One electron oxidation of chromium N,N-bis(diarylphosphino)amine and bis(diarylphosphino)methane complexes relevant to ethene trimerisation and tetramerisation

Article abstract of DOI:10.1039/b700559h

Complexes of the type [(diphosphine)Cr(CO)₄] (diphosphine = Ph₂PN(iPr)PPh₂, Ar₂PN(Me)PAr₂ or Ar₂PCH₂PAr₂ (Ar = 2-C₆H ₄(MeO)) have been synthesise

Full text of DOI:10.1039/b700559h

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
One electron oxidation of chromium N,N-bis(diarylphosphino)amine and
bis(diarylphosphino)methane complexes relevant to ethene trimerisation and
tetramerisation
Lucy E. Bowen, Mairi F. Haddow, A. Guy Orpen and Duncan F. Wass*
Received 15th January 2007, Accepted 31st January 2007
First published as an Advance Article on the web 9th February 2007
DOI: 10.1039/b700559h
Complexes of the type [(diphosphine)Cr(CO)4] (diphosphine = Ph2PN(iPr)PPh2, Ar2PN(Me)PAr2 or
Ar2PCH2PAr2 (Ar = 2-C6H4(MeO)) have been synthesised. In the solid state, these complexes show
tight phosphine bite angles in the range 67.82(4)◦ to 71.52(5)◦ and the nitrogen atom in
N,N-bis(diarylphophino)amine ligands adopts an almost planar (sp2) geometry. All of the complexes
are readily oxidised electrochemically or chemically to corresponding Cr(I) species. There is no evidence
for coordination of the pendant ether group in derivatives with Ar = 2-MeO-C6H4 in either Cr(0) or
Cr(I) species. Treatment of the [(diphosphine)Cr(CO)4] complexes with [NO]BF4 yields
[(diphosphine)Cr(NO)(CO)3]BF4. Removal of CO ligands to generate an oligomerisation-active species
is not observed with amine oxides but triethyl aluminium is effective in this role, and active catalysts can
be produced. The use of weakly coordinating anions seems crucial in achieving oligomerisation
catalysis.
Introduction
In recent years, several chromium catalysts have emerged capable
of the selective trimerisation of ethene to 1-hexene via a distinctive
metallacyclic mechanism.1–3 In 2002, we reported catalysts based
on chromium complexes of ligands of the type Ar2PN(Me)PAr2
(Ar = ortho-methoxy-substituted aryl group) with productivity
figures over an order of magnitude better than previous systems.2
This unprecedented performance led to interest both from a
mechanistic viewpoint and in ligand structural modification,3
the most significant subsequent development being the report
Scheme 1 Postulated mechanism of activation.
the possibility of accessing these important Cr(I) intermediates
via one electron oxidation from Cr(0) complexes (see Scheme 2),
from Bollmann and co-workers that relatively minor changes
particularly using analogues of well-known compounds of the type
to ligand structure and reaction conditions can lead to ethene
[(diphosphine)Cr(CO)4]. Our objectives are to more fully explore
the organometallic and coordination chemistry of such catalytic
relevant species and ultimately develop MAO-free activation
tetramerisation rather than trimerisation.4
Despite the growing importance of this catalyst family, the pre-
cise nature of active species remains to be defined, particularly with
routes for catalysis.
regard to the oxidation state of chromium through the catalytic cy-
cle. The most compelling evidence to date comes from Bercaw and
co-workers,3a and implicates a Cr(I)/Cr(III) manifold (Scheme 1).
Treatment of the Cr(III) species, [CrPh3(Ar2PN(Me)PAr2)] with
one equivalent of H(Et2O)2BAF (BAF = B[3,5-C6H3(CF3)2]4)
yields an active catalyst together with benzene (from protonation
of one phenyl ligand) and biphenyl (from subsequent reductive
elimination of the other two). One can envisage a similar modus
operandi for the more commonly used methyl aluminoxane (MAO)
Scheme 2 One electron oxidation route to Cr(I) intermediates.
activator; in this case the cocatalyst acting as an alkylating agent to
yield a chromium trimethyl intermediate, followed by abstraction
of a methide ligand and reductive elimination of ethane.
Although the structural and reaction chemistry of these
systems with Cr(III) centres is starting to be developed, Cr(I)
complexes are by contrast unexplored. We became interested in
Results and discussion
Synthesis and structural characterisation of
[Cr(CO)4(diphosphine)] complexes 4–6
Our studies have focused on three ligand derivatives (Fig. 1): 1,
having an isopropyl N-substituent, demonstrates the best activity
School of Chemistry, University of Bristol, Cantock’s Close, Bristol, UK
BS8 1TS. E-mail: duncan.wass@bristol.ac.uk
1160 | Dalton Trans., 2007, 1160–1168
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
Fig. 1 Ligand derivatives.
and selectivity for ethene tetramerisation;4 2, having ortho-anisyl
P-substituents, has proved to be the most successful ligand for
ethene trimerisation;2 3, having a methylene chelate bridge, is
inactive for oligomerisation reaction under the same conditions
that 1 and 2 give good activity,2 and we hoped that studying the
fundamental chemistry of this ligand by way of comparison may
provide insight to this surprising result.
Several synthetic routes to complexes of the general type
[Cr(CO)4(Ar2PN(R)PAr2)] have been reported.5 We found that
simple reaction of ligand with [Cr(CO)6] under reflux in toluene
proceeded most cleanly to give moderate (30–56%) yields of the
desired products 4–6. The carbonyl stretching region of IR spectra
for these products is typical for [M(CO)4L2] complexes, with often
only three bands distinct because of overlapping peaks. Compar-
ison of the data for 4 and 5 shows that, as expected, the presence
of electron-donating OMe substituents of 2 results in this being
the slightly more basic diphosphine (2005, 1919, 1888(br) cm−1 vs
2002, 1907, 1888, 1869 cm−1). Data for complexes 5 and 6 reveal
very similar stretching frequencies (2002, 1907, 1888, 1869 cm−1 vs
2002, 1906, 1888, 1860 cm−1), despite 5 having the more electroneg-
ative nitrogen-based bridge. Clearly, the fact that this –PN(Me)P–
backbone is almost planar (vide infra) with the delocalisation of
the potential nitrogen lone pair over the chelate results in a more
electron-rich diphosphine than might be first expected.
