Investigations of the Ligand Electronic Effects on α-Diimine Nickel(II) Catalyzed Ethylene Polymerization

The synthesis and characterization of a series of dibenzhydryl-based α-diimine Ni(II) complexes bearing a range of electron-donating or -withdrawing groups are described. Polymerization with ethylene is investigated in detail, involving the activator effect, influence of polymerization conditions on catalyst activity, thermal stability, polymer molecular weight and melting point. All of these Ni(II) complexes show great activity (up to 6 × 106 g of PE (mol of Ni)−1·h−1), exceptional thermal stability (stable at up to 100 °C) and generate polyethylene with very high molecular weight (Mn up to 1.6 × 106) and very narrow molecular weight distribution. In the dibromo Ni(II) system, the electronic perturbations exhibit little variation on the ethylene polymerization. In the Ni(acac) system, dramatic ligand electronic effects are observed in terms of catalytic activity and polyethylene molecular weight.


Introduction
The initial reports by Brookhart and coworkers showed that complexes of Ni(II) and Pd(II) bearing sterically hindered α-diimine ligands could generate high-molecular-weight polymer with high catalytic activity in ethylene polymerization and incorporate polar comonomers into polyolefins for the Pd(II) catalyst [1,2]. A key insight from these early studies is that bulky ortho-aryl substituents in the α-diimine ligand could retard the associative chain transfer process by sterically blocking the access of monomers to the metal axial positions [1,3], which is essential to achieve high polymerization activity and high polymer molecular weight. Additionally, the high catalytic activities can be achieved by the destabilization of the ground state species from steric effect. The degree of branching and branching distribution, which is regulated by the competition between chain propagation and chain walking, can be changed through the tailoring of ligand structures. Since Brookhart's seminal discovery, tremendous efforts have been made to explore new late-transition metal catalysts for olefin polymerization and copolymerization with polar monomers [3][4][5][6][7][8][9][10][11][12][13][14][15]. Despite these exciting features, one drawback of these catalysts is their low thermal stability, which greatly hinders their potential industrial applications (70-110˝C) [16]. It has been reported that α-diimine Pd(II) and Ni(II) catalysts undergo rapid decomposition over 60˝C because of increased associative chain transfer, C-H activation of the ligand, potential decomposition via in situ-generated metal hydride species and bis-ligation during the polymerization process [17][18][19][20].
Modifications of the N-aryl substituents and the ligand backbone have been very important to improve the thermal stability of α-diimine Pd(II) and Ni(II) catalysts. For example, Ionkin et al. reported reported that o-benzofuran-substituted α-diimine Ni(II) complexes (Scheme 1A,B) could maintain high activity even at 150 °C and generate high-molecular-weight polyethylene at 70 °C [21]. Subsequently, some cyclophane-based catalysts were shown to produce polyethylene with high activities and enhanced thermal stability [22][23][24][25][26]. In addition, catalysts bearing camphorquinonederived ligands displayed moderate stability at up to 80 °C [20,27]. Recently, dibenzhyldryl-derived ligand frameworks were used to generate highly stable catalysts [28][29][30]. For example, Long et al. reported the studies of α-diimine Ni(II) catalysts (Scheme 1C) containing dibenzhydryl moiety, which maintained high activities at temperatures up to 100 °C [31,32]. Our group showed that α-diimine Pd(II) complexes (Scheme 1D) bearing the dibenzhydryl moiety displayed high thermal stability and high activity in ethylene (co)polymerization, producing semicrystalline polyethylene and ethylene/methyl acrylate copolymer with high molecular weight and low branching density [33]. We also investigated the ligand electronic effect in that system. It was demonstrated that the ligand electronic effect played an important role on the polymerization properties in α-diimine catalyst systems. Guan et al. reported that higher polyethylene molecular weight and more linear topology were obtained with α-diimine Pd(II) catalysts (Scheme 1E) bearing electron-donating substituents. Catalysts bearing electron-donating substituents produced copolymers with higher methyl acrylate incorporation in ethylene-methyl acrylate copolymerization. Electron-donating substituents also led to more stable catalysts [34,35]. The same group studied the ligand electronic effect on a family of cyclophane-based α-diimine Pd(II) catalysts (Scheme 1F) [36]. The Pd(II) catalysts with electron-withdrawing substituents generated polymers with higher molecular weight than the electron-donating analogues, which was different from the previously studied acyclic α-diimine Pd(II) catalysts. Interestingly, there have been very few studies concerning the ligand electronic effects on α-diimine Ni(II) catalysts. Guan et al. studied some cyclophane-based α-diimine Ni(II) catalysts bearing different electron-donating or -withdrawing substituents, which showed very similar properties in ethylene polymerization (Scheme 1G) [36]. In this work, we synthesized and characterized a series of α-diimine Ni(II) complexes 1a-1c and 2a-2d It was demonstrated that the ligand electronic effect played an important role on the polymerization properties in α-diimine catalyst systems. Guan et al. reported that higher polyethylene molecular weight and more linear topology were obtained with α-diimine Pd(II) catalysts (Scheme 1E) bearing electron-donating substituents. Catalysts bearing electron-donating substituents produced copolymers with higher methyl acrylate incorporation in ethylene-methyl acrylate copolymerization. Electron-donating substituents also led to more stable catalysts [34,35]. The same group studied the ligand electronic effect on a family of cyclophane-based α-diimine Pd(II) catalysts (Scheme 1F) [36]. The Pd(II) catalysts with electron-withdrawing substituents generated polymers with higher molecular weight than the electron-donating analogues, which was different from the previously studied acyclic α-diimine Pd(II) catalysts. Interestingly, there have been very few studies concerning the ligand electronic effects on α-diimine Ni(II) catalysts. Guan et al. studied some cyclophane-based α-diimine Ni(II) catalysts bearing different electron-donating or -withdrawing substituents, which showed very similar properties in ethylene polymerization (Scheme 1G) [36]. In this work, we synthesized and characterized a series of α-diimine Ni(II) complexes 1a-1c and 2a-2d bearing the dibenzhydryl moiety and a range of electron withdrawing and donating substituents. The ligand electronic effects upon the catalytic activity and thermal stability, along with analysis of polymer molecular weight and melting points, were investigated.

