melt rheology and morphology of thermoplastic elastomers from polyethylene/nitrile-rubber blends:...

Melt Rheology and Morphology of ThermoplasticElastomers from Polyethylene/Nitrile-Rubber Blends:The Effect of Blend Ratio, Reactive Compatibilization,and Dynamic Vulcanization

JOSEPHINE GEORGE,1 K. RAMAMURTHY,2 K. T. VARUGHESE,3 SABU THOMAS1

1 School of Chemical Sciences, Mahatma Gandhi University, Priyadarshini Hills P. O., Kottayam, Kerala-686 560, India

2 Central Institute for Plastics Engineering and Technology, Chennai-600 032, India

3 Central Power Research Institute, Bangalore-560 080, India

Received 10 June 1999; revised 5 January 2000; accepted 19 January 2000

ABSTRACT: A study of the melt-rheological behavior of thermoplastic elastomers fromhigh-density polyethylene and acrylonitrile butadiene rubber (NBR) blends was carriedout in a capillary rheometer. The effect of the blend ratio and shear rate on the meltviscosity reveals that the viscosity decreases with the shear rate but increases withNBR content. Compatibilization by maleic anhydride modified polyethylene has nosignificant effect on the blend viscosity, but a finer dispersion of the rubber is obtained,as is evident from scanning electron micrographs. The melt-elasticity parameters, suchas the die swell, principal normal stress difference, recoverable shear strain, and elasticshear modulus of the blends, were also evaluated. The effect of annealing on themorphology of the extrudate reveals that annealing in the extruder barrel results in thecoalescence of rubber particles in the case of the incompatible blends, whereas thetendency toward agglomeration is somewhat suppressed in the compatibilized blends.© 2000 John Wiley & Sons, Inc. J Polym Sci B: Polym Phys 38: 1104–1122, 2000Keywords: blends; compatibilizer; dynamic vulcanization; morphology; melt viscos-ity; melt elasticity; flow-behavior index; die swell

INTRODUCTION

In recent years, the rate of new polymers beingintroduced into the market place has decreasedbecause of the high cost of research for developingnew chemistry and processes compared to theprobable return on this investment. As a result,increased attention has been focused on blends oralloys of existing polymers to expand the spec-trum of properties of established products and tomeet new market demands.

The use of elastomer–thermoplastic blends hasbecome increasingly important because the re-sulting thermoplastic elastomers (TPEs) havemany of the properties of elastomers and can beprocessed like thermoplastics. The recyclability ofsuch materials also makes them very attractive.The major drawback that limits the selection ofany two rubbers and plastics from blending is theincompatibility between the two phases that willresult in a blend with inferior properties. Thereare several methods to overcome this problem.For example, a block or graft copolymer can beadded with segments identical to those of thecomponent polymers. This will not bring aboutmolecular-level mixing but reduces the interfacial

Correspondence to: S. Thomas (E-mail: [email protected])Journal of Polymer Science: Part B: Polymer Physics, Vol. 38, 1104–1122 (2000)© 2000 John Wiley & Sons, Inc.

1104

tension and thereby permits better dispersion ofthe minor component in the matrix polymer. Also,the morphology is stabilized, and an overall im-provement in blend properties is obtained. Thus,with the help of compatibilizers, we are in a po-sition to select any two polymer pairs for blendingthat possess the properties to be obtained in thefinal product.

The usual processing techniques for these poly-meric materials include injection molding, extru-sion, and so forth. Therefore, it is very importantto study their rheological properties to optimizethe processing conditions and in developing newprocessing equipment and dies.

The importance of rheological studies in pre-dicting the flow behavior of polymer systems athigh temperatures and the shear conditions ofextrusion and injection-molding processes hasbeen described by several authors.1–3 Due to thegrowing importance of TPEs, the rheologicalcharacteristics of thermoplastic block copolymersand those of thermoplastic–elastomer blendshave also been studied extensively.4–14

The rheological behavior of polymer blends andthat of polymers containing gel and crosslinkedparticles have been reported by different researchgroups.15–22 The effects of rubber particles, bothblack and nonblack fillers, on the flow propertiesof polymer melts have also been studied.23–26

Danesi and Porter6 reported the rheological be-havior of polypropylene and ethylene–propylenerubber blends. The effect of annealing in the rhe-ometer barrel on the stability of the morphologywas also a subject of their study.

When polymer melt rheology is studied, theelastic effects should be given sufficient stressbecause they interfere with the product design.The elastic effects in a polymer melt include thedie-swell ratio, principal normal stress difference,recoverable shear strain, and elastic shear mod-ulus. Various factors affecting the elastic effectsin polymer melts have been reported by severalauthors.27–31

Blends of a specialty rubber such as acryloni-trile butadiene rubber (NBR) and a crystallinethermoplastic such as high-density polyethylene(HDPE) have drawn considerable attention. Thisis because the temperature range of NBR appli-cations could be increased through blending withHDPE. Improved oil resistance and ease of pro-cessing are other attractive features of such ablend. Because the system is incompatible, phe-nomena such as segregation, stratification, andphase inversion are expected as in other het-

erophase polymer systems.32–38 The changes inphysical properties with respect to the blend ratio,compatibilization, and dynamic vulcanization ofHDPE/NBR blends have been reported in our ear-lier publications.39–41 This article deals with therheological behavior of HDPE/NBR blends as afunction of the blend ratio and compatibilizer load-ing over a wide range of shear stresses and tem-peratures. Compatibilization does not have mucheffect on the flow properties but substantiallymodifies and stabilizes the blend morphology.

EXPERIMENTAL

Materials

HDPE with a density of 0.958 g/cc and a melt-flowindex of 7.5 g/10 min was supplied by M/s IndianPetrochemicals Corporation Ltd., Vadodara. NBRwith a density of 0.98 g/cc and an acrylonitrilecontent of 32% was purchased from M/s Synthet-ics and Chemicals, Bareilley, Uttar Pradesh. Allother ingredients, such as maleic anhydride anddicumyl peroxide, were laboratory reagent grade.

Preparation of the Blends

The blends were prepared in a Brabender Plasti-corder (PLE 331) with a cam-type mixer with arotor speed of 60 rpm and at 160 °C. First, HDPEwas melted for 2 min, then NBR was added, andthe mixing continued for 6 min. At the end of themixing cycle, the material was collected and madeinto sheets 2 mm thick in a compression-moldingpress at 170 °C with thermoplastic molds. Inthe compatibilized blends, a compatibilizer wasadded and melted for 1 min after the melting ofthe HDPE. Maleic anhydride modified polyethyl-ene (MAPE) was used as the compatibilizer. Thepreparation of the compatibilizer has been dis-cussed elsewhere.42–44 Dynamic vulcanizationwas carried out with sulfur (S), dicumyl peroxide(DCP), and a mixture of S and DCP. The S systemconsisted of 0.1 phr S, 2 phr tetramethyl thiuramdisulfide (TMTD), and 2.5 phr cyclohexyl benzthiazyl sulfenamide (CBS). In the peroxide sys-tem, the DCP concentration was varied from 1 to4 phr. For the mixed system, 0.05 phr S, 1 phrTMTD, 1.25 phr CBS, and 2 phr DCP were used.The crosslinking agents were added 2 min afterthe rubber was introduced, and mixing wasstopped after 3 min. The blends were designatedH100, H70, H50, H30, and H0, depending on the

MELT RHEOLOGY OF THERMOPLASTIC ELASTOMERS 1105

proportion of HDPE in the blend. In the case ofthe compatibilized blends, the letter M stands forMAPE, and the subscripts following M indicateits amount in the blend. For example, H70M2stands for a 70/30 blend of HDPE/NBR containing2% of the compatibilizer MAPE. In the dynami-cally vulcanized system, S, D, and SD stand forthe S, DCP, and mixed curing systems, respec-tively. Thus, the blend H50D1 stands for a 50/50blend of HDPE/NBR crosslinked with 1 phr DCP.

Melt-Flow Measurements

Melt-flow measurements were carried out in anInstron capillary rheometer, Model 3211. A cap-illary 25.468 mm long and 0.762 mm in diameterwas used. The angle of entry was 90° so as tominimize the end effects. The samples for testingwere cut into small pieces and fed into the barrelof the extrusion assembly without air entrap-ment. The melt height before extrusion was keptthe same in all the experiments. The materialwas extruded at 165, 175, and 185 °C through thecapillary at six different preselected crossheadspeeds. Forces corresponding to specific plungerspeeds were read on a dial. The forces and cross-head speeds were converted into apparent shearstress (tw) and shear rate (gw) at the wall withthe following equations involving the geometry ofthe capillary and plunger:

tw 5F

4Aplc/dc(1)

gw 53n9 1 1

4n9

32Qpdc

3 (2)

where F is the force applied at a particular shearrate, Ap is the cross-sectional area of the plunger,lc is the length of the capillary, and dc is thediameter of the capillary. Q, the volumetric flowrate, was calculated from the velocity of the cross-head and diameter of the plunger, and n9 is theflow-behavior index defined by

n9 5d~log tw!

d~log gwa!(3)

and was determined by the regression analysis ofthe values of tw and gwa obtained from the exper-imental data. gwa is the apparent wall-shear rate

calculated as32Qpdc

3 . The shear viscosity h was cal-

culated from tw and gw with the relation h 5 tw/gw.Extrudates emerging from the die of the capil-

lary were collected for die-swell measurements,with maximum care taken to avoid any furtherdeformations. The diameters of the extrudateswere measured at several points with a travelingmicroscope fitted with a micrometer. The die-swell ratio was calculated with the following re-lation:

Die-swell ratio

5Diameter of the extrudate ~de!

Diameter of the capillary ~dc!(4)

The principal normal stress difference (t11 2 t22)was calculated from the die-swell values withTanner’s equation45:

t11 2 t22 5 2tw@2~de/dc!6 2 2#1/2 (5)

Recoverable shear strain (gR) and apparent shearmodulus (G) were calculated from the followingequations46:

gR 5t11 2 t22

2tw(6)

G 5tw

gR(7)

Determination of the Crosslink Density

In dynamically vulcanized blends, the curing re-action can occur during mixing, molding, or ex-trusion. Still, most of the crosslinking reaction iscompleted during the first two stages, mixing andmolding, and the possibility of the curing reactionin the extrusion stage can be discarded. Thus, thecrosslink densities of molded samples were deter-mined. The molar mass between the crosslinks(Mc) of dynamically vulcanized samples was de-termined by equilibrium swelling in toluene47:

Mc 52rpVsf

1/3

ln~1 2 f! 1 f 1 xf2 (8)

where rp is the density of the polymer, Vs is themolar volume of the solvent, and f is the volumefraction of swollen rubber:

1106 GEORGE ET AL.

f 5~d 2 fw!rp

21

~d 2 fw!rp21 1 Asrs

21 (9)

where d is the swollen weight of the sample, f isthe volume fraction of the insoluble components,w is the initial weight of the sample, rs is thedensity of the solvent, and As is the amount ofsolvent absorbed.

The interaction parameter is given by

x 5 b 1Vs

RT ~ds 2 dp!2 (10)

where b is the lattice constant (0.34), Vs is themolar volume of the solvent, R is the gas con-stant, T is the temperature in kelvins, and ds anddp are the solubility parameters of the solvent andpolymer, respectively. The crosslink density (n)was measured from Mc as

n 51

2Mc(11)

Analysis of the Extrudate Morphology

The extrudate morphology was studied with scan-ning electron microscopy (SEM) (Phillips 500C).Cryogenically fractured samples were used. Therubber phase of the samples was extracted withchloroform. SEM pictures were taken at differentmagnifications after the surface was coated with

a gold–palladium alloy. The domain dimensionswere measured by image analysis. About 200 par-ticles were analyzed for the diameter measure-ment.

RESULTS AND DISCUSSION

Melt Viscosity

Melt-viscosity data of the HDPE/NBR binaryblends are presented in Figures 1 and 2 as thevariations of melt viscosity h with the shearstress and blend composition. The general natureof all the graphs in Figure 1 reveals that theviscosity decreases with the increase in shearstress, showing the pseudoplastic nature of theblend systems. At any given shear stress, the meltviscosity is lowest at a 0% NBR content (i.e., pureHDPE). When rubber is added to the plastic, theviscosity increases with NBR content at all shearstresses.

