and iodide anions on the reaction outcome (Table 1). The
best result was obtained by increasing the TBAI:1a ratio to
2:1 (entry 3). Unsatisfactory results were obtained using
Pd(OAc)2 (entry 9) or Pd2(dba)3, with or without PPh3
(entries 10 and 11), as well as decreasing the reaction
temperature (entries 4 and 5). K2CO3 proved superior to
Cs2CO3 or K3PO4 (entry 3 vs entries 12 and 13, respectively).
However, we subsequently found that using Cs2CO3 affords
slightly higher or similar yields in some cases (Table 2,
entries 2 and 4 vs entries 3 and 5, respectively). Therefore,
it seems advisible that the effectiveness of the bases be
evaluated each time. Substituting NaI for TBAI produced
3a in low yield (entry 6), and the use of KI in the presence
of catalytic amounts (entry 7) or 1 equiv (entry 8) of TBAI
proved unsuccessful as well.
Scheme 2
We set out to use the reaction of 1a with 2a as a probe
for evaluating the feasibility of the reaction. First attempts,
however, met with failure. For example, complex mixtures
(which we did not further investigate) were obtained under
conditions commonly used with vinyl triflates,5 aryl iodides,5
bromides,6 and triflates6 [Pd(PPh3)4, K2CO3, MeCN], and no
indole formation was observed. We envisaged that 2,3-
disubstituted indoles might be accessible by reacting arene-
diazonium tetrafluoroborates with 2-alkynyltrifluoroaceta-
nilides in the presence of TBAI provided that a iododediazoni-
ation reaction7 generating “ArPdI” species could be achieved
under the reaction conditions.
The optimal reaction conditions were then employed when
the reaction was extended to other arenediazonium salts and
2-alkynyltrifluoroacetanilides (Table 2). The corresponding
indole derivatives were isolated in good to excellent yields
with a variety of neutral and electron-rich arenediazonium
salts. With electron-poor arenediazonium salts similar or
better results could be obtained when a TBAI:1 1.5:1 ratio
was used (entries 4 and 5 vs entry 6). The reaction tolerates
a variety of useful substituents both in the starting alkyne
and the arenediazonium salt component, including bromo
and chloro substituents, nitro, cyano, keto, ester, and ether
groups. The ability to incorporate bromo and chloro sub-
stituents makes this reaction particularly attractive for
increasing the molecular complexity, for example, via
transition metal-catalyzed coupling reactions. Arenediazo-
nium salts containing ortho substituents such as methoxy and
methyl groups also give the corresponding indole products
in good to high yields (entries 10, 12, and 30). Formation of
3-unsubstituted indoles 4 was in some cases observed.
To demonstrate this concept, we treated 2 equiv of 1a with
1 equiv of 2a in the presence of 5 mol % of Pd(PPh3)4 and
2 equiv of K2CO3 in MeCN at 60 °C for 2.5 h using a TBAI:
1a ratio of 1.2:1. Pleasingly, 3a was isolated in 40% yield
along with a 50% yield of 4-iodoanisole, very likely formed
via a iododiazoniation reaction (Table 1, entry 1).
Table 1. Optimization Studiesa
As to the mechanism, we believe that the reaction proceeds
through a domino process that starts with a iododediazonia-
tion step (Scheme 3, path a).8,9 The oxidative addition of
the resultant aryl iodide to Pd(0) affords a σ-arylpalladium
iodide that coordinates to the C-C triple bond to give a
π-alkyne-σ-arylpalladium complex C. A subsequent intramo-
lecular aminopalladation step generates a σ-indolylpalladium
intermediate D from which the desired free N-H indole is
formed via reductive elimination and hydrolysis (not neces-
sarily in this order).3
yield (%)b
temp time
entry
[Pd]
additive (equiv) (°C)
(h)
3a 4a 5ac
1
2
3
4
5
6
7
8
9
Pd(PPh3)4 TBAI (2.4)
Pd(PPh3)4 TBAI (3)
60
60
60
40
50
60
60
60
60
2.5
40
50
19
2.75 52d
Pd(PPh3)4 TBAI (4)
1
12
69
Pd(PPh3)4 TBAI (4)
e
Pd(PPh3)4 TBAI (4)
3.45 52f
45
Pd(PPh3)4 NaI (2.2)/I2 (0.2)
Pd(PPh3)4 KI (4)/ TBAI (0.2)
Pd(PPh3)4 KI (4)/ TBAI (1)
Pd(OAc)2 TBAI (4)
3.5
30 11 58
2.75 20 35 45
3.25 32 19 49
The intervention of an alternative mechanism involving
the formation of σ-arylpalladium iodides through the reaction
of σ-arylpalladium cations E with iodide anions (Scheme 3,
path b) appears less likely. This view is supported by the
observation that no indole formation was observed when 1a
and 2a were subjected to the best conditions found substitut-
ing TBACl for TBAI, even prolonging the reaction time to
6 h (2a was recovered in 76% yield). If σ-arylpalladium
chlorides were generated via trapping of E10 with chloride
anions, the known reluctance of aryl chlorides to undergo
oxidative addition reactions with palladium catalysts coor-
dinated to triphenylphosphine11 should be circumvented and
indoles should form to some extent.
5
13 42 27
Pd2(dba)3,
d
10
11
12
13
PPh3
TBAI (4)
60
60
60
60
3
38g 13 42
Pd2(dba)3 TBAI (4)
Pd(PPh3)4 TBAI (4)
Pd(PPh3)4 TBAI (4)
4.25 48
30
58h tr 30
50i
35
1
5
6
a Unless otherwise stated, reactions were carried out at 60 °C on a 0.35
mmol scale using 2 equiv of 1a, 1 equiv of 2a, 2 equiv of K2CO3, and 5
mol % of [Pd] in 3 mL of anhydrous MeCN in the presence of a iodide
source. b Yields are given for isolated products. c Calculated on 1a. d 1.5
equiv of 1a was used. e At 40 °C. f At 50 °C. g 4 equiv of PPh3. h In the
presence of 2 equiv of Cs2CO3. i In the presence of 2 equiv of K3PO4.
We then started an optimization study to explore the
influence of bases, temperature, the source of Pd(0) species,
3280
Org. Lett., Vol. 12, No. 14, 2010