Comparison of Ziegler-Natta and metallocene ethylene elastomer products.

Polyolefins represent the largest class of polymers used throughout the world today. Annual consumption has soared to over 100,000 Kt per year. After the introduction of metallocene catalysts in the mid-1980s, it was suggested that metallocenes would revolutionize the polyolefin industry, but in reality, after almost 20 years of opportunity, only about 5% of the world's polyolefin production is made using metallocene catalysts (ref. 1). Although metallocenes distinguish themselves in a number of important segments, they do not significantly differentiate themselves from conventional Ziegler-Natta catalysts in the high-density commodity products (figure 1).

One area where metallocenes have made a significant impact is in the low-density, ethylene elastomer (EE) markets which include plastomers (ethylene [alpha]-olefin copolymers with density less than 0.910 g/cc) and EP(D)M products. Within this segment, metallocene products have grown to represent about 42% of the global EE production since their introduction only ten years ago (ref. 2). Part of the reason for metallocene growth in this segment is their ability to incorporate comonomer readily without significant loss of activity or molecular weight. Metallocenes clearly differentiate themselves at low density and high comonomer content when compared to conventional Ziegler-Natta catalysts.

The purpose of this article is to compare and contrast conventional Ziegler-Natta catalysts with metallocene catalysts, and to examine the implications these differences may have on their polyolefin products. Focus areas where metallocene products are expected to differentiate themselves include branching, sequence distribution and stereoregularity. Examples are given to illustrate how these features can influence product properties.

Results and discussion

General catalyst comparisons--structure

Ziegler-Natta catalysts are mixtures of metal halides and aluminum alkyl halide co-catalysts. They can combine in a variety of ways resulting in coordination complexes with multiple active sites (figure 2a). These catalysts produce products with a broader distribution of chains than one would expect from a single active center, and the resulting polymers generally have a broader distribution of properties. However, having broad product properties can be useful in some applications, and often modifiers can be added to reduce the multi-sited nature of conventional Z-N catalysts. It is not knowing the structure that limits their utility in product design. Not knowing the catalyst structure, and not being able to modify or design specific structures, makes it difficult to understand what features in the catalyst result in specific polymer properties.

[FIGURE 2 OMITTED]

In contrast, bridged metallocenes have well defined chemical structures that can be isolated and determined by x-ray crystallography (figure 2b). These structures have a single active polymerization site with uniform ligand environment that makes chains with uniform architecture and with a narrow distribution of properties. Knowing the metallocene structure allows one to study the relationship between its geometry and the product architecture. This is a powerful tool in the design of polymer properties. By modifying the metallocene structure, it is possible to tailor its product properties to meet very specific application needs (ref. 5).

Temperature stability

Conventional Z-N catalysts have poor temperature stability due to the loosely bound complex between catalyst and co-catalyst. They often lose a significant portion of their activity at temperatures as low as 60[degrees]C.

Many types of metallocene based catalyst systems have good high temperature stability and often remain active at temperatures over 200[degrees]C. High temperature stability allows higher polymerization temperatures and broadens the potential composition range. For example, products with higher melt temperatures can be kept in solution longer at higher process temperatures, and thereby extend the range of possible compositions.

Comonomer incorporation

In general, Ziegler-Natta catalysis polymerize higher [alpha]-olefins poorly. Rate constants for the polymerization of higher [alpha]-olefin drop considerably after propylene, and although butene plastomers made from vanadium Z-N catalysis are commercially available, they have low activity and limited availability. Efforts to force comonomer incorporation to higher levels often result in low molecular weight products or products with broad composition distribution.

Higher [alpha]-olefins are readily polymerized with metallocene catalysts, making possible a wide range of ethylene [alpha]-olefin copolymers with a broad range of compositions and properties. Butene, hexene and octene plastomers with up to 20 mole % comonomer are available from commercial metallocene processes.

Activity

For ethylene co-polymerizations, commercial Z-N catalysts have very low activity. Vanadium Z-N catalysts for EP(D)M production, for example, have catalyst activities around [10.sup.3] g polymer/g catalyst and below. Low activity results in high catalyst residues that require de-ashing to remove them from the polymer.

In contrast, metallocene catalysts show activity tip to [10.sup.6] g polymer/g catalyst under certain conditions. As a result, the catalyst residues are present in such low concentrations that de-ashing is not required. Elimination of the de-ashing step simplifies the process and allows product to be pelletized directly in extruders. Low density ethylene elastomer grades are now available in pellet form in addition to the conventional bales.

