Cobalt-bis(imino)pyridine complexes as catalysts for hydroalumination-isomerisation of internal olefins

Article abstract of DOI:10.1039/c6dt01113f

The insertion of α- and internal octenes (hydroalumination) and chain walking isomerisation at di-n-octylaluminium hydride [Al(Oct)₂H], catalysed by bis(imino)pyridine-Co complexes has been investigated by NMR spectroscopy. The Co-based catalysts promote efficient hydroalumination of 1-octene. Internal olefins are partially hydroaluminated, with isomerisation to the primary alkyls, but the catalyst responsible appears to deactivate rapidly. The reaction between the Co precatalysts and [Al(Oct)₂H] generates a Co-hydride species, likely to be a hydride bridged dinuclear Co and Al complex. This species is reactive towards α-olefins but inert towards internal olefins. In contrast to hydroalumination, the catalysts promote efficient hydroboration, where insertion and isomerisation of internal octenes goes to completion. The differences between the systems may be partially ascribed to formation of an active mononuclear Co catalyst in the borane system versus a less active Co/Al dinuclear complex in hydroalumination.

Full text of DOI:10.1039/c6dt01113f

Dalton
Transactions
View Article Online
View Journal
PAPER
Cobalt-bis(imino)pyridine complexes as catalysts
for hydroalumination–isomerisation of internal
olefins†
Cite this: DOI: 10.1039/c6dt01113f
Nandita M. Weliange,a David S. McGuinness,*a Michael G. Gardinera and Jim Patelb
The insertion of α- and internal octenes (hydroalumination) and chain walking isomerisation at di-n-octyl-
aluminium hydride [Al(Oct)2H], catalysed by bis(imino)pyridine–Co complexes has been investigated by
NMR spectroscopy. The Co-based catalysts promote efficient hydroalumination of 1-octene. Internal
olefins are partially hydroaluminated, with isomerisation to the primary alkyls, but the catalyst responsible
appears to deactivate rapidly. The reaction between the Co precatalysts and [Al(Oct)2H] generates a Co-
hydride species, likely to be a hydride bridged dinuclear Co and Al complex. This species is reactive
towards α-olefins but inert towards internal olefins. In contrast to hydroalumination, the catalysts promote
efficient hydroboration, where insertion and isomerisation of internal octenes goes to completion. The
differences between the systems may be partially ascribed to formation of an active mononuclear Co
catalyst in the borane system versus a less active Co/Al dinuclear complex in hydroalumination.
Received 22nd March 2016,
Accepted 9th June 2016
DOI: 10.1039/c6dt01113f
www.rsc.org/dalton
Introduction
There is currently much interest in the development of new
routes to linear α-olefins (LAOs),1–5 in particular the selective
formation of specific chain lengths (1-hexene, 1-octene), which
are employed as comonomers for polyethylene production.6–10
While most approaches rely on ethylene oligomerisation, a
potential alternative we have recently started investigating is
the contrathermodynamic isomerisation of internal olefins to
α-olefins.11 Mixtures of internal olefin isomers may be pre-
pared by unselective dehydrogenation of abundant and in-
expensive n-alkanes.12 Such an approach has potential
advantages over direct functionalisation of alkanes (alkane
activation), as the α-olefins so prepared could be used directly
in a great number of existing industrial processes.
Scheme 1 Proposed metal-hydride based cycle for contrathermo-
dynamic isomerisation of olefins.
Our investigations in this area have focussed on the individ-
ual steps of the cycle shown in Scheme 1 as a route to convert
internal olefins to α-olefins.13–15 The cycle relies on the ten-
dency of some transition metal and group 13 metal secondary
alkyls to rearrange to the thermodynamically preferred
primary alkyl by way of chain walking. In the final step,
β-hydrogen elimination together with rapid removal of the
olefin furnish the α-olefin selectively. For boron-hydrides the
first two steps (insertion, isomerisation) are reasonably
facile,16,17 but we found the elimination step18 to be unsuc-
cessful.13 For alkylaluminium the situation is reversed and it
was demonstrated that α-olefins may be removed from tri-n-
alkylaluminium with very high selectivity (>97%). However, in
this case the insertion and isomerisation of internal olefins
was unsuccessful.14 As a consequence, the use of transition
metal catalysts to promote insertion and isomerisation has
been investigated. Metal catalysed hydroalumination of
alkenes and alkynes (Scheme 2) has been widely studied19–21
and in the simplest cases can be considered as an insertion
into a transition metal hydride followed by transalkylation with
aluminium. While the majority of catalysed hydroaluminations
aSchool of Physical Sciences – Chemistry, Private Bag 75, Hobart 7001, Australia.
E-mail: david.mcguinness@utas.edu.au
bCSIRO Energy, 71 Normanby Rd, Clayton North 3169, Australia
†Electronic supplementary information (ESI) available: Additional experimental
and theoretical details, supporting NMR spectra, and optimised geometries and
energies of all stationary points. See DOI: 10.1039/c6dt01113f
This journal is © The Royal Society of Chemistry 2016
Dalton Trans.

View Article Online
Paper
Dalton Transactions
merisation of 1-octene to 2-octene was observed. For this
reason all further experiments were conducted in toluene or
benzene. Complex 4, with increased steric bulk on the nitrogen
donors, was somewhat less active, although still an effective
catalyst. Under identical conditions, complete (>97%) conver-
sion was achieved after 4 h (10 mol% catalyst). For this reason,
complexes 5 and 6 were employed in further experiments.
