Transition Metal–(μ-Cl)–Aluminum Bonding in α-Olefin and Diene Chemistry

Abstract

Olefin and diene transformations, catalyzed by organoaluminum-activated metal complexes, are widely used in synthetic organic chemistry and form the basis of major petrochemical processes. However, the role of M–(μ-Cl)–Al bonding, being proven for certain >C=C< functionalization reactions, remains unclear and debated for essentially more important industrial processes such as oligomerization and polymerization of α-olefins and conjugated dienes. Numerous publications indirectly point at the significance of M–(μ-Cl)–Al bonding in Ziegler–Natta and related transformations, but only a few studies contain experimental or at least theoretical evidence of the involvement of M–(μ-Cl)–Al species into catalytic cycles. In the present review, we have compiled data on the formation of M–(μ-Cl)–Al complexes (M = Ti, Zr, V, Cr, Ni), their molecular structure, and reactivity towards olefins and dienes. The possible role of similar complexes in the functionalization, oligomerization and polymerization of α-olefins and dienes is discussed in the present review through the prism of the further development of Ziegler–Natta processes and beyond.

1. Introduction

In the past few decades, catalytic transformations of olefins and conjugated dienes have become mainstream processes of the petrochemical industry: polyolefins are major multi-tonnage plastics [1,2]; polydienes are still indispensable in rubber manufacturing [3,4]. Advanced technologies of polyolefins and diene rubbers are based on the coordination polymerization of α-olefins, buta-1,3-diene, and isoprene [1,2,4,5,6,7,8,9,10,11]. Selective coordination di-, tri- and tetramerization of ethylene [12,13,14,15] are also industrially important processes. Coordination oligomerization [16,17], as well as the hydro- and carboalumination [18] of higher α-olefins, are currently lab-based processes, even though they show strong industrial prospects.

Supported metal salts (primarily chlorides) or metal complexes as pre-catalysts, and organoaluminum compounds as co-catalysts, are used in the majority of these processes. After activation, the reaction mixtures contain metal alkyls and alkylaluminum chlorides. In coordination polymerization, mechanistic understanding of the reaction mechanism as a coordination/insertion of α-olefin or diene molecule at the metal–alkyl (or metal–hydride) center in stable ligand environments is generally accepted [13,16,19,20,21]. For group 4 metal polymerization catalysts, the conventional Cossee–Arlman mechanism has been expanded by including interactions between the metal center and bulky counterions (XMAO⁻, B(C₆F₅)₄⁻, anionic supports, etc.) [22,23,24,25,26,27,28,29,30,31,32,33,34,35], and binuclear Zr₂ complexes [34,36]; however, the direct participation of R₂AlCl in the formation of catalytic species remains hypothetic, even with the results of numerous experimental and theoretical studies [37,38,39,40,41,42,43]. It is also an open question whether M–(μ-Cl)–Al species participate in the coordination polymerization of 1,3-dienes.

The current view of the problem is presented in Figure 1. Despite the many reviews on α-olefin and 1,3-diene functionalization, polymerization, and oligomerization, factual information about the possible role of M–(μ-Cl)–Al bonding is either absent or presented only fragmentary. In particular, M–(μ-Cl)–Al bonding was mentioned in reviews devoted to V-catalyzed olefin polymerization [44] and the Ni-catalyzed oligomerization of olefins [45,46]. In our recent review [17], we touched upon the possible role of Zr–(μ-Cl)–Al complexes in the selective dimerization of higher α-olefins, catalyzed by (η⁵-C₅H₅)₂ZrCl₂ (Cp₂ZrCl₂) and ansa-zirconocene dichlorides in the presence of 1–10 eq. methylalumoxane (MAO). The formation of Zr–(μ-Cl)–Al species during zirconocene-catalyzed hydro- and carboalumination of α-olefins is also not in doubt; however, the specific role of M–(μ-Cl)–Al in other catalytic process remains unclear [18]. To date, the comprehensive mechanistic concept, which takes into consideration M–(μ-Cl)–Al bonding in reaction intermediates and transition states of the catalytic process, has only been developed for the Cr-catalyzed trimerization of ethylene with the use of pyrrole/Cr/Et₂AlX (X = Cl, Et) systems [13,47,48].

In this study, we summarized disparate facts, presented in scientific periodicals, with the aim of demonstrating the fact that metal complexes, containing M–(μ-Cl)–Al fragments, are or may be directly involved in olefin and diene transformations as reagents or catalytic species. The review is divided into sections by the type of metal in the complex core.

2. Complexes of Ti

2.1. The Tebbe Reagent and Similar Ti Complexes

The Ti–Al heterobimetallic complex of the formula Cp₂Ti(μ-Cl)(μ-CH₂)AlMe₂ (Ti01) was first synthesized by Tebbe et al. in 1978; the reaction of Cp₂TiCl₂ with AlMe₃ resulted in the formation of Ti01 with 80–90% yields [49]. The synthesis of Ti01 was complicated by side reactions (Scheme 1), making it difficult to isolate the samples suitable for X-ray diffraction (XRD) analysis, and as a result, the molecular structure of Ti01 was only determined in 2014 [50] (Figure 2a). The difference between interatomic distances (d(Ti–CH₂) < d(Al–CH₂), d(Ti–Cl) > d(Al–Cl)) considers Ti01 as a complex of Cp₂Ti=CH₂ and Me₂AlCl. In this way, in Ti01, Me₂AlCl serves as a stabilizing agent for the Ti(IV) carbene complex, that can be removed by Lewis base.

A structural analog of the Tebbe reagent, Ti02, was synthesized by the reaction of the N¹,N¹,N²,N²-tetramethylethane-1,2-diamine (TMEDA) complex (TMEDA)TiCl₃(THF) with [(TMEDA)ZnI]₂CH₂ [51]. Analysis of the molecular structure of Ti02 (Figure 2b) showed that the value of d(Ti–CH₂) does not differ from that in Ti01. Notably, a series of chlorine-free analogs of Tebbe reagent was described in [52].

Close structural analogs of Ti01 were synthesized by the reactions of (η⁵-C₅H₄Me)CpTiClMe and (η⁵-1,2,4-C₅H₂Me₃)CpTiClMe with Me₃Al (Scheme 2, complexes Ti03 and Ti04, respectively) or by the interaction of (η⁵-C₅H₄Cl)CpTiCl₂ with Me₃Al (Scheme 2, complex Ti05) [53]. Treatment of Cp₂TiCl(CH=CHMe) with ⁱBu₂AlH resulted in the formation of complex Cp₂Ti(μ-Cl)(μ-CHEt)AlⁱBu₂ [54]; the product was not isolated, but has been identified by NMR spectroscopy. The attempt of Halterman and Ramsey to synthesize the chiral Tebbe reagent from 1,1’-binaphthyl-bridged bis(η⁵-tetrahydroinden-2-yl) Ti(IV) dichloride failed [55].

Ti01 has found its main applications for carbonyl methylenation similarly to phosphonium ylides in the Wittig reaction [52]. The reactions of Ti01 with carbonyl compounds are rapid and selective, even in the presence of other reactive groups (for example, –COOR [56], –C=NOR [57]), but this subject is beyond the scope of our review. However, α-olefins can react with Ti01: as shown by Grubbs et al., this reaction resulted in the formation of titanacyclobutanes, and allenes formed titanacyclobutanes with exocyclic C=C bonds (Scheme 3a) [58,59,60]. Titanacyclobutanes are reactive towards α-olefins, as demonstrated by Grubbs et al.; this reversible process proceeds through the formation of Cp₂Ti=CH₂ [61] (Scheme 3b). Ti–Al complexes Ti03–Ti05 reacted with ^tBuCH=CH₂ similarly, with a formation of the mixtures of isomeric titanacyclobutanes [53].

A degenerate metathesis reaction between titanium methylidene unveiled from Ti01 and unactivated alkenes, followed by acid hydrolysis, was recently proposed by Frederich et al. [62] as a direct method for the Markovnikov hydromethylation of alkenes (Scheme 4). The reaction is site-specific; the reactivity of olefins decreases in the order α-olefins > methylenealkanes > trisubstituted alkenes.

2.2. Single-Site Ti Catalysts of the Polymerization and Oligomerization of α-Olefins

In early experiments on the single-site polymerization of ethylene, conducted by Shilov’s group, the occurrence of cationic species during the reaction of Cp₂TiCl₂ with Et₂AlCl was regarded as an experimental criterion of the rapid catalytic process [63,64]. In the opinion of Shilov, after the intermediate formation of complex Cp₂TiClEt·EtAlCl₂, partial dissociation of the complex occurs, and Cp₂TiEt⁺ is active in polymerization. For the most part, this view has not fundamentally changed in terms of the reaction mechanism, and for sandwich and half-sandwich complexes of Ti(IV), the polymerization of α-olefins includes stages of π-coordination of alkene at the Ti–Alkyl cation and insertion of the alkene via the four-membered Cossee–Arlman transition state. Possible termination events for the growing polyolefin chain are β-hydride transfer to monomer, β-hydride elimination, and Ti → Al alkyl transfer (Scheme 5) [65,66]. The main problems arising during the use of Ti(IV) sandwich complexes derive from side reactions that deactivate the active titanocene catalyst via reduction of the Ti(IV) cationic center [65]. Stabilization of the active Ti(IV) species was achieved via ligand design, with the development of half-sandwich constrained-geometry complexes (CGCs) [67,68,69,70] and different Ti(IV)-based post-metallocene catalysts [5,71].

In the following model experiments on activation of Ti(IV) chloro complexes by aluminum alkyls, Ti–(μ-Cl)–Al-containing products were detected by ESR and electronic absorption spectroscopy. Thus, in the 1980s, Mach’s group studied the formation and relative stability of Ti–Al half-sandwich complexes with the formula CpTiAl₂Cl_8–xEt_x and (η⁶-C₆H₆)TiAl₂Cl_8–xEt_x, and demonstrated their reduced stability by increasing the number of Et fragments, x [72,73]. In 2004, Bonoldi et al. studied the activation of CpTiCl₃ and (η⁵-C₅Me₅)TiCl₃ (Cp*TiCl₃) by different aluminum alkyls, MAO and AlMe₃/[CPh₃][B(C₆F₅)₄] [74]. ESR spectral studies have shown that during the reaction of LTiCl₃ with Me₃Al at an Al/Ti ratio of 10:1, after 3 h, Ti(III) complexes are formed with 12% and 8% yields, respectively. This was explained by the more electron-rich nature of Cp* in comparison with Cp, which hindered an increase in the electron density by reduction (Ti(IV) + e⁻ → Ti(III)). In the presence of MAO, complex Ti06 was formed (Scheme 6). This complex demonstrated moderate catalytic activity in the polymerization of styrene, although it is the cationic complex Ti07 that was considered as an active polymerization catalyst.

