The effect of ring substitution on the O-neophyl rearrangement of 1,1-diarylalkoxyl radicals. A product and time-resolved kinetic study

Article abstract of DOI:10.1021/jo0502448

A product and time-resolved kinetic study of the effect of ring substitution on the reactivity of 1,1-diarylalkoxyl radicals has been carried out. The radicals undergo an O-neophyl shift to give the isomeric 1-aryl-1-aryloxyalkyl radicals from which the corresponding aromatic ketones are formed. The rearrangement rate constants are influenced by ring substitution, increasing in the presence of electron-withdrawing substituents and decreasing in the presence of electron-donating ones. From the results of product and kinetic studies, the following migratory aptitudes have been obtained: 4-trifluoromethylphenyl > phenyl ? 4-methylphenyl > 4-methoxyphenyl. Excellent Hammett-type correlations between the σ⁺ substituent constants and both the visible absorption band maxima and the rearrangement rate constants have been obtained. The experimental results indicate that the rearrangement is governed by electronic effects in the starting 1,1-diarylalkoxyl radicals, whereas the stability of the rearranged carbon-centered radical plays a minor role, in line with a reactant-like transition state, strongly supporting the hypothesis that the O-neophyl rearrangement of 1,1-diarylalkoxyl radicals proceeds through a concerted mechanism.

Full text of DOI:10.1021/jo0502448

The Effect of Ring Substitution on the O-Neophyl Rearrangement
of 1,1-Diarylalkoxyl Radicals. A Product and Time-Resolved
Kinetic Study
Carla S. Aureliano Antunes,1a Massimo Bietti,*,1a Gianfranco Ercolani,1a
Osvaldo Lanzalunga,1b and Michela Salamone1a
Dipartimento di Scienze e Tecnologie Chimiche, Universita` “Tor Vergata”, Via della Ricerca Scientifica,
1 I-00133 Rome, Italy, and Dipartimento di Chimica, Universita` “La Sapienza”, P.le A. Moro,
5 I-00185 Rome, Italy
bietti@uniroma2.it
Received February 7, 2005
A product and time-resolved kinetic study of the effect of ring substitution on the reactivity of
1,1-diarylalkoxyl radicals has been carried out. The radicals undergo an O-neophyl shift to give
the isomeric 1-aryl-1-aryloxyalkyl radicals from which the corresponding aromatic ketones are
formed. The rearrangement rate constants are influenced by ring substitution, increasing in the
presence of electron-withdrawing substituents and decreasing in the presence of electron-donating
ones. From the results of product and kinetic studies, the following migratory aptitudes have been
obtained: 4-trifluoromethylphenyl > phenyl = 4-methylphenyl > 4-methoxyphenyl. Excellent
Hammett-type correlations between the σ+ substituent constants and both the visible absorption
band maxima and the rearrangement rate constants have been obtained. The experimental results
indicate that the rearrangement is governed by electronic effects in the starting 1,1-diarylalkoxyl
radicals, whereas the stability of the rearranged carbon-centered radical plays a minor role, in
line with a reactant-like transition state, strongly supporting the hypothesis that the O-neophyl
rearrangement of 1,1-diarylalkoxyl radicals proceeds through a concerted mechanism.
SCHEME 1
The 1,2-migration of an aryl group in free radicals,
known as the neophyl rearrangement, has been the
subject of several studies.2 This process is believed to
occur via a spiro[2,5]octadienyl radical intermediate as
shown in Scheme 1. However, despite these studies the
mechanism of the neophyl rearrangement is still not
entirely understood, the main question being whether the
bridged radical is an intermediate or a transition state
in this process.2
radical is a true intermediate in the neophyl-like rear-
rangement of the 1,1-diphenylethoxyl radical.5 The decay
kinetics of the 1,1-diphenylethoxyl radical were found to
match the formation kinetics of the rearranged 1-phe-
noxy-1-phenylethyl radical, leading to the suggestion that
the lifetime of the bridged radical intermediate must be
significantly smaller than 100 ns.
What concerns alkoxyl radicals instead is an analogous
O-neophyl (or neophyl-like) rearrangement leading to an
isomeric carbon-centered radical that has been observed
only for radicals bearing at least two aryl substituents,
such as the triphenylmethoxyl and 1,1-diphenylethoxyl
radicals.2-4 By analogy, the rearrangement is believed
to occur via a 1-oxaspiro[2,5]octadienyl radical (Scheme 2).
Recently, on the basis of the results of laser flash
photolysis experiments and AM1-UHF calculations, Banks
and Scaiano proposed that the 1-oxaspiro[2,5]octadienyl
More recently, Grossi and Strazzari provided, via EPR
spectroscopy, evidence for the involvement of a 1-oxaspiro-
(4) For alkoxyl radicals bearing a single phenyl substituent such
as the benzyloxyl and cumyloxyl radicals, aryl migration would lead
to carbon-centered radicals that are significantly less stable than those
formed from both the 1,1-diphenylethoxyl and triphenylmethoxyl
radicals. Accordingly, with the former radicals the 1,2-phenyl shift does
not compete with hydrogen atom abstraction, and â-fragmentation
reactions.