Scheme 2 Metal-catalysed hydroalumination.
involve terminal olefins, a number of these catalysts also
promote olefin isomerisation.22–24 Following on from an
investigation into a zirconocene based catalyst,15 we were
drawn to the work of Chirik and others who have
demonstrated that Fe and Co complexes of bis(imino)-
pyridine ligands are effective catalysts for hydroboration and
hydrosilylation.25–30 This work followed on from earlier
studies by the same group,31 and Budzelaar,32 which showed
that these systems are effective alkene hydrogenation cata-
lysts. In particular, of relevance to our work is a recent report
that the Co system promotes hydroboration of internal
olefins with simultaneous isomerisation to the primary
alkyls.26 Herein we present our investigation of bis(imino)-
pyridine–Co complexes as catalysts for hydroalumination of
both α- and internal olefins.
1
The hydroalumination reaction may be monitored by H NMR
spectroscopy (Fig. 1), which is the principal method of
analysis used throughout. Along with the disappearance of
the 1-octene olefinic signals, changes in the speciation of
[Al(Oct)2H] are also observed as it is consumed. The Al–H signal
for trimer 3, which is the major form of [Al(Oct)2H] at the start
of the reaction, decreases and a new Al–H signal at 4.0 ppm
corresponding to dinulear 2 appears. The changing speciation
of aluminium is consistent with the conversion of [Al(Oct)2H]
to [Al(Oct)3], and has been observed previously in the uncata-
lysed hydroalumination of 1-octene.14 After 15 minutes the
signals due to 1-octene and Al–H are practically absent. At the
start of the reaction, signals attributed to a Co–H species gen-
erated from the precatalyst (−9.70 ppm for 4, −10.20 ppm for
5 and 6) are also observed which disappear as hydroalumination
Results and discussion
The catalysed hydroalumination reaction has been studied
with di-n-octylaluminium hydride [Al(Oct)2H] and the inser-
tion of n-octenes, both 1-octene and internal double bond
isomers. [Al(Oct)2H] was prepared by thermal elimination of
1-octene from [Al(Oct)3] under vacuum. The reaction may be
taken to ca. 80% conversion without decomposition.14 As
such, the aluminium hydride reagent employed in our studies
has
a nominal composition of 70–80% [Al(Oct)2H] and
20–30% [Al(Oct)3]. At this composition in non-coordinating
solvents the actual speciation of aluminium corresponds to
complexes 1–3 (R = nOct) as the major species, with 3 the most
abundant.14 We have tested three bis(imino)pyridine cobalt
precatalysts in this work, 4–6. It was envisaged that the
dichloro complexes 4 and 5 could be alkylated in situ by the
action of alkylaluminium. At the same time, the effect of a pre-
formed alkyl complex was compared by testing complex 6.
When [Al(Oct)2H] is treated with 1-octene (in the absence of
a catalyst), insertion to yield [Al(Oct)3] proceeds very slowly at
room temperature, and heating is required to progress the
reaction at a useful rate. We previously demonstrated that at
75 °C 1-octene inserts with a half-life of around 30 minutes.14
In contrast, in the presence of catalytic 5 or 6 (5–10 mol%),
insertion proceeds rapidly at room temperature and is essen-
tially complete (>97%) within 15 minutes. Catalyst loadings as
low as 1% may be employed, but doing so slows the reaction,
which now reaches completion overnight. Coordinating sol-
vents such as thf and acetonitrile caused a complete loss of
Fig. 1 1H NMR spectrum of the Al–H and olefinic region during reac-
hydroalumination activity. In the case of acetonitrile, an iso- tion of [Al(Oct)2H] with 1-octene catalysed by 10 mol% 5.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2016

View Article Online
Dalton Transactions
Paper
proceeds. Identical signals resulting from 5 and 6 indicate that Table 1, which shows that the level of conversion achieved is
the different precatalysts (with different oxidation states) lead dependent upon the catalyst loading. With 10 mol% cobalt
to the same species. The probable identity of the Co-hydride present around half of the octenes insert before the reaction
species, together with its role in catalysis, is discussed in more ceases. The same result was achieved when complex 6 was
detail below.
employed as catalyst. Increasing the catalyst loading beyond
While insertion of 1-octene is rapid and signals due to the 10% leads to very little improvement, for instance 19 mol% of
α-olefin disappear, the reaction does not go to 100% com- 5 led to only 56% insertion. Likewise, increasing the reaction
pletion. Instead minor signals due to internal olefins (2-, 3- temperature up to 55 °C also led to no improvement in conver-
and 4-octene) are observed at the end of the reaction (indicated sion, and in this case the appearance of multiple 1H NMR
in Fig. 1). The amount present corresponds to ca. 98% hydro- signals in the ligand aromatic region might indicate catalyst
alumination (2% free olefin). This is ascribed to an equili- decomposition (discussed below). No improvement in the
brium concentration of free olefin, and in support of this the results was noted if the reaction was carried out under neat
same composition is obtained if [Al(Oct)3] is treated with conditions.
