Mechanistic Studies

81

82

Figure 2.37 Variable-temperature ¹H NMR spectra of Hf-1/MAO in toluene-d8 (4.80-5.85 ppm, 25-80 ºC).

The ¹H NMR spectrum of Hf-1/MAO/NBE shows a weak but detectable signal at 25 °C at δ = 8.49 ppm (d, J = 8 Hz) and 8.58 (d, J = 8 Hz) again assignable to Ar-H in the “open” methylated-Hf (Figure 2.38). The two doublets are probably a consequence of Ar-H in the mono- and di-methylation at Hf. Ethylene (δ = 5.24 ppm at 40 °C) can only be observed at T > 25 °C. Methane (δ = 0.16 ppm at 25 ºC) is produced in the range of 25-80 ºC (Figure 2.38). The above-mentioned terminal vinyl group at δ = 4.92 and 5.73 ppm (CH2=CH-CH2-R species) remains visible in the range of 25-80 °C. It changes only slightly in intensity while ROMP-derived poly(NBE) (δ = 5.43 ppm (trans) and 5.26 (cis)) formed slowly but steadily (Figure 2.39). Thus, upon activation with MAO, Hf-1 must form a VIP-active cationic species that quickly consumes ethylene that is present in the system. α-H elimination results in the formation of a Hf-alkylidene, which upon reaction with E via cross metathesis forms a Hf-methylidene (Hf=CH2) and vinyl-terminated oligoethylenes. These oligomers are visible in the spectra in spite of the invisible Hf=CH2. Hf=CH2 then starts the ROMP of NBE. The low ROMP propensity of Hf-1 explains why only VIP-derived poly(NBE)-co-poly(E) is produced in the copolymerization of E with NBE.

83

Figure 2.38 Variable-temperature ¹H NMR spectra of Hf-1/MAO/NBE in toluene-d8

(25-80 ºC).

Figure 2.39 Variable-temperature ¹H NMR spectra of Hf-1/MAO/NBE in toluene-d8

(4.80-5.85 ppm, 25-80 ºC).

84

Figure 2.40 ¹H NMR spectra of catalyst/MAO and catalyst/MAO/NBE in toluene-d8 at 60 ºC (4.7-5.9 ppm).

85

As outlined above, the boron in Zr-1 is fully coordinated to the nitrogen both in the solid state and in solution (Figure 2.5 and 2.13). Upon addition of MAO or MAO/NBE, the borane remains mostly tetracoordinated, however, some tricoordinated species become visible, too (Figure 2.14-2.15). This rather small amount of tricoordinated borane translates into a small amount of free pyridine, which in the presence of MAO is capable of starting the ROMP of NBE at elevated temperatures (Table 2.10).

However, in case ethylene is present at the same time, no ROMP-derived structures are observed (Table 2.13). Obviously, similar to Hf-1, most of the catalyst forms a cationic alkyl complex after activation with MAO, resulting in poly(NBE)VIP-co-poly(E).

Upon treatment of Zr-1 with MAO at -60 °C, methane (δ = 0.26 ppm at -60 °C, 0.17 ppm at 20 °C) starts to evolve, a process that becomes more visible with increasing temperature. Clearly, the corresponding cationic species RR’Zr⁺-CH2-Al(CH3)-O- forms from RR’Zr⁺-CH3 and MAO (Figure S38-S39). Zr-1/MAO at 60 °C also shows vinyl-terminated oligoethylenes (Figure 2.40). In the presence of NBE, small amounts of poly(NBE)ROMP starts to form at -10 °C (cis:trans = 80:20) again accompanied by the formation of methane (Figure S40-S41). No Zr-alkylidene is observed, probably because its concentration is too low or the signal is overlapped by Ar-H. Zr-1/MAO/NBE at 60 °C again shows terminal vinyl groups, however, with somewhat different chemical shift from that of the oligoethylenes and multiplicity for the signal at δ = 5.67 ppm (Figure 2.40). This terminal vinyl group is considered to be adjacent to a 1,3-cyclopentylen ring, -(c-1,3-C5H8)-CH=CH2, which is believed to result from the reaction of a Zr-methylidene with NBE referring to the reaction sequence of Zr-B/MAO/NBE.^[108] In addition, signals for poly(NBE)ROMP become visible (Figure 2.40).

In contrast to Zr-1, the pyridyl group in Zr-2 is not coordinated to the boron in the solid state (Figure 2.7). In solution, equally substantial and comparable amounts of tri- and tetracoordinated boranes are observed, both in the absence and presence of MAO and NBE (Figure 2.17, S1-S2). Upon activation with MAO, an active cationic species forms via release of methane (Figure S42). Larger fractions of free pyridine favor α-H elimination and in the presence of NBE predominantly cis-poly(NBE)ROMP

forms even at -50 °C (Figure S43). This is why Zr-2/MAO in the presence of both E and NBE produces poly(NBE)VIP-co-poly(NBE)ROMP-co-poly(E). No methane is visible even at elevated temperature. Again, no Zr-alkylidene is observed.

86

Scheme 2.6 Proposed mechanism for the switch between VIP and ROMP.

87

The proposed mechanism for the formation of poly(NBE)VIP-co-poly(NBE)ROMP -co-poly(E) (path A: a switch from VIP to ROMP; path B: a switch from VIP to ROMP combined with from ROMP to VIP) is shown in Scheme 2.6. In path A, upon addition of MAO, the cationic species A-II forms from a metal dichloro or dialkyl complex A-I.

Potentially, A-II can be stabilized by the coordination of the pyridine nitrogen. A-II can further react with MAO to produce methane and A-III, i.e. RR’Zr⁺-CH2-Al(CH3)-O-, which is inactive in the polymerization.^[109-112] A-II can incorporate E and NBE to form A-IV in which E insertion is favored. This is in line with the high diffusivity of E in the resulting E-NBE copolymers. In path A, the catalyst initiates first vinyl-insertion copolymerization of E with NBE. Insertion of E or NBE in A-IV followed by an α-H elimination[3,96,113-114]

promoted by the pyridine nitrogen through a six-membered transition state produces the Zr-alkylidene (A-V), which is a ROMP-active species. As outlined earlier,^[2,18] substantial steric congestion and excess NBE are required to induce the switch from VIP to ROMP. In that respect, α-H involved in the elimination process of path A is possibly from the last insertion NBE molecule. A-V can now initiate the ROMP of NBE to generate A-VI, which is able to further start ROMP of NBE and form A-VII.

In path B, the initial formation of zirconium alkylidene (Zr=CHR′, B-I) from the cationic species A-II is hypothesized. This process proceeds by an α-H abstraction probably induced by the pyridyl group. The opening of the N-B bond generates a sufficient fraction of free pyridine moiety that induces α-H elimination. In line with that,

11B NMR shows apart from the parent tetracoordinated B atom, the formation of a tricoordinated B species indicates the dissociation of the N-B bond. Thus, the free pyridine moiety is believed to form through the generation of a tricoordinated boron species in the presence of MAO and MAO/NBE and initiate α-H elimination.

Additionally, the immediate formation of ROMP-type poly(NBE) was observed in the

1H NMR in the presence of MAO and NBE (Figure 2.40) and VIP-type poly(NBE) was invisible in the NBE homopolymerization (Table 2.10). Therefore, this system must experience the formation of the metal-alkylidene. B-I is able to initiate the ROMP of NBE to form B-II. In the presence of E, B-I can yield a Zr-methylidene (Zr=CH2, B-III) via cross metathesis, which is also ROMP-active. The previous observation that high NBE concentrations (Table 2.13, entries 12 and 14) are required to form any ROMP-derived sequences strongly suggests that cross

88

metathesis of any Zr-alkylidene with E to form Zr-methylidene is the predominant reaction in these systems. The transient Zr-methylidene (B-III) is presumably exhausted to most extent as a result of its low stability, e.g., via bimolecular decomposition resulting in the formation of ethylene.^[108] This accounts for the low ROMP propensity and the low productivity of these pre-catalysts upon activation with MAO in the E-NBE copolymerization. As already outlined, this also accounts for the finding that no polymer is obtained in the copolymerization of E with NBE by the action of [Zr-2(CH3)]⁺, [Zr-2(Bn)]⁺ or [Zr-2(CH2SiMe3)]⁺[B(C6F5)4]^- in the absence of AlⁱBu3. B-III is able to start the ROMP of NBE. As observed for Zr-B/MAO/NBE^[108]

and Zr-1/MAO/NBE, a terminal vinyl group (B-IV) formed via reaction of Zr=CH2 with NBE can be clearly seen in the ¹H NMR at δ = 4.86 and 5.67 ppm (Figure 2.40). B-II and B-IV are in principle capable of further initiating the ROMP of NBE to form B-V.

Competition between the insertion of NBE and the re-formation of the N-B bond is proposed. The α-H addition to B-V generates the cationic VIP-active species B-VI, which is now able to insert E and NBE to form B-VII and again E incorporation is favored. The formation of vinyl terminals at PE sequences in poly(NBE)VIP -co-poly(NBE)ROMP-co-poly(E) produced by Zr-2(CH2SiMe3)2/MAO could be explained by β–hydride elimination in B-VII. Alternatively, α–H elimination in B-VII followed by cross metathesis with E also provides a terminal vinyl group at PE in the block copolymers.

Both mechanistic routes (path A and path B) are capable of producing E-NBE copolymers with both ROMP- and VIP-type sequences. Although it is difficult to distinguish them, path B appears more reasonable. The explanation is that the quaternary carbon (A-V and A-VI) was not observed in the resulting E-NBE copolymers. Moreover, VIP-type NBE-NBE diads in the copolymers produced by Zr-2 and Zr-2(CH2SiMe3)2/MAO were invisible, which reveals the low propensity of the formation of A-VI from A-V as a result of highly steric hindrance around the metal center. Instead, B-IV was observed for Zr-B and Zr-1 in the presence of MAO and NBE^[108] (Figure 2.40). In spite of the invisible B-IV for Hf-1 and Zr-2 in the presence of MAO and NBE, poly(NBE)ROMP was indeed found even at low temperature which must proceed by the formation of metal-alkylidene and then initiate the ROMP of NBE, which reveals the initial formation of a metal alkylidene is plausible. In addition, ethylene was observed (Figure 2.40), which is considered to be generated by

89

bimolecular decomposition of the Zr-methylidene B-III. Further implication comes from the vinyl terminals at PE sequences in the resulting block copolymers produced by Zr-2(CH2SiMe3)2/MAO (δ = 5.9, 5.0 ppm in ¹H NMR and δ = 139.3, 114.1 ppm in

13C NMR). This vinyl end group is a consequence of β–hydride elimination or α-H elimination in combination with cross metathesis with E. This process is plausible in path B rather than in path A since the occurrence of this process in path A would result in either separate polymers instead of block copolymers (e.g., β–hydride elimination in VIP-active species A-IV) or the invisible terminal vinyls at PEs in the copolymers (e.g., cross metathesis of E with A-VI).

The proposed switch from ROMP to VIP, i.e. α-H addition, is remarkable, since for such a step the pyridinium (Py-H⁺) moiety formed in course of α-H abstraction must be stable in the presence of MAO at least for a short time. In contrast to AlⁱBu3 and most probably because of its size, MAO neither effectively blocks the pyridine moiety nor quickly deprotonates the Py-H⁺ moiety, which explains both for the ROMP propensity in the presence of MAO and the ROMP-inactivity in the presence of AlⁱBu3. In fact, as outlined earlier,^[2] higher MAO concentrations result in larger fractions of ROMP-derived units in E-NBE copolymers. Any additional α-H elimination in course of the copolymerization would regenerate the ROMP-active species, however, in view of the high propensity of the system to undergo cross metathesis with E (vide supra), only very few additional ROMP-derived poly(NBE) sequences can be expected to form (presumably < 1% with respect to the initial amount of pre-catalyst).

Instead, because of the instability of the Zr-methylidenes at elevated temperature, polymerization quickly comes to an end, which is indeed observed. High NBE concentrations were found to stabilize the metal alkylidene in Zr-2 and to encourage ROMP of NBE resulting in high proportion of ROMP-derived poly(NBE) units in the copolymers (Table 2.13, entries 12 and 14). Vice versa, ethylene pressures > 4 bar shift the reaction towards VIP (Table 2.13, entries 15-16).

All together, the data presented here are in line with the previous proposal,^[2-3,18]

which show that high NBE concentrations promote the ROMP process. In addition, a crowded ligand sphere around the metal as found in Zr-2 and Zr-2(CH2SiMe3)2

favors the α-H elimination/addition process. The low propensity of pyridine to coordinate to boron in Zr-2 and Zr-2(R)2 (R = CH3, Bn,CH2SiMe3) clearly stems from the sterics provided by the mesityl groups at boron, which together with the η⁵

-90

tetramethylcyclopentadienyl (Cp*) ligand simply prevents any extensive coordination.

The Cp* ligand as a six-electron donor can sterically and maybe also electronically stabilize the metal alkylidene^[115] and thus promote the ROMP of cyclic olefins.

Whether the strong Lewis-base character of the Cp* moiety favors α-proton abstraction remains speculative. With respect to the sterics, the CH2SiMe3 group in the system Zr-2(CH2SiMe3)2/MAO is believed to provide substantial steric hindrance and to promote α-H abstraction. In view of the above discussed sterics, it is also not surprising at all that Hf-1 with minor steric constraints around the metal shows no tendency to switch from VIP to ROMP in the presence of E regardless of both VIP- and ROMP-derived poly(NBE) obtained in NBE homopolymerization (Table 2.10, entries 3-4). This and the fact that Zr-3 without a 6-(2-BR2-phenyl)pyrid-2-ylamido group does not show any ROMP-activity, neither for NBE nor for E-NBE, strongly support an involvement of the borylamino group that promotes an α-H abstraction process. Finally, apart from high NBE concentrations (catalyst:NBE > 1:5000), the bulky ligand sphere (mesityl groups at boron, Cp* and alkyl ligands at metal) and the crucial 6-(2-BR2-phenyl)pyrid-2-ylamido motif, copolymerizations must be MAO-co-catalyzed catalyst, which ever, must allow for α-H elimination in order to observe ROMP-derived structures. This is in most copolymerization systems not the case. In fact, particularly industrial large-volume systems sometimes contain substantial amounts of aluminum alkyls, which not only promote chain transfer and increase productivity^[116-118] but also effectively prevent the formation of any ROMP-active sites by blocking any Lewis-basic groups.

The mechanistic studies here clearly support the formation of a Zr-alkylidene in the ROMP of NBE via pyridine-induced α-H abstraction and re-generation of a VIP-active species via an α-H addition. Other complexes bearing the 6-[2-BR2 -phenyl]pyrid-2-ylamido motif such as Zr-1, Hf-1, Ti-2, Zr-2(CH3)2 and Zr-2(Bn)2 can in principle also be capable of promoting the ROMP of NBE in the presence of E but might be too unstable toward E to allow for any ROMP-derived poly(NBE) structures under such conditions.

While implications on the copolymerization of E with NBE using “standard”

metallocenes are clear, the low polymerization activity remains a challenge. The highly constrained geometry of Zr-2 and Zr-2(CH2SiMe3)2 and the resulting low tendency to undergo α-H elimination together with the high propensity to undergo

91

cross-metathesis with E account for a very low productivity. In spite of improved activity of Ru-benzylidene catalysts in olefin metathesis, Zr-2(Bn)2/MAO shows no tendency to promote α–H elimination in the presence of E. The plausible approach to boost productivity is an internal olefin that forms a more stable alkylidene in course of cross metathesis with Zr-alkylidene. Additionally, internal olefins are unable to undergo VIP.