Structural, transport, magnetic and magnetoelectric properties of CaMn_1−xFe_xO_3−δ (0.0 ≤ x ≤ 0.4)

Abstract

The structural, magnetic, transport and magnetoelectric properties of parent and Fe doped CaMn_1−xFe_xO_3−δ (0.0 ≤ x ≤ 0.4) manganites are investigated using synchrotron X-ray, Raman, SQUID and nova control impedance analyzer. The Fe doped composition x = 0.3 shows a strong Maxwell–Wagner effect, and quite high positive magnetocapacitance (MC) of ∼ 8.45% at room temperature and at low magnetic field 7.8 kG has been observed. Rietveld refinement of synchrotron X-ray diffraction patterns suggests (i) a structural transformation from orthorhombic to cubic crystal system and (ii) an increase in lattice parameters with the substitution of Fe at the Mn site. Sintering at 1300 °C stabilizes the doping of higher ionic radii Fe⁺³ (0.645 Å)/Fe⁺⁴ (0.585 Å) atoms at the Mn⁺⁴ (0.53 Å) site in CaMn_1−xFe_xO_3−δ. The Magnetization data show a transformation of the G type of antiferromagnetic arrangement of Mn⁺⁴ electrons spins in CaMnO₃ into a paramagnetic spin type arrangement with Fe substitution. The AC conductivity of the Fe doped compositions decreases more than two orders of magnitude in comparison to CaMnO_3−δ.

---

Introduction

A strong coupling between the magnetic and electronic degrees of freedom makes mixed valence (Mn⁺⁴/Mn⁺³) containing perovskite manganites to show various important properties like colossal magnetoresistance, metal to insulator transition, ferromagnetic to paramagnetic transition, and charge ordering phenomenon.^1,2 Magnetic exchange interactions can be controlled by the ratio of Mn⁺⁴ (0.53 Å)/Mn⁺³ (0.645 Å) mixed valence states, involving the e_g hopping of electrons, explained using a double exchange (DE) mechanism in perovskite manganites.³ It has been found that operative DE mechanism due to the redox interactions of Mn⁺³/Mn⁺⁴ facilitated by the redox pair Ru⁺⁴/Ru⁺⁵ in perovskite manganites.⁴ When a smaller Ca⁺² atom (1.34 Å) is doped at the Sr⁺² (1.44 Å) site in Sr_1−xCa_xMnO₃, the crystal structure changes from orthorhombic to tetragonal to cubic.⁵ Recently, a large magnetodielectric response around the magnetic phase transition was reported in perovskite containing Mn and Ni or Fe ions, GdMnO₃, TbMnO₃ and DyMnO₃.^6–9 A large increase in the dielectric constant was also reported in the phase separated Pr_0.67Ca_0.33MnO₃ at low temperatures.¹⁰ A dielectric resonance was observed in perovskite manganite La_2/3Ca_1/3MnO₃ at 0.5 T magnetic field in the paramagnetic regime just above the ferromagnetic temperature at 270 K.¹¹ Moreover, Rairigh et al. observed a colossal magnetocapacitance coinciding with a regime of phase separation between the magnetic metal and charge ordered insulating phase at low temperatures.¹² Several systems show a magnetocapacitance (MC) effect without the presence of spontaneous electric polarization and magnetization.^13,14 However, these MC effects are below 1% under a strong magnetic field of several tesla. The origin of such effects can be understood through the Maxwell–Wagner relaxation model developed for inhomogeneous semiconductors.¹⁵ According to Maxwell–Wagner effects, the presence of inhomogeneity or contact effects in materials can enhance the dielectric constant and yield dielectric relaxation in the absence of intrinsic dipolar relaxation.^16,17 It has been found that the Maxwell–Wagner effect can also yield a magnetocapacitance without multiferroicity and saturated magnetization, provided the material exhibits an intrinsic magnetoresistance.¹⁵ A recent theoretical prediction suggests the possible occurrence of a magnetocapacitance effect even in non magnetic composite media.¹⁸ Materials with magnetocapacitance properties, can have various potential applications.¹⁹ However, the challenge is to find a material which shows magnetocapacitance within the ideal frequency range ω < 1 MHz at room temperature for a magnetic field (H) < 1 Tesla. Perovskite manganite materials may play a major role and might be a potential candidate for applications in magnetic sensors, miniaturization of antenna, spin charge transducers and microwave tunable filters.^1,2 In this manuscript, we show the doping of Fe at the Mn site in CaMnO_3−δ up to 40%. Further, Fe doping induced structural transformation, magnetic, AC transport studies of parent and Fe doped manganites CaMn_1−xFe_xO_3−δ (0.0 ≤ x ≤ 0.4), and quite a high magnetocapacitance (MC) value at low magnetic field, low frequency and at room temperature is reported.

