescence signals due to OH radicals were monitored
3
evacuated sampling tube (inner volume of about 50 cm ), and
at a right angle against both the excitation light and the
photolysis light (or flow tube for the DF method), and focused
by quartz lenses and a concave mirror, and detected by a
photomultiplier tube. The scattered light of the excitation light
supplied to a stainless steel column (9.6 mm i.d., 4 m long)
packed with Silicone DC 702 through a six-way switching
cock. Nitrogen was used as the carrier gas. The middle frac-
tion of the main peak (chromatogram was always monitored
by TCD detector) was collected through a four-way switching
cock into a trap cooled with liquid nitrogen. The sample of
CF3CF2CH2OH thus obtained showed purity of 99.935%.
(
and photolysis light source for LP and FP methods) was
reduced by a monochromator (Jarrell-Ash, Monospec 25,
360G/mm, 25 cm focal length). The output signal of the
2
photomultiplier tube was amplified by a preamplifier, and
accumulated (usually 400-600 shots) by a multichannel scaler/
averager (Stanford, SR430), or averaged (usually 128 shots by
a digital storage oscilloscope Gould-4050), and stored in a
microcomputer for further data processing. Trigger pulses for
YAG laser, multichannel scaler/averager (or digital storage
oscilloscope), and photolysis light source were generated by a
delay generator (Stanford, DG535). This computer was used to
control the delay generator (for LP and FP methods), pulse
controller (to determine the position of sliding injector for DF
method), and the other computer (see later).
Results and Discussion
According to the gas chromatographic analysis made before
the kinetic measurement, the purities of the samples of
fluorinated alcohols as supplied are 99.994, 99.56, and 99.999%
for CF CH OH, CF CF CH OH, and CF CH(OH)CF , respec-
3
2
3
2
2
3
3
tively. The individual impurities were not characterized though.
At any rate, since the impurity levels of CF CH OH and
3
2
CF CH(OH)CF in particular are extremely low, it is obvious
3
3
that the effects of impurities on the measurement of the OH
rate constants may be negligibly small for these samples. We
began with measuring the OH reaction rate constant for
CF3CH2OH and CF3CH(OH)CF3.
Gas Handling and Measurements. Various gas flow rates
were measured and controlled by calibrated mass flow control-
lers. In particular, the flow rate of fluorinated alcohol vapor
was directly measured and controlled by the calibrated mass
flow controller. For gaseous materials, calibration of the mass
flow controllers was made with a gas meter or a soap-film flow
meter. For liquid materials (fluorinated alcohols), calibration
of the mass flow controller was made by measuring the time-
pressure relationship in the vessel with a known volume using
a capacitance manometer, during which the sample vapor was
supplied to the evacuated vessel through the mass flow con-
troller. The total gas pressure of the reactor was monitored by
using a capacitance manometer (MKS Baratron), and was kept
constant by an electrically controlled exhaust throttle valve
located downstream of the reactor. The temperature of the
Figure 3 shows a typical example of a pseudo-first-order OH
decay plot obtained with the DF method for various CF3CH2-
OH concentrations. Since these plots show a linear relationship,
the pseudo-first-order rate constant (kobs) can be derived from
the slopes of the straight lines by least-squares fit to each decay
plot. In Figure 4, the observed pseudo-first-order rate constants
are plotted against CF3CH2OH concentration. In the case of the
DF method, a small correction factor (usually 1 to 3%) is applied
to each pseudo-first-order rate constants to account for the axial
3
diffusion effects. The plotted points are distributed along a
straight line, and the bimolecular rate constant for the reaction
of OH with CF3CH2OH can be derived from the slope by the

OH Reaction Rate Constants for Fluorinated Alcohols
J. Phys. Chem. A, Vol. 103, No. 15, 1999 2667
3 2
Figure 3. Pseudo-first-order decay of OH for various CF CH OH
-
1
concentrations. DF-LIF method. P ) 5 Torr, U ) 9.8 m s , T ) 298
Figure 5. Plot of the pseudo-first-order rate constants corrected for
1
4
-3
K. [CF
3 2
CH OH]/10 molecule cm ; (O), 0; (]), 2.39; (b), 3.93; (0),
wall loss, kobs - k
w 3 2 3 3
, against (O) CF CH OH, and (0) CF CH(OH)CF
5
.51; (3), 7.08; (4), 8.65.
concentrations at 298 K. DF-LIF method.
