Extruded co-continuous polyester–polyalkene– carbon black composites Influence of polyester glass transition on electrical properties I. Pillin, S. Pimbert, J.-F. Feller, and G. Levesque Co-continuous extruded polyester–carbon black filled polyalkene blends exhibit a slightly positive temperature coefficient, which depends on the nature of the polyester. From studies on various polyesters (poly(butylene terephthalate), poly(ethylene terephthalate), and some of their blends), it appears that the positive temperature coefficient becomes higher when the blends are heated above the glass transition temperature of the polyester. Polyalkene melting induces a hysteresis effect between heating and cooling steps. Attempts to catalyse grafting through enhanced transesterification also promote polyester degradation and alter the morphologies of the blends. PRC/1830 © 2002 IoM Communications Ltd. The authors (
[email protected]) are in the Laboratoire Polyme`res et Proce´de´s, Universite´ de Bretagne-Sud, Rue de Saint-Maude´, BP 92116, 56321 Lorient, France. Manuscript received 12 March 2002; accepted in final form 4 June 2002.
INTRODUCTION Generally, polymers are good insulators. However, the presence of conductive fillers such as carbon black (CB) or metallic particles can produce conducting materials with variable conductivity. A minimum concentration of conductive filler appears to be necessary to obtain a conductive material. This behaviour is usually described by percolation theories and the critical content required to produce conductivity is called the percolation threshold.1,2 In 1945, Frydman discovered the positive temperature coefficient (PTC) phenomenon in CB filled polyethylene (PE),3 in which the material’s resistivity increases with increasing temperature. This behaviour is due to matrix expansion,4 or variation in crystallinity5,6 and also to the nature of the fillers. Kolher4 concluded that the non-linearity observed in the relationship between resistivity and temperature results from a difference in expansion coefficient between the insulating matrix and the conductive filler network. Polymer expansion increases CB particle spacing and induces increasing gaps between conductive particles, leading to local disconnection in the conductive network. According to Heaney,5 the PTC effect in semicrystalline polymers can be attributed mainly to melting of the crystalline phase, leading to a greater dilution of CB aggregates and consequently to the observed conductivity decrease. In two phase polymer blends, the CB is often localised preferentially in one of the polymer phases and/or at the interface, so that the conductivity of the blend depends mainly on blend morphology. In co-continuous extruded blends, the CB content required to reach the percolation threshold is generally reduced. In polyethylene–poly(ethylene-co-vinyl acetate)–CB blends (PE–EVA–CB), variations in conductivity with blend composition are observed because CB particles are located preferentially in one phase.7 In these blends, the percolation threshold occurs between 3·6 and 4·2% CB, whereas 8 and 20% are needed in pure DOI 10.1179/146580102225005018
prc0001830 24-06-02 11:36:11 Rev 14.05 The Charlesworth Group, Huddersfield 01484 517077
PE and EVA, respectively.8 The percolation threshold is reduced from 5% in high density (HD) PE to 3% in 45HDPE–55PS co-continuous blends and to 0·4% when the CB is localised near the interface after a specific thermal treatment. Moreover, in some cases, the introduction of CB into immiscible polymer blends is able to prevent coalescence of drops and morphologies are stabilised.9,10 Poly(butylene terephthalate) (PBT) is well known for its good mechanical properties and high melting temperature (~225°C).11 Recently, the authors used PBT in blends with CB filled poly(ethylene-co-ethylacrylate) (EEA) to obtain co-continuous materials. The electrical conductivity and PTC behaviour of these systems were investigated. A change in the slope of the curve of log resistivity versus temperature was observed in the temperature range corresponding to the glass transition of PBT.1 Compared with PBT, poly(ethylene terephthalate) (PET) is characterised by a higher glass transition temperature; PBT and PET are