Effect of sodium cation on metallacycle β-hydride elimination in CO2-ethylene coupling to acrylates

Article abstract of DOI:10.1002/chem.201304196

The catalytic conversion of carbon dioxide and olefins into acrylates has been a long standing target, because society attempts to synthesize commodity chemicals in a more economical and sustainable fashion. Although nickel complexes have been known to successfully couple CO₂ and ethylene for decades, a key β-hydride elimination step has proven a major obstacle to the development of a catalytic process. Recent studies have shown that Lewis acid additives can be used to create a lower-energy pathway for β-hydride elimination and facilitate a low number of catalytic turnovers. However, the exact manner, in which the Lewis acid promotes β-hydride elimination remains to be elucidated. Herein, we describe the kinetic and thermodynamic role that commercially relevant and weakly Lewis acidic sodium salts play in promoting β-hydride elimination from nickelalactones synthesized from CO₂ and ethylene. This process is compared to a non-Lewis acid promoted pathway, and DFT calculations were used to identify differences between the two systems. The sodium-free isomerization reaction gave a rare CO ₂-derived β-nickelalactone complex, which was structurally characterized. A sodium scramble: Sodium cations can be used to facilitate a β-hydride elimination reaction in CO₂-derived nickelalactones, which is key to the production of acrylates. A related isomerization reaction can also occur without Lewis acid to produce rare unstabilized γ-nickelalactones, but with a higher activation energy and less favorable thermodynamics than the sodium-containing counterpart (see scheme).

Full text of DOI:10.1002/chem.201304196

DOI: 10.1002/chem.201304196
Full Paper
&
Lactones
Effect of Sodium Cation on Metallacycle b-Hydride Elimination in
CO2–Ethylene Coupling to Acrylates
Dong Jin,[a] Paul G. Williard,[a] Nilay Hazari,[b] and Wesley H. Bernskoetter*[a]
Abstract: The catalytic conversion of carbon dioxide and
olefins into acrylates has been a long standing target, be-
cause society attempts to synthesize commodity chemicals
in a more economical and sustainable fashion. Although
nickel complexes have been known to successfully couple
CO2 and ethylene for decades, a key b-hydride elimination
step has proven a major obstacle to the development of
a catalytic process. Recent studies have shown that Lewis
acid additives can be used to create a lower-energy pathway
for b-hydride elimination and facilitate a low number of cat-
alytic turnovers. However, the exact manner, in which the
Lewis acid promotes b-hydride elimination remains to be
elucidated. Herein, we describe the kinetic and thermody-
namic role that commercially relevant and weakly Lewis
acidic sodium salts play in promoting b-hydride elimination
from nickelalactones synthesized from CO2 and ethylene.
This process is compared to a non-Lewis acid promoted
pathway, and DFT calculations were used to identify differen-
ces between the two systems. The sodium-free isomerization
reaction gave a rare CO2-derived b-nickelalactone complex,
which was structurally characterized.
Introduction
the coupling of CO2 and ethylene has received considerable at-
tention.[4] Hoberg and co-workers were among the first to
report a well-defined coupling of these gases by using the bis-
(dicyclohexylphosphino)ethane (dcpe) ligand with nickel-1,5,9-
The functionalization of CO2 into value-added chemicals has
the potential to profoundly impact the cost and sustainability
of many consumer products.[1] Although it is unlikely that the
transformation of CO2 into chemicals will directly mitigate cli-
mate change originating from anthropogenic release, CO2 is
a nearly ideal C1 feedstock for the production of commodity
chemicals.[2] The primary barrier to widespread utilization of
CO2 is its significant thermodynamic stability. One method to
circumvent this obstacle is coupling CO2 with higher-energy
small molecules, which can provide reduction potential to help
drive the CO2 functionalization. A limited number of large-
scale commodity chemical processes already employ this tech-
nique, including the synthesis of urea from CO2 and ammo-
nia.[3] Another attractive target, which could dramatically
expand this methodology, is the functionalization of CO2 with
olefins to produce acrylates, a family of polar monomers used
heavily in fibers, fabrics, and super water absorbent polymers.
Currently, acrylates are industrially synthesized by the oxida-
tion of propylene, a non-renewable carbon source, cost of
which tightly controls that of most acrylates. As a result, a po-
tential alternative route to acrylic acid and its derivatives by
cyclododecatriene
to
produce
the
g-nickelalactone,
[Ni(dcpe)(kC,kOÀCH2CH2COO] (1-glactone).[5] Although this
seminal contribution demonstrated CÀC bond formation be-
tween CO2 and ethylene, 1-glactone and most subsequently
described nickelalactones species have been described as
unable to undergo the b-hydride elimination reaction necessa-
ry to form acrylate.[6]
In the last five years, several groups have found that excess
methylating agents can induce b-hydride elimination from g-
nickelalactone complexes giving limited quantities of acry-
late.[7] However, the use of methylating agents results in the
formation of oxidized metal species unsuitable for further cou-
pling of CO2 and ethylene. More recent investigations by Lim-
bach and co-workers have developed an exciting two-stage re-
action, in which strong external bases react directly with nick-
elalactone complexes to produce multiple equivalents of acry-
late in the presence of sodium.[8] Unfortunately, in these sys-
tems, the direct reaction between the external base and CO2
hinders a single-stage catalytic process, and the exact role of
the sodium cation is unclear. Recently, our group has reported
that the relatively strong Lewis acid tris(pentaflurophenyl)bor-
on can promote b-hydride elimination from a 1,1’-bis(diphenyl-
phosphino)ferrocene (dppf)-supported g-nickelalactone spe-
cies.[9] This activation by Lewis acid suggested that catalytic ac-
rylate formation could be achieved by use of an electron-ac-
cepting co-catalysts, though this proved improbable for the
dppf–nickel system, because the CO2–ethylene coupling did
not occur. Therefore, we were interested in exploring the role
[a] D. Jin, Prof. P. G. Williard, Prof. W. H. Bernskoetter
Department of Chemistry
Brown University, Providence, RI 02912 (USA)
E-mail: wb36@brown.edu
[b] Prof. N. Hazari
Department of Chemistry
Yale University, New Haven, CT 06520 (USA)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201304196.
