Filler-content Dependence of Dielectric Properties of Low-Density Polyethylene/MgO Nanocomposites

Transcription

1 Extended Summary 本文は pp Filler-content Dependence of Dielectric Properties of Low-Density Polyethylene/MgO Nanocomposites Toshiaki Kikuma Student Member (Waseda University, Norikazu Fuse Student Member (Waseda University, Toshikatsu Tanaka Senior Member (Waseda University, Yoshinao Murata Member (J-Power Systems, Yoshimichi Ohki Member (Waseda University, Keywords : polyethylene, nanocomposite, conduction current, space charge, permittivity Recently, polymer nanocomposites (NCs) are gaining new attraction as insulating materials, since their superior dielectric properties are being recognized. In the present paper, basic dielectric properties of NCs of low-density polyethylene (LDPE) and magnesium oxide (MgO) nano-fillers with various filler contents are examined. The samples tested were LDPE/MgO NC sheets with a thickness of about 1 µm, which contain 1, 5, and 1 phr (parts per hundred parts of resin by weight) MgO particles with average and max diameters of 5 and nm, respectively. As a reference, LDPE without MgO particles was also used. Instrumental analyses were carried out by ultraviolet-visible absorption spectroscopy (UVA), differential scanning calorimetry (DSC), X-ray diffraction (XRD) spectroscopy, and Fourier-transform infrared absorption (FT-IR) spectroscopy. Permittivity was measured at temperatures from to 1 C for frequencies from 3 to 1 5 Hz. Conduction current was measured under dc electric field of 5 kv/mm. Space charge distribution was measured at room temperature by the pulsed electroacoustic method. Optical transparency first increases by the addition of MgO fillers up to 1 to 5 phr, and then it decreases. Thermal properties and morphology are insensitive to the filler addition, as far as those analyzable by XRD, FT-IR, and DSC are concerned. Figure 1 shows relative permittivity measured at 3 Hz as a function of temperature. The permittivity shows minimum at 1-phr addition, and it increases with a further increase in filler content. Figure shows temperature dependence of the conductivity. The conductivity lowers by the filler addition up to 5 phr, but increases at 1 phr, at temperatures from 5 C to 9 C. Figure 3 shows total positive homocharge amount remaining in the bulk after short-circuit, that followed 1-min dc voltage application, as a function of time. It is the largest in the sample with 1-phr fillers. Fig.. Conductivity (S/m) : Base LDPE, : 1 phr MgO, : 5 phr MgO, : 1 phr MgO Temperature dependence of the corrected conductivity Charge (nc) : Base LDPE, : 1 phr MgO, : 5 phr MgO, : 1 phr MgO Fig. 1. Relative permittivity measured at 3 Hz as a function of temperature : Base LDPE, : 1 phr MgO, : 5 phr MgO, : 1 phr MgO Fig. 3. Total charge amount remaining after short-circuit as a function of time -7-

2 Paper Filler-content Dependence of Dielectric Properties of Low-Density Polyethylene/MgO Nanocomposites Toshiaki Kikuma * Norikazu Fuse * Toshikatsu Tanaka ** Yoshinao Murata *** Yoshimichi Ohki * Student Member Student Member Senior Member Member Member This paper reports measurement results of instrumental analyses, permittivity, conduction current, and space charge distribution profiles observed in low-density polyethylene/mgo nanocomposites with various filler contents, done with the intention to clarify the filler-content dependence of dielectric properties. The permittivity shows the lowest value in the sample with 1-phr fillers and then increases monotonically with an increase in filler content. The conductivity measured under dc electric field lowers by the filler addition up to 5 phr at temperatures from 5 to 9 C. In all the samples, positive homocharge is formed in the vicinity of the anode, and its amount is largest at 1 phr. The most adequate filler content for the electrical insulation purpose is around 1 to 5 phr. Keywords : polyethylene, nanocomposite, conduction current, space charge, permittivity 1. Introduction Polymer nanocomposites (NCs), in which only a small quantity of nm-sized inorganic fillers are dispersed, have been attracting much attention as new engineering plastic materials since various mechanical and thermal properties are improved (1). Recently, NCs are gaining new attraction as insulating materials, since their superior dielectric properties are being recognized (1)-(4). In the present paper, basic dielectric properties of NCs of low-density polyethylene (LDPE) and magnesium oxide (MgO) nano-fillers with various filler contents are examined.. Experimental The samples tested were LDPE/MgO NC sheets with a thickness of about 1 µm, which contain 1, 5, and 1 phr (parts per hundred parts of resin by weight) MgO particles with average and max diameters of 5 and nm, respectively (5). As a reference, LDPE without MgO particles was also used. They are called PE-, PE-1, PE-5 and PE-1, according to their particle contents. Instrumental analyses were carried out by ultraviolet-visible absorption spectroscopy (UVA), differential scanning calorimetry (DSC), X-ray diffraction (XRD) spectroscopy, and Fourier-transform infrared absorption (FT-IR) spectroscopy. Permittivity was measured at temperatures from to 1 C for frequencies from 3 to 1 5 Hz using the samples with an upper and a lower evaporated circular Au electrode 5 mm in the * Department of Electrical Engineering and Bioscience, Waseda University 3-4-1, Ohkubo, Shinjuku, Tokyo ** Graduate School of Information, Production and Systems, Waseda University -7, Hibikino, Wakamatsu, Kitakyushu *** J-Power Systems Corporation 5-1-1, Hitaka-cho, Hitachi effective diameter. Conduction current was measured using the samples with an upper and a lower evaporated circular Au electrode mm in the effective diameter. Space charge distribution was measured at room temperature for the samples sandwiched between an Al electrode and a semiconductive (SC) polymer electrode by the pulsed electroacoustic method with an acoustic sensor attached to the Al electrode. 3. Results and Discussion 3.1 Instrumental Analyses Figure 1 shows the UVA spectra. The unfilled sample PE- shows relatively large absorption. Compared to PE-, PE-1 is more transparent. However, the samples with higher filler contents PE-5 and PE-1 become less transparent. This result is consistent with observation by eyes. Namely, PE-1 looks clearest, while PE-1 is the haziest with milky white in color. The reason why PE- is translucent seems likely that light scattering occurs at the interfaces between crystalline and amorphous regions. It is known that nanofillers in polymer act as nuclei for crystal formation (6). Therefore, addition of well-dispersed fillers should make the crystal size smaller. If PE-1 has crystals smaller than the wavelengths of visible light, it becomes transparent. In the samples with higher quantities of fillers, PE-5 and PE-1, there is a possibility that the fillers act as light scattering centers. There is also another possibility that mixing of fillers was not sufficient enough to obtain optically uniform films in the cases of PE-5 and PE-1 due to their high contents of fillers. Whatever the reason is, the transparency becomes better first and worse later as the filler content increases. This tendency that a certain property is improved only by a proper addition of a small quantity of fillers seems important since a similar tendency is observed in other experimental results such as permittivity and conductivity, as will be discussed later. 17 IEEJ Trans. FM, Vol. 16, No.11, 6

