Structural, electrical and magnetic properties of Bi-substituted Co2MnO4

Materials Science and Engineering B 163 (2009) 48–56

Contents lists available at ScienceDirect

Materials Science and Engineering B journal homepage: www.elsevier.com/locate/mseb

Structural, electrical and magnetic properties of Bi-substituted Co2 MnO4 N.E. Rajeevan a , Ravi Kumar b,∗ , D.K. Shukla c , P.P. Pradyumnan a , S.K. Arora d , I.V. Shvets d a

Department of Physics, University of Calicut, Kerala 673635, India Materials Science Division, IUAC, New Delhi 110067, India c Department of Physics, Aligarh Muslim University, Aligarh 202002, India d CRANN, School of Physics, Trinity College Dublin, Dublin 2, Ireland b

a r t i c l e

i n f o

Article history: Received 24 January 2009 Received in revised form 31 March 2009 Accepted 2 May 2009 PACS: 61.10.Nz 72.80.Ga 77.22.−d 77.22.Gm 77.80.−e 77.80.Bh Keywords: Spinel oxide Ferroelectric Relaxor behaviour dc conductivity ac conductivity Ferrimagnetism and multiferroic

a b s t r a c t Structural, electrical and magnetic properties of single phase Bix Co2−x MnO4 (0 ≤ x ≤ 0.3) spinel materials synthesized by solid state reactions were studied. All the samples exhibit single phase with cubic spinel structure (space group Fd3m) and the lattice parameter increases with the Bi-substitution. The grain size was also observed to increase with the Bi-substitution. All the samples exhibit the semiconducting behaviour and overall resistivity decreases with the increase in the Bi-substitution. The dc as well as ac conductivity data were analyzed in the light of various conductivity models. The dc conductivity data is explained using variable range hopping (VRH) model. All the samples show diffused ferroelectric (FE) transition and follow the Debye-type relaxation, whereas ferroelectric transition temperature (TC ) increases with the Bi-substitution. The ac conductivity calculated from the dielectric data as a function of temperature and frequency demonstrate the cross-over from small polaron tunneling (SPT) to correlated barrier hopping (CBH) type conduction in these materials. The influence of cation composition on the magnetic properties of Bix Co2−x MnO4 (0 ≤ x ≤ 0.3) mixed cubic spinel system has been studied by dc magnetization. Soft magnetic type behaviour was observed in the Bi-substituted samples, showing magnetic dilution and the increased value of saturation magnetization suggests the presence of canted spin structure due to the incorporation of Mn and Bi. Nevertheless, ferrimagnetic (FM) nature of the Co2 MnO4 is preserved in the Bi-substituted samples. The coexistence of ferroelectricity and ferrimagnetism in these materials is attributed to the off centering of cations that result in non-centrosymmetric arrangement and canted spin structure. These materials are promising candidates for multiferroic applications. © 2009 Elsevier B.V. All rights reserved.

1. Introduction Mixed transition metal oxides with general formulae AB2 O4 (A = Zn, Fe, Co, Ni, Mn, Mg, Cd, etc.; B = Co, Fe, Cr, Al, Ga, Mn, etc.) where A and B represent divalent and trivalent cations respectively, are time honored systems owing to their interesting magnetic and electric properties, which can be tuned accurately by changing the chemical composition through cation redistribution [1–5]. The crystal structure of a cubic spinel AB2 O4 belongs to Fd3m (site symmetry Oh 7 ) space group. Each unit cell contains 8[AB2 O4 ] formula units and therefore 32 O2− ions. This close packing contains 64 tetrahedral interstices and 32 octahedral interstices coordinated with O2− ions. The charge distribution of the normal spinel is represented by [A2+ ]8a [B2 3+ ]16d [O4 2− ]32e , in which Wyckoff positions 8a denote the tetrahedral sites and 16d the octahedral sites surrounded by O2− ions at 32e sites. Empty interstitial space is

∗ Corresponding author. Tel.: +91 11 26893955; fax: +91 11 26893666. E-mail address: [email protected] (R. Kumar). 0921-5107/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.mseb.2009.05.003

comprised of 16 octahedral sites (16c) and 56 tetrahedral sites (8b and 48f). All spinel-like structures are characterized by an oxygen parameter u having a value around 0.375. For u = 0.375, the arrangement of O2− ions is exactly a cubic close packing (FCC structure). In actual spinel lattice, there is deviation from the ideal pattern, with u > 0.375, resulting in a deformation of oxygen tetrahedrons and octahedrons [6,7]. Depending on the species and stoichiometry of cations occupying A and B sites, some of the spinel phase may show multiferroic property. In case of magnetic spinel chromites, appearance of ferroelectricity has been explained based on the conical spin modulation [8]. Spin related ferroelectricity was also observed in many distorted perovskite type magnetic systems [9,10]. Bismuth based magnetic perovskites were investigated by many research groups and their ferroelectric and ferromagnetic behaviour was attributed to the covalent bonding between bismuth and oxygen atoms [11,12]. For the mixed spinel oxide, LiFeTiO4 , the dispersion behaviour in dielectric property and conductivity are thermally activated quantities and are attributed to the processes controlled by an activation energy related to the hopping of mobile cations [13,14].