Chem. Eur. J. 2014, 20, 3205 – 3211
3205
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
of Lewis acid activators in promoting the key b-hydride elimi-
nation reaction in g-nickelalactone complexes generated from
CO2–ethylene coupling. Herein, we describe the role of sodium
cation in promoting b-hydride elimination from a nickelalac-
tone. We believe that this is the weakest Lewis acid, which is
likely to be sufficient for this purpose. Furthermore, both theo-
retical and experimental studies have been used to compare
the Lewis acid promoted pathway with a non-Lewis acid pro-
moted pathway, and the role of the sodium cation has been
elucidated. Studies of the non-Lewis acid promoted pathway
for b-hydride elimination have led to the characterization of
the first example of an unactivated b-nickelalactone synthe-
sized directly from CO2–ethylene coupling.
to observations using the tBu substituted congener, bis(di-tert-
butylphosphino)methane.[8]
The convenient isolation of the g-nickelalactone complexes
enabled study of the role of Lewis acids in promotion of b-hy-
dride elimination reactions, in systems which in principle could
be catalytically relevant due to their facile synthesis from CO2
and ethylene. A particular focus was placed upon identifying
a Lewis acid, which binds reversibly to metallactones creating
the potential for a catalyzed elimination reaction. Inspired by
Limbach and co-workers’ empirical observation that exogenous
sodium salts were essential to acrylate formation from g-nickel-
alactone complexes in highly basic environments, the effect of
F
sodium tetrakis[3,5-bis(trifluoromethyl)phenyl]borate (NaBAr4 )
was investigated.[8] Treatment of an ethereal or arene solution
F
of 1-glactone with NaBAr4 immediately resulted in coordina-
Results and Discussion
tion of the sodium cation to the metallactone to produce
1-glactoneNa (Figure 3). The influence of the Lewis acid was
evident by a subtle shift of the lactone methylene resonances
A modest number of g-nickelalactone complexes derived from
CO2 and ethylene have been reported since Hoberg’s et al. ini-
tial discovery,[8] but given its historical significance to the field,
1-glactone and a closely related analogue using bis(dicyclohex-
ylphosphino)methane (dcpm; 2-glactone) were selected for in-
vestigation.[11] In our hands, the g-nickelalactone species were
best synthesized by treatment of Ni(COD)2 (COD=1,5-cyclooc-
tadiene) with diphosphine in the presence of excess CO2 and
ethylene (Figure 1). This permits use of a convenient and com-
mercially available nickel source. The successful preparation of
Figure 1. Synthesis of g-nickelalactone species from CO2–ethylene coupling.
Figure 3. Sodium activation and isomerization of nickelalactone species.
1-glactone was confirmed by comparison to previously report-
ed spectral data, as well as by X-ray diffraction (Figure 2). The
Ni center in 1-glactone displays a distorted square-planar ge-
ometry typical for g-nickelalactones,[8] with the five-membered
lactone ring close to planarity. The related complex 2-glactone
was also characterized by NMR spectroscopy and combustion
analysis. Interestingly, 2-glactone proved quite stable to isola-
tion in the absence of a CO2 or ethylene atmosphere, contrary
1
in the H NMR spectrum to d=0.82 and 2.31 ppm from 0.89
and 2.60 ppm, respectively. In the solid-state structure
(Figure 2), 1-glactoneNa is a dimer with each sodium bridging
between the carboxylate oxygen atoms of one metallacycle
and the exocyclic carboxylate oxygen atom of another. Nota-
bly, the Ni(1)ÀO(2) bond length in 1-glactoneNa (1.937(2) ꢁ) is
slightly elongated compared to that of 1-glactone (1.880(3) ꢁ),
suggesting a mild weakening of the nickel–oxygen bond,
which may assist in ring opening and b-elimination.[9] Interest-
ingly, heating 1-glactoneNa in THF at 558C for several hours in-
duced partial isomerization to the sodium-activated b-nickela-
lactone, 1-blactoneNa, with a Keq of 0.28(2). This process is for-
mally equivalent to b-hydride elimination, followed by subse-
quent 2,1-insertion from a transient nickel(II) acrylate hydride
intermediate. Van’t Hoff analysis of the interconversion be-
tween 1-glactoneNa and 1-blactoneNa gave parameters of
DS8=15(2) JmolÀ1 K and DH8=8.5(2) kJmolÀ1 suggesting
a small entropic preference for 1-blactoneNa. Analogous iso-
merization behavior was observed for the dcpm congener,
Figure 2. Molecular structures of 1-glactone (left), 1-glactoneNa (middle),
and 1-blactone (right) with ellipsoids at 30% probability level. All hydrogen
atoms, counteranions, and co-crystallized solvent molecules are omitted for
clarity.
