Threshold energies and unimolecular rate constants for elimination of HF from chemically activated CF3CH2CH3 and CF3CH2CF3: Effect of CH3 and CF3 substituents at the β-carbon and implications about the transition state

Article abstract of DOI:10.1021/jp972854k

Chemically activated CF₃CH₂CF₃ was prepared with 104 kcal/mol of internal energy by the combination of CF₃CH₂ and CF₃ radicals, and chemically activated CF₃CH₂CH₃ was prepared with 101 and 95 kcal/mol by combination of CF₃ and CH₂CH₃ radicals and by combination of CF₃CH₂ and CH₃ radicals, respectively. The experimental rate constants for unimolecular 1,2-dehydrofluorination were 1.2 × 10⁵ s^-1 for CF₃CH₂CF₃ and 3.2 × 10⁶ s^-1 for CF₃CH₂CH₃ with 95 kcal/mol and 2.0 × 10⁷ s^-1 with 101 kcal/mol of energy. Fitting the calculated rate constants for HF elimination from RRKM theory to the experimental values provided threshold energies, E₀, of 73 kcal/mol for CF₃CH₂CF₃ and 62 kcal/mol for CF₃CH₂CH₃. Comparing these threshold energies to those for CF₃CH₃ and CF₃CH₂Cl illustrates that replacing the hydrogen of CF₃CH₃ with CH₃ lowers the E₀ by 6 kcal/mol and replacing with CF₃ or Cl raises the E₀ by 5 and 8 kcal/mol, respectively. The CF₃ substituent, an electron acceptor, increases the E₀ an amount similar to Cl, suggesting that chlorine substituents also prefer to withdraw electron density from the β-carbon. As the HF transition state forms, it appears that electron density flows from the departing hydrogen to the β-carbon and from the β to the α-carbon, to the α-carbon from its substituents, but the α-carbon releases most of the incoming electron density to the departing fluorine. The present work supports this scenario because electron-donating substituents, such as CH₃, on either carbon would reduce the E₀ as they aid the flow of negative charge, while electron-withdrawing substituents such as Cl, F, and CF₃ would raise the E₀ for HF elimination because they hinder the flow of electron density.

Full text of DOI:10.1021/jp972854k

J. Phys. Chem. A 1998, 102, 5393-5397
5393
Threshold Energies and Unimolecular Rate Constants for Elimination of HF from
Chemically Activated CF3CH2CH3 and CF3CH2CF3: Effect of CH3 and CF3 Substituents at
the â-Carbon and Implications about the Transition State
Heather A. Ferguson, John D. Ferguson, and Bert E. Holmes*
Department of Chemistry, Lyon College, BatesVille, Arkansas 72501
ReceiVed: September 2, 1997; In Final Form: NoVember 17, 1997
Chemically activated CF3CH2CF3 was prepared with 104 kcal/mol of internal energy by the combination of
CF3CH2 and CF3 radicals, and chemically activated CF3CH2CH3 was prepared with 101 and 95 kcal/mol by
combination of CF3 and CH2CH3 radicals and by combination of CF3CH2 and CH3 radicals, respectively.
The experimental rate constants for unimolecular 1,2-dehydrofluorination were 1.2 × 105 s-1 for CF3CH2CF3
and 3.2 × 106 s-1 for CF3CH2CH3 with 95 kcal/mol and 2.0 × 107 s-1 with 101 kcal/mol of energy. Fitting
the calculated rate constants for HF elimination from RRKM theory to the experimental values provided
threshold energies, E0, of 73 kcal/mol for CF3CH2CF3 and 62 kcal/mol for CF3CH2CH3. Comparing these
threshold energies to those for CF3CH3 and CF3CH2Cl illustrates that replacing the hydrogen of CF3CH3 with
CH3 lowers the E0 by 6 kcal/mol and replacing with CF3 or Cl raises the E0 by 5 and 8 kcal/mol, respectively.
The CF3 substituent, an electron acceptor, increases the E0 an amount similar to Cl, suggesting that chlorine
substituents also prefer to withdraw electron density from the â-carbon. As the HF transition state forms, it
appears that electron density flows from the departing hydrogen to the â-carbon and from the â to the R-carbon,
to the R-carbon from its substituents, but the R-carbon releases most of the incoming electron density to the
departing fluorine. The present work supports this scenario because electron-donating substituents, such as
CH3, on either carbon would reduce the E0 as they aid the flow of negative charge, while electron-withdrawing
substituents such as Cl, F, and CF3 would raise the E0 for HF elimination because they hinder the flow of
electron density.