Metallocene plastomers

The first commercial metallocene elastomers were plastomers, low density ethylene [alpha]-olefin copolymers that were too soft to make efficiently and cleanly in the gas phase. Although ethylene-butene (EB) plastomers made from vanadium Z-N catalysts were made at the time (early 1990s). the plastomer business did not really flourish until metallocenes made it possible to incorporate high levels of comonomer with high activity at reasonable molecular weight. Metallocenes made it possible to use higher [alpha]-olefins for better toughness, to achieve lower density for better impact properties and to maintain higher molecular weights than previously possible.

Figure 3 illustrates the point. An analytical method that coupled size exclusion chromatography (SEC) for mwd analysis with Fourier transform infrared spectroscopy (FTIR) for composition analysis bore this out. In the SEC-FTIR method, the molecular weight distribution of chains was fractionated and each fraction was then analyzed for composition by FTIR (ref. 6). Figure 3a illustrates how conventional Z-N copolymer has the comonomer concentrated in the low molecular weight chains, while the high molecular weight chains are ethylene rich. The SEC-FTIR of the metallocene product has consistent composition distribution across the molecular weight distribution (figure 3b).

[FIGURE 3 OMITTED]

Today we continue to exploit the metallocene advantage by developing grades with even lower density and higher molecular weight. New plastomer grades available include two low-density ethylene-octene (EO) plastomers, a 0.5 MI, 0.860 g/cc grade and a 3.0 MI, 0.860 g/cc grade. Low density plastomer grades have a number of interesting applications. For example, we recently showed that low density plastomers improve impact properties in automotive TPO formulations (ref. 7), figure 4.

[FIGURE 4 OMITTED]

Melallocene EP(D)M

EP(D)M has been made in solution processes with vanadium Z-N catalysts since 1961. Although activities, conversions and process temperatures are generally low, vanadium Z-N catalysts are inexpensive, and they show reasonable conversion of the bulky dienes used in these terpolymerizations. Over time, the conventional Z-N EP(D)M products have developed into a large, stable business with over 900 Kt volume.

Metallocenes can be designed that make high molecular weight EP(D)M with reasonable diene conversion much like vanadium Z-N catalysts. However. some basic differences exist between the catalysts that provide an opportunity to broaden the range of possible EP(D)M products. In addition, Z-N catalysts make products that have advantages over metallocene products in some applications.

Future directions

Chain architecture control with metallocenes

Three areas where metallocenes are expected to distinguish themselves in ethylene elastomer products include long chain branching (LCB). sequence control and stereoregularity.

Branching control

Elastomers are generally soft with low crystallinity. Their melt properties arise from the entanglements of chains. High molecular weight chains, with more entanglements, have better elastic properties and higher toughness. As elasticity increases, the ability to process these chains becomes more difficult. Often a balance has to be found between the required properties and the ability to process the materials. Adding long chain branching to elastomer chains is one way to help achieve this balance (ref. 8).

Metallocenes differ from vanadium Z-N catalysts in that they often terminate through a mechanism called [beta]-hydride elimination (figure 5a). [beta]-hybrid elimination results in unsaturated chain ends that can reinsert into propagating chains, forming a long chain branch (figure 5b). LCB has a dramatic influence on the rheology of the chain by reducing its coil dimensions for a given molecular weight. Low levels of LCB can often improve processing by shear thinning at high shear rates, yet maintain tensile properties at low shear rates. Since metallocenes make products with terminal unsaturation, it is possible to find conditions that influence the level of LCB and thereby to make easier process mg grades (ret. 9).

[FIGURE 5 OMITTED]

This ability is unique to metallocenes since vanadium Z-N catalysts have low activity, low turnovers and do not terminate through a mechanism that results in unsaturation. As a result, Z-N copolymers are generally linear. This difference in branching content is an example of how Z-N products and metallocene products can have similar specifications but slightly different properties. Often these differences compliment each other. In applications where high molecular weight and high elasticity are important, a Z-N product may work better, and in applications where processing is important, a metallocene product may be the answer.

Sequence distribution control

Since vanadium Z-N catalysts incorporate comonomer poorly, they have even more difficulty inserting comonomer adjacent to another comonomer. Comonomer diads and triads are rare, and comonomer tends to be isolated between ethylenes and very well distributed along the ethylene chain. When comonomer is dispersed well, the sequence distribution tends toward alternating as comonomer content approaches 50 mole %, and polymerization activity tends to fall sharply as comonomer content is forced higher.

