Dielectric properties of CaCu 3 Ti 4 O 12 ceramics with Zn substitution for Cu

Introduction

Dielectric materials have been widely used in various electronic applications. High dielectric constant oxide materials are of considerable interest due to their potential application in microelectronics. Recently, CaCu₃Ti₄O₁₂ (denoted as CCTO) of perovskite structure has attracted much attention because it possesses a very high dielectric constant with so called ‘giant dielectric effect’.^1–5 The CCTO ceramics were reported to exhibit an extraordinarily high dielectric constant of about 10⁴–10⁵ at room temperature over wide frequency (10²–10⁶ Hz). This dielectric behaviour has been explained by the intrinsic permittivity of CCTO ² and extrinsic effects of grin boundary.^3,6 However, the effect of internal barrier layer capacitor is commonly accepted to interpret the high permittivity of CCTO ceramics by now. ⁴ The existence of polarisation effects at insulating grain boundaries between semiconducting grains or other internal barriers is responsible for the colossal dielectric constant.

Preparation method and doping have great influences on the microstructure and dielectric properties of CCTO ceramics. Many efforts have been made to improve their dielectric properties by tailoring the compositions of CCTO. The addition of CaTiO₃ (Ref. 7 and 8) or ZrO₂ (Ref. 9) could reduce the dielectric loss at low frequencies (f<10⁴ Hz) by increasing the resistance of the barrier layers. Small amount of transition ions, such as Mn, Cr and V substitution for the octahedral Ti and La substitution for Ca site can yield changes of the magnetic and dielectric properties.^10,11 Recently, much work based on partial substitution of Cu with transition ions has been shown to increase the dielectric constant of CCTO dramatically.^12–18 Among these dopants, Zn has been reported to improve dielectric response and sinterability in a certain extent due to the modification of mixed valent structure.^15–18 This may be ascribed to the complex valence of copper ions changed with Zn doping in CCTO with Jahn–Teller effect as reported. ¹³ In these literatures, the Zn substitution for Cu was only up to x = 0·1 in CaCu_3–xZn_xTi₄O₁₂ due to the relative lower sintering temperature (1000°C). However, the solubility can be elevated by increasing the sintering temperature. This work studies the influences of Zn substitution for Cu in CCTO on the dielectric properties for the sintering temperature at1100°C.

Experimental

CaCu_3–xZn_xTi₄O₁₂ (x is from 0 to 1·0) ceramic samples were prepared by a conventional two-step solid state sintering process. The reagent grade of CaO, CuO, ZnO and TiO₂ powders were mixed thoroughly by ball milling with zirconia balls for 24 h using deionised water as liquid medium. The milled powders then were put in an alumina crucible and calcined at 900°C in air for 6 h. The calcined powders were reground with 1·5 wt-% of polyvinyl alcohol and were sieved through a 63 μm mesh after drying. Powders were then uniaxially pressed into discs of 19·5 mm in diameter and 1·5 mm in thickness under 100 MPa. The discs were sintered in air from room temperature to the intermediate temperature at 600°C for 2 h with heating rate of 2°C min⁻¹. The samples were kept heating to 1100°C for 8 h with heating rate of 5°C min⁻¹ and furnace cooled to room temperature.

The X-ray diffraction (XRD) analysis of the sintered samples was conducted using a diffractometer machine (MAC Science M18XHF) with Cu K_α irradiation and Ni filter. The density of the sintered samples was obtained by the Archimedes method. The microstructure of surface of the sample was examined using a scanning electron microscope (JEOL-5600 SEM). The top and bottom surfaces of the samples were pasted with Ag using a screen printing method and then fired at 600°C for 30 min for the measurements of the dielectric properties. The capacitance C_p and the dissipation factor tanδ of the samples were measured at room temperature with a parallel plate capacitor coupled to a precision LCR meter (HP 4284A). This measurement was conducted at an ac voltage of 1 V in the frequency range from 20 Hz to 1 MHz. The real dielectric constant ϵ′ was obtained from capacitance C_p and the dimension of the sample (the thickness and the area of the electrode). The complex impedance was obtained by the equation Z* = 1/(iωC*), where C* = C_p(1−itan δ) and ω = 2πf.

Results and discussion

Figure 1 illustrates the XRD patterns for the samples with different Zn content sintered at 1100°C for 8 h. All the samples showed the predominant phase of perovskite CCTO from the XRD analysis. However, the sample without Zn substitution (x = 0) shows a small amount of rutile TiO₂ second phase as in Fig. 1a. The traces of the TiO₂ second phase disappeared for the sample with Zn substitution as in Fig. 1b–d. Nevertheless, Zn₂TiO₄ second phase was observed for the sample with highly Zn substitution (x≥0·8) as in Fig. 1e and f.

OPEN IN VIEWER

1

X-ray diffraction patterns of samples sintered at 1100°C as function of x value for CaCu_3–xZn_xTi₄O₁₂ (▾: CCTO phase; ▽: TiO₂ rutile; ◊: Zn₂TiO₄)

The Zn substitution on Cu sites in CCTO resulted in a small shift of XRD peaks towards lower diffraction angle as in Fig. 2. The lattice parameter of the sample versus Zn content determined from XRD peaks is shown in Fig. 3. There is an increase in the lattice parameter from 7·387 to 7·397 Å as the Zn content increases. This is attributed to the ionic size effects by the fact that Zn²⁺ (r∼0·74 Å) has larger ionic radius than Cu²⁺ (r, ∼0·62 Å). The shift of the diffraction peak and dissolution of second phase (TiO₂) from XRD analysis would thus confirm that Zn can largely substitute for Cu into the lattice of CCTO crystal lattice up to x = 0·4.

OPEN IN VIEWER

2

Relative intensity of XRD peaks of (422) plane to (220) plane of CCTO phase as function of x value for CaCu_3–xZn_xTi₄O₁₂

OPEN IN VIEWER

3

Expanded XRD patterns of (422) reflection of CCTO phase for samples as function of x value indicating shift of peaks

The SEM images of the surface morphologies of the sample sintered at 1100°C for 8 h were shown in Fig. 4. Abnormal grains (>50 μm) with very fine grains (<5 μm) segregating at the grooved grain boundaries were observed for pure CCTO ceramics as in Fig. 4a. It was reported that the fine grains present at the grooved grain boundaries are of Cu rich phases.^19,20 These small Cu rich grains segregated at the grooved grain boundaries lead the exaggerated grains to be non-stoichiometry in composition of CCTO. The non-stoichiometry in CCTO would thus give the formation of TiO₂ secondary phase as in Fig. 4a. The overall grain size did not change significantly with small amount of Zn substitution for Cu for x<0·8 as shown in Fig. 4b and c. However, it was found that the growth of small grains at the grooved grain boundaries was lessened as the Zn content increased as in Fig. 4c–e. The Zn substitution for Cu would cause more homogeneous and stoichiometric in composition without the formation of TO₂ second phase. With further increases of Zn content, the average grain size was found to decrease largely to the lowest reading of ∼1·5 μm for the x = 1·0. This is attributed to retarding the grain growth by the formation of Zn₂TiO₄ second phase.

OPEN IN VIEWER

4

Photographs (SEM) of surface of samples with a x = 0, b x = 0·2, c x = 0·4, d x = 0·8 and e x = 1·0 for CaCu_3–xZn_xTi₄O₁₂

