.
Angewandte
Communications
Table 1: Optimization studies for the formation of benzo[b]fluorene 2a.
second alkyne terminus. Substrate 1 f with an electron-
donating (p-Me) group on the aromatic ring worked effi-
ciently and furnished the corresponding product 2 f in 74%
yield, whereas substrate 1g with an electron-withdrawing
substituent (p-Br) gave the cyclization product 2g in only
34% yield along with 37% of naphthyl ketone 3g. The low
yield observed for 2g is attributed to the reduced nucleophi-
licity of the aromatic ring. The ortho-bromo-substituted
substrate 1h did not give the expected product, presumably
owing to steric effects. Interestingly, this method could be
used for the synthesis of the thienyl-fused polycycle 2i by
using the substrate 1i with a thienyl sustituent (R3). The
reactions proved to be quite general with respect to sub-
stitution of R2 on the first alkyne terminus, since aryl and alkyl
groups were all suitable for this substituent, thereby showing
a wide diversity of the products. Functional groups such as
-Me, 3,4,5-trimethoxy, 1-naphthyl, -Br, -CF3, and cyclopropyl
groups were well tolerated (2j–2n, 2p). The reactions with
bulky tBu-substituted diynes were also satisfactory and lead
to 2q and 2r in good yields. Furthermore, substrate 1s,
carrying a fluoro substituent on the parent phenyl ring, was
also compatible for this transformation, and 66% yield for
product 2s was realized. It should be noted that in some cases
as indicated in Table 2, the addition of molecular sieves was
necessary. In the absence of molecular sieves, these reactions
were either not clean or resulted in lower yields of the desired
products. The structure of benzo[b]fluorene was further
confirmed by X-ray crystallographic analysis of 2s.[11]
Entry Catalyst (mol%)
Solvent
t
Yield of 2a[a]
[h]
[%]
1
2
3
4
A (3)
A (3)
A (5)
A (5)
CH2Cl2
ClCH2CH2Cl
toluene
benzene
o-xylene
toluene
1
1
2
1
2
77[b]
77[b]
78
74
60
5
A (5)
6[c]
7[d]
8
[JohnphosAuCl]/AgPF6 (5)
[CyJohnphosAuCl]/AgSbF6 (5) toluene
[PPh3 AuCl]/AgSbF6 (5)
[tBu3PAuCl]/AgSbF6 (5)
AgSbF6 (5)
2.5 59
4
1
1
12
[e]
[e]
[e]
–
–
–
toluene
toluene
toluene
9
10
NR[f]
[a] Yields of isolated product. [b] Contaminanted with a small amount of
impurity. [c] Johnphos=(2-biphenyl)di-tBu-phosphine. [d] CyJohn-
phos=(2-biphenyl)di-cyclohexyl-phosphine. [e] Several products were
formed. [f] NR=No reaction.
combination of [JohnphosAuCl] with AgPF6 gave a lower
yield of 59% (Table 1, entry 6). We also examined the
influence of various phosphine ligands in this transformation.
CyJohnphos, PPh3, and P(tBu)3 all failed to give clean
reactions (Table 1, entries 7–9). The results indicate that the
nature of the phosphine ligand plays an important role in
controlling the chemoselectivity of this reaction. AgSbF6
alone did not promote any transformation (Table 1,
entry 10). For comparison, we also prepared the OAc-
protected substrate 3-phenyl-1-[2-(phenylethynyl)phenyl]-
prop-2-ynyl acetate. It was found that several products were
formed catalyzed by catalyst A (5 mol%) in toluene by using
this substrate; among these products was naphthyl ketone
phenyl(3-phenylnaphthalen-2-yl)methanone 3a, which was
derived by hydration of the oxocarbenium ion intermediate
and was isolated in 31% yield. The results imply that the
protecting group on the alcohol can strongly affect the
reaction pathway.
To understand the reaction mechanism, we tried to
determine and isolate the possible reaction intermediates.
To our delight, when tBu-substituted compound 1r was used
as a substrate, the carbonate product 4r was obtained in 43%
yield in the presence of catalyst A (1 mol%), together with
33% of decarboxylated product 2r [Scheme 2, Eq. (1)].
With the optimized reaction conditions in hand, the scope
of this cascade cyclization reaction was investigated. As
shown in Table 2, the method is applicable to a wide range of
suitably substituted 1,6-diynyl carbonates. We first examined
the substituent effect (R1) on the carbonate groups. The allyl,
alkyl, and benzyl groups are all compatible under the
cyclization conditions, leading to products 2a–2c and 2e in
66–85% yields. When R1 is a sterically more demanding alkyl
group, the efficiency of this reaction decreased. For example,
the methyl-substituted substrate 1b afforded 2b in 85% yield.
When the methyl group in the carbonate moiety was replaced
by an iPr group, the yield of the cyclization product 2c
decreased to 66%. Substrate 1d with a tert-butyl substituent
could not deliver the desired benzofluorene product, but only
the product 2-naphthyl ketone 3a in 77% yield. Clearly the
bulkiness of the substituent on the carbonate moieties had
a detrimental effect on the reactivity. Next, we investigated
the electronic effects of the arene substitution (R3) on the
Scheme 2. Determination of the reaction intermediates.
Control experiments with 4r catalyzed by catalyst
A
(5 mol%) in the presence of 5 ꢀ MS indicated that only
a trace amount of 2r was formed at room temperature;
however, upon heating a toluene solution of 4r at 708C, the
decarboxylation indeed occurred to give 2r in 70% yield
[Scheme 2, Eq. (2)]. Without the gold catalyst, no reaction
occurred. The results disclosed that gold could catalyze the
decarboxylative etherification of benzylic carbonates. The
reason why a higher reaction temperature than in the one-pot
procedure was required is not clear yet.
A crossover experiment was also performed. Treatment of
a 1:1 mixture of two diynes, 1a and 1r, bearing different
protecting groups under the catalytic conditions shown in
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 6493 –6497