Experimental section

Polycrystalline samples of CaMn_1−xFe_xO_3−δ (0.0 ≤ x ≤ 0.4) were prepared using a high temperature solid state reaction route. A stoichiometric amount of CaCO₃, MnO₂ and Fe₂O₃ were thoroughly mixed in an agate mortar and calcined at 950 °C for 48 hours keeping the heating rate at 10 °C min⁻¹ and cooling rate at 5 °C min⁻¹. After calcination, the powders were pelletized into 10 mm diameter and sintered in an ambient atmosphere at 1300 °C for 24 hours keeping the heating rate at 10 °C min⁻¹ and cooling rate at 5 °C min⁻¹. We noticed the colour of the CaMnO₃ pellet was gray while the Fe doped composition pellets were black in colour. The synchrotron X-ray powder diffraction patterns were recorded at room temperature using synchrotron radiation at the Indian beamline, Photon Factory (PF), KEK, Japan. A synchrotron X-ray beam (wavelength – 1.09 Å) from the bending magnet of the PF storage ring was first focused horizontally with a focusing mirror and then monochromatized using a double-crystal monochromator. The beam was then further collimated with a set of beam-defining slits having horizontal openings of 2 mm and 0.2 mm in the vertical direction. The sample was mounted onto an 8-circle goniometer (Huber, Germany) at the focal point of the focusing mirror of the beamline. The sample was mounted horizontally and the scattered beam was collected by a single-channel scintillation detector mounted at a distance of 380 cm onto the 2 h arm of the goniometer. A slit of 1.5 mm (horizontal) by 0.25 mm (vertical) was mounted just before the detector to increase the signal-to-background ratio. Rietveld refinements were done using Fullprof software. Raman measurements on the pelletized samples were performed using a Raman microprobe instrument consisting of a Jobin-Yvon T64000 spectrometer equipped with a microscope which allows a spatial resolution on the sample of about 1 μm. The Raman signal was detected by a multichannel CCD detector. An excitation wavelength of 514.5 nm was used with 3.0 mW power to carry out the experiments. We recorded data for each sample for 5 minutes in the range of 90–900 cm⁻¹ at room temperature. To improve the resolution of the closely spaced peaks, high-resolution scans were recorded using the triple additive configuration, with a spectral resolution better than 1 cm⁻¹. Iodometric titration was performed on 10–30 mg of the samples to determine the mean valency of Mn and Fe. The sample was weighed in a round neck vessel. An excess of potassium iodide with distilled water was added. The solution was flushed with Ar to avoid air oxidation of the excess iodide. After adding HCl, the sealed vessel was heated to dissolve the sample. The sample was cooled and titrated with standardized sodium thiosulphate. A freshly prepared Starch solution was used as an indicator. This gave the mean valency of Mn and Fe and the amount of oxygen was then calculated. Magnetization of the samples was recorded using SQUID in the temperature range 10–300 K at 1 Tesla applied magnetic field. Magnetocapacitance (MC) measurements were recorded using a nova control-Impedance analyzer in the frequency range from 1 Hz to 40 MHz and using a DC electromagnet. Magnetocapacitance (MC%) and magnetoloss (ML%) calculated by their respective formulas; (C(H) − C(0)) × 100/C(0), (D(H) − D(0)) × 100/D(0), where C(H)/D(H) is the capacitance/dielectric loss in the presence of a DC magnetic field in kilo Gauss and C(0)/D(0) is the capacitance/dielectric loss in the absence of a magnetic field.

Results and discussion

Synchrotron X-ray diffraction (XRD) patterns for CaMn_1−xFe_xO_3−δ (0.0 ≤ x ≤ 0.4) are shown in Fig. 1. Parent composition x = 0.0 shows some extra peaks indicated by *, which are due to the oxygen nonstoichiometry in the lattice. We also found a grey colour of our parent CaMnO_3−δ sample, which is the signature of oxygen nonstoichiometry in a CaMnO₃ lattice.^20–23

Fig. 1 Synchrotron X-ray diffraction patterns for CaMn_1−xFe_xO_3−δ (0.0 ≤ x ≤ 0.4). * shows the extra peaks due to oxygen nonstoichiometry. Note the absence of extra peaks for the x = 0.2, 0.3 and 0.4 compositions.

Reller et al.²⁰ showed that in the ideal perovskite structure, the creation of one anion vacancy (removal of one oxygen) causes the reduction of two Mn⁺⁴ cations to two Mn⁺³ cations and simultaneously two MnO₆ octahedra are transformed into two MnO₅ square-pyramids. The number of oxygen vacancies, i.e., the value of δ in the stoichiometric formula CaMnO_3−δ represents, therefore, a direct measure for the ratios Mn⁺⁴/Mn⁺³ and MnO₆ octahedra/MnO₅ square-pyramids, respectively. We found the oxygen content was greater than 2.5 in the x = 0.0 composition after comparing with the XRD pattern of CaMnO_2.5.^20–23 Manganites are known to have mixed valences of Mn and in the presence of oxygen deficiency, the Mn⁺⁴ valence state reduces into Mn⁺³ for charge valence in lattice.²⁴ In the presence of the Jahn–Teller (JT) ion Mn⁺³, lattice distortion is expected in CaMnO_3−δ. In fact, the new peaks shown in the X-ray diffraction pattern by * clearly indicate the distortion of the lattice due to the Jahn–Teller ion Mn⁺³ and the presence of MnO₆ octahedra along with the MnO₅ square-pyramids.^20,21 We compared the XRD patterns with the all possible phases (CaO, MnO₂,Mn₂O₃, Mn₃O₄-JCPDS XRD patterns given in ESI†) that could be an impurity phase but we found no similarities to any of them. So, conclusively, the new peaks in the x = 0.0 composition were generated due to the presence of the Jahn–Teller ion Mn⁺³ and MnO₅ square-pyramids. Similarly, the Fe doped composition x = 0.1 also showed the presence of extra peaks in its diffraction patterns.²⁵ With Fe doping, these extra peaks are reduced for x = 0.1, which shows less distortion in the lattice and this might be due to the presence of the non JT Fe⁺³ (t_2g³ e_g²) valence state in the lattice. Further, we observed a shift in the peaks towards lower theta values with increasing Fe content, which shows an increase in the lattice constants with increasing Fe content. The increased lattice constant for x = 0.1 in comparison to the x = 0.0 composition and equal ionic radii of Mn⁺³ (0.645 Å) and Fe⁺³ (0.645 Å) suggest that the oxygen content in both compositions is >2.9. We found systematic changes in the diffraction patterns with increases in Fe doping, which infers a change in the structure. The compositions (x = 0.0, 0.1 and 0.2) show a number of super lattice reflections (R, M, X), which are characteristic peaks for the orthorhombic crystal structure. The peaks (210), (102) and (112) are similar to the X, M and R point distortions found in Mn and Zr based perovskite materials.^22,26,27 The super lattice reflections R, M and X in CaMn_1−xFe_xO₃ (x = 0.0, 0.1 and 0.2) originate due to the rotation of the MnO₆ octahedra in the orthorhombic perovskite manganites. Reflections due to the out of phase or (−) rotations of MnO₆ octahedra are called R-point distortions and in-phase or (+) rotations of the octahedra are termed M-point reflections. X-point distortions occur when the R-point and M-point distortions operate in unison. These super lattice reflections are similar to the Glazer tilt system (a⁺b⁻b⁻), which corresponds to the orthorhombic crystal system. On increasing the Fe substitution at the Mn site, the M-point reflections are no longer observed in CaMn_0.7Fe_0.3O₃. The XRD pattern of the x = 0.2 composition resembles the XRD patterns of manganite compounds with the Pnma space group.²⁸ The XRD pattern of composition x = 0.2 shows no extra peaks which means there is no oxygen deficiency in the lattice. While, for the x = 0.3 composition, we neither observed the (R, M, X) super lattice reflections nor any extra peaks. The XRD patterns resemble the diffraction patterns of Imma space group manganite compounds. The XRD peaks of composition x = 0.4 are regularly spaced without the extra peaks, which shows the higher symmetry of the crystal system. The absence of extra reflections in x = 0.2, 0.3 and 0.4 shows that there is no oxygen nonstoichiometry and only Mn⁺⁴/Fe⁺⁴ valence states present in the lattice. We did Rietveld analysis of these compositions and found that the x = 0.2 composition crystallizes in Pnma, x = 0.3 in Imma and x = 0.4 in the cubic space group Pm3̄m (Fig. 2 and Table 1). It is reported in the literature that if the tolerance factor value [t = (r_A + r_O)/√2(r_B + r_O)] is close to 1 then the perovskite structure will be cubic. If t is ≤0.95 then it will be orthorhombic. In this case, the tolerance factor was found to be 1 for the x = 0.4 composition only when Fe is present in the lattice as Fe⁺⁴. Fe⁺⁴ has an ionic radius of 0.585 Å and Fe⁺³ has an ionic radius of 0.645 Å. So, if Fe⁺³ is present in the lattice then the tolerance factor value does not support the cubic crystal structure for the x = 0.4 composition. The results of the synchrotron XRD patterns show a change in the crystal system on increasing the Fe content at the Mn site in CaMn_1−xFe_xO₃. These observations are similar to the phase transitions in Sr_1−xCa_xMnO₃ perovskites,^5,29 where the smaller atomic Ca atom was doped at the bigger atomic Sr site. Our reported results are different from the reported results in the literature by Liu et al. where they reported a Pnma orthorhombic crystal structure up to x = 0.35 doping of Fe.³⁰ The difference in the structural properties was found due to different (i) precursors used, (ii) the synthesizing method and (iii) the sequence of sintering temperatures. The crystal structure of perovskites is very sensitive to sintering temperatures. The calcination and sintering temperatures, quenching, heating and cooling rates, sintering times, sintering in oxidation/reduction/ambient atmospheres, are some of the factors by which the structure, oxygen stoichiometry, crystal structure and physical properties of perovskite can be controlled. To prepare the bilayer manganites (La, Sr/Ca)₃Mn₂O₇, which are a stack resembling spin valve devices due to the Mn–O–Mn network separated by insulating (La, Sr/Ca)₂O₂ spacer layers along with the c-axis, precursors have to be sintered at >1400 °C.³¹ If they are sintered at a lower temperature than this then layered manganite does not form. Briatico et al. purposefully prepared oxygen deficient CaMnO_3−δ (0 ≤ δ ≤ 0.34) to explore the electrical resistivity in oxygen deficient compounds.³² In our case, we sintered at a high temperature (1300 °C) to stabilize the rare valence state phase in the Fe doped CaMnO_3−δ compositions.

