Communication
tions are applicable under both copper and silver co-catalysis,
regardless of the nature of the halide (Table 1, entries 3–6).
However, aryl iodides are better coupling partners under the
Ag-mediated reaction conditions, whereas the Cu-catalyzed
transformation works well with aryl bromides (Table 1, entries 3
and 6). We then observed that two equivalents of acid 1a are
required to achieve an efficient Pd0/AgI-catalyzed decarboxyla-
tive reaction (Table 1, entries 3 and 7). On the other hand, low-
ering the amount of Cu2O from stoichiometric to catalytic
(20 mol%) amounts, along with decreasing the amount of acid
1a from 2 to 1.5 equivalents, did not affect the efficiency of
the Pd0/CuI-catalyzed decarboxylative process (Table 1, en-
tries 8–10). Interestingly, the optimized protocols were also
easily scaled up from 0.3 to 1.5 mmol without a large decrease
in yield (Table 1, entries 3 and 6).
Figure 1. Possible mechanisms for the decarboxylative cross-coupling of
azine N-oxides with aryl halides.
To determine the catalytic cycle that operates between the
conventional decarboxylative arylation and the protodecarbox-
ylation/CÀH arylation (Figure 1, cycles A and C or B and C),
direct CÀH arylations of quinoline N-oxide were performed
under both copper- and silver-mediated decarboxylative aryla-
tion procedures.[24] The expected product, 3a, was obtained in
a very poor 12% yield with p-tolylbromide as a coupling part-
ner under the copper-mediated procedure, whereas with silver
as co-catalyst, C2 arylated 3a was synthetized in 17 and 93%
yields with p-tolylbromide and -iodide, respectively. With these
results, we can suggest that both decarboxylative coupling
methods proceed mainly through the conventional decarboxy-
lative cross-coupling mechanism pathway (Figure 1, cycles A
and C).
Table 1. Pd-catalyzed decarboxylative cross-coupling of quinaldic acid N-
oxide 1a under various reaction conditions.[a]
Entry
Additive [(equiv)]
p-Tol-X
Solvent
Yield [%][b]
1
2
3
4
5
6
7
8
9
Ag2CO3 (1.0)
Cu2O (1.0)
Ag2CO3 (1.0)
Ag2CO3 (1.0)
Cu2O (1.0)
Cu2O (1.0)
Ag2CO3 (1.0)
Cu2O (1.0)
Cu2O (0.5)
Cu2O (0.2)
I
Br
I
Br
I
Br
I
Br
Br
Br
DMF
DMF
56
40
91, 79[d]
78
1,4-dioxane
1,4-dioxane
1,4-dioxane
1,4-dioxane
1,4-dioxane
1,4-dioxane
1,4-dioxane
1,4-dioxane
With the optimized conditions in hand, the scope of the de-
carboxylative coupling was undertaken by using quinaldic acid
N-oxides 1a and 1b with a broad range of aryl halides, bearing
electron-donating or -withdrawing groups (Scheme 2). We
were pleased to observe the formation of the desired C2-
arylated products, 3b–k and 4a,b, in reasonable to good
yields, with (hetero)aryl iodides under the silver-mediated pro-
cess, as well as with (hetero)aryl bromides and chlorides under
copper co-catalysis (Scheme 2). Remarkably, the electronic and
steric hindrance effects of both coupling partners have no in-
fluence on the success of the decarboxylation reaction.
To further examine the versatility of this methodology, the
decarboxylative arylation was attempted with the isoquinolinic
acid N-oxides 5 and 6 (Scheme 3). Unfortunately, the desired
product could not be detected under the above optimized
conditions owing to a lack of solubility of the acid in 1,4-diox-
ane. Nevertheless, switching the solvent to DMF afforded the
expected arylated isoquinolinic N-oxides 7 and 8 in good
yields under the silver-mediated process (Scheme 3, [Eq. (1)]).
Remarkably, the transformations were regiospecific at the car-
boxy-position site. The exclusive formation of C3-arylated iso-
quinoline N-oxide 7, from the 3-carboxyisoquinoline acid N-
oxide 6, proves that the silver-mediated decarboxylative cross-
coupling of isoquinolinic acid N-oxides 5 and 6 proceeds only
through the conventional mechanism (Figure 1, cycles A and
C).
40
85, 50[d]
78[c]
83[c]
85[c]
80[c]
10
[a] PdBr2 (10 mol%), PCy3HBF4 (10 mol%), Cs2CO3 (3 equiv), 1a (2 equiv),
2a–b (0.3 mmol), additive, solvent (1.5 mL), 1508C. [b] Yield based on iso-
lated product after flash chromatography. [c] 1.5 equiv of 1a. [d] Scale up,
1.5 mmol of halide under optimized conditions.
catalytic systems, in DMF at 1508C, depending on the nature
of the halide.[2] The formation of desired product 3a was ob-
served in 40–56% yield under both co-catalysis conditions by
using PCy3HBF4 (Cy=cyclohexyl) as a ligand and Cs2CO3 as
a base (Table 1, entries 1 and 2).
At this stage, the major challenge to improve the yield was
to circumvent the competitive protodecarboxylation side reac-
tion. In accordance with the observations of Glorius et al.,[22]
we found that the use of 1,4-dioxane as a solvent dramatically
improved the yields under both silver and copper co-catalysis
(Table 1, entries 1–3 and 6). After screening various parameters
(e.g. ligands, palladium sources, and bases),[23] the best results
were obtained by using PdBr2 (10 mol%) in the presence of
Cs2CO3 with PCy3HBF4 (10 mol%) (Table 1, entries 3–6). We
were pleased to find that these decarboxylative coupling reac-
Moreover, the unselective C1/C3 arylation of isoquinoline N-
oxide 9 by direct CÀH arylation with p-tolyliodide 2a under
Chem. Eur. J. 2014, 20, 3610 – 3615
3611
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim