A R T I C L E S
Rousseaux et al.
°C (entry 3), complete conversion of 1a was reached, and BCB
2 was obtained in 89% GC yield (83% isolated yield). No trace
of the protodehalogenated product 2′, the usual side product in
these reactions, was observed. Other catalysts and solvents
provided different outcomes for the reactions of 1a and 1b. In
particular, tri-tert-butylphosphine was by far the most efficient
ligand with 1a among a panel of monophosphine ligands (entries
3-9), whereas bulkier monophosphine ligands such as JohnPhos,
Q-Phos, and P(o-tol)3 gave good results with 1b.5d Thus, PtBu3
appears to have the optimal electronic and steric properties
(Tolman cone angle ) 182°) for the intramolecular C-H
arylation of chloride 1a.13,14 Palladium acetate (entry 3) was
also the most active palladium precatalyst (see entries 10 and
11 for examples of other Pd sources). Finally, solvents other
than DMF could be employed for the arylation of 1a, including
N,N-dimethylacetamide (DMA) and N-methylpyrrolidone (NMP)
(entries 12 and 13), but DMF gave the highest yield of isolated
BCB. Interestingly, using mesitylene as the solvent with 30 mol
% of pivalic acid as an additive, a combination that gave good
results in other intramolecular C(sp3)-H arylations (Vide infra),6
led to irreproducible yields of 2 (entry 14).
The optimal set of conditions was applied to the C(sp3)-H
arylation of a wide range of aryl and heteroaryl chlorides (Table
2). As anticipated, the presence of the chlorine atom greatly
facilitated the synthesis of most arylation substrates due to the
better availability of starting materials and to the reduced
influence of steric hindrance.15 For instance, dichloride 15 was
prepared in one step from a commercially available material
using a stoichiometric amount of reagents, whereas four steps
and an excess in reagents were required for the corresponding
bromide.5d Moreover, in some cases, for example chlorides 13,
27, and 29, the corresponding aryl bromide would be almost
inaccessible by typical synthetic routes.
The standard arylation conditions were found to be compatible
with a wide range of substituents on the benzylic carbon (entries
1-6) and the aromatic ring (entries 7-12), giving rise to the
corresponding BCBs in 65-87% yield. In the case of substrates
with mutiple arylation sites (entries 4 and 5), the reaction showed
complete regioselectivity for arylation at the methyl group. In
the case of dichloride 15 (entry 8), careful monitoring of the
reaction conversion allowed BCB 16 to be obtained in good
yield after 30 min, whereas with prolonged heating a mixture
of 16 and the corresponding dechlorinated product was obtained.
Arylation of chloride 19 gave rise to a 3.6:1 mixture of BCB
isomers 20a,b (entry 10). The formation of isomer 20b, which
was also observed from the corresponding aryl bromide,5d
presumably arises from a palladium migration within the
catalytic cycle.16 Indeed, DFT calculations have demonstrated
that reductive elimination to form the strained BCB four-
membered ring is highly energetic and thus is in competition
with palladium migration.5d The reaction conditions were also
effective for heteroaryl chlorides. Indeed, chloropyridine 25
(entry 13), chloroquinoline 27 (entry 14), and chloroindole 29
(entry 15) yielded valuable heterocycles that are typically
unaccessible through more traditional synthetic methods. For
quinoline 27 (entry 14), the reaction was conducted at a lower
temperature (110 °C) to avoid the formation of a side product
arising from cyclobutene ring-opening.17 From indole 29 (entry
15), a 2:5 separable mixture of isomers 30a and 30b, arising
from cyclization at the C-5 and C-3 position of the indole
nucleus, respectively, was obtained. To the best of our knowl-
edge, compound 30b is surprisingly the first isolated molecule
with the 3,4-dihydro-1H-cyclopenta[c,d]indole scaffold, which
is structurally related to the orchid alkaloid dendrobine.18 This
isomer most likely arises from palladium migration, similar to
that observed with aryl chloride 19 (entry 10). Research is
ongoing to further study this rearrangement in detail.
