Catalyst deactivation by β-hydride elimination: Olefin and alkyne insertion into arenido-nickel(II) bonds

Article abstract of DOI:10.1002/ejic.200901179

Square planar arenido-(triphenylphosphane)nickel(II) complexes (3) containing a N,O-chelate ligand are catalysts for the carbon monoxide/ethene copolymerisation reaction. Pathways for catalyst deactivation have been elucidated by investigating the reactions of such complexes with aliphatic unsaturated compounds like olefins and alkynes. We have shown that the double or triple bond, respectively, inserts into the nickel-carbon, bond followed, by β-hydride elimination resulting in aryl-substituted olefins and. allenes which have been identified by means of GC/MS. The remaining nickel forms a bis(N,O-chelate ligand)nickel complex, the crystal structure of which has been determined. The nickel atom is coordinated, in a square-planar manner by two ligands. Additional coordination of two neighbouring complexes via their nitrile groups completes the coordination of the nickel to a distorted, octahedron and forms a two-dimensional metal-organic framework (MOF).

Full text of DOI:10.1002/ejic.200901179

FULL PAPER
DOI: 10.1002/ejic.200901179
Catalyst Deactivation by β-Hydride Elimination: Olefin and Alkyne Insertion
into Arenido–Nickel(II) Bonds
Monika M. Lindner,[a] Udo Beckmann,[a] Eva Eichberger,[a] Guido J. Reiß,[a] and
Wolfgang Kläui*[a]
Keywords: Homogeneous catalysis / Insertion / Nickel / Alkenes / Alkynes / Hydride elimination
Square planar arenido-(triphenylphosphane)nickel(II) com-
plexes (3) containing a N,O-chelate ligand are catalysts for
the carbon monoxide/ethene copolymerisation reaction.
Pathways for catalyst deactivation have been elucidated by
investigating the reactions of such complexes with aliphatic
unsaturated compounds like olefins and alkynes. We have
shown that the double or triple bond, respectively, inserts
into the nickel–carbon bond followed by β-hydride elimi-
nation resulting in aryl-substituted olefins and allenes, which
have been identified by means of GC/MS. The remaining
nickel forms a bis(N,O-chelate ligand)nickel complex, the
crystal structure of which has been determined. The nickel
atom is coordinated in a square-planar manner by two li-
gands. Additional coordination of two neighbouring com-
plexes via their nitrile groups completes the coordination of
the nickel to a distorted octahedron and forms a two-dimen-
sional metal-organic framework (MOF).
Ethylene polymerisation can be achieved with similar
nickel catalysts, e.g. with complexes comprising N,O-chelate
ligands instead of a P,O-ligand, like the ones described by
Brookhart with an anilinotropone ligand (2, Figure 1).[3]
The SHOP catalysts and the anilinotropone complexes
of Brookhart show structural similarity to the nickel com-
plexes 3a and 3b described in this work (Figure 2) and per-
form similar reactions like insertion of olefins and success-
ive β-hydride elimination, which will be discussed below.
The complexes discussed here contain a N,O-chelate ligand
and were found to catalyse – unlike the SHOP catalysts –
the copolymerisation of ethene and carbon monoxide to
yield polyketones.[4]
Introduction
The Shell Higher Olefins Process (SHOP) is one of the
most important catalytic processes of industrial relevance
performed with nickel complexes.[1,2] The worldwide pro-
duction capacity amounts to more than 106 t/a forming lin-
ear α- and internal olefins of a chain length between C8 and
C20. SHOP includes oligomerisation of ethene as well as
isomerisation of olefins and olefin metathesis. The oligo-
merisation of ethene is performed with the nickel catalyst 1
(see Figure 1).
Figure 1. SHOP catalyst 1 and the ethene polymerisation catalyst
2 studied by Brookhart.
Figure 2. Nickel catalysts active in the copolymerisation reaction
of ethene and carbon monoxide.
[a] Institut für Anorganische Chemie und Strukturchemie,
Lehrstuhl I: Bioanorganische Chemie und Katalyse, Heinrich-
Heine-Universität Düsseldorf,
Universitätsstraße 1, 40225 Düsseldorf, Germany
E-mail: klaeui@uni-duesseldorf.de
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.200901179.
Polyketones are synthesised on an industrial scale with a
catalyst system based on palladium.[5–11] However, since the
palladium catalyst remains in the polymer and is therefore
2352
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 2352–2360

Olefin and Alkyne Insertion into Arenido–Nickel(II) Bonds
lost, nickel complexes could provide a low cost alternative.
