EPDM-metallocene plastomer blends for W&C.

The market for EPDM based distribution cable in the U.S. has grown roughly from 10% in 1983 to over 30% today (ref. 1). This increase has resulted from both overall market growth for distribution cable and replacement of cross-linked polyethylene (XLPE) with EPDM based cable. EPDM cable users cite long service life (30+ years), better thermal stability at higher operating temperatures (105[degrees]C continuous use temperature rating) and flexibility that permits ease of installation as some of the factors for its selection (ref. 1). These superior qualities combined with inherent water tree resistance and low crystallinity make EPDM based insulations more attractive than XLPE (refs. 2 and 3). For these reasons, the market for EPDM based cable is expected to grow significantly in the foreseeable future.

Recently, a new advanced EPDM synthesized with conventional catalysts, designated as Vistalon 1703P, has been introduced for wire and cable applications (refs. 4 and 5). The polymer, which contains vinyl norbornene (VNB) as the termonomer, has a unique combination of polymer attributes such as low metal residues, a high degree of long chain branching and low diene content. These attributes, when applied to a medium voltage insulation compound, translate into the desirable performance characteristics of low dielectric loss, compound extrusion at fast rates without surface melt fracture and excellent heat aging properties. The key polymer performance requirements for medium voltage cable insulation are listed in table 1 alongside the polymer attributes needed to meet them. The advanced performance characteristics of V1703P are shown in figure 1 relative to other commercial electrical EPDM grades made with 1,4-hexadiene (HEX) and ethylidene norbornene (ENB).

[FIGURE 1 OMITTED]

In addition to conventional EP(D)M, metallocene plastomers such as ethylene-butene (EB) and ethylene-octene (EO) copolymers also have attributes suitable for wire and cable applications (refs. 6-10). By the term plastomers, we refer to a new family of ethylene higher [alpha]-olefin copolymers in the density range of 0.865-0.905 g/cc. These products, made with a single-site metallocene catalyst, have a uniform compositional distribution, where every polymer chain has about the same composition, and a narrow molecular weight distribution ([M.sub.w]/[M.sub.n] [approximately equal] 2.0) compared to products synthesized from conventional multi-site catalysts. Plastomers have found wide acceptance in automotive applications as impact modifiers for polypropylene resins in thermoplastic olefin compounds, and are now being applied in other market segments such as construction and consumer goods.

In this article, we explore blends of conventional EPDM elastomers and metallocene plastomers in a filled compound designed for medium voltage wire and cable insulation. Using a Mixture Design of Experiments approach, the influence of varying the levels of EPDM, plastomer and filler on compound performance attributes are investigated. Optimal compounds that have the most favorable combination of cure properties, mechanical properties, beat aging properties and processability are identified and validated by experimentation. The results of this study will be of significant interest to wire and cable compounders pursuing the development of novel compounds with enhanced properties and potentially lower cost.

Experimental

Materials

The characteristics of the EPDM and plastomer grades used in this study are shown in tables 2a and 2b, respectively. Three commercial grades of EPDM were used in this study, Vistalon 1703P, Nordel 2722 and Vistalon 8731. Each grade is synthesized with a different diene termonomer; vinyl norbornene, 1,4-hexadiene and ethylidene norbornene, and are referred to here as EPDM-VNB, EPDM-HEX and EPDM-ENB, respectively. All EPDM grades were made with conventional Ziegler-Natta catalysts, and have been applied in commercial wire and cable applications.

Two commercial plastomer copolymers having similar melt index (MI) and density were used; Exact 4033 containing butene as comonomer (EB) and Exact 8201 containing octene as comonomer (EO). These polymers are closer to the polymer characteristics of the EPDM grades typically used in medium voltage insulation. Plastomer polymer attributes, such as low density and low MI, translate into good flexibility required for cable installation and sufficient melt strength to limit sagging during cable manufacture. Both grades were made with a single-site metallocene catalyst and therefore have narrow molecular weight distribution and compositional distribution.

Polymer characterization

Polymer molecular weights are measured from both conventional GPC using differential refractive index detector (DRI) and low angle laser light scattering (LALLS) detectors. Lower moments of the molecular weight distribution, such as number average molecular weight (Mn), are obtained using DRI. Higher moments, weight average molecular weight (Mw) and Z average molecular weights (Mz), are obtained from light scattering.

The relative degree of branching of the EPDM terpolymers is quantified using the branching index (BI). This index is calculated from (i) [M.sub.w], LALLS; (ii) [M.sub.w], DRI; (iii) viscosity average molecular weight [M.sub.v], DRI; and (iv) inherent viscosity (IV) measured in decalin at 135[degrees]C. The branching index is defined by:

BI = ([M.sub.v, br] * [M.sub.w], DRI)/([M.sub.w], LALLS * [M.sub.v], DRI) where [M.sub.v, br] = k [(IV).sup.1/a] and `a' is the Mark-Houwink constant (= 0.759 for EPDM in decalin at 135[degrees]C). The branching index for a linear polymer is 1.0, and decreases with increasing levels of branching.

