Is

Sa

b

c

d

a

ARRAA

KCCSMI

1

tagtpeoop

I

i

s

0d

Journal of Alloys and Compounds 477 (2009) 706–711

Contents lists available at ScienceDirect

Journal of Alloys and Compounds

journa l homepage: www.e lsev ier .com/ locate / ja l l com

mpedance spectroscopy and morphology of SrBi4Ti4O15 ceramics prepared byoft chemical method

.K. Routa,d,∗, Ali Hussianb, J.S. Leeb, I.W. Kimc, S.I. Wooa,∗∗

Department of Chemical and Biomolecular Engineering, KAIST, South KoreaSchool of Materials Science and Engineering University of Ulsan, South KoreaDepartment of Physics, University of Ulsan, South KoreaDepartment of Applied Physics, BIT, Mesra, Ranchi, India

r t i c l e i n f o

rticle history:eceived 31 August 2008eceived in revised form 23 October 2008ccepted 24 October 2008vailable online 6 December 2008

eywords:eramicshemical synthesisintering

a b s t r a c t

In this work, we have synthesized polycrystalline SrBi4Ti4O15 (SBiT) ceramics by soft chemical method.These ceramics were structurally characterized by analysis of X-ray diffraction (XRD) patterns, indicatesthat SBiT ceramics present an orthorhombic structure. Scanning electron micrograph shows that the grainsexhibit a plate like morphology. Dielectric relaxations of the SBiT ceramics were investigated in the tem-perature range 100–700 ◦C. Using the Cole–Cole model, an analysis of the dielectric loss with frequencywas performed, assuming a distribution of relaxation time. The presence of the peaks in temperaturedependent dielectric loss indicates that the hoping of charge carriers is responsible for the relaxation.Impedance studies shows a non-Debye type relaxation, and relaxation frequency shift to higher side withincrease in temperature. A significant shift in impedance loss peaks towards higher frequency side indi-

icrostructurempedance spectroscopy

cates conduction in material and favoring the long range motion of mobile charge carriers. The Nyquistplot shows overlapping semicircles, for grain and grain boundary of SBiT ceramics. The frequency depen-dent ac conductivity at different temperatures indicates that the conduction process is thermally activatedprocess and the spectra follow the universal power law. The hopping frequency shifts towards higher fre-quency side with increase of temperature, below which the conductivity is frequency independent. Thevariation of dc conductivity confirms that the SBiT ceramics exhibits negative temperature coefficient of

h tem

tBrRt1pw

resistance behavior in hig

. Introduction

The bismuth layer-structured ferroelectrics (BLSFs) have a crys-al structure containing interleaved bismuth oxide (Bi2O2)2+ layersnd pseudo-perovskite blocks which contains BO6 octahedral andenerally formulated as (Bi2O2)2+ (Am−1BmO3m+1). In these nota-ions A represents a mono-, bi- or trivalent ion, B denotes a tetra-,enta- or hexavalent ion, and m is the number of BO6 octahedral in

ach pseudo-perovskite block (m = 1–5) [1]. The physical propertiesf this ceramic are strongly affected by the structure and morphol-gy [2]. Sr-based layered perovskite is currently one of the mostromising candidates for a new generation of nonvolatile ferroelec-

∗ Corresponding author at: Department of Applied Physics, BIT, Mesra, Ranchi,ndia. Tel.: +91 94370 85441.∗∗ Corresponding author at: Department of Chemical and Biomolecular Engineer-ng, KAIST, South Korea. Tel.: +82 42 869 3918.

E-mail addresses: drskrout@gmail.com, skrout@bitmesra.ac.in (S.K. Rout),iwoo@kaist.ac.kr (S.I. Woo).

satpvafipfa

925-8388/$ – see front matter © 2008 Elsevier B.V. All rights reserved.oi:10.1016/j.jallcom.2008.10.125

perature.© 2008 Elsevier B.V. All rights reserved.

ric random access memories (NvFRAM) devices [3]. Among severalLSF materials, SrBi4Ti4O15 (SBiT) is extensively studied by manyesearchers from possible applications in piezoelectric device [4].ecently, much attention has been paid to SBiT due to its high Curieemperature, barrier type property, large and stable 2Pr after up to011 cycles, and anisotropic physical properties [4–6]. Because of itsroperties and performances, SBiT, prepared by various methods,as investigated by several other researchers [7,8].