Fig. 2 Rietveld refinement results of synchrotron X-ray diffraction pattern for CaMn_0.6Fe_0.4O₃. The experimental data points are shown as circles, and the calculated fit and difference curve are shown as solid lines. Tick marks indicate the calculated reflection positions. Bragg R factor and R_F were found to be 6.1 and 10.5, respectively.

Table 1 Lattice parameters as determined by Rietveld refinement of synchrotron X-ray diffraction patterns

CaMn1−xFexO3−δ	x = 0.2	x = 0.3	x = 0.4
Space group	Pnma	Imma	Pm3̄m
a (Å)	5.30149(9)	5.3146(2)	3.7683(9)
b (Å)	7.49622(11)	7.5178(3)	3.7683(9)
c (Å)	5.2956(2)	5.2959(2)	3.7683(9)

---

Raman spectroscopy was efficiently applied for monitoring the structural changes and Jahn–Teller disorder in the manganites at microscopic levels. In the mixed-valence manganites, Jahn–Teller distortion is found due to the presence of one e_g electron of Mn⁺³ and the oxygen sub lattice remains highly distorted. Due to these effects, when the Raman spectra are compared to the undistorted structure’s spectra then the following effects arise; (i) broadening of some or all of the first-order Raman lines, (ii) activation in the Raman spectrum of otherwise forbidden phonon modes, and (iii) the appearance of additional broad Raman bands.^33–35 According to lattice dynamics calculations, orthorhombic perovskite manganite (space group Pnma/Pbnm) shows the allowed 24 Raman active modes while orthorhombic perovskite manganite (space group Imma) show12 Raman active modes. Since there are no Raman-active lattice vibrations predicted in the ideal cubic Pm3̄m perovskite, all phonon features in the spectra of cubic perovskite-like manganites originate from either coherent or incoherent lattice distortions.^33–35

Fig. 3 shows the unpolarized Raman spectra of CaMn_1−xFe_xO₃ (x = 0.0, 0.1, 0.2, 0.3 and 0.4) at room temperature in the spectral range 100 to 900 cm⁻¹. The Lorentzian fittings of the Raman spectra were done to find out the exact position of the peaks (Fig. 3). These peaks are tabulated in Tables II and III (ESI†). From a crystallographic point of view, stoichiometric Ca²⁺Mn⁴⁺O₃ has a crystal structure closest to that of a perfect crystal. As Mn⁴⁺ (t_2g³) is a non-Jahn–Teller (JT) ion, the six Mn–O bonds in the MnO₆ octahedra are almost equal and the real structure (with space group Pnma) is fixed mainly by the octahedral tilts. The Raman active modes in the orthorhombic Pnma structure originate from the cubic zone boundary points R_c, M_c and X_c. While, MnO₆ octahedra deforms due to the JT ion Mn³⁺ (t_2g³ e_g¹) in LaMnO₃ (four shorter and two longer Mn–O bonds exists).^34,35 CaMnO₃ and LaMnO₃ show the presence of mixed valence Mn³⁺/Mn⁴⁺ when prepared in open air at high temperatures.³⁶ The Raman spectra of CaMnO_3−δ show characteristic high intensity peaks due to the vibrations of both Mn⁺⁴–O₆ (463, 483 cm⁻¹) and Mn⁺³–O₆ (607 cm⁻¹) octahedra along with the other peaks (Fig. 3 and 4). This shows that JT distortion is present in the lattice due to the presence of Mn⁺³ along with Mn⁺⁴. These results have also been found by Wang et al. in CaMn₂O₄ where Mn remains in the Mn⁺³ valence state.³⁷

Fig. 3 Raman spectroscopy measurements for CaMn_1−xFe_xO_3−δ (0.0 ≤ x ≤ 0.4). Circles show the recorded data while the red line shows the whole fitting, green lines show the position of individual peaks. Images recorded by microscope show the samples surface while recording the data at room temperature.

Fig. 4 Combined results of synchrotron X-ray diffraction and Raman spectroscopic measurements for CaMn_1−xFe_xO_3−δ (0.0 ≤ x ≤ 0.4).

In CaMn_1−xFe_xO_3−δ, Raman peaks for x = 0.0–607 cm⁻¹ and x = 0.1–643 cm⁻¹ were observed due to the presence of Mn⁺³–O and Fe⁺³–O vibrations, while Raman peaks for x = 0.0–463, 483 cm⁻¹, and x = 0.1–444, 470 cm⁻¹ were observed due to the presence of Mn⁺⁴–O vibrations (Fig. 3 and 4). The compositions x = 0.2, x = 0.3 and x = 0.4 show Raman peaks due to Mn⁺⁴–O vibrations at 476 cm⁻¹, 474 cm⁻¹, 473 cm⁻¹ consequently (Fig. 3 and 4). This shows that the presence of the Mn⁺³ valence state is only in composition x = 0.0. The compositions x = 0.2, x = 0.3 and x = 0.4 do not show the presence of the Mn⁺³ valence state. The characteristic Mn⁺⁴–O vibrations at ∼460 and ∼480 cm⁻¹ are merged in compositions x = 0.2, x = 0.3 and x = 0.4, which might be due to the change in space group with increasing Fe content (Fig. 3 & 4). To find the presence of Fe–O vibrations in the CaMn_1−xFe_xO_3−δ (0.1 ≤ x ≤ 0.4) compositions, we compared the Raman spectra of these compositions with the Raman spectra of CaFeO₃ (CFO) and La_0.33Sr_0.67FeO₃ (LSFO).^32,38 The intense peak of the FeO₆ vibration found at 707 cm⁻¹ and 705 cm⁻¹ in the Raman spectra of CFO and LSFO implies that some JT-type distortion is present at room temperature in CFO, indicating the existence of a JT Fe⁺⁴ ionic state along with the non-JT Fe⁺³ valence state.³⁸ In this study, we found the presence of Fe–O vibration modes at 715 cm⁻¹ in the x = 0.1 composition, 708 cm⁻¹ in the x = 0.2 composition, 709 cm⁻¹ in the x = 0.3 composition and 706 cm⁻¹ in the x = 0.4 composition (Fig. 3 and 4). The Fe–O vibration modes show that there is a presence of the Fe⁺⁴–O vibration mode for x = 0.2, x = 0.3 and x = 0.4 while the x = 0.1 composition shows the vibrations of Fe⁺³–O₆ octahedra. These observations are similar to the results that have been found using synchrotron XRD patterns, where compositions x = 0.2, x = 0.3 and x = 0.4 do not show any oxygen deficiency in the lattice. Fig. 4 summarises the results of the synchrotron X-ray diffraction and Raman measurements. Further, we followed the procedure adopted by Vazquez et al. to obtained the mean valency of Mn and Fe in the perovskites.³⁹ Using iodometric titration, we found that the oxygen content was 2.92 in our parent composition CaMnO_3−δ which further increased to 2.96 in the x = 0.1 composition. The oxygen content was found to range from 2.98 to 3.01 in the x = 0.2, 0.3 and 0.4 Fe doped compositions. This shows that Fe is present in the +4 valence state in CaMn_1−xFe_xO₃.

Fig. 5 shows the magnetic moment vs. temperature (T) graph for CaMn_1−xFe_xO_3−δ (0.0 ≤ x ≤ 0.3). All samples were first cooled to 10 K then magnetization was recorded in a 1 Tesla magnetic field. Magnetization of CaMnO_3−δ follows the Curie–Weiss law χ = C/(T − θ_p) (where χ = M/H) between room temperature and 140 K. The Weiss constant θ_p was found to be negative (inset in Fig. 5), which indicates the presence of antiferromagnetic interactions in the parent composition CaMnO_3−δ. The magnetic transition temperature (Neel temperature T_N) from paramagnetic to antiferromagnetic was found to be 125 K in our non stoichiometric bulk CaMnO_3−δ. The magnetic transition was found to be more broad in our non stoichiometric bulk CaMnO_3−δ than it has been observed in stoichiometric CaMnO₃.³⁶ In stoichiometric CaMnO_3−δ, G-type antiferromagnetic order indicates antiferromagnetic domain spin canting effects. In zero field cooled conditions at low temperatures, antiferromagnetic domains will be oriented randomly. When the field is increased, those domains for which antiferromagnetic polarization is perpendicular to the field dominate the magnetic response. When temperature is raised, most of the domains reorientate and the decreasing antiferromagnetic order parameter will yield smaller static spins to be canted, thereby leading to a decrease in magnetization in stoichiometric CaMnO_3−δ.

Fig. 5 Magnetic properties of CaMn_1−xFe_xO_3−δ (0.0 ≤ x ≤ 0.4). Inset shows the inverse magnetization vs. T graphs.