Variations of the melt viscosity as a function ofblend composition at different shear rates isshown in Figure 2. From the figure, it is clear thatthe increase of the viscosity with the rubber con-tent becomes prominent only at lower shear rates.At higher shear rates, only a marginal increase ofthe viscosity with the rubber content occurs. This

Figure 1. Effect of the blend ratio on the shear-vis-cosity/shear-stress plots of HDPE and NBR blends.

Figure 2. Variation of the shear viscosity with theweight percentage of NBR in HDPE/NBR blends at fivedifferent shear rates.


can be explained on the basis of (1) relaxationprocesses and (2) structure buildup, which arepredominant only under conditions of low shear.Munstedt48 extensively studied the flow behaviorof acrylonitrile butadiene styrene copolymer(ABS) with different rubber contents. He calcu-lated the yield-stress values of each composition.Yield stress is the critical threshold shear stressthat should be exceeded to initiate flow. He foundthat the yield stress increases with an increase inthe rubber content. In his studies with ABS, Mun-stedt reached the conclusion that in the range oflow shear stresses where relaxation processestake place, a remarkable influence of the rubber,depending on its concentration, is observed on theflow behavior.

Another reason for the observed behavior isstructure buildup occurring at regions of lowshear. White et al.49 and Metzner50 reported thatparticle–particle interactions begin to dominateat progressively lower deformation rates. Lewisand Nielsen51 studied particle agglomeration oc-curring at lower shear rates. They correlated theviscosity increase with a reduced packing fractionresulting from particle agglomeration. As theshear rate increases, the agglomerated structureis broken down, and the flow curves closely followthe behavior of virgin polymers. The results ofthese studies reveal that in extrusion and injec-tion molding (regions of high shear rate), nostrong influence of the rubbery phase can be de-tected. In the low shear-stress region, where re-laxation process could take place, a remarkableinfluence of the rubber content on viscosity can beobserved.

It is also evident from Figure 2 that the varia-tion of the viscosity with the blend composition isnonlinear with negative deviation with respect tolinear extrapolation between HDPE and NBR ex-tremes. In polymer blends, the viscosity dependson the interfacial thickness and interface adhe-sion in addition to the characteristics of the com-ponent polymers. This is because in polymerblends there is an interlayer slip along with theorientation and disentanglement on the applica-tion of shear stress. When a shear stress is ap-plied to a blend, it undergoes an elongational flow.If the interface is strong, deformation of the dis-persed phase is effectively transferred to the con-tinuous phase. However, in the case of a weakinterphase, interlayer slip occurs, and, as a re-sult, the viscosity of the system decreases.

Utracki and Sammut52 showed that positive ornegative deviations of the measured viscosity

from that calculated by the log additivity rule isan indication of strong or weak interactions be-tween the phases of the blend. According to therelation,

ln~happ!blend 5 Oi

Wi ln~happ!i (12)

where Wi is the weight fraction of the ith compo-nent of the blend. They indicated that immiscibleblends show negative deviations due to the heter-ogeneous nature of the components. Thus, theobserved negative deviation is due to the incom-patibility between the phases and interlayer slip.

Various models have been applied to accountfor the deviations of the blend viscosities from theideal behavior. The models that have been ap-plied include the parallel model, the series model,the Hashin upper limit and lower limit model,53

and the model suggested by Mashelkar et al.54

The relevant expressions are given respectively ineqs 13 to 17:

h 5 h1f1 1 h2f2 (13)

1h

5f1

h11

f2

h2(14)

hmix 5 h2 1f1

1h1 2 h2

1f2

2h2

(15)

hmix 5 h1 1f2

1h2 2 h1

1f1

2h1

(16)

ln~hmix!

5f1~a 2 1 2 gf2!ln h1 1 af2~a 2 1 1 gf1!ln h2

f1~a 2 1 2 gf2! 1 af2~a 2 1 1 gf1!

(17)

where h1 and h2 are the viscosities of phases 1and 2 and f1 and f2 are their volume fractions.

a and g are calculated with the following equa-tions:

a 5f2

f1(18)

1108 GEORGE ET AL.

where f1 is the free-volume fraction of the compo-nent HDPE and f2 is the free-volume fraction ofthe component NBR. Moreover,

f 5 fg 1 af~T 2 Tg! (19)

where fg is 0.025, and

af 5B

2.303 3 C1C25

0.9 6 0.32.303 3 17.44 3 51.6 (20)

g 5b

f1(21)

where b is the interaction parameter.For the calculations, the value of g was varied

to obtain best-fit values with experimental re-sults. The blend viscosities were calculated withthe Mashelkar model with g 5 20.01. This valueof g corresponds to an interaction parameter b of20.001675 according to eq 21.

The viscosity values calculated with the vari-ous models and the experimental values obtainedare shown in Figure 3. It is clear from the figurethat the experimental values are very close tothose of the series model.

Effect of Dynamic Vulcanization

The different crosslinking systems under investi-gation are peroxide, S, and a mixture of S andperoxide. A representative model for the changesaccompanying the dynamic vulcanization isshown in Figure 4. Before vulcanization, the rub-ber particles are coarsely distributed in the plas-tic phase. With S vulcanization, there is only amarginal reduction in the phase dimensions, butwith mixed and peroxide vulcanization, an in-creased extent of crosslinking and particle break-down occur, depending on the effectiveness of thevulcanizing system.

The type of crosslinking system on the shear-stress/shear-viscosity behavior of a 50/50 blend ofHDPE and NBR is presented in Figure 5. Fromthe figure, it is clear that all the blends show apseudoplastic behavior. Another interesting ob-servation is that at a particular shear stress, theshear viscosity is the highest for the DCP-curedsystem, followed by the mixed- and S-cured sys-tem. A schematic model illustrating the types ofcrosslinks formed in each vulcanization system ispresented in Figure 6. In the DCP crosslinkedsystem, networks are formed through strongCOC linkages, whereas in the S-cured system,the crosslinking is through S, which is relativelyweak at high temperatures. In the mixed-curingsystem, both COC and COSOC linkages areformed.