2. Synthesis of Indanes. In a previous report, some of us
described the synthesis of indanes from aryl bromides by
intramolecular C(sp3)-H arylation at the terminal carbon of an
isopropyl group.5b For instance, the reaction of bromide 31b
using 5 mol % Pd(OAc)2 and 20 mol % F-TOTP as catalyst in
DMF at 100 °C furnished a 4:1 mixture of indane 32a (as a
single diastereoisomer) and the dehydrogenation product 33
(Table 3, entry 1). This catalyst was found to be ineffective in
the reaction of the corresponding aryl chloride 31a, even at 140
°C (entry 2). A rapid ligand screen (entries 3-9) revealed that
trialkylphosphines with cone angles between 160 and 170° have
optimal steric properties for this reaction,14 with PCy3, PCyp3,
and PiPr3 giving rise to an inseparable diastereoisomeric mixture
of 32a:32b (dr ) 4:1 in favor of the cis diastereoisomer 32a)19
in high yield and with no trace of olefin 33 (entries 5-7).
Interestingly, lower selectivity of indane vs olefin product with
concomitant higher diastereoselectivity was obtained with PtBu3,
compared to less bulky trialkylphosphines (entry 3). Tricyclo-
pentylphosphine, which gave the highest isolated yield of
diastereoisomers 32a,b (92%, entry 6), was retained for further
scope studies (Vide infra). Decreasing the temperature to 120
°C led to incomplete conversion (entry 9). Finally, replacing
DMF by other solvents such as DMA or mesitylene (with 30
mol % of pivalic acid as an additive)6 gave lower yields of
indane products.
The optimal conditions were applied to a range of other aryl
and heteroaryl chlorides (Table 4). Again, the ease of synthesis
of most arylation substrates was significantly improved com-
pared to that of the corresponding aryl bromides. The presence
of electron-withdrawing (entries 2 and 3) or -donating (entry
4) substituents on the aromatic ring did not affect the efficiency
of the reaction, and the corresponding indanes were obtained
in very good yields as 3.3:1 to 4:1 diastereoisomeric mixtures,
always in favor of the cis diastereoisomer.19 Interestingly, no
isomerized product arising from palladium migration was
observed in the reaction of aryl chloride 34 (entry 2), in contrast
to BCB formation from the analogous substrate 19 (Table 2,
entry 10). This observation indicates that reductive elimination
to form the indane five-membered ring is faster than palladium
migration. Remarkably, the intramolecular arylation of homo-
chiral trans-dimethylcyclopentane compounds 40 and 42 fur-
nished tricyclic products 41 and 43, respectively, in very good
yield and as a single stereoisomer (entries 5 and 6). The reaction
(13) (a) Tolman, C. A. Chem. ReV. 1977, 77, 313. (b) Brown, T. L.; Lee,
K. J. Coord. Chem. ReV. 1993, 128, 89. (c) Clavier, H.; Nolan, S. P.
Chem. Commun. 2010, 46, 841.
(17) Jefford, C. W.; Bernardinelli, G.; Wang, Y.; Spellmeyer, D. C.; Buda,
A.; Houk, K. N. J. Am. Chem. Soc. 1992, 114, 1157.
(14) Tolman cone angles θ:13 JohnPhos, 246°; CyJohnPhos, 226°; P(o-
tol)3, 194°; PtBu3, 182°; PCy3, 170°; MePtBu2, 161°; PiPr3, 160°; PnBu3,
132°.
(18) (a) Yamamura, S.; Hirata, Y. Tetrahedron Lett. 1964, 5, 79. (b)
Inubushi, Y.; Sasaki, Y.; Tsuda, Y.; Yasui, B.; Konita, T.; Matsumoto,
J.; Katarao, E.; Nakano, J. Tetrahedron 1964, 20, 2007. (c) Onaka,
R.; Kamata, S.; Maeda, T.; Kawazoe, Y.; Natsume, M.; Okamoto, T.;
Uchimaru, F.; Shimizu, M. Chem. Pharm. Bull. 1964, 12, 506.
(19) Relative configurations were ascribed by NOESY experiments.
(15) See the Supporting Information for detailed experimental procedures.
(16) Reviews on palladium migrations: (a) Ma, S.; Gu, Z. Angew. Chem.,
Int. Ed. 2005, 44, 7512. (b) Shi, F.; Larock, R. C. Top. Curr. Chem.
2010, 292, 123.
9
10708 J. AM. CHEM. SOC. VOL. 132, NO. 31, 2010