Design, Syntheses and Application of New Phosphine and Dithiophosphinate Complexes of Nickel: Catalyst Precursors for the Oligomerization of Ethylene

Article abstract of DOI:10.1021/ic970798w

Several new nickel(II) and nickel(0) complexes have been prepared which are active and selective for the oligomerization of ethylene to higher α-olefins. Nickel(II) complexes of the type NiCl(PR₃)S₂PR'₂ (R = Ph, Bu; R' = Me. Ph) contain one hound halide ion, a monodentate phosphine, and a bidentate chelating dithiophosphinate ligand. The single-crystal X-ray structure of NiCl(PPh₃)S₂PMe₂ is reponed. Crystal data: triclinic, space group P1 (No. 2), a = 10.288 (2) A?, b = 13.205 (3) A?, c = 16.157 (2) A?, α = 82.63(1)°, β = 88.29(1)°, γ = 78.87(2)°, Z = 4. Final R and R_w values were 0.042 and 0.048, respectively. There are two independent molecules A and B in the unit cell. In each case the molecular structure is based on a distorted square planar Ni(II) center subtended with a slightly asymmetrically bound chelated dithiophosphinate ligand with a relatively tight binding angle (S-Ni-S 87.4°). Molecule A is slightly more distorted from the square planar geometry (in that the P(V) of the dithiophosphinate chelate deviates more significantly from the coordination plane) than B. The bond distances (Ni-S (av) (trans to Cl) 2.205(2) A?, Ni-S (av) (trans to the phosphine ligand) 2.254(5) A?, Ni-Cl (av) 2.179(5) A and Ni - P (ave) 2.021(3) A?) do not differ significantly in the two molecules. Also prepared were the nickel(0) complexes Ni(Ph₂P-N=C(H)Ph)₄ and Ni(Ph₂PCH₂P(Ph₂)NC₆F ₄CN)₂. Each of the complexes is more active for the oligomerization of ethylene than are comparable conventional complexes of nickel with typical phosphine ligands. Batch reactions of the nickel complexes in combination with a 150 times molar excess of Et_nAlCl_3-n in toluene oligomerized ethylene at temperatures below 25 °C under 0.1-2.8 MPa of ethylene and a lesser amount of inert gas. The catalyst mixtures have high activity (for example, a rate of 7.64 × 10⁵ mol of C₂(mol of cat)^-1 h ^-1 was observed for Ni(Ph(H)NCPPh₂)₄ with excess diethylaluminum chloride cocatalyst) and high selectivity to α-olefins, in particular I-butene. The selectivity to 1-butene increases with increasing olefin pressure.

Full text of DOI:10.1021/ic970798w

Inorg. Chem. 1998, 37, 757-763
757
Design, Syntheses and Application of New Phosphine and Dithiophosphinate Complexes of
Nickel: Catalyst Precursors for the Oligomerization of Ethylene
Ronald G. Cavell,*,† Beth Creed,‡ Lucio Gelmini,† David J. Law,† Robert McDonald,†
Alan R. Sanger,† and Arpad Somogyvari‡
Department of Chemistry, University of Alberta, Edmonton, AB, Canada T6G 2G2, and NOVA
Chemicals Ltd., 2928 16th Street NE, Calgary, AB, Canada T2E 7K7
ReceiVed June 26, 1997
Several new nickel(II) and nickel(0) complexes have been prepared which are active and selective for the
oligomerization of ethylene to higher R-olefins. Nickel(II) complexes of the type NiCl(PR3)S2PR′2 (R ) Ph, Bu;
R′ ) Me, Ph) contain one bound halide ion, a monodentate phosphine, and a bidentate chelating dithiophosphinate
ligand. The single-crystal X-ray structure of NiCl(PPh3)S2PMe2 is reported. Crystal data: triclinic, space group
P1h (No. 2), a ) 10.288 (2) Å, b ) 13.205 (3) Å, c ) 16.157 (2) Å, R ) 82.63(1)°, â ) 88.29(1)°, γ ) 78.87(2)°,
Z ) 4. Final R and Rw values were 0.042 and 0.048, respectively. There are two independent molecules A and
B in the unit cell. In each case the molecular structure is based on a distorted square planar Ni(II) center subtended
with a slightly asymmetrically bound chelated dithiophosphinate ligand with a relatively tight binding angle (S-
Ni-S 87.4°). Molecule A is slightly more distorted from the square planar geometry (in that the P(V) of the
dithiophosphinate chelate deviates more significantly from the coordination plane) than B. The bond distances
(Ni-S (av) (trans to Cl) 2.205(2) Å, Ni-S (av) (trans to the phosphine ligand) 2.254(5) Å, Ni-Cl (av) 2.179(5)
Å and Ni-P (ave) 2.021(3) Å) do not differ significantly in the two molecules. Also prepared were the nickel(0)
complexes Ni(Ph2P-NdC(H)Ph)4 and Ni(Ph2PCH2P(Ph2)NC6F4CN)2. Each of the complexes is more active for
the oligomerization of ethylene than are comparable conventional complexes of nickel with typical phosphine
ligands. Batch reactions of the nickel complexes in combination with a 150 times molar excess of EtnAlCl3-n in
toluene oligomerized ethylene at temperatures below 25 °C under 0.1-2.8 MPa of ethylene and a lesser amount
of inert gas. The catalyst mixtures have high activity (for example, a rate of 7.64 × 105 mol of C2(mol of cat)-1
h-1 was observed for Ni(Ph(H)NCPPh2)4 with excess diethylaluminum chloride cocatalyst) and high selectivity
to R-olefins, in particular 1-butene. The selectivity to 1-butene increases with increasing olefin pressure.
Introduction
design shows that advancements in activity and selectivity are
still attainable.4,5 Recent development of catalysts with activities
which are at least 2 orders of magnitude higher than existing
commercial (e.g. SHOP) catalysts, Figure 1, have been
reported.2-7 However, in many cases these more active catalyst
systems also effect isomerization of the products to the
thermodynamically favored, but undesirable, internal olefins.4
In contrast, some of the new complexes herein reported are both
highly active and highly selective catalysts for the formation
of 1-butene and higher R-olefins.