known to be miscible in the amorphous phase.12 The present work is devoted to the modification of the previously investigated PBT–CB filled EEA composite by using PET or a PBT–PET blend as the matrix instead of PBT. First, attention was given to PBT–PET blends with variable PET contents to characterise the effects on the glass transition temperatures and crystallinities of these blends. The conductive properties of PBT–PET–CB filled EEA copolymer composite were then studied. Dibutyl tin oxide, which is known to be a good catalyst for the transesterification reaction,13–17 was also used to graft the polyester matrix onto EEA and to modify blend morphologies. EXPERIMENTAL Materials The PBT used was Vestodur 3000 from Hu¨ls, which has a Viscosity number (phenol/1,2-dichlorobenzene Plastics, Rubber and Composites 2002 Vol. 31 No. 7 1
2 Pillin et al. Extended co-continuous polyester–polyalkene–carbon black composites
1 : 1) of 165±7 cm3 g−1; density at 23°C d=1·31± 0·03 g cm−3; melting temperature T =221–226°C; m glass transition temperature T = 45°C; Young’s g modulus E2200 MPa; M =94 000 g mol−1 (data w from supplier). The PET was Arnite D04-300 from DSM (France) with an intrinsic viscosity of 82 cm3 g−1; density (23°C) d=1·34 g cm−3; melting temperature T =250°C; glass m transition temperature T =82·5°C; melt volumetric g index MVI=30·9 cm3/10 min at 250°C; 2·16 kg load. The EEA containing carbon black (EEAC) was supplied by Borealis (E/EA ratio=85 : 15, EEA/CB ratio=63 : 37).* Mixing procedure The polymers were melt mixed in an internal mixer (Brabender, 50 EHT) controlled by a Lab-Station driven by the Brabender software Winmix. The same conditions were used for each blend: 280°C and 10 rev min−1 for 15 min. The PBT and PET were dried for 24 h at 210°C before blending. No atmosphere control was used during blending. The PBT–EEA blend ribbons were extruded using a twin screw extruder (Brabender, DSK 42/6) controlled by a Lab-Station with a screw rotation speed of 30 rev min−1 and temperatures of 150, 240, and 260°C from hopper to slot die (4×50 mm2). The temperatures were 150, 260, and 280°C for the PET– PBT–EEA blends. The PBT–PET–EEAC blends were prepared in two steps: first, extrusion of a 30PBT–70PET compound, followed by grinding, then blending with EEAC at a 60 : 40 ratio to obtain strips for conductivity measurements. The 40 wt/wt ratio of EEAC was used for each blend. Electrical property measurements Test samples (10×1×0·2 cm3) were cut from the extruded bands. A 2 V continuous voltage V was applied between the sample extremities, along the length of the specimen. The current I was measured and the resistivity calculated from the relation V =RI. Measurements were performed in a heated oven under thermal cycling between 40 and 140°C (heating rate 0·6 K min−1). The test apparatus was controlled by a computer for data acquisition and treatment (Fig. 1). Curves of resistivity r and power P v temperature T were recorded. In Fig. 1, the graphs obtained during thermal cycling have double curves because of the thermal lag caused by melting and crystallisation of the polymers. Differential scanning calorimetry Thermograms were obtained using a Perkin Elmer Pyris 1 differential scanning calorimeter with the Pyris V 3·0 software for data collection and treatment. Calibration was performed with indium and tin in the temperature range +15 to +350°C; the base line was checked every day. Aluminum pans with holes in their lids were used and the mass of the samples was ~10 mg. All samples were first heated to 250°C (for PBT) or 280°C (for PET) for 5 min to remove * All ratios are wt/wt unless otherwise indicated. Plastics, Rubber and Composites 2002 Vol. 31 No. 7 prc0001830 24-06-02 11:36:11 Rev 14.05 The Charlesworth Group, Huddersfield 01484 517077
1 Apparatus for resistivity measurements on extruded samples