For plastomers, the melt index ratio (MIR) is commonly used to assess the extent of processability. It is defined as the ratio of the melt index measured with a 21 kg weight to a 2.16 kg weight. A polymer with a higher MIR is prone to more shear thinning, indicative of enhanced processability.

Polymer viscoelastic properties were obtained using the Rubber Process Analyzer 2000 instrument from Alpha Technologies. The experiments were performed in an oscillatory mode at temperatures ranging from 80[degrees]C to 125[degrees]C, constant strain of 14% (1[degrees]arc) and oscillating frequency ranging from 0.21 to 209 rad/sec. The elastic modulus (G'), the loss modulus (G") and the complex viscosity ([eta] *) were measured at varying frequencies.

Compounding and testing

Table 3 shows a typical medium voltage cable insulation compound, known in the wire and cable industry as 3728. The additive ingredients used in our formulations are identical to the ingredients listed in table 3. The blends of plastomer, EPDM and silane treated calcined clay were mixed in a 1,600 cc internal mixer using a volumetric fill factor of 75%. The density of the compounds ranged from 1.04 to 1.23 gm/cc, depending on the proportion of EPDM, plastomer and filler in the formulation. A conventional mixing procedure (polymer first followed by clay and additive ingredients) was used. The total mixing time was seven minutes. The clay was added in three stages at different time intervals for effective incorporation. The dump or discharge temperature of the compounds was typically 120[degrees]C. The compounds discharged from the mixer were sheeted out on a two roll mill. The peroxide curatives were added on the mill and ingested into the compound.

The computer program Design Expert was used for formulating the mixture experiments and performing optimization. The compounds were press cured for 20 minutes at 165[degrees]C. Compound cure, mechanical, heat aging and electrical properties were determined according to ASTM procedures shown in table 4.

Compound processability measurements were conducted on a single screw extruder. The length to diameter (L/D) of the extruder screw is 20/1, the compression ratio is 2/1. A constricted die with a land length of 0.132 cm and diameter of 0.170 cm was selected for producing cylindrical extrudates. Extrusion was performed over a range of screw speeds, varying from 25 to 100 rpm.

The smoothness of the extrudates was analyzed using a surface roughness tester. The instrument is equipped with a diamond stylus that moves over the surface of the extrudate under examination and records the surface irregularities. The imperfections appear as ridges or "shark skin" on the extruded strands. The peak to valley depth of these irregularities is an indication of the roughness of the extrudate. This vertical distance measured by the stylus is designated as [R.sub.t] (mm). A high value of [R.sub.t] indicates that the surface is rough and prone to inch fracture. The inverse of [R.sub.t], 1/[R.sub.t], is used as a measure of the surface smoothness of the extrudates.

Results and discussion

Polymer properties

The structure characteristics and molecular weights of the EPDM terpolymers and plastomer copolymers used in this study are listed in tables 2a and 2b, respectively. The EPDM polymers differ widely in their degree of branching. EPDM-HEX and EPDM-ENB have moderate levels of branching (BI [approximately equals] 0.5), while EPDM-VNB is highly branched (BI = 0.1). The high degree of branching in EPDM-VNB is realized through the incorporation of the VNB termonomer. The mechanism of branching in VNB is through Ziegler polymerization of the pendant double bond, which facilitates extensive branching at low diene levels in the polymer. In contrast, branching in EPDM-ENB and EPDM-HEX is predominantly through cationic coupling of the termonomer, which restricts the extent of branching at low levels of unsaturation. The EPDM-VNB was used as the EPDM component in the mixture DOE.

Two metallocene plastomers, EB and EO, with similar melt index ([approximately equals] 1.0 dg/10 min.) and density ([approximately equals] 0.88 g/cc) were selected for this study. The plastomer resins have low melt index (i.e., high molecular weight), which provides melt strength in the compound, and low density that makes the cable flexible over a wide range of operating temperatures. The plastomers are polymerized with [alpha]-olefins of varying side chain length and amount. The EO polymer, which has a longer [alpha]-olefin side chain length than EB, requires fewer comonomer units to disrupt the ethylene crystallinity. Therefore, for the same density, the EO plastomer will have a lower mole fraction octene compared to mole fraction butene in EB. The plastomers also vary in MIR, MIR of EO > EB. These differences enable us to study the impact of (1) size and concentration of [alpha]-olefin and (2) differences in MIR of the polymers on compound properties and processing.

The size and concentration of [alpha]-olefin in the polymers significantly affects the degree to which polymer chains entangle. The number of entanglements (i.e., entanglement density) in the polymer has a profound effect on compound properties such as filler acceptability, maximum curemeter torque [M.sub.H], adhesion (i.e., tack), blend compatibility and physical properties (ref. 11). The plateau modulus, [G.sub.N.sup.0], is a fundamental viscoelastic property that represents the degree of chain entanglement in the polymer. The plateau modulus is defined according to the theory of rubber elasticity as:

[G.sub.N.sup.0] = [rho]RT/[M.sub.e]

where [rho] stands for the density, R is the universal gas constant, T the absolute temperature and [M.sub.e] is the entanglement molecular weight (i.e., the chain molecular weight between two entanglement junctions). [G.sub.N.sup.0] is independent of molecular weight ([M.sub.n]) and molecular weight distribution (mwd).

