Modified giant dielectric properties of samarium doped CaCu3Ti4O12 ceramics

Materials Research Bulletin 47 (2012) 2257–2263

Contents lists available at SciVerse ScienceDirect

Materials Research Bulletin journal homepage: www.elsevier.com/locate/matresbu

Modified giant dielectric properties of samarium doped CaCu3Ti4O12 ceramics Prasit Thongbai a,*, Bundit Putasaeng b, Teerapon Yamwong b, Santi Maensiri c a

Integrated Nanotechnology Research Center (INRC), and Department of Physics, Faculty of Science, Khon Kaen University, Khon Kaen 40002, Thailand National Metal and Materials Technology Center (MTEC), Thailand Science Park, Pathumthani, 12120, Thailand c School of Physics, Institute of Science, Suranaree University, Nakhon Ratchasima 30000, Thailand b

A R T I C L E I N F O

A B S T R A C T

Article history: Received 14 March 2012 Received in revised form 8 May 2012 Accepted 30 May 2012 Available online 9 June 2012

Effects of Sm3+ substitution on the microstructure and dielectric properties of CaCu3Ti4O12 ceramics were investigated. The grain size of CaCu3Ti4O12 ceramics was greatly decreased by doping with Sm3+, resulting from the ability of Sm3+ to inhibit the grain growth rate. This result can cause a decrease in the dielectric constant (e0 ) and loss tangent (tan d) of CaCu3Ti4O12 ceramics. Interestingly, high dielectric permittivity (e0  10,863) and low loss tangent (tan d  0.043 at 20 8C and 1 kHz) were observed in the Ca0.925Sm0.05Cu3Ti4O12 ceramic. Nonlinear electrical properties of CaCu3Ti4O12 ceramics were modified by doping with Sm3+. The dielectric relaxation behavior of Sm-doped CaCu3Ti4O12 ceramics can be well ascribed based on the internal barrier layer capacitor model of Schottky barriers at the grain boundaries. ß 2012 Elsevier Ltd. All rights reserved.

Keywords: A. Ceramics C. X-ray diffraction D. Dielectric properties D. Electrical properties

1. Introduction Recently, CaCu3Ti4O12 (CCTO) and related materials (ACu3Ti4O12, where A = La2/3, Bi2/3, Na1/2Bi1/2, Na1/2La1/2, Na1/2Y1/2, etc.) gained considerable attention due to its abnormal high dielectric constant (e0  103–105) over the temperature range from 100 to 600 K and its apparent nonlinear current–voltage (current density–electric field, J–E) properties [1–17]. By considering the dielectric and nonlinear electrical properties of these perovskite-type materials, it is believed that these ceramics are promising materials for many applications. Normally, the degree of electronic device miniaturization utilizing capacitive components is decided by the e0 value of a dielectric material. The size of such devices can be reduced by replacing dielectric layers with a relative high-dielectric material. However, the loss tangent (tan d) of CCTO and related materials is still too large (tan d > 0.05 at 1 kHz), making them unsuitable for capacitor applications. Therefore, reduction of tan d value is an important and urgent issue requiring investigation. The dielectric properties of Ln-doped CCTO ceramics have been widely investigated, where Ln is lanthanide metal ions such as La3+ [18–23], Y3+ [22,23], Gd3+ [22–24], Eu3+ [23,25], and Nd3+ [26]. All of these doping ions have an influence on the microstructure and dielectric properties as well as the nonlinear current–voltage behavior of CCTO ceramics. Some doping ions can reduce

* Corresponding author. Tel.: +66 84 4190266; fax: +66 43 202374. E-mail addresses: [email protected], [email protected] (P. Thongbai). 0025-5408/$ – see front matter ß 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.materresbull.2012.05.058