complex 5 (10 mol%). We have previously demonstrated a
As is the case for 1-octene insertion, the hydroalumination
similar equilibrium concentration of free octene in the uncata- of internal octenes may be monitored by 1H NMR spec-
lysed reaction,14 except in that case the olefin remains as troscopy, where again the changes in the Al-hydride region are
1-octene due to the absence of an isomerisation catalyst. consistent with partial conversion of [Al(Oct)2H] to [Al(Oct)3]
Another difference is that the amount of free octenes (ca. 0.6% (ESI, Fig. S2 and S3†). Aluminium hydride signals are still
at room temperature) is lower in the uncatalysed reaction. observed once the reaction has stalled, although after an
While the presence of a catalyst should not change the posi- extended period (>24 h) these signals together with Co-hydride
tion of the equilibrium in itself, it does so in this case because slowly disappear despite incomplete olefin insertion. Interest-
internal olefins are more stable than α-olefins, thus shifting ingly, the proportion of olefins that do insert are isomerised to
the equilibrium slightly towards free olefin. The fact that the primary Al-octyls, with no evidence for sec-octyl aluminium
uninserted octenes are isomerised to internal olefins also species by 1H NMR or quenching with D2O. It therefore
demonstrated the isomerisation activity of the Co/Al system, in appears that the catalyst is successful for internal olefin to
analogy to that established by Chirik in hydroboration.26
Despite this presence of low levels of free internal octenes, remains active.
terminal alkyl isomerisation–hydroalumination while it
the hydroalumination reaction leads to the energetically pre-
The precise reason for the behaviour of the system towards
ferred primary n-octyl groups on [Al(Oct)3]. We demonstrated internal olefins is unclear. One possibility, that certain octene
previously that sec-octyl groups on Al can be detected by signal isomers are unreactive, appears not to be the case. Firstly, the
splitting in the Al-hydride region of the 1H NMR spectrum, composition of octenes which are not hydroaluminated is
and also by quenching with D2O followed by 13C NMR analysis quite similar to the starting composition (Table 2), indicating
of deuterium coupling in the 1-, 2-, 3- and 4-position of the the absence of large differences in reactivity (apart from
octyl chain.14 sec-Octyl substitution is not detected by either of
these experiments when the reaction is catalysed by Co (ESI,
Fig. S1†).
Table 1 Hydroalumination of internal octenes with [Al(Oct)2H] (1 : 1) in
The lifetime of the catalyst was evaluated by adding
the presence of catalytic 5
additional 1-octene and [Al(Oct)2H] over a day. As illustrated in
Amount cat.
(mol%)
Extent
(% insertion)
Table S2,† hydroalumination continues as additional 1-octene
and [Al(Oct)2H] are added, albeit more slowly. The reduced
rate of hydroalumination is not necessarily due to catalyst de-
activation, but rather may be a result of the lower overall cata-
lyst loading as more substrate is added (Table S2†). In a
separate experiment, it was found that even after sitting for
2 days at room temperature, addition of fresh substrate leads
to continued hydroalumination. Thus, the catalyst appears
quite stable with 1-octene as the olefin, although the situation
is different with internal octenes (below).
Entry
Octene
1
2
3
4
5
6
1
4
10
19
10
10
Mixa
15
27
48
56
41
43
Mixa
Mixa
Mixa
trans-4-
trans-2-
a Thermodynamic mixture of n-octenes, see text.
In order to explore the behaviour of the system with respect
to internal olefin insertion, a thermodynamic mixture of
n-octenes was used as has been employed in our prior studies
(3.0% 1-octene, 10.3% trans-4-, 32.9% cis/trans-3- and cis-4-,
39.9% trans-2- and 13.9% cis-2-octene). While insertion of this
mixture is initially rapid, the reaction in all cases reaches
a maximum conversion over 15–30 minutes before stalling.
Table 2 Composition of reactant octenes versus uninserted octenes (%)a
1-Octene trans-4 cis-2 trans-2 Otherb
Entry
1. Reactant composition 3.0
2. Uninserted
10.3
15.2
13.9 39.9
8.8 37.8
32.9
38.2
—
a Corresponding to experiment in entry 3 of Table 1. b cis- and trans-3
A number of representative experiments are presented in and cis-4-octene.
This journal is © The Royal Society of Chemistry 2016
Dalton Trans.

View Article Online
Paper
Dalton Transactions
1-octene). Secondly, the degree of insertion of tran-4-octene ionisation), but this was likewise unsuccessful. While we have
and trans-2-octene are very similar (Table 1), suggesting that little insight into the speciation of the majority of the Co in
the double bond position does not play a significant role in solution, it is worth mentioning the possibility of complexes
the extent of reaction. Another possibility is that the observed containing ligand centred radicals. It is well known that bis
extent of reaction represents an equilibrium position. (imino)pyridine ligands often display non-innocent redox
However, given that the inserted olefins are isomerised to the behaviour,34–37 particularly under reducing conditions (such
energetically preferred primary alkyls, it might be expected as we have here with excess alkylaluminium hydride). In such
that the reaction would be driven to near completion, as cases the ligand is reduced to the monoanionic radical. Hence
observed for 1-octene. Addition of further [Al(Oct)2H] to a paramagnetism from unpaired electrons on both the metal
stalled reaction does not lead to more octene inserting, as it centre and the ligand is possible.
should for an equilibrium. Furthermore, there is no evidence
Focussing on the diamagnetic component, integration of
for isomerisation of unreacted trans-4- or trans-2-octene the signals for bis(imino)pyridine and Co-hydride consistently
(entries 5 and 6, Table 1), which would be expected if an equi- give a ligand : hydride ratio of 1 : 2, suggesting a LCoH2
librium was established.
complex has formed (full spectra and additional discussion
Currently, the best explanation for this behaviour is a cata- are included in the ESI†). While the Co-hydride signal appears
lyst deactivation pathway which affects the active species as a single peak, the ligand resonances show that two very
responsible for internal olefin insertion and isomerisation. similar species are present, a major and a minor component.