The plateau modulus of EPDM, EB and EO polymers was calculated using two different techniques (ref. 12) based on viscoelastic data of G'(storage modulus), G" (loss modulus), [G.sup.*] (complex modulus), tan [delta] (damping factor = G"/G') versus frequency at varying temperatures and a constant strain of 14%. The transition from terminal zone to the plateau zone is distinct for polymers that have a high molecular weight and narrow mwd (ref. 13). With increasing mwd, the width of rubbery plateau decreases, making experimental determination of [G.sub.N.sup.0] more difficult. For highly branched polymers such as EPDM-VNB, the viscoelastic response of the polymer is already in the plateau region. The experimental methods do not work in this case. Since [G.sub.N.sup.0] is primarily dependent on the composition (ethylene content), a less branched EPM copolymer with similar composition as the EPDM-VNB was used in the analysis.

Table 5 shows the calculated plateau modulus results for the different polymers. There are some subtle differences in the calculated values using the methods, but the comparisons between the polymers are consistent. From the data in table 5, the plateau modulus for the EO polymer is significantly lower in comparison to EPM and EB. This reflects a lower density of entanglements relative to EPM and EB. The longer side branches of the octene comonomer make the EO polymer chains more "bulky" (i.e., has greater steric hindrance) in comparison to the more "slender" EPM and EB chains (ref. 14). Consequently, the EO polymer chains arc more stiff (i.e., more restricted chain conformation) and, in turn, entangle to a less degree. This difference in the degree of entanglements of the polymers, as will be discussed in more detail later, has a profound effect on compound properties.

Studies have shown that the presence of LCB can have a significant impact on the processing of polyolefins (ref. 15). Long chain branches, in contrast to the shorter [alpha]-olefin side branches, are long enough to entangle with surrounding chains (i.e., > [M.sub.e]) and thus impact the rheological behavior of the polymer. Previously, we have shown that EPDM-VNB has better processability, demonstrated by extrusion at fast rates with smooth surface, than EPDM-HEX and EPDM-ENB. Viscosity-shear rate data on the three EPDM polymers at 125[degrees]C is shown in figure 2a. EPDM-VNB exhibits a greater degree of shear thinning than EPDM-HEX and EPDM-ENB, resulting in improved processability at higher shear rates. The large decrease in viscosity is consistent with the concept that branched chains are restricted in their ability to reptate (i.e., snake-like motion of stress relaxation and diffusion) as compared to more linear chains. The level of long chain branching in the polymers is consistent with the surface smoothness hierarchy demonstrated by the polymers in compound, EPDM-VNB > EPDM-HEX > EPDM-ENB.

[FIGURE 2 OMITTED]

Figure 2b shows the viscosity-shear rate behavior for EB and EO at 125[degrees]C. The EO sample shear thins, lower viscosity at higher shear rates, to a greater extent than EB, which is more Newtonian in its flow properties. The melt index ratio (MIR) is used as a general measure of the sensitivity of a polymer to shear. The EO polymer has a higher MIR, as shown in table 2b, than the EB polymer, signifying that it has a higher level of long chain branching. The EO polymer, which is more non-Newtonian in its flow characteristics, is anticipated to have better processability than EB. The effect of short chain branches (comonomer effect) is not taken into account in this discussion, but may also be important (ref. 16).

Mixture experimental design

Design of Experiments (DOE) is an efficient process for exploring the functional relationship between the measured response (e.g., compound properties) and the controllable factor (e.g., polymer, filler). These mathematical relationships, which are based on statistical regression models, provide the basis for optimizing formulations to meet the compounder's objectives. DOEs generally fall into two broad categories (refs. 17 and 18):

* Factorial designs (including response surface designs);

* mixture designs.

Factorial designs are commonly used for screening multiple factors that are likely to influence the response, and isolating the vital few from the trivial many. Full factorial designs consist of all combinations of factors set to high and low levels. When a large number of factors is present, only a small fraction of the experiments needs to be completed in order to identify the main factors and the interaction effects between factors. When the response depends on the relative proportions of the factors, and not on the amount of the factors per se, factorial designs are not very meaningful (ref. 19). For these situations, mixture designs are more appropriate. In a mixture experiment, the sum of all the factors in the mixture must add up to a set total (e.g., 100%).

In this article, mixture designs arc used to (1) explore the impact of plastomer level (0 - 100 phr) and plastomer type (EB vs. EO) on compound properties and (2) determine the optimal combination of ingredients (EPDM, plastomer and filler) that yields a compound with the most favorable balance of properties.

Blends of EPDM-VNB and plastomer were formulated using a three component mixture design. The design basis is shown in table 6. Table 7 shows the list of design variables and the ranges over which they were varied. The EPDM-VNB and plastomer were allowed to vary over the range of 0-100 parts per hundred rubber (phr), while the clay was allowed to vary between 30-70 phr. Formulations with low levels of clay were included in this study to explore the possibility of developing a low filler compound with improved electrical properties (e.g., low dielectric loss). The rest of the ingredients in the formulation (antioxidant, vinyl silane, zinc oxide, red lead, paraffin wax, low density polyethylene and peroxide) were all held constant at the levels shown in table 3.