significantly tan d of CCTO ceramics. La3+ substitution into CCTO can reduce tan d to be less than 0.03 [19]. However, the reduction of tan d is usually accompanied by a decrease in e0 value (2000– 3000). Generally, the perovskite structure can be considered highly flexible. For ACu3Ti4O12 compounds (e.g., CCTO) [27], the octahedra have tilted to produce square planar sites. As a result, the structure of CCTO becomes very rigid. The space for the Ca2+ cations in CCTO structure is therefore essentially fixed. As well known, the ionic radius of doping ions is one of the most important parameters to determine the incorporation site. It was found that lattice parameters of some closely related Ln2/3Cu3Ti4O12 (especially for Ln = Eu3+ (7.390 A˚), Gd3+ (7.388 A˚), and Sm3+ (7.394 A˚)) compounds and CCTO (7.391 A˚) are nearly the same in value [27]. This indicates that Eu3+, Gd3+, and Sm3+ ions can perfectly substitute into Ca2+ sites in CCTO structure. In spite of intensive research of Ln-doped CCTO ceramics, there is no report on the dielectric properties of Sm3+-doped CCTO ceramics. Substitution of Sm3+ into CCTO ceramics may have an impact on the microstructure of CCTO ceramics and improves the dielectric properties of CCTO ceramics. Thus, the aim of this work is to study the effect of Sm3+ doping ions on the microstructure and dielectric properties as well as the electrical response in CCTO ceramics. In this work, we found that substitution of Sm3+ into CCTO ceramics has remarkable effects on both the microstructure and dielectric response in CCTO ceramics. The grain growth of CCTO ceramics was inhibited by doping with Sm3+, resulting in changes of dielectric properties. The possible origin of the giant dielectric properties and the dielectric relaxation behavior were explained based on the electrically heterogeneous microstructure.

P. Thongbai et al. / Materials Research Bulletin 47 (2012) 2257–2263

2258

2. Experimental procedure In this work, Ca13x/2SmxCu3Ti4O12 (x = 0, 0.05, 0.1, and 0.2) was prepared by the conventional solid state reaction method. CaCO3 (99.9% purity), Sm2O3 (99.99% purity), CuO (99.9% purity) and TiO2 (99.9% purity) were employed as starting raw materials. The raw materials were mixed homogenously by balls milling in ethanol for 24 h and calcined at 800 8C for 6 h. The calcined powders were ground and pressed into pellets of 9.5 mm diameter and 1 mm thickness by a uniaxial compression at 200 MPa. Finally, these pellets were sintered in air at 1080 8C for 3 h. The sintered Ca13x/ 2SmxCu3Ti4O12 ceramics with x = 0, 0.05, 0.1, and 0.2 were abbreviated as CCTO, CCTO-Sm1, CCTO-Sm2, and CCTO-Sm3 samples, respectively. X-ray diffraction (XRD) (Philips PW3040, The Netherlands) and scanning electron microscopy (SEM) (Hitachi S-3400, Japan) coupled with energy-dispersive X-ray spectrometry (EDS) were used to characterize the phase formation and microstructure of the Ca13x/2SmxCu3Ti4O12 ceramics. The dielectric properties of the samples were measured using an Agilent E4980A Precision LCR Meter over the frequency ranging from 102 to 106 Hz and at the oscillation voltage of 0.5 V. The measurements were performed over the temperature ranging from 70 to 150 8C. Each measured temperature was kept constant with an accuracy of 1 8C. Current– voltage measurements were made at room temperature using a high voltage measurement unit (Keithley Model 247). The rise rate of the source voltage is 0.45 V s1. Prior to measurements, Au electrodes were sputtered on each pellet face at a current of 25 mA for 8 min using an Polaron SC500 sputter coating unit. The breakdown electric field (Eb) was achieved at J = 1 mA cm2. Numerical values for the non-linear coefficient (a) were obtained by a linear regression of a log(J) vs. log(E) plot within the range of validity of the I = V equation 2 in the range of 1–10 mA cm . The complex impedance (Z*) was calculated from the relation,

e ¼ e0  ie00 ¼

1 1 ; ¼ ivC 0 Z  ivC 0 ðZ 0  iZ 00 Þ

(1)

where e0 and e00 are, respectively, the real (dielectric constant) and imaginary parts (dielectric loss) of the complex permittivity (e*). Z0 and Z00 are the real part and imaginary part p offfiffiffiffiffiffiffi Z*, respectively. v is the angular frequency (v ¼ 2p f ) and i ¼ 1. C 0 ¼ e0 S=d is the empty cell capacitance, where S is the sample area, d is the sample thickness, and e0 is the permittivity of free space, e0 = 8.854  1012 F/m.