We assume that in the reaction of the complexes 3a and
The nickel complexes shown here are the best ones to date 3b with olefins and alkynes NiL2 (5) is most likely formed
in terms of efficiency to catalyse the synthesis of poly- via a nickel-hydride intermediate resulting from a β-hydride
ketones. In order to increase the efficiency and to under- elimination (see below).
stand the reaction we varied the substituents on the N,O-
NiL2 (5) is a highly insoluble paramagnetic compound
chelate ligand.[12] To inhibit decomposition, which occurs the crystal structure of which was determined by X-ray dif-
during the catalysis, and thus to prolong the lifetime of the fraction studies (Figure 3). The complex NiL2 (5) crystal-
catalyst it is of vital interest to investigate the pathways of lises in the centrosymmetric, orthorhombic space group
decomposition.
Pbca, with the nickel being the centre of symmetry. The
The reactions of the complexes 3a and 3b with aliphatic nickel(II) atom is coordinated primarily in a square planar
olefins and alkynes in the absence of carbon monoxide and manner by two N,O-chelate ligands. The coordination
the products thereof will be discussed in this work.
sphere is completed by two nitrile functions of two neigh-
bouring complexes resulting in a distorted octahedral envi-
ronment. By this axial interaction of the nitrile nitrogen to
an adjacent nickel centre a two-dimensional “metal-organic
framework” is formed (see Figure 4). The polymeric struc-
ture consists of layers, which are composed of combs of
four nickel centres each as depicted in Figure 4. The dis-
tance between two consecutive planes of nickel atoms corre-
sponds to the half of the cell constant c, c/2 = 10.015 Å.
Each comb of the framework is incompletely filled with two
perfluorinated side chains of the ligands, resulting in a
Results and Discussion
Complexes 3a and 3b (Figure 2) react with unsaturated
compounds such as aliphatic olefins and alkynes. For our
investigation we used 1-heptene, 1-hexene, 2-hexyne, 2-hep-
tyne as well as cyclopentene and cyclohexene and elucidated
the reaction pathway leading to the formed products. In all
cases the yellow solution of the complexes 3a and 3b dis-
solved in the mentioned unsaturated compounds turned
cloudy and colourless with time. The reaction of the com-
plexes 3a and 3b with 1-heptene and 1-hexene, respectively,
is finished within days and with 2-hexyne and 2-heptyne
within hours at ambient temperature.
Characterisation of the Precipitated Reaction Products
The whitish precipitate formed was characterised by IR
spectroscopy and mass spectrometry and turned out to be
a mixture of the bis(N,O-chelate ligand) nickel complex
NiL2 (5) (Scheme 1), triphenylphosphane and tri-
phenylphosphane oxide, respectively. The identity of NiL2
(5) was confirmed independently by synthesising it in 93%
yield from nickel(II) bromide and 4,4,5,5,6,6,6-heptafluoro-
3-oxo-2-[pyrrolidin-(2Z)-ylidene]hexanenitrile,[12] the N,O-
Figure 3. Molecular structure of complex NiL2 (5), only basal coor-
dination plane shown. All but the fluorine atoms (arbitrary radius)
chelate ligand HL (4), in the presence of a base (Scheme 1). are shown at 50% probability level.
Scheme 1. Synthesis of the bis(N,O-chelate ligand)-nickel complex NiL2 (5).
Eur. J. Inorg. Chem. 2010, 2352–2360
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2353

M. M. Lindner, U. Beckmann, E. Eichberger, G. J. Reiß, W. Kläui
FULL PAPER
rhombic shape with the angles being 105.88° and 74.12°, Table 1. Selected interatomic distances [Å] and angles [°] in NiL2
(5) and 3a.[4]
respectively. The axial Ni–nitrile bond length (2.11 Å) is sig-
nificantly longer than distances for such bonds documented
in the literature. A bis(N,O-chelate ligand)-nickel complex,
with the N,O-chelate ligand being an α-iminoketone, and
NiL2 (5)
3a
Ni(1)–N(1)
Ni(1)–O(1)
2.031(4) Å
2.034(3) Å
2.108(4) Å
1.143(6) Å
1.926 Å
1.918 Å
–
two acetonitrile ligands in trans-position exhibits a Ni–NC Ni(1)–N(2)C(10)
bond length of 2.05 Å.[13] The distance of the Ni–N bond in
N(2)ϵC(10)
1.135 Å
Angle of the planes of adjacent
molecules of the complex
Bite angle N(1)–Ni(1)–O(1)
[Ni(MeCN)6]2+ ranges from to 2.07 to 2.08 Å with various
counteranions.[14,15] The coordination bond between nitro-
gen and nickel is the only position in this framework, which
is flexible enough to be adjusted. Apparently, to achieve the
required space for the two C3F7 chains in the comb of the
framework the bond between nickel and the nitrogen atom
of the nitrile group is elongated.
82.35°
91.96(13)°
–
91.13°
two-dimensional coordination polymer, making up combs
of four copper centres with a coordination of the nitrile
group to an adjacent copper centre.
Characterisation of the Reaction Products in the Solution
The products of the reaction of 3a and 3b with cyclic
(cyclopentene and cyclohexene) and non-cyclic (1-heptene,
1-hexene, 2-hexyne and 2-heptyne) unsaturated compounds
were analysed with GC/MS.
