Journal of ELECTRONIC MATERIALS, Vol. 40, No. 11, 2011
DOI: 10.1007/s11664-011-1743-5 Ó 2011 TMS
Fabrication and Properties of Novel Polyetheretherketone/ Barium Titanate Composites with Low Dielectric Loss R.K. GOYAL,1,2 V.V. MADAV,1 P.R. PAKANKAR,1 and S.P. BUTEE1 1.—Department of Metallurgy and Materials Science, College of Engineering, Shivaji Nagar, Pune 411 005, India. 2.—e-mail:
[email protected]
Dielectric, thermal, and microhardness properties of high-performance barium titanate (BaTiO3)-filled polyetheretherketone (PEEK) composites were studied. BaTiO3 was varied from 0 vol.% to 67 vol.% in the PEEK matrix. The dielectric constant of the composites measured at 1 MHz increased approximately 14-fold. There was no dispersion in the dielectric constant with frequency between 10 kHz and 15 MHz. The Lichtenecker equation and modified Lichtenecker equation agreed well with the experimental data. The dissipation factor of the composites varied from 0.0056 to 0.0096. Scanning electron microscopy showed uniform dispersion of BaTiO3 in the matrix. The microhardness of the composites increased by more than 2.5-fold compared with pure PEEK. The coefficient of thermal expansion measured below and above the glass-transition temperature was reduced by up to 56%. These results make these composites promising candidate high-temperature organic substrates. Key words: Polymer–matrix composites (PMCs), electrical properties, thermal properties, powder processing
INTRODUCTION Barium titanate is a ferroelectric ceramic with the perovskite structure. It shows five crystal structures (hexagonal, cubic, tetragonal, orthorhombic, and rhombohedral) depending upon temperature. Except for the cubic structure, all the others exhibit ferroelectric properties. Its dielectric constant varies between 250 and 5000, depending on its grain size, purity, crystallographic direction, measuring temperature range, and method of preparation.1–3 It has low coefficient of thermal expansion, and high thermal and chemical stability.1,4 However, it is brittle and requires high processing temperatures. To overcome these drawbacks, several BaTiO3-filled composites with polymers such as epoxy,5,6 polyimide,4,7 polystyrene,8 cyanoethyl ester of polyvinyl alcohol (CEPVA),9 and polyvinylidene fluoride (PVDF)10,11 have been studied. These composites could be used for embedded capacitors,5 electronic packaging,6,9 capacitors/piezoelectric transducers,4 multilayer ceramic capacitors,1 microwave substrate applications,12 and piezosensitive and chem(Received March 28, 2011; accepted August 17, 2011; published online September 20, 2011)
2240
ocapacitive sensors.10,13 The dielectric constant of the composites can be tailored as per the requirement of the user. For example, the dielectric constant of 50 vol.% BaTiO3-filled epoxy composite prepared by roll coating was 24 (100 kHz).5 It was further improved significantly using a solution method.6 Similarly, 50 vol.% BaTiO3-filled polyimide composites prepared by a solution method and in situ polymerization showed dielectric constants of 23 and 35, respectively.4 According to Popielarz et al.,14 the dielectric constants of 30 vol.% BaTiO3filled 1,14-tetradecanedioldimethacrylate, trimethylolpropane triacrylate, and poly(ethylene glycol) diacrylate composites are 18, 24, and 40, respectively. In other words, for the same BaTiO3 loading, the dielectric constant of the composite strongly depends on the polarity of the polymer matrix. As a polymer matrix, polyetheretherketone (PEEK) has been widely used in many applications due to its outstanding thermal, mechanical, and chemical properties and high continuous-use temperature (250°C). Moreover, it is also an attractive material for electrical/electronic applications due to its self fire-retardant nature, high electrical resistance, and moisture resistance.15,16 Despite these properties,
Fabrication and Properties of Novel Polyetheretherketone/Barium Titanate Composites with Low Dielectric Loss
there are no reports about the dielectric, thermal expansion, and mechanical properties of high-performance PEEK/BaTiO3 composites, which are needed for high-temperature applications. In view of the above, for the first time, we have studied the dielectric properties of environmentally benign (bromine-free) PEEK/BaTiO3 composites fabricated by conventional hot pressing. The dielectric properties were studied as a function of BaTiO3 volume fraction and frequency. The experimental results were also correlated with the Lichtenecker and modified Lichtenecker equations. The composites were also evaluated for morphology, coefficient of thermal expansion (CTE), and microhardness. EXPERIMENTAL PROCEDURES Materials Commercial PEEK (grade 5300PF) donated by Gharda Chemicals Ltd. Panoli, Gujarat, India, under the trade name GATONEä PEEK was used as the matrix. It has a reported inherent viscosity of 0.87 dL/g measured at a concentration of 0.5 g/dL in H2SO4. BaTiO3 (BT) purchased from Aldrich Chemicals Company was used in the as-received condition. An SEM image of BT particles is shown in Fig. 1a. The differential particle size distribution of BT determined using a GALAI CIS-1 laser particle size analyzer is shown in Fig. 1b. It can be seen that the BT particle size ranges from 0.4 lm to 1.8 lm with a mean of 0.87 lm. As reported elsewhere,15 the mean particle size and particle size range of PEEK powder were 25 lm and 4 lm to 49 lm, respectively.