with grain sizes of about 40–50 mm [Fig. 2(a)]. As seen in Fig. 2(b), the CCTO-Sm1 sample exhibits an abnormal grain growth microstructure. Some grains grow to sizes of 10–40 mm, which is significantly greater than the mean grain size (1–2 mm). As seen in Fig. 2(c) and (d), the CCTO-Sm2 and CCTO-Sm3 samples exhibit finegrained microstructure (1–2 mm) with porous microstructure, as observed in the insets. The amount of pores in the microstructure of the CCTO-Sm3 sample is likely to be larger than that of the CCTOSm2 sample. These results mean that Sm3+ doping ions reduce the grain growth rate and solidification process of CCTO ceramics during the sintering process. Fig. 2(e) and (f) illustrates EDS spectra of CCTOSm1 and CCTO-Sm3 samples, confirming the presence of Sm in the microstructure of Ca13x/2SmxCu3Ti4O12 ceramics. The corresponding peak of Sm is clearly observed in the EDS spectrum of the CCTOSm3 sample. Generally, the grain growth in a ceramic is primary driven by grain boundary (GB) mobility, which is controlled by the diffusion of ions, atoms, and/or charge species of the grain across the GB. For CCTO ceramics, the sintering process is correlated with the liquid phase sintering mechanism [28]. The presence of liquid phase during the sintering enhances the diffusion rate of ions across the GB. The capillary force produced by the liquid phase has a major contribution on the solidification process. The presence of the liquid phase in CCTO during sintering process can cause an increase in the grain growth and densification rates. Thus, a great decrease in grain size of Ca13x/2SmxCu3Ti4O12 ceramics is caused by the ability of Sm3+ to inhibit the grain growth rate. The decrease in grain size of Ca13x/2SmxCu3Ti4O12 ceramics is therefore due to a reduced GB mobility (or GB diffusion coefficient). It was also found that substitutions of La3+, Gd3+, Nd3+, and Eu3+ for Ca2+ in CCTO ceramics can cause a large decrease in the mean grain size [18,20,22,23,26]. The large decrease in grain size of Ca13x/ 2SmxCu3Ti4O12 ceramics may be related to the solute drag mechanism, which is an important approach for reducing the GB mobility [29]. Segregation of some Sm3+ and other Ln3+ doping ions at the GBs can produce a dragging effect on the motion of GB [23]. Furthermore, the reduced GB mobility may be due to the fact that the segregation of the additive can change the structure and composition of surfaces and interfaces, resulting in changes in GB and surface diffusion coefficients. Fig. 3(a) shows the frequency dependence of e0 at 20 8C for the Ca13x/2SmxCu3Ti4O12 ceramics. The values of e0 at 1 kHz and 20 8C for the Ca13x/2SmxCu3Ti4O12 samples are summarized in Table 1. Obviously, e0 decreases with increasing concentration of Sm3+ doping ions. The decrease in e0 is associated with the decrease in the grain size. This result observed in Sm-doped CCTO ceramics is

(422) (440) (433)

(620)

(400) (013) (222) (321)

(211)

Intensity (a.u.)

Fig. 1 shows the XRD patterns of Ca13x/2SmxCu3Ti4O12 ceramics sintered at 1080 8C for 3 h, confirming the formation of the CaCu3Ti4O12 phase (JCPDS card no. 75-2188) in all ceramic samples. The main phase of CCTO is detected in all XRD patterns. The lattice parameters of the sintered ceramics were calculated by using Cohen’s least mean square method, and found to be 7.392, 7.392, 7.394, and 7.394 A˚ for the CCTO, CCTO-Sm1, CCTO-Sm2, and CCTO-Sm3 ceramic samples, respectively. These values are comparable to the values reported in literature [1] and JCPDS card no. 75-2188 for CCTO (7.391 A˚). These values are also comparable to the lattice parameter of Sm2/3Cu3Ti4O12 (7.394 A˚) [27]. Thus, a small change of the lattice parameters for CCTO and Ca13x/2SmxCu3Ti4O12 is due the fact that ionic radii of Ca2+ and Sm3+ are nearly the same in value. The surface morphologies of the Ca13x/2SmxCu3Ti4O12 ceramics are demonstrated in Fig. 2. Significant changes in the microstructure of the Ca13x/2SmxCu3Ti4O12 ceramics are clearly observed. Pure CCTO ceramic shows a large-grained microstructure

(220)

3. Results and discussion

(d) (c) (b) (a)

20

30

40

50

60

70

80

90

2θ (Degree) Fig. 1. XRD patterns of Ca13x/2SmxCu3Ti4O12 ceramics: (a) CCTO, (b) x = 0.05, (c) x = 0.1, and (d) x = 0.2.