NO2 and H2 concentrations, linear flow velocity of the gas, and
the total pressure. Figure 5 summarizes the observed data of
(kobs - kw) for CF3CH2OH and CF3CH(OH)CF3 plotted against
concentrations. From each set of the plot of kobs versus reactant
concentration, the background value (kw) is subtracted, and the
results are shown in Figure 5 as functions of CF3CH2OH and
CF3CH(OH)CF3 concentrations. Both plots give linear relation-
ships with relatively small scatters. Thus, it is concluded that
the present rate constants derived from the least-squares fits are
not affected by any of the experimental factors such as total
pressure, linear velocity of gas, or H2 and NO2 concentrations.
Figure 6 shows a typical example of a pseudo-first-order OH
decay plot obtained with the LP method for various CF3CH2-
OH concentrations. In Figure 7, the observed pseudo-first-order
rate constants, which are derived from the slopes of the straight
lines by least-squares fit to each decay plot, are plotted against
CF3CH2OH concentration. In the case of FP method, both of
the pseudo-first-order plots and the kobs versus reactant con-
centration plots give linear relationships similar to the ones for
the DF and LP methods shown in Figures 3, 4, 6, and 7. In the
case of LP and FP methods, the intercept of kobs versus reactant
concentration for zero reactant (kd) is attributed to the diffusion
of OH radicals from the viewing zone, and is partially caused
by the reaction of OH with impurities contained in the gas
mixture. The kd was again dependent on the experimental
conditions. For each set, the value of kd was subtracted from
kobs to plot the latter against reactant concentrations. Figure 8
summarizes the example of observed data of (kobs - kd) for
CF3CH2OH and CF3CH(OH)CF3 plotted against concentrations
obtained with LP method. The plots of kobs - kd versus reactant
concentration give a linear relationship, where the scatter of
the data points is about the same as in the case of the DF method
shown in Figure 5. In the case of the FP method, the linearity
and the scatter of the plots of kobs - kd versus reactant
concentration is similar to the ones for the DF and LP methods
shown in Fiures 5 and 8. Thus, it can be said that the rate
constants obtained for the LP and FP methods are free from
Figure 4. Plot of the observed pseudo-first-order rate constant kobs
against the CF CH OH concentration. DF-LIF method. P ) 5 Torr, U
9.8 m s , T ) 298 K.
3
2
-
1
)
linear least-squares fit to the observed data. The small intercept
observed when there is no reactant (kw) is attributable to the
wall loss of OH radicals and also to a gas-phase reaction as
OH + NO + M f HNO + M
2
3
Hence, the intercept depends on the experimental conditions.
To ensure that there are no systematic errors, the experiments
were repeated at intervals ranging from several days to a few
months under a variety of experimental conditions including

2668 J. Phys. Chem. A, Vol. 103, No. 15, 1999
Tokuhashi et al.
Figure 6. Pseudo-first-order decay of OH for various CF
3
CH
CH
OH]/10 molecule cm ; (O), 0; (b), 0.78; (]), 1.55; (9), 2.39; (0),
.15; (3), 4.00; (4), 4.74.
2
OH
2
-
Figure 8. Plot of the pseudo-first-order rate constants corrected for
concentrations. LP-LIF method. P ) 40 Torr, T ) 298 K. [CF
3
1
5
-3
wall loss, kobs - k
d 3 2 3 3
, against (O) CF CH OH, and (0) CF CH(OH)CF
concentrations at 298 K. LP-LIF method.
3
TABLE 1: Rate Constants for the Reactions of OH Radicals
with Fluorinated Alcohols at Room Temperature
1
4
fluorinated
alcohol
k × 10
(cm molecule s-1
3
-1
)
techniquea referenceb
CF
3
CH
2
OH
10.7 ( 0.5
DF-LIF
LP-LIF
FP-LIF
FP-RF
LP-LIF
DF-LIF
LP-LIF
FP-LIF
LP-LIF
LP-LIF
LP-LIF
DF-LIF
LP-LIF
FP-LIF
this work
this work
this work
[4]
9
9
9
9
.86 ( 0.41
.68 ( 0.23
.55 ( 0.71
.44 ( 0.96
[5]
CF
3
CF
2
CH
2
OH
10.5 ( 0.5
this work
this work
this work
9
9
8
9
9
.78 ( 0.37
.74 ( 0.35
.66 ( 0.68
.29 ( 0.56
.29 ( 0.77
c
[5]
d
[5]
e
[5]
CF
3
CH(OH)CF
3
2.58 ( 0.12
this work
this work
this work
2
2
.65 ( 0.20
.34 ( 0.12
a
DF, discharge flow; LP, laser photolysis; FP, flash photolysis; LIF,
b
laser-induced fluorescence; RF, resonance fluorescence. The quoted
errors of this work and ref 4 represent 95% confidence level, and 2 σ
for ref 4. c Measured at total pressure of 50 Torr. Measured at total
d
e
pressure of 75 Torr. Measured at total pressure of 100 Torr.
are not considered. The systematic errors in our experiments
are estimated to be less than (10%. As is apparent in Table 1,
the rate constants measured by the three different methods agree
with each other within the estimated uncertainties. The literature
data of the OH with CF3CH2OH are listed in Table 1. The
Figure 7. Plot of the observed pseudo-first-order rate constant kobs
against CF CH OH concentration. LP-LIF method. P ) 40 Torr, T )
98 K.
3
2
2
4
,5
present results agree with the ones in references within the
estimated uncertainties.
the daily fluctuation of experimental conditions such as the total
pressure, residence time of gas mixture, H2O and N2O concen-
trations, laser power, and discharge energy of the Xe flash lamp,
at the ordinary experimental conditions.
