Applied Physics A

, Volume 112, Issue 4, pp 975–984

Effect of processing on dielectric properties of (0.95)PbZr0.52Ti0.48O3–(0.05)BiFeO3

Authors

  • Subhash Sharma
    • Department of Physics and Materials Science and EngineeringJaypee Institute of Information Technology
  • Vikash Singh
    • Department of Physics and Materials Science and EngineeringJaypee Institute of Information Technology
  • Om Parkash
    • Department of Ceramic EngineeringInstitute of Technology Banaras Hindu University
    • Department of Physics and Materials Science and EngineeringJaypee Institute of Information Technology
Article

DOI: 10.1007/s00339-012-7458-5

Cite this article as:
Sharma, S., Singh, V., Parkash, O. et al. Appl. Phys. A (2013) 112: 975. doi:10.1007/s00339-012-7458-5

Abstract

An effort has been made to synthesize solid solution of a composition with x=0.05 in the system (1−x)PbZr0.52Ti0.48O3–(x)BiFeO3 by the sol–gel method. XRD patterns of the pure PZT and BFO modified PZT samples have shown single phase formation. The effects of substitution of BFO on dielectric properties of PZT have been studied in the frequency range 10 Hz to 100 kHz and temperature range from RT to 773 K. The density was optimized by sintering the BFO modified PZT samples at different temperatures in four batches, S1, S2, S3 and S4. PE hysteresis loop measurements for all the samples have shown almost saturated polarization. It has been observed that sample S2, sintered at 950 °C, exhibits superior dielectric properties of the four samples. The occurrence of weak ferromagnetism, observed in the MH hysteresis loop, indicates coupling between ferroelectricity and magnetism. Impedance analysis has revealed that all the samples, sintered at different temperatures, have a different grain resistance. A large change in ac conductivity around Tc has been observed in all the samples.

1 Introduction

Since the discovery of perovskite ferroelectrics, the solid solutions of the PbZrO3–PbTiO3 system have attracted worldwide attention due to its potential technological properties, such as high dielectric constant and piezoelectric coefficients. A complex phase diagram of this system PbZrO3–PbTiO3 shows that the former one is antiferroelectric (AFE) and the latter one is ferroelectric (FE) [1]. Depending on the Zr/Ti ratio, it can coexist in the form of orthorhombic phase with antiferroelectric polar order and rhombohedral phase with ferroelectric polar order (Ps〈111〉). For piezoelectric applications, the compositions near the morphotrophic phase boundary are chosen, due to their superior piezoelectric properties, while for the pyroelectric applications, a composition near the antiferroelectric and ferroelectric phase boundary is preferred.

BiFeO3 (BFO) with perovskite structure is one of the well known multiferroics in which ferroelectric and magnetic ordering (G-type antiferromagnetism) coexist up to quite high temperature. This material exhibits ferroelectricity up to 830 °C and antiferromagnetism up to 370 °C. The coexistence of two important physical phenomena concurrently makes this material useful for actuator, sensor and multistage data storage applications [2]. The occurrence of ferroelectricity and antiferromagnetism makes this material interesting from the view point of fascinating physics involved in coupling of these two phenomena. In the first process, the concept of d°-ness (dn, where n=0 and n is the number of electrons in the d sub-shell) is playing an important role and spontaneous polarization arises due to off center displacement of center cations [3]. On the other hand, the second process (where n≠0) requires partially filled d orbitals which is responsible for the magnetic ordering and hence spontaneous magnetization. The spontaneous polarization of bulk BFO was expected to be 90–110 μC/cm2 [4]. Even after the large spontaneous polarization, these materials do not exhibit a saturated hysteresis loop, which is due to large leakage current. This leakage current may be due to the fact that Bi evaporates during sintering, leading to non-stoichiometry and creating charge defects [5]. To overcome this leakage problem and reduce defects, ion substitution with other oxides has been found to be useful [6, 7].

