Absolute rate constants for the addition of the 2- (methoxycarbonyl)propan-2-yl and the 3,3,3-trifluoracetonyl radicals to alkenes in solution

Article abstract of DOI:10.1002/(SICI)1522-2675(20000315)83:3<658::AID-HLCA658>3.0.CO;2-Z

Absolute rate constants and some of their Arrhenius parameters were obtained by time-resolved electron spin resonance (ESR) spectroscopy for the addition of the 2-(alkoxycarbonyl)propan-2-yl and 3,3,3-trifluoroacetonyl (= 3,3,3-trifluoro-2-oxopropyl) radicals to a variety of mono- and 1,1- disubstituted alkenes. Their analysis shows that the addition of 2- (alkoxycarbonyl)propan-2-yl is mainly governed by the exothermicity of the reaction with slight modifications by nucleophilic and electrophilic effects giving rise to an overall ambiphilic behavior. In contrast, large electrophilic polar effects dominate the addition of the 3,3,3- trifluoroacetonyl (= 3,3,3-trifluoro-2-oxopropan-2-yl) radical, as it is expected from its large electron affinity. For both radicals, the activation energies are well-predicted by analytic equations for the enthalpic and polar terms. A comparison of the rate data of 2-(alkoxycarbonyl)propan-2-yl with the homo- and copolymerization rate constants of the propagating radical of methyl methacrylate shows that the additions of these structurally related low- and high-molecular-weight radicals to alkenes are governed by very, similar effects.

Full text of DOI:10.1002/(SICI)1522-2675(20000315)83:3<658::AID-HLCA658>3.0.CO;2-Z

658
Helvetica Chimica Acta ± Vol. 83 (2000)
Absolute Rate Constants for the Addition of the
2-(Methoxycarbonyl)propan-2-yl and the 3,3,3-Trifluoracetonyl Radicals
to Alkenes in Solution
by Torsten Zytowski, Bernhard Knühl, and Hanns Fischer*
Physikalisch-chemisches Institut der Universität Zürich, Winterthurerstrasse 190, CH-8057 Zürich
Absolute rate constants and some of their Arrhenius parameters were obtained by time-resolved electron
spin resonance (ESR) spectroscopy for the addition of the 2-(alkoxycarbonyl)propan-2-yl and 3,3,3-
trifluoroacetonyl (ˆ 3,3,3-trifluoro-2-oxopropyl) radicals to a variety of mono- and 1,1-disubstituted alkenes.
Their analysis shows that the addition of 2-(alkoxycarbonyl)propan-2-yl is mainly governed by the
exothermicity of the reaction with slight modifications by nucleophilic and electrophilic effects giving rise to
an overall ambiphilic behavior. In contrast, large electrophilic polar effects dominate the addition of the 3,3,3-
trifluoroacetonyl (ˆ 3,3,3-trifluoro-2-oxopropan-2-yl) radical, as it is expected from its large electron affinity.
For both radicals, the activation energies are well-predicted by analytic equations for the enthalpic and polar
terms. A comparison of the rate data of 2-(alkoxycarbonyl)propan-2-yl with the homo- and copolymerization
rate constants of the propagating radical of methyl methacrylate shows that the additions of these structurally
related low- and high-molecular-weight radicals to alkenes are governed by very similar effects.
1. Introduction. ± Additions of C-centered radicals to CˆC bonds are very useful
reactions and have many synthetic applications [1]. On variation of radical and
substrate, the rate constants change by many orders of magnitude, and, since the
beginning of quantitative radical chemistry, this is understood to be caused by a
complex interplay of polar, steric, and enthalpic effects [2]. To better separate these
effects, we have determined a large number of rate constants and their temperature
dependencies for the addition of several C-centered radicals to many mono- and 1,1-
disubstituted alkenes in liquid solution by time-resolved electron spin resonance (ESR)
and related techniques [3].
Recently, these data led to a comprehensive and predictive analysis [4] in terms of
the state or configuration diagram [5]. We showed that, for all additions to mono- and
1,1-disubstituted alkenes, the frequency factors are within narrow ranges, and that, in
general, the rate constants increase, and the activation energies decrease with
increasing exothermicity of the reaction, as is predicted by the Evans-Polanyi relation-
ship. In addition, for many radicals, the activation energies are lowered in a
multiplicative way by polar factors which depend on the radical and the alkene
.
substituents. Thus, the easily oxidizable radicals tert-butyl (Me3C ), hydroxymethyl
.
.
(CH2OH), and 2-hydroxypropan-2-yl (Me2COH) show a strong nucleophilic addition
behavior towards most alkenes, such that even the omnipresent enthalpic factor is
.
.
partly obscured. The addition rate constants of the methyl (CH3), the benzyl (PhCH2),
and the cumyl radical (PhCMe2) follow the reaction enthalpy order more closely and
exhibit only weak nucleophilic effects. Enthalpic effects also dominate the additions of
.
.
.
the 2-cyanopropan-2-yl (Me2CCN), the cyanomethyl (CH2CN), and the (tert-

