Nickel promoted functionalization of CO2 to anhydrides and ketoacids

Article abstract of DOI:10.1039/c4dt01221f

The reductive functionalization of carbon dioxide into high value organics was accomplished via the coupling with carbon monoxide and ethylene/propylene at a zerovalent nickel species bearing the 2-((di-t-butylphosphino)methyl)pyridine ligand (PN). An initial oxidative coupling between carbon dioxide, olefin, and (PN)Ni(1,5-cyclooctadiene) afforded five-membered nickelacycle lactone species, which were produced with regioselective 1,2-coupling in the case of propylene. The propylene derived nickelacycle lactone was isolated and characterized by X-ray diffraction. Addition of carbon monoxide, or a combination of carbon monoxide and diethyl zinc to the nickelacycle lactone complexes afforded cyclic anhydrides and 1,4-ketoacids, respectively, in moderate to high yields. The primary organometallic product of the transformation was zerovalent (PN)Ni(CO)₂. This journal is

Full text of DOI:10.1039/c4dt01221f

Dalton
Transactions
View Article Online
View Journal | View Issue
PAPER
Nickel promoted functionalization of CO2 to
anhydrides and ketoacids†
Cite this: Dalton Trans., 2014, 43,
15990
Zoe R. Greenburg, Dong Jin, Paul G. Williard and Wesley H. Bernskoetter*
The reductive functionalization of carbon dioxide into high value organics was accomplished via the
coupling with carbon monoxide and ethylene/propylene at a zerovalent nickel species bearing the 2-((di-
t-butylphosphino)methyl)pyridine ligand (PN). An initial oxidative coupling between carbon dioxide,
olefin, and (PN)Ni(1,5-cyclooctadiene) afforded five-membered nickelacycle lactone species, which were
produced with regioselective 1,2-coupling in the case of propylene. The propylene derived nickelacycle
lactone was isolated and characterized by X-ray diffraction. Addition of carbon monoxide, or a combi-
nation of carbon monoxide and diethyl zinc to the nickelacycle lactone complexes afforded cyclic anhy-
drides and 1,4-ketoacids, respectively, in moderate to high yields. The primary organometallic product of
the transformation was zerovalent (PN)Ni(CO)2.
Received 25th April 2014,
Accepted 11th September 2014
DOI: 10.1039/c4dt01221f
www.rsc.org/dalton
Introduction
Rising energy costs and climatological changes associated
with carbon combustion have heightened society’s sensitivity
to its dependence on non-renewable fossil fuels. Though
less pervasive, the same reliance on coal, natural gas, and
petroleum which constrains our energy supply is also found in
the production of numerous carbon based commodity chemi-
cals.1 These sustainability and economic influences have
driven research into the utilization of inexpensive and renew-
able carbon feedstocks to satisfy the demand for industrial
chemicals.2 The functionalization of carbon dioxide has
emerged as a feedstock of particular interest due to its
immense abundance, low toxicity, and identity as a green-
house gas emission.3 Unfortunately the utilization of CO2 in
chemical and related applications is currently limited to
roughly 130 MT year−1, largely due to a lack of reactions which
surmount the strong kinetic and thermodynamic stability
of the gas.4 Our laboratory and others have been working to
expand this scope of chemical CO2 utilization by developing
transition metal mediated processes which couple CO2 with
light olefins.5
Fig. 1 Hoberg’s nickelalactone synthesis from CO2.
inspired considerable effort to establish catalytic syntheses of
acrylates from CO2 and ethylene which could alter the renew-
ability of acrylic materials such as super water absorbing poly-
mers.1 Recently researchers have reported progress in this
endeavor by using strong external bases and Lewis acids to
promote acrylate formation from diphosphine nickelalactone
complexes.5,7 However, the reticence of nickelalactone species
to undergo β-hydride elimination processes is still a limiting
factor in the utilization of this coupling reaction.5 Thus, develop-
ing reactions of nickelalactone complexes which further func-
tionalize the metallacycle without necessitating β-elimination
may be an attractive method to leverage low-valent nickel
species which couple CO2 with light olefins.
The ring-expansion of nickelalactone complexes by inser-
tion of organic unsaturates has previously been observed for a
handful of diamine and diphosphine supported complexes.6
In most cases, saturated 5-membered nickelalactone species
were treated with an alkyne, activated olefin, or carbon dioxide
to afford a 7-membered metallacycle which was hydrolyzed to
obtain carboxylate products.6 Unfortunately, the addition of
acid makes development of these reactions into catalytic pro-
cesses a significant challenge given the incompatibility of
Bronsted acids and low-valent nickel. An alternative approach
is the insertion of strong field ligands which not only functio-
nalize the five-membered nickelalactone species, but also
induce elimination of up-converted organics derived from
The chemistry of CO2–olefin coupling has been an area of
interest since the seminal reports of C–C bond formation
between CO2 and ethylene by Hoberg and Carmona in the
mid-1980’s.6 In particular, Hoberg’s isolation of a nickelacycle
lactone (also termed nickelalactone) complex (Fig. 1) has
Department of Chemistry, Brown University, Providence, RI 02912, USA.