Chem. Eur. J. 2014, 20, 3205 – 3211
www.chemeurj.org
3206
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
with a Keq of 0.13(3) (558C) between 2-glactoneNa and 2-blac-
toneNa.[14]
elalactones compared to the respective sodium adducts,
a strong qualitative thermodynamic preference for 1-glactone
over 1-blactone was still observed.
Over the course of our investigations of 1-glactone, it was
observed that this species undergoes reductive decoupling to
expel CO2 and ethylene at elevated temperature, as has been
previously reported.[5] Yet surprisingly, when the thermolysis of
1-glactone was conducted under CO2 and ethylene at 558C for
one day, partial conversion to an unactivated b-nickelalactone,
1-blactone, was observed even in the absence of Lewis acid
(Figure 4). This remarkable species was characterized by a com-
The faster equilibration between nickelalactones with
a sodium cation present offers a significant window into the
role of Lewis acid in promoting the key b-elimination step pro-
posed for catalytic CO2–ethylene coupling routes.[6] To gain
greater insight into this process, DFT calculations were per-
formed. In the calculations, the cyclohexyl substituents on
dcpe were replaced with methyl groups (denoted by the addi-
tion of ’ after the compound number, so 1 with methyl sub-
stituents on the phosphines is written as 1’). The validity of
this model was confirmed by performing selected calculations
with more sterically hindered substituents on the phosphine li-
gands.[11] For structures including sodium ions, three THF mole-
cules and two oxygen atoms of the metallactone molecule
were used to form the coordination sphere around sodium.
Consistent with the observed trend, the calculations indicated
a stabilization of the b-nickelalactone by sodium. The DFT
model predicted a 1.84 kcalmolÀ1 preference for 1’-glactone
over 1’-blactone, whereas 1’-blactoneNa was 9.77 kJmolÀ1
more stable than 1’-glactoneNa (Figure 6). In the optimized
structures of 1’-glactone and 1’-glactoneNa, both five-mem-
bered metallactone rings are nearly planar, closely matching
the experimental structures. In contrast, the inclusion of
a sodium ion substantially alters the planarity of the four-mem-
bered metallacycle. Sodium coordination increases the 19.28
torsion angle computed for 1’-blactone to 29.68 in 1’-blactone-
Na. Thus, the origin of the stabilization of b-nickelalactone by
sodium appears to be amelioration of ring strain associated
with the four-membered metallacycle.
Figure 4. Formation of an unstabilized b-nickelalactone complex.
bination of multinuclear 1D and 2D NMR spectroscopy, as well
as single-crystal X-ray diffraction. The solid-state structure of
1-blactone is a rare example of an unactivated four-membered
metallactone (i.e., no stabilizing group is present on the exocy-
clic carbonyl oxygen) and the first such species produced from
CO2–olefin coupling (Figure 2).[12] The metrical parameters
show a clear distinction between the carbon–oxygen bond
lengths of 1.214(6) (C(1)ÀO(1)) and 1.322(6) ꢁ (C(1)ÀO(2)) indi-
cating strong localization of the double bond. The Ni(1)ÀO(2)
distance of 1.900(3) ꢁ is analogous to those observed in relat-
ed activated b-nickelalactone structures.[7] Substantial pucker-
ing of the four-membered lactone ring was also observed with
a torsion angle of approximately 208. Notably, isomerization of
2-glactone to its corresponding unactivated b-nickelalactone
was not observed under analogous conditions.[11]
The pathway for the gamma-to-beta isomerization and the
kinetic influence of the sodium cation were also probed com-
putationally (Figure 6). Related calculations by Limbach and
co-workers have proposed that complexes featuring agostic in-
teractions play a key role in facilitating isomerization between
cationic nickelalactone species, in which the metallactone was
stabilized by methylation.[8] The key intermediate in that
system was an olefin hydride complex, and a similar structure
was located for the interconversion of 1’-glactone/1’-blactone
and 1’-glactoneNa/1’-blactoneNa (Figure 6).[11] However, due to
a relatively flat potential energy surface and a sharp energetic
well leading to the formation of 1’-glactone and 1’-blactone
(or 1’-glactoneNa and 1’-blactoneNa), it was not possible to
optimize structures for the agostic intermediates without plac-
ing a constraint on the NiÀO bond, except in the case of
1’-bagosticNa.[14] Nevertheless, comparison of the relative ener-
gies of the nickel–olefin hydride intermediates suggests
a lower-energy isomerization pathway in the presence of
Non-equilibrium mixtures of 1-glactone and 1-blactone were
also generated by treating a dcpe/Ni(COD)2 solution with acryl-
ic acid, which first gave [Ni(dcpe)(h2-CH2=CHCO2H)] (1-AA), fol-
lowed by isomerization to an approximate 2:1 initial ratio of
the metallacycles favoring 1-blactone (Figure 5).[11] Unfortunate-
ly, attempts to determine the equilibrium constant between
the unactivated nickelalactones by thermolysis under CO2 and
ethylene were obviated by partial sample degradation into
a mixture containing [Ni(dcpe)(kO,kO-CO3)] and other nickel
compounds.[13] Despite the lower stability of unactivated nick-
Figure 5. Alternative generation of 1-glactone/1-blactone mixture from acrylic acid.