the effects of thermal history. All of the temperatures measured from a peak value (T , T ) were determined c m to within less than ±0·5°C. Non-isothermal crystallisation and melting temperatures T and T , respectively, were determined c m from the crystallisation peak values in experiments at ±20 K min−1 heating/cooling rates. Subsequent melting temperatures were obtained from the melting peak maxima measured at a heating rate of 20 K min−1. Melting enthalpies were determined using constant integration limits. Dynamic mechanical measurements The dynamic mechanical analyser used was a DMA 2980 from TA Instruments. Film samples (0·2 mm thickness, 10 mm width, 30 mm length) were tested in tensile mode. Measurements were performed from −80°C to 100°C at a heating rate of 3 K min−1 with a strain amplitude of 5 mm and a frequency of 1 Hz. Scanning electron microscopy Blend morphologies were observed in samples fractured in liquid nitrogen, after EEA extraction in boiling xylene. The corresponding surfaces were vacuum metallised and examined by scanning electron microscopy, using a JEOL JSM-6031F controlled by the SemAfore pro3.0 software. The accelerating voltage was 9 kV. RESULTS AND DISCUSSION Thermal properties of PBT–PET blends Before studying more complex blends, it was necessary to characterise PBT–PET blends formed from the
Pillin et al. Extended co-continuous polyester–polyalkene–carbon black composites 3
2 Glass transition temperature of PBT–PET blend v PET content
selected samples. Variations of PBT–PET blend glass transition temperature as a function of blend composition are shown in Fig. 2. At any given PBT/PET ratio, only one glass transition is observed, intermediate between the T values of PBT and PET. These data g agree with the results of Escala et al.12 and correspond to the formation of blends that are partially miscible, at least in the amorphous phase. These PBT–PET blends can then be used as matrix materials with adjustable glass transition temperatures. Results on blend crystallisation under the conditions stated are shown in Fig. 3: only one non-isothermal crystallisation temperature is observed, whatever the PET content is, whereas two melting peaks, near the melting temperatures of PBT and PET, are observed up to 60% PET. It can be concluded that the two crystallisation peaks are superimposed because a shoulder is observed in several DSC curves. Nevertheless, the 50PBT–50PET blend offers the lowest crystallisation temperature, corresponding to the slowest crystallisation rate. For PET contents increasing from 20 to 60%, it can be seen that the melting temperature of PBT increases whereas the melting temperature of PET decreases (Fig. 4). This result can be attributed to lower crystallisation temperatures due to the presence of thinner crystalline lamellae. These data are in good agreement with the previously postulated miscibility of PBT with PET in both the amorphous and molten states.12 In Fig. 5, half crystallisation times t are reported 1/2 as a function of isothermal crystallisation temperatures T . The data show that pure PET exhibits the c fastest overall crystallisation kinetics, owing to a larger undercooling due to its higher melting temperature.
3 Non-isothermal crystallisation temperatures of PBT–PET blend v PET content
The 50PBT–50PET blend offers half crystallisation times that are not very different from the values observed for pure PBT in the 205–210°C temperature range. By contrast, the crystallisation rate of the 30PBT– 70PET blend is higher than the crystallisation rate of pure PBT and gives t values that are very near 1/2 value for value. This would allow shorter processing times than 50PBT–50PET blends, which exhibit both the lowest anisothermal crystallisation temperature and the slowest crystallisation rate. Finally, the 30PBT–70PET blend appears to be a good compromise between a high enough glass transition temperature, a melting temperature that is slightly lower than that of PET, and a crystallinity ratio ~25% lower than in pure PET or PBT. This blend has been used as a matrix of conductive polymer composite (CPC) instead of PBT. Thermomechanical properties of PBT– PET–EEAC blends Substitution of pure PBT with a 30PBT–70PET blend as the matrix gives a system with a higher nonisothermal crystallisation temperature (194·8°C rather than 188·5°C) and faster crystallisation rate than the pure PBT matrix. The melting point increases from 222·2°C in pure PBT to 240·2°C in the 30PBT–70PET blend (Table 1), which could allow slightly higher service temperatures for the final heating device. The thermomechanical properties of extruded PBT– PET–EEAC blends have been compared in order to study the influence of the matrix: PET, PBT, or 30PBT–70PET. Measurements were performed in
