Photodeoxygenation of Dibenzothiophene Sulfoxide
J. Am. Chem. Soc., Vol. 119, No. 1, 1997 101
interaction of the oxygen atom with the π system of 2 or an
excited state whose reactivity merely effectively mimics O(3P).
An important issue is the strength of the S-O bond and its
relationship to the excited state energies. Fluorescence and
phosphorescence measurements place the singlet and triplet
energies of 1 at 82 and 61 kcal/mol, although there is some
variation with solvent.18 Experimental heats of formation are
available for many sulfides, but only a few sulfoxides: DMSO,
a few other simple dialkyl sulfoxides, and Ph2SO. From these,
and the heat of formation of O(3P), bond energies of 85-90
kcal/mol can be established.40 These are several kcal/mol more
than what is available from the relaxed singlet of 1 and make
the dissociation mechanism appear unlikely, especially at 77
K. However, in a recent computational study,41 it was shown
that the aromaticity of 2 effectively lowers the S-O bond energy
of 1 down to the range of 74-78 kcal/mol, now well under the
energy of the singlet excited state. (This bond weakening effect
is larger and larger still for benzothiophene sulfoxide and
thiophene sulfoxide, respectively.)
The pattern of quantum yields is qualitatively understood as
a reflection of the reactivity of the solvent with the nascent O
atom. There exists a competition between recombination with
2, diffusive separation, and reaction of the O atom with solvent.
Intuitively, it is reasonable to expect that tetrahydrothiophene,
cyclohexene, and DMSO would be qualitatively more reactive
with O(3P) than the other substrates, as one anticipates reactions
analogous to an electrophilic triplet carbene. An increased
reactivity of the solvent with O(3P) then leads to an increase in
Photolysis of N2O in hydrocarbon solutions with 185 nm light
produces mainly O(1D).51 A very slight selectivity for secondary
or tertiary hydroxylation Vs primary was attributed to minor
amounts of ground-state O(3P). Hydroxylation of cis-decalin
also proceeded with retention of configuration (95% at 0 °C).
All of these point to direct insertion, in contrast to the results
in Table 2. Once the number of hydrogens are taken into
account, there is approximately 1 order of magnitude selectivity
for secondary hydroxylation over primary and a similar, if
slightly lower, preference for tertiary hydroxylation over
secondary. Taken with the isotope effect of about 6 observed
for both products of cyclohexane, a stepwise oxidation pathway
of hydrogen abstraction followed by recombination (or dispro-
portionation) is suggested.
Similarly, the partial randomization of stereochemistry in the
epoxidation of octene and â-methyl styrene isomers suggests
that the reaction is stepwise, though the intermediate still has a
fairly short lifetime, as gauged by the incomplete equilibration
of stereochemistry. Significant cis/trans isomerization of the
alkene under the reaction conditions was observed for cis-â-
methylstyrene, but for none of the others. Even in that case,
however, the percentage of trans-epoxide far exceeded the
percentage of trans-olefin.
For all the alkenes, epoxidation is favored over hydroxylation.
However, one example in particular deserves attention. In the
photolysis of 1 in allylbenzene, both cinnamyl alcohol and 21
were identified. Assuming for the moment that the initial
oxidant is O(3P), this is rationalized by the presence of a discrete
allyl-type radical 19.
the observed Φrxn
.
We are not aware of any clean method to generate O(3P) in
solution, aside, perhaps, from photolysis of 1. This means
ordinary chemical tests such as those used to probe for 1O2 are
not available. While atomic oxygen, due to its importance in
combustion and atmospheric chemistry, has been investigated
repeatedly in the gas phase, data regarding its solution phase
reactivity are very limited, particularly with regard to organic
substrates. Nonetheless, the products in Table 2 are consistent
with intuitive expectations for O(3P), which we would expect
to behave like an electrophilic triplet carbene. All of the
observed products can be explained on the basis of a reaction
sequence that begins with hydrogen abstraction or addition of
O to an olefin (or arene).