(1) (a) Universita` “Tor Vergata”. (b) Universita` “La Sapienza”.
(2) For an excellent review on radical aryl migration reactions, see:
Studer, A.; Bossart, M. Tetrahedron 2001, 57, 9649-9667.
(3) Falvey, D. E.; Khambatta, B. S.; Schuster, G. B. J. Phys. Chem.
1990, 94, 1056-1059.
(5) Banks, J. T.; Scaiano, J. C. J. Phys. Chem. 1995, 99, 3527-3531.
10.1021/jo0502448 CCC: $30.25 © 2005 American Chemical Society
Published on Web 04/12/2005
3884
J. Org. Chem. 2005, 70, 3884-3891

Effect of Ring Substitution on O-Neophyl Rearrangement
SCHEME 2
[2,5]octadienyl radical as an intermediate in the rear-
rangement of both the 1,1-diphenylethoxyl and the
triphenylmethoxyl radicals.6 The alkoxyl radicals were
generated by photolysis of the parent alcohols in the
presence of ceric ammonium nitrate (CAN). Quite inter-
estingly, in the same study no evidence was obtained for
the formation of a bridged radical but only of a peroxyl
radical when the 1,1-diphenylethoxyl radical was gener-
ated by photolysis of the parent 1,1-diphenylethyl per-
oxide.7 It was proposed that CAN plays a key role in the
formation and stabilization of the bridged radical inter-
mediate: the coordination of the alcoholic oxygen to the
ceric ion increases the electrophilicity of this group
favoring the cyclization process. This is an observation
that, however, raises some doubt on the involvement of
free alkoxyl radicals in the CAN-induced photooxidation
of arylcarbinols, indicating that the mechanism of the
neophyl-like rearrangement of alkoxyl radicals is also
still not entirely understood and that additional informa-
tion in this respect is certainly needed.
In this context, product studies have also shown that
in the rearrangement of triarylmethoxyl radicals the
product distributions can be influenced by the presence
of ring substituents. Thus, products deriving from the
migration of both the 4-nitrophenyl and 4-benzoylphenyl
groups in (4-nitrophenyl)diphenylmethoxyl and (4-ben-
zoylphenyl)diphenylmethoxyl radicals, respectively, were
formed in greater amounts than those deriving from the
migration of the unsubstituted phenyl group;8-10 whereas
the 4-methylphenyl group showed a similar migratory
aptitude as that of the phenyl group.11 These results have
been interpreted in terms of the activation toward
intramolecular free radical attack exerted by the presence
of ring substituents. A similar activation was predicted
in the presence of methoxy ring substituents,8,12 but
unfortunately no clear information in this respect is
available.9 These findings also suggest that the rates of
1,2-aryl shift are expected to be influenced by ring
substitution. Along these lines, it seemed particularly
interesting to study the effect of both electron-withdraw-
ing and electron-releasing ring substituents on the
reactivity of 1,1-diarylalkoxyl radicals. For this purpose,
we have carried out product and time-resolved kinetic
studies on the reactivity of 1,1-diarylethoxyl radicals
1•-3• and cyclopropyldiarylmethoxyl radicals 4•-8• (Chart
1). The use of the cyclopropyl derivatives is a consequence
CHART 1
of the instability of both 1-(4-methoxyphenyl)-1-phenyle-
thanol and 1,1-di(4-methoxyphenyl)ethanol, which un-
dergo water elimination very easily.13 Cyclopropyl(4-
methoxyphenyl)phenylmethanol (5a) and cyclopropyl[bis(4-
methoxyphenyl)]methanol (6a) (precursors of radicals 5•
and 6•) are instead stable under the experimental condi-
tions employed.
As a matter of completeness, product studies have been
carried out for all the 1,1-diarylalkanols (1a-8a: precur-
sors of the 1,1-diarylalkoxyl radicals 1•-8•), even though
the reactions of symmetric substrates 1a, 3a, 4a, 6a, and
8a do not provide information on the effect of ring
substituents.
Results
Product Studies. The 1,1-diarylalkoxyl radicals 1•-
8• have been generated photochemically by visible light
irradiation of CH2Cl2 solutions containing the parent 1,1-
diarylalkanol (1a-8a), (diacetoxy)iodobenzene (DIB) and
I2. It is well established that under these conditions the
DIB/I2 reagent converts alcohols into intermediate hy-
poiodites,14 which are then photolyzed to give alkoxyl
radicals, precursors of the observed reaction products.15-17
With 1,1-diarylalkanols, the corresponding alkoxyl radi-
cals so formed are expected to rearrange to give isomeric
carbon-centered radicals, from which, after efficient
iodine trapping,18,19 the product alkyl aryl ketones are
formed (Scheme 3).
(6) Grossi, L.; Strazzari, S. J. Org. Chem. 2000, 65, 2748-2754.
