• Keine Ergebnisse gefunden

Polymerizations with late transition metal complexes in aqueous system

1. Introduction

1.1. Olefin polymerization catalyzed by late transition metal complexes

1.1.2. Polymerizations with late transition metal complexes in aqueous system

The functional group tolerance of late transition metal catalysts allows for polymerization in polar reaction media, as illustrated by the SHOP. In polymer chemistry, the most prominent polar medium is water. It is employed on a large scale in emulsion polymerization, affording polymer dispersions. Water as the dispersing medium offers a unique combination of advantages:22

- A high heat capacity, which facilitates an effective transfer of the heat of polymerization.

- A high polarity that results in a perceptibly different miscibility with many monomers and polymers compared to organic solvents.

- The viscosity of an aqueous latex is essentially lower than of a solution of the polymer in an organic solvent. Thus, high polymer yields can be obtained during emulsion polymerization in a given amount of reaction volume.

- Dispersions of hydrophobic polymer particles can be effectively stabilized by surfactants in water towards aggregation.

- Water is non-flammable and nontoxic.

- Water is cheap.

- Water is particularly environmentally friendly.

About 7% of the annual plastics production is obtained as aqueous polymer dispersions.

Until today, aqueous polymer dispersions for coatings and paints are produced industrially by free radical routes exclusively.23 By contrast to radical polymerization, catalytic polymerization allows for a control of microstructure, and correspondingly materials properties, over a wide range. Therefore, catalytic polymerizations in emulsion have been studied recently.24 , 25 By polymerization of ethylene with Ni (II) salicylaldiminato complexes, high molecular weight polyethylene dispersions can be prepared.24b By appropriately tailoring the catalyst via remote substituents of N-terphenyl moieties in the salicylaldimine (Figure 1-2), linear, semicrystalline or highly branched, amorphous polymer can be obtained. The latter goes at the expense of

molecular weight, however.24c

R1

O Ni

N Me

R3 R4 L

R1 R1 R1

R2 R2

Figure 1-2. Salicylaldiminato complexes.

Recently, particles with sizes <10 nm were obtained using water-soluble salicyaldiminato Ni(II) complexes (Figure 1-3).24e, 29 Particles sizes in this range are difficult to access by other polymerization mechanisms, e.g. industrial free radical emulsion polymerization affords particles with sizes between 50 nm and 1 µm.26 Otherwise, very small polymer particles of 10-30 nm size have been prepared by means of catalyst microemulsions.27 Microemulsion as an approach is, indeed, not per se general since microemulsions often only form under specific conditions that require large quantities of alcohols, or other organic solvents, and surfactants. An aqueous solution of a hydrophilic catalyst clearly represents the highest possible initial degree of dispersion of the catalyst precursor, and is, therefore, of great interest in view of the synthesis of small particles. Dispersions of nanocrystals are of interest for applications such as, for example, delivery of poorly soluble drugs, controlled release of active molecules, and homogeneous incorporation of functional molecules such as dyes into solid materials or as carriers in aqueous multiphase catalysis.28 Also, semicrystalline particles in this size regime can contribute to fundamental understanding of crystallization in confinement,29 crystallinity can be a source of anisotropy, and such particles can serve as crystalline mesoscopic building blocks for ultrathin films.30,31

Figure 1-3. Cryo-TEM micrograph of nanoscale polyethylene crystals in aqueous dispersion.

(from ref. 29)

1.2. Living polymerizations with transition metal complexes 1.2.1. General considerations

Living catalytic polymerizations are a topic of current interest, and the development has recently been reviewed comprehensively.56 In contrast to late transition metal complexes, catalysts based on early transition metals are less prone to β-hydride eliminations. Thus, living polymerizations can be carried out with various catalyst systems (albeit some late transition metal catalysts also polymerize in a living fashion). Living polymerizations32 are desirable due to their capability to enchain monomer units without termination. Precise molecular weight control as well as the synthesis of a wide array of polymer architectures are possible.33 Living methods also allow the synthesis of end-functional polymers if special initiation and/or quenching methods are employed. Nevertheless, living polymerizations have the limitation that each catalyst only forms one chain, in contrast to common alkene polymerization catalysts that can produce thousands of chains each as a result of chain transfer. In terms of material properties, living polymerization methods allow for the synthesis of new materials from a basic

set of available monomers.

