Dielectric behavior of CaCu3Ti4O12 electro-ceramic doped with La, Mn and Ni synthesized by modified citrate-gel route

The effect of La3+, Mn2+ and Ni2+ doped calcium copper titanate, CaCu3Ti4O12 (CCTO), at higher concentrations (CR1 and CR2 with 5 mol% and 10 mol%, respectively), has been examined by semi-wet route at relatively lower temperature. This semi-wet route employs citrate-nitrate gel chemical method using TiO2 solid powders. X-ray diffraction (XRD) analysis confirms the formation of single phase in the doped samples sintered at 900 °C for 8 h. Scanning electron micrographs (SEM) show that the average grain size for CR2 is larger than that of CR1 composition. The energy dispersive X-ray spectroscopy (EDX) is used to study the percentage compositions of different ions present in both ceramics. Dielectric constant (ɛr) and dielectric loss (tanδ) values of CR1 are comparatively higher than those of CR2 ceramic at all measured frequencies and temperatures. The nature of temperature-dependent relaxation behavior of the ceramics is also studied by impedance, modulus spectroscopic analysis and confirms Maxwell-Wagner relaxation.


Introduction
Recently, much attention has been paid to an unusual cubic perovskite oxide material CaCu 3 Ti 4 O 12 (CCTO), due to its high dielectric constant (10 4 -10 5 ) and good stability over a wide temperature range from 100 K to 600 K [1][2][3][4]. The large dielectric constant is very unusual because it does not show ferroelectric behavior. Giant dielectric constant of CCTO has been attributed to Maxwell-Wagner effect [5] at the interface of grains and grain boundaries in earlier studies, while Lunkenheimer et al. [6] believed that it is due to the electrode polarization effect. The materials with high ε r can store more charge and the smaller electronic circuits can be designed. It is considered as a very potential material for applications in microelectronics, especially in capacitive components and varistors. Partial substitution of appropriate cations in CCTO is considered as an effective method to improve the physical properties of the materials. The radius, valancy and co-ordination number of an element are important parameters to determine the site it occupies  in the parent compound and the composition range of solid solution formation [7]. It is well known that the dielectric properties of CCTO are strongly dependent upon the processing condition as well as doping effects [8,9]. A lot of research work has been done on partial substitution of Cu and Ti ions in CCTO ceramic in order to improve the dielectric properties for understanding the origin of dielectric response in these materials [10,11]. In the previous work, various substitutions were made on different sites of CCTO to investigate the effect of substitution on dielectric properties, e.g., Ca by La [12], Cu by Mn [13] and Ti by Ni [14]. In the present communication, La 3+ , Mn 2+ and Ni 2+ cations doped calcium copper titanate, CCTO, at higher concentration level (5 mol% and 10 mol%), were synthesized by semi-wet route at relatively lower temperature and abbreviated as CR1 and CR2 for 5 mol% and 10 mol%, respectively. The surface morphology, dielectric properties, relaxation behavior of the doped ceramics have been examined.

Experimental procedures
Analytical reagent grade chemicals Ca(NO 3 (TiO 2 ) and citric acid, all having purity of 99.95% obtained from Merck, India, were used as starting materials. Standard solutions of metal nitrates were prepared using distilled water. Solutions of metal nitrates in stoichiometric amount of these metal ions were mixed in a beaker. Calculated amount of TiO 2 and citric acid equivalent to metal ions were added to the solution. The solution was heated on a hot plate using a magnetic stirrer at 70-90 ℃ to evaporate water and dried at 100-120 ℃ in hot air oven for 12 h to yield a blue gel. The gel was calcined in air at 800 ℃ for 6 h in a muffle furnace. The resultant mixtures were ground into fine powders using a pestle and a mortar. Cylindrical pellets were made using a hydraulic press. The pellets were sintered at 900 ℃ for 8 h in air. In order to perform the electrical measurements, silver paste was applied to both sides of the circular face of the ceramic pellets, then dried at 600 ℃ for 20 min, and cooled naturally to room temperature. The crystalline phases of the sintered samples were identified using an X-ray diffractometer (Rich-Siefert, ID-3000) employing Cu Kα radiation.