(7) In this EPR experiment, carried out in N2-saturated MeCN
solution at 230 K, the peroxyl radical is described to be formed by
oxygen quenching of either the bridged radical or the rearranged
1-phenoxy-1-phenylethyl radical. However, due to the very short
lifetime suggested for the former radical (see ref 5), quenching of this
radical by traces of oxygen is expected to be slower than the formation
of the 1-phenoxy-1-phenylethyl radical. As a consequence, the peroxyl
radical (if formed) is likely to derive from the reaction of this radical
with oxygen.
SCHEME 3
(8) Bartlett, P. D.; Cotman, J. D., Jr. J. Am. Chem. Soc. 1950, 72,
3095-3099.
(9) (a) Starnes, W. H., Jr. J. Am. Chem. Soc. 1968, 90, 1807-1815.
(b) Starnes, W. H., Jr. J. Am. Chem. Soc. 1967, 89, 3368-3369.
(10) Neckers, D. C.; Linden, S.-M.; Williams, B. L.; Zakrzewski, A.
J. Org. Chem. 1989, 54, 131-135.
(11) Kharasch, M. S.; Poshkus, A. C.; Fono, A.; Nudenberg, W. J.
Org. Chem. 1951, 16, 1458-1470.
(12) Kamata, M.; Komatsu, K. Tetrahedron Lett. 2001, 42, 9027-
9030.
J. Org. Chem, Vol. 70, No. 10, 2005 3885

Aureliano Antunes et al.
TABLE 1. Yields and Product Distributions (Q)a
Observed after Irradiation of Argon-Saturated CH2Cl2
Solutions Containing 1,1-Diarylalkanols 1a-8a, DIB, and
SCHEME 4
b
I2
were thus employed for the reactions of 1,1-diarylalkanols
2a-8a. The product yields observed after irradiation of
argon-saturated CH2Cl2 solutions containing substrates
1a-8a, DIB, and I2 are collected in Table 1.
In the same table, the product distributions, expressed
in terms of the molar ratios (Q) of ring-substituted and
unsubstituted acetophenones or aryl cyclopropyl ketones
observed in the reactions of nonsymmetric 1,1-diarylal-
kanols 2a, 5a, and 7a, are also reported.
With substrates 1a, 3a, 4a, 6a, and 8a, exclusive
formation of the corresponding acetophenone or aryl
cyclopropyl ketone was observed. The reactions of non-
symmetric 1,1-diarylalkanols 2a, 5a, and 7a led instead
to the formation of both the unsubstituted and ring-
substituted ketone. In particular, the reaction of 2a led
to the formation of comparable amounts of 4-methylac-
etophenone and acetophenone (Q ) 1.05). Cyclopropyl
4-methoxyphenyl ketone was instead the major reaction
product observed in the reaction of 5a accompanied by
cyclopropyl phenyl ketone (Q ) 1.64), whereas with 7a
cyclopropyl phenyl ketone was the major reaction product
accompanied by cyclopropyl 4-trifluoromethylphenyl ke-
tone (Q ) 0.54).
Time-Resolved Studies. The 1,1-diarylalkoxyl radi-
cals 1•-8• have been generated in MeCN by 266-nm laser
flash photolysis (LFP) of the parent tert-butyl peroxides
1p-8p (Scheme 4), which in turn have been synthesized
by reaction of 1,1-diarylalkanols 1a-8a with tert-butyl
hydroperoxide in the presence of p-toluenesulfonic acid
according to a slight modification of a previously de-
scribed procedure (see Supporting Information).21,22
All the alkoxyl radicals display a broad absorption
band in the visible region of the spectrum, which is
unaffected by oxygen and whose position and intensity
is influenced by the presence of ring substituents falling
between 490 and 640 nm, in line with previous observa-
tions on the spectral properties of ring-substituted aryl-
carbinyloxyl radicals.22-24 The visible absorption band
was observed to decay in the nanosecond to microsecond
a Molar ratio between ring-substituted and unsubstituted alkyl
aryl ketone. b [DIB] ) 22 mM, [I2] ) 10 mM, [substrate] ) 10 mM,
irradiation wavelength ) 480 nm, irradiation time ) 15 min, T )
20 °C. c No significant variation of Q was observed after 30 min
of irradiation. d Irradiation time ) 10 min. e Irradiation time ) 5
min. At longer irradiation times, formation of an unidentified
product was observed.
Argon-saturated CH2Cl2 solutions containing the 1,1-
diarylalkanol (1a-8a) (10 mM), DIB (11-40 mM), and
I2 (10-40 mM) were irradiated with visible light (λmax
≈
480 nm) at T ) 20 °C for 15 min. The irradiation time
was chosen in such a way as to avoid complete substrate
conversion. After workup, the reaction products were
identified by GC (comparison with authentic samples)
and GC-MS and quantitatively determined, together
with the unreacted substrate, by GC using bibenzyl as
internal standard. The reaction of 1,1-diphenylethanol
(1a) was carried out under a variety of experimental
conditions (see Supporting Information), leading in all
cases to the formation of acetophenone as the exclusive
product.20 Convenient experimental conditions were found
using a 2.2:1:1 DIB/I2/1a ratio, and analogous conditions
(13) Grummitt, O.; Marsh, D. J. Am. Chem. Soc. 1948, 70, 1289-1290.