Table 1 Thermal properties of extruded 60PE–40EEAC blends EEA
Polyester
Crystallisation
Melting
Polyester
T , °C c
DH, J g−1
T , °C m
PBT 30PBT–70PET PET
81·9 81·9 81·9
−26·7 −22·0 −16·7
97·3 97·6 97·9
Crystallisation
Melting
DH, J g−1
T , °C c
DH, J g−1
T , °C m
DH, J g−1
28·2 21·7 18·0
188·5 194·8 197·8
−50·3 −47·3 −49·2
222·2 240·2 243·2
54·7 54·8 59·0
Plastics, Rubber and Composites 2002 Vol. 31 No. 7 prc0001830 24-06-02 11:36:11 Rev 14.05 The Charlesworth Group, Huddersfield 01484 517077
4 Pillin et al. Extended co-continuous polyester–polyalkene–carbon black composites
4 Melting temperatures of PBT and PET v PET content in PBT–PET blends
6 log resistivity v temperature for 60PBT–40EEA extruded samples; straight line corresponds to PBT mechanical transition (DMA)
with −33·0°C in pure EEA. This could be correlated to variations of polymer molecular mobility in the blends as shown by large differences in viscosity between pure EEA and EEA–polyester blends observed through rheological measurements.18
5 Overall crystallisation half-time t v isothermal 1/2 crystallisation temperature for PBT, PET, and PBT–PET blends
the longitudinal direction on extruded samples and the resulting data are given in Table 2. It can be seen that all these blends offer quite high storage modulus values, both at −80°C and 20°C, without any significant differences due to the choice of matrix. On the other hand, the glass transition temperature of the matrix increases from 66·5 to 96·6 and 97·5, respectively, for pure PBT, 30 : 70 blend, and pure PET. The EEA relaxation temperature (b transition) seems to be slightly lower in PET containing blends: T =−26·7, b −41·3, and −38·3°C, respectively in PBT–EEAC, PBT–PET–EEAC, and PET–EEAC blends, compared Table 2 Thermomechanical properties of extruded 60PE–40EEAC blends* Storage modulus, MPa
T from tan d , g max °C
Polyester
−80°C
20°C
EEA
Polyester
PBT 30PBT–70PET PET
2900 2900 2500
1550 1850 1550
−26·7 −41·3 −38·3
66·5 96·6 97·5
*frequency=1 Hz, strain=5 mm, heating rate 3 K min−1. Plastics, Rubber and Composites 2002 Vol. 31 No. 7 prc0001830 24-06-02 11:36:11 Rev 14.05 The Charlesworth Group, Huddersfield 01484 517077
Electrical properties Resistivity measurements as a function of temperature are shown in Fig. 6 for 60PBT–40EEAC blends (stabilised cycle). The main change in the slope of the curve during heating seems to correspond to the change in the thermal expansion coefficient of PBT during heating through its glass transition. The small hysteresis behaviour observed between the heating and cooling steps is correlated to the difference between EEA melting and crystallisation temperatures, which induces morphological modifications in the CB filled polyalkene phase. However, the low crystallinity of EEA (~9%) could explain why the resistivity variations observed during EEA phase transitions remain relatively small when compared with the general tendency, which is clearly controlled by the thermal expansion of the polyester matrix, and particularly when compared to what is observed in PBT–PE conductive blends for example. In Fig. 7, the normalised resistivity log(r /r ) T 40 is expressed as a function of temperature. For the three matrixes used, the curves present quite similar shapes, and show an inflexion slightly above the glass transition of the polyester matrix. Straight line extrapolations show inflexions in the slope near 70°C (PBT), 83°C (70PBT–30PET), and 87°C (PET). Between 100 and 110°C, all heating curves exhibit a ‘positive hump’ corresponding to melting of the EEA, which is the phase in which the conductive CB is dispersed.1 For higher temperatures, the slopes of curves are very similar for PBT or PET matrixes, whereas the PBT– PET blend, which has a lower crystallinity ratio, gives a slightly higher slope. In these CB filled polyester–polyalkene thermoplastic blends, it seems that most of the observed temperature induced conductivity variations are controlled directly by the glass transition and crystallinity of the polyester matrix, whereas melting of the polyalkene has only a limited influence, but leads to significant hysteresis.