Design points were selected using a mathematical procedure called d-optimal algorithm that evenly distributes the experimental points in the design space. The d-optimal protocol generates 14 design points for each EPDM/plastomer blend, corresponding to a quadratic model that was selected to fit the response. The quadratic model is of the form:

[Y.sub.i] = [SIGMA] [A.sub.i] [X.sub.i] + [SIGMA][SIGMA], [B.sub.ij] [X.sub.i] * [X.sub.j (i [not equal to] j])

where Y is the response variable and X represents the factors. This model captures both main effect terms (A coefficients) and the two factor interaction terms (B coefficients), Only terms deemed statistically significant at the 95% level were included in the regression models. Four experimental points are used for determining lack of fit and pure error. The triangular design space, with EPDM, plastomer and clay as the vertices of the triangle, is shown in figure 3. The unshaded region inside the triangle illustrates the boundaries of the design space. All 14 design points were selected from either the perimeter or interior of the design space. Compounding and testing of each design point was performed as described in the experimental section.

[FIGURE 3 OMITTED]

Tables 8 and 9 show the regression models relating the compound properties to the design variables, along with the appropriate [R.sup.2] statistic for the EB and EO compounds, respectively. The adjusted [R.sup.2] statistic factors in the number of terms in the model relative to the number of design points. This statistic provides an estimate of the fraction of the overall variation accounted for by the model. An adjusted [R.sup.2] of 0.95 implies that the mod-el can explain 95% of the variation. The predicted [R.sup.2] statistic is a measure of how the regression model fits each experimental data point in the design. This statistic is calculated by removing one of the design points from the design space, and calculating the model without the design point. The new model is then used to estimate the removed data and measure the residual. A difference of less than 0.2 between adjusted [R.sup.2] and predicted [R.sup.2] is considered acceptable. For most properties, both adjusted [R.sup.2] and predicted [R.sup.2] values are greater than 0.9, and the difference between them less than 0.2, implying that the models have good predictive properties.

Model design simulations: Impact of plastomer level and type on compound properties

To assess the impact of plastomer level and type on compound properties, design simulations were performed using the regression models listed in tables 8 and 9, for both EB and EO based blends, respectively. This was accomplished by varying the plastomer level in increments of 10 phr from zero (all EPDM) to 100 (all plastomer) at a constant filler level. In the simulations, the level of silane treated calcined clay was fixed at 60 phr and the total polymer blend phr level (EPDM + plastomer) was held constant at 100 phr. All other compound ingredients are held constant at the levels listed in table 3. In some figures, experimental results are overlaid on top of the model simulations. These experimental results were not included in the development of the regression models and are shown for model validation purposes only.

Extrusion results

EPDM-VNB has exceptional processability in compound as demonstrated by fast extrusion rates without melt fracture. The superior processability of EPDM-VNB, relative to EPDM-HEX and EPDM-ENB, can be attributed to the highly branched architecture of the polymer. Figure 4 shows a model simulation demonstrating the effect of blending plastomers with EPDM-VNB on extrudate roughness ([R.sub.t]) in a 60 phr filled compound (i.e., 3728). Experimental results are superimposed on top of the model predictions. Compounds containing up to roughly 20 phr EB in the blend extrude with smooth surface. In contrast, for the EO based compounds, blends containing as much as 60 phr EO extruded free of melt fracture. Blends containing 60 to 80 phr EO extruded with surface roughness similar to EPDM-HEX, and blends containing > 80 phr EO extruded with surface roughness more similar to EPDM-ENB. Magnified images (14x) of the extrudate surfaces of the EPDM-VNB, EPDM-HEX and (50:50) EPDM-VNB/EO based compounds are shown in figure 5. The improvement in processability of the EO based compounds versus the EB formulations is attributed to differing molecular architectures. As discussed previously, the higher MIR of EO polymer furnishes enhanced processability.

[FIGURES 4-5 OMITTED]

Cure properties

EPDM-VNB demonstrates very efficient crosslinking, fast cure rate and high cure state in the 3728 compound. The impact of blending plastomers with EPDM-VNB on compound cure properties is shown in figure 6. Blending plastomer with EPDM-VNB results in a reduction in cure rate. This is not entirely unexpected since a key step during the peroxide curing is the generation of radicals along the polymer backbone followed by subsequent coupling of these to form crosslinks. With EPDM-VNB, the principal mechanisms for polymer chain radical generation are free-radical addition of a peroxy radical across the VNB double bond and hydrogen abstraction by a peroxy radical (refs. 20 and 21). The ease of hydrogen abstraction generally follows the order expected from hydrogen lability; allylic > tertiary > secondary > primary. Increasing the amount of plastomer in the blend dilutes the concentration of VNB double bonds and VNB allylic hydrogens in the compound, resulting in reduced cure rate.