(14) Courtneidge, J. L.; Lusztyk, J.; Page`, D. Tetrahedron Lett. 1994,
35, 1003-1006.
(15) Aureliano Antunes, C. S.; Bietti, M.; Lanzalunga, O.; Salamone,
M. J. Org. Chem. 2004, 69, 5281-5289.
(16) Sua´rez, E.; Rodriguez, M. S. In Radicals in Organic Synthesis;
Renaud, P., Sibi, M. P., Eds.; Wiley-VCH: Weinheim, Germany, 2001;
Vol. 2, pp 440-454.
(17) Togo, H.; Katohgi, M. Synlett 2001, 5, 565-581.
(18) Minisci, F.; Recupero, F.; Gambarotti, C.; Punta, C.; Paganelli,
R. Tetrahedron Lett. 2003, 44, 6919-6922.
(19) Hook, S. C. W.; Saville, B. J. Chem. Soc., Perkin Trans. 2 1975,
589-593.
(20) No evidence for the formation of phenols was obtained in the
product analysis experiments.
3886 J. Org. Chem., Vol. 70, No. 10, 2005

Effect of Ring Substitution on O-Neophyl Rearrangement
TABLE 2. Visible Absorption Band Maximum
Wavelengths and Decay Rate Constants for
1,1-Diarylalkoxyl Radicals 1•-8• Measured in
Argon-Saturated MeCN Solution at T ) 22 °C
radical
λmax (vis)/nma
k/s-1 b,c
1•
2•
3•
4•
5•
6•
7•
8•
535
540
550
535
610
640
520
490
2.8 × 106
2.4 × 106
2.0 × 106
2.0 × 106
5.1 × 105
2.6 × 105
5.0 × 106
1.8 × 107
a Visible absorption band maximum of the 1,1-diarylalkoxyl
radical. b Measured following the decay of the 1,1-diarylalkoxyl
radical at the visible absorption maximum. c The decay rate
constants were not affected by the presence of oxygen.
FIGURE 1. Time-resolved absorption spectra observed after
266-nm LFP of 3p (1.8 mM) in an argon-saturated MeCN
solution at 55 ns (b), 125 ns (O), 300 ns (9), and 1.4 µs (0)
after the 8-ns, 10 mJ laser pulse.
7•. This species undergoes a first-order decay (inset a),
accompanied by a corresponding buildup of optical den-
sity at 280 nm (an isosbestic point is visible at 380 nm)
whose decay is significantly accelerated in the presence
of oxygen (inset b), which again is assigned to the
rearranged carbon-centered radical(s).
The decay of the 1,1-diarylethoxyl radicals (1•-3•) and
cyclopropyldiarylmethoxyl radicals (4•-8•) was measured
spectrophotometrically by monitoring the decrease in
optical density at the corresponding visible absorption
maxima and was found to follow first-order kinetics. The
corresponding rate constants thus obtained are reported
in Table 2 together with the visible absorption band
maximum wavelengths of the 1,1-diarylalkoxyl radicals
1•-8•.
Discussion
The results of product studies collected in Table 1 show
that the reactions of 1,1-diarylalkoxyl radicals lead to the
exclusive formation of acetophenones or aryl cyclopropyl
ketones. In particular, with nonsymmetric 1,1-diarylal-
kanols 2a, 5a, and 7a, formation of both the unsubsti-
tuted and ring-substituted ketone was observed. Forma-
tion of these products can be rationalized in terms of the
competition between the two possible 1,2-aryl shifts in
the intermediate 1,1-diarylalkoxyl radicals (paths a and
b in Scheme 5: R ) Me, cyclopropyl).
Quite interestingly, when R ) cyclopropyl (substrates
4a-8a) the 1,2-aryl shift leads to the formation of a 1,1-
disubstituted cyclopropylcarbinyl radical. The exclusive
formation of aryl cyclopropyl ketones in the reactions of
these substrates clearly indicates that iodine trapping
of the rearranged radical occurs significantly faster than
cyclopropyl ring opening, in agreement with the relatively
low rate constants measured for ring opening of R-phenyl-
substituted cyclopropylcarbinyl radicals.25 Moreover, in
the reactions of substrates 4a-8a no â-scission of the
C-cyclopropyl bond in the corresponding alkoxyl radicals
FIGURE 2. Time-resolved absorption spectra observed after
266-nm LFP of 7p (1.67 mM) in an argon-saturated MeCN
solution at 102 (b), 170 (O), 250 (9), and 430 ns (0) after the
8-ns, 10 mJ laser pulse. Insets: (a) First-order decay of radical
7• monitored at 520 nm and (b) corresponding first-order
buildup of absorption at 280 nm, measured under argon and
oxygen.
time scale depending on ring substitution, leading in all
cases to a corresponding strong increase in absorption
in the UV region of the spectrum, assigned to the
rearranged carbon-centered radical formed by aryl shift
in the 1,1-diarylalkoxyl radical.3,5,24
As an example, Figure 1 displays the time-resolved
spectra observed after LFP of an argon-saturated MeCN
solution containing peroxide 3p, showing after 55 ns
(filled circles) a broad absorption band centered at 550
nm, which is assigned to the radical 3•.