General Information
All manipulations of air-and moisture-sensitive materials were performed under a dry Nitrogen atmosphere using standard Schlenk techniques or in a glove-box. Deuterated solvents for NMR were dried and distilled prior to use. Nitrogen was purified by passing through a MnO oxygen-removal column and an activated 4 Å molecular sieve column. Ethylene was purified by passing through an Agilent oxygen/moisture trap. The 1 H and 13 C NMR spectra were recorded on a Bruker Ascend™ 400 spectrometer (Bruker, Karlsruhe, Germany) at ambient temperature unless otherwise stated. The chemical shifts of the 1 H and 13 C NMR spectra (Bruker, Karlsruhe, Germany) were referenced to TMS. The 19 F chemical shifts are reported relative to external CFCl 3 . Coupling constants are in Hz. Elemental analysis was performed by the Analytical Center of the University of Science and Technology of China. Mass spectra were recorded on a P-SIMS-Gly of Bruker Daltonics Inc. (EI+, Bruker Daltonics Inc., Billerica, USA). X-ray Diffraction data were collected at 298(2) K on a Bruker Smart CCD area detector (Bruker, Karlsruhe, Germany) with graphite-monochromated Mo Kα radiation (λ = 0.71073 Å). Gel permeation chromatography (GPC) was carried out at 150˝C by a PL-GPC 220 high-temperature gel permeation chromatography (PL, Shropshire, UK). The 1,2,4-Trichlorobenzene (TCB) was used as solvent at a flow rate of 1.0 mL¨min´1, and the calibration was made using polystyrene standard and was corrected for linear polyethylene by universal calibration using the Mark-Houwink parameters of Rudin: K = 1.75ˆ10´2 cm 3 /g and R = 0.67 for polystyrene and K = 5.90ˆ10´2 cm 3 /g and R = 0.69 for polyethylene. Dichloromethane, toluene, THF, and hexanes were purified by solvent purification systems. The α-diimine ligands were prepared according to reported procedures [33]. All other reagents were purchased from commercial sources and used without purification.

