α-Diimine Ni-Catalyzed Ethylene Polymerizations: On the Role of Nickel(I) Intermediates

Abstract

Nickel(II) complexes with bidentate N,N-α-diimine ligands constitute a broad class of promising catalysts for the synthesis of branched polyethylenes via ethylene homopolymerization. Despite extensive studies devoted to the rational design of new Ni(II) α-diimines with desired catalytic properties, the polymerization mechanism has not been fully rationalized. In contrast to the well-characterized cationic Ni(II) active sites of ethylene polymerization and their precursors, the structure and role of Ni(I) species in the polymerization process continues to be a “black box”. This perspective discusses recent advances in the understanding of the nature and role of monovalent nickel complexes formed in Ni(II) α-diimine-based ethylene polymerization catalyst systems.

1. Introduction

Currently, a variety of highly branched elastomeric polyethylenes (PEs) is produced by co-polymerization of ethylene with linear α-olefins such as propene, 1-butene, 1-hexene, and 1-octene [1]. The synthesis of branched PE via ethylene homopolymerization is more attractive both from economic and technological perspectives, since such approaches do not require the use of expensive pure α-olefins. Brookhart and co-workers introduced Ni(II) complexes with N,N-bidentate α-diimine ligands [2], capable of producing branched PE from ethylene as the only feedstock. Since this pioneering discovery, the area has advanced greatly, with the modifications of α-diimine ligands (by tuning both the backbone and substituents) and adjusting proper polymerization conditions providing access to polymers with desired molecular (molecular weight and molecular weight distribution, degree of branching) and mechanical characteristics [3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20].

For the rational design of Ni(II) α-diimine catalysts with the desired properties (activity, thermal stability, molecular weight characteristics and microstructure of the produced PEs), detailed understanding of the reaction mechanism and of the structure–properties relationships is a prerequisite. The key role of Ni(II)-alkyl complexes as active sites of ethylene polymerization was disclosed by Brookhart and co-workers almost 20 years ago [21], and has been extensively corroborated in further studies [22,23,24]. However, although Ni(I) species have been found to be ubiquitous in such systems, their nature and role have remained unclear until recently [25,26,27,28,29,30,31,32]. In the present perspective, we survey the existing experimental data, related to the structure of the monovalent nickel species formed in Ni(II) α-diimine-based catalyst systems, and discuss their possible roles in the ethylene polymerization process.

2. Cationic Ni(II) α-Diimine Complexes in Ethylene Polymerization

In 1999, Brookhart and co-workers introduced the first family of cationic Ni(II) complexes of the type [LNi^IIMe(X)]⁺[BAr′₄]⁻ (where L = L¹ or L⁵, X = H₂O or Et₂O, and Ar′ = 3,5-C₆H₃(CF₃)₂) (Figure 1, Scheme 1) [21]. According to the results of in situ ¹H NMR experiments, ethylene addition to the samples containing [LNi^IIMe(X)]⁺[BAr′₄]⁻ led to ligand X displacement by C₂H₄ molecule, followed by monomer insertion into the Ni^II-CH₃ bond and further enchainment of the Ni^II-alkyl moiety. The reaction between [L¹Ni^IIMe(Et₂O)]⁺[BAr′₄]⁻ and C₂H₄ is shown in Scheme 1.

These findings clearly demonstrate the key role of cationic Ni(II)-alkyl complexes as the active species of ethylene polymerization. Further studies of agostic interactions between the Ni(II) center and the β-H proton of the Ni^II-alkyl moiety in the cationic parts of the ion pairs [L¹Ni^IIR]⁺[BAr′₄]⁻ (R = Et, Pr, ⁱPr) corroborated the “chain-walking” mechanism leading to formation of branched PE [33,34] (Scheme 2), previously proposed by Fink and Mohring for Ni(II) aminobis(imino)phosphorane catalysts [35].