Modifying the piezoelectric properties by suitable substitutions at different atomic sites in PZT is not a new idea. In recent years, there has been growing interest in multifunctional properties of the solid solutions in the ABO3–BiFeO3 systems which show the coexistence of two or more ordering parameters such as magnetic, ferroelectric, and/or ferroelastic [8, 9]. Coupling between any two of these in the same material not only make the system interesting but also play an important role in technological devices. It is, therefore, worthwhile to study their electrical and magnetic properties in detail. Until now, only a few reports of the solid solution compositions, formed between ABO3–BFO, are available in the literature, such as BiFeO3–PbTiO3 [10, 11], BiFeO3–PbFe0.5Nb0.5O3 and BiFeO3–BaTiO3 [9, 12] which has shown interesting properties. It has been reported that it is difficult to work with these systems due to the complexity of the process of solid solution formation. A strong tendency to form significant amounts of non-perovskite phases or impure phases creates complications in synthesizing these systems.

Solid solution formation of BFO–PZT (BiFeO3–PbZr0.60Ti0.40O3) system has been studied by Choudhary et al. [13] and has shown interesting electrical properties. A typical composition of lead zirconium titanate Pb(Zr0.52Ti0.48)O3, a composition close to the morphotropic phase boundary, has been studied widely [14] due to its high piezoelectric coefficients and dielectric constant. This material undergoes a phase transition from ferroelectric to paraelectric at a high temperature (Tc=390 °C). Recent studies in this field have expressed the need of the materials which can exhibit superior piezoelectric properties at temperatures for T≥400 °C for the use as high temperature sensor in automobile and aerospace engineering applications. It may be considered as one of the most suitable piezoelectric materials for making solid solutions with multiferroic BiFeO3 or BFO as it has been reported to have Tc≈830 °C. The addition of BFO in PZT is expected to enhance Tc in the solid solutions of PZT–BFO system. In view of the above, it is worthwhile to study the system (1−x)PZT–(x)BFO in detail. In this paper, the effect of 5 % BFO substitution on dielectric properties and detailed investigations on the electrical properties of a typical composition (0.95)PbZr0.52Ti0.48O3–(0.05)BiFeO3, abbreviated as PZT–5BFO, are reported. Here, we use the complex impedance spectroscopy to identify the various contributions to the overall dielectric behavior. In general, the overall dielectric properties arise due to inter grain, intra-grain and electrode processes. Impedance analysis has emerged as a powerful tool to separate out the above contributions [15, 16]. In view of this, impedance spectroscopy has been carried out to extract the information of grain and grain boundaries of the sample with the help of impedance plane and spectroscopic plots [17].

2 Experimental

Effort has been made to synthesize a solid solution of a typical composition with x=0.05 in the system (1−x)PZT–(x)BFO. For this, PZT(52/48) precursor solution (with 5 mole% excess Pb) was prepared by the sol-gel method using Pb acetate Tri hydrate [Pb(CH3CO2)2⋅3H2O], Zr n-propoxide [Zr(OCH2CH2CH3)4], and Ti iso-propoxide {Ti[OCH(CH3)2]4} as the starting materials and 2-methoxyethanol (CH3OCH2CH2OH) and acetic acid as the solvents. BiFeO3 precursor solution with 5 mol% excess of Bi prepared by sol-gel method using Bi nitrate penta-hydrate [Bi(NO3)3⋅5H2O] and Fe(III) nitrate nonahydrate [Fe(NO3)3⋅9H2O] as the starting materials, and 2-methoxyethanol and acetic acid [CH3COOH] as the solvent. Transparent solutions for both end members (PZT and BFO) are obtained separately. These solutions were mixed together in proper proportions, with constant stirring at room temperature for 5 h which transforms it to gel, a three dimensional network of polymeric precursor. Drying of the gel was carried out at 100 °C for 24 h. The dried gel of PZT pure and PZT–5BFO samples were ground and calcined at 800 °C for 4 h in an alumina crucible. The calcined powders were pelletized into the disc shape at optimized pressure using polyvinyl alcohol as a binder (2 wt% PVA in distilled water) with the help of a coaxial hydraulic press. The pellets of pure PZT were sintered at 1100 °C for 2 h and the pellets of PZT–5BFO were divided into four batches; S1, S2, S3 and S4 each containing three pellets. Samples in the batches S1, S2, S3 and S4 were sintered at 900 °C, 950 °C, 1000 °C, and 1050 °C, respectively. The phase formation and crystallinity of calcined powders were analyzed by X-ray diffraction using Cu–Kα (λ=1.5405 Å) radiations. Dielectric and ferroelectric measurements were carried out using a LCR meter (HIOKI 3522-50) and PE Loop Tracer (MARINE INDIA LTD.), respectively.

