1

e-Polymers 2009, no. 106

http://www.e-polymers.org

ISSN 1618-7229

PA6/NBR blends: improvement of processability and mechanical properties Ana C. O. Gomes,1* Bluma G. Soares,1 Marcia G. Oliveira,2 Marcelo F. L. Oliveira,2 Caio M. Paranhos3

1*Instituto de Macromoléculas - Universidade Federal do Rio de Janeiro, Ilha do Fundão, 21945-970 Rio de Janeiro, RJ, Brasil; fax: +55 21 25627207; e-mail: acogomes@ima.ufrj.br 2Divisão de Processamento e Caracterização de Materiais, Instituto Nacional de Tecnologia, Av. Venezuela, 82, CEP: 20081-312, Rio de Janeiro, Brasil. 3Departamento de Engenharia de Materiais, Universidade Federal de São Carlos, Rodovia Washington Luiz, km 235, CEP 13565-905, São Carlos, Brasil. (Received: 21 July, 2008; published: 12 September, 2009)

Abstract: This work deals with the study of the processability of blends based on PA6 and NBR. The aim is to achieve reduced PA 6 melting temperatures in order to avoid degradation of the NBR phase during processing. The effect of different additives in PA has been studied. For this purpose PA6 was blended with nonylphenol, EVA 18, EVA 33 and EVAMA. Among several masterbatches based on PA6; the sample containing 5% of EVAMA presented the best combination of processability and mechanical properties. The influence of additives in the processing behavior, crystallinity of the PA6 phase, mechanical and physico-chemical properties are discussed in the present work.

Introduction

Thermoplastic vulcanizate elastomers (TPVs) are the most promising materials in terms of multi-applications and reprocessability. The idea is to combine the processability of thermoplastics and the elasticity of rubbers in a unique material. Thus, it is possible to achieve new properties without expensive process to develop a new polymer. The most common way to produce TPVs, nowadays, is through dynamic crosslinking of the elastomer phase. Dynamic vulcanization was first related by Gessler and Haslett [1] and later by Coran and Patel [2], and has been considered the best way to produce TPVs. Mehrabzadeh [3, 4] describes a long history where the crosslinking of the rubber phase can improve properties such as tensile strength, fatigue resistance, resistance to attack by fluids and permanent deformation.

Thermoplastics like PA have high modulus, good mechanical strength, dimensional stability at elevated temperature and chemical resistance to many moderately polar and non-polar organic species. NBR is a rubber used for special purposes, because of its both good oil and abrasion resistance [5, 6]. Such blend combinations are expected to have excellent hot oil resistance and good strength properties, particularly at elevated temperatures [2, 7].

Despite this fact, few papers have been found in literature, regarding blends of PA6 and NBR, probably because of the high temperature required to process PA6, which may cause rubber degradation.

2

Nevertheless, there is still interest in this kind of TPVs, since some papers can be found in literature in recent years. Kumar and co-workers [7] studied the effect of blend ratio, dynamic crosslinking, compatibilization with chlorinated polyethylene and temperature on the flow behavior of NBR/PA6-PA6.6 blends. They observed the increase of viscosity with the increase on NBR compatibilized content. Compatibilized blends present higher viscosity, more sensitivity for temperature changes, better morphology and low level of die-swell. Chowdhury and co-workers [8] investigated the use of carboxylated NBR as compatibilizer for blends formed by NBR and PA6/6.6/6.10 copolymer. They observed the enhancement of tensile properties, hardness, swelling in oil and set properties. The good performance of carboxylated NBR was confirmed by dynamic-mechanical analysis. The strategy of using a PA copolymer with lower melt temperature reduces the possibility of high temperature applications. This feature is not attractive in terms of commercial applications, principally in automotive parts.

Another alternative has been the use of hydrogenated NBR (HNBR). Bhowmick et al [9] studied the compatibilization of HNBR/PA6 blends with carboxylated NBR, amine-terminated NBR and liquid carboxylated NBR. Evaluation of interfacial parameters shows that a new layer of compatibilizer between phases is formed, whose thickness depends on the type and concentration of the compatibilizer. Piglowisk et al [10] studied the effect of HNBR modified with oxazoline in HNBR/PA6 blends. Differential scanning calorimetry analysis shows partial miscibility between phases and the impact strength is increased in compatibilized blend. Another work of Bhowmick [11] showed the influence of dynamic crosslinking in the interfacial parameters of HNBR/PA6 blends. The cure process decreases the interfacial thickness. However, the alternative use of HNBR has a drawback. This saturated rubber is a different material, with different properties, less elastomeric and more expensive.

The use of additives has been confirmed as a good strategy of modification of polymers and blends properties. The suitable use of additives can generate a tailor-made polymeric material, adjusted for specific applications, without the need of modification of the main polymeric chain.

For industrial application of melt mixing in order to obtain TPVs, masterbatch dilution method seems to be favorable as compared to the direct mixture of TPVs components. In a first step both thermoplastic and rubber are individually mixed with its respective additives. In the second step the direct mixture of the obtained masterbatches occurs in order to get well dispersed additives within the matrix. The present work relates the effect of different additives on the processability and mechanical properties of NBR/PA6 TPVs prepared by masterbatch technique. Results and Discussion PA6 masterbatches

Figure 1 shows the mixing torques recorded during the preparation of PA6 masterbatches. Torque measurements can give interesting information about the material, as breaking of polymer chains, reaction between phases or crosslinking [12]. The delay time observed in the beginning of blending can indicate lose of adherence between molten mass and the rotor, due the low viscosity of the additives. Consequently, torque curves were not used as parameter to evaluate rheological differences between PA6 masterbatches.

3

Fig. 1. Torque curves for PA6 masterbatches. PA masterbatches take different times to reach maximum torque value. These differences occur due to the incorporation of the additives into PA6 mass. It is observed that the time for incorporation is higher for masterbatch containing 5 phr of nonylphenol (NP). It was verified that NP has adhered to the chamber walls and rotors, thus preventing the polymer shearing.

Fig. 2. Stock temperature curves for PA6 masterbatches. The sample with ethylene-co-vinyl acetate copolymer modified with maleic anhydride functional groups (EVAMA) showed shorter time of incorporation. This could be a result of the chemical reaction between anhydride groups graft onto ethylene-co-vinyl acetate copolymer (EVA) backbone and amino groups of PA6. In fact, there are several citations in literature of the reaction between these polymers at the interface

4

of the blend [13-21]. Another possibility to explain the fastest increase in mixing torque is the affinity of maleic anhydride (MA) group to the metal of the chamber [13].

Figure 2 shows the stock temperature for PA6 masterbatches. The frictional heat caused by the input mechanical energy during the mixing raises the stock temperature. Temperature variation can be associated to the addition of additives at room temperature and their heat capacities. The addition of EVAMA, however, results in smaller decrease of temperature than other EVA copolymers. This fact also can be associated to the EVAMA/PA6 reaction. The heat generated by the reaction can be responsible for the smaller decrease in temperature.

The role of the additives in this study is to decrease melting temperature of PA6 in order to improve the processability and to avoid a further thermal degradation of NBR phase during melt mixing. Table 1 shows thermal parameters as melt temperature

(Tm) and melt enthalpy (Hm), for PA6 masterbatches (data collected from the second heating cycle). It is possible to note that the Tm values are very similar for all samples.

However there is a significant difference in Hm. It seems that the presence of 10% of NP act as nucleating agent for PA6. This result is in accordance with that observed by Zang and co-authors [22] in polypropylene/polyoxyethylene–nonylphenol blends. In the case of EVA, higher content of VA contributes to enhance the formation of hydrogen bonds. The strong interaction between NH–C=O groups can lead to a rearrangement of molecular conformation [23]. This fact can explain the increase of crystalline content of PA6 masterbatches containing EVA. The observed decrease of

Hm in PA6/EVAMA confirms the strong reaction between the components [13-21].

As a matter of fact, decrease of Hm, in semi-crystalline polymers indicates a decreasing on the crystallinity degree of the system. Lower crystallinity can result in less energy consumption by the internal mixer for melting and processing the material. Tab. 1. Thermal parameters and MFI values for PA6 masterbatches.