Fig. 1. SEM image (a) and particle size distribution (b) of BaTiO3 powder.
2241
Composite Preparation PEEK and BT powders were first dried in a vacuum oven at 200°C for 24 h. Then, appropriate weights of both powders were taken, varying BT from 0 wt.% to 90 wt.% (67 vol.%), and mixed together mechanically using a mortar and pestle for 2 h. The resultant mixed powder was further dried in an oven for a few hours. Then, the PEEK/BT composites were fabricated using hot pressing. The dried powder was filled in a tool steel die with 13 mm diameter. The powder was heated at an average heating rate of 9°C/min under a pressure of 45 MPa to a maximum temperature of 380°C. After a soaking period of 30 min to 60 min, the samples were aircooled to a temperature of 100°C and then ejected. Six different compositions containing 0 wt.%, 50 wt.%, 60 wt.%, 70 wt.%, 80 wt.%, and 90 wt.% BT in PEEK matrix were fabricated and characterized. For calculating the theoretical properties, the volume fraction (Vf) of BT particles required for a given weight fraction (Wf) can be determined using the equation Vf = Wf/[Wf + Wm(qf/qm)], where, qf is the density of the particles, and Wm and qm are the weight fraction and density of PEEK matrix, respectively. The wt.% and vol.% of BT added into the PEEK matrix are presented in Table I. Characterization of Composites The theoretical density of the samples was calculated by the rule of mixtures using the density of BT as 5.85 g/cc and of PEEK as 1.30 g/cc. The experimental density was measured by Archimedes’s principle, where the volume is measured by floating in an absolute alcohol medium. To examine the
2242
Goyal, Madav, Pakankar, and Butee 5
3
By Vol.
0 50 60 70 80 90
0 18.18 25.0 34.15 47.06 66.67
(b) 6 3 9
R2 = 0.9482
2
12
1
15 0
degree of dispersion of BT in the matrix, samples were prepared using the method described elsewhere.16 The prepared samples were coated with platinum using a sputter-coater (JEOL JEC560) and studied using SEM (JEOL JSM 6360A). The microhardness of well-polished samples was measured using a Vickers hardness tester (Future Tech Corp. FM-700, Tokyo, Japan) at a constant load of 100 g and a dwell time of 15 s. Average values of six readings are reported as the microhardness of the samples. The out-of-plane (through-thickness direction) linear CTE of the samples was determined using a PerkinElmer DMA 7e in thermomechanical analyzer mode. A 100 mN force was applied to ensure that the probe was in good contact with the sample. Before measuring the CTE, the samples were annealed in a vacuum oven at 260°C for 2 h. Then an annealed sample was held under pressure for 5 min and heated to 250°C at a heating rate of 5°C/min in argon atmosphere. The thermal strain, i.e., the ratio of change in length to the original sample length, was obtained during the second heating cycle. It increased nonlinearly with temperature. CTE values were calculated from the slope taken over specific temperature ranges of 50°C to 100°C (i.e., below glass transition) and 170°C to 220°C (above glass transition). The dielectric constant was determined using a Wayne Kerr Electronics precision impedance analyzer (6515B, UK) at frequencies varying from 10 kHz to 15 MHz at 30°C. The dielectric constant (e) was evaluated by the relation e = Ct/e0S, where S is the surface area, t is the thickness of the dielectric material, and e0 is the permittivity of free space (8.854 9 1012 F/m). The dissipation factor was obtained directly from the instrument. RESULTS AND DISCUSSION Density Figure 2 shows the theoretical and experimental density of the PEEK composites. PEEK has an experimental density of 1.30 g/cc. The experimental density of the composites containing up to 34 vol.% BT (BT-70) was close to the theoretical density, indicating that the samples were almost porosity free. However, the experimental density of the
Porosity (%)
By Wt.