More recently, structurally related cationic Ni(II) complexes of “real” catalyst systems such as L²Ni^IIBr₂/MAO, L²Ni^IIBr₂/MMAO (Scheme 1 and Scheme 3; MAO—methylalumoxane, MMAO—methylalumoxane, modified by AlⁱBu₃), and L³Ni^IIBr₂/Al₂R₃Cl₃ (R = Me or Et) were identified by NMR spectroscopy [24,36]. Complex [L²Ni^II-^tBu]⁺[MeMMAO]⁻ was observed in the L²Ni^IIBr₂/MMAO catalyst system at −20 °C, whereas in the L²Ni^IIBr₂/MAO one, heterobinuclear dimethyl-bridged congener [L²Ni^II(μ-Me)₂AlMe₂]⁺[MeMAO]⁻ was detected under the same conditions (Scheme 3). Both these species could play the role of direct precursors of the active sites of ethylene polymerization. However, in the absence of ethylene, [L²Ni^II-^tBu]⁺[MeMMAO]⁻ and [L²Ni^II(μ-Me)₂AlMe₂]⁺[MeMAO]⁻ reduce to the monovalent nickel compounds at temperatures higher than −20 °C [24], which suggests that complexes of nickel(I) can also exist at higher temperatures, i.e., under practical polymerization conditions. Therefore, it is important to monitor the formation (and subsequent transformations) of nickel(I) species in “real” catalyst systems and reveal their possible role(s) in the catalytic cycle.

3. Ni(I) α-Diimine Complexes in Ethylene Polymerization

Despite the generally recognized role of Ni(II)-alkyl complexes as the active sites of ethylene polymerization, the nature and role of Ni(I) complexes in polymerization continues to be a subject of extensive studies [25,26,27,28,29,30,31,32,37,38]. Because of the lower tolerance (compared with Ni(II) counterparts) of Ni(I) α-diimine complexes to trace amounts of moisture and oxygen, the number of examples of their successful isolation and characterization has been limited [31,32,39,40,41,42,43].

In 2007, Reiger and co-workers synthesized two dinuclear monovalent nickel complexes [(L⁴Ni^IBr)₂MgBr₂(THF)₂] and [(L⁴Ni^IBr)₂] (Figure 2) and tested them in ethylene homopolymerization [32]. According to the results of catalytic studies, [(L⁴Ni^IBr)₂MgBr₂(THF)₂] and [(L⁴Ni^IBr)₂] did not display any detectable polymerization activity, but adding an excess of AlMe₃ converted them into “polymerization-active species”. The mechanism of formation of the active Ni(II) species was not established; however, the authors hypothesized that the latter could be accounted for by disproportionation of the type Ni^I = 1/2Ni^II + 1/2Ni⁰.

In 2015, Gao and Mu and co-workers isolated a neutral heterobinuclear nickel(I) complex [L⁵Ni^I(μ-Br)₂AlMe₂] from the L⁵Ni^IIBr₂/AlMe₃ reaction mixture (Scheme 4, Al/Ni = 4/1) [31]. When activated with excess AlMe₃ or MAO, [L⁵Ni^I(μ-Br)₂AlMe₂] displayed high ethylene polymerization activity, comparable with those of L⁵Ni^IIBr₂/AlMe₃ and L⁵Ni^IIBr₂/MAO systems. It was proposed that the active Ni(II) center was formed by the intramolecular electron transfer from Ni(I) to the redox non-innocent α-diimine ligand to form the [L^5(•−)Ni^II(μ-Me)₂AlMe₂] species (Scheme 4).

Chirik and co-workers prepared mixed-valence Ni(I)−Ni(II) ion pairs [L¹Ni(μ-H)]₂[BAr′₄] and [L⁴Ni(μ-H)]₂[BAr′₄] (Ar’ = 3,5-C₆H₃(CF₃)₂) that demonstrated catalytic activity in the polymerization of hexene regioisomers (1-hexene, trans-2-hexene, trans-3-hexene) when activated with MAO or Et₂AlOEt (Figure 3) [42].