Chem. Eur. J. 2014, 20, 3205 – 3211
www.chemeurj.org
3207
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 6. Energies of intermediates in the isomerization of a) 1’-glactone to 1’-blactone and b) 1’-glactoneNa to 1’-blactoneNa.
sodium. Relative to the respective g-nickelalactones, the
nickel–olefin hydride intermediate was almost 40 kJmolÀ1
lower in energy when activated by sodium. Furthermore, no
direct pathway for b-hydride elimination was identified.[6b] The
dramatic stabilizing effect of the sodium cation is likely from
compensation for the negative charge built up on the carbox-
ylate. This would also explain the higher temperatures necessa-
ry for isomerization of 1’-glactone/1’-blactone compared to
methylated nickelalactone species, in which no negatively
charged carboxylate is present.[8]
sion of ring stain and 2) it lowers the barrier to isomerization
by stabilizing the charge on high energy carboxylate anion in-
termediate. Our results are consistent with the hypothesis that
as the Lewis acidity of the promoter increases and a stronger
acid–oxygen bond is formed, the process of b-hydride elimina-
tion becomes more facile. Thus, a key for effecting catalytic
CO2–olefin coupling to acrylate is the use of Lewis acid co-cat-
alysts, which can promote b-hydride elimination, but still bind
weakly enough to allow its removal from substrate. The
sodium cation, particularly desirable for the synthesis of
sodium acrylate, appears to be one of the weakest Lewis acids,
which may be utilized, because the non-promoted pathway
occurs with only mildly greater difficulty.
Conclusion
We have observed a gamma-to-beta isomerization between
a diphosphine-supported nickelalactone with and without
sodium present, which has permitted a combined experimen-
tal and theoretical elucidation of the role of this highly attrac-
tive Lewis acid in the key b-hydride elimination step required
for catalytic acrylate formation. The sodium cation is important
in isomerization for two reasons: 1) it provides thermodynamic
stabilization of the b-nickelalactone structure through suppres-
Experimental Section
General
F
NaBAr4 was prepared according to the previously described proce-
dure.[15] All air- and moisture-sensitive manipulations were carried
out by using standard vacuum line, Schlenk, cannula, or glove-box
Chem. Eur. J. 2014, 20, 3205 – 3211
www.chemeurj.org
3208
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
techniques. Volatile chemicals were dried with 4 ꢁ molecular sieves
and distilled prior to use. All chemicals were purchased from Al-
drich, VWR, Fisher Scientific, Strem, or Cambridge Isotope Labora-
tories. Solvents were dried and deoxygenated by using literature
procedures.[16]
extracted with THF to remove trace nickel metal particulates. The
extract was then dried to give 0.208 g (89%) of 1-glactone with
good purity. The material may be further purified by layering n-
pentane on a concentrated THF solution and chilling at À358C.
The spectral data presented below are in agreement with those
1
1H, 13C, 19F, and 31P NMR spectra were recorded on Bruker DRX 400,
previously reported by Hoberg and co-workers. H NMR (400 MHz,
1
Avance 300 and 600 MHz spectrometers. H and 13C NMR chemical
[D6]benzene): d=0.89 (m, 2H, (PCy2)), 1.00 (m, 2H, Ni-a-CH2), 1.02–
1.65 (42H, PCy2), 2.02 (m, 2H, PCH2CH2P), 2.20 (m, 2H, PCH2CH2P),
2.85 (m, 2H, Ni-b-CH2); 13C{1H} NMR (100 MHz, [D6]benzene): d=
11.64 (Ni-a-CH2), 24.70, 26.42, 27.13–27.47, 29.17 (PCy2-CH2), 29.45
(PCH2CH2P), 29.70 (PCH2CH2P), 33.80, 35.30 (PCy2-CH), 37.76 (Ni-b-
CH2), 188.83 (CO2); 31P{1H} NMR (162 MHz, [D6]benzene): d=63.65
(bs s, 1P, PCy2), 69.79 (br s, 1P, PCy2); elemental analysis calcd for
C29H52NiO2P2 (553.36): C 62.94, H 9.47; found: C 63.21, H 9.77.
shifts are referenced to residual protio solvent signals; 19F and
31P NMR chemical shifts are referenced to the external standards
C6H5CF3 and H3PO4, respectively. In some cases, the inequivalent
31P nuclei have coupling constants too small to be observed due
to the combination of 2J and 3J couplings, these are denoted as
singlets. Probe temperatures were calibrated by using ethylene
glycol, as was previously described.[17] Unless otherwise indicated,
spectra were obtained at ambient temperature. Elemental analyses
were performed at Robertson Microlit Laboratories, Inc., in Madi-
son, NJ or Atlantic Micro Labs in Norcross, GA. X-ray crystallograph-
ic data were collected on a Bruker D8 Quest diffractometer. Sam-
ples were collected in inert oil and quickly transferred to a cold
gas stream. The structures were solved from direct methods and
Fourier syntheses and refined by full-matrix least-squares proce-
dures with anisotropic thermal parameters for all nonhydrogen
atoms. Crystallographic calculations were carried out by using
SHELXTL. CCDC-961246 (1-glactone), CCDC-961247 (1-blactone),
and CCDC-961248 (1-glactoneNa) contain the supplementary crys-
tallographic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
Synthesis of [Ni(dcpm)(kC,kOÀCH2CH2COO)] (2-glactone)
A heavy-walled glass reaction vessel (50 mL) was charged with
dcpm (0.309 g, 0.757 mmol), Ni(COD)2 (0.207 g, 0.753 mmol), and
THF (ca. 6 mL). On a high-vacuum line, ethylene (420 Torr in
101 mL) followed by carbon dioxide (700 Torr in 101 mL) were ad-
mitted to the reaction mixture at À1968C through a calibrated gas
bulb. After the mixture was stirred for one day at 508C, the vola-
tiles were removed in vacuo. The residue was washed with diethyl
ether and extracted with THF to remove trace nickel metal particu-
lates. The extract was then dried to give 2-glactone (0.190 g, 47%)
as orange powder containing minor amounts of [Ni(COD)(dcpm)].
