Designing EP(D)M structure to achieve exceptional processibility and product attributes.

Molded elastomeric parts are often employed in applications such as brake seals in automobiles. Such use subjects the rubber parts to extreme low ambient temperatures, for example, in northern latitudes. At those low temperatures, the rubber parts must retain much of their original flexibility to insure proper function. Accordingly, low temperature specifications for most automobile parts are fixed by the extreme ambient conditions. For this reason, EP(D)M polymers used in these applications must have good low temperature resistance to compression set, in addition to meeting certain other specifications. These, for example, include high tensile strength, modulus and elongation, tear strength, good heat aging, good mold flow and lack of adhesion to the mold wall.

It is well known (ref. 1) that the low temperature compression set of ethylene, alpha-olefin, diene polymers is controlled by the glass transition temperature (Tg). Tg is a function of polymer crystallinity, which in turn is controlled by polymer composition (ethylene content) and compositional distribution. Crystalline polymers have a well-defined melting/freezing peak in differential scanning calorimetry (DSC). This peak generally tends to reach insignificant levels at ethylene content less than about 60 wt.%. Polymers below this ethylene content are generally considered "amorphous."

This article will review our experimental results which indicate that within this amorphous region, differences in compositional distribution and ethylene content can have a significant effect on the low temperature compression set.

Experimental

EP(D)M synthesis

For EP(D)M prepared in a solution process, the compositional heterogeneity is governed by the choice of catalyst system (ref. 1). For example, vanadium components combined with ethylaluminum sesquichloride (SESQUI) exhibit kinetics typical of a single active species, resulting in narrow compositional and molecular weight distribution. This is also true of most metallocene based catalysts. On the other hand, the same vanadium components combined with diethylaluminum chloride (DEAC) produce multiple active species, resulting in both a broad compositional and molecular weight distribution.

Molecular weight distribution can be further influenced by either the use of multiple reactors in series or the presence of long chain branching (LCB). The effect that choice of catalyst and co-catalyst can have on the degree of long chain branching is well documented in the literature (ref. 1). Following the scheme described in this reference, one could readily see that DEAC polymers, while broad due to the multiple catalyst sites, tend to be linear, while the SESQUI polymers have higher propensity for LCB. Usually, the degree of LCB is controlled through the addition of a modifier. The role of LCB through the diene with metallocene catalysts depends on the choice of metallocene. The details of such LCB are not well understood and in general tend to be proprietary due to the nascent nature of that field of activity.

Product concept

Low molecular weight (or low Mooney viscosity, ML) and low ethylene content EP(D)M terpolymers based on ethylidiene norbornene (ENB) as the diene, are typically used in molding compounds. A low polymer Mooney viscosity (25-30 ML @ 1+4 125 [degrees] C) and a broad molecular weight distribution are important polymer attributes required for good mold flow, for example, in injection molding. Polymers of such general description have adequately met the industry needs for the past three decades. However, looking ahead, the requirements will be more exacting, with a narrower window for polymer structure design. Therefore, a more precise understanding and definition of this window is essential.

The product concept, therefore, is to design a polymer that would exceed the attributes of existing commercial EP(D)M without processibility debits.

Table 1 shows the polymer characteristics of two commercial polymers (Royalene 52 and Yistalon 2504) widely used in molding applications (hereafter referred to as R and V) and V5504, a new polymer from Exxon Chemical (hereafter referred to as MDV) (ref. 3). These three polymers were used in this comparative study. Polymer V is made with a multi-sited catalyst. Polymer MDV is made with a single-sited catalyst. The nature of the catalyst used to make polymer R is not known, though our data would suggest that it is single-sited.

Table 1 - polymer characteristics

Polymer   Ethylene    ENB        Mooney       Polydis-   CD
           (wt.%)    (wt.%)   viscosity, ML   persity,
                               (1+4), 125      Mw/Mn
                               [degrees] C

V           57.5      4.7          26           3.2      B
R           55.5      5.5          26           3.0      N
MDV          46       4.7          26           2.6      N

Compositional distribution

This is determined by the traditional solvent-non-solvent technique described here. Five grams of polymer are carefully dissolved in 500 ml of n-hexane. An aliquot of isopropyl alcohol is titrated into the flask. The precipitated polymer is collected, dried and characterized for ethylene and ENB composition by FT-IR and for molecular weight by gel permeation chromatography - low angle laser light scattering (GPC-LALLS). This procedure is repeated to obtain additional cuts from the hexane solution until the entire polymer is recovered.