Table 1 summarizes the OH reaction rate constants for CF3-
CH2OH and CF3CH(OH)CF3 at room temperature (298 K)
obtained by using DF, LP, and FP methods. Here, the error
limits of our results are at the 95% confidence level derived
from the linear least-squares fit to the plot of first-order rate
constant versus reactant concentration, and the systematic errors
As was stated at the beginning of this section, gas chromato-
graphic analysis showed that the purity for the CF3CF2CH2OH
sample is 99.56%, which is not very good. To examine the effect
of reactive impurities on the measured rate constant of OH with
CF3CF2CH2OH, the sample of CF3CF2CH2OH was subjected
to purification process by means of a gas chromatographic
method. The OH reaction rate constants for CF3CF2CH2OH at
298 K measured by using the original sample (99.56% purity)
-
13
-14
were (1.10 ( 0.06) × 10 , (9.82 ( 0.34) × 10 , and (1.02

OH Reaction Rate Constants for Fluorinated Alcohols
J. Phys. Chem. A, Vol. 103, No. 15, 1999 2669
attached to the neighboring carbon atom by C2F5 group is small
for the OH radical rate constant.
On the other hand, a method of predicting the OH reaction
rate constants has been developed by Atkinson and co-worker
based on the structure activity relationship (SAR).6 This
method itself can be applied to a wide range of reactions such
as the H-atom abstraction from C-H and O-H bonds, OH
addition to double and triple bonds, OH addition to aromatic
rings, and OH interaction with N-, S-, and P-containing groups.
However, it has been reported that the method does not work
well for the compounds such as CHF3, CH3CF3, CHF2CF2CF2-
-8
8
9
CF2H, and fluorinated ethers. For example, the predicted rate
8
constants for CF3CH2OH at 298 K is around four times larger
than the present result. The cause of disagreement between the
predicted and measured rate constants for fluorinated ethers can
be attributed to the fact that the SAR method considers only
9
next-neighbor atomic groups. For halogenated ethers, the
agreement is improved by introducing new substituent groups
8
such as -OCF3. On the other hand, DeMore has proposed to
consider the influence of the third group when two atoms are
1
0
already attached to the C-H bond. For F-, Cl-, and Br-
containing C1-C5 alkanes, the rate constants predicted by
DeMore’s method agree well with the experimental data.
However, the OH reaction rate constants of fluorinated alcohols
such as those reported in the present paper cannot be well
predicted by the existing methods.
Figure 9. Arrhenius plots of the reaction rate coefficients for OH with
CF CH OH, CF CF CH OH, and CF CH(OH)CF . The solid lines
3 2 3 2 2 3 3
represent least-squares fit to the data. The error bars represent 95%
confidence level from linear least-squares analysis. (O), DF-LIF; (0),
LP-LIF; (]), FP-LIF.
In this paper, we have reexamined the prediction method of
the H-atom abstraction rate constants from C-H and O-H
bonds. The method described here is similar to the ones such
-
13
cm molecule s-1 for DF, LP, and FP
3
-1
6-8
10
(
0.04) × 10
as reported by Atkinson
and DeMore. Here, the total
methods, respectively. On the other hand, the values measured
reaction rate constant for the H abstraction is obtained as the
sum of contributions from each C-H and/or O-H bonds present
in the molecule.
for the purified sample (99.935% purity) were (1.05 ( 0.05) ×
-
13
-14
-14
3
1
0
, (9.78 ( 0.37) × 10 , and (9.74 ( 0.35) × 10
cm
-
1
-1
molecule
s
for DF, LP, and FP methods, respectively.