The material may be further purified by layering n-pentane on
a concentrated THF solution and chilling at À358C. Alternatively,
2-glactone may be obtained by charging a scintillation vial (20 mL)
with dcpm (0.152 g, 0.372 mmol), Ni(COD)2 (0.103 g, 0.374 mmol),
and THF (ca. 3 mL). After the solid was dissolved, acrylic acid
(29 mL, 0.423 mmol) was injected into the reaction mixture. The
mixture was stirred at RT for 30 min, and the desired product was
isolated as described above. 1H NMR (400 MHz, [D6]benzene): d=
1.00 (m, 2H, Ni-a-CH2), 1.03–1.66 (40H, PCy2, (PCy2)), 1.95–2.08 (6H,
PCH2P & PCy2), 2.65 ppm (m, 2H, Ni-b-CH2); 13C{1H} NMR (100 MHz,
[D6]benzene): d=8.24 (Ni-a-CH2), 25.83, 26.25, 27.28–27.43, 28.67,
29.27, 29.56, 29.69, 30.23 (PCy2-CH2), 34.41, 34.57 (PCy2-CH),
37.37 ppm (Ni-b-CH2), one quaternary signal was not located;
31P{1H} NMR (162 MHz, [D6]benzene): d=À7.11 (d, 2JPÀP =23.3 Hz,
Computational details
All geometry optimizations were performed by using Gaussian 09,
Revision A.02.[18] Initial geometries were obtained by using the co-
ordinates from X-ray structures when possible. The M06 L function-
al was used in all calculations after evaluation of a variety of differ-
ent functionals (see the Supporting Information). The LANL2DZ
basis set was used for Ni and the 6–31G+ +(d,p) basis set was
used for all other atoms. The LANL2DZ pseudo-potential was used
for Ni. For calculations involving systems containing sodium,
QMMM calculations were performed by using ONIOM(M06 L:UFF),
as implemented in Gaussian. In these calculations, three THF mole-
cules were used to complete the coordination sphere around
sodium, which also included the oxygen atoms from the lactone.
These THF molecules were modeled at the MM level, with the ex-
ception of the oxygen atoms directly bound to the sodium, which
were part of the QM section of the molecule. Frequency calcula-
tions were performed on all optimized structures to ensure that
they were true minima. Solvent was modeled by using the integral
equation polarizable continuum model (THF), as implemented in
Gaussian. All energies presented are Gibbs free energies with sol-
vent corrections.
2
1P, PCy2), 17.73 ppm (d, JPÀP =23.3 Hz, 1P, PCy2); elemental analysis
calcd for C28H50NiO2P2 (539.33): C 62.35, H 9.34; found: C 62.08, H
9.13.
F
Synthesis of [Ni(dcpe)(kC,kOÀCH2CH2COONa)][BAr4 ]
(1-glactoneNa)
A scintillation vial (20 mL) was charged with 1-glactone (0.058 g,
F
0.105 mmol), NaBAr4 (0.099 g, 0.112 mmol), and THF (ca. 3 mL).
After the mixture was stirred at RT for 10 min, the volatiles were re-
moved in vacuo. The residue was washed with benzene and ex-
tracted with diethyl ether. Concentrating the solution, layering
with n-pentane and chilling to À358C gave 1-glactoneNa (0.145 g,
96%) as orange crystals. 1H NMR (400 MHz, [D5]bromobenzene):
d=0.82 (m, 2H, Ni-a-CH2), 1.1–1.72 (44H, PCy2), 1.94, 2.02 (m, 4H,
Synthesis of [Ni(dcpe)(kC,kOÀCH2CH2COO)] (1-glactone)
This species was prepared by modification of procedure described
by Hoberg and co-workers by using Ni(COD)2 as the metal
source.[18] A 50 mL heavy-walled glass reaction vessel was charged
with dcpe (0.178 g, 0.421 mmol), Ni(COD)2 (0.118 g, 0.429 mmol),
and THF (ca. 6 mL). On a high-vacuum line, ethylene (5 equiv,
390 Torr in 101 mL) followed by carbon dioxide (5 equiv, 390 Torr
in 101 mL) was admitted to the reaction mixture at À1968C
through a calibrated gas bulb. After the mixture was stirred for
one day at 508C, the volatiles were removed in vacuo. The residue
was washed with diethyl ether (ca. 3 mL) and toluene (1 mL), and
F
PCH2CH2P), 2.31 (m, 2H, Ni-b-CH2), 7.63 (s, 4H, BAr4 ), 8.18 ppm (s,
8H, BAr4 ); 13C{1H} NMR (100 MHz, [D5]bromobenzene): d=12.11
F
(Ni-a-CH2), 25.31–25.78, 26.57–26.93, 28.60, 28.88, 29.37, 29.80
(PCy2-CH2), 33.49, 35.43 (PCy2-CH), 36.67 (Ni-b-CH2), 117.48, 123.26,
F
134.86, 162.11 ppm (BAr4 )M; one aryl and one quaternary signal
was not located; 31P{1H} NMR (162 MHz, [D5]bromobenzene): d=
60.62 (s, 1P, PCy2), 70.93 ppm (s, 1P, PCy2); 19F NMR (C6D5Br): d=
Chem. Eur. J. 2014, 20, 3205 – 3211
www.chemeurj.org
3209
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
À62.95 (s); elemental analysis calcd for C61H64BF24NaNiO2P2
Observation of [Ni(dcpm)(kC,kOÀCH(CH3)COO)] (2-blactone)
(1439.56): C 50.89, H 4.48; found: C 50.94, H 4.71.
A scintillation vial (20 mL) was charged with dcpm (0.052 g,
0.127 mmol) and Ni(COD)2 (0.035 g, 0.127 mmol). After acrylic acid
(10 mL, 0.146 mmol) was injected into the vial, THF (ca. 3 mL) was
added in as the solvent. The mixture was stirred at RT for 30 min,
and the volatiles were removed in vacuo. The residue was washed
with diethyl ether (ca. 8 mL) and dried to give 2-blactone and
2-glactone mixture (0.057 g, 83%) highly favoring the g-nickelalac-
F
Synthesis of [Ni(dcpm)(kC,kOÀCH2CH2COONa)][BAr4 ]
(2-glactoneNa)
A scintillation vial (20 mL) was charged with 2-glactone (0.040 g,
0.074 mmol), NaBAr4 (0.073 g, 0.082 mmol), and THF (ca. 3 mL).
F
After the mixture was stirred at RT for 10 min, the volatiles were re-
moved in vacuo. The residue was washed with benzene and ex-
tracted with diethyl ether. Layering the concentrated diethyl ether
solution with n-pentane and chilling at À358C gave 2-glactoneNa
(0.103 g, 97%) as orange crystals. 1H NMR (400 MHz,
[D5]bromobenzene): d=0.71 (m, 2H, Ni-a-CH2), 1.04–1.94 (m, 44H,
PCy2), 2.01 (m, 2H, Ni-b-CH2), 2.13 (m, 2H, PCH2P) 7.63 (s, 4H,
BAr4 ), 8.18 ppm (s, 8H, BAr4 ); 13C{1H} NMR (100 MHz,
[D5]bromobenzene): d=7.42 (Ni-a-CH2), 25.13, 25.63, 26.31, 26.66–
26.93, 28.28, 29.26, 29.34 (PCy2-CH2), 29.09 (PCH2P), 34.28, 34.52
(PCy2-CH), 36.17 (Ni-b-CH2), 117.48, 123.26, 134.86, 162.11 ppm
1
tone. Spectral data for 2-blactone: H NMR (400 MHz, [D6]benzene):
d=1.29 (m, 3H, Ni-b-CH3), 1.03–2.15 (48H, PCy2& PCH2CH2P),
2.18 ppm (m, 1H, Ni-a-CH); 31P{1H} NMR (162 MHz, [D6]benzene):
d=À1.89 (d, 2JPÀP =2.9 Hz, 1P, PCy2), 21.92 ppm (d, 2JPÀP =2.9 Hz,
1P, PCy2); partial 13C NMR taken from 1H–13C HSQC (400 MHz,
[D6]benzene): d=18.4 (Ni-b-CH3), 26.7 ppm (Ni-a-CH).
F
F
F
Observation of [Ni(dcpe)(kC,kOÀCH(CH3)COO)Na)] [BAr4 ]
(1-blactoneNa)
A
scintillation vial (20 mL) was charged with 1-AA (0.126 g,
F
(BAr4 ), one aryl and one quaternary signal not located;
F
0.228 mmol), NaBAr4 (0.206 g, 0.233 mmol), and THF (ca. 5 mL).
After the mixture was stirred at RT for two days, the volatiles were
removed in vacuo. The residue was washed with benzene and ex-
tracted with diethyl ether to give 1-blactoneNa and 1-glactoneNa
mixture (0.320 g, 98%). Compound 1-blactoneNa may also be gen-
erated by heating solutions of isolated 1-glactoneNa at 558C for 5–
10 h. The mixture was characterized by 1 H, 31P{1H}, 1H–13C HSQC,
31P{1H} NMR (162 MHz, [D5]bromobenzene): d=À8.70 (d, JPÀP
=
2
35.5 Hz, 1P, PCy2), 19.24 ppm (d, JPÀP =35.5 Hz, 1P, PCy2); 19F NMR
2
([D5]bromobenzene): d=À62.95 ppm (s).
Synthesis of [Ni(dcpe)(h2-C,CÀCH2=CHCO2H)] (1-AA)
A
scintillation vial (20 mL) was charged with dcpe (0.129 g,
1
and H–1H COSY NMR spectroscopy without isolation. Spectral data
0.305 mmol), Ni(COD)2 (0.084 g, 0.305 mmol), and THF (ca. 4 mL).