The effect of the DCP dosage on the flow prop-erties of a 50/50 blend of HDPE and NBR is

Figure 4. Schematic model illustrating the type ofvulcanizing system on the morphology of the resultingblend.

Figure 3. Experimental and theoretical values of theshear viscosity as a function of NBR content in HDPE/NBR blends at a shear rate of 54.8 s21.


shown in Figure 7. All the flow curves follow thesame pattern of decreasing shear viscosity withincreasing shear stress. The viscosity increasesalmost linearly with the peroxide concentration.This is due to the variation in the extent ofcrosslinking with the DCP concentration.

The variation of the crosslink density with theDCP dosage of dynamically vulcanized H50 con-taining different doses of DCP is presented inTable I. Thus, with increased levels of DCP, thecrosslink density also increases and is reflected inthe increased viscosity.

The variation of the viscosity with the crosslinkdensity is pictured in Figure 8. From the figure, itis clear that an increase in the crosslink densityresults in a corresponding increase in the viscosity.

Effect of the Compatibilizer

The effect of compatibilizer incorporation on therheological properties of a 70/30 blend of HDPEand NBR is shown in Figure 9. Compatibilizerincorporation has no significant effect on the flowbehavior.

The effect of the compatibilizer concentrationon the viscosity of a 70/30 blend of HDPE andNBR is shown in Figure 10. From the figure, it isclear that the viscosity increases slightly with thecompatibilizer loading. In an incompatible blend,interlayer slippage occurs, and the resultant vis-cosity is lower than the calculated values. How-ever, when a compatibilizer is added, the situa-tion changes. Now, the interfacial conditionhighly favors compatibility. The compatibilizer,being located at the interface, brings down theinterfacial tension and greatly promotes inter-mixing, which normally leads to increased viscos-ity. However, in this case, no significant increasein the viscosity could be detected. The increase inthe viscosity of binary polymer blends upon theaddition of a compatibilizer was reported by Wil-lis and Favis55 and our research group.56 The

Figure 5. Type of crosslinking system on the shear-stress/shear-viscosity plots of 50/50 blends of HDPEand NBR.

Figure 6. Representative model showing the type ofcrosslinks formed in (a) sulfur, (b) mixed, and (c) per-oxide-cured systems.

1110 GEORGE ET AL.

compatibilizer action is more evident from themorphological observations.

Morphological Observations

To study the effect of the blend ratio, compatibi-lizer incorporation, and shear rate on the state ofthe dispersion of the rubbery phase, SEM studieswere carried out. The samples were etched withchloroform to dissolve the rubbery phase.

Effect of the Blend Ratio

The effect of the blend ratio on the state of thedispersion of the rubber particles of HDPE/NBRblends at a shear rate of 164 s21 is shown inFigure 11(a–c). It is clear from the figure that upto a 50% rubber content forms dispersed domains.The size of the rubber particles increases withincreasing NBR content, and at 70% NBR, a co-continuous morphology is observed. This is due to

the higher proportion of NBR and the lower meltviscosity of HDPE.

Effect of the Extrusion Rate

Figure 12(a,b) shows the extrudate cross-sectionmorphology of H70 at two different shear rates,16.4 and 1640 s21, respectively. From the figure,it is clear that there is a significant size reductionwith an increased shear rate. This is because at ahigh shear rate, the rubber particles in the blendare elongated at the entrance of the capillary.This structure, being unstable, tends to break upinto smaller droplets, thereby decreasing the do-main size and leading to the formation of a finedispersion in the HDPE matrix.

A schematic model for the particle breakup inthe extruder barrel is presented in Figure 13. Asubstantial deformation can be observed at theentrance of the capillary, and, thereafter, a severebreakdown, depending on the shear rate, can alsobe observed.

The particle-size distribution of H70 at threedifferent shear rates is shown in Figure 14. Fromthe figure, it is clear that the distribution patternhas substantially narrowed at higher shear rates.

Figure 7. Effect of the DCP concentration on theshear-stress/shear-viscosity plots of 50/50 blends ofHDPE and NBR.

Table I. Variation of the Crosslink Density (n) withthe DCP Dosage of Dynamically Vulcanized H50

DCP Dosage (phr) n

1 2.5407 3 1024

2 3.4938 3 1024

3 4.3330 3 1024

4 4.8188 3 1024

Figure 8. Shear-viscosity versus crosslink-densityplots of H50 containing different levels of DCP.


This indicates that severe particle breakdown oc-curs at high shear rates. The polydispersity indi-ces are 1.63, 1.26, and 1.26 at shear rates of 16.4,164, and 1640 s21, respectively.

The effect of the shear rate on the number-average particle size is shown in Figure 15. Fromthe figure, it is clear that a significant particle-size reduction occurs at high shear rates.

Effect of the Compatibilizer Loading

The effect of the compatibilizer loading on themorphology of a 70/30 blend of HDPE and NBR ata shear rate of 164 s21 is shown in Figure 16(a–c).The compatibilizer loadings are 2, 5, and 10%MAPE, respectively. There is a substantial de-crease in the domain size of the dispersed phasewith the compatibilizer loading. Also, compatibi-lization makes the particle-size distribution moreuniform. This indicates that MAPE brings downthe surface energy between the NBR and HDPEphases and aids in the finer dispersion of NBR inthe matrix polymer.

The particle-size-distribution curves of theblend containing different levels of the compati-bilizer are shown in Figure 17 as the number

fraction versus the particle size. The particle-size-distribution curves narrow with the compatibi-lizer incorporation. The effect of the compatibi-lizer loading on the average domain size of H70 ispresented in Figure 18. It is evident from thefigure that a substantial reduction in the particlesize is observed up to 5% MAPE; thereafter, theparticle size levels off. The leveling is an indica-tion of interfacial saturation. The decrease in theparticle size and the narrowing of the distributioncurves with the compatibilizer concentration areclear indications of the interfacial modification.The compatibilizer action is further evident fromthe morphologies of the annealed and unannealedsamples.