Indeed, olefin insertion catalysts have always been inferior to other chain-growth polymerizations in one respect. While tremendous advances in living/controlled polymerization have been reported by using anionic,34 cationic35 and radical-based36 polymerization, until recently there existed a comparative lack of living olefin polymerization systems. A significant number of advances have been discovered not until the last half decade. Seven generally accepted criteria for a living polymerization are:37

1) Polymerization proceeds to complete monomer conversion, and chain growth continues upon further monomer addition.

2) Number average molecular weight (Mn) of the polymer increases linearly as a function of conversion.

3) The number of active centers remains constant for the duration of the polymerization.

4) Molecular weight can be precisely controlled through stoichiometry.

5) Polymers display narrow molecular weight distributions, described quantitatively by the ratio of the weight average molecular weight to the number average molecular weight (Mw/Mn = 1).

6) Block copolymers can be prepared by sequential monomer addition.

7) End-functionalized polymers can be synthesized.

Only few polymerization systems, whether ionic-, radical-, or metal-mediated, that are claimed to proceed by a living mechanism have been shown to meet all of these criteria.

Common features of alkene polymerization catalyst systems are chain transfer and elimination reactions that terminate the growth of a polymer chain and result in the initiation of a new polymer chain by the catalyst. For example, in metallocene catalysts, consecutive alkene insertion into the metal carbon bond connecting the catalyst and polymer chain38 proceeds until β-hydrogen elimination and/or transfer to monomer occurs.39 When alkylaluminum cocatalysts

monomer addition methods for block copolymer synthesis futile. Several strategies have been devised to decrease the rate of chain termination relative to that of propagation such that living systems can be formed. The first consideration in many cases is simply lowering the polymerization temperature of an ordinary nonliving catalyst system to achieve living or at least controlled behavior. Since β-hydrogen elimination process is unimolecular while propagation can be bimolecular, a lowering in temperature can more adversely affects elimination processes relative to enchainment. However, particularly for semicrystalline polymers, precipitation of polymers from solution at low temperatures can hinder the controlled nature of a living polymerization. Therefore, a second strategy for discovering living systems is to design new catalysts through empirical modification and/or rational methods.40 By creating species that are unreactive for common termination reactions even at ambient temperature, living catalysts have been devised. A specific consideration is to eliminate the use of alkyl aluminum cocatalysts that give the potential for chain-transfer reactions. In this regard, the development of weakly coordinating anions has made significant advances in living olefin polymerization possible.41

1.2.2. Living propylene polymerization

syndiotactic polypropylene sPP

isotactic polypropylene iPP

atactic polypropylene aPP

regioirregular polypropylene rir-PP

Figure 1-4. Microstructures of propylene homopolymers.

By comparison to ethylene, the prochirality of this monomer gives rise to significantly greater complexity in the resultant polymers by introducing the variables of stereo- and regioregularity. The tacticity of the polymer is intimately related to its bulk properties, with

atactic polypropylene (aPP; Figure 1-4) being an amorphous material with limited industrial uses (e.g. adhesives, sealants, and caulks) and syndiotactic polypropylene (sPP) and isotactic polypropylene (iPP) being semicrystalline materials with relatively high Tm values of 150 and 165 °C respectively. Only a very few syndiospecific propylene polymerization catalysts have been discovered to date; this fact coupled with its lower crystallinity have limited the commercial impact of sPP. On the other hand, numerous catalysts, both heterogeneous and homogeneous, are capable of isospecific propylene polymerization. iPP possesses highly desirable mechanical properties (durability, chemical resistance and stiffness).