Helvetica Chimica Acta ± Vol. 83 (2000)
659
.
butoxycarbonyl)methyl (CH2CO2CMe3) radical, but these species are also slightly
.
electro- or ambiphilic. Finally, malonyl radicals (HC(CO2R)2) exhibit a clear rate
enhancement by electrophilic polar effects as do perfluoroalkyl [6] and dicyanomethyl
[7] radicals.
To date, our series included two tertiary radicals with electron-donor substituents
.
.
(Me3C and Me2COH) but only one tertiary species with an electron-acceptor
.
substituent (Me2CCN) and only one type of radical (a cyclic and an open-chain malonyl
radical) with a clear electrophilic addition behavior. Therefore, the series was extended
here to the addition of 2-(alkoxycarbonyl)propan-2-yl species, namely 2-(methoxy-
.
carbonyl)propan-2-yl (Me2CCO2Me) and, for one alkene, the 2-(tert-butoxycarbon-
.
yl)propan-2-yl (Me2CCO2CMe3) radical. The latter radicals are also models for the
technically important propagating radical of methacrylates. Further, we investigated
the addition of the 3,3,3-trifluoroacetonyl (ˆ 3,3,3-trifluoro-2-oxopropan-2-yl) radical
.
(CF3COCH2) to obtain a further example for a strongly electrophilic radical.
2. Methods and Results. ± The arrangements and procedures for steady-state and
time-resolved ESR have been described in detail in earlier work [3][4].
2.1. Steady-State Experiments. 2.1.2. 2-(Methoxycarbonyl)propan-2-yl and 2-(tert-
Butoxycarbonyl)propan-2-yl. tert-Butyl isobutyrate was synthesized according to a
literature procedure [8], but with N,N-dimethylaniline instead of pyridine as base and
omission of the ether solvent. The product was washed five times with 10% (v/v)
H2SO4, conc. K2CO3, and H2O, was dried (CaSO4) and found to be sufficiently pure (by
GC). The radicals were generated via H-atom abstraction by tert-butoxy radicals from
methyl isobutyrate (Fluka, >99% (GC)) and tert-butyl isobutyrate (Eqns. 1 and 2),
and the best results were obtained with solutions of 1.8m distilled di(tert-butyl)
1
peroxide (DTPB; Merck, e ꢀ 7 m 1 cm at l ꢀ 270 nm, >98% (GC)) in 5.9m methyl
isobutyrate or in 4.2m tert-butyl isobutyrate.
.
(Me3CO)2 ‡ hn ! 2 Me3CO
(1)
.
.
Me3CO ‡ R H ! Me3COH ‡ R
1
At room temperature, the H-abstraction reaction is rather fast (kH ˆ 2.5 ´ 105 m 1 s
[9]), and it was not resolved in time even at lower temperatures. ESR Spectra re-
corded during steady-state photolysis revealed the expected formation of the radicals
.
R ˆ 2-(methoxycarbonyl)propan-2-yl and 2-(tert-butoxycarbonyl)propan-2-yl by ab-
straction from the tertiary C-atom of the isobutyrates. The ESR spectra of these
.
radicals are quite complex, because there is a hindered rotation about the C CO bond
[10] which renders the two Me groups inequivalent, and because of the appreciable
second-order splittings. However, for moderate resolution 2-(methoxycarbonyl)pro-
.
pan-2-yl (Me2CCO2Me) shows a simple septuplet of quadruplets as presented in the
inset of Fig. 1. The hyperfine coupling constants a(6 Hb) ˆ (21.58 Æ 0.03) G and
a(3 Hd) ˆ (1.25 Æ 0.03) G at 294 K compare well with earlier literature data (g ˆ
2.00316, a(3 Hb-cis) ˆ 21.49 G, a(3 Hb-trans) ˆ 21.68 G, and a(3 Hd) ˆ 1.29 G at 233 K
.
[11]). The same holds for 2-(tert-butoxycarbonyl)propan-2-yl (Me2CCO2CMe3) which
has an average a(6 Hb) ˆ (21.51 Æ 0.03) G at 294 K ([10]: g ˆ 2.00329, a(3 Hb-cis) ˆ
21.44 G, a(3 Hb-trans) ˆ 21.58 G at 233 K).The spectra revealed additional small signals

660
Helvetica Chimica Acta ± Vol. 83 (2000)
arising from photoreduction products of the esters. The concentrations of these side
products were at least 20 times lower than those of the dominant radicals, and,
therefore, they were neglected in the kinetic analyses. Also, the possible b-scission of
the tert-butoxy radical [12] was ignored, because methyl radicals were not detected in
the whole temperature range of 233 to 333 K.
Fig. 1. Concentration of the 2-(methoxycarbonyl)propan-2-yl radical vs. time a) in the absence of alkene and b) in
the presence of 62 mm acrylonitrile. Bottom traces are residuals of fits. The inset shows a steady-state ESR
spectrum of the radical.
Solutions containing acrylonitrile showed ESR spectra of radicals formed by
addition of the ester-derived radicals to the alkene CH2 group. Their magnetic
properties, a(Ha) ˆ 20.04 G, a(2 Hb) ˆ 18.60 G, a(14N) ˆ 3.40 G, g ˆ 2.0027, and
a(Ha) ˆ 19.85 G, a(2 Hb) ˆ 18.61 G, a(14N) ˆ 3.37 G, g ˆ 2.0028 for the adducts of 2-
(methoxycarbonyl)propan-2-yl and 2-(tert-butoxycarbonyl)propan-2-yl, respectively,
are typical for the addition products of fairly large radicals [11]. They deserve no
further comment, but the radical structures ensure that the addition occurs highly
predominantly at the unsubstituted C-atom, as has previously been found for many
other radicals [3][4].
.
2.1.2. 3,3,3-Trifluoroacetonyl. The radical CF3COCH2 was generated by photolysis
of dibenzoyl peroxide (Eqn. 3) and subsequent H-abstraction from trifluoroacetone in
1,1,2-trichloro-1,2,2-trifluoroethane (Freon 113, Aldrich, 99%). The best signal-to-
noise ratio of the desired radical and minimum side products were obtained with
solutions containing 10 mm dibenzoyl peroxide (Aldrich, e ꢀ 2400 mol 1 cm at
1
.
273 nm) and 1m 1,1,1-trifluoroacetone. Since the PhCO2 radical decarboxylates within
.
200 ns [13] the CF3COCH2 radical is presumably formed by the H-abstraction of the