P. Thongbai et al. / Materials Research Bulletin 47 (2012) 2257–2263

2259

Fig. 2. SEM images of surface morphologies for (a) undoped CCTO, (b) CCTO-Sm1, (c) CCTO-Sm2, and CCTO-Sm3 ceramic samples; insets of (c) and (d) reveal high magnification images of SEM for CCTO-Sm2 and CCTO-Sm3 samples, respectively. (e and f) EDS spectra of CCTO-Sm1 and CCTO-Sm3 samples, respectively.

similar to those observed in literature for other Ln-doped CCTO ceramics [18–26]. This observation can be well ascribed based on the internal barrier layer capacitor (IBLC) model [28],

e0eff ¼

egb A t gb

;

(2)

where e0eff , egb , A, and tgb are the effective dielectric constant, dielectric constant of the GB, an average grain size, and the thickness of GB, respectively. Fig. 3(b) demonstrates the frequency dependence of tan d at 20 8C for the Ca13x/2SmxCu3Ti4O12 samples. The value of tan d (at 1 kHz and 20 8C) of the CCTO sample was reduced by doping with Sm3+. As shown in Table 1, the dielectric properties of the CCTO-Sm1 sample with e0  10,863 and tan d  0.043 are superior among these four samples. The CCTO-Sm1 ceramic may have potential for capacitor application. The variation of e0 in the

temperature range from 70 to 150 8C (at 1 kHz) for all ceramic samples is demonstrated in the inset of Fig. 3(b). e0 is nearly temperature independent in the range of about 70 to 80 8C. As revealed in Fig. 4, e0 value at frequency lower than 10 kHz increases with increasing temperature. The strong increase in e0 observed at 0.1 kHz is similar to that observed in relaxorferroelectric oxides. Recently, it was clearly proved that this type of relaxor-like behavior is an artifact induced mainly by a non-Ohmic sample-electrode contact impedance [10]. At high frequency and low-temperature range, the decrease in e0 is attributed to the dielectric relaxation process [3,5,9]. Generally, the occurrence of these this dielectric relaxation is particularly valuable for investigation of the possible physical mechanisms related to dielectric response(s) in materials. Therefore, to study the low-temperature dielectric relaxation behavior, dielectric data as a function of frequency at a low temperature range

P. Thongbai et al. / Materials Research Bulletin 47 (2012) 2257–2263

2260

(a)

0.1 kHz 1 kHz 10 kHz 100 kHz 200 kHz

30000

4

10

25000

ε'

ε'

20000 3

15000

10

10000

CCTO CCTO-Sm1 CCTO-Sm2 CCTO-Sm3

2

10

2

3

10

5000

o

@ 20 C 4

10

10

10

5

10

@ CCTO-Sm1

0

6

-80

-40

0

Frequency (Hz) 10 0

120

160

Temperature ( C) Fig. 4. Temperature dependence of e0 at different frequencies for CCTO-Sm1 ceramic.

4

3 CCTO CCTO-Sm1 CCTO-Sm2 CCTO-Sm3

2

CCTO CCTO-Sm1 CCTO-Sm2@ 1 kHz CCTO-Sm3

be followed the Arrhenius law,   E t ¼ t 0 exp a ; kB T

-100 -50 0 50 100 150 o Temperature ( C) -1

10

o

@ 20 C -2

10

2

10

3

4

10

5

10

6

10

10

Frequency (Hz) 0

Fig. 3. Frequency dependence of (a) e and tan d at 20 8C for Ca13x/2SmxCu3Ti4O12 ceramics; inset of (b) shows the temperature dependence of e0 at 1 kHz for Ca13x/2SmxCu3Ti4O12 samples.

were fitted by a modified Debye relaxation model, Cole–Cole model [30],

e ¼ e0  ie00 ¼ e1 þ

es  e1 1 þ ðivt Þ1a

(3)

;

where es and e1 are the limiting values of the real part of the permittivity for frequencies below and above the relaxation frequency, t is the average relaxation time of electric dipole moment, and a is constant value representing the distribution of relaxation times. We found that the frequency dependence of e0 at a low-temperature range for the Ca13x/2SmxCu3Ti4O12 ceramics can be well fitted by the Cole–Cole model, as demonstrated in Fig. 5 for the CCTO-Sm1 sample. According to the fitted results, values of t at different temperatures for all samples were achieved and found to