Density (g/cc)
BT-0 BT-50 BT-60 BT-70 BT-80 BT-90
(a)
4
% BaTiO3 in PEEK Matrix Sample Code
0
(a) Theo. density (b) Exp. density
Table I. Weight % versus volume % of BaTiO3 in PEEK matrix
20
40
60
80
Vol % BaTiO 3 in PEEK matrix
Fig. 2. Theoretical and experimental density of PEEK/BaTiO3 composites. Solid line passing through filled triangles represents the linear trend for porosity with correlation factor R 2 = 94.8%.
47 vol.% and 67 vol.% composites was decreased by 11.6% and 14.1%, respectively. This may be attributed to the presence of voids, as confirmed from SEM images. Figure 2 also shows that voids or porosity increased linearly with increasing volume fraction of BT in the matrix. A good linear trend was achieved between the porosity values with correlation factor R2 > 94%. Scanning Electron Microscopy (SEM) of Composites Figure 3 shows SEM images of PEEK composites containing 18 vol.%, 34 vol.%, 47 vol.%, and 67 vol.% BT. It can be clearly seen that the BT particles are uniformly well distributed in the 18 vol.% and 34 vol.% composites. There are no aggregates or voids in these samples. Figure 3a, b shows an average size of BT particles below 1 lm. This supports the mean particle size determined using the GALAI CIS-1 laser particle size analyzer. However, Fig. 3c, d shows severe aggregates (2 lm) and voids in the samples. This is due to the fact that, as the BT particle loading increases, the interparticle distance decreases, which results in aggregates of BT particles. These aggregates resulted in voids and hence an experimental density lower than the theoretical density. Similarly, small aggregates of BT and voids were reported for 50 vol.% BT-filled epoxy composites.6 In contrast, 70 vol.% BT-filled polyimide composite prepared by high-shear spin coating and 50 vol.% surface-modified BT-filled polyimide composites did not show aggregates. According to those authors, this was due to the excellent compatibility between the polyimide and the BT particles.4 Microhardness of the Composites Figure 4 shows the microhardness of the composites as a function of BT content. The microhardness
Fabrication and Properties of Novel Polyetheretherketone/Barium Titanate Composites with Low Dielectric Loss
2243
Fig. 3. SEM images of composites with BaTiO3 content of: (a) 18 vol.%, (b) 34 vol.%, (c) 47 vol.%, and (d) 67 vol.%.