The hypothesis of the coordination of R₂AlCl (Scheme 7a) with the possibility of the partial ionization of the active center, followed by the coordination/insertion of ethylene, was put forward by Jensen et al. in 1994 [75]. This hypothesis was based on the assumption that single-site Ziegler–Natta catalysts, obtained by the reaction of TiCl₄ with AlR₃, contain active Ti(IV) species. Quantum chemical optimization of the model complex MeCl₂Ti(μ-Cl)₂AlH₂·CH₂=CH₂ at the DFTG level showed weak coordination of the ethylene molecule (d(Ti–C) > 3 Å) [76]; no stationary point due to the π-coordination of ethylene was found for this system. Similar results were obtained by the same research group during more thorough theoretical studies [77]; however, the authors have not abandoned the idea of the direct participation of Ti–(μ-Cl)–Al species in the catalytic process. Notably, one year earlier, Sakai reported the results of ab initio studies of the formation and further transformations of exactly the same molecule, MeCl₂Ti(μ-Cl)₂AlH₂·CH₂=CH₂, in comparison with the MeTiCl₃/CH₂=CH₂ system [78]. First, Sakai was able to find the local minima corresponding to π-complex MeCl₂Ti(μ-Cl)₂AlH₂·CH₂=CH₂ (Scheme 7b). Second, even more interestingly, no stationary point was found for π-complex MeTiCl₃∙CH₂=CH₂, and the activation energy barrier of ethylene insertion for the MeTiCl₃/CH₂=CH₂ system was 20.5 kcal·mol⁻¹ higher than that for MeCl₂Ti(μ-Cl)₂AlH₂·CH₂=CH₂. Sakai explained it in this way: in the rate-limiting transition state of the ethylene insertion, the π-electrons in ethylene move to the Ti–C(ethylene) region to form the new Ti–C bond. The electrons, belonging to the Ti–C(methyl) bond, move to the C–C (methyl–ethylene) region to form a new C–C bond (push-pull mechanism, black arrows). In the Ti–(μ-Cl)₂–Al complex, the Ti–Cl² bond, opposite to the ethylene coordination side for the Ti center, increases in length with electron transfer to the Cl² atom along the reaction pathway (red arrow in Scheme 7b). The original π-electrons move to the new Ti–C bond region (opposite to the Ti–Cl² bond) easily (red arrow), this process is similar to S_N2 reaction. At the same viewpoints, the Ti–Cl¹ bond, opposite to Ti–C(methyl), decreases in length, and the electrons move from Cl¹ to Ti and from Ti–C(methyl) bond to C(methyl) (blue arrows). Therefore, the Al–Cl (or Ti–Cl) bonds switching mechanism facilitates the push–pull mechanism of the insertion.

It can be assumed that similar single-site catalytic species are formed when using TiCl₄/AlCl₃ on a SiO₂ support, activated by Et₂AlCl [79,80]. This catalyst was studied in the oligomerization of dec-1-ene, which resulted in the preferential formation of trimers (by wt.%). The authors did not discuss the nature of the catalytic species (including oxidation state of Ti), confining themselves to the analysis of oligomer’s microstructure. However, studies of the impact of AlCl₃/Ti ratio on catalytic activity and oligomer distribution have shown the presence of an optimum, suggesting Ti–(μ-Cl)–Al contacts. Note that dec-1-ene oligomers, obtained with the use of this Ti/Al catalytic system, were regioirregular and contained nonnegligible proportion of tail-to-tail and head-to-head monomer units, differing from structurally homogeneous products of zirconocene-catalyzed oligomerization (see Section 3.2). This fact, and uniform molecular weight distribution (MWD) of the oligomer mixtures, suggest this Ti/Al catalyst to be qualitatively distinct from conventional heterogeneous Ziegler–Natta catalysts (see Section 2.4).

The results of Jensen et al. [75] were augmented with the studies of Ti ether complexes by Bulychev’s group [81]. Here, it has clearly been established that in the presence of ethers, R₂AlCl reduces Ti(IV) to Ti(III). A conventional activator of single-site polymerization catalysts, MAO, was completely inactive towards Ti(III) etherates. In contrast, when using Et₂AlCl or 3:1 Et₂AlCl/Bu₂Mg mixture as activators, the ionic complex of Ti(III) [(15-Crown-5)TiCl₂]⁺[AlCl₄]⁻ (Ti08) demonstrated moderate activity in ethylene polymerization. The catalytic activity of Ti08 decreases with the increasing Al/Ti ratio; therefore, it can be assumed that only partial alkylation of the Ti–Cl bonds occurs. However, there was no experimental proof of Ti–(μ-Cl)–Al bonding in catalytic species discussed in [81].

There are limited studies devoted to the participation of Ti–(μ-Cl)–Al complexes with proven structure in catalytic polymerization and oligomerization of α-olefins. In 2012, Duchateau et al. reported the synthesis, structural characterization, and ethylene polymerization performance of heterobimetallic Ti–Al–pyrrolyl complexes (η⁵-2,5-Me₂C₄H₂NAlCl₂Me)TiCl₂Me (Ti09), (η⁵-2,3-Me₂C₈H₄NAlCl₂Me)TiCl₂Me (Ti10), and (η⁵-3,4,5,6-C₁₂H₁₂NAlCl₂Me)TiClMe₂ (Ti11) [82]. These complexes were obtained by the treatment of TiCl₄ with equimolar amounts of Me₃Al and the corresponding pyrrole ligands (Scheme 8).

Molecular structures of Ti09–Ti11 were determined by XRD analysis (Figure 3). All molecules exhibited a distorted tetrahedral piano stool configuration. The Ti(1)–CEN_pyrrole bond distances (CEN_pyrrole is a centroid of the pyrrole ring) are identical in Ti09 and Ti10 (d(Ti(1)–CEN = 2.062 Å), but are slightly shorter than the Ti(1)–CEN_pyrrole bond distance in Ti11 (d(Ti(1)–CEN = 2.088 Å). Although formally Ti09–Ti11 were assigned to η⁵-type complexes, in Ti10 and Ti11, titanium has a stronger interaction with the N atom, resulting in an η⁴,κ-bonding mode for Al–indolyl and Al–carbazolyl ligands.

Upon activation with MAO at 30 °C, Ti09 produces a single-site catalyst affording ultra-high-molecular weight polyethylene (UHMWPE, M_w = 2300 kDa, Đ_M= 2.5). When Ti09 was activated with [Ph₃C]⁺[B(C₆F₅)₄]⁻ and TIBA (50 eq.) at 30 °C, bimodal distribution of the polymer was detected, indicating the formation of two distinct active species. A similar bimodal distribution and significant reduction in M_n were observed when MAO-activated polymerization was carried out at 60 °C. MAO-activated Ti–Al complexes Ti10 and Ti11 displayed higher catalytic activity compared with Ti09, and also produced UHMWPE. Complex Ti10 showed the same bimodal behavior when activated with [Ph₃C]⁺[B(C₆F₅)₄]⁻ and TIBA.

To clarify polymerization mechanism, the reaction of Ti09 with B(C₆F₅)₃ was investigated to determine which of the methyl groups (Ti–Me or Al–Me) would be abstracted. The ¹H NMR spectrum clearly showed that B(C₆F₅)₃ selectively abstracts Al–Me group with the formation of a cationic complex Ti09′, containing a Ti–(μ-Cl)–Al fragment (Scheme 9). The ability of the supposed cationic species Ti09′ to polymerize ethylene was tested by treating Ti09 with B(C₆F₅)₃ under ethylene pressure (20 bar) at room temperature, and the rapid formation of polyethylene was detected.

DFT optimization of the putative structures of the cationic species and their ethylene adduct showed that a constrained type of geometry with the Ti–(μ-Cl)–Al fragment is the most stable configuration for Ti09′. This saturated complex Ti09′ should not be an active catalyst; either release of the bridged Cl (Type A structure in Scheme 9) or η⁵→μ ring slippage of the pyrrole moiety (Type B structure in Scheme 9) may result in a coordinatively unsaturated complex to coordinate and polymerize ethylene. The DFT calculations showed that the formation of a Type A complex requires 12.2 kcal·mol⁻¹, and no local minima were found for the Type B product of η⁵→μ ring slippage.

In this way, intramolecular Ti–(μ-Cl)–Al bonding has a strong impact on the catalytic behavior of Ti complex. The other extreme is weak Ti⋯Cl–Al coordination; de Bruin et al. [83] carried out a computational study of the influence of ‘chlorinated’ MAO anions on the catalytic behavior of complex Ti12 which represents an efficient pre-catalyst of the selective trimerization of ethylene with the formation of hex-1-ene [14,15,84]. The key intermediates of the reaction are presented in Scheme 10, although the energetically unfavorable process of but-1-ene elimination from intermediate M2 (see below) is not shown.

For DFT calculations, one of many possible MAO structures [85,86] was selected, namely, the cage (MeAlO)₆Cl⁻ model (Figure 4a), negatively charged due to the coordination of one Cl⁻, abstracted from Ti12 during activation. Two different potential free energy reaction profiles have been modeled (Figure 4b): the first only takes into account the cationic catalyst (cationic profile), whereas the second describes the energy changes in the presence of a chlorinated MAO (zwitterionic profile). The reaction starts with the activated catalyst M1 (two ethylene molecules coordinated to the cationic Ti(II) center). After oxidative coupling (M2), a third ethylene molecule coordinates and inserts with a formation of the seven-membered metallacycle (M3). This species can undergo a β-hydride transfer followed by reductive elimination (M4A), which, in turn, releases hex-1-ene. Alternatively, a fourth ethylene may coordinate to M3 and insert to yield a nine-membered metallacycle M4B. This latter species can then undergo a ring-opening reaction to yield 1-octene (M5A), or a fifth ethylene coordination and insertion can take place. The calculated barrier for β-hydride elimination in M2 (but-1-ene formation) was ~35 kcal·mol⁻¹. Based on a comparison of cationic and zwitterionic profiles, it can be concluded that the presence of (MeAlO)₆Cl⁻ does not impact the selectivity of ethylene oligomerization. Based on the results of modeling, the authors concluded that the presence of MAO model does not necessarily need to be taken into account, and the cationic system alone is sufficiently representative.

As reported by Blakemore et al., the reaction of the Ti(IV) complex (η⁶-C₉H₇)TiCl₂(N=P^tBu₃) with Et₃Al resulted in the formation of Ti(III)-Al complex Ti13 [87]. The molecular structure of Ti13 is presented in Figure 5a. Both complexes catalyzed the polymerization of ethylene being activated by Et₃Al in the presence of a solid superacid. During subsequent research [88], the complexes Ti14 and Ti15 of the formula (η⁶-C₉H₇)Ti (N=P^tBu₃)(μ-Cl)₂AlR₂ (R = Me and ⁱBu, respectively) were obtained, and the molecular structure of Ti15 (Figure 5b) was virtually identical to the molecular structure of Ti13. However, in experiments on the polymerization of ethylene, catalytic activities of the complexes decreased in the order Ti15 > Ti13 > Ti14. In this way, the nature of the substituents at Al atoms influenced the catalyst’s properties, thus indirectly confirming the presence of Ti–(μ-Cl)–Al bonding in catalytic species.

2.3. Single-Site Ti Catalysts of Polymerization of Dienes

Ti(IV) halogenides, activated by aluminum alkyls, represent the first-generation catalysts of coordination polymerization of 1,3-dienes [8,89]. However, despite half a century of research, molecular structure of the catalytic species, formed as a result of the reaction of LTiCl_n with R_mAlCl_3–m, is not entirely clear. In examining the reaction mechanisms of diene polymerization, the ‘pure’ cationic concept is still actual [90].

The participation of Ti–(μ-Cl)–Al species in the coordination polymerization of conjugated dienes was proposed by Mach et al., who studied the catalytic activity of η⁶-arene Ti complexes with AlCl₃ and alkylaluminum chlorides [91,92,93]. Similar complexes of (η⁶-C₆H₆)Ti[(μ-Cl)₂AlCl₂]₂ (Ti16) and (η⁶-C₆Me₆)Ti[(μ-Cl)₂AlCl₂]₂ (Ti17) were obtained and structurally characterized in the late 1970s and early 1980s by Thewalt et al. [94,95]. The molecules of Ti16 and Ti17 exhibit square pyramidal geometry, with the titanium atom above the rectangle of the bridging Cl atoms and with the π-bonded arene at the apex. When Ti16 was used as a single-component catalyst of the oligomerization of buta-1,3-diene, (1Z,5E,9E)-cyclododeca-1,5,9-triene was formed with 92% selectivity (Scheme 11) [91]. The authors assumed that the high stereospecificity of the cyclotrimerization indicates that the reaction proceeds by a sterically controlled coordination mechanism. The experimental results, i.e., the approximate second-order dependence of the reaction rate on buta-1,3-diene concentration, the low concentrations of all the intermediates participating in the catalytic steps and no interaction between the catalyst and final product, suggest that cyclotrimerization may proceed via the replacement of C₆H₆ by two molecules of buta-1,3-diene, followed by the formation of the active species Ti16′ (Scheme 11). In the presence of EtAlCl₂, Et₂AlCl, and Et₃Al, catalytic activity of complex Ti16 increased due to the substitution of non-bridged Cl atoms by Et groups [92]. In the presence of a large excess of Et₃Al, complete deactivation of the catalyst was observed. Complex Ti17 demonstrated low activity due to higher stability of the (η⁶-C₆Me₆)–Ti bond in comparison with the (η⁶-C₆H₆)–Ti bond [93]. However, the substitution of non-bridged Cl atoms by Et groups proceeded stepwise with maximum preservation of the molecular symmetry—by an example of the complex (η⁶-C₆Me₆)Ti[(μ-Cl)₂AlClEt]₂ (Ti17′), for which XRD analysis was performed. This substitution resulted in a manyfold increase in the catalytic activity [93].

In conclusion, we note that Ti complexes are not actually considered as prospective catalysts of the polymerization of conjugated dienes, and further comprehensive studies of the reaction mechanism seem unlikely.

2.4. Heterogeneous Ti Ziegler–Natta Catalysts

The TiCl₃/R₃Al system, called the ‘1st generation Ziegler–Natta catalyst’, was one of the early efficient catalysts of α-olefin polymerization [96]. It is for these catalysts that the common mechanism of the coordination polymerization of α-olefins was proposed by Cossee and Arlman [97,98]. This conventional mechanism (Scheme 12) involves the participation of Ti(III)–Alkyl species and is similar to the mechanism presented in Scheme 5, except that the Ti atom has zero charge and an oxidation state of +3.

In approaching the topic of modern and efficient Ziegler–Natta (ZN) catalysts, tributes should be paid to the work of Rodriguez and van Looy, who formulated heterobimetallic concepts of the reaction mechanism to explain the impact of alkylating agents on stereospecificity of TiCl₃ catalyst. This concept was based on an assumption of the presence of Ti centers with two potential vacancies on the Ti atom at the surface of TiCl₃ (coordination number of the Ti atom CN_Ti is equal to 4) [99]. According to this concept (Scheme 13), after alkylation by R₃Al, a Ti–(μ-Cl)–Al complex is formed (CN_Ti = 5). π-Coordination and subsequent insertion of the α-olefin molecule results in an alkyl complex, which further isomerizes with occupation of the position previously occupied by the R group, thus forming the isotactic polyolefin.

It has been nearly 70 years since the pioneering works of Ziegler and Natta [100,101], and there have been several generations of ZN catalysts. At present, the polymerization of α-olefins with the use of heterogeneous Ti-based MgCl₂-supported Ziegler–Natta catalysts (TMCs) is the most important industrial chemical process in the polyolefin industry [102,103,104,105,106,107,108]. However, despite the importance and impact of TMCs, these catalyst systems have not yet been completely explored. TMCs comprise the MgCl₂ support, the active Ti(III) [11,109,110], and Ti(II) [11,111] species; organoaluminum species formed from trialkylaluminum alkylating agent R₃Al, and Lewis base donors. Due to the practical importance of TMCs, thus far, thousands of articles and dozens of reviews have been published; in this section, we restrict ourselves to listing several examples of the possible or proven (very rarely) participation of R₂AlCl species in TMC-catalyzed polymerization.

Catalytic processes occur on the TMC surface. Pre-catalysts represent TiCl₄ adsorbed on MgCl₂ crystallites, and the binding strength and coordination environment of the Ti atom depend on the crystal structure of the MgCl₂ support. In their fundamental study [110], Cavallo et al. assumed the MgCl₂ bulk to be in the α-crystalline phase with the surface comprising (104) lateral cuts (Figure 6a) that have five-coordinate magnesium centers. The removal of one MgCl₂ unit results in a new (110) surface that can coordinate the titanium species (Figure 6b). Vanka et al. proposed the simpler model structure of the MgCl₂ support, close in meaning to Cavallo’s model [102] (Figure 6c). In their subsequent publication, Cavallo et al. upgraded their MgCl₂ model by (104)-facets presenting low-energy step-defects as potential sites for the strong coordination of TiCl4 [103].

After the absorption of TiCl₄, followed by the reaction with R₃Al, Ti(III)–Cl and Ti(III)–R species are formed [102]. Recent experimental DR UV/vis and NEXAFS data coupled with DFT simulation indicated that the majority of Ti sites in MgCl₂/TiCl₄ and MgCl₂/ethyl benzoate/TiCl₄ pre-catalysts are reduced by Et₃Al mostly to monomeric pentacoordinated Ti(III) Cl₅ species and, to a minor extent, to alkylated Ti(III) Cl₄Et species. For MgCl₂/dibutyl phthalate/TiCl₄ pre-catalyst, Et₃Al additionally promoted the formation of small TiCl₃ clusters with low catalytic activity [112].

The above-mentioned articles considered two-component systems MgCl₂/TiCl₃, but their results are relevant to the subject of our review, determining Ti-containing species capable of Ti–(μ-Cl)–Al bonding. However, real TMCs may contain absorbed organoaluminum species. During the study of the activation of the TiCl₄/MgCl₂ pre-catalyst by Et₃Cl, Oct₃Al, and Et₂AlCl with the use of magic-angle spinning (MAS) ²⁷Al NMR [113], Potapov et al. detected three types of alkylaluminum chloride species. The changes in ²⁷Al NMR MAS spectra of the TMC before and after polymerization indicated that part of the Al compounds was removed from the catalyst surface or that the coordination environments of Al in these compounds became disordered. In the authors’ view, the existence of Al compounds unremovable from the catalysts surface indicates the presence of inactive Ti–(μ-Cl)–Al species; in other words, R₂AlCl only acts as a catalyst inhibitor.

The same or a similar point of view received further development in a number of studies. In their theoretical study of the active sites in TMC [102], Vanka et al. used a simplified model of the MgCl₂ surface (Figure 6c) and optimized the geometries of different TI(III)–Et catalytic species that are formed during the polymerization of ethylene, including the presence of the EtO⁻ ligand, ester donors, or Et₂AlCl. In the absence of additives, the calculated difference in free activation energies ΔG^≠ for the propagation and termination stages was only 1.2 kcal·mol⁻¹. In the presence of esters, the difference increased to 12–13 kcal·mol⁻¹ and was only 3 kcal·mol⁻¹ when Et₂AlCl served as a donor. In the latter case, Ti–(μ-Cl)–Al coordination resulted in a substantial increase in the ΔG^≠ of ethylene insertion (15.8 kcal·mol⁻¹, see Figure 7). The results of this simulation explained how polymers can be successfully produced by TMCs without alkoxy groups and donor molecules [106]. In summary, the authors suggested that the possible role of organoaluminum activators is not limited by the reduction of Ti(IV); R₂AlCl can act as a donor due to its relatively weak Ti–(μ-Cl)–Al binding.

In 2017, Linnolahti et al. reported the first quantum chemical description of the initial steps of olefin polymerization on TMCs involving all the relevant catalyst components—absorbed Ti chlorides, electron donor, and R₃Al [114]. They demonstrated that TiCl₄ is coordinated on the (104) surface of MgCl₂ as a binuclear Ti₂Cl₈ and on the (110) surface as a mononuclear TiCl₄ (CN_Ti = 6 for both binding modes). Et₃Al coordinates to the MgCl₂ surface via an unsaturated Cl atom to initiate Ti alkylation. After alkylation via the Ti–(μ-Cl)–Al transition state on the (104) surface, Et₂AlCl is released from the MgCl₂ surface and dimerizes (exergonic process). However, in the presence of dimethyl phthalate as a model donor molecule at the (110) surface, the formation of Ti–(μ-Cl)₂–Al and then Ti–(μ-Cl)(μ-Et)–Al species becomes energetically favorable (Figure 8). However, even in this case, the process is finished by Et₂AlCl desorption. However, DFT modeling, presented in [114], closes by examining alkylation stage with a formation of Ti(IV)–Et species, the further reduction with a formation of active Ti(III) centers, as well as ability of these centers to conduct Ti–(μ-Cl)–Al bonding, remains open.

A recent study by Boisson [115] provided answers to certain questions about the role of organoaluminum activators in TMC-catalyzed polymerization. Based on the results of the study of the microstructure and MWD of polyethylenes, the authors suggested that Ti–(μ-Cl)–Al bonding may need to be considered in addressing the issue of the nature of ZN catalysts.

In [116], Busico’s group reported the results of an extensive experimental and theoretical investigation of the latest-generation commercial TMC. During this study, the researchers proposed a new function of R₂AlCl species, namely, interactions with absorbed internal donors that provide a higher degree of stereocontrol: for the insertion of prop-1-ene at the catalytic species formed by alkylation and reduction of the TiCl₄ precursors in panels (a) and (b) of Figure 9, the calculated ΔG_re/si values were ≈0 and 1.5 kcal·mol⁻¹, respectively; the latter was in good agreement with the experimental data.

At the end of this section, it is appropriate to mention the most recent studies by Bahri-Laleh’s group [117,118]. In [117], ethylene coordination/insertion at different TiEt and TiEt–(μ-Cl)₂–Al sites (Figure 10a–d) was modeled. Ethylene coordination at the Ti center on the (110) surface was thermodynamically more favored in comparison with it on the (104) surface due to the more acidic nature of 4-coordinated surface Mg atoms in (110). π-complexation of the ethylene molecule was more exergonic in Al-doped catalysts. The activation energy values for the insertion of ethylene molecule were 8.3 (a), 11.6 (b), 6.1 (c), and 9.5 (d) kcal·mol⁻¹. In this way, Al-doping increased the energy barrier needed for olefin insertion. In addition, the authors proposed that absorption of the Ti chloride at the (104) surface should not be ignored when analyzing distribution of the active sites, which is in contrast to the established paradigm. However, because TMCs represent a very striking example of the ‘catalytic black box’, any substantial expansions of the existing worldview can only be welcomed.

In [118], the catalytic activity of TMCs when using different organoaluminum activators (Et₃Al, Et₂AlCl, EtAlCl₂, and ⁱBu₃Al) was studied experimentally, and the effect of Ti–(μ-Cl)–Al bonding on catalytic activity was evaluated theoretically for Ti centers at the (110) MgCl₂ surface and these four organoaluminum compounds. From the results of the modeling, Et₂AlCl was the best Al component among others.

Concluding this section of the review, we have to note that, with the sole exception of complexes Ti09–Ti11 [82], the question of the participation of Ti–(μ-Cl)–Al species in catalytic transformations of α-olefins still has no clear answer. Irrelevance of this question to Ti-based post-metallocenes (too high CN_Ti) and fundamental unknowability of the TMC (the most important Ti-based catalyst) leave room for speculations, not complete answers. However, in the field of Zr-based polymerization catalysts we are facing a slightly better situation.

3. Complexes of Zr

Zirconocenes have a special place among Zr-based catalysts used in α-olefin chemistry. The hydro- and carboalumination of α-olefins [18], single-site polymerization with the use of MAO and perfluoroborate activators [119], selective dimerization [37,41,42] and oligomerization [16,17] of higher α-olefins—this is not a complete list of the processes with relevance for laboratory and industry. Historically, two groups of zirconocene-based catalytic species stand apart by the criterion of the charge of the catalytic Zr center. The first, zero-charged, group is traditionally considered as catalysts of hydroalumination, carboalumiation, and cyclocarboalumination of α-olefins and related reactions, limitedly used in the laboratory [18]. The second, ‘cationic’ species, represents highly efficient single-site α-olefin polymerization catalysts with a great potential of industrial applications [1,7,120,121]. Therefore, it is no surprise that the second group of the Zr catalysts attracted and continue to attract the most research attention; however, the participation of Zr–(μ-Cl)–Al species in catalytic process has only been proven experimentally for ‘neutral’ zirconocene species.

3.1. Zirconocene Chemistry: Non-Charged Complexes

The mechanisms of zirconocene-catalyzed hydroalumination and carboalumination of α-olefin were discussed ten years ago in the review by Parfenova et al. [18]. In this section, we consider only some key studies, published before 2010, and discuss the results of recent studies in more detail.