Helvetica Chimica Acta ± Vol. 83 (2000)
661
phenyl radical. ESR Spectra recorded during continuous photolysis showed the
.
expected triplet of CF3COCH2 together with other smaller signals (see inset in Fig. 4).
.
.
(C6H5CO2)2 ‡ hn ! 2(1 f)C6H5CO2 ‡ 2 f C6H5 ‡ 2 f CO2 f298 K ˆ 0.29 (3)
.
.
C6H5CO2 ! C6H5 ‡ CO2
(4)
.
.
C6H5 ‡ RH ! C6H6 ‡ R
(5)
The latter are compatible with the 1,1,2-trifluoro-1,2-dichloroethyl radical [14] formed
by Cl-atom abstraction from the solvent by phenyl radicals. These amount to less than
8% of the total signal intensity. Since also only traces of chlorobenzene were found by
GC as products in the photolyzed mixture, the Cl-atom abstraction was again neglected
.
in the further analysis. At room temperature, the CF3COCH2 radical has the coupling
constants a(2 Ha) ˆ (19.22 Æ 0.05) G and a(3Fg) ˆ (0.58 Æ 0.05) G. The g factor g ˆ
2.0047 Æ 0.0001 was measured at 297 K with a solution of di(tert-butyl) peroxide and
1,1,1-trifluoroacetone in Freon 113 relative to methyl (g ˆ 2.0025 [30]), which appears
in the same spectrum. These parameters agree very well with those of other a-carbonyl-
substituted methyl radicals [11].
Spectra obtained with alkene-containing solutions revealed secondary radicals from
additions to the alkene CH2 groups. The adduct to acrylonitrile exhibits a small long-
range coupling to two protons which was the same in the presence and the absence of
trifluoroacetone in the solution. Hence, the radical is probably the adduct of the phenyl
.
and not of the CF3COCH2 radical. This means that, in this case, the addition of phenyl
to acrylonitrile is faster than the H-abstraction from CF3COCH3, and the same may be
true for the other alkenes. Therefore, the spectra of the adduct radicals did not prove
.
the addition of CF3COCH2 directly.
2.2. Time-Resolved Experiments. 2.2.1. 2-(Methoxycarbonyl)propan-2-yl and 2-(tert-
Butoxycarbonyl)propan-2-yl. Fig. 1 shows kinetic traces obtained at 294 K for
2-(methoxycarbonyl)propan-2-yl during intermittent photolysis for an alkene free
(trace a) and for an alkene-containing (trace b) solution. According to the usual
analysis [3][4], trace a is governed by a second-order decay, which is due to radical self-
termination (t2 ꢀ 800 ms) and is slightly perturbed by a first- or pseudo-first-order
process (t10 ˆ (80 Æ 40) ms). This is attributed to reactions with the starting compounds
and/or products. The same behavior was found for 2-(tert-butoxycarbonyl)propan-2-yl
in the absence of alkenes. Rate constants for the addition of the 2-(alkoxycarbonyl)-
propyl radicals were determined from the pseudo-first-order lifetimes t1 obtained for
different alkene concentrations. Trace b in Fig. 1 corresponds to an acrylonitrile
concentration of 62 mm and led to a shorter t1 ˆ 5.53 ms and t2 ˆ 772 ms. Plots of t1 1 vs.
the alkene concentrations were linear and lead to the addition rate constants ka from
the slopes. Fig. 2 shows two examples. The non-zero value of t1 1 for [A] ˆ 0 represents
the side reactions mentioned above. Experiments at different temperatures then gave
the activation parameters for the additions by fits of the Arrhenius expression to the
data (Fig. 3).
All results obtained for the 2-(alkoxycarbonyl)propan-2-yl radical are presented in
Table 1, where the alkenes are ordered according to increasing reactivity. We also give
the ranges of the applied alkene concentrations and the number of evaluated kinetic

662
Helvetica Chimica Acta ± Vol. 83 (2000)
Fig. 2. Pseudo-first-order plot for the addition of the 2-(methoxycarbonyl)propan-2-yl radical to methyl
methacrylate and acrylonitrile at 294 K
experiments. The addition to several gaseous alkenes of our standard set [3][4] was not
measured, because these were not sufficiently soluble in the present solvent.
2.2.2. 3,3,3-Trifluoroacetonyl. Kinetic traces obtained during intermittent photolysis
in the absence (a) and presence (b) of an alkene are presented in Fig. 4. In trace a, the
decay is dominated by the second-order self-termination of the radical (t2 ˆ 500 ms),
which is again slightly perturbed by a slow first- or pseudo-first-order process (t10 ˆ
.
14.5 ms). The latter is attributed to a minor reaction of CF3COCH2 with the solvent
.
and/or the starting compounds. From the life-time t2, the rate constant of CF3COCH2
self-termination at room temperature was obtained as 2kt(298 K) ˆ (1.23 Æ 0.05) ´ 1010
m
1 s by a standard calibration procedure vs. the tert-butyl self-termination rate
1
1
constant of 2kt(297 K) ˆ 3.4 ´ 109
m
1 s in tetradecane [15]. The self-termination
.
.
constant of CF3COCH2 is similar to that of the CH3 radical in the same solvent
2kt(298 K) ˆ (1.65 Æ 0.2) ´ 1010 m 1 s 1 [30], and it is on the order of magnitude expected
for a diffusion-controlled process.
With alkene present (Fig. 4, trace b), the first-order life-time decreases as expected
from the addition reaction. As before, the addition rate constants were determined from
the slopes of the linear plots of t1 1 vs. the alkene concentrations, and examples are given
in Fig. 5. Unfortunately, for this radical, the temperature dependence could not be
measured, because the low boiling points of the solvent and of trifluoroacetone prevented
measurements at elevated temperatures, whereas dibenzoyl peroxide was found to be
not sufficiently soluble below room temperature. Table 2 shows the rate constants, the
number of evaluated kinetic traces, and the variation range of the alkene concentration.

Helvetica Chimica Acta ± Vol. 83 (2000)
663
Fig. 3. Rate constants for the addition of the 2-(methoxycarbonyl)propan-2-yl radical to acrylonitrile and to 1,1-
dichloroethene at various temperatures and fits to the Arrhenius law
3. Analysis and Discussion. ± 3.1. Rate Constants. Apart from the propagation rate
constants and the copolymerization parameters of methyl methacrylate, which are
related to the addition rate constants of the 2-(alkoxycarbonyl)propan-2-yl radicals,
and which are discussed later, there are only very few literature data with which our
rate constants can be compared. Gilbert et al. [16] used an indirect steady-state ESR
method to measure the addition rate constant of the 2-carboxypropan-2-yl radical
.
(Me2CCO2H) to acrylic and methacrylic acid at pH 2 in H2O and at room temperature.
1
1
They report k ˆ 5 ´ 105 m 1 s and k ˆ 7 ´ 105 m 1 s , respectively, and these values are
by more than two orders of magnitude larger than our results (Table 1). Though H2O as
a solvent normally accelerates additions in comparison to less polar media [3o][4b],
and the different radical and alkene substitutions may also cause slight changes of the
rate constants, we believe that the results of Gilbert et al. are not very likely to be
tenable. The order of magnitude of their findings disagrees not only with ours but also
with that expected for the addition rate constants of the propagating radical of methyl
.
methacrylate (vide infra). For CF3COCH2 and related species there are not earlier
addition rate data.
A comparison of the rate constants of the 2-(alkoxycarbonyl)propan-2-yl radicals
of Table 1 with those of other radicals in our series [3][4] shows that these radicals
are only moderately selective, since the rate constants at room temperature vary
with alkene substitution by only a factor of 550. Overall, their reactivity is also