increased with increasing BT content. It increased due to the resistance to the plastic deformation of the PEEK matrix due to the hard BT particles. The microhardness of the 18 vol.%, 34 vol.%, and 47 vol.% composites increased from 24 kg/mm2 for pure PEEK matrix to 38 kg/mm2, 46.8 kg/mm2, and 60.7 kg/mm2, respectively. However, the lower hardness (i.e., 46.6 kg/mm2) of the 67 vol.% composite than that of the 47 vol.% composite may be attributed to the presence of voids, as confirmed by density and SEM analysis. Dielectric Properties of Composites Figure 5a, b shows the dielectric constant of the PEEK/BT composites as a function of logarithm of frequency and BT content, respectively. It can be seen clearly from Fig. 5a that the dielectric constant of the composites remains stable with varying frequency. Similar trends were reported by Devaraju et al.7 and Dang et al.8 The dielectric constant of pure PEEK measured at 1 MHz is 3.3. It can be seen from Fig. 5b that the dielectric constant of the
composites increases with increasing BT content. The increased dielectric constants are due to the higher polarization or dielectric constant of BT than that of the PEEK matrix. The dielectric constant and dissipation factor of the 47 vol.% composite are 26 and 0.0069, respectively. Similarly, the dielectric constant of 50 vol.% BaTiO3-filled epoxy and polyimide composite was 305 and 25,7 respectively. In contrast, the dielectric constant of 50 vol.% BT-filled CEPVA composites was 89. According to those authors, the significant increase in dielectric constant was due to the good dispersion of BT in the CEPVA matrix.9 It is to be noted that the higher dielectric constant of CEPVA/BT composites is probably due to the higher dielectric constant (22) of CEPVA. For example, the net increase in dielectric constant of CEPVA/BT (50 vol.%) composite is only 4-fold (i.e., 89/22 = 4.05), which is a much smaller improvement compared with the 8-fold improvement (i.e., 26/3.3 = 7.9) for the PEEK/BT (47 vol.%) composite. Similarly, a 9-fold improvement in dielectric constant was reported for polyimide/BT (50 vol.%) composite.4 Liang et al.
2244
Goyal, Madav, Pakankar, and Butee
reported a maximum dielectric constant value of 40 at 60 vol.% ceramic loading, and thereafter a decrease in dielectric constant was found due to the presence of porosity.17 However, in the present study the dielectric constant of 67 vol.% BT-filled PEEK measured at 1 MHz was 46.5 (i.e., a 14-fold improvement). This high dielectric constant is found despite the 14% porosity shown in Fig. 2. This might be because of better dispersion of the smaller BT particles in the matrix. The dissipation factor (tan d), which is the ratio of the real dielectric constant to the imaginary dielectric constant, limits the frequency selectivity. Figure 6 shows tan d of PEEK composites as a function of BT content at various frequencies. Tan d measured at 1 MHz decreased from 0.0076 for pure PEEK to 0.0059 for the 34 vol.% composite. This decrease is due to the lower dissipation factor of BaTiO3 compared with that of pure PEEK. However, it increased on further increasing the content of BT in the PEEK matrix, i.e., to 0.0069 and 0.0096 for the 47 vol.% and 67 vol.% composites, respectively. This is probably due to the voids and
aggregates present in these composites as confirmed by SEM analysis. In contrast, the tan d of polyimide/ BT measured at 1 MHz was 0.07,7 which is 10-fold higher than the PEEK/BT composites. At 15 MHz, a maximum dissipation factor of 0.043 was observed for pure PEEK. This may be attributed to the relaxation of PEEK molecules. However, it decreased to 0.014 for the 67 vol.% composite. Some authors have reported that high alternating electric fields probably ionize the air present in voids (e.g., in composites) and generate plasma which might give rise to such high dissipation factors.18–21 Several theoretical models such as series, parallel, Hashing–Shtrikman, Lichtenecker, modified Lichtenecker, and an effective-medium theory have been used to correlate the dielectric constant of ceramic-dispersed polymer composites. However, the Lichtenecker Eq. (1) and modified Lichtenecker Eq. (2) are the most commonly used. log e ¼ Vf log ef þ ð1 Vf Þ log em ;
(1)
log e ¼ log em þ ð1 kÞVf log ðef =em Þ;
(2)
60
0.05 R2 = 0.9035
50
Dissipation factor
2
Microhardness (kg/mm )
70
40 30 20
1 MHz 15 MHz
0.03 0.02 0.01
10 0 0
20
40
60
0
80
0
20
Vol % BaTiO 3 in PEEK matrix
40
60
80
Vol% BaTiO3 in PEEK matrix
Fig. 4. Vickers microhardness of PEEK/BT composites.
Fig. 6. Dissipation factor of PEEK/BaTiO3 composites.
