Magnetic control of relaxor features in BaZr0.5Ti0.5O3 and CoFe2O4 composite

Magnetic control of relaxor features in BaZr0.5Ti0.5O3 and CoFe2O4compositeMuhammad Usman, Arif Mumtaz, Sobia Raoof, and S. K. Hasanain Citation: Appl. Phys. Lett. 102, 112911 (2013); doi: 10.1063/1.4795726 View online: http://dx.doi.org/10.1063/1.4795726 View Table of Contents: http://apl.aip.org/resource/1/APPLAB/v102/i11 Published by the American Institute of Physics. Related ArticlesRe-entrant relaxor behavior of Ba5RTi3Nb7O30 (R=La, Nd, Sm) tungsten bronze ceramics Appl. Phys. Lett. 102, 112912 (2013) Electrostrictive and relaxor ferroelectric behavior in BiAlO3-modified BaTiO3 lead-free ceramics J. Appl. Phys. 113, 094102 (2013) Abnormal polarization switching of relaxor terpolymer films at low temperatures Appl. Phys. Lett. 102, 072906 (2013) Quenching-induced circumvention of integrated aging effect of relaxor lead lanthanum zirconate titanate and(Bi1/2Na1/2)TiO3-BaTiO3 Appl. Phys. Lett. 102, 032901 (2013) Relaxor behavior of ferroelectric Ca0.22Sr0.12Ba0.66Nb2O6 Appl. Phys. Lett. 102, 022903 (2013) Additional information on Appl. Phys. Lett.Journal Homepage: http://apl.aip.org/ Journal Information: http://apl.aip.org/about/about_the_journal Top downloads: http://apl.aip.org/features/most_downloaded Information for Authors: http://apl.aip.org/authors

Downloaded 22 Mar 2013 to 111.68.96.57. Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions

Magnetic control of relaxor features in BaZr0.5Ti0.5O3 and CoFe2O4 composite

Muhammad Usman, Arif Mumtaz,a) Sobia Raoof, and S. K. HasanainDepartment of Physics, Quaid-i-Azam University, Islamabad 45320, Pakistan

(Received 19 January 2013; accepted 5 March 2013; published online 21 March 2013)

We report the effect of magnetic field on the dielectric response in a relaxor ferroelectric and

ferromagnetic composite (BaZr0.5Ti0.5O3)0.65-(CoFe2O4)0.35. Relaxor characteristics such as

dielectric peak temperature and activation energy show a dependence on applied magnetic fields.

This is explained in terms of increasing magnetic field induced frustration of the polar nanoregions

comprising the relaxor. The results are also consistent with the mean field formalism of dipolar

glasses. It is found that the variation of the spin glass order parameter q(T) is consistent

with increased frustration and earlier blocking of nanopolar regions with increasing magnetic field.VC 2013 American Institute of Physics. [http://dx.doi.org/10.1063/1.4795726]

Ferroelectric materials have been known for quite a long

time and are acquiring increasing importance in modern

electronic industry where their unique properties are being

utilized in capacitors, piezoelectric transducers, and many

other interesting applications.1 However, more recently two

developments have taken place in this field, which are in-

triguing from the perspective of basic and applied physics.

The first of these is the development of relaxor ferroelectrics

that exhibit glassy features in their dielectric response,

similar in many ways to their magnetic counterparts, spin

glasses.2,3 Second, the development of composite materials

that include multiferroic materials whereby one component

is ferroelectric and the other is ferromagnetic.4 A number of

studies have been reported on the magnetic field dependence

of ferroelectric multiferroics where the dielectric response is

controlled by applied magnetic fields, with manifest applica-

tions.5 It is understood that the magnetoelectric coupling in

such composite systems arises from the lattice strains devel-

oped by the magnetic component on the application of a

magnetic field.6 However, comparable studies on relaxor

multiferroics that highlight the effects of magnetic fields on

the dielectric, and in particular the relaxor features, have not

been reported to date. Typical characteristics of relaxor fer-

roelectrics are attributed to the formation of polar nanore-

gions within which the ferroelectric order is uniform,7 while

charge, size, or structural disorder effectively leads to a

breakdown of the otherwise long range ordered regions into

these ordered and interacting/noninteracting nano-regions.