Gao and co-workers prepared and characterized cationic bis-ligated complex [(L⁵)₂Ni^I][B(C₆F₅)₄], which displayed moderate ethylene polymerization activity when activated with AlMe₃ (Figure 3) [37].

Generation of monovalent nickel complexes in the catalyst systems in situ via one-electron reduction of Ni(II) precursors seems to be an easier approach to the investigation of the role of Ni(I) species in the ethylene polymerization process, provided that side products of reduction or (and) traces of the starting material do not significantly affect the catalytic behavior. Various one-electron reducing agents can be successfully used for the quantitative conversion of Ni(II) complexes to the corresponding Ni(I) counterparts [43].

Long and co-workers investigated the effect of addition of cobaltocene Cp₂Co (a widely used one-electron reducing agent) on the catalytic properties of L⁵Ni^IIBr₂/MAO system [44]. It was found that gradual increase of the Co/Ni ratio from 0 to 1 was accompanied by a 30% decrease of the degree of branching (from 114 to 88 branches/1000 C), while the catalytic activity did not significantly change. To establish the change of nickel spin state in the course of the reaction of L⁵Ni^IIBr₂ with cobaltocene, an EPR spectrum of the sample L⁵Ni^IIBr₂/Cp₂Co (Co/Ni = 1, toluene) was recorded. Single resonance at g = 2.342, characteristic of Ni(I) (S = 1/2) species, with uncoupled spin localized at the metal center, was detected at −196 °C. The structure of this complex was not disclosed.

Another original approach to one-electron reduction of Ni(II) center in the L⁵Ni^IIBr₂/MAO system is based on using photoreductants as stoichiometric one-electron donors, followed by exposing to visible light [45]. Adding the photoreductant (fac-Ir(ppy)₃ = tris[2-phenylpyridinato-C²,N] iridium(III)) to the system L⁵Ni^IIBr₂/MAO allowed achieving partial reduction, with the degree of reduction depending of the exposure time. Like in the case of L⁵Ni^IIBr₂/Cp₂Co/MAO system, the reduction of the Ni(II) center to the monovalent state leads to a drop of the degree of branching (from 111 to 93 branches/1000 C) of the resulting PE.

The explanation for the observed PE branching variation was suggested in 2020 by Roy with coworkers [38]. Computational studies have shown that Ni(II) species of the type L^5(•−)Ni^II(μ-Me)₂AlMe₂ rather than L⁵Ni^I(μ-Me)₂AlMe₂ congeners operate in the catalyst system L⁵Ni^IIBr₂/Cp₂Co/MAO. By analogy with [L¹Ni^IIR]⁺ species, observed by Brookhart and co-workers, it is assumed that complexes with the proposed structure L^5(•−)Ni^IIR act as the active sites of ethylene polymerization in these systems. Although the uncoupled electron was preferentially localized at the α-diimine ligand, small fraction of electron density at the metal center weakened the agostic interaction between Ni and β-H proton of the growing polymeryl chain, thus suppressing the β-hydride abstraction and hence restricting the degree of chain-walking, responsible for the formation of branches (Scheme 2).

Using EPR spectroscopy in situ, Petrovskii and co-workers documented rapid and quantitative reduction of the Ni(II) center to Ni(I) state upon activation with 20 equiv. of MAO in toluene at room temperature [29]. The resulting catalyst system was active in ethylene polymerization, with the ethylene consumption rate profile typical for the majority of homogeneous polymerization catalysts, with high initial polymerization rate (90 kg·mol_Ni⁻¹·bar⁻¹·h⁻¹) followed by gradual rate decay. The EPR spectra witnessed resonances of two types of paramagnetic species. The nearly axial frozen-solution signal (g₁ = 2.22, g₂ ≈ g₃ = 2.08) was tentatively assigned to a heterobinuclear L⁵Ni^I(μ-X)₂AlX₂ complex (the nature of X group was not specified), whereas a multiplet at g = 2.0 with partially resolved hyperfine structure (hfs) from one aluminum and two nitrogen nuclei was ascribed to aluminum species L^5(•−)AlX₂, the product of irreversible α-diimine ligand transfer to the co-catalyst.

The nature of monovalent nickel species formed upon Ni(II) α-diimines activation with MMAO and AlMe₃ were extensively studied using ¹H, ²H NMR and EPR spectroscopy in situ by Soshnikov and Talsi and co-workers [25,26,27]. When L⁴Ni^IIBr₂ complex, containing bulky o-ⁱPr-substituents, was activated with AlMe₃, two types of paramagnetic heterobinuclear Ni(I) complexes, L⁴Ni^I(μ-Br)(μ-Me)AlMe₂ and L⁴Ni^I(μ-Me)₂AlMe₂, were detected and characterized by NMR and EPR (Scheme 5, top). Using ²H-enriched AlMe₃ made it possible to assign the key ¹H and ²H NMR resonances of the AlMe₂ moiety of L⁴Ni^I(μ-Br)(μ-Me)AlMe₂, which confirmed its heterobinuclear nature. In the case of L²Ni^IIBr₂ with α-diimine ligand L² containing less bulky Me-substituents (Figure 1), only dimethyl-bridged heterobinuclear complex L²Ni^I(μ-Me)₂AlMe₂ was detected in the L²Ni^IIBr₂/AlMe₃ reaction solution (Scheme 5, bottom).

Using MMAO as activator for L⁴Ni^IIBr₂ led to a mixture of heterobinuclear complexes L⁴Ni^I(μ-Br)(μ-Me)AlR₂ and L⁴Ni^I(μ-Me)₂AlR₂ at Al/Ni = 25 (R = Me or ⁱBu), with only the latter species existing in the reaction solution at high Al/Ni ratios (Al/Ni > 100). In the case of the L²Ni^IIBr₂/MMAO sample, only L²Ni^I(μ-Me)₂AlR₂ was present, even at Al/Ni = 25. These results are in line with the hypothesis of Petrovskii and co-workers on the nature of Ni(I) species formed in LNi^IIBr₂/MAO catalyst systems.

Besides Ni(I) complexes, well-resolved EPR multiplets of L^2(•−)AlR₂ were detected in the systems L²Ni^IIBr₂/AlMe₃ and L²Ni^IIBr₂/MMAO (Figure 4). The propensity to the α-diimine ligand scrambling (from Ni to Al) strongly depends on the ligand structure. Indeed, the concentration of L^4(•−)AlR₂ in the systems L⁴Ni^IIBr₂/AlMe₃ and L⁴Ni^IIBr₂/MMAO was significantly higher than that of L^2(•−)AlR₂ in the systems L²Ni^IIBr₂/AlMe₃ and L²Ni^IIBr₂/MMAO at the same Ni/Al molar ratio. Combining L²Ni^IIBr₂ with AlⁱBu₃ at +25 °C leads to complete ligand transfer from Ni to Al even at relatively low Ni/Al ratios (Ni/Al ≥ 10). The system L²Ni^IIBr₂/AlⁱBu₃ showed no activity in ethylene polymerization, thus providing indirect evidence in favor of the catalyst deactivation via L^2(•−)AlR₂ formation.

Dimethyl-bridged nickel(I) complexes L²Ni^I(μ-Me)₂AlR₂ (R = Me or ⁱBu) ultimately formed both in the L²Ni^IIBr₂/MMAO and L²Ni^IIBr₂/AlMe₃ catalyst systems at high Al/Ni ratios (Al/Ni ≥ 500). To elucidate the role of these species in the ethylene polymerization process, a series of catalytic experiments was performed. Both L²Ni^IIBr₂/MMAO and L²Ni^IIBr₂/AlMe₃ displayed high ethylene polymerization activity, yielding branched PE (

Table 1. Ethylene polymerization data for the catalyst systems L²Ni^IIBr₂/MMAO and L²Ni^IIBr₂/AlMe₃ ^a.

