Synthesis of Alkylidenecyclobutanones
COMMUNICATION
spectively (Table 2, entries 9 and 10). When VDCPs 1l and
1m were used, in which the cyclopropane methyl groups of
1k were replaced by protons or cyclohexyl groups, the de-
sired cyclobutanone products 3l and 3m were obtained in
23 and 39% yield, respectively (Table 2, entries 11 and 12),
and the remaining staring materials 1l and 1m could be re-
covered during the purification by silica gel column chroma-
tography. The X-ray crystal structure of 3m is presented in
the Supporting Information. In the case of VDCP 1n, which
has methoxy substituents on its aromatic rings, the conjugat-
ed triene product 3n was formed in 83% yield instead of
the alkylidenecyclobutanone, presumably due to the elec-
tronic properties of the methoxy groups (Table 2, entry 13).
The details regarding this rearrangement can be found in
our previous work, and a plausible mechanism for the for-
mation of 3n has been proposed in Scheme SI-2 in the Sup-
porting Information.[15] Using VDCP 1o as the substrate
provided product 3o in 55% yield under identical condi-
tions (Table 2, entry 14); a plausible mechanism for its for-
mation has been proposed in Scheme SI-2 in the Supporting
Information. VDCP 1p could also be used as the reactant
under the standard conditions to afford the corresponding
product 3p in 92% yield (Table 2, entry 15). When VDCP
1q was used as the substrate, in which the cyclopropane ring
contained an ethoxy substituent, the oxidation reaction gave
a complex mixture of products (Table 2, entry 16). Notably,
using the tetrasubstituted allene 1r as the substrate (a com-
pound that is not categorized as a VDCP), no reaction oc-
curred under the standard conditions (Table 2, entry 17).
Subsequently, we explored the reaction mechanism for
the formation of alkylidenecyclobutanone derivatives. Ini-
tially, a 31P NMR spectroscopic tracing experiment was car-
ried out by employing VDCP 1a as the substrate. As shown
in Figure 1, after a reaction time of 5 min, a signal at d=
29.1 ppm was observed, and has been assigned to an in situ
generated {AuI} species coordinated to substrate 1a. As the
reaction proceeded, a new signal at d=28.8 ppm appeared
after 15 min, and this became the only signal observed after
50 min (Figure 1). We hypothesized that this new signal cor-
responds to a newly formed {AuI} species coordinated to
pyridine 4a, which can also be considered an efficient cata-
lyst in this oxidation reaction. Consequently, the gold com-
plex 5 was synthesized by following a slightly modified liter-
ature procedure.[16a] Its structure has been unambiguously
determined by X-ray diffraction. The ORTEP diagram is
shown in Figure 2 and the CIF data are presented in the
Supporting Information.[16b] The 31P NMR spectroscopic
chemical shift for compound 5 was indeed observed at d=
28.8 ppm (see the Supporting Information for details).
Figure 1. 31P NMR spectroscopic tracing experiment (400 MHz, CDCl3,
85% H3PO4): a) 31P NMR spectrum of [PPh3AuSbF6] (d=30.7 ppm);
b) the reaction was carried out for 5 min; c) the reaction was carried out
for 15 min; d) the reaction was carried out for 25 min; e) the reaction
was carried out for 30 min; f) the reaction was carried out for 50 min.
Figure 2. ORTEP diagram (at the 45% probability level) of the gold com-
plex 5 containing a CHCl3 molecule.
Figure 3. The calculation results indicate that the cyclobu-
tyl–gold carbene intermediate C is more stable, by more
than 30 kJmolÀ1, than the intermediates A, D, and E, which
have similar energies to one another. Notably, the structure
of the cyclopropyl–gold complex B, which possesses a vinylic
cation adjacent to the cyclopropane ring, could not be ob-
tained from theoretical calculations, indicating that it may
not exist. The intermediate F, which is also expected to
react with water to produce 3a’ (see Scheme SI-1 in the Sup-
porting Information), is less stable than other intermediates
and presumably can be easily rearranged to form the more
stable intermediate C,[18] which is critical to the formation of
the final product 3 (Scheme 2).
In order to further reveal the reaction mechanism, we in-
vestigated the thermodynamic stabilities of several inter-
mediates (Figure 3) that may be involved in all possible re-
action pathways. In the calculations at the B3LYP/6-31+G*/
Lanl2DZ level (for details, see the Supporting Information),
PMe3 was used instead of PPh3 because previous theoretical
work revealed that this simplification is reasonable.[17] The
relative energies for each intermediate are shown in
On the basis of these experimental and theoretical results,
a plausible reaction mechanism is proposed in Scheme 3. In-
itially, the coordination of vinylidenecyclopropane 1a to the
Chem. Eur. J. 2012, 00, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
3
&
ÞÞ
These are not the final page numbers!