Fig.
2
Molecular structure and numbering scheme for [Cr(CO)4-
Crystals of 4, 5 and 6 suitable for X-ray diffraction were
grown from dichloromethane at −30 ◦C. The structure of a
partially deuterated analogue of compound 5 has been previously
reported.6
(Ph2PN(iPr)PPh2)] 4 (Hydrogen atoms omitted for clarity; ORTEP plots
created at 50% probability).
Compound 4 crystallises with two molecules in the asym-
metric unit. The phenyl rings of the two crystallographically
independent molecules adopt similar conformations (see Fig. 2,
and Table 1), but the orientation of the iso-propyl group on
the nitrogen is different (torsion angles P1–N1–C5–H5 = 30.4◦
and P3–N2–C36–H36 and 75.1◦). This may be linked to the
difference in the fold angle between the CrP2 and P2N planes
in the two molecules: 10.9 and 4.9◦. The pyramidality of the
nitrogen atoms N1 and N2, (as measured by the sum of the
angles around the nitrogen) is the same within error (356.9◦),
and implies sp2 hybridisation at N. The Cr–CO bond lengths trans
to the phosphorus atoms are slightly shorter than those trans to
other carbonyls, consistent with the weaker trans-influence of the
Cr–P bond.
is small. However, the presence of the methoxy group results in
˚
a short intramolecular O1 · · · O7 distance (3.02 A). The carbonyl
groups on the metal are bending away from the PNP ligand, but
not to any greater extent than those in 4, where the methoxy groups
are not present. The Cr–P1–N1–P2 ring is almost flat, with a fold
angle of 2.6◦, and the nitrogen is again almost planar (sum of
angles around nitrogen is 356.6◦).
The dichloromethane solvate of 6 crystallises with two
molecules of complex and two molecules of dichloromethane
in the asymmetric unit (see Fig. 4, and Table 3). The crystallo-
graphically independent complex molecules adopt a similar (but
not identical) conformation, each with an approximate mirror
plane perpendicular to the plane of the Cr–P–C–P ring. The
methoxy groups adopt a conformation where two on one face
of the Cr–P–C–P ring point towards the metal, and two point
away from it. The Cr–P–C–P rings are much less planar than the
nitrogen equivalent, with fold angles of 16.9 and 15.8◦ in the two
independent molecules, but the Cr–P bonds are much the same
length as in those complexes with PNP ligands. The P–C bond is
The dichloromethane solvate of 5 crystallises with one molecule
of the complex and one molecule of dichloromethane in the
asymmetric unit (see Fig. 3, and Table 2). Neither the Cr–P bonds
nor the Cr–CO bond lengths differ significantly from those in 4,
indicating that the effect of the methoxy groups on the Cr–P bond
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 1160–1168 | 1161
©

View Article Online
◦
◦
˚
˚
Table 2 Selected bond lengths (A) and angles ( ) for 5
Table 1 Selected bond lengths (A) and angles ( ) for 4
Cr1–C1
Cr1–C2
Cr1–C3
Cr1–C4
Cr1–P1
Cr1–P2
P1–N1
P1–C8
P1–C14
P2–N1
P2–C20
P2–C26
1.878(3)
1.891(3)
1.862(3)
1.858(3)
2.3533(11)
2.3461(11)
1.714(2)
1.822(3)
1.833(3)
1.712(2)
1.828(3)
1.827(3)
Cr2–C32
Cr2–C33
Cr2–C34
Cr2–C35
Cr2–P3
Cr2–P4
P3–N2
P3–C39
P3–C45
P4–N2
1.885(3)
1.883(3)
1.858(3)
1.863(3)
2.3518(10)
2.3326(9)
1.715(2)
1.834(2)
1.827(3)
1.713(2)
1.824(2)
1.821(2)
Cr1–C2
Cr1–C3
Cr1–C4
Cr1–C5
Cr1–P1
Cr1–P2
P1–N1
P1–C6
P1–C13
P2–N1
P2–C20
P2–C27
1.892(3)
1.879(3)
1.868(3)
1.853(3)
2.3683(10)
2.3556(12)
1.710(2)
1.836(3)
1.833(3)
1.707(2)
1.844(3)
1.837(3)
P4–C51
P4–C57
P1–Cr1–P2
P1–N1–P2
N1–P1–Cr1
N1–P2–Cr1
67.82(4)
99.86(11) P3–N2–P4
95.45(8)
95.76(8)
P3–Cr1–P4
67.87(3)
99.44(11)
95.87(7)
P1–Cr1–P2
P2–N1–P1
N1–P1–Cr1
N1–P2–Cr1
68.44(3)
102.05(12)
94.46(8)
N2–P3–Cr2
N2–P4–Cr2
96.60(7)
94.98(8)
N1–P1–C8
N1–P1–C14
C8–P1–C14
N1–P2–C20
106.14(11) N2–P3–C39
106.09(11) N2–P3–C45
101.39(12) C45–P3–C39
111.73(11) N2–P4–C51
104.69(10) N2–P4–C57
102.88(11) C51–P4–C57
106.41(10)
109.51(11)
102.18(11)
105.85(11)
107.78(11)
100.83(11)
161.29(16)
−24.1(2)
88.2(2)
N1–P1–C6
N1–P1–C13
C6–P1–C13
N1–P2–C20
109.27(12)
105.71(13)
104.00(13)
103.84(12)
105.50(13)
101.34(14)
164.6(2)
N1–P2–C26
N1–P2–C27
C20–P2–C26
Cr1–P1–C8–C9
Cr1–P1–C8–C13
Cr1–P1–C14–C15
Cr1–P1–C14–C19
Cr1–P2–C20–C21
Cr1–P2–C20–C25
C20–P2–C27
Cr1–P1–C6–C7
Cr1–P1–C13–C14
Cr1–P2–C20–C21
Cr1–P2–C27–C28
−92.0(2)
Cr2–P3–C39–C40
Cr2–P3–C39–C44
Cr2–P3–C45–C46
83.8(2)
76.1(2)
−6.9(3)
−54.2(3)
176.82(17) Cr2–P3–C45–C50
−85.6(2)
−69.6(2)
−88.4(2)
Cr2–P4–C51–C52
Cr2–P4–C51–C56
18.0(2)
82.7(2)
−165.67(16)
−105.8(2)
72.9(2)
◦
˚
Cr1–P2–C26–C27 −160.19(16) Cr2–P4–C57–C62
Table 3 Selected bond lengths (A) and angles ( ) for 6
Cr1–P2–C26–C31 28.3(2) Cr2–P4–C57–C58
Cr1–C2
Cr1–C3
Cr1–C4
Cr1–C5
Cr1–P1
Cr1–P2
P1–C1
P1–C6
P1–C13
P2–C1
1.875(3)
1.882(3)
1.845(3)
1.868(3)
2.3626(16)
2.3683(17)
1.848(3)
1.824(3)
1.825(3)
1.846(2)
1.824(3)
1.823(3)
Cr2–C35
Cr2–C36
Cr2–C37
Cr2–C38
Cr2–P3
Cr2–P4
P3–C34
P3–C39
P3–C46
P4–C34
P4–C53
P4–C60
1.885(3)
1.889(3)
1.865(3)
1.846(3)
2.3662(16)
2.3690(17)
1.847(3)
1.832(3)
1.824(3)
1.848(3)
1.835(3)
1.829(3)
P2–C20
P2–C27
P1–Cr1–P2
P2–C1–P1
71.52(5)
96.92(12) P3–C34–P4
94.52(9)
94.37(9)
106.42(11) C34–P3–C39
102.89(12) C34–P3–C46
106.74(13) C39–P3–C46
104.25(11) C34–P4–C53
107.64(12) C34–P4–C60
104.86(12) C53–P4–C60
152.44(17) Cr2–P3–C39–C40
P3–Cr2–P4
71.55(5)
97.04(12)
94.59(9)
C1–P1–Cr1
C1–P2–Cr1
C1–P1–C6
C6–P1–C13
C1–P1–C13
C1–P2–C20
C1–P2–C27
C20–P2–C27
Cr1–P1–C6–C7
Cr1–P1–C13–C14
C34–P3–Cr2
C34–P4–Cr2
94.49(9)
104.88(12)
108.37(13)
103.69(12)
107.09(12)
106.49(12)
103.65(12)
158.35(16)
68.1(2)
Fig.
3 Molecular structure and numbering scheme for [Cr(CO)4-
66.9(2)
Cr2–P3–C46–C47
(Ar2PN(Me)PAr2)] Ar = 2-C6H4(MeO) 5 (Hydrogen atoms omitted for
Cr1–P2–C20–C21 −157.45(17) Cr2–P4–C53–C54 −152.20(17)
Cr1–P2–C27–C28 Cr2–P4–C60–C61 −64.2(2)
−64.1(2)
clarity; ORTEP plots created at 50% probability).
longer than the P–N bonds in the cases above, and consequently
the PCP angle is slightly smaller than the PNP angle. The bite
angle (P–Cr–P) for the PCP ligand is slightly larger (at 71.5◦) than
those in the PNP complexes (which range from 67.5◦ to 68.4◦).
that the oxidation of [Cr(CO)4(diphosphine)] (diphosphine =
Ph2PCH2CH2PPh2, dppe or Ph2PCH2PPh2, dppm) with NOPF6
or AgClO4 produces a 17 electron species [Cr(CO)4(diphosphine)]+
which is isolable but unstable. Cyclic voltammetry indicates that
[Cr(CO)4(dppm)] undergoes a reversible one electron oxidation.7
Cyclic voltammetry of 4, 5 and 6 reveals a clean reversible one-
electron oxidation in each case with E1/2 of 0.79, 0.53 and 0.42 V
vs SCE, respectively. As expected, the more electron rich complex
5 is more easily oxidised than 4. Comparison of 5 and 6 shows that
the bis(diarylphosphino)methane complex is more easily oxidised.
Electrochemistry and one electron oxidation of 4–6
In general, complexes of the type [Cr(CO)4(diphosphine)] can
be oxidised to the Cr(I) species, albeit these 17 electron com-
plexes are relatively unstable with respect to disproportionation
(Scheme 3). For example, Connelly and co-workers have reported
1162 | Dalton Trans., 2007, 1160–1168
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
Scheme 4 Synthetic route to complexes 7–9.
Although the Cr(I) complexes formed decompose slowly with time
under a nitrogen atmosphere (t1/2 ca. 24 h at room temperature)
they are somewhat more stable than the analogous dppm or dppe
complexes (t1/2 ca. 4 h at room temperature). Monitoring this
decomposition by IR spectroscopy shows the Cr(0) species 4–6 are
re-formed, suggesting the previously observed disproportionation
pathway is also operating in this case.
Synthesis and structural characterisation of
[Cr(CO)3(NO)(diphosphine)]BF4 complexes 10–11
Reaction of complexes of the general formula [M(CO)4-
(diphosphine)] (M = Cr, Mo, W) with nitrosyl salts leads to either
oxidation of the metal or formation of cationic M(O) nitrosyl
complexes depending on the reaction conditions used.7 As an alter-
native potential route to Cr(I) complexes, 4–5 were treated with one
equivalent of [NO]BF4 in MeOH/toluene according to Scheme 5.
Characterisation data indicates that in fact cationic Cr(0) com-
plexes of the general formula [Cr(CO)3(NO)(diphosphine)]+ are
formed (10–11). This is consistent with the previous observation
of the formation of analogous compounds with other diphosphine
ligands under similar reaction conditions.
Scheme 5 Synthetic route to complexes 10–11.
Fig.
4 Molecular structure and numbering scheme for [Cr(CO)4-
Crystals of 11 suitable for X-ray diffraction were grown
from dichloromethane at −30 ◦C. The tetrafluoroborate salt
11 crystallises with one molecule of complex and anion in the
asymmetric unit (see Fig. 5 and Table 4). The tetrafluoroborate
anion is disordered over two orientations. The nitrosyl and
carbonyl groups are ordered, with the nitrosyl trans to one of the
(Ar2PCH2PAr2)] Ar = 2-C6H4(MeO) 6 (Hydrogen atoms omitted for
clarity; ORTEP plots created at 50% probability).
˚
phosphorus atoms. The Cr–NO bond [1.774(3) A] is much shorter
than the Cr–CO bond lengths, and is reflected in the longer trans
˚
˚
Cr–P bond to this [2.4295(11) A cf. 2.3816(9) A trans to CO]. The
Cr–P1–N1–P2 ring is again very flat; the fold angle is 2.4◦ with
a near-planar nitrogen (sum of angles around N is 357.5◦). The
conformation of the methoxy groups differs from that in 5, with
three of the four methoxy groups pointing towards the metal. This
Scheme 3 Synthetic route to complexes 4–6.
Chemical oxidation of 4–6 with acetyl ferrocinium tetrafluo-
roborate is also facile in each case, yielding cationic Cr(I) salts
7–9 (Scheme 4). Monitoring of the reactions by IR spectroscopy
shows the expected shift to higher wavenumber for bands in
the carbonyl region (for example, 2002, 1907, 1888, 1869 cm−1
to 2082, 2026, 1965 (br) cm−1 for 5 to 8), consistent with less
backbonding to CO ligands for the higher oxidation state species.
It is noteworthy that the CO stretching bands observed for 8 and 9
are still those expected for complexes of the type [Cr(CO)4L2]; there
is no evidence for coordination of pendant methoxy substituents.
˚
again results in a short O3 · · · O7 distance of 2.99 A and a short
˚
C2 · · · O7 distance of 2.88 A.
CO ligand substitution and catalysis with complexes 4–9
There is no evidence for coordination of pendant methoxy
groups in the complex family based on 2, in contrast to the Cr(III)
chemistry of this ligand.3a,6 It seems that the carbonyl ligand of the
complexes described to date are simply not sufficiently labile to be
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 1160–1168 | 1163
©

View Article Online
◦
˚
displaced by this donor group. With this in mind, we attempted
to remove one or more CO fragments from complexes 4–9 by
treatment with 10 equivalents of amine oxides (pyridine N-oxide
and trimethylamine oxide) or irradiation. In all cases, there was
no evidence of CO displacement by IR spectroscopy. Perhaps not
surprisingly given this observation, our attempts to use these Cr(I)
complexes as catalytic precursors were unsuccessful. Treatment
of 7–9 with 1 bar ethene at room temperature for 3 h yields no
1-hexene as determined by GC. Addition of 10 equivalents of
pyridine N-oxide and trimethylamine oxide has no effect on this
result. The use of more forcing conditions, 40 bar ethene at 60 ◦C,
is also unsuccessful in the presence or absence of these amine
oxides.
More success came from the use of triethyl aluminium as
a CO ligand scavenger. Treatment of complexes 4–9 with 300
eq. of this reagent resulted in the loss of all peaks in the
coordinated CO region of the IR spectrum and formation of
a homogeneous orange–brown solution. To date we have been
unable to characterise the species formed, and CO removal by
action of the triethyl aluminium as a Lewis acid or alkylating
agent are both possibilities. We have tested this method in catalytic
ethene oligomerisation experiments (Table 5).
Potential catalysts were formed by treatment of complexes 4–
6 with one equivalent of oxidising agent, followed by 300 eq.
of triethyl aluminium. At 40 bar ethene pressure and 60 ◦C no
activity is observed with acetyl ferrocenium tetrafluoroborate, i.e.
complexes 7–9. Unoxidised complexes 4–6 themselves are also
inactive under these conditions. Reasoning that the tetrafluorob-
orate anion in these complexes may coordinate too strongly to the
chromium centre or fluorinate the highly reactive species formed,
we finally moved to a more stable and weakly coordinating anion,
[B(C6F5)4]−. Treatment of 4–6 with one equivalent of the clean
and powerful one-electron oxidising agent [NAr3][B(C6F5)4] (Ar =
4-C6H4Br) gives complexes with identical IR spectra in the CO
region to 7–9. Following addition of triethyl aluminium, low
oligomerisation activity is observed under the same conditions for
complexes 4 and 5. As previously observed with MAO activation,
complex 4 gives both 1-hexene and 1-octene, complex 5 gives high
selectivity to 1-hexene and complex 6 is inactive.
Table 4 Selected bond lengths (A) and angles ( ) for 11
Cr1–N2
Cr1–C2
Cr1–C3
Cr1–C4
Cr1–P1
Cr1–P2
P1–N1
P1–C5
P1–C12
P2–N1
P2–C19
P2–C26
1.774(3)
1.944(4)
1.928(4)
1.840(3)
2.4295(11)
2.3816(9)
1.700(2)
1.808(4)
1.815(3)
1.693(2)
1.816(3)
1.824(3)
P1–Cr1–P2
P1–N1–P2
N1–P1–Cr1
N1–P2–Cr1
N1–P1–C5
N1–P1–C12
C5–P1–C12
N1–P2–C19
67.55(3)
104.07(13)
93.22(9)
95.11(8)
108.58(15)
106.57(13)
102.46(14)
107.75(14)
106.50(13)
105.80(13)
17.4(3)
N1–P2–C26
C19–P2–C26
Cr1–P1–C5–C6
Cr1–P1–C12–C13
Cr1–P2–C19–C20
Cr1–P2–C26–C27
−70.5(3)
56.3(3)
66.5(2)
Conclusions
A range of low oxidation state chromium complexes of diphos-
phine ligands relevant to selective olefin oligomerisation catalysis
have been prepared. The almost planar (sp2) geometry of nitrogen
in N,N-bis(diarylphosphino)amine derivatives suggests delocali-
sation of the expected nitrogen lone pair over the ligand PNP
Fig. 5 Molecular structure and numbering scheme for [Cr(NO)(CO)3-
(Ar2PN(Me)2PAr2)][BF4] Ar = 2-C6H4(MeO) 11 (Hydrogen atoms and
tetrafluoroborate counterion omitted for clarity).
