Organic Letters
Letter
electronic effect of substituents on the aryl ring (R1) had a
marginal influence on the yield (1u−1aa). A direct cleavage of
the free hydroxyl group was also achieved employing this
catalytic system (1ab). The reactions of substrates with OMe
and OTBS as the leaving groups were much less efficient (1ac
and 1ad). In the case of substrates bearing α-acyl groups, the
catalytic system afforded the deoxygenation products efficiently
(1ae−1ai), albeit a relatively longer reaction time was required
for these cases.
Scheme 3. Proposed Mechanism
To further showcase the synthetic utility, the catalytic
system was applied to depolymerization of a polymeric lignin
β-O-4 ketone model (4), producing the ketone monomer (5)
in 78% yield (eq 1). The C−O bond cleavage reaction could
be scaled up to gram scale smoothly, affording the
corresponding ketone and phenol in 93% and 71% yields,
respectively (eq 2).
ligated cobalt−boryl species (A) is generated by treating the
PDI-CoCl2 complex with a base and B2pin2.8d,18,19 Following
the coordination of the substrate to the cobalt center, insertion
of a CO group into the Co−B bond forms species B, which
undergoes 1,2-Brook type rearrangement to produce species C.
Then, β-O elimination yields alkenyl boron ether and cobalt-
alkoxyl species (D).20 Finally, regeneration of A occurs via a
rate-limiting transmetalation process between D and B2pin2.
DFT calculations were carried out to improve our
understanding (Figures 2 and 3). Upon the formation of
Co(I)−boryl species A,8d its complexation with 1a to form
IN1 is almost energetically neutral (Figure 2). The insertion of
the carbonyl functionality of 1a into the Co−B bond of A
occurs via TS1 facilely with an activation barrier of 10.1 kcal/
mol, giving rise to intermediate IN2 exergonically. The energy
for regioisomeric insertion is much higher than that for TS1
(Figure S5). To undergo further transformation, the 1,2-Brook
type rearrangement would occur first to generate the alkyl Co
species. This process is realized via a stepwise reaction by the
generation of a three-membered ring intermediate IN3 via
TS2, requiring a barrier of only 6.5 kcal/mol and being slightly
endergonic. From zwitterionic species IN3, C−B bond
cleavage via TS3 is very easy and leads exergonically to alkyl
Co(I) complex IN4 in which the Co atom is associated
strongly with the O atom with a Co−O distance of 2.08 Å. The
dissociation of the Co−O interaction is favorable as IN5 would
be much lower in energy than IN4. Prior to the β-oxygen
elimination, complex IN6, in which the phenoxide oxygen is
associated with the Co, should be formed from IN5 by bond
rotation. From IN6, C−O bond cleavage occurs via TS4 with a
barrier of 12.2 kcal/mol, leading to alkenyl boron ether IN7
and Co(I)-phenoxide highly exergonically. The final product
could be formed by further protonation of IN7.
The regeneration of catalytic species A from the Co(I)-
phenoxide (LCo-OPh) was next investigated. The direct
metathesis of LCo-OPh with B2Pin2 involves two steps (Figure
3a). The first step is phenoxide transfer. In this step, the Co−O
bond is cleaved via TS5, from which the phenoxide moiety is
transferred to the boron moiety with a barrier of 22.7 kcal/mol
to form zwitterionic intermediate IN8 energonically. In IN8,
the B1−O distance is obviously longer than the normal B−O
bond (1.60 Å in IN8 vs 1.36 Å in IN11) while the B1−B2
distance is almost unchanged (1.76 Å in IN8 vs 1.71 Å in
B2Pin2). Prior to the second step of boryl transfer, IN8 should
first isomerize to IN9. In the latter complex, the negative
charge on B2 is stabilized by both Co (Co−B2 = 2.35 Å) and
B1 (B1−B2 = 1.89 Å) and the B1−O interaction is enhanced
To explore the mechanistic details of this transformation, we
conducted quantitative kinetic studies (Figure 1) to determine
Figure 1. Quantitative kinetic studies. (a) Plot of kin vs PDI-CoCl2
concentration. (b) Plot of kin vs B2pin2 concentration. (c) Plot of kin
vs 1a concentration.
the role of the substrate (1a), B2pin2, and catalyst at the rate-
determining step (RDS). Measurements of the initial rates
(kin) for the reaction of 1a with different concentrations of
PDI-CoCl2 and B2pin2 showed increases in the rates. Plots of
kin versus the concentration of PDI-CoCl2 (Figure 1a) and
B2pin2 (Figure 1b) gave two linear curves (slopes of 1.91 ×
10−2 and 1.45 × 10−2 M min−1, respectively), which suggested
a first-order rate dependence on catalyst and B2pin2. However,
an inverse correlation was found between kin and the
concentration of 1a (Figure 1c). These quantitative kinetic
studies indicated that B2pin2 should be involved in the RDS
and the reaction is slowed by excess substrate 1a.
On the basis of the quantitative kinetic studies and literature
report12 on insertion of a CO group into the Cu−B bond, a
plausible catalytic cycle is proposed in Scheme 3. First, a
C
Org. Lett. XXXX, XXX, XXX−XXX