In the case of 1-heptene, 1-hexene, 2-hexyne and 2-hept-
yne we found that β-hydride elimination through insertion
of the unsaturated compound into the arenido–nickel bond
takes place at ambient temperature. This was proved by the
β-hydride elimination products found in the GC/MS spec-
tra. This is similar to the SHOP oligomerisation where eth-
ene is repeatedly inserted into the nickel–hydride bond,
which is formed in a first step. The α-olefin is then likewise
released via β-hydride elimination.[1,2] In contrast to the
SHOP process our catalysts 3a and 3b are deactivated by
β-hydride elimination after the first insertion of a higher
olefin.
Figure 4. Perspective view of four units of NiL2 (5). Hydrogen and
fluorine atoms are omitted for clarity (arbitrary radius of all
atoms).
The formation of this two-dimensional coordination
The insertion of ethene into 3a has been investigated as
polymer explains why NiL2 (5) is only soluble in strong do- well. Here we also find the β-hydride elimination product,
nor solvents like dmso.
2-methyl styrene, showing a mass spectrum with the molec-
Figure 4 shows the square-planar coordination planes of ular ion m/z 118 ([M]·+). However, if 3a is reacted with
four adjacent complex molecules with coplanar Ni atoms 70 bar of ethene at room temp., the polymerisation of eth-
in the polymer. The planes intersect nearly rectangularly at ene is significantly faster than the β-hydride elimination so
ca. 82.35°.
that polyethylene is produced in high yields.
The IR spectrum of NiL2 (5) supports the findings of
For 1-heptene, according to the pathway depicted in
the crystal structure. The absorption of the CϵN stretching Schemes 2 and 3 five isomers are possible depending on
vibration at 2238 cm–1 is remarkable as it is shifted to how the olefin is orientated towards the nickel centre and
higher frequencies compared to 2212 cm–1 in the corre- the arenido ligand before inserting into the nickel–carbon
sponding vibration in the free ligand. The CϵN stretching bond. 1-Heptene can bind to nickel with C-1 (Scheme 2) or
vibration of 3a also amounts to only 2216 cm–1. This con- C-2 (Scheme 3). After olefin insertion β-hydride elimination
firms a coordination via the nitrogen atom of the nitrile can occur in two different ways, denoted A and B in
group to the nickel centre of a neighbouring complex as a Scheme 3. Of the five isomers one contains a terminal
coordination at the nitrile group results in a higher stretch- double bond (6a) and two have an internal double bond
ing frequency.[16] The crystal structure shows that the CϵN (7a, 7b and 8a, 8b), the latter showing E/Z isomerism.
bond length in NiL2 (5) is significantly longer than the one
in 3a (see Table 1).
In the case of the 2-toluenido complex 3a all of the five
proposed isomers were found in the gas chromatogram and
In the literature the crystal structure of a similar bis- identified by means of EI mass spectra. These isomers ap-
ligand copper(II) complex is documented, with the N,O-che- pear in the gas chromatogram with different retention times
lating ligand being 3-oxo-3-phenyl-2-(pyrrolidine-2-ylidene)- (see Supporting Information) and they all show mass spec-
propanenitrile.[17] Like NiL2 (5), the compound forms a tra with the molecular ion m/z 188 ([M]·+). Four of the EI
2354
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 2352–2360

Olefin and Alkyne Insertion into Arenido–Nickel(II) Bonds
Scheme 2. Mechanism of the β-hydride elimination with 1-heptene, insertion of the double bond resulting in a bond to the nickel atom
via C-1.
Scheme 3. Mechanism of the β-hydride elimination with 1-heptene, insertion of the double bond resulting in a bond to the nickel atom
via C-2, all possible isomers are shown (“a” isomers originate from the reaction with complex 3a, “b” isomers from the reaction with
complex 3b).
spectra differ only in the intensity of the fragments, while ted in the literature.[18] In 1969 Gerrard and Djerassi investi-
the fifth shows a different fragmentation pattern. We as- gated the EI mass spectra of a series of isomeric phenyl-
sume that this EI spectrum belongs to 6a, while the other heptenes and found that they exhibited double bond mo-
four spectra can be related to the other isomers, E-7a, Z- bility upon electron impact and therefore all isomers
7a, E-8a and Z-8a. The reasons for this will be discussed showed nearly identical mass spectra. The examined 1-
below.
phenylheptenes underwent skeletal and hydrogen rearrange-
Although the isomers 7a and 8a vary in the position of ments; for instance, the base peak [M – C4H9]+ of these
the double bond in the alkyl chain, their EI spectra show spectra results from a rearrangement of the molecule ion
an identical fragmentation pattern, differing solely in the including a 1,3-hydrogen shift. For this reason the spectrum
intensity of the fragments. This allows no correlation of the of 7a does not show a peak for [M – C5H11]+, also evident
isomers to the spectra. This phenomenon is well documen- for the rearrangement process.