E-mail: wb36@brown.edu
†Electronic supplementary information (ESI) available. CCDC 997717–997719.
For ESI and crystallographic data in CIF or other electronic format see DOI:
10.1039/c4dt01221f
15990 | Dalton Trans., 2014, 43, 15990–15996
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
CO2. Herein, we describe the coupling of CO2 and light olefins C22H36NPNi: C, 65.37; H, 8.98, N, 3.47. Found: C, 64.99; H,
1
to generate new amino-phosphine nickelalactones with good 8.28; N, 3.38. H NMR (C6D6): δ 1.15 (d, 11.6 Hz, 18H, PtBu),
regioselectivity, and the use of strong field carbon monoxide 2.20 (m, 4H, CH2–COD), 2.44 (m, 4H, CH2–COD), 2.67 (d, 2H,
to afford valuable cyclic anhydride and 1,4-ketoacid products.
6 Hz, 2H, CH2–P), 2.80 (m, 2H, CH2–COD), 4.21 (m, 2H, CH–
COD), 4.82 (m, 2H, CH–COD), 6.35 (t, 1H, py), 6.51 (d, 1H, py),
6.80 (t, 1H, py), 8.95 (d, py). 31P {1H} NMR (C6D6): δ 68.5 (s).
Partial 13C NMR taken from 1H–13C HSQC (C6D6): δ 29.39
(PtBu), 31.04, 31.75 (CH2–COD), 33.26 (P–CH2), 78.99, 82.37
(CH–COD), 121.22, 121.65, 129.98, 151.02 (py).
Methods
General considerations
All manipulations were carried out using standard vacuum,
(PN)Ni(C2H4) (2a). A J. Young NMR tube was charged with
Schlenk, cannula or glovebox techniques. Ethylene and carbon 0.030 g (0.07 mmol) of 1a in approximately 500 µL of benzene-
dioxide were purchased from Corp Brothers and stored over d6. On a high vacuum line 5 equivalents of ethylene were
4 Å molecular sieves in heavy walled glass vessels prior to use. added via a calibrated gas bulb at −196 °C. The reaction
Argon and nitrogen were purchased from Corp Brothers and was allowed to warm to ambient temperature, immediately
used as received. 2-((di-t-butylphosphino)methyl)pyridine (PN) changing color from dark blue to dark pink. Full conversion
was prepared according to literature procedure.8 All other to 2a was confirmed by NMR spectroscopy. 1H NMR (C6D6):
chemicals were purchased from Aldrich, VWR, Strem, Fisher δ 1.13 (d, 12 Hz, 18H, PtBu), 2.36 (br, 4H, ethylene), 2.65 (d,
Scientific or Cambridge Isotope Laboratories. Volatile, liquid 6 Hz, 2H, CH2–P) 6.23 (t, 1H, py), 6.63 (d, 1H, py), 6.93 (t, 1H,
chemicals were dried over 4 Å molecular sieves and distilled py), 9.35 (d, py). 31P {1H} NMR (C6D6): δ 71.99. Partial 13C NMR
prior to use. Solvents were dried and deoxygenated using taken from 1H–13C HSQC (C6D6): δ 28.10 (PtBu), 31.04, 33.61
literature procedures.9
(C2H4), 33.24 (P–CH2), 121.51, 122.45, 132.77, 151.99 (py).
(PN)Ni(C2H3CH3) (2b). A J. Young NMR tube was charged
1H, 13C, and 31P NMR spectra were recorded on Bruker DRX
400 Avance, 300 and 600 MH Avance spectrometers. 1H and with 0.0075 g (0.032 mmol) of Ni(COD)2, 0.076 g (0.029 mmol)
13C chemical shifts are referenced to residual solvent signals; of PN ligand and approximately 500 µL of benzene-d6. On
31P chemical shifts are referenced to an external standard of a high vacuum line 10 equivalents (140 torr in 28.9 mL) of
H3PO4. Probe temperatures were calibrated using ethylene propylene were added via calibrated gas bulb at −196 °C. The
glycol and methanol as previously described.10 Unless other- reaction was allowed to warm to ambient temperature, quickly
wise noted, all NMR spectra were recorded at 23 °C. IR spectra changing color from dark blue to an intense purple. Analysis
were recorded on a Jasco 4100 FTIR spectrometer. GC-MS data by NMR showed near quantitative conversion to 2b, with two
were recorded using a Hewlett-Packard (Agilent) GCD 1800C isomers apparent by 31P NMR in a 2 : 1 ratio. Only the major
1
GC-MS spectrometer. X-ray crystallographic data were collected isomer was resolved in the H NMR and HSQC NMR spectra.
1
on a Bruker D8 QUEST diffractometer. Samples were collected The sample reverted to 1 upon removal of propylene. H NMR
in inert oil and quickly transfered to a cold gas stream. The (C6D6):
δ
1.15 (d, 12 Hz, 18H, PtBu), 1.89 (br, 3H,
structures were solved from direct methods and Fourier syn- H2CvCHCH3), 2.18 (br, 1H, H2CvCHCH3), 2.43 (br, 1H,
theses and refined by full-matrix least-squares procedures with H2CvCHCH3), 2.64 (br, 2H, CH2–P), 3.13 (br, 1H,
anisotropic thermal parameters for all non-hydrogen atoms. H2CvCHCH3) 6.26 (m, 1H, py), 6.63 (m, 1H, py), 6.93 (m, 1H,
Crystallographic calculations were carried out using SHELXTL. py), 9.34 (m, 1H, py). 31P {1H} NMR (C6D6): δ 72.41 (minor),
CIF data for 1, 2c, 3b has been deposited with the CCDC under 68.83 (major). Partial 13C NMR taken from 1H–13C HSQC
numbers 997718, 997717, and 997719, respectively. Elemental (C6D6):
δ
23.95 (H2CvCHCH3), 29.94 (PtBu), 32.09
analyses were performed at Atlantic Microlab, Inc., in Nor- (H2CvCHCH3), 33.67 (CH2–P), 121.83, 122.75, 132.88, 151.41
cross, GA. All geometry optimizations were performed using (py), H2CvCHCH3 not located due to broadening from
Gaussian 09 Revision A.02,11 using the M06 l functional. The exchange with free propylene (exchange confirmed by 2D
LANL2DZ basis set was used for Ni and the 6-31G++(d,p) basis NOESY with 300 ms mixing time).
set was used for all other atoms. The LANL2DZ pseudo-poten-
tial was used for Ni. Frequency calculations were performed on with a vacuum adapter was charged with 0.066 g (0.278 mmol)
all optimized structures to ensure that they were true minima. of Ni(COD)2, 0.076 (0.276 mmol) of PN ligand and approxi-
(PN)Ni(C2H3CN) (2c). A 25 mL round bottom flask fitted
(PN)Ni(COD) (1). A 20 mL scintillation vial was charged mately 10 mL of toluene. On a high vacuum line 2 equivalents
with 0.058 g (0.26 mmol) of Ni(COD)2 and approximately 5 mL (76 torr in 101 mL) of acrylonitrile were added via calibrated
of tetrahydrofuran. The solution was chilled to −35 °C and gas bulb at −196 °C. The reaction was allowed to warm to
0.074 g (0.31 mmol) of 2-((di-t-butylphosphino)methyl)pyri- ambient temperature, quickly changing color from dark blue
dine (PN) in diethyl ether was added slowly over 2 hours to orange. After stirring for 3 hours, the volatiles were
turning the solution dark blue. After stirring at ambient temp- removed, the orange residue rinsed with pentane and extracted
erature for a further 30 minutes, the volatiles were removed, with 5 mL of 1 : 1 tetrahydrofuran–diethyl ether. The orange
and the residue extracted with pentane. Concentrating and solution was chilled to −35 °C overnight to afford 0.043 g of 2c
chilling the pentane extraction to −35 °C overnight afforded as orange powder. Concentrating the mother liquor produced
49 mg (75%) of 1 as dark blue crystals. Anal. Calcd for a second 0.022 g fraction of product for a total yield of 0.065 g
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 15990–15996 | 15991

View Article Online
Paper
Dalton Transactions
(69%). Anal. Calcd for C17H27NiN2P: C, 58.49; H, 7.80; N, 8.02. microsyringe. The tube was quickly removed from the inert
Found: C, 58.09; H, 7.49; N, 8.02. 1H NMR (C6D6): δ 0.93 (d, atmosphere box and chilled to −196 °C. On a vacuum line 260
7.4 Hz, 18H, PtBu), 1.89 (br, 1H, H2CvCHCN), 2.16 (br, 1H, torr carbon monoxide in a 28.9 mL volume was placed over the
H2CvCHCN), 2.19 (br, 1H, H2CvCHCN), 2.51 (br, 2H, CH2– sample and allowed to equilibrate for 5 minutes. The J. Young
P), 6.27 (m, 1H, py), 6.50 (m, 1H, py), 6.79 (m, 1H, py), 9.29 (m, tube was sealed and thawed in warm water. The sample was
1H, py). 31P {1H} NMR (C6D6): δ 77.52. Partial 13C NMR taken allowed to react at ambient temperatures overnight, then
from 1H–13C HSQC (C6D6): δ 11.92 (H2CvCHCN), 28.51 (PtBu), treated with excess HCl gas. The solution was dried under
28.80 (H2CvCHCN), 32.38 (CH2–P), 122.07, 122.23, 134.86, vacuum and the extracted into CDCl3 where the 1,4-ketoacid
152.15 (py).