Before discussions of hydroalumination and carboalumination processes, it should be noted that the reaction of L₂ZrCl₂ with R₃Al until recently was considered as the natural and only possible pathway of the activation of metallocene dichlorides. When studying this reaction, it was found that the nature of the η⁵-ligand significantly affects the completion of the process. As shown by Beck and Brintzinger [122] for alkyl exchange reactions (Equation (1)), the values of apparent equilibrium constants K_obs were 0.005 (L = η⁵-C₅Me₅), 0.016 (L = η⁵-C₅H₄^tBu), 0.19 (L = η⁵-C₅H₄^tSiMe₃), 0.49 (L = Cp) and 1.2 (L₂ = rac-Me₂Si(Ind)₂). Both steric and electronic factors affect this equilibrium, and in particular, decreased electron density at the Zr atom appears to favor uptake of the Me group.

2 Cp₂ZrCl₂ + Al₂Me₆ ⇆ 2 Cp₂ZrMeCl⋯AlMe₂Cl

(1)

The difference in equilibria constants for this reaction can be explained based on the results of DFT modeling, as shown by Linnolahti et al., who showed that catalyst alkylation step by AlMe₃ has two viable routes. AlMe₃ coordinates to the metallocene either from the side or from the front, leading to corresponding alkylation reaction pathways. The kinetically favorable route is dependent on the ligand structure of the catalyst: the front pathway is preferred for Cp₂ZrCl₂ (Figure 11) but is sterically hindered for Me₂C(Cp)(Flu)Zrl₂, thereby preferring the side pathway [123]. Similar results have also been obtained in a more recent study by Kumawat and Gupta [124]. Zr01 was also detected by IR spectroscopy during studies of the activation of Cp₂ZrCl₂ by MAO [125].

Dimethyl derivatives of zirconocenes are formed during the second stage of the process. The intermediate complex of the formula Cp₂ZrMeCl⋯AlMe₃ (Zr01) was detected by ¹H NMR at −85 °C; however, at −60 °C, the characteristic signal of Zr–Me group had already disappeared [126].

The presence of direct Zr–(μ-Cl)–Al bonding in Cp₂ZrMeCl complexes with R₃Al has clearly been proven by Barron et al. [127]. They showed that even Me₂Al can coordinate with Cl (d(Al–Cl) = 2.46 Å, d(Zr–Cl) = 2.52 Å, Figure 12) with the formation of a relatively stable complex Cp₂ZrMeCl⋯AlMe₃ (Zr01′).

It is obvious that the reactions of L₂ZrCl₂ with ⁱBu₂AlH or ⁱBu₃Al may result in the formation of Zr–Al hydrides, which is accompanied by isobutylene elimination in the latter case [128]. The rate and equilibria of the reaction with ⁱBu₃Al also strongly depend on the nature of the η⁵-ligand fragment L₂; for instance, Ph₂C(Cp)(Flu) forms a mono-alkyl complex Ph₂C(Cp)(Flu)Zr(ⁱBu)Cl even with a 50-fold excess of ⁱBu₃Al, whereas Cp₂ZrCl₂ reacts with an excess of ⁱBu₃Al with the formation of [Cp₂ZrH₂·AlⁱBu₃]₂ (Zr02) [128], and reaction with ⁱBu₂AlH results in Zr03 [129] (Scheme 14). In the presence of ⁱBu₂AlCl, Zr02 forms the Zr–Al₃ complex Zr04; ring-substituted bis(cyclopentadienyl) L₂ZrCl₂ complexes demonstrate similar behavior [129]. For Zr ansa-complexes, the formation of L₂ZrCl(μ-H)₂AlⁱBu₂ species was detected.

Efficient Cp₂ZrCl₂-catalyzed hydroalumination uses ⁱBu₂AlH or its equivalent ⁱBu₃Al as Al sources due to the ease of isobutylene elimination with the formation of Al–H bonds (Scheme 15a). A greatly simplified hypothetical mechanism of this reaction, proposed by Negishi [130], was substantially developed and supplemented by new experimental and theoretical results by Dzhemilev and Parfenova group. In [131], they proposed a hydroalumination mechanism involving the intermediate formation of Cp₂Zr(μ-Cl)(μ-H)Al (Zr05) and Cp₂Zr(μ-H)₂Al (Zr06) species that are able to perform π-coordination and insertion of the α-olefin molecule. In the same study, a dimer of the Zr–Al complex (Zr07) was detected, and excess ⁱBu₃Al or ⁱBu₂AlH resulted in the formation of Zr–Al₂ and Zr–Al₃ complex hydrides (Zr03, Zr08, Zr09; the latter may contain μ-Cl fragments) (Scheme 15b). It was later discovered that in comparison with other Cp₂Zr-based complexes (including the Schwartz’s reagent Cp₂ZrHCl), Zr03 demonstrated the highest catalytic activity in hydroalumination (Scheme 15c) [132,133]. Probably, Cl-containing organoaluminum fragments in Zr–Al hydride complexes accelerate the intramolecular exchange between bridging and terminal H atoms and shift the equilibrium between the dimeric and active monomeric forms.

The validity of Scheme 15 was confirmed in follow-up studies of the reaction of different L₂ZrCl₂ complexes with ⁱBu₂AlH [134]; Zr08-type complexes were found to be active species of the hydroalumination process. Zr08 species are formed via Zr–(μ-Cl)–Al intermediate Zr06′; therefore, the presence of Cl in the catalytic system is important for the hydroalumination process. These experimental results were in line with the results of a number of theoretical studies on the mechanism of L₂ZrCl₂-catalyzed hydroalumination performed by Parfenova et al. In [135], the formation of Cp₂ZrCl₂-based catalytic species was analyzed in detail. The main results of the modeling matched Scheme 15; in particular, by the optimization of the molecular structure of Zr06 that was found to be relatively stable (Figure 13). In accordance with Scheme 15, the kinetic model for the reaction steps was developed [136,137].

In [138], α-olefin interactions with catalytically active centers were studied by DFT and ab initio calculations. It was shown that Cp₂ZrHCl, Zr06, Zr08, and Zr09 are able to perform coordination and insertion of the α-olefin molecule; the activity decreases in the order Cp₂ZrHCl > Zr06 > Zr08 > Zr09. The results of the modeling correlate with experimental data on Cp₂ZrCl₂-catalyzed hydroalumination, considering the equilibria between Zr07 (isolated active catalyst) and Zr06 (highly reactive intermediate) and low solubility of Cp₂ZrHCl that slows down the reaction. DFT modeling of the final stage of hydroalumination (Zr → Al alkyl transfer) was conducted for Cp₂ZrCl(ⁿPr)–ⁱBu₂AlX system (X = H, Cl, ⁱBu) [139]. It was shown that the reaction of ⁱBu₂AlH needed no activation energy and proceeded through the coordination of ⁱBu₂AlH both outside and inside of the C–Zr–Cl angle, and led to the Cp₂ZrHCl⋯Al ⁱBu₂Pr association, which further dissociated into the Al ⁱBu₂Pr and AlⁱBu₂ⁿPr. Transmetallation in Cp₂ZrCl(ⁿPr)⋯ClAl ⁱBu₂ occurs via the coordination of ⁱBu₂AlCl to the inside of the C–Zr–Cl angle with the formation of a bridging complex Cp₂ZrCl(μ-Cl)(μ-ⁿPr)Al ⁱBu₂, followed by intramolecular ligand exchange that results in Cp2ZrCl₂⋯AlⁱBu₂ⁿPr association which further decomposes into AlⁱBu₂ⁿPr and Cp₂ZrCl₂. It was also shown that the transmetallation of the propyl group in Cp₂ZrCl(ⁿPr)⋯AlⁱBu₃ is less probable due to the high activation barrier (ΔG^≠ = 31.9 kcal·mol⁻¹). In this way, the transmetallation rates of Cp₂ZrCl(ⁿPr) under the action of ⁱBu₂AlX decrease in the order ⁱBu₂AlCl > ⁱBu₂AlH > ⁱBu₃Al [139].

Current views on the mechanism of the reaction of L₂ZrCl₂ with ⁱBu₃Al were substantially expanded in the recent study by Conley et al. [140]. When studying reactions of [CH₂CH₂(η⁵-C₉H₆)₂]ZrCl₂ with 12 eq. of organoaluminum compound, they detected the formation of Zr–(μ-Cl)–Al complex Zr10, presumably through β-Me elimination/carboalumination stages (Scheme 16). Notably, the π-complex in this Scheme represents the Zr(II) complex.

Cp₂ZrCl₂-catalyzed carboalumination was discovered by Van Horn and Negishi [141]. This reaction was also considered in the review by Parfenova [18] and in the more recent review of Negishi [142]. In most studies on catalytic carboalumination, Me₃Al and (n-Alkyl)₃Al have been used to avoid β-hydride elimination. Cp₂ZrCl₂ is not an efficient carboalumination catalyst (Scheme 17) [143,144]. As shown by Parfenova et al., the efficiency of carboalumination can be increased by the rational design of η⁵-ligands in the L₂ZrCl₂ pre-catalyst [145].

When using Me₃Al as an activator, both carboalumination and hydroalumination processes can take place in the presence of other L₂ZrCl₂ complexes. In 2018, Parfenova et al. [126] studied the ligand exchange in 15 different zirconocenes, and the oct-1-ene reactivity of 11 of them (Scheme 18a) for L₂ZrCl₂–Me₃Al systems. Based on the results of NMR spectral studies and catalytic experiments, a mechanism of the L₂ZrCl₂-catalyzed reaction of alkenes with Me₃Al was proposed (Scheme 18b). In the first step, complex 1 reacted with Me₃Al to yield L₂ZrClMe (5). The subsequent reaction of complex 5 with Me₃Al yielded intermediate 7 (Zr01, see above), in which the Zr–Me bond is more polarized due to associations with electron-deficient Me₃Al. As shown above, complex Zr01 exists in the associated state at temperatures below 230 K. At room temperature, it dissociates to Cp₂ZrMeCl and Me₃Al and its concentration becomes negligibly low; thus, the alkene does not react. Switching to the catalytic reaction shifts the equilibrium towards Zr01, and products 8–11 are formed in the system.

Notably, Cp₂ZrMe₂ demonstrated low catalytic activity in the reaction with oct-1-ene and Me₃Al. Complex 6 is not formed in the reaction of Cp₂ZrCl₂ with Me₃Al; thus, a key role is played by Zr01-type intermediates. The degree of association of L₂ZrMeCl with Me₃Al also depends on the π-ligand environment of Zr and the nature of the solvent. In this way, realization of the process presented in Scheme 18b requires the presence of Cl in the reaction mixture. In other words, the presence of the Zr–(μ-Cl)–Al structural fragment in key catalytic species and the alkene insertion would, most likely, be accelerated in the L₂ZrMe(μ-Cl)AlMe₃ active complex.

Catalytic reactions were conducted in two different solvents (CH₂Cl₂ and toluene) using a L₂ZrCl₂/alkene/AlMe₃ ratio of 1:50:60. The results of the experiments are presented in

Table 1. Conversion and product yields in the reaction of the oct-1-ene with Me₃Al in the presence of 2 mol% L₂ZrCl₂ (L₂ZrCl₂/alkene/Me₃Al = 1:50:60, 20 °C, 24 h) [126].

L2ZrCl2	Solvent	Conversion, %	Product Yield, %
			8	9	10	11
1a	CH2Cl2	92	3	14	7	68
	toluene	69	3	21	7	38
1b	CH2Cl2	84	11	14	7	52
	toluene	39	9	9	9	12
1c	CH2Cl2	68	53	8	7	-
	toluene	44	15	14	14	1
1d	CH2Cl2	24	<1	5	5	13
	toluene	13	<1	4	6	2
1i	CH2Cl2	89	31	19	-	39
	toluene	70	38	14	10	8
1j	CH2Cl2	0	-	-	-	-
	toluene	11	9	-	-	2
1k	CH2Cl2	2	0.6	0.3	0.7
	toluene	4	<1	1.6	1	1
1l	CH2Cl2	30	20	1	<1	9
	toluene	99	63	2	-	34
1m	CH2Cl2	16	15	-	<1	<1
	toluene	26	22	-	1	3
1n	CH2Cl2	3	1	-	-	-
	toluene	35 1	13	-	-	-
1o	CH2Cl2	17	16	<1	<1	-
	toluene	2	46	2	<1	4
Cp2ZrMe2	CH2Cl2	16	6	4	5	1
	toluene	9	3	3	<1	<1

¹ Dimers and oligomers formed.

