ACS Catalysis
Research Article
bimetallic catalysis (mechanisms 3 and 4), (E)-2 was
stereoselectively formed, very likely proceeding via a
bispalladium complex.28 A novel type of acetylide bridging
for bis-Pd complexes29 was found in putative catalytic
intermediates. Our studies suggest that a double μ−κ:η2
acetylide bridging allows for a cooperative bimetallic syn-
carbometalation, while in the above-mentioned previous
bimetallic carbometalation approaches, only (Z)-2 was
accessible by anti-additions. Contrariwise, depending on the
reaction conditions, it is also possible to form regioisomer 3 as
major product with the same catalyst. This regiodivergent
outcome is explained by a monometallic pathway for the
formation of 3.
Scheme 1. Common Mechanisms for Alkyne Dimerizations
and Comparison to This Work
RESULTS AND DISCUSSION
■
Optimization of the Selective Formation of (E)-2. The
chloride-bridged ferrocene-based palladacycle precatalysts
[FIP-Cl]2 (FIP = ferrocenyl imidazoline palladacycle) and
[FBIP-Cl]2 (FBIP = ferrocenediyl bisimidazoline bispallada-
cycle), which were previously developed by our group30,31 and
are accessible in high yields and in few steps, were studied as
precatalysts. They were activated by different silver salts for
chloride anion exchange by our previously described
protocols.30,31 The activated complexes were initially inves-
tigated in the dimerization of 1a in chlorobenzene at 50 °C
(Table 1). The best results were achieved by activation with
AgOAc. Using 5.0 mol % of precatalyst, a mixture of all three
isomers (E)-2a, (Z)-2a, and 3 was formed in low yields and
with a moderate preference for (E)-2a. With the FIP catalyst, a
higher yield (36%, entry 1) and a higher selectivity for
dimerization versus tri/oligomerization was noticed (48%
conversion), whereas FBIP showed higher catalytic activity
(99% conversion), but the product mixture was much more
complex (yield of dimer: 20%, entry 2).
A solvent screening was then performed, and with
acetonitrile, a higher activity, dimerization yield, and selectivity
for (E)-2a was found (entry 3). Improved selectivities were
attained by slightly reducing the reaction temperature to 45 °C
(entry 4) and using acetic acid (0.5 equiv) which was added to
speed the protonolysis of generated Pd−vinyl bonds (entry 5).
A similar result was obtained with NaOAc as additive (entry
6), probably due to generation of HOAc in the course of the
catalytic cycle. Adding water (5.0 equiv) instead, this effect was
not found (entry 7). Adding both NaOAc and H2O, yield and
selectivity were improved (entry 8). The product yield was
further increased by additional HOAc (entry 9). Surprisingly,
by reducing the precatalyst loading to 2 mol %, yield and
selectivity were improved to synthetically useful values (entries
10 and 11). This is at least partly explained by no detectable
alkyne trimerization as side reaction with the decreased catalyst
amount. Finally a yield of 91% and an isomeric ratio (E)-2a/
(Z)-2a/3 of 94/5/1 was accomplished. No hydrogenation
products were detected that might point to the formation of
metal hydride intermediates.
and a phosphinic acid to provide (E)-2 which is believed to
proceed via hydrometalation.23
In contrast to these studies employing neutral ligands,
Shaughnessy et al. studied the use of acetate-bridged aliphatic
C,P-palladacycle catalysts, which were effective in (E)-selective
couplings of propargylic alcohols and amines.27 Under the
reported reaction conditions, an acetate/acetylide-bridged
dimeric palladacycle 4 was identified (Figure 1). By X-ray
analysis the acetylide ligand was found to undergo an
unprecedented μ−κ1-coordination to both Pd(II) centers.
Herein we report a regiodivergent dimerization of alkynes 1
using a ferrocene-based C,N-palladacycle catalyst. In contrast
to the above-mentioned previous studies making use of
Scope of the (E)-Selective Dimerization. The con-
ditions described in Table 1/entry 11 were then used to
investigate the applicability to other terminal alkyne substrates
(Table 2). Using aromatic alkynes, substituents displaying
different electronic effects were all accommodated. Aromatic
alkynes bearing σ- (entries 3, 4, 9, 11, 12) and π-acceptors
(entry 5) thus provided useful to good yields and good to
excellent stereo- and regioselectivities. With an unsubstituted
Ph residue, a nearly quantitative yield and high selectivities
Figure 1. μ−κ1-Acetylide-bridged bis-Pd complex found in the
alkyne−alkyne coupling by Shaughnessy et al.27
5497
ACS Catal. 2021, 11, 5496−5505