ACS Catalysis
Research Article
suggesting that the reaction might not undergo via redox of the
Ru center (see Figure S6 in the Supporting Information). Next,
the di-deuterated substrate 2a-D2 together with the unsatu-
rated ketone 13 was subjected to the reaction conditions under
N2 (Figure 5C). Interestingly, appreciable amount of
deuterated compound 14 was detected. Thus, we speculated
that the deuterium in substrate 2a-D2 might be captured to
form the Ru-D species, which could reduce unsaturated ketone
13 through Michael addition.16 We also conducted a control
experiment by carefully measuring the amount of oxygen used
in the reaction in a sealed tube (Figure 5D). It was found that
0.5 equiv of molecular oxygen (i.e., 1 equiv of oxygen atom)
was sufficient to drive the reaction to completion.
In order to understand the role of chloride anions, various
silver salts including AgSbF6, AgBF4, and AgOTf were added in
order to displace the chloride counteranion (Figure 6A).17 It
was found that the reaction efficiency was affected by the
basicity of the counteranions and chloride showed the better
catalytic activity than other less coordinating anions. In the
absence of ZW2, the silver salts alone were also tested to
displace the chloride from RuCl3·xH2O and only trace amount
of oxidized product was observed. These results suggest that
the catalytic efficiency of the protocol was not done by simply
removing the strongly coordinating chloride counterion from
the ruthenium center.
Next, we attempted to probe the role of ZW2 in the
reaction. Phosphonium salt PBr and tosyl amide NLi, which
are the segmented catalysts of ZW2, were also evaluated
(Figure 6B). A sluggish reaction was observed when either of
these components or their 1:1 mixture was used, indicating
that the site-isolated cation and anion in ZW2 might work
synergistically. The ZW2·HCl complex was used instead of
ZW2 in the reaction, and the oxidation gave no conversion,
suggesting that the basicity of ZW2 is pivotal to activate
ruthenium trichloride at the initial stage in order to achieve
high catalytic performance. Another set of experiment was
carried out with ZW2 or ZW2·HCl alone as the catalyst, and
no reaction was observed. This result reveals that the organic
(ZW2) and inorganic (Ru−O species) components are both
necessary in the catalytic system. Potassium carbonate was
added to the reaction in order to capture the acidic proton on
ZW2·HCl that was in situ generated in the reaction. Reaction
yield of the desired product dropped dramatically, implying
that the acidic proton is also playing a crucial role. When 4 Å
molecule sieves were added to remove moisture, the catalytic
performance diminished significantly, which indicates the
important role of water.
We suspected that the N−H−Cl moiety in the ZW2·HCl
complex might interact with the alcohol substrate via a
hydrogen bond.18 Indeed, after adding 4-methylbenzyl alcohol
(2b) to a solution of ZW2·HCl in CDCl3, it was found that
protons of the TsN−H moiety of ZW2·HCl (both the N−H
and the aromatic region) shifted upfield significantly while the
α-H of phosphonium (Ha) shifted downfield (Figure 6C). At
the same time, protons at the benzyl group (Hc and Hd) of the
substrate shifted upfield. These chemical shifts could be
corresponding to the putative complex A that contains
hydrogen bonds between O−H in the substrate and H−Cl
in ZW2·HCl.
Figure 6. (A−C) Studies on the role of zwitterion.
selective oxidation of diol 2ae, and the mono-oxidized product
3ae was obtained exclusively.
2.4. Mechanistic Studies. Several control experiments
were conducted to shed light on the mechanism. Kinetic
isotopic effect (KIE) experiments were performed to reveal the
process of C−H cleavage (Figure 5A). Mono-deuterated
substrate 2a-D was subjected under the standard oxidation
conditions, and the ratio of kH/kD of product 3a was found to
be 5.75, indicating that the C−H abstraction of substrate 2a
might be the rate-determining step.12 Radical clock experi-
ments were conducted with substrate 2af under the standard
conditions (Figure 5B). The desired ketone product 3af was
obtained quantitatively and no cyclopropane ring-opening
product 12 was detected, suggesting that the reaction is
unlikely to go through the putative benzylic radical
intermediate 11. Furthermore, no signal relevant to the
redox of Ru was observed in the cyclic voltammetry study,
A plausible mechanism was established by piecing the
abovementioned experimental results (Figure 7). First, the
amide anion in ZW2 might coordinate with the ruthenium
center of ruthenium trichloride, causing the dissociation of
3503
ACS Catal. 2021, 11, 3498−3506