was quantified by NMR spectroscopy. Yield 46% from 3a; 14%
(PN)Ni(κC,κO-CH2CH2COO) (3a). A 50 mL Schlenk flask from 3b. The product may be further purified by sequential
was charged with 0.200 g (0.00073 mol) of Ni(COD)2, approxi- aqueous base-acid extraction.
mately 15 mL of tetrahydrofuran and placed under 1 atm
of ethylene. A tetrahydrofuran solution containing 0.172 g
(0.00073 mol) of PN ligand was added via syringe and stirred
for 20 minutes at ambient temperature. CO2 was then
Results and discussion
admitted to the reaction by sparging the solution for
The three decade history of nickel promoted coupling between
10 minutes. The reaction mixture was stirred overnight, and
CO2 and light olefins, in particular ethylene, has seen varied
the product, a light yellow powder, was collected by cannula
success with diamine and diphosphine ligands, but little
filtration to obtain 245 mg (91%) of 3a following rinsing with
investigation into the use of amino-phosphine chelates.12 Our
attention was drawn to amino-phosphine ligands due to the
dissymmetry in electron donation imposed on the metal
center which, when matched to dissymmetry in the C,O
chelate of a nickelalactone, could enhance opportunities for
regio- or even enantioselectivity in CO2–olefin coupling reac-
additional tetrahydrofuran. Anal. Calcd for C17H28NiNPO2: C,
55.48; H, 7.67; N, 3.81. Found: C, 55.81; H, 7.00; N, 3.47. 1H
NMR (acetone-d6): δ 0.86 (m, 2H, Ni-α-CH2) 1.43 (d, 12 Hz,
18H, tBu), 2.08 (m, 2H, Ni-β-CH2), 3.42 (d, 9 Hz, 2H, CH2–P),
7.29 (t, 1H, py), 7.56 (d, 1H, py) 7.87 (t, 1H, py), 8.86 (d, 1H,
py). 31P {1H} NMR (acetone-d6): δ 71.01 (s). Partial 13C NMR
tions. The disparate trans-influence between the strong field
taken from 1H–13C HSQC NMR (CD2Cl2): δ −1.7 (Ni-α-CH2),
phosphine coordination and weaker field amine coordination
29.1 (PtBu), 39.1 (Ni-β-CH2), 32.8 (CH2–P), 122.0, 122.2, 137.7,
should affect a preference to orient the nickelalactone with the
149.3 (py). 13C {1H} NMR of 3a labeled with 13CO2 (acetone-d6):
weak trans-influence carboxylate opposite the phosphine and
the stronger field methine group opposite the amine. This net
effect would allow judicious choice of the amino-phosphine
ligand substituents to reliably control the regioselectivity in
CO2–olefin coupling through steric interactions between the
phosphine and olefin substituents. This arrangement would
still leave the amino group tunable to permit ample access to
the metal for CO2 activation.
δ 185.51 (CO2).
(PN)Ni(κC,κO-CH2CH2(CH3)COO) (3b). A J. Young NMR
tube was charged with 0.028 g of PN ligand and 0.019 g of
Ni(COD)2 in tetrahydrofuran. On a high vacuum line, 3 equi-
valents of propene were added via calibrated gas bulb at
−196 °C. The sample was thawed and allowed to stand at
ambient temperature for 1 hour where it changed to a deep
pink solution. Then an additional 3 equivalents of CO2 were
added via calibrated gas bulb at −196 °C and the sample
The pyridyl-phosphine ligand, 2-((di-t-butylphosphino)-
methyl)pyridine (PN), was selected as a target for this methodo-
logy due to the large steric impetus of the t-butyl phosphine
allowed to stand overnight at ambient temperature. Monitor-
ing by 31P NMR spectroscopy showed complete disappearance
substituents matched with the relatively open pyridyl frag-
of 2b after that time, and then 30 mg (76%) of 3b was collected
ment.13 This chelate has recently been employed on heavier
as a yellow powder by filtration. Anal. Calcd for C18H30NiNPO2:
group 10 metals for cross-coupling catalysis and dioxygen
1
C, 56.58; H, 7.91; N, 3.67. Found: C, 57.09; H, 7.57; N, 3.39. H
insertion reactions.8,14 Coordination of PN to nickel was
NMR (CD2Cl2): δ 0.56 (br, 1H, Ni-α-CH2) 1.04 (br, 3H, Ni-
accomplished by treatment with a toluene solution of bis(1,5-
cyclooctadiene)nickel (Ni(COD)2) to afford modest yields of
γ-CH3), 1.22 (br, 1H, Ni-α-CH2), 1.35 (d, 14 Hz 18H, tBu), 2.36
(br, 1H, Ni-β-CH), 3.14 (br, 2H, CH2–P), 7.17 (br, 1H, py), 7.33
(PN)NiCOD (1) as a dark blue-purple solid upon crystallization
(br, 1H, py) 7.69 (br, 1H, py), 8.90 (br, 1H, py). 31P {1H} NMR
from pentane (eqn (1)).