N.E. Rajeevan et al. / Materials Science and Engineering B 163 (2009) 48–56

Koops [15] explained the dispersion of dielectric properties using a phenomenological theory in which the dielectric spinel oxide is assumed to consist of well-conducting grains separated by poorly conducting layers, with current flowing along the parallel alignments of grains. Further the dielectric behaviour of MgAl2 O4 spinel was explained by Maxwell–Wagner interfacial type polarization [16], which was in agreement with Koop’s theorem. The geometrical frustration of the lattice degrees of freedom is considered as the origin of dispersive dielectric behaviour of cubic sulpho-spinels [17]. There exist various spinels, which exhibit variety of electrical and magnetic properties. In the present study, our basic aim is to design a cubic spinel in which the magnetic and ferroelectric properties can coexist. In light of this Co3 O4 is chosen, a very interesting material, which exhibits an antiferromagnetic ordering at very low temperature (TN ∼ 20 K) and two non-equivalent sites of Co. In this material, the octahedrally coordinated Co3+ exists in low spin state and the antiferromagnetic ordering is due to weak Co2+ –Co2+ interactions. In the above milieu, it will be interesting to study the effect of Bi-substitution on cobalt based manganites because their electrical and magnetic properties are influenced strongly by the change in cation distributions induced by the competition towards the octahedral preference between the pair of Co2+ /Co3+ vs. the new substituted cation [18,19]. The semiconducting properties of these spinels can be described as in the case of manganites, where the conductivity reaches a maximum value due to the hopping of electrons when the numbers of cations with different valence states are equal for the octahedral/tetrahedral sites. More precisely in the spinel compounds, the difference in valence state at two different sites by a unit charge is found to favour electrical conduction [20,21]. Recently, we have reported [22] on the strong magnetoelectric coupling in Bi-substituted Co2 MnO4 , which persists up to 200 K. Coexistence of ferroelectric and magnetic properties in these materials suggests a strong potential for multiferroic applications. In this work, we present in detail the structural, electrical and magnetic properties of Bi-substituted Co2 MnO4 . 2. Experimental The samples were synthesized by standard solid state reaction technique from the pre-calcined mixture containing bismuth oxide (Bi2 O3 ), cobalt oxide (Co3 O4 ) and manganese oxide (MnO2 ) with purity >99.99%. The compositions were calcined at 820 ◦ C for 24 h. All the calcined compositions were uniaxially dye-pressed into pellets of size 10 mm in diameter and 2 mm in thickness. Sintering was performed at 1000 ◦ C for 24 h, with intermediate grinding. Room temperature powder X-ray diffraction (XRD) studies of the samples were performed using a Bruker D8 X-ray diffractometer with Cu K˛ radiation in 2 range of 15–75◦ . Morphological and microstructural features were investigated with a scanning electron microscope, SEM Hitachi (Japan) model No.: S3400N and quantitative analysis was made with an electron beam microanalyser equipped with an energy dispersive X-ray analysis (EDAX) capability. Temperature and frequency dependent dielectric measurements of the pellets were performed with a HP4192 precision LCR meter. For the dielectric measurement, surfaces of the pellets were polished and silver coated, which were kept at 200 ◦ C for 2 h for stabilization of electrodes. The dc conductivity measurements were performed by standard four-probe technique in the temperature range 150–400 K. Temperature was controlled with ±50 mK accuracy using a Lakeshore temperature controller. The magnetization measurements were carried out using the VSM option of PPMS set up within a temperature range, 5–300 K in zero field cooled (ZFC) and field cooled (FC) modes under a constant magnetic field (H = 1 kOe). Field dependent magnetization measurements (M–H) were performed at various fixed temperatures.

49

Fig. 1. X-ray diffraction pattern for the samples Bix Co2−x MnO4 (x = 0, 0.1, 0.2 and 0.3) measured at 300 K. Inset shows the variation of ‘a’ with x.