1.2.2.1. Vanadium acetylacetonoate catalysts

Vanadium compounds have been found to promote living olefin polymerization. In 1960s, Natta and coworkers discovered that vanadium tetrachloride with Et2AlCl at -78 °C in the presence of propylene furnished syndio-enriched polypropylene.42 More than a decade later, Doi reported [V(acac)3]/Et2AlCl as a living catalyst at -65 °C for the synthesis of syndio-enriched PP ([r] 0.81). At temperatures even slightly above -65 °C, the living behavior was greatly diminished as evidenced by an increase in molecular weight distribution (Mw/Mn = 1.37–

1.45 at -48 °C). Moreover, the activity was extremely low, only 4 % of the catalyst was active.

By the addition of anisole or by replacing the acetylacetonoate ligands with 2-methyl-1,3-butanedionato ligands, the number of active centers could be increased considerably.

Although the non-group 4 early metal complexes exhibit some characteristics of a living polymerization, a great part of the aforementioned criterion is still not fulfilled. Additionally, the polymerizations have to be carried out at very low temperature so that the productivity was very low. High molecular weight polymers with narrow molecular weight distribution could not be obtained in many cases.

1.2.2.2. Metallocene-based catalysts

Some of the metallocene-based catalysts which promote living propylene polymerization are given in Figure 1-5.

Figure 1-5. Metallocene catalyst precursors for living propylene polymerization.

Bochmann and co-workers reported that 6 activated with B(C6F5)3 at -20 °C produces atactic, high molecular weight PP (Mn = 789 000 g/mol, Mw/Mn = 1.4) that exhibits elastomeric properties.43 The polymerization showed a linear increase in molecular weight with time. Living propylene polymerization has also been observed for bis- Cp type Group IV metallocenes at low temperatures. Fukui and co-workers have reported that [Cp2ZrMe2] (7) activated with B(C6F5)3 at -78 °C produces PP with Mw/Mn = 1.15 (Mn = 9 400–27 300 g/mol).44 The polymerization shows a linear increase in Mn with time. While at -50 °C, the molecular weight distribution broadens (Mw/Mn = 1.55) for the PP produced by 7/B(C6F5)3, the hafnium analogue (8) produces polymer with Mw/Mn = 1.05–1.08 (Mn = 6 000–34 300 g/mol). Under similar conditions, a titanium complex bearing a linked monocyclopentadienyl-amido ligand, 9/B(C6F5)3 produces

low molecular weight (Mn = 9 800–19 900 g/mol) syndio-enriched PP ([rr] 0.49) with a narrow molecular weight distribution (Mw/Mn = 1.15–1.17).45 Also, the polymerization shows a linear increase in Mn with polymer yield. Complex 10, when activated with “dried” MMAO (dMAO) (free of trimethylaluminium) at 0 °C, produced polymer with even higher tacticity ([rr] 0.93) and the molecular weight distribution was 1.45.46 Employing an indenyl-based ligand 11 and dMAO at 0 °C in toluene affords quasi-living propylene polymerization.47 The PP produced is iso-enriched ([mm] 0.40), has Mw/Mn = 1.38 (Mn = 48 900 g/mol) and shows a linear increase in Mn with time.

In summary, the living character of the metallocene-based catalysts is only retained at low temperature. In general, the molecular weight is relatively low or a broader molecular weight distribution is observed.

1.2.2.3. Bis(phenoxyimine)titanium catalysts

In 1999, Fujita and co-workers reported on a class of group IV complexes bearing chelating phenoxyimine ligands (Figure 1-6). When activated with MAO, these complexes showed extremely high activity for ethylene polymerization.48

In contrast to the non-fluorinated analogous 13,49 the incorporation of fluorinated N-aryl moieties into the bis(phenoxyimine) ligand framework could provide catalyst precursors for the syndiotactic and living polymerization of propylene. When activated with MAO at 0 °C, 14 (Figure 1-6) produced highly syndiotactic PP ([rrrr] 0.96).50 The polymerization exhibited a linear increase in Mn with PP yield and while this trend deviated somewhat at higher molecular weights, polydispersities remained low (Mw/Mn = 1.11) for Mn up to 100 000 g/mol. Living behavior was further exemplified by a good correlation between measured and theoretical Mn

values and by the ability to make block copolymers (see 1.2.4). Remarkably, living behavior was also exhibited at 20 °C (Mn = 100 000 g/mol, Mw/Mn = 1.13).