(4)

where t0 is the pre-factor, Ea is the activation energy required for relaxation process, kB is Boltzmann constant, and T is absolute temperature. The data fitted using Eq. (4) are shown in the inset of Fig. 5. The Ea values for the CCTO, CCTTO-1, CCTO-2, and CCTTO-3 samples were calculated from the slope of the plots and summarized in Table 1. These results indicate that Sm3+ substitution has a slight effect on the thermally activated relaxation process in CCTO ceramics. The calculated Ea values are comparable to those reported in literature for CCTO and related ceramics [4,9,16,26]. Although replacing of Ca2+ with Sm3+ ions shows an insignificant effect on the value of Ea, this substitution has a remarkable influence on the average relaxation time of electric dipole moment in CCTO ceramics, as shown in the inset of Fig. 5. At a certain temperature, t decreases with increasing Sm3+ content. Nowadays, the interfacial polarization at insulating interfaces inside CCTO ceramics is likely accepted to be the origin of the giant dielectric response in CCTO and related ceramics. The possible insulating interfaces presented in CCTO polycrystalline ceramics concludes of GBs [2,4,5,10,12–14,16,28], domain boundaries [31],

CCTO-Sm1

10000 8000 -14

6000 Ln( τ)

10

ε'

(b)

10

tanδ

80 o

ε'

10

40

CCTO CCTO-Sm1 CCTO-Sm2 CCTO-Sm3

-15

o

-16

4000 Table 1 The values of e0 , tan d (at 1 kHz and 20 8C), relaxation activation energy (Ea), activation energy for the GB conduction, non-linear coefficient (a), and breakdown field (Eb) for the Ca13x/2SmxCu3Ti4O12 samples. Sample

e0

tan d

Ea (eV)

Egb (eV)

a

Eb (V cm1)

CCTO CCTO-Sm1 CCTO-Sm2 CCTO-Sm3

13,596 10,863 1,981 526

0.071 0.043 0.050 0.015

0.096 0.078 0.080 0.094

0.764 0.674 0.663 0.658

5.30 4.63 5.29 –

694.2 444.0 990.3 >2000

-17 3.8

2000 10

4.0

4.2

4.4

4.6

4.8

5.0

-1

1000/T (K ) 4

10

-70 C o -60 C o -50 C o -40 C o -30 C o -20 C

5

10

6

Frequency (Hz) 0

Fig. 5. Frequency dependence of e in temperature range from 70 to 20 8C for CCTO-Sm1 ceramic fitted by the Cole–Cole relaxation model (solid lines). Inset shows Arrhenius plots of relaxation time (t) for the Ca13x/2SmxCu3Ti4O12 samples.

P. Thongbai et al. / Materials Research Bulletin 47 (2012) 2257–2263

18000

o

1200

-Z'' (k Ω.cm)

-Z'' (k Ω.cm)

@ 20 C

12000

@ 20 C

900 600 CCTO CCTO-Sm1 CCTO-Sm2 CCTO-Sm3

300 0 0

50

100

6000

CCTO CCTO-Sm1 CCTO-Sm2 CCTO-Sm3

100 Hz 0 0

500

150

Z' (kΩ.cm)

1000

1500

2000

2500

Z' (kΩ.cm) Fig. 6. Impedance complex plane plot (Z*) measured at temperature of 20 8C for Ca13x/2SmxCu3Ti4O12 ceramics; inset shows an expanded view of high frequency data close to the origin.

and stacking fault defect [8,32,33]. For the IBLC electrically structure model [5,28], the dielectric response in CCTO ceramics is attributed to the interfacial polarization at GBs. All behaviors related to the electrical response at GBs should be ascribed based on this model. In this model, the average dielectric relaxation time can be expressed as [3,5],

t ¼ Rg C gb ;

(5)

where Rg and Cgb represent the resistance of grains and the capacitance of GBs, respectively. To clarify the effect of Sm3+ substitution on the average relaxation time, the values of Rg and Cgb were elucidated by using an impedance spectroscopy technique. The impedance complex plane plots (Z* plots) at a temperature of 20 8C and their expended view of high frequency data close to the origin for all samples are shown in Fig. 6 and the inset, respectively.