Many homogeneous oligomerization processes employ tolu-
ene or alkane solvents, in part because of the excellent solubility
of ethylene in these solvents. However, this also requires that
the catalyst precursor be soluble in this medium. To provide
appropriate solubility and activity, we have constructed two
types of catalyst precursors using a selection of our previously
reported phosphorus-nitrogen ligands8,9 that contain a range
Linear R-olefins, derived from basic olefins such as ethylene
and propylene, are extremely useful chemicals with a large range
of commercial applications, e.g., biodegradable detergents,
lubricants, plasticizers, and as formulation comonomers in a
wide variety of polyolefins. Their selective preparation by
catalytic oligomerization of ethylene has been of significant
interest to both academic and industrial chemists for many years.
The catalytic oligomerization of ethylene to yield higher
R-olefins has most commonly employed complexes of titanium,
nickel, and, to a lesser extent, zirconium. Much of the activity
and selectivity of these complexes has been effected by the
application of suitably designed ligands or ligand combinations,
and by optimization of catalyst/cocatalyst (typically EtnAlCl3-n
or methylaluminoxane) ratios. An excellent review covering
literature to 1989 summarizes the background literature con-
cerning the development and reaction mechanisms of these
catalysts.1 Although it had appeared that a “state of the art”
had been reached with nickel catalyst systems such as those
used in the Shell Higher Olefins Process (SHOP),2,3 recent
interest involving subtle ligand modifications and new ligand
(3) Keim, W.; Kowalt, F. H.; Goddard, R. H.; Kru¨ger, C. Angew. Chem.,
Int. Ed. Engl. 1978, 17, 466-467; Angew. Chem. 1978, 90, 498.
(4) Abeywickrema, R.; Bennett, M. A.; Cavell, K. J.; Kony, M.; Masters,
A. F.; Webb, A. G. J. Chem. Soc., Dalton Trans. 1993, 59-68.
(5) Ostoja Starzewski, K. A.; Born, L. Organometallics 1992, 11, 2701-
2704.
(6) Bogdanovic, B.; Henc, B.; Karmann, H. H.; Nu¨ssel, H.; Walter, D.;
Wilke, G. Ind. Eng. Chem. 1970,62, 34-44.
(7) (a) Keim, W. Polymer Sci. Ser. A 1994, 36, 1372-1379; Vysokomol.
Soedin., Ser. A 1994, 36, 1644-1652. (b) Hirose, K.; Keim, W. J.
Mol. Catal. 1992, 73, 271-276.
† University of Alberta.
‡ NOVA Chemicals Ltd.
(1) Skupi’nska, J. Chem. ReV. 1991, 91, 613-648.
(2) Bauer, R. S.; Chung, H.; Glockner, P. W.; Keim, W. U.S. Patent 3,-
644,563, 1972.
S0020-1669(97)00798-2 CCC: $15.00 © 1998 American Chemical Society
Published on Web 02/04/1998

758 Inorganic Chemistry, Vol. 37, No. 4, 1998
Cavell et al.
Scheme 1
Figure 1. Some examples of effective nickel homologation catalysts.
The cocatalyst is typically AlClEt2 (DEAC). Compound a is a typical
SHOP catalyst prepared from nickel(0) precursors and Ph2PCH2CO2H
or P-ylides [e.g., Ph3PdCH-C(O)Ph]. See also refs 1, 5, and 7.
of aryl and alkyl organic substituents on the backbone. We
have also explored the applications of nickel dithiophosphi-
nates10,11 and dithiophosphinates in combination with other
ligands as potential catalyst precursors.
Results
Scheme 2
(1) Preparation and Structures of Active Complexes.
Nickel(0) catalyst precursors such as Ni(PPh3)4 are well-known
and can be readily prepared from starting materials such as
Ni(cod)2.12-14 Similarly, treatment of Ni(cod)2 with the phos-
phinimine ligand Ph2P-NdC(H)Ph, 1,9 in 2 and 4 M ratio
combinations respectively yielded complexes 2 and 3 as air-
and moisture-sensitive dark brown solids (Scheme 1).
Treatment of Ni(cod)2 with 2 equiv of the phosphoranimine
phosphine ligand 48 gave the new nickel complex 5. In a direct
comparison, complexes 2 and 3 exhibited good hydrocarbon
solubility whereas complex 5 would only dissolve in warm
methylene chloride and was insoluble in toluene. The poor
solubility in organic solvents of metal complexes of tetrafluo-
rocyanobenzene-substituted ligands has been noted previ-
ously.15,16 The 31P NMR spectra of complexes 2 and 3 were
unremarkable, being singlets at 74.0 and 73.2 ppm, respectively.
Both the solid state and solution IR spectra of 2 and 3 also
show only one absorption for the CdN stretch, at 1622 and
1601 cm-1, respectively, further confirming the chemical and
geometrical similarities of these nickel/phosphine complexes.
The second class of complex targeted was the series of square
planar nickel(II) complexes with a variety of phosphine and
dithiophosphinate10 ligands. Few complexes of this type have
been reported (most examples are homoleptic),10,11 although
some structurally related analogues are known.17,18 The syn-
thetic path to such complexes is straightforward. Treatment of
NiCl2(PPh3)2 with the dithiophosphinate salts NH4S2PPh2 and
NaS2PMe2 (Scheme 2) yielded the purple microcrystalline
complexes 6 and 7 by elimination of NH4Cl or NaCl, respec-
tively, and PPh3. Free triphenylphosphine was clearly identified
(by 31P{1H} NMR at δ -5.8 ppm) in the reaction filtrates after
removal of the product. Complexes 6 and 7 each exhibit
excellent hydrocarbon solubility and are structurally similar to
the less readily prepared NiCl(PR3)(sacsac) complexes which
are known to be highly active oligomerization catalysts.4
The 31P{1H} NMR spectra of 6 and 7 are of considerable
interest. At room temperature the only resonance observed for
either complex is that of the dithiophosphinate moiety at
approximately δ 75-80 ppm (assigned by comparison to the
31P NMR spectra of corresponding homoleptic Ni(S2PR2)2
complexes; e.g. the 31P chemical shift of an authentic11 sample
3
of Ni((EtO)2PS2)2 is +82.5 ppm). No JPP coupling was
observed. The signal for the triphenylphosphine component is
not visible at ordinary temperatures; however, cooling the CD2-
Cl2 solution to -90 °C shows, as illustrated in Figure 2, the
bound triphenylphosphine component. Once again, there is no
(8) Katti, K. V.; Cavell, R. G. Organometallics 1989, 8, 2147-2153.