3. Results and discussion 3.1. Structural and morphological characterizations The X-ray diffraction (XRD) patterns obtained at room temperature for Bix Co2−x MnO4 , with x = 0.0, 0.1, 0.2 and 0.3 are shown in Fig. 1. Indexing and refinement of XRD patterns indicate the presence of a single-phase polycrystalline structure for the synthesized materials. The X-ray patterns confirm the existence of spinel structure with the reflection arising from 1 1 1, 2 2 0, 3 1 1, 2 2 2, 4 0 0, 4 2 2, 3 3 3, 4 4 0 and 5 3 3 planes. All the samples have good crystallinity and can be indexed with the Joint Committee on Powder Diffraction Standards (JCPDS) file 80-1544, in face centered cubic (FCC) structure. The lattice parameters ‘a’ obtained from the analysis of XRD pattern for the samples with cubic unit cell for different compositions are given in Table 1. From Fig. 1, the shift of planes towards lower 2 with the increase in Bi-substitution is clearly observed, indicating the lattice expansion due to the substitution of Co3+ by Bi3+ . This could be attributed to the fact that ionic radii of Bi3+ (1.17 Å) are greater than that of Co3+ (0.65 Å) and Mn3+ (0.785 Å). From the XRD results we infer that Bix Co2−x MnO4 is having single phase cubic spinel structure (space group Fd3m) and the lattice constant increases almost linearly with Bi content up to 0.2, obeying Vegard’s law [23], as shown in the inset of Fig. 1. However, for x = 0.3, a small increase in ‘a’ is noticed (did not follow the linear behaviour), which suggests that the system has approached to the

Table 1 Lattice constant, average grain size, ε & tan ı (at 100 kHz and 300 K), and ferroelectric transition temperature (at 100 kHz) for Bix Co2−x MnO4 . Composition

Lattice constant, a (Å)

Grain size, SEM (␮m)

ε

tan ı

FE, TC (K)

x = 0.0 x = 0.1 x = 0.2 x = 0.3

8.27268 8.31045 8.32425 8.32796

1.75 3.82 5.90 7.42

541 186 944 620

2.596 1.593 2.085 2.707

– 366 412 420

50

N.E. Rajeevan et al. / Materials Science and Engineering B 163 (2009) 48–56 Table 2 The compositional analysis of Bix Co2−x MnO4 , extracted from EDAX. Nominal composition

Bi

Co

Mn

Actual composition

Bi0.0 Co2.0 MnO4 Bi0.1 Co1.9 MnO4 Bi0.2 Co1.8 MnO4 Bi0.3 Co1.7 MnO4

0.00 0.09 0.18 0.26

2.0 1.9 1.8 1.7

1.0 1.0 1.0 1.0

Co2 MnO4 Bi0.09 Co1.9 MnO4 Bi0.18 Co1.8 MnO4 Bi0.26 Co1.7 MnO4

solubility limit of Bi in the system. In the above circumstances, it is not unfair to say that only a small amount of Bi3+ can be added considering its solubility in the spinel system. Fig. 2 shows the SEM micrographs of the Bix Co2−x MnO4 (0.0 ≤ x ≤ 0.3) samples taken at room temperature. The SEM images show a uniformly developed grain morphology and dense microstructure. The micrographs of the samples with x > 0 (Bi-substituted) showed relatively greater homogeneity in the microstructure along with increased grain size. Histograms shown in the inset of each SEM micrographs quantitatively represent the increased grain size (∼1–8 ␮m) due to Bi-substitution. This suggests that Bi-substituted compositions have higher fraction of bigger crystalline grains. The stoichiometry of the above compositions was calculated by means of an electron beam microanalyser (EDAX) attached to SEM and is shown in Table 2. It is important to mention that, for higher Bi concentration, stoichiometry of the sintered composition is found to deviate slightly from that of the nominal composition.

3.2. dc conductivity studies The dc conductivity measured in the temperature range 150–400 K exhibits semiconducting behaviour, with resistivity reducing from 600 -cm to 300 -cm at room temperature, on Bisubstitution (Fig. 3). For the sample with x = 0, the resistivity is of the order of k-cm. The decrease in resistivity can be attributed to the Bi-induced stabilization of the solid solution instead of less conducting pure phase and an increase of grain size that reduces the number of grain boundaries. Another consideration that can be made to support the increased dc conductivity in Bi substituted system is slight oxygen deficiency (see Table 2). The basic conduction mechanism is taken as the hopping of electrons between non-equivalent valence states existing between octahedral and tetrahedral sites. The dc conductivity behaviour can be explained

Fig. 2. SEM micrographs of Bix Co2−x MnO4 , for x = 0.0, 0.1, 0.2 and 0.3. Insets show the respective histograms of grain sizes. Fig. 3. The plot of dc resistivity as a function of temperature for Bix Co2−x MnO4 (x = 0, 0.1, 0.2 and 0.3) with x = 0.0, 0.1, 0.2 and 0.3.