Introduction
view is consistent with experimental observations2-5 that CF3,
Cl, and F increase E0’s, but a CH3 group has the opposite effect.
For CH2FCH3, CHF2CH3, and CF3CH3, the calculations6 also
predict that the â-carbons acquire electron density, probably
from the attached H’s, and transfer a portion of this charge to
the R-carbon as the HF departs; the changes in atomic charges
on the â-carbon are -0.03e, -0.07e, and -0.09e, respectively.
Although the ab initio results6 did not test the effect of â-carbon
substituents, because both carbons acquire electron density it
is reasonable to speculate that â-carbon and R-carbon substit-
uents would have the same effect on the E0.
From the E0(HF) for CF3CH3 (68 kcal/mol)7 and CF3CH2Cl
(76 kcal/mol)3 it is seen that replacing an H with a Cl at the
â-carbon raises the E0. The E0(HF)’s for two series of
hydrofluoroethanes show the same general trends for successive
fluorine for hydrogen substitution at the â-carbon: CH2FCH3
(57 kcal/mol),7,8 CH2FCH2F (61.5),9 and CH2FCHF2 (68)10-12
and CHF2CH3 (61),13 CHF2CH2F (68),10-12 and CHF2CHF2
(74.5).10,12 This illustrates that, with one F at the R-carbon,
successive substitution at the â-carbon raises the E0 by 4.5-
6.5 kcal/mol and, with two F’s at the R-carbon, the E0 increases
by about 7 kcal/mol. It appears that additional fluorines at the
R-position magnify the effect of F substituents at the â-position.
The ab initio calculations6 support this supposition because the
change in charge at the â-carbon is most pronounced, -0.09e,
when there are three F’s at the R-carbon. This suggests that a
CF3CH2X series (X ) substituent) would be the best to
investigate because the effect of â-carbon substituents should
be the most pronounced. Although F and Cl substituents on
the â-carbon increase the E0(HF), it is not known whether they
For several years we have been measuring1-5 unimolecular
rate constants for the 1,2-dehydrohalogenation of chemically
activated hydrofluorocarbons and hydrochlorofluorocarbons, and
by fitting rate constants calculated using the RRKM theory to
the experimental values, threshold energies, E0’s, have been
determined. By altering substituents and measuring their effect
on the rate constant and on the threshold energy barrier, it should
be possible to better understand the flow of electron density as
the four-centered transition state for chloro- and fluoroalkanes
forms and also to understand whether substituents act as electron
donors or acceptors.
Recent work2-5 by this laboratory for HF elimination has
shown that CF3, Cl, and F substituents at the R-carbon all raise
the threshold energy, and because CF3 accepts electron density,
the Cl and F substituents presumably have a similar inductive
effect. The R-carbon contains the halogen that is eliminated,
and in this paper, it is the first carbon in the chemical formula.
According to theoretical calculations by Toto, Pritchard, and
Kirtman6 for HF elimination from hydrofluorocarbons, the
R-carbon acquires electron density from the substituents and
from the â-carbon as the transition state forms. This might
suggest, contrary to observations, that electron-withdrawing
substituents would help disperse the electron density accumulat-
ing on the R-carbon, thus stabilizing the transition state and
lowering the E0. However, the calculations also predict that
electron density flows from the substituents to the R-carbon and
to the departing F, and this suggests that an electron-withdrawing
group would hinder the transfer of electron density to the
R-carbon, thereby raising the threshold energies. This latter
S1089-5639(97)02854-5 CCC: $15.00 © 1998 American Chemical Society
Published on Web 02/18/1998

5394 J. Phys. Chem. A, Vol. 102, No. 28, 1998
Ferguson et al.
are donating or withdrawing electron density as the transition
state forms because F and Cl can function in either manner. To
determine the behavior of the F and Cl, the H at the â-carbon
will be replaced with a CF3 substituent, an electron withdrawer,
and a CH3, an electron donor. In addition, this comparison
should provide information about the development of atomic
charges on the carbons as the transition state forms. In this
paper we will report unimolecular rate constants and E0’s for
HF elimination from CF3CH2CF3 (X ) CF3) and from CF3-
CH2CH3 (X ) CH3) that will ascertain the electronic behavior
of the F and Cl substituents at the â-carbon. This information
should also be a useful benchmark when ab initio calculations
appear that test the effect of â-carbon substituents.