This species undergoes a first-order decay that is
accompanied by a corresponding buildup of optical den-
sity at 260 and 320 nm assigned to the 1-(4-methylphe-
noxy)-1-(4-methylphenyl)ethyl radical (an isosbestic point
is visible at 345 nm). As expected for a carbon-centered
radical, a rapid decay of the 260- and 320-nm bands was
observed in the presence of oxygen.
Figure 2 displays instead the time-resolved absorption
spectra observed after LFP of an argon-saturated MeCN
solution containing peroxide 7p. The spectrum recorded
after 102 ns (filled circles) shows a very weak absorption
band centered at 520 nm, which is assigned to the radical
(21) Hendrickson, W. H.; Nguyen, C. C.; Nguyen, J. T.; Simons, K.
T. Tetrahedron Lett. 1995, 36, 7217-7220.
(22) Baciocchi, E.; Bietti, M.; Salamone, M.; Steenken, S. J. Org.
Chem. 2002, 67, 2266-2270.
(23) Avila, D. V.; Lusztyk, J.; Ingold, K. U. J. Am. Chem. Soc. 1992,
114, 6576-6577.
(24) Avila, D. V.; Ingold, K. U.; Di Nardo, A. A.; Zerbetto, F.; Zgierski,
M. Z.; Lusztyk, J. J. Am. Chem. Soc. 1995, 117, 2711-2718.
(25) Halgren, T. A.; Roberts, J. D.; Horner, J. H.; Martinez, F. N.;
Tronche, C.; Newcomb, M. J. Am. Chem. Soc. 2000, 122, 2988-2994.
J. Org. Chem, Vol. 70, No. 10, 2005 3887

Aureliano Antunes et al.
SCHEME 5
4•-8• occurs, as shown by the absence of diaryl ketones
among the reaction products. This result is in line with
the relatively low stability of the cyclopropyl radical.26-28
step where radical-stabilizing effects play a minor role.
In other words, the transition state must be characterized
by a small degree of carbon radical character.
The spectral data collected in Table 2 confirm that the
position of the visible absorption band of arylcarbinyloxyl
radicals is red-shifted by electron-donating ring substit-
uents and blue-shifted by electron-withdrawing ones, as
previously shown for ring-substituted cumyloxyl radi-
cals.22-24 An excellent Hammett-type correlation between
the experimentally measured excitation energies corre-
sponding to the visible absorption maxima of cyclopropyl-
diarylmethoxyl radicals 4•, 6•, and 8• and the σ+ substitu-
ent constants is obtained (Figure 3), in line with that ob-
served previously for ring-substituted cumyloxyl radicals.22
The results collected in Table 1 show that while the
4-methylphenyl and phenyl groups are characterized by
a similar migratory aptitude (Q ) 1.05), in agreement
with the results of previous studies,11 opposite behaviors
are instead observed in the presence of methoxy (5a) or
trifluoromethyl (7a) ring substituents: the phenyl group
migrates better than the 4-methoxyphenyl one (Q ) 1.64)
while 4-trifluoromethylphenyl migrates better than phen-
yl (Q ) 0.54). This behavior is in contrast with the
predicted activation toward intramolecular free radical
attack exerted by the presence of both electron-with-
drawing and electron-releasing ring substituents.8,12
A possible explanation for this behavior may be that
the reaction is influenced by the stability of the rear-
ranged carbon-centered radical. Along this line, it is well-
known that a methoxy substituent stabilizes carbon-cen-
tered radicals, whereas the presence of a trifluoromethyl
one determines a slight destabilization, as clearly shown
for example for benzyl and methyl radicals.29-33 Accord-
ingly, the product distributions observed in the reactions
of radicals 5a and 7a would indicate in both cases
preferential formation of the most stable carbon-centered
radical: the (4-methoxyphenyl)phenoxycyclopropylcarbinyl
radical over the (4-methoxyphenoxy)phenylcyclopro-
pylcarbinyl one from 5a, and the phenyl(4-trifluorom-
ethylphenoxy)cyclopropylcarbinyl radical over the phenoxy-
(4-trifluoromethylphenyl)cyclopropylcarbinyl one from
7a. However, the experimental rate constants displayed
in Table 2 showing that the trifluoromethyl substituted
radicals 7• and 8• are significantly more reactive than the
methoxy substituted ones 5• and 6•, and in particular that
8• is more reactive than 7• whereas 6• is less reactive than
5•, are in sharp contrast with this hypothesis. From these
resuts, it can be reasonably concluded that, in the
O-neophyl rearrangement of 1,1-diarylalkoxyl radicals,
the stability of the rearranged carbon-centered radical
is not important, an observation that points toward a
transition state of the rearrangement rate-determining
FIGURE 3. Hammett-type correlation between the experi-
mentally measured energies at the visible absorption maxima
of radicals 4•, 6•, and 8• and σ+ substituent constants. From
the linear regression: slope ) 0.45 eV-1, r2 ) 0.9966.