3 Results

X-ray diffraction (XRD) patterns of pure PbZr0.52Ti0.48O3 (PZT) and PZT–5BFO solid solutions are shown in Fig. 1(a, b). The diffraction pattern of pure PZT has been indexed on the basis of tetragonal structure with P4mm space group symmetry according to JCPDS file (No. 33-0784). No traces of other phases or impurity were formed (Fig. 1a). The diffraction peaks in XRD pattern of the calcined powder of PZT–5BFO are found to exactly match with the diffraction data of pure PZT sample (Fig. 1b), except a minor impurity peak (*), which is found to have disappeared in the XRD patterns of all the samples when sintered at higher temperatures. For example, the XRD pattern of S4 sample, sintered at the highest temperature say 1050 °C, is shown in Fig. 1c. The crystal structure of PZT–5BFO sample (only calcined sample) was analyzed by Rietveld analysis using powder diffraction data (Fig. 1d). This reveals that the composition possesses a tetragonal structure (P4mm). The pattern factor (Rp), the weighted pattern factor (Rwp) and the expected pattern factor (Rexp) were 16.8 %, 21.2 %, and 11.1 %, respectively. Unit cell constants and other parameters are given in Table I. The average crystallite size, determined using the Scherrer formula from X-ray line broadening of PZT–5BFO, was found to be ∼33 nm.
https://static-content.springer.com/image/art%3A10.1007%2Fs00339-012-7458-5/MediaObjects/339_2012_7458_Fig1_HTML.gif
Fig. 1

XRD patterns (a) pure PZT, (b) PZT–5BFO sample (calcined at 800 °C) , (c) S4 (PZT–5BFO sample (sintered at 1050 °C), (d) Rietveld plot for calcined sample (* impurity phase)

Figure 2 shows the energy dispersive X-ray spectrum (EDX) of the pure PZT and PZT–5BFO ceramics. Spectra are taken at a few selected positions of the sample to confirm the same constituents.
https://static-content.springer.com/image/art%3A10.1007%2Fs00339-012-7458-5/MediaObjects/339_2012_7458_Fig2_HTML.gif
Fig. 2

EDX spectra of (a) pure PZT, (b) PZT–5BFO (S4) sample

Figure 3 shows the variation of relative dielectric constant (εr) and tanδ (dielectric loss) of PZT–5BFO with logf at room temperature (RT) for all the samples S1, S2, S3 and S4. It has been observed from these plots that the dielectric behavior of these samples is almost similar. Dielectric constant has been found to decrease with increasing frequency in the frequency range of measurement. The dispersion in εr and tanδ has been found to be maximum in the low frequency region in all theses samples. These samples have shown a dielectric loss nearly 1 in the frequency range <1 kHz and a very small value (<0.15) in the frequency range 1 kHz to 100 kHz. From theses plots, it has been clearly observed that sample S2 exhibits the highest value of dielectric constant over the observed frequency range. When the data are compared with pure PZT, it is found that the dielectric constant of pure PZT sample is higher than the S2 sample (Inset of Fig. 3a).
https://static-content.springer.com/image/art%3A10.1007%2Fs00339-012-7458-5/MediaObjects/339_2012_7458_Fig3_HTML.gif
Fig. 3

εr and tanδ vs. logf plots for all the samples at room temperature. (Inset) Comparison between S2 and PZT

Figure 4 shows the variation of dielectric constant and tanδ with temperature at 50 kHz, 75 kHz, and 100 kHz for all samples S1, S2, S3 and S4, respectively. The dielectric behavior in all the samples is found to be similar. The dielectric plots for all the samples show a dielectric peak at around 713 K (440 °C) which is frequency independent. Figure 5 shows plots of dielectric constant and tanδ with temperature for pure PZT and the results are compared with S2 sample. For pure PZT, a dielectric peak is observed at 683 K, which is in agreement with the reported data [18]. Substitution of BFO in PZT for x=0.05 has lead to the increase in the transition temperature from 683 K to 713 K. From tangent loss vs. temperature (tanδ vs. T) plots, an anomaly at around 663 K (390 °C) is observed which is followed by a rapid rise in dielectric loss with temperature for T>713 K. The room-temperature values of dielectric constant and loss for pure PZT and PZT–5BFO at 100 kHz have been tabulated (see Table 1). The value of tangent loss at room temperature for pure PZT is found to be smaller (tanδ∼0.005) than PZT–5BFO (0.02).
https://static-content.springer.com/image/art%3A10.1007%2Fs00339-012-7458-5/MediaObjects/339_2012_7458_Fig4_HTML.gif
Fig. 4

Variation of εr with temperature for the samples (a) S1, (b) S2, (c) S3 and (d) S4

https://static-content.springer.com/image/art%3A10.1007%2Fs00339-012-7458-5/MediaObjects/339_2012_7458_Fig5_HTML.gif
Fig. 5

Comparison of εr with temperature for pure PZT and S2 sample

Table 1

Lattice parameters, values of dielectric constant and dielectric loss at 100 kHz, remanent and saturation polarization (Pr and Ps)

Materials

a (Å)

c (Å)

c/a

ϵr(RT)

tan(δ) (RT)

Pr (μC/cm2)

Ps (μC/cm2)

PZT

3.9995

4.1474

1.037

774

0.005

15.95

31.24

S2

3.9986

4.1398

1.035

371

0.019

10.139

16.498

S4

3.9975

4.1334

1.033

360

0.020

6.403

11.640

The plots of log(1/εr−1/εmax) vs. log(TTc) are shown in Fig. 6. According to Isupov, the diffuse phase transition (DPT) should be described by the relation [19]
$$ \frac{1}{\varepsilon _{r}} = \frac{1}{\varepsilon _{\mathrm{max}}} + \frac{(T - T_{c})^{\gamma }}{2\varepsilon _{\mathrm{max}}\delta ^{2}} $$
(1)
where the parameter ‘γ’ represents the diffuseness of the phase transition. Treatment of dielectric data above transition temperature in the form of log(1/εr−1/εmax) vs. log(TTc) has shown a linear behavior whose slope (γ) has been observed to be between 1 and 2. The value of γ for pure PZT is also determined and found to be 1.70 (see inset of Fig. 6 for sample S2).
https://static-content.springer.com/image/art%3A10.1007%2Fs00339-012-7458-5/MediaObjects/339_2012_7458_Fig6_HTML.gif
Fig. 6

log(1/ε–1/εmax) vs. log(TTc) for the samples S1, S2, S3, S4 and pure PZT (inset of S2)

Figure 7 shows a polarization–electric field (PE) hysteresis loop for all the samples S1, S2, S3 and S4. Polarization data have been generated by a Sawyer–Tower circuit. All the samples have shown nearly saturation polarization under an applied field of 60 kV/cm. The saturation polarization (Ps) and remanent polarization (Pr) of few of these samples are given in Table 1. The inset of Fig. 7 shows a magnetization vs. magnetic field (MH) curve for only a typical S2 sample. It has been observed that substitution of 5 % BFO in PZT (PZT–5BFO) has resulted in poor magnetization.
https://static-content.springer.com/image/art%3A10.1007%2Fs00339-012-7458-5/MediaObjects/339_2012_7458_Fig7_HTML.gif
Fig. 7