664
Helvetica Chimica Acta ± Vol. 83 (2000)
Table 1. Absolute Rate Constants (k) at (294 Æ 1) K, Frequency Factors (A), and Activation Energies (Ea) for
the Addition of the 2-(Methoxycarbonyl)propan-2-yl Radical to Alkenes (CH2ˆCXY)
X
Y
c [mM]
n
k [m 1 s 1]a)
Ea [kJ mol 1]a)b)
log(A/m 1 s 1)a)
(T [K])
H
AcO
MeO
AcO
EtO
Me3Si
Cl
1.0 ± 4.0
1.28
2.0 ± 4.0
1.0 ± 4.0
1.0 ± 2.0
0.5 ± 1.0
14
8
18 (3)
50 (25)
58 (2)
75 (25)
75 (20)
150 (35)
1150 (70)
35.6
33.1
32.7
32.1
32.1
30.4
19.8 (12)
25.3
21.9 (8)
25.3
Me
Me
H
H
Me
H
13
12
12
10
39
CO2Me
0.125 ± 0.400
(255 ± 315)
0.239 ± 0.487
(253 ± 323)
0.050 ± 0.110
0.050 ± 0.200
(248 ± 325)
0.046 ± 0.133
0.025 ± 0.100
0.050 ± 0.100
(253 ± 318)
0.041 ± 0.110
0.010 ± 0.033
(251 323)
0.069 ± 0.083
(295 ± 315)
6.6 (‡ 2/ 3)
6.9 (‡ 1/
Cl
Cl
34
1170 (30)
)
H
H
CHO
CN
10
23
1800 (300)
2650 (100)
23.3
3710 (250)22.4
4450 (200)
5500 (400)
24.2
21.0 (4)
7.2 (‡ 1/ 1)
6.7 (‡ 3/ 3)
Me
Me
H
CO2Me
CN
Ph
17
15
13
22.0
16.5 (12)
21.5
21.2
16.6 (5)
20.0
Me
Ph
Ph
Ph
20
13
6050 (250)
10200 (450)
7.0 (‡ 1/ 1)
7.5 (‡ 2/ 2)
H
CNc)
23
4400 (200)
21.8 (10)
21.8
a) Standard deviations in units of the last digital number are given in brackets. b) Activation energies without
c
errors were calculated from the rate constants with log(A/m 1 s 1) ˆ 7.5.
) Addition of the 2-(tert-
butoxycarbonyl)propan-2-yl radical.
rather low. It is slightly higher than, but closely resembles, that of the 2-cyanopropan-2-
yl radical, is higher than that of the resonance-stabilized benzyl radical, and is
considerably lower than that of the (tert-butoxycarbonyl)methyl and the methyl
radicals.
.
The addition rate constants of CF3COCH2 vary with alkene substitution by a factor
of about 250, i.e., this radical is also not very selective. However, its reactivity is high. It
.
resembles that of the malonyl radical, and to most alkenes the CF3COCH2 radical adds
.
even faster than the CH3 radical. From the substitution of the radicals and in view of
our earlier experience, these trends are reasonable, and they will be analyzed in the
following in greater detail.
3.2. Frequency Factors. Table 1 reveals that the frequency factors for the addition of
the 2-(alkoxycarbonyl)propan-2-yl radicals alkenes are all in the narrow range of 6.6 <
1
log(A/m 1 s ) <7.5 and show no specific dependence on the alkene substituents.
Therefore, as also found for other radicals [3][4], the spread is presumably caused by
error-compensation effects of the Arrhenius parameters. The average frequency factor
1
is log (A/m 1 s ) ˆ 7.0 Æ 0.3. Within the error limits, this is within the average of log (A/
1
m
1 s ) ˆ 7.5 Æ 0.8, suggested earlier for tertiary radicals [3][4], and ensures the
general validity of the latter value. To obtain the activation energies for all alkenes of
1
Table 1 from the rate constants, the average log(A/m 1 s ) ˆ 7.5 was used.

Helvetica Chimica Acta ± Vol. 83 (2000)
665
Fig. 4. Concentration of the 3,3,3-trifluoroacetonyl radical vs. time a) in the absence of alkene and b) in the
presence of 3.7 mm trimethyl(vinyl)silane. Bottom traces are residuals of fits. The inset shows a steady-state ESR
spectrum of the radical.
.
For CF3COCH2 additions, the frequency factors could not be measured. However,
in previous studies [3][4], we have established an average frequency factor for the
1
addition of primary radicals to alkenes as log(A/m 1 s ) ˆ 8.5 Æ 0.3, and have now used
this value to convert the rate data of Table 2 to activation energies for further
discussion.
That the frequency factors are similar within one order of magnitude indicates
rather similar transition-state structures for the various additions, as is also known from
theoretical work [17]. While the newly forming bond is still long (210 ± 240 pm), the
angle of attack is already tightly prescribed, and there is a considerable out-of-plane
deformation at the two newly bonded C-atoms. Moreover, the dependence of the
frequency factors on radical structure that emerged from the data is also reasonable.
The tertiary radicals should loose more motional freedom than primary radicals on
approach to the transition structure, because the rotation of methyl or other substituent
groups becomes more hindered. This provides a more negative activation entropy for
the tertiary radicals in comparison to the primary species and, hence, the lower
frequency factor.
3.3. Activation Energies. The narrow range of the frequency factors for the additions
means that the variation of the rate constants with the alkene substitution is to the
larger part due to variations of the activation energy. Previously, this has also been
found for other C-centered radicals [3][4]. To rationalize the individual substituent
effects, we shall now seek correlations of the activation energies with the reaction

666
Helvetica Chimica Acta ± Vol. 83 (2000)
Fig. 5. Pseudo-first-order plot for the addition of the 3,3,3-trifluoroacetonyl radical to 2-chloropropene and
methacrylonitrile at 295 K
Table 2. Absolute Rate Constants for the Addition of the 3,3,3-Trifluoroacetonyl Radical to Alkenes
(CH2ˆCXY) in Freon 113 at (296 Æ 1) K
X
Y
c [mM]
n
k [104 m 1 s 1]a)
Ea [kJ mol 1]b)
H
H
H
H
H
H
Me
H
H
CN
H
CHO
Cl
CO2Me
Me
1.5 ± 6.0
1.1 ± 4.7
1.4 ± 6.4
0.9 ± 3.4
0.9 ± 3.7
0.7 ± 2.3
0.4 ± 1.5
0.5 ± 2.2
0.9 ± 2.2
0.5 ± 1.9
0.3 ± 1.1
0.2 ± 1.1
0.2 ± 0.8
0.1 ± 0.8
0.04 ± 0.18
0.02 ± 0.09
0.02 ± 0.09
0.01 ± 0.13
14
20
17
15
15
24
16
15
14
23
27
20
19
19
12
13
17
14
3.5 (3)
3.9 (2)
6.4 (6)
9.5 (2)
12.5 (5)
19 (1)
19 (1)
20 (7)
21 (1)
28 (1)
33 (2)
50 (2)
59 (3)
120 (60)
400 (10)
630 (70)
690 (30)
770 (60)
22.6
24.0
21.1
20.1
19.3
18.4
18.4
18.3
18.1
17.4
17.0
16.0
15.3
13.8
10.8
9.7
CN
AcO
Me3Si
Et
Cl
Cl
AcO
CO2Me
Me
Ph
Ph
H
Cl
Me
Me
Me
Me
H
Ph
Me
9.5
9.2
Ph
a) Standard deviations in units of the last digital number are given in parenthesis. b) Calculated from the rate
constants with log(A/m 1 s 1) ˆ 8.5.

Helvetica Chimica Acta ± Vol. 83 (2000)
667
enthalpies Hr and, for polar effects, with the radical- and alkene-ionization energies IE
and electron affinities EA. Steric effects of the alkene substituents on the activation
energies are not considered, since they should be small for additions at the
unsubstituted carbon, and an inspection of the data does not reveal any trends.
In our earlier analysis [4b] of many data for the addition of C-centered radicals to
mono- and 1,1-disubstituted alkenes, which included those presented here, we have
shown that the activation energies Ea are well-described by Eqn. 6.
Ea ˆ (50 ‡ 0.22 Hr)Fn Fe
(6)
The first term in brackets represents the Evans-Polanyi relationship, according to
which the activation energy decreases with increasing exothermicity ( Hr) of the
addition. It describes the activation energies alone where polar effects are absent and is
an upper limit. The polar factors Fn and Fe (0 < Fn, e < 1) decrease the activation energy.
The nucleophilic part Fn depends on the energy gap between the ground-state
configuration of the reactants (RA, where R denotes the radical and A the alkene) and
‡
the charge-transfer configuration R A reduced by a Coulomb interaction in the
transition state geometry Cn, i.e., on IE(R) EA(A)
Cn, and on an interaction
parameter gn between the configurations. Its electrophilic analog Fe (0 < Fe < 1)
depends on IE(A) EA(R) Ce and a corresponding interaction parameter ge. Both
polar factors Fn and Fe are restricted to values between 0 and 1, because the polar
effects can only reduce the reaction barriers. As reasonable functions for Fn and Fe, we
have suggested Eqns. 7 and 8 and have presented suitable ranges for the parameters C
and g [4b].
Fn ˆ 1 exp [ ((IE(R) EA(A) Cn)/gn)2]
Fe ˆ 1 exp [ ((IE(A) EA(R) Ce)/ge)2]
(7)
(8)
Moreover, it has been pointed out [4] that the energies Hr, EA(A), and IE(R) are
themseves interrelated. Thus, substituents that make an alkene electron-deficient and
increase the alkene electron affinity often also stabilize the resulting adduct radical, i.e.,
render the reaction more exothermic. On the other hand, substituents that make
alkenes more electron-rich and decrease their ionization energies often decrease the
exothermicity. Nucleophilic radicals have low ionization energies. For these the effects
of alkene substituents that increase the exothermicity (enthalpic effect) and simulta-
neously increase the electron affinity (polar effect) reinforce each other. Hence,
nucleophilic radicals are generally highly selective and also particularly reactive
towards the electron-deficient alkenes. Electrophilic radicals have high electron
affinities. For these, the favorable polar effect runs opposite to favorable enthalpic
substituent effects. Therefore, the radicals are not very selective. However, they are
more reactive than radicals that exhibit only the same enthalpic effects, since Fe is
smaller than one. Radicals that are on the borderline of a nucleophilic and an
electrophilic behavior may exhibit an ambiphilic nature and react particularly fast with
both strongly electron-deficient and electron-rich alkenes.
In a schematic way, Fig. 6 shows how the nature of the species reveals itself in a plot
of the activation energies with the reaction enthalpy. It must also be mentioned that in

668
Helvetica Chimica Acta ± Vol. 83 (2000)
such plots the Ph-substituted alkenes are often outlyers if the polar effects are
important. They show higher activation energies than other compounds with similar
reaction enthalpies, similar ionization energies, and electron affinities, because the
charge delocalization in the transition structure diminishes both the Coulomb
attraction C and the interaction parameters g, and this leads to smaller rate
enhancements by polar effects.
Fig. 6. Schematic representation of the dependence of the activation energy Ea on the reaction enthalpy Hr for the
addition of alkyl radicals to mono- and 1,1-disubstituted alkenes for a) pure enthalpy control, and additional
b) nucleophilic, c) electrophilic, and d) ambiphilic polar effects
To apply Eqn. 6, the reaction enthalpies Hr of the addition step have to be known.
As before [3][4], they have been estimated, because experimental data are not
available. We consider the reaction sequence in Eqn. 9, from which the reaction
enthalpy can be written as a combination of molar heats of formation, hf, and bond
dissociation energies, BDE, of stable compounds (Eqn. 10), where A stands for the
.
.
.
.
R
H ‡ CH2 ˆ CXY! R‡ H‡ CH2ˆCXY! H‡ RCH2CXY! RCH2CHXY
(9)
.
Hr(R) ˆ hf(RAH) hf(RH) hf(A) ‡ BDE(RA H) BDE(R H)
(10)
alkene. It is obvious from Eqn. 10 that it includes the effects of stabilization of the
attacking and resulting radical and of the alkene. Values for hf(RAH) and
BDE(RA H) are known for many compounds with the residue methyl (RˆCH3)
but not for the radicals considered here. Therefore, we start from the most recent
.
and re-evaluated values of Hr [3o] for CH3 radical additions to the alkenes and
increment these for the substitution of methyl by 2-(methoxycarbonyl)propan-2-yl and