Pillin et al. Extended co-continuous polyester–polyalkene–carbon black composites 5
7 Normalised resistivity for several 60PE–40EEAC blends, heating curve only: (+) PBT–EEAC; (#) PBT–PET–EEAC; (×) PET–EEAC
Morphologies of multiphase conductive polymer composites Influence of matrix nature The morphologies of extruded samples were examined after fracture in liquid nitrogen and extraction of EEA with boiling xylene. In Fig. 8, scanning electron micrographs of the remaining polyester matrix are shown for PBT, 30PBT–70PET, and PET blends. The morphologies are quite similar: the number and size of the cylindrical solvent etched domains seem to remain nearly constant. Influence of dibutyltin oxide Small amounts of dibutyltin oxide (DBTO), a well known transesterification catalyst, were introduced before extrusion into polyester–EEAC composites in
order to evaluate its influence on blend morphologies and consequently on the electrical properties of the CPC. Addition of DBTO has no influence on the melting and crystallisation temperatures of EEA and polyester (Table 3). Thermomechanical measurements, made at −80°C in the extrusion direction, show that small amounts of catalyst (0·01% DBTO) induce an increase of storage modulus from 2900 MPa without DBTO to 4400 and 4300 MPa, respectively, for catalysed PBT–EEAC and PBT–PET–EEAC blends (Table 4). These modifications provide evidence of grafting or crosslinking reactions between ester groups of both the polyester matrix and EEA copolymer. At 20°C, for the 60PBT–40PET–EEAC blend, the storage modulus increased from 1850 MPa without catalyst to 2200 MPa with 0·01% catalyst. Moreover, a recent study on PBT–EEA blends19gave evidence for a limited miscibility of the PBT short chains with EEA copolymer, inducing an increase of storage moduli. After EEA extraction, SEM observations on these DBTO catalysed blends (Fig. 9) show important modifications in the morphology. The diameters of the EEA domains are much larger, resulting from the catalyst action which promotes coarser morphologies. These results contrast with those of an earlier study,16 in which catalyst addition results in more homogeneous morphologies and a better phase dispersion. However, some polyester molecular degradation in the presence of DBTO could be responsible for the evolution of the morphology, as a result of changes in viscosity ratio between the two polymer phases. Resistivity measurements conducted on these samples show that the addition of DBTO to a 60PBT– 40EEAC blend induces an increase in the material’s conductivity (Fig. 10), without influencing the overall shape of the curve. Such behaviour seems to be correlated to the modification of morphology, and particularly to the presence of larger EEAC domains, which create the electrically conductive network.
Table 3 Thermal properties of extruded 60PE–40EEAC blends with transesterification dibutyltin oxide Crystallisation Polyester
Bu SnO, % 2
T
PBT PBT PBT PBT 30PBT–70PET 30PBT–70PET
0 0·01 0·05 0·2 0 0·01
81·9 81·6 81·6 81·2 81·9 80·9
c,EEA
, °C
Melting
DH, J g−1
T , °C c,PBT
DH, J g−1
T , °C m,EEA
DH, J g−1
T
−26·7 −27·5 −25·5 −23·7 −22·0 −25·1
188·5 188·2 188·5 189·8 194·8 192·8
−50·3 −43·7 −52·2 −56·0 −47·3 −30·3
97·3 97·7 98·3 98·6 97·6 98·3
28·2 29·5 27·5 24·5 21·7 29·9
222·2 224·3 224·6 225·9 240·2 240·9
m,PBT
, °C
DH, J g−1 54·7 48·4 52·3 56·8 54·8 31·6
Table 4 Thermomechanical properties of extruded 60PE–40EEAC blends with dibutyltin oxide* Storage modulus E∞, MPa
tan d
max
, °C
Polyester
Bu SnO, % 2
at −80°C
at 20°C
EEA
Polyester
PBT PBT 30PBT–70PET 30PBT–70PET
0·05 0·2 0 0·01
2700 3800 2900 4300
1600 2100 1850 2200
−30·3 −30·1 −41·3 −35·9
69·1 71·6 96·6 92·6
*f=1 Hz, strain=5 mm, heating rate=3 K min−1. Plastics, Rubber and Composites 2002 Vol. 31 No. 7 prc0001830 24-06-02 11:36:11 Rev 14.05 The Charlesworth Group, Huddersfield 01484 517077
6 Pillin et al. Extended co-continuous polyester–polyalkene–carbon black composites
8 Scanning electron micrographs after EEA extraction: a 60PBT–40EEAC blend; b 60(30PBT–70PET)– 40EEAC blends; c 60PET–40EEAC blend (×3000)
9 Scanning electron micrographs after EEA extraction: a 60PBT–40EEAC blend; b 60PBT–40EEAC with 0·5% DBTO; (a,b×1000; c,d×10 000)
CONCLUSIONS Substitution of pure PBT with a 30PBT–70PET blend as the matrix in co-continuous CB filled composites gives a higher transition temperature, together with a higher melting temperature and a decrease in crystallinity. The relationships between resistivity and temperature follow quite similar patterns for the three matrixes used, showing slight inflexions in the slope above the glass transition of the polyester matrix. Linear extrapolation gives inflexion points near 70°C (PBT), 83°C (PBT–PET blend), and 87°C (PET).