Although the concentrations of individual impurities contained
in the purified sample is about one-seventh of the original
sample, the measured rate constants for the two samples with
three different methods all agree with each other within the
estimated uncertainties. For the purified sample of CF3CF2CH2-
OH, even if the OH rate constants of impurities are as large as
ktotal
)
∑ [k(CH)
F(-X) F(-Y) F(-Z) MP ] +
I
I
I
III
∑ [k(CH)
F(-X) F(-Y) MP ]
I
II
II
∑ [k(CH)
F(-X) ] +
III ∑ [k(OH)
F(-X)I]
-
11
3
-1 -1
where, k(CH) and k(OH) are the rate constants per C-H and O-H,
and F(-X), F(-Y), and F(-Z) are the substituent factors for
substituent groups -X, -Y, and -Z. The subscript I for
substituent factor, F(-X) indicates that only one substituent
group -X is attached to the target C-H. The subscript II and
III for F(-X) indicate that the same two or three atoms (or
atomic groups) are attached to the C-H. The multiplier, MPII,
MPIII, represent the synergistic effect when different two or three
substituent groups are attached to the C-H. Here, when -F
atoms and substituent groups containing fluorine atoms such
as -CF3, -CHnF(3-n) (n ) 1,2), -CF2-, and -CHF- are
attached to the target C-H, these numbers are considered to
evaluate the multiplier. On the other hand, when -H atom,
(1.0 × 10
cm molecule s ), the influence of impurities
on the measured rate constants may be 6.5% or less, while for
the original sample the influence is estimated to be 44% at the
worst case. Thus, it is concluded that the effect of impurities
on the measured rate constant for purified CF3CF2CH2OH is
negligibly small. As shown in Table 1, our results of the OH
reaction rate with CF3CF2CH2OH agree with the ones in ref 5
within the estimated uncertainties. At lower temperatures,
however, since the effect of reactive impurities on the rate
constant measurement may become large if the corresponding
activation energies are small, the purified sample of CF3CF2-
CH2OH was used in the further kinetic measurements.
The OH reaction rate constants of CF3CH2OH, CF3CF2CH2-
OH, and CF3CH(OH)CF3 over the temperature range 250-430
K, measured using different methods are listed in Tables 2, 3,
and 4, respectively. The experimental conditions of individual
measurements are also listed in these tables. Figure 9 shows
the Arrhenius plots for the OH reactions for these fluorinated
alcohols. The Arrhenius plots give linear relationships in the
temperature range examined. The differences among the results
obtained by the three different methods are small. The Arrhenius
rate parameters for the OH reactions of the fluorinated alcohols
are listed in Table 5. The preexponential factors and activation
energies of CF3CH2OH and CF3CF2CH2OH both have very
similar values. This shows that the substitution of a CF3 group
-CH3, -CH2-, -CH< are attached to the C-H, this number
is not considered for the evaluation of the multiplier. For
example, the estimated rate constant of CF3CHFCHF2 (HFC-
236ea), which include two different active sites, is obtained from
the following equation:
k ) k(CH) F(-CF ) F(-F) F(-CHF ) MP +
3 I
I
2 I
III
k(CH) F(-CHF-) F(-F) MP
II
I
II
The rate parameters, k(CH), k(OH), substituent factors (F(-X)),
and multipliers (MPII, MPIII) are determined from the literature
data shown in Table 6 by using a nonlinear least-squares

2
670 J. Phys. Chem. A, Vol. 103, No. 15, 1999
Tokuhashi et al.
2
OH] range no. of
TABLE 2: Experimental Conditions and Results for Measurements of OH Radicals with CF
3
CH
[CF
2
OHa
CH
1
4
c
temperature,
K)
k × 10
U or ht range
pressure range,
(Torr)
3
methodb
(cm molecule s-1
3
-1
)
(m s or s)
-1
× 10 (molecule cm-3)
-15
expts
(
2
50
LP
FP
LP
DF
LP
FP
DF
FP
LP
DF
LP
FP
5.66 ( 0.22
6.01 ( 0.23
7.30 ( 0.38
10.7 ( 0.5
9.86 ( 0.41
9.68 ( 0.23
14.0 ( 0.5
12.4 ( 0.4
17.8 ( 0.5
26.7 ( 0.8
24.4 ( 1.0
26.1 ( 0.7
0.26-0.35
0.32-0.41
0.22-0.34
6.6-10.6
0.32-0.47
0.24-0.55
8.0-9.9
0.25-0.41
0.25-0.35
11.9-14.1
0.26-0.34
0.26-0.34
20-40
20-40
20-40
4-5
20-40
10-40
4-5
20-40
20-40
4-5
20-40
20-40
1.04-9.38
0.82-9.11
0.67-9.58
0.15-0.91
0.46-4.74
0.70-9.53
0.16-0.93
1.15-8.71
0.90-9.34
0.11-0.57
0.98-9.70
0.72-9.33
6
6
6
8
6
6
6
6
6
6
6
6
2
2
73
98
3
31
3
4
75
30
a
The quoted errors represent 95% confidence level from linear least-squares analysis. b DF, discharge flow; LP, laser photolysis; FP, flash
c
photolysis. Linear flow velocity for DF method; ht , residence time for FP and LP methods.