Z ¼

Rgb 1 þ ðivt gb Þ1a

5

1.5x10

4

2x10

4

1x10

(6)

;

where tgb = RgbCgb and the parameter a is constant (0 < a  1). As shown in Fig. 7, complex impedance plane plots in the temperature range of 120–150 8C for Ca13x/2SmxCu3Ti4O12 ceramics can be fitted by the Cole–Cole equation. By fitting the experimental data to Eq. (6), we obtained the values of Rgb and Cgb at different temperatures. It was found that the values of Cgb are nearly temperature independent for all samples. The Cgb values of the CCTO, CCTO-Sm1, CCTO-Sm2, and CCTO-Sm3 samples were found to be 2.65, 1.40, 0.26, and 0.01 nF, respectively. Cgb decreases with increasing Sm3+ doping concentration. According to Eq. (5), a remarkable decrease in t value for the Ca13x/2SmxCu3Ti4O12 ceramics is therefore attributed to the reduction of Cgb.

-Z'' (Ω.cm)

-Z'' (Ω.cm)

In the absence of sample-electrode effect, Rg and GB resistance (Rgb) at particular temperatures for a polycrystalline ceramic are generally determined from the diameter of two semicircular arcs observed in Z00 vs. Z0 plots at high and low frequency data, respectively [3,28]. In some cases, only the semicircular arc at a low-frequency range was observed. In these cases, Rg can be estimated from a nonzero intercept on the Z0 axis at high frequencies [28]. As seen in Fig. 6, it is likely that the Rgb values of the CCTO-Sm1, CCTO-Sm2, and CCTO-Sm3 samples tends to increase with increasing the Sm3+ doping concentration. As demonstrated in the inset, it is found that Rgb of the CCTO ceramic sample is larger than that of the CCTO-Sm1 sample; whereas, Rg values of all samples are changed slightly with Sm3+ content. According to Eq. (5), a decrease in t value may be related to the decrease in Cgb for Ca13x/2SmxCu3Ti4O12 ceramics. To investigate the electrical properties of GBs, the impedance data were fitted by the Cole–Cole equation for the complex impedance plane plot [3,34],

(a)

4

3x10

2261

(b)

5

1.2x10

4

9.0x10

4

6.0x10

4

3.0x10

0

0.0

0

4

1x10

4

4

2x10

4

3x10

0.0

4x10

4

5.0x10

5

1.0x10

5

1.5x10

5

2.0x10

Z' (Ω.cm)

Z' (Ω.cm) -8

(c)

ln(σgb) S/cm

-10 -12 -14 CCTO CCTO-Sm1 CCTO-Sm2 CCTO-Sm3

-16 -18 -20 2.32

2.36

2.40

2.44

2.48

2.52

2.56

2.60

-1

1000/T (1/K ) Fig. 7. Impedance complex plane plots in the temperature range of 120–150 8C for (a) CCTO-Sm1 and (b) CCTO-Sm2 samples; solid lines are the fitted curves using Eq. (6). (c) Arrhenius plot of GB conductivity (sgb) of Ca13x/2SmxCu3Ti4O12 ceramics.

P. Thongbai et al. / Materials Research Bulletin 47 (2012) 2257–2263

2262

30 CCTO CCTO-Sm1 CCTO-Sm2 CCTO-Sm3

20

-2

J (mA.cm )

25

15 10 5 0 0

300

600

900

1200

1500

1800

-1

E (V.cm ) Fig. 8. Non-Ohmic characteristics (J vs. E) of Ca13x/2SmxCu3Ti4O12 ceramics at room temperature.

As shown in the inset of Fig. 7(c), we also found that the grain boundary conductivity, sgb = 1/Rgb, follows the Arrhenius law,

s gb ¼ s 0 exp

  Egb ; kB T

(7)

where s0 is pre-exponential term and Egb is the activation energy for conduction at the GBs. The values of Egb for the CCTO, CCTOSm1, CCTO-Sm2, and CCTO-Sm3 samples were calculated from the slope of plots of ln sgb vs. 1000/T, and found to be 0.764, 0.674, 0.663, and 0.658, respectively. The Sm3+ doping ions have a slight influence on the electrical conduction at the GBs. These values are comparable to reported values of 0.600, 0.720, and 0.639–0.672 eV for the grain boundaries of the CCTO [28], Bi2/3Cu3Ti4O12 [3], and Na1/2La1/2Cu3Ti4O12 [15] ceramics, respectively. In addition to the dielectric and electrical properties, the effect of Sm3+ substitution on the nonlinear current–voltage behavior of CCTO ceramics was investigated. As shown in Fig. 8, the CCTO and Sm-doped CCTO ceramics exhibit non-Ohmic properties. a and Eb of the ceramic samples were calculated from these curves and summarized in Table 1. The a values of the CCTO, CCTO-Sm1, and CCTO-Sm2 samples were found to be 5.30, 4.63, and 5.29, respectively. The Eb values were 694.2, 444.0, and 990.3 V cm1, respectively. Note that the nonlinear J–E behavior of the CCTO-Sm3 sample cannot be observed in the range of 0–2000 V cm1. According to the dielectric and nonlinear current–voltage properties, we observed that the dielectric properties of CCTO ceramics

Cu2p3/2 Photoemission intensity (a.u.)