In O-deficient CaMnO_3−δ at low temperatures, the magnetic transition was found to be broad and more suppressed even in a 1 Tesla applied field (Fig. 5).^32,36 This might be interpreted as the Mn⁺³–O–Mn⁺⁴ clusters having their net spin aligned with the overall antiferromagnetic polarization axis. Thus the spin system in the ground state is formed by the antiferromagnetic super exchange between Mn⁺⁴ ions and the antiferromagnetic interaction between Mn⁺³ and Mn⁺⁴ ions. In an external field, the preferred alignment of these clusters would be along the external field direction. Thus, the clusters hinder the field induced transverse reorientation of the antiferromagnetic domains. With increasing temperature, magnetization therefore manifests the transverse response that accompanies the antiferromagnetic order parameter decreases. Temperature destabilizes the G-type magnetic phase due to the weakening of the antiferromagnetic interaction between the Mn⁺³ and Mn⁺⁴ ions, which is caused by the enhancement of the double exchange and formation in the antiferromagnetic phase with chains of the ferromagnetically coupled Mn⁺³ and Mn⁺⁴ ions. The substitution of 10% Fe at the Mn site in CaMnO_3−δ causes melting of the antiferromagnetic state. These results are similar with the reported magnetic and transport properties of La_0.7Ca_0.3Mn_1−xFe_xO₃,⁴⁰ where it has been observed that ∼10% doping of Fe at the Mn site destroys the double exchange operating between the Mn⁺³ and Mn⁺⁴ ions via O. The compositions x = 0.1 and 0.2 of CaMn_1−xFe_xO_3−δ (0.0 ≤ x ≤ 0.4) show small kink at 100 K and 94 K in the magnetization data. Below ∼100 K, magnetization increases with decreasing temperature. The kink in magnetization data originated due to the competition between antiferromagnetic and paramagnetic states with the substitution of Fe at the Mn site in CaMnO_3−δ. For 30% Fe substitution at the Mn site, we did not observe any kink in magnetization and the moment continuously increased up to 20 K. We obtained the effective paramagnetic moment (P_eff) from the curie constant values using the Curie–Weiss law. We also calculated the P_eff(cal) value for the x = 0.2 and 0.3 compositions (4.20μ_B and 4.30μ_B) by a formula given by Taguchi et al.,⁴¹ both P_eff and P_eff(cal) values are close to the value obtained by Das et al.⁴² This shows that in our compositions (x = 0.2, 0.3 and 0.4), Fe and Mn are present in the +4 valence state. This result is similar to that which has been obtained by X-ray absorption spectroscopy (XAS) and magnetic measurement by Das et al. and synchrotron diffraction, chemical titration, tolerance factor analysis and Raman spectroscopic studies from our results.

Impedance measurements for CaMn_1−xFe_xO_3−δ (0.0 ≤ x ≤ 0.4) were carried out in the frequency range 10 Hz to 10 MHz. The impedance data of all these compounds were plotted and analyzed by frequency dependent AC conductivity. The frequency dependence of the conductivity for the CaMn_1−xFe_xO_3−δ (x = 0.0, 0.1, 0.2, 0.3 and 0.4) compositions at room temperature is shown in Fig. 6. The conductivity data of Parent CaMnO_3−δ show a decrease in conductivity with increasing frequency. This behavior is totally different from the observed behavior of conductivity for Fe doped samples. The Fe doped CaMn_1−xFe_xO_3−δ (x = 0.1, 0.2, 0.3 and 0.4) compositions show an increase in conductivity with increasing frequency. At low frequencies, the x = 0.3 composition shows a decrease in conductivity of more than two orders of magnitude while the other compositions x = 0.1, x = 0.2, x = 0.4 also show a decrease in conductivity of more than one order of magnitude in comparison to the x = 0.0 composition. This large decrease in conductivity of Fe doped compositions is mainly due to the absence of Mn⁺³ (t_2g³ e_g¹) in the lattice of the Fe doped compositions. In CaMnO_3−δ, it is known that the presence of double exchange interactions enhances the conductivity because of the hopping of one e_g electron of Mn⁺³ (t_2g³ e_g¹) in Mn⁺³–O–Mn⁺⁴ lattice.³² However, according to the band picture of Fe and Mn in perovskites proposed by Ahn et al., the bottom of the Mn e_g band is at the same level or higher than the top of the Fe e_g band.⁴³ In the Fe doped compositions x = 0.2, 0.3 and 0.4 containing Mn⁺⁴ (t_2g³ e_g⁰) and Fe⁺⁴ (t_2g³ e_g¹) valence states, the Mn e_g band is empty and the e_g band of Fe is less than half filled so hopping of an electron between the Fe e_g band and Mn e_g band is energetically forbidden. The lack of hopping of the e_g electron results in a decrease in conductivity with Fe doping at the Mn site in CaMnO₃.

Fig. 6 AC transport properties of CaMn_1−xFe_xO_3−δ (0.0 ≤ x ≤ 0.4) samples at room temperature.

