Facile preparation and activation of high-productivity single-site nickel catalysts for highly linear polyethylene

Article abstract of DOI:10.1002/hlca.200290016

A easy-to-prepare series of nickel complexes are reported, which, upon activation with near-stoichiometric alkylating agents, polymerize ethylene with turnover number (TON) of over 10⁵, which brings the catalyst into the range of commercial single-site catalysts based on early transition metals. However, in contrast to the usual metallocenes, the catalyst shows the same activity in coordinating solvents such as THF and dioxane as it shows in toluene. Moreover, the polymer produced is highly linear, which, together with the long catalyst lifetime, suggests that the combination of the nickel catalyst and a Lewis acid suppresses formation of nickel hydrides.

Full text of DOI:10.1002/hlca.200290016

Helvetica Chimica Acta Vol. 85 (2002)
4337
Facile Preparation and Activation of High-Productivity Single-Site Nickel
Catalysts for Highly Linear Polyethylene
by Marco Pickel, Thomas Casper, Annette Rahm, Christian Dambouwy, and Peter Chen*
Laboratorium f¸r Organische Chemie, ETH Z¸rich, CH-8093 Z¸rich
and
Thales Technologies AG, Technoparkstr. 1, CH-8005 Z¸rich
Dedicated to Professor Dieter Seebach on the occasion of his 65th birthday
A easy-to-prepare series of nickel complexes are reported, which, upon activation with near-stoichiometric
alkylating agents, polymerize ethylene with turnover number (TON) of over 105, which brings the catalyst into
the range of commercial single-site catalysts based on early transition metals. However, in contrast to the usual
metallocenes, the catalyst shows the same activity in coordinating solvents such as THF and dioxane as it shows
in toluene. Moreover, the polymer produced is highly linear, which, together with the long catalyst lifetime,
suggests that the combination of the nickel catalyst and a Lewis acid suppresses formation of nickel hydrides.
Introduction. Polymerization of ethylene and other olefins with transition-metal
catalysts is a process of considerable academic as well as industrial interest1). The
original Ziegler Natta catalyst systems heterogeneous catalysts activated in situ
have been continually improved, leading to highly active systems that produce nearly
100 million tons/year of polyolefins. Originally conceived as homogeneous models for
the heterogeneous catalysts, single-site catalysts (SSC), principally metallocenes, have
evolved into commercially important catalysts for high-performance polymers in which
the fine degree of molecular-level control afforded by the structurally well-defined
organometallic complexes can be used to produce superior products, e.g., stereoregular
polymers. Nevertheless, the metallocene and related catalysts of the current generation
of SSCs suffer from several disadvantages. First, the activated catalysts are exceedingly
sensitive to poisoning by impurities that contain coordinating heteroatoms. Second, the
metallocenes themselves are expensive compounds, being prepared in a multistep
process. Lastly, the activation of metallocenes is typically done with methylalumoxane
(MAO) partially hydrolyzed AlMe3 which plays a triple role as sacrificial drying
agent, alkylating reagent, and Lewis acid. Molar excesses of MAO over metallocene
between 100 and 1000 are normal, which becomes a substantial cost factor [2] despite
the low unit cost of MAO. An ideal alternative would be a SSC based on an inexpensive
late transition metal that would show greater tolerance for heteroatoms, removing the
need for MAO as a sacrificial drying agent. A neutral, as opposed to cationic, complex
would not need MAO in its role as Lewis acid. Alkylation by inexpensive
stoichiometric metal-alkyls would be desirable. Lastly, the well-established coordina-
1
)
A complete overview of Ziegler Natta polymerization can be found in [1].

4338
Helvetica Chimica Acta Vol. 85 (2002)
tion chemistry of late transition metals should simplify the synthesis of the catalyst. The
use of neutral, late transition metal complexes as polyolefin catalysts has been limited,
though, by the decidedly modest turnover number (TON) of known systems, i.e., TON
of ca. 103 as opposed to the 105 106 for Ti and Zr-based catalysts. We report the
synthesis and polymerization reactions of a family of simple NiII coordination
complexes, 1 3, containing at least one salicylaldimine ligand.
These catalysts are easily prepared from readily available Ni complexes and salts,
and can be activated by near-stoichiometric amounts of almost any metal-alkyl or
hydride to form a SSC that tolerates heteroatoms in that it shows the same high activity
for ethylene polymerization in THF or dioxane solution as it does in toluene. Moreover,
in contrast to previous systems in the literature, at least one member of the new family
is chemically robust. It shows high stability and long catalyst lifetime after activation,
leading to outstanding catalyst productivity, with turnover numbers (TON) on the
order of 105 in the production of highly linear polyethylene.
Experimental. All manipulations were done under Ar with either standard Schlenk techniques or in an
MBraun −Unilab× glove box unless otherwise specified. Analysis by 1H-NMR was performed with a Varian
Mercury 300 spectrometer. Electrospray ionization mass spectrometry (ESI-MS) was performed with a
Finnigan MAT LCQ Deca ion-trap mass spectrometer with a dynamic nanospray source. Infrared (IR) spectra
were recorded with a Perkin-Elmer 1600 Series FTIR. Toluene, THF, and 1,4-dioxane solvents were purified by
distillation from alkali metal. Ethylene, N35 (> 99.95%) grade, was purchased from Air Liquide and used as
received. MeLi (Fluka, 1.6m in Et2O), and BuLi (Fluka, 2.7m in heptane) were titrated prior to use to establish
their true concentration. Methylalumoxane (MAO, Witco) was used as received. The titer of MAO, in terms of
methide equivalents, varied from batch-to-batch and according to manufacturer, so the MAO results should be
considered to be less quantitatively accurate. BH3 ¥ THF complex was purchased from Fluka (1m in THF) and
used as received.
Polymerizations were performed in two different reactors. The −small× reactor was constructed from thick-
walled, 25-mm OD Pyrex pressure tubes (Ace Glass Inc.) fitted with a Bourdon-tube manometer on a stainless-
steel head fitted with high-pressure valves (Whitey SS-43MA-S4, specified for up to 200 bar). The soln. was
magnetically stirred with or without external temp. control. The stirring requires some comment; the soln. in the
reactor tube was stirred with a 50-mm long, vertically-oriented, magnetic stir-bar spinning at 1500 rpm so that
the soln. formed a thin, rapidly moving annulus extending up the walls of the reactor tube. This measure was
needed to ensure rapid transfer of ethylene from the gas-phase into soln. When stirring was inefficient, mass
transport, rather than chemical kinetics, limited the rate of the overall reaction. The rate of polymerization was
then quantified either by the amount of polyethylene produced at the end of the reaction, or by the (near)
isobaric rate of ethylene consumption. For the former measurement, the polymerization reaction was quenched
by release of the ethylene overpressure and pouring of the soln. into a large excess (typically 10 Â ) of acidified
methanol to precipitate polyethylene. The solid material was collected by filtration on a glass frit, and then dried
under vacuum to constant weight. Measurement of the polymerization rate by the (near) isobaric rate of
ethylene consumption was done by periodic isolation of the reactor tube from its constant-pressure ethylene

Helvetica Chimica Acta Vol. 85 (2002)
4339
source and measurement of the time needed for the pressure to drop by a fixed amount, usually 0.5 bar or less. In
cases of pressure changes, or at the onset of a reaction, measurements were started only after 5 min. to exclude
artifactual pressure changes due to re-equilibration between the gas- and solution-phase. The integrity of the
apparatus was tested up to 6 bar and found to be tight negligible pressure drop for periods up to 48 h, so it can
be assumed that the measured pressure drop is due to polymerization. The duration of polymerization was
limited by either the catalyst lifetime or, more usually in these experiments, the increase in viscosity of the soln.
For the more active systems under the conditions used in this series of experiments, the soln. became viscous
enough to essentially stopreaction after ca. 2 h.
Polymerizations in larger reaction volumes were conducted in a custom-made stainless-steel autoclave
(DC2, NWA GmbH), designated further is the −large× reactor. The mechanically-stirred inner volume of 100 ml
was temp.-controlled by means of a digital thermometer connected to an electrical heater. Injection of reagents
and withdrawal of samples could be done under pressure the entire autoclave was tested up to 300-bar total
pressure. The larger-scale polymerizations were conducted with the similar absolute amounts of catalyst and
activator as in the small reactors, but with solvent volumes between 50 and 100 ml. With the reactor connected to
a constant pressure ethylene source, ethylene consumption could be monitored by means of pressure drop upon
brief isolation of the reactor from the source. The gauge, however, was not suitable for quantitative work at an
accuracy comparable to that in the small reactors, but isobaric ethylene consumption could nevertheless be used
as an indicator that the catalyst was still producing polymer.
Synthesis of 1 from [Ni(cod)2]. A suspension of 1.21 g (4.41 mmol) [Ni(cod)2] in 150 ml of Et2O was cooled
to À1168 and 380 ml (534 mg, 1.0 equiv.) of allyl bromide were added with stirring. The mixture was allowed to
warm to r.t. and stirred until all [Ni(cod)2] had dissolved. The resulting dark red soln. was cooled to À58 and
1 equiv. (1.65 g) of the Na salt of the salicylaldimine ligand, prepared by condensation of the salicylaldehyde and
2,6-diisopropylaniline, both from Aldrich and used as received, was added in 150 ml of THF. The mixture was
allowed to reach r.t. and stirred for 18 h. After the soln. was filtered, the solvent was removed under reduced
pressure. Recrystallization of the brown residue from EtOH yielded between 0.8 1.1 g (42 60%) of 1 as brown
crystals. X-Ray diffraction by single crystals of 1 yielded a structure that is described in the Supplementary
Material (available upon request). IR (KBr): 3603m, 2959m, 1618s, 1597s, 1536s, 1457m, 1430s, 1384m, 1352m,
1
1
1328s, 1236m, 1204m, 1166s, 937w, 870s, 544m. H-NMR (CDCl3): 7.12 (dd, J ˆ 2J ˆ 7.3, 2 H); 7.02 (d, J ˆ 7.3,
4 H); 6.84 (d, J ˆ 11.6, 4 H); 6.55 (s, 2 H); 4.52 (sept., J ˆ 6.7, 2 H); 2.1 (s, 6 H); 1.96 (d, J ˆ 6.7, 12 H); 1.17 (d,
J ˆ 6.7, 12 H); 0.78 (s, 18 H); À8.3 (s, 2 H). ESI-MS (THF/CH3CN, c ˆ 0.1 mg/ml, spray voltage 2.5 kV, 150 ml/h,
‡
capillary voltage 10 V, tube lens offset 15 V): 851 857 (100, [M ‡ H] , cluster of Ni isotopes), 833 (65, [M ‡
‡
H À 18] ), 440 (35), 424 (45), 407 (55), 394 (32), 351 (30).
Synthesis of 1 from NiBr2 ¥ 6 H2O. A soln. of 219 mg (1.0 mmol) NiBr2 ¥ 6 H2O, 351 mg (1 equiv.) of
salicylaldimine ligand and 147 ml (150 mg, 1 equiv.) of DBU in 10 ml of toluene was heated to reflux for 4.5 h.
After cooling to r.t., the amount of solvent was reduced to 5 ml under reduced pressure and the residue was
diluted with 5 ml of pentane. The resulting mixture was filtered and the solvent removed. The brown residue was
recrystallized from Me to yield 31 123 mg (7.3 30%) of 1 as brown crystals, identical to those described above.
Synthesis of
2 from [Ni(acac)2]. A soln. of 512 mg (2.0 mmol) [Ni(acac)2] and 704 mg (1 equiv.)
salicylaldimine ligand in 20 ml of mesitylene was heated to reflux. The resulting excess of acetylacetone was
removed by azeotropic distillation with the solvent for 4 h; 10 ml of fresh solvent were added to the mixture two
times as the distillation proceeded. The remaining solvent was removed under reduced pressure. The solid
residue was dissolved in pentane and filtered through Celite. Removal of the solvent and recrystallization from
pentane yielded 203 355 mg (40 70%) of 2 as a brown powder. IR (KBr): 2963s, 2870m, 1618m, 1581s, 1523s,
1464s, 1431s, 1382s, 1356m, 1331m, 1164m, 777w, 764w. 1H-NMR (CDCl3): 7.37 7.04 (m, 5 H); 6.69 (dd, 1J ˆ 2.2,
2J ˆ 1.0, 1 H); 5.17 (s, 1 H); 4.32 (sept., J ˆ 6.8, 2 H); 2.23 (s, 3 H); 1.47 (d, J ˆ 6.8, 6 H); 1.44 (s, 6 H); 1.36 (s,
9 H); 1.21 (d, J ˆ 6.8, 6 H). ESI-MS (EtOH, c ˆ 0.1 mg/ml, spray voltage 2.7 kV, 150 ml/h, capillary voltage
‡
À20 V, tube lens offset À70 V): 508 (100, [M ‡ H] , cluster of Ni isotopes), 352 (30), 198 (37).
Synthesis of
3
from 1,2-Dimethoxyethanenickel(2‡) Dibromide. To
a soln. of 1.0 g (2.8 mmol)
salicylaldimine ligand in 10 ml of THF, a soln. of 1 equiv of MeLi in Et2O (1m) was added. After the evolution
of gas had ceased, 439 mg (0.5 equiv.) of NiII-dibromide 1,2-dimethoxyethane (DME) complex dissolved in
10 ml of THF were added, and the resulting mixture was stirred for 18 h. After the solvent was removed under
reduced pressure, the brown residue was dissolved in toluene and the soln. was filtered. Removal of the toluene
and recrystallization by slow evaporation from pentane yielded 600 mg (56%) of dark brown crystals. IR (KBr):
1
2962s, 1619m, 1580s, 1534s, 1459m, 1386m, 1362m, 1322m, 1234m, 1206w, 1166s, 1098w, 935w, 764w. H-NMR
(CDCl3; due to the paramagnetic properties of the compound, all signals were very broad and no coupling was
registered): 54.5 (s, 2 H); 43.5 (s, 2 H); 24.8 (s, 6 H); 22.8 (s, 2 H); 20.8 (s, 2 H); 12.7 (s, 2 H); 9.8 (s, 18 H); 9.0 (s,

4340
Helvetica Chimica Acta Vol. 85 (2002)
6 H); 5.8 (s, 6 H); 2.4 (s, 6 H); 1.4 (s, 4 H); À1.2 (s, 6 H); À6.9 (s, 2 H). ESI-MS (THF (abs.), c ˆ 0.1 mg/ml,
‡
spray voltage 2.5 kV, 150 ml/h, capillary voltage 21 V, tube lens offset À10 V): 759 (100, [M ‡ H] , cluster of Ni
isotopes), 444 (12), 424 (14), 410 (11), 352 (59).
In situ Preparation and Polymerization by the Catalytic Formulation from (CH3OCH2CH2OCH3)NiCl2 (4).
To a soln. of 20 mg (53.4 mmol) Na salt of the salicylaldimine in 10 ml of THF, 11.7 mg (1 equiv.) NiII chloride
DME complex were added. The resulting soln. was stirred for 20 min at r.t. and 305 ml (1 equiv.) of 0.175m soln.
of MeLi in THF were added. After stirring at an ethylene pressure of 4 bar for 30 min, the soln. was worked up
as usual. The yield of polyethylene was 1.8 g.
Preparation of Grubbs Catalyst 5. Catalyst 5 was prepared according to the published procedure [3].
Spectroscopic data agreed well with published values.
Results. Three compounds, representative of a general structural motif, were
prepared. The m-OH bridged complex, 1, is an uncommon structure for NiII, although m-
alkoxy-bridged complexes are known. Characterization of 1 by X-ray crystallography
on single crystals shows an approximately C2-symmetric structure (Fig. 1). Particularly
1
diagnostic is the chemical shift of the OH H-atom in the H-NMR spectrum. The H-
atom lies in the shielding region 2,6-diisopropylphenyl group, directly over the ring,
giving rise to the very large upfield shift to À8.3 ppm.
Fig. 1. X-Ray structure of 1, showing non-H-atoms. The m-bridging OH groupH-atoms are situated directly
above the aromatic ring of the Schiff base, leading to the characteristic upfield shift for those H-atoms in the
1H-NMR. A detailed description of the structure is available from the Cambridge Crystallographic Data Centre
(CCDC), deposition No. CCDC-198689.
Interestingly, 1 appears to be an unusually favorable structure, being produced by
nearly any synthetic route in which the components, i.e., NiII and salicylaldimine ligand,
are reacted in the presence of traces of H2O. The structure was assigned by X-ray
diffraction of single crystals. Compound 1 is a brown complex that is stable under
normal handling conditions in ambient air. It can be stored as a solid in a sealed vessel
for months without noticeable change. Associated b-diketonate-alkoxo-NiII complexes
with presumed structures similar to that of 1 have been reported by Baranwal and
Mehrotra [4]. Analogous dinuclear NiII [5] and PdII [6] complexes with m-aryloxy
bridging ligands have also been published. Perhaps most relevant are two structures, 6

