Dielectric behaviour of CaCu3Ti4O12-epoxy composites

Dielectric behaviour of CaCu₃Ti₄O₁₂-epoxy composites

Leandro Alfredo RamajoI, ^*; Miguel Angel Ramírez^II; Paulo Roberto Bueno^II; María Marta Reboredo^I; Miriam Susana Castro^I

^IInstitute of Research in Material Science and Technology – INTEMA, National Research Council of Science and Technology of Argentina – CONICET, University of Mar del Plata, Juan B Justo 4302 (B7608FDQ)

^IIChemistry Institute of São Paulo State University – UNESP, R. Francisco Degni, s/n Bairro Quitandinha, 14800-900 Araraquara - SP, Brazil

ABSTRACT

The dielectric behavior of composite materials (epoxy resin – barium titanate and epoxy – CCTO) was analysed as a function of ceramic amount. Composites were prepared by mixing the components and pouring them into suitable moulds. In some compositions, the matrix was reduced by tetrahydrofuran (THF) incorporation. Samples containing various amounts of ceramic filler were examined by TG/DTA and scanning electron microscopy analysis. Dielectric measurements were performed from 20 Hz to 1 MHz and 30 to 120 °C. It was demonstrated that the epoxy – CCTO composites possessed higher permittivity than classic epoxy – BaTiO₃ composites. However, the low resin permittivity prevailed in the composite dielectric performance.

Keywords: polymer-matrix composites (PMCs), electrical properties, composites

1. Introduction

In recent years, there has been an increasing interest on high dielectric constant flexible particulate composites made up of a ferroelectric ceramic and a polymer for high density energy storage and capacitor applications¹. However, the dielectric constant of such polymer based composites is rather low (about 50) because of the lower dielectric constant of the matrix (usually below 10)^1-5. For instance, in BaTiO₃/epoxy composites, though BaTiO₃ has relatively high dielectric constant (>1000), the effective dielectric constant (e_eff) of the composite was as low as 50, even when the highest possible volume fraction of ceramics was incorporated². As the volume fraction of ceramics increased, the composite, unfortunately, lost its flexibility.

A new generation of ultrahigh dielectric materials such as CaCu₃Ti₄O₁₂ (CCTO) can be used in order to obtain composites with better performance⁶. A number of theoretical studies and experimental observations has attempted to elucidate the remarkable (ultra high) dielectric properties of CCTO perovskite-like material. These materials have demonstrated to have a dielectric constant as high as 50,000.

In this work, the dielectric performance of epoxy – CCTO composites was studied. Systems were prepared mixing components with a solvent (tetrahydrofuran, THF) in order to reduce the polymer viscosity and to pour them into suitable moulds. Relaxation phenomena as a function of frequency, temperature, and filler volume fraction were analysed.

2. Experimental

Epoxy DER 325 (Dow Chemical) was chosen because of its good dielectric properties ( e= 4.5, tand = 0.0082). DEH 324 (Dow Chemical) was the curing agent (12.5 phr) and THF (Dorwil Chemical) was used as solvent (9 phr).

CaCu₃Ti₄O₁₂ (CCTO) polycrystalline ceramics were prepared by solid-state reaction. All the starting materials used were of analytical grade: CaCO₃ (Aldrich, 99.99%), TiO₂ (Aldrich, 99.8%), and CuO (Riedel, 99%). These materials were ball milled in an alcohol medium for 24 hours in a polyethylene bottle, using zirconium balls. After this, the slurry was dried and thermally treated at 900 °C in air atmosphere for 12 hours. In order to compare the influence of CCTO addition on the epoxy resin, BaTiO₃ composites were prepared. In this way, commercial barium titanate, BaTiO₃ (TAM Ceramics Inc.) was used as filler. It was doped with 0.6 mol (%) of Nb₂O₅ to modify its dielectric properties. Powders were mixed in isopropilic alcohol by agitation at 6000 rpm during 5 minutes. Afterwards, alcohol was eliminated by heating at 65 °C until constant weight was achieved. The powder was thermally treated at 1350 °C for 180 minutes using a heating and a cooling rate of 3 °C/min. The powder was milled using a planetary mill with ZrO₂ balls (Fritsch, Pulverisette 7) for 90 minutes in isopropilic medium.