.

The reaction between Et₃Al and Cp₂ZrCl₂ has a complex mechanism, which occurs in several elementary steps [146]. When using 1 eq. of Et₃Al, the Zr–(μ-Cl)–Al complex Zr11 is formed. With an excess of Et₃Al, C–H activation takes place to yield a relatively unstable species, Zr12, which is subsequently converted to a more stable Zr13 via the secondary C–H activation along with a smaller amount of Zr14 (Scheme 19). The molecular structure of Zr13 was determined by XRD [147] (Figure 14). In the presence of α-olefin, the highly reactive species Zr12 is further transformed into a cycloalumination product [148]. The mechanism of this reaction was studied by DFT modeling, which confirmed the importance of Zr–(μ-Cl)–Al bonding for all key reaction stages [149,150,151,152].

An interesting complex, Zr15 with a Zr–(μ-Cl)–Al fragment, was formed as a result of the reaction of the Cp₂Zr buta-1,3-diene complex with Me₂AlCl (Scheme 20) [153].

Zr analogs of the Tebbe reagent Cp₂Zr(μ-CHR)(μ-Cl)AlR′₂ were obtained (a) by the reaction of Cp₂Zr(Cl)CH=CHR with R′₂AlH or (b) by the reaction of Cp₂ZrHCl with R′₂AlCH=CHR [154,155]. The molecular structure of one of these complexes with R = CH₂^tBu and R′ = ⁱBu (Zr16) was proven by XRD (Figure 15). Similar complexes have not demonstrated any synthetic or catalytic prospects.

3.2. Zirconocene Cationic Complexes in Oligomerization and Polymerization of α-Olefins

Generally accepted mechanisms of the activation of zirconocene pre-catalysts L₂ZrCl₂ depend on the type of activator. When using MAO, methylation results in a L₂ZrMeCl complex, which eliminates Cl⁻ under the action of MAO. However, when using a slight excess of MAO, Me₃Al (inevitably present in MAO due to its dynamic nature), the intermediate active cation forms the inactive cationic complex Zr17, which can be reactivated when using a large excess of MAO. Notably, MAO has limited ability to activate L₂ZrCl₂; the nature of the primary η⁵-ligand environment (L₂Zr fragment) affects the rate and equilibria of the formation of the catalytic species [122,156]. In the presence of trialkylaluminum and perfluoroaryl borates, an alkylzirconocene cation is formed through the alkylation of L₂ZrCl₂ with AlR₃, followed by the elimination of one Zr–alkyl under the action of B(C₆F₅)₃ or by the alkyl abstraction or protonation of Zr–alkyl by [Ph₃C]⁺ or [PhNMe₂H]⁺ counterions of [B(C₆F₅)₄]⁻, respectively (Scheme 21).

The negative role of Cl⁻ in the zirconocene-catalyzed polymerization of α-olefins when using MAO as an activator was demonstrated by Cramail et al. in the framework of the simplified MAO model [157]; however, further studies have shown that not everything is so unambiguous, primarily because of the incompleteness of the early conceptions of the MAO structure. In recent years, this issue was clarified; essentially, the presence of Cl-containing MAO species as counter-ions was confirmed experimentally [28,125,158,159], In particular, the molecular ion with an m/z ratio of 1395 (Figure 16) was detected when analyzing the activation of Cp₂ZrCl₂ by MAO. Evidently, it is the Al–Cl fragment in this molecule that can coordinate at the Zr catalytic center. This model of ‘chlorinated’ MAO was developed on the basis of previous mass spectrometry studies [26].

In [23], Linnolahti et al. proposed a simple and visual model of the structure of an MAO cluster, suitable for DFT modeling. They demonstrated that the chlorination of MAO leads to the overall facilitation of catalyst activation processes.

There are no examples of cationic catalytically active zirconocene species with Zr–(μ-Cl)–Al fragments whose molecular structures have been proven by XRD. It is worth pointing out here that cationic Zr complexes with [MeAl(2-C₆F₅C₆F₄)₃]⁻ (Zr18) [160] and [FAl(2-C₆F₅C₆F₄)₃]⁻ ions (Zr19) [22] have been separated and analyzed by XRD. As can be seen in Figure 17, the differences between Zr–(μ-X) and Al –(μ-X) are equal to 0.48 and 0.33 Å for X = C and X = F, respectively. Whether Al(C₆F₅)₃ is capable of Cl abstraction from L₂CpCl(Alkyl) complexes is now an open question.

In recent publications [161,162], the mechanism of the activation of L₂ZrCl₂ pre-catalysts by MAO was significantly revised with new experimental data that point to the intermediate formation of [L₂Zr(μ-Cl)₂AlR₂]⁺ cationic species via the generation of R₂Al⁺, followed by their reaction with L₂ZrCl₂. At the same time, an alternative activation pathway was proposed for the reaction of L₂ZrCl₂ with [ⁱBu₂Al][B(C₆F₅)₄], namely, Cl⁻ abstraction with a formation of L₂ZrCl⁺ species [33].

Notably, the cationic complex {rac-[Me₂Si(η⁵-C₉H₆)₂]Zr(μ-Cl)₂AlⁱBu₂}⁺ (Zr20) was formed by the reaction of {rac-[Me₂Si(η⁵-C₉H₆)₂]Zr(μ-H)₃(AlⁱBu)₂}⁺ and characterized by NMR spectroscopy [163]. The complex {rac-[Me₂Si(η⁵-C₉H₆)₂]Zr(μ-Cl)₂AlMe₂}⁺ (Zr21) was similarly obtained by the reaction of {rac-[Me₂Si(η⁵-C₉H₆)₂]Zr(μ-Cl)₂AlMe₂}⁺ with Me₃AlCl [164]. XRD studies of type [L₂Zr(μ-Cl)₂AlR₂]⁺ complexes are mentioned in [33], referring to conference papers [165,166]; however, the Cambridge Crystallographic Data Centre (CCDC) has no information about these compounds. The structure of {rac-[Me₂Si(η⁵-C₉H₆)₂]Zr(μ-Cl)₂AlMe₂}⁺ was proven by a combination of chemical modification and XRD analysis of the Zr(III) neutral complex rac-[Me₂Si(η⁵-C₉H₆)₂]Zr(μ-Cl)₂AlMe₂ (Zr22) obtained by the reduction of Zr21 (Figure 18). The complexes Zr20–Zr22 were inactive in α-olefin polymerization.

In the presence of [Ph₃C][B(C₆F₅)₄], the activation of L₂ZrCl₂ by ⁱBu₂AlH results in the formation of complex cationic hydrides L₂Zr(μ-H)₃(AlⁱBu₂)₂⁺ via a L₂ZrCl(μ-H)₂AlⁱBu₂ intermediate [163]. However, the reaction of L₂ZrCl₂ (L₂ = Ph₂C(Cp)(Flu)) with TIBA and a [PhNMe₂H][B(C₆F₅)₄] activator proceeds in a complex pathway, allegedly involving Zr–(μ-Cl)–Al species (Scheme 22) [128].

Zirconocene-catalyzed selective dimerization of α-olefins (Scheme 23) represents an interesting and practically important catalytic process, for which the participation of Zr–(μ-Cl)–Al is quite possible. Discovered in 1987 by Slaugh and Schoenthal [167], and first published in scientific periodicals by Christoffers and Bergman [37,38], this reaction was further studied by Erker [168], Janiak [169,170], Kissin [171,172], and Longo [173]. The main dimer of undec-1-ene was obtained by Samela et al. using MAO-activated Cp₂ZrCl₂ [174], but the reaction products were mistakenly recognized as undec-1-ene trimers. Methylenealkanes can be used in the synthesis of single-component poly-α-olefin oil basestocks [171,175,176] and other value-added chemicals [169,171,172,177,178,179,180].

By the mid-2010s, only Cp₂ZrCl₂ [37,38] and [Me₂Si(η⁵-C₅H₄)₂]ZrCl₂ [173] were efficiently used as pre-catalysts in selective α-olefin dimerization. In comparative studies of a series of Zr complexes, activated sequentially by ⁱBu₃Al and MAO, in the dimerization of hex-1-ene (Scheme 23b) Nifant’ev et al. established the formula of complex Zr23 [39,181] that demonstrated high selectivity and activity in the dimerization of different α-olefins, including sterically hindered substrates [40]. Notably, the selectivity of dimerization increased with a decrease in the dihedral angle between cyclopentadienyl rings. During these experimental studies, increases in the selectivity of dimerization when adding R₂AlCl were detected. A similar ‘chlorine effect’ was previously demonstrated by the chemists Idemitsu Kosan [182,183], and later by Parfenova et al. [184]. As was shown by Parfenova’s group, Zr₂ bimetallic species are also active in selective dimerization [185]; however, the use of these in situ forming species is not of interest in practice.

The ‘chlorine effect’ was also noted by Christoffers and Bergman [38]; they suggested the retention of Cl at the Zr catalytic center (Cp₂ZrR⁺⋯ClMAO⁻ species) on the one hand, but on the other, proposed Cp₂ZrH⁺ as the actual catalyst. Based on experimental results, Nifant’ev and Ivchenko proposed a new mechanistic hypothesis, expanding the scope of the conventional cationic mechanism of the zirconocene-catalyzed oligomerization of α-olefins [39,40]. This hypothesis implies the participation of Zr–Al catalytic species Zr24 and Zr25 capable of reversible insertion of the α-olefin molecule. After the second α-olefin insertion, the reversible coordination of R₂AlX fragment can facilitate irreversible β-hydride elimination with the formation of methylenealkanes that are inert towards Zr24 and Zr25 (Scheme 24). It was also supposed that R₂AlCl should demonstrate the best efficiency as a selective ‘limiter’ of the degree of polymerization (DP_n) to 2.

These hypotheses were supported by additional experiments and DFT calculations [41,42]. In [41], oligomerization of hex-1-ene with the use of Cp₂ZrCl₂ pre-catalyst was studied experimentally, and DFT optimizations for all possible reaction pathways of prop-1-ene oligomerization with and without the involvement of R₂AlX (R = Me, ⁱBu; X = H, Cl, Me) were performed. The key stage of the α-olefin formation was chain termination after insertion of the second α-olefin molecule. For the simple cationic model, this process is carried out on the mechanism of β-hydride transfer to monomer, whereas with the assistance of the R₂AlX β-hydride, elimination occurs. In the latter case, Al atom demonstrates a cooperative effect (Figure 19). For R = Me, the values of the activation barriers ΔG^≠ of β-hydride elimination stage were 17.7 (H), 13.6 (Cl), and 16.2 (Me) kcal/mol. Thus, β-hydride elimination is the most affected by Me₂AlCl coordination at the Zr atom.

In the study by Nifant’ev et al. [42], oct-1-ene oligomerization with the use of pre-catalysts Cp₂ZrCl₂, Cp₂ZrMe₂, O[SiMe₂(η⁵-C₅H₄)]₂ZrCl₂ (Zr23), and its dimethyl derivative Zr23′ was studied, DFT modeling was conducted for both zirconocenes using but-1-ene as a model α-olefin. The main theoretical results with regard to Cp₂Zr-derived species were in line with the results of prior research [41], whereas in the case of Zr23-based catalytic species, additional Zr–O and Al–O interactions played a significant role in the catalytic process, stabilizing the reaction intermediates and lowering the activation barriers (Figure 20 and Figure 21).

Thus, if zirconocene-catalyzed dimerization requires the a minimal excess of MAO (less than 10 eq.) and zirconocenes of the cyclopentadienyl type (Cp₂ZrCl₂, Zr23), the catalytic oligomerization of α-olefins (Scheme 25a) is usually conducted at higher Al_MAO/Zr ratios. However, these ratios are hardly capable of the complete ‘fixation’ of R₂AlCl (MAO as a ‘sponge’ for organoaluminum compounds in the reaction mixture). On the other hand, the greater diversity of zirconocenes used in this reaction still attracts researchers’ attention [17,186,187,188,189,190,191,192]. The activation of L₂ZrCl₂ by R₃Al and perfluoroaryl borates was also found to be efficient in α-olefin oligomerization. This type of activation was used in the recent study by Nifant’ev et al., who reported that –CH₂CH₂– bridged indeno[1,2-b]indole ansa-complexes (Scheme 25b) demonstrate high efficiency in the synthesis of lightweight oligomers of dec-1-ene (DP_n = 3–5) with uniquely homogeneous molecular structures [43].