TABLE 3: Experimental Conditions and Results for Measurements of OH Radicals with CF
3
CF
2
CH
2
OHa
1
4
c
temperature,
K)
k × 10
U or ht range
pressure range,
(Torr)
[CF
3
CH
2
OH] range
no. of
expts
3
-1
-1
× 10 (molecule cm-3)
-15
(
methodb
(cm molecule s-1
)
(m s or s)
2
50
LP
FP
LP
DF
LP
FP
FP
LP
LP
FP
6.42 ( 0.27
6.31 ( 0.21
8.01 ( 0.27
10.5 ( 0.5
9.78 ( 0.37
9.74 ( 0.35
12.1 ( 0.4
17.6 ( 0.7
23.4 ( 0.9
23.6 ( 0.7
0.29-0.51
0.32-0.40
0.33-0.44
6.6-8.1
0.25-0.33
0.33-0.52
0.35-0.43
0.27-0.35
0.26-0.33
0.24-0.33
20-40
20-30
15-40
5
20-30
15-40
15-40
15-30
15-30
20-30
0.54-4.92
0.38-4.48
0.67-4.75
0.11-0.76
0.56-4.77
0.48-4.57
0.41-4.85
0.39-3.75
0.54-3.83
0.29-3.82
6
6
7
7
5
5
6
6
6
6
2
2
73
98
3
3
4
31
75
30
a
The quoted errors represent 95% confidence level from linear least-squares analysis. b DF, discharge flow; LP, laser photolysis; FP, flash
c
photolysis. U, linear flow velocity for DF method; ht , residence time for FP and LP methods.
TABLE 4: Experimental Conditions and Results for Measurements of OH Radicals with CF
3 3
CH(OH)CF a
1
4
c
temperature,
K)
k × 10
U or ht range
pressure range,
(Torr)
[CF
3
CH
2
OH] range
no. of
expts
3
-1
-1
× 10 (molecule cm-3)
-15
(
methodb
(cm molecule s-1
)
(m s or s)
2
50
LP
FP
LP
DF
LP
FP
DF
FP
LP
DF
LP
FP
1.40 ( 0.07
1.43 ( 0.05
1.79 ( 0.11
2.58 ( 0.12
2.65 ( 0.20
2.34 ( 0.12
3.45 ( 0.14
3.09 ( 0.14
4.83 ( 0.21
7.54 ( 0.29
7.11 ( 0.30
7.23 ( 0.24
0.30-0.42
0.32-0.43
0.31-0.45
5.4-7.2
0.32-0.42
0.19-0.57
5.9-8.8
0.25-0.39
0.26-0.35
8.4-11.2
0.26-0.36
0.26-0.34
20-60
20-40
20-70
5-6
0.78-9.24
0.66-9.38
0.79-9.06
0.21-2.31
1.24-9.20
0.43-9.76
0.17-1.83
1.29-9.15
0.77-9.36
0.19-1.45
1.33-9.17
0.75-9.69
7
6
5
7
7
10
6
5
7
7
2
2
73
98
20-60
20-60
5-6
3
31
20-40
20-40
4-6
3
4
75
30
20-40
20-40
7
6
a
The quoted errors represent 95% confidence level from linear least-squares analysis. b DF, discharge flow; LP, laser photolysis; FP, flash
c
photolysis. U, linear flow velocity for DF method; ht , residence time for FP and LP methods.
TABLE 5: Arrhenius Rate Parameters for the Reactions of
The following three estimation methods in addition to the
OH Radicals with CF
3 2 3 2 2
CH OH, CF CF CH OH, and
method described above were examined: (i) The synergistic
effects of the -H atom, the -CH3 group, etc., are taken into
consideration together with the -F atom and the fluorine-
containing groups, (ii) the substituent factors of the -H atom
are not taken into consideration (assuming F(-H)I ) F(-H)II
a
CF CH(OH)CF over the Temperature Range 250-430 K
3
3
A ×
molecule s-1
k298 ×
3
molecule s-1
1
2
3
14 b
fluorinated
alcohol
10 (cm
E/R,
(K)
10 (cm
purity
(%)
-
1
-1
)
)
CF3CH2OH
CF3CF2CH2OH
CF3CH(OH)CF3 0.699 ( 0.156 990 ( 70
2.00 ( 0.37
1.40 ( 0.27
890 ( 60
780 ( 60
10.0 ( 0.4
10.2 ( 0.4
2.52 ( 0.11
99.994
99.935
99.999
)
F(-H)III ) 1), and (iii) the synergistic effects of the -F atom,
the -CF3 group, etc., are not taken into consideration (assuming
MPII ) MPIII ) 1). As a result, it has been found that the best
estimation is obtained if the first method is employed.
a
The quoted errors represent 95% confidence level from linear least-
b
squares analysis. Rate constant at 298 K.
