A R T I C L E S
Gagnier and Larock
an aryl or an alkyl group work well (entries 2 and 3). Aryl
iodides 7 and 9 both reacted under the standard palladium
conditions to provide the desired indanone products 8 and 10
in 100 and 60% yields, respectively. Unfavorable steric inter-
actions between the phenyl group and the organopalladium
intermediate may occur during the cyclization, accounting for
the lower yield with starting material 9 (see the later mechanistic
discussion). Styryl derivative 11, which contains an internal
carbon-carbon double bond, failed to cyclize to the desired
indanone (entry 4). Presumably, steric interactions between the
aryl- or acylpalladium intermediate and the methyl group do
not allow the palladium to add across the carbon-carbon double
bond. Starting materials, which contain electron-donating meth-
oxy groups on the arene, cyclize to the desired indanones (entries
5 and 6). However, when a methoxy group is placed ortho to
the iodide, the yield drops substantially, presumably due to steric
hindrance to oxidative addition (entry 6).
derivatives. In our system, a proton is incorporated into the
product to form the indanone or 2-cyclopentenone. One possible
source of this proton might be the solvent DMF. Therefore, an
experiment was carried out replacing DMF with DMA under
the standard palladium conditions and employing 1 as the
starting material. It was discovered that indanone (6) was still
produced in a near quantitative yield in the same 8 h reaction
time. A second possible proton source is water. Water could
be introduced into the system by either n-Bu4NCl or DMF,
because both of these materials are hygroscopic. Another
experiment was carried out with 1, in which 4 equiv of D2O
was added (eq 4).
In addition to reactions which form 5,6 ring-fused ketones, a
few examples have been attempted which would lead to a 5,5
ring-fused system. However, both heterocyclic iodide 17 and
vinylic triflate 19 produced simple carboxylic acid derivatives,
rather than the desired cyclic ketones (entries 7 and 8). The
failure of these substrates to cyclize to the ketones can be
attributed to the ring strain introduced in trying to form a 5,5
ring-fused system. The anticipated acylpalladium intermediate
may not be able to properly coordinate to the nearby olefin and
subsequently add across the carbon-carbon double bond.
Instead, the corresponding hydrolysis products 18 and 20 are
produced.
The indanone product recovered in a 95% yield contained
55% deuterium in the 2-position, and no deuterium incorporation
was found in the 3-position. In a separate experiment, it was
found that indanone itself will incorporate 55% deuterium at
the 2-position under the standard palladium conditions when 4
equiv of D2O is added. Therefore, the results from the
experiment in eq 4 only suggest that the palladium intermediate
formed after addition across the carbon-carbon double bond
does not directly incorporate a proton from water. The deuterium
is apparently only introduced later, after the indanone product
has been formed.
Vinylic derivatives, which contain a triflate, bromide, or
iodide, are able to cyclize to 5,6 ring-fused cyclopentenones
efficiently (entries 9-11). Vinylic triflate 21 reacted under the
standard palladium conditions to produce cyclopentenone 22
in a 95% yield after 12 h (entry 9). The corresponding vinylic
bromide 23 and vinylic iodide 24 also gave high 87 and 86%
yields, respectively, of cyclopentenone 22 (entries 10 and 11).
However, the reaction times for both 23 and 24 increased to 72
h. It should also be noted that vinylic bromide 23 required 10
mol % of Pd(dba)2, instead of 10 mol % Pd(OAc)2, as the
palladium catalyst (procedure B). When 23 was allowed to react
using Pd(OAc)2 as the catalyst, palladium black precipitated
out of solution after only 1 h, and only a trace of indanone 22
was detected after 72 h. Precipitated palladium black was only
found when reacting vinylic bromides and no other starting
materials. Vinylic iodides 25 and 27, which were synthesized
from R- and â-tetralone, respectively, also produced high 85
and 98% yields, respectively, of the corresponding cyclopen-
tenones (entries 12 and 13).
A 5,7 ring-fused cyclopentenone has also been prepared from
vinylic bromide 29 in a 98% yield (entry 14). Finally, acyclic
vinylic iodide 31 afforded cyclopentenone 32 in a 70% yield
after 24 h (entry 15).
The mechanism for this process was not clearly understood
initially, because the products differed significantly from those
reported in Negishi’s earlier work.8,10 Negishi’s products (see
eq 2) are presumably derived from the â-hydride elimination
of an organopalladium intermediate to form indenones or
insertion of a second CO under high pressures to form ester
In another mechanistic reaction, styryl derivative 34, which
contains two deuteria at the terminal positions of the olefin,
was allowed to react under the standard palladium conditions
(eq 5). Indanone was obtained in an overall 86% yield. It was
found that 35% deuterium was incorporated into the 3-position
and 12% deuterium was incorporated into the 2-position. The
deuterium incorporation in the 3-position suggests that a
deuterium is migrating from the 2-position. The relatively low
percentage of deuterium at the 2-position can be explained by
deuterium-proton exchange occurring with the H2O present
under the reaction conditions, after the 2,3-dideuterioindanone
is formed. From these mechanistic studies, the following
mechanism for this transformation is proposed (Scheme 2): (1)
Pd(OAc)2 reduction to the active palladium(0) catalyst; (2)
oxidative addition of the aryl iodide to Pd(0); (3) coordination
and insertion of carbon monoxide to produce an acylpalladium
intermediate; (4) acylpalladation of the neighboring carbon-
carbon double bond; (5) reversible palladium â-hydride elimina-
tion and re-addition to form a palladium enolate; and (6)
protonation by H2O to produce the indanone and a Pd(II) salt,
which is once again reduced to Pd(0).
Indenone was also prepared and reacted under the standard
palladium conditions to see if it can be reduced to indanone.
After 12 h, the only material found in the reaction mixture was
indenone itself, which was recovered in an 80% yield. This
(10) Negishi, E.; Coperet, S. M.; Ma, S.; Mita, T.; Sugihara, T.; Tour, J. M. J.
Am. Chem. Soc. 1996, 118, 5904.
9
4806 J. AM. CHEM. SOC. VOL. 125, NO. 16, 2003