Fig. 7 & 8 present the frequency dependent magnetocapacitance (MC) and magnetoloss (ML) behavior at room temperature at selected magnetic fields from 1 kG to 7.8 kG of the CaMn_1−xFe_xO_3−δ (x = 0.0 and 0.3) compositions. The Parent composition in Fig. 7 shows the maximum negative magnetocapacitance value of −2% for a low magnetic field of 5 kG at low frequency values. At high frequencies, the MC values approach zero. Fig. 8 shows the MC and ML behavior of the x = 0.3 composition. The composition x = 0.3 shows positive MC values and it has been found a maximum of ∼8.45% for a 7.8 kG magnetic field at low frequency at room temperature. This type of MC property can be explained by the Maxwell–Wagner (M–W) capacitor model, where two leaky capacitors are in series and one of the leakage components is magnetically tuneable. In this case, the oxygen deficient Mn⁺⁴–O–Mn⁺⁴ lattice is one component and the Fe doped oxygen deficient lattice is another component. It has been shown in the literature that the oxygen deficient lattice of CaMnO_3−δ can be tuned with the use of a magnetic field and Fe doping blocks the double exchange mechanism which operates in manganites.^14,32,36 Lattice resistance increases with Fe doping in the Mn⁺⁴–O–Mn⁺⁴ lattice and it acts as a resistive layer. An important feature to note is that at low frequency the maxima of the magnetocapacitance and magnetoloss curves do not exactly coincide with each other which would be expected in the case where magnetocapacitance is completely driven by extrinsic effects. This shows that the MC of ∼8.45% at room temperature and at the low magnetic field of 7.8 kG is due to intrinsic effects. The magnetocapacitance in the x = 0.3 composition is due to the local magnetoresistance in the oxygen deficient Mn⁺⁴–O–Mn⁺⁴ lattice and the presence of Fe in the lattice. In the presence of Fe, the whole sample does not show magnetoresistance but local magnetoresistance might originate in the lattice which has induced magnetocapacitance. This behavior of the MC data is a basic illustration of the Maxwell–Wagner effect, where Fe in the Mn⁺⁴–O–Mn⁺⁴ lattice mixes the longitudinal, transverse, real and imagined modes of the response to create an unexpected rich behavior, i.e. dielectric relaxation and dielectric resonance.

Fig. 7 Magnetocapacitance (MC) and magnetoloss (ML) properties of CaMnO_3−δ.

Fig. 8 Magnetocapacitance (MC) and magnetoloss (ML) properties of CaMn_0.7Fe_0.3O₃. Note the ∼8.5% MC at room temperature at a low magnetic field of 7.8 kG.

Conclusions

In conclusion, we have demonstrated that Fe doping at the Mn site in perovskite CaMnO_3−δ brings novel effects; (i) a change in the space group, (ii) mixed valences of Fe⁺⁴ and Mn⁺⁴ for the x = 0.2, 0.3 and 0.4 compositions, (iii) antiferromagnetic magnetic ordering turning into paramagnetic, (iii) a large decrease in AC conductivity at room temperature and (iv) quite a high positive magnetocapacitance at room temperature at low magnetic field and at low frequencies. Transport studies show an insulator type of behavior and a conductivity value was observed more than two orders of magnitude less in x = 0.3 compared to x = 0.0. A large magnetocapacitance (MC ∼ 8.45%) at room temperature (300 K) at low magnetic field (7.8 kG) was found in CaMn_0.7Fe_0.3O_3−δ. A large MC value shows the strong magnetoelectric effects in the x = 0.3 composition are due to the presence of Fe in the Mn⁺⁴–O–Mn⁺⁴ lattice through a combination of local core magnetoresistance and the Maxwell–Wagner effect.

Acknowledgements

B. Singh thanks the Department of Science and Technology, India for financial support and the Saha Institute of Nuclear Physics, India, for facilitating the experiments at the Indian Beamline, Photon Factory, KEK, Japan. B. Singh also thanks Prof. S. S. Manoharan for magnetic and Prof. Rajeev Gupta for providing a Raman facility to collect Raman data in IIT Kanpur, India.