[FIGURE 6 OMITTED]

The reduction in cure rate of EB based compounds in comparison to EO based compounds is due to the lower tertiary hydrogen content in EO (lower [alpha]-olefin content) versus EB. Figure 6b shows that the cure state (MH-ML) of the EO based formulations is lower relative to corresponding EB based compounds. We speculate that the lower plateau modulus (lower entanglement density) of EO relative to EB results in reduced cure state (ref. 11). Although the cure properties of the blends have decreased relative to EPDM-VNB, the cure rate and cure state are comparable to EPDM-ENB and EPDM-HEX across the majority of blend compositions.

Physical and mechanical properties

Figure 7 illustrates the variation of compound mechanical properties as a function of plastomer content in the blend. The increase in hardness, 100% modulus and tensile strength with increasing plastomer content, is consistent with the higher crystalline plastomer replacing the less crystalline EPDM-VNB polymer. Furthermore, the plastomers contain fewer [alpha]-olefin comonomer units in the polymer chains, resulting in longer ethylene sequences. Under stress conditions, crystallinity may be induced through the alignment and ordering of these sequences referred to as strain induced crystallization, resulting in enhanced modulus and tensile properties in comparison to EPDM (ref. 11). The increase in elongation to break with increasing plastomer content is consistent with the reduction in the cure state or crosslink density (figure 6b. This difference also applies while comparing the EO compounds with the corresponding EB formulations.

[FIGURE 7 OMITTED]

Heat aging results

EPDM-VNB has exceptional long-term heat aging properties. The superior heat aging performance of EPDM-VNB is the result of lower unsaturation (0.9 wt. % diene) present in the polymer in comparison to EPDM-HEX (4.0 wt. % diene) and EPDM-ENB (3.3 wt. % diene). Plas-tomers, with no diene unsaturation and lower tertiary hydrogen content (lower [alpha]-olefin content), are also anticipated to have exceptional heat aging properties. The blends, as expected, display exceptional heat aging properties. The EO based compounds exhibit a slight improvement in aged properties relative to EB compounds. This improvement is attributed to a lower tertiary hydrogen content (lower [alpha]-olefin content).

Design optimization and validation

The objective of design optimization is to search for specific formulations within the confines of the design space that satisfy multiple performance criteria. Using a graphical optimization technique, a region is created within the design space where all property specifications are satisfied. This range of operability, often described as a "sweet spot." is achieved by overlaying the contour plots of each response and shading out the areas that do not meet one or more of the property specifications. The property specifications for the wire and cable insulation compound are shown in table 10. These specifications are based on minimum requirements cited in industry standards and on historical laboratory results. Comparing the overlay contour plots of the EB and EO based blends in figure 8, it is readily apparent that EO blends have a much larger design window that satisfy all specifications. The larger sweet spot for the EO based blends is mainly attributed to the enhanced processability of EO versus EB.

[FIGURE 8 OMITTED]

The next step is to find the optimum compound within the sweet spot region that best meets the desired product specifications. This call be achieved by utilizing a numerical optimization technique revolving an objective function called "desirability" (refs. 22 and 23). Each response is assigned a desirability function (d) that ranges from zero (does not meet target) to 1 (completely meets target).

The key to multiple optimization is to maximize or minimize a single objective function (D), which denotes the overall desirability. The function D is expressed as a geometric mean of the desirability functions of each of the response variables as shown below:

D = [([d.sub.1] x [d.sub.2] x [d.sub.3] x ......[d.sub.n]).sup. 1/n] = [([Pi][d.sub.i]).sup.1/n] where n is the total number of responses.

From the above expression, if any of the individual responses ([d.sub.i]) = 0, then the overall D value would be set to 0. An importance rating can also be assigned to the individual responses ([d.sub.i]); where a response such as aged elongation (rating = 4) is considered to be more important than hardness (rating = 2). Table 10 shows the optimization goals, lower and upper limits, set forth for each of the responses, along with the importance rating specified for each response. Optimization using desirability is by and large a trade off that searches for the best formulations that satisfy a ma-jority of the performance objectives. When multiple performance requirements are specified, it is virtually impossible to obtain a value of one for all the individual respon-ses. Thus, the objective function D is never a one: values between 0.3 and 0.7 are typical.

The formulation with the highest overall desirability value, based on simultaneous numerical optimization of the multiple performance requirements, is listed in table 11, and is shown on the overlay contour plot in figure 8. Property predictions were experimentally verified by compounding the optimum formulation in the laboratory (average of three individual mixes) and comparing the predicted properties with the average experimental values. This comparison is shown in figure 9. The experimental properties match well with the predicted properties and therefore validate the regression models. The optimal compound has a lower filler content ([approximately equals] 40 phr clay) than the 3728 medium voltage insulation formulation (60 phr clay), and yet is expected to have a good balance of cure properties, mechanical properties, heat aged properties and processability. These results demonstrate the potential to design compounds at lower clay levels, based on blends of EPDM and plastomers, for improved electrical properties (e.g., dielectric loss and dielectric constant). The electrical properties for the optimal compound are shown in table 12 and are lower than maximum industry specifications cited in table 1.