Complex impedance spectroscopy (CIS) is a flexible tool forimultaneous electrical and dielectric characterization of materi-ls. This powerful technique has been widely used to characterizehe dielectric behavior of single crystal, polycrystalline, and amor-hous ceramic materials [9]. The experimental data contains threeariables, real and imaginary component of the electrical responsend frequencies together with four inter related electrical response

ormalisms: impedance (Z*), admittance (Y*), relative permittiv-ty (ε*), and electric modulus (M*). Consequently, data can beresented in many formats; in practice, it is often found that dif-

erent formats may highlight different electrical characteristics ofsample. It is the most commonly used experimental technique to

S.K. Rout et al. / Journal of Alloys and Compounds 477 (2009) 706–711 707

atactwttt(fe

2

Sasassaaatoriomo(mpbuS1qdr

3

Sopp

F

[iwtla

semdp(

affe1pDtgsamples, but it is probably true to say that the influence of theseeffects are not clearly defined. In order to carry out a more thor-ough analysis of permittivity and their temperature dependence, itis necessary to separate the contributions of the various grain and

Fig. 1. XRD patters of SrBi4Ti4O15 ceramics powder sintered at 1175 ◦C for 4 h.

nalyze the dynamics of the ionic movement in solids. Contribu-ion of various microscopic elements such as grain, grain boundarynd interfaces to total dielectric response in polycrystalline solidsan be identified by a reference to an equivalent circuit, which con-ains a series of array of parallel RC elements. The purpose of thisork are: (a) synthesis of SBiT ceramic at relatively lower tempera-

ure to avoid the volatility of Bi3+, (b) apply the variable-frequencyechnique of impedance spectroscopy to SBiT ceramic over a wideemperature range, even across the transition temperature, andc) characterize the grain and grain boundary resistance using CIS,rom which one can calculate their capacitance and can model thequivalent circuit.

. Experimental

The basic materials used in the synthesis of SrBi4Ti4O15 (SBiT) ceramic werer(NO3)2, Bi(NO2)3, TiO2 (d10 = 027 �m, d50 = 0.35 �m, and d90 = 0.48 �m) and oxaliccid. All the chemicals were more than 99.5% pure. An aqueous solution of 0.2 Mtrontium nitrate and bismuth nitrate using deionized water was prepared. Anppropriate amount of TiO2 was added to a 0.2-M oxalic acid solution and was ultra-onicated for 10 min to break TiO2 agglomeration and kept on stirrer for continuoustirring to ensure suspension. The solution containing alkaline earth metal nitratend bismuth nitrate were added drop wise into the suspension of TiO2 in oxaliccid solution under stirring. Finally the pH of the solution was maintained at 7 bydding ammonia solution. All of the above experiments were carried out at roomemperature. This process precipitated alkaline earth metal oxalate on the surfacef fine TiO2 particles by nucleation. Similar types of ceramic synthesis have beeneported earlier by the author for BaTi0.6Zr0.4O3 [10] and for BaTiO3 [11]. The result-ng precipitate was filtered out and washed repeatedly with isopropyl alcohol. Thebtained powder was calcined at 750 ◦C for 4 h and 950 ◦C for 4 h with intermediateixing and grinding. The SBiT ceramics were structurally characterized by analysis

f X-ray diffraction (XRD) patterns and the lattice parameters were calculated [12]Xpert MPD, Philips). Surface morphology was studied using a scanning electron

icroscope (SEM, S-4200, Hitachi, Japan). For electrical property measurements,ellets were prepared with 2 wt.% polyvinyl alcohol (PVA) solution addition as ainder, and was sintered at 1175 ◦C for 4 h. The density and porosity was evaluatedsing Archimedes’ principle and found to be nearly 97% of the theoretical density.ilver electrodes were applied on the opposite disk faces by dc sputter (Cressington08, Cressington, Inc., USA). Dielectric measurements were carried out over the fre-uency range 10 Hz to 10 MHz using an impedance analyzer (HP 4192A, USA). Theielectric impedance data was collected at an interval of 25 ◦C while heating at aate of 1 ◦C/min.

. Results and discussion

Fig. 1 shows the room temperature powder XRD pattern ofBiT ceramic recorded by using Cu K� radiation. The preparationf Bi layered ceramics is always troubled by the coexisting of ayrochlore phase, which is observed due to the thermal decom-osition reaction caused by the vaporization of bismuth species

Ft

ig. 2. Microstructure of SrBi4Ti4O15 ceramics powder sintered at 1175 ◦C for 4 h.

13]. In this work, there is no such evidence of the existence of anympurity phase or its quantity is too small to be detected. All peaks

ere indexed in the orthorhombic symmetry. The lattice parame-ers were found to be a = 5.4507, b = 5.4376 and c = 40.9841 Å for SBiTayered ceramic. The observed lattice parameters were in excellentgreement with that reported elsewhere [14].

Fig. 2 shows the SEM micrograph of SBiT ceramic prepared byoft chemical method. Fig. 2 shows that the grains of SBiT ceramicsxhibit a plate like morphology. It was also found from the SEMicrograph that the grains of different sizes are homogeneously

istributed. Similar grain morphology was observed in SBiT pre-ared by other methods [15]. The average lengths to thickness ratioaspect ratio) of the grain are found to be more than 15.

As a first step in data analysis, relative permittivity values ε′

t 10 kHz, 100 kHz and 1 MHz were measured in traditional fixedrequency format (figure not shown). The result showed a normalerroelectric to paraelectric transition around 525 ◦C. This ferro-lectric transition temperature for SBiT has been reported since960 [16] where author observed the extrapolated Curie Weiss tem-erature, T0, was at least 20 ◦C lower than the Curie temperature.ifference between the value of T0 and Tc have been noticed many

imes in ceramics which are attributed to the presence of differentrain size, porosity, and grain boundary impedances [17,18] in the

ig. 3. Frequency dependent real part of the complex permittivity of SBiT at differentemperatures.

708 S.K. Rout et al. / Journal of Alloys and Compounds 477 (2009) 706–711

Ff

grt

rieott(suiidtiFrf

tt

R

I

E

A

a

t

wm8lbstca

Z

Fp

wcf

ε

Tc

(

(

(

asfbsTttd

R

Hs

fnti

ig. 4. Frequency dependent dielectric loss (tan ı) at different temperatures (highrequency data are shown only).

rain boundary components. The importance of making this sepa-ation is shown in Fig. 3, in which the permittivity data, at severalemperatures is plotted as a function of frequency.

The plot showed three different behaviors in whole frequencyange. In the low frequency range the permittivity increases withncrease in temperature. This may be caused due to grain boundaryffect and presence of all type of polarization in the materials. Thebserved high value of permittivity, both at low frequency and highemperature probably due to some kind of surface layer formed byhe diffusion of the electrode materials. In the high frequency range3–10 MHz) the permittivity showed a dispersion. That dispersionhifts toward lower frequency side with increase in temperaturep to 525 ◦C and again move towards higher frequency side with

ncrease in temperature in the paraelectric phase. Similar behav-or was observed in paraelectric BaTiO3 [19] ceramic. Around thisispersion frequency dielectric loss spectrum also gives a peak andhat peak frequency follows similar behavior as that of permittiv-ty. The frequency dependence dielectric loss spectrum shown inig. 4 implies that the hopping of charge carriers plays an importantole in their transport processes because a loss peak is an essentialeature of the charge carrier hopping transport [20].