Notes and references

1. S. Jin, T. H. Tiefel, M. McCormack, R. A. Fastnacht, R. Ramesh and L. H. Chen, Science, 1994, 264, 413.
2. Colossal Magnetoresistance, Charge ordering, and related properties of manganese oxides, ed C. N. R. Rao and B. Raveau, World Scientific, Singapore, 1998.
3. C. Zener, Phys. Rev., 1951, 82, 403.
4. B. Singh, S. S. Manoharan, M. L. Rao and S. P. Pai, Phys. Chem. Chem. Phys., 2004, 6, 4199.
5. Q. Zhou and B. J. Kennedy, J. Solid State Chem., 2006, 179, 3568.
6. N. Hur, S. Park, P. A. Sharma, J. S. Ahn, S. Guha and S.-W. Cheong, Nature, 2004, 429, 392.
7. N. S. Rogado, J. Li, A. W. Sleight and M. A. Subramanian, Adv. Mater., 2005, 17, 2225.
8. P. Padhan, H. Z. Guo, P. LeClair and A. Gupta, Appl. Phys. Lett., 2008, 92, 022909.
9. A. K. Kundu, R. Ranjith, B. Kundys, N. Nguyen, V. Caignaért, V. Pralong, W. Prellier and B. Raveau, Appl. Phys. Lett., 2008, 93, 052906.
10. S. Mercone, A. Wahl, A. Pautrat, M. Pollet and C. Simon, Phys. Rev. B: Condens. Matter Mater. Phys., 2004, 69, 174433.
11. J. Rivas, J. Mira, B. Rivas-Murias, A. Fondado, J. Dec, W. Kleemann and M. A. Señaris-Rodríguez, Appl. Phys. Lett., 2006, 88, 242906.
12. R. P. Rairigh, G. S. Bhalla, S. Tongay, T. Dhakal, A. Biswas and A. F. Hebard, Nat. Phys., 2007, 3, 551.
13. G. Lawes, A. P. Ramirez, C. M. Varma and M. A. Subramanian, Phys. Rev. Lett., 2003, 91, 257208.
14. R. Tackett, G. Lawes, B. C. Melot, M. Grossman, E. S. Toberer and R. Seshadri, Phys. Rev. B: Condens. Matter Mater. Phys., 2007, 76, 024409.
15. G. Catalan, Appl. Phys. Lett., 2006, 88, 102902.
16. J. C. Maxwell, Treatise on Electricity and Magnetism, Dover, New York, 1991.
17. R. J. Wagner, Ann. Phys., 1913, 40, 817.
18. M. M. Parish and P. B. Littlewood, Phys. Rev. Lett., 2008, 101, 166602.
19. Y. S. Shin and S. O. Park, Microwave and Optical Technology Letters, 2010, 52, 2364.
20. A. Reller, J. M. Thomas and D. A. Jefferson, Proc. – R. Soc. Edinburgh, Sect. A: Math., 1984, 394, 223.
21. K. Vijayanandhini and T. R. N. Kutty, J. Mater. Sci.: Mater. Electron., 2009, 20, 445.
22. C. C. K. Chiang and K. R. Poeppelmeier, Mater. Lett., 1991, 12, 102.
23. B. Singh and S. S. Manoharan, Mater. Lett., 2011, 65, 2029.
24. M. N. Iliev, M. V. Abrashev, V. N. Popov and V. G. Hadjiev, Phys. Rev. B: Condens. Matter Mater. Phys., 2003, 67, 212301.
25. M. Hernando, L. Miranda, A. Varela, K. Boulahya, S. Lazar, D. C. Siclair, J. M. G. Calbet and M. Parras, Chem. Mater., 2013, 25, 548.
26. Q. Zhou and B. J. Kennedy, J. Phys. Chem. Solids, 2006, 67, 1595.
27. C. J. Howard, K. S. Knight, B. J. Kennedy and E. H. Kisi, J. Phys.: Condens. Matter, 2000, 12, L677.
28. C. C. K. Chiang and K. R. Poeppelmeier, Mater. Lett., 1991, 12, 102.
29. K. Vjayanandhini and T. R. N. Kutty, J. Mater. Sci.: Mater. Electron., 2009, 20, 445.
30. X. J. Liu, Z. Q. Li, P. Wu and E. Y. Jiang, Solid State Commun., 2007, 142, 525.
31. I. Guedes, J. F. Mitchell, D. Argyriou and M. Grimsditch, J. Raman Spectrosc., 2000, 31, 1013.
32. J. Briatico, B. Alascio, R. Allub, A. Butera, A. Caneiro, M. T. Causa and M. Tovar, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 53, 14020.
33. V. Dediu, C. Ferdeghini, F. C. Matacotta, P. Nozar and G. Ruani, Phys. Rev. Lett., 2000, 84, 4489.
34. M. V. Abrashev, J. Backstrom, L. Borjesson, V. N. Popov, R. A. Chakalov, N. Kolev, R.-L. Meng and M. N. Iliev, Phys. Rev. B: Condens. Matter Mater. Phys., 2002, 65, 184301.
35. J. Xu, J. H. Park and H. M. Jang, Phys. Rev. B: Condens. Matter Mater. Phys., 2007, 75, 012409.
36. Z. Zeng, M. Greenblatt and M. Croft, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 59, 8784.
37. Z. Wang, S. K. Saxena and J. J. Neumeier, J. Solid State Chem., 2003, 170, 382.
38. S. Ghosh, N. Kamaraju, M. Seto, A. Fujimori, Y. Takeda, S. Ishiwata, S. Kawasaki, M. Azuma, M. Takano and A. K. Sood, Phys. Rev. B: Condens. Matter Mater. Phys., 2005, 71, 245110.
39. C. V. Vázquez, M. C. Blanco, M. A. L. Quintela, R. D. Sánchez, J. Rivas and S. B. Oseroff, J. Mater. Chem., 1998, 8, 991.
40. J. R. Sun, G. H. Rao, B. G. Shen and H. K. Wong, Appl. Phys. Lett., 1998, 73, 2998.
41. H. Taguchi, Phys. B, 2002, 311, 298.
42. N. Das, I. A. Dhiman, A. K. Nigam, A. K. Yadav, D. Bhattacharyya and S. S. Meena, J. Appl. Phys., 2012, 112, 123913.
43. K. H. Ahn, X. W. Wu, K. Liu and C. L. Chein, J. Appl. Phys., 1997, 81, 5505.

---

Footnote

† Electronic supplementary information (ESI) available: JCPDS-XRD data and tables of Raman spectroscopic data. See DOI: 10.1039/c5ra03565a

---