Eur. J. Inorg. Chem. 2010, 2352–2360
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2355

M. M. Lindner, U. Beckmann, E. Eichberger, G. J. Reiß, W. Kläui
FULL PAPER
Hence the four β-hydride elimination products of 3a with
1-heptene, E-7a, Z-7a, E-8a and Z-8a, are distinctly charac-
terised by very similar EI mass spectra, all with a molecular
ion m/z 188 ([M]·+) and a base peak m/z = 131, [M –
C4H9]+, resulting from an allyl or an alkyl cleavage. The
fragmentation in the EI spectrum of 7a and 8a is consistent
with the EI mass spectrum of 1-phenylhept-1-ene.[19]
The EI mass spectra of the β-hydride elimination prod-
ucts of 3b with 1-heptene, 7b and 8b, can be interpreted
analogously. Fragmentation pattern and mass differences
are identical.
The EI mass spectrum of the isomer 6a (see Scheme 2) Figure 5. Ethene polymerisation catalyst 9 studied by Grubbs.[22]
shows a different fragmentation pattern, also with a molec-
ular ion m/z 188 ([M]·+), but a base peak of m/z = 132,
on the insertion of 1-heptene into the nickel–carbon bond
which is evoked by a McLafferty rearrangement and corre-
nor on the β-hydride elimination. This is in favour of an
sponds to [M – C4H10]·+. The fragmentation conforms to
associative mechanism in which the complex and the coor-
the EI spectrum of 2-phenylhept-1-ene.[20]
dinating olefin react to an intermediate five-coordinate spe-
cies before the olefin inserts into the arenido-nickel bond.
The GC/MS analysis of the reaction of 3a with 1-heptene
shows that all five possible isomers 6a, E-7a, Z-7a, E-8a
One of the referees of this article has pointed out that we cannot
exclude a dissociative mechanism with the N,O-chelate ligand act-
ing as a hemilabile ligand.[23] During this pathway one donor func-
tion of the anionic and rather rigid N,O-chelate ligand would re-
versibly unblock a coordination site to which the incoming olefin
can coordinate. In case of a dissociation of the O-donor function
the olefin would be coordinated trans to the arenido ligand and the
complex would still have to isomerise to allow insertion into the
arenido–nickel bond.
and Z-8a are formed, as discussed above. The reaction of
the mesitylenido complex 3b with 1-heptene yields only two
of the five possible isomers.
It is probably the steric hindrance of the arenido ligand
that inhibits the reaction pathway shown in Scheme 2 and
leads to the formation of only one pair of E/Z isomers
through the pathway of Scheme 3. The influence of the sec-
ond ortho-methyl group would also account for the slower
decomposition of the mesitylenido complex 3b in compari-
son to the 2-toluenido complex 3a. Which of the two E/Z
isomer pairs is formed (7b or 8b) cannot be stated as there
is no apparent reason which rotamer could be favoured.
To unambiguously approve the aforesaid findings, we
conducted the same experiment using 1-hexene instead of
1-heptene. We found all five corresponding isomers from
the reaction with 3a and two corresponding isomers from
the reaction with 3b, as expected.
Heinicke has reported on related catalytically active
nickel(II) complexes with anionic P,O-chelate ligands in lieu
of our N,O-ligand. With trimethylphosphane a catalytically
five-coordinate nickel complex of type [NiMe(PMe3)2-
(P,O)][24] could even be prepared and structurally character-
ised.
Reaction of Complexes 3a and 3b with Alkynes Leads to
the Formation of Allenes
Complexes 3a and 3b react likewise with 2-hexyne. The
insertion of alkynes into nickel(II)-arenido bonds is de-
scribed in the literature, although only with sterically de-
Triphenylphosphane in Excess Does Not Inhibit β-Hydride
Elimination
Fast β-hydride elimination counteracts the polymerisa- manding alkynes, which do not show β-hydride elimi-
tion of olefins. Added triphenylphosphane might prolong nation.[25,26] The formation of allenes through β-hydride eli-
the lifetime of the catalyst by suppressing β-hydride elimi- mination has been reported on silver(111) surfaces where
nation or it might prolong the lifetime of the nickel-hydride allyl moieties are activated to afford 1,2-propadiene.[27]
species to allow the repeated insertion of olefins. Different
Analogously to the mechanistic pathway of the reaction
effects of added triphenylphosphane are documented in the with 1-heptene described above the alkyne can be inserted
literature. According to Brookhart[21] added triphenylphos- into the nickel–carbon bond in two different ways (see
phane has no influence on the polymerisation of ethene Schemes 4 and 5), forming two isomeric allenes. If the 2-
with catalyst 2, while Grubbs[22] states that additional tri- hexyne molecule binds with atom C-2 to the nickel atom,
phenylphosphane suppresses the polymerisation with nickel as depicted in Scheme 4, a terminal allene (10a, 10b) is
complex 9 (Figure 5).