Fujita and co-workers reported that 15/MAO was also living for propylene polymerization at room temperature, producing polymer with Mn = 28 500–108 000 g/mol and Mw/Mn = 1.07–1.14.51 While low molecular weight oligomers (Mn = 2000 g/mol, Mw/Mn = 1.08) exhibited 98% rr triads, higher molecular weight samples exhibited lower tacticites ([rr] 87%).

A catalyst using 15 activated with a supported cocatalyst has also been prepared and screened for propylene polymerization. Polypropylene formed using 15/MgCl2/i-BunAl(OR)3-n had narrow PDIs (Mw/Mn = 1.09–1.17, Mn = 53 000–132 000 g/mol) and the polymerization exhibited a linear increase in Mn with reaction time.52

In the last few years, many new bis(phenoxyimine)-based titanium complexes have been synthesized and screened for propylene polymerization activity. Attempts to improve upon early complexes led to new catalysts with both independent and simultaneous variation on the N-aryl and phenoxide moieties. Of note, a complex with a trimethylsilyl group ortho to the oxygen of the phenoxide moiety, 16 (Figure 1-7), has been shown to produce sPP with very high melting temperatures (Tm up to 156 °C).53 Polymerizations with 16/MAO conducted at 0 and 25 °C produced polymer with very narrow molecular weight distributions (Mw/Mn as low as 1.05, Mn = 24 700–47 000 g/mol) while raising the reaction temperature to 50 °C increased the PDI (Mw/Mn

= 1.18–1.23, Mn = 25 500–35 100 g/mol).

Figure 1-7. Bis(phenoxyimine) titanium complexes with various phenoxide substituents.

Changing the aforementioned ortho position to a larger triethylsilyl group (17) did not significantly affect the peak melting temperature or molecular weight distribution of the PP formed relative to 16.54 However, employment of a methyl (18) or isopropyl (19) group in the ortho position resulted in a substantial loss of stereocontrol with 18/MAO and 19/MAO both producing amorphous PP which exhibited fairly narrow polydispersities (Mw/Mn = 1.2).

Complex 20 which has a tert-butyl group at the ortho position of the phenoxide moiety and a methyl group para to that position, has also been synthesized. At 25 °C, 20/MAO behaved similarly to the related catalyst 15/MAO (Figure 1-6) yielding sPP with a peak melting temperature of 140 °C and a narrow molecular weight distribution (Mw/Mn = 1.11, Mn = 16 500 g/mol).

Studies on the effect of the fluorination pattern of the N-aryl ring have been conducted and it has been revealed that complexes bearing the 2,4-di-tert-butyl phenoxide moiety require at least one ortho fluorine on the N-aryl ring to exhibit living propylene polymerization behavior.55 As the amount of fluorination of the N-aryl moiety is decreased from the perfluoro complex 14 (Figure 1-6) to the monofluoro complex 21 (Figure 1-8), activities for propylene polymerization with the corresponding MAO-activated catalysts decreased while the

Figure 1-8. Bis(phenoxyimine) titanium complexes with varying N-aryl fluorine substitution patterns.

The PP produced by 23/MAO is highly tactic ([rrrr] 0.95), while 22/MAO produces polymer that is somewhat less stereoregular ([rrrr] 0.83). Notably, tacticity drops off significantly upon moving to the monofluorinated complex 21 which affords polypropylene with [rrrr] 0.52. Lastly, increasing the level of fluorination beyond that of 14 by introducing a trifluromethyl group at the para-position of the N-aryl moiety led to an increase in activity. It was reported that 25/MAO was 1.5 times more active than 14/MAO for propylene polymerization but was still able to provide highly syndiotactic polymer ([rrrr] 0.91) with a narrow molecular weight distribution (Mw/Mn = 1.13).56

Simultaneous changes to both the phenoxide and N-aryl moieties, relative to 14 (Figure 1-6) have also been made. For example, 26 (Figure 1-9) exhibits ortho fluorination on the N-aryl moiety and iodine substituents on the phenoxide moiety. When activated with MAO at 25 °C, 26 was reported to produce amorphous PP which exhibited a narrow molecular weight distribution (Mw/Mn = 1.17, Mn = 200 000 g/mol).57

Figure 1-9. Bis(phenoxyimine) titanium complexes.

Complexes 27-28 employ 3,5-difluorophenyl N-aryl groups and substituents smaller than tert-butyl in the ortho position of the phenoxide moiety. Both 27-28/MAO were shown to polymerize propylene at 0 °C to produce high molecular weight polymer (Mn up to 240 000 g/mol) that is amorphous ([rrrr] 0.48).58 For both of these catalysts, there was no evidence of β-H transfer and the polymer produced in each case exhibited a narrow molecular weight distribution (Mw/Mn = 1.13–1.16). This finding was remarkable at the time in that both complexes lack ortho fluorines on the N-aryl moiety.

One recent variation to the bis(phenoxyimine) complexes involves the coordination of two different phenoxyimine ligands to one titanium center. Using pooled combinatorial approach, several heteroligated complexes employing both a ‘‘living’’ and ‘‘non-living’’ ligand were independently synthesized and screened for propylene polymerization. While the molecular weight distribution of the resulting polymers formed from the mixed catalysts fell between those of the corresponding homoligated catalysts, the activity was significantly higher in some cases. For example, PP produced with 13/MAO (Figure 1-6) exhibited a broad PDI (Mw/Mn = 1.41) and a turnover frequency of 42 h-1 while 14/MAO exhibited a narrow PDI

turnover frequency of 760 h-1. Syndiotactic polymer ([rrrr] 0.91] was formed with this heteroligated catalyst.

Figure 1-10. Heteroligated bis(phenoxyimine) titanium complex.

1.2.2.4. Bis(phenoxyketimine)titanium catalysts

Bis(phenoxyimine) titanium complexes produce sPP through a chain-end control mechanism despite the fact that the catalyst precursors are C2-symmetric in the solid state and in solution. The reason for this phenomenon stems from a proposed ligand isomerization event that provides enantiomeric coordination sites. It had been proposed that placing a substituent at the imine carbon of the phenoxyimine ligand could prevent this isomerization and lead to the formation of iPP.59 Ketimine complexes 30-32 (Figure 1-11) were synthesized and while they were shown to polymerize ethylene in a living fashion upon activation, they proved to be sparingly active for propylene polymerization.59-60 It was reasoned that reduction of the size of the ortho substituent of the phenoxide moiety could provide a sterically less encumbered active site potentially enabling higher activities for propylene polymerization. Complexes 33-37 were synthesized and screened for propylene polymerization.59 Upon activation with MAO at 0 °C, each complex produced PP with a narrow molecular weight distribution (Mw/Mn = 1.12–1.17, Mn = 2 700–35 400 g/mol) and 35/MAO was shown to exhibit a linear increase in PP molecular

weight as a function of yield. The tacticities of the resulting polymers differed. Amorphous, moderately isotactic PP was produced by 33/MAO and 34/MAO ([mmmm] 0.45) while 35/MAO produced PP with higher tacticity ([mmmm] 0.53) and a peak melting temperature of 69.5 °C. This level of tacticity was increased ([mmmm] 0.61, Tm 96.4 °C) when the polymerization was run at 20 °C. Having R1= H, 36/MAO exhibited a significant loss of stereocontrol ([mmmm] 0.08). The aldimine analogue of 35 (37) in which R3= H, furnishes atactic PP with Mn = 123 100 g/mol and Mw/Mn = 1.13.

Figure 1-11. Bis(phenoxyimine) titanium complexes.