The microstructures of the fractured surfaces were examined using a scanning electron microscope (SEM, JEOL JSM5410). The energy dispersive X-ray analyzer (EDX, Kevex, Sigma KS3) was used for elemental analysis of the sintered samples. The dielectric and impedance data of the doped CCTO ceramics were collected using the LCR (inductance, capacitance and resistance) meter (PSM 1735, Newton 4th Ltd, U.K.) with variation in frequency (100 Hz-5 MHz) and temperature (300-500 K).

Results and discussion
The X-ray diffraction (XRD) patterns of CR1 and CR2 ceramics sintered at 900 ℃ for 8 h are shown in Fig. 1. The doped CCTO precursor dry powders were calcined at 800 ℃ for 6 h and then sintered at 900 ℃ for 8 h. XRD data were indexed on the basis of cubic unit cell similar to the undoped CCTO. The similarity of the doped X-ray data to the undoped CCTO (JCPDS 75-2188), confirms the formation of a single-phase material. There is no evidence of the presence of a secondary phase in the ceramic. The lattice parameter and unit cell volume were determined using the least-square refinement method. The value of lattice parameter and its unit cell volume for both compositions are given in Table 1. The lattice parameter increases with the dopant concentration  increasing. The slight increase in lattice parameter and unit cell volume is due to the small difference in ionic radii of the substituent ions and the host ions [15].
To investigate the role of microstructure and composition on the dielectric properties of these ceramics, SEM and EDX studies were performed for both materials. Figure 2 shows the SEM images of the fractured surface of both compositions sintered at 900 ℃ for 8 h. The SEM image for CR1 shows faceted-packed morphology having grain size in the range of 1-2 µm. The SEM image for CR2 shows faceted grains with cubical appearance morphology. It is clearly seen that the microstructure of the doped CCTO ceramic changes significantly with the value of x increasing. The grain size for CR2 composition is found to be in the range of 1-3 µm and there is an indication of agglomeration of a secondary phase along the grain boundaries. This is due to the transition of copper oxide into the liquid phase during sintering treatment [16,17]. This suggests that the specimens prepared in the present study have a polycrystalline structure. From Figs. 2(a) and 2(b), it seems that the grain size increases remarkably on the increasing of the dopant concentration x. The modification of grain size affects the dielectric response in CCTO as reported earlier, but it should not be greater than 20 µm [18]. The grain size of these ceramics synthesized by this method is relatively smaller than those synthesized by the conventional ceramic methods [19,20].
The EDX spectra obtained for CR1 and CR2 (Fig. 3 In both compositions, the major ions are present as per stoichiometric ratio and confirm the purity of CR1 and CR2 ceramics prepared by the semi-wet route.   The temperature dependence of dielectric constant (ε r ) and dielectric loss (tanδ) of the doped CCTO samples under different frequencies between 100 Hz and 100 kHz are shown in Figs. 4 and 5. It is noted from Fig. 4 that the dielectric constant of CR1 is ~ 1164 at room temperature (310 K) and 100 Hz, which increases continuously with temperature increasing and becomes more rapid beyond 425 K. Dielectric constant of CR2 is ~ 510 at room temperature and 100 Hz, and also increases with temperature increasing as shown in Fig. 5. Fig. 4 The variation of dielectric constant (ε r ) and dielectric loss (tanδ) with temperature at a few selected frequencies for CR1 ceramic. It is found that dielectric constant (ε r ) at low frequency is quite higher than that at high frequency. The result is similar to that of dielectric properties of CCTO as a function of temperature in the lower-frequency region [21]. The dielectric constants of CR1 and CR2 ceramics at 10 kHz and 100 kHz are almost independent in the temperature range of 300-425 K, and then increase steadily. It is also observed that the dielectric constants of CR1 are always higher than those of CR2 ceramic for all measured frequencies. This is due to the larger grain size of CR2 than CR1. It is noted that the tanδ values of CR1 and CR2 ceramics at room temperature and 100 kHz are 0.46 and 0.02, respectively. The tanδ values for CR2 remain constant in the temperature range of 310-425 K, and then increase steadily with the rise in temperature at high frequency (100 kHz). The plots of tanδ at low frequency show anomaly at 480 K.