formed. On the contrary, according to Scheme 5 an internal
Additional experiments with excess triphenylphosphane allene is generated, with R/S enantiomers (11a, 11b), when
added to the reaction of complexes 3a and 3b, respectively, binding occurs at atom C-3.
with 1-heptene were carried out. In both cases the corre-
Both isomers are found in the gas chromatogram of 3a
sponding β-hydride elimination products were formed, re- with 2-hexyne (10a and 11a) as well as in the gas chromato-
gardless of the amount of triphenylphosphane present. Ap- gram of 3b with 2-hexyne (10b and 11b) and were identified
parently an excess of triphenylphosphane has no influence by their EI mass spectra.
2356
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 2352–2360

Olefin and Alkyne Insertion into Arenido–Nickel(II) Bonds
Scheme 4. Mechanism of the β-hydride elimination with 2-hexyne, insertion of the triple bond resulting in a bond to the nickel atom via
C-2.
Scheme 5. Mechanism of the β-hydride elimination with 2-hexyne, insertion of the triple bond resulting in a bond to the nickel atom via
C-3.
The EI mass spectra characterise all four isomers dis-
tinctly. The EI spectrum of isomer 10a shows a molecular
ion of m/z 172 ([M]·+) with the base peak of m/z = 129
resulting from [M – C3H7]+. Consistent with this spectrum
is the EI mass spectrum of 10b, which is only shifted by the
mass of two methyl groups to a molecular ion m/z 200
([M]·+). The EI mass spectrum of 11a also has a molecular
ion of m/z 172 ([M]·+). Here the base peak m/z = 143 is
evoked by the fragment [M – C2H5]+. The spectrum of 11b
is in accordance with this spectrum. The parent compound
Scheme 6. Rearrangement of 10a.
2-phenyl-2,3-hexadiene causes a similar EI spectrum.[28]
The reaction of the 2-toluenido complex 3a with 2-
hexyne leads to a further product with an EI spectrum very blocked by a methyl group. A possible mechanistic explana-
similar to the one of isomer 10a, which only differs in the tion could be a metal-catalysed cycloisomerisation of the
intensity of the fragments. This product is absent in the re- allene 10a. This could take place through activation of the
action of 3b in 2-hexyne as is evident from the comparison distal π-bond of the allene by coordination of the electro-
of the two gas chromatograms. Especially the intensity of philic nickel(II) present in the solution, followed by an endo
the molecular ion m/z 172 in this spectrum is a lot higher attack of the nucleophilic aromatic ring and formation of
indicating a more stable molecule than 10a. This can be the product through protodemetallation of the intermedi-
explained by a rearrangement of 10a to 12a as shown in ate.[32] The EI spectrum of 12a is consistent with the EI
Scheme 6.
mass spectrum of 1-propyl-3H-indene.[33]
Such rearrangements are long since documented in the
The reaction of 2-heptyne with the complexes 3a and 3b,
literature.[29–31] This rearrangement cannot occur with the respectively, lead to products corresponding to the ones
β-hydride elimination product of the mesitylenido complex with 2-hexyne. The EI spectra of the β-hydride elimination
3b, as in 10b the ortho position of the aryl substituent is products (Figure 6) thereof are in accordance with the β-
Eur. J. Inorg. Chem. 2010, 2352–2360
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2357

M. M. Lindner, U. Beckmann, E. Eichberger, G. J. Reiß, W. Kläui
FULL PAPER
Figure 6. Isomers found in the reaction of 2-heptyne with the complexes 3a and 3b, respectively.
hydride elimination products of 2-hexyne, just shifted by
No sign of β-hydride elimination was found for the reac-
one CH2 unit. Also the ratio of the intensities in the gas tion of the complexes 3a and 3b with cyclopentene and cy-
chromatograms are very alike.
clohexene. Instead of β-hydride elimination products we
Apart from the β-hydride elimination products we also only found the arenido ligand in 2-methylphenol and 2,4,6-
found 2-methylphenol and 2,4,6-trimethylphenol as conse- trimethylphenol, respectively, which suggest an oxidative
quence of a different reaction mechanism. The alcohols of cleavage of the nickel–carbon bond. In fact the decomposi-
the arenido ligand are probably formed by oxidative cleav- tion of the complexes took much longer in the absence of
age of the nickel–carbon bond. This reaction is rather slow oxygen or did not take place at all.
and when the experiments in 1-heptene are carried out un-
It is probably the steric demand or the steric rigidity that
der nitrogen, the oxidative pathway is suppressed to a large hinders the insertion of the cycloolefins and the β-hydride
extent. According to the gas chromatograms the main reac- elimination.