Table 5 Catalytic results
Compounda
Oxidising agentb
Productivity/g g−1 h−1
1-hexene (wt%)
1-octene (wt%)
4
5
6
4
5
6
[AcFc]BF4
[AcFc]BF4
[AcFc]BF4
[NAr3][B(C6F5)4]
[NAr3][B(C6F5)4]
[NAr3][B(C6F5)4]
0
0
0
210
710
0
—
—
—
80
>99
—
—
—
—
20
0
—
a Conditions: 5 mmol compound, 1 eq. oxidising agent, 300 eq. triethyl aluminium, toluene diluent, 40 bar ethene, 60 ◦C; b AcFc = acetyl ferrocinium,
Ar = 4-C6H4Br.
1164 | Dalton Trans., 2007, 1160–1168
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
structure and leads to a relatively electron rich diphosphine. We
propose this is likely to be a factor in the exceptional performance
of these ligands in catalysis. Cr(I) complexes can be accessed
via electrochemical or chemical oxidation; in derivatives with
pendant ether groups there is no evidence for coordination of these
groups, in contrast to higher oxidation state Cr(III) derivatives.
Although the (lack of) lability of CO ligands must be taken into
account, this provides indirect evidence for hemilabile behaviour in
these ligands during catalytic cycles. Two factors seem important
in generating active catalysts using this methodology. Firstly,
efficiently removing the strongly bound CO ligands-to-date, an
excess of triethyl aluminium is the only method we have found
to be effective. A role for this reagent as a poison (oxygen and
water) scavenger is also possible. Secondly, the counter-anion used
must be stable and weakly coordinating, in parallel with early
transition metal olefin polymerisation chemistry, [B(C6F5)4]− is
which was isolated by filtration and dried in vacuo to yield a
yellow solid (287 mg, 0.46 mmol, 44%). X-Ray-quality crystals
were formed from a concentrated dichloromethane solution at
0 ◦C.
1H NMR (CDCl3, 300 MHz): d = 0.62 (d, 6 H, CH3, JHH = 6.8
Hz), 3.51 (sept, 1 H, CH, JHH = 7.0 Hz), 7.19–7.90 (m, 20 H, ArH);
31P NMR (CDCl3, ref-H3PO4, 121 MHz): d = 114 (s); 13C NMR
(CDCl3, 67.9 MHz): d = 1.1 (s, CH3), 23.6 (s, CH), 128.1 (d, JCP
= 5.5 Hz, CH), 128.5 (s, JCP = 5.5 Hz, CH), 130.6 (s, CH), 132.0
(t, JCP = 15.9 Hz, CH), 132.9 (d, JCP = 21.9 Hz, CH), 127.2.(s,
CP); IR (CH2Cl2): v = 1889 (s, br) (C≡O), 1919 (s) (C≡O), 2006
(s) (C≡O) cm−1; Elemental Analysis: C31H27CrNO8P2 calcd (%) C
62.95, H 4.60, N 2.37, found C 62.60 H 4.62 N 1.90.
Synthesis of [Cr(CO)4(Ar2PN(Me)PAr2)] Ar = 2-C6H4(MeO) 5.
An analogous method to that for 4 was followed using chromium
hexacarbonyl, [Cr(CO)6], (300 mg, 1.4 mmol) and 2 (500 mg, 0.96
mmol). The product was obtained as a yellow solid (380 mg,
0.56 mmol, 56%). X-Ray-quality crystals were formed from a
concentrated dichloromethane solution at 0 ◦C.
1H NMR (CDCl3, 300 MHz): d = 2.42 (s, 3 H, CH3), 3.59 (s,
12 H, OCH3), 6.83–6.88 (m, 8 H ArH), 7.06–7.10 (m, 4 H, ArH),
7.24–7.32 (m, 4 H, ArH); 31P NMR (CDCl3, ref-H3PO4, 121 MHz):
d = 102 (s); 13C NMR (CDCl3, 75 MHz): d = 51.6 (s, CH3), 55.0 (s,
OCH3), 110.7 (s, CH), 120.0 (t, JCP = 5.5 Hz, CH), 131.6 (s, CH),
129.8 (t, CH), 133.1 (t, JCP = 11.0 Hz), 159.7 (s, CP); IR (CH2Cl2):
v = 1869 (s) (C≡O), 1888 (s) (C≡O), 1907 (s) (C≡O), 2002
(s) (C≡O) cm−1; Elemental Analysis: C33H31CrNO8P2·1/2CH2Cl2
calcd (%) C 55.45, H 4.45, N 1.93, found C 55.17, H 4.85, N 1.88.
−
successful whereas BF4 is not. Of industrial relevance, although
low productivity compared to MAO-activated systems is observed,
this is offset by the much lower cost of triethyl aluminium
compared to MAO. These results also give more weight to a
Cr(I)/Cr(III) manifold during the catalytic cycle. We are now
focussing on the synthesis of related complexes with potentially
more labile ligands.
Experimental
General techniques
All procedures were carried out under an inert (N2) atmosphere us-
ing standard Schlenk line techniques or in an inert atmosphere (Ar)
glovebox. Chemicals were obtained from Sigma Aldrich or Fisher
Scientific and used without further purification unless otherwise
stated. All solvents were purified using an Anhydrous Engineering
Grubbs-type solvent system. Ligands 1 and 2 were synthesised
according to literature methods2,4 and 3 was obtained in a method
analogous to that reported for other bis(diarylphosphino)methane
derivatives.8 [NAr3][B(C6F5)4] was synthesised according to the
method of Nataro et al.9
Synthesis of [Cr(CO)4(Ar2PCH2PAr2)] Ar = 2-C6H4(MeO) 6.
An analogous method to that for 4 was followed using chromium
hexacarbonyl, [Cr(CO)6], (170 mg, 0.60 mmol) and 3 (300 mg,
0.77 mmol). The product was obtained as a yellow solid (120 mg,
0.17 mmol, 28%). X-Ray-quality crystals ◦were formed from a
concentrated dichloromethane solution at 0 C. 1H NMR (CDCl3,
400 MHz): d = 2.33 (s, 3 H, CH3), 3.56 (s, 12 H, OCH3), 6.65–
6.78 (m, 4 H ArH), 6.85–6.95 (m, 4 H, ArH), 7.05–7.28 (m, 8
H, ArH); 31P NMR (CDCl3, ref-H3PO4, 121 MHz): d = 19.9 (s);
13C NMR (CDCl3, 100 MHz): d = 22.8 (s, CH3), 55.1 (s, OCH3),
110.6 (s, CH), 120.1 (dd. JCP = 5.4 Hz, JCP = 15.4 Hz, CH), 128.2,
129.0, 131.5 (s, CH), 133.7 (dd, JCP = 6.88 Hz, JCP = 6.88 Hz,
CH), 160.2.(dd, JCP = 10.2 Hz, JCP = 10.2 Hz, CP), IR (CH2Cl2):
v = 1861 (s) (C≡O), 1888 (s) (C≡O), 1906 (s) (C≡O), 2002 (s)
(C≡O) cm−1; Elemental Analysis: C33H30CrO8P2 calcd (%) C 59.29,
H 4.52, found C 59.81, H 4.11.