Sequence distribution can be predicted by measuring the product of a catalyst's reactivity ratios, [r.sub.e][r.sub.c], where each reactivity ratio is a measure of the relative preference of a growing chain end to the addition of incoming monomer. Thus

(1) [r.sub.e] = [k.sub.ee]/[k.sub.ec]

and

(2) [r.sub.c] = [k.sub.cc]/[k.sub.ce]

where [k.sub.ee] is the rate constant associated with adding ethylene to an ethylene-ended chain and [k.sub.ec] is the rate constant for adding comonomer to an ethylene-ended chain. Likewise, [k.sub.cc] is the rate constant for adding comonomer to a comonomer-ended chain and [k.sub.ce] is the rate constant for adding ethylene to a comonomer-ended chain. Vanadium Z-N catalysts tend to have very small [r.sub.e] values, since [k.sub.ec] is small and [k.sub.ce] is large. The product of the reactivity ratios, [r.sub.e][r.sub.c], will have a value less than one when comonomer incorporation is poor and sequence distribution tends toward alternating. It will have a value close to one when the catalyst has no preference for ethylene or comonomer, and the sequence distribution is random. [r.sub.e][r.sub.c] will be greater than one if the catalyst tends to add comonomer to comonomer and ethylene to ethylene, making the sequence distribution blocky.

For metallocenes, where only one active site has influence over sequence distribution, the product of the reactivity ratios can be calculated from the [sup.13]C NMR using the method of Kakugo (ref. 10).

(3) [r.sub.e][r.sub.c] = 4(EE)(CC)/[(EC).sup.2]

where (EE) is the molar concentration of ethylene-ethylene diads. (CC) is the molar concentration of comonomer-comonomer diads and (EC) is the concentration of ethylene-comonomer diads.

When vanadium Z-N catalysts are modified to make narrow cd and mwd, they approach single-sitedness, and the product of the reactivity ratios can be measured by NMR. Values reported in the literature for these catalysts are around 0.25 and support the premise that sequence distributions tend toward alternating in vanadium Z-N copolymers (ref. 11).

Although many metallocenes tend to make random distributions of ethylene [alpha]-olefin copolymers, a range of values for the product of the reactivity ratios is reported in the literature (ref. 12). A metallocene's preference for comonomer can be influenced somewhat through ligand design, either through electronic interaction or steric interaction with comonomer, and this is important in product design because the sequence distribution can affect the crystallization behavior and ultimately influence mechanical properties.

Figure 6 illustrates the point. Two metallocenes were used in a series of ethylene copolymerizations. M-[C.sub.2] is a bridged bis-metallocene with [C.sub.2] symmetry, and M-[C.sub.s] is a bridged bis metallocene with Cs symmetry. The products of their reactivity ratios were measured by [sup.13]C NMR as 0.56 and 0.36, respectively. Therefore. M-[C.sub.s] is expected to have well dispersed comonomer, shorter ethylene run lengths and lower crystallinity at a given composition relative to M-[C.sub.2]. Differential scanning calorimetry (DSC) confirms that M-[C.sub.2] has both more crystallinity as measured by heat of fusion (J/g) and earlier onset of crystallinity.

Since the literature values of the product of the reactivity ratios are even lower for vanadium Z-N products, it was expected that they would display even less crystallinity and later onset temperatures. However, DSC melt temperatures for these products overlay the M-[C.sub.2] curve. The higher than expected crystallinity is probably the result of the multi-sited nature of the Z/N catalyst and broader composition distribution of the product.

It is interesting to note that through the careful choice of metallocene it was possible to match the thermal behavior of the conventional product. M-[C.sub.2]-like metallocenes could be used to design "drop-in" grades for conventional products and M-[C.sub.s]-like metallocenes could be used to make low density grades where comonomer is more evenly distributed along the chain.

Stereoregularity

Z-N catalysts and metallocenes are each capable of making stereoregular isotactic polypropylene. But it took an understanding of the well-defined catalyst structure in metallocenes to really sort out the mechanism of stereoregular polylnerization. The relationship between metallocene symmetry, comonomer orientation during insertion and chain architecture is well documented (ref. 6). Metallocenes have been designed that enchain prochiral monomers like propylene in atactic, isotactic, syndiotactic and hemi-isotactic orientations. Even though each is homo-polypropylene, their properties are vastly different. The orientation of the monomers has a huge effect on coil dimensions, crystallinity, viscosity, compatibility, entanglement and almost every other aspect of polymer physics.

Although stereoregular polymers are generally thought of as homopolymers like polypropylene, the short comonomer runs found in ethylene copolymers are still stereoregular (ref. 13). Three ethylene-propylene (EP) copolymers with similar composition, about 65 wt. % ethylene, were made using an isospecific metallocene, a syndiospecific metalloccne and an aspecific vanadium Z-N catalyst. [sup.13]C NMR showed that the isospecific metallocene product contained meso propylene diads, the syndiospecific metalloccne product had only rac propylene diads and the Z-N product had a few random propylene diads (figure 7).