Entry	Activator	Activity b	Mnc	Mw/Mn	Branches/1000 C d
1	MMAO	3.7	39	1.9	40 ± 2
2	AlMe3	2.7	39	2.0	36 ± 2
3 e	MMAO	1.5	38	2.3	39 ± 2
4 f	AlMe3	1.3	46	2.2	36 ± 2

^a Polymerization conditions: heptanes (200 mL) for entries 1, 3 and toluene (100 mL) for entries 2, 4; time 60 min; T = 50 °C; P(C₂H₄) = 5 bar; 2.0 μmol of pre-catalyst L²Ni^IIBr₂; Al/Ni = 500. ^b In 10⁶ g PE/(mol_Ni·h). ^c In 10³ g/mol; measured by GPC. ^d Measured by ¹H NMR. ^e Complex L²Ni^IIBr₂ was mixed with 20 eq. of MMAO in heptanes, stored during 1 h at 25 °C, and then introduced in the reactor. ^f Complex L²Ni^IIBr₂ was mixed with 10 eq. of AlMe₃ in toluene, stored during 1 h at 25 °C, and then introduced in the reactor.

).

The time profiles of the rates of ethylene consumption by the systems L²Ni^IIBr₂/MMAO and L²Ni^IIBr₂/AlMe₃ were typical for the majority of homogeneous polymerization catalysts, with high initial polymerization rate followed by gradual activity decay (Figure 5a). The systems containing L²Ni^I(μ-Me)₂AlR₂ (R = Me or ⁱBu), generated in situ by preliminary quantitative reduction of L²Ni^IIBr₂ with 10 equiv. of AlMe₃ or 20 equiv. of MMAO, demonstrated a dramatically different catalytic behavior.

Indeed, one can distinguish three regions in ethylene consumption time profiles of the L²Ni^I(μ-Me)₂AlR₂/MMAO and L²Ni^I(μ-Me)₂AlMe₂/AlMe₃ systems: the initial activity increase from zero to maximum values within 10–15 min, followed by stationary ethylene consumption during the next 10 min, which eventually turns into gradual decline (Figure 5b). The measured (average) catalytic activities of the systems L²Ni^I(μ-Me)₂AlR₂/MMAO and L²Ni^I(μ-Me)₂AlMe₂/AlMe₃ were twice as low as those of the systems L²Ni^IIBr₂/MMAO and L²Ni^IIBr₂/AlMe₃. Crucially, the molecular characteristics of the polymers obtained were virtually the same (Table 1), thus witnessing the same, single-site, catalytically active species in both cases. In contrast to the results of Long and co-workers [44], complete reduction of Ni(II) to Ni(I) had no effect on the degree of PE branching. We believe that the observed kinetic peculiarities (namely, the initial induction period, Figure 5b) and polymer properties (Table 1) could be accounted for by the disproportionation pathway Ni^I = 1/2Ni^II + 1/2Ni⁰, previously proposed by Reiger and co-workers [32], rather than by intramolecular electron transfer L²Ni^I(μ-Me)₂AlR₂ → L^2(•−)Ni^II(μ-Me)₂AlR₂.

Therefore, the L²Ni^I(μ-Me)₂AlR₂ species observed in L²Ni^IIBr₂/MMAO and L²Ni^IIBr₂/AlMe₃ should be considered as catalyst resting state rather than deactivation products. To rationalize the mechanism of transformation of L²Ni^I(μ-Me)₂AlR₂ resting states into the Ni(II) active sites in the presence of ethylene, the system L²Ni^IIBr₂/MMAO/C₂H₄ was studied by EPR spectroscopy [27]. Before ethylene addition to the sample L²Ni^IIBr₂/MMAO (Al/Ni = 25), the resonances of L²Ni^I(μ-Me)₂AlR₂ with nearly axial g-factor anisotropy (g₁ = 2.208, g₂ = 2.060, g₃ = 2.050) were detected in the frozen solution (−196 °C) EPR spectrum (Figure 6a). Noticeably, the perpendicular component (g_┴ = g₂ ≈ g₃) displayed a partially resolved hfs from two nitrogen atoms of the α-diimine ligand (a(¹⁴_N) = 1.06 mT). Upon the addition of ca. 600 equiv. of C₂H₄ at −70 °C, partial conversion of L²Ni^I(μ-Me)₂AlR₂ to a new Ni(I) complex A with dramatically different frozen-solution EPR-parameters (g₁ ≈ g₂ = 2.34, g₃ ≈ 1.99) was observed (Figure 6b). At temperatures higher than −70 °C, ethylene was rapidly consumed (with PE formation), accompanied by disappearance of A and restoration of the initial concentration of L²Ni^I(μ-Me)₂AlR₂.