(b)
(a) 60
50
67 vol%
40 30 47 vol% 20 34 vol%
Dielectric constant
10 kHz
50
Dielectric constant
100 KHz
0.04
40
100 KHz 1 MHz
30 15 MHz 20
10
25 vol%
10
0 vol%
0
0 10
100
1000
10000
Frequency (kHz), Log scale
100000
0
20
40
60
Vol% BaTiO 3 in PEEK matrix
Fig. 5. Dielectric constants for composites as a function of (a) frequency and (b) BT content in the PEEK matrix.
80
Fabrication and Properties of Novel Polyetheretherketone/Barium Titanate Composites with Low Dielectric Loss
2245
best fit to the Lichtenecker equation. More precisely, the predicted dielectric constant of BT powder with a mean particle size of 0.92 lm was found to be 560 and 275 based on the Jayasundere-Smith equation and Lichtenecker equation, respectively.5 Moreover, the dielectric constant of BT films with a grain size less than about 1.0 lm is approximately 1000.3,23 The lower dielectric constant of BT powder than that of BT bulk was attributed to the presence of voids, stresses, and defects.3,5 Hence, in the present study, dielectric constant values of 1000, 560, and 275 were selected for correlation purposes. It can be seen that the Lichtenecker equation agrees well with the experimental data when we substitute 275 for the dielectric constant of the BT powder. However, it overestimates the data when higher values (i.e., 560 and 1000) of dielectric constant are substituted. For the higher dielectric constant of BT powder, the modified Lichtenecker equation has been used by many authors. Figure 7b, c shows the correlation of the modified Lichtenecker equation with the experimental data. It is interesting that the modified Lichtenecker equation correlates well with the experimental data for k = 0.2 and a
where e, em, and ef are the dielectric constants of the composite, PEEK matrix, and BT filler, respectively. The k is an empirical fitting constant for the composites. Its value for most well-dispersed polymer/ ceramic composites is about 0.3.9 Sonoda et al.22 reported k values in the range of 0.22 to 0.30, and their investigation showed that it decreased as the carbon chain length of the surfactant increased. It is to be noted that the modified Lichtenecker equation fits the data better than the Lichtenecker equation because it uses an empirical fitting constant. Figure 7a shows a correlation between the experimental dielectric constant and values predicted from the Lichtenecker equation. Three predicted lines are drawn by considering three dielectric constants (1000, 560, and 275) of BT. The dielectric constant of bulk BT varies between 470 and 5000, depending on its grain size, purity, sintering temperature, crystallographic direction, measuring temperature, and method of preparation.1–4 However, the dielectric constant of BT powders has not been measured by any direct method so far. Recently, Cho et al. predicted a dielectric constant of BT powders in the range of 100 to 600 using the
Dielectric constant, Log scale
(a) 2.5
2.0
1.5
1.0
Exp. dielectric constant Lichtenecker Eqn ( =275) Lichtenecker Eqn ( =1000)
0.5
Lichtenecker Eqn ( =560) 0.0 0
0.2
0.4
0.6
0.8
Vol% BaTiO3 in PEEK Matrix
(b)
(c) 2.0 Dielectric constant, Log scale
Dielectric constant, Log scale
2.0
1.5
1.0 Exp. dielectric constant k=0.2, =560
0.5
k=0.3, =560 0.0 0
0.2
0.4
0.6
Vol% BaTiO 3 in PEEK composites
0.8
1.5
1.0 Exp. dielectric constant 0.5
k=0.2, =1000 k=0.3, =1000
0.0 0
0.2
0.4
0.6
0.8
Vol% BaTiO3 in PEEK composites
Fig. 7. Correlation between experimental dielectric constant with values predicted from (a) Lichtenecker equation with dielectric constant of BT of 1000, 560, and 275, (b) modified Lichtenecker equation with dielectric constant of BT of 560, and (c) modified Lichtenecker equation with dielectric constant of BT of 1000.