3+

Cu

Cu

CCTO-Sm3 CCTO-Sm2 CCTO-Sm1

CCTO 930

935

4. Conclusions In conclusion, the dielectric and electrical properties of Ca13x/2SmxCu3Ti4O12 ceramics were investigated. The microstructure analysis revealed that the mean grain size tended to decrease with increasing Sm3+ content, which was ascribed to the ability of Sm3+ doping ions to inhibit grain boundary mobility. This result can cause a decrease in the values of e0 and tan d. It was found that the Ca0.925Sm0.05Cu3Ti4O12 ceramic exhibited high e0 of 10,863 and low tan d of 0.043 at 20 8C and 1 kHz. The dielectric properties of Ca13x/2SmxCu3Ti4O12 ceramics were ascribed based on the IBLC model. The non-Ohmic properties of CaCu3Ti4O12 ceramics were modified by doping with Sm3+. Acknowledgments This work was financially supported by the Thailand Research Fund (TRF MRG5480045), the Commission on Higher Education (CHE), and Khon Kaen University, Thailand. References

2+

Cu +

can be improved by doping with a small amount of Sm3+; whereas, the nonlinear properties are improved by doping with high concentration of Sm3+. The oxidation states of polyvalent cations in Ca13x/2SmxCu3Ti4O12 ceramics were investigated by XPS technique. As demonstrated in Fig. 9, the XPS spectra of Cu2p regions for all ceramic samples were divided into three peaks by using Gaussian–Lorentzian profile fitting. Major peaks observed at the binding energies of about 933 eV correspond to Cu2+ [35– 37]. Additionally, the peaks’ relatively higher binding energies of about 935 eV indicate the presence of Cu3+ in these ceramic samples [35,36]. The presence of Cu3+ is similar to that observed in Mg-doped CCTO ceramics [36]. Another set of peaks was observed at relatively lower binding energies of about 931 eV indicate the presence of Cu+ [37]. The existence of Cu+ is electrically compensated by the occurrence of oxygen vacancies in ceramics during high-temperature sintering processes. For CCTO ceramics, at a high-sintering temperature, some Cu ions decomposed from the lattice and then segregated at the GBs. Cu deficiencies inside the grains lead to the presence of Cu vacancies. To keep the electrical charge balanced, some Cu2+ ions were oxidized to Cu3+ ions, as observed in XPS spectra in Fig. 9. The presence of Cu+ and Cu3+ may have an influence on the conductivity of grains. However, it is still difficult to obtain an accurate explanation for the origin of the semi conductivity in CCTO ceramics because the presence of these ions in very small amount would be sufficient to induce n-type semiconducting grains of CCTO ceramics.

940

945

950

Binding Energy (eV) Fig. 9. XPS spectra of Ca13x/2SmxCu3Ti4O12 ceramics samples.