3 Dielectric Properties of LDPE/MgO NCs Absorption Coefficient (cm -1 ) PE- PE-1 PE-5 PE Photon Energy (ev) : PE-, : PE-1, : PE-5, : PE-1 Fig. 1. Ultraviolet-visible absorption spectra 3 Absorbance Wavenumber (cm -1 ) : PE-, : PE-1, : PE-5, : PE-1 Fig. 3. FT-IR spectra 1 Intensity θ (deg) : PE-, : PE-1, : PE-5, : PE-1 Fig.. XRD spectra. Note: All the spectra overlap with each other almost completely throughout the whole range Figure shows XRD spectra of the four samples. Note that all the spectra overlap with each other almost completely throughout the whole range. The peaks at 1.4 and 3.7 seen in all the samples are due to the (1) and () planes of LDPE respectively, while the one at 43. seen in PE-5 and PE-1 is due to the () plane of MgO. The appearance of the MgO peak in PE-5 and PE-1 is simply due to its high content of fillers. Therefore, no specific feature emerges by the addition of fillers. This is in contrast to polyamide nanocomposites in which a new crystalline peak appears (7)(8). Likewise, no features appear by the filler addition both in FT-IR and DSC spectra as shown in Figs. 3 and 4, respectively. Note that subtle differences in magnitude of FT-IR or DSC spectra among the samples are not important, since they could be due to trivial reasons such as a difference in sample thickness. To summarize the results shown in Figs. 1 to 4, no apparent change is induced by the filler at least beyond the scale detectable by the present instrumental analyses, although the crystalline size seems to become smaller. 3. Permittivity Figure 5 shows the relative permittivity as a function of frequency measured at temperatures from C to 1 C. Figure 6 shows the relative permittivity as a function of temperature measured at 3 Hz, replotted from Fig. 5. As the sample temperature becomes higher, the permittivity lowers monotonically in all the samples. As for the filler-content dependence, PE-1 has smaller permittivity values than PE- as shown in Fig. 6, although the difference is small. This result might appear doubtful, when the fact that the relative permittivity of MgO (=9.7 (9) ) is much higher than that of LDPE (=.3) is taken into account. The permittivity then increases according to the order of filler content in PE-5 and PE-1. This trend is always Heat Flow (mw) : PE-, : PE-1, : PE-5, : PE-1 Fig. 4. DSC spectra the case regardless of the measurement frequency. A similar tendency was also confirmed by a different research group with different measurement apparatus (1). Namely, the permittivity also showed the lowest at 1-phr addition and it increased with a further increase in the filler content for a similar set of four samples taken from a different lot measured by an LCR meter (1). Therefore, it is reasonable to assume that the result shown in Fig. 6 is reliable. It has been reported for many NCs that permittivity is decreased by the addition of fillers despite a higher permittivity of the fillers (11)(1). There is a possibility that the low permittivity is an indication of a good NC (13). In this context, Fig. 6 indicates that PE-1 is the best NC. 3.3 Conductivity Figure 7 shows decay characteristics of the current in each sample after dc voltage application measured at 5 C. The dc voltage applied was adjusted so that the electric field was 5 kv/mm for the samples with a thickness from 9 to 18 µm. The current magnitude is found to be PE-5 < PE-1 PE-1 < PE-. In order to examine the temperature dependence in more detail, the current was measured according to the following procedures. First, dc electric field of 5 kv/mm was applied to the sample at 5 C for 1 min. Since the current reached a stable steady-state conduction value, the sample temperature was lowered to 5 C at a rate of 6 C/min while the voltage remained on. Then, after the sample was held at 5 C for 5 min, it was heated to 9 C at a rate of 6 C/min and held at that temperature for 1 min. Finally, the sample was cooled to 5 C. Figure 8 (a) shows the conduction current behavior observed for PE- during the temperature increasing and decreasing periods. As can be seen, the current while the temperature is increasing is lower than that during the temperature decreasing. As shown in Fig. 6, the permittivity decreases as the sample temperature 電学論 A,16 巻 11 号,6 年 173