Such catalytic behavior differs from both the activity and selectivity of conventional zirconocenes; therefore, additional research was carried out. Surprisingly, the addition of ⁱBu₃Al to the solution of Zr26 did not result in Cl → ⁱBu exchange at the Zr atom. One can assume that the dissolution of Zr26 occurs through the formation of the Zr26–ⁱBu₃Al complex with weak Cl–Al coordination. Notably, a similar complexation of L₂ZrCl₂ with metal alkyls was recently discussed by Kumawat and Gupta in their study on the DFT modeling of zirconocene activation and chain transfer [124].

At the first stage of the activation, Zr26 does not exhibit substantial conversion to Zr–Al hydrides under the action of TIBA. However, after the addition of [PhNHMe₂][B(C₆F₅)₄], a rapid reaction proceeded with the sedimentation of oily low-soluble product. After the addition of dec-1-ene, no oligomerization was observed. Rapid oligomerization was detected when the activation of Zr26 by TIBA and borate were conducted in the presence of molecular hydrogen, and light-brown crystals of Zr27 were formed at the end of the reaction. After the addition of dec-1-ene to the reaction mixture, oligomerization started again. It turned out that Zr27 alone was inactive in dec-1-ene oligomerization in the presence of H₂; however, when TIBA was added to the Zr27 suspension in toluene, the rapid oligomerization of dec-1-ene occurred. The final product of the catalyst transformation of Zr27 was separated and characterized by NMR (Figure 22) and XRD analysis (Figure 23). The spectral view of Zr27 remained unchanged after 7 days, which indicates high stability of the cationic complex in the solvating solvent (THF).

The experimental fact of the ‘recovery’ of Zr–(μ-Cl)₂–Al complex Zr27 clearly indicates that the mechanism of heterocene-catalyzed oligomerization went beyond the conventional cationic mechanism of the zirconocene-catalyzed polymerization of α-olefins. There is a very realistic chance that it is the retention of Zr–Cl–Al coordination that provides structural homogeneity of dec-1-ene oligomers when using heterocene catalysis.

Comparison of the hydrogen response during ethylene polymerization, demonstrated by dichloro- and dimethyl ansa–complexes [(η⁵-C₉Me₆)SiMe₂(η⁵-C₅H₄)]ZrX₂ (X = Cl, Zr28; X = Me, Zr29), immobilized on solid MAO, may also be seen as indirect evidence of Zr–(μ-Cl)–Al bonding in catalytic species: Zr28-based catalysts turned out to be inert to molecular hydrogen [193]. The relative stability of Zr–(μ-Cl)–Al bonding in catalyst precursor, L₂Zr(R)-(μ-Cl)-AlR’₃ complexes, essentially depends on the electrophilicity of the Al atom. With the introduction of the electron acceptor fragments, e.g., when using ⁱBu₂Al(OC₆F₅) [194], an ion pair is easily formed, as evidenced by polymerization experiments.

On the basis of the latest investigations [33,43,161,162], it can be concluded that the difference in the mechanisms of the L₂ZrCl₂ + AlR₃ reactions in the absence and presence of MAO (or perfluoroaryl borates) consists of various degrees of ‘chlorination’ for the first-stage products. In the absence of MAO or borate, alkyl transfer from Al to Zr occurs with the formation of LZrR₂⋯ClAlR₂ complexes with weak Zr⋯Cl bonds (if R = ⁱBu, the transformation to Zr–Al hydrides results in heterometallic complexes with stronger Zr–(μ-H)–Al bonds). When using MAO or perfluoroaryl borates, resulting R₂Al⁺ species rapidly react with LZrCl₂ to form cationic [LZr(μ-Cl)₂AlR₂]⁺, thus creating certain preconditions for the involvement of R₂AlCl in further reactions with α-olefins.

4. Complexes of V

Vanadium-catalyzed α-olefin polymerization was previously reviewed by van Koten et al. in 2002 [195]; by Gambarotta in 2003 [196]; by Redshaw in 2010 [197]; by Wu and Li in 2011 [198]; by Nomura and Zhang in the same year [199]; and by Phillips et al. in 2020 [44]. In these reviews, the focus was on different types of the coordination compounds of V, pre-catalysts of α-olefin polymerization. The mechanisms of polymerization in these reviews were fragmented; however, in contrast to reviews on group 4 metal Ziegler–Natta and single-site catalysts, the importance of V–(μ-Cl)–Al bonding in pre-catalysts and in catalytic species was not ignored (Scheme 26). Here, in strong contrast to Ti- and Zr-based pre-catalysts, a large number of V-based catalysts are effectively activated by R₂AlCl of RAlCl₂, as opposed to ineffective R₃Al and MAO. In this section, we discuss several examples of V–(μ-Cl)–Al complexes that have been missed in previous reviews, or that had been served poorly, as well as new examples of these complexes.

Among older publications on the subject of V-catalyzed polymerizations of α-olefins, the feature article of Zambelli et al. [200] merits special attention. The authors reinvestigated three models of catalytic complexes V05–V07 proposed in the literature for VCl₃/R₂AlCl catalyst systems (Scheme 27). DFT modeling of the insertion of ethylene molecule was conducted for P = Me according to the conventional Cossee–Arlman scheme. Thermodynamic data for ethylene insertion were close for all three complexes, but significant differences were observed for values of the activation barrier of insertion that were 16, 16, and only 1.7 kcal·mol⁻¹ for V05, V06, and V07, respectively. DFT modeling of prop-1-ene insertion explained the observed syndiotacticity of prop-1-ene polymerization.

As can be seen in Scheme 26 and Scheme 27, V(III) species are active in polymerization. Obviously, the mechanisms of the deactivation of V centers are important for the development of efficient catalysts. It is thought that the most likely route for deactivation would be reactions with the organoaluminum compound. However, in 2002, Gambarotta et al. demonstrated an alternative pathway of such deactivation via disproportionation of the V(III) complex V08 with the formation of V(II)/V(III) (V09) and V(IV) (V10) species [201] (Scheme 28, Figure 24); AlCl₃ in this process acts as Lewis acid and it cannot be ruled out that organoaluminum co-catalysts can promote catalyst deactivation in a similar way.

Another important aspect of V catalysis in α-olefin and diene polymerization was demonstrated by Centore et al. [202], who studied the catalytic activity of the pre-catalyst [(κ¹-ⁱPrN=C(Ph)NⁱPr)V(O)Cl(μ-Cl)]₂ using different organoaluminum activators. They showed that the amidinate ligand can easily be removed from the vanadium in the reaction with organometallic cocatalysts, and the active species are closely related to those obtained from other vanadium precatalysts such as VCl₄ (for example, V07). This result imposes certain restrictions on the ligand-oriented design of V-based polymerization catalysts. Stronger V–ligand bonding, such as in V03 [203], provides a retention of the base ligand environment. Polyphenolate ligands can also provide stability of the catalytic species [204]; in particular, activation of the pre-catalyst V11 (Scheme 29) resulted in the formation of V(IV) complex V11′, which is active in ethylene polymerization. V(III) complexes with bidentate N,N-chelating iminopyrrolyl ligands also demonstrated stability in the base ligand environment and relatively high catalytic activity in the polymerization of ethylene when using Et₂AlCl as an activator; in the presence of Et₃Al or MAO, PE was formed in trace amounts [205]. The complexes with tridentate iminopyrrolyl and tetradentate bis(iminopyrrolyl) ligands [206], well as with bidentate phenoxy-phosphine ligands [207] and with tridentate 2,6-bis(diphenylhydroxymethyl)pyridyl ligand [208], have demonstrated similar behavior. However, in the last case, when comparing V(III) and V(V) derivatives in Et₂AlCl-activated polymerization, the V(V) complex demonstrated higher activity. Apparently, even when using the same Cl-containing activator, the nature of the active site essentially depends on the nature of the polydentate ligand.

V(III) complexes bearing salicylaldiminato ligands (similar to V04), synthesized by Lee et al. [209,210], also showed high activity when using Et₂AlCl as an activator. In their later work [211], these same authors reported the results of theoretical studies of the mechanism of ethylene polymerization with V04-type complexes. Actually, V04 was chosen for DFT calculations, together with the consideration of Et₂AlCl-free species, in the modeling of ethylene insertion in the framework of the Cossee–Arlman mechanism (Figure 25). In the first step, the ethylene molecule coordinates to the vacant site in V04, forming a π-complex (–18.1 kcal·mol⁻¹). Subsequently, ethylene inserts into the vanadium–carbon bond via the four-membered cyclic transition state which has an activation energy barrier of 14.7 kcal·mol⁻¹ relative to the π-complex. The overall ethylene insertion was found to be highly exothermic (–27.1 kcal/mol), and insertion product resembled the starting structure of V04. For hypothetical cationic active species [(PhN=CHC₆H₄O)VEt(THF)]⁺, ethylene complexation was less exothermic and the insertion barrier was higher by 5.7 kcal/mol. An additional argument in favor of the V–(μ-Cl)₂–Al model was a clear correlation between the results of modeling and polymerization experiments for V(III) complexes with substituted salicylaldiminato ligands.

The results of relatively recent studies on V-catalyzed polymerization also confirm a distinct ‘chlorine effect’. In this regard, the study by Białek and Bisz [212] deserves a separate mention. In the polymerization of ethylene, ONNO-type bis(phenolates) V12 (Scheme 30) were manyfold more active in the presence of Et₂AlCl in comparison with MAO and perfluoroaryl borates. In ethylene/oct-1-ene copolymerization, the difference in activities decreased when using Et₂AlCl and ³Bu₃Al/[Ph₃C][B(C₆F₅)₄], but higher comonomer incorporation was observed in the first case.

In 2018 [213], Talsi et al. reported the results of the study of α-diimine (V13) and bis(imino)pyridine (V14) trichloro complexes of V(III) (Scheme 30) with the use of MAO, Me₂AlCl, Me₂AlCl/[Ph₃C][B(C₆F₅)₄], and Me₃Al/[Ph₃C][B(C₆F₅)₄] activators. Through careful NMR experimentation, they demonstrated the formation of V13-based cationic complexes without V–(μ-Cl)–Al bonding, whereas V14 formed V(Me)–(μ-Cl)₂–Al species V14′. Notably, when Me₂AlCl alone was used, an inactive V(Cl)–(μ-Cl)₂–Al complex was the main reaction product.

In a series of recent publications [214,215,216], Nomura et al. presented the results of the study of the activation of V complexes with the use of V K-edge X-ray absorption near-edge structure analysis (XANES) and extended X-ray absorption fine structure (EXAFS) analysis. Regarding V(V) pre-catalysts of (imido)VCl₂(OAr), (imido)VCl₃, and similar types, as well as VOCl₃, these studies clearly indicated the presence of V(III) species in the reaction mixture after activation by the organoaluminum compounds Me₂AlCl, Et₂AlCl, and EtAlCl₂. In this way, the attribution of V03 (Scheme 26) to active species should be corrected. However, no significant changes in either the oxidation state or the basic geometry were observed when (imido)vanadium(V) complexes were treated with MAO. In this way, the role of R₂AlCl is in both the formation and stabilization of catalytic species.

In conclusion, it should be noted that the chemistry of V-based single-site catalysts of α-olefin and diene polymerization primarily focuses on post-metallocene-type complexes. V-based catalysts exhibit some similarities to Ti(III) Ziegler–Natta catalysts (see V07 in Scheme 27); however, half-sandwich and sandwich complexes of V have not found wide application. What is more interesting is the possible role of V–(μ-Cl)–Al bonding in the catalytic behavior of Cp-ligated complexes of vanadium. Recent qualitative research of the catalytic activity of supported Cp₂V, carried out by Liu et al. [217], sheds some light on the issue. The authors suggested the formation of cationic CpV–O–Si(μ-O)₃(silica) species under the action of Et₂AlCl, and proposed a non-trivial mechanism of ethylene coordination/insertion with a marked V–Al cooperative effect (Figure 26). The stages of chain initiation and release were also studied. The activation barrier of the chain propagation was found to be ~16 kcal·mol⁻¹.