Remarkably, the parallel component (g_‖ = g₃ ≈ 1.99) of the frozen solution EPR spectrum of A displayed hfs from only one nitrogen atom (a(¹⁴_N) = 1.51 mT). This is evidence of dramatic reorganization of the first coordination sphere of the paramagnetic center in the course of L²Ni^I(μ-Me)₂AlR₂ conversion into A. Quantitative return of A into L²Ni^I(μ-Me)₂AlR₂ after ethylene consumption, apparently, reflects reversible ethylene coordination to Ni, with the α-diimine ligand remaining coordinated to Ni in both complexes. Based on the above data, A was tentatively assigned to an ethylene adduct of the type L²Ni^I(C₂H₄)R (R = Me or ⁱBu). Further studies are planned to reliably establish the structure of this adduct and evaluate its chemical reactivity.

4. Conclusions

Complexes of nickel(II) with bidentante N,N-donor α-diimine ligands, introduced by Brookhart and co-workers [2], have established themselves as promising catalysts for the preparation of valuable polymeric products, such as branched elastomeric polyethylene and copolymers of ethylene with polar monomers. Meanwhile, despite considerable efforts, detailed understanding of the catalytic mechanism has not been achieved as yet.

In their pioneering studies, Brookhart and co-workers disclosed the key role of cationic nickel(II)-alkyl complexes of the type [LNi^IIR]⁺[A]⁻ (L = α-diimine ligand, R = polymeryl, [A]⁻ = counter-ion) in the polymer chain growth. However, in-depth investigations witness that monovalent nickel species may prevail in the catalyst systems under “real” polymerization conditions. For example, in situ NMR and EPR studies of the Ni(I) species formed in “real” catalytic systems (LNi^IIBr₂/MMAO and LNi^IIBr₂/AlMe₃; L = α-diimine ligand) have demonstrated the prevalence of neutral heterobinuclear complexes of the type LNi^I(μ-Me)₂AlR₂ (where R = Me or ⁱBu) in the reaction solution. However, until recently, their role in the catalytic process has remained unexplored.

The observations of the reactivity of α-diimine Ni(I) complexes (either isolated or generated in the reaction mixture in situ) toward ethylene (affording branched PEs) indicated that monovalent nickel species should not be discounted as potential active species or, more plausibly, their direct precursors.

The mechanism of Ni(I) conversion to the active divalent state in the course of ethylene polymerization continues to be debated controversially. On the one hand, intramolecular one-electron transfer from the Ni(I) to the redox non-innocent α-diimine ligand (LNi^I(μ-Me)₂AlR₂ → L^(•−)Ni^II(μ-Me)₂AlR₂) can yield the catalytically active, formally “Ni(II)” sites. On the other hand, the occurrence of disproportionation of the type Ni^I = 1/2Ni^II + 1/2Ni⁰ in the course of polymerization cannot be excluded, either. Further studies are needed to distinguish between these possibilities. In either case, however, these complexes of monovalent nickel can play important roles as precursors of active species and/or catalyst resting state rather than the products of the catalyst deactivation. We foresee further progress in understanding the mechanistic landscape of α-diimine nickel(II)-based catalysts of ethylene polymerization, and try to contribute to it in the near future.