2
K.S. DeGlopper et al. / Polyhedron xxx (2016) xxx–xxx
O
O
O
O
[Ni0], bipy
R
Ph
N
H
N
R
R'
2Zn
N
+
N Ph
ZnEt
1,4-dioxane
95 o
C
R'
Et
[Ni]0 CO
[Ni]0
O
O
R: 10 examples
R': 14 examples
Up to 96% yield
F
A
O
O
Ph
Scheme 1. Nickel-mediated decarbonylative cross coupling of imides with
diorganozinc reagents.
Ph
N
N
ZnEt
[Ni] Et
[Ni]
O
O
nickel precatalysts are commercially available and used without
purification or prepared according to procedures provided. All phthal-
imides were obtained commercially or prepared via the condensation
of phthalic acid with the appropriate amine in either refluxing
toluene (Dean-Stark conditions) or in refluxing acetic acid [15].
General method for catalytic decarbonylative coupling with Et2-
Zn will be illustrated with a specific example. 2,20-Bipyridine (bipy)
(15.6 mg, 0.010 mmol) and N-phenylphthalimide (1) (238 mg,
0.90 mmol) were combined in an oven dried 25 mL round bottom
flask and transferred into an inert atmosphere glove box, where Ni
(COD)2 (COD = 1,5-cyclooctadiene) (24.8 mg, 0.090 mmol) was
added. The flask was sealed with a septum and removed from
the glove box. DMSO (2.5 mL) was added via syringe, followed by
B
O
Et2Zn
CO
E
Ph
N
ZnEt
[Ni] Et
D
O
Et2Zn
N
Ph
[Ni]
C
CO
Scheme 2. Mechanistic hypothesis.
Et2Zn (70 lL, 0.67 mmol) also via syringe. The dark solution was
active nickel(0) complex A undergoes a formal oxidative addition
into the phthalimide to generate metallacycle B. This species
undergoes transmetallation with Et2Zn to generate D and subse-
quent decarbonylation to provide complex E. Reductive elimina-
tion and acid workup yields the ortho-substituted benzamide
and regenerates nickel(0) complex F. In a potential alternative,
metalacycle B may undergo decarbonylation prior to transmetalla-
tion which would generate complex C before proceeding to E. Inde-
pendent of the specific pathway, it is hypothesized that the
strength of the Ni–CO bond in the final nickel(0) species F prevents
dissociation, and the complex is not sufficiently nucleophilic to
induce oxidative addition to achieve catalysis [16–18]. These fac-
tors contribute to the requirement for stoichiometric nickel to
mediate the transformation.
In analyzing reaction conditions, we identified several means of
altering reaction conditions that had the potential to invoke cata-
lyst turnover, including changing the ligand, solvent and substitu-
tion on the imide, with the initial intent of weakening the final Ni–
CO bond to induce dissociation and achieve the regeneration of a
catalytically active nickel species.
Standard reaction conditions for the stoichiometric decar-
bonylative coupling included the use of 1 equiv Ni(COD)2, 1.1 equiv
bipy (2,20-bipyridyl) and 1.2 equiv Et2Zn in 1,4-dioxane at 95 °C for
16 h. The reaction can also be performed utilizing a Ni(II) precursor
such as Ni(acac)2 (acac = acetylacetonate); however this requires
additional diorganozinc reagent to reduce the metal center to a
catalytically active Ni(0) species. In an early attempt at catalysis,
N-phenyl phthalimide was combined with 20 mol% Ni(COD)2 and
21 mol% bipy under otherwise identical conditions, yielding 19%
of the anticipated benzamide coupling product, consistent with
the optimized stoichiometric reaction.
Characteristics of the reaction solvent clearly influence catalytic
turnover. The initially developed reaction conditions utilized 1,4-
dioxane heated at 95 °C in an oil bath. Using 10 mol% of catalyst,
a significant range of solvents were examined, with significant
variability in polarity, boiling point and tendency to coordinate
to transition metals. These solvents were tested under otherwise
identical reaction conditions at temperatures near their boiling
points. High boiling solvents generally resulted in greater catalyst
turnover as did solvents with a strong ability to coordinate a metal
center. Based on our current mechanistic hypothesis, it is believed
that solvent molecule coordination assists in the dissociation of
carbon monoxide from the metal center to regenerate nickel(0)
then suspended in a 100 °C oil bath and allowed to stir for 16 h.
Following reaction, the reaction was cooled to room temperature,
the septum was removed and the reaction mixture was diluted
with Et2O (15 mL). The addition of 2 M aq HCl (15 mL) quenched
the reaction, which was then extracted with Et2O (3 ꢀ 15 mL).
The combined organic layers were washed with brine (15 mL),
dried over MgSO4 and concentrated under reduced pressure to
yield crude 2-ethyl-N-phenylbenzamide (2).
2.2. Purification and analysis
Products of the above procedure were analyzed via GC/MS (Agi-
lent 6890 GC with Agilent 5973 mass selective detector), IR (Bruker
Alpha, diamond ATR), and NMR spectroscopy (Bruker 400 MHz
Avance III). 1H and 13C NMR spectra were acquired using standard
acquisition parameters and are referenced to TMS. Purification was
achieved through column chromatography (10:1 hexane:ethyl
acetate) and thoroughly characterized.
The general method for in situ IR spectroscopy for investigation of
the reaction progress will be illustrated with a specific example. A
10-mL flask was sealed with a septum and placed under argon atmo-
sphere. 1,4-Dioxane (2 mL) was added to the flask via syringe. The
solvent was heated in a 95 °C oil bath. N-Pentafluorophenylphthal-
imide (3) (157 mg, 0.050 mmol) and 2,20-bipyridine (8.6 mg,
0.055 mmol) were weighed out into an oven-dried, two-necked 10-
mL flask with a stir bar and transferred into an inert atmosphere
glovebox, where Ni(COD)2 was added to the flask, and the flask
was sealed with a septum. While utilizing a flow of argon from a bal-
loon to preserve the inert atmosphere, in situ IR probe was inserted
into one of the necks of the two-necked flask. The flask was then
flushed with argon (5ꢀ). Next, 1,4-dioxane was transferred from
the 10-mL flask to the two-necked flask via syringe. The two-necked
flask was immersed in a 95 °C oil bath. A background spectrum was
acquired. For the next 10–15 min, a spectrum was taken every one
minute. Finally, diethyl zinc (70 lL, 0.67 mmol) was added to the
flask via syringe. For the remainder of the reaction (1–3 h), a spec-
trum was taken every minute.
3. Results and discussion
As presented in our initial report, our mechanistic hypothesis
for the decarbonylation reaction is illustrated in Scheme 2. The