ACS Catalysis
Research Article
13). Additionally, the reaction of styrene with DBpin produced
d-3a′ with the same D incorporation at the β-position (51%)
as acetophenone 1a did (eq 14). These results are consistent
with the intermediate of an alkene in the deoxygenative
borylation. With regard to the control of regioselectivity, it is
worth noting that the switch of regioselectivity was observed
when PPh2Me was employed instead of P(nBu)3 under the
standard conditions A, providing 3a′ as the major product (eq
15). Likewise, the linear alkylboronate 2a was dominant when
P(nBu)3 was utilized as the ligand under the standard
conditions B (eq 16). These results showed that the
regioselectivities were mainly controlled by the ligands.
We then turned our attention to investigate the triboration
mechanism. The vinylboronate 6b and 1,1-vinyldiboronate 6c
were obtained in 23 and 14% yields, respectively, when 1a was
treated with 2 equiv of B2pin2 (eq 17). Moreover, the
triboration of styrene gave the product 4a in 77% yield under
the standard conditions C (eq 18). The treatment of
vinylboronate 6b with the standard conditions C furnished
the triboronate 4a in 62% yield (eq 19), and no desired 4a was
formed in the absence of [RhCl(cod)]2 (eq 20). Furthermore,
the vinyldiboronate 6c was subjected to the standard
conditions C, providing the triboronate 4a in 79% yield (eq
21). These results illustrated that the formation of triboronates
involved the Rh-catalyzed deoxygenation of ketones to alkenes,
double dehydrogenative borylation of alkenes, and hydro-
boration (or diboration−protodeboration) of vinylboronates.
Based on the results of mechanistic studies, we propose a
plausible mechanism for this regiodivergent deoxygenative
borylation of ketones (Figure 3). As demonstrated in Figure 3,
the Rh−B bond generated from [RhCl(cod)]2 and B2pin2 in
the presence of the base63,64 undergoes a sequential reaction
process including 1,2-insertion, β-hydride elimination, and β-
oxygen elimination to give alkenes (cycle II), which is
consistent with our previous work. The resulting alkenes
then went through Rh-catalyzed hydroboration with HBpin to
afford alkylboronate products, in which the regioselectivities
were controlled by the ligands. In the presence of P(nBu)3
(cycle I), the styrene first coordinated with Rh complex B to
form intermediate F, which then underwent the anti-
Markovnikov hydroboration (migratory insertion and reduc-
tive elimination) to yield the linear alkylboronate 2. In cycle III
(L = PPh2Me), the coordination of styrene with Rh complex
B′ formed the intermediate F′. The subsequent migratory
insertion of the Rh−H bond, oxidative addition of HBpin, and
reductive elimination gave the Markovnikov product 3 and
regenerated Rh−H B′.
As expected, the hydroboration mechanism consistently/
simply involves insertion of olefin into [Rh]−H bond, followed
by oxidative addition of HBpin and then C−B bond-forming
reductive elimination. Interestingly, in each pathway, the olefin
insertion product [Rh]−alkyl does not undergo a direct one-
step σ-bond metathesis with HBpin to give an alkyl-Bpin
product. Instead, a two-step process consisting of oxidation
addition and reductive elimination is found. Clearly, an [Rh]−
C σ bond involving an sp3-carbon center considerably
compromises its accessibility to σ-bond metathesis with
HBpin. On the contrary, boryl, alkyl, and hydride ligands are
strongly σ-electron releasing, which are able to stabilize an
Rh(III) oxidation state, as evidenced by the reported fac-
[(PMe3)3Rh(B(cat))3]65 and fac-[Rh(H) (Bpin)2(PEt3)3],63
which promotes the two-step process.
From Figure 4, we found that the C−B bond-forming
reductive elimination is very facile, a result due to which the
presence of the empty p-orbital on boron promotes the
migration of the alkyl ligand to the boron center of the boryl
ligand.66 The elimination process is basically an empty-p-
orbital-assisted 1,2-migration of alkyl to the boron center of the
boryl ligand.
The DFT results also show that the olefin insertion is the
regioselectivity-determining step. The olefin insertion process
(into the Rh−H bond) can be formally viewed as a
nucleophilic attack of the M−H σ-bond on one olefinic
carbon. Therefore, among the two olefinic carbons, the less π-
electron-rich carbon is expected to be preferentially attacked
during the migratory insertion process.67−69 Indeed, for the
case when the phosphine PPh2Me is used, the regioselectivity
preference shown in Figure 4b is consistent with this
commonly accepted view as the phenyl substituent on styrene
is π-electron accepting. For the case when the phosphine PEt3
is used, the regioselectivity preference shown in Figure 4a
reverses. The steric effect provides a reasonable explanation for
these observations. In PPh2Me, the relative orientation of two
planar phenyl substituents allows to create a less sterically
hindered pocket for styrene to approach the rhodium metal
center to facilitate the insertion. In PEt3 (or PnBu3), such a
flexibility is no longer possible. The argument here is also
consistent with the fact that the insertion barriers calculated for
the PEt3 case (Figure 4a) are all noticeably greater than those
calculated for the PPh2Me case (Figure 4b).
In the energy profiles shown in Figure 4, the parts associated
with HBpin oxidative addition and C−B reductive elimination
resemble each other in both the PPh2Me and PEt3 cases,
suggesting that the electronic properties of the primary versus
secondary alkyl play dominant roles in the relative preference
between the two insertion modes.
In our previous work, we have computationally studied and
validated cycle II in Figure 3 related to the deoxygenation
process, leading to the generation of the intermediate styrene.
To investigate how phosphine ligands affect the regioselectivity
in the hydroboration of the intermediate styrene (cycles I and
III in Figure 3), we performed density functional theory
(DFT) calculations at the ωB97X-D level of theory (employed
in the previous work),29 considering [Rh]−H ([Rh] =
(PR3)3Rh) as the active species, which can be easily generated
in the presence of the HBpin reagent. Figure 4 shows the
energy profiles calculated for hydroboration of styrene with
HBpin catalyzed by (PEt3)3Rh−H (Figure 4a) and
(PPh2Me)3Rh−H (Figure 4b; here we used Et as the model
for nBu for theoretical simplicity and computational cost
reduction).
CONCLUSIONS
■
In summary, we have developed an efficient Rh-catalyzed
regiodivergent deoxygenative borylation of ketones for the
synthesis of linear and branched alkylboronates as well as
triboronates. This protocol represents the first example of
regiodivergent preparation of anti-Markovnikov and Markov-
nikov alkylboronates from readily available starting materials
other than alkenes and offers an important complement to the
existing protocols for their synthesis. In addition, this method
features mild reaction conditions, good functional group
tolerance, and broad substrate scope. The utilities of this
approach were also demonstrated by the gram-scale reactions
and various transformations of alkylboronates. Preliminary
9502
ACS Catal. 2021, 11, 9495−9505