Intramolecular nucleophilic addition of aryl bromides to ketones catalyzed by palladium

Full text of DOI:10.1021/ja000415k

J. Am. Chem. Soc. 2000, 122, 4827-4828
Table 1. Intramolecular Catalytic Arylpalladation of Ketones
4827
Intramolecular Nucleophilic Addition of Aryl
Bromides to Ketones Catalyzed by Palladium
Long Guo Quan, Mouad Lamrani, and Yoshinori Yamamoto*
Department of Chemistry, Graduate School of Science
Tohoku UniVersity, Sendai 980-8578, Japan
ReceiVed February 3, 2000
Among many synthetically useful palladium-catalyzed trans-
formations,1 a reaction of uncommon pattern is the palladium-
catalyzed nucleophilic addition of organic halides to ketones and
aldehydes, although such addition shown in eq 1 is the most
frequently encountered reaction in traditional main-group metal
chemistry. Nowadays, however, the reaction of this type is being
opened in palladium chemistry; vinylpalladium halides, generated
from aryl halides, alkynes, and catalytic amounts of palladium,
underwent intramolecular nucleophilic addition to aryl ketones,2a
aldehydes,2b and nitriles.3 We now wish to report that the
intramolecular nucleophilic addition of aryl bromides 1 to ketones
proceeds very smoothly in the presence of palladium catalyst to
give the corresponding cyclic alcohols 2 in good to high yields
(eq 2). This is a Grignard-type reaction using a palladium
catalyst.4
The reaction of the o-bromophenyl ketone 1a was examined
in the presence of Pd(OAc)2 (5 mol %) and KOAc (2 equiv) under
argon in DMF (0.25 M), as a logical extension of the previous
findings.2 However, a complex mixture of products was obtained,
and none of the cyclization product 2a was detected. Accidental
addition of 1-hexanol to the catalyst system brought about the
formation of small amounts of 2a. Then, we investigated the effect
of phosphine ligands. Among many ligands examined, tricyclo-
hexylphosphine gave the best result: 41% yield of 2a using Pd-
(OAc)2/P(Cy)3/KOAc/1-hexanol. Next, we searched a better base
and found that insoluble Na2CO3 gave the best result: 73% NMR
yield of 2a using Pd(OAc)2 (5 mol %)/P(Cy)3 (10 mol %)/Na2-
CO3 2 equiv)/1-hexanol (5 equiv). Other palladium catalysts, such
as Pd(PPh3)4, Pd2(dba)3‚CHCl3, and PdCl2(PPh3)2, were not
a Isolated yield except otherwise indicated. b As a by-product, the
dehalogenated aryl ketones were formed in 10-35% yields. For
example, in entry 1, PhCH2CH(CH3)COPh 3a was produced in 20%
yield. c The stereochemistry was determined by NOE experiments (see
Supporting Information). “Cis” stands for the cis stereochemical relation
d
between Me and OH group. 1H NMR yield and here 3a was obtained
in 57% yield. e 10 mol % of PBu3 was used in place of PCy3.
effective.5 In the absence of bases, no cyclization products were
obtained. The cyclization was quite inefficient in the absence of
1-hexanol. 1-Hexanol was recovered after the reaction was over,
and other aliphatic alcohols such as 1-pentanol and 1-propanol
were equally effective.6
The results of the cyclization of various bromo ketones are
summarized in Table 1. The arylpalladation of 1a in the presence
of Pd(OAc)2 (5 mol %), PCy3 (10 mol %), Na2CO3 (2 equiv)
and 1-hexanol (5 equiv) under argon atmosphere in DMF (0.25
M) at 135 °C gave the indanol 2a in 69% isolated yield as a
single diastereomer along with 20% yield of the corresponding
dehalogenation product 3a7 (entry 1). The use of the iodide 1a′,
instead of the bromide 1a, decreased the yield of 2a from 69%
to 41% and increased the yield of 3a from 20 to 57% (entry 2).
No reaction took place in the case of the corresponding chloride.
The arylation of the ketone 1b, which contains no hydrogen R to
(1) For reviews, see for example: (a) Trost, B. M.; Verhoeven, T. R.
