Angewandte
Communications
Chemie
Table 1: Optimization of the reaction conditions.[a]
Entry
(ArS)2
Yield [%][b]
d.r.[c]
1
2
3
4
5
6
5a
5b
5c
5d
6a
6b
6c
6c
10
13
48
7
80
71
88
87
67:33
72:28
69:31
64:36
72:28
72:28
72:28
72:28
7
8[d,e]
[a] Reaction conditions: N-tosyl vinylaziridine (1, 0.10 mmol), styrene
(0.12 mmol), disulfide (0.005 mmol). [b] Combined yield of the diaste-
reomers, as determined by 1H NMR spectroscopy with mesitylene as an
internal standard. [c] The diastereomeric ratio (trans/cis) was determined
by 1H NMR analysis of the crude material. [d] The reaction was carried
out for 2 h with 6 mol% of the disulfide. [e] Isolated yield.
Scheme 3. Proposed catalytic cycle and catalyst-decomposition path-
ways.
species reacts with an alkene to afford carbon radical II,
which cyclizes to furnish 2 with concomitant regeneration of
the thiyl radical. To elucidate the cause of catalyst deactiva-
tion, we scrutinized the residual products of the reactions
described above. The isolation and identification of minor by-
products revealed that the catalyst was trapped mainly by two
different pathways. One decomposition pathway for the
catalyst is the abstraction of a hydrogen atom from the
intermediate aminyl radical I to give allylic sulfide 3. Another
is the formation of bis(arylthiolated) 4, which is formed by the
addition of the thiyl radical to the alkene, followed by
a reaction of the thus generated carbon radical III with
another thiyl radical or the disulfide.[15] Whereas the former
pathway (I!3) is hard to circumvent, we reasoned that the
latter pathway leading to 4 could be prevented by appropriate
catalyst design.
vinylaziridine 1, the catalyst loading was slightly increased to
6 mol%. Under these optimized reaction conditions, the
reaction afforded the desired pyrrolidine in 87% yield within
2 h (entry 8). Although the reaction proceeds in a variety of
solvents, benzene was chosen for this study because of the
high solubility of the catalyst therein and the low probability
of undesired hydrogen abstraction from the solvent.
We examined the scope of the reaction by subjecting
a variety of styrene derivatives to the optimized reaction
conditions (Scheme 4).[16] Even though we observed that the
position of the functional group on the aryl moiety did not
affect the reaction yield (products 2c,d), the electronic
features of the styrenes were found to have a substantial
effect on their reactivity. Whereas the cyclization with
electron-rich styrene derivatives reached completion within
2 h (products 2e,f), the reaction of electron-deficient alkenes,
such as 4-bromostyrene, required longer reaction times and/
or higher catalyst loadings (products 2h–j). The catalytic
system was applicable to a-substituted styrenes, which
afforded the corresponding pyrrolidines with a quaternary
center in good yield and with modest diastereoselectivity
(products 2k–m). The use of a-silyloxy and a-alkoxy styrenes
enabled the incorporation of a protected tertiary alcohol
moiety in the pyrrolidine ring (products 2n,o). The reactivity
of these electron-rich styrenes was found to be higher than
that of a-alkyl styrenes. Notably, the reaction also proceeded
with a b-substituted styrene to give the corresponding
trisubstituted pyrrolidine 2p. A trans configuration was
ascertained for the stereochemical relationship between the
methyl and phenyl group of 2p; thus, a mixture of just two out
of four possible diastereomers was obtained. The E/Z
isomerization of b-methylstyrene is faster than the reaction
under the previously described optimal reaction conditions,
thus justifying the use of isomeric mixtures.
We envisaged that the use of a sterically hindered thiyl
radical should prevent the second addition of a thiyl radical or
disulfide (pathway III!4), thereby making the thiyl radical
available for the catalytic cycle. To validate this hypothesis,
we screened several aryl disulfides. For example, the 2,6-
dimethylphenyl and 2,4,6-triisopropylphenyl disulfides 5a
and 5b generated pyrrolidine 2b merely in low yields
(Table 1, entries 1 and 2), and the use of electron-deficient
tris(trifluoromethyl)phenyl disulfide 5c promoted the reac-
tion only modestly (entry 3). No improvement in conversion
was observed when 2,6-diarylphenyl disulfides, such as 2,6-
dimesitylphenyl disulfide 5d, were used (Table 1, entry 4).
Consequently, we focused our attention on the introduction of
sterically more demanding silyl groups at the 2,6-positions of
the aryl disulfide. Even though the use of trialkylsilyl-
substituted catalysts led to merely modest conversion at
best (data not shown), the introduction of triphenylsilyl
groups (disulfide 6a) resulted in a dramatic improvement in
the yield to 80% (Table 1, entry 5). To further improve the
catalytic activity, we fine-tuned the electronic properties of
the 2,6-bis(triphenylsilyl)aryl disulfide 6 by replacing the
substituent at the para position of the catalyst (Table 1,
entries 5–7). These experiments provided us with the optimal
catalyst 6c, which contains triphenylsilyl groups at the ortho
positions and a trifluoromethyl group at the para position of
the principal aryl moiety. To secure full consumption of
2
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
These are not the final page numbers!