(CD2Cl2): δ 71.97 (s). Partial 13C NMR taken from 1H–13C
HSQC NMR (CD2Cl2): δ 9.42 (Ni-α-CH2), 21.15 (Ni-γ-CH3), 29.46
(PtBu), 33.45 (CH2–P), 44.18 (Ni-β-CH), 122.33, 122.66, 137.75,
149.93 (py). 13C {1H} NMR of 3b labeled with 13CO2 (acetone-
d6): δ 186.71 (CO2).
General procedure for the formation of 1,4-ketoacids
ð1Þ
A J. Young NMR tube was charged with 0.020 g of either 3a or
3b in approximately 0.5 mL of THF. The solution was frozen in
(PN)NiCOD was characterized by NMR spectroscopy,
a liquid nitrogen cold well and 15 µL of neat Et2Zn injected via elemental analysis, and single crystal X-ray diffraction. The
15992 | Dalton Trans., 2014, 43, 15990–15996
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
suggesting a rapid exchange at ambient temperature. Propy-
lene species 2b also exhibited two resonances in the 31P NMR
spectrum in an approximate 2 : 1 ratio, likely originating from
isomers which vary in the orientation of the asymmetric
alkene. Complex 2c was further characterized by single crystal
X-ray diffraction (Fig. 4). The solid-state structure reveals an
approximate square planar geometry about the nickel with the
–CN substituent oriented away from the large PtBu2 ligand.
Additionally, the olefinic C(2)–C(3) bond length of 1.46(1) is
substantially elongated from a typical double bond indicative
of a strong metallacyclopropane resonance contribution. This
feature is also observed in other structurally characterized
nickel–acrylonitrile complexes.15
Following the characterization of complexes 2a–c, the olefin
species were used to explore the potential for coupling with
carbon dioxide. Treatment of 2a and 2b with 1 atm CO2 over
12 hours at ambient temperature resulted in successful oxi-
dative coupling to afford (PN)Ni(κC,κO-CH2CH2COO) (3a) and
(PN)Ni(κC,κO-CH2CH(CH3)COO) (3b) in good yields (Fig. 3).
The acrylonitrile complex 3c proved unsuitable for CO2 acti-
vation even under 50 bar at 50 °C, likely due to the poor
nucleophilicity engendered by the –CN substituent. Complexes
3a and 3b were both isolated as yellow solids with 3b showing
greater solubility in non-polar solvents than 3a. In acetone-d6,
Fig. 2 Molecular structure of 1 at 30% probability ellipsoids. All hydro-
gen atoms are omitted for clarity. Select bond distances (Å) and angles
(°): Ni(1)–P(1) 2.1948(4), Ni(1)–N(1) 1.993(1), P(1)–Ni(1)–N(1) 87.62(4).
solid state structure (PN)NiCOD of (Fig. 2) reveals a slightly dis-
torted tetrahedral coordination sphere common for four-co-
ordinate nickel 1,5-cyclooctadiene species. The dissymmetry in
the PN ligand steric environment is self-evident from the ellip-
soid plot and the five-membered metallacycle created by the
ancillary ligand is modestly buckled with the bridging methyl-
ene carbon (C6) displaced from the mean plane of the other
metallacycle atoms by 0.32 Å.