The dependence of conversion on catalyst loading is consistent The minor species diminishes as the amount of [Al(Oct)2H] is
with this. As noted above, a slow disappearance of Al-hydride increased, approaching a single species at an Al : Co ratio of
and Co-hydride signals does occur after octene insertion stops, 10 : 1. On the basis of the NMR data presented in the ESI,†
but this process appears to be secondary to the more rapid de- and theoretical modelling (energetics, NMR predictions) we
activation of internal olefin catalysis; the metal-hydride signals postulate the complexes to be Co/Al dinuclear complexes 7a
are still present when octene insertion has ceased. This and 7b (Scheme 3). Unfortunately, attempts at isolation were
suggests the Co-hydride species observed at −10.20 ppm is not unsuccessful.
active towards internal olefins, meaning some other minor or
This formulation is supported by resonances at −0.6 ppm
transient species must be responsible for internal olefin cataly- with a characteristic14 peak splitting pattern of an Al–
sis. Interestingly, the system remains active for 1-octene hydro- CH2CH2R group, and a singlet at −0.93 ppm assigned to the
alumination once internal octenes have ceased inserting. methyl group in 7a. Complex 7a is the minor species and
Thus, adding a stoichiometric amount of 1-octene to a stalled would result from the methyl group in the reactant complex 6.
reaction leads to complete insertion of the α-olefin and con- This species disappears as excess [Al(Oct)2H] is introduced. It
sumption of the metal-hydrides.
is also absent if the reaction is carried out with the dichloro
In order to investigate the identity of the Co-hydride species complex 5 (only 7b is observed). Dinuclear complexes of the
observed during catalysis, the reaction between complex 6 same structure but with iso-butyl substitution appear to be
and [Al(Oct)2H] has been studied by NMR spectroscopy. produced when 6 is treated with [Al(iBu)2H], as shown in the
Formation of a Co-hydride signal at −10.20 ppm is observed ESI† (in this case, the presence of two similar species is also
when complex 6 is treated with even half an equivalent of revealed by a shoulder on the Co–H signal). The relevance
[Al(Oct)2H]. The position of this signal does not change as more of this dinuclear complex to catalysis is not clear. While it
[Al(Oct)2H] is introduced. It is clear however, from comparison
of integrated bis(imino)pyridine ligand resonances with those
of the aluminium alkyl, that only a fraction of the bis(imino)
pyridine–Co added is observable by 1H NMR following
addition of [Al(Oct)2H]. At an Al : Co ratio of 1 : 1, approxi-
mately one third of the ligand is observed, whereas at an
Al : Co ratio of 11 : 1, the ligand resonances are only 10% of
that expected. While the reaction solutions are dark purple in
colour, they are homogeneous and precipitation is not the
cause of signal loss. It was noted that addition of the Co com-
plexes did cause signal broadening in the NMR spectra (see for
instance Fig. S2†), suggesting the possible formation of para-
magnetic Co complex(es). This is confirmed through analysis
by the Evans NMR method,33 which reveals a downfield shift
of the solvent peak in the presence of the metal complex (refer
to ESI† for details). Attempts at isolation or crystallisation of
the complexes from solution were invariably unsuccessful,
probably partly due to the long n-octyl group substitution. The
Scheme 3 Reaction of Co–Me 6 with [Al(Oct)2H] to produce Co/Al
solution was also analysed by mass spectroscopy (electrospray dinuclear species 7a,b.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2016

View Article Online
Dalton Transactions
Paper
disappears with complete insertion of 1-octene, it is unreactive tendency of bis(imino)pyridine ligands to promote metal-
towards internal olefins. There is a real possibility that to-ligand electron transfer,35–37 and have shown that singlet
another unobserved Co species is responsible for catalysis, Co(I)–alkyls of this ligand are best described as low spin Co(II)–
which certainly appears to be the case at least with internal alkyl antiferromagnetically coupled to a ligand radical anion
olefins. This species does however illustrate a tendency to (overall singlet biradical).47 Likewise, for the reactant and
form inert Co/Al hydride-bridged clusters when aluminium product we found the closed shell singlet (RBLYP) solution to
alkyls are present, a tendency also noted for Zr catalysed be unstable, with the broken symmetry UBLYP open shell
hydroalumination.15,38–42 This may be one reason for the poor singlet stable and lower in energy. The triplet spin state was
performance of this system in hydroalumination, relative to higher in energy in each case. More detail on the relative spin
hydroboration.
state energies is provided in the ESI.† The free energy surface
For comparative purposes, and to verify our experimental for the isomerisation is shown in Fig. 2. In contrast to the
conditions and octene isomer mixture, we have carried out reactant and product, the intermediates and transition struc-
hydroboration with pinacolborane (HBPin) and catalyst 6 tures were stable as closed shell singlets, and open shell
under otherwise identical conditions. Insertion and isomeri- singlet solutions could not be found. This is analogous to the
sation of all octene isomers to yield the n-octylboronic ester situation found in bis(imino)pyridine–Co catalysed olefin
occurs readily (>95% after 60 min), as reported by Chirik hydrogenation.32 Additionally, for [LCo(sec-butyl)] a stationary
with cis- and trans-4-octene.26 The use of a relatively high point containing a β-agostic interaction was located in which
catalyst loading in this work (10 mol%) reveals a clean con- the closed shell singlet is stable. This is 13 kJ mol−1 higher in
version of the Co catalyst to a new species which can be energy than the open shell species (no β-agostic interaction),
assigned as [LCo(n-octyl)] (8, Scheme 4). In particular, the and is the local minimum associated with the transition struc-
high field resonance for the ketimine methyl group ture for elimination. The barrier to elimination from [LCo(sec-
(−1.53 ppm in 8 cf. −1.20 ppm in 6) is consistent with a bis butyl)] is quite low at 51 kJ mol−1, and is followed by a simi-
(imino)pyridine Co(I)–alkyl complex, as is a multiplet at larly low barrier for 1,2-insertion of butene to generate the
−1.21 ppm attributed to the β-hydrogens of the alkyl [LCo(n-butyl)] complex. The primary alkyl complex is the most
chain.43–46 The full 1H NMR spectrum with assignments is stable species on the free energy surface, and shows that a
shown in the ESI, Fig. S13 and S14.† The greater simplicity of monomeric Co-hydride should catalyse hydrometallation with
the spectra in the case of pinacolborane hydroboration likely concomitant isomerisation to the primary alkyl. In light of
arises because boranes are less prone to forming alkyl or this result, it is probable that a simple monomeric Co-
hydride bridged clusters with transition metals (in compari- hydride/alkyl is not present in the alkylaluminium system,
son to aluminium). This tendency may also be the reason although it may be responsible for catalysis in the borane
why catalysed olefin insertion is more successful in the case system.