[FIGURE 7 OMITTED]

It has been shown that isotactic polypropylene and syndiotactic polypropylene have different melt viscosity (ref. 14). To show how comonomer orientation can affect copolymer properties, the viscosity of EP copolymers made with isospecific and syndiospecific metallocenes was compared to isotactic and syndiotactic homopolypropylene (figure 8). It was found that the copolymers followed the same trend in viscosity as their homopolymer counterparts. As a result, it is possible to tailor the melt viscosity of EP copolymers to meet the needs of a specific application through the proper choice of metallocene.

[FIGURE 8 OMITTED]

Summary

Metallocenes have had a huge impact on the ethylene elastomers segment of the polyolefins industry. Metallocene products have fueled strong growth in plastomers for over ten years and have made inroads into the EP(D)M markets more recently. In addition, new metallocene products, called specialty elastomers, promise to accelerate growth in this emerging market.

Although metallocene products dominate the plastomers market due to the advantages they provide, the distinctions in metallocene EP(D)M products are more subtle, and in some applications, conventional Z-N products are favored. Evidence has been presented here to illustrate the advantages in each of the conventional and metallocene product platforms in EP(D)M.

Finally, three areas in which metallocenes distinguish themselves have been discussed in terms of the new opportunities they represent. These include using metallocenes to control long chain branching content, to control comonomer sequence distribution and to control stereoregularity. Added control over chain architecture promises to provide customers with a powerful new tool in ethylene elastomer grade design.

Figure 1--global polyolefin consumption

         total Ziegler/Natta     total metallocene
         product consumption     product consumption

1991
1992
1995           0.3%
1996           0.5%
1997           0.8%
1998           1.1%
1999           1.4%
2000           1.6%
2001           1.9%
2002           2.5%
2003           3.3%
2004           4.5%

Table made from bar graph.

References

(1.) K. Sinclair, "Global polyethylene demand update," STA Research Report to ExxonMobil Chemical. July 2003.

(2.) Data compiled from Reference 1 and internal market assessment reports.

(3.) F. Esposito, "ExxonMobil plans Louisiana plant," Plastics News, June 25, 2003.

(4.) S. Srinivas, S. Datta and G. Racine, "Novel polvolefin specialty, elastomer polymers," Society of the Plastics Industry, National Plastics Exposition, June 25, 2003.

(5.) H.H. Brintzinger, D. Fisher; R. Muelhaupt, B. Rieger and R. Waymouth, "Stereospecific olefin polymerization with chiral metallocene catalysts." Angewandte Chemie, Int. Ed. Eng., 1995, 34(11), 1,143.

(6.) A. Faldi, J.B.P. Soares, D.A. Higgins and C. Pavlick, "Characterization of the combined molecular weight and composition distribution of ethylene/alpha-olfein copolymers," ExxonMobil Company Report #PLT.2BA. 98. 1998.

(7.) M.G. Williams. B.A. Harrington and T.M. Millet: "New low-density Exact plastomers as high performance impact modifiers," presentation to the Automotive TPO Conference, Detroit MI, October, 2003.

(8.) D.J. Lohse, S.T. Milner, L.J. Fetters, M. Xenidou, N. Hadjichristidis, R.A. Mendelson. C.A. Garcia-Franco and M.K. Lyons, "Well-defined model long chain branched polyethylene 2--melt rheological behavior," Macromolecules 2002, 35(8), 3,066.

(9.) A. Malmberg, J. Liimatta, A. Lehtinen and B. Loefgren, "Characteristics of long chain branching in ethylene polymerization with single site catalysts," Macromolecules 1999, 32(20), 6,687.

(10.) Kakugo et. al., Macromolecules, 1982, 15, 1,150.

(11.) C. Cozewith, G. Ver Strate, Macromolecules, 1971, 4, 482.

(12.) See for example: Galimberti et. al. Macromolecules 1998, 31, 3,409.

(13.) For peak assignments in EP copolymers see: I. Tritto. Z.-Q. Fan, P. Locatelli, M.C. Sacchi, I. Camurati and M. Galimberti, Macromolecules 1995, 28, 3,342.

(14.) A. Eckstein, J. Suhm, C. Friedrich, R.-D. Maier, J. Sassmannshausen, M. Bockmann and R. Mulhaupt, "Determination of plateau moduli and entanglement molecular weights of isotactic, syndiotactic and atactic polypropylenes synthesized with metallocene catalysts," Macromolecules 1998. 31, 1,335.

Bruce A. Harrington and M. Glenn Williams, ExxonMobil Chemical