The optimized rate parameters, k(CH), k(OH), substituent factors,
and multipliers obtained from the first method are listed in Table
7. The estimated rate constants using optimized rate parameters
are shown in Table 6. The experimental data are reproduced
with an average error of 14.7%. The ratio between the estimated
and experimental rate constants, kest/kexp, have been found to
range from 0.50 to 1.42. For fluorinated alcohols studied here,
procedure to minimize the sum of squares of relative error
between estimated (kest) and experimental (kexp) OH rate
constants, (kest - kexp)/kexp. Here, all parameters except one are
independent and can be determined by the least squares
procedure. F(-CH3) was set equal to unity, as was done by
Atkinson.6
-8

OH Reaction Rate Constants for Fluorinated Alcohols
J. Phys. Chem. A, Vol. 103, No. 15, 1999 2671
TABLE 6: Comparison between Estimated and Experimental OH Rate Constants at 298 K
est cm molecule s-1
3
-1
k
exp cm molecule s-1
3
-1
k
est/kexp
reference
[11]
[11]
[11]
[6]
[6]
[11]
[11]
[11]
compound
k
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
15
13
13
12
12
13
12
12
12
13
13
14
16
14
14
15
15
15
14
15
14
14
14
15
15
16
15
14
14
15
15
14
15
15
-15
CH
4
8.30 × 10
3.70 × 10
7.56 × 10
1.54 × 10
2.37 × 10
8.05 × 10
3.75 × 10
3.24 × 10
6.32 × 10
1.00 × 10
1.04 × 10
2.51 × 10
2.74 × 10
1.00 × 10
1.64 × 10
1.93 × 10
5.14 × 10
4.91 × 10
2.08 × 10
1.07 × 10
7.25 × 10
3.47 × 10
9.28 × 10
1.18 × 10
5.02 × 10
3.97 × 10
8.15 × 10
1.60 × 10
1.57 × 10
4.59 × 10
7.45 × 10
6.01 × 10
3.37 × 10
8.88 × 10
8.30 × 10
1.00
1.37
0.69
0.60
1.00
0.89
1.10
0.61
1.13
1.00
1.02
1.00
1.00
0.94
0.83
1.01
0.90
1.21
1.24
0.82
0.65
0.97
0.50
0.73
0.97
1.17
0.89
1.00
1.06
1.00
0.85
1.42
0.98
1.07
-13
C
C
2
H
H
6
8
2.70 × 10
-12
3
1.10 × 10
-12
n-C
4
H
10
10
2.55 × 10
-12
i-C
4
H
2.37 × 10
-13
CH
3
OH
OH
OH
OH
CH OH
CF CH
CH(OH)CF
(23)
(32)
F (41)
9.00 × 10
-12
2
C H
5
3.40 × 10
-12
n-C
i-C
3
H
7
5.30 × 10
-12
3
H
7
5.60 × 10
[11]
-13
CF
CF
CF
3
2
1.00 × 10
this work
this work
this work
[12, 13]
[12, 14]
[10, 13]
[14]
-13
3
3
2
2
OH
1.02 × 10
-14
3
2.52 × 10
-16 a
CHF
3
2.74 × 10
-14 a
CH
CH
2
F
2
1.06 × 10
-14 a
3
1.97 × 10
-15
CHF
CHF
2
CF
CHF
(134a)
FCHF (143)
CF (143a)
3
(125)
1.90 × 10
-15
2
2
(134)
5.70 × 10
[15
-15 a
CH
CH
CH
CH
CH
CH
2
2
3
2
3
3
FCF
3
4.07 × 10
[15, 16]
[17, 18]
[12, 14, 19]
[18]
[12, 16, 20]
[12, 13]
[12, 21]
[12, 22]
[12]
[12, 21]
[22]
[22]
[13, 23]
[21]
[22]
-14 a
2
1.67 × 10
-15 a
3
1.30 × 10
-13
FCH
CHF
CH
2
F (152)
(152a)
1.12 × 10
-14 a
2
3.57 × 10
-13 a
2
F (161)
(227ea)
CHFCHF (236ea)
CH CF (236fa)
FCF CHF (245ca)
CHF CHFCHF
CF CHFCH
CHF CF CF
CH CH CF
CF CH CH
1.84 × 10
-15 a
CF
CF
CF
3
CHFCF
3
1.61 × 10
-15 a
3
3
2
5.20 × 10
-16
2
3
3.40 × 10
-15 a
CH
2
2
2
9.20 × 10
-14
2
2
(245ea)
1.60 × 10
-14
3
2
F (245eb)
1.48 × 10
-15a
2
2
2
CF
3
2
H (338pcc)
(356ffa)
4.58 × 10
-15
CF
CF
CF
CF
3
3
3
3
2
2
8.75 × 10
-14
2
2
2
F (356mcf)
4.22 × 10
-15 a
CHFCHFCF
2
CF
CF
3
(43-10mee)
3.43 × 10
[13, 23]
[22]
-15
CF CH CH CF
2
2
2
2
3
(55-10mcff)
8.30 × 10
a
Averaged value of references.
