Organometallics
Communication
ruthenium catalyst (HG1) (Scheme 1a). The original
conditions used for installing CAAC-5 ligands (KHMDS/
Scheme 1. Synthesis of Ru−CAAC-6 Catalysts 3a−d
Figure 2. Solid-state structures of 3c, 3d, and 4a with thermal
ellipsoids drawn at 50% probability. For clarity, hydrogen atoms have
been omitted. Selected bond lengths (in Å): for 3c: C1−Ru1
1.963(5), C23−Ru1 1.839(4), O1−Ru1 2.354(4), Cl1−Ru1
2.353(2), Cl2−Ru1 2.344(1); for 3d: C1−Ru1 1.960(2), C22−Ru1
1.836(2), O1−Ru1 2.370(1), Cl1−Ru1 2.3408(9), Cl2−Ru1
2.3634(8); for 4a: C1−Ru1 1.942(6), C19−Ru1 1.839(6), O1−
Ru1 2.294(5), Cl1−Ru1 2.330(2), Cl2−Ru1 2.335(2).
increased significantly (2.354 Å for 3c and 2.370 Å for 3d)
compared with an analogous Ru−CAAC-5 complex (4a, 2.294
Å),20 which is consistent with the greater σ-donating ability of
the CAAC-6 ligands. The lengths of the Ru−C bonds between
the CAAC ligand and the metal center remains similar (1.963
Å for 3c and 1.960 Å for 3d vs 1.942 Å for 4a), which can be
rationalized by an offsetting steric and electronic effect. The
CAAC−Ru−O angles for the synthesized Ru−CAAC-6
complexes (175.1° for 3c and 172.4° for 3d) were also
observed to be smaller than those in a five-membered analogue
(178.5° for 4a).
Catalyst initiation rates were then probed by reacting 3a−d
with 30 equivalents of butyl vinyl ether in benzene at 30 °C
(Table 1). Catalysts bearing an N-diisopropylphenyl (N-Dipp)
a
1
a
Determined by H NMR analysis.
Table 1. Initiation Study for Ru−CAAC-6 Catalysts 3a−d
THF) led to clean generation of free carbene 2a through
deprotonation.20 However, complete decomposition of 2a was
observed over several hours after HG1 was added, and only a
5% yield of the desired complex 3a was observed (Scheme 1a,
entry 1). As free CAAC-6 ligands (e.g., 2a) are more prone to
decompose than CAAC-5 ligands, conditions that do not
rapidly decompose the free carbenes are crucial for the
preparation of 3a. Accordingly, solvents, bases, and other
reaction parameters were evaluated. Using benzene in place of
THF successfully furnished 3a in 42% yield (entry 2);
however, the conversion could not be further improved due
to decomposition of 2a (entries 3 and 4). While the use of an
alkoxide base led to instant carbene decomposition (entry 5),
we were delighted to find that the deprotonation of 1a with
LiHMDS followed by the removal of LiBF4 by filtration greatly
increased the carbene’s stability (entry 6). Finally, when the
number of equivalents of 1a and the reaction temperature were
slightly increased, 3a was formed in 78% yield (entry 8).
Interestingly, the use of KHMDS instead of LiHMDS led to a
much lower yield under otherwise identical conditions, despite
the fact that both KBF4 and LiBF4 salts were removed.
Under the optimized conditions, four representative Ru−
CAAC-6 catalysts, 3a−d, were successfully prepared (Scheme
1b). We were unable to further reduce the sterics on the
nitrogen (e.g., with an N-mesityl group) or the α-carbon (e.g.,
with gem-dimethyls) of the carbenes because of difficulty in
synthesizing the CAAC salts through HCl-promoted cycliza-
tion.23 Single crystals were obtained for 3c and 3d, leading to
their unambiguous characterization through X-ray diffraction
(XRD) analysis (Figure 2). The Ru−O bond lengths were
a
Initiation rates were determined by measuring the decrease in the
b
benzylidene signal using 1H NMR spectroscopy. Data adopted from
ref 20.
substituent initiated slower than those with an N-methyl-
isopropylphenyl (N-Mipp) substituent by an order of
magnitude (e.g., 3a vs 3c), indicating the significance of the
N substituents in catalyst performance. In contrast, sub-
stituents on the α-carbon of the ligands appear to have less
influence on the catalyst initiation rate (e.g., 3a vs 3b). In
addition, some Ru complexes with CAAC-6 ligands initiate
faster than those with CAAC-5 ligands. For example, 4a−c,
which are among the most active Ru−CAAC-5 catalysts for
ethenolysis in the previous report, initiate at lower rates than
3c and 3d even at elevated temperature.20 We believe that the
observed faster initiation of many Ru−CAAC-6 catalysts likely
can be traced to a weakened Ru−O bond compared with that
in Ru−CAAC-5 complexes, as suggested by XRD analysis.24
B
Organometallics XXXX, XXX, XXX−XXX