(a) PE hysteresis loops for samples S1, S2, S3 and S4 at room temperature (upper inset). Comparison of PE curve for pure PZT with S2 sample, and (lower inset) MH hysteresis loop for typical sample S2 at room temperature

Figure 8 shows the variation of ac conductivity with temperature for samples S1, S2, S3 and S4. It has been observed from lnσac vs. 1000/T plots for these samples that in the low temperature range (≤500 K), the conductivity is almost independent of temperature and above this temperature, it increases with temperature up to Tc, i.e. 713 K. The behavior of the conductivity of all these samples (S1, S2, S3 and S4) is almost similar in the different temperature regimes. However, when σac of S2 sample is compared with pure PZT (see inset of Fig. 8), it appears significantly different. It is observed that conductivity for PZT–5BFO is enhanced by the order of 5, i.e. at room temperature PZT shows conductivity ∼10−12 s/m and PZT–5BFO shows ∼10−7 s/m. This large variation in the conductivity may be due to BFO substitution. At transition temperature, the order of variation in the conductivity is reduced to ∼3 (see inset of Fig. 8).
https://static-content.springer.com/image/art%3A10.1007%2Fs00339-012-7458-5/MediaObjects/339_2012_7458_Fig8_HTML.gif
Fig. 8

Variation of ac conductivity with temperature for sample S1, S2, S3 and S4. (Inset) comparison between S2 and PZT

In order to investigate the effect of processing on the dielectric properties, observed in εr vs. T plots of these samples around Tc, and whether it is a pure bulk phenomenon [9] or space charge effect, impedance spectroscopy has been carried out. For this, Cole–Cole plots (Z″ vs. Z′) over the frequency range 10 Hz to 100 kHz and spectroscopic plots (Z″, Z′ with logf) plots at a few selected temperatures have been studied [15, 20]. The variations of Z″ with Z′ over the frequency range from 10 Hz to 100 kHz for all the samples at a few selected temperatures are shown in Fig. 9. A circular arc starts appearing from 350 °C onward for S2 and 375 °C onward for other samples (not shown). For temperatures T≥400 °C, a clear circular arc in each sample has been obtained (see Fig. 9a–c). These plots have been fitted with single depressed circular arcs which are passing through origin. The intercepts of these arcs on the Z′ axis give the resistive component ‘R’ of the parallel RC circuit elements. The value of capacitive components ‘C’ has been determined using ωRC=1 (where ω=2πf, f is frequency) at the frequency where Z″ shows maxima in Z″ vs. Z′ plots at all temperatures. These contributions may be considered as the contribution from the grain or bulk in terms of resistance and capacitance (Rg and Cg) [17]. The arc, corresponding to the grain boundary contribution, was not observed in the existing data. It has been observed from Fig. 9a that the value of intercept of an arc on Z′ axis (i.e. Rg) at 400 °C is maximum for S1 and, for S2 and S3 it is nearly same and minimum also. For S4, it is lower than S1 and higher than S2 and S3. At higher temperatures like 450 °C and 500 °C, the values of grain or bulk resistance ‘Rg’ are found to be different for each sample (see Fig. 9b and c).
https://static-content.springer.com/image/art%3A10.1007%2Fs00339-012-7458-5/MediaObjects/339_2012_7458_Fig9_HTML.gif
Fig. 9

Impedance plots for sample S1, S2, S3 and S4 at (a) 400 °C, (b) 450 °C, and (c) 500 °C