The ceramic powders were added to the epoxy resin at different volume fractions and then suitably blended using an ultrasonic mixer (Sonic vibra-cell 150 W) for 4 minutes. THF was introduced to reduce the viscosity of the mixture with filler fractions between 5 and 15 vol (%). Each mixture was poured into glass moulds and cured at 100 °C for 2 hours. Afterwards, samples were analysed by thermal gravimetric technique (TGA, Shimadzu TGA-50) in a nitrogen atmosphere and at a heating rate of 10 °C/min from room temperature to 800 °C. Density (r) was measured by Archimedes' method and theoretical density (r_T) was calculated using -Equation 1.

where r_p is the filler density, r_m is the matrix density, and V is the volume fraction of filler.

Finally, for dielectric measurements, samples were painted with a silver paste. Dielectric measurements were performed using a Hewlett Packard 4284A Impedance Analyser from 20 Hz to 1 MHz between 20 and 120 °C.

3. Results and Discussion

The experimental and theoretical densities of CCTO composites for different filler amounts are shown in Table 1. The difference between the experimental (r_E) and theoretical (r_T) density values is fundamentally due to the presence of pores into the composite, which are produced during the mixing process when air flow is restricted due to the high viscosity of the system.

Micrographs obtained by scanning electron microscopy of composites are shown in Figure 1. Regions without filler or trails of micro-porosity are observed. Moreover, particles with low size and relative fine distribution size can be seen. In both samples, particle distribution is not homogeneous and there is a lot of particle agglomeration due to a bad dispersion of particles during the mixing step.

Figures 2 and 3 show real and imaginary permittivity parts as a function of frequency and temperature, respectively. As it can be expected, real permittivity rose as the ceramic volume fraction increases. Values as high as 10e₀ were obtained with the addition of only 15 vol (%) of particles. Besides, the real permittivity increased as the temperature did (Figure 3) due to the higher mobility of polymer chains. Moreover, it was observed that imaginary permittivity decreased at low frequencies (Figure 3b), because of a relaxation process^7-8.

Changes in the permittivity values as a function of frequency are attributed to dielectric relaxations. These are more pronounced at low frequencies and high temperatures due to micro-Brownian motion of the whole chain (segmental movement). Nevertheless, these changes are also affected by the interfacial polarization process known as Maxwell-Wagner-Sillar, which exists in heterogeneous dielectric materials and is produced by the traveling of charge carriers⁹.

In order to study the frequency and temperature dependence of relaxation processes, electrical modulus was used. Figure 4 shows the real and imaginary parts of the electrical modulus obtained through Equation 2⁷ as a function of frequency and temperature, respectively.

In Figure 4 it can be seen that M´ values increased with frequency. Nevertheless, peaks in M´´ values were developed at this same frequency range, indicating the appearance of a relaxation process ( relaxation). The maximum of M´´ decreased when filler amount increased. Relaxations peaks were displaced to higher frequencies, since relaxation processes were influenced by the interfacial polarization effect which generated electric charge accumulation around the ceramic particles and the displacement of peak⁸ as the particle content increased.

Real permittivity values were fitted through the Lichtenecker model (Equation 3). In Figure 5 the experimental data of real permittivity values at 30 °C and 2500 Hz, as a function of the filler volume fraction, are plotted. From the fitting, it was obtained that real permittivity of BaTiO₃, CCTO, and the pure resin were of 2280, 9000, and 4.5, respectively^{7, 10}.

