1 Introduction

Polyolefins are the most widely used class of polymers with a global consumption of approximately 211 million metric tons with an average price on the US market of $1,422.00 per metric ton [1]. Notwithstanding “green” concerns, polyolefins still surpass biopolymers in performance and energy efficiency [2]. While biopolymers rank high in terms of green design metrics, their production incurs relatively large environmental impact. On the other hand, polyolefins rank first in terms of life cycle assessment. Polyethylene is the simplest (and cheapest) polyolefin, having the general formula (–CH2–)n. It is typically a semicrystalline material, with mixture of interconnected crystalline and amorphous regions. In terms of microstructure, different polyethylenes are commercially available, with completely dissimilar chemical, physical, and mechanical properties. Polyethylene is conventionally classified as either: (i) high-density polyethylene (HDPE), (ii) low-density polyethylene (LDPE), or (iii) linear low-density polyethylene (LLDPE). HDPE has small amounts of branching (<1 %), with a density of 0.941 g/cm3, LDPE contains a mixture of long (>C6) and short branches (methyl, ethyl, butyl) with a density of about 0.92 g/cm3, while LLDPE has a high content of short branches (<C6) and a density less than 0.925 g/cm3 (Fig. 1) [3]. The type and degree of branching underlie many of the differing HDPE, LDPE, and LLDPE physical properties because they affect crystallinity and the extent of amorphous character [4].

Fig. 1
figure 1

Examples of commercial polyethylenes

Full size image

Historically, polyolefins were produced using highly active heterogeneous catalysts prepared from supported groups 4 or 6 halides, activated by aluminum alkyls. In the 1950s, the first olefin polymerization catalyst was reported by Phillips Petroleum constituted essentially of Cr/silica or Cr/silica–alumina (Fig. 2) [5–7]. This catalyst system, modified and optimized, is used today by various companies around the world, and accounts for a large share (~40 to 50 %) of the world’s HDPE production. In 1963 Karl Ziegler and Giulio Natta received the Nobel Prize in Chemistry “for their discoveries in the field of the chemistry and technology of high polymers” [8]. Heterogeneous Ziegler–Natta catalysts are binary combinations of supported TiCl4 and AlR3 (R = alkyl, aryl, hydride) on MgCl2 (Fig. 2), and are still employed on a large scale to produce isotactic polypropylene (iPP) and HDPE [9–12]. In the late 1980s, molecule-based group 4 metallocene homogeneous olefin polymerization catalysts emerged from discoveries by Walter Kaminsky [13–15]. A methylaluminoxane (MAO, Fig. 3) cocatalyst, a partially hydrolyzed trimethyl aluminium reagent, was found to be an efficient activator for these homogeneous group 4 metallocene catalysts (Fig. 3) [13–15]. This highly active homogeneous system allows for tuning of the product polymer microstructure (M w, PDI,  % comonomer incorporation, tacticity) exclusively by varying the organic ancillary ligands surrounding the group 4 metal [14, 16–24].

Fig. 2
figure 2

Proposed structures for Phillips catalyst (left) and supported Ziegler–Natta catalyst on MgCl2 (right)

Full size image
Fig. 3
figure 3

Selected examples of metallocene precatalysts and activating cocatalysts

Full size image

Another major advance in olefin polymerization catalysis was the independent discovery in the 1990s by the Marks group [25, 26] at Northwestern and Ewen and co-workers at Exxon Chemical Company [27] of catalytically active, isolable, and structurally well-defined, 1:1 catalyst/cocatalyst ion pairs derived from metallocene dialkyls and the organo-Lewis acid B(C6F5)3 (Fig. 4) [28–30].

Fig. 4
figure 4

Molecular structure of the mononuclear ion-pair polymerization catalyst [(1,2-(CH3)2C5H3)2ZrCH3]+[CH3B(C6F5)3]−. Adapted with permission from J Am Chem Soc (1991) 113:3623–3625 [26]. Copyright 1991 American Chemical Society

Full size image

Detailed thermodynamic/thermochemical and NMR molecular dynamics studies of the ion-paired catalytic systems show that [LnM-R]+[MeB(C6F5)3]− polymerization properties are sensitive to metal ancillary ligation, R substituent, borane acidity, and solvent polarity. These trends can be understood in terms of the ability of the metallocene ancillary ligands to stabilize the cationic charge, the homolytic M–CH3 bond dissociation enthalpies, and the electron withdrawing power of the borane substituents. These new “single-site” catalytic systems are far better defined and more rationally tunable in terms of catalytic activity and product selectivity, structure, thermodynamics, and mechanism, which enable the discovery of new catalysts, cocatalysts, deeper mechanistic understanding of both the homogeneous and heterogeneous systems, and macromolecules with dramatically different properties, and finally, large-scale industrial processes [31].