N.E. Rajeevan et al. / Materials Science and Engineering B 163 (2009) 48–56

Fig. 4. log  dc T vs. 1000/T plots for Bix Co2−x MnO4 (x = 0, 0.1, 0.2 and 0.3). The dashed line fit to the data obtained using Eq. (1).

using the Arrhenius relation [24]:

 E  act

dc T = 0 exp −

(1)

kB T

where  0 is the pre-exponential factor and is a characteristics of the material depending on lattice parameter and phonon frequency, kB is Boltzmann’s constant, T is absolute temperature and Eact is the activation energy associated with the hopping process. Eact calculated from the plots of log  dc T vs. 1000/T (Fig. 4) for all the compositions is given in Table 3, and is found to decrease with Bisubstitution in the higher temperature region. The deviation from linearity in the log  dc T vs. 1000/T plots at low temperatures, 290 K for x = 0.0 and 212 K for x = 0.3 (see Fig. 4), indicates that the small polaron hopping model is valid only at relatively higher temperatures for lower Bi-substitution and over a wider temperature range for higher Bi-substitution. Table 3 Activation energy values calculated from dc conductivity, relaxation loss, for Bix Co2−x MnO4 . Composition

dc activation energy (eV)

Loss relaxation activation energy (eV) (100–1100 kHz)

x = 0.0 x = 0.1 x = 0.2 x = 0.3

0.1813 0.1651 0.1628 0.1591

– 0.1511 0.1492 0.1493

51

Fig. 5. log  dc vs. 1/T1/4 plots for Bix Co2−x MnO4 (x = 0, 0.1, 0.2 and 0.3).

In the low temperature region, the conduction may follow the variable range hopping (VRH) model [25–27]. According to VRH Model, the dc conductivity ( dc ) is proportional to exp (−T0 /T)1/4 , T0 being the characteristic temperature. Fig. 5 shows the variation of log  dc vs. T−1/4 which fits linearly rather well for x = 0, indicating the validity of variable range hopping mechanism. According to VRH theory [25], kB T0 = 18˛3 /N(E), where N(E) is the density of states and 1/˛ corresponds to the localization length. From Fig. 5, we have found that the value of T0 increases with Bi-substitution (T0 1/4 varying from 100 for x = 0.0–118 for x = 0.3), which suggests the modification of effective localization length and the density of states. 3.3. Dielectric constant and dielectric loss Fig. 6(a)–(d) shows the dielectric constant (ε ) as a function of temperature (T) at various frequencies (100 kHz–1.1 MHz), for Bix Co2−x MnO4 (0.0 ≤ x ≤ 0.3) samples. It is clearly evident from Fig. 6(a) that for x = 0 composition, the dielectric constant remains almost constant up to 204 K at 100 kHz, then increased sharply with the increase in temperature. Within the investigated temperature range, no peak is observed in ε vs. T. With the Bi-substitution (x = 0.1) the overall dielectric constant values are increased and the temperature up to which the dielectric constant remains constant is also increased to T = 230 K (for 100 kHz). However, a well defined broad peak-like structure (FE transition) appears at a temperature

52

N.E. Rajeevan et al. / Materials Science and Engineering B 163 (2009) 48–56

Fig. 6. The plot of dielectric constant, ε as a function of temperature for Bix Co2−x MnO4 (x = 0.0, 0.1, 0.2 and 0.3). The top inset shows the composition dependence of ε .

TC = 366 K. This peak temperature is found to shift towards a higher temperature with the increase in frequency. With a further increase in the Bi-substitution (x = 0.2), the dielectric behaviour is modified and exhibits two transitions, first transition is observed at 307 K and the second one at 412 K, for 100 kHz. The peak-like structures in temperature dependent ε and the shift in transition temperature (corresponds to ε m ) towards higher temperature with the increase in frequency indicate that these materials exhibit relaxor type ferroelectric behaviour. The first transition has strong relaxor behaviour, whereas the second one does not change appreciably with the frequency. It has been observed that the dielectric constant falls slightly for higher Bi-substituted compound (x = 0.2 and x = 0.3). Also, for x = 0.3, low temperature FE transition is found to disappear and high temperature transition became almost nonrelaxor type. This may be due to the polarization associated with the 6s2 lone pair electrons of Bi, and stabilization of grain growth. Notable broadening at FE transition indicates a diffusive phase transition. Inset of Fig. 6 shows the value of the dielectric constant as a function of Bi-substitution measured at 300 K and 100 kHz. The temperature dependence of the dielectric loss, tan ı, for all the samples at selected frequencies (100 kHz–1.1 MHz) is shown in Fig. 7. At low temperature (below 200 K), tan ı is found to be almost independent of frequency for all the samples. However for a given composition and temperature, ε and tan ı were found to decrease with the increase in the frequency. The overall higher values of tan ı are observed in all Bi-substituted samples, which may be due to an increase in dc conductivity of the materials with the Bisubstitution. The temperature, at which the peak in tan ı appears,

Fig. 7. The plot of tan ı as a function of temperature for Bix Co2−x MnO4 (x = 0, 0.1, 0.2 and 0.3). The top inset shows the composition dependence of tan ı and lower insets are depicting the loss relaxation.