In this connection, it has been recently suggested that
arylcarbinyloxyl radicals are characterized by a certain
degree of internal charge transfer, with the aromatic ring
bearing a partial positive charge and the oxygen atom a
partial negative one;23 the visible absorption band of
arylcarbinyloxyl radicals has been attributed to a π f
π* transition, which determines an increase in the extent
of charge separation in the radical on going from the
ground state to the excited state.22,23
This charge separation may be reasonably rationalized
in terms of the contribution of negative hyperconjugation
structures, such as those shown in Scheme 6, to the
stability of the starting arylcarbinyloxyl radical.
(26) Heinrich, M. R.; Zard, S. Z. Org. Lett. 2004, 6, 4969-4972.
(27) Luo, Y.-R. Handbook of Bond Dissociation Energies in Organic
Compounds; CRC Press: Boca Raton, FL, 2003.
(28) Walborsky, H. M. Tetrahedron 1981, 37, 1625-1651.
(29) Pratt, D. A.; DiLabio, G. A.; Mulder, P.; Ingold, K. U. Acc. Chem.
Res. 2004, 37, 334-340.
(30) Wen, Z.; Li, Z.; Shang, Z.; Cheng, J.-P. J. Org. Chem. 2001, 66,
1466-1472.
What concerns the kinetic data instead is the rate con-
stant measured for the decay of the 1,1-diphenylethoxyl
radical (1•) to give the 1-phenoxy-1-phenylethyl radical,
k ) 2.8 × 106 s-1, which is very similar to those measured
previously for the same radical.3,5,24 The observation that
(31) Wu, Y.-D.; Wong, C.-L.; Chan, K. W. K.; Ji, G.-Z.; Jiang, X.-K.
J. Org. Chem. 1996, 61, 746-750.
(32) Jackson, R. A.; Sharifi, M. J. Chem. Soc., Perkin Trans. 2 1996,
775-778.
(33) Henry, D. J.; Parkinson, C. J.; Mayer, P. M.; Radom, L. J. Phys.
Chem. A 2001, 105, 6750-6756.
3888 J. Org. Chem., Vol. 70, No. 10, 2005

Effect of Ring Substitution on O-Neophyl Rearrangement
SCHEME 6
the rate constant measured for decay of the cyclopropyldi-
phenylmethoxyl radical (4•), k ) 2.0 × 106 s-1, is com-
parable with that measured for 1•, indicates that replace-
ment of the methyl group with a cyclopropyl one has only
a negligible effect on the reactivity of these radicals.
In contrast with the results obtained for the â-scission
of ring-substituted cumyloxyl radicals, where the rate
constants were found to be essentially independent from
the nature of the substituent,22 with 1,1-diarylalkoxyl
radicals the decay rate constants are significantly influ-
enced by ring substitution. Accordingly, as compared to
the unsubstituted radicals 1• and 4•, an increase in rate
is observed in the presence of electron-withdrawing
substituents while the presence of electron-donating ones
determines a rate decrease, with radical 1• reacting at a
slightly higher rate than 3• and radicals 6•, 4•, and 8•
displaying the following relative reactivities: 0.13:1:9.
This reactivity order is in agreement with the results of
product studies described above for the nonsymmetric
substrates 2a, 5a, and 7a and from which the following
migratory aptitudes can be derived: 4-trifluorometh-
ylphenyl > phenyl = 4-methylphenyl > 4-methoxyphenyl.
Quite importantly, a good Hammett-type correlation
is also obtained between the experimental rate constants
measured for the rearrangement of cyclopropyldiaryl-
methoxyl radicals 4•, 6•, and 8• and the σ+ substituent
constants (Figure 4), in line with that observed for the
visible absorption maxima of the same radicals shown
in Figure 3.
This observation provides additional support to the
hypothesis of the existence of an internal charge transfer
in the ground state of 1,1-diarylalkoxyl radicals, indicat-
ing moreover that efficient stabilization of the partial
positive charge is also responsible for the effect of ring
substitution on the rearrangement rate constants. In
particular, the observation of a positive slope is in line
with a reaction that involves a decrease in the extent of
positive charge on the aromatic ring on going from the
starting 1,1-diarylalkoxyl radical to the transition state.
By combining the Q ratios measured for the reactions
of 1,1-diarylalkanols 5a and 7a reported in Table 1 with
the decay rate constants measured for cyclopropyldiaryl-
methoxyl radicals 4•, 5•, and 7• reported in Table 2, we
can obtain the partial rate constants for migration of the
unsubstituted (kH) and 4-X substituted ring (kX: X ) H,
OMe, CF3). These data are collected in Table 3.
FIGURE 4. Hammett-type correlation between the experi-
mental rate constants measured for the rearrangement of
cyclopropyldiarylmethoxyl radicals 4•, 6•, and 8• and the σ+
substituent constants. From the regression analysis: F ) 1.38;
r2 ) 0.984.
TABLE 3. Partial Rate Constants for the Migration of
the Unsubstituted and 4-X Substituted Ring in
Cyclopropyldiarylmethoxyl Radicals 4•, 5•, and 7•
radical
Qa
k/s-1 b
kH/s-1 c
kX/s-1 d
4•
5•
7•
-
1.64
0.54
2.0 × 106
5.1 × 105
5.0 × 106
1.0 × 106
3.2 × 105
1.75 × 106
1.0 × 106
1.9 × 105
3.25 × 106
a Molar ratio between ring-substituted and unsubstituted aryl
cyclopropyl ketone (see Table 1). b Measured following the decay
of the 1,1-diarylalkoxyl radical at the visible absorption maximum
(see Table 2). c Rate constant for migration of the unsubstituted
ring. d Rate constant for migration of the 4-X (X ) H, OMe, CF3)
substituted ring.
stituted and unsubstituted rings in cyclopropyldiaryl-
methoxyl radicals 4•, 5•, and 7• and σ+ substituent con-
stants are obtained (see Figure S1 in the Supporting In-
formation, plots a and b, respectively). The small F values
(0.93 and 0.56, respectively) are in agreement with reac-
tions involving little charge separation. Comparison be-
tween these values indicates, as expected, that migration
of the 4-X substituted ring is more sensitive to substitu-
ent effects than migration of the unsubstituted one.
Quite interestingly, Beckwith observed that the neo-
phyl rearrangement of alkenylaryl radicals is favored by
the presence of electron-withdrawing substituents on the
migrating aromatic ring,34 a behavior that is similar to
that observed in the present case and that was rational-
ized in terms of the nucleophilic character of alkyl
radicals. However in this study, rather than simply
invoking an unexpected nucleophilic character for the 1,1-
diarylalkoxyl radicals, the experimental results discussed
above seem more in line with the hypothesis that
stabilizing effects on the starting 1,1-diarylalkoxyl radi-
cal, expressed in terms of the negative hyperconjugation
contributions described in Scheme 6, play an important
role.35 According to this picture, in a nonsymmetric 1,1-
diarylalkoxyl radical, the oxygen radical will preferen-
tially attack the aromatic ring that is less involved in
the resonance stabilization, that is, the ring characterized
by the smallest extent of hyperconjugation: a hypothesis
The partial rate constants indicate that the presence
of the ring substituent influences the migration rate of
both aromatic rings in the 1,1-diarylalkoxyl radical. Thus,
as compared to 4•, migration of the unsubstituted ring is
favored by the presence of a trifluoromethyl substituent
as in 7• and disfavored by the presence of a methoxy
substituent as in 5•.
Again, excellent Hammett-type correlations between
the partial rate constants for migration of the 4-X sub-
(34) Abeywickrema, A. N.; Beckwith, A. L. J.; Gerba, S. J. Org.
Chem. 1987, 52, 4072-4078.
J. Org. Chem, Vol. 70, No. 10, 2005 3889

Aureliano Antunes et al.
SCHEME 7
(Chart 2, TS-I). This picture is clearly more in line with
the experimental evidences presented above.38
This hypothesis is in contrast with the suggestion of
Banks and Scaiano that the 1-oxaspiro[2,5]octadienyl
radical is a true intermediate in the neophyl-like rear-
rangement of the 1,1-diphenylethoxyl radical.5 In this
context, it is important to point out that this suggestion
was not supported by any direct experimental evidence
but only by the results of AM1-UHF calculations, which,
however, are not expected to provide sufficiently reliable
results. Unfortunately, theoretical calculations carried
out at a higher lever of theory, which may provide
additional information on the mechanism of the O-
neophyl rearrangement of 1,1-diarylalkoxyl radicals, are
not presently available.
CHART 2
In conclusion, the results of product and time-resolved
kinetic studies on the O-neophyl rearrangement of ring-
substituted 1,1-diarylalkoxyl radicals strongly support
the hypothesis that the reaction proceeds through a
concerted mechanism and not via formation of an inter-
mediate 1-oxaspiro[2,5]octadienyl radical as previously
proposed.3,5,6 The rearrangement is governed by electronic
effects in the starting 1,1-diarylalkoxyl radicals, ex-
plained in terms of negative hyperconjugative contribu-
tions, whereas the stability of the rearranged carbon-
centered radical plays a minor role.
that fully accounts for the results of kinetic and product
studies. In addition, the reactivity order observed in the
series of the symmetric cyclopropyldiarylmethoxyl radi-
cals 4•, 6•, and 8• (8• > 4• > 6•) reflects their differences
in stability, which, accordingly, are expected to follow the
opposite order (6• > 4• > 8•), with the ring-dimethoxylated
radical 6• characterized by the greatest extent of reso-
nance stabilization.
Experimental Section
Materials. Spectroscopic-grade MeCN was used as received.
CH2Cl2 was purified prior to use by column chromatography
over basic alumina. (Diacetoxy)iodobenzene (DIB), iodine, 1,1-
diphenylethanol (1a), cyclopropyldiphenylmethanol (4a), and
cyclopropyl-4-methoxydiphenylmethanol (5a) were of the high-
est commercial quality available and were used as received.
Details of the synthesis of 1,1-diarylalkanols 2a, 3a, 6a-8a,
the tert-butyl 1,1-diarylethyl peroxides (1p-3p), and tert-butyl
cyclopropyldiarylmethyl peroxides (4p-8p) are given in the
Supporting Information. The purity of the 1,1-diarylalkanols
(1a-8a) employed in the product studies was always g99%.