[FIGURE 9 OMITTED]

Conclusions

Blends of conventional EPDM and metallocene plastomers in a typical medium voltage insulation compound (60 phr clay) offer the following advantages relative to corresponding EPDM:

* Significant mechanical property enhancements (i.e., tensile strength and elongation to break):

* comparable or enhanced processability at high plastomer loadings (up to 70 phr);

* option to formulate to lower clay levels without compromise in processability or mechanical properties;

* excellent heat aging properties.

Plastomers exhibiting more shear thinning (EO) exhibited better processability in compound as manifested by extrusion at fast rates without surface melt fracture.

Results from mixture design of experiments demonstrate that lower filled compounds (< 60 phr clay), based on blends of EPDM and plastomers, can satisfy many performance targets for medium voltage cable insulation.

Table 1 - EPDM polymer requirements for medium voltage wire and cable

Requirement                  ICEA specification for filled EPDM
                             compound (ref. 24)

Good electrical properties   Dielectric constant < 4.0
                             Dissipation factor < 1.5%
Good processability          Surface roughness ([R.sub.t])<75
                               [micro]m *
Good physical properties     Tensile strength > 8.2 Mpa
                             Elongation to break > 250%
Superior heat aging          Aging after 14 days @ 121[degrees]C:
                             > 80% tensile retention
                             > 80% elongation retention
No gels                      No gels > 0.254 mm

Requirement                  Polymer attribute

Good electrical properties   Low metal residues
Good processability          Long chain branching
Good physical properties     High ethylene content
Superior heat aging          Low unsaturates in polymer backbone
No gels                      Gel free synthesis

* Minimum requirement determined in laboratory, not an ICEA
specification
Table 2a - EPDM characteristics

EPDM     Diene      Diene             Mooney          Ethylene
                  structure          viscosity        (wt. %)
                                (1+4)@125[degrees]C

V1703P    VNB    C=C[H.sub.2]           25               77
V8731     ENB    C=C[H.sub.3]           26               76
N2722     HEX    C=C-C-C=C-C            26               76

EPDM       Diene    Mn x    Mw/Mn   Branching
          (wt. %)   1,000             index

V1703P      0.9      36      29.9      0.1
V8731       3.3      35       6.0      0.35
N2722       4.0      39       3.6      0.62
Table 2b - plastomer characteristics

Plastomer     Co-         Mooney        Melt      Co-      Density
            monomer     viscosity      index    monomer    (g/cc)
                          (1+4)        (dg/10   (mole %)
                      @125[degrees]C   min.)

Exact       Butene          28          0.8       12.4      0.880
4033
Exact       Octene          17          1.1       8.9       0.882
8201

Plastomer    % crys-    Mn x    Mw/Mn            Melt
            tallinity   1,000                   index
                                                ratio
                                        ([I.sub.21]/[I.sub.2])

Exact          19        54      2.2              17
4033
Exact          19        45      2.4              35
8201
Table 3 - medium voltage cable compound

Ingredient                      Formulation (phr)

Polymer                                100
Surface-treated calcined clay          60
Vinyl silane                           1.0
Red lead                               5.0
Antioxidant                            1.5
Low density polyethylene               5.0
Wax                                    5.0
Zinc oxide                             5.0
Dicumyl peroxide (40% active)          6.5
Table 4 - test producers

Test                            Test method           Units

Mooney viscosity              ASTM D 1646-99       Mooney units
Cure characteristics:         ASTM D 2084-95
  [M.sub.L]                                            dN-m
  [M.sub.H]                                            dN-m
  ts2                                                  Min.
  tc90                                                 Min.
  Cure state,([M.sub.H]-                            dN-m/min.
    [M.sub.L])
  Cure rate                                            dN-m
Hardness                      ASTM D 2240-91       Durometer A
100% modulus                   ASTM D 412-92           MPa
Tensile strength               ASTM D 412-92           MPa
Elongation                     ASTM D 412-92
Heat aged                    ExxonMobil method     % retention
  elongation                 (14 days aging @
                              150[degrees]C)
Surface rough-               ExxonMobil method       Microns
  ness ([R.sub.t])          (Mitutoyo Surftest
                                  SV-500)
Electrical properties:         ASTM D 150-98
  Dissipation factor       (Samples aged in 90[degrees]C
  Dielectric constant         water, 600 V AC          None
                              source @ 60 Hz)
Table 5 - plateau modulus, [G.sub.N.sup.0] (Mpa)

Polymer   Marvin Oser method    Cross over modulus method

EPM              1.30                     1.50
EB               1.37                     1.49
EO               1.03                     1.09
Table 6 - design description

D optimal mixture design
Three mixture variables
14 total design experiments
- Replicates for lack of fit and pure error = 4
Optimization using desirability function
Table 7 - design basis