Helvetica Chimica Acta ± Vol. 83 (2000)
669
.
CF3COCH2, i.e., according to Eqn. 11. This procedure implies that BDE(RA H) is
.
not influenced by the added radical R.
.
.
Hr(R) ˆ Hr(CH3) ‡ hf(RA H) hf(CH3AH) hf(RH) ‡ hf(CH4)
BDE(R H) ‡ BDE(CH3 H)
(11)
For the estimation of the Hr value for addition of the 2-(methoxycarbonyl)propan-
2-yl radical to ethene, the following (gas phase) standard heats of formation
1
1
are available [18]: CH4:
(CH3)CHCO2CH3:
74.5 kJ mol , CH3CH2CH3:
104.5 kJ mol ,
1
456 kJ mol , but hf seems to be unknown for
CH3CH2C(CH3)2CO2CH3. It can be estimated as 511 kJ mol from the tabulated
1
1
1
hf((CH3)3CCO2CH3) ˆ 491 kJ mol by addition of the increment of 20 kJ mol
for substitution of a Me by an Et group, which was derived [3o] from various related
1
pairs of molecules found in [18]. The bond dissociation energies are 439 kJ mol for
CH3 H [19] and 379 kJ mol for H C(CH3)2CO2CH3 [30]. Insertion of these data
1
into Eqn. 11 shows that the addition of 2-(methoxycarbonyl)propan-2-yl radical to
.
ethene is by 35 kJ mol 1 less exothermic than that of CH3. This difference is assumed to
hold also in solution and for the other alkenes. Of course, the estimation procedure
1
leads to errors which may be as large as 10 kJ mol . However, the low exothermicity of
the additions immediately suggests a rather low reactivity of 2-(alkoxycarbonyl)pro-
pan-2-yl radicals as it is observed.
.
For CF3COCH2, the same procedure was applied to obtain Hr values for the
1
addition to ethene: hf(CF3COCH3) ˆ 812 kJ mol
is known [18], and
1
BDE(H CH2COCF3) ˆ 408 kJ mol was calculated according to the procedure of
1
Cumming and Kebarle [20]. hf(CF3COCH2CH2CH3) ˆ 853.7 kJ mol was obtained
from hf(CF3COCH3) and from the difference of 41.7 kJ mol 1 between the heats of
formation of CH3COCH3 and CH3COCH2CH2CH3 [18]. With the data given above for
.
CH4 and CH3CH2CH3, and Eqn. 11, these energies render the addition of CF3COCH2
.
1
to ethene by 16 kJ mol less exothermic than that of CH3. Again, this difference
in Hr was assumed to hold also for the other alkenes, and the small difference with
.
.
CH3 allows in part for the observed rather high rate constants of CF3COCH2
additions.
To discuss the polar factors, we also need the ionization energies, IE, and the
electron affinities, EA, of the alkenes and of the radicals. For the alkenes, these data
have been listed in our earlier work [3o]. For 2-(methoxycarbonyl)propan-2-yl and
.
CF3COCH2, the IEs and EAs were calculated by a high-level ab initio procedure as
IE ˆ 7.7 eV, EA ˆ 1.34 eV, and IE ˆ 10.9 eV, EA ˆ 3.18 eV, respectively [21], and, for
.
CF3COCH2, an experimental value of EA ˆ 2.625 eV is also available [22]. In
comparison to other radicals [3][4], the 2-(methoxycarbonyl)propan-2-yl radical has
an IE which is larger than that of strongly nucleophilic radicals, such as 6.7 eV for tert-
butyl, and an EA lower than those of strongly electrophilic species, such as 1.8 eV for
malonyl. Hence, the addition reactions of this radical should not be subject to very
.
strong polar effects. On the other hand, the EA of CF3COCH2 is larger than that of the
electrophilic malonyl radical (1.8 eV), and therefore large electrophilic polar effects
are expected. In combination with the high exothermicity, this explains why
trifluoroacetonyl is often even more reactive than the methyl radical.

670
Helvetica Chimica Acta ± Vol. 83 (2000)
Fig. 7 shows the activation energies calculated from the rate constants as described
above for the addition of the 2-(methoxycarbonyl)propan-2-yl radical and plotted
against the reaction enthalpies. Also shown is the line expected from Eqn. 6 for
complete enthalpy control and negligible polar effects. A linear correlation of the data
1
with the reaction enthalpies gave Ea/kJ mol 1 ˆ 47.9(3.0) ‡ 0.24(0.03)Hr/kJ mol (rˆ
0.87), which is very reasonable and not far from the general Evans-Polanyi line.
Clearly, and in view of the error limits of the experimental data and the estimated
reaction enthalpies, the activation energies point to a dominant enthalpy control. No
correlation was found between Ea or log (k) and the alkene ionization energies, which
excludes large electrophilic polar effects, but there is also a reasonable correlation with
the alkene electron affinities which is probably caused by the interrelation of Hr and
EA. However, in Fig. 7 there are noticeable deviations from the ideal line to lower
values for both the strongly electron-deficient (e.g., methyl methacrylate, methacry-
lonitrile) and the electron-rich alkenes (ethoxyethene, 2-methoxypropene). This
means additional polar effects, and the comparison of Fig. 7 with Fig. 6 suggests that the
general addition behavior can be termed ambiphilic. Yet, the polar effects are small and
1
decrease the activation energies by less than 10 kJ mol .
.
The corresponding plot of Ea vs. Hr for CF3COCH2 is shown in Fig. 8. Now the
deviations from the line expected for complete enthalpy control are larger, and they are
particularly large for electron-rich alkenes such as 2-methylpropene. If the Ph-
substituted alkenes are disregarded, since they should lead to weaker polar effects, the
Fig. 7. Activation energies Ea for the addition of the 2-(methoxycarbonyl)propan-2-yl radical to alkenes as a
function of the reaction enthalpy Hn. Solid line as expected for complete enthalpy control.

Helvetica Chimica Acta ± Vol. 83 (2000)
671
trend of the deviations follows that predicted in Fig. 6 for an electrophilic addition
behavior. This agrees with the large EA of the radical and is supported by a very
reasonable correlation of log (k) with the alkene IEs. Division of the activation
energies by the enthalpy term of Eqn. 6 gives the electrophilic factor Fe. This is plotted
in Fig. 9 against the energy gap between the ground-state (RA) and the charge-transfer
‡
configuration (R A ). Also shown are lines calculated with Eqn. 8 and appropriate
Coulomb and interaction terms, which accommodate the data and are given in Fig. 9. In
comparison with other radicals [4b], the parameters are reasonable, and, as expected,
the weaker polar effects for Ph-substituted alkenes manifest themselves in smaller
Coulomb and interaction terms.
Fig. 8. Activation energies Ea for the addition of the 3,3,3-trifluoroacetonyl radical to alkenes as a function of the
reaction enthalpy Hn. Solid line as expected for complete enthalpy control.
Finally, we have calculated the activation energies with Eqns. 6 and 7. For
.
CF3COCH2, the parameters Ce and ge were as given in Fig. 9, and the experimental
electron affinity was used. For the polar effects of 2-(methoxycarbonyl)propan-2-yl, we
used Cn ˆ 6 eV, gn ˆ 1.5 eV, and Ce ˆ 4.5 eV, ge ˆ 2 eV for the non-Ph-substituted, and
smaller values Cn ˆ 5.5 eV, gn ˆ 0.75 eV, and Ce ˆ 4 eV, ge ˆ 2 eV for the Ph-substituted
alkenes. These parameters are similar to those found for other radicals [4b], and the
comparison of the calculated and experimental activation energies in Fig. 10
demonstrates the equations, and the parameters describe the data within a maximum
1
deviation of ca. 5 kJ mol .