4 (a) PE (b) PE (c) PE-5 (d) PE-1 : C, : 3 C, : 4 C, : 5 C, : 6 C, : 7 C, : 8 C, : 9 C, : 1 C Fig. 5. Relative permittivity as a function of frequency at temperatures from to 1 C : PE-, : PE-1, : PE-5, : PE-1 Fig. 6. Relative permittivity measured at 3 Hz as a function of temperature increases. Namely, the sample capacitance changes during the temperature increasing and decreasing, which induces displacement current with the directions opposite and identical to the conduction current, respectively. Based on the results shown in Fig. 6, the displacement current is calculated to be about 7 pa for all the samples. Figure 8 (b) shows the curves shifted by adding or subtracting 7 pa to those shown in Fig. 8 (a). The two current curves in Fig. 8 (b) agree well with each other as long as they are above about 1 pa, the limit of reliable measurement. After the current curves measured for all the samples were corrected by the above-mentioned method, their conductivity values were calculated and are shown in Fig. 9. The order of the conductivity, PE-5 < PE-1 PE-1 < PE-, has been confirmed Current (pa) : PE-, : PE-1, : PE-5, : PE-1 Fig. 7. Decay characteristics of absorption current at 5 C after dc voltage application again. This result indicates that the optimum filler content is around 1 to 5 phr. A similar tendency was already observed in the above-mentioned transparency and permittivity measurements. Namely, for these properties, the addition of a small quantity of fillers is optimum. Furthermore, similar decrease in conductivity by adding a proper quantity of nanofillers has been reported for various NCs such as polyimide and polyamide NCs (4)(8). There is a possibility that molecular motion, which assists hopping-like carrier transport, is suppressed by a strong interaction at the matrix/filler interfaces. Note that lattice vibrations or phonons help charge carriers escape from traps by hopping, regardless of the carriers being electronic (i. e. electrons or holes) or ionic. Introduction of zigzag transport paths by the fillers could be a 174 IEEJ Trans. FM, Vol. 16, No.11, 6

5 Dielectric Properties of LDPE/MgO NCs Current (pa) (a) Before correction Charge Density (C/m 3 ) cathode (Al) anode (SC) Thickness (µm) (a) PE-1 (cathode: Al, anode: SC) Current (pa) (b) After correction : Increasing temperature, : Decreasing temperature Fig. 8. Conduction currents in PE- as a function of temperature Conductivity (S/m) Fig : PE-, : PE-1, : PE-5, : PE-1 Temperature dependence of the corrected conductivity reason, but this would be almost ineffective for electrons. 3.4 Space Charge Distribution Figure 1 shows space charge distribution profiles measured for PE-1 right after, 3-min after, and 1-min after the initiation of voltage application when the anode was either SC or Al. The applied voltage was dc and was adjusted so that the average electric field became 5 kv/mm according to the sample thickness from about 88 to 1 µm. Plenty of positive homocharge is seen before the anode regardless of its material. Similar presence of positive homocharge has been frequently reported for LDPE (14). The charge density with a triangular or trapezoidal distribution over a range from about -1 µm on the abscissa to about +15 µm consists not only of the charge on the anode induced by the voltage application, but also of the charge on the anode induced by the homocharge in the bulk. Since the induced charge has the Charge Density (C/m 3 ) - -4 anode (Al) cathode (SC) Thickness (µm) (b) PE-1 (anode: Al, cathode: SC) : right after, : 3-min after, : 1-min after the start of voltage application Fig. 1. Time-dependent changes in space charge distribution during voltage application Charge (nc) : PE-, : PE-1, : PE-5, : PE-1 Fig. 11. Amount of charge on the Al anode in each sample by the voltage application time polarity opposite to the homocharge, it reduces the total charge on the anode. By this reason, the charge amount observed on the anode is decreased as shown in Fig. 1. Therefore, the amount of homocharge present in the bulk can be estimated by the total amount of the anode charge. Figure 11 shows change in the total anode charge by the voltage application time for all the samples in the case of the Al anode that was set near the acoustic sensor, since the charge distribution was recorded more clearly in this case than in the other case where the anode was remote from the sensor. It is clear from Fig. 11 that the homocharge formation is the most active in PE-1. Space charge in a sample is generally more easily observed 電学論 A,16 巻 11 号,6 年 175