TABLE 7: Derived Contribution Factors for OH Rate
Constant Estimations
TABLE 8: The Synergistic Effect of the Substituent Factor
F(-X)
parameter
value
F(-X)II
/
F(-X)III/
3
2
F(-X)
F(-X)
I
F(-X)II F(-X)
0.0194 0.859 0.000653
1.33 0.0000862
0.529 0.715
0.00577 0.000416 12.5
-) 0.00771 0.00280 47.1
I
F(-X)III
F(-X)
I
-
-
12 a
13 a
k
k
(CH)
(OH)
3.18 × 10
1.18 × 10
0.150
F(-H)
0.151
0.192
0.123
0.715
F(-H)
I
F(-F)
0.0888 0.0105
1.00a
0.529
F(-H)II
F(-H)III
0.0194
3
F(-CH )
0.000653
3
F(-CF )
F(-CHF
b
F(-CH
F(-CH
F(-CH
F(-CH
3
3
3
2
)
)
)
I
1.00
2
II
III
0.529
0.715
0.675
0.171
3.71
0.0888
0.0105
0.0000862
0.00577
0.000416
0.00771
0.00280
0.0426
a
By definition.
-)
I
The -OH group of alcohol remarkably increases the C-H
reactivity. The substituent factors for the -F atom and fluorine-
containing groups are smaller than ones for the -H atom and
-CH3 indicating that they decrease the reactivity of the C-H
group when substituted for -H atom or -CH3. In particular,
the -CF3 group most diminishes the C-H reactivity. The
substituent factor of -CF2- is almost equivalent to that of
-CF3. In fact, there is no significant difference between the
reactivities of CF3CH2OH and CF3CF2CH2OH. For -CHF2,
F(-CH<)
F(-OH)
I
I
F(-F)
I
F(-F)II
F(-F)III
F(-CF
F(-CF
3
)
I
3
)
II
F(-CHF
F(-CHF
2
)
)
I
2
II
I
2
F(-CH F)
F(-CF
2
-)
I
0.00688
0.0121
-
CH2F, and -CHF-, the C-H reactivity is decreased with
F(-CHF-)
MPII
I
increase of the number of F atom. The multipliers for the
synergistic effect between different substituent groups (MPII and
MPIII) are larger than unity, and MPIII is larger than MPII. Thus,
when two or three different substituent groups are attached to
C-H, the degree of the decrease in the C-H reactivity by
individual substituent groups becomes small.
10.0
78.1
MPIII
a
Units are cm molecule s-1. b By definition.
3
-1
the estimated rate constants reproduce the experimental values
within a factor of 1.00 to 1.02.
The synergistic effects of substituent factor, F(-X), are
Inspection of Table 7 reveals some general features of the
substitution effects. The fact that the substituent factors of -CH3
and -CH2- are close to unity indicates that these substituent
groups increase the reactivity of C-H compared with others.
summarized in Table 8. The substituent factors, F(-H) , F(-
III
F) , F(-CH ) , and F(-CHF ) , are determined by using the
III
3 III
2 II
data of CH4, CHF3, i-C4H10, and CHF2CHFCHF2 (HFC-245ea),
respectively, and are irrespective of other compounds. Since

2672 J. Phys. Chem. A, Vol. 103, No. 15, 1999
Tokuhashi et al.
2
F(-CF3)II/F(-CF3)I is larger than unity, the degree of decreas-
ing the C-H reactivity per one -CF3 decreases if two -CF3
groups are attached together to the C-H. A similar tendency is
also noted for the relationship between F(-F)I and F(-F)II, and
that between F(-CHF2)I and F(-CHF2)II. This is in accord with
the fact that the corresponding multipliers are larger than unity.
As for F(-F)I, F(-F)II, and F(-F)III, F(-F)III is by far smaller
than the others. It seems that the reactivity of OH with CHF3
cannot be explained in line with other molecules. A similar
tendency is also recognized in the relations among F(-H)I,
F(-H)II, and F(-H)III.
Acknowledgment. The present investigation has been
performed with the support of the New Energy and Industrial
Technology Development Organization.