It can be seen that the real permittivity of both systems is appropriately fitted by the model. The real permittivity of CCTO systems was higher than BaTiO₃ composites, throughout the volume fraction range. Nevertheless, considering that dielectric permittivity of CCTO is 9000 e₀¹⁰ against 2500 e₀ of BaTiO₃⁷, composites permittivities did not show great differences at the same volume fraction due to the resin influence (4.5e₀) on the composite permittivity⁷. Moreover, heterogeneous particle distribution could have affected composite permittivity values.

4. Conclusions

On the one hand, real permittivity was influenced by filler volume fraction. Ceramic particles produced a rise in permittivity and had more influence on samples with high volume fraction. On the other hand, porosity was very low even though in the composites with higher amount of filler.

Resin had influence on the imaginary part of permittivity and generated relaxation processes near their T_g on all composites, while ceramic content only influenced the real part of permittivity. Interfacial polarization processes known as Maxwell-Wagner-Sillar were generated by particles. These processes produced an accumulation of charges on the interface which displaced peaks to higher frequencies.

Dielectric constants of CCTO composites were higher than the ones of BaTiO₃ / epoxy composites. However, the low resin permittivity prevailed in the composite dielectric performance.

The real permittivity variation with the filler content could be fitted by the Lichtenecker model.

Acknowledgments

This work was supported by the National Research Council of Science and Technology of Argentina (CONICET) and Brazilian research funding agencies FAPESP, CNPq and CAPES. It was also achieved thanks to material donation from Dow Chemical Argentina through the management of Ariadna Spinelli and Alfredo Fahnle.

Received: September 13, 2007; Revised: January 27, 2008

• 1. Bai Y, Cheng Z-Y, Bharti V, Xu HS, Zhang QM. High-dielectricconstant ceramic-powder polymer composites. Applied Physics Letter 2000; 25 (76): 3804-3806.
• 2. Kuo D-H, Chang C-C, Su T-Y, Wang W-K, Lin B-Y. Dielectric behaviour of multi-doped BaTiO₃ epoxy composites. Journal European of Ceramic Society 2001; 21(9): 1171-1177.
• 3. Gregorio R, Cestari JM, Bernardino FE. Dielectric behaviour of thin films of b-PVDF/PZT and b-PVDF/BaTiO₃ composites. Journal of Materials Science 1996; 11(31): 2925-2930.
• 4. Chan HLW, Chan WK, Zhang Y, Choy CL. Pyroelectric and piezoelectric properties of lead titanate/polyvinylidene fluoride-trifluoroethylene 0-3 composites. IEEE Trans. Dielectr. Electr. Insul 1998; 4(5): 505-12.
• 5. Dias CJ, Das-Gupta DK. Inorganic ceramic/polymer ferroelectric composite electrets, IEEE Trans. Dielectr. Electr. Insul 1996; 5(3): 706-734.
• 6. Bueno PR, Ramírez MA, Varela JA, Longo E. Dielectric spectroscopy analysis of CaCu₃Ti₄O₁₂ polycrystalline systems. Applied Physics Letter 2006; 89(19): 191117-3.
• 7. Ramajo L, Reboredo M, Castro M. Dielectric response and relaxation phenomena in composites of epoxy with BaTiO₃ particles. Composites Part A 2005; 36(9): 1267-1274.
• 8. Tsangaris G, Kouloumbi N, Kyvelidis S. Interfacial Relaxation Phenomena in Particles Composites of Epoxy Resin with Copper or Iron Particles. Materials Chemistry and Physics 1996; 44(3): 245-250.
• 9. Psarras G, Manolakaki E, Tsangaris GM. Electrical Relaxation in Polymeric Particulate Composite of Epoxy Resin and Metal Particles. Composites Part A. 2002; 33(3): 375-384.
• 10. Ramírez MA, Bueno PR, Longo E, Varela JA. Non-Ohmic and dielectric properties of a Ca₂Cu₂Ti₄O₁₂ polycrystalline system. Applied Physics Letter. 2006; 89(21): 212102-3.

Publication Dates

• Publication in this collection
  28 Apr 2008
• Date of issue
  Mar 2008

History

• Received
  13 Sept 2007
• Accepted
  17 Jan 2008