shifts with frequency. Insets of Fig. 7 for the doped samples show the plots between frequency vs. 1000/T i.e. Arrhenius relation, which again confirms the typical relaxor type behaviour. Loss relaxation energy obtained from these Arrhenius plots is given in Table 3. The strong frequency dependent dielectric properties are attributed to the interaction among the free carriers (electrons or holes) with potential barriers at grain boundaries, resulting in the enhancement of conductivity [28]. Among the various contributions to polarization such as electronic, ionic, dipolar and space charge polarizations, the contribution of space charge depends upon the purity of the system. In the present case, Bi ions are likely to occupy octahedral positions and create bonding defects. Capacitance measurements on these samples have been carried out in the high frequency range starting from 100 kHz to 1.1 MHz, greater than the frequencies corresponding to the time constants (RC) suggested by the Catalan [29], to rule out the possibility contribution from the Maxwell–Wagner interfacial polarization towards capacitance which is dominant at low frequencies. At high frequencies most of the systems exhibits the intrinsic capacitance and need not be an artifact effect. In a paper by Wang et al. [30], two kind of capacitive contribution above 100 K in TbMnO3 system by grains (dipolar effect i.e. hopping of charge carriers within the grain) and grain boundary effect (i.e. internal layer barrier capacitor, IBLC) was observed and the system follows the universal dielectric response (UDR). In UDR model, localized charge carriers hopping between spatially fluctuating lattice potentials not only produce the conductivity but also give rise to dipolar

N.E. Rajeevan et al. / Materials Science and Engineering B 163 (2009) 48–56

53

with the composition (shown in the top insets of Figs. 6 and 7). It is due to the stabilization of structure with the increase in Bisubstitution up to x = 0.2. Also there is chance of the formation of vacancies at the octahedral sites as the Bi content increases. In our earlier report [22], the capacitance vs. voltage (C–V) measurements of the Bix Co2−x MnO4 with x = 0–0.3 (using an applied voltage of 15 V) showed a typical butterfly-shaped loop in the C–V curves (very narrow for x = 0), which demonstrates the ferroelectric nature of the substituted samples. Similar to the dielectric constant (ε ) measurement, the C–V curves show diffuseness that increases with x, and may be attributed to the strain induced by Bi-substitution, owing to its larger ionic radius compared to the other cations. 3.4. ac conductivity

Fig. 8. Plot of log (fε ) vs. log f for the Bix Co2−x MnO4 (x = 0, 0.1, 0.2 and 0.3) at room temperature.

The ac conductivity ( ac ) was studied over a frequency range of 100 kHz–5 MHz, for temperature varying from 150 K to 425 K. The  ac was calculated from the dielectric data, using following equation: ac = ε0 ε ω tan ı

effects. If it is true, at a given temperature a linear behaviour should be obtained in the plot of log (fε ) vs. log (f), where f is the frequency. Fig. 8 represents the log (fε ) vs. log (f) plot for Bix Co2−x MnO4 samples at room temperature and in the range 100 kHz–1.1 MHz, where exact straight line behaviour is observed for undoped system i.e. Co2 MnO4 , whereas for doped samples it is not the case. This analysis straight away rules out the possibility of any kind of conductivity contribution in dielectric constant as per UDR in Bi-substituted system. Again it is well known that the exponential increasing background in loss tangent implies that the background is associated with the hopping conductivity. The loss tangent vs. temperature shown for Bi-substituted Co2 MnO4 does not show the exponential increase up to a reasonably high temperature and further support to the fact that these samples do not follow the UDR behaviour. From the above mentioned results it is evident that the present system exhibit relaxor ferroelectric behaviour for all the samples. To understand the nature of relaxor behaviour, the well known method is to construct the Cole–Cole plot [31]. In this conventional method ε is plotted as a function of ε (imaginary part of dielectric constant, i.e. dielectric loss factor) to get information on the distribution of relaxation times. For the Debye-type relaxor behaviour the data fits into a semicircular curve. However, the Cole–Cole plots for Bix Co2−x MnO4 (0.0 ≤ x ≤ 0.3) samples are not semicircular (not shown here), owing to conductivity of the samples [32]. From the plots representing the temperature dependence of the real part of the dielectric constant, ε at different frequencies (Fig. 6), it is observed that ε varies very slowly with temperature up to a temperature  D /2, where  D is the Debye temperature [33]. Above  D /2, the values of ε for all the samples increase with increase of temperature and decreases with an increase in the frequency. Such a behaviour again indicates the Debye-type dielectric relaxation process. Frequency dependent shift in ε m (dielectric constant maxima) towards the higher temperature was also exhibited by the samples, up to x = 0.3, as expected in relaxor ferroelectric materials. The dielectric constant, ε of the Bix Co2−x MnO4 samples with x = 0, was about an order of magnitude smaller as compared to samples with x = 0.1 and 0.2. The FE transition was not visible from the ε vs. T plots, within the high temperature limit (450 K) of our experimental setup. In other samples (Bix Co2−x MnO4 with x = 0.1, 0.2 and 0.3), FE transition was observed within the high temperature limit of the setup, with increased diffuseness. Finally, the variation in the dielectric constant and dielectric loss appears to be non-linear

(2)

 ac is found to increase for the Bix Co2−x MnO4 (0.0 ≤ x ≤ 0.3) samples with a greater Bi content especially in the high frequency region (Fig. 9). This indicates that the movement or hopping of the charge carriers is influenced by the neighborhood, i.e. enhanced by presence of Bi3+ ions. At higher frequencies (∼2 MHz),  ac is decreased

Fig. 9. The plot of log  ac vs. log ω at 300 K, for Bix Co2−x MnO4 (x = 0, 0.1, 0.2 and 0.3).

54

N.E. Rajeevan et al. / Materials Science and Engineering B 163 (2009) 48–56

for all compositions (not shown here), indicating the inability of hopping of charges to follow the high frequency of the applied field. The total conductivity of the material at a frequency can be written as: total (ω) = dc + ac (ω)

(3)

where hopping conduction arising from Bi-substitution is responsible for the enhancement of both dc and ac conductivities. The dc conductivity is the ω → 0 limit of  total (ω). For semiconductors and disordered systems the ac conductivity follows the power law behaviour [34]: ac (ω) = A ωs

(4)

where A is temperature dependent constant and s is the frequency exponent (s ≤ 1). From the log–log plots drawn for  ac vs. ω (Fig. 9) for the different compositions at room temperature, the slope directly provides the value of dimensionless frequency exponent, s, and it is clear that frequency exponent s varies with Bi content. The ac conductivity also shows frequency dependent dispersion similar to dielectric behaviour discussed earlier. In the case of disordered solids,  ac is an increasing function of frequency. The temperature dependence of frequency exponent s is shown in Fig. 10. Frequency exponent s is a measure of correlation between  ac and frequency. For random hopping of carriers, s should be 0 (frequency independent  ac ) and tends to 1, as correlation between

Fig. 11. Zero field cooled (ZFC) and field cooled (FC) magnetization data plotted as a function of temperature of Bix Co2−x MnO4 (x = 0, 0.1. and 0.3). Insets show the respective temperature dependence of 1/M .

 ac and ω increases. Qualitatively small polaron tunneling (SPT), usually associated with an increase of s with the increase in temperature indicates the activated behaviour of polarons, which is independent of intersite separation. On the other hand correlated barrier hopping (CBH) shows a decrease in s with increasing temperature, indicating the thermally activated behaviour of electron transfer over the barrier between two sites having their own Coulombic potential wells. Temperature dependence arising from the correlation between barrier height and intersite separation are the characteristics of CBH model [35]. From the temperature dependence plots of s (Fig. 10), it is clear that for pure Co2 MnO4 , the frequency exponent, s increases up to 260 K and then decreases. For this sample, SPT model is suitable up to 260 K at which maximum correlation is exhibited between  ac and ω. Above 260 K, CBH model is valid, where dc conductivity contribution is not negligible. For the Bi-substituted samples also, SPT and CBHs models seem to fit, which explain the conduction mechanism with better correlation at higher temperature ∼ 320 K. 3.5. Magnetization measurements

Fig. 10. Temperature dependence of frequency exponent, s for Bix Co2−x MnO4 (x = 0, 0.1, 0.2 and 0.3).

Temperature dependence of the field cooled (FC) and zero field cooled (ZFC) magnetization of Bix Co2−x MnO4 , with x = 0.0, 0.1 and 0.3, in the presence of an applied magnetic field of 0.1 T are shown in Fig. 11. It has been found that these samples obey the Curie–Weiss law and exhibits ferrimagnetic transition in the temperature range 180–186 K depending on Bi content with a divergence of FC and ZFC at lower temperatures (∼143–147 K). Below the

N.E. Rajeevan et al. / Materials Science and Engineering B 163 (2009) 48–56

55

to originate from Mn3+ /Mn4+ ions occupying the octahedral sites. In general, the saturation magnetization Ms shows a monotonic increase with increasing grain size. Bi-substituted samples are having large grain size resulting the increased Ms . Even though weak, super exchange interaction between A sites (Co2+ –Co2+ ) through O2− and Co3+ ions will have an effect on magnetic properties. In the present series, a sizeable portion of Co3+ is replaced by Mn3+ and Bi3+ ions. The exchange interaction is strongly influenced by the greater ionic radius of Bi3+ , that distorts the oxygen octahedron and causes A-site frustration, resulting in an incommensurate magnetic ordering in the system. In other words the presence of Bi and Mn at the cation sites modifies the magnetic behaviour of the parent compound Co3 O4 because of the competition among intrasite (JAA and JBB ) and intersite (JAB ) interactions that leads to the magnetic frustration. This disrupts the antiferromagnetic ordering in Co2+ sublattice and results in ferrimagnetism [37,38]. Analysis of temperature dependence of inverse susceptibility showed that Curie–Weiss temperature,  CW is highly negative and found varied with Bi-substitution, indicating the frustration of antiferromagnetic interactions in the compound, besides the ferrimagnetic character, which sets at TC . So the compounds show ferrimagnetism, controlled by Bi-substitution. 4. Conclusions

Fig. 12. Isothermal magnetization hysteresis of Bix Co2−x MnO4 (x = 0, 0.1 and 0.3) at 5 K and 120 K.

divergence point the FC magnetization increases linearly, whereas ZFC curve exhibit a wide maximum around 133 K. Inverse susceptibility (1/M ) obtained for these samples are plotted in the insets of Fig. 11, which revealed that the ferrimagnetic transition occurred at slightly varied temperature depending on Bi concentration. Isothermal magnetization hysteresis for Bix Co2−x MnO4 (0.0 ≤ x ≤ 0.3) samples have been performed at various temperatures below the transition temperature. In Fig. 12 the M–H plots at T = 5 and 120 K for Bix Co2−x MnO4 (0.0 ≤ x ≤ 0.3) samples are shown, hysteresis loops clearly depicting ferrimagnetic behaviour. Larger loop area at lower temperature indicates increased magnetization. At all measured temperatures, saturation magnetization (MS ) is found to be increasing with the increasing Bi content. Remnant magnetization (MR ) observed in these compounds was not so high. The coercive field (HC ) was found to decrease with Bi-substitution owing to a larger grain size, which results in the reduction of uncompensated spins. The appearance of ferrimagnetism in this compound may be attributed to either the canting of the antiferromagnetically ordered spins by the structural distortion [36] or a breakdown of the balance between the antiparallel sublattice magnetization of Co2+ due to the substitution of Mn or Bi ions with different valence states [37]. Regarding the magnetic behaviour of spinel oxides, there are mainly three types of magnetic interactions possible between the ions at A-site and B-site through the intermediate oxygen ions (O2− ) via super exchange interaction. It has been verified experimentally that these interaction energies are negative favouring antiferromagnetism when the d orbitals of the metal ions are half filled or more than half filled, while a positive interaction accompanied by ferrimagnetism appears when d orbital is less than half filled. So ferrimagnetism in the samples prepared is considered

(1) Polycrystalline bulk samples of Bix Co2−x MnO4 (0 ≤ x ≤ 0.3) have been successfully synthesized by standard solid state reaction technique. (2) The XRD analysis of Bix Co2−x MnO4 (0 ≤ x ≤ 0.3) indicates that all samples exhibit a single phase nature. All the samples are indexed in cubic spinel structure and lattice parameter ‘a’ is found to increase with Bi content. Analysis of XRD data and SEM micrographs shows that Bi content x > 0.3 is difficult to accommodate, while retaining the cubic spinel structure. (3) The temperature dependent dielectric data represents a diffused ferroelectric phase transition; with an increasing ferroelectric TC as the Bi content increases. Bi-substituted samples have moderately high dielectric constant depicting better ferroelectric characteristics along with Debye-type relaxation behaviour. (4) The dispersion of ac conductivity has been estimated in terms of frequency exponent s, which varies with temperature and is explained using small polaron tunneling (SPT) model and correlated barrier hopping (CBH) model. (5) The dc conductivity analysis reveals the semiconducting nature of the spinel compounds. The hopping of charges between cations with different valence states at the octahedral sites is considered to be the origin of electrical conductivity of the Bisubstituted Co2 MnO4 . (6) The dc magnetization measurements show that all the samples are ferrimagnetically ordered, with TC < 190 K and saturation magnetization increases with increasing the concentration of Bi3+ ions due to the redistribution of cations at the octahedral sites and consequent frustration in magnetic ordering. Bi-substitution in cobalt manganite spinel introduces structural distortion and modifies magnetic exchange interaction, which effects both ferroelectric and ferrimagnetic transition. The coexistence of ferroelectric and magnetic properties proposes the candidature of this material for the future multiferroic application. Acknowledgements Authors are thankful to Dr. Amit Roy, Director IUAC, New Delhi for his keen interest and encouragement to this work, and DST,

56

N.E. Rajeevan et al. / Materials Science and Engineering B 163 (2009) 48–56

India for the financial support under the project No. SR/S2/CMP0051/2007. One of the authors (N.E.R.) is thankful to IUAC, New Delhi, UGC, India, DCE and KSCSTE of State Govt. of Kerala and Z.G. College, Kerala, India for their support to carry out the research work. SKA and IVS are grateful to the Science Foundation of Ireland for the financial support under the project no. 06/IN.1/I91. References [1] B. Antic, G.F. Goya, H.R. Rechenberg, V. Kusigerski, N. Jovic, M. Mitric, J. Phys.: Condens. Matter 16 (2004) 651. [2] S. Kumar, Alimuddin, R. Kumar, A. Dogra, V.R. Reddy, A. Banerjee, J. Appl. Phys. 99 (2006) 08M910. [3] N. Fujiwara, H. Yasuoka, Y. Ueda, Phys. Rev. B 57 (1981) 3539. [4] J.L. Dormann, M. Nogues, J. Phys.: Condens. Matter 2 (1990) 1223. [5] G. Srinivasan, E.T. Rasmussen, R. Hayes, Phys. Rev. B 67 (2003) 014418. [6] E.J.W. Verwey, E.L. Heilmann, J. Chem. Phys. 15 (1947) 174. [7] Y. Ikedo, J. Sugiyama, H. Nozaki, H. Itahara, J.H. Brewer, E.J. Ansaldo, G.D. Morris, D. Andreica, A. Amato, Phys. Rev. B 75 (2007) 054424. [8] Y. Yamasaki, S. Miyasaka, Y. Kaneko, J.P. He, T. Arima, Y. Tokura, Phys. Rev. Lett. 96 (2006) 207204. [9] T. Kimura, T. Goto, H. Shintani, K. Ishizaka, T. Arima, Y. Tokura, Nature 426 (2003) 53. [10] T. Goto, T. Kimura, G. Lawes, A.P. Ramirez, Y. Tokura, Phys. Rev. Lett. 92 (2004) 257201. [11] D.K. Shukla, S. Mollah, R. Kumar, P. Thakur, K.H. Chae, W.K. Choi, A. Banerjee, J. Appl. Phys. 104 (2008) 033707. [12] W. Prellier, M.P. Singh, P. Murugavel, J. Phys.: Condens. Matter 17 (2005) R803. [13] A.K. Sianou, G.A. Stergioudis, K.G. Efhtimiadis, O. Kalogirou, I.A. Tsoukalas, J. Alloys Compd. 392 (2005) 310. [14] M.A. Arillo, M.L. Lopez, E. Perez-Cappe, C. Pico, M.L. Veiga, Solid State Ionics 107 (1998) 307.

[15] C.G. Coops, Phy. Rev. 83 (1951) 121. [16] R. Sarkar, S.K. Das, G. Banerjee, Ceram. Int. 25 (1999) 485. [17] J. Hemberger, P. Lunkenheimer, R. Fichtl, H.A. Krug von Nida, V. Tsurkan, A. Loidl, Nature 434 (2005) 364. [18] A.C. Tavares, M.A.M. Cartaxo, M.I. da Silva Pereira, F.M.A. Costa, J. Solid State Electrochem. 5 (2001) 57. [19] M.H. Mendonca, M.I. Godinho, M.A. Catarino, M.I. da Silva Pereira, F.M. Costa, Solid State Sci. 4 (2002) 175. [20] E.D. Macklen, J. Phys. Chem. Solids 47 (1986) 1073. [21] J.L. Martin deVidales, P. Garcia Chain, R.M. Rojas, E. Vila, O. Garefa-Martinez, J. Mater. Sci. 33 (1998) 1491. [22] N.E. Rajeevan, P.P. Pradyumnan, R. Kumar, D.K. Shukla, S. Kumar, A.K. Singh, S. Patnaik, S.K. Arora, I.V. Shevts, Appl. Phys. Lett. 92 (2008) 102910. [23] X. Liu, F. Gao, C. Tian, Mater. Res. Bull. 43 (2008) 693. [24] V. Massarotti, D. Capsoni, M. Bini, G. Chiodelli, J. Solid State Chem. 131 (1997) 94. [25] N.F. Mott, J. Non-Cryst. Solids 1 (1968) 1. [26] M. Viret, L. Ranno, J.M.D. Coey, Phys. Rev. B 55 (1997) 8067. [27] K. De, R. Ray, R.N. Panda, S. Giri, H. Nakamura, T. Kohara, J. Magn. Magn. Mater. 288 (2005) 339. [28] O. Raymond, R. Font, N. Suarez-Almodovar, J. Portelles, J.M. Siqueiros, J. Appl. Phys. 97 (2005) 084107. [29] G. Catalan, Appl. Phys. Lett. 88 (2006) 102902. [30] C.C. Wang, Y.M. Cui, Y.W. Zhang, Appl. Phys. Lett. 90 (2007) 012904. [31] K.S. Cole, R.H. Cole, J. Chem. Phys. 9 (1941) 341. [32] F.A. Grant, J. Appl. Phys. 29 (1957) 176. [33] A.H. Elsayed, A.M. Haffz, Egypt. J. Solids 28 (2005) 53. [34] S.R. Elliot, Adv. Phys. 36 (1987) 135. [35] J. Wang, A. Scholl, H. Zheng, S.B. Ogale, Viehland, R. Ramesh, Science 307 (2005) 1203b. [36] K. Ueda, H. Tabata, T. Kawai, Appl. Phys. Lett. 75 (1999) 555. [37] T. Suzuki, H. Nagai, M. Nohara, H. Takagi, J. Phys.: Condens. Matter 19 (2007) 145265. [38] R.N. Bhowmik, R. Ranganathan, R. Nagarajan, J. Magn. Magn. Mater. 299 (2006) 327.