Table 1: Optimization studies on the rearrangement of 4a.[a]
The connectivity of 5a can easily be explained by the
mechanism[13] depicted in Scheme 3. After coordination to the
triple bond (A),[14] the gold catalyst induces an electrophilic
Entry Catalyst
t [h] Yield [%]
1[b]
2[c]
3
4
5
[Ph3PAuNTf2]
42
3
48
42
42
1
8
48
1
87
58
68
19
23
91
93
18
[NACAuCl]/AgSbF6
[NACAuCl]/AgNTf2
AuCl
AuCl3
6
[IPrAuCl]/AgNTf2
[IPrAuCl]/AgNTf2
[IPrAuCl]/AgNTf2
[IPrAuCl]/AgNTf2
[MePhosAuCl]/AgNTf2 24
AgNTf2
7[d]
8[e]
9[f]
10
11
12
85
92
n.r.
Scheme 3. Proposed mechanism for the gold(I)-catalyzed rearrange-
ment of the substrate 4a to the 2,7-disubstituted benzo[b]furan 5a.
42
48
p-TsOH
decomposition of the substrate
[a] Reaction conditions: Substrate (200 mmol), [Au] (2 mol%), [Ag]
(2 mol%), 1.1 equiv iPrOH, CH2Cl2 (3 mL), RT, in air. The reaction was
monitored by TLC. [b] Substrate (300 mmol). [c] Substrate (100 mmol).
[d] 0.5 mol% of catalyst. [e] 0.1 mol% of catalyst. [f] Without iPrOH. Tf=
trifluoromethanesulfonyl, n.r. =no reaction.
attack at the most nucleophilic position of the furan ring, the
2-position (B). After this 5-endo-dig cyclization, a Wagner–
Meerwein shift delivers the intermediate C with a more stable
carboxonium ion. Rearomatization of the furan by deproto-
nation and protodeauration delivers D, and a subsequent
aromatization by elimination of silanol affords the final
product 5a.
Crucial for the substituent “castling” is the selective
migration of the sp3-carbon atom rather than the sp2-carbon
atom in the spiro intermediate B, which probably originates
from stabilization of the positive charge in the transition state
of the 1,2-shift of the oxygen atom from the 2- to the 3-
position (carboxonium-like stabilization). Other gold-cata-
lyzed conversions involving furan-derived arenium intermedi-
ates have so far only shown a 1,2-shift from the 3- to the 2-
position in E[15] or a rearomatization by opening of a three-
catalyst loading to 0.5 mol% gave a respectable 93% yield of
the product after 8 h (Table 1, entry 7). With only 0.1 mol%
of catalyst, the yield dropped to 18% after 48 h, no more
conversion was observed after that time (Table 1, entry 8).
Without iPrOH as the additive, the yield was reduced by 6%
(Table 1, entry 9). An even better yield of 92% was obtained
with the phosphane ligand MePhos, but the reaction needed
24 h (Table 1, entry 10). Control experiments with silver(I)
gave no conversion (Table 1, entry 11), while with p-toluene-
sulfonic acid (p-TsOH) the usual slow decomposition of the
furan was observed (Table 1, entry 12).
Initially one could have expected the benzofuran to be the
product of a normal hydroarylation of the furan ring in the 3-
position[3b] and a subsequent aromatization by elimination of
silanol.[11] This would deliver the 2,4-disubstituted benzofuran
6a. A safe assignment of the structure of anellated disub-
stituted benzofurans by NMR spectroscopy is difficult, but
more reliable results were obtained by X-ray crystal structure
analysis of the benzofuran product.[12] Thus, the 2,7-disubsti-
tuted structure of 5a was proven unambiguously (Figure 2).
membered ring in the spiro intermediate of type
F
(Scheme 4).[16]
Apart from the intriguing mechanism, we were also
interested in the scope of this conversion. A series of
substrates 4a–4i was easily available by the reaction of
furfurals with propargyl bromide and zinc, silyl protection of
Figure 2. Left: Conceivable structure of benzofuran 6a. Middle: Solid-
state molecular structure of 5a. Right: Solid-state molecular structure
of the sulfur-containing product 5i. Thermal ellipsoids at 50%
probability.
Scheme 4. Furyl-derived Wheland-type intermediates according to
Kirsch et al. and Schmalz and co-workers. Top: R1 =aryl, alkyl;
R2 =alkyl; R3 =aryl, alkyl; bottom: R1 =aryl, alkyl; R2 =aryl, alkyl;
Nu=R3O.
Angew. Chem. Int. Ed. 2011, 50, 5762 –5765
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5763