References and Notes
(1) “Scientific Assessment of Ozone Depletion: 1994”, UNEP/WMO,
Global Ozone Research and Monitoring Project - Report No.37.
(2) Burrows, J. P.; Wallington, T. J.; Wayne, R. P. J. Chem. Soc.,
Faraday Trans. 2 1983, 79, 111.
(3) Kaufman, F. J. Prog. React. Kinet. 1961, 1, 1.
(
4) Wallington, T. J.; Dagaut, P.; Kurylo, M. J. J. Phys. Chem. 1988,
9
2, 5024.
In this report, the synergistic effects due to multiple sub-
stituent groups (MPII and MPIII) were obtained irrespective of
the magnitude of individual substituent factors. However, as is
seen in Table 8, the synergistic effect of substituent groups
seems to depend on the magnitude of substituent factors. Thus,
it seems necessary to consider the following items in order to
obtain the estimated rate constants for a wider range of
compounds with enough accuracy. First, the synergistic effect
between different substituent groups may depend on the
magnitude of individual substituent factors. Therefore, the
improvement of the accuracy of estimation can be expected by
considering the magnitude of individual substitution effects to
synergistic effect between the substituent groups. Second, it is
necessary to consider the method of obtaining the substituent
factors from individual constituent atoms, though in this work
the substituent factors were treated for groups in such cases as
(
5) Inoue, G.; Izumi, K.; Lozovsky. V. A. Third International Confer-
ence on Chemical Kinetics, Gaithersburg, MD, 1993; pp 182-183.
(6) Atkinson, R. Chem. ReV. 1986, 86, 69.
(
(
(
7) Atkinson, R. Int. J. Chem. Kinet. 1987, 19, 799.
8) Kwok, E. S. C.; Atkinson, R. Atmos. EnViron. 1995, 29, 1685.
9) Zhang, Z.; Saini, R. D.; Kurylo, M. J.; Huie, R. E. J. Phys. Chem.
1
992, 96, 9301.
(10) DeMore, W. B. J. Phys. Chem. 1996, 100, 5813.
(11) Atkinson, R.; Baulch, D. L.; Cox, R. A.; Hampspn, R. F., Jr.; Kerr,
J. A.; Troe, J. J. Phys. Chem. Ref. Data 1989, 18, 881.
12) Hsu, K. J.; DeMore, W. B. J. Phys. Chem. 1995, 99, 1235.
13) Schmoltner, A. M.; Talukdar, R. K.; Warren, R. F.; Mellouki, A.;
Goldfarb, L.; Gierczak, T.; McKeen, S. A.; Ravishankara, A. R. J. Phys.
Chem. 1993, 97, 8976.
14) Talukdar, R. K.; Mellouki, A.; Gierczak, T.; Burkholder, J. B.;
McKeen, S. A.; Ravishankara, A. R. J. Phys. Chem. 1991, 95, 5815.
15) DeMore, W. B. Geophys. Res. Lett. 1993, 20, 1359.
16) Gierczak, T.; Talukdar, R. K.; Vaghjiani, L. V.; Lovejoy, E. R.;
Ravishankara, A. R. J. Geophy. Res. 1991, 96, 5001.
(
(
(
(
(
-
CH3 and -CF3. This means, however, that the long-range
(17) Barry, J.; Sidebottom, H.; Treacy, J.; Franklin, J. Int. J. Chem. Kinet.
995, 27, 27.
1
effects of substitution are taken into consideration. As a result,
it may become possible to estimate the rate constants of such
molecules as ethers with sufficient accuracy.
(
(
18) Martin, J.-P.; Paraskevopoulos, G. Can. J. Chem. 1983, 61, 861.
19) Orkin, V. L.; Huie, R. E.; Kurylo, M. J. J. Phys. Chem. 1996, 100,
8
907.
(20) Liu, R.; Huie, R. E.; Kurylo, M. J. J. Phys. Chem. 1990, 94, 3247.
(21) Zhang, Z.; Padmaja, S.; Saini, R. D.; Huie, R. E.; Kurylo, M. J. J.
Phys. Chem. 1994, 98, 4312.
22) Nelson, D. D., Jr.; Zahniser, M. S.; Kolb, C. E.; Magid, H. J. Phys.
Chem. 1995, 99, 16301.
23) Zhang, Z.; Kurylo, M. J.; Huie, R. E. Chem. Phys. Lett. 1992, 200,
230.
Finally, the temperature dependency of the reaction rate
constants was not examined in the present study. However, as
has been done by Atkinson et al., the temperature dependency
of the rate constants may be reproduced by applying Arrhenius’s
expression to k(CH) and k(OH), and exp(EX/T) to the substituent
factor F(-X).
(
(