Both hydrofluoropropanes were prepared by chemical activa-
tion, and two different reactions were used to produce CF3CH2-
CH3 at two different levels of energy. The chemically activated
CF3CH2CH3* (the asterisk denotes chemical activation), con-
taining 95 kcal/mol of internal energy, was formed by the
combination of CF3CH2 and CH3 radicals produced from the
photolysis of CF3CH2I and CH3I in the presence of Hg2I2.14
The important reaction channels are as follows:
5b or 8b plotted versus inverse pressure should produce a linear
relationship with an intercept of zero. The slope of the D/S
versus 1/P plot equals kHF/kM, which is converted to kHF by
calculation of kM using collision theory.
Experimental Section
For the CF3CH2CH3 chemical activation study, vessels with
a volume range of 14.85-2115.4 cm3 containing 44.6-89.8
µmol of CF3I and CH3CH2I or CH3I and CF3CH2I were
photolyzed 5-120 min at room temperature using a high-
pressure Oriel 6137 mercury lamp. The Pyrex vessels contained
Hg2I2 which removes iodine atoms to aid in the production of
halocarbon radicals.14 All samples were prepared on a grease-
free vacuum line using an MKS 270 electric manometer. A
Perkin-Elmer 3920 GC equipped with a flame ionization
detector and an 1/8 in. by 2 m stainless steel Porapak R column
was used for product analysis. An initial GC temperature of
60 °C followed by immediate temperature programming of 2
°C/min to a final temperature of 210 °C was used for both
methods of activating CF3CH2CH3. For the CF3I and CH3CH2I
system, elution times (in minutes) were generally as follows:
CF3CF3 (4.5), C2H6 (5.2), CH2dCH2 (5.2), CF2dCHCH3 (17),
CF3CH2CH3 (23), CF3I (28), C4H10 (30), and CH3CH2I (58).
The C2H6 and CH2dCH2 arise from the disproportionation
reaction between two ethyl radicals. For the CH3I with CF3-
CH2I system, elution times (in minutes) were typically C2H6
(5.2), CF2dCHCH3 (17), CF3CH2CH3 (23), CF3CH2CH2CF3
(37), CH3I (43), and CF3CH2I (49). The data were collected
and integrated with a Shimadzu Chromatopac CR5A integrator.
2CF3CH2I + Hg2I2
2CH3I + Hg2I2
9
hν8 2CF3CH2‚ + 2HgI2
hν8 2CH3‚ + 2HgI2
(1a)
9
(1b)
(2)
‚CH3 + ‚CH3 f CH3CH3
CF3CH2‚ + CF3CH2‚ f CF3CH2CH2CF3*
CF3CH2‚ + ‚CH3 f CF3CH2CH3*
(3)
Products were identified by comparison of GC retention times
with authentic samples, and the purity and identity were verified
by a Hewlett-Packard 5890/5791 GC/MS equipped with a 50
m PONA capillary column, except for CF2dCHCH3. The
CF2dCHCH3 was identified by its expected pressure depen-
dency of a chemically activated, unimolecular decomposition
system and by the following mass spectrum: (C3H3F2+) m/z )
77 (RA ) 100%), (C3H4F2+) 78 (60.4%), (CF2H+) 51 (37.3%),
(C3F2H+) 75 (18.6%), (C3H2F2+) 76 (14.1%), and (CF+) 31
(7.3%).
(4)
Reaction 4 produced chemically activated CF3CH2CH3* which
eliminated HF (reaction 5a) or, at increased pressures, was
stabilized through collision (reaction 5b).
kHF
CF3CH2CH3* 8 CF2dCHCH3 + HF
(5a)
(5b)
[M]8 CF3CH2CH3
kM
Because an authentic sample of CF2dCHCH3 was unavail-
able, the calibration factor for [CF2dCHCH3]/[CF3CH2CH3] was
estimated by comparing CF3CHdCH2 and CF3CHdCF2 with
CF3CH2CH3. The calibration factors for [CF3CHdCH2]/[CF3-
CH2CH3] and [CF3CHdCF2]/[CF3CH2CH3] were 0.953 ( 0.017
and 1.07 ( 0.07, respectively, using six trials of five mixtures
of the products and reactants to replicate reaction mixtures.
These values were then averaged to obtain the estimated
calibration factor of 1.01 for [CF2dCHCH3]/[CF3CH2CH3].
Photolysis of CH3CH2I and CF3I produced CH3CH2 and CF3
radicals that combined to give CF3CH2CH3, containing 101 kcal/
mol, reaction 6. For simplicity, the self-reactions of ethyl or
CF3 radicals are not shown.
CF3‚ + ‚CH2CH3 f CF3CH2CH3*
(6)
The decomposition or stabilization of CF3CH2CH3* is the same
as shown in reactions 5.
Reaction 7 is the preparation of chemically activated CF3-
CH2CF3 containing 104 kcal/mol of internal energy, which is
formed by photolysis of CF3CH2I and CF3I. The unimolecular
decomposition of CF3CH2CF3* is shown in reaction 8a.
For the chemically activated CF3CH2CF3 system, the CF3 and
CF3CH2 radicals were generated by photolysis of about 12 µmol
each of CF3I and CF3CH2I in Pyrex vessels containing Hg2I2
and ranging in volume from 11.31 to 2115.4 cm3. Samples
were treated as described for the CF3CH2CH3 systems, except
that the photolysis times varied from 2 to 45 min. Reaction
mixtures were analyzed using a Shimadzu GC-14A gas chro-
matograph equipped with 1/8" x 2 m Haysep A column coated
with 5% Ni. With an initial temperature of 80 °C for 5 min
followed by temperature programming at 2 °C/min to a final
temperature of 167 °C, the retention times were C2F6 (2.5),
CF2dCHCF3 (19), CF3I (27), CF3CH2CF3 (32), CF3CH2CH2-
CF3 (44), and CF3CH2I (70). A Shimadzu Chromatopac CR5A
integrator measured the product yields. Reaction products were
identified by comparing the GC retention times and mass spectra
CF3‚ + ‚CH2CF3 f CF3CH2CF3*
(7)
(8a)
(8b)
kHF
CF3CH2CF3* 8 CF3CHdCF2 + HF
kM
[M]8 CF3CH2CF3
The ratio of the yield of the decomposition (D) products from
reaction 5a or 8a and the stabilization (S) product from reaction

HF Elimination from CF3CH2CH3 and CF3CH2CF3
J. Phys. Chem. A, Vol. 102, No. 28, 1998 5395
Figure 1. A plot of [CF2dCHCH3]/[CF3CH2CH3] versus reciprocal
pressure for the four-centered elimination of HF from chemically
activated CF3CH2CH3 with 95 kcal/mol (circles) and 101 kcal/mol
(squares) of internal energy. The slope is 1.94 Torr, the intercept is
0.004 37, and the correlation coefficient is 0.983 at the higher energy,
and the values are 0.321 Torr, 0.0188, and 0.979, respectively, for CF3-
CH2CH3 with 95 kcal/mol of energy.
Figure 2. A plot of [CF3CHdCF2]/[CF3CH2CF3] versus reciprocal
pressure for the four-centered elimination of HF from chemically
activated CF3CH2CF3; the slope is 0.0117 Torr, the intercept is 0.0118,
and the correlation coefficient is 0.991.
TABLE 1: Experimental Unimolecular Rate Constants for
1,2-HF Elimination from CF3CH2X (X ) H, Cl, CH3, and
CF3), the Average Energy of the Chemically Activated
Molecule, and the Threshold Energy
with commercial samples. The GC detector calibration factor
for [CF3CHdCF2]/[CF3CH2CF3] was 1.0085 ( 0.0110.
molecule
E (kcal/mol)
kHF (s-1
)
E0 (kcal/mol)
Results and Discussion
a
CF3CH3
102
97.5
95
101
104
3.7 × 108
2.8 × 106
3.2 × 106
2.0 × 107
1.2 × 105
68
76
62
62
73
CF3CH2Clb
CF3CH2CH3
CF3CH2CH3
In this section the experimental unimolecular rate constants
will be discussed first, and they will be converted to units of
s-1 using collision cross sections. Next the average energy of
the chemically activated molecules will be calculated, and the
RRKM model will be developed. Then the threshold energies
will be estimated by iteration of the RRKM calculations until
the calculated and experimental rate constants agree at the
energy of activation. The ab initio calculations of Toto et al.6
will then be summarized for CF3CH3, and the threshold energies
for a series of CF3CH2X will be compared to determine the
effect of substituents on the threshold energies. Finally, the
trends in E0’s will be analyzed to help understand the nature of
the 1,2-HF elimination transition state.
c
c
c
CF3CH2CF3
a Reference 7. b Reference 3. c This work.
-333.5 kcal/mol, and adjustment to 0 K gave -329.9 kcal/
mol. For the methyl radical the ∆Hfo(0 K) ) 35.6 kcal/mol,18
and for the ethyl radical the ∆Hfo(0 K) ) 31.0 kcal/mol.19 Data
recommended by Rodgers20 were adopted for the CF3 and the
CF3CH2 radicals; -111.06 and -123.5 kcal/mol at 298 K were
corrected to -110.4 and -120.9 kcal/mol at 0 K, respectively.
The ∆Horxn(0 K) were -90.2 and -96.1 kcal/mol for reactions
4 and 6, respectively; addition of 4.8 kcal/mol thermal energy
for the two radical reactants gives an E ) 95 kcal/mol for
reaction 4 and E ) 100.9 kcal/mol, which was rounded to
101 kcal/mol, for reaction 6. The ∆Horxn(0 K) was -98.6 kcal/
mol for reaction 7; addition of 4.9 kcal/mol thermal energy gives
E ) 103.5 kcal/mol, which was rounded to 104 kcal/mol. An
E value exceeding 100 kcal/mol for the fluoropropanes agrees
with recent ab initio calculations21,22 of the carbon-carbon bond
dissociation energies for a series of fluoroethanes that predicted
the increased ionic character in the C-C bond for CF3CH3
would cause the dissociation energy to be 621 to 1222 kcal/mol
higher than for ethane. We also note that the E for reaction
6 is 5.9 kcal/mol higher than that for reaction 4, which agrees
with the ab initio predictions.21,22
A goal is to extract the threshold energies for 1,2-HF
elimination by comparing experimental unimolecular rate
constants with rate constants calculated using the RRKM theory
and then compare the E0’s for a series of CF3CH2X molecules
(X ) H, Cl, CF3, and CH3) to determine the effect of â-carbon
substituents on the E0’s. To our knowledge, thermal activation
experiments have not been reported for CF3CH2CF3 and CF3-
CH2CH3 so RRKM models were developed and adjusted to a
common thermal preexponential factor (partition function form)
of (9 ( 1) × 1012 s-1 per reaction path at 800 K (Table 2).
Models previously developed for the other CF3CH2X systems,
CF3CH2Cl3 and CF3CH3,7 had been parametrized to a thermal
preexponential factor within the expected limits so the E0’s from
The unimolecular rate constants for elimination of HF from
CF3CH2CH3 and CF3CH2CF3 were determined from the D/S
versus 1/P plots. Figure 1 has a slope of 1.9 Torr for loss of
HF from CF3CH2CH3 containing 101 kcal/mol of internal
energy, E , and a slope of 0.32 Torr for CF3CH2CH3 when
E ) 95 kcal/mol. Figure 2 has a slope of 0.0117 Torr for
loss of HF from CF3CH2CF3. Rate constants in pressure units
were converted to s-1 using temperature-independent collision
diameters for CH3I (5.0 Å), CF3I (5.6 Å), CH3CH2I (5.6 Å),
CF3CH2I (6.2 Å), CF3CH2CH3 (5.6 Å), and CF3CH2CF3 (5.7
Å). Collision diameters for CF3I and CH3I are known,2,3 and
the collision diameters for the other species were estimated by
determining the increase in collision diameters for the replace-
ment of H by an I (+1.21 Å), by substitution of a CF3 group
for a CH3 group (+0.6 Å), and by the effect of an additional
CH2 in a homologous series of alkanes (+0.55 Å).15 The kHF’s
for CF3CH2CH3 with E ) 101 and 95 kcal/mol and for CF3-
CH2CF3 were 2.0 × 107, 3.2 × 106, and 1.2 × 105
s
,
-1
respectively.
The average energy of the chemically activated fluoropro-
panes, E , is the enthalpy for reactions 4, 6, and 7 at 0 K plus
the thermal energy of the reactants. The enthalpy of formation
at 298 K for CF3CH2CH3 has not been reported so it was
estimated as -181 kcal/mol from group additivity schemes,16
and adjustment to 0 K gave -175.5 kcal/mol. At 298 K the
enthalpy of formation for CF3CH2CF3 has been measured17 as

5396 J. Phys. Chem. A, Vol. 102, No. 28, 1998
Ferguson et al.
TABLE 2: Summary of Experimental Rate Constants and RRKM Models for 1,1,1,3,3,3-Hexafluoropropane and
1,1,1-Trifluoropropane
molecule
CF3CH2CF3
activated complex,
elimination of HF
molecule
CF3CH2CH3
activated complex,
elimination of HF
vibrational frequencies, cm-1
and (degeneracies)
3011 (2)
3028 (1)
2954 (5)
2949 (4)
1272 (9)
879 (3)
576 (6)
332 (4)
128 (2)
66 (1)
1352 (5)
974 (6)
717 (3)
530 (4)
320 (5)
100 (2)
1340 (10)
920 (4)
568 (3)
349 (3)
241 (1)
215 (1)
1376 (8)
917 (6)
605 (3)
349 (3)
241 (1)
215 (1)
moments of inertia I‡/I
reaction path degeneracya
preexponential factor,b s-1
Eo, kcal/mol
0.94
8
0.969
4
9.0 × 1012
9.0 × 1012
73
104
62
95, 101
E , kcal/mol
ka(exptl), s-1
1.2 × 105
3.2 × 106 , 2.0 × 107
3.7 × 106, 1.2 × 107
ka(calcd), s-1
1.3 × 105
a Hindered rotor treated as a torsion. b Partition function form for unit reaction path degeneracy at 800 K.
these models were accepted without modification. Vibrational
frequencies for CF3CH2CH3 were estimated from CH3CH2-
CH3,23,24 CF3CF2CF3,25 CH2FCH2CH3,26 and CF3CH2CF3.27,28
The vibrational frequencies for CF3CH2CF3 are known,27,28 and
the frequencies for the four-member transition states were
assigned from previous models2-5 with the ring puckering
adjusted to give a preexponential factor within the desired range.
A torsional model was assumed giving reaction path degenera-
cies of 8 for CF3CH2CF3 and 4 for CF3CH2CH3, and an exact
count was used for the sum of states for the activated complexes.
Table 1 contains experimental rate constants, the E ’s, and the
E0’s for the CF3CH2X series.
The calculated rate constant matched the experimental values
with E0(HF) ) 62 kcal/mol for CF3CH2CH3 at both energies
of activation, and with E ) 104 kcal/mol for CF3CH2CF3 the
E0(HF) ) 73 kcal/mol (see Tables 1 and 2). The largest source
of error is likely the thermochemistry used to calculate the E ,
and the reported accumulated uncertainty is (4 kcal/mol. If
the E ’s are lowered by 4 kcal/mol, then the calculated rate
constant would be too low by about a factor of 1.9. To restore
agreement between the calculated and the experimental rate
constants, the threshold energies would need to be reduced by
2 kcal/mol, which increases the RRKM rate by a factor of 2.3.
On the basis of this analysis, we estimate the uncertainty in
E0’s is (2 kcal/mol. Cadman et al.29 also developed an RRKM
model for CF3CH2CH3 and compared the calculations to
unpublished experimental work for the formation of CF3CH2-
CH3 by radical combination and by insertion of a CH2 into the
C-H bond of CF3CH3. Their RRKM model is based on a
thermal preexponential factor considerably outside the range
we adopted, and their thermochemistry gives E ’s that differ
by 5 kcal/mol from the present results, so their higher E0 (68
kcal/mol) is not unexpected.
predict that although fluorine substituents are electron-
withdrawing and maintain a partial negative charge, they actually
provide electron density to the R-carbon as the transition state
forms.
At the R-carbon, the experimental E0’s are increased by 7-8.5
kcal/mol as an H from CHF2CH3 (E0 ) 61 kcal/mol)7,8 is
replaced by an F (CF3CH3, E0 ) 68 kcal/mol),7 is replaced by
a CF3 (CF3CF2CH3, E0 ) 68.5 kcal/mol),5 or is replaced by a
Cl (CClF2CH3, E0 ) 69.5 kcal/mol),2 suggesting that F’s, CF3’s,
and Cl’s exhibit similar behaviors at the transition state. Since
a CF3 group is known to be electron accepting, the F and Cl
substituents on the R-carbon must also tend to withdraw electron
density. Replacing the H from CHF2CH3 with a CH3 (CH3-
CF2CH3, E0 ) 54 kcal/mol)5 lowers E0 by 7 kcal/mol, which is
consistent with methyl’s ability to donate electron density. At
the R-carbon it appears the magnitude of the effect of an electron
donor (CH3) is nearly equal to the effect of the electron acceptors
(CF3, F, or Cl). The calculation6 predicts that fluorine substit-
uents on the reactant are strong electron acceptors (each F of
CF3CH3 contains -0.31e), but chlorine substituents are nearly
neutral or very slightly electron donors (each Cl of CCl3CH3
contains +0.01e or +0.02e). Apparently, the Cl and F
substituents exert similar inductive effects as the transition state
forms, even though they have different electronic behavior in
the reactant. When relating variation of the E0’s to the
calculated atomic charges on the carbon skeleton, it is assumed
that the substituents exert mainly an inductive, resonance, or
other electronic effect on the atomic charges that develop as
the transition state is formed from the reactant.
At the â-carbon, the replacement of the H by a CF3 or by a
Cl substituent for the series of CF3CH2X molecules (see Table
1) raised the threshold energy by 5 and 8 kcal/mol, respectively,
and a CH3 substituent lowered the threshold energy by 6 kcal/
mol. The effect of the CF3 group on the E0 is in the same
direction and of similar magnitude as a Cl substituent, suggesting
that they attempt to remove electron density from the â-carbon.
It does appear that at the â-carbon a Cl substituent might have
a slightly greater effect than a CF3. The influence of a â-carbon
fluorine substituent on the E0 for CF3CH2X (X ) F) has not
been determined, but the series of fluoroethanes presented in
the Introduction suggests that fluorine for hydrogen substitution
for a series of mono- and difluoroethanes raises the E0 by 4-7
kcal/mol. The ab initio calculations6 predict that both the R-
and â-carbons acquire electron density with a net gain that is
nearly identical (-0.09e and -0.12e, respectively), and this is
Toto et al.6 calculated molecular and transition-state atomic
charges for CF3CH3 and for the four-centered HF-elimination
transition state. The charge on the â-carbon also includes the
atomic charges carried by the two hydrogens which are not being
eliminated. It is predicted that electron density flows from the
departing H to the â-carbon, from the â-carbon to the R-carbon,
from the fluorine substituent to the R-carbon, and from the
R-carbon to the F being eliminated. Although the R-carbon
and the â-carbon maintain different signs in their partial charges,
both the R-carbon and the â-carbon acquire more electron
density than they donate, giving them a net gain in electron
density of -0.09e and -0.12e, respectively. These results

HF Elimination from CF3CH2CH3 and CF3CH2CF3
J. Phys. Chem. A, Vol. 102, No. 28, 1998 5397
in agreement with the experimental observations that the
magnitude of the effect on E0 by electron-donating or -accepting
substituents are similar at each carbon. The ab initio calcula-
tions6 focused on substituent effects at the R-carbon and did
not test â-carbon substituents; however, because methyl, chloro,
and trifluoromethyl substituents cause similar changes in the
E0’s at both the R- and â-carbons, it is tempting to speculate
that â-carbon substituents are also forced to provide electron
density as the HF departs.
R-carbon bringing the H with it. During this stepwise process
electron density might flow to the â-carbon as the H moves to
the lone pair on the F, then to the R-carbon, and finally to the
departing F. This picture was suggested30 to account for both
the relative large release of energy to translational motion and
the H/D kinetic isotope effect. A process where the first step
is repulsion between the F and the R-carbon causing the F to
drag the H along as it departs is also possible, but it is not clear
this would reproduce the observed isotope effect.
A number of issues about the transition state remain. It would
be useful to have ab initio calculations that determined whether
CH3 and CF3 substituents at the R- and â-carbons would produce
E0’s consistent with the experimental results. Also, it would
be helpful to have calculations for the series of CF3CH2X
molecules, listed in Table 1, that address the flow of electron
density for â-carbon substituents and that determine whether a
CF3 substituent actually releases electron density to the carbon
skeleton as the HF departs. It would be interesting to
experimentally measure if the effect of a methyl substituent is
counterbalanced by an electron acceptor such as CF3, F, or Cl
or if the effect of one type of substituent is dominant. It would
also be interesting to measure whether experimental results for
HCl loss are similar to those found for HF elimination when
CH3 and CF3 substituents are present at the R- or â-carbons.
We have begun to further experimentally and theoretically
explore these and other issues regarding the 1,2-dehydrohalo-
genation reaction.
In summary, we have measured unimolecular rate constants
for 1,2-HF elimination from chemically activated CF3CH2CH3
and CF3CH2CF3 and found that replacement of the H of CF3-
CH3 with a methyl substituent lowers the threshold energy
barrier by about 6 kcal/mol while replacement with a CF3
substituent raises the barrier by about 5 kcal/mol. The Cl and
F substituents have the same effect on the E0 as a CF3 group,
suggesting halogen substituents would also attempt to remove
electron density from the â-carbon. The picture of the HF
elimination transition state that emerges for CF3CH3 is one
where electron density flows to the â-carbon from the departing
hydrogen and from its other substituents. A portion of this
charge is transferred from the â-carbon to the R-carbon. The
R-carbon also receives electron density from the electron-rich
fluorine substituents, but the R-carbon releases most of this
electron density to the departing fluorine. The present experi-
mental results are consistent with this movement of electron
density because electron-donating substituents, such as CH3, on
either carbon would aid the flow of negative charge and reduce
the E0, while Cl, F, and CF3, which are electron-withdrawing
substituents, would hinder the flow of electron density and raise
the threshold energy for HF elimination. This view is also
consistent with a previous suggestion30 that the 1,2-elimination
is a stepwise process: first the H is rapidly transferred to a lone
pair on the departing F, the sp3 hybridized carbons begin to
relax to an sp2 geometry, and then the F recoils from the
Acknowledgment. We are pleased to acknowledge financial
support provided by the National Science Foundation (CHE-
9508666) and the Arkansas-NASA EPSCoR Program.
References and Notes
(1) Paisley, S. D.; Holmes, B. E. J. Phys. Chem. 1983, 87, 3042.
(2) Jones, Y.; Holmes, B. E.; Duke, D. W.; Tipton, D. L. J. Phys. Chem.
1990, 94, 4957.
(3) Rakestraw, D. J.; Holmes, B. E. J. Phys. Chem. 1991, 95, 3968.
(4) McDoniel, J. B.; Holmes, B. E. J. Phys. Chem. 1996, 100, 3044.
(5) McDoniel, J. B.; Holmes, B. E. J. Phys. Chem. A 1997, 101, 1334.
(6) Toto, J. L.; Pritchard, G. O.; Kirtman, B. J. Phys. Chem. l994, 98,
8359.
(7) Chang, H. W.; Craig, N. L.; Setser, D. W. J. Phys. Chem. 1972,
76, 954.
(8) Cadman, P.; Day, M.; Trotman-Dickenson, A. F. J. Chem. Soc. A
1970, 2498.
(9) Kerr, J. A.; Timlin, D. M. Trans. Faraday Soc. 1971, 67, 1376.
(10) Phillips, D. C.; Trotman-Dickenson, A. F. J. Chem. Soc. A 1968,
1144.
(11) Follmer, D. W.; Pritchard, G. O. Int. J. Chem. Kinet. 1974, 6, 573.
(12) Holmes, B. E.; Setser, D. W.; Pritchard, G. O. Int. J. Chem. Kinet.
1976, 8, 215.
(13) Kim, K. C.; Setser, D. W.; Holmes, B. E. J. Phys. Chem. 1973,
77, 725.
(14) Holmes, B. E.; Paisley, S. D.; Rakestraw, D. J.; King, E. E. Int. J.
Chem. Kinet. 1986, 18, 639.
(15) Hippler, H.; Troe, J.; Wendelken, H. J. J. Chem. Phys. 1983, 78,
6709.
(16) Joshi, R. M. J. Macromol. Sci.-Chem. 1974, A8, 861.
(17) Kolesov, V. P.; Ivanov, L. S.; Skurstov, S. M. Russ. J. Phys. Chem.
1975, 45, 303.
(18) Pamidimukkala, K. M.; Rogers, D.; Skinner, G. B. J. Phys. Chem.
Ref. Data 1982, 77, 5877.
(19) Chen, Y.; Rauk, A.; Tschuikow-Roux, E. J. Phys. Chem. 1990,
93, 1187.
(20) Rodgers, A. S. In Fluorine-Containing Free Radicals: Kinetics
and Dynamics of Reactions; Root, J. W., Ed.; ACS Symposium Series No.
66; American Chemical Society: Washington, DC, 1978; pp 269-313.
(21) Rayez, M.; Rayez, J.; Sawersyn, J. J. Phys. Chem. 1994, 98, 11342.
(22) Martell, J. M.; Boyd, R. J.; Shi, Z. J. Phys. Chem. 1993, 97, 7208.
(23) Durig, J. R.; Guirgis, G. A.; Li, Y. S. J. Chem. Phys. 1981, 74,
5946.
(24) Nanair, H.; Guirgis, G. A.; Durig, J. R. Spectrochim. Acta 1993,
49A, 2039.
(25) Pace, E. L.; Plaush, A. C.; Samuelson, H. V. Spectrochim. Acta
1966, 22, 993.
(26) Durig, J. R.; Godbey, S. E.; Sullivan, J. F. J. Chem. Phys. 1984,
80, 5983.
(27) Stolivik, R.; Bakker, P. J. Mol. Struct. 1991, 244, 27.
(28) Burger, H.; Pawelke, G. Spectrochim. Acta 1979, 35A, 517, 525.
(29) Cadman, P.; Kirk, A. W.; Trotman-Dickenson, A. F. J. Chem. Soc.,
Faraday Trans. 1 1975, 72, 996.
(30) Holmes, B. E.; Setser, D. W. J. Phys. Chem. 1978, 82, 2461.