672
Helvetica Chimica Acta ± Vol. 83 (2000)
Fig. 9. Electrophilic polar factor Fe for the addition of the 3,3,3-trifluoroacetonyl radical to alkenes
In conclusion, the addition of 2-(alkoxycarbonyl)propan-2-yl radicals to alkenes is
mainly governed by the reaction enthalpy. It is comparatively slow because there is a
considerable radical stabilization, but there are also slight ambiphilic polar effects. The
.
CF3COCH2 radical adds to alkenes rather fast, because there is little radical
stabilization, i.e., the exothermicity is high, and, for many alkenes, the reactions are
additionally facilitated by strong electrophilic polar effects. With reasonable param-
eters, the predictive equations developed in an earlier analysis [4b] describe the
activation energies rather well.
3.4. 2-(Methoxycarbonyl)propan-2-yl and the Propagating Radical of Methyl
.
Methacrylate. The 2-(methoxycarbonyl)propan-2-yl (Me2CCO2Me) radical is structur-
.
ally related to the propagating radical of methylmethacrylate (RCH2C(CH3)CO2CH3,
R ˆ polymer chain), and it is expected that the reactivities of these two radicals are
related. Recently, the homopropagation rate constant of methyl methacrylate has been
1
critically assessed [23]. Its temperature dependence is governed by log(A/m 1 s ) ˆ
1
1
6.43 and Ea ˆ 22.36 kJ mol and leads to a rate constant k ˆ 287 m 1 s at 294 K. 2-
(Methoxycarbonyl)propan-2-yl adds to methyl methacrylate with log(A/m 1 s ) ˆ 7.5,
1
Ea ˆ 22.4 kJ mol 1 and k294 ˆ 3710 m 1 s 1 (Table 1). Obviously, the activation energies
of the two additions are very similar, but the frequency factor and the rate constant are
lower for the polymer radical, and this remains also valid when the uncertainties of our
1
activation parameters (2.8 kJ mol for Ea and 0.5 for log (A)) are considered. There is
evidence for polymeric radicals that the propagation constants decrease substantially
with increasing chain length [23]. This is confirmed by the present findings, and the

Helvetica Chimica Acta ± Vol. 83 (2000)
673
Fig. 10. Experimental vs. calculated activation energies for the addition of the 2-(methoxycarbonyl)propan-2-yl
radical (squares) and of the 3,3,3-trifluoroacetonyl radical (circles) to alkenes. Solid line represents Ea(exp) ˆ
Ea(calc).
similarity of the activation energies indicates moreover that the chain-length effect may
be primarily due to a decreasing frequency factor [24].
Further, one can compare the ratios of rate constants for the addition of the 2-
(methoxycarbonyl)propan-2-yl radical to different alkenes with the co-polymerization
parameters r1 of methyl methacrylate. These are defined as r1 ˆ kp/ka, where kp is the
rate constant for the addition of the propagating radical to methyl methacrylate, and ka
holds for the addition to another monomer. Many co-polymerization parameters of
methyl methacrylate are available [25], but, for several monomers, they show a
considerable spread, and reasonable averages had to be chosen for a comparison.
Fig. 11 shows a plot of log(r1) vs. relative rate constants of 2-(methoxycarbonyl)pro-
pan-2-yl (k(methacrylate))/k(alkene)) as calculated from the data of Table 1. There is
an excellent linear correlation which demonstrates that the additions of the two radicals
are governed by the same (enthalpic and polar) factors. 1,1-Diphenylethene is an
obvious exception, and a redetermination of the rather early value of r1 for this alkene
[26] seems now necessary.
We thank the Swiss National Science Foundation for financial support.

674
Helvetica Chimica Acta ± Vol. 83 (2000)
Fig. 11. Experimental co-polymerization parameters of methyl methacrylate vs. relative rate constants of the 2-
(methoxycarbonyl)propan-2-yl radical
REFERENCES
[1] a) B. Giese, ꢀRadicals in Organic Synthesis: Formation of Carbon-Carbon Bondsꢁ, Pergamon Press, Oxford,
1986; b) D. P. Curran, in ꢀComprehensive Organic Synthesisꢁ, Eds. B. M. Trost, I. M. Fleming, Oxford, 1991,
Vol. 4; c) J. Fossey, D. Lefort, J. Sorba, ꢀFree Radicals in Organic Chemistryꢁ, Wiley, New York, 1995; d) B.
Giese, Angew. Chem. 1985, 97, 555; Angew. Chem., Int. Ed. 1985, 24, 553; e) B. Giese, Angew. Chem. 1983,
95, 771; Angew. Chem., Int. Ed. 1983, 22, 573; f) F. Minisci, E. Vismara, F. Fontana, Heterocycles 1989, 28,
489; g) D. N. Curran, N. A. Porter, B. Giese, ꢀStereochemistry of Radical Reactionsꢁ, VCH, Weinheim-New
York, 1996; M. P. Sibi, N. A. Porter, Acc. Chem. Res. 1999, 32, 163.
[2] a) C. Walling, ꢀFree Radicals in Solutionꢁ, Wiley, New York, 1957, and refs. cit. therein; b) W. A. Pryor,
ꢀIntroduction to Free Radical Chemistryꢁ, Prentice-Hall, Englewood Cliffs, 1966; c) J. M. Tedder, J. C.
Walton, Acc. Chem. Res. 1976, 9, 183; Adv. Phys. Org. Chem. 1978, 18, 51; Tetrahedron 1980, 36, 701; d) J. M.
Tedder, Angew. Chem. 1982, 94, 433; Angew. Chem., Int. Ed. 1982, 21, 401.
[3] a) K. Münger, H. Fischer, Int. J. Chem. Kinetics 1985, 17, 809; b) H. Fischer, in ꢀSubstituent Effects in
Radical Chemistryꢁ; Eds. H. G. Viehe et al., Reidel, Dordrecht, 1986; c) F. Jent, H. Paul, E. Roduner, H.
Heming, H. Fischer, Int. J. Chem. Kinetics 1986, 18, 1113; d) I. Beranek, H. Fischer, in ꢀFree Radicals in
Â
Synthesis and Biologyꢁ, Ed. F. Minisci, Kluwer, Dordrecht, 1988; e) K. Heberger, M. Walbiner, H. Fischer,
Â
Angew. Chem., Int. Ed. 1992, 31, 635; f) K. Heberger, H. Fischer, Int. J. Chem. Kinetics 1993, 25, 249; g) M.
Walbiner, J. Q. Wu, H. Fischer, Helv. Chim. Acta 1995, 78, 910; h) J. Q. Wu, H. Fischer, Int. J. Chem.
Kinetics 1995, 27, 167; i) J. Q. Wu, I. Beranek, H. Fischer, Helv. Chim. Acta 1995, 78, 194; j) C. Blättler,
Ph. D. Thesis, University of Zürich, 1995; k) S. N. Batchelor, H. Fischer, J. Phys. Chem. 1996, 100, 9794;
l) H. Rubin, H. Fischer, Helv. Chim. Acta 1996, 79, 1670; m) T. Zytowski, H. Fischer, J. Am. Chem. Soc.
1996, 118, 437; n) R. Martschke, R. D. Farley, H. Fischer, Helv. Chim. Acta 1997, 80, 1363; o) T. Zytowski,
H. Fischer, J. Am. Chem. Soc. 1997, 119, 12869; p) A. Salikhow, Ph. D. Thesis, University of Zürich, 1997;
q) M. Weber, H. Fischer, Helv. Chim. Acta 1998, 81, 770; r) R. Martschke, Ph. D. Thesis, University of
Zürich, 1999.