ComprehensiVe Organometallic Chemistry; Wilkinson, G.; Stone, F. G. A.,
Abel, E. W., Eds.; Pergamon Press: New York, 1982; Vol. 8, p 799. (b)
Collman, J. P.; Hegedus, L. S.; Norton, J. R.; Finke, R. G. Principles and
Applications of Organotransition Metal Chemistry; University Science
Books: Mill Valley, CA, 1978. (c) Tsuji, J. Palladium Reagents and Catalysts;
John Wiley: Chiester, 1995. (d) Negishi, E.; Cope´ret, C.; Ma, S.; Liou, S.-
Y.; Liu, F. Chem. ReV. 1996, 96, 365.
(2) (a) Quan, L. G.; Gevorgyan, V.; Yamamoto, Y. J. Am. Chem. Soc.
1999, 121, 3545. (b) Gevorgyan, V.; Quan, L. G.; Yamamoto, Y. Tetrahedron
Lett. 1999, 40, 4089.
(3) Larock, R. C.; Tian, Q.; Pletnev, A. A. J. Am. Chem. Soc. 1999, 121,
3238.
(4) Intramolecular coupling of less hindered ketones with aryl halides using
excess amounts of magnesium is known. See: (a) Cacchi, S.; Palmieri; G. J.
Organomet. Chem. 1985, 282, C3. We investigated the cyclization reaction
of 1a using magnesium under the same reaction conditions as those mentioned
in the literature. The product 2a was obtained in only 26% 1H NMR yield
along with complex by-products. Furthermore, the Mg reaction of 1g did not
proceed at all. Intramolecular coupling of cycloketones with aryl halides using
2 equiv of Bu3SnSiMe3 and 2 equiv of R4NX is known. See: (b) Mori, M.;
Kaneta. N.; Shibasaki, M. J. Organomet. Chem. 1994, 464, 35.
(5) Ph3P ligand was not suitable to the present reaction. For example, not
only Pd(PPh3)4 but also Pd(OAc)2/PPh3 system was not effective. On the other
hand, aliphatic phosphine ligands were suitable: Pd(OAc)2/PBu3 was also
effective.
(6) MeOH and EtOH were not examined since their boiling points were
lower than the reaction temperatures.
(7) Reduction of aryl halides to the corresponding dehalogenated arenes
in the presence of palladium catalyst and base, see: (a) Zask, A.; Helquist, P.
J. Org. Chem. 1978, 43, 1978. (b) Tamaru, Y.; Yamamoto, Y.; Yamada, Y.;
Yoshida, Z. Tetrahedron Lett. 1979, 16, 1401.
10.1021/ja000415k CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/29/2000

4828 J. Am. Chem. Soc., Vol. 122, No. 19, 2000
Communications to the Editor
ox
Scheme 1
more negative redox potentials E2red/E2 ) -417/-244 mV
(Figure 2 in SI), which may be attributed to a Pd(II) species
complexed by CO32- and DMF. The resulting Pd(II) species was
not stable and led spontaneously to a species detected by its
irreversible oxidation peak C at 1389 mV (Figure 2). The
oxidation peak current of C increased with time concomitantly
with decrease of the reduction peak current of A′. After 2 h under
argon, both peaks disappeared, and no electroactive species was
detected by CV: black precipitates were observed which presum-
ably corresponded to Pd black. When P(Cy)3 (2 equiv) was added
together with Na2CO3 (40 equiv), complete reduction of Pd(II)
to Pd(0) was observed after 57 min: The peak A′ was changed
to a new irreversible oxidation peak D at less positive potential
1000 mV, due to the electron-donating effect of P(Cy)3 which
coordinates to the Pd(0) species (Figure 3 in SI). The resulting
Pd(0), peak D, was stable for a prolonged period of time, and as
can be seen from Figure 3 there was no Pd(II) species under these
conditions, showing that the catalytic cycle starts by the oxidative
insertion of this Pd(0) to the aryl-bromide bond of 1. To the above
mixture containing only Pd(0), 1b (20 equiv) and 1-hexanol (100
equiv) were added, leading to a change of the previous voltam-
mogram. The peak D was replaced by a new redox couple E and
F, E1red/E1ox ) -635/-201 mV, corresponding to a Pd(II) species,
and in the meantime a new Pd(0) species was detected by its
peak J at 1108 mV (Figure 4 in SI). The peak E most probably
corresponds to either 5 or 6, and J corresponds to 7. The pattern
of voltammograms of Figures 1-3 did not change significantly
by addition of 1-hexanol.
