MAO- and Borate-Free Activating Supports for Group 4 Metallocene and Post-Metallocene Catalysts of α-Olefin Polymerization and Oligomerization

Abstract

Modern industry of advanced polyolefins extensively uses Group 4 metallocene and post-metallocene catalysts. High-throughput polyolefin technologies demand the use of heterogeneous catalysts with a given particle size and morphology, high thermal stability, and controlled productivity. Conventional Group 4 metal single-site heterogeneous catalysts require the use of high-cost methylalumoxane (MAO) or perfluoroaryl borate activators. However, a number of inorganic phases, containing highly acidic Lewis and Brønsted sites, are able to activate Group 4 metal pre-catalysts using low-cost and affordable alkylaluminums. In the present review, we gathered comprehensive information on MAO- and borate-free activating supports of different types and discussed the surface nature and chemistry of these phases, examples of their use in the polymerization of ethylene and α-olefins, and prospects of the further development for applications in the polyolefin industry.

1. Introduction

Polyolefins still remain the most valuable plastics, with an annual production of ~200 Mt [1]. Most polyolefins are manufactured using Ziegler-Natta titanium catalysts and Phillips chromium catalysts [2,3,4,5,6]. Group 4 metal single-site metallocene and post-metallocene catalysts outperform industrial catalysts on the parameters of activity and homogeneity of ethylene/α-olefin copolymers [2,3,4,7,8]. The advantages of single-site Group 4 catalysts can be primarily attributed to the nature of the catalytic species that represent cationic alkyl-metal centers in basic ligand environments, which have a high ability to π-complexation and insertion of α-olefin molecule [9] (Scheme 1a). Active sites of highly efficient Group 4 metal catalysts are stabilized by weakly coordinating anions. These anions are formed during the activation of pre-catalysts by methylalumoxane (MAO) [10,11] or by trialkylaluminums in combination with B(C₆F₅)₃ or perfluoroaryl borates [9,11,12,13] (Scheme 1b) and other, less commonly used, reagents [14,15,16,17]. High catalytic performance of single-site Group 4 metal catalysts is usually observed when using 10³–10⁴ molar equivalents of MAO; perfluoroarylborates can be applied in equal amounts [11,18]. Among trialkylaluminums, Me₃Al can act as catalyst poison, especially in the case of zirconocenes, due to the formation of dormant L₂Zr-(μ-Me)₂-AlMe₂⁺ species [19,20,21], and one of the functions of excess of MAO under homogeneous conditions is to be a ‘Me₃Al sponge’. When using zirconocenes and perfluoroaryl borates, activation mechanisms and the structure of catalytic species essentially depend on the type of R₃Al. In particular, pre-activation of LMCl₂ by ⁱBu₃Al results in the formation of heterometallic hydrides [22,23,24,25], which form cationic species during reaction with borates [26,27]; in the case of Et₃Al, metallacycles can be formed [28,29] (Scheme 1c). The type of borate activator can also have an impact on the activity, e.g., activation by [PhNMe₂H][B(C₆F₅)₄] is followed by the formation of PhNMe₂ capable of coordination at the catalytic center [20]. An important aspect of homogeneous ⁱBu₃Al/perfluoroaryl borate activation is the reaction of ⁱBu₃Al with borate. As early as 1998, catalyst deactivation caused by the reaction of [CPh₃][B(C₆F₅)₄] with ⁱBu₃Al was observed [30]; this process leads to CPh₃H, isobutylene, and highly reactive ⁱBu₂Al⁺ [31]. When using ⁱBu₃Al/[PhNHMe₂][B(C₆F₅)₄], isobutane was detected in reaction products [32]. Therefore, the mechanism of activation by ⁱBu₃Al requires considering the direct participance of cationic ⁱBu₂Al⁺ species, which were isolated and characterized recently [33].

Scheme 1. (a) Polymerization of α-olefins on Group 4 metal single-site catalysts; (b) activation of pre-catalysts; (c) neutral and cationic species formed during two-stage activation of zirconocenes by R₃Al and perfluoroaryl borates. L₂—bis(η⁵-ligand), ≠—transition state.

The use of a large excess of organoaluminum compounds often leads to hydroalumination [34,35], formation of inactive Zr–CH₂CH₂–Al species [36] (ⁱBu₃Al), and cycloalumination (Et₃Al) [37,38] side processes. The problems of side reactions during homogeneous activation can be resolved during the adaptation of Group 4 catalysts to actual polyolefin technologies that are tuned for use in gas-phase or slurry processes and require supported catalysts [39]. Heterogenization of single-site catalysts was and is a crucial challenge for the polyolefin industry, and this problem was intensively studied over the past three decades [39,40,41,42,43]. The most evident way to solve this problem is based on the reaction of MAO with active support (silica, alumina, etc.) [44,45,46,47,48,49,50,51] or on functionalization of the support by perfluoroaryl borate or aluminate fragments [52,53,54] with a formation of negatively charged particles that provide binding with catalytically active cationic species through Coulomb attraction (Scheme 2a); stabilization of the catalyst occurs through the same mechanism as for homogeneous catalysts [39,40,44]. The studies of immobilization by covalent binding of the ligand precursors with the support surface (Scheme 2b) were also carried out with varying results, this type of heterogenization also requires MAO or borate activators [39,55,56,57,58,59]. Finally, a number of research groups investigated acidic supports that can react with Group 4 metal complexes with a formation of active cationic catalytic species (Scheme 2c) [46,60]. Activating supports are particularly attractive since they do not require MAO or borates (only R₃Al are needed as alkylating agents and scavengers), thereby reducing the overall cost of the catalysts.

Scheme 2. Heterogenization of Group 4 single-site catalysts: (a) using silica/MAO (or perfluoroaryl borate) supports; (b) via immobilization of the ligand precursors; and (c) using activating supports.

The very idea of the activating supports for Group 4 metallocenes and post-metallocenes is based on early studies of the activation of homoleptic alkyl derivatives of Ti(IV) and Zr(IV) by Al₂O₃ and SiO₂; the activating effect of metal oxides was attributed to the reaction of surface–OH groups of the support (S) with MR₄ that resulted in the formation of (S)(–O)₂MR₂ species (Scheme 3a) [61,62,63,64,65,66]. Note that the chemical nature of catalytically active (≡Si–O)₂ZrBn₂ species, formed on SiO₂ surface, was confirmed experimentally just in 2013 [67]. Activation of bis(arene) Ti(0) and Zr(0) complexes by Al₂O₃ presumably proceeds via oxidative addition of a surface O–H bond to give a divalent supported species (Scheme 3b) [65,66,67,68,69].

Scheme 3. Early studies: activation of Group 4 metal alkyls (a) and bis(arene) complexes (b) by silica and alumina.

The first experimental proof of the ability of support to activate metallocenes with a formation of highly active cationic catalytic species according to Scheme 2c comes from the studies of Marks and colleagues [70,71,72]. In 1985, they investigated the reaction of (η⁵-C₅Me₅)₂ThMe₂ with dehydroxylated γ-Al₂O₃ (DA); the transfer of methyl groups from Th to surface Al sites with a formation of cationic species was detected using ¹³C cross-polarization/magic angle spinning (CP MAS) NMR [70]. In 1988, they showed that DA is an efficient activator for (η⁵-C₅H₅)₂ZrMe₂, (η⁵-C₅Me₅)₂ZrMe₂, and (η⁵-C₅Me₅)ZrMe₃ in ethylene polymerization [71]; Zr→Al migrations of Me group were also confirmed by ¹³C CP/MAS NMR experiments [65,69].

In the early 1990s, Soga and Kaminaka [73,74,75] studied the propylene polymerization activity of different zirconocene dichlorides activated by AlR₃, with SiO₂, Al₂O₃, MgCl₂, MgF₂, CaF₂, or AlF₃ supports. The absence of activity, observed for SiO₂ and MgO, was attributed to a lack of acidity of these oxides compared with the other materials. MgCl₂ was then studied as an activating support for Ti(IV)-based post-metallocenes [76,77,78,79] and zirconocenes [73,80,81]. Since unmodified silica and alumina provided only partial activation of metal complexes [82], in the current century, the focus was on sulfated metal oxides, studied mainly by the scientific group of Marks [83,84,85,86,87], and fluorinated metal oxides, studied by the scientists from Elf [88], Chevron Phillips [89,90,91,92,93,94,95,96,97,98], INEOS [99,100,101], and Total in cooperation with the scientific group of Boisson [102,103,104,105,106,107].

The use of activating supports in the catalytic chemistry of α-olefins has been discussed in a number of review papers and book chapters. The first promising results in the surface chemistry of sandwich complexes of actinides and Group 4 metals have been discussed by Marks and Eisen in the early 1990s [72,108]. Some issues regarding metallocene activation by silica [109], and the difference in activating chemistry and efficiency between silica and alumina [110], were discussed by Copéret et al. The importance of achieving supported catalyst structural uniformity was also demonstrated in a review from Wegener et al [111]. The chapters, devoted to MgCl₂-supported single-site catalysts [112] and activation of metallocenes by solid acids [113], were published in the collection ‘Tailor-Made Polymers, Via Immobilization of Alpha-Olefin Polymerization Catalysts’ (2008). In 2018, activating supports for metallocenes and related catalysts were briefly discussed by Hoff in a chapter of the ‘Handbook of Transition Metal Polymerization Catalysts’ [114]. Activating supports were also mentioned in recent reviews on supported catalysts [41,42,43,46,60].

In the present work, we summarized and analyzed published data on activating supports of different types and compared mechanisms and efficiency of Group 4 metallocene and post-metallocene activation and reactivity. The prospects of the use of activating supports in the polyolefin industry are also discussed.

2. Metal Halides

2.1. MgCl₂

MgCl₂ is a widely used support for conventional titanium–magnesium Ziegler–Natta catalysts (TMZNCs) of third and subsequent generations [115,116,117,118,119,120]. In view of the industrial importance of TMZNCs, efficient methods for controlling the particle size, porosity, and morphology of MgCl₂ have been developed, and it is not surprising that MgCl₂ was studied as a support for Group 4 metal single-site catalysts [39,112]. MgCl₂ supports may also have the advantage of easier fragmentation than silica or alumina supports, thereby facilitating polymerization and improving the morphology of polymer particles formed. Whilst TMZNCs are not the subject of our review, the role of MgCl₂ as an activating support for absorbed Ti catalytic species is not in doubt. According to current theoretical concepts about the nature of TMZNCs, pre-catalysts represent TiCl₄ adsorbed on MgCl₂ crystallites, and binding strength depends on the crystal structure of the MgCl₂ support. In their fundamental works, Cavallo and colleagues assumed the MgCl₂ bulk to be in the α crystalline phase with the surface comprising (104) and (110) lateral cuts that have Mg centers with coordination numbers (CN_Mg) of 5 and 4, respectively (Figure 1) [116,121].

Figure 1. Schematic representation of a MgCl₂ monolayer fragment developed by Cavallo et al. The (104) and (110) lateral faces are indicated. The Mg atoms are colored orange. The Cl atoms above and below the Mg plane are dark and light green colored, respectively. Reprinted with permission from [116]. Copyright (2012) American Chemical Society.

In terms of the reaction mechanism, the method of the activation, i.e., the introduction of R₃Al after or before/simultaneously with the interaction of the pre-catalyst with MgCl₂, could be essential. In the latter case, the chemistry of R₃Al on the MgCl₂ surface is an important point in examining the activation of Group 4 metal complexes. Using ²⁷Al MAS NMR spectroscopy, Potapov and colleagues have studied organoaluminum compounds Et_nAlCl_3−n supported on a highly dispersed MgCl₂, prepared by the reaction of Mg with n-BuCl [122]. In the ²⁷Al MAS NMR spectrum of the sample Et₂AlCl/MgCl₂ they detected a characteristic signal of Al species with coordination number (CN_Al) = 5. An attempt to record the ²⁷Al MAS NMR spectrum of Et₃Al/MgCl₂ failed; however, analysis of the spectrum obtained after treatment with Bu₂O allowed to propose the formation of dimeric Et₂Al(μ-Et)₂AlEt₂ species on the MgCl₂ surface (CN_Al = 5). However, the reaction of MgCl₂ with R₃Al can not be called fully studied; a lot of research on TMC activation was focused on R₃Al interactions with Ti species [123,124] without analyzing concurrent processes on the MgCl₂ surface.

In the 1990s, a number of MgCl₂-supported Group 4 metal complexes were studied in polymerization experiments after activation by alkylaluminum compounds. Kaminaka and Soga activated rac-[1,2-ethylenebis(η⁵-4,5,6,7-tetrahydroinden-1-yl)]ZrCl₂ (rac-[EBTHI]ZrCl₂, Zr1) by R₃Al (R = Me, Et) in the presence of MgCl₂ [73]. The supported catalyst was prepared by the grounding of the support and Zr1 in toluene in the presence of R₃Al. In propylene polymerization, supported catalysts were significantly inferior to homogeneous Zr1/MAO by activity, but at the same time, polypropylene (PP), formed when using a supported catalyst, had higher isotacticity. During a comparative study of (η⁵-C₅H₅)₂ZrCl₂ (Cp₂ZrCl₂, Zr2) and [Me₂C(η⁵-fluoren-9-yl)(η⁵-C₅H₄)]ZrCl₂ (Me₂C(Flu)(Cp)ZrCl₂, Zr3), activated by Me₃Al in the presence of commercial MgCl₂, atactic and syndiotactic polypropylenes were obtained; both zirconocenes have demonstrated high activities comparable with the activity of MAO-activated complexes [74]. In further studies, the mechanism of the activating of Cp₂ZrX₂ (X = Cl, Me, Et) by Zr→Mg transfer of X with a formation of Cp₂ZrX⁺ was proposed [75]. Note that with an increase of the Al:Zr ratio, catalytic activity goes through the maximum; this was attributed to R₃Al binding with the MgCl₂ surface. The binding of Cp₂Zr fragments with MgCl₂ surface is rather weak: so, for example, Ochędzan-Siodłak and Nowakowska showed that the addition of MAO to the Zr2/MgCl₂/R₃Al system results in the formation of a homogeneous catalyst [125].

Soga and colleagues have studied the catalytic behavior of CpTiCl₃ (Ti1), supported on MgCl₂ and activated by ⁱBu₃Al, in the polymerization of propylene [126]. Surprisingly, when using a MgCl₂-supported catalyst, isotactic PP was obtained. However, in a later study, Kang and colleagues reported a lack of activity when using CpMCl₃/MgCl₂ (M = Ti, Zr) pre-catalysts prepared in the presence of 2-ethylhexanol [127].

Bis- and mono-cyclopentadienyl Ti(IV) complexes Cp₂TiCl₂ (Ti2) and Ti1 were supported on MgCl₂ by physical (grinding) or chemical (in the presence of Et₃Al) methods [128]. UV-vis spectra showed coordination of Cl ligands of Ti1 and Ti2 to surface Mg Lewis acidic sites (LAS), which were more pronounced for Ti1. After activation by Et₃Al or MAO, supported catalysts were studied in propylene polymerization. For both Ti complexes, isotactic PP was obtained, and the authors attributed this fact to the formation of Ti(III) catalytic species.

The first study of the activation of Group 4 post-metallocenes using MgCl₂ and different alkylaluminums in propylene polymerization was conducted by Soga and colleagues in 1997 [129] on the example of β-diketonate L₂TiCl₂ complexes Ti3–Ti6 (Scheme 4a). Pre-catalysts were prepared by the reaction of Ti3–Ti6 with MgCl₂ in toluene, and the activation by alkylaluminums was performed in n-heptane. The polymerization rate when using Ti3 decreased in the following order: MAO > Me₃Al > Et₃Al > ⁱBu₃Al > Et₂AlCl from 300 to 17 kg∙mol⁻¹∙h⁻¹; mainly atactic PP was formed in all experiments. Ethylene/propylene copolymerization on the example of Ti6 as a pre-catalyst with and without diisopropyldimethoxysilane as an ‘external donor’ was studied the following year [76]. The nature of the catalytic species, formed by Ti3–Ti6 after activation, remained unclear, and the fact that the activating support mechanism is applicable to these catalysts did not.

Scheme 4. (a) β-Diketonate Ti(IV) complexes studied in propylene polymerization after activation by R₃Al/MgCl₂ [129]; (b) FI catalysts studied in ethylene polymerization using Et₃Al/MgCl₂ activator [79].

Bis(phenoxy-imine) Group 4 metal complexes (named FI catalysts), combined with MAO or borate activators, display unique polymerization activities in α-olefin polymerization [130,131,132,133,134,135]. Because of the high productivity of FI catalysts and the broad spectrum of accessible polymers, there has been significant interest in developing supported catalysts for industrial applications. The use of harmless and cost-competitive MgCl₂ as an activating support seemed a profitable and promising idea. In 2004, the Fujita group reported the results of the study of catalytic behavior of the complexes Ti7 and Zr4 (Scheme 4b), activated by Et₃Al in the absence and in the presence of mechanically pulverized MgCl₂, in the polymerization of ethylene. At 50 °C and an ethylene pressure of 0.9 MPa, no polymer formed in the absence of MgCl₂, whereas in the presence of MgCl₂, Ti7 and Zr4 demonstrated activities of 0.32 and 9.10 kg∙mmol⁻¹∙bar⁻¹∙h⁻¹ [79]. The authors proposed that MgCl₂ acts as a Lewis acid to form a cationic active species from the ethylated FI catalysts and becomes an anionic species stabilizing cationic active species. Under identical conditions, metallocenes Zr2 and Ti2 provided only a trace amount of PE (<0.1 kg∙mmol⁻¹∙bar⁻¹∙h⁻¹).

2.2. MgCl₂/THF Adducts

Supported Cp₂TiCl₂/MgCl₂ pre-catalyst, prepared by the precipitation of Ti2 and MgCl₂ solution in THF into pentane, polymerized ethylene after activation by ⁱBu₃Al or Et₂AlCl, and no activity was detected after treatment of Ti2 by ⁱBu₃Al in the absence of MgCl₂ [136]. Et₂AlCl was found to be a more efficient activator in comparison with ⁱBu₃Al, and the catalysts showed a steady-state kinetic behavior. The authors explained these results by Ti→Mg transfer of the alkyl group with a formation of active Cp₂TiR⁺ species.

The MgCl₂(THF)₂/Et₂AlCl phase was proposed as a support for Zr2 by Nowakowska and colleagues [125,137], who assumed the formation of the catalyst precursor Cp₂Zr(μ-Cl)(μ-Et)Al(μ-Cl)(μ-THF)Mg on the surface of MgCl₂(THF)₂, which was activated by MAO without desorption. Activation of this catalyst by R₃Al was not studied.

2.3. MgCl₂/ROH Adducts

The products of the reaction of MgCl₂/ROH adducts with alkylaluminum compounds are another type of MgCl₂-based activating supports. Evidently, their chemical nature is qualitatively different from the chemical nature of the product of the reaction of MgCl₂ with R₃Al.

Cho and Lee prepared adducts MgCl₂·4MeOH and MgCl₂·3.33EtOH, and thermal treatment of the former complex resulted in the formation of MgCl_x(OMe)_y species [138]. Active supports were obtained by the treatment of the adducts with Me₃Al, Et₃Al, ⁱBu₃Al, or MAO (comparative experiments). A more efficient binding with alkylaluminums was observed for MgCl₂·4MeOH due to the formation of Mg–(μ-OMe)–Al species. The proposed impregnation mechanism for Zr2, based on ²⁷Al NMR spectral data, is presented in Scheme 5. Supported catalysts were studied in ethylene/hex-1-ene copolymerization after activation by MAO; therefore, the results of the research MgCl₂(ROH)_x/R₃Al system cannot be regarded as a full-fledged activating support.

Scheme 5. Proposed mechanism for the impregnation of the metallocene catalysts over the recrystallized MgCl₂ [138].

Subsequent studies of the use of MgCl₂ supports obtained by the reaction of MgCl₂/EtOH adducts with AlⁱBu₃, AlEt₃, or AlEt₂Cl have demonstrated high efficiency of ⁱBu₃Al and Et₃Al pre-treated supports in activation of Ti2, Zr2, and Me₂Si(η⁵-C₅Me₄)(N^tBu)TiCl₂ (Ti8) in the absence of MAO and borates [80]. In the homopolymerization of ethylene, catalytic activities reached 300 kg∙mol⁻¹∙bar⁻¹∙h⁻¹, and PE of spherical morphology (Figure 2) and relatively high dispersity Đ_M = 2.4–2.7 were obtained. Note that Ti2/MgCl₂∙2.1EtOH/ⁱBu₃Al catalyzed formation of PE with Đ_M = 2.0, which confirms the uniformity and stability of the active species. The activity of rac-[1,2-ethylenebis(η⁵-inden-1-yl)]ZrCl₂ (rac-[EBI]ZrCl₂, Zr5)/MgCl₂∙2.1EtOH/Et₃Al was 5 kg∙mol⁻¹∙bar⁻¹∙h⁻¹, and PE with narrow molecular weight distribution (Đ_M = 2.1) was formed.

Figure 2. Scanning electron micrograph of polyethylene prepared using Cp₂TiCl₂ (Ti2)/MgCl₂ 1.1EtOH/Et₃Al catalyst. Reprinted with permission from [80]. Copyright (2004) Wiley-VCH Verlag GmbH and Co.

Supports, prepared by the reaction of MgCl₂/EtOH with Et₃Al (Et₃Al/EtOH = 2), were used for the activation of a series of zirconocenes and hafnocenes in ethylene polymerization [81]. The investigation of substituent effects was first carried out using the support of composition MgCl₂∙0.30AlEt_2.46(OEt)_0.54 (Table 1). Catalyst immobilization was obtained by contacting the support overnight with a metallocene solution in toluene; ethylene polymerization was carried out at 50 °C with an ethylene pressure of 5 bar. The ethylene mass flow remained essentially constant during polymerization, indicating negligible decay in catalyst activity.

Table 1. Ethylene polymerization using (RCp)₂ZrCl₂ and (C₅H_nR_5–n)₂ZrCl₂ immobilized on an MgCl₂-based support [81] ¹.

Zr Comp.	Cp Ring Substitution	Activity, kg∙mol−1∙ bar−1∙h−1	Tm, °C	Degree of Crystallinity χc	Mn, kDa	Mw, kDa	ĐM
Zr2	H	160	134.5	53.5	88	352	4.0
Zr6	Et	260	135.1	55.6	113	394	3.5
Zr7	n-Pr	2760	135.4	58.5	95	242	2.5
Zr8	iPr	360	135.6	51.7	194	485	2.5
Zr9	n-Bu	3680	135.3	56.9	120	271	2.3
Zr10	tBu	120	135.1	53.7	126	425	3.4
Zr11	n-Pentyl	1120	134.5	57.9	98	302	3.1
Zr12	n-Dodecyl	700	135.6	53.8	122	344	2.8
Zr13	1,3-Me2	140	135.2	47.1	172	474	2.8
Zr14	1-Me-3-n-Bu	880	135.9	54.3	217	467	2.2
Zr15	1,2,4-Me3	120	136.1	50.1	184	430	2.3
Zr16	Me4	140	139.2	47.3	119	392	3.3
Zr17	Me5	180	139.2	47.3	148	391	2.6

¹ Polymerization conditions: 500 mL of light petroleum, 100 mg of immobilized catalyst (1 μmol Zr), 1 mmol of AlEt₃, 50 °C, ethylene pressure of 5 bar, 1 h.

As can be seen in Table 1, low activities were obtained for (RCp)₂ZrCl₂ with R = H or Et, but longer-chain alkyl substituents (Zr7, Zr9) provided much higher activities. Introducing branching into the alkyl substituent (Zr8, Zr10) led to poor catalyst performance. According to the study of the authors, the increased steric bulk imposed by branched substituents evidently impedes effective activation of the magnesium chloride support, but the causes of abnormal activities of Zr7 and Zr8 in comparison with Zr6 remain unclear. The analogous hafnocenes were also investigated; Cp₂HfCl₂ (Hf1) demonstrated an activity of 40 kg∙mol⁻¹∙bar⁻¹∙h⁻¹, whereas with (n-BuCp)₂HfCl₂ (Hf2), the activity was 240 kg∙mol⁻¹∙bar⁻¹∙h⁻¹. An increase in the number of Me substituents had no effect on the activity (Zr13, Zr15–Zr17). Mixed-ligand zirconocenes (RCp)(Cp)ZrCl₂ with R = n-Pr, n-Bu, and n-pentyl (Zr18–Zr20) demonstrated activities of 740, 1020, and 560 kg∙mol⁻¹∙bar⁻¹∙h⁻¹, respectively, i.e., a smoothed but clear maximum of activity was observed for R = n-Bu in (RCp)(Cp)ZrCl₂ with equally unclear reasons. The variation of the composition of MgCl₂/EtOH complexes using Zr7 showed a better performance for support, prepared from MgCl₂∙1.1EtOH (4300 kg∙mol⁻¹∙bar⁻¹∙h⁻¹), and a higher EtOH content resulted in a lowering of the activity. When ball-milled MgCl₂ was used as a support, activity was 720 kg∙mol⁻¹∙bar⁻¹∙h⁻¹. When using ⁱBu₃Al instead of Et₃Al, more thermally stable catalysts were obtained, and the activity of the Zr7-based catalyst was 12,680 kg∙mol⁻¹∙bar⁻¹∙h⁻¹. It can be assumed that the higher efficiency of ⁱBu₃Al in comparison with Et₃Al is due to the possibility of side reactions after the formation of unstable ZrEt₂ derivatives (Scheme 1c) [37,38]. Note that the use of (n-PrCp)₂ZrMe₂ (Zr7′) instead of dichloro complex Zr7 with different organoaluminum co-catalysts did not result in changes of activity and PE characteristics, signifying that the limiting factor in the generation of active species in this system is not the alkylation stage. UV-vis spectral studies showed that the uptake of Zr2 from the solution was much slower than that of Zr6 or Zr7, indicating a beneficial effect of an electron-donating alkyl substituent.

The nature of the active species in MgCl₂-supported catalysts is still far from being resolved. Reasoning by analogy with TMZNCs, two possible coordinations were proposed for Cp₂ZrMe₂ (Zr2′, Figure 3) [81]. The formation of alkyl-zirconocenium cations was attributed to the abstraction of the Me group by acidic Mg centers. Another important factor was the complexation of R₂AlOEt with Zr2 or Zr2′; supports with relatively low contents of residual ethoxide would therefore be expected to give higher activities for immobilized zirconocenes, as observed in [81]. In our opinion, the coordination of Zr2′ on the MgCl₂ surface, presented in Figure 3, is a very rough model: Mg–(μ-OEt)–Al species, formed during the reaction of MgCl₂/EtOH with R₃Al, have little choice other than to participate in the formation of active species with Zr2 or Zr2′.

Figure 3. Possible modes of coordination of a zirconocene on (a) a (110) and (b) a (100) face of MgCl₂. Reprinted with permission from [81]. Copyright (2008) American Chemical Society.

It should also be noted that the reaction of MgCl₂/EtOH with Ti2 in the absence and in the presence of R₃Al is a complex process with a formation of diverse reaction products: as shown in 2016 by Sobota and colleagues [139], during this reaction, Ti2 loses η⁵-ligands, and heterometallic polynuclear clusters are formed. Given the high nucleophilicity of Mg and Al alkoxides, the decomposition of metallocene fragments is also possible for Zr complexes, but this issue has not yet been studied.

The MgCl₂/ROH/R′₃Al system was also successfully used for the activation of Group 4 post-metallocenes. Nakayama et al. [77,78,79] prepared an MgCl₂/2-ethyl-1-hexanol adduct, which is soluble in n-decane; treatment of this adduct with ⁱBu₃Al resulted in the formation of an MgCl₂/ⁱBu_mAl(OR)_n support, which is used for activation of bis(phenoxy-imine) titanium complexes Ti7 and Ti9–Ti11 (Scheme 6). At 50 °C and an ethylene pressure of 0.9 MPa, catalytic activities were 36.3, 20.8, 36.0, and 26.2 kg∙mmol⁻¹∙h⁻¹, respectively. M_w of PE was 230–1170 kDa, with broadened molecular weight distribution (Ð_M = 2.4–3.5). When using MAO as an activator, PE of the worst morphology was obtained (Figure 4).

Scheme 6. Bis(phenoxy-imine) titanium complexes, studied in ethylene polymerization using MgCl₂/ⁱBu_mAl(OR)_n support [77,78,79].

Figure 4. Photographs of the polyethylenes formed with complex Ti7 using (a) MgCl₂/ⁱBu_mAl(OR)_n or (b) MAO as an activator. Reprinted with permission from [78]. Copyright (2003) Elsevier B. V.

To confirm the single-site nature of the catalyst using an MgCl₂/ⁱBu_mAl(OR)_n activator, copolymerization of ethylene with propylene was performed [78,79]. Complex Ti10 produced a copolymer with a propylene content of 29.3 mol% with an activity of 28.2 kg∙mmol⁻¹∙h⁻¹ and a narrow MWD (Ð_M = 1.70), thus confirming homogeneity of the catalytic species [78,79]. The studies of Ti12 and Ti13, containing C₆F₅ fragment (Scheme 7a), in the polymerization of propylene resulted in the following: Ti12-based system catalyzed living propylene polymerization at 25 °C and atmospheric pressure. The living nature was confirmed by the linear relationship between M_n and polymerization time and by the narrow MWD of the PP obtained (M_n 53 kDa, Đ_M = 1.09). The complex Ti13 activated by MgCl₂/ⁱBu_mAl(OR)_n produced highly syndiotactic PP ([rr] = 97%, T_m = 155 °C).

Scheme 7. Bis(phenoxy-imine) titanium (a) and zirconium (b) complexes, studied in polymerization using MgCl₂/ⁱBu_mAl(OR)_n supports [79,140].

When studying bis(phenoxy-imine) Zr complexes Zr4 (Scheme 4b) and Zr21–Zr24 (Scheme 7b) using MgCl₂/ⁱBu_mAl(OR)_n activator in ethylene polymerization, a catalytic activity of 1820 kg∙mmol⁻¹∙h⁻¹ (Zr22) was achieved. For Zr21 and Zr22, MAO was a less effective activator in comparison with MgCl₂/ⁱBu_mAl(OR)_n [79], and PE had a very high molecular weight (up to M_v = 3 MDa) with an exceptionally high bulk density value of 0.47 g∙mL⁻¹ due to excellent spherical morphology of polymer particles.

The main findings and prospects of FI catalysts when using MgCl₂/R′_nAl(OR)_3−n activating supports were presented and discussed by Fujita and colleagues in conceptual work [140] that summarized and complemented the results of previous studies [77,78,79]. In the case of FI catalysts, MgCl₂-based supports pose a real competition to MAO, particularly with regards to polymer morphology (Figure 5).

Figure 5. SEM image of the spherical particle PEs produced with MgCl₂-supported Zr-FI catalyst Zr23. Reprinted with permission from [140]. Copyright (2006) Wiley-VCH Verlag GmbH and Co.

2.4. MgCl₂/SiO₂ Supports

To prepare MgCl₂/SiO₂ bisupport particles, Chung and colleagues used agglomeration of silica gel under the action of aq. MgCl₂ [141]. After separation and drying at 80 °C, bisupport particles were treated with Me₃Al and Zr2 in toluene media. The supported catalyst demonstrated an activity of 1013 kg∙mol⁻¹∙bar ⁻¹∙h⁻¹ in the polymerization of ethylene. One might suppose that the organoaluminum activator in this experiment represents MAO, which was formed by hydrolysis of Me₃Al; however, the comparative experiment with MAO was characterized by a qualitatively different kinetic profile with rapid deactivation of the catalyst.

A prospective approach to MAO- and borate-free supported catalysts was proposed by Kissin and colleagues who prepared catalysts by mixing Zr5, Bu₂Mg, and Et₂AlCl solutions (Zr/Mg/Al ratio ~1:10:30), followed by addition of the mixture to Davison-grade 955 silica [142]. In ethylene/hex-1-ene copolymerization at 90 °C and ethylene pressure of 1.3 MPa, catalytic activities up to 7800 kg∙mol⁻¹∙h⁻¹ were achieved, and the hex-1-ene content in the copolymer was 4.4 mol%. The authors assumed that the reaction of Bu₂Mg with Et₂AlCl results in the formation of MgCl₂ that can abstract R^– or Cl^– from the alkylated Zr complexes, and that the resulting anionic species have a broadly distributed negative charge to stabilize rac-[EBI]ZrR⁺ active species.

In 2018, Ko and colleagues described binary support prepared from MgCl₂/EtOH adducts and silica [143]. Subsequent treatment with various alkylaluminum compounds and Zr9 resulted in the formation of supported catalysts for ethylene/hex-1-ene copolymerization. Among alkylaluminum compounds, Et₃Al₂Cl₃ demonstrated the highest efficiency, but the hex-1-ene content was ~1 mol% in all experiments. The scheme of catalyst immobilization, presented in [143], looks doubtful as it suggests the formation of a Zr–(μ-O)–Al bond that hinders the formation of active catalytic species.

2.5. Metal Fluorides

Comparative experiments on the activation of Zr1 by Me₃Al using MgF₂, CaF₂, and AlF₃, conducted by Soga et al. [75], demonstrated relatively high activities in ethylene polymerization for Group 2 metal fluorides, which were comparable with the activities when using MgCl₂ support. Similar results were obtained for Zr3, which was supported on MgF₂ in the presence of Me₃Al [75].

The absence of further research on metal fluoride-activated single-site Group 4 metal can be attributed to the greater development of MgCl₂-based catalysts, widely used in Ziegler–Natta industrial processes. As a result, we have what we have, and the very idea of the activating supports currently neglects the use of metal fluorides as a basis for catalyst-containing particles. At the same time, positive results of the use of fluorine-modified metal oxides have demonstrated the advantages of fluorine-containing supports (see below); therefore, the use of metal fluorides as catalyst’s supports seems undervalued.

3. Silica and Metal Oxides

3.1. Silica SiO₂

Activation of Zr1 [73,75], as well as Zr2 and Zr3 [74], by Me₃Al using SiO₂ support in propylene polymerization failed. According to the study of the reaction of (η⁵-C₅Me₅)TiMe₃ (Ti14′) with SiO₂ and Me₃Al, the formation of Ti–Al and Ti–Al₂ heterobimetallic species was proposed [144]; however, ethylene uptake by this catalytic system was only mentioned without any experimental data for comparison.

Chemisorption of Me₃Al on dehydrated silica material MCM-41 (Mobil Composition of Matter No. 41 [145]) of hierarchical microstructure was studied by Anwander et al. [146]. In the reaction product, a SiMe/AlMe population ratio of 0.45, as well as the formation of Lewis acidic centers, were detected in model reactions using Y[N(SiHMe₂)₂]₃(THF)₂; Group 4 metal complexes on this support were not explored. In 2001, a similar study was conducted by Sano and colleagues [147]. After calcination at 500 °C and thermal pre-treatment at 280 °C (the content of Si–OH groups was ~3 mmol∙g⁻¹), MCM-41 was treated with Me₃Al (4 mmol per 1 g of the support) and, after separation, calcined at 500–800 °C in vacuo; AlMCM-41 supports were obtained in this way. Propylene polymerization was conducted at 40 °C using Zr1, and ⁱBu₃Al was used as an additional activator/scavenger. The use of MCM-41 in place of the AlMCM-41 provided no polymer. In the case of AlMCM-41, catalytic activity passed through the maximum for support and calcined at 700 °C in vacuo.

Detailed investigation of the solid–liquid reaction of silica with Me₃Al by ¹³C, ²⁷Al, and ²⁹Si MAS NMR [148] showed that AlMe_n moieties constitute about 70% of the total amount of surface-attached Me groups. The amounts of Si–OMe, SiMe₃, and SiMe₂ moieties were approximately equal to each other and totally amounted to ~30% of Me groups. Note that the formation of Si–OMe fragments implies the presence of Si–Al bonds in the reaction product (not detected by NMR); however, the formation of Si–Al bonds was only mentioned in this work, without direct experimental proof.

The reaction of mesoporous silica SBA-15 with ⁱBu₃Al(Et₂O) resulted in the formation of well-defined (Si_SO)₂AlⁱBu(Et₂O) and (Si_SO)₃SiⁱBu species via Al→Si transfer of ⁱBu group [149]. More in-depth and comprehensive studies of the reaction of SBA-15 with Et₃Al [150], ⁱBu₃Al [151,152], and Et₂AlCl [153] resulted in the detection of a variety of Al₁ and Al₂ species, which were chemically bonded with silica surface via Si_SOAl fragments. Among the products formed after treatment with ⁱBu₃Al, (Si_SO)₂AlH and (Si_SO)₃SiH species were detected [152]. The formation of Si–Al bonds was not discussed in these later works, which would make the very idea of Si–Al bond formation during the reaction of silica with R₃Al questionable.

In conclusion, we note that the interaction of silica with organoaluminum compounds was used as a first stage of the synthesis of functionalized (e.g., fluorinated) activating supports (see Section 5).

3.2. γ-Al₂O₃

The γ-Al₂O₃ represents the metastable spinel-type cubic phase of alumina obtained by the thermal dehydration (calcination) of aluminum hydroxides and oxyhydroxides [154]. Its bulk structure (Figure 6a) contains an fcc sublattice of oxide ions that generates octahedral and tetrahedral interstices that accommodate Al ions [155,156]. The precursor of γ-Al₂O₃ is a well-organized and stable boehmite (AlOOH) that contains a sublattice of cubic close-packaged O²⁻ anions with interstitially situated Al³⁺ cations [157] (Figure 6b).

Figure 6. (a) Optimized structure of γ-alumina. Reprinted with permission from [155]. Copyright (2008) American Chemical Society. (b) Crystal structure of boehmite. Reprinted with permission from [157]. Copyright (2021) International Union of Crystallography.

Dehydroxylated and partially dehydroxylated γ-alumina surfaces demonstrate an explicit Lewis acidic character due to the presence of coordinatively unsaturated surface Al sites [158,159,160,161]. From a coordination chemistry point of view, the accessibility of both Brønsted acidic sites (BASs) >Al–OH and Lewis acidic sites (LAS) >Al– on γ-Al₂O₃ surface offers a unique complexation environment (Figure 7). The nature of these sites changes with the level of hydroxylation, which is defined and achieved by pre-treatment temperature and duration [159,160,161].

Figure 7. Top views of (a) fully dehydrated (100), (b) fully dehydrated (110), (c) hydrated (110), and (d) hydrated (111) terminations of γ-Al₂O₃, showing different types of surface Al atoms. X(μ-OH)_w(OH)_y represents an Al atom with a total coordination number of X, including w bridging μ-OH and y terminal OH groups. Only atoms in the top surface layers are shown for clarity. Reprinted with permission from [161]. Copyright (2014) American Chemical Society.

During the complex experimental and theoretical study of the surface reactivity of γ-Al₂O₃ [160], Wischert et al. showed high efficiency of Al LAS in N₂ absorption and reactivity of Al,O Lewis acid−base pairs in heterolytic dissociation of H₂ and CH₄ molecules with a formation of Al–H and Al–Me species. The maximum site density was observed for alumina, thermally treated at 700 °C, thus confirming the significance of the process on (110) termination of γ-Al₂O₃ (see below).

The marked difference in reactivity of γ-Al₂O₃, calcined at different temperatures, was demonstrated in a number of works, based on the use of the complexes MR₄. Partially dehydroxylated at 500 °C, γ-Al₂O₃ was grafted with Zr(CH₂^tBu)₄, and reaction products contained 2 CH₂^tBu per Zr atom [162]. Combined experimental and theoretical (Figure 8) studies allowed the authors to propose the reaction mechanism that includes the formation of (Al_SO)₂ZrR₂ species that further react with the alumina surface through the transfer of one of its R ligands onto an adjacent Al_S Lewis center, providing a cationic surface complex [(Al_SO)₂ZrR]⁺[(Al_S)R]^–. This alkyl transfer was confirmed by a combination of ¹³C CP MAS NMR and chemical shift calculations. In further theoretical and experimental works [158,163], the formation and chemistry of hydride species [(Al_SO)₂Zr(H)(μ-H)-Al_VI] and [(Al_SO)₂Zr(H)(μ-R)Al_VI] along with cationic [(Al_SO)₂Zr(H)]+ species stabilized by tetrahedral aluminum hydrides [(Al_IV-H)^–] by hydrogenolysis of [(Al_SO)₂ZrR]⁺[(Al_S)R]^– was proposed. During the formation of these hydrides or the hydrogenolysis of alkanes, the alkanes transformed into lower homologues to provide methane and ethane as final products. These processes go beyond the subject of our review but seem relevant to the current problem of chemical recycling and upcycling of polyolefins [164].

Figure 8. {[(Al_SO)₂Zr(CH₂^tBu)⁺][Al_S(CH₂^tBu)^–]} structure optimized on the model surface. Reprinted with permission from [162]. Copyright (2006) American Chemical Society.

Grafting of Hf(CH₂^tBu)₄ on γ-Al₂O₃ resulted in different products depending on pre-treatment temperature (Scheme 8): the temperature of 350 °C was not high enough to provide the ability of γ-Al₂O₃ to react with Hf alkyl as an activating support [159]. These results were in line with previous experimental and theoretical studies of the Hf(CH₂^tBu)₄/γ-Al₂O₃ system [165].

Scheme 8. Grafting reaction of Hf(CH₂^tBu)₄ on γ-Al₂O₃ at room temperature [159].

Another important aspect of the chemistry of γ-Al₂O₃ as an activating support also involved the own reactivity of γ-Al₂O₃ towards R₃Al. As shown in the work of Mazoyer et al. [166], the reaction of calcined γ-Al₂O₃ with ⁱBu₃Al results in the formation of the surface Al–ⁱBu species, and in the presence of molecular hydrogen, Al–H species are formed. It was found that Al–H species can initiate the polymerization of ethylene (with low activities, ~0.55 kg∙mol⁻¹∙h⁻¹). However, it can be assumed that surface Al–R and Al–H species can participate in the formation of Group 4 metal catalytic centers on the surface of γ-Al₂O₃ when using organoaluminum activators.

The first study of the activation of Group 4 metallocenes by γ-Al₂O₃ was conducted by Marks and colleagues in 1988 [71]. In their experiments, highly dehydroxylated alumina (DA, 150 m²∙g⁻¹, ~0.1 surface OH∙nm⁻²) or partially dehydroxylated alumina (PDA, 150 m²∙g⁻¹, ~4 surface OH∙nm⁻²) were mixed with solutions of Zr2′, (η⁵-C₅Me₅)₂ZrMe₂ (Zr17′), and (η⁵-C₅Me₅)ZrMe₃ (Zr25′) in n-pentane, with subsequent elimination of the solvent in He flow and treatment by molecular hydrogen. Hydrogenation of propylene was studied at −63 °C, and turnover frequencies N_t (s⁻¹) amounted to 0.3, 0.2, and 1.1 (DA) and 0.1, 0.06, and 0.3 (PDA) for Zr2′, Zr17′, and Zr25′, respectively. Relative activities in ethylene polymerization were for Zr2′ >> Zr17′ ≈ Zr25′. The percentage of active sites, determined by their coordination with CO, was found to be 4% for Zr2′/DA and 12% for Zr25′/DA.

Activation of Zr1 by Me₃Al and Et₃Al using γ-Al₂O₃ support in propylene polymerization showed moderate efficiency, which was comparable with the efficiency of MgCl₂ support [73,75]. A similar pattern was observed for Zr3, whereas the catalytic activity of Zr3 when using γ-Al₂O₃ was four times lower than the activity in the presence of MgCl₂ (42 vs. 172 kg∙mol⁻¹∙h⁻¹) [74]. Ti1/ γ-Al₂O₃ pre-catalyst, prepared by the reaction of a toluene solution of Ti1 with γ-Al₂O₃ (110 °C, 3 h), demonstrated low activity in propylene polymerization after activation by ⁱBu₃Al [126]. Apparently, this was due to the low calcination temperature for γ-Al₂O₃ used (200 °C). During comparative studies of the activation of Zr9 by different metal oxides, treated by ⁱBu₃Al, γ-Al₂O₃ demonstrated a moderate efficiency relative to other supports [167].

¹³C CP MAS NMR and EXAFS studies of the reaction of Zr25′ with γ-Al₂O₃ (calcined at 500 °C) showed the presence of Al–O–ZrMe₂(η⁵-C₅Me₅) species; some of the surface complexes underwent further interaction with the γ-Al₂O₃ surface, either involving bridging methyl groups between Zr and Al or full Zr→Al transfer of a methyl group [82]. This sample showed moderate activity in ethylene polymerization (~20 kg∙mol⁻¹∙bar⁻¹∙h⁻¹). The results of the ¹³C CP MAS NMR study of the reaction of Zr2′ with γ-Al₂O₃ confirmed the formation of cationic Cp₂Zr⁺ species, which bonded with alumina via Al–O–Zr covalent or Coulombic interactions [82]. Obviously, only the latter Cp₂ZrMe⁺ species can be considered catalytically active, but their content was low (activity in ethylene polymerization was only 3 kg∙mol⁻¹∙bar⁻¹∙h⁻¹). In [82], the authors proposed the common scheme of the structure–activity relationship for Zr25′ and Zr2′, supported by γ-Al₂O₃ (Figure 9), but the potential catalytic activity of some complexes, presented in this figure, appears doubtful since direct Si–O–Zr or Al–O–Zr bonding will hardly lead to the formation of highly active catalytic species.

Figure 9. Structure–activity relationship for Zr25′ and Zr2′, supported on γ-Al₂O₃. Reprinted with permission from [82]. Copyright (2001) American Chemical Society. The number of “+” symbols reflects catalytic activity according to the assessment of the authors of [82], from very low (+) to very high (+++++).

A more thorough and accurate theoretical study of the structural and catalytic properties of Zr2′, chemisorbed on dehydroxylated γ-Al₂O₃, was conducted by Marks and colleagues [155]. They developed a γ-Al₂O₃ surface model and studied the interactions of Zr2′ with various surface coordination sites. The results of this modeling were compared and contrasted with data for the homogeneous [Cp₂ZrMe⁺][MeB(C₆F₅)₃^–] system. Since theoretical and experimental data for γ-Al₂O₃ indicate that the (110) surface predominates (70–80% of the total area), the modeling was focused on the (110) surface (Figure 10a). The optimized alumina (110) surface was found to exhibit significant rearrangement of Al and O ions relative to the bulk; in particular, bulk octahedral Al centers became pseudotetrahedral (Al_IV in Figure 10a) while bulk tetrahedral Al centers became pseudotrigonally planar on the surface (Al_III in Figure 10a). The surface O ions were found to have either μ³-O and μ²-O geometries. The μ³-O species were bound to Al_III and Al_IV surface ions and to the octahedral Al_VI bulk ion, while the μ²-O species were bound to the Al_IV and to the octahedral Al_VI bulk ions.

Figure 10. (a) Aluminum and oxide sites exposed on the optimized γ-Al₂O₃ (110) surface; (b) dioxo- and oxo-bridged zirconocenium coordination on μ³-O and μ²-O surface sites. Reprinted with permission from [155]. Copyright (2008) American Chemical Society.

The acidic Al sites activate Zr2′ via Zr–Me scission with a Zr→Al transfer of the Me group to the surface; solid-state NMR data showed that the methide group had been placed on the Al_III center at a distance of ~5–8 Å from the metallocenium center. The possible ion pair interactions involving Cp₂ZrMe⁺ and γ-Al₂O₃ surface were scrutinized to search for the most stable configurations for the complexes at μ³-O and μ²-O sites (Figure 10b). Comparison of the calculated enthalpies of ion pair formation (ΔH_form) and separation (ΔH_ips) for chemisorbed species and [Cp₂ZrMe⁺][MeB(C₆F₅)₃^–] (Table 2) showed the stronger coordinative capability of the μ²-O sites vs. the μ³-O sites. Moreover, interactions found in oxo-bridged configurations were significantly stronger than those in the dioxo-bridged configurations. For [Cp₂ZrMe⁺][MeB(C₆F₅)₃⁻] the adduct formation was less exothermic, but the value of ΔH_ips was comparable to that in the heterogeneous μ³-O dioxo-bridged complex.

Table 2. Calculated ion pair formation enthalpies ΔH_form and heterolytic ion pair separation enthalpies ΔH_ips (kcal∙mol⁻¹) for activation of Cp₂ZrMe₂ on the Al₂O₃ (110) surface and with the use of B(C₆F₅)₃ [155].

Complex of Cp2ZrMe+	d(Zr–O), Å 1	ΔHform, kcal∙mol−1	ΔHips, kcal∙mol−1
μ3-O dioxo-bridged	2.82	−16.7	77.2
μ3-O oxo-bridged	2.58	−21.7	82.1
μ2-O dioxo-bridged	2.34	−51.7	112.2
μ2-O oxo-bridged	2.13	−73.3	127.5
[MeB(C6F5)3−] adduct	2.56 2	−1.7	77.0

¹ In the case of dioxo-bridged structures, a mean distance is given. ² The value of d(Zr–C) for Zr–(μ-Me)–B fragment.

π-coordination and insertion of ethylene molecule required only partial ion pair separation. These reaction steps were modeled for all four coordination modes of chemisorbed Cp₂ZrMe⁺ and for [Cp₂ZrMe⁺][MeB(C₆F₅)₃⁻] (Figure 11).

Figure 11. Calculated ethylene enchainment energy profiles for μ²-O coordinated (a) and μ³-O coordinated (b) Cp₂ZrMe⁺ species; the energy profile for [Cp₂ZrMe⁺][MeB(C₆F₅)₃⁻] is given for comparison (c). Reprinted with permission from [155]. Copyright (2008) American Chemical Society.

As can be seen in Figure 11, π-coordination and insertion of ethylene molecule were easier for μ³-O coordinated species even in comparison with [Cp₂ZrMe⁺][MeB(C₆F₅)₃⁻]. Therefore, only a fraction of the surface sites in dehydroxylated γ-Al₂O₃-supported metallocenes are catalytically significant. However, those catalytically significant surface sites exhibit greater catalytic activities than their homogeneous analogs. In addition, it can be assumed that the chemistry of modified alumina supports (see Section 5) may involve surface modification with a decrease in the content of μ²-O sites.

Ti(III) complex Ti15 with the formula [Ti(nacnac)(CH2^tBu)₂] (nacnac = [Ar]NC(Me)CHC(Me)N[Ar], Ar = 2,6-(CHMe₂)₂C₆H₃) [168] was grafted onto alumina that was partially dehydroxylated at 700 °C. This reaction is accompanied by the release of 0.54 equivalents of neopentane per initial surface OH group, which confirms Al–O–Ti bonding with the support. The obtained complex was found to be active in ethylene polymerization, which expands the understanding of the mechanism of α-olefin polymerization by the notion that d¹ Ti alkyls are possible active sites in the heterogeneous Ziegler–Natta polymerization catalysts.

3.3. Silica-Alumina SiO₂/Al₂O₃

Due to their combined Lewis and Brønsted acidities, amorphous SiO₂/Al₂O₃ are widespread supports for multifunctional heterogeneous catalysts. The local environment of the acid sites is essential for activation, and in 2010, Chizallet and Raybaud [169] proposed bridging Si–(OH)^…Al fragments as an active BAS (Scheme 9a).

Scheme 9. (a) Model for Brønsted acidic site (BAS) in silica-alumina [169]; (b) the structures of the products of the reaction of Zr25′ with SiO₂/Al₂O₃ [82].

During the ¹³C CP MAS NMR and EXAFS studies of the reaction of Zr25′ with SiO₂/Al₂O₃ [82], different products were detected (Scheme 9b). The main product was the complex with a Si–O–Zr bond, which was observed when studying the reaction of Zr25′ with silica (see Section 3.1), but the minor products can be considered as trinuclear species with more or less degrees of charge separation. However, experiments on ethylene polymerization showed zero catalytic activity.

In the activation of Zr9, ⁱBu₃Al-treated SiO₂–Al₂O₃ demonstrated moderate efficiency, which is comparable with the efficiency of Al₂O₃ if specific surface areas of the supports are considered [167].

3.4. Clay Minerals and Related Systems

In 2002, Weiss and colleagues published the results of the study of activation of Group 4 metallocenes by kaolin (vacuum-dried) and montmorillonite (pre-heated at 200 °C) as inorganic carriers with the use of Me₃Al and ⁱBu₃Al as activators [170]. In view of the results of preliminary investigations [171], the authors postulated the mechanism of activation, presented in Scheme 10. Note that this mechanism is quite correlated with the results of the activation of metallocenes by silica/R₃Al (no activity) [73,74,75].

Scheme 10. Formation of heterogeneous Ti and Zr metallocene catalysts [170,171].

Using kaolin/Me₃Al as an activating support, in the polymerization of ethylene (50 °C, 10 bar), activities of 5.2, 4.3, and 1.5 kg∙mmol⁻¹∙h⁻¹ were achieved for Zr2, Cp₂ZrHCl (Zr2″), and Ti2, respectively. On montmorillonite/R₃Al support Zr2 demonstrated activities up to 14 kg∙mmol⁻¹∙h⁻¹, and ⁱBu₃Al was found to be a more efficient co-catalyst in comparison with Me₃Al. In propylene polymerization, Zr1/montmorillonite/ⁱBu₃Al ([Al]/[Zr] = 3200) and Zr1/montmorillonite/Me₃Al ([Al]/[Zr] = 800) catalysts were even more active than the homogeneous Zr1/MAO system ([Al]/[Zr] = 1000); the activities amounted to 60.3, 6.8, and 5.7 kg∙mmol⁻¹∙h⁻¹, respectively. However, the isotacticity of PP, obtained on supported catalysts, was lower (80 and 83% vs. 89%). Another noteworthy result was obtained when studying the activation of homo- and bimetallic Ti complexes Ti16 and Ti17: at [Al]/[Zr] = 2000, the montmorillonite/ⁱBu₃Al activator was much more efficient than MAO (Figure 12).

Figure 12. Activities of homogeneous and heterogeneous ethylene polymerization of catalysts Ti16 and Ti17. Reprinted with permission from [170]. Copyright (2002) Elsevier B. V.

Montmorillonite/Et₃Al system was also studied by Nakano and colleagues in activation of rac-Me₂Si(2-Me-4-Ph-1-Ind)₂ZrCl₂ (Zr26) [172]. After pre-heating at 200 °C and treatment with Et₃Al (support 1), 2,6-dimethylpyridine was added in different amounts (supports 2 and 3). Supports 1–3 were reacted with ⁱBu₃Al and Zr26. UV-vis spectra of the catalysts showed the formation of cationic Zr species only for support 1, containing highly acidic centers.

Since a specific surface area of the support is of great importance for catalytic activity, Sun and Garcés suggest acid lamellar aerogel prepared by exfoliation of clay in acidic media [173]. Supported catalysts prepared by the reaction of aerogel with R₃Al and Zr2′ demonstrated activities of 29–69 kg∙mmol⁻¹∙h⁻¹. Since ⁱBu₃Al was the best activator, and Me₃Al was the worst, the mechanism of activation can not be interpreted through the formation of alumoxanes. The use of spray-dried silica/clay agglomerates for activation of Zr9 and Ti8 has been patented by W. R. Grace & Co. [174]. When using 2–6 mmol of ⁱBu₃Al or Me₃Al and 12–120 μmol of pre-catalyst per 1 g of support, activities of 0.3–2 kg_PE∙g_Cat⁻¹∙h⁻¹ were achieved in ethylene polymerization.

Acid-treated commercial montmorillonite clays (K10, K30) and alumina-modified SBA-15 were studied by Cuenca and colleagues in 2010 [175]. After treatment with Me₃Al or Et₃Al, these supports (2.5 g) were mixed with (η⁵-C₅Me₅)(η⁵-C₅H₄SiMe₂CH₂CH=CH₂)ZrCl₂ (Zr27, 0.007 μmol) and ⁱBu₃Al (1 mL) in toluene and studied in polymerization of ethylene (1 bar). Alumina-modified SBA-15 was inactive, whereas all montmorillonite/R₃Al samples were proven to be activating supports with the best performance of 19.1 kg_PE∙mmol_Zr⁻¹∙h⁻¹ for K10/Me₃Al. The authors proposed that montmorillonite/R₃Al supports react with dichloro complex Zr27 in two pathways (Scheme 11). The first one was the ligand exchange and subsequent alkyl/halide abstraction; the support reacts as an organo-Lewis-acidic cocatalyst similar to MAO (Scheme 11a). An alternative activation process involves protonolysis of the M–R bond on Brønsted acidic centers (Scheme 11b). The Lewis and Brønsted acidities of the supports were determined using pyridine as a probe molecule by monitoring the bands for pyridine absorbed on BAS and LAS at 1540 and 1450 cm⁻¹, respectively. It was found LAS comprised the main fraction of the acid sites, detected in montmorillonite and montmorillonite/R₃Al. Treatment with R₃Al resulted in a decrease in LAS content, and it was notable that BASs were still detected, though in small numbers, after R₃Al treatment.

Scheme 11. Different activation pathways of a pre-catalyst complex considering the different nature of the acidic sites in the activating support: (a) via double alkylation; (b) via single alkylation and proton abstraction [175].

As a final note, the use of clay minerals did not receive further development due to the creation of more efficient activating supports (See Section 4 and Section 5).

3.5. Other Metal Oxides

A comparative study of the activation of Zr9 by different metal oxides (Al₂O₃, TiO₂, AlPO₄, TiO₂–Al₂O₃, SiO₂–Al₂O₃, TiO₂–SiO₂, and H–Y zeolite), calcined at 500 °C and treated with ⁱBu₃Al, in the polymerization of ethylene was conducted by Ikenaga et al. [167]. These oxides were also tested in the cracking of cumene to evaluate their acidity. Among the catalysts under study, the AlPO₄-supported catalyst showed the highest activity (230 kg∙mol⁻¹∙h⁻¹). The activity of the zeolite-supported catalyst was two times lower. Catalysts supported on other oxides demonstrated even lower activities. No correlation between the activity and the surface area of the oxides was observed. In cumene cracking (in view of the correlation between the cracking rate and Brønsted acidity), the highest average conversion was obtained using H–Y zeolite. Although SiO₂–Al₂O₃ showed the second-highest cracking activity, this oxide was ineffective as an activator for metallocene. Moreover, the most effective activator of polymerization, AlPO₄, showed negligible activity in the cracking of cumene (Figure 13).

Figure 13. Ethylene polymerization activity using acidic oxide-supported (n-BuCp)₂ZrCl₂ (Zr9), with acid properties as determined by cumene cracking: (1) polymerization (bar): Zr9 = 4.6 μmol, acidic oxide = 20 mg, TIBA = 0.52 mmol (Al/Zr = 115), temperature = 60 °C, and ethylene pressure = 0.70 MPa; (2) cracking (bent line): reaction temperature = 400 °C, catalyst = 0.10 g, carrier (He) = 30 mL∙min⁻¹, and pulse size = 0.5 μL; and (3) BET surface area of each oxide presented in parenthesis. Reprinted with permission from [167]. Copyright (2007) Elsevier B. V.

FT-IR spectral studies of pyridine absorption on Al₂O₃ and AlPO₄ showed more Lewis acidity of the former phase, and the absorption band of pyridinium ions, derived from BAS, was not observed in both cases. Under optimized conditions, the activity of Zr9 reached 820 kg∙mol⁻¹∙h⁻¹ (M_n = 10.0 kDa, Ð_M = 2.5). At the same time, activities of Zr2 and Zr1 were 18 and 66 kg∙mol⁻¹∙h⁻¹, respectively, and in the latter case PE with Ð_M = 5.3 was obtained. The formation of active sites on the AlPO₄ was attributed to the reaction of surface –OH groups with ⁱBu₃Al [167]. Nb₂O₅ was found to be ineffective in the activation of Zr25′; only Nb–O–ZrMe₂(η⁵-C₅Me₅) species have been detected [82].

Modification of silica by MgO, LiOH, or NaOH resulted in the formation of new phases capable of activation of rac-[SiMe₂(η⁵-4,5,6,7-tetrahydroinden-1-yl)]ZrCl₂ (Zr28)/ⁱBu₃Al in the polymerization of ethylene and propylene [176]. The highest catalytic activities were detected when using SiO₂/MgO support, they comprised 70 kg_PE∙g_Zr⁻¹∙h⁻¹ and 340 kg_PP∙g_Zr⁻¹∙h⁻¹. The acidity of the surfaces was estimated using a set of indicators, and acid strength in the range of 5.6 < pKa < 3.0 was detected for SiO₂/MgO. The acid strengths of SiO₂ and MgO and Mg(OH)₂ were found to be weak [177]; therefore, in SiO₂/MgO phase, SiO₂ and MgO have undergone a reorganization of the bonding structure with a formation of new acidic sites that are required for high activity of zirconocene in propylene polymerization. The role of the basic anionic sites had been also raised in [176], without mechanistic interpretation.

4. Sulfated Metal Oxides and Related Systems

More than 40 years ago, it was shown that the acidic sulfate treatment of metal oxides such as ZrO₂, TiO₂, SnO₂, and Fe₂O₃ increases the surface acidity and catalytic activity in the processes involving carbenium ions [178]. The Hammett acidity function (H₀) for superacid sites of sulfated metal oxides (SMO) was up to <−16. The common methods of the preparation of SMO are the treatment of precipitated hydroxides with H₂SO₄ with subsequent calcination, thermal decomposition of metal sulfate hydrates, or treatment of crystalline oxides with H₂SO₄. The features of the methods, as well as specific reaction and thermal treatment conditions, depend on the type of metal oxide and result in supports with different degrees of sulfation, –OH group content, and porosity.

Most of the studies of Group 4 metal complexes, supported on sulfated metal oxides, have been conducted by Marks and colleagues, and their significant proportion relates specifically to the catalytic hydrogenation of arenes [179,180,181]. Along with that, attention has been paid to the nature of the catalytic species and coordination polymerization of α-olefins, which matters for our review as well.

4.1. Sulfated ZrO₂

The preparation of active sulfated ZrO₂ (ZrS) is based on the calcination of sulfate-doped zirconium hydroxide or zirconium sulfate hydrates [182,183]. Active SZ is based on tetragonal zirconia (t-ZrO₂) as the main phase. ZrS, which is based on the more stable monoclinic ZrO₂, is much less active than tetragonal ZrS [184]. High-resolution transmission electron microscopy (HRTEM) studies of ZrS showed that the presence of SO₄²⁻ groups stabilizes small tetragonal ZrO₂ crystallites and induces the formation of well-faceted particles with the (110) plane as the most abundant crystallographic plane [185]. Periodical DFT calculations and FT-IR spectral studies showed a broad variety of structures for sulfated surfaces on t-ZrO₂ [184].

In 1998, Ahn and Marks reported the preparation of ZrS by thermal decomposition of Zr(SO₄)₂∙4H₂O with subsequent thermal activation under high vacuum (5 × 10⁻⁶ Torr) at 300, 400, and 740 °C [83]. ZrS³⁰⁰ contained strong BAS/LAS and weak BAS, ZrS⁴⁰⁰ contained both strong BAS and LAS, and ZrS⁷⁴⁰ contained mostly LAS. Next, Zr2′ and Zr25′ were adsorbed by acid supports from n-pentane solutions. The complex Zr2′ demonstrated high activity in hex-1-ene hydrogenation, and catalytic activity correlated with strong BAS populations. Less sterically hindered Zr25′ showed a dramatic enhancement in hydrogenation activity when supported on ZRS^400:, for example, it mediated fast hydrogenation of benzene at 25 °C and 1 bar H₂. Zr2′/ZrS⁴⁰⁰ and Zr25′/ZrS⁴⁰⁰ also catalyzed ethylene homopolymerization at 25 °C with activities of 1.5 and 40 kg∙mol⁻¹∙bar⁻¹∙h⁻¹, respectively. ¹³C CP MAS NMR spectra of the Cp₂Zr(¹³CH₃)₂/ZrS⁴⁰⁰ sample showed the presence of the signals of cationic Zr–Me⁺ species, without the signals of the transferred methide group. The authors have proposed that the sulfated zirconia activates Group 4 metallocenes via Zr–C protonolysis with a formation of ion pairs (Scheme 12). This reaction mechanism is qualitatively different from the Zr→Al migratory mechanism that is characteristic for the γ-Al₂O₃ activating support.

Scheme 12. Activation of Cp₂ZrMe₂ by sulfated zirconia (ZrS) [83].

More thorough investigation of the efficiency of ZrS⁴⁰⁰ supports in activation of different Group 4 metal complexes [84] showed that catalytic activities are reduced in a raw Zr(CH₂Ph)₄ > Zr(CH₂SiMe₃)₄ > Zr(CH₂^tBu)₄ > Zr25′ for both ethylene and liquid propylene polymerizations. ¹³C CP MAS NMR spectral studies of Zr2′/ZrS⁴⁰⁰ and Zr25′/ZrS⁴⁰⁰ showed the presence of both μ-oxo and cationic species, the signals of transferred methide groups were not reliably detected. These results confirm weaker Lewis acidity of sulfated zirconia than that of dehydroxylated γ-Al₂O₃.

The further studies of Group 4 polymerization catalysts by Marks’ group [186] were focused on well-known post-metallocene Hf complex Hf3 (Figure 14a), exhibiting excellent catalytic activity in ethylene homopolymerization and ethylene/α-olefin copolymerization [187,188,189,190,191,192]. In the ¹³C CP MAS NMR spectrum of Hf3/ZrS the downfield-shifted broad signal at δ = 65.5 ppm (Figure 14b, D) can be assigned to a “cation-like” electron-deficient LHf¹³CH₃⁺ species, as evidenced by the marked downfield shift from δ = 60.2 ppm for Hf3 pre-catalyst (Figure 14b, A). A similar downfield-shifted signal was observed at δ 65.8 ppm in the homogeneous [LHf¹³CH₃]⁺[MeB(C₆F₅)₃]^– [193]. DFT modeling of chemisorption of Hf3 on the ZrS surface yielded an average distance d(Hf–O) of 2.14 Å that is slightly elongated in comparison with typical Hf–OR bonds (1.92–2.03 Å).

Figure 14. (a) Immobilization of Hf3 (LHfMe₂) on ZrS support; (b) ¹³C CPMAS NMR spectra (100 MHz; repetition time, 5 s; contact time, 2 ms; 10 kHz spinning speed) of (A) LHfMe₂, 876 scans; (B) LHfMe⁺/ZrS⁻, 10,000 scans; (C) ¹³C labeled LHf(¹³CH₃)₂ (60 scans); and (D) LHf¹³CH₃⁺/ZrS⁻, 1600 scans. *—Rotational sidebands, #—impurities. Reprinted with permission from [186]. Copyright (2018) American Chemical Society.

Such close contact could not fail to affect the catalytic activity. In ethylene homopolymerization, activities of Hf3/ZrS and homogeneous Hf3/[Ph₃C]⁺[B(C₆F₅)₄]⁻ were 4.5 and 6260 kg∙mol⁻¹∙bar⁻¹∙h⁻¹, and activities in ethylene/oct-1-ene copolymerization amounted to 3.9 and 25,600 kg∙mol⁻¹∙bar⁻¹∙h⁻¹. And, it is highly significant that oct-1-ene incorporation in the transition from homogeneous to supported catalyst has fallen from 50 to 2.6 mol% [186]. Another important difference between homogeneous and supported catalysts consisted in opposite directions of first ethylene insertion, at Hf–Aryl and Hf–Me bonds, respectively.

In the search for more efficient catalytic systems, pyridylamido complexes Hf4 and Hf5 (Scheme 13) were synthesized and studied [194]. Being supported on ZrS, Hf4 and Hf5 have demonstrated higher activities in comparison with Hf3: 15.7 and 22.7 vs. 4.5 kg∙mol⁻¹∙bar⁻¹∙h⁻¹ in ethylene homopolymerization; 11.5 and 16.6 vs. 3.9 kg∙mol⁻¹∙bar⁻¹∙h⁻¹ on ethylene/oct-1-ene copolymerization.

Scheme 13. Pyridylamido complexes of Hf, studied in activation and polymerization on sulfated zirconia [194].

4.2. Sulfated Al₂O₃

Sulfated alumina (AlS), prepared by the impregnation of γ-Al₂O₃ with 1.6 M H₂SO₄ and subsequent 3 h calcination at 550 °C, was studied as an activating support for (C₅Me₅)₂ZrMe₂ (Zr17′) by Marks and colleagues in 2003 [85]. ¹³C CP MAS NMR spectroscopy revealed that chemisorption proceeds via Zr–Me bond protonolysis with a formation of cationic (C₅Me₅)₂ZrMe⁺ species. In ethylene polymerization experiments (toluene, 60 °C, 10 bar), catalytic activities of Zr complexes were 2100 (Zr(CH₂Ph)₄), 1200 (Zr(CH₂SiMe₃)₄), 1100 (Zr25′), 160 (Zr17′), and 80 (Zr2′) kg∙mol⁻¹∙h⁻¹ [85]. Under similar conditions, activities of these complexes, supported on γ-Al₂O₃, were an order of magnitude lower.

In 2010, Marks’ research group revisited the study of the activation of half-sandwich and sandwich Zr(IV) complexes by AlS. In particular, they prepared nanosized sulfated alumina (n-AlS) by calcination of commercial 25 nm γ-Al₂O₃ to remove residual surfactant, surface modification with H₂SO₄, and final calcination; TEM analysis indicated that the nanoparticles were loosely aggregated nanorods (diameter∼10 nm, lengths∼10–30 nm) [195]. When studying the chemisorption of Zr17′ on n-AlS by ¹³C CP MAS NMR spectroscopy, both (C₅Me₅)₂ZrMe⁺ and Al–Me species were detected in contrast with AlS. Experiments on ethylene (co)polymerization on Zr(CH₂Ph)₄/n-AlS and Zr25′/n-AlS showed higher catalytic activities in n-heptane relative to activities in toluene, which is generally not surprising.

Increased understanding of chemisorption of Zr(IV) complexes Zr2′ and Zr25′ on AlS was the result of further research with the use of Zr X-ray absorption spectroscopy (XAS) and DFT modeling [179]. DFT modeling reveals two types of sulfate species, sites S_A (formed by double-exchange/condensation with the surface and does not afford an acidic proton) and S_B (formed by single exchange/condensation reaction with the surface, preserving one acidic proton, which is then transferred to an Al-O surface site) as well as (Al)_nOH hydroxyl groups, with the OH coordination to three or two Al ions. Zr–Me protonolysis is favored at the (Al)₃OH site, with a formation of Cp₂ZrMe⁺ or (C₅Me₅)ZrMe₂⁺ species coordinated at sulfate groups. DFT modeling at site S_B and experimental studies with the use of Zr K-edge extended X-ray absorption fine structure (EXAFS) and X-ray absorption near-edge structure (XANES) methods yielded a d(Zr–O) of 2.37 Å, which was significantly longer than in typical covalent Zr–O bonds. Very similar results were obtained when studying (C₅Me₅)Zr(CH₂Ph)₂⁺ species [180] (Figure 15).

Figure 15. Energy-minimized computed chemisorbed Cp*Zr(CH₂Ph)₂⁺ catalyst structures on sulfated alumina surface between site S_A (a) and S_B (b) (Cp*—η⁵-C₅Me₅). Reprinted and corrected with permission from [180]. Copyright (2015) American Chemical Society.

A very important and interesting aspect of the use of AlS is associated with the emergence of polymerization stereoselectivity for chemisorbed, initially achiral pre-catalysts such as Zr25′ [196]. At 25 °C, Zr25′/AlS produced polypropylene and poly(hex-1-ene) with 89% and >95% isotacticity, respectively. The requisite Zr25′/AlS C₁ symmetry was attributed to a combination of AlS surface topography and sterically imposing groups beyond the active site. DFT confirmed that propylene coordination and insertion are favored at the re enantioface (Figure 16), which results in isotactic polymerization via a ‘back-skip’ mechanism.

Figure 16. (a) re and si face propylene insertion transition states with Zr25′/AlS; (b) computed Gibbs free energy profiles for si vs. re face propylene activation/insertion at the Zr25′/AlS catalytic center. Reprinted with permission from [196]. Copyright (2021) Wiley-VCH Verlag GmbH and Co.

Chemisorption on AlS and catalytic activity of post-metallocene constrained-geometry Group 4 metal complexes Zr29′–Zr31′ and Ti18 (Scheme 14) in (co)polymerization of ethylene were investigated by Williams and Marks in 2009 [197]. ¹³C CP MAS NMR spectral studies showed that chemisorption of Zr complexes proceeds with a formation of cationic LZrMe⁺ species in two reaction pathways, via Zr–Me bond protonolysis and Zr→Al methide transfer: in contrast with activation of Zr17′, surface Al–Me fragments were detected during activation of constrained-geometry complexes.

Scheme 14. Constrained-geometry Group 4 metal complexes, selected for activation with sulfated alumina [197].

Experiments on polymerization of ethylene (10 mL of toluene, 5 bar) were conducted with the use of supported catalysts and under homogeneous conditions with the use of perfluoroborate activator [Ph₃C]⁺[B(C₆F₅)₄]⁻ (Table 3). In all cases, supported catalysts were significantly less active than the homogeneous catalysts, presumably due to steric hindrance from the ancillary ligand-support interaction blocking the ethylene approach to the active site. However, the use of supported catalysts resulted in the formation of high-MW PE, whereas borate-activated Zr30′ and Zr31′ produced polyethylene waxes. Ti complex Ti15 was more active than Zr catalysts. For ethylene/1-hexene copolymerizations at 25 °C, binuclear complex Ti18/AlS was found to be the most active (16.4 kg∙mol⁻¹∙h⁻¹), followed by Zr31′/AlS, Zr30′/AlS, and Zr29′/AlS (1.24, 1.21, and 0.54 kg∙mol⁻¹∙h⁻¹, respectively). In this way, the cooperative ‘binuclear effect’, observed previously by Marks and colleagues for bimetallic Group 4 constrained-geometry complexes in homogeneous conditions [198,199,200], does not occur when using supported catalysts.

Table 3. Ethylene polymerization on supported (AlS) and nonsupported constrained-geometry catalysts [197].

Catalyst	T, °C	[M], μmol	Reaction Time, h	Activity, kg∙mol−1∙h−1	Tm of PE, °C
Zr29′/AlS	25	4.46	3	1.21	138
Zr30′/AlS	25	4.02	3	3.51	138
Zr30′/AlS	60	2.85	3	4.93	137
Zr31′/AlS	25	5.60	3	1.66	139
Zr31′/AlS	60	4.46	3	1.19	137
Ti18′/AlS	25	2.36	3	14.4	136
Ti18′/AlS	60	2.19	3	20.6	137
Zr30′/BA 1	25	10.0	0.75	127	n. d. 2
Zr31′/2 BA	25	5.0	1.25	87	n. d.

¹ Homogeneous polymerization, BA—[Ph₃C]⁺[B(C₆F₅)₄]⁻. ² No data, low-MW PE obtained (M_n = 610 and 760 Da, respectively).

During activation of a highly efficient industrialized pyridylamido-naphthyl Hf pre-catalyst Hf3 (see Figure 17) by AlS, both protonolysis and methide Hf→Al transfer were detected [194]. Hf EXAFS spectrum of Hf3/AlS allowed to determine the ligand environment of Hf as two Hf−N (2.30 and 2.08 Å), one Hf−C (2.26 Å), one Hf−Me (2.21 Å), and one Hf−O_support (2.06 Å). Comparative studies of supported Hf3 and pyridylamido complexes Hf4 and Hf5 showed that when using AlS these complexes demonstrate close catalytic activities (35.1–46.8 kg∙mol⁻¹∙bar⁻¹∙h⁻¹ in ethylene homopolymerization, 16.6–19.4 kg∙mol⁻¹∙bar⁻¹∙h⁻¹ in ethylene/oct-1-ene copolymerization), which was not observed when using ZrS as an activating support [194]. Oct-1-ene incorporation was also dependent on the pre-catalyst and support used (Figure 17).

Figure 17. Effects of pre-catalyst structure and support on (a) activity and (b) oct-1-ene incorporation in ethylene copolymerization using the indicated surface molecular catalysts. Reprinted with permission from [194]. Copyright (2021) American Chemical Society.

A rather high efficiency of AlS as activators for Group 4 metallocenes could not fail to interest chemists and chemical engineers employed in the polyolefin industry. AlS was the subject of a number of patents, filed by Chevron Phillips. In particular, in ethylene polymerization with the use of zirconocene Zr9, AlS, and Et₃Al activators, catalytic activities up to 3.23 kg∙g_AlS⁻¹∙h⁻¹ have been reached [92]. However, it should be borne in mind that catalytic efficiency of metallocene/AlS/R₃Al systems is critically dependent on many factors such as the characteristics of Al₂O₃, method of sulfation, [SO₄²⁻]/m_alumina ratio, conditions of calcination, nature of Group 4 metal complex and R₃Al, and others. Therefore, for example, the highest activity of 3.23 kg∙g_AlS⁻¹∙h⁻¹ was achieved when AlS was prepared by impregnation of (NH₄)₂SO₄ solution onto Ketjen grade B alumina (1.5 mmol of sulfate per 1 g of alumina), followed by vacuum-drying and calcination in air at 550 °C [92]. The role of the pre-catalyst can be illustrated by the data from the patent [201] (Scheme 15, Table 4), the maximum productivity of 17.3 kg∙g_Cat⁻¹∙h⁻¹ was achieved when using Zr36/ⁱBu₃Al in homopolymerization of ethylene in the absence of H₂. In this patent, only calcining conditions are given (heating to 550 °C with a 240 °C∙h⁻¹ ramp rate in air and 3 h hold at 550 °C).

Scheme 15. Metallocenes studied in ethylene polymerization using AlS/R₃Al support [201].

Table 4. Ethylene polymerization activities of Zr complexes, supported on sulfated Al₂O₃ (30 bar, isobutane) [201].

Zr Complex	T, °C	Time, min	AlS Weight, mg	Zr Comp. Weight, mg	R3Al/ mmol	PE, g	Activity, kg∙mmol−1∙h−1	Mn, kDa	Mw, kDa
Hf6	90	60	100	3.0	iBu3Al/0.5	294	80.8	296	911
Hf6	105	35	100	3.0	iBu3Al/0.2	203	95.6	239	730
Hf7	90	60	100	3.0	iBu3Al/0.5	252	68.1	315	972
Hf7	105	33	100	3.0	iBu3Al/0.2	186	91.3	318	843
Zr32	80	30	100	1.0	iBu3Al/0.25	315	455.5	305	754
Zr32	90	30	105	1.0	iBu3Al/0.5	295	426.5	263	639
Zr32	90	30	104	1.0	Bu3Al/0.5	320	462.7	278	708
Zr33	90	30	101	1.0	iBu3Al/0.5	272	400.9	223	591
Zr33	90	30	108	1.0	Bu3Al/0.5	211	311.0	314	750
Zr34	90	60	50	2.0	Bu3Al/0.5	158	60.4	311	772
Zr34	90	30	100	2.0	iBu3Al/0.25	255	195.1	205	637
Zr35	90	60	100	2.0	iBu3Al/0.25	42	14.0	ins.
Zr35	105	60	100	2.0	iBu3Al/0.25	63	21.1	ins.
Zr36	90	16	50	1.0	iBu3Al/0.5	232	655.2	70	183
Hf8	90	60	100	3.0	iBu3Al/0.5	294	82.4	108	375
Hf8	100	60	100	3.0	iBu3Al/0.5	369	103.4	97	267

For the synthesis of PE composites, the chemists of Chevron Phillips proposed the use of binary mixtures of metallocene catalysts, the first being the catalyst of the formation of low-MW PE (M_w 5–50 kDa, for example, (η⁵-Indenyl)₂ZrCl₂ (Zr37), and the second being the formation of high-MW PE (M_w 200–1200 kDa, for example, Zr33 [202]. AlS support was prepared by the reaction of bohemite (‘Alumina A’ from W. R. Grace & Co., Columbia, MD, USA) with aq. (NH₄)₂SO₄, followed by vacuum-drying and calcination at 400–600 °C. The combination of binary pre-catalyst, AlS, and ⁱBu₃Al allowed to obtain ethylene/hex-1-ene copolymers with unique mechanical and rheological characteristics that may be employed for the production of large diameter pipes. Polymerization experiments were conducted with the use of molecular hydrogen (350 ppm with respect to the ethylene flow) at 92–95 °C. Maximum catalytic activities amounted to 2.84 and 3.16 kg_PE∙g_Cat⁻¹∙h⁻¹ for ethylene homo- and copolymerizations, respectively.

The importance of pre-treatment of AlS by R₃Al was the subject of the patent [203]: when using metallocene dichlorides, such pre-treatment resulted in significant increase of the catalytic activity. From the mechanistic viewpoint, pre-treatment of AlS by organoaluminum compounds results in the substitution of all BAS by –OAlR₂ fragments, and this modified support retains the ability to absorb LZrCl₂ via ligand exchange and Zr→Al alkyl transfer to LAS. However, from the analysis of scientific periodicals, this issue remains unexplored. An efficient, but incomprehensible, method to increase the efficiency of metallocene/AlS/R₃Al systems, proposed in the patent [204], is based on using simultaneous mixing of the catalyst components that results in ~30% improvement of catalytic activity. This effect still remains unexplored with the use of modern analytical methods.

4.3. Other Sulfated Metal Oxides

Sulfated tin (IV) oxide microparticles were prepared by sedimentation of Sn(OH)₄ gel at pH 8, elimination of residual Cl^– by aq. NH₄OAc, and drying and reaction with 3M H₂SO₄, followed by calcination at 500 °C for 3 h [86]. ¹³C CP MAS spectral study of chemisorbed Zr17′ showed the formation of cationic species, Sn_S–O–Zr(C₅Me₅)₂Me were detected in trace amounts. In ethylene polymerization experiments (toluene, 60 °C, 10 bar), catalytic activities of Zr complexes were 660 (Zr(CH₂Ph)₄), 98 (Zr25′), 15 (Zr17′) and 8.6 (Zr2′) kg∙mol⁻¹∙h⁻¹.

In 2004, Marks and colleagues undertook a comparative study of sulfated metal oxides ZrS, AlS, SnS, FeS, and TiS as activators for Zr2′ [87]. Solid-state ¹³C NMR spectroscopic analysis of the Cp₂Zr(¹³CH₃)₂ adsorbate complexes reveals that each support affords a different ratio of μ-oxo and cationic species. The chemical shift of the signals of Me groups in cationic species (δ = 37.1, 35.7, 37.6, and 33.8 ppm for ZrS, SnS, FeS, and TiS, respectively) correlated with the efficiency of activating supports. In ethylene polymerization experiments (60 °C, 10 bar, using 30 mg of the catalyst), Zr25′ and Zr(CH₂Ph)₄ have demonstrated wide spectra of catalytic activities depending on the type of support (Table 5); sulfated zirconia was the most efficient activating support.

Table 5. Ethylene polymerization activities of Zr complexes, supported on sulfated metal oxides [87].

Support	Zr Complex	[Zr], molZr/mg	Time, min	Activity, kg∙mol−1∙h−1	Tm of PE, °C	Active Site % 1
ZrS	Zr25′	1.43 × 10−7	2	770	133.1	~65
ZrS	Zr(CH2Ph)4	1.37 × 10−7	0.75	2500	134.0
AlS	Zr25′	3.9 × 10−8	10	1100	133.6	87 ± 3
AlS	Zr(CH2Ph)4	3.9 × 10−8	8	2100	134.2
SnS	Zr25′	4.4 × 10−8	60	98	136.1	61 ± 5
SnS	Zr(CH2Ph)4	4.4 × 10−8	30	660	135.1
FeS	Zr25′	9.0 × 10−8	30	24	134.2	22 ± 2
FeS	Zr(CH2Ph)4	9.0 × 10−8	7	510	135.2
TiS	Zr25′	1.85 × 10−7	120	1.2	138.4	63 ± 9
TiS	Zr(CH2Ph)4	1.85 × 10−7	60	63	138.0

¹ Determined by poisoning Zr25′/SMO ethylene polymerization with neopentyl alcohol.

5. Fluorinated Metal Oxides and Related Systems

As was shown in Section 3, some metal oxides have demonstrated the ability to activate Group 4 metallocenes and post-metallocenes, which is primarily due to the high Lewis acidity of the metal centers on the support surface. The obvious way to increase the Lewis acidity consists in the replacement of the oxygen atom in the ligand environment of the acidic metal center by a more electronegative element that is fluorine, without options. On the other hand, partial chlorination can also be effective.

However, in any case, the starting oxide should already be acidic. And this is the reason why fluorinated silica was not used in catalysis [205] and, in particular, has demonstrated zero efficiency as an activating support for zirconocenes [89,90]. On the contrary, in the case of alumina and especially silica/alumina, surface modifications by introducing halogen (F, Cl) resulted in supports with high acidity and bright prospects for use in the polyolefin industry.

5.1. Fluorinated Al₂O₃

During the studies of gas-phase disproportionation of fluorochloroalkanes with the use of γ-Al₂O₃ catalyst, fluorinated alumina Al₂O₃(F) was formed [206]. γ-Al₂O₃ alone was inactive and required pre-treatment with HF (or fluorocarbon); in this way, it is Al₂O₃(F) that is a true catalyst. Although chemical analyses reveal significant amounts of F in the solid after this activation procedure, nearly no structural changes of the γ-Al₂O₃ were detected by X-ray diffraction (XRD) [206]. Only after a few days, crystalline α-AlF₃ becomes detectable. Other approaches to Al₂O₃(F) are based on ‘wet’ methods, the reaction of γ-Al₂O₃ with NH₄F or with NH₄HF₂ in aqueous media [207], and the content of F in the reaction product is easily determined by γ-Al₂O₃/[F⁻] ratio. The reaction of Al₂O₃ with HF (Equation (1)) is exergonic (ΔG = −329 kJ∙mol⁻¹), and the formation of AlF₃ during the ‘fluorination’ of alumina should not be ruled out.

Al₂O₃ + 6 HF → 2 AlF₃ + 3 H₂O

(1)

In this way, both prolonged gas-phase fluorination or an excess of fluoride when being used in a ‘wet’ process may result in a high fluorine content with a formation of AlF₃ after calcination. The problem is that the most stable crystalline phase of AlF₃, namely α-AlF₃, does not exhibit the properties of strong Lewis acid since the surfaces derived from α-AlF₃ are covered completely with basic fluoride anions, irrespective of the plane indexes [208]. It is relevant to remind here that crystalline AlF₃ was found to be a low-efficient activating support for Zr1 in the presence of Me₃Al in the early work of Soga and Kaminaka [75]. It is quite possible that these unsuccessful experiments were conducted with large-sized crystallites of α-AlF₃, whereas AlF₃ showed high Lewis acidity only as β-AlF₃ crystal modification or in the amorphous form [208,209]. Efficient methods of the synthesis of high-surface amorphous AlF₃ have been developed in the early 2000s [210,211]. These methods consist of a fluorolytic sol–gel reaction starting from aluminum isopropoxide, followed by an activation step under the flow of a fluorinating gas. A similar method was also used for the preparation of Al₂O₃(F) with relatively low [F]/[Al] ratios (0.2–0.5) by the reaction of Al(OCH(Me)Et)₃ with HF in ^IPrOH with subsequent hydrolysis and calcination [212]. Sol–gel methods are too complex, and fluorination of γ-Al₂O₃ seems preferable for the production of highly acidic Al-based supports.

However, the chemical nature of the surface of Al₂O₃(F) is not entirely clear. Back in the 2000s, Fischer et al. performed a solid-state NMR study of fluorinated γ-Al₂O₃ that showed the presence of the species with ^VIAl(O_6−nF_n) (n = 1, 2 or 3) environments [213]. With an excess of F⁻, the remaining weakened Al–O–Al bonds anchoring ^VIAl(O₃F₃) to the bulk of alumina were eventually broken with the formation of AlF₃∙3H₂O.

The nature of LAS and BAS formed on the alumina surface during fluorination and the efficiency of fluorinated Al₂O₃ supports in the activation of zirconocenes Zr2 and Zr2′ were studied by Panchenko et al. [214]. Four samples of Al₂O₃(F) with different F content, were prepared by treatment of γ-Al₂O₃ (specific surface area 196 m²∙g⁻¹, pore volume 0.58 cm³∙g⁻¹) with aqueous solutions of NH₄F and subsequent calcination in air at 500 °C for 2 h. Then, the supports were treated with a toluene solution of Zr2 or hexane solution of Zr2′ using a ratio of 0.05 g Zr per gram of support. Ethylene polymerization was conducted at 80 °C and an ethylene pressure of 10 bar; in the case of Zr2/γ-Al₂O₃ and Zr2/Al₂O₃(F), ⁱBu₃Al was added.

Determination of LAS and BAS on the surfaces of γ-Al₂O₃ and Al₂O₃(F) was conducted using additionally dehydroxylated samples (500 °C, 0.02 Torr, 1 h) by FT-IR studies of the adsorption of CO (LAS) and pyridine (BAS). The concentrations of LAS and BAS (C_S) were estimated from the integrated intensity of the CO absorption bands in the region 2175–2240 cm⁻¹ and pyridinium ion absorption bands with a maximum of 1535–1550 cm⁻¹ using Equation (2).

$$C_S=\frac{A}{A_0\cdot p\cdot {10}^{-3}}$$

(2)

where C_S—concentration of acidic sites (μmol∙g⁻¹), A—integrated absorption of the examined band (cm⁻¹), A₀—integral adsorption coefficient for CO or pyridine adsorbate concentration 1 μmol∙g⁻¹ (cm∙μmol⁻¹), and p—amount of the sample per 1 cm beam, g∙cm⁻². The values of A₀ for CO absorption bands at 2215, 2210, 2202, 2198, and 2180 cm⁻¹ are 1.15, 1.10, 1.02, 0.95, and 0.8 cm∙μmol⁻¹, respectively; for pyridinium, the ion absorption band A₀ is equal to 3 cm∙μmol⁻¹.

The strength of BAs was characterized by the proton affinity (PA), which was calculated from the IR spectra of adsorbed pyridine by Equation (3). Characteristics of the supports are given in Table 6.

$$PA=\frac{\lg (3400-{\nu }^{cg})}{0.0023}-51$$

(3)

where 3400 cm⁻¹ is the frequency of N–H stretching vibration for free pyridinium ion, and ν^cg is the position of the gravity center of N–H stretching vibration for pyridinium ion on the support surface.

Table 6. Characteristics of the supports and ethylene polymerization results [214].

Support	CS of Type L1 and L2 LAS, μmol∙g−1	CS of Type L3 LAS, μmol∙g−1	CS of BAS, μmol∙g−1	PA, kJ∙mol−1	Zr2–Based Catalyst			Zr2′–Based Catalyst
					Zr wt% (μmol∙g−1)	PE Yield, g	Activity, kg∙mol−1∙h−1	Zr wt% (μmol∙g−1)	PE Yield, g	Activity, kg∙mol−1∙h−1
γ-Al2O3	60, 660	–	35	1380	1.21 (133)	3.3	25	4.07 (447)	4.5	9
Al2O3(1%F)	380	20	4	1280	0.62 (68)	11	162	1.93 (212)	6.4	30
Al2O3(3%F)	180	50	2	1190	0.60 (66)	20	303	2.53 (278)	46	166
Al2O3(5%F)	120	50	7.3	1190	0.62 (68)	61	895	2.53 (278)	149.5	538
Al2O3(7%F)	100	50	6.1	1190	0.66 (73)	30	414	2.53 (278)	51	183

Aluminas, calcined at temperatures above 500 °C, contain four types of LAS. In the IR spectra of adsorbed CO, these LAS are characterized by absorption bands at 2178–2182 (L-1), 2185–2195 (L-2), 2203–2210 (L-3), and 2225–2240 cm⁻¹ (L-4). The L-1, L-2, and L-3 LAS correspond most likely to coordinatively unsaturated aluminum atoms in an octahedral environment, while L-4 corresponds to those in a tetrahedral or pentahedral environment. Weak LAS of the L-1 and L-2 types were found on the surface of Al₂O₃(F) samples. In addition, the spectra of CO, adsorbed on Al₂O₃(F), contain absorption bands at 2215–2208 cm⁻¹ corresponding to new, stronger sites (L-3) (Table 6). LAS of maximum strength (absorption band 2215 cm⁻¹) were observed in the Al₂O₃(5%F) sample. As shown by Corma et al. [215], the addition of HF to Al₂O₃ can affect the strength of LAS via direct introduction of one F⁻ (L-2) or two F⁻ (L-3) into the coordination sphere of aluminum. As can be seen in Table 6, the introduction of fluorine into Al₂O₃ sharply decreases the total amount of LAS but increases the content of strong type L-3 LAS. In Al₂O₃(F) samples with a fluorine content of 3–7 wt%, strong BAS with PA 1190 kJ/mol were detected. The concentration of these sites was 2–7.3 μmol∙g⁻¹ with a maximum in the Al₂O₃(5%F) sample.

The content and strength of basic sites at the support surfaces were estimated by analyzing FT-IR spectra of adsorbed CDCl₃, and it was shown that fluorination of Al₂O₃ results in formation of weak basic sites.

The catalysts prepared with the use of dimethyl zirconium complex Zr2′ were active in the ethylene polymerization without additional cocatalyst ⁱBu₃Al, indicating the formation of the active sites containing a Zr–Me bond by the direct interaction of zirconocene with LAS and BAS of the support. In the polymerization run with ⁱBu₃Al, the activity was only slightly higher (620 kg_PE∙mol_Zr⁻¹∙h⁻¹) than that without ⁱBu₃Al (538 kg_PE∙mol_Zr⁻¹∙h⁻¹). The catalyst Zr2′/Al₂O₃ showed low activity; the activities of Zr2′/Al₂O₃(F) were noticeably higher with a maximum observed for Zr2′/Al₂O₃(5%F). In the presence of ⁱBu₃Al, the activity of Zr2/Al₂O₃(F) catalysts was higher than that of Zr2′/Al₂O₃(F). This result was attributed to the formation of inactive μ-oxo-like zirconium compounds during the reaction of Zr2′ with BAS of the support. The activity of Zr2/Al₂O₃(F) catalysts also depended on fluorine content in the support; Zr2′/Al₂O₃(5%F) showed a maximum activity (Table 6) [214]. In this way, Al₂O₃(F) was a more effective activating support in comparison with γ-Al₂O₃.

The above-discussed works are of more theoretical than practical value. Al₂O₃(F) was used for the activation of Cr-based catalysts of ethylene polymerization in 1992 [216]. In the early 2000s, the chemists of Phillips Petroleum (subsequently Chevron Phillips) paid special attention to fluorinated alumina and related phases as activating supports for Group 4 metallocenes. In particular, Al₂O₃(F) was prepared by the ‘wet’ method from Ketjen B alumina (Akzo Nobel, surface area 340 m²∙g⁻¹, pore volume 1.78 mL∙g⁻¹) with the use of NH₄HF₂ as an F⁻ source at different [F]/[Al] ratios [89]. After calcination at ~500 °C in dry air, these supports were tested in combination with Zr9 and Et₃Al in ethylene homopolymerization. Activities of 240–950 g_PE∙g_Cat⁻¹∙h⁻¹ were achieved (while the activity was 6.9 g_PE∙g_Cat⁻¹∙h⁻¹ when using γ-Al₂O₃ in a comparative experiment), and the relationship between efficiency of the support and fluorine content had a distinct maximum.

Preparation of Al₂O₃(F) from Ketjen B alumina by ‘wet’ method (optimal calcining temperature 600 °C) and from HPV alumina (W. R. Grace & Co., surface area 500 m²∙g⁻¹, pore volume 2.8 mL∙g⁻¹ after pre-calcining at 600 °C) by heating in fluidizing N₂ and n-C₆F₁₄ vapor at 600 °C was described in [92]. In the latter case, the support turned black presumably due to carbon deposition. In ethylene polymerization (90 °C, 38 bar) with Zr9 and Et₃Al, the highest activities of 1.25 and 1.27 kg_PE∙g_Cat⁻¹∙h⁻¹ were achieved for supports, which were prepared by ‘wet’ and gas-phase fluorination. After recalcining of ‘black’ support in the air at 600 °C, during which the black color turned back to white, catalytic activity increased to 2.18 kg_PE∙g_Cat⁻¹∙h⁻¹.

5.2. Fluorinated SiO₂/Al₂O₃

Silica/alumina appears not to be an efficient activating support per se (see Section 3.3); however, the chemical binding of fluorine proved to be able to fundamentally increase the ability of supports to provide activation of Group 4 metal complexes. When compared with other activating supports, it is this type of support that showed the best efficiency, which puts fluorinated SiO₂/Al₂O₃ on the top of the activating supports for Group 4 metallocenes and post-metallocenes. Deep and comprehensive corporate studies of this type of support have been undertaken (and apparently are ongoing) for Chevron Phillips (previously Phillips Petroleum), Total, and INEOS, which are the leading polyolefin manufacturers.

These studies were aimed at finding the most effective solutions in the development of MAO- and borate-free activating supports, and were based on well-defined and commercially available inorganic phases with known porosity and surface area. However, these studies were conducted in fundamentally different directions, namely, fluorination of the commercial silica/alumina phases and surface treatment of commercial silicas by organoaluminums with subsequent thermal treatment and fluorination (optionally, the latter stage was omitted when alkylaluminum fluorides were used). Consequently, these different approaches (Scheme 16) are discussed separately in our review.

Scheme 16. Two distinct approaches to fluorinated SiO₂/Al₂O₃. (a) Fluorination of surface –OH groups in SiO₂/Al₂O₃; (b) Surface modification of SiO₂ by aluminum alkyls, followed by oxidation and fluorination. The content of Si–OH and Al–OH surface fragments depends on calcination temperature and duration.

Note that some recent patents protect the somewhat unexpected type of activating supports, namely, fluorinated silica-grafted alumina (Al₂O₃/g-SiO₂(F), see Section 5.2.3).

5.2.1. Fluorination of SiO₂/Al₂O₃ (SiO₂/Al₂O₃(F))

The common ‘wet’ method of the fluorination of silica/alumina, which was studied since the early 2000s by the chemists of Phillips Petroleum, is completely similar to the ‘wet’ method of the preparation of fluorinated γ-Al₂O₃. In early experiments, the samples of SiO₂/Al₂O₃ MS 13-110 (W. R. Grace & Co., 13 wt% alumina, surface area 400 m²∙g⁻¹, pore volume 1.2 mL∙g⁻¹) were treated by aqueous NH₄HF₂ at different [F]/[Al] ratios [89]. After calcination at different temperatures, the supports were used for activation of the Zr9/Et₃Al catalyst. The results of polymerization experiments have demonstrated that both [F]/[Al] ratio and calcination temperature affect the polymerization activity, and optimal calcination temperatures vary greatly for supports, which are prepared at different [F]/[Al] ratios (Figure 18). In the absence of fluorinated SiO₂/Al₂O₃ support, the Zr9/Et₃Al system was inactive.

Figure 18. The function of catalyst activity on calcination temperature for Zr9/Et₃Al system and SiO₂/Al₂O₃(F) activating supports, prepared with different NH₄HF₂ loadings. Designed based on the data presented in [89].

Fluorinated SiO₂/Al₂O₃ (MS 13-110, W. R. Grace & Co.), prepared under optimized conditions, was studied in the polymerization of ethylene using Zr2/Et₃Al with and without pre-contacting of Zr2/Et₃Al with hex-1-ene before interaction with an activating support [93]. The reaction of Zr2 with Et₃Al and hex-1-ene (toluene/n-heptane, 30 min) prior to charging to the support resulted in a marked increase of activity (2932 vs. 1640 g_PE∙g_Cat⁻¹∙h⁻¹). Under similar conditions, bis(η⁵-fluorenyl) and (η⁵-fluorenyl)-Cp ansa-zirconocenes Zr36 and Zr37 (Scheme 17a) have demonstrated lower activities.

Scheme 17. (a) Structures of zirconocene pre-catalysts and catalytic activities in ethylene polymerization (in brackets, g_PE∙g_Cat⁻¹∙h⁻¹) of the corresponding heterogeneous supported catalysts, prepared with the use of fluorinated SiO₂/Al₂O₃ and Et₃Al; (b) formation of cyclic organoaluminum species during ‘pre-contacting’ of Zr2/Et₃Al with hex-1-ene [93].

An increase in catalytic activity was attributed in [93] to the formation of aluminum metallocyclic species (Scheme 17b). Despite the experimentally proven fact of the formation of similar species during the reaction of Zr2 with Et₃Al [28,29], their possible role in further activation and polymerization processes is unclear.

The addition of arylboronic acids to the solutions of Zr2 and Zr2′, followed by mixing with SiO₂/Al₂O₃(F)/ⁱBu₃Al activator, resulted in a marked increase of catalytic productivity in the case of Zr2′ [95]. One can assume that Zr2′ reacts with ArB(OH)₂ with a formation of Cp₂Zr–O–B(OH)Ar species, but the role of similar species in further transformations is unclear too.

Very promising results were obtained when Ketjen Grade B alumina was subjected to ‘wet’ (with final calcining at 500 °C) or gas-phase fluorination (at 600 °C), followed by gas-phase chlorination (CCl₄, 600 °C). The activities of the Zr9/Et₃Al system on these supports in ethylene polymerization reached 3.1 and 6.3 kg_PE∙g_Cat⁻¹∙h⁻¹, respectively [92].

5.2.2. Treatment of SiO₂ by Organoaluminums with Subsequent or Simultaneous Fluorination (SiO₂/g-Al₂O₃(F))

The synthesis of type SiO₂/g-Al₂O₃(F) supports and their use for activation of Group 4 metallocenes and post-metallocenes are described and discussed in a number of patents (owned by Elf Atochem/Total and INEOS) and Total in cooperation with the scientific group of Boisson [99,100,101,102,103,104] and only a few scientific articles [105,106,107].

The method of the preparation of SiO₂/g-Al₂O₃(F) from silica, developed by Elf Atochem/Total, comprises the following stages (Scheme 18):

Scheme 18. (a) Thermal decomposition of (NH₄)₂SiF₆; (b) preparation of SiO₂/g-Al₂O₃(F) [107].

• Treatment of silica with organoaluminum compound in the inert organic solvent;
• Calcination of the product, at the final stage—in the presence of O₂;
• Fluorination using (NH₄)₂SiF₆ and final calcination.

Fluorination is based on the reaction of HF, which was released at a high temperature (typically 450 °C) from the decomposition of the (NH₄)₂SiF₆. An alternative method of fluorination is based on the reaction of silica with Et₂AlF, followed by calcination. When using Et₂AlF, there is no need for fluorination by (NH₄)₂SiF₆; a sufficient amount of fluorine is introduced at the stage of the reaction between silica and Et₂AlF.

An alternative approach to SiO₂/g-Al₂O₃(F) was developed by INEOS [99,100]. The method of INEOS is based on the reaction of SiO₂ with fluorinated alkoxides of alkylaluminum, followed by calcination (see below).

In Table 7, we tried to summarize the data on the synthesis of SiO₂/g-Al₂O₃(F) supports and the results of their use for activation of Group 4 metal complexes in α-olefin polymerization. In addition to the zirconocenes Zr2, Zr3, Zr5, Zr9, Zr26, and Zr37 and constrained-geometry complex Ti8, already mentioned in our review, the complexes rac-[Me₂Si(η⁵-inden-1-yl)]ZrCl₂ (Zr38) and rac-[Me₂Si(η⁵-2-Me-4,5-benzoinden-1-yl)]ZrCl₂ (Zr39) have been studied. The structures of complexes, investigated in combination with SiO₂/g-Al₂O₃(F) activating supports, are presented in Scheme 19.

Table 7. The types and characteristics of SiO₂; reagents and conditions of the preparation of SiO₂/g-Al₂O₃(F) activating supports and their catalytic applications.

Silica/Specific Surface Area, m2∙g−1 / Pore Volume, mL∙g−1/ Particle Size, μm	Organo- Aluminum Compound	Stages 2, 3; Calcination Regime	Stage 3: Fluorinating Reagent/3rd Calcination Regime	Pre-catalyst/wt% or (mmol∙g−1)	R3Al	Polymer/ Activity, g∙gCat−1∙h−1 (kg∙mol−1∙h−1)	Ref.
Grace 332/300/1.65/70	(BuO)2AlOSi (OEt)3	20→130 °C, 1 h 130→450 °C, 1 h 450 °C, 1 h	6 wt% (NH4)2SiF6 / 20→130 °C, 1 h 130→450 °C, 1 h 450 °C, 1 h	Zr2/(53)	iBu3Al	PE/—(900)	[88]
				Zr2/(53)	Et3Al	PE/—(490)	[88]
				Zr37/(53)	iBu3Al	PE/—(6720)	[88]
				Zr38/(53)	iBu3Al	PE/—(2020)	[88] 1
Grace 332/300/1.65/70	Et3Al	20→130 °C, 1 h 130 °C, 1 h 130→450 °C, 1 h 450 °C, 4 h	5 wt% (NH4)2SiF6/ 30→450 °C, 2 h 450 °C, 2 h	Zr5/n. d.	iBu3Al	PE/39 (3400)	[107]
Grace 332/300/1.65/70			10 wt% (NH4)2SiF6/ 30→450 °C, 2 h 450 °C, 2 h	Zr5/n. d.	iBu3Al	ethylene—hex-1-ene (20 mol%)/550 (75,000)	[107]
Crossfield ES70X/276/1.54/53				Zr5/n. d.	iBu3Al	ethylene—hex-1-ene (20 mol%)/725 (91,000)	[107]
Grace 332/300/1.65/70	Et2AlF	20→130 °C, 1 h 130 °C, 1 h 130→450 °C, 1 h 450 °C, 1 h	–	Zr38	iBu3Al	PE/820 (–)	[102]
Grace 332/300/1.65/70				Zr39/1.0	iBu3Al	PP/700 (–) 2	[102,107]
Grace 332/300/1.65/70				Zr5/	iBu3Al	ethylene—hex-1-ene (20 mol%)/110 (19,000)	[107]
MS-3030/320/3/85				Zr5/0.4	iBu3Al	PE/160 (–)	[105]
MS-3040/420/3/80				Zr5/0.4	iBu3Al	PE/320 (–)	–”–
SIL 3/800/1/15				Zr5/0.4	iBu3Al	PE/0 (–)	–”–
SIL 19/700/2.2/15				Zr5/0.4	iBu3Al	PE/280 (–)	–”–
SIL 20/630/3.2/100				Zr5/0.4	iBu3Al	PE/400 (–)	–”–
SIL 30/300/3/100				Zr5/0.4	iBu3Al	PE/250 (–)	–”–
SP9 446/520/1.9/53				Zr5/0.4	iBu3Al	PE/1080 (–)	–”–
Grace 948/290/1.7/58				Zr5/0.4	iBu3Al	PE/700 (–)	–”–

¹ In [88], ethylene/hex-1-ene copolymerization results for Zr5/ⁱBu₃Al/AS are also mentioned without meaningful results. ² The percentage of isotactic pentads mmmm = 96%.

Scheme 19. Group 4 metal complexes, investigated with SiO₂/g-Al₂O₃(F) activating supports.

Polymerization experiments with Zr5 were conducted at 80 °C with ethylene solutions in n-heptane (6 wt%); the [ⁱBu₃Al]/[Zr] ratio was ~2000 [105]. Ethylene/hex-1-ene copolymerizations were also performed; the activities of supports have demonstrated similar trends. In both series of experiments, catalytic activities depended significantly on the types of silica used for the preparation of SiO₂/g-Al₂O₃(F). If the average pore diameter was too small (e.g., support based on silica SIL 3), there was no activity. However, no clear correlations between silica specific surface/pore volume and catalytic activity have been found. The comparison of polymerization kinetic curves for Zr5 supported on SiO₂/g-Al₂O₃(F) and on solid MAO (SMAO) showed noticeable deactivation of the catalysts prepared with activating supports [105].

The results of the study of SiO₂/g-Al₂O₃(F) with different fluorine contents (Table 3) as activating supports for Zr5 and Zr39 were reported by Boisson’ group in 2013 [107]. The results of four comparative polymerization experiments are included in Table 7; however, a number of the findings of this comprehensive study deserve a separate discussion. As shown in the Table 8, a significant variation of the Al content was detected for SiO₂/g-Al₂O₃(F), whereas increasing the amount of (NH₄)₂SiF₆ from 5 wt% to 20 wt% led to higher F content in the support. The treatment of silica with a higher amount of (NH₄)₂SiF₆ resulted in a degradation of the silica grain.

Table 8. Preparation and elemental analysis of activating supports [107].

Support	Silica 1	Alkylaluminum	(NH4)2SiF6 wt%	Al wt%	F wt%
AS1	Grace 332	Et3Al	5	3.72	1.18
AS2	Grace 332	Et3Al	5	4.75	1.74
AS3	Grace 332	Et3Al	10	3.67	3.31
AS4	Grace 332	Et3Al	20	3.81	5.30
AS5	Grace 332	Et2AlF	–	4.56	2.21
AS6	Grace 332	Et3Al	10	4.98	3.60
AS7	ES70X	Et3Al	10	3.01	3.98
AS8 2	ES70X	Et3Al	10	3.20	2.20
AS9 3	ES70X	Et3Al	5	3.26	2.34

¹ Properties of silicas are given in Table 2. Before treatment with alkylaluminum, silicas were calcined at 0.01 mbar and 450 °C, [Si–OH] = 1.3 mmol∙g⁻¹. ² Pre-calcining temperature was 200 °C. ³ Pre-calcining temperature was 300 °C.

As can be seen in Scheme 18, SiO₂/g-Al₂O₃(F) should contain Al–F bonds; however, the formation of Si–F bonds on the surface of SiO₂/g-Al₂O₃(F) is theoretically possible. Figure 19 shows that the spatial distribution of Al and F for the support AS1 (see Table 8) appears homogeneous; therefore, F atoms are mainly bonded to Al atoms. ²⁷Al MAS NMR analysis of AS2, AS3, and AS4 after exposure to wet air showed road resonances at 50, 32, and 3.5 ppm for AS2 and the appearance of the signal at −15 ppm for AS3. This signal can be attributed to AlF₃∙3H₂O (the reference spectrum of this complex was also recorded). The signal at −15 ppm became the main for AS4. ¹⁹F MAS NMR spectrum of AS7 showed the presence of the signal of AlF₃∙3H₂O (−8 ppm) and the intence signal at 12 ppm. This last resonance is close to the signal assigned to ^VIAl(O₅F) (9 ppm) and ^VIAl(O₄F₂) (20 ppm) by Fischer et al. [213].

Figure 19. EDX analysis of the support AS1. Reprinted with permission from [107]. Copyright (2013) American Chemical Society.

The support AS1 was studied as an activator for the Zr5/ⁱBu₃Al system in ethylene polymerization and ethylene/hex-1-ene copolymerization (~20 mg AS1, 0.5 μmol Zr5, 1 mmol ⁱBu₃Al, 4 bar ethylene, 80 °C). High activities (3.4–110 kg∙mmol⁻¹∙h⁻¹, 39–900 g∙g_Cat⁻¹∙h⁻¹) and a remarkable hex-1-ene activation effect were observed. T_m of polymer decreased from 133 °C (PE) to 104 °C (copolymer with 56 mol% of hex-1-ene). Copolymerization experiments with AS2–AS7 supports and Zr5/ⁱBu₃Al showed a high increase in activity as the wt % of F is increased on the carrier. The support AS5 prepared using treatment of the silica with Et₂AlF showed comparable properties to support AS2 made using 5 wt % of (NH₄)₂SiF₆, which had a similar content of F. In separate polymerization tests at 10 bar of ethylene (500 mL of heptane, 1 μmol Zr5, 2 mmol ⁱBu₃Al, 73 mg of AS8, 80 °C, 6 mL of hex-1-ene), productivity of 3200 g∙g_Cat⁻¹∙h⁻¹ was achieved without the catalyst’ leaching, as evidenced by perfect morphology of the polymer obtained (Figure 20).

Figure 20. Polyethylene particles obtained using the catalyst Zr5/ⁱBu₃Al/AS8 (heptane, 10 bar, 6 mL of hexene, 80 °C). Reprinted with permission from [107]. Copyright (2013) American Chemical Society.

During the comparison of Zr5/AS/ⁱBu₃Al and Zr5/MAO catalysts, better incorporation of hexene was observed with the activating support. MWD and TREF analysis data indicated the presence of multiple active sites for the Zr5/AS/ⁱBu₃Al catalytic system. Similar comparative studies were also performed for other Group 4 metal complexes Zr9, Zr26, Zr3, and Ti8 (Scheme 19) with the use of supports AS6 and AS7 (Table 9).

Table 9. Activation of a range of Group 4 metal complexes using AS6, AS7, and MAO [107] ¹.

Comp	Activator	Hex-1-ene mol%	Activity, kg∙mmol−1∙h−1	Activity, g∙gCat−1∙h−1	Mw, kDa	ĐM	Tm, °C
Zr5	AS6 (20 mg)	20	75	550	97	3.7	111
Zr5	AS7 (20 mg)	20	91	725	125	3.6	112
Zr9	MAO (2000 eq)	0	150		206	2.9	134
Zr9	MAO (2000 eq)	20	76		160	3.0	134
Zr9	AS6 (21 mg)	20	9.1	71	102	3.0	124
Zr9	AS7 (23 mg)	0	28	142	146	2.3	131
Zr9	AS7 (21 mg)	20	22	462	94	3.1	119
Zr26 2	MAO (2000 eq)	0	140		704	5.1	140
Zr26 2	MAO (2000 eq)	20	27		154	2.1	90
Zr26 2	AS6 (20 mg)	20	24	196	346	5.2	81
Zr26 3	AS7 (43 mg)	0	4.2	29	557	3.6	131
Zr26 2	AS7 (26 mg)	20	32	189	193	3.3	84
Zr3	MAO (2000 eq)	0	2.6		161	3.7	131
Zr3 2	MAO (2000 eq)	20	9.5		51	2.0	121
Zr3 2	AS6 (48 mg)	20	33	224	144	4.0	116
Zr3	AS7 (41 mg)	0	2.9	23	342	4.1	135
Zr3 2	AS7 (42 mg)	20	43	338	122	3.5	111
Ti8	MAO (2000 eq)	0	1.1		173	3.3	138
Ti8 2	MAO (2000 eq)	20	3.9		15	1.9	100
Ti8 2	AS6 (108 mg)	20	38	287	186	3.3	103
Ti8	AS7 (112 mg)	0	3.3	24	1022	5.7	133
Ti8 2	AS7 (112 mg)	20	12	95	356	5.5	101

¹ 300 mL heptane, 4 bar ethylene, 80 °C, 60 min, MAO 10 wt %. For Zr5, Zr9, and Zr26, [Zr] = 0.5 μmol, [Al] = 1 mmol; Zr3 [Zr] = 2 μmol, [Al] = 4 mmol; and for Ti8 [Ti] = 2.5 μmol, [Al] = 5 mmol. ² 15 min. ³ [Zr] = 1 μM.

The complex Zr9 showed a poor ability to insert hexene into the solution, but its behavior was highly improved when using the activating support. It should also be noted that activities (per mmol) of the ⁱBu₃Al/AS-activated complexes Zr3 and Ti8 in ethylene/hex-1-ene copolymerization were higher than activities of Zr3/MAO and Ti8/MAO. Better incorporation of hex-1-ene was also detected in all experiments with SiO₂/g-Al₂O₃(F).

At first glance, SiO₂/g-Al₂O₃(F) has demonstrated promising characteristics of the activating supports for zirconocenes; the catalysts obtained could have competed with conventional SiO₂/MAO supports. However, more rapid decay of the polymerization rate when using Zr5 on an activating support in comparison with Zr5/SMAO [105] has encouraged the research group of Boisson to perform a more thorough study of the kinetic behavior of SiO₂/g-Al₂O₃(F)-activated Zr5 and Zr9 in the polymerization of ethylene [106]. In this study, three types of silica (Table 10) were used for the preparation of activating supports by the method described in [107]. The properties of fluorinated support M3703/079A, prepared from Grace Sylopol 952, are presented in the fourth row of Table 10. The complexes were pre-contacted with the supports for 10 min before introduction in the reactor. In all the reactions, around 60 mg of AS was used and the content of zirconocene was 0.4 wt%. Polymerization experiments were conducted at 80 °C and an ethylene pressure of 10 bar, in 0.1M hex-1-ene solution in n-heptane.

Table 10. Characteristics of the silica used for the synthesis of the SiO₂/g-Al₂O₃(F) supports [106].

Silica	Pore Volume, mL∙g−1	Specific Surface Area, m2∙g−1	Particle Size, μm	Mean Pore Diameter, Å
Grace Sylopol 952	1.8	290	58	246
Grace Sylopol 948	1.7	290	58	232
SP9_446	1.9	520	53	131
M3703/079A	1.9	240	53	240

The study of the evolution of the M_w as a function of the reaction time using Zr5/AS/ⁱBu₃Al showed that the M_w and Ð_M decrease as the reaction rate progresses. From the deconvolution results, it was shown that the rate of the formation of high-MW fraction tends to decrease significantly after 15 min of reaction, which corresponds to the beginning of the catalyst deactivation. Replacement of ⁱBu₃Al by Et₃Al had no effect. However, Zr9 seemed to respond quite differently to ⁱBu₃Al and Et₃Al. In the absence of a sufficient quantity of alkylaluminum, the catalyst became active, but rapidly reached a maximum rate that was followed by total deactivation of in 20–30 min. In the case of ⁱBu₃Al as a co-catalyst, the addition of approximately 500–1000 mol Al/mol Zr was sufficient to stabilize the reaction rate at an intermediate value close to the maximum rate for an hour. These general trends were also observed with Et₃Al as a co-catalyst to Zr9, although as with the Zr5, much less Et₃Al was required than ⁱBu₃Al. Also, in the case of Et₃Al, the stable reaction rates that were obtainable were higher than those that were realized with ⁱBu₃Al. However, an increase in the concentration of Et₃Al has led to a significant decrease in activity due to the formation of heterobimetallic dormant species [106]. The results of the comparative study of two common pre-catalysts Zr5 and Zr9 showed that the influence of the pre-catalyst nature on catalytic properties of the complex system, comprising Group 4 metal complex, SiO₂/g-Al₂O₃(F) and alkylaluminum are still not understood.

Besides the η⁵-ligand (L) environment in L₂ZrX₂, the nature of X can also affect the catalytic activity. Therefore, for example, [EBTHI]ZrF₂ (Zr1″) demonstrated higher activity than Zr1 and [EBTHI]ZrMe₂ (Zr1′) under the same reaction conditions [103].

The high efficiency of SiO₂/g-Al₂O₃(F) activating supports caused competing polyolefin companies to search for alternative synthetic approaches to these phases to circumvent the Elf Atochem/Total patents. Such a search had proved successful during the studies of the chemists of INEOS who proposed the reaction of silica with Et₂Al(OR) (R = C₆F₅, CH₂CF₃, CH(CF₃)₂), followed by calcination, as an alternative synthetic pathway to SiO₂/g-Al₂O₃(F) [99] (Scheme 20). The [F]/[Al] ratios in these supports amounted to 1.96–2.26; these values were determined by the temperature of the pre-calcining of SiO₂ (CS2050 from PQ Corp.) in the range of 250–450 °C.

Scheme 20. Preparation of SiO₂/g-Al₂O₃(F) activating supports by the method of INEOS [99]. (a) Preparation of fluorinated alkoxide or aryloxide; (b) Treatment of the silica surface followed by calcination in an inert atmosphere and in the air.

These supports were compared with activating support SiO₂/g-Al₂O₃(F), which was prepared by the method of Total with the use of (NH₄)₂SiF₆ ([F]/[Al] = 1.68) [217], in copolymerization of ethylene and hex-1-ene in slurry and gas-phase experiments. Constrained-geometry complex (C₅Me₄SiMe₂N^tBu)Ti(η⁴-1,3-pentadiene) (Ti8″) was selected as a pre-catalyst. In slurry polymerization experiments, Ti8″ was pre-mixed with ⁱBu₃Al ([Al]/[Ti] = 50) and added to 0.1 g of SiO₂/g-Al₂O₃(F) (Ti loading of 30 μmol/g_AS), and polymerization was conducted in 1.7 L of isobutane containing 0.25 mmol ⁱBu₃Al at 90 °C; an H₂/ethylene ratio of 0.2 mol% and ethylene pressure of 10 bar were maintained for 1 h. The samples of SiO₂/g-Al₂O₃(F), which were prepared with the use of Et₂Al(OC₆F₅) and (NH₄)₂SiF₆, have demonstrated close activities (2.0–3.0 kg_PE∙g_Cat⁻¹∙h⁻¹). Before gas-phase polymerization experiments, solid catalysts were prepared by the above-described pre-mixing of the catalyst components, followed by washing with n-hexane and drying in vacuo. H₂/ethylene and hex-1-ene/ethylene ratios were maintained at 0.13 and 0.55 mol%, respectively, during polymerization at 80 °C and 10 bar. In these experiments, catalytic activities were close to 0.5 kg_PE∙g_Cat⁻¹∙h⁻¹ [99]. Comprehensive rheological studies of copolymers [100] showed that novel supported catalysts provide the formation of polymers with improved rheological properties. In particular, higher viscosities were observed at lower shear rates, providing better bubble stability and lower viscosities at higher shear rates, resulting in better processability in the extruder.

The further development of the activating supports of SiO₂/g-Al₂O₃(F) type by INEOS [101] consisted of the treatment of calcined SiO₂ (Sylopol 332, W. R. Grace & Co.) by 2,4,6-trimethylpyridine and POCl₃, followed by vacuum-drying (and optionally additional heating at 130 °C) and high-temperature fluorination with (NH₄)₂SiF₆. The comparison of these supports with SiO₂/g-Al₂O₃(F), prepared by the method of Total [217], in activation of Zr5 and Ti8″ for ethylene/hex-1-ene copolymerization showed marked increase in catalytic activities: 927 vs. 340 g_PE∙g_Cat⁻¹∙h⁻¹ for Zr5 and 500 vs. 410 g_PE∙g_Cat⁻¹∙h⁻¹ for Ti8″. As can be seen in Scheme 19, a relatively small number of Group 4 metal complexes have been studied in MAO- and borate-free ethylene (co)polymerization using SiO₂/g-Al₂O₃(F) supports. It can be assumed that the studies of other pre-catalysts can provide a breakthrough in the further development of single-site supported catalysts, which are suitable for use in the polyolefin industry.

5.2.3. Silylation of Alumina with Subsequent Fluorination (Al₂O₃/g-SiO₂(F))

‘Inverted’ phase Al₂O₃/g-SiO₂ can be prepared, for example (so-called ‘SIRAL’ phases), by the hydrolysis of a solution of aluminum hexanolate in hexanol with deionized water at 90 °C, followed by mixing the filtered alumina suspension with a solution of orthosilicic acid, and spray-drying. Hydrothermal treatment (5 h at 180 °C) before spray-drying results in the formation of phases with higher acidity [218]. Modification of alumina through silica addition creates (i) strong LAS through isomorphous substitution of tetrahedral Si⁴⁺ by Al³⁺ ions and (ii) a small quantity of highly acidic BAS, which are thought to be bridged OH species similar to those found in zeolites [218]. The surface structures of SIRALs with different [Al]/[Si] ratios are schematically presented in Figure 21.

Figure 21. Model depicting the surface composition of PURAL SB (A), SIRALs 1.5–5 (B), SIRALs 10–20 (C), SIRALs 30–40 (D), SIRALs 60–80 (E), and SIRALs 90–100 (F). Reprinted with permission from [218]. Copyright (2000) Elsevier B. V.

In the first patent of Chevron Phillips in this field [94], Al₂O₃/g-SiO₂(F) was prepared by the impregnation of commercial SIRAL 28M (Sasol) Al₂O₃/g-SiO₂ with NH₄HF₂ in methanol with subsequent calcining. In ethylene polymerization on Zr32/ⁱBu₃Al, Al₂O₃/g-SiO₂(F) provided better activity; it was far superior to that of Al₂O₃/SiO₂(F) and almost twice that of sulfated alumina [94]. In ethylene homopolymerization on Zr40 (Scheme 21), activated by ⁱBu₃Al, catalytic activity depended on the alumina to silica ratio, varying from 1.69 kg_PE∙g_Cat⁻¹∙h⁻¹ (fluorinated γ-Al₂O₃) to 7.30 kg_PE∙g_Cat⁻¹∙h⁻¹ (Al₂O₃/g-SiO₂(F) with Al₂O₃/SiO₂ = 1.5:1 by weight) [94]. Undoubtedly of interest are the results of the comparative study of catalytic activity of Zr5/ⁱBu₃Al and Zr40/ⁱBu₃Al on different activating supports in ethylene/hex-1-ene copolymerization (Table 11).

Scheme 21. Group 4 metallocene pre-catalysts studied in olefin polymerization with the use of Al₂O₃/g-SiO₂(F) activating supports.

Table 11. Catalytic activities of zirconocenes Zr5 and Zr40 on different activating supports in ethylene/hex-1-ene copolymerization and the values of tan δ for the copolymers obtained [94].

Activating Support (Al2O3/SiO2 by Weight)	Zr5		Zr40
	Activity, kg∙gCat−1∙h−1	Tan δ	Activity, kg∙gCat−1∙h−1	Tan δ
Sulfated Al2O3 (–)	0.18	2.89	1.27	14.43
Fluorinated Al2O3 (–)	0.02	1.34	3.11	4.12
Chlorinated Al2O3 (–)	0.07	1.88	0.36	1.53
SiO2/Al2O3 (0.15:1)	0.15	1.17	1.61	7.05
Al2O3/g-SiO2(F) (2.6:1)	9.10	5.78	6.35	5.44
Sulfated Al2O3/g-SiO2(F) (2.6:1) 1	6.69	3.69	4.94	9.99
Phosphated Al2O3/g-SiO2(F) (P) 2	6.41	5.00	5.13	8.79

¹ Al₂O₃/g-SiO₂ was treated by MeOH solution of NH₄HF₂ and H₂SO₄ with subsequent calcining at 600 °C for 3 h. ² H₃PO₄ was used instead of H₂SO₄.

This table also contains the tan δ (the ratio of loss and storage moduli, determined at 190 °C by small-strain (10%) oscillatory shear measurements, the shear frequency of 0.1 s⁻¹) related to the presence of long-chain branches (LCB) in copolymer (higher tan δ means that the polymer relaxes easily, with little storage of the strain, and that the polymer has relatively lower LCB).

The synthesis of Al₂O₃/g-SiO₂(F) having varied [Al]/[Si] ratios was also described in [94]. Pre-calcined (600 °C) Alumina A (W. R. Grace & Co., surface area ~300 m²∙g⁻¹, pore volume 1.3 mL∙g⁻¹) was treated with different amounts of Si(OEt)₄ in MeOH and further with NH₄HF₂. The product was calcined in N₂ at 600 °C. The highest activity for the Zr40/ⁱBu₃Al system in ethylene polymerization (6.3 kg∙g_Cat⁻¹∙h⁻¹) was observed for Al₂O₃/g-SiO₂(F) with Al₂O₃/SiO₂ ratio of 7.3 by weight.

Very high catalytic activities in the polymerization of ethylene were achieved for the Zr40/ⁱBu₃Al system when using fluorinated-chlorinated silica-coated alumina as an activating support [96]. This support was prepared by the treatment of Alumina A (W. R. Grace & Co., surface area 300 m²∙g⁻¹, pore volume 1.2 mL∙g⁻¹) with Si(OEt)₄/ⁱPrOH, followed by drying and calcining at 500–900 °C. The chlorination stage (4 h) involved injecting and vaporizing CC1₄ into the gas stream used to fluidize the Al₂O₃/g-SiO₂ during calcination at 500 °C every five minutes. The fluorination step was conducted in a similar way using tetrafluoroethane within 4.5 h. The maximum activity of 17.83 kg_PE∙g_Cat⁻¹∙h⁻¹ was demonstrated by Zr40, which was activated by ⁱBu₃Al and Al₂O₃/g-SiO₂(F,Cl) containing 4 wt% of Cl and 7 wt% of F at 95 °C and 28 bar of ethylene [96]. A higher efficiency of Al₂O₃/g-SiO₂(F,Cl) in comparison with other activating supports in ethylene polymerization was also detected for Zr32, Zr35, and Hf9 [98].

Another interesting application of the Al₂O₃/g-SiO₂(F,Cl) support was the activation of the mixture of pre-catalysts Hf6 and Zr41 in ethylene/hex-1-ene copolymerization [96]. The amounts of 1.2 mg of Hf6, 1.4 mg of Zr41, and 150 mg of support were used, and copolymerization was conducted under similar conditions, with the addition of 5 g of hex-1-ene and 175 ppm of the molecular hydrogen. In this experiment, the catalytic activity of 2.19 kg_PE∙g_Cat⁻¹∙h⁻¹ was achieved. In propylene polymerization experiments (70 °C, 31 bar) using the Zr42/ⁱBu₃Al/Al₂O₃/g-SiO₂(F,Cl) system, catalytic activities reached up to 5.6 kg_PP∙g_Cat⁻¹∙h⁻¹ [96].

Particularly noteworthy are relatively recent studies of ‘mixed’ catalytic systems, conducted by the chemists of Chevron Phillips [219,220]. The very idea of these studies was to develop a heterogeneous system containing both Ziegler–Natta and zirconocene catalysts. As described in [219,220], Al₂O₃/g-SiO₂(F) was prepared from Alumina A by treatment with Si(OEt)₄/ⁱPrOH, followed by drying, calcining at 600 °C, fluorination by NH₄HF₂/MeOH, and calcination. The Ziegler–Natta catalyst was prepared by the reaction of Al₂O₃/g-SiO₂(F) with Bu₂Mg (toluene, 90 °C, 3 h), followed by treatment with TiCl₄ (90 °C, 3 h). Activation of zirconocenes Zr33, Zr40, and Zr41 by ⁱBu₃Al and this support gave highly efficient catalysts of the copolymerization of ethylene with hex-1-ene [219,220]. When using Zr40/ⁱBu₃Al at 90 °C and 27 bar of ethylene, catalytic activities were 49.4 and 44.9 kg∙g_Cat⁻¹∙h⁻¹ in the absence and in the presence (880 ppm) of the molecular hydrogen. In the absence of zirconocene, Al₂O₃/g-SiO₂(F)-based Ziegler–Natta catalyst was less active (26.5 and 10.8 kg∙g_Cat⁻¹∙h⁻¹ in the absence and in the presence (1000 ppm) of the H₂).

Another method of the preparation of ‘Ziegler–Natta’ Al₂O₃/g-SiO₂(F) catalysts/supports was based on the interaction of Al₂O₃/g-SiO₂(F) with solutions of active metal chlorides and MgCl₂ in THF, followed by centrifugation and drying in vacuo [219,220]. When TiCl₄ was used as active metal chloride, a very active catalyst was obtained (5.27 kg∙g_Cat⁻¹∙h⁻¹); however, after absorption of the Zr40/ⁱBu₃Al catalytic, activity tripled (~15 kg∙g_Cat⁻¹∙h⁻¹). Zirconocenes Zr33 and Zr41 were studied in combination with this type of active support with the aim of obtaining polyethylenes having a broad spectrum of characteristics.

The third method of the preparation of Ti/Mg Al₂O₃/g-SiO₂(F), proposed in [219,220], is close to the common method of the preparation of TMZNCs by the reaction of Mg(OEt)₂ with TiCl₄ [221]: Mg(OEt)₂ reacted with Al₂O₃/g-SiO₂(F) in toluene, with subsequent reflux with TiCl₄. Zirconocenes Zr33 and Zr41 with this type of Ti/Mg Al₂O₃/g-SiO₂(F) support have demonstrated maximum activities of 22.1 and 11.7 kg∙g_Cat⁻¹∙h⁻¹, respectively, in copolymerization of ethylene with hex-1-ene in the presence of 100 ppm of H₂. In almost all samples of TMZNC/zirconocene catalysts, supported on Al₂O₃/g-SiO₂(F), the [Zr]/[Ti] ratio was ~3:10, from which it follows that the zirconocene component of the binary catalyst is far more active.

5.2.4. Fluorinated SiO₂/Al₂O₃ in Oligomerization of Higher α-Olefins

SiO₂/Al₂O₃(F), prepared by the ‘wet’ method from MS13-110 silica/alumina (W. R. Grace & Co., 13 wt% Al₂O₃, surface area 400 m²∙g⁻¹, pore volume 1.2 mL∙g⁻¹), was used as an activating support for the series of Group 4 metallocenes (17 examples) in a comparative study of the catalytic activity in oligomerization of oct-1-ene [97]. ⁱBu₃Al was used as an organoaluminum activator for L₂MCl₂ before the addition of the activating supports. The structures of the most active pre-catalysts are presented in Scheme 22; the results of the oligomerization experiments are given in Table 12. As can be seen from this table, TOF of 10⁵ h⁻¹ and higher can be achieved when using SiO₂/Al₂O₃(F), which is comparable to activities of the best oligomerization catalysts under homogeneous conditions with perfluoroborate activation [222,223].

Scheme 22. Group 4 metallocene pre-catalysts that have demonstrated high efficiency in the oligomerization of oct-1-ene when activated by SiO₂/Al₂O₃(F) and ⁱBu₃Al [97].

Table 12. Oligomerization of 1-octene, catalyzed by zirconium complexes, and activated by TIBA and fluorinated aluminosilicate [97].

Cat.	[Mon]/[Zr] Ratio	T, °C	TOF, h−1	KV100	VI
Zr2	1.1·105	90	3.1·103	8.9	211
Zr5	4.2·105	115	4.3·104	136	210
Zr6	1.5·106	70	1.3·105	130	222
Zr8	3.1·105	110	3.4·104	62	186
Zr9	5.6·105	105	4.1·104	45	175
Zr35	1.1·105	90	9.0·103	23	169
Zr43	4.3·105	120	3.1·104	159	214
Zr44	4.4·105	120	3.6·104	132	200
Zr45	1.0·105	100	5.2·103	10.3	194
Hf10	5.5·105	90	2.5·104	8.3	157

However, this topic has not received further development. It can be assumed that SiO₂/Al₂O₃(F)-based α-olefin oligomerization catalysts have a fundamental flaw resulting from the high acidity of the support. The consequence of such acidity is a possibility of cationic skeletal rearrangements during oligomerization, which leads to the deterioration of the characteristics of α-olefin oligomers in terms of their use for the production of engine oil feedstocks.

5.3. Fluorinated Compositions of SiO₂ or Al₂O₃ with Other Metal Oxides

The preparation of fluorinated supports, based on silica and Group 4 metal oxides, was described in the patent of Phillips Petroleum [90]. The silica/titania, prepared by the cogellation method (8 wt% of TiO₂, surface area of ~450 m²∙g⁻¹) was calcined at 600 °C, treated with NH₄HF₂, and calcined at different temperatures. The temperature of the final calcination had a significant effect on the efficiency of these phases when being used as activating supports for the Zr9/Et₃Al catalyzed polymerization of ethylene. Therefore, for example, when using supports calcined at 600 and 450 °C, catalytic activities were 1164 and 2837 g_PE∙g_Cat⁻¹∙h⁻¹_. Note that fluorinated TiO₂ had demonstrated the lack of activation in Zr9/Et₃Al catalyzed polymerization of ethylene. Under the same conditions, the samples of fluorinated silica/zirconia have demonstrated a higher efficiency in comparison with fluorinated silica/titania (up to 5 kg_PE∙g_Cat⁻¹∙h⁻¹).

Impregnation of alumina by Zr(OBu)₄ with subsequent calcination (dry N₂) and fluorination by perfluorohexane at 600 °C resulted in the formation of fluorinated alumina/zirconia with low own activity in ethylene polymerization after treatment with Et₃Al (35 g_PE∙g_Cat⁻¹∙h⁻¹) [91]. When using fluorinated alumina/zirconia as an activating support for Zr9, the catalyst productivity increased to 1.38 kg_PE∙g_Cat⁻¹∙h⁻¹, which was comparable to the efficiency of the fluorinated alumina support [91]. The presence of high-MW ‘Ziegler–Natta’ PE, formed on fluorinated alumina/zirconia catalytic centers, without the participance of metallocene, allows to consider the product of polymerization as a prospective PE composite.

5.4. Chlorinated Metal Oxides

Ketjen Grade B alumina and Davison Grade 952 silica were chlorinated by the reaction with CCl₄ vapor at 600 °C. After the addition of a toluene solution of Zr9 and Et₃Al, a moderately active ethylene polymerization catalyst was formed by chlorinated alumina (~10³ g_PE∙g_Cat⁻¹∙h⁻¹), whereas chlorinated silica was found to be inactive [91]. Impregnation of alumina by Zr(OBu)₄ with subsequent calcination (dry N₂) and chlorination resulted in a solid phase, which was active in ethylene polymerization after treatment with Et₃Al; the addition of Zr9 resulted in a further increase of activity (862 and 1484 g_PE∙g_Cat⁻¹∙h⁻¹, respectively). The efficiency of chlorinated silica/zirconia support in the activation of Zr9 was moderate [91].

The studies of other methods of the chlorination of HPV alumina were reported in the Chevron Phillips patent [92]; SO₂Cl₂ at 300 °C and AlCl₃ at 250 °C were found to be fairly efficient chlorinating agents (activities of 459 and 401 g_PE∙g_Cat⁻¹∙h⁻¹ in ethylene polymerization when using Zr9/Et₃Al). Chlorination of alumina by CCl₄ under optimized conditions allowed to reach the catalytic activity of 2.8 kg_PE∙g_Cat⁻¹∙h⁻¹ for Zr9/Et₃Al [92].

6. Other Inorganic Supports

There are only a few articles that address patents devoted to the study of inorganic activating supports distinct from the supports discussed above. In 1995, Soga and colleagues demonstrated the ability of solid heteropolyacids H₃[PMo₁₂O₄₀] and H₅[PMo₁₀V₂O₄₀] to activate Zr2 and Zr5 in the polymerization of ethylene and propylene, respectively [224]. Catalytic activities were low (up to 23 and 3.7 kg∙mol_Zr⁻¹∙h⁻¹). Dimethyl derivative Zr2′ was inactive, on the basis of which they concluded that activation of zirconocene pre-catalyst L₂ZrCl₂ proceeds through Cl abstraction.

Impregnation of the silica, alumina, or SiO₂/Al₂O₃ by different metal chlorides with subsequent calcination allowed to obtain solid supports with different activating efficiency [225,226,227]. Active supports were prepared from Ketjen Grade B alumina using Cu(II), Sn(IV), Ag(I), Nb(V), Mn(II), W(VI), La(III), Nd(III), Sb(V), Zn(II) [225], Mo(VI) [226], and Ni [227] salts. The activation of Zr9/Et₃Al in comparative experiments on the homopolymerization of ethylene (38 bar, 90 °C) allowed the identification of the most efficient systems, which were based on Sn, Ag, and Zn chlorides. When the loading of ZnCl₂ was 20 wt% of alumina, after calcining with CCl₄ and impregnation with Zr9/Et₃Al, the catalytic activity of 11.8 kg∙g_Cat⁻¹∙h⁻¹ was achieved [225].

Cation-exchanged fluorotetrasilicic mica (Mⁿ⁺-mica, Mⁿ⁺ = Na⁺, Mg²⁺, Fe³⁺, Co²⁺, Ni²⁺, Cu²⁺, Zn²⁺) were prepared and employed as activating supports in the ethylene polymerization or ethylene/hex-1-ene copolymerization (7 bar, 60 °C) in the presence of L₂ZrCl₂/ⁱBu₃Al [228]. Mⁿ⁺-mica were calcined at different temperatures, and 100 mg samples of the support were contacted with 20 μmol of Zr2 or Zr9 with subsequent treatment by ⁱBu₃Al. Catalytic activities in Zr2 catalyzed homopolymerization of ethylene substantially depended on the nature of Mⁿ⁺ (Fe³⁺-mica demonstrated the highest productivity of 174 g_PE∙g_Cat⁻¹∙h⁻¹) and the temperature of calcination (200 °C was the optimal value). The authors proposed that the active sites were probably formed by the reactions of Cp₂ZrCl₂ and the exchanged cations Mⁿ⁺ since the activity was significantly increased by replacing Na⁺ ions with M²⁺ (M²⁺ = Mg²⁺, Co²⁺, Ni²⁺, Cu²⁺, Zn²⁺) and Fe³⁺. This work is of particular interest, but it is not applicable in practice due to the extremely low activity of the catalysts.

7. Organochemical Functionalization of Inorganic Surfaces and Organic Polymers

7.1. Organochemical Functionalization of Inorganic Phases

The very idea of low-cost surface modification of SiO₂, Al₂O₃, SiO₂/Al₂O₃, and other inorganic phases by C₆F₅Br and its analogs with a formation of perfluoroaryl-grafted oxides was patented by W. R. Grace & Co. in 1999 [229]. However, this patent did not include any data on catalytic activity.

In the early 2000s, the activation of post-metallocenes Zr46–Zr49 (Scheme 23) by surface-modified heteropolyacids and Et₃Al was studied by Fujita’s group [230]. The activating support of the formula (Ph₃C)_mH_n[PMo₁₂O₄₀]∙8H₂O (average m/n~2:1) was obtained by the reaction of H₃[PMo₁₂O₄₀]∙28H₂O with Ph₃CCl in acetone with subsequent evaporation, washing by toluene, hexane, and drying at 100 °C under vacuum. After the addition of 10 μmol of Zr complex to the mixture of 5 μmol of the support and 1 mmol of Et₃Al, fast ethylene polymerization was observed under atmospheric pressure and 25 °C, and activities of Zr46–Zr49 were 1.02–5.64 kg∙mol_Zr⁻¹∙h⁻¹; complex Zr49 demonstrated the highest activity. In the cases of Zr47 and Zr48, bimodal PEs were obtained, whereas complex Zr49 catalyzed the formation of PE with abnormally narrow MWD (Đ_M = 1.45). The nature of the supports and activation mechanisms were not studied and discussed in [230]. One can assume that the reaction of (Ph₃C)_mH_n[PMo₁₂O₄₀]∙8H₂O with Et₃Al results in the formation of soluble dimeric ethylalumoxane and Mo–O–AlEt₂ species and does not affect Ph₃C⁺—[PMo₁₂O₄₀]^m− interactions. Activation of L′ZrCl₂ proceeds through alkylation by Et₃Al and subsequent Zr→CPh₃ ethyl transfer with a formation of L₂ZrEt⁺ and 1,1,1-triphenylethane. This activation mechanism is very similar to the conventional mechanism of the activation of Group 4 metal complexes by [Ph₃C]⁺[B(C₆F₅)₄]⁻ [9].

Scheme 23. Zirconium complexes, studied in ethylene polymerization using (Ph₃C)_mH_n[PMo₁₂O₄₀] as an activator [230].

In 2007 [231], Jones and colleagues proposed the further development of the idea of solid activating supports, which consisted in surface modification of silica (mesoporous SBA-15) by perfluoroalkyl–SO₃H groups using fluorinated sultone precursor [232] (Scheme 24a). In the presence of Me₃Al, this support activated pre-catalyst Zr17′ in ethylene polymerization (25 °C, 4 bar), yielding productivities up to 1000 kg_PE∙mol_Zr⁻¹∙h⁻¹ without reactor fouling. During the study of the activation mechanism, possible ways of MAO formation and direct interaction of –SO₃H groups with Zr17′ were ruled out, which led to the structure of the activated catalyst presented in Scheme 24b.

Scheme 24. (a) Synthesis of tethered Brønsted acid sites by the reaction of silica with perfluorinated sultone; (b) hypothetical activated catalyst [231].

In 2011, Taoufik and colleagues showed that the treatment of silica surface with ⁱBu₃Al in Et₂O afforded a well-defined bipodal surface species [(≡SiO)₂Al(ⁱBu)(Et₂O)] [149] (Scheme 25). In further studies, this material was treated with C₆F₅OH (2.5 eq.) in benzene in the presence of one eq. of N,N-diethylaniline; repeated washing and vacuum-drying resulted in obtaining of activating support AS-1 (Scheme 25) [54]. The structure of the surface species was confirmed by diffuse reflectance infrared Fourier transform (DRIFT) spectroscopy and ¹H, ¹³C, ¹⁹F, and ²⁷Al CP-MAS NMR.

Scheme 25. Preparation of a bis(pentafluorophenoxy)aluminate activator [54].

The catalytic performance of the support was studied in ethylene polymerization with and without hex-1-ene, using common Zr5 and Zr9 pre-catalysts and ⁱBu₃Al. Reference tests were performed with silica-supported MAO (sMAO). The results of the experiments are presented in Table 13 [54].

Table 13. Ethylene (co)-polymerization results from the combination of AS-1 and sMAO solid activators with zirconium metallocene complexes Zr5 and Zr9 (80 °C, 4 bar, [ⁱBu₃Al] = 1 mmol, heptane, 30 min) [54].

Catalyst	Mol% hex-1-ene	Alsurface/ Zr Ratio	Wt% 1	[Zr], μmol	Activity, g∙gCat−1∙h−1	Activity, kg∙molZr−1∙h−1
Zr5/AS-1	–	9.6	1.18	2.0	250	8800
Zr5/AS-1	19.5	9.3	1.22	2.05	656	22,500
Zr9/AS-1	–	9.0	1.21	2.0	130	4300
Zr9/AS-1	19.5	9.0	1.21	2.0	133	4400
Zr5/sMAO	–	120	1.56	2.1	200	5370
Zr5/sMAO	19.5	120	1.56	2.1	620	16,600
Zr9/sMAO	–	148	1.32	2.0	326	11,400
Zr9/sMAO	19.5	148	1.55	2.0	230	8600

¹ Weight percent of metallocene in the supported metallocene catalysts.

The Zr5/AS-1 system was found to be more active in comparison with Zr5/sMAO; a major comonomer effect was observed. The initial activity was higher when using the activating support AS-1, but deactivation was observed during the course of polymerization. This deactivation could be ascribed to aryloxide transfer to Zr.

During the search of weakly coordinating surface anions, Conley and colleagues studied the reaction of Al(OC(CF₃)₃)₃∙PhF with calcined silica in perfluorohexane that resulted in the formation of highly acidic [≡SiOH^… Al(OC(CF₃)₃)₃] species [233] (Scheme 26a). Two years later, they published the results of further studies of this reaction [234]. It turned out that the reaction of Al(OC(CF₃)₃)₃∙PhF with partially dehydroxylated silica in fluorobenzene solvent, which induces proton transfer, results in the formation of [≡SiOAl(OC(CF₃)₃)₂(O(Si≡)₂) species (A in Scheme 26b) with extremely high Lewis acidity.

Scheme 26. (a) Formation of Brønsted (a) and Lewis (b) acidic sites during the reaction of Al(OC(CF₃)₃)₃∙PhF with silica; (c) activation of Cp₂Zr(¹³CH₃)₂ [234]. The key intermediates A, B, C and D are discussed below.

The ²⁷Al{¹H} MAS NMR spectrum of A contained a single signal that simulates as an isotropic chemical shift (δ_iso) of 74 ppm and a quadrupolar coupling constant (C_Q) of 18.0 MHz; DFT optimization of A predicted C_Q = 18.7 MHz. The calculated fluoride ion affinity (FIA) for A was 528 kJ∙mol⁻¹, which is significantly larger than the calculated FIA for B(C₆F₅)₃ (448 kJ∙mol⁻¹. The reaction of A with Cp₂Zr(¹³CH₂)₂ was complex. A ¹³C{¹H} CP-MAS NMR spectrum of Cp₂Zr(¹³CH₃)₂/A contains four signals at 38 (Cp₂Zr–¹³CH₃⁺ in species B, Scheme 26b), 23 (species C), 3 (Si–¹³CH₃ in species D), and –11 (Al–¹³CH₃ in species B and D) ppm. The pressurizing of Cp₂Zr(¹³CH₃)₂/A in cyclohexane by 20 bar of ethylene resulted in the formation of PE with an activity of 78.6 g_PE∙g_Cat⁻¹∙h⁻¹ [234]. The findings of this study are of great significance for understanding the chemistry of Si/Al-based activating supports and for developing new-generation MAO- and borate-free supported Group 4 metallocene and post-metallocene catalysts.

7.2. Functionalized Organic Polymers

Nafion (Scheme 27) is the first of a class of synthetic polymers with ionic properties; its molecular structure comprises poly(tetrafluoroethylene) backbone and perfluorovinyl ether groups terminated with sulfonate groups as side chains [235]. The presence of highly acidic –SO₃H groups and permeation for gases led the chemists of Mitsui Petrochemicals to the suggestion that Nafion can be used as an activating support for L₂ZrCl₂/R₃Al systems [236]. This assumption was confirmed by ethylene polymerization experiments (75 °C, pressure not indicated) using Zr5/ⁱBu₃Al; however, substantially cheaper sulfated polymer Amberlist (Scheme 27) showed comparable ability to activate zirconocene. The maximum activity of this catalyst can be roughly estimated as ~2 kg∙g_Cat⁻¹∙h⁻¹. Note that MgCl₂-supported Zr5/ⁱBu₃Al under the same conditions was about an order of magnitude less active [236].

Scheme 27. Structures of sulfated polymers used as activating supports for the Zr5/ⁱBu₃Al system [236].

8. Ionic Liquids

Ionic liquids (ILs) have long attracted the attention of researchers as carrier phases for different catalysts [237]. If the reaction proceeds in a non-polar medium, the use of ILs simplifies the separation of the products and recycling of the catalysts. In particular, ILs have already found applications in the cationic oligomerization of α-olefins [238,239,240,241,242]. Polymerization of ethylene and α-olefins with the use of coordination catalysts and ILs can be considered to be a heterogeneous process. The term ‘support’ implies belonging to the solid phase, so that the inclusion of ILs in this review is not entirely appropriate. However, ILs represent a convenient reaction medium for the study of activation processes, and we chose to include this short section in the present review.

In 1990 [243], Carlin and Wilkes reported the results of the study of ambient-temperature chloroaluminate molten salt AlCl₃∙MEIC (MEIC—1-ethyl-3-methylimidazolium chloride) as applied to activation of Ti2. When bubbling ethylene at 1 atm through 1.1:1.0 AlCl₃∙MEIC (6.8 g) containing 22 mg Ti2 and 0.118 g Al₂Me₃Cl₃, 20 mg PE was obtained. Similar Zr (Zr2) and Hf (Hf1) complexes were inactive under similar conditions. The authors proposed the formation of the species of the formula Cp₂TiMe-(μ-Cl)-AlCl₃ that are able to dissociate with a formation of active Cp₂TiMe⁺ species. The inertness of Zr2 and Hf1 was attributed by the authors to the higher strength of the M–Cl bond for M = Zr, Hf in comparison with Ti, but this explanation contradicts some reports, e.g., [244].

From 2007 to 2013, a number of articles on the subject of metallocene activation using ILs were published by the research group of Ochędzan-Siodłak [245,246,247,248,249]. In [245], the catalytic behavior of Ti2 in ethylene polymerization with 1-n-butyl-3-methylimidazolium tetrachloroaluminate [BMIM]⁺[AlCl₄]⁻ when using MAO, Et₂AlCl, and Et₃Al as alkylating agents was investigated. During polymerization experiments (n-hexane, 30 °C, 1 bar), a marked difference in the distribution of PE between IL and n-hexane phases was observed (Figure 22). At the beginning of the polymerization, the ionic liquid phase becomes white and swells considerably as the polyethylene appears. After 5–10 min, the hexane becomes a white suspension as the polyethylene is progressively shifted from the ionic liquid phase. The best results were obtained using Et₂AlCl, where the greatest amount of the polymer is in the hexane phase, whereas the amount of the polymer in the ionic liquid decreases and slowly reaches a constant level. After separation, this [BMIM]⁺[AlCl₄]^–/Ti2/Et₂AlCl phase maintained the ability to catalyze ethylene polymerization, at 30 °C and 1 bar activities, based on the weights of PE collected from the n-hexane phase, which were 9.0 and 8.8 kg_PE∙mol_Ti⁻¹∙h⁻¹ for freshly prepared and recycled catalysts, respectively. When using 1-ethyl-3-methylimidazolium tetrachloroaluminate and Ti2, the distribution of PE significantly deteriorated; even in the presence of Et₂AlCl, the main fraction of PE remained in the IL phase [246].

Figure 22. Influence of the kind of alkylaluminum compound and the polymerization time on the PE yield obtained using the Cp₂TiCl₂ catalyst in biphasic [BMIM]⁺[AlCl₄]⁻/hexane polymerization. The PE yield is presented separately for both phases. The alkylaluminum compound: (a) MAO, (b) Et₂AlCl (ionic liquid)/MAO (hexane), (c) AlEt₂Cl. Polymerization condition: Ti = 0.06 μmol, Al = 4 mmol, 1 bar. Reprinted with permission from [245]. Copyright (2007) Springer Nature.

The yield and distribution of PE between IL and aliphatic phases substantially depended on the nature of IL and the duration of polymerization. Comparative studies of ILs [C₈-mim]⁺[AlCl₄]⁻, [C₈-β-mpy]⁺[AlCl₄]⁻, and [C₈-γ-mpy]⁺[AlCl₄]⁻ (Scheme 28) as activators for Ti2/EtAlCl₂ in ethylene polymerization showed the advantage of pyridinium-based ILs both in activity and in PE distribution at a low Al/Ti ratio of 67 (Table 14) [247]. The best results were obtained with use of [C₈-mim]⁺[AlCl₄]⁻, for which the optimal Al/Ti molar ratio was 133. In all experiments, the use of the larger amounts of the EtAlCl₂ activator had a disadvantageous impact on the polymer transfer to the hexane phase. It was also shown that an active catalyst was formed and immobilized in IL phases, and no catalyst leakage was observed [247].

Scheme 28. Structures of ionic liquids studied as activators for Ti2/EtAlCl₂ in the polymerization of ethylene [247].

Table 14. Influence of the kind of the ionic liquid (IL) and the [AlEtCl₂]/[Ti2] molar ratio on PE yield in the biphasic IL/hexane polymerization of ethylene (30 °C, 5 bar).

IL	[AlEtCl2]/ [Ti2] Ratio	Reaction Time, h	PE Yield, kgPE∙molTi−1
			IL Phase	Hexane Phase	Total
[C8-mim]+ [AlCl4]−	67	1	10.7	46.7	57.3
	100	1	14.4	62.0	76.4
	133	1	17.0	101.0	118.0
	167	1	78.0	85.7	163.7
	100	2	13.7	146.7	160.3
	133	2	48.3	147.0	195.3
[C8-β-mpy]+ [AlCl4]−	67	1	8.0	68.0	76.0
	100	1	76.7	36.7	113.3
	133	1	101.7	41.0	142.7
	67	2	65.7	37.6	103.3
	133	2	103.3	105.0	208.3
[C8-γ-mpy]+ [AlCl4]−	67	1	5.3	59.8	65.1
	100	1	94.3	42.0	136.3
	133	1	171.5	37.5	209.0
	67	2	8.9	81.1	90.1
	133	2	42.0	236.7	278.7

The replacement of n-octyl substituent in [C₈-mim]⁺[AlCl₄]^– by the PhCH₂CH₂– fragment was found to be suitable to immobilize Ti2/EtAlCl₂, and PE was gathered mainly in the hexane phase [248]. An increase in the reaction time resulted in a considerable increase in the amount of PE and improved the PE transfer from the IL to hexane. Also, PhCH₂CH₂-substituted IL was more stable at high temperatures in comparison with [C₈-mim]⁺[AlCl₄]⁻.

In 2014, Ochędzan-Siodłak and Dziubek came up with the idea of the immobilization of Group 4 metal catalysts on IL-modified silica (SIL) [249]. The modification of SiO₂ was based on the reaction of surface Si–OH groups with (EtO)₃Si-finctionalized imidazolium chloride, followed by treatment with AlCl₃ and Et₂AlCl (Scheme 29). A comparative study of the pre-catalysts Ti2, bis[N-(salicylideno)anilinato]-titanium(IV) dichloride (Ti19), and N,N′-ethylenebis[5-chloro-salicylideneiminato]titanium(IV) dichloride (Ti20) showed that titanocene dichloride Ti2 catalyzed slurry polymerization of ethylene (30 °C, 5 bar, n-hexane) with a higher efficiency (activity up to 7200 kg_PE∙mol_Ti⁻¹∙h⁻¹) in comparison with other pre-catalysts (180 kg_PE∙mol_Ti⁻¹∙h⁻¹ for Ti19, 410 kg_PE∙mol_Ti⁻¹∙h⁻¹ for Ti20).

Scheme 29. Structures of ionic liquids studied as activators for Ti2/EtAlCl₂ in the polymerization of ethylene [247].

The Ti2/Et₂AlCl/SIL catalytic system was of spherical morphology (Figure 23a) and provided a formation of PE particles that have the shape of porous granules (Figure 23b). During polymerization experiments using Ti2/Et₂AlCl/SIL, a linear PE was obtained (T_m = 138–141 °C, degree of crystallinity 62–82%). Therefore, for example, in the experiment, during which the highest catalytic activity was achieved, the resulting PE had the following characteristics: M_w = 478 kDa, Ð_m = 3.1, T_m = 141 °C, and bulk density of 298 g∙dm⁻³.

Figure 23. SEM images of the SIL-metallocene catalytic system (a) and the obtained polyethylene (b). Reprinted with permission from [247]. Copyright (2014) Elsevier B. V.

9. Conclusions and Outlook

In our review, we summarized scientific periodicals and the patent literature related to the study of the MAO- and borate-free activating supports for Group 4 metallocenes and post-metallocenes and their use in catalytic polymerization of ethylene and α-olefins. In the frameworks of the current views on the mechanism of single-site polymerization catalysis, activation of Group 4 metallocenes and post-metallocenes lies in the formation of alkyl-metal cations, which are capable of α-olefin coordination/insertion. The mechanism of the formation of similar cations depends on the chemical nature of the activating support. If the support contains only strong Lewis acidic sites, activation proceeds via M(IV)→support alkyl migration (MgCl₂, γ-Al₂O₃ calcined at high temperatures). If the support contains mainly Brønsted acidic sites, activation proceeds via protonolysis of M(IV)–alkyl fragment, and catalytic activity depends on the strength of M(IV)–O(support) bonding (sulfated alumina). In the case of weak Brønsted acidic sites such as ≡Si–OH, inactive ≡Si–O–M(IV) species are formed. These three types of processes on the surface activating support are clearly represented in Figure 24 by an example of Cp₂ZrMe₂ activation [180].

Figure 24. Proposed structures of chemisorbed Cp₂ZrMe₂ on the surfaces of dehydroxylated Lewis acidic metal oxides (a), highly Brønsted acidic sulfated metal oxides (b), and weakly Brønsted acidic hydroxylated metal oxides (c). Reprinted with permission from [180]. Copyright (2015) American Chemical Society.

The surface chemistry of the activation of Group 4 metal complexes was studied by different methods, including MAS NMR, EXAFS, FT-IR spectroscopy, and DFT modeling. Among activating supports, γ-Al₂O₃, sulfated alumina, fluorinated alumina, and fluorinated alumina-grafted silica are the most thoroughly studied in terms of their surface structure and activation chemistry. Experimental and theoretical investigations in this field were carried out by scientific groups led by T. Marks, C. Boisson, V. Zakharov, M. Conley, and others. Among well-studied activating supports, sulfated alumina was in the top position in catalytic activity, and values of more than 17 kg∙g_Cat⁻¹∙h⁻¹ have been reported [201].

However, recent studies of the chemists of Chevron Phillips led to the development of a new type of activating support, namely, fluorinated silica-grafted alumina, Al₂O₃-g-SiO₂(F). It is noteworthy that the sequential high-temperature chlorination and fluorination of Al₂O₃-g-SiO₂(F,Cl) phases with even higher efficiency. A number of patents describe the use of these phases for the activation of Group 4 metallocenes and hybrid catalytic systems, which contain both single-site and Ziegler–Natta polymerization catalysts. Under optimal conditions and with proper selection of pre-catalysts, activities up to 50 kg∙g_Cat⁻¹∙h⁻¹ were achieved. Similar activating supports have not been thoroughly studied with respect to understanding mechanisms of activation and the nature of the active species and counterions. The presence of Cl in most active Al₂O₃-g-SiO₂(F,Cl) supports allows us to draw some analogies with other Group 4 metal catalysts with the obvious, still unclear, but undervalued role of M–(μ-Cl)–Al bonding in α-olefin polymerization [250].

It should also be noted that the total number of Group 4 metallocenes and post-metallocenes, studied in ethylene and α-olefin polymerization with the use of activating supports, is relatively small, with only a few dozen pre-catalysts. That is orders of magnitude less than the total number of pre-catalysts studied with the use of conventional MAO and borate activators. It is safe to assume that further search and design of Group 4 metallocenes and post-metallocenes, which are best suited for the use of activating supports and common alkylaluminums, will allow the creation of efficient heterogeneous catalysts for the production of advanced polyolefins. During this search, model reactions of α-olefin oligomerization [251,252,253] with simple end-group analysis and DP_n control as well as focusing on metal complexes, capable of homogeneous activation using R₃Al [254], deserve particular attention.