6 Charge Density (C/m 3 ) Al electrode SC electrode Thickness (Normalized) : SC anode, : Al anode Fig. 1. Space chrge distributions in PE-1 observed right after short-circuit following the 1-min application of dc voltage and decrease, respectively. () Thermal properties and morphology are insensitive to the filler addition, as far as those analyzable by XRD, FT-IR, and DSC are concerned. (3) The permittivity shows minimum at 1-phr addition, and it increases with a further increase in filler content. (4) The conductivity lowers by the filler addition up to 5 phr, but increases at 1 phr, in the temperature range from 5 C to 9 C. (5) The amount of positive homocharge in the vicinity of the anode is the largest in the sample with 1-phr fillers. (6) From the facts mentioned above, the addition of a small quantity of MgO fillers is very effective in decreasing the conductivity and permittivity. (Manuscript received Jan. 31, 6, revised July 7, 6) 1 Charge (nc) after the sample was short-circuited than during the voltage application as long as the charge has a relatively long life. Figure 1 shows space charge distribution profiles observed in PE-1 right after it was short-circuited, following the 1-min application of dc voltage. Note that the abscissa representing the thickness is normalized, because the sample thickness was slightly different for the two profiles. In both cases of the SC and Al anodes, the presence of positive homocharge in the bulk and negative countercharge on both electrodes is clear. Figure 13 shows the change in the total amount of positive homocharge in 1 minutes after the short-circuit, observed for each sample with the Al anode. As clearly shown in Fig. 13, the charge amount is in the order, PE-1 > PE- > PE-5 PE-1. The addition of 1-phr fillers shows again a unique characteristic. Presence of homocharge in the vicinity of electrodes usually suppresses carrier injection from the electrodes, by reducing the electric field intensity. Therefore, there is a possibility that the abundance in homocharge in PE-1 is a reason of its low conductivity. 4. Conclusion : PE-, : PE-1, : PE-5, : PE-1 Fig. 13. Total charge amount remaining after short-circuit as a function of time Instrumental analyses and measurements of dielectric properties were carried out for nanocomposites of low-density polyethylene and MgO. As a result, the following features have become clear. (1) Optical transparency first increases by the addition of MgO fillers up to 1 to 5 phr, and then it decreases. Miniaturization of crystal sizes and light scattering induced by the filler addition seem to be responsible for the transparency increase References (1) T. Tanaka, G. C. Montanari, and R. Mülhaupt : Polymer Nanocomposites as Dielectrics and Electrical Insulation Perspectives for Processing Technologies, Material Characterization and Future Applications, IEEE Trans. Dielect. Elect. Insulation, Vol.11, No.5, pp (4) () M. Kozako, N. Fuse, Y. Ohki, T. Okamoto, and T. Tanaka : Surface Degradation of Polyamide Nanocomposites Caused by Partial Discharges using IEC (b) Electrodes, IEEE Trans. Dielect. Elect. Insulation, Vol.11, No.5, pp (4) (3) G. C. Montanari, D. Fabiani, F. Palmieri, D. Kaempfer, R. Thomann, and R. Mülhaupt : Modification of Electrical Properties and Performance of EVA and PP Insulation through Nanostructure by Organophilic Silicates, IEEE Trans. Dielect. Elect. Insulation, Vol.11, No.5, pp (4) (4) Y. Cao, P. C. Irwin, and K. Younsi : The Future of Nanodielectrics in the Electrical