Considering the relaxor ferroelectric as a system with

in-built randomness of coupling between polar nanoregions,

we can envisage a further disorder in the composite due to

the randomness of the couplings between the polar-

nanoregions and the magnetic component. In the current

work, we address the question of how the applied magnetic

fields affect the relaxor properties of a typical relaxor-

ferrimagnetic composite and go on to show that these effects

can be described within the framework of the mean field the-

ories of spin glasses applied to relaxor systems.

The base ferroelectric material selected for this study is

the well-studied perovskite Barium Titanate BaTiO3

(BTO),8–10 which can form solid solution with additives

such as Zr, Sn, or Sr.7,11 Relaxor properties are introduced

on the addition of sufficient concentration of Zr (substituting

for Ti), this results in the successive phase transitions of BTO

(cubic to tetragonal, tetragonal to orthorhombic, and finally

orthorhombic to rhombohedral) being pinched to a single, dif-

fuse phase transition.12,13 For Zr concentrations greater than

25%, typical relaxor behavior sets in.8–10 This is manifested

by a broad and frequency dependent peak in the dielectric

constant, as a function of temperature. The peak temperature

Tm exhibits frequency dependence shifting to higher tempera-

tures with increasing frequencies. The relaxor composition

chosen was BaZr0.5Ti0.5O3, which is a type II relaxor where

the relaxor behavior is strain mediated.14 As for the ferromag-

netic component, we have selected CoFe2O4 (CFO) due to

its excellent magnetostrictive properties. The composite

ratio (BZT:CFO) studied in this work was 65:35. The

Ba(Ti0.5Zr0.5)O3 (BZ50) and composite ceramic samples

(BaZr0.5Ti0.5O3)0.65-(CoFe2O4)0.35 (BZ65C35) were prepared

by conventional sintering process described elsewhere.15

The temperature dependence of the dielectric spectra of

BZ50 (pure) and BZ65C35 (composite) in the frequency

range 0.2 to 500 kHz is shown in Fig. 1. For both samples

BZ50 and BZ65C35, the real component of the permittivity e0

passes through a broad maximum displaying the important

characteristic of a relaxor transition. Equally importantly,

strong frequency dispersion is also evident in both samples

with a clear shift in the position of the maxima with fre-

quency. With increasing frequency the peak position shifts to

higher temperature. Comparing the data for different frequen-

cies, we note that the data are coincident down to about 150 K

below which they separate out. We also note that the maxi-

mum value of dielectric constant (e0m) decreases with increas-

ing frequencies for both the compositions. Furthermore, the

peak temperature of the imaginary or loss component (e00m)

was also frequency dependent, increasing with increasing fre-

quencies (not shown). These observations correspond to a

typical relaxor and agree with the reported trend in BZ50.7 In

the case of the composite sample Fig. 1(b), it is evident that

the overall value of the dielectric constant has decreased as

a)Author to whom correspondence should be addressed. Electronic mail:

[email protected]

0003-6951/2013/102(11)/112911/5/$30.00 VC 2013 American Institute of Physics102, 112911-1

APPLIED PHYSICS LETTERS 102, 112911 (2013)

Downloaded 22 Mar 2013 to 111.68.96.57. Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions

compared to the pure BZ50 sample and the peaks in e0 have

shifted to lower temperatures. The decrease of the dielectric

constant and the peak temperature in the case of composites

may be attributed to a lack of correlation between the nanopo-

lar regions or to poor contacts due to the formation of a dead

layer between the grains.16 It has also been shown that the

position of the dielectric maxima and other relaxor features

may be affected by the sample preparation conditions.17

However, the relaxor behavior shown by the BZ50 sample is

essentially preserved in the composite sample BZ65C35.