Variables (phr)
Ingredients                           Min.    Max.
EPDM-VNB                               0      100
Plastomer                              0      100
Translink 37 clay                     30       70
EPDM-VNB + plastomer + Translink 37 clay = 160 phr
Constants (phr)
Agerite MA =1.5, Drimix A-172 silane =1.0, Zinc oxide =
5.0. ERD 90 (red lead) = 5.0, paraffin wax 1236 = 5.0,
Escorene LD 400 = 5.0, DiCup 40 KE = 6.5
Table 8 - responses and predicted R-squared values (EPDM-VNB/EB blends)

Response                            Correlation

Cure rate                           0.18[V] + 0.59[P] + 1.59[C] -
                                      0.014[P][C]
Cure state ([M.sub.H]-[M.sub.L])    0.18[V] + 0.36[P] + 0.67[C]-
                                      3.84[V][C]
Hardness (durometer A)              0.49[V] + 0.53[P] + 0.65[C] +
                                      5.13[V][P]
100% modulus                        5.22[V] + 6.45[P] + 48.16[C] -
                                      0.40[V][C]-0.39[P][C]
200% modulus                        -0.83[V] + 1.91[P] + 29.83[C]
300% modulus                        2.28[V] + 4.55[P] + 31.77[C]
Tensile strength                    6.25[V] + 13.84[P] + 19.42[C]
Elongation at break                 3.24[V] + 4.78[P] - 1.58[C] -
                                      0.015[V][P]
Aged elongation                     2.02[V] + 3.38[P]- 0.54[C]
  14 days @ 150[degrees]C
Surface roughness ([R.sub.t])       -0.22[V]+1.27[P]-0.80[C]-
                                      5.33[V][P]+0.01 [V][C]

Response                           [R.sup.2]   Adj. [R.sup.2]

Cure rate                            0.97           0.96
Cure state ([M.sub.H]-[M.sub.L])     0.93           0.92
Hardness (durometer A)               0.93           0.91
100% modulus                         0.98           0.98
200% modulus                         0.98           0.97
300% modulus                         0.98           0.97
Tensile strength                     0.98           0.97
Elongation at break                  0.97           0.96
Aged elongation                      0.94           0.92
  14 days @ 150[degrees]C
Surface roughness ([R.sub.t])        0.99           0.98

Response                           Pred. [R.sup.2]

Cure rate                               0.95
Cure state ([M.sub.H]-[M.sub.L])        0.89
Hardness (durometer A)                  0.89
100% modulus                            0.96
200% modulus                            0.96
300% modulus                            0.94
Tensile strength                        0.96
Elongation at break                     0.94
Aged elongation                         0.89
  14 days @ 150[degrees]C
Surface roughness ([R.sub.t])           0.97

[V] = EPDM-VNB; [P] = EO; [C] =surface treated calcined clay
Table 9 - responses and predicted R-squared values (EPDM-VNB/EO blends)

Response                           Correlation

Cure rate                          0.28[V] + 0.29[P] + 1.35[C] -
                                   1.60[V][P] - 9.58[P][C]
Cure state ([M.sub.H]-[M.sub.L])   0.31[V] + 0.31[P] + 1.34[C] -
                                   5.20[V][C] - 8.95[P][C]
Hardness (durometer A)             0.48[V] + 0.51[P] + 0.48[C] +
                                   3219E-04[V][P] + 1.862E-03[V][C] +
                                   2.004E - 03[P][C]
100% modulus                       4.70[V] + 6.44[P] + 43.69[C] -
                                   0.35[V][C] - 0.33[P][C]
200% modulus                       0.075[V] + 2.95[P] + 27.46[C]
300% modulus                       3.40[V] + 5.46[P] + 28.33[C]
Tensile strength                   5.22[V] + 20.42[P] + 21.05[C] -
                                   0.14[P][C]
Elongation at break                2.92[V] + 4.78[P] - 1.23[C]
Aged elongation                    2.02[V] + 3.38[P] - 0.54[C]
  14 days @ 150[degrees]C
Surface roughness ([R.sub.t])      0.090[V] + 0.22[P] + 0.017[C] -
                                   5.33[V][P]

Response                           [R.sup.2]   Adj. [R.sup.2]

Cure rate                            0.99           0.99
Cure state ([M.sub.H]-[M.sub.L])     0.99           0.99
Hardness (durometer A)               0.97           0.95
100% modulus                         0.99           0.98
200% modulus                         0.99           0.99
300% modulus                         0.99           0.99
Tensile strength                     0.93           0.90
Elongation at break                  0.98           0.98
Aged elongation                      0.97           0.96
  14 days @ 150[degrees]C
Surface roughness ([R.sub.t])        0.94           0.92

Response                           Pred. [R.sup.2]

Cure rate                               0.98
Cure state ([M.sub.H]-[M.sub.L])        0.98
Hardness (durometer A)                  0.90
100% modulus                            0.97
200% modulus                            0.98
300% modulus                            0.98
Tensile strength                        0.81
Elongation at break                     0.97
Aged elongation                         0.95
  14 days @ 150[degrees]C
Surface roughness ([R.sub.t])           0.87

[V] = EPDM-VNB; [P] = EO; [C] = surface treated calcined clay
Table 10 - optimization criteria for medium voltage
compound