5. Complexes of Cr

5.1. Selective Oligomerization of Ethylene Catalyzed by Cr Complexes

The selective oligomerization of ethylene with the use of Cr catalysts, resulting in the formation of hex-1-ene and oct-1-ene, is becoming an important contemporary industrial process [14]. The mechanism of the selective trimerization of ethylene [13,47,48,218] qualitatively differs from the Cossee–Arlman mechanism of polymerization and non-selective oligomerization of α-olefins by the coordination of two olefin molecules and the intermediate formation of metallacyclic species. As a result, when using selective trimerization catalysts, isomeric decenes are formed [219,220] (Scheme 31). The reaction intermediates and products, presented in this scheme, only provide a general idea of the Cr-catalyzed trimerization. Cr catalysts, unlike half-sandwich Ti complexes (Scheme 10), are of a more complex nature, selective tetramerization is possible, and the oxidation numbers of Cr reaction intermediates still are subject of discussion.

The Cr-catalyzed selective tri- and tetramerization of ethylene has been the subject of numerous reviews which have discussed both the practical and theoretical aspects of the reaction [13,14,47,48,84,221,222,223,224,225,226,227]. Pyrrole-based Cr catalysts were historically the first system for the synthesis of hex-1-ene [228], which is widely used in the petrochemical industry. Other groups of homogeneous Cr catalysts of tri-/tetramerization often outperform Cr/pyrrole systems on catalytic activity, but research is still ongoing.

At the very beginning of the study of Cr oligomerization catalysts, a substantial ‘chlorine effect’ was detected: Cr/pyrrole systems only demonstrated high activity in the presence of Et₂AlCl/Et₃Al mixtures of the activators. This fact was explained successfully from a mechanistic point of view (see Section 5.2 below), and the possible importance of Cr–(μ-Cl)–Al bonding in Cr-based catalytic species was not completely ignored during the ligand-oriented design of new single-site oligomerization catalysts, accompanied by oligomerization experiments.

5.2. Cr Complexes with Pyrrole and Similar Ligands

Discovered in the late 1980s by Reagan [228], catalytic systems containing Cr(III) 2-ethylhexanoate, 2,5-dimethyl-1H-pyrrole, Et₂AlCl, and Et₃Al were optimized to a 1:3:8:11 ratio during the further research by Phillips Petroleum and Mitsubishi [229,230,231,232,233]. This catalyst is commonly known as the Chevron–Phillips ethylene trimerization system. In augmenting the main idea of the metallacyclic mechanism, the scientists of Sasol Technology proposed a novel mechanistic concept, which included the formation of Cr–(μ-Cl)–Al catalytic species [234]. They suggested that the Cr(II) center coordinates two ethylene molecules. Then, a Cr(IV) metallacycle is formed, subsequent π-coordination/insertion of the ethylene molecule results in a seven-membered metallacycle which is further subjected to reductive eliminative intramolecular β-hydrogen migration to the ζ-carbon followed by the coordination of two ethylene molecules with hex-1-ene release (Scheme 32).

During DFT modeling, unsubstituted pyrrole was used as a ligand, with the consideration of possible η⁵- and κ¹-coordination. In addition, optimizations were performed for catalytic species with and without Me₃Al coordination (‘Cl’ and ‘Me₃AlCl’ models, respectively). Based on the results of the modeling, the triplet spin states for both Cr(II) and Cr(IV) were predicted to be the ground state. When comparing Cl and Me₃AlCl models, significant lowering of the activation energy of the rate-limiting step by 11.3 kcal·mol⁻¹ was found for the catalytic species with Cr–(μ-Cl)–Al bonding.

In follow-up studies of pyrrole-based catalytic systems, Gambarotta, Duchateau et al., after failed attempts to isolate single crystals of 2,5-dimethyl-1H-pyrrole derivatives, obtained characterizable complexes of 2,3,4,5-tetrahydro-1H-carbazole [235]. In particular, by the treatment of [CrCl₃(THF)₃] or [CrCl₂(THF)₂] with a mixture of the ligand and Et₃Al, the square-planar Cr(II) complex Cr01 was obtained as a blue paramagnetic crystals. When chromium(III) 2-ethylhexanoate was used, the presence of AlEt₂Cl was crucial, and the reaction afforded the new paramagnetic Cr(I) complex Cr02. Complex Cr02 was also obtained by the reduction of Cr01 using potassium (Scheme 33). The molecular structures of Cr01 and Cr02 are presented in Figure 27.

Complex Cr01 catalyzed the polymerization of ethylene. In methylyclohexane, complex Cr02 was an unprecedented single-component trimerization catalyst, producing hex-1-ene with only trace amounts of higher oligomers. The structure of the possible active complex Cr03 is also presented in Scheme 33. The role of Cr–(μ-Cl)–Al bonds in the formation of the ligand environment of the catalytic center in Cr03 seems clear.

One year later, the same research team reported the synthesis of the 2,3,4,5-tetramethyl-1H-pyrrole-based Cr–(μ-Cl)–Al complex Cr04 [236] (Scheme 34). Similarly to complex Cr02, in methylyclohexane, complex Cr04 turned out to be the highly active ethylene trimerization catalyst without any activators. Apparently, the dimer of Cr04 (Figure 28) dissociates in the solution with the retention of Cr–(μ-Cl)–Al coordination in the monomeric complex with an easily recognizable ‘constrained geometry’ structural motif.

Taking into account the results of theoretical [234] and experimental [235,236] studies of Cr/pyrrole systems, Budzelaar offered his own mechanistic concept, based on the DFT modeling of Cr/1H-indole catalytic systems [237]. The modeling results indicated that, in addition to the Cr(II)/Cr(IV) cycle (Scheme 33), Cr(I)/Cr(III) cycle, based on [(Indol-1-yl)⋯AlMe₂(μ-Cl)Cr] species (Cr05), the role of the Cl atom was to stabilize metallacyclic intermediates and to block additional coordination of the ethylene molecule at the stage of hex-1-ene formation.

Finally, to determine the problem of the oxidation states of the Cr catalytic center in the Chevron–Phillips ethylene trimerization system, Liu et al. carried out a detailed theoretical study of the model catalyst species Cr06 and Cr07 (Figure 29) based on 2,5-dimethyl-1H-pyrrole [238]. Optimizations showed that the retention of Cr–(μ-Cl)–Al bonding (strong or weak, d(Cr–Cl)) was changed in the interval of 2.5–4.0Å in all key reaction intermediates and transition states. Calculated free energy profiles (Figure 29) clearly showed the preference of the choice of Cr06 as a model catalytic species—free activation energies were 19.0 and 31.4 kcal·mol⁻¹ for Cr06 and Cr07-based reaction pathways, respectively. In this way, the mechanistic concept of Cr(I)/Cr(III) catalytic cycle received additional confirmation.

Recently, the activation of the Chevron–Phillips ethylene trimerization system was studied by Tromp et al. using catalytic and spectroscopic (XAS, EPR, UV–vis) experiments under industrial conditions [239]. It was found that 2,5-dimethyl-1H-pyrrole reacts with Et₂AlCl и Et₃Al with the formation of [2,5-dimethyl-2H-pyrrole]⋯AlEt₃ and [μ-2,5-dimethyl-1H-pyrrole](μ-Cl)(AlEt₂)₂ complexes, and their reaction with Cr(III) 2-ethylhexanoate results in polymeric Cr(II) pyrrolyl species containing Et₂AlCl fragments. However, the structure of real catalytic species remains unknown.

5.3. Cr Aminodiphosphine Complexes

Among others, Cr(III) derivatives of chelating ligands of the formula R¹R²P–N(R³)–PR⁴R⁵ (PNP–Cr complexes) represent pre-catalysts of the selective tri-/tetramerization of ethylene which are characterized by high catalytic activity and stability over time [14,47,84,221,222,223,224,225,240]. Usually, the activation of PNP–Cr pre-catalysts is performed by the reaction of L^PNPCrCl₃ complexes (L^PNP—PNP ligand) or mixtures of L^PNP with Cr(III) salts by MAO. The use of R₃Al/perfluoroaryl borate systems is rarely used [241] despite its undoubted merits. Evidently, the presence of R₂AlCl species, arising as a result of the reaction between L^PNPZrCl₃ and organoaluminum activators, implies the possible formation of Cr–(μ-Cl)–Al species. The results of studies of PNP–Cr catalysts in view of this possibility are summarized and discussed below in chronological order.

When studying the activation of [CyN(PPh₂)₂]CrCl₃ by Me₃Al in toluene (Cy—cyclohexyl, Scheme 35), Gambarotta, Duchateau et al. separated and characterized a cationic complex of Cr(II), Cr08 [242]. L^PNP does not react with CrCl₂(THF)₂; nevertheless, in the presence of excess of Me₃Al, Cr08 was obtained in a high yield. XRD studies confirmed the molecular structure of Cr08 (Figure 30). Being activated by MAO, this complex was highly active in the oligomerization of ethylene with the formation of oct-1-ene as the main reaction product.

The authors also proposed that the catalytic behavior of Cr08 and the failure of L^PNP to ligate to a CrCl₂ moiety in the absence of Me₃Al suggest that the acquisition of a second ligand and cationization are central to the stabilization of a Cr(II) catalyst precursor. Considering that excess MAO is necessary for the activation, they also speculated that the dicationic [L^PNP₂Cr]²⁺ complex may be the actual catalytically active species. However, other hypotheses, involving the possibility of further abstraction of the [PNP] ligand from the ‘dormant’ catalyst precursor Cr08, cannot be ruled out [242]. For our part, we will merely add that in the latter case, the Cr(II) cation becomes open for both the π-coordination of ethylene and for Cr–(μ-Cl)–Al bonding with organoaluminum components of the catalytic system.

PNP–Cr complexes containing diphosphine ligands with two amine groups at one of the phosphorus atoms (L^PNPN) are of interest to researchers due to the ability of similar pre-catalysts to be activated by Et₃Al, without the use of MAO [243]. Evidently, such unconventional reactivity forces the assumption that L^PNPN can form catalytic species which differ from L^PNP-based species. In [244], Müller et al. reported a study of the interaction of Et₃Al and Me₃Al with the Ph₂PN(ⁱPr)P(Ph)NHⁱPr ligand. During this reaction, rearrangement to the [ⁱPrNP(Ph)P(Ph)₂=NⁱPr ligand occurred. The complex of this ligand with CrCl₃ was inactive in oligomerization. After preparation of the active model pre-catalyst [Ph₂PN(ⁱPr)P(Ph)NⁱPr]Cr(Cp)Cl, it was concluded that the active catalytic species Cr09 exhibits a binuclear nature (Scheme 36). During further research [245], it was shown that the PNPN–Cr system, generated by the reaction of Cr(III) acetylacetonate (Cr(acac)₃) with L^PNP and Et₃Al, is inactive in oligomerization. At the same time, the addition of Cl-containing compounds of various nature resulted in the selective trimerization of ethylene.

Attempts to isolate PNPN-based Cr–Al species have been successful; Cr₂Al₂ complexes Cr10–Cr13 (Scheme 37) were obtained by the reaction of L^PNPNCrCl₃ with Et₃Al, or by the interaction of Al derivatives of L^PNPN with CrCl₂(THF)₂ [246]. Molecular structures of all complexes were determined by XRD (see Figure 31 for an example). It was found that Cr10 is inactive in ethylene oligomerization, indicating that Cr10 is not a self-activating complex and that the oxidation state of Cr(II) is not sufficiently low to form hex-1-ene. Hence, the formation of Cr09-type catalytic species needs dissociation and further reduction.