a carbonyl group, proceeded smoothly to give the indanol 2b in
82% yield (entry 3). The aliphatic ketone 1c underwent the
arylation similarly to produce 2c in 76% yield (entry 4). Even
when the sterically bulky naphthyl bromide 1d was employed as
a substrate, the reaction proceeded without any problem to afford
the cyclized product 2d in 76% yield with a high diastereomeric
selectivity (entry 5). The biaryl ketones 1e and 1f reacted very
smoothly to give the corresponding cyclized products 2e and 2f,
respectively, in high yields (entries 6, 7). The aryl ketone 1g
bearing an electron-donating group at the position para to bromine
gave the cyclized product 2g in 75% yield (entry 8). The
usefulness of this methodology was further demonstrated in the
preparation of six-member ring products. Thus, the aryl ketones
1h and 1i underwent the cyclization to produce the corresponding
cyclized products, 2h as a single diastereomer and 2i, respectively,
in good to moderate yields (entries 9, 10). The diketone 1j having
an electron-withdrawing group as a tether, gave 9-hydroxy-9-
phenylanthrone 2j8 in 89% yield (entry 11). The reaction of the
aldehyde 4a under the same reaction conditions as those above
gave a complicated mixture of products, and the reaction of the
corresponding chloride 4b resulted in the recovery of the starting
material.
To understand the crucial role of insoluble Na2CO3, we carried
out the experiment with Li2CO3 more soluble in DMF. The use
of Li2CO3, instead of Na2CO3, did not cause any reaction of 1b;
it was recovered. The CV of Li2CO3 in DMF led to a nonrevers-
ible oxidation peak K at -45 mV (Figure 5 in SI). The addition
of P(Cy)3 to this solution gave immediately a new peak L at
-1618 mV which corresponded to the reduction of (Cy)3PdO.
It was confirmed by GC-MS analysis that (Cy)3PdO was
produced by mixing P(Cy)3 with Li2CO3 in DMF at room
temperature. Most probably, the electrochemical oxidation of
2-
CO3 would produce O2 which may oxidize the phosphine to
the corresponding phosphine oxide.9 Under the reaction conditions
of the palladium-catalyzed reaction of 1b, more soluble Li2CO3
would react readily with P(Cy)3 and therefore there would be no
ligand which stabilizes Pd(0) complex. In the case of Na2CO3,
the reaction with P(Cy)3 was very slow due to its insolubility,
but Pd(II) is electron-deficient and thus would be more easily
reduced by Na2CO3 than P(Cy)3. The use of 1-hexanol is not yet
clear, but it does not directly participate in the redox process of
Pd since the above voltammograms did not change significantly
in the presence or absence of 1-hexanol. One possibility is that
the alkoxide exchange between 6 and 1-hexanol, leading to 2 and
less bulky HexOPdX, might facilitate the reaction of Pd(II) to
Pd(0).
To obtain a partial evidence for the proposed mechanism shown
in Scheme 1, we carried out the electrochemical studies of the
reaction system. A Pd(II) species in DMF solution of Pd(OAc)2
was detected electrochemically as its quasireversible peak A (and
ox
B) in CV at E1red/E1 ) -381/-236 mV (Figure 1 in supple-
mental information). Immediately after 40 equiv of Na2CO3 was
added, which was exactly same condition as the synthetic process,
no change was observed in the pattern of the above voltammogram
perhaps due to low solubility of the base in the medium. Then,
the peak A (and B) was shifted to a new peak A′ (and B′) with
Extension of the Grignard-type reaction using a palladium
catalyst is under active investigation in our laboratries.
Supporting Information Available: Spectroscopic and analytical data
for compounds 2a-e, g-i (PDF). This material is available free of charge
via the Internet at http://pubs.acs.org.
(8) Barnett, W. E.; Needham, L. L.; Powell, R. W. Tetrahedron, 1972, 28,
419.
2-
(9) The following oxidation of CO3
produces oxygen:
CO3 -2e8 CO + O2 (or CO2 + 1/2O2)
JA000415K
2-