Effect of Annealing

To study the effect of annealing on the phasemorphology, a few samples were annealed in theextruder barrel at 175 °C for 0.5 h. The effect ofannealing on the morphologies of H70 and H70M10at a shear rate of 164 s21 are shown in Figure19(a,b), respectively. From Figure 19(a), it is evi-dent that the particle size has slightly increasedafter annealing [cf. H70 shown in Figure 11(a)].This is because phase coalescence (agglomera-tion) occurs when an incompatible blend is an-nealed in the extruder barrel. However, the ten-

Figure 10. Effect of the compatibilizer concentrationon the shear viscosity of a 70/30 blend of HDPE andNBR.

Figure 9. Shear-stress/shear-viscosity plots of a70/30 blend of HDPE and NBR containing differentlevels of compatibilizer.

1112 GEORGE ET AL.

dency toward particle agglomeration is highlysuppressed in the compatibilized blend, as is ev-ident from Figure 19(b) [cf. unannealed H70M10pictured in Figure 16(c)]. This observation indi-cates that compatibilization by MAPE not onlyreduces the particle size but also stabilizes theblend morphology. Stabilization of the morphol-

ogy during processing and service is very essen-tial in heterophase polymer blends to ensureproper product performance. This study indicates

Figure 11. Scanning electron micrographs (3400) of(a) 70/30, (b) 50/50, and (c) 30/70 blends of HDPE andNBR indicating the morphological changes of the sam-ples extruded at a shear rate of 164 s21.

Figure 12. Photomicrographs (3400) of a 70/30 blendof HDPE and NBR extruded at shear rates of (a) 16.4and (b) 1640 s21 illustrating particle breakdown athigh shear rates.

Figure 13. Illustrative model indicating the particlebreakdown of an incompatible blend at different shearrates in a capillary during extrusion.


that compatibilization by MAPE stabilizes themorphology of HDPE/NBR blends.

The particle-size-distribution curves of incom-patible H70 and compatibilized H70 (H70M10) be-fore and after annealing are shown in Figure 20.The number-average particle size of H70 in-creased from 6.29 to 8.09 mm after annealing. Thepolydispersity index also shows a slight increaseafter annealing. However, annealing has no sig-nificant effect on the particle size of the compati-bilized blend. In fact, the polydispersity index hascome down slightly after annealing. This clearlyindicates the effectiveness of the compatibilizer inmodifying the interface.

As discussed in the previous section, the parti-cle size can be brought down in two ways. One isincreasing the shear rate of an incompatibleblend and another is adding a compatibilizerwithout the shear rate being increased. Let ushave a look at the morphological stability in bothcases. This can be studied by an analysis of thechanges in the interfacial tension in both cases.The interfacial tension was calculated with theWeber equation:

gAnhc

g125 4Fhd

hcG k

(22)

where g is the shear rate, An is the particle size,g12 is the interfacial tension, and hd and hc are

the viscosities of the dispersed and continuousphases, respectively. In the present case, k is 0.84because hd is greater than hc. The interfacial-tension values obtained with this method are veryhigh, and we cannot consider the values absolute.However, the interfacial tension values calcu-lated with the Weber equation can be successfullyapplied to study the variation of the interfacialtension with either the shear rate or compatibi-lizer concentration.

Plots of the interfacial tension versus the shearrate and the particle size versus the shear rateare presented in Figure 21. It is clear from thefigure that particle-size reduction occurs at highshear conditions. Also, at high shear rates theinterface is highly unstable, as is evident from thehigh interfacial-tension values. This means thatsignificant size reduction achieved at high shearrates will prevail only under the processing con-ditions, and on service or on storage, particle co-alescence will occur.

The variations of the particle size and the in-terfacial tension with the compatibilizer loadingare shown in Figure 22. The shear rate at whichthe investigations are made is 164 s21. It is clearfrom the figure that the particle-size reduction

Figure 15. Variation of the number-average particlesize with the shear rate of H70.

Figure 14. Effect of the shear rate on the particle-size-distribution curves of a 70/30 blend of HDPE andNBR.

1114 GEORGE ET AL.

and interfacial-tension reduction occur simulta-neously with the compatibilizer loading. The in-terfacial-tension reduction and the relatively un-changed morphology even after annealing of thecompatibilized blends indicate the morphologicalstability of the compatibilized blends.

The chemistry of the compatibilizer formationis shown in Figure 23. The added compatibilizerMAPE has a polar part and a long hydrocarbon

Figure 16. SEM pictures (3400) illustrating the ef-fect of the compatibilizer loading on the dispersion ofthe rubber phase of a 70/30 blend of HDPE and NBRcontaining (a) 2%, (b) 5%, and (c) 10% MAPE at a shearrate of 164 s21.

Figure 17. Particle-size-distribution curves of thesamples shown in Figure 16.

Figure 18. Variation of the particle size with thecompatibilizer loading of H70.


part. When the compatibilizer is added to incom-patible HDPE/NBR blends, it acts by locating atthe interface. The hydrocarbon part of the com-patibilizer is miscible with the HDPE phase, andthe polar part is miscible with the NBR phase,thereby reducing the interfacial tension betweenthe phases, permitting a finer dispersion duringmixing. It also provides stability against phasesegregation and results in improved adhesion.

Effect of Temperature on Viscosity

To get an idea about the temperature sensitivityof the blends, Arrhenius-type plots were made.Arrhenius plots of H100, H70, and H50 at threedifferent temperatures, 165, 175, and 185 °C, andat a shear rate of 548 s21 are shown in Figure 24.The data points for all the systems lie on straightlines. The activation energies of flow, calculatedfrom the slopes of these lines, are given in Table

II. The activation energy is lowest for HDPE, andit increases with increased rubber content. Theactivation energy of a material provides valuableinformation on the temperature sensitivity of thematerial. Therefore, such information is highlyFigure 19. Scanning electron micrographs (3400) il-

lustrating the effect of annealing in the extruder barrelon the morphology of (a) incompatible H70 and (b) H70

containing 10% MAPE at a shear rate of 164 s21.

Figure 20. Particle-size-distribution curves of unan-nealed and annealed samples of H70 and H70M10.

Figure 21. Effect of the shear rate on the interfacialtension and particle size of a 70/30 blend of HDPE andNBR.

1116 GEORGE ET AL.

useful in selecting the processing temperature ofpolymeric materials.

Shear-Rate/Temperature Superposition Curve

Anand57 and Mendelson58,59 reported the shear-rate/temperature superposition method to deter-mine the shift factor and used the method topredict the viscosities as a function of tempera-ture. To apply this technique to this HDPE/NBRsystem, flow curves were constructed at 165, 175,and 185 °C as the shear stress versus the log-shear rate (Fig. 25). With 175 °C as a referencetemperature, the average values of the shift fac-tor (aT) are obtained by considering a shear stressin the flow curve of the reference temperature and

shifting the point to coincide with the flow curvesof other temperatures at a constant shear rate,and aT is calculated with the following equation:

aT 5tw~ref!tw~T!

(23)

In this equation, tw(ref) is the shear stress at thereference temperature. tw(T) is the shear stressat a given temperature. To minimize error, theaverage values of aT are obtained by three differ-ent shear rates being chosen. A master curve isconstructed by the plotting of the modified shearstress (aT 3 tw) versus the log-shear rate (log gw)(Fig. 26). It is clear that the points at differenttemperatures fall on the reference-temperaturecurve.

Flow Behavior Index (n*)

The shear-thinning fluids in the non-Newtonianregion exhibit an n9 value less than 1, and it

Figure 22. Effect of the compatibilizer loading on theinterfacial tension and particle size of a 70/30 blend ofHDPE and NBR.

Figure 23. Chemistry of the formation of the com-patibilizer.

Figure 24. Arrhenius plots for a 70/30 blend ofHDPE and NBR at a shear rate of 548 s21.

Table II. Activation Energy (DE) of the Flow ofSelected HDPE/NBR Blends

Sample Code DE (kJ/mol)

H100 4.685H70 5.798H50 12.054


becomes smaller as the shear-thinning deviationfrom Newtonian behavior becomes greater. Thus,n9 may be considered a measure of the non-New-

tonian character of a polymer melt and has beencalled the flow-behavior index of the material.

The values of n9 at three different tempera-tures of a few blend compositions are shown inFigure 27. It is clear from the figure that n9 in-creases with increased temperature in all casesexcept in NBR. This is because as the tempera-ture is increased, the viscosity/shear-rate depen-dence becomes more linear. The decrease in thevalue of n9 with increased temperature in thecase of NBR may be due to the cyclization andcrosslinking of NBR at higher temperatures, lead-ing to reduced flowability. It has been reported60

that conventional unsaturated diene-type nitrilerubbers tend to embrittle (case-harden) when ex-posed to temperatures above 149 °C because of (1)oxidative crosslinking or (2) crosslinking by oilscontaining sulfur-bearing additives.

Melt Elasticity

In a capillary measurement, several phenomenamay be observed in the extruded strand that aremore intimately related to the polymer melt elas-ticity than to the melt viscosity but are extremelyimportant in terms of the commercial processingcharacteristics of different resins. The first one isthe die swell or elastic memory; it has very greatpractical importance in processing operations.The die swell originates primarily61–63 from arecoverable elastic deformation of the flowingpolymer at the entrance of the capillary that par-tially decays on relaxation during the flowthrough the capillary but causes swelling of theextrudate once the die-wall restriction is re-moved. To a lesser extent, relaxation of the nor-mal stress generated in the capillary and the ve-

Figure 25. Flow curves of a 70/30 blend of HDPE andNBR illustrating the variation of the shear stress withthe log-shear rate at three different temperatures.

Figure 26. Master curve of the modified shear stressversus the log-shear rate as a function of temperature.

Figure 27. Effect of the blend ratio and temperatureon the flow-behavior indices of HDPE/NBR blends.

1118 GEORGE ET AL.

locity rearrangements at the capillary exit arealso accounted for by die swell.

Effect of the Blend Ratio and Shear Rate on theDie-Swell Ratio

The die-swell ratio of the extrudates at three dif-ferent shear rates as a function of the blend ratiois shown in Figure 28. The die-swell ratio of pureNBR is not reported as it gives irregular extru-dates. From the figure, it is clear that in all thecases the die swell increases with the shear rate.This is because at high shear rates, the polymermolecules can not respond to the rapidly changingstresses and the stored elastic energy is greater.Once the die-wall restriction is removed, the ex-cess energy is released and manifested as dieswell. Also, it is evident from the figure that thedie swell increases with NBR content at a partic-ular shear rate. This can be explained by the factthat rubber is more elastic than plastic.

Effect of the Compatibilizer Loading on theDie-Swell Ratio

The die-swell behavior of the compatibilizedblends presented in Figure 29 is quite interesting.From the figure, it is clear that the die swellincreases upon the addition of 2% MAPE but,thereafter, decreases with an increase in the

MAPE concentration. This can be explained onthe basis of morphological changes associatedwith the compatibilizer loading. From the SEMpictures, it is clear that H70 contains a few largenonuniformly distributed rubber particles, andthe matrix effect is the major factor contributingtoward die swell. Upon the addition of MAPE, themorphology changes drastically. A large numberof uniformly distributed rubber particles can beseen in the compatibilized blends, and here thedispersed domains play a major role toward re-coverability behavior. This accounts for thehigher de/dc value of H70M2. On the further ad-dition of the compatibilizer, the size of the dis-persed domains becomes finer and finer. The finerparticles are less deformable,64 so there is lessrecoverability, which is reflected in the lower die-swell values of H70M5 and H70M10. As in othercases, for a particular blend composition, the dieswell increases with the shear rate.

Principal Normal Stress Difference (t11–t22)

The principal normal stress difference of theblends extruded at a shear rate of 164 s21 and at175°C is shown in Table III. HDPE shows thelowest value of the principal normal stress differ-ence. The addition of NBR increases the t11–t22values of HDPE. t11–t22 is a measure of the elas-ticity of the blends. As NBR is added, the elastic-

Figure 28. Effect of the blend ratio and shear rate onthe die-swell values of HDPE/NBR blends.

Figure 29. Effect of the compatibilizer loading on thedie-swell values of a 70/30 blend of HDPE and NBR.


ity increases, and there is a corresponding in-crease in the values of the principal normal stressdifference.

Recoverable Shear Strain (gR)

Bogue and White65 suggested the use of the pa-rameter recoverable shear strain, gR, for describ-ing and distinguishing the fluid elasticity of dif-ferent viscoelastic fluids as a function of the shearstress. The gR values shown in Table III indicatethat these values increase with an increase in theNBR content, which is expected. On the process-ing side, gR is very important as the extrudatedistortion tendency reduces with decreasing meltelasticity.66–69 The results of this study revealthat the processability of NBR is improved byblending with HDPE.

Elastic Shear Modulus (G)

From the data shown in Table III, it is clear thatthe highest value of G corresponds to that of pureHDPE and progressively decreases with the ad-dition of NBR.

Extrudate Behavior

The appearance of the extrudates at two differentshear rates is shown in Figure 30. From the fig-ure, it is clear that the extrudate distortion ten-dency increases with the shear rate. At a lowshear rate, the extrudate has a smooth surface;however, at a higher shear rate, the surface be-comes rougher. The rubber content of the blendalso plays a major role in determining the surfacecharacteristics. As the rubber content increases,the surface roughness also increases. Several fac-tors contribute toward surface irregularity. It hasbeen conclusively shown by photographic tech-niques70,71 that a fracturing or breaking of theelastically deformed flowing polymer stream oc-curs at the entrance to the capillary itself at some

critical shear stress. Another factor contributingtoward extrudate distortion is the successivesticking and slipping of the polymer layer at thewall in the capillary.72,73 Moreover, there may bean effect at the exit as well.

CONCLUSION

Melt-flow studies indicate that HDPE/NBRblends exhibit pseudoplastic behavior. The pseu-doplasticity decreases with increased tempera-ture in all the cases except NBR. Morphologicalobservations indicate a two-phase system, withthe rubber existing as the dispersed phase up to50 phr rubber content and, thereafter, as a cocon-tinuous morphology. The negative deviations ofthe blend viscosities are another indication of in-compatibility. Theoretical model studies havebeen carried out to correlate the variation of theblend viscosity with the composition. Experimen-tal values are very close to those of the seriesmodel. The flow curves of the compatibilizedblends are basically similar to those of the incom-patible blends. The compatibilizing action is evi-dent from the changes associated with compatibi-

Table III. Melt Elasticity of HDPE/NBR BlendsExtruded at a Shear Rate of 164 s21 at 175°C

SampleCode t11 2 t22 (Nm22) G (Nm22) gR

H100 21.28 3 105 6.69 3 105 1.26H70 45.79 3 105 4.54 3 105 2.25H50 110.30 3 105 2.41 3 105 4.78H30 127.60 3 105 2.64 3 105 4.92

Figure 30. Extrudate-deformation behavior of HDPE/NBR blends at two different shear rates.

1120 GEORGE ET AL.

lizer incorporation. The important changes are (1)a decrease in interfacial tension values, (2) a fine-ness of the morphology, (3) a slight increase in theblend viscosity, and (4) a narrowing of the parti-cle-size-distribution curves. Annealing studies re-veal that the incompatible blend shows a strongtendency toward particle agglomeration (coales-cence) in the extruder barrel. However, the co-alescence tendency is highly suppressed in com-patible blends, which is a critical requirement fora good compatibilizer. The measurement of theactivation energies reveals that the processabilityof NBR can be improved by blending with HDPE.There is no change in the generic nature of theflow curves of dynamically vulcanized samples.However, it registers an increase in the viscositywith the extent of crosslinking. Thus, the moreeffective and comparatively stronger COC bondsformed in the peroxide-cured system exhibit thehighest viscosity, followed by the mixed- and S-cured systems. The smoothness of the extrudatesurfaces at all shear rates indicates that underconditions of high shear (e.g., injection molding orextrusion), no strong influence of the rubber con-tent on the flow properties can be observed. Thus,by the blending of HDPE with NBR, productswith a wide range of applicability and improvedprocessability can be prepared.

The authors are thankful to R. Janardhan for helpingthem in carrying out the rheological measurements.

REFERENCES AND NOTES

1. Brydson, J. A. Flow Properties of Polymer Melts,2nd ed.; George, Godwin: London, 1981.

2. White, J. L. Rubber Chem Technol 1969, 42, 257.3. White, J. L. Rubber Chem Technol 1977, 50, 163.4. Kraus, G.; Childers, C. W.; Gruver, J. T. J Appl

Polym Sci 1967, 11, 1581.5. Kraus, G.; Gruver, J. T. J Appl Polym Sci 1967, 11,

2121.6. Danesi, S.; Porter, R. S. Polymer 1978, 19, 448.7. Ramos De Valle, L. F. Rubber Chem Technol 1982,

55, 1341.8. Goettler, L. A.; Richwine, J. R.; Wille, F. J. Rubber

Chem Technol 1982, 55, 1448.9. Koshy, A. T.; Kuriakose, B.; Premalatha, C. K.;

Thomas, S. J Appl Polym Sci 1993, 49, 901.10. Kuriakose, B.; De, S. K. Polym Eng Sci 1985, 25,

630.11. Thomas, S.; Kuriakose, B.; Gupta, B. R.; De, S. K.

Plast Rubber Process Appl 1986, 6, 101.12. Akhtar, S.; Kuriakose, B.; De, P. P.; De, S. K. Plast

Rubber Process Appl 1987, 7, 11.

13. Varughese, K. T. J Appl Polym Sci 1990, 39, 205.14. Oommen, Z.; Premaletha, C. K.; Kuriakose, B.;

Thomas, S. Polymer 1997, 38, 5611.15. Plochocki, A. P. Polymers Blends; Paul, D. R.; New-

man, S., Eds.; Academic: New York, 1978; Vol. 2,Chapter 21.

16. Folt, V. L.; Smith, R. W. Rubber Chem Technol1973, 46, 1193.

17. Averopoulos, G. N.; Wissert, F. C.; Biddison, P. H.;Bohn, G. G. A. Rubber Chem Technol 1976, 49, 93.

18. Utracki, L. A. Polym Eng Sci 1983, 23, 602.19. Plochocki, A. P. Polym Eng Sci 1983, 23, 618.20. Nakajima, N.; Collins, E. A. J Rheol 1978, 22, 547.21. Montes, S. A.; Ponie-Velez, M. A. Rubber Chem

Technol 1983, 56, 1.22. Miller, S. Prepr Pap (London) 1978, 8.23. Munstedt, H. Polym Eng Sci 1981, 21, 259.24. Berens, A. R.; Folt, V. L. Trans Soc Rheol 1967, 11,

95.25. Lobe, V. M.; White, J. L. Polym Eng Sci 1979, 19,

617.26. White, J. L.; Czarnecki, L.; Tanaka, H. Rubber

Chem Technol 1980, 53, 823.27. Huang, D. C.; White, J. L. Polym Eng Sci 1979, 19,

609.28. Mendelson, R. A.; Finger, F. I. J Appl Polym Sci

1975, 19, 1061.29. Newman, S.; Trementozzl, Q. A. J Appl Polym Sci

1965, 9, 3071.30. Overdiep, W. S.; Van Krevelen, D. W. J Appl Polym

Sci 1965, 9, 2779.31. Cogswell, F. N. J Non-Newtonian Fluid Mech 1977,

2, 37.32. Van Oene, H. J Colloid Interface Sci 1972, 40, 448.33. Starita, J. Trans Soc Rheol 1972, 16, 339.34. Han, C. D.; Kinn, J. W.; Chen, G. J. J Appl Polym

Sci 1975, 19, 2831.35. Lee, B. L.; White, J. L. Trans Soc Rheol 1975, 19,

481.36. Han, C. D.; Yu, T. C. Polym Eng Sci 1972, 12, 81.37. Han, C. D. J Appl Polym Sci 1973, 17, 1289.38. Han, C. D.; Kinn, Y. W. J Appl Polym Sci 1974, 18,

2589.39. George, J.; Joseph, R.; Thomas, S.; Varughese,

K. T. J Appl Polym Sci 1995, 57, 449.40. George, J.; Prasannakumari, L.; Koshy, P.; Va-

rughese, K. T.; Thomas, S. Polym Plast TechnolEng 1995, 34(4), 561.

41. George, J.; Varughese, K. T.; Thomas, S. Polymer2000, 41, 1507.

42. Coran, A. Y.; Patel, R. Rubber Chem Technol 1983,56, 1045.

43. Coran, A. Y.; Patel, R. (Monsanto). U.S. Patent4,299,931, Nov. 10, 1981.

44. Coran, A. Y.; Patel, R. (Monsanto). U.S. Patent4,355,139, Oct. 19, 1982.

45. Tanner, R. I. J Polym Sci A-2 1970, 14, 2067.


46. Han, C. D. Rheology in Polymer Processing; Aca-demic: New York, 1976; p 115.

47. Border, G. Structural Investigations of Polymers;Ellis Horwood: New York, 1991.

48. Munstedt, H. Polym Eng Sci 1981, 21, 259.49. White, J. L.; Czarneck, L.; Tanaka, H. Rubber

Chem Technol 1980, 53, 823.50. Metzner, A. B. J Rheol 1985, 29, 739.51. Lewis, T. B.; Nielsen, L. E. Trans Soc Rheol 1968,

12, 421.52. Utracki, L. A.; Sammut, P. Polym Eng Sci 1981, 28,

1405.53. Hashin, Z. In Second Order Effects in Elasticity,

Plasticity and Fluid Dynamics; Reiner, M.; Abir, S.,Eds.; MacMillan: New York, 1964.

54. Sood, R.; Kulkarni, M. G.; Dutta, A.; Mashelkar,R. A. Polym Eng Sci 1988, 28, 1, 20.

55. Willis, J. M.; Favis, B. D. Polym Eng Sci 1988, 28,1416.

56. George, S.; Ramamurthy, K.; Anand, J. S.; Groe-ninckx, G.; Varughese, K. T.; Thomas, S. Polymer1999, 40, 4325.

57. Anand, J. S. Int Plast Eng Technol 1994, 1, 25.58. Mendelson, R. A. Polym Eng Sci 1976, 16, 690.59. Mendelson, R. A. Polym Eng Sci 1968, 8, 235.

60. Rubber Technology, 3rd ed.; Morton, M., Ed., VanNostrand Reinhold: New York, 1987; Chapter 11.

61. Clegg, P. L. In The Rheology of Elastomers; Manson,P.; Wookey, N., Eds.; Pergamon: London, 1957; p 174.

62. Arai, T.; Aoyama, H. Trans Soc Rheol 1963, 7, 333.63. Bagley, E. B.; Storey, S. H.; West, D. C. J Appl

Polym Sci 1963, 7, 1661.64. Gupta, A. K.; Jain, A. K.; Maiti, S. N. J Appl Polym

Sci 1989, 38, 1699.65. Bogue, D. C.; White, J. L. Engineering Analysis of

Non-Newtonian Fluids; NATO Agardograph No.144, 1970.

66. Gupta, A. K.; Purwar, S. N. J Appl Polym Sci 1985,30, 1777.

67. Han, C. D. Rheology in Polymer Processing; Aca-demic: New York, 1976; Chapter 5.

68. Han, C. D.; Lamonte, R. R. Polym Eng Sci 1971, 11,385.

69. Han, C. D.; Lamonte, R. R. Polym Eng Sci 1972, 12,77.

70. Tordella, J. P. J Appl Phys 1956, 27, 454.71. Bagley, E. B.; Birks, A. M. J Appl Phys 1960, 31,

556.72. Tordella, J. P. J Appl Polym Sci 1963, 7, 215.73. Benbow, J. J.; Lamb, P. SPE Trans 1963, 3, 7.

1122 GEORGE ET AL.

melt rheology and morphology of thermoplastic elastomers from polyethylene/nitrile-rubber blends:...

Documents