Tm

(±1.5 °C)

Hm

(± 2.0 J/g)

MFI (± 0.1 g/10

min)

Processing Energy (± 0.70

J)

Final Torque (± 0.15 Nm)

Pure PA 222.0 59.3 19.6 29.64 4.05 PA/NP5 221.9 57.4 21.4 22.49 4.67

PA/NP10 222.0 69.9 21.8 17.28 5.14 PA/EVA18 221.3 46.3 19.2 19.97 5.59 PA/EVA33 221.9 57.4 17.4 20.64 5.13 PA/EVAMA 222.2 45.3 16.2 19.27 6.17

The energy expend to make a blend (processing energy) may be a good parameter to compare processability. This value is obtained by integration of the area below torque curve. As shown in Table 1 the presence of NP causes a decrease in processing energy, in comparison to pure PA sample. This fact can be attributed to its lubricant action [24]. The processing energy values for EVA28 and EVA33 show no significant difference. The addition of EVA causes a decrease in processing energy too, owing its action of processing aid. Our research group has reported the ability of EVA in decreasing the torque and consequently the viscosity of the system, during melt mixing [25-27]. EVAMA also causes a decrease in processing energy of EVAMA masterbatch. As previously cited, lower processing energy can be interesting

5

for obtaining the aimed material. It is expected to diminish the consumption of energy to melt and to process the blend. Therefore less shear and thermal degradation of NBR phase is achieved. These possible advantages confirm EVAMA as a good additive for the purposes of this work.

Table 1 also presents other processing parameters like final torque and melt flow index, both are related to the masterbatch viscosity. The values of final torque did not present significant differences. Meanwhile MFI values vary accordingly to the additive type. EVA based additives produce masterbatches with higher viscosity than pure PA 6. This increase can be associated with EVA composition (VA amount) and chemical reaction between EVAMA and PA6 as earlier discussed. The higher viscosity presented by the masterbatch containing EVAMA can be interesting for morphology development in a further blend with NBR. It is known that difference between the viscosities of blend components is an important parameter to evaluate the development of final morphology [24]. Lower difference between the viscosities of blend components improves the shear between phases favoring the breaking of the particles.

An extended study of NBR masterbatches was performed previously. Several additives were tested: estearic acid, WB222 (a mixture of high-molecular weight, aliphatic, fatty acid esters and condensation products), viernol (a mixture of low molecular weight polyesters), nonylphenol and dioctyl phthalate. However the results obtained do not change significantly the final properties of blends. It was concluded meanwhile that the use of additives in NBR phase could not be justified, for the blend system studied. PA6 / NBR blends

A set of blends was prepared based on NBR/antioxidant (AO) masterbatch (10 phr of

Naugard antioxidant) with all PA6 masterbatches, keeping a composition of 50:50 wt%. The blends nomenclature consists in a “B” letter after the additive name. Figure 3 shows the mixing torque of blend samples. It is possible to observe differences between the time of initial increase of torque, where PA6 masterbatches are beginning to be melted and plastified. There are also differences in load peak of NBR masterbatch, which can be associated to the interaction between PA6 masterbatches and NBR. This interaction was previously related in literature [26-28].

The delay on the increase of torque indicates that all samples containing additives showed a tendency to attach to the Brabender rotor. That is, more time is necessary to observe the same behavior in blend where PA is not additived.

In the case of EVAMAB, this fact can be justified by the affinity of maleic groups with the metal of rotors. Blends with NP take more time to reach maximum torque when NBR was added. These blends have also the lower final torque values (Table 2). This behavior can be related to the higher degree of crystallinity of the PA6 phase in the presence of NP. After 3 min of mixing, the torque is reduced by the addition of the crosslinking system. The crosslinking of elastomeric phase causes a torque increase, reaching a stabilized value. The addition of AO does not cause variation in torque profile. Figure 4 shows the stock temperature curves of blend samples. It is seen that higher viscosity of NBR generates higher frictional heat during processing, leading to equalization of the stock temperature.

6

Fig. 3. Torque curves for NBR/PA6 blends.

Fig. 4. Stock temperature curves for NBR/PA6 blends. Table 2 shows the processing parameters and MFI values for blends. It can be noted differences in processing energy and final torque values. About the mixing time, EVAMAB shows the lowest value of processing energy. When the crosslinking process is considered, NP10B is the more favorable chose to reach a TPV material, as well as the low final torque. These differences, however, are not sufficient to justify the choice of the additive. It is necessary to consider the effect in mechanical properties.

The comparison between MFI values for samples before and after injection molding can be useful to evaluate the reprocessability of materials. EVAMAB presents the

7

lowest MFI value, as well as the smallest difference of MFI value after injection. This means that the aging effect of the injection process was smaller for this sample. Tab. 2. Processing parameters and MFI values for NBR/PA6 blends.

Processing Energy up to 3

min ( 0.70 J)

Total Processing

Energy ( 0.70 J)

Final Torque

( 0.15 Nm)

MFI (± 0.1 g/10 min)

Before injection

After injection

MFI

PAB 32.13 55.73 6.09 3.1 2.4 -0.7 NP5B 24.68 47.85 5.84 4.0 3.3 -0.7

NP10B 26.16 47.10 5.90 3.2 2.0 -1.2 EVA18B 30.80 56.59 6.71 4.5 2.5 -2.0 EVA33B 31.58 57.89 7.00 4.2 3.4 -0.8 EVAMAB 24.14 50.14 6.00 2.5 2.4 -0.1

Table 3 shows tensile strength and physico-chemical properties for blends. The residual mass after extraction and swelling results indicate that a better crosslinking process occurred for blends with EVA copolymers. Tensile strength, meanwhile, is better to EVAMA, with better stress at break and modulus values. Therefore, the PA6 / EVAMA / NBR was chosen as the best choice for continues our studies. Tab. 3. Tensile strength and physic-chemical properties for NBR/PA6 blends.

Residual mass in toluene

( 3%)

Residual mass in

acetic acid

( 3%)

Mass gain in MEK

( 3%)

(MPa)

(%)

E (MPa)

PAB 100 52 45 16.8 1.2 191 21 290 40

NP5B 92 46 80 15.1 0.9 159 7 210 23

NP10B 100 44 81 13.1 0.5 127 9 167 30

EVA18B 99 55 60 13.4 1.0 145 3 201 40

EVA33B 97 54 65 14.4 0.9 157 13 175 14

EVAMAB 100 50 63 16.0 1.2 156 11 212 11

The efficiency of dynamic vulcanization was evaluated taking into account the weight gain after immersion in MEK and also the residual mass after extraction with toluene (solvent to the NBR phase). It is worth to note that dynamic crosslinking was not fully efficient. The fact that all samples retains their integrity after extraction in both solvents suggests morphology is at least partially co-continuous. Residual mass of the extraction in acetic acid is larger than 50%, what means that some PA masterbatch mass is confined inside NBR phase. This is consistent with a partially co-continuous morphology.

Figure 5 shows the SEM image of NBR/PA6-EVAMA blend sample. Zones of co-continuous morphology can be noted, as well as some zones of dispersed phase morphology, confirming the assumptions discussed above of a partially co-continuous morphology. This morphology is not considered “reprocessable”, consequently this material still cannot be considered a TPV. It is important, however, to emphasize that the present work is a first study to evaluate a potential blend to reach a TPV material.

8

Fig. 5. SEM image of NBR/PA6-EVAMA blend where (a) is an example of a line dispersed domain of NBR inside a domain of PA6 and (b) is an example of small dispersed domain of PA6. Table 4 shows thermal parameters of the pure PA6, PA6-EVAMA and NBR/PA6 blends and the degree of crystallinity (Xc) calculated from DSC analysis. The degree of crystallinity of the samples were calculated from the ratio between the ΔHm of the sample and the ΔHm of 100% crystalline PA6 (ΔHm100% = 240.0 J g-1). [29] The difference between the degree of crystallinity Xc values of pure PA, PA6 / EVAMA and the blends shows that the presence of a second polymer can disorder the crystallization process of PA6, which is expressed in smaller Xc values. These results show that the presence of NBR causes changes in PA crystalline phase. Tab. 4. Thermal parameters and crystallinity of blends.