The (PN)NiCOD complex, either in isolated form or gener-
ated in situ, served as a source of “(PN)Ni0” for activation of
light olefins such as ethylene, propylene, and acrylonitrile. For
example, addition of excess ethylene gas (∼10 equiv.) to an
NMR sample of (PN)NiCOD in benzene-d6 resulted in an
immediate color change from blue-purple to an intense violet,
signaling conversion to an η2-ethylene complex, (PN)Ni(C2H4)
(2a) (Fig. 3). Analogous species were observed using propylene
(2b) and acrylonitrile (2c), though only 2c was successfully iso-
lated following evacuation of the reaction volatiles. Each of the
olefin species were characterized by NMR spectroscopy, with
1H NMR resonances for the vinyl hydrogens ranging between
1.89–3.13 ppm. Olefin species 2a and 2b exhibited broad 1H
NMR resonances for both the free and bound alkene,
1
the H NMR spectrum of 3a exhibited two multiple methylene
resonances for the metal lactone at 0.86 and 2.08 ppm, which
correlated to 13C NMR resonances of −1.7 and 39.1 ppm in
1
HSQC NMR experiments. Similarly, the H NMR spectrum of
3b in dichloromethane-d2 exhibited two diastereotopic hydro-
gens for the α-CH2 at 0.54 and 1.22 ppm as well as methine
and methyl resonances from the metallacycle at 2.36 and
1.04 ppm. 2D COSY and HSQC NMR spectra were used
to establish these assignments and the regioselectivity of
the CO2–propylene coupling confirmed by X-ray diffraction
(Fig. 5). The crystal sample of 3b suffered from moderate
twining problems, but the data offer clear indication of a
methacrylate-type structure. A notable feature of 3b structure
is the significant contortion of the metal lactone ring in which
C(2) is 0.24 Å removed from the mean plane. This contrasts
the highly planar arrangement observed for numerous unsub-
stituted γ-nickelalactone species.5,16 Geometry optimization
performed 3b shows a structure consistent with that observed
by X-ray diffraction.17 Additional calculations on the ground
Fig. 3 Substitution with light olefins and coupling with carbon dioxide.
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 15990–15996 | 15993

View Article Online
Paper
Dalton Transactions
nickel remain challenging, which led to the exploration of
other methods to leverage value-added products from nickela-
lactones 3a and 3b.
Carbonylation of 3a and 3b under moderate pressures of
CO (500 torr added at −196 °C) afforded high yields of succi-
ninc anhydride and methylsuccinic anhydride, respectively,
along with (PN)Ni(CO)2 (Fig. 6).19 The organic products
were identified by comparison to authentic samples and the
yields determined by NMR spectroscopy. Careful monitoring
of the reaction between 3a and CO in acetone-d6 by 31P NMR
spectroscopy indicated the formation of a low concentration
intermediate with a chemical shift of 68.20 ppm, which
decayed as the reaction completed. Repeating this experiment
using 3a derived from 13CO2 and 13CO resulted in splitting of
this 31P resonance into a doublet (2JC–P = 22 Hz) and afforded
two enhanced resonances in the 13C NMR spectrum at 252.44
(2JC–P = 22 Hz) and 174.75 ppm. This isotopic labeling also
caused alteration in the splitting pattern of two 1H NMR peaks
at 2.54 and 2.84 ppm. This spectral data led to tentative assign-
ment of the intermediate as the six-membered metallacycle
resulting from carbonyl insertion into the Ni–C bond of 3a
(Fig. 6), analogous to species observed by Rovis and Coates
during the dicarbonylation of succinic anhydride with nickel(0)
sources.
The preparation of succinic and methylsuccinic anhydride
from CO2 functionalization offers a potentially attractive
alternative to dehydration, oxidation, or bio-synthetic routes
for accessing these cyclic products which are used in paper
manufacturing and as pharmaceutical intermediates. Attempts
to conduct these reactions catalytically using CO2, olefin, and
CO were unsuccessful thermally, even at low CO partial press-
ures. This obstruction likely originates from the inability of the
olefin to coordinate to the substitutionally inert (PN)Ni(CO)2.
Such carbonyl substitution processes are often aided by high
energy photolysis which our laboratory is currently investi-
gating. However, the conversion of the nickelalactones to cyclic
anhydrides eliminates much of the value added from regio-
selective CO2–olefin coupling. Given this factor, our focus
Fig. 4 Molecular structure of 2c at 30% probability ellipsoids. All hydro-
gen atoms are omitted for clarity. Select bond distances (Å) and angles
(°): Ni(1)–P(1) 2.137(2), Ni(1)–N(1) 1.968(3), C(2)–C(3) 1.46(1), P(1)–Ni(1)–
N(1) 88.1(2).