Helvetica Chimica Acta Vol. 85 (2002)
4341
and 7, that have been previously reported [7]. The structures were originally assigned
on the basis of spectroscopic and magnetic-susceptibility data; 7 has subsequently been
crystallographically characterized.
Both binuclear Ni complexes show the m-OH groups. The Ni-centers in 6 show
tetrahedral coordination. An X-ray structure of 7 shows each Ni to be square planar.
More importantly with respect to the structure of 1, the complexes 6 and 7 indicate that
the m-OH groups are neither so basic that they are cleaved by relatively acidic groups
(in 6), nor so acidic that they themselves protonate off Ni alkyls (in 7).
Complexes 2 and 3 are related in that they are salicylaldiminato-NiII complexes in
which the remaining two coordination sites are occupied by a bidentate monoanionic
1
ligand. Characterization by H-NMR, ESI-MS, and IR support the structural assign-
ments. Complex 2 is formed by treating anhydrous [Ni(acac)2] with 1 equiv. of the
salicylaldimine ligand in mesitylene, followed by azeotropic distillation to remove the
more-volatile acetylacetone. Similarly, 3 is prepared by combination of the anhydrous
NiBr2 ¥ DME complex with 2 equiv. of the Li salt of the salicylaldimine ligand. Both 2
and 3 are stable in air in the solid state under normal handling conditions. Exposure to
traces of H2O in solution leads to slow conversion of 2 or 3 to 1, as can be seen by
1H-NMR.
All three complexes can be activated with respect to ethylene polymerization by
treating dilute solutions of 1 3 in toluene with near-stoichiometric amounts of BuLi,
MAO, BH3 ¥ THF, or MeLi in the presence of 1 75 bars of ethylene. Catalyst
productivity was found to be broadly similar for all three complexes with all of the
activators. As a control experiment, the catalyst productivity, as determined by total
polyethylene formed after workup, was compared to the productivity computed by
integrating the isobaric rate of ethylene consumption. The resulting graph (Fig. 2)
shows that the two are linearly correlated, making either measurement a suitable
indicator of total productivity.
The experiment validates the utility of the isobaric rate of ethylene consumption as
a measure of instantaneous catalyst activity at a given point of the reaction. The last two
activators listed above introduce Lewis bases that normally poison single-site catalysts
the borane includes 1 equiv. of THF, and MeLi comes as a dilute solution in THF due
to its insolubility in hydrocarbon solvents. Neutral NiII complexes of Grubbs and co-
workers [3] show broad tolerance for ethers, so the activation of 1 3 was attempted in
THF or 1,4-dioxane as solvent in the presence of ethylene. Not only did polymerization
proceed, but the productivity in the coordinating solvents was hardly different from
that in toluene. While one might expect some tolerance for heteroatoms in a late
transition metal SSC, insensitivity to strongly coordinating solvents is rare.

4342
Helvetica Chimica Acta Vol. 85 (2002)
Fig. 2. Correlation of the integrated isobaric ethylene consumption against isolated polyethylene after workup for
polymerizations in the −small× reactor. The linear relationshipindicates that both measurements may be used to
characterize catalyst activity. It also indicates that the ethylene is polymerized rather than oligomerized because
the polyethylene yield was measured after removal of all volatile components.
Interestingly, the activity of the catalyst depends strongly on the amount of the
alkylating or hydride-transfer agent. In general, the maximum activity occurs for
slightly more than 1 equiv. of alkylating agent per nickel (Fig. 3).
Fig. 3. Dependence of catalyst productivity on the mol-equiv. per Ni of the activator for the case of BuLi and
MAO. The solid lines represent polynomial fits to the data points. The concentration of BuLi was standardized
by titration. MAO was assumed to have its nominal concentration as delivered.

Helvetica Chimica Acta Vol. 85 (2002)
4343
With much more than 1 equiv., the activity rapidly goes to zero, as shown in Fig. 3.
For MAO, productivity of the catalyst shows a similar maximum, although the curve is
shifted to a higher number of equiv. In the case of MAO, it cannot be excluded that all
of the Me equiv. are not equally available; nevertheless, the general conclusion that a
large excess of alkylating (or hydride transfer) agent deactivates the catalyst is clearly
demonstrated. Comparable curves were obtained for all activators and catalysts 2 and
3. In all cases, near-stoichiometric amounts of activators gave maximum productivity;
more than a few equiv. resulted in no activity. One should note that complex 3, termed
×the 2 :1 complex× in the study of Grubbs and co-workers, was reported to be inactive in
either oligomerization or polymerization of ethylene when activated by MAO.
Presumably, the customary large molar excess of MAO was employed, which we have
confirmed to result in deactivation of the catalyst.
The catalytic formulation produced by activation of 1, and, to a lesser extent, 2 and
3, is chemically robust and long-lived. Fig. 4 shows the isobaric ethylene consumption
by activated 1 under 4 bar of ethylene at 258 and 408, as well as under adiabatic
conditions where the reactor temperature was allowed to rise due to the reaction
exothermicity. As is clear from the data, the activated catalyst is long-lived in that there
is no deactivation process, at least on the timescale of 1 h. The drop in activity that is
typically observed in the −small× reactors (10 20 ml of solvent) comes because the
solution becomes too viscous for efficient stirring, slowing and, eventually, preventing
Fig. 4. Isobaric ethylene consumption for polymerization of ethylene in the −small× reactor at 258, 408, and under
adiabatic conditions. In the latter case, the temperature typically rose from room temperature to over 508.
Standard conditions were 20 mg of catalyst 1 in toluene, activated with 1.2 equiv. (per Ni) BuLi and 4 bar
ethylene.

4344
Helvetica Chimica Acta Vol. 85 (2002)
transfer of ethylene from the gas-phase into solution. As further evidence for the
chemical stability of the catalyst, we find that the polymerization in the −small× reactor
could be interrupted by replacement of ethylene by CO2 (three evacuation refilling
equilibration cycles) for a period of time, typically ca. 1 h. Upon replacement of CO2
with ethylene, polymerization recommenced with no break in the activity of the catalyst
as measured by isobaric ethylene consumption. Furthermore, if the ethylene source is
closed off from the −small× reactor so that all ethylene in the reactor is consumed after
polymerization has commenced, the resulting solution will continue polymerization
upon re-introduction of ethylene after interruptions of up to ca. 1 h.
The pressure dependence of the polymerization reaction is easily seen in Fig. 5.
With the isobaric ethylene consumption instead of final polyethylene yield after
workup, the instantaneous activity can be ascertained. An approximately linear
dependence of activity on ethylene pressure was observed. The pressure dependence as
well as the catalyst lifetime could be further verified by bulk polymerization
experiments in the −large× reactor (50 100 ml of solvent). The Table summarizes
catalyst productivities in both reactors under a variety of conditions. Entries 10 15
were performed with the −large× reactor. The increased productivity in Entry 10
appeared to be consistent with the −small× reactor results when the higher temperature
and ethylene pressure were considered. With successive decreases in the amount of
catalyst, e.g., Entries 11 13, it becomes apparent that the TON of 22,000 represents the
chemically-limited productivity for 1 with MeLi activation even as ethylene pressure is
further increased. In particular, ethylene consumption ceased after 2 h in Entry 12
without solidification of the reaction solution (which happened, e.g., in Entry 10),
indicating that catalyst deactivation had finally become limiting.
Fig. 5. Pressure dependence of the isobaric ethylene consumption rate at two pressures under reaction conditions
as in Fig. 4 for a polymerization at room temperature

Helvetica Chimica Acta Vol. 85 (2002)
4345
Table. Polyethylene (PE) Production in the −Small× Reactor (Entries 1 9) and the −Large× Reactor (Entries 10
15) by Catalysts 1 4. The amount of activator is given in mol-equiv. per Ni. Turnover number (TON) is defined
as mol ethylene converted to PE divided by mol nickel. Note that catalyst 1 contains 2 mol Ni per mol catalyst.
Entry Catalyst ([mg]) Activator
Solvent ([ml]) P [bar] T [8] t [h]
PE [g] TON
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
1 (20)
1 (10)
2 (26)
3 (20)
1 (10)
1 (10)
2 (20)
1 (15)
4 (12)
1 (20)
1 (10)
1 (5)
BuLi (1.2 equiv.) toluene (15)
MAO (6.0 equiv.) toluene (15)
BuLi (1.9 equiv.) toluene (20)
MeLi (1.0 equiv.) THF (12)
MeLi (1.2 equiv.) THF (10)
MeLi (1.2 equiv.) dioxane (10)
MeLi (1.0 equiv.) dioxane (10)
4
4
3.5
4
4
4
4
4
4
25
40
25
25
25
25
25
25
25
55
40
40
50
40
40
2
1
2.4
3.3
4.5
2.3
1.0
1.0
3.6
2.2
1.8
1,900
5,100
3,100
3,200
1,600
1,600
3,200
2,300
1,200
22,000
22,000
22,000
22,000
63,000
104,000
1.5
4
0.5
0.5
2
1.5
0.5
2.5
2
BH3 (1.0 equiv.)
THF (10)
MeLi (1.0 equiv.) THF (10)
MeLi (1.2 equiv.) toluene (70)
MeLi (1.2 equiv.) toluene (50)
MeLi (1.2 equiv.) toluene (50)
MeLi (1.2 equiv.) toluene (50)
75
28
14
7
40
40
75
75
75
2
2.5
1 (5)
1 (5)
1 (2.3)
7
MAO (5 equiv.)
MAO (5 equiv.)
toluene (56)
toluene (45)
overnight 20
overnight 16
The final lines, Entries 14 and 15, show the same experiment, but with MAO instead
of MeLi as activator. MAO had, in −small× reactor trials, generally given higher
productivity than MeLi. The reactor was left overnight in these cases because the
polymerization continued to run for several hours. For both Entries 14 and 15, the
increased TON of 63,000 or 104,000, respectively, represent lower bounds because the
reactor was completely filled with a solid mass in each run, blocking the ethylene inlet.
It should be noted that the polymerizations under higher pressures occur in a regime
where ethylene is supercritical2), which affects the TON in that the viscosity of the
medium is reduced and diffusion is enhanced3).
The properties of the related catalyst 5, from Grubbs and co-workers, have been
reported in several publications [3] and a patent [10]. Similar systems 8, from Mecking
and co-workers [11], have also been reported. These catalysts show TON values that
range from a few hundred to a few thousand for 5, and similar levels for 8. The
particular Grubbs catalyst with almost the same salicylaldimine ligand as 1 shows a
TON of 1600. Although 1 3 resemble 5 and 8, the TON for 1 in toluene, activated with
MAO, exceeds 105. It is also clear that the TON can be further improved by alterations
in either the catalyst structure or process design; the filling of the reactor limited TON
in Entries 14 and 15. The very high productivity for 1 begins to reach the level for the
current generation of high-activity Ziegler Natta catalysts in which the TON values of
ca. 105 106 mean that catalyst residues are low enough to simply leave in the polymer.
The substantially higher TON for 1 3 derives not from higher turnover frequency
(TOF), but rather from a significantly longer catalyst lifetime with respect to
deactivation.
2
)
The critical temperature and pressure for ethylene is 98 and 50 bar, respectively [8]. Polymerization
throughout the reactor at higher temperature and pressure occurs because the supercritical ethylene fills
the entire volume of the reactor. However, as in the case of polymerization in toluene solution, the
reaction ceases, when the polymer concentration becomes so high as to inhibit transport.
For a review of polymerization in supercritical media, see [9].
3
)

4346
Helvetica Chimica Acta Vol. 85 (2002)
Given the structural similarity between 1 3 and 5, the increased lifetime and the
concomitantly increased TON must be due to the presence or absence of some other
component. As a control experiment, 5 was prepared and tested as shown in Fig. 6.
When activated by 1.5 equiv [Ni(cod)2] in THF, 5 began to consume ethylene after an
induction period of 12 min. In this particular example, ethylene consumption continued
for ca. 25 min, and then ceased. The same amount of 5, activated identically, but with
the addition of 10 equiv. of LiClO4, consumed ethylene at a significantly higher rate,
but, more importantly, continued to consume ethylene even after 1 h when viscosity
problems began, as usual, to impact catalyst productivity. While full quantitative
characterization is not yet complete, one can see qualitatively that the activated species
from 1 is longer-lived than those from 2 or 3, and that the lifetime is generally longer in
toluene solvent than in THF or dioxane. While there may be other factors that may also
be important, the activated forms of 1 3 likely owe their increased lifetimes to the
presence of Lewis acidic metal cations in low molar excess.
Fig. 6. Isobaric ethylene consumption by a solution of 10 mg (13.3 mmol) 5 and 6 mg (1.5 equiv.) [Ni(cod)2] in
&
^
10 ml of THF at an ethylene pressure of 4 bar, with ( ) and without ( ) the addition of 10 equiv LiClO4. Addition
of [Ni(cod)2] as a phosphine scavenger defines the zero in time.
The polyethylene produced in Entries 13 and 14 was characterized by GPC and
viscometry4). For the two samples, the molecular weight and polydispersity, Mw and
Mw/Mn, were 14,000 and 1.91, and 24,000 and 1.93, respectively. The catalyst of Grubbs
with almost the same ligand gave 11,400 and 1.8 for Mw and Mw/Mn. Interestingly, while
the catalyst of Grubbs produced polyethylene with 55 branches per 1000 C-atoms, the
polyethylene from Entries 13 and 14 was completely linear with no detectable
branching.
4
)
Polymer characterization was done at Bayer AG, Leverkusen, with a Waters 150C GPC (o-
dichlorobenzene, 1408) by differential refractometric and viscometric detection. Branching was
determined by comparison of Mark-Houwink parameters to suitable reference samples.

Helvetica Chimica Acta Vol. 85 (2002)
4347
Given that 1 3 are simple coordination compounds that are subsequently activated
with near-stoichiometric alkylating or hydride-transfer agents, one may question the
necessity of prior isolation of the catalysts at all. Entry 9 in the Table shows the results
for the in situ preparation of a catalytic formulation by reaction of the DME complex of
NiCl2, 4, with the Na salt of the salicylaldimine in THF, followed directly by MeLi
activation with no isolation or purification of intermediates. The productivity of the
catalyst is within a factor of 2 3 of that for the isolated and purified catalysts under
comparable conditions. The successful in situ catalyst preparation and activation may
present interesting practical possibilities, but also opens the way to high-throughput
screening of pooled catalyst libraries [12].
Discussion. Late-transition-metal SSCs have recently become the focus of intense
investigation due to the possibility that one may be able to circumvent some of the
diadvantages of the early-transition-metal catalysts, notably, metallocenes. The return,
especially to Ni, is ironic, given that the coordination oligomerization and polymer-
ization of ethylene was originally discovered a half-century ago with adventitious Ni as
the catalytic metal [13]. Several of the first- and second-row late transition metals, as
well as a large variety of ligands, have now been shown to yield moderate-to-very good
catalysts for olefin polymerization5).
Importantly, Brookhart and co-workers [15][16] have reported extensive mecha-
nistic studies of the NiII diimine catalysts, 9, in which it was established that the olefin p-
complex, directly observed by low-temperature NMR, is the resting state in the
catalytic cycle, i.e., the turnover-limiting stepis insertion of the olefin into the
metalÀalkyl bond. Similar catalysts have been reported by Gibson and co-workers
[17]. Also, Grubbs and co-workers [3][10], building on SHOP-like [18] systems from
Cavell and co-workers [19], Keim and co-workers [20] and Ostoja-Starzewski and Witte
[21], have shown that neutral NiII complexes can effectively polymerize ethylene with
substantially improved tolerance for heteroatoms. A thorough review of NiII systems
has been published [22]. The tolerance of NiII systems is sufficient for their use in
toluene/H2O emulsions, as has been shown by Mecking and co-workers [11], for the
production of polyolefin latexes.
5
)
A review of new catalysts and ligands is found in [14].

4348
Helvetica Chimica Acta Vol. 85 (2002)
Looking at the minimum mechanism for ethylene polymerization according to the
Cossee Arlman model [23], one sees that, aside from chain transfer, there are two
fundamental steps in the catalytic cycle: ethylene p-coordination and insertion
(Scheme 1). If insertion is turnover-limiting, as is observed for the Brookhart catalysts,
then the polymerization rate is necessarily zero-order in ethylene, as is indeed observed
for those systems. Tolerance for coordinating solvent is expected, though, because there
is no competition for a coordination site by solvent at the turnover-limiting transition
state. On the other hand, the turnover-limiting stepin metallocene-catalyzed polymer-
ization of ethylene is p-coordination of the incoming olefin; the resting state of the
catalyst is computed to be the metal alkyl with the fourth coordination site on the metal
filled by an agostic CÀH bond [24]. Accordingly, the kinetics of polymerization are
first-order in ethylene, which is desirable in practice, but any strongly coordinating
solvent can occupy the fourth site and inhibit turnover.
Scheme 1
One expects, therefore, that two desirable features, first-order kinetics and
tolerance for heteroatoms, should be mutually exclusive. Given that complexes 1 3
show both first-order kinetics with respect to ethylene and tolerance for heteroatoms,
the mechanistic picture must necessarily be more complicated than that shown in
Scheme 1.
Complex 5 of Grubbs and co-workers [3], and several other similar complexes
[11][25], are the closest analogs to 1 3. All of the complexes are neutral in their active
form, and are based on NiII; 1 3 and 5 have moreover the same salicylaldimine ligand.
Nevertheless, there are some differences in chemical behavior, which shed light on the
mechanism by which 1 3 operate. The increase in both activity and lifetime, when
LiClO4 is added to a polymerization catalyzed by 5, suggests that the complexation of
the Lewis-acidic alkali-metal cation, presumably to two O-atoms of OH or alkoxy
moieties may be largely responsible for the significant improvement in catalyst

Helvetica Chimica Acta Vol. 85 (2002)
4349
performance. This observation supports a proposed structure for the resting state in the
catalytic cycle for activated forms of 1 3, shown in Scheme 2 for which the additional
ligands (Lig) on the second metal (M ˆ Li, Al, B) are a (salicylaldiminato)Ni moiety,
acetylacetonato, or salicylaminato ligands, or solvent, or a combination thereof.
Scheme 2
In contrast to the tetrahedral d0 metallocenes, the resting state in Scheme 2 is a
square planar d8 complex, which most likely performs ligand exchange predominantly
according to an associative mechanism [26] involving coordination of the incoming
ethylene to form a five-coordinate intermediate, followed by departure of one of the
original four ligands. The pre-equilibrium in the proposed catalytic cycle the first step
in the associative ligand exchange yields overall first-order kinetics with respect to
ethylene, and the assignment of the turnover-limiting stepto the departure of an oxy
ligand is consistent with the acceleration of turnover by the presence of Li or oxophilic
Lewis acids. Moreover, unless a nucleophile, including solvent, were to coordinate
more strongly than does the entropically favored Li-bound aryloxy moiety, that
nucleophile should exercise no adverse effect on the turnover rate. The mechanism is
similar to the associative pathway proposed by Cavell and co-workers [19]. For the rate
acceleration in the polymerization reactions of 5 in THF with LiClO4, coordinated
THF occupies at least some of the −Lig× positions in Scheme 2. Heterobimetallic
complexes with m-bridging heteroatom ligands have been reported in several contexts,
providing precedent for the structural hypothesis in Scheme 2. Uhlig and co-workers
have characterized several FeII, CoII, and NiII acac complexes in which a second cation
is bound by m-bridging oxy ligands [27]. Even more relevant are examples by Floriani
and co-workers [28], and others [29] in which aryloxy ligands in a late transition metal

4350
Helvetica Chimica Acta Vol. 85 (2002)
salicylaldimato complex have bound a second electrophilic metal. Overall, the
postulated active species is chemically reasonable, and provides an explanation for
all of the experimental observations in this catalyst system.
The last issue concerns the characteristic of 1 that most strongly contributes to its
practical potential. The TON of ca. 105 achieved by 1 brings it into the range of
productivity needed for commercial manufacture of polyolefins. Given that TON for
similar complexes 5 and 8 are lower by 2 3 orders of magnitude Brookhart and co-
workers also mention [22] that none of his Ni-based SSCs shows TON over a few
thousand one is compelled to speculate as to the role of additional components in
producing this dramatic effect. While the experiments do not yet support an
unambiguous explanation for catalyst deactivation in 5 and 8, and, conversely, the
suppression of catalyst deactivation in 1, one can make a suggestion. The associative
mechanism proposed above works with a switching between four-coordinate and five-
coordinate intermediates. An alternative dissociative mechanism would work via three-
coordinate and four-coordinate structures. Even if an associative mechanism were to be
predominant, a small contribution by a dissociative mechanism would mean that three-
coordinate Ni alkyls would be present, at least to a small extent. If these intermediates
were to undergo b-hydride elimination to form Ni hydrides, as expected for three-
coordinate Ni-alkyls, then the presence of basic reagents, present as additives in 5 and 8,
could lead to deprotonation effectively reduction with concomitant deactivation.
Bidentate complexation of a Lewis acid may largely shut down the alternative
dissociative mechanism for polymerization, leading to longer catalyst lifetimes. The
observed stability of activated 1, even when no ethylene is present, supports the
hypothesis; other NiII catalysts typically deactivated irreversibly and formed black
precipitates when starved of ethylene. If b-hydride elimination is shut down, chain-
transfer could still occur for activated 1 by a b-hydride transfer mechanism without the
formation of hydrides. Because hydride intermediates have been suggested to play an
integral role in the chain-walking mechanism for branch formation [22], one would
expect 1 to form highly linear polyethylene as a necessary consequence of the
mechanistic hypothesis for the increased catalyst lifetime. The observation of linear
polyethylene with no detectable chain-branching provides additional support for the
postulated mechanism. Further work is underway to clarify the mechanistic aspects of
this new catalyst system.
Lastly, 5 and 8 have been shown to copolymerize ethylene with polar monomers as
well as norbornene6). One expects that 1 should do so as well. These studies are also
underway.
Conclusions. A family of NiII coordination complexes are reported for which
activation by stoichiometric amounts of ordinary metal alkyls produce efficient, highly
productive, polyolefin catalysts. The kinetic order and an unusually high tolerance for
heteroatoms leads to a proposed catalytic cycle in which the key species are
heteronuclear complexes involving NiII and a second electrophilic metal cation. The
second metal center, with associated bridging ligands functions to suppress catalyst
6
)
Preliminary results show that norbornene is, in fact, incorporated.

Helvetica Chimica Acta Vol. 85 (2002)
4351
deactivation, leading to TON exceeding 105. With further optimization, higher
productivities are likely to be achievable.
The work was supported by Thales Technologies AG, Z¸rich, Switzerland, and the Kommission f¸r
Technologie und Innovation, an agency within the Swiss Bundesamt f¸r Bildung und Technologie. Polymer
characterization by Dr. R. Holm and Dr. J. Schweer (Bayer AG, Leverkusen), as well as helpful discussions with
Prof. P. Pregosin (ETH-Z¸rich) are acknowledged.
REFERENCES
[1] −Ziegler Catalysts×, Eds. G. Fink, R. M¸lhaupt, H. H. Brintzinger, Springer-Verlag, Berlin, 1995.
[2] N. F. Brockmeier, −Polypropylene Reinvented Cost of Using Metallocene Catalysts×, in −Metallocene-
Catalyzed Polymers×, Eds. G. M. Benedikt, B. L. Goodall, Plastics Design Library, New York, 1998, p. 11 20.
[3] C. Wang, S. Friedrich, T. R. Younkin, R. T. Li, R. H. Grubbs, D. A. Bansleben, M. W. Day, Organometallics
1998, 17, 3149; T. R. Younkin, E. F. Connor, J. I. Henderson, S. K. Friedrich, R. H. Grubbs, D. A.
Bansleben, Science 2000, 287, 460.
[4] B. P. Baranwal, R. C. Mehrotra, Transition Met. Chem. 1978, 3, 220.
[5] H. Bˆnnemann, J.-D. Jentsch, Appl. Organomet. Chem. 1993, 7, 553.
[6] J. Campora, M. L. Reyes, T. Hackl, A. Monge, C. Ruiz, Organometallics 2000, 19, 2950.
[7] H.-F. Klein, H. H. Karsch, Chem. Ber. 1973, 106, 1433; D. I. Maharaj, L. A. Hall, Polyhedron 1988, 7, 2155;
E. Carmona, J. M. Marin, P. Palma, M. Paneque, M. L. Poveda, Inorg. Chem. 1989, 28, 1895.
[8] −CRC Handbook of Chemistry and Physics×, 75th edn., 1994.
[9] K. M. Scholsky, J. Supercrit. Fluids 1993, 6, 103.
[10] D. A. Bansleben, S. K. Friedrich, T. R. Younkin, R. H. Grubbs, C. Wang, R. T. Li, PCT Patent Application
WO 98/42664, October 1, 1998.
[11] A. Held, F. M. Bauers, S. Mecking, Chem. Commun. 2000, 301; F. M. Bauers, S. Mecking, Macromolecules
2001, 34, 1165; F. M. Bauers, S. Mecking, Angew. Chem., Int. Ed. 2001, 40, 3020.
[12] C. Hinderling, P. Chen, Angew. Chem., Int. Ed. 1999, 38, 2253; C. Hinderling, P. Chen, Int. J. Mass
Spectrom. Ion Proc. 2000, 195/196, 377; C. Hinderling, C. Adlhart, P. Chen, Chimia 2000, 54, 232; C.
Adlhart, P. Chen, Helv. Chim. Acta 2000, 83, 2192; M. A. O. Volland, C. Adlhart, C. A. Kiener, P. Chen, P.
Hofmann, Chem. Eur. J. 2001, 7, 4621.
[13] K. Ziegler, H.-G. Gellert, E. Holzkamp, G. Wilke, Brennstoffchemie 1954, 35, 321.
[14] G. J. P. Britovsek, V. C. Gibson, D. F. Wass, Angew. Chem., Int. Ed. 1999, 38, 428.
[15] L. K. Johnson, C. M. Killian, S. D. Arthur, J. Feldman, E. F. McCord, S. J. McLain, K. Kreutzer, A. M. A.
Bennett, E. B. Coughlin, S. D. Ittel, A. Parthasarthy, D. J. Tempel, M. S. Brookhart, PCT Patent
Application WO 96/23010, August 1, 1996; S. J. McLain, A. M. A. Bennett, E. B. Coughlin, D. S. Donald,
L. T. J. Nelson, A. Parthasarathy, X. Shen, W. Tam, Y. Wang, PCT Patent Application WO 97/02298,
January 23, 1997; L. K. Johnson, A. M. A. Bennett, S. D. Ittel, L. Wang, E. Hauptman, R. D. Simpson, J.
Feldman, E. B. Coughlin, PCT Patent Application WO 98/30609, July 16, 1998.
[16] L. K. Johnson, C. M. Killian, M. Brookhart, J. Am. Chem. Soc. 1995, 117, 6414.; L. K. Johnson, S. Mecking,
M. Brookhart, J. Am. Chem. Soc. 1996, 118, 267; C. M. Killian, D. J. Tempel, L. K. Johnson, M. Brookhart,
J. Am. Chem. Soc. 1996, 118, 11664; B. L. Small, M. Brookhart, A. M. A. Bennett, J. Am. Chem. Soc. 1998,
120, 4049; B. L. Small, M. Brookhart, Organometallics 1999, 32, 2120; S. A. Svejda, L. K. Johnson, M.
Brookhart, J. Am. Chem. Soc. 1999, 121, 10634.
[17] G. J. P. Britovsek, V. C. Gibson, B. S. Kimberley, P. J. Maddox, S. J. McTavish, G. A. Solan, A. J. P. White,
D. J. Williams, J. Chem. Soc., Chem. Commun. 1998, 849.
[18] W. Keim, F. H. Kowalt, R. Goddard, C. Kr¸ger, Angew. Chem., Int. Ed. 1978, 17, 466.
[19] S. Y. Desjardins, K. J. Cavell, H. Jin, B. W. Skelton, A. H. White, J. Organomet. Chem. 1996, 515, 233.
[20] S. Y. Desjardins, K. J. Cavell, J. L. Hoare, B. W. Skelton, A. N. Sobolev, A. H. White, W. Keim, J.
Organomet. Chem. 1997, 544, 163.
[21] K. A. Ostoja-Starzewski, J. Witte, Angew. Chem. 1987, 99, 76.
[22] S. D. Ittel, L. K. Johnson, M. Brookhart, Chem. Rev. 2000, 100, 1169.
[23] P. Cossee, J. Catal. 1964, 3, 80; E. J. Arlman, P. Cossee, J. Catal. 1964, 3, 99.
[24] J. C. W. Lohrenz, T. K. Woo, T. Ziegler, J. Am. Chem. Soc. 1995, 117, 12793; T. K. Woo, P. M. Margl, J. C. W.
Lohrenz, P. E. Blochl, T. Ziegler, J. Am. Chem. Soc. 1996, 118, 13021.

4352
Helvetica Chimica Acta Vol. 85 (2002)
[25] M. D. Fryzuk, X. Gao, S. J. Rettig, Can. J. Chem. 1995, 73, 1175; A. Tomov, J. P. Broyer, R. Spitz, Macromol.
Symp. 2000, 150, 53; J. Heinicke, M. He, A. Dal, H.-F. Klein, O. Hetche, W. Keim, U. Flˆrke, H.-J. Haupt,
Eur. J. Inorg. Chem. 2000, 431; M. A. Dubois, R. Wang, D. Zargarian, J. Tian, R. Vollmerhaus, Z. Li, S.
Collins, Organometallics 2001, 20, 663.
[26] F. Basolo, R. G. Pearson, Prog. Inorg. Chem. 1964, 4, 381; F. Basolo, R. G. Pearson, −Mechanisms of
Inorganic Reactions×, Wiley, New York, 1968.
[27] M. Dˆring, E. Uhlig, L. Dahlenburg, Z. Anorg. Allg. Chem. 1989, 578, 58; M. Dˆring, H. Gˆrls, E. Uhlig, K.
Brodersen, L. Dahlenburg, A. Wolski, Z. Anorg. Allg. Chem. 1992, 614, 65.
[28] E. Solari, F. Corazza, C. Floriani, A. Chiesi-Villa, C. Guastini, J. Chem. Soc., Dalton Trans. 1990, 1345; C.
Floriani, Pure Appl. Chem. 1996, 68, 1; E. Gallo, E. Solari, C. Floriani, A. Chiesi-Villa, C. Rizzoli, Inorg.
Chem. 1997, 36, 2178.
[29] K. Brychcy, K. Dr‰ger, K.-J. Jens, M. Tilset, U. Behrens, Chem. Ber. 1994, 127, 1817.
Received July 4, 2002