Helvetica Chimica Acta ± Vol. 83 (2000)
675
[4] a) H. Fischer, in ꢀFree Radicals in Biology and Environmentꢁ, Ed. F. Minisci, Kluwer Acad. Publ.,
Dordrecht, 1997; b) H. Fischer, L. Radom, Angew. Chem., in preparation.
[5] a) S. S. Shaik, Prog. Phys. Org. Chem. 1985, 15, 197; b) S. S. Shaik, H. B. Schlegel, S. Wolfe, ꢀTheoretical
Aspects of Physical Organic Chemistry. The SN2 Mechanismꢁ, Wiley, New York, 1992; c) S. S.Shaik, A.
Shukri, Angew. Chem. 1999, 111, 616, and res. cit. therein; d) A. Pross, S. S. Shaik, Acc. Chem. Res. 1983, 16,
363; e) A. Pross, ꢀTheoretical and Physical Principles of Organic Reactivityꢁ, Wiley, New York, 1995; f) A.
Salikhov, H. Fischer, Theor. Chem. Acc. 1997, 96, 114.
[6] D. V. Avila, K. U. Ingold, J. Lusztyk, W. R. Dolbier, Jr., H. Q. Pan, J. Am. Chem. Soc. 1993, 115, 1577; D. V.
Avila, K. U. Ingold, J. Lusztyk, W. R. Dolbier, Jr., H. Q. Pan, M. Muir, J. Am. Chem. Soc. 1994, 116, 99;
D. V. Avila, K. U. Ingold, J. Lusztyk, W. R. Dolbier, Jr., H. Q. Pan, J. Org. Chem. 1996, 61, 2027; D. V. Avila,
K. U. Ingold, J. Lusztyk, W. R. Dolbier, Jr., H. Q. Pan, Tetrahedron 1996, 52, 12351; M. D. Bartberger,
W. R. Dolbier, Jr., J. Lusztyk, K. U. Ingold, Tetrahedron 1997, 53, 9857.
[7] K. Riemenschneider, H. M. Bartels, R. Dornow, E. Drechsel-Grau, W. Eichel, H. Luthe, Y. M. Matter,
Y. W. Michaelis, P. Boldt, J. Org. Chem. 1987, 52, 205.
[8] ꢀOrganic Syntheses: Collective Vol. 3ꢁ, Ed. E. C. Horning, J. Wiley, N. Y., 1967, p. 141.
[9] J. E. Bennet, B. C. Gilbert, S. Lawrence, A. C. Whitwood, A. J. Holmes, J. Chem. Soc., Perkin Trans. 2 1996,
1789.
[10] Wu Lung-min, H. Fischer, Helv. Chim. Acta 1983, 66, 138.
[11] Landolt-Börnstein, ꢀMagnetic Properties of Free Radicalsꢁ, Ed. H. Fischer, New Series, Volumes II/1, II/9b,
II/17b, Springer, Berlin, 1965 ± 1986.
[12] M. Weber, H. Fischer, J. Am. Chem. Soc. 1999, 121, 7381.
[13] a) J. Chateauneuf, J. Lutzyk, K. U. Ingold, J. Am. Chem. Soc. 1988, 110, 2886; b) S. Yamauchi, N. Hirota, S.
Takahara, H. Sakuragi, K. Tokumaru, J. Am. Chem. Soc. 1985, 107, 5021; c) H. Misawa, K. Sawabe, S.
Takahara, H. Sakuragi, K. Tokumaru, Chem. Lett. 1988, 357.
[14] A. Faucitano, A. Buttafava, F. Faucitano Martinotti, J. Chem. Soc., Perkin Trans. 2 1981, 425.
[15] H.-H. Schuh, H. Fischer, Helv. Chim. Acta 1978, 61, 2130.
[16] B. C. Gilbert, J. R. Lindsay Smith, E. C. Milne, A. C. Whitwood, P. Taylor, J. Chem. Soc., Perkin Trans. 2
1993, 2025.
[17] M. W. Wong, A. Pross, L. Radom, J. Am. Chem. Soc. 1993, 115, 11050; Israel J. Chem. 1993, 33, 415; J. Am.
Chem. Soc. 1994, 116, 6284; J. Am. Chem. Soc. 1994, 116, 11938; M. W. Wong, L. Radom, J. Phys. Chem.
1995, 99, 8582; C. Gonzalez, C. Sosa, H. B. Schlegel, J. Phys. Chem. 1989, 93, 2435; R. Arnaud, D. Subra, V.
Â
Barone, F. Lelj, S. Olivella, A. Sole, N. Russo, J. Chem. Soc., Perkin Trans. 2, 1986, 1517; T. Fueno, M.
Kamachi, Macromolecules 1988, 21, 908; H. Zipse, J. He, K. N. Houk, B. Giese, J. Am. Chem. Soc. 1991, 113,
4324.
[18] a) S. G. Lias, J. E. Bartmess, J. F. Liebman, J. L. Holmes, R. D. Levin, W. G. Mallard, J. Phys. Chem. Ref.
Data 1988, 17, Suppl. 1; b) N. Cohen, J. Phys. Chem. Ref. Data 1996, 25, 1411.
[19] J. Berkowitz, G. B. Ellison, D. Gutman, J. Phys. Chem. 1994, 98, 2744.
[20] J. B. Cumming, P. Kebarle, Can. J. Chem. 1978, 56, 1.
[21] L. Radom, A. P. Scott, private communication.
[22] E. A. Brinkman, S. Berger, J. Marks, J. I. Brauman, J. Chem. Phys. 1993, 99, 7586.
[23] S. Beuermann, M. Buback, T. P. Davis, R. G. Gilbert, R. A. Hutchinson, O. F. Olaj, G. T. Russell, J.
Schweer, A. M. Van Herk, Macromol. Chem. Phys. 1997, 198, 1545.
[24] J. P. A. Heuts, R. G. Gilbert, L. Radom, Macromolecules 1995, 8771.
[25] R. Z. Greenley in ꢀPolymer Handbookꢁ, Eds. J. Brandrup, E. H. Immergut, 3rd edn. Wiley, New York, 1989.
[26] B. G. Elgood, B. J. Sauntson, Chem. Ind (London) 1965, 1558.
Received December 10, 1999