.
Angewandte
Communications
heating an ethanolic solution of [RuCl(h6-p-cymene)(3)]Cl in
the autoclave at 1508C resulted in the formation of trans-
[RuCl2(3)]. The 31P{1H} NMR spectrum (unlocked) of the
product solution of a typical catalytic run (Table 1, entry 11)
exhibited a set of mutually coupled triplets at 15.4 and
1.9 ppm (2JPP = 29.2 Hz) and a singlet at 10.3 ppm, which we
attribute to cis and trans isomers of [RuH2(3)]. We synthe-
sized this dihydride complex according to a previously
reported method[16] and found it to be catalytically active
(Table 1, entry 17). 31P{1H} NMR spectroscopic analysis of
catalytic solutions after longer reaction times (e.g. Table 1,
entry 12) also showed several unknown species at 40.9 (s),
33.4 (s), and 21.1 ppm (s); their presence suggests some
catalyst decomposition. The importance of the small bite
angle of the diphosphine is apparent by comparison of the
results obtained with trans-[RuCl2(L)] complexes of 1 and 2
(Table 1, entries 18 and 19).
least rapid exchange between free and bound hydrogen/
hydrides. Similar results were obtained when EtOD was used
as the substrate (Scheme 3b). Indeed, deuterium incorpora-
À
tion into C H/D bonds of “unreacted” ethanol was also
observed in this experiment; moreover, the labeling pattern
was very distinctive, with a 2:1 preference for monodeutera-
tion at the ethanol 2-position over the 1-position. A simple
ethoxide b-elimination/reinsertion mechanism would seem to
favor incorporation at the 1-position (Scheme 3c), and this
result suggests the intermediacy of an enolate p-bound to the
=
ruthenium center through the C C bond.
The remarkable feature of these catalysts is their
extremely high selectivity as compared to that of all previous
systems. Assuming a Guerbet-type mechanism, a key facet of
this selectivity must be the exertion by the catalysts of
extremely high levels of control over the base-catalyzed
acetaldehyde aldol reaction, so that only dimeric products are
obtained, rather than the usual mixtures of higher oligomers.
To study this hypothesis, we performed a series of aldol
condensation reactions for both acetaldehyde and butylalde-
hyde in the presence and absence of the ruthenium catalyst
(see Table S2 in the Supporting Information for a summary of
these experiments). Under analogous conditions to those of
the catalytic reactions but in the absence of ruthenium,
acetaldehyde was oligomerized with 100% conversion in 4 h
by NaOMe with little or no control, and the expected
dimerization product crotylaldehyde accounted for only
14.6% of the total product mixture; the rest of the products
were higher oligomers. By contrast, when the analogous
reaction was performed in the presence of [RuCl(h6-p-
cymene)(3)]Cl (0.1 mol%), 56.9% of the product was croty-
laldehyde, with 100% conversion. The same experiment with
butylaldehyde was also revealing: in the absence of ruthe-
Preliminary mechanistic studies, as well as the observed
higher-alcohol side products, support a Guerbet-type mech-
anism. Treatment of [RuCl(h6-p-cymene)(3)]Cl with NaOEt
under catalysis-like conditions led rapidly to the formation of
1
a species with a triplet signal in the H NMR spectrum at
À9.47 ppm (triplet, 2JPH = 33 Hz), consistent with a ruthenium
hydride. Addition of the ethanol substrate to this preactivated
catalyst led to butanol formation. This result suggests that
after initial formation of a ruthenium ethoxide, b elimination
leads to the formation of the active hydride and the aldehyde.
The addition of ethanol results in metathesis with the hydride,
loss of hydrogen, and the formation once again of the
ruthenium ethoxide to close the dehydrogenation cycle.
To investigate whether the hydrogen remains bound to the
metal center during the catalytic cycle prior to delivery for
later hydrogenation, we performed a catalytic reaction in the
presence of D2 (2.5 bar; Scheme 3a). Significant D incorpo-
ration (25% across all sites) in the butanol product was
observed, which suggests the presence of free hydrogen, or at
nium, 85% selectivity for a mixture of higher oligomers (C8+
)
was observed in 4 h, whereas in the presence of [RuCl(h6-p-
cymene)(3)]Cl (0.1 mol%), only 24% selectivity was
observed for oligomers in the same reaction time, and C4
species made up 76% of the product, including n-butanol
(19%). These results suggest that the ruthenium catalyst
biases the aldol condensation to give the desired C4 products
by increasing the rate of acetaldehyde coupling but reducing
the rate of aldol reaction with C4 or higher alcohols.
Aldehyde-hydrogenation experiments also revealed why
such high selectivity is observed. Under catalysis-like con-
ditions with [RuCl(h6-p-cymene)(3)]Cl and H2 (2.5 bar), no
conversion of acetaldehyde into ethanol was observed within
4 h; indeed, only coupling products were observed (C4, C6,
and C8), including n-butanol (31.8%). By contrast, butylalde-
hyde was hydrogenated to n-butanol with 48.9% conversion
in 4 h, and higher oligomers (C8+) made up only 19.4% of the
product. These results suggest a regime during catalysis in
which the aldol condensation of ethanol is favored over that
of C4 or higher aldehydes, but the hydrogenation of C4 or
higher aldehydes is favored over that of acetaldehyde. The
low, steady concentration of acetaldehyde that results allows
the reaction to proceed in a selective manner. It is tempting to
also propose an “on-metal” aldol condensation to account for
the unusually high selectivity of this step, especially since the
intermediacy of the required enolates is implied by our
Scheme 3. a,b) Deuterium incorporation into the n-butanol product
and ethanol substrate in the presence of D2 gas (a) and EtOD (b)
under catalytic conditions ([Ru]=[RuCl(h6-p-cymene)(3)]Cl). c) Ethox-
ide-elimination/reinsertion mechanism.
4
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
These are not the final page numbers!