2246
Goyal, Madav, Pakankar, and Butee
dielectric constant of the BT powder of 560. However, if we use a dielectric constant of BT of 1000, the modified Lichtenecker equation with k = 0.3 correlates well with the experimental data. This shows that the fitting parameter k also depends upon the dielectric constant value of the ferroelectric ceramic. In brief, by selecting an appropriate value of the dielectric constant of the BT powder and the empirical fitting constant (k), one can easily predict the effective dielectric constant of polymer/ ferroelectric ceramic composites for the design of composite materials. Thus it can be concluded from Fig. 7a that the appropriate dielectric constant of micron-sized (0.9 lm) BT powder is about 275 (not 500 or 1000). Once an appropriate dielectric constant of BT has been selected, there is no need to fit the experimental data using the modified Lichtenecker equation, because the modified Lichtenecker equation uses a tradeoff between the values of dielectric constant and k. Coefficient of Thermal Expansion (CTE) Figure 8 shows the thermal strain (DL/L0) of pure PEEK and the 47 vol.% composite measured along the out-of-plane (thickness) direction. The meeting
2.5
(a) Pure PEEK (b) PEEK/ BaTiO3 (47 vol%)
Thermal strain (%)
2
(a) 1.5
1
(b) 0.5
0
30
60
90
120
150
180
210
240
Temperature (°C) Fig. 8. Thermal strain (DL/L) of composites as a function of temperature (30°C to 250°C).
point of the extrapolated tangents drawn over the linear curves (i.e., between 50°C and 100°C and between 170°C and 220°C) of thermal strain versus temperature was assumed to be the glass-transition temperature (Tg) of PEEK. The Tg value obtained by this method for pure PEEK was 143°C, which is close to the reported value. It can be seen from Fig. 8 that the thermal strain of PEEK increased significantly above Tg. The lower thermal strain of PEEK below Tg is due to the fact that, below Tg, the free volume is too small to allow movement of molecules or segments. However, it is sufficient to permit bond vibrations, which results in lower thermal strain as compared with above Tg. To enable movement of molecules or segments of molecules from place to place, there should be a free volume into which these molecules or segments may move. Above Tg, there is sufficient energy for molecular movement and the free volume increases sharply with temperature, thus resulting in higher thermal strain. Figure 8 also shows that the thermal strain of the 47 vol.% composite is significantly decreased due to the restriction of polymer chain movement by BT particles with increasing temperature. Table II shows the correlation between the experimental CTEs and values predicted from the rule of mixtures (ROM). The ROM can be expressed as ac = am.(1 Vf) + afVf, where ac, am, and af represent the linear CTEs of the composite, matrix, and BT particle, respectively. Below and above Tg, the CTE of pure PEEK is 59.9 9 106/°C and 98.2 9 106/°C, respectively. The CTEs of BT below and above the Curie temperature (120°C) were considered to be 6.2 9 106/°C and 11.3 9 106/°C, respectively.1 It can be clearly seen from Table II that both CTEs decreased by 56% for the 47 vol.% composite and showed a CTE much smaller than the value predicted using the ROM. This is due to the fact that the ROM does not take into account the interaction between the BT particles and the PEEK matrix. As discussed in the SEM section, there was better interfacial adhesion between the BT particles and the PEEK matrix. Moreover, CTE was decreased due to the decrease in volume fraction of the PEEK matrix and the lower intrinsic CTE of BT compared with pure PEEK.1 An appreciable decrease in CTE, i.e., an increase in dimensional
Table II. Dielectric, mechanical, and thermal properties of PEEK/BaTiO3 composites CTE (31026/°C) Dielectric Properties (1 MHz) Sample Code BT-0 BT-80
>Tg (170–220°C)
Exp. Density (g/cc)
Dielectric Constant
Tan d
Microhardness (kg/mm2)
ROM
Exp.
ROM
Exp.