In normal ferroelectrics, the high temperature dielectric

behavior is described by the Curie-Weiss law (T>TCW),

1

e0¼ T � TCW

C; (1)

where C is the Curie-Weiss constant and TCW is the Curie-

Weiss temperature. The fit of our data to the above expres-

sion is displayed in Fig. 2 for the measuring frequency of

500 kHz. The temperature Tdev where the data starts deviat-

ing from the Curie-Weiss law is marked in Figs. 2(a) and

2(b). Tdev was about 210 K and 188 K for the BZ50 and

BZ65C35 samples, respectively.

However, a larger range of temperature from close to

the peak and extending to high temperatures can be

described in relaxors by the modified Curie-Weiss relation,

1

e0� 1

e0m¼ ðT � T0mÞ

c

C1

; 1 � c � 2: (2)

Here e0 is the value of the permeability at the temperature Twhile e0m is the value at the peak temperature while c is the

parameter that relates to the character of the phase transition.

For the normal ferroelectrics, c¼ 1 and corresponds to the

normal Curie Weiss Law while for relaxors it is reported to

lie between 1 and 2, with c¼ 2 describing the completely

diffuse phase transition (DPT). Finally C1 in Eq. (2) is the

modified Curie-Weiss constant.

The insets of Fig. 2 show a graph between ln[(1/e0)-(1/e0m)] and ln(T-T0m) at 500 kHz. By fitting the data using Eq.

(2), we obtain values of c to be 1.67 and 1.60 for BZ50 and

BZ50C35, respectively. These values confirm typical relaxor

behavior in the samples and correlate well with the reported

values in the literature. For example, Dixit et al.18 have

reported c varying in the range 1.3–1.7 for thin films of

BaZrxTi1�xO3 for (0.3< x< 0.7). Similarly, Maiti et al.19

reported c¼ 1.89 for ceramic of same composition.

Comparing the values of different parameters, e.g. c, T0m, and

Ea (see below) with those reported,7,18,19 we presume that

our samples may be placed in the moderate relaxor category.

Considering the very close gamma values of the two samples

in this light strongly suggests that they can be regarded as

being at the same level of disorder with regard to the nature

of the phase transition.

We now consider the behavior of the dielectric response

below the peak temperature. The observed decrease in the

dielectric constant at lower temperatures is typically associ-

ated with the freezing of the polar nano regions. These

regions are understood to begin forming at elevated tempera-

tures well above the dielectric peak.7 This freezing or long

FIG. 1. Temperature dependent dielectric spectra e0 of the (a) BZ50 (b)

BZ65C35. Arrow indicates the direction of increasing frequency. � 0.2, �

1, �10, �100, � 500 kHz.

FIG. 2. 1=e0 vs. temperature measured at 500 kHz. Straight line shows the

Curie-Weiss behavior at higher temperature. Inset shows modified Curie-

Weiss law fitted for the respective samples.

112911-2 Usman et al. Appl. Phys. Lett. 102, 112911 (2013)

Downloaded 22 Mar 2013 to 111.68.96.57. Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions

relaxation times are similar to spin glass freezing in mag-

netic systems where the spin may not have enough thermal

energy to overcome the competing interactions. Thus the dy-

namics of the polar nanoregions slow down as the tempera-

ture is lowered and can be represented by the Vogel-Fulcher

relation,11

f ¼ fo exp � Ea

kBðT0m � TVFÞ

� �: (3)

Here TVF is the static (f¼ 0) freezing temperature, Ea is the

activation energy for polarization fluctuation of an isolated

nanopolar region, fo is a characteristic attempt frequency,

while T0m is the peak temperature corresponding to the fre-

quency f. The data for T0m as a function of frequency in the

range 0.2–500 kHz are shown in Fig. 3. Also shown are the

fits to Eq. (3). These fits suggest a Vogel-Fulcher (VF)

dielectric relaxation in the samples. The fitting parameters

obtained are included in Table I and are comparable with

reported ones.7,19 The VF relation followed by these samples

essentially manifests the short range interactions between the

polar nanoregions, which control the fluctuations of the

polarization above the freezing temperature, analogous to

the case of spin glasses.

