Substituent effects and threshold energies for the unimolecular elimination of HCl (DCl) and HF (DF) from chemically activated CFCl2CH3 and CFCl2CD3

Article abstract of DOI:10.1021/jp9523128

Combination of CFCl₂ and methyl-d₀ and -d₃ radicals form CFCl₂CH₃-d₀ and -d₃ with 100 and 101 kcal/mol of internal energy, respectively. An upper limit for the rate constant ratio of disproportionation to combination, k_d/k_c, for Cl transfer is 0.07 ± 0.03 for collision of two CFCl₂ radicals and 0.015 ± 0.005 for CH₃ and CFCl₂ radicals. The chemically activated CFCl₂CH₃ undergoes 1,2-dehydrochlorination and 1,2-dehydrofluorination with rate constants of 3.9 × 10⁹ and 4.9 × 10⁷ s^-1, respectively. For CFCl₂CD₃ the rate constants are 8.7 × 10⁸ s^-1 for loss of DCl and 1.1 × 10⁷ s^-1 for DF. The kinetic isotope effect is 4.4 ± 0.9 for HCl/DCl and appears to be identical for HF/DF. Threshold energies are 54 kcal/mol for loss of HCl and 68 kcal/mol for HF; the E₀'s for the deuterated channels are 1.4 kcal/mol higher. Comparison of these threshold energies with other haloethanes suggests that for HF and HCl elimination the transition states are developing charges of different signs on the carbon containing the departing halogen and that chlorine and fluorine substituents exert similar inductive effects.

Full text of DOI:10.1021/jp9523128

3044
J. Phys. Chem. 1996, 100, 3044-3050
Substituent Effects and Threshold Energies for the Unimolecular Elimination of HCl (DCl)
and HF (DF) from Chemically Activated CFCl2CH3 and CFCl2CD3
J. Bridget McDoniel and Bert E. Holmes*
Department of Chemistry, Lyon College, BatesVille, Arkansas 72501
ReceiVed: August 8, 1995; In Final Form: October 18, 1995X
Combination of CFCl2 and methyl-d0 and -d3 radicals form CFCl2CH3-d0 and -d3 with 100 and 101 kcal/mol
of internal energy, respectively. An upper limit for the rate constant ratio of disproportionation to combination,
kd/kc, for Cl transfer is 0.07 ( 0.03 for collision of two CFCl2 radicals and 0.015 ( 0.005 for CH3 and CFCl2
radicals. The chemically activated CFCl2CH3 undergoes 1,2-dehydrochlorination and 1,2-dehydrofluorination
with rate constants of 3.9 × 109 and 4.9 × 107 s-1, respectively. For CFCl2CD3 the rate constants are 8.7 ×
108 s-1 for loss of DCl and 1.1 × 107 s-1 for DF. The kinetic isotope effect is 4.4 ( 0.9 for HCl/DCl and
appears to be identical for HF/DF. Threshold energies are 54 kcal/mol for loss of HCl and 68 kcal/mol for
HF; the E0’s for the deuterated channels are 1.4 kcal/mol higher. Comparison of these threshold energies
with other haloethanes suggests that for HF and HCl elimination the transition states are developing charges
of different signs on the carbon containing the departing halogen and that chlorine and fluorine substituents
exert similar inductive effects.
Introduction
chlorine is delocalized to the developing positive charge on the
R-carbon. A fluorine substituent’s high electronegativity is
thought3,4,6 to destabilize the transition state by inductive
withdrawal of electron density from the R-carbon, increasing
the E0 with sequential fluorine substitution.
Two of the classic mechanisms in organic chemistry1,2 are
E1 and E2 eliminations of a hydrogen halide from a haloalkane.
The transition state for first order eliminations, E1, involves
heterolytic fission of the carbon-halogen bond and formation
of a planar carbocation. The second order elimination mech-
anism, E2, is assisted by a base and involves a planar transition
state with an anti-configuration; i.e., the halogen and hydrogen
are removed from opposite sides of the C-C bond. In the gas
phase, the 1,2-dehydrohalogenation reaction has been assumed3,4
to involve a planar and polar four-centered transition state, and
it has been shown5 that syn-elimination governs the removal of
the hydrogen halide. For the past 3 decades3-8 experiments
have been designed to better understand the polarity of the 1,2-
dehydrohalogenation transition state, primarily by measuring
the effect of different substituents on the rate constant and on
the threshold energy barrier, E0. Most of the experimental
work5-7 has involved chloro- and fluoroalkanes because com-
plicated reaction chemistry can result from carbon-halogen
bond rupture with bromo- and iodoalkanes.
For chloroethanes the effect of chlorine substituents at the
R-carbon (the carbon containing the halogen) for the uni-
molecular elimination of HCl is a decrease of 2-3 kcal/mol in
E0(HCl) for successive replacement of hydrogen by chlorine.6
For fluoroethanes the effect of fluorine substituents at the
R-carbon is an increase of E0(HF) by 3 kcal/mol when one
R-hydrogen of fluoroethane is replaced by a fluorine.6 Substitu-
tion of the second R-hydrogen with fluorine results in an
additional 7 kcal/mol increase in E0(HF).6
Unimolecular rate constants7 measured for elimination of HF
and HCl from chemically activated CF2ClCH3 produced E0’s
inconsistent with the idea of common transition states and
different electronic effects for Cl and F substituents. For CF2-
ClCH3, the chlorine and fluorine substituents had the same effect
on E0 when HF was eliminated. Both substituents increased
E0(HF) so that E0’s for CF3CH3 (68 kcal/mol) and CF2ClCH3
(69.5 kcal/mol) are nearly identical. We should note that the
uncertainity in experimental threshold energies is normally 2
kcal/mol. For HCl elimination, fluorine substituents behave like
hydrogens; the threshold energies are identical at 55 kcal/mol
for both CH2ClCH3 and CF2ClCH3. To account for the behavior
of the chlorine substituents on E0(HF) and the fluorine substit-
uents on E0(HCl), we suggested7 that the transition states have
different polarity and that Cl and F substituents exhibit similar
electronic effects; i.e., both either accept or donate electron
density to the carbon skeleton.
Recently, Toto et al.9 reported ab initio quantum chemical
calculations designed to determine the geometry, bond orders,
atomic charges, and activation energies for the transition states
for 1,2-HX elimination (X ) F, Cl, Br, and I) from haloethanes.
Theoretical results will be an important aid in understanding
this class of reactions because of the ambiguous nature of
halogen substituent effects. In general, these calculations
predicted a very asymmetric transition state structure with bond
orders that are much stronger for C-C and C-H and much
weaker for C-X and H-X than previous models.6,10-12 The
calculated C-H and C-X bond orders and their relation to the
experimental kinetic isotope effect will be considered in the
Results and Discussion. For fluoroethane the calculated activa-
tion energy was within the experimental uncertainty, but the
Ea’s were generally 3-7 kcal/mol too high for all the halo-
ethanes. Identical Ea’s were predicted for HF and HCl and for
HBr and HI, contrary to observations.9
It has been suggested3,4,6 that the threshold energies for the
removal of HF from fluoroethanes exhibit trends opposite that
for loss of HCl from chloroethanes because the Cl and F
substituents exert different electronic effects on common
transition states. Specifically, the charge on the R-carbon
becomes more positive, and a partial negative charge develops
on the carbon containing the hydrogen, the â-carbon. The
decrease in E0(HCl)’s with successive Cl substitution is
explained3,4 by resonance stabilization as a lone pair of the
X Abstract published in AdVance ACS Abstracts, January 15, 1996.
Transition states for loss of HF and HCl were also computed
0022-3654/96/20100-3044$12.00/0 © 1996 American Chemical Society

Cl and F Eliminations from CFCl2CH3 and CFCl2CD3
J. Phys. Chem., Vol. 100, No. 8, 1996 3045
for all possible R-chloro-/fluoro-substituted ethanes to explore
the substituent effects and the transition states’ ionic character.
Theoretical activation energies agreed with the experimentally
observed increase in Ea for HF elimination with increasing
R-substitution but a decrease in Ea for HCl elimination. For
CF2ClCH3 a calculated difference in E0(HF) and E0(HCl) of
10.4 kcal/mol is close to the experimental7 difference of 14.5
kcal/mol. Two important findings of the calculations9 are that
(1) for elimination of either HCl or HF, the transition state
geometry, bond orders, and atomic charges are nearly indepen-
dent of substituents but (2) the bond orders and the developing
atomic charges for the HF and HCl elimination transition states
are considerably different. The former supports our use of
identical Arrhenius A-factors to develop common transition state
models for the RRKM calculations, regardless of the substituents
on the R-carbon. The latter point supports our contention7 that
different transition states exist for loss of HF and HCl. Results
and Discussion will consider the calculated atomic charges and
the substituent effects in greater depth.
This work will further test the effect of Cl and F substituents
on the E0’s for competitive 1,2-HF and 1,2-HCl eliminations
by preparing chemically activated CFCl2CH3 and measuring the
unimolecular rate constants. Threshold energies for HF and
HCl elimination will be obtained by matching the experimental
rate constants to rate constants calculated using the RRKM
theory. Comparison of these threshold energies to other halo-
substituted ethanes and to the predictions of Toto et al.9 should
provide additional information about the nature of halogen
substituents and the polar character of the transition states. The
deuterated analogue, CFCl2CD3, was also studied to obtain the
kinetic isotope effect.
uents on haloacetones14 and with wavelengths shorter than 313
nm.15 Chlorine radicals from reaction 1b will react with alkenes
produced by the loss of HCl and HF from the CFCl2CH3*
(reactions 5a and 5b) which would give unimolecular rate
constants that are too small. Atomic chlorine has been observed
to react with alkenes in similar chemical activation systems,16
and propene was an efficient competitor for its removal.
Therefore, a narrow band pass filter centered at 313 nm was
used to reduce the undesired C-Cl fission, and a scavenger of
propene was also used to protect the alkene products.
The CFCl2CH3* can undergo loss of HCl and HF or become
stabilized through collision (reactions 5a-c). The ratio of the
yield of the decomposition (D) product from reaction 5a and
5b and the stabilization (S) product yield, CFCl2CH3, plotted
kHCl
CFCl2CH3*
8 CFCldCH2 + HCl
8 CCl2dCH2 + HF
(5a)
(5b)
(5c)
kHF
[M]8 CFCl2CH3
kM
versus inverse pressure should produce a linear relationship at
higher pressures with a zero intercept. The experimental rate
constant, ka, equals kM[M](D/S), where kM[M] is the collision
rate. It is anticipated that kHCl . kHF so that when the yield of
CCl2dCH2 is measurable, the yield of CFCl2CH3 would be too
small to be observed; thus, only kHCl can be determined from a
plot of D/S versus inverse pressure. To obtain kHF, the branching
ratio of kHCl/kHF will be obtained from a plot of [CFCldCH2]/
[CCl2dCH2] versus inverse pressure using data collected at very
low pressures, where the yields of CFCldCH2 and CCl2dCH2
are the greatest. Similar plots constructed for the deuterated
molecule will provide kDCl, kDF, and the deuterium kinetic
isotope effect.
Chemically activated CFCl2CH3 (CFCl2CD3) is formed with
100 (101) kcal/mol of energy by combination of CFCl2 and CH3
(CD3) radicals. These radicals are generated by the photolysis
of 1,2-difluoro-1,1,2,2-tetrachloroacetone and acetone-d0 (-d6).
The following scheme summarizes the reactions expected for
this system, with the deuterated channels omitted. The asterisk
denotes a chemically energized species.
Experimental Section
hν8 2CFCl2 + CO
(1a)
Quartz vessels with a volume range of 17.13-992.3 cm3
containing 2.23 µmol of 1,1-difluoro-1,1,2,2-tetrachloroacetone,
30.1 µmol of acetone or deuterated acetone, 0.94 µmol of
propene, and a bath gas of 1.04 mmol of sulfur hexafluoride
were photolyzed using a high-pressure Oriel 6137 mercury lamp
and an Oriel 59154 narrow-band-pass filter centered at 313 nm.