1.2.3. Living ethylene polymerization

1.2.3.1. Non-group 4 metal polymerization catalysts

Apart from the group 4 metal complexes, examples with group 3, group 5 transition metals and also with late transition metals have been reported relating to the living ethylene

1.2.3.1.1. Group 3 metal polymerization catalysts (Y, lanthanides)

The potential of f-block metal complexes to serve as living olefin polymerization catalysts was demonstrated by Marks and co-workers in 1985.61 The dimeric bis(Cp*) hydride complexes of lanthanum, neodymium, and lutetium (38-40, Figure 1-12) were shown to polymerize ethylene with extremely high activities (up to 3040 g PE/mmol-1 min-1 atm-1). The PEs exhibited high molecular weights (Mn = 96 000-648 000 g/mol) with Mw/Mn = 1.37–4.46.

For the most part, the PDIs were lower than 2.0 (e.g. 61c exhibited Mw/Mn = 1.37–1.68), with higher values being due to inhomogeneities brought about by precipitation of the polymer, possible mass-transport effects and rate-limiting dissociation of the dimer. Further evidence in support of the living nature of ethylene polymerization catalyzed by 38-40 includes the observations that catalytic activity is maintained at room temperature for up to 2 weeks, Mn increases with increasing reaction time, and the number of polymer chains per metal center is consistently less than one which argues against chain-transfer followed by reinitiation.

Figure 1-12. Lanthanide and group 3 metal catalysts for living olefin polymerization.

Hessen and coworkers reported on the synthesis and ethylene polymerization activity of dialkyl(benzamidinate)yttrium complexes which displayed some characteristics of living behavior.62 When treated with [PhNMe2H][B(C6F5)4], 41 (Figure 1-12) catalyzed the formation of PE. The resultant polymers possessed narrow molecular weight distributions (Mw/Mn = 1.1–

1.2) and high molecular weights (Mw = 430 000–1 269 000 g/mol). It was shown that about 1.1

polymer chains per metal center were produced and this number was constant over the course of the 30 min. Taken together, these observations suggest that 41 [PhNMe2H][B(C6F5)4] is living for ethylene polymerization.

1.2.3.1.2. Group 5 metal polymerization catalysts (V, Nb, Ta)

Figure 1-13. Niobium and tantalum complexes for living ethylene polymerization.

Nomura and co-workers reported that arylimido(aryloxo)vanadium dichloride complexes activated with Et2AlCl exhibited characteristics of living ethylene polymerization.63 In particular, 42/Et2AlCl (Figure 1-13) furnished PE with a narrow molecular weight distribution (Mw/Mn = 1.42) and high molecular weight (Mn = 2 570 000 g/mol) at 0 °C. Examples of living ethylene polymerization by Group 5 metal complexes are not limited to those based on vanadium. In 1993, Nakamura and coworkers reported on the synthesis and ethylene polymerization behavior of cyclopentadienyl(η4-diene)tantalum complexes.64 Upon activation with MAO at temperatures of 20 °C or below, compounds 43-45 (Figure 1-13) furnished PEs with narrow molecular weight distributions (Mw/Mn = 1.4) with Mn = 8 600–42 900 g/mol. In a subsequent report, the analogous niobium complexes were shown to behave as living ethylene polymerization catalysts up to 20 °C.65 Compounds 46-49 (Figure 1-13) when activated with MAO in the presence of ethylene furnished polymers with narrow polydispersities (Mw/Mn = 1.05–1.30) and Mn = 5 100–

1.2.3.1.3. Chromium polymerization catalysts

Theopold and co-workers investigated ethylene polymerization catalyzed by chromium complexes bearing 2,4-pentane-N,N’-bis(aryl)ketiminato ((Ar)2nacnac) ligands as a potential model for ethylene polymerization by the heterogeneous Philips catalyst.66 When exposed to

Theopold and co-workers investigated ethylene polymerization catalyzed by chromium complexes bearing 2,4-pentane-N,N’-bis(aryl)ketiminato ((Ar)2nacnac) ligands as a potential model for ethylene polymerization by the heterogeneous Philips catalyst.66 When exposed to