Intramolecular Homolytic Substitution
FULL PAPER
al as the reaction went on. We thus used excess AIBN
(Table 1, entry 1), which proved beneficial, and the sultine
yield improved to 62%. Only traces of the biaryl com-
pounds I were isolated. I is thermally and chromatographi-
cally unstable. This may be due to the distortion introduced
by the biaryl in a seven-membered ring and was a recurring
problem throughout our trials. Thus, although formation of
the sultine was highly reproducible, formation of I was not
optimized further, and, when given, the yields of I are only
indicative.
spacer. This is not the case with sulfinates. Besides, the addi-
tional heteroatom is directly attached to the atom undergo-
ing substitution, and could conceivably be expected to have
an influence on the reactivity. We thus decided to probe this
issue in more detail.
Thiosulfinates were examined first, but unfortunately,
only degradation was observed under the reaction condi-
tions. We thus switched to sulfinamides with more positive
results (Table 2).
Neither the nature of the starting halide (iodine or bro-
mine) nor that of the radical mediator (TBTH or tris(trime-
thylsilyl)silane (TTMSS)) had a significant impact on the
isolated yields (compare for example Table 1, entries 3–4 to
Table 1, entry 1). The reaction still proceeded at room tem-
perature (Table 1, entry 2), but reduced 1a (X=H) was also
obtained (along with 14% of the biaryl I).
Table 2. Variation of the tethering heteroelement.
Similarly, introduction of substituents on the aromatic
leaving groups did not modify the outcome of the reaction
(Table 1, entries 5–7). In the case of o-nitro substrate 1c,
two equivalents of the reducing agent are required to reach
completion (Table 1, entries 5–6). This is most certainly due
to adventitious reduction of the nitro group by TBTH,
which consumes the mediator. We did not determine wheth-
er this was happening before or after cyclization. The
methyl substituents in the mesityl case were introduced to
hamper the radical 1,7-cyclization, and thus to favor the thi-
ophilic attack. However, the sultine yield remained the
same.
Entry
Substrate
R
LG
Product
Yield [%]
1
2
3
4
3a
3b
3c
3d
H
H
Me
Me
p-Tol
tBu
p-Tol
tBu
4a
4a
4c
4c
58
85
33
22
Sulfinamides followed the same trends as sulfinates. With
p-tolyl substituted sulfinamides, formation of the biaryl com-
pounds was again observed (Table 2, entries 1 and 3), but
the seven-membered cyclic sulfinamides were extremely un-
stable and could not be isolated. Introduction of the tert-
butyl substituent was beneficial (Table 2, entry 2), as before.
However, when the nitrogen atom was substituted by an
alkyl group, the yields dropped sharply (Table 2, entries 3–
4). This was especially surprising for the tert-butyl substitu-
ent, which led to a disappointing 22% yield of cyclic sulfina-
mide as the sole product of the reaction, potentially because
of an unwanted 1,5-hydrogen atom translocation a to the ni-
trogen atom.
Homolytic substitutions are primarily governed by the
bond dissociation energy (BDE) of the bond that is created
and the one that is broken. Because the alkyl–sulfur bonds
ꢀ
are much weaker than the aryl–sulfur bonds (the C S BDE
in tBu-SMe is 70.4 kcalmolꢀ1 compared with 85.4 kcalmolꢀ1
for Ph-SMe; it is 67.6 kcalmolꢀ1 in tBu-SO2Me compared
with 82.3 kcalmolꢀ1 for Ph-SO2Me[22]), we decided to re-
place the aromatic group with alkyl moieties. In this way,
the biaryl formation would no longer be an issue. We thus
selected four different substituents: methyl, tert-butyl,
benzyl, and trifluoromethyl (Table 1, entries 8–16).
Size of the ring: The next parameter we examined was the
size of the ring that was formed during the homolytic substi-
tution. Our initial experiments involved five-membered
rings, then we extended the cyclization to six-membered
rings (Scheme 1).
Those larger ring homologues could be obtained, albeit
much less efficiently, certainly reflecting the increased en-
tropic demand of the SHi transition state. Nonetheless, cyclic
six-membered sulfinates can be accessed through this reac-
tion.
Gratifyingly, this proved to be a key change, because sul-
tine 2 was obtained in good yields from all the substrates.
Thus, sulfinates do undergo smooth homolytic substitution,
and to date, they are the most highly oxidized sulfur mole-
cules to accommodate this pathway. The method appears
C
C
well suited to access interesting radicals (Me and CF3 for
example). The tert-butyl derivatives are particularly appeal-
ing, because they are easily available,[23] including their
enantiopure forms (see below).[24] The reactions were so effi-
cient that they did not require slow addition of the reducing
agent, and because rearomatization no longer followed, only
substoichiometric amounts of initiator were required.
Incoming radical: In contrast, the reaction of sulfinates
proved very tolerant of the incoming radical modification.
Electron-donating groups (Table 3, entries 1–5), as well as
an electron-attracting moiety (Table 3, entries 6–8) could be
installed on the aryl group with no modification of the ob-
served reactivity.
Nature of the tethering heteroatom: A less obvious feature
of our system was related to the presence of a heteroatom
in the ring that is forming. Generally, intramolecular homo-
lytic substitutions involving sulfur rely on a full carbon-atom
We next envisaged heteroaryl-derived radicals, which
proved more difficult to work with. Thienyl radicals led to
Chem. Eur. J. 2009, 15, 10225 – 10232
ꢃ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
10227