For 30 min photolyses at room temperature, the typical reactant
conversion rates were less than 5%. Lower pressure mixtures
devoid of SF6 and propene were required to measure the HCl/
HF branching ratio. All samples were prepared on a grease-
free vacuum line using an MKS 270 electronic manometer. A
(CFCl2)2CO
9
9
hν8 Cl + CFCl2COCFCl (1b)
CH3 + CH3 f CH3CH3
(2)
(3a)
(3b)
(4a)
(4b)
kd
CFCl2 + CH3
98 :CFCl + CH3Cl
kc
98 CFCl2CH3*
1
Shimadzu 14A FID GC with a /8 in. stainless steel column
kd
containing Hayesep A coated with nickel that had been reduced
by H2 from a 5.0% NiCl2 solution was used for analyses. Using
an initial temperature of 100 °C for 5 min followed by
temperature programming at 1 °C/min to a final temperature of
167 °C, the elution times (in minutes) were generally as
follows: SF6, 0.1 min; C2H6, 2; propene, 7; CH3Cl, 12;
CFCldCH2, 13; CFCl2H, 37; CFCl3, 40; CCl2dCH2, 43; CFCl2-
CH3, 49; acetone, 67; CFCl2CFCl2, 128; difluorotetrachloro-
acetone, 325. The data were collected and integrated with a
Shimadzu Chromatopac CR5A integrator.
Products were identified by comparison of GC retention times
with commercial samples and verified by a Hewlett-Packard
5890/5791 GC/MS equipped with a 50 m PONA capillary
column. The CCl2dCH2 could not be identified by mass
spectrometry due to its very low yield, but the CFCldCH2,
CFCl2 + CFCl2
98 :CFCl + CFCl3
kc
98 CFCl2CFCl2*
The fate of the :CFCl carbene may be dimerization to
CFCldCFCl, similar to the reactions of the :CF2 carbene,7 or it
may be removed by reactions with the CFCl2 or CH3 radicals.
Reactions 3a and 4a are disproportionations, transferring a Cl
from CFCl2 to CH3 and to CFCl2, respectively. The combina-
tion reactions (3b and 4b) form the chemically activated species.
These rate constant ratios, kd/kc, will be reported. Photoactivated
difluorotetrachloroacetone reacts either by breaking the C-C
(1a) or the C-Cl bond (1b).13 Rupture of the C-Cl bond
increases in importance with increasing number of Cl substit-

3046 J. Phys. Chem., Vol. 100, No. 8, 1996
McDoniel and Holmes
ratio, a C-Cl stretch was the reaction coordinate, and two
C-C-Cl bends were reduced to 35 cm-1; this gave 1.0 × 1016
s-1 as the Arrhenius A-factor at 1000 K.22 The threshold energy
for C-Cl bond rupture for CFCl2CFCl2 is not known, so we
tested four E0’s (70, 75, 80, and 85 kcal/mol) that encompass
the range in the reported C-Cl bond dissociation energies for
CF2Cl2 (71.3-85 kcal/mol).23 Assuming kM ) 1.0 × 107 Torr-1
s-1 and an average energy of 100 kcal/mol for CFCl2CFCl2*,
the pressure when half of the CFCl2CFCl2* reacts is calculated
to be 60 Torr (E0 ) 70 kcal/mol), 7 Torr (E0 ) 75 kcal/mol),
0.5 Torr (E0 ) 80 kcal/mol), and 0.03 Torr (E0 ) 85 kcal/
mol). Figure 1 shows that kd/kc ) 0.07 at higher pressures and
doubles between 1 and 10 Torr. The RRKM results are
consistent with unimolecular decomposition of CFCl2CFCl2*,
causing the increase in kd/kc at lower pressures if the C-Cl bond
dissociation energy for CFCl2CFCl2 is between 70 and 80 kcal/
mol.
Figure 1. Plot of [CFCl3]/[CFCl2CFCl2] versus the log of reciprocal
pressure. The product ratio is kd/kc. The circles are for photolysis of
1,2-dichloro-1,1,2,2-tetrafluoroacetone with acetone-d0, and the squares
are with acetone-d6. Decomposition of chemically activated CFCl2-
CFCl2 may cause the upward curvature at lower pressures; see Results
and Discussion.
The other disproportionation/combination ratio that can be
determined is for the collision of CFCl2 and CH3 radicals; kd/kc
) k3a/k3b ) [CH3Cl]/[CFCl2CH3*] and [CFCl2CH3*] ) [CFCl2-
CH3] + [CFCldCH2] + [CCl2dCH2]. On the basis of 47 trials,
the kd/kc is independent of pressure and is constant at 0.015 (
0.005.
CCl2dCH2, and CFCl2CH3 did exhibit the expected pressure
dependence of a chemically activated, unimolecular decomposi-
tion system.
Relative calibration factors, which are needed for D/S, HCl/
HF branching ratios, and kd/kc, were measured as [CFCldCH2]/
[CFCl2CH3] ) 0.747, [CFCldCH2]/[CCl2dCH2] ) 1.15, [CFCl3]/
[CFCl2CFCl2] ) 7.57, and [CH3Cl]/{[CFCldCH2] + [CFCl2-
CH3]} ) 2.1. These relative responses were determined using
six to eight trials of three mixtures prepared to replicate typical
reaction yields and GC elution characteristics. Calibration
factors for the deuterated system were assumed to be the same
as the undeuterated compounds. Our response factors are
consistent with observations by Tschuikow-Roux et al. in that
increasingly halogenated methanes and ethanes exhibit reduced
response.17 Calibration factors vary18,19 with alteration of
hydrogen and air flame conditions; therefore, the response
factors reported here may not be applicable for other work. We
have also found20 that response factors for halocarbons, unlike
those for hydrocarbons, vary dramatically with small variations
in the fuel/air ratios.
Substituents should influence the kd/kc’s, so determining the
effect of F or Cl substituents for these bimolecular processes
might assist in interpreting these effects for the unimolecular
1,2-HX elimination reactions. Molecular orbital theory (MYFF
model),24 applied to the transfer of an H from C2H5 radicals to
methyl radicals, predicts that the transition state for dispropor-
tionation has ionic character, while it is neutral for combination.
Thus, the disproportionation reaction rate constant, but not the
combination rate, would be influenced by the electronic
character of substituents. Further, this model suggests that
electron withdrawing groups on the acceptor radical or electron
donating groups on the donor radical both cause an increase in
kd. In comparing a series of radical reactions, both kc and kd
cause a variation in kd/kc. Analysis of the substituent effects
of different halogens on kc compared to those on kd/kc allows
determination of the effects on kd. At room temperature, the
most reliable kc measurements show a decline in kc’s with
increasing halogenation with Cl’s have a greater effect than
fluorine: 3.6 × 1010 M-1 s-1 for CH3 radicals,25 1.4 × 1010
M-1 s-1 for CF3 radicals,26 and 0.2 × 1010 M-1 s-1 for CCl3
radicals.27
Results and Discussion
1. Disproportionation/Combination Rate Constant Ratios.
Two disproportionation/combination rate constant ratios (kd/kc)
can be determined from the product yields. The kd/kc’s will be
upper limits because the disproportionation product could also
be formed by radicals abstracting Cl from difluorotetrachloro-
acetone. For collision of two CFCl2 radicals, the kd/kc ) k4a/
k4b ) [CFCl3]/[CFCl2CFCl2*] and is constant at 0.07 ( 0.03 in
the pressure range from 1730 to about 10 Torr but clearly
increases at lower pressures, Figure 1. Because the combination
product is itself chemically activated, CFCl2CFCl2* may react
at lower pressures; presumably C-Cl bond rupture is the
pathway with the lowest energy barrier, reactions 6a and 6b.
We will now consider substituent effects on kd/kc and use
the MYFF model and the trends in kc to determine the effect of
Cl and F substituents on kd. Three sets of disproportionation/
combination reactions involving transfer of a chlorine between
halomethyl radicals will be discussed. The first compares CH3
+ CClF2 (kd/kc ) 0.24 per transferable Cl)7 with CH3 + CCl2F
(kd/kc ) 0.0075 per transferable Cl). It should be noted that
kd/kc ) 0.24 ( 0.05 for CH3 + CClF2 was considered tentative7
due to experimental complications. The radical listed first,
methyl, is the acceptor and the second, differing by a chlorine
and a fluorine, is the donor radical. The additional Cl when
CCl2F is the donor radical should reduce kc, causing kd/kc to
increase; however, it declines by a factor of 30. According to
the MYFF model, a donor radical would reduce kd if electron
withdrawing groups were present; for CCl2F to decrease kd/kc
compared to CClF2, the Cl would have to be more strongly
electron withdrawing than F. Secondly, substituent effects on
the acceptor radical can be probed by comparing CH3 + CCl2F
(kd/kc ) 0.0075 per transferable Cl) with CCl2F + CCl2F (kd/kc
) 0.0175 per transferable Cl), where the donor radical is
constant and the acceptor radical changes from CH3 to CCl2F.
kCl
CFCl2CFCl2*
9
8 CFCl2CFCl + Cl
(6a)
(6b)
[M]8 CFCl2CFCl2
kM
Reaction 6a would cause [CFCl3]/[CFCl2CFCl2] to increase as
pressure is reduced, and kd/kc doubles when half of the CFCl2-
CFCl2* reacts, i.e., when kCl ) kM[M]. RRKM theory was used
to calculate kCl; kM was obtained from collision theory, and these
gave [M], the pressure, when kd/kc doubled. Vibrational
frequencies for CFCl2CFCl2 were estimated from CFCl2CH3.21
The transition state model used 5.0 as the moments of inertia

Cl and F Eliminations from CFCl2CH3 and CFCl2CD3
J. Phys. Chem., Vol. 100, No. 8, 1996 3047
Figure 3. Experimental yield of [CFCldCH2]/[CCl2dCH2] (circles)
and [CFCldCD2]/[CCl2dCD2] (squares) versus reciprocal pressure. The
difficulty of measuring the very small CCl2dCH2 or CCl2dCD2 yield
is responsible for the scatter at higher pressures.
Figure 2. D/S versus reciprocal pressure plot for the four-centered
elimination of HCl from CFCl2CH3 (circles) and elimination of DCl
from CFCl2CD3 (squares). D is [CFCldCH2-d0 and -d2], and S is
[CFCl2CH3-d0 and -d3], respectively.
independent collision diameters of 5.9 Å (SF6), 5.0 Å (acetone),
6.0 Å (difluorotetrachloroacetone), 4.0 Å (propene), and 5.7 Å
(CFCl2CH3), the kM was 1.07 × 107 Torr-1 s-1 and gave kHCl
) 3.9 × 109 s-1 and kDCl ) 8.7 × 108 s-1. Collision diameters,
assumed invariant with deuterium substitution, are known33 for
SF6, acetone, and propene and were estimated for difluoro-
tetrachloroacetone and CFCl2CH3 by comparison to analogous
Again, the increase in the number of halogen substituents (in
this case on the acceptor radical) should reduce kc, causing an
increase for kd/kc. The kd/kc is a factor of 2 larger with CCl2F
versus CH3 as the acceptor radical, consistent with electron
withdrawing groups assisting the disproportionation reaction.
However, a decline in kc, an increase in kd, or a decline in kc
that is larger than the decline in kd could be responsible for the
small increase in kd/kc. A final comparison shows trends in
kd/kc counter to the previous set. For this case the acceptor
radical changes from CH3 to CF2Cl with the same donor radical,
CF2Cl. Comparing CH3 + CClF2 (kd/kc ) 0.24 per transferable
Cl)7 with CClF2 + CClF2 (kd/kc ) 0.085)7, the additional
halogens should decrease kc and increase kd/kc, but the rate
constant ratio declines by almost a factor of 3. In summary, it
appears that no clear trend emerges for the effect of Cl and F
substituents for disproportionation reactions, perhaps because
the MYFF model should not be applied to the disproportionation
reactions involving Cl transfer or to reactions that yield a carbene
(reactions 3a and 4a) rather than an alkene. It is also possible
that thermodynamic rather than kinetic factors determine the
product distribution or that the MYFF model does not accurately
represent disproportionation reactions.