tion pathway is the β-hydride elimination. The β-hydride
elimination with 2-hexyne is so much faster that the oxidat- added to complex 3a, dissolved in deuteriated benzene, and
ive cleavage of the nickel–carbon bond is not observed. the reaction was monitored for several weeks with NMR
In a further experiment an excess of cyclopentene was
1
We also performed the experiment with cyclopentene and spectroscopy. Neither the H NMR nor the 31P{1H} NMR
cyclohexene, respectively, to see whether β-hydride elimi- spectrum gave a hint for insertion of cyclopentene into the
nation can be suppressed when using cyclic olefins. A pre- nickel–carbon bond. Therefore according to these NMR ex-
condition for β-hydride elimination is the roughly coplanar periments an insertion of cyclopentene does not take place.
orientation of the metal, the α and the β carbon atom as
well as the hydrogen atom that is to be eliminated. This
orientates the β-H close to the metal.[34] A look at the inter-
mediate after the insertion of 1-heptene into the nickel–car-
bon bond (see Scheme 2) shows that the hydrogen atoms in
β-position can easily be brought into a conformation suit-
able for β-hydride elimination (Figure 7, 16). If the insertion
of cyclopentene or cyclohexene takes place, a conformation
that allows β-hydride elimination is difficult to achieve (Fig-
ure 7, 17) therefore a potential insertion product would be
more stable towards β-hydride elimination.
Conclusions
Both complexes 3a and 3b are air-stable in solid state
and do not decompose in toluene solution for several days,
or under heating to reflux for several hours, respectively.
They react relatively fast with olefins and alkynes. The reac-
tion of 3a and 3b with 1-heptene and 1-hexene, respectively,
is completed within days and with 2-hexyne and 2-heptyne
within hours at ambient temperature. We have shown that
1-heptene, 1-hexene, 2-hexyne and 2-heptyne insert into the
nickel–carbon bond of the square-planar nickel(II) com-
plexes described here and subsequently form β-hydride eli-
mination products comprising several isomers. The prod-
ucts of the insertion of 1-heptene and 1-hexene lead to ole-
fins, while the products of the insertion of 2-hexyne and 2-
heptyne produce allenes. Although β-hydride elimination is
widely known as a decomposition pathway of organo tran-
sition metal complexes the formation of allenes through β-
hydride elimination has only been described to occur on
silver(111) surfaces.[27]
The resulting nickel-hydride species decomposes to form
the bis(N,O-chelate ligand)nickel complex NiL2 (5), the
crystal structure of which has been determined. Upon ad-
Figure 7. Conformation of the intermediate after the insertion of
the olefin into the nickel–carbon bond (16 1-heptene, 17 cyclo-
pentene) suitable for β-hydride elimination.
2358
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 2352–2360

Olefin and Alkyne Insertion into Arenido–Nickel(II) Bonds
Crystals were obtained via slow diffusion in a u tube. A u tube
with a diameter of 12 mm and a length of 200 mm was filled with
methanol up to one third and frozen with liquid nitrogen. On one
side a methanolic solution of nickel(II) bromide (0.2 mmol, 0.04 g)
and on the other side a methanolic solution of 4 (0.400 mmol,
0.125 g) and triethylamine (0.4 mmol, 0.04 g) were added. After ap-
prox. three weeks violet crystals had formed on the bottom of the
u tube.
dition of excess triphenylphosphane to the reaction with 1-
heptene the β-hydride elimination following the insertion
into the nickel–carbon bond is not suppressed. The stability
of the complexes in solution significantly increases from the
2-toluenido complex to the mesitylenido complex by at
least an order of magnitude when exposed to air as well as
under inert conditions.
Compared to our complexes the SHOP catalyst 1 is able
to form a stable nickel-hydride intermediate, which is neces-
sary for a repeated insertion of olefins. Complexes 3a and
3b do not repeatedly insert higher olefins as no stable
nickel-hydride complex is formed, the latter being a precon-
dition for oligomerisation. The observed β-hydride elimi-
nation apparently takes place much faster than the insertion
of the olefins. However, the nickel complexes 3a and 3b are
catalytically active in the strictly alternating copolymeris-
ation of carbon monoxide and ethene.[4] Obviously here the
insertion of CO is much faster than the β-hydride elimi-
nation. We assume that catalyst deactivation i.e. the ter-
mination of the copolymerisation reaction is the result of a
β-hydride elimination.
Crystal data for C20H12F14N4NiO2 were collected with Stoe-CCD
diffractometer at 293 K (graphite-monochromated Mo-Kα radia-
tion, λ = 0.71073 Å) M = 665.05, orthorhombic centrosymmetric
space group Pbca, a = 9.664(2) Å, b = 12.797(3) Å, c = 20.029(4) Å,
V = 2477.1(10) Å3, 28156 reflections measured, R(int) = 0.0403,
2167 unique reflections, 1809 observed reflections [IϾ2σ(I)], 187
parameters, 9 restraints, R1 [IϾ2σ(I)] = 0.0695, wR2 (all reflec-
tions) = 0.1281, shift/sumax = 0.048, completeness (2thetamax
=
50°): 0.99.5%, δρmax = 0.908 eA–3.
