Metallocene-catalyzed alkene polymerization and the observation of Zr-allyls

Abstract

Single-site polymerization catalysts enable exquisite control over alkene polymerization reactions to produce new materials with unique properties. Knowledge of catalyst speciation and fundamental kinetics are essential for full mechanistic understanding of zirconocene-catalyzed alkene polymerization. Currently the effect of activators on fundamental polymerization steps is not understood. Progress in understanding activator effects requires determination of fundamental kinetics for zirconocene catalysts with noncoordinating anions such as [B(C₆F₅)₄]⁻. Kinetic NMR studies at low temperature demonstrate a very fast propagation rate for 1-hexene polymerization catalyzed by [(SBI)Zr(CH₂SiMe₃)][B(C₆F₅)₄] [where SBI is rac-Me₂Si(indenyl)₂] with complete consumption of 1-hexene before the first NMR spectrum. Surprisingly, the first NMR spectrum reveals, aside from uninitiated catalyst, Zr-allyls as the sole catalyst-containing species. These Zr-allyls, which exist in two diastereomeric forms, have been characterized by physical and chemical methods. The mechanism of Zr-allyl formation was probed with a trapping experiment, leading us to favor a mechanism in which Zr-polymeryl undergoes β-H transfer to metal without dissociation of coordinated alkene followed by σ-bond metathesis to form H₂ and Zr-allyl. Zr-allyl species undergo slow reactions with alkene but react rapidly with H₂ to form hydrogenation products.

Over the past 20 years new catalyst technologies have reinvigorated polyolefin chemistry by rapidly expanding new polyolefin materials and technology. The so-called “single-site” catalysts based on homogeneous transition metal complexes form the leading edge of new catalyst technologies (1). Benefits of single-site polymerization include access to a broader range of catalyst structures through rational syntheses, improved control of polymer molecular weight distributions, and compatibility with flexible solution-based polymerization processes. Perhaps more significantly, single-site catalysts enable the discovery and commercial production of novel polyolefin materials, such as “blocky” polymers, with unprecedented properties (2–4). From an academic viewpoint, homogeneous polymerization catalysts offer marvelous opportunities for probing reaction mechanisms and rational design that cannot be achieved with heterogeneous catalysts.

Catalysts based on group 4 metals (see examples in Fig. 1) exhibit the most attractive combination of activity, selectivity, and generality to a wide variety of α-olefins. Many of the initial discoveries of well defined single-site catalysts centered on metallocenes; however, the so-called nonmetallocenes account for many recent catalyst discoveries (5–7). The most active group 4 catalysts pair a cationic metal center with a noncoordinating anion, such as [MeB(C₆F₅)₃]⁻, [B(C₆F₅)₄]⁻, and the products of methide abstraction by methylalumoxane oligomers. Although methylalumoxane is widely used in industry, borane and trityl salt activators form isolable counteranions better suited for mechanistic studies.

Fig. 1.

Common group 4 polymerization catalyst precursors and activators. MAO, methylalumoxane; CGC, constrained geometry catalyst.

As with all polymerization reactions, the phases of chain growth include initiation, propagation, and termination (Fig. 2). The corresponding elementary rate laws and kinetic constants completely describe the catalytic kinetics and the distribution of polymer products (8). We have shown that a combination of active site counting methods, quenched flow reaction kinetics, and polymer end group analysis applied to 1-hexene polymerization catalyzed by B(C₆F₅)₃-activated catalyst, [(EBI)Zr(Me)][MeB(C₆F₅)₃] [where EBI is rac-C₂H₄(indenyl)₂], lead to rigorous characterization of the kinetic rate laws (8–11).

Fig. 2.

Elementary polymerization steps.

Direct observation of the propagating catalyst by low-temperature NMR spectroscopy has provided unique mechanistic insights. Polymerization of 1-hexene catalyzed by [(EBI)ZrMe][MeB(C₆F₅)₃] at −40°C enables direct characterization of growing Zr-polymeryl species, monitoring of monomer consumption kinetics, determination of instantaneous active site counts, and measurement of termination rates. Finer details of the polymerization mechanism, such as demonstration that chain growth occurs in a continuous rather than intermittent manner (8, 12), and the mechanism of chain-end epimerization of Zr-polypropenyl (10) have been resolved through NMR studies.

Rare polymerization events can profoundly influence catalyst activity by allowing catalyst to pool in less active, “dormant” forms and also affect polymer properties. Rare events include 2,1-insertion to form secondary metal alkyls, chain-end epimerization, and the formation of metal allyl complexes. It has been suggested that secondary metal alkyls formed by 2,1-insertions are less reactive than primary metal alkyls and accumulate as dormant catalytic sites not actively involved in polymerization (13). However, we recently demonstrated unexpectedly high reactivity of model secondary alkyls at low temperature with ethene, propene, and H₂ (14), but the general applicability of these results to alkene polymerization is uncertain (15) and calls for further investigation.

Conversion of actively propagating catalyst into dormant metal allyl species has been proposed as a consequential rare event for polymerization catalysts (16–19). Metal allyl complexes of early transition metals and lanthanides have been known for quite some time (20, 21), but their implication in bulk polymerization reactions was first proposed about a decade ago by a group from Union Carbide (18). Although the formation of cationic Zr-allyls was observed by mass spectrometry in the gas phase by Richardson and coworkers (16, 17), much of the current speculation on the presence of Zr-allyl species during polymerization relies on the observation of H₂ during propene polymerization and the presence of internal vinylidenes in the resulting polymer (18), theoretical studies (22–25), and reactivity studies with model Zr-allyl complexes (22). A mechanism shown in Fig. 3that accounts for the generation of both internal vinylidenes and H₂ features a Zr-allyl species formed from β-H elimination of Zr-polymeryl (i) followed by σ-bond metathesis of an allylic CH bond of the Zr-alkene species (ii) and dissociation of H₂ (iii). Insertion of alkene into this Zr-allyl complex adds internal vinylidenes to the polymer (iv).

Fig. 3.

Proposed formation of dormant Zr-allyl catalyst sites and internal vinylidenes. (i) β-H elimination. (ii) σ-Bond metathesis. (iii) H₂ dissociation. (iv) α-Alkene insertion.

Formation of Zr-allyl species may impact polymerization reactions in several ways. Resconi has suggested that Zr-allyl complexes may be important intermediates in the mechanisms of chain epimerization (19), but studies by Yoder and Bercaw (26) and our own group (10) demonstrate that alternate pathways can accommodate epimerization observations. It is more commonly accepted that Zr-allyls may form a dormant catalytic site and reduce polymerization activity. Computational studies by Ziegler and coworkers (23, 25) demonstrate that M-allyl (M is Ti, Zr) formation by the pathway shown in Fig. 3 is possible for both a Ti-constrained geometry catalyst and a zirconocene catalyst (24). Their calculations also show that alkene insertion into the Zr-allyl complex is slower than normal propagation (24). Prosenc and colleagues (22) have also shown computationally and experimentally that insertion into a model Zr-allyl complex is at least one order of magnitude slower than normal propagation. Recently, Collins and coworkers (27) identified a Zr-allyl complex as a byproduct of propene polymerization catalyzed by B(C₆F₅)₃-activated [Me₂C(Cp)(indenyl)ZrMe₂]. Observation of such Zr-allyl species is an important step in assessing the importance of metal allyl formation in polymerization reactions. Ultimately, the kinetics of Zr-allyl formation (steps i–iii) and the rate of alkene insertion into the Zr-allyl complex (step iv) determine the significance of these species in polymerization.

Herein we report recent studies involving the effect of counterions on propagation rates, resulting polymer properties, and catalyst speciation. In particular, we show that under similar polymerization conditions a catalyst with the noncoordinating anion [B(C₆F₅)₄]⁻ forms two Zr-allyl species that are not observed with the more coordinating [MeB(C₆F₅)₃]⁻. We also report the characterization of these Zr-allyl complexes by chemical and NMR spectroscopic methods. Additionally, we discuss possible mechanisms for this Zr-allyl formation and demonstrate the dormant natures of these Zr-allyl species.

Results and Discussion

Comparison of B(C₆F₅)₃ and [CPh₃][B(C₆F₅)₄] as Activators for Zr-Catalyzed Polymerization.

Our initial investigations into activator effects focused on comparing B(C₆F₅)₃ and [CPh₃][B(C₆F₅)₄] as cocatalysts with (EBI)ZrMe₃. Using quenched-flow methods we find that [(EBI)ZrMe][B(C₆F₅)₄] is ≈20 times more active for polymerization of 1-hexene than [(EBI)ZrMe][MeB(C₆F₅)₃] (11). However, we consider the rate constant determined for the trityl borate-activated system (125 M⁻¹·s⁻¹ at 20°C) to be a low estimate because of overcounting of active catalyst sites. Trityl salts are known to form μ-Me zirconocene dimers upon incomplete activation (28, 29). We have observed dimers of the form [(EBI)Zr(Me)(μ-Me)(polyhexenyl)Zr(EBI)]⁺ due to trapping of active catalyst sites by (EBI)ZrMe₂ as shown in Fig. 4. Quenching with MeOD converts any metal alkyl into a d-labeled alkane (30). Because the μ-Me dimers are less active for alkene insertion than monomeric species but react similarly with quenching reagents, the concentration of high-activity sites is overestimated.

Fig. 4.

MeOD quenching scheme cannot distinguish between active sites and μ-Me dimers.

Preliminary attempts to directly observe active sites by NMR spectroscopy during polymerization at low temperature were not successful for the [(EBI)ZrMe][B(C₆F₅)₄] catalyst (31). In each attempt polymerizations occurred too rapidly to be monitored by ¹H NMR. After complete consumption of the 1-hexene, no propagating species could be identified in the ¹H NMR spectrum or by ¹³C NMR spectroscopy using 1-¹³C-1-hexene. These early attempts at direct observation did demonstrate extremely fast propagation and very low levels of catalyst initiation.

Studying cationic zirconocene catalysts with noncoordinating anions is complicated by difficulties in preparing isolable catalyst ion pairs. However, a recently reported synthesis of the asymmetric (C₁ point group) zirconocene (SBI)Zr(Me)(CH₂SiMe₃) [where SBI is rac-Me₂Si(indenyl)₂] (1) enables clean activation with either B(C₆F₅)₃ or [CPh₃][B(C₆F₅)₄] to form the ion pairs [(SBI)Zr(CH₂SiMe₃)][MeB(C₆F₅)₃] (2a) or [(SBI)Zr(CH₂SiMe₃)][B(C₆F₅)₄] (2b), respectively (32). For the [CPh₃][B(C₆F₅)₄] activator, cosolvents such as ortho-difluorobenzene or chlorobenezene-d₅ are required to prevent phase separation from toluene-d₈. Bochmann and coworkers (32) used 2a and 2b to polymerize propene and 1-hexene but reported that 2b-catalyzed polymerizations of 1-hexene curiously were nearly 10 times slower than our previous estimates (11). Such results led them to propose different transition states for 1-hexene and propene polymerization when a noncoordinating anion was used.

We previously have shown that the SBI-ligated precatalysts with a “sticky” anion {[(SBI)ZrMe][MeB(C₆F₅)₃] and 2a} yield propagation rates for 1-hexene and propene polymerization that are similar to the EBI-ligated catalysts {[(EBI)ZrMe][MeB(C₆F₅)₃]} with the same anion (9). Catalyst 2b with a noncoordinating anion is much faster, effecting complete consumption of 1-hexene before the first ¹H NMR spectrum can be acquired at −40°C. Such rapid reaction, consistent with a propagation rate constant >100 M⁻¹·s⁻¹, conflicts with the report by Bochmann and coworkers (32) but agrees with our previous estimate (11).

Although polymerization with 2b is too rapid for the kinetics to be determined by direct observation, several distinctive characteristics are apparent from the NMR data. We determined the extent of apparent initiation by comparison of the ¹H NMR integrations of the Zr-CH₂SiMe₃ signal versus the POL-SiMe₃ signal. Typically, only approximately one-fourth of the starting ion pair underwent initiation. This finding suggests that initiation is much slower relative to propagation for 2b than for 2a, for which >95% of the catalyst initiates. However, data collected at a variety of initial conditions give widely varying ratios k_p/k_i with the 2b catalyst using the formulas developed by Gibson and colleagues (33). Given the rapid rate of this polymerization, it is unlikely that mixing in the NMR tube is sufficiently fast, and the apparent slow initiation of 2b is an artifact of mass transport limitations. Full initiation of 2a is in contrast to ≈60% initiation for both [(EBI)ZrMe][MeB(C₆F₅)₃] and [(SBI)ZrMe][MeB(C₆F₅)₃] (9, 34). Thus, Zr-CH₂SiMe₃ is a better model for the propagating Zr-polymeryl than Zr-Me, presumably because of weaker pairing of the anion with Zr-CH₂SiMe₃.

Interestingly, regioerrors (i.e., 2,1-misinsertions) of the alkene are enchained in the polymer prepared with catalyst 2b, whereas catalyst 2a exhibits no enchained regioerrors. Two doublets at 33.6 and 31.2 ppm with 35.5-Hz splitting in the ¹³C NMR spectrum of polymerized 1-¹³C-1-hexene demonstrate regioerror enchainment. Unambiguous confirmation of carbon–carbon coupling between the two doublets is provided by a ¹³C INADEQUATE NMR experiment. Integration of the ¹³C NMR spectrum shows that approximately one regioerror occurs every 400 normal 1,2-insertions. In further contrast with catalyst 2a, no alkene end groups are observed by ¹H NMR, even after warming to room temperature.

Activator Effects on Catalyst Speciation.

For the [(EBI)ZrMe][MeB(C₆F₅)₃]-catalyzed polymerization of 1-¹³C-hexene, a ¹³C triplet at 88.5 ppm with ¹J_CH = 117 Hz was identified as the α-carbon of the propagating Zr-polyhexenyl species (9). Similar experiments with SBI catalyst 2a gave a similar signal at 94.0 ppm with ¹J_CH = 113 Hz. However, polymerization of 1-¹³C-1-hexene with 2b does not yield a single ¹³C NMR signal as expected for a single Zr-alkyl intermediate. Instead, two ¹³C NMR signals are observed at 84 and 80 ppm, indicating the presence of two species (3a and 3b). The results of selective ¹H-¹³C heteronuclear single quantum coherence (HSQC) experiments with excitation of each ¹³C signals at δ80 and δ84 are shown in Fig. 5. Clearly observed are the diastereotopic geminal pairs of protons attached to each ¹³C label. Upon closer examination, the ¹J_CH coupling constants are observed to be 156 and 151 Hz for the 84-ppm signal and 163 and 152 Hz for the 80-ppm signal, much larger than expected for the sp³ CH bond of a Zr-alkyl species. The corresponding ¹H chemical shifts are too far upfield for a metal-coordinated alkene (35, 36), but they are reasonable for Zr-allyl species^† formed by the path presented in Fig. 6.

Fig. 5.

Selective ¹H-¹³C HSQC NMR experiments produce a ¹H NMR spectrum of protons coupled through ¹J_CH for each ¹³C signal of interest.

Fig. 6.

A Zr-allyl species is observed after polymerization of 1-hexene at low temperature.

The synthesis of Zr-allyl compounds by reaction of isobutylene with a cationic zirconocene has been reported (27). Addition of isobutylene to a solution of 2b at room temperature yields [(SBI)Zr(methallyl)]⁺ (4) and a single equivalent of SiMe₄ in addition to isobutylene oligomers. The ¹H NMR spectrum has four allylic signals at 4.00, 1.92, 0.84, and −0.17 ppm that display chemical exchange in 1D exchange spectroscopy experiments. Similar dynamic behavior due to η³-η¹-η³ allyl rearrangement has been observed for other cationic zirconocene allyls with the coordinating anion [MeB(C₅F₆)₃]⁻ (22). Other NMR experiments including ¹H-¹³C HSQC, 1D total correlation spectroscopy, and 1D NOESY allow for the full characterization of 4 as reported in Figs. 11–14 and Supporting Materials and Methods, which are published as supporting information on the PNAS web site. Comparison with the Zr-allyl species 3a and 3b generated during polymerizations confirms that the proton and carbon chemical shifts of the species after 1-hexene polymerization are consistent with the assignment of Zr-allyl species.

The chemical reactivity of the putative Zr-allyl species with MeOD provides further support for allyl formulation. Upon reaction with MeOD, the Zr-allyl species disappears, and a terminated vinylidene endgroup is observed as a singlet at 111 ppm in the ¹³C NMR spectrum. A very small 1:1:1 triplet is also observed at 21 ppm, which is consistent with the chemical shift of a d-labeled internal vinylene. All data are consistent with the two expected products of quenching a Zr-allyl species by MeOD, as shown in Fig. 7. Peak integrations reveal that the vinylidene species is formed nearly quantitatively, whereas only a trace amount of the internal vinylene is observed.

Fig. 7.

Addition of MeOD to Zr-allyl species forms vinylidene and vinylene polymer endgroups.

The presence of two Zr-allyl diastereomers is not surprising given that the Zr-allyl species contain three stereocenters, one at the metal and two at allylic carbons. A total of eight stereoisomers are possible: four diasteromers each comprising one pair of enantiomers. The four distinct diastereomers are shown in Fig. 8. We hypothesize that only two diastereomers are formed because of steric constraints. Potential interactions between the allylic propyl group and the indenyl ligand system lead us to favor the R,R and S,R diastereomers over the remaining two diastereomers.

Fig. 8.

Four possible diastereomers for Zr-polymeryl allyl species. Each diastereomer has a related enantiomer. The —SiMe₂— bridge is left out for clarity.

Mechanism of Zr-Allyl Formation.

Two related, but different, mechanisms for the formation of the observed Zr-allyl species are shown in Fig. 9. Mechanism A is the “H₂-generating path” and begins with β-H elimination followed by σ-bond metathesis with an allylic H to form a Zr(H₂)(allyl) species. Dissociation of H₂ then yields the observed Zr-allyl complexes. This mechanism was described above to explain the observation of H₂ and internal vinylidene unsaturation in propene polymerization reactions. A second possible mechanism, the “alkane-generating path,” also begins with β-H elimination but is followed by alkene dissociation. This terminated vinylidene endgroup then coordinates to another Zr-alkyl in solution. In this case, σ-bond metathesis of the allylic H and the Zr-alkyl form the observed Zr-allyl species and alkane. The two mechanisms are expected to impact polymerization reactions in different ways. Mechanism A, the H₂-generating path, should become more important as the alkene concentrations become low and alkene insertion is not as competitive with intramolecular σ-bond metathesis to form H₂. The released H₂ can act as a chain transfer agent, lowering the average molecular weight of the polymer. Pathway B is favored by high accumulation of vinylidene-terminated polymers produced by β-H transfer to metal.

Fig. 9.

Possible mechanistic pathways for Zr-allyl formation.

In principle, the H₂-generating (mechanism A) and alkane-generating (mechanism B) paths could be distinguished by examination of the products: H₂ is an expected byproduct of mechanism A and SiMe₄, and other saturated alkanes should be observed for mechanism B. For our experiments, the predominate Zr-alkyl species in solution is [(SBI)ZrCH₂SiMe₃]⁺ because of the low levels of initiation, so a major byproduct of mechanism B should be SiMe₄. However, control experiments determine that H₂ reacts rapidly with [(SBI)ZrCH₂SiMe₃]⁺ to form SiMe₄. Thus, both pathways likely will produce significant amounts of SiMe₄.

Another distinction between mechanisms A and B is that mechanism B requires dissociation of vinylidene-terminated polymer from the catalyst but mechanism A does not. We find from trapping experiments that a significant amount of alkene does not dissociate after the Zr-polymeryl undergoes β-H elimination. Given the relative dormancy of the Zr-allyl species (vide infra), the low extent of initiation suggests that Zr-allyl formation occurs after 1-hexene is consumed.

Trapping experiments were performed by using 2-methyl-1-pentene as the trap. Control experiments demonstrated that a single equivalent of 2-methyl-1-pentene reacts with 2b but that it does not polymerize. A mixture containing 100 equivalents of 1-¹³C-1-hexene and 10 equivalents of 2-methyl-1-pentene was added to catalyst 2b at −40°C. Under these conditions we expect 1-hexene to be the reactive alkene until it is nearly all consumed. Upon exhaustion of 1-hexene, 2-methyl-1-pentene is present in ≈10-fold excess to Zr. If a vinylidene endgroup were to dissociate after β-H elimination, the 2-methyl-1-pentene should trap the resulting Zr-H species, and one would observe free vinylidene in solution. With ¹³C-labeled 1-hexene, even small amounts of free vinylidene are detectable in the ¹³C NMR spectrum, but none is observed. This result suggests that metal-coordinated vinylidenes formed by β-H transfer to metal do not dissociate from the catalyst. We do observe a 3-fold decrease in the amount of 3a and 3b formed in the presence of 2-methyl-1-pentene, but this same decrease is seen when 2-methyl-1-pentene is added to preformed 3a and 3b. Such results imply that Zr-allyls 3a and 3b may undergo direct σ-bond metathesis with 2-methyl-1-pentene, forming a new metal allyl. Additionally, we have observed that 1-hexene reacts preferentially with Zr-alkyl species over Zr-allyl species (vide infra). If Zr-allyls 3a and 3b formed in the presence of excess 1-hexene, we expect that the alkene would react with uninitiated catalyst 2b and not species 3a and 3b, leading to a relatively high level of catalyst initiation. The observed low levels of initiation suggest that 1-hexene polymerization occurs rapidly and that 3a and 3b are formed after complete 1-hexene consumption. These data are most consistent with mechanism A but do not definitively rule out mechanism B.