Property                              Units

Cure rate                           dN-m/min.
Cure state, ([M.sub.H]-[M.sub.L])     dN-m
Hardness                             Durom. A
Tensile strength                       MPa
Elongation                              %
Aged elongation                     % retained
(14 days @ 150[degrees]C)
Surface roughness ([R.sub.t])        [micro]m

Property                                         Goal

Cure rate                           [greater than or equal to] 70
Cure state, ([M.sub.H]-[M.sub.L])   [greater than or equal to] 70
Hardness                                         86-92
Tensile strength                    [greater than or equal to] 8.2
Elongation                          [greater than or equal to] 250
Aged elongation                     [greater than or equal to] 80
(14 days @ 150[degrees]C)
Surface roughness ([R.sub.t])                     <75

Property                            Lower   Upper   Impor-
                                    limit   limit   tance

Cure rate                            70      max.     2
Cure state, ([M.sub.H]-[M.sub.L])    70      max.     2
Hardness                             86       92      2
Tensile strength                      8.2    max.     2
Elongation                          250      500      2
Aged elongation                      80      100      4
(14 days @ 150[degrees]C)
Surface roughness ([R.sub.t])         0       75      4
Table 11 - optimal compound
formulation

Optimal formulation   Weight %   Phr

EPDM-VNB                 28       60
EO                       18       40
Translink 37 clay        54       39
Total                   100      139
Table 12 - optimal formulation electrical properties

Dielectric constant
Original                       2.57
1 day                          2.30
7 days                         2.36
28 days                        2.32
Dissipation factor (%)
Original                       0.32
1 day                          0.62
7 days                         0.59
28 days                        0.61

References

(1.) J.H. Dudas and W.C. Cochran, IEEE Electrical Insulation Magazine, 15, 29, (1999).

(2.) S. Boggs and J. Xu, IEEE Electrical Insulation Magazine, 17, 1, (2001).

(3.) Niwa, T., Takahashi, T. Isaka, M. and Takehana, H., Conf. Rec. IEEE Int. Syrup. Electr. Insul. (1988).

(4.) P.S. Ravishankar and N.R. Dharmarajan, Rubber World, 219, 23 (1998).

(5.) N.R. Dharmarajan and P.S. Ravishankar, Electrical devices including ethylene, [alpha]-olefin, vinyl norbornene elastomeric polymers, U.S. Patent 5,674,613, Oct. 7, 1997.

(6.) M. Hendewerk and L. Spenadel, Proc. 1991 IEEE Power Eng. Soc., May 1992.

(7.) L. Spenadel, M.L. Hendewerk and A.K. Mehta, Electrical devices comprising polymeric insulating or semiconducting members, U.S. Patent 5,246, 783, Sept. 21, 1993.

(8.) N.R. Dharmarajan, P.S. Ravishankar and C.D. Burrage, Electrical devices including ethylene, [alpha]-olefin, vinyl norbornene elastomers and ethylene [alpha]-olefin polymers, U.S. Patent 5,763,533, June. 9, 1998.

(9.) N.R. Dharmarajan, P.S. Ravishankar and C.D. Burrage, Electrical devices including ethylene, [alpha]-olefin, vinyl norbornene elastomers and ethylene [alpha]-olefin polymers, U.S. Patent 5,952,427, Sept. 14, 1999.

(10). N.R. Dharmarajan, P.S. Ravishankar and C.D. Burrage, Electrical devices including ethylene, [alpha]-olefin, vinyl norbornene elastomers and ethylene [alpha]-olefin polymers, U.S. Patent 6,150,467, Nov. 21, 2000.

(11.) G. Ver Strate and D.J. Lohse, Science and Technology of Rubber, Second Edition, 95, 1994.

(12.) S. Wu, J. Poly. Sci., Polymer Physics Edition, 27,723 (1989).

(13.) W.W. Graessley, Advances in Polymer Science, 47, 67 (1982).

(14.) L.J. Fetters, et al., Macromolecules, 27, 17, (1994).

(15.) R.G. Larson, The structure and rheology of complex fluids, Oxford University Press, New York, 1999.

(16.) D.M. Kaylon, D.W. Yu and F.H. May, Polymer Engineering and Science, 28, 23 (1988).

(17.) M.J. Anderson and P.J. Whitcomb, Chemical Engineering Progress. 4, 63 (1998).

(18.) W.F. Smith, Today's Chemist at Work, 5, 2 (1995).

(19.) J.A. Cornell, Experiments with Mixtures, Second Edition, John Wiley, 1990.

(20.) R. C. Keller, Rubber Chem., Tech., 61, 238 (1988).

(21.) F.P. Baldwin, et al., Rubber Chem. and Tech., 43, 2 (1970).

(22.) G. Derringer and R. Such, Journal of Quality Technology, 12, 4 (1980).

(23.) M.J. Anderson and P.J. Whitcomb, Quebec Metallurgical Conference, Canadian Institute of Metallurgy, 1993.

(24.) Standard for Concentric Neutral Cables Rated 5,000 - 46,000 Volts, ANSI/ICEA S-94-649-1997.