CCDC-752362 contains the supplementary crystallographic 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.
General Procedure for the Reaction of Complexes 3a and 3b with
Unsaturated Compounds: 70 mg of the complex 3a or 3b, respec-
tively, were dissolved in degassed olefin or alkyne (8 mL), under
nitrogen and stirred until the solution became colourless and
cloudy. The precipitate was separated from the solution by centrifu-
gation and dried in vacuo. The solution was kept under nitrogen
before recording the GC/MS spectra.
Experimental Section
General Procedures: Infrared spectra were recorded on a FT-IR
Bruker IFS 66 spectrometer. EI-MS data and FAB-MS spectra
were recorded on a Finnigan MAT 8200 in a 3-nitrobenzylalcohol
matrix. GC/MS spectra were determined on a Thermo Finnigan
Trace GC-Ultra Trace DSQ comprising a column of 15 m length
with 0.25 mm in diameter and a DB5MS phase. Injection tempera-
ture was 220 °C, column temperature increased starting at 50 up to
250 °C with 20 °C min–1. %-Area is listed as intensity of the signal
in the gas chromatogram. One-dimensional NMR spectra were re-
corded at room temperature, proton and 31P NMR spectra on a
Bruker Avance DRX 200 spectrometer. Unless otherwise stated
chemicals were purchased from commercial sources, used as re-
ceived and dried and degassed by standard methods. 4,4,5,5,6,6,6-
Heptafluoro-3-oxo-2-[(2Z)-pyrrolidin-2-ylidene]hexanenitrile[12] (4)
and the complexes 3a[4] and 3b[35] were prepared according to lit-
erature procedures.
The reaction with excess triphenylphosphane was carried out as
described above, except that 5 equiv. triphenylphosphane was
added to the reaction mixture.
NMR Experiment with Cyclopentene: 3a (29.8 mg, 0.0417 mmol)
was dissolved in degassed [D6]benzene (0.6 mL) under nitrogen in
a NMR tube and degassed and dried cyclopentene (0.1 mL,
1 mmol) was added. The reaction was monitored during four
weeks.
Supporting Information (see also the footnote on the first page of
this article): GC and MS data of all reaction products and β-
hydride elimination products.
Acknowledgments
Bis{4,4,5,5,6,6,6-heptafluoro-3-oxo-2-[(2Z)-pyrrolidin-2-ylidene]-
hexanenitrilo}nickel(II) (5): Nickel(II) bromide (0.44 g, 2.0 mmol),
4,4,5,5,6,6,6-heptafluoro-3-oxo-2-[(2Z)-pyrrolidin-2-ylidene]hexane-
nitrile (4) (1.2 g, 4.0 mmol) and sodium methoxide (0.20 g,
4.0 mmol) are dissolved in 50 mL of methanol and refluxed until
the solution is clear and light green. The solvent is removed under
reduced pressure and the remaining solid is stirred in water over-
night. After filtration 1.24 g of a slightly violet solid were obtained
M. M. L. thanks the German National Academic Foundation
(“Studienstiftung des deutschen Volkes”) for a Ph. D. scholarship.
We thank Dr. H. Keck for helpful discussions concerning GC/MS
data.
[1] W. Keim, Angew. Chem. 1990, 102, 251–260; Angew. Chem. Int.
Ed. Engl. 1990, 29, 235–244.
[2] E. F. Lutz, J. Chem. Educ. 1986, 63, 202–203.
[3] F. A. Hicks, J. C. Jenkins, M. Brookhart, Organometallics 2003,
22, 3533–3545.
[4] W. Kläui, J. Bongards, G. J. Reiß, Angew. Chem. 2000, 112,
4077–4079; Angew. Chem. Int. Ed. 2000, 39, 3894–3896.
[5] A. Sommazzi, F. Garbassi, Prog. Polym. Sci. 1997, 22, 1547–
1605.
[6] C. Bianchini, A. Meli, Coord. Chem. Rev. 2002, 225, 35–66.
[7] E. Drent, Shell Internationale Research Maatschappij B. V., EP
121965, 1984.
(93%). IR (KBr disk): ν = 2985, 2883 (νC–H aliphatic), 2238
˜
(νCN), 1622 (νCO), 1555 (νC=C), 1432 (δC–H aliphatic), 1207
(νC–F) cm–1. MS (EI): m/z (%) = 665 (14) [M]·+, 664 (68) [M –
H]+, 495 (13) [M – C3F7H]+, 362 (16) [M – C10H6N2OF7]+, 304
(17) [M – C10H6N2OF7Ni]+, 192 (24) [M – C13H7N2OF14]+, 135
(100) [M – C13H6N2OF14Ni]+. MS (FAB+): m/z (%) = 665 (Ͻ 1)
[M]·+, 460 (6), 329 (34), 307 (100), 289 (60). C20H12F14N4NiO2
(665.06): calcd. C 36.12, H 1.82, N 8.43; found C 36.1, H 1.8, N
8.3.