state structure of 3a indicate a metal lactone ring distortion
comparable to 3b, suggesting this feature originates more
from an influence of the ancillary ligand than the alkylation of
the lactone.17
The regioselectivity for 1,2-coupling between propylene and
CO2 at ambient temperature (>90% by NMR) is remarkable
compared to the mixtures previously obtained from nickel
mediated α-olefin and CO2 oxidative cyclizations.6 In the case
of propylene, this selectivity is crucial to distinguishing
between precursors for methacrylate (methyl on the β-carbon)
and crotonate (methyl on the α-carbon) structures. The
observed preference for 1,2-coupling is particularly desirable
as methacrylate species are much more widely used monomers
with applications in durable thermoplastics such as Plexiglass,
optical fibers, and paints.18 As noted earlier, the β-hydrogen
elimination reactions required to liberate such acrylates from
Fig. 5 Molecular structure of 3b at 30% probability ellipsoids. Full view (left) and partial view (right). All hydrogen atoms and a cocrystallized solvent
molecule are omitted for clarity. Select bond distances (Å) and angles (°): Ni(1)–P(1) 2.136(6), Ni(1)–N(1) 2.00(1), Ni(1)–O(1) 1.87(1), Ni(1)–C(3) 1.94(3),
P(1)–Ni(1)–N(1) 85.1(4), O(1)–Ni(1)–C3(1) 84.2(8).
15994 | Dalton Trans., 2014, 43, 15990–15996
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
Fig. 6 Carbon monoxide insertion and organic product liberation from nickelalactone species.
shifted toward the liberation of γ-ketoacids, which are another
high value CO2 functionalization product.
Conclusions
Though the utilization of fossil fuel resources for meeting
energy and commercial chemical demands are quite different
in scale, both suffer for the same unsustainable paradigm. In
the commercial chemical industry, circumventing this limit-
ation will require the utilization of diverse renewable carbon
feedstocks including carbon dioxide. Our investigations
have discovered a new ligand–nickel platform which effectively
couples carbon dioxide with ethylene or propylene with good
regioselectivity. The addition of diethyl zinc and/or carbon
monoxide induces the elimination of high value organic pro-
ducts, including cyclic anhydrides and ketoacids. Interestingly,
this reaction also produces a zerovalent nickel dicarbonyl
product, which creates the possibility of reactivating the metal
for further coupling. Together, this sequence of reactions
opens potential pathways to synthetic procedures for accessing
commodity, biologically relevant, or chiral carboxylate pro-
ducts derived from CO2.
ð2Þ
Ketoacids are important precursors to multiple biosyn-
theses and are interesting components of bio-energy storage
mechanisms. Previous studies have shown that γ-ketoacids can
be obtained from the nickel catalyzed ring opening of cyclic
anhydrides in the presence of organonuclephiles, such as
diethyl zinc (eqn (2)). This led to investigation of the possible
synthesis of similar products sourced from CO2 and olefin
coupling. Initial attempts to liberate γ-ketoacids using 3a and
3b with 3 equiv. of diethyl zinc under 0.5 atm of CO resulted
in formation of the respective γ-ketoacids in low isolated yields
(26% for 3a; 14% for 3b) following hydrolysis (Fig. 6). In situ
monitoring of this reaction by 13C NMR spectroscopy using
13CO2 derived 3a revealed that the low yield of γ-ketoacid was
due, in part, to competitive formation of succinic anhydride.
To minimize this secondary reaction, the procedure was
repeated with a larger diethyl zinc to carbon monoxide ratio,
using 10 equivalents of Et2Zn under 25 torr of CO. This reac-
tion produced the zinc salt of the γ-ketoacid from 3a and 3b as
the sole 13C containing product by NMR spectroscopy. (PN)Ni-
(CO)2 again appeared as a prominent organometallic product,
obviating additional reaction. Acid hydrolysis and isolation
gave improved yields of 46% (from 3a) and 33% (from 3b).
While still modest, it is anticipated on the basis of the near
quantitative conversion observed by NMR spectroscopy that
the isolated yields from the acid/base extraction procedure
could improve substantially with larger scale reactions.
Acknowledgements
This material is based upon work supported by Brown Uni-
versity. We would like to thank Nilay Hazari for computational
assistance.
References
1 M. Cokoja, C. Bruckmeier, B. Rieger, W. A. Herrmann and
F. E. Kuhn, Angew. Chem., Int. Ed., 2011, 50, 8510; Renew-
able Raw Materials: New Feedstocks for the Chemical Industry,
ed. R. Ulber, D. Sell and T. Hirth, Wiley-VCH, Weinheim,
2011.
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 15990–15996 | 15995

View Article Online
Paper
Dalton Transactions
2 Developments and Innovations in Carbon Dioxide Capture and 10 J. Sandström, Dynamic NMR Spectroscopy, Academic Press,
Storage Technology. Vol 2: Carbon Dioxide Storage and Utilis-
ation, ed. M. M. Maroto-Valer, Woodhead, Cambridge, 2010.
3 Carbon Dioxide as Chemical Feedstock, ed. M. Aresta, Wiley-
VCH, Weinheim, 2011.
