CHEMCATCHEM
FULL PAPERS
The AgÀPd/C catalyst was comparable to other reported Pd-
based heterogeneous catalysts, such as Pd/AlO(OH)[14a] and
nano-PdÀV,[14e] in terms of the yield of the desired coupled
product. However, compared with the reported Pd/C catalyst
(66% yield over 5 mol% of Pd),[14c] our AgÀPd/C catalyst dem-
onstrated a superior catalytic activity (94% yield over 1 mol%
of Pd). Furthermore, no sacrificial hydrogen acceptor was re-
quired in our system. Hence, our catalytic system is a greener
process as less waste is generated.
the yield by using 1,4-dioxane with a huge excess (5 equiv) of
base (KOH or KOtBu) were not productive (Table 2, entries 3
and 4). We thus discontinued this approach. The use of the
stronger base KOtBu with toluene also did not improve the
yield (Table 2, entry 5), whereas the use of Cs2CO3 and LiOH
(Table 2, entries 6 and 7) resulted in lower yields of the desired
product. K3PO4 was ineffective in forming the quinoline
(Table 2, entry 8). Much of the unreacted starting material was
observed in the GC–MS spectrum of the reaction mixture; this
supported the hypothesis that K3PO4 was too weak a base to
form the corresponding enolate from 3a, which was required
to attack 2-aminobenzylaldehyde formed from 2-aminobenzyl
alcohol. Increasing the amount of K3PO4 to 5 equiv helped the
reaction greatly; this improved the yield from 14 to 43%
(Table 2, entries 8 and 9), but it still fell short of the 68% that
KOH could achieve. Increasing the catalyst loading to 2 mol%
of Pd with K3PO4 led to a slight increase in the product yield
(Table 2, entry 10). Similar to Step 1, the yields decrease with
decreasing concentration (Table 2, entries 11 and 12).
Modified Friedlꢀnder annulation of 2-aminobenzyl alcohol
with b-phenylpropiophenone
For the optimization of the second step, various conditions
were explored by using the model substrates 3a and 2-amino-
benzyl alcohol (Table 2). Substrates 3a and 2-aminobenzyl al-
cohol were reacted under an O2 balloon at 1258C for 21 h with
the AgÀPd/C catalyst (1 mol% of Pd) and 3 equiv of KOH to
form the product quinoline 4a in 68% yield (Table 2, entry 1).
Similar to Step 1, both toluene and 1,4-dioxane were examined
in this oxidative cyclization and we found that a slightly higher
yield was obtained with 1,4-dioxane (Table 2, entries 1 and 2).
This presented a conflict as the a-alkylation of ketones was
found to be better in toluene (Table 1, entries 1 and 2). Toluene
was chosen in the end because it gave less side products in
the first step, which thus reduces possible complications if
both steps were combined. Furthermore, attempts to boost
Compound 4a was produced in 68% yield even without the
AgÀPd/C catalyst in the presence of the starting materials, O2,
and KOH (Table 2, entry 13). Although transition-metal-free
quinoline syntheses have been reported,[13] these involved the
use of an organic hydrogen acceptor, such as benzophenone
or another equivalent of the ketone. The presence of a strong
base seemed to be the key to this catalyst-free reaction path-
way because the AgÀPd/C catalyst was still required if a weak
base was used, as demonstrated by the fact that changing the
catalyst loading of AgÀPd/C affected the efficiency of the reac-
tion in the presence of K3PO4. However, the AgÀPd/C catalyst
was still used in experiments in which strong bases were used
to ensure a fair comparison.
Table 2. Friedlꢀnder annulation of 2-aminobenzyl alcohol with 3a
catalyzed with AgÀPd/C.[a]
Comparison and selection of reaction conditions for one-pot
synthesis of polysubstituted quinolines
After studying the results of Steps 1 and 2, it was decided that
the best conditions that would enable the integration of both
reactions with maximum yield would be obtained by using tol-
uene, K3PO4, and Ar for Step 1 and toluene, KOH, and O2 for
Step 2. Although we had hoped to use the same base for both
reactions, KOH was very strong for the selective a-alkylation of
ketones and K3PO4 was very weak for the efficient formation of
quinoline. LiOH and Cs2CO3 were not the bases of choice for
Step 1 because they are more expensive than K3PO4. On
a 0.25 mmol scale, 0.5 mL of toluene was the optimal amount
of solvent required for both steps. The system was robust
enough so that linking Steps 1 and 2 involved adding 2-amino-
benzyl alcohol and KOH to the reaction vessel after the first
stage of the reaction was completed; no purification was re-
quired. The one-pot reaction of 1a and 2a to form 3a and fi-
nally 4a resulted in a yield of almost 70% (Table 3, entry 1),
which was similar to that obtained by the reaction of pure 3a
and 2-aminobenzyl alcohol to form 4a.
Entry
Base
Solvent
Isolated
yield [%]
1
2
3
4
5
6
7
8
KOH
KOH
toluene
68[b]
74
67
67
64
42
37
14
43
20
65
59
68
1,4-dioxane
1,4-dioxane
1,4-dioxane
toluene
toluene
toluene
toluene
toluene
toluene
KOH[c]
KOtBu[d]
KOtBu
Cs2CO3
LiOH
K3PO4
K3PO4
K3PO4
KOH
KOH
KOH
[e]
9
10[f]
11
12
13[i]
toluene[g]
toluene[h]
toluene
[a] Reaction conditions: 2-aminobenzyl alcohol (0.275 mmol), 3a
(0.25 mmol), AgÀPd/C (1 mol% of Pd), base (0.75 mmol), solvent (0.5 mL),
101.3 kPa of O2, 1258C, 21 h; [b] Average of two runs; [c] 1.25 mmol of
KOH was used; [d] 1.25 mmol of KOtBu was used; [e] 1.25 mmol of K3PO4
was used; [f] 2 mol% of the Pd catalyst was used; [g] 1 mL of toluene
was used; [h] 2 mL of toluene was used; [i] In the absence of the AgÀPd/
C catalyst.
The surface chemistry of the AgÀPd/C catalyst before and
after the one-pot reaction was analyzed by using XPS (Figur-
es S1 and S2). There was no change in the Ag 3d XPS peaks
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 277 – 283 279