â-Scission of Ring-Substituted Cumyloxyl Radicals
J . Org. Chem., Vol. 67, No. 7, 2002 2269
is in line with the results of theoretical calculations,
indicating a decrease in the visible band intensity on
going from the 4-methoxy to the 4-trifluoromethylcumyl-
oxyl radical.6 In contrast, in the case of 4-methyl, 4-meth-
oxy, and 3,4-dimethoxycumyloxyl radicals, the UV ab-
sorption is stronger than the tert-butoxyl one.6,11 Moreover,
the corresponding acetophenones strongly absorb in the
260-300 nm region (see, for example, Figure 1), and their
formation in small amounts shortly after the laser pulse
(<100 ns) by â-scission of the corresponding cumyloxyl
radicals prevents the observation of the tert-butoxyl
radical absorption band in the same spectral region.11
Rea ctivity. As already noted, in all cases the rate
constants for decay of the cumyloxyl radical are es-
sentially identical (within the experimental error) to the
rate constants obtained following the buildup of the
product acetophenone. Moreover, it is known that under
these conditions hydrogen atom abstraction by the cu-
myloxyl radicals from the solvent (MeCN or MeCN/H2O)
or from the parent peroxide (used in concentrations
between 1.1 and 5.5 × 10-3 M) is negligible.5c Therefore,
the first-order decay of the visible absorption band can
in all cases be assigned to the unimolecular â-scission
reaction leading to a methyl radical and the correspond-
ing acetophenone as described in Scheme 1. The value
of kâ obtained for the unsubstituted cumyloxyl radical
in MeCN (kâ ) 7.4 × 105 s-1, averaging kv_ and kV_ from
Table 1) is almost identical to that obtained previously
by Ingold and Lusztyk in MeCN [kâ ) (7.5-7.6) × 105
s-1], where the radical was generated by 266 or 308 nm
light.5c
Solvent effects on the rate constants for â-scission of
the cumyloxyl radical have been studied in detail;5b,c,12
the rate of â-scission increases with solvent polarity, an
effect that has been attributed to the increased stabiliza-
tion of the transition state for â-scission through in-
creased solvation of the incipient acetophenone product.
Because of the formation of the carbonyl group, the
transition state should have a higher dipole moment than
the cumyloxyl radical and is therefore more strongly
solvated.13 In line with this hypothesis, we observe that
kâ increases on going from MeCN to the more polar
MeCN/H2O 1:1 for all cumyloxyl radicals. For example,
for the unsubstituted cumyloxyl radical, kâ increases from
7.4 × 105 s-1 in MeCN to 2.7 × 106 s-1 in MeCN/H2O 1:1
(i.e. by a factor 3.6). Interestingly, the same factor is
observed on going from MeCN/H2O 1:1 to H2O (where a
value kâ ) 1.0 × 107 s-1 has been determined).12
expected to influence the stability of the reaction product
(substituted acetophenone), the observation of a negli-
gible effect on kâ seems to be in contrast with the proposal
of a product-like transition state for â-scission of the
cumyloxyl radical, based on the temperature dependence
of the secondary R-deuterium isotope effect in methyl
radical formation.14 However, it has been suggested that
in alkoxyl radical â-scission to give a carbonyl compound
and an alkyl radical, the predominant driving force for
cleavage is the stability of the alkyl radical,4d,15,16 whereas
the stability of the carbonyl product plays a minor role.
Exp er im en ta l Section
Ma ter ia ls. MeCN (Aldrich) of the highest available purity
was used as received. Water was obtained from a Millipore-
Milli-Q system. Dicumyl peroxide (Aldrich) was recrystallized
twice from methanol.
3-Chloro, 4-chloro, 4-trifluoromethyl, 4-methyl, 4-methoxy,
3,4-dimethoxy, and 2,5-dimethoxycumyl alcohols were pre-
pared by reaction of the corresponding arylmagnesium bromide
with acetone in anhydrous tetrahydrofuran, purified by column
chromatography (silica gel, eluent petroleum ether/ethyl ac-
etate 3:1), and identified by GC-MS and 1H NMR.