After the solid dissolved, acrylic acid (25 mL, 0.364 mmol) was in-
jected into the reaction mixture. The mixture was stirred at RT for
one hour, resulting in precipitation of a yellow solid. The solid was
filtered and washed with THF (ca. 3 mL) to give [Ni(dcpe)(h2-
for 1-blactoneNa: 1H NMR (400 MHz, [D5]bromobenzene): d=0.82
(m, 3H, Ni-b-CH3), 1.10–2.23 (48H, PCy2& PCH2CH2P), 2.23 (m, 1H,
Ni-a-CH), 7.63 (s, 4H, BAr4 ), 8.18 ppm (s, 8H, BAr4 ); 31P{1H} NMR
(162 MHz, [D5]bromobenzene): d=67.38 (s, 1P, PCy2), 74.03 ppm (s,
1P, PCy2); 19F NMR (376 MHz, C6D5Br): d=À62.95 ppm (s); partial
13C NMR taken from 1H–13C HSQC (400 MHz, [D5]bromobenzene):
d=15.9 (Ni-b-CH3), 30.0 ppm (Ni-a-CH).
F
F
CH2CHCOOH)] (0.130 g, 82%) as
(400 MHz, [D5]bromobenzene): d=1.02–2.02 (48H, PCy2
a
yellow powder. 1H NMR
&
PCH2CH2P), 2.08 (br m, 1H, h2-CH2CH), 2.40 (br m, 1H, h2-CH2CH),
3.31 (br m, 1H, h2-CH2CH), 12.33 ppm (br s, 1H, COOH);
13C{1H} NMR (100 MHz, [D5]bromobenzene): d=20.11–27.21 (PCy2-
CH2& PCH2CH2P) 25.81 (h2-CH2CH), 29.61, 32.11 (PCy2-CH),
F
Observation of [Ni(dcpm)(kC,kOÀCH(CH3)COO)Na)][BAr4 ]
(2-blactoneNa)
37.51 ppm (h2-CH2CH), one quaternary COOH signal was not locat-
ed; 31P{1H} NMR (162 MHz, [D5]bromobenzene): d=61.30 (d, JPÀP
A.J. Young NMR tube was charged with 2-glactone (0.020 g,
2
=
F
0.037 mmol), NaBAr4 (0.035 g, 0.040 mmol), and [D6]acetone (ca.
48.1 Hz, 1P, PCy2), 72.50 ppm (d, 2JPÀP =48.1 Hz, 1P, PCy2).
0.5 mL). The mixture was heated at 558C for one day, showing an
approximately 1:8 mixture of 2-blactoneNa and 2-glactoneNa. The
1
1
1
mixture was characterized by H, 31P{1H}, H–13C HSQC, and H–1H
Synthesis of [Ni(dcpe)(kC,kOÀCH(CH3)COO)] (1-blactone)]
COSY NMR spectroscopy without isolation. Spectral data for 2-blac-
This species may be generated by either thermolysis of 1-glactone
or 1-AA. Because the isomerization from 1-AA provides a higher ki-
netic yield of 1-blactone, it was used in isolation of small quantities
of the pure compound. A scintillation vial (20 mL) was charged
with dcpe (0.092 g, 0.218 mmol), Ni(COD)2 (0.060 g, 0.218 mmol),
and benzene (ca. 15 mL). Acrylic acid (16 mL, 0.233 mmol) was in-
jected into the reaction mixture. After the mixture was stirred at RT
for one day, the volatiles were removed in vacuo. The residue was
washed with n-pentane and extracted with toluene to give mixture
of 1-blactone and 1-glactone (0.050 g). The mixture was redis-
solved in THF and layered with pentane to give pure 1-blactone
1
3
toneNa: H NMR (400 MHz, [D6]acetone): d=0.86 (dd, JHÀH 6.9 Hz,
JPÀH 6.9 Hz, 3H, Ni-b-CH3), 1.89 (m, 1H, Ni-a-CH), 1.28–2.34 (PCy2&
PCH2P), 7.68 (s, 4H, BAr4 ), 7.80 ppm (s, 8H, BAr4 ); 31P{1H} NMR
F
F
2
(162 MHz, [D6]acetone): d=À4.15 (d, JPÀP 26.9 Hz, 1P, PCy2),
19.85 ppm (d, JPÀP 26.9 Hz, 1P, PCy2); partial 13C NMR spectrum
2
taken from 1H–13C HSQC (400 MHz, [D6]acetone): d=16.2 (Ni-b-
CH3), 26.3 ppm (Ni-a-CH).
Acknowledgements
1
(0.015 g, 12%) as orange crystals. H NMR (400 MHz, [D6]benzene):
We gratefully acknowledge support by the National Science
Foundation under the Centre for Chemical Innovation “CO2 as
a Sustainable Feedstock for Chemical Commodities” (Grant No.
CHE-1240020).
d=0.98–1.85 (44H, PCy2), 1.47 (m, 3H, Ni-b-CH3), 2.05–2.16 (m, 4H,
PCH2CH2P), 2.30 ppm (br m, 1H, Ni-a-CH); 13C{1H} NMR (100 MHz,
[D6]benzene): d=16.62 (Ni-b-CH3), 26.17–26.57, 27.06–27.57, 29.14–
29.85, 30.33, 31.26 (PCy2-CH2), 28.11 (Ni-a-CH), 33.87, 36.60 (PCy2-
CH) 180.21 ppm (CO2); 31P{1H} NMR (162 MHz, [D6]benzene): d=
64.36 (d, JPÀP =8.7 Hz, 1P, PCy2), 71.09 (d, JPÀP =8.7 Hz, 1P, PCy2);
elemental analysis calcd for C29H52NiO2P2 (553.36): C 62.94, H 9.47;
found: C 63.21, H 9.21.