2. Experimental Unimolecular Rate Constants. Photolysis
of fluorinated acetones have been a useful precursor for
fluoromethyl radicals,28-30 but chloroacetones are impractical
because C-Cl bond rupture competes with the desired carbon-
carbon bond fission.14,31 Difluorotetrachloroacetone has rarely
been used to generate CFCl2 radicals32 for kinetic studies;
although, we found it to be an effective source for CFCl2 radicals
despite production of Cl atoms. In our system, the Cl radicals
attack the alkenes formed in reactions 5a and 5b, apparently
decreasing the unimolecular rate constants. Propene was used
to scavenge atomic chlorine;16 as it was added at a constant
total pressure of 1000 Torr, the D/S increased from 0.19 with
no propene, to become constant at 0.44 when the propene to
total acetones ratio was 1:9. Maximum inhibition of the
mechanism of radical attack was assumed. Some alkene may
still be removed by reaction with atomic Cl; therefore, the rate
constants are considered to be lower limits, and the E0’s
represent upper limits. In the kinetic runs, a smaller amount
of propene was used because it was not completely resolved
from CFCldCH2 in the GC analysis at higher propene concen-
trations.
molecules.33 The rate constants for loss of HF, 4.9 × 107 s-1
,
and DF, 1.1 × 107 s-1, were determined from the HCl/HF
branching ratio of 80 in the low-pressure region, Figure 3.
Considerable scatter exists in Figure 3 due to the difficulty
in measuring the very small yield of CCl2dCH2. The intercepts
for the D/S plots of Figure 2 are negative for both the CFCl2-
CH3 and the CFCl2CD3 systems. Causes for this deviation could
include inaccurate measurements of the pressures, incorrect
calibration response factors for the flame ionization detector,
or an error in the product yields for CFCldCH2 or CFCl2CH3.
An MKS electronic manometer with a reported uncertainty of
0.08% was used in pressure measurements. The pressures
would have to be adjusted by over 100 Torr to give a zero
intercept. This corresponds to greater than 10% variation from
most pressures and, thus, cannot be the source of this error.
Alternate calibration factors were also tested to determine
the value that would produce a zero intercept. Modifying the
calibration factors changed the slope but only slightly altered
the intercept; for example, reducing the calibration factors by
10% increased the intercept from -0.016 to -0.014, while the
slope decreased by about 10%. Increasing the calibration factor
by 10% decreased the intercept from -0.016 to -0.017 and
the slope increased by about 11%. The calibration factor would
have to be in error by at least a factor of 2 to produce an
intercept close to zero. The actual [CFCldCH2]/[CFCl2CH3]
calibration factor was 0.630 with 1 standard deviation being
0.051. Because reasonable adjustments in the calibration factor
did not give significant improvement, an inaccurate response
factor is not responsible for the nonzero intercepts.
A final explanation would be the loss of some of the
decomposition product, excess CFCl2CH3, or incorrect measure-
ments of their yields. The integration for the CFCldCH2 peak
area was the least reliable because the GC analysis did not
completely resolve it from the very large propene peak. This
was particularly problematic at the highest pressures, where the
yield of CFCldCH2 approached zero. The CFCldCH2 yield
could also be low due to its removal by chlorine atoms. A
typical CFCldCH2 peak area from the digital integrator was
50 000 counts for higher pressures and over 100 000 counts for
lower pressures. For chemically activated CFCl2CH3*, addition
The unimolecular rate constants for loss of HCl and DCl were
determined from standard D/S plots; Figure 2 has a slope of
363 Torr for HCl and 81 Torr for DCl. Using temperature

3048 J. Phys. Chem., Vol. 100, No. 8, 1996
McDoniel and Holmes
TABLE 1: Summary of Experimental Rate Constants and RRKM Models for 1,1-Dichloro-1-fluoroethane
molecules
activated complexes, elimination of
CFCl2CH3
CFCl2CD3
HCl
DCl
HF
DF
vibr freq, cm-1 (degeneracies)
2998 (3)
1423 (3)
1070 (4)
750 (1)
591 (1)
403 (3)
266 (3)a
2246 (3)
1021 (5)
818 (3)
591 (1)
403 (3)
276 (2)
150 (1)
2991 (2)
1478 (2)
1123 (3)
823 (4)
610 (1)
403 (3)
317 (2)
0.955
2227 (2)
1223 (3)
925 (2)
774 (2)
457 (3)
389 (2)
345 (3)
0.955
2991 (2)
1452 (3)
1048 (3)
794 (2)
571 (3)
389 (2)
277 (2)
0.970
2227 (2)
1481 (2)
904 (4)
543 (4)
389 (2)
304 (2)
264 (1)
0.970
moments of inertia, Iq/I
reaction path degeneracya
preexponential factor,b s-1
E0, kcal/mol
4
4
2
2
7.0 × 1012
8.5 × 1012
9 × 1012
11 × 1012
54
55.4
68
69.4
E , kcal/mol
100
101
100
101
ka(exptl), s-1
3.9 × 109
8.7 × 108
4.9 × 107
1.1 × 107
ka(calcd), s-1
4.0 × 109
1.9 × 109
4.5 × 107
1.9 × 107
[kaH/kaD](exptl)
4.4 ( 0.9
2.1
4.4 ( 0.9
2.4
[kaH/kaD](calcd)
a Hindered rotor treated as a torsion. b Partition function form for unit reaction path degeneracy at 800 K.
of 11 250 counts to the CFCldCH2 yield for each analysis gave
a zero intercept. For CD3CFCl2*, the addition of 900 counts
to the decomposition product’s digital integrator counts would
result in an intercept through the origin. These adjustments in
the CFCldCH2 yield would cause an increase in both the rate
for kHCl and kDCl of 32.8 and 31.3%, respectively, but would
not change the kinetic isotope effect. Increasing the rates 30%
would require E0’s 1 kcal/mol lower than the recommended
values.
position of CFCl2CH3. To facilitate comparisons, the thermal
preexponential factors for loss of HF and HCl and the RRKM
models were developed using the same methodology employed
in earlier work.7 Vibrational frequencies for CFCl2CH3 are
known,21 and those for CFCl2CD3 were estimated from CFCl2-
CH3 by comparing frequencies of CF3CH3 with CF3CD3.36,37
For the four-member transition state, the frequencies were
assigned from previous models6,7,30 with the ring puckering
adjusted to produce the desired preexponential factor; see Table
1. The reaction path degeneracies are based upon a torsional
model.
The average energy, E , of CFCl2CH3* is normally deter-
mined from the enthalpy of reaction 3b. We anticipate that
E ≈ 100 kcal/mol based on the E for chemically activated
CH3CF3 and CFCl2CH3.7,38,39 A value near 100 kcal/mol is in
accord with a recent ab initio calculation40 of the bond
dissociation energies for a series of fluoroethanes in which the
bond energies for CH3CX3 (X ) halogen) are predicted to be
6 kcal/mol higher than for ethane because of the increased ionic
character in the C-C bond. The enthalpy of formation at 298
K for CFCl2CH3 is not available; it was estimated as -83.1
kcal/mol from group additivity schemes,41 and adjustment to 0
K gave -82.6 kcal/mol. For the methyl radical ∆Hf°(0) ) 35.9
kcal/mol,42 and for the CFCl2 radical a recent measurement43
at 298 K of -22 ( 2 kcal/mol was corrected to -23.9 kcal/
mol at 0 K. ∆H°rxn for reaction 3b at 0 K is -94.6 kcal/mol,
and addition of 3.8 kcal/mol thermal energy for CFCl2CH3 and
an Ea ) 1 kcal/mol for radical combination give an estimated
E ) 99.4 kcal/mol. Considering that the uncertainty in the
experimental ∆Hf°’s is at least (2 kcal/mol, we have selected
100 kcal/mol as the E for chemically activated CFCl2CH3.
Using E ) 100 kcal/mol, the calculated rate constants match
the experimental values with E0(HCl) ) 54 kcal/mol and
E0(HF) ) 68 kcal/mol. Table 1 is a summary of the RRKM
models, the rate constants, and the isotope effects.
The experimental rate constants from Figures 2 and 3 give
the kinetic isotope effect, kH/kD ) 4.4 ( 0.9 for HCl/DCl; HF/
DF appears to be identical. Other chemically activated tri-
substituted haloethanes with one primary and two secondary
deuterium isotope effects gave kH/kD slightly above 3,7 making
the present system one of the largest kH/kD’s reported.
Attempts to replicate the large deuterium kinetic isotope effect
required RRKM models with nearly complete fission of the
C-H bond; a bond order near 0.1 has been recommended6,34
for chemically activated haloethanes. In addition, the 37Cl/
35Cl thermal isotope effect35 is very small, suggesting only
moderate lengthening of the C-Cl bond and little Cl motion in
the reaction coordinate. The MO calculation of Toto et al.9
furnishes bond orders between 0.36 and 0.7 for C-H and
between 0.02 and 0.12 for C-X, and it is not clear that these
would match the large experimental deuterium isotope effect
and the small isotope effect for chlorine. Increasing numbers
of R-halogens increase the kH/kD’s for chemically activated
haloethanes,7 suggesting that the C-H bond weakens. By
contrast, the MO calculations9 show a monotonic increase in
the C-H bond order for HCl elimination: 0.512 (chloroethane)
to 0.700 (1,1,1-trisubstituted ethanes) and the large bond order
indicates little motion of H as the transition state forms. The
C-H bond orders for the HF elimination transition states are
calculated to be approximately 0.39, independent of R-halogena-
tion. This suggests a much greater deuterium isotope effect
for loss of HF versus HCl, but this has not been experimentally
confirmed.
4. Threshold Energies, Substituent Effects, and Transi-
tion State Atomic Charges. Toto et al.9 calculated the
activation energies for CFCl2CH3, and because the E0’s are
normally 2 kcal/mol lower, we have reduced the computed Ea’s
by that amount. For CFCl2CH3, this gives E0(HCl) ) 56.4 kcal/
mol and E0(HF) ) 63.9 kcal/mol, which are 2.4 higher and 4.1
kcal/mol lower, respectively, than our experimental numbers.
This agreement would be considered satisfactory because the
experimental uncertainty is (2 kcal/mol. However, a more
compelling comparison is the difference in the E0’s, because,
for both the experiment and the calculations, relative results
3. RRKM Calculated Unimolecular Rate Constants and
Threshold Energies. Primary goals of this work are to
determine threshold energies for elimination of HCl and HF
from CFCl2CH3 and to compare these with E0’s for other
haloethanes in order to understand the effect of F and Cl
substituents. The E0’s were ascertained by constructing RRKM
models and varying threshold energies until the calculated rate
constants were in agreement with the experimental ka’s. Ar-
rhenius parameters have not been reported for thermal decom-

Cl and F Eliminations from CFCl2CH3 and CFCl2CD3
J. Phys. Chem., Vol. 100, No. 8, 1996 3049
are more reliable than absolute values. The calculated difference
in the two threshold energies is 7.5 kcal/mol, which is about
half the experimental result of 14 kcal/mol and is clearly outside
the experimental uncertainty unless there exists a very large
undetected experimental error. For HF elimination, replacing
H’s with F’s results in an increase in the threshold energies
from CH2FCH3 (58 kcal/mol) to CHF2CH3 (61 kcal/mol) to CF3-
CH3 (68 kcal/mol). Replacing H’s with Cl substituents has the
same effect upon the E0(HF) as replacement of H’s with F’s;
E0’s for CF3CH3, CF2ClCH3, an CFCl2CH3 are, within the
experimental error, identical at 68, 69.5, and 68 kcal/mol,
respectively. Therefore, for HF elimination, Cl and F substit-
uents are equivalent; they increase E0.
Figure 4. Change in the calculated atomic charges for the HCl and
HF elimination transition states for CFCl2CH3 from Table 15 in ref 9.
The absolute charges on each atom of the reactant molecule are given
in the text.
For HCl elimination from both mono- and dichlorinated
haloethanes, replacing an H with an F does not appreciably alter
the threshold energies. The E0(HCl)’s for CH2ClCH3 and CF2-
ClCH3 are identical at 55 kcal/mol and are 52 and 54 kcal/mol
for CHCl2CH3 and CFCl2CH3, respectively. The small decline
in E0 with additional chlorines could arise from steric consid-
erations. Steric crowding present in the molecule would be
reduced with elongation of the C-Cl bond so that successive
chlorine substitution would raise the energy for the reactant more
than for the transition state, thereby reducing E0. When HF is
eliminated, the steric repulsions involving the Cl substituents
in the molecule and the transition state structures are the same
so there is no effect on the threshold energy.
have different charge characteristics. According to the calcula-
tions,9 the R-carbon is becoming more positively charged when
HCl is lost (the transition states’ R-carbons are 0.24 ( 0.04
more positiVe than in the reactant) but less positively charged
when HF is lost (the transition states’ R-carbons are 0.11 (
0.05 more negatiVe than in the reactant). Hiberty12 calculated
atomic charges for chloroethane very close to those of Toto et
al.9 To be consistent with our trends in the E0’s when H is
replaced by Cl and/or F, both halogen substituents would need
to be very slightly electron donating (almost inductively neutral)
when HCl is lost, but both Cl and F must be more strongly
electron donating when HF is eliminated. Figure 4 shows the
change in the atomic charges calculated for each atom of CFCl2-
CH3 when HCl and HF are removed.
It appears that Cl or F substituents raise the barrier for loss
of HF about 3-7 kcal/mol per halogen substituent; however, F
substituents have little effect on the barrier for loss of HCl. There
are at least two possible explanations for these observations:
(1) Halogen substituents have the ability to release or remove
electron density from their neighbor; fluorine is both a stronger
σ acceptor and π donor than chlorine. A complex interplay of
these two tendencies could be responsible for the observed trends
when multiple halogens are present so that analysis of substituent
effects on the threshold energies is not straightforward. If this
is the case, much more experimental and theoretical work will
be needed to unravel the intricacies of halogen substitution. (2)
Halogen substituents exert similar inductive effects, and the HCl
and HF transition states differ.7 Opposing trends in the E0’s
could result from an R-carbon that has little additional charge
development for HCl elimination but significant change in the
charge density when HF is removed. If there is minimal change
in the charge on the R-carbon for HCl elimination, then the
nature of the substituent would not influence E0 and H, F, and
Cl substituents would have similar threshold energies. This is
exactly the observed trend in the E0’s if Cl’s steric size accounts
for the decline in E0 with successive Cl substitution. We cannot
determine whether the R-carbon develops a more positive or
more negative charge as the transition state forms because it is
unclear if halogen substituents donate or remove electron
density.
The charges calculated for the CFCl2CH3 reactant are negative
on the F substituent (-0.26) with the two Cl’s being almost
neutral (-0.02). The R-carbon holds a partial positive charge
(+0.16) due to these electron withdrawing substituents. As the
transition state forms, electrons flow to the halogen being lost
and it develops a greater negative charge. In turn, all R-sub-
stituents become more positive as they feed electron density to
the R-carbon; the Cl substituents actually develop a greater
positive charge, and the F becomes less negative (see Figure
4). These changes on the leaving atom and on the substituents
occur irrespective of the identity of the halogen being eliminated.
However, the change in the R-carbons’s charge does depend
upon the halogen being lost. For HCl, the R-carbon becomes
more positive as the departing Cl removes electron density in
preparation for leaving with the hydrogen. For HF elimination,
the R-carbon accepts electron density from the substituents and
becomes negatively charged. This analysis obviously indicates
different transition states and would result in opposing trends
in the E0’s for substituents that donate electron density. For
loss of HF, the F and Cl substituents would intensify the
developing negative charge on the R-carbon and raise the E0’s.
By contrast, for loss of HCl, the donation of electron density
would stabilize the R-carbon’s additional positive charge and
decrease the threshold energy.
In analyzing substituent effects, it is essential that the change
in the electron density as the reactant forms the transition state
be considered. A carbon with a large atomic charge that
maintains the same charge in the transition state will not
experience any variation in E0 with substitution of halogen for
hydrogen, as the effect on the transition state is countered by a
corresponding change in the stability of the reactant. Several
conclusions of Toto et al.9 differ from ours; it appears that they
considered only the absolute and not the evolution of the atomic
charges.
Although the calculated charges9 are consistent with the trends
observed for CFCl2CH3 and CF2ClCH3, the magnitude appears
too large when HCl is removed. At the R-carbon the change
in the charge is greater for HCl versus HF elimination,
suggesting that Cl and F substituents would have a greater
influence on the E0(HCl) than E0(HF), counter to experimental
observations. Perhaps further refinement of the calculations
would improve agreement.
The movement of electrons just described is reminiscent of
models that involve three pairs of electrons.5,12 The electron
pair of the C-X bond shifts toward the departing halogen,
increasing its negative charge. A lone pair on the halogen shifts
The atomic charges computed9 for all possible chloro-/fluoro-
substituted ethanes will now be analyzed to determine if the
calculations support the proposal that the two transition states

3050 J. Phys. Chem., Vol. 100, No. 8, 1996
McDoniel and Holmes
toward the hydrogen as the H-X bond forms and the bond pair
of the C-H bond moves to become the π-bond between the
carbons.
References and Notes
(1) Hughes, I.; et al. J. Chem. Soc. 1937, 1271.
(2) Hughes, I.; et al. J. Chem. Soc. 1935, 244; 1927, 997.
(3) Maccoll, A. Chem. ReV. 1969, 33, 69.
We also note that the calculations9 suggest that there is almost
no change in electron density at the â-carbon when HCl is lost,
but the â-carbon becomes slightly more negatively charged at
the transition state for HF elimination. This suggests that
â-substituents will not alter E0(HCl). By contrast, electron
withdrawing substituents at the â-carbon will increase and
electron donating substituents will decrease E0(HF).
Assuming that these calculated atomic charges for the
R-carbon are of the correct sign, the models require both Cl
and F to provide electron density to the R-carbon. Comparing
the observed effects of the Cl and F substituents with CF3, a
known electron withdrawer, or CH3, capable only of providing
electron density, the sign on the R-carbon can be confirmed.
Experiments are in progress in our laboratory to provide
threshold energies for systems with R-CF3 or CH3 substituents,
and we will begin determining E0’s for 1,2-dehydrohalogena-
tions involving â-carbon substituents.
(4) Robinson, P. J.; Holbrook, K. A. Unimolecular Reactions; Wiley:
New York, 1972.
(5) Paisley, S. D.; Holmes, B. E. J. Phys. Chem. 1983, 87, 3042.
(6) Kim, K. C.; Setser, D. W. J. Phys. Chem. 1978, 78, 2166.
(7) Jones, Y.; Holmes, B. E.; Duke, D. W.; Tipton, D. L. J. Phys. Chem.
1990, 94, 4957.
(8) Egger, K. W.; Cocks, A. T. In The Chemistry of the Carbon-
Halogen Bond; Patai, S., Ed.; Wiley: London, 1973; Part 2, Chapter 10.
(9) Toto, J. L.; Pritchard, G. O.; Kirtman, B. J. Phys. Chem. 1994, 98,
8359.
(10) Haugen, G. R.; Benson, S. W. Int. J. Chem. Kinet. 1970, 2, 235.
(11) Kato, S.; Norokuma, K. J. Chem. Phys. 1980, 73, 3900.
(12) Hiberty, P. C. J. Am. Chem. Soc. 1975, 97, 5975.
(13) Bowles, R.; Majer, J. R.; Robb, J. C. Trans. Faraday Soc. 1962,
58, 1541.
(14) Haszeldine, R. N.; Nyman, F. J. Chem. Soc. 1961, 3015.
(15) Majer, J. R.; Robb, J. C.; Al-Saigh, Z. Y. J. Chem. Soc., Faraday
Trans. 1 1976, 72, 1697.
(16) Setser, D. W. J. Am. Chem. Soc. 1968, 90, 582.
(17) Tschuikow-Roux, E.; Faraji, F.; Niedzielski, J. Int. J. Chem. Kinet.
1986, 18, 513.
(18) Proksch, E.; Gehringer, P.; Szinovatz, W. J. Chromatogr. Sci. 1979,
17, 568.
(19) Blades, A. T. J. Chromatogr. Sci. 1973, 11, 267.
(20) Runyan, T. M.; Forrest, R. P.; Holmes, B. E. To be published.
(21) Durig, J. R.; Wurrey, C. J.; Bucy, W. E.; Sloan, A. E. Spectrochim.
Acta 1976, 3, 175.
(22) Tsang, W. J. Phys. Chem. 1986, 90, 414.
(23) Kumaran, S. S.; Lim, K. P.; Michael, J. V.; Wagner, A. F. J. Phys.
Chem. 1995, 99, 8673.
(24) Minato, T.; Yamabe, S.; Fujimoto, H.; Fukui, K. Bull. Chem. Soc.
Jpn. 1978, 1, 51.
(25) (a) Slagle, I. R.; Gutman, D.; Davies, J. W.; Pilling, M. J. J. Phys.
Chem. 1988, 92, 2455. (b) Wagner, A. F.; Wardlaw, D. M. J. Phys. Chem.
1988, 92, 2462.
(26) Arthur, N. L.; Davis, K. J. Aust. J. Chem. 1989, 42, 581.
(27) Danis, F.; Caralp, F.; Veyret, B.; Loirat, H.; Lesclaux, R. Int. J.
Chem. Kinet. 1989, 21, 715.
(28) Pritchard, G. O.; Bryant, J. T. J. Phys. Chem. 1968, 72, 1603.
(29) Perona, M. J.; Bryant, J. T.; Pritchard, G. O. J. Am. Chem. Soc.
1968, 90, 4782.
Conclusions
Unimolecular rate constants measured for the unimolecular
1,2-dehydrohalogenation of chemically activated CFCl2CH3
gave threshold energies indicating that fluorine and hydrogen
substituents are comparable for HCl elimination; fluorine’s
electronic effects are minimal, and E0 is only slightly affected.
For HF elimination, Cl substituents behave the same as F; they
raise E0 by 3-7 kcal/mol per halogen. A possible explanation,
contrary to the traditional view,3,4,6,8 is that the transition states
for 1,2-dehydrochlorination and 1,2-dehydrofluorination develop
partial charges at the R-carbon, the carbon containing the
departing halogen, that have different signs, and the Cl and F
substituents exert similar inductive effects; both either provide
or withdraw electron density from the R-carbon. The absolute
sign of the partial charge on the R-carbon cannot be determined
due to the 2-fold nature of halogen substituents; they are both
π donors and σ acceptors of electron density. Recent ab initio
calculations9 support dissimilar transition states for loss of HF
and HCl; assuming they furnish the correct sign for the partial
charge developing on the R-carbon, the Cl and F substituents
are electron donors. However, experimental confirmation for
the sign of the charge emerging as the reactant is converted to
products would be beneficial.
(30) Kim, K. C.; Setser, D. W.; Holmes, B. E. J. Phys. Chem. 1973,
77, 725.
(31) Majer, J. R.; Olavesen, C.; Robb, J. C. J. Chem. Soc. A 1969, 893.
(32) Lesclaux, R.; Caralp, F. Int. J. Chem. Kinet. 1984, 16, 1117.
(33) Chan, S. C.; Rabinovitch, B. S.; Bryant, J. T.; Spicer, L. D.;
Fujimoto, T.; Lin, Y. N.; Pavlon, S. J. Phys. Chem. 1970, 74, 3160.
(34) Dees, K.; Setser, D. W. J. Chem. Phys. 1968, 49, 1193.
(35) Christie, J. R.; Johnson, W. D.; Loudon, A. G.; MacColl, A.;
Mruzek, M. N. J. Chem. Soc., Faraday Trans. 1 1975, 71, 1937.
(36) Lafon, B.; Nielson, J. R. J. Mol. Spectrosc. 1966, 21, 175.
(37) Nielson, J. R.; Claassen, H. H.; Smith, D. C. J. Chem. Phys. 1950,
18, 1471.
(38) Marcoux, P. J.; Setser, D. W. J. Phys. Chem. 1978, 82, 97.
(39) Chang, H. W.; Craig, N. L.; Setser, D. W. J. Phys. Chem. 1972,
76, 954.
(40) Rayez, M.; Rayez, J.; Sawersyn, J. J. Phys. Chem. 1994, 98, 11342.
(41) Joshi, R. M. J. Macromol. Sci., Chem. 1974, A8, 861.
(42) Pamidimukkala, K. M.; Rogers, D.; Skinner, G. B. J. Phys. Chem.
Ref. Data 1982, 77, 5877.
Acknowledgment. We are pleased to acknowledge financial
support provided by the National Science Foundation (Grant
CHE-91222327) and by the Donors of The Petroleum Research
Fund, administered by the American Chemical Society. J.B.M.
is grateful for summer research stipends from the Council on
Undergraduate Research-CURSOR and the Science Information
Liaison Office-Summer Undergraduate Research Fellowship
Program from the State of Arkansas. Assistance from Dr.
Stephen Kempson was helpful during early stages of the work.
(43) De More, W. B.; Sander, S. P.; Golden, D. M.; Hampson, R. F.;
Kurylo, M. J.; Howard, C. J.; Ravishankara, A. R.; Kolb, C. E.; Molina,
M. J. JPL Publication 92-20; Jet Propulsion Laboratory: Pasadena, CA,
1992.
JP9523128