The plots of variation of dielectric constant (ε r ) and dielectric loss (tanδ) of these samples with frequency at a few selected temperatures are shown in Figs. 6 and 7. As seen from the figures, dielectric constants of both ceramics decrease with the increase in frequency.
The dielectric constant of CR1 is always higher than that of CR2 ceramic in the measured frequency range of 100 Hz-5 MHz. The rate of decrease of ε r with frequency increases with temperature increasing. It is observed from Fig. 6 that the dielectric loss also decreases with frequency increasing. As we increase the temperature, the dielectric loss also increases in low-frequency regions. It is observed from Fig. 7 that the broadening of the peaks shows typical relaxation behavior found in CR2 ceramic at low frequency with temperature increasing. These peaks shift to lower-frequency region with temperature decreasing [22]. This relaxation behavior arises due to space charge polarization present in these ceramics. The grain boundary barrier layer capacitor consists of conducting or semi-conducting grains and insulating grain boundary layers. Therefore, relaxation behavior arises whenever two phases with different electrical conductivities are in contact. Fig. 6 The variation of dielectric constant (ε r ) and dielectric loss (tanδ) vs. log f for CR1 ceramic at a few selected temperatures. Fig. 7 The variation of dielectric constant (ε r ) and dielectric loss (tanδ) vs. logf for CR2 ceramic at a few selected temperatures.
The investigation of dielectric relaxation and formation of barrier layer in this ceramic is confirmed by complex plane impedance (Z * ) and complex plane modulus (M * ) analysis, which separates the contributions of grains, grain boundaries and electrode specimen interface to the total resistance and capacitance, respectively, and correlates the electrical behavior of the ceramic to its microstructure [23]. This structure can be modeled using an equivalent impedance circuit that consists of two parallel resistancecapacitance (RC) elements in series. One pair of the parallel RC elements represents the grain boundary layers, while the other pair of parallel RC elements represents the grains. The capacitance C gb and resistance R gb of the grain boundary layers are usually much larger than the capacitance C g and resistance R g of the grains. Therefore, the relaxation time τ gb = R gb C gb of the grain boundary layers is much larger than that of the grains τ g = R g C g . In this case, two parallel RC elements in series produce two semicircular arcs in a complex impedance plane Z * plot [24] and the impedance can be calculated from the following equations: and g g gb gb g g b 2 2 g g gb gb Z' is the real part of the complex impedance and Z" is the imaginary part. The large arc is due to the grain boundary responses at the low frequencies, and the small arc is due to the grain responses at the high frequencies. At the maximum of an arc ωτ = 2πƒ max τ = 1, where ω = 2πƒ, ƒ is the frequency of the applied field. In actual measurements, it is difficult to obtain both semicircular arcs in one Z * plot, especially for the small arc associated with the grains. Its ƒ max at room temperature is well beyond the highest available frequency of impedance analyzers used. Figure 8 shows the complex impedance plane Z * plots of CR1 and CR2 ceramics obtained at temperatures 313 K and 413 K, respectively. The grain boundary response of CR1 dominates at 313 K as shown in the inset of Fig. 8(a) respectively, at room temperature (313 K). The resistivity of grain boundary (R gb ) for CR1 and CR2 are 3.25×10 5 Ω·cm and 12.36×10 5 Ω·cm at 413 K, respectively. It is observed from Fig. 8(b) and its inset that the grain resistivity of CR2 are 70.64×10 5 Ω·cm and 10.68×10 5 Ω·cm at 313 K and 413 K, respectively. It is observed that on increasing the dopants concentration, the resistivity of CR2 is higher than that of CR1 at both temperatures. In the case of CR1, the existence of single arc can be taken to interpret both the grain and grain boundary conduction processes have identical time constant at both temperatures, while CR2 shows the semi-circle arcs