We now turn to the most important feature of the current

work which is the influence of magnetic fields on the dielec-

tric properties of the composite sample. In these measure-

ments, the sample was cooled in a desired applied magnetic

field (0, 5, or 7 kOe) and the capacitance of the sample was

recorded as the temperature was lowered. The dielectric con-

stant for various applied magnetic fields and f¼ 500 kHz is

plotted in Fig. 4.

The first prominent feature apparent in the figure is the

decrease of the dielectric constant with increasing field and

this is evident over a wide temperature range. Note that this

effect of the magnetic field appears even at temperatures

well above the peak temperature and is nonvanishing, albeit

slightly decreasing, at the maximum temperature (300 K)

studied. The second feature is that all the data for the differ-

ent applied fields merge at a certain temperature below the

respective dielectric maxima. However, the most interesting

observation evident here is a significant shift in the dielectric

peak position to higher temperatures with increasing mag-

netic field. To verify the relationship of this shift with the

applied field, we show the effect of the field on the imaginary

part of the dielectric constant as well. As is evident these

data also display a systematic shift of dielectric loss peak

with the application of magnetic field. Note also that we did

not observe any effect of magnetic field in the dielectric

response of pure (non-magnetic) BZ50 samples. This obser-

vation testifies to the crucial role played by the magnetic

component of the composites in the above described effects.

We also note that the differences between the zero field and

applied field data persist for temperatures above the peak.

This is not surprising considering the various reports7 that

confirm the existence of polar nanoregions well above the

relaxor peak temperature and considering that in our temper-

ature range (T< 300 K) the ferrite is magnetically ordered.

We expect that for T>Tdev, the onset temperature for devia-

tion from the Curie-Weiss behavior, the individual nanore-

gions are expected to experience the effects of the magnetic

field induced strains. For T<Tdev the interacting nanopolar

regions would experience the effects of the magnetic field

induced strains and this would be reflected in the field de-

pendence of the dielectric constant for T>T0m.

Similar experiments were done for other frequencies as

well, covering the full range from 0.2–500 kHz. These

experiments demonstrate similar effects of applied magnetic

field on the dielectric constant. While detailed analysis of the

field dependence in the high temperature region will be

reported elsewhere, here we report the results of fitting the

Vogel-Fulcher law to the variation of the peak temperatures

with frequency, with non-zero applied magnetic field. The

values of the obtained fit parameters are shown in Table I.

We note that the values of the activation energies increase

very significantly with applied field suggesting that the

FIG. 3. lnf vs. T0m curves of � BZ50 and � BZ50C35 samples. Lines repre-

sent the fitted curve using Vogel Fulcher law.

TABLE I. Parameters calculated from Vogel Fulcher fitting for BZ50 and

BZ65C35.

Parameters fo (Hz) Ea (meV) TVF (K)

BZ50 8.40� 109 94 33.6

BZ50C35jH¼0 kOe 5.93� 108 39 61.5

BZ50C35jH¼5 kOe 5.66� 109 63 51

BZ50C35jH¼7 kOe 6.59� 1010 84 48

FIG. 4. Dielectric spectra ðe0; e00Þ of BZ65C35 taken in the presence of �

0 kOe, � 5 kOe, and � 7 kOe magnetic field.

112911-3 Usman et al. Appl. Phys. Lett. 102, 112911 (2013)

Downloaded 22 Mar 2013 to 111.68.96.57. Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions

activation of the dipolar orientations becomes more difficultas the applied magnetic field increases. This observation sug-

gests that in the presence of the applied magnetic field there

is an increased randomization of the interactions between the

nanopolar regions. As the field is applied and increased, the

resulting strains within the magnetic component are trans-

mitted to the coupled ferroelectric part, with randomness

both in their magnitude and directions. The net effect of

these randomized strains, we understand, is to make a uni-

form orientation of the various nanopolar regions more diffi-

cult, leading to enhanced frustration and consequent

blocking of the dipoles at elevated temperatures. We quan-

tify these speculative arguments by referring to the mean

field description of relaxors2,20 where the glass order param-

eter q is defined as q � hPiPji, where Pi and Pj are the

neighboring nanopolar regions dipole moments. In this

description, q can be taken to effectively represent the frac-

tion of the total clusters (nanopolar regions) that are blocked

and form a part of the glassy state. Applying the Sherrington

and Kirkpatrick (SK) infinite rage model for the spin glass to

the dipolar glasses, the order parameter q can be calculated

from2,21

v ¼ Cð1� qÞT � TCWð1� qÞ ; (4)

where C and Tcw are Curie Weiss constant and temperature

obtained from high temperature fitting data of the dielectric

constant in zero and applied magnetic fields, respectively. The

values of q(T) thus obtained at various fields are shown in Fig. 5.

It is noticeable from these data that the value of q for all

the three cases starts from zero at a temperature of about

162 K close to the temperature Tdev where the respective

e0(T) data begin to deviate from the pure Curie-Weiss behav-

ior. The vanishing of q at T¼ Tdev has been reported in the

literature.2 We note also that on the low temperature side the

values of q for all three cases merge in the temperature range

50–60 K close to the TVF determined and described earlier

(Table I). The merging of all the curves (H� 0) at low tem-

peratures (�50 K) may be associated with the freezing of the

nanopolar regions for T� TVF. Most significantly we note

that, for TVF< T< T0m, as the field increases, the value of qincreases for a fixed temperature. This suggests that our data

is consistent with the mean field description whereby in

increasing magnetic fields the randomness of the interactions

increases, thereby increasing the value of q of the composite

and implying that a greater fraction of the nanopolar regions

are blocked due to the application of the field.

Concluding, we have related our results to the descrip-

tion of relaxors using the analogy with spin glasses. In this

context, our main observations, viz. increase of the freezing

temperature and the decrease of the dielectric constant with

applied magnetic fields find a consistent explanation. The

field acts in this picture to increase the randomness of the

system by creating varying magnetostrictive strains that

affect various nanoregions in different and random ways.

The consequent increased randomness of the intercluster

interactions is expected to lead to a higher freezing tempera-

ture, in analogy with the case of spin glasses where the freez-

ing temperature is proportional to the variance of the

exchange constant distribution (Tf�DJ). The increase of this

order parameter with increasing magnetic field finds a con-

sistent explanation within this theory where the order param-

eter is a function of the scaled temperature T/DJ, where DJ is

a parameter characterizing the exchange or dipolar interac-

tion randomness. At a fixed temperature and in increasing

field (and randomness) the scaled temperature decreases, and

q is seen to increase. Finally, our studies show that the ferro-

electric relaxor behavior may be controlled by external mag-

netic fields essentially due to the transfer of strain from the

magnetic phase to the relaxor phase and the effects can be

described within the mean field theory of dipolar glasses.

The authors would like to thanks the Higher Education

Commission of Pakistan for the financial support under the

project Development and Study of Magnetic Nanostructures.

1U. Weber, G. Greuel, U. Boettger, S. Weber, D. Hennings, and R. Waser,

J. Am. Ceram. Soc. 84(4), 759 (2001).2D. Viehland, S. J. Jang, L. Eric Cross, and M. Wuttig, Phys. Rev. B

46(13), 8003 (1992).3D. Viehland, J. F. Li, S. J. Jang, L. E. Cross, and M. Wuttig, Phys. Rev. B

43(10), 8316 (1991).4C. A. F. Vaz, J. Hoffman, C. H. Ahn, and R. Ramesh, Adv. Mater.

22(26–27), 2900 (2010).5H. Zheng, J. Wang, S. E. Lofland, Z. Ma, L. Mohaddes-Ardabili, T. Zhao,

L. Salamanca-Riba, S. R. Shinde, S. B. Ogale, F. Bai, D. Viehland, Y. Jia,

D. G. Schlom, M. Wuttig, A. Roytburd, and R. Ramesh, Science

303(5658), 661 (2004); J. Hemberger, P. Lunkenheimer, R. Fichtl, H. A.

K. von Nidda, V. Tsurkan, and A. Loidl, Nature 434(7031), 364 (2005);

W. Bai, X. Meng, T. Lin, X. Zhu, L. Tian, J. Sun, and J. Chu,

Ferroelectrics 410(1), 50 (2010).6Y. S. Koo, K. M. Song, N. Hur, J. H. Jung, T. H. Jang, H. J. Lee, T. Y. Koo,

Y. H. Jeong, J. H. Cho, and Y. H. Jo, Appl. Phys. Lett. 94(3), 032903 (2009).7T. Maiti, R. Guo, and A. S. Bhalla, J. Am. Ceram. Soc. 91(6), 1769

(2008).8R. A. Cowley, S. N. Gvasaliya, S. G. Lushnikov, B. Roessli, and G. M.

Rotaru, Adv. Phys. 60(2), 229 (2011).9A. A. Bokov and Z. G. Ye, J. Mater. Sci. 41(1), 31 (2006).

10V. V. Shvartsman and D. C. Lupascu, J. Am. Ceram. Soc. 95(1), 1 (2012).11X. G. Tang, X. X. Wang, K. H. Chew, and H. L. W. Chan, Solid State

Commun. 136(2), 89 (2005).12D. Hennings, A. Schnell, and G. Simon, J. Am. Ceram. Soc. 65(11), 539

(1982).13J. Ravez and A. Simon, Eur. Phys. J.-Appl. Phys. 11(1), 9 (2000).

FIG. 5. Variation of order parameter q(T) at different applied magnetic

field. Solid arrow indicates the direction of increasing applied magnetic field

(� 0 kOe, � 5 kOe and � 7 kOe).

112911-4 Usman et al. Appl. Phys. Lett. 102, 112911 (2013)

Downloaded 22 Mar 2013 to 111.68.96.57. Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions

14T. Maiti, R. Guo, and A. S. Bhalla, Ferroelectrics 425(1), 4 (2011).15I. Kagomiya, Y. Hayashi, K.-I. Kakimoto, and K. Kobayashi, J. Magn.

Magn. Mater. 324(15), 2368 (2012); C. M. Kanamadi, J. S. Kim, H. K.

Yang, B. K. Moon, B. C. Choi, and J. H. Jeong, Appl. Phys. A-Mater. Sci.

Process. 97(3), 575 (2009).16R. Lin, T.-B. Wu, and Y.-H. Chu, Scr. Mater. 59(8), 897 (2008).17M. Bogs, H. Beige, P. Pitzius, and H. Schmitt, Ferroelectrics 126 (1),

197 (1992); M. D. Glinchuk, V. A. Stephanovich, B. Hilczer, J.

Wolak, and C. Caranoni, J. Phys.: Condens. Matter 11(32), 6263

(1999).18A. Dixit, S. B. Majumder, R. S. Katiyar, and A. S. Bhalla, J. Mater. Sci.

41(1), 87 (2006).19T. Maiti, R. Guo, and A. S. Bhalla, J. Appl. Phys. 100(11), 114109 (2006).20A. R. Akbarzadeh, S. Prosandeev, E. J. Walter, A. Al-Barakaty, and L.

Bellaiche, Phys. Rev. Lett. 108(25), 257601 (2012).21S. Nagata, P. H. Keesom, and H. R. Harrison, Phys. Rev. B 19(3), 1633 (1979).

112911-5 Usman et al. Appl. Phys. Lett. 102, 112911 (2013)

Downloaded 22 Mar 2013 to 111.68.96.57. Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions