Titanium (IV) and Nickel (II) Catalysts Based on Anilinotropone Ligands

Historically, the development of highly active olefin polymerization catalysts has been a trigger for creating new polymers which impact on our daily lives in countless beneficial ways [1-5]. A recent instance is the development of group 4 metallocene catalysts that exhibit very high ethylene polymerization activities [2,3]. Based on the highly active group 4 metallocene catalysts, high performance linear low-density polyethylene (LLDPE), isotactic poly-(propylene) (iPP) and syndiotactic polystyrene (sPS) etc., have been developed [6]. Therefore, much effort has been directed towards the development of highly active catalysts, following the group 4 metallocene catalysts. In consequence, quite a few highly active catalysts based on both early and late transition metal complexes have been developed [7-14]. There are, however, only a few examples of titanium complexes displaying high ethylene polymerization activities [15-19], though titanium metal is the major player in highly active heterogeneous Ziegler-Natta catalysts. Accordingly, further researches have been conducted on titanium catalysts with the intention of developing the highly active titanium catalysts and applying them to the polymerization of ethylene.


Introduction
Historically, the development of highly active olefin polymerization catalysts has been a trigger for creating new polymers which impact on our daily lives in countless beneficial ways [1][2][3][4][5]. A recent instance is the development of group 4 metallocene catalysts that exhibit very high ethylene polymerization activities [2,3]. Based on the highly active group 4 metallocene catalysts, high performance linear low-density polyethylene (LLDPE), isotactic poly-(propylene) (iPP) and syndiotactic polystyrene (sPS) etc., have been developed [6]. Therefore, much effort has been directed towards the development of highly active catalysts, following the group 4 metallocene catalysts. In consequence, quite a few highly active catalysts based on both early and late transition metal complexes have been developed [7][8][9][10][11][12][13][14]. There are, however, only a few examples of titanium complexes displaying high ethylene polymerization activities [15][16][17][18][19], though titanium metal is the major player in highly active heterogeneous Ziegler-Natta catalysts. Accordingly, further researches have been conducted on titanium catalysts with the intention of developing the highly active titanium catalysts and applying them to the polymerization of ethylene.
As a result of ligand-oriented catalyst design research, Sandaroos and coworkers [19][20][21] described the catalytic performance of new titanium complexes 1-9 containing aminotropone chelate ligands for ethylene polymerization (Scheme 1). The following subsections detail the results observed in these works.

Search for acquiring the fundamentally active ligand
Among the typical catalyst components, ligands play a predominate role in the polymerization process. During electron exchange between metal and monomer, ligands help the metal to balance its electron density with receiving electrons from the coordinated ethylene through the metal and releasing electrons whenever required to facilitate the ethylene insertion process (Figure 1). Accordingly, to obtain a highly active catalyst, the existence of ligands with a notable balance between their electron donating and withdrawing, evidenced by calculation of energy gap between the HOMO (the highest occupied molecular orbital) and LUMO (the lowest unoccupied molecular orbital) of them, is a predominate requirement [22]. For comparison, the energy gap between the HOMO and LUMO of three well-known ligands, namely phenoxy-imine [15,[23][24][25][26][27][28][29], pyrrolide-imine [15,[27][28][29], indolideimine [30] and the new aminotropone chelate ligand [19] was studied using densityfunctional theory (DFT). Because of the reasonable energy gap between the HOMO and LUMO of aminotropone (2.6 eV), it was theoretically offered as a fundamentally active ligand (Scheme 2).  H  19  6  Ethyl  H  19  7  isoporpyl  H  19  8  phenyl  methyl  19,20  9 phenyl t-Butyl 19,20 Scheme 1. Titanium (IV) complexes based on anilinotropone ligands 1-9.

Figure 1.
Presentation of the olefin coordination on the transition metal [19].

Substituent effects
This section concentrates on the steric features of bis(aminotropone)Ti complexes compared with those of bis(phenoxyimine) Ti complexes (Ti-FI catalysts). Unlike Ti-FI catalysts, which require sterically demanding substituents in ortho position to the phenoxy oxygen atom in order to exhibit high ethylene polymerization activity [23][24][25][26], the activity of bis(pyrrolideimine) Ti catalysts (Ti-PI) decreases with more congestion adjacent to the anionic N of the PI ligand because there is not enough space for monomer insertion [16]. The steric bulk in Ti-FI catalysts is thought to afford effective ion separation between the cationic active species and an anionic co-catalyst, resulting in enhancement of the catalytic activity. Accordingly, bis(aminotropone) Ti complexes 1, 8 and 9 with a series of subtituents (H, methyl and t-Butyl) adjacent to the carbonyl were prepared and examined as ethylene polymerization catalysts ( Figure 2; Table 1, entries 1, 37 and 40) [19]. The polymerization results provided information on the potential of bis(aminotropone) Ti complexes for ethylene polymerization, and additional information about the effect of the substituent adjacent to the carbonyl on catalytic performance. The bis(aminotropone) Ti complex 9, which possesses the t-Butyl group, proved to be a poor catalyst for ethylene polymerization under the conditions employed (activity < 100 Kg PE mol cat -1 h -1 ). Instead, the bis(aminotropone)Ti complex 1 with an H atom adjacent to the carbonyl formed polyethylene (PE) with Mw 98000 and a high activity of 8000 Kg PE mol cat -1 h -1 . Additionally, catalyst 8, which possesses the methyl group, exhibited almost moderate activity (activity 3520 kg PE mol cat -1 h -1 ). These facts show that, for bis(aminotropone) Ti complexes 8 and 9, the methyl and t-Butyl groups in close proximity to the carbonyl evidently provide unfavorable steric hindrance to ethylene polymerization. This is in sharp contrast to Ti-FI catalysts, which require steric congestion near the polymerization center in order to display high activity. The activity obtained with complex 1 (8000 Kg PE mol cat -1 h -1 ) is extremely high for a Ti complex with no cyclopentadienyl (Cp) or PI ligand(s) [16,[23][24][25][26]. In fact, the activity exceeds that displayed by Ti-FI catalyst 10 under analogous conditions (3480 Kg PE mol cat -1 h -1 ) ( Figure 2) [22]. Since the substituent on the amine-N is situated near the active site, the bis(aminotropone)Ti complexes 1-7, containing three types of substituents, aromatic ring (catalyst 1-2), aliphatic ring (catalyst [3][4][5] and alkyl groups (catalyst 6-7) on amine-N were prepared and evaluated, with methylaluminoxane (MAO) activation, for the polymerization of ethylene to find a correlation between the N-aryl substituent and ethylene polymerization behavior. The polymerization results are presented in Table 1 which also includes the results for Ti-FI catalyst 10, Cp2TiCl2 and Cp2ZrCl2 as references (compare entries 1, 7, 12, 17, 35-36 and 40-43). The basic trend observed is that an increase in the steric bulk of the substituent in each of the three types of catalyst enhances catalytic activity. A reasonable hypothesis is that the sterically more encumbered substituent on the amine-N part of bis(aminotropone) Ti complexes induces more effective ion separation between the cationic active species and an anionic cocatalyst. In addition, it gives better steric protection to the anionic N from Lewis acidic compounds present in a polymerization medium, leading to increased catalytic activity.
Comparison of catalysts 1-2 with 3-5 revealed dramatically increased activities for 3-5, which is thought to be due to a good match between electronic and steric effects of them. Therefore, it can be concluded that the catalytic activity is influenced by both electronic and steric effects. In other words, a good integration between the steric and electronic properties of the catalyst is needed to reach the highest activities. It should be noted that catalyst 5, which includes a cyclooctyl group on the amine-N, gives an activity of 27200 Kg PE mol cat -1 h -1 , which compares favorably with Cp2TiCl2 or Cp2ZrCl2 combined with MAO. Additionally, in each of catalysts presented in Table 1, sterically more encumbered substituents on the amine-Ns generally afford higher-molecular-weight PEs (compare entries 1, 7, 12, 17 and 35-36). The steric congestion provided by the substituent, which probably reduces the rate of chain transfer more significantly than that of chain propagation, is responsible for the enhancement of the product molecular weight [30].
Additionally, the activity of the catalysts 1 and 8 and Mw values of resultant polymers increased with increasing the monomer pressure (entries 1-3 and 37-39). This behavior is mainly due to the high concentration of the monomer close to the catalyst active centers.
The large amount of hydrogen concentration could slightly increase the activity of the catalyst 1, 8 and 9 ( Figure 3) [20]. A reasonable explanation for this effect might be increase of homogeneity of polymerization system and return of catalytically less reactive species, such as those resulting from 2,1-insertions, to the catalytic cycle through their fast hydrogenation [20].
Moreover, as it can be seen in Figure 4 [20], polydispersity and Mn of the polymer obtained using the catalyst 1 increases with time.

DFT studies of catalyst
Since a bis(aminotropone) Ti catalyst contains a pair of non-symmetric bidentate ligands, it potentially displays five isomers (A-E) arising from the coordination modes of the two ligands in an octahedral geometry (Scheme 3) [19]. DFT studies suggested that bis(aminotropone)Ti catalyst 1 assumed isomer A, with a trans-carbonyl-Os, cis-amine-Ns and cis-Cls disposition. Additionally, DFT calculations were performed on a methyl cationic complex (an initial active species generated from bis(aminotropone) Ti catalyst 1 with MAO) in the presence of ethylene, to obtain information about the structure of the catalytically active species ( Figure 5) [19] .The calculations revealed that an ethylene-coordinated cationic species assumed an octahedral geometry with a trans-carbonyl-Os, cis-amine-Ns and cis-Me/coordinated ethylene disposition, which fulfills the pivotal requirement for a high efficient catalyst, i.e., a growing polymer chain and a coordinated-ethylene group in the cisposition. An inspection of the calculated structure indicated that the phenyl group on the amine-N was located in close proximity to the active site, suggesting the substituent on the amine-N is the strategic substituent vis-a`-vis catalyst design.
As discussed, unlike catalyst 1, catalysts 8 and 9, with the methyl and t-Butyl groups adjacent to the carbonyl-Os, display very low productivity in the polymerization of ethylene. The calculated structure of methyl cationic complexes originating from 1 and 9 is displayed in Figure 6 [19]. The t-Butyl group of complex 9 seems to provide steric congestion near the polymerization center, which diminishes the rate of chain propagation. This is probably because it obstructs ethylene from gaining access to the active site and subsequent insertion into the Ti-carbon bond. On the other hand, DFT studies show that the active species derived from bis(aminotropone) Ti catalyst 1 possesses higher electrophilicity  at the Ti center compared with that of Ti-FI catalyst 10 (Mulliken charge of the Ti in atomic unit, catalyst 1=2.212 [19], catalyst 10=2.005 [16]). Considering that the active species derived from a typical metallocene (Cp2TiCl2) has a Mulliken charge (Ti) of 1.308 [16], bis(aminotropone) Ti catalyst 1 exhibits particularly high electrophilicity. Therefore, a bis(aminotropone) Ti catalyst generates a catalytically active species that has higher electrophilicity than a Ti-FI catalyst.

Introduction
Olefin polymerization catalysts based on early-metal metallocene complexes were introduced nearly two decades ago and are seeing increasing commercial utilization [31][32][33]. Their single-site nature makes them attractive for ligand tailoring, and as a result these catalysts have been modified in innumerable ways to enhance polymerization activity, improve catalyst lifetime, increase the α-olefin/ethylene reactivity ratio in copolymerizations, and control microstructures of polypropylene and other poly-α-olefins [31][32][33]. A drawback of metallocene and classical Ziegler catalysts is the extreme oxophilicity of the early-metal center. This oxophilicity renders metallocenes inactive toward most functionalized monomers and highly sensitive to polar solvents and impurities. The sensitivity of metallocenes to polar substituents is largely responsible for an increase in interest in late-transition-metal complexes as olefin polymerization catalysts over the past several years [34][35][36][37]. Late-metal catalysts complement early-metal catalysts in several ways.
Much of efforts has focused on cationic systems of α-diimine ligands (Figure 7) as well as catalysts incorporating closely related neutral ligands [38,[40][41][42]48,49,[51][52][53]. The α-diimine Ni complexes are very reactive toward ethylene and α-olefins and, as noted above, produce polymers with unique microstructures. The cationic Ni catalysts are electrophilic and sensitive to protic solvents. Functional group compatibility is diminished relative to Pd analogues; however, it should be noted that copolymerization of ethylene and alkyl acrylates has been achieved at high ethylene pressures (> 500 psig) and temperatures above 80 °C [47,54].  In view of the high sensitivity of the cationic Ni and Pd catalysts, there has been substantial interest in developing neutral Ni catalysts to overcome these limitations. Several examples of neutral Ni catalysts based on SHOP-type (Shell Higher Olefin Process) ligands have been reported, although productivities and molecular weights are often low [45,50,55,56]. Neutral Ni catalysts modeled after α-diimine complexes, containing bulky ortho-disubstituted arylimine functionality, have been reported. Catalysts based on salicylaldimines (Figure 8), have been described by both the DuPont [57] and the Grubbs groups [58,59]. The most active systems contain either electron-withdrawing nitro substituents in the aromatic ring [57] or bulky substituents at C-3, with a 9-anthracenyl group being most effective [58,59]. In the latter case, activities of 1.3 ×10 5 kg PE mol Ni -1 h -1 and lifetimes in excess of 6 h have been observed at 45-50 °C [58]. Brookhart [60,61] designed a ligand which incorporated the key elements of the sixmembered chelate salicylaldimine ligand but which would lead instead to a five-member chelate. He chose for this purpose the 2-anilinotropone moiety because it contained the desired anionic N,O chelate, a hindered N-aryl group, and complete conjugation between the N and O. Several variations can be introduced by using hindered or electron-poor Naryl group and modified tropone skeleton. This family of ligands were synthesized using the palladium catalyzed cross-coupling of aniline with 2-triflatotropone [60][61][62]. The corresponding highly neutral nickel catalyst is almost high active in ethylene polymerization and it does not require an activator (Figure 9). To determine the effects of varying the ortho-aryl substituents on ethylene polymerization, in an effort a series of anilinotropone ligands and corresponding nickel complexes was synthesized (Scheme 4). Among these catalysts, they incorporating alkyl substituents at the 2-and 6-positions of the N-aryl ring generated high-molecular weight polyethylenes with monomodal molecular weight distributions of ca. 2 (or less at low temperatures).
Under optimized conditions (Table 2), catalyst 14 bearing isopropyl substituents at the 2and 6-positions of the N-aryl ring can produce PE, without the addition of a cocatalyst, with a TOF of 8.8 ×10 3 kg of PE mol Ni -1 h -1 in a 10 min run. Additionally, 14 produced PE with a substantially higher Mn when compared with the salicylaldimino type bearing the anthryl substituent in the ortho position (89.6 vs. 54 kg mol -1 ). Nevertheless 14 showed a short lifetime when compared to 2-salicylaldimino type bearing the anthryl substituent [62].
As can be seen in Table 3, with the exception of the run at 40 °C, the Mn value of the resultant polymers decreases with increasing temperature. The degree of branching steadily increases from 40 to 100 °C (8-67 branches per 1000 carbon atoms). This implies that the lower molecular weights obtained at higher temperatures are the result of an increase in the ratio of the chain transfer rate relative to the chain propagation rate, a normal feature of such polymerizations. The catalyst TOF is at a maximum at 80 °C. Ethylene pressure has also been shown to have dramatic effects on the resultant PE. Increasing the pressure from 1.01 to 41.34 bar at 80 °C caused a decrease in the branching number from 113 to 41 branches per 1,000 C [62].

Catalyst
Substitution of the 2-(2,6-diisopropylanilino)tropone ligand with either phenyl (25) or naphthyl (26) groups resulted in a small increase in productivities and lifetimes at 80 °C and 13.78 bar. However, at 40 °C these catalysts exhibited much longer lifetimes (t1/2›1 h) and higher total turnover numbers could be achieved relative to 80 °C polymerizations [62].  Table 2. Selected ethylene polymerization data of the Ni (II) catalysts based on anilinotropone a [62] Molecular weights of the polyethylenes increased with pressure, which suggests that chain transfer at least in part occurs through classical β-H elimination rather than chain transfer to monomer. The catalyst decay product is the Ni (II) bis-ligand complex, whose formation must be initiated by reductive elimination of the ligand from a Ni (II) species [62].

Catalyst P (bar) T (°C) TOF b Mw Mw/Mn
Similar to other natural nickel ethylene polymerization catalysts which have already been reported for oligomerization of α-olefins [64,65], oligomerization of 1-hexene at 40 and 60 °C by catalyst 14 led to low yields of oligomers with degrees of polymerization of 21 and 16, respectively [62]. Branching numbers are 147 and 152 branches per 1000 carbon atoms, indicating a very minor amount of chain straightening via a 2,1-insertion and chain walking [39].
Salicylaldimine-based neutral Ni (II) catalyst bearing electron-withdrawing NO2 group displayed high activity for polymerization of ethylene and resulted in high MW polymer with fewer branches than the unnitrated catalysts [63]. Accordingly, Brookhart [66] prepared anilinotropone-based neutral Ni (II) catalysts 27-29 (Scheme 5) to examine the effect on catalyst activity and stability of adding strong electron-withdrawing NO2 groups to the tropone and N-aryl ring.
As can be seen in   All PEs produced by nitro-substituted catalysts 27-29 had monomodal GPC traces (PDIs between 2 and 4) and lower branching densities than those of parent catalysts 13 and 14 (entry 4 vs 6 and entries 9,10 vs. 12).
Catalyst 27 was able to incorporate vinyltrimethoxysilane or 1-hexene into polyethylene, but at a reduced rate compared to ethylene homopolymerization [66]. 1-Hexene or 1-octene oligomerization was also studied with catalysts 27 or 28 [66]. The oligomers were produced with low activities and lower amount of branching than expected (for 1-hexene, the expected number of branches after 2,1-insertion without chain walking is 166 branches/1000C and for 1-octene is 125 branches/1000C), indicating minor chain walking of catalysts during oligomerization.