2
2
Keywords: density functional calculations · fixation of carbon
dioxide · metallacycles · nickel
Chem. Eur. J. 2014, 20, 3205 – 3211
www.chemeurj.org
3210
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[1] a) M. Cokoja, C. Bruckmeier, B. Rieger, W. A. Herrmann, F. E. Kuhn,
Angew. Chem. 2011, 123, 8662; Angew. Chem. Int. Ed. 2011, 50, 8510;
b) Renewable Raw Materials: New Feedstocks for the Chemical Industry
(Eds.: R. Ulber, D. Sell, T. Hirth), Wiley-VCH, Weinheim, 2011.
[2] Developments and Innovations in Carbon Dioxide Capture and Storage
Technology, Vol 2: Carbon Dioxide Storage and Utilisation (Ed.: M. M.Mar-
oto-Valer), Woodhead, Cambridge, 2010.
[11] See the Experimental Section for further details.
[12] a) S. Baba, T. Ogura, S. Kawaguchi, H. Tokunan, Y. Kai, N. Kasai, Chem.
Commun. 1972, 910; b) L. Pandolfo, G. Paiaro, G. Valle, P. Ganis, Gazz.
Chim. Ital. 1985, 115, 65; c) W. A. Herrmann, U. Kusthardt, A. Schafer, E.
Herdtweck, Angew. Chem. 1986, 98, 818; Angew. Chem. Int. Ed. Engl.
1986, 25, 817; d) M. Gꢂmez, M. Paneque, M. L. Poveda, E. Alvarez, J. Am.
Chem. Soc. 2007, 129, 6092.
[3] M. Aresta, A. Dibenedetto, in Industrial Utilization of Carbon Dioxide; De-
velopments and Innovation in Carbon Dioxide Capture and Storage Tech-
nology (Ed.: M. M. Maroto-Valer), Woodhead, Cambridge, 2010,
pp. 377–410.
[4] a) R. Alvarez, E. Carmona, D. J. Cole-Hamilton, A. Galindo, E. Gutierrez-
Puebla, A. Monge, M. L. Poveda, C. Ruiz, J. Am. Chem. Soc. 1985, 107,
5529; b) A. Galindo, A. Pastor, P. Perez, E. Carmona, Organometallics
1993, 12, 4443; c) C. Collazo, M. del Mar Conejo, A. Pastor, A. Galindo,
Inorg. Chim. Acta 1998, 272, 125; d) R. Fischer, J. Langer, G. Malassa, D.
Walther, H. Gorls, G. Vaughan, Chem. Commun. 2006, 2510; e) W. H.
Bernskoetter, B. T. Tyler, Organometallics 2011, 30, 520.
[5] H. Hoberg, D. Schaefer, J. Organomet. Chem. 1983, 251, C51.
[6] For related computational assessments, see: a) I. Papai, G. Schubert, I.
Mayer, G. Besenyei, M. Aresta, Organometallics 2004, 23, 5252; b) D. C.
Graham, C. Mitchell, M. I. Bruce, G. F. Metha, J. H. Bowie, M. A. Buntine,
Organometallics 2007, 26, 6784.
[7] a) C. Bruckmeier, M. W. Lehenmeier, R. Reichardt, S. Vagin, B. Rieger, Or-
ganometallics 2010, 29, 2199; b) S. Y. T. Lee, M. Cokoja, M. Drees, Y. Li, J.
Mink, W. A. Herrmann, F. E. Kuhn, ChemSusChem 2011, 4, 1275; c) P. N.
Plessow, L. Weigel, R. Lindner, A. Schafer, F. Rominger, M. Limbach, P.
Hofmann, Organometallics 2013, 32, 3327.
[13] P. Mastrorilli, G. Moro, C. C. F. Nobile, M. Latronico, Inorg. Chim. Acta
1992, 192, 189.
[14] See the Supporting Information for further details.
[15] N. A. Yakelis, R. G. Bergman, Organometallics 2005, 24, 3579.
[16] A. B. Pangborn, M. A. Giardello, R. H. Grubbs, R. K. Rosen, F. J. Timmers,
Organometallics 1996, 15, 1518.
[17] J. Sandstrçm, Dynamic NMR Spectroscopy, Academic Press, New York,
1982.
[18] Gaussian 09, Revision A.02, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E.
Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Men-
nucci, 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. Ko-
bayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyen-
gar, J. Tomasi, M. Cossi, N. Rega, J. M. 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, ꢃ. Farkas, J. B. Foresman, J. V. Ortiz, J. Cio-
slowski, and D. J. Fox, Gaussian, Inc., Wallingford CT, 2009.
[8] 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, M.
Limbach, Chem. Eur. J. 2012, 18, 14017.
[9] D. Jin, T. J. Schmeier, P. G. Williard, N. Hazari, W. H. Bernskoetter, Organo-
metallics 2013, 32, 2152.
[10] a) H. Hoberg, Y. Peres, C. Kruger, Y. H. Tsay, Angew. Chem. Int. Ed. Engl.
1987, 26, 771; b) H. Hoberg, A. Ballesteros, A. Sigan, C. Jegat, D. Barhau-
sen, A. Milchereit, J. Organomet. Chem. 1991, 407, C23.
Received: October 28, 2014
Published online on February 12, 2014
Chem. Eur. J. 2014, 20, 3205 – 3211
www.chemeurj.org
3211
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim