Living polymerization of olefins with ansa-dimethylsilylene(fluorenyl)(amido)dimethyltitanium-based catalysts

Abstract

This article reviews the living homopolymerization and copolymerization of propene, 1-alkene and norbornene with ansa-dimethysilylene(fluorenyl)(amido)dimethyltitanium, Me₂Si(η³-C₁₃H₈)(η¹-N^tBu)TiMe₂ and its derivatives, correlating the effects of cocatalysts, solvents, polymerization conditions and the substituents of the fluorenyl ligand with catalytic features, such as livingness, initiation efficiency, propagation rate, syndiospecificity and copolymerization ability. The synthesis of novel olefin block copolymers and their catalytic synthesis are also introduced using this living system.

Main

A living polymerization system, in which neither chain transfer nor deactivation occurs, affords polymers with predictable molecular weights and narrow molecular weight distributions (MWDs). Living polymerization techniques are utilized for the synthesis of terminally functionalized polymers and block copolymers. Another feature of living polymerization is its simple kinetics. Because neither chain transfer nor deactivation occurs in living polymerization, the number of polymer chains (N) is equal to the number of active centers ([C^*]), and the number-average molecular weight (M_n) directly reflects the propagation rate. In living polymerization, the propagation rate can be evaluated from the M_n value and the polymerization time. As chain transfer via β-elimination and/or transmetalation with trialkylaluminum as a cocatalyst is inevitable in conventional Ziegler–Natta catalytic systems; homogeneous V(acac)₃/Et₂AlCl (where acac is acetylacetonate or its analog) had previously been the only catalytic system that produced syndiotactic-rich living polypropylene (PP) at −78 °C.^{1, 2}

The development of metallocene catalysts has enabled us to produce a variety of uniform olefin copolymers and to control the stereoregularity of poly(1-alkene)s.^{3, 4} The success of metallocene catalysts has stimulated research on transition metal complexes for olefin polymerization,^{5, 6, 7} so-called single-site catalysts, which has also resulted in various transition metal complexes for the living polymerization of olefins.^{8, 9}

We have also developed a living polymerization system using ansa-dimethysilylene(fluorenyl)(amido)dimethyltitanium, Me₂Si(η³-C₁₃H₈)(η¹-N^tBu)TiMe₂ (1), combined with a suitable cocatalyst. The catalytic system allows not only a syndiospecific living polymerization of propene and a 1-alkene but also a living homopolymerization and copolymerization of norbornene with a 1-alkene. We investigated the effect of cocatalyst, solvent and the ligand of the titanium complex on propene polymerization in terms of the livingness of the system. We also synthesized novel olefin block copolymers with the living system. This article reviews the characteristics of 1-based catalysts for tailor-made polyolefins.

Synthesis and structures of titanium complexes

Synthesis of titanium complexes

The titanium complexes that appear in this article are shown in Scheme 1. The ansa-monocyclopentadienylamido ligand for olefin polymerization catalyst was originally introduced by Bercaw and colleagues¹⁰ in a scandium complex, and the titanium complex with ansa-monocyclopentadienylamido ligand was first reported by Okuda¹¹ The commercial application of group IV ansa-monocyclopentadienylamido complexes was developed by Dow and Exxon and was named constrained geometry catalysts (CGCs) after the structure of the complexes. CGCs are superior in the copolymerization of ethene with 1-alkene or styrene compared with bis(cyclopentadienyl) group IV metal complexes and have been applied for the production of ethene-based copolymers such as linear low-density polyethylene. Ansa-Me₂Si(η⁵-C₅Me₄)(η¹-N^tBu)TiMe₂ (CGC-Ti in Scheme 1) is a representative CGC.¹²

Scheme 1

Alkyltitanium complexes used in this study.

We succeeded in the synthesis of the fluorenyl derivative 1 of CGC-Ti, the synthetic routes of which are illustrated in Scheme 2. We originally synthesized 1 via route II from TiCl₃ and the ligand via multisteps with very low yields.¹³ We applied route I,¹⁴ which starts from TiCl₄, the ligand and four molar equivalent of MeLi, to give 1 through a one-pot synthesis and succeeded in the synthesis of 1 and Zr and Hf analogs (1-Zr and 1-Hf) with 42–43% yields.¹⁵ The fluorenyl derivatives 2–4,^{16, 17} the indenyl derivative¹⁸ and CGC-Ti were also synthesized by this procedure with 20–60% yields. All of these complexes were isolated as crystals, and their molecular structures were determined by single-crystal X-ray analyses.

Scheme 2

Synthetic routes of 1.

The ORTEP drawings of CGC-Ti and 1 are illustrated in Figure 1. These complexes are activated via the abstraction of one of the two Me groups as Me^– to form coordinatively unsaturated cationic Ti–Me species, as shown in Equation (1). The polymerization begins by the coordination of monomer followed by the migratory insertion into the Ti–Me bond, and the propagation proceeds via successive monomer coordination and insertions.

Figure 1

ORTEP diagrams of CGC-Ti and 1.

The bond lengths from Ti to the cyclopentadienyl or the fluorenyl ligand determined by X-ray crystallography are summarized in Table 1. In CGC-Ti, the bond lengths between Ti and five carbons of the C₅Me₄ group are in the range 2.28–2.46 Å, indicating the η⁵-coordination of the C₅Me₄ ligand. In contrast, the bond lengths between Ti and the fluorenyl ligand in 1 become elongated as follows: Ti-C¹ (2.25 Å) < Ti-C² (2.42 Å) < Ti-C³ (2.57 Å). The other Ti fluorenyl derivatives, 2–4, show the same tendency, indicating the η³-coordination of the fluorenyl ligand in the Ti complexes. In addition, the planar fused benz-groups of 1–4 are less bulky than the Me groups of CGC-Ti. These structural characteristics produce the open coordination site in the ansa-dimethylsilylene(fluorenyl)(amido)dimethyltitanium complexes, 1–4, as compared with CGC-Ti. The difference in the bond length between Mt–C¹ and Mt–C³ in 1 is 0.32 Å, whereas that between Mt–C¹ and the Zr and Hf analogs 1-Zr and 1-Hf is 0.27 Å and 0.26 Å, respectively, indicating that the tendency for η³-coordination of the fluorenyl ligand is weakened in 1-Zr and 1-Hf.

Table 1 Bond lengths between metal and cyclopentadienyl ligand in CGC-Ti and 1–4

Polymerization of propene and higher 1-alkene

Effects of cocatalyst

Propene polymerization was conducted by activating 1 with methylaluminoxane (MAO), Ph₃CB(C₆F₅)₄ or B(C₆F₅)₃ in toluene at 40, 0 and −50 °C by in-batch polymerization, where a certain amount of monomer was first charged. The results are shown in Table 2. In the case of Ph₃CB(C₆F₅)₄ and B(C₆F₅)₃, trioctylaluminum (Oct₃Al) was added as a scavenger. MAO showed the highest yields at 40 and 0 °C to give low molecular weight polymers with M_w/M_n values of 2.5 and 1.8, respectively. The N value was determined from the M_n value and the polymer yield. The ratios of N to the Ti complex used (N/Ti) were 85 and 38, respectively, indicating the occurrence of chain transfer reaction. In the ¹³C nuclear magnetic resonance (NMR) data of the produced polymers, the resonances of terminal isopropyl group were observed in addition to the resonance of regioregular main-chain propene sequence. Any other resonance was not detected. The results clarified that the propene polymerization with this system proceeded via 1, 2-addition highly regiospecifically, and the chain transfer with Me₃Al in MAO selectively occurred as shown in Scheme 3.¹³

Table 2 Effects of activators on propene polymerization with 1^a

Scheme 3

Propene polymerization with 1-MAO.

When Ph₃CB(C₆F₅)₄ or B(C₆F₅)₃ was used, propene polymerization proceeded although the yield was very low. The N/Ti values were close to or <1 irrespective of the polymerization temperature, indicating the suppression of chain transfer reactions compared with MAO. The polymers obtained with B(C₆F₅)₃ at −50 °C showed an M_w/M_n value of 1.2, which suggests the living polymerization with 1–B(C₆F₅)₃. The time course of the polymerizations with B(C₆F₅)₃ and Ph₃CB(C₆F₅)₄ at −50 °C is shown in Figure 2.

Figure 2

Kinetic profiles of propene polymerization with 1 activated with B(C₆F₅)₃ (•, ○) and Ph₃CB(C₆F₅)₄ (▪, □). Polymerization conditions: Ti=B=40 μmol, Oct₃Al=800 μmol, toluene=30 ml, propene=3.5 g, temperature=−50 °C: time dependence of (a) conversion, (b) M_n, (c) M_w/M_n and N/Ti.¹⁹

The conversion increased almost linearly with increasing polymerization time when B(C₆F₅)₃ was used, whereas it saturated below 10% when Ph₃CB(C₆F₅)₄ was used (Figure 2a). The time dependences of the M_n values were similar to that of the conversions, namely the M_n value in the B(C₆F₅)₃ system was proportional to polymerization time, whereas that in the Ph₃CB(C₆F₅)₄ system showed a saturating tendency as shown in Figure 2b. These results indicate that the living polymerization of propene proceeded with 1–B(C₆F₅)₃, whereas deactivation occurred in 1–Ph₃CB(C₆F₅)₄ at −50 °C.¹⁹

The N/Ti value was almost constant during the polymerization, irrespective of the cocatalyst employed (Figure 2c). The N/Ti values indicate that about a half of the total Ti was active during the polymerization in 1–B(C₆F₅)₃, whereas only about 15% of Ti was converted into active species in 1–Ph₃CB(C₆F₅)₄.

Because the N/Ti values were constant during the polymerization, the slope of M_n versus polymerization time plots in Figure 2b corresponds to the propagation rate constant of each system. The initial slopes indicate that the Ti species activated with Ph₃CB(C₆F₅)₄ was more than 10 times as active as that activated with B(C₆F₅)₃, although it was quickly deactivated.

These phenomena are interpreted as an interaction between the cationic Ti active species and the counter anion.²⁰ B(C₆F₅)₃ partly abstracts the alkyl ligand to form ‘cation-like’ active species as shown in Equation (2), whereas Ph₃CB(C₆F₅)₄ gives the non-coordinating counter anion of [B(C₆F₅)₄]⁻ according to Equation (3). The former is less active but more stable in producing a living polymer.

Because the activation process with B(C₆F₅)₃ in Equation (2) is an equilibrium reaction, excess B(C₆F₅)₃ should be favored to increase the initiation efficiency. The effects of additional B(C₆F₅)₃ were clearly observed, as shown in Table 3 (Runs 10–13). A small excess of B(C₆F₅)₃ improved the conversion by a factor of 4.5, accompanied by the increase in the M_n value by a factor of 3.5; however, the N/Ti value increased by only a factor of ∼1.3. The further addition of B(C₆F₅)₃ increased the M_n value much more than the N/Ti value. It should be noted that the polymerization was conducted in a batch-type operation in which the increase of the M_n value is underestimated. These results imply that the main effect of excess B(C₆F₅)₃ is to enhance the propagation rate.²¹

Table 3 Effects of B(C₆F₅)₃ and Oct₃Al on propene polymerization with 1^a

Oct₃Al, which had previously been regarded as a scavenger, was also found to enhance the propagation rate. The polymer was obtained even in the absence of Oct₃Al, although the propagation rate was very low (Runs 14 and 15): the M_n value of the polymer reached 5100 after 12-h polymerization. The addition of Oct₃Al (Al/Ti=10) greatly promotes the propagation rate to give the polymer with an M_n value of 102 000 in 30 min (Run 16).

The enhancement of propagation rate by the addition of B(C₆F₅)₃ and Oct₃Al can be interpreted by observing Scheme 4. An electron-deficient group 13 compound (R₃M, where M is B or Al) can interact with the ion pair of the active species (A) in two ways. One is the coordination with the counter anion to give the coordinatively unsaturated Ti cationic species (B), and the other is the formation of hetero-bimetallic species with the cationic Ti species (C). If coordination to the counter anion selectively occurs, which should be the case of B(C₆F₅)₃ and Oct₃Al, the activity is increased by the formation of highly active coordinatively unsaturated Ti species B.

Scheme 4

Plausible reactions of B(C₆F₅)₃ and trialkylaluminum with the active Ti species.

When the hetero-bimetallic species C is formed, two possibilities can be considered. If the coordination of R₃M to the Ti species is in equilibrium, the chain transfer to R₃M should occur. If C is very stable, the activity should be suppressed. The formation and the stability of C should depend on R in R₃M as well as on the ligand of the Ti complex. The effects of various R₃Al were, therefore, investigated in the propene polymerization with 1–B(C₆F₅)₃ at −50 °C (Table 4).²² The polymerization features strongly depended on the added R in R₃Al. The polymer yield decreased in the following order: Oct₃Al⩾Et₃Al≫ⁱBu₃Al>Me₃Al=0. Although Et₃Al and Oct₃Al afforded polymers in similar yields, the M_n value with Et₃Al was one order of magnitude lower than that with Oct₃Al to give the N/Ti value of 7.2. The result shows the high chain transfer ability of Et₃Al, indicating that the coordination of Et₃Al to the cationic Ti species is in equilibrium under the present conditions. The M_n value with ⁱBu₃Al was also one order of magnitude lower than that with Oct₃Al, but the N/Ti value was almost the same as that with Oct₃Al. In addition, the M_w/M_n value with ⁱBu₃Al was 1.3. These results indicate that the living polymerization also proceeded in the presence of ⁱBu₃Al although the rate enhancement by ⁱBu₃Al was one order of magnitude lower than that by Oct₃Al. The very poor polymer yield with Me₃Al can be interpreted by the formation of the stable hetero-binuclear species.

Table 4 Additive effects of trialkylaluminum on propene polymerization with 1-B(C₆F₅)₃ at −50 °C^a

As described above, it was proved that the living polymerization of propene proceeded with 1 activated by B(C₆F₅)₃ at −50 °C in toluene, and the propagation rate was enhanced by the addition of excess B(C₆F₅)₃ and/or a suitable amount of Oct₃Al.

Activation with Me₃Al-free MAO

McConville and Scollard²³ demonstrated the living polymerization of higher 1-alkene at 23 °C with [Ar’N(CH₂)₃NAr’]TiMe₂ (Ar’=2, 6-ⁱPrC₆H₃ and so on) activated by an equimolar amount of B(C₆F₅)₃. They clarified, however, that the catalytic system was deactivated in the absence of monomer accompanied by evolving methane, according to Scheme 5.²⁴ They also reported that the use of MAO in place of B(C₆F₅)₃ produced more thermally stable active species, although chain transfer to Me₃Al selectively occurred.²⁵

Scheme 5

Deactivation of [Ar’N(CH₂)₃NAr’]TiMe₂–B(C₆F₅)_3.

In the propene polymerization with 1, MAO showed higher activity than with B(C₆F₅)₃ at 0 °C and 40 °C to give low molecular weight PP via selective chain transfer to Me₃Al shown in Scheme 3. If the deactivation of the 1–B(C₆F₅)₃ system at elevated temperature occurs via the reaction between the cationic Ti species and [MeB(C₆F₅)₃] similar to the [Ar’N(CH₂)₃NAr’]TiMe₂–B(C₆F₅)₃ system, the deactivation process can be eliminated by the use of MAO in place of B(C₆F₅)₃. In addition, if Me₃Al is absent in MAO, the chain transfer reaction can also be eliminated to produce a living polymerization system. We therefore prepared Me₃Al-free MAO by vacuum drying a toluene solution of MAO, followed by washing with hexane and applying it as a cocatalyst.²⁶ Thereafter, the MAO treated via this procedure was denoted as dMAO, an abbreviation of dried MAO. Figure 3 shows the ¹H NMR spectra of MAO and dMAO, which indicates that dMAO does not contain free Me₃Al.

Figure 3

¹H NMR spectra of MAO (a) and dMAO (b) in benzene-d₆.²⁶

Propene polymerization was conducted with 1 at 0 °C using MAO or dMAO as a cocatalyst in a semi-batch system, in which the propene pressure was kept at atmospheric pressure during the polymerization, and the consumed monomer was continuously supplied.²⁶ Hence, the polymerization rate can be monitored by the flow rate of propene introduced. CGC-Ti was also used for comparison.²⁷ The results are shown in Table 5. The polymerization time was changed in each experiment according to the catalytic activity because the produced polymer interfered with effective stirring.

Table 5 Effects of activator on propene polymerization with 1 and CGC-Ti at 0 °C^a

The activities with dMAO were one order of magnitude higher than that with MAO irrespective of the titanium complexes employed. The M_n values were also one or two orders of magnitude larger in the dMAO systems than the MAO systems. The polymers obtained with the dMAO systems showed an M_w/M_n value of ∼1.2, suggesting the living nature of propene polymerization with 1 and CGC-Ti using dMAO as a cocatalyst.

To investigate the livingness of the 1–dMAO and CGC-Ti–dMAO systems, propene polymerization was conducted at 0 °C in toluene with batch operation by changing the amount of monomer fed. Regardless of the titanium complex, polymerization proceeded quantitatively within 1 h. Figure 4 displays the relation between the yield and the M_n value. The M_n value increased against the yield in both systems. The linearity was, however, better in the 1–dMAO system than in the CGC-Ti–dMAO system. In the 1–dMAO system, the N/Ti value was almost constant, whereas it gradually increased in the CGC-Ti–dMAO system, accompanied by an increase in the M_w/M_n value. The results indicate that the chain transfer reaction occurred in the latter system under these conditions. The post-polymerization also supports the formation of living PP with 1–dMAO at 0 °C. The use of dMAO in place of B(C₆F₅)₃ was found to raise the living polymerization temperature of 1 to 0 °C.

Figure 4

Plots of M_n and N/Ti against yield in propene polymerization with dMAO: ▪, □; CGC-Ti: •, ○, 1. Polymerization conditions: 1 h; for other conditions, see Table 5.

The lower activity and lower molecular weight observed with MAO can be ascribed to the presence of Me₃Al, which forms hetero-binuclear species C in Scheme 4. The M_n and N/Ti values with MAO and dMAO in Table 5 indicate that chain transfer with Me₃Al frequently occurred in the 1–MAO system but rarely in the CGC-Ti–MAO system.

Activation with R₃Al-free modified MAO (MMAO)

As the active species in Ti complexes for olefin polymerization is an ion pair, as shown in Equation (1), the activity, namely [C^*] and the propagation rate constant (k_p) should depend on the solvent employed. In living polymerization, the [C^*] value is equal to the N value determined from the polymer yield and the M_n value; the k_p value can be directly evaluated from the M_n value and the polymerization time.

McConville and Scollard investigated the effects of solvents in the living polymerization of 1-alkene with [Ar’N(CH₂)₃NAr’]TiMe₂–B(C₆F₅)₃ by replacing a part of the 1-alkene with CH₂Cl₂ or toluene.²³ The replacement with CH₂Cl₂ increased the M_n value by a factor of ∼30, which can be attributed to the charge separation of the ion pair by the polarity of CH₂Cl₂, although the initiation efficiency was reduced. The replacement with toluene suppresses the activity because of the formation of the Ti η⁶-toluene species.

We succeeded in the living polymerization of propene with 1–dMAO in toluene at 0 °C, the kinetics of which should also depend on the solvent employed. However, the solubility of MAO in saturated hydrocarbons such as hexane is not good and further decreases after the removal of Me₃Al to obtain dMAO. MMAO, which is the condensation product of water and the mixture of Me₃Al and ⁱBu₃Al, possesses higher solubility than MAO and is easily soluble in hexane. We therefore removed the remaining Me₃Al and ⁱBu₃Al in MMAO by repeated vacuum drying and successive redissolving in heptanes to obtain R₃Al-free MMAO (dMMAO).²⁸ Figure 5 shows the ¹H NMR spectra of MMAO and dMMAO in tetrahydrofuran-d₈. The contents of each Al species were calculated from the relative intensities in the ¹H NMR spectra, as shown in Table 6, which indicates that the treatment reduced the contents of Me₃Al and ⁱBu₃Al to ∼0.2 mol% of the total Al species.²⁹

Figure 5

¹H NMR spectra of MMAO (a) and dMMAO (b) in THF-d₈.²⁹

Table 6 Composition of MMAO and dMMAO

Propene polymerization by a batch method was then conducted with 1 activated by dMAO or dMMAO in heptane (Table 7).³⁰ The dMAO system produced the polymer at a 65% conversion with a broad MWD (M_w/M_n=2.10) for 1-h polymerization, whereas the dMMAO system quantitatively produced the polymer with a narrow MWD (M_w/M_n=1.32). The broad MWD using dMAO in heptane is mainly attributed to the low solubility of dMAO: dMMAO is perfectly soluble in heptane, whereas dMAO is sparingly soluble in heptane under the polymerization conditions. The heterogeneity of the precipitated dMAO should have caused the broadening of MWD. We also succeeded in the living polymerization of propene with the [Ar’N(CH₂)₃NAr’]TiMe₂–dMMAO (Ar’=2,6-ⁱPr₂C₆H₃) system in heptane at 0 °C.²⁸

Table 7 Effect of activator on propene polymerization with 1 in heptane^a

As the potential of dMMAO as a cocatalyst in heptane was confirmed, we investigated the effect of solvents, namely toluene, chlorobenzene (CB) and ortho-dichlorobenzene (o-DCB), on propene polymerization with 1–dMMAO by a batch method conducted over 1 h.³⁰ The polymerizations quantitatively proceeded, irrespective of the amount of propene charged and the solvent employed. The produced polymers showed narrow MWDs of 1.14∼1.39. The M_n values are plotted against the polymer yields in Figure 6, which form a straight line through the origin regardless of the solvent employed.

Figure 6

Plots of M_n versus polymer yield in propene polymerization with 1–dMMAO at 0 °C in heptane, toluene, CB and o-DCB. Polymerization conditions: Ti=20 μmol, dMMAO=8.0 mmol, solvent=30 ml, propene=1.26 g, temperature=0 °C, 1 h. Post-polymerization conditions: propene=1.26 g, 1 h.³⁰

To confirm the livingness of the systems, we conducted post-polymerization in heptane and o-DCB, which have the lowest and the highest dielectric constant, respectively, among the solvents employed. The values are located on the same straight line, as shown in Figure 6. These results indicate that the living polymerization of propene proceeded in all of the solvents with 1–dMMAO, and the initiation efficiency was independent of the solvent employed.

The effects of the solvents on the kinetic features were then investigated using a semi-batch method. Propene polymerization was conducted in heptane and o-DCB under atmospheric pressure; the other conditions were the same as those of the batch method, except that 5 μmol of 1 was employed. The kinetic profiles thus obtained are shown in Figure 7.

Figure 7

Rate-time profiles of propene polymerization with 1–dMMAO at 0 °C in heptane and o-DCB. Polymerization conditions: Ti=20 μmol, norbornene=1.2 M, total volume (toluene + norbornene)=30 ml, temperature=20 °C, 3 min.³⁰

Even when 5 μmol of Ti was employed, the o-DCB system showed very high activity, and stirring was stopped within 3 min. In contrast, the consumption rate of propene in the heptane system gradually increased and reached a steady value.

Because the initiation efficiency was the same in heptanes as it was in o-DCB, as shown in Figure 6, the high activity in o-DCB can be ascribed to a high propagation rate. The reason why the propagation rate increased can be explained by the enhancement in the separation of the active Ti cation and the dMMAO-derived anion in the polar solvent. Fink et al.³¹ reported the linear relationship between the polarity of the solvent and the polymerization rate. Our results clearly indicate that the rate enhancement with the polar solvent is ascribed to the increase in the k_p value.

Structure of PP

The ¹³C NMR spectra of the PPs obtained with 1, CGC-Ti and (η⁵-C₅Me₅)Ti(CH₂Ph)₃ activated by dMAO in toluene are illustrated in Figure 8. The contents of steric triad and chemical inversion were determined from the spectra, and the results are summarized in Table 8.

Figure 8

¹³C NMR spectra of PPs obtained with dMAO as a cocatalyst: (a) (η⁵-C₅Me₅)Ti(CH₂Ph)₃; (b) CGC-Ti and (c) 1.

Table 8 Microstructures of PPs obtained with cyclopentadienylalkyltitanium–dMAO systems

Both the regioregularities and stereoregularities of the PPs depended on the titanium complexes used. (η⁵-C₅Me₅)Ti(CH₂Ph)₃ gave a slightly sydiotactic-rich polymer with more than 10% regioirregular units, whereas CGC-Ti produced an almost atactic polymer with about 1% of regioirregular units, as reported by Waymouth and colleagues.³² In contrast, 1 was highly regiospecific and produced a syndiotactic-rich polymer, the syndiotacticity of which was almost the same as that of the polymers obtained with the V-based living polymerization catalyst at very low temperature.^{1, 33} Waymouth and colleagues³² found that the corresponding Zr complex, ansa-Me₂Si (η⁵-C₁₃H₈)(η¹-^tBuN)ZrCl₂, combined with MAO yielded exceptionally regioregular PP and reported that the high regiospecificity could be ascribed to the Zr atom or the fluorenyl ligand because they could not isolate the corresponding Ti complex. The high regiospecificity of 1 implies that the fluorenyl ligand has an important role in the control of regiospecificity.

Living PPs were obtained in various solvents by 1–dMMAO. The high regiospecificity of 1 was maintained in all the solvents used, but the stereospecificity was significantly affected by the solvent. The ¹³C NMR spectra of the PPs in methyl regions are illustrated in Figure 9. The resonance of the syndiotactic pentad, rrrr, became stronger in the following order, CB = o-DCB < toluene < heptane.

Figure 9

125 MHz ¹³C NMR spectra of methyl regions of PPs obtained with 1–dMMAO at 0 °C in heptane, toluene, CB and o-DCB.³⁰

The steric pentad distributions determined from the spectra are shown in Table 9, in which the data of the PP obtained by 1–dMAO in toluene are also indicated for comparison. The pentad values indicate that the syndiotacticity of the PPs obtained with dMAO and dMMAO was almost the same but depended on the solvents used. Fink et al.³¹ studied propene polymerization with ansa-Me₂C(Flu)(Cp)ZrCl₂–MAO in toluene/CH₂Cl₂ mixture and found that the rrrr pentads of the produced PPs monotonously decreased from 89% in pure toluene to 42% in pure CH₂Cl₂. They proposed that the decrease in the stereospecificity in the polar solvents may be caused by the isomerization of the solvent-separated zirconocene species via migration of the growing polymer chain without monomer insertion. Marks and Chen³⁴ also reported the effect of cocatalysts and solvents (toluene and o-DCB) on the kinetics and the stereospecificity of propene polymerization with ansa-Me₂C(Flu)(Cp)ZrMe₂. Busico et al.³⁵ reported a detailed kinetic model of propene polymerization with ansa-R₂C(Flu)(Cp)ZrX₂ (where R is Me and Ph, and X is Me and Cl) on the basis of the microstructure of the produced polymers and discussed the effect of cocatalysts and solvents (toluene and bromobenzene). These papers report that the decrease in syndiospecificity with the C_s-symmetry zirconocene catalysts in polar solvent is ascribed to the ‘chain migration without monomer insertion (skipped insertion)’ and not to the ‘selection miss of enantioface’ (Scheme 6).³⁶

Table 9 Steric pentad distribution of PPs obtained with 1–dMMAO in various solvents

Scheme 6

Enantiomorphic site-controlled syndiospecific polymerization by C_s symmetric catalysts.

The change in the syndiospecificity observed by the present Ti complex can be also interpreted by the separation of an active ion pair. The dMMAO–polar solvent system produces a solvent-separated ion pair and, hence, allows the growing chain to migrate between the two enantiotopic sites on the Ti cation without monomer insertion. In contrast, the dMMAO–heptane system favors the contact ion pair and prohibits the migration of the growing chain.

Actually, the number of stereodefects (•••rmrr•••) arising from the ‘chain migration’ increased in the following order, as shown in Table 9: heptane (10%) < toluene (19%) < o-DCB (24%) < CB (26%). The number of stereodefects (•••rmmr•••) arising from the ‘selection miss’ is almost independent of the solvent employed (4∼5%). We can therefore conclude that the enhancement of syndiospecificity in heptane is ascribed to the suppression of the ‘chain migration’ and not to the selectivity of the prochiral face.

Substituent effects on the fluorenyl ligand

The use of dMMAO facilitated the living polymerization of propene in heptane, which afforded PP with the highest syndiotacticity among the solvents employed. We therefore investigated the propene polymerization with 2, 3 and 4 in heptane activated by dMMAO.^{16, 17} The results of semi-batch polymerizations are summarized in Table 10; the polymerizations were terminated by the addition of acidic methanol before the polymer produced began to interfere with the stirring. When the polymerization was too fast to precisely evaluate the activity, the amount of dMMAO was also reduced.

Table 10 Propene polymerization with 1–4 activated by dMMAO in heptane

The activities of 2,7-^tBu₂-substituted 2 and 3,6-^tBu₂-substituted 3 with Al/Ti=400 were 2280 and 2320 kg-PP·mol-Ti⁻¹ h⁻¹, respectively, which were approximately three times higher than that of the non-substituted 1. The activities were reduced to 787 kg-PP·mol-Ti⁻¹ h⁻¹ and 916 kg-PP·mol-Ti⁻¹ h⁻¹ with Al/Ti=200. These results indicate that the rate enhancement by the introduction of the ^tBu groups is mainly ascribed to an electronic effect, although the 3,6-position is slightly superior to the 2,7-position. Complex 4, which is the 2,3,6,7-^tBu₄-substituted analog of 1, showed the highest activity of 11 400 kg-PP·mol-Ti⁻¹ h⁻¹ among the four complexes, even with Al/Ti=100. The activity of 3 with Al/Ti=100 was 326 kg-PP·mol-Ti⁻¹ h⁻¹.

The effects of the Al/Ti ratio on propene polymerization with 3–dMMAO (Runs 29, 31 and 32) indicate that a large excess of dMMAO is necessary for a high propagation rate because the activity increased at a constant N/Ti value. Similar phenomena were also observed in propene polymerization with 1–B(C₆F₅)₃, where excess B(C₆F₅)₃ or the addition of Oct₃Al enhanced the propagation rate.

The M_n values of the PPs obtained were as high as 15–20 × 10⁴ with an M_w/M_n value of 1.5–1.7, except that obtained with 4, and the N/Ti values were 0.50–0.65. Post-polymerization, in which the same amount of monomer was sequentially added after the first polymerization was completed, was then conducted with 2 and 3 at 0 °C and 25 °C. The gel permeation chromatography (GPC) curves of the produced PPs are illustrated in Figure 10, and the results are summarized in Table 11. The post-polymerizations proceeded quantitatively, and the GPC curves were shifted to higher molecular weight regions while maintaining the MWD narrow and the N/Ti value constant regardless of the Ti complex and the polymerization temperature. The results indicate the living nature of these systems. The MWDs of the batch polymerizations (Table 11) were narrower than that of the semi-batch polymerization (Table 10), which suggests that the broader MWD should be partly due to the physical inhomogeneity of the system caused by the heat of polymerization and/or ineffective stirring. The slow initiation in heptanes, as observed in Figure 7, should also broaden the MWD.

Figure 10

GPC curves of PPs obtained in post-polymerization: (a) 2–dMMAO at 0 °C; (b) 2–dMMAO at 25 °C; (c) 3–dMMAO at 0 °C and (d) 3–dMMAO at 25 °C.¹⁶

Table 11 Post-polymerization of propene with 2 and 3^a

The MWD of the PP obtained with 4 after 30-s polymerization was very broad (M_w/M_n=2.9), although the N/Ti value was only 0.50. Batch polymerization of propene with 4 was conducted in heptane at 0 °C by changing the amount of propene charged. The polymerizations proceeded quantitatively to produce polymers with broad M_w/M_n values of ∼3. The M_n values of the produced polymers are plotted against the polymer yields in Figure 11, which shows a straight line through the origin. The results indicate that the chain transfer reactions did not occur during the polymerization, although the MWDs were broad. Post-polymerization was then conducted with 4–dMMAO in heptane at 0 °C.

Figure 11

Plots of M_n against yield in propene polymerization with 4–dMMAO in heptane.¹⁷

The GPC curves before and after the post-polymerization are illustrated in Figure 12 together with the results of the polymerizations. Although the polymerizations proceeded quantitatively during both stages, the M_n values did not double but only slightly increased and were accompanied by a narrow MWD. The peaks of the GPC curves remained almost unchanged, and the decrease in both the low and high molecular weight fractions produced a narrower MWD during the second stage. Considering the good linear relationship between the yield and M_n in Figure 11, we can conclude that chain transfer selectively occurred, probably due to the remaining ⁱBu₃Al in dMMAO after all of the monomers had been consumed. The narrower MWD during the second stage can be explained by the smooth initiation by the chain-transferred active species. The broad MWD by 4, as compared with the other Ti complexes, can be partly ascribed to the high propagation rate of 4, which makes the effect of slow initiation on the MWD more significant. In addition, the extremely high activity should induce physical inhomogeneity in the reactor, namely temperature and propene concentration, because of the insufficient removal of heat and stirring.

Figure 12

GPC curves of PPs before and after post-polymerization with 4–dMMAO.¹⁷

The substituent of the fluorenyl ligand also affected the stereospecificity of the Ti complexes. Table 12 summarizes the steric pentad distribution of the PPs obtained in heptane at 0 °C. The syndiotactic pentad (rrrr) of the PPs decreased in the following order: 3 (0.86)>2 (0.69)>1 (0.60)>4 (0.45). Considering the similar geometry of 3 and 4 in the coordination site of propene, the syndiospecificity of 4 being the lowest is very strange.

Table 12 Steric pentad distribution of PPs obtained by Ti complexes–dMMAO at 0 °C

Miller et al.^{37, 38, 39} reported that the diethyl-ether-coordinated Zr analog of 4, in which the fluorenyl ligand is coordinated to Zr in η¹-mode 5, showed excellent abilities for highly syndiospecific polymerization of propene and 1-alkenes as well as copolymerization of ethene with higher 1-alkenes. We obtained a tetrahydrofuran (THF)-coordinated Ti analog 6, which conducted copolymerization of ethene and dicyclopentadiene with high activity.⁴⁰ These results suggest the possibility of the isomerization of 4 during the polymerization, as shown in Scheme 7.⁴¹ The PP obtained with 4 contained a considerable amount of mmmm and mmmr, suggesting the presence of isotactic blocks in the syndiotactic sequence. The isotactic block might be produced by the C₁-symmetric Ti species formed by the isomerization of 4.

Scheme 7

Isomerization mechanism of 4.

The diagnostic stereodefects in the syndiotactic sequence, ‘rmmr’ arising from the ‘monomer miss-insertion’ and ‘rmrr’ arising from the ‘chain migration’ without monomer insertion, are reduced as follows: rmmr, 1 (0.04)=2 (0.04)>4 (0.03)>3 (0.02); and rmrr, 4 (0.16)>1 (0.10)>2 (0.05)>3 (0.02). Thus, the high syndiospecificity of 3 can be ascribed to the reduced tendency for isomerization in addition to the high enantioselectivity.

The syndiospecific living polymerization of propene was also achieved by Fujita and colleagues^{42, 43} and Coates and colleagues⁴⁴ using bis(phenoxyimine)titanium derivatives activated with MAO. The active species of this catalyst is an octahedral Ti with a 2,1-inserted propagation chain in which the chain-end chiral carbon isomerizes the chirality of the Ti species from one to the other in each monomer insertion to produce the syndiotactic polymers. Fujita and colleagues⁴⁵ succeeded in improving the syndiospecificity (rr=94%) by tuning the bis(phenoxyimine) ligand. The activity of propene polymerization with this catalytic system was significantly low, probably due to this stereospecific polymerization mechanism, as it conducted the living polymerization of ethene at a very high activity.⁴⁶

One of the advantages in our syndiospecific living polymerization is the high propagation rate for propene polymerization. We therefore investigated the polymerization of a higher 1-alkene with 3–dMMAO at 0 °C in toluene.⁴⁷ The results are summarized in Table 13. Although the activity was reduced as the length of 1-alkene was extended, an activity of 9500 kg-polymer·mol-Ti⁻¹ h⁻¹ was observed in 1-dodecene polymerization to produce a polymer with an M_n value of 308 000 after a 1-min polymerization. As the livingness of the system was confirmed by post-polymerization, the turnover frequency of each polymerization was determined from the M_n value evaluated by SEC-MALS, the molecular weight of the monomer and the polymerization time. The turnover frequency values obtained indicate that the propagation rate of 1-alkene decreased with δ the length of 1-alkene, but reached a definite value above that of 1-decene. Combined with the turnover frequency values of propene (26 s⁻¹) and 1-hexene (8 s⁻¹) obtained at different concentrations of the monomer and the catalyst, the relative reactivity of 1-alkene with 3–dMMAO was determined as follows:

Table 13 Polymerization of higher 1-alkenes with 3–dMMAO at 0 °C^a

The stereoregularity of the poly(1-alkene)s was syndiotactic with rr values of 85–88%, irrespective of the length of 1-alkene.

Synthesis of stereoblock PP

The syndiospecificity of the living ansa-dimethylsilylene(fluorenylamido)dimethyltitanium-based catalyst depends not only on the fluorenyl ligand but also on the polymerization conditions, such as solvent, temperature and monomer concentration, which enables us to synthesize tailor-made stereoblock copolymers composed of crystalline syndiotactic and amorphous PP sequences. Diblock copolymer was synthesized with 1–dMMAO at 0 °C via propene polymerization in heptane (rr=72%), followed by post-polymerization after the addition of CB (rr=44%). The block copolymer (M_n=94 700, M_w/M_n=1.27) did not show any melting temperature (T_m) in the differential scanning calorimetry measurement, although the corresponding PP blend showed a T_m of 94 °C.⁴⁸ The use of 3 in place of 1 produced a crystallizable block copolymer: first polymerization in heptane, rr=89%, T_m=123 °C; second polymerization in heptane/CB, rr=50%; total, M_n=119 000, M_w/M_n=1.29, T_m=119 °C, rr=69%.⁴⁹ Diblock copolymer was also synthesized by inducing a temperature jump in toluene: first polymerization at −20 °C, rr=90%, T_m=123 °C, ΔH=69 Jg⁻¹; second polymerization at 25 °C , rr=60%; total, M_n=101 000, M_w/M_n=1.32, T_m=121 °C, ΔH=30 Jg⁻¹, rr=75%.⁵⁰ Diblock and triblock copolymers were synthesized via propene pressure variation, the GPC curves of which are shown in Figure 13.⁵¹ The dependence of stereospecificity on propene pressure was reported by Shmulinson and Eisen⁵² using a Ti benzaminate complex activated by MAO, where the stereospecificity was interconverted between being isospecific and non-stereospecific.

Figure 13

GPC curves of stereoblock PPs with 3–dMMAO via propene pressure variation.⁵¹

Polymerization of norbornene

Norbornene can be polymerized via three different mechanisms: a cationic mechanism, ring-opening metathesis and coordination insertion (vinyl-type addition). Among these mechanisms, the vinyl-type addition produces polynorbornene that possesses special properties, such as high thermal stability, high glass transition temperature (T_g), high transparency and low birefringence, because of its constrained saturated carbon skeleton. Such polymeric materials have created a higher demand on optical plastics in data storage and microelectronics technologies.

Coordination-insertion polymerization of norbornene was studied with TiCl₄/R₃Al in the early 1960s, and a number of zirconocenes have been studied since the 1980s.^{53, 54} All of these catalysts, except for (η⁵-C₅H₅)Ti(OBz)₃–MAO, have shown very low catalytic activity, and the products produced are insoluble in common organic solvents.

We have found that ansa-dimethylsilylene(fluorenylamido)dimethyltitanium 1-based catalysts, which conduct living polymerization of 1-alkene, are also excellent catalysts for norbornene polymerization when combined with a suitable cocatalyst.^{55, 56} Table 14 shows the results of norbornene polymerization with 1 and CGC-Ti activated by dMAO in toluene. Complex 1 showed good activity above 20 °C to give polynorbornene with high molecular weights and narrow MWDs via coordination insertion. In contrast, CGC-Ti showed very poor activity under the same conditions to give a polymer that was insoluble in NMR (tetrachloroethane-d₂) and GPC (o-DCB) solvents, even at elevated temperature.

Table 14 Polymerization of norbornene with 1 and CGC-Ti using dMAO as cocatalyst at different temperatures^a

Table 15 shows the effects of using a cocatalyst on norbornene polymerization with 1 in toluene at 20 °C. Ph₃CB(C₆F₅)₄ combined with Oct₃Al showed the highest activity (∼5000 kg-polymer·mol-Ti⁻¹ h⁻¹) among the cocatalysts employed. The conversion reached 95% within 2 min. MMAO and dMAO showed good activities (∼1000 kg-polymer·mol-Ti⁻¹ h⁻¹), but MAO showed very poor activity because of Me₃Al in MAO. All of the produced polymers possessed narrow MWDs, and the N/Ti values were below 1. It should be noted that MMAO gave the narrowest MWD of 1.07, which was most likely because of the good solubility of MMAO, indicating that dMMAO is not necessary for the living polymerization of norbornene with 1. Chain transfer to R₃Al may be suppressed at the hindered propagation chain end during the norbornene polymerization process.

Table 15 Polymerization of norbornene with 1 using different cocatalysts^a

The effects of the amount of MMAO on norbornene polymerization were investigated using 1 (Figure 14). Both the polymer yield and the M_n value linearly increased with the Al/Ti ratios and were accompanied by a narrowing of the MWD, which indicates that a large excess of MMAO is necessary to achieve a high propagation rate. The same phenomenon was observed in propene polymerization with 3–dMMAO as described above. The narrow MWD of the produced polynorbornene, as compared with PP, may be ascribed to the fast initiation during norbornene polymerization.

Figure 14

Effects of Al/Ti ratio on norbornene polymerization with 1–MMAO. Polymerization conditions: Ti=20 μmol, dMMAO=8.0 mmol, solvent=30 ml, propene=1.26 g, temperature=0 °C, 1 h. Post-polymerization conditions: propene =1.26 g, 1 h.⁵⁶

The polynorbornenes obtained were amorphous, soluble in halogenated solvents and stable at temperatures up to 420 °C. A polynorbornene film was prepared by casting the polymer solution in CB. The transmittance of the film (thickness=54 μm) was 93% or higher in the range from 350 nm to 700 nm (Figure 15).

Figure 15

Transmittance of polynorbornene film (Run 52) in ultraviolet–visible region (350–700 nm).⁵⁶

Copolymerization of norbornene with 1-alkene

Cycloolefin polymers and cyclic olefin copolymers, which consist of a rigid acyclic polymer backbone, are attractive materials because of their good heat and chemical resistance as well as their low dielectric constants, non-hygroscopicity and high transparency. Cycloolefin polymer is synthesized via ring-opening metathesis polymerization of norbornene derivatives followed by hydrogenation of C=C bonds along the main chain, whereas cyclic olefin copolymer is synthesized via the copolymerization of ethene and cycloolefins such as norbornene (Scheme 8).

Scheme 8

Synthetic routes of cycloolefin polymer (COP) and cyclic olefin copolymer (COC).

One of the advantages of cyclic olefin copolymers is that their polymer properties can be easily controlled by the comonomer content and comonomer sequence. The development of early transition metal single-site catalysts enables us to synthesize poly(norbornene−alt−ethene) as well as poly(norbornene−ran−ethene), the T_g value of which can be controlled over a wide range.⁵⁷ A number of late transition metal-based catalysts have been reported to produce polynorbornene via coordination insertion with good activity, but they do not produce a copolymer with ethene because of the facile β-hydrogen transfer after ethene insertion.

The other advantage of cyclic olefin copolymers is that their polymer properties can also be modified by the comonomer employed. However, the number of catalysts that copolymerize norbornene with propene or higher 1-alkene is very limited, and their activity and/or the molecular weight of the produced polymers is not sufficient.^{58, 59, 60, 61, 62, 63}

Ansa-dimethylsilylene(fluorenyl)(amido)dimethyltitanium-based catalysts were found to be active not only for 1-alkene polymerization but also for norbornene polymerization. We therefore applied these catalysts for copolymerization. We first conducted ethene–norbornene copolymerization with 1 and CGC-Ti activated by dMAO in toluene at 40 °C (Table 16).⁶⁴ Both complexes gave copolymers with high molecular weights and narrow MWDs. The activity of CGC-Ti was, however, one order of magnitude lower than that of 1, and CGC-Ti required a high norbornene/ethene ratio in feed to incorporate norbornene in the copolymers compared with 1.