955

[1] M.A. Subramanian, D. Li, N. Duan, B.A. Reisner, A.W. Sleight, J. Solid State Chem. 151 (2000) 323. [2] S.-Y. Chung, I.-D. Kim, S.-J.L. Kang, Nat. Mater. 3 (2004) 774. [3] J. Liu, C.-G. Duan, W.-G. Yin, W.N. Mei, R.W. Smith, J.R. Hardy, Phys. Rev. B 70 (2004) 144106. [4] L. Liu, H. Fan, P. Fang, X. Chen, Mater. Res. Bull. 43 (2008) 1800. [5] P. Lunkenheimer, S. Krohns, R. Fichtl, S.G. Ebbinghaus, A. Reller, A. Loidl, Eur. Phys. J. Spec. Top. 180 (2010) 61. [6] H. Ren, P. Liang, Z. Yang, Mater. Res. Bull. 45 (2010) 1608. [7] Z.Y. Liu, X.M. Li, J.Q. Wu, J. Am. Ceram. Soc. 95 (2012) 476. [8] W.C. Ribeiro, R.G.C. Arau´jo, P.R. Bueno, Appl. Phys. Lett. 98 (2011) 132906. [9] W. Somphan, N. Sangwong, T. Yamwong, P. Thongbai, J. Mater. Sci.: Mater. Electron. 23 (2012) 1229. [10] M. Li, D.C. Sinclair, A.R. West, J. Appl. Phys. 109 (2011) 084106. [11] A. Rajabtabar-Darvishi, W. .Li, O. Sheikhnejad-Bishe, L. Wang, X. Li, N. Li, W. Fei, Trans. Nonferrous Met. Soc. China 21 (2011) 400. [12] J. Yuan, Y.-H. Lin, H. Lu, B. Cheng, C.-W. Nan, J. Am. Ceram. Soc. 94 (2011) 1966. [13] T. Li, K. Fang, J. Hao, Y. Xue, Z. Chen, Mater. Sci. Eng. B 176 (2011) 171. [14] J.L. He, F.C. Luo, J. Hu, Y.-H. Lin, Sci. China Technol. Sci. 54 (2011) 2506.

P. Thongbai et al. / Materials Research Bulletin 47 (2012) 2257–2263 [15] P. Thongbai, T. Yamwong, S. Maensiri, Mater. Res. Bull. 47 (2012) 432. [16] P. Thongbai, B. Putasaeng, T. Yamwong, S. Maensiri, J. Mater. Sci.: Mater. Electron. 23 (2012) 795. [17] J. Zhao, J. Liu, G. Ma, Ceram. Int. 38 (2012) 1221. [18] B.S. Prakash, K.B.R. Varma, J. Mater. Sci.: Mater. Electron. 17 (2006) 899. [19] L. Feng, X. Tang, Y. Yan, X. Chen, Z. Jiao, G. Cao, Phys. Stat. Sol. (a) 203 (2006) 22. [20] B. Cheng, Y.-H. Lin, J. Yuan, J. Cai, C.-W. Nan, X. Xiao, J. He, J. Appl. Phys. 106 (2009) 034111. [21] S. Jin, H. Xia, Y. Zhang, Ceram. Int. 35 (2009) 309. [22] C. Mu, H. Zhang, Y. Liu, Y. Song, P. Liu, J. Rare Earths 28 (2010) 43. [23] L. Liu, L. Fang, Y. Huang, Y. Li, D. Shi, S. Zheng, S. Wu, C. Hu, J. Appl. Phys. 110 (2011) 094101. [24] R. Kashyap, O.P. Thakur, R.P. Tandon, Ceram. Int. 38 (2012) 3029. [25] T. Li, Y. Xue, Z. Chen, F. Chang, Mater. Sci. Eng. B 158 (2009) 58. [26] Y. Wang, L. Ni, X.M. Chen, J. Mater. Sci.: Mater. Electron. 22 (2011) 345.

2263

[27] M.A. Subramanian, A.W. Sleight, Solid State Sci. 4 (2002) 347. [28] T.B. Adams, D.C. Sinclair, A.R. West, J. Am. Ceram. Soc. 89 (2006) 3129. [29] M.N. Rahaman, Ceramic Processing and Sintering, second ed., Marcel Dekker, New York, 2003. [30] K.S. Cole, R.H. Cole, J. Chem. Phys. 9 (1941) 342. [31] T.-T. Fang, C.P. Liu, Chem. Mater. 17 (2005) 5167. [32] P.R. Bueno, R. Tararan, R. Parra, E. Joanni, M.A. Ramirez, W.C. Ribeiro, E. Longo, J.A. Varela, J. Phys. D: Appl. Phys. 42 (2009) 055404. [33] W.C. Ribeiro, E. Joanni, R. Savu, P.R. Bueno, Solid State Commun. 151 (2011) 173. [34] P. Thongbai, S. Tangwancharoen, T. Yamwong, S. Maensiri, J. Phys.: Condens. Matter 20 (2008) 395227. [35] W.Z. Yang, M.S. Fu, X.Q. Liu, H.Y. Zhu, X.M. Chen, Ceram. Int. 37 (2011) 2747. [36] M. Li, G. Cai, D.F. Zhang, W.Y. Wang, W.J. Wang, J. Appl. Phys. 104 (2010) 074107. [37] R. Yu, H. Xue, Z. Cao, L. Chen, Z. Xiong, J. Eur. Ceram. Soc. 32 (2012) 1245.