Infra-red spectra were recorded on a Perkin-Elmer 1600 series
FTIR Spectrometer in dichloromethane. Bands are characterised
as strong (s), moderate (m), weak (w) and/or broad (br). NMR
spectra were recorded on a JEOL ECP 300 spectrometer at
300 MHz (1H) and 121 MHz (31P) and a JEOL delta 400 at
1
1
200.6 MHz (13C {1 H}), in deuterated solvent. H and 13C { H}
NMR spectra are referenced chemical shifts relative to high
frequency of residual solvent and 31P NMR spectra are referenced
relative to high frequency of 85% H3PO4. Microanalyses were
carried out by the Microanalytical Laboratory of the School of
Chemistry at the University of Bristol.
Electrochemistry of 4–6
Electrochemical studies were carried out using an EG&G model
273A potentiostat linked to a computer using EG&G Model
270 Research Electrochemistry software in conjunction with a
three-electrode cell. The working electrode was a platinum disc
(1.6 mm diameter) and the auxiliary electrode a platinum wire.
The reference was an aqueous saturated calomel electrode (SCE)
separated from the test solution by a fine porosity frit and an agar
bridge saturated with KCl. Solutions were 1.0 × 10−3 mol dm−3
Synthesis of [Cr(CO)4(diphosphine)] complexes 4–6
Synthesis of [Cr(CO)4(Ph2PN(iPr)PPh2)] 4. Toluene (40 mL)
was added to chromium hexacarbonyl, [Cr(CO)6], (350 mg, 1.6
mmol) and 1 (500 mg, 1.2 mmol) and the stirred mixture was
heated under reflux for 48 h. The solution was cooled to 0 ◦C and
filtered to remove excess [Cr(CO)6]. Solvent was removed under
reduced pressure and the product extracted into dichloromethane
(10 mL). Methanol (20 mL) was added to precipitate the product,
in the test compound and 0.1 mol dm−3 in [NBun ][PF6] as the
4
supporting electrolyte, the solvent used was CH2Cl2. Under these
conditions, E◦ for one electron oxidation of [Fe(g –C5H5)2] and
5
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 1160–1168 | 1165
©

View Article Online
5
[Fe(g –C5Me5)2], added to the test solutions as internal calibrants
Removal of CO ligands
are 0.47 and −0.08 V, respectively. Unless specified all E◦ values
are at scan rate, v, of 200 mV s−1.
Attempted reaction with amine oxides. The same general pro-
cedure was repeated for all compounds 4–9 using both trimethyl-
N-oxide and pyridine-N-oxide, but is described here for complex
8. No evidence for CO removal or substitution was found in any
case.
Complex 5 (20 mg, 30 lmol) and acetyl ferrocinium tetraflu-
oroborate (12 mg, 35 lmol) were dissolved with stirring in
toluene (20 mL), yielding a purple solution of 8 as confirmed
by IR spectroscopy. To this, a solution of freshly sublimed
trimethylamine-N-oxide (38 mg, 0.5 mmol) in toluene (10 mL) was
added. After stirring for 2 h at room temperature, IR spectroscopy
indicated only the presence of 8 and trace quantities of 5. The
solution was heated under reflux for 18 h but IR spectroscopy
again indicated only approximately equal quantities of 8 and 5.
Synthesis of [Cr(CO)4(diphosphine)]BF4 complexes 7–9
Synthesis of [Cr(CO)4(Ph2PN(iPr)PPh2)]BF4 7. Complex 4
(20 mg, 0.034 mmol) and acetyl ferrocinium tetrafluoroborate
(15 mg, 0.045 mmol) were dissolved in dichloromethane (5 mL) to
give a dark purple solution. The Schlenk tube was covered with foil
to reduce exposure of the reaction mixture to light. The solution
was stirred at room temperature for 2 h and was monitored by
infra-red spectroscopy. After 24 h, infra-red spectroscopy revealed
approx. 50% of the product had converted back to 4. Attempts
to isolate 7 by addition of diethyl ether (20 mL) resulted in
precipitation of black intractable products.
IR (CH2Cl2): v = 1963 (s, br) (C≡O), 2033 (s) (C≡O), 2087 (s)
(C≡O) cm−1.
Reaction with triethyl aluminium. The same general procedure
was repeated for all compounds 4–9 but is described here for
complex 8. Complex 5 (20 mg, 30 lmol) and acetyl ferrocinium
tetrafluoroborate (12 mg, 35 lmol) were dissolved with stirring
in toluene (20 mL), yielding a purple solution of 8 as confirmed
by IR spectroscopy. To this, a solution of triethyl aluminium in
toluene (1.9 M, 300 eq.) was added. After ca. 1 min at room
temperature, IR spectroscopy indicated no peaks in the CO region
and an orange–brown solution was formed.
Synthesis of [Cr(CO)4(Ar2PN(Me)PAr2)]BF4 Ar
=
2-
C6H4(MeO) 8. An analogous method to that for 7 was
followed using 5 (20 mg, 0.029 mmol) and acetyl ferrocinium
tetrafluoroborate (12 mg, 0.036 mmol). After 24 h, infra-red
spectroscopy revealed approx. 40% of the product had converted
back to 5. Again, attempted isolation led to intractable products.
IR (CH2Cl2): v = 1965 (s, br) (C≡O), 2026 (s) (C≡O), 2083 (s)
(C≡O) cm−1.
Ethene oligomerisation catalysis
Synthesis of [Cr(CO)4(Ar2PCH2PAr2)]BF4 Ar = 2-C6H4(MeO)
9. An analogous method to that for 7 was followed using 6
(20 mg, 0.030 mmol) and acetyl ferrocinium tetrafluoroborate
(12 mg, 0.036 mmol). Again, attempted isolation led to intractable
products.