At the beginning of the 1990s, new generations of “non-metallocene” [32–35] catalysts based on various phenoxyiminato ligands were reported by Fujita and co-workers (Group 4, Fig. 5) [36–42] and Grubbs and co-workers (Group 10, Fig. 5) [43–46]. In 1995, Brookhart and co-workers reported high activity α-diimine Ni and Pd complexes which are able to co-polymerize ethylene with polar monomers (Fig. 5) [47–53]. In addition, in early 2000, Drent and co-workers [54] at Shell reported that neutral Pd(II) catalysts generated in situ from phosphonium-sulfonate ligands copolymerize ethylene and methyl acrylate to produce linear copolymers (Fig. 5) [55]. Especially in the case of group 4, these new catalysts and cocatalyst/activators have achieved great success in the production of advanced polyolefin materials [4]. The new systems are able to control product molecular weight, polydispersity, comonomer enchainment level and pattern, the tacticity of poly(α-olefins) [17, 56, 57], copolymerization of olefins with polar comonomers (group 10) [54, 55, 58–62], and the catalytic synthesis of block copolymers by processes such as chain shuttling polymerization [63–67].

Fig. 5
figure 5

Examples of non-metallocene olefin polymerization catalysts

Full size image

Heterogeneous olefin polymerization catalysts offer many distinct attractions, such as thermal robustness, high activity, high degrees of coordinative unsaturation, high product molecular masses, and in some cases, high iso-selectivity. However, understanding structure/reactivity/selectivity relationships in these systems is severely complicated by uncertainties in the active site structure(s) and the percentage(s). For all these reasons, grafting well-defined, single-site homogeneous molecular catalysts on surfaces has emerged as a powerful tool to create new catalysts, to characterize surface species, and thus to establish unambiguous structure/activity relationships [68–73]. For example, it has been shown that well-defined zirconium hydrides supported on alumina are highly active catalysts for alkane hydrogenolysis (Fig. 6a) [74] and silica-supported tungsten [75–78] and molybdenum complexes (Fig. 6b) [79–82] are effective catalysts for alkane metathesis [69, 71, 77, 83–85].

Fig. 6
figure 6

Selected examples of supported organometallic catalysts

Full size image

In many cases, the catalytic activities of the supported organometallics far surpass those achievable with the analogous metal complexes in homogeneous system. Also, the structural nature of the molecule-derived adsorbates can be well characterized by a combination of techniques such as solid state NMR, X-ray photoelectron spectroscopy (XPS), vibrational spectroscopy (FT-IR), X-ray absorption spectroscopy (XAS), and density functional theory (DFT) calculations. However despite this progress, many important mechanistic and structural details remain unclear such as (i) the nature of both the catalytically active and catalytically inactive surface species created on chemisorption, including the degree to which the original ligation is preserved; (ii) the number and type of bonds established to the oxide support and the most reactive oxide surface sites for the grafting process; and (iii) the adsorbate structures which maximize catalytic activity. Thus, comparing the key aspects of the adsorbate–oxide surface interaction with the relevant features in well-characterized solution-phase analogues offers an opportunity to better understand the distinctive aspects of the surface organometallic chemistry.

This review focuses primarily on recent research at Northwestern in which organo-groups 4 and 5, and organoactinide complexes are supported on dehydroxylated γ-alumina [86, 87] and sulfated metal oxides [88–90]. The combined application of solid-state NMR spectroscopy, periodic DFT calculations, and metal X-ray absorption spectroscopy indicates formation of organometallic cations having a largely electrostatic Ln-M+···surface− interaction. It will be also seen that these species are highly active catalysts for α-olefin polymerization and arene hydrogenation, with nearly ~100 % of active site in the case of sulfated metal oxide supports.

2 Discussion

2.1 Dehydroxylated γ-Alumina Supports

Dehydroxylated γ-alumina (DA) and partially dehydroxylated γ-alumina (PDA) surfaces show pronounced Lewis acidic character due to coordinatively unsaturated surface Al sites [74, 91–93] that are reminiscent of analogous MAO cocatalysts and structurally well-defined perfluoroarylborane and borate catalyst activators [28–30]. From a coordination chemistry point of view, the accessibility of both Brønsted and Lewis acid sites on alumina offers a unique complexation environment. DA and PDA substrates are prepared by heating γ-alumina (Scheme 1) with consequent elimination of water molecules and formation of Al–O–Al moieties [74, 91–93]. Reaction of organometallic precursors with PDA or DA yields highly active catalysts for both olefin and arene hydrogenation and olefin polymerization (vide infra). Heterolytic M–C bond scission is the primary chemisorption mechanism for highly Lewis acidic surfaces [73]. Here the hydrocarbyl anion is transferred to the surface (Lewis acid sites) to give rise to a cationic organometallic adsorbate (Scheme 1, Pathway A). In marked contrast to these results, chemisorption of organometallic precursors on SiO2, Al2O3, and SiO2–Al2O3 surfaces having appreciable coverage by weakly acidic OH groups, predominantly yields covalently bound, poorly electrophilic species via M–CH3 protonolysis with CH4 evolution (Scheme 1, Pathway B) [72]. While these sites can be characterized by high-resolution solid state NMR and X-ray adsorption fine structure spectroscopy (EXAFS), they display minimal catalytic turnover in the absence of added activators (e.g., MAO or B(C6F5)3), and the fraction of catalytically significant sites is unknown [94, 95].

Scheme 1
scheme 1

Pathway A: synthesis of DA dehydroxylated γ-alumina, PDA partially dehydroxylated γ-alumina, and supported cationic organometallic catalyst. Pathway B: supported neutrally charged μ-oxo organometallic complex on γ-alumina

Full size image

The 13C CPMAS NMR spectrum of Cp2*Th(13CH3)2 chemisorbed on DA [96] shows the transfer of a methide anion from the actinide center to a quadrupolar Al Lewis acid site on the surface. The Al-13CH3 resonance appears at the characteristic upfield position (δ = −5 ppm), whereas the downfield shifted Th-13CH3 signal at δ ~60 ppm indicates formation of a “cation-like” electron-deficient organothorium species as in Scheme 1 Pathway A [97–99]. No evidence of Cp* protonation or redox processes are observed. Interestingly, homogeneous solutions of Cp2*Th(CH3)2 are minimally active for ethylene polymerization and olefin hydrogenation [100–104]. On the other hand, on adsorption DA results in a profound increase in catalytic activity [86, 105]. However, by means of quantitative poisoning experiments using either H2O or CO as probes, it was shown that only ~4 % of the Cp2*Th(CH3)2/DA sites are ethylene polymerization/olefin hydrogenation active [86, 96, 104, 106–110].

In order to test the generality of the supported organoactinide chemistry discussed above, group 4 zirconocenes have also been supported on dehydroxylated γ-alumina [111]. Analogous to the 13C CPMAS NMR experiments performed with Cp*2Th(13CH3)2, Cp2Zr(13CH3)2 supported on DA produces cationic complexes via methide transfer to the surface, while the same precursor supported on PDA yields both μ-oxo (Scheme 1 Pathway B) and cationic species, indicative of the presence of both weak Brønsted and Lewis acid sites. Measurements of evolved CH4 during the supporting reaction find similar yields for both thorium and zirconium species on DA, while lower yields were observed for zirconium species on PDA, likely due to the increased protonolytic stability of Zr-alkyl bonds [112]. The catalytic activity of the group 4 complexes was tested for propylene hydrogenation and ethylene polymerization. For those supported on DA, activity was generally found to be approximately 30 % of that of related organoactinide catalysts. However, in marked contrast to organoactinides, organozirconium complexes on PDA also showed reactivity for propylene hydrogenation. Active site measurements for organozirconium complexes on DA indicated that for Cp2Zr(CH3)2/DA, ~4 % of sites are active, and for Cp*Zr(CH3)3/DA, ~12 % are active, roughly mirroring the previous results for organoactinides on DA.