For a more detailed analysis we have adopted impedance spec-roscopic technique. Electrical ac data may be presented in any ofhe four interrelated formalism:

elative permittivity (ε∗) = ε′ − jε′′, (1)

mpedance (Z∗) = Z ′ − jZ ′′ = 1/jωC0ε∗, (2)

lectric modulus (M∗) = M′ + jM′′ = 1/ε∗, (3)

dmittance (Y∗) = Y ′ + jY ′′ = jωC0ε∗, (4)

nd

an ı = ε′′/ε′ = M′′/M′ = Z ′/Z ′′ = Y ′/Y ′′, (5)

here ω = 2�f is the angular frequency, C0 = ε0Ad−1 is the geo-etrical capacitance, j = √−1, ε0 is the permittivity of free space,

.854 × 10−14 F cm−1, d and A are the thickness and area of the pel-et. It is assumed that the impedance date can be represented ideallyy an equivalent circuit consisting of two parallel RC elements ineries. This circuit is used very widely with materials whose proper-

ies are some combination of bulk and grain boundary impedancesonnected in series. The impedance for this circuit can be describes

∗ = [R−1b + jωCb]

−1 + [R−1gb + jωCgb]

−1, (6)

mareh

ig. 5. Variation of real part of impedance of SBiT with frequency at different tem-eratures.

here the subscripts b and gb refer to bulk and grain boundary. Inases where Cb < Cgb and Rb < Rgb the real part of ε*, can be calculatedrom Eqs (1) and (6) as

′ =ε−1

gb + (ωRbCb)2(ε−1b ε−1

gb )

ε−2gb + (ωRbCb)2(ε−1

b ε−1gb )

2. (7)

he frequency dependence of ε′ for this circuit has the followingharacteristics:

a) A low frequency plateau (ε′ when ω → 0) is given by (Fig. 3)

ε′ = εgb. (8)

b) A dispersion centered on a frequency, fm, which also corre-sponds to a maximum in ε′′ and for which (Fig. 4)

fm = [2�εgb(ε−1b + ε−1

gb )RbCb]−1 = [2�Rb(Cb + Cgb)]−1. (9)

c) A high frequency plateau (when ω → ∞) is given by (Fig. 3)

ε′ = (ε−1b + ε−1

gb )−1 = εH. (10)

The high frequency and low frequency region in Fig. 3 are labeledccordingly. It is worth mentioning here that low frequency regionhould also show a plateau like behavior but in the present caseailed to detect within the limit of the impedance analyzer. It coulde noticed from Fig. 4 that high frequency plateau and the disper-ion move to higher frequency side with increase of temperature.his behavior is because Cb decreases with increase of tempera-ure above Tc and resistance (both Rb and Rgb) is expected to behermally activated and decrease with temperature with Arrheniusependence of the form

= ˇ exp(Ea/kBT). (11)

ence, the dispersion centered on a frequency, fm, increases at theame time as the value of εb decreases with increase in temperature.

Fig. 5 shows the variation of the real part of impedance (Z′) withrequency at various temperatures. It is observed that the mag-itude of Z′ decreases with increase in both frequency as well asemperature, indicating an increase in ac conductivity with risen temperature and frequency. The Z′ values for all temperatures

erge at high frequency. This is due to the release of space charges a result of reduction in barrier properties of the material withise in temperature [21,22] and may be a responsible factor for thenhancement of ac conductivity of the material with temperature atigh frequencies. Further, at low frequency Z′ values decrease with

S.K. Rout et al. / Journal of Alloys and Compounds 477 (2009) 706–711 709

Ft

irIgia

(t≥iiTtpm[

aar

�

woc

w

�

wacf

riarc(eawdNa2o

ig. 6. Variation of imaginary part of impedance of SBiT with frequency at differentemperatures.

ncrease in temperature show negative temperature coefficient ofesistance (NTCR) type behavior similar to that of semiconductors.t can also be noticed from Fig. 5 that the real part of impedance (Z′)ives a dip prior to merge at high frequency and that decreases withncrease in temperature in the paraelectric phase. This dip may bessociated with the charge carrier hopping and supports Fig. 4.

Fig. 6 shows the variation of the imaginary part of impedanceZ′′) with frequency at different temperatures. The curves show thathe Z′′ values reach a maximum peak (Z ′′

max) for the temperature400 ◦C and the position of Z ′′

max shifts to higher frequency withncrease of temperature (inset of Fig. 6). This shift in frequency max-ma indicates active conduction through the grain boundary [21].he magnitude of Z ′′

max also decreases with increase in tempera-ure. The merger of Z′′ values in the high frequency domain may beossibly an indication of the accumulation of space charges in theaterial at low frequency and at higher temperature (see Fig. 3)

22].Fig. 7 shows the variation of relaxation time with inverse of

bsolute temperature (103T−1). In a relaxation system, the relax-tion time (�) can be calculated from Z′′ versus log f plot using theelation

= 1/ω = 1/2�f , (12)

max

here fmax is the relaxation frequency. It is observed that the valuef � decreases with increase of temperature, which is a typical semi-onductor behavior. The activation energy (Ea) of this compound

Fig. 7. Log � vs. 103T−1 of the compound.

fFaafuahtbfbf(tmcecp

tAc

Fig. 8. Complex impedance plot of SBiT at different temperatures.

as calculated from Arrhenius relation:

= �0 exp(−Ea/kBT), (13)

here �0 is the pre exponential factor, kB is the Boltzmann constantnd T is the absolute temperature. The value of activation energyalculated from the slope of log � versus 103T−1 (see Fig. 7) wasound to be 1.4 eV.

Fig. 8 shows the impedance data taken over a wide frequenciesange over at several temperatures as a Nyquist diagram (complexmpedance spectrum). It is observed that with increase of temper-ture the slope of the lines decreases and their curve shift towardseal (Z′) axis and at temperature 450 ◦C and onwards two semicir-les could be traced with different values of resistance for grainRg) and grain boundary (Rgb). Hence grain and grain boundaryffects could be separated at these temperatures. These values of Rg

nd Rgb are obtained from the intercepts of the traced semicirclesith Z′ axis. It can be clearly noticed that the values of Rg and Rgbecreases with rise in temperature, which obviously indicates theTCR character of SBiT and supports Fig. 5. The capacitances (Cg

nd Cgb) due to these effects can be calculated using the relation�fmRC = 1, where fm is the frequency corresponds to the maximumf semicircle.

Resistance data for both Rb and Rgb along with their relaxationrequencies are summarized in the form of an Arrhenius plot inig. 9(a) and (b). At all temperatures, Rb Rgb and Rb and Rgb givelinear Arrhenius plot (see Fig. 9(a)) with activation energy 1.0 eVnd 1.46 eV respectively. The grain and grain boundary relaxationrequencies are recorded from the maxima of the semicircles andsed for calculation of activation energy of grains and grain bound-ries. The activation energy for grain boundaries is found to beigher than that of grains indicating higher resistive behavior thanhat of the grains. The activation energy calculated using the grainoundary resistance data (1.43 eV, see Fig. 9(a)) and their relaxationrequency (1.45 eV, Fig. 9(b)) are almost equal. Comparing the grainoundaries activation energy with the activation energy calculatedrom the frequency dependence imaginary part of the resistanceFig. 7) one can conclude that the grain boundaries resistances arehe dominating character for the overall electrical behavior of the

aterial. These observed low values of activation energy in SBiTeramic compared to other perovskites, may be due to the pres-nce of charges carriers inside the grains and some extrinsic chargearrier created due to the use of silver electrode at elevated tem-

eratures [19].

The ac electrical conductivity was obtained in accordance withhe literature [20], relation �ac = d/AZ′, where d is the thickness andis the surface area of the specimen. The log–log plot of electrical

onductivity versus frequency at different measuring tempera-

710 S.K. Rout et al. / Journal of Alloys and Compounds 477 (2009) 706–711

Fr

twotkswcdt

FS

ocbio0swpi

steTt

O

at

ig. 9. Arrhenius plot for (a) grain and grain boundary resistance, and (b) theirelaxation frequency derived from complex impedance plots.

ure is shown in Fig. 10. The plot shows conductivity increasesith increase of temperature. Frequency independent behavior

f the conductivity in the low frequency region is observed buthat become sensitive at high frequency region, which generallynown as hopping frequency, shifted towards higher frequency

ide with increase of temperature. In the higher frequency regionhere the conductivity increases is caused due to the hopping of

harge carrier in finite clusters. Frequency independent ac con-uctivity has been observed in the high temperature, indicateshe long range movement of mobile charge carriers. The three

ig. 10. Variation of ac conductivity with frequency at different temperatures forBiT.

sabtcsrmB[p

i

�

wgatwth

Fig. 11. Variation of dc conductivity as a function of temperature.

rder jump of conductivity in the studied temperature range indi-ates the enhancement of mobile charge carrier through the grainoundary which well supports the conclusion drawn from complex

mpedance spectra. The high frequency variation of �ac found tobey universal Jonscher’s power law behavior, �ac = Kωs [23], with≤ s ≤ 1, where ω is angular frequency of ac field, in the frequency

ensitive region. Nature of the plot shows a decreasing trend of sith increase of temperature. The model based on classical hop-ing of electrons over barrier predicts a decrease in the value of the

ndex s with increase in temperature.In perovskite ferroelectric materials oxygen vacancies are con-

idered to be one of the mobile charge carriers and mostly initanates, the ionization of oxygen vacancies create conductionlectrons, a process which is defined by Kroger-Ving notation [24].he excess electron and oxygen vacancies are formed in the reduc-ion reaction,

xO → 1

2O2 + V ••

O + 2e− (14)

nd they may be bond to Ti4+ in the form Ti4+ + e′ ↔ Ti3+. The forma-ion of oxygen vacancies, can be due to the three different chargetates: neutral (Vx

O) state, which it is able to capture two electronsnd it is neutral in the lattice, singly ionized (V •

O) state and dou-le ionized (V ••

O ) state, which it is not trap any electron and it iswofold positively, can be thermally activated, thus enhancing theonduction process. Double charge oxygen vacancies (V ••

O ) are con-idered to be most mobile charge in perovskite and play importantole in conduction. In addition to this the chemical inhomogeneity,ay be due to the difference in the ionic environment of Sr2+ and

i3+ [25,26]. Furthermore the sharing in the A site of perovskite andBi2O2]2+ slabs may also contributes to the conduction mechanism,iezoelectric response and photoluminescent properties [27–30].

The dc conductivity of the sample was evaluated from thempedance spectrum using the relation

dc = d/RA, (15)

here R is the resistance of the component (resistance of grain orrain boundary), d the thickness of the pellet and A is the electroderea. An Arrhenius plot of �dc against inverse of absolute tempera-ure is shown in Fig. 11. It follows a typical semiconductor behaviorith NTCR characteristics. At all temperature the conductivity of

he grain is higher than that of grain boundaries confirming theigh resistive behavior of the grain boundaries.

and C

4

csoocTnfReddttctNs

A

afirtR

R

[

[[[

[

[

[[[[[[

[

[[[[

[

[28] B.J. Kennedy, Y. Kubotab, B.A. hunter, Ismunandar, K. Kato, Solid State Comm.

S.K. Rout et al. / Journal of Alloys

. Conclusions

Polycrystalline SrBi4Ti4O15 (SBiT) ceramics were prepared suc-essfully by a soft chemical method. XRD patterns analysishowed that all the SBiT ceramics crystallize in a perovskite-typerthorhombic structure. The plate like grains morphologies werebserved with aspect ratio more than 15. Impedance analysis indi-ated the presence of grain and grain boundary effect in SBiT.he sample showed dielectric relaxation which is found to beon-Debye type and the relaxation frequency shifted to higher

requency side with increase in temperature. At all temperatures,b Rgb and Rb and Rgb give a linear Arrhenius plot with activationnergy 1.0 eV and 1.43 eV respectively. The Nyquist plot and con-uctivity studies showed the NTCR character for SBiT. The frequencyependent ac conductivity at different temperatures indicated thathe conduction process is thermally activated process and the spec-ra follow the universal power law. Long range mobility of chargearrier are playing effective role in conduction process at elevatedemperature. The activation energies from the impedance loss andyquist plot suggest oxygen vacancies/ions or defects are respon-

ible for long range motion of charge carriers.

cknowledgements

S.K. Rout is pleased to acknowledge Department of Sciencend Technology, Government of India, New Delhi, for providingnancial support through BOYSCAST postdoctoral fellowship. Theesearch is supported through Brain Korea 21 (BK21) project underhe Ministry of Education and Human Resources Development

epublic of Korea.

eferences

[1] H. Irie, M. Miyayama, T. Kudo, J. Appl. Phys. 90 (2001) 4089.

[[

ompounds 477 (2009) 706–711 711

[2] T. Jardiel, A.C. Caballero, M. Villegas, J. Ceram. Soc. Jpn. 116 (2008) 511.[3] J.F. Scott, Ferroelectric Rev. 1 (1998) 1.[4] A.Z. Simões, L.S. Cavalcante, E. Longo, J.A. Varela, C.S. Riccardi, B. Mizaikoff, Appl.

Phys. Lett. 90 (2007) 082910.[5] M.V. Gelfuso, J. Am. Ceram. Soc. 82 (9) (1999) 2368.[6] R.Q. Chu, Z.J. Xua, Z.G. Zhub, G.R. Li, Q.R. Yin, Mater. Sci. Eng. B 122 (2005)

106.[7] A.Z. Simões, M.A. Ramírez, C.S. Riccardi, E. Longo, J.A. Varela, J. Alloys Compd.

455 (2008) 407.[8] S. Luo, Y. Noguchi, M. Miyama, T. Kudo, Mater. Res. Bull. 36 (2001) 531.[9] K. Prabakar, S.P. Mallikarjun Rao, J. Alloys Compd. 437 (2007) 302.10] S.K. Rout, T. Badapanda, E. Sinha, S. Panigrahi, P.K. Barhai, T.P. Sinha, Appl. Phys.

A 91 (2008) 101.11] J. Bera, D. Sarkar, J. Electroceram. 11 (2003) 131.12] T.J.B. Holland, S.A.T. Redfern, Miner. Mag. 61 (1997) 65.13] A.Z. Simões, L.S. Cavalcante, C.S. Riccardi, J.A. Varela, E. Longo, J. Sol–Gel Sci.

Technol. 44 (2007) 269.14] A.V. Murugan, A.B. Gaikwad, V. Samuel, V. Ravi, Ceram. Int. 33 (2007)

569.15] M. Kimura, K. Shiratsuyu, A. Ando, T.S. Suzuki, Y. sakka, J Am. Ceram. Soc. 90

(2007) 1463.16] E.C. Subbarao, J. Phys. Chem. Solids 23 (1962) 665.17] H.C. Graham, N.M. Tallan, K.S. Mazdiyasni, J. Am. Ceram. Soc. 54 (1971) 548.18] T.T. Fang, H.L. Hsieh, F.S. Shiau, J. Am. Ceram. Soc. 76 (1993) 1205.19] N. Hirose, A.R. West, J. Am. Ceram. Soc. 79 (1996) 1633.20] Alo Dutta, T.P. Sinha, Phys. Rev. B 76 (2007) 155113.21] A. Kumar, N.M. Kumari, R.S. Katiyar, J. Alloys Compd., doi:10.1016/j.jallcom.

2008.01.130.22] Lily, K. Kumari, K. Prasad, R.N.P. Choudhary, J. Alloys Compd. 453 (2008)

325.23] A.K. Jonscher, J. Mater. Sci. 16 (6) (1881) 2037.24] C. Ang, Z. Yu, Z. Jing, P. Lunkenheimer, A. Loidl, Phys. Rev. B 61 (2000) 3922.25] J. Zhu, X.B. Chen, J.H. He, J.C. Shen, Phys. Lett. A 362 (2007) 471.26] A.Z. Simões, C.S. Riccardi, L.S. Cavalcante, E. Longo, J.A. Varela, B. Mizaikoff, Acta

Mater. 55 (2007) 4707.27] M.E. Fuentes, A. Mehta, L. Lascano, H. Camacho, R. Chianelli, J.F. Fernandez, L.

Fuentes, Ferroelectrics 269 (2002) 159.

126 (2003) 653.29] L. Kumari, J.H. Lin, Y.R. Ma, J. Phys. D: Appl. Phys. 41 (2008) 025405.30] D.P. Volanti, L.S. Cavalcante, E.C. Paris, A.Z. Simões, D. Keyson, V.M. Longo, A.T.

de Figueiredo, E. Longo, J.A. Varela, F.S. De Vicente, A.C. Hernandes, Appl. Phys.Lett. 90 (2007) 261913.