Power Industry, IEEE Trans. Dielect. Elect. Insulation, Vol.11, No.5, pp (4) (5) Y. Murata, Y. Sekiguchi, Y. Inoue, and M. Kanaoka : Investigation of Electrical Phenomena of Inorganic-filler/LDPE Nanocomposite Material, Proce. of 5 International Symposium on Electrical Insulating Materials, Vol.3, pp (5) (6) S. Murase, A. Inoue, Y. Miyashita, Y. Nisio, and N. Kimura : Structural Characteristics and Moisure Sorption Behavior of Nylon-6/Clay Hybrid Films, J. Polymer Science Part B, Vol.4, No.6, pp () (7) K. Fujimoto, M. Yoshikawa, S. Katahira, and K. Yasue : Crystal Structure of Nylon6/Inorganic Layered Silicate Nanocomposite Film, Kobunshi Ronbunshu, Vol.57, No.7, pp () (in Japanese) (8) N. Fuse, M. Kozako, T. Tanaka, and Y. Ohki : Effects of Mica Fillers on Dielectric Properties of Polyamide Nanocomposites, 5 Annual Report Conference on Electrical Insulation and Dielectric Phenomena, pp (5) (9) D. R. Lide, Ed. : CRC Handbook of Chemistry and Physics, 74 th Edition, , p.1(41), CRC Press, Boca Raton, Ann Arbor, London, Tokyo (1993) (1) T. Tanaka, A. Nose, Y. Ohki, and Y. Murata : PD Resistance Evaluation of LDPE/MgO Nanocomposite by a Rod-to-Plane Electrode System, International Conference on Properties and Applications of Dielectric Materials (to appear in 6) (11) N. Fuse, M. Kozako, T. Tanaka, S. Murase, and Y. Ohki : Possible Mechanism of Superior Partial-Discharge Resistance of Polyamide Nanocomposites, 4 Annual Report Conference on Electrical Insulation and Dielectric Phenomena, pp.3-35 (4) (1) J. C. Fothergill, J. K. Nelson, and M. Fu : Dielectric Properties of Epoxy Nanocomposites Containing TiO, Al O 3, and ZnO Fillers, 4 Annual Report Conference on Electrical Insulation and Dielectric Phenomena, pp (4) (13) T. Tanaka : Dielectric Nanocomposites with Insulating Properties, IEEE Trans. Dielect. Elect. Insulation, Vol.1, No.5, pp (5) (14) G. Chen, T. Y. G. Tay, A. E. Davies, Y. Tanaka, and T. Takada : Electrodes and Charge Injection in Low-density Polyethylene -Using the Pulsed Electroacoustic Technique-, IEEE Trans. Dielect. Elect. Insulation, Vol.8, No.6, pp (1) 176 IEEJ Trans. FM, Vol. 16, No.11, 6

7 Dielectric Properties of LDPE/MgO NCs Toshiaki Kikuma (Student Member) received the B. Eng. degree in 5 from Waseda University and is currently in his M. Eng. degree course. Yoshinao Murata (Member) received the B.S. degree in physics from the Science University of Tokyo in 199. In the same year, he joined Hitachi Cable Ltd. Since then, he has been engaged in research and development of insulating materials for power cables. In 1, he moved to J-Power Systems Corporation and is presently a researcher of the Research & Development Center of that company. Norikazu Fuse (Student Member) received the B. Eng. and M. Eng. degrees in 3 and 5, respectively, from Waseda University and is currently in his Ph.D. degree course. He is also a Research Fellow of the Japan Society for the Promotion of Science. Toshikatsu Tanaka (Senior Member) is Professor at Graduate School of Information, Production, and Systems, Waseda University. He received the M. Sc. and Ph.D. degrees in materials science from Osaka University, in 1964 and 1968, respectively. He served as Vice President at CRIEPI from 1995 to 1, where he had worked for 37 years since He is a recipient of IEEJ A-Society Award, IEEE-DEIS Whitehead Award, Minister of Science and Technology of Japan Prize, and other awards. He is an IEEE Fellow. Yoshimichi Ohki (Member) is Professor at the Department of Electrical Engineering and Bioscience, Waseda University. He received the B. Eng., M. Eng., and Dr. Eng. degrees in 1973, 1975, and 1978, respectively, all from Waseda University. He joined the teaching staff of Waseda University in 1976 and is presently Professor. He was a Visiting Scientist at the Massachusetts Institute of Technology from 198 to 84. He is a recipient of IEEE-DEIS Forster Award, Best Paper Award of IEEJ ( times), and other awards. He is an IEEE Fellow. 電学論 A,16 巻 11 号,6 年 177