4 W. B. Tolman, Carbon Dioxide Reduction and Uses as a
Chemical Feedstock. Activation of Small Molecules: Organo-
metallic and Bioinorganic Perspectives, Wiley-VCH,
Weinheim, 2006, pp. 1–35.
5 (a) M. L. Lejkowski, R. Lindner, T. Kageyama, G. E. Bodizs,
P. N. Plessow, I. B. Muller, A. Schafer, F. Rominger,
P. Hofmann, C. Futter, S. A. Schunk and M. Limbach,
Chem. – Eur. J., 2012, 18, 14017; (b) C. Bruckmeier,
M. W. Lehenmeier, R. Reichardt, S. Vagin and B. Rieger,
Organometallics, 2010, 29, 2199; S. Y. T. Lee, M. Cokoja,
M. Drees, Y. Li, J. Mink, W. A. Herrmann and F. E. Kuhn,
ChemSusChem, 2011, 4, 1275; (c) J. M. Wolfe and
W. H. Bernskoetter, Dalton Trans., 2012, 41, 10763;
(d) P. N. Plessow, L. Weigel, R. Lindner, A. Schafer,
F. Rominger, M. Limbach and P. Hofmann, Organometal-
lics, 2013, 32, 3327; (e) W. H. Bernskoetter and B. T. Tyler,
New York, 1982.
11 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone,
B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato,
X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng,
J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda,
J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao,
H. Nakai, T. Vreven, J. A. Montgomery Jr., J. E. Peralta,
F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers,
K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand,
K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar,
J. Tomasi, M. Cossi, N. Rega, N. J. Millam, M. Klene,
J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo,
R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin,
R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin,
K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador,
J. J. Dannenberg, S. Dapprich, A. D. Daniels, Ö. Farkas,
J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox,
Gaussian 09, Revision A.02, Gaussian, Inc., Wallingford CT,
2009.
Oganometallics, 2011, 30, 520; (f) S. Y. T. Lee, A. A. Ghani, 12 H. Hoberg, A. Ballesteros, A. Sigan, C. Jegat, D. Barhausen
V. D’Elia, M. Cokoja, W. A. Herrmann, J. M. Basset and and A. Milchereit, J. Organomet. Chem., 1991, 407, C23.
F. E. Kuhn, New J. Chem., 2013, 37, 3512; (g) Y. Zhang, 13 P. G. Edwards I. A. Fallis and B. S. Yong, British Patent
B. S. Hanna, A. Dineen, P. G. Williard and
W. H. Bernskoetter, Organometallics, 2013, 32, 3969.
Application GB, 2378182 A, 2003.
14 K. A. Grice and K. I. Goldberg, Organometallics, 2009, 28,
6 (a) H. Hoberg and D. Schaefer, J. Organomet. Chem., 1983,
953.
251, C51; (b) H. Hoberg, D. Schaefer, G. Berukhart, 15 L. J. Guggenberger, Inorg. Chem., 1973, 12, 499.
C. Kruger and M. Romao, J. Organomet. Chem., 1984, 266, 16 (a) J. Langer, D. Walther, A. Malassa, M. Westerhausen and
203; (c) H. Hoberg, Y. Peres, C. Kruger and Y. H. Tsay,
Angew. Chem., Int. Ed. Engl., 1987, 26, 771.
7 (a) J. Dong, P. G. Williard, N. Hazari and
W. H. Bernskoetter, Chem. – Eur. J., 2014, 20, 3205;
H. Gorls, Eur. J. Inorg. Chem., 2010, 275; (b) J. Langer,
D. Walther and H. Gorls, J. Organomet. Chem., 2006, 691,
4874; (c) J. Langer, H. Gorls, R. Fischer and D. Walther,
Organometallics, 2005, 24, 272.
(b) J. Dong, T. J. Schmeier, P. G. Williard, N. Hazari and 17 See ESI.†
W. H. Bernskoetter, Organometallics, 2013, 32, 2152.
8 L. Zhang, D. Peng, X. Leng and Z. Huang, Angew. Chem.,
Int. Ed., 2013, 52, 3676.
18 W. Bauer Jr., Methacrylic Acid and Derivatives, Ullmann’s
Encyclopedia of Industrial Chemistry, Wiley-VCH, Weinheim,
2011.
9 A. B. Pangborn, M. A. Giardello, R. H. Grubbs, R. K. Rosen 19 The identify of (PN)Ni(CO)2 was confirmed by independent
and F. J. Timmers, Organometallics, 1996, 15, 1518. preparation from addition of CO to 1.
15996 | Dalton Trans., 2014, 43, 15990–15996
This journal is © The Royal Society of Chemistry 2014