For a typical sample S2, the grain capacitance (Cg) and resistance (Rg), are determined at few temperatures from Z″ with Z′ plots (not shown separately) over the temperature range 350°C–500 °C. The electrical parameters (Cg and Rg) have been plotted with temperature shown in Fig. 10. It shows maximum capacitance at temperature ∼450 °C (Fig. 10a) which is close to the peak temperature of dielectric transition in εr vs. T plots (Fig. 4). Resistive contributions of grains (Rg) have also been plotted with temperature as lnRg vs. 1000/T (Fig. 10b). The activation energy (Eg) was determined from Arrhenius linear fitting of the plot which is 1.75 eV. For the same sample (S2), plots of Z″ and Z′ vs. logf at a few temperatures are shown in Fig. 11. The variation of Z″ with frequency shows broad maxima which shift to high frequency side with increasing temperature. Z′ has been found to decrease with increasing frequency and the position of Z″ maxima coincides with the inflection point of the Z′ variation. This frequency refers to characteristic relaxation frequency for the dipoles which are expected to be formed in the materials [21]. With increasing temperature, the relaxation frequency shifts to the high frequency side. This indicates that there exists a relaxation phenomenon [22]. Characteristic relaxation times ‘τ’ have been determined from the positions of the peak maxima in Z″ vs. logf at various temperatures, using the relation ωτ=1. The plot of log τ vs. 1000/T has been plotted (Fig. 12) which is found to be linear indicating that τ obeys the Arrhenius relationship:
$$ \tau = \tau_{0}e^{\frac{E_{\mathrm{rel}}}{kT}} $$
(2)
https://static-content.springer.com/image/art%3A10.1007%2Fs00339-012-7458-5/MediaObjects/339_2012_7458_Fig10_HTML.gif
Fig. 10

(a) Cg vs. temperature plot for typical sample S2 (b) lnRg vs. 1000/T plot for S2 sample

https://static-content.springer.com/image/art%3A10.1007%2Fs00339-012-7458-5/MediaObjects/339_2012_7458_Fig11_HTML.gif
Fig. 11

Z′ and Z″ vs. logf plots for S2 at 350 °C, 375 °C, 400 °C, 425 °C , 450 °C, and 500 °C

https://static-content.springer.com/image/art%3A10.1007%2Fs00339-012-7458-5/MediaObjects/339_2012_7458_Fig12_HTML.gif
Fig. 12

lnτ vs. 1000/T plot for a typical sample ‘S2’

Here τ0 is the characteristic time constant and Erel is the activation energy for the relaxation process. The value of Erel, obtained by least square fitting of the data, is found to be 1.32 eV.

4 Discussion

The perovskite structure can be described by the tolerance factor \(\mbox{`$t$'} = ( r_{A} + r_{0} )/\sqrt{2} ( r_{B} + r_{0} )\) where rA, rB and r0 are the ionic radii of A, B cations and oxygen anion. For perfect cubic structure, ‘t’ should be 1, if 0.89<t<1, the compound is inclined to form a cubic structure; for 0.8<t<0.89, it is more likely to form a distorted perovskite structure. In the present case for pure PZT, using the ionic radii given by Shannon and Prewitt [23], the value of ‘t’ is found to be 1.01, indicating the tetragonal structure. For BFO substituted PZT (PZT–5BFO), it is slightly reduced i.e. 1.004. This is supported by our results obtained from the XRD analysis. The lattice parameters, calculated from the diffraction data, indicate that pure PZT is found to have tetragonal structure with ratio c/a∼1.037 and for PZT–5BFO (S2) sample, this ratio (c/a) is ∼1.032. A slightly decreased c/a ratio for S2 sample seems to originate from reduced tetragonality due to BFO which has rhombohedral structure. The smaller ionic radii of Bi3+ (1.30 Å, extrapolated from the lower coordination) and Fe3+ (0.64 Å) than Pb2+ (1.49 Å) and Zr4+/Ti4+ (0.72 Å/ 0.68 Å), respectively, may be responsible for change in lattice constants. Using lattice parameter and atomic mass in appropriate proportion, the theoretical density (ρthe) for PZT–5BFO has been determined. The experimental density was calculated using geometrical dimensions and mass of the three pellets in each batch S1, S2, S3 and S4. The average % density for each of the sample was determined. It is found that sample S2 shows maximum experimental density, i.e. 94 % of the theoretical value. Nearly similar density (95 %) is obtained in the pure PZT sample. It is to be noted that the dielectric constant at room temperature of pure PZT and S2 sample is ∼774 and 371 respectively (Table 1). The dielectric constant of pure PZT is reduced on 5 % BFO substitution in PZT and tangent loss is increased. XRD patterns of PZT–5BFO sample, calcined at 800 °C and samples sintered at various temperatures are similar. A very small shift in the position of the diffracted peaks has been observed for the sample sintered at 1050 °C (see inset of Fig. 1b and c) which may be due to slight distortion in the structure caused by evaporation of Bi, however, the crystal structure remains the same and energy dispersive spectroscopy (Fig. 2b) has confirmed the presence of Bi, Fe along with Pb, Zr/Ti and O in S4 sample (sintered at 1050 °C).

