afforded product [D1]-6a in 95% yield (NMR), and no
deuterium scrambling was detected, thereby suggesting that
À
no cleavage of the alkynyl C H bond is involved. To further
confirm this conclusion, a crossover experiment was carried
out using an equimolar mixture of [D1]-1a and 1e. Both [D1]-
6a and 6e were obtained (see the Supporting Information),
and the crossover product was not detected by 1H NMR
spectroscopy, thus indicating that no alkyne–vinylidene
rearrangement is involved. This is in contrast to the proposed
vinylinde intermediates in [{Cp*IrCl2}2]-mediated activation
of simple terminal alkynes toward external OH and NH
nucleophiles.[14] A plausible mechanism is given in Scheme 4.
Scheme 3. Reactivity of iridium azomethine complexes.
reactive electrophiles such as CD2Cl2 can slowly react with 2a
to furnish product [D2]-5 (72%), which was characterized by
NMR spectroscopy and ESI-MS.
The lability of this azomethine ligand bodes well for
catalytic synthesis of azomethines. Excitingly, [{IrCp*Cl2}2]
proved to be a highly efficient catalyst for cyclizing 1a into
azomethine ylide 6a in 96% yield (NMR) and 84% yield
(isolated), and even a 1 mol% loading of the catalyst is
sufficient at room temperature. In accordance with the poor
reactivity of [{RhCp*Cl2}2] in stoichiometric reactions with
1a, [{RhCp*Cl2}2] failed to catalyze the cyclization of 1a. In
comparison, [{Ru(p-cymene)Cl2}2] (2 mol%) is an active
catalyst, and the azomethine product was generated in
comparably high yield (92%), albeit after a longer reaction
time. Under the iridium catalysis, nitrone alkynes 1a–e
bearing various substituents on both the aryl ring and the N,
cyclized smoothly. Thus free azomethine ylides 6a–e were
successfully isolated (70–84%) as low-melting solids after
column chromatography [Eq. (3)]. In the IR spectrum of 6a,
the signal for the C–O group appears at 1570 cmÀ1, and in the
Scheme 4. Proposed catalytic cycle for the synthesis of azomethines.
Upon coordination, the alkyne unit undergoes attack by the
nitrone group in six-exo-dig selectivity (step a). Subsequent
À
N O cleavage generates an a-oxo carbenoid intermediate
(step b). Attack of the amine group at this electrophilic
carbene affords a C-bound azomethine complex (step c), and
C-bound to O-bound isomerization furnishes intermediate
[D1]-2a (step d; R = tBu). The catalytic cycle is completed
when the azomethine ligand (either C bound or O bound) is
substituted by an incoming nitrone alkyne substrate.
The reactivity of azomethine ylides at the oxygen center
has been explored. Stirring a CH2Cl2 solution of azomethine
6a and metal complexes [{IrCp*Cl2}2] and [{Ru(p-cyme-
ne)Cl2}2] (0.5 equiv) resulted in nearly quantitative formation
of complex 2a and 3, respectively (Scheme 5). However, no
coordination occurred when [{RhCp*Cl2}2] and [{Ir(cod)Cl}2]
(cod = 1,5-cyclooctadiene) were used. These results indicated
the donor capacity of azomethines is sufficient to substitute
bridging chlorides, but binding is limited to rather electro-
philic metals. Furthermore, the nucleophilicity of the O atom
of azomethines was revealed in the reaction of 6a with HCl
and MeI, wherein protonation and methylation products were
isolated, respectively, in high yields (Scheme 5).
13C NMR spectrum (CD2Cl2)
168.0 ppm) was detected for the carbonyl carbon atom.
These indications all point to a certain amount of double
bond character.[13]
Complex 2a is the only iridium species observed in the
catalytic cyclization of 1a. In addition, when complex 2a
(1 mol%) was employed as a catalyst for the cyclization of 1a,
azomethine 6a was obtained at a rate comparable to that
when using [{IrCp*Cl2}2] as a catalyst, thus supporting the
intermediacy of an O-bound azomethine complex in the
catalytic cycle. Deuterium labeling experiments were carried
out to probe the mechanism. Cyclization of [D1]-1a (deu-
terated at the alkynyl position) using [{IrCp*Cl2}2] (1 mol%)
a low-field signal (d =
Generic azomethines are known to undergo [3+2] dipolar
addition with p bonds, thus representing a powerful method
for the consctruction of complex azacycles.[15] Independently,
the groups of Shin[3b] and Liu[5b] have reported that gold
Angew. Chem. Int. Ed. 2011, 50, 7791 –7796
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7793