Periodic density functional calculations has been performed to investigate the structural and catalytic properties of the organozirconium precatalyst Cp2Zr(CH3)2, chemisorbed on a model dehydroxylated γ-alumina (110) surface [113]. Two different prototypical surface oxide environments, namely μ3-O and μ2-O, can interact with the Cp2ZrCH3 + adsorbate (Fig. 7). The interaction of the Cp2ZrCH3 + adsorbate species with the μ2-O sites is far stronger than that with the μ3-O sites due to the greater coordinative unsaturation of the former. Moreover, the interaction with the μ3-O sites is weaker than that in the parent homogeneous Cp2ZrCH3 + H3CB(C6F5) −3 ion pair structure. The energetics of catalytic processes for the chemisorbed Cp2ZrCH3 + sites for ethylene polymerization was examined at both μ2-O and μ3-O environments and compared to the analogous homogeneous catalyst. A Cossee enchainment mechanism proceeds via ethylene π-complex formation and an α-agostic assisted transition state to yield γ- and β-agostic insertion products. The overall kinetics of enchainment are closely correlated with the energetics of π-complex formation, and it is suggested that the differing kinetic behaviors of the surface-bound Cp2ZrR+ species on the various Al2O3 coordination sites and the analogous homogeneous species reflect differences in the olefin π-complex stabilization energies. Specifically, the activation energy for olefin insertion rises in the order μ3-O site <homogeneous Cp2ZrCH3 + H3CB(C6F5) −3 catalyst <μ2-O site.

Fig. 7
figure 7

a Dioxo- and b oxo-bridged zirconocenium coordination on an alumina μ2-O site. c Dioxo- and d oxo-bridged zirconocenium coordination on a μ3-O alumina surface site. Adapted with permission from J Am Chem Soc (2008) 130:16533–16546 [113]. Copyright 2008 American Chemical Society

Full size image

2.2 Sulfated Metal Oxides

Sulfated metal oxides (SMOs) have received significant attention in the past as environmentally friendly alternatives for superacid catalyzed reactions, such as the skeletal isomerization of alkanes [114] and the production of biodiesel [115]. To this end, a variety of oxides have been sulfated, such as zirconia (ZrS, H0 = −16), alumina (AlS, H0 = −14.6), stannia (SnS, H0 = −18), hematite (FeS, H0 = −13), and titania (TiS, H0 = −14.6), [116] typically by exposure to a sulfation reagent, the most common of which is sulfuric acid, and subsequent calcination. As supports for single site catalysts, sulfated metal oxides are particularly well-suited due to the presence of a generally large concentration of strong Brønsted and Lewis acid sites, which yield highly active ‘cationic’ metal adsorbate centers (vide supra). Similar to the studies discussed previously with PDA and DA, the use of model adsorbates (Fig. 8) along with 13C CPMAS NMR techniques have been valuable in elucidating the surface character of these SMOs-organometallic complexes/catalysts (Fig. 9).

Fig. 8
figure 8

Molecular precursors for chemisorption on SMOs to produce anchored catalysts

Full size image
Fig. 9
figure 9

Proposed structures of zirconocenium complex supported on sulfated metal oxides

Full size image

Sulfated zirconia (ZrS) was one of the first SMOs investigated by our group [117–119]. Initial studies of the surface characteristics of ZrS were performed by the chemisorption of 13C labeled Cp*2Th(13CH3)3 (1*) [118]. The 13C CPMAS spectrum of 1*/ZrS compared to that of 1*/DA revealed important differences in the chemisorption pathways. Specifically, the peak resulting from methide transfer is substantially diminished, suggestive of only a small density of Lewis sites on the ZrS surface. The shift of the Th-13CH3 + resonance to lower field also denotes a more electrophilic ‘cationic’ metal center (Scheme 1, Pathway A), resulting from the greater surface acidity and suggesting a more active catalytic site (vide infra). The minimal presence of hydrocarbyl transfer to the surface and increased electrophilicity of the adsorbate metal center are also confirmed with adsorbates CpTi(13CH3)2/ZrS (2*/ZrS), Cp*Zr(13CH3)2/ZrS (3*/ZrS), Cp2Zr13CH3/ZrS (4*/ZrS), and Zr(tBu)3/ZrS. 1*/ZrS is also observed to have the most downfield shifted resonance for Zr-13CH3 + when compared with homogeneous analogues, indicating the highly electrophilic character of the metal center (Scheme 2).

Scheme 2
scheme 2

Synthesis of monometallic CGC-Zr catalysts supported on sulfated metal oxides

Full size image

Other SMOs, including AlS [117, 120–123], SnS [117], FeS [117], and TiS [117] were analyzed in a similar way and found to yield analogous surface species. In comparing a given series, it is noted that the magnitude of the shift in resonances between the precursor and the adsorbate qualitatively correlates with the electronic unsaturation of the electrophilic species (Table 1). For 4* chemisorbed on AlS, SnS, and TiS, mainly ‘cationic’ species (A), with small amounts of μ-oxo (B) present in SnS, are produced via protonolytic chemisorption of 4*, with methide transfer remaining below the detection limits, implying the presence of primarily strong Brønsted sites. For 4*/ZrS, however, μ-oxo species are formed, with protonolysis by weak Brønsted acid sites dominating, and a small, downfield shifted resonance for ‘cationic’ species also observed. Nanoparticles of AlS (n-AlS) [121] yield similar structural results to bulk AlS, except for a larger presence of methide transfer, indicative of a higher density of Lewis acid sites. n-AlS also enables increased loadings of Zr due to minimized internal surface area.

Table 1 13C CPMAS NMR parameters for model adsorbates on various SMOs
Full size table

Adsorption of organo-group 4 complexes on AlS was also studied by extended X-ray absorption fine structure (EXAFS) and periodic density function theory (DFT) computation in order to gain a more realistic understanding of the structural characteristics of the adsorbates on SMOs. [120]. Zr K-edge EXAFS data were collected under strictly anhydrous/anaerobic conditions and, due to the structural complexity of the supported moieties, difference spectra (before and after chemisorption) were utilized, since changes in the metal ligation are of primary interest. For 4*/AlS, the Zr–Osupport bond length is found to be 2.37 ± 0.02 Å, which is significantly elongated from typical literature covalent Zr(IV)-OR bond lengths, 2.000 ± 0.002 Å [120], supporting the 13C CPMAS NMR characterization of loose ion pairing between the ‘cationic’ metal adsorbate and the negative charged-delocalized surface. This is further supported by DFT calculations, which, based on two potential types of sulfate site (Fig. 10), find mean Zr–Osupport distances of 2.22 and 2.42 Å. As for the interaction between the adsorbate and the surface, the EXAFS difference spectra reveal a CN = 2.1, such that the negative charge on the surface is delocalized between two proximate oxygen atoms that both bind to the metal center. After exposure to benzene under hydrogenation conditions, DFT calculations predict insertion of the benzene between the Zr+···O¯ bond, which finds close analogy to the ion pair weakening and anion displacement in activated homogeneous early transition metal polymerization catalysts upon olefin approach/activation.

Fig. 10
figure 10

Energy-minimized computed chemisorbed catalyst structures for: a Cp*Zr(CH3) +2 coordination to the S=O groups of the sulfated alumina surface at SA sites. b Cp*Zr(CH3) +2 coordination to the S=O groups of the sulfated alumina surface at SB sites. Distances in Å. Avg. = average distance. Zr = purple, C = olive, H = blue, O = red, S = orange, Al = yellow-green. Adapted with permission from Proc Nat Acad Sci USA (2013) 110:413–418 [120]

Full size image

In the interest of expanding the variety of supported organometallic moieties, constrained geometry catalysts (CGC, Fig. 11) [124] were also chemisorbed onto AlS [122]. CGCs are valued for their increased thermal stability and favorable incorporation of comonomers during polymerization, resulting in long chain branching polymers, due to more open coordination spheres. The formation of cationic surface species was further verified by a distinct color change, from colorless to yellow (for organozirconium complexes) or colorless to orange (for organotitanium complexes), which is analogous to the color change observed during homogenous activation with organo-Lewis acids such as B(C6F5)3. For the bimetallic 9/AlS, both the number of metal centers activated (one or both) and the type (protonolytic or methide transfer) of chemisorption permits the exact structural characterization of this adsorbate (Fig. 12). The integration of the 9/AlS 13C CPMAS NMR resonances suggests that the majority of adsorbate species have both metal centers activated, while the integration of the Zr-13Me:Al-Me resonances in the 13C CPMAS NMR is found to be 1:0.6, whereas a ratio of ~ 3:1 is expected for activation of a single metal center. Thus, both metals in the bimetallic precursor are likely activated, approximately 40 % by metal-hydrocarbyl protonolytic chemisorption and 60 % by methide transfer to the surface.

Fig. 11
figure 11

Mononuclear and binuclear CGC-zirconium catalysts

Full size image
Fig. 12
figure 12

Potential binding modes of 9/AlS or 10/AlS. As determined by the type and concentration of acid sites: (i) Single metal center protonolytically chemisorbed on a strong Brønsted acid site; (ii) Single metal center chemisorbed via methide transfer on a Lewis acid site; (iii) Both metal centers protonolytically chemisorbed on proximate strong Brønsted centers; (iv) Both metal centers chemisorbed via methide transfer on proximate Lewis acid sites; (v) One metal center protonolytically chemisorbed, the other chemisorbed by methide transfer