at both temperatures but overlapped. The overlapped semicircle arcs between grain and grain boundary may be due to the molten grain and the formation of undefined grain boundaries [25]. The variation of the imaginary part of the impedance Z″ with frequency at a few selected temperatures is shown in Figs. 9 and 10 for CR1 and CR2, respectively. The inset figures are the expanded view of Figs. 9 and 10 at high temperature. It clearly shows that the relaxation peaks are observed at all measured temperatures for both ceramics. In the case of CR2 ceramic, they show relaxation peaks at low frequency and high frequency. The peaks are suppressed and slightly shifted to high-frequency region on increasing temperature, which confirms the existence of a temperature-dependent dielectric relaxation. The dispersion of curves in the low-frequency region at different temperatures is very clear and appears to be  merging at higher frequency with temperature. This behavior is apparently due to the presence of space charge polarization effect at lower frequency which is definitely eliminated at higher frequency. In order to investigate the temperature dependence of relaxation processes in these ceramics, the electric modulus studies were also carried out. The complex electric modulus (M * ) is defined in terms of the complex dielectric constant (ε * ) and is represented as where M', M" and ε', ε" are the real and imaginary parts of the electric modulus and dielectric constants, respectively. The imaginary parts of the modulus at various temperatures are calculated using Eq. (5) for these ceramics. Figs. 11 and 12 show the plots of imaginary part of electric modulus (M") with variation of frequency at few selected temperatures for CR1 and CR2. The relaxation M″ peaks are observed which shift to higher-frequency region with temperature increasing and give a direct evidence of temperature-dependent relaxation.
The impedance data were also used to evaluate the relaxation time of the electrical phenomena in CR1 and CR2 ceramics using the relation τ =1/ω=1/2πf, where f is the relaxation frequency. The nature of variation of the relaxation time (τ) with reciprocal of temperature (1/T) is shown in Fig. 13. The activation energy in the relaxation process is determined by the temperaturedependent relaxation time constant τ gb which obey Arrhenius law given by the following equation: where E a is the activation energy involved in the dielectric relaxation process; τ 0 is the pre-exponential factor; k B is Boltzmann constant; and T is the absolute temperature. The grain boundary activation energy evaluated from the slope of lnτ gb against 1/T curve is found to be 0.42 eV and 0.38 eV for CR1 and CR2, respectively. This value agrees with the value reported earlier for Maxwell-Wagner relaxation arising from the interfacial polarization for CCTO ceramic which confirms the temperature dependence of dielectric properties [26,27].

Conclusions
Calcium copper titanate ceramics, CaCu 3 Ti 4 O 12 (CCTO), doped with La 3+ , Mn 2+ and Ni 2+ at higher Fig. 11 Electric modulus plots of imaginary M" vs. frequency at a few selected temperatures for CR1 sintered at 900 ℃ for 8 h. Linear fit R concentration (5 mol% and 10 mol%) were synthesized by semi-wet route at relatively lower temperature. XRD confirms the single-phase formation of the ceramics and the scanning electron micrographs show their surface morphology with grain size in the range of 1-3 µm. The stoichiometry of the ceramics is confirmed by EDX studies. With the increase in concentration of the dopant ions in CCTO ceramics, the values of dielectric constant as well as the dielectric loss decrease. Impedance analyses of both the ceramics show that grain boundary resistance dominates the total resistance of the ceramics. The dielectric constant responses of both the ceramics arise due to the space charge polarisation. The value of resistance of the grain boundary is found to be several orders of magnitude larger than that of the grain. The nature of temperature-dependent relaxation behavior of the ceramics studied by impedance and modulus spectroscopic analyses confirms Maxwell-Wagner relaxation.