of boron, as a monomeric Co-hydride should be more reactive
towards olefins.
The energetics of sec-alkyl to n-alkyl conversion at Co has
been studied theoretically with the BLYP density functional.
Given the uncertainty in the identity of the actual active
species, and the fact that the dinuclear complex observed is
unreactive towards internal olefins, we have not attempted to
model hydrometallation occurring at this Co resting state.
Instead, the reactivity of a mononuclear Co–alkyl has been
studied, beginning with [LCo(sec-butyl)] and isomerisation to
[LCo(n-butyl)]. Budzelaar and others have demonstrated the
Fig. 2 Free energy profile for isomerisation of [LCo(sec-butyl)] to
[LCo(n-butyl)] (BLYP/BS2//BLYP/BS1, kJ mol−1, 298 K, L = 2,6-(2,4,6-
Me3C6H2NvCMe)2C5H4N). The structures shown in blue are singlet
biradicals, and are unstable on the restricted BLYP surface. Structures in
Scheme 4 Hydroboration of internal octenes catalysed by complex 6.
black are closed shell singlets on the restricted BLYP surface.
This journal is © The Royal Society of Chemistry 2016
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Catalysed hydroalumination of octenes. A representative
example corresponds to the insertion of 1-octene into
Summary and conclusion
Herein we have studied the use of bis(imino)pyridine–Co com- [Al(Oct)2H] with 5.6 mol% 5 as catalyst: inside a glove box,
plexes as catalysts for insertion of α- and internal olefins complex 5 (1.5 mg, 2.8 μmol) was suspended in 0.5 mL of
(hydroalumination), and chain walking isomerisation with C6D6 in a J-Young valve NMR tube. 1-Octene (8.0 μL, 51 μmol)
[Al(Oct)2H]. The reaction between the Co precatalysts and was added followed by [Al(Oct)2H] (80% hydride, 20%
[Al(Oct)2H] generates a Co-hydride species as the only observed [Al(Oct)3], 20.0 μL, 51 μmol of hydride). The solution became
Co complex. This complex has been assigned as a hydride dark green upon adding the hydride, changing to a purple
bridged Co/Al dinuclear complex. This species, or another colour as the reaction proceeded. The reaction was monitored
derived from it, is active towards α-olefin, however is ineffec- by 1H NMR, in which the disappearance of 1-octene and
tive for internal olefin hydroalumination. Internal olefins are Al–H signals is observed (Fig. 1). The reaction was practically
partially hydroaluminated, with isomerisation to the primary complete within 15 min, as discussed in the text. The
alkyls, but the catalyst responsible appears to deactivate NMR assignments for [Al(Oct)3] which is formed by hydro-
rapidly.
alumination have been reported previously.15 Assignment of
The behaviour of the catalyst with aluminium alkyls is in the signals resulting from the catalyst are as detailed below
contrast to that in hydroboration chemistry, where insertion (reaction of complex 6 with [Al(Oct)2H]). Insertion of internal
and isomerisation of internal olefins goes to completion.26 octenes may be monitored likewise.
The success of the borane system may hinge on the formation
D2O quench. Following olefin hydroalumination unreacted
of a simple mononuclear Co catalyst which is active for iso- octenes were removed under vacuum. The residue was dis-
merising hydroboration, as indicated by our NMR analysis of this solved in C6D6 and quenched with D2O. The solution was
reaction. In the case of Al-hydride, it seems the system is more passed through a filter of anhydrous Na2SO4 and analysed by
complex. While a Co/Al dinuclear complex is identified, much 13C NMR spectroscopy (Fig. S1†).
of the Co is not observed. The formation of stable hydride-
Reaction of complex 6 with [Al(Oct)2H] and formation of
bridged transition metal–aluminium clusters, and their inac- complex 7. Different ratios of [Al(Oct)2H] : 6 lead to different
tivity towards olefins, is a problem we15 and others38–42 have ratios of complex 7a and 7b (Fig. S5–S7†). The following repre-
noted previously in studies of zirconocene catalysed hydro- sentative procedure is for a 1 : 1 ratio, which corresponds to
alumination. Furthermore, the formation of alkyl-bridged M/Al approximately 30% 7a. Addition of further [Al(Oct)2H] to this
dinuclear species is well known to hinder catalysed sample favours complex 7b, as discussed in the text. Inside a
carboalumination.48–51 This tendency may represent the main glove box, complex 6 (5.6 mg, 13 μmol) was dissolved in
stumbling block in transition metal catalysed isomerisation– 0.40 mL of C6D6 in a J-Young valve NMR tube. To the resulting
hydroalumination. Further studies of alternate catalysts are in purple solution, [Al(Oct)2H] (5.0 μL, 13 μmol) was added with
progress.