The hydroxylation reactivity reported in Table 2 is similar
to that observed for Cytochrome P-450 models, much of which
can be described in terms of Groves’ “oxygen rebound”
mechanism.42-50 Essentially, hydroxylations are described as
a net hydrogen abstraction followed by delivery of a hydroxyl
radical to the substrate radical, as would be anticipated for a
true triplet oxene. (In contrast to the current results, however,
the P-450 model epoxidations are usually stereospecific.)
We presume that benzene is oxidized to phenol by stepwise
addition of O to the ring, forming benzene oxide. The latter is
thought to rearrange to phenol under the reaction conditions.
The small isotope effects are consistent with the initial step of
addition to the ring. Similar isotope effects were recently
obtained for the hydroxylation of xylene isotopomers by
microsomal cytochrome P-450.52 The 18O-labeling experiments
clearly show that the O atom derives from 1. We find no direct
evidence of benzene oxide or of oxepin. However, it is well
known that benzene oxide/oxepin readily isomerizes to phenol
(sometimes along with other products) in the presence of
Bro¨nsted acid, heat, or light.53,54 The less than quantitative yield
of phenol may reflect undetected products derived from other
processes available to the benzene oxide/oxepin. Indeed, the
percentage recovery of phenol differed for benzene and benzene-
D6, directly indicating an isotope-dependent branching prior to
phenol formation.
(40) Herron, J. In The Chemistry of Sulfoxides and Sulfones; Patai, S.,
Rappoport, Z., Stirling, C. J. M., Ed.; John Wiley and Sons Ltd.: New
York, 1988; pp 95-105.
(41) Jenks, W. S.; Matsunaga, N.; Gordon, M. J. Org. Chem. 1996, 61,
1275-1283.
For comparison to previous results with O(3P), the best data
are the rate constants of Bu¨cher and Scaiano,31 who generated
O(3P) by photolysis of pyridine N-oxide. This compound
(42) References 43-50 are meant as leading sources, not an exhaustive
list.
(43) Meunier, B. Chem. ReV. 1992, 92, 1411-1456.
(44) Weber, L.; Hommel, R.; Behling, J.; Haufe, G.; Hennig, H. J. Am.
Chem. Soc. 1994, 116, 2400-2408.
(51) Varkony, T. H.; Pass, S.; Mazur, Y. J. Chem. Soc., Chem. Commun.
1975, 457-458.
(45) Higuchi, T.; Shimada, K.; Maruyama, N.; Hirobe, M. J. Am. Chem.
Soc. 1993, 115, 7551-7552.
(52) Hanzlik, R. P.; Ling, K.-H. J. J. Am. Chem. Soc. 1993, 115, 9363-
(46) Nam, W.; Valentine, J. S. J. Am. Chem. Soc. 1993, 115, 1772-
9370.
1778.
(53) Vogel, E.; Gu¨nter, H. Angew. Chem., Int. Ed. Engl. 1967, 6, 385-
401.
(47) Groves, J. T.; Subramanian, D. V. J. Am. Chem. Soc. 1984, 106,
2177-2181.
(48) Groves, J. T.; Viski, P. J. Org. Chem. 1990, 1990, 3628-3634.
(49) Groves, J. T.; Gross, Z.; Stern, M. K. Inorg. Chem. 1994, 33, 5065-
5072.
(54) Jerina, D. M.; Witkop, B.; McIntosh, C. L.; Chapman, O. L. J. Am.
Chem. Soc. 1974, 96, 5578-5580.
(55) Hata, N.; Tanaka, I. J. Chem. Phys 1962, 36, 2072-2077.
(56) Ogawa, Y.; Iwasaki, S.; Okuda, S. Tetrahedron Lett. 1981, 22,
2277-2280.
(50) Tsuchiya, S.; Seno, M. Chem. Lett. 1989, 263-266.