1.30 3.05
3.3 26
0.0076 0.0069
24 60.7
59.9 34.45
59.9 26.0
98.2 57.25
98.2 43.3
Fabrication and Properties of Novel Polyetheretherketone/Barium Titanate Composites with Low Dielectric Loss
stability, of the composites is an important property for embedded capacitors to be used for high-temperature applications. CONCLUSIONS High-performance PEEK composites containing 0 vol.% to 67 vol.% BaTiO3 were successfully fabricated by hot pressing at 380°C and 45 MPa. The void content increased with increasing BaTiO3, particularly above 47 vol.%. SEM showed good dispersion and interaction of the BaTiO3 particles in the PEEK matrix. The microhardness increased by more than 2.5-fold for the 47 vol.% composite. However, it decreased on further increase of BaTiO3 loading in the matrix due to the presence of voids. The dielectric constant increased from 3.3 for pure PEEK to 46.5 for the 67 vol.% composite. The dissipation factor decreased from 0.0076 for pure PEEK to 0.0056 for the 34 vol.% composite, and thereafter it increased slightly with increasing BT. The high dielectric constant and low dissipation factor with negligible frequency dependence make these materials promising candidates for application in high-temperature embedded capacitors and organic substrates. The modified Lichtenecker equation with a suitable fitting parameter correlated well for the whole range of composites. It was also found that the value of k depends upon the dielectric constant of the ferroelectric ceramic. Both below and above Tg, the CTE of PEEK was significantly decreased. Moreover, the experimental CTE was lower than predicted using the rule of mixtures. ACKNOWLEDGEMENT We are grateful to Prof. M.J. Rathod, Head of Department of Metallurgy and Materials Science, College of Engineering, Pune for his help.
2247
REFERENCES 1. Y. He, Thermochim. Acta 419, 135 (2004). 2. W. Luan, L. Gao, and J. Guo, Ceram. Int. 25, 727 (1999). 3. T. Kanata, T. Yoshikawa, and K. Kubota, Solid State Commun. 62, 765 (1987). 4. S.H. Xie, B.K. Zhu, X.Z. Wei, Z.K. Xu, and Y.Y. Xu, Compos. A Appl. Sci. 36, 1152 (2005). 5. S.D. Cho, S.Y. Lee, J.G. Hyun, and K.W. Paik, J. Mater. Sci. Mater. Electron. 16, 77 (2005). 6. L. Ramajo, M. Reboredo, and M. Castro, Compos. A Appl. Sci. 36, 1267 (2005). 7. N.G. Devaraju, E.S. Kim, and B.I. Lee, Microelectron. Eng. 82, 71 (2005). 8. Z.M. Dang, Y. Zheng, and H.P. Xu, J. Appl. Polym. Sci. 110, 3473 (2008). 9. R. Kota, A.F. Ali, B.I. Lee, and M.M. Sychov, Microelectron. Eng. 84, 2853 (2007). 10. R. Patil, A. Ashwin, and S. Radhakrishnan, Sensors Actuat. A Phys. 138, 361 (2007). 11. K.Y. Lin, F.S. Yen, and C.Y. Hwang, J. Mater. Sci. 36, 3809 (2001). 12. V.S. Nisa, S. Rajesh, K.P. Murali, V. Priyadarsini, S.N. Potty, and R. Ratheesh, Compos. Sci. Technol. 68, 106 (2008). 13. P. Oikonomou, K. Manoli, D. Goustouridis, I. Raptis, and M. Sanopoulou, Microelectron. Eng. 86, 1286 (2009). 14. R. Popielarz, C.K. Chiang, R. Nozaki, and J. Obrzut, Macromolecules 34, 5910 (2001). 15. R.K. Goyal, A.N. Tiwari, and Y.S. Negi, Eur. Polym. J. 41, 2034 (2005). 16. R.K. Goyal, A.N. Tiwari, U.P. Mulik, and Y.S. Negi, Compos. Sci. Technol. 67, 1802 (2007). 17. S. Liang, S.R. Chong, and E.P. Giannelis, 48th Electronic Components and Technology Conference (1999), p. 171. 18. T. Hu, J. Juuti, and H. Jantumen, J. Eur. Cer. Soc. 27, 2923 (2007). 19. M. Valant and D. Suvorov, Mater. Chem. Phys. 79, 104 (2003). 20. D. Sen, T. Mahata, A.K. Patra, S. Mazumder, and B.P. Sharma, J. Phys.: Condens. Matter 16, 6229 (2004). 21. J.M. Herbert, Ceramic Dielectrics and Capacitors (New York: Gordon and Breach Science, 1985). 22. K. Sonoda, J. Juuti, Y. Moriya, and H. Jantunen, Compos. Struct. 92, 1052 (2010). 23. A.K. Tripati, V. Chariar, T.C. Goel, and P.K.C. Pillai, Mater. Sci. Eng. B Adv. 25, 34 (1994).