mixing. The solution was analysed by 1H and 13C NMR spectro-
scopy, and the signals assigned according to the proposed
formation of 7a,b. Attempts to crystallise the complex(es) were
invariably unsuccessful. 1H NMR: δ 7.60 (1H, overlapping
triplets, J = 7.6 Hz, p-H, pya+b); 7.46 (2H, overlapping doublets,
J = 7.6 Hz, m-H, pya+b); 6.84 (4H, overlapping singlets, m-H,
Experimental
General
All manipulations were performed under an argon atmo- Ara+b); 2.27 (6H, s, o-CH3, Arb); 2.21 (6H, s, p-CH3, Ara); 2.04
sphere using standard Schlenk techniques or carried out in (12H, overlapping singlets, o-CH3, Ara+b); 1.18, 1.15 (6H, over-
a glovebox (instances where bis(imino)pyridine Co complexes lapping singlets, NvCCH3(a+b)); 1.2–1.6 (br, m, CH2, octyl);
are reactive towards nitrogen have been reported,43,45 and 0.90 (br, m, CH3, octyl) −0.54 (4H, t, J = 8.2 Hz, Al–CH2(b));
we found argon to be necessary in this work). Solvents were −0.61 (2H, t, J = 8.3 Hz, Al–CH2(a)); −0.91 (3H, s, Al–CH3(a));
purified by passage through an Innovative Technologies −10.20 (2H, s, Co–H(a+b)). 13C NMR (toluene-d8): δ 157.5
solvent purification system and, where appropriate, were (NvCCH3(b)); 157.3 NvCCH3(a)); 151.1 (o-C, pyb); 150.9 (o-C,
stored over a sodium mirror. 1-Octene was dried by distilla- pya); 147.1 (i-C, Ara+b); 135.2 (p-C, Ara+b); 130.5, 130.1 (o,m-C,
tion over sodium and stored over a sodium mirror. [Al(Oct)3] Ara+b); 122.2 (m-C, pyb); 122.1 (m-C, pya); 115.2 (p-C, pya); 115.1
and [Al(Oct)2H] (70–80% hydride, 20–30% [Al(Oct)3], quanti- (p-C, pyb); 32.4–32.7, 29.8–30.3, 23.1–23.3 (multiple signals,
fied by 1H NMR) and the mixture of internal n-octenes were CH2, octyl); 21.2 (p-CH3, Arb); 21.2 (p-CH3, Ara); 19.0 (o-CH3,
prepared as described previously.14 Complexes 4, 5 52 and Arb); 19.0 (o-CH3, Ara); 16.2 (NvCCH3(b)); 16.1 (NvCCH3(a));
6 45 were prepared as reported previously. NMR measure- 14.4 (multiple overlapping signals, CH3, octyl); 9.8 (Al–CH2(b)
ments were recorded on a Bruker Avance III HD NMR (Al–CH2 and Al–CH3 due to species 7b not observed).
)
spectrometer with
a
5
mm BBFO probe operating at
Catalysed hydroboration with HBPin and internal octenes.
400 MHz (1H) or 100 MHz (13C). GC-MS analyses were Inside a glove box, HBPin (7.5 μL, 52 μmol) and a mixture of
carried out on a Varian 3800 GC coupled to a Varian 1200 octene isomers (composition as per entry 1, Table 2, 8.0 μL,
triple quadrupole mass spectrometer in single quadrupole 51 μmol)) were dissolved in 0.40 mL of C6D6 in a J-Young valve
mode.
NMR tube. Complex 6 (2.4 mg, 5.1 μmol) was added and the
Dalton Trans.
This journal is © The Royal Society of Chemistry 2016

View Article Online
Dalton Transactions
Paper
reaction was followed by 1H NMR spectroscopy. The signals complexes should be considered approximate. Further analysis
due to formation of (n-Oct)BPin were as reported previously.26 of the relative spin state energies is included in the ESI.†
The signals attributed to complex 8 (Fig. S13 and S14†) are:
δ 10.24 (1H, t, J = 7.7 Hz, p-H, py); 8.09 (2H, d, J = 7.7 Hz); 7.09
(4H, s, m-H, Ar); 2.37 (6H, s, p-CH3, Ar); 2.04 (12H, s, o-CH3,
Ar); 0.61 (2H, m, Co–CH2); −1.21 (2H, m, Co–CH2CH2); −1.53
Acknowledgements
(6H, s, CvNCH3).
We thank the Australian Research Council for financial
support through grant DP1211558 and the Australian National
Computational Infrastructure (NCI) and Tasmanian Partner-
ship for Advanced Computing (TPAC) programs for provision
of computing resources. Dr James Horne and A/Prof. Noel
Theoretical methods
Theoretical calculations throughout this paper were carried Davies of the UTas Central Science Laboratory are thanked for
out with the Gaussian0953 program. Geometry optimisations assistance with NMR and GC-MS measurements respectively.
were performed without symmetry constraints using the Dr Curtis Ho and Robert Robinson are thanked for assistance
BLYP54,55 functional in combination with the Stuttgart– with experimental aspects.
Dresden56 (SDD) double-ζ valence basis set and effective core
potential (ECP) for Co and the 6-31G(d) basis set for all other
atoms (referred to as BS1). Dispersion effects were incorpor-
ated into the geometry optimisations by including Grimme’s
Notes and references
empirical dispersion correction with Becke–Johnson damping
(GD3BJ).57,58 Analytical frequency calculations were carried out
to verify structural optimisations and to obtain free energy cor-
rections at the same level of theory (298 K and 1 atm standard
state, free energy estimates will differ somewhat in solution).