Figure 5 shows the results of broadband dielectric permittivity for the samples in the range of 100 Hz to 1 MHz at room temperature. The dielectric permittivity was very stable over the frequency range for the sample without Zn substitution. This is typical as the dielectric response of pure CCTO ceramics. Colossal value of ϵ′ of ∼1·5×10⁴ was obtained for the sample without Zn substitution (x = 0). The dielectric constant of the sample increased largely with Zn substitution. The highest dielectric constant was 6·2×10⁴ at f = 100 Hz with x = 0·8. Similar result of the enhanced giant dielectric response has been reported profoundly concerned with the modified mixed valence structure. ¹⁸ In addition, one plateau can be observed corresponding to a specific dielectric permittivity at the frequency between 10³ and 10⁵ Hz for the sample with Zn substitution. The dielectric constant of the sample increased from ∼10 000 for the pure CCTO (x = 0) to ∼45 000 for x = 0·8 at 10⁴ Hz. However, the sample with x = 1·0 showed a substantially reduced dielectric constant of∼30 000 at 10⁴ Hz. This may be ascribed to the formation of Zn₂TiO₄ second phase. These materials may present a so called giant dielectric permittivity associated with an internal barrier layer capacitor effect.^21–23 According to this model, CCTO ceramics are composed of semiconducting grains (pure CCTO phase) and insulating grain boundary layers. The increase in the dielectric constant with Zn substitution may be attributed the formation of the insulating grain boundary layers with Zn rich phase such as Zn₂TiO₄ second phase as observed from XRD spectra.

OPEN IN VIEWER

5

Frequency dependence of dielectric constant ϵ′ as function of x value for CaCu_3–xZn_xTi₄O₁₂

The variation tendencies of dielectric loss with frequency for the samples are presented in Fig. 6. No significant changes in the dielectric constant and relaxation were observed in the frequency range f<1 MHz for the pure CCTO. However, it was found that the sample with Zn substitution has the smallest tan δ value at high frequencies (f, ∼10⁴ Hz). In addition, the dielectric loss of the samples increased largely with the x value larger than 0·8. The existence of Zn₂TiO₄ with semiconducting behaviour could be considered as responsible for the high dielectric loss at low frequency.

OPEN IN VIEWER

6

Frequency dependence of dielectric loss (tanδ) as function of x value for CaCu_3–xZn_xTi₄O₁₂

The complex impedance spectra of the samples as a function of frequency range of 100 Hz to 1 MHz are shown in Fig. 7. The resistivity of the sample without Zn substitution (x = 0) is ∼9·5×10⁵ Ω cm at f = 100 Hz. The segregated small grains with Cu rich phases formed at the grain boundary do not give the half-circles of impedance in the range of frequency for x = 0 and x = 0·2. On the other hand, the impedance spectra yield apparent semicircles for the samples with Zn substitution (x≥0·6) as shown in Fig. 7. Only one semicircle is found in these complex plots. This is attributed to the large difference of the resistivity between grain and grain boundary. The higher resistivity of the grain boundary dominates the overall conductivity and thus leads to the appearance of only one semicircle in the complex impedance plots. The semicircle diameter gives the electrical resistivity of the sample, and the maximum value corresponds to the relaxation frequency. The semicircle diameter of the sample decreased as the Zn content increased for the x from 0·6 to 1·0. This indicates that the Zn substitution for Cu would decrease the overall resistivity of the sample.

OPEN IN VIEWER

7

Complex impedance plot Z″ versus Z′ at room temperature as function of x value

Conclusion

CaCu_3–xZn_xTi₄O₁₂ (x = 0∼1·0) ceramics over a broad range of Zn substitutions were prepared by two-step solid state reaction method. The results showed that CCTO ceramics exhibited single phase with 0·2≤x≤0·6 and the change in lattice constant confirms that Zn substitute into the lattice of CCTO from the XRD analysis. The dielectric constant of the sample largely increased with Zn substitution in the frequency range (f<10⁴ Hz). The highest dielectric constant can reach to 6·2×10⁴ at f = 100 Hz with Zn substitution of x = 0·8. The impedance spectrum of the sample with X≥0·6 yields an apparent semicircle due to the electrical response of grain boundary with the presence of insulating grain–boundary barrier layers.