However, the studies listed above only contain suggestions about the nature of catalytic species in the systems containing PNP–Cr complexes and R₂AlCl. In 2016, Evans et al. reported the results of a study of [ⁱPrN(PPh₂)₂]CrCl₃ activation by Me₃Al [247]. The Cr K-edge XAFS spectrum after 1 min indicated the formation of [ⁱPrN(PPh₂)₂]CrClMe(μ-Cl)AlMe₃(THF) and then [ⁱPrN(PPh₂)₂]CrMe(μ-Cl)AlMe₃. Such species are evidently able to form Cr(II) cations under the action of MAO; however, given that PNP–Cr catalysts usually operate under relatively low Al_MAO/Cr ratios (~100), in our view, the possibility of the intermediate coordination of R₂AlCl at the Cr center seems possible.

When studying the use of perfluoroaryl borates for the activation of PNP–Cr pre-catalysts, Lee et al. proposed an original method for the synthesis of Zr–Al dicationic complex Cr14 [241] (Scheme 38). This complex was found to be highly active in the selective tetramerization of ethylene, but a fairly large amount of PE (1.7 wt.%) was concomitantly generated. Further ligand design (introduction of SiR₃ substituents in p-positions of the Ph rings in L^PNP) resulted in an increase in the activity and suppression of the PE formation.

Summarizing the above-mentioned research, in view of the huge number of the articles on the PNP–Cr-catalyzed tri/tetramerization of ethylene, the role of R₂AlCl⋯Cr bonding was not examined in most studies. In many cases, this was due to the methodology of the experiment on pre-catalyst preparation and activation, based on the mixing of a soluble Cr source (for example, Cr(acac)₃) and L^PNP, followed by treatment with MAO. This, among other things, may be due to the low solubility and varied composition of the commonly used starting complex CrCl₃(THF)₃, which complicates the separation and purification of L^PNPCrCl₃ pre-catalysts (for instance, we have been unable to solve another similar problem, presented in [248]). The complex [CrCl₂(μ-Cl)(THF)₂]₂, recently synthesized by Lee et al. [249], seems to be a more convenient and reliable starting compound for the synthesis of L^PNPCrCl₃. We can assume that more studies on the effect of the use of [CrCl₂(μ-Cl)(THF)₂]₂, or additional amounts of R₂AlCl, on catalytic activity of known ‘chlorine-free’ PNP–Cr systems, may lead to further improvements in the PNP-based catalysts of ethylene tri- and tetramerization.

5.4. Other Cr Complexes in the Single-Site Catalysis of Oligomerization and Polymerization

The range of chelating ligands for the synthesis of Cr-based oligomerization and polymerization of single-site catalysts is not limited by PNP-type compounds. Thus, for example, complex Cr15 was synthesized by the reaction of the ^tBuNPN^tBu dianion with Cr chlorides, followed by treatment with ⁱBu₃Al [250] (Figure 32). This complex was found to be a single-component catalyst of ethylene polymerization.

HN(CH₂CH₂PPh₂)- and HN(CH₂CH₂SR)₂-based Cr(II) and Cr(III) chloro complexes were synthesized by McGuinness et al. [251]: the treatment of [HN(CH₂CH₂PPh₂)]CrCl₂ by 1,4-diazabicyclo[2.2.2]octane (DABCO) resulted in dimer {[N(CH₂CH₂PPh₂)]Cr(μ-Cl)}₂ and its –SR analogs. Further studies of the activation of these complexes by MAO and Et₃Al/B(C₆F₅)₃ indicated that the Cr(II)/Cr(IV) cycle is preferable for this type of catalyst. When HN(CH₂CH₂SCy)₂ was treated with CrCl₂(THF)₂ and EtAlCl₂, complex Cr16 was obtained (Figure 33) [252]. This Cr(II) complex, after activation by MAO, was inferior to SNS Cr(III) complexes in catalytic activity in the trimerization of ethylene.

In 2011, the SNS ligand was modified by replacing CH₂ fragments near the N atom with SiMe₂ groups (HN(CH₂CH₂SR)₂, R = Cy, ^tBu, Ph). The reaction of HN(CH₂CH₂SCy)₂ with Et₂AlCl and Cr(III) or Cr(II) chlorides afforded the Cr(II) complex Cr17 (Figure 34) [253]. In the presence of MAO (500–1000 eq.), this complex catalyzed the oligomerization of ethylene (mainly trimerization). After abstraction of the N–H proton during the synthesis of the AlMe₂ analog of Cr17, complex Cr18 was obtained (Figure 35): its selectivity in the oligomerization of ethylene was even lower.

A similar pattern was observed when studying Cr complexes with tris- and bis-pyrazolyl ligands [254] (Scheme 39). Depending on the absence or the presence of the –NH– fragment in the bridging ligand, the reaction of LCrCl₃ with Me₃Al resulted in the formation of Cr–(μ-Cl)–Al complexes with different types of Al–N bonding. After treatment with 200 eq. MAO, the activities of LCrCl₃ and Cr–(μ-Cl)–Al complexes (for example, Cr19, Figure 36) in the oligomerization of ethylene were close, with the same oligomer distributions. In this way, under the action of MAO, organoaluminum chloride is eliminated finally and irreversibly.

In the early 2010s, a number of articles were devoted to the preparation and catalytic studies of Cr complexes with bi- and polydentate ligands of different nature, including the synthesis and characterization of Cr–(μ-Cl)–Al complexes: in particular, Cr derivatives of Ph₂PN^tBu [255], (Ph₂P)₂CHCO₂⁻ and (Ph₂P)₂C=C(NHR)O⁻ [256], 2,6-bis(CH₂PPh₂)pyridine [257], 2,6-bis(NH=PPh₂)pyridine [258], 2-(NHCH₂PPh₂)pyridine [259], and Ph₂C(1H-pyrrol-2-yl) [260]. In several studies, an explicit ‘chlorine effect’ was detected, although most of them included XRD data for isolated Cr–(μ-Cl)–Al molecules (Figure 37). However, as opposed to the relatively well-studied Chevron–Phillips ethylene trimerization system and its analogs, we are not even close to understanding the role of Cr–(μ-Cl)–Al bonding in the catalytic chemistry of Cr-based post-metallocenes.

6. Complexes of Ni

Complexes of Ni are intensively being studied as efficient catalysts for the coordination oligomerization (SHOP process and beyond [46,261,262]) and polymerization [5,263,264,265] of α-olefins. Ni-based polymerization catalysts are very different from early transition metal (Ti, Zr, and V) complexes due to the lower sensitivity of Ni centers to electron-donor fragments, thus providing the copolymerization of α-olefins with polar vinyl monomers [266,267]. Another attractive property of Ni-based catalysts is their chain walking capability [268], which enables them to obtain branched polyolefins of different architectures. The direct participation of Ni–(μ-Cl)–Al species in catalytic processes at the Ni center is a matter of discussion; however, in the case of Ni, similar additional interactions are of dubious value. We have restricted the discussion in this section to a few examples of Ni–(μ-Cl)–Al bonding suspected of being associated with α-olefin polymerization.

The mere presence of Ni–(μ-Cl)–Al bonding in coordination compounds formed during the activation of Ni(II) chloro complexes by organoaluminum has been known for a fairly long time. Thus, for example, the formation of Cy₃P(η³-allyl)Ni(μ-Cl)AlMeCl₂ (Ni01) was clearly proven by XRD analysis [269] (Figure 38).

However, EXAFS studies of the reaction mixture formed from NiCl₂(PEt₃)₂ and Me₃Al₂Cl₃ [270] showed the presence of C and P atoms in the first coordination shell of Ni; more distant shells included Al (d = 3.0 Å) and Cl (d = 4.4 Å), thus excluding Ni–(μ-Cl)–Al bonding. Similar spectral studies were continued by de Souza et al. [271], for example, with MAO-activated Ni(α-diimine)Cl₂. They showed that in the active catalyst species, the Ni(II) atom is surrounded by C and N atoms in the first shell and Cl atoms at a higher distance (~3.5 Å), and proposed that the Ni⋯Cl interaction plays an important, previously underestimated, role in the polymerization of olefins.

In a number of later studies, the catalytic activity of different complexes was studied with the use of R_xAlCl_3–x activators. Thus, for example, a number of N-(5,6,7-trihydroquinolin-8-ylidene)arylamino complexes of Ni(II) (Ni02 in Scheme 40) have demonstrated high activity in the oligomerization of ethylene with the formation of but-1-ene (100% C4, 79–98% α-C4 selectivity) [272]. Higher activities were achieved when using Me₃Al₂Cl₃ instead of MAO. The authors have limited themselves to descriptions of the experimental results, without adding mechanistic interpretations. Complex Ni03 with the 2-benzimidazolyl-N-arylquinoline-8-carboxamide ligand demonstrated lower activity and but-1-ene selectivity [273]; Et₂AlCl was the best activator. (Iminoalkyl)imidazole complex Ni04 showed similar behavior [274]. Et₂AlCl and MAO equally successfully activated 2-iminopyridyl complex Ni05 [275] and 4,5-bis(arylimino)pyrenylidene derivative Ni06 [276] (Scheme 40).

As shown in [278], when using EtAlCl₂ as an activator, ethylene oligomerization is complicated by Friedel–Crafts alkylation when using toluene as a reaction medium. Another side process with the participation of alkylaluminum chlorides is interaction with the polydentate ligand with the emergence of Ni coordination vacancies [279].

Recently, Soshnikov et al. have shown that α-diimine Ni(II) complexes under the action of R₂AlCl form the Ni(I) complex Ni07 with Ni–(μ-Cl)–Al fragments, representing the catalyst’s resting states [277]. It is quite possible that derivatives of similar species are essential to ensuring ‘living’ ethylene/α-olefin copolymerization with the formation of block-polyolefins [280].

7. Conclusions

In the present review, we have summarized and commented on the data pertaining to the participation of M–(μ-Cl)–Al bonding in the formation of transition metal complexes which are active in α-olefin (and diene) chemistry. Of particular interest were the polymerization and oligomerization of α-olefins. We hypothesize that the theory and practice of single-site polymerization and oligomerization, catalyzed by Ti, Zr, and Cr complexes, have suffered from the use of 10²–10⁴ equivalents of MAO for the activation of group 4 and 6 metal complexes in laboratory practice. In such conditions, consideration of the possibility of relatively weak M–(μ-Cl)–Al bonding makes no sense because MAO operates as an organoaluminum ‘sponge’. However, in recent years, with the growing interest in ‘low-MAO’ and ‘MAO-free’ catalytic processes, interest has re-focused on reversible M–(μ-Cl)–Al coordination on catalytic centers, as an additional factor affecting the chain propagation/chain termination balance, which appears to be reasonable.

The complex understanding of the nature of M–(μ-Cl)–Al bonds in metal complexes and reaction intermediates is significantly complicated by small amounts of fragmented data. For complexes with one bridging fragment, the (μ-Cl)–Al distance is not much different from the d(Al–Cl) value in R₂AlCl (and the M–(μ-Cl) distance is lengthened in comparison with d(M–Cl) in mononuclear chloro complexes). However, when forming two bridging fragments, the values of d((μ-Cl)–Al) and d(M–(μ-Cl)) converge in magnitude. Apparently, the former M–(μ-Cl)–Al species can be considered as reactive intermediates, whereas the latter M–(μ-Cl)(μ-X)–Al species represent dormant sites that can be separated and characterized. Obviously, lower stability of the M–(μ-Cl)–Al species complicates their identification.

Additionally, in contrast to alkyl, H and F fragments that have been established in the formation of M–X–Al catalytic species, M–Cl–Al bonding in the reaction intermediates cannot be detected by the only convenient and reliable method of analysis of diamagnetic species, i.e., NMR spectroscopy. Other spectral methods leave space for interpretation; XRD analysis has limited use. In this context, it is reasonable that quantum-chemical modeling has been used extensively in the studies of M–Cl–Al-containing molecules and processes which utilize them. Additionally, during studies of the catalytic processes, the possibility of the formation of the M–(μ-Cl)–Al species should not be ignored, at least in order to separate and characterize metal complexes exhibiting a non-trivial structure or, in the end, to formulate more efficient catalytic systems.