The occurrence of the dielectric peak in εr vs. T plots at 713 K of these samples (S1, S2, S3 and S4) indicates the phase transition from ferroelectric to paraelectric (FE–PE) transition similar to the phase in PZT(52/48) [18], however, at different temperatures (Fig. 4a–d). The frequency independent nature of this transition indicates normal ferroelectric behavior. Further, on carefully observing the dielectric plots of S1 sample (see inset), another kink appeared at ∼648 K. This anomaly is observed in all samples, but at slightly different temperatures. Such types of anomaly in dielectric behavior have been reported in recent years in the ABO3–BFO system [24] which may be attributed to the magnetic transition in BFO.

The occurrence of a broad peak in all these samples at ∼713 K may be characterized by analyzing the power exponent ‘γ’ from the plots of log(1/ε–1/εmax) vs. log(TTc) above Tc, which measures the degree of diffuseness of the transition. In the present study, the value of γ has been determined for all samples and it is found to be between 1 and 2, which indicates diffused nature of FE–PE transition in these materials. The diffused type transition may be explained by the compositional micro heterogeneities present in the material. From the hysteresis loops for all these samples, it has been observed that all the samples show nearly saturated polarization. Sample S2 has shown maximum remanent polarization i.e. Pr=10.14 μC/cm2 among the four samples, however, less than pure PZT. Decrease in Pr and increasing coercive field ‘Ec’ on BFO substitution in PZT may be explained as follows. (i) Introduction of Bi3+ at Pb2+ sites may enhance oxygen vacancies due to Bi evaporation, since the bond energy of Pb–O bond (3.9 eV) is higher than Bi–O bond (1.7 eV). As a result, formation of defect complexes like \(( V_{\mathrm{Bi}}''' - V_{\mathrm{O}}^{ \bullet \bullet } )'\) may take place which could act as a pinning of domain wall motion and contribute in reducing polarization and enhancing coercive field [20]. (ii) Meanwhile, the insertion of Fe at Ti/Zr sites breaks Ti–O–Ti linkages and hence reduces spontaneous polarization. Further, the magnetization curve (MH loop) for a typical sample S2 shows a weak magnetization (see inset of Fig. 7). The induced weak magnetization in sample S2 seems to be due to BFO substitution in PZT. This is similar to the magnetization curve observed in the BFO–10 % BT ceramic [9], however, the value of Mr is much smaller in our sample. In fact, BFO is an antiferromagnetic system due to the symmetric spiral spin structure [25], which may be suppressed in PZT–5BFO sample due to disruption of the BFO lattice network by the presence of Pb and Ti ions.