Observed Zr-Allyl Species Is Dormant Toward Alkene Insertion.

The reactivity of Zr-allyl species is important to understanding their importance in polymerization. To probe the reactivity of the Zr-allyl species observed during our studies, we successively added isotopomers of 1-hexene to 2 and monitored the reaction by ¹³C NMR (Fig. 10). Initially, 100 equivalents of 1-¹³C-1-hexene were added to 2 to generate the allyl species 3a and 3b. Upon addition of 50 equivalents of unlabeled 1-hexene, all alkene was consumed quickly, but the ¹³C NMR signals for 3a and 3b did not fade significantly. The ¹H NMR spectrum shows that the alkene reacted with uninitiated catalyst 2. A second addition of unlabeled alkene yielded the same results, whereas a last addition of 1-¹³C-1-hexene increased the intensity of the ¹³C NMR signals for 3a and 3b. These data demonstrate that 1-hexene reacts preferentially with the uninitiated ion pair 2 over the Zr-allyl complexes 3a and 3b. This result implies that the formation of metal allyl species deactivates homogeneous polymerization catalysts and is consistent with previous reports (22, 24).

Fig. 10.

Multiple additions of alkene demonstrate a preference to insert into 2b instead of the Zr-allyl species.

Addition of 1-hexene to preformed Zr-allyl demonstrates slow insertion of alkene into metal allyl species. Catalyst 2b was completely reacted with 2-methyl-1-pentene at −40°C to form a Zr-allyl. Subsequent addition of 1-hexene showed no polymerization by ¹H NMR over several hours, clearly demonstrating that alkene insertion into Zr-allyls is much slower than normal propagation.

Conclusions

The data presented herein explore differences between zirconocene catalysts with noncoordinating anions and weakly coordinating anions. Zr-allyls may form under catalytic conditions when noncoordinating anions are used, and these allyls are relatively slow to react further (i.e., are dormant sites). As a result, catalyst deactivation due to Zr-allyl formation may significantly affect polymerization activities, although we cannot quantify the frequency of allyl formation expected under more common steady-state catalytic conditions. We also note the presence of enchained regioerrors with these catalysts only in the presence of a noncoordinating anion.

Direct observation of 1-hexene propagation and Zr-allyl formation is impossible at this time because of extremely fast reaction rates. We are unable to determine unambiguously whether all of the 1-hexene is consumed before Zr-allyl formation. Experiments with an alkene trap strongly suggest that Zr-allyl formation occurs without alkene dissociation, but a full mechanistic understanding of this process is also limited by current kinetic methods. These are important but challenging problems, and new instrumentation and kinetic methods are needed to study these very fast reactions.

Materials and Methods

Synthesis of all catalyst precursors followed previously reported procedures. Full experimental procedures and spectroscopic data are reported in Supporting Materials and Methods.

1-Hexene Polymerization by [(SBI)Zr(CH₂SiMe₃)][B(C₆F₅)₄].

Zirconocene 1 and trityl tetrakis(pentafluorophenyl)borate were dissolved in a solvent mixture of 0.70 ml of toluene-d₈ and 0.20 ml of chlorobenzene-d₅. The resulting orange solution was placed in a 5-mm NMR tube and capped with a septum. The tube was cooled in a −40°C acetone/dry ice bath, and 1-hexene was added to the sample by syringe immediately yielding a dark brown solution. The tube was quickly placed in the NMR magnet for data acquisition.

Synthesis of [(SBI)Zr(methallyl)][B(C₆F₅)₄].

A model cationic Zr-allyl species was prepared by a method similar to a previous report (27). Zirconocene 1 and trityl tetrakis(pentafluorophenyl)borate were dissolved in a mixture of 0.70 ml of toluene-d₈ and 0.20 ml of chlorobenzene-d₅ and placed in a 5-mm NMR tube. At room temperature, isobutylene was added by syringe, and the orange solution slowly darkened. ¹H, 1D total correlated spectroscopy, 1D NOESY, and ¹H-¹³C HSQC NMR experiments were used to characterize the product. The 1D NOESY experiment demonstrates chemical exchange of the allylic protons.

Abbreviations

EBI

rac-C₂H₄(indenyl)₂

SBI

rac-Me₂Si(indenyl)₂

HSQC

heteronuclear single quantum coherence.