Angewandte
Communications
Chemie
using diynes as C4 units,[16] we surmized that the azepine
framework could be assembled by [3+2+2] cycloaddition of
diynes with 2H-azirines, in which the 2H-azirines would serve
as three-atom components. Indeed, examples of the forma-
tion of aza-seven-memberd rings from diynes are extremely
rare.[17] Our previous studies demonstrated that [3+2] rather
than [3+2+2] cycloadducts were preferentially constructed in
the reaction of diynes with methyleneaziridines, thus leaving
an alkyne unit intact.[16c] To our delight, the reaction of diynes
with 2H-azirines follows the [3+2+2] pathway by using an
appropriate ruthenium catalyst,[18] thus providing the desired
azepine architectures (Scheme 1c). This approach features an
À
unprecedented metal-catalyzed C C bond cleavage of 2H-
azirines, and the challenging construction of aza-seven-
membered rings from diynes. Herein, we report our prelimi-
nary results.
At the outset, the diyne 1a and 2H-azirine 2a were chosen
as model substrates for the optimization of the reaction
conditions [Eq. (1)]. The results are summarized in Table S1
in the Supporting Information. Some typical transition-metal
catalysts for cycloadditions were first investigated. Unfortu-
nately, no desired product was observed with our previously
developed iron catalyst[16a] (Table S1, entry 1). In addition,
cobalt,[19a] nickel,[19b] rhodium,[19c] and iridium[19d] catalysts
also failed to provide any new products (entries 2–5). How-
ever, the possibility that the above systems might be workable
could not be completely excluded if an in-depth investigation
is done. Gratifyingly, the product 3H-azepine 3a was obtained
in 71% yield using 10 mol% [Cp*Ru(COD)Cl] (entry 6;
COD = 1,5-cyclooctadiene, Cp* = C5Me5). The structure of
3a was unambiguously confirmed by X-ray crystal diffrac-
tion.[20] Various ruthenium (II) catalysts were then evaluated
in the reaction, however, none of them led to better results
(entries 7–11). Intriguingly, ruthenium(II) complexes lacking
either the Cp or Cp* ligand exhibited no appreciable catalytic
activity (entries 9–11). Subsequently, a screening of solvents
revealed that both DCE and DCM (DCE = 1,2-dichloro-
ethane, DCM = dichloromethane) showed the best perfor-
mance (entries 12–15). Furthermore, a simple inspection on
the reaction temperatures indicated that lowering the temper-
ature to 258C can afford a better yield (entry 17). A slight
improvement in product yield can be achieved by prolonging
the reaction time to 20 hours (80%, entry 19). It is note-
worthy that the dimerization of diynes was commonly
detected under ruthenium catalysis,[18d,e] thus slightly lowering
the yield of 3a. Remarkably, decreasing the catalyst loading
to 5 mol% displayed lower activity (entry 20).
Scheme 2. Substrate scope. Yields of isolated products are given.
electron-donating and electron-withdrawing groups on the
phenyl ring could be successfully introduced, thus providing
the corresponding polysubstituted azepines (3b–e, 3g–i) in
moderate to excellent yields. The R1 group is placed near to
the methylene unit (CH2) in the product, as identified by the
X-ray crystal diffraction of 3g.[20] However, the 2H-azirine 2 f,
bearing a strong electron-withdrawing aryl group, provided
3 f in low yield. 2H-azirines bearing an ortho-substituted
phenyl ring (2h, 2i) or 1-naphthyl group (2j) reacted
smoothly to afford the corresponding azepines in high
yields, thus suggesting that steric hindrance is well tolerated
(3h–j). Moreover, the 2-thienyl group was tolerated and
provided the desired product 3k albeit with moderate yield.
Replacement of the aryl group with an alkyl substituent also
led to the effective formation of the corresponding apezine
(3l), although a much lower yield was observed. Notably, the
reaction was sensitive to the electronic and steric effect of the
R2 substituent. 2H-azirine, having an electron-donating group
(Me) on the para position of the phenyl ring (2m), showed
higher reactivity than that bearing an electron-withdrawing
group (NO2, 2n). When a sterically more demanding 2-
methoxyphenyl-substitued 2H-azirine (2o) was employed,
a satisfactory yield can still be achieved.
Next, the reactions of various terminal diynes with 2H-
azirines were investigated. The sulfonamide-based diynes (1a,
1p) could undergo this cyclization and be transformed into
the corresponding cycloadducts in moderate yields. However,
malonate-tethered terminal diynes were found to be less
effective, thus affording the azepines in 30 and 24% yield,
With the optimal reaction conditions secured, we turned
our attention to the scope of this reaction. The variations of
1
=
the R group on the C N double bond moiety of 2H-azirines
were first examined. As highlighted in Scheme 2, both
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 2861 –2865