At low pressure. The same general procedure was repeated
for all compounds 4–9 using both in the absence or presence of
trimethyl-N-oxide or pyridine-N-oxide, but is described here for
complex 8.
IR (CH2Cl2): v = 1962 (s, br) (C≡O), 2033 (s) (C≡O), 2084 (s)
Complex 5 (20 mg, 30 lmol) and acetyl ferrocinium tetraflu-
oroborate (12 mg, 35 lmol) were dissolved with stirring in
toluene (20 mL), yielding a purple solution of 8 as confirmed
by IR spectroscopy. To this, a solution of freshly sublimed
trimethylamine-N-oxide (38 mg, 0.5 mmol) in toluene (10 mL)
was added. The solution was degassed under reduced pressure
and placed under an ethene atmosphere. The solution was stirred
for 2 h at 70 ◦C. After this time, the solution was allowed to cool to
room temperature and dilute hydrochloric acid (10%) added. The
organic layer was collected and washed with water, then dried over
magnesium sulfate. No oligomerisation products were detected by
GC.
(C≡O) cm−1.
Synthesis of [Cr(CO)3(NO)(diphosphine)]BF4 complexes 10–11
Synthesis of [Cr(CO)3(NO)(Ph2PN(iPr)PPh2)]BF4 10. Com-
plex 4 (72 mg, 0.15 mmol) and [NO][BF4] (30 mg, 0.26 mmol)
were dissolved in a toluene/methanol mixture (10 mL : 1 mL)
to give a yellow solution and a yellow precipitate. The solution
was stirred at room temperature for 1 h, during which time the
precipitate dissolved. The solution was stirred for a further 2 h and
then filtered and reduced in volume to about 5 mL under reduced
pressure. Diethyl ether was added (20 mL) and the solution was
stored at 0 ◦C overnight to give a yellow precipitate which was
isolated by filtration and dried in vacuo (35 mg, 0.061 mmol, 41%).
IR (CH2Cl2): v = 1760 (s) (N≡O), 2022 (s) (C≡O), 2092 (s)
(C≡O) cm−1;; Elemental Analysis: C30H27BCrF4N2O8P2 calcd (%)
C 60.71, H 4.59, N 4.72, found C 59.88, H 4.31, N 4.50.
At elevated pressure without triethyl aluminium. Reactions
were performed in parallel using a Baskerville 10 multicell auto-
clave. The same general procedure was repeated for all compounds
both in the absence or presence of trimethyl-N-oxide pyridine-N-
oxide, but is described here for complex 8.
Synthesis of [Cr(CO)3(NO)(Ar2PN(Me)PAr2)]BF4 Ar = 2-
C6H4(MeO) 11. An analogous method to that for 10 was
followed using 5 (140 mg, 0.20 mmol) and [NO][BF4] (50 mg, 0.40
mmol). A yellow microcrystalline product was isolated (73 mg,
0.095 mmol, 47%). X-Ray crystals were formed from a 1 : 1 v/v
dichloromethane/hexane solution at 0 ◦C.
IR (CH2Cl2): v = 1755 (s) (N≡O), 2024 (s) (C≡O), 2091 (s)
(C≡O) cm−1; Elemental Analysis: C33H31BCrF4N2O8P2 calcd (%)
C 49.76, H 4.05, N 3.63, found C 49.83, H 4.45, N 3.79.
Complex 5 (20 mg, 30 lmol) and acetyl ferrocinium tetraflu-
oroborate (12 mg, 35 lmol) were loaded into a reaction cell in
an inert atmosphere glovebox. The autoclave was removed from
the glovebox and toluene (4 mL) added. To this, a solution of
freshly sublimed trimethylamine-N-oxide (38 mg, 0.5 mmol) in
toluene (1.5 mL) was introduced via syringe. The reactor was
closed, heated to 60 ◦C and pressurised to 40 bar with ethene.
The reaction was stirred for 1 h. The reactor was allowed to cool
to room temperature and the pressure was slowly released. Dilute
1166 | Dalton Trans., 2007, 1160–1168
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
Table 6 Crystallographic data
Compound
4
5
6
11
Colour, habit
Size/mm
Empirical Formula
M
Pale yellow plate
0.40 × 0.40 × 0.02
C31H27CrNO4P2
591.48
Triclinic
¯
P1
10.795(2)
15.430(3)
17.984(4)
90.51(3)
106.82(3)
95.17(3)
2853.7(10)
4
Yellow block
0.48 × 0.30 × 0.07
C34H33Cl2CrNO8P2
768.45
Triclinic
¯
P1
10.850(2)
13.509(3)
13.657(3)
76.40(3)
67.08(3)
81.53(3)
1788.6(8)
2
Pale yellow block
0.38 × 0.28 × 0.16
C34H32Cl2CrO8P2
753.44
Triclinic
¯
P1
11.807(9)
17.767(11)
17.847(9)
80.50(6)
86.84(4)
74.80(6)
3563(4)
Yellow block
0.09 × 0.04 × 0.03
C32H31BCrF4N2O8P2
772.34
Monoclinic
P21/n
11.398(2)
19.022(4)
18.460(4)
90
101.33(3)
90
3924.2(14)
4
3.748
Crystal system
Space group
˚
a/A
˚
b/A
˚
c/A
a/◦
b/◦
c /◦
3
˚
V/A
Z
4
l/mm−1
0.55
0.609
0.609
T/K
100
173
173
100
Reflections: total/independent/Rint
Final R1
32614/13018/0.0429
0.0515
0.551, −0.350
18942/8213/0.0526
0.0512
1.219, −0.688
37832/16293/0.0295
0.0472
1.015, −1.151
29906/7312/0.0980
0.0480
0.445, −0.314
−3
˚
Largest peak, hole/e A
−
hydrochloric acid (10%) was added, the organic layer collected
and washed with water, then dried over magnesium sulfate. No
oligomerisation products were detected by GC.
Furthermore, the BF4 unit was also disordered, and has been
modelled as lying over two positions.
CCDC reference numbers 626712–626715.
For crystallographic data in CIF or other electronic format see
DOI: 10.1039/b700559h
At elevated pressure with triethyl aluminium. Reactions were
performed using a very similar general procedure to that described
above only the required amount of triethyl aluminium (1.9 M
solution in toluene) was added after the addition of the required
oxidising agent. Oligomerisation products were detected by GC;
quantified using mesitylene as a standard.
Acknowledgements
The Royal Society and the University of Bristol are thanked
for funding. Professor Neil G. Connelly is thanked for useful
discussions.
X-Ray crystallography
References
X-Ray diffraction experiments on the dichloromethane solvates
of 5·CH2Cl2 and 6·CH2Cl2 were carried out at 173 K on a
Bruker SMART diffractometer and an experiment on 4 was
carried out at 100 K on a Bruker SMART APEX diffractometer,
1 (a) J. T. Dixon, M. J. Green, F. M. Hess and D. H. Morgan,
J. Organomet. Chem., 2004, 689, 3641; (b) D. F. Wass, Dalton Trans.,
2007, 816; (c) J. R. Briggs, J. Chem. Soc., Chem. Commun., 1989, (11),
674; (d) W. K. Reagan, Phillips Petroleum Company, EP 0 417 477, 1991;
(e) F. J. Wu, Amoco Corporation, US Pat. 5811618, 1995; (f) R. D. Ko¨hn,
M. Hauffe, G. Kociok-Ko¨hn, S. Grimm, P. Wasserscheid and W. Keim,
Angew. Chem., Int. Ed., 2000, 39, 4337.
2 A. Carter, S. A. Cohen, N. A. Cooley, A. Murphy, J. Scutt and D. F.
Wass, Chem. Commun., 2002, (8), 858.
3 (a) T. Agapie, S. J. Schofer, J. A. Labinger and J. E. Bercaw, J. Am. Chem.
Soc., 2004, 126, 1304; (b) M. J. Overett, K. Blann, A. Bollmann, J. T.
Dixon, F. Hess, E. Killian, H. Maumela, D. H. Morgan, A. Neveling
and S. Otto, Chem. Commun., 2005, 5, 622.
4 A. Bollmann, K. Blann, J. T. Dixon, F. M. Hess, E. Killian, H.
Maumela, D. S. McGuiness, D. H. Morgan, A. Neveling, S. Otto,
M. Overett, A. M. Z. Slawin, P. Wasserscheid and S. Kuhlmann, J. Am.
Chem. Soc., 2004, 126, 14712.
5 (a) M. S. Balakrishna, T. K. Prakasha and S. S. Krishnamurthy,
J. Organomet. Chem., 1990, 390, 203; (b) J. Ellermann and W. Wend,
J. Organomet. Chem., 1983, 258, 21.
6 S. J. Schofer, M. W. Day, L. M. Henling, J. A. Labinger and J. E. Bercaw,
Organometallics, 2006, 25, 2743.
7 P. K. Ashford, P. K. Baker, N. G. Connelly, R. L. Kelly and V. A.
Woodley, J. Chem. Soc., Dalton Trans., 1982, 477.
˚
both using Mo-Ka X-radiation (k = 0.71073 A). A Bruker
PROTEUM diffractometer using Cu-Ka radiation (k = 1.54157
˚
A) was used for 11·(CH2Cl2). All experiments were performed
using a single crystal coated in paraffin oil mounted on a glass
fibre.10a All three diffractometers used a CCD area detector and
intensities were integrated10b from several series of exposures, each
exposure covering 0.3◦ in x. Absorption corrections were based
on based on multiple and symmetry-equivalent measurements
using SADABS V2.10,10c and structures were refined against all
2
Fo data with hydrogen atoms riding in calculated positions, with
isotropic displacement parameters equal to 1.5 times (methyl and
hydroxyl hydrogen atoms) or 1.2 times (all other hydrogen atoms)
using SHELXTL.10d Complex neutral-atom scattering factors
were used.10e Crystal and refinement data are given in Table 6.
For 11, the crystal also contained disordered solvent molecules
which could not be resolved. This was modelled using
SQUEEZE,10f which found the unit cell to contain an ex-
8 J. N. L. Dennett, A. L. Gillon, K. Heslop, D. J. Hyett, J. S. Fleming,
C. E. Lloyd-Jones, A. G. Orpen, P. G. Pringle, D. F. Wass, J. N. Scutt
and R. H. Weatherhead, Organometallics, 2004, 23, 6077.
9 A. R. O’Connor, C. Nataro, J. A. Golen and A. L. Rheingold,
J. Organomet. Chem., 2004, 689, 2411.
tra 166 electrons distributed over two voids, each of volume
3
˚
317 A . This correlates well with having four extra molecules
of dichloromethane per unit cell, one per asymmetric unit.
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 1160–1168 | 1167
©

View Article Online
10 (a) SMART diffractometer control software. SMART collections:
version 5.054, Bruker-AXS Inc., Madison, WI, 1997–1998; SMART
APEX collections: version 5.628, Bruker-AXS Inc., Madison, WI, 2005;
PROTEUM collections: version 5.625, Bruker-AXS Inc., Madison,
WI, 1997–2001; (b) SAINT V6.02 integration software (4), Siemens
Analytical X-ray Instruments Inc., Madison, WI, 1994; SAINT V6.28
(5), Siemens Analytical X-ray Instruments Inc., Madison, WI, 2001;
SAINT V6.41 (6), Siemens Analytical X-ray instruments Inc., Madison,
WI, 2002; SAINT V7.06A (11), Siemens Analytical X-ray Instruments
Inc., Madison, WI, 2003; (c) SADABS V2.10, G. M. Sheldrick,
University of Go¨ttingen, 2003; (d) SHELXTL program system version
5.1, Bruker Analytical X-ray Instruments Inc., Madison, WI, 1998;
(e) International Tables for Crystallography, Kluwer, Dordrecht, 1992,
vol. C; (f) SQUEEZE, incorporated into PLATON, A. L. Spek, (2005),
A Multipurpose Crystallographic Tool, Utrecht University, Utrecht,The
Netherlands.
1168 | Dalton Trans., 2007, 1160–1168
This journal is
The Royal Society of Chemistry 2007
©