trophenyl)-3-phenylisoquinoline (11) in comparable yields,
indicating very poor selectivity for this reaction. Considering
that 4-iodonitrobenzene gave the best result of any aryl iodide
in the palladium-catalyzed cross-coupling of N-tert-butyl-
o-(phenylethynyl)-benzaldimine (1) without CO to form 4-(4-
nitrophenyl)-3-phenylisoquinoline (11),5 the low yield of 10
and the poor selectivity between 10 and 11 apparently result
from the very similar reactivities of the ArPdI and ArCOPdI
intermediates with the alkynyl imine, both of which promote
cyclization to the isoquinolines. In an attempt to improve
the selectivity of the reaction and the yield of the desired
ketone, we have carried out three further experiments in
which we have increased the CO pressure1c and decreased
the reaction temperature (entries 9-11).7 We were pleased
to observe that these experiments provided higher yields of
the desired product 10 and better selectivity between the two
3,4-disubstituted isoquinolines 10 and 11. Using both a lower
temperature and higher CO pressure improved the yield of
the ketone product 10 and the 10/11/4 ratio significantly
(entry 11).
Scheme 1
We also investigated the reactions of ethyl 3-iodobenzoate
with imino alkynes containing different R groups at the end
of the triple bond. Imino alkyne 12 bearing a 1-cyclohexenyl
group afforded the corresponding 4-aroylisoquinoline 13 in
a good yield, 55% (entry 12). Imine 14 containing an n-butyl
group afforded the desired product 15 in a 64% yield (entry
13). However, N-tert-butyl-2-phenylethynyl-3-pyridinealdi-
mine did not afford any of the desired ketone product.
Acyl halides readily undergo oxidative addition to Pd(0)
to form acylpalladium intermediates RCOPdX, which sub-
sequently undergo a wide range of useful transformations.8
We have, therefore, studied the utility of benzoyl chloride
in our chemistry. Under 1 atm of CO (Table 1, entry 14)
and with no CO present (entry 15), neither reaction afforded
3,4-diphenylisoquinoline (15) at all, indicating that the
initially formed acylpalladium intermediate does not undergo
decarbonylation to the corresponding arylpalladium species
very easily.9 However, whether there is external CO or not
does make a difference in the yields of the product 7 and
the reaction rates. The reaction was complete after 48 h under
1 atm of CO and was not complete after the same amount
of time without CO. Better results were obtained using 1
atm of CO, in which case a 62% yield of ketone 7 was
obtained (entry 14).
(3) intramolecular nucleophilic attack of the nitrogen atom
of the imine on the activated carbon-carbon triple bond to
afford intermediate C, (4) reductive elimination to form the
carbon-carbon bond between the carbonyl group and the
isoquinoline ring in D and simultaneous regeneration of the
Pd(0) catalyst, (5) cleavage of the tert-butyl group from the
nitrogen to release the strain between the tert-butyl group
and the 3-phenyl group with simultaneous generation of the
3-substituted 4-aroylisoquinoline. Two competing processes
are (1) cyclization of the starting material by a thermal or
Pd(II)-catalyzed process to afford the 3-monosubstituted
product3 and (2) cyclization of the imine starting material
promoted by an arylpalladium intermediate to afford a
3-substituted 4-arylisoquinoline.5
The yields of ketones obtained by this process are
relatively independent of the nature of the substituents on
the aryl iodide, while the yields of 4-arylisoquinolines
obtained from arylation of these same imines5 are highly
dependent on the nature of the substituents present in the
aryl iodide. This is easily understood when one considers
that the key step in the present synthesis apparently involves
attack of an electron-deficient acylpalladium species on the
carbon-carbon triple bond. The nature of the substituents
present in the aroylpalladium intermediate is not going to
change their electronics as profoundly as they would the
electronics of the corresponding arylpalladium species. The
presence of steric hindrance in the aryl iodide is also less
likely to affect the yield in the carbonylative cyclization,
because of the presence of the carbonyl group in the
aroylpalladium intermediates.
The mechanism shown in Scheme 1 is proposed for this
process. It consists of the following key steps: (1) oxidative
addition of the organic halide to the Pd(0) catalyst, followed
by CO insertion, (2) the resulting acylpalladium intermediate
A coordinates to the alkyne triple bond to form complex B,
which activates the triple bond toward nucleophilic attack,
(7) During our optimization work, we found out that the reactions of
substrate 1 and 4-iodoanisole at 80 and 100 °C gave the same yields of
products 2, 3, and 4, although the reaction at 80 °C took a much longer
time to complete.
(8) Tsuji, J. Palladium Reagents and Catalysts; John Wiley & Sons:
Chichester, 1999; p 253.
(9) (a) Tsuji, J.; Ohno, K. J. Am. Chem. Soc. 1968, 90, 94. (b) Blaser, J.
U.; Spencer, A. J. Organomet. Chem. 1982, 233, 267. (c) Spencer, A. J.
Organomet. Chem. 1982, 240, 209. (d) Spencer, A. J. Organomet. Chem.
1984, 265, 323.
In summary, we have developed an efficient synthetic
approach for the carbonylative cyclization of N-tert-butyl-
o-(1-alkynyl)benzaldimines to the corresponding 3-substi-
tuted 4-aroylisoquinolines. A systematic and extensive study
of the scope and limitations of this process is currently under
investigation in our laboratory and will be reported in due
course.
Org. Lett., Vol. 4, No. 2, 2002
195
Dai, Guangxiu