The ethylene and ENB distribution results are shown in figures 1A and 1B respectively. The molecular weight of each fraction is shown in figure 2. From figure 1A, it is clear that polymer V has a broad compositional distribution (CD), with about a 15-point spread in ethylene values of the individual fractions. The highest ethylene fractions are clearly not amorphous (%C2 [is greater than] 60), though the average composition appears to be in the amorphous region. On the other hand, polymer MDV is very uniform in CD. Polymer R is between polymers V and MDV. The ENB distributions shown in figure 1B are less dissimilar, though polymer MDV clearly shows a narrow distribution. The average molecular weight, Mw, of each fraction shown in figure 2 reveals that the polymer R has a very high molecular weight fraction, which is also higher in ethylene content.

[GRAPHS OMITTED]

Compound evaluations

A typical 80 Shore A formulation used in molding applications (table 2) was used for the comparative evaluation of each of the polymers. Vulcanization was monitored in an oscillating disc rheometer (ODR) and physical properties were measured on press cured pads. The compression set resistance was measured at 150 [degrees] C and -40 [degrees] C. Data shown in table 3 clearly point to the exceptional low temperature performance of the MDV polymer compared to the three commercial candidates.

Table 2 - compound formulation

Ingredient                                     Phr

Polymer                                       100
N650 FEF carbon black                          65.0
Tri methyl hydroxy quinoline                    0.5
Polyethylene wax                                2.0
Zinc-2-mercaptototoylimidazole                  1.5
Zinc oxide                                      5.0
Trimethylol propane trimethyl acrylate          3.0
2,5-dimethyl-2, 5-di (t-butylperoxy) hexane     1.75
Table 3 - comparison of cure and physical properties

                                             V      R     MDV

Compound ML (1+4) @ 100 [degrees] C          91     92     88
Hardness durometer A                         78     78     77
100% modulus Mpa                            5.4    4.9    5.9
Tensile strength Mpa                        20.1   18.9   18.6
% elongation at break                       245    282    236
Trouser tear R.T. Pk. value Kn/m            6.6    8.5    6.1
Compression set, press cure, 17 min.
  22 hrs./150 [degrees] C/25% deflection     19     21     18
  24 hrs./-40 [degrees] C/25% deflection,
  instantaneous                             95.5   87.9   74.3

The role of polymer composition and CD can be further clarified by examining a set of polymers that are narrow in CD and differ only in ethylene value. This was carried out using a set of experimental polymers (EX1-EX4) shown in table 4. The low temperature compression set of these polymers is also shown in the same table. It is striking that as the ethylene value is lowered from 50 to 46, there is a dramatic improvement in compression set. Further reduction in ethylene value produces only a slight improvement in compression set. Another experimental polymer, EX5, was produced with a broad CD at very low ethylene content. As shown in table 4, this polymer was poor in compression set performance. From these data, it is clear that the exceptional low temperature compression set performance of polymer MDV can be achieved only by a combination of low ethylene and a narrow CD.

Table 4 - comparison of experimental
polymers for ethylene effect

Polymer   Ethylene    ENB        Mooney,       CD     Compression
           (wt%)     (wt%)      viscosity             set at -40
                                ML (1+4),           [degrees] C (%)
                             125 [degrees] C

Ex-1         50       5.1          17          N          92
Ex-2         46       4.1          28          N          73
Ex-3         40       4.3          25          N          67
Ex-4         39       4.7          26          N          68
Ex-5         48       3.8          27          B          93

Processibility

In molding applications, processibility manifests itself in the form of mixing and milling ease (time), as well as mold flow. To study the effect of polymer structure on processibility, the MDV polymer was compared with an experimental polymer of narrower molecular weight distribution (MWD). The MWD was characterized through Mooney relaxation (MLR) measurement (ref. 2) at various Mooney values. Lower MLR at a given Mooney will correspond to a narrower MWD. MLR-Mooney data for MDV and the experimental polymer are shown in table 5, along with the mixing time, milling time and ease of molding on a scale of 1-10. These data demonstrate that the typical narrow distribution obtained with single-sited catalysts results in poor mixing and milling performance, though mold flow was equivalent. Such a polymer should be sufficiently modified, either by introduction of long chain branching or the use of multiple reactors in series, to optimize processibility.

Conclusions

An EP(D)M with exceptional low temperature compression set performance in a typical molding formulation should possess an ethylene content [is less than] 50% and a very narrow compositional distribution. The narrow CD can be obtained using a single-site catalyst.

For optimum processibility as defined by mixing time, the narrow molecular weight distribution associated with a single-site catalyst should be modified either by introduction of long chain branching or the use of multiple reactors in series.

V5504, a recently commercialized polymer, designed to possess such an optimum structure, is ideally suited for molding applications (ref. 3).

References:

(1.) G. Ver Strate, Encyclopedia of Polymer Science and Eng., 6, (1986), pp. 522-564.

(2.) C.B. Friedersdorf and I. Duvdevani, Rubber World, January (1995), pp. 30-34.

(3.) N. Dharmajaram and R. Liotta, paper no. 33, presented at the meeting of the Rubber Division, ACS, Cincinnati, Ohio, October 17-20, 2000.