Organic Letters
Letter
selective isomerization/hydrocyanation of aliphatic internal
olefins without involving Lewis acid (Scheme 1c).
effective in the earlier reports of transfer hydrocyanation
reactions.3m,n,6,7 These studies failed to reveal a catalyst system
that could provide any reactivity. Gratifyingly, BiPhePhos-type
ligand L4, which was first reported by Bower in a C−H bond
functionalization reaction,9 gave the desired product 3a in 19%
yield with 94% linear regioselectivity (Table 1, entry 9).
Interestingly, the use of L5, the diastereomer of L4, resulted in a
prominent increase of the yield from 19% to 66%. However, the
regioselectivity completely reversed from linear to branched,
which implied that the ligand configuration was crucial to the
regioselectivity. When L6 was employed, no corresponding
hydrocyanated product was observed, which indicates that tBu
as the R1 substituent is necessary for the reactivity. We next
studied the steric hindrance effect of the R2 substituent in order
to improve the yield (Table 1, entries 12−15). Notably, L8
afforded the linear product in 83% isolated yield with 94:6
regioselectivity (Table 1, entry 13). It was found that the steric
hindrance of the R2 substituent strongly influenced the reactivity
of the reaction rather than the regioselectivity. No hydro-
cyanated product was observed, and the octene isomer mixture
was recovered when the 2,2′-biphenol type ligand was employed
(Table 1, entry 16, for more details; see the Supporting
Information). This result disclosed that the ligand bearing
BINOL is necessary to drive the reaction.10
In our initial investigations we explored the effect of a series of
phosphorus ligands on the model reaction of 2-octene (1a) with
acetone cyanohydrin (2)8 as HCN donor in the presence of
bis(1,5-cyclooctadiene)nickel as a precatalyst. Ligands L1−L3
and DIOP, which showed preliminary success in the asymmetric
hydrocyanation reactions,3d,e,h were examined. Unfortunately,
no corresponding hydrocyanated product was observed and the
octene isomer mixture was recovered (Table 1, entries 1−4).
Further investigations showed that commercially available
bidentate ligands (e.g., BINAP, Xantphos, DPEphos and
Dppe) exhibited no reactivity in the formation of nonanonitrile
(3a) (Table 1, entries 5−8). DPEphos and Xantphos were
Table 1. Ligand Effect on the Ni-Catalyzed Hydrocyanation
a
of 1a
After identification of suitable reaction conditions for the
hydrocyanation of aliphatic internal olefins to linear nitriles
(Table 1, entry 13), we next explored the generality of this
transformation (Scheme 2). 2-Olefins bearing diverse functional
groups, such as ether (1c), ester (1d), nitrile (1e), acetal (1f),
boron ester (1g), phosphate ester (1h), and phthalimide (1i),
were all applicable substrates. The corresponding alkyl nitriles
(3c−3i) were smoothly produced in good yields with excellent
linear regioselectivity. Various 3-olefins bearing challenging
functional groups, namely, unsaturated ketone (1k), secondary
amine (1n), unprotected alcohol (1o) and trisubstituted alkene
(1p), underwent efficient anti-Markovnikov hydrocyanation to
afford the desired nitriles (3k and 3n−3p) in 58−84% yield with
84−99% linear regioselectivity. It is worth mentioning that
functional groups with active hydrogens are not compatible
under Lewis acid conditions. Moreover, 4-olefins were also
tolerated and produced the corresponding products in moderate
to decent yields with good linear-to-branched ratio (Table 1c).
In addition, cyclic olefin (1r) could also generate the
corresponding product (3r) in acceptable yield. Internal olefin
1s, containing chenodeoxycholic acid, a drug for cholesterol
gallstones and hyperlipidemia, could also be applied. The nitrile
product 3s was obtained in 87% yield. Besides, the internal olefin
(1t) derived from estrone was subjected to our protocol, leading
to the desired product (3t) in moderate yield. Notably, even
when the “isomerizing” alkenyl was conjugated, β-substituted
styrene (1u) and acrylate (1v) proved to be viable substrates.
Finally, the results obtained from the Ni-catalyzed isomer-
ization/hydrocyanation of internal olefins prompted us to apply
this system to octene mixtures on gram scale. To our delight,
nonanonitrile was obtained in 92% yield with excellent
regioselectivity (Table 1g).
b
c
entry
ligand
L1
L2
3a yield (%)
3a l/b
1
2
3
4
5
6
7
8
0
0
0
0
0
0
0
0
/
/
/
/
/
/
/
/
L3
DIOP
BINAP
Xantphos
DPEphos
Dppe
L4
L5
L6
L7
L8
9
19
66
0
94/6
10
11
12
13
14
15
16
22/78
/
13
92 (83)
87/13
94/6
93/7
90/10
/
d
L9
L10
L11
48
50
0
To gain more insight into the isomerization/hydrocyanation
process, both isotope labeling and crossover experiments were
conducted (Scheme 3). (CD3)2C(OD)CN was used for 1g
under standard conditions. We observed deuterium incorpo-
ration at the C1 (0.71D), C2 (0.96D), C3−C4 (0.91D), C5
(0.23D), and C6 (0D) positions in d-3g-1 in 58% yield with 95:5
linear-to-branched ratio (Scheme 3a). It implies that the nickel
a
The reactions were carried out at 80 °C for 12 h with 1a (0.1 mmol),
Me2C(OH)CN (0.3 mmol), and toluene (0.3 mL) in the presence of
bis(1,5-cyclooctadiene)nickel (5 mol %) and ligand (5 mol %). The
yields were determined by GC analysis using n-decane as the internal
b
c
standard. The regioselectivities were determined by GC and GC−
d
MS analysis. Yield of isolated linear product.
487
Org. Lett. 2021, 23, 486−490
Gao, Jihui