Full size image

Following structural and chemical characterization of the chemisorption pathways, olefin homo- and copolymerizations underscore the catalytic utility of SMOs as activators and weakly coordinating anions. For the catalysts discussed here, the principal product of ethylene polymerization is HDPE with minimal branching and ultrahigh molecular weight, as observed in previous studies of organozirconium hydrocarbyls supported on alumina (vide supra). For ethylene homopolymerizations using ZrS and AlS as supports (118–120; 124), a number of adsorbates with varying electronic and coordinative saturation were assayed (Table 2). In general, polymerization activity is found to increase with increasing coordinative unsaturation, such that Zr(CH2Ph)4 > Cp*Zr(CH3)3 for every SMO. As for support effects, activity follows the trend ZrS ≥ AlS > SnS > FeS > TiS. Importantly, AlS is found to have an ethylene homopolymerization activity almost10x that of Cp*Zr(CH3)3/DA (1.2 × 105 g PE/mol Zr h), proving that chemisorption on SMOs yields more active ‘cationic’ centers, reflecting both the increased number of active sites and their ‘cationic’ nature (vide infra). Nanoparticle-AlS is also found to be an effective activator and high surface area support for the rapid production of high molecular weight PE and its performance is found to be solvent dependent, heptane > toluene, likely due to the coordinative inhibition of arenes at the electrophilic active sites [126]. Although solvent dependence was only tested on n-AlS supported catalysts, all SMO adsorbates are expected to exhibit similar behavior due to their mainly ‘cationic’ character.

Table 2 Ethylene homopolymerization data for SMO supported zirconium hydrocarbyls
Full size table

One of the principal attractions of SMOs is the high percentage of zirconium adsorbate sites that are catalytically significant. In situ kinetic poisoning experiments performed on the above catalysts reveal the high number of active sites, with >60 % for all SMOs except FeS attainable, compared to ~12 % for Cp*Zr(CH3)3/DA and <6 % for Zr(CH2CMe2Ph)4/PDA [112]. While there is no direct correlation between acidity and percentage active sites, for the SMOs discussed here, it appears that materials with H0 < −14 have a high percentage of active sites. It has been concluded that the exceptional catalytic activity of these zirconium hydrocarbyl adsorbate catalysts reflects an interplay of both the high percentage of active sites and the electrophilicity of such sites, such that both are required for an active catalyst. For example, TiS as a support has 63 ± 9 % active sites, but 13C CPMAS NMR reveals that most of these are of the active μ-oxo type (vide supra), thus the catalyst performs relatively poorly. The attraction of using SMOs as supports/activators thus lies in the fact that they provide both highly active electrophilic single catalytic sites, and due to their high acidity, nearly 100 % of these sites are catalytically active.

The CGC catalysts supported on AlS were also evaluated for ethylene homopolymerization, along with ethylene/1-hexene copolymerization [122]. Ethylene homopolymerizations are found to yield high molecular weight PE as determined by DSC, and to be highly crystalline with low branch densities, but in all cases, the corresponding supported catalysts have lower catalytic activity than the homogenous analogs, which is attributed to steric repulsion of the support and the adsorbate ligands. Analogously to homogeneous systems [127], the bimetallic Ti2/AlS is more active than bimetallic Zr2/AlS. Concerning ethylene/1-hexene copolymerizations, the supported catalysts are again less active than their homogenous equivalents. Minimal comonomer enchainment is observed (Tm depressed only 3−5 °C from the homopolymer) and it is likely that the steric bulk of the bimetallic Zr2 and Ti2 limits 1-hexene approach, thereby frustrating the bimetallic cooperativity observed in homogenous systems.

3 Conclusion and Prospectus

This Review analyzes the effects on olefin polymerization by single-site catalysts supporting them on various oxides and sulfated oxides. The well-established synthetic routes to achieve single-site heterogeneous catalysts and the diverse techniques (SS NMR, EXAFS, DFT calculations) for their characterization are described. Solid state NMR and EXAFS studies show that alkyl migration from the organometallic precursor to the Lewis acidic sites constitutes the principal adsorption pathway on dehydroxylated alumina (DA), whereas the chemisorption of the same organometallic molecules on highly Brønsted “super-acidic” sulfated metal oxides yields, via Zr-CH3 protonolysis with methane evolution, highly electrophilic adsorbate species. The resulting “cation-like” species, reminiscent of the homogeneous analogues, are highly active heterogeneous catalysts for olefin polymerization and hydrogenation, with generally high percentages of catalytically significant sites. In many cases, extraordinarily high percentages (~100 %) of catalytically significant sites can be attained, permitting accurate structure determination and isolation/characterization of reaction intermediates.