tert-Butyl cumyl peroxides were prepared by reaction of the
ring-subtituted cumyl alcohols with tert-butyl hydroperoxide
in the presence of p-toluenesulfonic acid, according to a slight
modification of a previously described procedure.17 A 5.0-6.0
M solution of tert-butyl hydroperoxide in decane was added
to a solution of the cumyl alcohol (tert-butyl hydroperoxide/
cumyl alcohol ≈ 1.2) in CH2Cl2 (purified by filtration through
an alumina column) containing p-toluenesulfonic acid (alcohol/
acid ≈ 10). The reaction mixture was stirred at room temper-
ature and the reaction was followed by TLC. tert-Butyl cumyl
peroxides were purified by column chromatography (Florisil,
1
eluent pentane) and identified by H NMR.17
1
ter t-Bu tyl 3-ch lor ocu m yl p er oxid e: H NMR (CDCl3) δ
1.22 (s, 9H, C(CH3)3), 1.53 (s, 6H, ArC(CH3)2), 7.18-7.34 (m,
2H, ArH), 7.42-7.44 (m, 2H, ArH).
1
ter t-Bu tyl 4-tr iflu or om eth ylcu m yl p er oxid e: H NMR
(CDCl3) δ 1.23 (s, 9H, C(CH3)3), 1.56 (s, 6H, ArC(CH3)2), 7.56
(s, 4H, ArH).
ter t-Bu t yl 3,4-d im et h oxycu m yl p er oxid e: 1H NMR
(CDCl3) δ 1.25 (s, 9H, C(CH3)3), 1.56 (s, 6H, ArC(CH3)2), 3.87
(s, 3H, OCH3), 3.90 (s, 3H, OCH3), 6.79-7.08 (m, 3H, ArH).
ter t-Bu t yl 2,5-d im et h oxycu m yl p er oxid e: 1H NMR
(CDCl3) δ 1.30 (s, 9H, C(CH3)3), 1.59 (s, 6H, ArC(CH3)2), 3.77
(s, 3H, OCH3), 3.79 (s, 3H, OCH3), 6.73-7.31 (m, 3H, ArH).
Tim e-Resolved LF P Stu d ies. The alkoxyl radicals of
interest were generated at room temperature by direct laser
flash photolysis (LFP) of symmetric and asymmetric perox-
ides, using a 248 nm excimer laser (KrF*, Lambda Physik
EMG103MSC), providing 20 ns pulses with energies between
5 and 60 mJ /pulse (output power of the laser). Optical detection
was employed, with a pulsed xenon lamp as analyzing light.
Wavelengths were selected using a monochromator. The time-
dependent optical changes were recorded with Tektronix 7612
and 7912 transient recorders, interfaced with a DEC LSI 11/
73+ computer, which also controlled the other functions of the
instrument and preanalyzed the data.18 Argon- or oxygen-
saturated solutions of the peroxides (A248 ≈ 0.3-0.6) in MeCN
or MeCN/H2O flowed through a 2 mm (in the direction of the
With respect to the substituent effects, the data in
Table 1 show that ring substitution does not influence
to a significant extent the rate constants for â-scission
of the cumyloxyl radicals. Thus, C6H5C(CH3)2O• and
4-MeC6H4C(CH3)2O• undergo â-scission with very similar
kâ values (7.4 × 105 and 7.1 × 105 s-1, respectively), while
a slightly higher reactivity is observed for 4-MeOC6H4-
C(CH3)2O•, 3,4-(MeO)2C6H3C(CH3)2O•, and for the cumyl-
oxyl radicals bearing electron-withdrawing ring substit-
uents (kâ ≈ 1.0 × 106 s-1).
Thus, it appears that no significant charge separation
(e.g., as that responsible for the spectral behavior) occurs
in the transition state for â-scission of the cumyloxyl
radicals. By considering that the ring substituents are
(14) Zavitsas, A. A.; Seltzer, S. J . Am. Chem. Soc. 1964, 86, 3836-
3840.
(15) Kochi, J . K. J . Am. Chem. Soc. 1962, 84, 1193-1197.
(16) Wilsey, S.; Dowd, P.; Houk, K. N. J . Org. Chem. 1999, 64, 8801-
8811.
(12) Neta, P.; Dizdaroglu, M.; Simic, M. G. Isr. J . Chem. 1984, 24,
25-28.
(13) We thank a reviewer for this suggestion.
(17) Hendrickson, W. H.; Nguyen, C. C.; Nguyen, J . T.; Simons, K.
T. Tetrahedron Lett. 1995, 36, 7217-7220.
(18) Faria, J . L.; Steenken, S. J . Phys. Chem. 1993, 97, 1924-1930.