Table 16 Copolymerization of ethene and norbornene with 1 and CGC-Ti activated by dMAO^a

The ¹³C NMR spectra of the copolymers containing 44 mol% of norbornene are shown in Figure 16. The spectrum of the copolymer obtained with CGC-Ti shows only eight major resonances, which can be assigned to the alternating and the isolated norbornene sequences, as reported earlier by McKnight and Waymouth⁶⁵ for a series of ansa-dimethylsilylene(cyclopentadienylamido)dichlotitanium derivatives activated by MAO. The copolymer produced with 1 showed a completely different spectrum, indicating the presence of a norbornene dyad and triad.

Figure 16

¹³C NMR spectra of ethene–norbornene copolymers obtained with (a) CGC-Ti and (b) 1 activated by dMAO. T, norbornene triad; RD, norbornene racemo dyad; MD, norbornene meso dyad.⁶⁴

Table 16 also shows the norbornene contents and the T_g values of the produced polymers determined by ¹³C NMR and differential scanning calorimetry. The copolymers obtained by 1 showed higher T_g values than that obtained by CGC-Ti because of the norbornene–norbornene sequences as well as the high norbornene contents. The monomer reactivity ratios of the ethene–norbornene copolymerization with 1–dMAO at 40 °C were determined to be r_E=5.6 and r_N,=0.08 by the Fineman–Ross method. The product of the reactivity ratios, r_Er_N=0.45, indicates the random nature of the copolymerization with 1–dMAO. When the copolymerization was conducted at 0 °C, the MWD of the produced copolymer became narrow, 1.16, although the activity was significantly reduced (Run 61).⁶⁶

The 1–dMAO system also enabled the copolymerization of norbornene with propene to give high molecular weight copolymers with narrow MWDs (Table 17).⁶⁷ A typical ¹³C NMR spectrum of the copolymer is illustrated in Figure 17, in which that of PP and polynorbornene obtained with the same catalyst are also shown as a reference. In the spectrum of the copolymer, the resonances assignable to the propene–norbornene sequence appeared. The comonomer contents were calculated from the relative intensity of the methyl groups with respect to the total carbons. The norbornene content was varied according to the norbornene/propene ratio in feed from 17 mol% to 71 mol%. Compared with the ethene copolymerization, low norbornene/comonomer ratio was sufficient to obtain a copolymer with a high norbornene content.

Table 17 Copolymerization of propene and norbornene with 1–dMAO^a

Figure 17

¹³C NMR spectra of (a) PP, (b) polynorbornene, (c) propene–norbornene copolymer and (d) DEPT 135 of the copolymer obtained with 1–dMAO.⁶⁷

The T_g of the copolymer is plotted against the norbornene content of the copolymer in Figure 18, which gives a straight line. The results indicate that the T_g value of the copolymer should be freely controlled in the range of the T_g values of PP and polynorbornene.

Figure 18

Plot of T_g versus norbornene content in the propene–norbornene copolymers.⁶⁷

The alkyl substituent on the fluorenyl group of 1 affected not only the activity but also the stereospecificity during propene polymerization, as described in the previous section. We therefore conducted propene–norbornene copolymerization with 1–4 activated by dMMAO (Table 18).⁶⁸ All of the complexes showed high activity during the copolymerization to produce copolymers with high molecular weights of 150 000–230 000 and narrow MWDs of 1.1–1.3 in a few minutes. In propene polymerization, 4 gave polymers with broad MWDs. The narrow MWD of the copolymerization can partly be ascribed to the fast initiation reaction with norbornene. The dependence of the activity on the titanium complex was the same as that for propene polymerization: 4≫3>2>1. All of the catalytic systems gave amorphous copolymers with high norbornene contents (>70 mol%) and thus high T_g values (>250 °C).

Table 18 Copolymerization of propene and norbornene with 1–, 2–, 3– and 4–dMMAO^a

The copolymerization activity of norbornene with propene was too high to maintain the norbornene conversion in low level, especially in the 4–dMMAO system, which prevented the precise evaluation of the monomer reactivity ratios during copolymerization. In addition, there is no report on living copolymerization of norbornene with higher 1-alkene. We therefore conducted copolymerization of norbornene and 1-octene with 1–4 activated by dMMAO at various comonomer feed ratios. Some of the results are shown in Table 19.

Table 19 Copolymerization of norbornene and 1-octene with 1–, 2–, 3– and 4–dMMAO^a

Although the order of the activity of the titanium complexes was the same, the activity for the 1-octene copolymerization was one order of magnitude lower than that for propene copolymerization. It should be noted that 4–dMMAO still exhibited a high activity of 1660 kg-polymer mol-Ti⁻¹ h⁻¹. All of the catalytic systems gave high molecular weight polymers with narrow MWDs, although 3 and 4 gave slightly broader MWDs. The same tendency was observed in the propene copolymerization (Runs 71 and 72).

To confirm the living nature of the copolymerization with 4–dMMAO, we conducted post-copolymerization, in which the same amount of norbornene–1-octene mixture was sequentially added after the first step copolymerization had been completed. The polymerization results and the GPC curves of the copolymers are shown in Figure 19, which implies that the copolymerization of norbornene and 1-octene with 4–dMMAO proceeded in a living manner. The MWD in the post-copolymerization was narrower than that in the copolymerization (Runs 79 and 80 in Table 19), indicating that the broad MWD of the latter stems from the physical inhomogeneity caused by the high activity of the catalyst, namely the temperature distribution in the reactor because of the heat of polymerization and/or the ineffective stirring due to the produced polymers.

Figure 19

GPC curves of poly(norbornene-co-1-octene) obtained in post-polymerization with 4–dMMAO at 20 °C.⁶⁸

The ¹³C NMR spectra of the copolymers containing ∼75 mol% of norbornene obtained with each catalytic system are shown in Figure 20, which indicates the production of poly(norbornene–co–1-octene) in all of the catalytic systems. Although the detailed assignments of the resonances are not clear, the shapes of the spectra are slightly dependent on the titanium complex used. As each Ti complex possesses different stereospecificity in 1-alkene polymerization, the differences in the ¹³C NMR spectra of the copolymers may be because of the differences in the stereoregularity of the copolymers.

Figure 20

¹³C NMR spectra of poly(norbornene-co-1-octene) obtained with 1 (a) (75 mol%), 3 (b) (73 mol%), 2 (c) (74 mol%) and 4 (d) (75 mol%) with almost the same norbornene content.⁶⁸

The T_g values of the copolymers, as determined by differential scanning calorimetry, are plotted against the norbornene content in Figure 21. The T_g value showed a linear relationship with the norbornene content in all of the catalytic systems, indicating the formation of uniform random copolymers regardless of the complex used. The slope of the straight line is, however, dependent on the titanium complex. The results also show that the microstructures of the copolymers are dependent on the complex used. Tritto and colleagues⁶⁹ reported that the stereoregularity of a norbornene sequence affected the T_g value of poly(norbornene-co-ethylene) with high norbornene contents.

Figure 21

Plots of T_g value against norbornene content in poly(norbornene–co–1-octene) obtained with 1 (a), 2 (b), 3 (c) and 4 (d).⁶⁸

The monomer reactivity ratios for each catalytic system were determined by the Fineman–Ross method, although the norbornene conversions were not maintained at a sufficiently low level, especially in the 4–dMMAO system because of its high activity for norbornene. Table 20 shows the monomer reactivity ratios (r_N=k_NN/k_NO and r_O=k_OO/k_ON) thus obtained. These values indicate a preference for the insertion of norbornene, regardless of the last inserted monomer unit. The product of the reactivity ratios (r_Nr_O=0.97) obtained with 3 indicates a tendency for the formation of a random copolymer, whereas the products of the reactivity ratios (r_Nr_O=2.5∼3.5) obtained with 1, 2 and 4 imply a preference for the formation of a homonorbornene sequence in the copolymer.

Table 20 Monomer reactivity ratios of 1-octene-NB copolymerization with 1–4 activated by dMMAO^a

We achieved highly active living copolymerization of norbornene and 1-alkene with the tetraalkyl substituted fluorenyl complex 4 activated by dMMAO. If one does not require livingness for copolymerization, the most simple complex 1 activated by Ph₃CB(C₆F₅)₄/Oct₃Al possesses a sufficient activity of 10³ kg-polymer mol-Ti⁻¹ h⁻¹ for the copolymerization of norbornene with 1-alkenes to give random copolymers with M_n values of 70 000∼90 000 and M_w/M_n values of 1.4∼1.7. The T_g value of the copolymer was controlled over a wide range by the norbornene content, irrespective of the 1-alkene used (Figure 22).⁷⁰

Figure 22

Plots of T_g versus norbornene content of poly(1-alkene–co–norbornene) obtained with 1–Ph₃CB(C₆F₅)₄/Oct₃Al.⁷⁰

The molecular weights of the copolymers were sufficiently high to prepare self-standing films by the solution-casting method. The transparency of the films was found to improve according to the length of 1-alkene and reached ∼90% with poly(1-decene−co−norbornene) in the ultraviolet–visible region (Figure 23).

Figure 23

Transparency of poly(norbornene−ran−1-alkene) films: norbornene content, 92–93 mol%; film thickness, 120 μm. (a) 1-hexene, (b) 1-octene and (c) 1-decene.⁷⁰

Synthesis of olefin block copolymers

The most important role of living polymerization is to synthesize block copolymers. It is therefore strongly encouraged to develop a living polymerization catalyst that is active for various monomers. the ansa-dimethylsilylene(fluorenylamido)dimethyltitanium derivatives 1, 2 and 3 activated by dMAO or dMMAO were found to enable the living polymerization of propene and higher 1-alkene, where the rate of polymerization and the syndiotacticity of the produced polymers can be controlled by the Ti complex and/or the polymerization conditions, such as temperature, solvent and monomer concentration. The catalytic system has been also proven to conduct homopolymerization and copolymerization of norbornene with propene or 1-alkene in a living manner. Thus, we can synthesize any kind of block copolymer composed of 1-alkene and norbornene with this catalytic system.

As an example of the application of these living systems, we synthesized syndiotactic PP−block−poly(propene−ran−norbornene) with 3–dMMAO according to the procedure shown in Scheme 9, and the results are summarized in Table 21.⁴⁹ The PP obtained in the first step showed an M_n value of 108 000 with an M_w/M_n of 1.36 and a melting point of 135 °C. After the copolymerization, the yields increased, as did M_n values (208 000–226 000); however, the M_w/M_n values decreased (1.21–1.32) regardless of the amount of norbornene added. The block copolymers showed T_m (133–135 °C) and T_g (93–311 °C) values that corresponded to that of the crystalline syndiotactic PP sequence and amorphous poly(propene−ran−norbornene) sequence, respectively, and the T_g values were controlled by the amount of norbornene added. These results suggest the formation of the expected block copolymers and the phase separation of the copolymers.

Scheme 9

Synthesis of block copolymer with 3–dMMAO.

Table 21 Sythesis of syn−PP–block-poly(propene−ran−norbornene) with 3–dMMAO^a

Although living polymerization is very useful for the synthesis of monodisperse polymers and tailor-made block copolymers, one catalyst molecule (initiator) is necessary for every polymer chain. The catalytic synthesis of monodisperse polyolefins has been achieved by the combination of a living polymerization catalyst and a large excess of main-group metal alkyls, such as R₂Zn and R₃Al, where the propagating polymer chain on the transition metal exchanges with the metal alkyls reversibly much faster than the propagation rate (Scheme 10). This kind of polymerization is referred to as coordinative chain transfer polymerization and the metal alkyls are referred to as chain-shuttling reagents.^{71, 72} A suitable combination of the catalysts, which differ in copolymerization ability, produced multiblock copolymers via the copolymerization of ethene with 1-octene in the presence of Et₂Zn as a chain-shuttling reagent. The synthesis of a multiblock copolymer through this mechanism is called chain-shuttling polymerization.⁷³ Ethene-based diblock copolymers were also synthesized by switching the monomer composition during coordinative chain transfer polymerization.⁷⁴

Scheme 10

Coordinative chain transfer polymerization.

The combination of a living polymerization catalyst and a suitable chain transfer reagent should give monodisperse polymers if the propagation rate is much faster than the chain transfer rate and chain transfer occurs only after all the monomer is consumed (Scheme 11). Mitani et al.⁷⁵ catalytically obtained monodisperse polyethene with an phenoxyimine-based (so-called FI) catalyst in the presence of a large excess of Et₂Zn by the sequential addition of the same amount of ethene. We proved this concept in propene polymerization with 3–dMMAO by controlling the amount of ⁱBu₃Al in dMMAO.⁷⁶ The concept was applied in the catalytic synthesis of block copolymer as shown in Table 22.⁷⁷ Block copolymerization of norbornene and propene was conducted with 3 activated by dMMAO(1.8) containing 1.8 mol% of ⁱBu₃Al, in which propene was introduced before norbornene had been consumed (Run 85). Further sequential addition of norbornene and propene in the same manner did not change the M_n value or the narrow MWD (Run 86), indicating the catalytic synthesis of polynorbornene−block−poly(propene−ran−norbornene)–block–PP. We can conclude that the catalytic synthesis of block copolymers is possible by the sequential addition of monomers in a highly active living polymerization catalyst combined with a suitable kind and amount of chain transfer reagent (Scheme 12).

Scheme 11

Catalytic synthesis of monodisperse polymers with highly active living polymerization catalyst.

Table 22 Block copolymerization of norbornene and propene with 3 activated by dMMAO(1.8)^a

Scheme 12

Catalytic synthesis of monodisperse block polymers with highly active living polymerization catalyst.

Conclusions

1 and its derivatives activated by dMAO or dMMAO were found to conduct the living polymerization of propene at 0 °C and 25 °C. Solvent effects were investigated with 1 by using dMMAO, which is readily soluble in heptane. The syndiospecificity was improved by decreasing the polarity of the solvent from CB to heptane, accompanied by decrease in activity. The syndiospecificity and activity were also shown to be dependent on the alkyl substituent of the fluorenyl ligand. The activity monotonously increased with the number of introduced alkyl groups, whereas the highest syndiospecificity was obtained with the 3,6-ⁱBu₂-substituted complex 3. Thus, highly active and syndiospecific living polymerization of propene was achieved with 3–dMMAO in heptane. The syndiospecificity of the catalysts also depended on the polymerization conditions, such as temperature or monomer concentration, a fact that was applied to the synthesis of stereoblock PPs. The catalytic system also conducted homopolymerization and copolymerization of norbornene and higher 1-alkene in a living manner with high activity in toluene. The 3–dMMAO system was also proven to be catalytic in the synthesis of monodisperse block copolymers in the presence of a suitable amount of ⁱBu₃Al via selective chain transfer to ⁱBu₃Al after monomer consumption. These characteristics of the ansa-dimethysilylene(fluorenyl)(amido)dimethyltitanium-based catalysts can be applied to develop novel cycloolefin copolymers.