(9) Reed, R. W. Ph.D. Thesis, University of Alberta, 1992.
(10) Haiduc, I.; Sowerby, D. B. Polyhedron 1995, 15, 2469-2521.
(11) Cavell, R. G.; Byers, W.; Day, E. D.; Watkins, P. M. Inorg. Chem.
1972, 11, 1598-1606.
3
conclusive evidence of JPP coupling, but both resonances are
broadened. Below -90 °C the complex precipitates from
solution (CD2Cl2 mp -97 °C). We suggest that the loss of the
triphenylphosphine signal at ordinary temperatures in 6 and 7
is most likely attributable to exchange phenomenona, in
particular the facile dissociation and recombination of the
phosphine ligands and the participation of chelate a monomer
rearrangement of the dithiophosphinate ligands. Although we
cannot exclude absolutely the participation of a square planar
(diamagnetic) a tetrahedral (paramagnetic) equilibrium, the fact
that no EPR signal was observed for solutions of these
(12) Ittel, S. D. Inorg. Synth. 1977, 17, 117-124.
(13) Cross, R. J., In Dictionary of Organometallic Compounds; J. E.
Macintyre, Ed.; Chapman and Hall: London, 1995; Vol. 3, pp 2676
(Entry Ni-00326).
(14) Parshall, G. W.; Ittel, S. D. Homogeneous Catalysis: The Applications
and Chemistry of Catalysis by Soluble Transition Metal Complexes;
2nd ed.; Wiley: New York, 1992.
(15) Law, D. J.; Bigam, G.; Cavell, R. G. Can. J. Chem. 1995, 73, 635-
642.
(16) Law, D. J.; Cavell, R. G. J. Mol. Catal. 1994, 91, 175-186.
(17) Maxfield, P. L. Inorg. Nucl. Chem. Lett. 1970, 6, 693-696.
(18) Cavell, K. J.; Masters, A. F. J. Chem. Res. 1983, 72-73.

Phosphine and Dithiophosphinate Complexes of Nickel
Inorganic Chemistry, Vol. 37, No. 4, 1998 759
Scheme 4
Figure 2. Temperature dependent 31P{1H} NMR spectra of the nickel
dithiophosphinate/phosphine complex 7 in CD2Cl2.
Scheme 3
Cl2 or toluene-d8 as the solvent, indicating an exchange
phenomenon with very low activation energy. However, the
1H NMR spectrum was less broadened. Consequently an
accurate ratio of signals for phenyl:n-butyl ratio could be made
1
by integration of the H NMR signals, the results of which
supported the formulation of complex 10 as the bis(tri-n-
butylphosphine)diphenyldithiophosphinate nickel chloride com-
plex. The 31P{1H} NMR spectra of 11 exhibited two distinct
and characteristic resonances in a ratio of 1:2, a sharp singlet
at δ +46.8 ppm, which is indicative of a bound dithiophosphi-
nate moiety, and a broad singlet at δ -2.0 ppm which can be
attributed to two tri-n-butylphosphine ligands in essentially
similar environments. It is suggested that the broadened nature
of the NMR spectra of these complexes is due to a facile
structural isomerism between square pyramidal and square
planar geometries, as depicted in Scheme 4. These types of
structural changes have been noted previously in nickel(II)
complexes, and interchange between four-, five- and six-
coordinate structures are known.20 Further evidence of isomer-
ism and exchange processes in these complexes was obtained
by treating 11 with triphenylphosphine. At -60 °C the 31P
NMR spectrum of 11 showed a sharp singlet at δ +47.0 ppm
(indicative of the dithiophosphinate moiety) plus several
broadened signals lying between δ +5.0 and δ -15.0 ppm,
suggesting an exchange process which scrambles the phosphine
ligands as postulated in Scheme 4. Attempts to recover a single
compound from this mixture were unsuccessful. The possibility
that 10 and 11 are four-coordinate cations which exist as chloride
salts has been ruled out by conductivity measurements in
nitromethane solution, which showed no electrolyte character
for either complex. These complexes exhibit excellent solubility
in most hydrocarbon solvents, except n-pentane, which is
attributed to the nonrigidity of the pendent hydrocarbon groups
on the tris-n-butylphosphine ligands.
complexes at room temperatures where the probability of a
paramagnetic species is greatest and also the fact that both the
1H and 31P NMR signals due to the dithiophosphinate moiety
are not significantly broadened nor are they shifted from values
typical of bound dithiophosphinates in diamagnetic square planar
Ni(II) complexes suggest that the exchange process is not of
this type. The possibility of the general lability of these
complexes leading to active, coordinatively unsaturated com-
plexes under catalytic conditions although not definitely estab-
lished can be postulated.
Treatment of NiCl2(PPh3)2 with 2 equiv of the diethoxy-
dithiophosphinate ion gives the previously reported green
crystalline solid complex Ni{S2P(OEt)2}2PPh3, 819 (Scheme 3).
This five-coordinate complex is also easily accessible by
treatment of the complex Ni{S2P(OEt)2}2, 9, with PPh3.
However, the triphenylphosphine ligand in 8 is very labile, and
the isolation of the pure complex (Scheme 3) requires experi-
mental care. The broad nature of the NMR spectra of 8 makes
it difficult to characterize the complex by this method; however,
X-ray diffraction analysis of crystals of the product obtained
from our reactions is in agreement with previously reported
crystallographic data for this complex.19
Treatment of NiCl2(PBu3)2 with either ammonium diphen-
yldithiophosphinate or sodium dimethyldithiophosphinate in
toluene gave the corresponding complexes 10 and 11 as very
air-sensitive purple oils (Scheme 4).
The reactions of the dithiophosphinate salts with the nickel-
(II) dichloride bis(phosphine) complexes proceed smoothly and
The 31P{1H} NMR spectra of 10 (R ) Ph) remained broad
and unresolved at all temperatures accessible using either CD2-
(19) Liu, S.; Lin, C.; Yu, Y.; Zhu, D.; Zu, Z.; Gou, S.; You, X. Acta
Crystallogr. 1991, C47, 43-45.
(20) Cotton, F. A.; Wilkinson, G. AdVanced Inorganic Chemistry, 4th ed.;
Wiley-Interscience, New York, 1980.

760 Inorganic Chemistry, Vol. 37, No. 4, 1998
Cavell et al.
Table 2. Selected Interatomic Distances (Å) and Angles (deg) for 7
molecule A
Ni(1)-Cl(1)
molecule B
Ni(2)-Cl(2)
2.175(2)
2.205(2)
2.259(2)
2.187(2)
2.022(2)
1.991(2)
1.800(7)
1.781(7)
1.832(5)
1.809(6)
1.830(5)
2.183(2)
2.206(2)
2.249(2)
2.183(2)
2.019(2)
2.002(2)
1.787(6)
1.784(7)
1.809(5)
1.817(6)
1.821(5)
Ni(1)-S(11)
Ni(1)-S(12)
Ni(1)-P(12)
S(11)-P(11)
S(12)-P(11)
P(11)-C(11)
P(11)-C(12)
P(12)-C(21)
P(12)-C(31)
P(12)-C(41)
Ni(2)-S(21)
Ni(2)-S(22)
Ni(2)-P(22)
S(21)-P(21)
S(22)-P(21)
P(21)-C(51)
P(21)-C(52)
P(22)-C(61)
P(22)-C(71)
P(22)-C(81)
Cl(1)-Ni(1)-S(11) 177.53(9) Cl(2)-Ni(2)-S(21) 176.38(8)
Figure 3. ORTEP21 drawing for one molecule of 7 in the crystal lattice
showing the atom numbering scheme. Hydrogen atoms on the phenyl
rings have been omitted for clarity. Atoms are shown as 20% ellipsoids.
Cl(1)-Ni(1)-S(12)
Cl(1)-Ni(1)-P(12)
S(11)-Ni(1)-S(12)
S(11)-Ni(1)-P(12)
90.44(7) Cl(2)-Ni(2)-S(22)
88.55(6) Cl(2)-Ni(2)-P(22)
87.44(6) S(21)-Ni(2)-S(22)
93.71(6) S(21)-Ni(2)-P(22)
89.71(7)
89.43(7)
87.98(7)
93.03(7)
Table 1. Crystallographic Experimental Details for 7
S(12)-Ni(1)-P(12) 172.64(7) S(22)-Ni(2)-P(22) 176.66(7)
Ni(1)-S(11)-P(11)
Ni(1)-S(12)-P(11)
86.04(8) Ni(2)-S(21)-P(21)
85.32(8) Ni(2)-S(22)-P(21)
86.00(8)
85.26(8)
A. Crystal Data
empirical formula
fw
C20H21ClNiP2O2
481.63
S(11)-P(11)-S(12) 100.53(9) S(21)-P(21)-S(22) 100.6(1)
S(11)-P(11)-C(11) 113.2(3)
S(11)-P(11)-C(12) 110.9(3)
S(12)-P(11)-C(11) 111.6(3)
S(12)-P(11)-C(12) 114.2(3)
C(11)-P(11)-C(12) 106.5(4)
Ni(1)-P(12)-C(21) 119.0(2)
Ni(1)-P(12)-C(31) 106.3(2)
Ni(1)-P(12)-C(41) 109.1(2)
C(21)-P(12)-C(41) 100.8(2)
C(31)-P(12)-C(41) 104.2(2)
P(12)-C(21)-C(22) 119.8(4)
P(12)-C(21)-C(26) 120.9(4)
P(12)-C(31)-C(32) 125.1(4)
P(12)-C(31)-C(36) 116.6(4)
P(12)-C(41)-C(42) 118.2(4)
P(12)-C(41)-C(46) 122.7(4)
S(21)-P(21)-C(51) 112.5(2)
S(21)-P(21)-C(52) 111.1(2)
S(22)-P(21)-C(51) 112.2(3)
S(22)-P(21)-C(52) 114.7(3)
C(51)-P(21)-C(52) 106.0(3)
Ni(2)-P(22)-C(61) 114.2(2)
Ni(2)-P(22)-C(71) 109.8(2)
C(61)-P(22)-C(71) 108.3(3)
C(61)-P(22)-C(81) 101.8(3)
C(71)-P(22)-C(81) 105.1(3)
P(22)-C(61)-C(62) 120.6(4)
P(22)-C(61)-C(66) 120.7(4)
P(22)-C(71)-C(72) 124.4(5)
P(22)-C(71)-C(76) 116.8(5)
P(22)-C(81)-C(82) 118.4(4)
P(22)-C(81)-C(86) 123.0(4)
crystal dimensions (mm)
0.55 × 0.20 × 0.08
crystal system
space group
unit cell parameters
a (Å)
b (Å)
c (Å)
triclinic
P1h (No. 2)
10.288 (2)
13.205 (3)
16.157 (2)
82.63 (1)
88.29 (1)
78.87 (2)
2135.9 (9)
4
R (deg)
â (deg)
γ (deg)
V (Å3)
Z
F
calcd (g cm-3
)
1.498
13.77
µ (cm-1
)
B. Data Collection and Refinement Conditions
diffractometer
Enraf-Nonius CAD4
radiation (λ [Å])
Mo KR (0.710 73)
molecules in the unit cell, the molecular structures are similar,
being derivatives of the typical Ni(II) square planar structure.
In molecule A, the P(V) of the dithiophosphinate chelate
deviates more significantly from the coordination plane and thus
is slightly more distorted from the square planar geometry than
is the case for B, but otherwise the structures are very similar
and the bond distances comparable. In both cases the nickel
center displays a slightly distorted square planar coordination
because of the tight bite angle of the chelated dithiophosphinate
ligand (S-Ni-S 87.4°) and the action of the bulky triph-
enylphosphine ligand (S-Ni-P 93.7°). However, smaller
dithiophosphinate chelate bite angles of 82.06(6)° have been
observed in other nickel(II) bis(dithiophosphinato) complexes22-24
and angles as tight as 75.2(1)° have been reported for other
metals,25 so the deviations are not unusual. The Ni-S bond
lengths, which are in the range 2.21-2.26 Å, are comparable
with those found for other square planar four-coordinate Ni(II)
dithiophosphinate complexes such as Ni[S2P(OCy)2]2 (2.22
Å).22-24
monochromator
incident-beam, graphite crystal
temperature (°C)
takeoff angle (deg)
detector aperture (mm)
crystal-to-detector distance (mm)
scan type
-50
3.0
(3.00 + tan θ)horiz × 4.00vert
173
θ-2θ
scan rate (deg min-1
scan width (deg)
)
6.7-1.4
0.60 + 0.344 tan θ
data collection 2θ limit (deg)
total data collected
range of absorption correction factors 0.8281-1.1395
50.0
7756 ((h, (k, +l)
total no. of unique data
no. of observations (NO)
final no. of parameters varied (NV)
Ra
7408
3994 (I > 3σ(I))
469
0.042
0.048
1.209
b
Rw
GOFc
a R ) ∑||Fo| - |Fc||/∑|Fo|. b Rw ) [∑w(|Fo| - |Fc|)2/∑wFo ]1/2
.
2
c GOF ) [∑w(|Fo| - |Fc|)2/(NO - NV)]1/2
.
efficiently in most solvents tested, even though both reactants
are in apparent suspension. We have found that certain solvent
combinations give good yields of the complexes, as noted in
the Experimental Section.
X-ray structure quality single crystals of the complex 7 were
obtained by slow diffusion of hexane into a saturated dichlo-
romethane solution of the complex. An ORTEP21 view of the
molecule with atom labeling is shown in Figure 3. Experimental
crystal analysis parameters and selected metrical results are
given in Tables 1 and 2. Although there are two independent
(2) Catalytic Activity. Preliminary Screening. The re-
ported complexes were first tested for oligomerization activity
in the presence of ethylene at 1 atm, with diethylaluminum
chloride (DEAC) as cocatalyst. These screening tests were
(22) Lin, C.-C.; Liu, S.-X.; Xu, Z.; Lin, J.-H.; You, X.-X. Jiegou Huaxue
(Struct. Chem.) 1987, 6, 186-188 (paper in English).
(23) Jones, P. E.; Ansell, G. B.; Katz, L. Acta Crystallogr. 1969, B25,
1939-1943.
(24) Porta, P.; Sgamellotti, A.; Viciguerra, N. Inorg. Chem. 1968, 7, 2625-
2629.
(21) Johnson, C. K.; ORTEP; Report ORNL No. 5138; Oak Ridge National
(25) Drake, J. E.; Mislankar, A. G.; Ratnani, R. Inorg. Chem. 1996, 35,
2665-2673.
Laboratories: Oak Ridge, TN, 1976.

Phosphine and Dithiophosphinate Complexes of Nickel
Inorganic Chemistry, Vol. 37, No. 4, 1998 761
Figure 4. Low-pressure (1 atm) uptake of ethylene at 23 °C by a
catalyst system formed from complex 11 and DEAC in toluene (initial
rate: 1090 mol of C2H4 per mol of Ni per h).
Figure 5. Uptake of ethylene by a catalyst solution formed by complex
11 and DEAC.
Table 3. Product Distributions for Selected Catalyst Precursor
Complexes
designed to provide an indication of the most reactive systems,
and, although oligomerization catalysts are generally used under
pressures in excess of 3 MPa, it was considered that any catalyst
of potential value in industrial applications would show some
activity at low pressure, thus by this approach relatively less
active complexes could be rejected. Although this process
improves the efficiency of catalyst testing, there always remains
the possibility that a system that is seemingly uninteresting at
low pressures will become highly catalytically active at the very
much higher pressures typical of operational use. The fluoro-
aromatic substituted complex 5 had such poor solubility in
toluene that it was not considered as a viable catalyst candidate.
The PNC complexes 2 and 3 showed excellent solubility but
low activity under 1 atm of ethylene. In contrast, the dithio-
phosphinate-substituted complexes 6-11 each exhibited moder-
ate to high activity under 1 atm pressure of ethylene at 23 °C,
with the methyl-substituted dithiophosphinate complex 11 giving
an initial turnover rate of 1090 mol of ethylene per mole of
complex per hour, Figure 4. Complexes giving encouraging
results were selected for more detailed screening, including
product analysis.
compd no.
C4
C6
C8
1-C4 cis- 2-C4 trans-2-C4
3
6
7
8
9
95.26
93.14
83.81
94.41
3.69
6.02
4.87
4.33
0.44 83.16
0.41 52.19
2.53 21.85
0.41 57.21
5.41
17.39
29.04
17.00
2.45
6.69
23.55
32.92
20.20
4.75
76.31 12.11 10.19 69.11
10
11
87.33
96.21
4.90
2.08
7.09 22.59
0.24 41.26
28.57
25.72
36.17
29.23
Table 4. Comparative Studies on Selected Nickel Catalyst
Precursors
compd
no.
rate × 105,
1-C4/∑C4,
wt %
mol/mol h
C4/∑C,b wt %
3
6
7
8
9
10
11
7.64a
26c
87.30
56.03
26.07
60.60
90.56
25.87
42.88
95.26
93.14
83.81
94.41
76.31
87.33
96.21
89c
6.45
2.82
13.2
12.8
a At 2.8 MPa. b ∑C ) total product. c Estimate.
High-Pressure Tests. Comparative tests for activity and
selectivity were performed for those materials passing the initial
screening. These comparative tests were performed at an
ethylene partial pressure of 1.6 MPa, an initial temperature of
15 °C, and a DEAC to nickel ratio of 150:1. The ethylene
uptake rate is considered to be the overall rate of reaction. A
plot of the ethylene flow rate against time was fitted to an
exponential curve and extrapolated to zero time in order to
establish the initial rate of reaction (Figure 5). Due to a rapid
and often large initial exotherm, isothermal data for the early
portion of the experiment were not available. The ethylene flow
rates for complexes 6 and 7 exceeded the flowmeter maximum
(i.e. 20 dm3/min). The process is very exothermic, and the
temperature rises rapidly to a value 30-55 °C above the initial
setpoint. In earlier tests, we had observed a deterioration of
catalyst lifetime at elevated temperatures and surmised that the
catalyst was heat-sensitive. Consequently, only rough estimates
could be made for the initial rates [(2-8) × 106 mol/mol h] for
these two catalyst systems. It is interesting to note that the
immediate coordination environment about Ni in complexes 6
and 7 is very similar to those reported by Masters et al.26 for
systems which provided catalyst activities in excess of 3.0 ×
106 mol/mol h. Furthermore, varying the phosphine substituents
affects only the catalyst activity whereas varying the substituents
on the chelate affected both the activity and the selectivity.26
Keim3 has suggested that catalytic activity in this system requires
that the nickel be in the +1 oxidation state, in a square planar
environment, and that the complex must contain a chelating
ligand. Preferably, one of the remaining ligands is labile, to
facilitate the formation of a vacant coordination site. Under
these circumstances, the chelate controls both catalyst activity
and selectivity while the relatively nonlabile ligand affects only
the overall activity of the catalyst.
The initial rates and selectivities for the catalyst systems
described herein, determined both as the weight fraction of
1-butene in the C4 stream and as the weight fraction of the C4
stream in the total product slate, are given in Tables 3 and 4.
These results are consistent with those reported by Keim et al.7
and Brown and Masters26 for related systems similar in some
respects to 6 and 7. It appears that in these four coordinate
complexes the aliphatic dithiophosphinates provide a more active
but less selective catalyst while in five-coordinate complexes
(cf. 10 and 11) the reverse applies. It is therefore surprising
that 9 shows such high activity and selectivity in apparent
contravention of Keim’s postulates.7 Even more surprising is
that the coordination of triphenylphosphine to 8 would yield a
more active catalyst. Consistent with the findings of Keim,7
(26) Brown, S. J.; Masters, A. F. J. Organomet. Chem. 1989,367, 371-
374.

762 Inorganic Chemistry, Vol. 37, No. 4, 1998
Cavell et al.
Experimental Section
Solvents were dried and distilled prior to use, and manipulations
were carried out using standard Schlenk techniques under an argon
atmosphere. NMR spectra were taken on a Bruker 200 MHz instrument
operating at 81.015 MHz for 31P referenced to H3PO4 and at 200.133
1
MHz for H referenced to SiMe4. Spectra were measured relative to
the 2H lock of the internal deuterated solvent. Positive shifts lie
downfield in all cases. Infrared spectra were taken on a Fourier
transform instrument using both microscope and dichloromethane cast
techniques. Mass spectra were taken using fast atom bombardment,
and the resultant ions were detected with a positive ion detector.
Microanalyses were carried out in the Department of Chemistry
Microanalytical Laboratory of the University of Alberta. The ligands
Ph2PNCHPh (1)9 and Ph2PCH2P(Ph2)NC6F4CN (4)8 were prepared as
previously described. NaS2PMe2‚2H2O27 was also prepared as previ-
ously described. The salt NH4S2PPh2 was prepared by treating a toluene
solution of Ph2PS2H with gaseous ammonia until precipitation of the
product occurred.27
Figure 6. Correlation of selectivity and reactivity with ethylene
pressure for the catalytic system formed from 3 and DEAC. The
activity-pressure relationship is given in Table 5.
Initial Screening for Catalytic Activity. A solution of the
appropriate complex (4 × 10-4 mol in 100 cm3 of toluene) was prepared
in a 200 cm3 Schlenk flask containing a magnetic stirrer. The flask
was fitted to a standard hydrogenation apparatus comprising a 500 cm3
buret and bulb filled with iPrOH/glycerol ballast, a mercury manometer,
and a vacuum/gas inlet tap. The system was then evacuated and purged
three times with ethylene. The solution was then stirred vigorously
until the ethylene had equilibrated with the toluene solvent. At this
point, diethylaluminum chloride solution (2 cm3, 1.8 M) was introduced
to the reaction vessel and the timing of the reaction started. Ethylene
uptake, indicated by upward movement of the ballast solution in the
buret, was measured at 1-min intervals until catalytic activity declined.
Parametric Experiments. All high-pressure ethylene oligomer-
ization tests were performed in a 1-dm3 Parr autoclave fitted with a
Magnadrive stirrer assembly. The reactor was preconditioned prior to
use by injecting 2 cm3 of cocatalyst (25 wt % in toluene) into the reactor
containing about 500 cm3 of toluene and stirring for at least 1 h
(overnight preferred). (This procedure removes any oxygen, moisture,
and any other potential reaction poisons adsorbed on the reactor walls.
Variable reactivity results are often obtained if such preconditioning
is not done.) After this preconditioning solution was removed from
the reactor, the reactor was rinsed with fresh toluene, dried, flushed
with N2, and loaded with 200 cm3 of degassed toluene. The reactor
was then brought to reaction pressure with ethylene. The appropriate
amount of cocatalyst dissolved in toluene (10 cm3) was injected using
an injection bomb (internal volume 30 cm3) pressurized to 5.6 MPa
with purified N2. The reactor contents were stirred for 5 min before
the catalyst solution (10 cm3) was similarly injected. The reactor was
operated in the semibatch mode at constant pressure with ethylene being
fed on demand. The rate of ethylene consumption and the total amount
of ethylene fed to the reactor were measured using a precalibrated mass
flowmeter and recorded automatically, as were continuous temperature
and pressure readings, by use of Labtech Notebook. The reaction was
terminated by injecting 10 cm3 of water in the same manner as the
catalyst and cocatalyst solutions. The volume of gaseous products
leaving the reactor at the end of the run was measured using a wet test
meter corrected to STP. These gases were collected and analyzed by
Table 5. Dependence of Activity on C2H4 Pressurea
C2H4 press, psig
activity, mol/mol h
1-C4/∑C4
30
60
110
215
400
2000
18 155
62 586
87 983
114 948
214 517
540 389
9.4
46.1
67.0
82.9
87.3
98.2
a For the catalytic system formed from 3 and DEAC.
however, the addition (i.e. coordination) of triphenylphosphine
shifts the product distribution to the shorter carbon oligomers.
In contrast to the behavior of the related system,26 we do not
find isomerization of the 1-butene products to be faster than
ethylene oligomerization. In all cases, the present system
provides an amount of 1-butene in the C4 fraction which is
greater than expected from thermodynamics. Moreover, there
is a relationship between this selectivity and activity (Figure 6)
which is not intuitively obvious. We have observed that both
the overall reaction rate and the selectivity to 1-butene increase
with the ethylene partial pressure (Table 5). Thus it may be
postulated that isomerization is not intimately associated with
the oligomerization reaction and that the rate of product
isomerization is considerably less than that of the ethylene
oligomerization process. However, 1-butene, the primary
product of oligomerization, undergoes slow isomerization to cis-
and trans-2-butene when in contact with the catalyst. Support
for the above postulate was obtained by contacting 1-butene
with complex 6 at 15 °C. Of the total product obtained, 83.6
wt % was the collection of butene isomers and the remainder
was octenes. The butene distribution was 2.2% 1-butene, 72.1%
cis-2-butene, and 25.7% trans-2-butene.
1
gas chromatography (GC, HP 5890 series II, 10 ft. × /8 in. Porapak
Conclusions
R, 80/100 mesh, 50 °C for 8 min then 10 °C/min to 250 °C) to
determine the total amount of ethylene consumed. The reactor contents
were drained and weighed, and an aliquot was analyzed by gas
chromatography (HP 5890 series II, 50 m × 0.2 mm PONA 50 µm
liquid phase, 30 °C for 8 min then 5 °C/min to 250 °C). The results
of the gas and liquid analyses were combined to determine the product
distribution (Table 3).
Several new nickel complexes in conjunction with common
cocatalysts such as DEAC provide active and selective catalysts
for the oligomerization of ethylene to 1-butene. Either Ni(II)
or Ni(0) complexes may be used, the former being more active
but the latter being more selective. It is notable that the complex
Ni(Ph2PNC(H)Ph)4 shows unusually high selectivity for oligo-
merization to R-olefins in contrast to previously evaluated
complexes based on standard phosphines, indicating the en-
hancements in activity which can accrue from the modification
of the ligands alone. Unlike other comparable highly active
oligomerization systems, isomerization does not play an influ-
ential role in the catalytic cycles of the present systems.
Nickel(0) Catalyst Precursors. Ni(Ph2P-NdC(H)Ph)4 (3). Bis-
(1,5-cyclooctadiene)nickel (0.48 g, 1.76 mmol) and Ph2P-NdC(H)-
Ph (2.023 g, 6.99 mmol) were stirred as a suspension in hexane in a
flask under an argon atmosphere at 0 °C. Dichloromethane was slowly
added to the suspension until dissolution occurred, whereupon a brown
(27) Day, E. D. Ph.D. Thesis, University of Alberta, 1972.

Phosphine and Dithiophosphinate Complexes of Nickel
Inorganic Chemistry, Vol. 37, No. 4, 1998 763
precipitate immediately formed. The mixture was allowed to warm to
room temperature, and the precipitate was filtered and dried under
vacuum to give 1 as a product sensitive to both air and moisture.
NMR: 31P (81 MHz), δ 74.0 ppm (vs H3PO4); 1H (200 MHz), δ 6.8-
8.1 (m, Ph and CH) ppm. IR ν(CdN): 1622 cm-1. MS (FAB) (%):
NiL4, 1216 (0.3), NiL3, 927 (6.4), NiL2, 638 (18.5), NiL, 349 (12.4).
Anal. Found: C, 72.6; H, 5.3; N, 4.2. Calcd for C76H64N4NiP4: C,
72.9; H, 5.4; N, 4.5.
diethyl ether (3 × 20 cm3) and dried under vacuum for 0.5 h. CH2Cl2
(150 cm3) was added to the filtrate, and the resulting mixture was further
filtered to remove any inorganic salts. The compound was then
recrystallized using a mixture of CH2Cl2/hexane. Yield: 5.85 g (81%).
NMR: 31P (CD2Cl2) (81 MHz) (-90 °C), δ 82.3 (s, br), 19.8 (s, br)
ppm. 1H (200 MHz), δ 7.3-7.8 (m, Ph), 1.9 (d, CH3, JPH ) 12.0 Hz)
ppm. MS (FAB) (%): M - Cl, 446 (79), Ni(S2PMe2), 184 (97). Anal.
Found: C, 50.25; H, 4.28; S, 13.37; Cl, 7.39. Calcd for [NiCl-
(PPh3)(S2PMe2)]: C, 49.9; H, 4.36; S, 13.3; Cl, 7.4.
Ni(PPh3)[S2P(OEt)2]2 (8). A mixture of NiCl2(PPh3)2 (1.00 g, 1.5
mmol) and NH4[S2P(OEt)2] (2.76 g, 1.5 mmol) in dichloromethane (150
cm3) was stirred under argon for 0.5 h. The solution was filtered to
remove inorganic salts. The dichloromethane was removed under
vacuum, and the resulting solid was redissolved in diethyl ether (20
cm3). Slow evaporation of the solvent to approximately 5 cm3 at room
temperature resulted in the precipitation of the desired product. The
compound was recrystallized using this method to remove any excess
PPh3 and was identified by comparison with the literature data.19
Yield: 0.69 g 65%. NMR: 31P (CD2Cl2) (81 MHz), δ 40 (br) ppm;
1H (200 MHz), δ 7.2-7.5 (m, Ph), 4.3 (m, CH2), 1.4 (t, CH3) ppm.
MS (FAB) (%): M, 692 (51), Ni(PPh3)2, 583 (43), [Ni(S2P(OEt)2)2],
429 (36).
Ni(cod)(Ph2P-NdC(H)Ph)2 (2). Bis(1,5-cyclooctadiene)nickel
(1.377 g, 5 mmol) and Ph2P-NdCHPh (2.9 g, 10 mmol) were stirred
as a suspension in hexane in a flask under an argon atmosphere at 0
°C. Dichloromethane was slowly added to the suspension until solution
occurred, and immediately a brown precipitate formed. The mixture
was allowed to warm to room temperature, and the precipitate was
filtered and dried under vacuum to give 2 as a product sensitive to
1
both air and moisture. NMR: 31P (81 MHz), δ 73.2 ppm; H (200
MHz), δ 2.5 (m, cod), 5.5 (m, cod olefinic), 7.0-8.1 (m, Ph) ppm. IR
ν(CdN)
: . MS (FAB) (%): [NiL2(cod)]-Ph, 669 (18.6),
1601 cm-1
Ni(cod), 166.8 (22.0) Ph2PNH2, 201 (51.0). Anal. Found: C, 72.6;
H, 3.6; N, 5.5. Calcd for C46H44N2NiP2: C, 72.3; H, 3.6; N, 6.0.
Ni(Ph2PCH2P(Ph)2dN-C6F4CN)2‚CH2Cl2 (5). Bis(1,5-cyclo-
octadiene)nickel (0.387 g, 1.4 mmol) and Ph2PCH2P(Ph2)N-C6F4CN
(1.53 g, 2.67 mmol) were stirred as a suspension in hexane (15 cm3)
in a flask under an argon atmosphere at 0 °C. Dichloromethane (5
cm3) was slowly added to the suspension until a brown solution formed,
from which a red precipitate was subsequently formed. The mixture
was allowed to warm to room temperature, and the precipitate was
filtered and dried under vacuum to give 5 as an analytically pure
product. MS (FAB) (%): NiL, 630.6 (3.1), L, 573 (4.8). Anal.
Found: C, 60.4; H, 3.6; N, 4.1. Calcd for C65H46N4F8NiP4: C, 60.5;
H, 3.6; N, 4.3.
NiCl(PBu3)2(S2PPh2) (10). A mixture of NiCl2(PPh3)2 (1.657 g, 3
mmol) and NH4[S2PPh2] (0.802 g, 3 mmol) in toluene (150 cm3) was
stirred under argon with for 0.5 h. The solution was then filtered to
remove inorganic salts. The toluene was removed under reduced
pressure to quantitatively yield the product as an extremely air-sensitive
purple oil. NMR: 31P (CD2Cl2) (81 MHz), very broad at all
temperatures; 1H (200 MHz) 6.9-8.1 (m, br, Ph), 0.7-1.8 (m, br, Bu).
NiCl(PBu3)2(S2PMe2) (11). A mixture of NiCl2(PBu3)2 (1.657 g, 3
mmol) and NaS2PMe2‚2H2O (0.553 g, 3 mmol) in toluene (150 cm3)
was stirred under argon for 2 h. The solution was then filtered to
remove inorganic salts. The toluene was removed under reduced
pressure to give a high (80-90%) yield of the product as an extremely
air-sensitive purple oil. NMR: 31P (CD2Cl2) (81 MHz), δ 46.8 (s, P),
Nickel(II) Dithiophosphinate Complexes. NiCl(PPh3)(S2PPh2)
(6). A mixture of NiCl2(PPh3)2 (1.645 g, 2.5 mmol) and NH4[S2PPh2]
(0.672 g, 2.5 mmol) in ethanol (150 cm3) was stirred under argon for
0.5 h. Diethyl ether (50 cm3) was added to the resulting solution, and
the mixture was stirred for a further 10 min, after which the solution
was filtered and washed with degassed diethyl ether (3 × 20 cm3).
CH2Cl2 (150 cm3) was then added to the filtrate, and the resulting
mixture was further filtered to remove any inorganic salts. The product
was recrystallized using a mixture of CH2Cl2/hexane. Yield: 1.18 g
(76%). Mp 140-155 °C, decomp. NMR: 31P (CD2Cl2) (81 MHz)
1
-2.0 (s, br, 2P) ppm; H (200 MHz), δ 0.7-1.9 (m, br, alkyl) ppm.
Acknowledgment. We thank the Natural Sciences and
Engineering Research Council of Canada and NOVA Chemicals
Ltd. for support of the work carried out at the University of
Alberta.
1
(-90 °C), δ 77.8 (br, s), 19.9 (br, s) ppm; H (200 MHz), δ 7.0-8.0
(m, Ph) ppm. MS (FAB) (%): M - Cl, 570 (26), NiS2PPh2, 308 (12).
Anal. Found: C, 60.2; H, 4.1; Cl, 5.5; S, 10.9. Calcd for C70H25-
ClNiP2S2: C, 59.5; H, 4.1; Cl, 5.9; S, 10.6.
NiCl(PPh3)(S2PMe2) (7). A mixture of NiCl2(PPh3)2 (10.00 g, 15
mmol) and NaS2PMe2‚2H2O (2.67 g, 15 mmol) in toluene (300 cm3)
was stirred under argon for 2.5 h. The solution was cooled to -20 °C
and filtered, and then the solid residue was washed with degassed
Supporting Information Available: Complete experimental details,
tables of atom coordinates, bond lengths and angles, anisotropic thermal
parameters, and hydrogen atom coordinates for the crystal structure of
7 (21 pages). Ordering information is given on any current masthead
page.
IC970798W