The magnetic moment µeff = 3.1 BM was determined in methanolic
solution by a standard Evans setup.
[8] E. Drent, J. A. M. v. Broekhoven, M. J. Doyle, J. Organomet.
Chem. 1991, 417, 235–251.
Eur. J. Inorg. Chem. 2010, 2352–2360
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2359

M. M. Lindner, U. Beckmann, E. Eichberger, G. J. Reiß, W. Kläui
FULL PAPER
[9]
[23] P. Braunstein, F. Naud, Angew. Chem. 2001, 113, 702–722; An-
gew. Chem. Int. Ed. 2001, 40, 680–699, and references cited
therein.
[24] J. Heinicke, M. He, A. Dal, H.-F. Klein, O. Hetche, W. Keim,
U. Flörke, H.-J. Haupt, Eur. J. Inorg. Chem. 2000, 431–440.
[25] P. Braunstein, T. M. Gomes Carneiro, D. Matt, F. Balegroune,
D. Grandjean, Organometallics 1989, 8, 1737–1743.
[26] W. Keim, A. Behr, B. Gruber, B. Hoffmann, F. H. Kowaldt, U.
Kürschner, B. Limbäcker, F. P. Sistig, Organometallics 1986, 5,
2356–2359.
E. Drent, M. C. T. de Kock, Shell Oil Company, US 5688909,
1997.
C. E. Ash, Int. J. Polymeric Mater. 1995, 30, 1–13.
E. Drent, P. H. M. Budzelaar, Chem. Rev. 1996, 96, 663–681.
U. Beckmann, E. Eichberger, M. Lindner, M. Bongartz, P. C.
Kunz, Eur. J. Org. Chem. 2008, 24, 4139–4147.
C. C. Lu, E. Bill, T. Weyhermüller, E. Bothe, K. Wieghardt,
Inorg. Chem. 2007, 46, 7880–7889.
I. Leban, D. Gantar, B. Frlec, D. R. Russell, J. H. Holloway,
[10]
[11]
[12]
[13]
[14]
[15]
[27] J.-H. Wang, C.-M. Chiang, J. Am. Chem. Soc. 2000, 122,
Acta Crystallogr., Sect. C 1987, 43, 1888–1890.
J. Pietikäinen, A. Maaninen, R. S. Laitinen, R. Oilunkaniemi,
J. Valkonen, Polyhedron 2002, 21, 1089–1095.
11521–11522.
[28] A. M. Caporusso, C. Polizzi, L. Lardicci, J. Org. Chem. 1987,
52, 3920–3923.
[16] K. Nakamoto, Infrared and Raman Spectra of Inorganic and
Coordination Compounds, 4 ed., Wiley-Interscience, New York,
1986.
[17] R. W. Saalfrank, H. Maid, F. Hampel, K. Peters, Eur. J. Inorg.
Chem. 1999, 1859–1867.
[18] A. F. Gerrard, C. Djerassi, J. Am. Chem. Soc. 1969, 91, 6808–
[29] D. Vorländer, C. Siebert, Ber. Dtsch. Chem. Ges. 1906, 39,
1024–1035.
[30] J. Salkind, A. Kruglow, Ber. Dtsch. Chem. Ges. 1928, 61, 2306–
2312.
[31] K. Brand, Chem. Ber. 1950, 83, 346–354.
[32] N. Krause, A. S. K. Hashmi, Modern Allene Chemistry, vol. 2,
Wiley-VCH, Weinheim, Germany, 2004, pp. 877–881.
[33] M. Adamzcyk, D. S. Watt, D. A. Netzel, J. Org. Chem. 1984,
49, 4226–4237.
[34] R. H. Crabtree, The Organometallic Chemistry of the Transition
Metals, 2 ed., Wiley-Interscience, New York, 1994.
[35] U. Beckmann, G. Hägele, W. Frank, Eur. J. Inorg. Chem. 2010,
in press. DOI: 10.1002/ejic.200901196.
6814.
[19] Y. Hu, J. Yu, S. Yang, J.-X. Wang, Y. Yin, Synlett 1998, 11,
1213–1214.
[20] V. Fiandanese, G. Marchese, F. Naso, I. Ronzini, Synthesis
1987, 11, 1034–1036.
[21] J. C. Jenkins, M. Brookhart, J. Am. Chem. Soc. 2004, 126,
5827–5842.
[22] T. R. Younkin, E. F. Connor, J. I. Henderson, S. K. Friedrich,
R. H. Grubbs, D. A. Bansleben, Science 2000, 287, 460–462.
Received: December 6, 2009
Published Online: April 26, 2010
2360
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 2352–2360