It has been observed in the temperature dependent conductivity plots (Fig. 8) that in the low temperature range i.e. up to 500 K, ac conductivity is almost independent of temperature, which reveals that hopping of charge carriers occurs over a short range i.e. in the local region [26]. Ac conductivity increases with increasing temperature in the intermediate temperature range i.e. 500 K<T<700 K and becomes maximum at Tc. This increase in the conductivity of these samples may be attributed partly to the increase in polarizability around Tc and partly to the creation of charge carriers at 500 K. Above Tc, the conductivity was found to be decreased, which is a typical characteristic of the conductivity anomaly around Tc. It has been observed that conductivity is ∼10−8 S/m at around 500 K which increases up to 10−2 S/m at around 700 K. Such a large rise in conductivity is not only expected due to increase in the polarizability as mentioned above but may be attributed to the creation of charge carriers at ∼500 K. Because of the high negative temperature coefficient of resistance, this material may also be a potential candidate for NTCR devices. It has been well established in the literature that the activation energies for singly and doubly ionized oxygen vacancies lie in the range of 0.3 eV–0.4 eV and 0.6 eV–1.2 eV, respectively [27]. Usually in the perovskite ferroelectrics like titanate ceramics, oxygen vacancies are considered to be one of the mobile charge carriers. It is reasonable to believe that BFO substitution in PZT enhances oxygen vacancies as a result of Bi evaporation. According to the Kroger–Vink notation of defects, Bi evaporation and the formation of oxygen vacancies associated with pair of electrons are given by
https://static-content.springer.com/image/art%3A10.1007%2Fs00339-012-7458-5/MediaObjects/339_2012_7458_Equ3_HTML.gif
(3)
https://static-content.springer.com/image/art%3A10.1007%2Fs00339-012-7458-5/MediaObjects/339_2012_7458_Equ4_HTML.gif
(4)
With increasing temperature, the ionization of oxygen vacancies generates conduction electrons [28] as per the following reaction:
$$ V_{\mathrm{O}}^{ \bullet } \quad\Rightarrow\quad V_{\mathrm{O}}^{ \bullet \bullet } + e' $$
(5)

It appears that below 500 K, singly ionized oxygen vacancies (\(V_{0}^{\bullet}\)) exist in the material. With increasing temperature above 500 K, these singly ionized oxygen vacancies are excited to the doubly ionized oxygen vacancies (\(V_{0}^{ \bullet \bullet }\)) leading to the creation of free charge carriers and hence the conductivity increases. The activation energy in this region has been determined by linear fitting of the conductivity data to the Arrhenius equation, which is ∼1.0 eV. In order to analyze the proper conduction mechanism, a detailed analysis of activities energy has been carried out using other data for a typical composition S2. The activation energy, determined from the linear region of the lnτ vs. 1000/T plot (see Fig. 12), has been found to be ∼1.32 eV. This amount of energy may be required for the dipoles to relax within the grains of the sample, known as Erel. It is noticed that the activation energy, determined from log Rg vs. 1000/T plots, is 1.75 eV. This amount of energy is expected to be equivalent to the sum of the activation energy required for the orientation of the dipoles (Erel) within the grains and energy required for hopping of the charge carriers (i.e. 0.43 eV) via singly ionized oxygen vacancies which causes re-orientation of dipoles.

5 Conclusions

Polycrystalline samples with composition (0.95) PZT–(0.05)BFO were synthesized by the sol–gel method. XRD results have shown the formation of single phase compound having tetragonal structure. Density optimization has revealed that sample S2, sintered at 950 °C, has maximum density (94 % of theoretical value). Substitution of 5 % BFO in PZT has been found to enhance Tc and given rise to weak ferromagnetism. However, Pr and Ps are reduced as compared to pure PZT. Observation of an anomaly in dielectric constant around the magnetic transition temperature of BFO confirms multiferroic magnetoelectric coupling. Frequency independent dielectric behavior around Tc reveals diffused type ferroelectric transition. All these samples exhibited a saturated hysteresis loop. It seems that conduction process below Tc is due to oxygen vacancies. The large change in conductivity from 10−8 to 10−2 S/m around Tc suggests that ionization of singly ionized oxygen vacancies is responsible for the increase in the conductivity above 500 K.

Acknowledgements

Financial support from Department of Science and Technology, Government of India is gratefully acknowledged. One of the authors, Mr. Subhash is also thankful to Jaypee Institute of Information Technology, Noida 201307, India for Teaching Assistantship.

Copyright information

© Springer-Verlag Berlin Heidelberg 2012