Reactions of Aromatic Selenoxides
A R T I C L E S
excitation wavelength was varied; significant wavelength de-
pendences of quantum yields have been noted for related
sulfoxides.3
than that calculated for (CH3)2SO at the G3 level of theory.48-50
If that trend holds through the dibenzothiophene/selenophene
series, the deoxygenation may very well be nearly thermoneutral
from the lowest triplet of DBSeO.
Discussion
The available evidence suggests that there is a common
deoxygenation mechanism in the photolyses of DBTO and
DBSeO. In particular, the product distributions for toluene
oxidation represent a common oxidation fingerprint, which may
be attributable to a common intermediate, taken to be O(3P).
Furthermore, the result that mixed photolyses using DBSeO and
Ph2SeO forms only DBSe strongly argues against the necessity
of dimer formation for deoxygenation, in direct analogy to
DBTO.3 The lack of observation of any isomer of DBSeO2 when
FPT degassing is used shows that the mechanism does not
involve direct disproportionation. The observation of photo-
chemistry at 77 K argues similarly against bimolecular interven-
tion, but even more convincingly, shows that no strongly
activated step is required for deoxygenation to occur from the
relevant excited state.
The disparity between Φ-DBSeO and Φ+DBSe indicates that
chemistry other than simple deoxygenation is occurring. Sec-
ondary photochemistry of DBSe was ruled out by control
experiments, so it must be concluded that there are parallel paths
taken by DBSeO*. We assume that FPT degassing is more
effective in eliminating all O2 from the samples than is Ar
flushing, and thus take those results as the O2-free standard. In
the absence of oxygen, the quantum yield for the formation of
Y is 0.05, while that for formation of DBSe is about 0.09,
meaning that in ethanol most, if not all, of the loss of DBSeO
is accounted for (Φ-DBSeO ∼ 0.15). In dichloromethane, it
appears that another pathway remains competitive.
The quantum yield for loss of DBSeO with Ar flushing was
considerably higher than with FPT degassing. We may speculate
that this comes about because the adventitious O2 that remains
after Ar flushing reacts (either directly or indirectly through a
chain mechanism) with a transient intermediate. One plausible
possibility is the biradical 12 resulting from R-cleavage, as
illustrated in Scheme 1.51,52 Both the results from Ph2SeO and
precedent within sulfoxide chemistry support the suggestion that
cleavage occurs in parallel with deoxygenation.
The observed byproduct Y, observed under freeze-pump-
thaw degassing conditions, is tentatively identified as the
selenenic ester 9, based on it having eight inequivalent,
nonexchangeable 1H signals in the aromatic region. The notation
9/Y is used to indicate that the assignment of the structure 9 to
Y is, while reasonable on many counts, not definitively
established. The structure of 9, however, is consistent with a
plausible competition between deoxygenation and R-cleavage
as primary photochemical mechanisms. After C-Se cleavage,
the so-obtained biradical 12 ought to reclose unimolecularly but
may close on either the Se (path A) or the O atom (path B), in
analogy to sulfinyl radical chemistry.12,41,53-56 The sulfenic ester
Clearly, the most salient practical result of this study is the
observation that DBSeO is an oxygen atom donor with many
advantageous properties of DBTO and a photochemical ef-
ficiency approximately 40 times higher. These studies suggest
a common intermediate, likely to be O(3P). If the simple scission
hypothesis stands up to the rigorous examination it deserves,
the quantum yield of approximately 0.1 for formation of DBSe
implies that DBSeO may be significantly more useful than
DBTO for studies of O(3P) chemistry in solution.
We may speculate that two factors contribute to the higher
quantum yield. First, Se is one row further down on the periodic
table from S, and a heavy atom effect is expected to contribute
to the probability of all intersystem crossing events. For the
parent DBTO, it is clear that there is insufficient energy in the
spectroscopic triplet for efficient S-O scission, leading to DBT
and O(3P). This has led us to suggest a mechanism in which
elongation of the S-O bond begins from the excited singlet
and that intersystem crossing is somehow coupled to that event.
One could imagine a direct spin-surface crossing along the
elongation pathway. Alternatively, the logical limit of this
mechanism is an initial step that forms an ion pair (consisting
of DBT•+ and O-) is followed by back-electron transfer, yielding
ground-state DBT and O(3P). In either event, the substitution
of Se for S might very well increase the probability of its
occurrence, either by a heavy atom effect or a lower ionization
potential. (The difference in IP between DBT and DBSe is only
0.08 eV, about 1.7 kcal/mol, but DBSe has the lower one.44) In
related work,29 it was shown that remote substitution of a Br or
I on the DBTO benzene nuclei resulted in a modest increase in
the quantum yield for deoxygenation, which was interpreted as
a heavy atom effect. These results are consistent with that
interpretation, but the ease of ionization of the aromatic nucleus
and small differences in excitation energy cannot be ruled out.
The other obvious reason the cleavage may be more efficient
comes from the expectation that the selenoxide bond should be
somewhat lower in energy than the sulfoxide bond, while the
electronic excitation energies are taken to be similar. The lack
of phosphorescence from DBSeO makes determination of a
triplet energy difficult, but there is little reason to think it will
vary much from the DBTO value of about 61 kcal/mol,
especially since neither the SO nor SeO is involved in the
aromatic system for those compounds.45 (The triplet energies
of DBSe and DBT are within 1 kcal/mol.46) We have begun a
study aimed at estimating the Se-O bond energy that will be
reported separately. However, we can report that the Se-O bond
energy calculated47 for (CH3)2SeO is about 12 kcal/mol lower
(43) We also note that the apparent quantum yield is somewhat lower in
deuterated benzene than in ordinary benzene. We attribute this to a low
level of quenching impurities in the deuterated solvent, which is not subject
to the same level of purification as the “spectrograde” ordinary benzene
we ordinarily use.
(44) Rodin, O. G.; Redchenko, V. V.; Kostitsyn, A. B.; Traven, V. F. J. Gen.
Chem. USSR 1988, 58, 1256-1261.
(45) Jenks, W. S.; Matsunaga, N.; Gordon, M. J. Org. Chem. 1996, 61, 1275-
1283.
(48) Curtiss, L. A.; Raghavachari, K.; Redfern, P. C.; Rassolov, V.; Pople, J.
A. J. Chem. Phys. 1998, 109, 7764-7776.
(49) Curtiss, L. A.; Raghavachari, K.; Redfern, P. C.; Pople, J. A. J. Chem.
Phys. 2000, 112, 7374-7383.
(50) Curtiss, L. A.; Redfern, P. C.; Rassolov, V.; Kedziora, G.; Pople, J. A. J.
Chem. Phys. 2001, 114, 9287-9295.
(51) Sato, T.; Yamada, E.; Akiyama, T.; Inoue, H.; Hata, K. Bull. Chem. Soc.
Jpn. 1965, 38, 5-1225.
(46) Zander, M.; Kirsch, G. Z. Naturforsch. 1989, 44A, 205-209.
(47) The G3Large basis set for Se is not currently implemented in Gaussian,
though it is available from PNNL.
(52) Sato, T.; Goto, Y.; Tohyama, T.; Hayashi, S.; Hata, K. Bull. Chem. Soc.
Jpn. 1967, 40, 2975-2976.
(53) Guo, Y.; Jenks, W. S. J. Org. Chem. 1995, 60, 5480-5486.
9
J. AM. CHEM. SOC. VOL. 126, NO. 49, 2004 16063