Standard Procedure for the Synthesis of Complexes 1a-1c
All complexes were prepared in a similar manner by the reaction of (DME)NiBr 2 with the corresponding ligand in dichloromethane. A typical synthetic procedure for complexes 1a-1c is as follows: ligand (1.0 mmol) and (DME)NiBr 2 (0.31 g, 1.0 mmol) (DME = 1,2-dimethoxyethane) were stirred in 10 mL of CH 2 Cl 2 overnight at room temperature. The solvent was removed, and the resulting powder was washed with hexane (10 mLˆ2) and dried under vacuum to obtain a brown solid.

General Procedure for Ethylene Polymerization
In a typical experiment, a 350 mL glass thick-walled pressure vessel was charged with required amount of aluminum activator (MAO (methylaluminoxane), AlEt 2 Cl, Al(i-Bu) 3 ), 48 mL toluene and a magnetic stir bar in the glovebox. The pressure vessel was connected to a high pressure polymerization line and the solution was degassed. The vessel was warmed to the desired temperature using an oil bath and allowed to equilibrate for 5 min. Then 1.0 µmol of nickel complex in 2 mL CH 2 Cl 2 was injected into the vessel via syringe. With rapid stirring, the reactor was pressurized and maintained at 9.0 atm of ethylene. After a desired amount of polymerization time, the vessel was vented and the polymer was precipitated in acidified methanol (methanol/HCl = 50/1) and dried under vacuum at 50˝C for 24 h.

Synthesis and Characterization of the Ni(II) Complexes
The ligands were prepared using the literature procedure in high yields without using column chromatography [33]. Nickel dibromo complexes of α-diimine ligands are typical catalytic precursors in olefin polymerization [1]. It should be noted that complex 1b with CH 3 substituent has been studied by Long et al. (See [31]). This complex was prepared and studied in this work in order to obtain accurate comparisons in ethylene polymerization. Nickel dibromo complexes (1a-1c) were prepared by reacting the ligands with 1 equiv. of (DME)NiBr 2 (Scheme 2). However, no conversion was achieved in the case of the electron-poor trifluoromethyl-substituted ligands. These Ni complexes were characterized by elemental analysis and mass spectrometry.
A single crystal of complex 1a was obtained by layering hexane onto the CH 2 Cl 2 solution at room temperature. The molecular structure of 1a was determined by X-ray diffraction analysis (Figure 1). In solid state, the nickel center adopts distorted tetrahedron geometry with a N1-Ni1-N2 angle of 81.1 (2)

Ethylene Polymerization Studies
All of the dibromo Ni(II) complexes (1a-1c) were highly active in ethylene polymerization when activated with organoaluminum activators (Table 1). For all of these complexes, the catalytic activities were increased with polymerization temperatures and the highest activity was observed at 100 °C (Table 1, Figure 2). Their activities are on the level of 10 6

Ethylene Polymerization Studies
All of the dibromo Ni(II) complexes (1a-1c) were highly active in ethylene polymerization when activated with organoaluminum activators (Table 1). For all of these complexes, the catalytic activities were increased with polymerization temperatures and the highest activity was observed at 100 °C (Table 1, Figure 2). Their activities are on the level of 10 6 g of PE (polyethylene) (mol of Ni) −1 ·h −1 at all Figure 1. Molecular structure of complex 1a (thermal ellipsoids are shown at the 30% probability level). Hydrogen atoms have been omitted for clarity. Solvent molecule (CH 2 Cl 2 ) was also omitted.

Ethylene Polymerization Studies
All of the dibromo Ni(II) complexes (1a-1c) were highly active in ethylene polymerization when activated with organoaluminum activators (Table 1). For all of these complexes, the catalytic activities were increased with polymerization temperatures and the highest activity was observed at 100˝C (Table 1, Figure 2). Their activities are on the level of 10 6 g of PE (polyethylene) (mol of Ni)´1¨h´1 at all polymerization temperatures. These complexes represent one of the most active and thermally stable α-diimine Ni(II) catalysts in ethylene polymerization. At 100˝C for complexes 1a-1c, the MAO activator gave higher polymerization activity, generating polyethylene with higher molecular weight and narrower molecular weight distribution than the AlEt 2 Cl activator (Table 1, Entries 4,5,9,10,14 and 15). Therefore, MAO was chosen as the activator in the dibromo Ni(II) system. Under almost all of the polymerization conditions, the molecular weight (M n ) of the generated polyethylene is higher than one million, representing a big advantage of this catalyst system. Moreover, the molecular weight was only slightly decreased at elevated temperatures. For example, the M n was slightly dropped from 164ˆ10 4 at 80˝C to 154ˆ10 4 at 100˝C for complex 1b, and the melting point was decreased from 43.8 to 39.1˝C (Table 1, Entries 8 and 9). These results suggest that the chain transfer is greatly suppressed by the bulky ligands even at elevated temperatures. However, the chain walking rate is increased with increasing temperature, leading to an obvious drop in polymer melting point. The resultant polyethylene obtained at 40˝C displayed a very narrow molecular weight distribution (PDI < 1.20) while relatively broad molecular weight distributions (PDI > 1.40) were observed at elevated temperatures. No clear correlation between the ligand electronic effects and the catalytic properties was observed in this system. This is similar with the cyclophane-based α-diimine Ni(II) system (Scheme 1G) [36].
The branching densities of the obtained polyethylenes were increased with polymerization temperatures, following the same trend as other α-diimine Ni(II) catalysts [3]. Interestingly, the branching densities were observed to be lower for the catalyst with electron-donating substituents. For example, with MAO at 100˝C, the branching density for polyethylene generated using catalyst 1a is 62/1000C (Table 1, Entry 4), whereas branching density of 74/1000C is observed for catalyst 1c (Table 1, Entry 14). In addition, activator AlEt 2 Cl led to slightly lower branching density than activator MAO.  (Table 1).
Since the dibromo nickel catalyst with CF3 substituent cannot be prepared, the Ni(acac) system could provide more systematic comparisons (Table 2). Complex 2b was selected to probe the influence of aluminum cocatalysts and polymerization conditions. In sharp contrast to complex 1b, MAO was not able to activate complex 2b for ethylene polymerization at 40 or 80 °C ( Table 2, entries 1 and 3). Triisobutylaluminum (TIBA) and AlEtCl2 cocatalysts were not effective either (Table 2, entries 2, 4 and 5). Luckily, cocatalyst AlEt2Cl was able to activate 2b for ethylene polymerization. Similar to complex 1b, the activity of complex 2b was increased with temperature in the range of 40-100 °C ( Table 2, entries 6-9). However, the increase in activity with temperature in complex 2b is much more dramatic than complex 1b. The activities and the polyethylene molecular weight of 2b are similar to the cases of 1b. These results suggest that it is difficult to activate the bulky catalysts in the Ni(acac) system. The right cocatalyst and high polymerization temperatures are necessary to generate active catalytic species.  Quite surprisingly, complex 2d containing the electron-withdrawing CF3 group showed the highest activity of 6.18 × 10 6 g of PE (mol of Ni) −1 ·h −1 ( Table 2, Entry 12), which is almost three times as much as that of 2a (2.14 × 10 6 g of PE (mol of Ni) −1 ·h −1 , Table 2, Entry 10) and twice as much as that of 2b (3.76 × 10 6 g of PE (mol of Ni) −1 ·h −1 , Table 2 (Table 1).
Since the dibromo nickel catalyst with CF 3 substituent cannot be prepared, the Ni(acac) system could provide more systematic comparisons (Table 2). Complex 2b was selected to probe the influence of aluminum cocatalysts and polymerization conditions. In sharp contrast to complex 1b, MAO was not able to activate complex 2b for ethylene polymerization at 40 or 80˝C ( Table 2, entries 1 and 3). Triisobutylaluminum (TIBA) and AlEtCl 2 cocatalysts were not effective either (Table 2, entries 2, 4 and 5). Luckily, cocatalyst AlEt 2 Cl was able to activate 2b for ethylene polymerization. Similar to complex 1b, the activity of complex 2b was increased with temperature in the range of 40-100˝C ( Table 2, entries 6-9). However, the increase in activity with temperature in complex 2b is much more dramatic than complex 1b. The activities and the polyethylene molecular weight of 2b are similar to the cases of 1b. These results suggest that it is difficult to activate the bulky catalysts in the Ni(acac) system. The right cocatalyst and high polymerization temperatures are necessary to generate active catalytic species. Quite surprisingly, complex 2d containing the electron-withdrawing CF 3 group showed the highest activity of 6.18ˆ10 6 g of PE (mol of Ni)´1¨h´1 (Table 2, Entry 12), which is almost three times as much as that of 2a (2.14ˆ10 6 g of PE (mol of Ni)´1¨h´1, Table 2, Entry 10) and twice as much as that of 2b (3.76ˆ10 6 g of PE (mol of Ni)´1¨h´1, Table 2, Entry 9) and 2c (2.80ˆ10 6 g of PE (mol of Ni)´1¨h´1, Table 2, Entry 11). However, complex 2d generated polyethylene with the lowest molecular weight (M n = 57.5ˆ10 4 ) among all complexes. Complex 2b produced polyethylene with the highest molecular weight (157ˆ10 4 ), which is ca. twice as much as that by 2c (81.0ˆ10 4 ) and three times as much as that by 2a (60.6ˆ10 4 ) and 2d. Complex 2a bearing the electron-donating OMe group produced polyethylene with the highest melting point. However, the melting point is only slightly influenced by the ligand electronic effect. The molecular weight distribution of the polyethylene produced by the Ni(acac) system with the AlEt 2 Cl activator is broader than that of the dibromo Ni(II)/MAO system. The polymerization at 100˝C revealed the following trends in activity: CF 3 > CH 3 > Cl > OCH 3 , molecular weight: CH 3 > Cl > OCH 3 > CF 3 and melting points: OCH 3 > CH 3 > CF 3 > Cl. Despite the much more dramatic differences in ethylene polymerization in the Ni(acac) system compared with the dibromo nickel system, no clear trend was observed. Similar with that observed in complexes 1a-1c, the polyethylene branching density was increased with polymerization temperature for complex 2b. There was no obvious trend for branching density among complexes 2a-2d. It is interesting to note the different performances between the dibromo nickel system and the Ni(acac) system. Theoretically, both systems should perform similarly after MAO or Et 2 AlCl activation, because of the same ligand structures and, correspondingly, the same active species. The differences we observed might originate from the differences in the activation/initiation processes or the different counter-anions in the two systems. Future work is required to better understand the ligand electronic effect in the α-diimine nickel catalyst system.

Conclusions
In conclusion, a series of dibenzhydryl-based α-diimine nickel complexes bearing a range of electron-donating and -withdrawing substituents were prepared. The aim is to systematically investigate the ligand electronic effects on the ethylene polymerization in this class of catalysts. The dibromo nickel-catalyzed ethylene polymerization was relatively insensitive to the electronic perturbation. However, electronic effects in the Ni(acac) system were pronounced, although no discernible trend was observed. The CF 3 -substituted complex 2d exhibited exceptionally high activity and stability at elevated temperatures. All of these complexes were highly active in ethylene polymerization with activities on the level of 10 6 g of PE (mol of Ni)´1¨h´1, generating polyethylene with molecular weight larger than one million. These complexes also showed great thermal stability, maintaining high activity and high polyethylene molecular weight at up to 100˝C.