The mechanistic investigation of the polymerization of ethylene
Combined DFT/stochastic studies were undertaken on the mechanism of ethylene polymerization catalyzed by the neutral Ni-anilinotropone catalysts [67]. Chain propagation and isomerization as well as influence of reaction conditions on the branching formation were investigated. Similarly to the case of nickel-salicylaldiminato catalysts the activation barriers for the insertion of ethylene in complexes with the trans alkyl group to the oxygen donor were found higher than those encountered in cis/trans isomerization. Interestingly, stochastic simulation allowed for establishing temperature and pressure dependence of the polymer microstructure. In agreement with experimental evidences the model predicts a decrease with the number of branches with the increase of pressure. Temperature dependence behaves oppositely as a result of an increase in secondary insertions with the temperature.
To get a full mechanistic picture of a typical neutral catalyst system, an elegant and in-depth NMR study of ethylene polymerizations catalyzed by the anilinotropone complexes was also performed [67]. Detailed information concerning the chain propagation process, the barrier to ethylene insertion, the nature, and dynamics of the intermediate Ni alkyl complexes, and the chain transfer and catalyst decay processes were obtained. The rate of insertion of ethylene into the Ni-phenyl bonds of both unsubstituted and aryl-substituted of anilinotropone nickel (II) complexes (14 and 25) was monitored by low temperature NMR spectroscopy. The ethylene insertion, initially, forms the corresponding [Ni(L)(PPh3)(CH2CH2Ph)] complexes determined by monitoring the change in concentrations of starting complexes. First-order rate constants were measured as a function of temperature and ethylene concentration.
Even at high ethylene concentrations, only PPh3 complexes were observed (no ethylene complexes were detected). As expected, rates accelerate with temperature, increasing about an order of magnitude between -10 and 10 °C. Insertion is dramatically inhibited by phosphine concentration. Thus, it was concluded that for these systems, the catalyst resting state(s) are an equilibrium mixture of phosphine and ethylene complexes ( Figure 10). Nevertheless, at high ethylene pressures, the equilibrium is shifted nearly completely to the side of the ethylene complex and TOF becomes independent of ethylene pressure (saturation conditions). The barriers to migratory insertion in (N-O)Ni(R)-(C2H4) complexes, from measurements of TOFs under saturation conditions, were determined to lie in the range of 16-17 kcal mol -1 , that is only about 2-3 kcal mol -1 greater than those observed for cationic diimine complexes.
The intermediate alkyl complexes have α-agostic interactions with dynamic behavior similar to the cationic alkyl diimine complexes. The barrier to β-H elimination and reinsertion is estimated in ca. 17 kcal mol -1 . Free energy barrier to nickel-carbon bond rotation in these complexes occurs with a barrier of 11.1 kcal mol -1 . These isomerization processes account for branched polyethylenes generated from these catalysts and resemble those previously observed with diimine catalysts.
Thermolysis of (N,O)Ni(hexyl)(PPh3) generates the nickel hydride complex and 1-hexene via β-H elimination, through two pathways, one (dominant) independent of phosphine concentration and one involving reversible loss of phosphine, and thus inhibited by PPh3. This process is a model for chain transfer and clearly explains why polymer molecular weights decrease with increasing phosphine concentration. The rate of propagation is retarded by increasing [PPh3], while the rate of chain transfer is unchanged, resulting in an increase in Rct/Rprop. Since chain transfer to monomer would exhibit a rate inhibition equal to that for Rprop with added PPh3 it was concluded that the major chain transfer route is a simple β-elimination process, and not chain transfer to monomer.
At much slower rates, reductive elimination of the free ligand occurred from the hydride complex, and is inhibited by added phosphine. Catalyst decay under polymerization conditions was shown to occur by a similar process to generate free ligand and a bis-ligand complex formed by reaction of free ligand with an active catalyst species (Figure11).