Between 100 and 110°C, EEA melting/crystallisation induces hysteresis correlated to the polyalkene low crystallinity ratio. For higher temperatures, the slope of the resistivity curve depends slightly on the crystallinity of the polyester matrix. Addition of small amounts of a transesterification catalyst to these blends induces grafting reactions, resulting in an increase in both storage modulus and conductivity. However, dibutyltin oxide does not promote finer morphologies as expected, although the grafting or crosslinking are responsible for modifications in the morphology.
REFERENCES
10 Influence of catalyst content on conductivity of 60PBT–40EEAC blend: (#) without DBTO; ($) 0·2% DBTO Plastics, Rubber and Composites 2002 Vol. 31 No. 7 prc0001830 24-06-02 11:36:11 Rev 14.05 The Charlesworth Group, Huddersfield 01484 517077
1. . , . . , . , and . : Polym. Eng. Sci., 41, 1124–1132, 2001. 2. . , . , . , . , . , . , . , and . . : Macromolecules, 1994, 27, 1972–1974. 3. . : Improvement in or relating to resistance elements having positive temperature characteristics, 16/11/1945, GB, 604 695. 4. . : Resistance element, 13/11/1962, 3 243 753. 5. . . : Physica A, 1997, 241, 296–300. 6. . : Polym. Eng. Sci., 1973, 13, 463–468. 7. . . , . , . . , . . , and . . : Macromolecules, 1998, 31, 6724–6726. 8. . . : J. Polym. Sci. B, Polym. Phys., 1999, 37, 1899–1910. 9. . , . , . , and . : Eur. Polym. J., 1991, 27, 1261–1264. 10. . , . , . , . , . , . , . . , and . : J. Phys D, Appl. Phys., 1999, 32, 1517–1525. 11. . . : ‘Thermal characterization of polymeric materials’; 1997, New York, Academic Press. 12. . and . . : Adv. Chem. Series, 1979, 176, 455–487.
1 2
Pillin et al. Extended co-continuous polyester–polyalkene–carbon black composites 7
13. . , . . , . . , and . : Polymer, 1994, 35, 758–764. 14. . , . , . . , and . : Int. J. Polym. Anal. Char., 1995, 1, 137–157. 15. . , . , . , . . , and . : J. Organomet. Chem., 1994, 481, 7–17. 16. . , . . , . , and . : J. Appl. Polym. Sci., 1997, 65, 2457–2469.
17. . , . . , . . , and . : Polymer, 1997, 38, 5085–5089. 18. . , . , and . : unpublished work, Universite´ de Bretangne-Sud, 2002. 19. . , . , . . , and . : Proc. 17th Annual Meeting, Montreal, Canada, May 2001, Polymer Processing Society.
Plastics, Rubber and Composites 2002 Vol. 31 No. 7 prc0001830 24-06-02 11:36:11 Rev 14.05 The Charlesworth Group, Huddersfield 01484 517077
3 4
Authors Queries Journal: Plastics, Rubber and Composites Paper: 1830 Dear Author During the preparation of your manuscript, the queries listed below have arisen. The text to which the queries pertain is indicated on the proof by a numbered box in the margin. Please answer the queries and return this form with your corrected proof. Many thanks for your co-operation. Query Refs
Query
1
What is this – a stanard, a patent or what? Please give full details
2
What is this – a stanard, a patent or what? Please give full details
3
Are these details correct?
4
Please give page or paper number
prc0001830 24-06-02 11:36:11 Rev 14.05 The Charlesworth Group, Huddersfield 01484 517077
Remarks