Product Analysis. All the reactions were carried out under
an argon atmosphere. Irradiations were performed with visible
light (10 × 15 W lamps with emission between 400 and 550
nm, λmax ≈ 480 nm). The reactor was a cylindrical flask
equipped with a water cooling jacket thermostated at T ) 20
°C. Irradiation times were chosen in such a way as to avoid
complete substrate consumption. In a typical experiment, a
solution of the alcohol (10 mM) in CH2Cl2 (5 mL) containing
(diacetoxy)iodobenzene (22 mM) and iodine (10 mM) was
irradiated for times ranging between 5 and 60 min under Ar
bubbling. The reaction mixture was then poured into water
and extracted with CH2Cl2 (2 × 10 mL). The combined organic
layers were washed with a 10% aqueous thiosulfate solution
(2 × 30 mL) and water (2 × 30 mL) and dried over anhydrous
sodium sulfate. Reaction products and unreacted substrate
were identified by GC-MS and quantitatively determined by
GC using bibenzyl as internal standard. Good-to-excellent
mass balances (g85%) were obtained in all experiments.
Laser Flash Photolysis Studies. Laser flash photolysis
experiments were carried out with a laser kinetic spectrometer
using the fourth harmonic (266 nm) of a Q-switched Nd:YAG
laser. The laser energy was adjusted to e10 mJ/pulse by the
use of the appropriate filter. A 3-mL Suprasil quartz cell (10
mm × 10 mm) was used for all experiments. Argon- or oxygen-
saturated MeCN solutions of the peroxides 1p-8p (between
It is thus possible to discuss the results of product and
time-resolved kinetic studies described above in terms
of the two possible mechanisms proposed previously for
the O-neophyl rearrangement of the 1,1-diphenylethoxyl
radical (Scheme 7, R ) Me, cyclopropyl, showing for the
sake of simplicity a ring-unsubstituted radical):2 the
stepwise mechanism, proceeding through the formation
of a bridged 1-oxaspiro[2,5]octadienyl radical intermedi-
ate (paths b and c), and the concerted mechanism (path
a).
In a stepwise mechanism, by reasonably assuming that
the first step is the rate-determining one5 and by
considering that the rearranged benzyl radical (III) is
significantly more stable than the parent 1,1-diaryla-
lkoxyl radical (I), the bridged 1-oxaspiro[2,5]octadienyl
radical intermediate (II) is expected to be at a relatively
higher energy than I.5,36 Thus, as a consequence of the
Hammond postulate,37 the transition state of the rate-
determining step will be close in structure to radical II,
exhibiting a pronounced cyclohexadienyl radical charac-
ter (Chart 2, TS-II). On the other hand, in a concerted
mechanism a reactant-like transition state that is close
in structure to radical I, displaying a certain extent of
charge separation and a limited cyclohexadienyl radical
character, is to be expected, a transition state where
formation of a new bond by attack of the oxygen radical
onto the aromatic ring has not proceeded significantly
(35) According to this picture, the observation that the rate constants
for â-scission of ring-substituted cumyloxyl radicals are essentially
independent from the nature of the substituent (see ref 22) may
indicate that substituent effects operate in the same direction on both
the cumyloxyl radical and the product acetophenone.
(36) Asensio, A.; Dannenberg, J. J. J. Org. Chem. 2001, 66, 5996-
5999.
(38) As a consequence of a reactant-like transition state, possible
differences in C-CAr BDE in the 1,1-diaryalkoxyl radicals determined
by the presence of ring substituents should not be important.
(37) Hammond, G. S. J. Am. Chem. Soc. 1955, 77, 334-338.
3890 J. Org. Chem., Vol. 70, No. 10, 2005

Effect of Ring Substitution on O-Neophyl Rearrangement
5 × 10-4 and 3 × 10-3 M, A266 ≈ 1.0) were used. All the
experiments were carried out at T ) 22 ( 0.5 °C under
magnetic stirring. Rate constants were obtained by averaging
4-8 values, each consisting of the average of at least three
laser shots, and were reproducible to within 10%.
Gastaldo (Helios Italquartz s.r.l.) for providing us with
a photochemical reactor.
Supporting Information Available: Details on the prod-
uct studies and on the synthesis and characterization of the
precursor 1,1-diarylalkanols and tert-butyl 1,1-diarylalkyl
peroxides. Hammett-type correlations between the partial rate
constants for radicals 4•, 5•, and 7• and σ+ substituent
constants. This material is available free of charge via the
Internet at http://pubs.acs.org.
Acknowledgment. Financial support from the Min-
istero dell’Istruzione dell’Universita` e della Ricerca
(MIUR) is gratefully acknowledged. We thank Prof.
Enrico Baciocchi for helpful discussion, Prof. Basilio
Pispisa and Dr. Lorenzo Stella for the use and as-
sistance in the use of a LFP equipment, and Mr. Luigi
JO0502448
J. Org. Chem, Vol. 70, No. 10, 2005 3891