Tm (± 1.5°C) Hm (± 2.0 J/g) Xc (± 1%)

Pure PA 227.8 93.4 39 PA6/EVAMA 222.0 45.3 30

PAB 220.9 29.3 12 NP5B 220.3 23.8 10

NP10B 221.5 23.2 10 EVA18B 218.7 27.6 11 EVA33B 217.9 23.9 10 EVAMAB 221.3 28.8 12

Figure 6 shows the loss modulus curves of the blend samples obtained from dynamic mechanical tests (DMA). It is possible to note the presence of loss modulus maximum for NBR at -18.4º C and 12.2º C for PA6, corresponding to the Tg of pure nitrile rubber and PA6, respectively. This behavior is typical of an immiscible blend. However, the damping peaks of some blends are slightly shifted, as can be noted in values seen in Table 5. Usually the nature of this shift is associated to an interfacial miscibility between NBR-PA phases [29-31]. The presence of the different additives does not cause an overall change on the molecular mobility of the blend, except in the EVAMAB sample.

It is seen that the maleic anhydride pendent groups increase the local free volume at the interphase leading to an increase on the molecular mobility of NBR phase. Similar effect was observed on decrease of Xc from PA6/EVAMA sample.

9

Fig. 6. Loss modulus of NBR/PA6 blends. The area under the loss modulus curves has been related to the activation enthalpy of relaxation of the backbone motion in magnitude. The average activation enthalpy of transition for polymers can be calculated from dynamic mechanical data using Eq. (1),

avga

gRG

H

TREELA

2

'´2

(1)

where LA is the area under the loss modulus curve, E’G and E’R are the storage modulus in glassy and rubbery state, respectively, (ΔHa)avg is the average enthalpy of the relaxation process [30-32]. The area under the loss modulus curves showed in Figure 6 is qualitatively smaller for the blends than that of the two homopolymers. Table 5 shows data extract from loss modulus curves and the calculated (ΔHa)avg.

Any restriction in the main chain mobility is expected to decrease the area under the loss modulus curve [30-32]. The decrease of the molecular mobility in all blend samples is directly influenced by the restriction of mobility caused by phase restriction inside the PA matrix. NBR can be considered a confined phase inside PA6 matrix, thus its phase mobility can be affected by the changes in matrix mobility due to the incorporation of additives.

The addition of NP only causes decrease in Ha when the amount is equal to 10 phr. As seen before in the results of torque and MFI, NP just acts efficiently as plasticizer in this proportion. The well known property of process aid of EVA18 shows its effect

in mobility increase of PA phase. EVA33 do not change significantly Ha. EVAMA, in

spite of reducing processing energy, increase Ha of PA phase.

10

Tab. 5. DMA parameters and calculated (Ha)avg from NBR/PA6 blends.

Tg ( 0.3°C) LA ( 200 MPa*K) (Ha)avg ( 300J)

NBR

phase PA

phase NBR

phase PA

phase NBR

phase PA

phase

NBR -18.3 - 4746.9 - 9313 -

Pure PA - 11.8 - 3126.4 - 4760

PAB -18.2 13.9 3206.9 1154.9 5405 2360

NP5B -17.5 13.2 3718.8 997.3 5592 2326

NP10B -17.5 13.7 3289.6 920.9 5188 2272

EVA18B -16.5 10.6 3195.2 1252.9 5436 2092

EVA33B -16.9 11.9 3285.4 1214.7 5620 2303

EVAMAB -22.1 9.8 3575.4 995.0 6164 2583

Once again, this behavior can be associated to a reaction between PA6 and EVAMA, leading to a bigger molecule, with more segmental stiffness. In fact, these results are consistent with those observed either by DSC.

In combination to the tensile strength data, EVAMA can be considered the best alternative to continue the research for a new TPV material. The sample obtained with this additive demonstrated the best combination of mechanical properties, crystallinity, efficiency of crosslinking system and morphology. Conclusions

A good balance between processability and properties in NBR/PA6 blends was achieved. The results show that the absence of additives in NBR phase and the addition of EVAMA in PA phase lead to the best composition for PA6/NBR blends. Process temperature could be reduced without decreasing the final properties of the blends. The crosslinking process, yet, must be optimized, and the time of mixture must be studied to reach a stabilized and well-defined morphology. Blending conditions can be adjusted for posterior compatibilization studies. Experimental Part Materials

NBR (28 wt% acrylonitrile), phenolic resin (PR 1045) and nonylphenol (NP) were of commercial grade and kindly supplied by Petroflex Ind. Com. S.A., Brazil. PA 6

(Radilon S natural) was kindly supplied by Radici Group, Italy. EVA with 18%wt (EVA18) and 33%wt (EVA33) of vinyl acetate were kindly supplied by Brasken S.A., Brazil. EVAMA was acquired from Proquimil Produtos Químicos Ltda (MA 0.8% w/w). AO was kindly supplied by Crompton Corporation of Brazil. Stannous chloride (SnCl2) was commercially purchased and used as received. Blend Preparation

All masterbatches and blends were prepared by melt mixing in a Brabender Plasticorder Rheometer internal mixer, with banbury rotor at 80 rpm.

11

The blends were grounded in appropriated mills and injected in a semi-industrial injection-molding machine Battenfeld Plus 35 in 110 bar mold pressure, 240 °C injection temperature, 80 bar injection pressure and 50 bar back pressure. All materials were kept in a vacuum oven at 100 °C until constant weight. PA masterbatch evaluation

Masterbatches of PA6 were prepared at 240 °C, being mixed along 3 min. Table 6 shows PA masterbatches compositions (quantities in phr). PA6 / NBR blends

Table 6 shows the composition of NBR/PA6 blends. The blends were prepared mixing together the PA6 masterbatches and NBR at 220 °C, in a 50/50 composition. The following sequence of preparation was used: PA masterbatch (1 min of processing), NBR (3 min), crosslinking system based on PR and SnCl2 (2 min), AO (in the final 30 seconds). All blends received 10 phr of AO, 5 phr of PR and 0,25 phr of SnCl2 (not shown in Table 6).

Tab. 6. PA6 masterbatches and 50/50 blends compositions.

Additive

PA

masterbatch Amount of additive

PA Pure PA6 - NP5 PA6/NP 5.0 NP10 PA6/NP 10.0

EVA18 PA6/EVA18 5.0 EVA33 PA6/EVA33 5.0 EVAMA PA6/EVAMA 5.0

Processability Characterizations

Melt flow index (MFI) analysis for PA6 masterbatches was made in a TMI model XNR 400, at a temperature of 235 °C, with a weight of 2.17 Kg. Blending torque was measured by Brabender Plasticorder Rheometer, during mixing procedures. Mechanical Characterization

Tensile strength was measured in an EMIC universal testing machine DL-2000, in accordance to DIN 53504 with 200 mm/min of speed, at room temperature (25 °C) and humidity level of 53%. Physico-Chemical Characterizations

The swelling test was performed in methyl-ethyl-ketone (MEK). Samples of approximately 200 mm2 were immersed in the solvent in individual flasks over a week. Selective extraction was performed in Soxhlet extraction system during 24 hours. The PA6 phase was extracted by hot acetic acid and uncrosslinked NBR phase was extracted by hot toluene.

12

Thermal Characterizations

Differential scanning calorimetry (DSC) analysis was performed in Perkin Elmer DSC 7, at temperature range from 40 ºC to 260 ºC, with heating and cooling rate of 10 ºC/min.

Dynamic mechanical analysis (DMA) were performed in a Q800 TA Instruments, in a temperature range from -50 °C to 80 °C, temperature ramp of 3 °C/min, at frequencies of 1, 10, 50 and 100 Hz, using the single cantilever-bending mode with a

deformation of 15 m. Scanning Electronic Microscopy

Scanning electron microscopy (SEM) was performed on a JEOL 5610 LV equipment with a voltage of 20 kV. The samples were cryogenically fractured and chemical stained by osmium tetroxide (OsO4), in order to selectively stain the unsaturated phase. The samples were then coated with carbon and analyzed with a backscattered electron detector. The stained phase (NBR phase) appears as the “brighter” one. Acknowledgements

CAPES, CNPq, FAPERJ, and FINEP/CNPq sponsored this work. We would like to thank Petroflex, INT, Radici and Crompton for the supplied materials. References

[1] Gessler, A. M.; Haslett Jr., William H. U.S. Patent 3037954, 1962. [2] Kumar, R.; George, K. E.; Thomas, S. J. Appl. Polym. Sci. 1996, 61, 2383. [3] Mehrabzadeh, M.; Burford, R.P. J. Appl. Polym. Sci. 1996, 64, 1605. [4] Mehrabzadeh, M.; Delfan, N. J. Appl. Polym. Sci. 2000, 77, 2057. [5] Coran, A. Y.; Patel, R. Rubber Chem. Technol. 1980, 53, 781. [6] George, J.; Neelakantan, N. R.; Varughese, K. T.; Thomas, S. J. Appl. Polym. Sci. 2006, 100, 2912. [7] Kumar, C. R.; Nair, S. V.; George, K. E.; Ommen, Z.; Thomas; S. Polym. Eng. Sci. 2003, 43, 1555. [8] Chowhury, R.; Banerji, M. S.; Shivakumar, K. J. Appl. Polym. Sci. 2006, 100, 1008. [9] Bhowmick, A. K.; Inque, T. J. Appl. Polym. Sci. 1993, 50, 2055. [10] Piglowski, J.; Gancarz, I.; Wlazlak, M. Polymer 2000, 41, 3671. [11] Bhowmick, A. K.; Inque, T. J. Appl. Polym. Sci. 1993, 49, 1893. [12] Rabello, M. “Polymers Additivation” (in portuguese), ArtLiber, São Paulo 2000. [13] Sundararaj, U.; Macosko, C. W. Macromolecules 1996, 28, 2647. [14] Marco, C.; Ellis, G.; Gómez, M. A.; Fatou, J. G.; Arribas, J. M.; Campoy, I.; Fontecha, A. J. Appl. Polym. Sci. 1997, 65, 2589. [15] Barra, G. M. O.; Crespo, J. S.; Bertolino, J. R.; Soldi, V.; J. Brazil. Chem. Soc. 1999, 10, 31. [16] Phan, T. T. M.; DeNicola Jr., A. J.; Schadler, L. S. J. Appl. Polym. Sci. 1999, 68, 1451. [17] Roeder, J.; Oliveira, R. V. B.; Gonçalves, M. C.; Soldi, V.; Pires, A. T. N. Polym. Test. 2002, 21, 815. [18] Barra, G. M. O.; Roeder, J.; Soldi, V.; Pires, A. T. N.; Agnelli, J. A. M. Polimeros 2003, 13, 95.

13

[19] Chow, W. S.; Ishak, Z. A. M.;Karger-Kocsis, J.; Apostolov, A. A.; Ishiaku, U. S. Polymer 2003, 44, 7427. [20] Bhattacharyy, A. R.; Ghosh, A. K.; Misra, A.; Eichhorn, K. J. Polymer 2005, 46, 1661. [21] Kusmono, Z. A.; Mohd Ishak, W. S.; Chow, T.; Takeichi, R. Eur. Polym. J. 2008, 44, 1023. [22] Zhang, Q.X.; Song, H.B.; Wang, S.Y.; Mo, Z.S. Polymer 2005, 46, 11820. [23] Tsebranko, I.A., Pakharenko, V.A. Fibre Chem. 1999, 31, 197. [24] Willems, R. C.; Ramaker, E. J. J.; van Dam, J.; de Boer, A. P. Polymer 1999, 40, 6651. [25] Soares, B.G.; Sirqueira, A.S.; Oliveira, M.G.; Almeida, M.S.M. Macromol. Symp. 2002, 189, 45. [26] Soares, B. G.; Almeida, M. S. M.; Guimarães, P. I. C. Eur. Polym. J. 2004, 40, 2185. [27] Soares, B.G.; Gomes, A.C.O.; Moreira, V.X.; Oliveira, M.G. Rubber Chem. Tech. 2005, 78, 1. [28] Oliveira, M. G.; Soares, B. G. J. Appl. Polym. Sci. 2003, 90, 2408. [29] Fornes, T.D.; Paul, D.R. Polymer 2003, 44, 3945. [30] Chang, M. C. O.; Thomas, D. A.; Sperling, L. H. J. Appl. Polym. Sci. 1987, 34, 409. [31] Lin, M. S.; Lee, S. T. Polymer 1997, 38, 53. [32] Perera, M. C. S; Ishiaku, U. S.; Ishak, Z. A. M. Eur. Polym. J. 2001, 37, 167.