Optimisations to minima consistently resulted in a persistent
low value imaginary frequency (νi = ca. 15i cm−1), which corres-
ponds to rocking or ‘breathing’ of the bis(imino)pyridine
ligand. Attempts to remove this by using an ultrafine or super-
fine integration grid in combination with tight convergence
criteria were unsuccessful, unexpectedly leading to an increase
in the number of imaginary frequencies in many cases. Given
the nature and magnitude of this imaginary mode, errors
are expected to be minor. All transition structures contained
an imaginary frequency that exhibited vibrational modes
1 P.-A. R. Breuil, L. Magna and H. Olivier-Bourbigou, Catal.
Lett., 2015, 145, 173.
2 A. Forestière, H. Olivier-Bourbigou and L. Saussine, Oil Gas
Sci. Technol., 2009, 64, 649.
3 H. Olivier-Bourbigou, A. Forestière, L. Saussine, L. Magna,
F. Favre and F. Hugues, Oil Gas Eur. Mag., 2010, 36, 97.
4 G. P. Belov and P. E. Matkovsky, Pet. Chem., 2010, 50,
283.
5 W. Keim, Angew. Chem., Int. Ed., 2013, 52, 12492.
6 J. T. Dixon, M. J. Green, F. M. Hess and D. H. Morgan,
J. Organomet. Chem., 2004, 689, 3641.
7 D. S. McGuinness, Chem. Rev., 2011, 111, 2321.
8 T. Agapie, Coord. Chem. Rev., 2011, 255, 861.
9 P. W. N. M. van Leeuwen, N. D. Clément and
M. J.-L. Tschan, Coord. Chem. Rev., 2011, 255, 1499.
consistent with the anticipated reaction pathway (further veri- 10 D. F. Wass, Dalton Trans., 2007, 816.
fied by optimisation of slightly relaxed transition structures 11 A. de Klerk, S. W. Hadebe, J. R. Govender, D. Jaganyi,
in each direction and IRC calculations). Single point energies
for all structures were calculated with the BLYP functional
A. B. Mzinyati, R. S. Robinson and N. Xaba, Ind. Eng. Chem.
Res., 2007, 46, 400.
with the quadruple-ζ valence def2-QZVP59,60 basis set and 12 P. R. Pujado, in Handbook of Petroleum Refining Processes,
SDD ECP on Co, and the 6-311+G(2d,p) basis set on all other
ed. R. A. Meyers, McGraw Hill, 3rd edn, 2004, p. 5.11.
atoms (referred to as BS2). The Gibbs free energy (ΔG) values 13 N. M. Weliange, D. S. McGuinness and J. Patel, Organo-
reported in this paper were obtained at the BLYP/BS2//BLYP/
BS1 level.
metallics, 2014, 33, 4251.
14 N. M. Weliange, D. S. McGuinness, M. G. Gardiner and
J. Patel, Dalton Trans., 2015, 44, 15286.
The formally Co(I) complexes investigated here were evalu-
ated in the singlet, triplet and quintet spin states. In each case 15 N. M. Weliange, D. S. McGuinness, M. G. Gardiner and
the singlet spin state was most stable, consistent with experi- J. Patel, Dalton Trans., 2015, 44, 20098.
ment. In the case of four-coordinate bis(imino)pyridine Co– 16 H. C. Brown, Organic syntheses via boranes, Wiley,
alkyl and Co-hydride complexes, the closed shell singlet solu- New York, 1975.
tion is unstable, and a broken symmetry open shell singlet 17 H. C. Brown and B. C. Subba Rao, J. Org. Chem., 1957, 22,
(biradical) solution is preferred, as found previously.47 The
1137.
presence of additional ligation, or a β-agostic interaction with 18 A. B. Mzinyati and D. Jaganyi, Ind. Eng. Chem. Res., 2009,
Co, leads to the closed shell singlet becoming stable. The ener- 48, 2770.
gies of the open shell singlets ([LCo(sec-butyl)] and [LCo(n-butyl)]) 19 U. M. Dzhemilev and A. G. Ibragimov, Russ. Chem. Rev.,
were corrected through spin projection according to the 2000, 69, 121.
method of Budzelaar,47 which leads to a lowering of energy for 20 U. M. Dzhemilev and V. A. D’yakonov, Top. Organomet.
these structures by ca. 8 kJ mol−1. The energies of these
Chem., 2013, 41, 215.
This journal is © The Royal Society of Chemistry 2016
Dalton Trans.

View Article Online
Paper
Dalton Transactions
21 L. V. Parfenova, L. M. Khalilov and U. M. Dzhemilev, Russ. 43 V. C. Gibson, M. J. Humphries, K. P. Tellmann, D. F. Wass,
Chem. Rev., 2012, 81, 524.
A. J. P. White and D. J. Williams, Chem. Commun., 2001,
2252.
44 V. C. Gibson, K. P. Tellmann, M. J. Humphries and
D. F. Wass, Chem. Commun., 2002, 2316.
45 M. J. Humphries, K. P. Tellmann, V. C. Gibson, A. J. P. White
and D. J. Williams, Organometallics, 2005, 24, 2039.
46 T. M. Kooistra, Q. Knijnenburg, J. M. M. Smits,
A. D. Horton, P. H. M. Budzelaar and A. W. Gal, Angew.
Chem., Int. Ed., 2001, 40, 4719.
47 Q. Knijnenburg, D. Hetterscheid, T. M. Kooistra and
P. H. M. Budzelaar, Eur. J. Inorg. Chem., 2004, 1204.
48 R. A. Petros and J. R. Norton, Organometallics, 2004, 23, 5105.
22 F. Asinger, B. Fell, R. Janssen, G. Zoche, E. W. Mueller and
F. W. A. G. K. Korte, US 3322806 (Shell Oil), 1967.
23 E. W. Muller, G. Zoche and F. W. A. G. K. Korte,
US 3475477 (Shell Oil), 1969.
24 E. J. M. de Boer, H. Van der Heijden, I. Oosterveld and
A. Van Zon, WO 2006/005762 (Shell), 2006.
25 J. V. Obligacion and P. J. Chirik, Org. Lett., 2013, 15, 2680.
26 J. V. Obligacion and P. J. Chirik, J. Am. Chem. Soc., 2013,
135, 19107.
27 J. V. Obligacion, J. M. Neely, A. N. Yazdani, I. Pappas and
P. J. Chirik, J. Am. Chem. Soc., 2015, 137, 5855.
28 C. C. H. Atienza, A. M. Tondreau, K. J. Weller, K. M. Lewis, 49 J. M. Camara, R. A. Petros and J. R. Norton, J. Am. Chem.
R. W. Cruse, S. A. Nye, J. L. Boyer, J. G. P. Delis and
P. J. Chirik, ACS Catal., 2012, 2, 2169.
29 A. M. Tondreau, C. C. H. Atienza, K. J. Weller, S. A. Nye,
Soc., 2011, 133, 5263.
50 M. Bochmann and S. J. Lancaster, Angew. Chem., Int. Ed.,
1994, 33, 1634.
K. M. Lewis, J. G. P. Delis and P. J. Chirik, Science, 2012, 51 M. Bochmann and S. J. Lancaster, J. Organomet. Chem.,
335, 567. 1995, 497, 55.
30 M. D. Greenhalgh and S. P. Thomas, Chem. Commun., 52 B. L. Small, M. Brookhart and A. M. A. Bennett, J. Am.
2013, 49, 11230. Chem. Soc., 1998, 120, 4049.
31 S. Monfette, Z. R. Turner, S. P. Semproni and P. J. Chirik, 53 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
J. Am. Chem. Soc., 2012, 134, 4561.
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, 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, Ö. Farkas,
J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox,
Gaussian 09, Revision C1, Gaussian, Inc., Wallingford CT,
2009.
32 Q. Knijnenburg, A. D. Horton, H. van der Heijden,
T. M. Kooistra, D. G. H. Hetterscheid, J. M. M. Smits, B. de
Bruin, P. H. M. Budzelaar and A. W. Gal, J. Mol. Catal. A:
Chem., 2005, 232, 151.
33 D. F. Evans, J. Chem. Soc., 1959, 2003.
34 B. de Bruin, E. Bill, E. Bothe, T. Weyhermüller and
K. Wieghardt, Inorg. Chem., 2000, 39, 2936.
35 P. H. M. Budzelaar, B. de Bruin, A. W. Gal, K. Wieghardt
and J. H. van Lenthe, Inorg. Chem., 2001, 40, 4649.
36 S. C. Bart, K. Chlopek, E. Bill, M. W. Bouwkamp,
E. Lobkovsky, F. Neese, K. Wieghardt and P. J. Chirik,
J. Am. Chem. Soc., 2006, 128, 13901.
37 B. Butschke, K. L. Fillman, T. Bendikov, L. J. W. Shimon,
Y. Diskin-Posner, G. Leitus, S. I. Gorelsky, M. L. Neidig and
D. Milstein, Inorg. Chem., 2015, 54, 4909.
38 L. V. Parfenova, A. V. Balaev, I. M. Gubaidullin,
L. R. Abzalilova, S. V. Pechatkina, L. M. Khalilov,
S. I. Spivak and U. M. Dzhemilev, Int. J. Chem. Kinet., 2007, 54 A. D. Becke, Phys. Rev. A, 1988, 38, 3098.
39, 333.
55 C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens.
39 L. V. Parfenova, S. V. Pechatkina, L. M. Khalilov and
U. M. Dzhemilev, Russ. Chem. Bull. Int. Ed., 2005, 54, 316.
40 L. V. Parfenova, R. F. Vil’danova, S. V. Pechatkina,
Matter, 1988, 37, 785.
56 D. Andrae, U. Häussermann, M. Dolg, H. Stoll and
H. Preuss, Theor. Chim. Acta, 1990, 77, 123.
L. M. Khalilov and U. M. Dzhemilev, J. Organomet. Chem., 57 S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem.
2007, 692, 3424. Phys., 2010, 132, 154104.
41 E. Y. Pankratyev, T. V. Tyumkina, L. V. Parfenova, 58 S. Grimme, S. Ehrlich and L. Goerigk, J. Comput. Chem.,
L. M. Khalilov, S. L. Khursan and U. M. Dzhemilev, Organo-
metallics, 2009, 28, 968.
42 E. Y. Pankratyev, T. V. Tyumkina, L. V. Parfenova,
2011, 32, 1456.
59 F. Weigend and R. Ahlrichs, Phys. Chem. Chem. Phys., 2005,
7, 3297.
S. L. Khursan, L. M. Khalilov and U. M. Dzhemilev, Organo- 60 F. Weigend, F. Furche and R. Ahlrichs, J. Chem. Phys., 2003,
metallics, 2011, 30, 6078. 119, 12753.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2016